Retinol binding protein 4 stimulates hepatic sterol regulatory element-binding protein 1 and increases lipogenesis through the peroxisome proliferator-activated receptor-γ coactivator 1β-dependent pathway


  • Min Xia,

    Corresponding author
    1. Guangdong Provincial Key Laboratory of Food, Nutrition and Health, Guangzhou, Guangdong Province, China
    2. Department of Nutrition, School of Public Health, Sun Yat-sen University (Northern Campus), Guangzhou, Guangdong Province, China
    • Address reprint requests to: Wenhua Ling or Min Xia, Guangdong Provincial Key Laboratory of Food, Nutrition and Health, Department of Nutrition, School of Public Health, Sun Yat-sen University (Northern Campus), Guangzhou 510080, Guangdong Province, China. E-mail: fax: +86-20-87330446

    Search for more papers by this author
  • Yan Liu,

    1. Guangdong Provincial Key Laboratory of Food, Nutrition and Health, Guangzhou, Guangdong Province, China
    2. Department of Nutrition, School of Public Health, Sun Yat-sen University (Northern Campus), Guangzhou, Guangdong Province, China
    Search for more papers by this author
  • Honghui Guo,

    1. Department of Nutrition, School of Public Health, Sun Yat-sen University (Northern Campus), Guangzhou, Guangdong Province, China
    Search for more papers by this author
  • Duan Wang,

    1. Guangdong Provincial Key Laboratory of Food, Nutrition and Health, Guangzhou, Guangdong Province, China
    2. Department of Nutrition, School of Public Health, Sun Yat-sen University (Northern Campus), Guangzhou, Guangdong Province, China
    Search for more papers by this author
  • Yun Wang,

    1. Department of Nutrition, School of Public Health, Sun Yat-sen University (Northern Campus), Guangzhou, Guangdong Province, China
    Search for more papers by this author
  • Wenhua Ling

    Corresponding author
    1. Guangdong Provincial Key Laboratory of Food, Nutrition and Health, Guangzhou, Guangdong Province, China
    2. Department of Nutrition, School of Public Health, Sun Yat-sen University (Northern Campus), Guangzhou, Guangdong Province, China
    • Address reprint requests to: Wenhua Ling or Min Xia, Guangdong Provincial Key Laboratory of Food, Nutrition and Health, Department of Nutrition, School of Public Health, Sun Yat-sen University (Northern Campus), Guangzhou 510080, Guangdong Province, China. E-mail: fax: +86-20-87330446

    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

  • Supported by grants the National Basic Research Program (973, 2012CB517506).

  • See Editorial on Page 477


Recent studies have revealed the essential role of retinol binding protein 4 (RBP4) in insulin resistance. However, the impact of RBP4 on aberrant lipogenesis, the common hepatic manifestation in insulin resistance states, and the underlying mechanism remain elusive. The present study was designed to examine the effect of RBP4 on sterol regulatory element-binding protein (SREBP-1) and hepatic lipogenesis. Treatment with human retinol-bound RBP4 (holo-RBP4) significantly induced intracellular triglyceride (TAG) synthesis in HepG2 cells and this effect is retinol-independent. Furthermore, RBP4 treatment enhanced the levels of mature SREBP-1 and its nuclear translocation, thereby increasing the expression of lipogenic genes, including fatty acid synthase (FAS), acetyl coenzyme A carboxylase-1 (ACC-1), and diacylglycerol O-acyltransferase 2 (DGAT-2). Stimulation of HepG2 cells with RBP4 strongly up-regulated the expression of transcriptional coactivator peroxisome proliferator-activated receptor-γ coactivator 1β (PGC-1β) at both the messenger RNA (mRNA) and protein levels. The transcriptional activation of PGC-1β is necessary and sufficient for the transcriptional activation of SREBP-1 in response to RBP4. The cyclic adenosine monophosphate (cAMP)-response element binding protein (CREB) was identified as the target transcription factor involved in the RBP4-mediated up-regulation of PGC-1β transcription as a result of phosphorylation on Ser133. Furthermore, in vivo RBP4 infusion induced SREBP-1c activation and consequently accelerated hepatic lipogenesis and plasma TAG in C57BL/6J mice, a phenomenon not observed in Ppargc1b knockout mice. Conclusion: These findings reveal a novel mechanism by which RBP4 achieves its effects on hepatic lipid metabolism. (HEPATOLOGY 2013;8:564-575)


acetyl coenzyme A carboxylase-1


cyclic adenosine monophosphate


cAMP response element-binding protein


diacylglycerol O-acyltransferase 2


fatty acid synthase


gluconeogenic enzyme phosphoenolpyruvate carboxykinase


peroxisome proliferator-activated receptor-γ coactivator 1β


peroxisome proliferator-activated receptor gamma


retinol binding protein 4


reverse-transcription polymerase chain reaction


sterol regulatory element-binding protein




very low density lipoprotein

Retinol binding protein 4 (RBP4), a protein that belongs to the lipocalin family, was initially known as a specific carrier for the delivery of retinol (vitamin A) in the circulation.[1, 2] It is encoded by the RBP4 gene, localized in chromosome 10q23-q24.[3] Hepatocytes are regarded as the major source of RBP4 secretion under normal conditions[4]; however, adipose tissue expresses a considerable amount of RBP4[5] and could make a substantial contribution to elevated serum RBP4 levels in insulin-resistant states.[6] RBP4 is identified as a new adipokine suggested to link obesity with its comorbidities, especially insulin resistance, type 2 diabetes (T2D), and certain components of the metabolic syndrome. Serum RBP4 concentrations are elevated in subjects with impaired glucose tolerance, T2D, and correlate inversely with insulin sensitivity in nondiabetic subjects with a family history of T2D.[7] Circulating RBP4 levels correlate with the degree of insulin resistance in these subjects and the relationship is independent of obesity.[8] Transgenic overexpression of human RBP4 or injection of recombinant RBP4 in normal mice induced insulin resistance.[10] Conversely, heterozygous or homozygous RBP4 knockout mice had improved insulin sensitivity.[10] Increased serum RBP4 decreased glucose transporter type 4 (Glut4) expression in adipocytes and induced expression of the gluconeogenic enzyme phosphoenolpyruvate carboxykinase (PEPCK) in hepatocytes, thus contributing to the impairment of systemic insulin resistance.[10] Recently, a high circulating level of RBP4 was demonstrated to associate with elevated liver fat accumulation in human studies.[11] Blocking RBP4 expression in the liver was sufficient to reduce lipid droplets and ameliorate high fat diet-induced hepatic steatosis in C57BL/6 mice, which confirmed the potential role of RBP4 in the regulation of lipid metabolism in liver.[14] However, the pathophysiological roles of RBP4 involved in the regulation of hepatic lipid metabolism and the underlying molecular mechanism has not yet been fully characterized.

Sterol regulatory element binding protein (SREBP) is known as a key lipogenic transcription factor controlling the biosynthesis of cholesterol, fatty acids, and triglyceride (TAG).[15, 16] Mammalian genomes have two separate SREBP genes (SREBF1 and SREBF2). SREBP-1 expression produces two different isoforms, SREBP-1a and -1c. These isoforms differ in their first exons due to the use of different transcriptional start sites for the SREBP-1 gene. SREBP-1c was also identified in rats as ADD-1. SREBP-1a and SREBP-1c preferentially stimulate the lipogenic process by activating genes involved in fatty acid and TAG synthesis. SREBP-2 encodes SREBP-2, which mainly controls cholesterol homeostasis by inducing genes required for cholesterol synthesis and uptake.[17, 18] The precursor of these three isoforms of SREBP (P-SREBP) is synthesized as an endoplasmic reticulum (ER) membrane-bound protein, which is transported from the ER to the Golgi and undergoes proteolytic processing to release the transcriptionally active nuclear form. The nuclear form of SREBP (N-SREBP) is translocated into the nucleus, where it binds to sterol regulatory elements (SREs) present in the promoters of target genes and activates the transcription of SREBP-responsive genes involved in lipogenic pathways, such as fatty acid synthase (FAS), acetyl coenzyme A carboxylase-1 (ACC-1), and diacylglycerol O-acyltransferase 2 (DGAT-2).[19] Thus, the dysregulation of SREBP-1 contributes significantly to the pathogenic hepatic biosynthesis of fatty acids and the metabolism of TAG.[22]

In this study we examined the effects of RBP4 on SREBP-1 activation and lipogenesis in vitro and in vivo. Our data reveal that RBP4 activates SREBP-1 through a peroxisome proliferator-activated receptor-γ coactivator 1β (PGC1β)-dependent pathway in HepG2 cells and contributes to increased hepatic lipogenesis in mice. Our findings highlight RBP4 as a potential target for therapeutic intervention in metabolic syndrome-related lipid disorders.

Materials and Methods

Reagents and Antibodies

Human retinol-bound RBP4 (holo-RBP4, NP_006735.2, Met 1-Leu 201) expressed in HEK293 cells with a C-terminal polyhistidine tag was obtained from Sino Biological (Beijing, China). The endotoxin content was below 1.0 EU per μg of the protein as determined by the Limulus amoebocyte assay. The recombinant human RBP4 consists of 194 amino acids after removal of the signal peptide and migrates as an ∼23 kDa protein as predicted. Detailed other reagents and antibodies used in this study are provided in the Supporting Materials and Methods.

Cell Culture

HepG2 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and antibiotics at 37°C in a humidified, 5% CO2 / 95% air atmosphere. The cells were incubated with human recombinant RBP4 in serum-free medium for the indicated time periods.

Small Interfering RNA (siRNA).

HepG2 cells were transfected at 40%-60% confluency with 100 nM Ppargc1b SMARTpool siRNA or siCONTROL nontargeting siRNA (Dharmacon, Lafayette, CO). The efficiency of transfection (>70%-80%; data not shown) was determined using siGLO RISC-free nontargeting siRNA (Dharmacon). The effectiveness of the siRNA treatment was assessed by measuring PGC-1β protein level by immunoblotting.

Animal Studies

Adult male C57BL/6 mice (Jackson Laboratory) and Ppargc1b−/− knockout (PPARGC1B−/−) mice on a C57BL/6 background were used for in vivo studies. Investigations were conducted in accordance with Guide for the Care and Use of Laboratory Animals prepared by Sun Yat-sen University. The detailed experiment protocol is provided in the Supporting Materials and Methods.

Statistical Analysis

Values are expressed as mean ± standard error of the mean (SEM). Statistical significance was evaluated using the unpaired two-tailed t test and among more than two groups by one-way analysis of variance (ANOVA). Differences were considered significant at P < 0.05.


RBP4 Stimulates De Novo Lipogenesis in HepG2 Cells

Cultured human HepG2 cells were treated with different doses of human recombinant retinol bound RBP4 (holo-RBP4) for 24 hours. The incubation of HepG2 cells with RBP4 resulted in a dose-dependent increase in intracellular de novo lipogenesis, as measured by [3H]-acetate incorporation into the lipid fraction (Fig. 1A). As a result, the cellular accumulation of TAG was increased 1.29-fold, 1.71-fold, and 2.19-fold compared to control, respectively, as determined by direct mass measurements (Fig. 1B) and Oil red O staining (Supporting Fig. S1A). In addition, TAG synthesis from [3H]-palmitate (Fig. S1B) and fatty acid oxidation (Fig. S1C) did not differ between control and RBP4-treated HepG2 cells, which suggests that TAG accumulation was due to enhanced fatty acid synthesis. The amount of RBP4 is not toxic to HepG2 cells as measured by Trypan blue staining (Fig. S1D).

Figure 1.

RBP4 promotes lipogenesis and lipid accumulation in hepatocytes HepG2 cells, maintained in DMEM containing 1% BSA, were treated with RBP4 (20, 40, or 80 μg/mL) or vehicle control for 24 hours. (A) Lipogenesis (from 3H-acetate) was evaluated. (B) TAG mass was measured as indicated and expressed as μg of lipid/mg of protein. (C,D) Primary rodent hepatocytes isolated from fresh mouse livers were incubated with holo-RBP4 (80 μg/mL) for 24 hours. (C) Lipogenesis and (D) TAG mass were measured as described above. *P < 0.05, **P < 0.01 versus control.

This finding was further confirmed in rodent primary hepatocytes. We treated primary mouse hepatocytes with human RBP4 at the 80 μg/mL dose for 24 hours. In accordance with results from HepG2 cells, we observed a 78% increase of RBP4 on lipogenesis (Fig. 1C) and 63% TG content (Fig. 1D) in RBP4-stimulated cells. Although the magnitude of the stimulatory effects of RBP4 in primary hepatocytes was not as large as with HepG2 cells, this was expected because primary hepatocytes are not as metabolically active as cultured HepG2 cells.

Since retinol has been demonstrated to possess many roles in regulating cellular function,[23, 24] whether the effect of RBP4 on lipogenesis is retinol-dependent needs to be determined. We found that retinol-free RBP4 (apo-RBP4) exerted the same effects on the de novo lipogenesis (Fig. S2A) and TAG accumulation (Fig. S2B) with holo-RBP4 in HepG2 cells, excluding the possibility that retinol participated in this process. We thus conducted the experiments using human holo-RBP4 throughout the study unless specified otherwise.

RBP4 Induces SREBP-1 Activation

SREBP-1 is the major isoform of SREBPs that primarily controls lipogenesis in hepatocytes.[22, 25] To test the hypothesis that RBP4-induced lipogenesis might be due to the induction of SREBP, we quantified the precursor and nuclear active forms of SREBP-1 and SREBP-2 by immunoblotting. Treatment of HepG2 cells with RBP4 produced a marked increase in the mature nuclear form of SREBP-1 (nSREBP-1) and a corresponding decrease in the levels of precursor SREBP-1 (Fig. 2A). Because SREBP-1 activity is thought to depend on its subcellular localization,[26] the effect of RBP4 on the SREBP-1 subcellular distribution was determined. As shown in Fig. 2B, RBP4 treatment markedly decreased cytosolic SREBP-1 but elevated nuclear SREBP-1 levels. We further determined the messenger RNA (mRNA) amounts of SREBP-1a, SREBP-1c, and SREBP-2 by real-time polymerase chain reaction (PCR). Moreover, the mRNA levels of SREBP-1c but not SREBP-1a were significantly increased in a dose-dependent manner in response to RBP4 treatment (Fig. 2C), despite the fact that the SREBP-1c to SREBP-1a ratio (1:2) of human HepG2 cells was much less than that of mouse hepatocytes (9:1).[27] However, RBP4 did not significantly affect the degree of SREBP-2 nuclear form (Fig. S3A) and its mRNA expression (Fig. S3B).

Figure 2.

RBP4 increases SREBP-1 activation in hepatocytes HepG2 cells were treated with RBP4 (20, 40, or 80 μg/mL) or vehicle control for 24 hours. (A) Total cellular protein was extracted for immunoblotting to detect the nuclear active form of SREBP-1 (nSREBP-1). β-Actin was used as the loading control. Data from three independent experiments are shown. P and N denote the precursor (∼125 kDa) and cleaved nuclear (∼68 kDa) forms of SREBP-1. (B) Enhanced nuclear translocation of SREBP-1 in response to RBP4 in HepG2 cells. Immunoblot analysis of SREBP-1 in cytoplasmic and nuclear extracts is shown. (C) HepG2 cells were treated with RBP4 at the indicated concentrations for 24 hours. Total RNA was extracted and used to determine the relative mRNA levels of SREBP-1a and SREBP-1c.

RBP4 Induces SREBP1c-Dependent Lipogenic Gene Transcription

To determine the functional effects of increased nuclear SREBP-1 translocation by RBP4, the gene expression of key target enzymes of SREBP-1 in the HepG2 cells was evaluated by quantitative reverse-transcription (RT)-PCR. As expected in the case of dynamically altered nuclear SREBP-1c, RBP4 dose-dependently increased the expression of endogenous lipogenic genes, including FAS, ACC1, and DGAT2, involved in fatty acid and TAG synthesis in HepG2 cells. Similar to the lack of an effect on nuclear SREBP-2, the expression of mRNAs encoding two key enzymes of cholesterol biosynthesis, 3′-hydroxylmethyl glutaryl coenzyme A reductase (HMGCR) and low-density lipoprotein receptor (LDLR), was not altered in RBP4-treated HepG2 cells (Fig. 3A). In contrast, RBP4 exerted less effect on the nuclear SREBP-2-mediated transcriptional activation of SRE-containing target genes, including the 4×SRE-Lucand LDLR-Luc reporter genes (Fig. 3B).

Figure 3.

RBP4 induces the expression of SREBP-1 target genes. (A) HepG2 cells were treated with RBP4 (20, 40, or 80 μg/mL) or vehicle control for 24 hours. The transcription of genes involved in TAG and cholesterol biosynthesis, including FAS, SCD1, HMGCR, and LDLR, was determined by real-time RT-PCR. After normalization to GAPDH, relative mRNA levels were expressed as the fold induction of control which was defined as 1. *P < 0.05; **P < 0.01 versus control. (B) HepG2 cells were cotransfected with pcDNA empty vector or the vector encoding myc-tagged nuclear active SREBP-2 (nSREBP-2), together with either 4×SRE-Luc or LDLR-Luc reporter constructs, as indicated. Forty-eight hours posttransfection, the cells were treated with RBP4 (80 μg/mL) for an additional 16 hours. The cells were then harvested and analyzed to measure the level of luciferase activity. Results represent the means ± SEM from at least three independent experiments. *P < 0.05 versus pcDNA. (C) HepG2 cells were cotransfected with empty plasmid pGL3, luciferase reporter plasmids containing WT human SREBP-1c promoters, or the mutant reporter with disrupted LXRE or SRE, together with the Renilla luciferase reporter plasmid pRL-SV40. Forty-eight hours posttransfection, the cells were treated with or without RBP4 (80 μg/mL) in serum-free DMEM for 16 hours. Results represent the means ± SEM from at least three independent experiments. *P < 0.05; **P < 0.01 versus untreated cells expressing the corresponding promoter.

We next mapped the human SREBP-1c promoter and identified the element responsible for RBP4 action. Transcriptional activation of the wildtype SREBP-1c promoter was markedly induced by RBP4 in HepG2 cells. Disruption of the LXRE and SRE motif in the same promoter diminished the level of basal transcription and prevented the further induction caused by RBP4 (Fig. 3C). These data show that the LXRE and SRE motifs are necessary for the RBP4-dependent induction of SREBP-1c transcription.

RBP4 Up-Regulates PGC-1β Expression

PGC-1β is a recently identified transcriptional coactivator closely related to lipid metabolism.[28] To evaluate the effects of RBP4 on PGC-1β expression, we exposed HepG2 cells to recombinant RBP4 for different time periods and examined the effects on PGC-1β expression. RBP4 treatment was found to cause a time-dependent increase in the expression of Ppargc1b mRNA as determined by northern blot (Fig. 4A) and quantitative RT-PCR (Fig. 4B). The levels of Ppargc1b mRNA, over the control at 8 hours, increased as early as 2 hours after the addition of RBP4 to the cells, increased as early as 2 hours after RBP4 treatment. Moreover, the PGC-1β transcript levels remained high throughout the 24-hour treatment period. The effects of RBP4 were further found to be dose-dependent (Fig. 4C). Northern blot analysis revealed a 50% increase in the Ppargc1b mRNA levels in cells treated with RBP4 (20 μg/mL) and a maximal four-fold increase over the control was seen with a concentration of 80 μg/mL. Increases of similar magnitudes in the PGC-1β mRNA levels could be confirmed by quantitative RT-PCR (Fig. 4D). RBP4 treatment also induces PGC-1β protein expression in a concentration-dependent fashion (Fig. 4E).

Figure 4.

RBP4 up-regulates PGC1β expression. (A,B) Time-dependent induction of PGC1β mRNA expression. Cultured HepG2 cells were incubated with RBP4 (40 μg/mL) for the indicated times. PGC1β expression was then evaluated by (A) northern blot and (B) quantitative real-time PCR. The results shown are representative of three independent experiments. The abundance of PGC1β mRNA in untreated control cells was defined as 1 and the amounts of PGC1β mRNA from RBP4-treated cells were then expressed as fold changes in this control value. *P < 0.05 or **P < 0.01 versus 0 hours. (C,D) Dose-dependent induction of PGC1β mRNA expression. HepG2 cells were treated with RBP4 for 24 hours at the indicated concentrations and total RNA was isolated for the analysis of PGC1β and GAPDH mRNA expression by (C) northern blot and (D) quantitative real-time PCR. *P < 0.05 or **P < 0.01 versus 0 μg/mL. (E) Dose-dependent induction of PGC1β protein expression by RBP4 was measured by western blot. The results are expressed as the means ± SEM or presented as the representative blots from three independent experiments.

RBP4 Increases PGC-1β Binding to the LXREs of SREBP-1c

To identify whether the inductive effect of RBP4 on the processing of SREBP-1 depended on PGC-1β, HepG2 cells were transfected with PGC-1β siRNA or scramble siRNA as a control. PGC-1β siRNA efficiently decreased the protein levels of endogenous PGC-1β (a 73 ± 11% decrease) in HepG2 cells (Fig. S4A). Notably, the knockdown of Ppargc1b substantially inhibited the increase of SREBP-1 nuclear form by RBP4 (Fig. S4B) and was associated with the suppression of hepatic lipogenic enzyme expression (Fig. S4C). Consequently, [3H]-acetate incorporation was significantly attenuated in PGC-1β siRNA-transfected HepG2 cells, indicating that PGC-1β is an important regulator of RBP4-mediated hepatic lipogenesis (Fig. S4D). To further confirm the contribution of PGC-1β to hepatic lipogenesis, we overexpressed PGC-1β in cultured human HepG2 cells. The overexpression of PGC-1β significantly enhanced the basal and RBP4-mediated transcriptional activity of SREBP-1c (Fig. S4E), confirming the important role of PGC-1β in RBP4-increased SREBP-1 activation in HepG2 cells.

Mutation of the LXRE on the promoter completely abolished the transcriptional activation of SREBP-1c by RBP4 (Fig. 5A), indicating that LXR binding to its response element on the SREBP-1c promoter is required to mediate the effects of RBP4. We performed chromatin immunoprecipitation (ChIP) assays to clarify whether the activation of hepatic lipogenic genes was associated with an increased ability of PGC-1β complexes to bind LXREs in the promoters of these genes. In fact, RBP4 treatment increased the recruitment of PGC-1β to the LXREs of the SREBP-1c gene. The antibody against RNA polymerase II recognizes both the nonphosphorylated and the phosphorylated extending forms of RNA polymerase II, and occupancy within the gene can be used as an indicator of transcriptional activity.[29] RNA polymerase II ChIP-PCR showed that this RBP4-induced PGC-1β recruitment was associated with increased RNA polymerase II occupancy in the gene encoding SREBP-1c (Fig. 5B). Moreover, in PGC-1β-silenced HepG2 cells, RBP4 did not affect the occupancy of either PGC-1β or RNA polymerase II on the SREBP-1c gene. RBP4 increased PGC-1β binding to LXREs and increased RNA polymerase II occupancy on hepatic lipogenic genes in HepG2 cells (Fig. 5C).

Figure 5.

PGC-1β mediates RBP4-induced SREBP-1c activation. (A) Cultured HepG2 cells were transfected with SREBP-1c Luc or SREBP-1c ΔLXRE Luc for 48 hours. The cells were then stimulated with RBP4 (80 μg/mL) for another 24 hours. The level of luciferase activity was assayed. The results are expressed as the means ± SEM and expressed as fold changes from three independent experiments. (B,C) ChIP assays were performed using antibodies against PGC-1β (B) or RNA polymerase II (C) on HepG2 cells treated with vehicle, RBP4 (80 μg/mL), or RBP4 in the presence of scrambled or PGC1β siRNA (si). Recovery was determined by real-time PCR using primers positioned at the LXRE of SREBP1c located in the gene body. IgG was used as a control.

RBP4 Induces Ppargc1b Transcription Through CREB

Previous observations indicated that CREB regulates lipid metabolism and induces the transcription of Ppargc1a, which encodes PGC1α.[30] We further examined the involvement of CREB in linking RBP4 and PGC-1β induction. To better understand the role of CREB in RBP4-mediated PGC-1β expression, we introduced in human HepG2 cells with CREB inhibitor KG-501, a small molecule which blocks cyclic adenosine monophosphate (cAMP)-induction of CREB-dependent target gene transcription through interference in the binding between the KID domain of CREB and KIX domain of CREB binding protein.[31] Pretreatment of HepG2 cells with KG-501 (10 μM) suppressed the Ppargc1b expression induced by RBP4 (Fig. S5A). Conversely, overexpression of CREB exaggerated the expression of Ppargc1b in the presence of RBP4 (Fig. S5B).

Putative analysis of CREB-binding elements identified in the 5′-flanking region of Ppargc1b (Fig. S5C). Furthermore, ChIP measurements indicated that the binding of CREB to the CRE elements 1, 2, and 3 located 3.0, 3.4, and 6.5 kb upstream, respectively, in the Ppargc1b promoter increased upon RBP4 treatment (Fig. 6A). These data support the notion that CREB is involved in the transcriptional activation of PGC-1β by RBP4.

Figure 6.

RBP4 induces Ppargc1b transcription through CREB. (A) ChIP analysis of CREB binding to CRE1 and CRE2 of the PGC-1β promoter. The experiments were repeated three times and representative results are shown. (B) CREB (phospho-Ser133) transcriptional activity was measured as indicated using an ELISA-based method. Data are expressed as the means ± SEM of three independent experiments. *P < 0.05 or **P < 0.01 versus 0 μg/mL. (C) Measurement of luciferase activity in HepG2 cells overexpressing Gal4-CREB or Gal4-CREB S133A in the absence or presence of RBP4 treatment. Data are expressed as the means ± SEM of three independent experiments. *P < 0.05 or **P < 0.01 versus 0 μg/mL of Gal4-CREB transfected hepatocytes.

Phosphorylation on Ser133 activates CREB to induce the transcription of target genes. We then evaluated CREB DNA-binding activity using an enzyme-linked immunosorbent assay (ELISA)-based transcription factor assay kit for detecting phosphorylated CREB (Ser133). Notably, RBP4 treatment caused a dose-dependent induction of CREB transcriptional activity (Fig. 6B). We further confirmed that phosphorylation by RBP4 at Ser133 in CREB affected its physiological function. We transfected expression constructs containing a Gal4 DNA-binding domain fused to either wildtype (WT) CREB (pGAL4-CREB) or CREB containing a point mutation at serine 133 (mutated to alanine [pGAL4-CREB S133A]) into HepG2 cells in the absence or presence of RBP4. RBP4 treatment increased pGAL4-CREB activity by ∼1.9 to 4.4-fold compared with the control but had no effect on pGAL4-CREB S133A activity (Fig. 6C). Moreover, KG-501 prevented RBP4-induced transcriptional activation of SREBP-1c promoter (Fig. S6).

RBP4 Increases Hepatic Lipogenic Gene Expression and Lipid Accumulation in Liver in Mice

We next studied whether RBP4 induces hepatic lipogenesis in a PGC-1β-dependent manner in a mouse model in vivo. For this purpose, WT mice and Ppargc1b−/− mice were treated with the recombinant RBP4 or vehicle (dialysate obtained from the final step of RBP4 purification) for 14 days. This resulted in a daily average serum level of human RBP4 ∼2.3 times higher than endogenous mouse RBP4. In agreement with the in vitro data, RBP4 injection strongly induced PGC-1β protein expression (Fig. 7A), promoted the activation processing of SREBP-1 (Fig. 7B), and SREBP-1c expression (Fig. 7C) in C57BL/6J mice. As a result, RBP4 increased the hepatic expression of lipogenic genes, including FAS, Acc1, and Dgat2, in the postprandial state in WT mice (Fig. 7D). In addition, liver TAG accumulation (Fig. 7E) and plasma TAG levels (Fig. 7F) were much higher in the RBP4-treated C57BL/6J mice than that in untreated mice. Notably, the effect of RBP4 on hepatic lipid metabolism in WT mice was not observed in Ppargc1b−/− mice (Fig. 7A-F).

Figure 7.

RBP4 induces hepatic SREBP-1c and lipogenesis in vivo. (A) RBP4 induced PGC1β protein expression in livers from WT mice but not Ppargc1b−/− mice. (B) The mature, active nuclear form of hepatic SREBP-1 is increased in RBP4-treated WT mice, and the increase is completely blocked in Ppargc1b−/− mice. (C) Hepatic SREBP-1c and (D) lipogenic enzymes, including FAS, ACC1, and DGAT2 levels were measured by quantitative RT-PCR. Data normalized to GAPDH are expressed as the fold change compared with control. *P < 0.05. (E) Hepatic TAG mass was determined directly and expressed as mg/g of liver. The results are representative of eight mice in each group. *P < 0.05. (F) Serum TAG concentration was determined as indicated. *P < 0.05. Data are expressed as the means ± SEM of six mice per group. *P < 0.05 or **P < 0.01 versus corresponding time period of Ppargc1b−/−.


The present study uncovered novel findings that RBP4 promotes lipogenesis in hepatocytes by way of PGC-1β-dependent SREBP-1 activation both in vitro and in vivo. The following data support this conclusion: (1) RBP4 induces SREBP-1 activation by regulating SREBP-1 expression and nuclear active forms. (2) At cellular level, SREBP-1 activation by RBP4 promotes the transcription of target lipogenic genes, thereby stimulating de novo lipogenesis in HepG2 cells. (3) SREBP-1 is a direct target of PGC-1β and RBP4 increases ppargc1b transcription through CREB. (4) RBP4 stimulates the SREBP-1c and its target genes expression, resulting in hepatic triglyceride accumulation and VLDL secretion in C57BL/6J mice but not in Ppargc1b−/− mice. Thus, the induction of PGC-1β-dependent SREBP-1 activation may represent a molecular mechanism by which RBP4 regulates hepatic lipogenesis.

In the present study we found a positive dose-response effect of RBP4 (ranging from 0 to 80 μg/mL) in inducing hepatocyte lipogenesis. The stimulatory effect of RBP4 on lipogenesis was present at all doses, to a maximum at 80 μg/mL RBP4. Although different clinical studies reported different serum or plasma levels of RBP4 due to different measurement procedures and conditions,[32] human plasma RBP4 levels in patients with metabolic syndrome usually range from ranged from ∼20 μg/mL to ∼90 μg/mL. A plasma concentration of 20 μg/mL RBP4 was reported in normal lean humans.[7, 8] Plasma concentrations of 40∼90 μg/mL RBP4 were the typical range documented in patients who were obese and diabetic individuals.[7] Based on these reports, we estimate that our in vitro dose of RBP4 may encompass the clinically relevant range of serum RBP4 concentrations in patients with metabolic syndrome. In addition, this dose of RBP4 is consistent with those in the previous reports.[10, 33]

RBP4′s most well-defined function is to deliver retinol to tissues and since retinol has many important effects on lipid metabolism,[23, 24] it would not have been surprising if the action of RBP4 on hepatic lipid lipogenesis was retinol-dependent. However, we found that apo-RBP4 elicits lipid lipogenesis as robust as that of holo-RBP4 in HepG2 cells. Apo-RBP4 with retinol added back had similar effects to the original holo-RBP4, confirming that the retinol-stripping process did not alter the function of RBP4. Thus, RBP4 can elicit a lipogenic process from hepatocytes in a retinol-independent manner.

It is curious that RBP4 has such a robust effect on HepG2 cells because hepatocytes themselves are an important source of this protein. Actually, hepatocytes are regarded as the major source of RBP4 under normal physiological status; however, adipose tissue may be a major site of RBP4 synthesis in insulin-resistant states. Therefore, adipose tissue may play an endocrine role by increasing the circulating concentrations of RBP4. We thus speculate that RBP4 may affect hepatocyte function in both an autocrine and endocrine manner. In the autocrine model, the RBP secreted by hepatocytes affects hepatocyte function. Future studies are necessary to determine the relative contribution of the secretion of RBP4 from the liver and adipose tissue in the regulation of lipid synthesis in hepatocytes.

SREBPs are known to be important transcription factors and play a central role in lipid homeostasis; however, there is yet no evidence that links SREBP to the lipogenic effect of RBP4. Our present results reveal that in HepG2 cells, stimulation with human recombinant RBP4 did not affect the protein expression of SREBP-1, but rather reduced the nuclear mature form of SREBP-1, thus leading to a potent induction of its target genes as well as lipid accumulation in vitro. However, SREBP-2 predominantly regulates genes controlling cholesterol homeostasis, such as LDLR, HMG coenzyme A reductase, and squalene synthase[36, 37] was not affected in response to RBP4. These results are consistent with previous findings that hepatic overexpression of SREBP-1 induced lipogenesis.[21] SREBP-1a and SREBP-1c mainly regulates the transcription of key enzymes associated with the biosynthesis of fatty acids and the lipogenic process, although in the HepG2 cells, there are more SREBP-1a isoforms than SREBP-1c, but only the SREBP-1c promoter is transcriptionally activated in response to RBP4 treatment. This is demonstrated by our results that the transcription level of SREBP-1c is greatly induced by RBP4 treatment, while SREBP-1a is not significantly changed. Thus, the induced protein levels of SREBP-1 under RBP4 treatment is mainly from induced SREBP-1c, not SREBP-1a. Therefore, SREBP-1c might be the primary isoform of RBP4 action. Deletion analysis of the SREBP-1c promoter further showed that activity of the WT SREBP-1c promoter, but not SRE or the LXRE deletion promoter, was induced by RBP4. This study indicates that RBP4 functions as a potential adipokine that controls SREBP-1c transcriptional activity through autoloop regulation by way of an SRE/LXRE motif-dependent mechanism. Further studies with the use of SREBP-1c knockout mice are necessary to prove this feedforward mechanism. On the contrary, RBP4 exerted less effect on SREBP-2 transcriptional activity, as demonstrated by nuclear SREBP-2-induced autoregulation and the transcription of HMG-CoA and LDLR. These results strengthened the conclusion that the induction of lipogenesis by RBP4 is mainly dependent on SREBP-1c.

We further found that PGC-1β is a critical effector downstream of RBP4, which mediates RBP4 effects on the induction of SREBP-1c and thus promotes hepatic lipogenesis. Originally, PGC-1β was identified as the closest homolog of PGC-1α and a cold-inducible coactivator that interacts with peroxisome proliferator-activated receptor gamma (PPARγ) in brown adipose tissue.[38] PGC-1β is most highly expressed in tissues with high oxidative metabolism, such as brown adipose tissue, cardiac muscle, and skeletal muscle.[38, 39] PGC-1β coactivates the SREBP and LXR families of transcription factors, as it induces a broad program of lipid metabolism, including de novo lipogenesis and lipoprotein secretion.[28]

Subsequent studies identified Ppargc1b as an essential regulator in the lipid synthesis pathway and a transcriptional coactivator of SREBP-1c induced by RBP4. Both loss- and gain-of-function experiments suggest that PGC-1β is involved in transcriptional activation of SREBP-1c in response to RBP4 treatment. The depletion of PGC-1β strongly abolished the inductive effects of RBP4 on lipogenic gene transcription. In contrast, the overexpression of PGC-1β potently enhanced RBP4-mediated lipogenic gene transcription. Thus, PGC-1β is primarily responsible for the lipogenesis effect of RBP4. Furthermore, we provide the novel findings that RBP4 stimulates Ppargc1b expression in HepG2 cells. RBP4 treatment increases PGC-1β mRNA expression in a dose- and time-dependent fashion in hepatocytes. RBP4 treatment was also found to increase PGC-1β protein expression. However, RBP4 had little effect on Ppargc1α, the other isoform of PGC-1. Several pieces of data link PGC-1β with the LXR pathway.[28] PGC-1β coactivates LXR on both a synthetic reporter gene containing multimerized binding elements and an endogenous promoter in an LXR ligand-dependent manner. More important, PGC-1β is recruited to the promoter region of cytochrome P450 7A1 (CYP7A1) and ATP binding cassette A1 (ABCA1) and activates the expression of these LXR target genes.[40] We show here that RBP4 increased the recruitment of PGC-1β to the LXREs of specific SREBP-1c target genes implicated in hepatic lipogenesis, leading to their up-regulation and enhanced de novo TAG synthesis. Thus, LXRE is permissive for lipogenesis by RBP4 in hepatocytes. Although studies in this field have not elucidated how LXR activates the pathways of lipid transport in hepatocytes, the ability of PGC-1β to modulate LXR target gene expression in cultured cells and in vivo suggests that PGC-1β elicits at least a proportion of this hyperlipidemia through the coactivation of LXR. Taken together, it is clear that PGC-1β couples these two important aspects of lipid metabolism in liver, i.e., lipid synthesis by way of the coactivation of the SREBPs and lipoprotein secretion by way of the coactivation of LXR and likely other transcription factors.

Next, we explored the potential underlying mechanism by which RBP4 augments PGC-1β transcription. CREB is a cellular transcription factor that binds to certain DNA sequences called CRE, thereby increasing or decreasing the transcription of downstream genes.[41] Our study implicates the activation of CREB as a mechanism by which RBP4 increases PGC-1β expression. The ChIP assay revealed the direct binding of CRE to a noncanonical CRE motif upstream of the transcription initiation site of PGC-1β. This binding was enhanced by RBP4 treatment. Further studies indicate that CREB Ser133 is the critical target involved in the transcriptional induction of Ppargc1b induced by RBP4.

In conclusion, our findings further suggest that the development of future therapeutic intervention(s) to modulate RBP4 levels may serve to ameliorate hepatic abnormal lipogenesis and lipid accumulation related to obesity and diabetes.