Fatty acids stimulate lipid accumulation in parallel with increased expression of adipose differentiation-related protein (ADRP) in liver cells. Although it is generally considered that the fatty acid effect on ADRP expression is mediated by peroxisome proliferator-activated receptors (PPARs), we identified here an additional molecular mechanism using the NMuLi mouse liver nonparenchymal cell line, which expresses PPARγ and δ but not α. Oleic acid (OA) and specific ligands for PPARγ and -δ stimulated ADRP expression as well as the −2,090-bp ADRP promoter activity which encompasses the PPAR response element (PPRE) adjacent to an Ets/activator protein (AP)-1 site. When the AP-1 site was mutated, OA failed to stimulate the activity despite the presence of the PPRE, whereas ligands for PPARγ and -δ did stimulate it and so did a PPARα ligand under the coexpression of PPARα. DNA binding of AP-1 was stimulated by OA but not by PPAR ligands. Because we previously demonstrated that Pycnogenol (PYC), a French maritime pine bark extract, suppressed ADRP expression in macrophages partly by suppression of AP-1 activity, we tested the effect of PYC on NMuLi cells. PYC reduced the OA-induced ADRP expression along with suppression of lipid droplet formation. However, PYC neither suppressed the OA-stimulated ADRP promoter activity nor DNA binding of AP-1 but, instead, reduced the ADRP mRNA half-life. All these results indicate that the effect of OA on ADRP expression requires AP-1 as well as PPRE, and PYC suppresses the ADRP expression in part by facilitating mRNA degradation. PYC, a widely used dietary supplement, could be beneficial for the prevention of excessive lipid accumulation such as hepatic steatosis.
- lipid droplet
- hepatic steatosis
- adipose differentiation-related protein
- activator protein-1
- peroxisome proliferator-activated receptor
cytosolic lipid droplets are physiologically important because they have specific function in various cell types or tissues, for example, as an energy reservoir in adipocytes, sites of storage and biosynthesis of eicosanoids in leukocytes, and sites of production of pulmonary surfactants in pneumocytes (44). It has been demonstrated that excessive lipid accumulation, not only in an adipose tissue but also in various nonadipose tissues, is closely related to a variety of pathological conditions, including insulin resistance, type 2 diabetes mellitus, cardiovascular diseases, and fatty liver or nonalcoholic steatohepatitis (2, 59), all of which are emerging as important clinical ands socioeconomical worldwide problems. Mechanisms of intracellular lipid droplet formation have, therefore, increasingly attracted clinical and scientific interests.
A variety of proteins are associated with intracellular lipid droplets (10). Among them, PAT family proteins, which comprise perilipin, adipose differentiation-related protein (ADRP), TIP47, S3-12 and OXPAT/MLDP, are implicated in the formation, stabilization, and metabolism of lipid droplets (8, 37). ADRP was first identified in the early stages of adipocyte differentiation (33). Later studies revealed that ADRP does not directly induce adipogenesis but instead facilitates uptake of fatty acids or cholesterol (4, 22), whereas forced expression of ADRP also stimulates lipid droplet formation (31). In contrast to perilipin, whose expression is relatively limited in adipocytes and steroidgenic cells, ADRP is ubiquitously expressed in a variety of cells and is a specific marker of lipid accumulation (26).
The liver is a central organ for lipid metabolism, and excessive lipid accumulation in hepatocytes is generally associated with various metabolic abnormalities related to metabolic syndrome. In particular, fatty acids from several different sources, such as dietary fat, fatty acids released from adipose tissues, and de novo hepatic lipogenesis, facilitate lipid accumulation in hepatocytes in conjunction with oxidative stress or inflammatory stimuli (2, 59). ADRP expression is upregulated in hepatic steatosis in humans and mouse models (43). Considering that fatty acids act as ligands for all peroxisome proliferator-activated receptor (PPAR) subtypes (19, 34), although with different propensities for interaction (64), it is reasonable to hypothesize that the effect of fatty acids on ADRP expression is, at least in part, mediated by PPAR activation. Actually, a PPAR response element (PPRE) mediating the fatty acid effect was identified in the promoter region of mouse and human ADRP genes (13, 57). In our previous study (62), we demonstrated that oleic acid (OA)-induced enhancement of the mouse ADRP promoter activity required the upstream Ets/activator protein (AP)-1 element in RAW264.7 macrophage-like cells. This suggests that the PPRE in the ADRP promoter alone is likely to be insufficient for responding to OA, instead requiring the Ets/AP-1 element as well. This finding remains to be fully confirmed in other cell types.
On the other hand, fatty liver formation was markedly prevented in mice whose ADRP expression was abrogated by gene knockout or antisense oligonucleotides (12, 30). Besides fatty liver, it has recently been reported that ADRP abrogation also protects from atherosclerosis development (48) and diet-induced insulin resistance (60). These findings strongly suggest that ADRP could be a promising target for the prevention of excessive lipid accumulation. In a previous study, we reported that Pycnogenol (PYC), an extract from French maritime pine bark, suppressed lipopolysaccharide (LPS)-induced ADRP expression in RAW264.7 cells (23). It is important to elucidate if PYC can suppress the ADRP expression and lipid droplet formation in liver cells, since this widely used dietary supplement could be a useful modality for prevention or treatment of hepatic steatosis.
In the present study, we aimed to further clarify the mechanism by which long-chain fatty acids, such as OA, simulate the ADRP expression in the NMuLi mouse hepatic nonparenchymal cell line. We also tested the effect of PYC on OA-induced ADRP expression and lipid droplet formation and disclosed a mechanism of action. We demonstrate here that OA-induced ADRP expression requires AP-1 function in addition to PPRE in the promoter, and PYC suppressed OA-induced ADRP expression along with lipid droplet formation, in part, through facilitating ADRP mRNA degradation.
MATERIALS AND METHODS
The mouse liver epithelial cell line NMuLi was obtained from Dainippon Sumitomo Pharma (Osaka, Japan). The cells were routinely cultured in 10-cm tissue culture dishes (Falcon 3003; Becton-Dickinson Labware, Franklin Lakes, NJ) in DMEM (Invitrogen, Carlsbad, CA) supplemented with 10% charcoal-treated FCS, 1% nonessential amino acids, and appropriate antibiotics. The charcoal treatment of the FCS did not adversely affect concentrations of fatty acids, triacylglyceride, and total cholesterol (data not shown). We preliminarily confirmed by trypan blue staining that concentrations of reagents we used in this study were those that did not affect cell viability.
Actinomycin D, GW-501516, 15-deoxy-Δ12,14-prostaglandin J2, astaxanthin, and curcumin were purchased from Sigma (St Louis, MI). Caproic acid and OA were purchased from Wako Pure Chemical Industries (Osaka, Japan), dissolved in water containing 5 mM BSA, and then diluted in the medium. Fenofibrate, bezafibrate, troglitazone, and pioglitazone were kindly provided by Kaken Pharmaceuticals (Tokyo, Japan), Kissei Pharmaceuticals (Matsumoto, Japan), Sankyo (Tokyo, Japan), and Takeda (Tokyo, Japan), respectively. All of these agents were dissolved in dimethyl sulfoxide. Antibodies to ADRP and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were purchased from PROGEN Biotechnik (Heidelberg, Germany) and Ambion (Woodward Austin, TX), respectively. PYC was generously provided by Horphag Research (Geneva, Switzerland). Synthesized oligonucleotides were purchased from Roche Diagnostics (Tokyo, Japan). [α-32P]dCTP (5,000 Ci/mmol) and [γ-32P]ATP (5,000 Ci/mmol) were purchased from GE Healthcare (Buckinghamshire, UK).
Northern blot analysis.
NMuLi cells were plated at 1 × 106 cells/dish and cultivated for 24 h, and then the medium was changed to fresh medium containing the test reagent(s), followed by culturing for indicated times. Total RNA was purified using a commercially available kit (Isogen; Nippon GENE, Tokyo, Japan). Full-length mouse ADRP cDNA was [α-32P]dCTP-labeled by a random priming method (Oligo Labeling kit; GE Healthcare) and used as a probe. Hybridization was performed according to the standard method (51); final washing was carried out at 65°C in 1× 150 mM NaCl, 10 mM NaH2PO4, and 1 mM EDTA (pH 7.4) containing 0.5% SDS for 2 × 20 min. In addition, mRNA expression levels were also evaluated by the real-time PCR as described below, and the results are shown in bar graphs in the Figs. 1–8.
Western blot analysis.
Whole cell extracts were prepared by directly dissolving the cells in 2× SDS loading buffer (100 mM Tris·HCl, pH 6.8, 20% glycerol, 4% SDS, 12% 2-mercaptoethanol, and 2% bromphenol blue) at 1:1 in volume. The proteins were heated at 95°C for 5 min, resolved by 10% SDS-PAGE, and electroblotted on a polyvinylidene difluoride membrane (Millipore, Tokyo, Japan) for 1 h at 100 V with a wet blotting apparatus (Bio-Rad, Hercules, CA) in Tris-glycine transfer buffer (25 mM Tris, 192 mM glycine, 20% methanol, and 0.1% SDS). The transfer was monitored with Kaleidoscope prestained standards (Bio-Rad). The membranes were blocked for 1 h at room temperature with 5% nonfat milk in PBS containing 0.1% Tween 20 (PBS-Tween 20). Next, the membranes were incubated with guinea pig ADRP antibodies (0.5 μg/ml in PBS-Tween 20) for 1 h. After being washed four times (5 min each) with PBS-Tween 20, the membranes were incubated with an appropriate secondary IgG-horseradish peroxidase conjugate (Santa Cruz Biotechnology, Santa Cruz, CA) diluted in PBS-Tween 20 (0.08 μg/ml) for 1 h, and then washed as above. The blots were developed using ECL Western blotting detection reagents (GE Healthcare) and exposed to Hyperfilm ECL (GE Healthcare) for ∼15 min. Protein levels were evaluated using NIH Image, and the results are shown in bar graphs in Figs. 1–8.
Construction of promoter-reporter plasmids.
Transient transfection and luciferase assay.
Reporter plasmids were introduced into NMuLi cells using Lipofectamine PLUS (Invitrogen) according to the manufacturer's instructions. Cells were seeded at 2 × 105 cells/well in a 12-well cell culture cluster (Costar 3513; Corning, Corning, NY) and grown for 24 h. Medium was then changed to serum-depleted DMEM, and the transfection mixture (2 μl lipofectamine and 5 μl PLUS reagent 5 in a total volume of 50 μl/well) containing 1 μg reporter plasmid and 0.4 μg pRL-TK plasmid (Promega, Madison, WI) was added to the cells and incubated for 3 h. The cells were then washed with serum-depleted DMEM and cultured in the DMEM containing the test reagent(s) for an additional 24 h. Cell lysates were collected, and luciferase activity was measured using a Dual-Luciferase Reporter Assay System (Promega). The transfection efficiency was normalized to the Renilla luciferase activity expressed by pRL-TK. In the experiments using PPARα coexpression, 0.5 μg of PPARα expression vector or a vacant vector (pCMX) was added to the transfection mixture. The PPARα expression vector was kindly provided by Dr. Yasutomi Kamei (Tokyo Medical and Dental University, Tokyo, Japan). Each transfection experiment was performed in triplicate and repeated at least three times.
Nuclear protein extracts.
Nuclear protein extracts were prepared using a method described (55). NMuLi cells were seeded at 2 × 106 cells/dish and then cultured for 24 h. The medium was then changed to that containing a test reagent or a vehicle and cultured for an additional 1 h. When PYC effect was tested, cells were pretreated with PYC for 1 h before adding OA. The cells were washed with PBS, harvested by gentle scraping, and then resuspended in 5 volumes of buffer A (10 mM HEPES-KOH, pH 7.9, 10 mM KCl, 1.5 mM MgCl2, and 0.1 mM EDTA) supplemented with 0.3 M sucrose and 0.5% Nonidet P-40. The cells were then homogenized by pipetting, and the suspension was layered on 1 ml buffer A containing 1.5 M sucrose, followed by centrifugation for 10 min (4°C, 15,000 g). After being washed with buffer A, the precipitated nuclei were suspended in 50 μl buffer B (20 mM HEPES-KOH, pH 7.9, 25% glycerol, 420 mM NaCl, 1.5 mM MgCl2, and 0.2 mM EDTA) and then left on ice for 20 min. The mixture was centrifuged for 20 min (4°C, 15,000 g), and the resultant supernatant containing nuclear proteins was aliquoted, snap-frozen in liquid nitrogen, and stored at −80°C until use (within 30 days). All solutions used were ice-cold and contained 0.5 mM dithiotheitol (DTT), 0.5 mM phenylmethylsulfonyl fluoride (PMSF), 2 mg/ml pepstatin A, and 2 mg/ml leupeptin. The protein concentration was determined with a commercially available reagent (Bio-Rad) using BSA as a control.
Electrophoretic mobility shift assay.
Electrophoretic mobility shift assay (EMSA) was performed as previously described (23, 29, 62). A synthesized sense oligonucleotide was 32P-labeled with T4 polynucleotide kinase (Takara, Shiga, Japan), double-stranded with the unlabeled antisense strand, and then purified using Chroma-Spin+TE-10 columns (Clontech, Palo Alto, CA). NMuLi nuclear extracts (10 μg) were first incubated in a 25-μl reaction volume for 20 min at 20°C with or without unlabeled competitor oligonucleotides (100-fold molar excess). The reaction buffer consisted of 10 mM Tris·HCl (pH 7.6), 50 mM KCl, 5 mM MgCl2, 1 mM DTT, 1 mM EDTA, 12.5% (vol/vol) glycerol, 0.1% Triton X-100, 8 μg/ml calf thymus DNA, and 50 mM PMSF. A radiolabeled probe (50,000 counts/min, ∼1.0 ng DNA) was then added and incubated for an additional 20 min at 20°C. DNA-protein complexes were analyzed on a 5% native polyacrylamide gel at 100 V for 3 h in Tris-borate-EDTA buffer. Thereafter, the gels were dried and subjected to autoradiography, and a specific protein-DNA complex was quatified by NIH Image if necessary.
Oil-red O staining.
NMuLi cells were plated in a four-well chamber slide (Nalge Nunc International, Naperville, IL) at 1 × 105/chamber and cultured in the presence or absence of 50 μg/ml PYC for 1 h. The cells were then cultured in medium containing 200 μM OA in the presence or absence of PYC for 96 h. The cells were washed with PBS, fixed in 10% formaldehyde, and stained with the neutral lipid dye Oil-red O (0.3% in 60% isopropanol), followed by extensive washes (38). Photomicrographs were generated using a digital camera (U-CMAD3; Olympus, Tokyo, Japan) connected to an Olympus BX51 microscope.
PCR and real-time RT-PCR.
Single-stranded cDNA was synthesized with the ReverTra Ace-α kit (Toyobo, Osaka, Japan) using 0.5 μg of the total RNA. The PPAR expression was determined by PCR amplification of 35 cycles with annealing temperature at 60°C using the following primers: PPARα (forward: TGGACACAGAGAGCCCCATC, reverse: TCGTACACCAGCTTCAGCCG), PPARγ (forward: TGGCCATTGAGTGCCGAGTC, reverse: CGCCTTGGCTTTGGTCAGCG), and PPARδ (forward: GCTCAATGGGGGACCAGAAC, reverse: GCTTAGAGAAGGCCTTCAGG).
Real-time RT-PCR was performed with a SYBR Green method using the iCycler iQ multicolor real-time PCR detection system (Bio-Rad) in 25-μl reactions [12.5 μl of 2× iQ SYBR Green supermix (Bio-Rad, Hercules, CA), 320 nM each primer, 5 μl of 1:20 diluted cDNA]. Sequences of the primers were as follows: ADRP (forward: CTGTCTACCAAGCTCTGCTC, reverse: CGATGCTTCTCTTCCACTCC) and GAPDH (forward: ACCACAGTCCATGCCATCAC, reverse: TCCACCACCCTGTTGCTTA). PCR efficiencies for all reactions were >0.90. Quantitative PCR results were expressed as relative fold induction compared with the housekeeping gene GAPDH.
All data were presented as means ± SE. Statistical differences were determined by one-way ANOVA. A P < 0.05 was considered to be significant.
Long-chain fatty acids stimulate the expression of ADRP in the NMuLi hepatic nonparenchymal cell line.
Several studies have demonstrated that long-chain fatty acids stimulate ADRP expression in different types of cells (21, 22, 52, 62), and they also act as ligands for all three PPAR subtypes (19, 34). We first determined the expression of PPAR subtypes in NMuLi cells used in this study. We found that this cell line expressed PPARγ and PPARδ but not PPARα (Fig. 1A). The expression pattern of these PPARs was not affected by the agents that we used in this study (data not shown). Although NMuLi is not of hepatic parenchymal origin, it has been used expediently as an alternative of a normal murine hepatocyte having some similar propensities to hepatic parenchymal cells (42). More importantly, we found it a merit that the cell line did not express PPARα because we could test an effect of exogenously introduced PPARα on ADRP promoter activity in the following analyses.
As in previous studies (21, 52, 62), the expression of ADRP mRNA and protein was increased significantly by OA (C-18) after 12 h incubation, but not by caproic acid (C-6) (Fig. 1B), and that the effect of OA was dose-dependent (Fig. 1C). We also confirmed that other long-chain fatty acids, such as palmitic acid (C-16), linoleic acid (C-18), and linolenic acid (C-18), enhanced the ADRP mRNA and protein levels (data not shown), and used only OA for further analyses.
The effect of OA on the ADRP expression was exerted at the transcriptional level, since the addition of actinomycin D completely abrogated OA-induced enhancement of ADRP mRNA expression (Fig. 2A). It has been demonstrated that triacsin C, a long-chain acyl-CoA synthetase inhibitor, inhibits synthesis of triacylglycerol and cholesteryl esters, thereby blocking foam cell formation in macrophages (45). We, therefore, tested the effect of triacsin C on OA-induced expression of ADRP mRNA. As shown in Fig. 2B, the stimulatory effect of low-dose OA (20 μM) was significantly potentiated in the presence of triacsin C. Because a higher dose of OA (100 and 200 μM) stimulated high mRNA expression possibly at maximum level, the additive effect of triacsin C was not evident. This result suggested that OA itself, but not triacylglycerol, stimulates the mRNA expression. We confirmed that the PPARγ ligands troglitazone (Fig. 2C), pioglitazone, and 15-deoxy-Δ12,14-prostaglandin J2 (not shown) and a PPARδ ligand, GW-501516 (Fig. 2C), enhanced ADRP mRNA expression. As expected, PPARα ligands, fenofibrate (Fig. 2C) and bezafibrate (data not shown), did not significantly stimulate the expression. Taken together with previous reports proving that fatty acids are potential ligands for PPARs (19, 34), these results suggest that OA stimulates ADRP expression in part through PPARγ and -δ activation in NMuLi cells.
Functional AP-1 site is requisite for OA action.
Next, we investigated the effect of OA on the mouse ADRP promoter activity. The −2,090-bp promoter region of the mouse ADRP gene encompasses a canonical PPRE, starting at −2,001 bp (13, 62). In addition, we demonstrated that an Ets/AP-1 composite element exists at −2,047 bp, which is important for phorbol myristate ester (PMA)- and LPS-induced enhancement of ADRP expression in macrophages (23, 62). The promoter region encompassing these elements is highly conserved between mouse and human ADRP genes (Fig. 3A). As shown in Fig. 3B, activity of the −2,090-bp ADRP promoter was enhanced by OA, whereas caproic acid had no effect, consistent with the effects on the mRNA expression.
In a previous study, we raised the possibility of functional interaction between Ets/AP-1 and PPRE in response to PMA and OA in RAW264.7 cells (62). We, therefore, tested the OA-induced promoter activity, using a series of deleted or mutated promoters derived from the wild-type −2,090-bp promoter. By deletion of all three elements (−1,889 bp), the basal promoter activity was reduced significantly, and the OA-induced enhancement was completely abolished (Fig. 3C). Basal activity of the PPRE mutation (Pmut) was almost comparable to the −2,090-bp promoter, whereas OA failed to stimulate, actually suppressed, the Pmut promoter activity (Fig. 3C), indicating that OA acts through the PPRE. The basal activity of promoters containing either AP-1 or Ets mutations (Amut or Emut) was reduced, suggesting that these elements might be important for basal promoter activity. Interestingly, OA-induced enhancement was almost completely diminished in the Amut promoter despite the presence of an intact PPRE, whereas OA stimulated the Emut promoter to the wild-type levels (Fig. 3D). These results indicate that AP-1, in addition to the PPRE, is indispensable for OA action on ADRP expression in this cell line.
We further tested whether intact AP-1 function is also required for the action of synthesized ligands specific for PPAR subtypes. Like OA, a PPARγ ligand (troglitazone) and a PPARδ ligand (GW-501516) stimulated activity of the wild-type −2,090-bp and Emut promoters but not −1,889-bp and Pmut promoters. In contrast to OA, however, these compounds also stimulated Amut promoter activity (Fig. 4, A and B). This finding was duplicated by a PPARα ligand (fenofibrate) when PPARα was coexpressed in the cells (Fig. 4C). These results indicate that, while specific ligands for all three PPAR subtypes and OA could stimulate ADRP promoter activity through PPRE, only OA requires intact AP-1 function.
This was investigated further using an EMSA. Use of an oligonucleotide probe containing the Ets/AP-1 element exhibited double complexes (complex A) by NMuLi nuclear extracts (Fig. 5A). Complex A formation was effectively inhibited by the addition of oligonucleotide competitors containing wild-type sequence (self), Emut, and AP-1 consensus sequence, but not by competitors containing Amut, double mutations in Ets/AP-1 sequence, and PU.1 consensus sequence. This result indicates that the Ets/AP-1 site is recognized only by the AP-1 family of transcription factors in NMuLi cells, unlike RAW264.7 cells in which AP-1 and PU.1 conjointly bind (23, 62). DNA binding of AP-1 family proteins was enhanced by OA but not by caproic acid and synthesized PPAR ligands (Fig. 5B). This finding further supported the involvement of AP-1 in the OA effect.
Collectively, these results indicate that the Ets/AP-1 element is important for the basal promoter activity, and more importantly that OA stimulates ADRP expression by acting through the PPRE in collaboration with the AP-1 site in the promoter in NMuLi cells. The Ets sequence also seems to be involved in basal promoter activity, although we could not identify protein binding to this site by EMSA. This problem requires further investigation.
PYC suppresses OA-induced ADRP expression and lipid droplet formation in NMuLi cells.
Previously, we demonstrated that PYC suppressed ADRP expression in RAW264.7 cells (23). We, therefore, tested the effect of PYC on ADRP expression and intracellular lipid droplet formation in NMuLi cells. PYC significantly suppressed the OA-induced increase in ADRP mRNA and protein levels in a dose-dependent manner, whereas it did not affect basal ADRP expression levels (Fig. 6A). PYC (50 μg/ml) almost completely inhibited OA-induced ADRP expression. OA markedly induced intracellular lipid accumulation, which was clearly visualized by Oil-red O staining (Fig. 6B, a vs. b). Consistent with the effect on ADRP expression, OA-induced lipid droplet formation was remarkably inhibited in the presence of 50 μg/ml PYC with no apparent change in cell viability (Fig. 6B, b vs. d). The suppressive effect of PYC on OA-induced ADRP expression was likely to be specific, because the antioxidants, curcumin and astaxanthin (25, 27), failed to antagonize the OA effect on mRNA expression (Fig. 7).
PYC suppresses ADRP expression by facilitating the mRNA degradation.
Previously, we demonstrated that PYC suppresses the enhancer activity of AP-1 and nuclear factor (NF)-κB, thereby leading to the suppression of LPS-induced ADRP expression in RAW264.7 cells (23). Specifically, LPS-stimulated activity of the −2,090-bp promoter was suppressed by PYC in RAW264.7 cells (23). Because of the importance of AP-1 function in OA-induced activation, we investigated the effect of PYC on the ADRP promoter and found that OA-stimulated activity of the −2,090-bp promoter was not suppressed at all in NMuLi cells (Fig. 8A). In addition, OA-induced enhancement of AP-1 binding to the Ets/AP-1 element was not inhibited by PYC (Fig. 8B).
To clarify the mechanism by which PYC suppresses the OA-induced ADRP mRNA expression, we presumed that a posttranscriptional mechanism, such as mRNA degradation, might be involved. Therefore, we tested whether PYC might shorten the half-life of ADRP mRNA. NMuLi cells were cultured in the presence or absence of 50 μg/ml PYC for 1 h and then incubated with OA in the presence of 5 μg/ml actinomycin D. Here we added OA in the medium since OA itself might affect the mRNA stability. The ADRP mRNA level was measured up to 12 h after stimulation. When PYC was present in the medium, a faster decrease in ADRP mRNA levels was observed when compared with the absence of PYC (Fig. 8C). GAPDH mRNA as a control was not affected by PYC (data not shown). These results were confirmed by quantitative PCR (Fig. 8D). The estimated half-life of the ADRP mRNA was ∼6 h in the presence of PYC and 11 h in the absence of PYC.
All of these results indicate that PYC suppresses OA-induced ADRP expression and lipid droplet formation in NMuLi cells. The mechanism involved is, in part, facilitation of ADRP mRNA degradation. The precise molecular mechanism of mRNA degradation still remains to be elucidated.
ADRP is a ubiquitously and constitutively expressed protein (9) that is rapidly degraded via the proteasome in the absence of lipids (39, 62, 63). A number of studies demonstrated high expression levels of ADRP in various pathological conditions in humans and animal models, including steatohepatitis or fatty liver (43), atherosclerosis (36), and diabetic nephropathy (40). Although it has been shown that several pathological or pharmacological stimuli enhance ADRP expression, the precise molecular mechanism regulating the expression is not fully elucidated. In previous studies, we and others demonstrated that an inflammatory signal, such as PMA or LPS, stimulates ADRP expression in macrophages (14, 23, 62), hepatocytes (46), and keratinocytes (15). We identified the Ets/AP-1 composite element in the mouse ADRP promoter and proved that it mediates, in part, the action of PMA or LPS, which enhances ADRP expression in RAW264.7 cells (23, 62).
Another important regulatory mechanism of the ADRP gene is a PPAR-mediated pathway that plays a central role in lipid homeostasis in various tissues. A number of reports demonstrated that specific ligands for PPARα (16, 18), PPARγ (7, 24, 52), and PPARδ (13, 49, 53, 61) upregulate the expression of ADRP, and functional PPRE was identified in human and mouse ADRP genes (13, 57, 62). In addition to specific PPAR ligands, it was shown that long-chain fatty acids stimulated ADRP expression in various cells (21, 58, 62). Because long-chain fatty acids act as ligands for all PPAR subtypes (19, 34), it is understandable that the effects of fatty acids are mediated through PPAR activation. Specifically, it was shown that the activity of ADRP promoters of human (57) and mouse (13) origin, both of which encompass the PPRE, is stimulated by OA and lipoprotein lipase-hydrolyzed very low-density lipoprotein, respectively. Here, we have demonstrated that OA and specific ligands for PPARγ and -δ enhanced ADRP expression and ADRP promoter activity in NMuLi hepatic nonparenchymal cells, which express PPARγ and PPARδ but not PPARα. Under the coexpression of PPARα, however, a PPARα ligand stimulated ADRP promoter activity in this cell line. Furthermore, all of these specific ligands and OA failed to stimulate activity of the promoter containing a PPRE mutation. These findings confirm an indispensable role of PPARs and PPRE in the regulation of ADRP gene.
We found, however, an important difference between effects of OA and specific PPAR ligands on ADRP promoter activity. Specific PPAR ligands stimulated the promoter activity containing an AP-1 mutation with an intact PPRE (Amut), whereas OA did not. This finding was preliminarily implicated in our previous report (62). In the process of characterizing the Ets/AP-1 element of the mouse ADRP promoter in RAW264.7 cells, we found that the synergistic effect of OA and PMA was abolished when the Ets/AP-1 site was mutated, even though the promoter contained an intact PPRE. As shown by EMSA, the Ets/AP-1 site was recognized only by the AP-1 family of transcription factors but not the Ets family proteins such as PU.1 in NMuLi cells. Therefore, it seems reasonable that only the AP-1 mutation abolished the OA-induced activation of the promoter. DNA binding of AP-1 was enhanced by OA but not by specific PPAR ligands and caproic acid. It was shown that long-chain fatty acids, but not short-chain fatty acids, activate membrane receptors such as GPR40 and GPR120 (11, 28, 32). Recently, it has been reported that OA induces AP-1-DNA binding activity through Src kinase-dependent extracellular signal-regulated kinase (ERK) 1/2 activation in breast cancer cells (54). It could be surmised that this kind of membrane receptor-mediated mechanism might also be responsible for the ADRP gene regulation by OA, since our preliminary experiments showed that a Src kinase inhibitor suppressed ADRP mRNA expression (data not shown).
Thus, we have shown here that functional AP-1 activity is indispensable for PPAR-mediated ADRP expression by OA but not by specific PPAR ligands. Although responsiveness of the human or mouse ADRP gene to fatty acids was demonstrated in previous reports (13, 57), the Ets/AP-1 site was involved in the promoter region but not modified in their experiments. As we proposed (62), the Ets/AP-1 element and PPRE may form a functional complex as a lipid sensor in the ADRP promoter. We surmise that possible factors connecting these two elements might involve the transcriptional coactivator CBP/p300. CBP/p300 was shown to be indispensable for PPARγ-mediated adipocyte differentiation (56). OA, which stimulates AP-1 activity, could efficiently recruit CBP/p300 (3, 5), and induce interaction with PPARγ-interacting protein (PRIP) or PRIP-interacting protein with methyltransferase domain (41, 50), in turn, enhancing to PPAR-mediated actions. This possibility is currently under investigation.
PYC is a natural extract from French maritime pine bark and widely used as a dietary supplement or a herbal drug. The main constituents are monomeric phenolic compounds (catechin, epicatechin, and taxifolin) and condensed flavonoids (procyanidines and proanthocyanidines) (47). We showed here that PYC suppressed OA-induced ADRP expression in parallel with lipid droplet formation in NMuLi cells. In our previous report, we demonstrated that PYC suppressed enhancer activity of AP-1 and NF-κB, thereby suppressing ADRP expression in RAW264.7 cells by directly acting on its promoter and via the inhibition of inflammatory cytokine production (23). Considering the functional importance of AP-1 in mediating the OA effect, we had presumed that PYC would suppress the −2,090-bp ADRP promoter activity, which was significantly suppressed by PYC in RAW264.7 cells. Contrary to our expectation, PYC did not suppress ADRP promoter activity at all. Instead, we found that PYC shortened the half-life of ADRP mRNA. Although dominant biological effects of PYC have been attributed to its antioxidant property (20, 47), it seems that the effect on the ADRP expression could not be necessarily ascribed to it. In this study and our previous report (23), we demonstrated that DNA binding of AP-1 or NF-κB, typical redox-sensitive transcription factors, was not inhibited by PYC. In addition, curcumin or astaxanthin, which are well-known antioxidant substances (25, 27), failed to antagonize the OA-induced increase of ADRP mRNA. One of the major posttranscriptional regulatory mechanisms of gene expression is degradation of mRNAs (17). Many kinds of stabilizing or destabilizing factors, which bind to AU-rich stretches of 3′-untranslated region (UTR) of mRNAs, have been identified (6). Berberine, a well-accepted herbal medicine in China, modulates mRNA stability. It activates the mitogen/extracellular-regulated kinase 1-ERK pathway and stabilizes low-density lipoprotein (LDL) receptor mRNA, resulting in a serum LDL-cholesterol lowering effect (1, 35). Although the presence of AU-rich stretches in 3′-UTR in ADRP mRNA are yet to be identified, it is possible that PYC antagonizes the effect of OA through modification of stabilizing or destabilizing factors that regulate the ADRP mRNA stability.
In conclusion, we have demonstrated here, using the NMuLi mouse hepatic nonparenchymal cell line, that the OA-induced ADRP expression requires not only a PPRE but also an AP-1 site on its promoter, and this mechanism is distinct from that of specific PPAR ligands. In addition, we demonstrated that PYC suppressed ADRP expression and lipid accumulation by facilitating ADRP mRNA degradation. This study is the first to report showing the efficient effect of PYC on intracellular lipid accumulation. PYC, a widely used dietary supplement, could be a unique tool for the prevention or therapy of excessive lipid accumulation such as fatty liver disease.
This work was in part supported by a Grant-in-Aid for Scientific Research, Japan Society for the Promotion of Science.
We thank Yoshiaki Komatsu, Diagnostic Laboratories, Kyushu University Hospital at Beppu, Japan, for excellent technical support.
- Copyright © 2009 the American Physiological Society