Disruption of transforming growth factor-β signaling by curcumin induces gene expression of peroxisome proliferator-activated receptor-γ in rat hepatic stellate cells

Shizhong Zheng, Anping Chen

Abstract

Activation of hepatic stellate cells (HSC), the major effectors of hepatic fibrogenesis, is coupled with sequential alterations in gene expression, including an increase in receptors for transforming growth factor-β (TGF-β) and a dramatic reduction in the peroxisome proliferator-activated receptor-γ (PPAR-γ). The relationship between them remains obscure. We previously demonstrated that curcumin induced gene expression of PPAR-γ in activated HSC, leading to reducing cell proliferation, inducing apoptosis and suppressing expression of extracellular matrix genes. The underlying molecular mechanisms are largely unknown. We recently observed that stimulation of PPAR-γ activation suppressed gene expression of TGF-β receptors in activated HSC, leading to the interruption of TGF-β signaling. This observation supported our assumption of an antagonistic relationship between PPAR-γ activation and TGF-β signaling in HSC. In this study, we further hypothesize that TGF-β signaling might negatively regulate gene expression of PPAR-γ in activated HSC. The present report demonstrates that exogenous TGF-β1 inhibits gene expression of PPAR-γ in activated HSC, which is eliminated by the pretreatment with curcumin likely by interrupting TGF-β signaling. Transfection assays further indicate that blocking TGF-β signaling by dominant negative type II TGF-β receptor increases the promoter activity of PPAR-γ gene. Promoter deletion assays, site-directed mutageneses, and gel shift assays localize two Smad binding elements (SBEs) in the PPAR-γ gene promoter, acting as curcumin response elements and negatively regulating the promoter activity in passaged HSC. The Smad3/4 protein complex specifically binds to the SBEs. Overexpression of Smad4 dose dependently eliminates the inhibitory effects of curcumin on the PPAR-γ gene promoter and TGF-β signaling. Taken together, these results demonstrate that the interruption of TGF-β signaling by curcumin induces gene expression of PPAR-γ in activated HSC in vitro. Our studies provide novel insights into the molecular mechanisms of curcumin in the induction of PPAR-γ gene expression and in the inhibition of HSC activation.

  • collagen
  • fibrogenesis
  • phytochemicals
  • signal transduction

hepatic stellate cells (HSC) are the primary source of excessive production and deposition of extracellular matrix (ECM) during hepatic fibrogenesis (3, 13). HSC activation, characterized by enhanced cell growth and overproduction of ECM, is triggered by the release of fibrogenic transforming growth factor-β1 (TGF-β1) from Kupffer cells and activated HSC (33). This process is coupled with up-expression of type I and II TGF-β receptors (14). In addition, HSC activation coincides with a dramatic reduction in expression of the peroxisome proliferator-activated receptor-γ (PPAR-γ) (17, 23, 26).

TGF-β signaling is initiated by binding of active TGF-β1 to type II TGF-β receptor (Tβ-RII), which leads to the phosphorylation and activation of type I TGF-β receptor (Tβ-RI) (1, 24). The latter, in turn, phosphorylates Smad2 or 3, which subsequently form a complex with Smad4 and migrate into the nucleus to regulate expression of target genes (27, 28). Recent studies showed different roles of Smad2 and Smad3 in deposition of ECM components and cell proliferation in rat HSC (35). TGF-β signaling via Smad3 played an important role in the morphological and functional maturation of HSC (35). In response to a variety of endogenous and exogenous agonists, the nuclear receptor PPAR-γ forms heterodimers with the retinoid X receptor and binds to peroxisome proliferator response elements (PPRE) in gene promoters to regulate the transcription of target genes (21). PPAR-γ is highly expressed in quiescent HSC in the normal liver (17, 23, 26). However, the level of PPAR-γ and its activity are dramatically reduced during HSC activation in vitro and in vivo (17, 23, 26). The stimulation of PPAR-γ activity by its agonists inhibits HSC proliferation and α1(I) collagen expression in vitro and in vivo (18, 26). Forced expression of exogenous PPAR-γ cDNA itself is sufficient to reverse the morphology of activated HSC to the quiescent phenotype (20).

Curcumin, the yellow pigment in turmeric, has received attention as a promising dietary supplement for the protection against fibrogenic insults (7, 29). We previously demonstrated that curcumin dramatically induced gene expression of PPAR-γ in activated HSC, which facilitated its trans-activation activity, leading to the inhibition of HSC proliferation, the induction of apoptosis, and the suppression of ECM production (37, 39). We recently observed that curcumin interrupted TGF-β signaling in activated HSC likely by suppressing gene expression of TGF-β receptors (39, 40).

The aim of this study is to elucidate the molecular mechanisms underlying the curcumin induction of PPAR-γ gene expression in activated HSC. We hypothesize that TGF-β signaling might negatively regulate gene expression of PPAR-γ in activated HSC. The induction of gene expression of PPAR-γ by curcumin in activated HSC might result, at least partially, from the disruption of TGF-β signaling. Results from this report support our hypothesis and provide novel insights into the mechanisms underlying the curcumin induction of PPAR-γ expression in activated HSC.

MATERIALS AND METHODS

Isolation and culture of hepatic stellate cells.

HSC were isolated from male Sprague-Dawley rats (∼200–250 g) as previously described (4). Cells were cultured in DMEM supplemented with FBS (10%). HSC aged at passage 4–8 were used for experiments. All of the experiments were performed in media with FBS (10%) unless specific indication. Curcumin (purity >94%) was purchased from Sigma (St. Louis, MO). Active human TGF-β1 was purchased from Cell Sciences (Canton, MA).

Western blotting analyses.

Whole cell extracts were prepared from preconfluent passaged HSC. SDS-PAGE with 10% resolving gel was used to separate proteins (25 μg/well). Separated proteins were detected by using primary antibodies and horseradish peroxidase-conjugated secondary antibodies (Santa Cruz Biotech). Protein bands were visualized by utilizing chemiluminescence reagent (KirKegarrd & Perry Laboratories).

Plasmids and transient transfection assays.

The PPAR-γ promoter luciferase reporter plasmid pPPAR-γ-Luc contain the 5′-flanking region (−2776 bp) of the PPAR-γ gene promoter in a luciferase reporter plasmid (11). The rest of the PPAR-γ promoter luciferase reporter plasmids with various sizes of the promoter region were derived from the parental pPPAR-γ-Luc. They were kindly provided by Dr. Johan Auwerx (Pasteur Institute, Lille, France) (11). The PPAR-γ activity reporter plasmid pPPRE-TK-Luc contains three copies of the PPAR-γ response elements from acyl-CoA oxidase gene linked to the herpes virus thymidine kinase promoter (−105/+51) and a luciferase vector, which was a gift from Dr. Kevin J. McCarthy (Louisiana State University Health Sciences Center in Shreveport). The cDNA expression plasmid pdn-Tβ-RII was a gift from Dr. Robert J. Lechleider (National Cancer Institute, Bethesda, MD), containing cDNA encoding the dominant negative form of Tβ-RII (9). The plasmid p3TP-Lux was a TGF-β-inducible luciferase reporter, containing the promoter of plasminogen activator inhibitor-1 (PAI-1) gene, kindly provided by Dr. Joan Massague (Memorial Sloan-Kettering Cancer Center, New York, NY). The cDNA expression plasmid psmad4-cDNA contains a full length of Smad4 cDNA in a cytomegalovirus-driven plasmid. It was kindly provided by Dr. Lechleider (9). Semiconfluent HSC in six-well cell culture plates were transiently transfected by using the LipofectAMINE reagent (Life Technologies, Grand Island, NY). Each sample (3 μg DNA/well) had triplicate in every experiment. Luciferase assays were performed as previously described (5). Transfection efficiency was determined by cotransfection of a β-galactosidase reporter, pSV-β-gal (0.5 μg/well) (Promega). β-Galactosidase activity was measured by a chemiluminescence assay kit (Tropix, Bedford, MA), according to the manufacturer's instructions. Results were combined from three independent experiments.

Site-directed mutageneses.

The plasmids pMut1, pMut2, and pMut1+2 were derived from pPPAR-γ-Luc with site-directed mutations in one of the two, or both Smad binding elements (SBEs). The site-directed mutageneses were prepared by Top Gene Technology (Montreal, QC, Canada). The sequence of 5′-GAC TAG TCT AGG CAA CAT GTC AAG ACA CAG T-3′ (−1522 to −1492), containing the two SBEs in pPPARγ-Luc, was respectively mutated to 1) 5′-GAC TAa TaT AGG CAA CAT GTC AAG ACA CAG T-3′ in pMut1, 2) 5′-GAC TAG TCT AGG CAA CAT GTC A At At A CAG T-3′ in pMut2, and 3) 5′-GAC TAa TaT AGG CAA CAT GTC A At At A CAG T-3′ in pMut1+2. The mutations were confirmed by DNA sequencing.

EMSA.

Preparation of nuclear extracts, evaluation of their integrity and performance of electrophoretic mobility shift assay (EMSA) were previously described (4). The following oligonucleotides, synthesized by INVITROGEN (Carlsbad, CA), were used as probes in EMSA. Letters with underline in P(ppar) and P(smad) are consensus SBEs. The sequence of the probe P(smad) contains the binding elements for the complex of Smad proteins originally from the promoter of human PAI-1 gene (10). The mutant probes P(M1), P(M2), and P(M1+2) are derived from the parental probe P(ppar) with side-directed mutations in one or both of the SBEs. P(ppar): 5′-GAC TAG TCT AGG CAA CAT GTC AAG ACA CAG T-3′; P(smad): 5′-GTG TCT GGC TAA ATG TCT GGC TTT TTG TCT GGC TCT CGA-3′; P(M1): 5′-GAC TAa TaT AGG CAA CAT GTC AAG ACA CAG T-3′; P(M2): 5′-GAC TAG TCT AGG CAA CAT GTC A At At A CAG T-3′; P(M1+2): 5′-GAC TAa TaT AGG CAA CAT GTC A At At A CAG T-3′.

RNA isolation and real-time PCR.

Total RNA was isolated by TRI-Reagent (Sigma), following the protocol provided by the manufacturer. Real-time PCR was carried out as recently described (6). mRNA fold changes of target genes relative to the endogenous GAPDH control were calculated as suggested by Schmittgen et al. (32). PPAR-γ primers used in real-time PCR were (F): 5′-ATT CTG GCC CAC CAA CTT CGG-3′; (R): 5′-TGG AAG CCT GAT GCT TTA TCC CCA-3′. Other primers were recently described (37, 39).

Statistical analysis.

Differences between means were evaluated by an unpaired two-sided Student's t-test (P < 0.05 considered as significant). Where appropriate, comparisons of multiple treatment conditions with controls were analyzed by ANOVA with the Dunnett's test for post hoc analysis.

RESULTS

Exogenous active TGF-β1 suppresses gene expression of PPAR-γ in passaged HSC.

To test the hypothesis that TGF-β signaling might negatively regulate PPAR-γ gene expression in HSC, passaged HSC were transfected with the PPAR-γ promoter luciferase reporter plasmid pPPAR-γ-Luc, which contains the 5′-flanking region of the PPAR-γ gene promoter (∼2.8 kb) in a luciferase reporter plasmid (11). Cells were subsequently treated with exogenous active TGF-β1 at indicated concentrations for 24 h. As shown in Fig. 1A by luciferase assays, exogenous TGF-β1 dose dependently reduced luciferase activity in these cells, suggesting the reduction of the promoter activity of PPAR-γ gene. To verify this observation, passaged HSC were treated with exogenous TGF-β1 at indicated concentrations for 24 h. Total RNA and whole cell extracts were prepared from these cells. As shown in Fig. 1, B and C, by real-time PCR and Western blotting analyses, respectively, exogenous TGF-β1 dose dependently reduced the abundance of PPAR-γ at levels of mRNA and protein in passaged HSC. However, exogenous TGF-β1 had no significant impact on gene expression of Tβ-RI and Tβ-RII in passaged HSC (Fig. 1, B and C). Because of a relatively low level of PPAR-γ in activated HSC, a rather large amount of whole cell proteins was used for Western blotting analyses (60 μg/well). These results collectively indicated that exogenous TGF-β1 suppressed gene expression of PPAR-γ in passaged HSC.

Fig. 1.

Exogenous active transforming growth factor-β1 (TGF-β1) suppresses gene expression of peroxisome proliferator-activated receptor-γ (PPAR-γ) in passaged hepatic stellate cells (HSC). Passaged HSC were treated with exogenous active TGF-β1 at indicated concentrations for 24 h. A: luciferase assays of cells transfected with pPPAR-γ-Luc. Luciferase activity in cells was expressed as relative units after β-galactosidase normalization. Values were expressed as means ± SD (n = 6). *P < 0.05 vs. cells with no treatment. The floating schema denotes the pPPAR-γ-Luc luciferase reporter construct in use and the application of TGF-β1 to the system. B: real-time PCR assays of the mRNA abundance of PPAR-γ and types I and II TGF-β receptors (Tβ-RI and Tβ-RII, respectively) (n = 3). GAPDH was used as an invariant control for calculating mRNA fold changes. C: Western blotting analyses of the protein levels of PPAR-γ, Tβ-RI, and Tβ-RII in whole cell extracts (n = 3). β-Actin was used as an internal control for equal loading. Representatives were shown here from 3 independent experiments.

Different treatment protocols with curcumin result in different effects of exogenous TGF-β1 on regulating the gene promoter activity and trans-activation activity of PPAR-γ in passaged HSC.

We recently observed that the treatment of passaged HSC with curcumin for 24 h significantly reduced gene expression of TGF-β receptors, leading to the interruption of the TGF-β signal pathway (39). Results in Fig. 1 demonstrated that TGF-β signaling suppressed gene expression of PPAR-γ in passaged HSC. It is plausible to assume that the disruption of TGF-β signaling by curcumin could result in the induction of gene expression of PPAR-γ in activated HSC. To evaluate the assumption, HSC were transfected with the PPAR-γ promoter luciferase reporter plasmid pPPAR-γ-Luc. Cells were then divided into two groups. One group was simultaneously treated with curcumin at 20 μM plus exogenous TGF-β1 at various concentrations for 24 h. The other group was pretreated with curcumin at 20 μM for 24 h before the addition of exogenous TGF-β1 at various concentrations for an additional 24 h. As shown in Fig. 2A, compared with the untreated control (the 1st column on the left side), curcumin itself increased, as expected, luciferase activity (the 2nd column on the left side), which was significantly eliminated by simultaneous addition of exogenous TGF-β1 in a dose-dependent manner. In great contrast, postaddition of TGF-β1 showed little impact on luciferase activity in cells pretreated with curcumin for 24 h (Fig. 2B). Taken together, these results demonstrated the differential effects of exogenous TGF-β1 on regulating the promoter activity of PPAR-γ gene in passaged HSC, mainly depending on whether cells were pretreated with curcumin.

Fig. 2.

Different treatment protocols with curcumin result in different effects of exogenous TGF-β1 on regulating the gene promoter activity and the trans-activation activity of PPAR-γ in passaged HSC. Passaged HSC were transiently transfected with pPPAR-γ-Luc (A and B) or pPPRE-TK-Luc (C and D). Cells in A and C were simultaneously treated with curcumin at 20 μM plus exogenous TGF-β1 at indicated concentrations for 24 h (TGF-β1 + curcumin). Cells in B and D were pretreated with curcumin at 20 μM for 24 h before the addition of exogenous TGF-β1 at indicated concentrations for an additional 24 h (curcumin/TGF-β1). Luciferase activity was expressed as relative units after β-galactosidase normalization. Values were expressed as means ± SD (n = 6). *P < 0.05 vs. cells with no treatment (left 1st column); **P < 0.05 vs. cells with curcumin at 20 μM, without exogenous TGF-β1 (left 2nd column). The floating diagram denotes the construct of pPPAR-γ-Luc, or pPPRE-TK-Luc, in use and the application of curcumin and TGF-β1 to the system.

To verify the differential effects of exogenous TGF-β1, passaged HSC were transfected with the PPAR-γ activity reporter plasmid pPPRE-TK-Luc. This plasmid contains three copies of PPRE in a luciferase reporter vector. The transfected cells were similarly divided into the two groups. As shown in Fig. 2C, simultaneous addition of exogenous TGF-β1 with curcumin caused a dose-dependent and significant reduction in luciferase activity, suggesting that exogenous TGF-β1 might reduce the trans-activation activity of PPAR-γ. However, postaddition of exogenous TGF-β1 to cells pretreated with curcumin for 24 h showed little effect on luciferase activity (Fig. 2D). PPAR-γ agonists are presumed to exist in the media with 10% of FBS (20, 25, 39). Therefore, the increase in the expression of PPAR-γ could lead to the enhancement of the PPAR-γ activity. Taken together, these results demonstrated that exogenous TGF-β1 showed different effects on regulating the gene promoter activity and the trans-activation activity of PPAR-γ in passaged HSC depending on when the cells were treated with curcumin.

Different treatment protocols with curcumin show different effects of exogenous TGF-β1 on the abundance of TGF-β receptors and PPAR-γ in HSC.

Curcumin reduces gene expression of TGF-β receptors, leading to the interruption of the TGF-β signal pathway (39). We, therefore, hypothesized that the differential effects of exogenous TGF-β1 observed in Fig. 2 might result from the interruption of TGF-β signaling by curcumin. It is, thus, expected that different treatment protocols, i.e., simultaneousness or pretreatment with curcumin, might result in different roles of exogenous TGF-β1 in regulating the level of TGF-β receptors, leading to different impacts on the abundance of PPAR-γ. To test the hypothesis, HSC were similarly divided into two groups. One group was simultaneously treated with curcumin (20 μM) plus exogenous TGF-β1 at various concentrations for 24 h. The other group was pretreated with curcumin (20 μM) for 24 h before the addition of exogenous TGF-β1 at various concentrations for an additional 24 h. As shown in Fig. 3, A and B, by Western blotting analyses, compared with the untreated control (the 1st lane on the left), curcumin, as expected, reduced the level of type I and II TGF-β receptors and increased the abundance of PPAR-γ in passaged HSC (the 2nd lane on the left). Simultaneous addition of exogenous TGF-β1 with curcumin dose dependently eliminated these effects by elevating the level of TGF-β receptors and reducing the abundance of PPAR-γ (Fig. 3A). In great contrast, the subsequent addition of exogenous TGF-β1 showed little, if any, impact on the abundance of TGF-β receptors and PPAR-γ in cells pretreated with curcumin for 24 h (Fig. 3B). These results collectively demonstrated that exogenous TGF-β1 showed differential effects on regulation of gene expression of these receptors mainly depending on different treatment protocols with curcumin. Together with the experiments in Fig. 2, these results supported our assumption that the disruption of TGF-β signaling by curcumin might result in, at least partially, the induction of gene expression of PPAR-γ in activated HSC.

Fig. 3.

Different treatment protocols with curcumin show different effects on the abundance of TGF-β receptors and PPAR-γ in HSC. Passaged HSC were simultaneously treated with curcumin at 20 μM plus exogenous TGF-β1 at indicated concentrations for 24 h (A), or pretreated with curcumin at 20 μM for 24 h before the addition of exogenous TGF-β1 at indicated concentrations for an additional 24 h (B). Whole cell extracts were prepared for Western blotting analyses (n = 3). β-Actin was used as an internal control for equal loading. Representatives were shown here from 3 sets of independent experiments.

The interruption of TGF-β signaling by dominant negative Tβ-RII stimulates the promoter activity of PPAR-γ gene in passaged HSC.

To further verify the role of TGF-β signaling in regulating gene expression of PPAR-γ, passaged HSC were cotransfected with the plasmid pPPAR-γ-Luc and the plasmid pdn-Tβ-RII at indicated concentrations. pdn-Tβ-RII contains the fragment of cDNA encoding the dominant negative form of type II TGF-β receptor (dn-Tβ-RII) (9). To determine the role of dn-Tβ-RII in interruption of TGF-β signaling, pilot experiments were performed. HSC in six-well cell culture plates were cotransfected with the TGF-β-inducible luciferase reporter plasmid p3TP-Lux with pdnTβ-RII at indicated doses. A total of 4.5 μg of plasmid DNA was used in each well for transfection, including 2 μg of p3TP-Lux, 0.5 μg of pSV-β gal, pdn-Tβ-RII at indicated doses, and the empty vector pcDNA. The latter was used to ensure the equal amount of total DNA in cotransfection assays. The amount of DNA of pdn-Tβ-RII plus pcDNA was equalized to 2 μg. Cells were then treated with or without exogenous TGF-β1 (10 ng/ml), or curcumin (20 μM), for 24 h. As shown in Fig. 4A, forced expression of dnTβ-RII cDNA dose dependently reduced luciferase activity, suggesting that dn-Tβ-RII might interrupt TGF-β signaling in passaged HSC. In addition, exogenous TGF-β1 significantly increased luciferase activity in cells cotransfected with p3TP-Lux, but without pdnTβ-RII. In great contrast, exogenous TGF-β1 could not stimulate luciferase activity in cells cotransfected with p3TP-Lux plus pdn-Tβ-RII (2 μg). Taken together, these results suggested that like curcumin (the last column on the right side), forced expression of dnTβ-RII cDNA might interrupt TGF-β signaling in passaged HSC.

Fig. 4.

Forced expression of dn-Tβ-RII cDNA interrupts TGF-β signaling, leading to the stimulation of the promoter activity of PPAR-γ gene in passaged HSC. Passaged HSC in 6-well cell culture plates were cotransfected with p3TP-Lux (A) or pPPAR-γ-Luc (B), plus pdn-Tβ-RII, containing dominant negative Tβ-RII cDNA in a cytomegalovirus-driven expression vector. A total of 4.5 μg of plasmid DNA was used in each well for transfection, including 2 μg of p3TP-Lux or pPPAR-γ-Luc, 0.5 μg of pSV-β gal, pdn-Tβ-RII at indicated doses, and the empty vector pcDNA. The latter was used to ensure the equal amount of total DNA in transfection assays. The amount of DNA of pdn-Tβ-RII plus pcDNA was equalized to 2 μg. Cells were treated with or without TGF-β1 at 10 ng/ml, or curcumin at 20 μM, for 24 h. Luciferase activity was expressed as relative units after β-galactosidase normalization. Values were expressed as means ± SD (n = 6). *P < 0.05 vs. cells cotransfected without pdn-Tβ-RII (left 1st column in the A and B). ‡P < 0.05 vs. cells cotransfected with no pdn-Tβ-RII, but treated with TGF-β1 (3rd column on the right side in the A). The floating diagram denotes the construct of p3TP-Lux, or pPPAR-γ-Luc, in use and forced expression of dn-Tβ-RII in the system.

To verify the role of disruption of TGF-β signaling in the induction of the gene promoter activity of PPAR-γ, passaged HSC were cotransfected with the PPAR-γ promoter luciferase reporter plasmid pPPAR-γ-Luc and pdnTβ-RII at indicated doses. Cells were treated with or without curcumin at 20 μM for 24 h. As shown in Fig. 4B by luciferase assays, forced expression of dn-Tβ-RII mimicked the stimulatory role of curcumin and caused a dose-dependent increase in luciferase activity. These results confirmed that the interruption of TGF-β signaling by overexpression of dominant negative Tβ-RII stimulated the gene promoter activity of PPAR-γ in passaged HSC, which further supported our hypothesis that TGF-β signaling might negatively regulate gene expression of PPAR-γ in activated HSC.

The two Smad binding elements in the PPAR-γ gene promoter serve as curcumin response elements and negatively regulate the promoter activity of PPAR-γ gene.

Promoter deletion assays were conducted to localize curcumin response element(s) in the promoter of PPAR-γ gene and to further elucidate the molecular mechanisms of curcumin in the induction of PPAR-γ gene expression. Passaged HSC were transfected with plasmids containing various sizes of the PPAR-γ gene promoter (Fig. 5A). After transfection, cells were treated with or without curcumin at 20 μM for 24 h. As shown in Fig. 5A by luciferase assays, curcumin significantly increased luciferase activity by 40 and 52.8% in cells transfected with the reporter plasmids with the promoter in the size of 2776 and 1843, respectively. However, loss of the fragment of nucleotides between −1843 and −908 in pPPAR-γ1000-Luc failed in responding to curcumin, suggesting the presence of curcumin response element(s) in this DNA fragment.

Fig. 5.

The 2 Smad binding elements (SBEs) in the PPAR-γ gene promoter serve as curcumin response elements and act as cis-inhibitory elements in regulating the promoter activity of PPAR-γ gene. Semiconfluent passaged HSC were transiently transfected with luciferase reporter plasmids with various lengths of the 5′-flanking region of PPAR-γ gene promoter. Luciferase activity was expressed as relative units after β-galactosidase normalization. Values were expressed as means ± SD (n = 6). *P < 0.05 vs. cells transfected with the same plasmid without indicated treatments. A: luciferase assays of cells transfected with plasmids containing various lengths of wild-type PPAR-γ gene promoter. Cells were subsequently treated with or without curcumin at 20 μM for 24 h. The numbers next to the bars are the increase in luciferase activity caused by curcumin. B: luciferase assays of HSC transfected with the parental plasmid pPPAR-γ-Luc (i.e., pPPAR), or mutant plasmids derived from pPPAR with site-directed mutations in one or both of the SBEs. Cells were subsequently treated with or without curcumin at 20 μM for 24 h. The numbers next to the bars are the increase in luciferase activity caused by the site-directed mutations in the distal SBE without curcumin, or by curcumin, respectively. C: luciferase assays of HSC transfected with the parental plasmid pPPAR, or mutant plasmids derived from pPPAR with site-directed mutations in 1 or both of the SBEs. Cells were then treated with or without exogenous TGF-β1 (10 ng/ml) for 24 h. The numbers next to the bars are the reduction in luciferase activity caused by exogenous TGF-β1.

Computer analyses of the DNA fragment (−1843 to −908) revealed two potential SBEs, i.e., GTCT and AGAC, within −1517 to −1514 and −1500 to −1497, respectively. GTCT and its palindrome AGAC were originally known to bind the Smad protein complex and to mediate TGF-β signaling in regulation of target gene expression (34, 38). To evaluate the role of the two SBEs in the curcumin induction of the gene expression, the plasmids pMut1, pMut2, and pMut1+2 with site-directed mutations in one or both of the two SBEs were generated from the parental plasmid pPPAR-γ-Luc. Passaged cells were transfected with these plasmids. Luciferase assays in Fig. 5B demonstrated that curcumin significantly increased, as expected, luciferase activity in cells transfected with wild-type pPPAR-γ-Luc (i.e., pPPAR). However, compared with pPPAR, the mutant pMut1 and pMut2 with mutations in one of the two SBEs showed a significant reduction in the induction of luciferase activity in response to curcumin (Fig. 5B). For example, compared with pPPAR (Fig. 5, A and B), mutations in the distal SBE in pMut1 resulted in the increase in luciferase activity from ∼40 to ∼23% in response to curcumin (Fig. 5B). A similar reduction in the induction of luciferase activity in response to curcumin was observed in pMut2 with mutations in the proximal SBE. Mutations in both of the SBEs failed in responding to curcumin in stimulating luciferase activity (Fig. 5B). Taken together, these results suggested that the SBEs might serve as curcumin response elements in regulating the promoter activity of PPAR-γ in HSC.

In addition, it is of interest to observe that mutations in one of the SBEs resulted in a significant increase in luciferase activity in cells with no curcumin treatment. For example, compared with the parental pPPAR, the mutant pMut1 with mutations in the distal SBE increased luciferase activity in cells without curcumin by ∼78% (Fig. 5B). There was no significant difference between pMut1 and pMut2 in the increase in luciferase activity in cells without curcumin. Mutations in both of the SBEs in pMut1+2 led to a further increase in luciferase activity in cells without curcumin, suggesting an additive role of the two SBEs. These results collectively suggested that the two SBEs might also act as cis-inhibitory elements in regulating the promoter activity of PPAR-γ gene (Fig. 5B).

To further verify the critical role of the two SBEs in regulating PPAR-γ gene promoter activity, HSC were transfected with PPAR-γ promoter luciferase reporter plasmids with or without site-directed mutations in one or both of the SBEs. Cells were subsequently treated with or without exogenous TGF-β1 for 24 h. As shown in Fig. 5C by luciferase assays, exogenous TGF-β1 significantly reduced luciferase activity by ∼39% in cells transfected with wild-type pPPAR. However, TGF-β1 showed a partial suppression on luciferase activity in cells transfected with in pMut1 by ∼27%. A similar partial suppression was also observed in cells transfected with pMut2 with mutations in the proximal SBE. TGF-β1 had no impact on luciferase activity in cells transfected with pMut1+2 with mutations in both of the SBEs. These results confirmed our above results and collectively indicated the pivotal role of the SBEs, as cis-inhibitory elements, in regulating the promoter activity of PPAR-γ in passaged HSC. In addition, these results further supported our assumption that the interruption of TGF-β signaling by curcumin might stimulate the promoter activity and induce gene expression of PPAR-γ in activated HSC in vitro. Taken together, our results suggested that both of the SBEs in the PPAR-γ gene promoter were the curcumin response elements, acting as the cis-inhibitory elements in regulating the promoter activity of PPAR-γ gene in HSC.

The complex of Smad proteins specifically binds to the curcumin response elements in the promoter of PPAR-γ gene.

Further experiments of EMSA were performed to identify factors that bound to the curcumin response elements in the promoter of PPAR-γ gene. Nuclear protein extracts were prepared from passaged HSC treated with or without curcumin at indicated concentrations for 24 h. A series of EMSA were conducted using the 32P-labeled oligonucleotide probe P(ppar) containing the two SBEs identified in the promoter of PPAR-γ gene. As shown in Fig. 6A, a clear and strong protein-probe complex was observed in nuclear extracts from passaged HSC (lane 1). The DNA binding activity of the protein(s) to the DNA probe was dramatically reduced by curcumin in a dose-dependent manner (lanes 25). Curcumin at both concentrations of 20 and 50 μM caused an apparent reduction in the density of the DNA-protein complex. We previously showed that curcumin caused a dose-dependent reduction in cell numbers by inducing cell growth arrest and apoptosis (37). Curcumin up to 100 μM is not toxic to cultured HSC (37). Since curcumin at 50 μM causes more cell death than that at 20 μM, it would be more difficult to harvest enough viable HSC for preparation of nuclear extracts for subsequent gel shift assays. In addition, prior animal studies indicated that the absorptive rate of dietary curcumin was relatively low (30). A lower concentration in in vitro experiments might be closer to the in vivo physical concentration. On the basis of these reasons, we chose curcumin at 20 μM for most of our experiments.

Fig. 6.

Curcumin significantly reduces the DNA binding activity of the smad protein complex to the curcumin response elements found in the gene promoter of PPAR-γ in passaged HSC. Passaged HSC were treated with curcumin at indicated concentrations for 24 h. Nuclear protein extracts were prepared for electrophoretic mobility shift assay (EMSA) by using the 32P-labeled probe P(ppar) with the 2 SBEs found in the PPAR-γ gene promoter. Representatives of EMSA were shown from 3 independent experiments. A: EMSA of nuclear protein extracts from HSC treated with various concentrations of curcumin. B: competition assays of nuclear protein extracts from HSC treated with or without curcumin (Cur) at 20 μM using a 10-, 25-, or 50-fold excess of the unlabeled P(ppar) (lanes 35), or the unlabeled probe P(smad) (lanes 68). C: supershift assays of nuclear protein extracts from HSC treated with or without curcumin at 20 μM using 2 μl of anti-Smad2, 3, or 4 antibodies (Santa Cruz Biotech).

To examine the DNA binding specificity of the protein(s) to the probe, as shown in Fig. 6B, competition assays were performed using a 10-, 25-, or 50-fold excess of the unlabeled probe P(ppar) (lanes 35), or the unlabeled probe P(smad) (lanes 68). The probe P(smad) contains three consensus SBEs observed in the promoter of human PAI-1 gene (10). It was found that the amount of the protein(s) binding to the 32P-labeled probe P(ppar) was competitively reduced by the cold probe P(ppar) or P(smad), respectively, in a dose-dependent manner. These results suggested that the protein(s) specifically bound to the two SBEs in the probe P(ppar).

To clarify the protein(s) that bound to the SBEs in the promoter of PPAR-γ gene, supershift assays were performed by adding anti-Smad2, 3, or 4 antibodies to the nuclear protein extracts. As demonstrated in Fig. 6C, anti-Smad 2 antibodies had no apparent impact on the probe-protein complex and caused no supershift. On the other hand, anti-Smad3 or 4 antibodies significantly diminished the probe-protein complex and caused an apparent supershift, respectively. These results suggested that the DNA binding complex might contain Smad3 and 4, but not Smad2.

Additional competition assays were performed using a 10-, 25-, or 50-fold excess of the unlabeled probe P(M1) with mutations in the distal SBE, or the probe P(M2) with mutations in the proximal SBE, or the probe P(M1+2) with mutations in both of the SBEs. As shown in Fig. 7A, the cold mutant probes P(M1) (lanes 35) and P(M2) (lanes 68) competed for the DNA binding complex with the 32P-labeled probe P(ppar) in a dose-dependent manner. In great contrast, the mutant probe P(M1+2) showed no apparent ability to compete with the hot wild-type probe (lanes 911). These results indicated that mutations in any one of the two SBEs could not stop the competition of the cold mutant probe with the hot wild-type probe for the DNA binding complex. In addition, these results suggested that both of the SBEs might share the same DNA binding complex, which contains Smad3 and 4 demonstrated in Fig. 6C.

Fig. 7.

The 2 SBEs in the promoter of PPAR-γ gene share the same DNA binding complex. Passaged HSC were treated with or without curcumin at 20 μM for 24 h. Nuclear protein extracts were prepared for competition assays in EMSA. Representatives of EMSA were shown from 3 independent experiments. A: a 10-, 25-, or 50-fold excess of the unlabeled P(M1), P(M2), or P(M1+2), competed, respectively, with the 32P-labeled probe P(ppar) for the DNA binding complex. B: a 10-, 25-, or 50-fold excess of the unlabeled P(M1), P(M2), or P(M1+2), competed, respectively, with the 32P-labeled probe P(M1) for the DNA binding complex.

Mutant probes P(M1), P(M2), and P(M1+2) were respectively 32P-labeled and used in additional competition assays of EMSA. As shown in Fig. 7B, like the wild-type probe P(ppar) in Fig. 6B, the 32P-labeled mutant probe P(M1) with mutations in the distal SBE formed an apparent complex with DNA binding proteins in EMSA (lane 1). Curcumin significantly reduced the level of the complex binding to the mutant probe (lane 2). The unlabeled probe P(M1) or P(M2) competed for the DNA binding complex in a dose-dependent manner (lanes 35 or 68, respectively). However, the cold probe P(M1+2) with mutations in both of the SBEs could not apparently compete for the DNA binding complex (lanes 911). Similar results were observed in competition assays using 32P-labeled probe P(M2) (data not shown). In addition, the 32P-labeled probe P(M1+2) with mutations in both of the SBEs showed, as expected, no complex with the DNA binding proteins (data not shown). Taken together, these results further suggested that the same DNA binding complex might specifically bind to both of the SBEs in the PPAR-γ gene promoter. However, our results could not tell whether the proteins equally bind to the two SBEs. In summary, our results suggested that the DNA binding complex, containing Smad3 and 4, specifically bound, presumably as a trans-inhibiting factor, to the two SBEs, as the cis-inhibiting elements, in the PPAR-γ gene promoter. The activity of the DNA binding complex was significantly inhibited by curcumin in passaged HSC.

Curcumin reduces the abundance of Smad4 in HSC, leading to the interruption of TGF-β signaling.

To further elucidate the mechanism of curcumin in the interruption of TGF-β signaling, we hypothesized that curcumin might inhibit the activity of the DNA binding complex in HSC by reducing the abundance of Smad3 or/and Smad4, in addition to reducing the level of TGF-β receptors. To test the hypothesis, passaged HSC were treated with curcumin at the indicated concentrations for 24 h. As shown in Fig. 8A by Western blotting analyses, curcumin significantly reduced the abundance of Smad4. However, the effect of curcumin on the level of Smad3, if any, is not dramatic.

Fig. 8.

Curcumin interrupts TGF-β signaling in HSC at least partially by reducing the abundance of Smad4. A: Western blotting analyses of the protein levels of Smad3 and Smad4 in whole cell extracts treated with curcumin at the indicated concentrations for 24 h (n = 3). β-Actin was used as an internal control for equal loading. Representatives were shown here from 3 independent experiments. B: transfection assays of HSC in 6-well cell culture plates cotransfected with p3TP-Lux, or pPPAR-γ-Luc, plus psmad4-cDNA. The cDNA expression plasmid psmad4-cDNA contains a full size of Smad4 cDNA in a cytomegalovirus-driven expression vector. A total of 3.5 μg of plasmid DNA was used in each well for transfection, including 2 μg of p3TP-Lux or pPPAR-γ-Luc, 0.5 μg of pSV-β gal, psmad4cDNA at indicated doses, and the empty vector pcDNA. The latter was used to ensure the equal amount of total DNA in transfection assays. The amount of DNA of psmad4cDNA plus pcDNA was equalized to 1 μg. After recovery, cells were serum starved for 24 h before the addition of FBS (10%) and curcumin (20 μM) for 24 h. Luciferase activity was expressed as relative units after β-galactosidase normalization. Values were expressed as means ± SD (n = 6). *P < 0.05 vs. cells cotransfected without psmad4cDNA (left 1st column). ‡P < 0.05 vs. cells cotransfected with no psmad4cDNA, but treated with curcumin (left 3rd column). The floating diagram denotes the construct of p3TP-Lux, or pPPAR-γ-Luc, in use and forced expression of pSmad4 in the system.

To verify the role of the reduction in the level of Smad4 in the curcumin interruption of TGF-β signaling, and in the inhibition of PPAR-γ gene promoter activity, HSC in six-well plates were cotransfected with the TGF-β-inducible luciferase reporter plasmid p3TP-Lux, or the PPAR-γ promoter reporter plasmid pPPAR-γ-Luc, and the Smad4 cDNA expression plasmid pSmad4. The latter contains a full length of Smad4 cDNA in a cytomegalovirus-driven plasmid (9). A total of 3.5 μg of plasmid DNA was used in each well for cotransfection, including 2 μg of p3TP-Lux, or pPPAR-γ-Luc, 0.5 μg of pSV-β gal and 1 μg of pSmad4 at indicated doses plus the empty vector pcDNA. The latter was used to ensure the equal amount of total DNA in cotransfection assays. As shown in Fig. 8B, compared with the controls (the 1st columns on the left side), forced expression of Smad4 cDNA significantly increased luciferase activity in cells transfected with p3TP-Lux (the black one in the 2nd columns on the left), and reduced luciferase activity in cells transfected with pPPAR-γ-Luc (the white one in the 2nd columns on the left). These results collectively suggested that forced expression of Smad4 might stimulate TGF-β signaling and suppress the gene promoter activity of PPAR-γ. Compared with the controls (the 1st columns on the left), curcumin, as expected, interrupted TGF-β signaling and stimulated the promoter activity of PPAR-γ, respectively (the 3rd columns on the left). Further experiments in Fig. 8B observed that forced expression of Smad4 cDNA dose dependently eliminated the inhibitory effect of curcumin on the TGF-β signaling pathway and on the gene promoter activity of PPAR-γ. These results collectively demonstrated the role of reducing the abundance of Smad4 by curcumin in the interruption of TGF-β signaling and the induction of the promoter activity of PPAR-γ in HSC.

DISCUSSION

HSC activation is coupled with the stimulation of TGF-β signaling and the reduction in gene expression and trans-activation activity of PPAR-γ. In the normal liver, TGF-β1 is mainly produced by Kupffer cells. In contrast, in the fibrotic liver, the level of TGF-β1 is selectively increased in activated HSC (8). Persistent stimulation of HSC by TGF-β1 plays a key role in hepatic fibrogenesis (2, 31). The activation of TGF-β signaling in passaged HSC might be stimulated by binding of TGF-β1 secreted from cultured HSC, as well as from FBS (10%) in the media, to TGF-β receptors, the level of which is significantly increased in activated HSC (14). Our laboratory recently showed (37, 39) that curcumin significantly induced gene expression of PPAR-γ in activated HSC in vitro, which facilitated the trans-activation activity of PPAR-γ. The latter was required for curcumin to suppress gene expression of TGF-β receptors and to interrupt TGF-β signaling in HSC (37, 39). These observations prompted us to assume an antagonistic relationship between PPAR-γ activation and TGF-β signaling in HSC, i.e., induction of TGF-β signaling might suppress PPAR-γ gene expression during activation of quiescent HSC, whereas activation of PPAR-γ might interrupt TGF-β signaling, leading to the inhibition of HSC activation. Our recent results suggested that the induction of gene expression of PPAR-γ by curcumin and activation of PPAR-γ interrupted TGF-β signaling likely by reducing gene expression of TGF-β receptors in passaged HSC (39, 40), which supported our above assumption. In the present study, we further hypothesized that TGF-β signaling might negatively regulate PPAR-γ gene expression in HSC. Results presented in Fig. 1 supported our hypothesis and demonstrated that exogenous TGF-β1 suppressed gene expression of PPAR-γ in passaged HSC. Our results are consistent with prior other observations. TGF-β showed an early stimulation and late inhibition of PPAR-γ gene expression in human aortic smooth muscle cells (15). Transcription factors Egr-1, AP-1, and Smads are involved in the regulation of PPAR-γ gene expression (15). In addition, TGF-β could execute its inhibitory effect on PPAR-γ and inactivated its activity by phosphorylation (19).

Our recent studies demonstrated that the treatment of passaged HSC with curcumin for 24 h significantly reduced gene expression of TGF-β receptors, leading to the interruption of the TGF-β signal pathway (39). Since activation of TGF-β signaling by exogenous TGF-β1 represses gene expression of PPAR-γ (Fig. 1), it is expected that the interruption of TGF-β signaling by curcumin might relieve the inhibitory role and elevate gene expression of PPAR-γ in activated HSC. Exogenous TGF-β1 is, therefore, further expected to show different effects on inhibiting gene expression of PPAR-γ in activated HSC, depending on when cells are treated with curcumin. Simultaneous addition of exogenous TGF-β1 and curcumin is expected to have an apparent effect on reducing the promoter activity of PPAR-γ in activated HSC. Exogenous TGF-β1 would instantly initiate its signaling by binding to its receptors. At that time, the expression of TGF-β receptors should have not been reduced by curcumin yet. In contrast, pretreatment of cells with curcumin for 24 h would reduce gene expression of TGF-β receptors, leading to the reduction in the bioavailability of the receptors to the ligand and to the interruption of TGF-β signaling. Therefore, after the pretreatment with curcumin for 24 h, subsequent addition of exogenous TGF-β1 is expected to show a much weaker effect on inhibiting the gene promoter activity of PPAR-γ. Results presented in Figs. 2 and 3 met our expectation and demonstrated different effects of exogenous TGF-β1 on regulating gene expression of TGF-β receptors and PPAR-γ in passaged HSC depending on when the cells were treated with curcumin. Additional experiments confirmed the role of the interruption of TGF-β signaling by overexpression of dominant negative Tβ-RII in regulating the gene promoter activity of PPAR-γ in passaged HSC. These results collectively suggest the negative role of TGF-β signaling in the regulation of PPAR-γ gene expression.

To elucidate the molecular mechanisms of TGF-β signaling in the regulation of PPAR-γ gene expression, promoter deletion assays and transfection assays in Fig. 5 localized two SBEs as cis-inhibitory curcumin response elements in regulating the promoter activity of PPAR-γ in HSC. The repressive effects of SBE were previously reported. A TGF-β inhibitory element was characterized as essential to mediate TGF-β repressive effects in the gene promoter of stromelysin 1 (22). A recent study observed a novel SBE, termed a repressive Smad binding element, within the TGF-β inhibitory element of the c-myc promoter (12). Depending on protein cofactors, the Smad protein complex could act as either a transcription activator or a repressor. Further experiments of EMSA in Figs. 6 and 7 demonstrated that the DNA binding complex, containing Smad3 and 4, specifically bound, presumably as a trans-inhibiting factor, to the two SBEs, as the cis-inhibiting elements, in the PPAR-γ gene promoter. Prior other experiments demonstrated different roles of Smad2 and Smad3 in deposition of ECM components and cell proliferation in rat HSC (35). TGF-β signaling via Smad3 played an important role in the morphological and functional maturation of HSC. The results in EMSA (Figs. 57) collectively indicated that the activity of the DNA binding complex to SBEs was significantly inhibited by curcumin in passaged HSC. These results supported our hypothesis that TGF-β signaling might negatively regulate PPAR-γ gene expression.

We previously demonstrated the mechanisms of curcumin in the interruption of TGF-β signaling by reducing gene expression of TGF-β receptors (39, 40). Our result was consistent with prior other studies. Gaedeke et al. (16) reported that curcumin blocked TGF-β profibrotic actions on renal fibroblasts through downregulation of Tβ-RII. In the present report, we further observed that curcumin dose dependently reduced the level of Smad4 in passaged HSC (Fig. 8). Because of a limited life span of primary HSC (4–8 passages), it is unlikely to generate and maintain a stable transfectant of primary HSC with high expression of Smad4 to prepare nuclear extracts for gel shift assays. On the other hand, transient transfection efficacy in primary HSC is relatively low (∼15%) (unpublished observation). Therefore, it is impractical to directly show the role of overexpression of Smad4 in reversing the curcumin inhibitory effect on the DNA binding activity by EMSA. However, our results from cotransfection assays in Fig. 8B indirectly supported the assumption by demonstrating that overexpression of Smad4 eliminated the curcumin inhibitory effects on TGF-β signaling and on the gene promoter activity of PPAR-γ in HSC.

It remains obscure whether a natural polyphenolic component, such as curcumin, could block a signaling pathway by antagonistically interacting with its agonist(s) or a receptor, such as Tβ-RII. However, it was recently proposed that EGCG, a major and active polyphenol in green tea extracts, might incorporate into cell surface membrane and nonspecifically bind to the ligand platelet-derived growth factor (PDGF), which results in the reduction in the availability of PDGF binding to PDGF receptors and in the level of receptor tyrosine phosphorylation (36). Additional studies are necessary to study mechanism of curcumin in the interruption of signaling pathways.

The results in the present report demonstrate the role of TGF-β signaling in negatively regulating gene expression of PPAR-γ in passaged HSC and indicate that the interruption of TGF-β signaling by curcumin might stimulate gene expression of PPAR-γ in activated HSC. It bears emphasis that our results do not exclude the possible involvement of any other mechanisms in the curcumin induction of PPAR-γ gene expression in activated HSC. The present results and prior observations collectively support our initial assumption of the antagonistic relationship between PPAR-γ activation and TGF-β signaling in HSC, i.e., induction of TGF-β signaling suppresses PPAR-γ gene expression during activation of HSC, whereas activation of PPAR-γ interrupts TGF-β signaling, leading to the inhibition of HSC activation. The present results provide additional insights into the molecular mechanisms of curcumin in the induction of PPAR-γ gene expression and in the inhibition of HSC activation.

GRANTS

The work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant RO1 DK-047995.

Footnotes

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

REFERENCES

View Abstract