Ligands of peroxisome proliferator-activated receptor-gamma block activation of pancreatic stellate cells.

Activated pancreatic stellate cells (PSCs) have recently been implicated in the pathogenesis of pancreatic fibrosis and inflammation. Peroxisome proliferator-activated receptor gamma (PPAR-gamma) is a ligand-activated transcription factor which controls growth, differentiation, and inflammation in different tissues. Roles of PPAR-gamma activation in PSCs are poorly characterized. Here we examined the effects of PPAR-gamma ligands on the key parameters of PSC activation. PSCs were isolated from rat pancreas tissue, and used in their culture-activated, myofibroblast-like phenotype. Activation of PPAR-gamma was induced with 15-deoxy-Delta12,14-prostaglandin J2 (15d-PGJ2) or with troglitazone. Expression of PPAR-gamma was predominantly localized in the nuclei, and PPAR-gamma was transcriptionally active after ligand stimulation. PPAR-gamma ligands inhibited platelet-derived growth factor-induced proliferation. This effect was associated with inhibition of cell cycle progression beyond the G1 phase. PPAR-gamma ligands decreased alpha-smooth muscle actin protein expression and alpha1(I) procollagen and prolyl 4-hydroxylase(alpha) mRNA levels. Activation of PPAR-gamma also resulted in the inhibition of inducible monocyte chemoattractant protein-1 expression. 15d-PGJ2, but not troglitazone, inhibited the degradation of IkappaB-alpha and consequent NF-kappaB activation. In conclusion, activation of PPAR-gamma inhibited profibrogenic and proinflammatory actions in activated PSCs, suggesting a potential application of PPAR-gamma ligands in the treatment of pancreatic fibrosis and inflammation.

Recently, star-shaped cells in the pancreas, namely pancreatic stellate cells (PSCs), 1 have been identified and characterized (1,2). They are morphologically similar to the hepatic stellate cells that play a central role in the inflammation and fibrogenesis of the liver (3). There is accumulating evidence that PSCs, like hepatic stellate cells, are responsible for the development of pancreatic fibrosis (1,2,4). In the normal pancreas, stellate cells are quiescent and can be identified by the presence of vitamin A-containing lipid droplets in the cytoplasm. In response to pancreatic injury or inflammation, they are transformed ("activated") from their quiescent phenotype into highly proliferative myofibroblast-like cells. This process involves changes in cell morphology and gene expression and is characterized by the gradual loss of retinoid content, increased expression of the cytoskeletal protein ␣-smooth muscle actin (␣-SMA), and synthesis of type I collagen and other extracellular matrix components. Many of the morphological and metabolic changes associated with the activation of PSCs in animal models of fibrosis also occur when these cells are grown in culture on plastic (1,2). It has also been suggested that PSCs may participate in the pathogenesis of acute pancreatitis (1,5).
The peroxisome proliferator-activated receptor-␥ (PPAR-␥) is a member of the nuclear hormone receptor superfamily originally reported to be expressed at high levels in adipose tissue and to play a critical role in adipocyte differentiation (6,7). Ligands of PPAR-␥ include oxidative metabolites of polyunsaturated fatty acids and prostaglandins of the J series such as 15-deoxy-⌬ 12,14 -prostaglandin J 2 (15d-PGJ 2 ) (8). Although the precise enzymatic pathway leading to 15d-PGJ 2 generation is not completely understood, it has been shown that 15d-PGJ 2 is produced in intact cells and organisms (9) and is likely to represent a physiological ligand for PPAR-␥. PPAR-␥ is also activated by antidiabetic drugs of the thiazolidinedione group such as troglitazone (10). There is accumulating evidence that PPAR-␥ is implicated as an important regulator of inflammatory and immune responses (11,12). PPAR-␥ ligands were shown to inhibit the production of nitric oxide and macrophagederived cytokines, i.e. tumor necrosis factor-␣ (TNF-␣), interleukin (IL)-1, and IL-6, at least in part by antagonizing the activation of transcription factors such as activator protein-1 (AP-1) and nuclear factor B (NF-B) (11,12). They also induce apoptosis in several types of cells including macrophages (13), fibroblasts (14), and endothelial cells (15). Although PPAR-␥ gene expression is reportedly observed in a variety of tissues in addition to the adipose tissue, little is known about the pathophysiological relevance of PPAR-␥ activation in PSCs.
In this study, we examined the effects of PPAR-␥ ligands on the activation of PSCs: proliferation, expression of ␣-SMA, expression of ␣1(I) procollagen and prolyl 4-hydroxylase(␣) genes, and monocyte chemoattractant protein 1 (MCP-1) production. We show here that PPAR-␥ ligands inhibit these parameters of PSC activation, suggesting a potential application of PPAR-␥ ligands in the treatment of pancreatic fibrosis and inflammation.
Cell Culture-Rat PSCs were prepared as previously described using Nycodenz solution (2). All animal procedures were performed in accordance with the National Institutes of Health Animal Care and Use Guidelines. Isolated stellate cells were cultured in Ham's F-12 containing 10% fetal bovine serum (ICN Biomedicals, Aurora, OH), penicillin sodium, and streptomycin sulfate. All experiments were performed using cells between passages two and five. Cells were grown on 96-well flat-bottom plates for cell proliferation assay and enzyme-linked immunosorbent assay or on tissue culture dishes for other experiments. On the day of experiment, cells were refed with fresh medium containing 0.1% fetal bovine serum, incubated for 6 h, and treated with experimental reagents in the presence of 0.1% fetal bovine serum.
Immunostaining-We assessed PPAR-␥ expression in activated PSCs by immunohistochemical staining. Culture-activated PSCs were grown directly on glass coverslips, and immunostaining for PPAR-␥ was performed using a streptavidin-biotin-peroxidase complex detection kit (Histofine Kit; Nichirei, Tokyo, Japan) according to the manufacturer's instruction. Incubations were performed at room temperature unless otherwise specified. Cells were fixed with ice-cold methanol for 10 min, and then endogenous peroxidase activity was blocked by incubation in methanol with 0.3% hydrogen peroxide for 30 min. After immersion in normal rabbit serum for 1 h, the slides were incubated with rabbit polyclonal anti-PPAR-␥ antibody diluted at 1:100 in phosphate-buffered saline or rabbit immunoglobulin G at 4°C overnight. The slides were incubated with biotinylated goat anti-rabbit immunoglobulin antibody for 45 min followed by peroxidase-conjugated streptavidin for 30 min. Finally, color was developed by incubating the slides for several minutes with diaminobenzidine (Dojindo, Kumamoto, Japan). Expression of ␣-SMA, type I collagen, and prolyl 4-hydroxylase(␣) was examined in a similar manner.
Luciferase Assay-The luciferase expression vector containing three PPAR-responsive elements (16) was kindly provided by Dr. Takashi Osumi (Himeji Institute of Technology, Hyogo, Japan). The vector contains three copies of PPAR responsive element from the promoter of rat acyl coenzyme A oxidase. For the luciferase assay, ϳ1 ϫ 10 6 PSCs were transfected with 2 g of the luciferase expression vector along with 40 ng of pRL-TK vector (Promega, Madison, WI) as an internal control, using LipofectAMINE reagent (Life Technologies, Inc.). After 24 h, the transfected cells were treated with 15d-PGJ 2 (at 5 M), troglitazone (at 10 M), or vehicle (0.1% dimethyl sulfoxide) for an additional 24 h. At the end of the incubation, cell lysates were prepared using a Pica Gene kit (Toyo Ink Co., Tokyo, Japan), and the light intensities were measured using a model Lumat LB9507 luminescence reader (EG&G Berthold, Bad Wildbad, Germany).
Cell Proliferation Assay-PSCs (ϳ30% density) were treated with 15d-PGJ 2 or troglitazone at various concentrations for 1 h and then stimulated with PDGF (at 10 ng/ml), fetal bovine serum (at 5%), or epidermal growth factor (at 5 ng/ml) for 72 h. Cell proliferation was assessed using a commercial kit (CellTiter nonradioactive cell proliferation assay, Promega) according to the manufacturer's instruction. Cell viability was determined by differences in absorbance at wavelength 570 versus 690 nm.
Analysis of Cell Cycle-The cell cycle of PSCs was analyzed by flow cytometry. Briefly, serum-deprived PSCs (ϳ60 -70% density) were treated with 15d-PGJ 2 (5 M), troglitazone (10 M), or its vehicle for 1 h and then exposed to 10 ng/ml PDGF. After 24 h, cells were harvested and washed twice with phosphate-buffered saline. Cells were suspended in phosphate-buffered saline solution containing 40 g/ml propidium iodide, 0.02% Triton X-100, and 50 g/ml ribonuclease A. Samples were incubated in the dark at room temperature for 30 min and stored at 4°C until analysis. Cell fluorescence was measured by FAC-SCaliber flow cytometer (Becton Dickinson Co. Ltd., Tokyo, Japan) and analyzed using ModFit LT software (Verity Software House, Topsham, ME) to determine the distribution of cells in the various phases of the cell cycle.
Enzyme-linked Immunosorbent Assay-Confluent PSCs were treated with PPAR-␥ ligands at indicated concentrations for 30 min and then stimulated with IL-1␤ (at 10 ng/ml) or TNF-␣ (at 10 ng/ml). After 24 h, cell culture supernatants were harvested and stored at Ϫ80°C until measurement. MCP-1 levels in the supernatants were measured by enzyme-linked immunosorbent assay (Endogen) according to the manufacturer's instruction.
Northern Blotting-Total RNA was isolated using RNeasy total RNA preparation kit (Qiagen, Chatsworth, CA). Ten g of total RNA was separated on a 1% agarose-2.2 M formaldehyde gel and transferred to a nylon membrane filter (Amersham Biosciences, Inc.). Blots were hybridized for 16 h at 42°C to the 32 P-labeled DNA probes of MCP-1, ␣1(I) procollagen, and prolyl 4-hydroxylase(␣) generated by polymerase chain reaction. Specific primer sets were as follows (listed as 5Ј-3Ј, sense and antisense, respectively): MCP-1, AGCCAGATGCAGTTAATGCC and GGAAAAGAGAGTGGATGCAT; ␣1(I) procollagen, CCTGCTGGAC-CCCGAGGAAAC and TCACACCAGTATCACCAGGT; prolyl 4-hydroxylase(␣), TACTTCCTCAGTGTTCAGCC and CATCCAGAGTTCTGT-GTGGT. The PCR procedure consisted of 30 cycles at 94°C (for 1 min), at 55°C (for 1 min), and at 72°C (for 1 min). The identity of the reverse transcription-polymerase chain reaction was confirmed by direct sequencing. After the hybridization, the filter was washed three times with 2ϫ standard saline citrate (3 M NaCl, 0.3 M sodium citrate) and 0.1% SDS at 42°C for 10 min. The washed filter was subjected to autoradiography at Ϫ80°C overnight.
Nuclear Extract Preparation and Electrophoretic Mobility Shift Assay-Nuclear extracts were prepared, and an electrophoretic mobility shift assay was performed as described previously (17). Doublestranded oligonucleotide probes for NF-B (5Ј-AGTTGAGGGGACTT-TCCCAGGC-3Ј) and AP-1 (5Ј-CGCTTGATGAGTCAGCCGGAA-3Ј) were end-labeled with [␥-32 P]ATP. Nuclear extracts (ϳ5 g) were incubated with the labeled oligonucleotide probe for 20 min at 22°C and electrophoresed through a 4% polyacrylamide gel. Gels were dried and autoradiographed at Ϫ80°C overnight. A 100-fold excess of unlabeled oligonucleotide was incubated with nuclear extracts for 10 min prior to the addition of the radiolabeled probe in the competition experiments.
Western Blotting-PSCs were treated with experimental reagents and lysed in SDS buffer (62.5 mM Tris-HCl at pH 6.8, 2% SDS, 10% glycerol, 50 mM dithiothreitol, 0.1% bromphenol blue) for 15 min on ice. The samples were then sonicated for 2 s, heated for 5 min, and centrifuged at 12000 ϫ g for 5 min to remove insoluble cell debris. Whole cell extracts (ϳ100 g protein) were fractionated on a 10% SDS polyacrylamide gel (Bio-Rad) and transferred to a nitrocellulose membrane (Bio-Rad). The membrane was incubated with the first antibody at 4°C overnight. After incubation with the secondary antibody (horseradish peroxidase-conjugated), proteins were visualized using an ECL kit (Amersham Biosciences, Inc.).
Statistical Analysis-Differences between experimental groups were evaluated by the two-tailed unpaired Student's t test for other studies. A p value of less than 0.05 was considered statistically significant.

Activated PSCs Expressed PPAR-␥-It has been shown that
PSCs are activated (express ␣-SMA) after ϳ48 h of culture on plastic (2). As mentioned earlier, all experiments used passaged stellate cells, which were considered activated. Indeed, the fibrillary staining of ␣-SMA was observed by immunostaining (Fig. 1A). We examined PPAR-␥ expression by immunostaining. Culture-activated PSCs (on day 7, passage 2) showed positive staining of PPAR-␥ predominantly in the nuclei (Fig. 1B). No immunoreactivity was observed when nonspecific rabbit immunoglobulin G was used (data not shown). We also investigated whether PPAR-␥ exerts transcriptional effects in culture-activated PSCs. In cells transiently transfected with a luciferase reporter plasmid consisting of three copies of the PPAR-responsive element (16), exposure to 15d-PGJ 2 (at 5 M) and troglitazone (at 10 M) induced an ϳ5-fold increase in the activity of the reporter gene (Fig. 1C). These results indicated that PPAR-␥ is expressed and transcriptionally active after ligand stimulation in activated PSCs.

PPAR-␥ Ligands Inhibited Proliferation of Activated
PSCs-It has been shown that PPAR-␥ ligands modulate cell proliferation in several types of cells (13)(14)(15). We examined whether PPAR-␥ ligands could modulate proliferation of PSCs. Consistent with the previous report (18), PDGF induced an approximately 7-fold increase of cell proliferation in serum-free medium after 72 h ( Fig. 2A). PPAR-␥ ligands (15d-PGJ 2 or troglitazone) inhibited PDGF-induced cell proliferation in a dose-dependent manner ( Fig. 2A). The inhibitory effects were significant starting at 1 M for 15d-PGJ 2 and at 2.5 M for troglitazone. At 5 M 15d-PGJ 2 or 10 M troglitazone, the stimulation of cell proliferation by PDGF was virtually abolished. In these experiments, 15d-PGJ 2 and troglitazone up to these concentrations did not affect the cell viability during the incubation as assessed by a trypan blue exclusion test (data not shown). However, when PSCs were treated with PPAR-␥ ligands above these concentrations, cytotoxic effects were observed during the incubation. Treatment with a different peroxisome proliferator, WY-14643, at the concentration (at 10 M) able to activate the PPAR-␣ isoform did not inhibit PSC proliferation, but it did inhibit PSC proliferation at the concentration (at 100 M) also able to activate PPAR-␥ (19) (Fig. 2B). We also tested the effects of PPAR-␥ ligands on cell proliferation in response to epidermal growth factor or 10% fetal bovine serum.
Serum-induced cell proliferation was inhibited by PPAR-␥ ligands, although the inhibitory effects were less evident than in cells treated with PDGF (Fig. 2C). In addition, the proliferative response induced by epidermal growth factor was inhibited by PPAR-␥ ligands (Fig. 2C), indicating that PPAR-␥ ligands inhibit proliferation of PSCs independently of the mitogen used. PPAR-␥ Ligands Induced G 1 Arrest-We analyzed the cell cycle in PSCs in the presence or absence of PPAR-␥ ligands. Exposure to PDGF was associated with a marked decrease in the percentage of cells in the G 0 /G 1 phase together with an increase in the number of cells in the S phase (Fig. 3). The addition of the PPAR-␥ ligands before PDGF reduced the number of cells in the S phase, and the percentage of cells in the G 0 /G 1 phase was similar to the percentage observed in untreated cells. Thus, the addition of PPAR-␥ ligands inhibited PDGF-induced progression of the cell cycle beyond the G 1 phase.
PPAR-␥ Ligands Decreased Expression of ␣-SMA and Type I Collagen-It has been shown that culture-activated PSCs express ␣-SMA and produce type I collagen. Indeed, ␣-SMA expression has been accepted as a marker of PSC activation (2), and in situ hybridization techniques showed that ␣-SMA-positive cells were the principal source of collagen in the fibrotic pancreas (4). In agreement with the result of immunostaining, ␣-SMA expression was confirmed in culture-activated PSCs by Western blotting, and the treatment of PSCs with PPAR-␥ ligands for 48 h significantly reduced ␣-SMA expression (Fig.  4A). We also examined the effects of PPAR-␥ ligands on the expression of ␣1(I) procollagen and prolyl 4-hydroxylase(␣) genes, both of which play a central role in the collagen synthesis. Type I collagen immunoreaction was detected predominantly intracellularly with the highest intensity in the perinuclear region (Fig. 4B). Prolyl 4-hydroxylase(␣) is also a key enzyme that catalyzes the formation of 4-hydroxyproline, an essential residue for the folding of the procollagen polypeptide chains into triple helical molecules (21). Expression of prolyl 4-hydroxylase(␣) was observed strongly around the nuclei, and the expression spread toward the periphery of the cells (Fig.  4C). Steady-state mRNA levels of ␣1(I) procollagen and prolyl 4-hydroxylase(␣) were high in culture-activated PSCs, and the levels were significantly decreased by PPAR-␥ ligands after a 24-h incubation (Fig. 4D).

PPAR-␥ Ligands Inhibited MCP-1 Expression-Activated
PSCs acquire the proinflammatory phenotype, and they may modulate the recruitment and activation of inflammatory cells. One candidate may be MCP-1, a potent chemoattractant for monocytes and T lymphocytes (22). Proinflammatory cytokines IL-1␤ and TNF-␣ induced MCP-1 production in PSCs (Fig. 5A). Both of the PPAR-␥ ligands decreased the inducible MCP-1 expression in a dose-dependent manner (Fig. 5A). The inhibitory effects were significant starting at 1 M for both reagents. At concentrations as high as 5 M 15d-PGJ 2 or 10 M troglitazone, the MCP-1 induction was virtually abolished. We also examined the effects of PPAR-␥ ligands on the MCP-1 gene expression by Northern blotting. Both IL-1␤ and TNF-␣ increased the level of MCP-1 mRNA, but the effect was inhibited in the presence of PPAR-␥ ligands (Fig. 5B), suggesting that PPAR-␥ ligands inhibited MCP-1 expression at least in part at the transcriptional level.
15d-PGJ 2 , but Not Troglitazone, Inhibited Degradation of IB-␣ and Consequent NF-B Activation-Because activation of NF-B and AP-1 is important for MCP-1 expression in several types of cells (23,24), we investigated the effects of PPAR-␥ ligands on the activation of these transcription factors. We first examined the effect of PPAR-␥ ligands on NF-B binding activity by electrophoretic mobility shift assay. Two NF-B-spe- cific DNA-protein complex formations were observed with nuclear proteins extracted from PSCs treated with IL-1␤ (Fig.  6A). Preincubation of PSCs with 15d-PGJ 2 decreased IL-1␤induced NF-B binding activity, but troglitazone did not. Phosphorylation and degradation of the inhibitory protein IB-␣ and the subsequent dissociation of this protein from NF-B are thought to be necessary for the activation (25). We also examined the effect of PPAR-␥ ligands on the level of IB-␣ by Western blotting. In agreement with the result of electrophoretic mobility shift assay, 15d-PGJ 2 , but not troglitazone, inhibited IL-1␤-induced degradation of IB-␣ (Fig. 6B). TNF-␣ also activated NF-B, and the activation was inhibited by 15d-PGJ 2 but not by troglitazone (data not shown). Neither 15d-PGJ 2 nor troglitazone affected IL-1␤-and TNF-␣-induced AP-1-specific binding activity (Fig. 7, data not shown). DISCUSSION Following pancreatic injury, PSCs undergo a transformation from quiescent cells to activated proliferating myofibroblastlike cells, which produce cytokines and extracellular matrix proteins. There is accumulating evidence that activated PSCs play principal roles in the pathogenesis of pancreatic fibrosis and inflammation (1,2,4,5). The present study demonstrated that two PPAR-␥ ligands, the endogenously produced prostanoid, 15d-PGJ 2 , and troglitazone, inhibited key parameters of PSC activation including cell proliferation, ␣-SMA expression, ␣1(I) procollagen and prolyl 4-hydroxylase(␣) gene expression, and MCP-1 production. Treatment with a different peroxisome proliferator, WY-14643, at the concentration able to activate the PPAR-␣ isoform did not inhibit PSC proliferation but did so at the concentration also able to activate PPAR-␥, suggesting that specific activation of PPAR-␥ is necessary to inhibit PSC proliferation. These inhibitory effects were not through the potential cytotoxic effects of PPAR-␥ ligands, because the concentrations of these ligands used in this study did not cause cell death.
Activated PSCs are the principal source of collagen, mainly type I, during pancreatic fibrosis. In this study, we have shown that PPAR-␥ ligands decreased the steady-state mRNA levels of ␣1(I) procollagen. The mechanism of type I collagen gene expression is cell-and stimulation-type specific, but the precise mechanism is unclear in PSCs. In hepatic stellate cells, there are several reports dealing with this topic. For example, serumstimulated ␣1(I) collagen gene expression via extracellular signal-regulated kinase (ERK) and c-Jun N-terminal kinase/ stress-activated protein kinase (JNK/SAPK) pathways through different regions of the 5Ј-upstream promoter sequence of the gene (26). Although the ERK-stimulatory signal was mapped to the most proximal nuclear factor-1and Sp-1 binding domains, a distal GC box located at Ϫ1484 to Ϫ1476 base pairs played a central role in receiving extracellular signals through the JNK/ SAPK pathway (26). JNK/SAPK and AP-1 activation were also required for the UV-and acetaldehyde-induced increase in FIG. 4. PPAR-␥ ligands inhibited expression of ␣-SMA, ␣1(I) procollagen, and prolyl 4-hydroxylase. A, PSCs were treated with 15d-PGJ 2 (at 5 M) or troglitazone (at 10 M) for 48 h. Then total cell lysates were prepared, and the level of ␣-SMA was determined by Western blotting. B and C, PSCs were grown directly on glass coverslips. Immunostaining for type I collagen (B) and prolyl 4-hydroxylase(␣) (C) was performed using a streptavidin-biotin-peroxidase complex detection kit. Original magnification: ϫ 20 objective. D, after treatment with 15d-PGJ 2 (at 2.5 or 5 M) or troglitazone (at 5 or 10 M) for 24 h, total RNA was extracted, and mRNA levels of ␣1(I) procollagen and prolyl 4-hydroxylase(␣) were determined by Northern blotting.  Ͻ 0.01 versus control). B, PSCs were treated with 15d-PGJ 2 (5 M) or troglitazone (10 M) for 30 min and then stimulated with IL-1␤ (10 ng/ml) or TNF-␣ (10 ng/ml). After 6 h, total RNA was extracted, and the levels of MCP-1 were determined by Northern blotting. ␣1(I) collagen gene expression (26,27). The UV-and acetaldehyde-responsive elements were located in the distal GC box, and the GC box was bound by a DNA-binding protein termed basic transcription-binding protein. On the other hand, the treatment of rat hepatic stellate cells with a 5-lipoxygenasespecific inhibitor reduced ␣1(I) procollagen mRNA transcript abundance, suggesting that leukotriene production might be involved in maintaining the high level of collagen production of the activated cell (28). Suppression of the gene transcription was localized to a nuclear factor-1 binding domain in the proximal promoter and an AP-2 binding domain adjacent to it. An increase in AP-2 binding adjacent to the nuclear factor-1 site was probably the transmodulator responsible for the suppression of the nuclear factor-1-dependent gene expression (28). It remains to be studied whether PPAR-␥ ligands inhibit ␣1(I) collagen gene expression in a similar mechanism.
The activation of PPAR-␥ has been shown to be sufficient to induce growth arrest as well as initiate adipogenesis in the exponentially growing fibroblast cell line (29). Moreover, retrovirally mediated ectopic expression of PPAR-␥ in the presence of PPAR activators decreases myoblast-specific gene expression such as ␣-SMA and directs myogenic cells into adipocytes lineage (30). PPAR-␥ activation was shown to lead to G 1 cell cycle arrest in fibroblasts by a mechanism involving down-regulation of protein phosphatase 2A (14), but this effect was not documented in other cell types (32), suggesting that different mechanisms might be involved. Mutual interactions between members of the nuclear receptor superfamily and members of other families of transcription factors have also been described (33,34). PPAR receptors, similar to the retinoic acid receptors and the glucocorticoid receptor, can modulate gene expression by antagonizing AP-1 activity involved in cell proliferation. Both 15d-PGJ 2 and the synthetic PPAR-␥ ligand BRL49653 inhibited AP-1 activities in a PPAR-␥-dependent manner in many cell types (11,35). However, PPAR-␥ ligands did not inhibit AP-1 binding activity in this study, suggesting an involvement of other mechanisms.
We have shown that proinflammatory cytokines IL-1␤ and TNF-␣ induce MCP-1 expression in activated PSCs. MCP-1 is a major chemoattractant, playing the central role in the recruitment of monocytes and lymphocytes to the site of inflammation (22). MCP-1 expression by myofibroblasts is increased in fibrous tissue sections from patients with chronic pancreatitis (36). Furthermore, recent studies have suggested that it also acts as a fibrosis-promoting chemokine; MCP-1 stimulated collagen gene expression via endogenous up-regulation of transforming growth factor ␤ in rat lung fibroblasts (37). Therefore, control of MCP-1 expression is an important therapeutic target for pancreatic fibrosis as well as inflammation. In this study, PPAR-␥ ligands inhibited cytokine-induced MCP-1 expression at the gene and protein levels. It has been shown that MCP-1 expression is controlled primarily at the transcriptional level, mainly through the activation of NF-B and AP-1 (23,24). For example, the NF-B-like binding site and the AP-1 binding site located 90 and 68 base pairs upstream of the transcriptional start site, respectively, are required for maximal induction of the human MCP-1 promoter by IL-1␤ in human vascular endothelial cells (24). On the other hand, the inhibition of monocyte activation by PPAR-␥ ligands has been shown to be associated with inhibition of the activation of NF-B and AP-1 (11,12). PPAR-␥ ligands were shown to inhibit the activation of inflammatory response genes such as IL-2, IL-6, IL-8, TNF-␣, and metalloproteases by negatively interfering with the NF-B and AP-1 signaling pathways (38). However, a previous study (39) showed that neither 15d-PGJ 2 nor troglitazone inhibited the TNF-␣-induced DNA-binding activity of both NF-B and AP-1 in hepatic stellate cells, suggesting a cell-type specific action of PPAR-␥ on the activation of transcription factors. In our current study, neither 15d-PGJ 2 nor troglitazone affected AP-1-specific binding activity. However, different effects were seen with 15d-PGJ 2 and troglitazone on the activation of NF-B. 15d-PGJ 2 , but not troglitazone, inhibited IB-␣ degradation and consequent NF-B activation. It is possible that 15d-PGJ 2 inhibited MCP-1 expression through the decreased activation of NF-B, whereas the inhibitory effect of troglitazone on MCP-1 expression is not mediated through altered activation of NF-B. The reason for the differences in effect is unknown at this time, but recent studies have shown that FIG. 6. 15d-PGJ 2 , but not troglitazone, inhibited degradation of IB-␣ and consequent NF-B activation. A, PSCs were treated with 15d-PGJ 2 or troglitazone at the indicated concentrations for 30 min and then stimulated with IL-1␤ (at 10 ng/ml). After 1 h, nuclear extracts were prepared, and specific binding activity of NF-B was assessed by electrophoretic mobility shift assay. B, PSCs were treated with 15d-PGJ 2 (at 2.5 or 5 M) or troglitazone (at 5 or 10 M) for 30 min, and then stimulated with IL-1␤ (at 10 ng/ml). After 30 min, total cell lysates were prepared, and the level of IB-␣ was determined by Western blotting.
FIG. 7. PPAR-␥ ligands did not alter IL-1␤-induced AP-1 activation. PSCs were treated with 15d-PGJ 2 or troglitazone at the indicated concentrations for 30 min and then stimulated with IL-1␤ (at 10 ng/ml). After 1 h, nuclear extracts were prepared, and specific binding activity of AP-1 was assessed by electrophoretic mobility shift assay. 15d-PGJ 2 , but not PPAR-␥ ligands such as troglitazone, was a direct inhibitor of IB kinase, which is responsible for NF-B activation (40,41). This effect may be due to the direct inhibition of IB kinase independently of PPAR-␥. On the other hand, a somatic PPAR-␥ mutation, R288H, showed a normal response to synthetic ligands but a greatly decreased response to natural ligand 15d-PGJ 2 (42), implying that there are different responses of PPAR-␥ between different ligands.
The intracellular events that signal the morphological transformation of the PSCs from a quiescent to a myofibroblast-like cells are still unclear, and it would be important to examine the role of PPAR-␥ in this process. In the liver, it has been shown that the activation of stellate cells is associated with the reductions in PPAR-␥ expression and PPAR-responsive element binding both in vivo and in vitro (39,43). In addition, PPAR-␥ ligands inhibited many key parameters of hepatic stellate cell activation in a manner similar to that shown in our current study in PSCs (39). PPAR-␥ levels were reduced as early as after 3 days in culture, when the expression of ␣-SMA is not yet clear in hepatic stellate cells. The decrease in PPAR-␥ at the early stages of hepatic stellate cell activation and the fact that PPAR-␥ ligands inhibit many characteristics associated with the activated phenotype of hepatic stellate cells suggest that PPAR-␥ may be involved in the maintenance of a quiescent phenotype. In addition, growth factors such as epidermal growth factor and PDGF have been shown capable of inhibiting PPAR-␥ expression via mitogen-activated protein kinase-mediated phosphorylation of PPAR-␥ in NIH3T3 cells (44). Further studies are necessary to determine whether PPAR-␥ plays similar roles in PSCs, in which PPAR-␥ is highly expressed and transcriptionally active even after activation in culture.
It would be of particular interest to examine whether PPAR-␥ ligands might provide new therapeutic strategies for pancreatic fibrosis and inflammation, especially taking into consideration its anti-inflammatory effects on leukocytes (11,12) and endothelial cells (35) as well as on PSCs. Thiazolidinediones are now widely used for the treatment of type 2 diabetes mellitus. It has been reported that PPAR-␥ ligands are effective in the treatment of experimental models of several proinflammatory diseases such as adjuvant-induced arthritis (20) and inflammatory bowel disease (31). Experiments designed to test these hypotheses are currently under way in our laboratory.