DLPC decreases TGF-β1-induced collagen mRNA by inhibiting p38 MAPK in hepatic stellate cells

Qi Cao, Ki M. Mak, Charles S. Lieber

Abstract

Dilinoleoylphosphatidylcholine (DLPC), the active component of polyenylphosphatidylcholine extracted from soybeans, decreases collagen accumulation induced by TGF-β1 in cultured hepatic stellate cells (HSCs). Because DLPC exerts antioxidant effects and TGF-β1 generates oxidative stress, we evaluated whether the antifibrogenic effect of DLPC is linked to its antioxidant action. In passage 1 culture of rat HSCs, TGF-β1 induced a concentration-dependent increase in procollagen-α1(I) mRNA levels and enhanced intracellular H2O2 and superoxide anion formation and lipid peroxidation but decreased GSH levels. These changes were prevented by DLPC. Upregulation of collagen mRNA by TGF-β1 was likewise inhibited by catalase and p38 MAPK inhibitor SB-203580, suggesting involvement of H2O2 and p38 MAPK signaling in this process. TGF-β1 or addition of H2O2 to HSCs activated p38 MAPK with a rise in procollagen mRNA level; these changes were blocked by catalase and SB-203580 and likewise by DLPC. α-Smooth muscle actin abundance in HSCs was not altered by TGF-β1 treatment (with or without DLPC), indicating that downregulation of procollagen mRNA by DLPC was not due to alteration in HSC activation. These results demonstrate that DLPC prevents TGF-β1-induced increase in collagen mRNA by inhibiting generation of oxidative stress and associated H2O2-dependent p38 MAPK activation, which explains its antifibrogenic effect. DLPC, an innocuous phospholipid, may be considered for prevention and treatment of liver fibrosis.

  • dilinoleoylphosphatidylcholine
  • oxidative stress
  • antioxidant
  • catalase
  • p38 inhibitor SB-203580

polyenylphosphatidylcholine(PPC), a mixture of polyunsaturated phosphatidylcholines extracted from soybeans, protects against alcoholic fibrosis and cirrhosis in baboons (38) and attenuates fibrosis induced by carbon tetrachloride or heterologous albumin in rats (40). The protection afforded by PPC against fibrosis was associated with decreased transformation of hepatic stellate cells (HSCs) into myofibroblast-like cells (41), implying that PPC may have a direct inhibitory effect on HSC fibrogenesis. Indeed, when PPC was tested for its antifibrogenic effect in cultured rat HSCs, it decreased acetaldehyde-stimulated collagen accumulation in the culture media (36). The inhibition was attributable, in part, to the stimulation of collagenase activity by its active component dilinoleoylphosphatidylcholine (DLPC), which constitutes 45–50% of the PPC extract (38). DLPC was also shown to mimic the inhibitory effect of PPC on platelet-derived growth factor stimulation of HSC proliferation in passaged culture cells, possibly through inhibition of the mitogen intracellular signaling cascade (5). Furthermore, we have recently reported that the addition of DLPC to cultured rat HSCs resulted in downregulation of mRNA levels of procollagen-α1(I) and of the tissue inhibitor of metalloproteinase (TIMP)-1 induced by TGF-β1, leading to lesser accumulation of collagen type 1 in the culture media (6). The mechanism by which DLPC opposes the induction of fibrogenesis by TGF-β1 has not yet been elucidated.

TGF-β1 is a potent profibrogenic cytokine that mediates tissue matrix homeostasis in cultured HSCs, particularly in the later phase of HSC activation (32, 63). Its fibrogenic action has been shown to be mediated, in part, by H2O2, which is increased by the cytokine (20). Exogenous addition of H2O2 to HSCs in culture elicited an upregulation of procollagen-α1(I) mRNA, mimicking that of TGF-β1 treatment. Furthermore, both TGF-β1- and H2O2-induced collagen gene expression were blocked by catalase, an H2O2 scavenger, and by the chemical antioxidant pyrrolidine thiocarbamate. These data are consistent with the role of reactive oxygen species (ROS) in the promotion of HSC fibrogenesis (8, 20, 35, 47, 51, 61).

In several cell types, including HSCs, H2O2acts as a second messenger in TGF-β1 signaling cascades (15,20, 48), which involve, in addition to the SMAD group of proteins (26, 44), the MAPKs. MAPKs are important signal-transducing enzymes, unique to eukaryotes, that regulate many cellular functions, including gene expression, immune response, cell proliferation, apoptosis, and response to oxidative stress. Four subgroups of the MAPK family have been identified, which include ERK1/2, JNKs, p38 proteins (p38 α/β/γ/δ), and ERK5 (reviewed in Refs. 10, 14, 50). Of these, the p38 MAPK signaling pathway is also involved in the induction of procollagen-α1(I) mRNA by TGF-β1 in rat glomerular mesangial cells (11), since inhibition of p38 activation with SB-203580, the selective pharmacological p38 inhibitor (13), prevents the induction of collagen mRNA. In human gingival fibroblasts, Ravanti et al. (55) showed that p38 activation is required for the induction of collagenase-3 expression by TGF-β1. There is also evidence that, in the cytokine-induced signaling pathway, ROS participate in the activation of p38 (reviewed in Ref. 50). Indeed, Clerk et al. (12) found that H2O2 directly activates p38 in rat hearts during ischemia-reperfusion, which contributes to myocyte hypertrophy. However, the involvement of oxidative stress-dependent p38 MAPK activation, leading to fibrogenesis triggered by TGF-β1 in HSCs, has not been evaluated. Understanding this oxidant-based signaling mechanism provides an opportunity for the evaluation of the therapeutic efficacy of antioxidant agents against liver fibrosis.

Since DLPC, like PPC, has antioxidant properties (1, 46), the present study was undertaken to evaluate its antioxidant effects against the generation of oxidative stress and the induction of procollagen mRNA by TGF-β1 in culture-activated HSCs and to assess its putative antioxidant action against H2O2-mediated activation of p38 and the associated enhanced induction of collagen mRNA. The effects of DLPC were studied in parallel with those of catalase and the p38 inhibitor SB-203580.

MATERIALS AND METHODS

Isolation of HSCs

Male Sprague-Dawley rats (Charles River Breeding Laboratories, Wilmington, MA), weighing 350–600 g and fed Purina chow and water ad libitum, were used for isolation of HSCs. Animal experimental procedures were approved by the Institutional Animal Care and Use Committee in compliance with the National Research Council “Guide for the Care and Use of Laboratory Animals.” Nonparenchymal cells were isolated from the liver by sequential in situ perfusion with collagenase and protease as described before (45). HSCs were separated from other nonparenchymal cells over a discontinuous two-layer Nycodenz gradient (11.4 and 17%; Sigma, St. Louis, MO) prepared essentially as described by Knook et al. (33). Isolated HSCs were seeded onto plastic tissue culture flasks at 0.9–1.1 × 106 cells/ml in DMEM containing 10% FCS, 2 mM l-glutamine, 100 IU penicillin, and 100 mg/ml streptomycin (GIBCO BRL, Rockville, MD). HSCs were incubated at 37°C in a 5% CO2-air humidified atmosphere. The medium was changed 24 h after plating and 48 h thereafter. Cells were grown to subconfluence and then trypsinized. In all experiments, these cells were subcultured in DMEM supplemented with 10% FCS and the antibiotics and used 3 days later (as passage 1 HSCs). By phase contrast and immunocytochemical microscopy, these cells revealed an activated HSC phenotype as indicated by positive staining for α-smooth muscle actin (α-SMA), desmin, and vimentin.

Treatment of HSCs

Passage 1 HSCs were incubated with serum-free DMEM containing 0.05% bovine albumin (vehicle control), TGF-β1 (8 ng/ml; Sigma), DLPC (10 μM; Avanti Polar Lipids, Alabaster, AL) (6, 36, 49, 53), TGF-β1 + DLPC, catalase (1,000 U/ml), TGF-β1 + catalase, SB-203580 (20 μM; Sigma), TGF-β1+ SB-203580, H2O2 (50 μM), H2O2 + catalase, or H2O2 + SB-203580. DMSO, as vehicle control for SB-203580, was added to the culture when the p38 inhibitor was tested. The concentration-dependent effect of TGF-β1 was determined by using concentrations of 4, 6, 8, and 16 ng/ml.

RNA extraction and Northern blot analysis.

HSCs were incubated with the above reagents for 24 h, and the cellular RNA was extracted with acid guanidinium isothiocyanate-phenol-chloroform and used for Northern blot analysis according to a standard technique (42). A cDNA probe for rat procollagen-α1(I) or β-actin (American Type Culture Collection, Manassas, VA) was labeled with [32P]dCTP by using a random priming DNA labeling kit (Amersham, Arlington Heights, IL). Levels of mRNA were quantified by measuring the intensity of the bands on X-ray film with the MCID image analysis system (Imaging Research, St. Catherines, ON, Canada).

Oxidative Stress Assessment

HSCs (3 × 105) in six-well culture plates were incubated with TGF-β1 (8 ng/ml) with or without addition of DLPC for 2 and 24 h, followed by redox-sensitive probes (described below) for 30 min in the darkness. Cells were washed once with PBS, trypsinized, and resuspended in PBS for fluorescence analysis as follows.

H2O2 generation.

2′,7′-Dichlorodihydrofluorescein diacetate (DCFH2-DA; Molecular Probes, Eugene, OR) was used to measure intracellular H2O2 generation by the method of Carter et al. (7). DCFH2-DA is freely permeable across cell membranes and is incorporated into hydrophobic lipid regions of the cell. H2O2 produced by the cell oxidizes DCFH2-DA to 2′,7′-dichlorofluorescein (DCF), the fluorescence of which is proportional to the H2O2 produced. DCFH2-DA was added to the culture at a final concentration of 20 μM, and DCF fluorescence was measured 30 min thereafter by flow cytometry with a FACScan cytometer (BD Immunocytometry Systems, San Jose, CA) at 488 nm for excitation and 525 for emission. Background reading from cells incubated without the probe was subtracted, and data were analyzed with Cell Quest software.

Superoxide anion generation.

Hydroethidine was used to measure superoxide anion generation by the method described by Carter et al. (7). Hydroethidine is freely permeable to cells and can be directly oxidized to ethidium bromide by superoxide anion produced by the cell. The loss of fluorescence in the cells is proportional to the superoxide anion generated. Hydroethidine (Molecular Probes) was added to the culture at a final concentration of 10 μM, and the fluorescence was measured 30 min thereafter by spectrofluorometry at 352 nm for excitation and 434 nm for emission.

Lipid peroxidation detection.

Lipid peroxidation was measured according to the method described by Kuypers et al. (34). Cis-parinaric acid (Molecular Probes) is a fluorescent polyunsaturated fatty acid that is incorporated into cellular membranes. Subsequent to peroxidative stress, cis-parinaric acid is degraded, resulting in decreased intensity. Hence the loss of fluorescence is proportional to lipid peroxidation. Cis-parinaric acid was added to the culture at a final concentration of 5 μM, and the fluorescence was measured by spectrofluorometry at 325 nm for excitation and 413 nm for emission.

Measurement of GSH.

GSH levels were determined by a Glutathione assay kit (Cayman, Ann Arbor, MI) according to the manufacturer's instructions.

p38 MAPK (Thr180/Tyr182) Phosphorylation Assay

Phosphorylation of p38 was assayed by using the components provided in the PhosphoPlus p38 MAPK antibody kit obtained from Cell Signaling Technology (Beverly, MA). HSCs (1 × 106) grown on 100-mm culture dishes were treated with the reagents (described above) for 0.5, 2, and 24 h. The cells were then lysed by adding 100 μl SDS sample buffer containing 62.5 mM Tris · HCl (pH 6.5), 2% wt/vol SDS, 10% glycerol, 50 mM DTT, and 0.1% wt/vol bromphenol blue. These cells were immediately scraped off the plates and transferred to microcentrifuge tubes and kept on ice. The content was sonicated for 2 s to shear the DNA and reduce the sample viscosity. A 20-μl sample (20 μg protein) was boiled for 5 min, cooled on ice, and then centrifuged for 5 min. The protein was resolved on a 12% SDS-PAGE gel and electroblotted onto a nitrocellulose membrane. After being washed and blocked with blocking buffer, the membrane was incubated with phospho-p38 (Thr180/Tyr182) rabbit polyclonal antibody, which detected p38 only when activated by dual phosphorylation at Thr180 and Tyr182, overnight at 4°C. The antibody was diluted at 1:1,000. After being washed, the membrane was incubated with horseradish peroxidase (HRP)-conjugated anti-rabbit IgG (1:2,000) and HRP-conjugated anti-biotin antibody (1:1,000) to detect the biotinylated protein markers. As a control for protein loading, the blot was subjected to immunoblotting for the corresponding nonphosphorylated p38 with anti-p38 antibody, which detected total p38. Immunoreactive proteins were visualized by the LumiGLO chemiluminescent reagents and then exposed to X-ray film. Signal intensities were quantified with MCID.

p38 MAPK Activity Assay

The kinase activity of p38 was assayed by detection of activating transcription factor (ATF)-2 phosphorylation by using the p38 MAP kinase assay kit provided by Cell Signaling Technology. HSCs (1 × 106 cells) on culture dishes were lysed in a buffer containing 20 mM Tris · HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM Na-pyrophosphate, 1 mM β-glycerophosphate, 1 mM Na3VO4, 1 μg/ml leupeptin, and 1 mM PMSF, scraped off into microcentrifuge tubes, and sonicated. The cell lysate (200 μl) was added to 20 μl monoclonal phospho-p38 (Thr180/Tyr182) antibody immobilized by cross-linking to agarose hydrazide beads and incubated overnight with gentle rocking at 4°C. The immunoprecipitated pellet was washed several times with a kinase buffer (25 mM Tris · HCl, pH 7.5, 5 mM β-glycerophosphate, 2 mM DDT, 0.1 mM Na3VO4, and 10 mM MgCl2) and then incubated with 2 μg ATF-2 fusion protein in the presence of 100 μM ATP and 50 μl kinase buffer for 30 min at 30°C. The reaction was terminated with 25 μl 3× SDS loading buffer (187.5 mM Tris · HCl, pH 6.8, 6% wt/vol SDS, 30% glycerol, 150 mM DDT, and 0.03% bromophenol blue). A 20-μl sample was analyzed on a 12% SDS-PAGE gel and electroblotted onto a nitrocellulose membrane. The kinase activity was assayed by detection of phosphorylated ATF-2 by using a phospho-ATF-2 (Thr71) antibody (1:1,000). After an overnight incubation with the primary antibody at 4°C, the membrane was incubated with HRP-conjugated anti-rabbit IgG (1:2,000) and HRP-conjugated anti-biotin antibody to detect the biotinylated protein markers. As a control for protein loading, the blot was subjected to immunoblotting for the corresponding nonphosphorylated ATF-2. Immunoreactive proteins were detected by using the LumiGLO chemiluminescent reagents and then exposed to X-ray film.

Western Blot Analysis of α-SMA

Passage 1 HSCs treated with TGF-β1 with or without DLPC for 24 h were lysed with 62.5 mM Tris · HCl (pH 6.8) buffer containing 1% SDS, 10% glycerol, and 20 mM DTT. The cell lysate was sonicated and boiled for 3–5 min. An aliquot of 10 μg cell protein was resolved on a 12% SDS-PAGE gel and electroblotted onto a nitrocellulose membrane. The membrane was incubated with a monoclonal anti-α-SMA (clone 1A4; Sigma) and anti-GADPH (Biodesign, Sacco, ME) as the primary antibodies, followed by alkaline-conjugated goat anti-mouse IgG as the secondary antibody. Immunoreactive proteins were visualized with an enhanced chemiluminescence kit (ICN, Irvine, CA) and then exposed to X-ray film. Signal intensities were analyzed by using MCID.

Protein Determination

HSC lysate protein content was determined using the BCA protein assay kit from Pierce (Rockford, IL).

Cell viability assay.

The viability of HSCs after treatment with the various reagents was determined by the colorimetric 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT) assay according to the method described by Imamura et al. (27) and Martinou et al. (43).

Statistics

Data are reported as means ± SE. Statistical analysis was performed by ANOVA followed by Student-Newman-Keuls tests for multiple comparisons between treatment groups using Instat (v. 3.01) and Sigma Stat (v. 2.0) software (Jandel Scientific, San Rafael, CA).P < 0.05 was considered significant.

RESULTS

Induction of Procollagen-α1(I) mRNA by TGF-β1: Concentration Effect and Inhibition by DLPC

Culture-activated HSCs expressed procollagen-α1(I) mRNA of 5.8 and 4.8 kb, as analyzed by Northern blot (Fig.1 A). TGF-β1 induced a concentration-dependent increase in collagen mRNA levels, reaching a 3.8-fold rise at 8 ng/ml TGF-β1. As shown in Fig. 1 B, DLPC at 10 μM concentration reduced the abundance of collagen mRNA elicited by TGF-β1 to almost the control level. These concentrations of TGF-β1 and DLPC were used in subsequent experiments.

Fig. 1.

Induction of procollagen-α1(I) mRNA by TGF-β1: effects of concentration and dilinoleoylphosphatidylcholine (DLPC). A: hepatic stellate cells (HSCs) incubated with different concentrations of TGF-β1 for 24 h were analyzed for collagen mRNA levels by Northern blot. Top: sample Northern blot showing 2 procollagen-α1(I) mRNA transcripts of 5.8 and 4.8 kb in the sizes expressed by HSCs. Bottom: histograms summarizing 3 separate Northern blot analyses. TGF-β1 increased abundance of collagen mRNA, with a maximal effect at 8 ng/ml TGF-β1. The intensity of the collagen mRNA bands was normalized to that of β-actin, and values are expressed relative to control (0 ng/ml TGF-β1), which was assigned a value of 1. * P < 0.05 and *** P < 0.001 vs. 0, 4, and 6 ng/ml. B: HSCs were incubated with TGF-β1 (8 ng/ml) without or with the addition of DLPC for 24 h. Induction of collagen mRNA by TGF-β1 was almost fully prevented by DLPC. DLPC in the absence of TGF-β1 had no effect on basal rate of collagen mRNA induction. The change was calculated as in A, with no treatment (TGF-β1 −; DLPC −) assigned a value of 1. Top: sample Northern blot.Bottom: histograms summarizing 3 separate analyses. *** P < 0.001 vs. other treatment groups.

Generation of Oxidative Stress by TGF-β1 and Inhibition by DLPC

The capacity of TGF-β1 to generate oxidative stress was determined by measuring H2O2 and superoxide anion formation, lipid peroxidation, and GSH levels in HSCs. Figure2 A shows that TGF-β1 already increased the intracellular H2O2 level ninefold after 2 h, and the increase was still fourfold after 24 h. The increases were abolished by DLPC treatment. Superoxide anion generation and lipid peroxidation were doubled by the cytokine after 2 h; the changes were prevented by DLPC (Fig. 2, B andC). No changes of these parameters were found after 24 h. The level of GSH in HSCs was decreased by 30% after 2 h, and the change was maintained at 24 h after TGF-β1 treatment; DLPC restored the decreases (Fig. 2 D). These data document the antioxidant property of DLPC.

Fig. 2.

Effects of TGF-β1, DLPC, or both on oxidative stress. HSCs treated with TGF-β1 (8 ng/ml), DLPC, or both for 2 and 24 h were assessed for oxidative stress. A: intracellular H2O2 generation. Flow cytometry analysis of dichlorofluorescence in HSCs is shown. TGF-β1 increased H2O2 generation 9-fold after 2 h, and the level declined to half after 24 h. DLPC completely abrogated the increase induced by the cytokine both at 2 and 24 h. Band C: decreased fluorescence of hydroethidine andcis-parinaric acid. This is proportional to increased superoxide formation and lipid peroxidation, respectively. TGF-β1 doubled superoxide anion generation and lipid peroxidation after 2 h. Addition of DLPC completely prevented changes caused by the cytokine. No such changes were detected after 24 h. Values are expressed relative to respective controls, assigned a value of 1.D: after 2 and 24 h, GSH levels were significantly reduced by TGF-β1. These were restored by the addition of DLPC. In all cases, DLPC alone had no effect on these parameters. * P < 0.05 and *** P < 0.001 vs. other treatment groups at 2 h; # P<0.05 vs. other treatment groups at 24 h.

Effects of Catalase and SB-203580 on TGF-β1-Induced Procollagen-α1(I) mRNA

When catalase or SB-203580 was added to cultured HSCs in the presence of TGF-β1, the increased procollagen mRNA level induced by the cytokine was reduced by half after 24 h (Fig.3), suggesting involvement of both H2O2 and p38 MAPK in the TGF-β1-mediated signaling of collagen mRNA expression. In the absence of the cytokine, catalase or SB-203580 had no effect on the basal mRNA levels. DMSO, in a concentration equivalent to that present in SB-203580, had no effect on procollagen mRNA levels, whether induced by TGF-β1 or not.

Fig. 3.

Effects of catalase and SB-203580 on induction of procollagen-α1(I) mRNA by TGF-β1. HSCs were treated for 24 h with TGF-β1 (8 ng/ml) without or with addition of catalase (1,000 U/ml) or the p38 inhibitor SB-203580 (20 μM). Increase in procollagen mRNA level induced by the cytokine was significantly reduced by either catalase or SB-203580. Catalase, SB-203580, or DMSO treatment alone had no effect on basal rate of collagen mRNA induction.Top: sample Northern blot. Bottom: histograms of Northern blot analyses. *** P < 0.001 vs. TGF-β1 −; ** P < 0.01 vs. TGF-β1 alone.

Phosphorylation of the p38 MAPK by TGF-β1: Effects of Concentration and Time

Fig. 4 A shows the induction of p38 phosphorylation by TGF-β1 in a concentration-dependent manner with a maximal effect (4-fold) at 8 ng/ml after 2 h. At this TGF-β1 concentration, a significant increased p38 phosphorylaton was detected beginning at 2 h, and the effect was maintained at 24 h (Fig. 4 B).

Fig. 4.

Induction of p38 MAPK phosphorylation by TGF-β1: effects of concentration and time. A: HSCs were incubated with TGF-β1 at different concentrations for 2 h, and cell lysates were subjected to Western blot analysis using the phospho-p38 MAPK or p38 antibodies (sample blots shown at top). The phospho-p38 antibody detected phosphorylation of p38, whereas p38 antibody detected total p38. Although changes in band intensities were apparent with the phospho-p38 antibody, no corresponding changes were detected with the p38 antibody. Bottom: histograms of Western blot analyses revealed increased phosphorylation of p38, with a maximal effect at 8 ng/ml TGF-β1. Intensity of phospho-p38 bands was normalized to that of p38 of the corresponding treatment groups, and phospho-p38 protein levels are expressed relative to control (0 ng/ml TGF-β1), assigned a value of 1. B: time course of induction of p38 phosphorylation by TGF-β1 in HSCs. At a TGF-β1 concentration of 8 ng/ml, a significantly increased p38 phosphorylation was detected, beginning at 2 h and maintained at 24 h after the cytokine treatment. +C, positive control consisting of cell extracts provided in the PhosphoPlus p38 antibody kit. ** P < 0.01 vs. 0 ng/ml TGF-β1.

Effects of Catalase, SB-203580, and DLPC on TGF-β1-Induced p38 MAPK Phosphorylation and its Kinase Activity

TGF-β1 treatment caused a fourfold increase in p38 phosphorylation after 2 h (Fig.5 A), with an associated increase in the p38 kinase activity (6.5-fold; Fig.6). Catalase completely blocked TGF-β1 stimulation of p38 phosphorylation (Fig. 5 A) and the kinase activity (Fig. 6), suggesting an H2O2-dependent mechanism. The stimulation of p38 activation by TGF-β1 was specific, because it was blocked by the p38 inhibitor SB-203580. Likewise, DLPC blocked these changes (Figs. 5 B and 6).

Fig. 5.

Effects of catalase, SB-203580, and DLPC on the induction of p38 MAPK phosphorylation by TGF-β1. HSCs were incubated with TGF-β1 (8 ng/ml) in the absence or presence of catalase, SB-203580, or DLPC for 2 h, and cell lysates were prepared for analysis of p38 phosphorylation by Western blot. Catalase or SB-203580 completely prevented phosphorylation of p38 induced by the cytokine (A). Likewise, DLPC almost completely blocked TGF-β1-induced phosphorylation of p38 (B). Note that DMSO by itself or in the presence of TGF-β1 had no effect on p38 phosphorylation. In all cases, no changes in intensity of p38 bands were detected by the p38 antibody. Top: sample Western blots. Bottom: histograms of Western blot analyses. Intensity of phospho-p38 bands was normalized to that of p38 of corresponding treatment groups, and values of phospho-p38 protein are expressed relative to control, assigned a value of 1. *** P < 0.001 vs. other treatment groups.

Fig. 6.

Effects of catalase, SB-203580, and DLPC on induction of p38 MAPK activity by TGF-β1. HSCs were incubated with TGF-β1 (8 ng/ml) with addition of DLPC, catalase, or SB-203580 for 2 h, and cell lysates were prepared for p38 kinase activity assay by immunoprecipitation with phospho-p38 antibody, followed by detection of phosphorylation of the substrate ATF-2 protein with phospho-ATF-2 (Thr71) antibody by Western blotting (top). As shown in histograms summarizing Western blot analyses (bottom), TGF-β1 significantly increased p38 kinase activity in HSCs. Catalase, SB-203580, or DLPC abrogated changes induced by the cytokine. Signal intensity of the ATF-2 bands was normalized to that of the corresponding ATF-2 protein, and levels of p38 kinase activity are expressed relative to control, assigned a value of 1. *** P < 0.001 vs. other treatment groups.

H2O2-Dependent Induction of Procollagen-α1(I) mRNA and its Inhibition by Catalase, SB-2035880, and DLPC

To more directly assess whether H2O2upregulates procollagen mRNA level, exogenous H2O2 was added to cultured HSCs. After 24 h, procollagen mRNA abundance was increased ∼2.5-fold (Fig.7). The induction of procollagen mRNA by H2O2 was blocked by catalase. When SB-203580 was added to the HSCs in the presence of the oxidant, this p38 inhibitor, like catalase, prevented the rise in procollagen mRNA level induced by H2O2 (Fig. 7), suggesting the requirement for the p38 signaling pathway in the induction. DLPC, like catalase and SB-203580, reduced H2O2-induced increase in procollagen mRNA level to the control value, possibly through its antioxidant action.

Fig. 7.

Induction of procollagen-α1(I) mRNA by H2O2 and its inhibition by catalase, SB-203580, and DLPC. Northern blot analysis of collagen mRNA levels in HSCs incubated with exogenous H2O2 (50 μM) revealed after 24 h a 2.5-fold increase in procollagen mRNA level. Addition of catalase, SB-203580, or DLPC significantly reduced collagen mRNA levels induced by the oxidant. Top: sample Northern blot. Bottom: histograms of Northern blot analyses. Intensity of the collagen mRNA bands was normalized to that of β-actin, and values are expressed relative to control, assigned a value of 1. *** P < 0.001 vs. other treatment groups.

Effect of H2O2 on p38 MAPK Phosphorylation and its Kinase Activity: Inhibition by Catalase, SB-203580, and DLPC

The capacity of H2O2 to activate the p38 MAPK in HSCs was evaluated. When the oxidant was added to the HSCs in culture, a threefold increase in p38 phosphorylation and a fourfold increase in the kinase activity were found after 2 h (Fig.8). Catalase and SB-203580 completely blocked the stimulation elicited by H2O2. The inhibition by catalase and the p38 inhibitor was mimicked by DLPC, further substantiating DLPC's involvement in the modulation of the p38 signaling by H2O2.

Fig. 8.

Stimulation of p38 MAPK phosphorylation and activity by H2O2 and their inhibition by catalase, SB-203580, and DLPC. HSCs were treated with H2O2 (50 μM), and cell lysates were used for analysis of p38 phosphorylation and activity by Western blot. H2O2 increased phosphorylation (A) and activity (B) of p38 after 2 h. Addition of catalase, SB-203580, or DLPC completely blocked these stimulations induced by H2O2. Top: sample Western blots. Bottom: histograms of Western blot analyses. *** P < 0.001 vs. other treatment groups.

α-SMA and Cell Viability After TGF-β1 and DLPC Treatment

Passage 1 cultured HSCs expressed an appreciable amount of α-SMA, a cytochemical marker of HSC activation. Treatment of HSCs with DLPC in the absence or presence of TGF-β1 for 24 h did not change the α-SMA content as analyzed by immunoblot (Fig.9 A). Furthermore, HSC viability did not differ between treatment groups, as determined by the MTT assay (Fig. 9 B).

Fig. 9.

Effects of TGF-β1, DLPC, or both on α-smooth muscle actin (α-SMA) in HSCs and cell viability. A: TGF-β1 treatment with or without DLPC of HSCs for 24 h did not change the abundance of α-SMA, indicating that HSC activation was not affected.Top: sample Western blot. Bottom: histograms of 3 separate blots after normalization with GADPH. B: such treatment also did not affect HSC viability, as determined by 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide assay.

DISCUSSION

The present study revealed that DLPC, which is the active component of PPC extracted from soybeans, prevents the induction of procollagen-α1(I) mRNA by TGF-β1 in cultured rat HSCs. The inhibitory effect of DLPC on collagen mRNA induction is attributable to its antioxidant action against the oxidative stress generated by TGF-β1 and the associated H2O2-dependent activation of p38 MAPK signaling, leading to downregulation of collagen mRNA. DLPC mimics the action of catalase, which scavenges H2O2, and that of SB-203580, which blocks p38 activation, and is as effective as these mediators against the induction of collagen mRNA elicited by TGF-β1. These findings explain the DLPC antifibrogenic action that is mediated by its antioxidant property.

We have previously described three major regulatory actions of TGF-β1 on collagen production by cultured HSCs (6). These include upregulation of mRNA levels of collagen type I and of TIMP-1, with increased concentration of the corresponding proteins in the presence of unchanged matrix metalloproteinase (MMP)-13. As a consequence, collagen accumulates in the culture media. DLPC exerts its antifibrogenic effect against TGF-β1 by decreasing not only mRNA levels of procollagen but also that of TIMP-1, without affecting MMP-13 mRNA levels; this leads to lesser collagen accumulation in the culture media. Furthermore, the specificity of the antifibrogenic property of DLPC has been verified by comparing its action with that of palmitoyllinoleoylphosphatidylcholine, the second most abundant component of the PPC extract (38), which had no effects on collagen synthesis and its accumulation in cultured HSCs (6, 53).

In the present study, we have identified a novel antifibrogenic action of DLPC. The oxidative stress generated by TGF-β1 is a general property of the cytokine, because it induces ROS formation not only in HSCs (15, 20) but in other cell types as well, including hepatocytes (30), lung fibroblasts (62), and osteoblasts (49). ROS are potent mediators of fibrogenesis in HSCs (9, 20, 35, 47, 51, 61). In vivo studies showed that PPC decreases oxidative stress induced by alcohol consumption in baboons (37) and prevents lipid peroxidation induced by CCl4 in rats (1). DLPC was also found to prevent oxidization of human low-density lipoproteins (46). We now demonstrate that DLPC prevents the generation of H2O2 and superoxide anion, prevents lipid peroxidation, and restores GSH levels in HSCs after TGF-β1 treatment, documenting its antioxidant action.

In conjunction with its blocking H2O2accumulation generated by TGF-β1 in HSCs, DLPC inhibits H2O2-mediated phosphorylation of p38 MAPK. This is associated with a reduction in its kinase activity, as revealed by the decreased phosphorylation of the transcription factor ATF-2, a substrate of p38 (29, 39, 54), which has been shown to be activated by the TGF-β1 signaling cascade via the p38 pathway (24, 58). The efficacy of DLPC inhibition of p38 activation mediated by H2O2 and the resulting procollagen mRNA downregulation were found to be equal to that of catalase or of the p38 inhibitor SB-203580. Previous studies described H2O2 as a signaling molecule for TGF-β1 induction of procollagen-α1(I) mRNA in a rat HSC line (20) and as a link between oxidative stress and enhanced collagen-α1(I) gene expression induced by acetaldehyde in passaged mouse HSCs (22). Importantly, the present study provides additional data showing that H2O2activates the p38 signaling pathway that leads to collagen mRNA upregulation, thereby suggesting that the p38 MAPK serves as a molecular link between H2O2 signaling and enhanced collagen mRNA induction by TGF-β1 in HSCs.

Our finding of the inhibition of TGF-β1-induced p38 activation by SB-203580, leading to downregulation of procollagen mRNA level, is consistent with the observation of Chin et al. (11) in glomerular mesangial cells, which is not unexpected because activated HSCs and mesangial cells share common characteristics of extracellular matrix production and vitamin A storage (4). Thus far, there is only limited information on the role of p38 signaling in HSC fibrogenesis. A recent study showed that inhibition of p38 by SB-203580 decreased procollagen-α1(I) mRNA level in HSCs, with a maximal effect in early primary culture (59). Another study reported that the constitutive p38 activity was higher in activated HSCs than in quiescent cells, suggesting a role for p38 in the activation process of HSCs (56). Since the abundance of α-SMA in passage 1 HSCs was not altered by treatment with TGF-β1, DLPC, or both, the activation of HSCs was not a factor involved in the mediation of the antifibrogenic and antioxidant effects of DLPC. Knittel et al. (31) also observed no changes in the level of α-SMA in culture-activated HSCs after TGF-β1 treatment.

Our data indicate that, even in their activated phenotype, HSCs are responsive to TGF-β1 stimulation of ROS generation, activation of p38 signaling, and, ultimately, upregulation of procollagen mRNA in a concentration-dependent manner, with a maximal 3.8-fold increase at 8 ng/ml after 24 h. This level of mRNA induction is commonly elicited in culture-activated HSCs (3, 65) and in the HSC line (CFSC-2G) derived from a rat CCl4-cirrhotic liver (20) induced by TGF-β1 or in mouse HSCs stimulated by acetaldehyde, a potent fibrogenic stimulator (22, 45). Casini et al. (9) found that human passaged HSCs responded to TGF-β1 with a concentration-dependent increase of procollagen-I and -III mRNA abundance and of their collagen accumulation in the culture. Recent studies by Dooley et al. (16, 17) described changes in the responsiveness to TGF-β1 of HSCs upon passage in culture. Whereas HSCs in the early stage of primary culture were responsive to TGF-β1 stimulation of collagen-α2(I) mRNA and inhibition of HSC proliferation, activated HSCs in the later stage of primary culture were minimally responsive to TGF-β1 treatment. The loss of responsiveness of activated HSCs to TGF-β1 was ascribed to reduced TGF-β ligand-binding activity and diminished DNA binding of intracellular SMAD proteins, despite the normal expression of TGF-β receptors I and II on the cell surface. A lower collagen gene transcriptional response to TGF-β1 in the rat CFSC-2G HSC line compared with fetal skin fibroblasts has also been reported (28), yet these same cells upon TGF-β treatment can produce an amount of H2O2 sufficient to lead to a significant increase in collagen mRNA level (20). In the present study, we have not compared the responsiveness to TGF-β1 stimulation of ROS generation between passage 1 activated and quiescent HSCs in early primary culture. However, De Bleser et al. (15) showed that activated HSCs generate H2O2 as well as quiescent cells after TGF-β1 treatment. Renal mesangial cells are responsive to exogenous TGF-β1, with ensuing fibrogenesis mediated by the p38 (11) or SMAD-transducing signaling pathway (52), even after many passages in culture. The loss of cellular response to the cytokine may be cell type and context dependent. In the case of p38 activation stimulated by TGF-β1 in activated HSCs, as observed in the present study, the signaling mechanism is most likely involved with the TGF-β-activated kinase 1 pathway, which has been shown to activate the MAPKs (58, 64), rather than the SMAD pathway. Indeed, the data reported by Schnabl et al. (60) suggested that the activation of ERK1/2 by TGF-β1 in HSCs is independent of SMAD signaling.

In this study, we focused on the p38 signaling pathway that is activated by H2O2 after TGF-β1 treatment and on its inhibition by DLPC. Although the JNK pathway of the MAPKs can also be activated in response to environmental stress and by TGF-β1 signaling (64), its involvement in the induction of collagen mRNA by TGF-β1 is less clear, with inhibition reported in HSCs at the early stage of primary culture and no effect at the later stage (59). In renal mesangial cells, JNK was not activated by TGF-β1 (11) or shown to play a role in TGF-β-1-mediated collagen-α1(I) expression (25). The ERK1/2 MAPK has been reported to be activated by TGF-β1 in HSCs (57), but its role in fibrogenesis was not examined.

In response to liver injury in general and fibrogenic stimuli in particular, HSCs are activated to proliferate (18, 21) and to transform (23, 41) into myofibroblast-like cells with an active fibrogenic phenotype, which play a major role in the onset and progression of liver fibrosis. Because HSCs cultured on plastic recapitulate the biological activation process of HSCs in vivo (19), they provide a model for the evaluation of potential therapeutic agents against liver fibrosis (66). The present study demonstrates the usefulness of culture-activated HSCs in conjunction with TGF-β1 treatment as a model for the evaluation of the efficacy of antifibrogenic agents, in particular compounds with potential antioxidant properties, as exemplified by DLPC.

In conclusion, our findings revealed a novel mechanism by which DLPC prevents the induction of fibrogenesis by TGF-β1 through inhibition of the oxidative stress generated by the cytokine and the associated H2O2-dependent activation of the p38 MAPK signaling pathway. This effect of DLPC may explain, at least in part, how PPC opposes fibrosis, and through this action DLPC or PPC may be useful for the treatment of alcoholic as well as nonalcoholic liver fibrosis.

Acknowledgments

This study was supported, in part, by National Institute on Alcohol Abuse and Alcoholism Grant AA-11115, the Department of Veterans Affairs, and the Kingsbridge Research Foundation.

Footnotes

  • Address for reprint requests and other correspondence: C. S. Lieber, Alcohol and Nutrition Research Center, Veterans Affairs Medical Center, Mount Sinai School of Medicine, 130 West Kingsbridge Rd., Bronx, NY 10468 (E-mail:liebercs{at}aol.com).

  • 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.

  • 10.1152/ajpgi.00128.2002

REFERENCES

View Abstract