Mechanisms of endotoxin-induced NO, IL-6, and TNF-α production in activated rat hepatic stellate cells: Role of p38 MAPK

Authors

  • Chinnasamy Thirunavukkarasu,

    1. Department of Surgery, Thomas E. Starzl Transplantation Institute, University of Pittsburgh, Pittsburgh, PA
    Search for more papers by this author
  • Simon C. Watkins,

    1. Department of Cell Biology, Thomas E. Starzl Transplantation Institute, University of Pittsburgh, Pittsburgh, PA
    Search for more papers by this author
  • Chandrashekhar R. Gandhi

    Corresponding author
    1. Department of Surgery, Thomas E. Starzl Transplantation Institute, University of Pittsburgh, Pittsburgh, PA
    2. Department of Pathology and VA Medical Center, Thomas E. Starzl Transplantation Institute, University of Pittsburgh, Pittsburgh, PA
    • Thomas E. Starzl Transplantation Institute, University of Pittsburgh, E-1540 BST, 200 Lothrop Street, Pittsburgh, PA 15213
    Search for more papers by this author
    • fax: 412-624-6666.


  • Potential conflict of interest: Nothing to report.

Abstract

Compelling experimental evidence indicates that the interactions between endotoxin and hepatic stellate cells (HSCs) can play a significant role in the pathogenesis of liver disease. Endotoxin-induced release of a multifunctional mediator NO (via inducible NO synthase) and the proinflammatory cytokines tumor necrosis factor α (TNF-α) and interleukin (IL)-6 by HSCs could be an important mechanism of pathological changes in the liver. However, the signaling mechanisms of these effects are poorly understood. In this study, we found that endotoxin causes activation of mitogen-activated protein kinases (MAPKs) (extracellular signal-regulated protein kinase [ERK] 1 and 2, p38, and c-Jun NH2-terminal kinase [JNK]) and nuclear factor κB (NF-κB) and production of H2O2 in culture-activated HSCs. However, only p38 and NF-κB were found to be responsible for the synthesis of NO, IL-6, and TNF-α. Exogenous H2O2 caused modest stimulation of TNF-α synthesis, did not affect the synthesis of NO or IL-6, and did not activate NF-κB or MAPKs. Inhibition of p38 and NF-κB activation by SB203580 and pyrrolidine dithiocarbamate, respectively, blocked endotoxin-induced H2O2, NO, TNF-α, and IL-6 synthesis. Inhibition of ERK1/2 and JNK phosphorylation did not alter these effects of endotoxin. Whereas SB203580 inhibited endotoxin-induced NF-κB activation, pyrrolidine dithiocarbamate did not affect p38 phosphorylation in endotoxin-stimulated cells. In conclusion, endotoxin-induced synthesis of NO, TNF-α, and IL-6 in HSCs is mediated by p38 and NF-κB, with involvement of H2O2 in TNF-α production. (HEPATOLOGY 2006;44:389–398.)

Physiologically, hepatic stellate cells (HSCs) play an important role in maintaining the liver's architecture by producing components of extracellular matrix and regulating sinusoidal blood flow by contractility. HSCs, the major storage site of retinoids, produce several potent mediators such as hepatocyte growth factor, tumor necrosis factor α (TNF-α), transforming growth factor β, and endothelin-1.1–3 During chronic liver injury, HSCs undergo phenotypical transformation (i.e., activation) to actively proliferating retinoid-deficient cells that express α-smooth muscle actin known as activated HSCs. Activated HSCs are responsible for fibrosis and sinusoidal component of portal hypertension by depositing excessive amounts of extracellular matrix of abnormal composition and increased contractility.1–3 During activation, HSCs lose expression of hepatocyte growth factor but express PDGF receptor, and produce several other growth factors, cytokines, and chemokines such as transforming growth factor α, transforming growth factor β, interleukin (IL)-6, TNF-α, monocyte chemotactic protein 1, macrophage inhibitory protein 2, and IL-8/cytokine-induced neutrophil chemoattractant4, 5 that participate in the initiation and progression of liver disease.

During liver injury or failure, circulating levels of gram-negative bacterial endotoxin are increased as a result of increased gut permeability due to inflammatory cytokines such as TNF-α and inadequate hepatic clearance.6–8 Endotoxin causes upregulation of the endothelin system in HSCs,9, 10 which plays a significant role in the development of portal hypertension as well as fibrosis.11–13 Endotoxin stimulates synthesis of inflammatory cytokines such as interferon-γ, TNF-α, IL-6 and IL-1β in resident macrophages (i.e., Kupffer cells) and infiltrating macrophages.14–16 Endotoxin, in conjunction with interferon-γ, TNF-α, and IL-1β, also stimulates the synthesis of a multifunctional mediator NO in Kupffer cells17 and hepatocytes.18 Work from several laboratories, including ours, has demonstrated that endotoxin stimulates synthesis of NO via induction of inducible NO synthase (iNOS),10, 19–22 and of TNF-α and IL-621, 22 in both quiescent (normal liver) and activated (fibrotic liver) HSCs. TNF-α and IL-6, as well as increased NO (directly or via a potent free radical peroxinitrite), can exert a variety of effects on the hepatic system. However, the mechanisms of endotoxin-induced expression of iNOS, TNF-α, and IL-6 in HSCs are not clearly defined. We report that these effects of endotoxin in culture-activated rat HSCs are initiated by the mitogen-activated protein kinase (MAPK) p38 and are mediated by nuclear factor κB (NF-κB).

Abbreviations

HSC, hepatic stellate cell; TNF-α, tumor necrosis factor α; IL, interleukin; MAPK, mitogen-activated protein kinase; ERK, extracellular signal-regulated protein kinase; JNK, c-Jun NH2-terminal kinase; NF-κB, nuclear factor κB; PDTC, pyrrolidine dithiocarbamate; iNOS, inducible NO synthase; mRNA, messenger RNA; ROS, reactive oxygen species; DCFH-DA, 2′,7′-dichlorohydrofluorescein diacetate; DCF, 2′,7′-dichlorofluorescein; LPS, lipopolysaccharide.

Materials and Methods

Chemicals or Reagents.

PD98059, SB203580, SP600125, and pyrrolidine dithiocarbamate (PDTC) were obtained from Calbiochem (La Jolla, CA). Anti-mouse–α-smooth muscle actin, anti-mouse–desmin, anti-rabbit–glial fibrillary acidic protein, and anti-rabbit extracellular signal-regulated protein kinase (ERK) 1 and 2 antibodies were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Anti-rabbit antibodies against P-ERK1/2, P-p38, p38, P-JNK, and c-Jun NH2-terminal kinase (JNK) were obtained from Cell Signaling Technology, Inc. (Beverly, MA). Monoclonal anti-rat iNOS antibody was obtained from Transduction Laboratories Co. (Lexington, KY). Peroxidase-linked anti-rabbit immunoglobulin G was obtained from Amersham-Pharmacia, Piscataway, NJ. Endotoxin (Escherichia coli lipopolysaccharide, serotype 0111:B4) and peroxidase-linked anti-mouse immunoglobulin G were obtained from Sigma Chemical Company (St. Louis, MO). Cell culture media, fetal bovine serum, horse serum, and antibiotics were obtained from GibcoInvitrogen (Carlsbad, CA). All other chemicals were obtained from standard sources.

Isolation and Culture of HSCs.

The protocols were approved by the Institutional Animal Care and Use Committee of the University of Pittsburgh in accordance with National Institutes of Health regulations. HSCs were isolated from male Sprague-Dawley rats (450-500 g) as previously described21, 22 via collagenase and protease digestion of the liver, followed by removal of hepatocytes and cell debris via low-speed centrifugation. HSCs were purified via density gradient centrifugation on 13.14% (w/v) Nycodenz, suspended in Dulbecco's modified Eagle medium containing penicillin/streptomycin and 10% fetal bovine serum/10% horse serum, and plated on uncoated plastic dishes at a density of 1.0 × 106/well (24-well plate). The viability (trypan blue exclusion) was greater than 97%. The cells were washed and placed in fresh medium on the following day. Purity as determined via vitamin A autofluorescence and immunolabeling with anti-desmin and anti–glial fibrillary acidic protein antibodies was >95%. Kupffer cell contamination as determined via immunostaining with mouse anti-rat–ED2 antibody (Serotec, Raleigh, NC) was ≤3%; no sinusoidal endothelial cell contamination was observed as determined via immunostaining with monoclonal SE-1 antibody (KMI Diagnostics/IBL America, Minneapolis, MN). HSCs were maintained for 10 days with renewal of the medium on alternate days; the activated phenotype was confirmed by α-smooth muscle actin expression.1, 3

Determination of NO.

The concentration of NO end products (nitrite and nitrate) was determined via the Griess method using a colorimetric assay kit (Canyan Chemicals, Ann Arbor, MI). Briefly, Griess reagent was added to the culture supernatant after treatment with nitrate reductase, and the optical density was determined at 550 nm.

Western Blot Analysis for iNOS and MAPKs.

The cells were washed twice with phosphate-buffered saline and lysed for 30 minutes in ice-cold RIPA buffer (Santa Cruz Biotechnology) containing 0.5 mmol/L PMSF, 25 μL/mL of protease inhibitor cocktail (Sigma; catalog no. p-8340), and 1 mmol/L sodium orthovanadate. The lysate was centrifuged at 15,000g for 10 minutes at 4°C. The supernatants containing 10 μg (iNOS) or 20 μg (MAPKs) protein underwent SDS-PAGE at 8% and 12%, respectively. The separated proteins were transferred onto Immobilon-P membranes (Millipore, Bedford, MA). The membranes were incubated for 2 hours at room temperature in 1% bovine serum albumin in TBS/0.1% Tween-20 to block nonspecific binding, incubated with the primary antibodies (1:1,000 dilution) for 2 hours at room temperature, washed 4 times, incubated in appropriate secondary antibodies (1:50,000 dilution) for 2 hours at room temperature, and washed 4 more times. Detection was achieved using an ECL chemiluminescence kit (Amersham-Pharmacia) and Kodak X-Omat XAR film (Eastman Kodak Co., Rochester, NY). To confirm equal loading, the membranes were stripped or the same amount of protein was separated on a separate gel to assess the expression of actin.

Determination of NF-κB Translocation.

For determination of nuclear translocation of NF-κB via Western blot analysis, nuclear and cytoplasmic extracts were prepared using extraction reagents from Pierce-Endogen (Rockford, IL). Nuclear (5 μg) and cytoplasmic (20 μg) proteins were separated via 10% SDS-PAGE. Western blotting was performed using anti–NF-κB p65 antibody (Santa Cruz Biotechnology; 1:1,000) and secondary antibody (1:20,000) as described above.

For immunohistochemical determination, cells were fixed in 2% paraformaldehyde for 30 minutes and permeabilized with 0.01% Triton-X100 for 10 minutes. After blocking with 2% bovine serum albumin, the cells were treated with mouse anti-rabbit NF-κB–p65 antibody for 60 minutes, washed, treated with CY3-conjugated goat–anti-rabbit secondary antibody (Santa Cruz Biotechnology), then washed again. The slides were then incubated with Hoechst reagent (30 seconds), and analyzed under Olympus BX51 fluorescence microscope.

Determination of iNOS and Cytokine Messenger RNA.

Semiquantitative reverse-transcriptase polymerase chain reaction assay was employed to determine the messenger RNA (mRNA) expression of iNOS, IL-6, and TNF-α as previously described.21 β-Actin mRNA expression was determined for comparison. The polymerase chain reaction primers specific for various complementary DNAs were as follows: iNOS, 5′AGAATGTTCCAGAATCCCTCCCTGGACA3′ (forward) and 5′GAGTGAGCTGGTAGGTTCCTGTTG3′ (reverse) (356 bp); TNF-α, 5′CACGCTCTTCTGTCTACTGA3 ′ (forward) and 5′GGACTCCGTGATGTCTAAGT3′ (reverse) (543 bp); IL-6, 5′GAAAGTCAACTCCATCTGCC3 ′ (forward) and 5′CATAGCACACTAGGTTTGCC3 ′ (reverse) (681 bp); and β-actin, 5′TTCTACAATGAGCTGCGTGTG3 ′ (forward) and 5′TTCATGGATGCCACAGGATTC3′ (reverse) (561 bp). The polymerase chain reaction products were resolved in a 1.5% agarose gel and stained with 1× SYBR Green I (FMC Bioproduct, Rockland, ME). The gels were scanned under blue fluorescence light using a phosphorimager and the band intensity was quantified using ImageQuant software (Molecular Dynamics, Sunnyvale, CA).

Measurement of Cytokines.

The culture medium was aspirated and clarified by centrifugation at 10,000g for 10 minutes. The concentrations of TNF-α and IL-6 in the supernatant were measured using ELISA kits from Pierce-Endogen.

Measurement of Intracellular Superoxide and H2O2.

Intracellular reactive oxygen species (ROS) generation was measured fluorometrically in cells loaded with hydroethidine and 2′,7′-dichlorohydrofluorescein diacetate (DCFH-DA), which are specific for the detection of superoxide and H2O2, respectively.23 Hydroethidine enters cells freely and is oxidized to ethidium in the presence of superoxide; the loss of fluorescence is proportional to the amount of superoxide generated. DCFH-DA is cell permeable and is hydrolyzed by cellular esterases to dihydro-2′,7′-dichlorofluorescin, the oxidation of which by H2O2 yields a fluorescent product, 2′,7′-dichlorofluorescein (DCF). The cells were washed and loaded with 10 μmol/L hydroethidine or DCFH-DA in serum-free Dulbecco's modified Eagle medium without phenol red and containing 0.1% bovine serum albumin for 30 minutes at 37°C, and washed 3 times. The cells were then challenged with lipopolysaccharide (LPS), and hydroethidine fluorescence was measured with a Cytofluor 2300 plate reader (Millipore, Vienna, VA) at excitation and emission wavelengths of 352 nm and 434 nm, respectively. The intensity of DCF fluorescence was determined at an excitation wavelength of 485 nm and an emission wavelength of 530 nm. The background fluorescence intensity in the control wells without DCFH-DA was subtracted in all cases. For flow cytometric determination, the cells were collected via trypsinization and washed twice with sterile phosphate-buffered saline, and the fluorescence was measured within 30 minutes using a Coulter Epics XL Flow Cytometer.

Statistical Analysis.

Data are presented as the mean ± SD. Each experiment was repeated at least 3 times using cells from different animals. Statistical significance was determined via ANOVA-Duncan analysis using SPSS software. A P value of <.05 was considered statistically significant.

Results

LPS-Induced Expression of iNOS, IL-6, and TNF-α.

Previously, we demonstrated that incubation of quiescent and activated rat HSCs with LPS for 24 hours increases the synthesis of NO (via induction of iNOS) and the proinflammatory cytokines IL-6 and TNF-α.21, 22 Experiments of the present study demonstrate concentration-dependent stimulation of the synthesis of NO, IL-6, and TNF-α by LPS, with significant increases occurring at the LPS concentration of 0.001 μg/mL (NO) and 0.01-0.1 μg/mL (IL-6 and TNF-α) (Supplementary Fig. 1; Supplementary material for this article can be found on the HEPATOLOGY website (http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html). The time-course experiment demonstrated an increase in the mRNA (Fig. 1A) and protein (Fig. 1B) expression of iNOS starting at 3 hours after stimulation with LPS. LPS also increased NO synthesis, with a statistically significant increase occurring at 3-12 hours and a robust increase at 24 hours (Fig. 1C). To rule out the possibility that a part of the effect of LPS may be due to contamination of Kupffer cells, which was minimal (<3%) in our HSC preparations, the cells were treated with 1 mmol/L gadolinium chloride for 24 hours prior to stimulation with LPS. No difference in LPS-induced NO, IL-6, or TNF-α synthesis was observed between the cells pretreated with gadolinium chloride and the vehicle (phosphate-buffered saline) (data not shown). LPS treatment also did not alter the state of activation as observed by unaltered expression of α-smooth muscle actin and PDGFRβ (data not shown).

Figure 1.

Time-course of the effect of LPS on iNOS gene expression and the release of NO, IL-6, and TNF-α. Cells were washed and placed in serum-free medium containing 1 μg/mL LPS. At the indicated time points, (A) gene and (B) protein expression of iNOS and (C) release of NO were determined. Data (reverse-transcriptase polymerase chain reaction and Western blotting) represent at least 3 separate experiments. Values in line graphs show the means ± SD of triplicate determinations from a representative experiment. *P < .001 versus control. **P < .01 versus control #P < .05 versus control. CT, control; LPS, lipopolysaccharide; iNOS, inducible NO synthase; M, 100 bp DNA ladder.

The time course of the effect of LPS on the mRNA expression of IL-6 and TNF-α demonstrated a significant increase at 3 hours and 1 hour, respectively. Whereas the mRNA expression of IL-6 was still rising at 24 hours, that of TNF-α decreased considerably by 12-24 hours (Fig. 2A). Over the 24-hour time period, no change in the mRNA expression of IL-6 or TNF-α was observed in unstimulated cells. The extracellular release of both cytokines in LPS-stimulated cells also increased time-dependently, with the earliest increase occurring at 6 hours for IL-6 and at 3 hours for TNF-α (Fig. 2B). No significant change in the release of IL-6 was observed in unstimulated cells, but there was an increase in the release of TNF-α, suggesting that the cells constitutively synthesize and release this cytokine.

Figure 2.

Time-course of the effect of LPS on gene expression and the release of IL-6 and TNF-α. Cells were washed and placed in serum-free medium containing 1 μg/mL LPS. At the indicated time points, (A) gene expression and (B) release of IL-6 or TNF-α were determined. Data (RT-PCR) represent at least 3 separate experiments. Values in line graphs show the means ± SD of triplicate determinations from a representative experiment. *P < .001 vs. control. **P < .01 vs. control. ***P < .05 vs. 3-hour time point. In panel A, the control is the 0-hour time point; no change in IL-6 or TNF-α mRNA expression was observed in the control (vehicle-treated) cells over the 24-hour period. LPS, lipopolysaccharide; CT, control; IL-6, interleukin-6; TNF-α, tumor necrosis factor α; M, 100 bp DNA ladder.

Role of NF-κB and MAPKs in LPS-Induced Synthesis of NO, IL-6, and TNF-α.

Several important mechanisms of eukaryotic cell regulation involve signal transduction via MAPKs and NF-κB.24, 25 To understand the signaling mechanisms of LPS-induced expression of iNOS, IL-6, and TNF-α in activated HSCs, we determined activation of ERK1/2, p38, JNK/stress-activated protein kinase, and NF-κB. LPS induced phosphorylation (i.e., activation) of ERK1/2 within 5 minutes of stimulation (Fig. 3A) and phosphorylation of p38 at 15 minutes (Fig. 3B); activation of JNK occurred at a later time (45 min) (Fig. 3C). These effects of LPS were concentration-dependent with significant activation of all of the signaling molecules occurring at 0.01 μg/mL LPS as demonstrated by Western blot analysis (Supplementary Fig. 2). Consistent with previous observations in culture-activated HSCs,26, 27 NF-κB was constitutively active in our cell preparations (Fig. 4). LPS increased nuclear translocation of the P65 subunit of NF-κB at 30 and 60 minutes, which returned to the basal level by 2 hours (Fig. 4A). LPS-induced nuclear translocation of the P65 subunit of NF-κB was also demonstrated immunohistochemically (Fig. 4B).

Figure 3.

Effect of LPS on the activation of MAPKs. Cells were washed and placed in serum-free medium containing 1 μg/mL LPS. At the indicated time points, cellular protein was extracted and subjected to electrophoresis; Western blotting was performed to determine phosphorylated (A) ERK1/2, (B) p38, and (C) JNK as described in Materials and Methods. Nonphosphorylated MAPKs were determined in the same membranes after stripping. The membranes were also stripped to determine actin expression to ascertain equal loading. LPS, lipopolysaccharide; ERK, extracellular signal-regulated protein kinase; JNK, c-Jun NH2-terminal kinase.

Figure 4.

Effect of LPS on the nuclear translocation of NF-κB. Cells were washed and placed in serum-free medium containing 1 μg/mL LPS. (A) At the indicated time points, nuclear and cytosolic proteins were extracted and subjected to electrophoresis and Western blotting to assess NF-κB p65. (B) The cells were treated as described above and were fixed at 45 minutes of LPS stimulation. Immunohistochemistry was performed as described in Materials and Methods. Photomicrographs of control and LPS-treated cells were taken under the same exposure and magnification. In control cells, the stain is localized mostly in the cytosol and some nuclei, whereas in LPS-treated cells significantly increased staining can be seen in the nuclei. LPS, lipopolysaccharide; CT, control. Original magnification, ×20.

The above results suggested involvement of MAPKs and NF-κB in the LPS-induced synthesis of NO, IL-6, and TNF-α in HSCs. To ascertain this, the cells were pretreated with the specific inhibitors SB203580 (p38 kinase), PD98059 (ERK1/2 kinase), SP600125 (JNK kinase), and PDTC (NF-κB). None of the inhibitors affected basal NO (Fig. 5A), IL-6 (Fig. 5B) and TNF-α (Fig. 5C) levels. However, inhibition of p38 kinase and nuclear translocation of NF-κB blocked LPS-induced synthesis of NO as well as IL-6 and TNF-α. Inhibition of ERK1/2 kinase caused only modest inhibition of LPS-induced IL-6 synthesis and did not affect the synthesis of NO or TNF-α. Inhibition of JNK kinase did not significantly affect any of these LPS effects (Fig. 5). These results suggest specific roles of NF-κB and p38 in LPS-induced NO, IL-6, and TNF-α synthesis.

Figure 5.

Effect of inhibitors of MAPKs and NF-κB on LPS-induced synthesis of NO, IL-6, and TNF-α. Cells were washed and incubated in serum-free medium containing SB203580 (10 μmol/L; p38), PD98059 (10 μmol/L; ERK1/2), SP600125 (10 μmol/L; JNK), and PDTC (50 μmol/L; NF-κB) for 30 minutes. LPS (1 μg/mL) was then added and the incubation continued for 24 hours. NO, IL-6, and TNF-α in the culture medium were then analyzed. Values are expressed as the mean ± SD of triplicate determinations from a representative experiment. *P < .001 versus vehicle. **P < .01 versus vehicle. #P < .05 versus vehicle. ##P < .05 versus control. ***P < .001 versus control. LPS, lipopolysaccharide; PDTC, pyrrolidine dithiocarbamate; TNF-α, tumor necrosis factor α.

Role of ROS in LPS-Induced Synthesis of NO, IL-6, and TNF-α.

Next, we determined whether LPS stimulates ROS production in HSCs, and whether this effect is associated with activation of NF-κB and p38 and stimulation of NO, IL-6, and TNF-α synthesis. As shown in Fig. 6A, LPS did not affect superoxide generation by HSCs. LPS increased generation of H2O2 time-dependently up to 45 minutes, which then tended to decrease at 75 minutes (Fig. 6B). Flow cytometric analysis also demonstrated stimulation of H2O2 but not of superoxide generation by LPS (Fig. 6C).

Figure 6.

Effect of LPS on ROS generation. Cells were washed 3 times and loaded with (A) hydroethidine or (B) DCFH-DA, respectively (both at 10 μmol/L final concentration) in serum-free Dulbecco's modified Eagle medium without phenol red for 30 minutes at 37°C and challenged with 1 μg/mL LPS. At the indicated time points, hydroethidine and DCF fluorescence were determined as described in Materials and Methods. *P < .001 versus control. **P < .01 versus control. (C) After loading with hydroethidine and DCFH-DA as described above, the cells were stimulated with LPS for 45 minutes and processed for SO and H2O2 determination via flow cytometry. The X axis represents hydroethidine or DCF fluorescence; the Y axis represents the number of events. DHE, hydroethidine; LPS, lipopolysaccharide; DCF, 2′,7′-dichlorofluorescein.

We then investigated if H2O2 acts as a downstream effecter of LPS-induced synthesis of NO, IL-6, and TNF-α. H2O2 did not stimulate the synthesis of NO (Supplementary Fig. 3A) or IL-6 (Supplementary Fig. 3B); however, it did stimulate the synthesis of TNF-α, though to a smaller extent compared with that by LPS (Supplementary Fig. 3B). The time-course experiment demonstrated that H2O2-induced TNF-α synthesis increased significantly after 6 hours of stimulation (Supplementary Fig. 4A). The effect of H2O2 on TNF-α synthesis was also concentration-dependent (Supplementary Fig. 4B). The cells tended to detach at and beyond 250 μmol/L H2O2, indicating toxicity at higher concentrations. Neither time-course nor concentration-dependent experiments showed a significant increase in NO or IL-6 synthesis in H2O2-stimulated cells (data not shown). Together, these results suggest that H2O2 may partly mediate LPS-induced TNF-α but not NO or IL-6 synthesis in HSCs. Considering the baseline level of H2O2 in these cell preparations, a possibility of its contribution to the constitutive synthesis of TNF-α cannot be ruled out.

Because inhibition of the activation of both p38 and NF-κB blocked LPS-induced synthesis of NO, IL-6, and TNF-α, we determined a possible relationship between these signaling molecules. The cells were pretreated with SB203580 or PDTC prior to stimulation with LPS. PDTC did not affect LPS-induced p38 activity (Fig. 7A), but SB203580 inhibited nuclear translocation of NF-κB in LPS-stimulated cells (Fig. 7B), suggesting that NF-κB activation occurs downstream and is regulated by p38 phosphorylation.

Figure 7.

Effect of p38 and NF-κB inhibitors and H2O2 on activation of p38 and NF-κB. (A, B) Cells were washed and incubated in serum-free medium containing PDTC (50 μmol/L; NF-κB inhibitor) or SB203580 (10 μmol/L; p38 kinase inhibitor) for 30 minutes, then stimulated with LPS (1 μg/mL) for 45 minutes. (C, D) Cells were washed and incubated in serum-free medium containing 1 μg/mL LPS or 10 and 100 μmol/L H2O2. After incubation for 45 minutes, the cellular, cytosolic, and nuclear extracts were prepared and Western blot analysis was performed as described in Materials and Methods and in the legends for Figs. 3 and 4. Veh, vehicle; PDTC, pyrrolidine dithiocarbamate; CT, control; LPS, lipopolysaccharide.

The time-courses of p38 phosphorylation (Supplementary Fig. 2), NF-κB nuclear translocation (Fig. 4), and H2O2 generation (Fig. 6) suggest interactions between these processes. Therefore, we assessed whether H2O2 stimulates p38 phosphorylation and NF-κB nuclear translocation. As shown in Fig. 7C-D, H2O2 did not stimulate p38 phosphorylation or NF-κB nuclear translocation. This led to the hypothesis that p38 and NF-κB activation may be required for LPS-induced generation of H2O2. To test this hypothesis, cells were pretreated with SB203580 and PDTC before stimulation with LPS. Both PDTC and SB203580 inhibited LPS-induced H2O2 formation (Fig. 8). Because SB203580 inhibited NF-κB activation (Fig. 7B), these results suggest that p38 activation is the primary initiator of LPS signaling that leads to the generation of NO, IL-6, and TNF-α.

Figure 8.

Effect of inhibition of NF-κB and p38 on LPS-induced H2O2 synthesis. Cells were washed and incubated in serum-free medium containing 10 μmol/L SB203580 or 50 μmol/L PDTC for 30 minutes. LPS (1 μg/mL) was then added and incubation continued for 45 minutes. H2O2 generated under various conditions was determined via flow cytometry as described in Materials and Methods. The X axis represents DCF fluorescence; the Y axis represents the number of events. LPS, lipopolysaccharide; PDTC, pyrrolidine dithiocarbamate.

Discussion

Originally identified as a major storage site of retinoids and subsequently a cell type responsible for hepatic fibrosis, HSCs have undergone extensive research over the past 15 years, yielding strong evidence that they play important roles in hepatic growth modulation, immune function, and inflammation. Considering their strategic location, it is apparent that HSCs can influence the functions of hepatocytes and sinusoidal cells by releasing biologically active mediators. LPS has been implicated in several pathological conditions because of its ability to induce expression of inflammatory mediators such as IL-1, IL-6, IL-8, and TNF-α in many cell types, including macrophages, endothelial cells, and fibroblasts.28, 29 Our recent work showing LPS-induced synthesis of IL-6 and TNF-α, and NO via iNOS in HSCs21, 22 suggests their major influence on hepatic hemodynamic modulation, inflammation, and growth. However, the intracellular pathways elicited by LPS in HSCs that lead to the synthesis and release of NO, IL-6, and TNF-α are not known. We present evidence that activation of p38 by LPS initiates signaling via NF-κB and ROS (H2O2) leading to the induction of iNOS and expression of IL-6 and TNF-α.

Activation of MAPKs and nuclear translocation of NF-κB are major signaling pathways that lead to a variety of cellular effects including growth, apoptosis, and generation of cytokines.24, 25, 30, 31 In the murine macrophage cell line RAW264.7, LPS is shown to cause activation of ERK1/2, p38, and JNK/stress-activated protein kinase.32 Moreover, LPS-induced TNF-α expression is suppressed in microglia by pretreatment with the p38 inhibitor SB203580,33 and some studies have shown regulation of iNOS induction by MAPKs.34, 35 LPS has also been reported to elicit iNOS expression and to stimulate the release of NO via NF-κB in RAW264.7 cells,32 N9 microglial cells,36 and human myometrial cells.37 In HSCs, although LPS caused activation of ERK1/2, p38 and JNK, only activated p38 was found to increase the expressions of iNOS, IL-6, and TNF-α, as indicated by their strong inhibition by SB203580. PDTC, an inhibitor of NF-κB activation, also blocked LPS-induced NO, IL-6, and TNF-α synthesis, equally strongly suggesting interactions between p38 and NF-κB in eliciting these effects. The time-course experiment demonstrated that p38 phosphorylation occurs somewhat earlier than nuclear translocation of NF-κB after LPS stimulation. However, subtle changes in these biochemical pathways that may have critical influence on the cellular responses may be missed in the results of Western blot analysis. Therefore, we determined their activation in cells preincubated with the respective inhibitors. By this approach, inhibition of p38 was found to block nuclear translocation of NF-κB in LPS-stimulated HSCs, suggesting that the former pathway initiates IκB kinase stimulation and separation of NF-κB from the NF-κB–IκB kinase complex for its nuclear translocation. Similar to our results, in a previous study inhibition of p38 was shown to inhibit LPS-induced nuclear translocation of NF-κB in neutrophils.38 Interestingly, nuclear translocation of NF-κB in peripheral neutrophils was found to be modulated by p38 in patients with acute lung injury.39 Extending these observations to hepatic pathology, it is likely that a similar relationship between LPS-induced p38 and NF-κB activation may have significant clinical relevance.

NF-κB is known to regulate the expression of proinflammatory cytokines and immune mediators.40, 41 These and several other effects of NF-κB on gene expression have made this one of the most extensively studied signaling molecules. The observations that (a) the most widely used inhibitors of NF-κB activation, N-acetylcysteine and PDTC, can act as antioxidants,41–43 (b) many agents that stimulate ROS production also activate NF-κB,41–43 and (c) in certain cell types, H2O2 stimulates activation of NF-κB,44 have prompted the hypothesis that ROS is a stimulus for activation of NF-κB. However, our results show that activation of NF-κB and the increase in H2O2, occur at about the same time and further demonstrate that NF-κB activation is required for the production of ROS. Although it can be argued that PDTC, the inhibitor used in this study, has a dual effect and thus can block both NF-κB activation and H2O2 synthesis, the inhibition of LPS-induced NO synthesis by PDTC and the inability of H2O2 to stimulate NF-κB activation or to elicit NO synthesis suggest that PDTC blocks LPS-induced NF-κB activation, which leads to the inhibition of H2O2 synthesis. Furthermore, the inhibition of NF-κB activation by PDTC in a cell-free system in which ROS production does not occur45 also supports the contention that in our experiments, PDTC inhibited H2O2 synthesis by blocking activation of NF-κB.

In several LPS-stimulated cell types, there is an association between ROS generation and activation of p38 and NF-κB. p38 was found to be activated by intracellular ROS as well as by exogenously administered H2O2.46–48 Mitochondrial ROS and p38 have been implicated in the elicitation of gene transcription responses to a variety of stimuli.49, 50 In cardiac fibroblasts, ROS-induced IL-6 expression was shown to be p38- and ERK-dependent.46 However, our results show that exogenous H2O2 does not stimulate p38. Moreover, exogenous H2O2 does not stimulate the synthesis of NO and IL-6, and its effect on the synthesis of TNF-α was modest compared with that of LPS. These results suggest that H2O2 does not regulate LPS-induced p38 activation. In contrast, both SB203580 and PDTC strongly inhibited LPS-stimulated synthesis of H2O2, NO, IL-6, and TNF-α, and blockade of p38 activation led to inhibition of LPS-induced nuclear translocation of NF-κB. Together, these observations indicate that LPS-induced synthesis of NO, IL-6, and TNF-α follows a systematic pathway of p38 activation and NF-κB activation, with a role for H2O2 in TNF-α production.

Our results show that H2O2 is partly responsible for LPS-induced synthesis of TNF-α, and activation of p38 is required for its generation. It is unclear as to why this signaling pathway uses H2O2 to exert a part of its effects. The precise role of ERK1/2 and JNK signaling stimulated by LPS is also unclear. However, considering the unique location and versatility of HSCs in orchestrating a variety of intercellular signaling processes, the roles of LPS-induced H2O2 synthesis and ERK1/2 and JNK signaling should be clarified by future investigations.

In conclusion, we have demonstrated that p38/NF-κB signaling in rat stellate cells is a primary mechanism of LPS-induced generation of NO, IL-6, and TNF-α that are known to be major players in hepatic hemodynamic regulation, inflammation, and immune responses.

Acknowledgements

We appreciate the excellent technical assistance of Kristin Anselmi.

Ancillary