Essential Role of STAT1 in Caspase-Independent Cell Death of Activated Macrophages through the p38 Mitogen-Activated Protein Kinase/STAT1/Reactive Oxygen Species Pathway

Mice.C57BL/6, C3H/HeJ, and C3H/HeN mice were purchased from Jackson Laboratory. STAT1 and IRF-1 knockout mice were from the Charles River Laboratory. All mice were maintained in a specific-pathogen-free condition at Samsung Biomedical Research Institute. Unless otherwise mentioned, all knockout mice were of the C57BL/6 background.

Mφ preparation and culture.Peritoneal Mφ were collected from 8- to 10-week-old mice 4 days after intraperitoneal injection of 3% thioglycolate (Difco), plated on plastic tissue culture plates, and incubated at 37°C for 3 h. Nonadherent cells were removed by three repeated washings with fresh Dulbecco's modified Eagle medium (DMEM), and the adherent Mφ were cultured overnight in DMEM supplemented with 10% fetal bovine serum, 100 U/ml penicillin, 100 μg/ml streptomycin, and 2 mM glutamine.

Antibodies and other reagents.Antibodies were obtained as follows. Monoclonal anti-TNF-α-neutralizing antibody, clone MP6-XT22.11, was produced from hybridoma culture and prepared as described previously (58). Anti-beta interferon (IFN-β)-neutralizing antibody was from R&D Systems. Anti-caspase 3, anti-STAT1, anti-pY701-STAT1, and anti-p-p38 antibodies were from Cell Signaling. Anti-pS727-STAT1 antibody was from Upstate. Anti-p38, anti-IRF1, anti-α-tubulin, anti-IκBα, rat control immunoglobulin G (IgG), and rabbit control IgG antibodies were from Santa Cruz Biotechnology. Anti-TLR4 monoclonal antibody was from eBioscience. Anti-caspase 8 and anti-gp91 antibodies were from BD Pharmingen. Anti-c-myc-antibody was from Sigma. Escherichia coli serotype 0111:B4 was purchased from Sigma. Recombinant mouse TNF-α, mouse IL-1β, mouse IFN-β, and mouse IFN-γ were from R&D Systems. The caspase inhibitors (z-VAD-fmk, z-DEVD-fmk, z-IETD-fmk, z-LEHD-fmk, and z-YVAD-fmk) and colorimetric caspase substrates (Ac-DEVD-pNA [N-acetyl-Asp-Glu-Val-Asp-p-nitroanilide] and Ac-IETD-pNA) were purchased from Enzyme Systems Products. Fluorogenic caspase substrates (Ac-DEVD-AMC [N-acetyl-Asp-Glu-Val-Asp-7-amino-4-methylcoumarin] and Ac-IETD-AMC) were obtained from BD Clontech. N-Monomethyl-l-arginine (NMMA) and SB203580 were obtained from Calbiochem. All other chemicals were from Sigma, unless stated otherwise.

Cell death assays. (i) Crystal violet assay.The crystal violet uptake assay was performed in 96-well plates 12 h after treatment. For crystal violet staining, the medium was removed and cells were washed with phosphate-buffered saline (PBS). Then, 100 μl of 0.2% crystal violet in 10% ethanol was added and incubated for 30 min. To remove excess dye, cells were washed several times with tap water and subsequently dried. Dye absorbed by the cells was eluted with 0.5% sodium dodecyl sulfate, and absorbance was then read at 570 nm.

(ii) TUNEL assay.DNA fragmentation was visualized by incorporation of fluorescent oligonucleotides, using the terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling (TUNEL) kit (Roche). Mφ were grown on Lab-Tek chamber slides (Nunc), and staining of apoptotic cells was performed 6 h after treatment as instructed by the manufacturer.

(iii) Annexin V-FITC/PI assay.Annexin V-fluorescein isothiocyanate (FITC) binds to exposed phosphatidylserine on apoptotic and necrotic cells, and propidium iodide (PI) gains entry into necrotic cells. For determination of apoptosis and necrosis, Mφ were treated for 12 h and analyzed by FACScan (Becton Dickinson), using the annexin V-FITC apoptosis detection kit from BD Pharmingen.

Caspase activation.Pancaspase activity was detected using CaspACE (Promega) in accordance with the manufacturer's instructions. Soon after, Mφ were pretreated with 2 μM FITC-VAD-fmk for 1 h in the absence or presence of 50 μM zVAD-fmk and then stimulated with 100 ng/ml LPS. In some experiments, Mφ were treated with etoposide (30 μM) in order to induce apoptosis. After the indicated times, cells were recovered, washed with cold PBS, and analyzed by FACScan using CellQuest software. Caspase 3-like and 8-like activities were measured using commercial caspase assay kits (BD Clontech) according to the supplier's instruction. In brief, fluorogenic caspase 3 (Ac-DEVD-AMC) or caspase 8 substrate (Ac-IETD-AMC) was incubated with LPS (100 ng/ml)- or staurosporine (5 μM)-treated cell lysates for 1 h. Liberated AMC was measured using a fluorometric plate reader. Activation of caspase 3 or caspase 8 was also confirmed using colorimetric caspase 3 (Ac-DEVD-pNA) or caspase 8 substrate (Ac-IETD-pNA).

Measurement of ROS.Intracellular ROS levels were determined as previously described with some modifications (43). For measurement of superoxide anions and peroxide, Mφ were loaded with 2 μM dihydroethidium (DHE) (Molecular Probes) and 5 μM 2′,7′-dichlorofluorescein diacetate (DCFH-DA; Molecular Probes), respectively, in phenol red-free medium for 30 min at 37°C. After each treatment, cells were washed with cold PBS and resuspended in 0.5 ml PBS supplemented with 1% fetal bovine serum. Accumulation of intracellular fluorescence was analyzed by a FACScan (Becton Dickinson).

ELISA.Supernatant from cultured Mφ was used to determine the amount of TNF-α, NO, or IL-1β released. Enzyme-linked immunosorbent assays (ELISA) were carried out according to the manufacturer's recommendation (R&D Systems).

Detection of surface TLR4.Mφ were washed in PBS-2.5% horse serum. Fc block was done with anti-FcγRII/III monoclonal antibody (Pharmingen) and 10% normal mouse serum. Cells were stained with anti-TLR4 monoclonal antibody and then with FITC-conjugated secondary antibody (Zymed) for analysis by a FACScan (Becton Dickinson).

Western blot analysis.Mφ were lysed in lysis buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1% Triton X-100, 5 mM EDTA, 0.5 mM Na₃VO₄, 50 mM NaF, 1 mM phenylmethylsulfonyl fluoride, and protease inhibitor cocktail [Roche]) for 20 min on ice. Cell debris was removed by centrifugation at 20,000 × g for 15 min at 4°C. Protein concentration in cell lysates was determined using a Bio-Rad protein assay kit. An equal amount of protein for each sample was separated by 8, 10, or 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and subsequently transferred to a Hybond enhanced chemiluminescence (ECL) nitrocellulose membrane (Amersham). After blocking with 5% skim milk, membranes were sequentially incubated with primary antibodies and then with horseradish peroxidase-conjugated secondary antibody (Amersham). This was followed up by ECL detection (Amersham).

DNA constructs and transfections.To assess the effects of STAT1-701, STAT1-727, and p38 MAPK on zVAD/LPS-induced cell death, Raw264.7 cells in six-well plates were cotransfected with 0.5 μg of pEGFP (BD Clontech) together with either 1.5 μg of mutant STAT1-Y701F, STAT1-S727A (gifts from J. E. Darnell, Jr, Rockefeller University) or dominant-negative p38 MAPK (T180A and Y182F; a gift from E. J. Choi, Korea University) using FuGENE 6 reagent (Roche). After 24 h of transfection, cells were treated with zVAD/LPS for 16 h. Percent cell death was determined by calculating the fraction of cells positive for both PI and green fluorescent protein (GFP) as a fraction of total GFP-positive cells. Raw264.7 cell lines stably expressing dominant-negative Rac1 (Rac1^N17) were prepared by cotransfection of pEXV control vector or pEXV-RacN17 (gifts from A. Hall, University College London, London, United Kingdom) together with pcDNA-Neo^R plasmids, and stable transfectants were selected for 3 weeks in the presence of G418 as previously described (32).

Chromatin immunoprecipitation.Mφ were exposed to LPS (100 ng/ml) or IFN-γ (100 U/ml) for 1 h, with or without zVAD pretreatment (50 μM) for 1 h. These cells were cross-linked with 1% formaldehyde for 15 min at 25°C and washed twice in PBS. Chromatin immunoprecipitation (ChIP) assays were performed using a ChIP assay kit (Upstate Biotechnology) in accordance with the manufacturer's instructions. Monoclonal anti-STAT1 antibody (Cell Signaling) was used for immunoprecipitations. Semiquantitative PCR was performed with recovered DNA or 1% input DNA samples using primer sets specific for the mouse gp91 (from −141 to +63 relative to the transcription start; forward, 5′-TTTGTTATGGGAACAGCCTTTC-3′; reverse, 5′-CAATGACAAAGATGGAGAGTCC-3′). PCRs of various cycle numbers were repeated using different amounts of templates to ensure that results were within the linear range of PCR.

Reporter assay.gp91 reporter activity was measured using the Dual-Luciferase reporter assay system (Promega). In brief, Mφ in 24-well plates were cotransfected with 0.5 μg of the wild-type gp91 luciferase reporter or the mutant gp91 luciferase reporter containing a mutation in the STAT1 binding site (gifts from A. Kumatori, Nagasaki University, Nagasaki, Japan) together with 0.02 μg of Renilla luciferase gene under the herpes simplex virus thymidine kinase promoter (pRL-TK, Promega) using FuGENE 6 reagent. After 24 h of transfection, cells were treated with LPS (100 ng/ml) or IFN-γ (100 U/ml) for 2 h, with or without zVAD pretreatment (50 μM) for 1 h. Reporter activities were presented as relative firefly luciferase activities of gp91 reporter normalized to Renilla luciferase activities.

Electrophoretic mobility shift assay.Nuclear extracts were prepared from Mφ treated with LPS (100 ng/ml) for the indicated times, with or without zVAD pretreatment for 1 h (50 μM). Electrophoretic mobility shift assays were then performed using the nuclear extracts after incubation with labeled probe containing the consensus NF-κB binding sequence (Promega) as previously described (57).

LPS-induced early caspase activation is required for Mφ survival.We initially studied the role of caspases in the death of activated Mφ. LPS, a well-known activator of Mφ, induced elongation and cell spreading, but no evidence of cell death was observed (Fig. 1A). Treatment with only zVAD-fmk, a membrane-permeable pancaspase inhibitor, did not cause any noticeable effect on the morphology or viability of Mφ. However, in the presence of zVAD-fmk, LPS-stimulated Mφ underwent massive cell death, as evidenced by morphological changes and a significant increase in the number of PI- or TUNEL-positive cells. This type of death process occurred as early as 2 h after treatment and reached over 75% by 8 h (data not shown). Interestingly, the majority of dying cells (>70%) showed necrotic changes which were characterized by cellular swelling, loss of membrane integrity, and apparently normal nuclei. However, the other subpopulation of dying cells (<30%) showed atypical apoptotic changes, such as cell shrinkage and moderate chromatin condensation but did not show DNA laddering on agarose gels (Fig. 1A and data not shown). Fluorescence-activated cell sorter (FACS) analysis of dying cells by annexin V/PI staining supported the premise that zVAD/LPS induced both necrosis-like and apoptosis-like changes, as judged by a significant increase in the percentage of cells that were exclusively positive for either annexin V or PI (Fig. 1B). To further investigate whether activation of caspases participates in Mφ survival, we labeled Mφ with the fluorescent caspase substrate FITC-VAD-fmk, since most known caspases can cleave the tripeptide VAD. Caspase-like activity triggered by LPS was detectable as early as 30 min after treatment and reached a maximum at 2 h and was completely blocked by pretreatment with zVAD-fmk (Fig. 1C and D). In control experiments, caspase activation was detected after etoposide treatment, an event that was also blocked by zVAD-fmk and was associated with apoptosis of Mφ (Fig. 1E). To further investigate whether a specific caspase(s) was involved in the Mφ survival, we tested more specific caspase inhibitors on the LPS-activated Mφ. The caspase 3-, 6-, or 7-like inhibitor DEVD-fmk and caspase 9-like inhibitor LEHD-fmk partially inhibited the survival of LPS-treated Mφ, whereas the caspase 8-like inhibitor IETD-fmk and caspase 1-like inhibitor YVAD-fmk had no influence on the survival of LPS-stimulated Mφ (Fig. 2A). Consistent with inhibitor studies, we observed that LPS induced a clear and transient activation of caspase 3 but not caspase 8 in Mφ, with its maximum effect at 2 h (Fig. 2B and C). The kinetics of caspase 3 activation was closely related to those of LPS-triggered caspase-like activity detected by fluorescent pancaspase substrate FITC-VAD. Furthermore, our results suggest that, in contrast to staurosporine, which caused strong apoptotic caspase 3 and 8 activation, LPS triggered a nonapoptotic caspase activation which may play a role in the survival of Mφ. Although we cannot rule out the involvement of other isoforms of caspase or another zVAD-sensitive molecule(s), these results indicate the requirement for caspase 3-like and 9-like activities in LPS-mediated survival of Mφ. Taken together, our results suggest that LPS induces early caspase activation, the abrogation of which rapidly leads to death of activated Mφ.

• Open in new tab
• Download powerpoint

FIG. 1.

Abrogation of LPS-induced early caspase activation leads to Mφ death. (A) Primary murine Mφ were treated with zVAD (50 μM), LPS (100 ng/ml), or zVAD/LPS. Cells were observed under a phase-contrast microscope after treatment for 8 h (top) or under a fluorescent microscope after treatment for 8 h and Hoechst 33258/PI double staining (middle) or after treatment for 4 h and TUNEL staining (bottom). (B) Mφ were treated for 8 h, as in panel A, and analyzed by FACS after staining with annexin V-FITC and with PI. (C) Mφ were loaded with FITC-VAD and left untreated or exposed to LPS for 0.5, 1, or 2 h. Cell fluorescence intensity was analyzed by FACS. FITC-VAD-loaded Mφ were exposed to LPS for 2 h (D) or etoposide (30 μM) for 4 h (E) with or without zVAD pretreatment for 1 h and analyzed as in panel C. Shown are the results of three independent experiments.

• Open in new tab
• Download powerpoint

FIG. 2.

Caspase 3-like and 9-like activities are triggered by LPS in Mφ. (A) Mφ were either left untreated or treated with LPS (100 ng/ml) for 20 h after pretreatment with the indicated caspase inhibitors for 1 h (50 μM). Cell viability is shown as a percentage of untreated cells. Mφ were exposed to LPS (100 ng/ml) or staurosporine (5 μM), with or without zVAD pretreatment for 1 h (50 μM), and activation of caspase 3 (B) and 8 (C) was analyzed by a caspase substrate assay using fluorogenic caspase substrates (top) and by Western blotting (bottom) at the indicated times. Active caspase fragments (act) were noted. LPS triggered the maximum caspase 3 activation at 2 h, indicated by an arrow. Shown are the results of three independent experiments.

Superoxide by caspase inhibition as an effector in zVAD/LPS-induced Mφ death.ROS have been previously implicated in necrotic or caspase-independent death in several types of cells (31, 33, 34, 39). To investigate whether ROS are produced by LPS or zVAD/LPS, we used the superoxide anion-specific probe DHE and the hydrogen peroxide/hydroxyl radical-specific probe DCFH-DA. Alone, LPS moderately increased the level of superoxide in Mφ, while combined treatment with zVAD/LPS profoundly increased the superoxide level (Fig. 3A). In contrast, peroxide levels were not affected or even decreased after treatment with zVAD/LPS but were, in fact, moderately increased by LPS alone (Fig. 3B). zVAD alone did not affect superoxide or hydrogen peroxide levels in Mφ (data not shown). We then tested the possible involvement of superoxide in zVAD/LPS-induced death of Mφ using selective ROS inhibitors. Intriguingly, tempol and tiron, specific inhibitors of superoxide, were remarkably capable of blocking both zVAD/LPS-induced Mφ death and superoxide production in a dose-dependent manner, individually (Fig. 3C to E and data not shown). However, other antioxidants, such as butylated hydroxyanisole, glutathione, N-acetyl-l-cysteine (data not shown),and uric acid, did not significantly attenuate zVAD/LPS-induced Mφ death. Catalase, a hydrogen peroxide-eliminating enzyme (data not shown) and NMMA, an NO synthase inhibitor, were also without significant effect. These results suggest that generation of superoxide, but no other ROS, critically contributes to zVAD/LPS-induced death of Mφ. Consistent with TLR4 as a pattern recognition receptor for LPS, Mφ from C3H/HeJ mice, whose TLR4 is inactive, but not their control C3H/HeN mice, were completely resistant to the killing by zVAD/LPS (data not shown). The superoxide level remained unaffected by LPS and zVAD/LPS treatment of Mφ from C3H/HeJ mice (Fig. 3F).

• Open in new tab
• Download powerpoint

FIG. 3.

Superoxide is responsible for zVAD/LPS-induced death of Mφ. (A) Mφ were either left untreated (control) or treated with LPS (100 ng/ml) for 2 h with or without zVAD pretreatment for 1 h (50 μM). Intracellular superoxide levels were measured by DHE staining. (B) Mφ were treated as in panel A, and intracellular peroxide was detected by DCFH-DA staining. (C) Mφ were treated with zVAD/LPS for 8 h in the absence or presence of tempol (10 mM), tiron (10 mM), butylated hydroxyanisole (BHA; 5 mM), glutathione (GSH; 5 mM), uric acid (2 mM), or NMMA (2 mM). Cell viability is shown as a percentage of untreated cells. (D) Mφ were treated with zVAD/LPS, as for panel C, with the indicated concentrations (5 to 20 mM) of tempol or tiron. Cell viability is shown as in panel C. (E) Mφ were treated with zVAD/LPS for 2 h with or without tempol (20 mM). Intracellular superoxide levels were measured by DHE staining. (F) Mφ from C3H/HeJ mice were treated and analyzed, as shown in panel A. Shown are the results of three independent experiments, and results in panels C and D are the means ± standard errors.

Next, we investigated the sources of superoxide which is responsible for zVAD/LPS-induced death of Mφ. ROS emanate from mitochondrial and nonmitochondrial enzymes (18, 40). When Mφ were pretreated with diphenylene iodonium (DPI) or 4-(2-aminoethyl)benzenesulfonylfluoride (AEBSF), inhibitors that block ROS generation by NADPH oxidase, both superoxide production (Fig. 4A and data not shown) and cell death (Fig. 4B) were significantly attenuated after treatment with zVAD/LPS. We obtained similar results with rotenone, an inhibitor of complex I of the mitochondrial respiratory chain (Fig. 4B), suggesting that superoxide largely comes from both NADPH oxidase and mitochondria.

• Open in new tab
• Download powerpoint

FIG. 4.

Role of NADPH oxidase and mitochondria in superoxide production from Mφ by zVAD/LPS. (A) Mφ were treated with zVAD (50 μM)/LPS (100 ng/ml) for 2 h in the absence or presence of DPI (10 μM). Intracellular superoxide levels were measured by DHE staining. (B) Mφ were treated with zVAD/LPS for 8 h in the presence of DPI (10 μM), AEBSF (250 μM), or rotenone (100 μM), and cell viability was then assessed by crystal violet assay. Shown are the results of three independent experiments, and the results in panel B are the means ± standard errors.

Cell-autonomous death of activated Mφ by zVAD/LPS.Because activated Mφ stimulated by LPS could secrete a significant level of proinflammatory, potentially cytotoxic molecules, including TNF-α, IL-1β, NO, and IFN-β (1, 59) (data not shown), we tested whether these molecules contribute to death and superoxide production of Mφ subsequent to treatment with zVAD/LPS. We first considered the possible involvement of TNF-α, a predominant cytokine released in response to LPS treatment. With this objective, we treated Mφ with zVAD/LPS in the presence of a monoclonal anti-TNF-α neutralizing antibody whose activity had previously been confirmed (58). Anti-TNF-α neutralizing antibody only slightly attenuated the death of Mφ treated with zVAD/LPS (17 to 25%) (Fig. 5A). We also evaluated TNF-α production from zVAD/LPS-treated Mφ. While TNF-α in combination with zVAD induced Mφ death in a dose-dependent manner, the amount of TNF-α (Fig. 5B) released from zVAD/LPS-stimulated Mφ (<0.5 ng/ml) was not sufficient to induce significant Mφ death or superoxide production (Fig. 5C and D). Collectively, these results suggest that TNF-α does not significantly contribute to zVAD/LPS-induced death of Mφ. These findings are consistent with the previously reported minor role of TNF-α in the caspase-independent death of dendritic cells (10) and RAW264.7 Mφ-like cells (26). Despite the structural and signaling similarities between TLR4 and IL-1R, IL-1β did not induce noticeable death of zVAD-pretreated Mφ up to 10 ng/ml (Fig. 5E). Consistently, IL-1β did not increase superoxide level in zVAD-pretreated Mφ (Fig. 5F). NO has been well known as one of the major contributors to Mφ death (30), and we confirmed a significant role of NO in Mφ death as induced by IFN-γ/LPS (data not shown). However, zVAD/LPS did not induce significant NO release (data not shown), nor did l-NMMA protect Mφ from zVAD/LPS-induced cell death (Fig. 3C) in a condition where it inhibited IFN-γ/LPS-induced Mφ death. Next, we determined the contribution of IFN-β in zVAD/LPS-induced death of Mφ, since a previous study had reported that LPS-induced TLR4 engagement produces IFN-β in a MyD88-independent, TIRAP-dependent manner (59). When we exposed Mφ to zVAD/LPS in the presence of anti-IFN-β-blocking antibody that abrogated IFN-β- or zVAD/LPS-induced tyrosine phosphorylation of STAT1 (Fig. 5G), zVAD/LPS-induced Mφ death was not affected by the antibody, indicating no appreciable role for IFN-β in Mφ death by zVAD/LPS (Fig. 5H). Taken together, these data strongly suggest that zVAD/LPS-induced death of Mφ is mostly due to intrinsic cell-autonomous signaling mediated by LPS through TLR4, rather than proinflammatory or cytotoxic molecules released from activated Mφ.

• Open in new tab
• Download powerpoint

FIG. 5.

Cell-autonomous death of activated Mφ by zVAD/LPS. (A) Mφ were exposed to zVAD (50 μM)/LPS (100 ng/ml) for 8 h with or without the indicated concentrations (2.5 to 10 μg/ml) of anti-TNF-α-blocking antibody. (B) Mφ were treated with LPS for the indicated periods of time with or without zVAD pretreatment for 1 h. TNF-α release was then determined by ELISA. (C) Mφ were exposed to the indicated concentrations (0.1 to 1 ng/ml) of TNF-α for 8 h with zVAD pretreatment for 1 h. (D) Mφ were either left untreated (control) or treated with TNF-α (500 pg/ml) for 2 h after zVAD pretreatment for 1 h. Intracellular superoxide levels were measured by DHE staining. (E) Mφ were exposed to the indicated concentrations (2.5 to 10 ng/ml) of IL-1β for 8 h with zVAD pretreatment for 1 h. (F) Mφ were either left untreated (control) or treated with IL-1β (10 ng/ml) for 2 h after zVAD pretreatment for 1 h. Intracellular superoxide levels were measured by DHE staining. (G) Mφ were treated with IFN-β (100 U/ml) for 30 min or zVAD/LPS for 2 h with or without neutralizing anti-IFN-β-blocking antibodies (10 μg/ml). STAT1 activation was analyzed using anti-phosphotyrosine 701-specific antibody. Identical samples were probed using specific antibody to detect total STAT1. (H) Mφ were exposed to zVAD/LPS for 8 h with or without the indicated concentrations (2.5 to 10 μg/ml) of anti-IFN-β-blocking antibody. Cell viability in panels A, C, E, and H was assessed by crystal violet assay. Shown are the results of three independent experiments. The results in panels A to C, E, and H are the means ± standard errors.

Upregulation of STAT1 signaling by caspase inhibition is responsible for zVAD/LPS-induced death of Mφ.As activated Mφ death was primarily mediated by ROS imbalance during caspase inhibition in a cell-autonomous fashion, we hypothesized that the key signaling molecule(s) required for this death process could be inactivated by caspases downstream of TLR4 engagement. Recent studies have revealed that signaling of apoptosis through TLRs critically involves the adapter molecule TRIF, but not MyD88, in bacterially infected murine Mφ (16, 47). Thus, we chose to investigate the signaling molecule(s) downstream of TRIF which could potentially be inactivated by caspases. Among these potential candidates (1, 2, 16), we focused on STAT1 because STAT1, whose induction or activation is known to promote cell death, is activated by TLR4 engagement (45, 59) and is also sensitive to proteolytic inactivation by caspases (25, 53). Despite the well-known role of STAT1 in several models of caspase-dependent cell death (6, 28, 44, 52, 57, 58), very little is known about the role of STAT1 in the caspase-independent death process. First, we examined the protein level of STAT1 and found that it was dramatically increased by zVAD/LPS but was decreased by LPS alone. The minimum level was reached at 2 h and increased thereafter (Fig. 6A), suggesting a reciprocal relationship between STAT1 levels and LPS-induced early caspase-like activity (Fig. 1C). When the same blot was exposed for longer intervals, cleavage forms of STAT1 (STAT1*) were detected in LPS-treated Mφ. Furthermore, the cleavage was nullified by zVAD pretreatment, indicating cleavage of STAT1 by activated caspases. We then studied the phosphorylation status of STAT1, which determines STAT1 activity. Notably, at tyrosine 701 and at serine 727, the phosphorylation of STAT1 was significantly increased in samples treated with zVAD/LPS. This indicates that caspase inhibition significantly upregulated both the protein level and the activity of STAT1 in activated Mφ (Fig. 6A). The reciprocal regulation of STAT1 and LPS-mediated early caspase activation was further strengthened by the observation that selective caspase inhibitor DEVD-fmk or LEHD-fmk moderately increased both STAT1 level and phosphorylation at serine 727 (Fig. 6B), consistent with the selective death of LPS-treated Mφ by these inhibitors (Fig. 2A). To determine the involvement of STAT1 signaling in Mφ death by zVAD/LPS, we employed Mφ from STAT1^−/− mice (38) and compared the results with those from their wild-type littermate controls. Intriguingly, Mφ from STAT1^−/− mice were clearly resistant to zVAD/LPS-induced cell death, as judged by light microscopy, TUNEL staining, annexin V/PI staining, or crystal violet assay (Fig. 6C to E). This resistance against cell death was not due to the changes in the expression of the TLR4 receptor (Fig. 6F). Consistent with the marked resistance of STAT1^−/− Mφ against zVAD/LPS-induced death, superoxide production after treatment with zVAD/LPS was abrogated in STAT1^−/− Mφ (Fig. 6G). Taken together, these results suggest a central role for STAT1 signaling in Mφ death and superoxide production through zVAD/LPS treatment.

• Open in new tab
• Download powerpoint

FIG. 6.

STAT1 signaling is responsible for zVAD/LPS-induced death of Mφ. (A) Mφ were treated with LPS (100 ng/ml) for the indicated time periods with or without zVAD pretreatment for 1 h (50 μM). Cell extracts were analyzed for total STAT1 by Western blotting (top). STAT1 cleavage products (STAT1*) were shown after long exposure. STAT1 activation was analyzed using the same samples employing antibodies specific for phosphorylated forms of STAT1 at tyrosine 701 or at serine 727 (bottom). (B) Mφ were treated with LPS (100 ng/ml) for 1 h with or without the indicated caspase inhibitors for 1 h (50 μM). Cell extracts were analyzed by Western blotting for total STAT1 (top) or STAT1 phosphorylation at serine 727 (bottom). (C to E) Mφ from STAT1^+/+ or STAT1^−/− mice were either untreated (control) or treated with zVAD (50 μM)/LPS (100 ng/ml) for 8 h. Shown are representative results observed by light microscopy (C; top), TUNEL assay (C; bottom), annexin V-FITC/PI staining (D), or crystal violet assay (E). (F) Mφ from STAT1^+/+ or STAT1^−/− mice were analyzed using antibodies specific for TLR4 or control IgG to detect the surface expression of TLR4 receptor. (G) Mφ from STAT1^+/+ or STAT1^−/− mice were exposed to LPS for 2 h with or without zVAD pretreatment for 1 h. Intracellular superoxide levels were measured by DHE staining (top), and peroxide was detected by DCFH-DA staining (bottom). Shown are the results of three independent experiments. The results in panel E are the means ± standard errors.

STAT1 activation by p38 MAPK is critical for zVAD/LPS-induced death of Mφ.Next, we investigated the activation mechanism of STAT1, whose signaling is critically involved in Mφ death. Phosphorylation of STAT1 at serine 727 is mediated by p38 MAPK (12, 27), while that of STAT1 at tyrosine 701 is mediated by JAKs through LPS-induced IFN-β (59) (Fig. 5G). Because IFN-β and tyrosine phosphorylation of STAT1 did not play appreciable roles in zVAD/LPS-induced Mφ death (Fig. 5H), we chose to study the role of p38 MAPK, which has been implicated in stress-induced cell death (4). zVAD/LPS treatment activated p38 MAPK in Mφ with faster kinetics when compared to LPS alone (Fig. 7A). We also studied whether inhibition of p38 MAPK activity could affect the cellular response of Mφ to zVAD/LPS treatment. A p38 MAPK-specific inhibitor, SB203580, effectively abrogated Mφ death and superoxide production by zVAD/LPS (Fig. 7B and C). STAT1 phosphorylation at serine 727 was also abrogated by SB203580 (Fig. 7D). The importance of STAT1 phosphorylation at serine 727 by p38 MAPK in zVAD/LPS-induced Mφ death was further investigated using two STAT1 mutants (STAT1-Y701F and STAT1-S727A) and a dominant-negative p38 MAPK. Overexpression of either STAT1-S727A mutant or dominant-negative p38 MAPK but not the STAT1-Y701F mutant resulted in much lower levels of Mφ death (Fig. 7E). Taken together, these results indicate a key role for p38 MAPK in STAT1 activation and superoxide production, which lead to Mφ death.

• Open in new tab
• Download powerpoint

FIG. 7.

STAT1 activation by p38 MAPK is critical for zVAD/LPS-induced death of Mφ. (A) Mφ were treated with LPS (100 ng/ml) for the indicated times, with or without zVAD pretreatment for 1 h (50 μM). p38 MAPK activation was analyzed using specific anti-phospho-p38 MAPK antibody. (B) Mφ were exposed to zVAD/LPS for 8 h, with or without SB203580 pretreatment (5 to 10 μM) for 1 h. Cell viability was then assessed by crystal violet assay. (C) Mφ were treated with zVAD/LPS for 2 h in the absence or presence of SB203580 (10 μM). Intracellular superoxide levels were measured by DHE staining. (D) Mφ were treated with LPS for 30 min, with or without pretreatment with zVAD and/or SB203580 (10 μM) for 1 h. STAT1 activation was analyzed using the anti-phosphoserine 727-specific antibody. The same blots were probed using anti-STAT1 or -p38 MAPK antibody for equal loading. (E) Raw264.7 cells overexpressing mutant STAT1-Y701F, STAT1-S727A, or dominant-negative p38 MAPK were treated with zVAD/LPS for 16 h. Cell death was then assessed by PI incorporation into the transfected cells. Shown are the results of three independent experiments. The results in panels B and E are the means ± standard errors.

STAT1 activation is mediated by ROS-stimulated p38 MAPK in a positive-feedback mechanism.Next, we sought the mechanism of p38 MAPK activation in Mφ treated with zVAD/LPS. On the one hand, ROS was abundantly produced by Mφ after zVAD/LPS treatment through p38 MAPK/STAT1 activation (Fig. 6G and 7C and D). On the other hand, p38 MAPK is reportedly activated by ROS (4). Collectively, these suggest a potential positive feedback between ROS production and p38 MAPK/STAT1 activation. Thus, we investigated the role of ROS in p38 MAPK activation leading to STAT1 activation and Mφ death. When Mφ were pretreated with tempol or DPI, which scavenges superoxide, p38 MAPK activation by zVAD/LPS was markedly attenuated, indicating that zVAD/LPS activates p38 MAPK in an oxidant-dependent manner (Fig. 8A). We also studied whether ROS could directly activate p38 MAPK in Mφ. The superoxide donor menadione activated p38 MAPK and phosphorylation of STAT1 at serine 727, but not at tyrosine 701 (Fig. 8B), which was followed by caspase-independent death of Mφ (data not shown). This is consistent with recent reports showing ROS-mediated cell death through the activation of p38 MAPK (11). Furthermore, Mφ from STAT1^−/− mice showed remarkable impairment of p38 MAPK activation by zVAD/LPS. This is most likely due to impaired generation of superoxide (Fig. 8C and 6G), further supporting the positive-feedback regulation between superoxide and p38 MAPK/STAT1.

• Open in new tab
• Download powerpoint

FIG. 8.

STAT1 activation is mediated by ROS-stimulated p38 MAPK in a positive-feedback mechanism. (A) Mφ were pretreated with zVAD (50 μM) for 1 h in the absence or presence of indicated concentrations of tempol or DPI and then treated with LPS (100 ng/ml) for 15 min. p38 MAPK activation was analyzed using specific anti-phospho-p38 MAPK antibody. (B) Mφ were treated with indicated concentrations of menadione for 15 min. p38 MAPK activation was analyzed as in panel A (top). STAT1 activation was analyzed using the same samples employing antibodies specific for phosphorylated forms of STAT1 at serine 727 or at tyrosine 701 (bottom). Mφ treated with IFN-β (100 U/ml) were used as a positive control to detect phosphorylated forms of STAT1 at tyrosine 701. (C) Mφ from STAT1^+/+ or STAT1^−/− mice were exposed to LPS for the indicated times, with or without zVAD pretreatment for 1 h. p38 MAPK activation was analyzed using specific anti-phospho-p38 MAPK antibody. Shown are the results of three independent experiments. (D) Mφ from STAT1^+/+ or STAT1^−/− mice were treated in duplicate with LPS for the indicated times, with or without zVAD pretreatment for 1 h. Reverse transcription-PCR analysis was performed in order to detect the expression of STAT1, gp91, and β-actin (top). Western blot analysis was done to detect the protein level of gp91 and tubulin (bottom). (E) Schematic representation of the gp91 gene (top). Mφ were exposed to LPS or IFN-γ (100 U/ml) for 1 h, with or without zVAD pretreatment for 1 h. Cells were cross-linked, and the fragmented chromatins were immunoprecipitated with antibody against STAT1. Immunoprecipitated chromatin DNAs and 1% of the cross-linked chromatin used in the immunoprecipitation (input) were amplified by PCR using primer sets specific for the promoter, as indicated by the arrow (bottom). An IFN-γ-treated sample was included as a positive control. (F) Mφ were transfected with the wild-type gp91 luciferase reporter or a mutant gp91 luciferase reporter containing a mutation in the STAT1 binding site. gp91 reporter constructs contain the region comprising bp −115 to +12 of the gp91 promoter. After transfection, cells were treated with LPS or IFN-γ for 2 h, with or without zVADpretreatment for 1 h. Relative firefly luciferase activities of gp91 reporter normalized to Renilla luciferase activities are shown. An IFN-γ-treated sample was included as a positive control. (G) Raw264.7 stable cell lines overexpressing a control vector or dominant-negative Rac1 (Rac1^N17) were treated with zVAD/LPS for 20 h. Cell viability was then assessed by crystal violet assay. (H) The same stable cell lines were untreated or exposed to zVAD/LPS for 30 min. STAT1 and p38 MAPK activation were analyzed using antibodies specific for phosphorylated forms of STAT1 at serine 727 and p38 MAPK, respectively. The expression of dominant-negative Rac1 (Rac1^N17) was analyzed using a specific anti-c-myc antibody. Shown are the results of two independent experiments. The results in panels F and G are the means ± standard errors.

Finally, we explored the potential mechanism of superoxide production by STAT1 activation. We studied the expression of gp91, a key component of the NADPH oxidase complex, which has been reported to have a STAT1 binding site on its promoter region (29). gp91 expression was upregulated by LPS and further upregulated by zVAD pretreatment; this upregulation was abrogated in STAT1^−/− Mφ (Fig. 8D). To further clarify that activated STAT1 binds to the gp91 promoter in vivo, we performed ChIP analysis. The binding of STAT1 to the gp91 promoter was slightly increased by LPS treatment and further enhanced by zVAD pretreatment (Fig. 8E). Similar to the results of gp91 expression (Fig. 8D), a luciferase activity assay further demonstrated that zVAD increased the luciferase activity of a gp91 reporter in LPS-treated Mφ, which was not observed using a gp91 reporter mutated on the STAT1 binding site (Fig. 8F). Since Rac1 GTPase is well known for its role as a regulator of NADPH oxidase activation (8), we investigated the effects of dominant-negative Rac1 (Rac1^N17) on the cell death commitment and associated STAT1 signaling by zVAD/LPS. Expression of Rac1^N17 but not a control vector significantly inhibited zVAD/LPS-mediated Mφ death (Fig. 8G), STAT1 activation at serine 727, and p38 MAPK activation (Fig. 8H). These results indicate that STAT1 may have an important role in the LPS-mediated induction of gp91, and NADPH oxidase activated through Rac1 may play an integral part in the superoxide production induced by zVAD/LPS. Taken together, our results suggest that the activation of STAT1 by phosphorylation at serine 727 is critically mediated by ROS-potentiated p38 MAPK. This, in turn, modulates intracellular oxidative stress through a positive-feedback regulation mechanism.