Portal de Periódicos da CAPES

Suppression of Inflammatory Cytokine Production by Carbon Monoxide Involves the JNK Pathway and AP-1

2003; Elsevier BV; Volume: 278; Issue: 39 Linguagem: Inglês

10.1074/jbc.m302942200

1083-351X

Danielle Morse, Søren Erik Pischke, Zhihong Zhou, Roger J. Davis, Richard A. Flavell, Torsten Loop, Sherrie L. Otterbein, Leo E. Otterbein, Augustine M.K. Choi,

Optimism, Hope, and Well-being

The stress-inducible protein heme oxygenase-1 provides protection against oxidative stress and modulates pro-inflammatory cytokines. As the sepsis syndrome results from the release of pro-inflammatory mediators, we postulated that heme oxygenase-1 and its enzymatic product CO would protect against lethality in a murine model of sepsis. Mice treated with a lethal dose of lipopolysaccharide (LPS) and subsequently exposed to inhaled CO had significantly better survival and lower serum interleukin (IL)-6 and IL-1β levels than their untreated counterparts. In vitro, mouse macrophages exposed to LPS and CO had significantly attenuated IL-6 production; this effect was concentration-dependent and occurred at a transcriptional level. The same effect was seen with increased endogenous CO production through overexpression of heme oxygenase-1. Mutation within the AP-1-binding site in the IL-6 promoter diminished the effect of CO on promoter activity, and treatment of macrophages with CO decreased AP-1 binding in an electrophoretic mobility shift assay. Electrophoretic mobility supershift assay indicated that the JunB, JunD, and c-Fos components of AP-1 were particularly affected. Upstream of AP-1, CO decreased JNK phosphorylation in murine macrophages and lung endothelial cells. Mice deficient in the JNK pathway had decreased serum levels of IL-6 and IL-1β in response to LPS compared with control mice, and no effect of CO on these cytokine levels was seen in Jnk1 or Jnk2 genedeleted mice. In summary, these results suggest that CO provides protection in a murine model of sepsis through modulation of inflammatory cytokine production. For the first time, the effect of CO is shown to be mediated via the JNK signaling pathway and the transcription factor AP-1. The stress-inducible protein heme oxygenase-1 provides protection against oxidative stress and modulates pro-inflammatory cytokines. As the sepsis syndrome results from the release of pro-inflammatory mediators, we postulated that heme oxygenase-1 and its enzymatic product CO would protect against lethality in a murine model of sepsis. Mice treated with a lethal dose of lipopolysaccharide (LPS) and subsequently exposed to inhaled CO had significantly better survival and lower serum interleukin (IL)-6 and IL-1β levels than their untreated counterparts. In vitro, mouse macrophages exposed to LPS and CO had significantly attenuated IL-6 production; this effect was concentration-dependent and occurred at a transcriptional level. The same effect was seen with increased endogenous CO production through overexpression of heme oxygenase-1. Mutation within the AP-1-binding site in the IL-6 promoter diminished the effect of CO on promoter activity, and treatment of macrophages with CO decreased AP-1 binding in an electrophoretic mobility shift assay. Electrophoretic mobility supershift assay indicated that the JunB, JunD, and c-Fos components of AP-1 were particularly affected. Upstream of AP-1, CO decreased JNK phosphorylation in murine macrophages and lung endothelial cells. Mice deficient in the JNK pathway had decreased serum levels of IL-6 and IL-1β in response to LPS compared with control mice, and no effect of CO on these cytokine levels was seen in Jnk1 or Jnk2 genedeleted mice. In summary, these results suggest that CO provides protection in a murine model of sepsis through modulation of inflammatory cytokine production. For the first time, the effect of CO is shown to be mediated via the JNK signaling pathway and the transcription factor AP-1. Heme oxygenase-1 (HO-1) 1The abbreviations used are: HO-1, heme oxygenase-1; LPS, lipopolysaccharide; IL, interleukin; AP-1, activator protein-1; C/EBP, CAAT/enhancer-binding protein; ELISA, enzyme-linked immunosorbent assay; EMSA, electrophoretic mobility shift assay; JNK, c-Jun NH2-terminal kinase; MAPK, mitogen-activated protein kinase. is a microsomal enzyme responsible for the degradation of heme, generating biliverdin, iron, and carbon monoxide (CO). Induction of HO-1 has been reported in response to a wide variety of oxidant stimuli, and this enzyme is a vital component of cellular adaptation to stress. There is growing interest in the role of CO in the anti-inflammatory and cytoprotective function of HO-1. It has recently been demonstrated that the administration of exogenous CO inhibits lipopolysaccharide (LPS)-induced production of tumor necrosis factor-α while increasing interleukin (IL)-10 production both in vitro and in vivo (1Otterbein L.E. Bach F.H. Alam J. Soares M. Tao Lu H. Wysk M. Davis R.J. Flavell R.A. Choi A.M. Nat. Med. 2000; 6: 422-428Crossref PubMed Scopus (1851) Google Scholar). This effect is independent of the guanylyl cyclase/cGMP pathway, and at least in the case of tumor necrosis factor-α, the effect is post-transcriptional. More recently, Fujita et al. (2Fujita T. Toda K. Karimova A. Yan S.F. Naka Y. Yet S.F. Pinsky D.J. Nat. Med. 2001; 7: 598-604Crossref PubMed Scopus (374) Google Scholar) demonstrated that inhaled CO can rescue HO-1-deficient mice from lethal ischemic lung injury, further strengthening the paradigm of CO as a direct mediator of cellular protection. The sepsis syndrome is the leading cause of death in intensive care units in the United States, and its incidence continues to rise. Despite an increasing incidence, little impact on associated mortality rates has been made over the past several decades (3Friedman G. Silva E. Vincent J.L. Crit. Care Med. 1998; 26: 2078-2086Crossref PubMed Scopus (500) Google Scholar). It has long been recognized that invading microorganisms induce the release of a large number of humoral and cellular pro-inflammatory mediators, causing a systemic inflammatory response syndrome (4Wenzel R.P. Pinsky M.R. Ulevitch R.J. Young L. Clin. Infect. Dis. 1996; 22: 407-412Crossref PubMed Scopus (123) Google Scholar). Elevated plasma concentrations of a variety of cytokines and chemokines such as tumor necrosis factor-α, IL-1, and IL-6 have been described in septic patients (5Arnalich F. Garcia-Palomero E. Lopez J. Jimenez M. Madero R. Renart J. Vazquez J.J. Montiel C. Infect. Immun. 2000; 68: 1942-1945Crossref PubMed Scopus (228) Google Scholar, 6Van der Poll T. van Deventer S.J. Infect. Dis. Clin. N. Am. 1999; 13: 413-426Abstract Full Text Full Text PDF PubMed Scopus (292) Google Scholar). IL-6 is one of the most reliably elevated markers in sepsis (7Selberg O. Hecker H. Martin M. Klos A. Bautsch W. Kohl J. Crit. Care Med. 2000; 28: 2793-2798Crossref PubMed Scopus (209) Google Scholar) and one of the few serum markers that has a demonstrable correlation with mortality rates in humans (8Spittler A. Razenberger M. Kupper H. Kaul M. Hackl W. Boltz-Nitulescu G. Fugger R. Roth E. Clin. Infect. Dis. 2000; 31: 1338-1342Crossref PubMed Scopus (94) Google Scholar). Knowing the anti-inflammatory effects of CO, we hypothesized that CO would protect against lethality in an animal model of sepsis. We further postulated that modulation of IL-6 production may contribute to the salutary effects of CO. We tested these hypotheses using a model of LPS-induced sepsis syndrome. Animals—Male C57BL/6 mice were purchased from Jackson Laboratory (Bar Harbor, ME); male Jnk1 –/– and Jnk2 –/– mice and the corresponding control mice were generated as previously described (9Kuan C.Y. Yang D.D. Samanta D.R. Davis R.J. Rakic P. Flavell R.A. Neuron. 1999; 22: 667-676Abstract Full Text Full Text PDF PubMed Scopus (775) Google Scholar). Mice were allowed to acclimate for 1 week with rodent chow and water ad libitum. Cell Culture Experiments—RAW 264.7 mouse peritoneal macrophages and MH-S mouse alveolar macrophages were purchased from American Type Culture Collection (Manassas, VA). Both cell types were grown in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and 100 μg/ml gentamycin in a humidified atmosphere of 5% CO2 in air with or without additional CO as described below. Cell culture experiments were conducted in triplicate. Stable Cell Transfection—RAW 264.7 cells were stably transfected with a pSFFV/HO-1 plasmid construct as previously described (1Otterbein L.E. Bach F.H. Alam J. Soares M. Tao Lu H. Wysk M. Davis R.J. Flavell R.A. Choi A.M. Nat. Med. 2000; 6: 422-428Crossref PubMed Scopus (1851) Google Scholar). IL-6 Promoter Constructs—The IL-6 promoter constructs were kindly provided by Drs. Alaina J. Ammit (Faculty of Pharmacy, University of Sydney, Sydney, Australia) and Oliver Eikelberg (Giessen University School of Medicine, Giessen, Germany). The 651-bp fragment of the human IL-6 promoter was subcloned (5′-KpnI, 3′-XhoI) into the pGL3-Basic luciferase reporter gene vector plasmid (Promega, Madison, WI) to give the parental pIL6-luc651 construct. Within pIL6-luc651, consensus transcription factor-binding sites were inactivated by site-directed mutagenesis, including AP-1, C/EBP-β, and NF-κB as previously described (10Ammit A.J. Lazaar A.L. Irani C. O'Neill G.M. Gordon N.D. Amrani Y. Penn R.B. Panettieri Jr., R.A. Am. J. Respir. Cell Mol. Biol. 2002; 26: 465-474Crossref PubMed Scopus (172) Google Scholar, 11Eickelberg O. Pansky A. Mussmann R. Bihl M. Tamm M. Hilbebrand P. Perruchoud A.P. Roth M. J. Biol. Chem. 1999; 274: 12933-12938Abstract Full Text Full Text PDF PubMed Scopus (169) Google Scholar). Transient Cell Transfection and Luciferase Assay—Transfection of RAW 264.7 cells was performed with LipofectAMINE 2000 (Invitrogen) according to the manufacturer's instructions. Briefly, cells were plated onto 12-well plates at a density of 2.5 × 105 cells/plate for 18 h and then transfected with 600 ng of IL-6 construct as well as 30 ng of pRL-TK control vector (Promega) to normalize transfection efficiencies. Following transfection, cells were cultured for 24 h and then treated with 1 μg/ml LPS. Cells were harvested after 8 h, and luciferase activity was assessed using the Dual-Luciferase reporter assay system (Promega) according to manufacturer's instructions. Carbon Monoxide Exposures—Mice or macrophages were exposed to compressed air with varying concentrations of CO (0–500 ppm). For cell culture experiments, the atmosphere included 5% CO2 for adjustment of pH. Details of CO administration have been previously described (1Otterbein L.E. Bach F.H. Alam J. Soares M. Tao Lu H. Wysk M. Davis R.J. Flavell R.A. Choi A.M. Nat. Med. 2000; 6: 422-428Crossref PubMed Scopus (1851) Google Scholar). Animal Experiments—Mice were exposed to 250 ppm CO or room air for 1 h prior to intraperitoneal administration of LPS, Escherichia coli serotype 0127:B8 (Sigma). In lethality experiments, a dose of 40 mg/kg LPS was administered, and animals were returned to the exposure chamber until 48 h had elapsed or until death had ensued; the time of death was recorded. In cytokine experiments, 1 mg/kg LPS was administered by intraperitoneal injection, and blood was collected after 5 h; the animals were killed at this time. Cytokine Analysis—Serum and media samples were analyzed with enzyme-linked immunosorbent assay (ELISA) kits purchased from R&D Systems (Minneapolis, MN) following the manufacturer's instructions. RNA Extraction and Northern Blot Analysis—Total RNA was isolated from RAW 264.7 macrophages using TRIzol lysis buffer (Invitrogen), followed by chloroform extraction. The details of this technique have been previously described (1Otterbein L.E. Bach F.H. Alam J. Soares M. Tao Lu H. Wysk M. Davis R.J. Flavell R.A. Choi A.M. Nat. Med. 2000; 6: 422-428Crossref PubMed Scopus (1851) Google Scholar). Electrophoretic Mobility Shift Assay (EMSA)—RAW 264.7 cells were exposed to LPS (1 mg/ml) alone or ambient CO (250 ppm) as described above. Nuclei were extracted as previously described (12Camhi S.L. Alam J. Wiegand G.W. Chin B.Y. Choi A.M. Am. J. Respir. Cell Mol. Biol. 1998; 18: 226-234Crossref PubMed Scopus (126) Google Scholar). A double-stranded oligonucleotide containing the consensus transcription factor-binding site for AP-1 (5′-CGCTTGATGAGTCAGCCGGAA-3′) was purchased from Promega. Labeling and DNA binding reactions were performed as previously described (12Camhi S.L. Alam J. Wiegand G.W. Chin B.Y. Choi A.M. Am. J. Respir. Cell Mol. Biol. 1998; 18: 226-234Crossref PubMed Scopus (126) Google Scholar). For the supershift assay, rabbit polyclonal antibodies against c-Jun (sc-44 and sc-1694), JunB, JunD, and c-Fos (Santa Cruz Biotechnology, Santa Cruz, CA) were incubated with samples after the initial binding reaction between nuclear protein extracts and the consensus oligonucleotide. Western Analysis of JNK Phosphorylation—Assay kits were purchased from New England Biolabs (Beverly, MA) and used following the manufacturer's instructions. Statistical Analysis—Data are expressed as means ± S.E. Differences in measured variables between the experimental and control groups were assessed with Student's t test, except for the analysis of dose response, which was done using analysis of variance and Bonferroni's correction. Statistical difference was accepted at p < 0.05. CO Protects against Lethality in a Mouse Model of the Sepsis Syndrome—We used a murine LPS model to assess the effect of inhaled low dose CO on survival. C57BL/6 mice were treated with a lethal dose of intraperitoneal LPS and subsequently exposed to 250 ppm CO or ambient air. As shown in Fig. 1, the CO-treated animals had significantly better survival than their untreated counterparts. A low dose of CO is thus able to rescue animals from lethality due to LPS. CO Modulates Pro-inflammatory Cytokine Levels in Vivo—As inflammatory cytokines are known to play a role in the pathogenesis of the sepsis syndrome, we chose to study the effects of CO on cytokine production in our murine model. It has previously been demonstrated by our laboratory (1Otterbein L.E. Bach F.H. Alam J. Soares M. Tao Lu H. Wysk M. Davis R.J. Flavell R.A. Choi A.M. Nat. Med. 2000; 6: 422-428Crossref PubMed Scopus (1851) Google Scholar) that CO attenuates LPS-induced production of tumor necrosis factor-α and augments the production of IL-10 in vivo. To assess whether the effect of CO extends to IL-6 and IL-1β, we administered a sublethal dose of LPS (1 mg/kg) to mice in the presence or absence of inhaled CO (250 ppm). Mice treated with CO had lower serum levels of IL-6 and IL-1β 5 h after LPS injection than control mice as measured by ELISA (Fig. 2). HO-1 and CO Decrease IL-6 Expression in Vitro—To confirm the effect of CO on IL-6 in vitro, we exposed two cell lines of mouse macrophages (RAW 264.7 and MH-S) to 1 μg/ml LPS in the presence or absence of CO and assessed IL-6 production by ELISA. No IL-6 was detectable in the cell medium from untreated macrophages or from macrophages treated with CO alone. Cells exposed to LPS had an elevated level of IL-6 in the medium after 5 h, but this increase in IL-6 was significantly attenuated by exposure to CO (Fig. 3, A and B). The effect of CO was dose-dependent over a range of 50–500 ppm (Fig. 3C). To determine whether increased intracellular CO production would have the same effect as exogenously administered CO, we exposed a macrophage cell line (RAW 264.7) overexpressing HO-1 to 1 μg/ml LPS and assayed for IL-6 by ELISA. The control cells were transfected with the neomycin gene. The control cells exhibited a much greater increase in IL-6 levels than did the HO-1-overexpressing cells (Fig. 3D), suggesting that augmented intracellular CO production has an effect similar to that of exogenous administration of CO. CO Suppresses IL-6 mRNA Expression in RAW 264.7 Macrophages—We assessed the expression of IL-6 mRNA in RAW 264.7 macrophages exposed to 1 μg/ml LPS in the presence or absence of CO. Cells were harvested 5 h after LPS treatment for RNA extraction and analysis by Northern blotting. As shown in Fig. 4, the untreated cells and those treated with CO alone did not have detectable IL-6 mRNA by Northern blotting. LPS caused a large increase in IL-6 mRNA production, and this was abrogated by treatment with CO. This demonstrates that CO affects IL-6 production in RAW 264.7 cells at a transcriptional level. CO Exerts Its Greatest Effect via AP-1—Regulation of IL-6 gene expression is controlled by the binding of transcription factors to known consensus sequences within the promoter region, including binding sites for AP-1, NF-κB, or C/EBP-β (10Ammit A.J. Lazaar A.L. Irani C. O'Neill G.M. Gordon N.D. Amrani Y. Penn R.B. Panettieri Jr., R.A. Am. J. Respir. Cell Mol. Biol. 2002; 26: 465-474Crossref PubMed Scopus (172) Google Scholar). To identify which cis-regulatory sequences might be responsible for down-regulation of IL-6 by CO, RAW 264.7 macrophages were transiently transfected with wild-type or mutant constructs of the human IL-6 promoter bound to a luciferase reporter. Four promoter constructs were used, three with site-directed mutations in the AP-1-, NF-κB-, or C/EBP-β-binding sequence and a parental construct. Luciferase activity was induced by LPS treatment of cells transfected with all mutant constructs to at least the same level as the parental construct. CO treatment decreased luciferase activity to a significant degree in all constructs except the AP-1 mutant construct (Fig. 5). CO Attenuates AP-1 Binding in Vitro—To confirm that CO can inhibit the activation of AP-1, EMSAs were performed using nuclear extracts from RAW 264.7 macrophages treated with LPS alone or with LPS and CO (250 ppm). Cells were harvested 30 min after LPS treatment. As shown in Fig. 6, this concentration of CO was capable of attenuating LPS-stimulated AP-1 binding in macrophages. To assess the specificity of the AP-1 complex and to identify the contributing family members, supershift experiments were carried out by adding antibodies against c-Jun, JunB, JunD, and c-Fos to the nuclear extracts before EMSA. Protein-antibody recognition can be visualized by a decrease in the mobility of the DNA-protein complex and a diminution of the AP-1 complex. Fig. 7 indicates that the AP-1 complex consists mainly of c-Fos and, to a lesser extent, JunB and JunD. Correspondingly, treatment of cells with CO appears to affect c-Fos to the greatest extent and JunB and JunD to a lesser extent. CO Inhibits JNK Phosphorylation after Stimulation with LPS—The involvement of AP-1 in the inhibition of IL-6 production prompted us to ask whether the JNK pathway might also be involved. To test this hypothesis, RAW 264.7 cells were stimulated with LPS (1 μg/ml) and exposed to either CO (250 ppm) or ambient air as described above. Cells were collected at times from 0 min to 3 h for Western blot analysis of JNK phosphorylation. JNK phosphorylation was increased by stimulation with LPS; this effect was maximum at 15 min and declined thereafter. JNK phosphorylation was inhibited by CO treatment (Fig. 8) at early time points (5 and 15 min), but this effect was lost by 30 min. The 1- and 3-h time points are not shown, but CO had no effect on JNK phosphorylation at these times. Deletion of the Jnk1 or Jnk2 Gene Inhibits IL-1 and IL-6 Production in Vivo—To investigate whether the JNK pathway may mediate the effect of CO on LPS-induced IL-6 and IL-1 production, we used mice deficient in either the Jnk1 or Jnk2 gene. Each gene is capable of making various isoforms of JNK, which are produced through alternate splicing, but the effect of single gene deletion is a decrease in total JNK activity (13Gupta S. Barrett T. Whitmarsh A.J. Cavanagh J. Sluss H.K. Derijard B. Davis R.J. EMBO J. 1996; 15: 2760-2770Crossref PubMed Scopus (1183) Google Scholar). The effect of Jnk1 deletion on overall JNK function has been shown to be greater than that of Jnk2 deletion (14Tournier C. Hess P. Yang D.D. Xu J. Turner T.K. Nimnual A. Bar-Sagi D. Jones S.N. Flavell R.A. Davis R.J. Science. 2000; 288: 870-874Crossref PubMed Scopus (1550) Google Scholar). The control, Jnk1 –/–, and Jnk2 –/– mice were injected with a non-lethal dose of LPS, and cytokine levels were measured at time 0, 1 h, and 5 h in the presence or absence of CO. IL-1β levels were significantly higher in the control mice than in either the Jnk1 –/– or Jnk2 –/– mice at 1 h, but by 5 h, this difference no longer existed. At the 5-h time point, however, the CO-treated wild-type mice had significantly lower levels of IL-1β than the control mice (as previously described), but CO had no effect on IL-β levels in the Jnk1 –/– and Jnk2 –/– mice (Fig. 9A). Similarly, IL-6 (which is slower to increase after LPS injection compared with IL-1β) was significantly less in the Jnk1 –/– and Jnk2 –/– mice than in the control mice at 5 h, and there was no effect of CO on the levels of IL-6 in the Jnk1 –/– and Jnk2 –/– mice (Fig. 9B). This indicates not only that the JNK pathway is involved in the production of IL-1β and IL-6 in response to LPS, but that the attenuation of this pathway diminishes or abrogates the effect of CO. Deletion of the Jnk1 Gene Abrogates the Survival Effect of CO in Response to a Lethal Dose of LPS—To determine whether deletion of the Jnk1 gene affects survival in response to LPS, these mice were injected with a lethal dose of LPS and subsequently exposed to room air or CO (250 ppm). At 24 h, the average length of survival was 28–29 h in both groups (Fig. 10), indicating that CO had no effect on survival. Interestingly, the Jnk1 –/– mice had a longer survival than the wild-type mice shown in Fig. 1. The sepsis syndrome remains the leading cause of death in intensive care units in the United States, and efforts to impact mortality have been largely unsuccessful (3Friedman G. Silva E. Vincent J.L. Crit. Care Med. 1998; 26: 2078-2086Crossref PubMed Scopus (500) Google Scholar). Evidence from human and animal studies demonstrates that HO-1 is induced by sepsis, and this induction confers a survival advantage. Both increased carboxyhemoglobin concentration and increased breath CO excretion have been observed in patients with sepsis (15Moncure M. Brathwaite C.E.M. Samaha E. Marburger R. Ross S.E. J. Trauma. 1999; 46: 424-427Crossref PubMed Scopus (41) Google Scholar, 16Zegdi R. Perrin D. Burdin M. Boiteau R. Tenaillon A. Int. Care Med. 2002; 28: 793-796Crossref PubMed Scopus (97) Google Scholar). Bacterial LPS administration to rats induces HO-1 expression (17Otani K. Shimizu S. Chijiiwa K. Morisaki T. Yamaguchi T. Yamaguchi K. Kuroki S. Tanaka M. Dig. Dis. Sci. 2000; 45: 2313-2319Crossref PubMed Scopus (32) Google Scholar), and in a rodent model of hemorrhagic shock, induction of HO-1 has been shown to be protective and inhibition of HO-1 deleterious (18Tamion F. Richard V. Bonmarchand G. Leroy J. Lebreton J.P. Thuillez C. Am. J. Respir. Crit. Care Med. 2001; 164: 1933-1938Crossref PubMed Scopus (72) Google Scholar). Similarly, in a murine cecal ligation model, inhibition of HO-1 leads to increased mortality (19Downard P.J. Wilson M.A. Spain D.A. Matheson P.J. Siow Y. Garrison R.N. J. Surg. Res. 1997; 71: 7-12Abstract Full Text PDF PubMed Scopus (32) Google Scholar). The mechanism by which HO-1 provides cellular protection remains a subject of avid investigation; all of the products of heme catabolism have been implicated, including bilirubin, ferritin (stimulated by the release of iron), and CO (20Oberle S. Schwartz P. Abate A. Schroder H. Biochem. Biophys. Res. Commun. 1999; 261: 28-34Crossref PubMed Scopus (62) Google Scholar, 21Yesilkaya A. Altinayak R. Korgun D.K. Gen. Pharmacol. 2000; 35: 17-20Crossref PubMed Scopus (21) Google Scholar, 22Otterbein L.E. Choi A.M.K. Am. J. Physiol. 2000; 279: L1029-L1037Crossref PubMed Google Scholar). Here, we have examined the effect of exogenously administered CO in a murine LPS model of septic shock and have shown that CO reduces mortality. There is a well described correlation between serum IL-6 levels and poor outcome in sepsis (8Spittler A. Razenberger M. Kupper H. Kaul M. Hackl W. Boltz-Nitulescu G. Fugger R. Roth E. Clin. Infect. Dis. 2000; 31: 1338-1342Crossref PubMed Scopus (94) Google Scholar, 23Santana C. Guindeo M.C. Gonzalez G. Garcia-Munoz F. Saavedra P. Domenech E. Acta Paediatr. 2001; 90: 1176-1181Crossref PubMed Google Scholar, 24Geppert A. Steiner A. Zorn G. Delle-Karth G. Koreny M. Haumer M. Siostrzonek P. Huber K. Heinz G. Crit. Care Med. 2002; 30: 1987-1994Crossref PubMed Scopus (103) Google Scholar, 25Heesen M. Obertacke U. Schade F.U. Bloemeke B. Majetschak M. Eur. Cytokine Netw. 2002; 13: 72-77PubMed Google Scholar, 26Gallagher J. Fisher C. Sherman B. Munger M. Meyers B. Ellison T. Fischkoff S. Barchuk W.T. Teoh L. Velagapudi R. Int. Care Med. 2001; 27: 1169-1178Crossref PubMed Scopus (45) Google Scholar), and increased IL-6 is an independent predictor of survival in human sepsis (27Rodriguez-Gaspar M. Santolaria F. Jarque-Lopez A. Gonzalez-Reimers E. Milena A. de la Vega M.J. Rodriguez-Rodriguez E. Gomez-Sirvent J.L. Cytokine. 2001; 15: 232-236Crossref PubMed Scopus (82) Google Scholar). In our mouse model, serum IL-6 and IL-1β were markedly attenuated by exogenous CO administration. The low cytokine levels could result from any indirect action of CO that improves the pathophysiology of sepsis, but our in vitro data suggest that CO has a direct effect on cytokine production. The attenuation of IL-6 production by exogenous CO in vitro occurred even at very low concentrations of CO (50 ppm). For comparison, this concentration of CO is close to the levels that can routinely be measured in the exhaled breath of healthy individuals who smoke (30 ppm) (28Haustein K.O. Krause J. Haustein H. Rasmussen T. Cort N. J. Int. Med. 2002; 252: 130-139Crossref PubMed Scopus (48) Google Scholar). Using Northern blotting, we have shown that the in vitro attenuation of LPS-induced IL-6 production occurs at a transcriptional level. Our study using IL-6 promoter constructs with mutations at each of three transcription factor-binding sites (NF-κB, AP-1, and C/EBP) suggests that the primary action of CO may be through modulation of AP-1. This interpretation must be made cautiously, however, as mutation of any single binding site had no effect on the overall level of promoter stimulation by LPS, implying that no single transcription factor is absolutely required for IL-6 induction in response to LPS. Also, there was a trend toward decreased IL-6 promoter activity after treatment with CO even in the AP-1 mutant construct, suggesting that inhibition of AP-1 binding may not wholly account for the effect of CO. Inhibition of NF-κB binding by CO has previously been reported (29Sarady J.K. Otterbein S.L. Liu F. Otterbein L.E. Choi A.M. Am. J. Respir. Cell Mol. Biol. 2002; 27: 739-745Crossref PubMed Scopus (125) Google Scholar), but our IL-6 promoter data suggest that this is not a major factor in our particular model. We have shown that the particular AP-1 components affected by CO include JunB, JunD, and c-Fos. Other investigators have shown that IL-10 inhibits IL-6 transcription in human monocytes via inhibition of AP-1-binding activity and reduced expression of c-Fos (30Dokter W.H. Koopmans S.B. Vellenga F. Leukemia (Baltimore). 1996; 10: 1308-1316PubMed Google Scholar). The similarity of the results of that study to our findings is intriguing given the recent report that HO-1 mediates the anti-inflammatory effects of IL-10 in mice (31Lee T.S. Chau L.Y. Nat. Med. 2002; 8: 240-246Crossref PubMed Scopus (914) Google Scholar). The JNK pathway is activated by cellular stress and is intimately linked with the regulation of AP-1 transcription activity. Ventura et al. (32Ventura J.-J. Kennedy N.J. Lamb J.A. Flavell R.A. Davis R.J. Mol. Cell. Biol. 2003; 23: 2871-2882Crossref PubMed Scopus (138) Google Scholar) recently reported that the JNK pathway is required for the regulation of AP-1 activity by tumor necrosis factor. Activated JNK is capable of binding the NH2-terminal activation domain of c-Jun, activating AP-1 by phosphorylating its component c-Jun. AP-1 can then translocate into the nucleus to promote transcription of downstream genes (33Davis R.J. Cell. 2000; 103: 239-252Abstract Full Text Full Text PDF PubMed Scopus (3666) Google Scholar). The association of JNK with AP-1, along with the previous implication of the JNK pathway in the down-regulation of macrophage cytokines by heat shock (34Wang Y. Li C. Wang X. Zhang J. Chang Z. Biochem. Biophys. Res. Commun. 2002; 269: 742-748Crossref Scopus (38) Google Scholar), made the JNK signaling cascade an attractive candidate as a target for CO. We have shown that, in a mouse macrophage cell line, CO can inhibit LPS-stimulated JNK activation at early time points. JNK activation began to decline between 15 and 30 min after LPS stimulation, implying that, in our in vitro system, cellular events mediated by JNK result from an early and non-sustained activation of this pathway. Our studies using Jnk gene-deleted mice suggest that the effect of CO on the JNK pathway is functionally significant, even if it is not sustained over time in vitro. The mice we used are deficient in either the Jnk1 or Jnk2 gene; deletion of both genes is embryonically lethal (14Tournier C. Hess P. Yang D.D. Xu J. Turner T.K. Nimnual A. Bar-Sagi D. Jones S.N. Flavell R.A. Davis R.J. Science. 2000; 288: 870-874Crossref PubMed Scopus (1550) Google Scholar). Although JNK activity is not completely abolished in these mice, it is considerably diminished, and the Jnk1 deletion has been shown to have a more profound effect on total JNK activity than the Jnk2 deletion (14Tournier C. Hess P. Yang D.D. Xu J. Turner T.K. Nimnual A. Bar-Sagi D. Jones S.N. Flavell R.A. Davis R.J. Science. 2000; 288: 870-874Crossref PubMed Scopus (1550) Google Scholar). Both Jnk1 –/– and Jnk2 –/– mice exhibited low levels of IL-1β in response to LPS stimulation at 1 h, but the levels rose to those of the control mice after 5 h. The rise in IL-1β to more normal levels may be due to the fact that JNK activity is not completely absent, and other stress-induced signaling pathways are intact in these mice. The IL-6 levels were not elevated in either the control or gene-deleted mice at 1 h, but at 5 h, the wild-type mice had higher levels of IL-6 than either the Jnk1 –/– or Jnk2 –/– mice. More interestingly, the Jnk1 –/– and Jnk2 –/– mice did not respond to exogenous CO in the same way as the wild-type mice; the levels of IL-1β and IL-6 were not affected by CO in these mice. The effect of CO in prolonging survival was not seen in the Jnk1 –/– mice either, and their length of survival at base line was increased compared with C57BL/6 mice. The results of these experiments with Jnk-deficient mice suggest that CO involves the JNK pathway in the suppression of inflammatory cytokine production. There have been no previous studies linking CO to the JNK pathway. Prior publications have demonstrated an in vitro effect of CO on the p38 MAPK pathway (1Otterbein L.E. Bach F.H. Alam J. Soares M. Tao Lu H. Wysk M. Davis R.J. Flavell R.A. Choi A.M. Nat. Med. 2000; 6: 422-428Crossref PubMed Scopus (1851) Google Scholar, 35Brouard S. Otterbein L.E. Anrather J. Tobiasch E. Bach F.H. Choi A.M. Soares M.P. J. Exp. Med. 2000; 192: 1015-1026Crossref PubMed Scopus (880) Google Scholar). (Importantly, these studies did not examine time points as early as 15 or 30 min after cell stimulation.) It is likely that CO has effects on multiple signaling pathways in various models, and as noted above, our results do not eliminate the possible involvement of alternative transcription factors and signaling pathways. Our data implicating the JNK pathway do, however, represent a novel insight into the mechanism of CO effects in biological systems. In summary, we have demonstrated for the first time a protective role for low concentration inspired CO in a murine model of sepsis and have shown that this is associated with decreased production of IL-1β and IL-6 in vivo and in vitro. We have also presented evidence that the effect of CO involves the JNK/AP-1 signaling pathway. We speculate that decreased LPS-mediated activation of JNK/AP-1 in the presence of CO leads to a dampening effect on inflammatory cytokine production and protection against sepsis. We thank Drs. Oliver Eikelberg and Alaina J. Ammit for generously donating the IL-6 promoter constructs.