Acute inflammatory lung injury occurs frequently in the setting of severe infection or blood loss. Accumulation of activated neutrophils in the lungs and increased pulmonary proinflammatory cytokine levels are major characteristics of acute lung injury. In the present experiments, we examined mechanisms leading to neutrophil accumulation and activation in the lungs after endotoxemia or hemorrhage. Levels of IL-1β, TNF-α, and macrophage inflammatory protein-2 mRNA were increased in lung neutrophils from endotoxemic or hemorrhaged mice compared with those present in lung neutrophils from control mice or in peripheral blood neutrophils from endotoxemic, hemorrhaged, or control mice. The transcriptional regulatory factors NF-κB and cAMP response element binding protein were activated in lung but not blood neutrophils after hemorrhage or endotoxemia. Xanthine oxidase inhibition, achieved by feeding allopurinol or tungsten-containing diets, did not affect neutrophil trafficking to the lungs after hemorrhage or endotoxemia. Xanthine oxidase inhibition did prevent hemorrhage- but not endotoxemia- induced increases in proinflammatory cytokine expression among lung neutrophils. Hemorrhage- or endotoxemia-associated activation of NF-κB in lung neutrophils was not affected by inhibition of xanthine oxidase. cAMP response element binding protein activation was increased after hemorrhage, but not endotoxemia, in mice fed xanthine oxidase-inhibiting diets. Our results indicate that xanthine oxidase modulates cAMP response element binding protein activation and proinflammatory cytokine expression in lung neutrophils after hemorrhage, but not endotoxemia. These findings suggest that the mechanisms leading to acute inflammatory lung injury after hemorrhage differ from those associated with endotoxemia.
Acute inflammatory lung injury, known clinically as the acute respiratory distress syndrome (ARDS),3 occurs frequently in patients with severe infections or blood loss (1, 2). Clinical studies investigating novel pharmacologic therapies for ARDS have included patients with multiple predisposing conditions, including hemorrhage and infection, assuming that there were common mechanisms leading to lung injury (3, 4, 5, 6). None of the antiinflammatory or antioxidant therapies have shown benefit (4, 5, 6). While it is possible that the therapeutic agents examined were not effective, an alternative explanation is that the mechanisms leading to acute lung injury in the patients studied were so heterogeneous that even though an agent may have had positive therapeutic effect in a limited group, such benefit was diluted out and not detectable in the overall study population.
Massive accumulation of neutrophils in the lungs and increased pulmonary proinflammatory cytokine levels are major characteristics of acute lung injury (7, 8). Proinflammatory cytokines, including IL-1β, TNF-α, and macrophage inflammatory peptide (MIP)-2, can be produced by resident pulmonary cell populations, including alveolar macrophages and vascular endothelium (9, 10). However, neutrophils that accumulate in the lungs after endotoxemia or hemorrhage also appear to be a significant intrapulmonary source of IL-1β and other immunoregulatory cytokines (11). Identification of lung neutrophils as a significant intrapulmonary source of IL-1β after hemorrhage or endotoxemia may be particularly important because several studies have shown that IL-1β is a major proinflammatory cytokine in bronchoalveolar lavages obtained from patients with acute lung injury (12, 13).
Binding elements for the transcriptional regulatory factors NF-κB and cAMP response element binding protein (CREB) are present in the enhancer/promoter regions of immunoregulatory cytokine genes, including IL-1β and TNF-α, and have important functions in modulating transcription of these genes (14, 15, 16). Increased activation of both NF-κB and CREB occurs in the lungs of animals after endotoxemia or blood loss (17, 18, 19). NF-κB is activated in alveolar macrophages from patients with ARDS (20). Inhibition of NF-κB activation in the lungs after either hemorrhage or endotoxemia is associated with decreased expression of proinflammatory cytokines and neutrophilic alveolitis (18, 19, 21, 22). Although in vitro culture of neutrophils with endotoxin results in activation of the trancriptional regulatory factor NF-κB (23), in vivo activation of NF-κB in neutrophils has not been reported.
In experimental models of acute lung injury, reactive oxygen intermediates (ROI) appear to contribute to the increased production of proinflammatory cytokines in the lungs, at least in part through participating in the activation of NF-κB and CREB (17, 19, 21, 22, 24, 25, 26). Pretreatment with antioxidants inhibits the elevations in pulmonary proinflammatory cytokine expression that normally follow hemorrhage or endotoxemia (19, 24, 25, 27). Similarly, antioxidant treatment or inhibition of the generation of xanthine oxidase-derived ROI prevents hemorrhage- or endotoxemia-induced activation of NF-κB in the lungs (19, 21, 22). ROI also are involved in hemorrhage-associated activation of CREB because inhibition of xanthine oxidase with tungsten feeding prevents the increased levels of transcriptionally active serine 133-phosphorylated CREB normally present in lung cell populations after blood loss (17).
In the present experiments, we examined the mechanisms by which lung neutrophils become activated in vivo to produce proinflammatory cytokines after hemorrhage or endotoxemia. Our studies demonstrate that xanthine oxidase, presumably through a CREB-associated mechanism, has an important role in increasing proinflammatory cytokines in lung neutrophils after hemorrhage, but not after endotoxemia. These findings indicate that the mechanisms leading to acute inflammatory lung injury after hemorrhage differ from those associated with endotoxemia.
Materials and Methods
Male BALB/c mice, 8–12 wk of age, were purchased from Harlan Sprague-Dawley (Indianapolis, IN). The mice were kept on a 12-h light/dark cycle with free access to food and water. All experiments were conducted in accordance with institutional review board-approved protocols.
Methoxyflurane was obtained from Pittmann-Moore (Mundelein, IL). Escherichia coli 0111:B4 endotoxin, collagenase, and DNase were obtained from Sigma (St. Louis, MO). The allopurinol-supplemented diet and tungsten-enriched molybdenum-deficient diet were purchased from ICN Biochemicals (Costa Mesa, CA). RPMI 1640 medium/25 mM HEPES/l-glutamine was obtained from BioWhittaker (Walkersville, MD). Percoll was purchased from Pharmacia (Uppsala, Sweden). Anti-B220 and anti-Thy 1.2 magnetic beads were obtained from Dynal (Lake Success, NY). Primers for IL-1β, TNF-α, or G3PDH were obtained from Clontech (Palo Alto, CA). MIP-2 primers were synthesized by Operon Technologies (Alameda, CA) using sequences kindly provided by Dr. David Baltimore (California Institute of Technology, Pasadena, CA). Amplitaq polymerase was purchased from Perkin-Elmer (Norwalk, CT). The Coomassie-Plus protein assay reagent was purchased from Pierce (Rockford, IL). Sequenase DNA polymerase was obtained from Amersham (Arlington Heights, IL). Anti-phosphorylated CREB antiserum was purchased from Upstate Biotechnology (Lake Placid, NY). These anti-phosphoCREB Abs recognize Ser133 phosphorylated CREB, but do not react with CREB that is not phosphorylated on Ser133 (28). Anti-p50, anti-p65, and anti-cRel antiserum were obtained from Santa Cruz Biotechnology (Santa Cruz, CA).
Models of hemorrhage and endotoxemia
The murine hemorrhage model used in these experiments was developed in our laboratory and reported previously (11, 17, 18). With this model, 30% of the calculated blood volume (∼0.55 ml for a 20-g mouse) is withdrawn over a 60-s period by cardiac puncture from a methoxyflurane-anesthetized mouse. The period of methoxyflurane anesthesia is <2 min in all cases. The mortality rate with this hemorrhage protocol is ∼12%.
To assess the effects of xanthine oxidase on cytokine expression and transcriptional factor activation, mice were pair-fed an allopurinol-supplemented diet (2.5 g/kg chow) or a normal control diet for 1 wk before hemorrhage or endotoxemia (29).
To assess further the effects of xanthine oxidase on cytokine expression and transcriptional factor activation, mice were pair-fed a tungsten-enriched, molybdenum-deficient diet (0.7 g sodium tungstate/kg chow) or a normal control diet for 3 wk before hemorrhage or endotoxemia (30).
Myeloperoxidase (MPO) assay
MPO activity was assayed by a modification of the method of Anderson and coworkers (11, 31). Excised lungs from three to four mice per treatment group were frozen in liquid nitrogen, weighed, and stored at −86°C. Lungs were homogenized for 30 s in 1.5 ml 20 mM potassium phosphate, pH 7.4, and centrifuged at 4°C for 30 min at 40,000 × g. The pellet was resuspended in 1.5 ml 50 mM potassium phosphate, pH 6.0, containing 0.5% hexadecyltrimethyl ammonium bromide, sonicated for 90 s, incubated at 60°C for 2 h, and centrifuged. The supernatant was assayed for peroxidase activity corrected to lung weight.
Isolation of neutrophils
Neutrophils from intraparenchymal pulmonary and peripheral blood cell suspensions were isolated by a modification of the technique of Sugarawa and coworkers as previously used in our laboratory (11, 32). In brief, the chest of the mouse was opened and the lung vascular bed was flushed with 3–5 ml of chilled (4°C) PBS injected into the right ventricle. Lungs were then excised, avoiding the paratracheal lymph nodes, and washed twice in RPMI 1640 medium/25 mM HEPES/l-glutamine with penicillin/streptomycin. Intraparenchymal pulmonary cell suspensions were isolated by collagenase digestion, using techniques previously described by our laboratory (11, 17, 18). In brief, the excised lungs were minced finely, and the tissue pieces were placed in RPMI 1640 medium containing 5% FCS, 20 U/ml collagenase, and 1 μg/ml DNase. Following incubation for 60 min at 37°C, any remaining intact tissue was disrupted by passage through a 21-gauge needle. Tissue fragments and the majority of dead cells were removed by rapid filtration through a glass wool column, and cells were collected by centrifugation.
Peripheral blood cell suspensions were isolated by techniques previously described by our laboratory. In brief, mice were anesthetized with methoxyflurane and then exsanguinated. Blood was withdrawn by cardiac puncture, collected into 5 U heparin, and centrifuged. After the plasma was removed, the red cells were lysed in Gey’s solution. The remaining cells were washed in RPMI and collected by centrifugation.
The pellets from either the intraparenchymal pulmonary cell suspension or peripheral blood cell suspension were resuspended in 2 ml PBS. If cells were to be used in the EMSA, pellets from lung suspensions from three mice or blood suspensions from two mice were pooled in PBS. The lung cell suspension was layered onto 5 ml Percoll of density 1.080 g/ml, which previously had been layered upon 5 ml Percoll of density 1.088 g/ml. The blood cell suspension was layered onto 5 ml Percoll of density 1.085 g/ml, which previously had been layered upon 5 ml Percoll of density 1.097 g/ml. After centrifugation at 600 × g for 25 min at 18°C, the neutrophil-rich fraction was collected from the interface between the two Percoll layers (i.e., at the 1.080 g/ml/1.088 g/ml Percoll density interface for lung neutrophils and at the 1.085/1.097 g/ml interface for blood neutrophils), and washed with RPMI. For EMSA, lung or blood neutrophils were pooled from four Percoll gradients. Neutrophils were further purified by the removal of T and B cells with anti-B220 and anti-Thy 1.2 magnetic beads by using the manufacturer’s protocol. Viability, as determined by trypan blue exclusion, was consistently >98%. Neutrophil purity, as determined by Wright’s staining cytospin preparations, was >98%.
The basic procedure for determining cytokine expression by semiquantitative PCR has been described previously by our laboratory (18, 27). Groups of 4–14 mice, with PCR results obtained from individual mice, were used for each experimental condition. In brief, after purified neutrophil populations had been lysed in 4 M guanidium thiocyanate/25 mM sodium citrate/0.5% sarcosyl/0.1 M 2-ME, mRNA was phenol extracted by the method of Chomczynski and Sacchi (33). cDNA was synthesized from the mRNA of 100,000 cells from each sample using Moloney murine leukemia virus reverse transcriptase and random hexamer oligonucleotides, according to the procedure of Kawasaki (34). Semiquantitative PCR was conducted with the cDNA from 10,000 cells/reaction. A single PCR master mix was prepared and aliquots used for reactions in all treatment groups for each experiment. Primers at 0.4 μM for IL-1β, TNF-α, or MIP-2 were mixed with primers for G3PDH, a housekeeping gene used as an internal control for standardization of PCR product. After Amplitaq polymerase was added to the cDNA/primer mixture at 85°C (to prevent primer-dimer formation), between 27 and 40 cycles of PCR were conducted (1 min, 95°C denaturation; 1 min, 60°C annealing; 1 min, 72°C extension). To detect amplified cDNA, the PCR product was analyzed by agarose gel electrophoresis. The number of PCR cycles were selected for each cytokine product so that the ethidium bromide-stained amplified DNA products were between barely detectable and below saturation. The gel was scanned with a gel documentation system (ImageStore 5000 with GelBase Windows software; Ultraviolet Products, San Gabriel, CA). Cytokine densitometric results were normalized to the G3PDH products.
Preparation of nuclear extracts
Nuclear extracts were prepared as previously described (17, 18). In brief, 3–9 × 106 peripheral blood neutrophils or intraparenchymal pulmonary neutrophils, pooled from 8 or 12 mice, were incubated on ice for 15 min in buffer A (35). After cytoplasm was removed from the nuclei by 15 passages through a 25-gauge needle, the nuclei were centrifuged at 4°C for 6 min at 600 × g. After the nuclear pellet was incubated on ice for 15 min in buffer C (35), the extract was centrifuged at 4°C for 10 min at 12,000 × g. The supernatant was collected, divided into aliquots, and stored at −86°C. Protein concentration was determined by using the Coomassie-Plus protein assay reagent standardized to BSA according to the manufacturer’s protocol.
Activation of the transcriptional factors, NF-κB and CREB, was determined by EMSA analysis, as described previously in our laboratory (17, 18). The κB DNA sequence of the Ig gene (36) and cAMP responsive element (CRE) conserved element (37) were used. Synthetic double-stranded sequences (with enhancer motifs italicized) were fill in labeled with [α-32P]dATP using Sequenase DNA polymerase (κB: 5′-TTTTCGAGCTCGGGACTTTCCGAGC-3′, 3′-GCTCGAGCCCTGAAAGGCTCGTTTT-5′; CRE: 5′-TTTTCGAGCTCTGACGTCAGAGC-3′, 3′-GCTCGAGACTGCAGTCTCGTTTT-5′).
DNA binding reaction mixtures of 20 μl contained 1 μg nuclear extract, 10 mM Tris-Cl, pH 7.5, 50 mM EDTA, 0.5 mM DTT, 1 mM MgCl2, 4% glycerol, 0.08 μg poly (dI-dC)·poly (dI-dC), and 0.7 fmol 32P-labeled double-stranded oligonucleotide. After the samples were incubated at room temperature for 20 min, they were loaded onto a 4% polyacrylamide gel (acrylamide/bisacrylamide 80:1, 2.5% glycerol in Tris-borate-EDTA) at 10 V/cm. Each gel was then dried and subjected to autoradiography.
For supershift studies, 1 μl anti-phosphorylated CREB antiserum was added to the DNA binding reaction just before the 20 min incubation. Similarly, 5 μl anti-p50, 1 μl anti-p65, or 5 μl anti-cRel antiserum was added to the reaction mixture containing the κB oligonucleotide. Binding of the Ab to the appropriate transcriptional factor was indicated by a supershift in the EMSA. Specificity of binding was also confirmed by incubation with a 500-fold excess of unlabeled oligonucleotide.
Because of inherent variability between groups of mice, for each experimental condition, the entire group of animals was prepared and studied at the same time. For each experimental condition, mice in all groups had the same birth date and had been housed together. Separate groups of mice were used for MPO assays, semiquantitative PCR, and EMSA. For semiquantitative PCR, cells were obtained individually from each animal and analyzed individually before calculating group data. Data are presented as mean ± SEM for each experimental group. One-way ANOVA and the Tukey-Kramer multiple comparisons test or Student’s t test was used for comparisons between data groups. Values of p < 0.05 were considered significant.
Effects of hemorrhage or endotoxemia on MPO activity in the lung
Both hemorrhage and endotoxemia produced significant increases in lung MPO activity that began within 15 min of blood loss or endotoxin injection and reached maximal values at 60 min. In mice hemorrhaged 1 h previously, lung MPO activity was more than triple that present in control unmanipulated mice (Fig. 1). Similarly, an almost 5-fold increase in lung MPO activity compared with controls was found 1 h after endotoxin administration (Fig. 1). Because the maximal influx of neutrophils into the lungs occurred 1 h after hemorrhage or endotoxemia, the same interval at which maximal NF-κB activation in intraparenchymal lung cells had previously (22) been shown to occur after blood loss, subsequent experiments examined lung neutrophils at this time point after hemorrhage or endotoxemia.
Hemorrhage or endotoxemia increase IL-1β, TNF-α, and MIP-2 expression in lung neutrophils
In previous experiments (11) using immunohistochemistry, we demonstrated that IL-1β was increased in lung, but not peripheral blood neutrophils, within 1 h of hemorrhage or endotoxemia. To examine more completely the effects of hemorrhage or endotoxemia on proinflammatory cytokine expression in lung neutrophils, we performed semiquantitative PCR for IL-1β, TNF-α, and MIP-2 in lung and peripheral blood neutrophils, isolated 1 h after either hemorrhage or endotoxemia.
Both hemorrhage and endotoxemia produced significant increases in IL-1β, TNF-α, and MIP-2 expression among lung neutrophils compared with that found in lung or peripheral blood neutrophils from control, unmanipulated mice (Figs. 2 and 3). In contrast, blood neutrophils showed little evidence of activation after either hemorrhage or endotoxemia. MIP-2 expression was significantly increased in blood neutrophils after hemorrhage, but not endotoxemia. However, the hemorrhage-induced increase in MIP-2 mRNA levels in blood neutrophils was less than that found among lung neutrophils. Although there was a trend toward increased TNF-α expression in blood neutrophils after endotoxemia, this was again less than that seen in lung neutrophils. Neither sham hemorrhage nor sham endotoxemia produced any alterations in cytokine expression among lung neutrophils.
Hemorrhage or endotoxemia activate CREB and NF-κB in lung neutrophils
Activation of the transcriptional regulatory factors NF-κB or CREB can enhance expression of IL-1β, TNF-α, and MIP-2 through interaction with binding domains in their promoters (14, 15, 16). Therefore, we used EMSA analysis to determine whether NF-κB or CREB activation might be a potential mechanism for the observed increases of proinflammatory cytokine expression in lung neutrophils after hemorrhage or endotoxemia.
Increased activation of both NF-κB and CREB occurred in lung, but not in blood neutrophils after hemorrhage or endotoxemia (Fig. 4). The hemorrhage- and endotoxin-induced NF-κB complex contained both p50 and p65 subunits as shown by supershifts when Abs to p65 or p50 were added to the reaction mixture.
Effects of xanthine oxidase inhibition on hemorrhage- or endotoxin-induced lung neutrophil accumulation and activation
In previous studies (26), we found that inhibition of xanthine oxidase prevented hemorrhage-induced increases in IL-1β and TNF-α expression among unfractionated intraparenchymal pulmonary cell populations. Possible mechanisms for this effect would include a role for xanthine oxidase in mediating neutrophil trafficking to the lungs or in activating lung neutrophils to produce proinflammatory cytokines. To investigate these possibilities, xanthine oxidase activity was inhibited by feeding mice an allopurinol-enriched or a control diet for 1 wk.
Blockade of xanthine oxidase activity by allopurinol feeding did not affect endotoxin- or hemorrhage-associated increases in lung MPO activity (Fig. 5). However, hemorrhage-induced increases in IL-1β, TNF-α, and MIP-2 mRNA levels among lung neutrophils were prevented in mice given an allopurinol-containing diet (Fig. 6). In contrast, xanthine oxidase blockade did not affect endotoxin-associated increases in proinflammatory cytokine expression by lung neutrophils.
To verify that xanthine oxidase blockade, and not other aspects of the allopurinol-enriched diet, was responsible for the observed inhibition of lung neutrophil activation after hemorrhage, we used tungsten feeding as an alternative method to block xanthine oxidase activity. As was the case with allopurinol feeding, administration of the xanthine oxidase-depleting, tungsten-enriched diet prevented hemorrhage-induced increases in proinflammatory cytokine expression by lung neutrophils.
Effects of xanthine oxidase inhibition on CREB and NF-κB activation in lung neutrophils after hemorrhage or endotoxemia
To examine more completely the mechanism by which xanthine oxidase inhibition prevented hemorrhage-induced increases in proinflammatory cytokine expression by lung neutrophils, we fed mice an allopurinol-enriched or control diet and then determined NF-κB and CREB activation in lung neutrophils. Hemorrhage-associated activation of NF-κB in lung neutrophils was not affected by allopurinol feeding (Fig. 7,A). In contrast, total levels of CREB were increased in nuclear extracts of lung neutrophils isolated from hemorrhaged animals fed a xanthine oxidase-blocking, allopurinol-containing diet compared with those given a control diet (Fig. 7,B). Supershift analysis showed increased levels of transcriptionally active, serine 133-phosphorylated CREB after hemorrhage in allopurinol-fed mice compared with those receiving a control diet (Fig. 7 C).
As was the case after hemorrhage, allopurinol feeding did not affect NF-κB activation in lung neutrophils after endotoxin administration (Fig. 7,D). However, in contrast to the situation with hemorrhage, where allopurinol feeding resulted in increased CREB activation, no such effect of xanthine oxidase inhibition on phosphorylated CREB levels was present in lung neutrophils from endotoxin-treated mice (Fig. 7 E).
To verify that xanthine oxidase inhibition was responsible for the observed increase in CREB activation in lung neutrophils after hemorrhage, mice were fed either a xanthine oxidase-depleting, tungsten-enriched diet or a control diet and then subjected to hemorrhage. The results with tungsten feeding were similar to those observed after providing an allopurinol-containing diet. In particular, as was the case with allopurinol feeding, increased amounts of phosphorylated CREB were present after hemorrhage in nuclear extracts from lung neutrophils of tungsten-fed mice as compared with those from animals given a control diet (Fig. 8,A). Tungsten feeding, like allopurinol, did not affect hemorrhage-induced NF-κB activation in lung neutrophils (Fig. 8 B).
Neutrophils that accumulate in the lungs play an important role in the development of acute inflammatory lung injury (1, 4, 6, 7). Our previous studies (11) found that the neutrophils that traffic to the lungs after blood loss or endotoxin administration produce increased amounts of IL-1β. The present experiments extend those observations by demonstrating that hemorrhage or endotoxemia also induce increased expression of TNF-α and MIP-2 in lung neutrophils. The expression of proinflammatory cytokines by lung neutrophils after hemorrhage or endotoxemia was greater than that seen in blood neutrophils, demonstrating that these events lead to compartmentalization of activated neutrophils in the lungs.
Previous studies (27) showed that xanthine oxidase-derived ROI contributed to hemorrhage-associated increases in IL-1β and TNF-α mRNA levels in the lungs. The present experiments demonstrate that enhanced expression of IL-1β, TNF-α, and MIP-2 in lung neutrophils is dependent on xanthine oxidase after hemorrhage, but not endotoxemia. These results indicate that intracellular signaling pathways initiated by hemorrhage and leading to increased expression of proinflammatory cytokines in lung neutrophils are distinct from those associated with endotoxemia. In particular, xanthine oxidase-derived ROI appear to be involved in hemorrhage-induced signaling pathways, but not in those initiated by endotoxemia.
In contrast to the role of xanthine oxidase in modulating hemorrhage-induced proinflammatory cytokine expression in lung neutrophils, inhibition of xanthine oxidase did not affect accumulation of neutrophils in the lungs after hemorrhage or endotoxemia. These results indicate that trafficking of neutrophils to the lungs after hemorrhage is distinct from their activation to express proinflammatory cytokines and suggest that neutrophils are activated after they arrive in the lungs. In our previous studies (38), we demonstrated that hemorrhage induced rapid increases of ICAM-1 and p-selectin on pulmonary endothelial cells, providing a possible mechanism contributing to the accumulation of neutrophils in the lungs. The present experiments would suggest that activation of pulmonary neutrophils then occurs by xanthine oxidase-dependent mechanisms after hemorrhage and by xanthine oxidase-independent mechanisms after endotoxemia.
Both NF-κB and CREB were activated in lung neutrophils after hemorrhage or endotoxemia. Whereas previous studies (23) have shown that in vitro culture of neutrophils with LPS resulted in NF-κB activation, these appear to be the first experiments to show that neutrophils in vivo also demonstrate activation of NF-κB after endotoxin exposure. CREB has not previously been demonstrated to be activated in neutrophils. In our previous studies (17, 18, 21, 22), activated NF-κB and CREB were present in intraparenchymal pulmonary cell populations collected after hemorrhage. Because neutrophils make up the majority of cells in these isolated lung cell populations, it is likely that the NF-κB and CREB signals that were observed after hemorrhage were primarily from neutrophils. Nevertheless, the present results do not eliminate the possibility that hemorrhage or endotoxemia produce activation of NF-κB or CREB in other lung cell populations.
Xanthine oxidase-derived ROI are involved in the activation of CREB and NF-κB in the lungs (17, 21, 22). Because binding sites for CREB and NF-κB are present in the promoters of the IL-1β, TNF-α, and MIP-2 genes, modification of the activity of these transcriptional factors may affect expression of the proinflammatory cytokines examined in these experiments. ROI have been postulated to affect NF-κB activation through promoting IκB-α degradation and enhancing NF-κB translocation to the nucleus (39, 40). However, in the present experiments, even though proinflammatory cytokine expression was decreased in hemorrhaged mice given xanthine oxidase-blocking diets, increased levels of nuclear NF-κB continued to be present in their lung neutrophils. Therefore, these results do not indicate that inhibition of xanthine oxidase prevented hemorrhage-induced proinflammatory cytokine expression through affecting steps involved in the movement of NF-κB to the nucleus.
Further enhancement in CREB activation, as manifested by increased amounts of transcriptionally active serine 133-phosphorylated CREB, was found in the nuclei of lung neutrophils from hemorrhaged, but not endotoxin-treated, mice fed xanthine oxidase-inhibiting diets. Because hemorrhage-associated translocation of NF-κB did not appear to be affected by xanthine oxidase blockade, our results would suggest that increased activation of CREB, through xanthine oxidase-dependent pathways initiated by hemorrhage, may have contributed to the decreased expression of proinflammatory cytokines in lung neutrophils. A possible mechanism through which CREB activation can inhibit proinflammatory cytokine transcription involves competition between NF-κB and phosphorylated CREB for the KIX region of the transcriptional coactivator CREB-binding protein (CBP)/p300 (41).
Activity of many inducible transcription factors, including NF-κB and CREB, is regulated through their interaction with cellular coactivators, such as CBP/p300 (41, 42, 43, 44). Coactivator molecules link promoter-bound transcription factors with the general transcriptional machinery. In addition to associating with NF-κB and CREB, CBP/p300 can also interact with TATA box-binding protein (43) and TFIIB (45), becoming part of the general transcriptional apparatus. The association between CBP/p300 and the p65 subunit of the NF-κB heterodimer occurs through a bivalent interaction consisting of a phosphorylation-independent and a protein kinase A-phosphorylation-dependent interaction (41). The phosphorylation-dependent interaction of NF-κB p65 with CBP/p300 involves the KIX region of CBP/p300, which is the same region responsible for binding serine 133-phosphorylated CREB (41, 45). Disruption of the interaction between CREB or NF-κB and CBP/p300 by mutagenesis significantly decreases the efficiency of NF-κB- or CREB-dependent transcription (41).
Interaction of the same domain in CBP/p300 with both CREB and NF-κB p65 provides a possible explanation for the results of the present experiments, in which xanthine oxidase blockade in hemorrhaged mice was associated with further increases in CREB activation, but decreased proinflammatory cytokine expression, even though nuclear levels of NF-κB continued to be elevated. In particular, the amounts of CBP/p300 present in the nucleus appear to be limiting for transcription, because increased expression of CBP/p300 in transfection experiments stimulates transcription (41). If phosphorylated CREB has a higher affinity for CBP/p300 than does NF-κB, then increased amounts of phosphorylated CREB could sequester CBP/p300, thereby inhibiting NF-κB-dependent transcription. Evidence for such a mechanism has been provided by Parry and Mackman (46, 47). In their experiments, activation of the protein kinase A signaling pathway inhibited NF-κB-mediated transcription by generating increased amounts of phosphorylated CREB. In those studies (46, 47), elevations in cAMP and activation of protein kinase A resulted in increased amounts of phosphorylated CREB and inhibited NF-κB-mediated transcription of TNF-α, endothelial leukocyte adhesion molecule-1, and VCAM-1, but did not prevent nuclear translocation of NF-κB heterodimers, consistent with the results of the present experiments.
The present results may have important implications for the design of future clinical trials investigating therapeutic agents for patients with ARDS. Previous studies of pharmacologic interventions for ARDS have included patients with diverse etiologies for acute lung injury (4, 5, 6). Yet, if neutrophil activation and the development of ARDS result from different signaling pathways in patients with endotoxemia compared with patients with massive blood loss, then, dependent on the site of action of the therapy being examined, one patient group may be benefited whereas the other would not. Depending on the composition of the overall study population, such benefit in a particular subgroup may be diluted out by larger numbers of unresponsive patients, causing the final study results to be negative.
Contemporary definitions of ARDS and acute lung injury are clinical and require decreased arterial oxygen concentrations with associated bilateral pulmonary infiltrates on chest radiographs (3). The present experiments, showing differing mechanisms for the activation of lung neutrophils after hemorrhage or endotoxemia, would suggest that subpopulations of patients with ARDS might be identified based on the activation of specific signaling pathways. The evolution of definitions of ARDS to include such mechanistic parameters may allow better targeting of pharmacologic interventions to patients who would be expected to benefit from modulation of specific signaling pathways.
This work was supported by Grant HL 50284 from the National Institutes of Health.
Abbreviations used in this paper: ARDS, acute respiratory distress syndrome; MIP, macrophage inflammatory peptide; CREB, cAMP response element binding protein; ROI, reactive oxygen intermediates; MPO, myeloperoxidase; CRE, cAMP responsive element; CBP, CREB-binding protein.