INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Gram-negative septicemia is frequently complicated by the occurrence of adult respiratory distress syndrome (ARDS) (1). In the pathogenesis of ARDS various mechanisms appear to play a role, such as hypoxemia, increased pulmonary capillary permeability, and a reduction in lung compliance due to changes in the surfactant system (2). Most of these pathogenetic changes are mediated by the activation of polymorphonuclear and mononuclear cells and by the occurrence of multiple cytokines and chemokines (3). In addition, it has been demonstrated that intra-alveolar fibrin deposition may play an important role as well (2, 4). This fibrin deposition appears to be related to enhanced bronchoalveolar procoagulant activity and a simultaneously occurring reduction in fibrin degradation, due to an impaired function of the bronchoalveolar fibrinolytic system (5, 6).

Further unraveling of the pathogenetic mechanisms responsible for bronchoalveolar fibrin deposition associated with septicemia seems relevant for the future development of modalities for prevention and treatment. Recently, it was shown that intravenous administration of low doses of endotoxin to chimpanzees provides a useful model for reproducing many of the inflammatory events that occur during the early phases of Gram-negative sepsis (7, 8). Moreover, chimpanzees show an identical response to endotoxin compared with humans, and the immunoreactivity of their coagulation system proteins are virtually identical to that in humans, which allows the use of sensitive immunoassays developed for human proteins to monitor activation and inhibition of coagulation and fibrinolysis in these animals (7, 9). Therefore, in this study we have employed this chimpanzee model to analyze bronchoalveolar coagulation and fibrinolysis upon the intravenous administration of endotoxin in order to understand the processes of local fibrin deposition in more detail.

Recently, we and others have shown that changes in the coagulation and fibrinolytic system in plasma during endotoxemia are mediated by several cytokines, particularly tumor necrosis factor alpha (TNF-), interleukin 1 (IL-1), and interleukin 6 (IL-6) (7, 10). Endotoxin-induced activation of coagulation in plasma could be eliminated by the administration of a monoclonal anti-IL-6 antibody, indicating a pivotal significance of IL-6 in the activation of coagulation. TNF- appeared to play a central role in the endotoxin-induced activation and subsequent inhibition of fibrinolysis in plasma, since these fibrinolytic effects could be completely blocked by anti-TNF- treatment strategies. Since both TNF- and IL-6 have been shown to be involved in the pathophysiology of ARDS (14, 15), we hypothesized that these cytokines might be involved as mediators in the endotoxin-induced changes in bronchoalveolar coagulation and fibrinolysis. Therefore, we have investigated the effect of monoclonal anti-TNF- or anti-IL-6 antibody administration on endotoxin-induced bronchoalveolar procoagulant and anti-fibrinolytic response.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Chimpanzees

Adult chimpanzees (Pan troglodytes) were housed at the New York University Laboratory for Experimental Medicine and Surgery in Primates (LEMSIP, Tuxedo, NY). The animals selected for study weighed between 35 and 70 kg, exhibited normal kidney as well as liver function, and had normal routine coagulation tests (activated partial thromboplastin time and prothrombin time). The study protocols were approved by the animal health and welfare committees of the primate center and were conducted according to the NIH guidelines (Guide for the Care and Use of Laboratory Animals, NIH 86-23, 1985).

Experimental Protocols

After sedation with ketamine chloride, the chimpanzees were intubated and maintained under general anesthesia with nitrous oxide and halothane throughout the experiment, which lasted 5 h. Supplemental oxygen was administered throughout the procedure. Arterial blood pressure, heart rate, and oxygen saturation were continuously recorded with an automated blood pressure device, and pulsoxymeter and rectal temperature were measured in 15-min intervals.

Animals received U.S. Reference Escherichia coli endotoxin (Lot EC-5 from E. coli 0113, Bureau of Biologics, Food and Drug Administration, Bethesda, MD, kindly provided by Dr. D. Hochstein) with a specific activity of 10 units (U) per nanogram (ng), which was administered as a bolus injection through intravenous tubing at a dose of 4 ng/kg. The dose of endotoxin chosen was found to be safe for administration to humans and chimpanzees and has been shown to induce reproducible systemic inflammatory responses (7, 8, 16). Clinical signs and symptoms are minimal. Control studies, i.e., investigation of bronchoalveolar lavage (BAL) fluids at equal time points from chimpanzees that were injected with saline instead of endotoxin, have demonstrated that the experimental procedures themselves do not elicit changes in the inflammatory and coagulation parameters under investigation (7).

Treatment

Six chimpanzees received a bolus injection of purified endotoxin only (control group). In a second group of four animals the injection of endotoxin was immediately followed by the administration of anti- TNF- monoclonal antibody (mAb) (provided by Bayer, Wuppertal, Germany), given as a bolus injection of 15 mg/kg body weight. The antibody is a murine TNF- neutralizing mAb (immunoglobin G₁ [IgG₁]) produced by hybridoma culture and purified from culture harvests by cell separation, polyethylene glycol precipitation, anion exchange, and size exclusion chromatography (17). The dose of the antibody required to reduce recombinant human TNF-induced cytotoxicity in the WEHI bioassay by 50% is 0.27 µg anti-TNF- mAb per nanogram of recombinant human TNF. Previous experiments have shown good cross-reactivity of the antibody with chimpanzee TNF (13).

A third group of four animals received a bolus intravenous injection of anti-IL-6 mAb (30 mg) immediately following the administration of endotoxin. Anti-IL-6 mAb (CLB.IL-6/#8) is a murine IgG₁ against human IL-6 that potently neutralizes IL-6 activity in biological assays (KD 6 × 10¹² M) (18). CLB.IL-6/#8 does not discriminate between human and chimpanzee IL-6 (10). The antibodies were produced by in vitro cultures and purified by protein A-sepharose chromatography.

Results of systemic cytokine measurements and inflammatory parameters as well as plasma measurements of blood coagulation and fibrinolysis in the chimpanzees included in this study have been reported previously (10, 13).

BAL and Processing of Lavage Fluid

Bronchoscopy and BAL were performed before the injection of endotoxin (baseline) and at 4 h after the injection of endotoxin. BAL was performed, passing a flexible fiberoptic bronchoscope (Model BF-B4, Olympus Corp., New York, NY) through the intratracheal tube into the trachea. The tip of the bronchoscope was wedged into a lower lobe subsegmental bronchus. The initial bronchoscopy (baseline) was performed on the right side and the second bronchoscopy (following treatment) on the left side. Seven successive 20-ml aliquots of 0.9% NaCl at 37° C were instilled and immediately aspirated by gentle suction (50 mm Hg). The first aliquot recovered was discarded and the remaining aliquots were pooled. The pool contained less than 1 × 10⁵ erythrocytes/ml. The pool was filtered through a siliconized gauze and centrifuged at 500 × g and 4° C for 10 min. The cell-free supernatant was 10-fold concentrated using Centriplus-10 concentrators (Amicon, Capelle a/d Ijssel, The Netherlands) and stored at 70° C until analysis. Cells were resuspended in phosphate-buffered saline (PBS) and counted manually in a counting chamber. Subsequently, cells were cytocentrifuged at 500 rpm for 2 min and stained with Jenner-Giemsa. A total of 2 × 250 cells were counted.

Assays

TNF- and IL-6 levels were determined using enzyme-linked immunosorbent assays (ELISA) (Medgenix, Brussels, Belgium).

Activation of coagulation in BAL fluid was determined by measuring specific markers for thrombin generation, i.e., prothrombin activation fragment F1+2 and thrombin-antithrombin (TAT) complexes, using their respective ELISAs (Behringwerke AG, Marburg, Germany). Antithrombin activity was measured using an amidolytic assay (19). To assess whether activation of the contact system contributed to the thrombin generation, factor XIIa-C1-inhibitor complexes and kallikrein-C1-inhibitor complexes were determined with radioimmunoassays, as previously described (20).

Activity of the fibrinolytic system was assessed by measuring plasminogen activator activity (PAA). To analyze the origin of the PAA, antigenic levels of plasminogen activators (tissue-type plasminogen activator [t-PA] and urokinase-type plasminogen activator [u-PA]) as well as inhibitors of plasminogen activators (plasminogen activator inhibitor [PAI] type 1 and type 2) were measured. PAA was measured by an amidolytical assay (21). Briefly, 25 µl of plasma was mixed to a final volume of 250 µl with 0.1 M TRIS HCl, pH 7.5, 0.1% (vol/vol) Tween-80, 0.3 mM S-2251 (Chromogenix, Mölndal, Sweden), 0.13 M plasminogen, and 0.12 mg/ml CNBr fragments of fibrinogen (Chromogenix). The results were expressed as percentage of baseline level. t-PA antigen and u-PA antigen were measured with their respective ELISA's (Asserachrom t-PA; Diagnostica Stago, Asnieres-sur-Seine, France and Gaubius Institute TNO, Leiden, The Netherlands). PAI-1 and PAI-2 were also measured by ELISA (TintElize PAI-1 and PAI-2, Biopool, Umea, Sweden).

To ascertain that changes in concentration of coagulation and fibrinolytic proteins in BAL fluid could not be ascribed to differences in total protein concentration between lavages, total protein and albumin were measured in BAL fluids by a spectrophotometric method using bicinchoninic acid and an immunoturbidimetric assay using anti-albumin A001 (Dako, Glostrup, Denmark), respectively. Also, the albumin BAL fluid/plasma ratio was calculated.

Statistical Analysis

All values are given as means ± SD. Means of groups were compared with analysis of variance using the Newman-Keul's test to correct for multiple comparisons. p Values less than 0.05 were considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell Counts, Protein, and Cytokines

In the BAL fluid obtained after administration of endotoxin, there was a small but insignificant increase in cell count (Table 1). This increase was not different between the treatment groups. BAL neutrophil count showed an increased percentage of neutrophils after the administration of endotoxin (from 4.2 ± 1.1% to 8.6 ± 1.4% in the endotoxin alone group, p < 0.05). This increase in neutrophil count was not different in chimpanzees receiving endotoxin in combination with anti- IL-6, but appeared to be somewhat reduced in animals treated with anti-TNF- (Table 1).

Analysis of total protein and albumin concentration in BAL fluids as well as the albumin BAL fluid/plasma ratio showed no statistically significant differences between the various groups (Table 1). Hourly blood gas analyses did not change significantly during the experiment and were similar in all experimental groups (data not shown).

Levels of TNF- and IL-6 were not detectable in BAL fluid obtained before the injection of endotoxin. As shown in Table 2, after the injection of endotoxin alone levels of TNF- and IL-6 were slightly elevated (4.5 ± 3.9 pg/ml and 10.2 ± 5.0 pg/ml, respectively, p = NS). After the administration of endotoxin in combination with the monoclonal anti-TNF- antibody, TNF- was not detectable in BAL fluid, whereas IL-6 levels were not significantly different compared with the administration of endotoxin alone. Infusion of the combination of endotoxin and anti-IL-6 antibody did not affect bronchoalveolar TNF- or IL-6 levels.

Coagulation

The injection of endotoxin resulted in a substantial increment in bronchoalveolar thrombin generation, as reflected by a 3.7-fold increase in levels of prothrombin activation fragment F1+2 (p < 0.01) and a 2.5-fold increase in thrombin-antithrombin levels (p < 0.05) (Figure 1). The level of antithrombin in BAL fluid was 0.29 ± 0.12 µg/ml at baseline and dropped to almost undetectable levels (0.02 ± 0.05 µg/ml) after the administration of endotoxin (p < 0.05). Sensitive markers for activation of the contact system, i.e., factor XIIa-C1-inhibitor complexes, and kallikrein-C1-inhibitor complexes were not detectable, neither before nor after injection of endotoxin (data not shown).

View larger version (12K): [in this window] [in a new window]   Figure 1.   Endotoxin-induced activation of bronchoalveolar coagulation. Levels of prothrombin activation fragment F1+2, thrombin-antithrombin complexes and antithrombin activity in bronchoalveolar lavage fluid before and 4 h after the intravenous injection of endotoxin. Mean values and SD are given, and statistical significance is indicated by: *p < 0.05; **p < 0.01.

The combined administration of endotoxin and monoclonal anti-TNF- antibody resulted in a reduction in the post-endotoxin level of F1+2 and TAT complexes (Figure 2). The level of F1+2 in BAL fluid after endotoxin and anti-TNF- was 38% lower compared with the level of F1+2 after endotoxin alone (p = 0.07), whereas TAT complex levels after endotoxin and anti-TNF- were 32% lower as compared with the administration of endotoxin alone (p < 0.05).

View larger version (25K): [in this window] [in a new window]   Figure 2.   Effect of anti-TNF- and anti-IL-6 treatment on endotoxin-induced activation of bronchoalveolar coagulation. Levels of prothrombin activation fragment F1+2 and thrombin-antithrombin complexes in bronchoalveolar lavage fluid before and 4 h after the injection of endotoxin alone (left bars), endotoxin in combination with monoclonal anti-TNF- antibody (middle bars), and endotoxin in combination with monoclonal anti-IL-6 antibody (right bars). Mean values and SD are given. Statistically significant differences between treatment with either anti-TNF- or anti-IL-6 antibody and controls (endotoxin only) are indicated by: *p < 0.05; **p < 0.01.

The administration of the combination of endotoxin plus anti-IL-6 monoclonal antibody resulted in a complete inhibition of endotoxin-induced thrombin generation. As shown in Figure 2, post-endotoxin BAL fluid levels of F1+2 and TAT complexes were 0.37 ± 0.07 nmol/L and 1.1 ± 0.4 µg/ml as compared with levels of 1.11 ± 0.2 nmol/L and 2.8 ± 0.4 µg/ml after the administration of endotoxin alone (p < 0.01 for both F1+2 and TAT complexes).

Fibrinolysis

Plasminogen activator activity in BAL fluid was mainly due to u-PA, since levels of t-PA were not detectable (Table 3). Administration of endotoxin elicited a four-fold increase in u-PA antigen levels in BAL fluid. As shown in Figure 3, in post- endotoxin BAL fluid a 77% reduction in plasminogen activator activity was found. Analysis of this reduction in PAA showed that this was caused by an 8.3-fold increase in levels of plasminogen activator inhibitor, mainly PAI-2 (Figure 3). Taken these data together, it can be concluded that the endotoxin- induced increase in PAI-2 completely blocked the u-PA-mediated fibrinolytic activity in post-endotoxin BAL fluid. Hence, the increase in the level of u-PA antigen in the BAL fluid after the administration of endotoxin almost exclusively reflected u-PA complexed with PAI-2.

View larger version (17K): [in this window] [in a new window]   Figure 3.   Endotoxin-induced effects on bronchoalveolar fibrinolysis. Levels of plasminogen activator activity in bronchoalveolar lavage fluid before and 4 h after the intravenous injection of endotoxin (upper panel ). The reduction in plasminogen activator activity appeared to be caused by the endotoxin-induced increase in levels of plasminogen activator inhibitors, PAI-1 (left bars) and PAI-2 (right bars), respectively (lower panel ). Mean values and SD are given. Statistically significant differences between pre- and post-endotoxin are indicated by: *p < 0.05; **p < 0.01.

The reduction in bronchoalveolar fibrinolytic activity after the administration of endotoxin could be prevented with the simultaneous infusion of anti-TNF- monoclonal antibody. As shown in Figure 4, this could be explained by the fact that the administration of anti-TNF- antibody appeared to block the endotoxin-induced increase in PAI-2 in BAL fluid. Post-endotoxin BAL fluid levels of PAI-2 in chimpanzees treated with endotoxin in combination with anti-TNF- were 3.4 ± 0.7 ng/ml, as compared with levels of 12.4 ± 2.2 ng/ml after the administration of endotoxin alone (p < 0.01). Administration of anti-TNF- also prevented the endotoxin-induced increase in bronchoalveolar u-PA (u-PA antigen before endotoxin 0.6 ± 0.1 µg/L and after endotoxin 0.8 ± 0.2 µg/L). In contrast, anti- IL-6 antibody did not affect the endotoxin-induced enhancement of bronchoalveolar PAI-2 and the associated reduction in fibrinolytic activity (Figure 4).

View larger version (31K): [in this window] [in a new window]   Figure 4.   Effect of anti-TNF- and anti-IL-6 treatment on endotoxin-induced effects on bronchoalveolar fibrinolysis. Levels of plasminogen activator activity and plasminogen activator inhibitor-2 in bronchoalveolar lavage fluid before and 4 h after the injection of endotoxin alone (left bars), endotoxin in combination with monoclonal anti-TNF- antibody (middle bars), and endotoxin in combination with monoclonal anti-IL-6 antibody (right bars). Mean values and SD are given. Statistically significant differences between treatment with either anti-TNF- and controls (endotoxin only) are indicated by: **p < 0.01.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We have attempted to analyze the mechanisms that play a role in the pathologic fibrin deposition that occurs in the evolution of ARDS. Therefore, we have adopted a chimpanzee model of endotoxemia, allowing us to study sensitive markers for activation of coagulation and fibrinolysis in BAL fluid after an intravenous endotoxin challenge. Subsequently, in this model the role of TNF- and IL-6 could be investigated.

Our findings indicate substantial bronchoalveolar thrombin generation after the intravenous infusion of endotoxin, as evidenced by an increase in prothrombin activation fragment F1+2 and TAT complexes present in BAL fluid. Measurement of albumin in BAL fluid indicated that no significant protein leakage occurred, which is understandable after the relatively mild endotoxin stimulus that was given to the chimpanzees. Hence, the elevated markers for thrombin generation appear to reflect local prothrombin activation and thrombin generation rather than transcapillary leakage from the systemic circulation.

Sensitive markers for activation products of the contact system could not be detected in BAL fluid after the administration of endotoxin. This implies that the activation of bronchoalveolar coagulation most probably proceeded via the tissue factor-factor VII(a) route. This is in agreement with previous reports, showing that in patients with (evolving) ARDS enhanced tissue factor-factor VIIa-dependent activation of factor X could be detected in BAL fluid (22, 23). Conceivably, the initiation of bronchoalveolar coagulation activation occurs at the surface of the (activated) mononuclear cells. In vitro bronchoalveolar procoagulant activity is dependent on tissue factor expression at the alveolar macrophage surface (24, 25). Interestingly, it has been shown that even low-grade endotoxemia, such as induced in our study, is associated with activation of systemic and alveolar mononuclear cells (16, 26, 27). Another mechanism contributing to the endotoxin-induced enhancement of procoagulant activity may be the impaired action of physiologic coagulation inhibitors. Levels of the major inhibitor of thrombin, i.e., antithrombin, were relatively low in BAL fluid and were undetectable after the intravenous administration of endotoxin. This may suggest that antithrombin is consumed as a result of enhanced thrombin generation and becomes completely depleted. This depletion may explain the moderate discrepancy between the increase in levels of F1+2 and TAT complexes. It is likely that the depletion of antithrombin may further contribute to the procoagulant activity in BAL fluid.

Simultaneously with the enhancement of bronchoalveolar coagulation activation, a significant depression of fibrinolytic activity occurred. Bronchoalveolar fibrinolytic activity (mainly urokinase-type plasminogen activator activity), sharply dropped, due to increased levels of plasminogen activator inhibitors. Interestingly, and in accordance with previous reports by others (10), the main plasminogen activator inhibitor appeared to be PAI-2. In the circulation PAI-2 is not an important physiologic inhibitor of fibrinolysis and can only be detected during pregnancy (probably released from the placenta) (28), but it appears that this fibrinolytic inhibitor also plays an important role in the bronchoalveolar compartment. Previous in vitro studies have shown that endotoxin-stimulated human alveolar macrophages are able to secrete relatively large quantities of PAI-2; hence, the increase in bronchoalveolar PAI-2 probably originates from activated macrophages (29). Bronchoalveolar PAI-1 was only minimally increased after the administration of endotoxin. However, in a number of previous reports PAI-1 levels in BAL fluids of patients with ARDS were increased to a more significant extent (30). It might be that in case of severe derangement of bronchoalveolar fibrinolysis, the role of PAI-1 becomes more prominent.

Cytokines represent the pivotal link between inflammation and changes in coagulation and fibrinolysis. A differential role of various cytokines in the endotoxin-induced alterations in coagulation and fibrinolysis has been established, with a central role of IL-6 in the activation of coagulation and for TNF- in the activation and subsequent depression of fibrinolytic activity (10, 13). A similar distinct mediatory role of these two cytokines in endotoxin-induced changes in bronchoalveolar coagulation and fibrinolysis was established by the present data.

Administration of anti-IL-6 mAb completely abolished the endotoxin-induced activation of bronchoalveolar thrombin generation, indicating that also the activation of bronchoalveolar coagulation also is dependent on IL-6. Anti-TNF- treatment only partly inhibited the activation of coagulation. Since TNF- is an important mediator for the occurrence of IL-6 in endotoxemia and sepsis (12, 16), we hypothesize that this effect of anti- TNF- treatment is indirectly caused by a reduction in post-endotoxin IL-6 levels. It has indeed been shown in previous reports that administration of anti-TNF- results in a reduction of endotoxin-induced enhancement of IL-6 levels (7, 13), although this effect could not be detected in the BAL fluids. It should be mentioned, however, that the administration of anti- TNF- did not affect endotoxin-induced activation of coagulation in plasma, whereas in the present study a small but significant inhibition of endotoxin-induced bronchoalveolar thrombin generation was detected. Hence, the proposed TNF--mediated effect on IL-6 appears to be more important at the bronchoalveolar level as compared with the systemic level. Alternatively, it might be that part of the bronchoalveolar procoagulant activity is directly mediated by TNF-.

The effects of endotoxin on bronchoalveolar fibrinolysis appeared to be due to TNF-, since anti-TNF- treatment completely blocked the endotoxin-induced fibrinolytic effects. Similar to the effect of anti-TNF- on endotoxin-induced effects on the blood fibrinolytic system, anti-TNF- treatment blocked the endotoxin-induced increase in bronchoalveolar plasminogen activator inhibitor (in the bronchoalveolar compartment for the major part PAI-2), thereby preventing the depression of bronchoalveolar fibrinolysis. Several reports have been published concerning the therapeutic or protective effect of anti-TNF- strategies in models of ARDS (31, 32). Based on the present observations, it could be speculated that part of the beneficial effect of the inhibition of TNF- is related to the inhibition of bronchoalveolar anti-fibrinolytic mechanisms that may play a role in the development of ARDS.

Although the present findings thus indicate central roles for IL-6 and TNF- in the pathogenesis of endotoxin-induced effects on bronchoalveolar coagulation and fibrinolysis, respectively, levels of these cytokines were only minimally elevated in post-endotoxin BAL fluid. Also, with the exception of anti-TNF- treatment, these levels were not affected by the administration of the respective monoclonal antibodies, which is in agreement with previous studies in humans with low-grade endotoxemia (8, 33, 34). Interestingly, these reports all indicate that despite the absence of a clear enhancement of bronchoalveolar cytokine levels, cytokine-mediated effects, such as activation of mononuclear cells, may occur. Since most bronchoalveolar coagulation and fibrinolytic effects appear to be mediated by mononuclear cells, the effect of endotoxin on systemic cytokine levels apparently causes the bronchoalveolar effects on coagulation and fibrinolysis without marked elevations of local cytokine levels. Also, studies focusing on other cytokines that are important for activation of coagulation and fibrinolysis (such as interleukin-1) could further add to our understanding of deranged bronchoalveolar hemostasis.

The present results, although in line with findings in models of a more severe systemic inflammatory response and associated lung injury, were observed in a model with a relatively mild inflammatory stimulus. The relevance for human disease, i.e., sepsis and associated ARDS, needs to be confirmed. However, most of the observations in our model correlate well with findings in patients with ARDS (22, 23, 30, 35). Hence, assuming that fibrin deposition is of importance for the pathogenesis of ARDS, we hypothesize that strategies directed at the improvement of the bronchoalveolar coagulative-fibrinolytic balance, or directed against the cytokine responses in infectious and inflammatory disease, might be useful for treatment or prevention of ARDS.

In conclusion, our studies show that systemic endotoxemia is associated with bronchoalveolar activation of coagulation and depression of fibrinolysis. The activation of coagulation appears to be mediated via the tissue factor-factor VIIa route, whereas the marked enhancement of plasminogen activator inhibitor, mainly PAI-2, is responsible for the fibrinolytic shutdown. The effect of endotoxin on bronchoalveolar coagulation and fibrinolysis is regulated by cytokines. IL-6 appears to play a central role in the enhancement of bronchoalveolar procoagulant activity, whereas TNF- is pivotal for the fibrinolytic effects. These findings may be helpful to develop improved treatment strategies for patients with or at risk for ARDS.