INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Recent clinical trials have shown that protective ventilation strategies improve survival in patients with acute respiratory distress syndrome (ARDS) (1). In fact, the recent ARDSnet trial, which compared ventilation with a tidal volume (VT) of 6 ml/kg and a VT of 12 ml/kg, was the first large trial to show a beneficial effect of any treatment on the final outcome of patients with ARDS (3, 5). This and the other recent studies suggested that ventilation with volumes or pressures that are too low to cause direct physical damage to the lung can still be high enough to harm the lungs through other mechanisms.

At present, there is no unequivocal expression to describe the ventilation of lungs with inappropriately high pressures or volumes. Ventilation refers to VT multiplied by frequency. However, although there is a commonly accepted term that describes an increased breathing frequency (i.e., hyperventilation), there is no such term that describes ventilation with increased VT. Therefore, we suggest use of the term "overventilation" (OV) to refer to this condition.

Although the mechanistic basis for the detrimental effect of OV is not completely clear, both experimental and clinical evidence suggest that under certain conditions ventilation may itself be a cause of the disease for which it is frequently used as a treatment (i.e., pulmonary and systemic inflammation). The overactivation of the immune system by ventilation has been called "biotrauma" (6). In support of this effect of ventilation, experiments with isolated mouse and rat lungs have shown that OV causes the release of inflammatory mediators into the systemic circulation (7, 8) and into the alveolar space (9). Very recently, it was also shown in ARDS patients that ventilation with reduced VT attenuates pulmonary inflammation as well as pulmonary and systemic cytokine levels (3, 4). All of these findings support the hypothesis that OV elicits an inflammatory response very similar to that caused by bacterial lipopolysaccharide (LPS). The host reaction to LPS is characterized not only by increased cytokine levels, but also by the formation of chemokines and the sensitivity of these effects to steroids (10), at least under experimental conditions. Examples of proinflammatory mediators whose response to LPS is attenuated by steroids include tumor necrosis factor (TNF) (11), interleukin (IL)-1 (11), IL-6 (11) and IL-8 (12), macrophage inflammatory protein (MIP)-1 (11) and MIP-2 (11), macrophage chemotactic protein (MCP)-1 (11) and intercellular adhesion molecule (ICAM)-1 (13). In addition, it has been shown that a pivotal proximal step in the activation of cells by LPS is translocation of the transcription factor nuclear factor (NF)-B to the nucleus (14). Accordingly many, though not all, of the antiinflammatory effects of steroids can be explained by their blocking of NF-B activation.

Therefore, it was the aim of the present investigation to test the hypothesis that both LPS and OV elicit inflammatory responses through a similar mechanism (i.e., activation of NF-B and subsequent formation of cytokines and chemokines). We also investigated the effects of the steroid dexamethasone in our model in order: (1) to show that ventilation-induced cytokine release can be suppressed by antiinflammatory agents; (2) to provide further evidence for a role of NF-B in ventilation-induced mediator release; and (3) to examine whether steroids might be considered as a subsidiary treatment during ventilation. Additionally, to exclude contaminations by LPS as a reason for the OV-induced responses, we also performed experiments with LPS-resistant C3H/HeJ mice. These mice have been known for more than 30 yr for their defective response to LPS, but only recently was the mechanistic basis for this defect (i.e., a mutation in the Toll-like receptor [Tlr]-4) discovered (15). Since then, the central role of Tlr in cellular responses to LPS has been clearly established (16, 17).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Mice

Female BALB/C mice were obtained from the breeding house of the Research Center Borstel, and female C3H/HeJ mice from Charles River Laboratories (Sulzfeld, Germany). All animals were used at a weight of 20 to 23 g. Mice were housed in accordance with the regulations and standards of the U.S. Department of Agriculture and the U.S. Department of Health and Human Services.

Materials

Pentobarbital sodium (Narcoren) was purchased from the Wirtschaftsgenossenschaft Deutscher Tierärzte (Hannover, Germany); low-endotoxin-grade bovine albumin was purchased from Serva (Heidelberg, Germany); dexamethasone and LPS (of Salmonella minnesota), were purchased from Sigma (Deisenhofen, Germany); and RPMI-1640 was purchased from BioWhittaker (Verviers, Belgium).

Isolated Perfused Mouse Lung Preparation

Mouse lungs were prepared and perfused essentially as recently described (8, 18). Briefly, lungs were perfused in a nonrecirculating fashion through the pulmonary artery at a constant flow of 1 ml/min, resulting in a pulmonary artery pressure of 2 to 3 cm H₂O. As a perfusion medium, we used RPMI medium lacking phenol red (37° C) and containing 4% low-endotoxin-grade albumin. Under control conditions, the lungs were ventilated by negative pressure (3 to 10 cm H₂O) at a rate of 90 breaths/min, resulting in a VT of about 200 µl or 340 µl for BALB/C and C3H/HeJ mice, respectively. Artificial thorax chamber pressure was measured with a differential pressure transducer (DP 45-24; Validyne, Northridge, CA), and the airflow rate was measured with a Fleisch-type pneumotachograph tube connected to a differential pressure transducer (DP 45-15; Validyne). Arterial pressure was monitored continuously by means of a pressure transducer (Isotec Healthdyne, Irvine, CA) that was connected to the cannula ending in the pulmonary artery. All data were transmitted to a computer and analyzed with Pulmodyn software (Hugo Sachs Elektronik, March Hugstetten, Germany). VT was derived by integration of the flow rate, and the data were analyzed by applying the formula P = 1/C · V + RL · dV/dT, where P is chamber pressure, C is pulmonary compliance, V is volume, and RL is lung resistance.

Experimental Design

Recently, we showed that both negative- and positive-pressure ventilation lead to the same extent of mediator release (8). From these findings we concluded that it is the transpulmonary pressure, and not, for example, the change in vascular pressure, that is responsible for OV-induced mediator release. In the present study we used negative-pressure ventilation (NPV) because the low perfusion pressures that result from NPV help to minimize hydrostatic edema formation. To distinguish between the effects of LPS and ventilation, and to exclude the possibility that the effects of OV were caused simply by introducing excessive LPS into the lungs, we performed some of our experiments in C3H/HeJ mice that have a defective response to LPS (15).

Ventilation was always pressure-controlled. The different experimental conditions at 2 min after baseline conditions are shown in Table 1. Throughout all experiments, an end-expiratory pressure (EEP) of 3 cm H₂O was applied in the ventilator chamber. In all experiments, the lungs were perfused and ventilated for 60 min under baseline conditions, with an end-inspiratory pressure (EIP) of approximately 10 cm H₂O, resulting in a VT of about 200 µl ( 9 ml/kg) in Balb/C and of 340 µl ( 15 ml/kg) in C3H/HeJ mice. The increased VT in C3H/HeJ mice was in accord with a greater pulmonary compliance in the C3H strain (19). After 60 min of perfusion under baseline conditions, lungs were divided into three groups. The control group was perfused and ventilated with 10 cm H₂O for another 150 min without any other treatment. In the two other experimental groups, after 60 min either the lungs received 50 µg/ml LPS or alternatively, EIP was raised to reach a VT of about 700 µl ( 32 ml/kg; OV; Table 1), and the required EIP of approximately 25 cm H₂O was maintained for the remainder of the experiment. These experiments were performed with BALB/C and in C3H/HeJ mice. Perfusate samples were obtained every 10 min directly from the catheter leading to the left atrium, and were immediately frozen and kept at 20° C until further analysis. In the dexamethasone experiments, the steroid was added at 10 µM to the perfusate after 15 min of perfusion under baseline conditions (i.e., 45 min before initiation of OV or administration of LPS).

Previous in vivo studies had shown that after intraperitoneal injection of LPS, NF-B was activated within 60 min (20). Therefore, lungs for the analysis of NF-B activation were exposed to OV, LPS, or control conditions for 60 min. Subsequently, the surrounding tissue, the heart, and the trachea were trimmed away and the lungs were flash frozen with liquid nitrogen and powdered with a pestle in the constant presence of liquid nitrogen.

Analysis of NF-B Activation by Electrophoretic Mobility Shift Assay

The preparation of nuclear extracts was done as described elsewhere (21). Oligonucleotides containing the NF-B binding sequence of the murine immunoglobulin- light-chain enhancer (sense oligonucleotide: 5'-AGCTCAGAGGGGACTTTCCGAGAGAGCT-3'; antisense oligonucleotide: 5'-AGCTCTCTCGGAAAGTCCCCTCTGAGCT-3') were obtained from MWG-Biotech (Ebersberg, Germany). A quantity of 1.25 pmol of the double-stranded oligonucleotide was end-labeled in the presence of -[³²P]deoxyadenosine triphosphate, using T4-polynucleotidkinase (Boehringer Mannheim, Mannheim, Germany) for 30 min at 37° C. Unincorporated nucleotides were removed by passage through a Nick-S-column (Pharmacia, Freiburg, Germany). A quantity of 7.5 fmol of labeled oligonucleotides was used in the DNA-binding reaction, which was done with 1 µg of crude nuclear extract. The binding reaction also contained 1 µg poly/deoxyinosine-deoxycytosine, 1 µg poly/deoxyadenine-deoxythymidine, 4% Ficoll, 1 mM dithiothreitol, 2 mM MgCl₂, 0.03% NP40, and 60 mM KCl. Reaction mixtures were incubated for 20 min at 4° C and separated by electrophoresis in 4% polyacrylamide gels containing 0.5× ethylenediamine tetraacetic acid (EDTA) (TBE: 45 mM Tris-borate, 1 mM EDTA). Gels were run at 200 V for 1.5 h, sealed, exposed overnight to a phosphor screen, and analyzed with a PhosphorImager (Molecular Dynamics, Krefeld, Germany).

Measurement of Cytokines and Chemokines

Analysis of cytokine and chemokine release from lungs into the perfusate was done with enzyme-linked immunosorbent assays (ELISAs). ELISA kits (detection limits in parentheses) for murine IL-6 (10 pg/ml), TNF- (10 pg/ml), KC (10 pg/ml), JE/MCP-1 (10 pg/ml), MIP-2 (5 pg/ ml), MIP-1 (5 pg/ml), IL-10 (30 pg/ml), interferon (IFN)- (10 pg/ml), eotaxin (5 pg/ml) and vascular endothelial growth factor (VEGF) (5 pg/ ml) were obtained from R&D Systems GmbH (Wiesbaden, Germany). All assays were performed according to the supplier's instructions.

Statistical Analysis

Animals were randomly allocated to each study group. For the electrophoretic mobility shift assays (EMSAs) and ELISAs, the experimenters were blinded to the experimental groups. Data are expressed as mean ± SD. The various time courses of cytokine/chemokine concentrations were analyzed on the basis of the area under the time-versus-concentration curve (AUC) between 60 min and 210 min, except for TNF (which began to appear rather late), for which the AUC was calculated from the data between 150 min and 210 min. Before analysis, the AUC data were checked for homoscedasticity (JMP 3.2.6; SAS Institute Inc., Cary, NC), and in cases of heteroscedasticity the data were log-transformed. The AUC data were evaluated by analysis of variance (ANOVA) followed by Tukey's test. If log-transformation did not result in homogenous variances, the AUC data were analyzed by ANOVA modified according to Welch (JMP 3.2.6), and single contrasts were then determined with Welch's t test. In this case the  level was adjusted for multiple comparisons through Hommel's procedure (22). A value of p < 0.05 was considered significant in all analyses.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experiments with BALB/C Mice

The time course of VT in the different experimental groups is shown in Figure 1. In control lungs and LPS-perfused lungs, VT remained stable throughout the experiments, with similar values in lungs perfused with dexamethasone alone or with dexamethasone/LPS (data not shown). In overventilated lungs the increased VT resulting from the increased EIP decreased over time, possibly because of the exhaustion of surfactant stores, the development of small atelectatic areas, or the formation of mild edema (8). However, it must be emphasized that this ventilation strategy does not result in severe edema or gross lung damage (8).

View larger version (17K): [in this window] [in a new window]   Figure 1.   VT in overventilated and LPS-treated lungs from BALB/C mice. Lungs were perfused for 60 min under control conditions (open squares; n = 4) before they were exposed to either 50 µg/ml LPS (open triangles; n = 4), overventilation (open circles; n = 5) or dexamethasone/overventilation (closed circles; n = 3). The different conditions are detailed in Table 1. Data are mean ± SD. The differences in EIP and VT between OV and the other conditions were highly significant (p < 0.01). The differences between OV and dexamethasone/OV were not statistically significant (p > 0.05).

The activation of NF-B was analyzed in lungs obtained after 60 min of LPS treatment or OV (Figure 2). As compared with control lungs, both LPS treatment and OV caused activation of NF-B in the lung tissue. In both cases the activation of NF-B was abolished by pretreatment with dexamethasone. In the subsequent experiments we studied the release of chemokines and cytokines whose genes are known to contain the NF-B consensus sequence (i.e., KC [23], MIP-2 [23], MCP-1 [23], MIP-1 [23], TNF- [24], and IL-6 [24]). In addition, we studied four mediators that lack an NF-B consensus sequence (i.e., IL-10 [24], interferon- [24], eotaxin [23], and VEGF [25]).

View larger version (41K): [in this window] [in a new window]   Figure 2.   NF-B translocation by OV and LPS, and its prevention by dexamethasone (DEX) in lungs from BALB/C mice. All lungs were perfused for 120 min. After 60 min of baseline perfusion, lungs were exposed for another 60 min to either OV or 50 µg/ml LPS. Dexamethasone was added after 15 min of perfusion. NF-B translocation was determined by EMSA. The NF-B band was abolished in the presence of unlabeled oligonucleotides (data not shown).

Perfusion of lungs from BALB/C mice under control conditions (EIP = 10 cm H₂O, no LPS) caused no or only minor release into the venous effluate of the studied -chemokines (Figure 3), -chemokines (Figure 4), or cytokines (Figure 5). The small basal release of most of the mediators studied was attenuated by pretreatment with dexamethasone. OV caused release of the -chemokines KC and MIP-2 (Figure 3), the -chemokines MCP-1 and MIP-1 (Figure 4), and the cytokines TNF and IL-6 (Figure 5). Pretreatment of the lungs with dexamethasone attenuated all of these responses (Figures 3-5). Perfusion with LPS in normally ventilated lungs resulted in a similar mediator profile to that observed in the overventilated lungs (Figures 3-5). For the chemokines KC, MIP-2, MCP-1, and MIP-1 (Figures 3 and 4), as well as for IL-6 (Figure 5), both the absolute amounts released and the time course of release were very similar in overventilated and in LPS-treated lungs. Differences were noted only in the case of TNF, the release of which appeared to start later in the overventilated than in the LPS-treated lungs (Figure 5). However, the absolute amount of TNF released was similar after LPS treatment and OV. As in overventilated lungs, dexamethasone pretreatment of LPS-exposed lungs attenuated the release of all mediators. The only exception to this occurred with LPS-induced MIP-2 levels, which in accord with previous observations in the lung (26), were unaffected by dexamethasone.

View larger version (34K): [in this window] [in a new window]   Figure 3.   OV- (top) and LPS- (bottom) induced alterations in -chemokine levels in lung perfusate from BALB/C mice. Lungs were perfused for 60 min under control conditions (open squares; n = 4) before they were exposed to either overventilation (open circles; n = 5; top) or 50 µg/ml LPS (open triangles; n = 4; bottom). Dexamethasone (dexamethasone control, solid squares; n = 3) was added 45 min before exposure to either LPS (solid triangles; n = 3; bottom) or overventilation (solid circles; n = 3; top). MIP-2 (left panel ) and KC (right panel ) levels in the perfusate were assessed every 30 min. Data are mean ± SD. The MIP-2 and KC levels in LPS-treated and in overventilated lungs were significantly different from those in control lungs (p < 0.05). In overventilated lungs, dexamethasone caused a significant reduction in MIP-2 and KC levels (p < 0.01). In LPS-treated lungs, dexamethasone reduced KC levels (p < 0.01), but not MIP-2 levels.

View larger version (31K): [in this window] [in a new window]   Figure 4.   OV- (top) and LPS- (bottom) induced alterations in -chemokine perfusate levels in BALB/C mice. Lungs were perfused for 60 min under control conditions (open squares; n = 4) before they were exposed to either 50 µg/ml LPS (open triangles; n = 4; bottom) or overventilation (open circles; n = 5; top). Dexamethasone (dexamethasone control (solid squares; n = 3) was added 45 min before exposure to either LPS (solid triangles; n = 3; bottom) or overventilation (solid circles; n = 3; top). MCP-1 (left panel ) and MIP-1 (right panel ) levels in the perfusate were assessed every 30 min. Data are mean ± SD. The MCP-1 and MIP-1 levels in LPS-treated and in overventilated lungs were significantly different from those in control lungs (p < 0.01). In both overventilated and LPS-treated lungs, dexamethasone caused a significant reduction in MCP-1 and MIP-1 levels (p < 0.01).

View larger version (30K): [in this window] [in a new window]   Figure 5.   OV- (top) and LPS- (bottom) induced alterations in TNF- and IL-6 levels in lung perfusate from BALB/C mice. Lungs were perfused for 60 min under control conditions (open squares; n = 4) before they were exposed to either 50 µg/ml LPS (open triangles; n = 4; bottom) or overventilation (open circles; n = 5; top). Dexamethasone (dexamethasone control (solid squares; n = 3) was added 45 min before exposure to either LPS (solid triangles; n = 3; bottom) or overventilation (solid circles; n = 3; top). TNF- (left panel ) and IL-6 (right panel ) levels in the perfusate were assessed every 30 min. Data are mean ± SD. The TNF- and IL-6 levels in LPS-treated and in overventilated lungs were significantly different from those in control lungs (p < 0.05). In both overventilated and LPS-treated lungs, dexamethasone caused a significant reduction in TNF- (p < 0.05) and IL-6 (p < 0.01) levels.

IL-10, IFN-, and VEGF were not detected under any condition investigated (data not shown). Eotaxin was present in small amounts (< 10 pg/ml) in control lungs, but these levels were not increased either by OV or by LPS.

Experiments with C3H/HeJ Mice

Since the responses observed in overventilated and LPS-treated lungs appeared so similar, we wanted to exclude the possibility that OV-induced responses were caused by LPS contamination introduced by the vigorous ventilation under these conditions. Therefore, we examined the effects of LPS and OV on NF-B translocation in C3H/HeJ mice, which are known to be resistant to LPS (15). We found that in C3H/HeJ mice, OV activated NF-B, whereas LPS did not (Figure 6). In order to show that this difference was also true for mediator release, we measured MIP-2 perfusate levels in C3H/HeJ mice exposed to overventilation or endotoxin in a separate set of experiments. In these experiments, OV caused an increase in MIP-2 perfusate levels that was absent in LPS-perfused lungs (Figure 8).

View larger version (23K): [in this window] [in a new window]   Figure 6.   NF-B-translocation by OV but not by LPS in lungs from C3H/HeJ mice. All lungs were perfused for 120 min. After 60 min of baseline perfusion, lungs were exposed for another 60 min to either OV or 50 µg/ml LPS. NF-B translocation was determined by EMSA.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ventilation with a reduced VT is the first and so far the only therapeutic approach that has been shown to be effective in attenuating mortality among patients with ARDS (3). This is a major breakthrough for the treatment of ARDS patients. Further progress in this field will depend on understanding the mechanistic basis of how ventilation can aggravate or ameliorate lung injury and possibly also systemic inflammation. The data presented here suggest that OV may represent a stimulus for the immune system similar to that elicited by bacterial LPS. This notion is supported by our finding that the effects of OV and LPS on NF-B activation, chemokine release, and cytokine release were nearly indistinguishable. These findings suggest that by causing biotrauma, OV may be as dangerous to patients as are bacterial infections. In addition, recent findings suggest that under certain conditions cytokines may promote the growth of bacteria (27). Together with the present findings, this opens the possibility that inappropriate ventilation strategies may contribute to ventilator-associated pneumonia.

To our knowledge, a detailed analysis of ventilation-induced NF-B activation, chemokine release, and the effects of steroid treatment on ventilation-induced inflammatory responses in whole lungs has not previously been reported. The amounts of all mediators released by OV were similar to those found after perfusion with high LPS concentrations (50 µg/ml), indicating that these amounts may be biologically relevant. The present findings suggest that one of the initial steps after OV is activation of the transcription factor NF-B. All of the genes of the cytokines and chemokines detected in response to OV contain the NF-B consensus sequence, whereas the genes of those mediators that were not increased (IL-10, IFN-, VEGF, eotaxin) lack this sequence. Thus, the mediator profile observed is consistent with NF-B playing an important role in the signal transduction elicited by OV. Further support for an important role of NF-B in OV-induced mediator release comes from the finding that pretreatment with steroids blocked both NF-B activation and mediator release. The mechanism by which OV activates NF-B is unknown, although the experiments with the C3H/HeJ mice in our study demonstrated that this mechanism is unrelated to Tlr-4, which is mutated in these mice (15) and which contributes importantly to LPS- induced signal transduction (16, 17). An attractive hypothesis for the way in which OV stimulates cells is suggested by the finding that stretching of lung cells in culture resulted in NF-B activation and release of IL-8 (KC is considered the murine analogue of IL-8) (12, 28, 29). Although in cell cultures, stretching per se did not cause the release of other chemokines or cytokines, such as TNF or IL-6, this still opens the possibility that overstretching of lung units is the initial signal for the release of some chemokines or cytokines. In accord with this, it has been shown that stretching of cells activates potassium channels (30), and that the nonspecific potassium channel blocker quinine prevents ventilation-induced as well as LPS-induced cytokine release in perfused mouse lungs (31). However, other mechanisms than stretching, such as repeated collapse and reopening of alveolar units, cannot be excluded as promoting cytokine or chemokine release.

Regardless of the exact mechanism that finally leads to activation of NF-B, it appears evident that OV causes activation of the immune system. The evidence for this includes the findings in the present study as well as the experimental and clinical studies discussed earlier. Further indirect evidence is provided by studies done with lung-lavaged rabbits in vivo, a model that was introduced to study the application of drugs for ARDS (32). Pulmonary injury in this model can be attenuated not only by surfactant, but also by protective ventilation strategies (33). High-frequency ventilation, as compared with conventional mechanical ventilation (CMV), prevented pulmonary inflammation and lung damage in this model (34), suggesting that in already injured lungs, CMV is injurious by activating the immune system. Furthermore, it was found that even during CMV in this model, animals were largely protected if they were made neutropenic (37) or if neutrophil adhesion was prevented by leumedins (38). These findings suggest that ventilation in preinjured lungs causes neutrophil influx and neutrophil-dependent injury in the lungs. Clearly, increased levels of chemokines would play an important role in this process. Thus, OV appears to be capable of inducing a complete inflammatory response by stimulating the production of cytokines and chemokines that in turn attract and activate neutrophils. However, the finding in the present study that OV caused release not only of -chemokines (which are primarily responsible for recruitment of polymorphonuclear neutrophils), but also of -chemokines, suggests that during OV, monocytes and lymphocytes may also be recruited to the lungs.

The model of the isolated lung used in the present study allows investigation of the effects of OV in the absence of confounding factors present in vivo, such as chest-wall mechanics, blood-derived leukocytes, or regulation by the nervous system. Thus, this model allows investigation of the response of the "pure" lung to ventilation. In addition, the signals generated in this model may be greater than those generated in vivo, because in lungs perfused in a nonrecirculating manner, mediators are probably not metabolized. We have previously shown in this model that OV causes release of TNF, IL-6, and prostacyclin in a dose-dependent manner (7). Those findings provided the first direct evidence that ventilation may be a trigger for the release of immune mediators. However, subsequent studies with healthy rats (33) and human subjects (39) demonstrated that OV does not as easily cause cytokine release in vivo in homogeneously injured lungs. This finding, which is certainly reassuring for maintenance of the normal state, suggests that in vivo factors, such as limitation of lung extension by the chest wall, prevent undue overstretching of the lungs and thus prevent activation of the immune system. In this regard, perfused lungs therefore represent a model of inhomogeneously injured lungs, which are typical in ARDS patients (40). The healthy parts of such inhomogeneously injured lungs can extend to a greater degree than is normally the case, owing to the lack of counterpressure by the atelectatic/ injured parts. This concept is supported by both experimental and clinical findings. Clinically, it was demonstrated in ARDS patients that ventilation with greater as opposed to smaller volumes increased cytokine release and pulmonary inflammation (3, 4). Experimentally, it was shown in vivo that ventilation strategies affect cytokine release only in preinjured and not in healthy lungs (41). However, it should be noted that the degree of OV that can be studied in vivo is limited by the pronounced accompanying decrease in blood pressure and hemodynamics. For instance, the highest VT studied by Chiumello and colleagues (41) in rats in vivo was 16 ml/kg, which is much less than the volumes that have been applied in isolated lungs (e.g., 40 ml/kg [9] or 32 ml/kg [present study]). Presumably, if these volumes were reached in healthy animals in vivo (i.e., at similar transpulmonary pressures), a similar cytokine response to that seen in the isolated lungs would be observed.

In summary, we have shown that OV triggers activation of NF-B and elicits release of -chemokines, -chemokines, and cytokines from perfused lungs in a qualitatively and quantitatively similar manner to that of LPS (except in C3H/HeJ mice). Since in the case of LPS these mediators are known to contribute to LPS-induced inflammation, comparable consequences may be assumed in the case of OV. Since dexamethasone interferes with this process by blocking NF-B activation, steroid treatment (perhaps by inhalation) might be considered as a subsidiary treatment during chronic mechanical ventilation.