INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: lung ischemia-reperfusion; pulmonary edema; endogenous nitric oxide

Endogenous nitric oxide (NO) plays an important role in vascular homeostasis, including regulation of vasomotor tone and maintenance of microvascular barrier properties (1, 2). In ischemia/reperfusion (I/R) injury, the role of NO is controversial. On the one hand, an impaired production of baseline NO under conditions of I/R is assumed to contribute significantly to endothelial dysfunction. This view is supported by several studies of I/R demonstrating protective effects of endogenous NO in settings including ischemic-reperfused dog, rabbit, and guinea pig hearts (3); rat liver (6); gastric mucosa (7); and feline small intestine (8). However, inhibition of biosynthesis of endogenous NO has also been shown to be beneficial, as in ischemic-reperfused mouse hearts (9), rat proximal tubules (10), rabbit skeletal muscles (11), and cat brains (12). In discussion of these findings, NO was suggested to exert adverse effects by reacting with superoxide to form the strong oxidant peroxynitrite (13). A further possibility is even a lack of effect of endogenous NO in I/R injury, as suggested by a study of ischemic-reperfused rat kidneys (14).

The role of endogenous NO in pulmonary I/R is also not well defined. In I/R injury following clinical lung transplantation, with respiratory failure caused by noncardiogenic, high-permeability edema and increased pulmonary vascular resistance, the exogenous supply of NO by inhalation of NO has been recommended (15), on the assumption of impaired endogenous synthesis of this vasoactive agent in the reperfused lungs (16). This view is further supported by several experimental studies showing that inhalation of NO upon reperfusion attenuates microvascular leakage (17), an effect that is related to the nonvasodilatory, antiinflammatory properties of this agent (21), and which depends on release of cyclic guanosine monophosphate (cGMP) (18). However, in one model addressing NO inhalation in I/R injury, NO was noted to be ineffective, whereas a cGMP analogue reduced the reperfusion injury (16, 22), thus pointing to possibly disadvantageous (peroxynitrite-related?) effects of NO in addition to its stimulation of soluble guanylate cyclase in the pulmonary circulation. In blood-perfused rat lungs, in contrast, endogenously produced NO was found to be protective against I/R injury (23). Moreover, there is little knowledge about the role of endogenous NO during the period of ischemia. However, the use of NO donor substances or cGMP analogues was found to be beneficial in organ preservation, which may at least in part be ascribed to support of the NO/cGMP axis during the period of lung ischemia (24, 25).

In the current study, we investigated the release of endogenous NO during the period of ischemia as well as upon reperfusion in isolated rabbit lungs, measuring both NO liberation into the air spaces and intravascular accumulation of NO degradation products. Lungs were exposed to both normoxic and anoxic ventilation during ischemia, since lung ischemia does not necessarily result in anoxia of the lung when ventilation is maintained. Moreover, studies were performed under conditions of vascular distension during ischemia, which is known to be protective against I/R injury, and conditions of vascular collapse (26, 27). An NO synthase (NOS) inhibitor and exogenously supplied superoxide dismutase (SOD) were used as pharmacologic tools. We provide evidence that: (1) pulmonary NO synthesis is not impaired after I/R regardless of vascular distension or collapse during ischemia; (2) endogenous NO is continuously synthesized during normoxic ischemia, and does not play a major role in the induction or mitigation of lung injury in this type of I/R challenge; and (3) endogenous NO synthesis is suppressed during anoxic ischemia, reappearing promptly upon reperfusion, and this endogenous NO release following anoxia apparently contributes to the triggering of vascular leakage, possibly through its interaction with oxygen-radical formation. These findings may be relevant for optimum donor-lung handling prior to transplantation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The model consisting of isolated rabbit lungs has been described previously (26). Lungs were perfused with sterile Krebs-Henseleit-hydroxyethylamylopectin buffer in a recirculating system at 37.5° C (flow rate = 100 ml/min; total volume = 150 ml), and were normoxically ventilated (21% O₂, 5% CO₂, 74% N₂; tidal volume 30 ml; respiratory rate 30 breaths/min; positive end-expiratory pressure = 1 cm H₂O). Microvascular pressures were determined with the arterial and venous double-occlusion technique. Lungs were suspended from a force transducer for weight monitoring. The capillary filtration coefficient (Kfc) and vascular compliance were determined by stepwise elevation of the venous pressure (26). Lung weight gain was calculated as the weight difference before and after each hydrostatic challenge maneuver.

Exhaled NO was detected by chemiluminescence (28). For monitoring of NO and NO metabolites (nitrite, nitrate, peroxynitrite) in buffer fluid, summarized as NO_x, an aliquot of lung effluent was continuously reduced to gaseous NO by vanadium chloride exposure, with the resulting NO measured by chemiluminescence (28). During ischemia, this procedure was temporarily suspended, and the intravascular release of NO_x was calculated from the difference between preischemic and postischemic intravascular NO_x concentrations. Data obtained for accumulated NO metabolites in the perfusate were corrected with an appropriate algorithm, since the recirculating perfusate was continuously diluted by replacing the aliquot used for NO detection with fresh buffer fluid. These corrected data were differentiated to obtain the actual rate of NO production.

After a control hydrostatic challenge, ischemia was initiated by stopping the perfusion (t = 0 min). During ischemia, lungs were continuously ventilated, and the arterial and venous catheters were clamped for maintenance of a positive intravascular pressure, which was initially adjusted to 6 mm Hg in all but one experimental group (see the following discussion). At the end of ischemia, perfusion was reestablished by gradually increasing flow over a period of 3 min. Hydrostatic challenges were performed at 30, 60, and 90 min after the beginning of reperfusion. Double-occlusion maneuvers were undertaken as indicated in Table 1. Lungs were treated as follows:

FI_O2 = 0.21 (n = 6): Lungs were exposed to 210 min of ischemia with normoxic ventilation.

FI_O2 = 0.21 + N^G-monomethyl-L-arginine L-NMMA (n = 5): L-NMMA (400 µM; Calbiochem, Bad Soden, Germany) was admixed with the perfusate 5 min before a 210-min period of ischemia with normoxic ventilation.

FI_O2 = 0 (n = 7): Lungs were exposed to 210 min of ischemia with anoxic ventilation (5% CO₂, 95% N₂). Ventilation was switched to normoxic conditions immediately before reperfusion in these and all other experiments involving ischemia with anoxic ventilation.

FI_O2 = 0 + L-NMMA (n = 7): L-NMMA (400 µM) was administered 5 min before 210 min of ischemia with anoxic ventilation.

FI_O2 = 0 + SOD (n = 6): After 210 min of ischemia with anoxic ventilation, SOD (100 U/ml; Sigma, Munich, Germany) was admixed with the perfusate immediately before reperfusion.

Controls (n = 5) were continuously perfused and normoxically ventilated without interventions. Kfc and microvascular pressure were measured as in the 210-min ischemia experiments.

FI_O2 = 0 without vascular distension (n = 5): Lungs were exposed to 90 min of ischemia with anoxic ventilation. In contrast to all other ischemia experiments, the intravascular pressure during ischemia was maintained at zero.

Data were expressed as mean ± SEM. Differences were analyzed by one-way analysis of variance, followed by the post hoc Student- Newman-Keuls test. Values of p < 0.05 were considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pulmonary Arterial Pressure

In controls, values of pulmonary arterial pressure (Ppa) were constant throughout the observation period (Figure 1). Inhibition of NO formation by L-NMMA did not significantly affect the baseline pulmonary vascular tone. After 90 or 210 min of ischemia, a moderate and rapidly transient increase in Ppa was noted upon reperfusion (Figures 1 and 6), and Ppa values were then slightly higher in the L-NMMA groups than in lungs not receiving L-NMMA (Figure 1).

View larger version (22K): [in this window] [in a new window]   Figure 1.   Ppa before and after ischemia. At time zero, lungs were exposed either to ischemia with anoxic ventilation (n = 7) or ischemia with normoxic ventilation (n = 6) for 210 min. Control lungs did not undergo ischemia, but were continuously perfused (n = 5). When indicated, L-NMMA (400 µM) was admixed with the buffer fluid 5 min before the onset of either anoxic (n = 7) or normoxic (n = 5) ischemia. SOD was administered to one group of anoxic lungs upon reperfusion (100 U/ml, n = 6). Values are given as mean ± SEM; error bars are missing when falling with a symbol.

View larger version (13K): [in this window] [in a new window]   Figure 6.   Ppa, Kfc, NO concentration in the exhaled gas mixture (given as ppb), and rate of intravascular NOx (NO and NO degradation products) accumulation (given as µmol/L perfusate/10 min) in rabbit lungs undergoing 90 min of ischemia in the absence of a positive intravascular pressure during the ischemic period. At time zero, lungs were exposed to 90 min of ischemia with anoxic ventilation (n = 5). Kfc is expressed as cm3 × s1 × cm H2O1 × g1 × 104. Values are given as mean ± SEM; error bars are missing when falling within a symbol.

Microvascular Pressure and Vascular Compliance

Microvascular pressures were only marginally increased at 3 min after reperfusion in ischemic lungs (Table 1). Subsequently, pressures returned to baseline values in all lungs undergoing ischemia, independent of the maintenance or absence of a positive intravascular pressure during ischemia (Table 1). Only very minor variations between the different experimental groups were observed, indicating that any contribution of the capillary filtration pressure to the edema formation encountered upon reperfusion may largely be neglected. Vascular compliances were not significantly differerent between control lungs and the different experimental groups (data not given in detail). During the 210-min ischemic period, the intravascular pressure, initially adjusted to 6 mm Hg, slowly declined to approximately 2 to 3.5 mm Hg in all groups when measured at the end of ischemia (Table 1).

Kfc and Weight Gain

After 210 min of ischemia, anoxically ventilated ischemic lungs displayed highly elevated Kfc values as assessed during the subsequent reperfusion period (Figure 2). In parallel with this, massive edema formation was noted (Table 1), requiring experiments to be discontinued after the measurement of Kfc at 30 min after the onset of reperfusion. Lungs exposed to ischemia with normoxic ventilation displayed a moderate increase in Kfc and moderate edema formation, which was not affected by pretreatment with L-NMMA. When L-NMMA was administered before the ischemic period to anoxically ventilated lungs, both the dramatic rise in Kfc and the severe edema formation were significantly attenuated, with values approaching those in lungs with ischemia with normoxic ventilation. Anoxic ventilated lungs treated with SOD upon reperfusion displayed a similar reduction in the increase in microvascular permeability, which was obvious from both the Kfc values and the monitoring of lung edema formation.

View larger version (19K): [in this window] [in a new window]   Figure 2.   Impact of L-NMMA and SOD on postischemic vascular leakage. Kfc was determined before ischemia and at 30-min intervals after reestablishing perfusion (210 min period of normoxic [n = 6] or anoxic [n = 7] ischemia). Control lungs were continuously perfused (n = 5). When indicated, L-NMMA was admixed with the buffer solution (400 µM) at 5 min before the onset of either anoxic (n = 7) or normoxic (n = 5) ischemia. SOD was administered to one group of anoxic lungs upon reperfusion (100 U/ml, n = 6). Values are given as mean ± SEM; error bars are missing when falling with a symbol. $p < 0.05 versus control, $$$p < 0.001 versus control, & p < 0.05 versus all other ischemic groups.

Exhaled NO

In nonischemic control lungs, NO was continuously released into the air space (Figure 3). Upon initiation of ischemia, exhalation of NO was virtually completely abolished in anoxic ventilated lungs. In contrast, in normoxic ventilated lungs, NO was still continuously exhaled during ischemia, but at a moderately lower level. Pretreatment with L-NMMA inhibited NO release nearly completely. Upon reestablishment of lung perfusion, NO exhalation was completely restored in all ischemic groups not exposed to L-NMMA, at levels comparable to preischemic levels. Significant differences in exhaled NO upon reperfusion among normoxic-ischemic, anoxic-ischemic +/ SOD, and nonischemic control lungs were not observed. In lungs exposed to 90 min of ischemia with anoxic ventilation and with collapsed vasculature, release of exhaled NO was similarly completely restored upon reperfusion.

View larger version (26K): [in this window] [in a new window]   Figure 3.   Impact of oxygen during ischemia, L-NMMA-pretreatment, and SOD on exhaled NO in rabbit lungs undergoing I/R. NO was continuously monitored in the exhaled gas mixture and is given as parts per billion (ppb). At time zero, lungs were exposed to ischemia with anoxic (n = 7) or normoxic ventilation (n = 6) for 210 min. Control lungs were continuously perfused (n = 5). When indicated, L-NMMA was admixed with the buffer solution (400 µM) at 5 min before the onset of either anoxic (n = 7) or normoxic (n = 5) ischemia. SOD was administered to one group of anoxic lungs upon reperfusion (100 U/ ml, n = 6). Values are given as mean ± SEM; error bars are missing when falling within a symbol.

During each hydrostatic challenge, a transient and completely reversible increase in NO exhalation of approximately 5 ppb occurred in all experimental groups except those receiving L-NMMA pretreatment. These values are not presented in Figure 3, in order to avoid overloading with excessive data points.

Intravascular NO_x

Before the onset of ischemia, progressive accumulation of NO_x in the recirculating perfusate was observed, with production rates per unit of time comparable to previously reported values (28). For evident reasons, no data points were obtained during the ischemia period because of interruption of recirculating perfusion. In all ischemic lungs not undergoing L-NMMA pretreatment, the rate of appearance of intravascular NO_x was virtually completely restored upon reperfusion, with initially even markedly higher levels in lungs subjected to normoxic ventilation during ischemia (Figure 4). However, after 30 min of reperfusion, the appearance of intravascular NO_x was comparable to that in controls in all ischemic groups. In ischemic lungs treated with the nonspecific NOS inhibitor L-NMMA immediately before ischemia, NO_x release was suppressed upon reperfusion. In ischemic lungs with collapsed vessels during ischemia because of the absence of positive intravascular pressure, the rate of NO_x appearance in the perfusate also returned to preischemic values upon reperfusion.

View larger version (22K): [in this window] [in a new window]   Figure 4.   Impact of oxygen during ischemia, L-NMMA-pretreatment, and SOD on rate of intravascular NOx (NO and NO-degradation products) accumulation before ischemia and after reperfusion in rabbit lungs. The rate of appearance of NOx is given in µmol/L perfusate/10 min. At time zero, lungs were exposed to ischemia with anoxic (n = 7) or normoxic-ventilation (n = 6) for 210 min. Control lungs were continuously perfused (n = 5). When indicated, L-NMMA was admixed with the buffer solution (400 µM) at 5 min before the onset of either anoxic (n = 7) or normoxic (n = 5) ischemia. SOD was administered to one group of anoxic lungs upon reperfusion (100 U/ml). Values are given as mean ± SEM; error bars are missing when falling within a symbol. * p < 0.05, ** p < 0.01 versus L-NMMA groups; #  p < 0.05 versus FIO2 = 0 + L-NMMA, % p < 0.05 versus controls, FIO2 = 0, FIO2 = 0 + SOD.

When calculating the difference in perfusate NO_x concentrations between time zero and 210 min (reestablishment of perfusion in the ischemic lungs), a total of  15 µmol/L was obtained for the control lungs (Figure 5). In all lungs undergoing anoxic ischemia and those that were pretreated with L-NMMA, this difference approached zero, indicating virtual absence of the appearance of intravascular NO_x during the ischemic period. Interestingly, in normoxic ventilated lungs, a significant accumulation of NO_x ( 8 µmol/L) was calculated with this approach, showing that NO_x formation continued during the ischemic period under conditions of normoxic ventilation.

View larger version (12K): [in this window] [in a new window]   Figure 5.   Impact of oxygen and L-NMMA-pretreatment on appearance of NOx (NO and NO-degradation products) in buffer solution during ischemia. The difference in perfusate NOx concentration between 0 min (onset of ischemia) and 210 min (reestablishment of perfusion) is given as µmol/L. At time zero, lungs were exposed to ischemia with anoxic (n = 7) or normoxic ventilation (n = 6) for 210 min. Control lungs were continuously perfused (n = 5). When indicated, L-NMMA was admixed with the buffer solution (400 µM) at 5 min before the onset of either anoxic (n = 7) or normoxic (n = 5) ischemia. SOD was administered to one group of anoxic lungs upon reperfusion (100 U/ml, n = 6). Values are given as mean ± SEM; error bars are missing when falling with a symbol. && p < 0.01 versus all other ischemic groups, $$$ p < 0.001 versus controls.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the standard protocol of 210 min of warm ischemia and subsequent reperfusion of rabbit lungs used in this study, the pulmonary vascular leakage response (increase in Kfc, hydrostatic maneuver-induced edema formation) represented the predominant pathophysiologic abnormality. As known from preceding studies with this model, the severity of the leakage response was dependent on the use of anoxic or normoxic ventilation during the ischemic period (26). In addition, the pulmonary vascular leakage that occurs under the study conditions is known to be accompanied by severe abnormalities in gas exchange in the reperfused lungs, including ventilation- perfusion mismatch and shunt flow (19), which were, however, not the focus of the present study. Additionally, a transient increase in Ppa is consistently observed upon reperfusion of the lungs. The extent of this pressor response is rather limited if the perfusate flow is gradually increased upon reestablishing perfusion, as in the present study. Clearly, the edema formation observed in the reperfused lungs is independent of this elevation in Ppa, since measurements of the microvascular pressure undertaken in parallel with the Kfc measurements demonstrated values in the normal range. Since there is no reason to assume a dramatic increase in capillary filtration area in the postischemic lungs, the manifold rise in Kfc values has to be attributed to a severe increase in the capillary permeability of these organs.

All lungs were ventilated during ischemia, which has been identified as a useful strategy of "biophysical protection" in previous investigations, attenuating I/R injury independently of oxygen supply (26, 29). Moreover, ongoing ventilation enabled continuous monitoring of NO release into the air spaces even during the period of ischemia. Notably, in normoxic ventilated lungs, exhalation of NO and intravascular accumulation of NO degradation products were maintained to a major extent. This is of interest because shear stress at the endothelial surface is considered a major contributor to baseline vascular NO synthesis, and such shear stress is apparently absent upon cessation of perfusion. However, in the lung, shear forces transduced by ventilation might also represent a trigger of baseline NO synthesis, and indeed, recent experiments with perfused rabbit lungs showed that parameters of ventilation do have more impact on pulmonary NO generation than do variables of perfusion flow (H. Schütte, unpublished data). In contrast to normoxic ischemia, lung NO synthesis was virtually fully abrogated during anoxic ischemia. The current study does, however, show that the beneficial effects of normoxic ventilation on I/R injury must be ascribed to the oxygen supply itself rather than to the difference in lung NO production during the ischemic period: when mimicking the suppression of pulmonary NO synthesis during normoxic ischemia through the use of L-NMMA (at a dose inducing maximum inhibition of NO synthesis in previous experiments), we encountered no further impairment of the increase in microvascular permeability. Thus, although the characteristics of lung NO synthesis discriminate between the models of normoxic ischemia (ongoing endogenous NO synthesis) and anoxic ischemia (fully suppressed NO synthesis), this difference may not account for the more severe vascular leakage noted under conditions of anoxic ischemia.

The 210-min ischemia experiments were performed with maintenance of a positive intravascular pressure during ischemia. In a previous study (26), this maneuver had a strong potency for attenuating I/R lung injury. In accord with these earlier findings, we found that ischemic tolerance of the isolated organ was far lower in the absence of this positive intravascular pressure: under these conditions, a short (90-min) period of ischemia with anoxic ventilation sufficed to induce a massive increase in microvascular permeability, the severity of which was comparable to that encountered in the 210-min ischemia group with maintenance of intravascular pressure. However, regardless of the presence or absence of positive intravascular pressure, reperfusion restored NO synthesis to a comparable degree in both groups. Therefore, the protective effect of intravascular pressure in I/R is obviously not linked to preserved NO synthesis, necessitating the assumption of an NO-independent mechanism in this strategy of "biophysical protection." This conclusion contrasts with that of Becker and colleagues, who described a similar protective effect of positive intravascular pressure on injury in ischemic ferret lungs, which was abolished by the NOS inhibitor L-nitro arginine methyl ester (27), suggesting a significant contribution of NO to this phenomenon. However, these investigations were performed in hyperoxic ventilated, nonreperfused lungs, which may involve mechanisms of injury different from those in our study.

The role of endogenous NO in I/R injury of systemic organs is controversial: blocking of NOS in several models of I/R injury has been shown to be both beneficial (9, 30) and deleterious (3). The adverse effects of NO that occur with the reperfusion of ischemic organs may be explained by the reaction of NO with I/R-induced superoxide to form peroxynitrite, which in turn can further decompose to the cytotoxic oxidants hydroxyl radical and nitrogen dioxide (13). Superoxide formation may occur under conditions of anoxia/reoxygenation when oxygen is reintroduced into ischemic tissue (31). However, anoxia reoxygenation of systemic organs is not necessarily analogous to lung I/R, when oxygen is still supplied to pulmonary tissues by ventilation. Zhao and colleagues provided evidence for the pulmonary generation of reactive oxygen species through distinctly different pathways in I/R or anoxia/ reoxygenation (32). The data of the present study, focusing on endogenous NO, are in accord with the view that, for the reasons given in the following sections, details of the I/R protocol, and particularly the presence or absence of oxygen delivered via ventilation, have a major impact on the role of endogenous NO in the severity of the vascular injury occurring in I/R.

First, in normoxic ventilated ischemic lungs, no evidence was found for a role of endogenous NO. In particular, the beneficial effect of ventilation per se, and of the maintenance of a positive intravascular pressure, may not be linked to the ongoing endogenous synthesis of NO during the ischemic period, as discussed earlier, and complete suppression of NO synthesis by L-NMMA during both the ischemia and the reperfusion periods did not affect the severity of the leakage response. We cannot, however, exclude the possibility that this lack of effect represents the "balanced sum" of some "protective" effect of endogenous NO during the period of ischemia and some "deleterious" effect of this NO upon reperfusion (see the subsequent discussion). To dissect such a putative phase-dependent differential role of endogenous NO, inhibition of NO synthesis would have to be restricted to either the period of ischemia or the period of reperfusion.

Second, in anoxic ventilated lungs, impressive protection against the leakage response was provided by blocking lung NO synthesis with L-NMMA. This finding is in accordance with that of Huang and colleagues, who similarly noted an attenuation of lung injury by inhibiting NOS under conditions of anoxic I/R, although Huang and colleagues did not directly address vascular permeability (33). The present study shows that in a protocol of anoxic ischemia, the effect of an NOS inhibitor is apparently restricted to the reperfusion period, since endogenous NO synthesis during ischemia was fully blocked by the absence of oxygen, and was immediately restored upon reperfusion and the change to normoxic ventilation. Interestingly, a comparable protective effect was also achieved when SOD was admixed with the buffer fluid immediately before reperfusion of the lungs. These findings may therefore support the hypothesis that reappearance of NO and a burst of superoxide generation upon reintroduction of oxygen into the system give rise to the formation of peroxynitrite and subsequently generated toxic reaction products. Such a view is further supported by a study of ischemic hearts demonstrating the intravasular appearance of peroxynitrite upon reperfusion (34).

Against the background of endogenous NO having a role in lung I/R injury that depends on details of the ischemia and reperfusion protocol, as demonstrated in the present study for the absence or presence of oxygen during ischemia, it is readily conceivable that this is also true for the exogenous supply of NO. Inhalation of NO was protective in several lung I/R studies (17, 18, 20), including the current model (19). However, lack of protection (22) or worsening of edema (35) have also been reported. It is obvious that the timing (see the previous discussion above) and dosage of exogenous NO administration are critical issues in this context. In addition, the spatial distribution of inhaled NO may differ significantly from that of endogenous NO, producing different efficacy profiles. This makes all the more needed a detailed analysis of the intrapulmonary NO-related reaction sequences that occur when exogenous NO is introduced as an additional agent into this system.

In conclusion, the present study demonstrates that endogenous formation of NO is preserved in lungs undergoing normoxic ischemia, but is entirely lost under conditions of anoxic ischemia. No evidence was, however, obtained that the ongoing synthesis of NO during normoxic ischemia might exert protective effects against the I/R-related leakage response. Moreover, the present study does not support the view that the effects of "biophysical" lung protection (e.g., from ongoing ventilatory movements or vascular distension during ischemia) might be related to enhanced endogenous synthesis of NO. In contrast, restoration of endogenous NO synthesis upon reperfusion of anoxic ischemic lungs obviously contributes to the severity of the vascular leakage response, since the latter is markedly attenuated by the presence of L-NMMA in this type of protocol. That exogenous SOD, introduced into the system immediately before reperfusion, is similarly protective points to a putative role of peroxynitrite formation resulting from the reappearance of NO, and to a burst of superoxide synthesis upon reperfusion.