INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Nitric oxide (NO) plays a pivotal role in the regulation of local blood flow and vasomotor tone in both the systemic and the pulmonary circulation (1). NO is synthesized by different isoforms of nitric oxide synthase (NOS) by oxidative deamination of the amino acid L-arginine (1). A constitutive isoform (cNOS), which is calcium/calmodulin-regulated, has been originally described in endothelial cells and may at least in part contribute to the maintenance of low pulmonary vascular resistance under physiologic conditions (3). A calcium-independent, inducible isoform (iNOS) can be activated by cytokines interleukin-1, interferon , tumor necrosis factor-, (4, 5) and bacterial endotoxin (5, 6), leading to NO overproduction and concomitantly inappropiate vasodilation, which is an important feature of bacterial sepsis (7, 8).

In addition, liberation of NO can be stimulated by bacterial exotoxins; recently, it was noted that the pore-forming hemolysin of Escherichia coli (HlyA) is a potent activator of NO liberation in isolated pulmonary endothelial cells (9). HlyA also induces acute pulmonary vascular abnormalities in isolated lungs, causing thromboxane generation and related pulmonary hypertension (10). Against this background, we investigated the influence of HlyA on pulmonary NO synthesis at the intact organ level. We employed buffer-perfused rabbit lungs, as this model allows continuous monitoring of pulmonary NO release in the absence of plasma proteins and circulating red blood cells (11).

In particular, the aim of our study was first to demonstrate that HlyA challenge induces a significant pulmonary release of NO in both the alveolar and the intravascular compartment, which occurs independently from the HlyA-induced vasoconstriction. Second, we investigated whether the NO precursor L-arginine mitigates the HlyA-induced pressor response by enhancing HlyA-stimulated NO release. Third, we examined different NOS inhibitors in this septic lung model for possible detrimental effects by suppression of NO release and potentiation of the HlyA-induced increase of pulmonary arterial pressure.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Reagents

Aminoguanidine was obtained from Sigma (Munich, Germany). N^G-monomethyl-L-arginine (L-NMMA) and 2-(2-aminoethyl)-2-thiopseudourea-dihydrobromide (AETU) were obtained from Calbiochem (Bad Soden, Germany). L-arginine hydrochloride (1 mol/L) was purchased from Pfrimmer, Kabi Pharmacia (Erlangen, Germany). BM 13.505 {Daltroban, 4-[2-(4-chlorbenzolsulfonylamino)-ethyl]-phenyl acetic acid} was a generous gift from Boehringer AG (Mannheim, Germany). All other chemicals were purchased from Merck (Darmstadt, Germany). E. coli hemolysin was prepared as described previously (12) and generously provided by S. Bhakdi (University Mainz, Mainz, Germany). HlyA doses were expressed in hemolytic units (HU) per milliliter.

Lung Model

The model of perfused rabbit lungs has previously been described in detail (overview in reference 13). Briefly, rabbits of either sex weighing 2.6 to 2.9 kg were anticoagulated with heparin and deeply anesthetized with a mixture of ketamine and xylazine. Tracheostomy was performed, and the animals were ventilated with room air, using a Harvard respirator (Hugo Sachs Elektronik, Germany) (tidal volume, 30 ml; frequency, 30/min; positive end-expiratory pressure, 1 cm H₂O). After midsternal thoracotomy, catheters were placed into the pulmonary artery and the left atrium, and perfusion with Krebs-Henseleit buffer was started. The buffer contained 120 mM NaCl, 4.3 mM KCl, 1.1 mM KH₂PO₄, 24 mM NaHCO₃, 2.4 mM CaCl₂ and 1.3 mM MgPO₄, as well as 2.4 g/L glucose. In parallel with the onset of artificial perfusion, ventilation was changed from room air to a gas mixture of 16% O₂, 5% CO₂ and 79% N₂, supplied by a gas-mixing chamber (Witt, Witten, Germany). After extensive rinsing of the lung vasculature, the lungs were recirculatingly perfused with a pulsatile flow of 100 ml/min (total volume, 150 ml); left atrial pressure was set at 2 mm Hg (referenced at the hilus). Lungs were suspended from a force transducer, and the whole system was equilibrated at 37° C. Lungs included in the study (1) had a homogenous white appearance with no signs of hemostasis, edema, or atelectasis; (2) had pulmonary atery (Ppa) and ventilation pressures in the normal range; (3) were isogravimetric during an initial steady-state period of at least 30 min.

Detection of Nitric Oxide

NO release into both the alveolar and the intravascular compartments was monitored online as described in a previous study (11). Briefly, an aliquot of the mixed expired gas (160 ml/min) was continuously sampled at the ventilator exhaust valve and transferred to a chemiluminescence NO analyzer (UPK 3100; UPK, Bad Nauheim, Germany). The detection limit of NO in gas was 1 ppb (parts per billion, vol/vol). Daily calibration was performed with certified gases (NO in oxygen-free nitrogen; Messer Griesheim, Herborn, Germany). For monitoring of buffer fluid NO and NO metabolites (NO, nitrite, nitrate and peroxynitrite, all summarized as NO_x), a small aliquot of the lung effluent (600 µl/min) was continuously transferred into a reaction vessel containing 80 ml of 0.1 M vanadium (III) chloride in 2 M HCl at 98° C while the total recirculating volume was kept constant by continuous addition of fresh buffer fluid. This vanadium solution quantitatively reduces the NO decomposition products back to NO. Arising NO was carried by oxygen-free nitrogen continuously flushed through the device (160 ml/min), which after passage of a liquid trap and an acidic vapor trap entered a second chemiluminescence detector. Calibration was performed with buffer fluids containing known concentrations of nitrite and nitrate. As the detection of exhaled NO was altered by massive lung edema formation, experiments were terminated when lung weight gain exceeded 12 g.

Analysis of NO_x Data

All NO and NO_x data were registered by personal computers at a rate of 1 Hz. As the total volume of recirculating buffer fluid was continuously diluted by replacing the aliquot used for NO detection, the data obtained for accumulated NO metabolites in the perfusate were corrected by an algorithm programmed in Turbo Pascal (Borland International, Munich, Germany) after termination of the experiment. These corrected data of accumulating NO products were differentiated to obtain the actual rate of NO production, which was expressed as µmol/ 3 min.

Experimental Protocol

After a steady-state period of 30 min, the perfusate was replaced by fresh buffer fluid, and all lungs were perfused for another 30-min interval before being challenged with HlyA. In parallel, measurement of intravascular and airway NO release was started. The following experimental groups were formed.

Group I. Lungs were exposed to low doses of HlyA (0.037 HU/ml) in the absence and presence of the NOS substrate L-arginine (100 µM).

Group II. Lungs were exposed to high doses of HlyA (0.183 HU/ ml) in the absence and presence of the NOS substrate L-arginine (100 µM).

Group III. HlyA challenge (0.22 U/ml) in the presence of the thromboxane-receptor antagonist BM 13.505 (5 µM) and 100 µM L-arginine.

Group IV. HlyA challenge (0.037 HU/ml) in the presence of the NOS inhibitor N^G-monomethyl-L-arginine (L-NMMA, 400 µM) and 100 µM L-arginine.

Group V. HlyA challenge (0.037 HU/ml) in the presence of the iNOS-specific inhibitor aminoguanidine (100 µM); experiments performed with and without L-arginine.

Group VI. HlyA challenge (0.037 HU/ml) in the presence of the iNOS-specific inhibitor AETU (in concentrations of 5 and 50 µM) and 100 µM L-arginine.

In Groups III to VI, each agent was admixed to the buffer fluid 20 min before HlyA challenge.

Data Analysis

Data were expressed as mean ± SEM. Differences were analyzed using the Wilcoxon-Mann-Whitney Test; p values less than 0.05 were considered to represent a significant difference.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Baseline Conditions

Under baseline conditions, the Ppa ranged between 6 and 10 mm Hg in all experiments. Progressive intravascular NO_x accumulation was observed as previously described (11). The production rate ranged between 0.15 and 0.24 µmol/L/3 min. In parallel, continuous exhalation of NO was noted, ranging between 33 and 62 ppb in the expired air. Administration of L-arginine (100 µM) enhanced both the intravascular and the exhalative liberation of NO significantly (Figures 1 and 2), whereas the baseline Ppa was not significantly affected (data not shown).

View larger version (19K): [in this window] [in a new window]   Figure 1.   Effects of low-dose HlyA challenge in absence and presence of L-arginine (L-Arg) on pulmonary arterial pressure (PAP), NO exhalation, intravascular release of NOx, and lung weight. HlyA (0.037 HU/ml) was administered at time zero. For pulmonary artery pressure and lung weight, the change from baseline as assessed at time zero is given. Data are given as mean ± SEM; error bars are missing when falling into symbols. Significant differences are indicated by the horizontal bars.

View larger version (20K): [in this window] [in a new window]   Figure 2.   Effects of high-dose HlyA challenge in the absence and presence of L-arginine (L-Arg) on pulmonary arterial pressure (PAP), NO exhalation, intravascular release of NOx, and lung weight. HlyA (0.183 HU/ml) was administered at time zero. For pulmonary artery pressure and lung weight, the change from baseline as assessed at time zero is given. Because of marked weight gain, experiments in the absence of L-Arg were terminated after 10 min. Data are given as mean ± SEM; error bars are missing when falling into symbols. Significant differences are indicated by the horizontal bars.

HlyA Challenge

Intravascular application of HlyA induced an immediate, dose-dependent increase of NO release in the alveolar as well as the intravascular compartments (Figures 1 and 2). In parallel, HlyA challenge induced an increase in the Ppa. This pressor response was followed by marked pulmonary edema formation upon use of the higher HlyA dose (0.183 HU/ml). In the presence of L-arginine, the HlyA-induced intravascular and exhalative NO release was further enhanced. Moreover, as evident from use of the higher exotoxin dose, coapplication of L-arginine slightly limited the HlyA-elicited pressor response and the concomitant edema formation.

Thromboxane Blockade

In the presence of the thromboxane receptor antagonist BM 13.505, the HlyA-induced pressor response was markedly reduced. Intravascular and exhalative NO liberation was, however, not affected, suggesting independency from vascular tone and shear stress (Figure 3). BM 13.505 did not affect baseline vascular tone or NO production in the absence of HlyA (data not given in detail).

View larger version (18K): [in this window] [in a new window]   Figure 3.   Influence of the thromboxane-receptor antagonist BM 13.505 on HlyA-induced pressor response and NO liberation. Lungs were either pretreated with BM 13.505 (5 µM) or received solvent only. Twenty minutes later, challenge with 0.22 HU/ml HlyA was performed (time zero); 100 µM L-arginine was present in all experiments. Relative data of pulmonary artery pressure, NO exhalation, and intravascular NOx liberation, as measured after HlyA application, are presented as mean ± SEM; error bars are missing when falling into symbols. Significant differences are indicated by the horizontal bar.

NOS Inhibition

Preapplication of the nonspecific NOS inhibitor L-NMMA markedly suppressed the baseline intravascular and exhalative NO liberation (Figure 4), whereas the pulmonary arterial pressure was not significantly affected. Subsequent HlyA challenge then induced only minute NO release in both the alveolar and the intravascular compartments. Under these conditions, the HlyA-induced pressor response was amplified manifold.

View larger version (23K): [in this window] [in a new window]   Figure 4.   Influence of the NOS inhibitor L-NMMA on HlyA-induced pulmonary hypertension and NO release. L-NMMA was preadministered at a dose of 400 µM, and HlyA challenge was performed at a dose of 0.037 HU/ml. For pulmonary artery pressure (PAP) and lung weight, the change from baseline as assessed at time zero is given. Data are presented as mean ± SEM; error bars are missing when falling into symbols. Significant differences are indicated by the horizontal bars.

Preapplication of aminoguanidine (100 µm) did not affect the baseline intravascular NO liberation and the Ppa, whereas the exhalative NO release displayed some minor decrease (Figure 5). Subsequent HlyA challenge provoked the release of NO into both the alveolar and the intravascular compartment both in the absence and presence of aminoguanidine pretreatment. Peak values were virtually identical in both groups, but there was some decay of NO exhalation in the further course of experiments in the aminoguanidine-pretreated lungs. In parallel, the HlyA-induced pressor response in the aminoguanidine group was slightly enhanced. In additional experiments, the effect of 100 µM aminoguanidine was investigated in the presence of 100 µM L-arginine. In these studies, no influence of aminoguanidine on Ppa and NO release both under baseline conditions and in response to HlyA was noted as compared with control values in the absence of aminoguanidine (data not given in detail).

View larger version (21K): [in this window] [in a new window]   Figure 5.   Influence of the iNOS inhibitor aminoguanidine on HlyA-induced pulmonary hypertension and NO release. Aminoguanidine was preadministered at a dose of 100 µM, and HlyA challenge was performed at a dose of 0.037 HU/ml. For pulmonary artery pressure (PAP) and lung weight, the change from baseline as assessed at time zero is given. Data are presented as mean ± SEM; error bars are missing when falling into symbols. Significant differences are indicated by the horizontal bar.

At a dose of 5 µM, AETU did not exert significant effects on baseline and HlyA-elicited NO release or the exotoxin-provoked pressor response (Figure 6). At 50 µM, AETU moderately reduced the baseline and the HlyA-provoked intravascular NO release as well as NO exhalation, and the HlyA-elicited increase in Ppa was slightly increased.

View larger version (24K): [in this window] [in a new window]   Figure 6.   Influence of the iNOS inhibitor 2-(2-aminoethyl)-2-thiopseudourea-dihydrobromide (AETU) on HlyA-induced pulmonary hypertension and NO release. AETU was preadministered at doses of 5 µM and 50 µM, and HlyA challenge was performed at a dose of 0.037 HU/ml. For pulmonary artery pressure (PAP) and lung weight, the change from baseline as assessed at time zero is given. Data are presented as mean ± SEM; error bars are missing when falling into symbols. Significant differences of 50 µM AETU as compared with absence of AETU are indicated by the horizontal bars.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In buffer-perfused rabbit lungs, low doses of the proteinaceous bacterial exotoxin HlyA markedly increased the baseline NO liberation into both the intravascular and the alveolar compartments. The rapidity of the mediator response and the inhibitory profiles of different blockers of NO synthases suggested that the signal transduction proceeded via activation of cNOS, directly related to exotoxin activity and not to secondary changes in shear stress. Manipulation of the HlyA-evoked NO biosynthesis demonstrated that the NO release reaction mitigated the vasoconstrictor response induced in parallel by this bacterial exotoxin.

As previously described for isolated rabbit lungs, perfused in the absence of any inflammatory agent, continuous biosynthesis of NO with release into both the alveolar and the intravascular space was noted (11, 14). Interestingly, the liberation of this short-lived vasodilator into both compartments was increased by approximately 20 to 30% upon admixture of the precursor amino acid L-arginine to the recirculating buffer fluid. This finding corresponds to the observation that oral administration of L-arginine with concomitant augmentation of the plasma levels of this semiessential amino acid increases the concentration of NO in exhaled air of normal human subjects (15). Substrate availability thus appears to contribute to the regulation of NO biosynthesis in the pulmonary vasculature and/or parenchyma, and the chosen dose of 100 µM fairly corresponds to in vivo plasma levels in both rabbits and humans (16, 17). Baseline NO synthesis has been suggested to contribute to the maintenance of overall low pulmonary vascular resistance under physiologic conditions (2). Increase of NO biosynthesis upon L-arginine supply and complete inhibition of NO formation by L-NMMA did not however, significantly affect the baseline pulmonary vascular tone in the current study. This observation may be explained by the fact that under control conditions, in the absence of inflammatory agents, additionally formed vasodilatory agents such as prostaglandins I₂ and E₂ and endothelium-derived hyperpolarizing factor may suffice to keep the pulmonary vascular resistance at its overall low physiologic level.

Conditions may fundamentally change during lung inflammatory events. In the present study we chose a model of septic lung injury by employment of the proteinaceous exotoxin E. coli hemolysin. Such exotoxins, produced by gram-negative and gram-positive bacteria, have received increasing attention as potentially important instigators of systemic inflammation and multiorgan dysfunction. The pathogenetic relevance of HlyA as a well characterized representative of these bacterial virulence factors has been established in animal models as well as in human infection (12). As previously noted, HlyA provoked marked pulmonary vasoconstriction in rabbit lungs, and preceding studies have documented that this vasoconstrictor response is largely due to enhanced thromboxane formation in response to the exotoxin (10, 18, 19). In addition, we now describe how HlyA evokes an impressive increase in the biosynthesis of nitric oxide. NO was immediately liberated into both the alveolar and the intravascular compartments upon challenge with HlyA. The response was dose-dependent and further amplified by the presence of L-arginine in the buffer fluid, and the maximal NO release detected under these conditions of "septic lung" surpassed the baseline levels of both intravascular and alveolar NO liberation.

In what way are HlyA challenge and vasoconstrictor response linked to the increase in NO biosynthesis? At first glance, the augmentation of Ppa and the related increase in shear stress might be responsible for the enhanced NO formation. Shear-stress dependency of NO release has been demonstrated for the coronary vascular bed (20), cultured macrovascular endothelial cells (21), aortic segments (22), and femoral arteries (23), but the relevance of this mechanism for the pulmonary circulation has not been fully elucidated. Indeed, the current results do not support this view. First, the NO release occurred in parallel to the onset of a rise in Ppa and reached a maximal level clearly before the maximal pressor response. Second, limitation of the Ppa elevation by a thromboxane antagonist did not affect the NO response to HlyA. And third, "mimicry" of the vasoconstrictor response by direct administration of the stable thromboxane analogue U46619, with provocation of Ppa levels surpassing 40 mm Hg, did not elicit any increase in exhalative or intravascular NO release (H. Schütte, unpublished data). Thus, the increase in lung NO biosynthesis in response to HlyA may not be attributed to nonspecific provocation by enhanced shear stress.

In previous studies from our group in porcine endothelial cells (9), significant release of NO by HlyA was demonstrated. In these cells, as well as in human neutrophils, HlyA was identified as a most potent inductor of the preformed phosphatidylinositol hydrolysis-related signal transduction pathway (24), with rapid (within 1 min) formation of inositol phosphate and diacylglycerol. This corresponds to the rapid release of NO in the present investigation. NO is also known to be synthesized in response to a number of ligands linked to phosphoinositide metabolism via receptor occupancy (for review see references 1 and 25), and inositol phosphate-triggered Ca²⁺ release is assumed to represent the link to subsequent activation of a Ca²⁺/ calmodulin-dependent NO synthase, as demonstrated for NOS activation by endothelin-3 (26), acetylcholine (27), and bradykinin (28). Thus, there is good evidence to assume that the HlyA-elicited signal transduction events directly give way to both the generation of vasoconstrictor agents such as thromboxane and the formation of the vasodilatory agent nitric oxide.

Previous studies addressing NO generation in septic lung models mostly employed lipopolysaccharides of gram-negative bacteria (endotoxin) as causative bacterial agents. In response to endotoxin challenge, enhanced NO synthesis was indeed noted, but the response occurred only after an initial delay time of at least 1 h and was shown to proceed via increased expression of iNOS and related bioactivity. In contrast, the exotoxin HlyA provoked an immediate increase in NO synthesis, which suggests a predominant contribution of cNOS instead of iNOS to this metabolic response. This is indeed supported by the experiments with different inhibitors. L-NMMA, which inhibits both the constitutive and inducible NO synthase (29), suppressed the basal as well as the HlyA-elicited release of NO. In contrast, the iNOS-selective inhibitors aminoguanidine (29, 30) and AETU (31) did not affect baseline or HlyA-elicited NO synthesis when employed in doses previously reported to be effective and approximately equipotent to the L-NMMA concentration currently used (100 µM aminoguanidine in the presence of L-arginine; 5 µM AETU) (30- 32). Studying these drugs in the absence of L-arginine (aminoguanidine) and using higher dosage (50 µM AETU), some moderate decrease in exotoxin-elicited NO release, but also of the baseline generation of this mediator was noted, thus indicating a slight inhibition of the constitutive NOS under these conditions. The efficiency of both inhibitors to suppress iNOS in the rabbit species was confirmed in separate experiments, demonstrating that 50 µM AETU as well as 100 µM aminoguanidine blocked LPS-induced iNOS-activation in rabbit lung parenchymal cells in culture (data not given in detail).

The HlyA-elicited NO generation is apparently operative to mitigate the exotoxin-evoked pressor response and concomitant edema formation. Both abnormalities were moderately reduced upon coapplication of L-arginine to enhance the NO synthesis, and they were dramatically amplified when blocking the NO generation by L-NMMA. This finding is reminiscent of the previous observation that lung NO formation limits the vasoconstrictor response to alveolar hypoxia (14, 33, 34). Thus, the fact that HlyA obtains access to signal transduction pathways leading to enhanced NO formation may be looked upon as "beneficial," as it limits the disadvantageous vasoconstrictor response and related capillary fluid filtration upon challenge with this bacterial agent. However, it should be kept in mind that the HlyA-elicited vasomotor events are accompanied by severe ventilation-perfusion mismatch in rabbit lungs (19), and NO has been implicated in the matching of ventilation and perfusion distribution (14, 33, 34). An "anarchic" formation of NO, i.e., triggered by an exotoxin rather than being physiologically regulated, may well contribute to inadequate vasodilation and thus ventilation-perfusion mismatch on a local level. The clarification of this issue will demand detailed analysis of the gas exchange conditions when manipulating NO formation in this model of septic lung.

In conclusion, the present study has demonstrated that a bacterial exotoxin, E. coli hemolysin, obtains access to signal transduction pathways resulting in cNOS-related immediate enhancement of NO synthesis. This feature within the response pattern to the exotoxin mitigates the dramatic vasoconstrictor response imposed by only low doses of HlyA. Nonselective inhibition of NO synthesis, as suggested for the treatment of systemic vasodilatory collapse in severe sepsis, must thus be assumed to worsen the pulmonary hypertension in case of involvement of proteinaceous exotoxins in the pathogenetic sequence. This reasoning is in line with the observation of an increase in pulmonary vascular resistance in humans with septic shock upon treatment with L-NMMA (35). Against the background of the present study, the employment of an iNOS-specific inhibitor might be advantageous to maintain the regulatory function of lung cNOS-mediated nitric oxide formation. This will, however, not allow one to discriminate between physiologic and "inflammatory" (directly exotoxin-provoked) cNOS activation as would be desirable for optimizing perfusion destribution and, thus, gas exchange.