INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Nitric oxide (NO) is an important mediator of a variety of physiologic and pathologic conditions. Its formation from one of the terminal guanidine nitrogen groups of the amino acid L-arginine is catalyzed by at least three distinct isoforms of the enzyme NO synthase (NOS). In endothelial and neuronal cells, the NOS is expressed constitutively. Under physiologic conditions, activation of the constitutive NOS in endothelial cells by shear stress results in the continuous release of NO, which induces vasodilation.

NO produced by the constitutive NOS is regarded as a key regulator of basal pulmonary vascular tone in both animals (1, 2) and humans (3). Production of the constitutive NOS, however, seems to become blocked at an early stage during states of systemic inflammation (4). Nevertheless, even larger amounts of NO are released under inflammatory conditions, by an inducible isoform of the NOS that can be expressed in endothelial cells, vascular smooth-muscle cells, and monocytes/macrophages in response to several of the well-known stimuli of inflammation (5). This overproduction of NO is regarded as being one of the key mechanisms in the pronounced systemic vasodilation of sepsis (6). For this reason, inhibitors of NOS have been given to animals (7, 8) and to a small number of human patients (9) under conditions of systemic inflammation, in which this treatment was found to reverse the pronounced systemic vasodilation and decrease in blood pressure associated with inflammation.

Inhibition of NOS may, however, elicit side effects on the pulmonary vasculature. Increased pulmonary artery pressure (Ppa) and pulmonary vascular tone after inhibition of NOS have been demonstrated in various experimental models (2, 12, 13). These increases may also affect the pulmonary fluid balance. It has been shown that a transient increase in Ppa induced by a bolus of the NOS inhibitor N-nitro-L-arginine methylester (L-NAME) in healthy and endotoxemic sheep was not associated with changes in pulmonary microvascular permeability to protein or in lymph flow (2). However, it is not clear whether this would also be the case if pulmonary hypertension were to last for a prolonged period.

The present study was done to test the effects of sustained pulmonary hypertension induced by inhibition of NOS on extravascular lung water in a hyperdynamic ovine model of systemic inflammation. It was hypothesized that endotoxemia as such would be associated with pulmonary edema. Inhibition of NOS is known to cause a decrease in the cardiac index (CI) (2), which is a determinant of the filtration coefficient. It was therefore expected that even sustained pulmonary hypertension would not aggravate the pulmonary edema caused by endotoxin. A third group of animals received NO by inhalation so that its therapeutic effects on pulmonary hypertension and edema could be studied; selective pulmonary vasodilation by NO has previously been shown to reduce pulmonary arterial pressure in various pathologic conditions (14, 15).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The study was done on 18 adult chronically instrumented ewes (weight: 37.3 ± 1.8 kg [mean ± SE]), with the approval of the Government Animal Research Committee of the city of Münster.

Preparation of Animals

Instrumentation was done with the sheep under ketamine anesthesia on the day before the experiment. After anesthesia had been induced intramuscularly, a flow-directed pulmonary artery thermodilution catheter (Model 83IF75, size 7 French; Baxter Deutschland, Unterschleissheim, Germany) was inserted percutaneously through a size 8.5 French introducer sheath via the right jugular vein, and was advanced into the pulmonary artery to measure mean pulmonary arterial pressure (Ppa), pulmonary capillary wedge pressure (PCWP), effective pulmonary capillary pressure (PCP_eff), central venous pressure (CVP), and body core temperature. A combined size 4 French fiberoptic thermistor catheter (Pulsion Medizintechnik, Munich, Germany), with an additional lumen for measuring mean arterial blood pressure (MAP), was positioned in the right femoral artery and advanced proximally by 25 cm. Correct positioning of the fiberoptic catheter was verified continuously by visual checking of the pulsatility of the reflected light signal.

After additional local anesthesia with mepivacaine 1%, a tracheostomy was performed and a tracheostomy tube (I.D. = 4 mm) was inserted. The small size of the tracheostomy tube enabled the sheep to breathe simultaneously through the tube and the airway proximal to the tracheostomy, and to thereby control body temperature by panting. The sheep were kept in metabolic cages with free access to food and water.

Data Collection

For the experiment, the catheters were connected to pressure transducers (Ohmeda, Erlangen, Germany) with continuous flushing devices. Zero calibration of the catheters was performed at the level of the hip joint of the front leg while the animals were lying down. All hemodynamic variables were later recorded with the animals lying down. Blood pressures and heart rates were read from physiologic recorders (Hellige, Freiburg, Germany). PCP_eff was derived by visual determination of the pressure inflection point of the pulmonary arterial pressure tracing during inflation of the pulmonary artery catheter balloon (16). Cardiac output (CO) was measured according to the pulmonary thermodilution technique with a CO computer (Pulsion Medizintechnik). The CI and systemic vascular resistance index (SVRI) were calculated with standard formulas. The animals' body surface area (BSA) was derived from body weight through the method of Guyton and colleagues (17). The pulmonary vascular resistance index (PVRI) and its precapillary and postcapillary components were calculated according to the following equations:

(1)

(2)

(3)

Measurement of Extravascular Lung Water

The extravascular thermal volume was measured through the thermal indocyanine green dye double-indicator dilution technique, and was regarded as being equivalent to the extravascular lung water (18). A bolus of 10 ml of indocyanine green (1 mg/ml in ice-cold 5% glucose solution) was injected into the right atrium. Both indicator dilution curves were then recorded on-line in the distal aorta, and were further processed with a commercially available computer (COLD system; Pulsion Medizintechnik).

Because movement of the animals was potentially a major obstacle to taking extravascular lung water measurements in the awake animal, we initially attempted to place the fiberoptic thermistor catheter in the carotid artery. It was not possible to achieve properly reproducible determinations of extravascular lung water with this approach. Consequently, the femoral artery was used to insert the fiberoptic thermistor catheter, and special precautions therefore had to be taken to prevent the animal from moving. Measurements were made in triplicate at random points throughout the respiratory cycle, and were accepted as valid only if all of the following five conditions were fulfilled:

1. The animal was lying down throughout the measurement, and was not coughing. To prepare the animals for the measurement, their eyes were covered with a blanket. Most sheep lay down immediately in response to this. Some had to be slightly pushed on their backs to make them lie down, particularly before endotoxin infusion was started. The animals remained in the prone position with their eyes covered for 20 min before the actual measurements were made. This ensured a calm and hemodynamically stable condition for a period long enough for completion of the measurements.
2. Body temperature had to be stable during the measurements. A routine baseline temperature check was performed by the COLD computer before every injection.
3. Following injection of the indicators, the temperature signal in the aorta was greater than 0.2° C and the dye signal was greater than 10 mg/L.
4. The reflected light signal was pulsatile throughout the measurement.
5. The displayed and printed indicator dilution curves were even (smooth).

Extravascular lung water values were corrected for body weight and were displayed as an extravascular lung water index (EVLWI).

Experimental Protocol

After the baseline data had been collected, the awake, spontaneously breathing animals (n = 18) received a continuous infusion of endotoxin (Salmonella typhi, lipopolysaccharide [LPS], 10 ng/kg/min) for 28 h. After 24 h, an airflow of 6 L/min was delivered through the tracheostomy tube, with the cuff remaining unblocked. One group of animals (L-NAME/air, n = 6) received L-NAME intravenously (25 mg/ kg + 5 mg/kg/h) and breathed air. A second group (L-NAME/NO, n = 6) was given the same dosage of L-NAME, and NO (40 ppm) was added to the airflow. The third group (NaCl/air, n = 6) received NaCl 0.9% and breathed air. After the final measurements at 28 h, the animals were anesthetized with ketamine and were killed by intravenous injection of a saturated KCl solution.

Data Analysis

Results are presented as mean ± SEM. Statistical analysis of the data was done at 0 h, 24 h, and 28 h, using analysis of variance (ANOVA) with a post hoc Scheffé's F test for differences within groups. An unpaired t test with Bonferroni's correction was used to evaluate differences between the groups from 24 h to 28 h. Differences were regarded as statistically significant at p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effects of Endotoxemia

The continuous infusion of endotoxin resulted in a hyperdynamic cardiovascular response characterized by a reduction in the SVRI, an increase in CI, and a trend toward a lower MAP in all groups at 24 h (Figure 1). These hemodynamic changes remained stable in the NaCl/air group until 28 h.

View larger version (22K): [in this window] [in a new window]   Figure 1.   Endotoxin (10 ng/kg/min) induced a hyperdynamic circulatory response with decreased SVRI and increased CI in all groups by 24 h. At 24 h, L-NAME (25 mg/kg intravenously, followed by 5 mg/kg/h) increased SVRI and MAP and normalized CI. The inhalation of NO (40 ppm in air) had no significant effect in the systemic circulation. *p < 0.05 versus 0 h (NaCI/air); **p < 0.05 versus 0 h (L-NAME/air); p < 0.05 versus 24 h (L-NAME/air); p < 0.05 versus 24 h (L-NAME/NO); p < 0.05 for differences between groups, as indicated in the figure.

Ppa was increased after 24 h of endotoxemia in all groups (Figure 2), and this was associated with a tendency toward an increased PCP_eff, although the difference reached statistical significance only in the L-NAME/air group (p < 0.05 versus 0 h). The PVRI and its precapillary and postcapillary components were at baseline level after 24 h (Figure 3). The hemodynamic alterations associated with endotoxemia in the pulmonary circulation were still present in the NaCl/air group after 28 h. The EVLWI increased from a normal baseline value of between 8 and 9 ml/kg by about 50% in all groups during the first 24 h of endotoxemia, and remained at this level in the NaCl/air group (Figure 4). Pa_O2 declined by 24 h to a statistically significant level only in the L-NAME/NO group (p < 0.05 versus 0 h).

View larger version (25K): [in this window] [in a new window]   Figure 2.   Ppa was increased during hyperdynamic endotoxemia in all groups, with concomitant changes in PCPeff. After 24 h, L-NAME (25 mg/kg intravenously, followed by 5 mg/kg/h) further increased Ppa and PCPeff in the L-NAME/air group. The inhalation of NO (40 ppm in air) not only prevented the changes induced by L-NAME, but further reduced Ppa values (p < 0.05 versus NaCI/ air). *p < 0.05 versus 0 h (NaCI/air); **p < 0.05 versus 0 h (L-NAME/ air); ***p < 0.05 versus 0 h (L-NAME/NO); p < 0.05 versus 24 h (L-NAME/air); p < 0.05 versus 24 h (L-NAME/NO); p < 0.05 for differences between groups, as indicated in the figure.

View larger version (23K): [in this window] [in a new window]   Figure 3.   PVRI and its precapillary and postcapillary components were at baseline level after 24 h of endotoxemia. After 24 h, L-NAME (25 mg/kg intravenously, followed by 5 mg/kg/h) further increased PVRI on both sides of the pulmonary circulation. The changes induced by L-NAME were prevented by the inhalation of NO (40 ppm in air). **p < 0.05 versus 0 h (L-NAME/air); p < 0.05 versus 24 h (L-NAME/air); p < 0.05 for differences between groups, as indicated in the figure.

View larger version (25K): [in this window] [in a new window]   Figure 4.   EVLWI was increased in all groups after 24 h of endotoxemia. After 24 h, L-NAME (25 mg/kg intravenously, followed by 5 mg/kg/h) further aggravated pulmonary edema in the L-NAME/ air group. The inhalation of NO (40 ppm in air) prevented this further increase in EVLWI and even reduced EVLWI below its 24-h value. Pulmonary edema was still present in all groups at 28 h. PaO2 values deteriorated slightly during administration of L-NAME and improved slightly during inhalation of NO (p < 0.05 versus 24-h values). However, PaO2 values already differed between groups at 24 h. *p < 0.05 versus 0 h (NaCI/air); **p < 0.05 versus 0 h (L-NAME/air); ***p < 0.05 versus 0 h (L-NAME/NO); p < 0.05 versus 24 h (L-NAME/air); p < 0.05 versus 24 h (L-NAME/NO); p < 0.05 for differences between groups, as indicated in the figure.

Effects of L-NAME

After the administration of L-NAME, the hyperdynamic circulatory pattern was reversed (Figure 1). By 28 h, the values of SVRI and MAP rose above their 24-h values and rose in comparison with those of the NaCl/air group (p < 0.05). At 28 h, the CI values in the groups receiving L-NAME were no longer significantly different from their baseline (0 h) values.

The pulmonary hypertension that had already been present at 24 h was further aggravated in the L-NAME/air group by 28 h (p < 0.05 versus 24-h L-NAME/air, and versus NaCl/air and L-NAME/NO) (Figure 2). As a result, the PCP_eff levels were higher in the L-NAME/air group during the infusion of L-NAME than in the other two groups (p < 0.05). The PVRI values increased markedly on both the precapillary and postcapillary sides of the circulation in the L-NAME/air group (Figure 3). The EVLWI values increased further during the administration of L-NAME in the L-NAME/air group (p < 0.05 versus 24 h and versus the other two groups between 24 h and 28 h) (Figure 4). Pa_O2 further declined during this aggravation of pulmonary edema (p < 0.05 versus 0 h).

Effects of Inhaled NO

The inhalation of NO had no statistically significant effects on the L-NAME-induced reversal of the hyperdynamic pattern in the systemic circulation (Figure 1). However, during the inhalation of NO there was a trend toward a smaller reduction of CI and toward a lower SVRI in the L-NAME/NO group than in the L-NAME/air group.

Ppa levels in the L-NAME/NO group declined to values that were even lower than those in the NaCl/air group, and which no longer significantly differed statistically from the baseline values at 0 h (Figure 2). During breathing of NO, PCP_eff showed a trend toward lower values than at 24 h. The marked increase in PVRI that was seen in the L-NAME/air group was prevented by the administration of NO (Figure 3).

The inhalation of NO prevented the increase in EVLWI that was seen in the L-NAME/air group (Figure 4). EVLWI even declined in the L-NAME/NO group by 28 h (p < 0.05 versus L-NAME/NO at 24 h), but was still increased in comparison with its value at 0 h. The decline in Pa_O2 that had been present at 24 h in the L-NAME/NO group (p < 0.05 versus 0 h) was no longer statistically significant at 28 h.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Determination of EVLWI in Awake Animals

The present report describes the first model in which EVLWI was measured repeatedly in awake animals breathing spontaneously. EVLWI has previously been measured in anesthetized, ventilated animals. Ventilation as such, however, can affect EVLWI, particularly if positive end-expiratory pressure is applied (19, 20) or if vasoactive mediators such as atrial natriuretic peptide (ANP) are released as a result of ventilation. ANP affects permeability in the pulmonary circulation (21). In the present study, EVLWI was measured using the transpulmonary double-indicator dilution technique. Simultaneous in vivo recording of both indicator dilution curves was done with a combined fiberoptic thermistor catheter, which was inserted through the femoral artery and advanced into the aorta. Measurement of EVLWI using the Edwards lung water computer included aspiration of blood and ex vivo photometric determination of extinction of the light signal from indocyanine green. The time delay between the in vivo measurement of the temperature signal and the ex vivo determination of the dye signal was found to explain a dependency on CO of lung water measurements with the Edwards lung water computer (22). Extravascular lung water measurements made with the new combined fiberoptic thermistor catheter do not have this limitation. The finding that EVLWI increased only in the L-NAME/air group, despite the decrease in CI in both groups that had received L-NAME, is in keeping with the absence of a dependency on CO of lung water measurements made with the new device.

In earlier investigations, extravascular lung water was evaluated only in anesthetized animals, since the movement of animals during the measurements would have profoundly disturbed the recording of the indicator dilution curves. As stated earlier, however, standards for the conditions of the animals and for correct technical performance were established in the present study to ensure reliable measurements.

The new model was used to study the way in which extravascular lung water is affected by persistent pulmonary hypertension caused by inhibition of NOS with and without inhalation of NO in hyperdynamic ovine endotoxemia. The main findings were that: (1) continuous infusion of endotoxin was associated with a hyperdynamic circulatory response characterized by decreased SVRI and increased CI; (2) nonselective inhibition of NOS reversed these hemodynamic changes, but was associated with pulmonary hypertension and the aggravation of pulmonary edema; and (3) inhalation of NO during inhibition of NOS reduced Ppa and EVLWI.

Effects of Endotoxemia

The ovine reaction to the continuous administration of endotoxin has been studied repeatedly (7, 12, 23), mainly in connection with the hyperdynamic circulatory response that mimics the hemodynamic pattern of human endotoxemia (24) and sepsis (25). Ppa was increased and PCP_eff showed corresponding trends at 24 h, which was at least in part explained by an increased pulmonary blood flow. Possible increases in pulmonary blood volume may have contributed to pulmonary hypertension, but were not measured in this study. PVRI was at baseline level at 24 h, suggesting that pulmonary vascular tone was not responsible for the increased Ppa.

EVLWI was elevated after 24 h of endotoxemia. This finding was in keeping with the findings in previous studies, which showed increased pulmonary lymph flows after 24 h of endotoxemia, suggesting a high transvascular fluid flux and the presence of pulmonary edema (2, 26).

Effects of L-NAME

Nonselective blockade of NOS reversed the hemodynamic changes of endotoxemia in the systemic circulation. The bolus of L-NAME used was selected on the basis of a previous study in which it was found to reverse systemic vasodilation without causing selective impairment of regional blood flows (7). The pulmonary vascular changes, however, showed a rather phasic pattern when only a bolus of the NOS inhibitor was administered (12, 27). To permit study of the effects of persistent pulmonary hypertension on lung water, a continuous infusion of L-NAME was added to the initial bolus. Under this regimen, CI at 28 h showed a trend toward slightly lower values than at baseline.

The increase in Ppa during blockade of NOS was due to pulmonary vasoconstriction, which occurred in both the precapillary and postcapillary compartments of the pulmonary vasculature. Pulmonary transvascular fluid filtration occurs on both the precapillary and postcapillary sides, with filtration on the postcapillary side predominating (28). This may explain why pronounced pulmonary edema developed after inhibition of NOS, which was associated with marked venoconstriction. EVLWI had already increased after 1 h of NOS inhibition. Lung lymph flow did not increase after a bolus of L-NAME in a previous study (2). This discrepancy may be due to the short peak in Ppa after a single bolus of NO inhibitor. Pulmonary hypertension may have to last for some time before the lymph flow increases. This explanation would be in keeping with the gradual increase in lymph flow that is seen in the model of hydraulic pulmonary venous occlusion. Even a marked increase in pressure on the venous side takes at least 1 h to produce significant increases in lymph flow (unpublished data).

Effects of NO

NO reduced Ppa values to levels that were even lower than in the endotoxin control animals. However, the PVRI in these two groups was practically identical. The reason for the lower Ppa in the L-NAME/NO group can be attributed to the lower CI following inhibition of NOS in these animals, even though the difference from the endotoxin controls did not reach statistical significance. This phenomenon provides further evidence that the increased Ppa during hyperdynamic endotoxemia may be primarily due to the increased pulmonary blood flow in this condition rather than to the presence of vasoconstrictors, even though concentrations of the latter are increased during hyperdynamic endotoxemia in sheep (29).

Inhibition of NOS not only induced precapillary pulmonary vasoconstriction, but also induced marked venoconstriction, which was reversed by NO. Our finding of a high responsiveness of the pulmonary veins of endotoxemic sheep to inhibition of NOS is in keeping with the data from an in vitro study by Nelson and colleagues (30). They reported a marked depression of the ex vivo responsiveness of pulmonary veins from septic sheep to norepinephrine in combination with high venous levels of cyclic guanosine monophosphate (cGMP). Both the sepsis-induced responsiveness to norepinephrine and the increase in cGMP levels could be partly reversed by NOS inhibition, suggesting both the presence of NOS activity and an effect of other vasodilatory substances on venous tone during sepsis. Moreover, our data are consistent with the findings of other investigators, who have reported a particularly vasodilatory role of NO in the pulmonary venous vasculature (31, 32), and who provide a mechanism explaining the way in which inhibition of NOS may lead to pulmonary edema.

EVLWI values decreased significantly during NO inhalation between 24 h and 28 h suggesting an edema-reducing effect of NO in addition to its elimination of the effect of NOS inhibition. Direct intergroup comparisons, however, did not reveal any significant difference in EVLWI between the NaCl/ air and the L-NAME/NO groups at 28 h. Nishina and associates ventilated rabbits with NO for 6 h during early endotoxemia (33). There was a trend toward lower wet/dry weight ratios in the NO-treated rabbits versus endotoxemia, although this was too small to reach statistical significance. Benzing and coworkers reported a decrease in pulmonary albumin flux during NO administration in patients with acute lung injury (34). A conclusive evaluation of selective pulmonary vasodilation in the treatment of pulmonary edema is not yet possible. It is conceivable, however, that inhalational vasodilators might serve this therapeutic purpose, particularly if the gain in Ppa and PCP_eff is high enough, if permeability is increased, and if the vasodilator is inhaled for a sufficiently long period.

Levels of CI in the L-NAME/NO group showed a trend toward higher values than in the L-NAME/air group during treatment with L-NAME. This may have been caused by the substantially reduced right ventricular afterload in the L-NAME/ NO group. A large increase in CO was recently documented in a patient with adult respiratory distress syndrome (ARDS) and acute right heart failure when selective pulmonary vasodilation was produced with NO (35). Improvement in right ventricular function during NO inhalation was also seen by Rossaint and colleagues, but without concomitant changes in CI (36). This discrepancy from the data reported here may have come from the much smaller reduction in Ppa in their investigation than in the experimental protocol used in the present study.

There was also some indication of a lower SVRI in the L-NAME/NO group than in the L-NAME/air group during inhalation of NO, even if the difference did not reach statistical significance. This contrasts with the general view that inhaled NO acts as a highly selective pulmonary vasodilator that fails to elicit systemic effects because of NO scavenging by hemoglobin. A reduction in SVRI would, however, be in keeping with a proposition made by Stamler's group (37, 38). They have suggested that hemoglobin is S-nitrosylated in the lung when red blood cells are oxygenated, and that NO is then released during arterial-venous transit to improve microcirculatory blood flow, particularly under hypoxic conditions. The vasorelaxant activity of S-nitrosohemoglobin is promoted by the erythrocytic export of S-nitrosothiols, which protect NO from scavenging by hemoglobin. A vasodilatory effect in the systemic circulation during inhalation of NO may have been demasked in our study, since the endogenous production of S-nitrosothiols can be inhibited by inhibition of NOS (37). This still very speculative explanation, however, needs to be addressed in further investigations, since this effect was not part of the primary hypothesis of the present study.

In summary, the present report describes a new model of persistent, L-NAME-induced pulmonary hypertension in spontaneously breathing sheep, in which EVLWI can be measured repeatedly. In this model, pulmonary edema was present during the hyperdynamic phase of endotoxemia, and was aggravated by L-NAME. NO reduced Ppa in L-NAME-treated animals to baseline levels, and reduced EVLWI. Further studies are needed to address the question of whether selective pulmonary vasodilation may serve as a therapeutic approach to reducing pulmonary edema. The new model of persistent pulmonary hypertension described here might also be used for intraindividual comparison of the effects of different short-lived pulmonary vasodilators.