INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Chronic lung ischemia occurs in chronic thromboembolic obstruction of the pulmonary arteries in humans and can be induced in experimental animals by chronic ligation of the pulmonary artery. Chronic pulmonary artery obstruction in humans or animals is associated with marked expansion of the systemic blood supply to the lung. Pulmonary thromboendarterectomy, which restores perfusion to the previously occluded zones, is becoming the standard treatment in patients with chronic pulmonary thromboembolic hypertension (1). Two significant complications can occur after pulmonary thromboendarterectomy, namely a persistent increase in pulmonary vascular resistance and acute high-permeability pulmonary edema in the reperfused lung areas (2, 3). We recently found similar alterations in lung permeability and pulmonary vascular resistance distal to chronic thromboembolic obstruction of the left pulmonary artery in pigs (4).

The endothelium contributes to the local regulation of pulmonary vascular tone (5) and lung microvascular permeability (6) by releasing nitric oxide (NO). NO is synthesized in the endothelium from L-arginine and molecular oxygen by the enzyme type III endothelial NO synthase (eNOS). The biosynthesis of NO and L-citrulline from L-arginine and molecular oxygen involves a series of oxidations that require several cofactors such as NADPH, tetrahydrobiopterin, flavin adenine dinucleotide (FAD), and flavin mononucleotide (FMN).

Conceivably, the function of the L-arginine/eNOS system may be impaired in chronic ischemia because of decreases in shear stress (7), energy metabolism, and availability of substrates or cofactors for eNOS (8). However, unlike other solid organs, the pulmonary endothelium may retain normal function in the face of chronic lung ischemia because the regional ventilation supplies oxygen and because expansion of the systemic blood supply may compensate for the decreases in shear stress and substrate delivery to the lung. If this is indeed the case, eNOS dysfunction may be less severe after chronic than after acute ischemia. The aims of the present study were therefore twofold. First, we measured ATP and lactate lung concentrations, taken as markers of ischemia, 5 wk and 2 d after left pulmonary artery ligation in piglets. Second, in the same model, we compared eNOS lung activity, eNOS protein concentration, and eNOS-dependent relaxation in postobstructive pulmonary arteries.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Thirty piglets (Large White, mean weight ± SEM, 22.3 ± 4.7 kg) were used. The study complied with the "Principles of Laboratory Animal Care" developed by the National Society for Medical Research and the Guide for the Care and Use of Laboratory Animals written by the Institute of Laboratory Animal Resources and published by the National Institutes of Health (NIH Publication No. 86-23, revised in 1985).

Study Groups

The piglets were randomly allocated to three groups (n = 10 in each group). The studies were done 2 d and 5 wk after left pulmonary artery ligation (2-d and 5-wk ligation groups, respectively) and 2 d after left pulmonary artery dissection without ligation (sham group).

Pulmonary Artery Ligation

Anesthesia was induced with intramuscular ketamine (100 mg/kg) and maintained with intravenous pentobarbital (10 mg/kg bolus, followed by a continuous infusion of 0.1 mg/kg/min). The animals were paralyzed with pancuronium (0.3 mg/kg). After endotracheal intubation, intermittent positive-pressure ventilation was provided (MMS 107 ventilator; Pall, France) at a tidal volume of 15 ml/kg, with a respiratory rate of 18 cycles/min and a fraction of inspired oxygen (FI_O2) of 0.5. Body temperature was kept constant at 37° C. A midline sternotomy was performed under sterile conditions, the pericardium was opened, and the intrapericardial left pulmonary artery was dissected and ligated using a nonabsorbable loop. The chest was closed, and the animals were allowed to recover.

Pulmonary Harvesting

The pigs were anesthetized as described previously. A midline sternotomy was performed. Left pulmonary artery pressure was measured 1 cm downstream from the ligation site by direct puncture. After hep- arinization, the animals were killed by exsanguination, and the lungs were rapidly removed from the chest.

Morphology

Gross anatomic examination of the left lung was performed with special attention to the bronchial and systemic circulation supplying the left lung. Large specimens were taken from the upper and lower lobes of the left lung and fixed by immersion in 10% neutral buffered formalin. After staining with hematoxylin-eosin, all biopsy specimens were examined by observers who were blinded to the study conditions.

ATP and Lactate Measurements

All biopsy specimens of fresh left lungs were immediately frozen at a depth of 1 mm using metal tongs chilled in liquid nitrogen, then stored at 80° C. ATP and lactate lung concentrations were measured as described by Date and associates (9). The specimens were pulverized in liquid nitrogen, and metabolites were extracted in 0.6 N perchloric acid. After centrifugation and neutralization, the supernatants were analyzed for metabolites using the enzymatic methods of Lowry and Passonneau (10). Data are given as micromoles per gram of frozen tissue.

Isolated Pulmonary Artery Ring Studies

At the end of each experiment, intrapulmonary arterial segments were dissected out and placed in warm Krebs-Henseleit buffer composed of (in mM) 118.3 NaCl, 4.7 KCl, 2.5 CaCl₂ · H₂O, 1.2 KH₂PO₄, 1.2 MgSO₄ · 7 H₂O, 25 NaHCO₃, 0.03 ethylenediaminetetraacetic acid (EDTA), and 11.1 glucose. Isolated pulmonary arteries were cleaned and cut into rings 3 to 4 mm in length (1 to 2 mm outer diameter). Three to four rings were obtained from each animal. The rings were then mounted on stainless steel hooks, suspended in 10-ml tissue baths, and connected to force displacement transducers (LB-5; Showa- sokki, Tokyo, Japan) for force change recording using a chart recorder (LR 4210; Yokogawa, Tokyo, Japan). The baths were filled with 10 ml of Krebs-Henseleit buffer and aerated at 37° C with a mixture of 95% O₂-5% CO₂. Pulmonary artery rings were initially stretched to produce a preload of 1 g of force, and were then allowed to equilibrate for 60 to 90 min. Pilot studies involving length tension analysis of pulmonary artery rings showed that this preload value provided optimal resting tension; in addition, it is similar to the optimal resting tension (1.060 ± 0.040 g) found by Liu and coworkers (11) in piglet pulmonary arterial rings. During this period, the Krebs-Henseleit buffer in the tissue baths was changed every 10 min. After incubation with indomethacin (10⁵ M) for 60 min, a concentration-response curve to phenylephrine (10⁹ M/L to 3 · 10⁴ M/L) or to the thromboxane analog U 46619 (10⁹ M/L to 10⁶ M/L) was constructed. The rings were then washed, and the developed force was allowed to return to the baseline values. Next, the rings were precontracted with phenylephrine to generate approximately 1 g of developed force. Once a stable contraction was obtained, cumulative doses of acetylcholine (10⁹ to 10⁴ mol/L) were added to the bath, and changes in endothelium- dependent relaxation were assessed. Other rings were then precontracted with U 46619 to generate approximately 1 g of developed force. Once a stable contraction was obtained, cumulative doses of the endothelium-dependent vasodilator calcium ionophore A23187 (10¹⁰ M/L to 3 · 10⁷ M/L), or of the endothelium-independent vasodilator sodium nitroprusside (10⁹ M/L to 10⁴ M/L) were added to the bath. These rings were washed again and allowed to equilibrate to baseline levels. To determine whether alterations in endothelium- dependent relaxation were related to decreased delivery of L-arginine in the ligated animals, the procedure was repeated after 30 min of incubation with L-arginine (10⁴ mol/L).

In addition to the change in force, responses were assessed based on determination of the concentration that produced 50% of the maximal response (EC₅₀) extrapolated from a plot of log concentration versus percentage of maximal response. The contractile responses to phenylephrine were expressed in absolute values (milligrams), and the maximal relaxation to acetylcholine and sodium nitroprusside was expressed as the percentage of the phenylephrine-induced precontraction: 0% indicated no relaxation and 100% relaxation equal in magnitude to the precontraction.

Drugs

L-Arginine hydrochloride, A23187 (calcium ionophore), phenylephrine, sodium nitroprusside, acetylcholine hydrochloride, and indomethacin were purchased from Sigma Chemical Co. (St. Louis, MO). U46619 was provided by Upjohn USA (Peapack, NJ).

Measurement of Angiotensin-converting Enzyme (ACE) and eNOS Activities

Because one aim of this study was to investigate the consequences of chronic lung ischemia on endothelium function, we measured two important endothelial cell protein activities located in different subcellular sites: the caveolae-protein eNOS activity and the non-caveolae-protein ACE activity (12).

Measurement of ACE activity. ACE activity was determined based on the amount of hippuric acid produced by hydrolysis of the synthetic substrate p-benzoyl-L-glycyl-L-histidyl-L-leucine (Hip-His-Leu; Bachem, Bubendorf, Switzerland). Pulmonary artery segments from left lungs prepared as described previously were homogenized, using a potter homogenizer (Braun, Saint Quentin Fallavier, France), at 4° C, in 1 ml cold phosphate-buffered saline (120 mmol/L NaCl, 2.7 mmol/L KCl, 10 mmol/L phosphate buffer, pH 7.4) containing 8 mmol/L of the detergent 3-[(3-cholamidopropyl)dimethylammonio]-1-propane-sulfonic acid (Sigma). The crude homogenate was centrifuged at 600 g for 10 min at 4° C. Five microliters of the supernatant were incubated at 37° C in a total volume of 250 µl buffer containing 5 mmol/L substrate, 100 mmol/L potassium phosphate (pH 8.3), and 300 mmol/L NaCl (13). The reaction was stopped by addition of 50 µl of 12% (wt/ vol) phosphoric acid. The amount of hippuric acid produced from the substrate was measured by high-performance liquid chromatography (14). One unit of ACE activity was defined as the amount of enzyme producing 150 µmol hippuric acid per minute. Total protein concentration was measured according to Bradford (15) using bovine serum albumin (BSA) (Sigma) as the standard and Coomassie brilliant glue G-250 (Bio-Rad, Ivery, France).

Measurement of lung NOS activities. Calcium-dependent and calcium-independent NOS activities were determined in left lung homogenates from the three groups. The lungs were quickly frozen in liquid nitrogen immediately after removal. Tissue was homogenized on ice, using an ultraturax blender (Kinematica, Lucerne, Switzerland), in 4 volumes of buffer containing 50 mM Tris-HCl (pH 7.4), 0.1 mM EDTA, 0.1 mM ethyleneglycol-bis-(-aminoethyl ether)-N,N'-tetraacetic acid (EGTA), 0.1% 2-mercaptoethanol, 1 µM leupeptin, 1 µM pepstatin A, and 1 mM phenylmethylsulfonyl fluoride. The homogenate was centrifuged at 100,000 g for 1 h at 5° C. To remove soluble proteins, the pellet was resuspended in homogenization buffer containing 1 mol/L KCl and allowed to stand on ice for 5 min before centrifugation at 100,000 g for 30 min at 5° C. The supernatant fraction was discarded and the pellet (membrane fraction) was resuspended in homogenization buffer containing the detergent CHAPS [{(cholamidopropyl)dimethylammonio}propanesulfonate] (20 mM/L), 1 mol/L KCl, and glycerol (10% vol/vol), then allowed to stand on ice for 30 min before centrifugation at 100,000 g for 30 min at 5° C. Activity of eNOS in the supernatant (membrane fraction) was determined by measuring the calcium-dependent conversion of [³H]L-arginine to [³H]L-citrulline in the reaction mixture. The enzyme extract (25 µl) was added to 200 µl of the reaction mixture containing 50 mmol/L Tris-HCl (pH 7.4), 10 µmol/L tetrahydrobiopterin, 1 mmol/L dithiothreitol (DTT), 10 µg/ml calmodulin, 4 µM FAD, 4 µM FMN, 2 µM L-arginine, 10³ cpm/µl L-[³H]arginine, and 1 mmol/L NADPH, with or without 1 mmol/L CaCl₂. After 40 min of incubation at 37° C, the reaction was stopped with 2 ml of a solution containing 20 mmol/L Na acetate, pH 5.5, 1 mmol/L L-citrulline, 2 mmol/L EDTA, and 0.2 mmol/L EGTA. The mixture was applied to a 1-ml Dowex AG 50WX8 column (Bio-Rad), and [³H]L-citrulline was eluted with 2 ml of distilled water. The radioactivity in the eluate was measured using liquid scintillation spectroscopy, and the concentration of protein in the enzyme extract was determined according to Lowry.

Lung eNOS Protein

Tissue was homogenized on ice using an ultraturax homogenizer (Kinematica) in 4 volumes of buffer containing 50 mmol/L Tris-HCl (pH 7.4), 0.1 mmol/L EDTA, 0.1 mmol/L EGTA, 0.1% 2-mercaptoethanol, 1 µmol/L leupeptin, 1 µmol/L pepstatin A, and 1 mmol/L phenylmethylsulfonyl fluoride, CHAPS (20 mmol/L) and allowed to stand on ice for 30 min before centrifugation at 3,000 g for 10 min at 5° C.

The supernatant was subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Proteins in the gel were transferred to a nitrocellulose membrane by electroblotting in a transblot Bio-Rad transfer apparatus, for 12 h at 4° C. Before the transfer, the gels, Whatman filter paper, and nitrocellulose membrane were soaked in electroblotting buffer (25 mM Tris-HCl; 193 mM glycine; 20% methanol, pH 8.0) for 15 min. After transfer, the membrane was blocked using 1× TBST (0.15 M NaCl; 10 mM Tris-HCl, pH 8.0; 0.05% Tween 20; and 5 % BSA) for 1 h at room temperature. The eNOS protein was detected by incubating the membrane overnight at 4° C with mouse polyclonal anti-eNOS (Interchim, Montlucon, France) diluted 1:1,000. The membrane was washed three times in 1× TBST. Specific protein was detected using a horseradish peroxidase-conjugated secondary antibody and ECL reagents (Amersham, Buckinghamshire, UK). eNOS immunoreactivity was quantified using a semi-automated image analysis device (NIH image 1.52) that quantifies both the area and the intensity of immunoreactive bands using a ScanJet II scanner (Hewlett-Packard, Boulogne, France) with DeskScan II (Hewlett-Packard) software. Results are reported in arbitrary units.

Statistical Analysis

All results are reported as means ± SEM. One-way analysis of variance followed by Fisher's test for between-group comparisons was done. Correlations between ATP and lactate concentrations were examined using simple linear regression. All statistical analyses were performed using Statview IV (Abacus Concept, Berkeley, CA). Values of p < 0.05 were considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Morphology

The gross appearance of the right lung parenchyma was unremarkable in all three groups.

Hemorrhagic zones separated by morphologically normal zones were observed in the left lung parenchyma in the 2-d ligation group, which was the only group with nonadherent luminal thrombi in the pulmonary arteries distal to the ligation site. Histologic examination of the lungs in this group consistently showed focal hemorrhage with extravasation of red cells into alveolar spaces.

The bronchial circulatory system was hypertrophied in all the 5-wk ligation animals. Pulmonary artery pressure downstream from the ligation site was measurable only in the 5-wk ligation animals (1.8 ± 0.6 mm Hg). Moreover, all animals in this group exhibited back-bleeding distal to the ligation. Altogether these observations indicate that bronchial blood flow effectively back-perfuses proximal pulmonary arterial bed in the 5-wk ligation group. The left lungs from the 5-wk ligation and sham groups were histologically normal.

ATP and Lactate Lung Concentrations

There were no differences between the sham, 2-d ligation, and 5-wk ligation groups regarding right lung concentrations of ATP or lactate. Left lung ATP concentrations (Figure 1) were lower in the 2-d than in the 5-wk ligation group, whereas ATP concentrations were higher in the sham group than in the two ligation groups.

View larger version (18K): [in this window] [in a new window]   Figure 1.   Left lung ATP concentrations (mean ± SEM) 2 d and 5 wk after left pulmonary artery (PA) ligation and 2 d after sham operation.

Left lung lactate concentrations had the inverse profile (Figure 2). Lactate concentrations were higher in the 2-d than in the 5-wk ligation group, but lower in the sham group than in the two ligation groups. Left lung ATP and lactate concentrations were linearly correlated (r = 0.73, p = 0.003).

View larger version (18K): [in this window] [in a new window]   Figure 2.   Left lung lactate concentrations (mean ± SEM) 2 d and 5 wk after left PA ligation and 2 d after sham operation.

ACE Activity

As compared with the sham group (4.33 ± 0.90 U/g), left pulmonary artery ACE activity was decreased 2 d (2.44 ± 0.44 U/g; p = 0.03) and 5 wk (2.86 ± 0.40 U/g; p = 0.01) after ligation. No significant differences were found between the two ligation groups or regarding the right pulmonary arteries.

Isolated Left Pulmonary Artery Ring Study

Maximal contraction to phenylephrine was similar in all three groups. The phenylephrine EC₅₀ was higher in the 5-wk ligation group than in the 2-d ligation and sham groups (8.7 · 10⁷ ± 1.1 · 10⁷ M versus 4.0 · 10⁷ ± 1.1 · 10⁷ M, p = 0.037 and 8.7 · 10⁷ ± 1.1 · 10⁷ M versus 2.6 · 10⁷ ± 1.1 · 10⁷ M, p = 0.0002, respectively). There was no difference between the 2-d ligation group and the sham group.

Maximal contraction to U 46619 was lower in the 5-wk ligation group than in the 2-d ligation and sham groups (3,279 ± 347 mg versus 5,149 ± 572 mg; p = 0.006 and 3,279 ± 347 mg versus 5,850 ± 388 mg; p = 0.0004, respectively). There was no difference between the 2-d ligation group and the sham group. The U 46619 EC₅₀ was similar in the three groups.

Relaxation in response to sodium nitroprusside was not affected by pulmonary artery ligation.

Maximal relaxation in response to acetylcholine was lower 2 d than 5 wk after left pulmonary artery ligation (Figure 3), whereas maximal relaxation in response to acetylcholine was higher in the sham group than in the two ligation groups (Figure 3).

View larger version (18K): [in this window] [in a new window]   Figure 3.   Relaxation (percent reduction of maximal contraction to phenylephrine) produced by acetylcholine stimulation of left PA rings derived from left lungs after 2-d and 5-wk PA ligation as compared with sham. Results are presented as mean ± SEM. The concentration- response curves in the two PA ligation groups were significantly reduced compared with the sham group.

The acetylcholine EC₅₀ was higher in the 5-wk ligation group than in the 2-d ligation and sham groups (3.23 · 10⁷ ± 6.1 · 10⁸ M versus 1.45 · 10⁷ ± 6.1 · 10⁸ M, p = 0.04 and 3.23 · 10⁷ ± 6.1 · 10⁸ M versus 1.29 · 10⁷ ± 5.9 · 10⁸ M; p = 0.03, respectively).

Maximal relaxation in response to calcium ionophore was decreased in the 5-wk ligation group only (Figure 4). No differences in the calcium ionophore EC₅₀ were seen among the three groups. The decreased relaxation to calcium ionophore 5 wk after pulmonary artery ligation was not restored by preincubation with L-arginine (Figure 4).

View larger version (21K): [in this window] [in a new window]   Figure 4.   Relaxation (percent reduction of maximal contraction to U 46619) produced by calcium ionophore stimulation of left PA rings derived from left lungs after 2-d PA ligation and 5-wk PA ligation (with and without addition of L-arginine (L-Arg) in the organ bath) as compared with sham. Results are presented as mean ± SEM. The concentration-response curves in the 5-wk PA ligation group were significantly reduced as compared with the sham and 2-d PA ligation groups. Addition of L-arginine did not restore normal relaxation in the 5-wk PA ligation group.

eNOS Activity

Left lung calcium-dependent NOS activity in the 5-wk ligation group was lower than in the sham group and 2-d ligation group. As compared with the sham group, this activity was increased 2 d after pulmonary artery ligation. Left lung calcium-independent NOS activity was similar in the three groups (Figure 5).

View larger version (16K): [in this window] [in a new window]   Figure 5.   Left lung calcium-dependent and calcium-independent NOS activities (mean ± SEM) 2 d and 5 wk after left PA ligation and 2 d after sham operation. *p < 0.0001, comparison between 2-d PA ligation and sham groups; **p < 0.0001, comparison between 2-d and 5-wk PA ligation groups; p = 0.0005, comparison between 5-wk PA ligation and sham groups.

Left Lung eNOS Protein Level

Left lung eNOS protein level in the 5-wk ligation group was lower than in the sham group and 2-d ligation group (Figure 6). As compared with the sham group, the eNOS protein level was increased 2 d after pulmonary artery ligation. Left lung calcium-dependent NOS activity and eNOS protein level were linearly correlated (r = 0.9, p < 0.0001): the higher the protein, the higher the activity (Figure 7).

View larger version (20K): [in this window] [in a new window]   Figure 6.   Western blot of eNOS protein content (mean ± SEM) of 2-d and 5-wk left PA ligation and 2-d sham surgery animals. Typical results in two animals of each group are shown on the left panel.

View larger version (16K): [in this window] [in a new window]   Figure 7.   Linear correlation between left lung calcium-dependent NOS activity and eNOS protein level (r = 0.9, p < 0.0001) in the 2-d and 5-wk left PA and 2-d sham surgery animals.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Because long-term pulmonary artery obstruction is associated with expansion of the bronchial circulation and development of extensive bronchopulmonary anastomoses, we reasoned that these changes may prevent chronic lung ischemia, thereby preserving eNOS function in the postobstructive pulmonary vessels. Contrary to this hypothesis, we found that eNOS activity in the left lung and eNOS-dependent relaxation of the left pulmonary arteries were more depressed 5 wk than 2 d after left pulmonary artery ligation. These differences contrasted with the higher ATP and lower lactate concentrations found in the left lung after 5 wk than after 2 wk of ligation, indicating less severe ischemia after the longer period of pulmonary artery obstruction.

In agreement with previous studies in dog and sheep (16, 17), we consistently found marked expansion of the systemic blood supply to the lung and back-bleeding distal to the site of pulmonary artery ligation with significant perfusion pressure in the left pulmonary arteries (indicating the presence of anastomoses between the bronchial and pulmonary circulations) in piglets studied 5 wk after pulmonary artery ligation. Because ventilation is maintained after pulmonary artery ligation, oxidative metabolism in the obstructed lung is dependent only on the provision to cells of metabolizable substrates present in the bloodstream. We consequently used ATP and lactate concentrations in the lung as markers of ischemia. To our knowledge, our study is the first to show that the ATP decrease and lactate increase in the lung were less marked 5 wk than 2 d after pulmonary artery ligation, indicating that the consequences of pulmonary artery obstruction on lung metabolism diminished over time. This finding is consistent with the speculation that the development of collateral pulmonary blood vessels may supply cells with the substrates they need during prolonged pulmonary artery occlusion.

Despite the lessening of ischemia-induced metabolic effects over time, eNOS function was impaired in the postobstructive pulmonary arteries 5 wk after ligation. Both receptor-mediated (acetylcholine) and non-receptor-mediated (calcium ionophore) endothelium-dependent relaxations were depressed after 5 wk, whereas after 2 d only the receptor-mediated response was decreased. This decrease in the endothelium- dependent response was not restored by administration of the NO precursor L-arginine, and the response to the non-endothelium-dependent vasodilator agonist sodium nitroprusside remained normal in both ligation groups. Taken in concert, these results indicate that the decrease in endothelium-dependent relaxation 5 wk after pulmonary artery obstruction was not due to a defect in the membrane receptors or in the signaling mechanisms activated by these receptors, to an interruption of the supply of L-arginine to endothelial cells, or to an inability of the smooth muscle to relax in response to NO, but was probably related to impaired NOS function. In support of this interpretation, we found that calcium-dependent NOS activity was decreased in the long-term ligation group, whereas calcium-independent NOS activity remained unchanged, indicating a decrease in eNOS activity. The function of eNOS has been studied in experimental models of acute but not of chronic lung ischemia. Our results in animals studied 2 d after pulmonary artery ligation are consistent with reports by Serraf and coworkers (18) and Nyhan and coworkers (19) of a selective decrease in receptor-mediated endothelium-dependent pulmonary artery relaxation after acute lung ischemia caused by cardiopulmonary bypass in dogs and pigs. This decrease contrasts with our finding of increased lung eNOS activity after acute ischemia, a result also reported by Lu and coworkers (6), and suggests that acute lung ischemia may cause selective alterations in receptor-mediated eNOS signaling. Decreased eNOS activity in the 5 wk ligation animals was most likely caused by a decrease in eNOS expression, as indicated by the proportional decreases in eNOS activity and protein level (Figure 7). Recent study by Le Cras and coworkers (7) found decreases in eNOS mRNA and protein expression in rat lung 4 wk after pulmonary artery stenosis. Thus, eNOS expression may be downregulated by the reduction in shear stress seen in pulmonary arteries distal to an obstruction. In our study as well as in the study by Le Cras and coworkers (7), eNOS expression was measured in lung parenchyma and eNOS- dependent relaxation in intraparenchymal large pulmonary arteries. Thus, eNOS protein does not necessarily reflect the changes that occurred at the level of the pulmonary arteries. However, ACE activity, another important endothelial protein marker (12), was decreased in the pulmonary arteries of the 5 wk ligation animals, suggesting that important endothelial protein functions were impaired in both the proximal and distal pulmonary vascular bed.

In our study, contraction of postobstructive pulmonary arteries in response to phenylephrine was normal, whereas response to the thromboxane analog U 46619 was decreased. Alteration in smooth muscle contractility is an unlikely mechanism because phenylephrine maximal contraction was unchanged. Moreover, compared with the 2-d ligation or sham groups, the U46619 concentration-response curves in the 5-wk ligation group only differed in the maximal effect (Emax) values, whereas EC₅₀, an indicator of the sensibility to the agonist, remained unchanged. Altogether these results suggest that the number rather than the function of the thromboxane receptors was decreased in the 5-wk ligation group pulmonary arteries. All these observations contrast with previous studies showing increased responsiveness of postobstructive pulmonary arteries to 5-hydroxytryptamine in dogs (20) and to endothelin in rats (21). Differences in species, agonists, or duration of ischemia may account for these discordant results.

Our study may have important clinical implications in patients with chronic pulmonary thromboembolic hypertension. First, pulmonary thromboendarterectomy done to reperfuse the obstructed pulmonary circulation has been followed by acute high-permeability edema in up to 60% of cases (22). We previously demonstrated that this high-permeability edema reflected an acute inflammatory response to ischemia-reperfusion (23). Because endogenous NO plays a key role in protecting against ischemia-reperfusion lung injury (6), we suggest that a decrease in NO release from the endothelium may further worsen reperfusion edema. Second, pulmonary vascular resistance has been reported to remain elevated after total surgical repermeabilization of pulmonary arteries in humans (24) or experimental animals (23). The mechanisms for this have not been elucidated but may include persistent eNOS function impairment.

In conclusion, we found that, despite relative conservation of the pulmonary aerobic metabolism, prolonged pulmonary artery obstruction decreased eNOS activity, eNOS protein level, and eNOS-dependent relaxation of postobstructive pulmonary arteries. Potential mechanisms may include downregulation of eNOS expression owing to reduced shear stress in the postobstructive vascular bed.