INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

IN THE NORMAL LUNG, changes in pulmonary microvessel tone in response to alveolar hypoxia and hypercapnia regulate the local distribution of pulmonary blood flow (4, 29, 32, 34). To the best of our knowledge, however, no thorough attempts have ever been made to determine whether these hypoxia- and hypercapnia-related regulatory mechanisms of pulmonary microvessel tone are significantly impaired in the injured lung. Several groups of investigators have attempted to obtain definite evidence for changes in pulmonary vessel diameter in response to hypoxia (9, 11, 14, 16, 21, 24) or hypercapnia (5, 6, 8, 16, 17, 23, 27, 28) on the basis of radiological and nonconfocal microscopic studies. Unfortunately, however, these studies mainly focused on events in small vessels outside pulmonary acini. We developed a real-time confocal laser luminescence microscope coupled to a high-speed video camera to precisely distinguish dynamic cell kinetics in the arterioles, venules, and capillaries of a single pulmonary acinus (33). Applying this elaborate confocal system to microvessel kinetics in rat acini, we studied intra-acinar microvessel constriction and dilation elicited by hypoxia and hypercapnia in the injured lung. Exact knowledge of microvessel reactions to oxygen is indispensable when patients with diseased lungs are treated with varying modes of artificial ventilation, because pulmonary gas exchange is not adequately maintained if hypoxic microvessel constriction deteriorates (4, 32, 34). The importance of clarifying the hypercapnia-induced microvessel response in diseased lungs has increased, because hypercapnic conditions are encountered when patients are treated by controlled ventilation with permissive hypercapnia (1, 13). Because of their clinical significance, we hoped to shed light on the following issues in injured lungs exposed to a hyperoxic environment that are generally encountered when patients having serious hypoxemia are artificially ventilated: 1) whether the hypoxia-induced vascular response in acinar microvessels is impaired; 2) whether the hypercapnia-induced acinar microvessel response is abnormal; 3) the relative importance of NOS- and COX-associated pathways for the impaired hypoxic response in acinar microvessels; and 4) whether the contribution of NOS- and COX-associated pathways to abnormal hypercapnic microvessel reactions differs qualitatively from their contribution to impaired hypoxic microvessel reactions.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Isolated perfused lung preparation. We prepared isolated perfused lungs from pathogen-free 8-wk-old male Sprague-Dawley rats weighing 250-300 g (n = 163). Animals were housed in either a normoxic (21% O₂; group N, n = 46) or hyperoxic (90% O₂; group H, n = 117) environment for 48 h. A detailed description of the isolated perfused lung preparation is provided elsewhere (31). Briefly, the isolated lung was fixed supine on a microscopic stage and perfused at a constant recirculating flow rate of 15 ml/min. Krebs-Henseleit solution with 3% BSA was used as the perfusate, and the hematocrit (Hct) was adjusted to 6.0 ± 0.5% by adding fresh blood from donor rats and erythrocytes stained with FITC, as described below. Blood remaining within the lung was expelled into the perfusion circuit during recirculation and used as part of the perfusate. To prevent the movement produced by artificial ventilation, the trachea was ligated in the end-inspiratory position, and gas exchange was maintained with an extracorporeal membrane oxygenator (ECMO; Merasilox-S; Senko, Tokyo, Japan). A gas mixture containing 21% O₂ and 5% CO₂ was used as the control gas flowing into the ECMO, allowing adjustment of perfusate PO₂ to 141 ± 3 mmHg, PCO₂ to 36 ± 2 mmHg, and pH to 7.42 ± 0.02. A warmed, humidified gas mixture containing the same gas composition as used for the ECMO was supplied to the lung surface to maintain a temperature of 37 ± 0.5°C and to avoid lung surface desiccation. Mean pulmonary arterial pressure (Ppa) was continuously monitored by force displacement of pressure transducers (TP-400T; Nihon Koden, Tokyo, Japan).

Experimental protocols. After pulmonary hemodynamics stabilized, the gas flowing into the ECMO and blown onto the lung surface was switched from the control gas to the following mixtures for 15 min: 1) hypoxic-normocapnic gas (2% O₂ and 5% CO₂ in N₂), which reduced PO₂ to 34 ± 2 mmHg without changing PCO₂ and pH levels in the perfusate; and 2) normoxic-hypercapnic gas (21% O₂ and 15% CO₂ in N₂), which increased perfusate PCO₂ to 95 ± 3 mmHg and decreased pH to 7.1 ± 0.04 without altering perfusate PO₂. Changes in Ppa from the baseline under given experimental conditions were used to measure overall vascular resistance changes in pulmonary circulation, including intra-acinar microvessels and large extra-acinar vessels. Measurements of pulmonary hemodynamic changes during hypoxia and hypercapnia were made at high and low basal tones of the pulmonary vasculature. High basal tone was established by administering synthetic thromboxane A₂ [TXA₂; ONO-11113 (9,11-epithio-11,12-methanothromboxane A₂); Ono, Tokyo, Japan] into the perfusate [TX(+) group]. Low basal tone measurements were made in the absence of TXA₂ [TX() group]. In the TX(+) experiments, 3.5 ml of TXA₂ solution prepared at 0.1 µg/ml was slowly added to the perfusate over 5 min to obtain a stable Ppa transition throughout observations. The final concentration of TXA₂ in the perfusate was adjusted to 5 ng/ml.

In the TX() experiments, hypoxia- or hypercapnia-induced changes in Ppa and intra-acinar microvessel diameters in normoxia-exposed lungs (group N) were estimated in the absence of addition of any exogenous agents (n = 18 and 20 hypoxia and hypercapnia rats, respectively), whereas the changes in hyperoxia-injured lungs (group H) were analyzed under four different experimental conditions: 1) condition C (n = 14 and 15 hypoxia and hypercapnia rats, respectively), in which no agents were administered; 2) condition L (n = 10 and 12 hypoxia and hypercapnia rats, respectively), in which a constitutive form of NO synthase (ecNOS) and an inducible form of NO synthase (iNOS) were concomitantly inhibited with N-nitro-L-arginine methyl ester (L-NAME; 100 µmol/l); 3) condition I (n = 9 and 11 hypoxia and hypercapnia rats, respectively), in which indomethacin (20 µmol/l) was used to inhibit both constitutive and inducible cyclooxygenase (COX-1 and COX-2); and 4) condition LI (n = 7 and 9 hypoxia and hypercapnia rats, respectively), in which measurements were made in the presence of both L-NAME and indomethacin.

In the TX(+) experiments, group N during hypoxia and hypercapnia was measured agent free, with neither L-NAME nor indomethacin administered (n = 4 hypoxia and hypercapnia rats each), whereas group H was measured under three different experimental conditions: 1) condition C (n = 5 hypoxia and hypercapnia rats each), in which no agents were administered; 2) condition L (n = 5 hypoxia and hypercapnia rats each), which was in the presence of L-NAME; and 3) condition I (n = 5 hypoxia and hypercapnia rats each), in which indomethacin was administered.

Real-time measurement of acinar microvessel kinetics. Erythrocytes were stained with FITC (Sigma, St. Louis, MO). Fresh rat blood was centrifuged at 1,000 rpm, and the buffy coat was discarded. The packed erythrocyte solution was diluted with PBS, and FITC was added to give a final concentration of 0.1 mg/ml. This solution was then incubated at 37°C for 30 min and washed three times in PBS. A 1-ml volume of FITC-labeled erythrocyte solution was added to the perfusate, enabling determination of the direction of flow in the pulmonary microcirculation. The microvessels from which FITC-erythrocytes entered the capillary network were defined as arterioles, and the microvessels into which FITC-erythrocytes flowed from the capillary network were defined as venules. Microvessel diameter was measured by adding 200 µl of 5% FITC-dextran, which has a molecular weight of 145,000 (Sigma). These microvessel events in the acinus were studied precisely by using a recently developed real-time confocal laser luminescence microscope (33). Fluorescent emission from the specimen was imaged onto a high-sensitivity charge-coupled-device camera with an image intensifier (EktaPro Intensified Imager VSG; Kodak, San Diego, CA). By applying an excitation wavelength of 488 nm emitted from a low-power argon laser (532-BSA04; Omnichrome, Chino, CA), confocal units enabled us to obtain apparently instantaneous images at 1,000 frames/s. The resulting field of view was 210 × 210 µm, corresponding roughly to the diameter of a single pulmonary microvessel adjacent to the terminal bronchiole (33, 34), yielding images creditably detecting events in a single acinus. We recorded confocal images at 250 frames/s with a high-speed video analyzer (EktaPro 1000 Processor; Kodak). The direction of FITC-erythrocyte flow in the acinar microcirculation was determined during replays of the videotapes at the normal video rate. Each replay frame represented events during a 4-ms interval. We measured the diameter of each microvessel by processing a confocal video image with a digital image analyzer (Quadra 840AV/Image 1.58; Apple, Cupertino, CA).

Measurements of NO-related metabolites in the perfusate. A 2-ml perfusate sample was centrifuged and frozen at 80°C immediately after collection. To measure NO production in the lung, we studied the total concentration of end products of NO metabolism, NO₂ and NO₃, by the modified method of Green et al. (10). To deproteinize the perfusate sample, a 500-µl sample was mixed with 100 µl of 35% sulfosalicylic acid. The sample was then centrifuged, and 300 µl of 5% NH₄Cl and 60 µl of 5% NaOH were added to 300 µl of the supernatant. The prepared sample was passed through a high-pressure Teflon column packed with copper-plated cadmium metal allowing reduction of NO₃ to NO₂. The effluent was mixed with the reagent containing sulfanilamide and N-(1-naphthyl)ethylenediamine in phosphoric acid, and the color of the product yielded by the diazotization reaction was examined at 0°C with a spectrophotometer (US 501; Unisoku, Osaka, Japan) at an absorbance wavelength of 546 nm.

Measurements of prostacyclin metabolite in the perfusate. A 2-ml perfusate sample collected in a tube containing indomethacin and EDTA was centrifuged and frozen at 80°C until extraction. Measurements were made by extracting 1 ml of the sample into medium containing 4 ml of ethyl acetate. The ethyl acetate layer was transferred to a second tube and evaporated to dryness with pure N₂ gas. The dried extract was reconstituted in the assay buffer, and the concentration of immunoreactive 6-ketoprostaglandin F₁ (6-keto-PGF₁), the stable metabolite of prostacyclin (PGI₂), was measured by ELISA (EIA kit; Cayman Chemical, Ann Arbor, MI).

Statistical evaluation. The statistical significance of differences was generally judged by one-way ANOVA followed by multiple comparison Scheffés analysis. Ppa increments, microvessel diameter changes, and perfusate concentrations of NO and PGI₂ metabolites on hypoxia and hypercapnia during exposure to the same agent were compared by applying an unpaired t-test. The effects of TXA₂ on the indexes, including baseline Ppa, microvessel diameter, and concentrations of vasoactive substances, were estimated by an unpaired t-test. Changes in Ppa, microvessel diameter, and NO and PGI₂ metabolites after a given stimulation were assessed by a paired t-test. Values are means ± SD, with P < 0.05 statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Pulmonary hemodynamics before hypoxic and hypercapnic stimulation in the absence of TXA₂. Because many different experimental conditions were employed, we defined "agent free" as a condition in which neither L-NAME nor indomethacin was added to the perfusate, irrespective of the presence or absence of TXA₂. "Baseline" is used to indicate conditions before hypoxic or hypercapnic stimulation.

Before exposure to hypoxia and hypercapnia, Ppa did not differ in groups N and H, and the values averaged 12.3 ± 2.5 mmHg. The baseline diameter of precapillary arterioles and postcapillary venules before hypoxia or hypercapnia did not differ between the groups, and ranged between 20 and 40 µm. Baseline capillary diameter was 6-7 µm, with no difference between the groups. Administration of L-NAME or indomethacin did not alter baseline Ppa or microvessel diameter.

Hypoxia- and hypercapnia-induced microvessel diameter changes in agent-free groups N and H in the absence of TXA₂. Hypoxia enhanced Ppa in association with a reduction in arteriole diameter in group N (Table 1; Figs. 1 and 2). Hypoxia-induced Ppa increments in group H, however, were conspicuously smaller than in group N (Table 1). Although arteriole diameter in group H before hypoxic stimulation did not differ from the diameter in group N, hyperoxia-injured arterioles did not constrict during hypoxia (Fig. 2). Neither venule nor capillary diameter in group N or H was altered by hypoxic stimulation (Fig. 2).

View larger version (145K): [in this window] [in a new window]   Fig. 1.   Confocal views of hypoxia-induced arteriole constriction and hypercapnia-induced venule dilation in an intact acinus without addition of any agents, including thromboxane A2 (TXA2; 20× objective). Top: arteriole before (A) and after (B) hypoxia. Bottom: venule before (C) and after (D) hypercapnia.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Arteriole (A) and venule (B) responses to hypoxia and hypercapnia in rats exposed for 48 h to 21% (group N) or 90% O2 (group H) in the absence of TXA2. Data are absolute diameter changes expressed in µm. C, agent free; L, N -nitro-L-arginine methyl ester (L-NAME) administration; I: indomethacin administration. LI: concurrent administration of L-NAME and indomethacin. * Active vasoconstriction or vasodilation; # significantly different from group N under condition C; + significantly different from group H under condition C; $ significantly different from group H under condition L; & significantly different from group H under conditions L and LI.

Exposure to hypercapnia increased Ppa in both groups, but the degree of the Ppa increment was significantly greater in group H (Table 1). Hypercapnia elicited distinct venule dilation in group N but no change in arteriole or capillary diameter (Figs. 1 and 2). This hypercapnia-induced venule dilation was not seen in group H (Fig. 2). Arteriole and capillary diameter did not change in group H during hypercapnia (Fig. 2).

Effects of NOS inhibition on hypoxia- and hypercapnia-induced microvessel diameter changes in group H in the absence of TXA₂. L-NAME administration yielded a significant Ppa increment during hypoxia, comparable to the increment in group N (Table 1). L-NAME concomitantly restored hypoxic arteriole constriction without any change in venule or capillary diameter (Fig. 2).

The hypercapnia-induced increase in Ppa in the presence of L-NAME was smaller than in its absence (Table 1). Contrary to our expectation, L-NAME augmented venule and arteriole dilation during hypercapnia (Fig. 2).

Effect of COX inhibition on hypoxia- and hypercapnia-induced microvessel diameter changes in group H in the absence of TXA₂. Although indomethacin restored arteriole constriction with hypoxia, its extent was smaller than that of L-NAME (Fig. 2). Indomethacin did not improve hypoxia-induced Ppa changes (Table 1), nor did it alter venule and capillary diameter after hypoxia (Fig. 2).

Compared with the agent-free results, administration of indomethacin alone did not alter the hypercapnia-induced Ppa increment in association with no change in arteriole, venule, or capillary diameter (Table 1; Fig. 2).

Effect of simultaneous inhibition of NOS and COX on hypoxia- and hypercapnia-induced microvessel diameter changes in group H in the absence of TXA₂. Compared with the results with L-NAME alone, no additional improvement in microvessel constriction or overall Ppa increment on hypoxia was seen in response to concurrent L-NAME and indomethacin administration (Table 1; Fig. 2).

The augmentation of hypercapnia-induced venule and arteriole dilation observed in group H in the presence of L-NAME alone was inhibited in the concomitant presence of L-NAME and indomethacin (Fig. 2). Furthermore, simultaneous administration of both agents enhanced the Ppa increment to a level equal to that under agent-free conditions (Table 1).

Effects of TXA₂ on microcirculatory hemodynamics in groups N and H during hypoxic and hypercapnic gas breathing. TXA₂ analog increased baseline Ppa by a comparable amount in groups N and H under agent-free conditions (increment in group N, 3.6 ± 3.2 mmHg; in group H, 2.2 ± 0.9 mmHg). Group H baseline Ppa was enhanced by 4.2 ± 4.7 mmHg by TXA₂ concomitant with L-NAME and by 2.2 ± 1.4 mmHg in the presence of TXA₂ and indomethacin. These increments were not different from the values observed under agent-free conditions.

TXA₂ distinctly augmented hypoxia-induced Ppa changes under all experimental conditions studied for group H, decreasing differences in the hypoxia-induced overall pressor response observed in the absence of TXA₂ (Table 2). The extent of arteriole constriction on hypoxia in the presence of TXA₂ did not differ among the experimental conditions in group H (Fig. 3). The hypoxia-induced arteriole constriction in group H was not different from that obtained in group N (Fig. 3). Neither venule nor capillary diameter was modified by hypoxia under TXA₂ conditions in group H.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Arteriole (A) and venule (B) responses to hypoxia and hypercapnia in groups N and H in the presence of TXA2. Data are changes in diameter (µm). * Active vasoconstriction or vasodilation; # significantly different from group N under condition C; + significantly different from group H under condition C; $ significantly different from group H under condition L.

The hypercapnia-induced Ppa increments in groups N and H were larger in the presence of TXA₂ than in the absence of TXA₂ (Tables 1 and 2). The Ppa increase in response to hypercapnia with TXA₂ in group H was even larger than in group N (Table 2). The Ppa increase in the presence of L-NAME in combination with TXA₂ in group H was much smaller than obtained under agent-free conditions or in the presence of indomethacin (Table 2). The extent of hypercapnia-induced venule dilation in group N with TXA₂ did not differ from that obtained without TXA₂, whereas arteriole constriction was observed during hypercapnia in group N when TXA₂ was administered (Fig. 3). Hypercapnia-induced venule dilation was not seen in group H with TXA₂, but group H arterioles showed constriction in response to hypercapnia when TXA₂ was present (Fig. 3). Hypercapnic stimulation in the presence of TXA₂ evoked venule and arteriole dilation in group H when L-NAME was added (Fig. 3). In the presence of TXA₂, the hypercapnia-induced microvessel diameter changes with indomethacin were quantitatively the same as those observed when agent free (Fig. 3).

Perfusate concentrations of NO and PGI₂ in groups N and H in the absence of TXA₂. In the agent's absence, the baseline perfusate NO concentration in group H was 1.7× that in group N (Table 3). Hypoxia augmented NO production in both groups, but hypercapnia did not. L-NAME suppressed baseline NO production and the increase after hypoxia in group H.

The baseline concentration of perfusate 6-keto-PGF₁ under agent-free conditions was higher in group H than in group N (Table 4). Agent-free 6-keto-PGF₁ production was enhanced during hypoxia and hypercapnia in both groups. Indomethacin reduced the baseline 6-keto-PGF₁ concentration in group H to a level comparable to that in group N (Table 4). Indomethacin inhibited the 6-keto-PGF₁ increase after hypoxia and hypercapnia in group H. The influence of L-NAME on 6-keto-PGF₁ production in group H differed depending on whether the conditions were hypoxic or hypercapnic, i.e., hypoxia augmented 6-keto-PGF₁ production in the presence of L-NAME, but hypercapnia did not (Table 4).

View this table: [in this window] [in a new window]   Table 4.   Perfusate concentration changes in 6-keto-PGF1 during exposure to hypoxia and hypercapnia in the absence of TXA2

Perfusate concentrations of NO and PGI₂ in groups N and H in the presence of TXA₂. Perfusate concentrations of NO metabolites before and after hypoxia obtained in the presence of TXA₂ did not differ from those in the absence of TXA₂ in groups N and H. Qualitatively the same trend was observed in hypercapnic experiments. L-NAME reduced NO concentrations in both groups.

Concentrations of 6-keto-PGF₁ under baseline conditions were higher in the presence of TXA₂ than in the absence of TXA₂ in groups N and H (Table 5). Both hypoxia and hypercapnia in the presence of TXA₂ enhanced 6-keto-PGF₁ production under agent-free conditions in group H, but the enhancement was inhibited by indomethacin (Table 5). In the concomitant presence of TXA₂ and L-NAME, hypoxic stimulation augmented 6-keto-PGF₁ production in group H, but hypercapnic stimulation did not (Table 5).

View this table: [in this window] [in a new window]   Table 5.   Perfusate concentration changes in 6-keto-PGF1 during exposure to hypoxia and hypercapnia in the presence of TXA2

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Critique of methods. We perfused the isolated lung with buffer containing a small quantity of fresh blood, just enough to delineate the outlines of intra-acinar microvessels on confocal observation. This is because we wanted to preclude constriction and dilation elicited by vasoactive substances released from leukocytes and platelets in the blood. As shown by Barnard et al. (3) and Wilson et al. (30), however, the extent of NO and prostaglandin (PG) production in the lung may be closely related to vascular wall shear force, which depends on perfusate viscosity, and in turn is mainly governed by the Hct. They (3, 30) examined the vascular response to NOS inhibitors and perfusate concentrations of COX-associated products and reported that NO and PG production in the intact lung is enhanced in parallel with increasing perfusate viscosity. Their findings (3, 30) suggest that our experiments were substantially flawed in assessing the important roles of vasoactive compounds in pulmonary microvessels because of the significantly lower Hct than in vivo. To overcome these obstacles, we estimated microvessel kinetics in both the presence and absence of the stable TXA₂ analog, allowing shear force along the microvessel wall to change. In fact, TXA₂ augmented baseline Ppa, the increments in Ppa after hypoxic or hypercapnic stimulation, and perfusate concentrations of 6-keto-PGF₁ in group N compared with the values obtained in the absence of TXA₂ (Tables 2 and 5). Although the overall perfusate concentrations of NO were not increased by addition of TXA₂ in group H, the L-NAME-associated Ppa increments in group H during hypoxia were much larger in the presence of TXA₂ than in the absence of TXA₂ (Tables 1 and 2). These results suggest that addition of TXA₂ increases NO production locally in the lung, but this is not always reflected in the overall circulating medium. These findings appear to be consistent with those reported by Barnard et al. (3) and Wilson et al. (30).

The direction of the microvessel diameter changes in response to hypercapnia under various experimental conditions in groups N and H did not differ qualitatively according to whether TXA₂ was present or absent. Meanwhile, fine differences in hypoxia-induced microvessel diameter changes between agent-free conditions and conditions in which L-NAME and/or indomethacin were present but TXA₂ was absent were not observed in the presence of TXA₂ (Figs. 2 and 3). These findings seem to indicate that vasodilator production in the lung is actually augmented by TXA₂ but that its effect on microcirculatory kinetics is masked by the constrictive effect of TXA₂, especially during hypoxic stimulation. This suggests that the lung preparation perfused with low-Hct medium in the absence of TXA₂ is superior to the preparation in the presence of TXA₂ when attempting to detect subtle roles of vasodilators in modifying the vascular reactivity of acinar microvessels. Further study, however, is necessary to determine whether our experimental results regarding microvessel kinetics can actually be applied in vivo, where the Hct is much higher than in the present experiments. Unless otherwise specified, the discussions below focus on the acinar microvessel response to hypoxia and hypercapnia without TXA₂.

Abnormal response of hyperoxia-injured microvessels to hypoxia. Attenuated hypoxia-induced arteriole constriction in hyperoxia-injured acini was caused independently by increased production of vasodilating NO and PGs, because inhibition of either NOS or COX allowed arterioles to regain contractility during hypoxic stimulation (Fig. 2). The importance of NOS and COX was supported by the increased concentrations of NO and PGI₂ in the perfusate excreted from the hyperoxia-injured lung (Tables 3 and 4). The vasodilating substances yielded by NOS, however, appear to be more significant in blunting hypoxia-induced arteriole constriction than those yielded by COX (Fig. 2). Hypoxia-induced overall Ppa changes differed quantitatively between L-NAME and indomethacin; i.e., L-NAME markedly improved Ppa changes after hypoxia in hyperoxia-injured lungs, whereas indomethacin did not (Table 1). Because the overall Ppa increment reflects the sum of the pressor response imposed by all intra-acinar and extra-acinar vasculature (34), microvascular analysis, together with Ppa measurements, indicates that NOS upregulation in hyperoxia-injured lungs occurs along both intra-acinar and extra-acinar arterioles, whereas COX upregulation is confined to intra-acinar regions. Our findings are qualitatively consistent with in vitro studies reported by Liao et al. (20) and Lee et al. (19). Liao et al. (20) demonstrated an increase in ecNOS mRNA in bovine pulmonary artery endothelial cells exposed to hyperoxia, and Lee et al. (19) reported that exposure to hyperoxia upregulated COX in perinatal rat lung cells.

Abnormal response of hyperoxia-injured microvessels to hypercapnia. In intact acini, the hypercapnia-induced vasodilation was restricted to venules (Figs. 1 and 2) in association with an increase in perfusate 6-keto-PGF₁ but not in NO metabolites (Tables 3 and 4). We recently demonstrated that vasodilating PGs are the main substances mediating hypercapnia-induced venule dilation in intact rat acini, whereas NO has little impact on it (34). In the present study, we found that hypercapnia-induced venule dilation in hyperoxia-injured lungs was largely attenuated under agent-free conditions but that it was unexpectedly restored by NOS inhibition, irrespective of the presence or absence of TXA₂ (Figs. 2 and 3). These phenomena were no longer observed when NOS and COX were inhibited concurrently (Fig. 2). Taken together, our findings strongly suggest that COX activity causing hypercapnia-induced venule dilation is distinctly concealed, probably by COX inhibition by NO. This concept may explain the functional equivalence between the agent-free conditions and conditions in which COX is inhibited by indomethacin, neither of which caused constriction or dilation in hyperoxia-injured venules on hypercapnia (Fig. 2). Because COX is a hemoprotein containing iron, NO is expected to alter COX activity by interacting with iron at the active site. Several groups that used purified COX-1 and COX-2 enzymes have reported dual effects of NO on COX activity in vitro (7, 15, 18, 22, 25, 26), i.e., inhibition and excitation of COX by NO. COX-1 activity is maintained by ferric irons, and COX-2 activity is maintained by ferrous irons (15, 26). Tsai et al. (26) demonstrated that NO is a strong ligand for ferrous irons and that COX-2 (ferrous form) activity is almost completely inhibited by the presence of NO under anaerobic conditions. Kanner et al. (15) found that ferric COX-1 activity is largely inhibited by NO because of its ability to reduce the ferric enzyme to the ferrous form and because of competition for iron sites available for exogenous ligands. These studies indicate that NO inhibits the activity of the respective COX isoform. Landino et al. (18) reported that NO significantly enhances the activity of purified COX-1 and COX-2 when both NO and superoxide production is augmented. This is because peroxynitrite, a coupling product of NO and superoxide, specifically modulates both COX isoforms, indicating that NO-exciting COX activity must be taken into account in oxidative states. Given our experimental conditions, hyperoxia exposure is expected to enhance the generation of reactive oxygen species (2, 12, 31) and NO in the lung (Table 3), suggesting that both inhibitory and excitatory effects of NO on COX isoforms must be taken into account under our experimental conditions. Our findings of restoration in hypercapnia-induced venule dilation by NOS inhibition (Fig. 2) are consistent with the notion that the inhibitory effect of NO on COX activity exceeds its excitatory effect, at least in hyperoxia-injured venules under hypercapnia. Negative modulation of COX pathways by NO may not, however, completely explain the hypercapnia-induced changes in perfusate 6-keto-PGF₁ concentrations during NOS inhibition in hyperoxia-injured lungs. Hypercapnia did not increase perfusate 6-keto-PGF₁ concentrations when NOS was inhibited (Table 4). This requires NO-dependent COX excitation, suggesting that the inhibitory effect of NO on COX is confined to intra-acinar regions where venules are located, whereas the excitatory effect of NO on COX may dominate in other lung regions.

Although we previously demonstrated that intact arterioles are not sensitive to hypercapnia (34), the present study showed that hyperoxia-injured arterioles dilated on exposure to hypercapnia when NOS was inhibited but that this arteriolar dilation was lost when NOS and COX were simultaneously inhibited (Fig. 2). These findings suggest that hyperoxia-injured arterioles basically become more reactive to vasodilating PGs, but this is generally masked by the negative effect of excessive NO on the COX pathway. The unexpected reduction in overall Ppa increment during hypercapnia with L-NAME in hyperoxia-injured lungs may be attributable to enhanced venule and arteriole dilation under L-NAME conditions (Fig. 2; Table 1). Interestingly, negative crosstalk between NOS and COX pathways does not appear evident during hypoxic stimulation (Fig. 2; Table 1), indicating that hypercapnia per se and/or hypercapnia-induced acidosis is indispensable for eliciting NO-dependent reduction in COX activity in acinar microcirculation. To the best of our knowledge, this is the first study demonstrating that negative modulation of COX pathway by NO plays an important role in distorting microvascular response to hypercapnia in acinar regions injured by hyperoxia exposure.

Our findings have valuable clinical implications, i.e., NOS inhibitors administered while treating patients with injured lungs by permissive hypercapnia improves gas exchange by reducing alveolar edema via relieving venule constriction. Meanwhile, the improvement in hypoxia-induced arteriole constriction by NOS inhibitors is counteracted by enhanced hypercapnia-induced arteriole dilation. COX inhibitors may not be as effective as NOS inhibitors, because COX inhibition is latently achieved by enhanced NO production in injured acinar microvessels.

In conclusion, although the diameter of intact intra-acinar arterioles varies with changes in ambient PO₂, this is blunted in the microcirculation of lungs injured by exposure to hyperoxia. The blunted arteriole responsiveness to O₂ is caused by enhanced production of both NO and PGs along microvessel walls, with NO seeming to be more potent than PGs. Although the hypercapnia-induced vasodilator response of hyperoxia-injured venules is basically the same as that of normal venules, this is apparently masked by the inhibitory effect of excessive NO on COX activity. Injured arterioles show a greater vasodilator response to hypercapnia than normal arterioles, but this is also masked by the negative interaction between NOS- and COX-dependent pathways. Crosstalk between these two pathways is important in hypercapnia with both NOS and COX upregulated, whereas this mechanism has little influence in hypoxia.

Perspectives