Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Acute pulmonary thromboembolism is a common occurrence, with more than 500,000 episodes and 140,000 deaths attributed, at least in part, to this problem each year in the United States (1). To investigate the mechanism of lung injury in acute pulmonary embolism, we made nonhypoxic lung ischemia by clamping the unilateral pulmonary artery of a rat and observed the pulmonary circulation by means of a fluoro-imaging technique.

It is well known that ischemia depletes adenosine triphosphate (ATP) in cells (2, 3), which leads to the activation of xanthine oxidase (3). The subsequent reperfusion supplies oxygen to the ischemic cells, causing the activated xanthine oxidase to release superoxide anion and other toxic oxygen metabolities (3, 4). These toxic oxygen metabolites play a major role in ischemia-reperfusion (I/R) injury (4). Although reperfusion after pulmonary artery occlusion is associated with cytokine release (5, 6), neutrophil infiltration in lungs (5), microvascular injury (7), and pulmonary edema (7, 8), the events during ischemia are not fully understood, especially the role of the neutrophil.

I/R injury was attenuated by a monoclonal antibody (mAb) against intercellular adhesion molecule-1 (ICAM-1) (9, 10). I/R upregulates ICAM-1 expression (11, 12). Monoclonal antibodies against Mac-1 and ICAM-1 reduced I/R-induced leukocyte adherence (10, 13). It was also reported that neutrophil adherence dependent on Mac-1 and ICAM-1 activates the neutrophil respiratory burst (14). It was reported that phospholipase A₂ was activated during ischemia in the intestine (15) and I/R in the kidney (16). Phospholipase A₂ leads subsequent platelet-activating factor (PAF) synthesis.

PAF enhances the neutrophil respiratory burst (17). PAF itself also causes neutrophil superoxide anion production (20).

On the other hand, the lung ischemia is not synonymous with hypoxia if ventilation continues, supplying oxygen to the airway. Several reports have indicated that reperfusion is not necessary to cause lung injury in buffer-perfused lungs (24). Therefore, we hypothesized that reperfusion is not necessary in lung injury of live animals, and nonhypoxic ischemia without reperfusion alone may cause oxygen radical production and pulmonary microvascular injury. However, there is a lack of direct evidence of oxygen radical production and lung injury in vivo. In this study we applied a digital fluoro-imaging technique, which permits detection of hydrogen peroxide (H₂O₂) directly in the intact rat pulmonary circulation (27), to the nonhypoxic ischemia rat lung. We also determined the role of ICAM-1 and PAF using ICAM-1-neutralizing mAb 1A29 and PAF receptor-specific antagonist CV6209.

In this study, H₂O₂ was detected directly in leukocytes within the capillaries of the nonhypoxic ischemia rat lung. We also demonstrated that pulmonary artery occlusion alone increased microvascular permeability (MVP) of the rat lung and that discontinuing the oxygen supply from the airway in advance of pulmonary artery occlusion, ICAM-1 blocking, and PAF receptor antagonist attenuated both H₂O₂ production with the MVP change.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Reagents

2',7'-dichlorofluorescein-diacetate (DCFH-DA), fluorescein isothiocyanate (FITC)-dextran conjugate (molecular weight 70,000), and Evans blue were purchased from Molecular Probes, Inc. (Eugene, OR), Sigma Chemical Co. (St. Louis, MO), and Wako Chemicals (Tokyo, Japan), respectively. The PAF-specific antagonist CV6209 (28) was kindly provided by Takeda Pharmaceutical Co. (Osaka, Japan).

Production and Purification of mAb against Rat ICAM-1, 1A29

Hybridoma clone 1A29 (29) was kindly provided by Dr. Miyasaka of Osaka University (Osaka, Japan). Hybridomas were grown in RPMI 1640 containing 10% fetal calf serum (FCS; Gibco, New York, NY), 50 mM 2-mercaptoethanol, 100 U/ml penicillin, and 100 mg/ml streptomycin. A BALB/c mouse was immunized with an intraperitoneal injection of 2-3 × 10⁶ 1A29 cells. Ascites were collected and mAb 1A29, which recognizes rat ICAM-1, was purified using an Ampure^TM PA kit (Amersham K.K., Tokyo, Japan), according to the manufacturer's instructions. For the neutralizing experiment, mAb 1A29 was dialyzed against phosphate-buffered saline (PBS) (pH 7.0) and kept at 4°C until the experiment.

Animal Preparation and Observation of Intact Pulmonary Circulation

Male Wistar rats weighing 200 to 300 g were anesthetized by an intraperitoneal injection of pentobarbital sodium (50 mg/kg). To eliminate artifacts due to movements, and to block spontaneous breathing movements, paralysis was induced with 0.15 mg/kg of pancuronium bromide and maintained with smaller doses. The right femoral vein was cannulated with a polyethylene tube for infusion of agents. The rat was continuously infused with saline at a rate of 3 ml/kg/h. The femoral artery was cannulated to continuously monitor systemic arterial pressure. A polyethylene cannula (diameter 2.5 mm) was inserted into the trachea. The rat was ventilated with a mechanical ventilator (EVM-50A; Aika, Tokyo, Japan) delivering a respiratory volume/ min of 0.5 ml/g at 80/min, FI_O2 = 0.21. A median incision was made which extended to the left lateral thoracotomy along the eighth rib. An interlobar site of the left upper lobe was gently bonded to the glass chamber with the bonding agent Alonalpha (Toa Gosei Chemical Industry, Tokyo, Japan). The intact pulmonary microcirculation was observed by intravital fluorescence microscopy (BH2-FRC; Olympus, Tokyo, Japan). Fluorescence images were displayed on a video monitor (Sony, Tokyo, Japan) through a silicon-intensified target camera (C2741-08; Hamamatsu Photonics, Shizuoka, Japan) and recorded on a video cassette recorder (EDV900; Sony).

In Vivo Detection and Measurement of H₂O₂ in the Lung

We visualized and measured H₂O₂ production in the intact pulmonary circulation by a digital fluoro-imaging technique (27). On the day of the experiment, DCFH-DA compound was suspended in physiological saline. The rats were infused with 1 mg/body of DCFH-DA suspension intravenously. Fifteen minutes after DCFH-DA infusion, the intact pulmonary microcirculation was observed by fluorescence microscopy at excitation and emission wavelengths of 490 and 530 nm (excitation filter BP490, dichroic mirror and barrier filter DM500; Olympus). We used a C-mount lens (×3.3) and an objective lens (×4) (SPlan; Olympus) for image analysis, and ×40 (ULWCD Plan; Olympus) for high-power field observation. When the image was recorded, ventilation was stopped for 5 s at the end of the expiration to eliminate lung movements. For high-power microscopy, 0.5 ml of 0.01% FITC-dextran conjugate (molecular weight 70,000) was also injected to achieve more detailed anatomic information. After the experiment, the recorded images were digitized and recorded on a hard disk using an image digitizing card (FRM512; Photoron, Tokyo, Japan), and analyzed with an NEC 9801 VX personal computer (NEC, Tokyo, Japan) and image-analyzing software (RIPP2; Photoron). The image resolution was 512 × 512 with 256 grey levels. The 2',7'-dichlorofluorescin (DCF) fluorescent area in the image was selected by RIPP2 image-analyzing software and expressed as the number of pixels in five images. We defined this value as H₂O₂ production in pulmonary microcirculation. The specificity of the DCF reaction with H₂O₂ was confirmed by blocking in the presence of catalase. (Four rats were infused with 5,000 U/kg of catalase 20 min before left pulmonary artery occlusion.)

Experimental Groups

The rats were divided into the following groups:

1. The left pulmonary artery occlusion (PAO) group (n = 5). After recording of the baseline DCF images, lung ischemia was induced by clamping of the left pulmonary artery. Because the bronchial circulation can attenuate the I/R lung injury in the presence of bronchopulmonary anastomoses (30, 31), we confirmed by direct microscopic visualization that the blood flow in the pulmonary circulation was abolished. The rat was ventilated with a mechanical ventilator during the experiment, so that left lung ischemia was induced without hypoxia. DCF fluorescent images were then recorded every 15 min for 90 min.

2. The left main bronchus occlusion (BO) group (n = 5). After the recording of baseline DCF images, lung collapse was induced by clamping the left main bronchus. After that, DCF fluorescence images were recorded at 90 min.

3. BO + PAO group (n = 5). The left main bronchus was occluded 60 min before left pulmonary artery occlusion. After occlusion of the left pulmonary artery, DCF fluorescence images were recorded at 90 min.

4. Control group (n = 5). DCF fluorescence images were recorded every 15 min for 90 min.

5. 1A29 group (n = 4). The rats were infused with 2 mg/kg of mAb 1A29 against rat ICAM-1 20 min before left pulmonary artery occlusion. After occlusion of the left pulmonary artery, DCF fluorescence images were recorded at 90 min.

6. N-IgG group (n = 4). The rats were infused with 2 mg/ kg of nonspecific mouse IgG 20 min before left pulmonary artery occlusion. After occlusion of the left pulmonary artery, DCF fluorescence images were recorded at 90 min.

7. CV6209 group (n = 4). The rats were infused with 1 mg/ kg of PAF receptor-specific antagonist CV6209 20 min before left pulmonary artery occlusion. After occlusion of the left pulmonary artery, DCF fluorescence images were recorded at 90 min.

8. Catalase group (n = 4). The rats were infused with 5,000 U/kg of catalase (Tokyo Kasei Industry, Tokyo, Japan) 20 min before left pulmonary artery occlusion. After occlusion of the left pulmonary artery, DCF fluorescence images were recorded at 90 min.

Lung Tissue Oxygen Content

The lung oxygen content of rats in the PAO and BO + PAO groups were measured with a tissue PO₂ monitor (PO₂-100; Inter Medical Co., Ltd., Osaka, Japan). The needle probe (POE-10N; Inter Medical Co.) was inserted into the left upper lobe of each rat lung. The probe was placed where the oxygen content in the lung was lowest during the baseline recording. The oxygen content of the lung was continuously recorded during the experiment. Occlusion of the left pulmonary artery or the main bronchus was performed after recording the baseline value.

Lung MVP

Pulmonary MVP in the eight groups was measured using a modification of the Evans blue dye extravasation technique was previously described (32). Animals received 1,000 U/kg of heparin to prevent coagulation. Ten mg/kg of Evans blue (pH 7.4) was injected into the femoral vein cannula just prior to occlusion. At the time of death, a blood sample was taken from the right ventricle and plasma was removed after centrifugation. The lungs were then perfused free of blood with 20 ml of 0.9% normal saline. Lungs were removed from the thoracic cavity and weighed. Evans blue was extracted from pulmonary tissues after homogenization in 3 ml of 0.9% normal saline. This volume was added to 2 vol deionized formamide and incubated at 60°C for 12 h. The supernatant was separated by centrifugation at 2,000 × g for 30 min. Evans blue in the plasma and lung tissue was quantified by dual wavelength spectrophotometric analysis as described by Linderkamp and colleagues (33). This method corrects the sample absorbance at 620 nm for the absorbance of contaminating heme pigments by the following formula: corrected absorbance at 620 nm = (actual absorbance at 620 nm)  (1.426 [absorbance at 740 nm] + 0.03).

The Evans blue in the pulmonary tissues was then normalized to tissue weight. A permeability index (PI) was calculated by dividing the corrected pulmonary tissue Evans blue absorbance at 620 nm/g of lung tissue by the corrected plasma Evans blue absorbance at 620 nm. The PI reflects the degree of extravasation of Evans blue into the extravascular pulmonary compartment.

Statistics

Data from various groups were expressed as means ± SEM. To determine the significance of differences between the control and multiple experimental groups, one-way analysis of variance in combination with Dunnette's multiple comparisons test was used. Statistical significance was defined as P < 0.05.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

H₂O₂ Production in the Nonhypoxic Ischemia Rat Lung

Figure 1 shows the DCF fluorescence images from a single representative experiment. DCF fluorescence was negligible in the pulmonary circulation of the control rat (Figure 1A), but marked in the nonhypoxic ischemia (PAO group) rat (Figure 1B). The spot of DCF fluorescence under low magnification was consistent with several round fluorescent spots in the pulmonary capillary under high-power viewing (Figure 1C). The diameter of these spots was 5- 10 µm. Red blood cells were not stained with DCF fluorescence (data not shown); therefore, these DCF fluorescence spots must be attributable to leukocytes. H₂O₂ production in the PAO group increased rapidly 15 min after pulmonary artery occlusion compared with the control group (P < 0.05) and continued to increase in a time-dependent manner (P < 0.05) (Figure 2).

View larger version (73K): [in this window] [in a new window]   Figure 1.   DCF fluorescence images of the rat pulmonary circulation of the control group and the PAO group. The rats were injected with 1 mg/body of DCFH-DA suspension. Ninety minutes after the left pulmonary artery was ligated, DCF fluorescence images were recorded through a silicon-intensified target TV camera. Minimal fluorescence was detected in the control group (panel a). A significant amount of fluorescence (white spots) was detected in the PAO group (panel b). Panel c shows a high-magnification image of DCF fluorescence on the FITC-dextran microangiography in the PAO group. Many globular fluorescent spots attributable to leukocytes were seen in the pulmonary microcirculation. The diameter of these spots was 5 to 10 µm, and the spots were always in the capillaries and did not move during the observation.

View larger version (17K): [in this window] [in a new window]   Figure 2.   Time course of H2O2 production in response to left pulmonary artery occlusion. Using a computer image analyzing system, H2O2 production was determined by counting the pixels of DCF fluorescence in the image of the pulmonary circulation through a ×4 objective lens and ×3.3 C-mount lens. Lung ischemia was induced by clamping the left pulmonary artery. The rats were ventilated with a mechanical ventilator during the experiment, so that left lung ischemia was induced without hypoxia (PAO group). H2O2 production in the PAO group increased rapidly compared with that in the control group (P < 0.05). Values are expressed as means ± SEM; n = 5 for each group. *Significant difference compared with the control group (P < 0.05).

Tissue Oxygen Contents in Lung

The oxygen content in the lung tissue was measured by a tissue Pt_O2 monitor in the PAO and BO + PAO groups. Figure 3 illustrates the typical time course of Pt_O2 in the lung. After the left pulmonary artery was clamped, Pt_O2 increased immediately to almost 100% above baseline (Figure 3A) in the PAO rat lung. After the left main bronchus was clamped, Pt_O2 decreased to 20 mm Hg above baseline; and after clamping of the left pulmonary artery, Pt_O2 immediately decreased to almost 0 mm Hg (Figure 3B). Therefore, we were able to remove the oxygen almost completely from the pulmonary circulation in the BO + PAO rat lung.

View larger version (11K): [in this window] [in a new window]   View larger version (8K): [in this window] [in a new window]   Figure 3.   Effect of pulmonary artery occlusion and main bronchus occlusion on lung oxygen content. To determine the effect of left main bronchus occlusion on lung oxygen content (PtO2) in the nonhypoxic ischemia rat lung, we ligated the left main bronchus 60 min before left pulmonary artery occlusion (BO + PAO group). Panels A and B depict PtO2 of the rats in PAO group and BO + PAO group, respectively. PtO2 was measured by means of a tissue PtO2 monitor. The figures show the typical continuous PtO2 recording in both the PAO and BO + PAO groups. We repeated these experiments at least three times for each group. After clamping of the left pulmonary artery, PtO2 increased immediately to almost 100% above baseline (panel A) in the PAO rat lung. After clamping of the left main bronchus, PtO2 decreased to 20 mm Hg below baseline; and after clamping of the left pulmonary artery, PtO2 immediately decreased to almost 0 mm Hg (panel B). PAO, left pulmonary artery occlusion; BO, left bronchus occlusion.

Inhibition of H₂O₂ Production in the Nonhypoxic Ischemia Rat Lung

Figure 4 shows the inhibition of H₂O₂ production due to pulmonary artery occlusion with various treatments. H₂O₂ production at 90 min was higher than that in the control group (P < 0.05). These in BO and BO + PAO groups were lower than that in the PAO group (P < 0.05). The purpose of the bronchus occlusion 60 min prior to pulmonary artery occlusion (BO + PAO group) was to remove the oxygen from pulmonary circulation. According to the measurement of Pt_O2, the oxygen content was almost zero in the BO + PAO group. These results indicate that an oxygen supply from airway is required for H₂O₂ production due to pulmonary artery occlusion. Although nonspecific mouse IgG did not affect the H₂O₂ production due to pulmonary artery occlusion, ICAM-1 blocking with mouse mAb 1A29 against rat ICAM-1 inhibited it (P < 0.05). PAF receptor-specific antagonist CV6209 also inhibited H₂O₂ production due to pulmonary arterial occlusion (P < 0.05). To confirm the specificity of the DCF reaction with H₂O₂, four rats were infused with 5,000 U/kg of catalase 20 min before left pulmonary artery occlusion. The DCF fluorescence was almost completely inhibited in these rat lungs. These results indicate that H₂O₂ production due to pulmonary artery occlusion was ICAM-1 and PAF dependent.

View larger version (22K): [in this window] [in a new window]   Figure 4.   Inhibition of H2O2 production in the nonhypoxic ischemia rat lung. H2O2 production in various groups at 90 min was measured by the DCF fluoro-imaging technique described previously. To determine the effect of ICAM-1 blocking on H2O2 production in the nonhypoxic ischemia rat lung, the rats were infused with 2 mg/kg of monoclonal antibody against rat ICAM-1 (mAb 1A29) or nonspecific IgG 20 min before left pulmonary occlusion, respectively, in the 1A29 group or N-IgG group. Furthermore, to determine the effect of PAF receptor-specific antagonist CV6209 on H2O2 production in the nonhypoxic ischemia rat lung, the rats were infused with 1 mg/kg of CV6209 20 min before left pulmonary artery occlusion in the CV6209 group. To confirm the specificity of the DCF reaction with H2O2, the rats were infused with 5,000 U/kg of catalase 20 min before left pulmonary artery occlusion in the catalase group. We also measured the DCF fluorescence of the pulmonary circulation in the BO group and BO + PAO group. Values are expressed as means ± SEM. *Significant difference compared with the control group (P < 0.05).

Lung MVP

To evaluate the microvascular injury in experimental groups, we measured the PI using the Evans blue dye extraction method. PI in the PAO group was higher than that in the control, BO, BO + PAO, 1A29, CV6209, and catalase groups (P < 0.05; Figure 5). Therefore, the pulmonary artery occlusion increased MVP, whereas oxygen removal from the airway in advance of pulmonary artery occlusion almost completely inhibited the increase in MVP. Furthermore, ICAM-1 blocking, PAF inhibition, and oxygen radical scavenging with catalase attenuated this MVP change.

View larger version (17K): [in this window] [in a new window]   Figure 5.   Lung MVP in various groups. To evaluate the microvascular injury in experimental groups, we measured the PI using the Evans blue dye extraction method. Experimental groups were described in Figures 2, 3, and 4. Values are expressed as means ± SEM. *Significant difference compared with the control group (P < 0.05).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

To determine H₂O₂ production in the nonhypoxic ischemia rat lung, we applied our digital fluoro-imaging technique (27) in vivo. DCFH-DA, which has been used for measuring H₂O₂ production in individual leukocytes, endothelial cells, and myocytes in vitro (14, 34), is hydrolyzed in the cell into nonfluorescent 2',7'-dichlorofluorescein (DCFH). DCFH is rapidly oxidized to highly fluorescent DCF in the presence of H₂O₂ (35). Previously, it has been impossible to directly demonstrate oxygen free radicals in ischemic organs, because the neutrophils cannot be isolated from the ischemic circulation and oxygen free radicals are unstable. Usually, enzymatic scavenging systems for oxygen free radicals in the plasma rapidly inactivate these radicals (37). However, our digital fluoro-imaging technique allowed us to directly detect H₂O₂ production in the rat ischemic pulmonary circulation. In this study, H₂O₂ production was detected in leukocytes within the pulmonary capillaries only in the nonhypoxic ischemia rat lung. This H₂O₂ production was associated with an increase in MVP, which was observed only in the rats with nonhypoxic ischemia. Oxygen removal from the airway in advance of ischemia almost completely inhibited H₂O₂ production as well as the increase in MVP. ICAM-1 blocking with mouse mAb 1A29 against rat ICAM-1 and the antagonist for PAF receptor significantly inhibited the H₂O₂ production as well as the increase in MVP. These results indicate that nonhypoxic ischemia without reperfusion alone causes production of oxygen-derived free radicals and lung microvascular injury, and that the oxygen for the radical production is provided from the airway in the nonhypoxic ischemia rat lung. Furthermore, ICAM-1 and PAF contribute to these reactions.

Although the in vivo DCFH oxidation method was discussed in our previous paper (27), whether the H₂O₂ produced by leukocytes in the nonhypoxic ischemia rat lung is extracellular or intracellular remained to be determined. While DCFH accumulates in cells, Bass and coworkers demonstrated that intracellular DCFH was oxidized by extracellular reagent H₂O₂, and that catalase inhibited DCF fluorescence in the neutrophil stimulated by phorbol myristate acetate (35). We also demonstrated that the pretreatment with catalase inhibited DCF fluorescence due to pulmonary artery occlusion. Catalase must scavenge extracellular H₂O₂ around the leukocyte because it cannot penetrate the cell membrane. We speculate that extracellularly generated H₂O₂ diffuses into the intracellular space, where H₂O₂ oxidizes the intracellular DCFH and converts nonfluorescent DCFH into highly fluorescent DCF.

Even though the MVP was increased, the wet-to-dry weight ratio in the PAO group did not change (data not shown). After clamping of the pulmonary artery, no blood was supplied from the circulation and therefore the lung weight never increased. To measure the true extravascular lung water, the Evans blue dye extravasation technique was used (32) to evaluate the MVP. With this method it is impossible to avoid underestimating the MVP, because once the pulmonary artery is clamped, no blood (Evans blue) is supplied from the circulation. Also, the values are influenced by changes in the vascular surface area due to vasoconstriction or vasodilation, especially in the setting of hypoxia. Although there was no statistically significant difference between the PI in the control group and those in the BO and BO + PAO groups, the mean values in the BO and BO + PAO groups were lower than those in the control group (Figure 5). Although there are factors that make evaluation of the MVP with this method difficult, this study showed that pulmonary artery occlusion increased MVP and that the oxygen removal from the airway, ICAM-1 blocking, PAF receptor antagonist, and radical scavenging with catalase inhibited MVP change due to pulmonary artery occlusion.

Ischemia causes ATP depletion in the cells (2, 3), especially when pulmonary glycogen stores are minimal (38). The ATP level in the lung decreased within 30 min of pulmonary artery clamping to 27% of the baseline level (2). ATP depletion activates xanthine oxidase (3) and disturbs calcium homeostasis by the extrusion of intracellular calcium by the plasma membrane calcium ATPase. This leads to elevation of cytosolic calcium concentration and cell membrane damage (39). During I/R injury, reperfusion supplies oxygen to the ischemic cells and activated xanthine oxidase releases superoxide anion and toxic oxygen metabolites in endothelial cells (3, 40). In contrast, in the nonhypoxic ischemia rat lung, oxygen is always supplied from the airway, so changes which normally occur during reperfusion may occur during ischemia. Superoxide anion and other toxic oxygen metabolites are generated in the lung (26, 41), especially in the endothelial cells. Although H₂O₂ production in endothelial cells was not detected, our results do not contradict the model of oxygen radical production in endothelial cells. The sensitivity in our system may not be sufficient to detect oxygen radical production by endothelial cells. After being generated in endothelial cells, intracellular oxygen radicals promote the neutrophil and endothelial cell interaction (36) and subsequent various inflammatory changes.

It remains unclear how neutrophils floating in the pulmonary capillaries are stimulated in the nonhypoxic ischemia rat lung. We demonstrated in this study that ICAM-1 blocking with mAb 1A29 and PAF receptor blocking with antagonist CV6209 inhibited both H₂O₂ production and MVP changes. After pulmonary artery occlusion, no leukocytes in the systemic circulation should be sequestrated to the pulmonary circulation, because no blood is supplied after pulmonary artery occlusion. Therefore, CV6209 and mAb 1A29 inhibited oxygen radical generation independently of an effect on leukocyte sequestration. These results indicate that the oxygen radical generation in the pulmonary circulation is not due to leukocyte sequestration but to the respiratory burst of the leukocyte in the pulmonary capillaries, and this respiratory burst is induced by PAF and leukocyte adhesion to the pulmonary capillary endothelial cell via ICAM-1. Many investigators demonstrated that I/R upregulates ICAM-1 expression (11, 12). Furthermore, Sellak and associates demonstrated that oxygen radical rapidly increases endothelial ICAM-1 ability to bind neutrophils without detectable upregulation (42). Entman and coworkers reported that the adherence of neutrophil through Mac-1 and ICAM-1 enhanced the neutrophil respiratory burst (14). It was reported that phospholipase A₂ was activated during ischemia in the intestine (15) and I/R in the kidney (16). This change and subsequent PAF synthesis may occur in the nonhypoxic ischemia rat lung. It is well known that PAF primes neutrophils for enhanced superoxide production by adherent neutrophil (17). PAF itself also causes neutrophil superoxide anion production (20). These reports support our results. Then the superoxide anion produced by the PAF-stimulated neutrophils, which adhere to endothelial cell through ICAM-1, are spontaneously, or by catalase, converted to H₂O₂. Subsequently, this H₂O₂ increases lung MVP (43).

In our experiment, hypoxia in advance of ischemia almost completely inhibited H₂O₂ production and the increase in MVP. However, it was reported that hypoxia may worsen the ischemic injury (25) in the buffer-perfused lung. The discrepancy between our experiment and this report may be due to the absence of blood in the buffer-perfused lung. Blood contains numerous antioxidants and antiproteases, and red blood cells have a scavenger receptor for interleukin 8 and other chemokines. In our in vivo studies, blood may have helped to protect the lung from injury. Many investigators have demonstrated that lung edema after reperfusion is more severe than during ischemia (5, 8). We propose two possible reasons for this phenomenon. First, the lung injury actually occurred during the period of pulmonary artery occlusion, but because of low microvascular pressures during this period, the lung edema after reperfusion was more severe than during ischemia, as Bishop and colleagues reported (24). Second, the lung injury during ischemia may be less than after reperfusion, because the number of activated neutrophils is limited in the pulmonary circulation during ischemia. After reperfusion, more neutrophils accumulate in the lung. Therefore, the lung edema after reperfusion is more severe than during ischemia in other reports.

In summary, these studies demonstrate that pulmonary artery occlusion alone increased H₂O₂ production in the lung and MVP, and that removal of oxygen from the airway, ICAM-1 blocking, and PAF receptor antagonist significantly inhibited both H₂O₂ production and the increase in MVP. H₂O₂ produced in this reaction was derived from leukocytes within the pulmonary capillaries. These results indicate that the nonhypoxic ischemia without reperfusion alone causes production of oxygen-derived free radicals and lung microvascular injury, and that the oxygen for the radical production is provided from the airway in the nonhypoxic ischemia rat lung. Furthermore, ICAM-1 and PAF contribute to these reactions.