INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Endotoxin derived from gram-negative bacteria causes lung injury (1, 2). Oxygen radical generated from neutrophils and macrophages play a major role in the pathogenesis of lung injury (3, 4). Although we demonstrated direct evidence of respiratory burst of the neutrophil in the pulmonary circulation of the endotoxin-infused rat in a previous study (4), the mechanism of the neutrophil respiratory burst in the bloodstream is still unclear.

On the basis of in vitro and in vivo data, the initial margination of neutrophil along endothelium is mediated by the selectin family (5, 6) in the inflammatory process. The selectins induce neutrophil rolling on the endothelium (5). Once neutrophil is rolling on the endothelium, the firm adhesion between neutrophil and endothelial cell occurs. It is generally believed that this adhesion is mediated by Mac-1 (CD11b/ CD18) on the neutrophil and intercellular adhesion molecule-1 (ICAM-1) on endothelial cells (7, 8). This adhesion leads to subsequent inflammatory processes such as neutrophil respiratory burst (4, 9).

The mechanisms of adhesion-dependent neutrophil respiratory burst have been investigated in vitro. Entman and colleagues (9) demonstrated that adherence dependent on Mac-1 and ICAM-1 activates the neutrophil respiratory burst. It was also reported that the activation of ₂-integrin induces neutrophil respiratory burst (10). However, the role of adhesion molecules on the neutrophil respiratory burst in an in vivo model has not been studied. Therefore, to determine the role of adhesion molecule ICAM-1 in the neutrophil respiratory burst, we applied our fluoro-imaging technique (4) to an endotoxin-infused rat lung and studied the ICAM-1 neutralization with monoclonal antibody against rat ICAM-1 in the same model. Using these models, we were able for the first time to demonstrate that neutrophil adhesion to the endothelial cell with ICAM-1 mediates neutrophil respiratory burst in vivo.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Production and Purification of Monoclonal Antibody against Rat Intercellular Adhesion Molecule-1, 1A29

Hybridoma clone 1A29 (13) was kindly provided by Dr. Miyasaka, Osaka University. Hybridomas were grown in RPMI 1640 containing 10% FCS (Gibco, Grand Island, 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 to 3 × 10⁶ 1A29 cells. Ascites was collected and monoclonal antibody (mAb) 1A29, which recognizes rat ICAM-1, was purified using an Ampure 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 Wister rats weighing 200 to 300 g were anesthetized by an intraperitoneal injection of pentobarbital sodium (50 mg/kg). To eliminate artifacts caused by movement 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 rats were 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 rats were 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 performed, which was extended to the left lateral thoracotomy along the eighth rib. An interlobar site of the left upper lobe was gently bonded to a glass chamber with the bonding agent Alonalpha. 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).

The Number of Sticking Leukocytes in the Pulmonary Microcirculation

The rats were infused with 0.1% acridine red (Chroma) in saline just before observation. The intact pulmonary microcirculation was observed by a fluorescence microscope with an excitation filter BP545 (Olympus), dichroic mirror and barrier filter O590 (Olympus). The images were recorded on a video cassette recorder as mentioned previously. After the experiment, the number of sticking leukocytes was counted directly on the monitor. We considered a 5 to 10 µm round fluorescent spot in the vessels as a leukocyte because an erythrocyte is not stained with acridine red. We defined the leukocyte that did not move for 60 s as a sticking leukocyte. In a previous report (4) we demonstrated that leukocytes did not stick to the venule but to capillaries in this model; therefore, we determined the number of leukocytes sticking to capillaries. The number of sticking leukocytes was expressed as the leukocytes number/monitor in the capillary.

In vivo Visualization and Quantification of Hydrogen Peroxide

We visualized and measured hydrogen peroxide (H₂O₂) production in the intact pulmonary circulation by a digital fluoro-imaging technique (4). On the day of the experiment, 2',7'-dichlorofluorescin-diacetate (DCFH-DA) (Molecular Probes, Inc., Eugene, OR) was suspended in physiological saline. DCFH-DA is hydrolyzed in the cell into nonfluorescent 2',7'-dichlorofluorescein (DCFH). DCFH is rapidly oxidized to highly fluorescent 2',7'-dichlorofluorescin (DCF) in the presence of H₂O₂. Therefore, we were able to detect the H₂O₂ production as a DCF fluorescence. The rats were injected with one mg/body of DCFH-DA. Fifteen minutes after injection, 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 expiration to eliminate lung movements. After the experiment, the recorded images were digitized and recorded on a hard disk using an image digitizing card (FRM512; Photoron, Tokyo, Japan), analyzed with a Macintosh Computer and image-analyzing software. The digitization of images was 512 × 512 resolution with 256 gray levels. The fluorescent area in the image was selected by image-analyzing software, NIH Image (Version 1.56), 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 the catalase.

Experimental Groups

We gave the rats a continuous femoral intravenous infusion of endotoxin at a rate of 4.5 mg/kg/h delivered by a syringe pump (Atom Microinfusion Pump 201; Atom, Tokyo, Japan) for 120 min (the endotoxin group). The rats were infused with physiological saline alone for 120 min (the control group). They received 2 mg/kg of mAb 1A29 20 min before intravenous infusion of endotoxin at a rate of 4.5 mg/kg/h for 120 min (1A29 group). To confirm the specificity of DCF reaction with H₂O₂, 5,000 U/kg of a catalase was given to the endotoxin- infused rats 20 min before continuous infusion of endotoxin (catalase group). The rats received 2 mg/kg of nonspecific mouse IgG 20 min before intravenous infusion of endotoxin at 4.5 mg/kg/h for 120 min (n-IgG group). Each group consisted of five rats.

H₂O₂ Detection with Cerium Technique by Electron Microscopy

Cerium technique has been used to detect H₂O₂ in electron microscopic studies (14, 15). We modified and applied this technique to detect H₂O₂ in the pulmonary circulation. Briefly, after infusion of the rats with endotoxin, a polyethylene cannula was inserted into the pulmonary artery from infundibulum of the right ventricle. Then, the blood was washed out with saline for 5 min. The lung was perfused with a prewarmed (37° C) cytochemical reaction medium consisting of 0.1 M TRIS-maleate buffer at pH 7.4, with 7% sucrose, 1 mM CeCl₃, and 10 mM 3-amino-1H-1,2,4-triazole for 10 min. Then the lung was perfusion-fixed with a 2% glutaraldehyde 0.1 M cacodyrate buffer at pH 7.4 for 5 min. The lung was postfixed in 1% OsO₄, dehydrated, and embedded in an epoxy resin. Ultrathin sections were prepared by an ultramicrotome (LKB 8800 Ultratome III; Bromma, Stockholm, Sweden) and examined by electron microscopy (JEOL 1200 EX; JEOL, Tokyo, Japan). The specificity of the cerium reaction with H₂O₂ was confirmed by blocking in the presence of the catalase.

Electron Microscopic Immunohistochemistry

After the experiment, the rats were killed with an overdose of pentobarbital sodium, the chest was opened, and the lungs were removed. The lungs were cut into 2 mm³ tissue blocks and immersion-fixed with a PLP solution for 2 h. After fixation, lung tissue blocks were rinsed in ice-cold PBS and then cryoprotected overnight by immersion at 4° C in PBS containing 2.3 M sucrose and 50% polyvinyl pyrolidone (Sigma Corp., St. Louis, MO). Ultrathin (80 to 100 nm) cryosections were obtained with an ultramicrotome (LKB 8800 Ultratome III) equipped with LKB 14800 Cryokit (Bromma). Ultrathin sections were collected on cryoprotectant droplets (50% sucrose) and transferred to formvar-coated grids. The sections were rinsed with TBS containing 10 mM glycine and immunolabeled with the primary antirat ICAM-1 mAb 1A29 at a 1/100 dilution in TBS containing 1% BSA for 1 h at room temperature. Then the sections were immunolabeled with a secondary polyclonal rabbit-anti-mouse IgG conjugated to 10 nm colloidal gold (Dako, Carpenteria, CA). After embedding in a mixture of 0.4% uranyl acetate and 3.6% polyvinyl alcohol, the sections were examined with a transmission electron microscope (JEOL 1200 EX; JEOL).

Immunogold Quantification

We analyzed the surface immunoreactivity of ICAM-1 on rat pulmonary endothelial cells and type I epithelial cells using a modification of the method of Burns and colleagues (16). Briefly, two to three lung tissue blocks were randomly sampled from 20 blocks that were available for each rat. Immunogold-labeled tissue sections from these blocks were viewed at a primary magnification ×2,000 on a transmission electron microscope. At this magnification, gold particles were not visible, but it was possible to identify regions of lung tissue in which the alveolar wall was not twisted or obscured by overlapping sections. Five alveolar wall segments in each rat were randomly photographed. Then gold particles were counted at magnification ×100,000. At this magnification, the microvascular endothelial cell and alveolar type I epithelial cell surfaces were identified, and gold particles were counted only on the plasma membrane. We also measured the length of the plasma membrane of the endothelial cell and the alveolar epithelial cell in the photograph. The gold particle density was calculated by dividing the total number of gold particles that were counted by the total length of plasma membrane examined.

TNF- Assay

After 2 h of endotoxin or saline infusion, the rat lung was immediately frozen with liquid nitrogen and stored at 80° C until TNF- assay. The lungs were lysed with the lysis buffer (20 mM TRIS-HCl at pH 7.4, sodium dodecyl sulfate 0.1%, 1% Triton-X 100, 1% sodium deoxycholate). Then TNF- content in the lungs was determined by Cytoscreen rat TNF- ELISA kit (Biosource International, Camarillo, CA) according to the manufacturer's instructions, using a Wellreader SK601 (Seikagaku Corp., Tokyo, Japan).

Statistics

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

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The Number of Sticking Leukocytes in the Pulmonary Microcirculation

To study the leukocytes sticking to the pulmonary circulation, we observed and measured the number of sticking leukocytes stained with acridine red by means of intravital fluorescence microscopy. In the endotoxin group, the leukocytes stuck to the capillaries (Figure 1A). In contrast, leukocytes did not stick to capillaries in the control group (Figure 1B). The number of leukocytes sticking to the capillaries in the endotoxin group was higher than that in the control group (p < 0.05) (Figure 2). Pretreatment of 2 mg/kg of mAb 1A29 did not inhibit the leukocyte sticking to pulmonary capillaries of endotoxin-infused rats (Figure 2).

View larger version (104K): [in this window] [in a new window]   Figure 1.   Images of leukocytes sticking to the pulmonary capillaries of the rat in the endotoxin group (panel A) and the control group (panel B). The rat in the endotoxin group was infused with endotoxin for 120 min, and the rat in the control group was infused with saline alone for 120 min. The bloodstream was stained with acridine red. There are many leukocytes sticking to capillaries (panel A) of the rat in the endotoxin group. There are few leukocytes in the pulmonary capillaries of the rat in the control group (panel B). Arrow = leukocyte; a = alveolus; c = capillary.

View larger version (25K): [in this window] [in a new window]   Figure 2.   Effect of anti-rat ICAM-1 monoclonal antibody on the number of sticking leukocytes in response to endotoxin in the rat pulmonary circulation. The number of sticking leukocytes was directly counted from a TV monitor through a ×40 objective lens and a ×3.3 C-mount lens. Solid bars, blank bars, and hatched bars show the number of sticking leukocytes in the endotoxin group, the control group, and 1A29 group, respectively. Five rats were determined for each group. The values are expressed as mean ± SEM. *Significantly different from the control group (p < 0.05).

In vivo Visualization and Quantification of H₂O₂

To determine the H₂O₂ production in the pulmonary circulation, we observed and measured DCF fluorescence in the pulmonary circulation of rats in each experimental group. The DCF fluorescence images from single representative experiments are shown in Figure 3. DCF fluorescence was negligible in the rat pulmonary circulation of the control group (Figure 3B), but it was marked in the endotoxin group (Figure 3A). The DCF fluorescent spot in low magnification consisted of a globular DCF fluorescent spot attributable to leukocyte in high magnification view (Figure 3C). There was a spindle-shaped DCF fluorescence adjoining the globular spot. This spindle-shaped DCF fluorescence can be attributed to the capillary endothelial cell because the distance between the spindle-shaped DCF fluorescence and the leukocyte DCF fluorescence was less than 1 µm. The amount of DCF fluorescence in the pulmonary circulation of rats in the endotoxin group was higher than that in the control group (p < 0.05) (Figure 4). Pretreatment with 2 mg/kg of mAb 1A29 almost completely inhibited DCF fluorescence caused by endotoxin infusion (p < 0.05) (Figure 4). However, nonspecific mouse IgG did not inhibit DCF fluorescence (Figure 4). To confirm the specificity of DCF reaction with H₂O₂, 5,000 U/kg of catalase was given to endotoxin-infused rats 20 min before continuous infusion of endotoxin. Five thousand U/kg of catalase pretreatment almost completely inhibited DCF fluorescence (p < 0.05) (Figure 4). To determine the kinds of leukocytes that generate H₂O₂, we applied an electron microscopic cerium technique. The cerium was deposited around the neutrophil. It was also seen around the adjoining endothelial cell in the endotoxin-infused rat (Figure 5). The endothelial cerium deposition can be spindle-shaped DCF fluorescence (Figure 3C). In the catalase-pretreated rat lung, there was no cerium deposition around the neutrophil. In summary, neutrophil generated H₂O₂ in the endotoxin-infused rat. H₂O₂ was also detected in the adjoining endothelial cells. H₂O₂ production was blocked by ICAM-1 neutralized with mAb 1A29.

View larger version (37K): [in this window] [in a new window]   Figure 3.   Image of DCF fluorescence of the endotoxin-infused rat and control rat pulmonary circulation. The rat was infused with endotoxin for 120 min, and a control rat was infused with saline alone for 120 min. Then, to detect H2O2 production, each rat was injected with DCFH-DA, and DCF fluorescence images were taken through a silicon-intensified target TV camera. Numerous fluorescence (white spots) were detected in panel A (the endotoxin group). A small amount of fluorescence was detected in panel B (the control group). A globular fluorescent spot attributable to leukocyte and spindle-shaped fluorescence was seen in high power view (panel C ). The diameter of the globular spot was 5 to 10 µm, and this spot was always in the capillary and did not move during observation. The distance between the spindle-shaped fluorescence and the leukocyte fluorescence was always less than 1 µm. Therefore, the spindle-shaped DCF fluorescence may be attributable to the capillary endothelium. L = leukocyte; arrowhead = spindle-shaped DCF fluorescence. Bar = 5 µm.

View larger version (27K): [in this window] [in a new window]   Figure 4.   Effect of anti-rat ICAM-1 monoclonal antibody on H2O2 production in response to endotoxin in the rat pulmonary circulation. Using a computer image analyzing system, H2O2 production was determined by counting pixels of DCF fluorescence in the image of the pulmonary circulation through a ×4 objective lens and a ×3.3 C-mount lens. To confirm the specificity of DCF reaction with H2O2, 5,000 U/kg of catalase was given to endotoxin- infused rats 20 min before continuous infusion of endotoxin (catalase group). Catalase pretreatment almost completely inhibited H2O2 production caused by endotoxin infusion. Five rats were determined for each group. The values are expressed as mean ± SEM. *Significantly different from the control group (p < 0.01).

View larger version (136K): [in this window] [in a new window]   Figure 5.   Electron microscopic H2O2 detection in an endotoxin- infused rat lung. After infusion of the rat with endotoxin, the rat lung was perfused with cerium containing reaction medium for 10 min (see details in METHODS). The lung was fixed and examined by electron microscopy. The specificity of the cerium reaction with H2O2 was confirmed by blocking in the presence of catalase. The cerium was deposited around the neutrophil (asterisk). It was also seen around the adjoining endothelial cell in the endotoxin-infused rat. This cerium deposition around the endothelial cell must correspond to the spindle-shaped DCF fluorescence seen in Figure 3C.

ICAM-1 Expression in the Lung

We determined the ICAM-1 expression in the lung in the control and the endotoxin groups immunohistochemically. Even in the control lung the alveolar wall was stained with mAb 1A29 against rat ICAM-1. The alveolar wall of the lung in the endotoxin group was also positively stained. There was no difference between the ICAM-1 expression intensity and the site in the lung in the control group and in the endotoxin group under light microscopic observation (data not shown). However, it was impossible to recognize the alveolar type I epithelial cell and the capillary endothelial cell under light microscopic observation. Therefore, we determined the ICAM-1 expression of the alveolar type I epithelial cell and the capillary endothelial cell with immunoelectron microscopy. The typical immunoelectron microscopy that demonstrates the ICAM-1 expression of the alveolar type I cell and the capillary endothelial cell in the endotoxin group is shown in Figure 6. ICAM-1 was expressed on the plasma membrane of the capillary endothelial cell and the alveolar type I epithelial cell in both the control and the endotoxin groups (Table 1). The ICAM-1 expression in the capillary endothelial cell was ten times lower than that in the alveolar type I cell in both groups (Table 1). Although there was no statistical difference between the ICAM-1 expression in the alveolar type I cell in the control group and the endotoxin group, the ICAM-1 expression of the endothelial cell in the endotoxin group was 1.5 times higher than that in the control group (p < 0.05, Table 1).

View larger version (51K): [in this window] [in a new window]   Figure 6.   Immunoelectron microscopic study of ICAM-1 expression in the lung (single experiment). Panels A and B show typical photographs in an endotoxin-infused rat lung. We quantified ICAM-1 expression in a capillary endothelial cell and alveolar type I epithelial cell by counting the number of gold particles on the plasma membrane (Table 1) . Panels A and B show a low magnification photograph and high magnification photograph, respectively. Arrow = gold particle; Ep = type I epithelial cell; En = endothelial cell.

TNF- Content in the Lung

TNF- content in the lung in the control and the endotoxin groups was measured by ELISA. TNF- in the endotoxin group was five times higher than that in the control group (Table 2).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We made a rat acute lung injury model with the intravenous infusion of endotoxin. In this model, neutrophils stuck to the pulmonary capillary and generated H₂O₂, which was detected by both a 2,7-dichlorofluorescin (DCF) fluoro-imaging technique and an electron microscopic cerium technique. Anti-rat ICAM-1 mAb 1A29 did not inhibit the leukocyte from sticking to the pulmonary capillary, but it did inhibit H₂O₂ generated by adherent neutrophils. These data indicate that ICAM-1 is not necessary for leukocyte adhesion to the pulmonary capillary but is required for neutrophil respiratory burst in the rat pulmonary circulation infused with endotoxin. This is the first study to demonstrate adhesion-dependent neutrophil respiratory burst in vivo.

We detected 2,7-dichlorofluorescin (DCF) fluorescence attributable to leukocyte respiratory burst in the rat pulmonary circulation infused with endotoxin. H₂O₂ sensitive, nonfluorescent compound, 2',7'-dichlorofluorescein (DCFH) is oxidized and converts to the highly fluorescent DCF under the presence of H₂O₂. Bass and colleagues (17) developed the DCFH oxidation method for detection of H₂O₂ produced by isolated neutrophil using flow cytometry, and many researchers have studied neutrophil respiratory burst with this powerful method (18). However, some investigators reported problems related to the specificity of DCFH to H₂O₂ (9, 21- 23). On the other hand, Lebel and colleagues (22) demonstrated that hydroxyl radical and superoxide anion did not react with DCFH, and H₂O₂-Fe²⁺-derived oxidant was mainly responsible for the oxidation of DCFH. Furthermore, many researchers have reported that catalase inhibits DCF fluorescence caused by reagent H₂O₂ and neutrophil respiratory burst in an in vitro and in vivo model (4, 17, 23). Therefore, a DCFH oxidation method for detection of H₂O₂ in in vitro and in vivo models can be generally accepted. In this study, catalase pretreatment almost completely inhibited DCF fluorescence (Figure 4). Although DCFH accumulates in cells, Bass and coworkers (17) demonstrated that intracellular DCFH was oxidized by extracellular reagent H₂O₂ and that catalase inhibited DCF fluorescence in the neutrophil stimulated by phorbol myristate acetate. 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, whereas H₂O₂ oxidizes the intracellular DCFH and converts nonfluorescent DCFH to highly fluorescent DCF. Therefore, we concluded that DCF fluorescence, which we detected in the rat pulmonary circulation, was due to leukocyte respiratory burst. Furthermore, we confirmed that neutrophil generates H₂O₂ by an electron microscopic cerium technique. These results are consistent with the DCF fluorescence results.

In the DCF fluorescence image and the electron microscopic cerium image, we demonstrated that the endothelial cell adjoining neutrophil has spindle-shaped DCF fluorescence and cerium deposition. These observations indicate that the endothelial cell adjoining neutrophil, which produced H₂O₂, has H₂O₂. There are two possible reasons that spindle-shaped DCF fluorescence and cerium deposition were seen in the endothelial cells. The first possibility is that the endothelial cells released H₂O₂. The second possibility is that the H₂O₂ generated by neutrophil diffused into the endothelial cells. Although several reports have suggested that endothelial cells could respond with generation of superoxide anion (27, 28), we assume that the origin of the H₂O₂ is not the capillary endothelial cells but the adherent neutrophil because the DCF fluorescence and cerium deposition were always observed at a site close to an adherent neutrophil. It was reported that the H₂O₂ generated from activated neutrophils diffused into the myocytes (9), and this report supports our observation that the H₂O₂ generated from neutrophils diffused into the endothelial cells.

We observed ICAM-1-independent neutrophil sticking to the pulmonary capillaries induced by endotoxin in this study. Although we did not demonstrate the mechanism of ICAM-1-independent neutrophil sticking, Erzurum and colleagues (29) investigated it in detail. Endotoxin induces CD14-mediated neutrophil f-actin reorganization and increases the neutrophil stiffness. These mechanical changes of neutrophil cause CD18-independent neutrophil retention in lung in an early phase. After that, CD11b/CD18-mediated neutrophil adhesion occurs (30) in a later phase. Therefore, we speculate that the same reaction must have occurred in our model. After that, the neutrophil stuck firmly to the pulmonary capillary endothelial cell via ICAM-1. However, we also need to consider the other neutrophil adhesion mechanisms without ICAM-1 such as a selectin family (5, 6) and ATP-induced adhesion (31).

ICAM-1 neutralization with monoclonal mouse antibody 1A29 against rat ICAM-1 did not inhibit the neutrophil from sticking to the pulmonary capillary, but it did inhibit the neutrophil H₂O₂ production. The dose of antibody 1A29 (2 mg/ kg) was necessary and sufficient. This result agreed with the previous report (32). Furthermore, the same dose of nonspecific IgG did not influence the effect of endotoxin infusion. These results indicate that the neutrophil adhesion with ICAM-1 mediates the neutrophil respiratory burst. The role of ICAM-1 on neutrophil respiratory burst has not been well investigated and is controversial. von Asmuth and colleagues (33) demonstrated that neither induction of increased expression of ICAM-1 on endothelial cells nor subsequent addition of specific mAb influenced H₂O₂ release by TNF-activated neutrophils in vitro. On the other hand, Entman and colleagues (9) demonstrated that adherence dependent on Mac-1 (CD11b/ CD18) and ICAM-1 (CD54) activates the neutrophil respiratory burst in vitro. Furthermore, Chihara and coworkers (34) reported that recombinant soluble ICAM-1 augmented eosinophil respiratory burst. To clarify the role of ICAM-1 on neutrophil respiratory burst, further investigation is needed. Meanwhile, there are many reports of the role of ₂-integrin, the ligand for ICAM-1, on neutrophil respiratory burst. It was reported that the cross-linking of ₂-integrin induces neutrophil respiratory burst (10, 11). Furthermore, Menegazzi and colleagues (12) reported that TNF- induces ₂-integrin-dependent neutrophil respiratory burst. In fact, TNF- in the endotoxin group was five times higher than that in the control group in this study (Table 2). Therefore, we speculate that endotoxin infusion caused TNF--induced ₂-integrin and ICAM-1-dependent neutrophil respiratory burst in the pulmonary circulation in our model.

To determine the ICAM-1 expression on the endothelial cell due to endotoxin infusion, we applied an immunoelectron microscopic technique. Using this technique, we were able to show the increase in ICAM-1 expression on the endothelial cell. In light microscopic immunohistochemistry, the alveolar wall was positively stained with ICAM-1, even in the control lung. Panes and colleagues (35) also reported that ICAM-1 was detected in the lung without any treatment. Under light microscopic immunohistochemistry, there seemed to be no difference between the ICAM-1 expression in the endotoxin-infused rat lung and that in the control lung. However, it was impossible to distinguish between ICAM-1 on the capillary endothelial cell and on the alveolar type I epithelial cell by light microscopy. Therefore, we determined the ICAM-1 expression of the alveolar type I epithelial cell and the capillary endothelial cell with immunoelectron microscopy. The ICAM-1 expression in the capillary endothelial cell was ten times lower than in the alveolar type I epithelial cell. These results are consistent with the previous report (16). Endotoxin infusion for 2 h did not increase ICAM-1 expression on the alveolar type I epithelial cell, but endotoxin infusion for 2 h did increase ICAM-1 expression on the capillary endothelial cell 1.5 times higher than in the control lung. These results indicate that the increase in the ICAM-1 on the endothelial cell was masked with the ICAM-1 expression on the alveolar type I epithelial cell. Therefore, we should be careful when evaluating the ICAM-1 expression in the lung.

In summary, neutrophils stuck to the pulmonary capillary and generated H₂O₂ in rat pulmonary circulation infused with endotoxin. Anti-rat ICAM-1 mAb 1A29 did not inhibit the leukocyte from sticking to the pulmonary capillary, but it did inhibit H₂O₂ production generated by adherent neutrophils. These data indicate that ICAM-1 is not necessary for leukocyte adhesion to the pulmonary capillary, but it is required for neutrophil respiratory burst in the endotoxin-infused rat. We have demonstrated for the first time adhesion-dependent neutrophil respiratory burst in vivo.