Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Microvascular leakage and mucosal edema are prominent components of airway inflammation and they appear to play important roles in the pathogenesis of asthma (1, 2). Evidence for the involvement of plasma leakage in this disease includes elevated albumin concentration in the sputum of asthmatic patients, and the presence of plasma proteins as major components in the thick mucus plugs that have been observed occluding the airways of patients who die of status asthmaticus. Local instillation of antigen into the bronchial mucosa of sensitive subjects induces acute swelling and narrowing of the airway that can be directly visualized using a bronchoscope. The thickened airway wall suggests the presence of edema fluid in the interstitial tissue. The swelling and inflammation may directly limit air flow, contributing to hyperresponsiveness, reducing mucociliary clearance, and causing epithelial damage and compromising surfactant functions in the lungs of these patients. The presence of plasma components in the airway lumen indicates that both microvascular and epithelial permeability barriers have been compromised and the fluid balance disturbed. The beneficial effects of corticosteroids in asthma may be partly due to the prevention of vascular leakage.

The development of late-phase eosinophilia after antigen challenge is also one of the prominent features of chronic airway inflammation associated with asthma. Abundant evidence suggests that early events in the transmigration of eosinophils and T-lymphocytes from blood vessels are controlled by the interaction of multiple adhesion molecules between these leukocytes and vascular endothelial cells (3). Subsequent locomotion within tissue is dependent upon the interaction of leukocyte adhesion molecules with extracellular matrix proteins such as fibronectin (5, 7). In situ experiments with human asthmatic tissues demonstrated that eosinophils and T-lymphocytes appear to share several common adhesion molecules such as intercellular adhesion molecule 1 (ICAM-1) and very late activation antigen 4 (VLA-4). Blockade of the adhesion molecules with specific monoclonal antibodies (mAbs) can inhibit adherence of these inflammatory cells to endothelial cells in vitro. The roles of these adhesion molecules have previously been examined in several in vivo models of antigen-induced eosinophilia and hyperresponsiveness in guinea pigs (8, 9), rats (10, 11) and mice (6, 12). However, the correlation between lung eosinophilia and the change of microvascular permeability has not been established.

Antigen-induced airway inflammation has been studied extensively in allergic animals and the brown Norway rat model has been well characterized (13). Brown Norway rats produce high levels of IgE in response to active immunization (22) and develop both early and late airway constriction responses after inhalation of antigen (18). These rats also demonstrate a persistent increase in airway responsiveness to inhaled methacholine (15, 17). There was a significant increase in eosinophils, lymphocytes, and neutrophils in the bronchoalveolar lavage (BAL) fluids of sensitized and challenged animals after 18 to 24 h. During the late-phase response, large quantities of sulfidopeptido leukotrienes were excreted through the bile of these animals (18). Pharmacologic studies have shown that the late-phase response and airway hyperresponsiveness could be blocked by dexamethasone and were partially inhibited by the leukotriene antagonist MK-571 (20, 21, 23).

Most lung inflammation studies in allergic animals have focused on airway hyperresponsiveness and infiltration of inflammatory cells into the lungs. Very little is known concerning the role of microvascular permeability changes in the lung injury process. Hui and coworkers (24) used Evans blue dye extravasation into the airway tissues of ovalbumin (OA)-sensitized and aerosol antigen-challenged guinea pigs to demonstrate that plasma leakage occurred mainly in the distal airways and correlated with air flow obstruction. By contrast, intravenously administered antigen in the same guinea pig model resulted in an increased Evans blue exudation in the main bronchus and this effect was blocked by antagonists of histamine and leukotrienes (25). In another study, Misawa and Chiba (26), using dinitrophenol (DNP)- Ascaris antigen-sensitized and -challenged Wistar rats, reported that the main bronchus was a site of enhanced Evans blue exudation and hyperresponsiveness to acetylcholine and serotonin. There are no data concerning changes in microvascular permeability in allergic brown Norway rat models.

We have previously reported (27, 28) the effects of mAbs against the adhesion molecules ICAM-1 and ₄ integrin in the antigen-induced accumulation of eosinophils and lymphocytes in the airway lumen of OA sensitized brown Norway rats. The object of this study was to examine the roles of these two adhesion molecules in antigen aerosol-induced vascular leakage. We report that the antigen-induced increase in microvascular permeability in the lung may be independent of the infiltration of eosinophils in the airway lumen.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Materials

¹²⁵I-labeled bovine serum albumin (BSA) and ⁵¹Cr were purchased from either Du Pont-NEN (Boston, MA) or Amersham Corporation (Arlington Heights, IL). Carrier BSA, OA, and dexamethasone were purchased from Sigma Chemical Company (St. Louis, MO). The mAb against rat ICAM-1 (clone 1A29) (29) was kindly provided by C. Lane and D. C. Anderson. Anti-rat ₄ integrin mAb (clone TA-2) was purchased from Endogen Inc. (Boston, MA) and the nonblocking control mAb (clone P-23) was kindly supplied by J. G. Geng. Dulbecco's phosphate-buffered saline (D-PBS) and Hanks' balanced salt solution (HBSS) were obtained from GIBCO-BRL Life Technologies, Inc. (Grand Island, NY).

Antigen-induced Lung Inflammation

Male brown Norway rats, 180-250 g (Harlan Sprague, Prattville, AL) were sensitized by subcutaneous injection with OA at a dose of 5 mg/kg in a suspension of 1 g/kg aluminum hydroxide in saline as described previously (13, 27).

Microvascular Permeability

After 14 days, rat plasma protein was labeled with ¹²⁵I- labeled BSA. The BSA solution was prepared by mixing 1 mg/ml BSA (Sigma) in saline with an appropriate amount of ¹²⁵I-labeled BSA (NEX-076, 4.3 µCi/µg) to achieve a concentration of 10 µCi/mg. Animals were anesthetized with methoxyflurane (Metofane; Pitman Moore, Washington Crossing, NJ) by inhalation, and a small incision was made at the base of the neck over the clavicle, exposing the external jugular vein at the point where it emerges from under the pectoralis superficialis muscle. Three-tenths of a milliliter (3.0 µCi) of the ¹²⁵I-labeled BSA solution was injected with a 25-gauge needle inserted through the pectoralis superficialis muscle into the jugular vein. The muscle seals the vein once the needle is withdrawn, thereby minimizing leakage of labeled blood. The incision was closed with cyanoacrylate glue and the animals allowed to recover for at least 1 h prior to challenge.

OA aerosols were generated with a SPAG-2 small-particle generator (ICN Pharmaceuticals Inc., Costa Mesa, CA). Animals were placed in a 12-station inhalation chamber and challenged with an aerosol containing 1% OA in saline. The air flow to the atomizer was 6 liters/min whereas that to the drying chamber was set at 8 liters/min, yielding a median particle size of 0.66 µm. Control animals were treated with aerosols of saline. Following challenge, the rats were returned to holding cages for appropriate periods to allow for the development of lung inflammation.

The vascular space in the inflamed lungs was determined by the dilution of ⁵¹Cr-labeled red blood cells (RBCs). Brown Norway rat RBCs were prepared from a total of 30 ml of blood drawn from several donor animals (about 10 ml/animal) drawn in syringes containing 0.5 ml of an anticoagulant solution consisting of heparin (final concentration, 20 U/ml in blood) and EDTA, pH 8.0 (final concentration, 1.5 µM in blood). The blood was centrifuged at 1,300 × g for 15 min and the plasma discarded. The packed RBC fraction was diluted to its original volume with saline containing the same concentration of heparin and EDTA and mixed with 300 µCi of Na₂⁵¹CrO₄. The RBC fraction was incubated for 30 min at 37°C, washed twice with Dulbecco's phosphate-buffered saline (D-PBS) containing glucose, heparin, and EDTA, and once with D-PBS without heparin and EDTA. The packed cells were then resuspended and 0.3 ml of the suspension was injected into the jugular vein of each rat, as described previously. Each animal received an average of 3.0 µCi of ⁵¹Cr.

The animals were anesthetized intraperitoneally with 1.5 g/kg urethane and the incision over the jugular was reopened to expose the vein. The ⁵¹Cr-labeled packed RBCs were injected into the vein and allowed to distribute throughout the vasculature for 5 min. The abdominal wall was opened and 3 ml of blood was withdrawn from the inferior vena cava into a syringe containing 0.1 ml of a 1:1 mixture of heparin (1,000 U/ml) and 0.3 M EDTA at pH 8.0 to prevent clotting. The animals were exsanguinated via a cut in the major abdominal blood vessels. The chest cavity was then opened and the lungs were removed. Major pulmonary vessels were clamped with small hemostats to minimize blood in the lungs. Hematocrits for the rat blood were determined using an Autocrit II hematocrit centrifuge. Subsequently, the blood and plasma (prepared by centrifuging a blood sample at 1,300 × g for 15 min) samples and lungs were transferred to plastic tubes for counting in a Packard (Downers Grove, IL) 800C -counting system. Because repeated counting of standards consistently indicated a 10% spillover of counts from the ⁵¹Cr channel into the ¹²⁵I channel, 10% of the ⁵¹Cr counts for that sample have been subtracted from the ¹²⁵I counts for all calculations. After counting, the lungs from each animal were transferred to tared aluminum planchets to obtain the wet weight. The planchets were heated in a 100°C oven overnight, where they dried to constant weight.

The hematocrit was either measured directly or calculated from the difference between ¹²⁵I in blood and ¹²⁵I in plasma.

(1)

The volume of blood remaining in the vascular space of the lung is

(2)

The volume of plasma remaining in the vascular space of the lung was calculated from the hematocrit.

(3)

where Htc is the hematocrit, either calculated or measured, and plasma V is the volume of plasma (in milliliters). The total plasma volume of the lung was calculated from the ¹²⁵I data:

(4)

The exuded plasma volume in the inflamed tissue was calculated as

(5)

Dexamethasone was suspended in vehicle 122 and rats were dosed with 1 ml of the suspension. Monoclonal antibodies and isotypic control antibody were injected intraperitoneally with three doses of 2 mg/kg at 12 and 1 h before, and 7 h after, antigen inhalation.

Infiltration of Inflammatory Cells

To evaluate the effect of the test mAb on migration of inflammatory leukocytes into the lungs, groups of rats were injected with nonradioactive albumin prior to challenge at the same time as the animals for edema studies were injected with radioactive albumin. Enumeration of BAL and blood leukocyte populations was performed according to procedures described previously (27). Briefly, rats were anesthetized with urethane, as described in the preceding section, the trachea was cannulated, and 5 ml of PBS instilled into the lungs. The thorax was massaged for 30 s before recovering the BAL fluid. The lavage procedure was then repeated with an additional 5 ml of PBS. Three milliliters of HBSS containing 10% fetal bovine serum (FBS) were added to the pooled BAL fluid. Following centrifugation at 150 × g for 10 min at 4°C, the cells were resuspended in HBSS with 10% FBS. Total numbers of BAL leukocytes were determined using a Coulter counter (Model ZM; Coulter Electronics, Hialeah, FL). Differential cell counts were performed on cytospin preparations made in a Shandon cytocentrifuge (Shandon Southern Instruments, Sewickley, PA) set at 45 × g for 10 min at room temperature. Cells were fixed and stained using Diff-Quik (American Scientific Products, McGaw Park, IL) and 200 cells were counted using standard morphologic criteria to classify them as eosinophils, neutrophils, lymphocytes, or mononuclear cells (alveolar macrophages).

Blood samples were taken from the inferior vena cava and mixed with EDTA-coated Vacutainer tubes. Total white blood cells and differentials were counted using a Technicon H1E system (Miles Diagnostics, Tarrytown, NY).

Measurement of mAb Concentrations in Plasma

Circulating levels of the mAbs 1A29, TA-2, and P-23, were measured with an ELISA for mouse IgG₁ (27).

Histologic Observations of Lung Tissue

Naive and sensitized brown Norway (BN) rats were challenged with aerosols of OA, as described previously, without receiving ¹²⁵I-labeled BSA in order to assess their lungs histologically for evidence of edema. Following lavage, the lungs were removed and fixed by inflating with 10% phosphate-buffered formalin via the tracheal cannula to a pressure of 15 cm of water. The entire left lung was embedded in paraffin, sectioned at 6 µm, and stained with hematoxylin, phloxine, and eosin. Lung sections were observed for gross evidence of edema.

Statistical Analysis

Statistical comparisons for edema parameters were made using a Student's t test. Data from total and differential cell counts were analyzed using one-way analysis of variance (ANOVA) on untransformed and rank-transformed data. All means are presented with calculated standard errors of the mean (SEM).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Assessment of Edema

When OA-sensitized rats were challenged by inhalation, a significant change in ¹²⁵I content occurred in the lungs after 24 h (Figure 1). Plasma volume in the lung increased by 231.3 ± 42.3% (P < 0.005) as compared with control values. The wet weight of the lungs also increased to 130% of control value. Because the dry weight of the lungs also increased, the wet weight-to-dry weight ratio did not change significantly. This indicated that the wet weight increase in the antigen-exposed lung was not only due to the leakage of water into the interstitial space, but was also due to the entry of plasma proteins, inflammatory cells, and possibly collagen and fibrin. The time course studies indicated that the ¹²⁵I plasma volume in the lung tissue did not increase significantly during the first 12 h subsequent to challenge. However, increased permeability was apparent at 24 h with a doubling of ¹²⁵I-labeled albumin in the lungs, which then quadrupled after 72 h. The lungs appeared to be approximately twice as large as those of the control animals at 72 h. The time course of plasma protein leakage paralleled the accumulation of eosinophils in the airway lumen (I. M. Richards and coworkers, unpublished data, 1996). The results demonstrate that inhaled antigen induces a late-phase microvascular leakage in the lung tissue that is associated with the infiltration of inflammatory cells into the airway lumen.

View larger version (12K): [in this window] [in a new window]   Figure 1.   Time course of antigen challenge-induced vascular leakage in brown Norway rats. Rats (10 animals per group) were sensitized with OA for 2 wk and challenged with antigen aerosol for 12 min. The zero value represents OA-sensitized animals challenged with saline. All other points represent the time after antigen challenge in sensitized animals at which lung extravascular plasma volume was assessed. Data are expressed as mean ± SEM of extravascular plasma volume (ml) per gram of dried lung in animals labeled with 125I-labeled BSA as described in MATERIALS AND METHODS.

Histologic Evaluation

Evidence of edema was further investigated through histologic observations of hematoxylin-eosin (H&E)-stained lung sections from control and antigen-challenged animals. Representative sections of these preparations are presented in Figure 2. Figure 2 (upper panel ) was taken from a control animal and shows a close association between the blood vessels and the surrounding airway tissue. The effects of the ovalbumin challenge on lung histology are presented in Figure 2 (lower panel ). Note the loosely packed fibrous tissue, scattered inflammatory cells, and space around the vessel which is evidence of hydrostatically induced injury due to leakage of plasma from the vessel, or edema. In addition, edema was evident in animals with noticeable lung inflammation as thickened alveolar walls, which may be caused, in part, by fluid in the potential space between the endothelium and epithelium (data not shown).

View larger version (126K): [in this window] [in a new window]   View larger version (138K): [in this window] [in a new window]   Figure 2.   Lung sections from control animal lungs (upper panel) and 24 h postantigen challenge (lower panel) of sensitized BN rats. Lungs were removed, fixed with buffered formalin, embedded in paraffin, sectioned, and stained with H&E as described in the text. Magnification is indicated by the inserted scale bar. In the control animals, note the thickness of the blood vessel- airway junction, which shows almost no space between endothelium and epithelium. Contrast the control with the 24-h inflamed lung in the lower image with the increased space between airway and blood vessel, which is filled with fluid, loose fibers, and inflammatory cells.

Effect of Dexamethasone on Lung Edema

The formation of edema, indicated by the accumulation of ¹²⁵I-labeled BSA in the lungs, at 24 h after OA challenge was dose-dependently inhibited by oral administration of dexamethasone 1 h prior to challenge followed by an additional dose at 7 h postchallenge as depicted in Figure 3. The plasma leakage observed at 24 h postchallenge was completely eliminated by doses as low as 0.3 mg/kg and the lowest dose tested, 10 µg/kg, resulted in an inhibition of 59.77 ± 9.9%. These results confirmed previous reports (13, 21, 28) that steroids inhibited other features of antigen-induced lung inflammation in this model. Furthermore, the airway microvascular leakage and edema response appear to be a glucocorticoid-sensitive component of the lung inflammation induced by inhaled antigen.

View larger version (15K): [in this window] [in a new window]   Figure 3.   Dose-related inhibition of antigen-induced pulmonary vascular leakage by dexamethasone in brown Norway rats. OA-sensitized and challenged rats were dosed with dexamethasone (per os) at 1 h before, and 7 h after, antigen inhalation. Data are expressed as the mean ± SEM (n = 10) of the percentage change from the saline-challenged controls. The extravascular volume for the vehicle-administered control animals is plotted separately for comparison.

Effect of Anti-Rat ICAM-1 mAb

When the sensitized rats were given intraperitoneally a cumulative 6-mg/kg dose of anti-rat ICAM-1 mAb, 1A29, in three equal doses at 12 and 1 h before, and 7 h after, antigen inhalation, the late-phase eosinophilia was significantly inhibited. Table 1 shows the numbers and percentages of eosinophils recovered in the BAL samples. These data indicate that 1A29 when administered systemically led to a significant reduction in the number (P < 0.05) and percentage (P < 0.01) of eosinophils recovered in the BAL fluid relative to those of the control antibody. The pattern of circulating leukocytes was altered depending on the antibody administered as indicated in Table 2. Anti-ICAM-1 antibody, 1A29, caused a 288.7 ± 59.2% elevation in neutrophils relative to the antigen-challenged control animals, which was highly significant (P < 0.01). When the percentages of leukocytes were examined, not only were the neutrophils elevated but there was a significant reduction in the number of circulating lymphocytes (P < 0.01). The percentage of circulating eosinophils was not changed by the treatment. The cell infiltration data confirmed the results of a previous study by Richards and coworkers (27), which showed that the anti-ICAM-1 mAb reduced the eosinophilia in OA-sensitized and challenged rats.

However, when the 24-h microvascular leakage response was measured, the data (Figure 4) showed that the anti-ICAM-1 mAb did not inhibit the antigen-induced increase in vascular permeability. The permeability experiments were performed in separate groups of 10 animals to avoid the complication induced by the lavage procedure. The extravascular plasma volume in sensitized rats challenged with saline was 1.95 ± 0.19 ml, which increased to 5.03 ± 0.34 ml after exposure to aerosol OA. The administration of nonblocking mAb P-23 had no effect and the administration of anti-ICAM-1 mAb 1A29 further increased the extravascular plasma volume to 7.66 ± 0.92 ml. The increase in edema response was significant (P < 0.03) when compared with the control antibody (P-23) group. The measurement of circulating plasma concentrations of 1A29 and the isotype-matched IgG₁ mAbs by ELISA indicated that identical concentrations of mAbs existed in the blood of the rats used in both the edema and the BAL experiment groups (Figure 5). From these data, it is clear that reduction of the antigen challenge-induced eosinophilia did not result in a parallel reduction in the increased microvascular permeability, and may even exacerbate the plasma leakage response.

View larger version (17K): [in this window] [in a new window]   Figure 4.   Effect of intraperitoneal administration of anti-ICAM-1 (1A29, upper panel ) or anti-4 integrin (TA-2, lower panel ) (total of 6 mg/kg, divided into three equal doses at 12 h before, 1 h before, and 7 h after, aerosol antigen challenge) on OA-induced pulmonary vascular leakage 24 h after antigen exposure in sensitized brown Norway rats. P-23, a nonblocking isotype-matched control mAb, was used at 6 mg/kg as a control. Data are expressed as described for Figure 1 (n = 10).

View larger version (13K): [in this window] [in a new window]   Figure 5.   Plasma concentrations of anti-ICAM-1 mAb (1A29), anti-4 integrin mAb (TA-2), and the nonblocking isotypic control (P-23) after in vivo administration in brown Norway rats. Groups of 10 OA-sensitized and challenged animals were injected intraperitoneally with 3 equal doses of each individual antibody. The rats were sacrificed in all experiments at 24 h after aerosol antigen challenge. The blood samples were collected shortly before sacrifice and the plasma levels of mAbs were analyzed for murine IgG1 by an ELISA. The results are plasma concentrations of each individual animal from the experiment indicated. The means were calculated from data points in each experimental group and are indicated by the horizontal line through each set of data points.

Effect of Anti-Rat ₄ Integrin mAb

Similar results were obtained when the OA-challenged, sensitized brown Norway rats were treated with an anti-₄ integrin mAb, TA-2, using the same dosing regime as that used with the anti-ICAM-1 mAb (cumulative doses, 6 mg/ kg). Administration of TA-2 resulted in a significant decrease in the OA-induced late-phase eosinophilia from 35.6 ± 6% and 36.3 ± 4% in both positive control and isotype-matched control groups, respectively, to 15.4 ± 4.2% in the TA-2-treated group (Table ). In this experiment, the total eosinophil counts in the BAL were significantly decreased. The effect of TA-2 on vascular leakage of plasma was similar to that of 1A29 (Figure 4); animals treated with anti-₄ integrin mAb did not show a significant difference from controls (P > 0.05). The control IgG₁, P-23, did cause a slight decrease in plasma leakage (8.38 ± 1.01 ml/g for OA/OA versus 6.38 ± 0.61 ml/g for P-23; P < 0.06) and the difference between the P-23 and TA-22 antibody-treated animals (10.86 ± 2.11 ml/g) indicated a significant increase in edema in the lungs after anti-₄ integrin mAb treatment (P < 0.03). However, when the extravascular plasma volumes in TA-2-treated rats were compared to those of the OA-sensitized and challenged control animals there was no statistical difference. Again, the circulating plasma concentrations of TA-2 and P-23 in the lung edema rats were identical to those that inhibited eosinophilia in the BAL experiments (Figure 5). The administration of TA-2 led to a significant decrease in migration of eosinophils (71.7 ± 12.2%; P < 0.01) into the lung following OA challenge of (Table 1B). As shown in Table , the anti-₄ integrin mAb caused an alteration in the pattern of circulating leukocytes resulting in an increase in the numbers of neutrophils (217.4 ± 43.7%; P < 0.01) and lymphocytes (225.4 ± 43.2%; P < 0.01). However, these changes reflect an increase in the total number of white cells in circulation because the percentage of these cells did not significantly change. Previous investigators have also shown that anti-₄ integrin mAbs induced increases in circulating leukocytes in the mouse (30), rat (28), and primate (31). As with anti-ICAM-1 mAb treatment, it is clear that the reduction of the antigen challenge-induced eosinophilia was not associated with inhibition of the increased microvascular permeability.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Our results demonstrated that aerosol antigen challenge of OA-sensitized brown Norway rats induced a late-phase glucocorticoid-sensitive plasma leakage and lung edema. The onset of plasma leakage roughly paralleled the accumulation of eosinophils in the bronchoalveolar lavage fluid. Administration of mAbs against adhesion molecule ICAM-1 or ₄ significantly reduced the eosinophils accumulating in BAL fluid, but rather than inhibiting plasma leakage into the lungs, these mAbs actually led to small but significant increases in plasma in the lungs. The leukocyte data reconfirmed the results obtained in our previous studies (27, 28).

The increase in vascular permeability and the leakage of plasma have been postulated to operate by at least two different mechanisms: (1) a direct action of inflammatory mediators on endothelial cells or (2) the action of inflammatory cells recruited to the site of inflammation. Leukocyte-dependent edema formation has been described in several animal models such as the contact sensitivity reaction in mice (32), lung injury after thermal trauma of skin or ischemia-reperfusion of hind limbs (33, 34), in acute rat lung injury induced by cobra venom factor or immune complexes (35), as well as in an air-pouch, allergic inflammation model in rats (36). In other models, the edema formation was clearly independent of the presence of inflammatory cells. For example, anti-CD18 mAbs effectively inhibited neutrophil accumulation induced by platelet-activating factor (PAF), zymosan-activated plasma (ZAP), and eosinophil accumulation in a passive cutaneous anaphylactic reaction in guinea-pig skin, but did not affect edema formation in the same site (37, 38). The administration of mAbs against CD11a or CD18 reduced the cellular influx in several rabbit models of uveitis (39) yet the mAbs were unable to reduce the associated increase in vascular permeability as measured by protein in the aqueous humor. Apparently, the plasma leakage and lung edema in the antigen-challenged rats appear not to be critically dependent on the presence of eosinophils in the airway lumen.

In addition to the evidence of leakage of ¹²⁵I-labeled albumin that we have presented here, histologic examination of the inflamed lung tissue demonstrates evidence of microvascular leakage with increased interstitial space apparent around the arterioles and alveoli. The antigen-challenged allergic rats displayed the greatest evidence of edema around small blood vessels, as demonstrated in Figure 2. The evidence of edema was most apparent in the interstitial spaces surrounding small blood vessels, where large voids were formed separating the surrounding airways in the lungs of OA-challenged sensitized rats. These areas were observed to be scattered throughout inflamed lungs but were rarely seen in the lungs of nonsensitized rats. These micrographs are analogous to those included in a treatise on the pathology of asthma (40) to illustrate edema in asthmatic lung. In a model of Ascaris challenge-induced pulmonary inflammation in sensitized guinea pigs (41), investigators reported histologic evidence of narrowing of the bronchial lumen due to infolding of the epithelial layer as evidence of peribronchial edema during the late-phase reaction. This may indicate a more advanced stage of edema than we observed in these studies. Species differences may also account for the different manifestations of this aspect of the inflammatory reaction observed in these antigen-driven models. However, it is clear from both of these observations that microvascular leakage and the ensuing tissue edema are significant contributors to the pathogenesis of antigen-induced airway inflammation. Additional histologic evidence for the involvement of microvascular leakage in pulmonary inflammation has been presented in a study in which sheep with pulmonary artery occlusion were dosed intravenously with histamine, resulting in perivascular, peribronchial, and alveolar edema (42).

Reports on the eosinophil-dependent changes of tissue microvascular permeability are scant. Time course or dose- response studies of antigen-induced airway inflammation in guinea pigs (43) or in humans (44) suggest that the recruitment of eosinophils is closely associated with plasma exudation. Yoshikawa and coworkers (45) showed that phorbol myristate acetate (PMA)-activated eosinophils induced increases in microvascular permeability in isolated and perfused rat lungs that could be blocked by inhibitors of eosinophil peroxidase. However, our data clearly demonstrate that the antigen-induced eosinophilia and plasma leakage in brown Norway rats may be separate events involving different mechanisms, even though they occurred in parallel. These results support the findings of previous studies of PAF or leukotriene B₄ (LTB₄)-induced eosinophilia in guinea pig skin (4) or IgE-mediated protein exudation in rats (46). Both of these studies showed a dissociation between eosinophilia and plasma leakage in inflammation. In the study of Bandeira-Melo and coworkers (46), the presence of eosinophils in the pleural space of rats was reported to play an anti-inflammatory role by downregulating allergen-induced plasma exudation. Furthermore, Reed and coworkers (47) reported that blockade of ₁ integrin function (VLA-4) in rat skin decreased interstitial fluid pressure and produced an edema response. These findings corroborate our observation that suppressing airway eosinophils by blocking cell adhesion may elevate plasma leakage in this model.

Several preformed and newly synthesized mediators including histamine, sulfidopeptide leukotrienes, platelet-activating factor, bradykinin, and tachykinins are known to induce plasma extravasation in the airways of animal species including the rat (48). Some of these mediators are produced by eosinophils and others are derived from mast cells, airway epithelial cells, sensory afferent nerves, or from plasma itself by the activation of the kinin and complement systems. For example, Martin and coworkers (18) showed that the level of sulfidopeptide leukotrienes excreted in the bile of sensitized brown Norway rats increased fourfold 4-8 h after antigen challenge. In vitro studies (49) revealed that alveolar macrophages, rather than eosinophils, were the likely source of the increased synthesis of sulfidopeptide leukotrienes. Furthermore, eosinophils present at the site of inflammation may play a downregulatory role in the hypersensitivity reaction. Specific enzymes contained in eosinophils are capable of inactivating mast cell-derived mediators and reduce their effects. Eosinophils contain histaminase (50), which inactivates histamine, and phospholipase D (51), which breaks down platelet-activating factor. Calcium ionophore-stimulated human or horse eosinophils transform sulfidopeptide leukotrienes to inactive species via peroxidase action (52, 53). Furthermore, granulocytes in general possess peptidases that break down peptido-leukotrienes (54). It is possible that the decreased eosinophil infiltration into the lung tissue by the antibodies partially eliminated modulatory factors leading to exaggerated plasma extravasation in response to the inflammatory mediators.

Previous investigators have studied the roles of ICAM-1 (10) or VLA-4, LFA-1, and Mac-1 (11) using mAbs in rat models of lung inflammation. In these studies with intravenous doses of 6 mg/kg, the mAbs inhibited bronchial hyperresponsiveness without reducing the influx of inflammatory cells in the lavage. Conversely, Milne and Piper (8) reported that two mAbs against CD18 at 3 mg/kg reduced the number of eosinophils recovered from BAL fluid of OA-sensitized and antigen-challenged guinea pigs, but only one of the two was effective in reducing bronchial hyperresponsiveness to acetylcholine or histamine. Although the correlation of BAL eosinophilia and hyperresponsiveness is a debatable issue, it is apparent that different protocols used in different laboratories may yield conflicting results. In fact, our preliminary studies using mAb 1A29 at an intravenous dose of 3 mg/kg failed to reduce either BAL eosinophilia or plasma leakage in the brown Norway rat model. In a separate study, the same mAb completely inhibited neutrophil infiltration at an intravenous dose of 1 mg/kg in an acute acetic acid-induced colitis in rats (55). Similarly, intravenous administration of anti-₄ integrin mAb TA-2 at 1 mg/kg also failed to affect either eosinophilia or plasma leakage responses in the brown Norway rat model. The failure of intravenous administration of mAbs to reduce cell infiltration into the airways may be related to their rapid clearance from the body. ELISA analyses of murine IgG in rat plasma showed the circulating level of 1A29 to be, on average, 2.34 µg/ml after three intraperitoneal 1-mg/kg doses of 1A29, which was much greater than the average of 0.44 µg/ml after one intravenous dose of 3 mg/kg over the same time period. At the effective intraperitoneal doses of 6 mg/kg, the mean plasma level of TA-2 at 24 h after antigen challenge was 4.29 µg/ml, which exceeded the concentration of 1 µg/ml required to completely inhibit the VLA-4-dependent adhesion of rat basophilic leukemia cell 1 (RBL-1) to human CS-1 fragment of fibronectin or recombinant human VCAM-1 in vitro (data not shown).

It was previously reported by Richards and coworkers that the same lot of anti-ICAM-1 antibody, 1A29 (27), when administered via the same dosing regime as that employed here, resulted in a significant reduction in the migration of eosinophils into the lungs of OA-challenged sensitized BN rats. The results reported here confirm the previous study indicating that anti-ICAM-1 causes a reduction in the late-phase eosinophilia and indicating that this antibody was effective in these animals.

It was noted that although treatment with 1A29 or TA-2 significantly decreased the numbers and percentage of eosinophils in the BAL samples after antigen challenge, they were less efficacious than dexamethasone. The mAb-treated animals still manifested a significant antigen-induced BAL eosinophilia when compared with the nonsensitized controls. Because eosinophils express multiple adhesion molecules, it is possible that the treatment with specific mAb limited the migration of only one population of eosinophils, while allowing others to migrate into the airway lumen. The plasma leakage may occur as long as there are inflammatory cells transmigrating through the endothelium. The effectiveness of dexamethasone in blocking both the eosinophilia and the plasma leakage may result from the complete elimination by the drug of cell migration into the lung.

In summary, we have shown that aerosol antigen challenge of OA-sensitized brown Norway rats induced an airway inflammatory response characterized by both elevated plasma leakage and an accumulation of eosinophils in the airway lumen. Blockade of the adhesion molecule ICAM-1 or ₄ integrin with mAbs partially limited eosinophil recruitment without reducing plasma leakage. The results suggest that increased plasma leakage occurs independently of BAL eosinophilia.