Introduction

Abstract Introduction References

Recruitment of leukocytes to inflammatory reactions in the systemic circulation is believed to be regulated by the sequential engagement of leukocyte-endothelial cell adhesion molecules. The selectins mediate "rolling" of leukocytes on the vascular endothelium (1), whereas integrins enable firm leukocyte adhesion and transmigration across the endothelium (2). The 2 integrins lymphocyte function-associated antigen (LFA)-1 (L2, CD11a/CD18) and Mac-1 (M2, CD11b/CD18) are particularly important in this process. LFA-1 is expressed by all leukocytes and binds to the immunoglobulin superfamily (IgS) molecules intercellular cell adhesion molecule (ICAM)-1 and -2 on the endothelium. In allergic inflammation, ICAM-1 is also expressed on the bronchial epithelium (3, 4). Mac-1 is expressed mainly by granulocytes and monocytes and binds to various molecules, including ICAM-1, fibrinogen, the complement component C3bi, factor X, and heparin and heparan sulfate (2, 5), the latter being a constituent of many cell-surface and matrix-associated proteoglycans. The 1 integrin very late activation antigen-4 (VLA-4) (41, CD49d/CD29) has also been shown to mediate adhesion and to serve as an alternative mechanism to CD11/ CD18 integrins for monocyte and lymphocyte migration (6).

There is accumulating evidence, especially in the case of the neutrophil, that the mechanisms for emigration in the pulmonary vasculature differ from those in the systemic circulation, and may involve selectin-independent (12) and CD18-dependent or -independent pathways, depending on the nature (13, 14) as well as the dose of the inflammatory stimulus (15). Eosinophils are the predominant inflammatory cells in the lungs in allergic asthma, and they may contribute to tissue damage and the development of nonspecific bronchial hyperresponsiveness (16). Although neutrophils are not commonly found in biopsy specimens of asthmatics (17), a role for neutrophils has been suggested in cases of sudden-onset fatal asthma (17, 18) and in several animal models of allergen-induced lung inflammation and bronchial hyperresponsiveness (19). It has been suggested that VLA-4, which is expressed by eosinophils but not normal human blood neutrophils (22), mediates selective eosinophil recruitment in allergic reactions by interacting with vascular cell adhesion molecule-1, an inducible IgS member on the vascular endothelium (25). VLA-4 also binds to an alternatively spliced connecting segment, CS-1, of the extracellular matrix protein fibronectin. This interaction may promote eosinophil survival (28) and augment bronchial narrowing via activation of 5-lipoxygenase pathways (29). Very recently, however, it was discovered that VLA-4 can also be expressed by human blood neutrophils under certain conditions (30). VLA-4 is also present at low levels on rat blood neutrophils, and it can mediate migration of these leukocytes to arthritic joints and to dermal inflammation in this species (31). It is not known, however, whether VLA-4 plays a role in neutrophil recruitment to allergic or any other form of lung inflammation.

During lung inflammation, leukocytes accumulate in the interstitium and migrate into the air space. This is a complex process that involves diapedesis through the blood vessel endothelium, directed movement in the interstitium, and eventually transmigration through the epithelial barrier. How these processes are regulated and what adhesion molecule mechanisms are involved at each step are currently not well defined. To investigate these mechanisms for eosinophil and neutrophil migration, we recently developed techniques for quantitating eosinophils and neutrophils in the lung parenchyma of rats during allergen-induced inflammation, using highly specific peroxidase enzyme assays optimized for the rat (32). Using this approach, we investigated the role of the 4 (CD49d) and 2 (CD18) integrins in eosinophil and neutrophil migration into the interstitium and into the air space by using adhesion function-blocking anti-4 and anti-2 monoclonal antibodies (mAbs) in a rat model of allergic lung inflammation (33, 34).

	Materials and Methods

Animals

Male, inbred, Brown Norway (BN/SSN) rats weighing 125 to 149 g were purchased from Harlan-Sprague Dawley, Inc. (Indianapolis, IN). The animal experiments were in accordance with protocols approved by our University Committee on Laboratory Animals.

mAbs

The mouse antirat 4 (CD49d) mAb TA-2 immunoglobulin G1 (IgG1) was produced by immunization of BALB/c mice with rat peritoneal lymphocytes as described previously (9). It blocks VLA-4-dependent lymphocyte and monocyte adhesion in vitro and migration in vivo (8, 9, 11, 35), and reacts with VLA-4 on rat neutrophils (31) and eosinophils (36). The TA-2 mAb was used as both an intact and F(ab')₂ form generated using standard acid-pepsin digestion of the mAb and analysis of fragments by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Antirat CD18 mAb WT.3 was a kind gift from Dr. M. Miyasaka (Osaka, Japan). This IgG1 mAb blocks adhesion function of both LFA-1 (CD11a/CD18) and Mac-1 (CD11b/CD18) (37). mAb B9 (IgG1), which reacts with pertussis toxin (38), was used as a nonbinding negative control.

Sensitization and Allergen Challenge of BN Rats

Ovalbumin (OA), grade V (Sigma Chemical Co., St. Louis, MO), was prepared at 2 mg/ml in 0.9% sterile, pyrogen-free NaCl and precipitated at a 1:1 ratio with Al(OH)₃ (45 mg/ml) (Imject Alum; Pierce, Rockford, IL), following the manufacturer's instructions. Rats were immunized with 1 mg OA (1 ml OA/Al[OH]₃ suspension) given subcutaneously at two sites on the back of the neck, and 10¹⁰ heat-killed Bordetella pertussis bacilli (gift from S. Halperin, Halifax, NS, Canada) in 0.5 ml saline given intraperitoneally (i.p.) as an adjuvant, following the sensitization procedure of Renzi and colleagues (39). Sham immunization omitted the OA in the Al(OH)₃. After 14 d, rats were placed in a Plexiglas chamber (21 liters) and an aerosol of 0.5% OA in saline was delivered for 1 h at 3 liters/ min by an ultrasonic nebulizer (Monaghan 670; Monaghan Co., Littleton, CO) at setting 7, with a mean particle size of 2.5 ± 1 µm.

In Vivo Treatment with mAbs

Rats were injected intravenously (i.v.) with mAbs 15 min before aerosol challenge. The anti-CD18 mAb (WT.3) and negative control mAb (B9) were given at a dose of 20 mg/ kg. TA-2 was used in a dose of 10 mg/kg, and for a separate group of animals as purified F(ab')₂ fragments in a dose of 10 mg/kg. There was no difference between a group of rats treated with negative control mAb B9 or a further control group that received no mAb. The results from these two groups were therefore combined.

Dissection and Lung Perfusion

All rats were killed 24 h after challenge. Rats were premedicated with 0.2 mg/kg atropine sulfate (Astra Pharma Inc., Mississauga, ON, Canada) injected i.p. 5 min before anesthesia with i.p. administration of ketamine (50 mg/kg; Warner-Lambert Canada Inc., Scarborough, ON, Canada) and xylazine (10 mg/kg, Rompun; Chemagro Ltd., Etobicoke, ON, Canada). The abdominal cavity was opened, a 25-gauge butterfly needle was inserted into the inferior vena cava, and 1 ml of blood was collected in ethylenediaminetetraacetic acid (EDTA) for blood cell counts and analysis of plasma mAb level. This was followed by injection of 100 U of heparin in 1 ml saline. The abdominal aorta was severed, and at the same time Tyrode's (pH 7.4, 37°C) solution was infused into the inferior vena cava. After 25 ml of Tyrode's solution, the chest was opened and another 25 ml were infused via the vena cava above the diaphragm, followed by 10 ml phosphate-buffered saline (PBS), pH 7.4, containing 0.1% EDTA. This protocol consistently cleared the lung vasculature of blood cells (40).

Bronchoalveolar Lavage (BAL) and BAL Cell Determination

The trachea was cannulated and the lungs were lavaged four times (7 ml/lavage) with cold (4°C), Ca²⁺- and Mg²⁺-free PBS containing 0.1% EDTA. Cells in the BAL fluid (BALF) were sedimented by centrifugation (10 min at 200 × g, 4°C) and resuspended in PBS. Total leukocytes were determined by counting on a hemocytometer using crystal violet stain, and eosinophils were counted using 0.05% phloxine B in 50% propylene glycol (Sigma) in water (41). Cytocentrifuge preparations of the BALF leukocytes were stained with Diff-Quik (Fisher Scientific, Mississauga, ON), and at least 200 cells were differentiated according to standard morphologic criteria. Total BAL eosinophils are reported on the basis of the phloxine B stain and were in good agreement with counts calculated from the differential on the stained cytocentrifuge preparations and the total leukocyte count. BALF protein content was estimated spectrophotometrically by absorbance (optical density, OD) at 280 nm, and 1 OD unit was estimated to be equivalent to 1 mg/ml total protein.

Quantitation of Lung Tissue Eosinophils and Neutrophils

Lung tissue eosinophils and neutrophils were examined as described previously (32). Briefly, after BAL the lungs were removed, separated into the individual lobes, and weighed after all extrapulmonary airway tissue was trimmed away. Samples of parenchyma from each lobe, approximating 20% of the total lung wet weight, were pooled, stored at 70°C, and later freeze-dried. For enzyme extraction, lyophilized samples were homogenized in 50 mM N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid (Hepes), pH 8.0, at 0.5% dry wt/vol with a pestle homogenizer (Talboys Engineering Corp., Emerson, NJ) and centrifuged at 10,000 × g for 30 min at 4°C, and the supernatant was discarded. The pellet was resuspended in 0.5% cetyltrimethylammonium chloride in distilled water to the original volume, rehomogenized, and centrifuged again as before. An aliquot of the supernatant was taken for analysis of eosinophil peroxidase (EPO) and myeloperoxidase (MPO) activity.

Lung extracts were diluted 1/10 in 50 mM Hepes, pH 8.0 (EPO-dilution buffer), or 10 mM citrate buffer, pH 5.0 (MPO-dilution buffer). Aliquots of 75 µl of each sample were pipetted into four wells of a 96-well tissue culture plate. Cold stop solution (4 N H₂SO₄, containing also 2 mM resorcinol for the EPO assay) was added to two wells (150 µl/well) to stop the reaction at t = 0 s (background OD). The EPO-substrate solution consisted of 50 mM Hepes (pH 8.0), 6 mM KBr, 3 mM ortho-phenylenediamine (OPD), and 8.8 mM H₂O₂. The MPO-substrate solution was 3 mM 3',5,5'-tetramethylbenzidine dihydrochloride, 120 µM resorcinol, and 2.2 mM H₂O₂ in distilled water. Substrate solution (75 µl) was added to each well, and the reaction was stopped after 30 s (EPO) or 2 min (MPO) with 150 µl of cold stop solution. The OD_490nm (EPO) or OD_450nm (MPO) was determined with a Thermomax microplate reader (Molecular Devices Corp., Menlo Park, CA). As an additional control, 75 µl of dilution buffer (without lung extract) was placed into four wells and 75 µl of substrate buffer was added, followed by 150 µl stop solution after 30 s or 2 min. No color reaction was observed in these controls. All reagents were used at room temperature and the reaction was carried out at 22°C. The enzyme activities of the lung samples were calculated by subtracting the mean background OD and are expressed as change of OD/min.

Standard curves for calculating the numbers of eosinophils and neutrophils in the lungs on the basis of the enzyme activities were developed as previously described (32). Briefly, 10 × 10⁶ BAL eosinophils were injected into a piece of noninflamed control lung, and this tissue was frozen, lyophilized, and extracted as described above. The EPO activity of the resulting extract was tested at different dilutions, and the activity was plotted against the theoretical eosinophil equivalents in the dilution of lung extract. This standard curve was used to estimate the eosinophil number in the test lung extracts and, on the basis of the weight, in the lungs. In the extract from the eosinophil- injected control lung, the MPO activity was also measured and correlated with the EPO activity. The resulting linear regression was used as a correction of EPO spillover (approximately 5 to 6%) to the MPO assay (activity) in lung extracts, as shown previously (32). The neutrophil content of the lungs was estimated again as previously, that is, by injection of purified neutrophils (in this study 53 × 10⁶) into control lung tissue, and then this tissue was lyophilized and extracted. Different dilutions of this extract were analyzed for MPO activity, and this standard curve was used for estimation of the neutrophil content of the lungs. There was no MPO activity detected using the EPO assay conditions (32).

Histology

Freshly removed lung tissue was fixed in 10% phosphate-buffered neutral formalin. After fixation, samples were cut from the hilus to the periphery and embedded in paraffin. Sections were cut at 5-µm thickness and stained with hematoxylin and eosin. For preparation of semithin sections, lungs were first perfused via the pulmonary artery and BAL was performed as usual. The lungs were then fixed in situ by 10 to 15 min of slow perfusion for whole lungs with 1% glutaraldehyde in cacodylate buffer (0.1 M, pH 7.3). Representative samples of lung tissue were embedded in epon (TAAB resin; Marivac Ltd., Halifax, NS, Canada), and semithin (2-µm) sections were prepared and stained with toluidine blue.

Enumeration of Blood Leukocytes

Blood leukocyte counts were determined on fresh EDTA blood after staining with crystal violet. Eosinophils were quantified with 0.05% phloxine B in 50% aqueous propylene glycol and counting in a hemacytometer (41).

Measurement of Circulating mAb Levels in Plasmas

Circulating levels of murine mAb in the plasmas of single mAb-treated rats, collected at the time of death, were measured by enzyme-linked immunosorbent assay (ELISA) on 96-well plates (Maxisorp-Nunc; Canadian Life Technologies Inc., Burlington, ON, Canada). Wells were coated overnight at 4°C with goat IgG antimouse -chain (5 µg/ml in PBS, pH 7.4), pre-adsorbed to rat serum proteins (Sigma), and blocked with 10% goat serum in PBS. Dilutions of samples and standard (mouse IgG in normal rat plasma) were added to the wells, incubated for 1 h, and washed. The wells were then incubated with horseradish peroxidase-conjugated goat antimouse IgG (0.1 µg/ml in PBS/5% goat serum and 5% normal rat serum). Bound enzyme was detected with OPD and H₂O₂ in 0.15 M citrate/phosphate buffer, pH 5.4. The reaction was stopped with 4 N H₂SO₄, and absorption was read at 490 nm with a Thermomax microplate reader (Molecular Devices Corp.).

Data Analysis

All data are reported as arithmetic means ± 1 SEM. The effects of mAb treatment were analyzed by the nonparametric Kruskal-Wallis test (42). When this test indicated significant differences between treatments, the Mann- Whitney U test was used to compare individual groups (42).

	Results

Effect of Anti-4 and Anti-2 mAbs on Inflammatory Cell Accumulation in BALF

OA exposure of sensitized BN rats, but not of sham-sensitized rats, led to a marked accumulation of neutrophils and eosinophils and a few lymphocytes (not shown) at 24 h after challenge. Treatment of rats with intact anti-4 (CD49d) mAb TA-2 had no significant effect on granulocyte accumulation. In contrast, TA-2 administered as a F(ab')₂ fragment significantly inhibited the accumulation of eosinophils (by 66%) and neutrophils (by 57%) in the BALF, as shown in Figure 1. Blocking 2 (CD18) integrins with the WT.3 mAb significantly inhibited eosinophil (by 54%) and neutrophil (by 55%) accumulation. Additional experiments with the F(ab')₂ form of WT.3 showed no difference between intact and F(ab')₂ fragments for this mAb. Combined treatment with WT.3 plus TA-2 F(ab')₂ led to a significantly stronger inhibition of both eosinophil and neutrophil accumulation. BAL eosinophil accumulation was virtually abolished (98% inhibition) and neutrophil accumulation was significantly further reduced (by 91%, Figure 1).

View larger version (33K): [in this window] [in a new window]   Figure 1.   Effect of anti-4 or -2 integrin mAb treatment on inflammatory cell accumulation in BALF after allergen challenge. Rats were injected i.v. with control mAb (B9), anti-4 (CD49d) mAb TA-2 in an intact as well as an F(ab')2 form, anti-2 (CD18) mAb WT.3, or a combination of WT.3 plus TA-2 F(ab')2 15 min before allergen challenge. BAL was performed at 24 h after challenge, and total BALF cells were enumerated and differentiated. +P < 0.05; *P < 0.01; #P < 0.001 compared with control mAb treatment or as indicated. n = 7-14, except for control mAb (n = 25).

Effect of mAb Treatments on Granulocyte Accumulation in Lung Parenchyma

To determine the role of 4 and 2 integrins in the accumulation of granulocytes in the lung parenchyma, tissue eosinophils and neutrophils were quantitated by specific EPO and MPO enzyme assays on lung tissue extracts as described in MATERIALS AND METHODS. The results in Figure 2 show that administration of intact anti-4 (CD49d) mAb TA-2 had no effect on either eosinophil or neutrophil accumulation. However, this mAb as an F(ab)₂ fragment partially but significantly inhibited (by 48%) the accumulation of eosinophils in lung parenchyma. The accumulation of neutrophils also appeared slightly reduced, although this did not reach significance (P = 0.2). Blockade of CD18 with the WT.3 mAb reduced eosinophil accumulation by 48%, whereas the combination of WT.3 plus TA-2 F(ab')₂ abolished eosinophil accumulation in the lung parenchyma to the level of the sham-sensitized control group (95% inhibition, Figure 2). This inhibition was significantly greater than that with WT.3 alone or with TA-2 F(ab')₂ alone. In contrast to the eosinophils, the accumulation of neutrophils in the parenchyma increased (by 55%; P < 0.05) with anti-CD18 mAb treatment, and this effect was even stronger in the WT.3 (anti-CD18) plus TA-2 F(ab')₂ (anti-CD49d) double-treated group.

View larger version (46K): [in this window] [in a new window]   Figure 2.   Effect of mAb treatment on accumulation of eosinophils and neutrophils in lung. Rats were given control mAb (B9), anti-CD49d mAb TA-2 in an intact as well as an F(ab')2 form, anti-CD18 mAb WT.3, or a combination of WT.3 plus TA-2 F(ab')2 i.v. 15 min before allergen challenge. At 24 h later, the pulmonary vasculature was perfused, BAL was performed, and eosinophils and neutrophils in parenchyma were quantitated by specific EPO or MPO enzyme assays. Values are total eosinophils or neutrophils in both lungs. +P < 0.05; *P < 0.01; #P < 0.001 compared with control mAb treatment or as indicated. n = 7-14, except for control mAb (n = 25).

We have recently reported in detail on the lung tissue histopathology in this model (40). At 24 h after challenge, there is an intense eosinophilic and to a lesser extent neutrophilic infiltration, especially in the perivascular and peribronchial spaces. Many eosinophilic cell aggregates are also observed. As shown in the semithin sections of perfusion-fixed lungs (Figure 3), the size and number of these aggregates were markedly reduced by treatment with mAb to CD18 (WT.3) (Figures 3C and 3D) or by mAb to CD49d (TA-2 F[ab]₂) (Figures 3E and 3F) alone. Essentially no aggregates were seen in the anti-CD18 plus anti-CD49d-treated group (Figures 3G and 3H). Thus, the histologic picture was in good agreement with the BALF and lung EPO results. In the anti-CD18 and the anti-CD18 plus anti-CD49d-treated group, there remained only a diffuse interstitial, alveolar neutrophilic leukocyte accumulation that was very different from the intense eosinophilic infiltrates and aggregates (not shown).

View larger version (123K): [in this window] [in a new window]   Figure 3.   Histology of pulmonary inflammation and effect of anti-2 (CD18) and anti-4 (CD49d) antibodies. Partially inflated lungs from OA-challenged rats were perfused free of intravascular leukocytes and then fixed with gluteraldehyde. Plastic-embedded semithin sections were stained with toluidine blue. (A) and (B) from a control mAb (B9)-treated rat show the characteristic peribronchial, perivascular, and some interstitial leukocytic infiltrate composed of  70% eosinophils, accompanied by many eosinophilic leukocyte aggregates as described in detail previously (40). (C) and (D) are from a rat treated with anti-CD18 (2) mAb (WT.3), and (E) and (F) from an anti-CD49d F(ab)2 mAb (TA-2)-treated animal. With either treatment, there is a notable decrease in leukocyte aggregates, although some peribronchial and perivascular leukocytes and inflammation are still present. The rat in (G) and (H) was treated with mAbs to both CD18 and CD49d prior to OA challenge. These sections show virtually no inflammatory-cell aggregates. Magnification: ×50 in A, C, E, and G; ×125 in B, D, F, and H.

Effect of mAb Treatments on BALF Protein Content

The BALF protein content is a measure of vascular and bronchoalveolar permeability and lung edema, a characteristic feature of allergic asthma (43), which can potentiate airway reactivity (44). We have recently reported that the BALF protein content significantly increases after allergen challenge of sensitized BN rats (40, 45). Anti-CD49d mAb (TA-2 F[ab']₂) significantly reduced BALF protein content (by 40%) at 24 h after challenge, as shown in Figure 4. Blockade of CD18 with WT.3 had no significant effect. The combination of mAb to CD18 plus CD49d further reduced BALF protein 24 h after allergen challenge, and this inhibitory effect was significantly greater than with anti-CD49d treatment alone.

View larger version (51K): [in this window] [in a new window]   Figure 4.   Effect of mAb treatment on BALF protein. Rats were treated i.v. with control mAb (B9), anti-CD49d mAb TA-2 in an intact as well as an F(ab')2 form, anti-CD18 mAb WT.3, or a combination of WT.3 plus TA-2 F(ab')2 15 min before allergen challenge. BALF was performed 24 h after allergen challenge, cells were sedimented by centrifugation, and the protein content in the supernatant was determined by absorbance at 280 nm. +P < 0.05; *P < 0.01 compared with control mAb treatment or as indicated. n = 7-14, except for control mAb (n = 25).

Effect of mAb Treatments on Blood Leukocyte Counts

The effect of mAb treatments on the blood eosinophil count is shown in Figure 5A. The eosinophil count was very low in rats that received no or control mAb (14 ± 4 eosinophils/µl). It increased significantly with anti-CD49d (intact IgG), anti-CD18, and with combined mAb treatment (anti-CD18 + anti-CD49d F[ab]₂). As shown in Figure 5B, the blood mononuclear cell counts were significantly increased in all groups compared with the control mAb group. Blood neutrophil counts (Figure 5B) were increased in all mAb-treated groups as well, with the greatest increases seen during anti-CD18 treatment alone and when combined with anti-CD49d mAb.

View larger version (35K): [in this window] [in a new window]   Figure 5.   Effect of mAb treatment on blood leukocyte counts. Rats were given control mAb (B9), anti-CD49d mAb TA-2 in an intact as well as an F(ab')2 form, anti-CD18 mAb WT.3, or a combination of WT.3 plus TA-2 F(ab')2 i.v. 15 min before allergen challenge. Blood samples were collected from the inferior vena cava at 24 h. Leukocytes were enumerated and differentiated by crystal violet and phloxine staining. +P < 0.05; *P < 0.01; #P < 0.001 compared with control mAb treatment or as indicated. n = 7-14, except for control mAb (n = 24).

Plasma mAb Levels at 24 h after Challenge

Two approaches were taken to determine whether saturating levels of mAbs were maintained over the 24-h course of the experiment. The concentration of murine mAb was measured in plasma of the single mAb-treated groups directly by ELISA, as described in MATERIALS AND METHODS. There were 72 ± 9.4 µg/ml of TA-2 (intact), 5 ± 0.6 µg/ml of TA-2 F(ab')₂, and 128 ± 15 µg/ml of WT.3 in the plasma. Additionally, immunofluorescence staining of blood leukocytes from untreated rats with dilutions of test plasma before and after addition of a saturating amount of the reference mAb was performed and the cells were analyzed by flow cytometry. These results show that maximum staining was achieved with plasma dilutions of 1:5 to 1:50, indicating that all plasmas contained concentrations of mAb well in excess of that required for saturation of antigen (results not shown).

	Discussion

In this study we show that both 4 (CD49d) and 2 (CD18) integrins play a role in eosinophil and neutrophil migration into the bronchoalveolar space in allergic inflammation. Blocking either of these adhesion molecules resulted in a significant but only partial reduction of eosinophil and neutrophil migration into the airways as assessed by BAL (Figure 1). However, when both CD49d and CD18 integrins were blocked, inhibition of eosinophil and neutrophil accumulation in the BALF was essentially complete. The inhibitory effects of the mAbs were not due to a reduction in blood leukocyte counts because the number of blood eosinophils, neutrophils, and mononuclear cells increased in all mAb-treated groups (Figure 5). The observed effects may be due to inhibition of eosinophil and neutrophil emigration from the blood or to inhibition of transepithelial migration into the BAL. In the case of eosinophils, the former mechanism appears to be involved because inhibition of eosinophil accumulation in lung parenchyma by anti-CD49d (F[ab]₂) and anti-CD18 mAbs, alone or in combination, paralleled the inhibition of eosinophil accumulation in BALF (Figures 1 and 2). However, the mechanisms of neutrophil and eosinophil recruitment to lung appear, in part, to differ on the basis of the differing effects of CD18 and CD49d blockade on accumulation of these leukocytes in parenchyma (Figure 2). This warrants further discussion, first in relation to CD49d or VLA-4 (4).

Recently, neutrophils in rat and human have been shown to express a low level of VLA-4 (CD49d/CD29) (30, 31). The role of neutrophil VLA-4 in neutrophil recruitment to lung has not been determined. Because neutrophil migration to lung is known to be CD18-independent under some circumstances (13, 14), VLA-4 on neutrophils has become an important alternate mechanism to investigate. To our knowledge, this is the first observation that VLA-4 (CD49d) plays a role in neutrophil recruitment to the air space in allergic lung inflammation. In contrast to the air space, as well as to eosinophil migration into parenchyma, no significant inhibitory effect of the anti-CD49d mAb (F[ab]₂) was observed for neutrophil accumulation in the parenchyma (Figures 1 and 2). This suggests that VLA-4 is involved in interstitial or transepithelial migration of the neutrophil rather than in the transendothelial migration process. This observation may seem surprising because neutrophils have been thought to lack VLA-4. However, VLA-4 on rat blood neutrophils has recently been shown to mediate neutrophil accumulation in cutaneous inflammation and in arthritic joints, especially when CD18 mechanisms are blocked (31). Human blood neutrophils express VLA-4 during hematopoiesis (24), and recently this was shown to be reexpressed upon stimulation with chemotactic factors in the presence of dihydrocytochalasin B or following transendothelial migration (30). An indirect effect of the anti-CD49d mAb on neutrophil recruitment to BALF (e.g., by modulating mast cell activation, as recently suggested [46]) cannot be excluded, although one would have expected such an effect to result also in decreased neutrophil recruitment to parenchyma.

In vivo blockade of CD49d with mAbs has been used in various animal models of antigen-induced lung inflammation and airway responses with varying results. Anti- VLA-4 treatment inhibited both eosinophil accumulation and the increase in bronchial hyperreactivity in a guinea pig model (47). In a sheep model, an anti-CD49d mAb inhibited late-phase allergic airway responses and hyperreactivity without having an effect on inflammatory cell accumulation (48). Blockade of CD49d with the intact IgG form of TA-2 mAb had no effect on the BALF inflammatory cell profile in two other studies in the BN rat, despite inhibition of the late-phase bronchoconstriction (36, 49). In the current study we attempted to define the mechanisms of eosinophil and neutrophil accumulation in the lung because of the poor correlation between attenuation of airway responses and leukocyte infiltration. Our results confirm that mAb TA-2 IgG has no effect on eosinophil recruitment, but they extend previous findings to show that F(ab)₂ fragment blockade of CD49d significantly inhibits this process (Figures 1 and 2). Furthermore, we also observed that CD11/CD18 integrins play an important role in eosinophil recruitment. In contrast, in a recent study by Richards and colleagues (50), intact TA-2 IgG mAb suppressed eosinophil accumulation in BALF and the lung tissue eosinophilia in a somewhat different BN rat model (see below), almost to the levels in control animals as assessed by histology. It is not clear why the effects of CD49d blockade on eosinophil accumulation vary in the same species with allergen inflammation and with the form of anti-CD49d mAb (IgG or F[ab]₂) used. In our hands, the anti-CD49d mAb TA-2 used as intact IgG had no inhibitory effects on granulocyte accumulation in BALF (Figure 1) or in the parenchyma (Figure 2). The use of F(ab')₂ fragments, however, resulted in consistent inhibition of eosinophil and neutrophil accumulation in BALF (Figure 1). The VLA-4 adhesion function-blocking properties of the TA-2 mAb (used as intact IgG) are well characterized in vitro (9) and in rat models in vivo (8, 11, 35). The discrepancy between the two different preparations of TA-2 in this model was surprising. In parallel experiments, the anti-CD18 mAb WT.3 was used as an F(ab')₂ fragment for comparison with intact WT.3 treatment. However, results with these two forms of the anti-CD18 mAb were comparable (not shown). Both mAb TA-2 and WT.3 are IgG1 isotypes. It has been shown that vascular cells, including endothelial cells, express some 4 (CD49d) integrin (51), and rat endothelial cells in culture bind the TA-2 mAb (T. B. Issekutz, unpublished observation). This may account for the differences between the intact and F(ab')₂ forms when inflammation is studied in certain vascular beds, such as the pulmonary microvasculature. In alveolar capillaries, leukocytes come into very close contact with endothelium because capillaries are smaller in diameter than capillaries in the systemic circulation (13). In this setting, Fc-mediated interactions with mAb may play a more prominent role than in peripheral vascular beds, or 4 (CD49d) integrin expression may be higher on lung capillaries. In contrast, no 2 (CD18) integrins are present on vascular endothelium.

Another factor possibly contributing to differences in results with anti-CD49d mAb treatment may be the local immune response and sensitization and challenge protocol, which may influence the biologic response. For example, Richards and associates did not use B. pertussis adjuvant for sensitization of the rats with OA, a treatment which might have led to a higher degree of sensitization in the rats used in our experiments (52). Antigen challenge in that study was for 5 min, compared with 1 h in our study. In the present study, 2.2 × 10⁶ neutrophils accumulated in BALF at 24 h after challenge, whereas in the study by Richards and coworkers, only about 2.7 × 10⁵ neutrophils were found at the same time and eosinophil numbers were about 30% lower compared with the response observed here. This suggests that there are differences in the nature and the extent of the inflammatory response after sensitization with B. pertussis adjuvant and prolonged allergen exposure, and this may have influenced the relative role of 4 (CD49d) integrins in granulocyte migration and thus also the effects of the anti-CD49d mAb TA-2.

To dissect further the adhesion molecules involved in eosinophil and neutrophil migration into the lung, the effect of CD18 integrin blockade with the WT.3 mAb was investigated (Figures 1 and 2). These results show that eosinophil and neutrophil accumulation in the BALF (Figure 1) and eosinophil accumulation in the parenchyma (Figure 2) are partially dependent on the CD18 pathway. The CD18 mechanism apparently serves as a functional alternate to CD49d because both had to be blocked to eliminate eosinophils and neutrophils in BALF and eosinophils in the parenchyma. In contrast, neutrophil accumulation in the lung parenchyma could not be inhibited by blocking CD18 alone or in combination with CD49d blockade (Figure 2). On the contrary, when CD18 alone was blocked or CD18 and CD49d were blocked simultaneously, accumulation of neutrophils in the parenchyma was significantly enhanced (Figure 2). It thus appears that the neutrophil emigration from the blood in this allergen model is largely CD18- and CD49d (VLA-4)-independent. The increase in neutrophil accumulation in the lung parenchyma was probably secondary to the blood neutrophilia (Figure 5) in the mAb-treated animals, which leads to increased delivery of these leukocytes to the lungs with more cells available for CD18- and VLA-4-independent emigration. It should be noted that despite a comparable relative increase in circulating blood eosinophils (Figure 5), the accumulation of eosinophils in the lung parenchyma was completely inhibited by CD18 plus CD49d blockade (Figure 2). These findings indicate that neutrophil recruitment mechanisms to this allergic lung inflammation are unique; and this contrasts with the mechanisms for neutrophil migration into skin and joint inflammation, where mAb to CD18 (WT.3) plus CD49d (TA-2) completely (in skin) or strongly (70 to 83% in joints) inhibit this process (31).

These findings show the uniqueness of neutrophil-lung interactions and raise the question of which additional mechanisms neutrophils might use to leave the pulmonary vessels. The small diameter of the pulmonary capillaries suggests that neutrophils come into close contact with the capillary endothelium, and in fact neutrophil emigration in the lungs seems to occur mainly in the capillary bed (13). Human neutrophils (53) also express a laminin receptor, VLA-6 (61), which would not have been blocked in these experiments. Laminin is the most abundant glycoprotein in basement membranes (54), and sequestered leukocytes in the pulmonary microvessels could adhere directly to exposed laminin at sites of increased vascular permeability. Under these conditions the platelet endothelial cell adhesion molecule-1, which is expressed at endothelial cell junctions and on granulocytes and platelets (55, 56) via homotypic binding or by serving as a counter-receptor for _V3 (57, 58), could also become involved in adhesion/migration events.

The increase in BALF protein content after challenge that was observed in this model (40) was reduced by the anti-CD49d mAb TA-2, and this inhibition was significantly greater when both CD49d and CD18 were blocked (Figure 4). In contrast, treatment to block CD18 integrins alone had little effect on the alveolar-capillary protein leak, suggesting that CD49d integrin-dependent mechanisms play an important role in this process. CD49d or 4 may not only mediate granulocyte migration, but also may be an important cofactor in eosinophil activation (29), survival (59), and even mast cell mediator release (46).

In conclusion, we have shown that both 2 (CD18) and 4 (CD49d) integrins mediate eosinophil migration to allergic lung inflammation. Both of these adhesion molecule mechanisms need to be blocked to inhibit eosinophil accumulation optimally in the lungs. However, despite comparable inhibition of neutrophil accumulation in the air space with CD49d and CD18 blockade, neutrophil migration into the lung parenchyma can proceed unabated, presumably via alternative mechanisms. The 4 integrin-dependent mechanisms in particular may mediate the alveolocapillary hyperpermeability following allergen exposure, perhaps in part due to their role as signaling molecules in leukocyte (eosinophil) and mast cell activation as well as in mediating granulocyte recruitment. On the basis of this study and that of others in allergic lung inflammation, the nature and extent of the immunoinflammatory response (sensitization, allergen dose, airway exposure, etc.) may in part determine which adhesion pathways are of particular importance. Further investigations are necessary to better understand the relationship between these variables, the mechanisms of leukocyte migration and activation in allergic lung inflammation, and how these relate to patients with asthma of differing severity.