Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Beta-2 integrins (lymphocyte function-associated antigen [LFA]-1 and Mac-1 [CD11b/CD18]) are important adhesion molecules involved in the migration of leukocytes, cell signaling, and subsequent secretory responses, and as such, are critical mediators of inflammation (1). A number of reports have examined the role of Mac-1 and LFA-1 in neutrophil adhesion and transmigration. Although surface levels of LFA-1 and Mac-1 are approximately equal on human neutrophils, Mac-1 can be mobilized from granules upon activation. It was initially suggested that Mac-1 was the dominant beta-2 integrin involved in neutrophil adhesion and migration. This was based on functional inhibition studies using monoclonal anti-Mac-1 antibodies both in vitro and in vivo (2, 3). However, subsequent investigation into the underlying mechanisms raised questions as to the relative importance of Mac-1 and LFA-1. In rat and rabbit models of inflammation, it appeared that LFA-1 played a more prominent role in neutrophil emigration than did Mac-1 (4). Further insight into the relative contributions of Mac-1 and LFA-1 in neutrophil migration came into light with the generation of Mac-1-deficient mice. Our group and others (7) have illustrated that mice deficient in Mac-1 have normal neutrophil migration into the inflamed peritoneum. Furthermore, neutrophil influx into the subcutaneous air pouch was significantly enhanced in the Mac-1-deficient mice compared with their wild-type counterparts (7). Taken together, these observations suggested that Mac-1 may play a less important role in extravasation of neutrophils and a more important role in extravascular adhesive events, including activation and degranulation leading to tissue dysfunction.

Mac-1 is also constitutively expressed on the surface of eosinophils and is known to mediate adhesion to either platelet-activating factor (PAF) or cytokine-stimulated endothelial cells in vitro (10). In addition, eosinophils isolated from allergic donors exhibit significant Mac-1-dependent adhesion to cytokine-activated endothelial cells (13). The role of beta-2 integrins in eosinophil emigration has also been studied in animal models in vivo. In a rat model of allergic inflammation, Schneider and coworkers (14) demonstrated that CD18 blockade significantly, but only partially, reduced airway and tissue eosinophil and neutrophil accumulation. Complete inhibition of eosinophil accumulation was only observed when both CD18 and very late antigen (VLA)-4 were blocked. The alpha subunit of CD18, however, was not determined in this study. In another study, blocking both LFA-1 and Mac-1 significantly reduced both the early and late airway inflammation associated with ovalbumin (OVA) challenge in Brown Norway rats (15). The individual effects of blocking either Mac-1 or LFA-1 alone were not examined. On the other hand, in another murine model, Bloehmen and colleagues (16) demonstrated that blocking LFA-1 only, but not Mac-1, inhibited tracheal hyperreactivity; yet blocking either LFA-1 or Mac-1 had no effect on eosinophil accumulation in the lung. These studies point to a controversial, yet potentially important, role for Mac-1 in allergic pulmonary inflammation.

The goal of this study was to directly and systematically examine the role of Mac-1 in eosinophil emigration and tissue dysfunction in vivo. We developed a murine model of allergic pulmonary inflammation, assessed eosinophil migration into the airways, accumulation in the lung parenchyma, and bronchial hyperreactivity in wild-type mice and Mac-1-deficient mice.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals

All animals used in this study were male mice weighing 20 to 35 g and were used between 6 and 10 wk of age. Mac-1-deficient mice were generated on a C57Bl/6 background by targeted gene deletion (8). We have demonstrated that Mac-1-deficient mice do not develop gross abnormalities in growth rate or tissue architecture of major organ systems. Flow cytometric analysis of peripheral blood stained with specific monoclonal antibodies revealed in homozygous mice that CD18 expression was 60% of wild-type levels, LFA-1 was not different from wild-type levels, and Mac-1 was undetectable (using monoclonal antibody M1/70). Mac-1-dependent functions of isolated neutrophils were absent.

Sensitization and Challenge Protocol

Allergic pulmonary inflammation was elicited by sensitizing and challenging animals with chicken egg OVA. Mice were systemically sensitized with an intraperitoneal injection of 0.5 ml 0.9% phosphate-buffered saline (PBS) containing 4 µg of OVA (Sigma Chemical Co., St. Louis, MO) adsorbed to 4 mg aluminum hydroxide (AlOH) on Days 0 and 14. Nonsensitized animals received an intraperitoneal injection of AlOH in saline on Days 0 and 14. Sensitized mice were lightly anesthetized with metophane (Pittman-Moore, Mundelein, IL) and intranasally challenged with 40 µl of saline containing 4 µg OVA. Control mice received intranasal saline. Mice were challenged either once on Day 14 or five times on Days 14 and 25 through 28. In some experiments, wild-type mice received a blocking, monoclonal anti-Mac-1 antibody (M1/70; 2 mg/kg; F[ab']₂ or an isotype-matched control antibody (SFR3DR5; 2 mg/kg; F[ab']₂). The antibodies were administered intranasally 30 min before each of five OVA challenges. For all experiments, parameters were measured 24 h after the last intranasal challenge, either on Day 15 or Day 29. This time point was chosen based on observations by Inman and associates (17), demonstrating significant eosinophil accumulation in perivascular, peribronchial, and parenchymal regions of the lung.

Interleukin-5 and Eotaxin-Induced Eosinophil Recruitment

To assess the role of Mac-1 in direct eosinophil migration in response to an eosinophil-specific chemokine, in another series of experiments, wild-type and Mac-1-deficient mice were treated with interleukin (IL)-5 (0.1, 0.5, or 1 µg; R&D Systems, Inc., Minneapolis, MN) intravenously to induce blood eosinophilia. At approximately 90 min after IL-5 injection (0.5 µg; optimal time and dose at which blood eosinophilia peaks), mice were challenged intranasally with eotaxin (5 µg/mouse R&D Systems Inc.). Bronchoalveolar lavage (BAL) was collected 4 h later, and lungs were harvested for assessment of eosinophil infiltration.

Pulmonary Function Test

Animals were anesthetized with 0.2 ml ketamine (200 mg/kg, intraperitoneal) and xylazine (10 mg/kg, intraperitoneal) in normal saline. A tracheostomy was performed, and respiratory resistance in response to varying doses of intravenous methacholine was measured in a whole body plethysmograph. Pulmonary resistance (RL) was measured for each animal, and an average dose-response curve was generated as previously described (18). The curves were analyzed by calculating the maximum increase in RL (MAX RL), which was defined as the difference between baseline (BASE RL) and the maximum value. The maximum response was determined by reaching a point where the two following measurements either remained the same or began to fall. The sensitivity was the dose of methacholine required to produce one-half MAX RL, and the reactivity was determined by dividing MAX RL by the sensitivity (18).

Blood and Plasma Collection

After pulmonary function testing, blood was collected by retro-orbital puncture. Approximately 1 ml of blood was collected from each animal. Total white blood cells were counted, and blood smears were made for leukocyte differentials. Plasma was then collected for assessment of anti-OVA immunoglobulin E antibodies in the passive cutaneous anaphylaxis reaction.

Passive Cutaneous Anaphylaxis Reaction

Plasma was obtained from all OVA- and saline-sensitized mice. Serial dilutions (0.13 to 0.02) of the serum samples were prepared, and 200 µl of each sample was injected intradermally into the shaved backs of control, untreated Sprague-Dawley rats. We have previously demonstrated that plasma from sensitized mice elicited the same reaction in Sprague-Dawley rats as it did in untreated C57Bl/6 mice (19). Therefore, rats were used for all passive cutaneous anaphylaxis (PCA) reactions, primarily because PCA was easier to detect and record. After 72 h, rats were challenged with an intracardiac injection of a solution containing 2.5 mg Evan's blue dye and 5 mg OVA in a total volume of 1.5 ml saline. The final reaction was read 60 min later as the highest dilution that produced a distinct blue region (Evan's blue extravasation) at the center of the injection site.

Bronchoalveolar Lavage

After blood was drawn, BAL fluid was collected by lavaging both lungs. Briefly, 0.5 ml saline was injected and aspirated through the tracheostomy tube for a total of three times. The total volume of BAL fluid collected from all mice ranged between 1 and 1.5 ml. Total BAL fluid cells were counted, and cytospin smears were made and stained with hematoxylin and eosin (H&E) to count differentials.

Lung Histology

After BAL fluid was collected, the lungs were inflated with 1 ml of fixative (PBS containing 4% paraformaldehyde), and the trachea was tied off. The lungs were then removed from the chest cavity and placed in fixative overnight. Tissues were embedded in paraffin and cut into 5-µm sections, and stained with H&E. Inflammatory infiltrates and lung architecture were assessed using light microscopy.

Electron Microscopy

For ultrastructural studies, lungs were fixed by inflation with 0.1 M PBS containing 2.5% glutaraldehyde (2 h, room temperature) and postfixed in 0.1 M PBS containing 1% osmium tetroxide (1 h, room temperature). This was followed by en bloc uranyl acetate staining, dehydration through ethanol, and embedding in LX112 resin (Polysciences Inc., Warrington, PA). Ultrathin sections (80 nm), stained with uranyl acetate and lead citrate, were viewed and photographed on a JEM 200CX electron microscope (JEOL USA, Inc., Peabody, MA).

Statistical Analysis

All data are presented as mean ± standard error of the mean (SEM). A one-way analysis of variance with Newman-Keuls multiple comparison test was used for comparisons. Statistical significance was set at P < 0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

All sensitized animals used in this study had serum anti-OVA antibody titers of at least 0.02 as assessed by a PCA reaction (data not shown). There was no evidence of anti-OVA antibodies in saline-sensitized animals.

Total number of circulating leukocytes was determined in allergic (OVA/OVA) wild-type (7.9 ± 1.7 × 10⁶ cells/ ml) and Mac-1-deficient mice (6.2 ± 0.8 × 10⁶ cells/ml), and allergic wild-type mice treated with an anti-Mac-1 antibody (5.9 ± 1.0 × 10⁶ cells/ml). There was no difference between the three groups. Also, baseline circulating leukocyte counts were not different between untreated wild-type and Mac-1-deficient mice. OVA sensitization and challenge induced a similar (two to threefold) increase in circulating eosinophils in both wild-type and Mac-1-deficient mice (data not shown).

The change in respiratory resistance in nonallergic (OVA/SAL) and allergic (OVA/OVA) wild-type and Mac-1-deficient mice is shown in Figure 1. (Note: the OVA/SAL group includes data from untreated animals as there was no difference between the two groups.) In this series of experiments, all sensitized animals received five intranasal challenges (Days 14 and 25 through 28), and respiratory resistance was measured 24 h after the last challenge (Day 29). The analysis of the data is shown in Table 1. There was no difference in BASE RL between the four groups of animals. Bronchial reactivity in nonallergic Mac-1-deficient mice was significant, greater than that of nonallergic wild-type mice. After OVA sensitization and five OVA challenges, bronchial reactivity increased in wild-type and Mac-1-deficient mice. It is noteworthy that the reactivity of Mac-1-deficient mice (OVA/OVA) was even further enhanced compared with their wild-type counterparts (Table 1). These observations demonstrate that Mac-1-deficient mice are significantly more responsive to OVA challenge than are allergic wild-type mice.

View larger version (16K): [in this window] [in a new window]   Figure 1.   Change in respiratory resistance over increasing doses of methacholine in wild-type nonallergic (OVA/SAL, n = 11; open circles) and allergic (OVA/OVA, n = 8; open squares), and Mac-1-deficient nonallergic (n = 6; solid circles) and allergic (n = 15; solid squares) mice after five OVA challenges.

In the next series of experiments, we asked whether the differences in bronchial hyperreactivity between allergic wild-type and Mac-1-deficient mice existed at an earlier time point, i.e., after only one antigen challenge. Both wild-type and Mac-1-deficient mice were sensitized as previously described and challenged only once on Day 14. Bronchial reactivity was then measured on Day 15. The change in respiratory resistance in response to increasing doses of methacholine is illustrated in Figure 2. Control (OVA/SAL) wild-type and Mac-1-deficient mice exhibited a similar baseline dose-dependent increase in respiratory resistance. After one OVA challenge, there was no apparent increase in respiratory resistance in either wild-type mice or Mac-1-deficient mice. These results are in contrast to those in Figure 1 where five OVA challenges significantly increased respiratory resistance in both groups of animals. These data suggest that one intranasal OVA challenge is not enough to increase respiratory resistance in either wild-type or Mac-1-deficient mice.

View larger version (13K): [in this window] [in a new window]   Figure 2.   Change in respiratory resistance over increasing doses of methacholine in wild-type nonallergic (OVA/SAL, n = 3; open circles) and allergic (OVA/OVA, n = 4; open squares), and Mac-1-deficient nonallergic (n = 6; solid circles) and allergic (n = 5; solid squares) mice after one OVA challenge.

BAL fluid was collected from all animals after assessment of bronchial hyperreactivity. The total number of leukocytes in the BAL fluid of control (OVA/SAL) and allergic (OVA/OVA) wild-type and Mac-1-deficient mice, after either one or five intranasal OVA challenges, is shown in Table 2. In control wild-type mice, there was no difference in the total BAL leukocytes whether the mice were challenged with saline once or five times. These values are also not different from nonsensitized, untreated control mice (data not shown). Total BAL leukocyte counts in control (OVA/SAL) Mac-1-deficient mice, after one or five challenges, are not different from each other or their respective wild-type counterparts. After one OVA challenge in wild-type mice, there is no change in the total leukocyte counts in the BAL fluid. After five OVA challenges, however, there is an apparent increase in the total leukocyte counts. On the other hand, in Mac-1-deficient mice, there is a significant increase in total BAL leukocyte counts after only one OVA challenge, and this is further enhanced after five OVA challenges. These data demonstrate that whereas wild-type mice respond to five OVA challenges, Mac-1-deficient mice respond to almost the same degree after a single challenge.

We also examined airway and tissue eosinophilia in control and allergic wild-type and Mac-1-deficient mice. There was no airway (Table 3) or tissue (Figures 3A and 3C) eosinophilia in control (OVA/SAL) wild-type or Mac-1-deficient mice. In wild-type mice, total airway (Table 3) and tissue eosinophilia did not change after one OVA challenge but they did increase significantly after five challenges (Table 3 and Figure 3B). In Mac-1-deficient mice, however, airway (Table 3) and tissue (Figure 4A) eosinophilia appeared after one OVA challenge. Airway eosinophilia observed in Mac-1-deficient mice after one challenge was comparable to that observed in wild-type mice after five challenges (Table 3). In addition, electron microscopic observations confirmed the presence of large numbers of eosinophils within the airway tissues after one challenge. Eosinophils were frequently observed emigrating from large vessels located in the bronchovascular bundles (BVB; Figure 4B). In Mac-1-deficient mice, after five challenges, the total number of BAL (Table 3) and tissue (Figure 3D) eosinophils was further enhanced and significantly greater than that observed in the respective wild-type mice. Total airway polymorphonuclear leukocytes (PMNs) and macrophages were also increased after five challenges in wild-type (PMNs: OVA/SAL = 3.5 ± 1 × 10⁴/ml, OVA/OVA = 12 ± 4 × 10⁴/ml; macrophages: OVA/SAL = 32 ± 1.4 × 10⁴/ml, OVA/OVA = 44 ± 2 × 10⁴/ml) and Mac-1-deficient mice (PMNs: OVA/SAL = 4.4 ± 1 × 10⁴/ml, OVA/OVA = 9.8 ± 2 × 10⁴/ml; macrophages: OVA/SAL = 29 ± 2.3 × 10⁴/ml, OVA/OVA = 61 ± 3 × 10⁴/ml ). In the tissue, large numbers of mononuclear cells were present and eosinophils constituted 95% of all granulocytes. These data clearly demonstrated that allergic Mac-1-deficient mice are more inflamed than are their wild-type counterparts.

View larger version (109K): [in this window] [in a new window]   Figure 3.   Histologic assessment of tissue eosinophilia in control (OVA/SAL) wild-type (A) and Mac-1-deficient (C) mice, and allergic (OVA/OVA, five challenges) wild-type (B) and Mac-1-deficient mice (D) mice. Note the accumulation of leukocytes around the BVB in allergic animals. Br = bronchioles; arrowheads = bronchial arteriole.

View larger version (85K): [in this window] [in a new window]   Figure 4.   Histologic assessment of tissue eosinophilia (A) and transmission electron microscopy (B) in allergic (OVA/OVA) Mac-1-deficient mice after one challenge. Note the eosinophil (arrowhead) migrating beneath the endothelium (arrows). Br = bronchioles.

In the next series of experiments, we asked whether blocking Mac-1 by a monoclonal antibody would yield similar results to those observed with the Mac-1-deficient mice. Allergic wild-type mice received either a blocking monoclonal anti-Mac-1 F(ab')₂ fragment or an isotype-matched control F(ab')₂ fragment before each of five OVA challenges. Bronchial hyperreactivity, BAL cellularity, and tissue histology were assessed 24 h after the last challenge. Figure 5 shows that neither a control nor an anti-Mac-1 antibody affected bronchial hyperreactivity in allergic wild-type mice; the data are not different from those observed in allergic wild-type mice that did not receive antibody (Figure 1). These data are in contrast to those in the Mac-1-deficient mice, which exhibited significantly greater bronchial hyperreactivity after five OVA challenges. These observations illustrate a clear difference between the use of a blocking antibody and Mac-1-deficient mice.

View larger version (14K): [in this window] [in a new window]   Figure 5.   Change in respiratory resistance over increasing doses of methacholine in allergic (OVA/OVA) wild-type mice treated with either a control (n = 7; open circles) or a blocking monoclonal anti-Mac-1 antibody (n = 12; solid circles).

The total number of leukocytes and eosinophils in the BAL fluid of allergic mice treated with either control or anti-Mac-1 antibody is shown in Table 4. It is noteworthy that the total number of BAL leukocytes in this experiment was greater than that observed in the wild-type mice after five challenges (Table 2). The reason for this difference is unknown. It should be noted, however, that the untreated allergic wild-type and Mac-1-deficient mice were studied with a different batch of OVA and before the antibody-treated animals. Hence, direct comparisons between antibody-treated and untreated mice are not possible. More importantly, as shown in Table 4, there was no difference between the total number of leukocytes in control antibody and anti-Mac-1 antibody-treated mice, suggesting that blocking Mac-1 at the time of challenge is not sufficient to inhibit the inflammatory response. Table 4 also shows the total number of BAL eosinophils. In allergic wild-type mice treated with a control antibody, there were significant numbers of eosinophils in the BAL, greater than those of control (OVA/SAL) wild type mice (Table 3). Administration of a blocking anti-Mac-1 antibody significantly reduced the total number of BAL eosinophils (Table 4) but had no effect on tissue eosinophilia (Figure 6). These data are in contrast to those in Table 4 where deletion of Mac-1 actually enhanced the total number of BAL eosinophils. Again, these observations suggest differences between antibody administration and gene deletion.

View larger version (92K): [in this window] [in a new window]   Figure 6.   Histologic assessment of tissue eosinophilia in allergic (OVA/OVA) wild-type mice treated with a control antibody (A) or a monoclonal anti-Mac-1 antibody (B).

In a final series of experiments, we asked whether direct eosinophil migration in response to a chemotactic agent would be altered in the Mac-1-deficient mice. Wild-type and Mac-1-deficient mice were treated intravenously with IL-5 to induce a blood eosinophilia and then challenged intranasally with eotaxin. We show in Table 5 that 4 h after eotaxin challenge, there was a significant increase in the total number of eosinophils in the BAL of both wild-type and Mac-1-deficient mice; however, there was no difference between the two groups. These data demonstrate that in response to an acute chemotactic factor, eosinophil migration is not altered in the Mac-1-deficient mice.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study, we directly and systematically examined the role of Mac-1 in a murine model of allergic pulmonary inflammation. We report for the first time that Mac-1-deficient mice develop a significantly enhanced allergic response compared with their wild-type counterparts. After five intranasal OVA challenges, the total number of eosinophils in the BAL fluid is significantly greater than that observed in allergic wild-type animals. Histologic analysis of the lung reveals large numbers of eosinophils around the BVBs, and pulmonary function tests reveal significantly enhanced bronchial hyperreactivity. Moreover, we observe that as early as 24 h after the first OVA challenge, Mac-1-deficient mice develop significant airway and tissue eosinophilia, whereas their wild-type counterparts do not. However, despite tissue and airway eosinophil infiltration, these mice do not exhibit airway hyperreactivity, as do the Mac-1-deficient mice after five challenges, suggesting that in this model, there is no direct correlation between airway eosinophilia and airway reactivity. We also examined the effect of an anti-Mac-1 antibody and observed that although airway eosinophilia is reduced, neither tissue eosinophilia nor bronchial hyperreactivity is affected. Finally, we report that eosinophil migration in response to acute eotaxin administration is not different between Mac-1-deficient and wild-type mice. These data indicate that during an adaptive, allergic immune response, but not acute inflammation, Mac-1-deficient mice are more inflamed than are their wild-type counterparts.

Although others have investigated a role for beta-2 integrins, this study is the first to directly examine the role of Mac-1 in allergic pulmonary inflammation in vivo. In a rat model of allergic inflammation, Schneider and coworkers (14) demonstrated that CD18 blockade significantly, but only partially, reduces airway and tissue eosinophil and neutrophil accumulation. Complete inhibition is only observed when VLA-4 is blocked in addition to CD18. An assessment of pulmonary inflammation and tissue dysfunction has not been performed in CD18-deficient mice. Blocking LFA-1 and Mac-1 significantly reduces both early and late airway inflammation associated with OVA challenge in rats and nonallergic inflammation in mice (15). The individual effects of blocking either LFA-1 or Mac-1 were not examined in these studies. Chin and colleagues (20) demonstrated that in mice, an anti-intercellular adhesion molecule (ICAM)-1 antibody significantly reduces the airway inflammation associated with OVA challenge. Gerwin and associates (21) demonstrated that mice genetically deficient in ICAM-2 develop prolonged airway hyperreactivity and eosinophilia in response to OVA challenge. Although the studies by Chin and coworkers (20) and Gerwin and colleagues (21) did not directly address the role for Mac-1, ICAM-1 and ICAM-2 are ligands for Mac-1 (22, 23), and the results may be extrapolated to suggest that ICAM-1 and ICAM-2 interactions with Mac-1 may promote allergic inflammation in the lung.

The mechanism underlying enhanced inflammation in Mac-1-deficient mice is not clear. Mac-1 deficiency does not influence circulating numbers of leukocytes under control, untreated conditions. An early increase in circulating eosinophils after antigen challenge may account for the enhanced eosinophil accumulation in allergic Mac-1-deficient mice. This contention is based on the observation that Mac-1-deficient, but not wild-type, mice develop blood, airway, and tissue eosinophilia as soon as 24 h after one challenge. In fact, the eosinophil burden in the Mac-1- deficient mice after one challenge is comparable to that observed in wild-type mice after five challenges. After five challenges, however, there is no difference between allergic wild-type and Mac-1-deficient mice. It is possible that Mac-1-deficient mice are simply more sensitive to intranasal OVA challenge and, therefore, are able to mobilize eosinophils more efficiently than are their wild-type counterparts. However, other possible explanations exist.

First, it is possible that in the absence of Mac-1, other adhesion molecules (LFA-1, alpha d, VLA-4) are upregulated and thereby account for the enhanced migration of eosinophils in the mutant mice. Although a systematic assessment of adhesion molecule expression on eosinophils from Mac-1-deficient mice has not been reported, it is known that levels of LFA-1 and L-selectin are not altered on neutrophils (peripheral blood and peritoneal) and macrophages from Mac-1-deficient mice (8, 9). Second, it has been suggested that Mac-1 may be an important endogenous regulator of apoptosis. Coxon and associates (9) demonstrated that in response to thioglycollate challenge, significantly greater numbers of neutrophils accumulate in the peritoneum. The investigators suggested that this is due to delayed apoptosis in the elicited neutrophils. If Mac-1 is important in eosinophil apoptosis, it is possible that the lifespan of extravasated eosinophils in Mac-1-deficient mice is prolonged, and therefore, there is greater accumulation in the lung interstitium and airspace. Third, it is not known whether there are any local microenvironmental changes in the lungs of Mac-1-deficient mice. Although lungs were not directly examined in their study, Rosenkranz and coworkers (24) reported that there is a significant reduction in the total number of resident mast cells in the peritoneum and skin of Mac-1-deficient mice. Mast cells are resident immunocytes, capable of storing and generating an array of proinflammatory molecules, including tumor necrosis factor-, PAF, oxidants, histamine, IL-10, and transforming growth factor-, thereby contributing significantly to the regulation of inflammation and tissue repair. However, compared with other species, very few mast cells are present in the murine lung (25), and we found no difference in the number or distribution of lung tissue mast cells between Mac-1-deficient and wild-type mice (data not shown).

Another important observation in this study is that acute eotaxin challenge induced similar eosinophil migration in both wild-type and Mac-1-deficient mice. These data suggest that direct eosinophil migration from the intravascular compartment into the lung tissue is not affected by Mac-1 deficiency and further support the contention that Mac-1 may play an important regulatory role during the adaptive immune response. For example, during the adaptive immune response, the local generation and/or balance of T helper (Th) 1 (interferon-gamma) versus Th₂ cytokines (IL-4 and IL-5) may be altered in the Mac-1-deficient mice, thereby enhancing eosinophil migration and accumulation in the lung. In this model, OVA challenge is known to elicit a Th₂ response, with IL-4 and IL-5 as the predominant cytokines driving eosinophil migration and activation (26). Although we have not directly measured Th₂ cytokine levels in wild-type and Mac-1-deficient allergic mice, it is conceivable that during the adaptive immune response, a shift in the Th₁ versus Th₂ profiles renders the Mac-1-deficient mice more susceptible to allergic inflammation. Further investigation into the assessment of cytokines and T-cell populations in normal and allergic Mac-1-deficient mice is warranted.

Based on our observations in the Mac-1-deficient mice, we anticipated that administration of an anti-Mac-1 antibody would enhance the OVA-induced allergic inflammation. However, the results were very different. Before each of five intranasal OVA challenges, wild-type mice received either a monoclonal anti-Mac-1 or an isotype-matched control antibody. We observed that an anti-Mac-1 antibody significantly reduces the airway eosinophilia, as assessed in the BAL fluid. These results are in direct contrast to those obtained in the Mac-1-deficient mice where airway eosinophilia is significantly enhanced after five OVA challenges. Although a clear explanation for these differences is unknown, it is conceivable that the antibody itself may have induced a secondary affect that directly influenced eosinophil migration into the airways. Another possibility is that the mutant mice have compensated for Mac-1-deficiency by upregulating other adhesive interactions, including VLA-4/vascular cell adhesion molecule-1 and/or LFA-1/ICAM-1. Finally, if in fact Mac-1 is involved in the adaptive immune response (sensitization phase), immunoneutralization of Mac-1 (antibody treatment) during the migration (challenge) phase would not be expected to yield similar results to those of the Mac-1-deficient mice. Although the anti-Mac-1 antibody reduces airway lumen eosinophilia, it has no effect on either tissue accumulation of eosinophils or airway hyperreactivity. It should be noted that the eosinophil burden in the airways (BAL fluid) of mice treated with the anti-Mac-1 antibody (31 ± 5 × 10⁴/ml) is similar to that in Mac-1-deficient mice after one OVA challenge (33 ± 7.7 × 10⁴/ml), yet the Mac-1- deficient mice do not exhibit enhanced airway reactivity. Again, these observations suggest that in this model, there may be no direct correlation between airway eosinophilia and airway reactivity.

Controversy as to whether airway eosinophilia is directly linked to airway reactivity in animal models of allergic pulmonary inflammation has been noted (15, 17, 27 29). It is becoming clear that there are important species and strain-specific differences in the extent and type of airway inflammation and subsequent association with airway reactivity. Other confounding factors include type of allergic antigen, route of administration, and duration of antigen delivery. All of these factors may influence the site of eosinophil emigration and accumulation, and the activation status of eosinophils, and thereby modulate airway reactivity. The airway reactivity in the Mac-1-deficient mice is significantly greater than that in wild-type mice after five but not one OVA challenge. It is possible that the eosinophils accumulating in the lung become increasingly activated with five OVA challenges and, therefore, contribute to increased airway hyperreactivity. This phenomenon may be exacerbated in the Mac-1-deficient mice. There is evidence to suggest that eosinophil accumulation in close proximity to the bronchial mucosa is critical to the development of airway reactivity (17). In the present study, we demonstrated that in addition to capillaries, eosinophils migrate directly out of large vessels and primarily accumulate around the BVB; eosinophil migration across the bronchial epithelium is not observed. In addition, our electron microscopic observations clearly show that eosinophil emigration induces a pronounced vasculitis, as evidenced by detachment of the endothelium from the underlying basal lamina. Although we have no direct evidence for this, preferential accumulation of eosinophils around the BVB may be related to increased antigen deposition at the bronchoalveolar junction (30, 31).

In conclusion, by using Mac-1-deficient mice, we have uncovered a potentially significant antiinflammatory role for Mac-1 in a murine model of allergic pulmonary inflammation. Further investigation into and delineation of the underlying mechanism of this regulatory role will undoubtedly enhance our understanding of the way in which Mac-1 modulates allergic inflammation.