Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Eosinophils are potent inflammatory cells recruited to inflamed tissue in various pathologic conditions such as allergy and asthma (1, 2). This recruitment involves the steps of eosinopoiesis, chemotaxis, cell activation and adhesion-molecule expression, rolling and firm adherence to endothelium, and finally migration through the vascular wall into tissue (3). Mediators and cytokines, which promote eosinophil chemotaxis and chemokinesis, probably play an important role in the first steps of eosinophil recruitment, but their actions in eosinophil transmigration must be further defined. Transmigration of eosinophils through the vascular wall implies many cellular functions, such as modification of the cell skeleton and digestion of extracellular matrix. This last phenomenon has been evaluated in lymphocyte and neutrophil transmigration, and it has been reported that these cells use proteinases, the matrix metalloproteinases (MMP)-9 and -2, to accomplish this function (4).

Recently, Okada and colleagues (7) investigated eosinophil transmigration through Matrigel (Becton Dickinson, Bedford, MA), a gel containing basement membrane components used as an in vitro model of basement membrane (8). They tested several mediators and cytokines known to be active eosinophil chemotactic factors. They first showed that platelet-activating factor (PAF) and eotaxin, in the presence of interleukin (IL)-5, IL-3, or granulocyte-macrophage-colony stimulating factor (GM-CSF), promoted eosinophil transmigration through basement membrane components in vitro. At optimal concentration, PAF was more potent than eotaxin, although the transmigration it induced remained rather weak, at  8% of the total number of eosinophils. Both PAF as chemotactic factor in the lower chamber, and an eosinophil-active cytokine as the activating factor in both chambers, were required for eosinophil migration. In the study by Okada and colleagues (7), leukotriene B4 (LTB₄), C5a, the chemokines regulated on activation, normal T-cell expressed and secreted (RANTES) and macrophage inflammatory protein (MIP)- 1, and IL-8 did not induce measurable eosinophil transmigration. Moreover, other cytokines, such as tumor necrosis factor (TNF)-, RANTES, IL-8, interferon (IFN)-, and IL-4 did not promote eosinophil transmigration in the presence of PAF. These data suggest that basement membrane transmigration requires specific eosinophil activation, which is optimized by the effects of a combination of specific mediators and cytokines.

Recently, Powell and coworkers (9) identified a new class of arachidonic acid metabolite, 5-oxo-6,8,11,14-eicosatetraenoic acid (5-oxo-ETE). 5-Oxo-ETE is formed from 5-hydroxy-6,8,11,14-eicosatetraenoic acid (5-HETE) by a highly specific dehydrogenase that is found in lymphocytes, monocytes, neutrophils, and eosinophils (9). 5-Oxo-ETE is a chemoattractant for neutrophils that is 100-fold more potent than 5-HETE (13). It activates neutrophils in many functions, but appears less potent than LTB₄ (13, 14). In contrast, 5-oxo-ETE is a potent and specific stimulator of human eosinophil migration, and is even more effective in this respect than other lipid mediators such as PAF, 5-oxo-15-hydroxy-ETE, and LTB₄ and LTC₄ (12). In vitro, 5-oxo-ETE activates many eosinophil functions and appears more potent in stimulating eosinophils than neutrophils (15, 16). In vivo, 5-oxo-ETE induces pulmonary eosinophilia in Brown Norway rats, further supporting the possibility that it may act as a potent physiologic mediator in the eosinophilic inflammatory process (17). Structure-activity studies suggest that 5-oxo-ETE acts through a specific receptor (13, 18, 19). However, this receptor has not been yet characterized.

Eosinophils extracted from carcinoma (20, 21) and blood (22) express MMP-9. Moreover, there is some evidence that MMP-9 levels are increased in the bronchial tissue of asthmatic subjects, and that eosinophils are implicated in this phenomenon (23, 24). In another study, Okada and colleagues (25) further showed that the combination of PAF and IL-5 appeared to induce MMP-9 activation in the supernatants of eosinophils migrating through a Matrigel membrane, indicating that this protease was involved in eosinophil transmigration through the basement membrane. However, their data also showed that other proteases could be involved in eosinophil migration through basement membrane components. An inhibitor of serine proteases decreased the eosinophil migration, suggesting that these proteases were also implicated in eosinophil migration.

Urokinase plasminogen activator (uPA) and its receptor, uPA-R (CD87), play an important role in a number of physiologic and pathologic extracellular degradative processes and remodeling (26, 27). This system promotes matrix degradation by generating plasmin, a serine protease, from the abundant zymogen plasminogen (26). Moreover, plasmin can act as a promoter of local stroma degradation by converting inactive zymogen pro-MMP into active MMP (26, 28). It was shown that uPA-R rapidly redistributes at the leading edge of migrating monocytes (29). Thus uPA-R plays a pivotal role by modulating and concentrating uPA activity at the required sites on the cell surface. Stahl and associates showed that uPA-R played an important role in melanoma cell invasion through artificial basement membrane, and that this could be strongly inhibited by a monoclonal anti-CD87 antibody (30). This receptor has also been observed on eosinophils (31, 32). Overall, these data suggest that the uPA-uPA-R system could be involved in eosinophil transmigration across the basement membrane.

The blood eosinophil count increases with increasing asthma severity (33). The eosinophils in asthma display features of activation: they are hypodense and present an increased capacity to release mediators (34, 35). Consequently, it is possible that blood eosinophils of asthmatic subjects better respond to chemotactic factors and migrate more efficiently through the vascular basement membrane barrier than do those of normal individuals. This would at least partly explain the greater number of eosinophils observed in asthmatic bronchial mucosa. In the present study we postulated that the effect of 5-oxo-ETE on eosinophil transmigration across basement membrane would be greater than that observed with PAF, since the chemotactic activity of 5-oxo-ETE on eosinophils was much greater than that of PAF. Consequently, we evaluated the effect of 5-oxo- ETE on eosinophil migration in comparison with that of PAF, investigated the modulating effect of IL-5 and the role of MMP and uPA-R in 5-oxo-ETE-induced eosinophil migration, and compared the response of normal and asthmatic eosinophils to these substances.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents

Human recombinant IL-5 was purchased from Peprotech, Inc. (Rocky Hill, NJ), 5-oxo-ETE from Cayman Chemical (Ann Arbor, MI), monoclonal mouse IgG₁ antibody to MMP-9 (Ab-1; clone 6-6B) from Calbiochem (San Diego, CA), monoclonal mouse IgG_2a antibody to CD87 (uPA-R) (clone 3936) from American Diagnostica Inc. (Greenwich, CT), mouse IgG_2a (isotypic control) from Pharmingen (San Diego, CA), fluorescein isothiocyanate (FITC)-conjugated goat antimouse IgG from Caltag (Burlingame, CA), and monoclonal antibody against CD16, conjugated to magnetic beads, from Miltenyi Biotec (Bergisch-Gladbach, Germany). Bovine serum albumin (BSA; fraction V), ethylenediaminetetraacetic acid (EDTA) and PAF (L--phosphatidylcholine) were purchased from Sigma Chemical Co. (St. Louis, MO); Dextran T-500 and Ficoll-Paque from Pharmacia LKB (Uppsala, Sweden); and RPMI 1640 medium, Hanks' balanced salt solution (HBSS) without calcium/magnesium, penicillin/streptomycin, and heat-inactivated fetal bovine serum (FBS) from Gibco BRL (Grand Island, NY). -Aminocaproic acid, a plasmin inhibitor, was purchased from Sigma. BB-3103, an inhibitor of MMP, was generously provided by British Biotechnology (Oxford, UK). The MMP-9 human enzyme immunoassay system was obtained from Amersham Pharmacia Biotech (Piscataway, NJ). According to information provided by Calbiochem, the monoclonal mouse IgG₁ against MMP-9 (Ab-1), clone 6-6B, is specific for MMP-9 and has not been shown to cross-react with other proteinases. The IC₅₀ values of BB-3103 for different MMPs are: collagenase: 2.0 nM; gelatinase A: 10.0 nM; gelatinase B: 7.0 nM; stromelysin: 30.0 nM, and matrilysin: 20.0 nM.

Evaluation of Subjects

Normal subjects, without a history of allergy or asthma, and asthmatic subjects meeting the criteria of the American Thoracic Society for the diagnosis of asthma (36) were recruited for the study. The asthmatic subjects had mild asthma defined by a morning, prebronchodilator FEV₁ > 85% of the predicted value, and requirement for a short-acting ₂-agonist on demand, or had moderate asthma, defined as a morning baseline FEV₁ < 85% or the use of a short-acting inhaled ₂-agonist more than three times a week. The inclusion criteria were stable treatment for more than 3 mo, no use of inhaled steroids over the 3 mo preceeding the study, no use of other drugs, and no other disease than asthma. Approval for the study was obtained from the local ethics committee. All subjects underwent blood sampling early in the morning. FEV₁ values were measured in the morning at least 8 h after any ₂-agonist inhalation, with a PFT II Vitalograph Spirometer (Vitalograph Ltd., Buckingham, UK), and were expressed as percentages of predicted values (37). Blood eosinophil counts were measured on a Coulter STKS (Model 809; Coulter Electronics, Hialeah, FL).

Blood Cell Processing and Eosinophil Purification

Venous blood (150 ml) was centrifuged to remove platelet-rich plasma, and the cell pellet was sedimentated for 30 min on Dextran 6%. Leukocytes were centrifuged on Ficoll-Paque for 20 min at 700 × g. The cell pellet containing the granulocytes was resuspended and red cells were lysed with distilled water. Eosinophils were purified according to the technique described by Hansel and coworkers (38) and modified in our laboratory (39). Briefly, granulocytes were resuspended in cold HBSS + 1% BSA (10⁶ cells/µl) and incubated with bead-conjugated monoclonal anti-CD16 (FcR III) antibody for 20 min at 4°C (100 µl/2 × 10⁸ cells). Cells were washed once in HBSS + 1% BSA (700 × g). The granulocyte suspension was depleted of neutrophils by negative selection, using a magnetic cell sorter (Miltenyi Biotec GmbH, Bergisch-Gladbach, Germany). An aliquot of the resulting cell suspensions was used to determine the total cell count (hemacytometer) and cell viability (trypan blue exclusion, Sigma), and to prepare smears for differential cell counts (Diff-Quik; Dade Diagnostics, Aguada, PR). The purity of the eosinophil preparations used in this study was greater than 98%, and the contaminating cells were mostly neutrophils. Eosinophils were resuspended in RPMI 1640 medium containing 25 mM 4-(2-hydroxyethyl)-1-piperazine-N'-2-ethanesulfonic acid (Hepes), 1% penicillin/streptomycin, and 10% FBS.

Experiment Design and Transmigration Assay

The migration of eosinophils through basement membrane components was evaluated with 24-well Biocoat Matrigel Invasion Chambers (Becton Dickinson). To optimally induce migration of eosinophils, we incubated eosinophils with or without IL-5 (10 ng/ml) for 30 min at 37°C in a 5% CO₂ atmosphere before the transmigration assay. Plates were incubated at 37°C under 5% CO₂ for 18 h except for kinetic studies, for which cells were incubated for 1 h to 18 h. To evaluate the number of cells that had migrated into the lower chambers, cells in both chambers were removed by aspiration, and remaining cells on both sides of the pore membranes were gently washed and harvested twice with cold RPMI/5 mM EDTA. Cells from each chamber were centrifuged (700 × g), resuspended in cold RPMI 1640, and counted with a hemacytometer. For each experimental condition, the percentage of transmigration was calculated as the number of cells in the lower chamber of the Matrigel Invasion Chamber divided by the number of cells in the lower chamber of a control invasion chamber without the Matrigel membrane.

The recovery of cells from the washing procedure was evaluated by calculating the number of cells in both chambers after transmigration, and comparing this with the number of cells before the transmigration assay. The recovery was always better than 95%. The control invasion chambers had the same pore size (8 µm) and density, and were made of the same material as the Biocoat Matrigel Invasion Chambers. When incubated in the presence of IL-5 (> 95%), most eosinophils migrated through the uncoated control invasion chambers.

To study the role of proteinases in transmigration, eosinophil suspensions (10⁶ cells/ml in complete RPMI 1640) were preincubated with or without monoclonal antibody (mAb) against MMP-9 (2 µg/ml) or CD87 (10 µg/ml) for 30 min at 37°C in 5% CO₂. Cells (500 µl) were added to the upper chamber and, in specific wells, BB-3103 at different concentrations, or -aminocaproic acid (500 µg/ml), was added directly to cell suspensions after IL-5 incubation. PAF (1 µM) or 5-oxo-ETE (1 µM) was used as a chemotactic factor and was added in the lower chambers. In some experiments, the PAF and 5-oxo-ETE diluents (dimethyl sulfoxide and ethanol, respectively) were used as controls and did not show any effects on eosinophil transmigration.

Flow-Cytometric Analysis of Eosinophil CD87 Expression

Eosinophils were incubated at 10⁶ cells/ml in complete RPMI 1640, and were stimulated for 30 min with or without recombinant IL-5 (10 ng/ml) at 37°C in 5% CO₂. Eosinophils were then washed once with cold 1× phosphate-buffered saline (PBS) supplemented with 0.1% BSA, and were labeled with monoclonal antibodies. Briefly, 250,000 cells/ 100 µl were incubated 45 min at 4°C with 1 µg/ml of mAb specific for CD87 or mouse IgG_2a mAb (isotype control). Cells were washed once with PBS + 0.1% BSA, resuspended in 100 µl of PBS + 0.1% BSA, and incubated for 45 min at 4°C with 1 µg/ml of FITC-conjugated monoclonal goat antimouse IgG antibody. Expression of CD87 on eosinophils (mean fluorescence) was analyzed with flow cytometry (EPICS ELITE ESP; Coulter Electronics).

Measurement of MMP-9 in Supernatants of Eosinophils

MMP-9 was measured in supernatants of eosinophils incubated on Matrigel membranes in the presence of RPMI 1640 alone, with FBS 10%, with FBS 10% and IL-5, or with FBS 10%, IL-5, and 5-oxo-ETE (1 µM). Some Matrigel chambers were incubated in absence of cells with RPMI plus FBS 10%, IL-5 and 5-oxo-ETE. The MMP-9 enzyme immunoassay sensitivity was 0.6 ng/ml.

Statistical Analysis

Means and SEM were determined for continuous variables. Mean values of quantitative variables were compared through a two-way analysis of variance (ANOVA) (randomized block design) for unpaired data, since normality and variance assumptions were met. Tests were adjusted when designs were unbalanced. A posteriori comparisons were performed with Tukey's method. Comparisons between 5-oxo-ETE and 5-oxo-ETE + IL-5 were made with Student's paired t test. All reported values of P are two-sided. The results were considered significant when values of P were < 0.05. The data were analyzed with the SAS statistical package program (SAS Institute, Inc., Cary, NC).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Characteristics of Subjects

Fifty-one subjects, consisting of 22 normal subjects (11 females and 11 males, age 32.2 ± 2.3 yr [mean ± SEM]) and 29 asthmatic subjects (15 females and 14 males, age 26.1 ± 1.9 yr) were recruited into the study. All were nonsmokers. The mean FEV₁ values of normal and asthmatic subjects were 105.1 ± 2.9% and 86.6 ± 3.7% predicted, respectively. Mean blood eosinophil counts were increased in the asthmatic group as compared with the normal group, at 0.48 ± 0.04 × 10⁹/liter and 0.15 ± 0.02 × 10⁹/liter, respectively (P < 0.0001).

Kinetics and Dose-Response of Eosinophil Transmigration to 5-Oxo-ETE

In the presence of IL-5, 5-oxo-ETE induced an important transmigration through Matrigel purified blood eosinophils from both normal and asthmatic subjects (Figure 1A). This effect increased over time, and plateaued at 6 h in both groups. The optimal eosinophil migration rate observed was 90.8 ± 2.2% for normal eosinophils and 94.1 ± 2.4% for those from asthmatic subjects. A very weak migration ( 3.2%) was observed in the presence of IL-5 alone (Figure 1B). The dose-response data showed that the maximal effect of 5-oxo-ETE was obtained with a concentration of 1 µM in both cell preparations. At a much lower concentration (0.001 µM), 5-oxo-ETE still induced significant eosinophil transmigration (14.6 ± 3.6%). As compared with a concentration of 1 µM, higher concentrations of 5-oxo-ETE (5 µM and 10 µM) decreased eosinophil transmigration but did not abolish it. In these two sets of experiments, the effects of 5-oxo-ETE on eosinophils from normal and asthmatic subjects were similar.

View larger version (16K): [in this window] [in a new window]   Figure 1.   Kinetics (A) and dose-response (B) of 5-oxo-ETE- induced eosinophil transmigration through Matrigel basement membrane. Eosinophils of healthy volunteers (open bars) and asthmatic subjects (dark bars) were preincubated with IL-5 (10 ng/ml) for 30 min at 37°C and added to the upper chamber. For the kinetic study, the cells were incubated for variable periods (1-18 h), with 5-oxo-ETE (1 µM) being added to the lower chamber (n = 4 in each group). In the dose-response study, 5-oxo-ETE was used at various concentrations and cells were incubated for 18 h (n = 5 for 5-oxo-ETE at 0 µM and 1 µM; n = 3 for 0.1 µM and 5 µM; n = 2 for 0.001 µM and 10 µM).

Comparison of the Effects of 5-Oxo-ETE and PAF on Eosinophil Transmigration

Figure 2 compares eosinophil transmigration obtained with IL-5 alone, PAF + IL-5, and 5-oxo-ETE + IL-5. As observed earlier, IL-5 induced a weak eosinophil migration (mean: 1.4%). As compared with IL-5 alone, PAF (1 µM) in the presence of IL-5 promoted significant and similar eosinophil migration in eosinophils from both normal and asthmatic subjects, of 7.7 ± 3.1% and 5.4 ± 1.7%, respectively (P = 0.0001), with a mean of 6.5 ± 1.7% for all subjects. However, this effect remained much weaker than that observed with 5-oxo-ETE at 1 µM and IL-5, which induced large and similar eosinophil migration in eosinophils from both normal and asthmatic subjects, of 82.0 ± 3.7% and 88.1 ± 3.8%, respectively (P = 0.0001, compared with IL-5 alone and PAF + IL-5). In fact, this migration was even slightly weaker than that observed with 5-oxo-ETE at 0.001 µM (Figure 1B).

View larger version (15K): [in this window] [in a new window]   Figure 2.   Comparison of the effects of 5-oxo-ETE (1 µM) and PAF (1 µM) on eosinophil transmigration through Matrigel basement membrane. Eosinophils of healthy volunteers (open bars) and asthmatic subjects (dark bars) were preincubated with IL-5 (10 ng/ml) for 30 min, added to the upper chamber, and incubated for 18 h at 37°C. 5-Oxo-ETE (1 µM) or PAF (1 µM) was added to the lower chamber. Bars identified by different letters are significantly different (ANOVA, P = 0.0001). Numbers under bars represent the number of subjects for each condition.

Modulation of 5-Oxo-ETE-Induced Eosinophil Migration by IL-5

In the experiments presented in the preceding sections, eosinophils were treated with both IL-5 and a chemotactic factor, either 5-oxo-ETE or PAF. Since Okada and colleagues (7) showed that IL-5, which alone did not induce eosinophil migration, was nevertheless essential for PAF-induced eosinophil migration, we evaluated the effect of IL-5 on 5-oxo-ETE cell migration. Interestingly, 5-oxo-ETE alone was capable of inducing a significant eosinophil migration in cells from normal and asthmatic subjects, and this effect was similar in both groups, at 63.4 ± 0.4% and 59.2 ± 7.0%, respectively (Figure 3). The addition of IL-5 at 30 min before the transmigration assay significantly increased the migration of cells from both normal and asthmatic subjects, 89.7 ± 1.5% and 95.9 ± 2.2%, respectively (P = 0.003).

View larger version (18K): [in this window] [in a new window]   Figure 3.   Modulation by IL-5 of 5-oxo-ETE-induced eosinophil transmigration through Matrigel basement membrane. Eosinophils of healthy volunteers (open circles) and asthmatic subjects (closed circles) were preincubated with or without IL-5 (10 ng/ml) for 30 min, added to the upper chamber, and incubated for 18 h at 37°C. 5-Oxo-ETE (1 µM) was added in the lower chamber. 5-Oxo-ETE induced a significant eosinophil transmigration, which was amplified by IL-5 (P = 0.003, paired Student's t test).

Modulation of Eosinophil uPA-R (CD87) Expression by IL-5

CD87 was constitutively present on eosinophils as measured with flow cytometry (RPMI mean fluorescence: 1.14 ± 0.1). IL-5 significantly increased the surface expression of CD87 (mean fluorescence: 1.52 ± 0.15; P = 0.0001, paired t test), and this effect was similar for eosinophils from normal and asthmatic subjects.

Modulation of 5-Oxo-ETE-Induced Eosinophil Migration by MMP and Serine Protease Inhibition, and MMP-9 and CD87 Blocking

The MMP inhibitor BB-3103 used at concentrations of 1 to 10 µM significantly and progressively inhibited the migration of normal eosinophils, from 79.0% to 45.9% (P = 0.0001; Figure 4). In cells from asthmatic subjects, the inhibition was significant, at 86.4% to 42.3%, and was similar to that observed with normal cells, but interestingly was almost optimal with the lowest inhibitor concentration.

View larger version (16K): [in this window] [in a new window]   Figure 4.   Modulation of eosinophil transmigration through Matrigel basement membrane by the metalloproteinase inhibitor BB-3103. Eosinophils from healthy volunteers (A) and asthmatic subjects (B) were preincubated with IL-5 (10 ng/ml) for 30 min at 37°C. The cells were incubated for 18 h in the upper chamber with or without various concentrations of BB-3103 (1 to 10 µM). 5-Oxo-ETE (1 µM) was added in the lower chamber. In each graph, bars identified by different letters are significantly different (ANOVA, P = 0.0001). Numbers under bars represent the number of subjects for each condition.

In other experiments, the inhibitory effects obtained with anti-MMP-9 antibody were measured and compared with those found with the MMP inhibitor (Figure 5). The monoclonal anti-MMP-9 antibody provided similar transmigration inhibition to that with the MMP inhibitor in both groups of subjects, at 54.3 ± 9.8% and 64.0 ± 3.2%, respectively. The presence of monoclonal anti-CD87 antibody also caused significant inhibition of eosinophil transmigration in both groups, at 51.8 ± 3.9% and 63.3 ± 2.0%, respectively, and the range of this inhibition was similar to that obtained with MMP inhibitor and monoclonal anti-MMP-9 antibody. However, the combination of MMP inhibitor and monoclonal anti-CD87 antibody did not provide better inhibition of eosinophil migration (50.9% for the combination versus 51.0% for BB-3103 and 52.9% for CD87 alone, n = 3). -Aminocaproic acid inhibited eosinophil transmigration (61 ± 5%; n = 4, two normal and two asthmatic subjects) as compared with the IL-5 + 5-oxo-ETE control condition (83 ± 3%). In these subjects, monoclonal anti-CD87 antibody decreased eosinophil transmigration to a mean of 63 ± 3%. These data suggest that inhibition of plasmin obtained with a plasmin inhibitor is similar to that with an antibody to CD87.

View larger version (20K): [in this window] [in a new window]   Figure 5.   Comparison of different inhibitors on 5-oxo-ETE- induced eosinophil transmigration through Matrigel basement membrane. Eosinophils of healthy volunteers (open bars) and asthmatic subjects (dark bars) were preincubated with IL-5 (10 ng/ml) and with or without anti-MMP-9 (2 µM) or anti-CD87 (10 µM) for 30 min. The cells were added to the upper chamber and incubated for 18 h at 37°C. In specific wells, BB-3103 was added directly to the upper chamber after incubation with IL-5. 5-Oxo-ETE (1 µM) was added in the lower chamber. Bars identified by different letters are significantly different (ANOVA, P = 0.0001). Numbers under bars represent the number of subjects for each condition.

Measurement of MMP-9 in Supernatants of Eosinophils

MMP-9 measured in supernatants of Matrigel chambers incubated without cells (RPMI + FBS 10% + 5-oxo-ETE), and with unstimulated (RPMI 1640 alone, RPMI 1640 ± FBS 10%) and stimulated (RPMI 1640 + FBS 10% and 5-oxo-ETE ± IL-5) eosinophils (n = 3 for each condition) remained below the assay sensitivity (all data  0.26 ng/ml).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

This study shows that 5-oxo-ETE, a potent eosinophil chemotactic factor, has a great capacity for promoting eosinophil transmigration through basement membrane components. In the presence of IL-5, this optimal effect on transmigration is much more important, by > 10-fold, than that observed herein or reported by Okada and colleagues with other chemotactic factors, notably PAF (7). Moreover, on a molar basis, 5-oxo-ETE is 1,000 times more potent than PAF, since it induced a sligthly greater transmigration at 0.001 µM than did PAF at 1 µM (Figures 1 and 2). To our knowledge, 5-oxo-ETE is the most powerful promoter of eosinophil migration, and this significant in vitro effect suggests that it could be an efficient mediator involved in tissue eosinophil recruitment in diseases such as asthma. Moreover, the present study shows that eosinophil transmigration is mediated by both MMP and the plasminogen-uPA-R systems. It suggests that a therapy that inhibits the activity of these enzymes could significantly decrease eosinophil recruitment, which is believed to be a significant event in the pathogenesis of many diseases (2).

The magnitude of eosinophil transmigration obtained with PAF in this study is comparable to that reported by Okada and colleagues (7), even though different techniques were used to measure cell transmigration. This observation further strengthens our results obtained with 5-oxo-ETE, which revealed it to be a much more potent inducer of eosinophil transmigration than PAF. The greater capacity of 5-oxo-ETE to promote transmigration could be due to its capacity to activate both proteases and cell chemokinesis. Indeed, 5-oxo-ETE has been shown to be a more potent chemotactic factor than PAF (15). Moreover, it could also induce greater gelatinase activity than PAF; however, in our experiments, we could not demonstrate a greater eosinophil gelatinase activity with 5-oxo-ETE (data not shown), or release of MMP-9 in supernatants of unstimulated or stimulated eosinophils incubated in Matrigel chambers with or without FBS or IL-5 and 5-oxo-ETE. These data are in accordance with those reported by Foissier and coworkers, who did not observe increased eosinophil MMP-9 expression in the presence of IL-5, PAF, or eotaxin (22). This expression was increased at the protein level but not at the messenger RNA level after stimulation with 10 µM phorbol myristate acetate. In contrast, Okada and colleagues (25) showed by zymography and immunoblotting that expression of MMP-9 was increased in culture media of eosinophils by PAF, IL-5, or both, but that its enzymatic activity was increased only in the presence of both PAF and IL-5. The reasons for the differences seen between these results and ours remain unclear. On the other hand, the mouse tumor-derived Matrigel matrix contains a certain amount of MMP-9 (40). However, this Matrigel- derived enzyme did not seem to interfere significantly with cell transmigration, since almost no transmigration was observed in the absence of 5-oxo-ETE (Figures 1 and 2).

In contrast to PAF, 5-oxo-ETE induced a significant cell migration in the absence of IL-5 priming. This could be explained in part by the greater capacity of 5-oxo-ETE to induce eosinophil proteinase activity and chemokinesis, as mentioned earlier. 5-Oxo-ETE could also prime eosinophils via the production of an autocrine cytokine such as GM-CSF or IL-5. IL-5 amplified the effect of 5-oxo-ETE, probably by priming.

The plasminogen-uPA-R system converts the proenzyme plasminogen to plasmin, which has important effects in tissue homeostasis. Plasmin is a potent serine protease that degrades, directly or indirectly by activation of MMP, the tissue extracellular matrix (26). Its role has been demonstrated in granulocyte infiltration into a fibrin matrix (41). In the present study, we showed that plasmin also plays an important role in eosinophil transmigration, since a monoclonal anti-uPA-R antibody and a plasmin inhibitor, -aminocaproic acid, significantly decreased eosinophil transmigration. These experiments do not allow a determination of whether plasmin is acting directly on extracellular matrix components or indirectly via MMP activation (26).

Interestingly, the inhibition of MMPs and of uPA-R provided significant and similar inhibition of eosinophil transmigration, but neither type of inhibition, nor their combination, completely blocked eosinophil transmigration. It could suggest that protease inhibition was incomplete and that sufficient protease activity was present for inducing cell transmigration. Indeed, the activity of these enzymes is localized and prevails on the cell membrane area engaged with the basement membrane through which passage is to occur (29). This area could be physically confined to a certain point beyond the reach of either drug or antibody inhibitors. This possibility could also explain why MMP-9 activity and protein are usually not measureable in the supernatants of transmigrating eosinophils.

On the other hand, other types of proteases could be produced by eosinophils and could be involved in the transmigration process. Indeed, the inability to demonstrate MMP-9 expression raises the possibility that the MMP inhibitors used in our experiments might act on additional MMPs or on membrane proteins containing a disintegrin and metalloproteinase domain (ADAM) (42). Although BB-3103 inhibits various MMPs, the Calbiochem anti-MMP-9 antibody is reported to be specific for MMP-9, suggesting that its anti-MMP blocking activity occurs mainly through the neutralization of MMP-9. To our knowledge, no prior demonstration of ADAM with MMP-9 activity has been described, nor has the presence of ADAM on eosinophils.

No difference was observed between the transmigration response of eosinophils from normal and asthmatic subjects to 5-oxo-ETE. This may seem surprising, since it has been suggested that 5-oxo-ETE acts differently on normodense and hypodense eosinophils (12, 15). Because the asthmatic subjects in our study had increased blood eosinophil counts and did not take inhaled steroids, we could conclude that their eosinophils were activated and hypodense. Nevertheless, in our experiments, these eosinophils had a similar transmigration response to 5-oxo-ETE as that of normal eosinophils. Consequently, our data do not suggest that eosinophils of asthmatic subjects had a greater response to 5-oxo-ETE than did cells of normal subjects. They also suggest that the 5-oxo-ETE receptor number and activity would be similar for eosinophils from normal and asthmatic subjects. Interestingly, the kinetic effects of BB-3103 suggest that MMPs of asthmatic subjects' eosinophils are more sensitive to inhibition. The reason for this effect remains undefined.

In conclusion, this study shows that besides being a potent eosinophil chemotactic factor, 5-oxo-ETE, an arachidonic acid metabolite released by many cell types, is also an efficient activator of eosinophil transmigration and much more potent than PAF. This effect seems to be mediated by various proteinases, notably MMP-9 and the plasminogen-uPA-R system produced by eosinophils. However, more data on 5-oxo-ETE modulation of MMP, the serine protease plasmin, and possibly ADAM are needed. The eosinophil transmigration promoter activity of 5-oxo-ETE should also be further confirmed in vivo (17), but it suggests that inhibition of 5-oxo-ETE or its activity on proteases constitutes a potential therapeutic avenue in diseases in which eosinophil recruitment is involved.