Stimulation of eosinophil IgE low-affinity receptor leads to increased adhesion molecule expression and cell migration

Stimulation of eosinophil IgE low-affinity receptor leads to increased adhesion molecule expression and cell migration. S. Lantero, G. Alessandri, D. Spallarossa, L. Scarso, G.A. Rossi. #ERS Journals Ltd 2000. ABSTRACT: Immunoglobulin binding on eosinophil surface receptors results in activation of these cells. Evaluating blood eosinophils from atopic subjects, it was investigated whether ligation of immunoglobulin E low-affinity receptor (FceRII/ CD23) with specific monoclonal antibodies (Mabs) resulted in enhanced eosinophil migration and adhesion molecule expression. Eosinophils from 20 subjects with allergic asthma (atopic individuals) and nine nonatopic normal individuals (controls) were purified using Percoll gradients. The effect of antihuman CD23 Mabs on: 1) eosinophil migration through human umbilical vein endothelial cells (HUVECs); and 2) eosinophil expression of the adhesion molecules leukocyte function-associated antigen-1 (LFA-1, CD11a/CD18), macrophage antigen-1 (Mac-1, CD11b/CD18) and very late activation antigen-1 (VLA-4, CD49d/CD29) was evaluated by specific Mab staining and flow cytometric analysis. As compared to controls, freshly isolated eosinophils from atopic individuals showed enhanced migration through HUVECs (p<0.05) and increased LFA-1 expression (p<0.01), but similar Mac-1 and VLA-4 expression (p>0.1 for both). In both controls and atopic individuals, eosinophil incubation with antihuman CD23 Mabs induced a dose-dependent increase in cell migration through HUVECs, significant at antihuman CD23 Mab concentrations of 5 mg.mL (p>0.05 for all). Similarly, incubation of the cells with antihuman CD23 Mabs induced dose-dependent upregulation of LFA-1 and Mac-1 expression, whereas no changes in VLA-4 expression were observed (p>0.1). Finally, the enhanced eosinophil migration induced by antihuman CD23 Mab stimulation was significantly inhibited by antihuman LFA-1 (8414% (meanSEM); p<0.01) and VLA-4 Mabs (4715%; p<0.05) but not by antihuman Mac-1 Mabs (p>0.1). In both atopic and control subjects, immunoglobulin E, low-affinity receptor stimulation induces functional changes in eosinophils characterized by increased eosinophil migration associated with enhanced late function antigen-1 and Mac-1 expression. Eur Respir J 2000; 16: 940±946. *Pulmonary Division and ***Blood Bank, G. Gaslini Institute, and **Dept of Cell Biology, National Cancer Research Institute, Genoa, Italy.

In allergic asthma, exposure to the sensitizing allergen leads to eosinophil recruitment and activation. Eosinophils have the potential to injure human lung tissues and their presence in the airway mucosa has been associated with degree of airflow limitation and morphological derangement of the bronchial epithelium. The current thought is, therefore, that eosinophils act as major effector elements in the pathogenesis of allergic asthma [1,2].
Eosinophil recruitment is a complex mechanism, which includes the expression of surface adhesion molecules on circulating cells, able to interact with their counterreceptors, expressed on vascular endothelial cells [3]. It has been demonstrated that T-helper 2 (Th2) cytokines (interleukin (IL)-3, IL-5 and granulocyte-macrophage colony-stimulating factor) and chemokines (eotaxin and regulated on activation, normal T-cell expressed and secreted are able to not only prime and activate eosinophils but also increase cell surface receptor and adhesion molecule expression on these cells [3±7]. In atopic asthma, a condition characterized by increased secretion of Th2 cytokines and chemokines, the expression of adhesion molecules on the eosinophil cell membrane is increased and appears to be involved in cell migration [8]. Besides inflammatory mediators, stimulation of surface receptors specific for immunoglobulins (Igs) also plays a significant role in modulating cell function [9,10]. Eosinophils are able to bind IgE through highand low-affinity receptors (FceRI and FceRII or CD23 respectively) [11,12], and, although the specific functions of FceRI and FceRII are still controversial [10,12], it has been demonstrated that anti-CD23 monoclonal antibodies (Mabs) or IgE/anti-IgE immune complexes are able to increase hydrogen peroxide and tumour necrosis factor production by eosinophils [9]. The observation that allergic patients with a high level of allergen-specific IgE show increased expression of eosinophil adhesion molecules and the reported correlation between serum allergen-specific IgE, the number of eosinophils infiltrating the airways and the severity of asthma [8] suggest that FceRII may modulate eosinophil locomotion. Against this background, the present study was designed to evaluate in vitro whether stimulation of the FceRII by antihuman CD23 Mabs enhanced eosinophil migration through human umbilical vein endothelial cells (HUVECs) and expression of the adhesion molecules leukocyte function-associated antigen-1 (LFA-1, CD11a/CD18), macrophage antigen-1 (Mac-1, CD11b/ CD18) and very late antigen-1 (VLA-4, CD49d/CD29) involved in eosinophil transendothelial migration.

Population
Eosinophils were obtained from nine nonatopic normal individuals (controls) (6±11 yrs; seven male and two female) and 20 atopic asthmatic subjects sensitized to house dust mites, as demonstrated by skin-prick test and radioallergosorbent test (Pharmacia) (atopic individuals) (6±14 yrs; 15 male and 5 female) (table 1). Asthma was defined according to the criteria of the American Thoracic Society [13]. Controls and atopic individuals had not suffered from respiratory infections in the previous 4 weeks and were not under any treatment other than b 2stimulants on an "as necessary" basis. The study was approved by the G. Gaslini Institute Ethical Committee and parents or tutors of all subjects gave their informed consent.
Because of the relatively high numbers of cells required, eosinophils collected from different individuals were often used to perform different kinds of experiment, as indicated below.

Eosinophil purification
Isolation of blood eosinophils was performed on discontinuous Percoll gradients, as previously described [14]. Briefly, to 10 mL heparinized blood, an equal volume of 6% dextran in 0.9% NaCl was added, mixed gently and incubated at 378C. After 30±40 min, the upper phase was collected, washed once in PBS, resuspended in 1.5 mL 1.070 g . mL -1 Percoll containing 5% FCS and layered on a discontinuous Percoll gradient with the following volumes (mL) and densities (g . mL -1 ) respectively: 1.5, 1.100; 3, 1.090: 3, 1.085; and 3, 1.080 [14]. The Percoll gradient densities were obtained by mixing nine parts of Percoll with one part of 103HBSS and then diluting the 90% Percoll solution with 13HBSS containing 5% FCS. The normodense eosinophils recovered from 1.090±1.095 g . mL -1 Percoll were washed twice in PBS and resuspended at 1310 6 eosinophils . mL -1 in complete medium. The hypodense population represented only a small proportion of the total eosinophil population and, because of the low number of cells, could not be used to perform any study. The eosinophils recovered were 70±80% pure (as established by Diff-Quik staining, with neutrophils as contaminant) and 90% viable, as determined by trypan blue dye exclusion test.

Evaluation of eosinophil migration through endothelial cell layer
The day before the assay, 5610 5 HUVECs were seeded on polycarbonate membranes with 5-mm pores coated with type I collagen, in order to reach confluence on the day of the assay, and incubated in EBM under 5% CO 2 conditions at 378C [15]. On the day of the assay, the membranes were placed in Boyden chemotaxis chambers with the HUVEC side up. The lower wells were filled with 50 mL complete medium, to evaluate random migration, or with 0.1 mg . mL -1 C5a in complete medium, as eosinophil chemoattractant [16]. The C5a concentration had been established in previous dose/response experiments on eosinophils from both controls and atopic individuals. Unstimulated or CD23-stimulated eosinophils (see below) in 100 mL complete medium at a concentration of 1310 6 cells . mL -1 were seeded in the upper chamber compartments, on the HUVEC side. After a 3-h incubation under 5% CO 2 conditions at 378C, the membranes were detached, the HUVECs peeled off the upper side and the eosinophils that had migrated through the HUVECs to the lower side were fixed, stained with Diff-Quick and counted using a light microscope. All the chemotactic conditions were tested in duplicate and the data expressed as number of eosinophils migrated in 10 high-power fields (HPFs, magnification 3400) [16].
The red fluorescence intensity of the CD16b-cells (eosinophils) was obtained for 10,000 acquired cells using flow cytometry (FACScan; Becton Dickinson, Milan, Italy) and expressed as mean fluorescence channel (MFC) [18]. To compare the fluorescence intensity of different samples, the cells were acquired under identical logarithmic amplifier settings and analysed using Lysis II software (Becton Dickinson). After conversion to linear fluorescence intensity units the mean background obtained using the control antibody (PE-conjugated antihuman Mab to CD3) was subtracted from the mean fluorescence intensity of the specifically stained cells to obtain a linear function of fluorescence intensity over a wide range [19].

Eosinophil stimulation by immunoglobulin E lowaffinity receptor ligation with antihuman monoclonal antibody to CD23
Eosinophils from eight controls and 14 atopic individuals were resuspended in complete medium, seeded in 96-well U-bottomed plates (Costar, Cambridge, MA, USA) in a total volume of 100 mL at 10 6 eosinophils . mL -1 . Flow cytometry on CD23-stained eosinophils was performed before the assays to demonstrate the expression of FceRII on eosinophils of all subjects who were the source of cells for these experiments. Preliminary sets of experiments demonstrated that the length of time required to detect antihuman Mab to CD23-induced modification of adhesion molecule expression on eosinophils was~3 h. The cells were, therefore, incubated with different concentrations of antihuman CD23 (0.5, 2.5, 5, 10 or 20 mg . mL -1 ) or, as control, antihuman Mab to CD45, which recognizes all leukocytes for 3 h under 5% CO 2 conditions at 378C. After incubation, chemotactic activity and expression of adhesion molecules were evaluated as described above. To evaluate the functional role of the different adhesion molecules in cell migration, eosinophils were preincubated in the presence or absence of antihuman Mabs to CD11a, CD11b or CD49d or mouse IgG, as isotype control, (10 mg . mL -1 ) for 30 min under 5% CO 2 conditions at 378C prior to the chemotaxis assay.

Statistical analysis
Data are expressed as meanSEM. Statistical comparisons between different cell culture conditions were performed using an unpaired t-test or the Mann-Whitney U-test, when appropriate [20]. Data were considered significant at a p-value of <0.05.

Eosinophil migration through human umbilical vein endothelial cells and adhesion molecule expression
In both controls and atopic individuals, C5a significantly enhanced eosinophil locomotion above random migration levels. However, as compared to controls, eosinophils from atopic individuals showed greater migration through HUVECs towards C5a (164 versus 224 eosinophils . 10 HPF -1 , p<0.05) ( fig. 1a). In addition, as compared to controls, increased membrane expression of LFA-1 was detected on eosinophils from atopic individuals (336 versus 625 MFC, p<0.01), whereas there were no difference in the expression of Mac-1 and VLA-4 (p>0.1) (fig. 1b). No staining was observed with the mouse IgG isotype control (data not shown). The process of eosinophil separation did not appear to influence the expression of adhesion molecules, since no differences in the levels of LFA-1, Mac-1 and VLA-4 expression were found when eosinophils in unseparated whole-blood samples or after discontinuous Percoll gradient separation, in both atopic individuals and controls (p>0.1 for both), were evaluated (data not shown).

Immunoglobulin E low-affinity receptor ligation activity on eosinophils
All of the atopic individuals and most of the controls (six of nine) expressed low but detectable levels of CD23 (p<0.05 for both versus CD3, the isotype control). The mean expression level was slightly but not significantly higher in atopic individuals compared to controls (p>0.1) ( fig. 2). Incubation of eosinophils with antihuman Mabs to CD23 induced a dose-dependent increase in cell migration towards 0.1 mg . mL -1 C5a in both controls and atopic individuals. In contrast, no changes in cell locomotion through HUVECs were detected when eosinophils were cultured in the presence of different concentrations of the control antihuman Mabs to CD45 (p>0.05 for all) (fig. 3). The increase in cell migration induced by antihuman Mabs to CD23 reached statistical significance at Mab concentrations of 5 mg . mL -1 in both controls (p<0.01) and atopic individuals (p<0.05). Interestingly, eosinophil stimulation with the antihuman Mab to CD23 also induced a dose-dependent increase in LFA-1 and   antihuman Mab to CD23-induced increase in LFA-1 expression reached statistical significance at Mab concentrations of 5 mg . mL -1 in atopic individuals and of 10 mg . mL -1 in controls ( fig. 4a and d). The increase in Mac-1 expression reached statistical significance at Mab concentrations of 10 mg . mL -1 in both atopic individuals and controls (figs. 4b and e).

Adhesion receptor activity and eosinophil migration through human umbilical vein endothelial cells
In order to evaluate the functional role of the different adhesion molecules in modulating cell migration, CD23stimulated eosinophils from five atopic individuals were preincubated with blocking Mabs directed against LFA-1 (CD11a), Mac-1 (CD11b) or VLA-4 (CD49d), or with a mouse IgG, as isotype control, and tested in the chemotaxis assay using C5a as chemotactic agent. Eosinophil migration through HUVECs was totally inhibited by antihuman Mabs to CD11a (284 versus 62 eosinophils . 10 HPF -1 for controls versus atopic individuals, p=0.001), and only partially by antihuman Mabs to CD49 (284 versus 156 eosinophils . 10 HPF -1 , p<0.05), whereas antihuman Mabs to CD11b had no modifying effect on cell locomotion (p>0.1) (fig. 5).

Discussion
Evaluating partially purified blood eosinophils in vitro, it was demonstrated that unstimulated cells from atopic asthmatic subjects show a significantly higher cell migration rate and increased expression of LFA-1 compared to eosinophils from controls. No differences in Mac-1 or VLA-4 expression were observed. Eosinophils from most controls and all atopic individuals showed similar CD23 expression by fluorescence-activated cell sorting (FACS) analysis, and cell stimulation with anti-IgE low-affinity receptor Mabs enhanced, in a dose-dependent manner, eosinophil migration through HUVECs and LFA-1 and Mac-1 (but not VLA-4) expression in both controls and atopic individuals. Finally, eosinophil migration through HUVECs was totally inhibited by the antihuman Mabs to LFA-1 and partially by the antihuman Mabs to VLA-4.
The finding that eosinophil migration and LFA-1 expression were increased in atopic asthmatic subjects is consistent with previous reports and with the concept that "preactivation" of circulating polymorphonuclear leukocytes occurs in allergic asthma [21,22]. These biological properties of circulating eosinophils may, at least partially, explain the substantial migration of these cells into target tissue that occurs during acute asthma attacks and shortly after natural or experimental allergen exposure [23].
Cytokines and chemokines are able to preactivate eosinophils and increase adhesion molecule expression and cell locomotion [3±7]. In the present study it was also demonstrated that stimulation of FceRII, with antihuman Mabs to CD23 enhanced eosinophil migration through HUVECs and adhesion molecule (LFA-1 and Mac-1) expression.
Although eosinophils also display other IgE receptors, such as FceRI and Mac-2/ebp, it appears that many IgEinduced eosinophil functions are mediated mainly by FceRII/CD23 [9]. In addition, the demonstration that, in atopic patients, greater LFA-1 expression corresponds to disease severity and to higher serum levels of allergenspecific IgE [8] further supports the hypothesis of a significant role of FceRII in allergic asthma. In contrast with a recent paper showing, by flow cytometry, no CD23 expression on eosinophils isolated from atopic subjects [10], detectable CD23 was found on eosinophils from all atopic subjects and the majority of the controls in the present study. Patient selection and/or methodological differences may account for the discrepancy in results between the present results and those previously reported [24].
Only eosinophils from controls showing CD23 expression by FACS analysis were used in the CD23 stimulation experiments; these responded in a dose-dependent manner to antihuman Mabs to CD23 stimulation by increasing both LFA-1 and Mac-1 expression. The greater effect of anti-CD23 on adhesion molecule expression on eosinophils from controls than atopic individuals could be due to the fact that eosinophils from atopic individuals are in an activated state [5] and their capacity to increase adhesion molecule expression is almost saturated. Although a dose/ response effect of CD23 stimulation on both eosinophil migration and adhesion molecule expression were obtained, no correlation between these biological functions was observed in controls and atopic individuals. As demonstrated for other surface molecules, receptor functions are related not only to the "amounts" of molecules expressed but also to their biological properties (affinity, avidity state, etc.) [8].
Although stimulation of FceRII induced an increase in LFA-1 and Mac-1 adhesion molecule expression, but not in VLA-4, the experiments performed in the present study with blocking Mabs suggest that, in addition to LFA-1, VLA-4 but not Mac-1 is involved in eosinophil transendothelial migration. LFA-1 is the adhesion molecule which is specifically involved in eosinophil migration through the endothelium [25,26] whereas VLA-4 appears to mainly determine the initial adherence process to the vessel wall, allowing the cells to slow down in the bloodstream [27,28]. Mac-1 was not efficient in the present experimental system in mediating eosinophil transendothelial migration. Indeed, the demonstration that, in asthmatic patients, Mac-1 expression is increased in bronchial eosinophils, as compared to blood eosinophils from the same subjects [29], suggests its involvement in processes that follow transendothelial migration, e.g. interaction with other airway cells in the tissues.
Adhesion mechanisms can be upregulated through increased molecule expression on the cell membrane or through conversion of the molecule from a "low" to a "high" avidity state.
Modifications of LFA-1 and Mac-1 expression, observed in the present study following FceRII stimulation, appear to be related, at least in part, to increased molecular density on the cell membrane. These changes were detected by flow cytometry 2±4 h after stimulation, a period of time typical of that required for the regulation of surface adhesion receptor density at the messenger ribonucleic acid level [28,30]. In addition to increased molecular density, qualitative changes in adhesion molecule function, which occur on a much shorter timescale (minutes), could also be involved in the observed enhancement of eosinophil migration through HUVECs [28,30].
In summary, in both atopic and control subjects, the functional presence of immunoglobulin E low-affinity receptor on blood eosinophils and its involvement in regulating the transendothelial trafficking of eosinophils was demonstrated.