Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Eosinophils accumulate selectively at sites of allergic inflammation. Numerous studies have now implicated specific adhesion molecules and chemokines in these responses (1). For example, the cytokine-inducible endothelial adhesion molecules vascular cell adhesion molecule-1 (VCAM-1, CD106) and intercellular adhesion molecule-1 (ICAM-1, CD54) contribute to aspects of eosinophil rolling, firm adhesion, and transendothelial migration as these cells leave the intravascular compartment to enter tissue sites. Eosinophil-active chemokines, such as those that bind to receptors such as CCR3, trigger migratory responses of these cells, in part by altering the function of integrin counter-receptors for VCAM-1 and ICAM-1 (2). Thus, it is felt that a coordinated series of adhesive and de-adhesive events are required to permit selective, directed leukocyte emigration. Ultimately, the cellular makeup of the inflammatory infiltrate is likely controlled at the tissue level by the pattern of endothelial activation and chemokine production (5).

A variety of in vitro assays have been developed to study the key steps involved in leukocyte recruitment. These include parallel-plate adhesion assays performed under conditions that mimic blood flow conditions, as well as those that examine leukocyte-endothelial adhesion, transendothelial migration, and chemotaxis (3, 6, 7). Indeed, it was these types of assays that initially implicated adhesion molecules and chemokines in selective eosinophil recruitment responses. Additional studies, performed in both animals and humans, have confirmed the roles of VCAM-1, ICAM-1, and CCR3-active chemokines in allergic inflammation (8).

Despite this emerging picture, it remains unclear exactly how and when chemokines actually contribute to the recruitment response. In one scenario, it has been hypothesized that tissue-generated chemokines get displayed on the luminal surface of endothelial cells adjacent to the site of inflammatory insult. Marginated leukocytes then probe the endothelial surface for the presence of chemokines that trigger diapedesis. In vivo, direct evidence for this proposed pathway has been lacking. In vitro, however, experimental models of blood vessels in which flowing leukocytes interact with immobilized endothelial adhesion molecules and chemokines have demonstrated that such conditions can result in rapid deceleration and attachment of rolling leukocytes (9). In contrast, other studies, including those with eosinophils, suggest that depending on the adhesion molecules present, addition of chemokines can result in either rapid attachment or rapid detachment (3, 4, 10, 11). Furthermore, studies suggest that conditions that enhance adhesion may actually inhibit migration (12), so that coordination of these events are required for directed migration of leukocytes out of the circulation into the extravascular compartment. Finally, the key signaling pathways mediating these chemokine- or integrin-dependent changes remain to be fully elucidated, although proteins such as Rho GTPases and mitogen-activated protein kinases (MAPK) have been implicated (13).

We hypothesized that chemokines facilitate leukocyte extravasation by shifting integrin usage in such a way as to reduce adhesion to luminal ligands such as VCAM-1 while simultaneously enhancing adhesion to pericellular ligands such as ICAM-1 (16). In an effort to test this hypothesis and further delineate whether eosinophils are capable of responding rapidly in this manner to chemokines, we developed a parallel-plate, flow-based adhesion assay in which various combinations of adhesion molecules (i.e., VCAM-1 and ICAM-1) and chemokines (e.g., eotaxin-2) are co- adsorbed to plastic surfaces. Using video microscopy, infused cells can be filmed in real time, and interactions between these various substrates can be subsequently quantified. By pretreating cells with various pharmacologic agents, the role of various signaling pathways can also be elucidated. Herein we present data demonstrating that when flowing eosinophils encounter a surface coated with VCAM-1, ICAM-1, and eotaxin-2, CCR3- and MAPK-dependent signals cause these cells to rapidly decrease their VCAM-1-mediated adherence while enhancing their ICAM-1 binding.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents

Human eotaxin-2 was kindly provided by Dr. John White (GlaxoSmithKline Pharmaceuticals, King of Prussia, PA). Human soluble recombinant VCAM-1 (sVCAM-1) and ICAM-1 (sICAM-1) were purchased from R&D systems (Minneapolis, MN). Bovine serum albumin (BSA) was purchased from Sigma Chemical Co. (St. Louis, MO). The MAPK inhibitor PD98059 (17) and the phosphatidylinositol 3 (PI3)-kinase inhibitor LY294002 (18) were purchased from Calbiochem (San Diego, CA).

Antibodies

The CCR3-blocking monoclonal antibody (mAb) 7B11 was kindly provided by Dr. Walter Newman (Millennium Pharmaceuticals, Inc., Cambridge, MA). An isotype-matched control mouse IgG2a mAb was purchased from Caltag (Burlingame, CA), as were blocking mouse IgG1 F(ab')₂ mAbs recognizing human VCAM-1 (clone 1G11B1) and ICAM-1 (clone MEM-111). An intact blocking mouse IgG1 mAb recognizing ICAM-1 (clone 84H10) was purchased from Coulter-Immunotech (Hialeah, FL), and another mouse IgG1 mAb recognizing eotaxin-2 (clone 61016.11) was purchased from R&D Systems. Peroxidase-conjugated affinity purified polyclonal rabbit anti-mouse IgG was purchased from Jackson ImmunoResearch Laboratories (West Grove, PA).

Preparation of Protein-Coated Plates

Tissue culture plates (35 mm; Corning, Corning, NY) were coated overnight at 4°C with 80-µl aliquots of optimal concentrations of proteins (3.4 µg/ml sVCAM-1, 10 µg/ml sICAM-1, or both in the presence or absence of 10 µg/ml eotaxin-2). Any remaining uncoated surfaces on the plates were then blocked with 1% BSA (Sigma) for 1 h at room temperature and kept with phosphate-buffered saline (PBS) containing 1 mM CaCl₂ and 1 mM MgCl₂ before use. To measure the amounts of immobilized proteins, coated plates were incubated with primary mAb recognizing VCAM-1, ICAM-1, or eotaxin-2 for 30 min at room temperature. After washing, the plates were incubated with 1 µg/ml peroxidase-conjugated rabbit anti-mouse IgG for 1 h, washed, then incubated with 75 µl of substrate (ABTS [2,2'-Azino-bis(3-ethylbenzothiazoline-6-sulfonic acid; Sigma]:hydrogen peroxide = 1,000:1) for 5 min at room temperature. Fifty microliters of each solution was transferred to a fresh 96-well plate with wells containing 50 µl of azide. The optical density for each well was measured at 405 nm using an enzyme-linked immunosorbent assay plate reader (Dynatech MR4000; Dynatech, Inc., McLean, VA), and was linear over a range of 0.08-10 µg/ml.

Isolation of Human Eosinophils

Human eosinophils were isolated from ethylenediaminetetraacetic acid-anticoagulated venous blood of donors with mild allergic rhinitis or asthma by Percoll (1.090 g/ml) density gradient centrifugation at room temperature and removal of CD16-positive cells (neutrophils) using immunomagnetic beads exactly as previously described (3). Eosinophil purity (based on the examination of Diff-Quik-stained cytocentrifugation preparations) was > 96%, and viability (by erythrosin B dye exclusion) was nearly 100%.

Eosinophil Adhesion under Flow Conditions

The assembled parallel plate flow assay system consisted of (1) a Plexiglas flow chamber (Glycotech, Rockville, MD) with inlet/ outlet ports, a vacuum line, and silicone gasket; (2) a Nikon TMS-F inverted phase contrast microscope with video capacity (Image Systems, Inc, Columbia, MD); (3) a high-resolution CCD camera (Hamamatsu, Japan); (4) a black and white high-resolution monitor and videocassette recorder (Sony Corp. of America, Park Ridge, NJ); and (5) a Harvard infuse/withdrawal syringe pump (Harvard Apparatus, South Natick, MA) (19). Before assemblage, the flow chamber was filled with media and all air removed from the system. The flow chamber was then inverted with the gasket in place and media placed on the flow path. A 35-mm tissue culture plate onto which VCAM-1, ICAM-1, and/or eotaxin-2 was adsorbed was then placed on top of the chamber and a vacuum created. Once assembled, the chamber and plate were placed on the microscope stage and the flow of cells initiated by the syringe pump attached to the outlet port, so that cells were drawn rather than pushed through the chamber. Eosinophils (2 × 10⁵ cells/ml in RPMI 1640 containing 0.2% BSA) were drawn at a constant flow of 0.5 dyn/cm² for 10 min, conditions previously shown to be optimal for integrin-mediated rolling (20). Throughout the infusion, cells were kept at 37°C by a water jacket that maintained the temperature of each portion of the experimental setup. In some experiments, eosinophils were pre-incubated with the CCR3-blocking mAb 7B11 or an isotype control mAb for 15 min before perfusion. In other experiments, protein-coated plates were pretreated with F(ab')₂ blocking mAb to VCAM-1 (3 µg/ml) or ICAM-1 (34 µg/ml) for 30 min before infusing eosinophils. To examine the role of PI3-kinases and MAPKs in the attachment assays, eosinophils were pre-incubated with LY294002 (10 µM) or PD98059 (100 µM), respectively, for 30 min at 37°C before perfusion across the chamber. These concentrations were chosen based on the literature or, for PD98059, were chosen to be 10-fold higher than the approximate IC₅₀ value of 10 µM for its ability to inhibit chemokine-induced ERK phosphorylation in human eosinophils (Ref. 21 and data not shown). Interactions between cells and plate were visualized in real time with video microscopy. Images were digitized from the videotape recorder and the number of adherent cells that accumulated for 20 s at 2-min intervals was determined by visual counts of the videotapes.

Statistical Analyses

All results are shown as mean ± standard error of the mean (SEM). Statistical analyses were performed using paired Student's t test. The level of significance was set at P < 0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Verification of Tissue Culture Plate Protein Coating

Because endothelial cells constitutively express ICAM-1 but only express VCAM-1 or CCR3-active chemokines when activated (1), we immobilized several different combinations of these molecules. Thirty-five-millimeter plates were coated with ICAM-1 alone, VCAM-1 alone, or both. In some cases, eotaxin-2 was co-adsorbed along with adhesion molecules. Table 1 shows that equivalent coating of wells was usually seen, except for a significant, ~ 50% reduction in the amount of ICAM-1 adsorbed in the presence of the other two proteins, and a significant, ~ 10% reduction in the amount of VCAM-1 adsorbed in the presence of eotaxin-2.

Eosinophil Accumulation on Immobilized Adhesion Proteins and Effect of Eotaxin-2 Co-Immobilization

Figure 1 shows rates of eosinophil accumulation on ICAM-1 alone, VCAM-1, or both, with or without co-immobilized eotaxin-2. As previously reported, we observed that eosinophils accumulated well on VCAM-1 but not ICAM-1 or BSA under flow conditions (3). Eosinophil accumulation on VCAM-1 was decreased by eotaxin-2, whereas it was increased on ICAM-1. The highest level of eosinophil accumulation was seen on surfaces co-coated with both VCAM-1 and ICAM-1. Interestingly, co-immobilization of eotaxin-2 with these two adhesion molecules had no net effect on eosinophil accumulation.

View larger version (17K): [in this window] [in a new window]   Figure 1.   Eosinophil accumulation on immobilized VCAM-1 (circles), ICAM-1 (triangles), or both (squares) in the presence (open symbols) or absence (filled symbols) of immobilized eotaxin-2. Eosinophils were infused at 0.5 dyn/cm2 for 10 min. Data represent mean ± SEM of three separate experiments. *P < 0.05.

Using F(ab')₂ blocking monoclonal antibodies, Figure 2 shows adhesion molecule usage during eosinophil accumulation on VCAM-1-, ICAM-1-, and/or eotaxin-2-coated plates. As in the previous figure, the number of accumulated eosinophils was the same on VCAM-1 plus ICAM-1-coated plates regardless of the presence of eotaxin-2. However, the presence of chemokine significantly changed adhesion receptor usage. For example, anti-ICAM-1 had no effect on eosinophil adhesion to VCAM-1 plus ICAM-1 in the absence of eotaxin-2. However, when ICAM-1 was blocked by antibody, eosinophil accumulation was decreased by eotaxin-2, presumably because the cells release from VCAM-1 but cannot bind to ICAM-1. In contrast, anti-VCAM-1 antibody inhibited eosinophil accumulation on VCAM-1 plus ICAM-1 in the absence of eotaxin-2. However, when VCAM-1 was blocked by antibody, eosinophil accumulation was not decreased by eotaxin-2, presumably because the cells can now bind to ICAM-1. Although the level of eosinophil accumulation was similar in the presence and absence of eotaxin-2, it appears that eosinophils adhered mainly using VCAM-1 in the absence of eotaxin-2 and adhered using ICAM-1 in the presence of eotaxin-2. Furthermore, the effect of immobilized eotaxin-2 on eosinophil accumulation under flow conditions was CCR3-dependent, because pretreatment of eosinophils with a CCR3 blocking mAb prevented the eotaxin-2-induced effects (Figure 3).

View larger version (22K): [in this window] [in a new window]   Figure 2.   Effect of ICAM-1 or VCAM-1 F(ab')2 antibodies on eosinophil accumulation on VCAM-1- and/or ICAM-1-coated plates in the presence (striped bars) or absence (filled bars) of immobilized eotaxin-2. Background binding to BSA-only-coated wells is also shown. Data represent mean ± SEM of three separate experiments. *P < 0.05.

View larger version (28K): [in this window] [in a new window]   Figure 3.   Effect of CCR3 blockade on eosinophil adhesion to VCAM-1-coated plates under flow conditions in response to immobilized eotaxin-2. Eosinophils were pre-incubated with a CCR3 blocking mAb (7B11) or an isotyped matched control mAb before being infused across a plate co-coated with VCAM-1 alone ( filled bars) versus VCAM-1 with eotaxin-2 (striped bars). Background binding to BSA-coated wells (open bar) is also shown. Data represent mean ± SEM of three separate experiments. *P < 0.05.

Role of MAPKs and PI3-Kinases in Eotaxin-2-Induced Changes in VCAM-1 Adhesion

Activation of G protein-coupled receptors, such as CCR3, is known to activate a number of downstream signaling pathways, including those involving serine/threonine kinases and PI3-kinases. To begin to examine the role of some of these kinases in our flow-based attachment assays, we tested the effect of MAPK inhibitors on eosinophil accumulation on co-immobilized VCAM-1 plus eotaxin-2 under flow conditions. Two different inhibitors, PD98059, a MAPK inhibitor that blocks activation of extracellular signal-related kinases (ERK1/2), and LY294002, a PI3-kinase inhibitor, were employed. As shown in Figure 4, PD98059, but not LY294002, inhibited the detachment of eosinophils from VCAM-1 induced by eotaxin-2 under flow conditions.

View larger version (31K): [in this window] [in a new window]   Figure 4.   The role of ERK1/2 and PI3-kinases in eotaxin-2- induced modulation of eosinophil accumulation on VCAM-1- coated plates. Eosinophils were pre-incubated with the ERK1/2 inhibitor PD98059 (100 µM), the PI3-kinase inhibitor LY294002 (10 µM), buffer alone, or an equivalent dilution of dimethyl sulfoxide (DMSO) in buffer for 10 min at 37°C. Experiments were performed in the presence (striped bars) or absence ( filled bars) of co-immobilized eotaxin-2. Background binding to BSA-coated wells (open bar) is also shown. Data represent mean ± SEM of three separate experiments. *P < 0.05.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The experiments presented herein have explored some of the mechanisms by which activation of eosinophils rapidly alters their integrin-dependent adhesiveness for specific ligands. A novel aspect of the assays used was that they examined interactions of eosinophils under flow conditions with specific combinations of adhesion molecules and chemokines adsorbed to plastic surfaces. Thus, the model more closely resembles the in vivo paradigm in which leukocyte-endothelial cell adhesive interactions are modulated by chemokines displayed on the luminal surface of activated endothelium. Using this model (Figure 1), it was demonstrated that in the absence of chemokines, eosinophils accumulate efficiently on a VCAM-1-coated surface, but not on surfaces coated with ICAM-1 or BSA. These results are consistent with those reported by this laboratory and others (3, 11). Optimal accumulation was observed with surfaces co-coated with both VCAM-1 and ICAM-1. When eotaxin-2 was co-immobilized with VCAM-1, reduced eosinophil accumulation was observed, whereas for ICAM-1, enhanced accumulation was seen. However, when eosinophils were allowed to accumulate under flow conditions on surfaces to which VCAM-1, ICAM-1, and eotaxin-2 were all adsorbed, levels of accumulation were similar to those in the absence of eotaxin-2. This seems to partially contradict prior reports using cytokine-activated endothelium, where eosinophil arrest was observed (4). However, a careful analysis of integrin usage in the presence or absence of eotaxin-2 co-immobilization revealed a shift from predominantly VCAM-1-dependent adhesion to a pattern that is more dependent on ICAM-1 (Figure 2), similar to that reported for T cells (22). Finally, using antibodies or pharmacologic antagonists, the work presented extends previous studies by showing that the CCR3-dependent changes in adhesion were mediated via MAP kinases such as ERK1/2, but not via PI-3 kinases (Figures 3 and 4). Taken together, these results indicate that under conditions where endothelial cells are activated to express ICAM-1, VCAM-1 and eotaxins (for example, IL-4 or IL-13 stimulation [23]), eosinophil attachment is facilitated. However, the presence of CCR3-active chemokines promotes a greater role for ICAM-1-mediated attachment.

It is important to note that changes in the levels of immobilized adhesion proteins and chemokines were monitored by enzyme-linked immunosorbent assay (Table 1). This allowed a determination of whether adsorption of more than one protein would significantly alter their levels on the plate surface. In general, similar amounts of proteins were adsorbed, regardless of the combination used for coating. A possible exception to this was a reduction of about 10% and 50%, respectively, in the amounts of eotaxin-2 and ICAM-1 adsorbed to the plate in the presence of multiple coat proteins. Despite this reduction, an enhanced chemokine-mediated attachment to ICAM-1 was still detectable, so the effect may in fact have been greater if equal amounts of ICAM-1 coating had been achieved. Another potential weakness of this experimental model is that it is difficult to determine whether the levels of coating achieved on the plastic surface resembles those that might normally be found on inflamed endothelium in vivo.

The ability of chemokine receptor activation to rapidly influence integrin function has been reported in a number of model systems, including those involving eosinophils (3, 4, 11). Some reports focusing on eosinophils suggest a CCR3-mediated, rapid yet transient upregulation of VLA-4-dependent adhesion (4), whereas others, including our own work, do not (3, 11). The reason for these differences is unclear. However, all appear to agree that CCR3 activation causes a sustained downregulation of 1 integrin function along with a concomitant enhancement of 2 integrin function. Although the exact interaction between G-protein-coupled chemokine receptor activation and regulation of integrin function remains to be elucidated, it appears that a variety of intracellular pathways such as Rho are involved (13). In addition, studies of receptors for chemokines and other chemoattractants have strongly implicated a number MAPKs, such as ERK1 and ERK2, in eosinophils and other cells (14, 15, 21). A novel finding from our observations is that the crosstalk between chemokine receptors and integrins can be blocked by PD98059, a selective antagonist for ERK1/2 (17), whereas the PI3 kinase inhibitor LY294002 (18) had no effect (Figure 4). Whether the changes in adhesiveness observed in the present and other studies are primarily due to changes in integrin affinity, avidity, or clustering remains to be elucidated.

In summary, we have shown that under flow conditions, eosinophils encountering adsorbed eotaxin-2 on a plastic surface together with integrin ligands such as ICAM-1 and VCAM-1 rapidly undergo CCR3-dependent, ERK1/2-dependent changes in adhesiveness, shifting their attachment preferences toward ICAM-1 and away from VCAM-1. The changes observed are in line with previous work demonstrating that initial eosinophil attachment under flow is primarily selectin- and VCAM-1-dependent (1). Subsequent chemokine-induced transendothelial migration across cytokine-activated endothelium is almost exclusively dependent on ICAM-1 and 2 integrins, with little or no involvement of VCAM-1 (6). Thus, triggering of CCR3 on attached, intravascular eosinophils may help to facilitate release from luminal VCAM-1 while simultaneously facilitating 2 integrin-ICAM-1 adhesive events, including diapedesis. These findings shed additional light on pathways by which chemokines, integrins and their ligands regulate eosinophil adhesion under flow conditions, and suggest that a switch between adhesive pathways occurs when circulating leukocytes undergo margination and transendothelial migration in vivo.