Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Eosinophils (EOS) are known to be important effector cells in allergic diseases and helminthic infections (1). Several cytokines, including interleukin (IL)-3, IL-5, and granulocyte-macrophage colony-stimulating factor (GM-CSF) are reported to regulate the development of EOS from hematopoietic stem cells, and to support their survival in vitro (1, 2). These cytokines are also capable of inducing EOS degranulation, and will dramatically enhance a variety of EOS functions, including adhesion, chemotaxis, leukotriene production, and killing of parasitic larvae (3). These effects, often referred to as "cytokine priming," occur along with a number of phenotypic changes in the cells, including reduced granulation, vacuolization, and expansion of their cytoplasm, leading to a reduction in cell density and the so-called hypodense EOS phenotype (4, 5). The number of hypodense cells has been correlated with allergic-disease severity (6, 7), and it is now believed that exposure to these priming cytokines in vivo may be responsible for these changes in allergic patients.

Because these phenotypic changes correlate with EOS activation and disease severity, attempts have been made to define cell-surface markers that can distinguish normodense, unactivated EOS from hypodense or activated EOS. In one study, hypodense EOS generated in vitro were found to have higher levels of the ₂-integrin subunit CD18, even though levels of the CD11  subunits were similar (8). Other cell-surface structures, such as human-leukocyte-associated antigen-DR (HLA-DR), CD16, and CD54 have been reported to appear on EOS after exposure to interferon- (IFN-) in vitro (9). In other studies, CD69 was found to be expressed on eosinophils from patients with parasitic diseases (12) or during culture with IL-3, IL-5, or GM-CSF (13). With the possible exception of CD69, which is expressed by many activated cell types (14, 15), no cell-surface marker has been found to be of use in detecting primed EOS.

We hypothesized that changes in cell-surface phenotype would accompany EOS priming. To test this hypothesis, we assembled a panel of monoclonal antibodies (mAbs) directed against a variety of cell-surface markers to screen for two types of activation markers: (1) those absent on freshly isolated peripheral blood EOS but present after in vivo activation or culture for up to 72 h with cytokines; or (2) markers normally expressed on fresh EOS that are increased during cytokine culture. These studies were made possible in part by our recent participation in the Fifth International Leukocyte Typing Workshop, in which a panel of over 450 mAbs was used to phenotype purified human EOS. We confirm that the surface marker CD69 is expressed on activated EOS and not on resting EOS, and report the novel observation that levels of CD44 are increased on EOS activated in vitro or in vivo.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Reagents

Bovine serum albumin (BSA) and hyaluronic acid were purchased from Sigma Chemical Company (St. Louis, MO). Human recombinant IL-3, IL-5, GM-CSF, and IFN- were purchased from R&D Systems (Minneapolis, MN).

Murine mAbs

The mAbs to CD11a (MHM24 [IgG1]) and CD18 (H52 [IgG1]) were generously provided by Dr. James Hildreth of the Johns Hopkins University School of Medicine, Baltimore, MD, and the mAb to CD16 (3G8 [IgG1]) was provided by Dr. Paul Guyre of the Dartmouth Medical School, Hanover, NH. The mAbs to CD44 (J173 [IgG1]) and CD49d (HP2/1 [IgG1]) were purchased from Immunotech Inc. (Westbrook, ME), the mAb to CD69 (Leu23 [IgG1]) was obtained from Becton-Dickinson Inc. (Mountain View, CA), and a phycoerythrin (PE)-conjugated mAb to CD9 was obtained from Research Diagnostics Inc. (Flanders, NJ). Irrelevant isotype control mouse mAbs were purchased from Coulter (Hialeah, FL). In addition, approximately 475 mAbs were provided from the Fifth International Workshop on Human Leukocyte Antigens; these reagents, along with EOS phenotyping data using these reagents, are described elsewhere (16, 17). Certain of these mAbs (e.g., recognizing CD69) are mentioned in the text.

Eosinophil Purification

For most experiments, EOS were purified from blood of mildly allergic donors or late-phase bronchoalveolar lavage fluid (BALF) samples through CD16-negative selection, using an immunomagnetic bead technique as previously described (18, 19). The purity and viability of EOS were  97%. For experiments comparing the phenotype of normodense and hypodense EOS, 20 ml of freshly drawn ethylenediamine tetraacetic acid (EDTA)-anticoagulated blood were diluted with 2.5 volumes of sterile 0.9% NaCl. This mixture was layered on two discontinuous density gradients consisting of 8 ml of Percoll (Pharmacia, Uppsala, Sweden) at 1.085 g/ml and 8 ml of Percoll at 1.090 g/ml in a 50-ml conical tube. These gradients were centrifuged at room temperature for 20 min at 400 × g in a swinging-bucket rotor. The cells were carefully collected from each interface and washed once. Erythrocytes were then removed through hypotonic lysis, and the total number of viable leukocytes was counted with a hemacytometer and Erythrosin B stain (Sigma). Differential cell counts were determined by examining 200 cells on cytocentrifuge preparations stained with Diff-Quik (VWR Scientific Products, Bridgeport, NJ).

Eosinophil Culture

EOS were cultured for up to 72 h in up to 100 ng/ml of IL-3, IL-5, GM-CSF, or IFN-, and their viability was determined as previously described (20). Briefly, EOS were suspended in RPMI 1640 medium (GIBCO BRL, Grand Island, NY) supplemented with 10% fetal bovine serum (High Clone Laboratories, Inc., Logan, UT), 100 U/ml penicillin G, 100 µg/ml streptomycin, and 0.25 µg/ml amphotericin B (GIBCO BRL). Cells were cultured at a density of 5 × 10⁵/ml for up to 72 h in the presence or absence of cytokines at 37°C in a humidified atmosphere containing 95% air and 5% CO₂. The cultures were done in 24-well sterile, flat bottom plates (Costar Corp., Cambridge, MA) previously coated with 1% BSA.

Cell Lines

The KU812 basophil-like cell line (21) was maintained in supplemented media, and was chosen for use in the present studies because flow-cytometric analyses revealed it to have comparable levels of CD44 to those on unstimulated EOS, and other analyses demonstrating functionally active CD44 on these cells, thus allowing their use as positive controls in the hyaluronate adhesion assays (see the subsequent discussion). The cells were maintained in RPMI 1640 medium (GIBCO BRL), supplemented with 10% fetal bovine serum (High Clone Laboratories), 100 U/ml penicillin G, 100 µg/ ml streptomycin, and 0.25 µg/ml amphotericin B (GIBCO BRL).

Flow-cytometric Analysis of EOS Surface Markers

Expression of various cell-surface markers on EOS was examined with indirect immunofluorescence and flow cytometry as previously described (19). Gating based on light scatter was used to eliminate debris and dead cells. Freshly purified EOS or EOS harvested after culture were incubated with saturating concentrations of mAb or an equivalent concentration of isotype-matched control mAb. Cells were washed and then incubated with saturating dilutions of R-PE-conjugated F(ab')₂ fragments of goat antimouse IgG antibody (Tago Inc., Burlingame, CA). For analysis of levels of CD44, CD49d, and CD69 on hypodense versus normodense EOS in which contamination with neutrophils occurred, granulocyte scatter along with dual-color labeling with CD9 were used to distinguish EOS (CD9⁺) from neutrophils (CD9) (17, 22). In brief, after incubating cells with saturating concentrations of unconjugated primary mAb (CD44, CD49d, CD69, or an isotype-matched control), the cells were incubated with fluorescein isothiocyanate (FITC)-conjugated goat antimouse IgG (Tago). After a blocking step with excess mouse IgG, cells were incubated with R-PE-conjugated mAb to CD9 (or an R-PE-conjugated, isotype-matched control mAb [Becton-Dickinson]). After fixation in 1% paraformaldehyde in phosphate-buffered saline (PBS), at least 5,000 cells were evaluated with an EPICS Profile flow cytometer (Coulter).

Hyaluronic Acid Adhesion Assays

Adhesion of ⁵¹Cr-labeled EOS and KU812 cells to immobilized hyaluronic acid was tested as described (23). Briefly, 96-well microtiter plates (NUNC Maxi-sorb immunoplates; PGC Scientific Corp., Gaithersburg, MD) were coated overnight at 4°C with 50-µl aliquots of hyaluronic acid (0.016 to 500 µg/ml) diluted in PBS. The wells were blocked with 3% heat-denatured BSA (65°C, 1 h), and 50-µl aliquots of ⁵¹Cr-labeled EOS or KU812 cells were added to wells in triplicate. Cells were allowed to adhere for 60 min at 37°C, after which the nonadherent cells were removed, adherent cells were lysed, and the radioactivity of adherent cell lysates was determined with a gamma counter. Percent adherence was calculated by comparing the radioactivity of adherent cell lysates with that of separate 50-µl aliquots of cell suspension.

In certain experiments, saturating concentrations of mAbs to CD44 or CD11a (control) were added to the wells simultaneously with cells, and were allowed to remain throughout the entire adhesion assay.

Statistical Analyses

Data are presented as mean ± SEM. Statistical significance was determined by one-way or two-way analysis of variance (ANOVA) as appropriate, and values were considered significant at P < 0.05.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Initial experiments were designed to determine whether we could identify any surface markers absent on freshly isolated peripheral blood EOS but present after activation in vivo or in culture for up to 72 h with cytokines. Thirty-five pools of 10 mAbs from the Fifth International Workshop on Human Leukocyte Antigens were prepared, using mAbs that did not react with normodense peripheral blood EOS (17). These pools were tested for binding to EOS from late-phase BALF, or to cultured EOS (5 ng/ml IL-5, 24 h); values for mean fluorescence intensity were compared with those determined with matched concentrations of control, irrelevant mouse IgG1. Three mAb pools were identified that gave values for fluorescence intensity that was greater in primed EOS than with fresh EOS. The 30 mAbs constituting these three pools were then tested individually for binding to EOS from late-phase BALF (Figure 1). In each pool, an mAb to CD69 was responsible for essentially all activity; similar results were obtained with IL-5-cultured peripheral blood EOS (24 to 72 h, data not shown). This confirms previous reports that CD69 is an EOS activation marker (12, 13, 26).

View larger version (24K): [in this window] [in a new window]   Figure 1.   EOS cultured with IL-5, or EOS isolated from late-phase BALF, express surface antigens not found on freshly isolated peripheral blood EOS. Thirty-five pools of 10 mAbs each that did not react with freshly isolated EOS were used to see whether any of these mAbs bound to EOS at levels higher than control IgG after culture (5 ng/ml IL-5, 24 h), or to EOS from late-phase BALF. Three such pools were identified (data not shown). The 30 mAbs constituting these three pools were tested individually for binding to EOS from late-phase BALF. In each mAb pool, an mAb to CD69 (either mAb C1.18 or UN6) was responsible for essentially all activity. Results are representative of at least three separate experiments.

In the next series of experiments, we screened for surface markers normally expressed on fresh EOS that were increased during activation. Initial experiments were performed to determine whether markers normally expressed on fresh EOS could be increased after culture with cytokines. A panel of 16 mAbs recognizing constitutively expressed markers on EOS for which we had sufficient remaining quantities of mAb for further studies (CD9, CD37, CD40, CD44, CD45, CD45RB, CD45RO, CD48, CD50, CD53, CD55, CD58, CD63, CD67, CD73, and CD81) were tested for binding to EOS after 24 h of culture with or without 10 ng/ml IL-5. These studies determined that CD44 was the only marker that increased with IL-5 culture in initial experiments (data not shown). All subsequent experiments were therefore focused on the effects of cytokines on CD44 expression; levels of CD11a were simultaneously analyzed as a control. Figure 2 shows the kinetics and concentration dependence of the effect of IL-5 on EOS CD44 and CD11a expression. Using two-way ANOVA, we found that IL-5 induced significant time- and concentration-dependent increases in CD44 expression (P < 0.0001 and P < 0.02, respectively), with an approximate doubling of surface expression at 72 h with 10 ng/ml IL-5. Levels of CD11a also showed statistically significant time- dependent increases (P < 0.0008), but no cytokine-dependent changes were observed. Essentially identical results were obtained with IL-3 (n = 3, P < 0.02) and GM-CSF (n = 2) over a similar range of concentrations (data not shown). In contrast, separate experiments showed that culture of EOS for up to 72 h with 1 to 100 ng/ml IFN-, another cytokine known to prolong EOS survival and alter EOS phenotype (9, 27), did not produce any time-dependent effect on CD44 levels, despite maintenance of similar levels of viability (Figure 3 and data not shown).

View larger version (23K): [in this window] [in a new window]   Figure 2.   Kinetics and concentration dependence of the effects of IL-5 on EOS surface expression of CD11a and CD44. EOS were cultured in the indicated concentrations of IL-5 for up to 72 h, and expression of CD11a (upper panel) and CD44 (lower panel) was determined with flow cytometry. Values are plotted as average mean fluorescence intensity ± SEM for four experiments. By two-way ANOVA, significant time- and concentration-dependent IL-5 effects on CD44 expression were seen (P < 0.0001 and P < 0.02, respectively). Levels of CD11a also demonstrated statistically significant time-dependent increases (P < 0.0008), but these changes were independent of IL-5 concentrations.

View larger version (22K): [in this window] [in a new window]   Figure 3.   Culture of EOS with IFN-, unlike IL-5, did not induce time-dependent increases in surface expression of CD44. EOS were cultured with 10 ng/ml IL-5 or 100 ng/ml IFN- for up to 72 h, and expression of CD44 was determined with flow cytometry. Values are plotted as average mean fluorescence intensity ± SEM for three experiments. Lower concentrations of IFN- also failed to alter CD44 levels on EOS (data not shown).

Because IL-3, IL-5, and GM-CSF cause EOS in culture to become hypodense (3), we determined whether CD44 expression was greater on freshly isolated hypodense EOS than on normodense peripheral blood EOS from several donors. EOS from peripheral blood were fractionated with a discontinuous Percoll gradient (specific gravity: 1.085/1.090), and contaminating neutrophils were distinguished from EOS by differences in expression of CD9 or CD16, as determined with dual-color flow cytometry. Figure 4 shows that levels of CD44 were from 3- to 8-fold higher in hypodense EOS, whereas levels of CD49d (4 integrin) and CD69 were similar in both hypodense and normodense EOS.

View larger version (22K): [in this window] [in a new window]   Figure 4.   Hypodense EOS express higher levels of CD44 than do normodense EOS. EOS were isolated by density-gradient centrifugation, and expression of CD44, CD69, and CD49d was determined with flow cytometry (n = 3 to 7).

To test the functional consequences of altered CD44 expression on EOS, we examined their ability to adhere to a known CD44 ligand, hyaluronic acid (24). Freshly isolated EOS, EOS cultured for 72 h with IL-5 and confirmed by flow cytometry to have upregulated CD44 expression (data not shown), and KU812 cells were tested for their ability to adhere to various concentrations of immobilized hyaluronic acid. Modest levels of binding to the BSA blocking protein was seen (which were completely inhibited by mAb to CD18, data not shown). No specific binding to hyaluronate was observed; indeed, greater amounts of immobilized hyaluronic acid had the apparent effect of inhibiting adhesion (presumably because of a smaller amount of BSA on the plates [28]), and CD44 mAb had no effect on adhesion (Figure 5 and data not shown). In contrast, concentration-dependent attachment of KU812 cells was observed, and their adhesion was completely inhibited with an mAb to CD44 (data not shown).

View larger version (28K): [in this window] [in a new window]   Figure 5.   Hyaluronic acid is a ligand for KU812 cells but not for freshly isolated or IL-5-cultured EOS. Freshly isolated EOS, EOS cultured for 72 h with 10 ng/ml IL-5 to upregulate CD44 expression, and KU812 cells were tested for their ability to adhere to various concentrations of immobilized hyaluronic acid. Values are plotted as means ± SEM for three experiments.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The goal of the present study was to utilize mAb reagents from the Fifth Leukocyte Typing Workshop to test the hypothesis that EOS surface phenotype would be qualitatively and/or quantitatively altered after activation in vitro or in vivo. Several types of EOS were examined, including those collected from late-phase BALF, hypodense EOS, and normodense EOS before and after culture with cytokines. In screening for markers expressed de novo on activated cells, CD69 was the only marker identified. This confirms several previous reports (13, 26). The function of CD69 on EOS and other cell types is not entirely clear, although engagement of this marker leads to EOS apoptosis in vitro (29), whereas in other cells, cytokine release is observed (30). Whether CD69 modulates eosinophilic inflammation in vivo is unknown.

In screening for markers whose constitutive expression is increased on activated EOS, CD44 was unique among a panel of 16 markers studied in that it displayed enhanced expression on hypodense EOS and EOS activated in vitro by certain cytokines. We were unable to confirm previous reports that CD40, CD54, and CD81 might also represent EOS activation markers (12, 31). Whether EOS in BALF express a similar or distinct phenotypic pattern of upregulated markers has been suggested (32), but was not examined in the present study. CD44 is a proteoglycan expressed on all leukocytes, and exists in multiple isoforms on other types of cells (34). CD44 functions in cell trafficking by mediating rolling and firm adhesion to matrix constituents, including hyaluronic acid (24, 35). Previous studies have shown that cell-surface levels of CD44 can be modulated both in vitro and in vivo. For example, monocyte expression of certain CD44 isoforms is increased during culture (36), and levels of CD44 were found to be increased on leukocytes in patients with rheumatoid arthritis (37). Shedding of CD44 through proteolytic pathways has also been observed (38, 39). In the present studies, exposure of EOS to IL-3, IL-5, or GM-CSF led to increased CD44 levels in a concentration- and time-dependent manner. These findings are similar to those obtained when B cells are cultured with IL-5 (40). This adds CD44 to the list of EOS cell-surface markers altered by exposure to these cytokines (e.g., CD11b and CD18, which increase, and L-selectin which decreases [32, 41]). Although exposure of EOS to IFN- results in similar levels of viability, and can induce expression of other surface markers such as CD16, HLA-DR, and intercellular adhesion molecule-1 (ICAM-1) (10, 11), it had no effect on EOS CD44 expression. This is in contrast to the ability of IFN- to enhance CD44 on other cell types, such as monocytes (34). Interestingly, levels of CD44 were 3- to 8-fold higher on hypodense than on normodense EOS. Thus, levels of CD44 on hypodense EOS appear to be greater than those produced by cytokine exposure in vitro, suggesting that additional factors, besides or in addition to these cytokines, may be involved in regulating EOS CD44 expression in vivo.

Surprisingly, levels of CD11a slowly but significantly increased in culture. These effects were independent of IL-5 concentration, and therefore appear to have been the result of cell culture. In fact, a similar trend in CD44 expression was seen in the absence of IL-5 (Figure 5). The reason for this culture-induced increase in expression is unclear. Similar changes in CD11a and CD44 were seen with IL-3 and GM-CSF, and background fluorescence measured with isotype-matched irrelevant control mAb did not change (data not shown). Therefore, the observed increases were not due to enhanced autofluorescence or nonspecific mAb binding. Of note, however, was the lack of an increase in CD44 when EOS were cultured with IFN- (Figure 3), suggesting that cytokines present in the medium can affect time-dependent changes in CD44 expression.

Little is known about the function of CD44 on human EOS. In one study, CD44 expression on developing CD34⁺ EOS progenitors first diminished and then increased during 5 wk of culture, and enhanced proliferation was seen if these precursors were cultured on hyaluronic acid (42). This in vitro eosinophilopoiesis was dependent on the presence of IL-3 and IL-5 in the cultures, and was blocked by mAb to CD44. Interactions of T cells with airway smooth-muscle cells via CD44 have been reported, and it has been hypothesized that leukocyte-smooth-muscle interactions occurring via CD44 may play a role in the pathophysiology of lung diseases such as asthma (43). However, despite increased expression of CD44, we were unable to detect CD44-dependent attachment to hyaluronic acid of normodense, hypodense (data not shown), or cytokine-cultured EOS. This may be due to the fact that CD44 on most leukocytes is heavily glycosylated and cannot bind hyaluronic acid unless it is deglycosylated with enzymes such as neuraminadase (44, 45). Therefore, additional studies are required to determine the biologic importance of altered CD44 expression on activated EOS.