Abstract
Migration from blood to tissue modulates eosinophil function, possibly through interactions with endothelial cells.
The effects of contact with and migration through endothelial cells on eosinophil expression of surface markers and release of leukotriene C4 were evaluated.
A small proportion (2.6%) of eosinophils spontaneously migrated through endothelial cell monolayers. Activation of endothelial cells by interleukin (IL)-4 or IL-1β slightly increased this migration (to 12.4%), which became much greater when a chemoattractant was placed in the lower chamber (84.3%). However, the chemotactic effect was downregulated by pretreating endothelial cells with interferon gamma (IFN-γ; 63.1%). At baseline, 5% of eosinophils expressed CD69; this increased to 30.7% in culture on untreated endothelial cells and to 50.9% on IL-1β-pretreated endothelial cells. This effect was mediated through intercellular adhesion molecule-1/CD11b interaction. Eosinophil migration through endothelial cells further increased CD69 expression to 63.9% and also increased CD35 expression from 83.3 to 91.3%. Upon stimulation, eosinophils that had migrated through endothelial cells produced more leukotriene C4 than control cells (872.4 and 103.9 pg·mL−1, respectively). Endothelial cell pretreatment with IL-4 or IL-1β further increased leukotriene C4 release (1,789.1 and 2,895.1 pg·mL−1, respectively), whereas pretreatment with IFN-γ decreased it (293.7 pg·mL−1).
These data show that in vitro interactions with endothelial cells upregulate eosinophil membrane receptor expression and mediator release and that these effects are differently modulated by T-helper cell type 1 and 2 cytokines. These eosinophil modulations may play an important role in asthma pathogenesis.
This study was supported by the Canadian Institutes of Health Research, Ottawa, Canada (MOP-49534).
Eosinophils are recruited to tissue in various pathological conditions 1. In asthma, eosinophils infiltrate the bronchial mucosa 2 and their counts in blood, sputum and bronchial mucosa correlate with indices of disease severity 3–4. Blood eosinophils are thought to circulate in a quiescent state until they migrate into the tissue 5. In healthy mucosa, eosinophils account for a small proportion of the cells, making it almost impossible to evaluate their activation status. In eosinophilic diseases in which their numbers increase, tissue eosinophils show increased expression of complement and immunoglobulin (Ig)G receptors compared to their blood counterparts 1, 5. They also express intercellular adhesion molecule (ICAM)-1, human leukocyte antigen (HLA)-DR and CD69, which are barely or not detectable on blood eosinophils 6–9. Tissue eosinophils show increased release of superoxide anion and leukotriene (LT)C4 compared to blood eosinophils 10, 11. Lung lavage eosinophils, obtained after segmental bronchoprovocation, show an increase in membrane receptors, adhesion, survival and generation of superoxide anions, which was not achieved by incubation of blood eosinophils with interleukin (IL)-5 and granulocyte macrophage-colony stimulating factor (GM-CSF) 12–14. These observations suggest that passage into tissue activates eosinophils, increasing their pro-inflammatory potential. The factors involved in this phenomenon have yet to be fully determined.
During their recruitment to tissue eosinophils pass through the endothelium. This is a complex process 1, 14, 15 and interactions with endothelial cells have been shown to modulate certain eosinophil functions 16. Activation of endothelium isan important step in inflammatory processes and occurs indisease models, including those for allergic diseases 17. Endothelial cells exposed to cytokines such as IL-1β, IL-4 and tumour necrosis factor-α increase their expression of ICAM-1 and vascular cell adhesion molecule (VCAM)-1 18, 19, which serve as ligands for eosinophil integrins. Consequently, cytokines may modulate the transendothelial migration of eosinophils and amplify endothelial cell-induced changes in eosinophil functions 19, 20.
In order to further clarify the role of endothelial cells in themodulation of eosinophils during their migration to tissue, the effects of eosinophil migration through a human umbilical vein endothelial cell monolayer on expression of cell-surface markers and liberation of LTC4, a powerful bronchoconstrictor and pro-inflammatory mediator, were evaluated. It was found that expression of CD69, an early marker of activation, and of CD35, a receptor for complement proteins, and LTC4 release were modulated by endothelial cells and that treatment of endothelial cells with T-helper cell type 1 and 2 cytokines differently modified their effect on eosinophils.
Materials and methods
Reagents
Platelet-activating factor (PAF), activated complement fraction 5a (C5a) and bovine serum albumin (BSA) were purchased from Sigma Chemical Co. (St Louis, MO, USA). Kits for LTC4 determination and 5-oxo-6, 8, 11, 14-eicosatetraenoic acid (5-oxo-ETE) were obtained from Cayman Chemical (Ann Arbor, MI, USA). Purified monoclonal mouse antibody IgG1κ and IgG2b (isotypic control), purified human anti-CD4, anti-CD16, anti-CD28, anti-CD86, anti-CD35, anti-CD69 and anti-HLA-DR and phycoerythrin-conjugated polyclonal antimouse antibodies were purchased from BD Biosciences (Mississauga, ON, Canada). In specific experiments, fluorescein isothiocyanate (FITC)-conjugated anti-CD69 antibodies were used. Monoclonal mouse antihuman CD16 antibodies were purchased from Miltenyi Biotec (Auburn, CA, USA) and human recombinant IL-1β, IL-4 and interferon gamma (IFN-γ) from Peprotech, Inc. (Rocky Hill, NJ, USA). Monoclonal mouse antihuman CD11b antibody (clone bear 1) was obtained from Beckman Coulter (Mississauga, ON, Canada) and fibronectin was from Roche Diagnostics (Laval, QC, Canada). Anti-ICAM-1 antibody (clone HA58) was purchased from BD Biosciences. Human umbilical vein endothelial cells cryopreserved from asingle donor in primary culture were purchased from Clonetics (San Diego, CA, USA).
Subject selection
Seven normal subjects (two male/five female, median age 31 yrs (range 18–55 yrs)) without a history of allergy or asthma and 10 asthmatics (five male/five female, median age23.5 yrs (range 19–35 yrs)) meeting the criteria of the American Thoracic Society for the diagnosis of asthma were recruited for the study 21. The asthmatic subjects had a morning prebronchodilator forced expiratory volume in one second (FEV1) of >85% of the predicted value and required only a short-acting β2-agonist on demand on less than four occasions per week. The inclusion criteria included stable treatment for >3 months, no inhaled steroids over the 3 months preceding the study, no use of other drugs and no disease other than asthma. Approval for the study was obtained from the local ethics committee and all subjects signed an informed consent form. FEV1 and provocative concentration of methacholine causing a 20% fall in FEV1 (PC20) were measured in the morning, ≥8 h after any β2-agonist inhalation. Median FEV1 for normal and asthmatic subjects were 98.5 (91–110) and 96 (86–102)% pred, respectively (p=0.03), and geometric mean PC20 were 43.9±19.7 and 1.1±0.9 mg·mL−1, respectively (p=0.006, Wilcoxon rank-sum test). Subjects underwent blood sampling early in the morning. Mean blood eosinophil counts were 0.2×109 and 0.3×109cells·L−1 for normal and asthmatic subjects, respectively.
Blood eosinophil purification
Eosinophils were purified as previously described 22. They were separated from neutrophils by negative selection using a magnetic cell sorter. An aliquot of the cell suspension was used to determine cell number (haemocytometer) and viability (trypan blue exclusion; Sigma Chemical Co.) and differential cell counts (Diff-Quik; Dade Diagnostics, Aguada, PR, USA). The purity of the eosinophil preparations used in this study was >98% and the contaminating cells were neutrophils and/or lymphocytes. Eosinophil viability was always >99%.
Endothelial cell culture
Endothelial cells were grown in endothelial growth medium supplemented with human recombinant epidermal growth factor, human fibroblast growth factor, vascular endothelial growth factor, ascorbic acid, long-R3 insulin-like growth factor-I, heparin, hydrocortisone, gentamicin and amphotericin B (Clonetics) and 10% foetal bovine serum (FBS) (Invitrogen Canada, Burlington, ON, Canada) in a 5% carbon dioxide (CO2) atmosphere at 37°C. When ∼80% confluent, cells were harvested, resuspended in fresh culture medium and seeded at a density of 2×105 cells·500 µL−1 on cell culture inserts (12-mm diameter polycarbonate membrane with 3.0-µm pores; BD Biosciences Labware, Mississauga, ON, Canada). The inserts were placed in 24-well culture plates (BD Biosciences Labware), culture medium without FBS (500 µL) was placed in the lower chamber to inhibit formation of an endothelial cell bilayer 23 and the cells cultured for 5 days. Endothelial cell monolayers were confirmed as being at confluence by toluidine blue staining on control inserts. All experiments were carried out on passage 4 cells. In specific experiments, endothelial cells were cultured on the bottom of 24-well culture plates.
Incubation of eosinophils with endothelial cells and transendothelial migration
Confluent endothelial cell monolayers were treated with IFN-γ (1,000 U·mL−1) for 72 h, IL-4 (100 U·mL−1) for 24 h or IL-1β (50 U·mL−1) for 4 h in a 5% CO2 atmosphere at 37°C. These cytokines were added to the upper compartment of the wells. Times of incubation and cytokine concentrations were chosen based on the results of previous studies 24–27. Thereafter, the upper and lower compartments of the wells were washed three times with Hank's balanced salt solution (HBSS) (37°C). Eosinophils (1×106 cells·mL−1 in Roswell Park Memorial Institute 1640 containing 10% FBS and 1% penicillin/streptomycin) were laid in inserts coated or not with endothelial cells. A potent eosinophil chemoattractant, 5-oxo-ETE (1 µM), was added to the lower chamber of some wells to induce migration 22. After a 4-h incubation, eosinophils recovered from the upper chambers served to evaluate the effect of contact with endothelial cells and cells that had migrated through endothelial cell-coated inserts under the action of 5-oxo-ETE served to evaluate the effect of transendothelial migration. In these two sets of experiments, inserts without endothelial cells served as controls. In specific experiments, to further evaluate the effects of endothelial cell contact in the presence or absence of 5-oxo-ETE and to prevent migration through endothelial cells, eosinophils were incubated in 24-well culture plates coated or not with endothelial cells. Moreover, in order to evaluate the role of ICAM-1/CD11b ligation in eosinophil/endothelial cell interactions, an anti-ICAM-1 antibody (20 µg·mL−1) was added tothe incubation medium. The migration rate is presented asthe percentage of cells in the upper chamber that migrate into the lower chamber. At the end of incubations, the viability of the eosinophils was always >98%.
Incubation on fibronectin- or anti-CD11b antibody-coated plates
In order to study the effect of CD11b and very late activation antigen-4 (VLA-4) ligation on eosinophil CD69 expression, 96-well culture plates (Nunc-Immuno Plate Maxisorp Surface; VWR International, Montreal, QC, Canada) were incubated with either 100 µL fibronectin (20 µg·mL−1), 100 µL anti-CD11b monoclonal antibody (20 µg·mL−1), both fibronectin and anti-CD11b antibody, or 100 µL HBSS (control) overnight at 4°C. At the end of this incubation, the wells were washed twice with HBSS and nonspecific protein binding was blocked by addition of 100 µL HBSS containing 1% BSA for 1 h at 37°C. Before use, the coated wells were washed with HBSS and eosinophils (1×105 cells·100 µL−1) were added to wells and incubated for 4 h at 37°C. Cells were then harvested and stained with FITC-conjugated antihuman CD69 as described below.
Cell-surface marker detection
Eosinophils were labelled (2×105 cells·100 µL−1 for 30 min at 4°C) with specific antibodies (1 µg·mL−1) directed against CD4, CD16, CD28, CD35, CD69, CD86 and HLA-DR or with their respective isotypic control, as previously described 28, 29. After incubation, cells were washed with HBSS containing 1% BSA and fixed in 4% paraformaldehyde (10 min at4°C). Phycoerythrin-conjugated antimouse secondary antibody (1 µg·mL−1) was then added (for 30 min at 4°C). In experiments with incubation on fibronectin- or anti-CD11b antibody-coated plates, FITC-conjugated anti-CD69 antibodies were used. Finally, eosinophils were washed, suspended in HBSS containing 1% BSA, kept at 4°C and immediately subjected to cytometric analysis using an EPICS® ELITE ESP flow cytometer (Beckman-Coulter, Miami, FL, USA). Mean fluorescence (MF) was measured on a logarithmic scale 28, 29 and the percentage of cells expressing a particular marker determined by counting the number of cells showing greater fluorescence activity than ≥96% of negative controls.
Leukotriene C4 assay
Eosinophils (1×106 cells·mL−1 in HBSS/CaCl2 (1.6 mM)) were incubated with PAF (1 µM) for 10 min at 37°C under 5% CO2 and thereafter with C5a (10 nM) under the same conditions. Tubes were then put on ice for 5 min to stop the reaction. Cells were centrifuged and LTC4 measured in the supernatants by quantitative enzyme immunoassay according to the manufacturer's recommendations.
Statistical analyses
Mediator and cell-surface marker expression data were analysed using a crossed-nested design (analysis of variance). Data are presented as mean±sem and results were considered significant when p was <0.05. There was no difference between the eosinophils of normal and asthmatic subjects for all measured parameters under all tested conditions. Their data were, therefore, pooled together for analysis and presentation.
Results
Migration through endothelial cells
Spontaneous eosinophil migration through uncoated inserts, untreated endothelial cells and IFN-γ-pretreated endothelial cells was 1.4±0.2, 2.6±0.3 and 2.1±0.4%, respectively (fig. 1⇓). Pretreatment of endothelial cells with IL-4 or IL-1β increased eosinophil migration to 11.5±0.9 and 12.4±0.8%, respectively (p=0.0001). Addition of 5-oxo-ETE to the lower chambers dramatically stimulated eosinophil migration through endothelial cells (84.3±1.0%; p=0.0001). Pretreatment of endothelial cells with IFN-γ decreased the effect of 5-oxo-ETE on eosinophil migration (63.1±1.3%; p=0.0001). In contrast, pretreatment of endothelial cells with IL-4 or IL-1β did not modify 5-oxo-ETE-induced migration (82.8±1.3 and 81.8±2.0%, respectively).
Eosinophil surface marker expression
Among eosinophils incubated on uncoated inserts and recovered from the upper chambers, the percentage of CD69-positive cells was 5.0±2.2% (0.06±0.02MF) and increased to 30.7±8.4% (0.34±0.11MF) on inserts coated with untreated endothelial cells (p<0.0001) (fig. 2a⇓). Contact of eosinophils with endothelial cells in culture plates in the presence of 5-oxo-ETE induced similar CD69 expression (27.2±6.2%; 0.2±0.04MF) to incubation on endothelial cells alone. Pretreatment of endothelial cells with IFN-γ or IL-4 did not modify eosinophil expression of CD69 (36.3±7.6% (0.35±0.06MF) and 45.4±7.3% (0.43±0.05MF), respectively). However, pretreatment of endothelial cells with IL-1β increased CD69-positive eosinophil percentages (50.9±10.8%; 0.58±0.14MF) (p<0.0001). Compared to cells incubated on endothelial cells, eosinophils of the lower chambers that had migrated through endothelial cells under the action of 5-oxo-ETE showed a further increase in CD69 expression (63.9±7.0%; 0.88±0.18MF; p=0.004) (fig. 2b⇓). Migration of eosinophils through cytokine-pretreated endothelial cells did not further modify CD69 expression compared to migration across untreated endothelial cells.
The effect of CD11b and VLA-4 ligation on eosinophil CD69 expression is presented in figure 3⇓. Eosinophils incubated on fibronectin showed low expression of CD69 (6.2±1.5%), similar to control wells (4.2±1.4%). In contrast, cells incubated on anti-CD11b antibody-coated wells exhibited increased expression of CD69 (27±6%; p=0.007; n=6). Addition of fibronectin did not modify this expression (26.8±11.5%). These values were similar to those observed with eosinophils incubated on untreated endothelial cells (fig. 2a⇑). Furthermore, in two experiments eosinophils were incubated on endothelial cells in the presence of an anti-ICAM-1 antibody. This antibody decreased endothelial cell-induced CD69 expression from 36.6 to 2.4%.
The expression of CD35 in eosinophils incubated in uncoated inserts (83.3±4.5%; 1.2±0.1MF), inserts coated with endothelial cells (85.5±2.6%; 1.4±0.2MF) or on endothelial cells in the presence of 5-oxo-ETE (85.5±3.8%; 1.8±0.3MF) were similar (fig. 4a⇓). Treatment of endothelial cells with cytokines did not modify eosinophil CD35 expression: IFN-γ 87.9±1.1%, IL-4 90.1±1.4%, and IL-1β 89.5±2.3% (p=0.38). Migration ofeosinophils through endothelial cells under the action of 5-oxo-ETE significantly increased CD35 expression (91.3±3.2%; 1.9±0.3MF) compared to cells incubated in inserts coated with endothelial cells and that had not migrated (p=0.0013) (fig. 4b⇓). Cytokine pretreatment of endothelial cells did not modify the CD35 expression of migrated eosinophils.
Purified blood eosinophil expression of CD4, CD16, CD28, CD86 and HLA-DR were 15.0±3.7, 2.0±1.0, 3.5±1.3, 6.9±2.8 and 1.5±0.3%, respectively (n=4–8). Contact with and migration through endothelial cell monolayers did not modify eosinophil expression of these surface markers (data not shown).
Leukotriene C4 release
Compared to eosinophils recovered from the upper chambers, cells of lower chambers that migrated through endothelial cells under the action of 5-oxo-ETE released more than eight times more LTC4 (103.9±30.2 and 872.4±292.5 pg·mL−1, respectively; p=0.0055) (fig. 5⇓). Migration through IFN-γ-pretreated endothelial cells decreased eosinophil LTC4 release (293.7±116.9 pg·mL−1; p=0.0006). In contrast, migration through IL-4- and IL-1β-treated endothelial cells further increased eosinophil LTC4 release (1,789.1±457.3 and 2,895.1±831.4 pg·mL−1, respectively; p=0.0006). Eosinophils incubated in endothelial cell-coated and uncoated culture plates released similar amounts of LTC4 (32.4±12.7 and 82.9±28.2 pg·mL−1, respectively; p=0.2, n=6). Moreover, under these conditions, addition of 5-oxo-ETE did not amplify the PAF/C5a-induced LTC4 release (32.4±12.4; n=3).
Discussion
Eosinophil emigration from blood to tissue and migration through endothelium may contribute to the observed activation status of tissue eosinophils compared to their blood counterparts. This study shows that in vitro contact with endothelial cells increases eosinophil CD69 expression and that eosinophil migration through an endothelial cell monolayer further increases CD69 expression and also augments CD35 expression and LTC4 release. Moreover, the endothelial cell effects on eosinophils are modulated by cytokines, IL-1β or IL-4 amplifying them and IFN-γ decreasing them.
In the present study, endothelial cell pretreatment with IL-4 or IL-1β but not IFN-γ facilitated the passage of eosinophils across endothelial cell monolayers. Since IL-4 andIL-1β upregulate ICAM-1 and VCAM-1 on endothelial cells 30, 31, these ligands may increase adhesion of eosinophils to endothelial cells via Mac-1 (a β2 integrin) and VLA-4, respectively, and, consequently, may promote eosinophil migration.
5-oxo-ETE, a potent chemoattractant 32, promoted the migration of most of the eosinophils. IFN-γ partially inhibits 5-oxo-ETE-induced eosinophil migration. The mechanism of this inhibition remains undefined. IFN-γ has no known effect on the expression of adherence molecules such as VCAM-1 but increases the expression of ICAM-1 on airway epithelium 33.
Induction of CD69 expression on eosinophils has been reported as an indication of eosinophil activation 34. The function of CD69 on eosinophils and other cell types is not entirely understood, although engagement of this marker leads to eosinophil apoptosis in vitro and cytokine release in other cell types 35. CD69 may be a coreceptor for eosinophil activation, or, as for platelets and monocytes, facilitate mediator release or degranulation 9, 36, 37. CD69 is barely detectable on freshly isolated eosinophils but is rapidly expressed after stimulation with numerous cytokines 9, 38–40. CD69 is also present on bronchoalveolar lavage eosinophils 8, 9, 34. The present data and the recent report of Yamamoto et al. 16 show that CD69 expression is increased by contact with and further increased after migration through endothelial cells. The induction of CD69 expression is also observed on eosinophils incubated with antibodies directed against CD11b and inhibited by an anti-ICAM-1 antibody, suggesting that binding of eosinophils to endothelial cells via ICAM-1/β2 integrin interaction is probably involved in endothelial cell-mediated eosinophil activation. The greater expression of CD69 on migrated eosinophils is probably due to the effect of migration through the endothelial cell monolayer, given that CD69 expression on eosinophils cocultured with endothelial cells in the presence of 5-oxo-ETE was not increased compared to that on eosinophils in contact with endothelial cells. Yamamoto et al. 16 showed that migration across IL-1β-pretreated endothelial cells stimulated eosinophil expression of CD69, CD54 and HLA-DR. The present study failed to show upregulation of CD69 and HLA-DR expression on eosinophils that had migrated across IL-1β-treated endothelial cells compared to those that had migrated through unstimulated endothelial cells. One possible explanation forthis discrepancy is the different eosinophil populations studied. Yamamoto et al. 16 analysed a small population ofeosinophils (<15% of total eosinophils placed on the endothelial cell monolayer) that migrated spontaneously through IL-1β-pretreated endothelial cells, whereas the present study analysed eosinophils that migrated through endothelial cells under the action of 5-oxo-ETE (>80% of total eosinophils). Moreover, Yamamoto et al. 16 measured MF on a linear scale, which greatly increases the sensitivity but also significantly decreases the specificity of measurements.
In the present study, CD35 was highly expressed on blood eosinophils. As shown by Walker et al. 27, this expression was not modified by contact with, but increased slightly after migration through, endothelial cells. Complement receptors such as CD35 are known to play a role in the activation and adhesion of neutrophils and eosinophils, and, during asthmatic reaction, expression of these receptors is elevated on circulating granulocytes 41. In contrast, other receptors, CD4, CD16, CD28, CD86 and HLA-DR, were barely expressed by blood eosinophils and their expression was not upregulated by migration through endothelial cells. These receptors are involved in interactions of eosinophils with mediators and other cells and are upregulated on tissue eosinophils 42. The present results suggest that eosinophil migration through endothelial cells is not involved in this upregulation.
The present study shows that passage of eosinophils through endothelial cells modulates LTC4 release. Munoz et al. 43 reported that incubation on IL-1β-treated endothelial cells but not endothelial cells alone augmented LTC4 release from stimulated eosinophils and that this augmentation was related to the process of eosinophil binding at the endothelial surface 43. Herein, the effect of incubation on cytokine-treated endothelial cells was not evaluated, but it was shown that migration of eosinophils through IL-1β- or IL-4-treated endothelial cells further increased LTC4 release. It is likely that part of this increase is mediated via upregulation of adhesion molecule expression on endothelial cells 30, 31. The inhibitory effect of endothelial cell pretreatment with IFN-γ is important (fig. 5⇑) and could also be mediated via modulation of adhesion molecule expression on endothelial cells. Moreover, the amount of LTC4 produced by eosinophils in vivo depends on its level of activation and eosinophil exposure to cytokines and mediators 10. Thus, during migration, exposure to cytokine(s)/mediator(s) expressed by endothelial cells might be involved in the modulation of LTC4 release observed after migration of eosinophils through endothelial cells. Previous reports showed that IL-1 and IL-4 increased GM-CSF production by endothelial cells, whereas IFN-γ decreased it 44–46.
Since blood eosinophils from asthmatic subjects are activated compared to normal eosinophils 1–5, the present authors postulated that asthmatic eosinophils might respond differently to migration through an endothelial cell monolayer. However, no difference was observed between normal and asthmatic blood eosinophils for the measured parameters. This may be related to the clinical state of the patients, given that the subjects recruited in the present study had very mild asthma. Blood eosinophils from subjects with more severe asthma might respond differently to interactions with endothelial cells.
In conclusion, the present study investigated many cell surface markers and release of leukotriene C4, a potent mediator. The data suggest that, in vivo, endothelial cells modulate specific eosinophil cell-surface marker expression and function, thus modifying eosinophil phenotype, and that these effects are themselves modulated by various cytokines and, at least in part, mediated via adhesion molecules. Therefore, evaluation of the activation status of eosinophils following their passage from blood to tissue and the increase in their capacity to interact with cells and extracellular matrix components is a very important step in understanding the role of these cells in tissue homeostasis.
Acknowledgments
The authors thank L. Trépanier for evaluating and recruiting the volunteers and J. Milot and F. Deschenes for blood sampling.
- Received November 7, 2002.
- Accepted January 21, 2003.
- © ERS Journals Ltd