Modulation of Eosinophil Chemotaxis by Interleukin-5 R. A. J. Warringa, R. C. Schweizer, T. Maikoe, P. H. M. Kuijper, P. L. B. Bruijnzeel, and L. Koenderman Department of Pulmonary Diseases, University Hospital Utrecht, Utrecht, The Netherlands, and Swiss Institute of Allergy and Asthma Research, Davos, Switzerland
Eosinophilia and eosinophil function are regulated by cytokines such as granulocyte/macrophage colonystimulating factor (GM-CSF), interleukin-3 (IL-3), and interleukin-5 (IL-5). We have investigated the modulatory role of IL-5 on N-formyl-methionyl-Ieucyl-phenylalanine (FMLP), neutrophil-activating factor (NAF/IL-8), platelet factor 4 (PF4), and cytokine-induced chemotaxis of eosinophils from normal individuals. These eosinophils show a small chemotactic response toward PF4 but not to NAF/IL-8 and FMLP. Preincubation of eosinophils with low concentrations of IL-5 caused significantly increased responses toward PF4 and induced a significant chemotactic response toward FMLP and NAF/IL-8. In marked contrast, IL-5 (or IL-3) priming of eosinophils from normal donors resulted in a strong inhibition of GM-CSF-induced chemotaxis. A similar decrease in the chemotactic response toward GM-CSF was observed in eosinophils derived from allergic asthmatic individuals. This finding suggests that the latter eosinophils may have had a prior exposure to IL-5 (or IL-3). Washing of the cells after priming did not abrogate the inhibition of the GM-CSF response. Our data indicate that at low concentrations IL-5 is an important modulator of eosinophil chemotaxis, causing selective upregulation or downregulation of chemotactic responses toward different agents.
Recently, attention has been focused on the role of cytokines in the recruitment and subsequent activation of eosinophils in the lung during the late-phase asthmatic reaction (for a review, see reference 1). The reason for. this interest is based on data that indicate the importance of T lymphocytes in the pathogenesis of bronchial asthma (2). These T cells infiltrate the lung tissue, and ultrastructural studies on bronchial biopsies of subjects with mild asthma have shown that these lymphocytes are activated (IL-2R +) and almost exclusively express the CD4 + phenotype (helper/inducer subset) (3). Demonstration of mRNA for interleukin-5 (IL-5) in bronchial mucosal biopsies from patients with allergic asthma indicate that part of these infiltrating T cells exhibit a Th,like phenotype (4, 5). This is further supported by the demonstration of elevated levels of IL-5 and, to a lesser extent, IL-3 and granulocyte/macrophage colony-stimulating factor (GM-CSF) in the blood of allergic asthmatic individuals (6). Many in vitro studies have shown that cytokines derived
(Received in original form March 2, 1992 and in revised form July 1, 1992) Address correspondence to: L. Koenderman, Ph.D., Department of Pulmonary Diseases, University Hospital Utrecht, P.o. Box 85500, 3500 GA Utrecht, The Netherlands. Abbreviations: N-formyl-methionyl-leucyl-phenylalanine, FMLP; granulocyte/macrophage colony-stimulating factor, GM-CSF; human serum albumin, HSA; interleukin, IL; neutrophil-activating factor, NAF/IL-8; platelet factor 4, PF4; recombinant human, rho Am. J. Respir. Cell Mol. BioI. Vol. 7. pp. 631-636, 1992
from T cells of the Th, phenotype have profound effects on eosinophil differentiation and function. GM-CSF, IL-3, and IL-5 are important growth and differentiation factors for eosinophils and prolong survival of eosinophils in vitro (7-10). Moreover, these cytokines can potentiate functional responses of eosinophils ("priming") such as enhancement of respiratory burst activation, antibody-dependent cytotoxicity, and migratory responsiveness (10-16). In contrast to eosinophils and B-cell subpopulations, other inflammatory cells lack receptors for IL-5 (17, 18). Therefore, many effects of IL-5 are eosinophil specific, These include terminal differentiation of eosinophil precursors and enhancement of eosinophil adhesion to endothelial cells (19, 20). Recent data indicate that besides the Th, cell, the mast cell may act as a storage pool for cytokines (21-23). These mast cell-derived cytokines include both the endothelialactivating cytokines IL-l, IL-4, and tumor necrosis factor-a and the eosinophil-"priming" cytokines GM-CSF, IL-3, and IL-5. Release of these cytokines is probably mediated via an antigen-mediated IgE activation (21, 22). This combination of upregulation of adhesion molecules on endothelial cells and eosinophil preactivation may cause recruitment of eosinophils. At present, little is known about mediators involved in eosinophil adhesion and subsequent migration. The contribution of platelet-activating factor and IL-5 in eosinophil adhesion to endothelial cells and migration is well established (13,20,24-27). However, other mediators are of potential interest and may participate in inflammation. In this study, we have focused on neutrophil-activating factor (NAF/IL-8) and the structurally related compound platelet
632
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 7 1992
factor 4 (PF4). NAF/IL-8 is produced and released into the circulation by activated endothelial and epithelial cells (28, 29). PF4 is released by platelets that are activated by allergen challenge in individuals with allergic asthma (30). Previously, we have shown that eosinophils from healthy individuals require cytokine priming to be responsive to NAF/IL-8 and N-formyl-methionyl-Ieucyl-phenylalanine (FMLP) (15, 16). Interestingly, eosinophils isolated from the peripheral blood of allergic asthmatic individuals exhibit a phenotype comparable to cytokine-primed cells in vitro (16). In this study, we examined the effects of IL-5 on eosinophil migration toward FMLP, NAF/IL-8, and PF4. In addition, interaction between GM-CSF, IL-3, and IL-5 in cytokine-induced eosinophil migration was studied in detail.
method makes use of the fact that, in marked contrast to neutrophils, eosinophils lack the epitope on FcRm recognized by the monoclonal antibody CLB-FcR-gran 1 directed against CD16 (33). As a result, highly purified eosinophils can be isolated by removing neutrophils coated with FcRgran 1 with immunomagnetic dynabeads. Briefly, neutrophils were coated with a monoclonal antibody against CD16 (CLB FcR-gran 1, 2 j-tg/107 cells/ml) during 30 min at o-c Hereafter, the cells were washed twice and subsequently coincubated head over head with beads (Dynal Beads; Dynal, Oslo, Norway) at a ratio of 1:2 (cells/beads) for 20 min at 4°C. Neutrophils were subsequently removed by a magnetic particle concentrator (MPCTH-l; Dynal). Eosinophil purity was always> 95 %, and the viability was 98 %.
Materials and Methods
Chemotactic Assay Eosinophil chemotaxis was measured by a modification of the method of Boyden (34) using a 48-well microchemotaxis chamber (Neuroprobe, Cabin John, MD). Chemotaxins or Hepes buffer (30 j-tl) were added to the lower compartments. Two filters (cellulose nitrate) were placed between lower and upper compartments. The lower filter had a pore width of 0.45 j-tm (type HA; Millipore Corp., Bedford, MA), and the upper filter a pore width of8 j-tm (SM 113, thickness 100 j-tm; Sartorius AG, Gottingen, Germany). Before use, the filters were soaked in Hepes buffer. Purified eosinophils preincubated with different cytokines or with Hepes buffer were placed in the upper compartments (25 j-tl of 5 X 106 cells/ml). Concentration ranges of the chemotactic agents (diluted in Hepes) FMLP, NAF, PF4, and cytokines were tested using Hepes buffer as a control. Eosinophils isolated from normal individuals were preincubated for 30 min at 37°C with cytokines (for optimal time for GM-CSF and IL-3 priming, see reference 15), before analysis of chemotactic activity was carried out. Pilot experiments showed that this preincubation time was also optimal for IL-5 priming. The chemotaxis chambers were subsequently incubated for 2.5 h at 3rc. Hereafter, the upper filters were removed, fixed in butanol/ethanol (20/80%, vol/vol) for 10 min, and stained with Weigert solution (composition 1% [vol/vol] hematoxylin in ethanol mixed with a 70 mM acidic FeCl 3 solution at a 1:1 ratio). The filters were dehydrated with ethanol, made transparent with xylene, and fixed upside down. The number of cells per 10 hpf was determined by light microscopy (magnification: X400). In this way, the number of cells that passed the upper filter (and migrated 100 j-tm) was determined.
Patients All allergic asthmatic individuals participating in this study met the American Thoracic Society's definition of asthma, including episodic airway obstruction and hyperresponsiveness to bronchoprovocation with histamine. Seven men and six women between 20 and 45 yr of age participated in the study. All patients were hypersensitive to house dust mite or cat allergens, had positive skin test reactions to these allergens, and a bronchial hyperreactivity for histamine (PC 20 hist ~ 4 mg/ml). In all patients, anti-asthma medication was stopped 10 days before the study except for {J-sympathomimetics (salbutamol inhaler), which were stopped 24 h before patients entered the study. At the start of the study, peripheral blood eosinophil counts ranged from 2 to 16%. All subjects gave their informed consent, and the study was approved by the hospital's ethics committee. Reagents PF4 and FMLP were purchased from Sigma Chemical Co. (St. Louis, MO). Ficoll-paque and Percoll were obtained from Pharmacia (Uppsala, Sweden). All other materials were reagent grade. All experiments were carried out in Hepes buffer, which contained 132 mM NaCl, 6.0 mM KCI, 1.0 mM csci, 1.0 mM MgS04 , 1.2 mM potassium phosphate, 20 mM Hepes, heparin (10 IU/ml), 5 mM glucose, and 1.0% human serum albumin (HSA) (wt/vol), pH 7.4. Cytokines Recombinant human GM-CSF (2.5 X 108 U/mg) and IL-3 (108 U/mg) were purchased from Genzyme (Boston, MA). Recombinant human IL-5 (106 U/mg) was purchased from Amersham (Buckinghamshire, UK). NAF (NAP-l/IL-8; 10 j-tg/ml) was from British Biotechnology (Oxford, UK). Stock solutions (10-7 M) of the cytokines were prepared in phosphate-buffered salt solution supplemented with 0.5% purified HSA and were stored at -70°C until use.
Statistical Analysis Statistical analysis of data was performed using the Complete Statistical System (CSS) by Statsoft (Tulsa, OK). Student's t test for paired or unpaired observations was used, accepting P < 0.05 as significant.
Cell Isolation Blood was obtained from the buffy coat of 500 ml (normal donors) or 100 ml whole blood (allergic asthmatic individuals). Mixed granulocytes were isolated as described before (31). Eosinophils were subsequently isolated by the method described by Hansel and colleagues (32). This isolation
Effect of IL-5 Preincubation on Chemotactic Responses of Eosinophils toward FMLP and NAF/IL-8 Previously, we found that the preincubation of eosinophils derived from normal donors with picomolar concentrations
Results
Warringa, Schweizer, Maikoe et al.: Effects of IL-5 on Eosinophil Chemotaxis
of GM-CSF or IL-3 is essential for the induction of a chemotactic response toward the well-known neutrophil chemoattractants FMLP and NAF/IL-8 (15). Here, we evaluated whether IL-5, a cytokine selectively acting on eosinophils, had the same potency. Figure IA shows the chemotactic response toward 10 nM FMLP for eosinophils from normal donors in the presence or absence of different concentrations of recombinant human (rh) IL-5. In these ex-
120
LL 0-
J:
o
90
,.-
.....
(J)
...J
I
60
0-
o Z Ci5
ow
30
BUFFER C
13
12
11
10
9
633
periments, human eosinophils were preincubated for 30 min at 37°C. Thereafter, the cells were transferred to the upper compartment of a Boyden chamber and chemotactic responses were measured. Eosinophils did not show a significant chemotactic response toward FMLP unless they were preincubated with picomolar concentrations (10pM to 1 nM) ofIL-5. Optimal priming ofIL-5 was observed at 100 pM. The magnitude of the optimal chemotactic response to FMLP ofIL-5-primed eosinophils was in the same range as the one measured for optimal PAF-induced eosinophil chemotaxis (16, 24-26). Figure IB shows that, as observed for FMLP, NAF/IL-8 does not induce a chemotactic response. However, preincubating eosinophils with rhIL-5 (optimal concentration, 1 nM) caused a dose-dependent induction of chemotaxis toward NAF/IL-8. Effect of IL-S Preincubation on Chemotactic Responses of Eosinophils toward PF4 Next to NAF/IL-8, we tested the possibility that the NAFrelated peptide PF4 had chemotactic activity for human eosinophils. Figure 2 (lower line) showed this compound to be a weak chemoattractant for eosinophils from normal donors, even at relatively high concentrations (0.1 J!M). The PF4induced chemotactic response was very low compared with PAF responses (15). However, a considerably enhanced response (upper line) was observed when eosinophils were preincubated with an optimal priming concentration ofIL-5.
-LOG [IL-5] (M) 120
LL 0-
J:
o ,... ..... (J) :::! ::I:
140
LL 0-
90
::I:
o
.....
en ...J
60
I
0-
oZ
?;
ow
70
0-
o (J)
105
,.-
(J)
30
ow
O'----'---'-----L..--'--...........-
BUFFER C
13
12
11
.........- - - L
10
9
-LOG [IL-5] (M) Figure 1. Effect of concentration ranges of rhIL-S on the optimal FMLP (10 nM) (panel A) and NAF/IL-8 (10 nM) (panel B) induced chemotactic response of eosinophils from normal individuals. The eosinophils were preincubated for 30 min at 3rC with Hepes buffer or dose ranges of IL-S before the chemotaxin was added. C = chemotactic response of eosinophils to the applied chemotaxin after preincubation in buffer for 30 min at 37°C. After a period of 2.S h, filters were collected and stained (see MATERIALS AND METHODS). The chemotactic response is expressed as the number of cells/lO hpf (magnification: x400). Mean values ± SEM of six individual experiments are presented. * P < O.OS; values are considered to differ significantly from the control value (C). The eosinophil purity in these experiments was 98 ± 1%.
35
0'---"---"----'----'-----'----'BUFFER 11 10 9 8 7
-LOG [PF 4] (M) Figure 2. Dose-response curve of PF4-induced chemotaxis of eosinophils isolated from normal individuals. Eosinophil chemotaxis toward different concentrations of PF4 (10 pM to 100 nM) was determined in the presence (solid circles) or absence (open circles) of rhIL-S (I nM). Hepes buffer was used as a control. The eosinophils were preincubated for 30 min at 37°C with IL-S or Hepes buffer before transfer to the Boyden chamber. Chemotaxis is expressed as the number of cells/lO hpf. Mean values ± SEM of eight experiments are presented. Eosinophil purities were 97 ± 2 %. Values indicated with * differed significantly from buffer control values; values indicated with ** differed significantly from the chemotactic response of unprimed eosinophils (P < O.oS, paired Student's t test).
634
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 7 1992
Furthermore, a marked increase in the sensitivity for PF4 was found after priming. The chemotactic response was already present at 10 pM of PF4, reaching plateau levels at 100 pM. The magnitude of this last response was similar to that measured for (optimal) PAF-induced eosinophil chemotaxis (15). Effect of IL-5 Preincubation on GM-CSF-induced Eosinophil Chemotaxis Previous studies showed high concentrations of the cytokines GM-CSF, IL-3, and IL-5 to be chemotactic for eosinophils (13-16). Recently, we compared this cytokine-induced chemotaxis of eosinophils from normal donors with that of eosinophils from allergic asthmatic individuals. Interestingly, the latter eosinophils showed a decrease of ::::80 % in their chemotactic response toward 10 nM GM-CSF, whereas their chemotactic responses toward IL-3 and IL-5 were unchanged (16). This observation and the fact that blood levels of IL-5 are elevated in the circulation of allergic asthmatics (6) prompted us to investigate the nature of this phenomenon. Eosinophils from normal donors were preincubated with optimal priming concentrations (see above) of GM-CSF, IL-3, or IL-5. Subsequently, eosinophil chemotaxis was measured toward these cytokines. Figure 3A shows the optimal chemotactic responses toward GM-CSF, IL-3, and IL-5 of eosinophils from normal donors preincubated with buffer or cytokines, and eosinophils from allergic asthmatic donors. Pretreatment with 100 pM GM-CSF hardly influenced the chemotactic response toward 10 nM of GM-CSF, whereas preincubation with 1 nM IL-5 or 100 pM IL-3 caused decreases of 65 and 69 %, respectively. This low chemotactic response toward GM-CSF is similar to the optimal GM-CSF response observed in eosinophils from allergic asthmatic patients. The optimal chemotactic responses toward IL-3 and IL-5 also seemed slightly decreased by preincubating the eosinophils with either GM-CSF, IL-3, or IL-5, but this was not statistically significant (not shown). To evaluate whether this modulation phenomenon was irreversible, experiments were repeated with eosinophils preincubated with GM-CSF, IL-3, or IL-5 for 30 min before removal of excess cytokines by washing cells with buffer (Figure 3B). The results of these experiments indicated that the inhibition of GM-CSF-induced chemotaxis was still present, although less pronounced.
Discussion Asthma is characterized by an influx of mostly eosinophils in and around the bronchioli (35-37). These infiltrated eosinophils are in an activated state as monitored by EG2+ expression, and many of these eosinophils are found to be degranulated (38). Because of parallels between eosinophil influx, the occurrence of bronchial hyperresponsiveness, prolonged bronchoconstriction, and many other symptoms associated with bronchial asthma, it is now assumed that eosinophil-derived products are in part responsible for asthmatic symptoms (35). The processes involved in eosinophil influx are largely unknown, but phenomena such as adhesion, diapedesis, and migration are important. These processes are known to be well regulated, and several additional events such as priming of inflammatory cells contribute to
200
LL
0J:
... .....
150
0
en ...J J:
100
0-
0
Z
en
0
50
W
0 Buffer
N
P
A
B
c
GM-CSF (10nM) 160
B
LL
0J:
...
120
0
Eosinophilia and eosinophil function are regulated by cytokines such as granulocyte/macrophage colonystimulating factor (GM-CSF), interleukin-3 (IL-3), and interleukin-5 (IL-5). We have investigated the modulatory role of IL-5 on N-formyl-methionyl-Ieucyl-phenylalanine (FMLP), neutrophil-activating factor (NAF/IL-8), platelet factor 4 (PF4), and cytokine-induced chemotaxis of eosinophils from normal individuals. These eosinophils show a small chemotactic response toward PF4 but not to NAF/IL-8 and FMLP. Preincubation of eosinophils with low concentrations of IL-5 caused significantly increased responses toward PF4 and induced a significant chemotactic response toward FMLP and NAF/IL-8. In marked contrast, IL-5 (or IL-3) priming of eosinophils from normal donors resulted in a strong inhibition of GM-CSF-induced chemotaxis. A similar decrease in the chemotactic response toward GM-CSF was observed in eosinophils derived from allergic asthmatic individuals. This finding suggests that the latter eosinophils may have had a prior exposure to IL-5 (or IL-3). Washing of the cells after priming did not abrogate the inhibition of the GM-CSF response. Our data indicate that at low concentrations IL-5 is an important modulator of eosinophil chemotaxis, causing selective upregulation or downregulation of chemotactic responses toward different agents.
Recently, attention has been focused on the role of cytokines in the recruitment and subsequent activation of eosinophils in the lung during the late-phase asthmatic reaction (for a review, see reference 1). The reason for. this interest is based on data that indicate the importance of T lymphocytes in the pathogenesis of bronchial asthma (2). These T cells infiltrate the lung tissue, and ultrastructural studies on bronchial biopsies of subjects with mild asthma have shown that these lymphocytes are activated (IL-2R +) and almost exclusively express the CD4 + phenotype (helper/inducer subset) (3). Demonstration of mRNA for interleukin-5 (IL-5) in bronchial mucosal biopsies from patients with allergic asthma indicate that part of these infiltrating T cells exhibit a Th,like phenotype (4, 5). This is further supported by the demonstration of elevated levels of IL-5 and, to a lesser extent, IL-3 and granulocyte/macrophage colony-stimulating factor (GM-CSF) in the blood of allergic asthmatic individuals (6). Many in vitro studies have shown that cytokines derived
(Received in original form March 2, 1992 and in revised form July 1, 1992) Address correspondence to: L. Koenderman, Ph.D., Department of Pulmonary Diseases, University Hospital Utrecht, P.o. Box 85500, 3500 GA Utrecht, The Netherlands. Abbreviations: N-formyl-methionyl-leucyl-phenylalanine, FMLP; granulocyte/macrophage colony-stimulating factor, GM-CSF; human serum albumin, HSA; interleukin, IL; neutrophil-activating factor, NAF/IL-8; platelet factor 4, PF4; recombinant human, rho Am. J. Respir. Cell Mol. BioI. Vol. 7. pp. 631-636, 1992
from T cells of the Th, phenotype have profound effects on eosinophil differentiation and function. GM-CSF, IL-3, and IL-5 are important growth and differentiation factors for eosinophils and prolong survival of eosinophils in vitro (7-10). Moreover, these cytokines can potentiate functional responses of eosinophils ("priming") such as enhancement of respiratory burst activation, antibody-dependent cytotoxicity, and migratory responsiveness (10-16). In contrast to eosinophils and B-cell subpopulations, other inflammatory cells lack receptors for IL-5 (17, 18). Therefore, many effects of IL-5 are eosinophil specific, These include terminal differentiation of eosinophil precursors and enhancement of eosinophil adhesion to endothelial cells (19, 20). Recent data indicate that besides the Th, cell, the mast cell may act as a storage pool for cytokines (21-23). These mast cell-derived cytokines include both the endothelialactivating cytokines IL-l, IL-4, and tumor necrosis factor-a and the eosinophil-"priming" cytokines GM-CSF, IL-3, and IL-5. Release of these cytokines is probably mediated via an antigen-mediated IgE activation (21, 22). This combination of upregulation of adhesion molecules on endothelial cells and eosinophil preactivation may cause recruitment of eosinophils. At present, little is known about mediators involved in eosinophil adhesion and subsequent migration. The contribution of platelet-activating factor and IL-5 in eosinophil adhesion to endothelial cells and migration is well established (13,20,24-27). However, other mediators are of potential interest and may participate in inflammation. In this study, we have focused on neutrophil-activating factor (NAF/IL-8) and the structurally related compound platelet
632
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 7 1992
factor 4 (PF4). NAF/IL-8 is produced and released into the circulation by activated endothelial and epithelial cells (28, 29). PF4 is released by platelets that are activated by allergen challenge in individuals with allergic asthma (30). Previously, we have shown that eosinophils from healthy individuals require cytokine priming to be responsive to NAF/IL-8 and N-formyl-methionyl-Ieucyl-phenylalanine (FMLP) (15, 16). Interestingly, eosinophils isolated from the peripheral blood of allergic asthmatic individuals exhibit a phenotype comparable to cytokine-primed cells in vitro (16). In this study, we examined the effects of IL-5 on eosinophil migration toward FMLP, NAF/IL-8, and PF4. In addition, interaction between GM-CSF, IL-3, and IL-5 in cytokine-induced eosinophil migration was studied in detail.
method makes use of the fact that, in marked contrast to neutrophils, eosinophils lack the epitope on FcRm recognized by the monoclonal antibody CLB-FcR-gran 1 directed against CD16 (33). As a result, highly purified eosinophils can be isolated by removing neutrophils coated with FcRgran 1 with immunomagnetic dynabeads. Briefly, neutrophils were coated with a monoclonal antibody against CD16 (CLB FcR-gran 1, 2 j-tg/107 cells/ml) during 30 min at o-c Hereafter, the cells were washed twice and subsequently coincubated head over head with beads (Dynal Beads; Dynal, Oslo, Norway) at a ratio of 1:2 (cells/beads) for 20 min at 4°C. Neutrophils were subsequently removed by a magnetic particle concentrator (MPCTH-l; Dynal). Eosinophil purity was always> 95 %, and the viability was 98 %.
Materials and Methods
Chemotactic Assay Eosinophil chemotaxis was measured by a modification of the method of Boyden (34) using a 48-well microchemotaxis chamber (Neuroprobe, Cabin John, MD). Chemotaxins or Hepes buffer (30 j-tl) were added to the lower compartments. Two filters (cellulose nitrate) were placed between lower and upper compartments. The lower filter had a pore width of 0.45 j-tm (type HA; Millipore Corp., Bedford, MA), and the upper filter a pore width of8 j-tm (SM 113, thickness 100 j-tm; Sartorius AG, Gottingen, Germany). Before use, the filters were soaked in Hepes buffer. Purified eosinophils preincubated with different cytokines or with Hepes buffer were placed in the upper compartments (25 j-tl of 5 X 106 cells/ml). Concentration ranges of the chemotactic agents (diluted in Hepes) FMLP, NAF, PF4, and cytokines were tested using Hepes buffer as a control. Eosinophils isolated from normal individuals were preincubated for 30 min at 37°C with cytokines (for optimal time for GM-CSF and IL-3 priming, see reference 15), before analysis of chemotactic activity was carried out. Pilot experiments showed that this preincubation time was also optimal for IL-5 priming. The chemotaxis chambers were subsequently incubated for 2.5 h at 3rc. Hereafter, the upper filters were removed, fixed in butanol/ethanol (20/80%, vol/vol) for 10 min, and stained with Weigert solution (composition 1% [vol/vol] hematoxylin in ethanol mixed with a 70 mM acidic FeCl 3 solution at a 1:1 ratio). The filters were dehydrated with ethanol, made transparent with xylene, and fixed upside down. The number of cells per 10 hpf was determined by light microscopy (magnification: X400). In this way, the number of cells that passed the upper filter (and migrated 100 j-tm) was determined.
Patients All allergic asthmatic individuals participating in this study met the American Thoracic Society's definition of asthma, including episodic airway obstruction and hyperresponsiveness to bronchoprovocation with histamine. Seven men and six women between 20 and 45 yr of age participated in the study. All patients were hypersensitive to house dust mite or cat allergens, had positive skin test reactions to these allergens, and a bronchial hyperreactivity for histamine (PC 20 hist ~ 4 mg/ml). In all patients, anti-asthma medication was stopped 10 days before the study except for {J-sympathomimetics (salbutamol inhaler), which were stopped 24 h before patients entered the study. At the start of the study, peripheral blood eosinophil counts ranged from 2 to 16%. All subjects gave their informed consent, and the study was approved by the hospital's ethics committee. Reagents PF4 and FMLP were purchased from Sigma Chemical Co. (St. Louis, MO). Ficoll-paque and Percoll were obtained from Pharmacia (Uppsala, Sweden). All other materials were reagent grade. All experiments were carried out in Hepes buffer, which contained 132 mM NaCl, 6.0 mM KCI, 1.0 mM csci, 1.0 mM MgS04 , 1.2 mM potassium phosphate, 20 mM Hepes, heparin (10 IU/ml), 5 mM glucose, and 1.0% human serum albumin (HSA) (wt/vol), pH 7.4. Cytokines Recombinant human GM-CSF (2.5 X 108 U/mg) and IL-3 (108 U/mg) were purchased from Genzyme (Boston, MA). Recombinant human IL-5 (106 U/mg) was purchased from Amersham (Buckinghamshire, UK). NAF (NAP-l/IL-8; 10 j-tg/ml) was from British Biotechnology (Oxford, UK). Stock solutions (10-7 M) of the cytokines were prepared in phosphate-buffered salt solution supplemented with 0.5% purified HSA and were stored at -70°C until use.
Statistical Analysis Statistical analysis of data was performed using the Complete Statistical System (CSS) by Statsoft (Tulsa, OK). Student's t test for paired or unpaired observations was used, accepting P < 0.05 as significant.
Cell Isolation Blood was obtained from the buffy coat of 500 ml (normal donors) or 100 ml whole blood (allergic asthmatic individuals). Mixed granulocytes were isolated as described before (31). Eosinophils were subsequently isolated by the method described by Hansel and colleagues (32). This isolation
Effect of IL-5 Preincubation on Chemotactic Responses of Eosinophils toward FMLP and NAF/IL-8 Previously, we found that the preincubation of eosinophils derived from normal donors with picomolar concentrations
Results
Warringa, Schweizer, Maikoe et al.: Effects of IL-5 on Eosinophil Chemotaxis
of GM-CSF or IL-3 is essential for the induction of a chemotactic response toward the well-known neutrophil chemoattractants FMLP and NAF/IL-8 (15). Here, we evaluated whether IL-5, a cytokine selectively acting on eosinophils, had the same potency. Figure IA shows the chemotactic response toward 10 nM FMLP for eosinophils from normal donors in the presence or absence of different concentrations of recombinant human (rh) IL-5. In these ex-
120
LL 0-
J:
o
90
,.-
.....
(J)
...J
I
60
0-
o Z Ci5
ow
30
BUFFER C
13
12
11
10
9
633
periments, human eosinophils were preincubated for 30 min at 37°C. Thereafter, the cells were transferred to the upper compartment of a Boyden chamber and chemotactic responses were measured. Eosinophils did not show a significant chemotactic response toward FMLP unless they were preincubated with picomolar concentrations (10pM to 1 nM) ofIL-5. Optimal priming ofIL-5 was observed at 100 pM. The magnitude of the optimal chemotactic response to FMLP ofIL-5-primed eosinophils was in the same range as the one measured for optimal PAF-induced eosinophil chemotaxis (16, 24-26). Figure IB shows that, as observed for FMLP, NAF/IL-8 does not induce a chemotactic response. However, preincubating eosinophils with rhIL-5 (optimal concentration, 1 nM) caused a dose-dependent induction of chemotaxis toward NAF/IL-8. Effect of IL-S Preincubation on Chemotactic Responses of Eosinophils toward PF4 Next to NAF/IL-8, we tested the possibility that the NAFrelated peptide PF4 had chemotactic activity for human eosinophils. Figure 2 (lower line) showed this compound to be a weak chemoattractant for eosinophils from normal donors, even at relatively high concentrations (0.1 J!M). The PF4induced chemotactic response was very low compared with PAF responses (15). However, a considerably enhanced response (upper line) was observed when eosinophils were preincubated with an optimal priming concentration ofIL-5.
-LOG [IL-5] (M) 120
LL 0-
J:
o ,... ..... (J) :::! ::I:
140
LL 0-
90
::I:
o
.....
en ...J
60
I
0-
oZ
?;
ow
70
0-
o (J)
105
,.-
(J)
30
ow
O'----'---'-----L..--'--...........-
BUFFER C
13
12
11
.........- - - L
10
9
-LOG [IL-5] (M) Figure 1. Effect of concentration ranges of rhIL-S on the optimal FMLP (10 nM) (panel A) and NAF/IL-8 (10 nM) (panel B) induced chemotactic response of eosinophils from normal individuals. The eosinophils were preincubated for 30 min at 3rC with Hepes buffer or dose ranges of IL-S before the chemotaxin was added. C = chemotactic response of eosinophils to the applied chemotaxin after preincubation in buffer for 30 min at 37°C. After a period of 2.S h, filters were collected and stained (see MATERIALS AND METHODS). The chemotactic response is expressed as the number of cells/lO hpf (magnification: x400). Mean values ± SEM of six individual experiments are presented. * P < O.OS; values are considered to differ significantly from the control value (C). The eosinophil purity in these experiments was 98 ± 1%.
35
0'---"---"----'----'-----'----'BUFFER 11 10 9 8 7
-LOG [PF 4] (M) Figure 2. Dose-response curve of PF4-induced chemotaxis of eosinophils isolated from normal individuals. Eosinophil chemotaxis toward different concentrations of PF4 (10 pM to 100 nM) was determined in the presence (solid circles) or absence (open circles) of rhIL-S (I nM). Hepes buffer was used as a control. The eosinophils were preincubated for 30 min at 37°C with IL-S or Hepes buffer before transfer to the Boyden chamber. Chemotaxis is expressed as the number of cells/lO hpf. Mean values ± SEM of eight experiments are presented. Eosinophil purities were 97 ± 2 %. Values indicated with * differed significantly from buffer control values; values indicated with ** differed significantly from the chemotactic response of unprimed eosinophils (P < O.oS, paired Student's t test).
634
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 7 1992
Furthermore, a marked increase in the sensitivity for PF4 was found after priming. The chemotactic response was already present at 10 pM of PF4, reaching plateau levels at 100 pM. The magnitude of this last response was similar to that measured for (optimal) PAF-induced eosinophil chemotaxis (15). Effect of IL-5 Preincubation on GM-CSF-induced Eosinophil Chemotaxis Previous studies showed high concentrations of the cytokines GM-CSF, IL-3, and IL-5 to be chemotactic for eosinophils (13-16). Recently, we compared this cytokine-induced chemotaxis of eosinophils from normal donors with that of eosinophils from allergic asthmatic individuals. Interestingly, the latter eosinophils showed a decrease of ::::80 % in their chemotactic response toward 10 nM GM-CSF, whereas their chemotactic responses toward IL-3 and IL-5 were unchanged (16). This observation and the fact that blood levels of IL-5 are elevated in the circulation of allergic asthmatics (6) prompted us to investigate the nature of this phenomenon. Eosinophils from normal donors were preincubated with optimal priming concentrations (see above) of GM-CSF, IL-3, or IL-5. Subsequently, eosinophil chemotaxis was measured toward these cytokines. Figure 3A shows the optimal chemotactic responses toward GM-CSF, IL-3, and IL-5 of eosinophils from normal donors preincubated with buffer or cytokines, and eosinophils from allergic asthmatic donors. Pretreatment with 100 pM GM-CSF hardly influenced the chemotactic response toward 10 nM of GM-CSF, whereas preincubation with 1 nM IL-5 or 100 pM IL-3 caused decreases of 65 and 69 %, respectively. This low chemotactic response toward GM-CSF is similar to the optimal GM-CSF response observed in eosinophils from allergic asthmatic patients. The optimal chemotactic responses toward IL-3 and IL-5 also seemed slightly decreased by preincubating the eosinophils with either GM-CSF, IL-3, or IL-5, but this was not statistically significant (not shown). To evaluate whether this modulation phenomenon was irreversible, experiments were repeated with eosinophils preincubated with GM-CSF, IL-3, or IL-5 for 30 min before removal of excess cytokines by washing cells with buffer (Figure 3B). The results of these experiments indicated that the inhibition of GM-CSF-induced chemotaxis was still present, although less pronounced.
Discussion Asthma is characterized by an influx of mostly eosinophils in and around the bronchioli (35-37). These infiltrated eosinophils are in an activated state as monitored by EG2+ expression, and many of these eosinophils are found to be degranulated (38). Because of parallels between eosinophil influx, the occurrence of bronchial hyperresponsiveness, prolonged bronchoconstriction, and many other symptoms associated with bronchial asthma, it is now assumed that eosinophil-derived products are in part responsible for asthmatic symptoms (35). The processes involved in eosinophil influx are largely unknown, but phenomena such as adhesion, diapedesis, and migration are important. These processes are known to be well regulated, and several additional events such as priming of inflammatory cells contribute to
200
LL
0J:
... .....
150
0
en ...J J:
100
0-
0
Z
en
0
50
W
0 Buffer
N
P
A
B
c
GM-CSF (10nM) 160
B
LL
0J:
...
120
0
