A parallel signal-transduction pathway for eotaxin- and interleukin-5-induced eosinophil shape change

Abstract

Interleukin-5 (IL-5) and eotaxin are the most important cytokines/chemokines responsible for regulating eosinophil locomotion and are known to play a co-operative role in the selective recruitment of eosinophils to inflamed tissues. Following exposure to chemoattractants, eosinophils undergo a series of events, including reorganization of actin filaments and subsequent rapid shape changes, culminating in chemotaxis. In this study we examined the signalling pathways for eosinophil shape change regulated by eotaxin and IL-5, primarily using a gated autofluorescence/forward-scatter assay. Eotaxin and IL-5 were able to elicit shape change with peaks at 10 and 60 min, respectively, and IL-5 triggered the shape change more efficiently than eotaxin. The pharmacological inhibitors of mitogen-activated protein kinase (MAP kinase) and p38 blocked both eotaxin- and IL-5-induced eosinophil shape change in a dose-dependent manner. In addition, depletion of intracellular Ca²⁺ and inhibition of protein kinase A (PKA) strongly reduced eosinophil shape change. In contrast, even when used at high concentrations, protein tyrosine kinase (PTK) inhibitors caused only a slight reduction in the ability to change shape. However, treatment with protein kinase C (PKC) inhibitors, such as GF109203X and staurosporine, resulted in a striking inhibition of eosinophil shape change by IL-5, but not eotaxin. Data from the inhibition of activation and chemotaxis of the extracellular signal-regulated kinases (ERK1/2) by the PKC inhibitors were also consistent with findings from the experiments on shape change. Collectively, two eosinophil-selective cytokines/chemokines probably regulate eosinophil shape change via a largely overlapping signalling pathway, with involvement of PKC restricted to the IL-5 signal alone.

Introduction

Allergic inflammatory disease, such as asthma, is often characteristically featured by infiltrated eosinophils in the airways, and they are thus thought to be key mediators of the allergic inflammation.¹ An array of evidence from clinical and experimental observations demonstrates a strong correlation among the number of eosinophils, damage of the airways and the development of airways hyper-reactivity.^2,3 As a small number of eosinophils are present in the circulation, a mechanism must be operating whereby selective accumulation of eosinophils is achieved in the tissues and blood of patients with eosinophilic inflammatory diseases. The chemotaxis of eosinophils involves a series of events, including an increase of transient calcium mobilization, activation of a variety of signalling enzymes, actin rearrangement, shape change and diapedesis.⁴ All these steps are so closely linked to one another that they cannot properly be considered in isolation.

It is becoming increasingly evident that eotaxin and interleukin (IL)-5 play, in concert, a major role in the selective accumulation of eosinophils.^5–7 Eotaxin selectively binds CC chemokine receptor 3 (CCR3), which is abundantly expressed on eosinophils, and thus acts as a principal chemokine for mediating eosinophil chemotaxis.^8–10 In addition, eotaxin activates eosinophil functions in vitro, including CD11b up-regulation, increase of reactive oxygen production, actin reorganization and Ca²⁺ mobilization, in a pertussis toxin (PTX)-sensitive manner.¹¹ Eotaxin-bound CCR3 transduces signals via activation of multiple signalling proteins, such as mitogen-activated protein kinases (MAP kinases), extracellular signal-regulated kinases (ERK1/2) and p38,^12,13 protein tyrosine kinases (PTK)¹⁴ and protein kinase C (PKC).¹⁵ On the other hand, IL-5 is the most prominent cytokine associated with the regulation of eosinophil functions. IL-5 induces differentiation,^16,17 chemokinesis,¹⁸ prolongation of survival,^19,20 respiratory burst,²¹ degranulation³ and priming of human eosinophils.^22–24 Furthermore, IL-5 enhances the level of blood eosinophils by facilitating the mobilization of eosinophils from the bone marrow pool into blood.5–^7,25,26 IL-5 binding to its receptor has recently been shown to activate phosphatidylinositol 3-kinase (PI 3-kinase), MAP kinases and PTK, including Lyn, Janus kinase 2 (Jak2), Fyn and Syk.^18,27–33 The activation of MAP kinases is initiated by a variety of signalling enzymes, such as PI 3-kinase, tyrosine kinases, PKC and G proteins, and acts both as an end-point, integrating the upstream signals, and as a starting point, to exert subsequent eosinophil functions. Therefore, the activation of MAP kinases is essential for eosinophil functions in response to chemoattractants, including eotaxin and IL-5, as well as small lipid mediators such as platelet-activating factor (PAF).

In a variety of inflammatory conditions, both IL-5 and eotaxin exist at the same time in the microenvironment.^5,34 Given the fact that these two cytokines/chemokines play co-operative roles in the regulation of eosinophil locomotion in vivo, it seems probable that an intracellular signal-transduction pathway exists to integrate the activation of these two different classes of receptors. In this study we examined a signal-transduction pathway for eosinophil shape change stimulated by eotaxin and IL-5, primarily using a method – a so-called recently developed gated autofluorescence/forward scatter (GAFS) assay – which is suitable for the measurement of eosinophil activation and determination of chemokine receptor usage.³⁵ Our results demonstrate that the eosinophil shape change induced by these two cytokines exhibits similar patterns in response to an intracellular Ca²⁺ chelator and a number of pharmacological inhibitors, such as phospholipase C (PLC), protein kinase A (PKA), PTK and MAP kinases. Although the shape change occurs through a largely overlapping signal-transduction pathway, PKC appears significantly to affect IL-5-, but not eotaxin-, induced phosphorylation of ERK1/2 and chemotaxis, as well as shape change.

Materials and methods

RPMI-1640, fetal bovine serum (FBS), phosphate-buffered saline (PBS), human serum albumin and Hanks' balanced salt solution (HBSS) were purchased from Gibco BRL (Grand Island, NY). Reagents for eosinophil preparation were purchased from Sigma (St. Louis, MO). Monoclonal anti-human CCR-3 antibody was a kind gift from Dr Koichi Hirai (Tokyo)³⁶ and anti-IL-5Rα antibody, recombinant human eotaxin and IL-5 were obtained from R & D Systems (Minneapolis, MN). AG-490, an inhibitor of Jak2 kinase, was obtained from RBI (Natick, MA). We purchased MEK (MAP ERK kinase) inhibitors, including PD098059, from RBI and U0126 from Tocris (Ballwin, MO). The p38 MAPK inhibitor, SB203580, the PTK inhibitor, genistein, and the PKC inhibitor, GF109203X, were obtained from Tocris. The PKA inhibitor, H-89, PTX and 1,2-bis(o-amino-phenyl)ethane-N,N,N′,N′-tetraacetic acid tetraacetoxymethyl ester (BAPTA/AM) were supplied from Biomol Research Laboratories, Inc. (Plymouth Meeting, PA). The PTK inhibitor, herbimycin A, and the PKC inhibitor, staurosporine, were purchased from Sigma. Diff-Quik solution was obtained from American Scientific Products (McGaw Park, IL).

Preparation of polymorphonuclear leucocytes and eosinophils

Blood donors included individuals who were both atopic and non-atopic, with eosinophils comprising between 10 and 20% of their peripheral blood leucocytes. Polymorphonuclear leucocytes (PMNL) were purified from the heparinized peripheral blood of volunteers. Briefly, after erythrocytes were sedimented in 6% dextran-dextrose in 0·1 m EDTA, the leucocyte-rich cell suspension was centrifuged at 300 g for 10 min at 4° and then washed with cold PBS. The leucocyte-rich suspension was then overlaid on 2 ml of a discontinuous Percoll gradient (specific gravity; 1·070–1·115) in [1,4- piperazinebis (ethane sulphonic acid)] (Pipes) buffer (25 nm Pipes, 110 mm NaCl, 5 mm KCl, 40 mm NaOH and 5·4 mm glucose) in 140 × 10 mm polystyrene tubes, and centrifuged at 1600 g for 8 min at 4°. A PMNL layer was removed from the Percoll gradient, and resuspended in HBSS containing 0·5% human serum albumin. Cytospin slides were stained with Diff-Quik to determine eosinophil purity and number. Eosinophils typically accounted for 60–90% of total PMNL, as determined by microscopic examination. For the assay of MAP kinase activation and chemotaxis, eosinophils were further purified by CD16-negative selection using a magnetic cell separator (MACS system; Becton-Dickinson, Mountain View, CA), as previously described.³⁷ The purity of eosinophils was ≈ 98%, as determined on the light microscopic examination of cytocentrifuge slides prepared by Diff-Quik staining.

Measurement of eosinophil shape change using flow cytometry

To determine eosinophil shape change we employed the GAFS assay, as previously described.³⁵ The purified PMNL were washed with tissue culture medium (TCM), RPMI-1640 containing 10% heat-inactivated FBS, and resuspended at a concentration of 2 × 10⁶ cells/ml. Aliquots of cells (2 × 10⁵) were treated with or without chemokines/cytokines in a final volume of 100 µl. The tubes were incubated at 37° in a CO₂ incubator for the indicated time intervals. After incubation, 500 µl of ice-cold 4% paraformaldehyde was added to terminate the reactions. The fixed cells were then analysed using a FACScan flow cytometer (Becton-Dickinson). The forward scatter was not affected by fixation, regardless of eosinophil stimulation. Setting of the FL-2 fluorescence channel (585 nm) allowed eosinophils to be distinguished from neutrophils, as eosinophils display higher autofluorescence than neutrophils.

Inhibition of eosinophil shape change by various inhibitors of signal-transduction pathways

We examined the effects of inhibitors for a variety of signalling molecules on eotaxin- or IL-5-driven eosinophil shape change using the GAFS assay. PMNL were preincubated at 37° for 30 min, with or without inhibitors, in a volume of 100 µl, and then stimulated with eotaxin (30 ng/ml) and IL-5 (10 ng/ml) for an additional 10 min and 60 min, respectively. The same procedure was used to confirm the specificity of eotaxin and IL-5 by adding the neutralizing antibodies to CCR3 and IL-5Rα, respectively. We also examined the effect of intracellular Ca²⁺ chelation on eosinophil shape change. Cells in Ca²⁺/Mg²⁺-free HBSS were pretreated with BAPTA/AM, or medium, at 25° for 30 min and were then exposed to eotaxin or IL-5. Following incubation, all samples were fixed in 4% paraformaldehyde at 4° for 5 min, washed with Ca²⁺/Mg²⁺-free PBS, and analysed by flow cytometry. Eosinophil shape change was expressed in two ways. When eotaxin and IL-5-induced eosinophil shape changes were compared with the unstimulated eosinophil shape change, a percentage of change in the forward scatter (FSC) was used to estimate the extent of increase in the shape change from unstimulated cells. In this case, the percentage was calculated as follows:

When inhibitors and antibodies were used to block the shape change, the relative FSC was calculated as follows:

Effects of the inhibitors alone on shape change were minimal, and their relative shape changes were compared with medium alone (assumed 100%): 99·0% for PTX (1 µg/ml); 100·9% for PD98059 (30 µm); 102·4% for U0126 (1 µm); 108·0% for SB203580 (1 µm); 99·7% for U73122 (1 µm); 95·9% for H89 (10 µm); 97·2% for herbimycin A (10 µm); 100·9% for genistein (10 µm); 102·5% for staurosporine (100 nm); 104·7% for GF109203X (10 µm); and 99·4% for BAPTA/AM (200 µm).

Immunoblot of phosphorylated ERK 1/2

Measurement of ERK1/2 phosphorylation was performed exactly according to the protocol previously described, except that an anti-phospho-ERK1/2 antibody was used.³³ The freshly purified eosinophils were suspended at a density of 2 × 10⁶ cells in 1 ml of HBSS containing 0·5% bovine serum albumin (BSA), pretreated with or without inhibitors for 30 min at 37°, and then exposed to eotaxin, IL-5 or control buffer at 37° for the indicated time intervals. The cells were viable after treatment with the inhibitors, as determined by Trypan Blue exclusion. The cell suspension was washed twice with ice-cold Buffer A (20 mm Tris, pH 7·4; 137 mm NaCl; 1 mm EDTA; 0·1 mm sodium orthovanadate; 10 mm sodium fluoride; 10 mmβ-glycerophosphate; 10 µg/ml aprotinin; and 10 µg/ml leupeptin) by centrifugation at 400 g for 8 min. The resulting cell pellets were permeabilized with Buffer B [1% Triton-X-100, 0·25% deoxycholate and 0·1% sodium dodecyl sulphate (SDS) in Buffer A] and incubated for 15 min on ice. The cell lysate was prepared from the supernatant by microcentrifugation at 10 000 g for 15 min. The protein in the supernatant was mixed with sample buffer, run on a Tris–Tricine gel (9·8%), and electrophoretically blotted to polyvinylidene difluoride (PVDF) membrane (Millipore, Bedford, MA). The membranes was blocked with 5% non-fat dry milk in TBST (10 mm Tris–HCl, pH 8·0; 150 mm NaCl; 0·05% Tween-20) and probed with a 1 : 1000 dilution of anti-phospho-p44/42 MAP kinase antibody (Cell Signalling Technology, Beverly, MA). Equal loading of proteins was confirmed by reprobing anti-p44/42 MAP kinase antibody (1 : 1000 dilution) (Cell Signalling Technology). After washing three times with TBST for 15 min, the membranes were incubated with an anti-rabbit immunoglobulin G (IgG), horseradish peroxidase (HRP)-linked antibody (1 : 5000 dilution) (Cell Signalling Technology), and developed by enhanced chemiluminescence (ECL) (Amersham, Pharmacia Biotech, Piscataway, NJ).

Chemotaxis

Chemotaxis assays were carried out in a 48-well Boyden chamber (Neuroprobe, Cabin John, MD). Eosinophils were resuspended at 2 × 10⁶ cells/ml in RPMI-1640 containing 10% FBS. A 27-µl aliquot of the indicated concentration of eotaxin or IL-5 was added to the lower chamber. A 5-µm pore size PVPF membrane (Osmonics, Minnetonka, MN) was overlaid, and then 50 µl of the eosinophil suspension that had been preincubated with medium or inhibitors for 30 min in a CO₂ incubator was added to the upper chambers. The lower chambers contained medium, eotaxin (100 ng/ml), or IL-5 (10 ng/ml). The chambers were subsequently incubated for 3 hr at 37°. The filter was then removed and the migrating cells in the lower chamber were counted. Chemotaxis was expressed as percentage migration, as follows:

Statistical analysis

The Wilcoxon signed rank test or the Student's t-test was applied to the measurement of eosinophil shape change induced by eotaxin and IL-5 in the presence or absence of inhibitors and antibodies. The difference was considered significant when the P-value was < 0·05. The results were expressed as mean ± standard error of the mean (SEM).

Results

Shape change of eosinophils by eotaxin and IL-5

Unstained human eosinophils are originally described to exhibit bright autofluorescence with excitation maxima at 370 nm and 450 nm and maximum emission at 520 nm,³⁸ and the shape change of a particular leucocyte can be readily evaluated by measuring change in FSC upon exposure to agonists. Sorting eosinophils by specific autofluorescence (FL2) and monitoring shape change by FSC, in combination, allow quantitative determination of eosinophil shape change by flow cytometry without separation of eosinophils in PMNL. This technique – a GAFS assay – has recently been used to measure the eosinophil activation and shape change induced by agonistic agents.^35,39 The GAFS assay was used throughout this study to measure eosinophil shape change. The FL2 autofluorescence histogram of unstimulated PMNLs was largely divided into two parts. The cells with a high autofluorescence on the fluorescence histogram were strongly immunoreactive to anti-CCR3 antibody (data not shown) and were therefore assigned as eosinophils, as demonstrated in a previous report.³⁵ Treatment of PMNL with eotaxin and IL-5 led to a selective increase in the FSC of high-autofluorescent cells, with little change observed in the FSC of low-autofluorescent cells (data not shown). Eotaxin and IL-5 were able to induce eosinophil shape change in both dose- and time-dependent manners (Fig. 1a,b). Eotaxin exhibited more rapid kinetics (with a peak at 5–10 min) than IL-5 (which had a maximum response at 30–60 min, depending on the donor) (Fig. 1a and data not shown). Hence, the subsequent experiments were carried out with 10- and 60-min incubations for eotaxin and IL-5, respectively. Although both eotaxin (30 ng/ml) and IL-5 (10 ng/ml) were able to induce a maximum eosinophil shape change, the shape-change response was consistently vigorous from IL-5-stimulated eosinophils of any donor (n > 20, P < 0·0001) (Fig. 1c). In addition, pretreatment with IL-5 for 60 min induced eosinophils to become more responsive to subsequent treatment with eotaxin, leading to an increase in FSC by 30–40% compared with IL-5 alone (data not shown). The priming effect of IL-5 has been previously observed in the synthesis of leukotriene C4 and chemotaxis mediated by chemoattractant, C^22,33,40 and in the eotaxin-induced chemotaxis and accumulation of eosinophils in the airways compartment.⁴¹ Preincubation of PMNL in TCM or Ca²⁺/Mg²⁺-free PBS immediately after purification did not affect the eotaxin-induced eosinophil shape change, whereas the use of TCM or Ca²⁺/Mg²⁺-containing PBS slightly augmented the shape change by IL-5 compared with Ca²⁺/Mg²⁺-free PBS (data not shown). To confirm the specificity of the eosinophil shape change by eotaxin and IL-5, PMNL were preincubated with a blocking anti-CCR3 mAb or anti-IL-5Rα mAb for 30 min, and treated with eotaxin or IL-5. The shape-change responses to eotaxin were almost completely blocked by 0·1 µg/ml of anti-CCR3 mAb (Fig. 2a). Similarly, the shape-change response to IL-5 was partly inhibited by anti-IL-5Rα mAb (Fig. 2b), indicating that eotaxin and IL-5 induce the shape change by binding their cognate receptors. The eotaxin-induced eosinophil shape change was almost completely blocked by a concentration of PTX as low as 0·1 µg/ml, while the inhibitor increased IL-5-induced eosinophil shape change (but this was not statistically significant), demonstrating that IL-5-induced eosinophil shape change occurs via a pathway that is independent of PTX-sensitive G protein (Fig. 2c). Both cytokine/chemokine effects were strongly abrogated by 0·5 µg/ml cytochalasin B (data not shown).

Figure 1.

Eosinophil shape change induced by eotaxin and interleukin (IL)-5. A total of 2 × 10⁵ polymorphonuclear leucocytes (PMNL) were incubated with either (a) eotaxin (0.3, 3 or 30 ng/ml) or (b) IL-5 (0.1, 1.0 or 10 ng/ml) for the indicated periods of time, and their shape changes were measured by flow cytometry. The percentage change in forward scatter (FSC) represents the FSC from stimulated eosinophils ÷ FSC from unstimulated eosinophils × 100 (n = 4, *P < 0.05, **P < 0.005, ***P < 0.0001). (c) The eosinophil shape changes induced by the two stimuli were compared (n > 20, ***P < 0.0001).

Figure 2.

Determination of the specificity of eosinophil shape change induced by eotaxin and interleukin-5 (IL-5). A total of 2 × 10⁵ polymorphonuclear leucocytes (PMNL) were preincubated with increasing concentrations of anti-CC chemokine receptor 3 (anti-CCR3) antibody (a), anti-IL-5Rα antibody (b) or pertussis toxin (PTX) (c) at 37° for 30 min. They were then stimulated with eotaxin (30 ng/ml) for 10 min or with IL-5 (10 ng/ml) for 60 min. The shape change was expressed as relative forward scatter (FSC) in which eotaxin- or IL-5-induced shape change in the absence of blocking antibody or PTX was assumed to be 100%. Data shown here represent the average of three independent experiments. *P < 0.05.

The effect of inhibitors of various signal transducers on eosinophil shape change

In order to examine the signalling pathways for eosinophil shape change by these two chemotactic cytokines, a number of inhibitors were used to block the activities of their specific signal transducers. PMNL were preincubated with an inhibitor for 30 min, and then stimulated with eotaxin (30 ng/ml) for 10 min or with IL-5 (10 ng/ml) for 60 min. The MEK inhibitors, PD98059 and U0126, inhibited eotaxin- or IL-5-induced eosinophil shape change in a dose-dependent manner. A similar pattern was observed with a p38 inhibitor, SB203580 (Fig. 3a,b). The extent of inhibition of the shape change by these inhibitors was also in a range comparable between IL-5- and eotaxin-stimulated eosinophils. Collectively, these results suggest the involvement of MAP kinases in both eotaxin- and IL-5-induced eosinophil shape change. A 20-µm concentration of BAPTA/AM, an intracellular Ca²⁺ chelator, was equally effective in inhibiting both eotaxin- and IL-5-mediated responses by 55–60% (n = 4, P ≤ 0·01) (Fig. 4a). Treatment with a PLC inhibitor, U73122, also resulted in a decrease of the shape change (Fig. 4b). A PKA inhibitor, H89, at 10 µm considerably blocked the eotaxin- and IL-5-induced shape changes, but the shape changes were refractory at 1 µm (Fig. 4c). PTK inhibitors such as herbimycin A (Fig. 5a) and genistein (Fig. 5b) caused a partial reduction of eotaxin- and IL-5-induced eosinophil shape change only at their highest concentrations. At 10 µm, herbimycin A or genistein, the shape change induced by eotaxin and IL-5 was inhibited by 40 ± 8% and 30 ± 5%, respectively (n = 3–4, P ≤ 0·05). In contrast, the treatment with Jak2 inhibitor, AG-490, did not inhibit the eosinophil responses (Fig. 5c). Thus, signalling pathways for eotaxin- and IL-5-mediated eosinophil shape changes appear largely to overlap. However, an unexpected result came from treatment with PKC inhibitors. GF109203X and staurosporine, both of which have a strong affinity for a number of PKC isozymes, inhibited the IL-5-induced shape change of eosinophils by 41·9 ± 6·4% (P < 0·05) and 46·3 ± 5·1% (P < 0·05) at 10 µm each. In contrast, these PKC inhibitors had little or no effect on the eotaxin-induced shape change (Fig. 6a,b).

Figure 3.

The effects of mitogen-activated protein kinase (MAP kinase) inhibitors on eotaxin- and IL-5-induced eosinophil shape change. Polymorphonuclear leucocytes (PMNL) were pretreated with different concentrations of PD98059 (3, 30 or 100 µm), U0126 (0.1, 1.0 or 10 µm) or SB203580 (10, 50 or 100 µm) at 37° for 30 min, then stimulated with eotaxin (30 ng/ml) for 10 min (a) or with interleukin-5 (IL-5) (10 ng/ml) for 60 min (b). Data shown here represent the average of at least three independent experiments. *P < 0.05, **P < 0.005.

Figure 4.

The effects of phospholipase C (PLC) and protein kinase A (PKA) inhibitors and depletion of intracellular Ca²⁺ on eotaxin- and interleukin-5 (IL-5)-induced eosinophil shape change. Polymorphonuclear leucocytes (PMNL) were pretreated with different concentrations of 1,2-bis(o-amino-phenyl)ethane-N,N,N′,N′-tetraacetic acid tetraacetoxymethyl ester (BAPTA/AM) (2, 20 or 200 µm), U-73122 (0.1, 1.0 or 10 µm) or H-89 (0.1, 1.0 or 10 µm) at 37° for 30 min, and then stimulated with eotaxin (30 ng/ml) for 10 min (a) or with IL-5 (10 ng/ml) for 60 min (b). Data shown represent the average of at least three independent experiments. *P < 0.01, **P < 0.005.

Figure 5.

The effects of tyrosine kinase inhibitors on eotaxin- and interleukin-5 (IL-5)-induced eosinophil shape change. Polymorphonuclear leucocytes (PMNL) were pretreated with the indicated concentrations of herbimycin A (a), genistein (b), or AG490 (c) at 37° for 30 min, then stimulated with eotaxin (30 ng/ml) for 10 min or with IL-5 (10 ng/ml) for 60 min. Data shown represent the average of at least three independent experiments. *P < 0.05.

Figure 6.

The effects of protein kinase C (PKC) inhibitors on eotaxin- and interleukin-5 (IL-5)-induced eosinophil shape change. Polymorphonuclear leucocytes (PMNL) were pretreated with the indicated concentrations of GF109293X (a) or staurosporine (b) at 37° for 30 min, and stimulated with eotaxin (30 ng/ml) for 10 min or with IL-5 (10 ng/ml) for 60 min. Data shown here represent the average of at least three independent experiments. *P < 0.05.

Blockage of phosphorylation of ERK1/2 by PKC and MEK inhibitors

To examine whether the differential responses to the PKC inhibitors would also be true for the phosphorylation of ERKs stimulated by eotaxin and IL-5, immunoblot analysis was carried out. The phosphorylation of ERK1/2 was previously shown to correlate with their activities in eosinophils.^13,14 The purified, unstimulated eosinophils exhibited a basal level of phosphorylated ERK2 (p42) with no phosphorylated ERK1 (p44). Upon stimulation with eotaxin or IL-5, the phosphorylation of ERK1 and ERK2 was induced and augmented, respectively. The peaks of ERK1/2 phosphorylation were observed at 1–5 min for eotaxin and at 5–15 min for IL-5, depending upon the donor. The phosphorylation of ERK1 by eotaxin and IL-5 was completely blocked by 50 µm PD98059 and 10 µm U1026, and ERK2 phosphorylation returned to basal levels following treatment with these inhibitors (Fig. 7a). There have been conflicting results regarding which ERK is the dominant phosphorylated form activated in eosinophils. It is possible that the form of phosphorylated ERKs observed in activated eosinophils is largely dependent upon which anti-phospho-ERK antibodies are used, rather than stimuli or eosinophil donors. Our result is consistent with previous findings in that the phosphorylation of ERK1 is triggered upon exposure to IL-5,^27,32,33 and that eotaxin increases the level of ERK2 phosphorylation.¹² U0126 is, in particular, known to have a similar inhibitory efficacy against MEK1 (IC₅₀ = 70 nm) and MEK2 (IC₅₀ = 60 nm) compared with PD98059, which has much greater affinity for ERK1 than ERK2.⁴² ERK1, of which phosphorylation was regulated by eotaxin and IL-5 to a greater extent than that of ERK2 in our study and others, appears to play a more important role in eosinophil functions. Surprisingly, 1 µm GF109203X significantly inhibited only IL-5-induced phosphorylation of ERK1/2, whereas the phosphorylation of the MAP kinases by eotaxin was not sensitive to the PKC inhibitor, even at 10 µm (Fig. 7b). This suggests that the PKC-mediated eosinophil shape change is restricted to stimulation with IL-5, and that this event may occur via the activation of ERK1/2.

Figure 7.

Inhibition of eotaxin- and interleukin-5 (IL-5)-induced phosphorylation of extracellular signal-regulated kinases (ERK1/2) by inhibitors of MEK (MAP ERK kinase) and protein kinase C (PKC). Purified eosinophils (2 × 10⁶ cells) were preincubated with PD98059 (50 µm), U0126 (10 µm) or GF109203X (1·0, 10 µm) at 37° for 30 min, then stimulated with eotaxin (100 ng/ml) for 1–5 min or with IL-5 (10 ng/ml) for 5–15 min. The phosphorylation of ERK1/2 was analysed by Western blot using anti-phospho-p44/42 mitogen-activated protein kinase (MAP kinase) antibody. Equal loading of proteins was confirmed by reprobing with anti-p44/42 MAP kinase antibody. Phorbol 12-myristate 13-acetate (PMA) (10 nm) always strongly induced ERK phosphorylation, but the result is omitted in this figure. Panel (a) shows inhibition by PD98039 and U0126, and is representative of four experiments. Panel (b) reveals the effect of GF109203X on phosphorylation and represents one of two experiments.

Chemotaxis

In order to confirm that the PKC inhibitor acts only on IL-5-driven eosinophil shape change, we carried out a chemotaxis assay. It should be mentioned that the IL-5-induced migration of eosinophils is shown to be the result of chemokinesis (random migration), not chemotaxis (directed migration) in a Boyden chamber assay, as previously described.¹⁸ The MEK inhibitor blocked the migration of eosinophils by eotaxin and IL-5, whereas PTX inhibited only eotaxin-mediated chemotaxis, thereby confirming the specificity of signalling. GF109203X specifically inhibited IL-5-mediated migration in a dose-dependent manner (57·0 ± 9·6% and 64·5 ± 7·4% at 1- and 10-µm GF109203X, respectively) (Table 1). In contrast, eotaxin-induced chemotaxis was not affected by treatment with the PKC inhibitor and was in agreement with the results from the shape change and phosphorylation of ERK1/2.

Table 1.

Inhibition of eotaxin- and interleukin-5 (IL-5)-mediated migration of eosinophils by GF109203X

	Experiment 1		Experiment 2		Experiment 3
Treatment	Eotaxin	IL-5	Eotaxin	IL-5	Eotaxin	IL-5
Control	290 ± 5	302 ± 3	317 ± 4	388 ± 4	344 ± 4	400 ± 1
PTX
0.1 µg/ml	187 ± 2	255 ± 2	196 ± 1	329 ± 3	56 ± 1	327 ± 3
1 µg/ml	96 ± 4	260 ± 1	100 ± 2	304 ± 3	38 ± 1	345 ± 4
PD98059
10 µm	176 ± 1	201 ± 3	209 ± 4	246 ± 2	181 ± 5	264 ± 2
50 µm	158 ± 1	146 ± 1	139 ± 1	162 ± 2	81 ± 1	173 ± 1
GF109203X
1 µm	271 ± 6	105 ± 2	305 ± 4	134 ± 1	350 ± 4	236 ± 1
10 µm	281 ± 17	87 ± 2	337 ± 1	88 ± 1	456 ± 1	191 ± 9

Purified eosinophils (1 × 10⁵ cells) were preincubated with inhibitors for 30 min at 37° and loaded into the upper chambers.

The lower chambers contained either eotaxin (100 ng/ml) or IL-5 (10 ng/ml). The chambers were incubated for an additional 3 hr, and the migrated cells (set up in triplicate) were counted and expressed as relative mean ± standard error of the mean (SEM). The migration of cells in the presence of medium alone was referred to as 100%. This data represent three of six to seven experiments.

PTX, pertussis toxin.

Discussion

Leucocyte trafficking is one of the most essential parts of inflammation. Chemoattractants stimulate leucocytes to induce a series of events, including an increase of transient calcium mobilization, activation of various signalling molecules, actin rearrangement and shape change, culminating in chemotaxis. Because shape change of a leucocyte must occur before or during its movement, it may be considered a necessary and/or preliminary step for chemotaxis to occur. In this study we have examined the signalling pathway for eosinophil shape change induced by eotaxin and IL-5, the two most important chemokines/cytokines that are known to selectively regulate the migration and function of eosinophils. Our data demonstrate that eotaxin induces eosinophil shape change in a dose- and time-dependent manner (Fig. 1a,b). In particular, 30 ng/ml eotaxin consistently induced the maximum shape change of eosinophils isolated from more than 20 donors, even though the optimal time-point for the maximum response varied, depending upon donors. We confirm that this type of assay is a very reliable and reproducible method for the measurement of eosinophil activation, as described in a previous report.³⁵ The extent and the time of eotaxin-induced eosinophil shape change are almost identical to those from the previous study. Given the fact that IL-5 regulates the mobilization of eosinophils in vivo and that it primes and activates eosinophil locomotion in vitro, it is not a surprise that IL-5 is able to elicit eosinophil shape change, although it has a delayed kinetics and more potency than eotaxin. The maximum shape change occurred at 30–60 min after treatment with 10 ng/ml IL-5. A similar gap in Ras-activation kinetics is also detected in IL-5 and PAF-stimulated eosinophils.²⁹ The ability of IL-5 to induce the robust shape change supports its prominent role in migrating eosinophils in vivo and in vitro.5–^7,41–44 Furthermore, the priming effect of IL-5 with subsequent eotaxin treatment on eosinophil shape change supports evidence for the co-operative roles in the migration of eosinophils from bone marrow to peripheral blood or to tissues.^7,41 The slow response to eosinophil shape change, however, is consistent with the fact that IL-5 is a relatively weak chemoattractant in vivo⁵ and thus may not be considered as the primary chemoattractant of eosinophils when extravasation of the leucocyte is urgently required.

Although both eotaxin and IL-5 act as excellent stimulators of eosinophil shape change, the structurally different nature of their receptors prompted us to examine whether there could exist a differential regulation to induce eosinophil shape change. Our results re-emphasize the central importance of MAP kinases in eosinophil shape change and chemotaxis (Fig. 3 and Table 1). The involvement of MAP kinase is expected because a number of previous studies have unambiguously documented its crucial participation in a variety of eosinophil functions. The ERK1/2 are shown to regulate survival, chemotaxis, IgG binding to the Fc receptor, degranulation, rolling and release of arachidonic acids.^{12–15, 27–30,32,33} p38 has also been shown to be a key regulator of such eosinophil functions, and can be activated by PAF, IL-5 and eotaxin.^13,32 The latter MAP kinase largely shares its roles in eosinophil functions with ERK1/2, yet it has been proposed to play preferential roles in the differentiation and cytokine production of eosinophils.³¹ Hence, our data showing that inhibitors of ERK1/2 and p38 block the shape change provide additional evidence that they act as critical members on eotaxin and IL-5-mediated regulation of eosinophil activity.

In this study, the treatment with PTK inhibitors such as herbimycin A and genistein (Fig. 5a,b) results in only a partial reduction of eotaxin- and IL-5-induced eosinophil shape change, even at 10 µm (40 ± 8% versus 30 ± 5%, respectively). Furthermore, the Jak2 inhibitor did not prevent shape change at all (Fig. 5c). These results are in marked contrast to previous findings in which these inhibitors, at concentrations similar to those used in the present study, effectively blocked shape change and chemotaxis as well as the activation of ERK1/2 induced by PAF, IL-5 or eotaxin.^14,18,32 As our data represents the average of at least three independent experiments using granulocytes with unknown in vivo priming from different donors, and there is strong concordance between assays (i.e. it has a small standard deviation), the relatively weak inhibition would hardly be reconciled with the results from the studies mentioned above. One study demonstrates that even though PAF and vasoactive intestinal peptide (VIP), both of which bind the same class of Gi-coupled receptors, are equally able to induce chemotaxis of eosinophils, their chemotactic responses to herbimycin A are drastically different from each other. PAF-mediated chemotaxis is insensitive to the inhibitor, whereas VIP-induced chemotaxis is augmented. Therefore, migration towards chemotactic agents with a similar nature may have different outcomes in response to inhibitors.⁴⁵

Other two similarities between eotaxin- and IL-5- induced eosinophil shape change are the responses to depletion of intracellular Ca²⁺ and a PKA inhibitor. As chelation of intracellular Ca²⁺ is known to inhibit IL-5-, but not PAF-, mediated migration of eosinophils, even though IL-5 does not trigger any change in the intracellular level¹⁸ of Ca²⁺, it could be speculated that an event associated with Ca²⁺ signalling might be able to help to clarify the difference between chemotaxis and chemokinesis in terms of signalling pathway. Our result indicates that intracellular Ca²⁺ is necessary for both IL-5- and eotaxin-induced shape change (Fig. 4a). These data favour the results from previous studies in which buffering intracellular calcium prevents the PAF and C5a-induced migration of eosinophils.⁴⁶ In addition, PAF-induced eosinophil ERK1/2 phosphorylation is shown to be reduced by BAPTA/AM.³² Therefore, the event of shape change seems to require a homeostatic level of Ca²⁺, which is essential for both chemotaxis and chemokinesis. Our data reveals that PKA is involved in both eotaxin- and IL-5-induced shape change of eosinophils, as a PKA inhibitor, H89, substantially abrogates the shape change at 10 µm, a concentration below which cell viability is influenced. There are previous studies for and against the role of PKA for eotaxin- and IL-5-mediated leucocyte functions. In one study, H89 inhibits eotaxin-driven adhesion of activated T cells,⁴⁷ whereas in the other it marginally inhibits PAF- and IL-5-induced migration of eosinophils at 1 µm, yet its effect at concentrations higher than 1 µm was not tested in that experiment.¹⁸ Our data indicates that PKA appears to be one component mediating eosinophil shape change by eotaxin and IL-5.

The parallel signalling from MAP kinases, PTKs, Ca²⁺ and PKA for eosinophil shape change induced by eotaxin and IL-5 comes to an end at PKC signalling. PKC enzymes have been implicated in a range of cellular functions within the eosinophil, although there is no direct evidence that either IL-5 or eotaxin activates eosinophil PKC. Rather, there are a few reports that PKC inhibitors modulate IL-5- or eotaxin-mediated eosinophil functions. Activation of the respiratory burst and production of a growth factor that are induced by IL-5 and/or eotaxin, are inhibited by PKC inhibitors.^48–50 Our result indicates that the involvement of PKC is strictly restricted to IL-5-induced eosinophil shape change, as evidenced by ≈ 40–50% inhibition at 100 nm and 10 µm of staurosporine and GF109203X (Fig. 6). A higher level of inhibition is even found in chemotaxis (57–65%) (Table 1) and ERK1/2 phosphorylation (≥ 50%) at the same concentrations (Fig. 7b). Our results are highly consistent with data from a study of Schweizer et al.¹⁸ In that study, staurosporine (200 nm) was shown to inhibit IL-5-induced chemokinesis, but not PAF-mediated chemotaxis. In contrast, GF109203X (1 µm), which is known to be more specific for PKC than staurosporine,⁵¹ only slightly affects eosinophil migration induced by IL-5 (≈ 15%). With respect to the latter result, the small inhibition of IL-5-induced chemokinesis by 1 µm GF109203X does not necessarily contradict our data because we also obtained a similar and minimal inhibition of shape change (15·0 ± 5·1%). An increased concentration (10 µm) of GF109203X considerably inhibited only IL-5-induced shape change (Fig. 6a). A similar consideration can be taken in terms of inhibitor dose used. Staurosporine at 10 nm does not inhibit PAF-induced ERK1/2 phosphorylation.³² The marginal, but significant, inhibition of IL-5-induced shape change at that concentration of staurosporine was also detected in our study. Therefore, our data are comparable to the results from those reports when specificity and the concentrations of inhibitors used are considered. The inhibition of shape change and migration by GF109203X or staurosporine, together with an absolute requirement of intracellular Ca²⁺ for shape change, suggests that eosinophils utilize a Ca²⁺-dependent signal-transduction pathway for migration. Once PKC is activated, a well-established pathway via Raf activates ERK1/2.^52,53

What brings about the differential PKC activation? Eotaxin and IL-5 activate PI 3-kinase in eosinophils.^29,32,54,55 Furthermore, the shape change induced by the two chemotactic cytokines was completely inhibited by U73122, a PLC inhibitor (Fig. 5b) and was dependent on intracellular Ca²⁺ (Fig. 4a). Therefore, a point at which eotaxin- and IL-5-mediated signalling pathways take their own way must be downstream of the PLC and Ca²⁺ effect. However, as products of PLC and Ca²⁺ act directly on eosinophil PKC activation,^56,57 the mechanism of such differential PKC activation is not easily explained. Nonetheless, differential PKC activation might be an answer to the difference in signalling pathway observed between chemotaxis and chemokinesis. Although chemotaxis and chemokinesis seem very similar in mechanistic features that require actin rearrangement, shape change and cell motility,⁵⁸ it is not clear whether chemokinesis (random movement) precedes chemotaxis (directed movement), or vice versa. Alternatively, an additional signal could override chemotaxis to lose directionality, leading to chemokinesis. If the latter is the case, however, a signal resulting from PKC activation by IL-5, in which the signal is not provided by eotaxin, would be a possible source of signal that allows the two phenomena to be discernible and somehow weighs chemokinesis over chemotaxis in this circumstance. In summary, the two important chemokines/cytokines that selectively regulate eosinophil locomotion and functions induce shape change of the cell via a largely overlapping signal-transduction pathway, but the differential involvement of PKC may provide an important clue to differentiate chemokinesis from chemotaxis.