Involvement of protein tyrosine kinases in activation of human eosinophils by platelet-activating factor

Abstract

Activation of human eosinophils by platelet-activating factor (PAF) involves multiple signal transduction pathways. Among these, protein kinase C has been demonstrated both to mediate respiratory burst and to suppress an alternative pathway of activation of respiratory burst and arachidonic acid metabolism in eosinophils. We utilized inhibitors of protein tyrosine kinases (PTK) to elucidate the role of PTK in PAF-induced activation of eosinophils. Eosinophils were isolated from peripheral blood of atopic donors and stimulated with PAF in the absence or presence of broad-spectrum PTK inhibitors – genistein or lavendustin A; an inhibitor of mitogen-activated protein (MAP) kinase activation – tyrphostin AG126; or an inhibitor of Janus kinase 2 (Jak2) – tyrphostin B42 (AG490). PAF induced superoxide anion (${\text{O}}_2^{-}$·.) generation, leukotriene C4 (LTC₄) release, intracellular calcium ion mobilization and tyrosine phosphorylation of multiple eosinophil proteins in a concentration-dependent manner. All of these responses were concentrationdependently inhibited by genistein; lavendustin A also exhibited potent inhibition of PAF-induced LTC₄ release. AG126 had no effect on either ${\text{O}}_2^{-}$·. generation or LTC₄ release, while AG490 inhibited both responses, albeit less effectively than genistein. We conclude that PAF activates PTK in human eosinophils and that this signalling pathway is involved in eliciting respiratory burst and leukotriene production. The specific PTK(s) involved are unknown but may include Jak2.

INTRODUCTION

Eosinophils are immune effector cells that are particularly associated with immunoglobulin E (IgE) -dependent reactions, such as parasite killing and allergy.¹ They are the distinctive infiltrating cells in human airways in asthma; eosinophil-derived basic proteins, lipid mediators, reactive oxygen species (ROS) and cytokines have been implicated in allergic and asthmatic reactions.²

The products of eosinophils stimulated with platelet-activating factor (PAF) have been demonstrated to induce contraction of airway smooth muscle^{3, 4} and to damage respiratory epithelium in vitro.^{5, 6} The former effect is mediated largely through the release of the cysteinyl leukotriene, LTC₄.^{3, 7} Activated eosinophils have also been shown to increase the responsiveness of bronchial smooth muscle to other contractile agents, ⁸ an action that can be mimicked in vivo by inhaled LTC₄.⁹ The toxic actions of activated eosinophils on respiratory epithelium appear to be mediated largely through a combination of ROS and granule proteins, particularly eosinophil peroxidase.^{5, 10} Thus, both eicosanoid lipid mediator production and other aspects of eosinophil activation, including ROS generation, are important in the eosinophilic inflammation occurring in asthmatic airways.

The cell signalling pathways through which inflammatory mediators activate eosinophils have only recently begun to be elucidated.¹¹ We have recently identified the role of protein kinase C (PKC) in the activation of eosinophil respiratory burst by PAF, measured as production of the ROS, superoxide anion radical (${\text{O}}_2^{-}$·), but also found PKC activation to be responsible for damping of alternative pathways of activation for both respiratory burst and eicosanoid production.¹² The aim of the present study was to determine the role of an alternative set of signal transduction enzymes – the protein tyrosine kinases (PTK) – in these responses in eosinophils. The actions of PTK inhibitors were assessed on ${\text{O}}_2^{-}$· generation and LTC₄ release, as well as on PAF-induced intracellular calcium concentration ([Ca²⁺]_i) elevations and protein tyrosine phosphorylations, to determine the involvement of PTK both in early cell signalling events and in the release of active molecules from the cells.

MATERIALS AND METHODS

Cells

Eosinophils were isolated from acid citrate phosphate-anticoagulated blood of atopic, non-asthmatic donors by negative immunomagnetic selection using anti-CD16-coated microbeads, as described previously.¹² Cell preparations contained > 90% eosinophils [95·5 ± 0·50% (mean ± SEM), n = 37 from 20 donors; contaminants mostly mononuclear cells] which were > 97% viable at the time of experimentation. Eosinophils were suspended in sterile-filtered HEPES-bovine serum albumin (BSA) buffer, as described previously.¹² Cell suspensions were stored on ice for up to 20 min before experimentation. All experiments were performed in HEPES-BSA buffer.

Respiratory burst measurements

Superoxide anion (${\text{O}}_2^{-}$·) generation was measured as the superoxide dismutase (SOD)-inhibitable reduction of ferricytochrome c, as described previously.¹² Results are expressed as nanomoles of cytochrome c reduced per 10⁶ cells in 15 min.

Leukotriene production

Cysteinyl leukotriene release was measured as described previously.¹² As cell suspensions were almost exclusively eosinophils, all cysteinyl leukotriene measured was assumed to be LTC₄.⁷ Results are expressed as femtomoles of LTC₄ released per 10⁶ cells in 5 min.

Measurement of intracellular Ca²⁺

Intracellular Ca²⁺ concentration ([Ca²⁺]_i) was measured in Fura-2-loaded eosinophils, as described previously.¹²

Measurement of protein phosphorylation

Phosphorylation of tyrosine residues on eosinophil proteins was assessed by Western blotting. Eosinophils (10⁶) were incubated for 10 min at 37° in the absence or presence of genistein prior to the addition of 1 µm PAF or an equal volume of HEPES-BSA buffer. Mixtures were incubated for 1–20 min, after which cells were precipitated by centrifugation. Supernatants were discarded; cell pellets were disrupted in 100 µl denaturing lysis buffer (composition: Tris–HCl, 20 mm; NaCl, 40 mm; Nonidet P-40, 1% v/v; deoxycholic acid, 0·5% w/v; ethylenediaminetetraacetic acid, 5 mm; leupeptin, 10 µg/ml; aprotinin, 5 µg/ml; phenylmethylsulphonyl fluoride, 2 mm; NaF, 50 mm; sodium pyrophosphate, 30 mm; sodium orthovanadate, 2 mm; pH 7·4) on ice for 10 min. Lysates were centrifuged at 1000 g for 5 min to precipitate unbroken nuclei and cell debris; supernatants were mixed 1:1 with 4% sodium dodecyl sulphate (SDS) sample buffer (composition: Tris–HCl, 250 mm; SDS, 9·2% w/v; glycerol, 40% v/v; 2-mercaptoethanol, 20% v/v; bromophenyl blue, 0·004% w/v; pH 6·8) and boiled for 5 min. Proteins in cell lysates (approx. 25 µg per sample) were separated by 7·5% polyacrylamide gel electrophoresis and blotted onto polyvinylidene difluoride membranes (400 mA for 1 hr). Tyrosine-phosphorylated protein bands were stained using anti-phosphotyrosine antibody 4G10 (1 µg/ml for 1 hr) and detected by enhanced chemiluminescence (ECL+, Amersham Corp., Arlington Heights, IL).

Statistical analysis

Data are expressed as arithmetic mean ± SEM or geometric mean with 95% confidence interval (CI) from the indicated numbers of experiments.

All statistical analyses were performed using instat® (graphpad™ Software, San Diego, CA). Groups were compared by repeated-measures anova. Comparisons between untreated (control) cells and cells pretreated with inhibitors were performed using Dunnett’s test for multiple comparisons; comparisons between points on concentration–response curves obtained in the absence and presence of inhibitors were made using Bonferroni-corrected Student’s t-tests. Indices of [Ca²⁺]_i changes were compared using paired Student’s t-tests. A probability < 0·05 was defined as significant.

Materials

Fura-2 acetoxymethyl ester (Fura-2/AM), lyophilized PAF, lavendustin A (RG14355), tyrphostin A10 (AG126) and tyrphostin B42 (AG490) were purchased from Calbiochem-Novabiochem GmbH [Bad Soden (Taunus), Germany]. Genistein and lavendustin A methyl ester were supplied by Alexis GmbH (Grünberg, Germany). Anti-FcγRIII (CD16)-coated microbeads and magnetic separation apparatus (MACS system) were purchased from Miltenyi Biotec GmbH (Bergisch Gladbach, Germany). Cysteinyl leukotriene enzyme immunoassay (EIA) kits were obtained from Cayman Chemical Co. (Ann Arbor, MI). The 4G10 anti-phosphotyrosine monoclonal antibody was purchased from Upstate Biotechnology Inc. (Lake Placid, NY). Acid citrate phosphate, cytochrome c (type VI, from horse heart) and SOD were supplied by Sigma Chemie (Deisenhofen, Germany). All other reagents were of analytical grade.

Genistein, tyrphostins, lavendustin A and Fura-2/AM were dissolved in dimethyl sulphoxide. Lyophilized PAF was dissolved at 1 mm in deionized water 15–30 min before use. All compounds were diluted to the desired concentration in HEPES-BSA buffer.

RESULTS

Effects of PTK inhibitors on PAF-induced ${\text{O}}_2^{-}$· generation

PAF (1 µm) induced a production of ${\text{O}}_2^{-}$· by human eosinophils that was consistently significantly higher than basal production (P < 0·05 in all sets of experiments). Both basal and PAF-induced ${\text{O}}_2^{-}$· generation were inhibited in a concentration-dependent manner by the non-selective PTK inhibitor, genistein, with a median inhibitory concentration (IC₅₀) in PAF-stimulated cells of 0·93 µm (geometric mean, 95% CI 0·18–4·8 µm; Fig. 1a).

Figure 1.

Effects of (a) genistein (n = 6), (b) tyrphostin AG126 (n = 3) and (c) AG490 (n = 3) on basal and PAF-induced ${\text{O}}_2^{-}$· generation in human eosinophils. Data are mean ± SEM from the indicated numbers of experiments conducted in duplicate. PAF-stimulated ${\text{O}}_2^{-}$· generation was significantly higher than baseline in all cases (P < 0·05). *P < 0·05, **P < 0·01, compared to control cells preincubated without inhibitors.

To determine which PTK(s) might participate in this response, two drugs with greater selectivity were studied. Tyrphostin AG126 had no significant effect on either basal or PAF-induced ${\text{O}}_2^{-}$· generation (Fig. 1b). In contrast, tyrphostin B42 (AG490) caused significant suppression of PAF-induced ${\text{O}}_2^{-}$· generation at concentrations in the range 0·1–10 µm (Fig. 1c).

Effects of PTK inhibitors on PAF-induced LTC₄ release

PAF induced a concentration-dependent release of LTC₄ from human eosinophils in the range 20 nm−20 µm that declined at higher concentrations (not shown). In the presence of 10 µm genistein, PAF-induced LTC₄ release was abolished (Fig. 2a). A second broad-spectrum PTK inhibitor, lavendustin A (as the cell-permeant methyl ester), also abolished PAF-induced LTC₄ release at 10 µm and inhibited substantially at 1 µm (Fig. 2b). Lavendustin A methyl ester could not be studied for effects on ${\text{O}}_2^{-}$· production owing to its deep orange colour, which interfered profoundly with the cytochrome c reduction assay.

Figure 2.

Effects of (a) genistein (n = 6), (b) lavendustin A (n = 3), (c) tyrphostin AG126 (n = 3) and (d) AG490 (n = 3) on PAF-induced LTC₄ release from human eosinophils. Data are mean ± SEM. *P < 0·05, **P < 0·01, ***P < 0·001, compared to responses to the same concentration of PAF in the absence of inhibitors.

Similarly to ${\text{O}}_2^{-}$· production, PAF-induced LTC₄ release was unaffected by AG126 (Fig. 2c) but was inhibited by approximately 50% by 10 µm AG490 (Fig. 2d).

Effects of genistein on [Ca²⁺]_i

The possibility of suppression of early cell signalling events by PTK inhibition was addressed by studying the effect of genistein on PAF-induced Ca²⁺ mobilization. Pre-incubation of eosinophils with 10 µm genistein for 10 min caused a small, non-significant reduction in baseline [Ca²⁺]_i and had no effect on peak PAF-induced [Ca²⁺]_i elevations (Fig. 3). The return of [Ca²⁺]_i to baseline levels after peaking, however, was accelerated in the presence of genistein (Fig. 3), with both plateau [Ca²⁺]_i (level 60 seconds after addition of PAF) and time taken for [Ca²⁺]_i to return to 10% of peak rise (90% recovery time: RT₉₀) being reduced significantly (Table 1).

Figure 3.

Effects of genistein on PAF-induced calcium mobilization in human eosinophils. Changes in [Ca²⁺]_i induced by PAF are shown in the absence and presence of 10 µm genistein. Data are mean from three experiments. Additional details are given in Table 1.

Table 1.

Effects of genistein pretreatment on indices of PAF-induced Ca² mobilization in human eosinophils

	Baseline (nm)	Peak (nm)	‘Plateau’ (nm)	RT90 (seconds)
Control	115 ± 5·59	330 ± 7·89	161 ± 6·70	132 ± 5·96
Genistein 10 µm	106 ± 15·0	321 ± 58·1	116 ± 15·3	48·8 ± 2·79

Data are mean ± SEM from three experiments. Baseline = mean [Ca²⁺]_i in the 60 seconds immediately preceding addition of 1 µm PAF; peak = maximum value of [Ca²⁺]_i after addition of PAF; ‘plateau’ = [Ca²⁺]_i 60 seconds after addition of PAF; RT₉₀= time taken for [Ca²⁺]_i to return from peak to within 10% of baseline.

Effects of genistein on PAF-induced protein tyrosine phosphorylation

PAF (1 µm) caused a time-dependent phosphorylation of tyrosine residues on a number of eosinophil proteins (Fig. 4a). There was notable, reversible tyrosine phosphorylation of proteins with apparent molecular masses of 30 000 and below, and of 63 000 and 109 000. Genistein caused a concentration-dependent inhibition of the PAF-induced phosphorylation of all of these proteins with similar potency (Fig. 4b,c). Within the range of concentrations of genistein that were effective in suppressing eosinophil functions, the degree of inhibition of PAF-induced tyrosine phosphorylation was less than 50% (Fig. 4c).

Figure 4.

PAF-induced tyrosine phosphorylation of eosinophil proteins. (a) Western blot of phosphotyrosine-containing proteins in unstimulated human eosinophils (left lane) and eosinophils activated with 1 µm PAF for 2, 5, 10 and 20 min. (b) Western blot of phosphotyrosine-containing proteins in unstimulated eosinophils (lane 1), and eosinophils activated with 1 µm PAF in the absence (lane 2) or presence of genistein (lanes 3–5). (c) Effects of genistein on tyrosine phosphorylation of selected eosinophil proteins, quantified by laser densitometry of Western blots. Data are mean of two experiments.

DISCUSSION

The objective of this study was to identify the role of PTK in mediating responses of human eosinophils to the inflammatory mediator, PAF. The broad-spectrum PTK inhibitor, genistein, was used to demonstrate PTK involvement in respiratory burst, LTC₄ production and Ca²⁺ mobilization responses, while two more selective inhibitors – AG126, which inhibits tyrosine phosphorylation of p42^MAPK, ¹³ and AG490, which inhibits the activity of Janus kinase 2 (Jak2)^{14, 15}– were studied to determine which PTKs might be involved in eosinophil activation.

PAF stimulated respiratory burst and LTC₄ release in eosinophils and also induced increases in cytoplasmic free calcium concentration [Ca²⁺]_i, which may be involved in the activation of one or both of these responses.^{16, 17} PAF also induced rapid, transient phosphorylation of tyrosine residues on multiple eosinophil proteins, indicating the activation of one or more tyrosine kinase enzyme(s). Genistein suppressed both ${\text{O}}_2^{-}$· generation and LTC₄ release in human eosinophils, demonstrating that tyrosine kinase activation is required for induction of both of these responses. A second broad-spectrum PTK inhibitor, lavendustin A, also potently inhibited PAF-induced LTC₄ release, confirming the dependence of this response on PTK.

We have demonstrated previously that an augmentation of PAF-induced eicosanoid (thromboxane and LTC₄) production in eosinophils by PKC inhibitors is accompanied by an increase in both peak and post-peak [Ca²⁺]_i.¹² Since this indicates possible involvement of intracellular calcium mobilization in PAF-induced LTC₄ production, we investigated the effect of genistein, at a concentration that inhibits this response, on the [Ca²⁺]_i signal following PAF stimulation of eosinophils. While peak [Ca²⁺]_i was unaffected by 10 µm genistein, the post-peak return to baseline was accelerated significantly, indicating that the prolonged [Ca²⁺]_i elevation depends upon PTK activation, while the induction of the initial [Ca²⁺]_i‘spike’ is apparently independent of PTK. The finding that a PKC inhibitor and a PTK inhibitor – drugs that have opposite actions on PAF-induced LTC₄ release – also have opposite effects on PAF-induced Ca²⁺ mobilization supports the hypothesis that the generation of arachidonic acid metabolites in eosinophils is dependent upon [Ca²⁺]_i elevation.

To identify which PTK(s) may be involved in transduction of PAF signals in eosinophils, we studied two more selective PTK inhibitors: tyrphostins AG126 and AG490. AG126 had no effect on either respiratory burst or LTC₄ release induced by PAF in eosinophils at a concentration close to that which abolishes lipopolysaccharide-induced tumour necrosis factor-α production in murine macrophages, suggesting that mitogen-activated protein kinase (MAPK) activation is unlikely to be required for transduction of PAF signals in human eosinophils. PAF has been demonstrated to induce transient activation of p42^MAPK (Erk2);¹⁸ it might therefore seem surprising that AG126, which prevents Erk2 activation, ¹³ was without effect on PAF-induced eosinophil responses. In fact, suppression of Erk2 activation by an inhibitor of MAPK/Erk kinase (Mek) causes no inhibition of respiratory burst responses in PAF-primed eosinophils.¹⁸ This finding indicates a similarity with leukotriene B₄-stimulated guinea-pig eosinophils, in which p42/p44 MAPK (Erk1/2) are activated but play no part in mediating either respiratory burst or arachidonic acid metabolism.¹⁹ It seems likely that PAF activates Erk2 in human eosinophils but that Ca²⁺ mobilization, respiratory burst and LTC₄ secretion are not dependent upon activation of this enzyme.

AG490, on the other hand, caused roughly 50% inhibition of both responses, indicating that both responses involve a PTK that is sensitive to inhibition by this tyrphostin. While AG490 has been demonstrated to exert inhibitory actions in leukaemic B lymphocytes through Jak2 inhibition, ¹⁴ it also inhibits receptor PTKs such as epidermal growth factor receptor kinase (EGFR-K) and EGFR-related kinases, ²⁰ a property it shares with genistein²¹ and lavendustin A²² but not AG126.^{23, 24} It should be noted, however, that AG490 is 30-fold more potent than genistein in inhibiting EGFR-K but was substantially less effective than genistein as an inhibitor of PAF-induced eosinophil activation. This suggests that the effect of AG490 is mediated by Jak inhibition while the additional effects of genistein and lavendustin A reflect inhibition of other PTKs that are insensitive to AG126 and AG490. Jak2 has been shown to be involved in interleukin-5 signalling in human eosinophils;²⁵ it is therefore a candidate for involvement in signalling in eosinophils activated with other G protein-coupled receptor stimuli.

Finally, the degree of inhibition by genistein of PAF-induced tyrosine phosphorylation observed in our experiments appears small in view of the drug’s effectiveness in suppressing PAF-induced eosinophil functions. While the ability of genistein to inhibit tyrosine phosphorylation by only 50% may indicate the presence of some non-genistein-sensitive PTK, these activities remain to be identified. However, the abolition of functional responses that we observed at concentrations of genistein producing only 50% inhibition of tyrosine phosphorylation is not unique: for example, AG490 produces near-total suppression of leukaemic lymphoblast proliferation at a concentration (5 µm) that causes only partial inhibition of protein tyrosine phosphorylation.¹⁴ It would appear that a high level of tyrosine phosphorylation is required to effect some cell functions and that a 50% inhibition of PTK activity can have profound effects on PTK-dependent functions.

In conclusion, PAF activates one or more PTK in human eosinophils and this activation is involved in the maintenance of intracellular Ca²⁺ elevations and in the generation of ${\text{O}}_2^{-}$· and of LTC₄ by these cells. The specific PTK enzymes responsible for evoking these responses remain to be identified but may include an AG490-sensitive Jak2.

Glossary

Abbreviations

CI

confidence interval

IC₅₀

median inhibitory concentration

Jak2

Janus kinase 2

LTC₄

leukotriene C₄

MAPK

mitogen-activated protein kinase

${\text{O}}_2^{-}$·

superoxide anion radical

PAF

platelet-activating factor

PKC

protein kinase C

PTK

protein tyrosine kinase

ROS

reactive oxygen species

RT₉₀

90% recovery time

SDS

sodium dodecyl sulphate

SOD

superoxide dismutase