Immunology Letters 68 (1999) 317–323

Differentiation of a human eosinophilic leukemic cell line, EoL-1:characterization by the expression of cytokine receptors, adhesion

molecules, CD95 and eosinophilic cationic protein (ECP)

C.K. Wong, C.Y. Ho, C.W.K. Lam *, J.P. Zhang, N.M. HjelmDepartment of Chemical Pathology, The Chinese Uni6ersity of Hong Kong, Prince of Wales Hospital, Shatin NT, Hong Kong, PR China

Received 18 January 1999; accepted 15 February 1999

Abstract

Purification of enough eosinophils for the study of allergic inflammation is difficult because eosinophils comprise only a smallpercentage of circulating leucocytes. A human eosinophilic leukemic cell line, EoL-1, has been considered to be an in vitroeosinophilic model. In the present study, the suitability of EoL-1 cells as an eosinophilic model was further investigated. EoL-1cells were induced to differentiate by dibutyryl cyclic AMP (dbcAMP). The expression of cell surface cytokines (IL-3, IL-5,GM-CSF) receptors, adhesion molecules (CD49d, CD11b), and CD95 (Fas) was investigated by flow cytometry. Expression ofeosinophilic cationic protein (ECP) was determined by fluorescence enzyme immunoassay (FEIA) and reverse transcription-poly-merase chain reaction (RT-PCR). EoL-1 cells could be differentiated into eosinophilic vacuole-containing cells by dbcAMP. Theywere found to express cell surface IL-3 and GM-CSF receptors, CD95 and CD49d. Treatment with dbcAMP could induce theexpression of CD11b but decrease the expression of CD95. Anti-CD95 antibody could induce their apoptosis. The differentiationof EoL-1 cells was accompanied by increase in release of ECP into the supernatant and total ECP synthesis. Differentiation ofEoL-1 cells also up-regulated the expression of mRNA for ECP and its level was parallel to the total amount of ECP synthesis.By virtue of their expression of haematopoietic cytokines receptors, adhesion molecules, CD95, synthesis and release of ECP,EoL-1 cells are suitable as an in vitro eosinophilic model for studying eosinophilic functions. © 1999 Elsevier Science B.V. Allrights reserved.

Keywords: EoL-1 cells; Eosinophil differentiation; Dibutyryl cyclic AMP; Eosinophilic cationic protein

1. Introduction

Eosinophils play important roles both in allergicinflammation and in the defense mechanisms againstparasitic infestation [1]. Delayed apoptosis ofeosinophils is a key pathogenic event for eosinophilia inallergic inflammation [2]. Eosinophils express CD95(Fas) on their cell surface and Fas-mediated signalstrigger the apoptosis of eosinophils [3]. The eosinophilicsecretory products such as major basic protein (MBP)[4], eosinophilic cationic protein (ECP) [5], eosinophilperoxidase (EPO) [6], leukotriene C4 [7], platelet-acti-vating factor [8] and substance P [9] are highly cyto-toxic causing tissue injury. ECP concentrations in

serum, sputum and bronchial alveolar lavage (BAL)fluid are elevated in asthmatic patients and correlatewith the severity of asthma [10–12]. In particular, ECPconcentrations in sputum have been shown to correlatewith the levels of cytokine IL-5 which activateeosinophils [13], delay of eosinophilic apoptosis, andthe phagocytosis of eosinophils by macrophages [14,15].Therefore, studies of the control mechanisms ofeosinophils for expression of cell surface cytokine re-ceptors, adhesion molecules, proteins mediating apop-tosis, ECP synthesis and secretion are important forelucidating the immunopathogenesis of allergicinflammation.

There has been recent progress in understanding thedifferentiation [16,17], apoptosis [18] and intracellularsignal transduction [19] of eosinophils. However, thedifficulty of obtaining enough purified eosinophils and

* Corresponding author. Tel.: +852-2-6322332; fax: +852-2-6365090.

E-mail address: waikeilam@cuhk.edu.hk (C.W.K. Lam)

0165-2478/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved.PII: S 0 1 6 5 -2478 (99 )00064 -4

C.K. Wong et al. / Immunology Letters 68 (1999) 317–323318

their short in vitro life span limited such research. Ahuman eosinophilic leukemic cell line, EoL-1, has beenestablished and partially characterised for some years[20]. It has been shown that EoL-1 cells can be inducedto differentiate into eosinophilic granule-containingcells by many stimuli [21]. Normal blood eosinophilsexpress haematopoietic cytokines receptors, adhesionmolecules for migration as adherent cells through tis-sues, CD95 for the induction of apoptosis by Fasligand, as well as cytotoxic granular proteins includingECP [22]. These are important characteristics ofeosinophils in allergic inflammation which have notbeen well studied in EoL-1 cells. This study thereforeattempts to further evaluate the suitability of EoL-1cells as an eosinophilic cell model. Cells were inducedto differentiate and the expression of cell surface cy-tokine (IL-3, IL-5, GM-CSF) receptors, adhesionmolecules (CD11b, CD49d), CD95 and ECP wasinvestigated.

2. Materials and methods

2.1. Reagents and antibodies

Mouse anti-human IL-5 and GM-CSF receptor achain, CD11b, CD49d and Fas (CD95) monoclonalantibodies were obtained from Pharmingen Inc., CA,USA. Mouse anti-human IL-3 receptor a chain mono-clonal antibody was obtained from Genzyme Diagnos-tics, MA, USA. Mouse IgG isotype and FITCconjugated goat anti-mouse IgG (Fc specific) were pur-chased from Sigma Chemical Co., MO, USA. Onehundred base-pair ladder marker was purchased fromAmersham Pharmacia Biotech, Uppsala, Sweden.

2.2. Cell culture

EoL-1 cells (Riken Cell Bank, Tsukuba Science City,Japan) was maintained in RPMI1640 medium (GibcoLaboratories, NY, USA) supplemented with 10%defined fetal bovine serum (Gibco) and 20 mMHepes buffer (Gibco) in 5% CO2−95% humidified air at37°C.

2.3. Endotoxin-free solutions

Cell culture medium was purchased from Gibcofree of detectable lipopolysaccharide (LPS) (B0.1 EU/ml). All other solutions were prepared using pyrogen-free water and sterile polypropylene plasticware. Nosolutions contained detectable LPS, as determined bythe Limulus amebocyte lysate assay (sensitivity limit 12pg/ml; Associates of Cape Cod, MA, USA).

2.4. Detection of morphological change

EoL-1 cells were prepared on slides with the use ofShandon Cytospin III (Shandon Inc., Pittsburgh, PA,USA) and stained with Hemacolor rapid blood smearstaining set (Merck, Darmstadt, Germany). The hema-color rapid color stained cells were examined by lightmicroscopy. The percentage of differentiated cells wasdetermined by counting at lease 100 cells.

2.5. Flow cytometry of cell surface antigens

Indirect immunofluorescent staining, followed byflow cytometry, was used to determine the cell surfaceantigens. EoL-1 cells were washed with Hank’s bal-

Fig. 1. Effect of dbcAMP on the morphological differentiation of EoL-1 cells. After EoL-1 cells were treated (a) without or (b) with 0.1 mMdbcAMP for 9 days with cell number re-adjustment to 5×105/ml every 3 days in 75 cm2 culture flasks, they were harvested and stained withhemacolor rapid blood smear staining set. The hemacolor rapid color stained cells were examined by light microscopy. Arrows denote cellscontaining refractile vacuoles.

C.K. Wong et al. / Immunology Letters 68 (1999) 317–323 319

Fig. 2. Expression of surface IL-3 and GM-CSF receptors a chain, CD11b, CD49d and CD95 on normal EoL-1 cells determined by flowcytometry. (a) IL-3 receptor a; (b) GM-CSF receptor a; (c) CD 11b; (d) CD 49d; and (e) CD 95. Cells were treated with either mouse IgG isotype(black profile) or goat anti-human monoclonal antibody (grey profile). The percentage of expression of each cell surface antigen was determinedby the enhanced fluorescence following incubation with mouse anti-human monoclonal antibody compared with cells incubated with mouse IgGisotype.

anced salt solution (HBSS) (Gibco) and resuspendedwith 0.3 ml cold HBSS. Human pooled serum wasadded to the cell suspension to block the nonspecificbinding sites. After cells had been incubated at 4°Cfor 30 min and washed with HBSS, either mouse IgGisotype or primary mouse anti-human IgG antibodywas added. The cells were incubated at 4°C for 30min and washed with HBSS before secondary FITC-conjugated goat anti-mouse IgG antibody [1:100 (v/v)]was added. After incubation at 4°C for 30 min andwashing with cold PBS with 0.1% BSA, the cells werefinally resuspended in PBS with 0.1% BSA and sur-face antigens were analyzed by flow cytometry (FAC-Scan flow cytometer, Becton Dickinson, CA, USA).A total of 10 000 events (gated to exclude nonviablecells) were collected in the log mode and expressed ashistograms of relative fluorescence intensity.

2.6. Measurement of apoptosis

EoL-1 cells (1×104/well) in 96 well microtiter platewere incubated with anti-CD95 antibody, clone 11(Coulter Immunotech, FL, USA) (5 mg/ml) for 6 h.Their apoptosis was quantitated by Cell Death Detec-tion ELISAPLUS (Boehringer Mannheim, Germany)which assayed nucleosomes in the cell lysate. Resultswere expressed as enrichment factor which is the ratioof O.D. for anti-CD95 antibody treated cell to O.D.for cells without treatment.

2.7. Preparation and assay of intracellular ECP

EoL-1 cells were lysed in normal saline (0.9%NaCl) containing 0.3% N-Cetyl-N,N,N-trimethylam-monium bromide (CTAB). The lysate was stored at

C.K. Wong et al. / Immunology Letters 68 (1999) 317–323320

Fig. 3. Effects of dbcAMP on the cell surface antigens on EoL-1 cells. Cells were treated with dbcAMP (0.1 mM) for 9 days with cell numberre-adjustment to 5×105/ml every three days in 75 cm2 culture flasks. The expression of cell surface antigens was determined by flow cytometry:(a) CD11b of dbcAMP treated cells; (b) CD95 of dbcAMP treated cells. Cells were treated with either mouse IgG isotype (black profile) or goatanti-human monoclonal antibody (grey profile). The difference of expression of CD11b, CD95 after treatment was calculated by comparing Fig.2(c) with Fig. 3(a), Fig. 2(e) with Fig. 3(b).

−20°C until assayed for ECP using fluorescence en-zyme immunoassay (FEIA) (Pharmacia and UpjohnDiagnostics AB, Uppsala, Sweden).

2.8. RT-PCR

RNAs were prepared from EoL-1 cells usingRNeasy Mini Kit (Qiagen GmbH, Hilden, Germany).RNA (5 mg) was reverse transcribed into first-strandcDNA using First-Strand cDNA Synthesis Kit(Amersham Pharmacia Biotech). The reaction mixturewas incubated at 37°C for 1 h, heated at 65°C for 10min, and promptly chilled in ice. The final cDNAproduct was diluted to 100 ml with Tris-EDTA buffer(pH 8.0). PCR was performed in a 50 ml reactionmixture containing 10 ml cDNA, 50 mM KCl, 10 mMTris–HCl, pH 9.0, 1.5 mM MgCl2, 0.1% triton X-100, 250 mM of each dNTPs and 1 unit Taq DNApolymerase (Promega Corporation, WI, USA) as de-scribed previously [23]. The amplification of humanECP was performed using 50 pmole of primer (5%-CT-CACAGGAGCCACAGC-3%; 5%-GGGCAGCGTAT-ACTTTGG-3%). The amount of RNA used wasnormalized by comparison with the amplification ofhuman b-action mRNA using primer (5%-AGCGGGAAATCGTGCGTG-3%; 5%-CAGGGTA-CATGGTGGTGCC-3%). The amplification reactioninvolved 30 cycles of denaturation at 94°C for 1 min,annealing at 62 and 60°C for ECP and b-actin re-spectively for 1 min, and extension at 72°C for 1 minusing DNA Thermal Cycler (Perkin Elmer Corp., CT,USA). PCR products were electrophoresed on 2%agarose gel and stained with ethidium bromide. Therelative intensities of electrophoretic bands were ana-lyzed with Fluor-S™ MultiImager (Bio-Rad, CA,USA).

3. Results

3.1. Morphological differentiation of EoL-1 cellsinduced by dbcAMP

As shown in Fig. 1, EoL-1 cells treated with db-cAMP were found to contain more refractile vacuolesthan control cells. These results suggested that gran-ule matrix was altered leading to the light microscopyappearance of small refractile ‘vacuoles’. The percent-age of differentiated cells containing refractile vacu-loes was found to increase from 15 to 40% upon thetreatment with dbcAMP.

3.2. Expression of surface antigens on EoL-1 cells

This is shown in Figs. 2 and 3. Using flow cytome-try, normal EoL-1 cells were found to express cellsurface IL-3 receptor a (75%), GM-CSF receptor a(84%), adhesion molecules CD49d (98%) and CD95(87%) (Fig. 2(a), (b), (d) and (e), respectively). How-ever, CD11b expression was not detected on normalEoL-1 cells (Fig. 2(c)). Nine-day treatment with db-cAMP could elevate the expression of CD11b by 27%(Fig. 2(c) and Fig. 3(a)) but decrease the expressionof CD95 by 26% (Fig. 2(e) and Fig. 3(b)).

3.3. Induction of apoptosis in EoL-1 cells byanti-CD95 antibody

Using Cell Death Detection ELISAPLUS, anti-CD95antibody treated EoL-1 cells were found to have anenrichment factor of 2.81, compared to 1.00 of un-treated cells. These results suggest that anti-CD95 an-tibody could induce EoL-1 cell apoptosis.

C.K. Wong et al. / Immunology Letters 68 (1999) 317–323 321

3.4. Effects of dbcAMP on extracellular and intracellularECP, and total ECP synthesis of EoL-1 cells

As shown in Fig. 4, dbcAMP could stimulate theincrease of the release of extracellular ECP and totalECP synthesis from day 0 to day 9 but decrease theamount of intracellular ECP of EoL-1 cells from day 3to day 9 compared to cells without treatment.

3.5. Effect of dbcAMP on the mRNA expression ofECP in EoL-1 cells

In order to investigate the effect of dbcAMP on theexpression of ECP gene in vitro, EoL-1 cells were

harvested on different days after the treatment withdbcAMP. Extracted mRNAs were reverse transcribedinto cDNA. Primers for b-actin and ECP were used forthe detection of mRNA of b-actin and ECP in the PCRreaction respectively. As shown in Fig. 5, dbcAMPtreated cells showed similar expression of b-actinmRNA from day 0 to 9. Treatment with dbcAMPcould continuously activate the ECP mRNA gene ex-pression from day 0 to day 9. The level of increase ofECP mRNA was found to be parallel to the amount oftotal ECP synthesis (Fig. 4).

4. Discussion

Allergic diseases such as allergic rhinitis, asthma andeczema are prevalent and have been increasing world-wide [24]. Eosinophilia has been well documented tooccur in association with allergies and parasitic infesta-tion [22,25,26]. Eosinophils contribute to airway hyper-responsiveness in asthma through the effects of highlybasic granule proteins on the bronchial epithelium in-cluding MBP, ECP and EPO [12]. IL-3 and GM-CSFhave been found to activate the secretion of ECP inhuman blood eosinophils [27]. Elevated haematopoieticcytokines such as IL-3, IL-5 and GM-CSF in allergicinflammation not only activate eosinophils but alsoprolong their life span by inhibiting their apoptotic celldeath [2]. The delayed apoptosis of eosinophils by thosethree cytokines eventually lower the phagocytosis ofeosinophils by macrophages and the subsequent re-moval of their cytotoxic granule proteins, thereby fos-tering the lysis of eosinophils to release cytotoxic ECP[28]. Although studies on the differentiation ofeosinophils have been investigated, the mechanisms in-volved the synthesis and secretion of ECP is still notwell understood.

In studies of allergic inflammation, purification ofeosinophils from peripheral blood is often a difficultfirst step since eosinophils comprise only 1–4% ofcirculating leucocytes and have a short in vitro lifespan. The human eosinophilic leukemic EoL-1 cell linehas been established and partially characterised forsome years [20] and considered to be a useful in vitromodel for studying human eosinophils [21]. It has beendemonstrated that these cells can be induced intoeosinophilic granule-containing cells by a number ofstimuli e.g. HIL-3-derived factor [16], prostaglandin E2

and forskolin [17] and butyric acid [29]. To furtherinvestigate the suitability of EoL-1 cells as aneosinophil model, we have found in this study thatEoL-1 cells expressed cell surface IL-3 and GM-CSFreceptors and CD49d (Fig. 2) as well as synthesized andsecreted eosinophilic cationic protein (Fig. 4), the essen-tial characteristics of eosinophils. In our study, wecould not detect IL-5 receptors by flow cytometry (data

Fig. 4. Effects of dbcAMP on intracellular and extracellular ECP andtotal ECP synthesis of EoL-1 cells. Exponentially growing EoL-1 cellsat a concentration of 5×105/ml were cultured with or withoutdbcAMP (0.1 mM) [Fig. 3(a), (b) and (c), control (�), dbcAMP (�)]for 9 days with cell number re-adjustment to 5×105/ml every 3 daysin 75 cm2 culture flask. The extracellular and intracellular ECP weredetermined on days 0, 3, 6, and 9 using FEIA. Total ECP synthesiswas the summation of extracellular and intracellular ECP. Resultswere expressed as arithmetic mean9SD for triplicate experiments.

C.K. Wong et al. / Immunology Letters 68 (1999) 317–323322

Fig. 5. RT-PCR analysis of ECP gene expression in EoL-1 cells. Exponentially growing EoL-1 cells at a concentration of 5×105/ml were culturedwith or without dbcAMP (0.1 mM) for 9 days with cell number re-adjustment to 5×105/ml every three days in 75 cm2 culture flask. EoL-1 cellswere harvested on days 0, 3, 6, 9 and total RNA was extracted, reverse transcribed and analyzed by PCR as described. (a) PCR products inagarose gel. Lanes 1–4 represented the expression of b-actin gene on indicated days. Lanes 5–8 represented the expression of ECP gene onindicated days. (b) Relative intensity of electrophoretic bands. (a)M 100 base-pair ladder molecular weight marker;1: day 0, b-actin; 2: day 3,b-actin; 3: day 6, b-actin; 4: day 9, b-actin; 5: day 0, ECP; 6: day 3, ECP; 7:; day 6, ECP; 8: day 9, ECP. (b) 1: day 0, b-actin; 2:day 3, b-actin;3: day 6, b-actin; 4: day 9, b-actin; 5: day 0, ECP; 6: day 3, ECP; 7: day 6, ECP; 8: day 9, ECP.

not shown) probably because its a chain is a mem-brane-penetrated glycoprotein [30]. However, we foundthat EoL-1 cells expressed CD95, the receptor whichcan be activated to induce cellular apoptosis [31]. EoL-1 cells can therefore be used as a suitable experimentalmodel for studying the apoptosis of eosinophils. Previ-ous studies have indicated that dbcAMP induced apop-tosis of EoL-1 cells, however, it did not suggest thatCD95 system was involved in dbcAMP-induced apop-tosis of EoL-1 cells [17]. In fact, our result demon-strated that the dbcAMP-induced differentiation ofEoL-1 cells decreased the expression of CD95 but anti-CD95 antibody could induce EoL-1 cells apoptosis.

Activated eosinophils not only circulate as non-ad-herent cells in blood but also transmigrate as adherentcells through endothelium and tissues. We found thatEoL-1 cells expressed the CD49d (Fig. 2 (d)) which canreact with integrin a4 chain to mediate binding tovascular-cell adhesion molecule-1 (VCAM-1). More-over, dbcAMP can stimulate EoL-1 cells to expressCD11b/macrophage antigen-1 (Mac-1) which functionsin cell–cell and cell–substrate interactions and as areceptor for intercellular cell adhesion molecule-1(ICAM-1) and ICAM-2. Therefore, normal EoL-1 cellscould be induced to differentiate for the display of thecharacteristics of eosinophils as adherent cells fortransendothelial migration.

Previous studies have shown that dbcAMP inducedeosinophilic differentiation and apoptosis of EoL-1 cells[17]. The present morphological staining supported thatdbcAMP can induce the vacuolation process in EoL-1cells. As human blood eosinophils contain refractile

vacuoles [32], therefore, dbcAMP could induce certaindegree of maturation of EoL-1 cells. It is becausedbcAMP is metabolized into butyrate and cyclicAMP(cAMP) in cells. cAMP is primarily responsible for thedifferentiation of EoL-1 cells into eosinophilic granule-containing cells [17].

After activation, eosinophils release cytotoxicproteins including ECP which have been shown topossess deleterious effects on tissues including thebronchial epithelium [33,34]. Recent studies havedemonstrated that activation of eosinophils and in-crease of the release of ECP occur in patients naturallyexposed to allergen. The elevated serum ECP concen-trations decline consequent to effective therapy. ECPmay therefore be useful in evaluating the treatment ofasthmatic patients and as a marker for the efficacy oftherapy [35]. Our results demonstrated that extracellu-lar ECP concentration and total ECP synthesis in-creased as the EoL-1 cells underwent differentiationinduced by dbcAMP. The determined level of ECP(about 10 ng/106 cells) synthesized by EoL-1 cells iscomparable with a previous report [36]. However, thelevel of ECP level synthesized by EoL-1 cells is muchlower than that of human blood eosinophils. The differ-entiation of EoL-1 cells induced the formation of acti-vated hypodense eosinophils. The hypodensity isrelated to the prodensity of the cells to synthesize andrelease ECP. The decrease of intracellular ECP bydbcAMP might be due to the enhanced degranulationand release of ECP. In terms of morphological appear-ance and the level of ECP synthesis, differentiatedEoL-1 cells are not exactly the same as normal humanblood eosinophils. However, EoL-1 cells can be applied

C.K. Wong et al. / Immunology Letters 68 (1999) 317–323 323

as a useful eosinophilic model for studying eosinophilicfunctions in view of the expression of haematopoieticcytokines receptors, adhesion molecules, CD95, synthe-sis and storage of ECP in EoL-1 cells,

Acknowledgements

This study was supported by a Chinese University ofHong Kong Direct Grant for Research (Project Code2040615) and a donation from Zindart (De Zhen)Foundation Ltd., Hong Kong.

References

[1] P.F. Weller, N. Engl. J. Med. 324 (1991) 1110–1118.[2] H.U. Simon, K. Blaser, Immunol. Today 16 (1995) 53–55.[3] K. Matsumoto, R.P. Schleimer, H. Saito, Y. Likura, B.S. Boch-

ner, Blood 86 (1995) 1437–1443.[4] W.W. Filley, G.M. Kephart, K.E. Holly, G.J. Gleich, Lancet 2

(1982) 11–15.[5] C.J.F. Spry, P.-C. Tai, J. Barkans, Int. Arch. Allergy Appl.

Immunol. 77 (1986) 252–254.[6] R.L. Olsen, C. Little, Biochem. J. 209 (1983) 781–787.[7] W.F. Owen, R.J. Soberman, T. Yoshimoto, A.L. Sheffer, R.A.

Lewis, K.F. Austen, J. Immunol. 138 (1987) 532–538.[8] G.J. Gleich, C.R. Adolfson, Adv. Immunol. 39 (1986) 177–253.[9] J.V. Weinstock, A. Blum, J. Walder, R. Walder, J. Immunol. 141

(1988) 961–970.[10] J. Bousquet, P. Chanez, J.Y. Lacoste, et al., N. Engl. J. Med. 323

(1990) 1033–1039.[11] J.V. Fahy, J. Liu, H. Wong, H.A. Boushey, Am. Rev. Respir. Dis.

147 (1993) 1126–1131.[12] D.S. Robinson, B. Assoufi, S.R. Durham, A.B. Kay, Clin. Exp.

Allergy 25 (1995) 1118–1127.[13] T. Adachi, S. Motojima, A. Hirata, T. Fukuda, S. Makino, Am.

J. Respir. Crit. Care Med. 151 (1995) 618–623.

[14] D.T. Dudley, L. Pang, S.J. Decker, A.J. Bridges, A.R. Saltiel,Proc. Natl. Acad. Sci. USA 92 (1995) 7686–7689.

[15] M. Stern, L. Meagher, J. Savill, C. Haslett, J. Immunol. 148(1992) 3543–3549.

[16] M. Morita, H. Saito, T. Honjo, et al., Blood 77 (1991) 1766–1775.[17] G. Tai, E.Y. Jung, Y. Horiguchi, M. Kawai, T. Heike, K.

Katamura, K. Furusho, M. Mayumi, Hematol. Oncol. 14 (1996)181–192.

[18] H.U. Simon, S. Yousefi, B. Dibbert, et al., Blood 92 (1998)547–557.

[19] M. Bracke, P.J. Coffer, J.W. Lammers, L. Koenderman, J.Immunol. 161 (1998) 6768–6774.

[20] H. Saito, A. Bourinbaiar, M. Ginsburg, et al., Blood 66 (1985)1233–1240.

[21] M. Mayumi, Leuk. Lymphoma 7 (1992) 243–250.[22] P. Desreumaux, M. Capron, Curr. Opin. Immunol. 8 (1996)

790–795.[23] C.K. Wong, K.N. Leung, M.C. Fung, K.P. Fung, Y.M. Choy,

Immunopharmacol. Immunotoxicol. 15 (1994) 347–357.[24] R.A. Goldstein, W.E. Paul, D.D. Metcalfe, W.W. Busse, E.R.

Reece, Ann. Intern. Med. 121 (1994) 698–708.[25] H. Mehta, W.W. Busse, Compr. Ther. 20 (1994) 651–657.[26] P.F. Weller, K. Lim, H.C. Wan, A.M. Dvorak, D.T.W. Wong,

W.W. Cruikshank, H. Kornfeld, D.M. Center, Eur. Respir. J. 9(1996) 109s–115s.

[27] P.C. Tai, C.J. Spry, Clin. Exp. Immunol. 80 (1990) 426–434.[28] P. Venger, C.G.B. Peterson, in: J. Morley, I. Colditz (Eds.),

Eosinophils in Asthma, Academic Press, New York, 1989, pp.163.

[29] H. Saito, T. Hayakawa, H. Mita, K. Akiyama, T. Shida, Int.Arch. Allergy Immunol. 100 (1993) 240–247.

[30] K. Takatsu, Yakugaku Zasshi 115 (1995) 570–583.[31] A. Druilhe, Z. Cai, S. Haile, S. Chouaib, M. Pretolani, Blood 87

(1996) 2822–2830.[32] P.C. Tai, C.J.F. Spry, Br. J. Haematol. 49 (1981) 219–226.[33] G.J. Gleich, N.A. Flavahan, T. Fujisawa, P.M. Vanhoutte, J

Allergy Clin. Immunol. 81 (1988) 776–781.[34] P. Venge, R. Dahl, K. Fredens, C.G. Peterson, Am. Rev. Respir.

Dis. 138 (1988) S54–S57.[35] S. Ahlstedt, Allergy Proc. 16 (1995) 59–62.[36] M. Morita, Y. Ohshima, H. Akutagawa, Y. Uenoyama, M.

Nambu, M. Mayumi, H. Mikawa, Curr. Med. Res. Opin. 13(1993) 163–174.

.