UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl)
UvA-DARE (Digital Academic Repository)
The border-crossing behavior of eosinophils and neutrophils in the lung
Zuurbier, A.E.M.
Publication date
2001
Link to publication
Citation for published version (APA):
Zuurbier, A. E. M. (2001). The border-crossing behavior of eosinophils and neutrophils in the
lung.
General rights
It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s)
and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open
content license (like Creative Commons).
Disclaimer/Complaints regulations
If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please
let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material
inaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letter
to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You
will be contacted as soon as possible.
CHAPTER II
Triple Role of Platelet-Activating Factor
in Eosinophil Migration across
Monolayers of Lung Epithelial Cells:
Eosinophil Chemoattractant and Priming
Agent and Epithelial Cell Activator
^ ^ _ _ _ _ C 'hapler II
Triple Role of Platelet-Activating Factor in Eosinophil Migration across
Monolayers of Lung Epithelial Cells: Eosinophil Chemoattractant and
Priming Agent and Epithelial Cell Activator
Lixin Liu. Astrid E.M. Zuurbier, Frederik P. J. Mul. Arthur J. Verhoeven, René Lutter*. Edward F. Knol and Dirk Roos
Central Laboratory of the Netherlands Blood Transfusion Service (CLB) and Laboratory for Experimental and Clinical Immunology, Academic Medical Center, I 'niversitv of Amsterdam. Amsterdam, The Netherlands
Department of Puhnonology, Academic Medical Center, ( 'niversitv of Amsterdam. Amsterdam, The Netherlands
Infiltration of eosinophils into the lung lumen is a hallmark of allergic asthmatic inflammation. To reach the lung lumen, eosinophils must migrate across the vascular endothelium, through the interstitial matrix, and across the lung epithelium. The regulation of this process is obscure. In this study, we investigated the migration of human eosinophils across confluent monolayers of either human lung H292 epithelial cells or primary human bronchial epithelial cells. Established eosinophil chenioattraetants (II.-8, RANTES, PAF, I.TBj and ( 5 a ) or activation of the epithelial cells with II.-I ß induced little eosinophil transmigration (<7% in 2 h). In contrast, addition of PAF in combination with ( 5 a induced extensive (>20%) transepithelial migration of unprimed and IL-5-primed eosinophils. Eosinophil migration assessed in a Boyden chamber assay, i.e. without an epithelial monolayer, was only slightly increased upon addition of PAF+C5a. Pre-incubation of eosinophils with the PAF-receptor antagonist WEB 2086 only inhibited migration of unprimed eosinophils towards PAF+C5a, whereas pre-incubation of epithelial cells with WEB 2086 abolished migration of both ll.-5-primed and unprimed eosinophils. This latter result indicated the presence of PAF receptors on epithelial cells. Indeed, addition of PAF to epithelial cells induced an increase in cytosolic free Ca2+, which was blocked by the PAF-receptor antagonists WEB 2086 and TCV-309. Our
results show that PAF induces permissive changes in epithelial cells, and that PAF acts as a chemoattractant and priming agent for the eosinophils.
Introduction
Allergic asthma is characterized by massive infiltration of eosinophils into the lung. Eosinophils are considered to play an important role in the pathogenesis of allergic asthma (1 ;2). After infiltration of the allergic reaction sites, eosinophils release toxic granule proteins causing tissue damage, and generate lipid mediators, which can influence the behavior of surrounding vasculature and airway smooth-muscle cells. To reach the airway lumen, circulating eosinophils must first extravasate; i.e. the cells initially roll on the endothelium, followed by firm adhesion to and migration across endothelial cells ( 1 ;3 ). Thereafter, the eosinophils migrate through the interstitial matrix and across the epithelium into the lung lumen (1:4:5). The migration to the sites of inflammation is presumed to be regulated at the levels of (a) adhesion receptors that mediate transient or firm adhesion to inflamed vascular endothelium and to extracellular matrix molecules, (b) activating factors (cytokines and chemokines) that induce migration and expression of adhesion molecules and their ligands. and (c) cells that are present at the inflammatory site and regulate the release of these actuating factors.
Triple Role oj'PAF in Fosmopliil Mip-ulion
The first step in the migration process is the interaction of circulating eosinophils \\ ith the vascular endothelium at the allergic reaction site. Endothelial cells actively regulate leukocyte infiltration in inflammatory tissues through different mechanisms, such as vasodilatation, expression of adhesion molecules, opening of intercellular junctions, and secretion of chemotactic factors (6). One can imagine that selective up-regulation and/or activation of adhesion molecules on eosinophils and endothelial cells by cytokines and the production of chemokincs and/or chemoattractants may promote specific eosinophil migration and their subsequent accumulation in tissues. However, migration studies have shown that the so-called "allergy-related" cytokines or chemokines are not selective in attracting eosinophils (3;7-9). These findings imply that the selective activation and attraction of eosinophils towards sites of allergic inflammation is the result of a complex interplay between different cell types, mediators and adhesion molecules.
Recently, attention has focused on the role of the epithelium in leukocyte infiltration into inflammatory sites (10-13). Besides the barrier function for protection against pathogen invasion, epithelial cells play an active role in the induction of inflammation through the expression of cellular adhesion molecules, such as ICAM-1 (14-16) and integrins ( 17). Another role of epithelial cells in lung inflammation involves the synthesis of a variety of proinflammatory mediators including RANTES, PAF and IL-S after IL-Iß orTNF-a stimulation (18), and GM-CSF after stimulation with eosinophil peroxidase (19).
Whereas eosinophil migration across endothelial monolayers has been described extensively ( 1 ;3;6;8;9), little is known about the mechanisms regulating eosinophil migration across epithelial monolayers. In the present study, we investigated human eosinophil migration across confluent monolayers of lung epithelial cell line cells (H292) and primary human bronchial epithelial cells (HBEC) in vitro. Substantial eosinophil transmigration was observed with combinations of PAF and C5a. or PAF and LTB4. but not with these chemoattractants alone. Besides the conventional potent
chemotactic role of PAF and the priming effect of PAF on eosinophils, we also found that a permissive change of the epithelial monolayers induced by PAF is pivotal for efficient eosinophil transepithelial migration.
Materials & Methods
Reagents
Recombinant human (rh) TNF-a was a gift of Dr. A. Creasy (Cetus. Oakland. CA). PAF, C5a. fMLP and LTB4 were purchased from Sigma Chemical Co. (St. Louis. MO). Rh RANTES was purchased
from Gibco Life Technologies (Gaithersburg, MD). The PAF-receptor antagonist WEB 2086 was a kind gift of Dr. II. Heuer (Boehringer Ingelheim. Ingelheim, Germany), and the PAF-receptor antagonist TCV-309 was purchased from Takeda Chemical Industries (Osaka. Japan)(20). RhlL-lß and rhlL-5 were purchased from Pepro Tech (Rock) Hill. N.I), and rhIL-8 and rhIFN-y were obtained from Boehringer Mannheim (Mannheim. German) ). WEB 2086 was dissolved in DMSO at 1000 times the final concentration and stored at -20°C. C5a, PAF, fMLP. RANTES. LTB4, I|>. 5,
IL-8 and IFN-y were dissolved in PBS supplemented w ith 0.5% (vv v) human serum albumin (USA). (CLB. Amsterdam. The Netherlands) and were stored at -20 C. HEPES medium contained 132 mM NaCI, 6.0 mM KCl. 1.0 mM CaCk 1.0 mM MgS04. 1.2 mM KH;P04. 20 mM HEPES, 5.5 mM
( 'kupier II
mixture of RPMI-1640 and Medium 199 (Gibco, Paisley. UK) supplemented with 0.5% (w/v) HSA. ELISA sample buffer consisted of PBS supplemented with 0.1 % (v/v) Tween-20 (Merck.
Schuchardt. Germany), 0.2 % (w/v) hexadecyl-trimethyl-arnmoniumbromide (Sigma). 0.2 % (w \ ) bovine serum albumin (BSA). (Sigma) and 20 mM EDTA.
Cell culture
The human lung adenocarcinoma-derived cell line H292 (American Type Culture Collection CRL 1848, Rockville, MD) (21 ) was grown in RPMI 1640 supplemented with 10% (v/v) heat-inactivated fetal calf serum (Gibco), 100 U/ml penicillin (Gibco), 100 pg streptomycin/ml (Gibco) and 2 mM glutamine (Merck). The doubling time of the H292 cells was approximately 24 hours. Primary human bronchial epithelial cells (HBEC) were obtained from bronchial tissues with microscopically normal appearance from patients undergoing thoracotomy. The cells were cultured in a keratinocyte medium (Keratinocyte-SKM, Gibco) ( 14). After confluence was reached, cell suspensions were obtained by proteolysis with trypsin/EDTA (Gibco). The activity of trypsin was inhibited immediately after detachment of the cells by means of Soybean trypsin inhibitor typc-ll (Sigma), and the cells were subsequently transferred for propagation.
The 4th-30th passages of H292 cells and 3rd-4th passages of HBEC were used for sub-culturing on polycarbonate membranes (3.0 p. m pore si/e, 12 mm diameter) of inverted transwells (Costar, Cambridge, MA), according to Parkos et al (22) with minor modifications (23). In brief, sterile polyoxymethylene polyacetal collars with an inner diameter equal to the outer diameter of the transwells and with a height of 13 mm were tightly fixed to the bottoms of the transwells. Epithelial cell suspensions were added to the inverted transwells and were allowed to attach for 1 8 hours. Thereafter, the collars were removed, and the transwells were placed upright in culture dishes and incubated for 5 days. In this way. monolayers of epithelial cells hanging underneath the membranes were obtained. Complete confluence of the inverted epithelial cell monolayers was reached as determined by May-Grunwald/Giemsa staining and light microscopy, and the confluence of the monolayers was confirmed by [3H]-inulin (Amersham, Amersham, UK) leakage experiments (24).
Granulocyte isolation
Blood was obtained from healthy volunteers. Granulocytes were purified from a buffy coat of 500 ml of blood by density gradient centrifugation oxer isotonic Pcrcoll (Pharmacia. Uppsala. Sweden) (25). After 1% sis of the erythrocytes in the pellet fraction with a cold lysis buffer containing 155 mM NH4CI, 10 mM KHCO3 and 0.1 mM EDTA (pH 7.4). the granulocytes were washed twice in PBS and
resuspended in HEPES medium without CaCI:.
Eosinophil purification
Human eosinophils were purified by means of the fMLP method (26). In brief, granulocytes in HEPES medium without CaCl: were incubated for 30 mm at 37°C to restore the initial density of the
cells. The cells were washed, were resuspended in PBS supplemented with 0.5 % (w/v) HSA and 13 mM trisodium citrate, and were incubated for 5 mm in a shaking water bath at 37°C. The incubation was continued for 10 mm after the addition of 10 nM fMLP to the cell suspension. Thereafter, the eosinophils were purified by centrifugation ( 15 mm. 1000 x g) over isotonic Pcrcoll ( 1.082 g/ml, pi I 7.4). were washed and were resuspended in HEPLS medium. The purity and viability of the
Triple Role ofPAI'in l-'osiiioplitl Migration
eosinophils were more than 95 %. Contaminating cells were mostly neutrophils. In some experiments, eosinophils were isolated by means of immunomagnetic removal of CD16-exprcssing cells (27).
Eosinophil transmigrât ion
Fresh medium was added to the transwells 4 hours prior to the start of the transmigration assay, and the transwells were washed twice just before starting the experiment. The lower compartment was filled with pre-warmed incubation medium with or without chemoattractants. Eosinophils (5x10' cells in 0.5 ml pre-warmed incubation medium) were placed in the upper compartment, and the transwells were incubated for 2 hours at 37°C w ith 5% CO: and maximal humidity. Whenever indicated,
eosinophils and/or epithelial cells were pre-incubated with 10 pM WEB 2086 for 5 min before the start of the transmigration assay. When epithelial cells were treated with WEB 2086, WEB 2086 was added to the lower compartment. WEB 2086 remained present throughout the experiment. After the 2-hour incubation, the upper and lower compartments were washed separately with incubation medium and ELISA sample buffer, respectively, and the fluids were collected. The cells in the collected fluids and in the excised membranes were lysed in ELISA sample buffer. The percentage of eosinophils that had transmigrated was determined by means of an ELISA for eosinophil cationic protein (ECP).
ECP quantification
The amount of ECP as a measure for the number of eosinophils in different cell preparations was determined by means of a slightly modified, previously described ELISA (28). In brief, specific polyclonal rabbit antiserum (poAb) against human ECP was obtained by immunization of a rabbit with highly purified human ECP (29). Human ECP was a kind gift of Prof. I. Olsson ( Lund. Sweden). The Ig fraction of the rabbit serum was isolated by ammonium-sulfate precipitation and was
biotinylated (30).
Culture plates with 96 wells (Maxisorb. Nunc, Roskilde, Denmark) were coated overnight at 4°C w ith 100 pi of 3 pg rabbit anti-ECP poAb/ml diluted in 0.1 mM NaHCO-,. After each incubation step. the plates were washed 3 times w ith PBS containing 0.1 % (v/\ ) BS A and Tween-20. The wells were blocked with block buffer, consisting of PBS supplemented with 0.2 % (w/v) BSA and 0.1% (v/'v) Tween-20, for 1 hour at 37°C. Highly purified ECP and samples were diluted in ELISA sample buffer, were added at a volume of 100 pi per well, and were incubated for 2 hours at 37°C. The wells were subsequently incubated with 100 pi of 1 pg biotin-conjugated anti-ECP poAb.ml diluted in block buffer for 90 min at 37"C. The biotin-labeled antibodies were allowed to bind
avidin-biotinylated alkaline phosphatase complex (DAKO A/S, Glostrup. Denmark) diluted in Tris-buffcrcd saline supplemented with 0.1 % (v/v) Tween-20 and 0.2% (w/v) BSA for 30 min at room temperature according to the manufacturer's description. Enzymatic activity was detected with I nig
phosphatase/ml substrate (Sigma 104. Sigma) dissolved in I M dicthanolamme, 0.5 mM MgCk 0.02 % NaN, (pH 9.8). The absorbance was measured after 60 mm in a Multiscan Multisoft microplate reader (Labsystems Oy, Helsinki, Finland) at 405 ran
The sensitivity of this assay ranged from 0.1 ng/ml to 10 ng/ml of ECP. We confirmed that this ECP ELISA assay was highly specific for eosinophils, i.e. no reaction was measured in lysatcs of human monocytes, lymphocytes or neutrophils. In addition, eosinophil migration quantified by means of the ECP ELISA and by cell counting in the lower compartment was comparable. The total amount of ECP added to the transwell system, as well as that in each compartment separately (upper
( 'haplcr II
compartment, lower compartment and membrane) was determined. The percentage of recover) was always more than 85 %. The percentage of eosinophils that had transmigrated was calculated from the amount of ECP detected in the lower compartment in relation to the total amount of added ECP. Alternatively, transepithelial migration of calcein-AM-labcled eosinophils was measured both bv ECP detection and by calccin quantification.
Calcein labeling
Eosinophils were labeled with calccin-AM (Molecular Probes, Eugene, OR) before the onset of the transmigration assay. The cells ( lOxlO'Vml) were labeled with 4 jig calcein-AM/ml diluted in HEPES medium for 45 min at 37°C. After labeling, the cells were washed 3 times with HEPES medium. The transmigration assay with calcein-AM-labeled eosinophils was performed as described above, except that HEPES medium was used instead of incubation medium. The concentration of calcein-AM in the upper compartment, lower compartment and membrane was measured with a spectrofluorimeter (Model RF-540, Shimadzu Corporation. Kyoto. Japan). The percentage of eosinophils that had transmigrated was calculated from the amount of fluorescence detected in the lower compartment in relation to the fluorescence of the total added calcein-AM-labeled eosinophils.
Chemotaxis assay
Chemotaxis in Boyden chambers was measured with a computerized image analyzer (Model Quantimet 600. Leica, Cambridge, UK) to quantify eosinophil migration into filters after an incubation of 1 hour (31).
Intracellular free Ca' concentration ([Cef ]',• measurement
For the [Ca24], measurement, cells (6-10 x lO'Vml in HEPES medium) in suspension were loaded with
I U.M indo-1/AM (Molecular Probes) for 40 mm at 37°C (32). The cells were washed, were resuspended in HEPES medium to the previous concentration and were kept in the dark at room temperature. Before being transferred to a cuvette, the indo-l/AM-loaded cells were diluted 10 times in HEPES medium and were pre-warmed for 5 min at 37°C. Fluorescence changes of the
magnetically stirred cells were monitored with a spectrofluorimeter (Model RF-540. Shimadzu Corporation), with 340 and 390 nm as excitation and emission wavelengths, respectively. To calibrate the mdo-1 fluorescence as a function of [Ca~ ],. all trapped indo-1 was saturated with Ca " by addition of 10 pM digitonm. after which the indo-l fluorescence was quenched w ith 0.5 m M MnCF. A dissociation constant of 250 nM for the indo-1-Ca" complex was used to calculate [Ca~ ], (33).
Statistical analysis
Results were expressed as the mean ± SEM of the number of different experiments mentioned in the legends. Results were analyzed with the Student's / test. Two-tailed p values were calculated, and p values exceeding 0.05 were considered not significant.
Triple Role o/ PAF m Eosinophil Migration
Results
Low eosinophil transepithelial migration induced by individual chemoattractants
Human eosinophil migration across continent monolayers of lung epithelial H292 cell line cells was analyzed in the physiological basolateral-to-apical direction (with the epithelial cells growing underneath a filter membrane). Eosinophil transepithelial migration reached its plateau after 2 hours of incubation (unpublished observations). The monolayers remained intact during the transmigration assay as checked by examination of the filter by light microscopy, and by determination of [ H]-inuhn diffusion through the monolayer before and after the transmigration assay.
Individual chemoattractants (C5a, PAF. IL-8. RANTES or LTB4) induced little transmigration of
IL-5-primed eosinophils, i.e. maximally 7% (Table I). Pretreatment of the epithelial monolayers with 5 ng IL-lß/ml for 4 hours (Table 1) or with TNF-u and/or IFN-y for 4 hours or 24 hours (not shown) did not induce this migration either. In these studies eosinophils that were isolated by means of removal of CD16-expressing cells showed a similar response. Moreover, quantification of eosinophil migration by means of ECP measurement yielded similar results as with calcein fluorescence measurement.
Chemoattractants Migration (% ECP)
Medium C5a (1CT8M) PAF (10'6M) IL-8 (108 M) RANTES (50 ng/ml) LTB4 (10"6M) IL-1|l(4h)(5ng/ml) 0.40 + 0.16(4) 4.69 + 1.19 (4) 6.48 + 0.82 (4)' 0.61 +0.14 (3) 1.81 + 0 . 2 5 ( 4 ) ' 0.60 + 0.14(3) 0.46 + 0.09(6)
Table I: Eosinophil transepithelial migration induced by individual chemoattractants Chemoattractants were added tu the lower compartment of the transwell system, except for II -l ß, winch was used for pretreatment of the epithelial cells for 4 h. The eosinophils were incubated with IL-5 (l(rln M) tor 30 min at 37°C before addition to
the upper compartment of the transwell system. Values are given in percentage of input LLP measured in the lower compartment (mean + SEM. (n) experiments). Asterisks indicate significance of difference with medium value (* p<0.05; ** p<0.01).
Enhancement of eosinophil transmigration by P. IF
In general, we observed that neither individual chemoattractants nor most combinations of
chemoattractants induced more than 10% of eosinophil transepithelial migration (Tables 1 & II). Onl> when PAF was combined with C5a, LTB4 or RANTES, higher percentages of eosinophils migrated across the monolayers ('fable II). The enhanced migration towards PAF/C5a (tip to 25%) was not only observed with IL-5-primed eosinophils but also with unprimed cells (Fig. 1 ). In addition, PAF/C5a enhanced the migration of lL-5-primed eosinophils across primary HBEC (C5a: 6.5%; PAF: 9.5%; C5a/PAF: 34.6%). This synergistic response was not observed when C5a was combined with LTB4 (Table II): thus, the enhancement must be due to the effect of PAF. Both PAF and C5a
play a chemotactic role in these transmigration assays, because the addition of cither of these agents to the upper compartment just before starting the migration assay resulted in a dosc-dependent inhibition of eosinophil transmigration towards PAF C5a in the lower compartment (not shown). PAF did not induce damage to the epithelial cell layers, as measured by light microscopy and by
Chemoattractants Migration (% ECP) Medium PAF + C5a PAF + LTB4 PAF + IL-1ß (4h) PAF + IL-8 PAF + RANTES RANTES + IL-8 C5a + LTB4 C5a + IL-1ß(4h) 0 . 3 8 + 0 . 0 6 ( 1 3 ) 2 4 . 8 5 + 3 . 8 4 ( 5 ) * * 1 4 . 4 9 + 4 . 4 5 ( 3 ) * 2.32 + 0.50 (3) 6.18 + 2.17 (4) 1 2 . 4 6 ± 2 . 3 3 (4) * 1.10 + 0.40 (3) 8.28 + 2.61 (3) 2.30 + 1.38 (3) ( 'Ihipicr II
T a b l e II: Eosinophil Iriinsepilhelinl migration induced by combinations of chemoattractants. For chemoattractants and eosinophil treatment, see Table I Values are given in percentage of input ECP measured in the lower compartment of the transwell system (mean ± SEM. (n) experiments) The chemoattractants were added to the lower compartment of the transwell svslem. Asterisks indicate significance of difference between eosinophil transmigration in response to a combination of chemoattractants in comparison to the sum of the responses to die individual chemoattractants ('1 able I); * p< 0.05, ** p< 0.02.
Figure I. Human eosinophil migration across monolayers of lung H292 epithelial cells induced by l'A F'and/or C5a. Unprinied eosinophils (open bars) or IL-5-primed eosinophils [solid bars) (10" cells/ml ) were added to the upper compartment of the transwell system, and chemotactie solutions of PAF ( 10* M I. ('5a ( 1 0s M ) or PAF plus ('5a were added to the lower compartment. Eosinophil transmigration was measured after a 2-h incubation at 37"C'. Asterisks indicate that eosinophil transmigration towards the combination of PAF and ('5a was significant!) higher than the sum of the results with PAF ami ('5a alone (p- 0.05). Data arc mean ± SEM of 4 experiments.
Hepes PAF C5a PAF/C5a
Figure 2. Migration of eosinophils in the Hoyden ( haniber Chemotaxis assa\ I 'hprimed (open bars ) or II -5-primed [solidbars) eosinophils (2x10" mil were added to the upper chambers and chemotactie solutions of ('5a l l n ' M i . PAF (10"* M) or PAF plus ('5a were placed m the lower chambers A I 50-j.tm thick filler (N u.m pore size) and a "slop" filter (0.45 urn pore si/el weie placed between the two chambers After 1 hour at 37°C, the 8-p.m filter was removed, fixed in butanol ethanol (20 80%, \ \ I and stained with Weigert solution Eosinophil migration was analyzed with a computerized Image Analyzer (Quantimet 600) h\ measuring the total distance ol all eosinophils that had migrated more than 10 urn into the 150-um filter. Results are mean ± SI:M ol"3 experiments.
Hepes C5a PAF PAF/C5a
Triple Role oj P. IF in Eosinophil Migration P/C+WEB+Eo+EPI P/C+WEB+EPI P/C+WEB+Eo P/C hepes P/C+WEB+Eo+EPI P/C+WEB+EPI P/C+WEB+Eo P/C hepes
^ ^ ^ ™ B
• " " ™ ^
i , , , ,
Figure 3. 77«' role of PAF in eosinophil migration across lung epithelial monolayers towards mixed clieniolaelic solutions of PAF (I O'6 M) and C5a (II)" M). Unprimed (open hars) and IL-5-primed (solid liars) eosinophils were tested for their response against this mixture (indicated by P/C). The PAF-receptor antagonist WEB 2086 (10 u.M) was incubated either with eosinophils or with epithelial cells, or with both, and remained present during the 2-h transmigration assay. Lung H292 epithelial cells (Figure 3a, mean ± SEM. n=4) and primary HBEC (Figure 3b. mean. n=2) were analyzed in the same experimental set-up. In Figure 3a asterisks mark the significant difference as compared to the control groups (C5a PAF) (*p<0.05; **p<0.01 ). In a control experiment. 0.8 % of DMSO was found to have no inhibitory effect
10 20 30 transmigration 40 50 2 c 5 0 0 -< * 200-100 A. • 1 min . - WEB 2086 + WEB 2086 time CM O 500-200 100 B. 1 min - WEB 2 0 8 6 • WEB 2 0 8 6
Figure 4. Changes in cylosolic free Ca' of epithelial cells in suspension in response lo PAF. 11292 epithelial cells (a) and primary HBEC (b) were suspended in HEPES medium, and the intracellular free Ca2* concentration (|('a:']l) was
measured as described in Materials anil Methods Epithelial cells in suspension showed a rapid increase in |Ca2'], after
PAF 110"M) addition (arrow ) In the low er graphs. Ill u\l \\ I IS 208Ó hail been incubated with the epithelial cells I-4 mm (37"('| before PAF addition.
Chapter II
Eosinophil Chemotaxis in a modified Borden chamber assay
To investigate the effect of PAF on eosinophil migration towards C5a in the absence of epithelial cells. Chemotaxis of eosinophils towards PAF. C5a or combinations was assessed. The total distance migrated by eosinophils in filters was determined in a Boyden chamber assay. Addition of PAF or C5a to the lower compartment, as well as IL-5-priming of eosinophils, was found to enhance migration, but these three factors together appeared to render the strongest migration stimulus (Fig. 2).
Effect of WEB 20S6 on eosinophil transepithelial migration towards PAF and C5a
The role of PAF in inducing eosinophil migration across epithelial monolayers towards C5a was further analyzed with the PAF-reccptor antagonist WEB 2086. When eosinophils were treated with WEB 2086, transmigration of unprimed eosinophils was inhibited, whereas transmigration ofIL-5-primed cells remained unaffected, even when higher concentrations (up to 25 u \ l ) of WEB 2086 were used (Fig. 3a). In contrast, the transmigration of both IL-5-primed and unprimed eosinophils was impaired when the epithelial monolayers were treated with WEB 2086 (Fig. 3a). WEB 2086 treatment of both eosinophils and epithelial cells did not yield more inhibition than WEB 2086 treatment of the epithelial cells alone. Eosinophil migration across monolayers of primary HBEC was comparably affected by WEB 2086 (Fig. 3b).
Increase in cytosolic free Cef concentration in epithelial cells induced by PAF
The inhibitory effect of WEB 2086 treatment of epithelial cells on transmigration of eosinophils indicated the existence of PAF receptors on lung epithelial cells. To investigate the response of epithelial cells to PAF binding, we measured the change in intracellular free Ca" concentration in the epithelial cells. Both H292 epithelial cells and primary HBEC showed a rapid [Ca""1 ], increase Lipon addition of 10"'' M PAF (Fig. 4). and this response was completely abolished by pretreatment of the cells with two distinct PAF-receptor antagonists, i.e. 10 p.M WEB 2086 (Fig. 4) or 0.1 pM TCV-309 (not shown). Addition ofC5a(10"8 M). IL-8(10"8M). fMLP(10"sM). WEB 2086 ( 10 uM or 25 u\1)
or TCV-309 (0.1 pM) to epithelial cells did not cause a Ca"' response (not shown). Moreover, human umbilical vein endothelial cells (HUV'EC) did not show a Ca" response upon addition of PAF. revealing a difference in this respect between epithelial and endothelial cells.
Discussion
Experimental set-up
In this study, we investigated the migration of human eosinophils across monolayers of human lung epithelial cell line H292 cells and primary HBEC. The H292 epithelial cell line was chosen because these cells form polarized confluent monolayers and functional tight junctions in a reproducible manner (24). Moreover, H292 cells resemble differentiated primary bronchial epithelial cells, for instance in the expression of mucins and cytokeratins (unpublished observations). In a pre\ ious study, we demonstrated that human neutrophil migration across lung epithelial cells is strongly dependent on the polarity of the epithelial cells, i.e. much more pronounced in the physiological basolateral-to-apical direction than in the opposite direction (23). Eosinophils exhibit the same preferential transepithelial migration, but to a lesser extent (unpublished observations). Therefore, the epithelial
Triple Rule ni I'AF in Fji\inoplul Migration
cells were routinely cultured in the inverted position, with the apical side towards the lower compartment of the transwcll system. The integrity of the epithelial monolayers was checked after each eosinophil migration assay, for eosinophils may release toxic components upon actuation (7:34-37). and thus may damage the epithelial monolayers. However, we did not detect a decreased integrity of the epithelial cell monolayer after eosinophil passage or after incubation with PAF.
Eosinophil migration was detected by means of an ECP ELISA (28). Quantification of this specific eosinophil marker excludes neutrophil interference in the determination of eosinophil migration. This is necessary because the eosinophil suspensions arc generally contaminated with 0-4 % neutrophils. Detection of eosinophil migration by cell counting or by calccin fluorescence measurement (38) yielded similar results as with ECP measurement. Thus, we found no indication for ECP release during eosinophil migration, or selective migration of ECP-rich or ECP-poor eosinophils.
Cliemotactic and priming role <>/ PAF
Transcpithelial migration of primed eosinophils was found to be low. E\cn PAF. the most potent chemoattractant, attracted less than 7 % of eosinophils. Resnick et al. (39) also observed low eosinophil migration across human intestinal epithelium towards individual chemoattractants. but in contrast, relatively high migration towards PAF. Apparently. PAF induces more eosinophil migration across intestinal epithelium than across bronchial epithelium.
We show that eosinophil migration across bronchial epithelium towards C5a, LTB4 or RANTES
is syncrgistically enhanced by PAF. This stimulatory effect was specific for PAF, but did not occur w hen PAF was combined w ith other chemoattractants. On the basis of these results, we predict that in human eosinophils PAF activâtes a signaling pathway that differs from those activated by C5a, LTB4
or RANTES.
Eosinophils need to be primed (e.g. b> IL-5) for induction of migration across epithelial cells, i.e. unprimed eosinophils hardly migrated across epithelial monolayers. However, PAF C5a induced massive migration of unprimed eosinophils across epithelial monolayers, and this migration could be inhibited by treatment of the eosinophils with the PAF-receptor antagonist WEB 2086. In contrast, treatment of IL-5-primed eosinophils with WEB 2086 did not inhibit transcpithelial migration towards PAF/C5a. This reveals the priming role of PAF that had diffused from the lower compartment to unprimed eosinophils in the upper compartment. In addition, PAF and C5a both induce Chemotaxis of eosinophils m a Boydcn chamber assay. Thus C5a is a chemoattractant for eosinophils, and PAF is not only a chemoattractant but also a primer for eosinophils.
Epithelium-activating role of I'AF
Our results indicate that the synergistic transcpithelial migration of eosinophils towards C5a/'PAF is partly due to the direct action of PAF on the epithelial cells. We found that the synergistic
enhancement by PAF was not observed when the eosinophils migrated in the absence of epithelial cells. Moreover, treatment of the epithelial monolayer with the PAF-receptor antagonist WEB 2086 strongly decreased the transepithelial migration towards C5a/PAF. Expression of PAF receptors in bronchial epithelial cells was confirmed by the finding that PAF induces a rapid increase in the intracellular Ca"* concentration ([Ca2*],) in epithelial cells, a response that was prevented by pretreatment of the epithelial cells with the PAF-receptor antagonists WEB 2086 or TCV-309. PAF receptors have previously been identified in epithelial cells derived from the chinchilla middle ear (40). feline, canine, guinea pig and cow trachea (41-44) and rabbit cornea (45). PAF induces an
( haplcr II
increase in [Ca~'J,, acts as a mucous secretagogue and decreases the ciliary beat frequency of the tracheal epithelial cells (41-43:46). Moreover, PAF receptors have been identified on human primary bronchial epithelial cells (47). and it has been shown that PAF induces up-regulation of the nuclear transcription factor activator protein-1 in these cells (47). Binding of PAF to the receptors in bronchial epithelial cells may cause transduction of signals leading to functional changes, such as the ability to permit eosinophil transmigration.
Together, these data strongly suggest that PAF induces permissive changes in epithelial cells thai favor eosinophil transmigration. However, the exact nature of the epithelial changes that allow eosinophil transmigration remains to be elucidated. The role of PAF is currently under investigation. Preliminary results suggest that the morphology of bronchial epithelial cells is unaffected by PAF. and that the monolayers do not become more "leaky" when PAF is present. Moreover, we found that ICAM-I expression is not up-regulated by PAF (4h incubation with PAF) (unpublished observation). Thus, PAF-treated bronchial epithelial cells do not appear to be activated or morphologically changed. PAF possibly induces more subtle changes in the epithelial cells. One possible explanation might be that tight junction resistance is decreased as a result of the PAF-mduced [Ca" ], increase. Regulation of tight junction resistance by [Ca~H], elevation has been shown in human intestinal
epithelial T84 cell line cells (48). Tight junction resistance in T84 cells is unaffected by PAF (39). because these epithelial cells do not express PAF receptors (49). However, inulin leakage experiments did not show increased permeability of the monolayers. Another possibility is that PAF induces changes in cortical actin. which affects cell-cell adhesion of epithelial cells and may result in augmented transmigration (50).
Differences between endothelial and epithelial cells
PAF did not induce an increase in [Ca2 ], in endothelial cells from umbilical veins (HUVEC), and
PAF did not enhance eosinophil migration across HUVEC towards C5a (6;5 I ) (not shown). Thus. HUVEC probably lack a functional PAF receptor. However, endothelial cells arc capable of PAF production after incubation with IL-lß or TNF-u (52:53). in contrast to epithelial cells (23). These results confirm the importance of the stimulatory effects of PAF on eosinophil transepithclial migration. Moreover, this demonstrates that different mechanisms control eosinophil migration across the endothelium and epithelium.
PAF did not enhance eosinophil transepithelial migration towards IL-8. It is known that 1L-8 is a very poor chemoattractant for eosinophils (9). We have previously shown that lung epithelial cells generate IL-8 after activation with IL-1 ß (23:54). In accordance, eosinophil migration across IL-1 ß-stimulated epithelial monolayers was not enhanced by PAF. These results suggest that IL-S is too weak to induce eosinophil migration even in case of primed eosinophils and PAF-activated epithelium.
Conclusion
Together, our results show that human eosinophils migrate massively across lung epithelial monolayers in response to a chcmotactic gradient of PAF in combination with a potent
chemoattractant (C5a or LTB4). In this process, PAF acts as a priming agent and as a chemoattractant
for the eosinophils. In addition, PAF induces transmigration-permissive changes of the epithelial cells. Clarification of this epithelial activation may help in the development of drugs to prevent eosinophil migration into the lungs and the concomitant damage to the epithelium.
Triple Role </j PAF ui l-~u\iiit>pliil Mrgniiinn
Acknowledgements
The linthors thank Dr I' S 1 liemstra and Dr. S. van Wetering lor providing the primär} human bronchial epithelial cells, and Dr. Arne Egesten and Dr Anton I ..I I ool for helpful discussions and technical assistance The authors also thank Prof. Inge Olsson (I und University. Swollen) for providing purified ECP.
Footnote
fins study was financially supported by the Netherlands Asthma Foundation (grant no. 32.93.13) and by the Nederlandse Fondsenwervingsacties Volksgezondheid (grant no. 93.044 VG).
References
I knol.F.F.. and Roos.l). 1996. Mechanisms regulating eosinophil extra\asation in asthma. Eur.RespirJ. 9:136s.
2. Weller.P.F., Lim.K., Wan.H.O. Dvorak.A.M.. Wong,D.T.W., Ci uikshank.W.W., Kornfeld.H.. and Center.D.M. 1996 Role of eosinophil in allergic reactions. Eur.RespirJ. 9:109s.
3. Moser.R., Fehr.J.. Olgiati.T. and Bruijnzeel, P.L.B. 1992. Migration of primed human eosinophils across cytokine-activated endothelial cell monolayers Blood79:2937.
4. l.ukacs.N.W.. Strietcr.R.M., and Kunkel.S.L. 1995 I eukocyte infiltration in allergic airway inflammation. Am.J.Respir.Cell Mol.Biol. 13:1.
5 Roman. J. 1996 Extracellular matrix and lung inflammation. Immunol.Res. 15:163.
6. Ebisawa.M., Liu.M.C, Yamada.T.. Kato.M.. Lichtenstein.L.M., Bochner.B.S., and Schleimer.R.P. 1994. Eosinophil transendothelial migration induced by cytokines: II. Potentiation of eosinophil transendothelial migration by eosinophil-active cytokines. J.Immunol. 152:4590.
7. Rot.A.. Krieger.M., Brunncr.T.. Bischoff.S.C, Schall.T.J.. and Dahinden,C.A. 1992 RAN I IS and macrophage inflammatory protein 1 [i induce the migration and activation of normal human eosinophil granulocytes. J.Exp.Med. 176:1489.
8. Kitayama.J.. Carr.M.W., Roth.S.J.. Buccola.J., and Springer.T.A. 1997 Contrasting responses to multiple chemotactic stimuli in transendothelial migration. J.Immunol. 158:2340.
9 Ebisawa.M., Yamada.T., Bickel.C, Klunk.D., and Schleimer.R.P. 1994. Eosinophil transendothelial migration induced by cytokines: III. Effect of the chemokine RANTES. J.Immunol. 153:2153. 10. Bloemen. P.C. M.. Nan den Tweel.M.C, Henricks.P.A.J.. Fngels.F.. Van de Velde.M.J.V..
Blomjous.F.J.. and Nijkamp.F.P. 1996. Stimulation of both human bronchial epithelium and neutrophils is needed for maximal interactive adhesion. AmJ.Physiol.Lung Cell.Mol.Physiol. 270:L80.
I 1. Godding, V., Stark.J.M.. SedgwickJ.B., and Busse.W.W. 1995 Adhesion of activated eosinophils to respiratory epithelial cells is enhanced by tumor necrosis factor-a and intcrleukm- 1 ß. Am.J.Respir.Cell Mol.Biol. \ï.^.
12. Rennard.S.I.. Rombcrger.D.J., SissonJ.H., Von Essen.S.G., Rubinstein.I., Robbins.R.A, and Spurzem,J.R. 1994 Airway epithelial cells: Functional roles in airway disease. Am.J.Respir.Crit.Care Med 150:S27.
13. Simon,R.II. and Paine III..R. 1995. Participation of pulmonary alveolar epithelial cells in lung inflammation. J.Lab.Clin.Med, 126:108.
14 Bloemen.P.C.M., Van den Tweel.M.C Henricks.P.A.J., Engels.F., Wagenaar.S.S., Rutten.A.A., and Nijkamp.F.P. 1993 Expression and modulation of adhesion molecules on human bronchial epithelial cells Am.J.Respir.Cell Mol.Biol. 9:586.
15. Cunningham.A.C. and Kirby.J.A. 1995. Regulation and function of adhesion molecule expression by human alveolar epithelial cells. Immunology 86 :279
16. l.i.X.C. Je\nikar.A.M., and Grant.D.R. 1997 Expression of functional ICAM-I and VCAM-1 adhesion molecules by an immortalized epithelial cell clone derived from the small intestine. Cell.Immunol. 175:58. 17. Sheppard.D. 1996. Epithelial integrins. BioEssays 18:655
18. Stellato.C Beek.L.A.. Gorgone.G.A.. Proud.D., Schall.T.J.. Ono.S.J.. Lichtenstein.L.M., and Schleimer.R.P. 1995. Expression of the chemokine RANTES by a human bronchial epithelial cell line: Modulation by cytokines and glucocorticoids. J.Immunol. 155:410.
Chapter II
19. Motojinip.S.. Adachi.T.. Manaka.K.I.. A r i m a . M . . Fukuda.T., and MakinoS. 1996. Eosinophil peroxidase stimulates the release of granulocyte- macrophage colony-stimulating factor from bronchial epithelial cells J.AIIergy Clin Immunol. 98:S216.
20. Terashita.Z., K a n a m u r a . M . . Takatani.M.. Tsushima,S.. Imura.Y.. and Nishikawa.K. 1992 Beneficial effects of TCV-309, a novel potent and selective platelet activating factor antagonist in endotoxin and anaphylactic shock in rodents J.Pharmacol.Exp. Titer. 260:748.
21. Banks-Schlegel.S.P., Gazdar.A.F., and Harris.C.C. 1985. Intermediate filament and cross-linked envelope expression in human king tumor cell lines. Cancer Res. 45:1 187.
22 Parkos.C.A., Delp.C, Arnaout.M.A.. and Madara.J.E. 1991. Neutrophil migration across a cultured intestinal epithelium. Dependence on a CD1 lb/CDl 8-mcdiated event and enhanced efficiency in physiological direction. J.Clin.Invest. 88:1605.
23. l.iu.L.. Mul.F.P.J.. Lutter.R.. Roos.!)., and Knol. F.. F. 1996. Transmigration of human neutrophils across lung epithelial cell monolayers is preferentially in the physiological basolateral to apical direction.
Am~J.Respir.Cell Mol.Biol. 15:771.
24 Van Schill'gaarde.M.. Van Alphen,1... Eijk.P.. Everts.\'.. and Dankert.J. 1995 Paracytosis of Haemophilus influenzae through cell layers of NCI- 11292 lung epithelial cells. Infect.Immun. 63:4729. 25. Roos.D.and De Boer.M. 1986 Purification and cryopreservation of phagocytes from human blood. In:
Methods Enzymol. Volume 132 Immunochemical Techniques Part .1 Phagocytosis and Cell-Mediated
Cytotoxicity (G Di Sabato and .1 Everse, eds.). Academic Press, New York 225.
26. Koenderman.l... Hamelink.M.L., Kok.P.T.M., Verhoeven.A.J.. and Bruijnzeel.PX.B. 1988 An improved method for the isolation of eosinophilic granulocytes from peripheral blood of normal individuals.
J.Leukoe.Biol. 44:79.
27 Hansel.T.T.. De \ ries.l.J.M.. IIÎ.T.. Rihs.S.. Waiul/.ilak.M.. Betz.S.. Blaser.K.. and Walker.C. 1991 An improved immunomagnetic procedure for the isolation of highly purified human blood eosinophils.
J.Immunol.Methods 145:105.
28. Reimert.C.M.. Venge,P., Khara/mi.-V, and Bendtzen.K. 1991. Detection of eosinophil calionic protein (ECP) by an enzyme- linked immunosorbent assay. .1.Immunol.Methods 138:285.
29. Olsson.l. and \ enge.P. 1974. Cationic proteins of human granulocytes. II. Separation of the canonic proteins of the granules of leukemic myeloid cells. Blood 44:235.
30. Gretch.D.R., Suter.M., and Stinski.M.F. 1987. The use of biotinylatcd monoclonal antibodies and streptavidin affinity chromatography to isolate herpesvirus hydrophobic proteins or glycoproteins.
Anal.Bioehem. 163:270.
31 Schweizer.R.C.. Van Kessel-Welmers.B.A.C, Warringa.R.A.J., Maikoe.T.. Raaijmakers.J.A.M.. I.ammers.J.-W.J. and Koenderman.l.. 1996. Mechanisms involved in eosinophil migration. Platelet-actuating factor-induced Chemotaxis and interleukin-5-induced chemokinesis are mediated by different signals. J.Leukocyte Biol. 59:347
32 Vossebeld.P.J.M.. Kessler..).. Von dem Borne.A.E.G.Kr., Roos.D.. and Verhoeven.A.J. 1995
Heterotypic FcyR clusters evoke a synergistic Ca: response in human neutrophils. J.Biol.Client. 270:10671.
33. Bijsterbosch.M.K.. Riglcv.K.P.. and Klaus.G.G. 1986. Cross-linking of surface immunoglobulin on B lymphocytes induces both intracellular C a:' release and Ca2* influx: analysis with indo-1
Biocltem.Biopliys. Res. Commun. 137:500.
34. Kay.A.B. 1991. Asthma and inflammation. J.AIIergy Clin.Immunol. 87:893.
35. Eda.R., Sugiyama.H.. Hopp.R.J.. Okada.C, Bewti a.A.K., and Townley.R.G. 1993. Inhibitory effects of formoterol on platelet-activating factor induced eosinophil Chemotaxis and degranulation. Int.Arch.Allergy
Immunol. 102:391.
36. Eda.R.. Townley.R.G., and Hopp.R.J. 1994. Effect of terfenadine on human eosinophil and neutrophil chemotactic response and generation of superoxide. Ann.Allerg) 73:154.
37. Takalu ji.S.. Ohtoshi.T.. Takizawa.H.. Tadokoro.K.. and Ito.K. 1996. Eosinophil degranulation in the presence of bronchial epithelial cells. Effect of cytokines and role of adhesion. J.Immunol. 156:3980. 38. Sunder-Plassmann.G.. Hofbauer.R.. Sengoelge.G.. and H o r l . W . I . I . 1996. Quantification of leukocyte
migration: improvement of a method. Immunol.Invest. 1-2:49.
39. Resniek,M.B.. Colgan.S.P.. Parkos.C.A.. Delp-Archer.C ., McGuirk.D., Weller.P.F.. and Madara.J.L. 1995. Human eosinophils migrate across an intestinal epithelium in response to platelet-activating factor.
Gastroenterology 108:409.
40. Wu.T., Rieves R.D.. L o g u n . C , and Shelhamer.J.H. 1995. Platelet-activating factor stimulates eicosanoid production in cultured feline tracheal epithelial cells. Lung 173:89.
41 Kahhak.L.. Roehe.A.. D u b r a y . C A r n o u x . C . and Benveniste.J. 1996. Decrease of ciliary beat frequency by platelet actuating factor: protective effect of ketotifen. Inflamm.Res. 45:234.
Triple Role of PAF in Eosinophil Migration
42 Kotulo.M.. Taniaoki.J.. I so no, K., Takeuchi.S., Ozawa.Y., ( hiyotani.A.. and Konno.K. 1994. Effect of platelet-effect factor on intracellular free calcium in cow tracheal epithelium. AmJ.Respir.Cell Mol.Biol. 3:278.
43. Rieves,R.D., GofU. Wu„T. Larivee.P., Logun.C. and Shelhamcr,J.H. 1992 Airway epithelial cell mucin release: immunologic quantitation and response to platelet-activating factor. AmJ.Respir.Cell Moi Biol. 2:158.
44. Taniaoki.J., Sakai.N.. IsonoK.. Kanemui a.T., Yamawaki.I., and Takizawa.T. 1991. Effects of platelet-activating factor on bioelectric properties of cultured tracheal and bronchial epithel ia. J.Allergy
Clin.Immunol. 6:1042.
45 Tao.V.. Ba/an.H.K.. and Bazan,N.G. 1996. Platelet-activating factor enhances urokinase-type plasminogen actuator gene expression in corneal epithelium. Invest.Ophthalmol. Vis.Sei. 37:2037.
46. Lin.J-, Juhn.S.K.. Adams.G.l... Giehink.G.S.. and Kim.Y. I 997 Dexamethasone inhibits mucous glycoprotein secretion via a phosphohpase A2-dependent mechanism in cultured chinchilla middle ear epithelial cells. Ado Otolaiyngol.fStockh.) 1 1 7:406.
47. I.eVan.T.D., Bloom.J.W., Adams.D.G., HenselJ.L., and llalonen.M. 1998. Platelet-activating factor induction of activator protein-1 signaling in bronchial epithelial cells. Mol.Pharmacol. 1:135.
48. Tai.Y.H., Flick.J.. I.evine.S.A.. Madara.J.I... Sharp.G.W.G., and Donowitz.M. 1996. Regulation of tight junction resistance in T-84 monolayers by elevation in intracellular Ca2+: a protein kinase C effect. J.Membr.Biol. 149:71.
49. Travis.S.P.L., ( rotty.B.. and JewelLD.P. 1995. Site of action of platelet-activating factor within the mucosa of rabbit distal colon. Clin.Sei. 88:51.
50 Hordijk,P.I.., Ten Kloostir.J.P.. Van der Kammen.R.A.. Michiels.F.. Oomcn.L.C '.. and ( ollard.J.G. 1997. Inhibition of invasion of epithelial cells by Tiaml-Rac signaling. Science 5342:1464,
51. Warringa.R.A.J., Mengelers.H.J.J., Raaijmakers.J.A.M.. Briiijnzeel.P.L.B. and Koenderman I.. 1993 Upregulation of formyl-peptide and interleukin-8-induced eosinophil Chemotaxis in patients with allergic asthma. J.Allerg}- Clin.Immunol 91:1 198
52. Kuijpers.T.W., Hakkert.B.C, Hart..M.H.I... and Roos.I). 1992. Neutrophil migration across monolayers of cytokinc-prestimulated endothelial cells: A role for platelet-activating factor and IL-8. J.Cell Biol
117:565.
53. Sniart.S.J. and Casale.I.B. 1994. TNF-a-induced transendothelial neutrophil migration is IL- 8 dependent. Am.J.Physiol.Limg Cell.Mo! Physiol. 266:L238.
54 Liu.l... Mul.F.P.J.. Knijpers.T.W. I.utter.R.. Roos.I).. and Knol.F.F. 1996 Neutrophil transmigration across monolayers of endothelial cells and airway epithelial cells is regulated by different mechanisms. Ann.NYAcad.Sci. 796:21.
