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.
C H A P T E R IV
Sequential Migration of Neutrophils
across Monolayers of Endothelial and
Epithelial Cells
Chapter II'
Sequential Migration of Neutrophils across Monolayers of Endothelial
and Epithelial Cells
Frederik P.J. M u l . A s t r i d E.M. Zuurbier. Hans Janssen". Jero Calafat , Sandra \ a n W e t e r i n g ' . Pieter S. Hiemstra , D i r k Roos and Peter L. H o r d i j k .
Frederik P.J. M u l and A s t r i d E.M. Zuurbier contributed equally to this w o r k
Central Laboratory of the Netherlands Blood Transfusion Service (CLB) and Laboratory for Experimental and Clinical Immunology. Academic Medical ('enter, ( diversity of Amsterdam, Amsterdam, The Netherlands
Division of Cell Biology, Netherlands Cancer Institute, Amsterdam. The Netherlands
' Dept. ofPulmonology, Leiden University Medical ( 'enter, Leiden, The Netherlands
In the course of granuiocyte-dominated Inn«; inflammation, granulocytes migrate across the endothelium and epithelium of the tun«; and cause severe tissue damage. To study this process in more detail, we developed a bi-layer transmigration model comprised of primary human endothelial and lung epithelial cells, simultaneously cultured on opposite sides of Transwell filters. Flection microscopical analysis showed that both the morphology of the cells and the expression of junctional proteins remained unaltered, and that matrix components were deposited onto the filter. Inti iguingly. neutrophil migration was more efficient across the bi-layers than across single epithelial monolayers, and did not differ from migration across single endothelial monolayers. Co-culture experiments showed that endothelial cells stimulated epithelial cells to release II.-6 and that epithelial cells enhanced release of IL-8 from endothelial cells. Together these data reveal bi-directional signaling and enhanced neutrophil migration in a transmigration model of primary human epithelial and endothelial cells.
Introduction
In lung inflammation and asthmatic reactions, circulating leukocytes, primarily neutrophils and eosinophils, migrate across the lung endothelium and the lung epithelium into the airway lumen. The process o f transmigration initially involves selectin-mediated rolling o f the granulocytes on the endothelial cell surface and is followed by f i r m , integrin-mediated adhesion to and passage across the endothelial cell layer. Subsequently, the granulocytes migrate through the interstitial matrix and across the epithelium into the lung lumen (1-5). Concomitant activation o f the granulocytes and release o f oxygen radicals and toxic proteins (e.g. eosinophil cationic protein or elastase) cause damage to the lung tissue.
Activation o f endothelial cells by inflammatory stimuli promotes leukocyte infiltration through increased cellular adhesion-molecule expression, increased vascular permeability and production o f chemoattractants ( 1 ; 6-12). The molecular basis underlying transendothelial migration has been w e l l described (9; 10; 13; 14). Granulocyte transmigration is triggered by various types o f chemoattractant. such as chemokines, e.g. Interleukin ( I L ) - 8 , lipid mediators, e.g. P A F , bacteria-derived peptides, e.g. f M L P a n d complement fragments, e.g. C5a (15-17). M i g r a t i o n o f granulocytes is inhibited by antibodies that either block ligand binding o f granulocyte integrins (9; 18; 19), integrin-associated proteins such as C D 4 7 . or lg-like adhesion molecules such as C D 3 I (20-25).
Migration o) Neutrophils across Endothelial and Epithelial Mono/avers
Similar to the endothelium, the epithelium also plays an important role in leukocyte infiltration at sites of inflammation. However, whereas a large number of molecules are implicated in the control of transendothelial migration, only few molecules are known to be involved in the transmigration across epithelial monolayers. These include the leukocyte |J;-integnn CD1 Ib/CDIS
and the glycoprotein CD47 ( 15; 26; 27). The epithelial ligand for the [i:-intcgrin has not yet been
firmly established (27). although adhesion of eosinophils to human bronchial epithelium was recently described to depend on CD18/ICAM-1 interaction (28). These results are in line with earlier reports of up-regulation of cellular adhesion molecules, such as ICAM-1, on activated epithelial cells (29-32). In addition, activated epithelial cells release a variety of proinflammatory mediators, chemokines and lipid mediators that may all modulate leukocyte infiltration (15; 26; 33-35). For instance, bronchial epithelial cells of asthmatic patients have been shown to produce increased levels of IL-I ß. 1L-6. 1L-8. granulocyte/macrophage-colony stimulating factor and IL-16 (35). Finally, the orientation of the epithelial monolayer is essential to allow efficient ;;; vitro transmigration. The physiologically relevant basolateral-to-apical migration of leukocytes is much more efficient than migration in the opposite direction, implicating the polarity of the epithelium as an additional regulatory factor of leukocyte transmigration (26).
Despite the large body of knowledge on migration across monolayers of endothelial or epithelial cells, possible interactions between these cell types and the resulting modulation of leukocyte transmigration have not been thoroughly studied. We have therefore developed a transmigration model in which we simultaneously culture monolayers of primary human lung epithelial cells and human umbilical cord vascular endothelial cells (HUVECs) on opposite sides of Transwell filters. We have characterized this model with respect to morphology and chemoattractant-induced transmigration of granulocytes, and compared the results with those obtained with single endothelial or epithelial cell monolayers. The main results of this study reveal a paracrine interaction between the endothelial and epithelial monolayers resulting in increased release of cytokines and chemokines and enhanced transmigration of neutrophils.
Materials and Methods
Reagents
PAF. C5a, fMLP and isoproterenol were from Sigma Chemical Co. (St. Louis, MO). Recombinant human (rh) IL-8 was purchased from PeproTech (Rocky Hill, NJ) and bFGF from Boehringcr Mannheim (Mannheim, Germany). Monoclonal antibody IB4 (CD18, IgGl) (36) was isolated from the supernatant of the hybridoma by precipitation with 50% saturated ammonium sulphate and subsequent protein-A affinity chromatography. CD31 antibody (IgGl ; monoclonal antibody HEC170) (37; 38) was also isolated from hybridoma supernatants as described above. CD 14 antibody (IgG 1 ; monoclonal antibody 8G3), HSA and fibronectin were obtained from the CLB (Amsterdam, The Netherlands). Fluorescent secondary antibodies were from Dako (Glostrup, Denmark). Vitrogen was obtained from Cohesion (Palo Alto, CA). Calcein-AM and FITC-dextran 3000 was from Molecular Probes (Eugene, OR). RPMI was from Gibco (Breda, The Netherlands).
Granulocyte isolation
Blood was obtained from healthy volunteers. Granulocytes were isolated from a buffy coat of 500 ml of blood by density gradient centrifugation over isotonic Pcrcoll (Pharmacia, Uppsala, Sweden) (39). After lysis of the erythrocytes in the pellet fraction with cold lysis buffer (155 mM NH4CI. 10 mM K.HCO, and
Chapter IV
0.1 mM EDTA. pH 7.4), the granulocytes (> 9 5 % neutrophils) were washed in PBS and resuspended in
HEPES medium ( 132 mM NaCl. 6.0 mM KCl, 1.0 mM C a C k 1.0 mM Mg,S04, 1.2 mM K H:P 04. 20
mM HEPES, 5.5 mM glucose and 0.5% (w/v) HSA. pH 7.4). This fraction is hereafter referred to as neutrophils.
Cell culture and experimental models
Endothelial cells. The human papilloma virus-immortalized HUVEC cell line (40) or freshly isolated, primary HUVECs (41 ) were cultured in HUVEC medium (RPMI 1640 supplemented with 10% (v/v) heat-inactivated human serum. 100 U/ml penicillin, 100 ug/ml streptomycin, 2 mM glutamine and 1 ng/ml bFGF) in culture flasks coated with 1 mg/ml fibronectin. The doubling time of the HUVECs was approximately 48 hours. At confluence, cell suspensions were obtained by trypsin/EDTA treatment. The
2"d-4lh passages of the primary HUVECs were used for sub-culturing on fibronectin-coated polycarbonate
Transwcll filters (3.0 |itn pore size, 12 mm diameter; Costar, Cambridge, MA). 150.000 HUVECs (in 0.5 ml culture medium) were added to the upper compartment and the Transwclls were cultured for another 4 days to obtain confluent HUVEC monolayers.
Epithelial cells. The human lung adenocarcinoma-dcrivcd cell line H292 (American Type Culture Collection CRL 1848) (42) was grown in RPMI 1640 supplemented with 10% (v/v) heat-inactivated human serum, 100 U/ml penicillin, 100 |ig/ml streptomycin and 2 mM glutamine, in uncoated culture flasks. The doubling time of the H292 cells was approximately 24 hours. At confluence, cell suspensions
were obtained by trypsin/EDTA treatment. The 4lh-30lh passages of H292 cells were used for
sub-culturing on the bottom side of Transwell filters according to Parkos et al (43) with minor modifications (26). In brief, a sterile polyoxymethylene polyacetal collar, with an inner diameter equal to the outer diameter of the Transwell insert and with a height of 13 mm, was tightly fixed to the bottom of the Transwcll insert. Subsequently, 80,000 H292 cells (in a volume of 0.5 ml culture medium) were allowed
to attach for 18 hours (5% CO:, 37°C). Thereafter, the collars were removed, and the Transwell inserts
were placed upright in 12-wcll culture dishes and incubated for 5 days.
Primary epithelial cells. Subcultures of primary human bronchial epithelial cells were obtained from bronchial tissues w ith macroscopically normal appearance from patients undergoing lobectomy or pneumcctomy for lung cancer. The cells were cultured in scrum-free keratinocytc medium (Kcratinocyte-SFM. Gibco) with 1 mM isoproterenol (44) in culture flasks coated with 10 ng/ml fibronectin. 30 tig ml vitrogen and 10 jig/ml BSA. After the monolayers had reached confluence, cell suspensions were obtained by mild trypsin/EDTA treatment (Gibco). The detached cells were washed once in PBS containing Soybean trypsin inhibitor type-ll (Sigma) before seeding. The 3rd-4th passages of the bronchial epithelial cells were used for culturing at the bottom side of Transwell filters. The inverted monolayers were created as described abo\ e for H292 cells, except that approximately 200,000 bronchial epithelial cells were added to coated Transwell filters, and cultured in 50% keratinocytc medium and 50% RPMI 1640 supplemented with 2.5 % HSA and 2 mM glutamine (final CaCf concentration 0.5
mM 1(50/50 medium).
Bi-lavcr model. The epithelial cells (75.000 cells well) were allowed to adhere to the bottom side of the Transwell filters as described above. After 1 day (11292) or 5-7 days (primary cells), the top side of the filters was coated w ith I mg ml fibronectin and 150.0(H) III IVECs were seeded. The bi-layers comprised
Migration of Neutrophils across Endothelial and Epithelial Monolayers
of cell line cells were cultured in HUVEC medium, and the bi-layers comprised of primary cells were cultured in 50/50 medium. The bi-layers were cultured for another 4 days to allow formation of confluent monolayers of epithelial and endothelial cells. To confirm confluence, the cells at either side of the Transwell filters were labeled by adding 4 ug/ml calcein-AM (45) in HEPES medium to the lower or upper compartment of the Transwell system. The filters were washed after 15 minutes, mounted on glass slides and inspected by fluorescence microscopy (Dialux, Leit/, Germany). Alternatively, cells on the filters were stained by May-Griinwald/Giemsa. The different monolayers were consistently found to reach confluence within the time frame of culture.
Co-culture model. The primary epithelial cells (75,000 ceils/well) were cultured for 5-7 days in serum-free keratinocyte medium with 1 mM isoproterenol on the bottom of culture plates coated with 10 ug/ml fibronectin, 30 ug/ml vitrogcn and 10 ug/ml BSA. Primary HUVEC (150,000 cells/well) were seeded on fibronectin-coated Transwell inserts and cultured in HUVEC medium for 4 days. The monolayers were subsequently washed and the inserts were placed in the wells with epithelial cells cultured on the bottom. The co-culture was cultured in 50/50 medium for another day. The two cell types were subsequently separated, washed and cultured separately in 50/50 medium for another day.
Electron microscopy
Transwell filters, w ith endothelial cells on the top of the Transwell filter and epithelial cells on the bottom side, were fixed with 2.5% glutaraldehyde (v/v) in 0.1 M cacodylate buffer (pH 7.2) for 1 hr and post-fixed in 1% (w/v) osmium tctroxide in the same buffer for 1 hr. The filters were subsequently block-stained with uranyl acetate, dehydrated and embedded in LX-1 12. Thin sections were block-stained with uranyl acetate and lead citrate and examined with a CM 10 transmission electron microscope (Philips. Eindhoven. The Netherlands).
C 'alcein-AM labeling and transmigration
The endothelial and epithelial cell line monolayers and bi-layers were cultured in HUVEC medium, and the primary endothelial and epithelial monolayers and bi-layers in 50/50 medium. Fresh medium was added to the Transwells 4 hours before the start of the assay. The Transwclls were washed tw ice w ith HEPES medium just before the start of the experiment. Neutrophils ( 10 ml) were labeled with 4 g ml calcein-AM in IIEPES medium for 45 min at 37°C prior to the start of the transmigration assay (45 ). After labeling, the cells were washed twice and resuspended in HEPES medium (final cell concentration 101' ml). Where indicated, neutrophils or monolayers of endothelial or epithelial cells were prctrcated for
15 mm w ith 10 ug'ml antibody to |i:-integnns, CD31 or to CD 14 as a control, followed by washing of the
cells with HEPES medium. Calcein-labeled neutrophils (0.5x 106cells) were placed in the upper
compartment, and chemoattractants were placed in the lower compartment. The chemoattractant concentrations in the lower compartment were PAF. 100 iiM: IMLP. 10 nM: IL-S. 10 nVl: C5a. 10 nM. The Transwells were incubated for 35 min at 37°C.
To quantify transmigration, cells in the upper and lower compartments and cells attached to the filter were separately lysed in lysis buffer (PBS supplemented w ith 0.1% (v/v) Tween-20, 0.2% (w/v) hexadecyl-trimethyl-arnmoniumbromide (Sigma). 0.2 % (w/v) BSA and 20 mM EDTA). The amount of fluorescence in each of these compartments was measured in a spectrofluori meter (Model RF-540.
Shimadzu Corporation. Kyoto, Japan: | x 485 nm; | M 525 nm) and related to the fluorescence of the
Chapter IV IL-6 und ILS quantification
The concentration of IL-6 and IL-8 in the supernatants, collected from both the upper and lower compartments of the different transmigration and co-culture models, was determined by ELISA (CLB) according to the manufacturer's instructions. The absorption was measured in a Multiscan Multisoft microplate reader (Labsystems Oy, Helsinki. Finland) at 450 nm.
Statistical analysis
Results were expressed as the mean + SEM of at least 3 independent experiments, performed with cells from different donors. Results were analyzed with the Student's paired /-test (indicated in the legend of the figures) (46). Two-tailed P values were calculated, and P values exceeding 0.05 were considered not sianificant.
Results
Morphological characterization of the bi-laxer model
We here describe a transmigration model in which sequential migration of neutrophils across human umbilical cord vascular endothelial cells (HUVEC) and human lung epithelial cells was studied (Fig. la). The results obtained with this model were compared to those obtained with either single endothelial or epithelial monolayers. The proper formation of confluent monolayers of
endothelial and epithelial cells on the same filter required careful titration of the number of cells that was seeded. This was especially true for epithelial cells, which showed a tendency to penetrate the pores of the Transwell filters when seeded at high density.
The electron microscopy studies showed that the endothelial as well as the epithelial cells displayed a normal morphology and well-developed cell-cell contacts, and they revealed the striking difference in thickness between the two cell layers (Fig. 1 b). On the basal surface of the epithelial cells hemidesmosomes were formed (Fig. 2) and filamentous structures could be seen between the plasma membrane and the filter, suggesting a basement membrane-like deposition (Fig. 2). Finally, neutrophils migrating through the intercellular junctions of the epithelial cells could also be visualized (Fig. 2).
Immunocytochemical staining for CD31 of the bi-layers. composed of the immortalized endothelial cells and the H292 cell-line epithelial cells, confirmed the localization of CD? I at the cellular junctions of the endothelial cells (47; 48). The staining was restricted to the cells on the topside of the
Transwell filter: i.e. the endothelial cells (data not shown). Similarly, staining of the bi-layers for E-cadherin revealed E-E-cadherin expression at intercellularjunctions of the epithelial cells, present exclusively on the bottom side of the filter (data not shown).
Migration of Neutrophils across Endothelial and Epithelial Monolayers
B
endothelium
epithelium
Figure 1. Schematic and morphological representation of the
bi-layer model, (a) Schematic representation of the bi-bi-layer model,
comprised of endothelial and epithelial cells cultured on the top. respectively, the bottom of a porous polycarbonate membrane.
(b) Morphological characterization of the bi-layer model comprised
of immortalized HUVEC and H292 epithelial cells. Transmission electron micrograph of a Transwell filter carrying the two monolayers. Indicated in the photomicrograph are the (liter (F). the thin monolayer of endothelial cells (en) and, underneath the filter, the relatively thick monolayer of epithelial cell (ep). The epithelial cells are cuboidal and occasionally grow through the pores (p) of the filter. Bar represents 1 urn.
Mai
I
ep
Figure 2. Morphological characterization of
the bi-layer model comprised of immortalized HI 'I EC ' and H292 epithelial cells (a)\ ligli
magnification of a contact area of an epithelial cell with the filter (F) showing
hemidesmosomes {large arrows)
characterized by two electron dense plaques
(small arrows) separated by an electron-lucent
space. Extracellular matrix deposition, represented by the fine filaments
(arrowheads), can be detected in the region
between the plasma membrane and the filters. Bar represents
0.5 urn (/)). An area of the epithelial cell layer (ep) is shown where a neutrophil (n). migrating through the epithelial cell-cell junction, can be seen Arrows indicate neutrophil granules. Bar renresents 1 uni.
Chapter II
Chenioattractant-induced transmigration
Having established the bi-layer model, we tested whether the characteristics of neutrophil migration across the bi-layers were different from those of migration across single monolayers. Transmigration of
neutrophils across single epithelial monolayers, as induced by a series of chemoattractants. was lower when compared to migration across single endothelial monolayers (Fig. 3a). This may be due to the relatively thick and compact epithelial monolayer (Fig. lb), being a more difficult barrier to cross. Interestingly, the percentage of neutrophils that migrated across the bi-layers equaled the percentage of cells that migrated across single endothelial monolayers, despite the presence of the additional epithelial monolayer. This was particularly evident for PAF. C5a and IL-8. FM LP already induced a relatively high migration of neutrophils across epithelial monolayers, and the migration across the bi-layer did not differ significantly from the migration in the other models.
The results obtained with the cell line cells were then compared to those obtained with primary endothelial and primary lung epithelial cells. Here, migration across single epithelial monolayers was equally efficient as migration across single endothelial monolayers (Fig. 3b). This result correlated with the higher basal permeability of primary epithelial cells, when compared to the 11292 cells (data not shown). Moreover, these primary cells appeared less cuboidal and more flattened than the H292 cells and may thus represent less of a barrier for the migrating neutrophils (data not shown). Importantly, similar to
Medium PAF ff.ILP IL
Figure 3. Neutrophil transmigration
in the different models. Migration of
neutrophils across endothelial monolayers cultured on the top of the fillers (open bars), epithelial monolayers cultured on the bottom ol the filter {hatched bars), and bi-layers
{tilled hars) (a)'\ ransmigration across
HUVLC cell line and 1I2')2 epithelial monolayers and bi-layers cultured in III IV|-.( ' medium »as measured inwards medium alone or in response to various chemotactic stimuli Data are mean i SLM of 3-8 independent experiments using cells trom different donors Student's paired /-test was performed to compare transmigration across the bi-layers \\ uh migration across the epithelial monolayers, -k : P<0.05.
(b) I ransmigration across primais
endothelial monolayers, primai") limit epithelial monolayers and bi-layers cultured in 50 50 medium was determined as in a. Data are mean + SIM ol'3-6 independent experiments Student's paired Mest was performed to compare Iransmigration across the bi-layers with migration across the epithelial monolayers, -k : P<0.05.
Migration of Neutrophils across Endothelial and Epithelial Monolayers
neutrophil migration across cell line monolayers and bi-layers, the migration towards PAF. C5a and IL-8 was significantly higher across the primary bi-layers than across primary epithelial monolayers, and again the migration towards flVILP was not significantly elevated across the primary bi-layers (Fig. 3b). These results show that neutrophil migration across bi-layers o f endothelial and epithelial cells, either cell line cells or primary cells, is increased in comparison to the migration across single epithelial monolayers.
The relative increase in neutrophil migration across the bi-layer. as compared to the migration across single epithelial monolayers, was not simply due to increased adhesion. The percentage o f neutrophils associated to endothelial monolayers cultured on filters w i t h a pore size which does not allow passage o f neutrophils (0.45 urn), was similar in the absence or presence o f epithelial cells on the bottom o f the filter (data not shown). The interaction o f neutrophils with the extracellular matrix deposited by endothelial cells onto the filter did not seem to play a role either, because the presence o f a deposited H U V E C matrix or fibronectin coating onto the topside o f the filter did not enhance subsequent transepithelial migration (data not shown).
Role of ßrintegrins and CD3I in neutrophil transmigration
Neutrophil transmigration in the three different models was almost completely CD18 dependent, since CD18-blocking antibodies inhibited more than 9 5 % o f the transmigration across endothelial and epithelial monolayers and across bi-layers (Fig. 4). Pretrcatment o f the endothelial cells w i t h a blocking antibody to ('1)3 I inhibited transmigration o f neutrophils across the bi-layers and across single H U V E C monolayers for 6 4 % and 4 7 % , respectively. The CD31 antibody did not block migration across single epithelial monolayers, which is explained by the absence o f CD31 on epithelial cells. These data show that the formation o f the bi-laver does not significantly alter the role o f prototypic adhesion molecules in neutrophil adhesion and transmigration. When neutrophils were pre-treated w ith the CD31 antibody, transmigration was inhibited and adhesion to the endothelial monolayers was increased (data not shown), possibly due to CD31-mediated activation o f [ i , and/or ß;-integrins on the surface o f neutrophils (22: 49).
100 o c o 60 40 20 Figure 4 \\ as mea: bi-layers aller pret Results a antiboth
flrïi
CD18 CD31. Rti/c for/3-integrins and ÇD3I in neutrophil transmigration. l'A )•-induced neutrophil transmigration ;ured across monolayers of IIUVIIC cell line (open bars), H2()2 epithelial cells {hatchedbars), and the
(filled bars i al'tei pretreatment of neutrophils with an antibod) to CD 18 {monoclonal antibody IB4) or
reatment of endothelial and epithelial cells with an antibody to CD31 (monoclonal antibody I IE( ' I 70). re represented as percentage of inhibition of neutrophil migration in the presence ol an irrelevant to C D I 4 . Data are mean of. respeclhely. 2 and 4 independent experiments.
Chapter 11 '
Role of IL-J ß, IL-6 and IL-8
Endothelial cells and, in particular, epithelial cells arc known to produce cytokines and chemokmes spontaneously, as well as after stimulation with inflammatory cytokines. It is known that 1L-I |i is produced by endothelial and epithelial cells and that it activates cither cell type, resulting in increased expression o f adhesion molecules and production o f various cytokines and chemokmes. e.g. I L - 6 and IL-S (6; 15; 19: 26).
In an initial effort to address the notion that these agents are involved in the efficient
transmigration o f neutrophils across the bi-layers, we measured the concentration o f I L - 1 ß, IL-6 and IL-8 in the culture supernatant by means o f E L I S A . IL-1|) was found in the supernatant o f primary epithelial cells, whereas in the supernatant o f primary endothelial cells hardly any 1L-1[S was detected (Fig 5a). Measurement o f the IL-6 concentration revealed that the supernatant o f monolayers o f primary endothelial cells. H292 epithelial cells and primary epithelial cells contained hardly any IL-6. Yet. the I L - 6 level in the supernatant o f the primary bi-layers was significantly higher (Fig. 6). Moreover, a substantial amount o f I L - 6 was detected in the supernatant o f primary epithelial cells that had previously been co-cultured w i t h primary endothelial cells, whereas the IL-6 level in the endothelial supernatant was low with or
200 150 100 50 epi endothelial cells
1
I endo I epithelial cells co-culture 3 -c 2B
epi endothelial cellsI
I endo I epithelial cells co-culture 70 60 50 40 30 20 10 0C
co-culture epithelial cellsFigure 5. Concentration ofinlerlenkins in the
supernatant ofpriman endothelial or primär) epithelial cells, with or u iilioul preceding co-culture with the other cell type. I he cells «ere first cultured
lor one da\ on filters with or without the oilier cell type on the bottom of the Transwell culture plate.
I'hereaftcr, the cells were refreshed, separated «hen indicated and incubated lor another day. Open bars. culture for one da) w ithout co-culture with the other cell type. I latched bars, culture for one da\ after preceding co-culture w ith the oilier cell type. Black bars, culture for one day alter preceding co-culture I he IL-lß (HI. II -d (b) and IL-8(c) concentration of the supcrnatanls was measured b\ means of f I ISA. The dala are mean ! SI \ l of J independent experiments
Migration oj Neutrophils across Endothelial and Epithelial Monolayers
without previous co-culture with epithelial cells (Fig. 5b). In contrast, the supernatant of H292 epithelial cells did not contain more IL-6 after co-culture with primär) endothelial cells (data not shown). Thus, the augmented IL-6 level in the supernatant of primary bi-layers appears to be due to increased IL-6 production by epithelial cells in response to soluble factors secreted by endothelial cells.
1L-8 was also hardly detectable in the supernatant of endothelial cell monolayers, and the concentration of IL-8 in the supernatant of bi-layers of primary cells was increased significantly in comparison to the concentration detected in monolayers of endothelial and epithelial cells. However, in contrast to IL-6. IL-8 was found in considerable amounts in the supernatant of primary epithelial cell monolayers (Fig. 6). This was not only true for primary epithelial cells (6.8 ng IL-8/ml), but also for H292 epithelial cells (3.5 ng IL-S/ml). Moreover, the IL-8 level in the supernatant of primary epithelial cells that had previously been co-cultured with primary endothelial cells was not elevated, whereas augmented IL-8 levels were detected in the supernatant of primary endothelial cells that had previously been co-cultured with primary epithelial cells (Fig. 5c). Thus, the augmented IL-S level in the supernatant of bi-layers appears to be the result of both the spontaneous epithelial IL-8 production and the stimulated endothelial IL-8 production. Together, these data show that co-culture of endothelial and epithelial cells results in increased production and/or release of cytokines and chemokmes. e.g. IL-6 and IL-8.
Figure 6. Concentration of IL-6 (open bars; and IL-S ("filled bars; in the
supernatant of primary endothelial or primary epithelial cells separately, or in combination, when endured as (i bi-layer. The IL-6 and IL-8 concentration
of the supematants after two days of culture was measured by means of an ELISA. The IL-6 data are mean ± SEM of 4 independent experiments. "I he II -S data are mean ± -SEM of 5 independent experiments. Student's paired t-test was performed to compare the concentration of IL-6 or IL-8 in the supernatant of bi-layers and epithelial monolayers. •*• : P <0.05.
endothelial cells epithelial cells bi-layer
Discussion
Prc\ IOLIS studies on in vitro leukocyte transmigration have relied on models that consisted of a single monolayer of endothelial, epithelial, mesothelial or even (transfected) fibroblast cells, cultured on porous membranes. These models ha\c pro\ ided important experimental e\ idence for the role of cell-adhesion molecules, cytokines, and chemoattractants in leukocyte transmigration (9; 10; 23: 26: 50). However, transmigration across a particular monolayer of endothelial cells is followed in vivo b> passage through the basement membrane and contact w ith a second cell type, e.g. epithelial cells in the lung or stromal cells in the bone marrow In an attempt to mimic this complex in vivo situation, we developed an experimental model to investigate granulocyte migration in the context of lung inflammatory disorders, using monolayers of endothelial and king epithelial cells, both primary cells and cell lines, cultured on opposite sides of the same Transwcll filter.
( 'liapter II
The characterization of this bi-layer model b\ electron microscop) and immunocytochemistry showed that cellular morphology, distribution of the junctional proteins CD31 and E-cadherin, and the integrity of the endothelial and epithelial monolayers was unaltered in the bi-layer model. In addition, the bi-layer barrier w as found to be less permeable to 3 kD FITC-conjugated Dcxtran than the single endothelial and epithelial monolayers (data not shown).
Epithelial cells were found to grow into pores (size 3 urn) of the filters. This phenomenon has also been described for endothelial cells by Mackarcl et al. (51 ). and is likely to be a general phenomenon. We found no indications for specialized structures or basement membranes at the /ones of contact between endothelial and epithelial cells. Moreover, the molecular basis of neutrophil migration in the bi-layer models appeared to be unaltered, i.e. the migration was largely mediated by ('>;-integnns and CD31. Yet, the transmigration of neutrophils in the bi-layer model was more efficient than in the single epithelial monolayer model, and equaled the migration in the single endothelial monolayer models. Several mechanisms may be implicated in this phenomenon.
We tested whether the endothelial monolayer in the context of the bi-layer model would represent a more adhesive surface for neutrophils, as compared to a single endothelial monolayer. However, neutrophil adhesion to the endothelium was not different in the bi-layer model, indicating that increased adhesion docs not occur. Currently, we cannot exclude that qualitative changes in neutrophil adhesion, i.e. the use of additional types of adhesion molecules other than |l:-integnns or CD31. play an important role in the migration in the bi-layer model. Future research will therefore include the analysis of the adhesion molecule repertoire on the endothelial cells in the absence or presence of epithelial cells.
Transendothclial migration may enhance leukocyte motility, thus facilitating subsequent passage across an epithelial monolayer. Such effects may. for example, result from the interaction w ith endothelial CD31, as CD31-mediated interactions have recently been shown to stimulate the rate of integrin-supported neutrophil migration (52). The interaction with the extracellular matrix at the basal side of the endothelial cells could enhance subsequent neutrophil transepithelial migration as well. However, the presence of a matrix deposited by endothelial cells did not alter the migration of neutrophils across the epithelial monolayer.
Finally, the endothelial and epithelial cells may influence each other such that neutrophil transmigration across both monolayers is enhanced, i.e. the epithelial cells may secrete soluble factors that promote neutrophil transendothelial migration, and. vice versa, endothelial cells may secrete factors that promote migration across epithelial monolayers. Our results indeed show that transendothelial migration is increased when epithelial cells arc co-cultured on the bottom of the Transwell culture plate for two days, and that transepithelial migration is increased when endothelial cells are co-cultured. Thus, physical contact between these cell types is not required for the increase in transmigration. Instead a paracrine interaction between the epithelial and the endothelial cells seems to be implicated in the increase of neutrophil transmigration.
Our present results further support this idea; i.e. co-culture of endothelial and epithelial cells dramatically increases the release of particular cytokines, an as yet undescribed phenomenon. For instance, the concentration of IL-6 was significantly and synergistically increased when primary endothelial and epithelial cells were co-cultured, due to stimulated epithelial IL-6 production. IL-6 enhances survival of neutrophils in vitro (53) and has been shown to decrease cell-cell associations of carcinoma cells (54). In addition, IL-8, a potent neutrophil chemoattractant that has been described to be secreted by epithelial and stimulated endothelial cells (35). was found to be significantly and
Migration o] Neutrophils across Endothelial and Epithelial Monolayers
stimulated endothelial IL-8 production. In general, we found that the primai) lung epithelial cells produce more II.-6 and IL-8 than did the H292 cell line cells, an effect that may also contribute to the relatively efficient migration across primary versus H292 epithelial monolayers.
IL-8 release induced by 1L-I has previously been shown to promote migration of neutrophils across monolayers of endothelial and lung epithelial cells (55; 56). In parallel, our data suggest that in the bi-layer model epithelial-derived IL-Iß induces IL-8 production by endothelial cells. Co-culture with epithelial cells increases the endothelial IL-S production, whereas the spontaneous IL-8 production of epithelial cells is unaffected by co-culture with endothelial cells. In addition, epithelial cells spontaneously produce high levels ofIL-1 ß, in contrast to endothelial cells, which hardly produced any IL-1 ß. Moreover, antibodies to IL-lß prevented almost completely the production of IL-8 in epithelial monolayer and bi-layer cultures (data not shown). These data show that IL-1 ß regulates epithelial IL-S secretion via an autocrine loop. Whether epithelial cell-derived IL-lß is indeed the initiating factor in the stimulated neutrophil migration across the bi-layers remains to be established.
The enhanced release of IL-6 and IL-8 may contribute to the increase in neutrophil migration across the bi-layer. This may occur as a result of enhanced Chemotaxis, but may also involve cvtokinc-mediated changes in the endothelial or epithelial monolayers. Regardless the mechanism involved, a strong chcmotactic stimulus was still required for neutrophil transmigration, as the spontaneous migration across the bi-layers remained as low as across the single monolayers.
Recently, a similar transmigration model for neutrophils was described, combining HUVECs with the A549 lung epithelial cell line (57). Although the experimental set-up of this work is similar to ours, these studies did not address the relative role for adhesion molecules in the migration or reveal any paracrine communication, involving cytokines or chemokines. between epithelial and endothelial cells. Moreover, in this model the migration across the bi-layers was not increased, when compared to migration across the individual monolayers. The differences with our findings may be related to the alveolar, rather than bronchial, origin of A549 tumor line, and these cells may behave differently with respect to production and release of soluble mediators.
In conclusion, our current results with the bi-layer model provide new insights in the molecular basis of neutrophil transmigration. This model adds new aspects to the research on leukocyte migration in king inflammatory disorders by providing an extra level of complexity, i.e. the "cross-talk' between monolayers of different cell types and the concomitant effects on leukocyte passage. These interactions between endothelial and epithelial cells are likely to be relevant for inflammatory disorders in the lung, and may also play a role in other tissues where endothelial and epithelial cell linings are in close proximity.
Footnote
( 'liapler IV
References
1 Moser.R., Fehr.J.. Olgiati.T., and Bruijn/eel.P.L.B. 1992. Migration of primed human eosinophils across cytokine-activated endothelial cell monolayers. Blood 79:2937.
2 Aalbers.R., De Monchy.J.G., Kauffman.H.F., Smith,M.. Hoekstra.Y., Vrugt.B., and Timens.W. 1993 Dynamics of eosinophil infiltration in the bronchial mucosa before and alter the late asthmatic reaction.
Eur.Respir.J. 6:840.
3. Lukacs.N.W., Strieter.R.M., and Kunkcl.S.L. 1995. Leukocyte infiltration in allergic airway inflammation. Am.J.Respir.Cell.Mol. Biol 13:1.
4. KnoI.E.F. and Roos.l). 1996. Mechanisms regulating eosinophil extravasation in asthma. Eur. Respir.J. 9:136s.
5. Roman,J1996. Extracellular matrix and lung inflammation. Immunol.Res. 15:163.
6. Kuijpers.T.W.. H a k k e r t . B . C . Hart.M.H.I... and Roos.l). 1992. Neutrophil migration across monolayers of cytokine-prestimulated endothelial cells: A role for platelet-activating factor and IL-S. J.Cell Biol
117:565.
Rot.A.. Krieger,M. Brunner.T.. Bischoff.S.C, Schall.T.J.. and Dahinden.C.A. 1992 R A N U S and macrophage inflammatory protein l a induce the migration and activation of normal human eosinophil granulocytes. J.Exp.Med. 176:1489.
8. Ebisawa.M., Yamada.T.. Bickel.C, Klunk.D., and Schleimer.R.P. 1994. Eosinophil transendothelial migration induced by cytokines: 111. Effect of the chemokine RANTES. J.Immunol. 153:2153 9. Springer.T.A. 1994, Traffic signals for lymphocyte recirculation and leukocyte emigration: the multislep
paradigm. Cell 76:301.
10. Bianchi.E.. Bender..!.R. Blasi.F.. and I'ardi.R. 1997. Through and beyond the wall late steps in leukocyte transendothelial migration. Immunol. Today 12:586
1 1. 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.
12. Kitayama.J., Mackav.C.R.. Ponath.P.I).. and Springer.T.A. 1998. The C-C chemokine receptor CCR3 participates in stimulation of eosinophil arrest on inflammatory endothelium in shear How. J.Clin.Invest. 9:2017.
13. Lampugnani,lVl.G. and Dejana.E. 1997 Interendothelial junctions: Structure, signalling and functional roles CuiT.Opin.Cell Biol. 5:674.
14. McNulty.C.A., Symon.F.A., and Wardlaw,A.J. 1999. Characterization of the integrin anil actuation steps mediating human eosinophil and neutrophil adhesion to chronically inflamed airway endothelium
AmJ.Respir.Cell. Mol.Biol. 20:1251.
15. Liu.l... Mul.F.P.J. Kuijpers.T.W.. Lutter.R., Roos.l).. and Knol.E.F. 1996 Neutrophil transmigration across monolayers of endothelial cells and airway epithelial cells is regulated by different mechanisms.
Ann.NYAcad.Sci. 796:21.
16. Ward.P.A. 1996. Role of complement, chemokines, and regulatory cytokines in acute lung injury. Ann.N)
Acad.Sei. 796:104.
17. Rollins.BJ. 1997. Chemokines. Blood 3:909.
18 Luscinskas.F.W.. Cubulsky.M.I.. K i e l y J . - M . , Peckins.C.S.. Davis.V.M.. and Cimbrone.M.A. 1991 Cytokine-activated human endothelial monolayers support enhanced neutrophil transmigration \ ia a mechanism invoK ing both endothelial-leukocyte adhesion 1 and intracellular adhesion
molecule-I. J.Immunol. 146:1617.
19. Hakkert.B.C'., Kuijpers.T.W., Leeuwenberg.J.F.M., Van Mourik.J.A., and Roos.D. 1991. Neutrophil and monocyte adherence to and migration across monolayers of cytokine-activated endothelial cells: the contribution ofCD18, E L A M - 1 , and VLA-4. Blood 78:2721.
20. Muller,W.A.. Weigl.S.A.. Deng.X.. and Phillips.D..M. 1993. PECAM-1 is required for transendothelial migration of leukocytes. J.Exp.Med. 178:449.
21 Cooper,!).. Lindberg.F.P.. Gamble.J.R.. Brown.E.J.. and Yadas.M.A. 1995 Transendothelial migration of neutrophils involves integrin- associated protein (CD47). Proc.Natl.Acad.Sci.l 'SA 92:3978.
22. Berman.M.E.. Xie.Y.. and Muller.W.A. 1996. Roles of platelet endothelial cell adhesion molecule-l (PECAM- 1, CD31) in natural killereell transendothelial migration and b: integrin actuation. J.Immunol.
156:1515.
23 Zocchi.M.R., Ferrero.E., Leone.B.E., Rovcre.P., Bianehi.F... Toninelli.F.. and Pardi.R.1996.
CD31/PECAM-1-driven chemokme-mdependent transmigration of human T lymphocytes. Eur.I.Immunol. 26:759.
Migration oj Neutrophils «< ross Endothelial unci Epithelial Monolayers
24 Christofldou -Solomidou,M., Nakada.M.T., WilliamsJ., Muller.W.A-, DeLisser,H.M. 1997 Neutrophil platelet endothelial cell adhesion molecule-1 participates in neutrophil recruitment at inflammatory sites and is down-regulated after leukocyte extravasation J.Immunol 158:4872.
25. Yong,K.L., \\ atts.M.. Shaun Thomas.N.. Sullivan,A., Ings.S., and I.inch.I).('. 1998. Transmigration of CD34-I cells across specialized and nonspecialized endothelium requires prior activation by growth factors and is mediated b> PKCAM-1 (CD31 ) Blood 4:1 196.
26. Liu.I... Mul.F.P.J.. Lutter,R., Roos.I)., and Knol.K.F. 1996. I ransmigration of human neutrophils across lung epithelial cell monolayers is preferentially in the physiological basolateral to apical direction.
AmJ.Respir.Cell.Mol.Biol. 15:771.
27. Parkos.C.A. 1997. Molecular events in neutrophil transepithelial migration. BioEssays 19:865. 28. Burke-Gaffney,A. and Hellewell,P.G. 1998. A CD18/ICAM-1 -dependent pathway mediates eosinophil
adhesion to human bronchial epithelial cells. AmJ.Respir.Cell.Mol.Biol. 19:408.
29 Cunningham.A.C. and Kirby.J.A. I 995. Regulation and function of adhesion molecule expression by human alveolar epithelial cells. Immunology 86:279.
30. Simon, R.H. and Paine III. K. 1995 Participation of pulmonary alveolar epithelial cells in lung inflammation. J.Lab.Clin.Med. 126:108.
3 1 Bloemen.P.G.M.. Van den Tweel.M.C., Henricks.P.A.J.. Engels,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. Am.I.Physiol.Lung Cell.Mol.Physiol. 270:L80.
32. Sheppard.I). 1996. Epithelial integrins. BioEssays 18:655.
33. Stellato C . I..A. Beck. O.A. Gorgone, ». Proud. T.J. Schall. S.J. Ono. L.M. Lichtenstein, and R.P. Schleimer. 1995 Expression of the chemokine RANTES by a human bronchial epithelial cell line: Modulation by cytokines and glucocorticoids. J.Immunol. 155:410-418.
34. Terajima M.. M. Yamaya, K. Sekizawa, S. Okinaga, T. Suzuki, N. Yamada, K. Nakayama,. T. Ohrui, T. Oshima, Y. N'umazaki. and H. Sasaki. 1997. Rhinovirus infection of primary cultures of human tracheal epithelium: role of ICAM-1 and II.-lb Am.J.Physiol.Lung Cell.Mol.Physiol. 273:L749-I.759.
35. Van der Velden V.H.J., H.F.J. Savelkoul, and M.A. Versnel. 1998. Bronchial epithelium: morphology, function, and pathophysiology in asthma. Eur.Cytokine Netty 9:585-597.
36. Wright S.I).. P.E. Rao. W.C. van Voorhis, I..S. Craigmyle, K. lida. M.A. Talie, E.F. Westberg, G. Goldstein, and S.C. Silveistein. 1983. Identification of the C3bi receptor of human monocytes and macrophages by using monoclonal antibodies. Proc.Natl.Acad.Sci.USA 80:5699-5703.
37. Van Mourik J.A.. O.C. Leeksma, J.H. Reinders, P.G. de Groot, J. Zandbergen-Spaargaren. 1985 Vascular endothelial cells synthesize a plasma membrane protein indistinguishable from the platelet membrane glycoprotein IIa. J.Biol.Chem. 20:1 1300-11306.
38. Van H.C.. J..M. Pilevvski. Q. /.hang. U.M. DeLisser, L. Romer, and S.M. Albelda.1995. Localization of multiple functional domains on human PECAM-1 (CD31 ) by monoclonal antibody epitope mapping. Cell
Adhes.Commun. 3:45-66.
39. Roos I).and M. de Boer. 1986 Purification and cryopreservation of phagocytes from human blood. In:
Methods Enzymol. Volume 132. Immunochemical Techniques Part J Phagocytosis and Cell-Mediated
Cytotoxicity ((i Hi Sabato ami J Everse, eds.), Academic Press. New York 225.
40. Fontijn R.. C. Hop. H.J. Brinkman. R. Slater. A. Westerveld, J.A. van Mourik. and H. Pannekoek. 1995. Maintenance of vascular endothelial cell-specific properties after immortalization with an amphotrophic replication-deficient retrov irus containing human papilloma virus 16 E6/E7 DNA. Exp.Cell
Res. 216:199-207.
41. Brinkman,H.J.. Mertcns.K., Holthuis..!.. Zwart-HuininkX.A., Grijm,K., and Van Mourik.J.A. 1994 fhe activation of human blood coagulation factor X on the surface of endothelial cells: a comparison with various vascular cells, platelets and monocytes. Br J. Haematol. 2:332.
42 Van SchilfgaardcM., Van Alphen,L., Eijk.P., Everts,V., and Dankert.J. 1995. Paracytosis of Haemophilus influenzae through cell layers of NCI- H292 lung epithelial cells. Infect.Immun. 63:4729 43. Parkos,C.A., Delp.C. Arnaout.M.A., Madara.J.L. 1991. Neutrophil migration across a cultured intestinal
epithelium. Dependence on a CD11 b/CDI 8-mediated event and enhanced efficiency in physiological direction. J.Clin.Invest. 88:1605.
44 Bloemen.P.G.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.
45. Akeson A.L. and Woods.C.W. 1993. A fluorimetric assay for the quantification of cell adherence to endothelial cells. J.Immunol.Methods 2:181.
46. Altman,D.G. 1991. Comparing groups - continuous data. In: Practical statistics for medical research. Chapman & Hall, London. UK 179.
Chapter II'
47. Ayalon.O.. Sahanai.H.. l.ampiignani.M.G.. Dejana.E.. and Geiger.B. 1994. Spatial and temporal relationships between Cadherins and PECAM-1 in col I-cell junctions of human endothelial cells. J. Cell Biol.
1:247.
48. Albelda.S.M.. Muller,W.A., Buck.C.A.. and Newman.P.J. 1991. Molecular and cellular properties of PECAM-1 (endoCAM/CD31 ): a novel vascular cell-cell adhesion molecule. J.Cell Biol. 5:1059. 49. Bcrman.M.E. and Muller.W.A. 1995 ligation of platelet/endothelial cell adhesion molecule 1
(PECAM-1/CD31 ) on monocytes and neutrophils increases binding capacity of leukocyte CR3 (CD1 lb V I ) 18).
J.Immunol. 154:299.
50. Zeillemaker.A.M., Mul.F.P.J., Hoynck van Papendreeht.A.A.G.M., Kuijpcrs.T.W., Roos.I).. LeguitP., Verburgh,H.A. 1995. Polarized secretion of interleukin-8 by human mesofhelial cells: A role in neutrophil migration. Inwiiinolog) 84:227.
51. Mackarel.A.J.. Cottell,D.C. Russcll.K.J.. Fitzgerakl.M.X., and O'C'onnor.C.M. 1999 Migration of neutrophils across human pulmonary endothelial cells is not blocked by matrix metalloproteinase or serine protease inhibitors. Am.J.Respir.Cell.Mol Biol. 20:1209.
52. Rainger.G.E.. Buckley.C. Simmons.D.I... and Nash.G.B. 1997. Cross-talk between cell adhesion molecules regulates the migration velocity of neutrophils. Curr.Biol. 7:316.
53. Daffern.P.J.. Jagels.M.A., and Hugli.T.E. 1999. Multiple epithelial cell-derived factors enhance neutrophil survival. Am.J.Respir.Cell.Mol.BioI. 21:259.
54. Tamm.l.. Cardinale.!., Krueger,J„ Murphy.J.S., May.L.T, and Sehgal.P.B. 1989. Interleukin 6 decreases cell-cell association and increases motility o I'ductal breast carcinoma cells. J.Exp.Med. 170:1649. 55. Bittleman.D.B. and ('asale.T.B. 1995. Interleukin-8 mediates interleukin-la-induced neutrophil
transcellular migration. Am.J.Respir.Cell.Mol.BioI. 13:323.
56. Carolan.E.J. and ('asale.T.B. 1996. Neutrophil transepithelial migration is dependent upon epithelial characteristics. Am.J.Respir.Cell.Mol.BioI. 15:224.
57. ('asale.T.B. and Carolan.E.J. 1999. Cytokine-induced sequential migration of neutrophils through endothelium and epithelium. Inflamm.Res. 1:22.
