Published ahead of print on January 7, 2005, doi:10.1165/rcmb.2004-0187OC
© 2005 American Thoracic Society DOI: 10.1165/rcmb.2004-0187OC A Three-Dimensional Cellular Model of the Human Respiratory Tract to Study the Interaction with ParticlesInstitute of Anatomy, Division of Histology, University of Bern, Bern, Switzerland; and Department of Veterinary Anatomy, University of Nairobi, Nairobi, Kenya Correspondence and requests for reprints should be addressed to Barbara M. Rothen-Rutishauser, Ph.D., Institute of Anatomy, Division of Histology, University of Bern, Baltzerstrasse 2, CH-3000 Bern 9, Switzerland. E-mail: rothen{at}ana.unibe.ch
A novel triple co-culture model of the human airway barrier was designed to simulate the cellular part of the air–blood barrier of the respiratory tract represented by macrophages, epithelial cells, and dendritic cells. When epithelial cells (A549 cells) were grown on filter inserts with pores of 3.0 µm in diameter in a two-chamber system, they formed monolayers with polarization into apical and basolateral domains. The epithelial cell cultures were then supplemented with human blood monocyte–derived macrophages and dendritic cells on the apical and basal aspect, respectively. The single-cell cultures as well as the triple co-cultures were characterized in terms of a number of typical features, for example, morphology of cell types, integrity of epithelial layer, and expression of specific cell surface markers (CD14 for macrophages and CD86 for dendritic cells). The interplay of epithelial cells with macrophages and dendritic cells during the uptake of polystyrene particles (1 µm in diameter) was investigated with confocal laser scanning and conventional transmission electron microscopy. Particles were found in all three cell types, although dendritic cells were not directly exposed to the particles. More investigations are needed to understand the translocation pathway.
Key Words: epithelial cells • macrophages • dendritic cells • co-culture • particle uptake
The wall of the respiratory system consists of a cascade of barrier components that protect against foreign material from gaining access to the lungs of the organism and reaching the immunocompetent cells. These include the surfactant film (1–3), the mucociliary system (4), highly phagocytic airway macrophages (AMs) (5, 6), and the epithelium with its tight junctions (7). However, despite those barriers, respiratory diseases are frequent and increasing (8) and more attention is being directed toward elucidating how and when the antigens evade these barriers. Insoluble particles deposited on the airways are largely cleared by the mucociliary action. However, not all particles deposited on the airways are removed by this mechanism. The fate of the deposited particles depends on the physical-chemical characteristics of the particles and the nature of their interaction with the surfactant film at the air–liquid interface. Particles deposited on the airways are displaced into the subphase below the surfactant film and may be coated with surfactant or surfactant components during the displacement process. As a result of the displacement, particles come into contact with the epithelium and AMs (2, 9, 10). Of particular importance and interest is how the antigens can reach the dendritic cells (DCs), which are present at the base of the epithelium and are the most competent antigen-presenting cells in the lung (11, 12). In vivo, AMs occupy the luminal aspect of the epithelium (6, 13), whereas immature DCs occupy the basal aspect of the epithelium, laying within the basement lamina and reaching to the tight junctions with fine cytoplasmic processes between the epithelial cells (14) (Figure 1).
By use of confocal laser scanning microscopy (CLSM) and transmission electron microscopy (TEM), it has been demonstrated that DCs are efficiently phagocytic for a variety of particles, such as polystyrene particles (15–17), puff ball spores, biodegradable microspheres (18), and Salmonella typhimurium (19). Although both AMs and DCs are derived from circulating blood monocytes, AMs are twice as phagocytic as immature DCs (17). Whereas much is known on the interaction of particles with DCs and with AMs, nothing is known about how the antigens reach the DCs or if AMs and epithelial cells are involved. Transport of the particles to the DCs presupposes their passage across the epithelium either through the epithelial cells or between the epithelial cells, i.e., through the tight junctions. However, the route they take is not yet known. It has been shown in an in vitro study with enterocytes that DCs send dendrites outside the epithelium, where they directly sample bacteria (20). Vermaelen and colleagues (21) recently found in an in vivo study that fluorescein isothiocyanate–conjugated macromolecules are transported to the tracheal lymph nodes by airway DCs after intratracheal instillation. Another in vitro model using mouse tracheal epithelial cells and mouse bone marrow DCs showed impaired migration of metalloproteinase-9–deficient DCs through tracheal epithelial tight junctions (22). Again, however, the mechanism with which the macromolecules passed through the epithelium to reach the DCs was not addressed. To study the interaction of particles that have been deposited at the airway wall with the cells of the human airway barrier, an in vitro model simulating the cellular airway barrier was generated and characterized. A triple co-culture model system composed of cuboidal epithelial cells, AMs, and DCs was established and evaluated in terms of its functional relevance to the in vivo tissue. For this purpose the human (alveolar) epithelial cell line A549, which originated from human lung carcinoma (23), was chosen. The A549 cell line is routinely used as an in vitro model of pulmonary cuboidal epithelium to study the interaction of environmental particles (24) and ultrafine particles (25) with the cells. Bilayer models with A549 and endothelial cells to study bacteria infections have also been described (26, 27), as well as co-cultures of A549 cells with an alveolar macrophage cell line to study the effect of lung surfactant phospholipids during the uptake of microspheres (28). In our study human AMs and DCs derived from human blood monocytes (29) were combined with A549 cells. The cultures of single cells as well as the triple co-cultures were characterized in terms of their typical features, e.g., morphology of the cells, integrity of the epithelial layer, and expression of specific cell surface markers (CD14 for AMs and CD86 for DCs). The interplay of epithelial cells with AMs and DCs during the uptake of polystyrene particles was investigated with CLSM and TEM.
A549 Cultures The A549 cell line (23) was obtained from American Tissue Type Culture Collection (LGC Promochem, Molsheim, France). Cells (passage number 10–70) were maintained in RPMI 1640 medium (w/25 mM Hepes; LabForce AG, Nunningen, Switzerland) supplemented with 1% L-Glutamine (LabForce AG), 1% penicillin/streptomycin (Gibco BRL, Invitrogen AG, Basel, Switzerland), and 10% fetal calf serum (PAA Laboratories, Lucerna-Chem AG, Lucerne, Switzerland). Cells were seeded at a density of 0.5 x 106 cells/ml on BD Falcon cell culture inserts (surface area of 4.2 cm2, pores with 3.0 µm in diameter, high pore density PET membranes for 6-er well plates; BD Biosciences, Basel, Switzerland). Inserts were placed in BD Falcon tissue culture plates (6-er well plates; BD Biosciences) with 2.5 ml medium in the upper and 3 ml in the lower chamber. Medium was changed twice weekly. Growth curves were determined by trypsinisation of the cells from the filter insert with trypsin-EDTA (Gibco BRL) and counting them with a Neubauer counting chamber.
AM and DC Cell Culture
Triple Cell Co-Cultures
DCs were harvested, suspended in 2 ml RPMI 1640, and 500 µl of the cell suspension was added to the basal side of the inserts turned upside down (Figure 2B). The petri dishes were covered and placed for 2 h into the incubator. Afterwards, the nonadherent DCs were removed and the inserts were placed back into the tissue plates. A quantity of 2.5 ml RPMI 1640 supplemented with 1% L-Glutamine, 1% penicillin/streptomycin, and 5% heat inactivated (pooled) human serum was added to the lower chamber. AMs were harvested, suspended in 2 ml RPMI 1640, and 500 µl of the cell suspension was added on the apical surface of the epithelial monolayer (Figure 2C). Cells were allowed to attach for 2 h, nonadherent cells were washed away, and 2 ml of RPMI 1640 supplemented with 1% L-Glutamine, 1% penicillin/streptomycin, and 5% heat-inactivated (pooled) human serum was added to the upper chamber. The triple co-cultures were kept for 24 h at 37°C in a 5% CO2 humidified atmosphere.
Particle Uptake
Transepithelial Electrical Resistance Measurements
Mannitol Transport Studies
Cell Labeling and Fixation Antibodies were diluted as follows in PBS: rabbit anti-human occludin 1:20 (71–1,500; Zymed, P. H. Stehelinand Cie AG, Basel, Switzerland), rabbit anti-human claudin-2 1:50 (51–6,100; Zymed), mouse anti-human E-cadherin 1:300 (Clone HECD-1, 13–1,700; Zymed), rabbit anti-human ZO-3 1:100 (36–4,100; Zymed), mouse anti-human CD14 1:20 (Clone UCHM-1, C 7673; Sigma), mouse anti-human CD86 1:20 (Clone HB15e, 36931A; PharMingen, BD Biosciences), goat anti-mouse cyanine 5 1:50 (AP124S; Chemicon, VWR International AG, Life Sciences), goat anti-rabbit cyanine 5 1:50 (AP187S; Chemicon), goat anti-mouse R-Phycoerythrin 1:20 (P 9670; Sigma), and rhodamine phalloidin 1:100 (R-415; Molecular Probes, Invitrogen AG, Basel, Switzerland). As a control, the specificity of the antibodies and the labeling procedure were tested with the secondary antibodies only.
Confocal Microscopy and Image Restoration
TEM
Characterization of the Cell Cultures The two-chamber cultures allow free access to two chambers that are separated by the epithelial layer. An advantage of this two-chamber system is the possibility of monitoring the integrity of the epithelial layer during the course of an experiment by measuring TEER (31). The cell number doubled and then leveled off within 5 d (Figure 3A). TEER measurements were performed at various times (from 4 h after seeding to Day 12) and the values remained stable from Days 3–12 (between 140 and 180  cm2) (Figure 3B). All experimental cell cultures were taken at Day 7 in culture (Figures 3A and 3B, arrows). The expression of adherens junctions (E-cadherin) as well as of the tight junctions was also confirmed by immunostaining of the transmembrane tight junction proteins occludin and claudin-2 (Figure 3C). The presence of the peripheral tight junction protein zonula occludens (ZO)-3 was also shown (Figure 3C). The staining for the adherens and tight junction proteins has been verified in three independent experiments. The morphology of A549 cells was studied with cells grown for 7 d on filter inserts in the two-chamber system, fixed, and stained for F-Actin. The A549 cells formed monolayers (Figures 4A and 4C). In long-term cultures, epithelial cells grown in a monolayer may traverse the membrane through the 3.0-µm pores and grow also at the bottom of the membrane. This phenomenon was also observed in our cultures, where a thin epithelial layer at the bottom of the membrane insert was seen (Figure 4B). The epithelial cells at the bottom were removed by turning the insert and abrading them with a cell scraper. The removed cells were pipetted away and the inserts were placed back into the tissue plates. Fresh medium was afterwards added to both chambers. CLSM pictures of actin-labeled cells showed no disturbance of the monolayer on the upper side of the insert (Figure 4C). Some pseudopods in the pores could still be observed (Figure 4D, arrows). In addition, TEER measurements before (146.4 [SD 18.2] cm2) and after removing cells from the basal side of the inserts (159.3 [SD 17.7] cm2) showed no difference, indicating that the integrity of the upper epithelial layer was not affected.
The maturation state of AMs and DCs can typically be characterized by the expression of specific surface molecules. The surface antigen expression was analyzed by immunostaining methods in combination with CLSM. CD14 was expressed on the surface of AMs (Figure 5A) and CD86 on the surface of DCs (Figure 5D). Three-dimensional reconstructions of the confocal data showed a round phenotype of AMs and DCs (Figures 5A and 5D, insets). To see if maturation can be induced, both cell types were incubated with lipopolysaccharide (LPS). The surface expression of the CD14 in AMs (Figure 5B) and CD86 in DCs (Figure 5E) was not different if compared with nontreated cells, but the shape changed from a round (Figures 5A and 5D, insets) to a more stretched form, and in addition many pseudopods have been formed (see insets in Figures 5B and 5E). Cells were incubated with fluorescently labeled polystyrene particles with a diameter of 1 µm to investigate the influence of particle uptake on the maturation of AMs and DCs. Cells were fixed and stained for the different surface markers afterwards. Uptake of particles was seen in AMs with CD14 expression at the surface (Figure 5C) and in DCs with CD86 expression at the surface (Figure 5F). The shape of DCs changed to a more stretched phenotype with thin pseudopods (Figure 5F, inset).
Evaluation of the Triple Cell Co-Cultures After 24 h of incubation the co-cultures were fixed and embedded in Epon. Ultrathin sections were observed under the TEM (Figure 6). The A549 cells formed a polar monolayer expressing tight junctions (Figure 6A, inset arrowhead) and lamellar inclusions (Figure 6). Cellular protrusions grew through the pores (Figure 6A, black arrow). The AMs were localized at the top of the epithelia, with their characteristic morphology. DCs were seen at the basal side of the insert, having a more stretched appearance. Epithelial cells touched the DCs via pseudopods without developing a junctional complex (Figures 6B, 6C, and 6D, white arrows).
To identify clearly AMs and DCs in the triple co-culture system, the expression of specific markers was also investigated with CLSM. For visualization of the cells, their actin-containing cytoskeleton was stained with rhodamine-labeled phalloidin. After 24 h in co-culture the cells were fixed, and for the immunostaining the inserts were cut with a scalpel into two pieces. One piece was turned upside down to allow the investigation of the lower cells. As it can be shown in z-projections as well as in three-dimensional reconstructions, CD14 positive AMs (Figures 7A and 7A', arrows) were localized on top of the epithelial monolayer. DCs at the lower surface of the insert expressed CD86 in the co-culture system as shown in Figures 7B and 7B' (arrowheads). DCs formed many pseudopods and were obviously more numerous as compared with AMs.
Use of the Triple Cell Co-Cultures to Study Particle–Cell Interactions It was of particular interest to study how particulate antigens deposited on the epithelium may reach the dendritic cells. The interaction of 1 µm fluorescently labeled polystyrene particles with the triple cell co-culture system was studied. Special attention was paid to the integrity of the epithelial layer, and during incubation with 1 µm polystyrene particles TEER measurements were performed at various times. Addition of particles did not influence the tight junctions when compared with control cultures (Figure 8). A control experiment was performed by addition of EDTA to a triple cell co-culture to open the tight junctions (30). This revealed a decrease of TEER to 28 cm2 within 2 h (Figure 8), whereas the control cultures and cultures treated with particles showed TEER values within a range from 120–190 cm2. The integrity of the cell monolayers during incubation with particles was further investigated using the paracellular marker 14C-mannitol (Table 1). The Papp value of A549 cells alone was 2.25 ± 0.4 x 10–6 cm/s, of co-cultures 0.93 ± 0.2 x 10–6 cm/s, and of co-cultures incubated for 24 h with 1-µm particles 1.35 ± 0.37 x 10–6 cm/s. As a control the Papp for MDCK cells, a well-described epithelial cell culture model (31) was determined and was 1.41 ± 0.13 x 10–6 cm/s.
To investigate particle uptake by and the interplay of the different cell types, CLSM was used in combination with immunofluorescence methods. After incubation with fluorescently labeled 1-µm particles, cells were fixed and stained for F-Actin in addition to the specific surface markers. After 4 h, most of the particles were attached to the surface of CD14-positive AMs (Figure 9A, arrowhead), only some particles were found within the cells (Figure 9A, arrow). However, when AMs were incubated with particles for 24 h, most of the particles were seen inside the cell (Figure 9B). After this period of time particles were also found in epithelial cells (Figure 9B, arrowhead), but never to the same extent as in AMs. No particles were seen between epithelial cells. Particles were also found at the bottom side of the membrane insert attached or within cells. Only few particles were seen attached to the surface of CD86-positive DCs (Figure 9C, arrowhead) after 4 h. However, after 24 h also DCs were filled up with particles (Figure 9D, arrow). In addition to CLSM analysis, co-cultures were fixed after 24 h of incubations with particles and embedded in Epon for TEM. It was confirmed that particles were within all three cell types (Figures 9E and 9F, arrows). Particles were found within the pseudopods of epithelial cells in the pores (Figure 9F, arrows), but also in TEM images, particles were never observed between epithelial cells.
AMs and DCs contribute to the airway epithelial barrier. AMs are found on the apical aspect of the epithelium and the DCs on the basal side. The mechanisms through which particles break the macrophage and epithelial barrier to reach the DCs is yet to be fully elucidated.
Most studies aimed at resolving this question have applied animal models. However, this approach is expensive and time consuming and does not give insights into the individual steps easily, in particular into particle–cell interactions. Therefore the use of in vitro models with human cells may give first insights into the mechanisms of particle–cell interaction at the airway epithelial barrier. We have chosen the A549 human epithelial cell line, which was proposed to be a standardized model to study lung epithelium (32). It has been shown that the A549 cells have many important biological properties of alveolar epithelial type II cells, such as membrane-bound inclusions, which resemble lamellar bodies of type II cells (33). Other ultrastructural characteristics common to type II cells have also been described, as for example distinct polarization, tight junctions, and extensive cytoplasmic extensions (25). In our study the A549 cells were grown on membrane inserts for 7 d with 3.0-µm pores in a two-chamber system where they formed a polarized monolayer. The expression and typcial localization of E-cadherin, a tissue-specific protein expressed at sites of cell–cell contacts and which is important for the formation of polarized epithelia (34), was verified. However, in a recent study it was reported that epithelial cell lines, as for example the A549 cell line, failed to develop TEER (35). This is in contrast to our studies, where the cells developed a TEER plateau of So far, in vitro co-cultures with two cell types have been described in literature. One model works with epithelial cells (A549 cells) and AMs to study effects of inhaled microbes (40). Two cell co-culture models with epithelial (A549 cells) and endothelial cells were established to examine events in the pathogenesis of bacteria (26, 27). To investigate the mechanisms of antigen delivery to the DCs, a co-culture system with epithelial cells and DCs has been used (20). This was a human enterocyte cell line combined with mouse DCs. However, in our study we used cells of human origin only. In this article a co-culture consisting of three cell types is described. It consists of epithelial cells together with blood monocyte–derived AMs and DCs in a two-chamber system. The AMs were located on top of the epithelial monolayer and the DCs were at the basal side of the insert, separated from the epithelial cells by the insert containing 3-µm pores. TEM analysis revealed that the DCs and epithelial cells made direct contacts by sending out protrusions through the insert pores, but did not develop cell junctions to the best of our knowledge. Particles that are deposited on the cellular airway barrier may be taken up by AMs and epithelial cells. The uptake of particles of 1 µm in diameter by AMs is known to occur by phagocytosis (41) that has been confirmed by our studies. We have shown that polystyrene particles of this size are taken up by AMs in single cell cultures as well as in the triple co-culture system. There seems to be a dose–response relationship: particles that are not removed by the AMs are likely to be taken up by epithelial cells (42). However, only little information is available on particle uptake by these cells. In our system, particles were found in A549 single cell cultures (data not shown) and also in the epithelial cells of the co-culture system. Particle uptake was time-dependent (because after 24 h more particles were found in AMs as well as in epithelial cells than after 4 h) and also cell type–dependent. We never observed as many particles in the epithelial cells as in AMs. This was an obvious but only qualitative observation; it is planned, however, to count the particles in the different cell types in future experiments. Particles were also observed in DCs in single-cell cultures, and they were seen in the triple-cell co-culture system, although the DCs were not directly exposed to particles in the latter system. Because the TEER did not decrease during the experiment, we assume that the tight junctions were not opened after addition of particles to the medium. This was also confirmed by permeation studies using 14C-mannitol, a substance that is routinely used as a paracellular marker. Compared with the Papp for MDCK cells (which was 1.41 ± 0.13 x 10–6 cm/s, a value that has already been described in the literature [43]), we suggest that in our cultures the A549 cells (alone or in the co-culture system) build a tight monolayer and that incubation with particles do not cause any changes in the properties of tight junctions. In the literature, the Papp values for mannitol in A549 cells are usually higher by a factor of 100 (43). However, as already seen with the TEM and CLSM images the tight junctions are very well organized, which is in agreement with the TEER measurements and the Papp values. The mechanism described by Rescigno and coworkers (20) is that DCs open the tight junctions between epithelial cells and push their dendrites through into the luminal aspect to directly take up bacteria. We did not observe any such process in our triple-cell co-culture system and we never observed particles between the epithelial cells, which suggests that other mechanisms must exist, through which particulate matter gains access to DCs. At this point, we postulate that particles are transported to the DCs via the epithelial cells because we found particles in them. The delivery of particles by AMs through the epithelial barrier to the DCs could also be considered, but we did not receive any evidence that such a mechanism exists. This is, however, something to be investigated further. In this article we have described a newly established triple-cell co-culture model of the cellular air–blood barrier. The three-dimensional model has been used to investigate particle uptake and translocation by the three cell types and their possible interplay during this process. The cell-to-cell communication of the three human cell types in the two-chamber system makes it a more realistic human tissue model than the standard monolayers. More investigations are needed, including quantification of the particle transfer from the air to the DCs to understand the pathway of antigens and the interplay of the different cell types involved.
The authors thank PD Dr. Marianne Geiser for stimulating discussions. They are also grateful to Barbara Tschirren and Beat Haenni for their excellent technical assistance, and Sandra Frank for the computer drawings. Thanks are also due to Carl Zeiss AG, Switzerland, who sponsored a partial amount for printing of the color pictures.
This work was supported by the Swiss National Science Foundation (Nr. 32-65352.01); the Swiss Agency for the Environment, Forests, and Landscape; and the Silva Casa Foundation. Conflict of Interest Statement: B.M.R.-R. has no declared conflicts of interest; S.G.K. has no declared conflicts of interest; and P.G. has no declared conflicts of interest. Received in original form June 8, 2004 Received in final form November 30, 2004
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
2 µM. Samples of 100 µl were collected every 10 min (total 60 min) from the lower chamber (i.e., the receiver compartment). Sample volumes were not replaced. Samples were mixed with 3 ml scintillation cocktail and analyzed in a scintillation counter (TriCarb 2000CA Liquid Scintillation Analyzer; Packard BioScience BV, Groningen, Netherlands). The appearance of the drug in the receiver compartment was plotted as a function of time. Apparent permeability coefficient (Papp) values were determined as follows: Papp = (dQ/dt)(1/(c0A)), where dQ/dt (mol/s) is the increase in the amount of drug in the receiver chamber per time interval, A (cm2) the growth area of the cell culture insert, and c0 (mol/ml) the initial drug concentration in the donor chamber. 
