Introduction

Abstract Introduction Materials and Methods Results Discussion References

Lymphocyte and monocyte accumulation in the bronchial wall and lumen is symptomatic of an inflammatory event and is typically observed in many chronic pulmonary disease states. It is likely that multiple, independent, and interdependent events mediate the recruitment of leukocytes from peripheral blood to airways. Initially, leukocytes must first adhere to vascular endothelium and extravasate into the extracellular matrix of the submucosa, an event that in itself is a multistep process (1, 2). Subsequently, leukocytes must migrate across the epithelial lining of the lung into the airways; monocyte-derived alveolar macrophages and lymphocytes in the airway lumen play a central role in the development and perpetuation of the pulmonary inflammatory response. Transepithelial movement of leukocytes into the conducting airways is of particular functional significance, because the epithelial cells of this location are likely to be first to encounter injurious inhaled agents and may modulate the cellular environment to promote leukocyte accumulation.

Given the shear forces necessary to maintain fluid flow within the systemic circulation, recruitment of leukocytes from the vasculature requires rapid, transient (selectin/ carbohydrate or 4 integrin-dependent), and sustained (integrin-dependent) adhesive interactions with endothelium. In contrast, the physical forces that restrict leukocyte interactions with endothelia are not a factor in the migration of leukocytes across mucosal surfaces. Although this would suggest that the adhesive mechanisms that regulate transepithelial emigration of leukocytes are limited to integrin-mediated events, some evidence exists that certain carbohydrates are important for neutrophil movement across intestinal epithelium (3). Importantly, integrin-mediated leukocyte interactions with endothelia appear to be distinct from those with certain epithelia; transendothelial emigration of neutrophils is CD11a/CD18 (LFA-1)-dependent (4), whereas transmigration of neutrophils across epithelium is exclusively CD11b/CD18 (Mac-1)-dependent (5, 6).

In addition to adhesion molecules, chemokines and other chemoattractants also play an important role in leukocyte-endothelial cell interactions. Although the exact pathophysiological functions of chemokines are not yet clearly defined, some studies have suggested a role for these small-molecular-weight chemoattractant cytokines in the directed recruitment and activation of leukocytes to sites of inflammation (7, 8). It has been proposed that chemokine binding to Gi protein-linked transmembrane receptors functions as a physiological "trigger" for the rapid activation of leukocyte surface integrins (1, 2). This notion is supported by in vivo data demonstrating that pertussis toxin, an inhibitor of G protein-mediated signal transduction, can block activation-dependent binding of lymphocytes to endothelium (9). Biologic sources for chemokines include T lymphocytes, monocytes, platelets, as well as cell types of nonhematopoetic origin. Airway epithelial cells can express chemokines for both monocytes and lymphocytes (10). The expression of chemokines by lung epithelium is inducible in response to a variety of stimuli and may play a role in emigration of leukocyte cell populations in vivo.

Animal studies suggest that the repertoire of circulating lymphocyte populations includes a fraction that selectively recirculates to lung lymphoid tissues (13). One approach to characterization of lymphocyte phenotype in the human lung is immunohistochemical analysis of leukocytes obtained via bronchoalveolar lavage (BAL). This method has demonstrated that the majority of T lymphocytes in BAL fluid (BALF) are CD45RO⁺, implying a memory/effector phenotype for this leukocyte population (16). Pulmonary infiltrating lymphocytes isolated directly from lung parenchyma are also predominantly CD45RO⁺ and CD45RA (17). Interestingly, it has been demonstrated recently that the expression of certain chemokine receptors is restricted to a subset of CD45RO⁺ cells within the T lymphocyte lineage (18). Additionally, T lymphocytes that chemoattract to either regulated on activation, normal T cell expressed and secreted (RANTES) or monocyte chemotactic protein-1 (MCP-1) in vitro are CD45RO⁺ (19, 20). This would then suggest that the capacity of certain T lymphocyte subpopulations to emigrate reflects the expression of appropriate chemokine receptors in addition to adhesion molecules.

To mimic the pathway by which human lymphocytes and monocytes migrate into the airways, we present here an in vitro transmigration model that utilizes the human transformed airway epithelial cell line BEAS-2B S.6. This in vitro system allows a focused investigation of the adhesive and chemoattractant components of airway epithelium, which may be important for transepithelial migration of lymphocytes and monocytes in vivo. Furthermore, it permits direct immunophenotyping of T lymphocyte populations that exhibit the capacity to migrate across airway epithelium.

	Materials and Methods

Abstract Introduction Materials and Methods Results Discussion References

Materials

Recombinant human interferon- (IFN-) was a generous gift from Dr. M. Shepard (Genentech, San Francisco, CA). Recombinant human tumor necrosis factor- (TNF-) was obtained from Biosource (Camarillo, CA). Pertussis toxin was obtained from List Biologicals Laboratories (Campbell, CA). A mutant pertussis holotoxin (PT9K 129G) was a kind gift from Dr. Rino Rappuoli (Chiron Vaccines, Siena, Italy).

Cell Culture

The BEAS-2B S.6 cell line was derived from normal human bronchial epithelial cells immortalized using an SV40-adenovirus 12 hybrid virus (21). The BEAS-2B S.6 cell line was a kind gift of Dr. Curtis Harris (Bethesda, MD). BEAS-2B S.6 cells were cultured in a serum-free bronchial epithelium growth medium (BEGM) (Clonetics, San Diego, CA). For migration assays, 3-4 × 10³ BEAS-2B S.6 cells were plated onto polycarbonate membranes of Transwell cell culture inserts (3 µm pore size, 6.5-mm diameter; Costar, Cambridge, MA). Depending on the orientation required for an experiment, cells were plated on either the top of membranes (apical surface up) or the bottom of membranes by inversion of Transwell inserts (apical surface down). Cells on inverted Transwell inserts were allowed to settle for 12 h, then flipped to 24-well plates holding the growth medium. BEAS-2B S.6 monolayers were cultured until confluent (approximately 7 d) before initiation of migration experiments. Complete confluence of monolayers was confirmed using a modified Giemsa stain (Sigma, St. Louis, MO) and phase-contrast microscopy.

The SCL-1 squamous cell line was derived from a cutaneous squamous cell carcinoma (22). SCL-1 cells were cultured in a serum-free keratinocyte growth medium (KGM; Clonetics). For transmigration assays, SCL-1 cells were cultured on Transwell inserts as described for BEAS-2B S.6 cells.

PBMC Transepithelial Migration Assay

Mononuclear cells were purified from peripheral blood from healthy human volunteers, using density gradient centrifugation as previously described (23). Just before initiation of migration assays, peripheral blood mononuclear cells (PBMC) were resuspended in preequilibrated/prewarmed (37°C) BEGM at a concentration of 1.5-2 × 10⁷ cells/ml. Approximately 1 h prior to initiation of migration assays, both upper and lower chambers of Transwell inserts containing either BEAS-2B S.6 or SCL-1 monolayers were changed to fresh, preequilibrated/prewarmed BEGM. In some experiments, PBMC transmigration across SCL-1 cells was performed with KGM; resulting migration was similar to that obtained by BEGM (data not shown). PBMC were added directly to upper chambers of Transwell inserts at a concentration of 1.5-2 × 10⁶ cells/0.1 ml/ insert. Transwell inserts were incubated at 37°C in 5% CO₂ for 3 to 4 h. Transmigration was quantified by collection of PBMC from lower chambers (cells that had fallen to the bottom of the 24-well plates) and counting with a standard hemocytometer.

For experiments utilizing pertussis toxin, PBMC were isolated as described and suspended in RPMI 1640 supplemented with 10% fetal calf serum at a concentration of 5 × 10⁶ cells/ml. PBMC were incubated with either pertussis toxin or mutant pertussis holotoxin (100 ng/ml) for 2 h at 37°C in 10% CO₂. Following treatment with native or mutant pertussis toxin, PBMC were washed twice with Hanks' balanced salt solution and resuspended in preequilibrated/ prewarmed BEGM as described previously.

Monoclonal Antibodies

Monoclonal antibody (mAb) anti-CD2 fluorescein isothiocyanate (FITC) (pan T/NK cell), CD19 FITC (B cell), CD4 FITC (T helper), CD4 biotin, CD8 biotin (T cytotoxic), CD45RO FITC, CD29 unconjugated (1 integrin), and negative control mouse IgG were obtained from Immunotech (Westbrook, ME). mAb anti-CD14 phycoerythrin (PE) (monocyte), CD26 PE, and secondary reagent streptavidin Cychrome were obtained from PharMingen (San Diego, CA). Dreg 56 conjugated (mouse antihuman L-selectin) was prepared in our own laboratory (24). The secondary reagent goat antimouse IgG FITC was obtained from Biosource (Camarillo, CA).

Cell Staining/Flow Cytometric Analysis

Following termination of transmigration assays, PBMC from upper and lower wells of Transwell inserts were collected and separately stained for analysis by flow cytometry. As a control, fractions of PBMC isolates were incubated in BEGM for the entire length of the transmigration assay; this fraction represents the starting population. A five-parameter analysis was performed on a FACScan flow cytometer (Becton Dickinson Immunocytometry Systems, San Jose, CA) with FITC, PE, and Cychrome fluorochromes. For each test sample, 2-3 × 10⁵ PBMC were stained; comparable numbers of cells were used for each test sample in an individual experiment. Immunofluorescence staining for multiparameter analysis was performed as previously described (25). All data was acquired in list mode, acquiring 30,000 to 40,000 events per sample. Data were analyzed by CELLQuest software (Becton Dickinson Immunocytometry Systems). To eliminate potential contamination of BEAS-2B S.6 cells in the analysis of migrated PBMC fractions, a large gate was placed around the entire PBMC population as determined by forward and slide light-scatter properties. Analysis of CD4⁺ or CD8⁺ lymphocytes was performed by using a light-scatter gate and a gate based on CD4⁺ or CD8⁺ fluorescence.

Statistics

Unless indicated, all data are presented as means ± SEM. Statistical significance was determined by Student's t test.

	Results

Abstract Introduction Materials and Methods Results Discussion References

PBMC Transmigration across Airway Epithelial Cell Monolayers

The effect of cell monolayer polarity on the capacity of PBMC to transmigrate across airway epithelium was initially examined. The BEAS-2B S.6 cell line was chosen for our in vitro model system on the basis of previously reported data, indicating that it behaves most like primary tracheobronchial epithelial cell cultures in response to proinflammatory mediators, secretion of cytokines, and expression of adhesion molecules (26). Also, it has been shown that BEAS-2B S.6 cell cultures exhibit significant transepithelial resistance and tight junction formation (32). Within a 3-h period of incubation, a limited number of PBMC were capable of migration across a BEAS-2B S.6 bronchial epithelial cell monolayer. The 3-h period of culture was initially used because it allowed for a number of leukocytes to migrate sufficient for accurate quantitation (per Transwell) and analysis by flow cytometry. As shown in a representative experiment using leukocytes obtained from one blood donor (Figure 1), basal-to-apical surface migration of PBMC across BEAS-2B S.6 cell monolayers was significantly more efficient than migration in the apical-to-basal direction (mean ± SEM percentage of migration: 5.40 ± 0.35 versus 0.51 ± 0.24, respectively; P < 0.001). Although the capacity for PBMC basal-to-apical migration across BEAS-2B S.6 monolayers was variable among blood donors used for individual experiments (n = 7; mean = 3.7%, range = 1.2-6.4% of starting PBMC), the overall efficiency of migration across BEAS-2B S.6 cells was consistently enhanced at least twofold in the physiologically appropriate direction (compared with the apical-to-basal direction).

View larger version (37K): [in this window] [in a new window]   Figure 1.   Effect of epithelial cell monolayer polarity (apical versus basal) on PBMC transmigration. PBMC were allowed to migrate across BEAS-2B S.6 or SCL-1 cell monolayers in either the apical-to-basolateral direction (apical) or basolateral-to-apical direction (basal). Transmigration was measured following a 3-h incubation period. The calculated percentage of transmigrated PBMC is based on the starting number of PBMC added to individual Transwell inserts. Each column represents the mean ± SEM percentage of transmigrated PBMC across three Transwell inserts. Data from one representative experiment (using one blood donor) of three carried out are shown. *P < 0.001 compared with apical-to-basolateral transmigration across BEAS-2B cell monolayers; **P < 0.001 compared with basolateral-to-apical transmigration across BEAS-2B cell monolayers.

Comparison of PBMC Transmigration across Airway Epithelium versus Cutaneous Epithelium

To determine if the ability of PBMC to transmigrate across epithelium is dependent on cell monolayer phenotype and not an artifact of gravity-induced movement across Transwell membrane inserts, PBMC migratory capacity across squamous epithelium of epidermal origin was compared with airway epithelium. We utilized the SCL-1 cutaneous squamous carcinoma cell line in our experiments because it has been reported to exhibit similar patterns of adhesiveness to T lymphocytes to those of primary keratinocytes (33). In contract with the BEAS-2B S.6 cell line, the SCL-1 cutaneous squamous carcinoma cell line supported only minimal transmigration, with no preference for the basal-to-apical direction (Figure 1). The capacity of PBMC to migrate across basal or apical surfaces of SCL-1 cell monolayers (mean ± SEM percentage of migration: 0.75 ± 0.22 and 0.50 ± 0.33, respectively) was significantly limited when directly compared with basal-to-apical surface migration across BEAS-2B S.6 cell monolayers (P < 0.001).

Effect of IFN- and TNF- on PBMC Transmigration across BEAS-2B S.6 Cell Monolayers

To determine if an inflammatory stimulus to airway epithelium can enhance PBMC transepithelial migration, the effect of proinflammatory cytokine treatment of BEAS-2B S.6 cells on subsequent PBMC migration was examined. BEAS-2B S.6 cell monolayers were treated with a combination of IFN- and TNF-, as both of these cytokines have been shown to upregulate chemokine and adhesion molecule expression by airway epithelium (27, 34). As shown in Figure 2, 24-h preincubation of BEAS-2B S.6 cell monolayers with IFN- (1,000 U/ml) and TNF- (10 ng/ml) resulted in significantly enhanced PBMC transmigration across BEAS-2B S.6 cell monolayers in the basal-to-apical direction compared with the control, untreated BEAS-2B S.6 cell monolayers (mean ± SEM percentage of migration: 6.96 ± 2.35 versus 3.44 ± 1.72, respectively; P < 0.05). Although IFN- and TNF- treatment of SCL-1 cell monolayers has previously been shown to enhance T lymphocyte adhesion to apical surfaces (33), these cytokines had minimal effect on PBMC transmigration across SCL-1 cells in the basal-to-apical direction (data not shown).

View larger version (51K): [in this window] [in a new window]   Figure 2.   Effect of IFN- and TNF- on PBMC transmigration across BEAS-2B S.6 cell monolayers. BEAS-2B S.6 cell monolayers cultured in Transwell inserts were incubated with IFN- (1,000 U/ml) and TNF- (10 ng/ml) for 24 h prior to addition of PBMC. One hour before the migration assay was started, IFN- and TNF- were removed from the cultures and replaced with fresh medium (no cytokine). PBMC were allowed to migrate across cell monolayers in the basolateral-to-apical direction for a period of 3 h. The calculated percentage of transmigrated PBMC is based on the starting number of PBMC added to individual Transwell inserts. Each column represents the mean ± SEM percentage of transmigrated PBMC from five experiments using four different blood donors. Each experiment utilized three to six Transwell inserts per variable. *P < 0.05 compared with control (no cytokine).

Phenotypic Analysis of PBMC Exhibiting Transepithelial Migratory Capacity

The distribution of PBMC populations following transmigration across BEAS-2B S.6 cell monolayers is shown in Table 1. In three separate experiments using three different blood donors, a large percentage (range = 66% to 90%) of the migrated PBMC population consisted of CD14⁺ monocytes, which were correspondingly depleted in the nonmigrated PBMC fraction. The percentage of CD19⁺ lymphocytes was slightly decreased in the migrated PBMC fraction (range = 2% to 5%, compared with 7% to 12% starting); this modest reduction of migrated CD19⁺ cells was consistently observed within all three experiments. For two of three experiments, the range of CD2⁺ cells migrated in PBMC was 21% to 24%; one blood donor exhibited unusually low CD2⁺ cell migration (5% of migrated PBMC). Because CD2 is on both T lymphocytes and natural killer (NK) cells, additional phenotypic analysis from two experiments (with two different blood donors not used in Table 1) showed that the range of CD56⁺ NK cells was 10% to 12% of the total migrated PBMC population (14%-15% of the total starting PBMC population were CD56⁺). Surface expression of CD3 was reduced on T cells following transmigration; therefore, it was not used as a reliable, quantitative measure of T lymphocyte populations in our study (data not shown). Downregulation of CD3 expression has been documented on freshly isolated T cells obtained via BAL (35, 36); modulation of cell surface CD3 levels may be reflective of recent activation events. Given the limited number of PBMC that migrated across SCL-1 monolayers, we were unable to perform phenotypic analysis of these populations by fluorescent-activated cell sorting (FACS).

Memory T Lymphocytes Preferentially Migrate across BEAS-2B S.6 Cell Monolayers

Immunophenotyping of T lymphocytes in BEAS-2B S.6 cell transepithelial migration assays showed that, on average, 8.8% of migrated PBMC populations were CD4⁺ cells (Table 2); and 2.2% were CD8⁺. In comparison, the mean percentages of CD4⁺ and CD8⁺ cells in the starting PBMC populations were 36% and 18%, respectively. The mean CD4⁺/CD8⁺ ratio for the staring PBMC population was 2.1 (range = 1.1 to 2.7); in contrast, the mean CD4⁺/ CD8⁺ ratio for transmigrated PBMC was 16.4 (range = 0.8 to 51.0). Thus, CD4⁺ cells exhibited a greater migratory capacity across airway epithelium in comparison with CD8⁺ cells. In the CD4⁺ cell subset, CD45RO⁺ cells were significantly enriched in transmigrated fractions compared with numbers found in the starting population (mean percentage positive: 42.1 starting versus 77.1 migrated; P < 0.01). CD45RO⁺ cells within the CD8⁺ cell subset were also significantly increased in transmigrated fractions (mean percentage positive: 31.4 starting versus 83.4 migrated; P < 0.001). The mean ratio of CD45RO⁺/CD45RO cells in migrated CD4⁺ or CD8⁺ subsets was correspondingly much higher than that in the starting PBMC population (mean ratio: 0.9 (CD4⁺) and 0.5 (CD8⁺) starting versus 5.3 (CD4⁺) and 8.3 (CD8⁺) migrated). Although IFN- and TNF- treatment appeared to enhance PBMC transmigration across BEAS-2B S.6 cell monolayers (Figure 2), these cytokines did not significantly alter the overall distribution of CD4⁺, CD8⁺, and CD45RO⁺ phenotypes in transmigrated PBMC populations from two experiments (compared with untreated BEAS-2B S.6 cell monolayers; data not shown).

Further FACS were performed to characterize the predominant CD4⁺ migrating T cell population (Figure 3). The memory/effector T cell phenotype has previously been characterized as CD45RO⁺, CD29⁺, and heterogenous for expression of CD62L (L-selectin) (24, 37). As demonstrated in Table 2, CD45RO⁺ cells are enriched within the migrated CD4⁺ fraction. Interestingly, the expression of L-selectin in the migrated CD4⁺ cell fraction was uniformly at an intermediate level, between the highly positive and negative peaks exhibited in the starting input and nonmigrated CD4⁺ populations. In addition, most of the migrated CD4⁺ cells exhibited increased expression of CD29. The majority of the CD4⁺ migrated fraction also showed increased expression for CD26 (dipeptidyl peptidase IV), which is upregulated on the surface of activated, but not resting T lymphocytes (38, 39).

View larger version (30K): [in this window] [in a new window]   Figure 3.   Phenotype of CD4+ cells following PBMC transmigration across BEAS-2B S.6 cell monolayers. Starting input PBMC (a), nonmigrated PBMC (b), and migrated PBMC (c) were separately collected following a 3-h transmigration across BEAS-2B S.6 cell monolayers. Recovered PBMC were incubated with antibodies as described in MATERIALS AND METHODS; stained PBMC were subsequently analyzed by flow cytometry. Histograms show surface expression of CD45RO, CD29, CD26, and CD62L on cells gated for light-scanner and CD4+ fluorescence. Data from one representative experiment (using one blood donor) of three carried out are shown.

Pertussis Toxin Inhibits PBMC Transmigration across BEAS-2B S.6 Cell Monolayers

Pertussis toxin has previously been shown to inhibit activation-dependent adhesion of lymphocytes to high endothelial venules and block transendothelial diapedesis of T cells in vitro (40); it is likely that this effect is mediated via adenosine diphosphate ribosylation of G proteins involved in signaling events following binding of chemokines to appropriate receptors on the surface of lymphocytes. To determine if transmigration of PBMC across airway epithelium is dependent on a similar signal transduction event, we treated PBMC with pertussis toxin before allowing transepithelial migration across BEAS-2B S.6 cell monolayers. As a control for nonspecific effects, a mutant pertussis holotoxin that specifically lacks adenosine diphosphate-ribosyltransferase activity was used in these experiments. As shown in Table 3, pertussis toxin treatment of PBMC significantly inhibited transepithelial migration by 83.5 ± 4.9% (mean ± SEM from three separate experiments) compared with mutant pertussis toxin-treated PBMC. In comparison with untreated PBMC, mutant pertussis holotoxin reduced transmigration by 23.1 ± 4.4% (mean ± SEM from three separate experiments), suggesting a slight ribosylation-independent effect and emphasizing the importance of appropriate controls in experiments utilizing pertussis toxin.

	Discussion

Abstract Introduction Materials and Methods Results Discussion References

In this study we have characterized the capacity of human PBMC to migrate across a barrier consisting of human lung epithelium. We found that migration of PBMC across a conducting airway epithelial cell monolayer occurs preferentially in a physiologically appropriate direction (basal-to-apical surface); moreover, this response is enhanced by epithelial cell exposure to the proinflammatory cytokines IFN- and TNF-. For the experiments described here, the human transformed bronchial epithelial cell line BEAS-2B S.6 served as an in vitro model of airway epithelium. Although we have not directly compared PBMC transmigration across BEAS-2B S.6 cells with primary tracheobronchial epithelial cells, a recent study by Liu and colleagues (6) has demonstrated basal-to-apical surface preferential migration of human neutrophils across both primary cells and the human airway mucoepidermal cell line H292. Using the BEAS-2B S.6 cell line, we have also found basal-to-apical epithelial surface preferential migration of human neutrophils (L. Miller and D. Hyde, in preparation). For investigation of the cellular and molecular mechanisms of leukocyte transepithelial migration, the use of a cell line is advantageous in that it limits the experimental variability obtained by primary epithelial cell preparations. Our observations here demonstrate the viability of the BEAS-2B S.6 cell line as a model of leukocyte transepithelial migration, and suggest that cell polarity is universally important for transepithelial migration of human leukocytes.

A direct comparison of leukocyte transmigration across airway epithelium versus cutaneous epithelium demonstrated that lung epithelium exhibits an enhanced capacity for transmigration of PBMC. It has been reported that IFN- and TNF- can dramatically upregulate T lymphocyte adhesiveness to either SCL-1 or primary keratinocyte cell cultures within a 30-min incubation period (33). Our experiments showed that movement of PBMC across SCL-1 squamous cell monolayers within a 3-h incubation period is quite limited, regardless of polarity and/or stimulation with proinflammatory cytokines. It is not known if the rate of leukocyte movement into the epidermis in vivo is inherently slower than movement into the airways; the results of this experiment suggest that the efficiency of leukocyte transepithelial movement is dependent upon the origin and/or nature of the epithelium in the body.

Stimulation of airway epithelium with the proinflammatory cytokines IFN- and TNF- can upregulate the expression of intercellular adhesion molecule-1 and leukocyte chemoattractants. We have found that treatment of BEAS-2B S.6 cells with a combination of TNF- and TNF- can upregulate PBMC transmigration in the basal-to-apical direction. We do not know if enhancement of PBMC migration occurs via upregulation of chemokine and/or adhesion molecule expression by BEAS-2B S.6 cells. It is unlikely that this mechanism occurs via alteration of tight junctions and subsequent nonspecific cellular "leakiness" of the epithelial cell monolayer; we found that the distribution of CD4⁺/CD8⁺/CD45RO⁺ cells in migrated PBMC populations remained consistent regardless of prior treatment of cultures with cytokine. Additionally, this distribution of CD4⁺/CD8⁺/CD45RO⁺ cells was not observed following PBMC transmigration across naked Transwell membranes in response to fresh or BEAS-2B S.6 cell-conditioned medium (L. Miller and E. Butcher, unpublished observation). The lack of change in the migrated T lymphocyte population (in response to IFN-/ TNF--treated epithelium) was not surprising, as it has been reported that the characteristic memory/effector profile of BALF T cells from normal lung is very similar to that found in subjects with pulmonary inflammatory disease (41).

Transepithelial migrated PBMC consisted of a large population of monocytes and a small population of lymphocytes and NK cells. The mechanisms of monocyte adhesion to epithelium are not known but may be similar to the leukocyte function-associated antigen (LFA)-1-ICAM-1 pathway, which can mediate transendothelial migration in vitro (42). Bronchial epithelial cells can express chemokines, which promote chemotaxis of monocytes, such as macrophage inflammatory peptide-1, MCP-1, and RANTES (10). It is possible that the observed enrichment of monocytes in the migrated PBMC population is related to levels of specific chemoattractants expressed by BEAS-2B S.6 cells; however, it is more likely that this phenomenon is primarily the result of the inherent motility of monocytes. Lymphocytes are less motile than monocytes, and it has been reported that the rate of lymphocyte migration to MCP-1 is much slower than that of monocytes (20, 43).

The CD4⁺ and CD8⁺ cells that migrated across BEAS-2B S.6 cell monolayers were enriched for the surface expression of CD45RO. In addition, the migrated CD4⁺ cell population was predominantly CD29⁺ and CD26⁺. CD62L surface expression on migrated CD4⁺ cells was of an intermediate level; a similar profile of CD62L has been reported for T lymphocytes that have migrated across human umbilical vein endothelial cell (HUVEC) monolayers (44, 45). This pattern of CD45RO⁺/CD29⁺/CD26⁺ expression has also been demonstrated on certain subsets of T cells that exhibit the capacity for transendothelial migration (44). Interestingly, reports of the transendothelial migration capacity of CD4⁺ and CD8⁺ subsets have been dissimilar. One group has shown preferential migration of CD8⁺/CD45RO/CD45RA⁺ and CD4⁺/CD45RO⁺ T cells, with the CD8⁺ population exhibiting greater transendothelial migratory capacity (44). In contrast, another study found that T cells that migrated across HUVEC monolayers were predominantly CD45RO⁺ and were enriched for CD4⁺ cells (47). In our studies of transepithelial movement of PBMC, we have found that CD4⁺ cells exhibited a much greater capacity to migrate across BEAS-2B S.6 cell monolayers than CD8⁺ cells. However, in comparison with results obtained by Berman and colleagues (47), the relative predominance of the CD4⁺ subset (as measured by the CD4⁺/CD8⁺ ratio) was more pronounced in PBMC following transepithelial migration. Nevertheless, the persistent observation of a CD45RO⁺ phenotype in transendothelial or transepithelial migrated CD4⁺ cell populations suggests that migrational competency (in this subset of T lymphocytes) is enhanced among previously activated cells.

Pertussis toxin treatment of PBMC almost completely negated transmigration across BEAS-2B S.6 monolayers. In contrast, neutrophil transmigration across BEAS-2B S.6 cells is only partially sensitive to pertussis toxin treatment (L. Miller and D. Hyde, in preparation). Our experiments implicate a significant role for chemokine/chemoattractant-mediated migration of PBMC across airway epithelium. As yet, we have not determined which chemokine(s) are responsible for migration of monocytes and lymphocytes across BEAS-2B S.6 cell monolayers. The MCP-1 chemokine is an attractive possibility because it not only is expressed by airway epithelium (10), but also can function as a chemoattractant for both monocytes and memory/effector T lymphocytes (20, 48). However, T lymphocyte migration to MCP-1 does not favor enrichment of either CD4⁺ or CD8⁺ cells (20), whereas our results indicate that transepithelial migration across BEAS-2B S.6 cells is selective for CD4⁺ cells. It is possible that airway epithelial cells express other chemokines or undetected attractants that act in concert.

In conclusion, our results demonstrate the viability of our in vitro transepithelial model for monocyte and lymphocyte migration into the airways. The capacity for T lymphocyte transmigration across airway epithelium is limited to subpopulations of memory/effector cells and is almost completely dependent on pertussis toxin-sensitive G protein signaling events. It is anticipated that subsequent studies will reveal the precise chemotactic and adhesive parameters that define monocyte and T lymphocyte transepithelial migration.