Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Pseudomonas aeruginosa is a typical opportunist pathogen that does not usually infect healthy tissues. Adherence to different receptors that are not available for bacterial binding in intact tissues may initiate and ultimately induce persistent infection in injured tissues (1).

Airways from cystic fibrosis (CF) patients are continually exposed to inflammatory insult and are particularly susceptible to chronic P. aeruginosa colonization and infection. The factors that contribute to the pathogenesis of P. aeruginosa infection in CF have not yet been completely defined but are likely attributable to the dysfunction of the CF transmembrane conductance regulator (CFTR) protein.

In normal human airway surface epithelium, CFTR is restricted to the apical plasma membrane of ciliated cells, and this localization constitutes a marker of cell differentiation (2, 3). CFTR with deletion of the amino acid Phe 508 (F508), resulting from the predominant mutation in CF gene, has an abnormal intracellular (IC) maturation and trafficking (4) and is unable to accumulate at normal levels at the apical plasma membranes. Mutated CFTR has most frequently been localized diffusely in the cytoplasm of epithelial respiratory cells (2, 5), as a consequence of its retention in the endoplasmic reticulum.

Pier and colleagues (6) have shown that epithelial respiratory cells actively uptake P. aeruginosa and that bacterial internalization correlates with membrane expression of wild-type CFTR. They suggested that P. aeruginosa can initiate infection in CF patients because airway epithelial cells expressing F508 mutant CFTR are defective in internalization of the bacterium, a process that may be an important host defense mechanism. In a further investigation, Pier and coworkers (7) have shown that the first extracellular domain of CFTR bound specifically to P. aeruginosa and that a synthetic peptide of this region inhibited bacterial internalization by cells from neonatal mouse lungs wherein epithelial cell invasion has been observed (8). DiMango and associates (9) also observed a higher P. aeruginosa internalization by normal epithelial respiratory cells than by cells from a CF transformed cell line. In contrast, Ko and colleagues (10) have recently shown significant P. aeruginosa entry into respiratory cells from a CF F508 cell line, suggesting that bacterial internalization does not depend on CFTR expression at host-cell membranes.

In our studies on P. aeruginosa interaction with human CF and non-CF epithelial respiratory cells in primary culture, we rarely observed IC P. aeruginosa (11). In contrast, we have recently shown that P. aeruginosa entry into epithelial cells is enhanced in repairing wounded monolayers and in noninjured monolayers treated with a calcium chelator to disrupt intercellular junctions (12).

In normal epithelia, junctional complexes between adjacent cells provide high resistance intercellular seals, forming a barrier between biologic compartments. Circumferential apical tight junctions also separate apical and basolateral domains of cell membranes, assisting in maintenance of epithelial cell polarity. Membranes of these two cell domains differ structurally, biochemically, and physiologically (13). By isolating basolaterally restricted proteins, tight junctions also restrict the repertoire of potential interactions that may occur between luminal bacteria and the epithelial cell surface.

During the repair of epithelial wounds, tight junctions are disrupted not only in the wound area but also between uninjured cells at some distance from the area of trauma (14, 15), and apical cell membranes of respiratory cells exhibit molecules usually restricted to basolateral membranes (16). Accordingly, exposure of basolateral receptors to bacterial ligands in poorly polarized epithelial cells or in an epithelium that has lost its barrier integrity may have accounted for the enhancement of P. aeruginosa internalization by repairing monolayers that we have observed (12). Fleiszig and coworkers (17) have recently reported, on different types of epithelial cells, that cell polarity affects susceptibility to P. aeruginosa invasion and cytotoxicity. However, in their study, no specific marker could assess the degree of epithelial respiratory cell differentiation and polarity.

To unravel the relationship between P. aeruginosa internalization by epithelial respiratory cells, cell polarity, and CFTR expression, we analyzed the interaction of two piliated strains of P. aeruginosa with human epithelial respiratory cells (HERC) in primary culture and with 16 human bronchial epithelial (HBE) 14o- cells (18) cultured on substrates allowing different degrees of cell polarity and differentiation. We also analyzed the role of cell barrier integrity on P. aeruginosa internalization.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Airway Epithelial Cell Culture

Primary cultures of HERC from nasal polyps were obtained as described by Chevillard and associates (19). Cells were cultured in serum-free defined RPMI 1640 culture medium (GIBCO BRL, Gaithersburg, MD) supplemented with 1 µg/ml insulin, 1 µg/ml transferrin, 10 ng/ml epidermal growth factor, 0.5 µg/ml hydrocortisone, 10 ng/ml retinoic acid, and antibiotics. Transformed cell line 16 HBE 14o- (18), kindly provided by Dr. D. C. Gruenert (University of California, CA), was cultured in 199-N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid (Hepes) medium containing 10% fetal calf serum and antibiotics. Both HERC and 16 HBE 14o- cells were seeded onto circular glass or Thermanox (Nunc, Inc., Naperville, IL) coverslips in 24-well microplates and cultured until confluence. For a few assays, cells were cultured in immersed conditions on porous polycarbonate membranes (Transwell 6.5-mm diameter; Costar, Cambridge, MA). In all assays, coverslips and the polycarbonate membranes were coated with thick type I collagen gels, prepared according to the technique of Chambard and colleagues (20), or coated with thin collagen I films prepared as described by Herard and coworkers (15). Collagen type I used to prepare thick gels and thin films was isolated from rat tendons and dissolved in 0.016 M acetic acid solution at 2.5 mg/ml. Thick collagen gels were obtained by reticulation of 250 µl of the type I collagen solution with 1 N sodium hydroxide. Thin collagen films were obtained by incubating 200 µl of a solution containing 20 µl of carbodiimide (Sigma Chemical Co., St. Louis, MO) at 100 µg/ml, 1 ml of the type I collagen solution, and 7 ml of distilled water for 3 h in microplate wells. Thereafter the solution was removed and the culture dishes were air-dried.

Immunolocalization of CFTR

Localization of CFTR protein at apical membranes of HERC and 16 HBE 14o- cells was used as a marker of cell differentiation and was carried out as described (5). Briefly, cells cultured on polycarbonate membranes coated with thick collagen gels or thin collagen films were embedded in Optimum Cutting Temperature compound (Tissue Tek; Miles, Elkhart, IN), immersed in liquid nitrogen for 5 min, and kept at 80°C. Thick sections (5 µm), obtained with a Reichert 2800 Frigocut (Cambridge Instrument, Nussloch, Germany), were deposited onto gelatin-coated glass slides and air-dried. Sections were rehydrated in 0.1 M phosphate-buffered saline (PBS), pH 7.4, and in PBS containing 1% bovine serum albumin (PBS-BSA) for 10 min and exposed to a 1:200 dilution of the 24-1 mouse monoclonal antibody raised against the amino acid sequence 1377- 1480 of the C-terminal domain of the CFTR protein (gift from Genzyme Corp., Cambridge, MA) for 1 h at room temperature. Negative controls were performed using nonimmune mouse immunoglobulin (Ig)G fractions. Cells were then rinsed and exposed to biotinylated antimouse antibody (Amersham, Arlington Heights, IL) at 1:50 for 1 h and to streptavidin-fluorescein isothiocyanate (FITC) (Amersham) at 1:50 for 30 min. Sections were counterstained with Harris hematoxylin solution, mounted in Citifluor antifading solution (Agar Scientific, Stansted, UK) and observed under an Axiophot microscope equipped with epifluorescence illumination (Axiophot Zeiss).

Integrity of Junctional Complexes

Lanthanum nitrate has been used as a tracer in transmission electron microscopic (TEM) studies of the permeability of biologic barriers. Because intact tight junctions are impermeable, and therefore exclude lanthanum, the identification of this electron-dense tracer in intercellular spaces indicates the permeability of the epithelial barrier (21). The permeability of respiratory cell cultures was investigated as described by Revel and Karnovsky (22). Briefly, confluent cells cultured on Thermanox coverslips coated with thick collagen gels or with thin collagen films were fixed for 1 h with glutaraldehyde at 2.5% in 0.1 M phosphate buffer (pH 7.4), washed with phosphate buffer and S-collidine buffer (pH 7.4), and postfixed for 2 h at room temperature with a 1:1 solution of lanthanum nitrate (Sigma) at 4% in S-collidine buffer (pH 7.8) and OsO₄ at 2%. Cells were quickly dehydrated through graded ethanol series and embedded in Epon. Ultrathin sections were observed with a Hitachi H 300 or a 906 Zeiss transmission electron microscope.

Transepithelial Resistance Measurements

HERC and 16 HBE 14o- cells were grown on porous polycarbonate membranes (Transwell) coated with thick collagen gel or thin collagen film, and their transepithelial resistance was measured using the Millicell-ERS Resistance system (Millipore, Bedford, MA). After calibrating the instrument against culture medium, one sterile electrode was placed inside the insert and the other was placed on the outside. Inserts with no monolayers served as blanks. The transepithelial electrical resistance (TER; ohms × cm²) was calculated from the following equation: (TER_sample  TER_blank) × surface area.

Immunolocalization of ZO-1 Protein

ZO-1 is a cytoplasmic protein closely associated with tight junction integrity (15). Cells cultured on glass coverslips coated with collagen gels or films were fixed with methanol at 20°C for 20 min, rinsed with PBS-BSA for 10 min, and incubated with a rat monoclonal antibody raised against ZO-1 protein (Biogenesis, Poole, UK) at 1:10 in PBS-BSA for 1 h. Cells were then exposed to biotinylated antirat antibody (Amersham) at 1:50 and to a streptavidin-fluorescein complex, mounted in Citifluor solution, and observed with a Zeiss Axiophot microscope.

Immunolocalization of 1 Integrins

The procedures of embedding of cells cultured on polycarbonate membranes coated with thick collagen gels or thin collagen films, cryofixing, and cryosectioning were similar to those used for immunolocalization of CFTR protein. Thick sections of the samples were exposed to a monoclonal antibody derived from hybridoma P5D2 against 1 integrin at 1:10 for 1 h. P5D2 anti-1 antibody was a gift from Dr. E. Warner (University of Minnesota, Minneapolis, MN). Negative controls were performed by using nonimmune mouse IgG fractions or by omitting the incubation step with the primary antibody. Cells were then rinsed, exposed to biotinylated antimouse antibody at 1:50 and to streptavidin-FITC at 1:50, mounted in Citifluor antifading solution, and observed.

Bacteria

Piliated PAK and PAO-1 P. aeruginosa strains were kindly provided by Dr. W. Paranchych (University of Alberta, Edmond, Canada) and Dr. A. Lazdunski (CNRS, Marseille, France), respectively. Poorly adhesive mutant PAK-N1, carrying a mutation in the rpoN gene required for the expression of the pilin structural gene (23), was a generous contribution from Dr. S. Lory (University of Washington, Seattle, WA) and was used, along with DH5 Escherichia coli, as a negative control. Bacteria were grown overnight at 37°C in trypticase soy broth (Difco Laboratories, Detroit, MI), harvested by centrifugation, and resuspended in 199 or in RPMI medium containing 25 mM Hepes (Sigma) to A₆₄₀ = 0.1, corresponding to 10⁸ colony-forming units (cfu)/ ml. Bacterial concentration was confirmed by quantitative culture on trypticase soy agar (TSA; Difco).

Immunocytochemical detection in cryosections of bacteria adherent to epithelial respiratory cells was performed using a rabbit antiserum raised against P. aeruginosa (Institut Pasteur, Marres-la-Conquette, France). Thereafter, cryosections were treated with an antirabbit IgG-biotin complex at 1:50 and with streptavidin-Texas Red (Amersham) at 1:50, mounted in Citifluor, and observed with a Zeiss Axiophot fluorescent microscope. In other assays, cell cultures were fixed with paraformaldehyde at 4% and adherent extracellular bacteria were distinguished from IC bacteria by in situ immunoperoxidase labeling of extracellular microorganisms and counterstaining epithelial respiratory cells and IC microorganisms with Giemsa stain, as described (12).

Bacterial Internalization Assays

Confluent cell cultures on 24-well microplates coated with collagen gels or films were exposed to 500 µl of P. aeruginosa suspensions containing 10⁸ cfu/ml. After 1 or 4 h at 37°C, supernatants were removed and cultures were rinsed and incubated for 1 h with gentamicin at 300 µg/ml in 199 or RPMI-Hepes medium to kill extracellular bacteria. Antibiotic-containing medium was removed and cells from at least three different wells were washed and lysed with sterile PBS containing 0.1% Triton X 100 (Sigma). Aliquots of cell lysates were serially diluted and plated on TSA to quantify viable IC bacteria.

Bacterial Survival

To determine whether bacteria could survive intracellularly, cells were exposed to bacterial suspensions for 1 h, rinsed, and exposed to gentamicin-containing medium for 1 h, as described previously. Thereafter, cells from at least three different wells were lysed and plated, whereas cells from other wells were incubated with the antibiotic-containing medium for additional periods (2, 4, 6, and 23 h) and then submitted to the lysing treatment. The efficiency of gentamicin in killing extracellular bacteria was systematically analyzed by plating the postinfection media on TSA.

Disruption of Tight Junctions

HERC or 160 HBE 14o- cultures on 24-well microplates coated with thick collagen gels or with thin collagen film were treated for about 10 min with ethylenediaminetetraacetic acid (EDTA) at 15 mM and monitored by inverted microscopy. Cells were then rinsed with culture medium and exposed to bacterial suspensions. Quantitative detection of IC bacteria was carried out by the gentamicin exclusion assay, as described previously.

Cell Viability after Bacterial Infection

Viability of 16 HBE 14o- cells harboring IC bacteria was assessed by the cleavage of methylthiazole tetrazolium (MTT) salt into an insoluble blue-colored formazan by the mitochondrial enzyme succinate-dehydrogenase of viable cells (24) and by the release of the cytoplasmic enzyme lactate dehydrogenase (LDH ) by cells presenting any loss of membrane permeability (25). For the MTT assay, cells were infected for 1 h and incubated with gentamicin-containing medium for different periods, as described in the bacterial survival assay. In parallel, noninfected control cells were also incubated with the gentamicin-containing medium. At different intervals, cells harboring IC bacteria and control noninfected cells were incubated with MTT solution (Sigma) at 1 mg/ml in phenol red-free culture medium for 1 h at 37°C. Cells were then rinsed with PBS and overlaid with 250 µl of isopropanol to solubilize formazan crystals formed in viable cells. Two 100-µl aliquots from each culture well were transferred to enzyme-linked immunosorbent assay microplates, and their absorbances were measured with an automatic microplate scanning spectrophotometer (Bio-Rad Laboratories, Richmond, CA) with a 550-nm test wavelength and a 690-nm reference wavelength. For the LDH assay, the supernatants of 16 HBE 14o- harboring IC bacteria for different periods and of noninfected cell cultures were recovered and their LDH activity was determined with a LDH kit (Sigma), following the manufacturer's instructions. LDH activity was also measured in lysates of noninfected cells, obtained by exposing cells to distilled water. LDH released by infected and noninfected cells was expressed as: (the activity in supernatants from experimental wells)/(the LDH activity in control cell lysates) × 100.

Statistical Analysis

All experiments were repeated three times. Results are presented as means ± standard deviation. A Student's t test was used to compare means, and P < 0.05 was considered significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

HERC and 16 HBE 14o- cells cultured on thick gels and thin films exhibited several different properties, confirming the influence of supporting substrates on the outcome of cell cultures.

Expression of CFTR Varies with the Culture Substrate

CFTR was detected at the apical membranes of many HERC cultured on thick gels (Figure 1A). In 16 HBE 14o- cells cultured under the same conditions, CFTR was detected both at apical membranes and in cell cytoplasm (Figure 1C). No CFTR could be detected in HERC cultured on thin collagen films (Figure 1B). In contrast, in 16 HBE 14o- cells cultured on thin films, CFTR was regularly observed but in cell cytoplasm only (Figure 1D). To assess the role of CFTR as a P. aeruginosa receptor, we carried out a double fluorescence assay in which CFTR was labeled in green, with FITC, whereas bacteria were labeled in red, with Texas Red. No bacteria adherent to HERC cultured on thick collagen gel could be colocalized with CFTR (Figure 1E).

View larger version (33K): [in this window] [in a new window]   Figure 1.   Immunofluorescence localization of CFTR protein in cryosections of HERC and 16 HBE 14o- cells cultured on thick collagen gels (A, C, and E) or on thin collagen films (B and D). (A) CFTR protein was detected at apical plasma membranes of many HERC (arrowheads), whereas in 16 HBE 14o- cells (C), CFTR protein was detected at both apical membranes (arrowheads) and in cell cytoplasm. (B) No labeling could be detected in HERC cultured on thin collagen films with the 24-1 antibody raised against the C-terminal domain of the CFTR protein, whereas in 16 HBE 14o- cells (D), CFTR was regularly detected but in cell cytoplasm only. (E) Double immunolabeling of HERC cultured on a thick collagen gel with 24-1 anti-CFTR antibody and anti-P. aeruginosa antibodies. CFTR appears labeled in green, whereas bacteria appear labeled in red. No adherent bacteria (arrows) could be colocalized with CFTR (arrowheads) expressed at cell apical membranes.

Polarity of Epithelial Respiratory Cells and Impermeability of Cell Cultures Depend on Supporting Substrates

Cells cultured on thick gels and on thin collagen films exhibited different degrees of polarity and permeability, according to the following criteria:

(1) Localization of 1 integrins. In both HERC and 16 HBE 14o- cells cultured on thick collagen gels, 1 integrins were restricted to lateral and basal cell membranes in contact with the supporting substrate (Figures 2A and 2C). In contrast, in both HERC and 16 HBE 14o- cells cultured on thin collagen films, 1 integrins were detected also at apical plasma membranes (Figures 2B and 2D).

View larger version (17K): [in this window] [in a new window]   Figure 2.   Immunofluorescence localization of 1 integrin in cryosections of HERC and 16 HBE 14o- cells cultured on thick collagen gels (A and C ) or on thin collagen films (B and D). In both HERC and 16 HBE 14o- cells cultured on thick collagen gels, 1 integrins were detected at lateral and basal cell membranes in contact with the supporting substrate (arrowheads); whereas in cells cultured on thin collagen films, 1 integrins were detected at apical membranes as well (arrowheads).

(2) Permeability of cultures to lanthanum nitrate. Cells cultured on collagen gels presented a dense precipitate of lanthanum at their apical membranes. As shown in Figures 3A and 3B, no precipitate was detected in the extracellular space between adjacent HERC and 16 HBE 14o- cells, suggesting that, under such growing conditions, cells were linked by tight junctions that prevented impregnation of intercellular spaces by lanthanum. In contrast, in cells cultured on thin collagen films, lanthanum was often detected in intercellular spaces, delineating adjacent cells (Figures 3C and 3D).

View larger version (189K): [in this window] [in a new window]   Figure 3.   TEM micrographs showing permeability of epithelial respiratory cells to lanthanum. In HERC (A) and 16 HBE 14o- cell (B) cultures on thick collagen gels (CG), lanthanum precipitates were seen exclusively at apical membranes of upper cells. Arrowheads in A and B point to the exclusion of lanthanum from the intercellular space between adjacent HERC and 16 HBE 14o- cells, respectively. In both HERC (C) and 16 HBE 14o- cells (D) cultured on Thermanox coverslip (TC) coated with collagen films, lanthanum penetrated into the intercellular spaces (arrowheads), delineating adjacent cells. Bars: A and C, 20 µm; B and D, 80 µm.

(3) TER. TER of both cells cultured on thick gels were significantly higher (P < 0.001 and 0.01 for HERC and 16 HBE 14o- cells, respectively) than resistance of cells cultured on thin collagen films (Table 1). TER assays also revealed that HERC cultures were significantly more sealed than 16 HBE 14o- cultures (P < 0.001).

(4) Expression of ZO-1 protein. Cells cultured on thick collagen gels often exhibited a positive pericellular labeling with a monoclonal antibody against ZO-1, a protein commonly found in junctional complexes (Figures 4A and 4B). The labeling of cells cultured on thin collagen films was usually faint and patchy (Figures 4C and 4D).

View larger version (84K): [in this window] [in a new window]   Figure 4.   Immunofluorescence localization of ZO-1 protein in HERC (A) and 16 HBE 14o- cell (C) cultures on thick collagen gels. The labeling was identified uniformly along membranes of adjacent cells (arrows), as well as in cell cytoplasm. In HERC (B) and 16 HBE 14o- cells (D) cultured on thin collagen films, ZO-1 labeling (arrows) was faint and restricted to membranes of a few cells.

P. aeruginosa Internalization by Respiratory Epithelial Cells Varies with the Degree of Cell Polarity and Culture Permeability

No IC bacteria could be detected by the gentamicin exclusion assay in polarized and tight HERC and 16 HBE 14o- cells cultured on thick collagen gels after cell exposure to bacterial suspensions for 1 or 4 h (Figure 5). In contrast, high bacterial concentrations were detected in lysates from poorly differentiated and poorly polarized HERC and 16 HBE 14o- cells cultured on thin collagen films (Figure 5). P. aeruginosa internalization by HERC was maximal after 1 h incubation, whereas bacterial entry into 16 HBE 14o- cells increased significantly (P < 0.05) after 4 h incubation. To assess the specificity of P. aeruginosa internalization, HERC and 16 HBE 14o- cells were incubated with two poorly adhesive bacterial controls centrifuged onto the cell monolayers, to circumvent their low adhesiveness. Concentrations of intracellular PAO-1 and PAK P. aeruginosa were significantly higher (P < 0.001) than concentrations of control P. aeruginosa PAK-N1 and E. coli DH5 in both HERC and 16 HBE 14o- cells (Figure 6). Because bacterial strains exposed to cell cultures were not equally uptaken, P. aeruginosa internalization by poorly polarized cells seems to be a selective process depending on the affinity of specific ligands from few bacterial strains for newly exposed cell receptors.

View larger version (20K): [in this window] [in a new window]   Figure 5.   Internalization of PAO-1 and PAK P. aeruginosa by epithelial respiratory cells. No IC bacteria were detected in HERC and 16 HBE 14o- cells cultured on thick collagen gels, even after incubation periods of 4 h. Increasing the incubation period of cells cultured on thin collagen films with bacteria from 1 to 4 h significantly increased the concentration of IC microorganisms in 16 HBE 14o- cells but did not modify bacterial internalization by HERC. Data represent means ± SD of four experiments carried out in triplicate.

View larger version (42K): [in this window] [in a new window]   Figure 6.   Internalization of different bacterial strains by epithelial respiratory cells cultured on thin collagen films. Concentrations of IC PAO-1 and PAK P. aeruginosa were significantly higher than concentrations of PAK-N1 P. aeruginosa and DH5 E. coli centrifuged at 1,000 × g for 10 min onto cell monolayers to circumvent their low adhesiveness. Data represent means ± SD of three experiments carried out in triplicate.

The presence of bacteria in the IC compartment of cells cultured on collagen gels or films was also investigated by immunoperoxidase labeling of extracellular bacteria and by counterstaining IC microorganisms with Giemsa stain. Bacteria were seen adherent to both HERC and 16 HBE 14o- cells cultured on thin films and to 16 HBE 14o- cells cultured on thick gels. In contrast, IC microorganisms could be observed only in cells cultured on collagen films (data not shown).

Disruption of Tightness of Epithelial Respiratory Cell Cultures Increases P. aeruginosa Internalization

P. aeruginosa PAO-1 internalization by HERC and by 16 HBE 14o- cells grown on collagen films was enhanced by factors of 7 and 3, respectively, when cells were pretreated with EDTA to disrupt intercellular junctions and further expose basolateral receptors to bacteria (Figure 7). Surprisingly, the concentration of IC PAK in cultures treated with EDTA did not differ from the concentration in control cells. No bacteria could be detected by the gentamicin exclusion assay in HERC or 16 HBE 14o- cells cultured on collagen gels after EDTA treatment. However, TEM observation showed that EDTA-treated cultures no longer excluded lanthanum nitrate but exhibited bacteria both in the intercellular spaces and the IC compartments (Figure 8).

View larger version (36K): [in this window] [in a new window]   Figure 7.   Effect of EDTA on P. aeruginosa internalization by human epithelial respiratory cells cultured on thin collagen films. Concentration of IC PAO-1 strain in both HERC and 16 HBE 14o- cells increased significantly after treatment of monolayers with EDTA to disrupt cell junctions further. No increase in concentration of IC PAK bacteria could be detected. Data represent means ± SD of three experiments carried out in triplicate.

View larger version (213K): [in this window] [in a new window]   Figure 8.   TEM micrograph showing that, after treatment with EDTA, 16 HBE 14o- cells cultured on thick collagen gels (CG) no longer excluded lanthanum that could be seen delineating adjacent cells (arrowheads). Bacteria were detected in the intercellular space, adherent to cellular basal membranes, as well as in the IC compartment (arrow). Bar: 50 µm.

Fate of P. aeruginosa after Internalization by Respiratory Cells

The ability of P. aeruginosa to survive in epithelial respiratory cells was assayed by a modification of the invasion test. As shown in Figure 9A, viable IC bacteria could be detected in poorly polarized cells as late as 24 h after infection. TEM examination of infected cells showed partial disruption of endosomal membranes and the presence of mitochondria surrounding the endocytic vacuole (Figure 9B). IC bacteria appeared to be intact because they maintained their rod shape, cytoplasmic ultrastructural organization, and electron density.

View larger version (55K): [in this window] [in a new window]   Figure 9.   (A) Time course of the concentration of IC bacteria in 16 HBE 14o- cells cultured on thin collagen films. Data represent means ± SD obtained in two experiments carried out in triplicate. (B) Transmission electron microscopy of cells infected with PAO-1 P. aeruginosa for 2 h, fixed in 2% glutaraldehyde, 4% paraformaldehyde, and 5mM CaCl2 in 0.1 M cacodylate buffer and postfixed in 1% OsO4, 5 mM CaCl2, and 0.8% potassium ferrocyanide (37). Note areas of lysis of endosomal membranes (arrowheads) and surrounding mitochondria (arrows). Bar: 20 µm.

Viability of Infected Respiratory Cells

To ascertain whether persistence of infection would affect host-cell viability, cells harboring IC bacteria for different periods were compared with control noninfected, gentamicin-treated cells in their mitochondrial enzymatic activity detected by the MTT test. As shown in Table 2, bacterial IC residence for up to 3 h seemed to stimulate host-cell metabolic activity because infected cells produced significantly more formazan than did noninfected cells. However, after 7 h of bacterial IC residence, the activity of infected cells decreased significantly. The effect of bacterial infection in host-cell viability was also investigated by LDH assay. As shown in Figure 10, the release of LDH by 16 HBE 14o- cells increased with the infection period, being significantly higher in infected cells than in control cells at 24 h after infection (P < 0.05). These data confirm that cell injury occurred as a consequence of P. aeruginosa internalization.

View larger version (18K): [in this window] [in a new window]   Figure 10.   Time course of LDH release by noninfected 16 HBE 14o- cells cultured on thin collagen films and by cells harboring IC PAK and PAO-1 P. aeruginosa for different periods. Data represents means ± SD obtained in a typical experiment carried out in quadruplicate. *P < 0.05.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Cultured mammalian cells provide a simple and easily controlled model for investigating bacterium-eukaryotic cell interaction (26). The type of surface on which cells are grown is an essential element in the differentiation and polarization of cultured epithelial cells. Supporting substrates also influence membrane receptors and cellular architecture, two factors that regulate gene expression and ultimately lead to the presentation of specific phenotypic characteristics. Thick collagen gels provide the extracellular matrix components and the optimal physical properties that activate cell-surface receptors and allow the assumption of a cuboidal polarized, rather than a flattened, epithelium (27). In the present study we have demonstrated that HERC and 16 HBE 14o- cells cultured on thick collagen gels are polarized inasmuch as they express CFTR protein at their apical membranes and express 1 integrins exclusively at their lateral and basal membranes in contact with supporting gels. The tightness of the cultures was demonstrated by pericellular labeling with the anti-ZO-1 protein, by their high transepithelial resistance, and by the exclusion of lanthanum nitrate. Although ciliated cells could not be detected in 16 HBE 14o- cultures, cells cultured on thick gels may be considered representative of cells usually found in airway surface epithelia. In contrast, cells cultured on thin collagen films were nondifferentiated and nonpolarized; they did not express CFTR protein at their apical membranes and did not establish tight junctions, thus allowing lanthanum nitrate to penetrate between adjacent cells. As a consequence, basolateral receptors (such as 1 integrins) were exposed at their apical membranes. In a recent study on the dynamics of the restoration of the epithelial barrier integrity during the wound-repair process, it was shown that repairing cultures of epithelial respiratory cells do not exclude lanthanum nitrate for up to 1 to 2 d after complete wound closure (15) and that, in contrast with cells from nonwounded areas of the monolayers, repairing epithelial respiratory cells do exhibit apical 1 integrins (16). Accordingly, because both HERC and 16 HBE 14o- cells cultured on thin collagen films are untight and nonpolarized, they can be considered representative of cells found in repairing respiratory epithelia.

A striking characteristic of P. aeruginosa is its inability to infect healthy tissues and its ablity to cause persistent infections in wounded tissues. We have previously shown that polarized HERC cultured on thick collagen gels are resistant to P. aeruginosa adherence (28), whereas in this present study we show that polarized 16 HBE 14o- cells are permissive to P. aeruginosa adherence. This difference in susceptibility to P. aeruginosa adherence between polarized HERC and 16 HBE 14o- cells may depend on alterations in the repertoire of apical receptors for bacterial ligands expressed by 16 HBE 14o- cells because the processes of cell transformation may produce several mutations and rearrangements in the cell genome (26). Moreover, cell lines that are repeatedly passed in culture may continue to develop new rearrangements (26). In contrast with polarized cells, poorly polarized HERC and 16 HBE 14o- cells cultured on thin collagen films were equally susceptible to bacterial entry and survival in the IC compartment of infected cells. These findings are consistent with our prior observations (12) and with electron microscopic studies carried out by Stern and colleagues (31) on the interaction of P. aeruginosa with wounded rabbit corneal epithelium. In their study, nonpolarized repairing cells were seen to engulf adherent bacteria by the formation of pockets surrounding the microorganisms. Bacteria could not adhere to or penetrate into intact epithelium. Recently, Fleiszig and associates (17) also reported that epithelial cell polarity affects susceptibility to P. aeruginosa invasion. Our results demonstrate that, in addition to cell polarity, the integrity of the epithelial barrier is a major determinant in cell susceptibility to P. aeruginosa invasion. The polarity of epithelial cells and the integrity of the epithelial barrier are likely to contribute to defense against P. aeruginosa infection by isolating basolateral proteins and restricting the repertoire of potential interactions that may occur between luminal bacteria and the epithelial cell surface. Available data suggest that different bacterial pathogens, such as Shigella flexneri (32, 33) and Listeria monocytogenes (34), preferentially interact with epithelial cells by means of basolateral ligands exposed following transient microdiscontinuities of epithelia. Perdomo and coworkers (33) and McCormick and colleagues (35) reported that neutrophil migration across intestinal cell monolayers produced reversible disruption of intercellular junctions, resulting in epithelial depolarization and localization of 1 integrin at apical membranes. Unmasking of 1 integrins facilitated cell invasion by Yersinia pseudotuberculosis (35).

While characterizing the fate of intracellular P. aeruginosa in nonpolarized respiratory cells, we observed bacteria in vacuoles presenting partially lyzed membranes similar to what we have described in human endothelial cells harboring IC PAO-1 P. aeruginosa (36). Also similar to our findings in infected endothelial cells, we observed mitochondria concentrated around the bacteria-containing vacuole. In their study on P. aeruginosa internalization by CF F508 epithelial respiratory cells, Ko and associates (10) also observed IC bacteria in contact with a mitochondrion. Mitochondria wrap around Legionella pneumophila, and Toxoplasma gondii-containing vacuoles have been associated with provision of energy necessary for parasite IC multiplication (37, 38). Because only minor P. aeruginosa multiplication in infected respiratory cells could be detected, the meaning of mitochondria accumulation around endocytic vacuoles remains to be determined.

Our finding that poorly polarized epithelial respiratory cells, which do not express CFTR protein at their apical membranes, do internalize P. aeruginosa after short incubation periods is in contrast with recent reports from Pier and coworkers (6, 7) but is in accord with results from Ko and colleagues (10). In this latter study, binding and internalization of P. aeruginosa by cells from two different CF F508 lines were significantly higher than binding and internalization by normal non-CF cells, suggesting that neither adherence nor internalization of P. aeruginosa depends on bacterial interaction with apically expressed CFTR protein.

This report demonstrates that nonpolarized epithelial cells do not exhibit CFTR at their apical membranes but are highly susceptible to P. aeruginosa adherence, invasion, and survival. By extrapolating our results to the in vivo setting, we suggest that during repair of wounded tissues, untight and nonpolarized respiratory epithelia may remain susceptible to bacterial paracellular penetration, interaction with cell basolateral receptors, and entry. In the IC compartment, bacteria are protected from host defense mechanisms and this localization may ultimately favor persistent infections.