The Efficiency of the Translocation of Mycobacterium tuberculosis across a Bilayer of Epithelial and Endothelial Cells as a Model of the Alveolar Wall Is a Consequence of Transport within Mononuclear Phagocytes and Invasion of Alveolar Epithelial Cells

Bacterial strains.

M. tuberculosis H37Rv was obtained from American Tissue Culture Collection (Rockville, Md.). Mycobacterium avium strain 101 was isolated from the blood of an AIDS patient (13). Mycobacterium bovis BCG strain Pasteur was a gift from Brigitte Gicquel (Institut Pasteur, Paris, France). All the strains were grown on Middlebrook 7H11 agar, and after selection of pure colonies they were transferred to 7H9 broth for 7 days at 37°C in 5% CO₂. Cultures were passed through a 23-gauge needle 10 times, and then the tube containing the suspension was kept on the bench for 5 min. The top half of the suspension was obtained and stained for viability using the LIVE-DEAD assay (Molecular Probes, Eugene, Oreg.) and used in the experiments (5). The suspension was constituted of dispersed bacteria.

Purification of monocytes.

Blood from purified protein derivative-negative donors was collected with heparin-containing tubes and submitted to a process of purification as previously described (2). Monocytes were then enriched by adherence to the plastic and subsequently resuspended by treatment of the monolayer with 0.1 M EDTA for 15 min. Monocytes were washed and resuspended in RPMI 1640 supplemented with 10% fetal bovine serum (FBS) and 2 mM L-glutamine. Monocytes in suspension in RPMI 1640 (Gibco BRL, Grand Island, N.Y.) in a Teflon jar were examined for viability by using trypan blue exclusion as described (2). Cells were approximately 95% viable.

Tissue culture.

A549 human type II alveolar epithelial cells were purchased from American Type Culture Collection. Approximately 10⁵ cells were suspended in RPMI 1640 medium supplemented with 10% FBS, and 10⁴ cells were added to each well of a 24-well tissue culture plate (Costar Corp., Cambridge, Mass.). Cells were allowed to grow to 100% confluence or, in some experiments, 80% confluence. The EAhy926 human endothelial cell line (a permanent endothelial cell line established by the hybridization of human umbilical vein endothelial cells and A549 cells) was kindly provided by Cora-Jean Edgell (University of North Carolina) (9). EAhy926 cells were grown in Dulbecco’s minimum essential medium (Difco, Detroit, Mich.) supplemented with 10% FBS, endothelial growth factor (Sigma Chemicals Co., St. Louis, Mo.), and 10 U of heparin per ml. The cells were cultured to complete confluence or, in some experiments, to 80% confluence. EAhy926 expresses human factor VIII-related antigen and produces prostacyclin.

Transwell-polarized monolayer.

A transwell insert (Costar) with 3.0-μm pores was placed in each well of a 24-well tissue culture plate (Costar). To construct a polarized monolayer, the transwell insert was inverted (i.e., the bottom part was placed towards the top), and 10⁵ EAhy926 cells were placed on the bottom side of the membrane and allowed to be established (usually 48 h). The cells were then allowed to grow to confluence (approximately 5 × 10⁸ A549 cells and 7 × 10⁸ EAhy926 cells). The transmembrane resistance was measured every 2 days until it reached approximately 520 ± 63 Ω/cm² for A549 cells and 486 ± 35 Ω/cm² for EAhy926 cells, and the integrity of the monolayers was determined with ¹²⁵I-inulin (ICN, Costa Mesa, Calif.) and, in some experiments, 0.1 ml of 1% blue dextran 2000 (Pharmacia, San Diego, Calif.). The content of the lower chamber after 2 h was retrieved, and the amount of radioactive material was measured in a gamma radiation counter or the optical density was determined.

To construct a bilayer culture, filters were seeded both in upside-down (top up) and upright positions. The A549 alveolar epithelial cells seeded in the upper part and EAhy926 cells were seeded in the bottom. Briefly, filters were placed in inverted position (upside down) and 10⁵ EAhy926 cells were seeded first in F-12 medium on the bottom side of the filter, supplemented with 10% FBS and endothelial growth factor (0.5%) (Sigma), and allowed to establish for 48 h. Then, the filter was inverted back to the upright position, and 10⁵ A549 cells were seeded in F-12 medium supplemented with 10% FBS. Endothelial growth factor was added to the bottom chamber every time the medium was changed (every 4 days). The confluence of the monolayers and the resistance were monitored daily. The bilayer was considered ready for use when the transmembrane resistance achieved approximately 546 ± 33 Ω/cm³, which took an average of 10 to 12 days. The permeability of the bilayer was measured as previously reported (21) and as described above.

Infection of monolayers and bilayers.

Monolayers were infected with 10⁷ bacteria (approximate ratio, 10 bacteria to one cell) for different periods of time. Then, the supernatant was removed, and the wells were extensively washed with Hank’s buffered salt solution (HBSS). Afterward, amikacin (200 μg/ml) was added for 2 h at 37°C to kill the extracellular bacteria remaining in the well. Amikacin at a concentration of 200 × the MIC kills the majority of bacteria in the wells and inhibits the adherence of the microorganisms that survive (3, 17). The monolayer was then lysed with 1% Triton X-100 in HBSS for 15 min, and the suspension was plated onto 7H11 agar after dilution.

In the case of bilayers, the medium was removed, the cells were washed once with HBSS, and medium and 10⁵ bacteria were added to the top chamber. Passage of the bacteria across the layers of cells was monitored by collecting the supernatant in the bottom chamber. Transmembrane resistance was also monitored during the course of the experiment.

In some assays, M. tuberculosis was used after passage in either A549 epithelial cells or human monocytes for 3 days. To obtain M. tuberculosis, A549 cells or human monocyte monolayers were infected with H37Rv or H37Ra (100 bacteria to one cell) for 1 h, and the extracellular bacteria were removed afterwards by two consecutive washings. The intracellular bacteria were allowed to replicate and were obtained after 3 days by a previously described method (5). The viability of the bacteria was determined by using the LIVE-DEAD assay (5), and the number of organisms and the purity of preparation were determined by plating bacteria on 7H11 agar and microscopic examination, as described (5). Bacteria obtained this way were kept at 4°C for 24 h before the assay. This period at 4°C did not change the characteristics of the inoculum (data not shown).

Translocation of monocytes.

To evaluate the role of blood monocytes and alveolar macrophages in the transport of M. tuberculosis from the alveolar space to the blood, we used infected and uninfected blood monocytes and measured their translocation across the bilayer (epithelial and endothelial cells) with the transwell system. Monocytes were used uninfected or infected with bacteria for 1 h (10 bacteria per monocyte) in suspension under constant rotation at 37°C. After 1 h, monocytes were centrifuged at 400 × g for 5 min, and the pellet was examined for viability and the approximate percentage of intracellular bacteria by trypan blue exclusion and acid-fast staining by phase-contrast microscopy, respectively. Only preparations that contained more than 90% viable monocytes were used in the experiment. In addition, only preparations that contained at least 60% of the monocytes infected (as determined by counting 300 cells in 10 fields) by one or more bacteria were used in the assays (as evidenced by phase-contrast microscopy with an image-enhancing system).

Infected monocytes were added to the top chamber in the transwell, and translocation was measured as the number of monocytes that crossed the bilayer over time. In some experiments, A549 epithelial cells in the transwell were infected with 10⁵ bacteria for 1 h, and then the supernatant containing extracellular bacteria was changed by the addition of fresh medium before monocytes were added.

Chemokine production and neutralization.

Previous work has shown that A549 cells produce interleukin 8 (IL-8) when infected with M. tuberculosis. To determine if two chemokines important for phagocytic cell migration, IL-8 and macrophage chemoattractant protein 1 (MCP-1), had a role in the chemotaxis of infected monocytes across the cell bilayer, we performed the assays described above in the presence and absence of anti-IL-8 (Biosource International, Camarillo, Calif.) and anti-MCP-1 (R & D Systems Minneapolis, Minn.) antibodies. In addition, the concentration of both chemokines in both the upper and bottom chamber supernatants was measured by assays purchased from Biosource International. Irrelevant antibodies, mouse anti-human immunoglobulin G (IgG) and rat anti-human IgG, were used as controls. Enzyme-linked immunosorbent assays had a sensitivity of 20 (MCP-1) and 5 (IL-8) pg/ml.

Role of receptors in the translocation of monocytes.

To determine the receptors used by monocytes to cross the alveolar wall, we carried out translocation assays with bilayer A549 epithelial cells and endothelial cells in the presence of mouse anti-human antibodies to CD11a (clone 12101, mouse IgG; Biosource), CD11b (clone L-MO-1, mouse IgG; Biosource), ICAM-1 (clone 84A6, mouse IgG; Biosource), anti-CD11c (clone 3.9, mouse IgG; Biosource), CD29 (clone PD-15, mouse IgG; Biosource), anti-VLA-4 (clone VD-46, Upstate Biotechnology, Inc., Lake Placid, N.Y.), anti-CD51 (clone 23C6, mouse IgG; Biosource), anti-CD26 (clone SMO, mouse IgM; Biosource), anti-CD14 (clone LO-MD-1, rat IgG; Biosource), and anti-CD47 (clone BRIC 126, mouse IgG; Biosource). Antibodies at three different concentrations were added 30 min before infection, and cells were incubated at 37°C. Then, the monolayers were infected for different periods of time, and translocation was measured as described above.

Statistical analysis.

Each experiment was repeated at least three times, and the results were analyzed using Student’s t test.

Does M. tuberculosis cross an A549 polarized monolayer?

To examine whether M. tuberculosis crosses a polarized intact A549 monolayer, we performed the assay in comparison to M. bovis BCG and measured translocation after different periods of time. As shown in Table 1, M. tuberculosis was capable of crossing the intact monolayer with greater efficiency than M. bovis BCG. Crossing by M. tuberculosis was significant only after 4 h of incubation. The transmembrane resistance was constantly measured and showed to be unaltered at 4 h. Five days after infection, 4% of the inoculum had translocated across the A549 membrane; however, the fall in the transmembrane resistance suggested that the integrity of the monolayer was compromised. In fact, by using blue dextran added to the upper chamber as a marker for membrane integrity, we were able to determine that the polarized membrane allowed the passage of a small amount of dextran.

M. tuberculosis invasion of endothelial cells.

The ability of M. tuberculosis to invade endothelial EAhy926 cells was first evaluated using a endothelial cell monolayer seeded on plastic. As shown in Table 2, M. bovis BCG and M. tuberculosis strains H37Rv and H37Ra invaded EAhy926 poorly, in agreement with the results reported with human lung endothelial cells (6). However, if M. tuberculosis was passed in A549 cells or human macrophages for 3 days, the efficiency of the invasion increased by more than 10-fold. The same behavior was not observed with M. bovis BCG (data not shown).

We then evaluated if the same sort of results were observed with monolayers of polarized EAhy926 cells. Cells were seeded upside down to mimic the contact between M. tuberculosis in the alveolar space and the basolateral surface of endothelial cells. As shown in Table 3, only M. tuberculosis previously passed in A549 cells was able to cross EAhy926 cells with significant efficiency.

Translocation across bilayer.

Translocation of M. tuberculosis across the bilayer was significantly greater than the translocation of M. bovis BCG. When M. tuberculosis used in the bilayer assay was derived from macrophages or A549 cells, the efficiency of translocation was significantly higher than with bacteria that were not passed in macrophages or A549 cells (Table 4), suggesting that an invasive phenotype might have emerged when the bacterium was exposed to the intracellular environment.

Mononuclear phagocytes translocate across the A549-EAhy926 bilayer.

To determine if mononuclear phagocytes would carry M. tuberculosis across the alveolar wall (a bilayer of epithelial and endothelial cells), we infected or did not infect human monocytes and added them to the upper chamber. Figure 1 schematically shows the model used. As shown in Table 5, infected monocytes cross the bilayer barrier very efficiently in contrast to uninfected ones. In some assays, A549 cells were infected with M. tuberculosis prior to the addition of the monocytes. Infection of A549 cells significantly increased monocyte translocation across the bilayer membrane (Table 5). Preliminary data (not shown) established that the passage of M. bovis BCG in A549 cells did not alter its ability to translocate.

Production of IL-8 and MCP-1.

To examine if infected A549 cells produced chemokines that induced monocyte migration, we determined the concentrations of IL-8 and MCP-1 in the supernatants of upper and lower chambers of a bilayer of A549 and EAhy926 cells (Fig. 1). As shown in Table 6, there was a significant increase of the chemokines in both chambers following infection. The lower chamber supernatant showed a greater augmentation of chemokine concentrations than the upper chamber.

To investigate the relative importance of both chemokines in the migration of monocytes across the bilayer, the bilayer was infected as described in Materials and Methods, followed by the addition of anti-IL-8 and anti-MCP-1 to both chambers, and then 10⁵ infected monocytes were added to the upper chamber and their translocation was determined. Figure 2 shows that the use of neutralizing concentrations of anti-MCP-1 but not anti-IL-8 significantly inhibited the migration of monocytes across the bilayer membrane.

Role of cell receptors in monocyte migration.

To identify surface molecules on A549 cells, EAhy926 cells, and monocytes that participate in the process of translocation across the bilayer, we used monoclonal antibodies against several molecules known to be expressed on the host cells and known to be involved in cell translocation. As shown in Table 7, monoclonal antibodies against CD11a, CD11b, and CD47 partially blocked monocyte translocation in a dose-response fashion (only a higher concentration is shown in Table 7). Use of anti-VLA-4 and anti-ICAM-1 had a limited but still significant effect on the crossing of monocytes.