Necrosis of lung epithelial cells during infection with Mycobacterium tuberculosis is preceded by cell permeation

Abstract

Mycobacterium tuberculosis establishes infection, progresses towards disease, and is transmitted from the alveolus of the lung. However, the role of the alveolar epithelium in any of these pathogenic processes of tuberculosis is unclear. In this study, lung epithelial cells (A549) were used as a model in which to examine cytotoxicity during infection with either virulent or avirulent mycobacteria in order to further establish the role of the lung epithelium during tuberculosis. Infection of A549 cells with M. tuberculosis strains Erdman and CDC1551 demonstrated significant cell monolayer clearing, whereas infection with either Mycobacterium bovis BCG or Mycobacterium smegmatis LR222 did not. Clearing of M. tuberculosis-infected A549 cells correlated to necrosis, not apoptosis. Treatment of M. tuberculosis-infected A549 cells with streptomycin, but not cycloheximide, demonstrated a significant reduction in the necrosis of A549 cell monolayers. This mycobacterium-induced A549 necrosis did not correlate to higher levels of intracellular or extracellular growth by the mycobacteria during infection. Staining of infected cells with propidium iodide demonstrated that M. tuberculosis induced increased permeation of A549 cell membranes within 24 h postinfection. Quantitation of lactate dehydrogenase (LDH) release from infected cells further demonstrated that cell permeation was specific to M. tuberculosis infection and correlated to A549 cellular necrosis. Inactivated M. tuberculosis or its subcellular fractions did not result in A549 necrosis or LDH release. These studies demonstrate that lung epithelial cell cytotoxicity is specific to infection by virulent mycobacteria and is caused by cellular necrosis. This necrosis is not a direct correlate of mycobacterial growth or of the expression of host cell factors, but is preceded by permeation of the A549 cell membrane and requires infection with live bacilli.

FIG. 1

(A) (inset) Optical densities of histone-DNA complexes released in the culture supernatants of epithelial cell monolayers exposed to the pore-forming detergent digitonin (at 1%). □, untreated cells; , 1% digitonin-treated cells. Data shown are means from two experiments run in triplicate. Error bars, standard deviations. (B) Optical densities of histone-DNA complexes released in the culture supernatants of infected epithelial cell monolayers. All monolayers were infected at an MOI of 10:1. □, M. bovis BCG; , M. tuberculosis strain Erdman; , M. tuberculosis strain CDC1551; , M. smegmatis. Data shown are means from three different experiments of triplicate monolayers. Error bars, standard errors. The values obtained were normalized by subtraction of background levels of necrosis of uninfected A549 cell monolayers at each time point.

FIG. 2

(A) (inset) Optical densities of histone-DNA complexes from the cell lysates of epithelial cell monolayers exposed to 0.005 mM camptothecin. □, untreated cells; , 0.005 mM camptothecin-treated cells. Data shown are means from two experiments run in triplicate. Error bars, standard deviations. (B) Optical densities of intracellular histone-DNA complexes from the cell lysates of infected epithelial cell monolayers. All monolayers were infected at an MOI of 10:1. □, M. bovis BCG; , M. tuberculosis strain Erdman; , M. tuberculosis strain CDC1551; , M. smegmatis. Data shown are means from four different experiments of triplicate monolayers. Error bars, standard errors. The values obtained were normalized by subtraction of background levels of apoptosis of uninfected A549 cell monolayers at each time point.

FIG. 3

(A) Optical densities of histone-DNA complexes released in the culture supernatants of infected epithelial cell monolayers with or without treatment using the bacterial protein synthesis inhibitor streptomycin. □, uninfected, streptomycin-treated cells; , M. tuberculosis CDC1551-infected, untreated cells; , M. tuberculosis CDC1551-infected, streptomycin-treated cells. (B) Optical densities of histone-DNA complexes released in the culture supernatants of infected epithelial cell monolayers with or without treatment with the eukaryotic protein synthesis inhibitor cycloheximide. □, uninfected, cycloheximide-treated cells; , M. tuberculosis CDC1551-infected, untreated cells; , M. tuberculosis CDC1551-infected, cycloheximide-treated cells. All monolayers were infected at an MOI of 10:1. Data shown are means from two different experiments of triplicate monolayers. Error bars, standard deviations. The values obtained were normalized by subtraction of background levels of necrosis of uninfected A549 cell monolayers at each time point.

FIG. 4

(A) CFU obtained from intracellular growth of Mycobacterium spp. within A549 cells. Time zero is 6 h postinfection. P < 0.01 for differences between all three mycobacteria at all time points tested, except that no difference was observed between any mycobacteria at 24 h, no difference was observed between M. tuberculosis strain Erdman and M. tuberculosis strain CDC1551 at 24, 48, 72, and 96 h, and no difference was observed between M. bovis BCG and M. tuberculosis strain Erdman at 120 h. (B) CFU obtained from extracellular growth of Mycobacterium spp. cocultured with A549 cells. P < 0.01 at 72 h for M. tuberculosis strain CDC1551 versus M. bovis BCG and at 24 h for M. tuberculosis strain Erdman versus M. tuberculosis strain CDC1551. ▴, M. tuberculosis strain CDC1551; ■, M. tuberculosis strain Erdman; ●, M. bovis BCG. All experiments were performed in duplicate and repeated twice. Error bars, standard deviations.

FIG. 5

LDH release from A549 cell monolayers infected with Mycobacterium spp. over 3 days. □, M. bovis BCG; , M. tuberculosis strain Erdman; , M. tuberculosis strain CDC1551. All monolayers were infected at an MOI of 10. All experiments were performed in triplicate. Error bars, standard deviations.

FIG. 6

Optical densities of histone-DNA complexes released in the culture supernatants of epithelial cell monolayers treated with inactivated preparations of M. tuberculosis strain CDC1551. , heat-killed M. tuberculosis; , γ-irradiated M. tuberculosis. All monolayers were treated with an inoculum equivalent to an MOI of 10. All experiments were performed in triplicate. Error bars, standard deviations. The values obtained were normalized by subtraction of background levels of necrosis of untreated A549 cell monolayers at each time point.

FIG. 7

(A) Optical densities of histone-DNA complexes released in the culture supernatants of epithelial cell monolayers treated with subcellular fractions of M. tuberculosis strain CDC1551. (B) Optical densities of intracellular histone-DNA complexes from the cell lysates of epithelial cell monolayers treated with subcellular fractions of M. tuberculosis strain CDC1551. □, CF; , CW; ▩, WCL. All monolayers were treated with a protein inoculum equivalent to an MOI of 100 bacilli per A549 cell. All experiments were performed in triplicate. Error bars, standard deviations. The values obtained were normalized by subtraction of background levels of necrosis and apoptosis of untreated A549 cell monolayers at each time point.