Bacillus anthracis spore movement does not require a carrier cell and is not affected by lethal toxin in human lung models

Abstract

The lung is the entry site for Bacillus anthracis in inhalation anthrax, the most deadly form of the disease. Spores escape from the alveolus to regional lymph nodes, germinate and enter the circulatory system to cause disease. The roles of carrier cells and the effects of B. anthracis toxins in this process are unclear. We used a human lung organ culture model to measure spore uptake by antigen presenting cells (APC) and alveolar epithelial cells (AEC), spore partitioning between these cells, and the effects of B. anthracis lethal toxin and protective antigen. We repeated the study in a human A549 alveolar epithelial cell model. Most spores remained unassociated with cells, but the majority of cell-associated spores were in AEC, not in APC. Spore movement was not dependent on internalization, although the location of internalized spores changed in both cell types. Spores also internalized in a non-uniform pattern. Toxins affected neither transit of the spores nor the partitioning of spores into AEC and APC. Our results support a model of spore escape from the alveolus that involves spore clustering with transient passage through intact AEC. However, subsequent transport of spores by APC from the lung to the lymph nodes may occur.

1. Introduction

The lung is the site of entry for Bacillus anthracis in inhalation anthrax, the most deadly form of the disease. B. anthracis does not, however, cause disease in the lung [1–3]. It is well established that the pathogen first escapes the alveolus then passes through the lymphatic system to the mediastinal lymph nodes and lymphatic duct with subsequent dissemination in the bloodstream [4–6]. Four pathways, three involving cells, for alveolar escape have been debated. Firstly, macrophages were originally thought to be the most important cell type for transport, based on animal studies showing that alveolar macrophages (AM) ingest spores and data in some models showing that B. anthracis virulence toxins damage or impair these cells [4,7]. However, other studies have suggested that dissemination from the lung occurs in macrophage-depleted mice [2], indicating that macrophages may not play a key role in alveolar escape. Secondly, some evidence suggests a more prominent role for dendritic cells (DC). DC with internalized spores were observed in the thoracic lymph nodes of mice, although other cells containing spores were also observed [1,8], and DC can sample the airway luminal surface [9]. Thirdly, despite these data suggesting that a migratory or “carrier” cell (often referred to as a “Trojan horse”) is required for dissemination, there is also evidence based on isolated cell culture models for the involvement of a transcellular route of dissemination through the lung epithelium [10]. Using the human lung epithelial cell line A549 and primary human small airway cells, both of which internalize spores, Russell et al. [10] showed that spores survived and were translocated from the apical to basolateral side of the cell. Spore internalization in A549 cells involves interaction of the spore BclA protein with cell integrin α2β1 and complement C1q [11]. These results suggest that spores may cross the alveolar epithelial barrier and disseminate without the assistance of migratory cells [10]. Finally, in the “Jailbreak” model, it is proposed that the clustering of spores followed by germination and production of various B. anthracis virulence toxins causes epithelial damage which permits free spore passage not involving cells [12]. It should be noted that AEC are stationary and not considered carrier cells, yet are tacit players in all four pathways because they form the alveolar epithelial barrier through which the spores must pass.

B. anthracis produces exotoxins that are important in pathogenicity. As described above, it has been proposed these toxins play a role in early stages of inhalation anthrax. These toxins consist of edema factor (EF) and lethal factor (LF) which combine with protective antigen (PA) to form edema toxin (ET) and lethal toxin (LT), respectively. PA binds to cellular receptors to permit entry of EF and LF into the host cell [13]. EF, an adenyl cyclase, causes edema upon injection and impairs immune responses [14]. LF is a metalloproteinase, inhibits immune responses, and induces apoptosis in susceptible cells [15–17]. Our laboratory has also shown that LT decreases barrier function and impairs tight junction formation in primary human AEC [18]. It has also been demonstrated that these toxins may play a role in dissemination of B. anthracis from the lung [19].

Taken together, the evidence suggests that macrophages, DC and AEC may all be involved in dissemination, and that LT may have a role in the process. However, the extent to which each cell type is involved in initial alveolar escape has not been examined, especially in humans. Similarly, the role of LT has not been explored in humans. Quantitation of spore uptake and determination of spore location in the lung after initial spore exposure can provide evidence about which cells are involved, if any, in movement of the pathogen across the alveolar epithelial barrier. Significant spore internalization in APC (AM or DC) in the absence of free spores or AEC internalization would support a Trojan horse model of alveolar escape and subsequent movement and dissemination through lymphatics. Spore movement, in the absence of significant AEC internalization, particularly when spore germination is allowed, would support a Jailbreak model of alveolar escape. AEC internalization, with spore movement, but without germination, would support that alveolar escape likely occurs without a Trojan horse cell and without a true Jailbreak scenario occurring. Alterations of this process by B. anthracis toxins would favor a toxin-sensitive carrier in this process.

We have previously shown that B. anthracis spores interact with a human lung organ culture model and trigger an innate immune response [20]. The model uses human lung tissue from surgical pathology specimens or from donor lungs rejected for transplant. The tissue is inflated with low-melting/low-gelling point agarose, which preserves lung architecture and maintains cell viability for 10 days [21]. In addition, innate immune cytokine responses are observed and evaluable in this model [20–23]. The advantages of this system include: All cell types in native human lung are present in their normal configuration; immune responses to human pathogens can be measured in a human system; and, multiple variables can be tested in one experiment in tissue from one subject [24,25]. Disadvantages of the model include the lack of vascular flow, the lack of respiratory motion, and the uncertainty of tissue availability.

We used this in vitro system to track early events including spore uptake by APC and AEC, spore movement through the model, spore partitioning between these two cell types, and spore partitioning between these cells and the extracellular space. We also tested whether one of the B. anthracis virulence toxins (LT) or toxin components (PA) altered spore partitioning between these cell types. We confirmed these findings in a lung epithelial cell line model.

2. Materials and methods

2.1. Preparation and culture of PCLS

Human lung was obtained from patients undergoing resection for cancer in accordance with protocols approved by the Institutional Review Boards of the University of Oklahoma Health Sciences Center, Veterans Administration Hospital, Integris-Baptist Hospital, St. Anthony’s Hospital, and Mercy Health Center, all of Oklahoma City, OK. Only tissue without tumor was used. Healthy, non-transplantable, human lung was also obtained from the International Institute for the Advancement of Medicine (IIAM, Jessup, PA). The preparation and culture of PCLS has been described previously [20,21,23]. Briefly, the subsegmental bronchi of fresh lung tissue were cannulated, and segments were gently inflated with 37 °C lung slice medium (LSM) containing 1.5% Sea Plaque agarose (Lonza). LSM consisted of minimal essential medium supplemented with 1.0 μg of bovine insulin/ml, 0.1 μg of hydrocortisone/ml, 0.1 μg of retinyl acetate/ml, 50 μg of gentamicin/ml, 100 U of penicillin/ml, 100 μg of streptomycin/ml, and 0.250 μg of amphotericin B/ml. After the agarose solidified, 1 cm diameter cores were prepared and sliced into 250 μm thick sections using a Krumdieck Tissue Slicer (Alabama Research and Development). Slices were incubated in 0.5 ml of LSM in 24-well plates at 37 °C in 5% CO₂. The LSM was changed at 2-day intervals and immediately prior to B. anthracis spore and toxin exposure.

2.2. Culture of A549 epithelial cells

The human alveolar type II epithelial cell line A549 from the American Type Culture Collection (CCL-185) was used in spore internalization experiments. A549 cells were propagated in Dulbecco’s modified Eagle medium containing 10% fetal bovine serum (FBS), 2 mM L-glutamine, and 80 μg/ml gentamicin (DMEM-10) at 37 °C in 5% CO₂. For experiments, cells were grown to 90% confluence on cover slips in DMEM-10.

2.3. Preparation and labeling of B. anthracis spores

B. anthracis, Sterne strain 7702 (pX01, pX02-), a gift of Dr. Jimmy Ballard, was grown at 37 C+ in Luria–Bertani broth and streaked onto AK Agar #2 sporulating slants. Bacteria were incubated for three weeks at 30 °C, and washed with chilled, deionized water. The resultant spore suspension was heated at 65 °C for 30 min and washed 5 times. The final spore pellet was resuspended in chilled, sterile deionized water, and the colony forming units per milliliter (cfu/ml) titer was determined by plate counts initially and prior to use.

Spores were labeled with Cy5 fluorescent dye for experiments by conjugation with Bis-reactive, N-hydroxysuccinimido ester Cy5 dye (GE) according to the manufacture’s protocol. Unconjugated dye was removed by repeated washes. Spore viability was not diminished by the fluorophore by plate counts. For PCLS, the titers (cfu/ml) of the spore preparations were determined by plate counts before each use, and for A549 cells the concentration (spores/ml) of the labeled spores was determined by counts of fluorescent spores using a hemocytometer.

2.4. Treatment of PCLS and A549 cells with B. anthracis toxin and spores

After overnight culture of PCLS, the slices to be harvested at a given incubation time were distributed to separate 24-well plates to minimize disturbing the spore-exposed slices during subsequent incubations, and the medium was replaced with fresh LSM containing 50 μg/ml gentamicin which prevents bacterial survival or replication and does not promote germination [20,26]. For treatments, the LSM was amended with 2 μg/ml protective antigen (PA) or 2 μg/ml lethal toxin (LT, 2 μg/ml PA plus 2 μg/ml LF, List Biological Laboratories). Negative control slices (“no treatment”, NT) received an equal volume of the protein diluent, sterile deionized water. PCLS were exposed to 1 × 10⁶ cfu/slice Cy5-labeled Sterne spores diluted in 10 μl LSM plus gentamicin with or without PA or LT. The 10 μl spore suspension was added over the entire slice surface to distribute the spores as evenly as possible, and the plates were carefully transported to the incubator to diminish spore loss from the upper slice surface. After incubation, the PCLS were washed with sterile Dulbecco’s phosphate buffered saline (PBS) to remove unbound spores and then fixed in fresh 2% paraformaldehyde. Following fixation, the PCLS were washed and stored at 4 °C in PBS +0.02% sodium azide.

A549 cell monolayers were grown overnight on cover slips, and the medium was replaced with DMEM-10 containing 50 μg/ml gentamicin for experiments. For anthrax toxin treatments, 2 μg/ml PA or 2 μg/ml LT (=2 μg/ml PA plus 2 μg/ml LF) was added to the DMEM-10. An equal volume of protein diluent was added to the negative control (NT) monolayers. A549 cells with or without PA or LT in the growth medium were exposed to Cy5-labeled Sterne spores mixed into the growth medium prior to addition at a multiplicity of infection (MOI) of 10 spores per cell, and incubated for the indicated times at 37 °C, 5% CO2. Afterward, the monolayers were washed with PBS and fixed for 15 min at room temperature in 2% paraformaldehyde. Following fixation, the A549 monolayers were washed two additional times with PBS and immunofluorescent stained or stored at 4 °C in PBS +0.02% sodium azide for later staining.

2.5. Immunofluorescent staining of PCLS

Fixed PCLS were washed twice in PBS, permeabilized for 2 h in 0.3% Triton-X in PBS at room temperature, washed with 0.05% Tween-20 and 0.01% saponin in PBS (wash buffer), and then blocked in 5% donkey serum (Dako) in wash buffer (blocking buffer). Blocking buffer was replaced with the anti-HLA-DR plus anti-pan-cytokeratin, or anti-ZO-1, or the anti-caveolin plus anti-CD31 (PECAM-1) primary antibody or negative control antibodies in blocking buffer (0.5 ml/slice) and incubated for 2 days at 4 °C. Then PCLS were washed 4 times in wash buffer, and twice in blocking buffer. PCLS were then incubated with fluorescent conjugated secondary antibodies in blocking buffer (0.5 ml/slice) for 2 days at 4 °C. PCLS were then washed 4 times at RT for 20 min and then stained with 4′, 6-diamidino-2-phenylindole (DAPI, 1.0 μg/ml) in PBS at RT for 1 h and washed in PBS. Whole mounts of PCLS were made using ProLong Gold Antifade Reagent (Thermo-Fisher). Primary antibodies and concentrations were as follows: Anti-human HLA-DR mouse monoclonal antibody, clone LN3 (Biolegend, 327002), 0.5 μg/ml; mouse monoclonal antibody IgG2b, clone X0944 (Dako, X0944), 0.5 μg/ml (isotype, negative control for anti-HLA-DR); anti-human pan-cytokeratin rabbit polyclonal antibody (Thermo-Fisher, 18-0059), 20.0 μg/ml; and normal rabbit polyclonal antibody IgG (Santa Cruz, Sc-2027), 20.0 μg/ml (negative control for anti-cytokeratin) or 6.5 μg/ml (negative control for anti-caveolin); rabbit anti-human caveolin (Santa Cruz), 6.5 μg/ml; mouse anti-human CD31 monoclonal antibody, clone JC70A (Dako, M0823), 10 μg/ml; mouse anti-human ZO-1 monoclonal antibody, clone ZO1-1A12 (Thermo-Fisher, 33-9100), 10 μg/ml; and mouse IgG₁ (Dako, X0931), 10 μg/ml (isotype, negative control for anti-ZO-1 and anti-CD31). The secondary antibodies and concentrations were as follows: Alexa Fluor 488 conjugated anti-mouse IgG2b goat polyclonal antibody (Jackson Immunoresearch, 115-545-207), 7.5 μg/ml, and Cy3 conjugated anti-rabbit IgG donkey polyclonal antibody (Jackson Immunoresearch, 711-166-152), 7.5 μg/ml; Cy3 conjugated anti-mouse IgG donkey polyclonal antibody (Jackson Immunoresearch, 711-166-150), 7.5 μg/ml, and AF488 conjugated anti-rabbit IgG donkey polyclonal antibody (Jackson Immunoresearch, 711-546-152), 7.5 μg/ml.

2.6. Immunofluorescent staining of Cy5-spores and A549 cells

To differentiate intracellular and extracellular bacterial endospores we used the method of Russell with modifications [27]. To detect extracellular spores, the fixed A549 monolayers were blocked at RT for 1 h in 5% fetal bovine serum (FBS) diluted in PBS. The blocking buffer was replaced with the primary antibody anti-Cy5 mouse monoclonal antibody (Santa Cruz, sc-166896), 1.0 μg/ml, diluted in 5% FBS in PBS (0.5 ml/well) and hybridization continued for 1 h at 37 °C. Then monolayers were washed 3 times at RT for 5 min in PBS, and fluorescent conjugated secondary antibody Alexa Fluor 488 conjugated anti-mouse IgG (H + L) goat polyclonal antibody (Thermo-Fisher, A11001), 5.0 μg/ml diluted in 5% FBS in PBS (0.5 ml/well) was added, and the cells were hybridized for 1 h at 37 °C. Finally, monolayers were washed and stained with DAPI (0.5 μg/ml) for 10 min at RT, washed, and then mounted using ProLong Gold Antifade Reagent.

To confirm the intracellular location of spores, A549 monolayers were immunoprobed for cytokeratin. Monolayers were stained to detect extracellular Cy5+ spores, fixed and washed as described above, then cells were permeabilized and blocked for 1 h at RT in 0.3% Triton-X and 5% FBS in PBS, and then washed three times with 0.05% Tween-20 and 5% FBS in PBS. Primary antibody solution, anti-human pan-cytokeratin rabbit polyclonal antibody (Thermo-Fisher), 5.0 μg/ml, diluted in 0.05% Tween-20 and 5% FBS in PBS (0.5 ml/well) was added and the cells were hybridized for 1 h at 37 °C. Monolayers were washed with 0.05% Tween-20 and 5% FBS in PBS, and incubated with secondary antibody solution Cy3 conjugated anti-rabbit IgG donkey polyclonal antibody (Jackson Immunoresearch, 711-166-152), 5.0 μg/ml diluted in 0.05% Tween-20 and 5% FBS in PBS (0.5 ml/well). Finally, monolayers were washed, stained with DAPI and mounted as described above.

2.7. Quantitation of spore uptake

Stacked confocal images of different, random X-Y coordinate areas of PCLS and A549 monolayers were obtained using a Zeiss LSM-710 multiphoton laser scanning confocal microscope. Detector gain was adjusted for each field imaged so that there were minimal saturated voxels per image and the entire dynamic range of the instrument was used. In short, the Imaris software was then used to record the X-Y-Z location and intensity of fluorescent signals detected in a stack of confocal images. The external or internal location of the Cy5+ spores relative to the HLA-DR (AF488+) or cytokeratin (Cy3+) positive cells was determined and tabulated for further analysis by Imaris. The method and software we employed has been used by other investigators to quantify phagocytosis [28,29]. The specific steps used for data acquisition and interpretation are as follows: From the raw images, surfaces were rendered on all HLA-DR (AF488+) and cytokeratin (Cy3+) positive objects using Imaris software (Bitplane). The parameters for surface rendering established the largest diameter as 8.0 μm and excluded background autofluorescence. Spores (Cy5+) were rendered as point spots (spheres) within a mask channel with a diameter of 1.7 μm with background subtraction. The number and intensities of spore voxels (volume pixels) located within surface object volumes (AEC and APC), was determined by applying the spore mask channel to the cell surface channels, which allowed the mask to serve as a punch-out template and associate the spore information with the cell-surface objects.

For quantitation of spore location (Fig. 3) and uptake by APC and AEC (Figs. 4 and 5) in PCLS, the spore mask was applied to the imaged field volume and to the HLA-DR+ and cytokeratin + surfaces, allowing the location and voxel intensity of the spore mask to be assigned with a specific surface-bounded volume or with the imaged field volume. All spore voxel intensities were summed in the channels of interest and expressed as a percent of total fluorescent intensity of spores within the volume of surface rendered per total spore fluorescent intensity within the volume of field imaged.

Fig. 3.

Spore association within APC and AEC is non-uniform. PCLS cultured in unamended LSM were exposed to Cy5-labeled spores for 8 h, immunoprobed, imaged and analyzed as described. (A and B) Surfaces rendered on cytokeratin+ (A) and on HLA-DR + objects (B) were color-coded based on the total fluorescent intensity of spores (Cy5 + signal) contained within the surface-bounded volumes. In the resultant heat maps of cell-specific spore association, red colored volumes contain relatively many spores and violet colored volumes contain no spores. The image data presented is representative of identically analyzed image sets from the triplicate donors.

Fig. 4.

Lethal toxin does not affect spore uptake by APC or AEC, and spores preferentially associate with AEC in PCLS. (A–D) The percent spores internalized within either cytokeratin + surfaces (AEC) (A and C) or in HLA-DR + surfaces (APC) (B and D) versus incubation time. PCLS in unamended LSM (NT) (A and B), in LSM containing 2 μg/ml protective antigen (PA), or in LSM continuing 2 μg/ml lethal toxin (LT) (PA + LF, 2 μg/ml each) (C and D) were exposed to Cy5+ spores for the indicated times, and then immunoprobed, imaged and analyzed as described. Toxin and spore exposure were concomitant. The % internalized spores = [Number of cell-associated spores/Total number of spores] × 100. The three no treatment (NT) experiments shown individually in panels A and B are summarized in panels C and D. For comparisons within cell types, there is no significant difference between the three treatments at any time point. However, between APC and AEC, internalization by APC is significantly lower than that of AEC for all treatments at all time points (p < 0.05). The data are expressed as the mean ± SEM of PCLS prepared from triplicate donor lungs.

Fig. 5.

Spore internalization by A549 lung epithelial cells is non-uniform and is not affected by lethal toxin. A549 lung epithelial cells were cultured in either unamended DMEM plus antibiotics (NT) or in DMEM plus antibiotics and containing either 2 μg/ml protective antigen (PA) or 2 μg/ml lethal toxin (LT) (PA + LF, 2 μg/ml each) and exposed to 10 MOI Cy5+ spores (red) for the indicated times. After fixation, the non-internalized spores were probed using an anti-Cy5 mouse monoclonal antibody followed by an AF488 anti-mouse IgG secondary antibody (green). The images shown are composite projections of the stacked confocal series used for subsequent analysis with Imaris software. (A and B) Confocal images showing that the majority of spores are extracellular early after exposure, and (D and E) the majority of spores are internalized late after exposure. Fully internalized spores (D–F) appear as red only, while external and partially internalized spores (A–C) appear both red and green or yellow. (A, B, D, and E) Stacked confocal images of A549 monolayers incubated with spores for 2 h (A and B) and 24 h (D and E). (B and E) Increased magnification of the highlighted insets in A and D, respectively. By 24 h incubation (D and E), internalized spores are non-uniformly distributed into a relatively few cells, as compared to the more homogeneous distribution of free spores seen at 2 h (A and B). (C and F) Immunostaining for cytokeratin (white filaments) permits confirmation of Cy5-spore internalization, as shown in the orthogonal (z-stack) projections of a representative confocal microscopy image through the spore focal plane after 24 h incubation. (G) Graph of the number of spores internalized per A549 cell versus incubation time from experiments depicted above. The data are expressed as the mean ± SEM of triplicate experimental replicates. Scale bars = 50 μm.

To count intracellular and extracellular spores in A549 cells (Fig. 6), spores were defined as a mask channel as they were for PCLS. Internalized spores were defined as those point spheres lacking any AF488 (green) total fluorescent intensity above background levels. Conversely, extracellular spores were those point spheres possessing any AF488 total fluorescent intensity above background levels. The data were expressed as the number of internalized spores (voxels) per cell.

Fig. 6.

Cell viability of PCLS and A549 cells is not reduced by lethal toxin. PCLS (A) and A549 (B) cells were treated with 2 μg/ml each of PA or LT with and without spore exposure. PCLS were exposed to 1 × 10⁶ cfu and A549 were exposed to 10 MOI B. anthracis Sterne strain spores. Negative controls (“no treatment”, NT) were treated with growth medium containing toxin and spore diluent (water) and growth medium containing 10 μM staurosporine was used as a positive control. Cytotoxicity (loss of cell viability) was assessed by the release of lactate dehydrogenase into the culture medium, and expressed as the percentage of LDH released (%LDH = [supernatant LDH activity/(supernatant LDH activity + lysate LDH activity)] × 100). LDH release from both PCLS (A) and A549 cells (B) was not significantly changed by LT, PA or spore exposure through 48 h incubation, although, compared to the NT control, cytotoxicity was significantly increased by staurosporine (*, p < 0.05). The data are expressed as the mean ± SEM of triplicate donor lung slices (A) or triplicate experimental replicates (B).

2.8. Lactate dehydrogenase cytotoxicity, assay

PCLS cultured in 24-well plates and A549 cells grown in 96-well plates were treated with 2 μg/ml of PA or 2 μg/ml LT with and without spore exposure, and negative controls (NT) were treated with growth medium containing toxin diluent (water) as described above (Section 2.4). PCLS were exposed to 1 × 10⁶ cfu and A549 were exposed to 10 MOI unlabeled B. anthracis Sterne strain spores. Staurosporine (10 μM) (Cell Signaling Technologies) diluted in growth medium was added as a positive control for cell death. The PCLS and A549 cells were incubated in a tissue culture incubator at 37 °C, 5% CO₂ for 2, 4, 8, 24, or 48 h, after which both culture supernatants and PCLS or A549 cell lysates were collected, prepared and stored at −80 °C. Triplicate samples of PCLS and A549 cells were tested at all conditions. LDH activity present in the medium supernatants and in the PCLS or A549 cell lysates was measured using the LDH-Cytotoxicity Assay Kit II (BioVision) at an absorbance of 500 nm using a Vmax kinetic microplate reader (Molecular Devices). Cytotoxicity, expressed as the percentage of LDH released, was calculated as follows: [supernatant LDH activity/(supernatant LDH activity + lysate LDH activity)] × 100.

2.9. Statistical analysis

Where applicable, the data are expressed as the means ± standard errors of the means (SEM). Statistical significance was determined by one way analysis of variance (ANOVA) with the Tukey–Kramer post hoc correction for multiple comparisons. A p-value of <0.05 was considered significant.

3. Results

3.1. APC and AEC can be distinguished in the human PCLS model of B. anthracis infection

Fresh normal human lung tissue was obtained, inflated with agarose, then cored and sliced to a thickness of 200–300 μm, as described in Materials and Methods. This method of preparing precision cut lung slices (PCLS) preserved alveolar architecture, including cell–cell tight junctions, endothelial cells, and epithelial cells. The boundaries of the alveolar space and integrity of the capillary bed were visualized by stains for the tight junction protein zonula occludens-1 (ZO-1), caveolin, and CD31 (Fig. 1A, B). PCLS were cultured and exposed to Cy5-labeled B. anthracis (Sterne) spores for various times of infection. PCLS were stained with antibody markers to identify AEC and APC (AM and DC, Fig. 1C). The APC in the PCLS were identified by staining with an antibody against the APC surface marker HLA-DR (Fig. 1C, D). AEC were identified by staining with an antibody against the epithelial marker cytokeratin (Fig. 1C, E). Spore location was determined by fluorescent signal in the Cy5 channel (Fig. 1F). For each anti-human antibody of interest, PCLS were stained with a negative control antibody of the same host and isotype and at the same concentration. The negative control staining produced minimal or no detectable fluorescence above background levels when scanned on detector gain settings without over saturation, indicating that the anti-human antibodies were specific for their target antigens. In addition, PCLS stained only with fluorescently labeled secondary antibodies produced minimal or no detectable fluorescence above background levels.

Fig. 1.

The alveolar architecture is preserved in the precision cut lung slice (PCLS) human lung organ culture model, both AEC and APC can be detected in human PCLS, and three-dimensional rendering of confocal image series permits quantitation of spores within APC and AEC and localization of spores. Fixed PCLS were immunoprobed with (A) anti-human ZO-1 to identify tight junctions (red) or (B) anti-human caveolin (green) and anti-human CD31/PECAM-1 (red) to identify endothelial cells. The primary antibodies were detected with fluorescently labeled secondary antibodies. Nuclei were stained with DAPI (blue). Whole lung slices were mounted and a stacked image was produced using laser scanning fluorescent confocal microscopy. (A and B) Both images show a large central alveolar space (AS) surrounded by interstitium (IS). PCLS cultured in unamended LSM were exposed to 1 × 10⁶ cfu/slice with Cy5-labeled B. anthracis, Sterne spores (gold) (C and F). After incubation for various times, PCLS were harvested, fixed, and immunoprobed to identify APC, AEC, and spores by laser scanning fluorescent confocal microscopy. (C and D) APC were identified with anti-HLA-DR antibody (green); (C and E) AEC were identified with anti-pan cytokeratin antibody (red), and (C) nuclear DNA was visualized using DAPI (blue). (C) Composite projection image of stacked confocal series showing APC (green), AEC (red), cell nuclei (blue) and B. anthracis, Sterne spores (gold) used for subsequent analysis with Imaris software and depicts a large central alveolar space (AS) surrounded by interstitium (IS). (G–J) The same confocal microscopy image series shown in C–F after surface rendering using Imaris software. Surface rendering creates enclosed volumes around the fluorescent channels of either APC (green) (G and H), AEC (red) (G and I), and spores (gold) (G and J). By rendering the spores into masked points, their location within the image stack and the number of spores co-localized within the APC and AEC volumes can be quantitated. This analysis is also used to plot the location of all spores within the three-dimensional space of the PCLS. Scale Bars = 100 μm.

In order to quantitate the number and location of the spores and cells in the PCLS, stacked confocal images were rendered using Imaris software. This resulted in conversion of the confocal image stacks into three-dimensional image data containing fluorescent intensity and location, and allowed for combined (Fig. 1G) or separate analysis of APC (Fig. 1H), AEC (Fig. 1I) or B. anthracis spore location (Fig. 1J) in the PCLS image.

3.2. Spores migrate through human lung in a cell-dependent and cell-independent manner

We assessed the location and partitioning of B. anthracis spores in PCLS by evaluating the three-dimensional position of extracellular and intracellular spores over time (Fig. 2). As depicted in Fig. 2A, lung cells and spores were imaged in 3 random, non-overlapping areas per PCLS using confocal fluorescent microscopy, and the images were processed using Imaris software. The total fluorescent intensities of the free spores, as well as spores internalized in APC and AEC, were plotted versus the vertical depth (Z) for the 5 times of exposure (Fig. 2B–F). The results showed that there was progression of the spores through the lung slice over time, the average value of the Z position significantly lengthening between 2 and 48 h incubation. However, the extent of partitioning of spores between AEC, APC, and the extracellular space did not significantly change. For example, at 2 h after exposure, 4.7% were internalized in APC; 13.8% were internalized in AEC; and 81.5% of the spores were free, that is, not associated with either AEC or APC. By 48 h in culture, 2.9% were in APC; 12.7% were in AEC; and 84.4% were not associated with either cell type. Movement of the individual spores, as indicated by the average Z position (Fig. 2G), did not require cell association, although location in PCLS also changed when spores were associated with AEC or APC. Clustering of spores appeared to only be possible within cells as indicated by the higher cell-associated fluorescent intensity at various Z locations (Fig. 2B–F).

Fig. 2.

Spores migrate through PCLS in a cell-dependent and cell-independent manner. (A) Fluorescent confocal imaging of immune-stained and mounted PCLS. Confocal microscopy was used to image three random, non-overlapping locations from within the mounted PCLS. The confocal image stacks measured 360 × 360 × 50 μm (x, y, and z axis, respectively), and Imaris software was used to analyze the surfaces rendered on HLA-DR+ (APC) or cytokeratin+ (AEC) objects as well as on the spores located within these volumes and within the image stack. Unless stated otherwise, all reported data is derived from the average of triplicate imaged areas within PCLS which were prepared from triplicate lung donors. PCLS cultured in unamended LSM were exposed to Cy5-labeled spores for 2, 4, 8, 24 and 48 h and immunoprobed, imaged and analyzed as described in the Experimental Section. (B–F) Total fluorescent intensity of all spores (blue diamonds), spores in APC (green squares) and spores in AEC (red triangles) plotted over vertical depth (Z position) within the imaged PCLS. The data shown in panels B–F is representative and is from a single PCLS from a single donor lung. (G) The mean vertical depth (Z position) of all spores (black bars), spores in APC (green bars) and spores in AEC (red bars) verses incubation time. For all three groups, significant differences in the Z position were observed between the two earliest times, 2 and 4 h, compared to the later times 8, 24 and 48 h (2, 4 h vs. 8, 24, 48 h, p < 0.05). The data are expressed as the mean ± SEM of PCLS prepared from triplicate donor lungs.

3.3. Spore uptake into AEC and APC is non-uniform

To further evaluate spore uptake by AEC and APC, the relative spore association within the cells was examined. Variations in spore association within cellular volumes was converted to a color map. Visual analysis revealed there was more variable spore internalization into AEC than into APC (Fig. 3A, B).

3.4. Spore internalization into AEC and APC and preferential uptake by AEC are not affected by LT and PA

PCLS were exposed to 1 × 10⁶ spores for various times of infection, and the degree of internalization into AEC and APC was measured using Imaris software. Concomitantly, additional slices were exposed to PA (2 μg/ml) or LT (PA 2 μg/ml + LF 2 μg/ml) for the duration of spore exposure at 4, 8 and 24 h. At all times of exposure, spore internalization was greater in AEC than APC (Fig. 4A, B, p < 0.05 for all time points). There was no effect of PA or LT treatment on spore internalization in AEC or APC (Fig. 4C, D). Furthermore, spore internalization was greater in AEC than APC in the presence or absence of PA or LT (Fig. 4C, D, p < 0.05 for 4, 8, and 24 h).

3.5. Spore uptake by A549 lung epithelial cells is nonuniform

Because the acquisition of human tissue and harvest of AEC is an expensive, time-consuming process, we wanted to counter-test our observations regarding spore internalization in a tissue culture model. We chose the human alveolar epithelial cell line A549 as the AEC alternative. Cells were grown to near confluence and exposed to 10 MOI of Cy5+ spores for various times. After exposure, external spores were labeled using mouse anti-Cy5 antibody followed by anti-mouse IgG AF488 + secondary antibody. This allowed for differentiation and quantitation of fully internalized spores (red only), which were not accessible to anti-Cy5 antibody, from external spores (red and green or yellow, Fig. 5A, B, D, E). The results were imaged using confocal microscopy, and rendered and analyzed by Imaris software. Confirmation that this technique distinguished external from internal spores was obtained by staining additional identically treated cells for cytokeratin, and colocalizing spores within the cells or demonstrating extracellular location by confocal microscopy (Fig. 5C, F). All spores colocalizing with cytokeratin were red only, while external spores were red and green or yellow. Visual inspection showed that spore internalization increased with time and was nonuniform (Fig. 5A, D). A few cells internalized many spores per cell, while most of the cells internalized few or no spores, a process that became more pronounced as incubation time increased from 2 to 24 h.

3.6. Spore uptake in A549 lung epithelial cells is not affected by LT or PA

We have previously shown that human AEC express anthrax toxin receptors, and that LT alters barrier function and tight junction protein expression in AEC [18], while human alveolar macrophages do not express anthrax toxin receptors and are unaffected by LT [30]. To determine the effect of LT and PA on spore internalization by A549 cells, these cells were exposed to a multiplicity of infection (MOI) of 10 spores per cell in the presence or absence of PA (2 μg/ml) or LT (PA 2 μg/ml + LF 2 μg/ml) for the duration of spore exposure at 1, 2, 4, 8 and 24 h. Spore internalization was increased significantly by 24 h of exposure (Fig. 5G). However, there was no statistically consistent effects of PA or LT on spore internalization (Fig. 5G).

3.7. Cell viability of PCLS and A549 cells is not reduced by LT or PA

LT exposure has been associated with a loss of cell viability due to apoptosis in some lung cell types but not others [15,18,30]. Therefore, we examined whether toxin exposure of PCLS and A549 cells was associated with cell death, as this may have affected the results.

PCLS (Fig. 6A) and A549 cells (Fig. 6B) were treated with PA (2 μg/ml), LT (PA + LF; 2 μg/ml each), or toxin diluent as a negative control in the presence or absence of spores as in the previous experiments. Staurosporine (10 μM), a strong inducer of apoptosis-mediated cytotoxicity, was used as a positive control. Cytotoxicity was measured as the percentage of LDH release as described. As expected, staurosporine exposure resulted in increased cytotoxicity as compared with other treatments. For PCLS (Fig. 6A), toxicity due to staurosporine was greater than that seen with all other treatments beginning at 24 h of exposure and continuing for the duration of the study. For A549 cells (Fig. 6B), toxicity due to staurosporine was greater than that seen with all other treatments beginning at 8 h of exposure and continuing for the duration of the study. For both PCLS and A459 cells, there was no significant increase in cytotoxicity due to spore treatment with or without PA or LT, or due to PA or LT treatment alone versus untreated cells at any time point.

4. Discussion

Dissemination of B. anthracis from the lung is a necessary step that the pathogen must take during inhalation anthrax infection to produce fatal disease [6]. In human inhalation anthrax, the time between spore exposure and symptoms can vary from days to weeks and is dose-dependent [6]. The period between exposure and symptoms offers a potential window for blocking progression of the disease. Therefore, understanding the initial progress of the pathogen after inhalation into the lung is important. At present, little is known about the survival of B. anthracis in the lung and what is responsible for the variable time between exposure and dissemination [1,2,4,7,8,10,31].

One of the key controversies in the initial progress of B. anthracis spores is whether the pathogen requires interaction with cells to escape the alveolar space and enter the interstitium of the lung. Cell-assisted escape could be via a Trojan horse APC, or directly through AEC [7,10,27]. Escape could also occur via a Jailbreak model where germinating spores produce exotoxins that damage the epithelial surface and allow for cell-independent escape [12]. It is also possible that several model types could be operating simultaneously or even synergistically. Our data and the work of others, suggest that no single model, Trojan horse, epithelial translocation (described by Russell), or Jailbreak, accounts for all phases of transit of the B. anthracis from the alveolar space through the mediastinal nodes to the bloodstream.

For example, a significant, ongoing question is whether B. anthracis toxins play a role in escape. This would require that germination occurs at the site of toxin action, and this was observed in alveolar epithelial cells in air-liquid interface culture [32], and in at least one mouse model [27]. However, that germination occurs is a controversial issue [2,33]. In any case, it was thought that anthrax toxins enable viable spores to escape the alveoli by inhibiting AM [7,15,34]. However, we have shown that human AM do not express functional receptors for the toxins and are not susceptible to immunosuppression, MEK cleavage, or apoptosis induction by LT [30]. In contrast, our work has shown that LT induces MEK cleavage and inhibits barrier function in primary human AEC [18].

We used a human lung organ culture model that contained all cells of the normal human lung, including APC and AEC in their native three-dimensional configuration, to investigate the initial uptake and partitioning of spores within these cells and in the extracellular space. The model contained alveolar units, allowing for examination of the initial phases of escape from the alveolus. We also investigated whether LT altered the distribution and uptake of spores in this model, and confirmed these findings in an epithelial cell line model.

We identified APC and AEC in our model, and tracked B. anthracis spore movement and partitioning between these cells. Migration of the cells through the lung organ culture model occurred and increased with time. Interestingly, the majority of the spores migrated outside of AEC or APC, and, thus, migration from the alveolar space did not require a cellular carrier in this model, and one may not be required for escape of the pathogen at this stage of infection. Furthermore, uptake did not occur evenly across the alveolar surface or within the alveolar epithelial cell line model. Instead, it occurred in clusters, with some areas and cells taking up large numbers of spores, while others were devoid of the pathogen. This may be due to upregulation of epithelial ligands for spore BclA following spore attachment. All these findings appear to be consistent with the predictions inherent in the Jailbreak model of pathogen escape from the alveolar epithelial surface, but there are significant differences. In our system germination and toxin production do not occur, and diffusion occurs in their absence. Also, we do not see evidence of cytotoxicity. Thus, clustering of spores alone in AEC may be sufficient to allow escape.

The data does not exclude the possibility that there is an APC carrier cell or Trojan horse associated with initial alveolar escape or further dissemination of the pathogen from the interstitium to the lymphatic channels of the lung and onward to the mediastinum, just that such a carrier is not necessary for the first step of alveolar escape. Indeed, the finding that spore position in potentially mobile APC changed with time is consistent with the initial stages of the Trojan horse model for dissemination from the lung, as is the observation that higher spore densities are present in these cells. However, our finding that most spores remain unassociated with cells suggests that the majority likely reach lymphatics outside of a cellular carrier. Further experiments with other models may be required to answer whether a carrier cell is important for transport beyond the interstitial space. Our finding that the initial stage of alveolar escape may not require this carrier cell is consistent with in vivo data from a mouse model of inhalation anthrax where spores were directly internalized into lung epithelial cells [27].

In these experiments, the effect of B. anthracis toxins on spore internalization by cells in human lung and in an alveolar epithelial cell model was also tested. Neither PA nor LT significantly affected spore internalization in either model, or spore partitioning between AEC and APC in the lung model. Also, toxins did not increase cellular cytotoxicity in either model. These findings suggest that LT and PA do not have a role in the initial escape from the alveolus. Our model uses externally added toxins because germination is not supported in PCLS and this difference from natural infection could affect our results. Despite these limitations, the results are consistent with the findings of others that demonstrated that early events in the lungs of mice and rabbits were not affected by the presence of toxins, although toxins may play a role in later events including pathogen survival and dissemination to mediastinal lymph nodes [19,34–36].

Our in vitro data suggest that B. anthracis spores migrate through the lung soon after exposure. The primary initial phase of spore movement from the alveolar space across the alveolar epithelial barrier does not require a cellular carrier. The presence or absence of PA or LT does not appear to affect these initial stages of infection, perhaps because spore migration starts before LT alters AEC permeability, even if it is present [18]. Therefore, toxin based approaches to inhibiting this stage of the infection are unlikely to be successful. This does not preclude strategies aimed at enhancing alveolar macrophage killing of spores, based on our previous work and those of others demonstrating that these cells take up spores, are resistant to toxins, and can kill early germinated spores [30,37]. Enhancing killing of the pathogen has shown promise in a mouse model of anthrax [38]. In contrast, toxins may play a role later in the rate of B. anthracis escape from the lung [19], and anti-toxin approaches may be effective in this stage of infection.