The fatal character of the infection caused by inhalation of Bacillus anthracis spores results from a complex pathogenic cycle involving the synthesis of toxins by the bacterium. We have shown using immunofluorescent staining, confocal scanning laser microscopy and image cytometry analysis that the alveolar macrophage was the primary site of B. anthracis germination in a murine inhalation infection model. Bacillus anthracis germinated inside murine macrophage-like RAW264.7 cells and murine alveolar macrophages. Germination occurred in vesicles derived from the phagosomal compartment. We have also demonstrated that the toxin genes and their trans-activator, AtxA, were expressed within the macrophages after germination.
Bacillus anthracis, the causal agent of anthrax, is a Gram-positive, spore-forming, aerobic, rod-shaped bacterium. Infection may result from intradermal inoculation, ingestion or inhalation of spores (Klein et al., 1966; Laforce et al., 1969; Friedlander et al., 1993). Humans may become infected by inhaling spore-contaminated aerosols generated by operations with domestic livestock. The transformation of a dormant spore into a vegetative cell, a key step in the pathogenic cycle of a spore-forming bacterium, enables bacteria to proliferate actively and to synthesize their virulence factors. B. anthracis is an extracellular pathogen. The vegetative forms are encapsulated and toxinogenic. The principal known B. anthracis virulence factors are an anti-phagocytic poly-γ-D-glutamic acid capsule (Green et al., 1985) and two exotoxins, the oedema (PA-EF) and lethal (PA-LF) toxins (Beall et al., 1962; Friedlander, 1986). The target cell binding domain, protective antigen PA (83 kDa), is common to two effector moieties: the oedema factor, EF (89 kDa), and the lethal factor, LF (90 kDa). Oedema factor is a calmodulin-dependent adenylate cyclase (Leppla, 1982) that induces an increase in intracellular concentrations of cyclic AMP in eukaryotic cells (Mock and Ullmann, 1993). It has been suggested that lethal factor is a Zn2+-metalloprotease (Klimpel et al., 1994). Anthrax lethal toxin causes death in laboratory animals (Ezzell et al., 1984) and has been implicated in anthrax pathogenesis (Stanley and Smith, 1961; Beall et al., 1962; Pezard et al., 1991). In vitro, the toxin causes cytolysis of primary macrophages (Mφs) and murine macrophage-like cells, such as J774A1 and RAW264.7 cells (Friedlander, 1986). At subcytolytic concentrations, lethal toxin induces the release of tumour necrosis factor (TNF-α) and interleukin-1β (IL-1β) from RAW264.7 cells (Hanna et al., 1993) and may generate reactive oxygen intermediates (ROIs) (Hanna et al., 1994). The anthrax lethal toxin has also been shown to have T-cell mitogenic activity (Guidi-Rontani et al., 1997). Therefore, the macrophage has a key role in B. anthracis pathogenesis: (i) as the first host cell interacting with B. anthracis spores via phagocytosis; (ii) as a cell that can induce host defence responses against the invading bacteria; and (iii) as the cell mediating cytotoxicity during anthrax infection. The pulmonary form of anthrax causes the rapid progression of the disease and often leads to death (Dalldorf and Beall, 1967; Fritz et al., 1995). Previous studies have suggested that inhaled spores are subjected to phagocytosis by macrophages and are carried by the lymphatic system to local mediastinal lymph nodes (Ross, 1957). The severity of infections with the pulmonary form of anthrax impelled us to investigate the involvement of the alveolar macrophages during the first stage of infection. Confocal scanning laser microscopy and image cytometry combined with a high-sensitivity, fluorescence-based reporter system made it possible to investigate host macrophage–pathogen interactions in single cells. In this study, we have shown that the alveolar macrophage is the primary site of B. anthracis germination in a murine inhalation infection model. We have also identified the cell compartment in which germination occurs and demonstrated the early onset of toxin gene expression.
Specific labelling of B. anthracis spores and bacilli
Spores and germinated spores of the toxinogenic, unencapsulated Sterne strain (7702), obtained after 2 h incubation at 37°C in tissue culture medium (RPMI-1640), were tested for indirect immunofluorescence with rabbit polyclonal sera raised against spores or bacilli and a rhodamine-conjugated secondary antibody (Fig. 1). Samples were also observed by phase-contrast microscopy. The antispore serum labelled dormant spores (Fig. 1A) and germinated spores (Fig. 1C), because it recognized the vestiges of outer envelope structures: the exosporium and spore outer coat (Tomcsik et al., 1959). These two structures persist in germinated anthrax spores and adhere to the bacterium (Hachisuka et al., 1966). In contrast, no labelling of dormant spores was observed with antibacillus serum (Fig. 1E), which only labelled germinated spores (Fig. 1G). These results demonstrated the specificity of the antibacillus serum and allowed us to distinguish unambiguously between dormant and germinated spores.
Germination of B. anthracis spores inside alveolar macrophages
Germination of the Sterne strain in the host was analysed after intranasal infection of Balb/c mice with 2.5 × 107 spores each. Germination is the conversion of a resistant, dormant spore into a sensitive, metabolically active form. Alveolar macrophages were collected from bronchial alveolar fluids (BAL fluids), and the germination of spores associated with BAL fluids was assessed by testing for a loss of heat resistance (30 min at 65°C) (Fig. 2). Bacterial counts without heat treatment corresponded to the total spore population (dormant and germinated spores). Viability after heat treatment was a measure of the population of dormant spores. In the cell-free supernatants of BAL fluids, the number of colony-forming units was the same whether or not there had been heat treatment and throughout the time course of experiment. In contrast, the alveolar macrophage fraction showed significant levels of germination after 3 h and 24 h. By 24 h, a large overall drop in the spore count was observed. This phenomenon may reflect a loss of viability of the germinated forms within alveolar macrophages.
The fate of the spores within macrophages was investigated further by immunofluorescence in a 5 h time course experiment (Fig. 3). Germinated spores were detected 30 min after infection (Fig. 3A and B). They appeared to be located within the perinuclear compartment of the macrophages (Fig. 3C–H).Their abundance demonstrated that the spores had been taken up rapidly by alveolar macrophages. Moreover, as can be seen at 5 h (Fig. 3I), there were no major changes in the morphology of cells harbouring germinated spores at this stage. Actin patches were visible as concentrated areas of intense immunofluorescence. F-actin fibres were visible in the cells, demonstrating cell integrity. Electron microscopy studies also showed ruffling activity of the infected macrophages (data not shown). Free dormant spores were detected in the BAL fluids after 30 min to 24 h. Neither free bacillary forms nor bacterial colonization of the lungs were observed. Thus, the primary site of germination of the inhaled spore was the alveolar macrophage.
Subcellular site of B. anthracis germination within the macrophage
The observation that B. anthracis spores germinated within the perinuclear area of macrophages prompted us to investigate the subcellular site of germination further. Samples were analysed by confocal scanning laser microscopy. As shown in 4Fig. 4A, germinated spores inside macrophages were found co-localized with F-actin-rich phagocytic cups. The uptake of B. anthracis spores therefore depended on F-actin assembly, which is essential for the phagocytic process. The distribution of phagolysosomal compartment and germinated spores on phagocytosis was visualized by double immunofluorescence confocal microscopy (Fig. 4B). The phagolysosomal compartment was detected via CD107a, a lysosome-associated membrane protein 1 (LAMP-1) (Kannan et al., 1996; Fig. 4B, bottom). Three hours after infection, we tested the co-localization by merging the confocal images (Fig. 4B, top) obtained by labelling the germinated spores and the phagolysosome. The rhodamine-labelled germinated spores were outlined in yellow. Thus, the extent of co-localization indicated that the germinated spores and phagolysosomes were located at the same position.
Toxin gene expression by B. anthracis within macrophages
To characterize the early events associated with the germination of B. anthracis, we studied the expression of genes encoding virulence factors such as the toxin trans-activator, AtxA (atxA) (Uchida et al., 1993), and the lethal factor, LF (lef ). B. anthracis strains carrying atxA::lacZ and lef ::lacZ transcriptional fusions were used. Image cytometry and a high-sensitivity, fluorescence-based reporter system provide a powerful new approach for detecting gene expression in single cells. This approach made it possible to detect the expression of a bacterial gene inducible in response to complex environments, such as the intracellular environment of host cells. We detected β-D-galactosidase activity inside the macrophage by taking advantage of the fluorescent lipophilic β-D-galactosidase substrate, 5-dodecanoylaminofluorescein di-β-D-galactopyranoside (C12FDG) (Zhang et al., 1991). These experiments were performed in vitro with the murine macrophage-like cell line RAW264.7, which has no fluorescent background in the presence of C12FDG. Spores were labelled by indirect labelling with antispore serum before infection, allowing detection within macrophages by immunofluorescence microscopy (Fig. 5B, D and F). No fluorescein isothiocyanate (FITC) fluorescence indicating β-D-galactosidase activity was detected in cells infected with the control strain without lacZ (Fig. 5A) despite extensive phagocytosis of spores (Fig. 5B). In contrast, fluorescence was clearly detected after 3 h in the atxA::lacZ (Fig. 5C) and lef ::lacZ recombinant strains (Fig. 5E). This FITC fluorescence always overlapped with the phycoerythrin fluorescence indicating germinated spores (Fig. 5D and F). Some spores were revealed only by phycoerythrin fluorescence, thereby indicating that β-D-galactosidase appeared during germination and did not pre-exist in the resting spore. These results showed a strong association between β-D-galactosidase activity and germinated spores, and reflected the specific expression of atxA and lef. Similar results were obtained with genes encoding the other two toxin components, PA and EF (data not shown).
We have shown by a combination of immunofluorescence staining and confocal scanning laser microscopy that the alveolar macrophages were the primary site of toxinogenic B. anthracis germination during infection by inhalation. Once within the alveoli of the lungs, there was efficient phagocytosis of the spores by the alveolar macrophages. Germinated spores were detected within the perinuclear compartment of macrophages after infection and showed the same distribution as the membrane marker, CD107, of lysosomes. Moreover, the alpha2-macroglobulin receptor, a marker of early/recycling endosome, did not co-localize with germinated spores (data not shown). These observations suggest that the spore-containing phagosome fused with the lysosome and that the spores were able to germinate as the phagosome fused with the lysosome in host cells. The entire population of germinated spores, detected by indirect labelling with antibacillus antibody, gave a high level of fluorescence inside polarized macrophages. This reflected the active change in coat ultrastructure during germination and showed that germination inside alveolar macrophages was efficient. To gain more insight into the molecular mechanisms after germination, we studied the expression of genes encoding virulence factors. We registered a rapid onset of expression for genes encoding virulence factors such as the toxin trans-activator, AtxA, the lethal factor, LF, the protective antigen, PA, and the oedema factor, EF. This demonstrates that germination was closely followed by the expression of the toxin genes. Therefore, B. anthracis is able to respond elaborately to sudden environmental changes inside the cell. These results led us to hypothesize a phagolysosome mechanism that triggers outgrowth via effective and perhaps specific germination factors. Such specific macrophage germination factors may be required for B. anthracis pathogenicity and may be critical for the establishment of pulmonary disease. The germination factor must bind tightly to specific site(s) within the spore to act as a trigger (Feavers et al., 1985). Therefore, these should presumably be specific receptor(s) for B. anthracis germination factors. Moreover, this result underlined the existence of inducer(s), belonging to the phagolysosome, that trigger the transcription of the genes encoding virulence factors such as AtxA, PA, LF and EF. This raises the question of the mechanisms by which these genes are expressed after exposure to stress stimuli (Chen et al., 1996; Rafie-Koplin et al., 1996). Toxins may play an essential role as soon as germination occurs and thereby contribute to bacterial survival.
By confocal microscopy, we showed that the uptake of B. anthracis spores into alveolar macrophages occurs via the recruitment of F-actin, an essential phenomenon for the phagocytic process that leads to membrane activity, resulting in the formation of a phagosome. Consequently, the pathogenic bacterium has to circumvent the oxidative burst resulting from microbicidal NADPH-oxidase activation, detoxifying and eliminating internally generated reactive oxygen species such as the superoxide anion. It is unknown whether superoxide dismutase or catalase (Beaman and Beaman, 1984) are produced by B. anthracis. We demonstrated that the task of phagosome receptors, regulating lysosomal docking and fusion, was accomplished by B. anthracis-containing phagosomes. So, B. anthracis emerged within the phagolysosome, one of the most effective antimicrobial environments. The germinated spore has then to deal with acid pH and lysosomal acid hydrolases to survive in a vegetative form. Various pathogenic strategies are used to ensure survival in the harsh internal environment of the phagolysosome. A major distinction can be made regarding the strategies used by pathogens to survive acid challenge. Pathogens may delay and reduce phagolysosome acidification. This is the approach taken by Salmonella typhimurium (Alpuche Aranda et al., 1993) and Histoplasma capsulatum (Groppe Eissenberg et al., 1993). Bacteria such as the acidophilic bacterium Coxiella burnetii have an absolute requirement for low pH to activate their metabolism (Hackstadt, 1983; Thompson and Williams, 1991). It is possible that changes in pH act as a stimulus for B. anthracis, leading to the transcription and translation of the acid resistance loci necessary for intracellular survival. Studies to identify B. anthracis virulence factors that enable the bacterium to survive in the hostile environment of the phagolysosome are now in progress in our laboratory. Spores germinate in vitro and grow to form fully encapsulated vegetative cells within 30 min (Ezzell and Abshire, 1995). Thus, future research will be directed towards understanding the contribution of the capsule to protection and survival after germination inside the macrophage. This factor may act as a barrier giving the bacterium resistance to bactericidal elements. We will investigate the contribution to pathogenesis of the three virulence factors, toxins and capsule, in our in vivo model.
Bacterial strains and plasmids
The following B. anthracis strains were used: 7702 (Sterne strain; pXO1+) (Pasteur Collection); RBAF143 (Sterne strain derivative containing pXO1 with lef ::lacZ fusion) (Sirard et al., 1994); 7702-XF1 (Sterne strain derivative carrying pXO1 with atxA ::lacZ ). RBAF143 and 7702-XF1 were generously provided by Jean-Claude Sirard (Unité Toxines et Pathogénie Bactériennes).
Preparation of spores
Bacteria were cultured in brain–heart infusion at 37°C with continuous shaking until these were approximately 1 × 10+9 cells ml−1. Bacteria were streaked onto NBY agar and incubated for 7 days at 30°C. Sterile distilled water (10 ml) was added to each bottle and a suspension produced. The spore suspension was incubated at 65°C in a water bath for 30 min to kill any remaining vegetative cells. The spores were collected by centrifugation and resuspended in water to give samples containing at least 108 cfu ml−1. These spore stocks were stored in 50% glycerol at 4°C.
Preparation of antisera against spores and bacilli
New Zealand White rabbits were injected subcutaneously with 109 inactivated spores [incubated in 4% (v/v) formaldehyde overnight at 37°C] or with 1010 dead bacilli [incubated in 4% (v/v) formaldehyde for 3 h at 37°C] in Freund's complete adjuvant. Three weeks later, the rabbits received an additional 109 inactivated spores or 1010 dead bacilli in incomplete adjuvant. Blood was removed from the animals via an ear vein. The IgG fraction was prepared from whole serum by DEAE Affigel Blue chromatography (Bio-Rad). Serum was applied to the column, and IgG was eluted in the unbound protein peak with 0.02 M Tris-HCl (pH 8.0), 0.028 M NaCl, 0.02% (w/v) NaN3. Fractions were pooled and precipitated with ammonium sulphate at 50% (w/v) saturation at 4°C, and the precipitate was suspended in phosphate-buffered saline (PBS). The protein was dialysed against PBS and stored frozen at −70°C.
Cells and cell culture
The murine macrophage-like cell line RAW264.7 was cultured as monolayers in RPMI-1640 supplemented with 50 μg ml−1 penicillin, 50 μg ml−1 streptomycin, 2 mM L-glutamine, 10% (v/v) newborn calf serum in an atmosphere at 90% humidity containing 5% CO2 at 37°C. Three days before use, the cells were detached by gentle scraping with a rubber policeman and seeded into 24-multiwell disposable trays containing the same medium at a density of approximately 50 000 cells per well. Phagocytosis was allowed to proceed for 1 h at 37°C in an atmosphere containing 5% CO2. At this point, the culture medium was replaced with RPMI-1640 containing 2.5 μg ml−1 gentamycin to kill the extracellular germinated spores and cytochalasin D (10−6 M) to prevent phagocytosis.
Intranasal infection of mice
Spore suspension (50 μl; 5 × 10+8 spores ml−1 H2O) was deposited in the nostrils of 7-week-old female Balb/c mice (Mus musculus; IFFA-CREDO) that had been slightly anaesthesized with ether. Infected female Balb/c mice were sacrificed at various times after infection. Mice (two per group) were killed, the tracheas were cannulated and BAL fluids removed by two consecutive washes with 0.5 ml of PBS. At t0, the efficiency of the recovery of spores/bacilli was determined in crude BAL fluids and corresponded to 64% of the inhaled doses. The total number of macrophages in each sample of BAL fluid was counted with a haematocytometer. Samples of 5 × 10+4 cells were cytocentrifuged for 10 min at 250 × g on glass slides with a cytofuge (Hettich). Spores/bacilli in the supernatant and cell fractions of BAL fluids were counted by plating on BHI with or without heat treatment (30 min at 65°C). Viability after heat treatment was taken as a measure of the population of dormant spores. The alveolar macrophage fraction gave the same number of cfu in the presence or in the absence of 2.5% saponin. All experiments were repeated at least twice. One representative experiment is shown.
Immunofluorescence and confocal microscopy
For all labelling, cells were buffered in 30 mM sodium phosphate in PBS (pH 7.4), fixed by incubation in 3.7% (v/v) paraformaldehyde (stock solution as described by Robertson et al., 1963) and 30 mM sucrose in PBS at room temperature for 20 min. Subsequent steps were performed at room temperature. Free aldehyde groups were quenched by incubation for 10 min with 50 mM NH4Cl in PBS. The cells were washed once in PBS supplemented with 1 mg ml−1 bovine serum albumin (PBS/BSA) and permeabilized by incubation for 5 min at 37°C in 0.2% Triton X-100 in PBS/BSA. Immunofluorescence labelling was performed as described by Bucci et al. (1992). Cells were incubated, when indicated, with the primary antibody for 1 h. They were washed twice, and the antibody was detected by incubating the cells for 1 h with labelled secondary antibodies. The cells were washed three times in PBS/BSA and once in PBS, then mounted in Fluoprep (BioMerieux).
Dormant and germinated spores were detected by indirect immunofluorescence using rabbit polyclonal antibodies directed against spores (used at 1:2000 dilution) or bacilli (used at 1:1000 dilution) and rhodamine-conjugated goat anti-rabbit IgG (H + l) (1:50; Kirkegaard and Perry Laboratories). The distribution of filamentous actin (F-actin) was analysed by incubating fixed cells suspended in PBS with Oregon green 488–phalloidin (1:50 dilution; Molecular Probes) for 1 h at 37°C in the dark. Late endosomal compartments were labelled with FITC-conjugated rat anti-mouse CD107a (1:50 dilution; Pharmingen).
Samples were examined in a Leitz Laborlux S fluorescence microscope (Leica) equipped with a PLAN 100/1.25 PHACO3 phase contrast objective. Two excitation cube units were used: one for detecting fluorescein (I3-PLOEMOPAK; Leica) and one for rhodamine (N2.1-PLOEMOPAK; Leica). Images were captured and analysed using a SWild MPS48/52 camera (Leica). Confocal laser cytometry analysis was performed with a confocal scanning fluorescence microscope (Wild Leitz) using an air-cooled argon–krypton laser (Leica). Images were recorded on Ilford 400 film.
Confocal laser microscopy and image cytometry
β-D-Galactosidase activity was detected inside the macrophage using the fluorescent lipophilic β-D-galactosidase substrate, 5-dodecanoylaminofluorescein di-β-D-galactopyranoside (C12FDG; 33 μM; Molecular Probes). Images were acquired with an ACAS 570 laser cytometer (Meridian) equipped with an argon laser (Coherent) run at 488 nm wavelength and 200 mW power. Fluorescence was analysed from a pseudocolour image after thresholds were set.
We are grateful to F. Brossier and J. C. Sirard for valuable discussions and for the generous gift of RBAF143 and 7702-XF1. We wish to thank A. Fouet for comments on this manuscript. We are greatly indebted to R. Hellio (Service de Microscopie Confocale, Institut Pasteur) for efficient assistance with the confocal microscope, and H. Kiefer (Service de Cytométrie Analytique et Préparative, Institut Pasteur) for efficient assistance with the ACAS 570 laser cytometry.