|Home | About | Journals | Submit | Contact Us | Français|
Anthrax toxins significantly contribute to anthrax disease pathogenesis, and mechanisms by which the toxins affect host cellular responses have been identified with purified toxins. However, the contribution of anthrax toxin proteins to dissemination, disease progression, and subsequent immunity after aerosol infection with spores has not been clearly elucidated. To better understand the role of anthrax toxins in pathogenesis in vivo and to investigate the contribution of antibody to toxin proteins in protection, we completed a series of in vivo experiments using a murine aerosol challenge model and a collection of in-frame deletion mutants lacking toxin components. Our data show that after aerosol exposure to Bacillus anthracis spores, anthrax lethal toxin was required for outgrowth of bacilli in the draining lymph nodes and subsequent progression of infection beyond the lymph nodes to establish disseminated disease. After pulmonary exposure to anthrax spores, toxin expression was required for the development of protective immunity to a subsequent lethal challenge. However, immunoglobulin (immunoglobulin G) titers to toxin proteins, prior to secondary challenge, did not correlate with the protection observed upon secondary challenge with wild-type spores. A correlation was observed between survival after secondary challenge and rapid anamnestic responses directed against toxin proteins. Taken together, these studies indicate that anthrax toxins are required for dissemination of bacteria beyond the draining lymphoid tissue, leading to full virulence in the mouse aerosol challenge model, and that primary and anamnestic immune responses to toxin proteins provide protection against subsequent lethal challenge. These results provide support for the utility of the mouse aerosol challenge model for the study of inhalational anthrax.
Bacillus anthracis is a gram-positive, spore-forming bacterium and the etiologic agent of anthrax (38). Three forms of the disease exist, dependent on the route of exposure. Cutaneous anthrax occurs when spores are introduced through a skin wound and gastrointestinal intestinal disease occurs after ingestion of spores. The third manifestation, inhalational anthrax, results after spores are inhaled into the lungs. Inhalational anthrax is the most severe form of the disease and is often associated with rapid disease progression and death (4, 25, 65). Although inhalational anthrax is extremely rare in humans, the use of anthrax as a biological weapon in the fall of 2001 highlighted the need to understand inhalational anthrax disease pathogenesis for the generation of improved therapeutics and vaccines (15).
Protective antigen (PA) is the primary component of the current U.S. licensed vaccine (anthrax vaccine absorbed [AVA]) (6). Functional anthrax toxins require the combination of PA with lethal factor (LF) for lethal toxin (LT) and PA with edema factor (EF) for edema toxin (ET). Genes encoding the toxin proteins are carried on the virulence plasmid pXO1 (44). Once produced by the bacterium, PA binds to the eukaryotic cell surface receptors tumor endothelium marker 8 (TEM8) and capillary morphogenesis protein 2 (CMG2), which then interact with host expressed LDL-receptor-related protein 6 (LRP6) (7, 55). After PA binds to the host cell receptor, it is cleaved by furin and forms heptamers capable of binding three EF and/or LF molecules. After uptake of the toxin complex into the cell by phagocytosis and subsequent acidification of the phagosome, the PA heptamer inserts into the membrane and mediates the translocation of the bound EF and/or LF into the cytoplasm of the host cell (64). The host cell repertoire that anthrax toxins can target is large, since TEM8 and CMG2 are expressed on a variety of cell types, including immune cells (3, 5). LF is a zinc-dependent metalloprotease capable of cleaving host cell mitogen-activated protein kinase kinases in the cell cytosol (19, 62). Cleavage of mitogen-activated protein kinase kinases by LT has been shown to interrupt host cell signal transduction, consequently inhibiting the expression of some cytokines (8). However, spore infection of primary cells with toxin-producing strains does not always result in the inhibition of cytokine production (9, 48). EF is an adenylate cyclase that increases intracellular concentrations of cyclic AMP, which also disrupts host cell responses (31, 40, 59). The majority of work describing the effects of anthrax toxins has been done using purified toxins in vitro and in vivo (1, 12, 39, 40, 45, 64). Limited information is available on the effects of anthrax toxins in the context of an infection (17, 42, 49, 52). Detailed reviews on the mechanisms of toxin entry and the effects of toxins on host cell activity are available (1, 3, 45, 58, 60).
Fully virulent B. anthracis also carries a second virulence plasmid, pXO2. The pXO2 plasmid harbors the genes for biosynthesis of the poly-d-glutamic acid capsule, which encases vegetative bacilli (28, 61). The function of capsule in disease pathogenesis has not been completely elucidated but is hypothesized to prevent bacterial phagocytosis and act as a nonimmunogenic surface (17, 56). Previously published reports have described the susceptibility of mice to B. anthracis strains that express pXO2, even in the absence of pXO1 (18, 27, 30), although the mechanism by which this occurs is unknown. Available data suggest that other animal species do not exhibit the same sensitivity to pXO1− pXO2+ strains as mice do, although published reports on this subject are limited (32, 33).
Rabbits and nonhuman primates, challenged with fully virulent spores, are often used as models to study human anthrax pathogenesis. However, these animals can be costly and difficult to obtain, and such studies require costly biocontainment facilities. Mice provide a practical model for studying disease pathogenesis, evaluating immune responses, and testing new vaccines and therapeutics. We recently described an inhalational murine model of anthrax that recapitulates the disease pathogenesis observed in rabbits and nonhuman primates challenged with fully virulent B. anthracis (42). In an effort to more fully validate this murine model and understand the contribution of anthrax toxins to disease pathogenesis in vivo, we used the murine aerosol challenge (42) and a series of isogenic toxin-deficient mutants (in a Sterne strain genetic background) (36) to examine the role of anthrax toxins in virulence, dissemination, disease progression, and the development of protective immunity. Our results show that functional LT is required for the establishment of disseminated disease and subsequent lethality and that ET contributes to that process but is not required. In addition, our data show that primary exposure to toxin components is required for immune responses that provide protection to a subsequent lethal challenge. The mice did not all exhibit elevated immunoglobulin titers in response to toxin proteins after primary exposure, but rapid anamnestic responses were evident and were likely involved in protection against secondary lethal challenge.
Male and female A/J mice were purchased from the National Cancer Institute (Bethesda, MD). Mice between the ages of 6 and 12 weeks of age were challenged as previously described (42, 49) with minimal modifications. Briefly, mice were exposed to aerosolized spores for 70 min by using a nose-only exposure system (CH Technologies, Westwood, NJ), with fresh air supplied for 8 min before and after spore exposure. The spore inoculum for each challenge, prepared from a toxin-deficient strain (triple knockout [TKO] [Δpag Δlef Δcya], double knockout [DKO] [Δlef Δcya], or Δpag strain), or the parent strain B. anthracis Sterne (strain 7702) contained 12 ml of 5 × 109 spores/ml in distilled water with 0.01% Tween 80. Generation of toxin-deficient strains was previously described, and the loss of toxin genes was confirmed with PCR using published primers (36; data not shown). Generation and purification of spores for all strains were carried out as previously described (22, 49). The 70-min aerosol exposure results in a retained dose in the lungs of 2 × 106 to 4 × 106 spores, which is approximately 10 times the 50% lethal dose (LD50) calculated for strain 7702 (42). For strain 7702 challenges, in which 1 LD50 was desired, groups of A/J mice were exposed to an aerosolized spore inoculum of 5 × 108 spores/ml for 45 min, with fresh air supplied for 8 min before and after the challenge. Immediately after any challenge, four mice were euthanized, and the lungs were removed and homogenized, serially diluted, and plated to determine the average number of spores retained in the lungs for that challenge. Values are reported as CFU. Mice were retained for survival or rechallenge studies or euthanized at various times postchallenge for tissue collection. For rechallenge studies, mice were exposed to aerosols of strain 7702 (inoculum at 5 × 109 spores/ml) for 70 min. All mice for these studies were housed and maintained at the Center for Biologics Evaluation and Research animal facility under the approval of the Institutional Animal Care and Use Committee.
Total serum immunoglobulin G (IgG) antibody titers to PA, LF, or EF were determined by using a quantitative anti-recombinant PA, LF, or EF enzyme-linked immunosorbent assay. Ninety-six-well microtiter plates (Immunolon 2HB; ThermoLabsystems, Franklin, MA) were coated with 100 μl of recombinant PA, LF, or EF (1 μg/ml)/well overnight at 4°C. Plates were then washed (phosphate-buffered saline plus 0.05% Tween) and blocked with 3% bovine serum albumin in phosphate-buffered saline for 1 h at 37°C. Plates were incubated with 100 μl of serially diluted (1:100 to 1:300,000 in blocking buffer) serum samples at 37°C for 1 h. Plates were then incubated for 30 min at room temperature with purified horseradish peroxidase-conjugated goat anti-mouse IgG (KPL, Gaithersburg, MD) diluted 1:1,000 in blocking buffer. Finally, the plates were incubated for 15 to 20 min at room temperature with 100 μl of ABTS [2,2′azinobis(3-ethylbenzthiazolinesulfonic acid); KPL). The reaction was stopped by adding 100 μl of ABTS peroxidase stop solution (KPL). Absorbance values were obtained by using a Molecular Devices (Sunnyvale, CA) VersaMax microplate reader at 405 nm. Samples were assayed in triplicate, and the endpoint antibody titers were expressed as the maximum dilution giving an absorbance of >0.2 (405 nm). The results are presented as the reciprocal of the dilution multiplied by the absorbance value.
Luminescent strains of B. anthracis (Sterne 7702 or mutants thereof) were created as follows. The luxCDABE operon from Photorhabdus luminescens was reconstructed by joining PCR fragments of the individual genes, or portions thereof, and in the process, providing each open reading frame with a strong gram-positive ribosome-binding site. The PCR template was pUTmini-Tn5kmlux (24). Fragments were assembled, without the introduction of unwanted additional restriction sites, through the use of the type IIs restriction enzyme BsaI, essentially as described by Stemmer and Morris (57). A luxBADCE operon fragment, thus created, and flanked by XhoI and KpnI was cloned between the XhoI and KpnI sites of pSS4030, a derivative of the temperature-sensitive vector pBKJ236 (36). In a second step, a XhoI-digested PCR fragment consisting of the entire open reading frame of BA1951 was cloned into the upstream XhoI site in the same orientation as the lux operon. BA1951 is the open reading frame driven by the L-19 (Pntr) promoter, which was identified by a promoter screen in B. anthracis as a strong, constitutive promoter (26). In addition, complementary oligonucleotides encoding the strong, consensus promoter trc-99 were inserted at the 3′ end of the BA1951 open reading frame. The resulting plasmid, pSS4530, was introduced into the Sterne 7702 strain by conjugation, and isolates in which the plasmid had integrated into the chromosome were selected by growth at 37°C with selection for erythromycin resistance, as described previously (36). This insertion is predicted to have no effect on the integrity of the BA1951 gene but places luxBADCE under the control of the L-19 and trc-99 promoters. It was noted that, upon initial plating, colonies that were either more or less luminescent arose at a low frequency. Serial passage of the more luminescent colonies to derive a highly luminescent strain led to the isolation of BA679. In a similar way, highly luminescent derivatives of the toxin deletion strains BA695 (Δcya), BA723 (Δlef), and BA781 (Δpag Δlef Δcya) (36) were derived to yield BA680, BA681, and BA682, respectively. The genetic mechanism by which more luminescent derivatives arose upon passage is unknown. However, these alterations exhibited adequate phenotypic stability to allow their use in animal challenges. Furthermore, bacteria recovered from infected tissues of challenged animals have demonstrated no diminution of either the level of luminescence or the percentage of bacteria expressing the highly luminescent phenotype (data not shown).
Images of mice were acquired by using the IVIS 100 in vivo imaging system (Xenogen). Mice were anesthetized by using 2.5% isofluorane mixed with oxygen using the XGI-8 gas anesthesia system supplied with the IVIS 100 (Xenogen). Images were acquired according to manufacturer's recommendations and analyzed by using Living Image 2.5 software (Xenogen).
For tissue collection, groups of mice were euthanized at various times postchallenge and lungs, cervical lymph nodes (cLNs), mediastinal lymph nodes (mLNs), diffuse and organized nasally associated lymphoid tissues (NALT), spleens, and livers were collected. A 0.3-g piece of liver was collected for determining CFU. Tissues used for CFU determination were homogenized in phosphate-buffered saline. In order to distinguish between spores and spores plus bacilli, a fraction of each tissue homogenate was heat treated (HT) at 65°C for 30 min to kill any vegetative bacilli. HT and untreated (UnT) samples were serially diluted and plated on brain heart infusion agar to determine CFU numbers. Each figure distinguishes between UnT and HT samples for discerning the number of spores plus bacilli (UnT) versus spores alone (HT). The limits of detection were 250 CFU for lungs, spleen, NALT, and liver and 10 CFU for cLNs and mLNs.
GraphPad Prism software (version 4.00; GraphPad Software, San Diego, CA) was used for all statistical analysis. A log-rank test was used to analyze differences in survival after aerosol challenge. One-way analysis of variance with Tukey's multiple comparison post-test was used for analyzing cytokine data.
Rabbits and nonhuman primates are often used as a model of anthrax disease of humans. Studies with these animals have shown that anthrax toxins (LT/ET) significantly contribute to anthrax pathogenesis and that PA-based immunity is critical for protection to lethal anthrax challenge. We have found that an inhalational murine model of anthrax in which complement-deficient A/J mice are challenged with Sterne strain spores recapitulates the disease pathogenesis observed in rabbits and NHP challenged with fully virulent B. anthracis (42). In an effort to further validate this murine aerosol challenge model of anthrax, we examined the requirement for toxin expression for lethality in this model. An isogenic collection of B. anthracis strains, derived from strain 7702 and containing in-frame, unmarked deletions in the genes encoding PA (Δpag), LF (Δlef), EF (Δcya), LF and EF (DKO, Δlef Δcya), or LF, EF, and PA (TKO, Δpag Δlef Δcya) were utilized for these studies. A/J mice were challenged by the inhalational route with spores prepared from mutant strains or the parent strain 7702, and survival was monitored for 10 days (Fig. (Fig.1A).1A). Mice were challenged with 1 × 106 to 5 × 106 spores of each strain, which represents approximately 10 to 20 LD50s for the parental strain 7702 (Fig. (Fig.1B).1B). Figure Figure1A1A shows that deletion of either the gene coding for PA (Δpag) or LF (Δlef) resulted in complete attenuation of disease (100% survival of challenged mice). Although the strain lacking EF (Δcya) was reduced in virulence compared to 7702, it was not completely attenuated. Taken together, these data indicate that ET contributes to pathogenesis during inhalational infection but is not absolutely required for lethality. In contrast, LT is required for lethality during inhalational anthrax. These data also show that nontoxigenic derivatives of the Sterne strain are not lethal to A/J mice. The requirement for toxins in the mouse aerosol challenge model provides an opportunity to evaluate the functional role of toxins and toxin components in vivo after aerosol exposure to spores.
Since mice challenged with the DKO (Δlef Δcya), TKO (Δpag Δlef Δcya), and Δpag strains survived primary aerosol exposure, we were provided with the opportunity to evaluate the relative contribution of the immune response to toxin proteins (PA or LF/EF) to protective immunity generated after inhalational exposure to spores. Groups of A/J mice were immunized by the aerosol route with spores prepared from each of the toxin mutant strains and the 7702 parent strain. For aerosol immunization, mice challenged with the toxin-deficient strains received 10 times the LD50 calculated for 7702; however, mice challenged with strain 7702 were exposed to spores at one LD50 in order to have survivors for rechallenge (Fig. (Fig.2A).2A). All of the mice challenged with a toxin-deficient strain survived the primary immunization challenge, and ca. 50% of mice challenged with one LD50 of strain 7702 survived primary immunization (data not shown). At 30 days after the primary aerosol exposure, all surviving mice, as well as a group of naive A/J mice, were challenged by the aerosol route with 10 LD50s of strain 7702 spores (Fig. (Fig.2B).2B). Survival was monitored for 10 days, as shown in Fig. Fig.2C.2C. Mice initially challenged with mutant strains that still encoded either PA or LF/EF (the Δlef Δcya [DKO] or Δpag strain, respectively) had a significantly higher survival rate than mice immunized with a strain lacking all toxin genes (Δpag Δlef Δcya [TKO]). Mice initially challenged with strain 7702, which produces all toxin proteins, exhibited the greatest protection since 94% of the mice survived secondary lethal challenge (Fig. (Fig.2C).2C). Mice initially challenged with the Δlef Δcya (DKO) or Δpag strain survived the secondary lethal challenge 81 and 68% of the time, respectively. However, only 39% of mice initially challenged with the Δpag Δlef Δcya (TKO) strain survived the lethal rechallenge. This survival rate was not significantly different from that observed for naive mice. Although mice immunized with the Δpag strain had fewer survivors after lethal rechallenge than mice immunized with strain 7702, this difference was not significant (P = 0.0605). Taken together, these data indicate that even in the context of an infection with spores, in which the host is exposed to a wide variety of spore and vegetative cell antigens, the development of protective immunity is dependent upon exposure to toxin components, PA and/or LF/EF.
Antibody responses to PA elicited after vaccination have been shown to contribute significantly to protection upon lethal exposure to B. anthracis (34, 50, 51). To determine whether titers of antibody to PA, LF, or EF were elevated after aerosol immunization, groups of mice were immunized by aerosol challenge. On days 14 and 28 after aerosol immunization with live organism, sera were collected and assayed for total IgG to toxin components (Fig. (Fig.3).3). Only titers to PA and LF were measured due to the small amount of sera collected from each animal. Figure Figure3A3A indicates that anti-PA IgG titers were elevated in mice exposed to strain 7702 and the DKO (Δlef Δcya Δlef Δcya) strain. The increase in anti-PA IgG titers was not significant between these two groups, likely due to the wide range of titers within each challenge group. Mice exposed to strains lacking PA (Δpag Δlef Δcya [TKO] and Δpag strains) did not show significant titers to PA, as a result of the lack of this antigen. Mice exposed to the Δpag strain, which produces EF and LF, did not show elevated titers to LF (Fig. (Fig.3B),3B), although these mice showed some level of protection against lethal aerosol rechallenge (Fig. (Fig.2A2A).
Mice that received an aerosol immunization with the spores of toxin-deficient Δlef Δcya (DKO) or Δpag strains did not exhibit elevated titers of PA- or LF-specific IgG antibody relative to mice that received an aerosol immunization with spores of the toxin-deficient Δpag lef cya (TKO) strain (Fig. (Fig.3).3). However, these mice were protected from a secondary lethal challenge, whereas mice that received an aerosol immunization with the TKO strain were not protected (Fig. (Fig.2C).2C). In order to further understand the contribution of antibodies to toxin proteins in protection, we measured the anamnestic antibody response after a secondary lethal challenge. We collected sera on day 28 after aerosol immunization and on days 3 and 5 after secondary lethal challenge and measured the titers of IgG antibody to PA, LF, and EF. Figure Figure44 shows that titers of IgG to PA, LF, and EF were elevated on day 28 in mice immunized with strain 7702, which is indicative of their survival rate to secondary lethal challenge (Fig. (Fig.2C).2C). Conversely, anti-PA IgG titers of mice immunized with the DKO (Δlef Δcya) strain were not elevated on day 28 after immunization but increased significantly by day 5 after lethal challenge (Fig. (Fig.4A).4A). This correlated with resistance to secondary challenge (Fig. (Fig.2C).2C). As previously noted, the percent survival of Δpag strain-immunized mice, following secondary lethal challenge, was less than for mice immunized with strain 7702, although the difference was not significant (P = 0.0605). This may be explained by the anamnestic antibody response to LF after secondary challenge (Fig. (Fig.4B),4B), since the anti-IgG titers increased slightly by day 5 after secondary challenge. As expected, titers to PA, LF, and EF were not elevated in mice immunized with the TKO (Δpag Δlef Δcya) strain, since these antigens were not present during priming (day 28, Fig. Fig.4).4). After secondary challenge, these mice succumbed to disease (Fig. (Fig.2C),2C), which is likely explained by the lack of a primary response to toxin components. Taken together, these data indicate that although titers of IgG to toxin proteins may not be significantly elevated after primary spore exposure, this does not necessarily correlate to a lack of protection to secondary lethal exposure. Priming of the immune response by the initial exposure can lead to a rapid, anamnestic response that contributes to protection. Initial exposure to toxin proteins is critical, since immunization with live spores that are deficient for toxin genes does not induce protective immunity.
To better understand the dissemination of B. anthracis after aerosol exposure and to investigate the contribution of the anthrax toxins to dissemination and persistence, we utilized an in vivo bioluminescence imaging system to monitor the location of vegetative bacilli in infected animals. This system allows tracking of an infection using a bacterial strain in which bioluminescent genes from P. luminescens are genetically engineered to be constitutively expressed by metabolically active bacterial cells. Sterne strain derivatives were generated, in which the bioluminescent genes under the control of the L-19 and trc-99 promoters were crossed onto the chromosomes of the 7702, Δlef, Δcya, and TKO (Δpag lef cya) strains. Groups of A/J mice were challenged by the aerosol route with 10 LD50s of the 7702-lux strain or each of the luminescent toxin KO strains and were imaged daily for bioluminescence (Fig. (Fig.5).5). Insertion of the lux operon into the chromosome and the expression of the lux operon did not affect the virulence of the strains as evidenced by the survival curves after challenge with these strains (see Fig. S1 in the supplemental material). A minimum of 16 A/J mice were challenged with each luciferase-expressing strain and were imaged daily. Figure Figure55 illustrates representative images from challenges of A/J mice with each of the luminescent toxin KO strains. Representative images of two mice challenged with 7702-lux are provided to illustrate that anthrax disease in mice challenged with B. anthracis spores by the aerosol route does not progress synchronously (Fig. (Fig.5A).5A). Bacilli were first noted in the NALT of all mice by day 1 postchallenge regardless of challenge strain. In strain 7702-challenged mice, progression to the cLNs was observed, followed by rapid dissemination to distal organs. In individual strain 7702-challenged mice, the day that progression to the cLNs was observed varied from day 1 to day 4 (data not shown). Once widespread bacteremia was apparent, mice succumbed to infection. Previously, emphasis has been placed only on the role of the lung and mLNs during inhalational anthrax disease progression, but use of the bioluminescent strain (Sterne-lux) with the in vivo imaging system led us to identify the NALT and cLNs as an additional route of dissemination following aerosol challenge. After challenge with LT-deficient strains (the Δlef-lux and TKO-lux strains), bacilli were observed in the NALT of challenged mice by day 1 postchallenge (Fig. 5C and D). However, luminescence was not observed in any tissue beyond the NALT, and luminescence in the NALT diminished and was lost over time in these mice. This result indicated that in the absence of LT expression, the infection was contained within the draining lymphoid tissue and eventually cleared. The observation that ca. 50% of mice challenged with the ET-deficient strain succumb to infection was reflected in these imaging experiments (Fig. (Fig.1A1A and Fig. Fig.5B).5B). Representative images of two mice challenged with the Δcya-lux strain are provided to illustrate the two outcomes of infection observed (Fig. (Fig.5B).5B). In the mice that ultimately survived infection (e.g., mouse 1, Fig. Fig.5B),5B), the progression observed was similar to that observed in mice challenged with the LT-deficient strains (Fig. 5C and D). In mice that succumbed to infection (e.g., mouse 2, Fig. Fig.5B),5B), the progression observed was similar to that observed in mice challenged with the parental 7702 strain (Fig. (Fig.5A5A).
In order to correlate studies in which luminescence was used to monitor disease progression with studies in which tissue CFU were enumerated, we exposed groups of A/J mice to aerosols of spores prepared from luciferase-expressing toxin-deficient strains (the TKO, Δlef, and Δcya strains) or the parent strain (strain 7702) and imaged the mice each day. Mice were sacrificed at one of three stages of infection and the numbers of spores plus bacilli (UnT) and spores only (HT) were enumerated for the lung, NALT, cLNs, mLNs, liver, and spleen (Fig. (Fig.6).6). Stage I was defined as the point in infection at which the luminescence was first observed in the NALT (Fig. (Fig.6A).6A). For all mice, regardless of challenge strain, stage I was reached by day 1 postinfection. Stage II was defined as the point at which luminescence is first observed in the cLNs (Fig. (Fig.6B).6B). For both strain 7702-challenged mice and mice challenged with the Δcya-lux strain, stage II was attained between day 1 and day 4 postchallenge. No animal challenged with either of the LT-deficient strains (TKO or Δlef strain) was observed to progress to stage II. Stage III was defined as the point in which infection beyond the cLNs was observed (Fig. (Fig.6C).6C). Due to spore deposition from the aerosol challenge, CFU were always present in the lungs of infected animals (Fig. (Fig.6).6). However, vegetative bacilli were only observed in the lungs of animals sacrificed at stage III, as evidenced by the reduction in CFU upon HT (Fig. (Fig.6C).6C). This is consistent with the observation that bacilli appear in the lung vasculature at late stages of disease progression after challenge with wild-type spores (42). Vegetative bacilli were also found in the lungs of mice challenged with the Δcya strain at the late stage of disease (Fig. (Fig.6C6C).
As shown in Fig. Fig.6,6, CFU were present in the NALT, cLNs, and to a lesser extent in the mLNs of all animals at stage I of infection regardless of toxin expression. The small reduction seen in CFU upon heat treatment suggests that mixtures of spores and bacilli were present in the NALT, cLNs, and mLNs. Overall, the presence or lack of toxin genes did not affect the initial dissemination to these draining lymphoid tissues. However, replication and persistence of LT-deficient strains was limited in these draining lymphoid tissues, unlike with the 7702 parent strain (Fig. (Fig.55 and and6).6). LT-deficient strains failed to persist in the NALT or replicate to high numbers in the cLNs or mLNs, as evidenced by the failure of these strains to progress to stage II.
Anthrax toxin expression appears to have a significant role in bacterial dissemination beyond the draining lymphoid tissue, since only mice challenged with the parent strain (strain 7702) or, at a lower frequency, mice challenged with the Δcya strain attained stage III of infection. Mice at this late stage of infection had high numbers of bacilli in the liver and spleen (Fig. (Fig.6C).6C). This observation is supported by experiments in which large numbers of mice challenged with the nonluminescent 7702 and nonluminescent LT-deficient strains were all sacrificed on day 5. In these experiments, liver CFU counts in mice exposed to aerosols of 7702 averaged 1.7 × 104 CFU, but bacilli were not observed in the livers of mice challenged with any toxin-deficient strain (see Fig. S2 in the supplemental material).
Although cases of humans contracting inhalational anthrax are rare, the 2001 mail attacks highlighted the potential threat posed by the use of this agent in a bioterrorist or biowarfare attack. Thoroughly investigating and defining the involvement of various anthrax virulence factors, including toxins, is critical for the development of new vaccines and therapeutics. While utilizing in vitro models to investigate the mechanism by which toxins hinder host cell signaling and responses is important, in vitro experiments utilizing purified components do not always accurately model the complex interactions between host and pathogen. Thus, the availability and use of animal models is essential to further our understanding of anthrax and the development and evaluation of new vaccines and therapeutics.
Rabbits and nonhuman primates often serve as models for studying anthrax pathogenesis following exposure to virulent pXO1+ pXO2+ strains of B. anthracis, and these models are thought to accurately reflect disease in humans (21, 25, 41, 51, 65). Mice can also serve as a model that is useful and affordable for the study of anthrax disease pathogenesis. We utilize a mouse model in which complement-deficient mice are challenged with an unencapsulated, toxigenic strain of B. anthracis (pXO1+ pXO2−). Confidence in any animal model of pathogenesis requires the accumulation of results from many studies that address different aspects of disease. In previous work we described disease progression, innate cytokine responses, and histological changes following aerosol challenge of complement-deficient mice with Sterne strain B. anthracis spores. Our previous studies demonstrated that the course of anthrax disease in complement-deficient mice (A/J) challenged with aerosolized Sterne spores is similar to that observed in rabbits and nonhuman primates challenged with fully virulent B. anthracis. More specifically, the bacterial dissemination and pathological changes observed in mice were similar to those observed in rabbits and nonhuman primates (42). We and others have demonstrated that this model can be used to evaluate toxin-based vaccines (23, 37), and in the present study we demonstrate that functional LT is required for disease progression in this mouse aerosol challenge model.
Previous studies have demonstrated that toxins are involved in virulence and that host exposure to toxin is required for the development of a protective immune response (10, 35, 43, 46, 47). Our results are consistent with these earlier findings. We have extended our study to include an examination of the contribution of the anthrax toxins to dissemination and persistence of B. anthracis following pulmonary infection in vivo and of the role of the toxin components in eliciting protective primary and secondary antibody responses.
Our examination of the contribution of the anthrax toxin proteins to the induction of protective adaptive immunity demonstrated that exposure to all three toxin proteins, in the context of a sublethal aerosol challenge, provided the highest survival rate to a subsequent, lethal aerosol challenge (Fig. (Fig.2C).2C). Primary exposure of mice to strains expressing either PA or LF/EF also affords the animals protection to secondary lethal challenge. Our results suggest that primary exposure to PA provided the greatest protection to secondary lethal challenge, since survival after rechallenge was not significantly different between strain 7702, which expresses all three toxin components, and the DKO (Δcya-lef) strain, which expresses only PA. Taken together, our results indicate that after primary aerosol exposure, responses to PA provide the greatest level of protection following secondary lethal challenge, but exposure to LF and EF contributes to protection as well. It is striking that although the host was exposed to a variety of spore and vegetative cell antigens after aerosol immunization, as evidenced by the detection of bacilli in the NALT and draining lymph nodes of challenged animals, development of protective immunity depended on expression of PA and/or LF/EF.
To further our understanding of the contribution of anthrax toxins to disease pathogenesis, we utilized an in vivo bioluminescence imaging system to monitor disease progression after aerosol spore exposure. Our results show that after challenge with 7702 spores, luminescence is first observed in the NALT, indicating that vegetative bacilli are present in sufficient numbers to be detected in this system (Fig. (Fig.5).5). At this same point in infection, bacilli and spores can be detected in the cLNs and to a lesser extent in the mLNs (Fig. (Fig.6A).6A). Taken together, these results indicate that following aerosol exposure to anthrax spores, spores not only are deposited in the lungs but also are taken up in the NALT. From the NALT, spores and/or bacilli presumably drain to the cLNs, and from the lungs, spores presumably traffic to the mediastinal lymph nodes. In our model, luminescence is next observed in the cLNs, indicating that sufficient proliferation has occurred at this site for the threshold level of detection to be reached. When luminescence was observed in the cLNs of live animals, it was also detectable in the mLNs of the same animal; however, in order to observe the luminescence in the mLNs, it was necessary to dissect the animal and expose the lymph nodes ex vivo for imaging (data not shown). Furthermore, when mice with luminescence in the cLNs were sacrificed and the CFU enumerated, CFU were observed in both the cLNs and the mLNs (Fig. (Fig.6B).6B). These results support the conclusion that dissemination occurs by both the nose and the lung following aerosol exposure to spores.
Although bacilli are observed by in vivo imaging in the NALT of all challenged animals as early as day 1 postinfection, the day at which luminescence is first observed in the cLNs varied from as early as day 2 to as late as day 4. We speculate that the delay in disease progression observed at this step is due to the engagement of the host innate immune response and the effort of the host to contain the infection. The appearance of luminescence in the cLN correlates with the failure of the host innate immune response to control the infection since 100% of the animals that exhibited luminescence in the cLNs progressed to broadly disseminated disease and death.
We investigated the contribution of each toxin to dissemination, replication, and bacterial persistence during inhalational anthrax and discovered that neither LT nor ET was required for the initial dissemination to the NALT or to the draining lymph nodes (Fig. (Fig.55 and and6).6). Work in the field has shown that both macrophages and dendritic cells can traffic spores to the draining lymph nodes (8, 11, 29, 54). Thus, it is not surprising that toxins would not be required for this step, since toxins would not be expressed prior to germination of spores and the outgrowth of bacilli. Our finding that toxins do not contribute to the initial dissemination to the draining lymph nodes is consistent with previous observations indicating that germination does not occur in the lungs and suggests that the location where toxins first have an opportunity to be expressed and have an impact on disease or antigen presentation is in the draining lymphoid tissues (14, 30, 42, 49). Although LT did not affect the initial appearance of B. anthracis spores and bacilli in the NALT and draining lymph nodes, it was required for persistence and replication to high numbers in both sites (Fig. (Fig.55 and and6).6). In animals challenged with strains lacking LT, luminescence was never observed in the cLNs, and the luminescence observed initially in the NALT diminished and was ultimately eliminated. This indicates that in the absence of LT, the bacterium cannot evade clearance by the host. This observation is consistent with the growing body of data generated utilizing in vitro approaches that suggests the anthrax toxins impair host innate immune responses, allowing the establishment of a productive infection (2, 3, 13, 16, 20, 53, 59, 60, 63). Our results provide the first in vivo demonstration that toxin expression is required early and that the bacterial infection does not progress beyond the draining lymphoid tissue in the absence of LT. LT may be the primary contributor to persistence and subsequent mortality, since challenge with the ET-deficient (Δcya) strain was still lethal in 60% of the challenged animals, but challenge with an LT-deficient (Δlef) strain was not lethal (Fig. (Fig.1A).1A). The results using in vivo imaging suggest that ET also contributes to evasion of the innate immune response in that all animals that survived challenge with the ET-deficient strain cleared the infection before high numbers of bacilli could be observed in the draining lymph nodes (Fig. (Fig.55 and and6).6). Taken together, our results demonstrate that LT is required for the progression of disease beyond the draining lymphoid tissues and that ET contributes to that process but is not essential. Ongoing and future experiments are directed toward the identification of LT-inhibited, innate immune functions essential for the control of B. anthracis infection.
Our results demonstrating the requirement for LT and the contribution of ET to disease progression provide further evidence of the relevance and usefulness of this mouse aerosol challenge model. We have demonstrated the role of the anthrax toxins in evasion of the host response and the establishment of disseminated disease in the context of an aerosol infection in mice. Further studies using the mouse model should allow for the characterization of the interaction between the bacillus, the secreted toxins, and the host innate immune response.
We thank Gopa Raychaudhuri for critical reading of the manuscript.
This project has been funded in whole or in part with Federal funds from the National Institute of Allergy and Infectious Diseases, National Institutes of Health, Department of Health and Human Services, under IAAY1-A1-6153-01.
Editor: V. J. DiRita
Published ahead of print on 27 October 2008.
†Supplemental material for this article may be found at http://iai.asm.org/.