|Home | About | Journals | Submit | Contact Us | Français|
Bacillus anthracis is a spore-forming, gram-positive organism that is the causative agent of the disease anthrax. Recognition of Bacillus anthracis by the host innate immune system likely plays a key protective role following infection. In the present study, we examined the role of TLR2, TLR4, and MyD88 in the response to B. anthracis. Heat-killed Bacillus anthracis stimulated TLR2, but not TLR4, signaling in HEK293 cells and stimulated tumor necrosis factor alpha (TNF-α) production in C3H/HeN, C3H/HeJ, and C57BL/6J bone marrow-derived macrophages. The ability of heat-killed B. anthracis to induce a TNF-α response was preserved in TLR2−/− but not in MyD88−/− macrophages. In vivo studies revealed that TLR2−/− mice and TLR4-deficient mice were resistant to challenge with aerosolized Sterne strain spores but MyD88−/− mice were as susceptible as A/J mice. We conclude that, although recognition of B. anthracis occurs via TLR2, additional MyD88-dependent pathways contribute to the host innate immune response to anthrax infection.
Bacillus anthracis is a spore-forming, gram-positive organism that is the etiologic agent of the various forms of anthrax infection. According to the current model of B. anthracis pathogenesis, following inhalation, anthrax spores enter the lungs and are phagocytosed by host alveolar macrophages and are subsequently carried to regional lymph nodes (11, 23). Spores germinate inside the host macrophages and become vegetative bacilli that are then released from the macrophages (9-11). Vegetative bacilli can produce a number of virulence factors including pXO1-encoded exotoxins, lethal toxin and edema toxin, and a pXO2-encoded poly-d-glutamic acid capsule (6, 23). Bacilli multiply in the lymphatic system, enter the bloodstream, and multiply to high levels (107 to 108 organisms per milliliter of blood) (11).
The innate immune response is the first line of defense against invading pathogens. One of the most important components of the innate response is comprised of the phagocytic cells such as the macrophages, dendritic cells, and neutrophils. These cells engulf and kill invading pathogens. Early in infection, the innate immune system likely plays a critical role in the host recognition and response to B. anthracis. The ability of phagocytic cells to recognize microbial products is mediated by the Toll-like receptors (TLRs). TLRs are a family of receptors that recognize pathogen-associated molecular patterns shared by large groups of microorganisms, thereby assisting in microbial recognition by the host immune system (3-5). Lipopolysaccharide (LPS) of gram-negative bacteria is a TLR4 ligand, whereas peptidoglycan (PGN) and lipoproteins of gram-positive bacteria are TLR2 ligands as reviewed in references 2 to 4 and 26. A central feature of this system is that binding of TLRs by pathogen-associated molecular patterns leads to activation of signaling pathways that are critical for induction of the early host defense against invading pathogens (2-4, 26). The TLRs can activate a common signaling pathway that leads to activation of NF-κB transcription factors and mitogen-activated protein kinase (MAPK) signaling molecules p38, extracellular signal-regulated kinase, and Jun N-terminal protein kinase (4). This pathway involves an adaptor protein, myeloid differentiation factor 88 (MyD88), which contains a Toll/interleukin-1 receptor (TIR) domain and a death domain (38). MyD88 is essential for responses against a broad range of microorganisms or their components which are recognized by TLR2, TLR4, TLR5, TLR7, or TLR9. TLR4-mediated signaling occurs via MyD88-dependent and MyD88-independent pathways in response to LPS whereas TLR2, TLR5, TLR7, and TLR9 signaling is MyD88 dependent (5, 14, 16, 20, 27). In addition, recognition of microorganisms can occur via TLR-independent pathways such as the intracellular recognition of gram-positive and gram-negative bacterial cell wall PGN by the nucleotide-binding oligomerization domain (NOD) family of proteins, including NOD1 and NOD2 (18, 19).
We hypothesized that the B. anthracis Sterne strain, a gram-positive organism, is recognized by TLR2. In the present study, we found that heat-killed B. anthracis (HKBa) activates TLR2 signaling as measured by NF-κB reporter gene activity in HEK293 cells expressing TLR2. HKBa stimulates TNF-α production in C3H/HeN, C3H/HeJ (which are functionally TLR4 deficient), and Tlr2tm1Kir (TLR2−/−) bone marrow-derived macrophages (BMDMs). Thus, HKBa activates TLR2 but a non-TLR2 signaling pathway also appears to be involved. In vivo studies demonstrated that TLR2−/− mice and TLR4-deficient mice are resistant to aerosol exposure by B. anthracis spores but MyD88−/− mice are susceptible. These results provide evidence for the presence of redundant pathways of TLR-mediated recognition of B. anthracis.
C3H/HeN mice were obtained from Charles River Laboratories, Inc. (Wilmington, MA). A/J, C3H/HeJ, C3H/HeOuJ, C57BL/6J, and Tlr2tm1Kir (TLR2−/−) mice were purchased from the Jackson Laboratory (Bar Harbor, ME). The TLR2−/− mice were backcrossed to C57BL/6J for nine generations at the Jackson Laboratory; C56BL/6J mice were obtained from the Jackson Laboratory and used as the control strain for TLR2−/− mice and for MyD88−/− mice. MyD88−/− mice were backcrossed to a C57BL/6J background for over eight generations and were kindly provided by Shizuo Akira (Research Institute for Microbial Diseases, Osaka University, Osaka, Japan) (1, 15) and were bred and maintained at the Center for Biologics Evaluation and Research.
Human embryonic kidney 293 (HEK293) cells and L-929 cells were obtained from the American Type Culture Collection (Manassas, VA). Murine BMDMs were prepared by sacrificing mice using CO2 per approved protocol 3383 of the University of Virginia Animal Care and Utilization Committee (ACUC) and protocol 2004-13 of the FDA Institutional ACUC. Mouse femurs were removed to obtain hematopoietic precursor cells. Cells were grown over 6 days in Dulbecco modified Eagle medium (DMEM; Mediatech, Inc.) plus 10% fetal calf serum (FCS; Gibco) plus 10% L-cell conditioned medium plus penicillin-streptomycin-l-glutamine (Gibco). L-cell conditioned medium was obtained by growing L-929 cells to confluence in DMEM with 10% FCS as per published protocols (8, 32, 34, 36). Cell culture incubations with HEK293 cells were carried out at 37°C, 5% CO2, and BMDMs were incubated at 37°C, 10% CO2.
Staphylococcus aureus PGN was obtained from Sigma Chemical Co. (St. Louis, MO). N-Palmitoyl-S-[2,3-bis(palmitoyloxy)-(2RS)-propyl]-cysteinyl-[S]-seryl-[S]-lysyl-[S]-lysyl-[S]-lysyl-[S]-lysine · 3HCl (PAM3CSK4) was purchased from EMC Microcollections (Tubingen, Germany). LPS from Escherichia coli was purchased from Sigma and purified using the protocol of Hirshfeld et al. (16a).
A single colony of freshly grown B. anthracis (Sterne strain 34F2 or 7702; pXO1+ pXO2−) on sheep's blood agar was inoculated into brain heart infusion (BHI; Fisher Scientific Company, Houston, TX) broth and grown at 37°C to an optical density at 600 nm (OD600) of 0.4 as described by Hsu et al. (17). CFU were determined by inoculating dilutions of the bacterial cultures onto sheep's blood agar plates, incubating the plates overnight at 37°C, and counting the colonies. To generate HKBa, the bacterial suspension was heated to 60°C for 60 min, centrifuged at 3,000 rpm for 5 min, washed three times with phosphate-buffered saline, and resuspended in medium (DMEM plus 10% heat-inactivated FCS). After the bacteria were heat killed, aliquots of cultures were plated on sheep's blood agar and incubated overnight at 37°C to confirm that there was no growth of HKBa.
Spores were prepared from B. anthracis strain 7702 (pXO1+ pXO2−) using the method described by Finlay et al. (12, 13). Briefly, nutrient agar (NA) plates were inoculated with overnight cultures and incubated overnight at 30°C. Colonies from the NA plates were used to inoculate nutrient broth (NB) cultures, which were grown overnight at 30°C with shaking. An aliquot of the culture was spread onto NA plates containing 5 μg/ml MnSO4. These plates were incubated overnight at 30°C followed by incubation at room temperature (~22°C) for 48 h in the dark. Colonies scraped from the surface of the agar were suspended in distilled water (dH2O) and heat treated at 65°C for 30 min to kill any viable vegetative cells. The spores were washed once with dH2O and stored at 4°C. These preparations were stained with modified Ziehl-Neelsen stain and examined microscopically to evaluate the presence and purity of spores. Purification of spores was performed using 58% (vol/vol) Renografin (Renocal-76 diluted in dH2O; Bracco Diagnostics, Princeton, NJ), prior to use. Spores were layered on the 58% Renografin and were spun at 4,000 × g for 30 min in a swinging bucket rotor. The spore pellet was washed twice with dH2O (6,000 × g for 30 min), and the spores were resuspended in dH2O. The final concentration of the stock spore solutions was adjusted to ~5 × 109 spores/ml.
Transfections using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) in serum-free medium were carried out with (i) pcDNA3.1 vector without insert (Invitrogen) as a control cotransfected with the NK-κBLuc reporter plasmid (Clontech, Palo Alto, CA), (ii) pcDNA3.1 vector containing human TLR2 construct corresponding to the entire coding region plus NF-κBLuc reporter construct (7, 35), or (iii) pcDNA3.1-TLR4 plus pEF6-MD-2 plus pRc/RSV-CD14 plus NF-κBLuc reporter construct (7, 35). All transfections also contained phRLTK (Promega) as an internal control for transfection efficiency. HEK293 cells were transfected in triplicate wells in a 24-well plate and then exposed 48 h after transfection to vehicle control, PGN (10 μg/ml) (21, 22), PAM3CSK4 (100 ng/ml) (24, 25), purified LPS (100 ng/ml or 1 μg/ml) (7), or HKBa (OD600 of 0.4) (17). Transfected cells were challenged for 6 h, and luciferase activity was measured using the dual luciferase assay (Promega). The data are representative of 5 to 17 separate experiments.
Limulus amebocyte lysate (LAL) assays (Charles River Laboratories, Inc., Wilmington, MA) were performed on freshly prepared, sterile BHI broth, DMEM, PGN, HKBa, and supernatants of cells incubated with HKBa.
Following challenge of murine macrophages in vitro, cell-free supernatants were harvested and frozen for later tumor necrosis factor alpha (TNF-α) cytokine assay by enzyme-linked immunosorbent assay (eBioscience, San Diego, CA). Cell viability was measured using a colorimetric assay per manufacturer's directions (Cell Counting Kit-8 [CCK-8]; Dojindo Molecular Technologies, Gaithersburg, MD).
MyD88−/−, TLR2−/−, C57BL/6J, and A/J mice (18 to 21 g) were exposed to aerosolized spores prepared from B. anthracis strain 7702 for 90 min per approved protocol 2004-13, FDA Institutional ACUC. A/J mice were used as a control strain that is susceptible to B. anthracis Sterne strain challenge (31, 37). The spore aerosol was generated using a six-jet Collison nebulizer equipped with a precious fluids jar containing an inoculum of 15 ml (5 × 109 spores per ml in dH2O; BGI Incorporated, Waltham, MA) as described by Pickering et al. (29, 30). The mice were exposed using a nose-only exposure system (CH Technologies, Westwood, NJ). Prior to exposure, mice were supplied with fresh air for 10 min to allow respiratory rates to normalize. One hour following exposure, four mice were euthanized via lethal injection of Avertin, and their lungs were homogenized and plated to determine the number of organisms that were retained (average retained dose). Ten mice per challenge group were followed for survival.
Statistical analyses were performed using Student's paired t test with 95% confidence intervals using GraphPad Prism 4.0 software (GraphPad Software, Inc., San Diego, CA).
Since HEK293 cells do not express endogenous TLR2 or TLR4, we used these cells to selectively express these TLRs. HEK293 cells were cotransfected with plasmids encoding human TLR2 or TLR4 (plus the cofactors MD-2 and CD14 for the latter) and an NF-κB reporter gene plasmid (7). As a control, HEK293 cells were transfected with plasmid pcDNA3.1 without insert and the NF-κB reporter gene plasmid. After 24 h, cells were challenged with buffer, the TLR2 ligands PGN or PAM3CSK4, the TLR4 ligand LPS, or HKBa for 6 h. As expected, PGN and PAM3CSK4 stimulated NF-κB reporter gene activity only in TLR2-expressing cells (Fig. (Fig.1A),1A), and LPS stimulated the NF-κB reporter gene activity only in the TLR4-expressing cells (Fig. (Fig.1B).1B). Exposure to HKBa resulted in a three- to fivefold activation of NF-κB reporter gene activity in cells expressing TLR2 (Fig. (Fig.1A).1A). We also observed a 1.5- to 3-fold HKBa stimulation of NF-κB reporter gene activity in HEK293 cells expressing TLR4 (Fig. (Fig.1B).1B). Given the likelihood of endotoxin in BHI growth medium being responsible for the TLR4 signaling, we tested the ability of polymyxin B to reduce or eliminate the TLR4 response. In the presence of polymyxin B, HKBa failed to stimulate NF-κB reporter gene activity over control values in TLR4-expressing HEK293 cells (Fig. (Fig.1B).1B). In contrast, addition of polymyxin B (30 μg/ml) to TLR2-expressing HEK293 cell samples had no effect on HKBa-stimulated NF-κB reporter gene activity (Fig. (Fig.1A).1A). Renilla luciferase activity was equivalent between buffer-treated and polymyxin B-treated cells, indicating that the polymyxin B was not toxic to cells. We also used a CCK-8 assay to confirm that polymyxin B (30 μg/ml) had no cytotoxic effect on HEK293 cells (data not shown). These data suggest that the observed 1.5- to 3-fold TLR4 stimulation by HKBa was attributable to the presence of endotoxin in the medium. LAL assays of the samples revealed that there was 1,100 to 7,600 pg/ml of endotoxin in the supernatants of the cells treated with HKBa in the absence of polymyxin B. In addition, LAL assay of sterile BHI broth yielded approximately 500 pg/ml endotoxin. We attempted to grow B. anthracis in cell culture medium (RPMI 1640 or DMEM) that was endotoxin free, but growth of the organism was minimal even after 72 h in these medium alternatives. Hsu et al. (17) recently demonstrated that exposure of murine BMDMs to the Sterne strain results in apoptosis that depends on signaling from LPS-responsive TLR4. Park et al. (28) further demonstrated that the cholesterol-dependent cytolysin, anthrolysin O, secreted from B. anthracis is a TLR4 ligand. Although B. anthracis products can activate TLR4 (28), our results do not support a role for TLR4 in recognition of HKBa. Our results demonstrate that HKBa is recognized by TLR2.
We tested the ability of HKBa to stimulate TNF-α production in C57BL/6J BMDMs. HKBa was prepared from cultures grown to an OD600 of 0.1, 0.2, 0.4, 0.6, and 0.8, and the level of TNF-α stimulated from the C57BL/6J BMDMs increased with increasing OD600 of the added HKBa up to the OD600 of 0.4 (data not shown). Use of HKBa prepared from cultures with OD600 above 0.4 resulted in the same level of TNF-α production (data not shown). Therefore, for all subsequent experiments, HKBa was prepared from cultures grown to an OD600 of 0.4. The stimulation by LPS and PAM3CSK4 when added at concentrations of 10 ng/ml, 100 ng/ml, and 1,000 ng/ml was measured, and the optimal concentration for stimulating TNF-α production in C57BL/6J BMDMs was found to be 100 ng/ml for both agonists (data not shown). This result is in agreement with previously published studies (7, 17, 21, 22, 24, 25). Therefore, for subsequent experiments, we used 100 ng/ml for LPS and PAM3CSK4.
We found that HKBa stimulation of C3H/HeN BMDMs, which express endogenous TLR2, TLR4, and other TLRs, resulted in TNF-α levels in the 4,000- to 5,500-pg/ml range (Fig. (Fig.2A);2A); these results were comparable to stimulation with the control agonists, LPS (100 ng/ml) and PAM3CSK4 (100 ng/ml) (Fig. (Fig.2A).2A). In comparison, only HKBa and the TLR2 agonist PAM3CSK4 stimulated TNF-α production in C3H/HeJ BMDMs (Fig. (Fig.2A),2A), which express a nonfunctional TLR4 due to a naturally occurring mutation in the TLR4 receptor (30a). Thus, stimulation of TNF-α production by HKBa is not dependent on TLR4 signaling, which is consistent with the HEK293 transfection data presented in Fig. Fig.11.
We next tested the ability of HKBa to stimulate TNF-α production in TLR2−/− and C57BL/6J BMDMs. We observed that there was TNF-α production in the HKBa-stimulated TLR2−/− and C57BL/6J BMDMs (Fig. (Fig.2B).2B). Addition of polymyxin B (10 or 30 μg/ml) to the samples had no effect on HKBa-stimulated TNF-α production in TLR2−/− BMDMs (data not shown). In contrast, we observed a complete absence of TNF-α response with the control TLR2 ligand PAM3CSK4 in the TLR2−/− BMDMs (Fig. (Fig.2B).2B). These data suggest that the presence of endotoxin in the culture broth preparation of the HKBa alone cannot explain the ability of HKBa to stimulate TNF-α production in TLR2−/− cells. The data support involvement of non-TLR2 signaling pathways in the response to HKBa. Taken together, these in vitro results suggest that innate immune recognition of B. anthracis vegetative cells occurs through the TLR2 signaling pathway but that TLR2-independent pathways also contribute to that recognition.
We next tested the effect of HKBa on TNF-α production in MyD88−/− BMDMs. We found that TNF-α production was absent in the PAM3CSK4-stimulated and HKBa-stimulated MyD88−/− BMDMs (Fig. (Fig.2C).2C). For LPS, there was low but measurable TNF-α production in the MyD88−/− BMDMs (Fig. (Fig.2C).2C). This latter effect was expected, since LPS can activate TLR4 signaling of MyD88-dependent and MyD88-independent pathways (20). These data indicate that HKBa is recognized by MyD88-dependent signaling pathways.
Given our in vitro findings that HKBa is recognized by TLR2, we tested the effect of B. anthracis spore aerosol challenge of TLR2−/− mice. In animal challenges using aerosolized spores prepared from the B. anthracis Sterne strain (7702), C57BL/6J mice are resistant to challenge (100% survival) and A/J mice are susceptible to challenge (0 to 20% survival). As expected, we found that spore challenge of A/J mice resulted in >95% mortality over a cumulative 10-day observation period (Fig. (Fig.3A).3A). In contrast, the C57BL/6J mice were resistant to aerosolized spore challenge with 100% survival (Fig. (Fig.3A).3A). Aerosolized spore challenge of the TLR2−/− mice resulted in 100% survival, indicating that the TLR2−/− mice were as resistant to infection as the C57BL/6J control strain (Fig. (Fig.3A).3A). This result suggested that, although B. anthracis is recognized by TLR2, recognition of B. anthracis following infection with spores may also be occurring through additional TLRs. In this series of experiments, we also tested effects of aerosolized spore challenge of C3H/HeJ mice, which are TLR4 deficient (30a). The C3H/HeJ mice were as resistant to spore challenge as were the C3H/HeOuJ control mice, with 100% survival observed in both strains (Fig. (Fig.3B3B).
Since a large number of TLRs require the cytoplasmic adaptor molecule MyD88 in order to induce activation of NF-κB and MAPKs, we can address our hypothesis that redundant TLR signaling pathways contribute to recognition of B. anthracis following aerosol infection with spores by challenging MyD88−/− mice. Therefore, we performed aerosolized spore challenge of MyD88−/−, A/J, and C57BL/6J mice and found that the MyD88−/− mice were as sensitive to aerosol challenge with B. anthracis spores as A/J mice. Challenge of both A/J mice and MyD88−/− mice resulted in 80% mortality over the 10-day postchallenge period (Fig. (Fig.3C).3C). As above, we found that the C57BL/6J control strain was resistant to challenge, with 100% survival (Fig. (Fig.3C).3C). This result demonstrates that MyD88-dependent signaling contributes to the recognition of B. anthracis following aerosol exposure to spores with protection from the infection.
Our data underscore the complex nature of innate immune recognition of B. anthracis infection by the host. In vitro, we found that HKBa is recognized by HEK293 cells that express TLR2 but not TLR4. However, we also found that HKBa is capable of stimulating TNF-α production in TLR2−/− BMDMs, with results that were similar to the positive-control ligand, LPS. This result indicates that at least one of the multiple B. anthracis-derived components in the HKBa preparations is recognized by TLR2. One or more additional components are recognized by a TLR2-independent pathway(s). Taken together, these data provide the first evidence that HKBa and, by a logical extension, vegetative B. anthracis are recognized by TLR2-dependent and non-TLR2-dependent pathways.
Having demonstrated that recognition of B. anthracis occurs through TLR2 in vitro, we next examined the outcome of aerosolized spore challenges in mice that were deficient in TLR2. In these studies, we found no difference in survival between TLR2−/− mice and the control C57BL/6J strain (100% survival). Two independent reports indicate that recognition of certain B. anthracis components can occur through TLR4 (17, 28). We, however, found no difference in survival between C3H/HeJ mice (which express a nonfunctional TLR4 receptor) and the C3H/HeOuJ control mice (100% survival) following aerosol challenge with B. anthracis Sterne strain spores. The observation, discussed above, that TLR2 and TLR4 can recognize B. anthracis or its secreted products indicated to us that there was the potential for redundancy in the innate immune recognition of B. anthracis infection. For that reason, we anticipated that the lack of either TLR2 or TLR4 alone would not result in increased susceptibility to spore challenge since the host is exposed to the full spectrum of B. anthracis-expressed components following aerosol challenge with spores. It remained a possibility, however, that the lack of susceptibility in the TLR2- and TLR4-deficient mice was due to the innate resistance of the parental mouse strain to challenge with B. anthracis. To address this issue and to demonstrate the importance of TLR-mediated recognition of infection, we performed spore challenges of MyD88−/− mice. MyD88 is an adaptor molecule required by multiple TLRs to initiate the signaling cascade that leads to the activation of NF-κB and MAPKs. In contrast to the outcomes observed when mice lacking a single TLR were challenged, spore aerosol challenge of MyD88−/− mice resulted in mortality comparable to that seen upon the concomitant challenge of susceptible A/J mice. These results support the hypothesis that host recognition and signaling occur through several TLRs following aerosolized spore infection. Further, our data support the idea that innate immune recognition of B. anthracis infection occurs through redundant pathways that are MyD88 dependent. Ongoing and future projects in the authors' laboratories are focused on the identification of the TLRs that are capable of recognizing specific B. anthracis components in vitro and the in vivo demonstration of the contribution of that recognition to protection.
This work was supported by National Institutes of Health/National Institute of Allergy and Infectious Diseases grants K08 AI50804 Mentored Clinical Scientist Award (M.A.H.), U54 AI057168 Mid-Atlantic Regional Center of Excellence in Biodefense (M.A.H. and T.J.M.), and R01 AI34358 (M.F.S.) and, in part, by R21 AI56113-01 (M.A.H.).
We thank Erik L. Hewlett and members of his laboratory, Gail W. Sullivan, William A. Petri at the University of Virginia, and members of the Mid-Atlantic Regional Center of Excellence in Biodefense for helpful and stimulating discussions. We thank Virginia Carl (University of Virginia) for advice and assistance with the transfection experiments.
Editor: J. T. Barbieri