Francisella tularensis is an obligate, intracellular bacterium that causes acute, lethal disease following inhalation. As an intracellular pathogen F. tularensis must invade cells, replicate, and disseminate while evading host immune responses. The mechanisms by which virulent type A strains of Francisella tularensis accomplish this evasion are not understood. Francisella tularensis has been shown to target multiple cell types in the lung following aerosol infection, including dendritic cells (DC) and macrophages. We demonstrate here that one mechanism used by a virulent type A strain of F. tularensis (Schu4) to evade early detection is by the induction of overwhelming immunosuppression at the site of infection, the lung. Following infection and replication in multiple pulmonary cell types, Schu4 failed to induce the production of proinflammatory cytokines or increase the expression of MHCII or CD86 on the surface of resident DC within the first few days of disease. However, Schu4 did induce early and transient production of TGF-β, a potent immunosuppressive cytokine. The absence of DC activation following infection could not be attributed to the apoptosis of pulmonary cells, because there were minimal differences in either annexin or cleaved caspase-3 staining in infected mice compared with that in uninfected controls. Rather, we demonstrate that Schu4 actively suppressed in vivo responses to secondary stimuli (LPS), e.g., failure to recruit granulocytes/monocytes and stimulate resident DC. Thus, unlike attenuated strains of F. tularensis, Schu4 induced broad immunosuppression within the first few days after aerosol infection. This difference may explain the increased virulence of type A strains compared with their more attenuated counterparts.
Francisella tularensis is a Gram-negative, obligate intracellular bacterium that was first identified as the agent of rabbit fever in the early 1920s (1). There are four subspecies of F. tularensis, mediasiatica, novicida, holartica (type B), and tularensis (type A). Type A is the most lethal for humans, with as few as 10 inhaled organisms capable of causing acute, lethal pneumonic disease (2, 3). F. tularensis is readily available in the environment and has been previously modified to be used as a biological weapon. These features have necessitated its classification as a category A priority pathogen. Despite identification of this bacterium in the early 1920s and subsequent studies on its pathogenesis over the last 80 years, little is understood about many aspects of immunity to F. tularensis, especially the early events following aerosol infection.
The majority of the information describing F. tularensis pathogenesis has been compiled using an attenuated type B strain known as live vaccine strain (LVS).3 LVS retains dose- and route-related virulence in mice, although it is relatively attenuated in humans. Using the mouse model, we and others have shown that LVS targets multiple cells, including dendritic cells (DC) and macrophages for infection and replication (4, 5, 6). In the lung, LVS initially targets pulmonary DC for replication. The infection of these cells results in “phenotypic” activation or increased expression of MHC class II (MHCII) and CD86 on the surface of DC (6). However, despite this phenotypic activation, DC from LVS-infected mice do not secrete proinflammatory cytokines, such as TNF-α and IL-12, typically associated with the “functional” maturation of DC. Furthermore, it has been shown that LVS actively suppresses the ability of both DC and macrophages to secrete these cytokines in response to secondary stimuli (6, 7). Given the central role of TNF-α and IL-12 in the resolution of Francisella infections, this partial suppression of primary APC is thought to contribute to the virulence of LVS in mice (8, 9, 10). Indeed, in agreement with its marked attenuation in humans, LVS stimulates productive proinflammatory responses following vaccination or in vitro infection of human cells (11, 12).
Other attenuated pathogens have a similar profile of aberrant activation of DC. For example, intranasal infection of mice with Bordetella pertussis results in a similar profile of phenotypic, but not functional, maturation of DC (13). Virus-like particles that are morphologically similar to intact virus but lack the components required for replication and assembly readily activate numerous cells, including NK cells and DC (14, 15). In contrast to these more attenuated pathogens and preparations, it has been shown that more virulent organisms interfere with multiple aspects of host/DC responses. Both the Ebola virus and Marburg viruses infect and replicate in DC but suppress the ability of these cells to stimulate T cells and secrete important anti-viral cytokines such as IFN-α (16, 17). It is believed that the targeting and suppression of DC during the first few hours of infection significantly hobbles the host response, allowing the pathogen to replicate, disseminate, and ultimately cause death (18).
To determine whether virulent type A F. tularensis induces a similar aberrant activation of host cells as that observed with LVS or executes a broader spectrum of immunosuppression, we examined early events in the pulmonary tissues following low-dose aerosol infection with a representative type A strain of F. tularensis, Schu4 (see Table I⇓). In the current study, we demonstrate that Schu4 replicates within multiple compartments of the pulmonary system, including the airways and draining lymph nodes. However, despite this rapid replication and dissemination (in direct contrast to LVS), Schu4 did not induce the phenotypic activation of pulmonary DC, macrophages, or DC in the lymph node draining the lung. Schu4 infection also did not induce the secretion of several proinflammatory cytokines, including TNF-α, IL-12, IL-10, and IL-1β, within 24–48 h of infection. The absence of proinflammatory cytokines could not be attributed to the loss of cells due to apoptosis. Rather, an overwhelming in vivo suppression of multiple host responses was observed in Schu4-infected mice. This active suppression included both the failure of resident DC to undergo activation and the inability of the host to mobilize effector cells in response to secondary stimuli (LPS). TGF-β, a potent immunosuppressive cytokine, was detected at both the site of the infection and in peripheral tissues within 24 h of infection, before the detectable dissemination of the bacterium. The administration of anti-TGF-β Abs in Schu4-infected mice increased the production of TNF-α and reduced bacterial loads in the lungs. These results suggest active suppression of host responses by Schu4, including the elicitation of TGF-β by the bacterium, facilitates its replication in the lung. Together, our data may in part explain the enhanced virulence associated with the type A strains of F. tularensis and could aid in the identification of novel candidates for vaccine or therapeutic countermeasures.
Materials and Methods
Specific pathogen-free, 6- to 8-wk old female C57BL/6 mice were purchased from The Jackson Laboratory. All mice were housed in sterile microisolater cages in the laboratory animal resources facility or in the Biohazard Research Building BSL-3 facility at Colorado State University (Ft. Collins, CO) and provided sterile water and food ad libitum. All research involving animals was conducted in accordance with animal care and use guidelines and animal protocols were approved by the Animal Care and Use Committee at Colorado State University.
F. tularensis Schu4 and Yersinia pestis strain MG05 were provided by Drs. J. Peterson and M. Schriefer (Centers for Disease Control, Fort Collins, Colorado). Y. pestis MG05 is a recent virulent isolate obtained from a fatal human case of plague (M. Schriefer, unpublished observations). Schu4 was cultured in modified Mueller-Hinton broth at 37°C with constant shaking overnight, aliquoted into 1 ml-samples, frozen at −80°C,and thawed just before use as previously described (5). Frozen stocks were titered by enumerating viable bacteria from serial dilutions plated on modified Mueller-Hinton agar as previously described (5). Y. pestis MG05 was cultured in brain-heart infusion (BHI) broth at 37°C with constant shaking overnight, aliquoted into 1-ml samples, frozen at −80°C, and thawed just before use. Frozen stocks were titered by enumerating viable bacteria from serial dilutions plated on BHI agar plates. For both Y. pestis and F. tularensis the number of viable bacteria in frozen stock vials varied <5% over a 10-mo period.
Infection of mice
Mice were infected with Schu4 via a whole body aerosol. Low-dose aerosol infections were performed as previously described with minor modifications (19). Briefly, mice were placed in a Glas-Col inhalation apparatus. The nebulizer compartment was filled with an 8-ml suspension of F. tularensis Schu4 at 106/ml in sterile PBS. The vacuum was set to 35 cubic feet per hour and the compressor was set to 19 cubic feet per hour. These settings allowed reproducible delivery of ∼50 CFU to the lungs over a 30-min exposure period. Initially, inoculum doses were confirmed by homogenizing the lungs of freshly infected mice and culturing the entire homogenate as described below. This inoculum routinely results in 100% mortality and a mean time to death of 5 days following infection (data not shown).
In some experiments mice were infected intranasally with Y. pestis MG05 as previously described (43). MG05 was cultured in BHI broth supplemented with 2 mM CaCl2 broth at 26°C with constant shaking overnight. A 100-μl aliquot was then transferred to fresh BHI broth and grown overnight at 37°C with constant shaking. Bacteria were diluted to achieve an inoculum of ∼1 × 105/ml in sterile PBS immediately before infection of the mice. The inoculum was titered by enumerating viable bacteria from serial dilutions plated on BHI agar plates. Mice were anesthetized i.p. with a 200-μl injection of 2.5% Avertin (Sigma-Aldrich). Forty microliters (∼4 × 104 or 100 LD50) of freshly grown and diluted Y. pestis MG05 was administered to the nares of each mouse. This dose routinely results in 100% lethality within 2.5 days of infection. In other experiments mice were anesthetized as described above and 5 μg per 40 μl ultrapure LPS (Invivogen) was administered to the indicated mice. Mice receiving PBS served as vehicle (untreated) controls.
Delivery of anti-TGF-β
Collection of airway, lung lymph node, and spleen cells
Airway cells were obtained by bronchoalveolar lavage (BAL) as previously described (20). Briefly, mice were euthanized by cervical dislocation and an 18-gauge catheter was immediately inserted into the trachea. Approximately 1.5 ml of ice-cold PBS was injected and then aspirated from the lungs. This was repeated three times. BAL cells from each mouse were pooled and then centrifuged at 1200 rpm for 5 min at 4°C. BAL cells were then resuspended in either complete RPMI or FACS buffer (PBS with 2% FBS and 0.05% sodium azide) before analysis. Lung cells were isolated as previously described (21). Briefly, lungs were excised, minced, and incubated in PBS supplemented with 5 mg/ml collagenase type 1A, 125 μg/ml DNase, and 250 μg/ml soybean trypsin inhibitor (all from Sigma-Aldrich) for 30 min at 37°C with 5%CO2. Tissues were triturated using a 5-ml syringe and an 18-gauge needle. Cells were then pelleted by centrifugation at 1200 rpm for 5 min and RBC were lysed with NHCl4. Cells were then washed twice in PBS and resuspended in FACS buffer before analysis. The draining lymph nodes (mediastinal lymph node (MLN)) from the lungs were aseptically removed, passed through a 70-μm cell strainer, washed twice, and resuspended in FACS buffer before analysis. Total live cells from the airways, lungs, and MLN were enumerated using trypan blue and a hemacytometer. For assessment of the ex vivo production of cytokines, the lungs and spleens were excised, weighed, and placed (one per well) in 24-well tissue culture plates. Tissues were then uniformly minced and 1 ml of complete RPMI was added to each well. Tissues were incubated at 37°C and 5% CO2 and supernatants were collected 24 h later and stored at −80°C before analysis for cytokines by ELISA (see below).
Enumeration of bacteria in organs
Lungs, spleens, and MLN were collected and homogenized in sterile PBS using a stomacher (Tekmar). Bacterial colony counts in each organ were determined by plating serial 10-fold dilutions of organ homogenate on modified Mueller-Hinton agar and incubating the plates at 37°C for 48 h.
Detection of Schu4 by immunohistochemistry (IHC) in lung tissue
Detection of intracellular Schu4 in vivo following infection was accomplished by performing IHC on lungs at various time points after infection. Briefly, lungs were collected and preserved in 10% buffered formalin. Following routine processing and embedding in paraffin, tissues were cut into 4-μm sections and mounted on Platinum Line StarFrost slides (Mercedes Medical). Tissue sections where deparaffinized with Histoclear (National Diagnostics) and rehydrated. The sections were then subjected to Ag retrieval by incubation for 15 min at 90°C in DakoCytomation Target Retrieval solution (pH 9.0). The sections were then treated with 3% H2O2 for 5 min and blocked with 0.1 M glycine in PBS for 15 min followed by blocking with 1% normal horse and goat serum in TBST with 1% BSA for 30 min. The slides were then incubated with a 1/4000 dilution of rabbit anti-Schu4 IgG (a gift from Dr. J. Peterson, Division of Vector-Borne Infectious Diseases/Center for Disease Control) in blocking buffer. The slides were then washed thoroughly followed by visualization of Ab binding using the Vectastain system (Vector Laboratories) and diaminobenzidine as the chromogen. Sections were counterstained with hematoxylin, mounted, and examined on an Olympus BX41 light microscope equipped with a QColor3 camera and associated QCapture Pro software.
Flow cytometry and analysis of BAL, lung, and MLN cells
DC were defined as CD11c+DEC-205+ cells that expressed variable levels of other surface determinants, including CD11b, Gr-1, B220, MHCII, CD86, and F4/80, as well as high side and forward scatter properties as previously reported (22, 23) Alveolar macrophages were identified as CD11b+F480+CD11c−DEC-205− cells with high forward and side scatter properties. Monocytes were identified as CD11b+GR-1+/−MHCII+/− cells that had forward and side scatter properties more characteristic of lymphocytes. Granulocytes were identified as GR-1+MHCII−CD11b+/− cells that displayed an elongated side scatter profile typical of granulocytic cells in blood. Isotype control Abs were included when analyses and panels were first being performed to assure the specificity of staining but were not routinely included with each experiment. Data were analyzed using Summit software (DakoCytomation). The percentage of each cell population was determined and then total cell numbers for each population were calculated from the total number of viable cells collected. Experimental and control groups consisted of 3–5 animals each. SEM and statistical significance between treatment groups were determined by ANOVA followed by Tukey-Kramer’s comparison of means.
Assessment of apoptosis
Pathological analysis of lungs
Lungs were harvested, fixed in 10% buffered formalin, and processed as described above. Tissue sections were stained with H&E and examined on an Olympus BX41 light microscope equipped with a QColor3 camera and associated QCapture Pro software.
Determination of secreted cytokines
Statistical differences between two groups were determined using an unpaired t test with significance set at p < 0.05. For comparisons between three or more groups, analysis was done by one-way ANOVA followed by Tukey’s multiple comparisons test with significance determined at p < 0.05.
Replication of Schu4 in pulmonary compartments
Virulent and attenuated strains of F. tularensis efficiently replicate in the lungs following pulmonary infection (2). However, the airways and draining lymph nodes also represent two important target organs for bacterial replication and dissemination following aerosol infections. The airways also contain a large number of APC that are essential for initiating immune responses against invading pathogens following migration to the draining lymph node (23, 24). Colonization and replication within these tissues may represent an important aspect of F. tularensis pathogenesis. Therefore, we assessed the ability of Schu4 to colonize and replicate in the airways and the draining lymph node, the MLN, following a low-dose aerosol infection. Schu4 readily colonized the airways and replicated over time in this pulmonary compartment (Fig. 1⇓). Interestingly, despite being the site of first contact, bacterial numbers in the airways were significantly lower than those detected in the lungs (p < 0.01) (Fig. 1⇓). In addition to colonization of the airways, we also saw rapid dissemination to several peripheral organs. Whereas previous reports have suggested that other type A strains fail to or only poorly disseminate from the lungs until 72 h after infection, we repeatedly observed 102–103 CFU of Schu4 in the spleen and, importantly, the MLN as early as 48 h after infection (Fig. 1⇓) (2). Collectively, these data suggest that Schu4 colonizes and replicates in a variety of pulmonary tissues, including the draining lymph nodes. Additionally, dissemination can occur at time points earlier than previously anticipated.
Schu4 does not induce increased expression of MHCII or CD86 on DC
DC represent a key cellular population that stimulates both innate and adaptive immunity. DC participation in and activation of the immune response is marked by their ability to rapidly secrete proinflammatory cytokines and increase the expression of surface receptors such as CD86 and MHCII that are central to T cell stimulation long before other host cells can respond. Many pathogens target DC as a mechanism to evade and disable host immunity. Given that Schu4 replicated in multiple pulmonary compartments that have abundant numbers of DC ready to detect pathogens, we next examined what effect Schu4 had on the phenotypic activation of lung and lymph node DC. We first determined which cell types Schu4 was targeting following low-dose aerosol infection. Lungs were analyzed by IHC for Schu4 24–72 h after infection. As described for LVS, Schu4 targets multiple cell types after aerosol infection including epithelial cells and APC (macrophages and DC) present in both the airways and the interstium of the lung (Fig. 2⇓, B–D) (6, 25, 26). Interestingly, by 72 h after infection Schu4 was detected as both an intracellular bacterium and cell-free in necrotic lesions (Fig. 2⇓D). Because Schu4 appeared to target many cell types, we next assessed the changes in MHCII and CD86 expression on DC and macrophages in the airways and lungs at various time points after low-dose aerosol infection. Neither DC nor macrophages changed the expression of MHCII or CD86 during the first 24 h after infection (data not shown). However, the lack of increased MHCII and CD86 on the surface of APC 24 h after infection may have been due to the relatively low numbers of bacteria in these two compartments (Fig. 1⇑). Therefore, we also analyzed cells isolated 48 h following Schu4 infection when higher numbers of bacteria (105–106) were present in the airways and lungs. Once again, we observed no significant differences in MHCII or CD86 expression on either DC or macrophage populations in Schu4-infected mice compared with uninfected controls 48 h after infection (Fig. 2⇓E).
The primary function of DC is to carry Ag to the draining lymph node for presentation to effector T cells. During this migration DC can also continue to increase the expression of MHCII and CD86. Thus, it was possible that activated DC, with an increased surface expression of MHCII and CD86, may have migrated to the MLN following aerosol infection with Schu4. To determine whether Schu4 infection resulted in the activation and subsequent migration of DC to the MLN, we analyzed cells isolated from the MLN 24–72 h after infection. Schu4 did not cause increased numbers of DC in the MLN at any time during the course of infection (Fig. 3⇓A, and data not shown). Schu4 infection also failed to significantly increase expression of MHCII and CD86 on DC in the MLN throughout infection and, in fact, reduced the expression of MHCII on DC in the lymph node 48 h after infection (p < 0.05) (Fig. 3⇓, B and C, and data not shown).
It is possible that the absence of DC activation may have been due to the inability of Schu4 to infect these cells. However, as shown above, Schu4 is promiscuous in the type of cells it targets following aerosol infection. Furthermore, we have recently shown that Schu4 readily infects and replicates in bone marrow-derived DC and freshly isolated airway APC (C.M. Bosio and S.L. Warner, submitted for publication). Thus, failure to infect these targets seems unlikely. Together, this data suggests that, unlike attenuated LVS, Schu4 does not induce the phenotypic activation of resident pulmonary DC, hampering the ability of these cells to activate lymphocytes following infection (6).
Schu4 does not induce the secretion of proinflammatory cytokines from pulmonary tissues
Pneumonic tularemia infections are often marked by the absence of early inflammatory responses in the lung despite growing numbers of bacteria in the lung and airways and the presence of clinical symptoms such as fever (27). Proinflammatory cytokines are critical for the development of pulmonary inflammation and their absence can represent a mechanism by which a pathogen can evade host immune responses. For example, TNF-α and IL-12p40 have been shown to be critical for the resolution and survival of Francisella infections (8, 9, 10). We have previously shown that the poor production of proinflammatory cytokines observed in murine LVS infection correlates with the dampened pathological changes at early time points after infection. Thus, we next examined the potential for Schu4 to elicit proinflammatory cytokines following aerosol infection. Airway cells and lungs were removed at the indicated time points after infection and cultured overnight as described in Materials and Methods. The presence of cytokines in the culture supernatant was then examined by ELISA. In contrast to pulmonary infections with a virulent strain of Y. pestis (MG05), Schu4 did not elicit the production of TNF-α, IL-12p40, or IL-10 within the first 48 h following aerosol infection (Fig. 4⇓, and data not shown). Additionally, although IL-1β has been observed both in vivo and in vitro following LVS infections, we were unable to detect this cytokine at any time point (24–72 h) after either in vivo or in vitro infection with Schu4 (12, 28) (data not shown). Therefore, despite the exponential replication of Schu4, this organism does not elicit the typical proinflammatory responses associated with acute pulmonary bacterial infections within the first 48 h of infection.
Schu4 does not induce the apoptosis of pulmonary cells
One possible explanation for the absence of proinflammatory cytokines and resident APC activation following Schu4 infection is that the bacterium induced the apoptosis of infected cells, rendering them unable to produce and/or secrete cytokines and respond to infection. To address this possibility, we next assessed the ability of Schu4 to induce apoptosis in pulmonary cells following in vivo infection. We first examined cellular apoptosis by staining airway and lung cells with annexin. Airway and lung cells were harvested as described above and analyzed for their ability to bind annexin via flow cytometry. Surprisingly, very few annexin-positive DC were observed throughout the course of infection (Fig. 5⇓A). The lack of annexin staining was not confined to DC, because even the majority of infiltrating granulocytes remained annexin negative up to 72 h after aerosol infection (Fig. 5⇓A).
Because the ability to bind annexin is also associated with cells undergoing necrosis, we also assessed apoptosis via in situ staining for cleaved caspase-3. Cleaved caspase-3 is a central molecule in apoptosis and its expression is required for the activation of several components of the apoptotic cascade (29). Lungs from uninfected and Schu4-infected mice were assessed for the expression of cleaved caspase-3 by the immunohistochemical staining of tissue sections sampled 24, 48, and 72 h after infection. In agreement with the data describing the ability of cells to bind annexin, very few cells were positive for cleaved caspase-3 throughout 72 h of infection (Fig. 5⇑, F and G). The scarce number of cells positive for cleaved caspase-3 were confined to areas of necrosis and appeared to be granulocytes or monocytes (Fig. 5⇑, F and G). Cleaved caspase-3-specific staining was confirmed following the preincubation of tissues with cleaved caspase-3-blocking peptide (data not shown). Pulmonary endothelial cells are also thought to represent a major site of Francisella infection and replication. Furthermore, it has been suggested that it is the infection and destruction of these cells that contributes to the dissemination of F. tularensis. However, an examination of the pulmonary endothelium throughout Schu4 infection revealed that these cells remain relatively intact, undisturbed, and free of apoptosis (Fig. 5⇑D). Therefore, the absence of pulmonary proinflammatory cytokines and resident DC activation does not appear to be due to the loss of cells caused by apoptosis.
Schu4 actively suppresses inflammatory responses in the lung
We and others have previously shown that LVS actively suppresses the ability of infected cells to respond to secondary stimuli as a mechanism of virulence in vitro (6, 7). However, to date the ability of Francisella, especially type A strains, to suppress proinflammatory responses in vivo has not been shown. Given the absence of inflammatory responses in the lungs of Schu4-infected mice during the first 48 h of infection, we next determined whether Schu4 merely failed to elicit cellular activation and proinflammatory cytokines or could actively inhibit the ability of the host to respond to other inflammatory stimuli. Ultrapure LPS was administered intranasally to mice 24 h after receiving a low-dose aerosol of Schu4. Twenty-four hours after the inoculation with LPS, airway cells were harvested and assessed by flow cytometry for the accumulation of inflammatory cells and the activation of resident APC. Lungs were also removed, processed, and examined for histopathological changes. As expected, LPS stimulated significant increases of CD86 on the surface of airway DC in uninfected controls compared with untreated mice (p < 0.01) (Fig. 6⇓, A and B). LPS also induced a marked increase in monocytes in the airways compared with untreated controls (p < 0.01) (Fig. 6⇓C). In contrast, LPS failed to induce significant changes in CD86 expression on airway DC of Schu4-infected mice compared with uninfected LPS-treated mice (Fig. 6⇓, A and B). Schu4-infected mice were also refractory to LPS-induced infiltration of monocytes into the airways compared with uninfected controls (Fig. 6⇓C).
It was possible that freshly immigrating cells had adhered to the airways of Schu4-infected mice treated with LPS. These cells would not be readily available in the BAL fluid for analysis. Thus, we also examined the lungs for changes in cellular populations and overall organ structure. LPS induced marked multifocal margination of monocytes and neutrophils in pulmonary vessels in uninfected mice (Fig. 6⇑, D and F). Transmigration and marked accumulation of these inflammatory cells was also noted in the perivascular stroma with a gradient of infiltration in the alveolar septae and spaces (Fig. 6⇑F). In marked contrast, Schu4-infected mice had mild and highly restricted infiltration of monocytes and granulocytes in the lungs following inoculation with LPS (Fig. 6⇑G). Together, these data suggest that Schu4 not only interferes with the ability of resident DC to respond to further stimulation via the increased expression of cell surface markers but also inhibits the ability of the host to properly mobilize critical cellular responses crucial for the resolution of infection at the site of inoculation.
Schu4 induces local and systemic production of TGF-β
There are many mechanisms by which inflammatory responses are suppressed and controlled in the host. TGF-β has previously been associated with broad immunosuppression and homeostatic processes in a variety of host tissues including the lung (30, 31, 32). Because Schu4 failed to elicit the production of proinflammatory cytokines and the activation of pulmonary DC and actively suppressed responsiveness to LPS, we hypothesized that Schu4 may induce the production of immunosuppressive cytokines, including TGF-β. To determine whether Schu4 could stimulate the production of immunosuppressive cytokines airways, lungs and spleens were removed and cultured overnight and the resulting supernatants were analyzed for TGF-β and IL-10 as described in Materials and Methods. IL-10 was not detected in any tissue at any time point following Schu4 infection (data not shown). However, 24 h after infection TGF-β levels were significantly greater in the lungs of Schu4-infected mice compared with uninfected controls (p < 0.05) (Fig. 7⇓A). Surprisingly, despite the absence of detectable bacteria at this time point TGF-β was also elevated in the spleens of Schu4-infected mice compared with the uninfected controls (Fig. 1⇑ and 7⇓A). The concentrations of TGF-β in all tissues tested returned to levels detected in uninfected controls by 48 h after infection and did not significantly change throughout the remaining course of infection (data not shown).
We next examined what effect the neutralization of TGF-β would have on the elicitation of proinflammatory cytokines and the control of Schu4 replication in the lungs 48 h after low-dose aerosol challenge. Mice were treated with anti-pan-TGF-β Abs, isotype control Abs, or PBS (untreated) immediately before challenge and 24 h after infection. Forty-eight hours after infection (24 h after the last administration of antibodies) mice were analyzed for the production of TNF-α by airway cells and loads of Schu4 in the lung and spleen. Administration of anti-TGF-β significantly increased the production of TNF-α by airway cells compared with untreated controls (p < 0.05) (Fig. 7⇑B). Treatment with anti-TGF-β also increased the production of TNF-α compared with mice treated with isotype control; however, this difference was not significant (Fig. 7⇑B).
Mice treated with anti-TGF-β Abs were also assessed for bacterial loads in the lungs following infection. Forty-eight hours after infection, mice treated with anti-TGF-β Abs had significantly fewer bacteria in the lungs compared with untreated mice (p < 0.05) (Fig. 7⇑C). Administration of anti-TGF-β also reduced bacterial loads in the lungs compared with isotype-treated mice, but these differences were not significant (Fig. 7⇑C).
These data suggest that the production of TGF-β may contribute to the active suppression observed in the lungs of Schu4-infected mice. Furthermore, although Schu4 appeared to be localized in the pulmonary tissues during the first 24 h after aerosolization, pulmonary infection resulted in the systemic production of TGF-β. Therefore, Schu4 may not only dampen immunity at the site of infection but may also interfere with protective immune responses at peripheral sites before the dissemination of the bacterium.
In this study we demonstrate that, unlike more attenuated strains, virulent type A F. tularensis actively suppresses early proinflammatory responses in the lung following aerosol infection (Table I⇑). Specifically, F. tularensis Schu4 failed to activate pulmonary DC and macrophages as measured by the changes in key cell surface receptors (MHCII and CD86). Schu4 also failed to induce the early secretion of several proinflammatory cytokines in the lungs and airways, including TNF-α, IL-12, and IL-1β, which are central to initiation of effective host immune responses and the control of Francisella infections (8, 9, 10). The lack of cellular activation could not be attributed to the apoptosis of pulmonary cells. In fact, the relative viability and health of multiple pulmonary cell types, including the endothelial cells, epithelial cells, and monocytes, were remarkable considering the rapidly increasing bacterial loads. Rather, we found compelling evidence of broad and pervasive active interference with early (first 24 h) proinflammatory responses following Schu4 infection. Our evidence also suggests that Schu4 modulates the pulmonary environment, in part via the induction of TGF-β, to favor anti-inflammatory conditions. Together, these data suggest a novel mechanism by which virulent F. tularensis infects, replicates, and disseminates in the host while evading host immune responses.
Many pathogens have evolved strategies to evade host defenses and in many cases can co-opt normal host physiology to gain an advantage following infection. Once such strategy is the induction of TGF-β. TGF-β has multiple functions, including maintaining homeostasis in the host as well as potent anti-inflammatory and antimicrobial activity. For example, the addition of TGF-β to Leishmania-infected macrophages completely abolished the ability of these cells to kill the infecting parasite (33). The neutralization of TGF-β in vivo following infection with Leishmania reduced parasite load and increased wound healing compared with untreated controls (34). In addition to its direct effect on controlling the replication of pathogens in host cells, investigators have also observed that TGF-β plays an important role in limiting the extravasation of granulocytes into various tissues, including the lung. A single administration of DNA encoding TGF-β suppressed the recruitment of eosinophils into the lung following infection with Cryptococcus neoformans, the influenza virus, and the respiratory syncytial virus (30). Interestingly, although this treatment limited lung pathology, it also significantly increased susceptibility to infection.
Considering the multifunctionality of TGF-β as a regulatory and suppressive cytokine and the ability of Schu4 to elicit TGF-β within the first 24 h of infection in both the lung and spleen (Fig. 7⇑), it was possible that this cytokine contributed to Schu4 replication in the host. Following the administration of anti-TGF-β Abs we observed decreased bacterial loads in the lungs concomitant with increased TNF-α in the BAL fluid. However, these differences were not significant compared with isotype-treated controls. Furthermore, there were no differences in the bacterial loads or concentrations of proinflammatory cytokines in the spleens of anti-TGF-β-treated animals compared with those in controls.
The inability of treatment with anti-TGF-β Abs to significantly increase TNF-α while reducing bacterial loads in the lung compared with isotype controls may reflect the relatively small contribution of this cytokine toward the pathogenesis of pneumonic tularemia. However, it is also possible that other molecules capable of modulating the host response work in conjunction with TGF-β to result in the profound inhibition of inflammation observed during the first 48 h of Schu4 infections. The interplay of TGF-β with other immunomodulators such as prostaglandins has been well established. For example, the inhibition of lung NK cells by alveolar macrophages is through the production of both TGF-β and PGE2 by the macrophages (35). The neutralization of TGF-β and the simultaneous inhibition of prostaglandins significantly increased the production of TNF-α and NO while inhibiting replication of M. tuberculosis in the lung (36). More recently Woolard et al. demonstrated that the elicitation of PGE2 by LVS contributes to the successful replication of the bacterium in macrophages while suppressing T cell responses in the lung (37). Thus, it is possible that Schu4 successfully modulates the pulmonary environment via the induction of TGF-β and PGE2. Studies designed to determine the role of PGE2 toward the pathogenesis of Schu4 infections are currently underway in our laboratory.
Many pathogens have developed mechanisms they use to evade host immunity to ensure their survival and transmission. The data presented here demonstrate multiple forms of immunosuppression by which virulent F. tularensis modulates that host response in the lung to its benefit. These mechanisms include the induction of TGF-β, active inhibition of DC responsiveness, and suppression of the host’s ability to recruit effector cells crucial for the resolution of tularemia. These studies also point out important considerations, e.g., the inability of the host to respond to therapeutic stimuli after infection, for the development of novel therapeutics intended to treat individuals exposed to aerosolized F. tularensis. Additional studies aimed at identifying the specific mechanism and bacterial components responsible for this suppression will be central to the development of novel vaccines and therapeutics for F. tularensis and other important pulmonary pathogens.
We thank Dr. Peter Henson for helpful discussions and Shayna Warner and Beth Stallman for excellent technical assistance.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 This work was supported by a grant form the Colorado State University College Research Council and by Rocky Mountain Regional Center of Excellence National Institutes of Health Grant U01 AI056487-01.
↵2 Address correspondence and reprint requests to Dr. Catharine M. Bosio at the current address: National Institutes of Health/National Institute of Allergy and Infectious Diseases/Rocky Mountain Laboratories, 903 South 4th Street, Hamilton, MT 59840. E-mail address:
↵3 Abbreviations used in this paper: LVS, live vaccine strain; BAL, bronchoalveolar lavage; BHI, brain-heart infusion; DC, dendritic cell; IHC, immunohistochemistry; MHCII, MHC class II; MLN, mediastinal lymph node.
- Received October 18, 2006.
- Accepted January 17, 2007.
- Copyright © 2007 by The American Association of Immunologists