ABSTRACT
Upregulation of the transcription factor T-bet is correlated with the strength of protection against secondary challenge with the live vaccine strain (LVS) of Francisella tularensis. Thus, to determine if this mediator had direct consequences in immunity to LVS, we examined its role in infection. Despite substantial in vivo gamma interferon (IFN-γ) levels, T-bet-knockout (KO) mice infected intradermally (i.d.) or intranasally (i.n.) with LVS succumbed to infection with doses 2 log units less than those required for their wild-type (WT) counterparts, and exhibited significantly increased bacterial burdens in the lung and spleen. Lungs of LVS-infected T-bet-KO mice contained fewer lymphocytes and more neutrophils and interleukin-17 than WT mice. LVS-vaccinated T-bet-KO mice survived lethal LVS intraperitoneal secondary challenge but not high doses of LVS i.n. challenge, independently of the route of vaccination. Immune T lymphocytes from the spleens of i.d. LVS-vaccinated WT or KO mice controlled intracellular bacterial replication in an in vitro coculture system, but cultures with T-bet-KO splenocyte supernatants contained less IFN-γ and increased amounts of tumor necrosis factor alpha. In contrast, immune T-bet-KO lung lymphocytes were greatly impaired in controlling intramacrophage growth of LVS; this functional defect is the likely mechanism underpinning the lack of respiratory protection. Taken together, T-bet is important in host resistance to primary LVS infection and i.n. secondary challenge. Thus, T-bet represents a true, useful correlate for immunity to LVS.
INTRODUCTION
In order to develop vaccines against intracellular pathogens, a more complete understanding of T cell functions and products that are required to survive lethal exposure is important. To address these issues, we have taken advantage of the murine model of infection with the Francisella tularensis live vaccine strain (LVS) to search for T cell mechanisms critical to protection. The outcome of F. tularensis LVS infection of inbred mice is route dependent; mice survive high doses of intradermal (i.d.) infection, resulting in protective immunity against secondary lethal intraperitoneal (i.p.) or intranasal (i.n.) challenge (1, 2). Recently, we identified a number of biological mediators that correlate with the relative strength of protection against secondary lethal challenge with LVS (3). F. tularensis vaccine candidates, including LVS, spontaneous colony morphology variants of LVS that express alternative chemotypes of Francisella lipopolysaccharide (LVS-G and LVS-R [4]), and heat-killed (HK) LVS, induce quantitatively different levels of protection in vivo against F. tularensis challenge of vaccinated mice. Further, cocultures containing Francisella-immune splenocytes from mice differentially vaccinated and then stimulated by LVS-infected macrophages were monitored for differential expression of the genes for immune mediators. Given their established importance in mediating protection against other intracellular pathogens, including LVS, it was not surprising that gamma interferon (IFN-γ) and tumor necrosis factor alpha (TNF-α) were identified among these mediators. However, we also identified correlates that have not previously been associated with protection against Francisella. For example, the Th1 transcription factor T-bet was more strongly upregulated in LVS-immune lymphocytes from spleens (3) and lungs (R. De Pascalis et al., submitted for publication) compared to those from mice immunized with LVS-R or HK LVS.
F. tularensis is an intracellular pathogen that causes tularemia in humans. Severe and sometimes fatal respiratory tularemia typically results from inhalation of the bacteria, whereas ulceroglandular tularemia following infection through skin is less severe and rarely results in death. F. tularensis LVS was attenuated from F. tularensis subsp. holarctica (type B) by serial in vitro and in vivo passage (5). Although used as an investigational vaccine, to date, F. tularensis LVS is not licensed for use in the United States.
Various studies examining the mechanisms of protection against lethal LVS challenge of mice have demonstrated that mice deficient in IFN-γ are highly susceptible to infection via the parenteral, pulmonary, and intravenous routes, and typically succumb to infection within a week (1, 6, 7). In addition to IFN-γ, TNF-α (1) and nitric oxide (8) are also important in controlling primary sublethal LVS infection. Following i.d. LVS infection of mice, IFN-γ is produced by a variety of cells, including natural killer (NK) cells, neutrophils, and dendritic cells (DCs) (9). Notably, IFN-γ production in mice is regulated by the transcription factor T-bet.
T-bet (also known as Tbx-21) was first described by Szabo et al. in 2000 as a novel T-box transcription factor expressed in T cells (10). CD4+ T cells in T-bet-deficient mice produce substantially less IFN-γ, and fail to differentiate into the Th1 lineage (11). However, T-bet expression is not restricted to T lymphocytes but is also present in B cells, DCs, and NK cells. Thus, in addition to regulating IFN-γ in CD4+ T cells, T-bet is required for optimal production of IFN-γ in CD8+ T cells and NK cells (12–14). Furthermore, T-bet also controls the ability of CD8+ T cells and NK cells to mediate their cytotoxic function. Outside of its function in regulating T and NK cell function, T-bet has been shown to be required for priming of DCs for antigen-specific CD4+ T cells and enhancing secretion of antigen-specific IgG2 by B cells (15, 16).
Many infection models demonstrate increased susceptibility in T-bet-knockout (KO) mice, including the Leishmania species Leishmania donovani (17) and L. major (11), Mycobacterium tuberculosis (18), Salmonella enterica serovar Typhimurium (19), and Mycoplasma pulmonis (20). The mechanism underlying increased susceptibility appears to depend on the model; for example, following L. major infection, T-bet-KO mice exhibited increased levels of Th2 cytokines, such as interleukin-4 (IL-4) and IL-5 (11), while during M. tuberculosis infection, CD4+ lung lymphocytes from T-bet-KO mice produced increased levels of the immunosuppressive cytokine IL-10 compared to the levels produced by M. tuberculosis-infected wild-type (WT) mice (18). To date, the role of T-bet in F. tularensis infection is unknown. Further, it is unclear whether T-bet has different roles in different tissues. Given the clear role of Th1 T cells in responding to Francisella infection, as well as the identification of T-bet as a potential correlate, in the current study, we directly examined the role of T-bet in F. tularensis LVS infection.
MATERIALS AND METHODS
Mice.Male C57BL/6J mice and B6.129S6-Tbx21tm1Glm/J (T-bet-KO)-deficient mice on a C57BL/6 background were purchased from The Jackson Laboratory (Bar Harbor, ME). Animals were housed in a barrier environment at the Center for Biologics Evaluation and Research, U.S. Food and Drug Administration, and were fed autoclaved food and water ad libitum. Procedures were performed according to approved protocols under the guidelines of the Center for Biologics Evaluation and Research Animal Care and Use Committee. Age-matched mice between 6 and 12 weeks old were used throughout the study.
Bacteria and growth conditions.F. tularensis LVS (ATCC 29684; American Type Culture Collection, Manassas, VA) and F. tularensis subsp. tularensis (strain SchuS4), obtained from Jeannine Peterson (Centers for Disease Control and Prevention, Fort Collins, CO), were cultured on modified Mueller-Hinton (MH) agar plates supplemented with ferric pyrophosphate, glucose, and IsoVitaleX (Becton, Dickinson and Company). All experiments utilizing SchuS4 were conducted within the Centers for Disease Control and Prevention-certified animal biosafety level 3/biosafety level 3 facility at Rocky Mountain Laboratories. Active mid-log-phase bacteria grown in MH broth were harvested and stored at −80°C; 0.5-ml aliquots were thawed periodically. LVS viability was measured using a LIVE/DEAD BacLight bacterial viability kit (Invitrogen), following the manufacturer's protocol. Viable bacteria were quantified by plating serial dilutions on MH agar plates that were then incubated for 2 to 3 days at 37°C in 5% CO2.
In vivo bacterial infections.C57BL/6J and T-bet-KO mice were infected with LVS at various doses delivered through the i.d., i.p., or i.n. route. Mice were given F. tularensis LVS diluted in sterile phosphate-buffered saline (PBS) with <0.01 ng/ml endotoxin either i.p. in 0.5 ml, i.d. in 0.1 ml, or i.n. in 20 μl, which was given in one nostril. For i.n. LVS infections, mice were anesthetized with 0.1 ml of a cocktail of ketamine HCl (1.5 mg/0.1 ml; Ketaject; Phoenix Pharmaceuticals, St. Joseph, MO) and xylazine (0.3 mg/0.1 ml; AnaSed; Lloyd Laboratories, Shenandoah, IA) diluted in sterile PBS and given intraperitoneally. For i.n. infections with SchuS4, mice were anesthetized with a single 70-μl injection of 12.5 mg/ml ketamine plus 3.8 mg/ml xylazine immediately prior to infection. Actual doses of inoculated bacteria were determined by simultaneous plate counts. The numbers of CFU in the organs of infected mice were determined at various time points. Mice were euthanized, and organs were aseptically removed and homogenized in a stomacher in 5 ml sterile PBS. Appropriate dilutions were plated on modified MH plates, and the bacterial CFU were enumerated 2 to 3 days later.
Depletion of neutrophils and IL-17 neutralization.Purified anti-Ly6G/Ly6C (RB6-8C5; National Cell Culture Center, Minneapolis, MN), which depletes neutrophils as well as some plasmacytoid dendritic cells and monocytes, or purified anti-LY6C (1A8; BioXCell, West Lebanon, NH), which primarily recognizes neutrophils, were used for in vivo depletion. Mice were treated by i.p. injection of 250 μg of either antibody 1 day before infection with LVS and on days 3 and 7 after infection. Depletion of neutrophils was monitored by flow cytometry using a panel of antibodies against cell surface proteins on myeloid cells combined with determination of light-scatter properties. For IL-17 neutralization, mice were treated i.p. with 100 μg purified anti-mouse IL-17A antibodies (clone eBioMM17F3; eBioscience, San Diego, CA) on the day of infection and every 2 to 3 days thereafter through day 10. Neutralization of IL-17A was monitored by obtaining blood 2 days after LVS infection and assessing the levels of IL-17A in serum by enzyme-linked immunosorbent assay (ELISA).
Blood and serum chemistry.Blood was collected from the femoral artery and heart of infected C57BL/6J and T-bet-KO mice. Whole blood was collected and diluted in 0.5 mg/ml EDTA to prevent clotting. Sera were also prepared from whole blood using Sarstedt serum gel microtubes (Fisher Scientific, Pittsburg, PA). Whole-blood and serum samples were sent to the Department of Laboratory Medicine, Clinical Center, National Institutes of Health, in Bethesda, MD, for analyses; a whole-blood complete blood count (CBC) was performed on a Cell Dynn 3700 analyzer (Abbott Diagnostics, Abbott Park, IL). Serum chemistry analyses were performed on a Siemens Dimension Vista 1500 analyzer (Siemens Healthcare Diagnostics, Tarrytown, NY); analytes tested in the serum are listed in Table 1.
Comparison of serum chemistry in LVS-infected WT and T-bet-KO micea
Preparation of lymphocytes.Spleens were aseptically removed and transferred to a sterile petri dish with PBS with 2% fetal bovine serum (HyClone, Logan, CT). The plunger of a 3-ml sterile syringe was used to homogenize the spleens, and the homogenate was passed through 40-μm-pore-size cell strainers. Cells were washed in PBS–2% fetal calf serum (FCS) and then resuspended in ACK lysing buffer (Invitrogen, Grand Island, NY) for lysis of red blood cells. Cells were again washed in PBS–2% FCS, and live cells were enumerated using trypan blue and a hemocytometer. Lungs were aseptically removed and transferred to a petri dish containing sterile PBS–2% FCS. Lungs were carefully sliced into small pieces and then incubated in 1.4 mg/ml collagenase (Gibco) plus 2 mg/ml ampicillin for 1 h at 37°C. The collagenase mixture containing lung tissue was removed, placed in a filter bag, and further homogenized with the plunger of a syringe. Cells were then collected and washed in PBS–2% FCS at 4°C. Cells were resuspended in 4 ml of PBS–2% FCS and gently placed on top of a lymphocyte separation medium gradient (LSM; Mediatech, Inc.). The gradients were centrifuged at 1,500 rpm for 20 min at room temperature, and the lymphocyte layer was collected. Lymphocytes were washed in PBS–2% FCS at 4°C and passed through a 40-μm-pore-size filter, and live cells were enumerated using trypan blue and a hemocytometer.
Flow cytometry analyses.Single-cell suspensions prepared from spleens or lungs were stained for a panel of murine cell surface markers and analyzed using a Becton, Dickinson LSR II flow cytometer (San Jose, CA) and FlowJo software (Tree Star, Inc.), as previously described (3). Briefly, cells were washed and resuspended in PBS–2% FCS. Nonspecific binding of antibodies was inhibited by blocking Fc receptors with anti-CD16/CD32 (Fc Block; BD Pharmingen) for 10 min on ice. Live/dead staining was performed using a commercially available reagent and following the manufacturer's protocol (LIVE/DEAD staining kit; Invitrogen). Following live/dead staining, cells were washed and stained for cell surface markers. Antibody concentrations were optimized separately for use in nine-color staining protocols, using appropriate fluorochrome-labeled isotype-matched control antibodies. The following antibodies were used: anti-B220 (clone RA3-6B2), anti-CD19 (clone 1D3), anti-T cell receptor β (anti-TCRβ; clone H57-597), anti-CD4 (clone RM4-5), anti-CD8α (H35-17.2), anti-NK1.1 (clone PK136), anti-CD11b (clone M1/70), anti-Gr-1 (clone RB6-8C5), and anti-CD11c (cloneHL3) (the antibodies listed above were purchased from BD Pharmingen). Each antibody was labeled with a variety of fluorochromes, as needed. Side-scatter (width and height) and forward-scatter (width and height) plots were used to gate on singlet events prior to all subsequent analyses.
In vitro coculture assays of T cell function.Bone marrow-derived macrophages (BMDMs) were prepared and cultured as previously described (21, 22). Briefly, bone marrow was flushed from the femurs of mice using complete Dulbecco modified Eagle medium supplemented with 10% heat-inactivated fetal bovine serum (HyClone), 10% L-929-conditioned medium, 0.2 mM l-glutamine, 10 mM HEPES buffer, and 0.1 mM nonessential amino acids. Cells were seeded in 24- or 48-well plates and incubated for 7 days 37°C in 5% CO2. Medium was replaced every 2 to 3 days. BMDMs were infected with F. tularensis LVS at a multiplicity of infection (MOI) of 1:20 (bacteria/macrophages). At 2 h following infection, the medium containing the bacteria was removed and replaced with medium containing 100 μg/ml gentamicin. Cells were incubated in the presence of gentamicin for 1 h to kill extracellular bacteria and then washed three times with PBS–2% FCS. Single-cell suspensions of lymphocytes derived from vaccinated mice (5 × 106/ml or as indicated) were added to LVS-infected macrophages. Where indicated, TNF-α-neutralizing antibodies (BD Pharmingen) were added at the initiation of cocultures at a concentration of 25 μg/ml. At 72 h after infection, supernatants were harvested and stored at −70°C. Adherent infected macrophages were lysed with sterile water, lysates were serially diluted in sterile PBS, and appropriate dilutions were plated on modified MH plates for bacterial enumeration.
Cytokine and nitrite measurements.Culture supernatants were assayed for IFN-γ, IL-12p40, IL-12p70, IL-6, IL-17, IL-5, and TNF-α using standard sandwich ELISAs. Antibody pairs and standards were purchased from BD Pharmingen, and the assays were performed according to the manufacturer's instructions. The absorbance was read at 405 nm on a VersaMax tunable microplate reader with a reference wavelength of 630 nm. Cytokine levels were measured by comparison to recombinant standard proteins using four-parameter fit regression in SoftMax Pro ELISA analysis software (Molecular Devices). The nitric oxide in culture supernatants was estimated using the Griess reaction following the manufacturer's protocol (Molecular Probes). Nitrite (NO2) was measured by comparison to serially diluted NaNO2 as a standard using four-parameter fit regression, as described above.
Characterization of antibody responses.The titers of specific anti-LVS serum antibodies were determined by ELISA as described previously (23). Briefly, Immulon 1 plates were coated with live LVS, washed, and blocked with 10% calf serum, and serum samples were serially diluted. In each assay, serum from naive mice was used as a negative control and serum from LVS-hyperimmune mice was used as a positive control. Horseradish peroxidase-labeled antibodies (anti-IgM; anti-IgG that detects IgG1, IgG2b, IgG2c, and IgG3 or specific subtype IgG2c and subtype IgG1) (Southern Biotech) were added, and ABTS [2,2′-azinobis(3-ethylbenzthiazolinesulfonic acid)] peroxidase substrate (Kirkegaard & Perry Laboratories) was used for color development. The endpoint titer was defined as the lowest dilution of serum that gave an optical density (OD) at 405 nm greater than the OD plus 3 standard deviations of the mean of the matched dilution of normal prebleed mouse serum and also greater than 0.025.
Histopathology.Tissue sections from livers, lungs, and spleens were fixed with 10% buffered formalin. Samples were then sent to American Histolabs, Inc. (Gaithersburg, MD), where the tissues were embedded in paraffin and sectioned for slides and the slides were stained with hematoxylin-eosin (H&E). Pathology slides were analyzed by a board-certified pathologist in a blinded fashion.
Statistical analyses.All statistical analyses were performed using Student's t test and InStat software (GraphPad Software Inc.). Survival results are expressed as Kaplan-Meier curves, and P values were determined using a log-rank test. P values of ≤0.05 were considered significant. The 50% lethal doses (LD50s) were calculated using the Reed-Muench method (24).
RESULTS
T-bet-KO mice are highly susceptible to F. tularensis LVS infection and exhibit increased bacterial replication.To examine the role of T-bet in host resistance to F. tularensis LVS, mice deficient in T-bet were inoculated with a range of doses of LVS via the i.d. or i.n. route. As expected, WT mice survived intradermal challenge with 104 and 105 CFU of LVS (data not shown and Fig. 1A). Moreover, LVS was not detected in any organ tested in WT mice that survived these challenge doses (data not shown). In contrast, T-bet-KO mice succumbed to infection following i.d. inoculation with doses of F. tularensis LVS that were sublethal in WT mice (104 or 105 bacteria) (Fig. 1A). The calculated i.d. LD50 (24) of T-bet-KO mice was <104 LVS bacteria, which is over 2 log units less than that of WT mice (LD50 > 106). We also examined the survival of T-bet-KO mice following i.n. infection of LVS. As shown in Fig. 1B, T-bet-KO mice were also more susceptible than WT mice to primary i.n. infection. The i.n. LD50 for WT mice following F. tularensis LVS infection was ∼3.5 × 103. A large proportion of T-bet-KO mice succumbed to infection with an i.n. dose of 102 bacteria, and the calculated LD50 was ∼2 × 102.
Susceptibility of T-bet-KO mice to F. tularensis LVS infection. WT and T-bet-KO mice (n = 5 mice/group) were inoculated with 105 or 104 F. tularensis LVS bacteria i.d. (A) or 103 or 102 F. tularensis LVS bacteria i.n. (B). The actual dose of infection was confirmed by simultaneous plate counts. The calculated LD50s for WT and T-bet-KO mice following infections either i.d. or i.n. are displayed next to the graph. These data are from a single experiment that is representative of three experiments of similar design, in which additional doses of 103 and 102 LVS bacteria i.d. or 101 and 100 LVS bacteria i.n. were tested in mice, but for clarity the results are not displayed here. All WT and T-bet-KO mice survived the inoculation with the additional doses. The results are expressed as Kaplan-Meier survival curves, and the data were analyzed using the log-rank test. The median time to death was significantly longer in WT mice infected with 105 and 104 LVS bacteria i.d. and 103 and 102 LVS bacteria i.n. than T-bet-KO mice (P ≤ 0.05).
Given the increased susceptibility of T-bet-KO mice to LVS infection, we next determined if this was due to a reduced ability of T-bet-KO mice to control bacterial replication in target organs, compared to WT control mice. Mice were infected with 105 LVS bacteria i.d. (Fig. 2A and 3A) or 103 LVS bacteria i.n. (Fig. 2B and 3B), doses that are lethal for T-bet-KO mice but sublethal for WT mice; and bacterial burdens and IFN-γ levels were measured in the spleens, livers, and lungs on days 3, 5, and 7 after infection. On day 3 after i.d. LVS infection, there were no significant differences in bacterial organ burdens in the spleens or livers of LVS-infected T-bet-KO mice and those of WT mice (Fig. 2A). In contrast, significantly higher numbers of bacteria were observed in the lungs of T-bet-KO mice than in those of WT animals on day 3 after infection (Fig. 2A). By day 7 after infection, there was a significant increase in the numbers of bacteria in all organs from LVS-infected T-bet-KO mice compared to those from WT mice (Fig. 2A). Following infection with 103 LVS bacteria administered via the i.n. route, significantly higher numbers of bacteria were detected in the lungs of T-bet-KO mice by day 7 after infection. Similarly, at this time point there were significant increases in the numbers of bacteria in the spleens and livers of i.n. LVS-infected T-bet-KO mice compared to those in the spleens and livers of WT mice (Fig. 2B). In addition, IFN-γ levels were measured in organ homogenates following i.d. and i.n. LVS infection. Interestingly, IFN-γ levels in the spleens, livers, or lungs of i.d. LVS-infected T-bet-KO mice were not significantly different from those in the spleens, livers, or lungs of i.d. LVS-infected WT mice (Fig. 3A). Surprisingly, IFN-γ levels were significantly higher in the spleens of T-bet-KO mice infected through the i.n. route; however, IFN-γ levels significantly increased in the lungs of WT mice on day 7 following i.n. LVS infection (Fig. 3B).
Bacterial organ burdens in WT and T-bet-KO mice following primary lethal F. tularensis LVS infection. WT and T-bet-KO mice were infected with F. tularensis LVS at a dose of 105 bacteria i.d. (A) or 103 bacteria i.n. (B). Three mice per group were euthanized at each time point on day 3, 5, or 7 after infection, and the numbers of CFU per organ were determined. The doses of infection were confirmed by simultaneous plate counts. CFU data are representative of those from five independent experiments of similar design. *, P ≤ 0.05 by Student's t test in a comparison of the numbers of CFU in WT and KO mice; **, P ≤ 0.01 by Student's t test in a comparison of the numbers of CFU in WT and KO mice.
Bacterial cytokine levels in WT and T-bet-KO mice following primary lethal F. tularensis LVS infection. WT and T-bet-KO mice were infected with F. tularensis LVS at a dose of 105 bacteria i.d. (A) or 103 bacteria i.n. (B). Three mice per group were euthanized at each time point on day 3, 5, or 7 after infection, and organ homogenates were analyzed for IFN-γ. Data are from the same experiment whose results are shown in Fig. 2 and are representative of those from two independent experiments. *, P ≤ 0.05 by Student's t test in a comparison of the numbers of CFU in WT and KO mice; **, P ≤ 0.01 by Student's t test in a comparison of the numbers of CFU in WT and KO mice.
We further examined the ability of T-bet-KO mice to clear sublethal i.n. LVS infection over a longer time frame. WT or T-bet-KO mice were infected with a sublethal dose (101 bacteria) of F. tularensis LVS by the i.n. route, and mice were sacrificed on days 3, 6, 11, and 32 after infection. Despite the low dose, LVS-infected T-bet-KO mice exhibited increased bacterial burdens in spleens, livers, and lungs early after infection (Fig. 4, days 3 to 11). Consistent with previous studies (25), by day 32, the amounts of bacteria in all organs of LVS-infected WT mice were below the limit of detection. By this later time point, LVS was not detected in T-bet-KO mice in either the livers or the lungs; however, there was still a low but detectable level of bacteria remaining in the spleens of LVS-infected T-bet-KO mice 32 days after infection. The concentrations of cytokines were measured in the spleens, livers, and lungs on days 6, 11, and 32. Although there were no significant differences in the levels of IFN-γ in the T-bet-KO mouse lungs throughout the course of this lower dose of i.n. infection (Fig. 5A), there were significantly higher levels of TNF-α on day 6 after infection (Fig. 5B). Furthermore, by day 11 there were elevated levels of IL-17 in the T-bet-KO mouse lungs and spleens compared with those in WT mouse organs (Fig. 5C). Although the levels of IL-17 were consistently elevated in LVS-infected T-bet-KO mouse lungs and spleens compared to the levels in the lungs and spleens from WT mice, these differences were not significant across repeated experiments. By day 32 after LVS infection, the concentrations of cytokines were similar to those in uninfected animals. No differences in the levels of IL-12 were noted throughout the course of infection (data not shown).
Bacterial organ burden in WT and T-bet-KO mice following primary sublethal F. tularensis LVS infection. WT and T-bet-KO mice were infected with a sublethal dose of 101 F. tularensis LVS bacteria i.n. Three mice per group were sacrificed on days 3, 6, 11, and 32 after infection, and the numbers of CFU per organ were determined and are expressed as the total number of CFU per organ. The dose of infection was confirmed by simultaneous plate counts. These data are representative of those from two experiments of similar design. *, P ≤ 0.05 by Student's t test in a pairwise comparison of the numbers of CFU in WT and KO mice; **, P ≤ 0.01 by Student's t test in a pairwise comparison of the numbers of CFU in WT and KO mice.
Cytokine levels in organ homogenates of WT and T-bet-KO mice following primary sublethal F. tularensis LVS infection. WT and T-bet KO mice were infected with a sublethal dose of 101 F. tularensis LVS bacteria i.n. Three mice per group were sacrificed on days 6, 11, and 32 after infection, and organ homogenates were collected. Concentrations of IFN-γ (A), TNF-α (B), and IL-17 (C) in lung homogenates were assessed on days 6, 11, and 32 after infection. These data are from the same experiment whose results are shown in Fig. 4 and are representative of those from two experiments of similar design. **, P ≤ 0.01 by Student's t test in a pairwise comparison of assays with WT and KO mice.
These observations raised the possibility that the excessive IL-17 and/or neutrophils (a common source of IL-17) produced in the absence of T-bet contributed to pathology or death. Although both contribute substantially to the survival of sublethal LVS infection in wild-type mice, we nonetheless performed small-scale experiments to examine the possibility that either was detrimental in T-bet-KO mice. T-bet-KO mice were treated with anti-IL-17A antibodies and then infected i.n. with 103 LVS bacteria. Although the time to death was extended slightly by anti-IL-17 treatment, differences were minor compared to the time to death for untreated mice, and all LVS-infected T-bet-KO mice succumbed to infection (data not shown). Similarly, mice treated with either RB6-8C5 or 1A8 antibodies were infected i.n. with a low dose of 102 LVS bacteria. However, neither WT nor T-bet-KO mice survived infection when neutrophils were depleted by either treatment (data not shown). These outcomes emphasize the complex roles, both positive and negative, that both IL-17A and neutrophils play in supporting survival after respiratory LVS infection. Nonetheless, the data are not consistent with the idea that either is a key mediator in decreasing the survival of LVS-challenged T-bet-KO mice when made in excess.
Loss of T-bet leads to differential cell recruitment to the lungs.To determine whether differences in cell recruitment were related to the increased susceptibility of T-bet-KO mice, we analyzed the spleen and lung cell populations 6 days after infection with F. tularensis LVS via the i.d. or i.n. route. WT and T-bet-KO mice were infected i.d. with 104 LVS bacteria, a dose that is lethal for T-bet-KO mice and sublethal for WT mice; organs were isolated on days 3 and 6; and cells from each group of mice were pooled. There were no major differences in the cell populations between WT and T-bet-KO mice on day 3 after infection (data not shown). However, by day 6 there was a substantial increase in the percentage (Fig. 6A) and total number (Fig. 6B) of neutrophils (CD45+ TCRβ− CD19− CD11b+ Gr1hi) in the lungs of LVS-infected T-bet-KO mice compared to those for WT mice. This dramatic increase in neutrophils was not observed in the spleens of T-bet-KO mice (data not shown). Thus, there was an obvious expansion or recruitment of the neutrophils to the lungs in the T-bet-KO mice between days 3 and 6 after LVS infection. There was also a notable decrease in both the proportion (Fig. 6A) and total numbers (Fig. 6B) of B cells (CD45+ TCRβ− B220+ CD19+), T cell subpopulations (CD45+ B220− TCRβ+ CD4+ CD8− and CD45+ B220− TCRβ+ CD4− CD8+), DCs (CD45+ TCRβ− CD19− CD11c+), and natural killer cells (CD45+ TCRβ− CD19− NK1.1+) in the lungs of i.d. LVS-infected T-bet-KO mice compared to the proportion and numbers in the lungs of their WT counterparts. In addition, the total numbers of macrophages in the lungs of LVS-infected T-bet-KO mice were increased compared to the numbers for WT mice (Fig. 6B), although this was not reflected in the proportion of macrophages (Fig. 6A). Similar to i.d. infection, the proportion and total numbers of DCs and NK cells were also decreased following i.n. infection (data not shown).
Quantification of lung cell subpopulations in WT and T-bet-KO mice after intradermal LVS infection. WT and T-bet KO mice were infected with 104 LVS bacteria i.d. On day 6 following infection, lung cells were harvested and assessed by flow cytometry for B cells (CD45+ B220+ CD19+), CD8+ T cells (CD45+ B220− TCRβ+ CD8+), CD4+ T cells (CD45+ B220− TCRβ+ CD4+), neutrophils (polymorphonuclear leukocytes [PMNs]; CD45+ B220− TCRβ− CD19− CD11b+ Gr1hi), macrophages (MΦ; CD45+ B220− TCRβ− CD19− CD11b+ Gr1−), DCs (CD45+ TCRβ− CD19− CD11c+), and NK cells (CD45+ TCRβ− CD19− NK1.1+). The proportions of cell types (A) and total cell numbers (B) were calculated on the basis of the total recovered cell yields, after application of the indicated gating scheme. Flow data comprised those for pooled cells obtained from the lungs of 3 mice per group, and the results of one experiment that is representative of four independent experiments are depicted.
In addition to analysis of the cells in organs, blood from WT and T-bet-KO mice infected with 103 LVS bacteria i.n. was analyzed by traditional clinical CBC analysis on day 6 after infection. Unlike the results from lungs, there was no increase in the levels of neutrophils in blood from LVS-infected T-bet-KO mice on day 6, and LVS-infected WT and T-bet-KO mice had similar proportions of monocytes, lymphocytes, and basophils in blood (data not shown). Of note, LVS-infected T-bet-KO mice had a considerably higher percentage of eosinophils in the blood than WT mice (∼16% and ∼6%, respectively), which may be indicative of a Th2 environment.
Finally, blood samples from WT and T-bet-KO mice infected with 103 LVS bacteria i.n. were collected on day 6 after infection, and serum was obtained for serum chemistry analyses. The analytes tested in the sera are listed in Table 1. The levels of the liver enzymes alanine aminotransferase (ALT), aspartate aminotransferase (AST), and creatine kinase (CK) appeared to be mildly elevated in the LVS-infected T-bet-KO mice compared to the levels in WT mice, although these differences were not statistically significant. However, there were no obvious differences in pathology in the livers of LVS-infected WT and T-bet-KO mice at this time point (data not shown).
Characterization of role of T-bet in secondary immune responses to F. tularensis LVS.T-bet was recently identified by our laboratory to be an immune mediator whose relative gene expression was more highly upregulated in lymphocytes from the spleens of LVS-immune mice than in lymphocytes from the spleens of less protected LVS-R- or HK-LVS-immune mice (3). We therefore examined the role of T-bet during secondary adaptive immune responses to LVS. WT and T-bet-KO mice were vaccinated by sublethal infection with either 103 LVS bacteria i.d. or 101 LVS bacteria i.n. At approximately 6 weeks following vaccination, mice were challenged i.p. with 2 × 106 LVS bacteria. Both the WT and T-bet-KO vaccinated groups survived secondary lethal i.p. challenge, while all naive mice died by day 4 (Table 2). We next examined the ability of T-bet-KO mice to survive intranasal secondary challenge. WT and T-bet-KO mice were vaccinated with sublethal doses either i.d. or i.n. and challenged with high doses of LVS i.n. (5 × 107 or 5 × 106 bacteria). As summarized in Table 2, naive WT and naive T-bet-KO mice succumbed to i.n. challenge with 5 × 107 LVS bacteria by day 6. WT mice vaccinated through either the i.d. or i.n. route largely survived high doses of secondary i.n. challenge; however, T-bet-KO mice vaccinated i.d. all died after receiving the highest i.n. challenge dose of 5 × 107 bacteria, and 40% of T-bet-KO mice died by day 6 following challenge with a dose of 5 × 106 bacteria i.n. Similar to i.d. vaccinated mice, all WT mice vaccinated i.n. survived secondary i.n. challenge, while 100% of T-bet-KO mice vaccinated i.n. died following challenge, with the highest dose of secondary challenge (5 × 107 bacteria) and 40% died following challenge with a lower dose (5 × 106 bacteria). These results indicate that T-bet is important for protection following i.n. lethal challenge with LVS, regardless of the route of vaccination. To further assess the role of T-bet during secondary challenge, mice were vaccinated with LVS i.d. and challenged subcutaneously with ∼50 CFU of the fully virulent strain F. tularensis SchuS4. All WT and T-bet-KO mice eventually succumbed to the challenge, but T-bet-KO mice died approximately 2 days earlier than WT mice. These experiments are limited, however, by the necessity of using mice on a C57BL/6 background; C57BL/6 mice remain quite susceptible to challenge with fully virulent Francisella even when vaccinated. Nonetheless, the decrease in time to death may suggest a role for T-bet during parenteral immune responses directed against SchuS4 in vaccinated mice (data not shown).
Survival of secondary LVS challenge by WT and T-bet-KO micea
Antibody production has been shown to play an important role in protection against F. tularensis LVS, and T-bet deficiency results in reduced production of antigen-specific IgG2 by B cells. We therefore examined the quantities and subclasses of anti-LVS antibodies following vaccination of WT and T-bet-KO mice. At approximately 3 weeks after sublethal i.d. or i.n. LVS vaccination, WT mouse serum contained moderate amounts of IgM and IgG anti-LVS antibodies; the main IgG subclass detected in the WT mouse serum was IgG2c, with essentially no detectable levels of IgG1. However, high levels of IgM and total IgG were detected following i.d. LVS vaccination of T-bet-KO mice compared to the levels detected in WT mice (Table 3). In addition, the IgG response shifted from the typical Th1 IgG2 response to a response consisting of IgG1, a Th2-associated subclass. There were no substantial differences between the total levels of serum anti-LVS IgG or IgM antibodies following i.n. LVS vaccination in T-bet-KO mice and WT mice; however, there was a clear shift in the response from IgG2 to IgG1 in LVS-vaccinated T-bet-KO mice.
Antibody production following F. tularensis vaccination of WT or T-bet-KO micea
LVS-immune T-bet-KO mouse splenocytes control intramacrophage LVS replication in vitro, but LVS-immune T-bet-KO mouse lung lymphocytes do not.We next examined the role of T-bet in controlling intramacrophage growth of F. tularensis LVS, the setting in which increased T-bet expression was first detected. Splenocytes were isolated from WT and T-bet-KO mice approximately 6 weeks after i.d. vaccination with LVS. WT or T-bet-KO splenocytes were cocultured with LVS-infected WT BMDMs, and recovery of bacteria was monitored 72 h after infection. LVS replicated exponentially in WT macrophages, and addition of naive WT splenocytes did not impact intramacrophage LVS growth (Fig. 7A; see Fig. S1A in the supplemental material). In contrast, splenocytes from both WT and LVS-immune T-bet-KO mice controlled intracellular LVS replication to a similar degree; the degree of control was dose dependent, and the relative efficiency was comparable (see Fig. S1A in the supplemental material). We further examined the production of mediators secreted in the supernatants of these cocultures, including TNF-α, IFN-γ, IL-5, IL-17, IL-6, and nitric oxide. LVS-immune T-bet-KO splenocytes produced very low levels of IFN-γ, and these levels were significantly lower than those from LVS-immune WT splenocytes (Fig. 7B). However, significantly increased levels of TNF-α were detected in cocultures containing LVS-immune T-bet-KO splenocytes (Fig. 7C). In addition to TNF-α, significantly increased levels of IL-17 (Fig. 7D) and IL-6 and IL-5 (data not shown) were also detected in supernatants from LVS-immune T-bet-KO lymphocytes compared to the levels detected in supernatants from LVS-immune WT lymphocytes. Consistent with control of intramacrophage LVS growth, the levels of nitric oxide were similar in cocultures containing lymphocytes from either LVS-immune WT or LVS-immune T-bet-KO mice (Fig. 7E). Similar to the degree of control of bacterial replication, cytokine levels were also dependent on cell number (see Fig. S1B to F in the supplemental material). Similarly, we also examined the function of splenocytes from WT and T-bet-KO mice vaccinated with LVS i.n. Consistent with the observation that both WT and T-bet-KO mice vaccinated with LVS i.d. survived i.p. LVS challenge (Table 2), splenocytes from WT and T-bet-KO mice vaccinated i.n. controlled LVS intramacrophage growth equally well (data not shown).
Control of F. tularensis LVS intramacrophage replication by LVS-immune splenocytes (spl). WT and T-bet-KO mice were infected i.d. with 103 LVS bacteria. Six weeks later, 2 mice from each group, including naive WT mice, were sacrificed, and splenocytes were isolated. WT BMDMs were infected with F. tularensis LVS at an MOI of 1:20, and 106/ml LVS-immune splenocytes were added to triplicate wells. Seventy-two hours later, intracellular bacteria were enumerated (A) and supernatants were assayed for IFN-γ (B), TNF-α (C), IL-17 (D), or nitric oxide (E). Overlay assays were also performed in the absence or presence of anti-TNF-α-neutralizing antibodies (25 μg/ml). Seventy-two hours later, intracellular bacteria were enumerated (F) and supernatants were assayed for TNF-α (G) or nitric oxide (H). Error bars represent the SDs of triplicate wells. Data are representative of those from three experiments of similar design. *, P ≤ 0.05 by the Student t test in a comparison of WT and KO cultures; **, P ≤ 0.01 by the Student t test in a comparison of WT and KO cultures; ***, P < 0.001. There was no significant difference (P > 0.05) between the recovery of bacteria in cocultures with LVS-immune WT splenocytes and LVS-immune T-bet-KO splenocytes.
In order to directly examine the role that TNF-α played in the capacity of LVS-immune T-bet-KO lymphocytes to control intramacrophage LVS replication, we deprived cultures of TNF-α by adding TNF-α-neutralizing antibodies to wells at the time of addition of LVS-immune splenocytes. Neutralization of TNF-α decreased the ability of both LVS-immune WT splenocytes and LVS-immune T-bet-KO splenocytes to control LVS intramacrophage growth, but had a slightly greater impact when added to cocultures with LVS-immune T-bet-KO splenocytes (Fig. 7F). Cytokine levels were also measured in the coculture supernatants. Although TNF-α levels were greatly reduced in LVS-immune T-bet-KO splenocytes in the presence of TNF-α-neutralizing antibody, low but detectable levels of TNF-α were still present (Fig. 7G). Addition of anti-TNF-α-neutralizing antibodies had little impact on the levels of IFN-γ or IL-17 (data not shown); however, nitric oxide levels were significantly decreased following TNF-α neutralization (Fig. 7H).
Because of the reduced ability of LVS-immune T-bet-KO mice to survive intranasal LVS secondary challenge, we next examined the ability of WT and T-bet-KO lung lymphocytes to control bacterial replication. WT and T-bet-KO mice were vaccinated i.d., and at approximately 6 weeks following vaccination, LVS-immune lung lymphocytes were isolated and added to LVS-infected macrophages. WT naive lung lymphocytes were not able to control LVS intramacrophage growth (Fig. 8A). LVS-immune WT lung lymphocytes controlled intramacrophage growth (Fig. 8A) at similar levels as LVS-immune WT splenocytes (compare to Fig. 7A). However, LVS-immune T-bet-KO lung lymphocytes were not able to control LVS growth nearly as well as WT lung lymphocytes (Fig. 8A). Supernatants from cocultures were collected and assayed for cytokines. Similar to cocultures containing splenocytes, LVS-immune T-bet-KO lung lymphocytes produced significantly decreased amounts of IFN-γ compared to those found in supernatants from cocultures with WT lung lymphocytes (Fig. 8B), and this was accompanied by significantly increased levels of IL-17 (Fig. 8D) and IL-5 (data not shown). Although the levels of TNF-α consistently tended to be higher in cocultures containing LVS-immune T-bet-KO lung lymphocytes (Fig. 8C), across multiple experiments the differences were not significant compared to the TNF-α levels produced by WT lung lymphocytes. The levels of nitric oxide were low and did not differ between WT and T-bet-KO cocultures (Fig. 8E). Thus, T-bet-KO lung lymphocytes are qualitatively different from T-bet-KO splenocytes. Taken together, these results indicate that survival of i.n. secondary challenge is at least partially dependent on T-bet. The mechanism of reduced protection against respiratory challenge may be reflected in the greatly reduced ability of T-bet-KO lung lymphocytes to control intramacrophage LVS growth.
Control of F. tularensis LVS intramacrophage replication by LVS-immune lung lymphocytes. WT and T-bet-KO mice were infected i.d. with 103 LVS bacteria. Six weeks later, 5 mice from each group, including naive WT mice, were euthanized, and lung lymphocytes were isolated. BMDMs from WT mice were infected with F. tularensis LVS at an MOI of 1:20, and 106/ml LVS-immune lung lymphocytes were added to triplicate wells. Seventy-two hours later, intracellular bacteria were enumerated (A) and supernatants were assayed for IFN-γ (B), TNF-α (C), IL-17 (D), or nitric oxide (E). Error bars represent SDs of triplicate wells. Data are representative of those from eight experiments of similar design. *, P ≤ 0.05 by the Student t test in a comparison of WT and KO cultures; **, P ≤ 0.01 by the Student t test in a comparison of WT and KO cultures; ***, P < 0.001.
DISCUSSION
Previous findings from our laboratory suggested that differential regulation of T-bet may correlate with the strength of vaccine-induced protection (3). In the current study, we examined the role of T-bet directly in the development of innate and adaptive immune responses against F. tularensis LVS. Our findings suggest that T-bet plays an important role in both primary and secondary protection against F. tularensis LVS infection. T-bet-KO mice were more susceptible to primary LVS i.d. and i.n. infection than WT mice, and exhibited greater bacterial burdens in the spleens, livers, and lungs throughout the course of infection. The lungs of the T-bet-KO mice appeared to be the organ affected the most; the lungs exhibited exacerbated recruitment of neutrophils; decreased B cells, T cells, DCs, and NK cells; and greatly increased concentrations of IL-17. Most importantly, LVS-immune lung lymphocytes, but not splenocytes, lacking T-bet were clearly impaired in their capacity to control LVS intramacrophage replication in vitro. These data suggest that the mechanism of reduced protection against respiratory challenge is due to deficits in this critical function in lung lymphocytes (but not splenic lymphocytes) in the absence of T-bet.
T-bet is a member of the T-box family of transcription factors and is expressed by a number of cell types, including T cells, B cells, NK cells, NK T cells, and DCs. Although T-bet is expressed in all these cell types, it is especially important for IFN-γ production in CD4+ T cells and NK cells, but not in CD8+ T cells, in vitro (11). In many circumstances, T-bet-deficient mice typically produce reduced levels of IFN-γ and nitric oxide, as wells as exhibit the establishment of a Th2 environment with respect to cytokine and antibody production. It is well established that IFN-γ plays a crucial role in immunity against F. tularensis, and IFN-γ-KO mice as well as IFN-γ receptor-KO mice do not survive any dose of LVS infection by any route (6, 26, 27). Thus, it is not surprising that T-bet-KO mice were also highly sensitive to LVS infection. However, in contrast to the strict requirement for IFN-γ in mediating protection against LVS, the role of T-bet is separable from IFN-γ production. T-bet-KO mice survived lower doses of LVS infection, and were still able to produce substantial amounts of systemic IFN-γ in vivo (Fig. 3), despite reduced production by T cells in vitro (Fig. 7 and 8). This suggests that T-bet-KO mice have compensatory mechanisms to produce IFN-γ, perhaps from cell types other than T cells, (3), which is a topic for future studies. T-bet has been shown to be dispensable for Th1 responses against Trypanosoma cruzi; however, it is critical in repressing the Th17 response (28). Similar to LVS-infected T-bet-KO mice, T-bet-KO mice infected with T. cruzi produced IFN-γ and IL-17. Infection of T-bet-KO mice with Mycoplasma pulmonis resulted in increased IFN-γ levels in lungs early during M. pulmonis infection compared to the levels in infected WT mice (day 1), and the levels of IFN-γ in lungs were equal to those in the lungs of infected WT mice on days 3 and 7 (20). In addition to T-bet, IFN-γ production is regulated by eomesodermin (Eomes) in both CD8+ and CD4+ T cells (29, 30). Thus, T-bet and Eomes may have redundant roles in regulating IFN-γ transcription. Both Eomes and T-bet control IFN-γ expression and expression of genes encoding granzyme B and CD122 (IL-2 and IL-15 receptor) by CD8+ cells (29). It is therefore possible that IFN-γ production is heavily dependent on Eomes during LVS infection.
Because IFN-γ levels are not substantially different in vivo throughout infection, other mechanisms likely play roles in the increased sensitivity of LVS-infected T-bet-KO mice. T-bet-KO mice have an altered repertoire of immune cells infiltrating the lungs compared to that of WT mice. This includes decreased levels of B cells, CD4+ and CD8+ T cells, NK cells, and DCs. Inappropriate trafficking and/or insufficient expansion of these cell types to the lungs may increase the susceptibility of T-bet-KO mice to LVS infection. In addition, there were increased numbers of neutrophils recruited to the lungs, as well as greater numbers of neutrophils and eosinophils circulating in the blood, of LVS-infected T-bet-KO mice compared to LVS-infected WT mice. Neutrophils, which are key innate immune cells, have been reported to be an important cell type that responds to early Francisella infection. For example, neutrophils comprise approximately half of the Francisella-infected lung cells 3 days after pulmonary infection (42); this recruitment is at least partially dependent on IL-17A (25, 31). Depletion of neutrophils in WT mice results in greatly increased susceptibility to i.d. LVS infection (26, 32, 33), as well as to i.n. LVS infection, of both WT and T-bet-KO mice (see Results). Similar to the F. tularensis life cycle in macrophages, in neutrophils, LVS can hinder NADPH oxidase assembly, escape from the phagosome, and persistence in the cytosol (34, 35).
On the other hand, following LVS pulmonary infection, matrix metalloproteinase 9-KO mice were found to exhibit reduced levels of neutrophils in the lungs, a condition which was associated with a reduction in bacterial burden and increased survival (36). Type 1 interferon receptor-KO mice exhibit increased resistance to F. novicida infection, and resistance was associated with the ability of type I interferons to downregulate the number of IL-17A-positive γ/δ T cells, thus diminishing recruitment of neutrophils to infected spleens (37). Thus, excessive recruitment of neutrophils to the lung may also contribute to Francisella pathogenesis.
T-bet negatively regulates Th17 differentiation, and mice deficient in T-bet or STAT4 often have decreased amounts of Th1 cells and increased numbers of Th17 cells (38–40). Thus, it was not surprising to detect high levels of IL-17 in the LVS-infected T-bet-KO mice following both primary infection and secondary challenge, as well as in cocultures containing LVS-immune T-bet-KO lymphocytes. However, neutralization efforts did not suggest that excess IL-17A was pathogenic in LVS-infected T-bet-KO mice (see Results). IL-17A is produced in LVS-infected mouse lungs by CD4+ T cells, CD8+ T cells, double-negative T cells, and γ/δ T cells (25, 31, 41). IL-17A-deficient mice exhibited a reduction in the proportion of lung neutrophils at early time points (day 4) after LVS infection; however, at later time points, the proportion of neutrophils in IL-17-deficient mice was equivalent to that in WT mice, and IL-17-KO mice exhibited increased bacterial organ burdens by day 7 (25). IL-17A can act together with IFN-γ to inhibit LVS intracellular growth in macrophages following sublethal i.n. LVS infection in WT mice (25); however, in T-bet-KO mice, the number of neutrophils and levels of IL-17 were greatly increased compared to those in WT mice. Because substantial amounts of IFN-γ are still present in LVS-infected T-bet-KO mice, the mechanisms and cells producing IFN-γ may differ between T-bet-KO and WT mice, thus altering the outcome of infection.
In addition to direct impacts on Th1 T cell development and IFN-γ production, T-bet also plays other important roles in both innate and adaptive immunity. Higher expression of T-bet in B cells has been shown to initiate IgG2 class switching (16). B cell expression of T-bet is induced in response to IFN-γ, as well as following stimulation through Toll-like receptor 9 and CD40, and leads to the induction of IgG2a, IgG2b, and IgG3 while repressing IgG1 and IgE. IgG1 and IgE mediate mast cell activation and increased Th2-associated cytokines, such as IL-4, IL-5, and IL-13, which can enhance eosinophil recruitment. Here, LVS-infected T-bet-KO mice displayed qualitatively different antibody responses following infection. Anti-LVS IgG2 antibodies have been shown to be partially important in defense against F. tularensis, particularly against secondary challenge (23). Similar to LVS infections, infection of T-bet-KO mice with M. pulmonis led to increased levels of IgG1 and lower IgG2c levels in the serum compared to those in M. pulmonis-infected WT mice. Interestingly, Salmonella Typhimurium-infected T-bet-KO mice do not develop IgG2 following infection but also do not exhibit class switching or production of IgG1 responses. Similar to M. pulmonis-infected mice, LVS-infected T-bet-KO mice produced essentially no IgG2c but produced high levels of IgG1, suggesting a Th2 environment. The roles of these altered antibody responses during secondary pulmonary challenge remain a subject for future studies. It is also possible that the high levels of anti-LVS antibodies observed after i.d. vaccination compensate for weaker Th1 T cell responses after i.p. challenge but are not useful (or are even detrimental) after i.n. challenge.
Following vaccination by either route, T-bet-KO splenocytes controlled LVS intracellular growth in in vitro cocultures as well as WT splenocytes did. Control of intramacrophage LVS growth by T-bet-KO splenocytes, in the face of poor IFN-γ production, is partially explained by increased levels of TNF-α. Neutralization of TNF-α largely reversed the ability of LVS-immune T-bet-KO splenocytes to control LVS intramacrophage growth, and reversal was associated with reduced nitric oxide production. Although neutralization of TNF-α was incomplete, the levels of nitric oxide were greatly reduced and the control of intramacrophage bacterial growth was lost. This suggests that in LVS-immune splenocytes lacking T-bet, TNF-α production compensated for the poor production of IFN-γ, and nitric oxide was produced in a largely TNF-α-dependent fashion. Also of note, the observation that LVS-immune splenocytes obtained after i.d. or i.n. vaccination are equivalent in controlling intramacrophage bacterial growth does not preclude the possibility that other parameters not measured here are quite different when different vaccination routes are used.
The mechanism of reduced protection against respiratory challenge is likely directly reflected in the greatly reduced ability of T-bet-KO lung lymphocytes to control intramacrophage LVS growth. LVS-immune WT lung lymphocytes controlled intramacrophage growth at similar levels as LVS-immune WT splenocytes; however, LVS-immune T-bet-KO lung lymphocytes were not able to control LVS growth nearly as well as WT lung lymphocytes. LVS-immune T-bet-KO lung lymphocytes produced significantly decreased amounts of IFN-γ, while they produced high levels of IL-5 and IL-17 compared to the levels found in supernatants from cocultures with WT lung lymphocytes. Previous studies showed that CD4+ lung lymphocytes from M. tuberculosis-infected T-bet-KO mice produce increased levels of the immunosuppressive cytokine IL-10 compared to the levels produced by CD4+ lung lymphocytes from M. tuberculosis-infected WT mice (18). The role of IL-10 in LVS-infected T-bet-KO lung lymphocytes is under further investigation.
Most importantly, unlike LVS-immune T-bet-KO splenocytes, LVS-immune T-bet-KO lung lymphocytes were unable to control intramacrophage LVS growth as well as WT lung lymphocytes did. These data suggest the mechanism underlying results from in vivo protection studies, which clearly indicated that T-bet-KO mice are unable to survive higher doses of secondary i.n. challenge. The results further indicate important functional differences in the contribution of T-bet to immune lymphocytes residing in different tissues. Details of the nature of these differential contributions will be important areas for future studies. Collectively, therefore, these results demonstrate T-bet-dependent control of primary Francisella infection by two key routes of infection. Adaptive secondary immunity to respiratory Francisella challenge appears to be particularly dependent on T-bet, further emphasizing its potential value as a useful correlate to predict vaccine-induced protective immunity.
We declare no financial or commercial conflict of interest.
ACKNOWLEDGMENTS
This work was supported in part by an interagency agreement with the National Institute of Allergy and Infectious Diseases (Y1-AI-6153-01/224-06–1322). This project was supported in part by an appointment to the Research Participation Program at the Center for Biologics Evaluation and Research administered by the Oak Ridge Institute for Science and Education through an interagency agreement between the U.S. Department of Energy and the U.S. Food and Drug Administration.
We are grateful to our CBER colleagues Siobhán Cowley, Sherry Kurtz, and Roberto De Pascalis for thoughtful and comprehensive reviews of the manuscript.
FOOTNOTES
- Received 3 December 2013.
- Accepted 8 January 2014.
- Accepted manuscript posted online 13 January 2014.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/IAI.01545-13.
- Copyright © 2014, American Society for Microbiology. All Rights Reserved.
