Despite the recognized role of the T-bet transcription factor in the differentiation of Th1 cells, T-bet-deficient mice can develop small numbers of IFN-γ-producing CD4 T cells. Although these are not sufficient to allow normal handling of some pathogens, T-bet-deficient mice do resolve infection with the intracellular pathogen Listeria monocytogenes. In contrast, we report that expression of T-bet is required for resistance to Salmonella infection. T-bet-deficient mice succumbed to infection with attenuated Salmonella and did not generate IFN-γ-producing CD4 T cells or isotype-switched Salmonella-specific Ab responses. Spleen cells from Salmonella-infected T-bet-deficient mice secreted increased levels of IL-10, but not IL-4, upon in vitro restimulation. A Salmonella-specific TCR transgenic adoptive transfer system was used to further define the involvement of T-bet expression in the development of Salmonella-specific Th1 cells. Wild-type Salmonella-specific CD4 T cells activated in T-bet-deficient recipient mice displayed no defect in clonal expansion, contraction, or IFN-γ production. In contrast, T-bet-deficient, Salmonella-specific CD4 T cells activated in wild-type recipient mice produced less IFN-γ and more IL-2 upon in vivo restimulation. Therefore, expression of T-bet by CD4 T cells is required for the development of Salmonella-specific Th1 cells, regulation of IL-10 production, and resistance to Salmonella infection.
Infection of susceptible mouse strains with Salmonella enterica serovar typhimurium (S. typhimurium) is the best available animal model to study human typhoid fever (1). Resistance to infection with attenuated S. typhimurium strains requires the activation of CD4 T cells and production of IFN-γ (2). It is likely that IFN-γ produced by CD4 T cells is required to directly activate Salmonella-infected macrophages and induce microbial killing (3, 4, 5). Unlike many other intracellular pathogens, Ag-specific B cells are also involved in resistance to Salmonella infection (6, 7, 8, 9), although the exact nature of this contribution is unclear. Isotype-switched B cell responses may mediate protective immunity by opsonizing extracellular bacteria (10), by presenting bacterial Ags to Salmonella-specific T cells (11), or by modulating dendritic cell Ag presentation to T cells (12).
It is known that intracellular pathogens can adapt local microenvironments to avoid or inhibit the microbicidal processes of host cells (13, 14). IFN-γ production by the innate or adaptive immune system can enhance pathogen killing and is often critically required for the resolution of infection with intracellular pathogens (2, 15). T-bet, a member of the T-box family of transcription factors, plays a major role in regulating IFN-γ production in numerous cell types (16, 17, 18, 19, 20, 21). T-bet expression in naive CD4 T cells, induced by both TCR ligation and IFN-γ itself, is essential for the differentiation of Th1 cells in vitro and in vivo (16, 17, 22). Furthermore, ectopic expression of T-bet in Th2 cells is sufficient to repress Th2 cytokine production and induce the expression of IFN-γ in vitro (16). Thus, T-bet plays a vital role in directing the development of IFN-γ-producing cells from naive precursors.
Despite the known importance of IFN-γ to host defense against intracellular pathogens, fewer studies have examined the role of T-bet in infectious disease models. In one study, T cells recovered from T-bet-deficient mice infected with Leishmania major produced less IFN-γ and increased levels of IL-4 and IL-5 upon restimulation (17), reinforcing the notion that T-bet is required for the development of Th1 cells and repression of Th2 cytokine production (16). Furthermore, T-bet-deficient mice developed large chronic lesions following L. major infection, consistent with the well-characterized requirement for Th1 cells to mediate parasite killing in this model (23). We have found a similar requirement for T-bet in handling infection with lymphocytic choriomeningitis virus and Mycobacterium tuberculosis but the former reflected impaired CD8 function (19, 24), and the specific involvement of CD4 cells in the latter has not been established (B. M. Sullivan et al., submitted for publication). Further, recent studies with another intracellular pathogen, Listeria monocytogenes, reported that T-bet-deficient mice were fully capable of resolving bacterial infection and displayed only a marginal impairment in the development of IFN-γ-producing CD4 T cells (25). Therefore, the contribution of T-bet expression to the development of pathogen-specific Th1 cells and host resistance to intracellular infection is unclear, and may depend upon the nature of the individual pathogen. It is possible that T-bet expression is critically required for Th1 development during a chronic parasite infection (17), but is dispensable during acute bacterial infection (25).
We show that expression of T-bet by Salmonella-specific CD4 T cells is necessary for the development of Th1 cells and resistance to Salmonella infection. T-bet-deficient mice were unable to control the growth of attenuated Salmonella and produced much higher levels of IL-10 in response to restimulation with Salmonella Ags. Our data demonstrate a requirement for T-bet in protection from acute bacterial infection, development of Th1 cells, and the repression of IL-10 secretion.
Materials and Methods
T-bet-deficient C57BL/6 mice have been described (17, 26) and were maintained in our facility by intercrossing homozygous pairs. SM1 Rag-deficient TCR transgenic mice expressing the CD90.1 or CD45.1 allele have been described (27). SM1 Rag-deficient, T-bet-deficient, homozygous CD90.1 congenic mice were produced by several rounds of intercrossing T-bet-deficient and SM1-Rag-deficient CD90.1 mice and screening progeny for the desired alleles. SM1 transgene and congenic marker expression were examined by staining peripheral blood with Abs to Vβ2, CD4, CD90.1, CD90.2, CD45.1, and CD45.2. C57BL/6 mice were purchased from the National Cancer Institute (Frederick, MD) and used at 8–16 wk of age. All mice were housed in specific pathogen-free conditions and cared for in accordance with University of Connecticut Health Center and National Institutes of Health guidelines.
Salmonella infection and bacterial colonization in vivo
S. typhimurium AroA−D− (BRD509) were grown overnight in Luria-Bertani broth without shaking, and diluted in PBS after estimation of bacterial concentration using a spectrophotometer. Mice were infected i.v. in the lateral tail vein with 5 × 105 bacteria and then closely monitored for signs of infection. In accordance with institutional animal care guidelines, moribund mice were sacrificed before death. In all infection experiments the dose of bacteria administered was confirmed by plating serial dilutions on MacConkey agar plates. To determine bacterial colonization in vivo, spleens from infected mice were homogenized in PBS and serial dilutions were plated onto MacConkey agar plates. After overnight incubation at 37°C, plates were counted and bacterial numbers calculated for each individual spleen.
Salmonella-specific serum Ab responses
Salmonella-specific Ab responses were determined as previously described (9). An overnight culture of BRD509 was washed twice in PBS and incubated at 65°C for 1 h to prepare heat-killed S. typhimurium (HKST).3 HKST was coated on 96-well microtiter plates (Costar) overnight at 4°C. Blood was recovered from the retro-orbital plexus of infected mice. Serum was prepared and added in serial dilutions to HKST-coated microtiter plates. After extensive washing, Salmonella-specific serum Ab was detected using mouse isotype-specific biotinylated Abs (BD Pharmingen) and ExtrAvidin peroxidase (Sigma-Aldrich).
Salmonella-specific ELISPOT and ELISA
Salmonella-specific IFN-γ-producing cells were quantified using the ELISPOT assay (28). Spleens were harvested from Salmonella-infected and uninfected mice, a single cell suspension was prepared, RBC were lysed, and remaining cells counted. Spleen cells were incubated in 96-well tissue culture plates with added serial dilutions of HKST for 48 h before being recovered and added to 96-well ELISPOT plates (Millipore) precoated with purified anti-IFN-γ (BioSource International) or anti-IL-4 (BD Pharmingen). After 24 h incubation in ELISPOT plates, cells were lysed and cytokine spots visualized using biotinylated anti-IFN-γ or IL-4 (BD Pharmingen), avidin peroxidase (Sigma-Aldrich), and 3-amino-9-ethylcarbazole substrate (Sigma-Aldrich). Cytokine spots were counted using a KS ELISPOT stereo microscope (Carl Zeiss), and the total number of cytokine-producing cells per spleen calculated using the cell dilution and total spleen cell counts. In some experiments supernatants were recovered after 24 h stimulation of spleen cells with HKST and analyzed for the presence of IL-10 using the OptEIA mouse IL-10 kit (BD Pharmingen).
In vivo HKST pulse assay
The HKST pulse assay has been described in detail previously (27). Cytokine production using this assay is Ag-specific and not due to nonspecific inflammatory stimulus. Salmonella-infected or naive mice were left untreated or injected i.v. with 1 × 108 HKST. Spleen cells were harvested 2 h later and rapidly surface stained at 4°C, before being fixed with formaldehyde, permeabilized using saponin (Sigma-Aldrich), and stained intracellularly using cytokine-specific Abs. Cytokine staining was analyzed using a FACSCalibur (BD Biosciences)
Adoptive transfer of SM1 T cells
Spleen and lymph node cells (inguinal, axillary, brachial, cervical, mesenteric, periaortic) were harvested from SM1 Rag-deficient, congenic (or SM1 Rag-deficient, T-bet-deficient, congenic) TCR transgenic mice and a single cell suspension was generated. A small aliquot of these cells was stained with Abs specific for CD4, CD90.1 (or CD45.1), and Vβ2 (BD Pharmingen), and the percentage of SM1 cells determined by flow cytometry. Before adoptive transfer, transgenic SM1 T cells were stained with the dye CFSE (29) by incubation for 10 min at 37°C. Following CFSE staining, 2 × 106 transgenic SM1 T cells were injected i.v. into the lateral tail vein of recipient mice.
Flow cytometric analysis
Spleen cells were incubated for 20–45 min at 4°C in Fc block (spent culture supernatant from the 24G2 hybridoma, 2% rat serum, 2% mouse serum, and 0.01% sodium azide) in the presence of different primary Abs. FITC-, PE-, CyChrome-, PE-Cy5-, or allophycocyanin-conjugated Abs specific for CD4, CD11a, CD45.1, CD69, CD90.1, IFN-γ, IL-2, IL-4, and TNF-α were purchased from eBioscience and BD Pharmingen. After staining, cells were fixed and analyzed by flow cytometry using a FACSCalibur. Data were analyzed using FlowJo software (Tree Star).
Tracking SM1 T cells and recalling cytokine production in vivo
At various times after Salmonella infection, spleen cells were harvested into Eagle’s Hank’s Amino Acids (EHAA) medium (Biofluids) containing 2% FBS and 5 mM EDTA. Cells (5 × 106/tube) were stained as described and the percentage of SM1 T cells per spleen was determined by flow cytometry. The total number of SM1 T cells per spleen was calculated by multiplying the percentage of SM1 cells by the total cell count for each spleen. For recall cytokine assays, adoptively transferred mice were immunized i.v. with 1 × 108 HKST and at various time points later injected i.v. with 200 μg of flagellin peptide 427–441 (VQNRFNSAITNLGNT) (30), plus 25 μg of LPS (Sigma-Aldrich). To examine cytokine production directly, spleens from injected mice were harvested 6 h after peptide/LPS injection. These spleen cells were immediately surface stained, fixed with formaldehyde, permeabilized using saponin (Sigma-Aldrich), and stained intracellularly using the cytokine Abs. Recall cytokine production therefore occurs in vivo and involves no ex vivo stimulation, as previously described (31).
T-bet-deficient mice fail to resolve infection with S. typhimurium
To examine the role of T-bet in the clearance of an acute bacterial infection, we infected wild-type (Wt) or T-bet-deficient (T-bet−/−) mice i.v. with 5 × 105 attenuated S. typhimurium and monitored infected mice for signs of disease. As expected using this attenuated strain (BRD509) (32), most Wt mice displayed only transient evidence of bacterial infection and survived for months afterward. In contrast, all Salmonella-infected T-bet−/− mice developed severe disease by 2 wk after administration of bacteria, eventually causing a moribund state that required euthanasia (Fig. 1 A). The kinetics of the disease process in T-bet−/− mice was strikingly reminiscent of Salmonella infection of class II-deficient, CD28-deficient, and IFN-γR-deficient mice (2, 9), suggesting T-bet-deficient mice have a major defect in the development of CD4 T cell adaptive immune responses.
Next we assessed the number of viable S. typhimurium in the spleens of Wt and T-bet−/− mice 2 and 3 wk after Salmonella infection. As expected, the spleens of Wt mice contained large bacterial burdens 2 wk after infection that subsequently declined over the following week (Fig. 1,B). In marked contrast, T-bet-deficient spleens contained slightly elevated bacterial burdens 2 wk postinfection that increased markedly by 3 wk (Fig. 1 B). Thus, in the absence of T-bet, mice are unable to control Salmonella replication in vivo and ultimately succumb to an infection that is resolved in Wt mice.
Mice lacking T-bet develop impaired T cell and B cell responses to S. typhimurium
To understand the susceptibility of T-bet-deficient mice to infection with attenuated Salmonella, we first examined the development of Salmonella-specific serum Ab responses after infection. As noted in previous studies (9), Wt mice developed elevated titers of Salmonella-specific IgM and IgG2a 2 wk after Salmonella infection (Fig. 2,A). In contrast, although T-bet−/− mice developed equivalent titers of Salmonella-specific IgM, they failed to develop a detectable Salmonella-specific IgG2a titer (Fig. 2,A). As noted previously, T-bet−/− mice display elevated levels of IgG1 and diminished levels of IgG2a in response to immunization (17). However, neither Wt or T-bet−/− mice developed significant titers of Salmonella-specific IgG1 or IgG2b after infection with attenuated Salmonella (Fig. 2 A and data not shown). Thus, T-bet-deficient mice have a profound deficiency in the development of Salmonella-specific IgG2a as might have been predicted, but surprisingly do not develop a compensatory IgG1 response.
Next we examined the development of Salmonella-specific Th1 and Th2 responses in Wt and T-bet−/− mice. Two different assays were used to examine the frequency of Salmonella-specific T cells that can produce IFN-γ in response to infection. Using an ELISPOT assay, the spleens of Wt mice were found to contain an elevated frequency of IFN-γ-producing cells that responded to HKST stimulation in vitro (Fig. 2,B). The frequency of IFN-γ-producing cells increased markedly between 2 and 3 wk after infection. We also examined Salmonella-specific CD4 T cell responses directly in vivo, using a simple in vivo recall assay that we previously described (27). CD4 T cells from Salmonella-infected Wt mice spontaneously secreted IFN-γ without restimulation (Fig. 3,A, Infected). As previously described (27), this response is amplified following a short 2-h i.v. pulse with HKST (Fig. 3,A, Infected plus HKST). All of this induced IFN-γ production comes from CD4 T cells that express high levels of CD11a, an indicator of previous Ag exposure (Fig. 3 B). Together these data are consistent with previous reports on the development of IFN-γ-producing CD4 T cells in Salmonella-infected mice (2, 30).
In contrast, using the ELISPOT assay, very few IFN-γ-producing cells were detected in the spleens of T-bet−/− mice at any time point following Salmonella infection (Fig. 2,B). In agreement with these data, CD4 T cells from Salmonella-infected T-bet−/− mice failed to produce detectable levels of spontaneous IFN-γ (Fig. 3). Furthermore, amplification of IFN-γ production by CD4 T cells was not detected in T-bet−/− mice in response to a short in vivo pulse with HKST (Fig. 3).
T-bet deficiency has been reported to cause a shift toward the production of Th2 cytokines by effector T cell populations under certain circumstances (17, 22). However, Salmonella-specific IL-4-producing cells were not detected in the spleens of Salmonella-infected T-bet-deficient mice using either the ELISPOT or in vivo HKST pulse assay (data not shown). Instead, as was observed in the setting of experimental autoimmune encephalomyelitis (EAE) (22), elevated IL-10 production was clearly detected in the supernatants of HKST-stimulated splenocytes recovered from T-bet−/−, but not Wt mice (Fig. 2 C). Therefore, T-bet deficiency prevents the development of CD4 T cells that secrete IFN-γ, enhances Salmonella-specific IL-10 production, but does not increase the frequency of IL-4-producing Th2 cells.
T-bet deficiency in non-CD4 T cells does not affect Salmonella-specific T cells
We examined in more detail the cellular source of T-bet that modulates the development of Salmonella-specific Th1 cells. T-bet expression regulates the production of IFN-γ in CD4 T cells (16), CD8 T cells (19), γδ T cells (18), dendritic cells (20), and NK cells (21). Furthermore, T-bet also regulates the development of cytotoxic effector responses (33) and isotype switching in B cells (34). Therefore, T-bet expression by cells other than CD4 T cells could influence the generation of IFN-γ-producing CD4 T cells and affect protective immunity to Salmonella infection. Of particular interest is the observation that T-bet regulated IFN-γ production by dendritic cells can contribute to the development of Th1 responses in vivo (20). We therefore examined the development of Wt Salmonella-specific Th1 cells that had been activated in a T-bet-deficient environment.
TCR transgenic Salmonella-specific SM1 T cells were adoptively transferred into Wt or T-bet-deficient recipients, and mice were infected the following day with attenuated Salmonella. After adoptive transfer, SM1 T cells could be found in the spleen and lymph nodes of Wt and T-bet−/− mice at similar frequencies indicating that there are no defects in T cell circulation in T-bet−/− recipients (Fig. 4,A, Transfer Only). Three days after Salmonella infection, SM1 T cells had expanded to a similar extent in the spleens of both Wt and T-bet−/− mice (Fig. 4,A, Day 3). SM1 T cells in both recipient strains had also undergone a similar number of cell divisions, as evidenced by CFSE dye dilution (Fig. 4,A, Day 3). These data indicate that T-bet deficiency in non-CD4 T cells does not affect the peak clonal expansion of Salmonella-specific T cells. However, it remained possible that T-bet deficiency affected T cell proliferation or survival at other time points. Therefore, in additional experiments we examined the full kinetics of SM1 T cell activation. However, no difference in the kinetics of SM1 expansion or contraction was noted in T-bet-deficient vs Wt recipients (Fig. 4, B and C). These data suggest that T-bet deficiency in non-CD4 T cells does not influence the initial activation of Salmonella-specific T cells during Salmonella infection.
Next we examined the development of IFN-γ-producing CD4 T cells in Wt or T-bet−/− recipients. SM1 T cells were adoptively transferred into both recipient strains and mice were then immunized i.v. with HKST the following day. We chose to immunize mice with HKST for this experiment because infection with attenuated Salmonella led to death of some mice within 2–3 wk (Fig. 1) and therefore complicated any analysis of effector cytokine production in infected mice. Furthermore, HKST immunization is actually the best immunization strategy in this model for the development of Salmonella-specific effector T cells and results in the generation of a large number of Ag-experienced SM1 T cells (27). To examine the recall response of SM1 T cells, some groups of previously adoptively transferred and HKST-immunized mice were injected i.v. with flagellin peptide 427–441 and LPS 20 days later. SM1 T cells, in both Wt or T-bet−/− recipients, expressed low levels of CD69 and did not produce significant amounts of IFN-γ. As expected, injection of peptide resulted in rapid activation of almost all SM1 T cells in the spleen of both recipients, as evidenced by increased expression of CD69 (Fig. 4,D). Furthermore, a significant proportion of SM1 T cells in both Wt and T-bet−/− recipients produced intracellular IFN-γ in response to peptide stimulation (Fig. 4 E). Naive SM1 T cells do not produce detectable levels of IFN-γ using this assay (27). Therefore, T-bet deficiency in non-CD4 T cells had no effect upon the development of Salmonella-specific IFN-γ-producing CD4 cells.
T-bet deficiency in Salmonella-specific CD4 T cells impairs Th1 development
We also studied the effect of intrinsic T-bet deficiency on the development of Salmonella-specific CD4 T cells. To do this, we generated T-bet-deficient, Rag-deficient, SM1 TCR transgenic mice that express the congenic marker CD90.1. We previously generated Wt, Rag-deficient, SM1 TCR transgenic mice expressing the congenic marker CD45.1 (35). Therefore, by adoptively transferring a similar number of both SM1 populations, we were able to visualize the activation of T-bet−/− and Wt SM1 T cells in the same C57BL/6 recipient.
Following adoptive transfer, T-bet−/− (CD90.1) and Wt (CD45.1) SM1 T cells were found at similar frequencies in the spleen and lymph nodes of recipient mice (Fig. 5 and data not shown). Thus, T-bet deficiency confers no selective advantage or disadvantage in the circulation of naive CD4 T cells through lymphoid tissues. Three days after Salmonella infection, both populations of SM1 T cells had undergone a similar amount of CFSE dye dilution and clonal expansion in vivo (Fig. 5,A). Furthermore, no difference was noted in the expansion and contraction kinetics of T-bet−/− and Wt SM1 T cells at any time point (Fig. 5 B). Thus, T-bet deficiency in CD4 T cells does not influence the proliferative or contraction program that occurs in response to Salmonella infection. In additional experiments, the adoptive transfer of T-bet−/− or Wt SM1 T cells into separate recipients produced similar results (data not shown).
As there was no defect in the proliferative program of T-bet-deficient SM1 T cells, we examined the capacity of these cells to develop effector cytokine production. We used the same immunization regimen described earlier. C57BL/6 mice were first adoptively transferred with a similar number of T-bet-deficient (CD90.1) and Wt (CD45.1) SM1 T cells and immunized i.v. the following day with HKST. Twenty days after HKST immunization, an elevated frequency of T-bet and Wt SM1 T cells was clearly detected in the spleen of immunized mice, compared with unimmunized mice (Fig. 6 A). Therefore, T-bet deficiency does not confer any selective advantage or disadvantage in the generation of an elevated number of CD4 memory cells.
To examine effector cytokine production directly in vivo, mice immunized 14 days previously with HKST were injected i.v. with flagellin peptide 427–441 and LPS, and spleen cells were harvested 6 h later. After surface staining to identify CD45.1 and CD90.1 SM1 populations, cytokine production was analyzed by intracellular staining. No cytokine production was detected by endogenous CD4 cells in response to peptide injection (data not shown). Peptide activated almost all T-bet−/− and Wt SM1 T cells in vivo, as evident by increased expression of CD69 in response to stimulation (Fig. 6,B). Thus, both Wt and T-bet−/− cells remain responsive to recall activation when encountering peptide in vivo. A sizable percentage of Wt SM1 T cells produced IFN-γ and TNF, but not IL-2, IL-4, or IL-6 in response to recall activation (Fig. 6,B and data not shown). This pattern of cytokine production is markedly different from naive SM1 T cells that predominantly secrete IL-2 (36). Therefore, Wt SM1 T cells have acquired effector functions after HKST immunization. T-bet−/− SM1 T cells produced IL-2 and TNF-α in response to peptide stimulation (Fig. 6,B). As expected, based upon previous reports (17), the percentage of T-bet−/− SM1 T cells producing IL-2 was much higher than that observed for Wt SM1 T cells. In addition, very few IFN-γ-producing, T-bet−/− SM1 T cells were detected after peptide stimulation (Fig. 6 B). Therefore, intrinsic T-bet deficiency in Salmonella-specific CD4 T cells inhibited the development of IFN-γ production and enhanced IL-2 production.
Few studies have examined the role played by the T-bet transcription factor in models of microbial infection. We report that T-bet-deficient mice are unable to control bacterial replication following infection with attenuated Salmonella. These data are in agreement with the previously described requirement for T-bet expression in resistance to chronic Leishmania infection (17), but contrast with data describing effective clearance of Listeria infection in T-bet-deficient mice (25). Our data extend previous work on the critical importance of T-bet in resistance to microbial infections and highlight the particular importance of T-bet expression during an acute bacterial infection. It seems unlikely that the contribution of T-bet expression to Th1 development is more important during a chronic vs an acute microbial infection.
There are a number of possible explanations for the differential susceptibility of T-bet-deficient mice to Listeria vs Salmonella infection. One possibility is that the background strain accounts for this difference. Our mice have been backcrossed to a C57BL/6 background whereas BALB/c or F1 (BALB/c × C57BL/6) were used in the Listeria study. It is possible therefore that genetic differences in the background strain amplify or diminish the effect of T-bet deficiency. Future studies using T-bet-deficient mice backcrossed to different inbred strains will be able to examine this issue directly.
An alternative possibility is that susceptibility to Salmonella infection in T-bet-deficient mice reflects the greater contribution of Ab to protective immunity in this particular model. T-bet-deficient mice display a profound defect in the production of Salmonella-specific IgG2a (Fig. 2), and Ab is known to be required for protective immunity to Salmonella (6, 7, 8, 9). In contrast Ab responses are thought to play a minor role in resistance to Listeria infection (37). However, despite the known contribution of Ab to protective immunity, B cell-deficient mice resolve infection with attenuated Salmonella (7, 8, 9). Therefore, it seems unlikely that a deficiency in switched Ab production can account for the increased susceptibility of T-bet-deficient mice to infection with attenuated Salmonella.
A more likely possibility is that the susceptibility of T-bet-deficient mice to Salmonella infection is due to a deficiency in T cell responses. Although both Listeria and Salmonella cause intracellular infections that require IFN-γ for clearance (2, 38, 39), the cellular source of IFN-γ production is likely to derive from different cell subsets in each case. During Salmonella infection, CD4 T cells undergo massive expansion, acquire effector functions, and are known to be critical for bacterial clearance (2, 6, 27, 40, 41). In contrast, both MHC class Ia- and class Ib-restricted CD8 T cells are more important for the eradication of Listeria infection (42, 43). CD8 T cells appear to display less dependence on T-bet expression for effector cytokine development (17, 44). Therefore it seems most likely that transcription factors other than T-bet can contribute to protective immunity during Listeria infection. An interesting possibility is that T-bet and Eomes, another T-box family member, serve somewhat redundant functions in the setting of Listeria infection (44). However, it should be noted that a difference in requirement for CD4 vs CD8 T cells does not explain why Th1 CD4 T cells can develop during Listeria infection but are absent during Salmonella infection (25). It therefore remains possible that intrinsic differences in T-bet dependency for Th1 development during Gram-negative and Gram-positive infections account for the contrasting susceptibility of T-bet-deficient mice to these two pathogens.
We found no effect of T-bet deficiency on the capacity of CD4 T cells to proliferate in response to Salmonella infection. This was true when T-bet deficiency was confined to CD4 T cells themselves or to non-CD4 T cells. Although we did note a slight trend for T-bet−/− SM1 T cells to reach a higher peak clonal expansion than Wt SM1 T cells, this difference never approached statistical significance (data not shown). This result contrasts with a recent report using the 2D2 TCR transgenic mouse on a T-bet−/− background, in which T-bet deficiency conferred greater proliferative potential on TCR transgenic T cells (22). It is possible that disparities in the affinity of the 2D2 and SM1 TCR for peptide MHC account for this difference in behavior of clonal CD4 populations in response to Ag.
In agreement with previous reports examining CD4 T cell responses (17, 45), IFN-γ production by Salmonella-specific CD4 T cells was severely inhibited in T-bet-deficient mice. In addition, we noted an increase in IL-2 production in T-bet-deficient SM1 effector cells. These data agree with the previously documented role of T-bet in driving IFN-γ production and suppressing IL-2 (16). Previous reports have also suggested that T-bet expression represses the production of Th2 cytokines (16, 22). Indeed, T-bet-deficient mice have a propensity to spontaneously develop airway disease resembling human asthma, tend to produce Th2 responses upon immunization, and resist the development of EAE (17, 22, 26). It was therefore surprising that we did not detect the development of Salmonella-specific IL-4 production, either in infected T-bet-deficient mice or in adoptively transferred TCR transgenic T-bet-deficient effector cells responding to HKST. It seems likely that even in the absence of IFN-γ, the inflammatory conditions during Salmonella infection are not favorable for the differentiation of classical Th2 cells.
In agreement with a recent study using T-bet-deficient mice in an EAE model, we noted enhanced production of IL-10 in spleen cells recovered from T-bet-deficient mice (22). The initial report describing T-bet noted that the transduction of myelin basic protein-specific TCR transgenic T cells with a T-bet expression vector did not affect IL-10 production (16). However, the in vitro conditions were probably not optimal for the development of IL-10-producing T cells and may not have revealed a role for T-bet in IL-10 repression. From our data and others, it seems likely that T-bet represses IL-10 production in developing CD4 T cells, but this remains to be confirmed. It is possible that increased IL-10 production by Salmonella-specific CD4 T cells contributes to the susceptibility of T-bet-deficient mice by inhibiting the microbicidal activity of macrophages that would otherwise be activated by inflammatory cytokines such TNF-α. Experiments are currently underway to examine the contribution of IFN-γ deficiency vs excess IL-10 production in T-bet-deficient mice in the development of susceptibility to Salmonella infection.
We used a Salmonella-specific TCR transgenic system that allowed us to examine the role of intrinsic vs extrinsic T-bet expression in the development of naive CD4 T cells into Th1 cells. These data show that T-bet deficiency in non-CD4 T cells was insufficient to affect the development of IFN-γ-producing CD4 T cells. Previous work demonstrated that IFN-γ production by dendritic cells is regulated by T-bet expression and is required for the development of Ag-specific Th1 cells in vivo. It appears from our data that this pathway is not a major contributor to the development of Salmonella-specific Th1 cells. However, it remains possible that the expression of T-bet deficiency by non-CD4 T cells serves to influence the development of CD4 IFN-γ production in other model systems, and may also affect other aspects of CD4 memory. Deficiency in T-bet expression by CD4 T cells themselves was sufficient to inhibit IFN-γ production and increase IL-2 production. These data are in agreement with a CD4 T cell intrinsic role for T-bet in repressing IL-2 production and enhancing IFN-γ production.
In conclusion, our data demonstrate a obligatory role for the expression of T-bet during the differentiation of Salmonella-specific Th1 cells in response to infection, and highlight the importance of this transcription factor in the development of protective immunity.
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.
This work was supported by National Institutes of Health Grants AI056172 and AI055743 (to S.J.M.) and AI56296 (to L.H.G.).
Abbreviations used in this paper: HKST, heat-killed Salmonella typhimurium; Wt, wild type; EAE, experimental autoimmune encephalomyelitis.