ABSTRACT
Interleukin-12 (IL-12) and IL-18 are both central to the induction of gamma interferon (IFN-γ), and various roles for IL-12 and IL-18 in control of intracellular microbial infections have been demonstrated. We used IL-12p40−/− and IL-18−/− mice to further investigate the role of IL-12 and IL-18 in control of Salmonella enterica serovar Typhimurium. While C57BL/6 and IL-18−/− mice were able to resolve attenuated S. enterica serovar Typhimurium infections, the IL-12p40−/− mice succumbed to a high bacterial burden after 60 days. Using ovalbumin (OVA)-specific T-cell receptor transgenic T cells (OT-II cells), we demonstrated that following oral infection with recombinant S. enterica serovar Typhimurium expressing OVA, the OT-II cells proliferated in the mesenteric lymph nodes of C57BL/6 and IL-18−/− mice but not in IL-12p40−/− mice. In addition, we demonstrated by flow cytometry that equivalent or increased numbers of T cells produced IFN-γ in IL-12p40−/− mice compared with the numbers of T cells that produced IFN-γ in C57BL/6 and IL-18−/− mice. Finally, we demonstrated that removal of macrophages from S. enterica serovar Typhimurium-infected C57BL/6 and IL-12p40−/− mice did not affect the bacterial load, suggesting that impaired control of S. enterica serovar Typhimurium infection in the absence of IL-12p40 is not due to reduced macrophage bactericidal activities, while IL-18−/− mice did rely on the presence of macrophages for control of the infection. Our results suggest that IL-12p40, but not IL-18, is critical to resolution of infections with attenuated S. enterica serovar Typhimurium and that especially the effects of IL-12p40 on proliferative responses of CD4+ T cells, but not the ability of these cells to produce IFN-γ, are important in the resolution of infection by this intracellular bacterial pathogen.
The induction of a gamma interferon (IFN-γ) response during infection with phagosome- or endosome-bound intracellular bacterial pathogens is considered pivotal in the resolution of such an infection (13, 29, 30). Antigen-specific CD4+ T cells are considered the most important source of this cytokine (13, 15), although natural killer (NK) cells may be the major producers early after infection when innate immune responses are activated (39). Production of IFN-γ and other cytokines by CD4+ T cells can be induced by antigen-independent signals, such as CD80/86, interleukin-12 (IL-12), IL-18, IL-6, IL-4, or IFN-γ (20, 31, 32, 41, 49), as well as through T-cell receptor (TCR) ligation by peptide-major histocompatibility complex class II (MHC-II) complexes (5). The cytokines IL-12 and IL-18 have been shown to act synergistically to induce IFN-γ production by T cells (32, 49), where IL-12 upregulates IL-18 receptor expression on IFN-γ-producing cells. However, IL-12 alone has the potential to drive Th1 cell differentiation (19), with IL-18 serving as a costimulatory signal for IFN-γ production in this setting (42). Although T cells are considered the primary target of IL-12 and IL-18, NK cells also respond to these cytokines in vivo (32, 49).
Bioactive IL-12 is a heterodimeric cytokine composed of p35 and p40 subunits and is produced mainly by professional antigen-presenting cells (APCs) (phagocytes and dendritic cells) in response to microbial stimulation (14, 18, 49). The p40 and p35 subunits combine in the endoplasmic reticulum to form bioactive IL-12p75, identified as the key factor in the induction of Th1 immune responses (11). IL-18 shares structural similarities with IL-1 and functional similarities with IL-12 (8, 33). IL-18 is inactive when it is initially synthesized and requires cleavage by caspase-1 to become active (8, 38). Unlike IL-12, which is produced only by cells of the immune system, IL-18 is also produced by keratinocytes (46), osteoblasts (51), pituitary gland and adrenocortical cells (6), and intestinal epithelial cells (48).
Salmonella enterica serovar Typhimurium is a facultative, intracellular bacterium that resides and replicates within macrophages of the reticuloendothelial system (3). Protection against S. enterica serovar Typhimurium infection requires CD4+ IFN-γ-producing T cells (13, 29). The role of IL-12 and IL-18 in control of Salmonella infections has been examined in natural human infections and in experimental infections in animal models. Humans who do not express the IL-12 receptor β chain or IL-12p40 are susceptible to serious infection by Salmonella serovars that are not usually associated with serious disease (53). Intravenous infection of innately susceptible BALB/c mice with a highly attenuated strain of S. enterica serovar Typhimurium, where IL-12 was neutralized by administration of an antibody, resulted in lethal infection, as well as decreased IFN-γ and inducible nitric oxide synthase production, decreased expression of MHC-II, and an increase in the production of the Th2 cytokine IL-10 (28). Neutralization of IL-12 in immunized mice at the time of challenge also abrogated clearance of a challenge infection (27). More recently, it was shown that in S. enterica serovar Typhimurium-infected IL-12p40−/− mice, IFN-γ levels were reduced and the bacterial load was increased, resulting in higher mortality rates (22). Other studies showed that IL-12p40−/− mice were unable to sustain an IFN-γ response, which was attributed to a reduction in MHC-II-mediated antigen presentation. These studies suggested that IL-12-independent pathways are also involved in activation of IFN-γ producing T cells (15).
The contribution of IL-18 to host resistance against Salmonella is generally considered secondary compared to the contribution of IL-12 (36). Neutralization of IL-18 reduces the survival time of BALB/c mice by 2 days after oral inoculation with wild-type S. enterica serovar Typhimurium and slightly reduces the production of IFN-γ in response to immunization with an attenuated strain (9). Neutralization of IL-18 in IL-12-deficient mice did not accelerate death from attenuated salmonellae, implying that IL-18 was acting through IL-12 and did not have further differentiated effector functions (9, 36). In innately resistant mice, IL-18 appears to play a more important role than it plays in susceptible mice, and it is crucial for the generation of IFN-γ responses in vivo (26). Recently, it was demonstrated that IL-18−/− mice were more susceptible to systemic infection with virulent S. enterica serovar Typhimurium (40). Collectively, these studies suggest that IL-18 could be important for the resolution of Salmonella infections.
We further investigated the immune response to infection with Δaro attenuated S. enterica serovar Typhimurium in mice lacking the IL-12p40 subunit or IL-18. We used transgenic antigen-specific T cells to investigate early antigen presentation in vivo and examined ex vivo production of IFN-γ by T cells during infection, as well as the role of macrophages in control of the infection.
MATERIALS AND METHODS
Mice.C57BL/6, IL-12p40−/− (referred to as IL-12−/− below) (24), IL-18−/− (47), and OT-II mice (1), all with a C57BL/6 background, were bred and maintained at the Animal Facility of The Department of Microbiology and Immunology, The University of Melbourne. Mice were age matched and used when they were 6 to 12 weeks of age. All animal experiments were approved by The University of Melbourne Animal Experimentation Ethics Committee.
Bacteria and plasmids.In this study we used an aroAD mutant of S. enterica serovar Typhimurium SL1344 (BRD509), a gift from G. Dougan (Sanger Institute, Cambridge, United Kingdom). The pTETtac4 expression plasmid (10) was used to generate plasmids pG and pGO by replacing the C-frag portion of pTETtac4 with green fluorescent protein (GFP) (pG) or with GFP fused to ovalbumin (OVA) epitopes OVA257-264 and OVA326-339 (pGO). BRD509 was transformed with pG or pGO by electroporation. For immunization of mice, bacteria were cultured for 24 h at 37°C in 500 ml Luria-Bertani (LB) broth to stationary phase, pelleted by centrifugation, and resuspended in 1 ml phosphate-buffered saline (PBS). Plasmids pTETtac4, pG, and pGO are stably maintained by BRD509 during growth in mice (10, 56, 57).
Immunization of mice.Mice were inoculated with approximately 1010 CFU S. enterica serovar Typhimurium by oral gavage under inhalation anesthetic (Penthrane; Abbot Laboratories, North Chicago, IL). Ten minutes before inoculation of S. enterica serovar Typhimurium, the stomach acid was neutralized by feeding the animals 0.1 ml of sodium bicarbonate (10%, wt/vol).
Viable counting of S. enterica serovar Typhimurium in organs.Organs were removed aseptically from mice, and tissue homogenates were prepared as described previously (56). Peyer's patches (PPs) were washed in PBS and incubated in PBS containing 100 μg/ml gentamicin for 1.5 h at 37°C to kill any extracellular bacteria. PPs were then washed again in PBS and homogenized in 5 ml PBS. The number of bacteria in each organ was determined by plating serial dilutions of tissue homogenates on LB agar plates containing the appropriate antibiotics. When applicable, each sample was plated on LB agar containing either streptomycin or streptomycin and ampicillin to determine whether plasmids pG and pGO were retained.
Antibodies.The following purified and/or fluorescently conjugated antibodies were purchased from BD Pharmingen: anti-IFN-γ-phycoerythrin (PE) (clone XMG1.2), anti-CD4-CyChrome (clone H129.19), anti-CD8-APC (clone 53-6.7), anti-T-cell receptor β-fluorescein isothiocyanate (FITC), anti-CD19-PE (clone 1D3), anti-I-Ab-PE (clone AF6-120.1), anti-CD3-FITC (clone 145-2C11), and purified anti-CD3 (clone 145-2C11). Carboxyfluorescein ester (CFSE) was purchased from Molecular Probes (Eugene, OR).
Detection of intracellular IFN-γ by flow cytometry.Spleens and mesenteric lymph nodes (MLNs) were removed from mice at different time points, and single-cell suspensions were prepared. Splenocytes (1 × 106 cells) were seeded in individual wells of a U-bottom 96-well plate that was coated with anti-CD3 monoclonal antibody (MAb). Cells were cultured for 5 h at 37°C and 6% CO2 in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, penicillin, streptomycin, glutamine, and 2-mercaptoethanol (DMEM-10) in the presence of 10 μg/ml brefeldin A (Sigma). Cells were then washed in PBS-brefeldin A and incubated with anti-CD4-CyChrome, anti-CD8-APC, and anti-CD3-FITC for 30 min on ice, washed again, and fixed with a 2% formaldehyde solution (Sigma). After washing to remove the fixative, the cells were permeabilized with PBS-0.5% (wt/vol) saponin (Sigma) and incubated with anti-IFN-γ-PE or a rat immunoglobulin G-PE control at room temperature for 30 min. After the final wash, binding of antibodies to cells was analyzed with a FACSCaliber using CellQuest software (Becton Dickinson, San Diego, CA).
Adoptive transfer of OT-II T cells.Peripheral lymph nodes and MLNs from naïve OT-II mice were collected in Hanks balance salt solution and teased apart using fine forceps. The resultant suspension was gently passed through a metal sieve, washed with Hanks balance salt solution-0.1% bovine serum albumin, and labeled with CFSE (Molecular Probes) as described previously (34). CFSE-labeled cells (5 × 106 cells) were transferred to sex-matched naïve C57BL/6 mice by tail vein injection. Approximately 50% of the transferred cells were CD4+ Vα2+ cells (OT-II T cells). One day after adoptive transfer, mice were immunized with S. enterica serovar Typhimurium as described above. At different time points after immunization, spleens, MLNs, and PPs were removed, and single-cell suspensions were prepared. Cells were labeled with anti-CD4-CyChrome antibody. CFSE fluorescence of adoptively transferred CD4+ T cells was detected using a FACSCaliber, and data were analyzed using CellQuest software. Dead cells were excluded from analysis using the propidium iodide exclusion method.
Bactericidal activity of macrophages.The peritoneal cavity of mice was washed with 10 ml ice-cold PBS, and the cells were collected by centrifugation. Cells were seeded in 24-well plates at a concentration of 1 × 106 cells/well in DMEM-10 and cultured overnight at 37°C and 5% CO2. Nonadherent cells were removed, and the adherent macrophages were infected with S. enterica serovar Typhimurium BRD509 at a multiplicity of infection of 10. After 1 h, the medium was replaced with DMEM-10 containing 40 μg/ml gentamicin to kill any extracellular bacteria. This medium was replaced with fresh DMEM-10 containing gentamicin 1.5 h later, and 1 and 4 h after that cells were lysed with 0.1% Triton X-100 and the numbers of bacteria present in the lysates were determined by viable counting.
In vivo macrophage depletion.Macrophages were depleted from mice by treatment with dichloromethylene diphosphonate (Cl2MDP)-loaded liposomes as described in detail previously (54, 56, 57). Cl2MDP was a kind gift from Roche Diagnostics GmbH (Mannheim, Germany). Briefly, mice were treated with 0.2 ml Cl2MDP liposomes 2 days before immunization to eliminate macrophages from the spleen and liver, and they subsequently received 50 μl Cl2MDP liposomes intravenously every 5 days for the duration of the experiment to remove newly immigrating macrophages.
Statistical analysis.Data were analyzed using Student's t test and were considered significantly different when the P value was <0.05.
RESULTS
Kinetics of ΔaroAD S. enterica serovar Typhimurium infection in IL-12- and IL-18-deficient mice.Groups of five mice were orally infected with 1010 CFU S. enterica serovar Typhimurium BRD509, and the bacterial load was measured over time by determining the viable counts in organ homogenates. Figure 1 shows that both C57BL/6 and IL-18−/− mice controlled the infection. C57BL/6 and IL-18−/− mice carried similar bacterial loads in the spleen and PPs throughout the infection, and the maximal loads occurred 2 to 3 weeks after infection (∼104 CFU in the spleen and ∼103 CFU in PPs). In contrast, IL-12−/− mice were unable to control the infection and succumbed 60 days after inoculation. In IL-12−/− mice, the bacterial load in the PPs steadily increased to 2.4 × 107 ± 1.8 × 107 CFU, and in the spleen the bacterial numbers reached 6 × 109 ± 3.5 × 109 CFU. In IL-12−/− mice, but not in C57BL/6 mice and IL-18−/− mice, infection with S. enterica serovar Typhimurium resulted in severe splenomegaly, and the spleen weights were up to 1.33 ± 0.4 g (n = 5) at 60 days after infection (data not shown). Flow cytometric analyses showed that most cells infiltrating the spleen of infected IL-12−/− mice were large granular cells (data not shown).
IL-12−/− mice cannot control infection with S. enterica serovar Typhimurium. C57BL/6 (open columns), IL-12−/− (filled columns), and IL-18−/− (cross-hatched columns) mice were orally inoculated with 1010 CFU S. enterica serovar Typhimurium ΔaroAD strain BRD509, and at the indicated time points animals were euthanized and the bacterial loads in the PPs (A) and spleens (B) were determined by viable counting. The data are the means and standard deviations for groups of five animals in one representative experiment out of three experiments performed.
In vivo proliferation of antigen-specific CD4+ T cells.Clearance of S. enterica serovar Typhimurium from infected mice is thought to depend on the activation of IFN-γ-producing CD4+ T cells (13, 29). In order to determine whether antigen-specific proliferation of CD4+ T cells was impaired in IL-12−/− mice, an in vivo proliferation assay was used (23). Mice were adoptively transferred with CFSE-labeled OVA-specific CD4+ OT-II T cells, and 1 day later they were infected with an S. enterica serovar Typhimurium BRD509 recombinant expressing GFP-OVA (BRD509/pGO) or GFP as a control antigen (BRD509/pG) or were fed 20 mg OVA. Three days after transfer of OT-II cells, flow cytometry was used to determine the proliferation of OT-II cells, as shown by loss of CFSE. Figure 2A shows that in the MLNs of C57BL/6 and IL-18−/− mice, OT-II cells had undergone approximately four divisions. No proliferation of OT-II cells was observed in IL-12−/− mice. To further analyze the proliferation of OT-II cells in vivo, a proliferation index was calculated by dividing the number of CFSElow cells by the total number of CFSE+ cells (Fig. 2A). As shown in Fig. 2B, the proliferation index was increased in C57BL/6 and IL-18−/− mice compared with IL-12−/− mice (P < 0.0001) following immunization with BRD509/pGO. There was no statistically significant difference in proliferation of OT-II cells between C57BL/6 mice and IL-18−/− mice (P = 0.36). No proliferation of OT-II cells was detected in mice immunized with BRD509/pG. All bacteria recovered from organs of infected mice retained the pG or pGO plasmid (data not shown). As a positive control, mice received an oral dose of 20 mg of whole OVA protein, which induced equivalent levels of proliferation of OT-II cells in C57BL/6, IL-12−/−, and IL-18−/− mice (Fig. 2A), indicating that the OT-II cells were able to proliferate in all strains of mice.
Reduced proliferation of antigen-specific CD4+ T cells in IL-12−/− mice. CFSE-labeled OT-II cells (5 × 106 cells) were adoptively transferred to C57BL/6, IL-12−/−, or IL-18−/− animals, and 1 day later the mice were immunized with either 1010 CFU BRD509/pGO or BRD509/pG or 20 mg OVA via the oral route. Mice were euthanized after 3 days, and proliferation of CFSE-labeled OT-II cells in the MLNs was examined using flow cytometry (A). A proliferation index was then calculated by dividing the number of CFSElow cells by the number of CFSE+ cells as indicated in panel A. The proliferation index for each animal is shown in panel B. The data are the data for one representative animal in a group of five animals in panel A and the results for one representative experiment out of two experiments performed in panel B.
Increased cytokine production in S. enterica serovar Typhimurium-infected IL-12−/− mice.We next examined whether the inability of IL-12−/− mice to control the infection was due to a lack of IFN-γ production. Single-cell suspensions from MLNs and spleens of infected and naïve animals were restimulated in vitro for 6 h with immobilized anti-CD3 to reactivate in vivo-activated T cells (25) and, following permeabilization, were labeled with an IFN-γ-specific MAb and analyzed by flow cytometry (Fig. 3A). No IFN-γ-producing cells were detected in unstimulated cultures (data not shown), and very low numbers of IFN-γ-producing cells were detected in anti-CD3-stimulated cells from naïve mice (Fig. 3A). Strong IFN-γ responses from both CD4+ and CD8+ lymphocytes were detected in the spleens and MLNs of infected mice, and approximately twice as many absolute numbers of CD4+ cells as CD8+ cells produced IFN-γ (Fig. 3B to E). Figures 3B and 3C show that the numbers of CD4+ IFN-γ-producing cells and CD8+ IFN-γ-producing cells were maximal in the spleens of all strains of mice around 25 days after infection, and there were no significant differences in the number of CD4+ IFN-γ-producing cells between strains of mice (Fig. 3B), even though the bacterial load in IL-12−/− mice was significantly increased compared with the bacterial loads in C57BL/6 and IL-18−/− mice (Fig. 1). In addition, no differences in the number of CD8+ IFN-γ-producing cells in the spleen were observed on days 12 and 25 after infection (Fig. 3C). However, on day 40 after infection, while the absolute numbers of CD4+ IFN-γ+ and CD8+ IFN-γ+ cells were reduced in all strains of mice compared to the numbers on day 25, there was an increased number of IFN-γ-producing cells in IL-12−/− mice compared with C57BL/6 and IL-18−/− mice (P < 0.05).
S. enterica serovar Typhimurium induced IFN-γ production. Mice were orally immunized with 1010 CFU BRD509, and at the indicated time points the mice were euthanized and the numbers of cells producing IFN-γ were determined using intracellular staining and flow cytometric analysis. (A) Labeling of CD4+ cells with either anti-IFN-γ antibody or isotype control antibody for one representative of five animals per group, measured on day 25 after infection. The numbers in the upper right quadrants are the percentages of CD4+ cells that were labeled with the antibody. (B to D) Absolute numbers of CD4+ (B and D) and CD8+ (C and E) cells producing IFN-γ in the spleens (B and C) and MLNs (D and E) of C57BL/6 (open columns), IL-12−/− (filled columns), and IL-18−/− (cross-hatched columns) mice. The data are the means and standard deviations for groups of five animals in one representative experiment out of three experiments performed.
The kinetics of the IFN-γ response in the MLNs of infected mice differed from the kinetics in the spleen, and increased numbers of cells were produced on days 12 and 25 after infection compared with day 40. The absolute numbers of CD4+ and CD8+ IFN-γ-producing cells in the MLNs of IL-12−/− mice were significantly increased compared with the numbers in C57BL/6 and IL-18−/− mice at day 40 after infection. No differences were detected between C57BL/6 and IL-18−/− mice.
Expression of IFN-γ receptor.The experiments described above showed that an increased number of CD4+ T lymphocytes from IL-12−/− mice produced IFN-γ, but the CD4+ T lymphocytes did not proliferate in response to antigenic exposure in vivo. Since IFN-γ is known to further activate T lymphocytes as well as bactericidal activities of macrophages, it was hypothesized that the inability of IL-12−/− mice to control the infection may be due to an inability of effector cells to respond to IFN-γ. Therefore, the expression of the IFN-γ receptor was analyzed by flow cytometry using a specific MAb. Figure 4A shows that equally high proportions of splenocytes (>95%) from C57BL/6, IL-12−/−, and IL-18−/− mice expressed the IFN-γ receptor at equivalent levels, and no differences were observed between naïve and infected mice. Further analysis showed that the percentage of spleen cells expressing the IFN-γ receptor and the level of expression (mean fluorescence intensity) of the IFN-γ receptor on spleen cells (Fig. 4B) did not change over the course of the infection and that the values were comparable for the different mouse strains. Similar results were obtained with MLN cells (data not shown).
Flow cytometric detection of expression levels of IFN-γ receptor. (A) Spleen cells from naïve and S. enterica serovar Typhimurium BRD509-infected C57BL/6, IL-12−/−, and IL-18−/− mice were labeled with an anti-IFN-γ receptor antibody (solid lines and gray shading) or an isotype control antibody (dashed lines) and analyzed by flow cytometry. The data are the results for one representative animal from three independent experiments performed with five animals per mouse strain. (B) Mean fluorescence intensity (MFI) of IFN-γ receptor expression on spleen cells of BRD509-infected C57BL/6 (open columns), IL-12−/− (filled columns), and IL-18−/− (cross-hatched columns) mice. The data are means and standard deviations for groups of five mice.
Role of macrophages in control of S. enterica serovar Typhimurium infection.We next investigated whether the bactericidal activity of macrophages from IL-12−/− animals differed from the bactericidal activity of macrophages from C57BL/6 mice or IL-18−/− mice. Macrophages obtained from the peritoneal cavity of mice were infected in vitro with S. enterica serovar Typhimurium BRD509, and the number of bacteria present in the macrophages was determined at several time points after infection. No difference was observed in the bacterial loads in macrophages from C57BL/6 mice, IL-12−/−, mice and IL-18−/− (Fig. 5A), suggesting that the bactericidal activities of the macrophages were comparable in these mouse strains.
Effect of in vivo depletion of macrophages on control of S. enterica serovar Typhimurium infection. (A) Peritoneal macrophages from C57BL/6 (open columns), IL-12−/− (filled columns), and IL-18−/− (cross-hatched columns) mice were infected with S. enterica serovar Typhimurium BRD509 at a multiplicity of infection of 10 for 1 h. Extracellular bacteria were killed by addition of fresh medium containing gentamicin, and 1 and 4 h later the numbers of bacteria present in the macrophages were determined by viable counting of cell lysates. The data are the means and standard deviations for triplicate cultures. (B and C) C57BL/6 (open columns), IL-12−/− (filled columns), and IL-18−/− (cross-hatched columns) mice were treated with either PBS (−) or Cl2MDP liposomes (+) before oral inoculation with 1010 CFU BRD509. At 2 weeks (B) and 4 weeks (C) after infection, mice were killed, and the number of bacteria present in the spleen was determined using viable cell counting. The data are the means and standard deviations for groups of five mice.
To further address the question whether macrophages are unable to control S. enterica serovar Typhimurium infection in IL-12−/− mice, mice were selectively depleted of macrophages using Cl2MDP liposomes (54, 56, 57). Mice were treated with Cl2MDP liposomes 2 days before infection with S. enterica serovar Typhimurium BRD509, and depletion was maintained for 4 weeks. This treatment resulted in depletion of macrophages from the liver and spleen, as demonstrated by histology (54, 56, 57; data not shown). At 2 and 4 weeks after infection, the bacterial colonization of PPs and spleens was determined by viable counting (Fig. 5). As shown previously (56), treatment of C57BL/6 mice with Cl2MDP liposomes during infection with S. enterica serovar Typhimurium BRD509 did not affect the bacterial load in the PPs or spleen. While the bacterial load was significantly increased in IL-12−/− animals compared with C57BL/6 and IL-18−/− mice, depletion of macrophages from IL-12−/− mice did not significantly alter the number of bacteria in the PPs or spleen. In contrast, treatment of IL-18−/− mice with Cl2MDP liposomes resulted in an increased bacterial burden in the PPs (data not shown) (P < 0.01) and spleen (P < 0.01) at both 2 and 4 weeks after infection with S. enterica serovar Typhimurium BRD509 (Fig. 5B,C).
DISCUSSION
Both IL-12 and IL-18 are central to the induction of IFN-γ production (18, 32). Several studies have demonstrated a role for IL-12 in control of intracellular microbial infections (27, 28, 35, 45, 50, 58, 59). In studies in which IL-12 was neutralized by injection of an MAb, it was found that both innately resistant and susceptible mice succumbed to infection with attenuated as well as wild-type S. enterica serovar Typhimurium strains, which was attributed to reduced levels of circulating IFN-γ (17, 27, 28). In this study, we used mice harboring a mutation in the IL-12p40 subunit (24), which therefore are unable to produce biologically active IL-12, as well as IL-18−/− mice. The results of our study confirmed that IL-12 is essential for control of infection of susceptible mice with an attenuated strain of S. enterica serovar Typhimurium, whereas IL-18 seems to be dispensable. Since IL-12 and IL-18 orchestrate the production of IFN-γ by T cells and since activation of IFN-γ-producing CD4+ T cells is considered pivotal for control of S. enterica serovar Typhimurium infections, we investigated the response of antigen-specific CD4+ T cells in Salmonella-infected mice. Recombinant S. enterica serovar Typhimurium cells expressing OVA epitopes were used, enabling the use of OVA-specific T-cell receptor transgenic mice (OT-II) to study in vivo responses of OVA-specific CD4+ T cells. This approach has been successfully employed by other workers to investigate the response of Salmonella (surrogate) antigen-specific T cells in vivo following infection of mice (2). Using this technique, we demonstrated that in contrast to the findings for IL-18−/− and C57BL/6 mice, Salmonella antigen-specific CD4+ T lymphocytes did not proliferate in vivo in the MLNs of S. enterica serovar Typhimurium-infected IL-12−/− mice. As a control, mice were immunized with 20 mg whole OVA protein, resulting in proliferation of OT-II cells in all strains of mice (Fig. 2A), suggesting that the lack of proliferation of OT-II cells in IL-12−/− mice was not due to an inability of OT-II cells to proliferate in response to antigen in IL-12−/− mice. The proliferation of OT-II cells in response to 20 mg OVA was more extensive than that induced by recombinant S. enterica serovar Typhimurium, most likely due to the amount of OVA administered compared to the amount of OVA epitopes that would be expressed by the recombinant S. enterica serovar Typhimurium strain. Using a comparable expression system, Bumann calculated that around 150 pg of GFP-OVA was expressed in the MLNs of infected mice (2). Alternatively, the reduced proliferation of OT-II cells during Salmonella infection compared with the levels detected following feeding of soluble OVA may also have been due to Salmonella-mediated inhibition of T-cell proliferation (52). Nevertheless, our results indicated that in the absence of IL-12p40, antigen-specific CD4+ T cells failed to proliferate during S. enterica serovar Typhimurium infection, suggesting that there is a defect in the early presentation of Salmonella-associated antigens and/or activation of T cells in IL-12−/− mice compared with C57BL/6 and IL-18−/− mice. This result may be explained by the fact that IL-12 is a stimulus not only for IFN-γ production by T cells but also for the growth of activated T cells (55).
We next investigated whether the lack of proliferation of antigen-specific CD4+ T cells correlated with reduced production of cytokines. To increase the sensitivity of detection of IFN-γ production by intracellular staining and flow cytometry, cells from infected and naïve control mice were briefly restimulated in vitro with immobilized anti-CD3, a protocol known to restimulate only the T cells that were preactivated in vivo (25). Indeed, this treatment did not stimulate IFN-γ production by T cells from naïve mice (Fig. 3A). Interestingly, these experiments revealed that the numbers of CD4+ and CD8+ T cells producing IFN-γ in the spleen and MLNs in response to infection with S. enterica serovar Typhimurium were equivalent during the first 4 weeks after infection and increased significantly at later time points compared with the numbers in C57BL/6 and IL-18−/− mice. The increased number of CD4+ and CD8+ T cells producing IFN-γ at day 40 may in part have been due to the increased bacterial load; however, at days 12 and 25 the bacterial loads in IL-12−/− mice were 100- to 1,000-fold higher than the loads in C57BL/6 and IL-18−/− mice, while the numbers of IFN-γ-producing cells at these time points were equivalent. Regardless, the most important finding of this study is that during infection of IL-12p40−/− mice with S. enterica serovar Typhimurium, the number CD4+ and CD8+ IFN-γ-producing cells was at least equivalent to the number of IFN-γ-producing CD4+ and CD8+ T cells detected in infected C57BL/6 mice and IL-18−/− mice.
We attempted to measure the production of IL-4 and IL-10 as well, but we were unsuccessful at detecting any cells producing IL-4 or IL-10 by intracellular staining. However, analysis of Salmonella-specific antibody titers showed a predominant immunoglobulin G2 response (data not shown) in all strains of mice, suggestive of a CD4 T-helper type 1 response. At present, it is unclear why the CD4+ T-cell cytokine response detected in IL-12−/− mice is insufficient to protect the mice from the lethality of S. enterica serovar Typhimurium infection. However, we did not measure the amount of circulating, secreted IFN-γ and other cytokines, and it may well be that while there were more CD4+ T cells producing IFN-γ, the secretion of the cytokines was impaired. Furthermore, although we were unable to detect production by T cells of other cytokines that counteract the effect of IFN-γ, such as IL-4 and IL-10 (12, 44), the circulating levels of such cytokines may be increased in IL-12−/− mice. Regardless, our results imply that the induction of Th1 immunity alone appears to be insufficient to protect IL-12−/− mice from lethal Salmonella bacteremia. In this regard, it has been demonstrated that control of growth of Mycobacterium tuberculosis in mice depends, at least in part, on IFN-γ-independent effects of CD4+ T cells (7). The results of this study suggest that during infection with S. enterica serovar Typhimurium, such IFN-γ-independent effects rely on proliferation of antigen-specific CD4+ T cells.
Macrophages are generally regarded as the major host cells for S. enterica serovar Typhimurium, and they are considered to be the principal effectors that restrict growth of salmonellae during infection (3, 16, 56); IFN-γ-mediated activation of macrophages is important for activation of bactericidal mechanisms of macrophages. Infection of IL-12p40−/− mice with ΔaroAD S. enterica serovar Typhimurium resulted in splenomegaly as a result of an influx of large granular cells. Since the expression levels of the IFN-γ receptor were equivalent in IL-12−/−, IL-18−/−, and C57BL/6 mice during the course of the infection, suggesting that cells from all strains of mice can be activated by IFN-γ, it was hypothesized that the failure of IL-12−/− mice to control the infection may not have been due to a lack of IFN-γ production or an inability to respond to IFN-γ but may have been the result of an inability of macrophages to kill bacteria. Results of previous studies have shown that in mice treated with an IL-12-specific MAb, Salmonella cells are present in spleen and liver macrophages, and it was suggested that the bactericidal capacity of macrophages in the absence of IL-12 is impaired (28). Other studies showed that in S. enterica serovar Enteritidis-infected IL-12p40−/− mice, formation of granulomas was impaired, which was compensated for by an influx of neutrophils (22). The results of our in vitro experiments indicated that macrophages obtained from the peritoneal cavity of IL-12−/− mice were as capable of controlling infection with S. enterica serovar Typhimurium as macrophages from C57BL/6 mice and IL-18−/− mice. We therefore further addressed the function of macrophages in in vivo control of S. enterica serovar Typhimurium infection by eliminating macrophages from mice using Cl2MDP liposomes, a treatment known to selectively remove macrophages and not other phagocytic cells, such as granulocytes or dendritic cells, from the spleen and liver (54, 56, 57). Cl2MDP liposomes were injected before and during the infection to ensure continuous depletion of newly recruited macrophages. Using this technique, we have previously shown that macrophages play a major role in the pathogenesis of virulent S. enterica serovar Typhimurium infections and are important effectors of immunity in vaccinated mice (56). These studies also demonstrated that removal of macrophages during vaccination with attenuated S. enterica serovar Typhimurium (ΔaroAD) did not affect the growth or clearance of the bacterium or the induction of protective immunity (56). Similarly, this study showed that elimination of macrophages did not significantly alter the bacterial load in C57BL/6 mice and, moreover, that removal of macrophages did not affect the bacterial load in IL-12−/− animals, suggesting that macrophages in IL-12−/− mice exerted no bactericidal or bacteriostatic effect. These findings confirm our previous finding that macrophages are not essential for control and elimination of ΔaroAD attenuated S. enterica serovar Typhimurium strains and also suggest that ΔaroAD S. enterica serovar Typhimurium may reside and replicate in host cells other than macrophages, such as liver cells (56). It is possible that the large influx of neutrophils detected in the spleens of IL-12−/− mice may help control the infection, stimulated by the large number of CD4+ T lymphocytes producing cytokines.
Recently, increased sensitivity of IL-18−/− mice to systemic infection with virulent S. enterica serovar Typhimurium strain SL1344 was reported, and by comparing the susceptibilities of mice deficient in either caspase-1, IL-1β, or IL-18, it was demonstrated that IL-18, generated by caspase-1, was most important for control of the systemic phase of S. enterica serovar Typhimurium infection (40). In contrast to this and other studies that have reported increased susceptibility to infection with virulent S. enterica serovar Typhimurium in the absence of biologically active IL-18 (9, 26, 40), we report here that IL-18−/− mice are as capable of controlling oral infection with ΔaroAD attenuated S. enterica serovar Typhimurium as C57BL/6 mice. The altered in vivo growth characteristics of the ΔaroAD strain compared with the virulent parent strain SL1344 result in decreased demands on the immune system for control of the infection, which most likely explains this discrepancy. Interestingly, we found that in contrast to the findings for C57BL/6 and IL-12−/− mice, the bacterial load was significantly (P ≤ 0.01) increased in the PPs (data not shown) and spleens of IL-18−/− mice following depletion of macrophages, suggesting that in the absence of IL-18, the presence of macrophages is important to limit growth of S. enterica serovar Typhimurium ΔaroAD strain BRD509. Control of the infection by macrophages may be either direct due to their bactericidal activities or through the production of cytokines and other inflammatory mediators, possibly induced by IL-18, that may activate other cells of the immune system, such as neutrophils or NK cells, which may kill the bacteria.
In addition to the promotion of Th1 responses and IFN-γ production, IL-12 and signaling through the IL-12 receptor regulate several other immune functions related to tissue homing and migration of T-helper lymphocytes, including upregulation of selectins on high endothelial venules (43) and induction of extracellular matrix proteins which facilitate binding of T cells (4). IL-12 also enhances the expression of several chemokine receptors, notably CCR5 and CXCR3 by naïve T cells and CCR1 by Th1 cells, which influence the migration of T cells into inflammatory sites (4). In addition to pairing with IL-12p35 to form IL-12p75, IL-12p40 forms homodimeric IL-12p40 (p402) and a heterodimer with p19, generating IL-23. While IL-23 has been shown to be important in chronic inflammation and formation of granulomas by recruiting and activating inflammatory cells, its role in control of intracellular pathogens is thought to be limited (21). Therefore, while we demonstrated in this study that the number of IFN-γ-producing CD4+ and CD8+ T lymphocytes is not reduced in the absence of IL-12p40, the lack of control of Salmonella infection may be due to reduced recruitment of T lymphocytes and other inflammatory cells to infected tissues.
In this and other studies, an attenuated S. enterica serovar Typhimurium strain (BRD509) was used that due to deletions in aroA and aroD is unable to synthesize aromatic compounds and is therefore thought to be attenuated for growth in mammalian tissues. Aromatic mutants of S. enterica serovar Typhimurium and S. enterica serovar Typhi have been intensively studied for use as live attenuated vaccines (37) and are normally cleared in 6 to 8 weeks after oral administration (56). In IL-12p40−/− mice or in mice treated with anti-IL-12 antibodies, a high bacterial load is detected in the PPs and spleen starting 2 weeks after infection, and the titer slowly increases over the next 6 weeks, until the animals die. The growth rate of the ΔaroAD S. enterica serovar Typhimurium strain is kept under control in the IL-12p40−/− mice compared with growth kinetics of virulent S. enterica serovar Typhimurium strains, even in IL-12p40−/− mice depleted of macrophages, and while antigen-specific CD4+ T cells fail to proliferate early in response to infection, IFN-γ production by T cells can be detected in S. enterica serovar Typhimurium-infected mice. Considering the continuous high bacterial burden in mice lacking IL-12, it is not clear why the animals die only 8 weeks after infection and not earlier, and this remains an important unanswered question that needs to be investigated in further studies.
ACKNOWLEDGMENTS
We thank M. Smyth (Peter MacCallum Cancer Centre, Melbourne, Australia) for kindly providing the IL-12p40−/− mice, S. Akira (Osaka University, Japan) for the generous gift of IL-18−/− mice, and F. Carbone (Department of Microbiology and Immunology, The University of Melbourne, Melbourne, Australia) for making OT-II mice available.
This work was financially supported by the NH&MRC and the Cooperative Research Centre for Vaccine Technology. O.L.C.W. was supported by a C. R. Roper Fellowship from The University of Melbourne and is currently an NH&MRC R. D. Wright Fellow.
FOOTNOTES
- Received 17 July 2007.
- Returned for modification 20 August 2007.
- Accepted 2 September 2007.
- Copyright © 2007 American Society for Microbiology
