INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Typhoid fever, a significant health care problem in a number of developing nations, is caused by infection with the gram-negative bacterium Salmonella enterica serovar Typhi (12). After ingestion of contaminated food or water, salmonellae are able to penetrate the gut epithelium through specialized M cells lining the Peyer's Patches (11). After crossing this intestinal barrier, bacteria are able to spread rapidly to visceral tissues, including the liver and spleen, where they reside primarily in tissue macrophages.

Infection of susceptible mice with S. enterica serovar Typhimurium provides a murine model for typhoid fever which bears many similarities to human serovar Typhi infection. This disease is ultimately fatal due to the inability of such mice to restrict bacterial growth in vivo. Numerous attenuated strains of serovar Typhimurium have been shown to stimulate protective immunity in susceptible mice when administered several weeks prior to challenge with a virulent strain (10, 29, 40, 43), and similar attenuated strains of serovar Typhi have been produced as potential vaccines in humans (20, 39).

The mechanisms of immunity to Salmonella have been studied in naive hosts following infection with attenuated organisms, or in vaccinated hosts following infection with virulent organisms.

Mice containing defects that affect cellular immunity are incapable of clearing infection with avirulent salmonellae. These include T-cell receptor , major histocompatibility complex class II and gamma interferon (IFN-) receptor knockouts, all of which succumb to Salmonella infections that are not lethal in normal mice (9, 41). Thus, the induction of IFN--producing CD4⁺ T cells is critical for the resolution of infection. In contrast, T-cell receptor or ₂-microglobulin knockout mice resolve such Salmonella infections (9, 41), indicating that class I-restricted T cells or T cells are not required. These data are in agreement with recent reports documenting that human genetic deficiencies in IFN- or interleukin-12 (IL-12) production or receptor signaling result in increased susceptibility to Salmonella infection (13).

Infection of vaccinated mice with virulent salmonellae has led to similar conclusions. The protective effect of vaccination with an attenuated Salmonella strain is limited if CD4⁺ T cells are depleted before challenge with a virulent strain (22, 26). In addition, the depletion of Th1-like cytokines such as IFN-, tumor necrosis factor alpha, and IL-12, using neutralizing antibodies after vaccination, exacerbates secondary infection (21, 22). Thus, the vaccine-induced resolution of infection with virulent salmonellae and the defense against primary infection with attenuated salmonellae require the activation of a Th1-like population of T cells. Such a requirement might be expected due to the intracellular localization of salmonellae in vivo, and indeed these results are similar to those described in disease models with other intracellular pathogens (14, 33).

In contrast, the role of antibody in immunity to Salmonella infection remains poorly defined. Vaccination with purified polysaccharide Vi antigen from serovar Typhi stimulates antibody production and provides protection in humans (1, 17). Because this vaccine would not be predicted to stimulate T cells, it is likely that antibody is responsible for protection in this case.

In the murine typhoid model, antibody has also been shown to contribute to the resolution of Salmonella infection under some circumstances. The passive transfer of immune serum can protect genetically resistant Ity^r mice but not susceptible Ity^s mice (6). In addition, resistant mice, which have an X-linked genetic deficiency in B-cell development (xid), are defective in the control of bacterial replication in vivo (27). This increased susceptibility to Salmonella in xid mice can be reversed by the passive transfer of immune serum prior to infection (27). Thus, in resistant mice there is a clear role for antibody in contributing to Salmonella immunity.

However, a role for antibody is less clear for innately susceptible mice. It has been reported that immunity to Salmonella infection can be transferred with T cells alone in these mice (24, 31, 32). However, others have suggested that T cells are not sufficient (5). In one report, the adoptive transfer of immunity to naive mice was achieved only when immune serum was cotransfered with immune splenocytes, suggesting that antibody plays a supplemental role to effective cellular immunity (23).

Here, we demonstrate that the effective clearance of a primary infection with attenuated serovar Typhimurium is critically dependent on costimulatory signals to T cells via CD28 but is completely independent of B cells. However, serum antibody produced by B cells during this initial infection is critical for subsequent protection against virulent salmonellae. Therefore, antibody is required for protection against virulent but not clearance of attenuated salmonellae.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Mice. Female BALB/c and nude mice on the BALB/c background were purchased from the National Cancer Institute, Frederick, Md., and used at 8 to 16 weeks of age. Homozygous CD28-deficient (37) and B-cell-deficient (µ knockout) mice on a BALB/c background were originally obtained from S. Reiner (University of Chicago) and were bred and housed under specific-pathogen-free conditions in our facility.

Bacterial strains and infection. Mice were infected with Salmonella as previously described (9). Virulent (SL1344) or aroA-attenuated (SL3261) serovar Typhimurium was grown overnight in Luria-Bertani medium and diluted in phosphate-buffered saline after estimation of bacterial concentration by spectrophotometry. Mice were infected intravenously with 5 × 10⁵ SL3261 or intraperitoneally with 10 to 10⁴ SL1344. These routes of challenge were preferred to oral challenge in order to avoid possible complications related to the development of Peyer's patches in immunodeficient mice (15, 42). The dose of viable bacteria administered was verified by plating out dilutions of the bacterial solution on MacConkey agar plates, which were incubated at 37°C overnight before counting colonies. In all experiments, the estimated bacterial concentration differed from the actual concentration by less than threefold (data not shown). Infected mice were monitored daily after infection for at least 2 months. Moribund animals (defined by unresponsiveness to gentle prodding) were euthanized in accordance with the University of Minnesota animal care guidelines.

Determination of bacterial growth in vivo. Organ homogenates from at least three mice per time point were incubated overnight on MacConkey agar plates as previously described (25). Colonies were counted the next day, and the number of bacteria per organ was calculated.

HKST preparation and Salmonella-specific antibody measurement. Heat-killed serovar Typhimurium (HKST) was prepared from an overnight culture of SL1344 that had been washed twice in phosphate-buffered saline before incubation at 65°C for 1 h. At various time points after infection with SL3261, blood was obtained from the retro-orbital plexus and serum was prepared. HKST-specific immunoglobulin M (IgM), IgG1, and IgG2a was measured by an enzyme-linked immunosorbent assay (ELISA) method (25). Briefly, antigen was coated on 96-well microtiter plates (Costar, Cambridge, Mass.), and serum was added. After washing, bound antibody was detected using a goat anti-mouse IgG2a horseradish peroxidase-conjugated antibody (Southern Biotechnology Associates, Birmingham, Ala.).

Adoptive transfer of immune serum. Blood was collected from BALB/c mice that had resolved infection with SL3261 at least 2 months previously. Serum was prepared, pooled, and stored at 20°C before being injected intravenously into mice 1 day before challenge with the virulent Salmonella strain, SL1344.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Costimulation via CD28 is required for the resolution of infection with attenuated serovar Typhimurium. It has been reported that T-cell costimulation via CD28 is required for the effective induction of Th2 but not Th1-like cells (18, 35). As Th1-like cells are required for the resolution of Salmonella infection (9), it was of interest to determine if CD28 costimulation is required for the induction of these cells. Gene-targeted mice lacking CD28 expression were therefore infected with the attenuated Salmonella strain, SL3261. As shown in Fig. 1, these mice failed to recover from infection whereas control mice effectively eliminated the bacteria. This defect appeared more prominent at later stages of infection, as there was greater bacterial growth in both the spleen and liver (Fig. 2). The time course of death and bacterial growth observed in CD28-deficient mice was very similar to that previously reported for mice lacking CD4⁺ T cells (9, 38). Thus, CD28 is critical for the induction of protective Salmonella-specific T cells during infection with attenuated serovar Typhimurium.

View larger version (9K): [in this window] [in a new window]   FIG. 1.   CD28-deficient mice succumb to infection with attenuated serovar Typhimurium. CD28-deficient and normal BALB/c mice were infected intravenously with 5 × 105 SL3261 and monitored daily until time of death. Data are pooled from two individual experiments using 7 CD28-deficient () and 10 normal () BALB/c mice. Autopsy confirmed that CD28-deficient mice had overwhelming Salmonella infection in both liver and spleen (data not shown).

View larger version (29K): [in this window] [in a new window]   FIG. 2.   CD28-deficient mice do not clear attenuated serovar Typhimurium from the spleen or liver. CD28-deficient and normal BALB/c mice were infected intravenously with 5 × 105 SL3261, and bacterial colonization of spleen (A) and liver (B) was assessed 2 and 4 weeks later. Data shown mean ± standard deviation of at least three mice per time point.

B cells are not essential to resolve infection with attenuated serovar Typhimurium. CD28 has been shown to be required for the induction of both T-cell responses and B-cell responses to T-dependent antigens (19). Thus, the susceptibility of CD28-deficient mice to Salmonella infection described above could be due to a defect in either humoral or cellular immunity, or a combination of both. To determine the role that humoral immunity plays in resistance to attenuated Salmonella, B-cell-deficient mice were infected with SL3261. In contrast to T-cell-deficient nude mice, which succumbed to infection, B-cell-deficient mice did not die from infection with attenuated Salmonella (Fig. 3) and showed no defect in the rate of clearance of bacteria from the spleen or liver (Fig. 4). Thus, B cells and/or antibodies are not required for the effective resolution of infection with attenuated Salmonella.

View larger version (12K): [in this window] [in a new window]   FIG. 3.   B-cell-deficient mice do not succumb to infection with attenuated serovar Typhimurium. Nude, B-cell-deficient, and wild-type BALB/c mice were infected intravenously with 5 × 105 SL3261 and monitored daily until time of death. Data are pooled from three individual experiments using 27 B-cell-deficient mice, and 24 wild-type, and 5 nude BALB/c mice. Autopsy confirmed that nude mice had overwhelming Salmonella infection in both liver and spleen (data not shown).

View larger version (18K): [in this window] [in a new window]   FIG. 4.   Naive and immune B-cell-deficient mice effectively clear attenuated serovar Typhimurium from spleen and liver. B-cell-deficient and normal BALB/c mice were infected intravenously with 5 × 105 SL3261 (naive mice). After bacteria were cleared from the spleens and livers of these mice, some were reinfected with the same dose of bacteria (immune mice). Bacterial colonization of spleen (A) and liver (B) was assessed in both groups at the indicated time points. Mean ± standard deviation values from at least four mice per time point are shown. Similar results were obtained in two separate experiments.

Primary infection of mice with attenuated serovar Typhimurium induces long-lasting antibody responses. While B cells are not required to clear infection with attenuated Salmonella strain SL3261, an antibody response was induced by exposure to this strain (Fig. 5). Salmonella-specific IgM was detected by ELISA as early as 1 week after infection with SL3261. Three to four weeks after infection, high titers of both HKST-specific IgG1 and IgG2a were present. The titer of Salmonella-specific IgM dropped after 5 weeks, whereas the high titer of anti-Salmonella IgG2a was maintained for at least 16 weeks after infection.

View larger version (13K): [in this window] [in a new window]   FIG. 5.   BALB/c mice infected with attenuated serovar Typhimurium produce serum IgM, IgG1, and IgG2a specific for HKST. BALB/c mice were infected intravenously with 5 × 105 SL3261 and bled at various time points later. Titers of anti-HKST IgM, IgG1, and IgG2a in serum were measured by ELISA. Mean ± standard deviation values for three to five mice per time point are shown.

Secondary clearance of attenuated Salmonella is more efficient with B cells. To determine if the antibody response induced by attenuated Salmonella contributes to the clearance of a secondary Salmonella infection, mice were infected a second time with SL3261 and bacterial growth in vivo was assessed. Mice were reinfected only after primary infection had been resolved, i.e., no earlier than 6 weeks after primary infection. At this time point there were no detectable bacteria in the liver or spleen (data not shown). As expected, the elimination of second infection with SL3261 from both the spleen and liver was greatly accelerated in wild-type mice compared to a primary infection (Fig. 4). This enhanced clearance is likely due to the combined action of memory T and B cells and immune serum. In immune mice lacking B cells, there was also an accelerated clearance of bacteria compared to primary infection (Fig. 4). This immunity cannot be due to B cells and thus most likely reflects the contribution of memory T cells to secondary clearance of bacteria. However, there was a slight but significant delay in this accelerated clearance, especially from the liver in mice lacking B cells (Fig. 4B), suggesting that at least a portion of the faster resolution during secondary infection can be attributed to the presence of B cells. This might be explained by the presence of a memory B-cell response or specific antibody at the time of secondary challenge in wild-type mice that accelerates the clearance of bacteria. Alternatively, the absence of B cells could adversely effect T-cell priming and therefore affect a secondary response indirectly. This is unlikely, as at 1 week after infection immune B-cell-deficient mice have about 1,000- to 10,000-fold fewer bacteria than naive B-cell-deficient mice, indicating that vaccination with attenuated Salmonella induces effective specific immunity in the absence of B cells.

Immune serum is required for vaccine-induced protection against virulent Salmonella. To determine if the contribution of B cells to secondary clearance of bacteria was necessary for protection against virulent Salmonella, vaccinated control and B-cell-deficient mice were challenged with the virulent strain, SL1344. As shown in Fig. 6, B cells were required for the protective immunity observed in wild-type mice, as mice lacking B cells rapidly succumbed to infection with different doses of virulent bacteria. To determine if this effect was due to the presence of Salmonella-specific antibodies or memory B cells, B-cell-deficient mice were injected with immune serum before rechallenge with SL1344. Such immune serum transfer to naive mice or the transfer of preimmune serum to vaccinated mice does not provide protective immunity (data not shown) (5, 23). However, as shown in Fig. 6B, the transfer of immune serum effectively reconstituted immunity to immune B-cell-deficient mice.

View larger version (20K): [in this window] [in a new window]   FIG. 6.   Immune serum is required for vaccine-induced resistance to virulent serovar Typhimurium. B-cell-deficient and normal BALB/c mice were vaccinated intravenously with 5 × 105 SL3261. (A) Groups of mice were challenged intraperitoneally with 100 or 10,000 virulent SL1344 bacteria and monitored daily for death; (B) immunized, B-cell-deficient mice were injected with 0.4 ml of immune serum 1 day prior to challenge with 100 SL1344 bacteria or were untreated. Data are from one of two separate experiments which together total 8 to 10 mice per group. Similar data were also obtained when vaccinated mice were rechallenged intravenously with SL1344 (data not shown).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Here, we demonstrate that the induction of protective cellular immunity during infection with attenuated Salmonella requires costimulation via CD28. These data are in direct contrast to a number of other infectious disease models where CD28 is thought to play a redundant role and account for only minor effects on the disease process (2, 37). The absolute requirement for CD28 in this model may be due to the intracellular localization of the bacteria, which may place a greater emphasis on effective CD4⁺ activation compared to these other infections. Alternatively this difference may be due to the lack of Salmonella products which are able to induce the expression of alternative costimulatory ligands on the antigen-presenting cells and thus reflect a difference between Salmonella products and those of other microorganisms.

Macrophages have been shown to upregulate B7 molecules when exposed to Salmonella extracts in vitro (8), indicating that the activation of T-cell responses requiring CD28 costimulation may be enhanced by bacterial products. Indeed, lipopolysaccharide (LPS) is often used as an adjuvant for protein antigens due to its capacity to induce costimulatory molecules and the production of inflammatory cytokines by antigen-presenting cells (16, 34, 36). These effects of LPS result in the production of an enhanced number of antigen-specific T cells, increased migration into B-cell follicles, and better help for antibody production (30). Furthermore, this adjuvant effect of LPS has been shown to be completely CD28 dependent (16).

CD28-mediated costimulation is known to have profound consequences for the development of humoral immunity, as B cells from mice lacking CD28 fail to isotype switch, do not make germinal centers, and do not acquire somatic mutations (7). Thus, it was possible that the susceptibility of CD28-deficient mice was actually due to an antibody, rather than a cellular, defect. However, as shown here, the effective resolution of infection with attenuated Salmonella in B-cell-deficient mice leads to the conclusion that the attenuating defect in CD28-deficient mice is related to their inability to effectively activate some other function of CD4⁺ T cells, perhaps IFN- production.

While the induction of both T and B cells has been previously recognized as important in Salmonella infection (23), there has been debate over the relative importance of B cells (4, 6, 24, 31, 32), with the majority of available data pointing to an essential role for CD4⁺ T cells and a minor role for antibody. This is understandable, as the macrophage tropism of Salmonella is very similar to that of other organisms such as Leishmania and Listeria, neither of which require antibody for protection (3, 33). Indeed, our data demonstrate that B cells are not required for the resolution of primary or secondary infection with attenuated Salmonella although antibody does accelerate clearance of secondary infection.

In contrast, the vaccine-induced resistance to infection with virulent Salmonella is critically dependent on the presence of antibody. This qualitative difference in antibody requirement between virulent and attenuated Salmonella can be explained by differences in the growth rate of these bacteria in vivo (9, 28, 29, 38) and the time required to generate an effective CD4⁺ T-cell response. Thus, the presence of serum antibody may be required to delay the rapid buildup of virulent bacteria in order to allow time for CD4⁺ T cells to become activated and macrophage activation to occur. Attenuated salmonellae while able to replicate in vivo, do so at a reduced rate (28, 29), thus allowing T-cell activation to occur before excessive bacterial growth has occurred.