Clonal Expansion of Antigen-Specific CD4 T Cells following Infection with Salmonella typhimurium Is Similar in Susceptible (Ity^s) and Resistant (Ity^r) BALB/c Mice

Adoptive transfer of T cells from TCR transgenic mice and construction of a recombinant S. typhimurium strain expressing chicken OVA.

We directly assessed the effects of Nramp-1 on T-cell activation by using a model system in which a population of chicken ovalbumin (OVA)-specific T cells from T-cell antigen receptor (TCR) transgenic mice could be physically tracked in vivo during infection with an S. typhimurium strain that expresses OVA. Because OVA-specific T cells are too rare to detect directly in the repertoire of normal mice, it was necessary to transfer into normal BALB/c recipients a small number of CD4 T cells from DO11.10 TCR transgenic mice that uniformly express a TCR specific for an OVA peptide-class II major histocompatibility complex molecule (18). The TCR expressed by DO11.10 T cells is not expressed on other T cells in BALB/c mice and can be detected with the KJ1-26 anticlonotypic antibody (8). DO11.10 T cells (2.5 × 10⁶) were injected intravenously into unirradiated BALB/c (Ity^s) mice (purchased from the National Cancer Institute, Frederick, Md.) or BALB/c.DBA2 Ity^r [BALB/c (Ity^r)] congenic mice (15, 23) (kindly provided by Bruce Zwilling, Ohio State University) as previously described (20).

A recombinant S. typhimurium strain expressing OVA was constructed by using the expression vector pYA3149 (19) (kindly provided by Roy Curtiss III, Washington University, St. Louis, Mo.). S. typhimurium x4550 lacks a functional aspartate β-semialdehyde dehyrogenase gene and thus cannot synthesize a cell wall unless diaminopimelic acid, the product of the reaction catalyzed by aspartate β-semialdehyde dehyrogenase, is provided in the medium. However, S. typhimurium x4550 can grow in the absence of diaminopimelic acid if complemented with plasmid pYA3149, which contains a functional aspartate β-semialdehyde dehyrogenase gene. The OVA coding sequence was inserted into the pYA3149 plasmid just downstream of the P_trc promoter to produce the pYA3149-OVA plasmid (Fig. 1A), which was when introduced into S. typhimurium x4550, producing bacteria that grow in the absence of diaminopimelic acid and constitutively express OVA (55 kDa) at a level of 70 μg of OVA/10⁸ bacteria as assessed by immunoblotting with anti-OVA antibody (Fig. 1B).

Recombinant S. typhimurium infection in BALB/c (Ity^r) and BALB/c (Ity^s) mice.

The infectivity of the recombinant S. typhimurium was tested in BALB/c (Ity^s) and BALB/c (Ity^r) mice that did or did not contain DO11.10 T cells. After subcutaneous inoculation, the number of bacteria in the draining lymph nodes increased rapidly (Fig. 2A) such that a stable level was achieved by day 3 postinfection. An identical pattern was observed in BALB/c (Ity^s) and BALB/c (Ity^r) mice whether or not DO11.10 T cells were present. The bacteria recovered at all time points retained expression of OVA (data not shown). Thus, BALB/c (Ity^r) mice did not clear the OVA-expressing S. typhimurium organisms from the lymph nodes faster than did BALB/c (Ity^s) mice and the presence of initially naive DO11.10 T cells did not alter the course of the infection.

In contrast, bacterial clearance from the spleen differed greatly between BALB/c (Ity^s) and BALB/c (Ity^r) mice. The number of viable bacteria in the spleens of BALB/c (Ity^s) mice increased rapidly to a plateau level on day 5 (Fig. 2B). About half of the infected BALB/c (Ity^s) mice succumbed to the infection by day 7 postinfection, and the survivors had high bacterial loads even 21 days postinfection. Although the bacteria also initially increased in the spleens of BALB/c (Ity^r) mice, the maximal level achieved on day 5 was about 10-fold less than that observed in BALB/c (Ity^s) mice (Fig. 2B). In addition, the number of bacteria in the spleen declined after day 5 and the bacteria became undetectable 21 days postinfection. The bacteria were cleared at equal rates from the spleens of BALB/c (Ity^r) mice that did or did not receive DO11.10 T cells prior to infection. These data demonstrated that BALB/c (Ity^r) mice were capable of clearing OVA-expressing S. typhimurium, at least from the spleen, and that the presence of a large population of initially naive OVA-specific DO11.10 T cells did not enhance the clearance rate.

DO11.10 T cells proliferate similarly in BALB/c (Ity^s) and BALB/c (Ity^r) mice during the early phase of infection with recombinant S. typhimurium expressing OVA.

The clonal expansion of DO11.10 T cells was monitored by flow cytometry in the lymphoid tissues of BALB/c (Ity^s) or BALB/c (Ity^r) recipients infected with OVA-expressing recombinant S. typhimurium. As previously shown (4, 12, 20), CD4⁺ KJ1-26⁺ T cells could not be detected in lymph node cells of mice that did not receive DO11.10 T cells (Fig. 3A). In contrast, a small population of CD4⁺ KJ1-26⁺ T cells was present in lymph nodes of mice that received DO11.10 cells (Fig. 3B). The numbers of DO11.10 cells detected after transfer were equivalent in BALB/c (Ity^s) and BALB/c (Ity^r) mice (Fig. 3E). The number of CD4⁺ KJ1-26⁺ DO11.10 T cells in the draining lymph nodes increased dramatically after subcutaneous infection with OVA-expressing S. typhimurium (Fig. 3C and E), such that a maximal level was achieved on day 5. This increase in DO11.10 T cells did not occur in recipients that were infected with control S. typhimurium that did not express OVA (Fig. 3D and E), indicating the specificity of the response. Accumulation of DO11.10 cells was also observed in the spleens of mice infected with OVA-expressing S. typhimurium (data not shown). An identical increase in the number of DO11.10 T cells in the draining lymph nodes was observed in BALB/c (Ity^s) and BALB/c (Ity^r) recipients 3 and 5 days after subcutaneous infection with OVA-expressing S. typhimurium. Twelve days into the infection, a time at which BALB/c (Ity^r), but not BALB/c (Ity^s), mice had cleared the bacteria from the spleen, the number of DO11.10 T cells in the draining lymph nodes was significantly lower in susceptible BALB/c (Ity^s) recipients than in resistant BALB/c (Ity^r) recipients (Fig. 3E). This difference was due mainly to the fact that the total number of cells in the lymph nodes was much greater in BALB/c (Ity^r) recipients than in BALB/c (Ity^s) recipients; the percentages of DO11.10 T cells present in the lymph nodes of the two groups were actually similar at this late time point. These results show that antigen-specific T cells expand and survive equally well in susceptible and resistant recipients, at least until late times after infection.

This conclusion was supported by results from bromodeoxyuridine (BrDU) labeling experiments. BALB/c (Ity^s) and BALB/c (Ity^r) recipients of DO11.10 T cells were infected subcutaneously with OVA-expressing S. typhimurium and injected daily with the thymidine analog BrDU (0.8 mg/day). Cohorts of mice were sacrificed on days 3 and 5 postinfection, and the percentage of DO11.10 cells (identified as viable CD4⁺ KJ1-26⁺ cells) that had incorporated BrDU into the DNA (detected by anti-BrDU antibody staining) was determined by flow cytometry. The percentage of BrDU⁺ DO11.10 T cells was 17% on day 3 and 35% on day 5 postinfection in both BALB/c (Ity^s) and BALB/c (Ity^r) recipients. The specificity of BrDU detection was shown by the finding that <1% of the DO11.10 T cells were BrDU⁺ at any time in BALB/c (Ity^s) or BALB/c (Ity^r) recipients infected with S. typhimurium that did not express OVA. These results demonstrate that antigen-specific T cells proliferate equally well at early times after infection in the two strains of mice.

IFN-γ production by DO11.10 cells recovered from BALB/c (Ity^s) and BALB/c (Ity^r) mice during infection with OVA-expressing S. typhimurium.

Since IFN-γ has been shown to be critical for control of S. typhimurium infection (10), attention was focused on the capacity of DO11.10 T cells to produce this lymphokine during infection. Draining lymph node cells from BALB/c (Ity^s) or BALB/c (Ity^r) recipients of DO11.10 T cells were harvested 5 or 10 days after infection and cultured in vitro with OVA peptide. On day 5 after infection, the same amount of IFN-γ was produced in response to the OVA peptide by lymph node cells from BALB/c (Ity^s) and BALB/c (Ity^r) recipients of DO11.10 T cells (group 2 in Fig. 4A). In contrast, 10 days after infection, OVA-stimulated lymph node cells from BALB/c (Ity^r) recipients made much more IFN-γ than did BALB/c (Ity^s) recipients (group 2 in Fig. 4B). However, the day 10 cultures contained different numbers of input DO11.10 T cells (Fig. 3E), and thus, it was important to determine whether this alone accounted for the differences in IFN-γ production by BALB/c (Ity^s) and BALB/c (Ity^r) recipients. Most of the IFN-γ detected was dependent on DO11.10 T cells because IFN-γ was barely detected from antigen-stimulated cultures of lymph node cells from mice that were infected with OVA-expressing S. typhimurium but did not receive DO11.10 T cells (group 4 in Fig. 4A). This dependence made it possible to estimate the amount of IFN-γ produced by each DO11.10 T cell by dividing the total amount of IFN-γ detected by the number of DO11.10 cells present in the culture. This analysis showed that 5 days after infection with OVA-expressing S. typhimurium, DO11.10 T cells from BALB/c (Ity^s) and BALB/c (Ity^r) recipients produced similar amounts of IFN-γ (group 2 in Fig. 4C), whereas 10 days after infection, DO11.10 T cells from susceptible BALB/c (Ity^s) recipients infected with OVA-expressing S. typhimurium produced about sixfold less IFN-γ than did cells from BALB/c (Ity^r) recipients (group 2 in Fig. 4D). This difference was specific to IFN-γ because no significant differences in antigen-stimulated interleukin-2 (IL-2), IL-5, IL-6, or tumor necrosis factor alpha production were observed (data not shown). No IFN-γ was produced by lymph node cells from recipients of DO11.10 T cells that were not infected (group 1 in Fig. 4A and C) or were infected with non-OVA-expressing bacteria (group 3 in Fig. 4A), consistent with the idea that naive T cells must differentiate in vivo in response to antigenic stimulation before acquiring the capacity to produce IFN-γ (34). Finally, recall IFN-γ production by DO11.10 T cells was similar in BALB/c (Ity^s) and BALB/c (Ity^r) recipients that had been primed 2 weeks earlier with soluble OVA in complete Freund’s adjuvant (data not shown), indicating that the reduced IFN-γ production observed in infected BALB/c (Ity^s) recipients was related to bacterial infection.

By direct tracking of antigen-specific cells, it was possible to show that the relative resistance to S. typhimurium infection conferred by the Ity^r allele could not be explained by superior clonal expansion of antigen-specific CD4 T cells in this environment at early times of infection. This finding was surprising because in vitro studies have shown that macrophages from resistant mice stimulate T cells more efficiently than those from susceptible mice (13). The discrepancy between the in vivo and in vitro results could be related to the fact that dendritic cells, which have not been reported to express Nramp-1, not macrophages, are the antigen-presenting cells that initiate clonal expansion in vivo (9). It is possible, however, that macrophages participate as antigen-presenting cells later in the response and that macrophages expressing the resistant form of Nramp-1 are better at stimulating previously activated T cells at this point. This could explain our finding that clonal expansion of antigen-specific T cells was reduced in BALB/c (Ity^s) recipients at late times after infection.

The basis for the retention of IFN-γ production potential by antigen-specific T cells in the resistant environment and its relationship to Nramp-1 function is not clear. The Nramp-1 product encoded by the Ity^r allele is clearly responsible for reducing the bacterial load late in infection, perhaps by regulating the bactericidal functions of macrophages (6, 25, 28). Lack of this activity in susceptible mice results in the accumulation of large numbers of bacteria, which may stimulate factors, e.g., IL-10 (16, 21), which are suppressive to IFN-γ-producing T cells. However, IL-10 is thought to inhibit all Th1 cytokines (17), and thus, exaggerated production of IL-10 cannot account for the specificity of the effect on IFN-γ production that we observed. It is possible that some factor specific to the IFN-γ pathway, perhaps IL-18 (27, 30), is preferentially produced by macrophages in the resistant environment. Alternatively, the stress of uncontrolled bacterial infection in BALB/c (Ity^s) recipients may lead to the production of glucocoriticoids that could inhibit IFN-γ production.

Our finding that the bacteria were cleared from the spleen but not the draining lymph nodes, even in BALB/c (Ity^r) mice containing an artificially high number of antigen-specific T cells, suggests that Nramp-1 is especially important for control of disseminated, blood-borne organisms. It was surprising that BALB/c (Ity^r) mice that had cleared the S. typhimurium organisms from the spleen but still contained viable bacteria in the lymph nodes did not display symptoms of salmonellosis. This suggests that the pathology of salmonellosis is related to systemic infection and that the immune response can achieve compartmentalized immunity that does not clear the organism but prevents disseminated disease.