|Home | About | Journals | Submit | Contact Us | Français|
CD4+ memory-phenotype T cells decline over time when generated in response to acute infections cleared by other components of the immune system. It was therefore of interest to assess the stability of CD4+ T cells during a persistent Salmonella infection, which is typical of persistent phagocytic infections that are controlled by this lymphocyte subset. We found that CD4+ T cells specific for Salmonella peptide:MHCII ligands were numerically stable for greater than a year after initial oral infection. This stability was associated with peptide:MHCII-driven proliferation by a small number of T cells in the secondary lymphoid organs that harbored bacteria. The persistent population consisted of multi-functional Th1 cells that induced PD-1 and became exhausted when transferred to hosts expressing the specific peptide:MHCII ligand in all parts of the body. Thus, persistent infection of phagocytes produced a CD4+ T cell population that was stably maintained by low-level peptide:MHCII presentation.
Expanded populations of antigen-specific T and B cells that persist long after pathogen clearance are responsible for immunological memory. These cells are useful to the host, providing protective immunity to subsequent challenge by the original microbe. The memory cell paradigm has been established most clearly for CD8+ T cells. CD8+ T cells expressing TCRs specific for microbe peptide:MHCI ligands proliferate extensively, producing a peak number of effector cells about a week after acute infection. About 90% of these effector cells then die by apoptosis, leaving a population of memory cells that is stably maintained by recurrent IL-15-driven homeostatic proliferation. Central memory CD8+ T cells (Tcm) recirculate through secondary lymphoid organs, producing new memory cells during secondary responses, while effector memory (Tem) cells located in non-lymphoid organs are immediately cytotoxic during secondary responses.
It is less clear whether the concept of stable immune memory applies to CD4+ T cells. Naïve CD4+ T cells expressing TCRs specific for microbe peptide:MHCII ligands proliferate extensively to produce a peak number of effector cells about a week after acute infections with Lymphocytic Choriomeningitis Virus (LCMV) or Listeria monocytogenes (1, 2). As in the case of CD8+ T cells, about 90% the effector cells then die, leaving a population of memory-phenotype cells, about half of which are Th1 cells and the other half follicular helper cell-like Tcm cells (3). Although both CD4+ and CD8+ T cells are maintained by recurrent IL-15-driven homeostatic proliferation, the proliferation rate is much lower for CD4+ memory T cells than for CD8+ memory T cells (2, 4). Thus, CD4+ memory T cells slowly decline in number after infection is cleared, probably because their lower rate of homeostatic proliferation cannot keep pace with their death rate.
Previous studies showing the decline of CD4+ memory T cells involved acute infections, which were completely cleared by CD8+ T cells and innate immune cells with no apparent involvement of CD4+ T cells (5, 6). Infections that are controlled by CD4+ T cells (7) are caused by microbes such as Mycobacterium, Leishmania, and Salmonella species, that establish persistent infections in the phagosomes of phagocytes. In these cases, CD4+ T cells contain, but never fully clear the microbes from the initial site of infection, while providing sterilizing immunity at all other body sites (8-10). Notably, clearance of the original infection is associated with loss of systemic immunity (11), indicating that the maintenance of protective CD4+ T cells depends on persistent antigen, as suggested by several recent studies (12-14). The requirement for persistent antigen presentation for CD4+ T cell maintenance goes against the idea derived from studies of CD8+ T cells that durable memory and immune protection develop after antigen is cleared while exhaustion ensues if it is not cleared.
Here, we assessed CD4+ T cell responses to the facultative intracellular bacterium Salmonella enterica that can infect mice and humans through the gastrointestinal tract. Mice expressing Nramp1, a protein that helps limit bacterial replication in phagocytes, develop a persistent infection following oral inoculation, which is controlled by IFN-γ-producing CD4+ T cells (15, 16). In the current study, we developed a model to study a p:MHCII-specific immune response to Salmonella enterica serovar Typhimurium (ST) infection in Nramp1-resistant (Nramp1Gly169/Gly169, referred to hereafter as Nramp1R) mice to understand how bacterial persistence and p:MHCII presentation affect the numerical and functional stability of protective CD4+ T cells.
C57BL/6 (B6), B6.SJL-PtprcaPep3b/BoyJ (CD45.1) mice, and 129SvJ (129) female mice were purchased from the Jackson labs. (B6 × 129) F1, (CD45.1 × 129) F1, and (Act-2W × 129) F1 mice were bred in our facilities. All mice were housed in specific pathogen-free conditions at the University of Minnesota and all experiments were conducted in accordance with institutional and federal guidelines.
Mice were pre-treated by 12-hour food deprivation or intragastric gavage with 100 μl 5% sodium bicarbonate solution, pH 9.0 prior to infection. They were then given 108 S. enterica serovar Typhimurium (ST) strain SL1344 or recombinant S. enterica serovar Typhimurium strain SL1344 OmpC-2W (ST-2W) by intragastric gavage. Enrofloxacin was included in the drinking water at 2 mg/ml in some cases.
ST was tagged chromosomally with the 2W peptide (EAWGALANWAVDSA) as previously described by Uzzau et al. 2001. Briefly, primers were designed with extension arms homologous to the 3’ portion of the OmpC gene, deleting the stop codon, and extending downstream from it. Additionally, a single FLAG sequence was included, for blotting purposes, before reintroduction of the stop codon while kanamycin resistance was introduced downstream of the newly incorporated stop codon. PCR products were generated by amplification from a template plasmid (pJM1) encoding the 2W peptide and FLAG epitope. PCR products were used directly for electrotransformation into ST containing the temperature-sensitive pKD46 plasmid, carrying arabinose inducible bacteriophage λ red genes. Bacterial suspensions in 10% glycerol were mixed with 0.5–1 μg of PCR product and incubated on ice for 30 min before transferring to a chilled 0.2-cm cuvette. Cuvettes were subjected to a single pulse of 12.5 kV/cm. After recovering for 1 hour at 37°C in SOC medium, bacteria were plated on LB agar plates supplemented with 50 μg/ml kanamycin. DNA sequencing was used to verify recombination.
Wild type and ST-2W bacteria were grown overnight in LB Broth. Bacteria were centrifuged and pellets were suspended in 1 ml of 1X CelLytic™ B cell lysis reagent (Sigma), containing 0.2 mg/ml lysozyme (RPI Corp.), 50 U/ml Benzonase® Nuclease (Sigma), and 25 μl of Protease Inhibitor Cocktail (Sigma). Cell extracts were centrifuged and supernatants containing soluble proteins were removed for SDS-PAGE analysis. Separated proteins were transferred to a nitrocellulose membrane then blocked in 5% BSA for 1 hour. Membranes were probed with mouse anti-FLAG M2 monoclonal antibody (Sigma) at 1 μg/ml in 1% BSA for 1 hour at room temperature with gentle agitation. The blot was washed and then probed with 1 μg/ml of a secondary goat anti-mouse conjugated to AlexaFluor-680 (Invitrogen) for 1 hour at room temperature. The blot was analyzed on an Odyssey® Imaging System (Li-Cor) at 700 nm.
Spleen and lymph node cells were harvested and made into single cell suspensions. 10% of the cell suspensions was removed for plating bacterial CFUs with the addition of 0.1% Triton. The remaining 90% of the samples were stained for 1 hour at room temperature with or 2W:I-Ab-streptavidin-allophycoyanin tetramer, enriched for tetramer bound cells, counted and labeled with antibodies as previously described (17). Cells events were collected on an LSRII or Fortessa flow cytometer (Becton Dickinson) and analyzed using FlowJo software (TreeStar).
BrdU (Sigma-Aldrich) was dissolved in PBS at a concentration of 10 mg/ml. In vivo BrdU labeling was performed by intraperitoneal injection of 1 mg of BrdU on day 50 of ST-2W infection. 12 hours later, spleens and MLN were harvested individually and BrdU incorporation into DNA was detected by intracellular staining according to the manufacturer's specifications (BD Biosciences).
ST-2W-infected mice were injected with 100 μg of 2W peptide. Two hours later, spleens and mesenteric lymph node cells were harvested in media containing 10 μg/ml brefeldin A. The resulting cell suspensions were fixed and permeabilized and stained with IFN-γ, TNF, and IL-2 antibodies as previously described (2).
CD4+ T cells were isolated from spleens and lymph nodes of (CD45.1 × 129) F1 mice infected 50 days earlier with ST-2W and purified using Miltenyi CD4 isolation kits. The cells were then washed in EHAA media, suspended to a concentration of 5 × 107 cells/ml, incubated with 5 μM CFSE (Invitrogen) at 37°C for 10 minutes, and washed with pre-warmed EHAA media before intravenous injection into (B6 × 129) F1 hosts. For cytokine experiments, purified CD4+ T cells were washed and injected intravenously without CFSE labeling. To control for experiment-to-experiment variability, the numbers of parked cells were normalized for the input number of donor-derived CD4+ T cells.
Statistical differences between normally distributed data sets were assessed in most cases using paired or unpaired two-tailed Student's t tests with Prizm (Graphpad) software. A Mann-Whitney test was used in one case where the data were not normally distributed (Fig. 4E).
Before monitoring the stability of CD4+ T cells during persistent ST infection, it was necessary to identity a relevant epitope. Although epitopes consisting of I-Ab molecules complexed with peptides from the FliC and SseJ proteins of ST have been described (18), the CD4+ T cell populations that recognize them are very small (17). The largest naïve CD4+ T cell population identified to date is specific for a peptide called 2W (17), which is a variant of MHCII I-E alpha chain peptide 52–68 (19). We therefore used the lambda Red-mediated recombination system to insert the 2W peptide coding sequence at the 3’ end of the OmpC gene in the ST chromosome (20). The resulting organisms, referred to hereafter as ST-2W, express an OmpC-2W fusion protein (Fig. 1A). As shown in Fig. 1B, oral inoculation of ST-2W bacteria into Nramp1R 129 mice resulted in an infection that peaked between 2-3 weeks in the mesenteric lymph nodes (MLN) and spleen, then declined to an undetectable level in the spleen by day 50 but persisted at low levels in MLN for hundreds of days as previously described by Monack and colleagues for wild-type ST organisms (16). Therefore, ST-2W organisms were capable of producing a persistent infection that was controlled by CD4+ T cells.
We next measured the number of 2W:I-Ab-specific cells over time after infection using a p:MHCII tetramer-based cell enrichment approach (21). Cells from MLN and spleens of infected 129 mice were stained with fluorochrome-labeled 2W:I-Ab tetramer and anti-fluorochrome-labeled magnetic beads and enriched on magnetized columns as previously described (17). 2W:I-Ab tetramer-binding cells were detected by flow cytometry among the CD4+ cells that bound to the column.
Uninfected 129 mice each contained about 100 2W:I-Ab-specific CD4+ T cells in MLN and about 300 in spleen. The majority of the cells in both locations were CD44low as expected for naïve T cells (Fig. 2A). A similar number and phenotype of 2W:I-Ab-specific CD4+ T cells was observed in these tissues in 129 mice infected with ST organisms, not expressing the 2W peptide (Fig. 2A). In contrast, 2W:I-Ab-specific CD4+ T cells increased dramatically in number and CD44 expression in MLN and spleen after ST-2W infection (Fig. 2A). 2W:I-Ab-specific T cells present on day 6 and 60 after ST-2W infection expressed large amounts of CD44 (Fig. 2A) and the Th1 lineage-defining transcription factor T-bet (Fig. 2B). Therefore, intragastric ST-2W infection of NrampR mice induced an early homogeneous Th1 population, some of which survived into the persistent phase of the infection.
We next studied the kinetics of 2W:I-Ab-specific T cells following intragastric infection of 129 mice to determine whether a stable cell population was produced. The naïve 2W:I-Ab-specific CD4+ T cell population of about 400 cells proliferated to produce 3×105 cells in MLN and 2×106 cells in the spleen two weeks after infection (Fig. 2C). The number of cells in both locations then underwent an 80% reduction that stabilized around day 40. Thereafter, the number of the 2W:I-Ab-specific CD4+ T cells remained consistent at 3×104 cells in MLN and 3×105 cells in the spleen. The cells in both locations were CD44high T-bethigh cells (Fig. 2B). These results show that ST p:MHCII-specific memory-phenotype cells are stably maintained during persistent ST infection.
A BrdU-labeling experiment was then performed to determine whether proliferation within chronically infected tissues could account for CD4+ T cell stability during the persistent phase of infection. 129 mice were injected with BrdU on day 50 after ST-2W infection, when viable bacteria were detected only in the MLN (Fig. 1B), and lymphoid organs were analyzed for BrdU incorporation 12 hours later as a measure of cellular proliferation. This short labeling period was chosen to maximize the chance that cells were labeled in the location where they divided. As shown in Fig. 3B, 5-10% of the 2W:I-Ab-specific CD4+ T cells located in the MLN labeled with BrdU, whereas only 2-3% of cells in the spleen were labeled. Therefore, the stability of the 2W:I-Ab-specific population correlated with a higher rate of proliferation in the persistently infected MLN.
If the 2W:I-Ab-specific population was maintained by proliferation of a subset of cells in persistently infected MLN, then clearance of the ST-2W bacteria would be expected to result in decline of the population. ST-2W-infected mice were cleared of infection by treatment with enrofloxacin antibiotic for 40 days, beginning at day 28 after ST-2W infection, as a test of this hypothesis. As shown in Fig. 4B, mice not treated with antibiotic had about twice as many 2W:I-Ab-specific T cells in their lymphoid organs on day 68 of ST-2W infection as mice that were treated with antibiotic.
We next determined whether p:MHCII recognition was the aspect of persistent infection that was required for the stability of the 2W:I-Ab-specific population. CFSE labeled 2W:I-Ab-specific cells from the spleens and lymph nodes of persistently infected (CD45.1 × 129) F1 mice were transferred into (B6 × 129) F1 mice with time-matched ST-2W or wild-type ST infections. The number of donor-derived CD45.1+ 2W:I-Ab+ CD4+ T cells was enumerated 1, 7, or 50 days after transfer to assess the stability of the transferred cells. ST-2W bacteria were not transferred along with the T cells, as evidenced by the fact that the 2W:I-Ab-specific cells of endogenous origin in wild-type ST-infected mice, receiving 2W:I-Ab-specific memory-phenotype cells, neither proliferated nor increased CD44 expression (Fig. 4C). As shown in Fig. 4D, the numbers of transferred 2W:I-Ab-specific cells were similar between days 7 and 50 in ST-2W-infected hosts, but were significantly lower on day 50 than on day 7 in hosts infected with wild-type ST bacteria lacking the 2W peptide.
The stability of 2W:I-Ab-specific cells in ST-2W-infected hosts was associated with proliferation. 2W:I-Ab-specific cells residing in ST-2W- or wild-type ST-infected hosts for 50 days underwent 1-3 divisions, probably in response to IL-15 (2). However, donor-derived 2W:I-Ab-specific cells that resided in ST-2W infected hosts for 50 days also contained cells that had divided greater than 7 times. Even though these cells accounted for 70% of the population 50 days after transfer, the fact they doubled at least 7 times during that period means they arose from less than 1% of the initial transferred population. These results are consistent with the possibility that periodic proliferation by a small number of cells in persistently infected MLN accounted for the long-term numerical stability of the 2W:I-Ab-specific cell population.
We then investigated whether persistent ST-2W infection affected the function of 2W:I-Ab-specific T cells. Mice infected with ST-2W 14, 45, or 100 days earlier were injected intravenously with 100 μg of 2W peptide to elicit lymphokine production by 2W:I-Ab-specific T cells, which was detected by direct ex vivo intracellular staining. 2W:I-Ab-specific T cells increased CD69 following 2W peptide injection and produced IFN-γ as expected based on their high expression of T-bet. While IFN-γ is essential for controlling persistent ST infection (16), the best correlates of a protective Th1 response for other infections are CD4+ T cells capable of simultaneously producing IFN-γ, TNF, and IL-2 (triple+) (22, 23). Interestingly, almost all of the 2W:I-Ab-specific cells also expressed TNF and a fraction of these cells also produced IL-2 (Fig. 5A). Thus, 2W:I-Ab-specific T cells were maintained in a highly multifunctional state during persistent ST-2W infection.
We next tested whether the low-level, localized p:MHCII presentation in the MLN of ST-2W infected mice that maintains CD4+ T cell numbers also preserved their function. We found that the overall composition of single (IFN-γ+), double (IFN-γ+, TNF+) and triple lymphokine-producing (IFN-γ+, TNF+, and IL-2+) cells was similar between donor-derived cells that resided for 50 days in ST-2W infected or uninfected hosts (Fig. 6C). These results indicate that periodic stimulation by 2W:I-Ab complexes during the persistent phase of infection did not significantly affect the cytokine production potential of the 2W:I-Ab-specific cells.
This result was surprising in light of studies indicating persistent stimulation by p:MHCI complexes results in functional exhaustion of CD8+ T cells (24). It was possible that 2W:I-Ab-specific cells avoided exhaustion in mice with persistent ST-2W infection because infected antigen presenting cells were too rare. We tested this possibility by assessing the fate of 2W:I-Ab-specific memory-phenotype cells after transfer into (Act-2W × 129) F1 transgenic mice that display 2W:I-Ab throughout the body due to ubiquitous expression of 2W peptide under the control of the Actb promoter (25). Indeed, 2W:I-Ab-specific cells induced by ST-2W infection completely lost the capacity to produce lymphokines in response to peptide challenge after transfer and residence for 50 days in ST-2W-infected (Act-2W × 129) F1 transgenic mice. This loss of function was accompanied by induction of surface PD-1, which is a marker of T cell exhaustion (Fig. 6A) (26). Therefore, 2W:I-Ab-specific cells became functionally exhausted when exposed to systemic 2W:I-Ab complexes.
Previous studies of acute L. monocytogenes and LCMV infections reached the conclusion that p:MHCII-specific CD4+ memory-phenotype T cells slowly decline after infections are cleared (1, 2). The numerical decline of murine CD4+ memory-phenotype T cells without their relevant antigen is at odds with the remarkable stability of CD8+ memory-phenotype T cells, IgM+ memory B cells and plasma cells (1, 27, 28), and challenges the classical definition of immune memory. The stability of the population of 2W:I-Ab-specific T cells in mice with persistent ST-2W infection could result from the periodic proliferation of a few memory-phenotype CD4+ T cells to produce short-lived effector cells coupled with the slow death of non-stimulated memory-phenotype cells. Numerical stability would be achieved in this case despite the fact that neither the short-lived effector cells nor the quiescent memory-phenotype cells are perfectly stable. A similar scenario has been reported for CD8+ T cells induced by chronic LCMV infection (29).
The numerical decline of CD4+ memory-phenotype T cells without p:MHCII presentation may be less surprising when considering the relationship between immune memory and protection. Immune memory is useful to the host only insofar as it contributes to immune protection. CD4+ T cells are not essential for immunity to acute L. monocytogenes and LCMV infections, which are controlled by CD8+ T cells specific for abundant p:MHCI ligands generated by these cytosolic intracellular microbes (5, 6). Thus, the decline of CD4+ T cells comes at no cost to the host in these cases because these cells do not provide a protective advantage. In contrast, in a case where CD4+ T cells play an essential role in controlling a persistent phagosomal ST infection, p:MHCII-specific CD4+ memory-phenotype T cells were numerically stable. The fact that this stability depended on persistent infection and p:MHCII presentation indicates TCR stimulation is the key signal needed for maintenance of CD4+ T cells. Therefore, the classical idea of memory cell stability without antigen does not have to apply to CD4+ T cells in cases where they are protective because the phagosomal infections that they control are persistent, thus providing a “constant TCR reminder” to maintain the T cell population.
In the cases of chronic LCMV, hepatitis C, and HIV infections, prolonged p:MHCI antigen stimulation eventually results in the exhaustion of CD8+ T cells (30). The fact that persistent ST-2W infection and p:MHCII presentation was required to maintain CD4+ T cells without exhaustion is probably related to the low level of peptide:MHCII antigen presentation during persistent ST infection. This is supported by the recent demonstration of CD8+ T cell function in the face of chronic HSV infection in the sensory ganglia (31). The restricted anatomical location of infection could also influence CD4+ T cell function during persistent ST infection. After the first two months after oral ST infection, the MLN were the primary site in the body to harbor viable bacteria. Thus, it is conceivable that CD4+ T cells periodically migrate into sites of low-level persistent infection, proliferate in response to p:MHCII ligands displayed by persistently infected phagocytes, and then migrate to other body sites to recover. This process could produce numerical stability by balancing the death of cells in the population and prevent the cells from becoming exhausted. These ideas are supported by our observations that MLN were preferential sites of CD4+ T cell proliferation in persistently infected mice and exposure to systemic p:MHCII complexes resulted in CD4+ T cells that could not produce lymphokines and expressed high levels of PD-1, the canonical marker for exhausted cells (30). Thus, although the lack of cytokine production by these cells could be due to an indirect mechanism like regulatory T cell suppression, it more likely relates to the direct mechanism of T cell exhaustion.
Persistent low-level p:MHCII presentation by phagocytes during ST infection was also associated with homogenous production of Th1 cells, unlike transient p:MHCII presentation during acute infections, which generate a mixture of Th1 cells and follicular helper cells (3, 32). The reason that follicular helper cells were not maintained during persistent ST infection could relate to lack of p:MHCII presentation by B cells (33).
Given the impressive lymphokine production capacity of 2W:I-Ab-specific cells in the face of persistent ST-2W infection, it was possible that periodic TCR stimulation was actually required to maintain multifunctional Th1 cells with maximal IFN-γ production. However, our results demonstrate that T-bet expression and IFN-γ production capacity was maintained at a high level in CD4+ memory-phenotype T cells that were parked in antigen-free hosts. However, a loss in the number of highly functional Th1 cells would be expected to reduce the capacity of CD4+ T cells to control a subsequent phagosomal ST infection. Such a loss could explain the concomitant immunity phenomenon studied in Leishmania-infected individuals. In this case, IFN-γ producing Th1 cells control the infection within phagocytes at the initial site of infection and prevent it from spreading to other parts of the body. Oddly however, the Th1 cells never eliminate the microbes from the initial site. Indeed, persistent infection at the original site is required for the Th1 cells to eliminate bacteria from other body sites after a second infection (11). Our results indicate that localized persistent infection and p:MHCII presentation is required to maintain a numerically stable CD4+ T cell population with maximal IFN-γ production capacity needed to control the infection at the initial site and eliminate it from other sites.
The fact that CD4+ T cells control persistent ST infection in phagocytes and presumably depend on p:MHCII presentation by these cells to be maintained as maximally functional cells and yet cannot eliminate the infection, is perplexing. One possibility is regulatory T cells within MLN restrain microbicidal activities of CD4+ effector T cells in the same location (34, 35). Alternatively, persistently infected phagocytes may produce IL-10, which can ameliorate some of the phagocyte-activating effects of IFN-γ made by CD4+ T cells engaging in cognate interactions.
Our results suggest that a CD4+ T cell-dependent vaccine for a phagosomal pathogen will have to produce a local and long-lasting depot of bacterial antigen to maintain the protective CD4+ T cell population, which would otherwise decline. The failure to maintain protective CD4+ T cells could explain why the effectiveness of the Bacillus Calmette-Guérin vaccine against tuberculosis wanes after the vaccine organisms are cleared from the body. To improve on this situation, it may be necessary to produce even longer-lasting antigen delivery systems, perhaps along the lines of those used for long term administration of contraceptives.
We thank J. Walter and R. Speier for technical assistance.
1This work was supported by grants from the US National Institutes of Health, R37-AI027998, R01-AI39614, R01-AI66016 (M.K.J.), and T32-GM008244, F30-DK093242 (R.W.N.).