FliC-Specific CD4+ T Cell Responses Are Restricted by Bacterial Regulation of Antigen Expression1
http://www.100md.com
免疫学杂志 2005年第12期
Abstract
Salmonella typhimurium, a facultatively intracellular pathogen, regulates expression of virulence factors in response to distinct environments encountered during the course of infection. We tested the hypothesis that the transition from extra- to intracellular environments during Salmonella infection triggers changes in Ag expression that impose both temporal and spatial limitations on the host T cell response. CD4+ T cells recovered from Salmonella immune mice were propagated in vitro using Ag derived from bacteria grown in conditions designed to emulate extra- or intracellular environments in vivo. Extracellular phase bacteria supported a dominant T cell response to the flagellar subunit protein FliC, whereas intracellular phase bacteria were unable to support expansion of FliC-specific T cells from populations known to contain T cells with reactivity to this Ag. This result was attributed to bacterial regulation of FliC expression: transcription and protein levels were repressed in bacteria growing in the spleens of infected mice. Furthermore, Salmonella-infected splenocytes taken directly ex vivo stimulated FliC-specific T cell clones only when intracellular FliC expression was artificially up-regulated. Although it has been suggested that a microanatomical separation of immune T cells and infected APC exists in vivo, we demonstrate that intracellular Salmonella can repress FliC expression below the T cell activation threshold. This potentially provides a mechanism for intracellular Salmonella at systemic sites to avoid detection by Ag-specific T cells primed at intestinal sites early in infection.
Introduction
The pathogen Salmonella enterica serovar typhimurium (S. typhimurium) is a facultatively intracellular pathogen capable of replication both outside and inside host cells. After ingestion by the host, Salmonella travel to the small intestine, penetrate the intestinal epithelium via M cells (1), and disseminate to systemic sites such as the spleen and liver to replicate within host cells (2, 3). Alternative pathways of dissemination have also been identified (4), possibly occurring as dendritic cells (DC)3 sample the gut lumen directly (5). To survive and replicate in the diverse host environments encountered during infection, Salmonella have the ability to regulate the expression of numerous genes in response to specific environmental conditions (6). For example, the two-component regulatory system PhoP/PhoQ activates or represses expression of over 40 genes in response to the environment encountered in the phagosomes of host cells (7, 8) and is essential for virulence (9, 10). In addition, many virulence genes are clustered into coordinately regulated Salmonella pathogenicity islands (SPIs) (11). SPI1 genes are required for invasion by extracellular bacteria (12), and SPI2 genes encode one set of gene products required for intracellular replication and systemic infection (13). SPI1 and SPI2 genes are regulated in response to a variety of environmental signals including, but not limited to osmolarity (14, 15, 16), pH (14, 15, 17), Mg2+ depletion (7), oxygen tension (14), phosphate starvation (18), bile (19), cationic microbial peptides (20), and intestinal short-chain fatty acids (21). In fact, it has been estimated that 20% of all Salmonella coding sequences, including those required for metabolism, ion transport, resistance to antimicrobial peptides, and motility, are differentially regulated by Salmonella during growth in macrophages (22).
To counter bacterial invasion, dissemination, and intracellular replication, the host deploys a variety of immune defense mechanisms. The intestinal epithelium presents a significant obstacle to infection; bacterial invasion induces a strong proinflammatory response in which cytokine release is followed by recruitment and activation of neutrophils, macrophages, and DCs (23). Adaptive immune responses are critical for clearance of Salmonella from the host; B cells contribute to protective immunity via Ab production and the ability to serve as APC (24), and both CD4+ and CD8+ T cell responses are generated during Salmonella infection (25, 26). CD4+ T cells play an essential role in adaptive immunity, and protection can be transferred to naive susceptible mice with CD4+ cells and sera from immune mice (27). Although the importance of CD4+ T cells in immune clearance of Salmonella infection is well established, the location and kinetics of T cell activation are only beginning to be understood.
The flagellar subunit protein FliC is a well-characterized target of host immune responses, and immunization of mice with purified FliC can protect against challenge with virulent organisms (28, 29). FliC stimulates innate immune responses via TLR5 (30, 31), and is a major proinflammatory component of bacterial invasion (32, 33). Flagellin elicits Ab production (29, 34) and has been previously identified as an important recall Ag for CD4+ T cells from protectively immunized mice (35) and humans (36). In addition, FliC expression is regulated by the PhoP/PhoQ regulon (22, 37). Together these data suggest the interesting possibility that FliC expression may be limited to early stages of infection before dissemination and intracellular replication (38, 39). Indeed, recent studies in mice using an adoptive transfer model have shown that FliC-specific T cells are activated only at intestinal sites during oral Salmonella infection, despite the presence of large numbers of bacteria at systemic locations such as the liver and spleen (40). Given the complex regulation of Salmonella virulence determinants during infection, we hypothesized that differential regulation of Ag expression could account for these observations.
In this study, we compared extracellular with intracellular phase Salmonella for the ability to stimulate CD4+ T cells recovered from protectively immunized mice. We demonstrate that the transition from extra- to intracellular environments triggers changes in bacterial Ag expression that directly affect recognition by Salmonella-specific T cells, and which could therefore impose temporal and spatial restrictions on activation of host immune responses. We present a model in which repression of Ag expression below the T cell activation threshold restricts stimulation of host immune cells and provides virulent Salmonella with one mechanism of escaping immune surveillance by Ag-specific CD4+ T cells.
Materials and Methods
For intracellular induction of FliC expression, S. typhimurium strain 14028 (wild type), FliC-null strain 14028 (BC698), and FliC-null strain 14028 carrying plasmid pMalEFliC (35) in which FliC expression is controlled by an isopropyl -D-thiogalactoside (IPTG)-inducible promoter (BC898) were used for infection experiments. Bone marrow-derived DCs isolated in the presence of GM-CSF according to standard methods (42) were infected at a multiplicity of infection of 100:1 with bacteria grown in N-minimal medium containing 8 μM Mg2+ (44), washed to remove extracellular bacteria after 2 h, and incubated an additional 3 h in medium containing 15 μg/ml gentamicin (and 100 μg/ml ampicillin if needed for plasmid maintenance) before use as APC. FliC expression was induced by addition of 2 mM IPTG to medium during incubation with gentamicin. For in vivo induction, 10 mM IPTG was added to mouse drinking water 3 days after i.p. infection (45), and spleens were harvested 24–48 h later. Low density splenocytes were isolated by brief incubation with collagenase, followed by separation in dense BSA gradient (42), and used as APC as described. Mice received 2 mg/ml ampicillin in drinking water for 48 h before, as well as throughout the duration of infection as required for plasmid maintenance. Presence of plasmid in strains recovered from infected tissues was confirmed by replica plating onto media containing ampicillin, and the CFU to APC ratio for all experiments ranged from 0.2 to 4.0.
Results
Differential stimulation of Ag-specific CD4+ T cells by extracellular and intracellular phase Salmonella
Salmonella replication, either extracellularly or inside host cells, requires coordinate regulation of many genes (6). We hypothesized that alteration of gene expression in response to changing environments during infection would impact production of bacterial Ags, and influence the antigenic specificity of responding CD4+ T cell populations. Therefore, we designed reagents that model bacterial Ag expression during extracellular or intracellular bacterial growth. Extracellular phase Ags (STEX) were prepared by growing S. typhimurium strain 14028 under conditions known to stimulate expression of proteins required for intestinal host cell invasion (14). Intracellular phase Ags (STIN) were approximated using a genetic approach; because the bacterial PhoP/PhoQ regulatory system is activated during growth inside host cells (7, 8), we used a mutant that over-expresses PhoP in its active, phosphorylated state (PhoPc). STIN and STEX Ags were processed and presented by naive splenocyte APC in assays further described below. A third strategy, using splenocytes from Salmonella-infected mice (IS) as APC directly ex vivo, allowed propagation of T cells responding to Ags expressed, processed, and presented only in vivo. Naive mice were orally infected with live, attenuated bacteria (S. typhimurium SL3261) to generate Salmonella-specific memory CD4+ T cells. These cells, primed throughout the course of infection by bacteria growing extracellularly in the gastrointestinal (GI) tract and intracellularly at systemic locations, were recovered from mice 60 days after immunization and used as input cells to generate T cell lines TEX, TIN, and TIS by selective amplification with STEX, STIN, or IS Ags, respectively (Fig. 1A).
After extracellular bacteria in the gut penetrate the intestinal epithelium (1) or are taken up by phagocytic cells directly from the lumen (5), intracellular Salmonella spread (4) and replicate at systemic sites such as the liver and spleen (2, 3). To test the hypothesis that extra- and intracellular phase bacteria stimulate different populations of immune CD4+ T cells, we examined the responses of immune T cell lines to intact, extracellular phase bacterial Ag (STEX) (Fig. 1B). CD4+ T cell line TEX had a 100-fold greater response to STEX as compared with T cell lines TIN and TIS. This indicated that Ag-specific T cells, presumably stimulated by Ags expressed only by extracellular phase bacteria, were present in line TEX but absent from lines TIN and TIS. In addition, the nearly identical responses of lines TIN and TIS indicated that T cells selectively amplified in response to intracellular phase bacteria grown in vitro (TIN) resemble those T cells responding to Ags expressed by bacteria growing intracellularly in vivo (TIS). These results suggested that the ability of bacteria to alter Ag expression during infection could affect immune recognition by the host.
Intracellular phase bacteria fail to support expansion of FliC-specific T cell populations in vitro
To more directly compare the antigenic specificities of T cell populations in each CD4+ T cell line, we tested the ability of SDS-PAGE fractionated STEX or STIN bacteria to stimulate immune T cells (Fig. 2A–F). As predicted by a previous study (35), line TEX (Fig. 2, first row), derived by selective amplification using STEX bacteria as Ag, responded to the fraction from STEX bacteria containing FliC (fraction 20, Fig. 2A). In contrast, line TEX displayed very weak responses to the corresponding fraction from STIN bacteria (fraction 20, Fig. 2B). Consistent with these results, STIN bacteria expressed substantially less FliC than STEX bacteria, as confirmed by Western blot analysis using anti-FliC Abs (Fig. 2, G and H). The reduced level of FliC expressed by STIN bacteria failed to support expansion of FliC-specific T cells from memory T cell populations, as T cell line TIN (Fig. 2, second row), derived by selective amplification using STIN bacteria as Ag, did not respond to any FliC-containing fraction (fraction 20, Fig. 2, C and D) or purified FliC protein (Fig. 2I).
Surprisingly, CD4+ T cell line TIS (Fig. 2, third row), derived by selective amplification with Salmonella-infected splenocytes as APC, responded to fractions 20 and 21 from STEX bacteria (Fig. 2E). However, equivalent responses of T cell line TIS to the corresponding fractions from STIN bacteria (fractions 20 and 21, Fig. 2F), which expressed markedly less FliC than STEX (Fig. 2, A, B, G, and H), suggested that TIS was responding to non-FliC Ags. This was confirmed using SDS-PAGE fractionated FliC-null bacteria, which contained Ags stimulatory for T cell line TIS that were in the approximate molecular mass range of FliC (50 kDa, Fig. 2J). The completely negative response of T cell line TIS to purified FliC protein (Fig. 2I) confirmed the absence of FliC-specific T cells in line TIS. The strong response of T cell line TEX to the FliC-containing fraction of STEX bacteria (Fig. 2A) and purified FliC (Fig. 2I) confirmed that FliC-specific CD4+ T cells were present in the original input responder population used to generate T cell lines TEX and TIS (see also Fig. 1). We therefore conclude that the expansion of FliC-specific T cell populations in vitro is restricted by the reduced FliC expression of intracellular phase bacteria.
Transcription of the gene encoding FliC is repressed during Salmonella growth inside macrophage phagosomes
Because FliC expression by STIN bacteria is below the threshold required for expansion of FliC-specific T cells in vitro (Fig. 2, C, D, and I), and maximal PhoP/PhoQ activation occurs in macrophage phagosomes (7, 8), we predicted that repression of fliC transcription during bacterial growth inside host cells would be PhoP/PhoQ-dependent. Therefore, we used GFP fluorescence to report transcription of fliC (PfliC::gfp) during Salmonella infection of macrophages. Wild-type Salmonella grown in vitro displayed equivalent fluorescence from either PfliC::gfp or Ppos::gfp (constitutive promoter) (Fig. 3, A and B). However, wild-type Salmonella inside of macrophages (Fig. 3, D, E, G, and H) repressed transcription of PfliC::gfp (Fig. 3E), whereas Ppos::gfp was constitutively expressed (Fig. 3D). In contrast, intracellular PhoP-null Salmonella, which lack a functional PhoP/PhoQ regulatory system (Fig. 3I) failed to repress PfliC::gfp (Fig. 3F). These data confirm that PhoP/PhoQ-dependent repression of fliC promoter activity occurs during bacterial growth in the macrophage phagosome.
Reduced expression of FliC in vivo limits activation of FliC-specific CD4+ T cells
Our observation that intracellular Salmonella repress expression of FliC, together with the fact that only extracellular phase bacteria are capable of supporting the expansion of FliC-reactive T cells in vitro (Fig. 2, A, B, and I), suggests that in vivo, FliC expression by intracellular Salmonella will fall below the T cell activation threshold. To confirm that FliC expression during bacterial growth in vivo is reduced, we used GFP fluorescence to report fliC transcription (PfliC::gfp) during Salmonella infection in mice. The absence of fluorescence indicated silencing of transcription from the fliC promoter in bacteria recovered from spleens of infected mice, although the same bacteria expressed GFP when grown in vitro (Fig. 4A, left column). In contrast, fluorescent bacteria were detected both in vivo and in vitro when a constitutive promoter controlled GFP expression (Fig. 4A, middle column). These results were confirmed by Western blot analysis: detection of bacterial DnaK verifies the presence of Salmonella, but FliC is undetectable in spleens of Salmonella-infected mice (Fig. 4B).
Because Salmonella growing in vivo repress FliC production, we predicted that splenocyte APC from Salmonella-infected mice would fail to stimulate FliC-specific CD4+ T cells. Therefore, splenocytes from mice infected with wild type or FliC-null Salmonella were used as APC directly ex vivo and compared with uninfected APC for their ability to stimulate Salmonella-specific T cell clones (Fig. 4C). APC from infected mice failed to stimulate FliC-specific T cells, however the T cells vigorously responded to exogenously added heat-killed FliC+ Salmonella presented by uninfected APC (Fig. 4C). As a control, infected APC or naive APC presenting exogenously added bacteria were equally stimulatory to a Salmonella-specific CD4+ T cell clone that recognizes both STIN and STEX bacteria (Fig. 4C). These results confirm that Salmonella growing in vivo at systemic sites specifically repress expression of FliC below the T cell activation threshold.
Artificial induction of FliC expression in vivo results in stimulation of FliC-specific T cells
If failure of FliC-specific T cells to detect and respond to bacteria growing in the spleens of infected mice is due to down-regulation of Ag expression, artificially increasing production of FliC during infection should render Salmonella-infected APCs stimulatory for FliC-specific T cells. Therefore, we performed a series of experiments using a strain in which an IPTG-inducible promoter controlled FliC expression. Bone marrow-derived DCs were infected in vitro with wild-type Salmonella (natural FliC regulation) or with BC898 (IPTG-regulated FliC expression) as described in Materials and Methods, and tested for the ability to stimulate Salmonella-specific CD4+ T cell clones (Fig. 5, A and B). In the absence of IPTG, FliC-specific T cell stimulation by infected APC did not differ from uninfected APC; after IPTG induction of bacterial FliC expression the response was similar to that stimulated by the presentation of heat-killed FliC+ Salmonella by uninfected APC (Fig. 5A). As before, these infected APC were capable of processing and presenting Ag to stimulate control T cell responses (Fig. 5B). Similar results were obtained during growth in host tissues when induction of FliC expression was performed in vivo (Fig. 5, C and D). Mice infected with BC898 (IPTG-regulated FliC expression) were provided with IPTG in their drinking water (45) and low-density splenocytes recovered from these infected mice were interrogated directly ex vivo for their ability as APC to stimulate Salmonella-specific T cell clones. FliC-specific T cells responded to infected APC only after IPTG exposure in vivo; without IPTG treatment, stimulation of T cells by infected APC was no different than with uninfected APC (Fig. 5C). Similar numbers of bacteria were recovered from infected spleens regardless of IPTG treatment (data not shown). These results confirm that exogenous up-regulation of FliC expression during intracellular bacterial replication allows FliC-specific CD4+ T cells to detect Salmonella-infected APC. Therefore we conclude that repression of FliC expression below the T cell activation threshold is sufficient to account for the absence of FliC-specific CD4+ T cell activation in liver and spleen observed during orogastric Salmonella infection (40). Taken together, our data support the hypothesis that bacterial regulation of Ag expression imposes both temporal and anatomical limits on host T cell responses: opportunities for T cell activation and subsequent effector function are restricted to those anatomical sites and stages of infection when bacterial expression of stimulatory Ag occurs.
Discussion
Infection with Salmonella results in a complex and dynamic interaction between pathogen and host. In this study we show that intracellular phase Salmonella repress expression of the protective Ag FliC below the threshold required for T cell activation, and that this repression is mediated by PhoP/PhoQ. We propose that during infection of a naive host with virulent Salmonella, the transition from extra- to intracellular environments triggers changes in Ag expression that subsequently influence both the location and timing of critical adaptive immune responses (Fig. 6 and see below).
Flagella, and the flagellar subunit FliC, are abundantly expressed on the surface of extracellular phase Salmonella and therefore serve as a large reservoir of ligands for both innate (30, 31) and adaptive (35) immune systems (see also Fig. 2). After bacterial invasion or translocation of flagellin across the intestinal epithelial cell layer, TLR5 stimulation results in a strong proinflammatory response (Fig. 6A) that is not induced by normal microbiota (32, 33, 46). Flagellin-stimulated secretion of chemokines such as IL-8 (31, 46, 47) and CCL20 (48) results in the recruitment of phagocytes, including DCs, to the small intestine. DCs are highly competent APCs (49) that can directly sample the gut lumen for Ags they present to immune effector cells (5). In addition, flagellin stimulates the maturation of both DCs and macrophages (50, 51). Thus, the proinflammatory response to flagellin, in combination with the proinflammatory response evoked by Salmonella-induced cell death (52), results in intestinal localization and activation of cells capable of priming adaptive immune responses (Fig. 6B). In our model, DCs (and possibly macrophages) in the Peyer’s patches initially prime CD4+ T cell responses to Ags, such as FliC, that are expressed by bacteria invading the host intestinal epithelium or captured from the gut lumen (STEX Ags) (Fig. 6C).
After uptake by phagocytes, a small number of bacteria disseminate inside host cells (4) (Fig. 6D) to systemic sites such as the liver and spleen, where intracellular bacteria repress expression of Ags such as FliC below the T cell activation threshold (Fig. 4C). The natural consequence of this repression is that Ags expressed only at early stages in infection (STEX Ags), before intracellular dissemination and replication, will prime CD4+ T cell responses that fail to recognize intracellular bacteria growing at systemic sites during later stages of infection (Fig. 6E). This idea is supported by oral Salmonella infections in which adoptively transferred FliC-specific T cells were activated at intestinal but not systemic sites, despite the presence of bacteria in all host tissues (40). A microanatomical separation of T cells away from APC was proposed to account for these observations, however our data indicate that bacterial Ag regulation represents an alternative explanation. Harvesting infected splenocytes disrupts splenic architecture, and yet these APCs fail to stimulate FliC-specific T cells (Figs. 4C and 5, A and C). The recovery of stimulatory capacity after exogenous up-regulation of FliC expression (Fig. 5, A and C) reinforces this conclusion. Thus, the repression of Ag expression below the T cell activation threshold during intracellular growth may functionally contribute to the ability of Salmonella to establish chronic carrier status in mice (53) and humans (54).
Flagellin is a potent stimulator of both innate and adaptive immune responses, and the expression of flagella is important for chemotaxis and invasion (reviewed in Ref. 55). In addition, a functional flagellar apparatus may also serve to facilitate bacterial pathogenesis independent of motility by altering expression or secretion of other virulence factors (56, 57, 58, 59, 60). It is noteworthy that studies of diverse pathogens often reveal coregulation of FliC with other virulence factors in response to environmental conditions (22, 61, 62, 63, 64). Through coordinate regulation of multiple genes, bacteria are able to precisely tailor expression of virulence factors and successfully infect their hosts. The global virulence regulon PhoP/PhoQ, activated by intracellular growth conditions (7, 8), is necessary for bacterial surface modifications required to resist innate immune defenses (65, 66). Our data demonstrate that Ag repression by intracellular bacteria occurs by a PhoP/PhoQ-mediated mechanism (Fig. 3) and activation of the PhoP/PhoQ regulon represses expression of surface Ags recognized by both innate and adaptive immune responses of the host (37). Many Salmonella mutants with dysregulated FliC expression have altered virulence (9, 63, 67). Furthermore, when a PhoP-activated promoter controls expression of FliC, virulence after oral infection of mice is decreased (our unpublished observations). Therefore, the linking of FliC expression to that of other virulence factors by PhoP/PhoQ regulation (in concert with other mechanisms) allows bacteria to maximize the contributions of flagellin to the establishment of infection while minimizing the availability of this Ag to stimulate host responses. It is significant that limiting host access to bacterial flagellin is a strategy used by many bacteria. For example, Helicobacter produces a flagellin that is inherently less stimulatory than that of other organisms, and limits secretion of flagellin monomer into culture supernatants (68). Many species, including Vibrio and Helicobacter cloak their flagella in an outer membrane sheath (69, 70). Finally, Pseudomonas strains recovered from cystic fibrosis patients with chronic infections have lost motility and are nonflagellated (71).
Our data support the hypothesis that during oral Salmonella infection of the naive host the initial T cell response is misdirected: T cells primed to respond to Ags expressed early in infection may fail to recognize intracellular bacteria growing at systemic sites. Thus, the timely development of host immune responses is critical for control of Salmonella infection. This is illustrated by the increased susceptibility of TLR4-deficient mice to Salmonella infection, in which bacterial burden increases as a result of delayed immune responses (72). It may not be necessary for a pathogen to completely abrogate host immune responses if delaying the response allows bacteria to multiply beyond the capacity of host defenses. We propose that repression of protective Ags like FliC, in concert with other genetically co-coordinated modifications of the bacterial surface (37), could facilitate bacterial growth in vivo by affording virulent Salmonella the opportunity for additional rounds of replication before the onset of effective adaptive immune recognition. A similar "decoy and delay" strategy is used during successful CMV infection: the dominant CD8+ T cell response is directed toward an epitope not presented in infected tissues in vivo (73).
The importance of timely host immune responses is underscored by the different outcomes for naive, susceptible mice when infected with virulent bacteria as compared with attenuated bacteria. AroA-attenuated Salmonella grow more slowly in vitro and in vivo (Ref. 74 and our unpublished observations) and are cleared by immunocompetent hosts, whereas virulent Salmonella rapidly overwhelm host defenses. The growth advantage of virulent bacteria is eliminated by prior immunization, as immune host T and B cell functions provide protection (75, 76). Interestingly, memory T cells express CCR6, the only known receptor for CCL20, a chemokine secreted by intestinal epithelial cells after stimulation by flagellin (77). Furthermore, Ag presented by DCs in the Peyer’s patches primes memory T cells that preferentially home to the gut (78). Therefore, in an immunized host it appears that memory T cells, primed to respond to early phase (STEX) Ags such as FliC, are recruited to the intestine where their resulting activation could facilitate the containment of bacterial replication and dissemination.
The gut mucosa presents a major barrier to Salmonella dissemination after oral infection. In addition to overcoming physical barriers such as the glycocalyx and mucus layers, Salmonella must also withstand chemical attacks from antimicrobial peptides secreted by the intestinal mucosa (79). Therefore, it is not surprising that most Salmonella do not progress farther than the intestinal lumen but are contained by host defenses in the small intestine or are eliminated. Although high doses of Salmonella (many times the oral LD50) have been used in experimental models of oral infection, near the oral LD50 only a few bacteria successfully escape to systemic circulation and colonization (80, 81). Based on the fact that FliC-specific CD4+ T cell activation occurs only at intestinal sites after oral infection (40), together with our observations that Salmonella express FliC only during extracellular phase growth (Fig. 2, A and G), we propose that priming of FliC-specific T cells must occur in the intestine where DCs, potent T cell activators, are recruited and activated by flagellin-mediated mechanisms. Therefore, the bacteria in the gastrointestinal tract, most of which will never reach systemic sites of replication, serve as a rich source of Ag for DC capture and subsequent activation of CD4+ T cells responding to STEX Ags (Fig. 6).
Finally, it is significant that intracellular Salmonella, although repressing the expression of Ags such as FliC, are not completely invisible to surveillance by CD4+ T cells derived from immune hosts (Figs. 4 and 5). This observation suggests that priming of CD4+ T cells specific for Ags expressed by virulent intracellular bacteria in the naive host occurs beyond the time limit for a successful adaptive response (Fig. 6, F and G). It is possible that CD4+ T cells recognizing intracellular phase bacteria respond to Ags that are expressed exclusively in vivo, or to Ags that are critical for bacterial survival and thus cannot be repressed. Based on our model, it is reasonable to predict that altered expression of STIN Ags could influence the final outcome of the host-pathogen interaction. For example, over-expression of STIN Ags at the extracellular phase of growth could facilitate early stage priming of potentially protective CD4+ T cell responses. In addition, purified STIN Ags may prove useful as vaccine candidates. Our model predicts that identification of these Ags and characterization of CD4+ T cell responses to them will provide additional useful insights into the relationship between Salmonella and its hosts.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by National Institutes of Health Grant AI47242.
2 Address correspondence and reprint requests to Dr. Brad T. Cookson, Departments of Laboratory Medicine and Microbiology, Box 357110, University of Washington Medical Center, 1959 N.E. Pacific, Seattle, WA 98195. E-mail address: cookson{at}u.washington.edu
3 Abbreviations used in this paper: DC, dendritic cell; SPI, Salmonella pathogenicity island; IPTG, isopropyl -D-thiogalactoside.
Received for publication February 15, 2005. Accepted for publication April 6, 2005.
References
Jepson, M. A., M. A. Clark. 2001. The role of M cells in Salmonella infection. Microbes Infect. 3: 1183-1190.
Richter-Dahlfors, A., A. M. Buchan, B. B. Finlay. 1997. Murine salmonellosis studied by confocal microscopy: Salmonella typhimurium resides intracellularly inside macrophages and exerts a cytotoxic effect on phagocytes in vivo. J. Exp. Med. 186: 569-580.
Salcedo, S. P., M. Noursadeghi, J. Cohen, D. W. Holden. 2001. Intracellular replication of Salmonella typhimurium strains in specific subsets of splenic macrophages in vivo. Cell. Microbiol. 3: 587-597.
Vazquez-Torres, A., J. Jones-Carson, A. J. Baumler, S. Falkow, R. Valdivia, W. Brown, M. Le, R. Berggren, W. T. Parks, F. C. Fang. 1999. Extraintestinal dissemination of Salmonella by CD18-expressing phagocytes. Nature 401: 804-808.
Rescigno, M., G. Rotta, B. Valzasina, P. Ricciardi-Castagnoli. 2001. Dendritic cells shuttle microbes across gut epithelial monolayers. Immunobiology 204: 572-581.
Lucas, R. L., C. A. Lee. 2000. Unravelling the mysteries of virulence gene regulation in Salmonella typhimurium. Mol. Microbiol. 36: 1024-1033.
Groisman, E. A.. 2001. The pleiotropic two-component regulatory system PhoP-PhoQ. J. Bacteriol. 183: 1835-1842.
Alpuche Aranda, C. M., J. A. Swanson, W. P. Loomis, S. I. Miller. 1992. Salmonella typhimurium activates virulence gene transcription within acidified macrophage phagosomes. Proc. Natl. Acad. Sci. USA 89: 10079-10083.(Lisa A. Cummings*, Sara L)
Salmonella typhimurium, a facultatively intracellular pathogen, regulates expression of virulence factors in response to distinct environments encountered during the course of infection. We tested the hypothesis that the transition from extra- to intracellular environments during Salmonella infection triggers changes in Ag expression that impose both temporal and spatial limitations on the host T cell response. CD4+ T cells recovered from Salmonella immune mice were propagated in vitro using Ag derived from bacteria grown in conditions designed to emulate extra- or intracellular environments in vivo. Extracellular phase bacteria supported a dominant T cell response to the flagellar subunit protein FliC, whereas intracellular phase bacteria were unable to support expansion of FliC-specific T cells from populations known to contain T cells with reactivity to this Ag. This result was attributed to bacterial regulation of FliC expression: transcription and protein levels were repressed in bacteria growing in the spleens of infected mice. Furthermore, Salmonella-infected splenocytes taken directly ex vivo stimulated FliC-specific T cell clones only when intracellular FliC expression was artificially up-regulated. Although it has been suggested that a microanatomical separation of immune T cells and infected APC exists in vivo, we demonstrate that intracellular Salmonella can repress FliC expression below the T cell activation threshold. This potentially provides a mechanism for intracellular Salmonella at systemic sites to avoid detection by Ag-specific T cells primed at intestinal sites early in infection.
Introduction
The pathogen Salmonella enterica serovar typhimurium (S. typhimurium) is a facultatively intracellular pathogen capable of replication both outside and inside host cells. After ingestion by the host, Salmonella travel to the small intestine, penetrate the intestinal epithelium via M cells (1), and disseminate to systemic sites such as the spleen and liver to replicate within host cells (2, 3). Alternative pathways of dissemination have also been identified (4), possibly occurring as dendritic cells (DC)3 sample the gut lumen directly (5). To survive and replicate in the diverse host environments encountered during infection, Salmonella have the ability to regulate the expression of numerous genes in response to specific environmental conditions (6). For example, the two-component regulatory system PhoP/PhoQ activates or represses expression of over 40 genes in response to the environment encountered in the phagosomes of host cells (7, 8) and is essential for virulence (9, 10). In addition, many virulence genes are clustered into coordinately regulated Salmonella pathogenicity islands (SPIs) (11). SPI1 genes are required for invasion by extracellular bacteria (12), and SPI2 genes encode one set of gene products required for intracellular replication and systemic infection (13). SPI1 and SPI2 genes are regulated in response to a variety of environmental signals including, but not limited to osmolarity (14, 15, 16), pH (14, 15, 17), Mg2+ depletion (7), oxygen tension (14), phosphate starvation (18), bile (19), cationic microbial peptides (20), and intestinal short-chain fatty acids (21). In fact, it has been estimated that 20% of all Salmonella coding sequences, including those required for metabolism, ion transport, resistance to antimicrobial peptides, and motility, are differentially regulated by Salmonella during growth in macrophages (22).
To counter bacterial invasion, dissemination, and intracellular replication, the host deploys a variety of immune defense mechanisms. The intestinal epithelium presents a significant obstacle to infection; bacterial invasion induces a strong proinflammatory response in which cytokine release is followed by recruitment and activation of neutrophils, macrophages, and DCs (23). Adaptive immune responses are critical for clearance of Salmonella from the host; B cells contribute to protective immunity via Ab production and the ability to serve as APC (24), and both CD4+ and CD8+ T cell responses are generated during Salmonella infection (25, 26). CD4+ T cells play an essential role in adaptive immunity, and protection can be transferred to naive susceptible mice with CD4+ cells and sera from immune mice (27). Although the importance of CD4+ T cells in immune clearance of Salmonella infection is well established, the location and kinetics of T cell activation are only beginning to be understood.
The flagellar subunit protein FliC is a well-characterized target of host immune responses, and immunization of mice with purified FliC can protect against challenge with virulent organisms (28, 29). FliC stimulates innate immune responses via TLR5 (30, 31), and is a major proinflammatory component of bacterial invasion (32, 33). Flagellin elicits Ab production (29, 34) and has been previously identified as an important recall Ag for CD4+ T cells from protectively immunized mice (35) and humans (36). In addition, FliC expression is regulated by the PhoP/PhoQ regulon (22, 37). Together these data suggest the interesting possibility that FliC expression may be limited to early stages of infection before dissemination and intracellular replication (38, 39). Indeed, recent studies in mice using an adoptive transfer model have shown that FliC-specific T cells are activated only at intestinal sites during oral Salmonella infection, despite the presence of large numbers of bacteria at systemic locations such as the liver and spleen (40). Given the complex regulation of Salmonella virulence determinants during infection, we hypothesized that differential regulation of Ag expression could account for these observations.
In this study, we compared extracellular with intracellular phase Salmonella for the ability to stimulate CD4+ T cells recovered from protectively immunized mice. We demonstrate that the transition from extra- to intracellular environments triggers changes in bacterial Ag expression that directly affect recognition by Salmonella-specific T cells, and which could therefore impose temporal and spatial restrictions on activation of host immune responses. We present a model in which repression of Ag expression below the T cell activation threshold restricts stimulation of host immune cells and provides virulent Salmonella with one mechanism of escaping immune surveillance by Ag-specific CD4+ T cells.
Materials and Methods
For intracellular induction of FliC expression, S. typhimurium strain 14028 (wild type), FliC-null strain 14028 (BC698), and FliC-null strain 14028 carrying plasmid pMalEFliC (35) in which FliC expression is controlled by an isopropyl -D-thiogalactoside (IPTG)-inducible promoter (BC898) were used for infection experiments. Bone marrow-derived DCs isolated in the presence of GM-CSF according to standard methods (42) were infected at a multiplicity of infection of 100:1 with bacteria grown in N-minimal medium containing 8 μM Mg2+ (44), washed to remove extracellular bacteria after 2 h, and incubated an additional 3 h in medium containing 15 μg/ml gentamicin (and 100 μg/ml ampicillin if needed for plasmid maintenance) before use as APC. FliC expression was induced by addition of 2 mM IPTG to medium during incubation with gentamicin. For in vivo induction, 10 mM IPTG was added to mouse drinking water 3 days after i.p. infection (45), and spleens were harvested 24–48 h later. Low density splenocytes were isolated by brief incubation with collagenase, followed by separation in dense BSA gradient (42), and used as APC as described. Mice received 2 mg/ml ampicillin in drinking water for 48 h before, as well as throughout the duration of infection as required for plasmid maintenance. Presence of plasmid in strains recovered from infected tissues was confirmed by replica plating onto media containing ampicillin, and the CFU to APC ratio for all experiments ranged from 0.2 to 4.0.
Results
Differential stimulation of Ag-specific CD4+ T cells by extracellular and intracellular phase Salmonella
Salmonella replication, either extracellularly or inside host cells, requires coordinate regulation of many genes (6). We hypothesized that alteration of gene expression in response to changing environments during infection would impact production of bacterial Ags, and influence the antigenic specificity of responding CD4+ T cell populations. Therefore, we designed reagents that model bacterial Ag expression during extracellular or intracellular bacterial growth. Extracellular phase Ags (STEX) were prepared by growing S. typhimurium strain 14028 under conditions known to stimulate expression of proteins required for intestinal host cell invasion (14). Intracellular phase Ags (STIN) were approximated using a genetic approach; because the bacterial PhoP/PhoQ regulatory system is activated during growth inside host cells (7, 8), we used a mutant that over-expresses PhoP in its active, phosphorylated state (PhoPc). STIN and STEX Ags were processed and presented by naive splenocyte APC in assays further described below. A third strategy, using splenocytes from Salmonella-infected mice (IS) as APC directly ex vivo, allowed propagation of T cells responding to Ags expressed, processed, and presented only in vivo. Naive mice were orally infected with live, attenuated bacteria (S. typhimurium SL3261) to generate Salmonella-specific memory CD4+ T cells. These cells, primed throughout the course of infection by bacteria growing extracellularly in the gastrointestinal (GI) tract and intracellularly at systemic locations, were recovered from mice 60 days after immunization and used as input cells to generate T cell lines TEX, TIN, and TIS by selective amplification with STEX, STIN, or IS Ags, respectively (Fig. 1A).
After extracellular bacteria in the gut penetrate the intestinal epithelium (1) or are taken up by phagocytic cells directly from the lumen (5), intracellular Salmonella spread (4) and replicate at systemic sites such as the liver and spleen (2, 3). To test the hypothesis that extra- and intracellular phase bacteria stimulate different populations of immune CD4+ T cells, we examined the responses of immune T cell lines to intact, extracellular phase bacterial Ag (STEX) (Fig. 1B). CD4+ T cell line TEX had a 100-fold greater response to STEX as compared with T cell lines TIN and TIS. This indicated that Ag-specific T cells, presumably stimulated by Ags expressed only by extracellular phase bacteria, were present in line TEX but absent from lines TIN and TIS. In addition, the nearly identical responses of lines TIN and TIS indicated that T cells selectively amplified in response to intracellular phase bacteria grown in vitro (TIN) resemble those T cells responding to Ags expressed by bacteria growing intracellularly in vivo (TIS). These results suggested that the ability of bacteria to alter Ag expression during infection could affect immune recognition by the host.
Intracellular phase bacteria fail to support expansion of FliC-specific T cell populations in vitro
To more directly compare the antigenic specificities of T cell populations in each CD4+ T cell line, we tested the ability of SDS-PAGE fractionated STEX or STIN bacteria to stimulate immune T cells (Fig. 2A–F). As predicted by a previous study (35), line TEX (Fig. 2, first row), derived by selective amplification using STEX bacteria as Ag, responded to the fraction from STEX bacteria containing FliC (fraction 20, Fig. 2A). In contrast, line TEX displayed very weak responses to the corresponding fraction from STIN bacteria (fraction 20, Fig. 2B). Consistent with these results, STIN bacteria expressed substantially less FliC than STEX bacteria, as confirmed by Western blot analysis using anti-FliC Abs (Fig. 2, G and H). The reduced level of FliC expressed by STIN bacteria failed to support expansion of FliC-specific T cells from memory T cell populations, as T cell line TIN (Fig. 2, second row), derived by selective amplification using STIN bacteria as Ag, did not respond to any FliC-containing fraction (fraction 20, Fig. 2, C and D) or purified FliC protein (Fig. 2I).
Surprisingly, CD4+ T cell line TIS (Fig. 2, third row), derived by selective amplification with Salmonella-infected splenocytes as APC, responded to fractions 20 and 21 from STEX bacteria (Fig. 2E). However, equivalent responses of T cell line TIS to the corresponding fractions from STIN bacteria (fractions 20 and 21, Fig. 2F), which expressed markedly less FliC than STEX (Fig. 2, A, B, G, and H), suggested that TIS was responding to non-FliC Ags. This was confirmed using SDS-PAGE fractionated FliC-null bacteria, which contained Ags stimulatory for T cell line TIS that were in the approximate molecular mass range of FliC (50 kDa, Fig. 2J). The completely negative response of T cell line TIS to purified FliC protein (Fig. 2I) confirmed the absence of FliC-specific T cells in line TIS. The strong response of T cell line TEX to the FliC-containing fraction of STEX bacteria (Fig. 2A) and purified FliC (Fig. 2I) confirmed that FliC-specific CD4+ T cells were present in the original input responder population used to generate T cell lines TEX and TIS (see also Fig. 1). We therefore conclude that the expansion of FliC-specific T cell populations in vitro is restricted by the reduced FliC expression of intracellular phase bacteria.
Transcription of the gene encoding FliC is repressed during Salmonella growth inside macrophage phagosomes
Because FliC expression by STIN bacteria is below the threshold required for expansion of FliC-specific T cells in vitro (Fig. 2, C, D, and I), and maximal PhoP/PhoQ activation occurs in macrophage phagosomes (7, 8), we predicted that repression of fliC transcription during bacterial growth inside host cells would be PhoP/PhoQ-dependent. Therefore, we used GFP fluorescence to report transcription of fliC (PfliC::gfp) during Salmonella infection of macrophages. Wild-type Salmonella grown in vitro displayed equivalent fluorescence from either PfliC::gfp or Ppos::gfp (constitutive promoter) (Fig. 3, A and B). However, wild-type Salmonella inside of macrophages (Fig. 3, D, E, G, and H) repressed transcription of PfliC::gfp (Fig. 3E), whereas Ppos::gfp was constitutively expressed (Fig. 3D). In contrast, intracellular PhoP-null Salmonella, which lack a functional PhoP/PhoQ regulatory system (Fig. 3I) failed to repress PfliC::gfp (Fig. 3F). These data confirm that PhoP/PhoQ-dependent repression of fliC promoter activity occurs during bacterial growth in the macrophage phagosome.
Reduced expression of FliC in vivo limits activation of FliC-specific CD4+ T cells
Our observation that intracellular Salmonella repress expression of FliC, together with the fact that only extracellular phase bacteria are capable of supporting the expansion of FliC-reactive T cells in vitro (Fig. 2, A, B, and I), suggests that in vivo, FliC expression by intracellular Salmonella will fall below the T cell activation threshold. To confirm that FliC expression during bacterial growth in vivo is reduced, we used GFP fluorescence to report fliC transcription (PfliC::gfp) during Salmonella infection in mice. The absence of fluorescence indicated silencing of transcription from the fliC promoter in bacteria recovered from spleens of infected mice, although the same bacteria expressed GFP when grown in vitro (Fig. 4A, left column). In contrast, fluorescent bacteria were detected both in vivo and in vitro when a constitutive promoter controlled GFP expression (Fig. 4A, middle column). These results were confirmed by Western blot analysis: detection of bacterial DnaK verifies the presence of Salmonella, but FliC is undetectable in spleens of Salmonella-infected mice (Fig. 4B).
Because Salmonella growing in vivo repress FliC production, we predicted that splenocyte APC from Salmonella-infected mice would fail to stimulate FliC-specific CD4+ T cells. Therefore, splenocytes from mice infected with wild type or FliC-null Salmonella were used as APC directly ex vivo and compared with uninfected APC for their ability to stimulate Salmonella-specific T cell clones (Fig. 4C). APC from infected mice failed to stimulate FliC-specific T cells, however the T cells vigorously responded to exogenously added heat-killed FliC+ Salmonella presented by uninfected APC (Fig. 4C). As a control, infected APC or naive APC presenting exogenously added bacteria were equally stimulatory to a Salmonella-specific CD4+ T cell clone that recognizes both STIN and STEX bacteria (Fig. 4C). These results confirm that Salmonella growing in vivo at systemic sites specifically repress expression of FliC below the T cell activation threshold.
Artificial induction of FliC expression in vivo results in stimulation of FliC-specific T cells
If failure of FliC-specific T cells to detect and respond to bacteria growing in the spleens of infected mice is due to down-regulation of Ag expression, artificially increasing production of FliC during infection should render Salmonella-infected APCs stimulatory for FliC-specific T cells. Therefore, we performed a series of experiments using a strain in which an IPTG-inducible promoter controlled FliC expression. Bone marrow-derived DCs were infected in vitro with wild-type Salmonella (natural FliC regulation) or with BC898 (IPTG-regulated FliC expression) as described in Materials and Methods, and tested for the ability to stimulate Salmonella-specific CD4+ T cell clones (Fig. 5, A and B). In the absence of IPTG, FliC-specific T cell stimulation by infected APC did not differ from uninfected APC; after IPTG induction of bacterial FliC expression the response was similar to that stimulated by the presentation of heat-killed FliC+ Salmonella by uninfected APC (Fig. 5A). As before, these infected APC were capable of processing and presenting Ag to stimulate control T cell responses (Fig. 5B). Similar results were obtained during growth in host tissues when induction of FliC expression was performed in vivo (Fig. 5, C and D). Mice infected with BC898 (IPTG-regulated FliC expression) were provided with IPTG in their drinking water (45) and low-density splenocytes recovered from these infected mice were interrogated directly ex vivo for their ability as APC to stimulate Salmonella-specific T cell clones. FliC-specific T cells responded to infected APC only after IPTG exposure in vivo; without IPTG treatment, stimulation of T cells by infected APC was no different than with uninfected APC (Fig. 5C). Similar numbers of bacteria were recovered from infected spleens regardless of IPTG treatment (data not shown). These results confirm that exogenous up-regulation of FliC expression during intracellular bacterial replication allows FliC-specific CD4+ T cells to detect Salmonella-infected APC. Therefore we conclude that repression of FliC expression below the T cell activation threshold is sufficient to account for the absence of FliC-specific CD4+ T cell activation in liver and spleen observed during orogastric Salmonella infection (40). Taken together, our data support the hypothesis that bacterial regulation of Ag expression imposes both temporal and anatomical limits on host T cell responses: opportunities for T cell activation and subsequent effector function are restricted to those anatomical sites and stages of infection when bacterial expression of stimulatory Ag occurs.
Discussion
Infection with Salmonella results in a complex and dynamic interaction between pathogen and host. In this study we show that intracellular phase Salmonella repress expression of the protective Ag FliC below the threshold required for T cell activation, and that this repression is mediated by PhoP/PhoQ. We propose that during infection of a naive host with virulent Salmonella, the transition from extra- to intracellular environments triggers changes in Ag expression that subsequently influence both the location and timing of critical adaptive immune responses (Fig. 6 and see below).
Flagella, and the flagellar subunit FliC, are abundantly expressed on the surface of extracellular phase Salmonella and therefore serve as a large reservoir of ligands for both innate (30, 31) and adaptive (35) immune systems (see also Fig. 2). After bacterial invasion or translocation of flagellin across the intestinal epithelial cell layer, TLR5 stimulation results in a strong proinflammatory response (Fig. 6A) that is not induced by normal microbiota (32, 33, 46). Flagellin-stimulated secretion of chemokines such as IL-8 (31, 46, 47) and CCL20 (48) results in the recruitment of phagocytes, including DCs, to the small intestine. DCs are highly competent APCs (49) that can directly sample the gut lumen for Ags they present to immune effector cells (5). In addition, flagellin stimulates the maturation of both DCs and macrophages (50, 51). Thus, the proinflammatory response to flagellin, in combination with the proinflammatory response evoked by Salmonella-induced cell death (52), results in intestinal localization and activation of cells capable of priming adaptive immune responses (Fig. 6B). In our model, DCs (and possibly macrophages) in the Peyer’s patches initially prime CD4+ T cell responses to Ags, such as FliC, that are expressed by bacteria invading the host intestinal epithelium or captured from the gut lumen (STEX Ags) (Fig. 6C).
After uptake by phagocytes, a small number of bacteria disseminate inside host cells (4) (Fig. 6D) to systemic sites such as the liver and spleen, where intracellular bacteria repress expression of Ags such as FliC below the T cell activation threshold (Fig. 4C). The natural consequence of this repression is that Ags expressed only at early stages in infection (STEX Ags), before intracellular dissemination and replication, will prime CD4+ T cell responses that fail to recognize intracellular bacteria growing at systemic sites during later stages of infection (Fig. 6E). This idea is supported by oral Salmonella infections in which adoptively transferred FliC-specific T cells were activated at intestinal but not systemic sites, despite the presence of bacteria in all host tissues (40). A microanatomical separation of T cells away from APC was proposed to account for these observations, however our data indicate that bacterial Ag regulation represents an alternative explanation. Harvesting infected splenocytes disrupts splenic architecture, and yet these APCs fail to stimulate FliC-specific T cells (Figs. 4C and 5, A and C). The recovery of stimulatory capacity after exogenous up-regulation of FliC expression (Fig. 5, A and C) reinforces this conclusion. Thus, the repression of Ag expression below the T cell activation threshold during intracellular growth may functionally contribute to the ability of Salmonella to establish chronic carrier status in mice (53) and humans (54).
Flagellin is a potent stimulator of both innate and adaptive immune responses, and the expression of flagella is important for chemotaxis and invasion (reviewed in Ref. 55). In addition, a functional flagellar apparatus may also serve to facilitate bacterial pathogenesis independent of motility by altering expression or secretion of other virulence factors (56, 57, 58, 59, 60). It is noteworthy that studies of diverse pathogens often reveal coregulation of FliC with other virulence factors in response to environmental conditions (22, 61, 62, 63, 64). Through coordinate regulation of multiple genes, bacteria are able to precisely tailor expression of virulence factors and successfully infect their hosts. The global virulence regulon PhoP/PhoQ, activated by intracellular growth conditions (7, 8), is necessary for bacterial surface modifications required to resist innate immune defenses (65, 66). Our data demonstrate that Ag repression by intracellular bacteria occurs by a PhoP/PhoQ-mediated mechanism (Fig. 3) and activation of the PhoP/PhoQ regulon represses expression of surface Ags recognized by both innate and adaptive immune responses of the host (37). Many Salmonella mutants with dysregulated FliC expression have altered virulence (9, 63, 67). Furthermore, when a PhoP-activated promoter controls expression of FliC, virulence after oral infection of mice is decreased (our unpublished observations). Therefore, the linking of FliC expression to that of other virulence factors by PhoP/PhoQ regulation (in concert with other mechanisms) allows bacteria to maximize the contributions of flagellin to the establishment of infection while minimizing the availability of this Ag to stimulate host responses. It is significant that limiting host access to bacterial flagellin is a strategy used by many bacteria. For example, Helicobacter produces a flagellin that is inherently less stimulatory than that of other organisms, and limits secretion of flagellin monomer into culture supernatants (68). Many species, including Vibrio and Helicobacter cloak their flagella in an outer membrane sheath (69, 70). Finally, Pseudomonas strains recovered from cystic fibrosis patients with chronic infections have lost motility and are nonflagellated (71).
Our data support the hypothesis that during oral Salmonella infection of the naive host the initial T cell response is misdirected: T cells primed to respond to Ags expressed early in infection may fail to recognize intracellular bacteria growing at systemic sites. Thus, the timely development of host immune responses is critical for control of Salmonella infection. This is illustrated by the increased susceptibility of TLR4-deficient mice to Salmonella infection, in which bacterial burden increases as a result of delayed immune responses (72). It may not be necessary for a pathogen to completely abrogate host immune responses if delaying the response allows bacteria to multiply beyond the capacity of host defenses. We propose that repression of protective Ags like FliC, in concert with other genetically co-coordinated modifications of the bacterial surface (37), could facilitate bacterial growth in vivo by affording virulent Salmonella the opportunity for additional rounds of replication before the onset of effective adaptive immune recognition. A similar "decoy and delay" strategy is used during successful CMV infection: the dominant CD8+ T cell response is directed toward an epitope not presented in infected tissues in vivo (73).
The importance of timely host immune responses is underscored by the different outcomes for naive, susceptible mice when infected with virulent bacteria as compared with attenuated bacteria. AroA-attenuated Salmonella grow more slowly in vitro and in vivo (Ref. 74 and our unpublished observations) and are cleared by immunocompetent hosts, whereas virulent Salmonella rapidly overwhelm host defenses. The growth advantage of virulent bacteria is eliminated by prior immunization, as immune host T and B cell functions provide protection (75, 76). Interestingly, memory T cells express CCR6, the only known receptor for CCL20, a chemokine secreted by intestinal epithelial cells after stimulation by flagellin (77). Furthermore, Ag presented by DCs in the Peyer’s patches primes memory T cells that preferentially home to the gut (78). Therefore, in an immunized host it appears that memory T cells, primed to respond to early phase (STEX) Ags such as FliC, are recruited to the intestine where their resulting activation could facilitate the containment of bacterial replication and dissemination.
The gut mucosa presents a major barrier to Salmonella dissemination after oral infection. In addition to overcoming physical barriers such as the glycocalyx and mucus layers, Salmonella must also withstand chemical attacks from antimicrobial peptides secreted by the intestinal mucosa (79). Therefore, it is not surprising that most Salmonella do not progress farther than the intestinal lumen but are contained by host defenses in the small intestine or are eliminated. Although high doses of Salmonella (many times the oral LD50) have been used in experimental models of oral infection, near the oral LD50 only a few bacteria successfully escape to systemic circulation and colonization (80, 81). Based on the fact that FliC-specific CD4+ T cell activation occurs only at intestinal sites after oral infection (40), together with our observations that Salmonella express FliC only during extracellular phase growth (Fig. 2, A and G), we propose that priming of FliC-specific T cells must occur in the intestine where DCs, potent T cell activators, are recruited and activated by flagellin-mediated mechanisms. Therefore, the bacteria in the gastrointestinal tract, most of which will never reach systemic sites of replication, serve as a rich source of Ag for DC capture and subsequent activation of CD4+ T cells responding to STEX Ags (Fig. 6).
Finally, it is significant that intracellular Salmonella, although repressing the expression of Ags such as FliC, are not completely invisible to surveillance by CD4+ T cells derived from immune hosts (Figs. 4 and 5). This observation suggests that priming of CD4+ T cells specific for Ags expressed by virulent intracellular bacteria in the naive host occurs beyond the time limit for a successful adaptive response (Fig. 6, F and G). It is possible that CD4+ T cells recognizing intracellular phase bacteria respond to Ags that are expressed exclusively in vivo, or to Ags that are critical for bacterial survival and thus cannot be repressed. Based on our model, it is reasonable to predict that altered expression of STIN Ags could influence the final outcome of the host-pathogen interaction. For example, over-expression of STIN Ags at the extracellular phase of growth could facilitate early stage priming of potentially protective CD4+ T cell responses. In addition, purified STIN Ags may prove useful as vaccine candidates. Our model predicts that identification of these Ags and characterization of CD4+ T cell responses to them will provide additional useful insights into the relationship between Salmonella and its hosts.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by National Institutes of Health Grant AI47242.
2 Address correspondence and reprint requests to Dr. Brad T. Cookson, Departments of Laboratory Medicine and Microbiology, Box 357110, University of Washington Medical Center, 1959 N.E. Pacific, Seattle, WA 98195. E-mail address: cookson{at}u.washington.edu
3 Abbreviations used in this paper: DC, dendritic cell; SPI, Salmonella pathogenicity island; IPTG, isopropyl -D-thiogalactoside.
Received for publication February 15, 2005. Accepted for publication April 6, 2005.
References
Jepson, M. A., M. A. Clark. 2001. The role of M cells in Salmonella infection. Microbes Infect. 3: 1183-1190.
Richter-Dahlfors, A., A. M. Buchan, B. B. Finlay. 1997. Murine salmonellosis studied by confocal microscopy: Salmonella typhimurium resides intracellularly inside macrophages and exerts a cytotoxic effect on phagocytes in vivo. J. Exp. Med. 186: 569-580.
Salcedo, S. P., M. Noursadeghi, J. Cohen, D. W. Holden. 2001. Intracellular replication of Salmonella typhimurium strains in specific subsets of splenic macrophages in vivo. Cell. Microbiol. 3: 587-597.
Vazquez-Torres, A., J. Jones-Carson, A. J. Baumler, S. Falkow, R. Valdivia, W. Brown, M. Le, R. Berggren, W. T. Parks, F. C. Fang. 1999. Extraintestinal dissemination of Salmonella by CD18-expressing phagocytes. Nature 401: 804-808.
Rescigno, M., G. Rotta, B. Valzasina, P. Ricciardi-Castagnoli. 2001. Dendritic cells shuttle microbes across gut epithelial monolayers. Immunobiology 204: 572-581.
Lucas, R. L., C. A. Lee. 2000. Unravelling the mysteries of virulence gene regulation in Salmonella typhimurium. Mol. Microbiol. 36: 1024-1033.
Groisman, E. A.. 2001. The pleiotropic two-component regulatory system PhoP-PhoQ. J. Bacteriol. 183: 1835-1842.
Alpuche Aranda, C. M., J. A. Swanson, W. P. Loomis, S. I. Miller. 1992. Salmonella typhimurium activates virulence gene transcription within acidified macrophage phagosomes. Proc. Natl. Acad. Sci. USA 89: 10079-10083.(Lisa A. Cummings*, Sara L)