Despite the importance of the GALT as an inductive site during exposure to microbial pathogens, relatively little is known about the early processes of T cell activation in this location. DCs have been shown to capture orally administered antigens in mucosal sites, and purified DCs from mucosal tissues can activate T cells in vitro (Fleeton et al., 2004b
; Hopkins et al., 2000
; Kunkel et al., 2003
; Liu and MacPherson, 1993
; Pron et al., 2001
). However, our data demonstrate an essential requirement for DCs in the induction of a mucosal T cell response in vivo and suggest a key role for PP DCs in the establishment of protective immunity to mucosal pathogens. The host defense against enteroinvasive pathogens likely involves a number of macrophage and DC subsets, each of which may present antigen. However, presentation in the absence of DCs is insufficient to support S. typhimurium
-specific T cell activation and clonal expansion in the mucosal immune system.
Various models have been proposed to explain microbial antigen acquisition by PP DCs (Fleeton et al., 2004a
; Iwasaki and Kelsall, 1999b
; Ravindran and McSorley, 2005
; Yrild and Wick, 2000
). Most predict that infection will stimulate directed migration of PP DCs away from the subepithelial dome and toward the interfollicular region (IFR). In marked contrast, our data reveal that a DC subset is rapidly recruited toward the subepithelial dome within hours of oral infection. Such migration is CCR6 dependant since CCR6+
PP DCs in homozygous GFP-knockin mice do not redistribute to the FAE after S. typhimurium
infection. Furthermore, this migration occurs despite the presence of a resident CX3
DC subset already situated beneath the FAE. Although initial reports suggested that CCR6 expression was required for constitutive DC migration to the subepithelial dome (Cook et al., 2000
; Varona et al., 2001
), a subsequent report using a more sensitive histological technique observed no defect in DC migration (Zhao et al., 2003
). Indeed, CD11c+
DCs were detected in the subepithelial dome region of the PPs of four different CCR6 deficient lines (Zhao et al., 2003
). Our data help reconcile these observations by demonstrating that two distinct DC subsets in PPs are distinguished by either expression of CCR6 or CX3
CR1. Both of these DC subsets accumulate in the subepithelial dome, one using a constitutive process and the other being dependent on CCR6 expression and local inflammation for migration. These two populations are distinct since CCR6+
DCs are not found in the lamina propria, and CX3
PP DCs do not coexpress CCR6.
While CCL20 is expressed by intestinal epithelium in response to Salmonella
exposure (Sierro et al., 2001
) and CCR6 expression characterizes the immune compartment of PPs, additional signals appear to organize the distribution of immune cells within PPs. For example, although PP B cells express CCR6, they do not aggregate beneath the FAE but appear distributed throughout the PP under homeostatic conditions. Therefore, despite the absolute requirement for CCR6 expression for DC recruitment to the dome, it seems likely that other chemokine receptor-ligand pairs play an additional role in defining the distribution of lymphocytes within the immune compartments of PPs.
DCs are located in the PP subepithelial dome and FAE of uninfected mice where they are ideally situated to capture incoming antigens. However, our data demonstrate that CCR6+
DCs in PPs are responsible for the early activation and proliferation of pathogen-specific T cell responses in the GALT. The role of the resident FAE CX3
DCs during adaptive host responses remains to be determined, but these DCs may be involved in steady-state antigen acquisition from luminal contents and tolerance induction under homeostatic conditions or innate defense activation in response to mucosal pathogens. A recent report demonstrated that close association of dendritic cells with intestinal epithelial cells generates a noninflammatory DC phenotype (Rimoldi et al., 2005
). It seems possible that PP CX3
DCs are ‘‘conditioned’’ in such a manner by their close association with epithelial cells, generating a need for CCR6+
DC migration toward the FAE for the induction of a pathogen-specific immune response in vivo.
Our experiments have not addressed direct Salmonella
entry into the lamina propria via invading the epithelium or DC processes that extend across the epithelial layer (Rescigno et al., 2001
; Vazquez-Torres et al., 1999
). Lamina propria CX3
DC engage the bacteria in the lamina propria and have been demonstrated to carry commensals and pathogens to MLNs (Niess et al., 2005
). Therefore, our data do not rule out the possibility that CX3
DCs are involved in mediating Salmonella
- specific T cell activation at later time points after infection or in other secondary lymphoid tissues.
CCR6-deficient mice are reported to have a reduced number of PP domes, B cells, and M cells, compared to wt mice (Lugering et al., 2005
; Varona et al., 2001
). However, differences in bacterial entry or replication do not explain defective T cell activation in CCR6-deficient mice since S. typhimurium
-specific T cells remained unresponsive across a large challenge dose range. Furthermore, bacterial entry into the PPs of CCR6-deficient mice at 12 hr after infection was similar to wt mice, demonstrating that initial entry of bacteria to the PP is not CCR6 dependant. From these data, it seems unlikely that CCR6+
DCs acquire bacteria by extending processes directly into the lumen, although we have not formally ruled out this possibility. Our data seem more consistent with a model where CCR6+
DCs engulf bacteria after bacterial M cell entry has already occurred.
CCR6 mediated migration to the FAE, and lack of SM1 T cell activation in the PPs correlates with enhanced susceptibility of CCR6-deficient mice to S. typhimurium infection. Surprisingly, we also detected slightly fewer bacteria in the PPs of CCR6-deficient mice 3 days after infection. It is possible that the absence of CCR6 directed T cell activation in the PP allows for greater bacterial dissemination away from the PPs by macrophages or other DC subsets, causing a small reduction in bacterial numbers in the PPs and enhanced bacterial loads in the liver. Thus, the primary function of activated Salmonella- specific T cells in the PPs may be to limit initial bacterial dissemination to other tissue sites. Alternatively, activated PP T cells may migrate to the liver and mediate bacterial killing outside the intestine. Whatever the mechanism, our data indicate that CCR6-dependent pathogen-specific T cell activation in PPs plays an important part of the defense against enteroinvasive pathogens. It remains possible that other DC subsets participate in T cell activation in other lymphoid tissues or at later times after infection. Indeed, we think it likely that defense against enteroinvasive pathogens will involve the simultaneous engagements of distinct DC subsets in both the PP and lamina propria, each of which may have a limited ability to mediate immediate or late activation of T cells or may participate in innate immune clearance.
In conclusion, PPs contain a distinct DC subset that is absolutely required to initiate rapid local S. typhimurium- specific CD4 T cell activation. CCR6 facilitates the attraction of this DC subset into the FAE and is essential for the recognition of mucosal pathogens by T cells in PPs. These CCR6-dependent adaptive immune responses are a critical component of the mucosal defense against entero-invasive pathogens.