Given the central role of intestinal dendritic cells (DCs) in the regulation of gut immune responses, it is not surprising that several bacterial pathogens have evolved strategies to prevent or bypass recognition by DCs. In this article, we will review recent findings on the interaction between intestinal DCs and prototypical bacterial pathogens, such as Salmonella, Yersinia, or Helicobacter. We will discuss the different approaches with which these pathogens seek to evade DC recognition and subsequent T cell activation. These diverse strategies span to include mounting irrelevant immune responses, inhibition of Ag presentation by DCs, and stretch as far as to manipulate the Th1/Th2 balance of CD4+ T cells in the bacteria’s favor.
The gastrointestinal epithelia perform essential roles in the degradation and absorption of food. Although the primary function of the intestine is the uptake of food-derived nutrients, this permeability also creates gateways for entry by incompletely degraded immunogens and pathogenic microbes. Accordingly, the intestinal immune system has evolved multiple strategies to protect the epithelia involving innate and adaptive immune mechanisms (1–4).
Evolution, largely through horizontal gene transfer, has yielded pathogenic bacteria that can overcome the barrier properties of the gut epithelia and the protection by the intestinal immune system (5). Virulence factors injected by bacteria into or near epithelial cells are responsible for directed uptake and transcytosis, paracellular entry, and local damage that result in epithelial or lamina propria and sometimes deeper, systemic infections. Although the interaction between different pathogenic intestinal bacteria, such as Salmonella, Yersinia, or Helicobacter and the innate immune system has been the subject of many studies, much less is known about how the bacteria interact with intestinal dendritic cells (DCs). Given the critical role of DCs in inducing, priming, and regulating antibacterial immunity, it is not surprising that pathogenic bacteria have developed distinct strategies to circumvent immune recognition by DCs. In this review, we will outline recent findings on the interaction between intestinal bacteria and DCs and discuss how some of these interactions can be regarded as sophisticated evasion strategies promoting bacterial colonization and/or dissemination.
Intestinal DC subtypes
DCs make up a heterogeneous group of cells, which differ significantly in phenotype and function (6, 7). A useful way of classifying different subsets is to divide them into tissue-resident DCs and those that reside primarily in lymphatic organs such as the spleen or the lymph nodes (LNs). Tissue-resident DCs are strategically positioned in the periphery where they sample a variety of Ags, including those of self- and microbial origin. Upon successful ingestion of Ag, DCs will undergo maturation and migrate to the local draining LNs (8). Through upregulation of MHC and costimulatory molecules, matured DCs convert into highly efficient APCs (9). Successful Ag presentation to CD4+ T cells requires recognition of cognate peptide in the context of MHC class II molecules, whereas epitopes presented on MHC class I molecules stimulate Ag-specific CD8+ T cells. It was long held that T cells specific for tissue-derived Ags were only primed by migratory DCs. However, it is now clear that migratory DC can also transfer their Ag to resident DCs for T cell stimulation, a process that most likely involves cross-presentation (10, 11). Even though the transfer of Ag from migratory DCs to those DCs residing only in the lymphatic tissues represents an interesting function, resident DCs are best known for their ability to capture and present Ags that are either present within the local microenvironment or, in terms of the spleen, are blood borne (8). The diverse nature of Ags (i.e., self-derived, food, commensals, and pathogenic microbes) requires that Ag presentation by DCs is very strictly regulated. To avoid activation of self-reactive T cells and to limit unnecessary responses, such as those against commensal flora, DCs can imprint tolerance onto T cells (12). Even though we have only a limited understanding of the precise mechanisms and the events that govern whether immunity or tolerance is induced, tolerance induction appears to depend on local factors (13, 14), the presence of regulatory cells (15, 16), and importantly, on which DC subset presents the Ag (11, 14).
With only a single layer of epithelial cells separating the external from the internal world amid the constant need for particle exchange, intestinal DC play a key role in maintaining intestinal homeostasis as well as governing protective immune responses against invading pathogens. To complete this complex task, intestinal DCs differ significantly with respect to subset composition, function, and location from other DC subsets (11, 13, 15, 17, 18). Traditionally, intestinal DCs have been differentiated based on the expression of CD11b and CD8α (CD11b−CD8α−, CD11b+ CD8α−, and CD11b−CD8α+). More recent studies indicate that the expression of αE-integrin (CD103) defines two functionally distinct subsets (14, 19–21). Whereas CD103− DCs are thought to initiate proinflammatory T cell immunity (22), CD103+ DCs have the potential to induce FOXP3+ regulatory T cell responses through the production of TGF-β and retinoic acid (14, 23, 24). However, it remains to be determined whether the recently observed dichotomy in CD11b expression (25) on CD103+ DCs (CD11b+ versus CD11b−) further divides this DC subset functionally and where CD103-expressing mesenteric LN-resident CD8α+ DCs fit into the picture (26).
One interesting observation concerning DC engagement with intestinal bacteria comes from studies showing that DCs extend protrusions between epithelial cells, enabling direct sampling of luminal Ags (27). Whereas this unique function was directly attributed to the expression of CX3CR1 (28), others have suggested that the signals required for the delivery of these protrusions extended into the lumen involve CCL20 secretion from a parenchymal, radioresistant cell type (29). These studies revealed that the development of protrusions depended on TLR2 and TLR4 expression and that the protrusions markedly decreased after broad-spectrum antibiotic treatment (29). From these observations, it was concluded that the radioresistant cells probably represent epithelial cells that are sensing luminal, microbe-derived pathogen-associated molecular patterns and, in response, release CCL20 into the lamina propria. The expression of E-cadherin on intestinal epithelial cells suggested that CD103 might play an important role in guiding lamina propria DCs close to the epithelial surface (25). However, more recent evidence indicates that CX3CR1 is expressed by CD103− DCs rather than CD103+ DCs in the lamina propria (25). This observation questions whether both DC subsets are equally efficient at sampling the intestinal lumen, whether this is a subset-specific function, or whether this unique function is exclusive to CX3CR1+CD103− DCs. Finally, on the basis of a study, which showed that lamina propria DCs can also be found in the intestinal lumen (30), it was recently suggested that sending out protrusions into the lumen might only be a first step toward entering the intestinal lumen. However, whether there is a sequential relationship between DCs sending out their protrusions and migration into the lumen remains to be elucidated.
Salmonella Typhimurium deceives and inhibits DCs
Salmonella enterica serovar Typhi is the causative agent of typhoid fever in humans. Similarly, S. enterica serovar Typhimurium (S. Typhimurium) causes a systemic salmonellosis in mice, an infection that is widely used to model typhoid fever (31). Following ingestion, bacterial invasion occurs either directly through the epithelial barrier or, more efficiently, through microfold cells (M cells), a specialized cell type interspersed between epithelial cells lining the intestinal surface (32). Once translocated to the lamina propria, the bacteria replicate extracellularly and eventually gain access to Peyer’s patches. One of the major virulence determinants of Salmonellae are host cell modulating effector proteins that can be injected directly into host cells through one of two types of type III secretion systems (TTSS). Salmonella pathogenicity island (SPI)-1– and SPI-2–encoded TTSS of S. Typhimurium play an important role in determining how the bacteria invade the host (33). Whereas SPI-1–encoded genes initiate extracellular Salmonella invasion via M cells and epithelial cells, SPI-2–dependent genes direct Salmonella invasion via DCs (34). In the absence of fully functional SPI-2, S. Typhimurium infection in mice occurs through direct invasion into epithelial cells, a SPI-1–encoded phenomenon that is termed “the classical pathway.” Under these circumstances, Salmonellae are taken up by M cells and intestinal epithelial cells (35), causing these cells to die very rapidly. This in turn promotes bacterial escape into the subepithelial compartment (36) and dissemination of the bacteria both locally and systemically. Surprisingly, the SPI-1–dependent pathway of invasion appears to be less important, because SPI-1–deficient Salmonellae still infect mice and cause colitis (34), although mucosal IgA responses are reduced (37). This observation raises the question as to what the biological relevance of the SPI-1–dependent invasion pathway might be. To address this question, it is worth having a closer look at the events that occur during the extracellular lifespan of the bacteria. To directionally navigate in the extracellular space, Salmonellae use a flagellin-based propulsion system (38). Flagellae are composed of numerous subunits that are attached to the bacterial surface via a “hook-system” and a basal body, which acts as a “motor” to drive flagella rotation (39). Of the major structural components, the flagellin subunit FliC is of particular interest to immunologists, because it is the prototypical ligand for TLR5 (40). Recognition of FliC by TLR5 expressed on epithelial cells and monocytes, for example, results in the secretion of proinflammatory mediators, such as IL-6, TNF-α, IL-1β, as well as the anti-inflammatory cytokine IL-10 (41). However, flagellin is not only a TLR agonist, but it is also attractive “bait” for DCs. Upon ingestion and processing within the DC, a peptide derived from its FliC subunit is presented in the context of MHC class II molecules to CD4+ T cells (42). In studies that were greatly facilitated by the generation of a TCR transgenic mouse line in which all CD4+ T cells are specific for FliC (43), analysis of FliC-specific CD4+ T cell responses during Salmonella infection has yielded some remarkable findings. Although FliC-specific CD4+ T cells rapidly upregulate the early activation marker CD69, undergo clonal proliferation, and gain effector function, these fully armed CD4+ T cells fail to protect the host (43). With flagellin being one of the most abundant bacterial proteins present in the infected host tissues, it is quite surprising that flagellin-specific CD4+ T cells are so remarkably inefficient. Elegant work by Cookson and colleagues (44) provides a simple yet striking explanation for this puzzling observation. Once Salmonella has reached the inside of a cell, the bacteria change their gene transcription profile substantially, and among these changes, the production of FliC is dramatically reduced. This makes biological sense, because the bacteria no longer require motility when residing intracellularly. With FliC-specific CD4+ T cells dominating the immediate T cell response to Salmonella (45), the shutdown of FliC production and therefore elimination of the their cognate Ag from the bacteria simply renders the bulk of the initial CD4+ T cell response functionally irrelevant. Thus, by directing DCs to induce T cell responses against Salmonella Ags that rapidly become irrelevant, Salmonella uses an intriguing strategy of deception and diversion (Fig. 1).
The interaction of S. Typhimurium with DCs is not limited to the regulation of key epitopes and their presentation to T cells. As with macrophages and monocytes (46, 47), it has been suggested that infected DCs can act as vehicles for Salmonella invasion and dissemination (48). Seminal work by Hapfelmeier et al. (34) has recently demonstrated that caecal inflammation following SPI-1–deficient Salmonella infection was entirely abrogated in DC-depleted mice, indicating that SPI-2–mediated entry of the host occurs exclusively via DCs. This, of course, creates a situation in which large quantities of Salmonella Ags are present within cells that are specialized to activate T cells to foreign Ags. One obvious approach at preventing infected DCs from using this Ag store for subsequent T cell activation is to trigger apoptosis in infected DCs. Indeed, Salmonella is quite effective at inducing apoptosis of DCs (49). However, bearing in mind that Salmonella also uses DCs to facilitate its dissemination (48), it would be evolutionarily unwise to kill off all infected DCs. Remarkably, Salmonella responds to this challenge by inhibiting Ag presentation by infected DCs. van der Velden and others (50–52) have shown that infected DCs are dramatically compromised in their ability to stimulate Ag-specific CD4+ and CD8+ T cells. This inhibitory action not only accounts for presentation of bacterial-derived Ags (52) but also extends to T cell proliferation following pulsing of DCs with cognate peptides (51, 52). This effect is dependent on bacterial protein synthesis, suggesting that a thus far unidentified inhibitory factor is actively produced within infected DCs (51). The precise mechanisms by which Salmonella inhibits Ag presentation so efficiently are unclear but might involve alterations in MHC expression (51), intracellular trafficking of MHC molecules along actin polymers (53), or their ubiquitination (54). Thus, by inhibiting Ag presentation of infected DCs, Salmonella has evolved a strategy with which to disguise its presence in the most potent APCs. Taken together, these findings indicate that Salmonella has developed distinct strategies to circumvent DC-mediated immune recognition. Considering the functional heterogeneity among the different DC subsets, it is tempting to speculate that distinct evasion strategies target different DC subsets. However, to date, the precise role different DC subsets play in response to Salmonella infection is largely unknown, and many questions remain unanswered. Clearly, future studies need to shed light into this exciting area of research.
Yersinia enterocolitica inhibits Ag presentation by DCs
Salmonellae are not the only intestinal pathogen that has developed means to prevent priming of bacteria-specific T cells by interfering with Ag presentation by DCs. Yersinia enterocolitica is a Gram-negative bacterium that causes food-borne acute and chronic gastrointestinal diseases (55). Ingested Y. enterocolitica are taken up by M cells, colonize Peyer’s patches, and disseminate to liver and spleen (56). Despite its predominantly extracellular lifestyle, the bacteria have a TTSS, allowing them to inject a number of effector proteins, including Yersinia outer proteins (Yops) into host cells. Notably, DCs are also targeted by Yops in vivo and injection of these bacterial effector molecules has been shown to inhibit Ag presentation by DCs (57). By interacting directly with the actin cytoskeleton, Yops inhibit the phagocytic activity of DCs (58) and, as proof of principle, transfection of DC with YopP has been shown to reduce uptake of OVA (59). Direct evidence that YopP can inhibit the ability of DCs to present Ag to CD8+ T cells in vivo was provided by Trülzsch et al. (60), who tracked the induction of CD8+ T cell responses to Yersinia encoding for the Listeria-derived Ag listeriolysin O (LLO). Upon infection of mice with wild-type Y. enterocolitica encoding LLO, they found that the frequency of IFN-γ–producing LLO-specific CD8+ T cells was greatly reduced when compared with mice challenged with Listeria monocytogenes. Importantly, priming of LLO-specific CD8+ T cells was reinstated when mice were infected with a mutant of Y. enterocolitica that lacks YopP. Given that similar amounts of LLO were transferred by the YopP mutants, these findings indicate the potent ability of Y. enterocolitica to actively inhibit the induction of Ag-specific T cells via injection of YopP. Even though the mechanisms involved in the inhibition of Ag presentation by Y. enterocolitica remain unresolved, it has been suggested that the transfer of Yops induces apoptosis of DCs, thereby simply reducing the number of DCs available for T cell stimulation (60, 61). Of note, other members of the Yersinia genus also appear to have developed special strategies to interact with DCs. For example, Y. pestis, the infamous causative agent of the plague, has been shown to infect DCs (62), whereas Y. pseudotuberculosis uses YopE to circumvent phagocytosis by DCs (63). Thus, direct inhibition of the ability of DCs to present Ag represents another interesting evasion strategy used by intestinal bacterial pathogens (Fig. 2).
Helicobacter pylori modulates CD4+ T cell differentiation
Another strategy to prevent immune recognition by DC is used by Helicobacter pylori. This Gram-negative bacterium is unique among the intestinal pathogens, because it inhabits the gastric mucosa (64). Infections with H. pylori almost always lead to chronic infection, and in some patients, this chronic infection can lead to gastritis, ulcers, or gastric cancer (65). Despite its location in a protective mucous layer situated between the acidic luminal content of the stomach and the epithelial gastric lining, several lines of evidence indicate an active interplay between the bacterium and DCs. First, DCs are recruited into the gastric mucosa very early on postinfection of mice with H. pylori (66). Second, H. pylori can productively infect bone marrow-derived DCs in vitro (67), and third, DCs in the gastric mucosa of H. pylori-infected individuals can send protrusions into the lumen with which they make direct contacts with H. pylori (67). Even though there is still an ongoing debate as to the exact role of CD4+ T cells during H. pylori infection, these cells are clearly placed at the center of the ensuing immune response (68, 69). A current concept stipulates that a strong bias of CD4+ T cells toward a Th1 phenotype is associated with attenuated H. pylori colonization (70, 71). Consistent with this notion, it appears that H. pylori has evolved strategies with which to prevent the induction of Th1-biased CD4+ T cells. Direct evidence for this comes from a study that compared H. pylori-induced IL-12 secretion by monocyte-derived DCs following short-term and sustained stimulation with fixed H. pylori. Whereas short-term stimulation induced DC maturation and strong IL-12 responses, sustained stimulation significantly impaired DC function and reduced the secretion of IL-12 (72). Lending further support to the notion that H. pylori skews the Th1/Th2 balance by reducing IL-12 secretion, Kao et al. (66) recently identified a nonproteinaceous factor in sonicates of H. pylori that has the ability to prevent IL-12 secretion by DC in vitro, indicating H. pylori may act to actively suppress DC function, as opposed to causing exhaustion through chronic stimulation. Although the precise nature of this factor remains enigmatic, this process likely involves the C-type lectin DC-SIGN, which is expressed by DCs. Fucose-motifs present in H. pylori inhibit IL-12 secretion by means of modulating DC-SIGN signalosome plasticity (73). Furthermore, interaction between DC-SIGN and Lewis Ags expressed in certain phase variants of H. pylori LPS has been shown to block the development of Th1 response (74). Even though much remains to be examined, the above-mentioned observations suggest that H. pylori manages to regulate Th1 immunity in its favor, by skewing DCs toward a cytokine profile that is more conducive toward Th2 responses, thus representing yet another strategy with which an intestinal pathogen evades DC recognition and the subsequent T cell response.
The above-discussed observations and findings clearly indicate that different bacterial pathogens have evolved sophisticated strategies to evade immune recognition by intestinal DCs. Most remarkable is that these evasion strategies are complex and often only target certain functional aspects that are relevant to the respective pathogen. Despite these exciting insights, very little is known about how these evasion strategies relate to different DC subsets. Given the functional heterogeneity among the subsets, these specific evasion strategies will most likely have to take into consideration the respective subset that is involved. For example, inhibiting Ag presentation by a subset that does not participate in the presentation of relevant Ags, or alternatively, modulating cytokine secretion by subsets whose function is the transfer of Ag rather than the priming of T cells would result in little evolutionary advantage. Thus, future research should revisit previously explored evasion strategies and resolve at which level and in which location of the complex DC network these strategies are most effective.
We apologize to all researchers whose work could not be discussed here due to space limitations. The critical review of the manuscript by Kirsty Short is gratefully acknowledged.
Disclosures The authors have no financial conflicts of interest.
Abbreviations used in this paper:
- M cell
- S. Typhimurium
Salmonella enterica serovar Typhimurium
Salmonella pathogenicity island
type III secretion system
Yersinia outer protein.