Getting to the guts of immune regulation.

Abstract

The mechanisms controlling the balance between tolerance and active immunity in mucosal tissues are critical, but not well understood. Specifically, the normal healthy gut exhibits tolerance both to beneficial antigens derived from food and to innocuous antigens derived from commensal flora, but will mount active and protective immune responses against detrimental and damaging antigens from invading gut pathogens. When such immune control mechanisms go awry, the consequences can be devastating. Loss of tolerance to food antigens can manifest as food allergies, such as coeliac disease, that may afflict about 1 in every 300 persons in several countries in the Western World. Likewise, loss of tolerance to commensal flora may manifest as inflammatory bowel disease. Conversely, the desirability of developing oral vaccines that are protective against pathogens can be confounded by a ‘default state of tolerance’ in the gut that is so robust that it has been used in recent years as a means to ameliorate autoimmune disorders. Oral administration of soluble protein that mimics the presumed target autoantigen is sufficient to induce systemic tolerance – a phenomenon first described almost a century ago, and commonly known as ‘oral tolerance’. In the laboratory, it is possible to actively immunize animals against orally administered antigens by combining soluble proteins with strong mucosal adjuvants such as cholera toxin (CT)1 or a heat-labile derivative from E. coli enterotoxin (LT).2 Such adjuvants, effective at overcoming oral tolerance, highlight the potential for developing efficient oral vaccines. However, because many of the early, toxin-based, experimental mucosal adjuvants are not appropriate for humans, appropriate modifications are still being sought. Inevitably, the quest to improve mucosal vaccines, coupled with the desire to make improved use of oral tolerance to treat auto-immune diseases or graft rejection, requires a better understanding of the events that contribute to the divergent handling of antigen by the mucosal immune system. While there is much still to learn, the paper by Smith et al.3 in this issue of Immunology makes some intriguing new observations that take us a step closer to understanding at which point in the immune response to antigen the critical decision making between an active vs. a tolerogenic response may take place. Smith et al.3 used the well-established DO11.10 transgenic (Tg) T-cell adoptive transfer system. In this approach, chicken ovalbumin (OVA)-specific Tg T cells are labelled with 5-(-6-)carboxyfluorescein diacetate, succinimidyl ester (CFSE), and adoptively transferred to recipient mice that are then challenged with antigen either in soluble form (for inducing oral tolerance) or coadministered with CT (for inducing oral immunization). The CFSE labelling permits both T-cell activation and T-cell division to be simultaneously assessed in various tissues [i.e. mesentereic lymph nodes (MLNs) and peripheral lymph nodes (PLNs)] of the recipient mice, following antigen challenge. Consistent with previous studies,4,5 the authors observed the induction of oral tolerance initially to provoke the ‘activation’ of T cells. This is an important point, since T-cell tolerance after antigen feeding is often wrongly assumed to result from ignorance or abstention from an immune response – quite different from the specific T-cell education and prolonged antigen-specific tolerance that arise in reality. The novel twist that Smith et al.3 have added to their studies is the direct comparison of the tolerogenic T-cell response with the active, adjuvant-assisted T-cell response. Although T cells appear to be activated with similar kinetics in either case, the T cells in the MLN of orally tolerized mice proliferated less compared with T cells in the MLN from orally primed mice. Conversely, the number of cell divisions made by T cells in peripheral tissues did not differ under conditions of priming or tolerance. Additionally, in response to lower doses of antigen, fewer cells acquired a memory phenotype in the PLNs than in the mucosal tissues. The authors concluded that similar downstream effector T-cell responses occur simultaneously in mucosal and peripheral tissues, irrespective of whether antigen is administered via tolerogenic or immunogenic routes, but that the capability to distinguish between the conditions of antigen feeding lies within the mucosal tissues, where tolerance is associated with a decrease in T-cell expansion. Although Smith et al.3 focused their efforts on studying the downstream effector T-cell response to orally administered antigen, the diverse local antigen-presenting cells (APCs) present in the intestine presumably play a major role in determining the T-cell response to antigen. Among such cells – epithelial cells, B cells, macrophages and dendritic cells (DC) – the DC in particular have been shown to contribute to both tolerogenic and immunogenic responses to mucosally delivered antigen.6,7 Indeed, although DC are classically thought of as the most potent and immunogenic of all APC, mucosal DC may in their native state preferentially behave as tolerogenic APC. Although expansion of DC in vivo has been shown to enhance T-cell tolerance, the precise mechanistic basis for this observed phenotype remains to be determined. The number of studies dealing with mucosal DC and their influence on T-cell reactivity is fairly low, and most have focused on Peyer's Patches (PPs). PPs are organized lymphoid structures located throughout the intestine and believed to be one of the primary sites for induction of mucosal immune responses. Interestingly, a number of groups have reported that distinct DC subsets exhibit particular localization preferences within the PP,8,9 perhaps suggesting that distinct DC contribute to the determination of the type of response required from the T cell. Many of the DC resident in the PP constitutively secrete immunosuppressive cytokines,10 and further up-regulate them under conditions of stimulation that would usually trigger IL-12 production by splenic DC (E. Williamson et al., manuscript submitted). The CD8α+ DC subset, which is particularly prominent in the PP,11 has been shown to be inefficient at stimulating T-cell proliferation12 and can actively suppress T-cell responsiveness through enzyme-mediated degradation of essential nutrients.13 Intriguingly, under normal conditions, the CD8α+ DCs are positioned within the T-cell-rich interfollicular regions of the PP, placing them in a unique position to directly and continuously dampen local T-cell responses to normal luminal antigens (Fig. 1, top panel). If these or other mucosal DC subsets are truly central players responsible for disseminating an immune response, it is likely to be necessary for these cells to be mobile. Supporting the notion that mucosal DCs may have a bona fide role as potent, travelling immune regulators of tolerance, the DCs that predominate in the gut appear to have a unique chemokine receptor expression profile. CD8α+ DCs primarily and constitutively express the chemokine receptors CXCR4 and CCR7.9,14 The ligand for CXCR4 is SDF-1, a chemokine constitutively expressed within organized lymph node tissue,15 and the ligands for CCR7, MIP-3β and ELC are abundant in T-cell areas of secondary lymphoid tissue.9,14 It is possible that, under homeostatic conditions, DCs in the PP (or other gut tissues) may load with antigen and continually traffic to T-cell-rich areas of both PP and lymph nodes in order to present antigen in the absence of activation (Fig. 1, top panel). Indeed, it has been demonstrated that intestinal DCs pick up orally administered protein antigen and drain from the gut in the absence of activation (16). Under such conditions, the proliferative T-cell response to orally administered antigen is unlikely to be robust – a point well illustrated by Smith et al.3 in this issue of Immunology. Model for dendritic cell and resident T-cell interactions during induction of oral tolerance or active immune responses in Peyer's patches (adapted from ref. 9). Under steady-state conditions (upper panel), myeloid DC and CD4+ CD25+ T-reg cells are co-localized to the subepithelial dome of PPs through CCR6/MIP-3α interactions, where the follicle-associated epithelium is known to produce MIP-3α. CD8α+ DCs, which secrete immunosuppressive factors (white halo), are predominately located within the T-cell-rich interfollicular region of PPs as a result of CXCR4/SDF-1 and CCR7/MIP-3β interactions. The co-association of the immunosuppressive or immunoregulatory CD8α+ DCs with T helper cells probably results in dampened T-cell responsiveness to innocuous luminal antigens in the T-cell follicular region and may even promote, via IL-10 secretion into the local microenvironment, the differentiation of Th0 cells into T-reg cells. Equally important in this model is the co-association of myeloid DCs with immunoregulatory CD4+ CD25+ T-reg cells in the subepithelial region at the site of antigen entry into the mucosa in the dome region of PPs. Antigens associated with myeloid DCs under non-inflammatory conditions would be presented to T cells in the absence of co-stimulation and in the presence of regulatory T cells, resulting in tolerance. When inflammatory stimuli are introduced to this system (lower panel), myeloid DCs become activated resulting in the up-regulation of co-stimulatory molecules as well as chemokine receptors such as CCR7. CCR7 expression promotes the trafficking of myeloid DCs to T-cell-rich regions of lymphoid organs and results in the presentation of antigens, together with the necessary co-stimulation, to induce active immune responsiveness. In contrast, CD8α+ DCs, which fail to respond to inflammatory chemotactic stimuli, will produce high levels of IFN-α when exposed to viruses and some bacterial products (blue halo), thus providing local antimicrobial activity. CD8α+ DCs retain CCR7 expression, however, and will continue to populate the T-cell-rich regions of lymph nodes. Hence, once antigen has been cleared by active immune responsiveness, immunosuppressive CD8α+ DCs may function to re-establish local immune homeostasis through T-cell suppression. Key: SED, subepithelial dome; HEV, high endothelial venule; SDF-1, stromal-derived factor 1; ELC, EBI-1 ligand chemokine; SLC, secondary lymphoid organ chemokine; MIP, macrophage inflammatory protein; FAE, follicle-associated epithelium. Under stimulatory conditions, the mucosal DCs would be required to present antigen to T cells with necessary co-stimulation. Although mucosal DCs may appear to be tolerogenic APCs by default, there are no apparent defects in the ability of this cell population to become activated, for example by CT or IL-1, and to convert to immunogenic APCs on demand.7,17 Upon stimulation, myeloid DCs, which usually are found in the subepithelial dome region of PPs under resting conditions, appear to lose their affinity to their resident site and migrate to the T-cell-rich regions of PPs9 (Fig. 1, bottom panel). It is conceivable that these DCs may continue to travel on to the MLN and elsewhere following activation. These activated DCs express high levels of co-stimulatory molecules and can present antigen in an immunogenic fashion to T cells, driving active immunity. This type of behaviour, following oral delivery of antigen in the presence of adjuvant, probably gives rise to the more aggressive expansion of antigen-specific T cells in the MLN, as discussed by Smith et al.3 An intriguing question for mucosal immunologists today revolves around how the gut revises its response in order to return to a tolerant state after potent immunization. It is interesting to note that, in contrast to the myeloid DCs that readily migrate upon activation, the more immunosuppressive CD8α+ DC subset, which routinely resides in the T-cell areas of organized lymphoid tissue, appears to down-regulate receptors for inflammatory chemokines upon activation,14 perhaps indicating a temporary suppression of tolerogenic mechanisms while there are signals indicating pathogenic challenge. Hence, one may speculate that, following a period of active expansion of antigen-specific T cells, the more suppressive CD8α+ DCs that have continued to populate the T-cell regions of secondary lymphoid organs gradually accumulate and take precedence over other DC types. Once the original offending antigen has been cleared, the presence of this unique DC type might re-establish immune homeostasis (Fig. 1, bottom panel). Clearly, the balance between tolerance and immunity is not solely regulated by the resident DCs directing naïve T cells to respond in a certain manner, or by the effector T cells that are generated. Other important cell types, for example CD4+CD25+ regulatory T-cell (‘T-reg’) populations that are a prominent feature at mucosal surfaces, must be considered.18 It has been shown in many animal models that T-reg cells are needed to prevent the development of intestinal inflammation,19–22 although the precise derivation of these cells remains controversial. Other regulatory T-cell populations have been demonstrated to develop or evolve following oral tolerance induction.23,24 It is highly probable that resident T-reg populations synergize with immunosuppressive CD8α+ DCs to prevent inappropriate immune responses to beneficial or innocuous luminal antigens. Of note, quantitative real-time PCR indicates high levels of CCR6 mRNA expressed by T-reg populations (T. George, Immunex Corporation, personal communication). Expression of CCR6 might facilitate recruitment of this cell type towards the mucosal surface and into close proximity with myeloid DCs, under the subepithelial dome of the PP. The combination of regulatory cells with potentially immunogenic myeloid DCs probably ensures that a necessary threshold of responsiveness is reached before the initiation of a potentially harmful immune response is unleashed. An additional player with the capacity to regulate the outcome of local T-cell responses is the intraepithelial T cell. Such cells may account for 10–20% of all human T cells, and, although they remain incompletely understood,25 it was recently shown that intraepithelial lymphocytes (IELs) in the skin actively suppress local infiltration of systemic, antigen-specific TCRαβ(+) T cells (26). In summary, immunoregulation of the intestinal immune response is likely to be controlled at many levels. Immunosuppressive mediators produced by local APC and regulatory T cell populations probably contribute to the default dampened response in the gut, but the suppressive APCs are readily converted to potent immunogenic APCs that may overcome the suppressive environment in order to promote effective immunization. The work of Smith et al.3 in this issue of Immunology demonstrates that the expansion of gut-associated T cells is at least one of the early targets of immuno-regulation. This provides an important experimental avenue for investigators to work backwards from T-cell proliferation in the MLNs to the precise molecules that provoke it, and will help identify the molecular difference between oral tolerance and oral vaccination.