Roundtrip Ticket for Secretory IgA: Role in Mucosal Homeostasis?¹

In humans, mucosal surfaces of the gut, the airways, the urogenital tracts, and the ducts of all exocrine glands are lined by epithelial layers that form tight barriers separating a rapidly changing external environment from a highly regulated internal compartment. Due to the vast surface they cover, these vulnerable cellular barriers are the most frequent portals of entry for microorganisms and exogenous materials. Mucosal surfaces are endowed with powerful defense mechanisms, which selectively process harmful or innocuous Ags to guarantee local homeostasis. Physicochemical and mechanical features, including mucus, glycocalyx, lactoferrin, peroxidase, defensins, trefoil peptides, and peristalsis, form an array of constitutive, nonspecific defenses acting in concert with an inducible, highly specialized immune system. The conjugated action of epithelial cells, the MALT, immunoreactive cells distributed within the mucosal tissue, Abs, and the commensal microbiota contribute to shape adaptive immunity (1). Fine tuning of effector mechanisms regulating productive immunity, oral tolerance, and immune ignorance is crucial to prevent bystander tissue damage, making the specific immune reaction even worse than the inducing effect caused by the exogenous agent (2).

In the intestine, organized secondary lymphoid tissues known as GALT comprise a variety of inductive sites required for Ag uptake, processing, and presentation for induction of mucosal responses. In mammalian gut, the major inductive site is constituted by Peyer’s patches. In mice, they form macroscopic, large clusters of lymphoid follicles distributed in the mucosa along the entire length of the small intestine (3). In humans, Peyers’ patches are described as macroscopic lymphoid aggregates that can be seen by the naked eye in the serosal side of the small intestine, with the greatest density found in the ileum (4, 5). Depending on the species, other GALT-like elements occur including small intestinal isolated lymphoid follicles (6), cryptopatches (7), and lymphocyte-filled villi (8). Lymph nodes found in the mesentery of the small intestine contribute to local immune responses (9), but should not be considered as GALT structures, because they are not capable to sample Ags directly from the lumen (10). Direct uptake by lamina propria dendritic cells displaying “snorkeling” transepithelial dendrites in intestinal villi (11) or phagocytosis by enterocytes bearing TLR-4 (12) are other pathways that govern Ag passage para- or transcellularly. However, ambiguities remain as to their anatomical and functional capacity to ensure all the steps (uptake, processing, T cell activation) leading to the induction of specific immune responses.

Murine Peyer’s patches contain a subepithelial dome (SED) region, underlying lymphoid follicles filled with naive and activated B cells, and interfollicular regions (IFR) rich in T cells. The surface of the dome region is covered by the follicle-associated epithelium (FAE) comprising M (microfold) cells (13), whose function is to selectively sample and transport luminal Ags, including proteins and microorganisms (14). The numbers of M cells vary between species, ranging from percentages as low as 2% of the FAE in humans (15) to an almost homogeneous population in the ruminant ileal Peyer’s patch (16, 17). The differentiation and uptake processes of M cells are poorly defined (18), mainly because their scarcity in the intestinal mucosa has hampered their cellular and biochemical studies. The existence of M cells outside the FAE capable to sample Salmonella, Yersinia, and gut bacterial Ags has been reported in intestinal villi in mice devoid of Peyer’s patches (19). Labeling with the novel M cell-specific marker secretory granule neuroendocrine protein-1 should help to establish the possibly related phenotype of such cells (20).

DC in the SED region are anatomically positioned to sample Ags from the intestinal lumen and are crucial to control the nature, quality, and intensity of the T cell response locally and timely (21). The phenotypic analysis of DC within tissues of the gastrointestinal tract has identified a plethora of distinct subtypes with different spatial distribution and effector functions (22). Intestinal DC have been shown to play a central role in the induction of local productive responses (23) or oral tolerance in the absence of any exogenous inflammatory signal (24). Populations of mouse DC isolated from Peyer’s patches, lamina propria, mesenteric lymph nodes (MLN), and even lung have a propensity to induce Th2 responses in T cell priming assays in vitro (25, 26). DC freshly isolated from Peyer’s patches promote elevated IL-10 mRNA expression (27), a particular feature not observed with DC recovered from spleen (28). Nonpathogenic Ag-loaded DC trafficking (29, 30) from the Peyer’s patch, the lamina propria, or the intestinal epithelium to the draining MLN is essential to the induction of oral tolerance at this obligatory site (31). Only MLN and Peyer’s patch DC imprint gut-homing specificity on T cells by inducing expression of α₄β₇ integrin, thus shaping the immune response at anatomical sites prone to encounter the cognate Ag (32, 33). These many criteria mark mucosal and more particularly Peyer’s patch DC with unique features, which most likely are essential for the processing of foreign Ag and subsequent immune imprinting under inflammatory or homeostatic conditions.

Cognate interactions among underlying APC (DC and macrophages) in the SED region, naive T cells residing in the IFR and B cells in the corona (mantle zone), lead to the generation of B cells with a high level of IgA and J chain expression (34). Ag-activated T and B cells emigrate from the inductive environment via the lymphatic drainage, circulate through the bloodstream, and home to effector sites represented by the mucosal lamina propria, the stroma of exocrine glands, and surface epithelia (10). Differentiation and expansion of IgA⁺ B cells occur during their spreading to the effector sites (35, 36), resulting in terminal maturation to plasma cells producing polymeric IgA (pIgA; mainly dimers). Multiple cytokines, including obligatory TGF-β and IL-10, IL-4, IL-5, and IL-6, are required to promote IgA class switching and maturation (37). Subsequent transport of pIgA across the epithelium of exocrine tissues is ensured by the polymeric IgR (pIgR); following cleavage at luminal surfaces, SIgA is released as a complex of pIgA and the cleaved extracellular portion of the pIgR defined as bound secretory component (SC) (Fig. 1).

IgA functions at three anatomical levels in relation to mucosal epithelium: 1) luminal SIgA Ab prevents adhesion and entry of Ag into the epithelium (38); 2) IgA Ab in the lamina propria binds and excretes Ag to the lumen (39); and 3) IgA Ab in transit through the epithelium can inhibit virus production (40) or neutralize proinflammatory Ags (41). An additional property of IgA is its inability to trigger the release of inflammatory mediators through receptors specific for its Fc domain (42, 43, 44). In support of this are two reports showing differential pattern of in vitro-derived DC activation after surface binding of IgA as compared with other Ig isotypes (45, 46). This is consistent with the multitask role of protecting against foreign substances and microbes, regulating the commensal microbiota, while at the same time not subjecting the mucosa to undue inflammation (47, 48).

SIgA exhibits the striking feature to adhere to the apical membrane of M cells (49, 50), promoting uptake of small amounts of Ab as reflected by the detection of gold-coated SIgA in the pocket of M cells (51). Refined analysis of the interaction of M cells with Ab molecules demonstrated that IgA, with or without bound SC, but not IgG or IgM, bound selectively to murine and human M cells (52). Adherence of IgA to M cells required domains Cα1 and Cα2, suggesting that no known IgA receptor (pIgR (53), Fcαμ receptor (54), FcαRI (55), and asialoglycoprotein receptor (56)) was involved in the process. The presence of SIgA in the M cell pocket and in processes that extend into the basal lamina indicated that the Ab was subsequently transported across the epithelium and brought in contact with the GALT. In face of the large excess of SIgA in the intestinal lumen, this suggests that the intrinsic passage of the Ab remains limited. In the context of immune complexes, we postulate that upon binding of the Ag, SIgA experiences (a) conformational change(s) that result(s) in increased binding capacity to the IgA receptor.

In vivo uptake of SIgA delivered into mouse ligated ileal loop containing a Peyer’s patch resulted in specific targeting to DC in the SED region and CD4⁺ T cells in the IFR (57). As for binding to M cells, the interaction of SIgA with DC was dependent on the IgA moiety. SIgA was internalized into DC, whereas the association with CD4⁺ T cells was limited to cell surface (57). Ex vivo, only DC isolated from Peyer’s patches and MLN showed selective binding and internalization mimicking the in vivo situation (K. Kadaoui and B. Corthésy, unpublished data); DC from the spleen and other lymph nodes did not exhibit any association, arguing for a specific receptor still to be identified.

To address the issue of whether transport of SIgA Ab across the M cell could trigger immune responses, rabbit SC was genetically engineered to deliver in association with pIgA a foreign epitope from Shigella flexneri invasin B to the intestinal lymphoid tissue (58). Oral administration to mice of the “antigenized” SIgA in the presence of cholera toxin evoked immune responses that included Abs against both invasin B and the SC carrier. This indicated that SIgA was able to survive the harsh gastrointestinal environment and provides a valuable mucosal delivery system for microbial and other peptidic Ags.

These observations raised the question of the functional consequences of the Peyer’s patch-mediated re-entry of SIgA across the intestinal mucosa. Such an issue was tackled following oral immunization of naive mice with engineered SIgA molecules consisting of mouse pIgA and human SC serving as a non-self-Ag. The recombinant Ab induced mucosal and systemic responses against human SC in the absence of any mucosal adjuvant (59). Engineered SIgA triggered production of human SC-specific Abs, human SC-dependent mixed Th1/Th2 type of responses, preserved or induced IL-10 and TGF-β expression in MLN, and migration and maturation of DC along the Peyer’s patch-MLN-spleen axis. By comparison with human SC adjuvantized with cholera toxin, engineered SIgA promoted less pronounced Ag-specific responses, marking this class of Ab as a weak intrinsic mucosal immunopotentiator. In other words, specific targeting of SIgA-based immune complexes to DC more prone to presentation than killing (as opposed to macrophages) induces low degrees of activation in a noninflammatory context favorable to preserve local homeostasis of the gastrointestinal tract. This causes partial activation of DC, which results in the local induction of both effector and regulatory T cell phenotypes. Interestingly, we have evidence that the privileged target of SIgA in Peyer’s patches are CD11c⁺/CD11b⁺ DC (K. Kadaoui and B. Corthésy, unpublished data), which in Peyer’s patches and MLN are poor producers of IL-12 but potent inducers of IL-10-secreting T cells (60) and IgA production from naive B cells (61). Taken together with studies which show that Peyer’s patch DC are less activated than DC that have migrated to the MLN (62), the neutralization by SIgA of microbial stimuli in Peyer’s patches can favor tolerogenic responses through capture by poorly activated, “immature” DC (Fig. 2). Whether internalization into APC in the form of immune complexes leads to selective processing of the Ag, as recently shown to explain differential activation with respect to self and microbial Ags (63), remains to be evaluated.

The possible significance of SIgA-based immune complexes in educating the mucosal immune system finds another illustration in a recent paper that studied protection against Salmonella mediated by innate SIgA (64). Naive pIgR^−/− mice were much more susceptible to infection with Salmonella than wild-type mice. This was accompanied by a more pronounced bacterial shedding in the feces of pIgR^−/− as compared with wild-type mice. Cohousing of orally inoculated mice with naive mice resulted in less infection when infected mice had a wild-type phenotype, suggesting that reduced shedding and opsonization by SIgA of excreted bacteria prevented invasion of a cohoused host. However, the contribution of perturbed intestinal permeability in pIgR^−/− mice (65) to facilitated invasion by Shigella was not examined. The authors interpreted the data in terms of limitation of the spread of pathogenic microorganisms throughout the population.

Moreover, natural SIgA Abs present in intestinal washes are reactive with commensal bacteria (66, 67). As they form luminal immune complexes with commensals (68), it is reasonable to assume that re-entering SIgA Abs capable to bind with DC in the SED region participate in the controlled translocation observed at the level of Peyer’s patches. Specific targeting to mucosal DC more prone to tolerance than activation would contribute to promote low reactivity against the microbiota. This adds to the observation that commensal bacteria associated with local DC in the SED region do not penetrate further than the MLN, resulting in the confinement of immune induction against the microbiota to the mucosa (69). Making the systemic immune system relatively ignorant of these organisms at this stage would permit adequate stimulation in the case of sepsis. Although the reason for limited dispersion is not clear, the commensal nature of associated molecular patterns might preclude full activation of DC locally. This might in turn restrict the spectrum and degree of production of IgA, achieving levels of humoral immune responses ensuring appropriate self-limitation, but not complete elimination, of the microbiota (70).

As a consequence of dissemination of activated mucosal B cells committed to IgA synthesis from GALT, SIgA Abs against intestinal Ags are also produced in distant exocrine glands (71). During late pregnancy, due to this “enteric-mammary link,” the milk contains SIgA Abs against Ags (microbes, microbiota, and food), which have once passed the mother’s gut. Ag coating by SIgA limits epithelial translocation and protects infants against neonatal infections (72). In humans suffering from severe Shigella dysentery, this view might apply to mothers who prevented the symptoms in their breast-fed infants, yet they kept growing Shigella from their feces (73). Similar conclusions were reached in the case of breastfeeding by CMV-contaminated mothers (74). Based on the evidence of SIgA re-entry into Peyer’s patch, an extended reading of the data would consist in postulating that SIgA-coated, neutralized bacteria might help paving the immune system of naive individuals within a whole population in the absence of global infection.

Neutralization of Ags by SIgA results in the blocking of bacteria attachment to mucosal glycolipids and glycoprotein receptors and subsequent entrance into mucosae. Carbohydrate side chains in SIgA have been shown to interact with Helicobacter pylori (75), Escherichia coli through type I fimbrial lectin (76, 77), and the ricin toxin (78). Free SC in milk is able on its own to prevent adhesion of enteropathogenic E. coli, of Clostridium difficile toxin A, and of Streptococcus pneumoniae CbpA to epithelial cells (79, 80), a mechanism that identically depends on carbohydrates abundantly present on the surface of SC (81). Hence, SC plays an important role in the protection of mucosal surfaces, and when bound to pIgA, serves as a companion potentializing the Ab molecular function through additional microbial scavenger functions. In addition to contributing to passive protection, breastfeeding actively stimulates the immune system of the offspring. Factors including lymphocytes, cytokines, hormones, lactoferrin, and anti-idiotypic Abs are presumably involved (82). In the context of immunomodulation, maternal SIgA, free or in the form of immune complexes, can be seen as a part of the arsenal that shapes the gastrointestinal immune system both in terms of defense or tolerization during initial exposure to non-self-antigenic structures.

It has been acknowledged for almost a century that oral feeding with proteins abolishes subsequent responses to systemic challenge with the same mixture of proteins. This phenomenon is referred to as oral tolerance, and has been well characterized in animal models (mostly rodents) using multiple soluble proteins (83). An intriguing possibility in the context of SIgA-based immune complexes would be that these latter contribute early in life to educate the mucosal immune system toward a tolerogenic profile. It makes a sense to speculate that maternal milk SIgA Abs passing across the epithelium direct associated innocuous Ags to DC, and prime CD4⁺ T cells instrumental to the establishment of intestinal tolerance to become low responders, to evolve to a regulatory T cell phenotype, or to be deleted. In species with abundant maternal IgG in breast milk, uptake of IgG-Ag immune complexes by epithelial cells can similarly achieve the transport of Ags to mucosal tolerogenic DC for appropriate hyporesponsive shaping. This is consistent with a study that showed that the IgG-specific Fc receptor (neonatal Fc receptor, FcRn) is expressed by epithelial cells in the intestine and airways, and can mediate IgG transport in both directions across epithelial barriers (84).

As far as regulation of responses against pathogens is concerned, it is conceivable that SIgA-based immune complexes play a role in triggering counteracting mechanisms necessary to return to steady-state conditions and maintenance of homeostasis. In this respect, cross-talk between DC and epithelial cells producing thymic stromal lymphopoietin conditions these latter to polarize T cells toward a Th2 phenotype (IL-6 and IL-10) and the production of mucosal IgA (85). SIgA present in the frame of secondary (recall) stimulation might in addition promote Ag presentation under neutralized conditions and subsequently participate in “educating” targeted DC to respond in a manner that limits the risk of overstimulating the local immune protection system. The complex array of multitask DC subsets present in Peyer’s patches (86), lamina propria (87), IFR, and draining MLN (31) justifies of such complicated regulatory pathways essential to the wellness of intestinal mucosae.