Inducing mucosal IgA: A challenge for vaccine adjuvants and delivery systems

- Regulation of Ig class switch for production of IgA

B cells can undergo Ig class switch recombination (CSR) and acquire the ability to produce IgA after CD40-CD40L ligation in the presence of TGF-β with contribution from other cytokines including IL-4, IL-5, IL-6, IL-10, and IL-21 (5–8) (Figure 1). Other costimulatory signals such as B cell-activating factor of the TNF family (BAFF), a proliferation-inducing ligand (APRIL), retinoic acid (RA), and nitric oxide also facilitate CSR for the production of IgA (7, 9, 10). BAFF and APRIL further enhance IgA responses by providing survival signals, and/or inducing plasma cell differentiation and IgA secretion (10, 11). Retinoic acid, a metabolite of vitamin A, plays an important role in the production of mucosal IgA since in addition to acting synergistically with IL-5 and IL-6 to induce IgA secretion (12), it also induces expression of gut-homing receptors on B cells (12) (Figure 1). Additional cytosolic factors and transcription factors contribute to the regulation of IgA class switching. For example, the noncanonical IκB kinases TANK-binding kinase 1 (TBK1) in B cells was shown to negatively regulate IgA class switching by attenuating noncanonical signaling via the transcription factor NF-κB (13). The cytosolic protein clathrin, which controls receptor-mediated endocytosis and internalization of receptors, also influences antibody isotype switching and production of IgA (14). Thus, absence of clathrin light chain in B cells enhances immunoglobulin switch to IgA, an effect consistent with a defective endocytosis of TGFβR2 and subsequent increased TGFβR2 signaling (14).

- T-dependent and T-independent mechanisms of IgA induction

IgA can be produced by naïve B cells in gut-associated lymphoid tissues (GALT) or nasopharyngeal-associated lymphoid tissues (NALT) in response to stimulation by commensal microbes, microbial pathogens, or after vaccination. Comparison of naïve B cells in the intestinal Peyer’s patches with IgA producing cells in the intestinal lamina propria have shown that a change of energy metabolism occurs during differentiation of B cells into IgA-producing cells (15). Hence, while both naïve and IgA-producing cells use the tricarboxylic acid cycle and fatty acid for energy production, glycolysis preferentially occur in IgA-producing cells (15). Immunoglobulin CSR and production of IgA can occur via T-independent or T-dependent mechanisms depending on the nature of antigen and the cellular origin of help received by B cells (Figure 1). High affinity IgA are generated in a T-dependent fashion, while IgA produced in a T-independent manner lack or only have limited antigen specificity due to limited somatic hypermutation. In the gut, IgA-producing cells are generated in germinal centers of the Peyer’s patches, but also in other secondary lymphoid tissues (cryptopatch and cryptopatch-derived isolated lymphoid follicle) and microbe-induced tertiary lymphoid tissues. This process is finely regulated and it has now been shown to require prolonged interaction of B cells with dendritic cells in the subepithelial dome (SED) of the Peyer’s patches and DC-mediated production of TGF-β via integrin αVβ8-mediated activation of TGF-β (16). Innate lymphoid cells play a role upstream of these interactions as they facilitate the maintenance of DCs in the SED (16).

Several T cell subsets are involved in the induction of IgA. Among these cells, follicular T helper (Tfh) cells (17) play a key role in both immunoglobulin CSR and somatic hypermutation in antigen recognizing regions and subsequent generation of high-affinity antibody producing plasma cells and memory B cells. The expression of CXCR5 by Tfh cells allow them to locate into B cell follicles, where they produce IL-21, an IgA promoting cytokine (5). Tfh were reported to differentiate from other Th cell subsets suggesting a plasticity of mucosal T cells and potential for rapid adaptation to the luminal environment and ultimately produce the mucosal IgA Abs needed to maintain homeostasis.

Fox3+ T cells were the first T cell subset shown to undergo differentiation into Tfh cells in the GALT to support IgA responses. Thus, following the adoptive transfer of Fox3+ T cells to T cell deficient mice, these cells could trigger formation of germinal centers in the Peyer’s patches (18). Formation of germinal centers was limited to Peyer’s patches and, to a lesser extend mesenteric lymph nodes, and the process was initiated by Fox3+ T cells that have lost expression of Foxp3. The re-differentiation of Fox3+ T cells appears to be gut-specific, since evidence of this process was not seen in the spleens or peripheral lymph nodes. Mechanisms leading to this re-differentiation of Foxp3 cells remains to be elucidated, but it is possible that gut-derived signals down regulate Foxp3 allowing the expression of the transcription factor Bcl6 and other characteristics of Tfh cells. The re-differentiation of GALT Fox3+ T cells into Tfh could not be confirmed in other studies (19). Nonetheless, a couple of reports suggest that a small subset of mucosal resident Treg can respond to T-dependent antigens and co-express Bcl6, upregulate CXCR5 and migrate into germinal centers to function as regulatory Tfh cells (20, 21).

Th17 cells are preferentially found in mucosal tissues and their presence in the intestine has been shown to depend on stimulation from commensal bacteria, such as Segmented filamentous bacteria (22, 23). Adaptive transfer studies have confirmed the tropism Th17 cells for mucosal tissues of the gut, but also demonstrated that in recipient host, these cells down-regulate their expression of Th17 characteristics (i.e., RORγt and IL-17A expression) to acquire characteristic of Tfh cells, including the transcription factor Bcl6, the surface molecules CxCR5, PD-1, and the ability to produce the IgA-promoting cytokine IL-21 (19). In another study, adoptive transfer of T cells lacking MyD88 into germ-free Rag−/− mice helped demonstrate that gut microbiota-mediated signaling through MyD88 in CD4⁺ T cells induce their differentiation into Tfh cells and promote the production of high-affinity SIgA (24).

A number of studies have now shown that immunoglobulin CSR and production of IgA can be induced in a T-independent manner in lymphoid structures of the GALT (i.e., Peyer’s patches, isolated lymphoid follicles, cryptopatches and mesenteric lymph nodes) (7, 25, 26). It was also suggested that IgA production could take place in the intestinal lamina propria independently of cognate T cell help following interactions of APRIL with the transmembrane activator and CALM interactor (TACI) on B cells (27, 28). It is important to indicate that a major difference between T-dependent and T-independent production of IgA is the source of cytokines that provide help to B cells. Innate lymphoid cells, which are present in high number in mucosal tissues, can produce the same pattern of cytokines than T helper cells and thus may play a central role in the T-independent production mucosal IgA (29). The plasmacytoid dendritic cells (pDCs) are a source of APRIL, BAFF, IL-10 and IL-6 (30), and thus, could be key players in this process. In addition to soluble cytokines, membrane-bound lymphotoxin β (LTα1β2) produced by RORγt⁺ ILCs was shown to be critical for T cell-independent IgA induction in the lamina propria (31). The list of cells potentially involved in the T-independent production of IgA continues to grow and now includes eosinophils, which are widely considered as proinflammatory and associated with allergy and eosinophilic gastrointestinal disorders (32). Eosinophils promote CSR and production of IgA by providing active TGFβ (32).

T-dependent IgA responses predominantly involved B2 cells, while T-independent IgA responses involve both follicular B2 cells and innate B1 cells that reside in the peritoneal cavity. A study that addressed the relative contribution of the B cell subsets to IgA responses toward commensal microbes has shown that most commensals elicit T-independent IgA responses by the orphan B1b and B2 cells, while only atypical commensal such as segmented filamentous bacteria elicits T-dependent IgA responses (33).

- Mucosal homing of IgA producing cells and routes of vaccine delivery

Conventional injected vaccines generally induce specific T cell responses in the bloodstream and serum IgG responses. In contrast with injected vaccines, needle-free vaccines administered via the oral, nasal, rectal, or epicutaneous routes have the potential to induce mucosal IgA, in addition to systemic immunity. In this regard, mucosal homing of immune effector cells is finely orchestrated by expression of mucosal addressins and homing receptors, and chemokine receptors. Retinoic acid (RA) produced by antigen-presenting cells in mucosal tissues provides key signals for expression of the gut homing receptor α4β7 (12, 42). Cytokines such as IL-5 and IL-6 enhance the effect of RA on α4β7. The cytokine IL-21 was shown to induce the expression of the gut-homing receptor α4β7, an effect enhanced by TGFβ1 and RA (5). The chemokine receptor CCR9 together with α4β7 represent the best-described factors regulating homing of immune cells into the gut (43). On the other hand, differentiation of IgA-producing cells in the absence of RA leads to induction of α4β1, L-sectin and CCR10, which target B cells to other mucosal compartments including the airways, salivary gland, reproductive mucosa and colon (43). Oral immunization more effectively targets mucosal DCs, which are rich in RA, and this leads to generation of IgA-producing cells capable of migrating into the small intestine. Rectal immunization, which induces expression of both CCR10+ and CCR9+, generates effector cells capable of migrating on both the small and large intestine. Non-cytokines factors such as Sphingosine1-phosphate can fine tune homing of effector B cells to the intestine (44). Similarly, epicutaneous immunization was reported to promote SIgA responses in the gut. It has been suggested that migration of antigen from the skin to Peyer’s patches and/or mesenteric lymph nodes could be a reason for induction of effector B cells expressing α4β7 after epicutaneous immunization. However, supplementation of epicutaneous vaccine with RA may be the most effective way to consistently induce α4β7+ cells IgA-producing cells capable of migrating to the gut (42)

IgA class switch occurs in the organized structures of the Nasopharyngeal-Associated Lymphoid Tissues (NALT) (45). However, nasal immunization, which induces α4β1, L-sectin and CCR10, is generally not considered an effective approach for inducing SIgA in the gut. This notion has been challenged by a report indicating that NALT DCs possess the machinery for induction of RA and expression of α4β7+ on effector B cells, but stimulation by microbiota was required for these functions to develop (46). The sublingual route is used effectively for delivery of medication and allergen-specific immune therapy in humans and animals (47–50). This route has recently emerged as a mode of vaccine delivery capable of eliciting both systemic and mucosal immune responses while avoiding many of the side effects and formulation concerns associated with oral or intranasal vaccination (51–53). Perhaps because both routes use cervical lymph nodes as inductive sites, sublingual immunization generally induces IgA responses with similar mucosal tropisms than intranasal immunization. The ocular route has also investigated for delivery of experimental vaccines (54, 55). This route of immunization was found to induce SIgA in the tears, but also in the airways and in the genito-urinary tract (54). The nature of homing receptors imprinted in the Tear-Duct-Associated Lymphoid Tissues (TALT) has not been studies in detail. The fact that nasal immunization also induces SIgA in ocular tissues suggests the existence of TALT-NALT cross-talk via the tear duct bridges between the ocular and nasal cavities (55). Hence, ocular vaccines more likely promote the same homing receptors induced by nasal immunization.

- Vaccine adjuvants for induction of SIgA

Earlier studies with cholera toxin (CT) and heat-labile toxin I (LT-I) from Escherichia coli have helped established key principles, which guide current research on mucosal adjuvants. These closely related molecules are AB-type toxins consisting of two structurally and functionally separate enzymatic A subunits and binding B subunits. The B subunit of CT (CT-B) binds to GM1 gangliosides, while the B subunit of LT-I (LT-B) binds to GM1 as well as GM2 asialo-GM1 gangliosides. The A subunits of these toxins are ADP-ribosyl transferases. Binding of the B subunits to gangliosides receptors on target cells allows the A subunits to reach the cytosol where they elevate cAMP via activation of adenylate cyclase. Although the mechanisms underlying the mucosal adjuvant activity of these enterotoxins remain to be fully understood, it is now established that they stimulate APCs to enhance expression of MHC class II and costimulatory molecules, but also IgA-promoting cytokines such as IL-1, IL-6, and IL-10. Their adjuvant activity also induces antigen-specific Th2 and Th17 cells, which support the production of IgA via secretion of IL-4, IL-6, IL-10, and IL-17A (56–58). Despite their efficacy as mucosal adjuvant for needle-free vaccines in experimental animal models, the inherent toxicity of ganglioside-binding bacterial enterotoxins such as CT and LT-I prevent their use in humans.

Approaches used to dissociate the toxicity and the adjuvanticity of the bacterial toxins included partial or complete inactivation of the enzymatic activity of CT (59) and LT (60). Studies with a recombinant derivative of CT, which targets the enzymatic subunit of CT (i.e., CTA1) to B cells (CTA1-DD) have demonstrated that targeting ADP-ribosyl transferase to the right cells could alleviate the unacceptable side effects (CNS targeting and inflammatory responses) of the native CT and LT (61). Our own studies have shown that Bacillus anthracis edema toxin, an adenylate cyclase which targets anthrax toxin receptors and increases cAMP, is a mucosal adjuvant for nasally administered vaccines (62). Edema toxin was found to be potentially safer, as it did not target CNS or promote massive inflammatory responses in airways and lung tissues after intranasal administration (62). We also reported that only minimal (<20%) of edema factor activity was needed for Bacillus anthracis edema toxin to act as a mucosal adjuvant and induce mucosal IgA Ab responses (63). Cyclic di-nucleotides that bind Stimulator of Interferon Gamma Genes (STING) were recently shown to be potential alternatives to cAMP-inducing bacterial toxins and derivatives as vaccine adjuvants. For example, STING ligands of bacterial origin, including 3′3′-cGAMP, c-di-AMP, and c-di-GMP, have been shown to effectively elicit mucosal and systemic immune responses following intranasal administration (64–66). More recently, targeting STING with 3′3′-cGAMP was found to be an effective strategy for enhancing the magnitude of immune responses and promoting IgA by sublingual immunization (67). It is worth indicating that induction of Th17 responses is a common feature of all the adjuvants mentioned above, which enhance IgA responses by increasing intracellular levels of the cyclic nucleotide cAMP or directly releasing select cyclic nucleotides (i.e., STING ligands) in the cytosol of immune cells. This point is of importance since Th17 cells are believed to be crucial for production of high-affinity T-dependent IgA (19).

SIgA responses can be induced by adjuvants that target specific innate signaling pathways, or specific immune cells and their products. For example, CpG ODNs which target TLR9 can act as mucosal adjuvants for nasal vaccines and promote SIgA in a Th1 environment. Studies with plasmid DNA expressing FLt3L have shown that increasing the number of DCs in mucosal inductive sites is a strategy for inducing SIgA and Th2 responses (68). Mucosal tissues contain high numbers of natural killer T cells, which upon stimulation rapidly secrete large amounts of Th1 (IFN-γ and TNF-α), and Th2 (IL-4, IL-5, and IL-13) cytokines (69, 70). They also express CD40L and induce the expression of costimulatory molecules (i.e., CD40, CD80, and CD86) (71–76). The NKT cell ligand α-Galactosylceramide (α-GalCer), originally isolated from a marine sponge, has now been shown to be an effective adjuvant for induction of SIgA by nasal (77) and oral vaccines (78). Mast cell activators represent another class of mucosal adjuvants described during the last decade (79). These studies have shown that the mast cell activator compound 48/80 promotes SIgA responses by stimulating the migration of DCs into the T-cell areas of the NALT (79), and the development of mixed Th1/Th2/Th17 responses (80). Taken together, these studies suggest that targeting cells present in high numbers in mucosal tissues and capable of instantly release several pro-inflammatory mediators is beneficial for induction of SIgA. This notion is consistent with a recent study that showed that this is not the case for adjuvant targeting/recruiting neutrophils, an immune cell subset primarily found in the systemic compartment. Thus, using sublingual vaccination with Bacillus anthracis edema toxin as adjuvant, we identified an inverse relationship between the ability to recruit neutrophils in sublingual tissues and cervical lymph nodes, and the development of SIgA responses (81). It is worth noting that TGF-β and IL-10, two of the key IgA-promoting cytokines antagonize rather than stimulate neutrophil functions. Furthermore, other have shown that depletion of neutrophils in guinea pigs with primary C. caviae ocular infections, reduced rather than enhanced the ocular pathology and that this correlated with increased titers of C. caviae-specific IgA in the tears (82).

This work was supported by National Institutes of Health grants AI18958 and DK101323