Rethinking mucosal antibody responses: IgM, IgG and IgD join IgA

Abstract

Humoral immune responses at mucosal surfaces have historically focused on IgA. Growing evidence highlights the complexity of IgA-inducing pathways and the functional impact of IgA on mucosal commensal bacteria. In the gut, IgA contributes to the establishment of a mutualistic host–microbiota relationship that is required to maintain homeostasis and prevent disease. This Review discusses how mucosal IgA responses occur in an increasingly complex humoral defence network that also encompasses IgM, IgG and IgD. Aside from integrating the protective functions of IgA, these hitherto neglected mucosal antibodies may strengthen the communication between mucosal and systemic immune compartments.

Mucosal surfaces harbour commensal bacteria, archaea, viruses, fungi and protozoa, collectively referred to as microbiota¹. The containment of these microorganisms by the immune system involves a flexible continuum between innate and adaptive responses to ensure commensal–host symbiosis². In general, mucosal commensals mutualistically interact with the host immune system, including antibodies produced by B cells from the adaptive immune system. By mucosal antibodies, we refer to immunoglobulins produced by B cell-rich mucosal inductive sites. The reciprocal interaction of these immunoglobulins with the mucosal microbiota is clearly exemplified by gut IgA³, which derives from B cells receiving stimulating signals from intestinal bacteria⁴. Conversely, IgA responses to intestinal bacteria shape their topography, composition, growth, invasiveness and immunometabolic functions⁴. Remarkably, IgA may control these functions by regulating bacterial gene transcription⁴. Indeed, some IgA-responsive genes regulate the expression of microbial factors with pro-inflammatory or anti-inflammatory function^5,6, whereas others modulate the release of microbial metabolites that shape commensal colonization niches within the host mucosa^7,8.

Antibody-mediated host–microbiota interactions occur in the context of multiple biological networks. Besides innate and adaptive immune effector molecules, these networks include commensals, mucus, epithelial cells as well as haematopoietic and non-haematopoietic cells^9,10. Recent studies have made important advances towards a more holistic understanding of mucosal humoral immunity, and the impact of IgA-secreting plasma cells on inflammation, cancer and neurodegeneration has begun to emerge^11–15. In this Review, we will discuss how IgA cooperates with IgM, IgG and IgD to maximize mucosal homeostasis and immunity.

Mucosal IgA

IgA mostly originates from gut B cell responses to commensals^3,16,17 and IgA responses develop through highly complementary T cell-independent (TI) and T cell-dependent (TD) pathways^4,18,19. While TI-induced IgA broadly targets non-invasive commensals, TD-induced IgA coats penetrant commensals and invasive pathogens^13,20,21. IgA can be detected in both gut mucosa and serum²². In humans, intestinal IgA is dimeric, whereas serum IgA is largely monomeric (Box 1). In mice, both gut and serum IgA are dimeric¹⁶. Gut IgA originates from long-lived plasma cells that also express an IgA-interacting polypeptide termed the joining chain^23,24. Besides stabilizing the IgA dimer, the joining chain serves as a ligand for the polymeric immunoglobulin receptor (pIgR), an IgA transporter expressed by mucosal epithelial cells²⁵. As discussed later, pIgR also transports pentameric IgM, but not monomeric IgG and IgD. Intraluminal IgA, called secretory IgA (SIgA), further includes a pIgR-derived fragment termed the secretory component²⁵. This polypeptide covalently binds to SIgA to augment its stability and provide mucus-anchoring sites^25,26.

Box 1 |. Species-specific properties of mucosal IgA, IgM, IgG and IgD.

Some properties of mucosal IgA, Igm, IgG and IgD antibodies differ in humans and mice. These differences may be due to the distinct molecular configuration of the immunoglobulin heavy chain locus, the distinct structure of the antibody molecule and/or discrepancies in microbiota-driven signals that shape the magnitude and quality of mucosal humoral immunity. IgA exemplifies these species-specific differences. As detailed in Box 2, mucosal IgA includes two subclasses, IgA1 and IgA2, in humans but not in mice^97,119. Accordingly, human IgA1 binds to Fcα receptor I (FcαRI), also known as CD89 (reF.³⁴), whereas mouse IgA does not bind due to the lack of an FcαRI homologue. of note, the FcαRI receptor is highly expressed by mucosal and hepatic phagocytes, including Kupffer cells, which facilitates the clearance of secretory IgA (SIgA)-coated bacteria that leak through the gut epithelium³⁴. As for Igm, gut plasma cells releasing this antibody are virtually absent in the mouse gut lamina propria, whereas they account for up to 20% of total plasma cells in the human gut lamina propria³⁰. As a consequence, the concentrations of free secretory Igm (SIgm) and the frequency of SIgm-coated bacteria are negligible in the intestinal lumen from mice compared with humans³⁰. The low magnitude of gut SIgm responses in mice could relate to B cell-intrinsic differences, but B cell-extrinsic differences linked to the gut microbiota could also be at work³⁰. Few data are available regarding gut IgG responses, but commensal-specific IgG1 responses detected in mice may be equivalent to human IgG4 (and IgG2) responses detected in humans¹²¹. Finally, while human mucosal IgD includes up to 80 V(D)J gene mutations and has a clear germinal centre derivation^138,157, mouse mucosal IgD is poorly mutated and seems to mostly stem from an early extrafollicular response¹⁴². of note, human IgD encompasses a longer hinge region compared with mouse IgD^97,137. O-glycans from this hinge region may mediate IgD interaction with mammalian galectins, such as galectin 9, and plant lectin proteins, such as jacalin, similar to O-glycans from IgA1 (reFs^142,158).

Having co-evolved with gut bacteria during human evolution, IgA follows patterns of mucosal production that are influenced by the microbiota^4,27. This collection of microbial communities is acquired from both the mother and the environment after birth but, in humans, does not reach adult-like complexity and stability until 2 years of age^4,27,28. The gradual maturation of the gut microbiota is paralleled by the progressively increased clonal diversity of gut IgA^4,27,29. This process leads to the extensive coating of gut bacteria by SIgA, particularly in the energy-rich milieu of mucus^4,13,30.

The mechanisms through which intestinal SIgA promotes host–microbiota symbiosis remain unclear, but include immune exclusion. By confining SIgA-coated commensals in the lumen of the gut, immune exclusion minimizes pro-inflammatory interactions of the gut immune system with a myriad of microorganism-associated molecular patterns, including Toll-like receptor (TLR) ligands, that are produced by endogenous commensals. Of note, SIgA can implement immune exclusion in a non-inflammatory manner by interacting with commensals without activating the complement cascade^4,31. However, in humans, SIgA can elicit inflammation by cross-linking the Fcα receptor I (FcαRI) on pro-inflammatory phagocytes following penetration of SIgA-coated commensals into the gut mucosa^32–34. This pro-inflammatory property of human SIgA further underscores the importance of immune exclusion in gut homeostasis. Of note, mice have no molecular equivalent of human FcαRI, but instead express a still poorly understood Fcα/μ receptor (Fcα/μR)³⁵.

SIgA further enhances host–microbiota symbiosis by targeting molecules implicated in the growth of gut commensals, including the metabolic enzyme serine hydroxymethyltransferase⁶. Moreover, SIgA promotes the agglutination of gut commensals, which limits their motility³⁶. Recently, agglutination has also been found to be involved in SIgA-mediated protection against intestinal pathogens. Indeed, aside from blocking or neutralizing key pathogenicity factors³⁷, SIgA repels rapidly dividing bacteria by cross-linking post-division daughter cells, thereby preventing their separation³⁸. This SIgA-mediated ‘enchainment’ strategy accelerates mucosal clearance of pathogens while promoting mucosal resilience of commensals³⁸.

Another mechanism whereby SIgA favours host–microbiota symbiosis is referred to as immune inclusion. In this process, the stable retention of commensals in the gut mucosa is facilitated by microbial regulatory systems that change the molecular architecture of commensals to invite their binding by SIgA⁷. This binding facilitates the colonization of defined mucosal niches by endogenous commensals through the exclusion of exogenous competitors⁷. Immune inclusion further involves anchoring of SIgA-coated commensals to the gut mucous layer lining the epithelial surface through a mechanism involving glycans from the secretory component of SIgA²⁶.

Finally, host–microbiota symbiosis is further ensured by the non-inflammatory nature of the dominant TD pathway that induces IgA in germinal centres from mucosal lymphoid follicles, including Peyer’s patches^37,39. This pathway involves IgA induction by bacteria-specific B cells that undergo stringent microbiota-regulated selection via follicular T cells expressing powerful anti-inflammatory molecules, such as the cytokine transforming growth factor-β (TGFβ) and the inhibitory co-receptor programmed cell death 1 (PD1)^40–42. Remarkably, human but not murine gut IgA includes two structurally distinct intestinal IgA1 and IgA2 subclasses (Box 2), but their relative contribution to host–microbiota symbiosis remains elusive.

Box 2 |. Mucosal IgA1 and IgA2 subclasses in humans.

Human IgA includes IgA1 and IgA2 subclasses. While IgA2 is mostly confined to the gut and is more abundant than IgA1 in the distal intestine, IgA1 is released in both systemic and mucosal compartments⁸⁶. The biology and relative contribution of IgA1 and IgA2 to mucosal immunity remain poorly understood. Structurally, IgA2 has a shorter hinge region compared with IgA1, which makes IgA2 more resistant to degradation by bacterial proteases^63,86. Conversely, IgA1 has a longer and heavily O-glycosylated hinge region that may increase its binding flexibility compared with IgA2 (reFs^63,86). Both IgA1 and IgA2 encompass an N-glycosylated Fc segment that outcompetes influenza viruses for binding to sialic acid receptors on target cells besides interacting with both joining chain and secretory component segments¹⁵⁹. of note, some IgA2 emerges from sequential IgA1-to-IgA2 class switch recombination (CSR), which targets Cα1 and Cα2 genes positioned at proximal 5′ and distal 3′ sites within the immunoglobulin heavy chain locus, respectively^71,160,161. In adults with stable exposure to the gut microbiota, B cells may express IgA2 by undergoing sequential IgA1-to-IgA2 CSR in addition to direct Igm-to-IgA2 CSR⁷¹. Sequential CSR could involve pre-existing IgA1⁺ memory B cells stimulated by T cell-dependent signals upon re-entry into mucosal germinal centres or by T cell-independent signals upon colonizing extra-germinal centre sites of the intestine. The specific stimuli required for IgA1 and IgA2 production remain poorly understood, but IgA1 may largely depend on T cell-dependent signals from CD40 ligand (CD40L), whereas IgA2 may also involve T cell-independent signals from innate CD40L-like factors such as a proliferation-inducing ligand (APRIl)^17,71. Functionally, IgA1 binds to the high-affinity Fcα receptor I (FcαRI) receptor, which has both activating and inhibitory functions in phagocytes³³. In addition or alternatively to FcαRI, IgA2 binds to dectin 1, which allows m cells to capture secretory IgA2-coated gut antigens⁵⁶. Similar to FcαRI, dectin 1 may deploy both activating and inhibitory signals in epithelial and myeloid cells.

How gut bacteria interact with mucosal IgA.

SIgA can recognize gut bacteria through multiple strategies, including specific interaction with the antigen-binding Fab segment of SIgA and non-specific interaction with the Fc and secretory component segments of SIgA^4,31. Protective recognition of a specific bacterial antigen usually involves high-affinity IgA from the TD pathway (Fig. 1). This humoral response often involves IgG in addition to IgA and mainly targets penetrant commensals colonizing the gut epithelium, including Escherichia coli and segmented filamentous bacteria^43–45. Remarkably, distinct microorganisms shape IgA-inducing T cell responses in different ways. While commensals such as Akkermansia muciniphila elicit T follicular helper (T_FH) cell-dominated IgA responses⁴⁶, segmented filamentous bacteria trigger T helper 17 (T_H17) cell-dominated IgA production, at least in mice^43,44,47. Finally, regulatory T (T_reg) cell-dominated IgA responses expand beneficial Firmicutes bacteria while constraining potentially harmful Proteobacteria, thereby modulating the composition of the gut microbiota⁴⁸. Yet the identity of the bacterial antigens targeted by each of these IgA responses remains elusive. One of these antigens is flagellin, a bacterial motility protein that requires innate recognition by TLR5 to elicit mucosal SIgA and systemic IgG production^49–51. By deploying cross-reactivity for flagellin epitopes across a broad spectrum of bacteria, flagellin-specific SIgA quenches the gut microbiota in cooperation with IgG by multiple mechanisms, including downregulation of flagellin-related gene expression⁴⁹. Of note, the lack of immune responses to flagellin caused by TLR5 deficiency increases breaching of the gut barrier by Proteobacteria and Firmicutes⁴⁹.

Fig. 1 |. Inductive pathways and protective strategies of intestinal IgA and IgM.

Gut IgA emerges from highly complementary T cell-dependent (TD) and T cell-independent (TI) pathways. In the TD pathway, microfold cells from the follicle-associated epithelium of Peyer’s patches capture naked or secretory IgA (SIgA)-coated commensals, which are then transferred to subepithelial dendritic cells (DCs) or B cells. These cells establish cognate interactions with T follicular helper (T_FH) cells in the subepithelial dome (SED) underneath the follicle-associated epithelium. After initiating IgM-to-IgA class switch recombination (CSR) in the SED, T_FH cell-activated B cells enter germinal centres (GCs) from Peyer’s patches to undergo IgA somatic hypermutation. In addition to plasma cells, this GC reaction generates memory B cells with high affinity for highly penetrant members of the microbiota. Although GCs from young individuals recruit de novo activated naive B cells, GCs from adults with a stable microbiota mainly recruit pre-existing memory B cells to generate ‘IgA-edited’ plasma cells. This second strategy permits intestinal GCs to continually adjust IgA responses to small changes of the microbiota. While memory B cells re-enter intestinal GCs, plasma cells home to the gut lamina propria to release dimeric IgA, which then translocates across the gut epithelium via the polymeric immunoglobulin receptor (pIgR). The resulting SIgA binds to intraluminal commensals. In humans, the TD pathway also generates secretory IgM (SIgM) clonally affiliated with some SIgA and both target a fraction of commensals. SIgA mediates immune exclusion but also immune inclusion and selection by anchoring some bacteria to metabolically regulated competitive niches via mucus-associated functional factors (MAFF), which exclude less beneficial microorganisms. Moreover, SIgA quenches the motility, growth and pro-inflammatory properties of commensals, regulates their metabolic output and promotes their sampling by M cells. In the TI pathway, extrafollicular B cells generate additional polyreactive and low-affinity SIgA that recognizes a broad spectrum of commensals. FDC, follicular dendritic cell; ILF, isolated lymphoid follicle.

Recognition of gut commensals by SIgA also involves the TI pathway (Fig. 1), which generates low-affinity SIgA that specifically binds to multiple bacterial taxa, at least in mice²¹. However, Fab-mediated specificity may not be mandatory⁸. Indeed, non-specific glycan-mediated interaction of Bacteroidetes thetaiotamicron with the Fc domain of SIgA has been recently shown to modulate bacterial expression of polysaccharide utilization loci collectively named mucus-associated functional factor (MAFF) genes⁸. These genes (Fig. 1) encode bacterial molecules that engender gut symbiosis between B. thetaiotamicron and mucus-embedded bacteria of the Firmicutes phylum⁸. By altering the metabolic profile of some of them, including Clostridiales, MAFF gene products promote competitive mucosal colonization of B. thetaiotamicron⁸. Thus, the function of specific commensals changes upon non-specific microbial binding to glycans from SIgA and this change promotes symbiosis of bacterial species required for gut homeostasis⁸. Along the same line, a sensor/regulatory system shapes the surface architecture of the human commensal Bacteroides fragilis to invite its binding by SIgA⁷. By facilitating glycan-mediated microbial adherence to epithelial cells⁷, SIgA facilitates the colonization of a stable mucosal niche by bacteria that exclude other competitors to promote robust host–microbiota symbiosis.

How gut antigens initiate mucosal IgA production.

The follicle-associated epithelium (FAE) of Peyer’s patches is a well-studied site of gut antigen entry⁵². FAE-like structures also mediate antigen entry into isolated lymphoid follicles in both the small and large intestines^52,53. The FAE lacks mucus-secreting goblet cells⁵², which facilitates antigen sampling by microfold cells positioned in the FAE (Fig. 1). Besides capturing unbound or SIgA-bound antigens via glycan-mediated interactions with glycoprotein 2, dectin 1 and, possibly, other carbohydrate-binding receptors^52,54–56, M cells deliver captured antigens to dendritic cells (DCs)^52,57, which then interact with T_FH cells to initiate high-affinity IgA production by B cells^52,58.

Soluble antigens are also captured by interepithelial projections from CX₃CR1⁺ macrophages, which elicit TD-induced IgA responses by transferring antigen to CD103⁺ DCs via connexin 43-expressing gap junctions⁵⁹. These CX₃CR1⁺ macrophages express receptors for antigen-complexed SIgA, which shape the quantity and quality of gut IgA production depending on the poly/monoreactive nature of SIgA⁶⁰. Gut CD103⁺ DCs can also receive soluble antigens from goblet cells via goblet cell-associated passages⁶¹. Finally, M cells can initiate TD IgA production by directly transferring antigen to B cells from the subepithelial dome (SED) of Peyer’s patches⁶².

Induction and regulation of mucosal IgA production.

Recent studies show that a complex network of cellular interactions jump-starts IgM-to-IgA class switch recombination (CSR) in extrafollicular B cells from the SED⁵⁸. TD induction of gut IgA is commonly thought to begin with cognate interaction of DCs with T_FH cells, which subsequently migrate to the interfollicular area of Peyer’s patches^18,63. There, cognate interaction of T_FH cells with B cells drives the germinal centre reaction⁶⁴, including IgM-to-IgA CSR⁶³. In addition to CD40 ligand (CD40L), this process requires TGFβ⁶³, which originates from T_reg cells or T_reg cell-derived T_FH cells^40,48,65. In this TD model, germinal centres are central to IgA induction, but new evidence now shows that the SED plays a central role (Fig. 1). Aside from CCL20-responsive CCR6⁺ B cells, the SED includes RORγt⁺ group 3 innate lymphoid cells (ILC3s) that establish lymphotoxin-dependent interaction with local CD11b⁺ DCs⁵⁸. This interaction supports the survival of CD11b⁺ DCs, which in turn initiate IgA induction in B cells⁵⁸. Indeed, DC-activated B cells receive integrin αvβ8-dependent signals that induce active TGFβ, which is essential for IgM-to-IgA CSR along with CD40L⁵⁸. In this model, CD40L derives from T_FH cells⁵⁸, but a contribution from ILC3s cannot be excluded a priori⁶⁶. By demonstrating the mandatory role of the SED in the initiation of IgM-to-IgA CSR, these new data expand our mechanistic understanding of the SED and, together with more recent findings, add a novel extrafollicular facet to the TD pathway leading to IgA induction and affinity maturation in the follicular environment of intestinal germinal centres^52,58,62,67. Of note, SED-based T_FH cells promote B cell expansion and initiate IgM-to-IgA CSR without eliciting B cell clonal competition, which instead takes place in the germinal centre⁶⁸.

Aside from eliciting CD11b⁺ DC survival in the SED, mouse ILC3s from gut-draining lymph nodes deploy MHC class II-dependent regulatory functions by acting as an innate checkpoint that constrains adaptive TD-induced IgA responses to colonic bacteria⁶⁹. These data integrate earlier studies showing impaired diversification and selection of germinal centre B cells and reduced SIgA-coated gut bacteria in mice with dysregulated T_FH cell expansion due to the lack of regulatory signals from the PD1 receptor^40,48.

The TI pathway mounts IgA responses complementary to those from the TD pathway (Fig. 1), in that TI-induced IgA controls a broader spectrum of bacteria^20,21. As opposed to the TD pathway, which relies on CD40L from T cells, the TI pathway relies on microbial TLR ligands, B cell activating factor (BAFF) of the tumour necrosis factor family, a proliferation-inducing ligand (APRIL) and other innate signals from the gut microbiota, DCs, stromal cells and epithelial cells^17,18,63. Recent mouse data show that endoplasmic reticulum stress from gut epithelium elicits a TI polyreactive SIgA response that protects against inflammation and involves peritoneal B1b cell differentiation into gut-homing IgA-secreting plasma cells⁷⁰. These findings echo earlier studies proposing a direct involvement of gut epithelial cells in TI induction of IgA^71,72.

Microbiota metabolites in mucosal IgA production.

The gut microbiota generates metabolites that derive from the breakdown of proteins, fat, fibres and other dietary constituents⁷³. These metabolites include short-chain fatty acids (SCFAs)^73–75, such as acetate, propionate and butyrate, and can modulate gut IgA production^73,76. Mouse studies show that SCFAs enhance both homeostatic and post-immune IgA responses through various mechanisms (Fig. 2). First, SCFAs accelerate B cell metabolism to generate more energy and building blocks, such as ATP and acetyl-CoA, required for IgA synthesis⁷⁷. Second, SCFAs engage molecular pathways that are directly or indirectly (that is, via T cells or DCs) implicated in gut B cell activation and differentiation, such as G protein-coupled receptor 43 (GPR43), GPR109A and mammalian target of rapamycin (mTOR)^74,75,77. Third, SCFAs increase gut B cell expression of genes involved in IgM-to-IgA CSR (for example, activation-induced cytidine deaminase (AID)), plasma cell differentiation (for example, BLIMP1) and IgA production⁷⁷. Given their small molecular size, SCFAs are readily absorbed by the gut to also influence systemic IgG production^73,77.

Fig. 2 |. Impact of gut metabolites on gut IgA and systemic IgG responses.

Fermentation of dietary fibres by gut bacteria generates short-chain fatty acids (SCFAs), including acetate (C2), propionate (C3) and butyrate (C4). Engagement of G protein-coupled receptor 43 (GPR43) on T cells or GPR109A on dendritic cells (DCs) as well as inhibition of histone deacetylase (HDAC) in T cells by SCFAs elicits anti-inflammatory immune responses poised on enhancing gut homeostasis. These responses include induction of DCs with enhanced ability to prime FOXP3⁺ regulatory T (T_reg) cells, increased differentiation and expansion of IgA-inducing FOXP3⁺ T_reg cells and augmented conversion of T_reg cells into professional IgA-inducing T follicular helper (T_FH) cells. Aside from increasing gut IgA responses by shaping the B cell-activating function of T cells, SCFAs augment gut IgA production by modulating B cells. Indeed, SCFAs increase the differentiation of IgA-secreting plasma cells by augmenting B cell expression of the class switch-inducing enzyme activation-induced cytidine deaminase (AID) and by increasing IgA synthesis as well as plasma cell differentiation. Due to their small molecular size, SCFAs can also reach systemic lymphoid organs via the general circulation after being absorbed by the gut mucosa. In systemic lymphoid organs, SCFAs enhance IgG responses by supporting the formation of IgG class-switched plasma cells that secrete IgG. In addition to SCFAs, the gut microbiota as well as gut host cells release ATP, which engages the ATP-gated purinergic receptor P2X₇ on T_FH cells. Engagement of P2X₇ by ATP constraints the expansion of T_FH cells in intestinal germinal centres, thereby increasing the stringency of the selection of IgA-expressing B cells specific to commensals. BCR, B cell receptor; CSR, class switch recombination; SIgA, secretory IgA; TGFβ, transforming growth factor-β.

SCFAs also generate IgA-inducing T_reg cells (Fig. 2) by inhibiting histone deacetylase and activating mTOR^78,79. This pathway also converts T_reg cells into IgA-inducing T_FH cells^40,80. In general, SCFAs activate gut T cell and B cell responses in a mucosal context poised on the preservation of homeostasis^75,79,81. Indeed, both SCFA-induced T_reg cells and IgA prevent mucosal inflammation⁸². Similarly to SCFAs, ATP from both host cells and commensals regulate TD-induced SIgA (Fig. 2). Indeed, T_FH cells express the ATP-gated ionotropic receptor P2X₇, which constrains mucosal T_FH cell expansion and the germinal centre reaction⁸³. Accordingly, transient depletion of ATP enhances gut IgA responses to oral vaccines⁸⁴.

While microbial metabolites influence IgA responses, gut IgA regulates microbial metabolites, including their penetration into systemic tissues (Fig. 2). In particular, SIgA curtails the exposure time of penetrant metabolites to the colonic mucosa by accelerating colonic transit of bacteria⁸⁵. This is a key finding given that gut metabolites shape the development and function of virtually every organ in our body⁸⁵. Thus, besides controlling the short-range effects of commensals in the gut mucosa, SIgA regulates the long-range impact of gut bacteria on systemic organs by constraining the penetration of their metabolites across the intestinal barrier.

Impact of age and microbial complexity on mucosal IgA.

Age is an important variable underpinning gut IgA responses. Indeed, the clonal architecture of gut IgA varies in parallel with age-dependent changes in the composition of the gut microbiota²⁹. Growing evidence is modifying the earlier notion that gut IgA emerges from naive follicular B cells that enter mucosa-associated lymphoid follicles^18,63,86. While this canonical TD pathway is certainly active in young individuals exposed to a progressively developing gut microbiota, adults with a stable microbiota may produce most of their high-affinity IgA via an alternative TD pathway involving re-entry of pre-existing gut IgA⁺ memory B cells into germinal centres^29,39. By stimulating immunoglobulin gene somatic hypermutation and plasma cell differentiation, this alternative TD pathway permits the adjustment of high-affinity SIgA responses to small antigenic changes in the gut microbiota^29,39. More dramatic changes would be met by high-affinity SIgA from the canonical TD pathway ^29,39, which involves de novo recruitment of naive B cells into germinal centres^18,63.

Age may also dictate the choice of IgA induction via high-affinity TD pathways over low-affinity TI pathways. These latter pathways are clearly at work in mice, where they generate a large pool of SIgA that binds to a broad spectrum of bacteria^17–20. Of note, TI-induced SIgA features antigen specificity²¹, but is of lower affinity compared with SIgA from TD-induced responses from mucosal germinal centres³⁷. Remarkably, the murine gut microbiota generates antibody-editing signals in intestinal naive B cells during a short time-window immediately after birth⁸⁷. In this way, the gut naive B cell repertoire would generate antigen recognition diversity independently of random immunoglobulin V(D)J diversification in the bone marrow⁸⁷. Should this be the case, gut B cells may acquire low-affinity specificity for commensal antigens very early in life^19–21.

In humans, who have a longer lifespan and a more complex microbiota than mice, the TD pathway is dominant over the TI pathway during adult life²⁵. However, TI signals may be at work even in adults, as they may induce IgM-to-IgA CSR in pre-existing gut IgM⁺ memory B cells³⁰. The TI pathway might play a more important role in neonates and children at a time when the gut microbiota is still immature²⁷. However, as soon as the gut microbiota acquires an adult-like configuration, humans develop a heavily germinal centre-based mode of gut IgA production^16,86. This (super-dominant) TD pathway may be needed by humans to impede any breach of the gut barrier by highly penetrant and relatively abundant commensals^3,4. Indeed, the human TD pathway relies on redundant gut-associated follicular structures that include Peyer’s patches, mesenteric lymph nodes and a very diverse set of isolated lymphoid follicles disseminated along both the small and large intestines^16,86. All of these follicles may be needed to support combined IgA and IgM responses to beneficial and yet potentially harmful microorganisms via intertwined TD and TI pathways^4,17,30. Accordingly, the ‘human-like’ gut microbiota from outbred wild mice is superior in improving host fitness and resistance to disease compared with the gut microbiota from inbred laboratory mice⁸⁸.

Mucosal IgA interaction with systemic IgM and IgG.

A fraction of gut IgA-secreting plasma cells may home to the bone marrow to induce serum accumulation of dimeric IgA with broad antimicrobial reactivity, at least in mice²². This IgA emerges from a TD pathway that requires gut colonization by commensal communities encompassing members of the Proteobacteria phylum²². Gut IgA may further converge with systemic IgM to provide cross-specific protection against Klebsiella pneumoniae, an opportunistic pathogen⁸⁹. Of note, human IgM and IgA were shown to recognize glycans associated with lipopolysaccharide O-antigen from different species of K. pneumoniae as well as other commensals⁸⁹. Thus, the recognition of minimal glycan epitopes expressed by various microorganisms could facilitate the control of a broad array of pathogens and commensals with a limited set of cross-protective IgA and IgM specificities released both at mucosal sites of antigen entry and systemically.

The main function of serum dimeric IgA of intestinal origin would relate to humoral defence against polymicrobial sepsis caused by systemic translocation of gut bacteria resulting from catastrophic disruption of the intestinal barrier²². Of note, commensal-reactive serum dimeric IgA would implement such defensive activity in cooperation with commensal-reactive serum monomeric IgG and would do so not only in mice but also in humans^90,91. However, species-specific differences between mouse and human serum IgA must be considered, including the fact that most human IgA is monomeric and therefore may not have a gut origin⁸⁶.

Aside from protecting against sepsis in adults, IgA may cooperate with IgG to promote gut homeostasis in neonates⁹². In humans, SIgA specific to gut commensals may be acquired by neonates upon migration of maternal IgA class-switched B cells from TD intestinal IgA-inductive sites to the glandular mammary tissue of the lactating mother via lymphatic vessels and the general circulation⁹³. Homing of intestinal IgA class-switched B cells to the breast would be followed by differentiation into IgA-secreting plasma cells and pIgR-mediated translocation of IgA across the epithelium of mammary excretory ducts into the milk⁹³. The resulting SIgA would then reach the gut mucosa of the suckling neonate and bind to the local microbiota, which largely derives from the mother as a result of her interaction with the neonate through the vaginal mucosa at the time of delivery and the subsequent symbiotic relationship with the neonate^92,93.

In mice, breast-derived IgG to commensals emerges from a microbiota-dependent TI pathway that activates maternal B cells via TLR2 and TLR4 (reF.⁹²). Plasma cells from this TI pathway secrete IgG2b or IgG3 subclasses and home to the mammary gland⁹². Subsequent IgG2b and IgG3 translocation across mammary ducts via the neonatal Fc receptor (FcRn) leads to IgG2b and IgG3 accumulation in maternal milk and then the neonatal gut⁹². Then, binding of IgG2b and IgG3 to the gut microbiota quells pro-inflammatory T cell and B cell responses in cooperation with TI-induced SIgA through elusive mechanisms^20,92. In addition to trapping commensals into the lumen of the neonatal gut, milk-derived IgG2b and IgG3 could promote bacterial sampling by epithelial cells via the FcRn, followed by FcRn-mediated bacterial clearance by phagocytes. Reversed transcytosis of IgG-coated bacteria would also reinforce the gut epithelium by expanding ILC3 populations that release the cytokine IL-22 (reF.⁹⁴). In the absence of maternal TI-induced IgG and SIgA, compensatory TD-induced neonatal IgG and SIgA responses would ultimately promote gut homeostasis⁹². Maternal antibodies also protect neonates against necrotizing enterocolitis, which mostly affects preterm neonates not sufficiently exposed to maternal milk⁹⁵. Indeed, milk-derived SIgA protects against necrotizing enterocolitis, which correlates with binding of maternal SIgA to gut bacteria in the first month of life⁹⁵. Before the onset of necrotizing enterocolitis, uncoated gut bacteria are enriched with potentially pathogenic Enterobacteriaceae compared with SIgA-coated bacteria⁹⁵. Thus, maternal SIgA shapes neonate–microbiota symbiosis by promoting gut homeostasis.

Mucosal IgM

IgM is the most ancient member of our antibody family⁹⁶ and the first antibody to appear during B cell ontogeny. IgM is also the first antibody released during a humoral response⁹⁷. Serum IgM forms pentamers associated with the joining chain and includes monoreactive IgM from conventional B cells, also termed B2 cells, and polyreactive IgM from innate-like B1 cells⁹⁸. The latter IgM is readily detected in mice, but may also exist in humans^99,100. In general, IgM generates a first line of humoral defence against pathogens, but may also contribute to tissue homeostasis¹⁰⁰.

General properties of mucosal IgM responses.

Secreted IgM exerts multiple immune effector functions. In particular, IgM activates complement as effectively as IgG, and this is instrumental for microbial clearance and initiation of inflammation¹⁰⁰. In addition to immunoactivating signals from complement receptors, IgM deploys immunoregulatory signals from Fcα/μR and Fcμ receptor (FcμR)^35,101. Although only immune cells express FcμR¹⁰¹, both haematopoietic and non-haematopoietic cells express Fcα/μR, including splenic innate-like marginal zone B cells and germinal centre-based follicular dendritic cells³⁵. Of note, Fcα/μR delivers regulatory signals that shape humoral immunity, whereas FcμR modulates both early and late phases of B cell development^35,101.

Additional IgM receptors have been identified in pathogens such as Toxoplasma gondii, Trypanosomatidae and Plasmodium falciparum¹⁰², which might facilitate evasion from protective IgM responses. Consistent with the complex function of IgM, patients with selective IgM deficiency and mice selectively lacking secreted IgM suffer from recurrent viral, bacterial and protozoal infections¹⁰³. Moreover, IgM-deficient patients show an increased incidence of IgE-mediated allergic and IgG-mediated auto immune disorders¹⁰³. Thus, similarly to mouse IgM, human IgM is likely involved in both immunoprotective and immunoregulatory functions.

Besides being released in the circulation by short-lived plasma cells from lymph nodes and the spleen as well as long-lived plasma cells from the bone marrow and the spleen^100,101,104, IgM enters mucosal secretions as secretory IgM (SIgM)²⁵, albeit at low concentrations. Indeed, similarly to dimeric IgA, pentameric IgM translocates across the mucosal epithelium upon binding the pIgR via the joining chain²⁵. The resulting secretory component fragment does not covalently bind to IgM as it does to IgA and, consequently, does not efficiently protect SIgM against proteolytic degradation^25,86.

Despite its relatively low stability, SIgM may provide mucosal protection very early in life, prior to the progressively dominant production of SIgA. Accordingly, SIgM at least partially compensates for the lack of SIgA in children and adults with selective IgA deficiency¹⁰⁵. Indeed, about 50% of these patients are asymptomatic, which correlates with increased SIgM release into the gastrointestinal mucosa^105,106. Similarly, IgA-deficient mice show no remarkable phenotype aside from an increase of IgM and certain IgG subclasses both systemically and at mucosal sites of antigen entry^20,107. Yet human intestinal SIgM and SIgA are not completely redundant, because SIgM shows only limited binding to Enterobacteriaceae compared with SIgA in patients with selective IgA deficiency¹⁰⁵. Moreover, SIgM binds to a broader set of gut bacteria compared with SIgA, which instead targets a more specific subset of commensals with a higher level of microbial coating compared with SIgM^105,108. Furthermore, human SIgM does not fully restore microbiota diversity and spatial organization in IgA-deficient patients, who show significant shifts in the relative abundances of specific microbial taxa in addition to decreased microbial diversity and increased gut colonization by species usually inhabiting the oropharynx^105,108,109.

Cooperation of mucosal IgM and IgA responses.

Besides coating the gut microbiota from IgA-deficient patients¹⁰⁵, SIgM may enhance gut bacterial diversity in healthy individuals. Indeed, human SIgM targets consortia of gut bacteria dually coated by SIgA³⁰. These bacteria are enriched in beneficial Firmicutes and show increased diversity compared with IgA-only coated bacteria³⁰. Given the mucus-binding properties of the secretory component²⁶, SIgM may cooperate with SIgA to stably retain beneficial commensals within gut mucus³⁰. Unlike humans, specific pathogen-free mice commonly used in laboratories lack SIgM-coated gut bacteria^20,30. This is consistent with the scarcity of gut IgM-secreting plasma cells in the mouse³⁰. However, both SIgM production and SIgM-coated bacteria increase in specific pathogen-free mice with chemically induced colitis, which may prevent lethal systemic dissemination of commensals across the damaged gut epithelium¹¹⁰. Accordingly, a combination of exogenous human plasma-derived secretory-like IgM and IgA has been shown to prevent bacterial dissemination and gut inflammation as efficiently as endogenous SIgA in mice¹¹¹.

In agreement with earlier studies¹¹², IgM-secreting plasma cells are particularly abundant in the human terminal ileum, where they account for approximately 10–20% of total plasma cells³⁰. Similarly to IgA⁺ plasma cells, gut IgM⁺ plasma cells retain surface antigen receptor expression, which could reflect their need to receive antigen-specific signals for their survival and/or function¹¹³. However, compared with IgA⁺ plasma cells, gut IgM⁺ plasma cells show decreased CD138 expression, which could reflect a less advanced maturation stage³⁰. In adults, gut IgM⁺ plasma cells are mutated and clonally affiliated to gut memory IgM⁺IgD⁻CD27⁺ B cells³⁰. These B cells colonize the intestinal mucosa early in life, mainly inhabit gut-associated lymphoid follicles and have a post-germinal centre immunoglobulin V(D)J gene mutation profile³⁰. Consistent with their affiliation to some IgA⁺ clones, gut memory IgM⁺IgD⁻CD27⁺ B cells undergo IgM-t o-IgA CSR in response to TI or TD signals³⁰. Thus, in a similar manner to IgA responses³⁹, gut IgM responses may involve IgM diversification from pre-existing IgM⁺IgD⁻CD27⁺ memory specificities rather than de novo recruitment of naive IgM⁺IgD⁺CD27⁻ B cells. This strategy would allow adults to rapidly accommodate SIgM to small and transient antigenic changes of the gut microbiota. It could also lead to the co-release of SIgM and SIgA with identical microbial specificity to maximize mucus retention of beneficial commensals.

In mice, germinal centre-derived splenic IgM⁺ memory B cells emerging from canonical gut inductive sites have been recently described¹¹⁴. This B cell population resembles human gut memory IgM⁺IgD⁻CD27⁺ B cells as well as human gut marginal zone-like IgM⁺IgD^lowCD27⁺ B cells from the gut SED¹¹⁵. Contrary to gut memory IgM⁺IgD⁻CD27⁺ B cells, gut marginal zone-like IgM⁺IgD^lowCD27⁺ B cells rarely clonally relate to IgA⁺ memory B cells¹¹⁵. Of note, it remains unclear whether gut marginal zone-like IgM⁺IgD^lowCD27⁺ B cells contribute to the induction of mucosal IgM⁺ plasma cells.

Mucosal IgG

IgG is more abundant than or as abundant as IgA in reproductive and lower respiratory tracts^86,112. In the upper respiratory tract, gut mucosa as well as mammary and lachrymal glands, IgG is less abundant than IgA, but its concentration increases in response to infection^{16,25,86,116,117}. In general, IgG enhances mucosal homeostasis in addition to controlling non-invasive and invasive mucosal bacteria, and does so by reaching the lumen of mucosal organs upon binding to the epithelial transporter FcRn¹¹⁸. In the lumen, monomeric IgG likely exerts its homeostatic and defensive functions in cooperation with polymeric SIgA and SIgM as well as monomeric IgD.

Mucosal IgG subclasses.

In humans, mucosal IgG responses are pro-inflammatory when they involve IgG1 and IgG3 subclasses with complement-activating and Fcγ receptor I (FcγRI)/FcγRIII signalling functions¹¹⁹. Conversely, IgG responses are mostly non-inflammatory when they involve IgG2 and IgG4 subclasses with little or no complement-activating and FcγRI/FcγRIII signalling functions¹¹⁹. Accordingly, gut inflammation elicits IgG1 and IgG3 (reF.¹²⁰), whereas a healthy gut induces IgG2 and IgG4 (reF.¹²¹). Remarkably, human IgG4 is equivalent to mouse IgG1 (reF.¹¹⁹), which targets a beneficial gut commensal called A. muciniphila⁴⁶. This mucosal IgG1 response requires T_FH cells, which select germinal centre B cells with higher affinity for antigen⁴⁶. Considering that A. muciniphila reaches intestinal epithelial cells by means of its mucus-degrading properties⁴⁶, gut IgG responses may predominantly target highly penetrant mucus-dissolving microorganisms positioning themselves in proximity to the gut immune system. These commensals usually exert important immune functions, but may need to be tightly controlled by highly specific IgG and IgA responses induced by T_FH cell-induced germinal centre B cells^44,122.

Protective functions of mucosal IgG in adults.

In both healthy humans and mice, IgG specific to commensals can be detected in the intestinal mucosa and in peripheral blood⁹¹, which raises the possibility that some intestinal IgG-secreting plasma cells home to the bone marrow. Of note, serum IgG shows non-overlapping specificity to the gut microbiota compared with SIgA⁹¹, which points to a specific protective role of IgG in the control of some, but not all, commensals. Accordingly, individuals with significant serum IgG to toxin A from Clostridium difficile are protected from recurrent C. difficile-associated disease, whereas individuals lacking this IgG response suffer from recurrent C. difficile-associated disease¹²³.

The highly protective value of IgG in the gut mucosa can be seen in mice infected with Citrobacter rodentium. Eradication of this attaching-and-effacing pathogen from the intestinal mucosa requires IgG, but not IgA or IgM, responses to critical bacterial virulence factors¹²⁴. The resulting IgG selectively opsonizes virulent bacteria, which are subsequently killed by neutrophils transmigrating from the lamina propria to the intestinal lumen^124–126. Recent mouse work has also revealed that the protective function of mucosal IgG is not confined to the gut. Indeed, selective intestinal symbionts have been shown to physiologically disseminate systemically to elicit specific IgG responses to antigens from Gram-negative bacteria⁹⁰. These IgG responses occur through a canonical TD pathway and confer protection against systemic infections by pathogens such as E. coli and Salmonella typhi⁹⁰. Thus, the microbiota is not rigidly compartmentalized in the gut lumen, but rather communicates with extra-intestinal lymphoid organs to enhance IgG-mediated protection against systemic infections.

Although IgG enhances the control of commensals and pathogens with pronounced mucosal invasiveness, dysregulated IgG responses also contribute to mucosal inflammation. Accordingly, patients with inflammatory bowel disease show increased concentrations of serum IgG against commensal bacteria¹²⁰. This pathogenic IgG response was recently shown to derive at least partially from gut plasma cells and to drive intestinal inflammation by eliciting T_H17 cell population expansion and neutrophil recruitment through a mechanism that involves pathogenic stimulation of gut macrophages via activating FcγRs¹²⁷. Altogether, the evidence currently available indicates that IgG responses to gut commensal bacteria originate from plasma cells predominantly induced by a TD pathway developing in both gut-associated and systemic lymphoid tissues.

Protective functions of mucosal IgG in neonates.

Mucosal IgG is also critical in neonates as it mediates immune adaptation to commensal colonization and immune defence against pathogens^94,126. IgG achieves these functions via multiple mechanisms. Early in life, trans-placental IgG (Fig. 3a) offers protection against common infections, which highlights the efficacy of maternal vaccination¹¹⁸. However, IgG can be also acquired from maternal milk via the intestinal mucosa (Fig. 3b). Unlike milk-derived SIgA, which can enhance mucosal protection by exerting immune exclusion without being microorganism-specific, milk-derived IgG needs to be pathogen-specific to protect against mucosal pathogens^93,126,128. Indeed, IgG from immunized mothers protects neonatal mice against C. rodentium or Heligmosomoides polygyrus^126,128. This IgG specifically recognizes virulence factors from the invading mucosal pathogen, including the adhesin intimin and the T3SS filament EspA encoded by the pathogenicity island termed the locus of enterocyte effacement from C. rodentium¹²⁶. In addition, pathogen-specific IgG needs to reach the neonatal gut to exert its protective function and can be acquired from either the maternal milk or the general circulation via the FcRn¹¹⁸.

Fig. 3 |. IgG in the neonatal gut mucosa and IgD in the aerodigestive mucosa.

a | Transfer of circulating maternal IgG across the placenta occurs via the neonatal Fc receptor (FcRn) on syncytiotrophoblasts and protects the fetus against microorganisms in addition to promoting gut immune maturation. b | Transfer of circulating maternal IgG into breast milk also occurs via the FcRn. This IgG derives from commensal-reactive plasma cells induced in the gut-associated lymphoid tissue (GALT). In addition to IgG, GALT-derived plasma cells release secretory IgA (SIgA) and secretory IgM (SIgM) into milk via polymeric immunoglobulin receptor (pIgR). Milk-derived IgG, SIgA and SIgM promote gut homeostasis and immunity in suckling neonates. c | Mucosal IgD mostly emerges from a follicular T cell-dependent pathway in the human aerodigestive tract. In this pathway, a fraction of IgM⁺IgD⁺ B cells undergo IgM-to-IgD class switch recombination (CSR) and very extensive IgD somatic hypermutation. The resulting IgM⁻IgD⁺ B cells are highly polyreactive and autoreactive, and further differentiate into IgD-secreting plasmablasts and plasma cells. While some of these cells release IgD locally, others colonize distal respiratory or glandular districts via the general circulation. Very few IgD-secreting cells home to or emerge from systemic organs, at least in humans. In addition to undergoing transcytosis by an unknown mechanism, secreted IgD binds to myeloid effector cells such as basophils and mast cells through a receptor comprised of galectin 9 and CD44. Soluble IgD recognizes aerodigestive commensals, pathogens and food proteins. In addition, engagement of basophil-bound and mast cell-bound IgD by antigen amplifies humoral IgG and IgE responses while constraining IgE-mediated basophil and mast cell degranulation. These responses may enhance mucosal homeostasis and immunity in cooperation with SIgA and IgG. Also, an extrafollicular IgD-inducing T cell-independent pathway may exist. Indeed, mice generate short-lived IgD responses very early after immunization. BAFF, B cell activating factor; T_H2 cell, T helper 2 cell; TNF, tumour necrosis factor.

Besides protecting against enteric pathogens, milk-derived IgG may promote intestinal immune maturation⁹⁴, tolerance to commensal microorganisms⁹² and even enhanced protection against allergies in neonates^129,130. Central to these functions is the FcRn, which bidirectionally transports IgG across epithelial cells from intestinal, pulmonary, genital and mammary mucosae¹¹⁸. Furthermore, the FcRn on syncytiotrophoblasts translocates circulating maternal IgG to the fetus across the placenta, thereby providing passive immune protection during perinatal life¹¹⁸. In both neonates and adults, IgG can undergo FcRn-mediated basolateral-to-apical epithelial trafficking to clear microbial virulence factors and control luminal pathogens^131–133. Of note, IgG can be retro-transferred from the lumen to the gut by the FcRn¹¹⁸. This pathway can implement the sampling of intraluminal IgG-bound antigens to subepithelial antigen-presenting cells^131,134. While promoting protection against enteric pathogens^131,134, dysregulated IgG transportation by the FcRn may precipitate gut inflammation¹³⁵. Indeed, the FcRn can initiate immunity and even deliver protective drugs or vaccines¹¹⁸, but can also be hijacked by IgG-binding pathogens¹³⁶.

Mucosal IgD

Although mostly known as a B cell antigen receptor, IgD also exists as a secreted antibody and, indeed, was first discovered in human sera^97,137. This IgD is released by IgM⁻IgD⁺ plasma cells largely derived from mucosal B cells that have undergone IgM-to-IgD CSR^97,138–142. In humans, this process occurs in lymphoepithelial organs from the aerodigestive mucosa (Fig. 3c), including palatine and pharyngeal tonsils^137,143. In mice, IgM-to-IgD CSR has been detected in nasal-associated lymphoid tissue as well as in submandibular and mesenteric lymph nodes^140,141,144. Human IgM⁻IgD⁺ plasma cells not only secrete IgD into the aerodigestive mucosa but also enter the circulation to seed the middle ear as well as the lachrymal, salivary and mammary glands^112,145,146. Remarkably, IgM-to-IgDCSR is rare or absent in systemic lymphoid tissues, but is readily detectable in mucosal districts exposed to airborne and oral antigens in addition to commensals^{139–142,144}.

In mice, signals from the microbiota drive IgM-to-IgD CSR in both nasal-associated lymphoid tissue and mesenteric lymph nodes¹⁴¹. The resulting secreted IgD from class-switched plasmablasts and plasma cells binds to both commensals and pathogens along with their products through both Fab-dependent and Fc-dependent mechanisms^97,139,141. The mechanism whereby IgD reaches the lumen of mucosal and glandular organs across epithelial cells remains unknown, but it might involve CD71, a transferrin receptor that also binds monomeric or polymeric IgA1, and tandem-repeat galectins such as galectin 4 or galectin 9, which bind monomeric IgD^{142,147–149}. The protective role of IgD is suggested by its increase in sera from patients with respiratory infections or chronic lung inflammation^97,137. Accordingly, nasal IgD-secreting plasma cells compensatorily increase in a subset of IgA-deficient patients, who suffer from increased airway infections²⁵.

Of note, secreted IgD features unique properties that other antibodies lack. First, neither human nor mouse IgD binds to an Fcδ receptor (FcδR), possibly because IgD emerged at a time of evolution when Fc receptors had not yet appeared^96,97,119. Second, human IgD is highly mutated (Box 3) and its protruding heavy chain complementarity-determining region 3 (CDR3) may target recessed epitopes essential for pathogen entry into cells^146,150,151. Despite lacking an FcδR¹¹⁹, secreted IgD binds to basophils and mast cells via a receptor complex that includes galectin 9 and the galectin 9-binding protein CD44 (reF.¹⁴²). Engagement of cell-bound IgD by antigen promotes tissue infiltration by basophils and augments basophil expression and release of antimicrobial peptides, chemokines and cytokines, including tumour necrosis factor (TNF) and BAFF as well as IL-4, IL-5 and IL-13 (Fig. 3c). These effects activate antimicrobial, pro-inflammatory and T helper 2 (T_H2) cell-mediated responses (Fig. 3c). In addition, IgD suppresses antigen-induced IgE-mediated basophil and mast cell degranulation (Fig. 3c). Accordingly, increased serum concentrations of allergen-specific IgD correlate with increased desensitization against IgE-mediated allergic reactions¹⁴².

Box 3 |. Structure and function of mucosal IgD in humans.

unlike Igm, but similarly to IgA, secreted IgD does not activate complement, raising the possibility that IgD evolutionarily diverged from Igm to implement non-inflammatory immunity⁹⁷. Instead of complement, mucosal IgD may recruit lectins, including galectin 9, a galactose-binding tandem-repeat lectin that deploys both activating and inhibitory immune functions, including T cell-modulating, mast cell-modulating and basophil-modulating functions¹⁴². Secreted IgD may deliver these effects by interacting with the proteoglycan CD44 and other signal-transducing proteins yet to be characterized¹⁴². Similarly to Ige, secreted IgD has a short half-life of 2–3 days generally ascribed to its long hinge region, which is uniquely encoded by two dedicated exons⁹⁷. As a result, secreted IgD is more susceptible to cleavage by bacterial proteases, but also has more binding flexibility, which may facilitate IgD interaction with low-density epitopes and polyvalent antigens, such as immune complexes^97,162. of note, the immune-augmenting properties of IgD are linked to O-glycans that resemble O-glycans from IgA1, which bind a plant lectin termed jacalin^142,158,163. These glycans may also bind mammalian lectins, including galectins¹⁴². Besides being heavily biased towards λ light chain usage, secreted IgD is highly enriched in polyreactivity and autoreactivity, probably due to its need to recognize aerodigestive antigens comprising both foreign and autologous epitopes, including highly conserved epitopes shared by microbial and host cells^146,150. The pronounced poly/autoreactivity of IgD from about 50% of tonsillar IgD class-switched B cells stems from the molecular configuration of its antigen-binding variable heavy chain (v_H) region, which is extremely mutated and often encoded by v_H3–30, v_H4–34 and J_H6 genes^146,150. These immunoglobulin genes generate a complementarity-determining region 3 (CDR3) domain with finger-like properties resembling those of CDR3 from broadly neutralizing IgG antibodies of certain HIv-1 patients¹⁵¹. Whether secreted IgD also broadly neutralizes certain aerodigestive antigens remains unknown.

IgD has existed for ~500 million years and, contrary to initial prediction, is extremely evolutionarily conserved across species⁹⁶. Indeed, IgD secretion by B cell-derived plasmablasts or plasma cells can be detected in all jawed vertebrates from fish to rodents and humans^96,142, suggesting an evolutionary advantage of this antibody in at least some humoral responses. Consistent with this possibility, in both mammals and teleosts, IgM-to-IgD CSR is microbiota-dependent and generates plasma cells releasing IgD specific to some commensals^141,152. These findings indicate that IgD might contribute to mucosal homeostasis. Considering the striking evolutionary perpetuation of IgD, the link of IgD secretion with inductive T_H2 cell-mediated signals and the T_H2 cell-amplifying and IgE-antagonistic functions of secreted IgD^{96,139,142,144,153}, IgD responses may occur in the context of a protective brand of T_H2 cell-mediated immunity possibly related to the maintenance of immune harmony in the antigen/allergen-rich milieu of the aerodigestive mucosa (Fig. 3c).

Given its ability to bind mucosal mast cells, basophils and possibly phagocytes^97,139, secreted IgD could facilitate the clearance of common environmental antigens, including allergens¹⁴². Of note, IgE may exert a similar function¹⁵⁴, which raises the possibility that IgD cooperates with IgE. This cooperation would include IgD-mediated signals that attenuate IgE-induced basophils and mast cell degranulation^139,142. Such inhibitory signals could result from competition of cell-bound IgD with cell-bound IgE for antigen binding as well as interference of the IgD receptor with signals from the IgE receptor, FcεRI^142,155,156. Whether dysregulated IgD responses contribute to IgE-mediated allergic disorders remains to be established.

Conclusion

Recent advances have added new details to the mechanisms underlying mucosal IgA responses and have further elucidated how these responses control commensal and pathogenic bacteria. Additional studies have started unveiling the impact of IgM and IgD on mucosal homeostasis and have revealed how IgG cooperates with IgA to optimize homeostasis, tolerance and immunity (TABLE 1). Future work will need to elucidate shared and unique molecular targets of IgA, IgM, IgG and IgD across commensal and pathogenic microorganisms as well as food antigens. This knowledge could open new avenues in the development of antibody-based therapies against infectious, inflammatory and allergic disorders.

Table 1 |.

Properties of human mucosal IgA, IgM, IgG and IgD

Feature	Antibody class
	IgA	IgM	IgG	IgD
Origin in the IgH locusa	First (IgA1) or second (IgA2) duplication unit	5′ of first duplication unit	First (IgG3, IgG1) or second (IgG2, IgG4) duplication unit	5′ of first duplication unit
Molecular formb	Oligomeric	Oligomeric	Monomeric	Monomeric
Subclassesc	IgA1, IgA2	None	IgG1, IgG2, IgG3, IgG4	None
Length of hinge regiond	Long (IgA1), short (IgA2)	Short	Intermediate (IgG1, IgG2, IgG4), long (IgG3)	Long
Glycosylation statuse	O-glycans and N-glycans (IgA1), N-glycans (IgA2)	N-glycans	N-glycans (IgG1, IgG2, IgG4), O-glycans and N-glycans (IgG3)	O-glycans and N-glycans
Polyreactivityf	Yes	Yes	Yes	Yes
Main inductive sites	PPs, MLNs, ILFs, MeLNs, CLNs	PPs, ILFs, MeLNs, CLNs, spleen	Spleen, PLNsg/PPs/tonsils	Tonsils
Main effector sitesh	Gut, lung	Gut, lung	Gut, lung, urogenital tract	Aerodigestive tracti
Presence in breast milk	Abundant	Some	Abundant	Some
Basolateral transporter	pIgR	pIgR	FcRn	Unknown
Apical transporter	CD71, dectin 1	Unknown	FcRn	Unknown
Immune receptors	FcαRI, Fcα/μR	FcμR, Fcα/μR	FcγRI, FcγRII, FcγRIII (IgG1, IgG3)	Galectin 9, CD44j
Complement activationk	No	Yes	Yes (IgG1, IgG3), no (IgG2, IgG4)	No
Inductive pathways	TD, TIl	TD, TI	Mostly TD	Mostly TD
CSR involvementm	Yes	No	Yes	Yes
Somatic hypermutation involvement	Yes	Yes	Yes	Yes
Inducing cytokines	TGFβ, IL-10, IL-21	IL-2, IL-10, IL-21	IL-4, IL-10, IL-21	IL-2, IL-4, IL-15, IL-21
Inducing TNF ligands	CD40L, APRIL, BAFF	CD40L, APRIL, BAFF	CD40L	CD40L, BAFF, APRIL
Immune memory	Long-lasting	Possibly long-lasting	Long-lasting	Unknown
Reactivity to microbiota	Yes	Yes	Some	Somen
Reactivity to pathogens	Yes	Yes	Yes	Possible
General functions	Gut homeostasis, control of commensals via immune exclusion and inclusion, protection against mucosal pathogens and their products	Gut homeostasis, control of commensals via immune exclusion, protection against mucosal pathogens	Control of some commensals, protection against mucosal pathogens and their products	Control of aerodigestive antigens, including some bacteria and potential allergens

APRIL, a proliferation-inducing ligand; BAFF, B cell activating factor; CD40L, CD40 ligand; CLN, cervical lymph node; CSR, class switch recombination; FcαRI, Fcα receptor I; Fcα/μR, Fcα/μ receptor; FcγR, Fcγ receptor; FcμR, Fcμ receptor; FcRn, neonatal Fc receptor; IgH, immunoglobulin heavy chain; ILF, isolated lymphoid follicle; MeLN, mediastinal lymph node; MLN, mesenteric lymph node; pIgR, polymeric immunoglobulin receptor; PLN, peripheral lymph node; PPs, Peyer’s patches; TD, T cell-dependent; TGFβ, transforming growth factor-β; TI, T cell-independent; TNF, tumour necrosis factor.

^aIn class-switched B cells, isotypes encoded by constant heavy chain (C_H) genes from the proximal first duplication unit (IgG3, IgG1, IgA1) tend to sequentially class switch to isotypes encoded by C_H genes from the second duplication unit (IgG2, IgG4, IgE, IgA2)¹⁵⁹.

^bOligomeric mucosal antibodies (IgA, IgM) transcytose across the epithelial barrier through the pIgR, whereas monomeric antibodies (IgG, IgD) transcytose through distinct molecular mechanisms.

^cIndividual subclasses bind to distinct cellular receptors. IgG2 and IgG4, which is equivalent to mouse IgG1, show little binding to FcγRs, which may contribute to their non-inflammatory properties. Instead, IgG1 and IgG3 can be highly pro-i nflammatory. However, this pro-inflammatory activity is balanced by Fc γRIIB, which delivers negative feedback signals to IgG-activated immune cells.

^dA longer hinge region is thought to make mucosal antibodies more susceptible to proteolytic degradation by commensals.

^eO-glycans, such as galactose residues, decorate antibodies with long hinge regions, which may permit the recruitment of ‘adaptor proteins’, such as galectins¹³⁹. These galactose-binding lectins may implement/optimize antibody binding to immune effector cells or even mucosal epithelial cells.

^fThe polyreactivity of mucosal antibodies is usually associated with autoreactivity and extensive hypermutation, possibly due to iterative re-entry of pre-existing memory B cells into mucosal germinal centres, which are chronically active. Polyreactivity and autoreactivity may be needed to recognize functionally essential and thus highly conserved microbial antigens, which entail epitopes shared by host cells. Somatic hypermutation may attenuate or even redeem the initially elevated autoreactivity of these antibodies and their precursors, thereby impeding pathological recognition of self-a ntigens.

^gPLNs include draining lymph nodes from respiratory and urogenital tracts.

^hThe abundance of each antibody in mucosal secretions is only partially known, partly because it is technically difficult to establish it in an accurate manner and partly because it has never been systematically explored. In the gut, IgA is the most abundant antibody and its daily release amounts to 2–3 g. In the serum, antibody concentrations have the following pattern: IgG > IgA > IgM > IgD, with IgA1 > IgA2 and IgG1 > IgG2 > IgG3 > IgG4. The half-l ife of these antibodies is 21–28 days (IgG), 5–6 days (IgM) and 2–3 days (IgD), but may differ in mucosal secretions.

ⁱThe aerodigestive mucosa includes the naso-oropharyngeal mucosa and the middle ear mucosa. Additional IgD effector sites entail the lachrymal glands, salivary glands, middle ear and, possibly, lungs.

^jThe IgD receptor remains poorly understood. Galectin 9 and CD44 belong to a multi-p rotein IgD receptor complex that may also encompass CD71.

^kThe scarce complement-activating function of IgA1, IgA2, IgG2, IgG4 and IgD may contribute to their non-inflammatory properties at mucosal sites, at least under physiological conditions. In the lumen of mucosal organs, these non-i nflammatory antibodies may be counterbalanced by pro-inflammatory IgM, which mediates complement-mediated bacterial killing, possibly to dampen the growth of excessively penetrant and aggressive commensals.

^lThe relative contribution of these pathways to mucosal antibody production likely changes with age. The TD pathway is dominant in the adult age and mostly involves the recruitment of pre-existing mucosal memory B cells into germinal centres^29,36.

^mHuman IgD may be released by two subsets of plasmablasts/plasma cells, one emerging upon IgM-t o-IgD CSR and the other characterized by alternative splicing of a long precursor V(D)J–Cμ–Cδ transcript¹³⁷.

ⁿMurine and fish IgD recognize a fraction of mucosal commensals^139,152, and human IgD might as well.

Glossary

Immune exclusion

Active impairment of microbial penetration through the gut epithelium that includes anchoring of intraluminal bacteria to mucus as well as attenuation of bacterial motility, growth and adhesion to gut epithelium by secretory igA.

Immune inclusion

Pro-microbial activity mediated by secretory igA and mucus that promotes the growth of beneficial bacteria while preventing potentially harmful microorganisms from colonizing the same mucosal niche.

Isolated lymphoid follicles

single intestinal lymphoid aggregates with or without germinal centres, disseminated along the small intestine in mice and both the small and large intestines in humans, which develop in association with the follicle-associated epithelium and function as dynamic immune reservoirs for T cell-independent or T cell-dependent induction of igA.

M cells

specialized antigen-capturing epithelial-like cells from the follicle-associated epithelium that initiate intestinal immune responses by transporting intraluminal antigens into Peyer’s patches and isolated lymphoid follicles with the help of dendritic cells, macrophages and B cells.

Somatic hypermutation

A germinal centre-associated process that promotes antibody affinity maturation by introducing V(D)J gene point mutations through a molecular machinery that includes the DNA-editing enzyme activation-induced cytidine deaminase.