Immunological Outcomes of Antibody Binding to Glycans Shared between Microorganisms and Mammals

Glycans, polymers of glycosidically linked sugars, are one of the most basic cellular components, and exist as carbohydrate-only entities as well as covalently attached modifications of proteins (glycoproteins) or lipids (glycolipids). In this article, we use glycan to indicate both oligosaccharides and polysaccharides. In mammals, glycans have diverse functions, such as marking apoptotic cells for clearance, immune self-/non-self-discrimination, cell–cell communication, and intracellular signaling (1, 2). Glycosylation defects in humans are linked to disease (3), and the expressed glycome can be altered during inflammation, cellular stress, as well as cancer (4). Although the combinatorial composition of a saccharide array can generate an immense number of structures, the composition of mammalian glycans is well-conserved (5). Some microbial pathogens including bacteria, fungi, and protozoans display mammalian-associated glycans on their surfaces as an evolutionary adaptation to evade detection by the host’s immune system (6). This property can also directly contribute to pathogenicity of these organisms (7, 8). In addition, expression of host-similar glycans by allergens may promote their engagement of innate receptors expressed by APCs and epithelial cells in the lung (9, 10).

In mammals, B cells and Abs that react with self-glycans exist naturally and function to promote homeostasis (11) by facilitating the clearance of dangerous and potentially inflammatory components, such as apoptotic cells (12), senescent RBCs (13), and metabolic products such as oxidized lipids (14). Aside from these homeostatic functions, naturally occurring Abs specific for mammalian glycoproteins or glycolipids recognize these structures when displayed by microorganisms, as well as allergens, and can facilitate their clearance (10, 12, 15, 16). Many cellular processes such as engagement of Siglec-G/CD22 (17), sequestration of autoreactive Ags (12), and induction of cellular anergy (18) exist to regulate and maintain autoreactive B cells within the B cell repertoire. In this review, we discuss the expression and interactions of B cells with selected glycan epitopes that are expressed on host cells, microbes, and allergens. These epitopes include N-acetylglucosamine (GlcNAc), sialyl-lacto-N-tetraose, and α-1,3-glucan. We then give some examples of how Abs to these glycans mediate housekeeping functions and provide protection against pathogens and allergens (Fig. 1, Table I).

Much has been written regarding Abs with extensive polyreactivity (19–21). The term polyreactive has generally been used to describe Ab reactivity with seemingly structurally unrelated Ag targets and has been attributed to both germline and somatically mutated gene-encoded contributions to their Ag-binding sites (22, 23). The polyreactive nature of some Abs has often been characterized by Ab binding to recombinant Ags or mimotopes in solid-phase ELISA-type assays or by Western blotting of denatured complexes from microbes, mammalian tissues, or cell extracts. These assays may also detect low-avidity binding of Abs against neoepitopes generated through processing of the Ag-containing material that may not normally be exposed on the native nondenatured molecules. By contrast, the representative Abs discussed in this review exhibit exquisite specificity for defined oligosaccharide moieties expressed as glycan epitopes. These structures are themselves similar among bacterial, fungal, and parasitic components, as well as allergens and multiple mammalian cell types. We and others have demonstrated that anti-GlcNAc Ab binding to bacterial cell wall peptidoglycan complexes is inhibited by soluble monomeric GlcNAc monosaccharide, but not its enantiomer N-Acetylgalactosamine (24, 25). Thus, these Abs are polyreactive only in the sense that they bind an identical or a very similar associated glycan epitope present in various molecular entities expressed by a wide range of living organisms.

Host-resembling epitope expression by microorganisms has been referred to as molecular mimicry, and infections caused by some of these organisms, especially viruses and bacteria, have been implicated as the precipitating event for development of autoimmunity (26–29). Most of the studies in this field imply that the immune response to these infectious organisms drives the development of autoreactive T or B cells. However, evidence for most of these claims is sparse, and some of the suggested immune mechanisms by which this occurs are debatable. The association between Streptococcus pyogenes infection and rheumatic heart disease (RHD) is commonly discussed, and although 30–40% of humans who recover from rheumatic fever experience development of RHD (30), the cellular mechanism driving this disease remains controversial. Some prominent studies have implicated anti-GlcNAc Ab activity in mice immunized with myosin or GlcNAc-BSA conjugates as being a causative agent of RHD (31, 32). However, the specificity of Ab binding to GlcNAc-protein conjugates was not demonstrated by inhibition with free monosaccharide. In addition, these T-dependent hybridoma Abs reactive with myosin and GlcNAc-BSA did not express the canonical VHJ606 gene segment characterizing the highly conserved Ab repertoire of GlcNAc-specific mAbs raised by immunizing mice with S. pyogenes (33). Further, other studies have clearly demonstrated that serum Abs from rabbits immunized with S. pyogenes cell wall that bound to heart muscle were only inhibitable with the insoluble, but not soluble (GlcNAc-containing), portion of the bacterial cell wall (34). Instead, many studies have shown that streptococcal M protein and cardiac myosin express similar epitopes and peptides (35–37). Therefore, T and B cells with specificity for streptococcal M protein that also react with myosin are most likely the initiating factor in S. pyogenes–induced RHD. Although pathogenic Abs against M protein are involved in RHD pathology, they are not likely to be directed toward defined GlcNAc epitopes as some studies suggest (31). Instead, these Abs may be cross-reactive, with neoepitopes possibly formed artificially by conjugating GlcNAc to BSA (31). In contrast, the GlcNAc residues on group A streptococcal carbohydrate are structurally ordered (38), and mAbs against GlcNAc peptide conjugates do not bind native S. pyogenes bacteria or the purified group A streptococcal polysaccharide (J.S. New and J.F. Kearney, unpublished observations). Therefore, studies demonstrating that Abs generated against bacterial glycan epitopes cause autoimmunity should be interpreted with caution.

Bacterial infections may drive the development of other autoimmune conditions such as Guillain–Barré syndrome (GBS) (39). GBS is characterized by neuromuscular paralysis caused by T and B cells with specificity for peripheral nerve myelin gangliosides (40). Gangliosides are sialylated glycolipids that are widely distributed in mammalian neuronal membranes, where they are sequestered in islands in the lipid bilayer (41, 42). In some rare cases, GBS can be preceded by an infection with ganglioside-bearing organisms (39) such as Streptococcus agalactiae (group B Streptococci) (43, 44), Mycoplasma pneumoniae (45), or Campylobacter jejuni (46–48), the most common organism implicated in this disease. Many C. jejuni strains have cell wall LPSs that contain sialylated gangliosides such as GM1, GT1a, GD1b, or GD3 that are identical to human gangliosides (47). It has been proposed that infection with C. jejuni can induce ganglioside-reactive Abs that cause host nervous system damage (46–49). C. jejuni–induced autoreactive Abs are specific for the oligosaccharide sequences attached to the polar groups in membranes, but sequence similarities and their membrane distribution may yield cross-reactivities with different gangliosides such as GM1, GT1a, GD1b, or GD3 (46, 48). Because of the sequestration of gangliosides in membrane islands (41), the reactivity of these autoantibodies with gangliosides would most likely depend on their density, distribution, and clustering patterns in host membranes. Although the similarity of these ganglioside-like epitopes on the C. jejuni LPS may explain some GBS-inducing pathology, development of GBS among individuals infected with C. jejuni (50) and other ganglioside-expressing organisms, such as CMV (51), varicella zoster virus, and EBV is low (52), making this cause and effect controversial (39). Therefore, in addition to Abs to gangliosides, there may be other unknown immune factors that contribute to incidence of disease.

In addition to C. jejuni, it has long been known that the type-specific capsular polysaccharides of group B streptococci (which will be discussed in detail later in this review) share structural similarities with epitopes on human glycoproteins (43) and the oligosaccharide head groups of mammalian membrane sphingolipids (43). The repeating units of the sialylated group B streptococci type III polysaccharide are identical to gangliosides GD3 and GT3 (43). Similarly, the group B streptococci type Ιb repeating unit (53) is identical to the oligosaccharide sialyl-lactoneotetraosyl-ceramide (44) and has the potential to share epitopes with other gangliosides, including GD1a, expressed on mouse TH2 cells and GD1α expressed preferentially on mouse TH1 cells (54). Anti–group B streptococcal Abs raised by immunization of mice with the group B streptococci type III or type Ib does not appear to cause neuromuscular paralysis or affect T cell functions (55). Another striking characteristic of these Abs is that their binding to target acidic polysaccharides is calcium dependent (44), such that these Abs may not bind to oligosaccharides expressed as gangliosides on host cells. The sialyllactose oligosaccharide epitope of the ganglioside GM3 is found on the plasma membrane of many mammalian cell types (56), is highly expressed in melanoma, and has been extensively studied as a potential antitumor target (57, 58). Conventional wisdom suggests that Abs against GM3 and other gangliosides may target and damage many GM3-bearing cells. However, binding of certain induced anti-GM3 mAbs occurs in an all-or-none fashion depending on the threshold density of GM3 exposed at the cell surface (59). These observations suggest that bacteria could possibly induce Abs that would bind to the target bacterial epitope without affecting functions of normal cells. Therefore, Ab recognition of cell-surface GM3 may depend on the density and spatial distribution of oligosaccharides on complex membrane-bound structures, and thus be a general characteristic regulating the relative binding of Abs to similar epitopes on bacteria and host cells. Although bacteria display epitopes that can also be expressed on mammalian cells, the implications that Abs to self-glycan epitopes cause autoimmunity should be interpreted carefully.

In this review, we discuss the immunological activities of Abs with specificity for GlcNAc, sialyl-lacto-N-tetraose, and α-1,3-glucan. Such Abs are detectable in the serum of mice, as well as humans, and are referred to as natural Abs (24). This terminology is often used to imply that these Abs are detectable in the absence of deliberate immunization or vaccination. Glycan epitopes inducing natural Abs may include those expressed by bacterial organisms comprising the microbiome, whereas others represent neoepitopes that are normally sequestered but can be exposed by altered glycosylation related to cell apoptosis or death. The B cell repertoire encoding natural Abs is generated early in life, and its individual clonal components can fluctuate throughout the lifetime of the host depending on exposure to self, microbiota, or environmental Ags expressing these epitopes (24). Mechanisms that both limit and maintain these autoreactive B cells within the B cell repertoire are discussed later in this review.

Diverse groups of bacteria, protozoans, and fungi use GlcNAc as a building block in glycan-based structures (60). Helminths such as Heligmosomoides polygyrus and Strongyloides stercoralis express GlcNAc-containing cellular and structural components (24, 61), and helminth exposure can drive the production of high-titer serum Abs against GlcNAc (24). In addition, the biochemical composition of bacterial cell walls is highly conserved (62). Gram-positive bacteria incorporate alternating units of GlcNAc and N-acetylmuramic acid connected by a β-1,4-glycosidic bond to assemble peptidoglycan for their cell walls (63). In addition to peptidoglycan, Gram-negative bacteria possess an outer membrane containing a unique LPS that promotes immune evasion by shielding peptidoglycan GlcNAc moieties from serum Abs and lectins (64). Certain Gram-positive organisms such as Streptococcus pneumoniae evade the host immune system by generating a serotype-specific cell wall–linked polysaccharide capsule that prevents complement-mediated opsonophagocytosis (65). Regardless of serotype, S. pyogenes (group A Streptococcus [GAS]) expresses a cell wall polysaccharide with a helical rhamnose backbone decorated with repetitive terminal GlcNAc residues (38) that is highly immunogenic in mice and humans. Infection with GAS induces Abs against GlcNAc (25), and these Abs are somewhat protective in mouse models of GAS infection (66–68). The development of a GlcNAc-based vaccine for GAS infection has been proposed for many years. However, the likelihood of pathogenic self-reactivity of such Abs and relatively poor protection in mouse models may have disfavored efforts to develop such a vaccine (38, 69). Because multiple bacterial species express variable levels of GlcNAc-containing molecules, efforts have also been made to design a GlcNAc-based vaccine that could protect against a broad array of organisms, including antibiotic-resistant Staphylococcus aureus (70). Recently, Genentech developed a novel anti-GlcNAc–based therapeutic agent in which human IgG1 anti-GlcNAc Abs were linked to the antibiotic rifalogue. In vivo, these modified Abs bound to S. aureus, and upon phagocytosis, endosomal activation of the antibiotic agent mediated killing of intracellular S. aureus (71). The conserved microbial expression of molecules containing GlcNAc epitopes suggests that these approaches may have wider implications for drug delivery to intracellular compartments for killing other GlcNAc-expressing pathogenic organisms, such as Mycobacterium tuberculosis.

GlcNAc is the basic subunit in the form of β-1,4 GlcNAc–linked residues constituting the homopolymer chitin (72), which is the second most abundant biopolymer on earth. Chitin is present in the cell walls of various fungi, including yeast (73, 74), and is a major component of crustacean and insect exoskeletons (75), as well as protozoan and insect gut walls (76, 77). The physicochemical properties of different chitins may vary, and chitin particles purified from crab shells or different fungal species elicit significantly varied proinflammatory cytokine responses (78). Chitin is also expressed by a wide variety of respiratory allergens such as house dust mite, cockroach, and many fungal species including Aspergillus fumigatus (9, 79). Chitinous particles derived from these organisms may be considered as carriers associated with protein allergens that act as cargo (9, 10). For example, the hyphal cell wall of A. fumigatus contains chitin as a structural element and bears the major allergens Asp f 1 and Asp f 2. Engagement of innate receptors by chitin-bearing particles can encourage the cellular entry to promote the processing of these allergenic proteins. Targeting the chitinous portion of these particles to prevent engagement of these innate receptors would also prevent entry of many different types of inflammatory elements associated with allergens, such as LPS, β-glucan, and proteases aside from allergenic proteins (80). Our laboratory has demonstrated that anti-GlcNAc Abs bind chitin particles and germinated A. fumigatus and significantly decrease their uptake by alveolar macrophages and APCs in the lung. In addition, neonatal exposure of mice to S. pyogenes results in increased levels of serum anti-GlcNAc Ab and an attenuation of A. fumigatus–induced allergic disease. These studies further demonstrated that i.v. delivered anti-GlcNAc Abs suppressed A. fumigatus–induced airway disease development (9). Serum Abs against GlcNAc occur naturally and may vary depending on the microbiome, blood type, as well as the history of infection (12). Accordingly, therapeutic manipulation of anti-GlcNAc Ab levels may protect against bacterial and protozoan infection, as well as respiratory and food allergies associated with chitin-bearing organisms.

Aside from GlcNAc expression in its cell wall, germinated A. fumigatus also expresses sialyl-lacto-N-tetraose and α-1,3-glucan (9), an epitope described later in this review. Sialyl-lacto-N-tetraose is also expressed by group B streptococci type 1b (44) and is a component of human milk oligosaccharides (HMOs) (81). Apart from causing allergic airway disease, A. fumigatus is the leading cause of invasive aspergillosis (I.A.) (82). I.A. results from an opportunistic infection and is common among immunocompromised individuals (82), and Aspergillus is commonly isolated from deep tissue wounds of soldiers wounded in combat (83, 84). Current I.A. treatments have a limited success rate (85), and an efficacious vaccine for preventing I.A. has not yet been developed (86). Although some fungal polysaccharides can induce protective Abs, they do so very poorly. This could be a result of multiple factors, including suppression of T and B cell activity (87), variation in epitopes expressed between hyphal and yeast states, as well as fungus-induced proteasomal degradation of Abs (88). Because the generation of Abs to fungal epitopes is difficult, we hypothesized that generating Abs to a bacterial epitope that is also expressed by fungi would stimulate the production of immunoprotective Abs. Mice vaccinated with a group B streptococci 1b strain containing sialyl-lacto-N-tetraose were protected in a model of disseminated aspergillosis (55). In addition, i.v. infusion of Abs against sialyl-lacto-N-tetraose (clone SMB19) or use of SMB19 B cell–transgenic mice expressing high levels of endogenous Abs against sialyl-lacto-N-tetraose without deliberate bacterial immunization resulted in protection. Interestingly, J558 B cell–transgenic mice with a high frequency of B cells specific for α-1,3-glucan and increased levels of Ab against α-1,3-glucan, which is also a major component of the A. fumigatus hyphal cell wall, were not protected against disseminated aspergillosis (55). Compared with Abs against α-1,3-glucan, the SMB19 Ab is unique because it preferentially binds the tip of the germinating hyphae (24, 55). Calcium pumps and channels regulate the growing tip of the hyphae, and dysregulation of these calcium gradients may inhibit hyphal growth (89). The hyphal tip of many other yeast and fungal species such as Candida albicans also contains sialyl-lacto-N-tetraose (Fig. 2). Therefore, targeting the hyphal tip may be an effective therapeutic for preventing many different types of invasive fungal diseases.

Group B streptococci 1b polysaccharide conjugate vaccine provides protection against group B streptococci 1b infection (90). Group B streptococci can cause severe infection among newborns, pregnant women, and the elderly (91). Currently, women who test positive for group B streptococci during pregnancy receive antibiotics during labor to prevent maternal and neonatal group B streptococcal infections. Glycoconjugated group B streptococcal vaccines have been developed to generate long-lasting immunity and to circumvent antibiotic use (90). Group B streptococcal vaccine generation was predated by the observation that maternal Abs against the group B Streptococcus capsular polysaccharide correlated with lower neonatal susceptibility to group B streptococcal disease (92). In clinical trials, pregnant women immunized with the glycoconjugate group B Streptococcus vaccine produced type-specific Abs that were also detectable in the newborn cord blood and in the infants for up to 2 mo of age. In these studies, group B Streptococcus carriage rates were also lower among adult females receiving the vaccine (90). Collectively, the effectiveness of the group B streptococcal polysaccharide conjugate vaccine for inducing human Abs that both protect against group B streptococci infections and also react with multiple fungi have motivated our efforts to repurpose group B Streptococcus–conjugate vaccines for protection against I.A. development in immunocompromised individuals.

In addition to being expressed on bacteria and fungi, lacto-N-tetraose and sialyl-lacto-N-tetraose are highly abundant oligosaccharides found in breast milk (81, 93–96). These HMOs can be advantageous by promoting growth of many beneficial Bifidobacterium species (97, 98) or as decoys for pathogenic diarrhea-causing organisms that use glycan-mediated attachment mechanisms for accessing the host immune system (95). However, HMOs may also promote pathogenicity as suggested by their ability to increase Staphylococcus epidermidis and S. aureus growth in vitro (99). Both of these staphylococcal organisms cause mastitis (100, 101), contaminate breast milk, and cause infections among infants (102–104). The exact mechanism by which the HMOs modulate the host immune system, apart from benefiting the gut microbiome (105), has not yet been determined.

The presence of lacto-N-tetraose in breast milk and the infant digestive system is of interest because group B Streptococcus 1b–vaccinated women and their children had Abs against sialyl-lacto-N-tetraose (90). Although these Abs can potentially react with an HMO epitope that is highly expressed in the mammary glands of pregnant females and digestive tracts of neonates, the vaccine did not have related adverse effects and was shown to be safe for use in women of childbearing age, including those who were pregnant (106, 107). This further supports observations that Abs against shared epitopes can reside in the body naturally without causing autoimmune manifestation.

α-1,3-Glucan is a linear α-1,3-linked homopolymer of glucose, and Abs against α-1,3-glucan are naturally occurring (24). In mice, B cells specific for α-1,3-glucan are enriched within the marginal zone and B1b cell populations (108). Anti–α-1,3-glucan Abs are detectable in human plasma (109); however, unlike the other Abs discussed in this review, our understanding of the levels and characteristics of human Abs against α-1,3-glucan is poor. Fungi, including yeast, are among the best documented expressers of the α-1,3-glucan epitope (74). Molecules expressing these epitopes can function as virulence factors when expressed by yeast (110) and fungal cell walls (74), or oral plaque-forming bacteria (111). However, commensal enteric organisms from mice can also express this molecule without any known pathogenic consequence (112). In fungi, such as Aspergillus nidulans, α-1,3-glucan accumulates during vegetative cell growth and is used as an endogenous carbon source during sexual development (113). Whereas in yeast, such as Cryptococcus neoformans, α-1,3-glucan is involved in anchoring the capsule to the cell wall (110). In some Histoplasma capsulatum strains, the α-1,3-glucan cell wall polymer may act as a shield for β-glucan, thereby evading immune clearance via recognition by Dectin-1, a mammalian β-glucan receptor (114); however, a receptor for α-1,3-glucan has not yet been identified. Although Abs against α-1,3-glucan were not protective in a mouse model of I.A. (55), it is possible that they could protect against noninvasive fungal and yeast infections similarly to Abs against β-1,3-glucan, which have been shown to inhibit growth of multiple fungi species (115).

Bacterial species expressing α-1,3-glucan include oral streptococci (111, 116), cyst-forming bacteria (117), and some enteric organisms isolated from mice (112). Planktonic Streptococcus mutans do not express α-1,3-glucan; however, when these organisms begin producing biofilms, a series of glucosyltransferases synthesize α-1,3-glucan from sucrose (118). α-1,3-Glucan provides structural stability for these biofilms and is critical for S. mutans attachment, aggregation, as well as accumulation on tooth surfaces (119). It has yet to be determined whether oral Abs against α-1,3-glucan could specifically prevent cariogenic biofilm formation. In addition, biofilms formed by other bacterial organisms or yeasts may also contain α-1,3-glucan. Aside from S. mutans, cyst-forming bacteria such as Azotobacter vinelandii and some commensal enteric organisms such as Enterobacter cloacae and Serratia liquefaciens can also express α-1,3-glucan under non-biofilm-forming conditions (112, 117). We and others have used α-1,3-glucan to study the immune response to T cell–independent Ags (108, 120–124). Textbook descriptions of T-independent responses state that these immune reactions do not result in the formation of memory; however, using the α-1,3-glucan epitope-bearing strain of Enterobacter (MK7), we demonstrated that α-1,3-glucan–specific IgM-secreting cells contribute to polysaccharide-specific memory Ab responses, which are maintained long term (108, 125).

Aside from bacteria and fungi, German cockroach (Blattella germanica) allergen also contains detectable α-1,3-glucan epitopes. These epitopes are expressed in the insect’s muscle and exoskeleton, and binding of anti–α-1,3-glucan is not exclusively due to microorganism contamination. Cockroach fecal pellets contain both allergenic protein Bla g 2 and α-1,3-glucan, and we suggest that particles bearing α-1,3-glucan epitopes can act as carriers of allergenic Bla g 2 proteins (109). Natural Ab-based therapeutics for treating cockroach allergy are attractive because children who are skin-prick test positive for cockroach allergens are more likely to have difficult-to-control asthma (126, 127) and visit the emergency department because of an asthmatic event (128) than children who are not allergic to cockroaches. Treatment of neonatal, but not adult, mice with purified α-1,3-glucan or an α-1,3-glucan–expressing MK7 Enterobacter strain resulted in suppressed development of cockroach allergy during adult life. α-1,3-Glucan–specific IgA-secreting cells were identified in the lungs of mice immunized with MK7 as neonates, but not as adults. By generating mice that are unable to make IgA responses to α-1,3-glucan, we confirmed that these B cells are responsible for protection against cockroach allergy (109). It has been shown that neutralization of bacteria and viral particles by Ag-specific IgA at mucosal sites is critical for preventing some diseases (129, 130). In addition, this mechanism may also be important for suppressing allergic disease, and selection of B cell clones capable of producing IgA during neonatal life may also be crucial in this process. Considering these, and other observations, we suggest that early exposure to α-1,3-glucan, in the form of a probiotic or natural colonization, may be sufficient to protect against mycosis and suppress fungal or cockroach allergy development for the lifetime of the individual.

As we have described in this review, B cells and their Ab products with the potential to react with self-glycans exist naturally and there are multiple mechanisms that maintain self-reactive B cells within the repertoire. These Abs against self-Ags are not limited to those mentioned in this review and also include Abs against major and minor blood group Ags, as well as Abs against the phospholipid epitope phosphorylcholine. One major unanswered question is: what mechanisms govern the maintenance of these potentially autoreactive B cells within the adult B cell repertoire?

The levels of Abs in human serum that react with self-glycans, such as GlcNAc, are much lower compared with those that react with non-self-ligands such as Gal-(α-1,3)-Gal (131) or anti-Neu5Gc (132). A possible explanation for this is that a tolerance mechanism may be involved in maintaining self-reactive B cells within the B cell repertoire. For example, engagement of CD22 or Siglec-G by sialic acid–bearing glycans dampens B cell signaling and can even cause B cell apoptosis (17). However, glycan engagement by the BCR along with additional signals through TLR or NOD receptors, such as after an encounter with a pathogenic organism, could break these tolerance-inducing mechanisms, allowing for autoreactive Ab production (133). Retention of B cells with reactivity for both self-Ags and microbial Ags may be evolutionarily advantageous because it drives the development and maintenance of a complete B cell repertoire (24).

In another suggested mechanism, epitopes similar among microorganisms and host may be sequestered in the host such that they are not normally exposed to BCRs at a level sufficient to initiate B cell activation. For example, GlcNAc subunits of mammalian glycans are usually capped by mannose or highly charged sialic acids, which prevent BCR engagement (12). These Ags, which are normally sequestered intracellularly, may be exposed only as part of the apoptotic program or released during necrotic cell death. Normally, immune disposal mechanisms clear these Ags rapidly and limit their ability to drive immune reactions (134). However, chronic inflammation and cancer induce states of glycan remodeling (4), during which these epitopes could possibly be exposed.

Our last suggested mechanism for the maintenance of autoreactive B cells in the repertoire involves B cell clonal anergy. Many self-reactive B cells are found in innate-like B1 and marginal zone B cell subsets in an activated state (135). Their heightened activation state most likely results from BCR engagement by self-Ag. Studies of the regulatory mechanism controlling B cell responses after exposure to self-Ag revealed that a proportion of B cells, but not T cells, exist in a state of anergy instead of being deleted. In this particular study, B cell unresponsiveness was manifest as a continuum with the highest indication of BCR engagement apparent in marginal zone B cells (18), consistent with our previous findings that the marginal zone B cell population is one of the main reservoirs of B cells that can respond to self- and microbe-associated epitopes (135).

Much of the dogma related to the tolerant and anergic state of B cells was originally generated with BCR transgenic or knock-in mice (136). The BCRs in many, but not all, of these models were derived from B cells initially isolated from mice that were repeatedly immunized with T cell–dependent Ags or from autoimmune mice that made somatically hypermutated and isotype-switched pathogenic Abs. Therefore, forced expression of transgenes among B cells in these models involves the expression of highly mutated high-affinity receptors that may not be reflective of those that would exist in a normal primary repertoire, where B cells would express mostly germline Ig genes. Thus, B cells expressing self-reactive BCRs may normally coexist peacefully in a system with self-reactive Ags because of various mechanisms such as Siglec-G/CD22 engagement, self-Ag sequestration, anergy, or even receptor editing (137). However, these tolerance mechanisms can be breeched upon engagement of microorganism-associated epitopes that coengage TLRs, enabling B cells to respond rapidly to the corresponding environmental organisms. Most likely, frequencies of these autoreactive B cell clones fluctuate throughout the lifetime of an individual depending on acquisition of and alterations in the microbiome, bacterial infection, and tissue injury.

Glycans can exist in different states, such as attached to lipids or proteins, and can be reshaped by posttranslational modification or metabolic or catabolic processes. Glycan epitopes are extensively distributed over various organisms across kingdoms (Fig. 1, Table I). Some of these glycans mediate primary biological functions such as self-/non–self-distinction and apoptotic cell removal. It is clear that some bacteria and allergens can display epitopes that are also expressed on host glycans. The bacterial expression may serve as an effort to evade host immune responses. However, Abs reactive with these shared glycans are common in mammals, where they mediate homeostasis within the host and provide defense against pathogenic organisms and allergens. Production of Abs against these glycans can vary with age, and Ab reactivity against glycans expressed commonly by both the host and these environmental organisms can be involved not only in protection against infectious diseases but also allergic diseases. Despite the association of some of these glycans with host molecules, anti–self-Abs exist naturally, and self-reactive B cells exist in equilibrium within the normal B cell repertoire as the result of highly regulatory and complex dynamics.

The authors have no financial conflicts of interest.