IgA nephropathy (IgAN) is an autoimmune disease in which IgA, exclusively of the IgA1 subclass, with an altered glycan moiety manifested as deficiency of galactose (Gd-IgA1) on its heavy-chain hinge region (HR) acts as an autoantigen which is recognized by ubiquitous, naturally occurring, anti-glycan antibodies to form nephritogenic circulating immune complexes (CIC). Some of these complexes deposit in the glomerular mesangium to induce kidney injury (1–7). Detailed analyses of CIC and the mesangial immune deposits of IgAN patients revealed the presence of under-galactosylated IgA1 in the polymeric (p) form with joining (J) chain (3, 8–22). The nephritogenic potential of CIC containing Gd-IgA1 isolated from IgAN patients was demonstrated by studies of their effect on mesangial cells (23–25) or that of complexes generated in vitro with poorly galactosylated myeloma pIgA1 and corresponding anti-glycan antibodies (26) on the proliferation and activation of human primary mesangial cells in culture. Importantly, complexes with molecular mass of ~700-1,000 kDa displayed stimulatory activity whereas smaller complexes did not (3, 25). Furthermore, intravenous injection of complexes comprised of human Gd-pIgA1 and recombinant IgG antibodies specific for this IgA1 into immune-deficient mice induced kidney pathology features and urinary abnormalities typical of IgAN (27).

IgAN is the leading cause of primary glomerulonephritis in many countries (1). However, its prevalence displays striking geographic, racial, and age-related distributions (1, 28–30). IgAN is common in most European countries, USA, and east Asia but is less frequent in South America, India, Bangladesh, Indonesia, Nepal, Pakistan, and other Asian countries and is rare in central Africa (29–31). Furthermore, there are marked racial differences in disease incidence. African Blacks, African Americans, and indigenous Australian Aborigines living in remote rural areas display a low incidence of IgAN (31–41). Because of these race-associated differences in the incidence of the disease and discovery of multiplex families with multiple affected members, it has been proposed that genetically modulated differences play an important role in the mechanisms of disease (42–48).

Investigation of a possible role of genetics in the development and expression of IgAN initially included linkage studies of multiplex pedigrees. Three loci on separate chromosomes have been identified, although the genes responsible for the linkage have not been defined (42, 47). Later, genome-wide association studies (GWAS), first in the United Kingdom and later with cohorts of patients and controls of European and east Asian ancestry, identified multiple loci associated with IgAN (now at least 30) (45, 48). The associated loci include a wide variety of genes, including some in the major histocompatibility complex involved in antigen processing and presentation. Other associated loci encompass genes involved in chemokine and B and T cell receptor signaling, regulation of the alternative complement pathway, genes encoding anti-microbial peptides α-defensins, and genes affecting NF-κB signaling, T cell–independent IgA class-switching, IgA plasma cell activation, IgA Fc receptor, and the O-glycosylation pathway (47, 49). Interestingly, analysis of 85 world populations performed by Kiryluk et al. (47) showed that a genetic risk score, based on 15 single-nucleotide polymorphisms, increased with progressive eastward and northward distance from Africa. Additional studies found a highly significant association between a genetic risk score and age at diagnosis; a greater genetic burden promoted an earlier onset of disease (46).

An alternative explanation of immunopathogenesis of IgAN based on the epidemiology and impact of Epstein-Barr virus (EBV) infection on the IgA system was recently proposed to elucidate the possible pathways in the geographic and race-associated differences in the prevalence of IgAN (50). The structural characteristics of IgA in CIC and mesangial deposits and phenotypic profiles of circulating IgA-secreting cells from IgAN patients revealed a remarkable concordance with results observed with in vitro EBV-infected B cells (18, 50–56), including the predominant secretion of pIgA1 with poorly galactosylated glycan chains ( Table 1 ). Therefore, we initiated studies of phenotypes of EBV-infected B cells from the peripheral blood of IgAN patients and African American and White controls (50) to address a potential role of EBV in the pathogenesis of IgAN. These cells were analyzed with respect to the maturation profiles and the expression of cell-surface homing markers, including those involved in the characteristic lymphoid tissue distribution, and the ability to produce pIgA1 with poorly galactosylated glycans (50). The EBV-infected IgA⁺ B cells from IgAN patients displayed phenotypic characteristics very similar to those of IgA⁺ B cells infected in vitro with EBV (50–56).

Studies of the composition of CIC and mesangial deposits revealed that they consist of IgA exclusively of the IgA1 subclass, IgG, C3 of the complement cascade, and sometimes IgM (2, 3, 10, 11). These findings prompted the search for the participating exogenous or possibly endogenous antigens involved in CIC formation. Although antigens of the microbial and food origin have been sought as components of CIC or immune complexes in the mesangium, no uniformly prevalent antigen was identified (57–60). The possible autoimmune character of IgAN was postulated because of the exclusive presence of the IgA1 subclass in complexes with IgG,C3, and soluble IgA Fc receptor (FcαRI or CD89) in the absence of other identifiable components (10). Recently it was reported that CD89 represents critical factor for mesangial proliferation in childhood IgAN (61). As discussed below, human and hominoid-primate IgA1 has, in contrast to IgA2, a unique HR of the α1 heavy chains characterized by an additional 13 amino acids that include Thr and Ser residues which may be glycosylated ( Figure 1 ) (62). Comparative evolutionary studies of the Ig HR clearly indicate a recent insertion of a gene segment encoding for the HR of IgA1 into phylogenetically older IgA2 (62). Furthermore, IgA1 from IgAN patients displayed an aberrant glycosylation pattern with the characteristic galactose (Gal) deficiency in the O-linked glycans in the HR ( Figures 1 , 2 ) as revealed by reactivities with relevant lectins or monoclonal antibodies specific for Gd-IgA1, or direct biochemical analyses of glycans of IgA1 eluted from glomerular immune complexes (2, 3, 7, 8, 12, 14, 18, 71, 76). Gd-IgA1 in CIC and in mesangial deposits is in the polymeric form, as demonstrated by the elution profiles of IgA from dissociated CIC, and reactivity of mesangial IgA1 with secretory component (SC) – the extracellular part of the polymeric IgA receptor (pIgR) expressed on epithelial cells – which binds exclusively pIgA and IgM with J chain (2, 21, 77–80).

Gd-pIgA1 is recognized by antibodies specific for terminal N-acetylgalactosamine (GalNAc) residues in the O-linked glycans in the HR of Gd-pIgA1 which may contribute to the conformational alteration of HR and generation of new antigenic determinants (2, 3, 21, 23, 77, 81–83). It is thus evident that the structurally unique HR of human IgA1 in IgAN patients ( Figure 1 ) resulted in the appearance of novel antigenic determinants. These epitopes are recognized by ubiquitous, naturally occurring, antibodies (21, 23, 84, 85) or those that probably evolve due to the affinity maturation (86), ultimately leading to formation of nephritogenic CIC that deposit in the glomerular mesangium (2, 3). The universal presence of IgG in mesangial deposits was demonstrated using anti-IgG nanobodies which selectively recognize potentially hidden antigenic determinants of IgG in mesangial immune complexes of IgAN patients (87). Furthermore, somatic mutations in the variable regions of heavy chains may modulate the affinity of autoantibodies against Gd-IgA1 (86). The presence of naturally occurring antibodies in sera of all healthy individuals, specific for the HR-containing Fab fragment of some IgA1 myeloma proteins, suggested that these IgA1 proteins exhibit unique antigenic determinants not present in IgA2 myeloma proteins (23, 86, 87). Importantly, parallel determination of such IgA1-specific antibodies revealed significantly increased levels in sera of IgAN patients (21, 77, 85). There are several sites in the HR of the IgA1 heavy chains that display a high degree of heterogeneity in their O-linked GalNAc without Gal (63, 64). The specific site(s) with the O-linked GalNAc that serves as the epitope(s) recognized by autoantibodies to form CIC that accumulate in the mesangium of IgAN patients has not been identified (65). Based on the reactivities of IgG anti-Gd-IgA1 antibodies with the HR fragments generated by the proteolytic cleavage of IgA1 myeloma protein, we propose that GalNAc on Thr228 and Thr233 may be involved (66).

However, the basis for the induction of these naturally occurring GalNAc-specific antibodies has not been conclusively established. Certain microorganisms, including EBV, express O-linked glycan chains on their surfaces which may induce such antibodies (88–93). Gp350/220 EBV complex is expressed on the virion surface at high density, thereby allowing efficient cross-linking of the CD21 and activating resting B cells to proliferate. Gp350 is heavily glycosylated with N- and O-linked glycans, together contributing 60% of its total molecular weight (88, 89).

The biochemical and biological activities of immune complexes containing Gd-pIgA1 and IgG were first reported after their characterization and isolation from sera of IgAN patients (3, 10, 21, 25, 77). The specificity of antibodies for GalNAc residues in O-linked glycans of the IgA1 HR was demonstrated by the inhibition of re-association of acid-dissociated immune complexes in the presence of other GalNAc-containing glycoproteins or free GalNAc (21). The biological properties manifested as proliferation of mesangial cells induced in vitro were dependent on the molecular mass of such complexes. Those of molecular mass ~700-1,000 kDa displayed the stimulatory effect whereas smaller complexes did not (3, 25, 26). Thus, we concluded that immune complexes composed of Gd-pIgA1 and GalNAc-specific IgG in the circulation of IgAN patients are nephritogenic and responsible for the clinical expression of IgAN (1–3, 6).

The pathways of differentiation of lymphocytes of B cell lineage into IgA-producing cells have been extensively studied using polyclonal stimulation with various cytokines, pokeweed mitogen (PWM), and EBV (51–55, 135, 147–151). EBV infects B cells in the earliest stages of their differentiation pathway ( Figure 4 ). Human pro-, pre-, immature, and mature B cells are EBV infectable due to the presence of an EBV receptor, CD21 (52). Interestingly, as EBV-infected B cells mature, the transition of pro– and pre–B cells into phenotypically characterized plasma cells is not accompanied by a parallel synthesis of Igs (52). This lack of production is due to the “sterile” differentiation pattern, with the failure of VDJ rearrangement (52). Such plasma cells contain abundantly expressed J chain that is otherwise involved in the polymerization of IgA and IgM (62). In cell culture, J chain is not secreted in the free form into the supernatants and remains strictly in the intracellular compartment (52). EBV infection of peripheral blood B cells induces their differentiation into lymphoblasts, plasmablasts, and plasma cells secreting Ig of all major isotypes (51). Although EBV induces intracellular production of IgA of both subclasses, only IgA1 is secreted into cell culture supernatants after extended incubation (53, 54). This finding is of considerable importance in IgAN because the IgA in elevated levels in patients’ sera is exclusively of the IgA1 subclass (1–3, 9). Furthermore, most of the secreted IgA1 is in the polymeric, J chain-containing, form (51, 55) and is Gal-deficient ( Table 2 ) (18).

The preferential synthesis of IgA1 that is Gal-deficient may be due to competition between IgA1 and the gp350 protein of EBV for galactosylation. Galactosylation of GalNAc residues in the HR of IgA1 involves the enzymatically-mediated transfer of Gal from the donor, UPD-Gal, to the recipient GalNAc residues (64). Importantly, the galactosylation of the IgA1 HR in EBV-infected plasma cells proceeds with the parallel production of EBV with its heavily O-glycosylated gp350 glycoprotein (67, 88, 89). Thus, it is conceivable that the GalNAc residues on the HR of IgA1 and gp350 of EBV compete for UDP-Gal as well as access to the requisite enzymes C1GalT1 and Cosmc, resulting in the reduced galactosylation of IgA1 HR ( Figure 2 ).

In EBV-infected cells, the activity and gene expression of several enzymes involved in the synthesis of the O-glycans in the IgA1 HR are altered (18). The activity of β1,3-galactosyltransferase that adds Gal to GalNAc is deceased. The galactosylation is further stressed by reduced expression of encoding gene, C1GalT1, and the gene encoding Cosmc, the chaperone for β1,3-galactosyltransferase that maintains its enzymatic activity. Furthermore, another enzyme, ST6GalNAc, exhibits increased activity and its gene is overexpressed. The resulting enhanced α2,6-sialylation prevents attachment of Gal to GalNAc-S/T in the IgA1 HR, thereby accentuating synthesis of Gd-IgA1 (75).

Recently, Dotz et al. (152) found that decreased sialylation of IgA1 is associated with decreased estimated glomerular filtration rate (eGFR) in patients with IgAN. Although this study used mass spectrometry to analyze serum total IgA containing IgA1 and IgA2 in monomeric and polymeric forms, the finding may provide a new biomarker for monitoring disease activity. Two earlier reports confirmed that Gd-IgA1-specific autoantibodies in IgAN patients bound more Gd-IgA1 after removal of sialic acid (23, 86). Based on the observation of Dotz et al, acute removal of sialic acid from Gd-IgA1 HR, for example due to infection by neuraminidase-secreting viruses (such as influenza) or bacteria (such as pneumococci), could contribute to increased amounts of Gd-IgA1 in the circulation, leading to enhanced binding of autoantibodies and formation nephritogenic CIC. This scenario may explain the clinical association of macroscopic hematuria with mucosal infection in patients with IgAN (3). Because mucosal infection induces a general inflammatory response that includes stimulation of IgA-secreting cells and Gd-IgA1 production and because desialylation of IgA is the natural catabolism of IgA, the postulated contribution of infection-mediated desialylation of Gd-IgA1 to disease activity in IgAN should be tested in future studies.

EBV profoundly influences the expression of receptors on infected B cells, with the preferential expression of those involved their homing to tonsils and the upper respiratory tract (50, 153, 154). Thus, EBV, as well as other viruses (155), direct the ultimate tissue distribution of these cells through expression of pertinent homing receptors. In the case of EBV infection, integrin α4β7 (LPAM-1) is induced in the tonsils of patients with infectious mononucleosis, thereby allowing B cells to home to the gastrointestinal mucosa-associated lymphoid tissue (GALT) (156).

EBV as well as other herpesviruses establish life-long and latent residence in target cells of the host and evade elimination (141, 142, 157–160). EBV infects only humans (157, 160). The ensuing clinical manifestations depend on the type and magnitude of the induced immune responses and age of the host (141, 157–160). Acute EBV infection of children usually remains clinically silent (141, 158–161). EBV is present in saliva to provide an easy means to spread the virus to uninfected individuals. EBV crosses the epithelial barrier of the oral cavity and nasopharynx to infect susceptible B cells to induce their proliferation and maturation to the Ig-secreting plasma cells or to establish persistent residence (141, 142, 157). The most important lymphoepithelial tissue susceptible to EBV infection is Waldeyer’s ring which includes the adenoids and palatine, tubal, and lingual tonsils with adjacent draining lymph nodes ( Figure 5 ) (141, 142, 157). EBV establishes persistent infection in resident, long-lived, memory B cells (142) and retains the ability to replicate in activated and differentiated plasma cells that produce antibodies of various isotypes and also release the virus ( Figure 4 ) (141, 143, 144). In these oropharyngeal lymphoid tissues, the virus can initiate a new round of B cell and perhaps epithelial cell infection that leads to further shedding into the saliva (141). However, the possibility of EBV replication in epithelial cells remains controversial; virus in complexes with antibodies may be also internalized through the Ig receptors expressed on epithelial cells (141). The frequency of EBV-infected B cells is highly variable, ranging from 5 to 3,000 infected cells/10⁷ memory B cells in Waldeyer’s ring and peripheral blood; other lymphoid tissues (spleen and mesenteric lymph nodes) contain at least a 20-fold fewer infected cells (141, 142, 157). The virus replicates in terminally differentiated B cells (plasma cells) in Waldeyer’s ring but only a few cells participate in viral production (143, 157). However, the life-span of such infected plasma cells in the upper respiratory tract and oral cavity has not been determined. In mucosal tissues and the bone marrow, IgA-producing plasma cells persist for a surprisingly long time, up to 10-20 years (113). Asymptomatic EBV reactivation in oral mucosa-associated lymphoid tissues occurs periodically in most healthy subjects (94). This event may be caused by reactivation of EBV-infected resting memory B cells upon their entry into lymphoid tissue and physiologic stimulation through the B cell receptor, leading to terminal differentiation into plasma cells and activation of the viral replicative cycle, associated with expression of viral glycoprotein gp350.

Due to the presence of pIgA-producing cells with J chain in the bone marrow of IgAN patients, the glomerular IgA may originate in the bone marrow as well as mucosal tissues (163–165). However, the altered glycosylation pattern and the possible presence of EBV in the bone marrow of IgAN patients has not been addressed. It is conceivable that in these patients there are indeed Gd-pIgA1λ-producing cells in the bone marrow in addition to the lymphoid tissue in the upper respiratory tract and, perhaps, other mucosal tissues. In these tissues, EBV remains associated with infected memory B cells. Upon the stimulation of these cells, probably by infection with other microorganisms, some differentiate into plasma cells that produce Gd-IgA1λ and release EBV.

IgA vasculitis (formerly known as Henoch-Schönlein purpura) is the most common vasculitis in children, characterized by leukocytoclastic inflammation and IgA in the small blood vessels in the skin, joints, intestines, and (in a minority of patients) kidneys (192–195). IgAN and IgA vasculitis with nephritis share some common clinical, laboratory, and pathology features, including increased levels of Gd-IgA1 in the circulation and accumulation of Gd-IgA1 in glomeruli, suggesting a related immunopathogenesis (17, 43, 196–199). In contrast, IgA vasculitis patients without nephritis have normal circulating levels of Gd-IgA1 (196, 197). Furthermore, serum Gd-IgA1 levels are elevated in many first-degree relatives of pediatric patients with IgAN and IgA vasculitis with nephritis (43, 196). Based on the mechanisms involved in the glycosylation of HR of IgA1 (196, 200) and the familial epidemiology of EBV infection, it is plausible to speculate that EBV is also involved in the aberrant glycosylation of IgA1 in IgA vasculitis with nephritis. Indeed, several case reports support this possibility (201–203). The occurrence of IgA vasculitis with nephritis in children seropositive for acute EBV infection has suggested a role for the virus in the acute syndrome (201–203).

Based on the pleiotropic impact of EBV infection on B cells as related to the stage of maturation of the IgA system, the production of Gd-pIgA1λ as the autoantigen, and the preferential expression of homing receptors specific for the upper respiratory tract, we propose a novel pathway in the immunopathogenesis of IgAN. Although almost all adults are infected with EBV, there are marked differences in the time of infection among individuals of diverse racial and, more importantly, socio-economic backgrounds. Most African Americans, African Blacks, and Australian Aborigines are infected with EBV in very early childhood (1-2 years of age) without overt clinical symptoms. At that time, the IgA system is physiologically immunodeficient, manifested as absent or low serum levels of IgA and a paucity of IgA-producing cells in lymphoid tissues. Consequently, when very young children are exposed to EBV, the virus enters B cells of non-IgA isotype. The ensuing humoral and cellular immune responses prevent subsequent infection of IgA⁺ B cells during EBV reactivation at older ages when IgA B cells are more numerous. EBV infection of IgA-secreting cells markedly increases the fraction of synthesized IgA that is Gd-pIgA1λ. Therefore, EBV infection at a very young age may significantly reduce the lifetime risk of development of IgAN. Thus, the low incidence of IgAN in the above specified populations may reflect immunological, age-related, genetic, and pronounced socio-economic differences from populations with higher incidences of IgAN with respect to the frequency of early acquisition of EBV infection.

JM conceived and designed the manuscript and wrote the original draft. All authors contributed to the article and approved the submitted version.