IgA nephropathy (IgAN) is an inflammatory kidney disease that is characterized by the deposition of immunoglobulin A (IgA) in the glomerular mesangium (Figure 1). It presents on a wide clinical spectrum (Table 1) and is the most common pattern of primary glomerular disease reported worldwide (Chembo et al., 2015; Hwang et al., 2010; Nakai et al., 2006). Incident patients are commonly young adults, and up to 40% progress to end-stage kidney disease (ESKD) within 20 years of diagnosis (Reid et al., 2011; Wyatt & Julian, 2013). To date, there are no approved disease-specific treatments.

Mesangial IgA deposition characteristic of IgAN

TABLE 1. Common presentations of IgAN Presentation Mode of detection Asymptomatic Detected incidentally on biopsy during kidney donation or post-mortem Non-visible haematuria ± mild proteinuria Detected on routine testing, for screening purposes or on unrelated medical assessments Visible haematuria Commonly within 72 hours of an upper respiratory tract infection Hypertension Detected on assessment of secondary causes of hypertension, usually following subsequent detection microscopic haematuria or impaired kidney function Chronic kidney disease Could be detected on routine testing, or incidentally during unrelated medical assessments Rapidly progressive glomerulonephritis The most severe presentation, characterized by rapid loss of kidney function in hours to days, culminating in ESKD Nephrotic syndrome Rare presentation of generalized oedema, greater than 3 g of proteinuria per day, hypoalbuminaemia and hyperlipidaemia

The paucity of current treatments stems from an incomplete understanding of the underlying disease mechanisms and the slowly progressive course of IgAN, which has historically made adequately powered clinical trials challenging and prohibitively expensive to conduct. This is rapidly changing. Evolving laboratory techniques and the acceptance of surrogate end points for ESKD by regulatory authorities have ushered in a new era of clinical trials in IgAN (Inker, Lesley A. et al., 2019; Inker, Lesley A., MD, MS et al., 2016). Contributing to our understanding is the growing evidence of the role of the gut mucosal immune system in this disease. In this review, we examine established immunological paradigms in IgAN, explore how mucosal immune responses can drive disease, and discuss how this is being leveraged for treatment.

1 THE FOUR-HIT HYPOTHESIS

There is a marked heterogeneity in the incidence and severity of IgAN between both individuals and geographical locations, suggesting that its pathogenesis is driven by genetic and environmental factors. Deposition of IgA in the glomerular mesangium, however, occurs universally in IgAN, causing varying degrees of inflammation and fibrosis (Allen et al., 1995,1999, 2001). The pathogenesis of IgAN has been framed by the four-hit hypothesis, which necessitates four events to occur for clinically significant disease to develop.

1.1 Hit 1: Poorly O-galactosylated IgA1

IgA is the primary immunoglobulin at mucosal surfaces, where it plays a key role in the host defence of a surface area of approximately 400 m2 that is under constant antigenic challenge. It is the most highly produced immunoglobulin in humans and exists in monomeric (mIgA) and polymeric (pIgA) forms (Kerr, 1990). mIgA is predominantly produced in the bone marrow and bears a typical immunoglobulin structure, with antigen binding and effector poles separated by a ‘hinge’ region (Figure 2) (Kerr, 1990). The structure of this hinge region differentiates the IgA1 and IgA2 subclass, the two isoforms produced in humans. The IgA1 hinge region is longer and usually possesses several O-linked glycans (Woof & Russell, 2011). The O-glycosylation process involves addition of N-acetylgalactosamine (GalNac) to threonine and/or serine residues catalysed by the action of N-acetylgalactosaminyltransferase 2 (GALNT2), the GalNac then undergoes O-galactosylation by core 1 ß1,3-galactosyltransferase (C1GALT1) facilitated by the chaperone protein Cosmc (Figure 2) (Knoppova et al., 2016).

Naturally occurring states of IgA in humans. Monomeric IgA2 has the structure of a typical immunoglobulin. It consists of two heavy chains and two light chains, with the antigen binding (Fab) and effector poles (Fc) separated by a hinge region. Monomeric IgA1 has a wider hinge region than IgA2 and is ordinarily equipped with several O-linked glycosylations. For these glycosylations to occur, GalNac is added to threonine or serine residues under the action of GALNT2, which then undergoes O-galactosylation by C1GALT1 facilitated by the chaperone protein Cosmc. Polymeric IgA most commonly consists of two IgA monomers bound by a peptide called ‘J-chain’. Secretory IgA is formed when the polymeric Ig receptor is cleaved following transcytosis, when IgA is trafficked to mucosal surfaces. The secretory component binds polymeric IgA and protects it from enzymatic degradation

Disruption of this galactosylation process is a defining feature of IgAN. Mucosal IgA1 is poorly O-galactosylated (which can be measured by lectin binding assays, which preferentially bind the exposed GalNac residues), and this form of IgA1 is enriched in the serum, urine and within glomerular deposits in IgAN (Allen et al., 2001; Moldoveanu et al., 2007; Suzuki, H. et al., 2016; Yeo et al., 2018). The appearance of poorly O-galactosylated IgA1 (also referred to as galactose-deficient IgA1, or gd-IgA1) is driven by dysregulation of C1GALT1 and Cosmc expression in B cells, influenced by genetic and environmental factors. A genome-wide association study (GWAS) identified alleles at a single locus spanning the C1GALT1 gene which associated with elevated levels of serum gd-IgA1 (Gale et al., 2017). However, IgD (the only other O-glycosylated immunoglobulin in humans) is normally galactosylated in IgAN, implying this dysregulation is not an intrinsic B-cell defect but instead one that occurs following class switch recombination (Smith et al., 2006).

Despite the universality of elevation of gd-IgA1 levels in IgAN, elevated levels alone are not sufficient to induce the disease. Increased levels of gd-IgA1 are not specific to IgAN, and in vitro, monomeric gd-IgA1 cannot induce the mesangial hypercellularity that is characteristic of this disease (Lin et al., 2009; Novak et al., 2005).

1.2 Hit 2: Anti-gd-IgA1 Antibodies with specificity for gd-IgA1 are elevated in IgAN and are associated with poor prognosis (Berthoux et al., 2012; Yanagawa et al., 2014). They may form through:

An autoantibody response against a ‘neo-epitope’ of exposed GalNaAc residues at the hinge region of gd-IgA1 (Berthoux et al., 2012).

A response to microbial GalNac, which through ‘molecular mimicry’ mis-targets gd-IgA1 (Novak et al., 2008).

The requisite for anti-gd-IgA1 antibodies in IgAN remains uncertain. While IgG co-localization in the glomerular mesangium is reported in IgAN, it is not universal, and there is little evidence of activation of the classical arm of the complement pathway (Figure 3) (Maillard et al., 2015; Rizk et al., 2019). Hit 2 may not, therefore, be critical, but could nevertheless contribute to disease should it occur.

The complement system. The complement system can be activated by three arms. The classical pathway is activated by antigen–antibody complexes. These complexes bind C1q inducing a conformational change which activates C1r to cleave C1s into an activated serine protease. C1s is then able to cleave C4 to C4a and C4b. C4b binds C2, allowing its cleavage by a serine protease, producing C2b. C4b and C2b bind together to generate a C3 convertase. The lectin binding pathway is activated by mannose moieties commonly found on microbial surfaces, which are bound by an MBL. This activates MASP1 and MASP2, leading to C4bC2b formation. The alternative pathway is a consistently activated pathway, fuelled by the hydrolysis of C3 thioester bonds. C3(H20) is bound by factor B, allowing it to be cleaved by factor D, to produce C3(H20)Bb – a C3 convertase. The common pathway is initiated by C3 convertases formation, which cleaves C3 to C3a and C3b. C3b is further acted upon by the C3 convertases to produce a C5 convertase, which cleaves C5 to produce C5a, a potent inflammatory mediator and C5b. C5b is serially bound by C5, C6, C7, C8 and C9 to form the membrane attack complex, which is capable of cell lysis. C3b can also be bound by factor B of the alternative pathway to form C3bB, which can be cleaved by factor D to produce the C3 convertase C3bBb. C3b activity is regulated by hydrolysis and factor I, which inhibit its activity

1.3 Hit 3: gd-IgA1 containing circulating immune complexes (gd-IgA1 CICs) gd-IgA1 CICs are elevated in the serum in IgAN and correlate with disease severity (Suzuki et al., 2014). CICs could form either as

Anti-gd-IgA1 IgG – gd-IgA1 CICs (Suzuki et al., 2014; Tomana et al., 1999)

Anti-gd-IgA1 IgA – gd-IgA1 CICs (Suzuki et al., 2014; Tomana et al., 1999)

gd-IgA1 self-aggregates (Kokubo et al., 1998)

gd-IgA1 – soluble CD89 CICs (Lechner et al., 2016)

CD89 is glycoprotein found primarily on the surface of myeloid cells, where it acts as an Fc receptor for IgA (Morton & Brandtzaeg, 2001). IgA binding can induce CD89 shedding, producing a soluble form of the receptor (sCD89) which is postulated to complex with gd-IgA1 to form one of the four possible gd-IgA1 CICs (Lechner et al., 2016; van Zandbergen et al., 1999).

Unlike monomeric gd-IgA1, gd-IgA1 CICs promote mesangial proliferation and inflammation in vitro, and their glomerular presence is a key driver of inflammation in IgAN (Novak et al., 2005; Tortajada et al., 2019). However, CIC deposition on its own is not sufficient for disease; cadaveric and kidney donor studies confirm a proportion of the population have gd-IgA1 immune-complex deposits without clinically significant kidney disease (Gaber et al., 2020; Waldherr et al., 1989). A fourth ‘hit’ within the kidney seems necessary to trigger inflammation.

1.4 Hit 4: Mesangial IgA1 deposition with variable inflammation and fibrosis Mesangial IgA1 deposition may occur through interactions with

The transferrin receptor (CD71)

Soluble CD89 (sCD89)

Transglutaminase-2

β-1,4-galactosyltransferase-1

Each of these has been reported to be upregulated in IgAN (Berthelot et al., 2012; Molyneux et al., 2017; Moura et al., 2005), and mouse models yield insights into how they may interact to promote mesangial IgA binding (Berthelot et al., 2012). In IgAN, IgA1 has been reported to bind and promote CD89 shedding from circulating myeloid cells, leading to circulating IgA1–sCD89 complexes (Lechner et al., 2016; van Zandbergen et al., 1999). In a humanized transgenic mouse model of IgAN, this complex was found to bind mesangial CD71, promoting both CD71 and transglutaminase-2 expression, both of which subsequently bind further IgA1–sCD89 complexes (Berthelot et al., 2012; Daha et al., 2013). These findings have been replicated in vitro (Berthelot et al., 2012). A role for β-1,4-galactosyltransferase-1 has also been demonstrated in vitro; mesangial cells incubated with β-1,4-galactosyltransferase-specific antibodies bound less IgA, secreted less IL-6 and demonstrated less IgA-induced mesangial cell phosphorylation of spleen tyrosine kinase (Molyneux et al., 2017).

Mesangial deposition of gd-IgA1 CICs can drive a local inflammatory response, resulting from cytokine release and activation of the complement system (Lai et al., 2009; Tortajada et al., 2019). The extent to which this occurs is variable, leading to heterogeneity in disease progression.

1.4.1 Inflammatory cytokines and proteinuric tubular damage

Binding of gd-IgA1 CICs prompts mesangial cell proliferation and extracellular matrix production in vitro, mirroring findings in IgAN (Launay et al., 2000; Novak et al., 2005). This stimulation leads to cytokine production, including TNFα, which disrupts podocyte function, compromising the glomerular barrier leading to proteinuric tubular damage (Huang et al., 2008; Lai et al., 2009; Trimarchi & Coppo, 2019). Tubulo-interstitial damage may also occur via filtration of mesangial cell–derived cytokines (termed glomerulo-tubular cross-talk; Chan et al., 2005).

1.4.2 Complement activation The complement system is a major driver of glomerular inflammation in IgAN (Tortajada et al., 2019). The system consists of a cluster of proteins present in plasma and extracellular fluid which can be rapidly activated in a cascade fashion (Vignesh et al., 2017). Activation can occur via three pathways (Figure 3):

The classical pathway: activated by antigen–antibody complexes

The alternative pathway: activates spontaneously at low rates, but also activated by moieties on microbial surfaces

The lectin pathway: activated by saccharide moieties on microbial surfaces

The alternative and lectin pathways drive inflammation in IgAN, with little evidence of classical pathway activity (Tortajada et al., 2019). C3 co-deposits with IgA in IgAN (Tortajada et al., 2019). This suggests alternative pathway activity and is supported by the frequent presence of the alternative pathway regulatory proteins factor H and properdin in IgAN biopsies (Bene & Faure, 1987; Zhang et al., 2009). Alternative pathway activation leads to local inflammatory damage, and in keeping with this, glomerular C3 deposition correlates with poor prognosis (Nam et al., 2020). The exact mechanism of alternative pathway activation in IgAN is unclear; although IgA1 may directly activate the alternative pathway, gd-IgA1 does not appear to be essential (Hiemstra et al., 1987; Russell & Mansa, 1989).

C4d deposition is also reported in IgAN (Maeng et al., 2013; Nam et al., 2020). Taken together with the absence of C1q and the presence of lectin pathway components [mannose binding lectin (MBL) and MBL-associated serine protease (MASP)] in IgAN biopsies, this strongly suggests lectin pathway activity without involvement of the classical pathway (Jennette, 1988; Roos et al., 2006). This activation ultimately triggers the common complement pathway, cleaving C3 leading to local inflammation (Figure 3), and as expected, glomerular C4d deposition correlates with poor outcomes (Nam et al., 2020; Roos et al., 2006). In a study of 60 IgAN patients in the Netherlands, 25% demonstrated lectin pathway activity (evidenced by MBL, MASP and C4d glomerular deposition); these patients had poorer renal function, more proteinuria and worse histology at presentation, all of which predict a poor prognosis (Barbour et al., 2019; Roos et al., 2006). These findings were supported by a subsequent Spanish study, which demonstrated poorer kidney survival in those with glomerular C4d deposition (28% vs 85% renal survival at 20 years) (Espinosa et al., 2014). There are plausible mechanisms by which the lectin pathway may be activated in IgAN; pIgA purified from those with IgAN activates the lectin pathway in vitro, likely through interaction with exposed GalNac resides (Beltrame et al., 2015; Roos, A. et al., 2001).

2 LEVERAGING THE FOUR-HIT HYPOTHESIS FOR TREATMENT

Evidence of complement activity is driving a surge in the interest of complement targeting therapeutics for IgAN. Among those currently being studied in clinical trials are cemdisiran, iptacopan (LNP023), IONIS-FB-LRx, avacopan and narsoplimab (OMS721) (summarized in Table 2).

TABLE 2. Complement-directed therapies undergoing clinical trials for IgAN Drug Trial identifier Mechanism of action Cemdisiran NCT03841448 C5 inhibitor (oligonucleotide RNA inhibitor) IONIS-FB-LRx NCT04014335 Factor B inhibitor (antisense oligonucleotide) LNP023 NCT03373461 Factor B inhibitor (small-molecule inhibitor) OMS721 NCT02682407 MASP-2 inhibitor (monoclonal antibody) Avacopan NCT02384317 C5a receptor blocker (small-molecule receptor blocker)

Other therapeutics attempting to leverage the four-hit hypothesis for treatment have not been successful. Rituximab, a B-cell–depleting monoclonal antibody directed against CD20, was investigated as a treatment for IgAN (Lafayette et al., 2017). Rituximab is used in the treatment of several glomerular diseases, including anti-neutrophil cytoplasm antibody (ANCA)-associated vasculitis and lupus nephritis, where it depletes circulating auto-antibodies and improves kidney outcomes (Fervenza et al., 2019; Rovin et al., 2012; Stone et al., 2010). Rituximab failed to be of value in IgAN and, curiously, had no effect on gd-IgA1 or anti-gd-IgA1 IgG levels (Lafayette et al., 2017). This finding raises questions about the origins of gd-IgA1 in IgAN.

3 BEYOND THE FOUR-HIT HYPOTHESIS: THE SIGNIFICANCE OF ANTIGEN–MUCOSAL INTERACTIONS

Interactions between mucosal immune cells and local antigens have demonstrable roles in the generation of polymeric gd-IgA1. Indeed, IgA-producing CD20-negative plasmablasts and plasma cells residing in gut-associated lymphoid tissue are protected from rituximab-induced depletion (Mei et al., 2010). Their role in gd-IgA1 production may account for the negative results of the rituximab trial.

Mucosal involvement in the pathophysiology of IgAN is suggested by a key set of observations:

Mesangial gd-IgA1 deposits in IgAN are polymeric (Bene & Faure, 1987; Waldherr et al., 1983)

Secretory IgA is found in IgAN biopsies and correlates with poor outcomes (Oortwijn et al., 2007)

Mucosal infections trigger IgAN flares

Microbiome variations associate with IgAN severity (Cao et al., 2018; De Angelis et al., 2014; Piccolo et al., 2015; Watanabe et al., 2017)

Microbial DNA motifs upregulate mucosal IgA production via B-cell activating factor (BAFF) and a proliferation inducing ligand (APRIL; Blaas et al., 2009; Katsenelson et al., 2007)

The efficacy of TRF-budesonide as a treatment for IgAN (Fellström et al., 2017)

Each observation has an immunological basis grounded in mucosal immune cell-antigen interactions, as discussed below:

3.1 Mesangial gd-IgA1 deposits in IgAN are polymeric

pIgA is predominantly manufactured by plasma cells residing in mucosal-associated lymphoid tissue (MALT) (Kerr, 1990). It is trafficked across the mucosal epithelium into secretions following binding to the polymeric Ig receptor (Perše & Večerić-Haler, 2019). At the luminal surface, this receptor is subsequently cleaved to form a stabilizing polypeptide chain (‘secretory component’), which binds to pIgA to form secretory IgA (sIgA). This protects pIgA from bacterial protease and host enzymatic degradation (Figure 2) (Perše & Večerić-Haler, 2019).

pIgA is elevated in serum in IgAN and is the major form of IgA1 to deposit in the mesangium, indicating a mucosal B-cell origin (Bene & Faure, 1987; Waldherr et al., 1983). The appearance of pIgA in the serum may be explained by ‘reverse trafficking’ from mucosal surfaces or by mis-homing of activated mucosal B cells to bone marrow (Barratt et al., 2007; Emmerson et al., 2011). Following antigen exposure, naïve mucosal B cells undergo class switch recombination from IgM/IgD to IgA mediated by T-cell–dependent and –independent mechanisms (Stavnezer et al., 2008). Activated B cells then enter the lymphatics and home to central lymphoid tissue, undergo maturation and home back to the lamina propria of the gut where they reside as IgA-producing plasma cells (Stavnezer et al., 2008). This process may be disrupted in IgAN, leading to activated B cells mis-homing to the bone marrow from where they may release ‘mucosal’ polymeric gd-IgA1 into circulation, eventually leading to IgA deposition in the kidney (Barratt et al., 2007).

The appearance of serum pIgA in IgAN may also occur due to a dysregulation of LIGHT, which is a lymphotoxin β receptor ligand (a member of the tumour-necrosis factor superfamily). LIGHT has roles in mediating apoptosis and inflammation, and mice which overexpress LIGHT demonstrate mesangial IgA and C3 deposition in association with proteinuria and haematuria. These mice also demonstrate intestinal inflammation, which correlates with increased IgA-producing MALT plasma cells, but inversely correlates with IgA trafficking across the gut mucosal barrier. This associates with increased serum pIgA, findings which were mirrored in patients with inflammatory bowel disease (Wang et al., 2004). LIGHT-mediated inflammation was thus proposed to disrupt pIgA trafficking across the intestinal mucosa and predispose to pIgA deposition in the kidneys.

3.2 Secretory IgA correlates with worse outcomes in IgAN

sIgA is solely produced at mucosal surfaces and is present in increased concentrations in the serum and glomeruli in IgAN, correlating with disease severity (Oortwijn et al., 2007; Zhang et al., 2008). In vitro studies suggest sIgA binds mesangial cells to promote inflammatory cytokine production, including TNFα, MCP1 and IL-6 (Liang et al., 2015). Co-localization of sIgA and MBL in IgAN biopsies has also been reported with plausible mechanisms for how the two interact (Oortwijn et al., 2008), suggesting sIgA invokes the lectin binding pathway to further mediate glomerular damage.

Of note, involvement of sIgA in IgAN may be race dependent. sIgA was present in the mesangium of 15% of Caucasian IgAN patients, but 100% of Chinese patients (Oortwijn et al., 2007; Zhang et al., 2008). This could partly account for heterogeneity in disease progression between the two groups.

Although the precise mechanism by which sIgA appears in the serum in IgAN is unclear, postulations are similar to those for pIgA. The presence of pIgA and sIgA in the serum and glomeruli of IgAN patients is highly suggestive of mucosal B-cell involvement in disease. This may account for the common observation of IgAN flares following mucosal infections.

3.3 Mucosal infections trigger IgAN flares

A common presentation of IgAN is visible haematuria following an upper respiratory tract infection (Table 1). This association can continue throughout the natural history of IgAN, with mucosal infections triggering disease flares.

It was postulated that this may be due to a direct effect of infectious agents on the glomerulus, supported by descriptions of microbial antigen presence in mesangial deposits in IgAN (including cytomegalovirus, adenovirus, herpes simplex, Haemophillus parainfluenzae and Staphylococcus aureus) (Gregory et al., 1988; Koyama et al., 2004; Nagy et al., 1984; Suzuki et al., 1994). M proteins expressed by Streptococcus pyogenes, a common upper respiratory tract pathogen, contain an IgA binding region which co-localized with mesangial IgA deposits in some paediatric IgAN cases (Schmitt et al., 2010). These M proteins were subsequently found to preferentially bind gd-IgA1 and synergistically increase mesangial proliferation and production of IL-6 and C3, compared to gd-IgA1 alone (Schmitt et al., 2014). However, direct interactions between infectious agents and the mesangium has not been reported consistently, and it does not account of the presence of pIgA deposits in IgAN.

Another plausible explanation may relate to pathogen–mucosal immune cell interactions, leading to increased mucosal pIgA1 production which subsequently inundates the kidneys during episodes of mucosal infection. Evidence for this arises from animal and in vitro models.

Mesangial expansion and IgA, IgG and C3 deposition in the ddY mouse model of IgAN were markedly reduced when mice were reared in germ-free conditions (Kano et al., 2021). When germ-free mice were subsequently exposed to environmental pathogens, substantial increases in mesangial IgA, IgG and C3 deposition were observed, in tandem with increased proteinuria, compared to mice that remained in germ-free conditions. Nasal-associated lymphoid tissue was found to be the primary site of gd-IgA1 production in these mice, compared to mesenteric lymph nodes and Peyer's patches. Gd-IgA1 production was enhanced by nasal immunization with CpG oligodeoxynucleotide (a highly antigenic component of bacterial cell walls) and toll like receptor (TLR) 9 interactions (Kano et al., 2021).

Furthermore, B cells isolated from the lamina propria of the colon from humans upregulate IgA production following activation of cell-surface pattern recognition receptors (PRR), including TLRs, with in vitro exposure to CpG oligodeoxynucleotides (Blaas et al., 2009). TLR ligation has been shown to suppress Cosmc expression in B cells isolated from IgAN patients, leading to increased gd-IgA1 production (Qin et al., 2008).

These animal model and in vitro findings mirror responses observed in IgAN patients; mucosal antigens produce exaggerated IgA responses in those with IgAN (Barratt et al., 1999). IgAN patients infected with the mucosal pathogen Helicobacter pylori produce an exaggerated antigen-specific polymeric gd-IgA1 response compared to healthy subjects, whereas responses to systemic antigen exposure, such as immunization with tetanus toxoid, produce a largely monomeric normally galactosylated IgA1 response in both healthy subjects and patients with IgAN (Smith et al., 2006).

These findings together suggest a mechanism by which mucosal, and not systemic, antigens could continually drive the production of polymeric gd-IgA1, which over time leads to progressive accumulation of immune-complex deposits in the kidneys.

3.4 Variations in the mucosal microbiome associate with IgAN severity

The exaggerated pIgA response to H. Pylori raises the possibility of mucosal microbiome alterations contributing to the pathophysiology of IgAN. The gut ‘microbiome’ refers to the genetic profile of the estimated 100 trillion microbes that reside in the digestive tract (Turnbaugh et al., 2007). The role of the microbiome in disease is being increasingly characterized, driven by advances in gene sequencing and mass spectrometry. The gut microbiome is altered by a range of factors including diet, medication and circulating hormone profiles, and alterations have also been associated with immune-mediated diseases (Turnbaugh et al., 2007).

Variations in the mucosal microbiome may be associated with IgAN severity. An analysis of the gut microbiome found greater representation of Streptococcaceae and Eubacteriaceae and lower Lactobacillacae in patients with progressive IgAN (De Angelis et al., 2014). Furthermore, tonsillar crypt microbiomes have been shown to differ significantly between IgAN patients and healthy subjects, further hinting at a role of mucosal pathogens in IgAN (Watanabe et al., 2017). Salivary and sub-gingival microbiomes are also altered in IgAN (Cao et al., 2018; Piccolo et al., 2015).

It is thus plausible that polymeric gd-IgA1 production in IgAN is driven by mucosal interactions between the host microbiome and the MALT. The mechanism by which this occurs is likely related to activation of PRRs, such as TLR, which promote IgA class switch recombination and modification of the glycosylation of mucosal IgA.

3.5 Microbial motifs upregulate mucosal IgA production via BAFF and APRIL

The activation of B cells requires antigen presentation and a co-stimulatory signal which may occur through T-cell–dependent or –independent means (Stavnezer et al., 2008). There is evidence that T-cell–independent B-cell activation is mediated by BAFF and APRIL, which are members of the TNF superfamily and produced by cells of the innate immune system in response to pathogen-associated molecular patterns (PAMPs) (Litinskiy et al., 2002; MacLennan & Vinuesa, 2002). BAFF and APRIL activate B cells in the presence of antigens through interaction with cell-surface receptors: the BAFF receptor (BAFF-R), B-cell maturation antigen (BCMA) and T-cell activator and calcium-modulating ligand interactor (TACI). Ligation of these receptors promotes IgA class switch recombination and B-cell survival (Litinskiy et al., 2002; MacLennan & Vinuesa, 2002; Stavnezer et al., 2008). There is mounting evidence that both BAFF and APRIL mediate the microbial-driven effects on IgA B-cell responses in the mucosa (Katsenelson et al., 2007; McCarthy et al., 2011).

Serum BAFF and APRIL are elevated in IgAN and correlate with disease severity. Mice overexpressing BAFF (BAFF-Tg mice) have more IgA-producing plasma cells in the lamina propria of the gut and have high serum levels of polymeric and gd-IgA1 (McCarthy et al., 2006, 2011). These mice also develop a severe IgA-mediated nephritis, which is not observed in BAFF-Tg mice that are IgA deficient (McCarthy et al., 2011). This exaggerated IgA response is reduced by nearly 100-fold when BAFF-Tg mice are raised in germ-free conditions and so lack a gut microbiome. These mice have no glomerular IgA deposits (McCarthy et al., 2011). Reintroduction of gut commensal organisms to BAFF-Tg mice rescues the exaggerated IgA response and consequent mesangial IgA deposition. Collectively, these data highlight the role of the gut microbiome in driving gd-IgA1 production in concert with BAFF, which may ultimately lead to IgAN.

Dendritic cells stimulated with CpG-ODN upregulate APRIL production, and when B cells are cultured in the resulting supernatant, significant increases in IgA production are seen which are APRIL mediated (Hardenberg et al., 2007). CpG-ODN also results in elevated IgA production in murine B cells, and this response is almost completely abrogated in TACI knockout mice (Katsenelson et al., 2007).

These findings are complemented by the results of a large meta GWAS in IgAN, which identified a panel of IgAN risk loci which were related to the regulation of gut immune and mucosal integrity (Kiryluk et al., 2014). This included TNFSF13, which encodes APRIL, and ITGAM and ITGAX, which code for integrins that appear on intestinal dendritic cells and modulate IgA-producing plasma cells in the murine gut (Kunisawa et al., 2013), as well as CARD9, VAV3, PSMB8, PSMB9, TAP1, TAP2, HLA-DQA1, HLA-DQB1, HLA-DRB1, LIF, OSM, HORMAD2, MTMR3, DEFA1, and DEFA3-6, all of which are linked with intestinal mucosal integrity or diseases (Kiryluk et al., 2014). Indeed, intestinal epithelial tight junctions were found to be compromised in a rat model of IgAN compared to controls, with a clear downregulation of tight junction proteins, including ZO-1, OCLN, and MUC2 (Peng et al., 2013; Zhou et al., 2020). Supporting this, intestinal permeability as measured by 51Cr-EDTA was markedly higher in patients with IgAN compared to healthy controls and correlated with proteinuria (Kovács et al., 1996). It is possible that increased intestinal permeability facilitates interactions between intestinal pathogens and mucosal immune cells.

Of note, the GWAS study also found an association between the frequency of these risk loci and local microbial diversity within the included cohorts (Kiryluk et al., 2014), suggesting that acquisition of these alleles was protective against microbial pathogens but increased the risk of developing IgAN. These observations may begin to explain the increased incidence of IgAN in the pathogen-rich and diverse geographies of East and South East Asia.

Taken together, these observations suggest that IgAN may occur as a result of an exaggerated mucosal polymeric gd-IgA1 response to mucosal antigens, derived both from the prevalent commensal microbiome and from intermittent exposure to environmental pathogens, mediated predominantly through PAMPS and PRRs and the downstream effects of BAFF and APRIL (Figure 4).