BRIEF REVIEW www.jasn.org

The Pathophysiology of IgA Nephropathy

Hitoshi Suzuki,*‡ Krzysztof Kiryluk,† Jan Novak,‡ Zina Moldoveanu,‡ Andrew B. Herr,¶ ʈ Matthew B. Renfrow,§ Robert J. Wyatt,** Francesco Scolari,†† Jiri Mestecky,‡ ʈ Ali G. Gharavi,† and Bruce A. Julian‡

*Department of Internal Medicine, Division of , Juntendo University Faculty of Medicine, Tokyo, Japan; †Department of Medicine, Columbia University, College of Physicians and Surgeons, New York, New York; Departments ʈ of ‡Microbiology, §Biochemistry and Molecular Genetics, and Medicine, University of Alabama at Birmingham, Birmingham, Alabama; ¶Department of Molecular Genetics, Biochemistry and Microbiology, University of Cincinnati College of Medicine, Cincinnati, Ohio; **Department of Pediatrics, University of Tennessee Health Sciences Center and the Children’s Foundation Research Center at the Le Bonheur Children’s Hospital, Memphis, Tennessee; and ††Second Division of Nephrology, Montichiari Hospital, University of Brescia, Montichiari, Italy

ABSTRACT Here we discuss recent advances in understanding the biochemical, immunologic, in the O-linked glycans in the hinge re- and genetic pathogenesis of IgA nephropathy, the most common primary glomer- gion of the heavy chain. Blood levels of a ulonephritis. Current data indicate that at least four processes contribute to similarly aberrantly glycosylated IgA1 development of IgA nephropathy. Patients with IgA nephropathy often have a are higher in patients with IgAN than in genetically determined increase in circulating levels of IgA1 with galactose-defi- healthy controls or patients with other cient O-glycans in the hinge-region (Hit 1). This aberrancy is, how- diseases. However, as we discuss ever, not sufficient to induce renal injury. Synthesis and binding of here, a high circulating load of galactose- directed against galactose-deficient IgA1 are required for formation of immune deficient IgA1 alone does not induce the complexes that accumulate in the glomerular mesangium (Hits 2 and 3). These renal injury. Rather, several sequential immune complexes activate mesangial cells, inducing proliferation and secretion of processes or hits are necessary for the , cytokines, and chemokines, which result in renal injury (Hit 4). clinical expression of IgAN. Recent genome-wide association studies identify five distinct susceptibility loci—in the MHC on chromosome 6p21, the complement factor H locus on chromosome 1q32, and in a cluster of on chromosome 22q22—that potentially influence these PATHOGENESIS OF IgAN processes and contain candidate mediators of disease. The significant variation in prevalence of risk alleles among different populations may also explain some of the Four processes come together to induce re- sizable geographic variation in disease prevalence. Elucidation of the pathogenesis of nal injury that culminates in IgAN: aber- IgA nephropathy provides an opportunity to develop disease-specific therapies. rant glycosylation of IgA1, synthesis of antibodies directed against galactose-defi- J Am Soc Nephrol 22: 1795–1803, 2011. doi: 10.1681/ASN.2011050464 cient IgA1, binding of the galactose-defi- cient IgA1 by the anti-glycan/glycopeptide antibodies to form immune complexes, IgA nephropathy (IgAN) was described tion. IgAN frequently recurs in allo- and accumulation of these complexes in histologically for the first time in 1968 by grafts; in contrast, kidneys from donors the glomerular mesangium to initiate renal Berger and Hinglais as les de´poˆts intercap- with subclinical IgAN are clear of IgA de- injury. We recently performed a genome- illaires d’IgA-IgG (intercapillary deposits posits shortly after transplantation into of IgA-IgG).1 Over the ensuing decades, recipients with non-IgAN renal dis- this renal disease has been recognized as eases.2 The glomerular IgA eluted from Published online ahead of print. Publication date the most common primary glomerulo- tissue specimens from patients with available at www.jasn.org. and has been shown to arise IgAN is exclusively of the IgA1 sub- H.S. and K.K. contributed equally to this work. from a systemic process wherein the kid- class, predominantly in the polymeric Correspondence: Dr. Bruce A. Julian, Division of neys are damaged as innocent bystand- form, and, most importantly, glycosy- Nephrology, Department of Medicine, THT 643, 1530 Third Avenue South, Birmingham, AL 35294. ers. The latter point is best illustrated by lated aberrantly. Specifically, this aber- Phone: 205-934-9045; Fax: 205-934-7742; E-mail: the experience with renal transplanta- rant IgA1 exhibits galactose deficiency [email protected]

J Am Soc Nephrol 22: 1795–1803, 2011 ISSN : 1046-6673/2210-1795 1795 BRIEF REVIEW www.jasn.org

wide association study (GWAS) that iden- 255 228 230 232 236 A -Pro-Val-Pro-Ser-Thr-Pro-Pro-Thr-Pro-Ser-Pro-Ser-Thr-Pro-Pro-Thr-Pro-Ser-Pro-Ser-Cys- tified five susceptibility loci for IgAN and provided molecular candidates for these processes.3

Hit 1: Hereditary Increase in Galactose-Deficient Circulating IgA1 As is the case for other immunoglobu- lins, IgA1 is glycosylated. An altered pattern of its glycosylation has been recognized as a potentially pathogenic abnormality in IgAN for nearly 20 years.4 The key feature is the deficiency of galactose in the hinge region of the IgA1 heavy chains. The hinge region of IgA1 extends by 13 amino acids longer than the hinge region of IgA2 and is found only in humans and higher pri- 5 mates (Figure 1A). It can carry up to B six relatively short and simple sugar α2,3 α2,3 chains, called O-linked glycans, each attached by glycosidic linkage to an ox- α2,6 β1,3 β1,3 β1,3 α2,6 β1,3 α2,6 ygen atom of a serine or threonine. These glycans are synthesized in stepwise Ser/Thr Ser/Thr Ser/Thr Ser/Thr Ser/Thr Ser/Thr fashion and include up to six different forms (Figure 1B). Patients with IgAN I II III IV V VI have increased circulating levels of IgA1 Figure 1. Human IgA1: hinge-region amino acid sequence (A) and possible glycan with abbreviated glycans composed of N- variants (B). (A) IgA1 contains up to six O-glycans per hinge region: five major sites are acetylgalactosamine (GalNAc), with or shown (in orange or magenta) and the sixth site is Thr233.9 Novel approaches using without sialic acid, that are devoid of a ga- IgA-specific bacterial proteases and lectin binding, and, more recently, high-resolu- lactose moiety, as in Figure 1B structures I tion mass spectrometry with electron capture and electron transfer dissociation, have and II. A GalNAc-specific lectin, Helix been used to determine O-glycan heterogeneity, the sites of glycosylation, and the 6,7,9,45,46 aspersa agglutinin, is frequently used in microheterogeneity at the individual sites. The model of intact IgA1 was generated from published crystal and solution structures of IgA1.47,48 N- and O- an ELISA to measure the amount of ga- glycans were modeled using the GlyProt server and related databases (http://www. lactose-deficient IgA1 (IgA1 with glycosciences.de), based on observed IgA1 glycoforms.9,49 For clarity, the O-glycans hinge-region O-glycan structures I and are shown with transparent spheres for each atom, and are colored orange for II in Figure 1B) after treatment of the GalNAc-galactose residues and magenta for GalNAc; the illustrated O-glycan distri- IgA1 with neuraminidase to remove bution was taken from a study by Takahashi et al.9 (B) The variants of O-glycans on terminal sialic acid.6 Recent high- circulatory IgA1. Galactose-deficient glycans present in elevated amounts in patients resolution mass spectrometry studies with IgAN are represented by structures I and II in magenta.6,19 Galactosylated demonstrate significant heterogeneity variants are in orange as structures III to VI. The largest O-glycan on circulatory IgA1 in the composition and attachment is a GalNAc-galactose with two sialic acids, i.e., tetrasaccharide, structure VI. IgA1 sites of these glycans.7–12 Determina- with GalNAc and sialylated GalNAc (structures I and II in magenta) is present at tion of the precise structural abnor- elevated serum levels in patients with IgAN due to the changes in expression and activity of specific glycosyltransferases, ST6GalNAcII, and C1GalT1.50 The stability of mality in IgAN is complicated by the C1GalT1 during translation is controlled by Cosmc, a foldase.51 Structure I is gener- multiplicity of O-glycosylation sites ated by a GalNAc-transferase52; structure II, by ST6GalNAcII53; structure III, from and combinatorial possibilities of gly- structure I by C1GalT154; and structures IV to VI, by sialyltransferases. Symbols: can variants. However, examination of rectangle, GalNAc; circle, galactose; diamond, sialic acid. IgA1-producing cells shows that aber- rant O-glycosylation in IgAN is driven by combined abnormalities in expres- (Figure 1), implicating complex regu- ficient IgA1 originate from and reside re- sion or activity of the glycosyltrans- latory defects.13 main uncertain. Circulatory IgA1 is pro- ferases involved in the sequential post- In patients with IgAN, the predominant duced mainly in the , whereas translational modification of IgA1 sites where the cells secreting galactose-de- aberrantly glycosylated IgA1 may be syn-

1796 Journal of the American Society of Nephrology J Am Soc Nephrol 22: 1795–1803, 2011 www.jasn.org BRIEF REVIEW

thesized in response to a mucosal infection, Hit 2: Circulating Antibodies susceptibility loci within the MHC com- and thus abnormalities in the mucosal re- Directed against plex.3 The strongest genetic effect was sponse to common microbial or food anti- Galactose-Deficient IgA1 observed for the MHC-II locus contain- gens may be involved in production of ga- Aberrantly glycosylated IgA1 in the ing the HLA-DQB1, DQA1, and DRB1 lactose-deficient IgA1.14 For example, blood of patients with IgAN is found genes. This effect appeared to be con- blood levels of galactose-deficient IgA1 di- nearly exclusively within immune com- veyed by a highly protective haplotype rected against mucosal pathogens are in- plexes bound to IgG or IgA1 antibodies. DRB1*1501-DQA1*0102-DQB1*0602. creased in patients with IgAN compared We have recently shown that these IgG Specifically, the DQB1*0602 allele re- with those of healthy controls.14 This pro- antibodies recognize GalNAc-contain- duced the odds of disease by over 50% cess may include dysregulated innate im- ing epitopes on the galactose-deficient per copy. This is a relatively common mune responses through Toll-like recep- hinge region O-glycans of IgA1, defining classical HLA allele, present in 10 to 20% tors.15 an autoimmune component to IgAN.19 of Europeans and 2 to 10% of Asians.22 Serum levels of galactose-deficient IgA1 Furthermore, these IgG autoantibodies The second independent genetic effect is are above the 90th percentile for healthy exhibit unique features in the comple- from a region encompassing two genes controls in as many as 70 to 80% of IgAN mentarity-determining region 3 (CDR3) encoding transporters associated with patients.6 Furthermore, 40 to 50% of first- of the variable region of their heavy processing (TAP1 and TAP2) degree relatives of IgAN patients have ele- chains.13 Specifically, the third position and two genes encoding components of vated levels comparable to that of patients, in CDR3 is typically serine in patients the immunoproteasome (PSMB8 and demonstrating significant heritability of with IgAN, a feature necessary for effi- PSMB9). These molecules process anti- this trait.16,17 The heritability of galactose- cient binding of the IgG to galactose-de- gens in the cytosol and transport them deficient IgA1 is observed in diverse racial ficient IgA1. Importantly, serum levels of into the endoplasmic reticulum for de- groups and is not explained by variation in IgG antibodies specific for galactose- livery to the cell surface in association serum IgA levels, suggesting that distinct deficient IgA1 correlated with disease with MHC-I molecules. It is not yet clear genetic mechanisms influence IgA1 glyco- severity, as assessed by the magnitude which of these four genes is involved in sylation and production. These data also of . It is not known whether the susceptibility to IgAN. Finally, the indicate that aberrant IgA1 glycosylation the CDR3 serine substitution origi- third locus of association is on chromo- precedes clinically overt disease and consti- nates from somatic mutations that some 6p21 and encodes MHC-II mole- tutes an inherited risk factor for the devel- arise during maturation of the anti- cules DPA1, DPB1, and DPB2, and thus opment of IgAN. Moreover, because most body-producing cells or from inherited also relates to the process of antigen persons with elevated levels of galactose- germline mutations. These antibodies presentation. Taken together, these deficient IgA1 do not exhibit clinical signs are also present in sera of healthy indi- new genetic findings strongly implicate of renal injury, this hereditary defect is in- viduals, albeit at lower levels. One can adaptive immunity in the pathogenesis sufficient to cause IgAN, implicating addi- postulate that these antibodies are pro- of IgAN, and define the genetic context tional pathogenic hits as described in this duced in response to bacterial or viral required for the recognition of galac- article.16 cell-surface GalNAc-containing glyco- tose-deficient IgA1 as an antigen and GWAS data have identified a major conjugates on commensal or infectious for generation of pathogenic anti- locus on chromosome 22q12.2 influenc- microorganisms and then cross-react glycan antibodies. ing susceptibility to IgAN.3 This locus is with galactose-deficient O-linked gly- also associated with variation in serum cans on IgA1. A predominant IgA1 au- Hit 3: Formation of Pathogenic IgA levels and has been previously asso- toantibody response19 directed against IgA1-Containing Immune ciated with risk of inflammatory bowel galactose-deficient IgA1 may explain Complexes disease,18 further implicating this inter- why some patients have IgA1 as the sole It is generally agreed that, in IgAN, the val in the regulation of mucosal inflam- Ig isotype in the glomeruli. It is to be mesangial cells represent the primary mation. Two cytokine genes within the noted, however, that the presence of target of pathogenic deposits formed by associated region, LIF and OSM, are ex- IgG in the renal biopsy specimens also circulating immune complexes (Figure 2, cellent positional candidates, as both are correlates with mesangial and endo- solid lines) or by lanthanic deposits of expressed in B cells and may participate capillary cellularity.20 aberrantly glycosylated IgA1, followed in the regulation of mucosal immunity. The strongest signals in the recent by binding of newly generated anti- It is not yet known if this locus also influ- GWAS for IgAN were localized within glycan antibodies to form immune ences aberrant IgA glycosylation. Further the MHC complex (encoding polypep- complexes in situ (Figure 2, broken studies, including resequencing of this tides that present to T lines).23 The presence of circulating locus and evaluation of its effect on se- cells),3,21 a region highly associated IgA1-containing immune complexes is rum levels of galactose-deficient IgA1, with risk for many autoimmune disor- not unique to patients with IgAN. Such will clarify causal variants and their role ders. Based on careful conditional anal- complexes can be detected in persons in the synthesis of IgA1. yses, we identified three independent without apparent renal disease, includ-

J Am Soc Nephrol 22: 1795–1803, 2011 Pathophysiology of IgAN 1797 BRIEF REVIEW www.jasn.org

Hit 1 Hit 2 ing healthy individuals and patients Increased circulating Production of unique with Henoch-Schoenlein galactose-deficient IgA1 anti-glycan antibodies without nephritis.19,24,25 The com- plexes in patients with Henoch- Schoenlein purpura without nephritis Hit 3 consist of IgA, but not IgG, and are of smaller mass than the complexes found Formation of pathogenic IgA1-containing circulating immune complexes in patients with IgAN. As these persons do not develop overt renal disease, it can be assumed that these IgA com- plexes are not nephritogenic. In con- Hit 4 trast, patients with Henoch-Schoenlein Mesangial deposition and activation of purpura with nephritis have larger cir- mesangial cells resulting in glomerular injury culating immune complexes contain- ing IgA and IgG.24 By analogy with other human diseases caused by im- mune complexes, it is likely that, in

Proliferation IgAN, the molecular proportion of an- ECM production tigens (galactose-deficient IgA1) and Cytokines antibodies (IgG or IgA1) determines Growth factors the size of the formed immune com- Mesangial cell plexes and, consequently, their rate of IgA1 complexes removal from the circulation, as well as biologic activity. The pathogenic circu- lating IgA1-IgG immune complexes in patients with IgAN are relatively large Cytokines (Ͼ800 kD) and thus may be excluded from entry into the hepatic space of Disse to reach the asialoglycoprotein receptor (ASGP-R) on hepatocytes, the normal catabolic pathway for circula- tory IgA1. As a result, these immune complexes enter the renal circulation. Due to the unique location of the mes- angium between the fenestrated endo- thelial lining of the capillaries and the glomerular basement membrane, the PodocytePod mesangium is prone to deposition of Figure 2. Proposed pathways involved in the pathogenesis of IgAN: multi-hit mechanism. immune complexes. While it is not Hit 1: Production of galactose-deficient IgA1 by a subpopulation of IgA1-secreting cells. IgA1 completely understood what deter- production may be affected by the IgAN-associated locus on chromosome 22q12.2.3 Hit 2: mines the entry of circulating immune Formation of anti-glycan antibodies with specific characteristics of the variable region of the complexes into the mesangium, the heavy chain that recognize galactose-deficient IgA1. Hit 3: Formation of immune complexes factors involved likely include the size from autoantigen (galactose-deficient IgA1) and O-glycan-specific antibodies. Hits 2 and 3 of immune complexes, their amount, 3 may be regulated by the three MHC loci on chromosome 6p21 associated with risk of IgAN. and local hemodynamic factors.26 The Hit 4: Deposition of pathogenic immune complexes in the mesangium, activation of mesangial biologic activity of large circulating cells, and induction of glomerular injury. Hits 3 and 4 may be affected by genotype at the complement factor H locus on chromosome 1q32 that regulates the alternative complement immune complexes with galactose- cascade.3 The first pathway assumes formation of immune complexes in the circulation and their deficient IgA1 increases in IgAN pa- subsequent mesangial deposition (solid lines).13,19,55,56 An alternative theory proposes that some tients during episodes of macroscopic of the aberrantly glycosylated IgA1 molecules are in the mesangium as lanthanic deposits (left .27 However, it is not known broken line) and are later bound by newly generated anti-glycan antibodies to form immune whether this increase in activity is due to complexes in situ (right broken line) that activate mesangial cells.23 ECM, extracellular matrix. greater production of galactose-deficient IgA1, anti-glycan antibodies, or other un- defined factors influencing the formation of these complexes and/or their composi-

1798 Journal of the American Society of Nephrology J Am Soc Nephrol 22: 1795–1803, 2011 www.jasn.org BRIEF REVIEW

tion.28,29 MHC risk alleles may participate binds immune complexes containing ga- These mechanistic issues have impor- in this step by influencing the efficiency of lactose-deficient IgA1, leading to en- tant clinical and therapeutic implications antigen presentation, recognition, and hanced expression of CD71.35,36 This because subclinical findings consistent processing, and subsequent activation of binding creates a positive feedback loop, with IgAN are common in the general autoreactive B cells. causing overexpression of CD71 on pro- population. Necropsy series found glo- liferating mesangial cells. However, it is merular IgA deposits in 2% and 4% of Hit 4: Mesangial Deposition of not known whether CD71 is the only re- individuals in Singapore and Germany, IgA1-Containing Immune ceptor that binds IgA1-containing im- respectively.39,40 Even more striking, a Complexes, Cell Activation, and mune complexes or whether it has a di- study in Japan showed that 16% of 510 Initiation of Glomerular Injury rect pathogenic role in IgAN. renal allografts at engraftment were af- The pathogenetic importance of im- Activation of the fected, of which 19 (3.7%) had a mesan- mune complexes has been shown by in in glomeruli augments the inflammatory gioproliferative nephritis.41 To date, the vitro studies. The glomerular injury of cascade and potentiates tissue injury in glycosylation pattern of this lanthanic IgAN histologically manifests as prolifer- IgAN. The immune complexes with IgA1 glomerular IgA has not been examined ation of mesangial cells and expansion of can activate complement via the alterna- to determine if it differs from that found extracellular-matrix components. The tive or lectin pathway. The pattern of gly- in IgAN patients. Such an analysis would detailed mechanisms of activation of cosylation of IgA1 and the molecular clarify whether the IgA deposits are clin- mesangial cells remain to be elucidated. mass of IgA1-containing immune com- ically silent because they have a different Nonetheless, cultured human mesangial plexes are also important factors in the composition that renders them relatively cells provide a convenient model for ability of IgA1 to activate the alternative inert or because there is an intrinsic hy- evaluating the biologic activities of IgA1- complement pathway.37 Accordingly, re- poresponsiveness of the kidney in which containing complexes. Immune com- nal biopsy specimens have usually de- they are deposited. plexes from patients with IgAN contain- tectable C3, while the components of the ing galactose-deficient IgA1 bind to the classical pathway, such as C1q, are typi- cells more efficiently than do uncom- cally absent. Our recent GWAS identi- POTENTIAL NEW DIAGNOSTIC plexed IgA1 or immune complexes from fied a major IgAN susceptibility locus AND PROGNOSTIC MARKERS healthy controls. Complexes with galac- within the complement factor H tose-deficient IgA1 induce cultured hu- (CFH) cluster on chromosome 1q32. In Blood man mesangial cells to proliferate, se- Products of CFH and its neighboring Based on the central role of galactose- crete extracellular matrix components, CFHR (CFH-related) genes participate in deficient IgA1 in the pathogenesis, and release humoral factors such as the modulation of the alternative path- Moldoveanu et al.6 investigated the TNF␣, IL-6, and TGF␤. These factors way by binding C3a and C5a convertases. value of serum levels of this protein as a can, in turn, alter podocyte gene expres- Mutations in CFH lead to uncontrolled diagnostic test. By receiver operating sion and glomerular permeability.30,31 In activation of the alternative pathway and characteristic (ROC) curve analysis, the contrast, uncomplexed galactose-defi- cause inherited forms of membranopro- serum level of galactose-deficient IgA1 cient IgA1 or relatively small immune liferative type II, a that provided a 0.77 sensitivity had a complexes (Ͻ800 kD) have no stimula- disease pathologically distinct from specificity of 0.90 to distinguish IgAN tory effect on cellular proliferation. IgAN. However, carriers of a common patients from healthy controls, while a The cellular receptors on mesangial deletion encompassing the neighboring level with a specificity of 1.00 had a sen- cells involved in the binding of IgA1 are CFHR1 and CFHR3 genes had an ap- sitivity of 0.44.6 Other groups have repli- not well characterized. IgA1-containing proximately 30% decreased risk of devel- cated these findings.4,42 Importantly, the immune complexes display a high affin- oping IgAN. The risk was almost 60% serum level of galactose-deficient IgA1 ity for the extracellular-matrix compo- lower in the rare individuals who carry may be significantly elevated long before nents fibronectin and type IV collagen in two copies of this deletion.3 The role of the diagnosis of IgAN (Olson S. et al., un- the mesangium, and preferentially bind CFHR1 and CFHR3 proteins in the reg- published observation). and activate mesangial cells. None of the ulation of complement cascade is cur- IgG specific for galactose-deficient well-known IgA receptors (CD89, poly- rently under active investigation. Based IgA1 represents another potential bio- meric Ig receptor, ASGP-R) and comple- on early experimental data, however, marker, as serum levels of this ment receptors (CR 1–3) have been con- CFHR1 and CFHR3 compete with CFH are significantly elevated in IgAN pa- firmed on human mesangial cells.32,33 for binding to C3b, the key activator of tients, and the levels correlate with pro- However, transferrin receptor (CD71), the terminal portion of the complement teinuria. ROC curve analysis indicates which is expressed on the surface of pro- pathway.38 Therefore, a relative loss of that when the specificity of the level of liferating human mesangial cells, can CFHR1 and CFHR3 may enhance the in- serum IgG antibody directed against ga- bind polymeric IgA1.34 Moreover, CD71 hibitory action of CFH and thus convey lactose-deficient IgA1 reached 0.95, the on human mesangial cells effectively protection against local . corresponding sensitivity was 0.88.13

J Am Soc Nephrol 22: 1795–1803, 2011 Pathophysiology of IgAN 1799 1800 RE REVIEW BRIEF ora fteAeia oit fNephrology of Society American the of Journal www.jasn.org

Table 1. Summary of the four hits involved in the pathogenesis of IgAN Putative Environmental Putative Genetic Factors Potential Novel Therapeutic Hit Pathogenic Process Potential Clinical Biomarkers Factors Involved Involved Approaches 1 Hereditary increase in circulating Potential role of mucosal Strong evidence for high Serum galactose-deficient IgA1 Suppression of synthesis of galactose-deficient IgA1 exposure to infectious heritability of serum level (HAA-based ELISA) galactose-deficient IgA1 or dietary antigens galactose-deficient Enzymatic boost of galactose IgA1 level transfer to IgA1 hinge-region Potential role of O-glycans chromosome 22q12.2 Suppression of sialylation of galactose-deficient O-glycans 2 Circulating antibody directed Potential role of mucosal Potential role of three Serum anti-glycan antibodies Alteration of processing and against galactose-deficient exposure to infectious MHC-II loci in antigen (dot-blot assay) presentation of galactose- IgA1 or dietary antigens presentation and deficient IgA1 O-glycopeptides humoral response to Specific B-cell depletion therapy galactose-deficient IgA1 O-glycans 3 Formation of pathogenic Unknown Unknown Circulating and/or urinary Competitive blockade of immune IgA1-containing immune immune complexes complex formation by complexes non-cross-linking anti-glycan antibodies or specific glycopeptides 4 Mesangial deposition of Unknown Protective effect of Circulating and/or urinary Suppression of the alternative IgA1-containing immune common deletion in complement degradation complement pathway complexes, cell activation and CFHR1 and CFHR3 products, or novel markers Targeted CHFR1/3 depletion initiation of glomerular injury of glomerular injury Blocking mesangial cell signaling

mScNephrol Soc Am J induced by nephritogenic IgA1-containing immune complexes HAA; Helix aspersa agglutinin, a lectin specific for terminal GalNAc 22: 7510,2011 1795–1803, www.jasn.org BRIEF REVIEW

In Urine rational therapies for IgAN (Table 1). ies directed against the aberrantly glyco- Urinary proteomics holds promise for devel- This model predicts that the generation sylated O-linked hinge-region glycans to opment of noninvasive tests for IgAN. A sub- of nephritogenic immune complexes form immune complexes, accumulation set of mesangial immune complexes appar- composed of galactose-deficient IgA1 of the complexes in the mesangium, and ently enters the urinary space. Aberrantly and anti-glycan autoantibodies initiates activation of mesangial cells. Genetic fac- galactosylated IgA1 within immune com- disease. Therefore, interventions that tors apparently influence the expression plexes has been found in the urine of patients can reduce generation of galactose-defi- of these hit mechanisms. Elucidation of withIgANbutnotinpatientswithnon-IgAN cient IgA1 or anti-glycan antibodies, or the pathogenesis of IgAN provides an proteinuric glomerular diseases.43 It is also block the interaction between these two opportunity to develop a disease-specific possible to develop a diagnostic test without a components to form nephritogenic im- therapy that heretofore has been missing. detailed knowledge of the pathogenesis, mune complexes, may prove effective. based on analysis of the urinary peptidome. This result can be achieved by enhancing 44 In a preliminary study, Julian et al. found enzymatic activity of glycosyltransferases DISCLOSURES that analysis of urine samples by capillary for synthesis of galactose-replete hinge- Supported in part by grants DK082753, electrophoresis, coupled online with mass region O-glycans to reduce the availabil- DK078244, DK083663, DK075868, DK080301, spectrometry, distinguished patients with ity of aberrantly glycosylated IgA1 for DK077279, DK071802, and K23-DK090207 from primary IgAN from patients with IgA-im- formation of nephritogenic immune the National Institutes of Health and a grant from mune-complex renal disease due to cirrhosis, complexes. Alternatively, generation of the IgA Nephropathy Foundation of America. even if clinical proteinuria was absent. non-cross-linking, monovalent reagents (single-chain antibodies) with high af- Genetic finity for GalNAc on galactose-deficient REFERENCES The GWAS of sporadic IgAN identified IgA1 could theoretically prevent binding five novel genetic variants with relatively of anti-glycan antibodies. Lastly, block- 1. Berger J, Hinglais N: Les de´poˆ ts intercapil- 3 strong protective effects against IgAN. ing the antibodies with a glycopeptide laires d’IgA-IgG (Intercapillary deposits of While these variants are all common, their may be another strategy to prevent for- IgA-IgG). J Urol Nephrol 74: 694–695, 1968 frequencies vary significantly across differ- mation of immune complexes (Table 1). 2. Silva FG, Chander P, Pirani CL, Hardy MA: ent continental populations and closely Similarly, interventions aimed at re- Disappearance of glomerular mesangial IgA parallel the prevalence rates of IgAN. For ducing immune complex deposition and deposits after renal allograft transplantation. Transplantation 33: 241–246, 1982 example, African populations, which have the downstream inflammatory signals 3. Gharavi AG, Kiryluk K, Choi M, Li Y, Hou P, the lowest reported prevalence of IgAN, may prove beneficial. The genetic studies Xie J, Sanna-Cherchi S, Men CJ, Julian BA, carry the most protective alleles, while identify the alternative complement Wyatt RJ, Novak J, He JC, Wang H, Lv J, Zhu Asians, who have the highest reported pathway as a prime candidate for inter- L, Wang W, Wang Z, Yasuno K, Gunel M, prevalence, have significantly fewer protec- vention, and predict that targeted deple- Mane S, Umlauf S, Tikhonova I, Beerman I, Savoldi S, Magistroni R, Ghiggeri GM, Bo- tive variants. The ROC analysis for a ge- tion of CFHR1 and/or CFHR3 would be dria M, Lugani F, Ravani P, Ponticelli C, Al- netic risk score based on these five alleles is tolerated and prove protective. More- legri L, Boscutti G, Frasca G, Amore A, Pe- estimated in the range of 0.60 to 0.63, with over, blocking of specific signaling path- ruzzi L, Coppo R, Izzi C, Viola BF, Prati E, the risk score explaining up to 7% of dis- ways induced in mesangial cells by Salvadori M, Mignani R, Gesualdo L, Berti- ease variation among Asian and Caucasian pathogenic IgA1-containing complexes netto F, Mesiano P, Amoroso A, Scolari F, Chen N, Zhang H, Lifton RP: Genome-wide populations. Identification of additional can be theoretically accomplished by association study identifies susceptibility loci genetic susceptibility loci in follow-up protein-kinase inhibitors, a class of drugs for IgA nephropathy. Nat Genet 43: 321– GWAS studies of Caucasian populations that is frequently used in the treatment of 327, 2011 will likely improve the predictive power of some types of . 4. Coppo R, Feehally J, Glassock RJ: IgA ne- the genetic risk score. Additionally, a com- phropathy at two score and one. Kidney Int 77: 181–186, 2010 posite risk score based on a combined as- 5. Mestecky J, Moro I, Kerr MA, Woof JM: sessment of circulating, urinary, and ge- CONCLUSIONS Mucosal immunoglobulins. In: Mucosal Im- netic disease biomarkers holds promise to munology, 3rd ed, edited by Mestecky J, ultimately provide a tool for the noninva- IgAN is an autoimmune renal disease Bienenstock J, Lamm ME, Mayer L, McGhee sive diagnosis of IgAN. arising from consequences of increased JR, Strober W, Amsterdam, Elsevier Aca- demic Press, 2005, pp 153–181 circulating levels of IgA1 with galactose- 6. Moldoveanu Z, Wyatt RJ, Lee J, Tomana M, deficient hinge-region O-glycans. How- Julian BA, Mestecky J, Huang WQ, Anreddy POTENTIAL APPROACHES FOR ever, this glycosylation aberrancy alone is S, Hall S, Hastings MC, Lau KK, Cook WJ, DISEASE-SPECIFIC THERAPY not sufficient to induce nephritis. For the Novak J: Patients with IgA nephropathy clinical manifestation of renal injury, have increased serum galactose-deficient IgA1 levels. Kidney Int 71: 1148–1154, 2007 The pathogenesis model (Figure 2) pro- several additional hits are required, in- 7. Renfrow MB, Cooper HJ, Tomana M, Kul- vides an opportunity to design and test cluding synthesis of circulating antibod- havy R, Hiki Y, Toma K, Emmett MR,

J Am Soc Nephrol 22: 1795–1803, 2011 Pathophysiology of IgAN 1801 BRIEF REVIEW www.jasn.org

Mestecky J, Marshall AG, Novak J: Determi- Barker CV, Woodford SY, Lifton RP, loci: A meta-analytic review of 497 popula- nation of aberrant O-glycosylation in the Mestecky J, Novak J, Julian BA: Aberrant tion studies. Hum Immunol 69: 443–464, IgA1 hinge region by electron capture dis- IgA1 glycosylation is inherited in familial and 2008 sociation Fourier transform-ion cyclotron sporadic IgA nephropathy. J Am Soc Neph- 23. Glassock RJ: The pathogenesis of IgA ne- resonance mass spectrometry. J Biol Chem rol 19: 1008–1014, 2008 phropathy. Curr Opin Nephrol Hypertens 280: 19136–19145, 2005 17. Kiryluk K, Julian BA, Wyatt RJ, Scolari F, 20: 153–160, 2011 8. Renfrow MB, MacKay CL, Chalmers MJ, Ju- Zhang H, Novak J, Gharavi AG: Genetic 24. Levinsky RJ, Barratt TM: IgA immune com- lian BA, Mestecky J, Kilian M, Poulsen K, studies of IgA nephropathy: Past, present, plexes in Henoch-Scho¨nlein purpura. Lancet Emmett MR, Marshall AG, Novak J: Analysis and future. Pediatr Nephrol 25: 2257–2268, 2: 1100–1103, 1979 of O-glycan heterogeneity in IgA1 myeloma 2010 25. Hall RP, Lawley TJ, Heck JA, Katz SI: IgA- proteins by Fourier transform ion cyclotron 18. Imielinski M, Baldassano RN, Griffiths A, containing circulating immune complexes in resonance mass spectrometry: Implications Russell RK, Annese V, Dubinsky M, Kugath- dermatitis herpetiformis, Henoch-Scho¨nlein for IgA nephropathy. Anal Bioanal Chem asan S, Bradfield JP, Walters TD, Sleiman P, purpura, systemic erythematosus and 389: 1397–1407, 2007 Kim CE, Muise A, Wang K, Glessner JT, other diseases. Clin Exp Immunol 40: 431– 9. Takahashi K, Wall SB, Suzuki H, Smith AD, Saeed S, Zhang H, Frackelton EC, Hou C, 437, 1980 Hall S, Poulsen K, Kilian M, Mobley JA, Ju- Flory JH, Otieno G, Chiavacci RM, Grund- 26. Sterzel RB, Lovett DH, Stein HD, Kashgarian lian BA, Mestecky J, Novak J, Renfrow MB: meier R, Castro M, Latiano A, Dallapiccola B, M: The mesangium and glomerulonephritis. Clustered O-glycans of IgA1: Defining mac- Stempak J, Abrams DJ, Taylor K, McGovern Klin Wochenschr 60: 1077–1094, 1982 ro- and micro-heterogeneity by use of elec- D, Silber G, Wrobel I, Quiros A, Barrett JC, 27. Novak J, Tomana M, Matousovic K, Brown tron capture/transfer dissociation. Mol Cell Hansoul S, Nicolae DL, Cho JH, Duerr RH, R, Hall S, Novak L, Julian BA, Wyatt RJ, Proteomics 9: 2545–2557, 2010 Rioux JD, Brant SR, Silverberg MS, Taylor Mestecky J: IgA1-containing immune com- 10. Wada Y, Dell A, Haslam SM, Tissot B, Canis KD, Barmuda MM, Bitton A, Dassopoulos T, plexes in IgA nephropathy differentially af- K, Azadi P, Backstrom M, Costello CE, Hans- Datta LW, Green T, Griffiths AM, Kistner EO, fect proliferation of mesangial cells. Kidney son GC, Hiki Y, Ishihara M, Ito H, Kakehi K, Murtha MT, Regueiro MD, Rotter JI, Int 67: 504–513, 2005 Karlsson N, Hayes CE, Kato K, Kawasaki N, Schumm LP, Steinhart AH, Targan SR, Xavier 28. Novak J, Moldoveanu Z, Renfrow MB, Khoo KH, Kobayashi K, Kolarich D, Kondo A, RJ, Libioulle C, Sandor C, Lathrop M, Be- Yanagihara T, Suzuki H, Raska M, Hall S, Lebrilla C, Nakano M, Narimatsu H, Novak J, laiche J, Dewit O, Gut I, Heath S, Laukens D, Brown R, Huang WQ, Goepfert A, Kilian M, Novotny MV, Ohno E, Packer NH, Palaima E, Mni M, Rutgeerts P, Van Gossum A, Zelenika Poulsen K, Tomana M, Wyatt RJ, Julian BA, Renfrow MB, Tajiri M, Thomsson KA, Yagi H, D, Franchimont D, Hugot JP, de Vos M, Mestecky J: IgA nephropathy and Henoch- Yu SY, Taniguchi N: Comparison of methods Vermeire S, Louis E, Cardon LR, Anderson Schoenlein purpura nephritis: Aberrant gly- for profiling O-glycosylation: Human Pro- CA, Drummond H, Nimmo E, Ahmad T, cosylation of IgA1, formation of IgA1-con- teome Organisation Human Disease Gly- Prescott NJ, Onnie CM, Fisher SA, Marchini taining immune complexes, and activation comics/Proteome Initiative multi-institu- J, Ghori J, Bumpstead S, Gwillam R, Tremel- of mesangial cells. Contrib Nephrol 157: tional study of IgA1. Mol Cell Proteomics 9: ling M, Delukas P, Mansfield J, Jewell D, 134–138, 2007 719–727, 2010 Satsangi J, Mathew CG, Parkes M, Georges 29. Novak J, Mestecky J: IgA Immune-complex. 11. Wada Y, Tajiri M, Ohshima S: Quantitation M, Daly MJ, Heyman MB, Ferry GD, Kirsch- In: Recent Advances in IgA Nephropathy, of saccharide compositions of O-glycans by ner B, Lee J, Essers J, Grand R, Stephens M, edited by Lai KN, Hong Kong, Imperial Col- mass spectrometry of glycopeptides and its Levine A, Piccoli D, Van Limbergen J, Cuc- lege Press and the World Scientific Pub- application to rheumatoid . J Pro- chiara S, Monos DS, Guthery SL, Denson L, lisher, 2009, pp 177–191 teome Res 9: 1367–1373, 2010 Wilson DC, Grant SF, Daly M, Hakonarson 30. Lai KN, Leung JC, Chan LY, Saleem MA, 12. Odani H, Yamamoto K, Iwayama S, Iwase H, H: Common variants at five new loci associ- Mathieson PW, Lai FM, Tang SC: Activation Takasaki A, Takahashi K, Fujita Y, Sugiyama ated with early-onset inflammatory bowel of podocytes by mesangial-derived TNF-␣: S, Hiki Y: Evaluation of the specific structures disease. Nat Genet 41: 1335–1340, 2009 Glomerulo-podocytic communication in IgA of IgA1 hinge glycopeptide in 30 IgA ne- 19. Tomana M, Novak J, Julian BA, Matousovic nephropathy. Am J Physiol Renal Physiol phropathy patients by mass spectrometry. J K, Konecny K, Mestecky J: Circulating im- 294: F945–F955, 2008 Nephrol 23: 70–76, 2010 mune complexes in IgA nephropathy consist 31. Lai KN, Leung JC, Chan LY, Saleem MA, 13. Suzuki H, Fun R, Zhang Z, Brown R, Hall S, of IgA1 with galactose-deficient hinge re- Mathieson PW, Tam KY, Xiao J, Lai FM, Julian BA, Chatham WW, Suzuki Y, Wyatt RJ, gion and antiglycan antibodies. J Clin Invest Tang SC: Podocyte injury induced by me- Moldoveanu Z, Lee JY, Robinson J, Tomana 104: 73–81, 1999 sangial-derived cytokines in IgA nephropa- M, Tomino Y, Mestecky J, Novak J: Aber- 20. Bellur SS, Troyanov S, Cook HT, Roberts IS: thy. Nephrol Dial Transplant 24: 62–72, rantly glycosylated IgA1 in IgA nephropathy Immunostaining findings in IgA nephropa- 2009 patients is recognized by IgG antibodies thy: Correlation with histology and clinical 32. Leung JCK, Tsang AWL, Chan DTM, Lai KN: with restricted heterogeneity. J Clin Invest outcome in the Oxford classification patient Absence of CD89, polymeric immunoglob- 119: 1668–1677, 2009 cohort. Nephrol Dial Transplant 26: 2533– ulin receptor, and asialoglycoprotein recep- 14. Smith AC, Molyneux K, Feehally J, Barratt J: 2536, 2011 tor on human mesangial cells. JAmSoc O-glycosylation of serum IgA1 antibodies 21. Feehally J, Farrall M, Boland A, Gale DP, Gut Nephrol 11: 241–249, 2000 against mucosal and systemic antigens in I, Heath S, Kumar A, Peden JF, Maxwell PH, 33. Novak J, Vu HL, Novak L, Julian BA, IgA nephropathy. J Am Soc Nephrol 17: Morris DL, Padmanabhan S, Vyse TJ, Mestecky J, Tomana M: Interactions of hu- 3520–3528, 2006 Zawadzka A, Rees AJ, Lathrop M, Ratcliffe man mesangial cells with IgA and IgA-con- 15. Suzuki H, Suzuki Y, Narita I, Aizawa M, PJ: HLA has strongest association with IgA taining circulating immune complexes. Kid- Kihara M, Yamanaka T, Kanou T, Tsukagu- nephropathy in genome-wide analysis. JAm ney Int 62: 465–475, 2002 chi H, Novak J, Horikoshi S, Tomino Y: Soc Nephrol 21: 1791–1797, 2010 34. Moura IC, Centelles MN, Arcos-Fajardo M, Toll-like receptor 9 affects severity of IgA 22. Solberg OD, Mack SJ, Lancaster AK, Single Malheiros DM, Collawn JF, Cooper MD, nephropathy. J Am Soc Nephrol 19: 2384– RM, Tsai Y, Sanchez-Mazas A, Thomson G: Monteiro RC: Identification of the transferrin 2395, 2008 Balancing selection and heterogeneity receptor as a novel immunoglobulin (Ig)A1 16. Gharavi AG, Moldoveanu Z, Wyatt RJ, across the classical human leukocyte antigen receptor and its enhanced expression on

1802 Journal of the American Society of Nephrology J Am Soc Nephrol 22: 1795–1803, 2011 www.jasn.org BRIEF REVIEW

mesangial cells in IgA nephropathy. J Exp 42. Shimozato S, Hiki Y, Odani H, Takahashi K, Renfrow MB, Herr AB: Analysis of IgA1 N- Med 194: 417–425, 2001 Yamamoto K, Sugiyama S: Serum under- glycosylation and its contribution to Fc␣RI 35. Moura IC, Arcos-Fajardo M, Sadaka C, Leroy galactosylated IgA1 is increased in Japa- binding. Biochemistry 47: 11285–11299, V, Benhamou M, Novak J, Vrtovsnik F, Had- nese patients with IgA nephropathy. 2008 dad E, Chintalacharuvu KR, Monteiro RC: Nephrol Dial Transplant 23: 1931–1939, 50. Suzuki H, Moldoveanu Z, Hall S, Brown R, Glycosylation and size of IgA1 are essential 2008 Vu HL, Novak L, Julian BA, Tomana M, for interaction with mesangial transferrin re- 43. Matousovic K, Novak J, Tomana M, Kulhavy Wyatt RJ, Edberg JE, Alarco´n GS, Kim- ceptor in IgA nephropathy. J Am Soc Neph- R, Julian BA, Mestecky J: IgA1-containing berly RP, Tomino Y, Mestecky J, Novak J: rol 15: 622–634, 2004 immune complexes in the urine of IgA ne- IgA1-secreting cell lines from patients with 36. Moura IC, Arcos-Fajardo M, Gdoura A, Le- phropathy patients. Nephrol Dial Transplant IgA nephropathy produce aberrantly gly- roy V, Sadaka C, Mahlaoui N, Lepelletier Y, 21: 2478–2484, 2006 cosylated IgA1. J Clin Invest 118: 629– Vrtovsnik F, Haddad E, Benhamou M, Mon- 44. Julian BA, Wittke S, Novak J, Good DM, 639, 2008 teiro RC: Engagement of transferrin recep- Coon JJ, Kellmann M, Zu¨rbig P, Schiffer E, 51. Ju T, Cummings RD: Protein glycosylation: tor by polymeric IgA1: Evidence for a posi- Haubitz M, Moldoveanu Z, Calcatera SM, Chaperone mutation in Tn syndrome. Na- tive feedback loop involving increased Wyatt RJ, Sykora J, Sladkova E, Hes O, Mis- ture 437: 1252, 2005 receptor expression and mesangial cell pro- chak H, McGuire BM: Electrophoretic meth- 52. Iwasaki H, Zhang Y, Tachibana K, Gotoh M, liferation in IgA nephropathy. JAmSoc ods for analysis of urinary polypeptides in Kikuchi N, Kwon YD, Togayachi A, Kudo T, Nephrol 16: 2667–2676, 2005 IgA-associated renal diseases. Electrophore- Kubota T, Narimatsu H: Initiation of O-glycan 37. Zhang W, Lachmann PJ: Glycosylation of IgA is sis 28: 4469–4483, 2007 synthesis in IgA1 hinge region is determined by required for optimal activation of the alternative 45. Gomes MM, Suzuki H, Brooks MT, Tomana a single enzyme, UDP-N-acetyl-␣-D-galac- complement pathway by immune complexes. M, Moldoveanu Z, Mestecky J, Julian BA, tosamine:polypeptide N-acetylgalactosaminyl- Immunology 81: 137–141, 1994 Novak J, Herr AB: Recognition of galactose- transferase 2. J Biol Chem 278: 5613–5621, 38. Fritsche LG, Lauer N, Hartmann A, Stippa S, deficient O-glycans in the hinge region 2003 Keilhauer CN, Oppermann M, Pandey MK, of IgA1 by N-acetylgalactosamine-specific 53. Raska M, Moldoveanu Z, Suzuki H, Brown R, Kohl J, Zipfel PF, Weber BH, Skerka C: An snail lectins: A comparative binding study. Kulhavy R, Hall S, Vu HL, Carlsson F, Lindahl imbalance of human complement regulatory Biochemistry 49: 5671–5682, 2010 G, Tomana M, Julian BA, Wyatt RJ, proteins CFHR1, CFHR3 and factor H influ- 46. Moore JS, Kulhavy R, Tomana M, Moldoveanu Mestecky J, Novak J: Identification and ences risk for age-related macular degener- Z, Suzuki H, Brown R, Hall S, Kilian M, Poulsen characterization of CMP-NeuAc:GalNAc- ation (AMD). Hum Mol Genet 19: 4694– K, Mestecky J, Julian BA, Novak J: Reactiv- IgA1 ␣2,6-sialyltransferase in IgA1-produc- 4704, 2010 ities of N-acetylgalactosamine-specific lec- ing cells. J Mol Biol 369: 69–78, 2007 39. Waldherr R, Rambousek M, Duncker WD, tins with human IgA1 proteins. Mol Immunol 54. Ju T, Brewer K, D’Souza A, Cummings RD, Ritz E: Frequency of mesangial IgA deposits 44: 2598–2604, 2007 Canfield WM: Cloning and expression in non-selected autopsy series. Nephrol Dial 47. Boehm MK, Woof JM, Kerr MA, Perkins SJ: of human core 1 ␤1,3-galactosyltransferase. Transplant 4: 943–946, 1989 The Fab and Fc fragments of IgA1 exhibit a J Biol Chem 277: 178–186, 2002 40. Varis J, Rantala I, Pasternack A, Oksa H, different arrangement from that in IgG: A 55. Julian BA, Novak J: IgA nephropathy: An Jantti M, Paunu ES, Pirhonen R: Immuno- study by X-ray and neutron solution scatter- update. Current Opin Nephrol Hypertens globulin and complement deposition in ing and homology modelling. J Mol Biol 13: 171–179, 2004 glomeruli of 756 subjects who had commit- 286: 1421–1447, 1999 56. Mestecky J, Tomana M, Moldoveanu Z, ted suicide or met with a violent death. 48. Herr AB, Ballister ER, Bjorkman PJ: Insights Julian BA, Suzuki H, Matousovic K, Ren- J Clin Pathol 46: 607–610, 1993 into IgA-mediated immune responses from frow MB, Novak L, Wyatt RJ, Novak J: The 41. Suzuki K, Honda K, Tanabe K, Toma H, Nihei the crystal structures of human Fc␣RI and its role of aberrant glycosylation of IgA1 mol- H, Yamaguchi Y: Incidence of latent mesan- complex with IgA1-Fc. Nature 423: 614– ecules in the pathogenesis of IgA ne- gial IgA deposition in renal allograft donors 620, 2003 phropathy. Kidney Blood Pres Res 31: 29– in Japan. Kidney Int 63: 2286–2294, 2003 49. Gomes MM, Wall SB, Takahashi K, Novak J, 37, 2008

J Am Soc Nephrol 22: 1795–1803, 2011 Pathophysiology of IgAN 1803