BRIEF COMMUNICATION www.jasn.org

Neonatal Fc Promotes Immune Complex– Mediated Glomerular Disease

† ‡ Florina Olaru,* Wentian Luo,* Hani Suleiman, Patricia L. St. John, Linna Ge,* Adam R. Mezo,§ † ‡ | Andrey S. Shaw, Dale R. Abrahamson, Jeffrey H. Miner, and Dorin-Bogdan Borza*¶

*Division of Nephrology, Department of Medicine, Vanderbilt University School of Medicine, Nashville, Tennessee; †Department of Pathology and Immunology, and |Renal Division, Washington University School of Medicine, St. Louis, Missouri; ‡Department of Anatomy and Cell Biology, University of Kansas Medical Center, Kansas City, Kansas; §Biogen Idec Hemophilia, Waltham, Massachusetts; and ¶Department of Microbiology and Immunology, Meharry Medical College, Nashville, Tennessee

ABSTRACT The neonatal (FcRn) is a major regulator of IgG and albumin homeostasis pathology induced by targeting systemically and in the kidneys. We investigated the role of FcRn in the development autoantigens at the dermal–epidermal of immune complex–mediated glomerular disease in mice. C57Bl/6 mice immunized junction, although this protection can be with the noncollagenous domain of the a3 chain of type IV collagen (a3NC1) de- overcome by excess autoantibodies.2–4 veloped albuminuria associated with granular capillary loop deposition of exoge- In kidneys, FcRn is expressed in podo- nous antigen, mouse IgG, C3 and C5b-9, and injury. High-resolution cytesandproximaltubularepithelialcells.5 imaging showed abundant IgG deposition in the expanded glomerular basement Overall, renal FcRn reclaims albumin but membrane, especially in regions corresponding to subepithelial electron dense facilitates elimination of IgG.6 Tubular deposits. FcRn-null and -humanized mice immunized with a3NC1 developed no FcRn mediates IgG transcytosis.7 Podo- albuminuria and had lower levels of serum IgG anti-a3NC1 antibodies and reduced cytes use FcRn to clear IgG from the glo- glomerular deposition of IgG, antigen, and complement. Our results show that FcRn merular basement membrane (GBM).8 promotes the formation of subepithelial immune complexes and subsequent glo- IgG accumulates in the glomeruli of 2 2 merular pathology leading to proteinuria, potentially by maintaining higher serum aged Fcrn / mice due to impaired clear- levels of pathogenic IgG antibodies. Therefore, reducing pathogenic IgG levels by ance of IgG from the GBM, and saturating pharmacologic inhibition of FcRn may provide a novel approach for the treatment of this clearance mechanism by excess ligand immune complex–mediated glomerular diseases. As proof of concept, we showed potentiates the pathogenicity of nephro- that a peptide inhibiting the interaction between human FcRn and human IgG ac- toxic sera in wild-type mice. Podocyte celerated the degradation of human IgG anti-a3NC1 autoantibodies injected into FcRn has been postulated to be involved FCRN-humanized mice as effectively as genetic ablation of FcRn, thus preventing in the clearance of immune complexes the glomerular deposition of immune complexes containing human IgG. (ICs) present in pathologic conditions 5 J Am Soc Nephrol 25: 918–925, 2014. doi: 10.1681/ASN.2013050498 such as membranous nephropathy. Ex- pression of FcRn in human is

TheMHC class I–like neonatal Fc receptor up by cells by pinocytosis bind strongly to – Received May 16, 2013. Accepted October 23, (FcRn), a heterodimer comprising a heavy FcRn at pH 6.0 6.5 in endosomes. FcRn- 2013. chain and b2-microglobulin light chain, is bound ligands are then recycled to the F.O. and W.L. contributed equally to this work. the major regulator of IgG and albumin plasma membrane, where they dissociate homeostasis.1 Perinatally, FcRn mediates at pH 7.4, whereas IgG and albumin not Published online ahead of print. Publication date the transfer of IgG from mother to off- bound to FcRn are targeted to available at www.jasn.org. spring, across the placenta in primates for degradation. FcRn is thought to pro- Correspondence: Dr. Dorin-Bogdan Borza, De- and trans-intestinally in suckling rodents. mote some autoimmune diseases because partment of Microbiology and Immunology, Meharry Medical College, 1005 Dr. D.B. Todd Jr. Throughout life, FcRn protects IgG and it protects pathogenic IgG from degrada- Boulevard, Nashville, TN 37208-3599. Email: 2 2 albumin from catabolism, explaining the tion. For instance, Fcrn / mice are re- [email protected] or [email protected] unusually long t1/2 and high serum levels sistant to passive transfer of arthritis Copyright © 2014 by the American Society of of these . IgG and albumin taken by K/BxN sera and autoimmune skin Nephrology

918 ISSN : 1046-6673/2505-918 J Am Soc Nephrol 25: 918–925, 2014 www.jasn.org BRIEF COMMUNICATION increased in various immune-mediated glomerular diseases.9 Given its role in IgG and albumin handling in the kidneys and systemically, FcRn can be expected to influence the development of immune- mediated kidney diseases at multiple lev- els. This conjecture awaits experimental verification. To determine the role of FcRn in IgG- mediated glomerular disease, we asked how FcRn deficiency alters the course of disease in mice immunized with the NC1 domain of a3typeIVcollagen(a3NC1). We chose this antigen because of its re- ported ability to induce disease in C57Bl/6 (B6) mice,10 corroborated in pi- lot studies (Supplemental Figure 1). 2 2 Fcrn / mice are hypoalbuminemic due to impaired albumin recycling,11 and also exhibit reduced urinary albumin excretion.12 As a control for this potential confounder, we used FCRN-humanized mice, which have normal because human FcRn recycles mouse al- bumin but not mouse IgG.13 All mice immunized with a3NC1 de- veloped circulating mouse IgG anti- a3NC1 antibodies, which reached the maximum titer about 6 weeks later and gradually declined thereafter. At all times, the levels of mouse IgG anti-a3NC1 anti- 2 2 bodies in sera from Fcrn / mice and FCRN-humanized mice were approxi- Figure 1. FcRn ablation reduces serum levels of mouse IgG anti-a3NC1 antibodies and prevents the development of albuminuria in a3NC1-immunized mice. (A) The left panel mately 50%–70% lower than those in 2 2 shows circulating mIgG anti-a3NC1 antibodies from C57Bl6 wild-type mice (○), Fcrn / wild-type mouse sera (Figure 1A). The re- mice (□), FCRN-humanized (hFCRN) mice (◇), and the control CFA group (△), which are sults were similar for mouse IgG1, IgG2b, assayed by indirect ELISA in plates coated with a3NC1 (100 ng/well). Mouse sera are di- and IgG2c anti-a3NC1 antibodies (Sup- luted 1:5000. The right panel shows the significance of circulating mIgG anti-a3NC1 an- plemental Figure 2). Wild-type B6 mice tibody differences among groups at week 12, as assessed by one-way ANOVA followed by immunized with a3NC1 started develop- Bonferroni post tests for pairwise comparisons. (B) The left panel shows that the urinary ing progressive albuminuria 8–10 weeks albumin creatinine ratio (mean6SEM) time course is monitored in C57Bl6 wild-type mice 2 2 later (Figure 1B). By week 14, the urinary (○), Fcrn / mice (□), and hFCRN mice (◇) immunized with a3NC1 (n=5–8miceineach albumin creatinine ratio increased approx- group, from two separate experiments). Mice in the control group (△) are immunized with imately 100-fold, and hypoalbuminemia adjuvant alone (n=9). The right panel shows the urinary albumin creatinine ratio 6 fi developed (Figure 1C). Urinary albumin (mean SEM) at 14 weeks, when mice are euthanized. The signi cance of differences 2 2 among groups is assessed by one-way ANOVA followed by Bonferroni post tests for excretion in Fcrn / mice and FCRN- pairwise comparisons. (C) The left panel shows SDS-PAGE analysis of serum (0.5 ml/lane) humanized mice immunized with and urine samples (2 ml/lane) from CFA-immunized control mice (a) and a3NC1-immunized fi 2 2 a3NC1 was not signi cantly higher than wild-type mice (b), Fcrn / mice (c), and hFCRN mice (d) collected at week 14. The right in adjuvant-immunized control mice. No panel presents a densitometric analysis of the relative levels of albumin in mouse serum mice developed renal failure (Supplemental samples showing that a3NC1-immunized wild-type mice developed hypoalbuminemia. Figure 3). *P,0.05 by two-tailed t test versus CFA-immunized wild-type mice; **P,0.01; At 14 weeks after a3NC1 immuniza- ***P,0.001. ns, not significant; WT, wild type. tion, kidneys examined by light micros- copy showed mild glomerular pathology, a3NC1-immunized DBA/1 mice with IC deposits surrounded by an expanded with few crescents and relatively little in- comparable albuminuria.14,15 Electron mi- GBM and effacement of podocyte foot flammation (Figure 2A), similar to croscopy showed extensive subepithelial processes in a3NC1-immunized B6

J Am Soc Nephrol 25: 918–925, 2014 FcRn Promotes Immune Complex–Mediated GN 919 BRIEF COMMUNICATION www.jasn.org

2 2 mice, whereas Fcrn / mice had fewer subepithelial deposits (Figure 2B, Supple- mental Figure 4). Immunofluorescence staining showed granular capillary loop deposition of mouse IgG, exogenous an- tigen, C3, and C5b-9, more intense in 2 2 wild-typemicethaninFcrn / mice and FCRN-humanized mice (Figure 2, Ca–Cp, Supplemental Figure 5). A loss of nephrin staining, indicative of podocyte injury, oc- curred in a3NC1-immunized B6 mice but 2 2 not in Fcrn / mice or FCRN-humanized mice (Figure 2, Cq–Ct). Because B6 mice immunized with bovine GBM NC1 hexamers have normal kidney function and histology despite linear GBM deposition of IgG auto- antibodies binding to mouse a345(IV) collagen (Supplemental Figure 1), the question arises as to what causes pro- teinuria in a3NC1-immunized mice. Because the clinical presentation, mor- phology, and effector mechanisms de- pend on where ICs are localized in the capillary wall, we compared IgG distri- bution in a3NC1-immunized mice and mice injected with anti-a3NC1 antibod- ies modeling anti-GBM autoantibodies. Figure 2. FcRn deficiency reduces formation of pathogenic subepithelial ICs. (A) Light The distribution and relative abundance microscopic evaluation of kidneys from adjuvant-immunized control mice (a) and a3NC1- 2 2 of mouse IgG, as imaged by immunoper- immunized wild-type mice (b) and Fcrn / mice (c) revealed few pathogenic changes and oxidase immunoelectron microscopy the absence of glomerular inflammation (periodic acid–Schiff staining). (B) Transmission and stochastic optical reconstruction mi- electron microscopy shows normal GBM (arrow) and podocyte foot processes in control croscopy (STORM), a method for super- mice (a), extensive subepithelial electron dense deposits (arrowhead), thickened GBM, and podocyte foot process effacement in a3NC1-immunized wild-type mice (b), and fewer IC resolution fluorescence microscopy, were 2 2 deposits in the Fcrn / mice (c). (C) Immunofluorescence analysis of kidneys from adjuvant- concordant. In a3NC1-immunized mice, immunized control mice (a, e, i, m, and q) and a3NC1-immunized wild-type mice (b, f, j, n, 2 2 IgG deposition was abundant in the areas and r), FcRn / mice (c, g, k, o, and s), and hFCRN mice (d, h, l, p, and t) evaluate the of expanded GBM and especially in re- deposition of mouse IgG (a–d), exogenous a3NC1 antigen stained by mAb RH34 (e–h), gions corresponding to the subepithelial mouse C3c (i–l), C5b-9 (m–p), and nephrin staining (q–t) at 14 weeks. Wild-type mice ex- dense deposits seen by routine electron hibit linear-granular GBM deposition of mouse IgG and granular GBM deposition of ex- 2 2 microscopy. By contrast, in mice injected ogenous antigen, C3, and C5b-9, which are attenuated in Fcrn / mice and hFCRN mice with a3NC1-specificanti-GBMmAb,the and essentially absent in control mice. Compared with control mice, a3NC1-immunized 2/2 IgG was confined to an ultrastructurally wild-type mice but not Fcrn or hFCRN mice exhibit a loss of nephrin staining, indicative – normal GBM that lacked subepithelial of podocyte injury. WT, wild type; EM, electron microscopy, PAS, periodic acid Schiff. fi 3 3 3 deposits (Figure 3). Original magni cation, 400inA; 2850 in B; 200 in C. Subepithelial ICs, a hallmark of hu- man membranous nephropathy (MN), form when IgG antibodies bind to po- a3NC1-immunized mice may be ex- mice and DBA/1 mice eventually de- docyte antigens, such as phospholipase plained by exogenous a3NC1 deposited velop crescentic GN by 26 and 10 weeks, A2 receptor (PLA2R) and neutral endo- in glomeruli acting as a planted anti- respectively.10,14 The combination of peptidase (NEP), or to planted antigens, gen.19 Alternatively, anti-a3NC1 anti- subepithelial ICs and crescentic anti- such as cationic BSA.16–18 Subsequent bodies in complex with a3NC1 antigen GBM GN was most recently expansion of the GBM, complement ac- may act as surrogate antipodocyte anti- described in a series of eight patients tivation, and podocyte injury by C5b-9 bodies, because a3NC1-containing ICs with circulating anti-a3NC1 autoanti- cause proteinuria. Although it is unex- bind to podocytes.20 After four immuni- bodies but undetectable anti-PLA2R au- pected, formation of subepithelial ICs in zations with a3NC1 monomers, B6 toantibodies.21

920 Journal of the American Society of Nephrology J Am Soc Nephrol 25: 918–925, 2014 www.jasn.org BRIEF COMMUNICATION

proteinuria.24 The slow decline in pro- teinuria is problematic for patients al- ready suffering from complications of nephrotic syndrome, who would benefit from ancillary therapies that reduce pathogenic IgG antibodies more rapidly. This may be achieved by inhibiting FcRn. One implementation of thisconcept is therapy with high-dose intravenous Ig (HD-IVIG). HD-IVIG accelerates the degradation of IgG by saturating FcRn,25 one of the mechanisms that ex- plain the beneficial effects of HD-IVIG therapy in some autoimmune diseases.3 In pregnant women with circulating anti-NEP alloantibodies mediating ante- natal MN, treatment with HD-IVIG re- duces the titers of IgG alloantibodies by Figure 3. Localization of IgG by high-resolution imaging. The localization of mouse IgG in approximately 30% within 2–3 weeks.26 – glomerular capillary walls of wild-type mice immunized with a3NC1 (A, C E), or in- However, HD-IVIG is inefficient, be- – travenously injected with anti-mouse a3NC1 IgG mAb 8D1 (B, F H) is determined by cause large amounts of IgG (1–2g/kg) immunoperoxidase electron microscopy (A and B) and STORM imaging (C–H). In A, the cause relatively modest reductions in GBM is irregularly thickened, and abundant electron dense peroxidase reaction product is fi present in discontinuous, subepithelial patterns beneath broadly effaced podocyte foot pathogenic IgG titers. Speci cFcRnin- processes (arrows). In B, the peroxidase reaction product is diffusely present throughout hibitors recapitulate this activity of HD- the GBM (arrowhead), but less abundant compared with A. Electron dense deposits are IVIG more effectively at lower doses. By absent, and podocyte foot process architecture appears normal. (C–E) By STORM imaging, reducing pathogenic IgG levels, func- anti-agrin (blue) identifies both normal and thickened areas of the GBM, both of which tion-blocking anti-FcRn mAbs amelio- contain dense accumulations of mouse IgG throughout (red). The electron microscopy rate experimental myasthenia gravis in correlation in E shows GBM staining with respect to the podocytes and endothelial cells. rats,27 and engineered IgG “Abdegs” (F–H) IgG mAb 8D1 (red) is present in the GBM, which shows no evidence of thickening. CL, that bind with high affinity to FcRn ame- capillary lumen; EM, electron microscopy En, ;Po, podocyte. liorate arthritis transferred by K/BxN serum.28 To assess the translational potential of In contrast to wild-type B6 mice, maintaining higher serum levels of path- our findings, we asked whether pharma- 2 2 congenic Fcrn / mice and FCRN- ogenic IgG (Supplemental Figure 6). cologic blockade of human FcRn can humanized mice did not develop albu- However, we cannot formally exclude a reproduce the effects of genetic FcRn minuria after a3NC1 immunization. Their possible pathogenic role of podocyte deficiency. Tothis end, FCRN-humanized 2 2 resistance to proteinuria was associated FcRn, whose stimulation by ICs may in- and Fcrn / mice were passively immu- with lower serum titers of anti-a3NC1 duce maladaptive signaling.9 Future nized with human IgG containing anti- IgG antibodies and reduced glomerular studies in mice with podocyte-specific a3NC1 (Goodpasture) autoantibodies. deposition of IgG, antigen, C3, and C5b-9. ablation of FcRn would address this pos- To inhibit human FcRn, we used a lysine Because C5b-9 is an essential mediator of sibility. analog of SYN1436 (Figure 4A),29 apep- podocyte damage and proteinuria by Our findings identify FcRn as a po- tide that binds with subnanomolar affin- subepithelial ICs,22,23 reduced comple- tential target for therapeutic interven- ity to human FcRn, thus preventing IgG ment activation potentially explains the tion in IC-mediated glomerulardiseases, binding.30 In vivo,SYN1436reducesIgG attenuated glomerular pathology in typically treated with nonspecificimmu- levels in cynomolgus monkeys by 80%.30 FcRn-deficient mice. The resistance of nosuppressants that are toxic and some- Serum anti-a3NC1 autoantibodies in FCRN-humanized mice indicates that times ineffective. More specific therapies FCRN-humanized mice treated with FcRn promotes IC-mediated glomeru- include ablation of B cells by rituximab. In anti-FcRn peptide, but not with control lar disease due to its interaction with patients with idiopathic MN who respond peptide, sharply decreased to the same 2 2 IgG rather than albumin. We propose to rituximab therapy, serum levels of anti- levels as in Fcrn / mice (Figure 4B), that FcRn promotes the development PLA2RIgGautoantibodiesdeclineovera and were no longer detected after 4 days. of subepithelial ICs and subsequent period of many months, and their dis- In mice, human IgG elicits murine anti- glomerular injury primarily by appearance is followed by resolution of human IgG antibodies, forming ICs that

J Am Soc Nephrol 25: 918–925, 2014 FcRn Promotes Immune Complex–Mediated GN 921 BRIEF COMMUNICATION www.jasn.org

previous publication,29 and a control peptide (Figure4A),weregenerouslyprovidedby Biogen Idec Hemophilia (Waltham, MA).

Animal Experiments C57Bl/6J, B6.12931-Fcgrt,tm1Dcr./Dcr 2 2 (Fcrn / ), and B6.Cg-Fcgrt,tm1Dcr. Tg (FCGRT)32Dcr (FCRN-humanized) mice were purchased from The Jackson Laboratory 2 2 (Bar Harbor, ME). Fcrn / mice have lower IgG and serum albumin levels than wild-type mice. In contrast, FCRN-humanized mice have lower IgG levels but similar serum albumin lev- els compared with wild-type mice.13 Mice were housed in a specific pathogen-free facility with free access to food and water. All procedures were approved by the Institutional Animal Care and Use Committee and conducted in ac- cordance with the Guidelines for Animal Care and Use Program of Vanderbilt University. Male mice (aged 6–10 weeks) were used for experiments. For active immunizations, a3NC1 (30 mgin50mlPBS)wasemulsified in an equal volume of CFA (Difco, Detroit, MI), and then injected subcutaneously at two sites on the back. Mice were boosted three timeswiththesameamountofa3NC1 in Figure 4. Pharmacologic blockade of human FcRn accelerates the catabolism of human IFA (Sigma-Aldrich, St. Louis, MO). In control IgG autoantibodies in FCRN-humanized mice. (A) Structure of a peptide that binds with mice, the antigen was replaced by PBS. Mice high affinity to human FcRn, competitively inhibiting its interaction with human IgG (top). The control peptide (bottom) containing D-amino acids does not bind to human FcRn. Pen, were regularly checked for signs of disease. Sar, and NMeLeu denote penicillamine, sarcosine, and N-methyl-leucine, respectively. (B) Blood and spot urine were collected as de- Serum level of human IgG anti-a3NC1 antibodies in FCRN-humanized mice treated with scribed.15,33 Mice were euthanized at 14 weeks 2 2 anti-FcRn peptide (▪) or control peptide (●)andinFcrn / (▲)micesera(n=3 in each after immunization. Blood and kidneys were group) is analyzed by indirect ELISA in plates coated with a3NC1 (100 ng/well). Mouse sera collected for further analyses. are diluted 1:500. (C) Kidney deposition of human IgG (a and b) and mouse IgG (c and d) in For IgG imaging studies, DBA/1J mice FCRN-humanized mice treated with control peptide (a and c) or anti-FcRn peptide (b and d) immunized with a3NC1, as described,15 were fl is evaluated by direct immuno uorescence staining. Treatment with anti-FcRn peptide euthanized 8 weeks later. Tomodel the glomer- prevents the glomerular deposition of ICs containing human IgG. ular deposition of anti-GBM IgG autoantibod- ies, DBA/1 mice were injected intravenously can deposit in glomeruli, as shown in ac- mediates the trans-placental transfer of with mouse IgG anti-a3NC1 mAb 8D1 (200 tive serum sickness models. Glomerular IgG from mother to the fetus, FcRn inhi- mg in sterile PBS) and euthanized the next day. For passive immunization experiments, deposition of ICs containing human IgG bition may be particularly attractive for 2/2 was abolished in mice treated with anti- preventing antenatal MN caused by ma- FCRN-humanized and FcRn mice were in- FcRn peptide, but not with control pep- ternal anti-NEP alloantibodies. jected intravenously in the tail vein with 1 mg tide (Figure 4C). Linear GBM deposition of human IgG containing anti-a3NC1 autoan- of human anti-GBM IgG was not ob- CONCISE METHODS tibodies at day 0. FCRN-humanized mice were served, because the epitopes recognized treated with anti-FCRN peptide or control peptide (10 mg/kg daily injected in the tail by Goodpasture autoantibodies are com- Materials vein) on days 1–4. After 14 days, mice were pletely inaccessible in the mouse GBM.31 Recombinant NC1 monomers of human euthanized and the deposition of human IgG These results provide proof of concept a3NC1 were expressed in human embryonic in kidneys was analyzed. that therapies targeting human FcRn ef- kidney 293 cells and purified as described.32 fectively lower serum levels of pathogenic Human IgG was purified from plasma ex- Evaluation of Kidney Function and human IgG autoantibodies, which could change fluid of a patient with Goodpasture Renal Histopathology be beneficial in patients with IgG-medi- disease.31 A peptide antagonist of human Urine albumin, urine creatinine, and BUN ated kidney diseases. Because FcRn also FcRn, described in detail as “peptide 3” in a were measured using commercial kits as

922 Journal of the American Society of Nephrology J Am Soc Nephrol 25: 918–925, 2014 www.jasn.org BRIEF COMMUNICATION described.15,33 Albuminuria was expressed as Sections were refixed for 20 minutes at with 30% sucrose in PBS. Tissue was covered the urinary albumin to creatinine ratio. The room temperature with 4% PFA, followed with Tissue Tek O.C.T. Compound (Electron proteins in serum and urine samples were by three washes in PBS, quenched using 50 Microscopy Sciences) and snap-frozen in separated by SDS-PAGE under nonreducing mM glycine in PBS, and blocked overnight at isopentane chilled in a dry ice-acetone bath. conditions and stained with Coomassie Bril- 4°C using 3% BSA in PBS. Primary antibody Cryostat sections, 30 mm thick, were cut at liant Blue. For light microscopy, portions of diluted in 3% BSA-PBS was applied on the 220°C, picked up on Thermanox coverslips mouse kidneys were fixed in 10% buffered sections and incubated overnight at 4°C. The (Nalgene Nunc International, Rochester, formalin, dehydrated through graded etha- next day, sections were washed 3320 minutes NY), and air dried at room temperature nols, embedded in paraffin, and 2 mm–thick with PBS at room temperature and then in- for 1 hour. Coverslips were blocked with sections were stained with periodic acid– cubated with the secondary antibodies diluted 0.5 M ammonium chloride in PBS for Schiff. To evaluate ultrastructural changes, in 3% BSA-PBS at room temperature for 2 30 minutes at room temperature in a humid transmission electron microscopy was per- hours. After washing the samples 3320 minutes chamber, and then incubated overnight at 4° formed as described.15 with PBS at room temperature, immunolabeled Cwithrabbitanti-mouseIgGconjugatedto sections were postfixed using 3% PFA and horseradish peroxidase (1:200 in PBS; Direct and Indirect 0.1% glutaraldehyde (Electron Microscopy Sigma-Aldrich). The next day, coverslips Immunofluorescence Sciences) in PBS and prepared for STORM were washed three times with PBS and Portions of snap-frozen mouse kidneys or imaging. For orientation, tissues were stained then refixed with 2% glutaraldehyde in 0.1 lung embedded in Tissue Tek O.C.T. Com- with hamster anti-agrin C terminus as a pri- M sodium phosphate buffer (pH 7.3) for 30 pound (Electron Microscopy Sciences, Fort mary antibody.34 Anti-total mouse IgG as well minutes at room temperature. Coverslips Washington, PA) were cryosectioned (5 mm), as donkey anti-hamster secondary antibodies were washed three times with buffer, and fixed in acetone for 10 minutes at 220°C, and (Jackson ImmunoResearch, West Grove, PA) the peroxidase histochemistry reaction was blocked with 1% BSA. For direct immuno- were using for STORM imaging. The secondary carried out for 1 hour at room temperature. fluorescence, frozen sections were stained antibodies were custom conjugated to After washing the sections three times with with Alexa Fluor 488–conjugated goat anti- Alexa647 reporter dye and either Alexa405 buffer, coverslips were postfixed in 2% os- mouse IgG (Invitrogen, Carlsbad, CA) or or Cy3 activator dyes. mium tetroxide in 0.1 M sodium phosphate FITC-conjugated goat anti-mouse C3c (Nordic STORM image acquisition was performed bufferinahumidchamberfor2hoursat Immunology, Tilburg, Netherlands). Primary using a custom-made setup as described.35,36 room temperature. Slips were then dehydra- antibodies were rat IgG mAbs RH34, specific Coverglass with the sections on it was inverted tedthroughagradedseriesofincreasing for human a3NC1 (generously provided by onto a slide containing a drop of imaging concentrations of ethanol and then 100% Dr. Yoshikazu Sado), rabbit anti–C5b-9 (Ab- buffer containing mercaptoethylamine along propylene oxide. Samples were then infiltra- cam, Cambridge, MA), and guinea pig anti- with an oxygen scavenger system, and cover- ted overnight in a 1:1 mixture of propylene nephrin (Progen, Heidelberg, Germany). glass edges were sealed with nail polish. Ap- oxide and Polybed 812 embedding resin, in- Secondary antibodies were Alexa Fluor proximately 10,000 images per channel were cluding DMP-30 (Polysciences, Warrington, 488–conjugated goat anti-rat and anti-rabbit captured and analyzed using custom software. PA). The next day, pieces of coverslips were IgG (Invitrogen). Sections were examined For quick-freeze deep-etch electron micros- flat-embedded in molds containing fresh under Nikon Eclipse E800 epifluorescence copy, the nail polish from the coverglass was Polybed 812 and DMP-30 and polymerized microscope. Photomicrographs were recor- removed after STORM imaging and the tissue overnight at 60°C. Ultrathin sections were ded with a charge-coupled device digital sections were fixed in 2% glutaraldehyde. Areas stained briefly with lead citrate and then camera, using the same exposure settings of the coverglass containing STORM-imaged viewed in a JEOL JEM-1400 transmission for each primary antibody. sections were cut, rinsed in dH2O, and deep electron microscope. frozen, and then etched and replicated with STORM Imaging approximately 2 nm platinum deposition. Af- Analysis of Circulating IgG Kidneys were taken from mice after perfusion ter dissolving the glass in concentrated hydro- Antibodies with 4% paraformaldehyde (PFA). After mul- fluoric acid, replicas were rinsed with distilled Circulating mouse IgG or human IgG anti- tiple washes, kidney pieces were impregnated water, picked up on Luxel grids (Friday Harbor, a3NC1 antibodies were analyzed by ELISA overnight at 4°C in a cryoprotectant solution WA), and photographed on a JEOL 1400 mi- in plates were coated overnight with a3NC1 of 2.3 M sucrose and 10% polyvinylpyrroli- croscope with attached AMT digital camera. (100 ng/well) and blocked with 1% BSA. done in 0.1 M PIPES (pH 7.2). Cryoprotected The glomeruli imaged by electron microscopy Mouse sera were diluted as indicated. Second- tissues were mounted on a metal sectioning were matched with the corresponding STORM ary antibodies were alkaline phosphatase– pin and frozen by immersion in liquid nitrogen. images, and the two images were superim- conjugated goat anti-mouse IgG or anti-human Sections were collected on carbon coating and posed using Adobe Photoshop. IgG (Rockland Immunochemicals, Gilberts- fresh glow discharge coverglass. Frozen tissues ville, PA) and horseradish peroxidase– were sectioned at approximately 200 nm in Immunoelectron Microscopy conjugated goat anti-mouse IgG1, IgG2b, and thickness on a Leica EM-FC6 ultracryomi- Strips of kidney cortex from mice that were IgG2c (Bethyl Laboratories, Montgomery, crotome equipped with a diamond knife. perfusion-fixedwith4%PFAwerecryoprotected TX).

J Am Soc Nephrol 25: 918–925, 2014 FcRn Promotes Immune Complex–Mediated GN 923 BRIEF COMMUNICATION www.jasn.org

Statistical Analyses 4. Sesarman A, Sitaru AG, Olaru F, Zillikens D, PM: Antenatal membranous glomerulonephritis Data are shown as the mean6SD. Statistical Sitaru C: Neonatal Fc receptor deficiency due to anti-neutral endopeptidase antibodies. N – analyses were performed using GraphPad protects from tissue injury in experimental Engl J Med 346: 2053 2060, 2002 epidermolysis bullosa acquisita. J Mol Med 17. Beck LH Jr, Bonegio RG, Lambeau G, Beck Prism software (version 5.01). The significance (Berl) 86: 951–959, 2008 DM, Powell DW, Cummins TD, Klein JB, of differences between groups was evaluated by 5. Haymann JP, Levraud JP, Bouet S, Kappes V, Salant DJ: M-type phospholipase A2 re- the t test (for two groups) or by one-way Hagège J, Nguyen G, Xu Y, Rondeau E, Sraer ceptor as target antigen in idiopathic mem- ANOVA,followed by post hoc tests for pairwise JD: Characterization and localization of the branous nephropathy. N Engl J Med 361: – comparisons. A P value,0.05 was considered neonatal Fc receptor in adult human kidney. J 11 21, 2009 Am Soc Nephrol 11: 632–639, 2000 18. Debiec H, Lefeu F, Kemper MJ, Niaudet P, to be statistically significant. 6. Sarav M, Wang Y, Hack BK, Chang A, Jensen Deschênes G, Remuzzi G, Ulinski T, Ronco P: M, Bao L, Quigg RJ: Renal FcRn reclaims al- Early-childhood membranous nephropathy bumin but facilitates elimination of IgG. JAm due to cationic bovine serum albumin. ACKNOWLEDGMENTS Soc Nephrol 20: 1941–1952, 2009 N Engl J Med 364: 2101–2110, 2011 7. Kobayashi N, Suzuki Y, Tsuge T, Okumura K, 19.BorzaDB,ZhangJJ,BeckLHJr,Meyer- We thank Ms. Jeanette Cunningham for Ra C, Tomino Y: FcRn-mediated transcytosis Schwesinger C, Luo W: Mouse models of performing the electron microscopy and Dr. of in human renal proxi- membranous nephropathy: The road less mal tubular epithelial cells. Am J Physiol Re- travelled by. Am J Clin Exp Immunol 2: 135– Shreeram Akilesh (University of Washington) nal Physiol 282: F358–F365, 2002 145, 2013 for helpful comments and discussions. 8. Akilesh S, Huber TB, Wu H, Wang G, 20. Borza CM, Borza DB, Pedchenko V, Saleem This work was supported by the National Hartleben B, Kopp JB, Miner JH, Roopenian MA,MathiesonPW,SadoY,HudsonHM, Institutes of Health (NIH) (Grants R01- DC, Unanue ER, Shaw AS: Podocytes use Pozzi A, Saus J, Abrahamson DR, Zent R, DK080799 to D.B.B. and R01-DK078314 FcRn to clear IgG from the glomerular base- Hudson BG: Human podocytes adhere to the ment membrane. Proc Natl Acad Sci U S A KRGDS motif of the alpha3alpha4alpha5 to J.H.M.), Satellite Healthcare (Norman 105: 967–972, 2008 collagen IV network. J Am Soc Nephrol 19: S. Coplon Extramural Grant to D.B.B.), and 9. Gan H, Feng S, Wu H, Sun Y, Hu R, Zhao Z, 677–684, 2008 the American Heart Association (Grant-in- Zhang Z: Neonatal Fc receptor stimulation 21.JiaXY,HuSY,ChenJL,QuZ,LiuG,CuiZ, Aid 12GRNT11480005to D.B.B.). F.O. has induces ubiquitin c-terminal hydrolase-1 Zhao MH: The clinical and immunological been supported, in part, by a postdoctoral overexpression in podocytes through acti- features of patients with combined anti- vation of p38 mitogen-activated ki- glomerular basement membrane disease fellowship (11POST7300008) from theAmerican nase. Hum Pathol 43: 1482–1490, 2012 and membranous nephropathy [published Heart Association South-East Affiliate. Electron 10. Hopfer H, Maron R, Butzmann U, Helmchen online ahead of print September 18, 2013]. microscopy was performed by the Washington U, Weiner HL, Kalluri R: The importance of Kidney Int doi:10.1038/ki.2013.364 University O’Brien Center for Kidney Disease cell-mediated in the course and 22. Cybulsky AV, Rennke HG, Feintzeig ID, Research supported by NIH Grant P30- severity of autoimmune anti-glomerular base- Salant DJ: Complement-induced glomerular ment membrane disease in mice. FASEB J 17: epithelial cell injury. Role of the membrane DK079333. The Electron Microscope Research 860–868, 2003 attack complex in rat membranous ne- Laboratory at the University of Kansas Medical 11. Chaudhury C, Mehnaz S, Robinson JM, phropathy. JClinInvest77: 1096–1107, Center is supported, in part, by NIH Grant P20- Hayton WL, Pearl DK, Roopenian DC, 1986 GM104936. The JEOL JEM transmission elec- Anderson CL: The major histocompatibility 23. Cunningham PN, Quigg RJ: Contrasting tron microscope used for immunoelectron complex-related Fc receptor for IgG (FcRn) roles of complement activation and its reg- binds albumin and prolongs its lifespan. J ulation in membranous nephropathy. JAm microscopy was purchased with funds from Exp Med 197: 315–322, 2003 Soc Nephrol 16: 1214–1222, 2005 NIH Grant S10-RR027564. 12. Koltun M, Nikolovski J, Strong K, Nikolic- 24. Beck LH Jr, Fervenza FC, Beck DM, Bonegio Paterson D, Comper WD: Mechanism of hy- RG, Malik FA, Erickson SB, Cosio FG, Cattran poalbuminemia in rodents. Am J Physiol DC, Salant DJ: Rituximab-induced depletion DISCLOSURES Heart Circ Physiol 288: H1604–H1610, 2005 of anti-PLA2R autoantibodies predicts re- 13. Stein C, Kling L, Proetzel G, Roopenian DC, sponse in membranous nephropathy. JAm None. de Angelis MH, Wolf E, Rathkolb B: Clinical Soc Nephrol 22: 1543–1550, 2011 chemistry of human FcRn transgenic mice. 25. Bleeker WK, Teeling JL, Hack CE: Acceler- – REFERENCES Mamm Genome 23: 259 269, 2012 ated autoantibody clearance by intravenous 14. Hopfer H, Holzer J, Hünemörder S, Paust HJ, immunoglobulin therapy: Studies in experi- Sachs M, Meyer-Schwesinger C, Turner JE, mental models to determine the magnitude 1. Roopenian DC, Akilesh S: FcRn: The neonatal Panzer U, Mittrücker HW: Characterization of and time course of the effect. Blood 98: Fc receptor comes of age. Nat Rev Immunol the renal CD4+ T-cell response in experi- 3136–3142, 2001 7: 715–725, 2007 mental autoimmune glomerulonephritis. 26. Nortier JL, Debiec H, Tournay Y, Mougenot 2. Akilesh S, Petkova S, Sproule TJ, Shaffer DJ, Kidney Int 82: 60–71, 2012 B, Nöel JC, Deschodt-Lanckman MM, Christianson GJ, Roopenian D: The MHC 15. Zhang JJ, Malekpour M, Luo W, Ge L, Olaru F, Janssen F, Ronco P: Neonatal disease in class I-like Fc receptor promotes humorally Wang XP, Bah M, Sado Y, Heidet L, Kleinau S, neutral endopeptidase alloimmunization: mediated autoimmune disease. J Clin Invest Fogo AB, Borza DB: Murine membranous ne- Lessons for immunological monitoring. Pe- 113: 1328–1333, 2004 phropathy: Immunization with a3(IV) collagen diatr Nephrol 21: 1399–1405, 2006 3. Li N, Zhao M, Hilario-Vargas J, Prisayanh P, fragment induces subepithelial immune com- 27. Liu L, Garcia AM, Santoro H, Zhang Y, Warren S, Diaz LA, Roopenian DC, Liu Z: plexes and FcgR-independent nephrotic syn- McDonnell K, Dumont J, Bitonti A: Amelio- Complete FcRn dependence for intravenous drome. JImmunol188: 3268–3277, 2012 ration of experimental autoimmune myasthenia Ig therapy in autoimmune skin blistering 16. Debiec H, Guigonis V, Mougenot B, Decobert F, gravis in rats by neonatal FcR blockade. JIm- diseases. J Clin Invest 115: 3440–3450, 2005 Haymann JP, Bensman A, Deschênes G, Ronco munol 178: 5390–5398, 2007

924 Journal of the American Society of Nephrology J Am Soc Nephrol 25: 918–925, 2014 www.jasn.org BRIEF COMMUNICATION

28. Patel DA, Puig-Canto A, Challa DK, Perez bodies induce murine glomerulonephritis: Holzman LB, Burgess RW, Miner JH: Disrup- Montoyo H, Ober RJ, Ward ES: Neonatal Fc Protection by quinary crosslinks locking cryptic tion of glomerular basement membrane receptor blockade by Fc engineering ameli- a3(IV) collagen autoepitopes in vivo. J Immunol charge through podocyte-specific mutation orates arthritis in a murine model. J Immunol 185: 3520–3528, 2010 of agrin does not alter glomerular permse- 187: 1015–1022, 2011 32. Kang JS, Colon S, Hellmark T, Sado Y, lectivity. Am J Pathol 171: 139–152, 2007 29. Mezo AR, Sridhar V, Badger J, Sakorafas P, Hudson BG, Borza DB: Identification of 35. Dani A, Huang B, Bergan J, Dulac C, Zhuang X: Nienaber V: X-ray crystal structures of mo- noncollagenous sites encoding specificin- Superresolution imaging of chemical synapses nomeric and dimeric peptide inhibitors in teractions and quaternary assembly of alpha in the brain. Neuron 68: 843–856, 2010 complex with the human neonatal Fc re- 3 alpha 4 alpha 5(IV) collagen: Implications 36. Suleiman H, Zhang L, Roth R, Heuser JE, ceptor, FcRn. JBiolChem285: 27694– for Alport gene therapy. J Biol Chem 283: Miner JH, Shaw AS, Dani A: Nanoscale pro- 27701, 2010 35070–35077, 2008 tein architecture of the kidney glomerular 30. Mezo AR, McDonnell KA, Hehir CA, Low SC, 33. Olaru F, Wang XP, Luo W, Ge L, Miner JH, basement membrane. Elife 2: e01149, 2013 Palombella VJ, Stattel JM, Kamphaus GD, Kleinau S, Geiger XJ, Wasiluk A, Heidet L, Fraley C, Zhang Y, Dumont JA, Bitonti AJ: Kitching AR, Borza DB: Proteolysis breaks tol- Reduction of IgG in nonhuman primates by a erance toward intact a345(IV) collagen, eliciting Present address: Dr. Adam R. Mezo, Eli Lilly and peptide antagonist of the neonatal Fc re- novel anti-glomerular basement membrane Company, Indianapolis, Indiana. ceptor FcRn. Proc Natl Acad Sci U S A 105: autoantibodies specificfora345NC1 hexamers. – – 2337 2342, 2008 J Immunol 190: 1424 1432, 2013 This article contains supplemental material online at 31. LuoW,WangXP,KashtanCE,BorzaDB:Alport 34. Harvey SJ, Jarad G, Cunningham J, Rops AL, http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ alloantibodies but not Goodpasture autoanti- van der Vlag J, Berden JH, Moeller MJ, ASN.2013050498/-/DCSupplemental.

J Am Soc Nephrol 25: 918–925, 2014 FcRn Promotes Immune Complex–Mediated GN 925 Supplementary Information:

Materials and Methods:

Susceptibility of C57Bl/6 mice to antibody‐mediated glomerular injury induced by immunization with NC1 domains of GBM collagen IV and selection of a mouse model. Commercially available FcRn‐/‐ mice were on the C57Bl/6 (B6) genetic background, which is among the most resistant to immune mediated nephritis.1 Although B6 mice lacking the inhibitory IgG receptor FcRIIB are highly susceptible to crescentic glomerulonephritis, for instance after passive immunization with human IgG anti‐GBM alloantibodies from patients with Alport post‐transplant nephritis, wild type B6 mice are resistant.2 To avoid the need for time‐consuming back‐crosses, we sought a mouse model in which antibody‐ mediated glomerular injury could be induced in wild type B6 mice. Previously, antibody‐mediated glomerulonephritis has been reported in B6 mice immunized four times with recombinant NC1 monomers of human α3(IV) collagen (3NC1),3 as well as mice immunized twice with 3‐5NC1 dimers from bovine GBM.4 In preliminary experiments, we compared the susceptibility of B6 mice to develop kidney injury after active immunization with recombinant human α3NC1 monomers and total NC1 hexamers from bovine GBM (containing 345NC1 hexamers). C57Bl/6J mice were purchased from The Jackson Laboratory and maintained in specific‐pathogen free facilities with access to water and food. For active immunization experiments, 4‐5 mice in each group were immunized at 8‐10 weeks of age with 3NC1 monomers (25 g) and NC1 hexamers from bovine GBM (100 g), prepared as previously described.5 The antigens were emulsified in complete Freund’s adjuvant (Sigma, St. Louis, MO) and injected subcutaneously at two sites. Mice received one (for both antigens) or three (for 3NC1) booster immunizations with antigen in incomplete Freund’s adjuvant. Measurements of blood urea nitrogen and urinary albumin‐creatinine ratio, and analysis of kidney histopathology by light microscopy and immunofluorescence microscopy were performed as described in the main text. All mouse studies were performed in accordance with the principles for humane treatment of lab animals and were approved by the IACUCs at Vanderbilt University.

References: 1. Xie C, Sharma R, Wang H, Zhou XJ, Mohan C: Strain distribution pattern of susceptibility to immune‐ mediated nephritis. J Immunol 172: 5047‐5055, 2004. 2. Luo W, Wang XP, Kashtan CE, Borza DB: Alport alloantibodies but not Goodpasture autoantibodies induce murine glomerulonephritis: protection by quinary crosslinks locking cryptic alpha3(IV) collagen autoepitopes in vivo. J Immunol 185: 3520‐3528, 2010. 3. Hopfer H, Maron R, Butzmann U, Helmchen U, Weiner HL, Kalluri R: The importance of cell‐mediated immunity in the course and severity of autoimmune anti‐glomerular basement membrane disease in mice. Faseb J 17: 860‐868, 2003. 4. Kitching AR, Turner AL, Semple T, Li M, Edgtton KL, Wilson GR, Timoshanko JR, Hudson BG, Holdsworth SR: Experimental autoimmune anti‐glomerular basement membrane glomerulonephritis: a protective role for IFN‐gamma. J Am Soc Nephrol 15: 1764‐1774, 2004. 5. Boutaud A, Borza DB, Bondar O, Gunwar S, Netzer KO, Singh N, Ninomiya Y, Sado Y, Noelken ME, Hudson BG: Type IV collagen of the glomerular basement membrane: Evidence that the chain specificity of network assembly is encoded by the noncollagenous NC1 domains. J Biol Chem 275: 30716‐30724, 2000.

1 Supplementary Figures:

Supplemental Figure 1. Susceptibility of wild type B6 mice to kidney disease induced by immunization with 3NC1 monomers and bovine GBM NC1 hexamers. A. At 10 weeks, urinary albumin creatinine ratio (ACR) was significantly increased in mice immunized twice (2x) or four times (4x) with α3NC1 monomers, but not in mice immunized with bovine GBM NC1 hexamers or CFA alone (control). B. Blood urea nitrogen levels remained in the normal range in all groups of mice. C. Kidneys from mice sacrificed at 10 weeks after the first immunization were observed by light microscopy and immunofluorescence microscopy. Light microscopy revealed minimal glomerular injury in mice immunized with CFA (a), bovine GBM NC1 hexamers (b) or α3NC1 monomers (c) (periodic acid‐Schiff staining, magnification 400x). Immunofluorescence showed linear GBM deposition of mouse IgG in mice immunized with bovine GBM NC1 hexamers (e) and linear‐granular GBM staining for IgG in mice immunized with 3NC1 (f, g). Weak mesangial IgG deposition was observed in CFA control mice (d). Granular C3c deposition along the GBM was only observed in mice immunized with 3NC1 monomers, more intense after four immunizations (k) than after two immunizations (j). GBM deposition of C3c was absent in mice immunized with CFA (h) and bovine GBM NC1 hexamers (i). Magnification 400x. Acid‐eluted antibodies from the kidneys of mice immunized with bovine GBM contained mouse IgG1, IgG2a and IgG2b autoantibodies binding to 345NC1 hexamers from mouse GBM (not shown). Based on these preliminary results, in the experiments described in the main text, wild type B6 mice were immunized with 3NC1 four times.

2

Supplemental Figure 2. Serum levels of mouse IgG1, IgG2b and IgG2c anti‐α3NC1 antibodies in 3NC1‐immunized mice. Circulating mIgG1 (A), mIgG2b (B) and mIgG2c (C) autoAbs from C57Bl/6 wild type mice (circles), FcRn ‐/‐ mice (squares), hFCRN mice (diamonds) and control CFA group (triangles) were assayed by indirect ELISA in plates coated with α3NC1 (100 ng/well). Mouse sera were diluted 1/5000.

3

Supplemental Figure 3. Normal levels of blood urea nitrogen (BUN) in 3NC1‐immunized mice. Renal function was evaluated by measuring blood urea nitrogen levels (mg/dl) in 3NC1‐immunized C57Bl6 wild type mice (circles), FcRn ‐/‐ mice (squares), hFCRN mice (diamonds) and control CFA‐ immunized mice (triangles), which remained in the normal range for the duration of the experiment.

4

** 0.8 *** ** 0.6

0.4 m of GBM m of 

IC deposits IC 0.2 per 0.0 e FA C typ ld FcRn-/- wi

2.5 ** ** 2.0 ns 1.5

m of GBM m of 1.0  0.5 per Foot processesFoot 0.0

e -/- FA yp n C t R ild Fc w

Supplemental Figure 4. Quantitation of electron‐dense subepithelial IC deposits and podocyte foot process effacement in wild type mice and Fcrn‐/‐ mice immunized with α3NC1. The number of subepithelial electron‐dense deposits per unit of GBM length was about 1.8‐fold greater in α3NC1‐ immunized wild type mice than in Fcrn‐/‐ mice (top). Compared to control mice immunized with CFA alone, the number of podocyte foot processes per unit of GBM length was decreased by about 30% in wild type mice immunized with α3NC1 as well as in Fcrn‐/‐ mice immunized with α3NC1 (bottom). To quantitate the ultrastructural features, the numbers of foot processes and electron dense deposits per length of GBM in capillary loops from at least three glomeruli per mouse were determined. The graphs depict the means and SEM. The significance of differences among groups was assessed by one‐way ANOVA followed by Bonferroni post tests for pairwise comparisons (ns, not significant; ** p< 0.01; ***p< 0.001).

5

Supplemental Figure 5. Glomerular deposition of C3c, exogenous antigen forming immune complexes and nephrin staining in 3NC1‐immunized wild type and FcRn‐deficient mice. ImageJ software was used to determine glomerular mean fluorescence intensity of C3c deposited in 3NC1‐ immunized wild type and FcRn‐deficient mice (A). To evaluate the glomerular deposition of exogenous 3NC1 antigen, mean fluorescence intensity was determined on kidney sections stained by mAb H31 using native conditions (B). Wild type mice immunized with 3NC1 monomers show glomerular nephrin loss compared to CFA controls, but not FcRn‐deficient mice (C). Means and SEM are shown for approximately 15 gloms (n=3‐5 mice in each group). The significance of differences among groups was assessed by one‐way ANOVA followed by Bonferroni post tests for pairwise comparisons (ns, not significant; * p<0.05; ** p< 0.01).

6

Supplemental Figure 6. The role of FcRn in immune complex‐mediated glomerular disease. Left: Recycling of IgG by endothelial and bone marrow (BM)‐derived cells maintains high serum IgG levels. IgG (as well as albumin, not shown) taken up into cells by fluid‐phase pinocytosis binds to FcRn at pH 6.0‐6.5 in endosomes and is then recycled to the plasma membrane, where it is released upon exposure to pH 7.4. In contrast, IgG not bound to FcRn enters a default lysosomal pathway and is degraded. In IgG‐mediated kidney disease, FcRn maintains high serum level of pathogenic IgG antibodies, promoting the development of glomerular ICs which can activate various effector pathways, damaging the glomerular filtration barrier and causing albuminuria. Right: In the absence of FcRn, IgG catabolism is accelerated, decreasing the serum levels of pathogenic IgG antibodies, which limits formation of subepithelial ICs and associated glomerular pathology.

7