Both Complement and IgG Fc Receptors Are Required forDevelopment of Attenuated Antiglomerular BasementMembrane Nephritis in Mice1

Marielle A. Otten,2* Tom W. L. Groeneveld,2* Roelof Flierman,* Maria Pia Rastaldi,†

Leendert A. Trouw,‡ Maria C. Faber-Krol,* Annemieke Visser,* Maria C. Essers,*Jill Claassens,§ J. Sjef Verbeek,§ Cees van Kooten,* Anja Roos,*¶ and Mohamed R. Daha3*

To elucidate the mechanisms of glomerulonephritis, including Goodpasture’s syndrome, mouse models are used that use heter-ologous Abs against the glomerular basement membrane (GBM) with or without preimmunization with foreign IgG from the samespecies. These studies have revealed the requirement of either Fc�R or complement, depending on the experimental model used.In this study, we provide evidence that both Fc�R and complement are obligatory for a full-blown inflammation in a novelattenuated passive model of anti-GBM disease. We demonstrate that administration of subnephritogenic doses of rabbit anti-GBMAbs followed by a fixed dose of mouse mAbs to rabbit IgG, allowing timing and dosing for the induction of glomerulone-phritis, resulted in reproducible complement activation via the classical pathway of complement and albuminuria in wild-type mice. Because albuminuria was absent in FcR-�-chain�/� mice and reduced in C3�/� mice, a role for both Fc�R andcomplement is postulated. Because C1q�/� and C4�/� mice lacking a functional classical and lectin pathway did developalbuminuria, we suggest involvement of the alternative pathway of complement. Anti-GBM glomerulonephritis occursacutely following the administration of mouse anti-rabbit IgG, and proceeds in a chronic fashion dependent on both Fc�Rand complement. This novel attenuated model allows elucidating the relative contribution of different mediator systems ofthe immune system to the development of renal injury, and also provides a platform for the assessment of different treatmentprotocols and evaluation of drugs that ultimately may be beneficial for the treatment of anti-GBM mediatedglomerulonephritides. The Journal of Immunology, 2009, 183: 3980 –3988.

A ntiglomerular basement membrane (GBM)4-mediatedglomerulonephritis plays an essential role in autoim-mune diseases such as Goodpasture’s syndrome. Al-

though the pathological processes that damage the kidney follow-ing deposition of anti-GBM Abs are incompletely understood,Goodpasture’s syndrome, a life-threatening renal disease, is char-acterized by rapidly progressive glomerulonephritis and a lineardeposition of Abs along the GBM (1). Binding of these autoanti-bodies to the GBM (anti-GBM Ab) leads to autoimmune injurycharacterized by strong complement activation (as evidenced bythe deposition of C3), leukocyte infiltration and proteinuria. Thiscan ultimately lead to crescent formation, scarring, and loss of

renal function (2). Until now, the only treatment available forGoodpasture’s syndrome is to slow down disease progression us-ing corticosteroids and plasmapheresis (1). However, no curativetherapy is available. Therefore, development of controlled exper-imental animal models is crucial.

To better understand the mechanism of anti-GBM disease, sev-eral animal models have been developed. Earlier experimentalstudies have mainly used rat models to investigate the pathology ofanti-GBM disease (3). However, to profit from animals that havegenetic deficiencies, a number of investigators switched to mice,which resulted into the development of two distinct experimentalmodels for anti-GBM disease. In the direct model, heterologousAbs against mouse GBM are injected (4, 5). However, to inducedisease, large amounts of Abs are required. Therefore, an addi-tional model was used in which mice were first preimmunized withheterologous IgG, followed a week later with injection of heter-ologous anti-GBM Ab. In this so-called accelerated model, flam-boyant inflammation takes place (6, 7). However, because thedegree of renal inflammation in the latter model is strongly depen-dent on the initial immune response against the IgG used for im-munization, and because this response is variable between indi-vidual mice, this model is difficult to control. Furthermore, due tothe injection of high amounts of Ab in the direct model, and dueto the preimmunization in the accelerated model, both modelscomprise two phases, a first, heterologous, phase which mimics theAb-mediated disease, and a second, autologous, phase, in whichmice have developed an immune response against the foreign Abs(8–10). As this autologous phase is independent of Ab-mediatedeffector mechanisms, these models are less adequate to studychronic effects of anti-GBM disease. Furthermore, these direct and

*Department of Nephrology, Leiden University Medical Center, Leiden, TheNetherlands; †Renal Research Laboratory, Fondazione Istituto Di Ricovero e Curaa Carattere Scientifico Ospedale Maggiore Policlinico and Fondazione D’Amicoper la Ricerca sulle Malattie Renali, Milano, Italy; and ‡Department of Rheuma-tology, §Department of Human Genetics, and ¶Department of Clinical Chemistry,LUMC, Leiden, The Netherlands

Received for publication April 23, 2009. Accepted for publication July 6, 2009.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This work was supported by NWO/Zon-MW (901-12-095) and the European Union(LSHB-CT-2004-504761).2 M.A.O. and T.W.L.G. contributed equally to this study.3 Address correspondence and reprint requests to Prof. Dr. M. R. Daha, Departmentof Nephrology, LUMC, Albinusdreef 2, 2333 ZA, Leiden, The Netherlands. E-mailaddress: m.r.daha@lumc.nl4 Abbreviations used in this paper: GBM, glomerular basement membrane; PAS,periodic acid-Schiff; WT, wild type; DIG, digitonin.

Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00

The Journal of Immunology

www.jimmunol.org/cgi/doi/10.4049/jimmunol.0901301

Published August 26, 2009, doi:10.4049/jimmunol.0901301

accelerated models differ in their mechanism of Ab-induced renalinjury.

Autoantibodies might provoke damage via distinct effectormechanisms. Firstly, Abs might cause injury due to the activationof the complement system (11). The complement system consistsof three pathways: the classical, the lectin, and the alternative path-way, which all merge at the level of C3. Activation of the com-plement system can lead to inflammatory processes, such as pro-duction of chemotactic fragments resulting in the attraction andactivation of leukocytes, formation of the lytic membrane attackcomplex (C5b-9), and release of inflammatory mediators (12),which together determine the degree of glomerular injury (13–15).Classically, IgG Ab have been shown to be able to activate theclassical pathway of complement. However, in a growing list ofAb-induced autoimmune diseases, such as Ab-mediated arthri-tis (16), subepidermal blistering disease (17), anti-phospholipidsyndrome (12), ischemia/reperfusion injury (18), and cryoglob-ulin-induced immune-complex glomerulonephritis (19), alter-native pathway involvement has been implicated. Secondly,Abs can also induce injury through activation of IgG Fc recep-tors (Fc�R) (20). IgG binding to Fc�R on immune cells canactivate immune effector functions including Ab-dependent cel-lular cytotoxicity, phagocytosis, and release of inflammatorymediators and reactive oxygen species by leukocytes whose ca-pacity to infiltrate the site of inflammation is strongly correlatedwith the induction of proteinuria in most models of anti-GBMdisease. Fc�R have been shown to be involved in numerousautoimmune diseases, such as Ab-induced arthritis (21), bullouspemphigoid (22), immune hemolytic anemia (23), and lupusnephritis (24). Furthermore, a potential role for both comple-ment and Fc�R has been shown in anti-GPI Ab-induced arthritis(25), and anti-C1q Ab-induced lupus nephritis (26).

For induction of anti-GBM disease, the direct model depends oncomplement activation, as was shown in C3�/� animals (27, 28).This was, at least partially, elicited via the classical pathway, asC4�/� mice (which lack an intact classical and lectin pathway)showed reduced numbers of inflammatory cells and renal injurycompared with WT mice (28, 29). However, a potential role for thealternative pathway of complement was postulated as well (29). Incontrast, the accelerated model is dependent on the presence ofactivating Fc�R (8, 10, 30), which might be explained by thestrong immune response elicited by the preimmunization (29). Inthe accelerated model, the complement system even seems to havea protective function as glomerular injury was more severe inC1q�/� and C3�/� mice (6, 7, 27). The latter might be explainedby the absence of C1q-mediated immune complex clearance, andC3-mediated Ag-antibody complex solubilisation (31), resulting inincreased glomerular IgG and/or immune-complex deposition, re-spectively (6, 7). Importantly, the complement system and Fc�Rwere not important for progression of disease in the second, au-tologous, phase of both models (27, 8, 10, 30), in which potentialroles for inflammatory cells other than neutrophils (i.e., T cells) aswell as angiotensin II were postulated (8, 9, 10).

To circumvent and to prevent an excessive (biphasic) immuneresponse, we have used a distinct approach to induce anti-GBMglomerulonephritis. In our attenuated passive mouse model, ad-ministration of subnephritogenic amounts of rabbit anti-mouse-GBM Ab (�GBM) is followed one week later with a predeter-mined concentration of mouse mAb (mAb) against rabbit IgG(Ms�Rb IgG). Using this approach, we succeeded to induce renalinjury (primarily assessed by development of albuminuria) that isdependent on both complement and triggering of Fc�R. Further-more, proteinuria occurs acutely following the administration ofMs�Rb IgG, and proceeds in a chronic fashion, which remains

dependent on both an intact complement system and activatingFc�R. This novel model, therefore, allows elucidation of the rel-ative contribution of different mediator systems of the immunesystem to the development of renal injury, and also provides a veryuseful platform for the assessment of different treatment protocolsand evaluation of drugs that ultimately may be beneficial for thetreatment of anti-GBM disease.

Materials and MethodsMice

C57BL/6 wild-type (WT) mice were purchased from Charles River Lab-oratories. The C1q�/�, C4�/�, and C3�/� mice were provided by MarinaBotto (Imperial College, London, U.K.) and Mike Carroll (Harvard Med-ical School, Boston, MA), respectively, and backcrossed for six genera-tions to the C57BL/6 background. The FcR-�-chain�/� mice, generated inC57BL/6 background, are a gift of Takashi Saito (Yokohama, Japan). Allstrains, including the FcR-�-chain�/� � C3�/� double-knockout strain,which was generated by crossing FcR-�-chain�/� mice with C3�/� mice,were maintained by the Department of Human Genetics and the Nephrol-ogy Department, as described before (26). All experiments were approvedby the Leiden University animal ethics committee, and performed accord-ing to institutional and national guidelines.

Antibodies

Rabbit anti-mouse GBM Abs (�GBM Abs) were obtained as described before(26). The mouse mAb (IgG2a) anti-rabbit IgG (Ms�Rb IgG, Fc-specific) wasproduced from hybridoma CRL-1753 (American Type Culture Collection) inIMDM (Lonza BioWhittaker) supplemented with 5% FCS, 0.05 mM 2-ME,0.5 ng/ml IL6 (Strathmann Biotec) and antibiotics. Ms�Rb IgG in the super-natant was concentrated, centrifuged at 10,000 rpm and purified by protein GSepharose chromatography (Amersham Biosciences). Presence of Ms�Rb IgGin the elution fractions was confirmed by ELISA. The positive fractions werepooled, concentrated and, after dialyzing against PBS, aliquotted, and stored at�20°C until further use. A fraction of Ms�Rb IgG was coupled to digoxigenin(DIG) as described before (32).

Induction of glomerulonephritis

At day 0, mice were injected i.v. with 0.5 mg �GBM Ab (200 �l in PBS).After 6 days, 1 mg of monoclonal Ms�Rb IgG was given i.p. Urine sam-ples were collected at day 0, 5, and 6 to measure albumin excretion in urine(albuminuria). Mice were sacrificed 24 h after administration of Ms�RbIgG (acute model), or one month after Ms�Rb IgG injection (chronicmodel). Thereafter, kidneys were collected, snap frozen and stored at�150°C for immunohistochemical analysis.

Albuminuria and creatinine

To collect urine, mice were placed in metabolic cages for 18 h. Urine wascentrifuged and subsequently stored at �20°C. Rocket immuno-electro-phoresis (protocol modified from Ref. 33) was used to quantify albuminlevels in urine. Albuminuria per mouse is plotted as mg/24 h. Urine cre-atinine levels were determined by a kinetic colorimetric assay using a com-mercially available kit (Creatinine Jaffe method, Roche diagnostics) and aCobas Integra 800 analyser (Roche).

Histological analysis

Mouse kidneys were sectioned into 3-�m slides to assess deposition ofcomplement components and IgG. The presence of rabbit IgG (�GBM Ab)in glomeruli was detected using a goat anti-rabbit IgG FITC Ab (NordicImmunological Laboratories). Mouse IgG was detected using goat anti-mouse IgG coupled to Oregon Green (Invitrogen). C1q deposition wasvisualized using a polyclonal rabbit anti-mouse C1q, which was coupled toDIG (26). C4 and C3 were detected using rat anti-mouse C4 and C3 mAb(Hycult Biotechnology) followed by DIG-coupled mouse anti-rat � mAb(His 8, provided by Dr. N.A. Bos, Department of Histology and Cell Bi-ology, Groningen University, The Netherlands), which was then detectedusing a (Fab�)2 sheep anti-DIG-FITC (Roche Diagnostics). The intensity ofstaining was scored by at least two independent observers, who wereblinded to the code of the sections. A score between 0 (no staining) and 3(maximal staining) was given. For each mouse, the score was determinedby averaging the scores given by the observers.

Leukocytes and granulocytes were visualized using an anti-CD45 Ab(AbD Serotec, Oxford, U.K.) and mAb Gr-1 (a gift from Dr. R. Toes,Department of Rheumatology, LUMC, The Netherlands), respectively, fol-lowed by incubation with goat anti-rat IgG, coupled to Alexa fluor 488

3981The Journal of Immunology

(Molecular Probes, Leiden, The Netherlands). Leukocytes were counted bya pathologist who was blinded to the code of the sections.

For light microscopy, the kidney was fixed in paraformaldehyde, em-bedded in paraffin, and 3-�m sections were stained with H&E, and periodicacid-Schiff (PAS). Evaluation of histopathologic changes was performedby a pathologist who was blinded to the code of the sections.

Statistics

Statistical differences were determined by one-way ANOVA using KruskalWallis statistics, Dunn’s correction for multiple comparison, and Chi-square analysis. Significance was accepted when p � 0.05.

ResultsSet up and evaluation of the attenuated mouse model foranti-GBM glomerulonephritis

To facilitate interactions with mouse effector systems in a naturalfashion, mice were injected with a fixed subnephritogenic amountof rabbit �GBM Ab (0.5 mg i.v.) that serves as an anchor on theGBM for the subsequent binding of mouse mAb anti-rabbit IgG(Ms�Rb IgG), which was given 6 days later. This Ms�Rb IgGbinds well to the �GBM Ab, as shown in Fig. 1. In in vivo Ms�RbIgG dose-finding pilot experiments, an injection of 1 mg Ms�RbIgG (i.p. at day 6) was chosen to be the best dose to induceglomerulonephritis.

To evaluate the manifestation of disease, as measured by al-buminuria, glomerular complement deposition, leukocyte in-flux, and glomerular pathology, C57BL/6 mice were injectedwith PBS (as a control) or with �GBM Ab, followed 6 dayslater with the injection of PBS (as a control) or Ms�Rb IgG.Albuminuria was determined in all groups 24 h after the secondinjection. Mice that received PBS or �GBM Ab alone, did notdevelop albuminuria (Fig. 2A). Only when both �GBM andMs�Rb IgG were given, overt albuminuria was detected, indi-cating that renal damage was specifically elicited by the injec-tion of the mouse Ab (Fig. 2A). Blood urea nitrogen, as well asurine creatinine levels were unaltered between these threegroups (data not shown).

FIGURE 1. Ms�Rb IgG binds to �GBM in vitro. Rabbit IgG �GBMIgG (1 �g/ml) was coated on a plate, followed by a titration of DIG-labeledmAb Ms�Rb IgG. Detection was completed using HRP-conjugated F(ab�)2

fragments of sheep anti-DIG, and ABTS as a substrate.

FIGURE 2. Induction of glomeru-lonephritis in WT mice. C57BL/6 micewere injected i.v. with either rabbit�GBM (0.5 mg in PBS) or PBS alone.After 6 days, PBS or 1 mg Ms�Rb IgGwas injected i.p. Urine was collectedusing metabolic cages and mice weresacrificed after 24 h. A, Urinary albu-min excretion was determined by rocketimmunoelectrophoresis. Data are ex-pressed as mg/24 h. B, Mouse kidneysof PBS, �GBM, or �GBM plus Ms�RbIgG injected mice were stained for thepresence of leukocytes. Data are ex-pressed as number of leukocytes perglomerulus. C, Mouse kidneys of PBS,�GBM, or �GBM plus Ms�Rb IgG in-jected mice were stained for the pres-ence of rabbit IgG, mouse IgG, C1q,C4, and C3 using specific Abs (magni-fication �200). D, Representative pic-tures of PAS stained kidney sections ofPBS, �GBM, or �GBM plus Ms�RbIgG injected mice (magnification�250). Example of a PAS-positive de-posit in an afferent arteriole of a �GBMcontrol mouse is shown (�, middle).PAS-positive deposit is present in thewhole glomerulus of a �GBM plusMs�Rb IgG injected mouse (right).

3982 DEVELOPMENT OF PROTEINURIA IN ANTI-GBM DISEASE

To determine binding of �GBM Ab and Ms�Rb IgG as well ascomplement deposition in mouse kidneys, mice were sacrificed24 h after injection of the Ms�Rb IgG, and kidney sections wereexamined using immunofluorescent stainings (Fig. 2C). Presenceof �GBM Ab was visualized using a rabbit IgG staining, whichshowed that rabbit IgG was equally present in the glomeruli of allthe mice that received �GBM Abs, and that this was distributed ina GBM-like pattern. Mice that received both �GBM as well asMs�Rb IgG showed a deposition of mouse IgG in a GBM-likepattern. Although glomeruli of mice that received �GBM alonewere positive for mouse IgG, staining was significantly less ascompared with mice that received both Abs (a score of 0.5 � 0.2vs 2.3 � 0.4; p � 0.01, respectively). These data indicate thatMs�Rb IgG colocalizes with rabbit IgG on the GBM in vivo.Glomerular rabbit or mouse IgG deposition was not detectable inmice that were injected with PBS alone.

Ab deposition was associated with complement activation inmouse glomeruli, as C1q, C4, and C3 were also deposited in glo-meruli in a GBM-like pattern (Fig. 2C). C1q and C3 depositionwere significantly increased in glomeruli of mice that receivedboth �GBM and Ms�Rb IgG compared with mice that received�GBM alone (a score of 2.4 � 0.3 vs 1.4 � 0.1 for C1q; p � 0.01,and a score of 3.1 � 0.35 vs 1.9 � 0.4 for C3; p � 0.01, respec-tively), whereas differences in C4 deposition were not significant.Control (PBS) mice had no detectable glomerular deposition ofC1q, C4, or C3.

At 24 h after the last injection, the group of mice that receivedboth �GBM and Ms�Rb IgG had an average of 2.5 leukocytes perglomerulus, which consisted of granulocytes and some macro-phages (Fig. 2B). These leukocyte numbers were not significantlydifferent from �GBM control mice ( p � 0.09). The influx of neu-trophils might have been missed as glomerular neutrophil influxpeaks at several hours, and resolves within 24 h in most models(4, 5, 9, 28, 34 –36), and this was also shown in a neutrophil-dependent direct model for anti-GBM disease (4, 28). Indeed, inan experiment in which mice were sacrificed within 15 h afterinjection of Ms�Rb IgG, significantly more leukocytes wereobserved in glomeruli of �GBM and Ms�Rb IgG experimentalmice, compared with controls (4.3 � 1.1 vs 1.6 � 0.4 cells/glomerulus; p � 0.01).

Furthermore, although glomerular sclerosis, endocapillary pro-liferation, and mesangial matrix expansions were not detected in

mice that had received both �GBM and Ms�Rb IgG, the percent-age of glomeruli with PAS-positive deposits, as well as the loca-tion of these PAS-positive deposits, was altered in mice that re-ceived �GBM plus Ms�Rb IgG, compared with mice that hadreceived �GBM alone (Fig. 2D): the average percentage of glo-meruli affected in �GBM control mice was 2 � 3.5% (n � 6)compared with 20.8 � 7.2% (n � 8) in mice that obtained both�GBM as well as Ms�Rb IgG. In addition, mice that had receivedboth �GBM plus Ms�Rb IgG had diffuse PAS-positive deposits,whereas the minor percentage of affected glomeruli of mice thatreceived �GBM alone showed deposits that were only present inafferent arterioles (Fig. 2D).

The attenuated anti-GBM model is dependent on bothcomplement activation and Fc�R

The underlying mechanisms of anti-GBM mediated renal injurywere evaluated by using C57BL/6 mouse strains, deficient for spe-cific complement components. WT, C1q�/�, C4�/�, and C3�/�

mice were injected with both �GBM and Ms�Rb IgG and thedevelopment of glomerulonephritis was assessed by measuring al-buminuria, and evaluation of complement deposition.

Compared with WT mice, development of albuminuria was sig-nificantly reduced in C3�/� mice, implying an important role forcomplement in development of renal injury (Fig. 3A). BothC1q�/� mice (which lack an intact classical complement pathway)as well as C4�/� mice (which lack a functional classical and lectinpathway) developed albuminuria similar to that observed in WTmice, indicating that neither C1q nor C4 are crucial for develop-ment of anti-GBM mediated glomerulonephritis (Fig. 3A). How-ever, C1q, the first component of the classical pathway, waspresent in glomeruli of mice deficient for C3 or C4 (Fig. 3B),whereas C4 deposition, which is downstream of C1q activation,was reduced to background in mice lacking C1q (Fig. 3C), indi-cating that the lectin pathway most likely does not play a signif-icant role. Furthermore, C3 deposition was present but clearly de-creased in glomeruli of C1q�/� and C4�/� mice (Fig. 3D), whichindicates that the classical pathway of complement is, at least par-tially, involved in the glomerular deposition of C3 in this model.All knockout mice lacked significant staining for the complementcomponent that was knocked out, confirming the specificity of thestaining (Fig. 3, B–D).

FIGURE 3. Role of complement activation in renal in-jury. A, Groups of WT, C1q�/�, C4�/�, and C3�/� micewere injected with rabbit �GBM followed 6 days later byeither Ms�Rb IgG or PBS as control. Urinary albuminexcretion was measured by rocket immunoelectrophoresisand expressed as mg/24 h. B–D, Kidneys of WT, C1q�/�,C4�/�, and C3�/� mice injected with both rabbit �GBMand Ms�Rb IgG were stained for the presence of C1q, C4,and C3. Semiquantitative analysis of complement deposi-tion was performed in the different strains. Fluorescenceintensity (range 0–3) was scored by at least two blindedobservers, and average score for each mouse was calcu-lated. �, p � 0.05 compared with WT; ��, p � 0.01 com-pared with WT.

3983The Journal of Immunology

To further unravel the underlying mechanism of anti-GBM me-diated glomerulonephritis, the role of Fc�R was evaluated in FcR-�-chain�/� mice, which lack all activatory IgG Fc receptors. Asshown in Fig. 4, FcR-�-chain�/� mice were completely protected

from development of albuminuria (Fig. 4), although all FcR-�-chain�/� mice exhibited glomerular depositions of Ab and com-plement (data not shown).

The attenuated anti-GBM model results in chronic renal injury

To investigate whether anti-GBM mediated glomerulonephritiscould proceed into a chronic fashion, and to evaluate whether al-buminuria would remain dependent on the presence on an intactcomplement system and on the presence of Fc�R, we injected WT,C3�/�, and FcR-�-chain�/� mice with �GBM followed 6 dayslater with Ms�Rb IgG, similar to experiments mentioned above.Furthermore, FcR-�-chain�/� � C3�/� double-knockout mice,which lack both Fc�R as well as an intact complement system,were included, as well. However, instead of sacrificing the animals24 h after injection of Ms�Rb IgG, mice were followed for 37days, and urine samples were collected three times a week. WTcontrol mice receiving �GBM Ab alone did not develop any al-buminuria during the whole period (Fig. 5A). In contrast, a rapidrise of albuminuria was observed in WT mice in the first two days

FIGURE 4. Role of Fc�R in development of albuminuria. FcR-�-chain�/� mice and WT mice were injected with rabbit �GBM, followed 6days later by an injection of Ms�Rb IgG. Albumin excretion in urine wasmeasured and expressed as mg/24 h. ���, p � 0.001.

FIGURE 5. Mice injected with�GBM and Ms�Rb IgG develop andmaintain albuminuria. WT, C3�/�,FcR-�-chain�/�, and FcR-�-chain�/� �C3�/� double-knockout mice wereinjected with rabbit �GBM and 6days later with Ms�Rb IgG (arrow inA). As a control, WT mice were in-jected with �GBM Ab alone (de-picted as “�GBM only”). Urine wascollected three times a week and an-alyzed for albumin content. Micewere sacrificed at day 37. A, Averagealbumin excretion is shown per group.The percentage of mice with mesan-gial matrix expansion (B), and glo-merular sclerosis (C) at day 37 areshown per group. D, Representativepictures of PAS stained kidney sectionsof �GBM or �GBM plus Ms�Rb IgGinjected mice (magnification �250).No major abnormalities were detectedin mice injected with �GBM only,whereas mesengial matrix expansion(lower left), segmental matrix accumu-lation (lower middle) and sclerosis (ar-row, lower middle) and tubular casts(white asterisks, lower right) are ob-served in �GBM plus Ms�Rb IgG in-jected mice.

3984 DEVELOPMENT OF PROTEINURIA IN ANTI-GBM DISEASE

after injection of Ms�Rb IgG, after which albuminuria declinedslowly, but remained significantly present up until day 37 (Fig.5A). Similar to what was observed in the acute setup, C3�/� miceshowed decreased albuminuria, whereas albuminuria was com-pletely absent in FcR-�-chain�/� mice. In C3�/� mice, albumin-uria was slightly above background up until day 37. As expected,albuminuria was absent in FcR-�-chain�/� � C3�/� double-knockout mice, as well (Fig. 5A).

At day 37, mice were sacrificed and kidneys were examinedhistologically. No differences were found in leukocyte numbers,which might be due to the late time point on which the mice weresacrificed (data not shown). However, clear histological injury wasnow observed in WT mice injected with �GBM and Ms�Rb IgG,as evidenced by significantly increased mesangial matrix expan-sion (Fig. 5, B and D), and segmental matrix accumulation andglomerular sclerosis (Fig. 5, C and D), as compared with mice

injected with �GBM alone. Furthermore, tubular casts, represen-tative for heavy proteinuria, was observed in 50% of WT animals,whereas tubular casts were not observed in �GBM control mice(Fig. 5D).

Mouse kidneys were furthermore stained for rabbit IgG(�GBM), mouse IgG (Ms�Rb IgG), and complement products. Inall mice, �GBM could still be observed in significant amounts inthe glomeruli, indicating that rabbit IgG was not cleared from theGBM. Furthermore, no significant differences were observed inrabbit IgG staining between WT (either injected with �GBMalone, or injected with both �GBM and Ms�Rb IgG) and knockoutmice (Fig. 6A, and data not shown). Mouse IgG staining was ob-served in all mice that received �GBM and Ms�Rb IgG, whereashardly any mouse IgG could be observed in mice injected with�GBM alone, indicating that Ms�Rb IgG remained anchored tothe �GBM Ab for over a month (Fig. 6A). At day 37, complement

FIGURE 6. Rabbit IgG, mouseIgG, and complement depositions arepresent at day 37. A, Groups of miceinjected with rabbit �GBM alone (de-picted as “�GBM”) or injected withboth �GBM and Ms�Rb IgG were sac-rificed after 37 days, and the presenceof rabbit IgG, mouse IgG, C1q, C4, andC3 was determined by immunofluores-cence (representative stainings, magni-fication �200). B–D, Quantitative anal-ysis of C1q, C4, and C3 deposition wasperformed on kidneys of the differentmouse strains. Fluorescence intensity(range 0–3) was scored by at least twoblinded observers, and the averagescore for each mouse was calculated. �,p � 0.05 compared with WT; ��, p �0.01 compared with WT.

3985The Journal of Immunology

deposition, as measured with C1q, C3, and C4, was low in controlmice (injected with �GBM alone) whereas all and knockout miceinjected with both Abs showed significant complement deposition(Fig. 6, B–D), indicating either low clearance of deposited com-plement or continuous complement activation. Compared with WTmice, C3 deposition was absent in C3�/� mice, thereby confirmingstaining specificity, and was significantly increased in FcR-�-chain�/� mice (Fig. 6D).

DiscussionIn this manuscript, we describe a novel model for anti-GBM me-diated glomerulonephritis, characterized by albuminuria, comple-ment activation, and involvement of Fc�R. The model is based onpassive administration of a fixed amount of rabbit �GBM Abwhich serves as an anchor for the subsequent binding of a mousemAb against rabbit IgG. In this way, the present model differs fromearlier described models where either high amounts of rabbit orgoat anti-GBM Ab were administered to naive mice (4, 5), or themodel where mice were first preimmunized with IgG of the anti-GBM Ab species, followed a week later with anti-GBM Ab (6, 7).The advantage of the model described in the present study is thatthe degree of proteinuria is dependent on the amount of Ab used.In pilot experiments we first determined the required amount ofrabbit �GBM Ab and the amount of Ms�Rb IgG, which resultedin a choice of 0.5 mg polyclonal rabbit Ab and 1.0 mg mono-clonal mouse Ab per mouse. Compared with the earlier de-scribed models, the present model results in a controllable de-gree of renal injury, which is not dependent on the immuneresponse of individual mice. Indeed, when higher doses ofmonoclonal mouse anti-rabbit Abs were used, a more severedisease occurred (data not shown). Therefore, this approachseems to circumvent potential variations in the strength of theimmune response against either rabbit immunoglobulins or col-lagen type IV fragments as required in the accelerated nephro-toxic nephritis model of anti-GBM disease (6, 7) and the modelof Goodpasture’s syndrome (37– 40), respectively.

The induction of disease was rapid, with an average proteinuriaof 5 mg/24 h in WT mice. Furthermore, a clear involvement ofclassical complement activation was observed with deposition ofC1q, C4, and C3 in glomeruli of mice receiving both rabbit �GBMand Ms�Rb IgG. Interestingly, mice with a deficiency in the clas-sical pathway (C1q�/� mice) or the classical and the lectin path-way (C4�/� mice) displayed almost the same degree of proteinuriaas compared with WT mice. This observation is compatible withthe finding that C3 deposition in these mice was still present, whileC4 and C1q deposition at the tissue level was clearly absent ordiminished (Fig. 3). Whereas C3 deposition in C1q�/� and C4�/�

mice was present but clearly reduced, C4 deposition in C1q�/�

mice was reduced to background, indicating that the classical path-way clearly contributes to complement activation, whereas the lec-tin pathway most likely does not play a significant role in thismodel. Although C1q�/� and C4�/� mice displayed a reducedlevel of C3 deposition, there still remained clear deposition of C3indicating a diversion of complement activation from the classicalto the alternative pathway, and therefore maintenance of the degreeof proteinuria in these knockout mice.

This is in contrast with an experimental mouse model for lupusnephritis, a glomerular immune complex disease which was shownto be mediated via the classical pathway of complement (26).However, similar observations have been made in other passiveAb-mediated models of experimental disease in mice, like thecollagen arthritis model (16), subepidermal blistering disease (17),anti-phospholipid syndrome (12), ischemia/reperfusion injury(18), and cryoglobulin-induced immune complex glomerulone-

phritis (19), where the alternative pathway was shown to be bothrequired and sufficient for disease induction.

The question is why a diversion of complement activation fromclassical to the alternative pathway is observed. For our observa-tion, which is in line with previous studies in another model ofanti-GBM mediated disease in mice (28), there are, at least, twoexplanations. The first explanation is that certain Abs, like rabbitAbs may elicit activation of the alternative pathway as has beenreported by several groups (41, 42). Additionally, studies in col-lagen Ab-induced arthritis have shown that mouse Abs can acti-vate the complement system in mice via the alternative pathway(16). A second explanation may be found in the observation thatmodified self tissue may lead to the exposure of intracellular Agswhich may directly activate the alternative pathway as exemplifiedby the binding of properdin to, for instance, DNA (43). Binding ofAbs to different cells such as endothelial cells and mesangial cellsmay lead to apoptosis with a flip-flop of the cell membrane. Thismechanism is associated with a loss of protection of the tissueinvolved against complement attack. Various studies have shownthat apoptotic cells are more prone to complement-recognitionthan their intact counterparts (44–46). Additionally, induction ofapoptosis is accompanied with direct binding of properdin (43,47), which then can serve as a focus point for the direct binding ofC3b. This will lead to local alternative pathway complement ac-tivation, as there is continuous turnover of C3 into C3b and C3d inthe circulation (48).

That C3 activation is critical for the development of proteinuriais evidenced by the observation that C3�/� mice display a clearlyreduced degree of proteinuria (Fig. 3A). C3 activation and depo-sition is associated with the further recruitment and engagement ofinfiltrating cells such as neutrophils and macrophages (Fig. 2B, anddata not shown). These cells can than directly interact via theirFc�R and complement receptors with deposited IgG and comple-ment fragments, respectively. Involvement of Fc�R in the induc-tion of renal injury was investigated in FcR-�-chain�/� mice,which have a complete absence of all stimulatory Fc�R, either oninfiltrating inflammatory cells or on resident kidney cells. FcR-�-chain�/� mice did not show any signs of proteinuria at 24 h ofdisease. In this respect, the present observations are discrepantfrom earlier reports in mice preimmunized with IgG from the spe-cies of the anti-GBM Abs (6, 7, 27). In these studies, it was dem-onstrated that there was a protective role for complement in thedegree of renal damage as was shown by proteinuria and otherparameters of renal injury. We reason that these differences aredependent on the severity of the renal damage i.e., the amount,time span, isotype, and the species-origin of the IgG that is in-volved in the ongoing process (36). At high amounts of IgG, com-plement activation is not obligatory, as has been observed bySheerin et al. (29). Therefore, we feel that the model we describein the present manuscript allows the further analysis of the con-tribution of both Fc�R and complement in the process of anti-GBM mediated renal damage in mice. These observations suggestthat modulation of Fc�R engagement or interference with the ac-tivation of complement may provide strategies for the control ofanti-GBM disease.

In additional experiments summarized in Fig. 5 and 6, we dem-onstrate that the attenuated model of anti-GBM disease has achronic course. In WT mice, proteinuria was persistently presentfor at least 37 days. Sustained proteinuria was dependent both oncomplement and the presence of activatory Fc�R. Although theabsence of Fc�R is a prerequisite for disease chronicity, the ab-sence of complement leads to a moderate disease over time. Im-munofluorescence analysis revealed that Ab deposition (both rab-bit IgG and mouse IgG) persisted over the whole period and that

3986 DEVELOPMENT OF PROTEINURIA IN ANTI-GBM DISEASE

this was accompanied with C1q, C4, and C3 deposition. Impor-tantly, even after 1 mo, mouse IgG deposition was not increased inglomeruli of control mice which received �GBM only (fluores-cent intensity was similar to irrelevant control staining), sup-porting the view that the injected amount of rabbit �GBM didnot induce an autologous immune response. It is of interest tonote that C3 deposition was higher in FcR-�-chain�/� mice, ascompared with the WT mice. This observation is compatiblewith the absence of Fc�R-mediated activation of infiltratingneutrophils. It has been reasoned that secretion of proteolyticenzymes by activated neutrophils can lead to degradation ofdeposited C3 fragments, which may therefore determine theamount of activated C3 that remains deposited at the site oftissue injury (49). Histological analysis of kidney tissue re-vealed that �50% of the WT mice developed sclerotic glomer-uli on day 37 of the disease (Fig. 5, C and D). This sclerosis wascompletely absent in double-knockout mice lacking both C3and Fc�R (FcR-�-chain�/� � C3�/� mice), while mice lackingeither Fc�R (FcR-�-chain�/� mice) or C3 (C3�/� mice) dis-played a reduced degree of sclerosis. Furthermore, mesangialmatrix expansion was observed in �65% of WT mice animals,and was reduced in mice lacking Fc�R (Fig. 5, B and D). How-ever, mesangial matrix expansion was increased in C3�/� mice(�80% of mice). This might be explained by the absence ofC3-mediated Ag-antibody complex solubilization (31), whichcan lead to the presence of more immune-complexes in glomer-uli, and thus an increased Ab-mediated injury. The latter mightalso explain the presence of mesengial matrix expansion inmice deficient for both Fc�R and C3. These pathological phe-nomena were not observed in the acute model at 24 h afterinjection of Ms�Rb IgG. This may be explained by the timeinterval of 24 h, which seems to be too short for developmentof glomerular expansion and sclerosis.

Taken together, our novel well controllable model of anti-GBM disease shows an interplay between humoral mechanismsincluding Abs and complement (the latter being partially re-sponsible for tissue injury and leukocyte attraction), and cellu-lar mechanisms. Together, these mechanisms lead to acute glo-merular injury, which can proceed into a chronic renal injurythat remains complement and Fc receptor dependent. These re-sults provide insight into the different mechanisms that play arole in anti-GBM disease, and, furthermore, this novel attenu-ated model may also provide a useful platform for the evalua-tion of drugs that may ultimately be beneficial for the treatmentof Goodpasture’s syndrome.

AcknowledgmentsWe thank Kyra Gelderman for technical assistance. Furthermore, we thankMelissa Mulder and Jos van der Kaa for delivery of knockout mice, andFred de Boer, Ben van der Geest, Peter de Jong, and Michel Mulders forexcellent animal care. We would like to thank Laura Anna Giardino andAlessandro Corbelli for expert assistance with histological analysis of kid-ney tissues.

DisclosuresThe authors have no financial conflict of interest.

References1. Salama, A. D., J. B. Levy, L. Lightstone, and C. D. Pusey. 2001. Goodpasture’s

disease. Lancet 358: 917–920.2. Hudson, B. G., K. Tryggvason, M. Sundaramoorthy, and E. G. Neilson. 2003.

Alport’s syndrome, Goodpasture’s syndrome, and type IV collagen. N. Engl.J. Med. 348: 2543–2556.

3. Salant, D. J., and A. V. Cybulsky. 1988. Experimental glomerulonephritis. Meth-ods Enzymol. 162: 421–461.

4. Feith, G. W., K. J. Assmann, M. J. Bogman, A. P. van Gompel, J. Schalkwijk,and R. A. Koene. 1996. Different mediator systems in biphasic heterologousphase of anti-GBM nephritis in mice. Nephrol. Dial. Transplant. 11: 599–607.

5. Robson, M. G., H. T. Cook, C. D. Pusey, M. J. Walport, and K. A. Davies. 2003.Antibody-mediated glomerulonephritis in mice: the role of endotoxin, comple-ment and genetic background. Clin. Exp. Immunol. 133: 326–333.

6. Robson, M. G., H. T. Cook, M. Botto, P. R. Taylor, N. Busso, R. Salvi,C. D. Pusey, M. J. Walport, and K. A. Davies. 2001. Accelerated nephrotoxicnephritis is exacerbated in C1q-deficient mice. J. Immunol. 166: 6820–6828.

7. Sheerin, N. S., K. Abe, P. Risley, and S. H. Sacks. 2006. Accumulation of im-mune complexes in glomerular disease is independent of locally synthesized c3.J. Am. Soc. Nephrol. 17: 686–696.

8. Park, S. Y., S. Ueda, H. Ohno, Y. Hamano, M. Tanaka, T. Shiratori, T. Yamazaki,H. Arase, N. Arase, and A. Karasawa. 1998. Resistance of Fc receptor- deficientmice to fatal glomerulonephritis. J. Clin. Invest. 102: 1229–1238.

9. Tarzi, R. M., and H. T. Cook. 2003. Role of Fc� receptors in glomerulonephritis.Nephron. Exp. Nephrol. 95: e7–e12.

10. Suzuki, Y., I. Shirato, K. Okumura, J. V. Ravetch, T. Takai, Y. Tomino, andC. Ra. 1998. Distinct contribution of Fc receptors and angiotensin II-dependentpathways in anti-GBM glomerulonephritis. Kidney Int. 54: 1166–1174.

11. Trouw, L. A., M. A. Seelen, and M. R. Daha. 2003. Complement and renaldisease. Mol. Immunol. 40: 125–134.

12. Girardi, G., J. Berman, P. Redecha, L. Spruce, J. M. Thurman, D. Kraus,T. J. Hollmann, P. Casali, M. C. Caroll, and R. A. Wetsel. 2003. ComplementC5a receptors and neutrophils mediate fetal injury in the antiphospholipid syn-drome. J. Clin. Invest. 112: 1644–1654.

13. Walport, M. J. 2001. Complement: second of two parts. N. Engl. J. Med. 344:1140–1144.

14. Kohl, J. 2006. The role of complement in danger sensing and transmission. Im-munol. Res. 34: 157–176.

15. Kohl, J. 2001. Anaphylatoxins and infectious and non-infectious inflammatorydiseases. Mol. Immunol. 38: 175–187.

16. Banda, N. K., K. Takahashi, A. K. Wood, V. M. Holers, and W. P. Arend. 2007.Pathogenic complement activation in collagen antibody-induced arthritis in micerequires amplification by the alternative pathway. J. Immunol. 179: 4101–4109.

17. Mihai, S., M. T. Chiriac, K. Takahashi, J. M. Thurman, V. M. Holers,D. Zillikens, M. Botto, and C. Sitaru. 2007. The alternative pathway of comple-ment activation is critical for blister induction in experimental epidermolysisbullosa acquisita. J. Immunol. 178: 6514–6521.

18. Thurman, J. M., and V. M. Holers. 2006. The central role of the alternativecomplement pathway in human disease. J. Immunol. 176: 1305–1310.

19. Trendelenburg, M., L. Fossati-Jimack, J. Cortes-Hernandez, D. Turnberg,M. Lewis, S. Izui, H. T. Cook, and M. Botto. 2005. The role of complement incryoglobulin-induced immune complex glomerulonephritis. J. Immunol. 175:6909–6914.

20. Nimmerjahn, F., and J. V. Ravetch. 2006. Fc� receptors: old friends and newfamily members. Immunity 24: 19–28.

21. Kagari, T., D. Tanaka, H. Doi, and T. Shimozato. 2003. Essential role of Fc �receptors in anti-type II collagen antibody-induced arthritis. J. Immunol. 170:4318–4324.

22. Zhao, M., M. E. Trimbeger, N. Li, L. A. Diaz, S. D. Shapiro, and Z. Liu. 2006.Role of FcRs in animal model of autoimmune bullous pemphigoid. J. Immunol.177: 3398–3405.

23. Clynes, R., and J. V. Ravetch. 1995. Cytotoxic antibodies trigger inflammationthrough Fc receptors. Immunity 3: 21–26.

24. Clynes, R., C. Dumitru, and J. V. Ravetch. 1998. Uncoupling of immune complexformation and kidney damage in autoimmune glomerulonephritis. Science 279:1052–1054.

25. Ji, H., K. Ohmura, U. Mahmood, D. M. Lee, F. M. Hofhuis, S. A. Boackle,K. Takahashi, V. M. Holers, M. Walport, and C. Gerard. 2002. Arthritis criticallydependent on innate immune system players. Immunity 16: 157–168.

26. Trouw, L. A., T. W. Groeneveld, M. A. Seelen, J. M. Duijs, I. M. Bajema,F. A. Prins, U. Kishore, D. J. Salant, J. S. Verbeek, and C. van Kooten. 2004.Anti-C1q autoantibodies deposit in glomeruli but are only pathogenic in combi-nation with glomerular C1q-containing immune complexes. J. Clin. Invest. 114:679–688.

27. Sheerin, N. S., T. Springall, K. Abe, and S. H. Sacks. 2001. Protection and injury:the differing roles of complement in the development of glomerular injury. Eur.J. Immunol. 31: 1255–1260.

28. Hebert, M. J., T. Takano, A. Papayianni, H. G. Rennke, A. Minto, D. J. Salant,M. C. Carroll, and H. R. Brady. 1998. Acute nephrotoxic serum nephritis incomplement knockout mice: relative roles of the classical and alternate pathwaysin neutrophil recruitment and proteinuria. Nephrol. Dial. Transplant. 13:2799–2803.

29. Sheerin, N. S., T. Springall, M. C. Carroll, B. Hartley, and S. H. Sacks. 1997.Protection against anti-glomerular basement membrane (GBM)-mediated nephri-tis in C3- and C4-deficient mice. Clin. Exp. Immunol. 110: 403–409.

30. Tarzi, R. M., K. A. Davies, M. G. Robson, L. Fossati-Jimack, T. Saito,M. J. Walport, and H. T. Cook. 2002. Nephrotoxic nephritis is mediated by Fc�receptors on circulating leukocytes and not intrinsic renal cells. Kidney Int. 62:2087–2096.

31. Miller, G. W., and V. Nussenzweig. 1975. A new complement function: solubi-lization of antigen-antibody aggregates. Proc. Natl. Acad. Sci. USA 72: 418–422.

32. Groeneveld, T. W., M. Oroszlan, R. T. Owens, M. C. Faber-Krol, A. C. Bakker,G. J. Arlaud, D. J. McQuillan, U. Kishore, M. R. Daha, and A. Roos. 2005.Interactions of the extracellular matrix proteoglycans decorin and biglycan withC1q and collectins. J. Immunol. 175: 4715–4723.

3987The Journal of Immunology

33. Tran, T. N., S. K. Eubanks, K. J. Schaffer, C. Y. Zhou, and M. C. Linder. 1997.Secretion of ferritin by rat hepatoma cells and its regulation by inflammatorycytokines and iron. Blood 90: 4979–4986.

34. Rops, A. L., M. Gotte, M. H. Baselmans, M. J. van den Hoven, E. J. Steenbergen,J. F. Lensen, T. J. Wijnhoven, et al. 2007. Syndecan-1 deficiency aggravatesanti-glomerular basement membrane nephritis. Kidney Int. 72: 1204–1215.

35. Tang, T., A. Rosenkranz, K. J. Assmann, M. J. Goodman, J. C. Gutierrez-Ramos,M. C. Carroll, R. S. Cotran, and T. N. Mayadas. 1997. A role for Mac-1 (CDIIb/CD18) in immune complex-stimulated neutrophil function in vivo: Mac-1 defi-ciency abrogates sustained Fc� receptor-dependent neutrophil adhesion and com-plement-dependent proteinuria in acute glomerulonephritis. J. Exp. Med. 186:1853–1863.

36. Giorgini, A., H. J. Brown, H. R. Lock, F. Nimmerjahn, J. V. Ravetch,J. S. Verbeek, S. H. Sacks, and M. G. Robson. 2008. Fc�RIII and Fc�RIV areindispensable for acute glomerular inflammation induced by switch variantmonoclonal antibodies. J. Immunol. 181: 8745–8752.

37. Bolton, W. K., L. Chen, T. Hellmark, J. Wieslander, and J. W. Fox. 2005. Epitopespreading and autoimmune glomerulonephritis in rats induced by a T cell epitopeof Goodpasture’s antigen. J. Am. Soc. Nephrol. 16: 2657–2666.

38. Bolton, W. K., A. M. Luo, P. L. Fox, W. J. May, and B. C. Sturgill. 1995. Studyof EHS type IV collagen lacking Goodpasture’s epitope in glomerulonephritis inrats. Kidney Int. 47: 404–410.

39. Chen, L., T. Hellmark, V. Pedchenko, B. G. Hudson, C. D. Pusey, J. W. Fox, andW. K. Bolton. 2006. A nephritogenic peptide induces intermolecular epitopespreading on collagen IV in experimental autoimmune glomerulonephritis. J. Am.Soc. Nephrol. 17: 3076–3081.

40. Sado, Y., A. Boutaud, and M. Kagawa. 1998. Induction of anti-GBM nephritis inrats by recombinant �3(IV)NC1 and �4(IV)NC1 of type IV collagen. Kidney Int.53: 664–671.

41. Reid, K. B. 1971. Complement fixation by the F(ab�)2-fragment of pepsin-treatedrabbit antibody. Immunology 20: 649–658.

42. Boxx, G. M., C. T. Nishiya, T. R. Kozel, and M. X. Zhang. 2009. Characteristicsof Fc-independent human antimannan antibody-mediated alternative pathway ini-tiation of C3 deposition to Candida albicans. Mol. Immunol. 46: 473–480.

43. Xu, W., S. P. Berger, L. A. Trouw, H. C. de Boer, N. Schlagwein, C. Mutsaers,M. R. Daha, and C. van Kooten. 2008. Properdin binds to late apoptotic andnecrotic cells independently of C3b and regulates alternative pathway comple-ment activation. J. Immunol. 180: 7613–7621.

44. Nauta, A. J., L. A. Trouw, M. R. Daha, O. Tijsma, R. Nieuwland,W. J. Schwaeble, A. R. Gingras, A. Mantovani, E. C. Hack, and A. Roos. 2002.Direct binding of C1q to apoptotic cells and cell blebs induces complement ac-tivation. Eur. J. Immunol. 32: 1726–1736.

45. Cole, D. S., T. R. Hughes, P. Gasque, and B. P. Morgan. 2006. Complementregulator loss on apoptotic neuronal cells causes increased complement activationand promotes both phagocytosis and cell lysis. Mol. Immunol. 43: 1953–1964.

46. Trouw, L. A., A. M. Blom, and P. Gasque. 2008. Role of complement and com-plement regulators in the removal of apoptotic cells. Mol. Immunol. 45:1199–1207.

47. Kemper, C., L. M. Mitchell, L. Zhang, and D. E. Hourcade. 2008. The comple-ment protein properdin binds apoptotic T cells and promotes complement acti-vation and phagocytosis. Proc. Natl. Acad. Sci. USA 105: 9023–9028.

48. Farries, T. C., P. J. Lachmann, and R. A. Harrison. 1988. Analysis of the inter-actions between properdin, the third component of complement (C3), and itsphysiological activation products. Biochem. J. 252: 47–54.

49. Westerhuis, R., S. C. van Straaten, M. G. van Dixhoorn, N. van Rooijen,N. A. Verhagen, C. D. Dijkstra, E. de Heer, and M. R. Daha. 2000. Distinctiveroles of neutrophils and monocytes in anti-thy-1 nephritis. Am. J. Pathol. 156:303–310.

3988 DEVELOPMENT OF PROTEINURIA IN ANTI-GBM DISEASE