Kidney International, Vol. 64 (2003), pp. 1406–1416

Predominant role of Fc�RIII in the induction of acceleratednephrotoxic glomerulonephritis

TAKAYUKI FUJII, YUKI HAMANO,1 SHIRO UEDA, BUNSHIRO AKIKUSA, SHO YAMASAKI,MAKOTO OGAWA, HIROMITSU SAISHO, J. SJEF VERBEEK, SHINSUKE TAKI,2 and TAKASHI SAITO

Department of Molecular Genetics, Department of Medicine and Clinical Oncology, Graduate School of Medicine, ChibaUniversity, Chiba, Japan; Department of Drug Information and Communication, Graduate School of Pharmaceutical Science,Chiba University, Chiba, Japan; Department of Pathology, Matsudo City Hospital, Matsudo, Japan; Department of HumanGenetics, Leiden University Medical Center, Leiden, The Netherlands; and Laboratory of Cell Signaling, RIKEN ResearchCenter for Allergy and Immunology, Yokohama, Japan

Predominant role of Fc�RIII in the induction of acceleratednephrotoxic glomerulonephritis.

Background. Nephrotoxic glomerulonephritis is induced bythe administration of antibody against the glomerular basementmembrane (GBM). We demonstrated previously that Fc recep-tors for immunoglobulin G (IgG) (Fc�R) play crucial roles inthe induction of accelerated nephrotoxic glomerulonephritisby using FcR�-deficient (�/�) mice. Since FcR��/� mice lackthe cell surface expression of two activating Fc�Rs, Fc�RIand Fc�RIII. The present study aims to identify the Fc�Rresponsible for the induction of nephrotoxic glomerulonephritis.

Methods. Accelerated anti-GBM glomerulonephritis was in-duced in Fc�RI�/�, Fc�RIII�/�, and FcR��/� mice by pre-immunization with rabbit IgG followed by inoculation of rabbitanti-GBM antibody. Histologic analysis and immunostainingof renal sections were performed.

Results. Fc�RI�/� mice as well as wild-type mice showedsevere glomerulonephritis with hypernitremia by the adminis-tration of anti-GBM antibody. In contrast, Fc�RIII�/� miceshowed much milder renal involvement, similar to FcR��/�mice. Histologically, Fc�RI�/� mice showed intracapillary pro-liferation, glomerular thrombosis, and crescent formation,whereas Fc�RIII�/� mice showed only glomerular hypercellu-lar changes. The depositions of anti-GBM antibodies, autologousantibodies and complement C3 along the GBM were equallyobserved among all three FcR�/� mouse types by immuno-staining.

Conclusions. Accelerated nephrotoxic glomerulonephritis isinduced predominantly through Fc�RIII but not Fc�RI.

1 Dr. Hamano’s current address is Program in Matrix Biology, Divisionsof Nephrology and Gastroenterology, Department of Medicine, BethIsrael Deaconess Medical Center and Harvard Medical School, Boston,Massachusetts.

2 Dr. Taki’s current address is Department of Organ Transplant, Gradu-ate School of Medicine, Shinshu University, Matsumoto, Japan.

Key words: Fc receptor, glomerulonephritis, anti-GBM antibody.

Received for publication December 13, 2002and in revised form April 9, 2003Accepted for publication May 15, 2003

2003 by the International Society of Nephrology

1406

Immune complex–mediated inflammation is a commonmechanism of various autoimmune diseases such as sys-temic lupus erythematosus, rheumatoid arthritis, vasculi-tis, and glomerulonephritis, including Goodpasture’s syn-drome. Deposition of autoantibodies and immune complexto the systemic or specific organ triggers immune com-plex–mediated inflammation and tissue injury. Immunecomplex–induced diseases had been thought to be medi-ated mainly by complement activation, which then re-cruited various inflammatory cells. These cells then in-duce the further inflammation cascade. However, recentanalyses of various Fc receptor (FcR)–deficient micehave shown that FcR plays crucial roles in the inductionof immune complex–triggered inflammations [1–7].

FcRs play important roles in the connection betweenhumoral and cellular responses by linking the antigen-specific interactions of antibodies to the nonspecific ef-fector mechanisms of FcR-bearing cells [8–10]. In themouse system, there are three classes of Fc receptorsfor immunoglobubulin G (IgG) (Fc�R), FcRs: Fc�RI(CD64) is the high-affinity receptor able to bind to mono-meric IgG (especially IgG2a), and the two low-affinityreceptors Fc�RII (CD32) and Fc�RIII (CD16), whichbind mainly to polymeric IgG and IgG within immunecomplex. In addition, Fc�R can be divided functionallyinto two groups: activating (Fc�RI and Fc�RIII) andinhibitory (Fc�RII) receptors. The two activating recep-tors, Fc�RI and Fc�RIII, are associated with the � chain(FcR�) homodimer. FcR� functions as signal transducerof IgG-binding signals and is also required for the cellsurface expression of both activating receptors. It hasbeen shown that FcR�-deficient (�/�) mice do not ex-press any functional activating Fc�R on the surface ofvarious cells [1, 11], although recent study on Fc�RI-deficient mice showed reduced but not diminished sur-face appearance of functional Fc�RI in FcR��/� mice[12]. Important functions of Fc�R have been demon-

Fujii et al: FccR-induced glomerulonephritis 1407

strated by showing that FcR��/� mice exhibited impair-ment in immune complex–mediated Arthus reaction, arepresentative type III hypersensitivity [13], and resistanceto experimental hemolytic anemia as well as thrombocy-topenia, a representative type II inflammatory sensitivity[2, 5]. Impaired Arthus reaction and low susceptibilityto IgG-induced hemolytic anemia have also been observedin Fc�RIII�/� mice [3, 14]. Recently, Fc�RI-deficientmice were produced [12, 15], and showed some func-tional impairments in IgG2a-immune complex–inducedphagocytosis and cytotoxity, as well as in antigen presen-tation. In vivo, Fc�RI�/� mice were partially protectedfrom IgG2a-induced autoimmune hemolytic anemia andsystemic anaphylaxis, suggesting that Fc�RI has its ownrole, which Fc�RIII is unable to compensate for in vivo[15]. The availability of these knockout mice of two acti-vating FcRs, Fc�RI and Fc�RIII, has enabled us to dis-sect and analyze the individual role in the FcR-depen-dent inflammatory and autoimmune diseases.

Human Goodpasture’s syndrome is an autoimmunedisease with the characteristic rapidly progressive glomeru-lonephritis and hemorrhagic pneumonitis. The presenceof autoantibodies on glomerular and alveolar basementmembranes and the deposition of the immune complexalong these basement membranes have been suggestedto be major pathogenic factors of the disease. For theanalysis of renal involvement in Goodpasture’s syndrome,an experimental model of anti-gomerular membrane dis-ease (GBM) glomerulonephritis, known as nephrotoxicserum glomerulonephritis or Masugi nephritis, has beenused. This model is a representative type II inflammatoryhypersensitivity and is induced by injection of anti-GBMantigens into rats, rabbits, or mice [16–20]. We havedemonstrated that the induction of the murine model ofaccelerated anti-GBM glomerulonephritis was depen-dent on Fc�R from the finding that anti-GBM glomerulo-nephritis is not induced at all in FcR��/� mice [11].In this system, wild-type mice exhibited severe renaldamage, such as crescent formation and glomerularthrombosis, whereas FcR��/� mice were completelyprotected from renal damage without any hypercellular-ity in the glomeruli [11].

In the present study, we examined the individual con-tributions of stimulatory Fc�RI and Fc�RIII using eachtype of Fc�R-deficient mouse to clarify the mechanismof the induction of accelerated Masugi nephritis by iden-tification of the responsible Fc�R. We demonstrated thatFc�RI�/� mice induced a quite similar disease to thatof wild-type mice, whereas Fc�RIII�/� mice exhibitedresistance to anti-GBM glomerulonephritis that wasrather similar to that of FcR��/� mice. These resultsindicate that Fc�RIII plays a predominant role in theinduction of anti-GBM glomerulonephritis.

METHODS

Mice

C57BL/6 mice were purchased from Shizuoka Labor-atory Animal Corp. (Hamamatsu, Japan). The estab-lishment and characteristics of FcR��/� mice withC57BL/6 background have been described previously[11]. Fc�RI�/� and Fc�RIII�/� mice were producedas described previously [3, 15] and backcrossed intoC57BL/6 for six and twelve generations, respectively.All mice were bred and maintained in our own animalfacility under specific pathogen-free conditions. Experi-ments were performed according to institutional guide-lines. Ten- to 14-week age-matched female mice wereused in all experiments.

Preparation of rabbit antimouse GBM antiserum

Rabbit anti-GBM antigens was prepared as previouslydescribed [11]. Briefly, glomeruli were obtained fromkidneys of C57BL/6 mice by sieving with No. 100 and250 meshes and disrupting by supersonic vibration. Thecellular fraction was removed from glomeruli by centrifu-gation three times at 2000g. The final precipitate wasdigested by trypsin at 37�C for 18 hours in 0.1 mol/L Trisbuffer (pH 8.2). After centrifugation, the supernatantwas collected and used as the “GBM antigen.” NewZealand White rabbits were immunized with 10 mg of theGBM antigen emulsified in complete Freund’s adjuvant(CFA)(Wako, Osaka, Japan) in the back six times weeklyand finally intravenously with 2.5 mg of GBM antigenalone. The sera obtained from the rabbits were mixedand used as rabbit antimouse GBM antibodies.

Induction of accelerated Masugi nephritis

Nephrotoxic glomerulonephritis was induced as pre-viously described [11]. Briefly, six female mice per groupwere immunized intraperitoneally with 250 �g of rabbitIgG (Cappel, Aurora, Ohio, USA) in CFA, followed byintravenous injection of 250 �L of rabbit antimouse GBMantibodies 4 days later. On days 7 and 14 after the injec-tion, serum and urine samples were collected to evaluaterenal function. Thereafter, all mice were sacrificed forthe histologic assessment on day 14 after the anti-GBMantibody injection. Moreover, to investigate infiltratingcells in the initial phase of glomerulonephritis, threefemale mice per group were sacrificed at 3 hours afteradministration of anti-GBM antibodies, according to theprevious study of polymorphonuclear cell (PMN) re-cruitment in anti-GBM nephritis [21–23].

Histologic analysis

Kidneys were removed, fixed in 10% buffered forma-lin, and embedded in paraffin. Paraffin sections werestained with hematoxylin and eosin (H&E) and periodicacid-Schiff (PAS) and evaluated by light microscopy. All

Fujii et al: FccR-induced glomerulonephritis1408

histologic analysis was performed in a blinded fashion.We utilized a modified scoring system of experimentalglomerulonephritis described by Raji et al [24, 25] andKanamaru et al [26]; briefly, the following light micro-scopic features were evaluated: (1) intracapillary cellularproliferation; (2) glomerulosclerosis, defined as the dis-appearance of cellular elements from the tuft, collapseof the capillary lumen, and folding of the GBM with en-trapment of amorphous material; (3) glomerular throm-bosis; and (4) extracapillary cellular proliferation (cres-cent). Morphologic changes for the first three criteriawere quantitated as a score of 1 to 4. A score of 1 wasequivalent to 25% of the glomeruli affected by the particu-lar morphologic change, and 4 represented involvementof 100% of the glomeruli. Intermediate values were as-signed a value of 0.5. An injury score was then obtainedby multiplying the severity score by the percentage ofglomeruli exhibiting the same severity of the change. Theextent of injury for each individual tissue specimen wasthen obtained by the addition of these distinct scores. Mor-phologic change for the fourth criterion was expresseddirectly as the percentage of glomeruli. Twenty equatori-ally sectioned glomeruli per kidney were assessed.

Immunofluorescence andimmunohistochemical staining

The kidneys obtained at autopsy were rolled in TissueTek 22-oxacalcitriol (OCT) (Miles, Inc., Elkhart, IN, USA)compound, snap-frozen in liquid nitrogen, and stored at�80�C. Frozen sections were cut at a thickness of 5 �mand fixed in acetone at �30�C. For the immunofluores-cence study, cryostat sections were stained with fluores-cein isothiocyanate (FITC)-conjugated goat antirabbitIgG, Cy3-conjugated antimouse IgG, and goat antimouseC3. As for immunohistochemical study, the cryostat sec-tions were quenched intrinsic peroxidase by H2O2 andincubated with horseradish peroxidase (HRP)-conju-gated rabbit antimouse IgG1 or IgG2a antibodies. Forvisualization, 3-3�-diaminobenzidine (DAB) was used assubstrate. For immunostaining of cell marker, those cryo-stat sections were stained with biotinylated rat antimouseneutrophil monoclonal antibodies (mAbs) (clone 7/4) forPMN, biotinylated anti-B220 antibodies for B cells, andbiotinylated anti-Thy1.2 mAbs for T cells. Macrophageswere stained with goat antimouse CD68 antibodies fol-lowed by biotinylated donkey antigoat IgG antibody.The number of these infiltrating cells in glomeruli wasdetermined by counting cells in glomeruli (more than 50per mouse) at each time point (at 3 hours and on day14 after administration of anti-GBM antibody).

Hematologic examination andmeasurement of albuminuria

Serum levels of blood urea nitrogen (BUN), creati-nine, and albumin were measured by standard methods

for blood chemistry: urease glutamate dehydrogenase(urease-GLDH) (Kyowa, Tokyo, Japan) for BUN, crea-tinase-peroxidase (CRTNase-POD) (Kainos, Tokyo, Ja-pan) for creatinine, and bromocresol green method foralbumin. Mouse urine was individually collected in ametabolic cage for 12 hours on days 7 and 14 after admin-istration of anti-GBM antibodies. Mice were allowedfree access to water but not food.

Measurement of specific antibodies in serum byenzyme-linked immunosorbent assay (ELISA)

Serum levels of antirabbit IgG antibodies and the anti-body against individual IgG subclasses (IgG1 or IgG2a)were measured by ELISA. A 96-well microplate wascoated with 10 �g/mL rabbit IgG (Sigma Chemical Co.,St. Louis, MO, USA) in 0.1 mol/L NaHCO3 at 4�C over-night, blocked, and applied with 50 �L/well of dilutedserum. After reaction at 37�C for 1 hour, the wells werewashed 5 times with phosphate-buffered saline (PBS)containing 0.05% Tween 20 (PBST) and incubated withbiotin-conjugated goat antimouse IgG-Fc (Sigma Chemi-cal Co.) antibody at room temperature for 1 hour. Afterwashing with PBST, the plate was incubated with avidin-peroxidase at room temperature for 30 minutes. For mea-surement of the specific mouse IgG1 or IgG2a antibody,palate was incubated with HRP-coupled antimouse IgG1or IgG2a antibody at room temperature for 1 hour, fol-lowed by washing and the addition of tetramethyl benzi-dine (TMB) as a substrate. Optical density at 450 nm wasmeasured by an ELISA reader.

StatisticsData are shown as mean � SD. Statistical differences

in each group for the level of albuminuria, blood chemis-try, and pathologic score were calculated using the Mann-Whitney U test. Counted numbers of the glomerularinfiltrating cells were assesed by Bonferroni/Dunn test.Differences were considered significant when P � 0.05.

RESULTSTo elucidate the individual roles of the two activating

Fc�Rs (Fc�RI and Fc�RIII) in the induction of anti-GBM glomerulonephritis, accelerated anti-GBM glo-merulonephritis was induced in wild-type, Fc�RI�/�,Fc�RIII�/�, and FcR��/� mice, all with C57BL/6background or backcrossed to C57BL/6 for at least sixgenerations. The mice were preimmunized with rabbitIgG and then injected with rabbit anti-GBM antibodies4 days later, and were analyzed on days 7 and 14 afteranti-GBM antibody injection. The differences betweenseveral clinical and pathologic alterations were com-pared among these mice.

Serologic examination of Fc�R�/� miceSerum markers of renal damage were evaluated on

days 7 and 14. BUN values were increased on days 7

Fujii et al: FccR-induced glomerulonephritis 1409

and 14 in Fc�RI�/� mice (111.1 � 82.8 mg/dL, 219.0 �138 mg/dL, respectively) in a manner similar to wild-type mice (105.2 � 42.2 mg/dL, 131.3 � 64.6 mg/dL, respec-tively). On the other hand, BUN values in Fc�RIII�/�mice (31.1 � 7.0 mg/dL, 40.3 � 17.2 mg/dL, respectively)were almost the same as those in FcR��/� mice (29.5 �8.0 mg/dL, 28.8 � 6.6 mg/dL), significantly lower thanthose of wild-type and Fc�RI�/� mice, and they stayedwithin normal range or just around the upper limit (Fig.1A). Much like the BUN values, those of serum creatinine(Fig. 1B) and total cholesterol (Fig. 1D) of wild-type andFc�RI�/� mice were similarly elevated and were signifi-cantly higher than those of Fc�RIII�/� and FcR��/�mice. In stark contrast these increases upon anti-GBMantibody injection, the serum albumin levels decreasedsignificantly in nephritic wild-type and Fc�RI�/� miceto a similar degree, whereas those of Fc�RIII�/� andFcR��/� mice were not significantly reduced (Fig. 1C).These results indicate that renal impairment was inducedin Fc�RI�/� mice similarly to wild-type mice, whileFc�RIII�/� mice were resistant in a manner similar toFcR��/� mice.

Albuminuria in Fc�R�/� mice

We then measured albuminuria on days 7 and day 14following the disease induction to analyze the correlationwith renal impairment (Fig. 2). In wild-type mice, mas-sive urinary albumin excretion was induced as comparedto FcR��/� mice [wild-type vs. FcR��/� mice (mg/day),10.9 � 4.5 vs. 3.4 � 3.1 on day 7, and 8.9 � 3.5 vs. 3.2 �2.8 on day 14]. In Fc�RIII�/� mice, urinary albuminexcretion was somewhat less than in wild-type on day 7,although this difference became insignificant by day 14(Fc�RIII�/� vs. wild-type mice, 3.8 � 5.1 vs. 10.9 � 4.5on day 7, 5.1 � 6.9 vs. 8.9 � 3.5 on day 14). In contrast,urinary albumin excretion in Fc�RI�/� mice was similarto that of wild-type mice (Fc�RI�/� vs. wild-type mice,7.1 � 2.4 vs. 10.9 � 4.5 on day 7, 9.3 � 1.8 vs. 8.9 � 3.5on day 14). These results showed that albuminuria iscorrelated with the serum markers in these FcR�/�mice—Fc�RI�/� and wild-type mice showed severe im-pairment while Fc�RIII�/� mice showed very mild al-teration, similar to FcR��/� mice, suggesting that albu-minuria reflects the renal injury in this disease model.

Histologic alterations in Fc�R�/� mice

We analyzed histologic severity of renal damage in aneffort to comfirm that the blood chemistry and protein-uria results really reflect the renal damage. Paraffin-embed-ded kidney sections of Fc�R�/� mice were stained withPAS. Wild-type and Fc�RI�/� mice showed severe re-nal damage with mesangial proliferation, inflammatorycell infiltration, glomerular thrombosis, and crescent for-mation in the glomeruli on day 14 after anti-GBM anti-body injection. In contrast, FcR��/� and Fc�RIII�/�

Fig. 1. Blood chemistry of mice injected with anti-glomerular basementmembrane (GBM) antibodies. Serum samples were collected from indi-vidual mice on days 7 and 14 after anti-GBM antibody administration.Serum levels of blood urea nitrogen (BUN) (A), creatinine (B), albumin(C), and total cholesterol (D) were measured as described in the Meth-ods section. Data are presented as mean � SD of six mice in eachgroup. Symbols are: wild-type (�); Fc�RI�/�( ); Fc�RIII�/�(�);FcR��/�( ); and negative control ( ) which are mice received rabbitimmunoglobulin G (IgG) immunization but no anti-GBM injection.*P � 0.05; **P � 0.01.

mice were protected from severe glomerulonephritis,and Fc�RIII�/� mice showed only mesangial prolifera-tion and cell infiltration (Fig. 3A). The severity of glo-merulonephritis was graded in reference to these fourparameters: (1) intracapillary cellular proliferation, (2)glomerular sclerosis, (3) glomerular thrombosis, and (4)

Fujii et al: FccR-induced glomerulonephritis1410

Fig. 2. Albuminuria induced by anti-glomerular basement membrane(GBM) antibody administration. Urine samples were collected fromindividual mice (N � six female) in a metabolic cage on days 7 and 14after anti-GBM antibody administration, and the amount of albumin-uria was measured. The numbers indicate mean � SD (mg/day). Abbre-viations are: WT, wild-type; NC, negative control, mice receiving rabbitimmunoglobulin G (IgG) immunization but no injection of anti-GBMantibodies. *P � 0.05.

crescent formation as 0 to 4 for each according to modi-fied criteria by Raji et al [24, 25] and Kanamaru et al[26]. As shown in Figure 3B, the histologic score revealedthat Fc�RI�/� mice developed severe glomerulonephri-tis as well as wild-type mice, whereas Fc�RIII�/� miceshowed much milder glomerulonephritis. However, it isnoteworthy that Fc�RIII�/� mice still showed morechanges in glomerular, particularly intracapillary, prolif-eration compared with FcR��/� mice (Fig. 3B). Fromthese results, it can be concluded that the severity ofhistologic renal damage corresponded to the level ofBUN and serum creatinine values. These findings dem-onstrate that Fc�RIII plays a dominant role in the induc-tion of accelerated anti-GBM glomerulonephritis andalso that Fc�RI may play some additional role, asFc�RIII�/� mice exhibited a minor degree of impair-ment when compared with FcR��/� mice.

Antibody production in Fc�R�/� mice after rabbitIgG immunization

In order to determine the reason for the significantdifferences in glomerulonephritis inducibility betweenFc�RI�/� and Fc�RIII�/� mice, we first measured theproduction of anti-rabbit IgG antibodies in each mouseby immunization with normal rabbit IgG to investigatewhether the mouse groups produce similar antibody re-sponses required for the disease induction. The serumtiter of antirabbit IgG antibody in the individual micewas measured by ELISA. All mouse groups exceptFcR��/� mice exhibited similar levels of the antibody

response against rabbit IgG, suggesting that Fc�RI�/�and Fc�RIII�/� mice can produce normal levels of anti-rabbit Ig antibody (Fig. 4) and the antibody productiondoes not explain the difference of the inducibility be-tween these two mice. In FcR��/� mice, the kineticsof antirabbit IgG formation were delayed as comparedwith wild-type, Fc�RI�/�, and Fc�RIII�/� mice. Theantibody level in FcR��/� mice was significantly lowerat the early phase (day 7) after anti-GBM antibody injec-tion (data not shown), but the titer then increased onday 14, although not to the level of the other mousegroups (Fig. 4). However, as described below, theamount of antibody may be sufficient to bind to anti-GBM antibody in the kidney.

Local deposition of anti-GBM andautologous antibodies

The renal sections of each experimental group werestained with FITC-conjugated antirabbit IgG antibodyfor anti-GBM antibody to exclude the possibility thatthe resistance of FcR��/� and Fc�RIII�/� mice toglomerulonephritis was caused by insufficient binding ofrabbit antimouse GBM antibodies to the GBM (Fig. 5).The renal sections in all experimental groups werestained with FITC-antirabbit IgG antibodies almostequally along the GBM in a linear fashion after theinjection of antimouse GBM antibody. This proves thatthe resistance of Fc�RIII�/� and Fc R��/� mice toglomerulonephritis cannot be attributed to the failure ofthe binding of anti-GBM antibody to the GBM. Further-more, to detect the deposition of mouse antirabbit IgGantibody against rabbit IgG that had been bound tothe GBM in the autologous phase, renal sections werestained with Cy3-conjugated antimouse IgG antibody(Fig. 6). As described above, the serum level of antirabbitIgG antibody was lower in FcR��/� mice than in theother groups, but the deposition of mouse antirabbitIgG antibody in renal tissue was similarly observed. Thedifference of autologous antibody deposition in the renaltissue of FcR��/� mice was much less than expectedin the view of the serum level. Therefore, the titer differ-ence of anti-GBM antibody does not appear to be thereason for the failure of glomerulonephritis developmentin FcR��/� mice.

Fc�RI and Fc�RIII have different affinities for differ-ent IgG subclasses, which comprise immune complexes.Particularly, Fc�RI interacts preferentially with IgG2a-containing immune complex, while Fc�RIII prefers tointeract with IgG1 immune complex. Therefore, we inves-tigated such subclasses of autologous antibodies. In im-munohistochemical staining of renal sections, both IgG1and IgG2a autologous antibodies against rabbit anti-GBM antibody were equally deposited along the GBM,and we could not find a preference of deposition amongthe experimental groups (Fig. 7).

Fujii et al: FccR-induced glomerulonephritis 1411

Fig. 3. Histopathologic alterations of kidneysof Fc�R�/� mice by anti-glomerular base-ment membrane (GBM) antibody injection.(A ) Light microscopic appearance of the renaltissue stained with periodic-acid Schiff (PAS).Histopathlogic examination of kidneys inwild-type (A), Fc�RI�/� (B ), Fc�RIII�/�(C ), FcR��/� (D ) mice on day 14 after anti-GBM antibody injection or without injectionof anti-GBM antibody (E ), negative control.Wild-type (A) and Fc�RI�/� (B) miceshowed marked tissue damage such as intra-capillary cellular proliferation, glomerularthrombosis, and crescent formation. In con-trast, Fc�RIII�/� mice (C) exhibited mainlyhypercellular changes and were preventedfrom severe glomerulonephritis. FcR��/�mice (D) appeared almost normal. (B) Scoringof renal histology for intracapillary cellularproliferation (A), glomerular thrombosis (B),glomerular sclerosis (C), and crescent forma-tion (D), at day 14 after anti-GBM antibodyinjection. The scoring protocol is described inthe Methods section. Data are represented asmean � SD for six mice in each group. Sym-bols are: wild-type (�); Fc�RI�/�( );Fc�RIII�/�(�); FcR��/�( ); and negativecontrol ( ). **P � 0.01.

Fujii et al: FccR-induced glomerulonephritis1412

Fig. 4. Production of antirabbit immunoglobulin (Ig) antibody uponimmunization with rabbit IgG. Serum titer of antirabbit IgG antibodyin individual mice was measured by enzyme-linked immunosorbentassay (ELISA). On day 14 after rabbit IgG immunization, serum levelsof anti-rabbit IgG antibody were almost the same among wild-type (�),Fc�RI�/� (�) and Fc�RIII�/� (�) mice, except for a little lowerlevel in FcR��/� mice (�).

Fig. 5. Deposition of rabbit antimouse glomerular basement mem-brane (GBM) antibody along the GBM. Immunofluorescence stainingof the kidney in wild-type (A), Fc�RI�/� (B), Fc�RIII�/� (C), andFcR��/� (D) mice on day 14 after anti-GBM antibody injection. Therenal sections from these mouse groups were equally stained with fluo-rescein isothiocyanate (FITC)-conjugated antirabbit immunoglobulinG (IgG) in a linear pattern (200 magnification).

Fig. 6. Deposition of autologous mouse antirabbit immunoglobulin(Ig) antibody along the glomerular basement membrane (GBM). Immu-nofluorescence staining of the kidney in wild-type (A), Fc�RI�/� (B),Fc�RIII�/� (C), and FcR��/� (D) mice on day 14 after anti-GBMantibody injection. The renal sections from these mouse groups wereequally stained with Cy3-conjugated antimouse IgG antibody (200magnification).

Deposition of C3 on GBM in Fc�R�/� mice

In addition, we investigated whether there was any dif-ference in the binding of complement components in renaltissue among the experimental groups after injection ofanti-GBM antibody to clarify the contribution of comple-ments to the induction of the disease. Renal sections werestained with FITC-conjugated antimouse C3 (Fig. 8). Theresults showed that the C3 staining was equally observedon the GBM in the sections in a linear fashion in all themouse groups. It was demonstrated that the depositionof complement C3 was equally deposited regardless ofthe mouse type, and C3 binding is not responsible forthe inflammatory cascade leading to renal injury.

Characterization of glomerular infiltrating cells

To identify the type of inflammatory cells involvedin the induction of glomerulonephritis, we performedimmunohistochemical staining of infiltrating/recruitedcells in glomeruli at 3 hours (initial phase) and day 14(late phase) after administration of anti-GBM antibody(Table 1). Whereas PMN were recruited into glomeruliat the initial phase of glomerulonephritis in wild-type andFc�RI�/� mice, the PMN recruitment was minimallyobserved in Fc�RIII�/� mice. The recruitment of othercells, such as macrophage/monocytes, T cells and B cells,was not remarkable. At the late phase of glomerulo-

Fujii et al: FccR-induced glomerulonephritis 1413

Fig. 7. Immunohistochemical staining for immunoglobulin (Ig) G1 andIgG2a antibodies on renal tissues in all Fc�R�/� mouse groups. Thekidneys in wild-type (A and E ), Fc�RI�/� (B and F ), Fc�RIII�/�(C and G ) and FcR��/� (D and H ) mice on day 14 after anti-GBMantibody administration were stained with antimouse IgG1 Ab (A toD), or anti-mouse IgG2a antibody (E to H). Both antimouse IgG1 andIgG2a antibodies in the glomeruli were equally stained in each group(200 magnification).

nephritis, the number of those inflammatory cells, includ-ing PMN and macrophages, in the glomeruli was not dif-ferent among the experimental groups. The difference inthe initial infiltration of FcR-bearing PMN/macrophagesis involved in the differential induction of glomerulo-nephritis between Fc�RIII- and Fc�RI-deficient mice.

Fig. 8. Deposition of complement C3 along the glomerular basementmembrane (GBM). Immunofluorescence staining of the kidney in wild-type (A), Fc�RI�/� (B), Fc�RIII�/� (C ), and FcR��/� (D) miceon day 14 after anti-GBM antibody injection. The renal sections fromthese mice were stained equally with fluorescein isothiocyanate (FITC)-conjugated antimouse C3 antibody (200 magnification).

DISCUSSION

By establishing and analyzing FcR��/� mice, we dem-onstrated in the previous study that the activating Fc�Rs,Fc�RI and/or Fc�RIII, play a crucial role in the inductionof accelerated nephrotoxic glomerulonephritis. The pres-ent study advanced the analysis a step further in the identi-fication of the responsible Fc�R, revealing that the induc-tion of this model of glomerulonephritis is mediatedlargely through Fc�RIII from the finding that Fc�RI�/�mice showed severe renal damage at a similar level towild-type mice, whereas Fc�RIII�/� mice showed muchmilder changes compared with wild-type and Fc�RI �/�mice.

Fc�RI and Fc�RIII are both activating Fc�Rs associ-ated with a homodimer of the � subunit, but their affinitiesfor different subclasses of Ig and immune complex aredifferent [27]. As for the binding capacity to immune com-plex, Fc�RI preferentially interacts with IgG2a-containingimmune complex, but not IgG1 immune complex [28].In contrast, Fc�RIII can interact with IgG1 immune com-plex in addition to IgG2a. Contribution of either of theactivating Fc�Rs, Fc�RI or Fc�RIII, to the induction ofdiseases appears to vary depending on the disease model.In a serum-induced passive arthritis model (KRN model),it has been shown that anti-glysoxyl-phosphatidylinositol(GPI) antibodies induce the disease and Fc�RIII is the

Fujii et al: FccR-induced glomerulonephritis1414

Table 1. Various cells infiltration into the glomeruli in anti-glomerular basement membrane (GBM) glomerulonephritis

Wild-type Fc�RI�/� Fc�RIII�/� FcR��/� NC

Initial phase (3 hours)PMN 1.87�0.24a,b 1.82�0.09a,b 0.61 �0.08 0.45�0.11 0.38�0.05M/Mo 0.37�0.05 0.31�0.01 0.28�0.06 0.24�0.10 0.15�0.02T cell 0.05�0.01 0.07�0.04 0.06�0.01 0.06�0.03 0.04�0.01B cell 0.12�0.06 0.12�0.06 0.08�0.06 0.14�0.07 0.13�0.07

Late phase (14 days)PMN 0.79�0.15 0.83�0.12 0.53�0.16 0.59�0.14 0.05�0.01M/Mo 1.75�0.09 1.64�0.22 1.61�0.33 1.64�0.07 0.13�0.09T cell 0.15�0.06 0.10�0.01 0.13�0.04 0.13�0.09 0.05�0.02B cell 0.15�0.02 0.14�0.04 0.17�0.05 0.16�0.07 0.04�0.01

Various glomerular infiltrating cell counts in wild-type, Fc�RI�/�, Fc�RIII�/�, FcR��/� and NC mice. Cells were counted at 3 hours and 14 days afteradministration of anti-GBM antibody. Data are expressed at the mean � SD per glomeular cross-section. Abbreviations are: M/Mo, macrophage/monocyte; PMN,polymorphism cells.

aP � 0.01 vs. Fc�RIII�/� micebP � 0.01 vs. FcR��/� mice

dominant effector Fc�R for the induction probably throughactivation of neutrophils and mast cells. Fc�RIII maypreferentially function on these cells because IgG1 appearsto be dominated among anti-GPI antibodies [29, 30].Similarly, Fc�RIII plays an essential role in the inductionof a collagen-induced arthritis model [31]. In contrast,even in a model for arthritis, it has been reported thatFc�RI play a prominent role in inducing antigen-inducedarthritis [32].

In our actively induced nephrotoxic glomerulonephri-tis model, Fc�RIII plays a critical role in disease induc-tion. It is possible that qualitative and/or quantitativedifferences in immune depositions in glomeruli betweenFc�RI�/� and Fc�RIII�/� mice may influence the sus-ceptibility for anti-GBM glomerulonephritis. However,we failed to detect any apparent difference in the amountof autologous antibodies (mouse antirabbit IgG anti-body) bound to the heterologous antibody (rabbit anti-mouse GBM antibody) that had been deposited onto theGBM. Immunohistochemically, IgG1 and IgG2a antibod-ies against anti-GBM antibody were equally depositedalong the GBM among the experimental groups. In addi-tion, the degree of complement C3 deposition in theglomeruli of Fc�RI�/� and Fc�RIII�/� mice was similarto that of wild-type mice. From these results, the differen-tial sensitivity between Fc�RI�/� and Fc�RIII�/� micein the induction of glomerulonephritis may not be attrib-utable to the differential binding of different subclassesof anti-GBM IgG antibodies. However, since the compo-sition of IgG subclasses of anti-GBM antibody within theimmune deposits was hardly determined, we cannot ex-clude the possibility that the characteristics of the immunedeposits containing different IgG subclasses in local re-gions of glomeruli as well as kinetics of the binding areassociated with differential susceptibility to anti-GBMglomerulonephritis between Fc�RI�/� and Fc�RIII�/�mice.

Another possibility is that the differential suscepti-bility for glomerulonephritis between Fc�RI�/� and

Fc�RIII�/� mice may be due to the differences in theFc�R-bearing effector cells in the glomeruli in thesemice, and they are responsible for inducing glomerulo-nephritis. Immune complex–mediated disease may resultfrom the activation of residential tissue inflammatorycells and/or the recruitment of circulating monocytes orneutrophils to the sites of immune complex deposition.In the skin, resident FcR-bearing mast cells are necessaryfor initiation of immune complex–triggered Arthus reac-tion or vasculitis [4, 6, 33], whereas FcR-bearing alveolarmacrophages are likely to initiate the inflammatory re-sponse of immune complex–mediated alveolitis in thelung [7, 34]. In the kidney, it has been shown that bonemarrow–derived cells are required for anti-GBM glo-merulonephritis by bone marrow transfer experiments[35], and neutrophils and macrophage have been thoughtto have a potency to induce immune complex–mediatedglomerulonephritis.

In the classical model of anti-GBM glomerulonephritis[16, 18–21], heterologous anti-GBM antibody inducesglomerular damage through initial binding to the GBM,followed by activation of complements with PMN infil-tration into the glomeruli at the early heterologous phase.Deposition of fibrin and necrosis of the glomerular capil-laries were observed in severe cases [36]. There are sev-eral bodies of evidence to support the idea that neutro-phils play a critical role in the progression of renal injury.First, neutrophil-depleted mice or beige mice lackingneutrophil-neutral proteinase activity are protected fromglomerular injury [21, 37, 38]. Second, it has been sug-gested that the recruitment of glomerular neutrophils ismediated through Fc�R [21]. Third, the PMN influx bothat early and late phases of renal damage was minimal inFcR��/� mice [11, 22], and the accumulation of glomer-uler neutrophils was reduced in Fc�RIII�/� andFcR��/� mice [23, 39]. Consistent with these prevousstudies, our present study demonstrates that the PMNrecruitment was significantly suppressed in Fc�RIII�/�mice as compared with wild-type mice and Fc�RI�/�

Fujii et al: FccR-induced glomerulonephritis 1415

mice (Table 1). Taken together with the fact that Fc�RIis not expressed on neutrophils in the murine system [15],it is likely that Fc�RIII on glomerular neutrophils mayplay an important role in the recoginition of anti-GBMantibody on glomeruli and the induction of acceleratednephrotoxic glomerulonephritis.

Macrophages have also been thought to be importanteffector cells in glomerulonephritis [40]. In the rat modelof anti-GBM glomerulonephritis, macrophage accumu-lation is related to renal injury [41, 42]. However, inmouse model of accelerated Masugi nephritis, our obser-vation in Table 1 revealed that the infiltration of macro-phages into the glomeruli was not remarkable at theacute phase. In contrast, macrophage accumulation wasobserved also in FcR��/� mice at the late phase, sug-gesting that this accumulation is FcR-independent [22].Macrophages express both Fc�RI and Fc�RIII on thesurface, and produce proinflammatory cytokine such asinterleukin 1 (IL-1) and tumor necrois factor- (TNF-)and various chemokines upon the interaction of theseFcRs with immune complex. Since Fc�RI is occupied bymonomeric IgG2a in vivo, newly formed IgG2a-con-taining immune complex can hardly access Fc�RI [43],and only a high titer of specific IgG2a antibody canoccupy Fc�RI permanently [15, 44]. Since Fc�RIII playsan essential role in anti-GBM glomerulonephritis asdemonstrated in the present study, the pathogenic effectof macrophages, if there is a contribution, may be alsobe induced mainly through Fc�RIII.

In addition to the obvious requirement of bone mar-row–derived PMN and macrophage, it might be still possi-ble that intrinsic renal cells, especially activated mesangialcells, are involved in the induction of glomerulonephritisbecause they also express Fc�Rs [45] and mesangial cellsproduce proinflammatory cytokine through Fc�R [46, 47].It has been reported that the administration of anti-GBMantibody induced the expression of Fc�R, particularlyFc�RIII, in mesangial cells at the initial phase of glomer-ulonephritis, which is required for neutrophil recruit-ment and the production of chemokines such as mono-cyte chemoattractant protein-1 (MCP-1) from these cells.Together with the observation that in vitro stimulationof mesangial cells by immune complex induces MCP-1secretion through Fc�RIII [48], Fc�RIII activation by resi-dent kidney cells might be involved in the disease.

In our anti-GBM glomerulonephritis model, Fc�RIII�/�mice were similarly resistant to FcR��/� mice in termsof the induction of glomerulonephritis. However, it has tobe stressed that the former were not completely preventedfrom renal injury when compared with the latter. Thisrenal damage includes intracapillary proliferation andintravascular thrombosis. This observation in Fc�RIII�/�mice may be attributed to the involvement of eitherFc�RI or an unknown receptor other than Fc�RI andFcεRI, which is assembled with the � homodimer. Since

interferon-� (IFN-�) enhances the expression of Fc�RI,increased contribution of Th1 cells and macrophages inthe disease model might promote the contribution ofFc�RI. The greater contribution of macrophages overneutrophils may be correlated with the involvement ofFc�RI in any disease model. We are now generatingFc�RI/RIII double-deficient mice with the same back-ground, which should provide a clear answer to this ques-tion. In addition, we are trying to identify the cells actu-ally responsible for the induction of glomerulonephritisby cell transfer experiments.

As shown in this study, identification of the responsibleFc�R by the use of various combinations of Fc�R-defi-cient mice and responsible Fc�R-bearing effector cellsby cell transfer into such knockout mice will unveil com-plicated immune complex–mediated diseases, includingnephrotoxic glomerulonephritis.

ACKNOWLEDGMENTS

We thank Dr. K. Takase for discussion, Dr. S.-M. Chen for antibodypreparation, Ms. M. Sakuma and Ms R. Shiina for technical assistance,and Ms. H. Yamaguchi and Ms Y. Kurihara for secretarial assistance.This work was supported by Grants-in-Aid for Scientific Researchfrom the Ministry of Education, Science and Culture, and the Ministryof Health, Labor and Welfare Japan.

Reprint requests to Takashi Saito, Ph.D., Department of MolecularGenetics, Graduate School of Medicine, Chiba University, 1-8-1 Ino-hana, Chuo-ku, Chiba 260-8670, Japan.E-mail: saito@med.m.chiba-u.ac.jp

REFERENCES

1. Takai T, Li M, Sylvestre DL, et al: FcR gamma chain deletionresults in pleiotrophic effector cell defects. Cell 76:519–529, 1994

2. Clynes R, Ravetch JV: Cytotoxic antibodies trigger inflammationthrough Fc receptors. Immunity 3:21–26, 1995

3. Hazenbos WL, Gessner JE, Hofhuis FM, et al: Impaired IgG-dependent anaphylaxis and Arthus reaction in Fc gamma RIII(CD16) deficient mice. Immunity 5:181–188, 1996

4. Sylvestre DL, Ravetch JV: A dominant role for mast cell Fcreceptors in the Arthus reaction. Immunity 5:387–390, 1996

5. Sylvestre DL, Clynes R, Ma M, et al: Immunoglobulin G-medi-ated inflammatory responses develop normally in complement-deficient mice. J Exp Med 184:2385–2392, 1996

6. Watanabe N, Akikusa B, Park SY, et al: Mast cells induce auto-antibody-mediated vasculitis syndrome through tumor necrosis fac-tor production upon triggering Fc gamma receptors. Blood 94:3855–3863, 1999

7. Clynes R, Maizes JS, Guinamard R, et al: Modulation of immunecomplex-induced inflammation in vivo by the coordinate expres-sion of activation and inhibitory Fc receptors. J Exp Med 189:179–185, 1999

8. Ravetch JV: Fc receptors: Rubor redux. Cell 78:553–560, 19949. Deo YM, Graziano RF, Repp R, et al: Clinical significance of IgG

Fc receptors and Fc gamma R-directed immunotherapies. ImmunolToday 18:127–135, 1997

10. Daeron M: Fc receptor biology. Annu Rev Immunol 15:203–234,1997

11. Park SY, Ueda S, Ohno H, et al: Resistance of Fc receptor-deficient mice to fatal glomerulonephritis. J Clin Invest 102:1229–1238, 1998

12. Barnes N, Gavin AL, Tan PS, et al: Fc gamma RI-deficient miceshow multiple alterations to inflammatory and immune responses.Immunity 16:379–389, 2002

13. Sylvestre DL, Ravetch JV: Fc receptors initiate the Arthus reac-

Fujii et al: FccR-induced glomerulonephritis1416

tion: Redefining the inflammatory cascade. Science 265:1095–1098,1994

14. Meyer D, Schiller C, Westermann J, et al: Fc gamma RIII(CD16)-deficient mice show IgG isotype-dependent protection toexperimental autoimmune hemolytic anemia. Blood 92:3997–4002,1998

15. Ioan-Facsinay A, de Kimpe SJ, Hellwig SM, et al: Fc gamma RI(CD64) contributes substantially to severity of arthritis, hypersensi-tivity responses, and protection from bacterial infection. Immunity16:391–402, 2002

16. Masugi M: Uber das Wesen der spezifischen Veraderungen derNiere und der Leber durch das Nephrotoxin bzw das Hepatotoxin.Zugleich ein Beitrag zur Pathogenese der Glomerulonephritis undder eklamptischen Leber Erkrankung. Beitr Path Anat 91:82–112,1933

17. Benoit FL, Rulon DB, Theil GB, et al: Goodpasture’s syndrome.A clinicopathologic entity. Am J Med 37:424–444, 1964

18. Lerner RA, Glassock RJ, Dixon FJ: The role of anti-glomerularbasement membrane antibody in the pathogenesis of human glo-merulonephritis. J Exp Med 126:989–1004, 1967

19. Couser WG, Darby C, Salant DJ, et al: Anti-GBM antibody-induced proteinuria in isolated perfused rat kidney. Am J Physiol249:F241–F250, 1985

20. Shigematsu H: Glomerular events during the initial phase of ratMasugi nephritis. Virchows Arch B Cell Pathol 5:187–200, 1970

21. Schrijver G, Bogman MJ, Assmann KJ, et al: Anti-GBM nephritisin the mouse: role of granulocytes in the heterologous phase. Kid-ney Int 38:86–95, 1990

22. Suzuki Y, Shirato I, Okumura K, et al: Distinct contribution ofFc receptors and angiotensin II-dependent pathways in anti-GBMglomerulonephritis. Kidney Int 54:1166–1174, 1998

23. Coxon A, Cullere X, Knight S, et al: Fc gamma RIII mediatesneutrophil recruitment to immune complexes. A mechanism forneutrophil accumulation in immune-mediated inflammation. Im-munity 14:693–704, 2001

24. Raji L, Azar S, Keane W: Mesangial immuneinjury, hypertension,and progressive glomerular damage in Dahl rats. Kidney Int 26:137–143, 1984

25. Raji L, Dalmasso AP, Staley NA, et al: Renal injury in DOCA-salt hypertensive C5-sufficient and C5-deficient mice. Kidney Int36:582–592, 1989

26. Kanamaru Y, Nakao A, Mamura M, et al: Blockade of TGF-betasignaling in T cell prevents the developement of experimentalglomerulonephritis. J Immunol 166:2818–2823, 2001

27. Ravetch JV, Kinet JP: Fc receptors. Annu Rev Immunol 9:457–492, 1991

28. Gavin AL, Hamilton JA, Hogorth PM: Extracellular mutationsof non-obese diabetic mouse Fc gamma RI modify surface expres-sion and ligand binding. J Biol Chem 271:17091–17099, 1996

29. Ji H, Ohmura K, Mahmood U, et al: Arthritis critically dependenton innate immune system players. Immunity 16:157–168, 2002

30. Matsumoto I, Staub A, Benoist C, et al: Arthritis provoked bylinked T and B cell recognition of a glycolytic enzyme. Science286:1732–1735, 1999

31. Stahl TD, Andren M, Martinsson P, et al: Expression of Fcgamma RIII is required for development of collagen-induced ar-thritis. Eur J Immunol 32:2915–2922, 2002

32. Van Lent PL, Nabbe K, Blom AB, et al: Role of activatory Fcgamma RI and Fc gamma RIII and inhibitory Fc gamma RII in

inflammation and cartilage destruction during experimental anti-gen-induced arthritis. Am J Pathol 159:2309–2320, 2001

33. Baumann U, Chouchakova N, Gewecke B, et al: Distinct tissuesite-specific requirements of mast cells and complement compo-nents C3/C5a receptor in IgG immune complex-induced injury ofskin and lung. J Immunol 167:1022–1027, 2001

34. Baumann U, Kohl J, Tschernig T, et al: A codominant role ofFc gamma RI/III and C5aR in the reverse Arthus reaction. JImmunol 164:1065–1070, 2000

35. Tarzi RM, Davies KA, Robson MG, et al: Nephrotoxic nephritisis mediated by Fc gamma receptors on circulating leukocytes andnot intrinsic renal cells. Kidney Int 62:2087–2096, 2002

36. Assmann KJ, Tangelder MM, Lange WP, et al: Anti-GBM ne-phritis in the mouse: severe proteinuria in the heterologous phase.Virchows Arch A Pathol Anat Histopathol 406:285–299, 1985

37. Feith GW, Assmann KJ, Bogman MJ, et al: Lack of albuminuriain the early heterologous phase of anti-GBM nephritis in beigemice. Kidney Int 43:824–827, 1993

38. Schrijver G, Schalkwijk J, Robben JC, et al: Antiglomerularbasement membrane nephritis in beige mice. Deficiency of leuko-cytic neutral proteinases prevents the induction of albuminuria inthe heterologous phase. J Exp Med 169:1435–1448, 1989

39. Tang T, Rosenkranz A, Assmann KJ, et al: A role for Mac-1(CDIIb/CD18) in immune complex-stimulated neutrophil functionin vivo: Mac-1 deficiency abrogates sustained Fc gamma receptor-dependent neutrophil adhesion and complement-dependent pro-teinuria in acute glomerulonephritis. J Exp Med 186:1853–1863,1997

40. Nikolic-Paterson DJ, Atkins RC: The role of macrophages inglomerulonephritis. Nephrol Dial Transplant 16:3–7, 2001

41. Holdsworth SR, Neale TJ, Wilson CB: Abrogation of macro-phage-dependent injury in experimental glomerulonephritis in therabbit. Use of an antimacrophage serum. J Clin Invest 68:686–698,1981

42. Boyce NW, Holdsworth SR: Macrophage-Fc-receptor affinity:role in cellular mediation of antibody initiated glomerulonephritis.Kidney Int 36:537–544, 1989

43. Unkeless JC, Eisen HN: Binding of monomeric immunoglobulinsto Fc receptors of mouse macrophages. J Exp Med 142:1520–1533,1975

44. Fossati-Jimack L, Ioan-Facsinay A, Reininger L, et al: Markedlydifferent pathogenicity of four immunoglobulin G isotype-switchvariants of an antierythrocyte autoantibody is based on their capac-ity to interact in vivo with the low-affinity Fc gamma receptor III.J Exp Med 191:1293–1302, 2000

45. Radeke HH, Gessner JE, Uciechowski P, et al: Intrinsic humanglomerular mesangial cells can express receptors for IgG com-plexes (hFc gamma RIII-A) and the associated Fc epsilon RIgamma-chain. J Immunol 153:1281–1292, 1994

46. Hora K, Satriano JA, Santriago A, et al: Receptors for IgGcomplexes activate synthesis of monocyte chemoattractant peptide1 and colony-stimulating factor 1. Proc Natl Acad Sci USA 89:1745–1749, 1992

47. Gessner JE, Radeke HH, Uciechowski P, et al: Activation-depen-dent expression of Fc gamma-receptors on glomerular mesangialcells. Immun Infect 23:67–69, 1995

48. Radeke HH, Janssen-Graalfs I, Sowa EN, et al: Opposite regula-tion of type II and III receptors for immunoglobulin G in mouseglomerular mesangial cells and in the induction of anti-glomerularbasement membrane (GBM) nephritis. J Biol Chem 277:27535–27544, 2002