ADONIS 030094759100086N Scand. J. Immunol. 34, 81-89, 1991

Deposition of IgA is Associated with MacrophageInflux in the Kidney of Rats

R. K. STAD. W. M. .1. M. BOGERS. Y. MUIZERT. L. A. VAN ES &M. R. DAHA

Department ol" Ncphrology. University Hospital Leiden. Leiden. The Netherlands

Stad RK, Bogers WM JM. Muizert Y. van Es. LA. Daha M R. Deposition of IgA is Associatedwith Macrophage Influx in the Kidney of Rats. Scand J Immunol 1991:34:81-89. 1991

In the present study we treated rats with N-nitrosodimethylamine (DMN) or D-galaetosaminc(GALN) to achieve increased circulating IgA levels in rats. GALN-treatcd rats showed a six-foldinerease in serum IgA levels afler the firsl inlraperitoneal (i.p.) injection, whereas a 10-foldincrease after a second i.p. injection of GALN was seen. DMN-lreated rats showed a three-foldincrease in serum IgA levels. No differences were observed in IgG and IgM levels between treatedand non-treated rals. Sequential renal biopsies analysed by immunolluorescence exhibitedmesangial deposits of IgA wilh different intensities of C3 deposition. Rats trealed with GALNshowed more IgA deposition in the kidney ihan DMN-treated rats. The IgA deposition togetherwith C3 was more prominent in rats treated with GALN ihan in rats treated with DMN. Thedeposition ofC3 together wilh IgA was associated with an influx of monocyles as delected by ED-1. an antibody direcled againsi a rat monocyte marker. These studies provide evidence thai aninerease in serum IgA levels is assoeialed wilh depositii>n of IgA in ihe kidney and ihal IgA has aninflammatory potential,

R. K. Stad. Department of Nephrology. Bttlg. t. C3-P. University Hospitat Leiden. PO Box 9600.2300 RC Leiden. Vie Nethertands

Primary IgA glomerulonephritis (IgAGN), firsldescribed by Berger [1], is characterized by thepredominant mesangial deposition of IgA in thekidney. In order to ckirify the pathogenesis ofIgAGN. several experimental models have beendevised [2-7]. Recently, studies [8] demonstratedglomerular IgA deposits and increased serum IgAlevels in rats treated with N-nitrosodimethyla-mine (DMN) or D-galactosamine (GALN).

In rats polymeric IgA (plgA) is rapidly trans-ported from blood to bile by hepatocytes [9. 10].It wa.s suggested that liver damage induced byDMN. GALN [8] or carbon tetrachloride [3]resulted in decreased elimination of plgA byhepatocytes. A reduced clearance of IgApolymers could facilitate their persistence inblood and therefore iheir deposition in the kid-ney. It has been reported that in the rat. IgAimmune complexes can cause acute lung injuryand that this injury is dependent on the presenceof macrophages [11-13]. These studies suggest apathogenetic role for macrophages in IgA-mediated lung injury.

It has been shown that in humans, macro-

phages/monocytes play an important role in anumber of proliferative glonierulonephrilidis,including essential cryoglobulinaetnia. systemiclupus erythematosus and Henoch-Schonlein Pur-pura [14. 15]. In contradistinction to previousstudies [14. 15]. Yoshioka [16] recently showedincreased glomerular monocyte infiltration inpatients with IgA nephropathy. In the presentstudy we treated rats with DMN or GALN toachieve increased circulating IgA levels. In addi-tion, we investigated the inflammatory potentialof IgA in the rat kidney.

MATERIALS AND METHODS

Materiats. D-galactosamine, N-niirosodimethy la-mine, y-phenylene-diamine (Sigma Chemical Co.. StLouis, MO. USA). Sephacryl S-.̂ OO (Pharmacia. Wocr-den). NBCS (Gibco. Grand Island. NY. USA) slrepla-vidin peroxidase conjugate, slrcpliividin fiuoresceinisoihiocyanate (FITC) conjugate (Zymed LaboratoriesInc.. San Francisco. CA. USA) and mcrthiolale (Koch-Light Laboralories Lid, Colnbrook, Bucks, England)wore purchased as indicated.

Treatment protoeots. Following a slight modification

81

82 R. K. Stud et at.

of the approach described by Ogata [8]. experimentalIgA nephropathy was induced in inbred male Wislarrats weighing 225-250 g. Eight rals received an intraper-itoneal injection of 35 mg/kg body weighi of DMN as aL75"/ii solution in saline [al day ()]. in addition, theinjections were repeated al day 7 and day 14. One ratdied alter 9 days of treatment. The rats were sacrificed21 days after the start of ihe experimeni. Similarly eightrats received an intraperiloneal injection of 1.2 g/kgbody weight of GALN as a 30% solution in saline. Thisinjection was repeated after 7 days. Because GALN ismore loxic lo the animals ihan DMN. rals were killed 9days after the start of the experiment. Eight control ratsreceived intra peritonea I injections ol" the same volumeof phosphate-buffered saline (PBS) and were killed 21days afler ihc siart of ihe experiment. All biopsies ofGALN-lreated rats were taken on day 5 and day 9.Biopsies of DMN- and PBS-trcated rats were taken 5.9.14 and 21 days afler the siari of ihe experimeni.

ELISA for quantification of rat lg.-\ and rat IgM. Forquaniificalion of rat serum IgA or serum IgM the wellsin microiitre plates (Tilcrtek Flow Laboratories. Zwa-nenburg. The Netherlands) were eoated wilh allinity-purified rabbii IgG anii-rat IgA or antigen-specificrabbit IgG anti-rat IgM (both produced in our labora-tory) for 2 h at 37 C. .M'ler washing, the plalcs wereincubated wilh serial dilutions of serum samples for 1 hat 37 C. Because in rats almost 98'1'r. of ihe immuno-globuiins have the kappa lighl chain, the plates wereirealed with biotin-labcllcd monoeional mouse anti-ratkappa lighl chain [17] for IgA quantilicaiion. or wilhbiotin-labellcd polyclonal rabbit anli-ral IgM for IgMquantification. The plales were then treated with perox-idase-labcllcd streptavidin. and fj-phenylene-diamine asa substrate. Between the incubation sieps Ihe plateswere washed thoroughly with PBS 0.05",, Tween 20.The amount of reacted substrate was delected using aTiterlek Muhiskan (Flow Laboratories. Zwanenburg.The Netherlands) al OD 492 after ihc reaction had beenstopped with I N H;SO^. A standard curve wasconstructed using dilutions of rat IgA or rat IgM withknown conceniraiion. The amount of rai IgA or ratIgM in the samples was determined from ihe slandardcurve.

Quantification of rat IgG, Serum IgG was quantifiedby the single radial imtnunodilTusion method as de-scribed by Maneini et at. |18]. In short, radial immuno-diffusion plates were prepared using rabbii anli-rat IgGserum. Serum samples were diluted appropriately lo fitthe determined range ol" measurement. The scrum IgGconcentrations were obtained by standardizing theassay using dilutions of normal rat serum (NRS) with aknown amount of rat IgG.

Size anatysis of rat serum IgA. For size analysis ofserum IgA. 2 ml of four normal rat and four GALN-treated rat sera were applied to a ! .5 x 90 em SephacrylS-300 column equilibrated wilh veronal-buffered saline(VBS) containing 0.15 .w NaCl and 2 mM EDTA. RatIgA wtis detected in Ihe fractions by ELISA. Therefore,fractions wore eqtiaily diluted in PBS 0.05"/,i Tween-1% A new-born calfserum (NBCS) so direct compari-son of serum IgA between normal and GALN-treatedrats was possible. Rat IgM and ral IgG in the sera wasused as a reference.

Imnntnoftuorescence studies. Ral renal lissue wassnap-froxen in precooled isopentane and slorcd at- 7 0 Cor in liquid nitrogen. The biopsies were cut witha eryostat inlo 4-;jm sections and. after air drying,acetone-fixed for 10 min al 22 C. The slides werewashed twice for 7 min in PBS atid stained with FITC-conjugated rabbii anti-rai igA (S.iJ mg.ml; I-'.P ratio2.5; working dilution I; 100) and C3 (5.0 mg/ml; F/Pratio 2.6; working dilution I; IOO) [ i 9]. Rat IgG and IgMwere evaluated Indireclly by using biotin-conjugaiedrabbit anti-rat IgG and IgM followed by FITC-conju-gated streptavidin. For delecting macrophages and IgAsimultaneously, iwo-coiour immunofluoreseence wasperformed. Macrophages were deteeted with the use ofmnu.se anti-rat macrophage ED-1 (ascites diluted 1:500in PBS4-0.5"^ bovine serum albumin (BS.A). kindlyprovided by Dr C. D, Dijkstra. Free University ofAmsterdam. Amsterdam. The Netherlands) [20] andtetramethylrhodamin isothiocyanate (TRlTC)-labelledgoat anti-mouse IgG. As a control, eryostat sectionswere stained with secondary antibody alone. Theglomerular infiux of macrophages was determined bycounting the total numberofmacrophages in at leasl 20glomerular .sections per animal and expressed as theaverage number of macrophages per glomerulus. Thepresenee of granulocytes and T cells was evaluatedusing the W3,'I3 monoclonal antibody [21| followed byFITC-conjugated goat anti-mouse IgG. Between theincubation steps the slides were washed three limes for 7min in PBS. Photographs were taken on Scotch Chrome640-film. ASA-400 on a Leil/ microscope equipped witha 4-mm BG 38 + 5-mm BG 12 filler for FITC and aninterference 560-nm-f2-mm BG filter forTRITC,

Proieimiria. Proteinuria was assessed by measuring24-h urine samples containing 0.1 '".i Merthiolate by thebiuret method using rormiil iiorsc serum as a proteinstandard.

RESULTS

Serum immunoglobutin levels

The serum concentralions of IgA. IgG andIgM were measured in GALN-., DMN- or PBS-ireated rals over the period of study. GALN-Irealed rats showed a six-fold increase in IgAlevels in relation to control animals (0.32 + 0.15versus 0,05 ±0.01 mg/ml) afler the first intraperi-loneal (i.p.) injeetion, whereas there was a 10-foldincrease (0.53 ±0.17 mg/ml) after a seeond i.pinjection. DMN-trealed rats showed a three-fr'increase in IgA levels (0.13 ± 0.03 mg/ml) (Fig. I).No significant changes in IgG (PBS: 0.93 ±0.36;GALN; 1.18±0.39; DMN; 1.I2±O.25 mg/ml)and IgM (PBS: 0.24±0.06; GALN: 0.20±0.09;DMN; 0.20±0.07 mg/ml) levels were observedbetween treaied and non-ireated rals.

IgA-associated macrophage influx 83

day 2 day S day 7 day 9 day 14

FIG, I. Concentrations orserum igA (in mg/ml) in rats treated with GALN (O).DMN (D) and PBS (•) at 0, 2. 5. 7. 9 and 14 days after the start of iheexperiment. Each bar represents the mejin (+SD) of eight animals.

Size analysis of .serum IgA

Fraclionalion of normal ral sera by gelehromatography showed a distribution of IgAwhich consisted of 19% monomers and 81%polymers. Fraclionation of sera from rals Irealedwilh GALN revealed a substantial increase ofpIgA as compared to mIgA (97'Vi. and .1"/i,respectively) (Fig. 2).

Histological studies

Immunofluoreseent microseopy demonstratedmesangial IgA depo.silion of both GALN- andDMN-treated rats. However, rats irealed with

GALN showed more IgA deposition in thekidney than DMN-trealed rats (Fig. 3. C and E).No differences in IgG and IgM intensities wereobserved between treated and control animals(not shown).

The glomeruli of bolh GALN- and DMN-treated rals showed mesangial deposition of C3.GALN-treated rats showed more C3 depositionthan DMN-trealed rals (Fig. 3. D and F).Control rats showed no mesangial deposition ofIgA and C3 {Fig. 3. A and B). In addition,staining of sequential cryostal seetions for IgAand C3 showed a co-localization of bolh rat IgAand rat C3 in GALN-treated rats (Fig. 4). In

900

600-

300-

ao 1 GO

F R A C T t O N S

FIG, 2. Fractionation of serum of GALN- (•) and PBS- (O)treated rals by Sephacryl S-.lOO chromatography. In order tocompare the results with each other, 2 ml of each scrum wasapplied to the column. The tractions were then assessed for IgAin a dilution of 1/250 by ELISA.

IgA-associated macrophage infiu.x 85

FIG. 4. Immuno lluorescence staining for IgA (A) and d (B) in sequential kidney sections in glomeruli of GALN-treatedrats. Co-localization for IgA and C3 is noted. (All x 500),

FIG. 5, Immunofluorescence staining for macrophages (see arrows) in glomeruli obtained from PBS- (A) and GALN-(Bj treated rats. (All, x 500).

GALN-treated rats, the mesangial depositioti ofIgA together with C3 was associated with aninflux of monocytes as detected by ED-1. anantibody directed against a rat monocyte marker

(Fig. 5). In addition. ED-I positive cells were alsopositive for IgA as assessed by double immuno-fluorescence staining (Fig. 7). No fluorescencewas observed when cryostat sections were stained

FIG. 3. Immunofluorescence staining for IgA in glomeruli of separate kidney sections from rats treated with PBS (A),GALN ( O and DMN (E). Mesangial C3 deposition is noted mainly in GALN-lreated rats (D) and to a lesser extent inDMN-treated animals (F). C3 deposits are not seen in glomeruli of control rats injected with PBS (B), Theextragiomerular C3 staining observed in B. D and F is frequently encountered in kidney sections of older rats showingbasa! C3 production by proximal tubular epithelial cells. (All, x 500).

86 R. K. Stacletal.

— .2

day 0 day 5

FtG. 6. Number otmaciophages pergiomcrulus + SD(vertical lines) in GALN-(Ei). DMN- (D) and PBS- (•) treated rats. Macrophages were identilied bymonoclonal antibody ED-1 and quantified 0, 5 and 9 days after the start of theexperiment. Statistically significant differences iP<0.00\) were obtainedbetween day 0 (O). day 5 (*) and day 9 (*) Ibr the GALN group (two-tailed t-test).

with secondary antibody alone. Quantification ofmonocytes per glomerulus revealed a significantincrease of these cells in G.ALN-treated r<its.whereas no change in the number of monocyleswas detected in DMN-and PBS-treated rats (Fig.6). T cells and granulocytes were not detected inthe glomeruli of PBS-, GALN- or DMN-treatedrats.

The relevant results of the hislological part ofthe whole study are summarized in Table I.

Prateinuria

There weri; no significant differences in urinaryprotein levels among GALN-. DMN- or PBS-treated rats.

DISCUSSION

II is assumed that the liver plays a central role inthe homeoslasis of IgA [22-25]. In rats, liverdamage induced by int rape ritoneat administra-tion of GALN or DMN resulted in increasedcirculating levels of IgA and deposition of IgA inthe kidney [8]. Similar data have been found instudies where rats had been rendered cirrhotic bythe administration of carbon tetrachloride [3] andin rats which had undergone bile dtict ligation [6].In addition, it has been shown that IgA immunecomplexes can cause acute lung injury [12. 13]characterized by an influx of macrophages.Therefore, in an experimental model we exam-

ined the capacity of IgA lo attract macrophagesin the rat kidney.

In agreement with previous studies [8] we alsoobserved that injection ofGALN or DMN in ratsresulted in increased serum IgA levels (Fig. 1).The increase of serum IgA was more prominent inGALN-treated rats whereas DMN treatmentresulted in prolonged increase of circulating IgAlevels. In histological studies of liver tissue sec-tions performed by O '̂̂ ata et al.. it was shown thatDMN treattnent induced necrosis of both hepato-cytes and sinusoidal cells [8]. However, GALNtreatment restilted in hepatocelkilar necrosis asdemonstrated by Ogata [S) and Lichr [26]. Thiscould explain the observation that at day 7 serumIgA levels in GALN-treated rats gradually de-creased lo control levels. The administration ofGALN or DMN did not influence IgG and IgMserum levels.

It is generally accepted that a substantialfraction of rat scrum IgA consists of polymers[27-.1O]. A major fraction (81'^) of normal ratserum IgA was eluted from Sephacryl S-300 at aposition compatible with a molecular size largerthan 7S IgA. The elution pattern of serum IgA ofrats treated with GALN showed an increase ofIgA in high molecular weight fractions (Fig. 2). Inrodents, small IgA immune complexes arc clearedvia the plgA receptor {or membrane-bound secre-tory component) on hepatocytes [31]. This recep-tor mediates transport of intact plgA or smallIgA immune complexes containing pIgA into thebile (9. 32]. Thus, the increase of plgA in the

Ig A-as.fociated macrophage influ.x 87

j . 7. Double immunofiuurcscciiccsliiining fur Ig.A (A) and m:icrophtiges(B) in glomeruli obtaintd form GALN-treated rats, demonstrating ED-I posiiive cells containing rat IgA (sec arrow; all. x 500).

1. Summary of immunofluorescence studies ofPBS-. GALN- and DMN-treated rats

Rats{«) IgG IgM IgA C3 M0* T PMN+

PBS (8) - - - _ _GALN (8) - - + + + + + +DMN(8)J - - + + _

•Macrophages stained by ED-1 monoclonal :inti-body.

t T cells and polymorphonuclear cells stained byW3,/I3 monoclonal antibody.

J One rat died at day 9 during the experiment.

circulation may be attributed to hepatocellulardamage, as was also proposed by Gormly [3] andOgata [8].

Immunofluoresccnl studies revealed both IgAand C3 deposition in the glomeruli of rats treatedwilh DMN or GALN (Figs 3 and 4). The IgAdeposition together with C3 was more prominentin rats treated with GALN than in rats treatedwith DMN. These findings are at variance withthe results of Ogata [8] who showed markedglomerular IgA deposition only in the DMNmodel. Although the same doses of the toxicchemicals were injected, the dilTerence in outcomemay be explained by the repeated Injections ofGALN and DMN. As a result, the difference ofserum IgA levels between GALN- and DMN-treated rats is more prominent in our study. Thus

differences in circulating IgA levels might accountlor the findings in the immunofluorescencestudies.

The capacity of IgA to induce inflammation issuggested by the studies of Johnson et al. [11 -13].He reported that IgA immune complexes cancause acute lung injury. This model of injuryrequires the participation of the complementsystem and is characterized by an influx ofmacrophages. In vivo and in vitro studies sug-gested that lung injury was mediated by oxygen-derived metabolites produced by these cells. Inaddition. Rits el al. have shown that rat IgA isable to activate the complement system by thealternative pathway in a homologous system [33].The degree of complement activation is depen-dent on the molecular composition of IgA.Furthermore, Gorter et al, have shown thatpolymorphonuclear leucocytes are activatedupon incubation with staphylococci coated withIgA [34]. In addition, they also demonstratedbinding, intcrnalization and degradation of IgAby rat peritoneal macrophages [35]. In our studywe found IgA containing intraglomerular macro-phages as shown by two-colour imnuinofluores-cence.

In experimental models of IgAGN induced byeither bile duct ligation [6] or administration ofcarbon tetrachloride [3], mesangial IgA depositstogether with C3 were found. However, thesemodels did not examine whether there was a

88 R. K. Siail et ai

glomerula'' infiltration of inflammatory cells suchas macrophages. In the GALN model- increasedserum IgA levels resulted in mesangial depositionof IgA (Figs I and 3). In this model it is not clearwhether pIgA alone is deposited. This can only beapproached in a model in which quantitativedeposition of IgA in its various forms can bestudied separately. We are currently working onsuch a model. The co-localization of rat IgA andrat C3 {Fig. 4) was associated with an influx ofmonocytes (Fig. 6). Whether this influx is a directeffect of IgA or mediated indirectly via comple-ment is not clear at the moment. At present nospecific antibodies which recognize activationproducts of rat C3 are available including ratproperdin. Therefore it remains impossible forthe time being to establish whether the C3 seen inthe glotiicruli is intact C3 or activated C3.

In several studies no participalion of mono-cytes in the pathogenesis of IgAGN was observed[14. 15]. However, it has recently been reported byYoshioka [16] that monocytes were the pre-dominant cell type infiltrating glomcrtili inIgAGN and Henoch-Schonlein Purpura. In addi-tion, the studies described in the present paperand the studies of Johnson [11-13] have shownthat IgA deposition can lead to the increasedpresence of intraglomerular monocytes and anincreased number of macrophages in the lungrespectively. Whether an increased number ofglomerular monocytes results in glomerularinjury remains to be established.

In IgAGN. IgA is the major antibody compo-nent found in the mcsangium. As is suggested byEmancipator [36], co-deposits of IgG and IgMantibodies might be involved in thcncphritogcne-sisof IgAGN. Although glomerular IgCi and IgMwere not detectably different between control andtreated rats we cannot exclude participation ofminute amounts of IgG and IgM in the inductionof influx of macrophages. Furthermore, theabsence ofincreased deposition oflgG and IgM isin concordance with the unchanged levels ofcirculating IgG and IgM.

In conclusion, disturbance of liver function byD-galactosamine leads to increased levels of pIgAand deposition of IgA in the kidney. This IgAdeposition together with C3 was associated withan increased number of intraglomerular mono-cytes. These studies cotild be relevant for ourunderstanding of the pathogenesis of IgA neph-ropathy in humans.

ACKNOWLEDGMENTS

The authors would like to thank Mrs H. C. N,Kappcllc-de Vries. Mrs C. J. M. van der Voort-Van Dijk and Ms M. L. Kluiters for typing themanuscript. This study was supported in part bythe Foundation for Medical Research {MEDI-GON). which is subsidized by the NetherlandsOrganization for the Advancement of PureResearch (NWO).

REFERENCES

1 Berger J. Hinglais N. Lcs depots intercapillairesd'IgA [gG. J Urol Ncphrol 1968:74:694-5.

2 Emancipator SN. Gallo GR. Lamm ME. Experi-mental IgA nephropathy induced h.v oral immuniza-tion. J Exp Med 19H:i;L')7:572-82.

3 Gormly AA. Smith PS. Seymour AE. Clarkson AR,Woodruffc AJ. IgA giomcrular deposits in experi-mental cirrhosis. Am J Pathol 19X1:104:50-4.

4 Imai H. Nakamoto Y. Asakura K. Miki K., YasudaT. Miura AB. Spontaneous glomerular IgA deposi-tion in d(!Y mice: An animal model of IgA nephritis.Kidney Int 1985:27:756-61.

5 Isaacs K. Miller F. Lane B. Experimental model forIgA nephropathy. Clin Immunol Iinmunopathol1981:20:419 26.

6 Mclvin T. Burke B. Michael AF. Kim Y. Experi-mental IgA nephropathy in bile duct ligated rats.Clin Immunol Immunopathol 1983:27:369 77.

7 Rifai A. Small PA, Jr, Tcaguc PO. Ayoub EM.Experimental IgA nephropathy. J Exp Med1979:150:1161 73.

8 Ogata I. Fujiwara K. Nishi T, Kuwata S. Ohta Y.Nosaka K. Oka H. Conlribution of hepatic reticu-loendothelial system to glomerular IgA depositionin rat liver injury. Am J Pathol I988:I.M:411-44.

9 Jackson GDF. Lcmaitre-Coelho I. Vaerman JP,Beckers A. Biizin H. Rapid disappearance fromserum of intravenously injected rai myeloma IgAand its secretion into hile. Eur J Immunnl1978:8:123 6.

10 Vaerman JP. Lemaitre-Coelho I. Jackson GDF.Role of the liver in the rat intestinal s-lgA system.Adv Exp Med Biol, 1978:107:233 9.

! I Johnson KJ, Wilson BS. Till GO, Ward PA. Acutelimg injury in ral caused by immunoglohulin Aimmune complexes. J Clin Invest 1984:74:.358 69.

12 Johnson KJ, Ward PA. Kunkel RG. Wilson BS.Mediation of IgA induced lung injury in the rat: roleof macrophages and reactive oxygen products. LabInvest 1986:54:499-506.

13 Warren JS, Kunkel RG, Johnson KJ. Ward PA.Comparative O: responses of iung macrophiigcs andhlood phagoeytic cells in the ral: possible relevancelo IgA immune complex induced lung injury. LabInvest 1987:57:311 20.

14 Fcrrario F, Castiglione A.Colasanti G. BarhianodiBelgioioso G, Bcrtol S, D'Amico G, The detection

IgA-associated macrophage infiux 89

of monocytes in human giomerulo nephritis. KidneyInt 1985:28:513-19.

15 Nolasco FEB. Cameron JS. Hartley B, Coelho A,Hildreth G. Reuhen R. Intraglomerular T cells andmonncyles in nephritis: Sludy with monoclonalantibodies. Kidney Int 1987:31:1160 6.

16 Yoshioka K. Takemura T. Aya N. Akano N,Miyamoto H. Maki S. Monocyte infiltration andcross-linked fibrin deposition in IgA nephritis andHenoch-Schonlein purpura nephritis. Clin Nephrol1989:32:107 12.

17 Querinjean P, Bazin H. Beckers A. Deckers CHeremans JF. Milstein C. Transplantable immuno-globulin-sccreting tumours in rats. Purification andchemical characterization of four kappa chainsrrom LOU/W.sl rats. Eur J Biochem [972:31:354-9.

18 Mancini G. Carbonara AO. Heremans JF, Im-munochemical quantilation of antigens by singleradial immunoclifiiision. Int J Immunochem1965:2:235-54.

19 Daha RM. Stuffers-Heiman M, Kijlstra A, van EsLA. Isolation and characterization of the thirdcomponent of rat complement. Immunology1979:36:63-70.

20 Dijkstra CD. Dopp EA, Johng P. Kraal G. Theheterogeneity of mononuclear phagocytes in lym-phoid organs: distinct macrophage subpopulationsin the ral recognized by monoclonal antibodiesEDI, ED2and ED3. Immunology 1985:54:589-99.

21 Williams AF. Galfre G. Milstein C, .Analysis of cellsurfaces by xenogeneic myeloma-hybrid antibodies:differentiation antigens of rat lymphocytes. Cell1977:12:663-73.

22 Bogers WMJM. Gorter A. Stuurman ME. van EsLA. Daha MR. Clearance kinclics and tissue distri-bution of aggregated human serum IgA in rats,immunology 1989:67:274 80,

23 Brown WR. Kloppel TM, The liver and IgA:Immunological. cell biological and clinical implica-lions, Hepatology 1989:9:763 84.

24 Conley ME. Delacroix DL. Intravascular andmucosa! immunoglobuJin A: Two separate hutrelated systems of immune defense. Ann Intern Med1987:106:892-9.

25 Mestecky J. McGhee JR. Immunoglobulin A (IgA):

Molecular and cellular interactions involved in IgAbiosynthesis and immune response. Adv Immunol1987:40:153-245.

26 Lichr H, Grun M. Endotoxins in liver disease. In:Popper H, SchafTner. F. eds, Progres.s in liverdiseases. Vol. VI. New York: Grune and Stratton,1979:313-26.

27 Stechschulte DJ. Austen KF. Immunoglobuiins ofrat colostrum. J Immunol 1970:104:1052-62.

28 Vaerman J-P. Heremans JF. The IgA system of theguinea pig. J Immunol 1972:108:637 48.

29 Vaerman J-P. Andre C. Bazin H. Heremans JF.Mesenteric lymph as a major source of serum IgA inguinea pigs and rats. Eur J Immunol 1973:3:580-4.

30 Van Breda Vriesman PJC, Feldman JD. Rat -/Mimmunoglobulin: isolation and st)me biologicalcharacteristics. Immunochem 1972:9:525-34.

31 Underdown BJ, Schif!" MJ. Immunoglobulin A:Strategic defense active at the mucosal surface. AnnRev Immunol 1986:4:389 407.

32 Peppard J. Orlans E, Payne AWR. Andrew E, Theelimination of circulating complexes containingpolymeric IgA by excretion in Ihc bile. Immunology1981:42:83-9.

33 Rits M, Hiemstra PS, Bazin H. van Es LA. VaermanJP. Daha MR. Activation of rat complemeni bysoluble and insoluble rat IgA immune complexes.Eur J Immunol 1988:17:1873 80.

34 Gorter A. Hiemslra PS. Leijh PCJ, vd Sluys ME, vdBarselaar MTh. van Es LA. Daha MR. IgA- andsecretory IgA opsonized S. Aureus induce a respira-tory burst and phagocytosis by polymorphonuclearleucocytes. Immunology 1987:61:303 9.

35 Gorter A. Hiemstra PS. Klar-Mohamad N. van EsLA. Daha MR. Binding, internalization and degra-dation of soluble aggregates of human secretory IgAby resident rat peritoneal macrophages. Immuno-logy 1988:64:703-8.

36 Emancipator SN. Lamm ME. Biology of disease.IgA nephropathy: pathogenesis of t:ie most com-mon form of glomertilonephritis. Lab Invest1989:60:168-83.

Received 27 December 1990Accepted in revised form 7 March 1991