primary IgA nephropathy. circulating immune complexes in patients ...

0022-1767/88/1411-0l25$02.00/0 THEJOURNAL OF f M M U N O L o C Y Copyright 0 1988 by The American Assoclation of Immunologists

Voi. 141. 125-130, No. 1. July 1. 1988 Prfnted in U . S . A .

POLYMERIC IgA AND IgA RHEUMATOIDFACTOR DECREASE THE CAPACITY OF SERUM TO SOLUBILIZE CIRCULATING IMMUNE COMPLEXES IN PATIENTS WITHPRIMARY IgA NEPHROPATHY' PAOLO F. SCHENA,2*ANGEL0PASTORE," RENATO A. SINICO,'VINCENZO ALESSANDRO FORNASIERIt

MONTINARO,"

AND

Primary IgA nephropathy (IgAN) is characterized in the renal damage of human IgAN. by the presence ofimmunecomplexes (IC), high A putative pathogenic role has also been ascribed to levels of polymeric I@ (PI@), and I@ rheumatoid the presence of large amounts of true pIgA, Some being factor (RF)in the blood. The impaired -Pacity of in the form of IC, in the serumof patients with IgAN (6, Seruxn to solubilize IC in the presence Of 7). Circulating lymphocytes from secretory tissues after values ofC hemolyticactivity as as hi@ serum viral stimulation could lead to the production of large levels Of c3*c4*and proprdin factor have led us amounts ofpIgA which could than be deposited in the to whether PI@ and RF from patients renal mesangium, as judged by its affinity for human with I g A N where capable of inhibiting the capacity free secretory component (8, 9). of normal human serum to solubilize immune preIgA RF has been found by some investigators (10, 1 1) cipitates (BSA-anti-BSA) preformed at equivalence. in serum samplesfrom patients withIgAN and, recently, The results showed a significant reduced mean capacity ofserum from patients with I ~ toNsolu- pIgA with RF activity has been described in these patients bilize "in vitro" immune precipitates (p < 0.001)and (12). Finally, elevated serum levels Of antibody idiotypes significanthighmean levels of iP 0.001) and glYadine, and casein) IgA RF (p 0,005)in the blood. Increasing amounts against c0mmOn dietary Ag have been observed in patients withIgAN (13, 14). of P I ~ A inhibited solubilization of IC in the fluid The possible relationship between these factors has phase, and inhibitory activity was also shown by the I g A RF.There were inverse correlations between been debated over the last few years, but without any pIgA and the capacityof serum to solubilize IC (r = pathogenic role of these factorsbeing elucidated. -0.36:p 0.05). and between I@ RF and the cornIt is known that fresh human serum solubilizes preplement-mediatedsolubilization (t = 4 - 5 7 ; p formed immune precipitates"in vitro" through the c SYSo*ool)-It is that PI@ and RF may be tem (15). Furthermore, it has been established that C responsible for the impaired complement-mediated modifies the size of vitro,99because large amounts Of Serumand the persistence Of of activated fragments of C3 (C3b). incorporated in the uble nephritogenic IC in the Of patients with lattice of IC, cause the disruption of IC into smaller agprimary IgAN.

gregates. This CMS seems to be mostly dependent upon the integrity of the alternative C pathway (16).Impairprimary IgAN3 (Berger's disease) is characterized by the ment of this function is due tolow levels of C in patients C3 and other Ig deposits with systemic diseases (lupus and vascupresence of IgA and frequently in the glomerular mesangium. litis) (17). whereas IgM RF inhibits the CMS in serum A pathogenic role has been attributed to the circulating from patients with rheumatoid arthritis (18) and essenIgA IC, and some investigators (1-4) have correlated their tial mixed cryoglobulinemia (19).Impaired CMS has been levels with episodes of gross hematuria. Nevertheless, reported in Serum from patients with primary IgAN in others (5)have shown that IgA-IC persist irrespective of the Presence of normal levels of c (20).The Present clinical nephritis, while IgG and IgM-IC may be implicated study. therefore, wasUx"taken to determine the Serum factors which were present in IgAN patients and were capable of inhibiting the capacity of normal human Received for publication November 24. 1987. Accepted for publication April 8 , 1988. serum to solubilize "in vitro" the immune precipitates The costs of publication of this article were defrayed in part by the preformed at equivalence. payment of page charges. This article must therefore be hereby marked aduertlsernent in accordance with 1 8 U.S.C. Section 1734 solely to indiThe findings showthat high levels ofpIgA and IgA RF cate this fact. are present in the blood of patients withIgAN. In addition, 'Thls work was supported in part byGrant 86-960 from Minister0 della Pubblica Istruzione and by Grants 85-00541 and 86-00145 from there is a reduction in the capacity Of serum to the Consiglio Nazionale delle Richerche (CNR). Roma. Italy. IC in a high percentage of thesepatientsinthepresence of normal values of hemolytic c and high levels of 'Address all correspondence and reprint requests toProf. F. Paolo

Italy.

Schena. Chair of Medical Therapy, Polyclinic, Piazza G. Cesare 1 1 , 70124 Bari. components. C A relationship exists between pIgA and AbbreviationSused in this paper: W N . primary I g A nephropathy:IC, the capacity of serum to solubilize IC. Increasing amounts immune complexes: pIgA.polymeric I g A : RF, rheumatoid factor: CMS, complement-mediatedsolubilization:VBS, Veronal-bufferedsaline: NHS. Of pigA inhibit the CMS* inhibitory activity I s normal human sera; mIgA. monomeric IgA. shown by IgA RF, suggesting that pIgA and IgA RF are

125

126

REDUCED CMS IN IgA NEPHRITIS

fromhuman whey by affinity chromatography. Briefly, human monoclonal IgM were covalently coupled to Sepharose 4 8 (Pharmacia). The IgM-Sepharose was incubated for 2 h at 4°C with diluted whey. after which the mixture was poured into a chromatographic column. The absorbed protein was eluted with 1.0 M potassium MATERIALS AND METHODS thiocyanateand 0.01 M potassiumphosphate, pH 7.0. dialyzed Patients. We studied 29 patients (22 male and 7 female) with against PBS, and filtered on a SephacrylS200 (Pharmacia)column. primary IgAN. The diagnosis was based on the presence of IgA in Two peaks wereobtained, the second containing puresecretory the glomerular mesangium, with or without other Ig and C3, on component. as judged by immunoelectrophoresis and double immudirect immunofluorescence of renal biopsy specimens. All patients nodiffusion. using anti-whole human milk (Nordic Immunology, Tilhad negative tests for nuclear antibody, hepatitis B Ag, cryoglobu- burg, The Netherlands) and anti-secretory component (Dakopatts. lins. and impaired liver function. The age of the patients ranged Copenhagen, Denmark) antisera. from 17 to 64years. Microtiter plates (Dynatech Deutschland GmbH. Plochingen. W. Controls. Blood samples from medical students, staff members, Germany) were coated overnight a t 4°C with a 6 pg/ml solution of and healthy blood donors. matched for age and sex with IgAN papurified free secretory component in 100 pl of PBS followed by PBS tients. were usedas controls. Neitherpatients nor controls had takencontaining 0.2% human serum albumin(Sigma, St. Louis, MO) and steroids or immunosuppressive drugs during the time of the study. 0.2% gelatin (Difco Laboratories, Detroit, MI). Individual column Serum samples. Venous blood samples were allowed to clot in a fractions or samples were addedto thewells and incubatedovernight sterile pyrogen-free test tube (Vacutainer. Becton Dickinson. Ruth- a t 4°C. After three washes with PBS-Tween. the bound pIgA were erford, N J ) at 37°C for 2 hand were centrifuged at 2000X g a t room measured by the addition of peroxidase-conjugated F(ab'), fragment temperature for 20 min. Serum sampleswere storedin small aliquots goat anti-human IgA anti-serum followed by four more washes with at -70°C until tested. All serologic studies were performed in double the specific substrate. Thirty minutes later, the color reactions was blind trials. measured by a Titertek Multiskan [Flow Laboratories, Rockville, MD) Preparation of radiolabeled immune precipitates (BSA-antia t 405 nm. The results were expressed as the OD readings. The BSA). BSA (Sigma Chemical Co, St. Louis, MO) was radiolabeled with mean value of serum pIgA in healthy subjects was 0.22 +. 0.09 (SD). Iz5I by the chloramine-T method (21).Sp. act. was1 x lo6 cpmpg of Zsotation and detection of IgA RF. IgA RF wasisolated from the protein. Increasing amounts of radiolabeled BSA were incubated serum of a patient with IgAN by incubation with human 1gG covawith constant amounts of rabbit anti-BSA IgG eluted by DEAE lently bound to Sepharose4B (Pharmacia, Uppsala. Sweden)and by cellulose chromatography. The mixtureswere lefta t 37°C for l h and elution with 0.1 M glycine-HC1at pH 2.8. IgG and IgM were removed at 4°C overnight. After centrifugation at 2000x g for 20min a t 4°C. by incubating the eluates with protein A Sepharose CL-4B (Pharthe precipitatesobtained a t equivalence (molar antibody to Ag ratio macia. Uppsala, Sweden)and with anti-IgM, respectively. The purity = 5.1) were washed once in cold 0.01 M PBS a t pH 7.2. and resusof IgA RF was checked by double immunodtffusion using anti-whole pended in VBS supplemented with 0.15mM Ca2+and 0.05 mM Mg2+ human serum, anti-IgG. -1gA.-1gM. -K and - X chain antisera (Beha t pH7.2. They were dispersedthrough a 26-gauge needleand stored ringwerke AG. Marburg. W. Germany) and by ELISA ( 12).To avoid in aliquots containing2 pg of immune precipitates. interpretative problems, only monomeric IgA RF, as judged by gel CMS. The capacity of serum to solubilize immune precipitates chromatography, wasused in the solubilization process.The 1gA RF ('"I-BSA-anti-BSA) was tested by using the technique described assay was performed with the ELISA technique as described previpreviously (20). Thesolubilization test wascarried out by incubating ously (12, 22). with minor modifications.Microplate wells (Dynatech 2 pg of a 1/2 dilution of the serum sample in theVBS a t pH 7.2 for Deutschland GmbH, Plochingen. W. Germany) were coated overnight 2 h a t 37°C. The reaction was stopped by adding 1 ml of ice-cold a t 4°C with a 10 pglrnl solution of normal human IgG in 100 plof saline. After centrifugation a t 2000 X g for 20 min at 4°C. the PBS. The well were washed three times with PBS-Tween and then were blocked by incubation withPBS containing0.2% HSA and 0.2% radioactivity of both the pellet and the supernatant was determined. The percentage of solubilization was calculated by the following gelatin for 2 h a t room temperature. After three more washes, 100 formula: (total cpm - pellet cprn)/total cpm X 100. All tests were plof either standards or samples were added, and the plates were carried out in duplicate, and when the results of duplicate tubes incubated in a moist chamber a t room temperature for 18 to 20 h. varied by more than 5%of the mean, the assay wasrepeated. The The standardsused were multimeric IgA RF purified fromthe serum of a patient with IgaM as described above a n a monoclonal polymric coefficients of variation (standard deviation -SD- in percent of the arithmetic mean) for one sample tested intriplicate on a single day IgAk purified from the serum of a patient with IgAk-IgG essential mixed cryoglobulinemia by gel chromatography at acid pH. Plates and for four samples tested on 3 different dayswere 0.85 and 2.85, were washed three times in PBS-Tween, and 100 pl of peroxidaserespectively. Heat-treated serum (56°C 30 min), used as negative conjugated F(ab')a fragment goat anti-human serum IgA (Cappel). control, showed a spontaneous solubilization (2 to 8%)that was diluted 1/2000 in PBS-Tween-HSA 0.1%. were added to the wells. assumed as background. To better compare the results of different tests, the values of normal pathologic serum sampleswere expressed After incubation for 1 h a t 37°C. the wells were washed four times, as percentage of solubilization of a normal human serum pool fol- and 100 plof the substrate were added (ABTS-Hydrogen Peroxide. lowed by correction of the background according to the following Kirkegaard and Perry Lab. Inc. Gaithersburg. MD). After incubation for 30 min a t room temperature, the absorbance of each well at 405 formula: nm was read with a Titertek Multiskan (Flow Laboratories. Rockppt BSA with 56°C serum ville, MD) interfacedto a n Apple IIE computer to automate the - ppt BSA with serum sample estimations. The results were expressed a s pglml. The mean value x 100 % PNHS = of IgA RF in healthy subjects was 7.7 f 5.0 (SD). ppt BSA with 56°C serum Detection of IgA and 1gG antibodies to BSA. Micro-ELISA plates - ppt BSA with serum pool (Dynatech) were coated overnight a t room temperature with 50 ml BSA (Sigma). The plates. treated as in the abovementioned where ppt = precipitate and PNHS = normal human serumpool. The mean value for the CMS f SD = 95.9 k 4.4 and wasobtained from technique, were incubated with samples for 3 h a t room temperature. After washing, 50 pl of peroxidase-conjugated goat anti-human a-or the values found in 40 serum samples from healthy blood donors. y-antiserum (Cappel Laboratories, Downingtown. PA.) diluted I / The results were expressed as a percentage of the CMS obtained 1000 were added. After incubation for 1 h a t 37°C. the wells were from a pool of NHS. Isohtlon of d g A a n d pIgA. IgA were purified from the serumof washed and the substrate added. After incubation for 30 min a t read a patient with lgAN (and no IgA RF activity) by incubation with a room temperature, the absorbanceof each well at 405 nm was rabbit anti-human IgA antiserum (Dako. Immunoglobulins A/S. Co- with Titertek Multiskan (Flow Laboratories, Rockville. MD). Positive penhagen. Denmark)covalently bound to Sepharose4 8 (Pharmacia. controls consisted of patients with celiac disease and high titers Of Uppsala, Sweden) and by elution with 0.1 M glycine-HCl, pH 2.8. anti-BSA antibodies (23). Gel chromatography study.This wasperformed with a n Ukrogel The eluate was dialyzed against PBS, concentrated by Aquacide, and ACA 22 column (2.5 x 80 cm. LKB. Bromma, Sweden) equilibrated its purity was checked by immunoelectropheresis. Subsequently, mIgA and pIgA were purified by HPLC. Molecular with 0.01 M PBS (pH 7.4). Human IgG (m.w. 150,000) andIgM (m.w. sieving HPLC was performed by using a TSK 4000 column connected 900,000) were utilized as m.w. markers. Three milliliters of each serum sample were eluted with the same buffer and IgG. IgA, IgM to a HPLC pump (LKB. Broma, Sweden). The HPLC system was previously calibrated with IgM (m.w. 900,000)and IgG (m.w. were assayed in the collected fractions with a laser nephelometer utilizing the appropriate monospecific antisera (Behring. Scoppito. 150.000). Flow rate was 0.6 ml/min by using 0.1 M sodium phosItaly). Fractions containing I g A were pooled according to their m.w. phate (pH 6.8).The collected fractions were assayed for totalIgA and (150.000 to 400.000: 400,000 to 700,000; 700.000 to 1 . 0 0 ~ . 0 0 ~ : pIgA by the techniquedescribed below. =-1.000,000) and concentrated (lox): the presence of pIgA and IgA Detection of plgA. The secretory component solid phase assay RF was assayed with theELISA technique. Aliquots of pooled fracwas used to detect pIgA (12).The secretory component was purified

responsible for the presence of insoluble nephritogenic

IC in the blood of patients with primaryIgAN.

"

127

REDUCED CMS IN IgA NEPHRITIS

tions were mixed 1 :2 with NHS as a C source, and levels of IC-like mean value of the CMS in 42 serum samples from 29 material were evaluated by a conglutinin-binding assays. patients with primary IgAN was significantly low ( p < Inhlbftton of the CMS. In order to estimate if eluted sera contained 0.001). In contrast,the meanvalues for CHlOO and substances inhibiting CMS. 30 $1 of each poolof fractions were mixed with 200 pl of NHS (diluted 1/2 with VBS). The mixture was APl 00 did not differ significantly from the mean values immune precipitates observed in the controls. The mean serum levels of C3, then incubated with2 pg of 1251-BSA-anti-BSA and the serumsolubilization capacity was evaluated as mentioned C4, and B were significantly increased( p < 0.001). above. The possibility that pIgA may inhibit the CMS was studied by Serum plgA and antibodies.Serum samplesfrom pamixing increasing amounts of purified pIgA with 200 pl of NHS tients with IgAN showed high mean values of pIgA ( p < (diluted 1/ 2 with VBS). Each mixture was incubated 37°C at with 2 pgof preformed radiolabeledimmunoprecipitates. The kinetic curves0.001)and IgA RF ( p < 0.005) (Fig. 1). IgG and IgA antiof the serum solubilization capacity were obtained by sampling 25 BSA antibodies were absent in all serum samples (data plof the mixture a t selected time (10, 20, 30.60.and 90 min) and not shown). mixing them with 500 plof ice-cold PEE to stop the solubilization Inhibitory actiuityof pIgA and IgA RF on the CMS of reaction. After centrifugation a t 2000 x g for 20 min at 4°C. the supernatants andpellet were assayed for '"1 activity. The percent- a N H S . Figure 2 shows the inhibitory effect of increasing age of CMS was calculated by the following formula: (total cprn - amounts ofpIgA on the CMS of NHS. Inhibition of the pellet cpm/total cpm) x 100. The kinetic curve of the NHS, treated CMS was observed when increasing amounts ofpIgA in the same way but without plgA. was used as control. Other methods. Laser nephelometry was used for the measure- were added to the serum in presence of the BSA-antiment of serum IgG. IgA, and IgM. The hemolytic activities of the BSA immune precipitates. A dose-dependent inhibition classical (CH1 and 00)alternative(AP100) Cpathway were measured of this function was seen. In contrast, mIgA in the same by single radial immunodiffusion methods. as described elsewhere amounts did not inhibit the CMS (Fig. 3).In order to verify (20). C3. C4. and properdin factor B serum levels were determined that thiseffect on CMS was due to direct C consumptlon, by a radial immunodiffusion technique in agar plates(24). Circulating IC were detected by the amplified conglutinin binding assay the anti-complementary activity of purified pIgA was using monospecific anti-IgG, anti-IgA, and anti-IgM antisera (Dako, determined. The percentage of hemolytic activity of the Immunoglobulins A/S. Copenaghen. Denmark] in accordance with classical and alternative C pathway of NHS was measthe technique of Barnet et al. (25).Values were expressed in ng of the bound 1251-proteinA. ured after theaddition of purified pIgA at different conStatfstfcal analysis. Data from the groups of patients are given centrations (50 and 150 pdml). Polymeric IgA had neglias the meank SD. Significances of dffferences between groups were gible direct anti-complementary activity (data not determined by the Student's t-test and the Mann-Whitney U-test. Correlation coefficients wereobtained by using the Spearman ranktest. RESULTS

Capacity of serum to solubilize immune precipitates in vitro. The results of the CMS study and theC activity values are given in Figure 1 and Table I, respectively. The

10 20 30

60

Incubation (*in)

Ffgure 2. Klnetics of the inhibitory effect of varying amounts of plgA on the CMS of BSA-anti-ESA complexes. Two hundred microliters of NHS diluted 1/2with VBS were added to50 p&ml of plgA (A). to 150 p g h l of pIgA (0).The percentage ofCMS was obtained at different intervals according tothe standard procedure. Each point represents the mean and S D of four determinations. Percentage ofCMS in fresh NHS 10) and heated NHS (0). o(5

PI@

0.001 0.001 Figure I . The values of CMS. pIgA and IgA RF in serum samples from 29 patlents with primary IgA nephropathy. Values are expressed as percentages of pooled NHS (%PNHS) for CMS, as OD for pIgA and as p g / ml for IgA RF. The enclosed areas represent the normal range In controls (mean & 2 SD). Bars indicate the mean values in patients. p values are

'YI

shown. TABLE I Study of the C system in 29 IgAN patfents Method

No. of Serum

lgAN Patlent9 Normals"

CH 100

18

93.25 20.1

APlOO c3

16 33

99.1217.0

c4 €3

120.9+ 33 33 119.2

110.6520.1 37.1 f

20.5

P*

99 2 25 101 2 1 2 96-Cl5 90 2 17 89 2 25

NS NS

0.001 0.001

0.001

Means 2 standard deviations are expressed as percentages of pooled NHS. Student's t-test.

1ncub.tia (.in)

Ftgure 3. Kinetics of the effectof varying amounts of mlgA ( A , 50 p~ ml: 0. 150 &ml) on the solubllizatlon process of Immune complexes IESA-anti-BSAl. Procedureand amounts were carried m"_"a t 2s in "_ Fiu~rlre - -0" - 3 ~~

" " ~"

"

128

REDUCED CMS IN I g A NEPHRITIS

shown).Inhibition of the CMS wasalsofoundwhen increasing amountsof IgA RF (15,30,and 60 pg/ml) were added to the NHS in thepresence of immune precipitates (BSA-anti-BSA) formed at equivalence. Theinhibitory activity was dose-dependent (Fig. 4). There was a significant inverse correlation between the values of CMS and the serumconcentration of pIgA (r = -0.36;p < 0.05; Fig. 5a).However, there were some serum samples in which the CMS was very low as well as the pIgA serum level and vice versa. There wasa n inverse correlation between the levels of IgA RF and the CMS values (r = -0.57; p