IgA1-containing immune complexes in IgA nephropathy differentially affect proliferation of mesangial cells

Kidney International, Vol. 67 (2005), pp. 504–513

IgA1-containing immune complexes in IgA nephropathydifferentially affect proliferation of mesangial cells

JAN NOVAK, MILAN TOMANA, KAREL MATOUSOVIC, RHUBELL BROWN, STACY HALL, LEA NOVAK,BRUCE A. JULIAN, ROBERT J. WYATT, and JIRI MESTECKY

Departments of Microbiology, Medicine, and Pathology, University of Alabama at Birmingham, Birmingham, Alabama;Department of Pediatrics, University of Tennessee Health Sciences Center and Children’s Foundation Research Center at the LeBonheur Children’s Medical Center, Memphis, Tennessee

IgA1-containing immune complexes in IgA nephropathy dif-ferentially affect proliferation of mesangial cells.

Background. Sera of patients with IgA nephropathy (IgAN)contain circulating immune complexes (CIC) composed ofgalactose-deficient IgA1 complexed with antiglycan antibod-ies. The role of these CIC in the pathogenesis of IgAN is notknown.

Methods. We studied how proliferation of cultured mesan-gial cells (MC) is affected by CIC prepared from sera of IgANpatients and healthy control subjects using size-exclusion chro-matography. CIC-containing fractions were added to serum-starved MC in culture, and cell proliferation was measuredusing 3H-thymidine incorporation. The results were confirmedby staining MC using an antibody against proliferating cell nu-clear antigen.

Results. The incubation of starved MC with serum fractionswith Mr 800 to 900 kD, rich with galactose-deficient IgA1, stimu-lated proliferation, while fractions with smaller complexes wereinhibitory. Furthermore, CIC-containing larger molecular massfractions isolated from serum of an IgAN patient collectedduring an episode of macroscopic hematuria stimulated MCproliferation more than CIC obtained during a subsequent qui-escent phase. To examine the role of IgA, we removed IgA1from serum before fractionation. The resultant IgA1-depletedfractions were devoid of stimulatory IgA-CIC. Sera of IgANpatients were also fractionated after addition of desialylatedgalactose-deficient polymeric IgA1 to form additional immunecomplexes. Supplementation with a small quantity of this IgA1increased cellular proliferation in assays using serum fractionsof Mr ≥800 to 900 kD; uncomplexed IgA1 did not affect MCproliferation significantly. In contrast, supplementation with alarger quantity of this IgA1 inhibited cellular proliferation inassays using serum fractions of Mr 700 to 800 kD.

Dr. Matousovic’s current address is the Department of InternalMedicine, School of Medicine 2, Charles University, Prague, CzechRepublic.

Key words: glomerulonephritis, glycosylation, O-linked glycans, im-munoglobulin A (IgA), autoimmunity.

Received for publication October 10, 2003and in revised form May 19, 2004, and July 29, 2004Accepted for publication August 18, 2004

C© 2005 by the International Society of Nephrology

Conclusion. Overall, these findings suggest that CIC contain-ing aberrantly glycosylated IgA1 affect proliferation of MC invitro and, thus, likely play a role in the pathogenesis of IgAN.

IgA nephropathy (IgAN), the most common typeof glomerulonephritis worldwide, is characterized bymesangial deposits of IgA1, often with codeposits of C3,and IgG or IgM, or both [1–5]. Proliferation of mesan-gial cells (MC) and expansion of extracellular matrix arefound in patients with mild clinical disease, but progres-sive glomerular and interstitial sclerosis leads to end-stage renal disease in 30% to 40% of patients within20 years after diagnosis [1]. IgAN recurs in approximately50% to 60% of renal allografts [6–8]; in contrast, immunedeposits in a kidney transplanted inadvertently from adonor with subclinical IgAN to a patient with non-IgANrenal disease clear within several weeks [9], suggestingthat the cause of IgAN is extrarenal.

Serum or plasma of IgAN patients often exhibitsincreased levels of IgA-containing circulating immunecomplexes (CIC) [10–12]. Several lines of evidence sug-gest that the mesangial immune deposits likely originatefrom CIC: (1) IgA1, but not IgA2, is present in CIC ofmost IgAN patients [10, 13] and in the mesangial deposits[3]; (2) shared idiotypic determinants are expressed onCIC and in mesangial deposits [14] (however, without adisease-specific idiotype [15]); (3) IgAN-CIC [16–19] arecharacterized by galactose (Gal)-deficient hinge-regionO-linked glycans of IgA1 [12, 20], and the same Gal de-ficiency was also detected in IgA1 in the mesangium [21,22]; and (4) Gal-deficient IgA1 and IgA-IgG CIC arealso found in sera of Henoch-Schonlein purpura (HSP)patients, but only in those with renal involvement [23, 24].

Recent reports from several laboratories support theproposal that O-linked glycans in the hinge regionof some IgA1 molecules in the circulation of IgANpatients are deficiently galactosylated [12, 16–20, 23,25–36]. In the absence of Gal, the terminal sugar is

504

Novak et al: Stimulation of mesangial cells by IgA1 complexes from IgAN patients 505

N-acetylgalactosamine (GalNAc) [12, 20]. Subsequently,these sugar moieties or hinge-region glycopeptides [37,38] are recognized by naturally occurring antibodieswith antiglycan or antihinge region peptide specificities[20, 38], and CIC are formed [36]. The circulating Gal-deficient IgA1 is exclusively in CIC and is mostly aJ-chain-containing polymer (p) [20]. Our earlier studyshowed that the IgAN-CIC containing Gal-deficientIgA1 bound to MC more efficiently than uncomplexedIgA, and that a greater amount of CIC from an IgANpatient bound to MC than CIC from a healthy controlpatient [39]. To study activation of MC in vitro, CIC(700–1000 kD) were isolated from sera of IgAN patientsand healthy volunteers by size-exclusion chromatogra-phy, fractions were added to serum-starved MC in cul-ture, and cellular proliferation was measured using 3H-thymidine incorporation. Our results support the hypoth-esis that aberrantly glycosylated IgA1-containing CIC areinvolved in the pathogenesis of IgAN.

METHODS

Patients and control subjects

Serum samples were obtained from 12 biopsy-provenIgAN patients (mean age 36.8 years, 9 males, 3 females;serum creatinine ranged from 0.7 to 5.1 mg/dL, dipstickurinary protein from 0 to 4+, and blood from 0 to 4+) and5 healthy control subjects (mean age 38.2 years, 3 males,2 females; normal serum creatinine levels and urine neg-ative by dipstick for protein and blood). Seven IgANpatients had serum samples collected during an acuteepisode of macroscopic hematuria, and one of them hadan additional sample collected several weeks after res-olution of the macroscopic hematuria. The study wasapproved by the Institutional Review Board; informedwritten consent was obtained before collecting samples.

Cell cultures and proliferation assays

Human MC (2 different preparations of primary cells)were purchased from BioWhittaker (Walkersville, MD,USA). Cells from passages 3 to 4 were maintained inRPMI 1640 supplemented with 20% fetal calf serum(FCS), L-glutamine (2 mmol/L), penicillin G (100 U/mL),streptomycin (0.1 mg/mL) in humidified 5% CO2 atmo-sphere at 37◦C [39]. Purity of MC was assessed by cellmorphology (stellar shape) and immunohistochemicalfeatures, including positive staining for vimentin, con-firmed by Western blot and negative staining for factorVIII–related antigen and cytokeratin (to exclude contam-ination with endothelial and epithelial cells, respectively).

Proliferation experiments were conducted in 24-welltissue culture plates seeded with MC. At 85% to 95% con-fluence, MC were serum-starved (in a medium contain-ing 0.5% FCS) [39] for 24 hours before the experiment.Serum fractions were pooled (each two-total 0.4 mL; seebelow), filter-sterilized, mixed with an equal volume of

the medium containing 1% FCS (to obtain final concen-tration 0.5% FCS), and incubated in duplicates (0.4 mLin each well) with MC in humidified 5% CO2 atmo-sphere at 37◦C for 20 hours. The culture medium aloneand medium supplemented with platelet-derived growthfactor (PDGF; R&D Systems, Minneapolis, MN, USA)(10 ng/mL) were used as negative and positive controls,respectively. For the last 4 hours of incubation, 100 lL 3H-thymidine (1 lCi; PerkinElmer, Wellesley, MA, USA) inthe culture medium was added to each well. MC in eachwell were washed two times with PBS to remove free3H-thymidine [each wash was filtered using MultiScreen-HA filter plates (Millipore, Bedford, MA, USA) to cap-ture detached cells]. Washed cells were lysed with 0.2 mL0.3 mol/L NaOH, then each well was washed with 0.1 mL5% acetic acid, and all washes and the filter were com-bined and added to 5 mL scintillation liquid in a scintilla-tion vial. The radioactivity was determined using a Wallacliquid scintillation counter 1409DSA (PerkinElmer). Av-erage values were calculated from duplicates for eachserum fraction and expressed either directly as counts perminute (cpm) or relatively to the negative control (cpmof sample/cpm of the control) as relative proliferation.Thus, ratios greater than 1.0 indicate stimulation, whileratios below 1.0 indicate inhibition.

Isolation of serum fractions enriched with CIC

Serum fractions enriched with CIC were isolated byprecipitation of serum from 3 IgAN patients and 3 nor-mal healthy control subjects with 7% polyethylene gly-col (PEG) [40] and from 9 IgAN patients and 4 normalhealthy control subjects by size-exclusion chromatogra-phy. For the latter procedure, serum (0.5 mL) was fil-tered using 0.45 lm filter (Pall Corporation, Ann Arbor,MI, USA) and fractionated on a calibrated Superose 6column (600 × 12 mm; Amersham Biosciences Corpora-tion, Piscataway, NJ, USA) in phosphate-buffered saline(PBS) [20, 39]. Fractions containing proteins of appar-ent molecular mass over 700 kD were collected, andthe aliquots were analyzed for IgA, IgG, HAA-bindingIgA1, and IgA-IgG complexes using enzyme-linked im-munosorbent assay (ELISA). The profiles of IgA, IgG,and HAA were essentially identical with that shown inFigure 5 of our previous publication [20]. The remainderof each two fractions was pooled (total 0.4 mL), filter-sterilized, and added to MC as described above.

Depletion and supplementation of serum IgA

Serum (0.6 mL) was incubated overnight at 4◦C withimmobilized jacalin (EY Laboratories, San Mateo, CA,USA) or anti-IgA1 monoclonal antibody (H69-7.1) [20]to remove IgA1 or with immobilized Protein G (SigmaChemical Co., St. Louis, MO, USA) to remove IgG. Then,the serum sample was filtered using 0.45 lm filter (PallCorporation) and fractionated as described above.

506 Novak et al: Stimulation of mesangial cells by IgA1 complexes from IgAN patients

To supplement serum with Gal-deficient IgA1, 20 lgor 50 lg of IgA1 proteins (a naturally Gal-deficientmyeloma pIgA1 (Mce) or a jacalin-affinity and size-exclusion chromatography-purified pIgA1 from an IgANpatient) were added to the serum samples from an IgANpatient to form additional immune complexes in vitro.After overnight incubation at 4◦C, the serum sample wasfiltered using 0.45 lm filter (Pall Corporation) and pro-cessed as described above.

ELISA

Serum fractions were analyzed by ELISA for IgA, IgG,HAA-binding IgA1, IgA-IgG complexes, and IgG sub-classes bound to IgA [20, 39]. Biotin-labeled HAA waspurchased from Sigma, goat biotin-labeled antihumanIgG- and IgA-specific antibodies were purchased fromSouthern Biotechnology Associates, Inc. (Birmingham,AL, USA), and an F(ab’)2 fragment of anti-IgA (heavychain-specific) was purchased from ICN Biomedicals, Inc.(Irvine, CA, USA). IgG subclass-specific antibodies usedin our assays have been previously described [20].

Immunohistochemical staining

MC were grown on Lab-Tech chambered slides (Nalge/Nunc, Hoperville, IL, USA), serum-starved as describedabove, and incubated with control medium or stimu-lated with CIC-containing serum fractions or PDGFfor 24 hours. MC were then washed, fixed in 3%paraformaldehyde, and stained with antibody directedagainst proliferating cell nuclear antigen (PCNA; Dako,Carpinteria, CA, USA) to detect proliferating cells orstained with the DeadEnd colorimetric TUNEL (TdT-mediated dUTP Nick-End Labeling) system (Promega,Madison, WI, USA) to detect apoptotic cells. For PCNAstaining, antigen heat-retrieval was used [41].

Sections of renal biopsies from two patients with IgAN(different patients than those who provided serum sam-ples; diffuse mesangial proliferation without crescents),two patients with diffuse lupus nephritis (WHO classIV), three patients with non-IgA proliferative glomeru-lonephropathy (one minimal, one moderate, one dif-fuse), and one patient with minimal change disease wereprocessed for PCNA and TUNEL staining as describedabove.

Statistical analysis

Comparison between group means was performed us-ing Student t test. P values equal to or less than 0.05 wereconsidered statistically significant.

RESULTS

Proliferation of MC in the presence of sera from IgANpatients and control subjects

Native sera from IgAN patients and healthy controlsubjects stimulated proliferation of serum-starved MC,

Table 1. Proliferation of human mesangial cells stimulated with 5%native human serum from four IgAN patients and four healthy

control subjects, Gal-deficient pIgA1 protein (Mce; 10 lg), and PDGF(10 ng/mL)

Sample Thymidine incorporationb cpm

Negative controla 9490PDGF 29,711Gal-deficient pIgA1 8564IgAN sera 78,625 ± 30,578Normal healthy control sera 68,457 ± 6940

aRPMI 1640 medium with 0.5% FCS, L-glutamine (2 mmol/L), penicillin G(100 U/mL), streptomycin (0.1 mg/mL).

bAverage or average ± SD.

1.1

1.0

0.9

0.8

0.7

0.6

HA

A r

eact

ivity

of s

erum

IgA

,O

D 4

05 n

m

100,00010,0003H-thymidine incorporation, cpm

Fig. 1. Proliferation of human MC stimulated with 5% human serumfrom IgAN patients (serum samples were randomly selected from pa-tients with serum creatinine range 1.1 to 4.5 mg/dL, no apparent hema-turia, and proteinuria 0 to 3+) correlated with the amount of Gal-deficient IgA (r = 0.832, P < 0.05). Number of samples did not allowany analyses of possible correlation between clinical status and prolifer-ation. Proliferation of MC was assessed using 3H-thymidine incorpora-tion after 20 hours; Gal-deficiency of IgA was analyzed by determiningthe binding of HAA lectin using ELISA. Each value represents meanfrom duplicate assay.

as determined by 3H-thymidine incorporation (Table 1).Sera from IgAN patients showed somewhat higher stim-ulation than sera from control subjects, but the differencewas not statistically significant. Because the formation ofsublytic complement complexes (C5-9) can activate MC[42–44], we also tested MC proliferation in the presenceof heat-inactivated sera. Under these conditions, MC pro-liferation was by 15% to 28% greater when compared tothe stimulation with native (e.g., untreated) sera (resultsnot shown), possibly due to aggregation of immunoglob-ulins. MC responded to a positive control, PDGF, by atleast a 2- to 3-fold increase in 3H-thymidine incorpora-tion. In contrast, Gal-deficient pIgA1 only slightly re-duced 3H-thymidine incorporation.

Stimulation with sera from IgAN patients increasedproliferation proportionally to the levels of Gal-deficientIgA1 (measured as HAA binding) in the samples(Fig. 1). This may be related to our observation that


6

4

2

0

Rel

ativ

e pr

olife

ratio

n

15 25 35 45Fraction number

1000 kD 700 kD

Fig. 2. Proliferation of MC after stimulation with fractions from nativeserum collected from an IgAN patient during an episode of macroscopichematuria (closed circles) and then later at a quiescent stage (open cir-cles). Serum samples (0.5 mL) were fractionated on a calibrated columnof Superose 6; 0.25-mL fractions were collected, every two consecutivefractions were pooled and filter-sterilized before adding in duplicates tothe culture of serum-starved MC in 24-well plates. After 20 hours incu-bation, 3H-thymidine was added and the culture incubated for another4 hours. MC were then washed and lyzed, and 3H-thymidine incorpo-ration was measured using a liquid scintillation counter. Each valuerepresents an average from a duplicate. CIC of 800 to 900 kD collectedat the time of macroscopic hematuria stimulated MC proliferation morethan CIC of similar size collected after macroscopic hematuria had re-solved. Furthermore, MC proliferation-stimulating CIC of >900 kDwere detected only at the time of macroscopic hematuria.

Gal-deficient IgA1 bound to MC with higher affinity thannormally glycosylated IgA1 [39]. Because Gal-deficientIgA1 is present in CIC [12, 20], we isolated immunecomplexes from three IgAN patients and three controlsubjects using precipitation with 7% PEG and exam-ined their effects on MC proliferation. 3H-thymidineincorporation after stimulation with 1 to 2 lg proteinwas about 2-fold greater for samples from IgAN pa-tients than from healthy control subjects. However, be-cause of inconsistency of results in further experiments,and because PEG precipitates serum proteins other thanjust immune complexes [40], we tested MC responsesto serum fractions from IgAN patients and control sub-jects enriched with IgA-containing CIC prepared by size-exclusion chromatography.

Gal-deficient IgA1-containing CIC fractionated fromserum of an IgAN patient collected during an episodeof macroscopic hematuria stimulated MC proliferationmore than CIC obtained during a subsequent quiescentphase

A serum sample from an IgAN patient was collectedat the time of macroscopic hematuria and several weeksafter resolution; both samples were fractionated on a cal-ibrated column of Superose 6 and tested as previously

14

12

10

8

6

4

2

0

Rel

ativ

e pr

olife

ratio

n

15 25 35 45

Fraction number

14

12

10

8

6

4

2

0

Rel

ativ

e pr

olife

ratio

n

1000 kD 700 kD

A

B

Fig. 3. Relative proliferation of human MC stimulated with fractionsof sera from four healthy control subjects (open symbols; A) and fourIgAN patients without macroscopic hematuria (closed symbols; B).Serum samples (0.5 mL) were fractionated on a calibrated column ofSuperose 6, and fractions processed and analyzed as described in Fig. 2.Each value represents an average from a duplicate.

described. CIC of 800 to 900 kD from the first samplestimulated MC proliferation more than CIC from thesecond sample (Fig. 2). Furthermore, MC proliferation-stimulating CIC of >900 kD were detected in only thefirst sample. These results indicated that there may bechanges in the CICs associated with clinical activity ofthe disease.

Stimulation of MC proliferation by CIC-containingfractions from sera of IgAN patients and healthy controlsubjects

To determine proliferative response of MC incubatedin the presence of serum fractions from IgAN patientsand healthy control subjects, we subjected sera from fourIgAN patients without macroscopic hematuria and fourhealthy control subjects to size-exclusion chromatogra-phy on a calibrated Superose 6 column. In each exper-iment, we compared one IgAN patient’s sample withone control sample. Fractions from two IgAN patientsand from one control containing Gal-deficient IgA1-CIC with molecular mass 800 to 900 kD stimulated MCproliferation (Fig. 3). To compare more samples in one


2

1

0

Rel

ativ

e pr

olife

ratio

n


1000 kD 700 kD

Fig. 4. Proliferation of human mesangial cells after stimulation withfractions from sera from four IgAN patients with macroscopic hema-turia. Serum samples (0.5 mL) were fractionated on a calibrated columnof Superose 6 and fractions analyzed as described in Fig. 2. All four sam-ples appeared to have similar pattern of CIC: inhibitory CIC of about700 to 800 kD and stimulatory CIC of about 800 to 900 kD.

experiment, we pooled Superose 6 fractions containingCIC of 800 to 900 kD from sera of three IgAN patientswithout macroscopic hematuria and from three healthycontrol subjects. The results showed that the samplesfrom IgAN patients induced, on average, 37% higher3H-thymidine incorporation in MC compared to the sam-ples from healthy control subjects.

When MC proliferative responses were determined inthe presence of CIC-containing fractions from sera offour IgAN patients with active macroscopic hematuria,two types of IgA-containing CIC, one stimulatory andthe other inhibitory, were detected (Fig. 4). Notably, allfour samples exhibited proliferation activity of 800 to900 kD fractions and inhibitory activity of ∼700 to 800 kDfractions. The fractions with proliferation-stimulating ac-tivity in all samples (800–900 kD) corresponded to CICcontaining most of the Gal-deficient IgA1 (detected asHAA-reactive IgA by ELISA) [20].

Results obtained using 3H-thymidine incorporationwere confirmed with immunohistochemical staining ofthe cells with anti-PCNA antibody. Staining of serum-starved MC (cultured in RPMI 1640 + 0.5% FCS) showedabout 50% cells weakly expressing PCNA; stimulationwith Gal-deficient IgA1-containing CIC (800–900 kD)increased the number of MC expressing PCNA (93%were PCNA-positive with moderate to strong intensity).As a positive control, PDGF-stimulated MC showed thatmore than 98% of them stained intensely for PCNA. Wealso examined cell apoptosis using TUNEL staining, butdid not find any changes with MC incubated with vari-ous fractions, including the CIC inhibiting MC prolifera-tion. Thus, the inhibition of 3H-thymidine incorporationby 700 to 800 kD CIC appeared to be a true inhibition ofproliferation rather than acceleration of apoptosis.

Interestingly, we observed similar patterns of PCNAstaining (increased numbers of PCNA-positive MC;Fig. 5) and TUNEL staining (few positive cells) in themesangia of the renal biopsies from two IgAN patients.This observation suggests that MC both in vitro and inthe glomerular mesangia of IgAN patients respond toIgAN-CIC by enhanced proliferation with no apparentinduction of apoptosis.

Role of IgG and IgA in MC-stimulating/inhibitingactivity

The MC-stimulating as well as MC-inhibiting CIC con-tained IgA and IgG, as determined by ELISA and pre-viously published information [12, 20, 39], but there wasno direct evidence implicating these components in thestimulation of MC proliferation. The stimulatory frac-tions contained IgG bound to IgA1 predominantly ofthe IgG3 subclass, followed in descending order by IgG1,IgG2, and traces of IgG4. The inhibitory fractions con-tained relatively more IgG bound to IgA1, predominantlyof the IgG3 and IgG1 subclasses.

To examine a possible stimulatory effect of this IgA,we fractionated an aliquot of serum from an IgAN pa-tient before and after adsorption on jacalin-agarose toremove IgA1 before fractionation. Removal of IgA1 ex-tinguished the MC-stimulating and dampened the inhibit-ing activities (Fig. 6A). Furthermore, we eluted the boundIgA-CIC from jacalin with melibiose and observed stim-ulatory activity associated with CIC of apparent molec-ular mass about 800 to 900 kD after fractionation usingsize-exclusion column (Fig. 6A). Removal of IgA1 us-ing immobilized monoclonal antibody specific for IgA1had an effect similar to that of jacalin—removal of MC-stimulating activity.

To determine the role of IgG in the interactions ofCIC with MC, we examined the proliferation of MC inthe presence of serum from which the free and com-plexed IgG were removed by affinity chromatographyon agarose-conjugated protein G. Adsorption on proteinG removed part of the inhibitory activity from an IgANserum (Fig. 6B). In contrast, protein-G adsorption re-moved most of the inhibitory, as well as part of the stim-ulatory, activities from the serum of a healthy controlsubject (Fig. 6C).

These data suggested that MC activation by CICwas dependent on the composition and size of thesecomplexes. Most Gal-deficient IgA1 was present in thefractions stimulating MC proliferation, while inhibitoryfractions were rich in IgG. The finding that protein G re-moved very little of the stimulatory activity and only aportion of the inhibitory activity from IgAN sera, whilejacalin or anti-IgA antibody removed both stimulatoryand inhibitory activities, may reflect differential three-dimensional organization of IgG and IgA in these CICand, thus, differential availability for various Ig receptors.


BA

Fig. 5. Representative pictures of PCNA staining of glomeruli from renal biopsies of patients with a minimal change disease (A) and IgAN (B).PCNA-positive nuclei in both figures are marked by arrows.

Gal-deficient IgA1 added to serum increasesMC-stimulating activity

To extend the data from studies with IgA1-depletedsera, Gal-deficient IgA1 was added to serum from anIgAN patient. In addition to CIC, this serum con-tained an excess of IgG and IgA1 antibodies withspecificity for Gal-deficient O-linked hinge region gly-cans; therefore, additional complexes were formed be-fore the fractionation. Two types of desialylated IgA1were used: a naturally Gal-deficient myeloma pIgA1(Mce) and a jacalin-affinity-purified and size-exclusionchromatography-purified pIgA1 from an IgAN patient.In both instances, we added 20 lg of each IgA prepara-tion to 0.5 mL serum. After incubation, the serum sampleswere fractionated and the activities of fractions tested asbefore. Supplementation of IgAN serum with pIgA1 iso-lated from the same patient increased MC proliferation,mostly with 800 to 900 kD CIC (Fig. 7A). Addition of20 lg myeloma IgA1 (Mce) increased MC proliferationactivity that was associated with fractions containing 800to 900 kD CIC, as well as with fractions containing 900to 1000 kD CIC (Fig. 7B). ELISA revealed that thesefractions contained newly formed IgG-IgA complexes(Fig. 7C). No increase in proliferation was detected when10 lg IgA1 myeloma protein was used instead of CIC(Table 1).

In contrast, supplementation of serum with a largeramount of IgA1 (50 lg) resulted in the increase of in-hibitory activity of the serum fractions containing low-molecular-weight immune complexes (Fig. 7B). Thisresult suggested that the antigen/antibody ratio plays animportant role in the biological properties of the resultantimmune complexes.

DISCUSSION

IgA-containing immune deposits in the renalmesangium associated with mesangial proliferativeactivity and matrix expansion are the hallmark of IgAN.We have examined with in vitro experiments the bindingand proliferative responses of MC to IgA1-containingCIC and uncomplexed IgA1. It is known that in patientswith IgAN the circulating IgA1 with aberrant hingeregion glycans is bound to IgG or IgA1 with antiglycanspecificity [20]. Our earlier studies have shown that vari-ous forms of uncomplexed myeloma IgA1 (monomeric,polymeric, with intact or enzymatically modified hingeregion glycans) bound to human MC in vitro withrelatively low affinity and did not alter significantly theirproliferation. IgA1-CIC prepared by PEG precipitationor by size-exclusion chromatography bound to MCwith considerably higher affinity and, in some instances,


2

1

0

Rel

ativ

e pr

olife

ratio

n

C

15 25 35 45

Fraction number

2

1

0

Rel

ativ

e pr

olife

ratio

n

B

2

1

0

Rel

ativ

e pr

olife

ratio

n

A

1000 kD 700 kD

Fig. 6. Proliferation of human MC after stimulation with fractionsfrom native serum (open symbols) from an IgAN patient and the sameserum after adsorption on immobilized jacalin lectin to remove IgA1(closed symbols; A). The jacalin-eluted sample was also fractionatedand the fractions were added to MC (filled triangles, A). One aliquotof serum (0.5 mL) from an IgAN patient with macroscopic hematuriawas directly fractionated on a calibrated column of Superose 6 whilethe other aliquot was first adsorbed on jacalin-agarose to remove IgA1and then fractionated. MC were grown and the fractions analyzed asdescribed in Fig. 2. Both stimulatory CIC and inhibitory CIC (althoughthis native serum sample did not exhibit as large an inhibitory effect asother samples) were removed by jacalin adsorption. The jacalin-elutedsample was also fractionated and the fractions retained their abilityto stimulate MC proliferation (triangles, A). Proliferation of humanmesangial cells after stimulation with fractions from native serum (opensymbols) of an IgAN patient and the same serum after adsorption onimmobilized Protein G (closed symbols; B). Protein G adsorption,

stimulated proliferation. When assayed by 3H-thymidineincorporation, proliferation of MC incubated with IgA1-CIC from IgAN patients and healthy control subjects didsignificantly differ, consistent with a previous report [45].However, further experiments demonstrated two typesof IgA-CIC: stimulating and inhibiting MC proliferationin vitro. Thus, experiments using whole serum or totalCIC cannot be clearly interpreted without separatingIgA-CIC according to their molecular masses.

To our knowledge, this is the first report of two types offunctionally distinct IgA1-containing CIC that differen-tially affect quiescent MC. While it is not clear whetherboth types of CIC play a role in vivo, the quality (e.g., sizeand composition) may determine the biological activityof these complexes. For example, smaller CIC may beefficiently catabolized in the liver while larger CIC mayescape hepatic catabolism and instead penetrate into therenal mesangium.

We undertook additional studies to define whetherproteins other than IgG and IgA1 (such as acute phaseproteins) are present in the active fractions. First, usingsodium dodecyl sulfate-polyacrylamide gel electrophore-sis (SDS-PAGE) and proteomics, we have identified a2macroglobulin and its precursor, fibronectin, and hap-toglobin 2 precursor (J. Novak et al, unpublished results).These proteins were in all three types of fractions (e.g.,stimulatory, inhibitory, and having no effect on MC pro-liferation), indicating that they were not likely to alterMC proliferation and may be in complexes or aggregatesof various molecular masses. The same proteins remainedin the nonbinding fractions after removing IgA1 by affin-ity chromatography, and the resultant IgA1-depletedfractions did not alter MC proliferation. Furthermore,the IgA1-containing preparations of CIC eluted fromthe affinity (anti-IgA1) column and analyzed by SDS-PAGE did not contain detectable levels of proteins otherthan immunoglobulins. Thus, based on these prelimi-nary results, both stimulatory and inhibitory activitiesof serum fractions with IgA1-containing CIC may beattributed mostly to different properties of these com-plexes rather than to the presence of proteins other thanimmunoglobulins.

The importance of Gal-deficient IgA1 in the stimula-tion of MC proliferation is obvious: depletion of IgA1-CIC abolished the proliferation-stimulating activity. Onthe other hand, uncomplexed Gal-deficient IgA1 did notincrease MC proliferation, suggesting that to stimulate

by removing free and complexed IgG, diminished the inhibitory effect.Proliferation of human mesangial cells after stimulation with fractionsfrom native serum (open symbols) of a healthy control and the sameserum after adsorption on immobilized Protein G (closed symbols; C).Protein G adsorption removed most of the inhibitory CIC and some ofthe stimulatory CIC.


0.5

1.0

0

IgG

-IgA

CIC

, OD

405

nm

C


2

3

4

1

0

Rel

ativ

e pr

olife

ratio

n

B

2

1

0

Rel

ativ

e pr

olife

ratio

n

1000 kD 700 kD

A

Fig. 7. Proliferation of MC after stimulation with fractions from nativeIgAN serum (open symbols) and IgAN serum supplemented with 20lg polymeric desialylated IgA1 from the same IgAN patient (closedsymbols) (A). Proliferation of MC after stimulation with fractions fromIgAN serum supplemented with 20 lg (closed circles) or 50 lg (opentriangles) desialylated polymeric Gal-deficient IgA1 myeloma protein(Mce) to form additional immune complexes in vitro (B). IgA-IgG com-plexes in individual fractions of native serum (open symbols) and IgANserum supplemented with 20 lg desialylated polymeric Gal-deficientIgA1 myeloma protein (closed symbols) was determined by ELISA(C). Newly formed IgG-IgA complexes were detected in the fractionsstimulating MC proliferation.

MC proliferation it must be bound in immune complexes.Indeed, supplementation of serum with Gal-deficientIgA1 formed more CIC and thereby increased stimula-tory activity. These observations suggest a requirementfor additional factor(s) (e.g., potential ligands other than

IgA1, such as IgG) that allows activation of MC byCIC but not by uncomplexed IgA1. This model impliescross-linking of different receptors. An alternative modelassumes that only in CIC does the IgA1 conformationpermit cross-linking of a single type of receptor. How-ever, in either case, both the quality and quantity of IgA1-contaning CIC apparently determine the effect on MC invitro. These conclusions were further supported by theobservation that while supplementation of serum witha smaller amount (20 lg) of desialylated Gal-deficientpIgA1 increased the stimulatory activity of serum frac-tions containing CIC with high-molecular mass, supple-mentation with a larger amount (50 lg) increased the in-hibitory activity of serum fractions containing CIC withlow-molecular mass (Fig. 7).

While our data suggested a role of IgA1 in IgA-containing CIC from IgAN patients for binding to MC[39] and consequential stimulation of MC proliferation(this paper), the nature of the mesangial receptor(s) re-mains unresolved [46]. A current consensus proposes thatthis receptor(s) differs from the Fca receptor (CD89), theasialoglycoprotein receptor (ASGPR), and the polymericIg receptor [47–50]. Two other receptors have emergedas possible candidates: CD71 (transferrin receptor) [51,52] and Fca/l receptor [53, 54]. We have confirmed ex-pression of CD71 mRNA in proliferating MC in culture(unpublished data and [55]). CD71 binds only IgA1 [51],while MC bind both IgA subclasses [39, 50], suggestingexistence of another IgA receptor. Theoretically, it maybe the Fca/l receptor [54]. In contrast, pIgA1 from IgANpatients induced MC to secrete macrophage migrationinhibitory factor and TNF-a, likely through an unidenti-fied IgA receptor different from CD89, ASGPR, CD71,or Fca/l [56]. While the nature of the IgA1-CIC recep-tor on MC is unknown, we speculate, based on the factthat IgA1-containing CIC but not free IgA1 stimulatedMC proliferation, that a coordinated binding and cross-linking of two IgA (or other) receptors may be requiredto trigger cellular proliferation. Further studies are neces-sary to clarify the expression and pathogenic importanceof these candidate receptors and their potential splicingvariants.

CONCLUSION

IgA-containing CIC with the highest content of Gal-deficient IgA1 stimulated proliferation of MC in vitro.While it is not clear why CIC of different sizes differ-entially altered MC proliferation, we propose that fac-tors such as molecular form and quantity of Gal-deficientIgA1, number and localization of GalNAc epitopes in thehinge region, isotype and affinity of the antiglycan anti-bodies, and handling and disposal of the formed CIC areamong the important factors that affect quality and quan-tity of CIC and their pathogenic potential. Understanding


the mechanisms by which these pathogenic CIC form andactivate MC [57–59] may provide a rational basis for thedesign of future interventional strategies.

ABBREVIATIONS

Nonstandard abbreviations: CIC, circulating immunecomplexes; GalNAc, N-acetylgalactosamine; HAA, He-lix aspersa lectin; HSP, Henoch-Schonlein purpura; IgAN,IgA nephropathy; MC, mesangial cells; PCNA, prolif-erating cell nuclear antigen; PEG, polyethylene glycol;TUNEL, TdT-mediated dUTP nick-end labeling.

ACKNOWLEDGMENTS

This work was supported by grants DK 57750 andDK 61525 from the National Institutes of Health, andfrom General Clinical Research Centers of the Uni-versity of Alabama at Birmingham M01 RR00032 andthe University of Tennessee Health Sciences CenterM01 RR00211. The authors express their appreciationto Catherine Barker and Sue Woodford (University ofAlabama at Birmingham), and Dr. Kimberly Fisher andSandra Grimes (University of Tennessee) for help withcollecting clinical samples, Rose Kulhavy (University ofAlabama at Birmingham) for providing IgA1 myelomaproteins, and Dr. Jiri Radl (TN, Leiden, The Netherlands)for the monoclonal antibody specific for IgA1.

Reprint requests to Jan Novak, University of Alabama at Birming-ham, Department of Microbiology, 845 19th Street South, Room 734,Birmingham, AL 35294.E-mail: [email protected]

REFERENCES

1. EMANCIPATOR SN: IgA nephropathy and Henoch-Schonlein syn-drome, in Heptinstall’s Pathology of the Kidney, edited by JennetteJC, Olson JL, Schwartz MM, et al, Philadelphia, Lippincott-RavenPublishers, 1998, pp 479–539

2. BERGER J, HINGLAIS N: Les depots intercapillaires d’IgA-IgG (in-tercapillary deposits of IgA-IgG). J Urol Nephrol 74:694–695, 1968

3. CONLEY ME, COOPER MD, MICHAEL AF: Selective deposition ofimmunoglobulin A1 in immunoglobulin A nephropathy, anaphy-lactoid purpura nephritis, and systemic lupus erythematosus. J ClinInvest 66:1432–1436, 1980

4. JULIAN BA, TOMANA M, NOVAK J, et al: Progress in the pathogenesisof IgA nephropathy. Adv Nephrol 29:53–72, 1999

5. NOVAK J, JULIAN BA, TOMANA M, et al: Progress in molecular andgenetic studies of IgA nephropathy. J Clin Immunol 21:310–327,2001

6. COPPO R, AMORE A, CIRINA P, et al: IgA serology in recurrentand non-recurrent IgA nephropathy after renal transplantation.Nephrol Dial Transplant 10:2310–2315, 1995

7. JULIAN BA, SAID M, BARKER CV: Allograft loss in IgA nephropathy.J Am Soc Nephrol 9:91A, 1998

8. ODUM J, PEH CA, CLARKSON AR, et al: Recurrent mesangial IgAnephritis following renal transplantation. Nephrol Dial Transplant9:309–312, 1994

9. SILVA FG, CHANDER P, PIRANI CL, et al: Disappearance of glomeru-lar mesangial IgA deposits after renal allograft transplantation.Transplantation 33:241–246, 1982

10. CZERKINSKY C, KOOPMAN WJ, JACKSON S, et al: Circulating immunecomplexes and immunoglobulin A rheumatoid factor in patients

with mesangial immunoglobulin A nephropathies. J Clin Invest77:1931–1938, 1986

11. COPPO R, BASOLO B, PICCOLI G, et al: IgA1 and IgA2 immune com-plexes in primary IgA nephropathy and Henoch-Schonlein nephri-tis. Clin Exp Immunol 57:583–590, 1984

12. TOMANA M, MATOUSOVIC K, JULIAN BA, et al: Galactose-deficientIgA1 in sera of IgA nephropathy patients is present in complexeswith IgG. Kidney Int 52:509–516, 1997

13. COPPO R, BASOLO B, MARTINA G, et al: Circulating immune com-plexes containing IgA, IgG and IgM in patients with primary IgAnephropathy and with Henoch-Schonlein nephritis. Correlationwith clinical and histologic signs of activity. Clin Nephrol 18:230–239, 1982

14. GONZALES-CABRERO J, EGIDO J, MAMPASO F, et al: Characteriza-tion of circulating idiotypes containing immune complexes andtheir presence in the glomerular mesangium in patients with IgAnephropathy. Clin Exp Immunol 76:204–209, 1989

15. VAN DEN WALL BAKE AWL, BRUIJN JA, ACCAVITTI MA, et al: Sharedidiotypes in mesangial deposits in IgA nephropathy are not disease-specific. Kidney Int 44:65–74, 1993

16. ANDRE PM, POGAMP P, CHEVET P: Impairment of jacalin binding toserum IgA in IgA nephropathy. J Clin Lab Anal 4:115–119, 1990

17. MESTECKY J, TOMANA M, CROWLEY-NOWICK PA, et al: Defec-tive galactosylation and clearance of IgA1 molecules as a pos-sible etiopathogenic factor in IgA nephropathy. Contrib Nephrol104:172–182, 1993

18. ALLEN AC, HARPER SJ, FEEHALLY J: Galactosylation of N- and O-linked carbohydrate moieties of IgA1 and IgG in IgA nephropathy.Clin Exp Immunol 100:470–474, 1995

19. HIKI Y, HORII A, IWASE H, et al: O-linked oligosaccharide on IgA1hinge region in IgA nephropathy. Fundamental study for precisestructure and possible role. Contrib Nephrol 111:73–84, 1995

20. TOMANA M, NOVAK J, JULIAN BA, et al: Circulating immune com-plexes in IgA nephropathy consist of IgA1 with galactose-deficienthinge region and antiglycan antibodies. J Clin Invest 104:73–81, 1999

21. ALLEN AC, BAILEY EM, BRENCHLEY PEC, et al: Mesangial IgA1 inIgA nephropathy exhibits aberrant O-glycosylation: Observationsin three patients. Kidney Int 60:969–973, 2001

22. HIKI Y, ODANI H, TAKAHASHI M, et al: Mass spectrometry provesunder-O-glycosylation of glomerular IgA1 in IgA nephropathy.Kidney Int 59:1077–1085, 2001

23. ALLEN AC, WILLIS FR, BEATTIE TJ, et al: Abnormal IgA glycosyla-tion in Henoch-Schonlein purpura restricted to patients with clinicalnephritis. Nephrol Dial Transplant 13:930–934, 1998

24. LEVINSKY RJ, BARRATT TM: IgA immune complexes in Henoch-Schonlein purpura. Lancet 2:1100–1103, 1979

25. ALLEN AC, TOPHAM PS, HARPER SJ, et al: Leucocyte b1,3 galacto-syltransferase activity in IgA nephropathy. Nephrol Dial Transplant12:701–706, 1997

26. ALLEN A, FEEHALLY J: IgA glycosylation in IgA nephropathy. AdvExp Med Biol 435:175–183, 1998

27. ALLEN AC, BAILEY EM, BUCK KS, et al: O-glycosylation of mesangialIgA1 in IgA nephropathy. J Am Soc Nephrol 10:506A, 1999

28. COPPO R, EMANCIPATOR S: Pathogenesis of IgA nephropathy: Es-tablished observations, new insights and perspectives in treatment.J Nephrol 7:5–15, 1994

29. COUSER WG: Glomerulonephritis. Lancet 353:1509–1515, 199930. HIKI Y, KOKUBO T, IWASE H, et al: Underglycosylation of IgA1 hinge

plays a certain role for its glomerular deposition in IgA nephropa-thy. J Am Soc Nephrol 10:760–769, 1999

31. ALLEN AC: Abnormal glycosylation of IgA: Is it related to the patho-genesis of IgA nephropathy? Nephrol Dial Transplant 10:1121–1124, 1995

32. BAHARAKI D, DUEYMES M, PERRICHOT R, et al: Aberrant glycosy-lation of IgA from patients with IgA nephropathy. Glycoconj J13:505–511, 1996

33. LEUNG JCK, POON PYK, LAI KN: Increased sialylation of polymericimmunoglobulin A1: Mechanism of selective glomerular depositionin immunoglobulin A nephropathy? J Lab Clin Med 133:152–160,1999

34. FLOEGE J, FEEHALLY J: IgA nephropathy: Recent developments. JAm Soc Nephrol 11:2395–2403, 2000

35. MESTECKY J, TOMANA M, MATOUSOVIC K, et al: Heterogeneity of


carbohydrate moieties of IgA1 molecules from IgA nephropathypatients and normal individuals. Nephrology 3:85–89, 1997

36. MESTECKY J, NOVAK J, JULIAN BA, et al: Pathogenic potential ofgalactose-deficient IgA1 in IgA nephropathy. Nephrology 7:S92–S99, 2002

37. KOKUBO T, HIKI Y, IWASE H, et al: Exposed peptide core of IgA1hinge region in IgA nephropathy. Nephrol Dial Transplant 14:81–85, 1999

38. KOKUBO T, HASHIZUME K, IWASE H, et al: Humoral immunity againstthe proline-rich peptide epitope of the IgA1 hinge region in IgAnephropathy. Nephrol Dial Transplant 15:28–33, 2000

39. NOVAK J, VU HL, NOVAK L, et al: Interactions of human mesangialcells with IgA and IgA-containing circulating immune complexes.Kidney Int 62:465–475, 2002

40. CROWLEY-NOWICK PA, CAMPBELL E, SCHROHENLOHER RE, et al:Polyethylene glycol precipitates of serum contain a large proportionof uncomplexed immunoglobulins and C3. Immunol Invest 25:91–101, 1996

41. KERBY JD, VERRAN DJ, LUO KL, et al: Immunolocalization of FGF-1 and receptors in human renal allograft vasculopathy associatedwith chronic rejection. Transplantation 62:467–475, 1996

42. SCHONERMARK M, DEPPISCH R, RIEDASCH G, et al: Induction of medi-ator release from human glomerular mesangial cells by the terminalcomplement components C5b-9. Int Arch Allergy Appl Immunol96:331–337, 1991

43. LOVETT DH, HAENSCH GM, GOPPELT M, et al: Activation of glomeru-lar mesangial cells by the terminal membrane attack complex ofcomplement. J Immunol 138:2473–2480, 1987

44. COUSER WG, PIPPIN JW, SHANKLAND SJ: Complement (C5b-9) in-duces DNA synthesis in rat mesangial cells in vitro. Kidney Int59:905–912, 2001

45. FUJII K, MULLER KD, CLARKSON AR, et al: The effect of IgA immunecomplexes on the proliferation of cultured human mesangial cells.Am J Kidney Dis 16:207–210, 1990

46. MONTEIRO RC, VAN DE WINKEL JG: IgA Fc receptors. Annu RevImmunol 21:177–204, 2003

47. DIVEN SC, CAFLISCH CR, HAMMOND DK, et al: IgA induced acti-vation of human mesangial cells: Independent of FcaR1 (CD 89).Kidney Int 54:837–847, 1998

48. WESTERHUIS R, VAN ZANDBERGEN G, VERHAGEN NA, et al: Humanmesangial cells in culture and in kidney sections fail to express Fcalpha receptor (CD89). J Am Soc Nephrol 10:770–778, 1999

49. LEUNG JCK, TSANG AWL, CHAN DTM, et al: Absence of CD89,polymeric immunoglobulin receptor, and asialoglycoprotein recep-tor on human mesangial cells. J Am Soc Nephrol 11:241–249, 2000

50. BARRAT J, GREER MR, PAWLUCZYK IZ, et al: Identification of a novelFca receptor expressed by mesangial cells. Kidney Int 57:1936–1948,2000

51. MOURA IC, CENTELLES MN, ARCOS-FAJARDO M, et al: Identificationof the transferrin receptor as a novel immunoglobulin (Ig)A1 recep-tor and its enhanced expression on mesangial cells in IgA nephropa-thy. J Exp Med 194:417–425, 2001

52. HADDAD E, MOURA IC, ARCOS-FAJARDO M, et al: Enhanced expres-sion of the CD71 mesangial IgA1 receptor in Berger disease andHenoch-Schonlein nephritis: Association between CD71 expres-sion and IgA deposits. J Am Soc Nephrol 14:327–337, 2003

53. SHIBUYA A, SAKAMOTO N, SHIMIZU Y, et al: Fca/l receptor mediatesendocytosis of IgM-coated microbes. Nature Immunol 1:441–446,2000

54. MCDONALD KJ, CAMERON AJM, ALLEN JM, et al: Expression ofFc a/l receptor by human mesangial cells: A candidate receptor forimmune complex deposition in IgA nephropathy. Biochem BiophysRes Commun 290:438–442, 2002

55. MOURA IC, ARCOS-FAJARDO M, SADAKA C, et al: Glycosylation andsize of IgA1 are essential for interaction with mesangial transfer-rin receptor in IgA nephropathy. J Am Soc Nephrol 15:622–634,2004

56. LEUNG JC, TANG SC, CHAN LY, et al: Polymeric IgA increases the syn-thesis of macrophage migration inhibitory factor by human mesan-gial cells in IgA nephropathy. Nephrol Dial Transplant 18:36–45,2003

57. COPPO R, AMORE A: Aberrant glycosylation in IgA nephropathy(IgAN). Kidney Int 65:1544–1547, 2004

58. AMORE A, CIRINA P, CONTI G, et al: Glycosylation of circulatingIgA in patients with IgA nephropathy modulates proliferation andapoptosis of mesangial cells. J Am Soc Nephrol 12:1862–1871, 2001

59. JULIAN BA, NOVAK J: IgA nephropathy: An update. Curr OpinNephrol Hypertens 13:171–179, 2004

Documents

IgA1-containing immune complexes in IgA nephropathy differentially affect proliferation of mesangial cells