Kidney International, Vol. 50 (1996), pp. 1946—1957

Enhanced production of glomerular extracellular matrix in a newmouse strain of high serum IgA ddY mice

ERI Muso, HARUYOSHI YOSHIDA, Eui TAKEUCHI, MASATOMO YASHIRO, HIR0YuKI MATSUSHIMA,ATSUSFH OYAMA, KATSUO SUYAMA, TAKAHIDE KAWAMURA, TADASHI KAMATA, SHIGEKI MIYAwAKI,

SHozo Izui, and SHIGETAKE SASAYAMA

Third Division, Department of Internal Medicine and Department of Pathology, Kyoto University, and Department of Nephrology, Kitano Hospital, Osakaand Research Laboratories, Nippon Shinyaku Co. Ltd., Kyoto, Japan; and Department of Pathology, CMU, University of Geneva, Geneva, Switzerland

Enhanced production of glomerular extracellular matrix in a newmouse strain of high serum IgA ddY mice. To investigate the relationshipbetween high serum levels of IgA and glomerular lesions, selective matingwas performed in high serum IgA ddY mice, a murine model ofspontaneously developing mesangioproliferative glomerulonephritis mim-icking human IgA nephropathy. The selection and mating of high IgA ddYmice were accomplished when the mice were three to four months old. Inthe 12th generation of high IgA ddY (HIGA) mice, significantly higherlevels of serum IgA from 10 age weeks to 60 weeks (P < 0.0002 to 0.000 1)were observed in comparison with BALB/c mice. Relatively high protein-uria was observed at 40 weeks of age, although hematuria was consistentlynegative. Microscopic observations of renal tissue disclosed a markedglomerular mesangial matrix increase and a reduction of cell proliferationwith age by both semiquantitative and morphometric analyses withmoderate tubulointerstitial damage. These mesangial matrices werestained markedly by antisera for collagen type IV and by fibronectin, butnot by collagen type I. Localization of TGF-f3 protein was also detected inthe mesangium of the HIGA mice. The positive mesangial IgA depositionwas maintained consistently by this mating procedure and became moremarked with age. Size analysis of IgA from ten pooled HIGA mice aged50 to 60 weeks revealed dominant polymeric IgA in sera and dimeric IgAin glomerular eluates. Clonal analysis of serum IgA disclosed heteroge-neous spectrotypes in a wide pH range (4.5 to 6.5), in contrast to verylimited spectrotypes in the acidic pH range (4.5 to 5.2) of IgA in theglomerular eluates from these mice. The analyses of retroviral gp7Oantigen involvement in the HIGA mice disclosed a significant increase ofserum levels of gp7O anti-gp7O immune complexes with age, with norelationship to the severity of glomerular gp7O deposition. Northern blotanalysis of renal tissue revealed markedly high mRNA expression ofcollagen type I, IV, fibronectin and TGF-/3 even in 10-week-old HIGAmice in comparison with BALB/c mice. The expression became moresignificant in 60-week-old animals. The genetic background required toinduce the expansion of IgA-producing B-cell clones is suggested to beclosely related to the increased gene expression of TGF-13, which inducesenhanced glomerular extracellular matrix (especially fibronectin) accumu-lation in HIGA mice, being possibly mediated by the mesangial depositionof dimeric and highly acidic IgA. This newly established strain may providea model for investigating the relationship between progressive glomerularsclerotic lesions and the induction of pathogenic IgA in human IgAnephropathy.

Received for publication October 26, 1995and in revised form July 12, 1996Accepted for publication July 15, 1996

© 1996 by the International Society of Nephrology

Since its first description by Berger [1, 2], immunoglobulin A(IgA) nephropathy has been shown to be the most commonglomerular disease worldwide [3]. It eventually progresses toend-stage renal failure in 30% to 35% of patients [4]. Thehistological criteria for the diagnosis of IgA nephropathy arepredominant mesangial IgA deposition with frequent associationof the deposition of complement 3 (C3). Microscopically, variousmesangial lesions are observed, from mild mesangial cell prolif-eration to marked mesangial expansion with or without extracap-illary lesions such as adhesion and crescent formation. Theincreased extent of global glomerular sclerosis is related toend-stage renal failure [5].

Serologically, from 35% to 50% of patients with IgA nephrop-athy have elevated serum IgA levels [6, 7]. Although this elevationof IgA is not always correlated with the severity and activity of thedisease [8], the significantly high incidence of elevated serumlevels of IgA, especially of the IgAl subclass derived from bonemarrow [9], may be related to the genetic background of thisdisease as well as to a higher production of IgA by B lymphocytes[10].

To investigate the pathogenesis of this renal disease, variousattempts have been made to develop an animal model. Immuni-zation with injection of antigens [11] or the direct injection ofpreformed immune complexes containing IgA [12] have provideda chance to elucidate the pathogenic role of multimers of lgAcontaining the J-chain. However, these IgA injections do notinduce the chronic lesions, especially those with sclerotic renaldamage, which are frequently observed in human IgA nephropa-thy.

In contrast, animal models that spontaneously develop mesan-gioproliferative lesions with IgA deposition more closely resemblethe human disease, although it takes much time to obtain overtglomerulonephritis. DdY mice were originally reported to de-velop spontaneous mesangioproliferative glomerulonephritis withIgA deposition after the age of 40 weeks [13]. Further investiga-tions have revealed the association of retroviral surface mem-brane glycoprotein (gp7O) antigen on the mesangial depositionwith a higher frequency of gp7O-containing immune complexes inthe sera of about 50% of these mice [14]. In addition to a highfrequency of elevated serum IgA levels with aging, these miceshow dominant polymeric or dimeric size and acidic charge IgA as

1946

Muso et at: ECM expansion in high igA strain of ddY mouse 1947

specific characteristics of serum and glomerular eluate IgA [151,which closely resemble those of human IgA nephropathy [16].

However, the marked individual differences of phenotype aris-ing from the crude genotype sometimes disturb the preciseinvestigation of pathological background in this mouse strain. Toestablish a genetically uniform model in which characteristicallyhigh serum IgA levels are maintained as in human IgA nephrop-athy, the selection and mating of ddY mice with high serum levelsof IgA have been performed. This procedure was recently foundto be successful in establishing a new ddY mouse strain, named"HIGA," showing consistently high IgA levels in the 12th gener-ation of mice at a young age (manuscript submitted).

The present study was performed to investigate whether or notthis newly established high IgA strain consistently maintains andexpresses the serological and immunopathological characteristicsresembling human IgA nephropathy. The renal expression oftransforming growth factor-n (TGF-) and its mRNA was inves-tigated, as well as that of extracellular matrix (ECM), collagentypes I and IV and fibronectin, since TGF-13 plays a crucial role inpromoting glomerular lesions through the accumulation of ECMleading to glomerular sclerosis [17—201.

Methods

Solid-phase IgA assay

IgA in the sera and in the samples fractionated by HPLC andIEF were measured by sandwich ELISA, using anti-mouse IgAantibodies (Cappel Laboratories Cochranville, PA, USA), asdescribed previously [15].

Urinaiy examination

Urinary protein and occult blood were examined when the micewere sacrifled, using Multisticks (AMES, Tokyo, Japan). Theresults were evaluated semiquantitatively from (—) to (+ + +) asfollows: urinary protein, (—) negative, (±) trace, (+) 30 mgldl,(++) 100 mg/dl, (+++) 300 mg/dl; occult blood, (—) negative,(±) trace, (+) < 20 red blood cell (RBC)/jxl, (++) 21 — 50

RBC/d, (+++) 50 250 RBC/jxl.

Establishment of high IgA ddY mice

Selective mating of high IgA ddY (HIGA) mice was performedas follows (manuscript submitted). Briefly, 52 female and maleddY mice of an outbred strain at eight weeks old (Japan SLC Inc.,Shizuoka, Japan) were used as the foundation colony (generation0: GO). They were maintained at Nippon Shinyaku Co. Ltd.(Kyoto, Japan). After verifying that there was a strong tendencyfor the GO mice with high serum lgA levels at the age of fourmonths to show higher serum levels of IgA with aging than did thegroups of mice with low IgA levels at the same age, six single-pairs(4 months old) with high serum IgA levels were selected andmated. The first generation (G1) showed higher serum IgA levelsat the age of four months than did those of GO, although markedindividual differences were noted (data not shown). An average ofsix single-pair matings was performed with 4-month-old mice untilG4, and thereafter with 3-month-old mice until the 12th genera-tion (G12), Throughout the mating procedure, the mice were fednormal laboratory chow and were not subjected to any experi-mental procedure to induce IgA immune responses. We desig-nated this newly established, high IgA strain of ddY mice asHIGA mice.

Pathological analysis of renal tissues

A portion of renal specimens from HIGA and BALB/c miceaged 10, 25, 40, and 60 weeks were fixed in Doubosque Brazilsolution and embedded in paraffin. Sections (2 /.tm) were stainedwith hematoxylin and eosin (HE) and periodic acid-Schiff (PAS).Semiquantitative microscopic analysis was performed with thegrading of (—) normal, (+) mild, (++) moderate, and (+++)severe being employed to denote the extent of mesangial cellproliferation and mesangial matrix expansion. Tubulointerstitialdamage was also evaluated from (—) to (+ + +) in the areas where100 glomeruli were counted, as follows: (—) no lesion of cellinfiltration and fibrosis, (+) single focus of lesion, (++) morethan two isolated foci, (+++) diffuse infiltration and fibrosis.Positivity for global glomerular sclerosis was defined by thepresence of more than one sclerotic glomerulus in the total countof 100 glonieruli. The presence of cellular crescent formation wasalso examined in 100 glomeruli.

To quantitatively evaluate the glomerular cell number per arearenal tissue sections obtained from ten HIGA mice (5 of each sex)at age 10, 25, 40, and 60 weeks, and six BALB/c mice (3 of eachsex) at age 10 and 60 weeks were randomly selected. With animage analyzer (Luzex 3U, Nikon Ltd., Tokyo, Japan), 10 selectedglomerular sections containing both vascular and urinary poleswere counted, and the cell numbers per area of the glomerulartuft are expressed as the number per 1000 /xm2. To evaluate theage-related changes of glomerular size in HIGA mice, the areas offifty randomly selected glomeruli in each section from the samemice as those used for the cell count were measured; the meanarea of larger half glomeruli in each section was compared inanimals of different ages.

Immunopathological examination

Serial immunofluorescence studies of IgG, IgA, 1gM, C3, andretroviral surface antigen glycoprotein (gp7O) were performed aspreviously described [13]. The grades of this deposition were alsoevaluated semiquantitatively, from absent (—) to severe (+ + +).All the histological examinations were performed independentlyby two investigators who had no knowledge of the experimentaldata. The staining of ECM was also performed for 40-week-oldHIGA and BALB/c mice by the indirect method using specificantibodies: polyclonal rabbit anti-mouse collagen type I antibody(Chemicon International Inc., Temecula, CA, USA), polyclonalrabbit anti-mouse collagen type IV antibody (Becton Dickinson,Bedford, MA, USA), polyclonal rabbit anti-mouse fibronectinantibody (Chemicon). Second antibody was FITC-labeled poly-clonal goat anti-rabbit IgG (Zymed Laboratories, San Francisco,CA, USA). The detection of TGF-/3 was performed by the directmethod using polyclonal rabbit antibody raised against a peptidecorresponding to amino acid residues 328 to 353 mapping withinthe carboxy terminal region of human TGF-131 (Santa CruzBiochemistry Inc., Santa Cruz, CA, USA) [21] that was labeledwith FITC in our laboratory. This antibody was proven to reactwith both human and murine TGF-pl [22].

Electron microscopical examination

Small blocks of the renal cortex were fixed in 2% glutaralde-hyde and then in 1% osmotic acid, after which they wereembedded in Epon 812. Ultrathin sections were then treated with

1948 Muso et al: ECM expansion in high IgA Strain of ddY mouse

Age(weeks) BALB/c N HIGA N p°

10 7.77 1.23" 17 23.0 2.37 19 0.000225 13.69 1.63 16 98.5 5.17c 14 <0.000140 22.67 2.07 22 103.38 3.29° 21 <0.000160 68.79 5.36 12 101.95 4.06° 21 <0.0001

uranyl acetate and lead citrate and observed by electron micros-copy (HITACHI H7100, Hitachi Ltd., Tokyo, Japan).

Charge and size analysis of IgA in sera and glomerular eluates

To examine the charge and size characteristics of serum andglomerular IgA, sera and glomerular eluates from ten HIGAmice, aged 50 to 60 weeks, wcre analyzed as described previously[15]. Briefly, the elution from isolated glomeruli, using a sievingtechnique [23], was performed by incubation for two hours with0.02 M citrate buffer, pH 3.5 [24]. After centrifugation, thesupernatant was obtained and dialyzed against phosphate-buff-ered saline (PBS), pH 7.4.

The electric charge characteristics of serum and glomerular IgAwere analyzed by agarose gel isoelectric focusing (IEF) using a flatbed apparatus electrophoresis unit (Pharmacia Fine Chemicals,Uppsala, Sweden). The gel solution contained 1% agarose IEF, 4M urea, 10% NP4O, 11% sorbitol, and 5% Ampholite (pH 3 to 10;LKB Productor, Bromma, Sweden). The cathodic electrode solu-tion was I M NaOH and the anodic solution was 0.5 M acetic acid.After focusing was performed for three hours at 1,500 V at 7°C,each sample gel was cut into 16 pieces and the sample solution ineach gel piece was eluted in 500 jxl of distilled water and dialyzedwith PBS overnight.

Size fractionation of the pooled sera and eluates was performedby I-IPLC on a 7.5 mm x 60 cm SW column (Tosoh, 5 to 1,000kDa fractionation range; Tosoh Ltd., Tokyo, Japan) with a flowrate of 0.5 mI/mm. Sixty fractions of each sample were collected in0.1 M phosphate buffer, pH 7.0, containing 0.1 M NaCI. Humangamma globulin (158 kDa) and thyroglobulin (670 kDa) wereused as molecular weight markers.

Measurement of serum gp7O and anti-gp7O immune complexes(gp7O IC)

The measurement of gp7O IC was performed as describedpreviously [25]. Briefly, IC-containing fractions in the sera ofHIGA mice were precipitated with 10% polyethylene glycol(PEG). The precipitates, redissolved in PBS, were added to thewells of a plastic plate that had been coated with goat anti-Rauscher gp7O antibodies and incubated. The plate was washedwith PBS and then further incubated with horseradish-peroxidase-conjugated rabbit anti-mouse immunoglobulin solution at roomtemperature for one hour. The wells were then washed andhydrogen peroxide and ortho-phenylenediamine solution wereadded as substrate. The absorbance was measured at a wavelengthof 490 nm. As the standard, gp7O IC at different concentrationspreformed between purified gp7O which had been obtained fromlipopolysaccharide (LPS)-injected New Zealand mice [26] and

10 15 9 4 2 0 0 18 0 11 5 2 025 14 10 4 0 0 0 14 1 3 8 2 040 15 10 3 2 0 0 21 1 0 8 8 4°60 20 9 10 1 0 0 20 0 5 9 5 1

anti-gp7O antibody were similarly applied, and absorbance wasplotted to construct the standard line.

Northern blot analysis of extracellular matrix and TGF-p

RNA from the kidneys of BALB/c and HIGA mic aged 10 and60 weeks was extracted by the acid guanidium thiocyanate-phenol-chloroform method [27]. Ten micrograms of total RNA wasfractionated in 0.66 M formaldehyde-1% agarose gel and trans-ferred to a nylon membrane (Biodyne PALL, Glen Cove, NY,USA) by capillary action. The following eDNA probes werelabeled with 32P using a random primed DNA labeling kit(Boehringer Mannheim Biochemica, Mannheim, Germany): forcollagen type I, a 3.0 kb EcoRI mouse collagen type I ul c-DNAfragment (a gift from Dr. C. Shiota of Wakayama PrefecturalSchool of Medicine); for collagen type IV, a 0.83 kbAval and PstImouse collagen type IV s1 c-DNA fragment (a gift from Dr. Y.Yamadaof, NIH, Bethesda, MD, USA) [28]; for TGF-p, a mouse0.96 kb c-DNA fragment (a gift from Dr. C. Shiota); for fibronec-tin, a mouse 2, 3 kb cDNA fragment (a plasmid from JCRB); andfor GAPDH, a human 1.2 kb DNA fragment (ATCC, MD, USA)[29]. The densities of each lane of the autoradiographs wereanalyzed with an NIH-Image analyzer.

Statistical analysis

The quantitative densitometric measurements of the Northernblots were normalized to the signal for GAPDH mRNA. Allvalues are expressed as means SE. Statistical significance wasdetermined by Student's 1-test for the comparison of two sets ofunpaired data, and the Scheffe F-test for multiple comparison.P < 0.05 was considered statistically significant.

Results

Serial analysis of serum IgA concentration of HIGA mice

Serial analysis of serum IgA at G12 of HIGA mice showedsignificantly higher levels of serum IgA, from as early as the age of10 weeks compared with BALB/c mice (P < 0.0002), and levels ofIgA were even more significantly marked from age 25 weeks (P <0.0001). These levels were maintained until age 60 weeks (P <0.05; Table 1). However, BALB/c mice also showed relatively highserum levels of IgA in comparison with younger mice.

Urinary findings of HIGA miceTable 2 shows the serial findings of proteinuria and hematuria

in HIGA mice. The number of mice showing proteinuria more

Table 1. Serum IgA concentration in BALB/c and HIGA miceaccording to age

Table 2. Serial urinary findings of HIGA mice

Hematuria° Proteinuria"Ageweeks N — N —

a Comparison between BALB/c and HIGA mice.h Serum IgA concentration (mean SE) mg/dl

Significantly higher than in 10-week-old HIGA mice (P < 0.001)

a Semiquantitative analysis: —,negative; trace; +, < 20 RBc/pl; + +,2—SO RBc/td; +++, 50-250 RBC/jJ

"Semiquantitative analysis: —, negative; trace; +, 30 mgldl; + +, 100mg/dl; +++, 300 mg/dl

P < 0.005 versus the mice with 10 and 25 weeks of age

Muso Ct al: ECM expansion in high IgA strain of ddY mouse

Table 3. Serial semiquantitative analyses of microscopical findings in HIGA and BALB/c mice

1949

MiceAge

weeks N

Mesangial lesionsCrescentformation

.Positivemic5 %

Globalsclerosis

.Positivemic&'%

Tubulointerstitial lesionCell proliferation Matrix expansion— + ++ +++— + ++ +++ — + ++ +++

HIGA10 19 1 7 10 1 7 12 0 0 9(47) 13(68) 16 3 0 025 14 0 7 7 0 0 9 5 0 7(50) 10(71) 9 5 0 040 21 0 5 12 4 0 6 12 3 3 (14) 20 (95) 0 14 6 1

60 21 5 9 5 2 0 6 13 2 4(19) 21 (100) 1 16 3 1

BALB/c10 17 9 8 0 0 17 0 0 0 0(0) 1(6) 17 0 0 025 18 11 7 0 0 14 4 0 0 0(0) 2(11) 18 0 0 040 22 12 10 0 0 4 17 1 0 0(0) 1(5) 20 2 0 060 12 5 6 1 0 1 7 3 0 0(0) 3(3) 10 20 0

Total number of miceb Number of mice with more than one glomerulus showing crescent formation or global sclerosis in total of 100 glomeruli

than 30 mgldl was relatively increased. There was no markedhematuria in HIGA mice throughout the life course.

Aging study of immunohistopathological findings of HIGA mice

Microscopic findings. Serial microscopic findings were examinedat the ages of 10, 25, 40, and 60 weeks in both HIGA and BALB/cmice. The results of the semiquantitative analyses are summarizedin Table 3. The analysis of mesangial lesions showed mild tomoderate cell proliferation at a young age in the HIGA mice. Atthe age of 60 weeks, 67% of the HIGA mice showed only mild tominimum cell proliferation. Analysis of the mesangial matrixrevealed a significant expansion with aging. At more than 40weeks of age, 90% of the HIGA mice showed moderate to markedmesangial matrix expansion in contrast to the BALB/c micegroup, in which only one mouse of 22 at the age of 40 weeks and25% of the mice at 60 weeks showed moderate mesangial matrixexpansion (Fig. 1). These tendencies were quantitatively sup-ported with findings of a significantly low cell number per area in60-week-old mice and a relatively higher number of cells inyounger mice. In contrast, in BALB/c mice, the average cellnumber per glomerular area did not change with age (Fig. 2A).The quantitative analysis of glomerular size in HIGA micerevealed a significant increase with age (Fig. 2B). The significantdecrease of cell number and the marked enlargement of glomer-ular size with age indicated massive expansion of the mesangialmatrix.

As shown in Table 3, the number of the HIGA mice withglomeruli showing crescent formation was relatively high inyounger mice. In the 40- and 60-week-old HIGA mice, only a fewshowed glomeruli with crescent formation. However, global gb-merular sclerosis was observed even in 10-week-old HIGA mice.It should be noted that in the HIGA mice the number of micewithout any global sclerosis decreased with age, and in mice agedmore than 40 weeks, all but one showed global sclerosis. As fortubulointerstitial lesions, focal cellular infiltration and fibrosiswere frequently observed in HIGA mice compared with BALB/cmice. A tubular degenerative change with vacuolation was ob-served more frequently in female HIGA mice from a younger age,although there was no correlation between the degree of thistubular damage and glomerular matrix expansion including en-largement of glomeruli (data not shown).

Immunofluorescence findings

As shown in Figure 3, in HIGA mice, there was no significantdifference in the grades of deposition of IgG, M, and C3 amongmice at 25, 40, and 60 weeks of age. In contrast, IgA depositionwas apparent even from 25 weeks of age. It should be noted thatthe frequency of HIGA mice with severe IgA deposition washighest at 60 weeks of age (Fig. 4). As for gp7O deposition, theratio of glomerular gp7O-positive mice was not significantlychanged with age, although the number of mice that showed verysevere glomerular gp7O deposition was highest in the 60-week-oldmice. The immunofluorescent analysis of ECM for 40-week-oldHIGA and BALBIc mice revealed a marked increase of glomer-ular fibronectin staining (Fig. 5 E, F) and mildly higher staining ofcollagen type IV (Fig. 5 C, D) in aged HIGA mice in comparisonwith BALB/c mice. It is also noted that fibronectin, which isnormally observed mainly in glomeruli, was markedly stained ininterstitium in HIGA mice. Although focal interstitial staining ofcollagen type I was noted in HIGA mice, glomerular staining ofthis collagen was consistently negative both in BALB/c and HIGAmice (Fig. 5 A, B). There was little individual differences in thisECM staining among the two groups of mice. Although thestaining of TGF-j3 was relatively faint, its apparent localization atthe mesangial area was observed dominantly in HIGA mice (Fig.6).

Electron microscopic findings

As shown in Figure 7, marked expansion of the mesangial areawith moderate electron dense deposits was observed in 60-week-old HIGA mice. In contrast, no mesangial cell proliferation wasobserved. In addition to mesangial lesions, thickening of thecapillary wall was also observed in aged HIGA mice. This capillarywall thickening was associated with segmental nodular formation,which has been reported as a marker of aging [30].

Charge and size characteristics of serum and glomerulus-boundIgA in aged HIGA mice

To examine the change in clonal and size characteristics ofserum and glomerular IgA by the selective mating, we performedIEF analysis and HPLC fractionation of IgA in 50- to 60-week-oldHIGA mice. As previously reported, in original ddY mice,

• S'S

• a

aaI.- -

I

-v.4,

S

S S IS

V

S

I

j'54

-_

7'S. S

II IS

S

IS

S

a.S

S-e

1950 Muso et a!: ECM expansion in high IgA strain of ddY mouse

Fig. 1. Microscopic photographs of glomendi inHIGA (a, b, c, d) and BALBIc (e, f, g, h) miceat 10 (a, e), 25 (c, f), 40 (c, g) and 60 weeks ofage (d, h). Marked enlargement of glomerularsize with matrix expansion was recognized withage in HIGA mice. In aged HIGA mice,particularly those aged 60 weeks, a significantdecrease in the mesangial cell population wasnoted. PAS staining (X680).

acid-restricted clonal expansion and prominent polymeric anddimeric IgA were shown in sera and, more particularly, in theglomerular eluate [15]. As shown in Figure 8, analysis of the clonalcharacteristics of serum IgA in aged HIGA mice disclosed aremarkably wide pH range from 4.2 to 6.5 (Fig. 8A). In contrast,the pH range of glomerular IgA was always limited to a highlyacidic range, of 4.5 to 5.2 (Fig. 8B).

Size fractionation disclosed a significant increase of polymericand dimeric IgA in the sera (Fig. 8C) and dominant dimeric IgAin the glomerular eluate (Fig. 8D). Monomeric IgA alwaysshowed the smallest peak in both the sera and the glomerulareluates.

Gp70 IC in sera of HIGA

As shown in Figure 9, although gp7O was not detectable in theserum in HIGA mice at 10 weeks old, it was detected at 25 weeksold, and increased significantly with age, with relatively large

individual differences. It was not correlated with the serum levelsof IgA and IgG.

Northern blot analysis of ECM and TGF-(3

Figure 10 shows the Northern blot analysis of collagen type I(A), type IV (B), fibronectin (C), TGF-13 (D) and GAPDH (E)mRNA levels in the renal tissues of HIGA and BALB/c mice aged10 and 60 weeks. The mRNA levels of collagen types I and IV andfibronectin were markedly high in HIGA mice from age 10 weekscompared with those of BALB/c mice. The intensity of theseexpressions in the HIGA mice did not increase significantly withage. The mRNA levels of TGF-13 were significantly increased inHIGA mice from age 10 weeks, and this expression increasedsignificantly with age. There was no marked difference of GAPDHmRNA levels between the two types of mice or between youngand old mice. It should be also noted that there were no marked

Muso et al: ECM expansion in high IgA strain of ddY mouse 1951

Fig. 2. Morphometric analyses of average cellnumber per area (A) and glomerular size (B) inthe renal tissues of selected HIGA mice (•) at10, 25, 40 and 60 weeks of age and in 10-and

° 60-week old-BALBIc mice (0). A significantlylow cell number per area was recognized in

o § aged HIGA mice compared not only withyounger HIGA mice but also with aged BALB/

o c mice, in which no remarkable changeoccurred with age. Although the glomerularsize increased with age in both strains of mice,significant enlargement of size was recognizedonly in aged HIGA mice compared withyounger mice (*P < 0.05, **p < 0.01, ***p <0.005, < 0.0001).

individual differences of mRNA levels, being particularly notablewith TGF-/3, among HIGA mice at 10 or 60 weeks of age.

Discussion

Our newly established high serum IgA strain ddY (HIGA) miceshowed increasing glomerular enlargement with age with signifi-cant matrix expansion, in contrast to decreased cellularity, onsemiquantitative observation. Electron microscopic analysis re-vealed marked mesangial widening, mostly occupied by a matrix-like substance and electron-dense deposits, with relatively raremesangial cells. These findings were supported by morphometricanalysis in which glomerular size was measured and glomerularcell number per area was counted. Immunopathological studiesrevealed a significant increase of mesangial fibronectin stainingwith marked IgA deposition. Furthermore, Northern blot analysisof renal tissue demonstrated markedly high mRNA expressions of

not only ECM, including collagen types I and IV and fibronectin,but also TGF-p, the protein of which was detected as expressed atthe mesangial area by IF staining. In contrast to the relatively highproteinuria observed in 40 weeks of age, hematuria was notconsistently recognized as in the original ddY mice [13]. Theseresults suggested that in HIGA mice, the already high productionand accumulation of extracellular matrix (ECM) especially in theglomerulus were enhanced by the high IgA selective matingprocedure we used via the increase of TGF-f3 production.

The immunostaining of ECM in aged HIGA and BALB/c micerevealed intense expression of fibronectin in glomeruli of HIGAmice, whereas collagen type IV was only relatively increased andcollagen type I was not detected in glomeruli, although partialincreased localization was noted in interstitium. These findingsindicated that among the significantly increased renal mRNAexpressions of these three types of ECM, fibronectin mRNA is

A I **** II I

SS

• :: :

I **r___.II ****PJI0 0

eJgI

00

(\1

E

a)

a)E

(a

a)E0

B60005500

5000

4500

4000

3500

3000

2500

2000 -

:

.ii

2422

w 20= 18

1614a)o 12

10

8

4W

*C.2 +C')00.a)

(a

a)E00)0a)-a

0

10 25 40 60 10 60 10 25 40 60 10

Age, weeks Age, weeks

A B C

::h ::h :ff

•:10 25 40 60

.:.. S.. ..:.

•••• ::r

:u:: up

. ..

zzr nu

::h: .::

dill

10 25 40 60E

10 25 40 60D

44

+ ... •5• •• •..

S • •••

"r •u: u

••• ::: '::10 25 40 60 10 25 40 60

Age, weeks

Fig. 3. Semiquantitative analysis byimmunofluorescence staining of the glomerulardeposition of immunoglobulins, C3, and gp7O in10-, 25-, 40- and 60-week-old HIGA mice. Therewere no significant differences in the grades ofIgG (A), 1gM (C) and C3 (D) deposition amongthe 25-, 40- and 60-week-old mice. In contrast,the number of mice with intensely positive IgAdeposition increased with age (B). Although thenumber of glomerular gp7O-positive mice didnot increase significantly with age, the numberof mice that showed very severe glomerulargp7O deposition was highest at 60 weeks.

U.

r'a

1952 Muso et al: ECM expansion in high IgA strain of ddY mouse

Fig. 4. Immunofluorescence photographs ofglomerular IgA deposition in 10-week-old (A) and60-week-old (B) HIGA mice. Severe mesangialand paramesangial deposition of IgA was notedin aged HIGA mice (><990).

Fig. 5. Immunofluorescent photographs ofcollagen type I (A, B), collagen type IV (C, D)and fibronectin (E, F) in renal tissues of BALB/c(A, C, E) and HIGA (B, D, F) 40-week-oldmice. Markedly intense fibronectin staining anda mild increase of collagen type IV stainingwere noted in the glomeruli of HIGA mice.Collagen type I was not consistently detected inglomeruli of either strain, although a focalincrease of interstitial staining was observed inHIGA mice (B) (X680).

suspected to be the main contributor to the progression ofglomeruloscierosis. Collagen type I, which is normally expressedin interstitium, is frequently expressed in sclerosing glomeruli withintensive inflammatory change, and is sometimes recognized asthe marker of abnormal transformation of glomeruli due to active

glomerulonephritis [31]. No abnormal localization of glomerularcollagen type I in the HIGA mice of the present study indicatedthat phenotypic change of glomerular cells did not occur as aresults of this mating protocol, but the enhancement of ECMformation of glomerular cells was induced. This might be the

me

Muso et al: ECM expansion in high IgA strain of ddY mouse 1953

Fig. 6. Glomerular TGF-j3 localization in 40-week-old mice. An apparent expression of TGF-p in the mesangial area was noted in HIGA mice (B) incontrast to the very faint expression in BALB/c mice (A). Reproduction of this figure in color is made possible by a grant from Nippon Shinyaku Co.Ltd., Kyoto, Japan.

Fig. 7. Electron microscopical photograph of theglomenilus of a HIGA mouse at age 60 weeks.Marked expansion of the mesangial matrixarea, with electron-dense deposits, localizedmainly in the paramesangial area (arrow), wasnoted. The mesangial cell number was relativelylow. Along the capillary wall, in addition to thediffuse thickening, some nodular formation ofelectron dense-deposits was observed.

cause of the relatively moderate and slowly progressive character-istics of the glomerulosclerosis in the HIGA mice, of which only 3and 2 mice showed several lesions at 40 and 60 weeks of age,respectively.

The findings of glomerular cellularity were somewhat contraryto those in the original ddY mice. in the original ddY mice,mesangial cell proliferation was relatively increased with age, andwas sometimes accompanied by active cellular crescent formation,although there were substantial individual differences among mice[141. In contrast, in the HIGA mice the glomerular cellularity aswell as crescent formation were significantly decreased with age.The quantitative study of glomerular cells per area clearly showedsignificant decreases of the cells at age 60 weeks not only incomparison with younger HIGA mice but also in comparison withnormal BALB/c mice of the same age. It is also noteworthy that,at 10 weeks, the cellularity of the HIGA mice was relatively higherthan that in normal young BALB/c mice. Although a preciseanalysis of whether this increase in glomerular cells originatedfrom circulating cells or resident glomerular cells was not per-

formed, the abnormally decreased cellularity with age indicatedthat this mating protocol may have induced some factor thatsuppressed the proliferation of resident glomerular cells, possiblymesangial cells.

TGF-/3, which has been reported to play a crucial role in thedevelopment of glomeruloscierosis in anti-Thy 1 glomerulone-phritis and diabetic nephropathy [32, 33], has potent regulatoryeffects in inflammation, immune response, and, especially, inextracellular matrix production [34]. Several in vitro investigationsin which TGF-j3 was employed to stimulate cultured renal paren-chymal cells showed direct evidence of the induction of extracel-lular matrix production of collagen type IV, fibronectin, laminin,and biglycan [17, 18]. In IgA nephropathy patients, TGF-f3 wasshown to be expressed in the glomerulus of biopsied tissue by animmunofluorescence technique that employed polyclonal andmonoclonal antibodies [35]. Increased TGF-p mRNA expressionhas also been observed not only in glomerular cells but also ininfiltrative monocyte/macrophages by in situ hybridization [36].These reports emphasize that the significant correlation between

10 weeks 60 weeks

0 St —

fl

s41I II

HIGA BALB/c

I...It

HIGA BALB/c

Bfl$SR.IP

1954 Muso et al: ECM expansion in high IgA strain of ddY mouse

0.0110 25 40 60

Age, weeks

Fig. 9. Serum levels of gp70 anti-gp7O antibody immune complexes mea-sured by PEG precipitation methods, in the sera of 10-, 25-, 40- and60-week-old HIGA mice. Marked increases of IC levels were noted,particularly in the 40- and 60-week-old mice. Relatively high individualdifferences were seen (*P < 0.05, < 0.0001).

TGF-13 expression and renal damage indicates the pathologicalrole played by this cytokine in the progression of glomeruloscle-rosis in IgA nephropathy. TGF-p has some controversial effectson the proliferative response of mesangial cells in vitro, dependingon the species, cell density, incubation period, and transformationof the cells [37, 38]. Haberstroh and colleagues [38] have demon-strated that the proliferation of non-transformed cultured ratmesangial cells was stimulated by TGF-/3 after 48 hours, whereasthe incubation of quiescent cells with TGF-j3 for 24 to 48 hours

Fig. 10. Northern blot analysis of collagen type I (A), collagen type IV (B),fibronectin (C) and TGF-/3 (D) mRNA levels in renal tissues from HIGA andBALB/c mice JO and 60 weeks old. Levels of GAPDH mRNA (E) servedas the housekeeping gene. The mRNA levels of all the above extracellularmatrix components and of TGF-/3 were markedly high in HIGA mice,even from age 10 weeks, compared with those in BALB/c mice. Eachvertical lane represents the mRNA expression of each mouse.

had an inhibitory effect on cell proliferation in subconfluent cellsand no effect in confluent cells. In contrast, in a transformedmurine mesangial cell line, TGF-p exhibited an antiproliferativeeffect. These effects were confirmed to be TGF-/3-specific sincemonoclonal antibodies against TGF-f3, added simultaneously,attenuated these reactions. Considering these effects of TGF-/3 onmurine cell lines, we speculated that the significantly higherexpression of TGF-f3-mRNA in renal tissue obtained from youngand aged HIGA mice in comparison with normal BALB/c mice

C

670k 158k1.8

1.6

1.4'1 1.200

0.8C) 0.6

0.40.2

0

1 .4

1.201

0o 0.8

0.6

0.4

0.2

0

B3 5 7 9 11 13 15Fraction number

9

8

7

6

5

4

3

9

8

7

6

5

4

3

in0IC)

to0

10C)

0.8

0.6

0.4

0.2

0

0.5

0.4

0.3

0.2

0.1

0 70 80 90 100 110D Fraction number

120

670k 158k

1 3 5 7 9 11 13 15 0 70 80 90 100 110

Fig. 8. Clonal (A, B) and size (C, D) analysis ofIgA in pooled sera and glomenilar eluates fromten HJGA mice age 50 to 60 weeks. Isoelectricfocusing in agarose gel showed markedexpansion of the serum IgA pH range, thisbeing 4.2 to 6.5 (A). In contrast, glomerulareluate IgA had a consistently limited acid pHrange of 4.2 to 5.2 (B). Size analysis by HPLCfractionation revealed a significant increase ofpolymeric and dimeric IgA in the sera (C) anddominant dimeric IgA in the glomerular eluates(D).Fraction number

120

I **I **

Fraction number

I I100

• .S• •

_10 : •I

SISISI0 T .1N-0 SI SI SIC) I •l .1E I •I 'I•l .1 .1I gI •

0.1 II SII

.1• S

Muso et al. ECM expansion in high IgA strain of ddY mouse 1955

and the apparent localization of TGF-j3 in mesangium in HIGAmice was closely related to the marked ECM production andaccumulation, and abnormally decreased cellularity in the glomer-uli in aged HIGA mice.

In ddY mice, as we have previously reported, significantlyelevated serum IgA was noted after 40 weeks of age althoughthere were marked individual differences [151. This finding, takentogether with the consistently high serum IgA levels and wide pHrange of serum IgA determined by IEF analysis in HIGA mice,leads us to presume that the mating protocol performed in thisstudy selected a strain in which IgA-bearing B-cell clones, withelevated production or secretion of IgA, were greatly increased. Itis also highly suspected that these phenomena are closely relatedto a higher expression of TGF-p, since TGF-p is not only aregulatory factor in promoting ECM production but also acts asan immunoregulatoiy cytokine in inducing the IgA isotype switch-ing of mature IgA-producing B cells, while it also plays asuppressive role in IgG and 1gM secretion [39, 40]. Furtherinvestigation of the cellular immunity system and TGF-j3 produc-tion in HIGA mice may provide much information not only of thebackground mechanism on the high serum IgA levels, but also onthe regulatory abnormalities of glomerular ECM accumulationcausing glomerular sclerotic lesions.

The increasing IgA deposition with age in HIGA mice demon-strated that IgA increased by this mating protocol may maintainthe nephritogenic characteristics shown in the original ddY mice.As we have reported previously, the clonal characteristics ofglomerulus-bound IgA in ddY mice, analyzed by IEF, revealed arelatively acidic pH range in comparison with serum IgA [151. InHIGA mice at the age of 50 to 60 weeks, a remarkably wide pHrange of serum IgA was observed. In contrast, the pH range ofglomerulus-bound IgA was always limited to the acidic range. Thisfinding suggests that, although this selective mating protocolinduced the polyclonal expansion of IgA-producing B cells inthese mice, the nephritogenic IgA is always oligoclonally acidic, aspreviously observed in human IgA nephropathy [16]. As for thesize characterization, the significant increase of polymeric anddimeric serum IgA and the dominant dimeric glomerulus-boundIgA in HIGA mice resemble the characteristics of IgA in theoriginal ddY mice [15]. It is also suggested that polymeric size orpolymerization of IgA is essential for IgA accumulation in themesangium, which is also sometimes observed in human igAnephropathy [161. These results indicate that the clonal expansionof specific B cells that secrete nephritogenic IgA, resemblinghuman IgA nephropathy, was possibly induced and maintained bythis mating protocol.

Considering the findings that glomerular IgA deposition wasnot apparent at the age of 10 weeks when 50% of the HIGA micealready express some glomeruli with global sclerosis, high serumIgA is not necessarily the leading cause of global glomerularsclerosis. However, the marked expansion of mesangium associ-ated with increasing IgA deposition with age indicated that thespecific IgA can be at ributed to the progression of the glomeru-losclerosis in HIGA mice. Furthermore, rare glomerular lesionsof control BALB/c mice that showed moderately high serum levelsof IgA with aging indicated the specific increase of nephritogenicIgA in HIGA mouse.

The serial increase of the serum levels of gp7O IC with age inHIGA mice, measured by the PEG precipitation method, were ofconsiderable interest. In our previous report of original ddY mice

[141, levels of circulating gp7O were measured in pooled sera andrevealed various amounts of serum gp 70 in mice as young as 12weeks without any apparent increase with age. However, theadsorption test performed with pooled sera, using columns coatedwith antibodies against mouse IgA and IgG, revealed that about50% of gp7O detected in the aged mice (aged 40 weeks) was theimmunoglobulin-bound form, while in contrast, no gp7O IC wasdetected in younger ddY mice (aged 12 weeks). The significantincrease of gp7O IC in HIGA mice in the present study wasconsistent with these results in ddY original mice, although thedetection was performed in individual sera, As indicated by theimmunohistopathological findings of the lymphoid tissue, in ddYmice [14], the source of gp7O antigen in HIGA mice appears to bein the lymphoid tissues and the secretion of the antigen was at ayounger age. The probable accelerated production of gp7O-specific antibodies with age in these mice did not appear to bechanged by the selective mating protocol performed in this study.

As for the glomerular deposition of gp7O antigen, in ourprevious study of ddY mice [14], we speculated that circulatinggp7O IC was deposited in the glomerulus rather than the directdeposition of gp7O antigen, since there was no significant corre-lation between the serum levels of gp7O and the extent ofglomerular gp7O deposition in those mice. Other investigators,using the same antibodies as ours, have since reported that thepathogenic role of gp7O in ddY mice was negligible, as they couldnot detect this antigen in some strains [41]. In our newly estab-lished HIGA strain, although there was no correlation betweenthe grade of glomerular deposition of gp7O and the amount ofserum IC, it is possible that the serum IC detected by our PEGprecipitation method may have been involved, to some extent, inthe mesangial lesion. The limited localization of gp7O depositiononly in the mesangial matrix in ddY and HIGA mice suggests thatlarger size IC molecules were trapped in the mesangium. How-ever, IC detected by the PEG precipitation method may includesmaller as well as larger molecules, and this may be the cause ofthe discrepancy between the glomerular findings and the serumgp7O IC amount. This result was somewhat in contrast to thefindings of renal lesions in lupus-prone New Zealand mice, whichshowed significant glomerular deposition of gp7O dominantlyalong the capillary wall, this being well correlated with serumlevels of gp7O IC [4 1—44].

In summary, the selective mating protocol by which we estab-lished HIGA mice appears to have maintained and augmentedthe original nephritogenic characteristics, especially those relatedto glomerular sclerotic lesions. The genetic background againstwhich the clonal expansion of IgA-producing B cells is inducedappears to be closely related to glomerular ECM accumulation, aphenomenon which is possibly mediated by the overexpression ofTGF-f3. Further investigations of the relationship between thesetwo events in this newly developed HIGA strain may help toelucidate the pathogenic background of the progressive glomer-ular sclerosis that occurs in IgA nephropathy patients.

Acknowledgments

This work was partly supported by a Grant-in-Aid for General ScientificResearch from the Japanese Ministry of Education, Science, and Cultureand by a grant from the Swiss National Foundation for Scientific Re-search. Publication of Figure 6 in color was made possible by a grant fromNippon Shinyaku Co. Ltd., Kyoto, Japan. We are indebted to Prof. K.Ooshima, and Dr. C. Shiota, Department of Pathology, Wakayama

1956 Muso et al. ECM expansion in high IgA strain of ddY mouse

Prefectural School of Medicine for the gift of the eDNA prove of collagentype I and TGF-J3 and for helpful suggestions; to Dr. Y. Yamada for thegift of the eDNA probe of collagen type IV; to Dr. H. Shibata, ResearchLaboratories, Nippon Shinyaku Co. Ltd., Kyoto, for assistance in thebreeding of mice and to Mr. M. Kurino, Division of Clinical Pathology,Kyoto University Hospital, Kyoto, for assistance with electron microscopy.We would also like to thank Ms. Y. Sueno for her technical assistance andMs. T. Yoshinaga for preparing the manuscript.

Reprint requests to En Muso, M.D., Third Division, Department of InternalMedicine, Faculty of Medicine, Kyoto University, 54 Kawaracho, Shogoin,Sakyo-ku, Kyoto 606, Japan.

Appendix

Abbreviations are: C3, complement 3; ECM, extracellular matrix;ELISA, enzyme-linked inimunosorbent assay; GAPDH, glyceraldehyde-3-phosphate-dehydrogenase; HIGA, high IgA ddY strain; HPLC, highperformance liquid chromatography; IC, immune complexes; IEF, isoelec-tric focusing; LPS, lipopotysaccharide; PAS, periodic acid-Schiff; PBS,phosphate buffered saline; PEG, polyethylene glycol; RBC, red blood cell;TGF-13, transforming growth factor /3.

References

1. BuRGER J, HINGLAIS N: Intercapillary deposits of IgA-IgG. J UrolNephrol Paris 74:694—695, 1968

2. BERGER J: IgA glomerular deposits in renal disease. Transplant Proc1:939—944, 1969

3. JULIAN BA, WALDO FB, RIFAI A, MESTECKY J: IgA nephropathy, themost common glomerulonephritis worldwide. A neglected disease inthe United States? Am J Med 84:129—132, 1988

4. CHAUVEAU D, DROZ D: Follow-up evaluation of the first patients withIgA nephropathy described at Necker Hospital. Contrib IVephrol104:1—5, 1993

5. McCoy RC, ABRAMOWSKY CR, TISHER CC: lgA nephropathy. Am JPathol 76:123—144, 1974

6. D'AMlco G: Clinical features and natural history in adults with IgAnephropathy. Am J Kidney Dis 12:353—357, 1988

7. WOODROFFE AJ, CORMLY AA, MCKENZIE PE, WoorroN AM,THOMPSON AJ, SEYMOUR AE, CLARKSON AR: Immunologic studies inIgA nephropathy. Kidney mt 18:366—374, 1980

8. BERGER J: Idiopathic mesangial deposition of IgA (chapt 32), Ne-phrology, edited by HAMBURGER J, CROSNIER J, GRUNFELD J-P, NewYork, Wiley-Flammarion, 1979, pp 535—541

9. LAYwARD L, ALLEN AC, HAYrERSLEY JM, HARPER SJ, FEEHALLY J:Elevation of IgA in IgA nephropathy is localized in the serum and notsaliva and is restricted to the IgA 1 subclass. Nephrol Dial Transplant8:25—28, 1993

10. CAGNOLI L, BELTRANDI E, PASQUALI 5, Bicu R, CASADEI-MALDINIM, Rossl L, ZUCCHELLI F: B and T cell abnormalities in patients withprimary IgA nephropathy. Kidney mt 28:646—651, 1985

11. LAMM ME, EMANCIPATOR SN, GALLO GR: Relevance of an experi-mental model to clinical IgA nephropathy. Cont rib Nephrol 40:32—36,1984

12. RIFAI A, SMALL P JR, AYOUB EM: Experimental IgA ncphropathy.Factors governing the persistence of IgA-antigen complexes in thecirculation of mice. Contrib Nephrol 40:37—44, 1984

13. IMAI H, NAKAMOTO Y, ASAKURA K, MIK! K, YASUDA T, MIuRA AB:Spontaneous glomerular IgA deposition in ddY mice: An animalmodel of IgA nephritis. Kidney mt 27:756—761, 1985

14. TAKEUCHI E, Dol T, SHIMADA T, MUSO E, MARUYAMA N, YOSHIDAH: Retroviral gp7O antigen in spontaneous mesangial glomerulone-phritis of ddY mice. Kidney mt 35:638—646, 1989

15. Muso E, YOSHIDA H, TAKEUCHI E, SHIMADA T. YASIIIRO M, SUO-IYAMA T, KAWAI C: Pathogenic role of polyclonal and polymcric IgAin a murine model of mcsangial proliferative glomcrulonephritis withIgA deposition. Clin Exp Immunol 84:459—465, 1991

16. MONTEIRO RC, HALBWACI-IS-MECARELLI L, ROQUE-BARREIRA MC,NOEL LH, BERGEIS J, LESAVRE P: Change and size of mesangial IgAin IgA nephropathy. Kidney mt 28:666—671, 1985

17. BORDER WA, OKVDA S, LANG VINO LR, RUOSLAHTI E: Transforminggrowth factor-/I regulates production of proteoglycans by mesangialcells. Kidney mt 37:689—695, 1990

18. NAKAMURA T, MILLER D, RUOSLAHTI E, BORDER WA: Production of

extracellular matrix by glomerular epithelial cells is regulated bytransforming growth factor-/Il. Kidney mt 41:1213—1221, 1992

19. OKUDA 5, LANGUINO LR, RUOSLAHTI E, BORDER WA: Elevatedexpression of transforming growth factor-beta and proteoglycan pro-duction in experimental glomerulonephritis. Possible role in expan-sion of the mesangial extracellular matrix. J Clin invest 86:453—462,1990 lerratum, J Clin Invest 86:2175, 1990}.

20. TAMAKI K, OKUDA S, ANDO T, IwAM0T0 T, NAKAYAMA M, FUJISHIMAM: TGF-/31 in glomerulosclerosis and interstitial fibrosis of adriamy-cm nephropathy. Kidney mt 45:525—536, 1994

21. DERYNIK R, JARRE1-F JA, CHEN EY, EATON DH, BELL JR, ASSOIANRK, ROBERTS AB, SPORN MB, GOEDDEL DV: Human transforminggrowth factor-/I complementary DNA sequence and expression innormal and transformed cells. Nature 316:701—705, 1985

22. TAN CEL, CHAN VSW, YONG RYY, VIJAYAN V, TAN WL, CHONGSMCF, Ho JMS, CHENG HH: Distortion in TGF/31 peptide immuno-localization in biliary atresia: Comparison with the normal pattern inthe developing hyman intrahepatic bile duct system. Pathollnt 45:815—824, 1995

23. KREISBERG JI, HOOVER FL, KARNOVSKY MJ: Isolation and character-ization of rat glomerular epithelial cells in vitro. Kidney mt 14:21—30,1978

24. WOODROFFE A, WILLSON CB: An evaluation of elution techniques inthe study of immune complex glomerulonephritis. J Immunol 118:1788—1 794, 1977

25. MERINO R, SHIBATA T, DE-KOSSODO 5, I 5: Differential effect ofthe autoimmune Yaa and lpr genes on the acceleration of lupus-likesyndrome in MRL/MpJ mice. EurJlmmunol 19:2131—2137, 1989

26. KENNEL Si: Purification of a glycoprotein from mouse ascites fluid byimmunoaffinity chromatography which is related to the major glyco-protein of murine leukemia viruses: Immunologic and structuralcomparison with purified viral glycoproteins. J Biol Chem 251:6197—6204, 1976

27. CHOMCZYNSKI P, SACCHI N: Single-step method of RNA isolation byacid guanidinium thiocyanate-phenol-chloroform extraction. AnalBiochem 162:156—159, 1987

28. OBERBAUMER I, LAURENT M, SCHwARZ U, SAKURAI Y, YAMADA Y,VOGELI G, Voss T, SIEBOLD B, GLANVILLE RW, KUHN K: Amino acidsequence of the non-collagenous globular domain (NC1) of the alphaI (IV) chain of basement membrane collagen as derived fromcomplementary DNA. Eur J Biochem 147:217—224, 1985

29. Tso JY, SUN XH, K.Ao TH, REECE KS, WU R: Isolation andcharacterization of rat and human glyceraldehyde-3-phosphate dehy-drogenase cDNAs: Genomic complexity and molecular evolution ofthe gene. NucI Acid Res 13:2485—2502, 1985

30. DUAN HJ, NAGATA T: Glomerular extracellular matrices and anionicsites in aging ddY mice: A morphometric study. Histochemistiy99:241—249, 1993

31. Y05HIOKA K, TOHDA M, TAKEMURA T, AKANON, MATSUARA K,O05HIMA A, MAKI S: Distribution of type I collagen in human kidneydiseases in comparison with type III collagen. J Pathol 162:141—148,1990

32. BORDER WA, OKUDA 5, LANOUINO LR, SPORN MB, RUOSLAHTI E:Suppression of experimental glomerulonephritis by antiserum againsttransforming growth factor /31. Nature 346:371—374, 1990

33. YAMAMOTO T, NAKAMURA T, NOBLE NA, RUOSLAHTI E, BORDERWA: Expression of transforming growth factor beta is elevated inhuman and experimental diabetic ncphropathy. Proc NatI Acad SciUSA 90:1814—1818, 1993

34. BORDER WA, RUOSLAHTI E: Transforming growth factor-/I in disease:The dark side of tissue repair. J Clin invest 90:1—7, 1992

35. LAI KN: Increased mRNA encoding for transforming growth factor-/Iin CD4+ cells from patients with IgA nephropathy. Kidney mt46:862—868, 1994

36. YOSHIOKA K, TAKEMURA T, MURAKAMI K, OKADA M, HINO S,MIYAMOTO H, MAKI 5: Transforming growth factor-/I protein andmRNA in glomeruli in normal and diseased human kidneys. LabInvest 68:154—163, 1993

37. MACKAY K, STRIKER U, S1'AUFFER JW, Doi T, AGODOA LY, STRIKER

Muso et at: ECM &pansion in high IgA strain of ddY mouse 1957

GE: Transforming growth factor-/3. Murine glomerular receptors andresponses of isolated glomerular cells. J Gun Invest 83:1160—1167,1989

38. HABERSTROH U, ZAHNER G, DISSER M, TI-IAISS F, WOLF 0, STAHLRA: TGF-13 stimulates rat mesangial cell proliferation in culture: Roleof PDGF /3-receptor expression. Am J Pathol 33:F199—F205, 1993

39. COFFMAN RL, LEBMAN DA, SHRADER B: Transforming growth factor

beta specifically enhances IgA production by lipopolysaceharide-stimulated murine B lymphocytes. J Exp Med 170:1039—1044, 1989

40. SONODA F, MATSUMOTO R, HIT0sHI Y, Istui T, SUGIM0T0 M, ARAKIS, TOMINAGA A, YAMAGUCHI N, TAKATSU K: Transforming growthfactor /3 induces IgA production and acts additively with interleukin 5for IgA production. J Exp Med 170:1415—1420, 1989

41. SHIMIzU M, TOMINO Y, ABE M, SHIRAI T, K0IDE H: Retroviralenvelope glycoprotein (gp7O) is not a prerequisite for pathogenesis of

primary immunoglobulin A nephropathy in ddY mice. Nephron62:328—331, 1992

42. Y05HIKI T, MELLORS RC, STRAND M, AUGUST JT: The viral envelopeglycoprotein of murine leukemia virus and the pathogenesis ofimmune complex glomerulonephritis of New Zealand mice. J Exp Med140:1011—1027, 1974

43. Nii Y, MARUVAMA N, OHTA K, YOSHIDA H, HIR05E S, SHIRAi T:Genetic studies of autoimmunity in New Zealand mice. Association ofcirculating retroviral gp7O immune complex with proteinuria. Immu-nol Left 2:53—58, 1980

44. MARUYAMA N, FURUKAwA F, NAKAL Y, SASAKI Y, OHTA K, OzAxi S,HIROSE S, SHIRA! T: Genetic studies of autoimmunity in New Zealandmice. IV. Contribution of NZB and NZW genes to the spontaneousoccurrence of retroviral gp7O immune complexes in (NZB X NZW)Fl hybrid and the correlation to renal disease. J Immunol 130:740—746, 1983