Search

Focal segmental glomerulosclerosis (FSGS) is a common cause of end-stage renal disease. Albuminuria is a risk factor for FSGS and is influenced by environmental, genetic, and sex-specific factors. Podocytes play a central role in the development of albuminuria, but the precise relationship between early glomerular and podocyte-associated damage and albuminuria is unclear. Furthermore, experimental findings demonstrate a sex difference in development of albuminuria and FSGS. We investigated the early glomerular changes in male Munich-Wistar-Frömter (MWF) rats, which spontaneously develop albuminuria, and male albuminuria-resistant spontaneously hypertensive rats (SHR). In addition, since female MWF rats are protected from overt proteinuria and progressive renal disease, we compared the phenotypic changes in podocytes during early development of albuminuria in male and female MWF rats. In male MWF rats, glomerular hypertrophy preceded the onset of albuminuria and was greater than in male SHR. Albuminuria developed starting at 6 wk of age and coincided with focal and segmental loss of podoplanin, increased expression of desmin, entrapment of albumin in affected podocytes, and focal and segmental foot process effacement at the ultrastructural level. Other podocyte-associated molecules, such as nephrin and zonula occludens 1, were unaffected. Early glomerular hypertrophy and podocyte damage did not differ between male and female MWF rats. Our data show for the first time that albuminuria in male and female MWF rats is preceded by glomerular hypertrophy and accompanied by focal and segmental loss of podoplanin when FSGS was not yet present.

Although it is clear that genetic predispositions play a role in progressive glomerulosclerosis, identification of specific genes is difficult because of natural genetic heterogeneity among individuals. We have reported a differential susceptibility to progressive glomerulosclerosis after induction of experimental glomerulonephritis anti-Thy-1 nephritis in Lewis rat substrains. Glomerular lesions in Lewis/Møllegard rats resolve spontaneously, whereas Lewis/Maastricht (Lew/Maa) rats develop progressive glomerulosclerosis. This predisposition for progressive glomerulosclerosis is governed by unknown genes that are expressed by renal cells. Here, differential gene expression analysis using a rat complementary DNA micro array revealed neuronal activity-regulated pentraxin (Narp) as a candidate gene involved in the remodeling or progression of damaged glomeruli. Glomerular Narp mRNA expression was monitored during disease in both Lewis sub strains. Immunohistochemistry revealed that Narp protein is exclusively expressed in Lew/Maa glomeruli 7 and 14 days after induction of anti-Thy-1 nephritis. Double-immunofluorescent staining showed that proliferating mesangial cells and parietal epithelial cells (PECs) at sites of adhesion to podocytes are partially Narp-positive, whereas podocytes fail to express Narp. Immunohistochemistry in nephritic Wistar, unilaterally nephrectomized Wistar and Sprague-Dawley rats showed that Narp protein is present only in strains that develop progressive glomerulosclerosis but never in strains that show remodeling. We conclude that Narp is a predictor for anti-Thy-1 nephritis-induced glomerulosclerosis and its expression by PECs may be involved in the progression to glomerulosclerosis.

Extracapillary proliferation and fibrinoid necrosis are the main diagnostic glomerular lesions in renal biopsy specimens of patients with idiopathic systemic vasculitis. Neither the incidence nor the correlation between extracapillary proliferation and fibrinoid necrosis in renal biopsy specimens from patients with systemic vasculitis has been systematically evaluated. By means of a three-dimensional analysis, we made a topographic reconstruction of the distribution of extracapillary proliferation and fibrinoid necrosis in affected glomeruli and tested different biopsy-processing protocols to optimize histopathologic analysis in clinical practice. Paraffin blocks of renal biopsy specimens from six patients diagnosed with systemic vasculitis were completely and serially sectioned in 2-microm thick sections and stained with the Gomori trichrome method. Glomeruli were scored per section for the presence of fibrinoid necrosis and extracapillary proliferation. Subsequently, a three-dimensional reconstruction was obtained for 87 glomeruli. In only one glomerulus did fibrinoid necrosis occur without extracapillary proliferation; in 51%, a combination of the two lesions was found; in 22%, extracapillary proliferation occurred in the absence of fibrinoid necrosis; and 26% did not show either lesion. Using the standard protocol from our department (ie, evaluation of 20 consecutive sections in various stainings), the chance of finding extracapillary proliferation was 100% and that of finding fibrinoid necrosis was 73%. If 5 sections stained with the Gomori trichrome were added, the latter percentage increased to 86%. Using skip-serial sections, even better results (87% to 92%) were obtained, with four skips as the best option (92%). In conclusion, our finding that fibrinoid necrosis rarely occurs in the absence of extracapillary proliferation may imply that both lesions are etiologically related. In addition, our observations indicate that the incidence of fibrinoid necrosis may be underestimated in clinical practice, depending on the number of sections evaluated.

The purpose of this article is to review a set of recently obtained data concerning matrix and matrix adhesion molecules in renal disease. Our goal is not to cover the entire topic, but rather to focus on findings obtained with an experimental model for chronic lupus nephritis, evoked in mice by inducing graft-versus-host disease (GVHD). The overall aim of these studies was to investigate the role of adhesion molecules as targets for autoantibodies, in the recruitment of inflammatory cells, and in the accumulation of matrix in kidney disorders. In addition, we set out to discover how matrix proteins in renal diseases differ from normal matrix molecules both quantitatively, in their increased frequency, and qualitatively, in their intramolecular structure. The advances in understanding and methodology described in this review imply a substantial capability for greater insight into the pathogenesis of kidney disease; for making better use of renal biopsies, such as in applying competitive reverse-transcriptase-polymerase chain reaction (RT-PCR) in RNA analysis for matrix; and in developing more effective treatment strategies for patients with kidney disease.

Accumulation of matrix proteins is a prominent feature of diabetic nephropathy. Glomerular visceral epithelial cells (GVECs) are important contributors to extracellular matrix (ECM) production in the glomerulus. Factors involved with increased accumulation of ECM proteins are high glucose, angiotensin II (ANG II), and transforming growth factor (TGF)-beta. Therefore, we investigated the effects of high glucose and ANG II on fibronectin and TGF-beta production by human GVECs in vitro. We found that ANG II had no effect on the production of fibronectin and TGF-beta by GVECs. Using reverse transcriptase-polymerase chain reaction analysis, no ANG II receptor could be detected on these cells. However, high glucose induced a twofold increase in fibronectin (P < 0.01) and a three- to sixfold increase in TGF-beta (P < 0.001) production. Similar results were obtained by analyzing the mRNA levels of fibronectin (increased 2.7-fold) and TGF-beta (increased 3.5-fold). Addition of increasing concentrations of rTGF-beta to control cells resulted in increased fibronectin production. Neutralizing antibodies against TGF-beta significantly reversed the increase in fibronectin protein and mRNA caused by high glucose back to control levels. We conclude that high glucose concentrations stimulate the synthesis of fibronectin and that this effect is mediated by induction of TGF-beta. These results suggest that in diabetic nephropathy, high glucose levels play a role in changing the matrix composition of the glomerular basement membrane through induction of TGF-beta. Our results indicate that a contribution to this process by an effect of ANG II on GVECs seems unlikely.

General agreement exists on the correlation of renal insufficiency with the severity of tubulointerstitial abnormalities in the biopsy. This could not be shown for the severity of glomerular pathology by semiquantitative methods. The relation between renal function and glomerular pathology was therefore evaluated in patients with various kidney diseases using quantitative measurements. Fifty-five patients and ten controls were studied. Histomorphometric measurements of renal biopsies were obtained for both frozen and paraffin sections and were correlated with serum creatinine values. The frozen sections were stained with an indirect immunoperoxidase technique using antibodies against collagen types I, IV, V, VI, laminin, fibronectin, decorin and heparansulphate proteoglycan core protein. The contribution of each component to the composition of the mesangial extracellular matrix was scored with a semiquantitative technique. All glomerular histomorphometric indices correlated with the severity of renal insufficiency expressed as serum creatinine at the time of biopsy. However, quantitative estimates of the glomerular deposition of periodic acid-Schiff positive extracellular matrix seemed to be the most important structural correlate of renal function (r = 0.524, p = 0.0001). Semiquantitative estimates of interstitial extracellular matrix accumulation were less correlated with renal function (r = 0.370, p = 0.01). The composition of the mesangial extracellular matrix did not differ in patients and controls, with the exception of the extent of laminin staining, which was significantly higher in patients. The extent of fibronectin staining in patients correlated with the severity of glomerular structural abnormalities. This study demonstrates that the severity of renal insufficiency in a variety of renal diseases correlates with the severity of glomerular pathology, when quantitative scoring is applied.

Proteoglycans metabolically labelled with [35S]sulphate and [3H]glucosamine or [3H]leucine were isolated from the incubation medium and cell layer of human adult mesangial cells and glomerular visceral epithelial cells using sequential DEAE chromatography purification steps followed by gel-filtration chromatography. The proteoglycan composition of each peak was analysed by treatment with HNO2, chondroitinase ABC or chondroitinase AC followed by chromatography on Sephadex G-50 columns. Heparan sulphate proteoglycan (HSPG) and dermatan sulphate proteoglycan were detected in both the culture medium and cell layer of mesangial cells. Culture medium of glomerular visceral epithelial cells contained HSPG and a second proteoglycan with the properties of a hybrid molecule containing HS and chondroitin sulphate (CS). The cell layer contained HSPG and CSPG. Detailed analysis of the hybrid molecule revealed that it had an apparent molecular mass of 400 kDa. SDS/PAGE of hybrid molecules, after treatment with heparitinase and chondroitinase ABC, revealed a core protein of 80 kDa. Using 1.8% polyacrylamide/0.6% agarose-gel electrophoresis, we deduced that the HS and CS were independently attached to one core protein. Because glomerular-basement-membrane HSPG is thought to be derived from mesangial cells and glomerular visceral epithelial cells and this molecule is involved in several kidney diseases, we investigated its synthesis in more detail. Anti-(rat glomerular-basement-membrane HSPG) monoclonal antibodies (JM403) and anti-(human glomerular-basement-membrane HSPG) polyclonal antibodies (both antibodies known to react with the large basement-membrane HSPG, perlecan) reacted strongly with HSPG obtained from both mesangial cells and glomerular visceral epithelial cells. However, the hybrid molecule did not react with these antibodies, suggesting that the HS side chain and the core protein were different from glomerular-basement-membrane HSPG. To quantify HS we performed an inhibition ELISA using mouse antibodies specific for glomerular-basement-membrane HS glycosaminoglycan side chains. Glomerular visceral epithelial cells produced significantly higher levels of HS (between 197.56 and 269.40 micrograms/72 h per 10(6) cells) than mesangial cells (between 29.8 and 45.5 micrograms/72 h per 10(6) cells) (three different cell lines; n = 3; P < 0.001). HS production by these cells was inhibited by cycloheximide, revealing that it was synthesized de novo. Expression of perlecan mRNA, demonstrated using reverse transcriptase PCR, was different in the two cell types. We conclude that glomerular visceral epithelial cells and mesangial cells have characteristic patterns of proteoglycan production. Glomerular visceral epithelial cells produced a hybrid proteoglycan containing CS and HS independently attached to its core protein.(ABSTRACT TRUNCATED AT 250 WORDS)

In a previous study, susceptibility for Trypanosoma brucei-related glomerulopathy in mice was shown to be dependent on non-major histocompatibility complex genes. Glomerular disease in this model could not be explained by the production of autoantibodies alone. In order to analyze which part of the defense system, in addition to the B-cell compartment, is involved in the development of this infection-related glomerular disease, groups of athymic (BALB/c rnu/rnu), splenectomized, or macrophage-depleted BALB/c mice were inoculated with T. brucei parasites. Polyclonal B-cell activation, invariably observed in infected BALB/c mice, was absent in BALB/c rnu/rnu mice. Glomerular disease in athymic mice, however, as defined by albuminuria and deposition of immune complexes, was not different from that seen in euthymic infected BALB/c mice. Splenectomy prior to inoculation of parasites led to a decreased incidence of albuminuria in 40% of the animals, whereas splenectomy 21 days after inoculation reduced albuminuria significantly, suggesting a role for spleen cells in the induction of glomerular disease. After macrophage depletion with liposome-encapsulated dichlorodimethylene-diphosphonate, infected BALB/c mice developed significantly higher albuminuria levels for a period up to 2 weeks after depletion. Therefore, it was concluded that the development of T. brucei-related glomerular disease is independent of thymus-matured T cells, while the involvement of macrophages in the development of proteinuria is inhibitory rather than disease inducing. Spleen cells other than thymus-dependent T cells, B cells, and macrophages should be investigated for their role in the pathogenesis of this glomerulopathy.

Diabetic nephropathy is characterized by albuminuria which proceeds to overt proteinuria. The highly negatively stained HS side chain of heparan sulphate proteoglycan (HSPG) is a major determinant of the charge-dependent permeability of the GBM. We set out to study the presence of HS and HSPG in the GBM of patients with diabetic nephropathy using newly developed monoclonal antibodies, and to compare HSPG expression to the expression of other previously investigated glomerular extracellular matrix compounds. Immunohistochemically, glomerular extracellular matrix components were analysed in 14 renal biopsies of patients with diabetic nephropathy and compared with those of normal control subjects. Monoclonal antibodies used were: JM403 against the HS side chain of GBM HSPG and JM72 against the HSPG-core protein. Also, a polyclonal antiserum (B31) against human GBM-HSPG-core protein was used. Additionally, antibodies were used against collagen types I, III, IV and against alpha 1 (IV)NC, alpha 3(IV)NC and fibronectin. Staining was scored for intensity and for staining pattern by four independent observers who had no previous knowledge of the sample origin. No glomerular staining was seen for collagen type I. Collagen type III was present in some diabetic nodules. Anti-collagen type IV showed a decreased GBM staining in patients with diabetic nephropathy (p = 0.04). With anti-alpha 1 (IV)NC no changes in GBM staining intensity were observed; with anti-alpha 3 (IV)NC brilliant GBM staining was seen in both groups. Increased mesangial staining (p = 0.003) was seen with anti-collagen type IV in biopsies with nodular lesions.(ABSTRACT TRUNCATED AT 250 WORDS)

Transforming growth factor-beta (TGF-beta) is a multifunctionally acting peptide with a broad spectrum of cellular targets. The biologic function of this peptide is currently widely studied. These studies reveal remarkable effects on proliferation, differentiation, migration, and production of many cell types. Among these effects, the regulatory function of TGF-beta in extracellular matrix homeostasis is thought to represent a major part of its action, mediating various other effects. Because disturbance of the regulation of extracellular matrix production is involved in various diseases, TGF-beta was expected to play a role in pathologic accumulation of extracellular matrix. By now, involvement of TGF-beta in the development of pathologic matrix production is most clear in experimental glomerular diseases. In this review, various lines of evidence for such an involvement are presented. After a general introduction about the molecular structure of TGF-beta and its biologic function, the regulation of extracellular matrix production by TGF-beta is discussed, with special emphasis on the glomerulus. Finally, the involvement of TGF-beta in the development of glomerulosclerosis is considered. Because in vivo inhibition of TGF-beta action in animal models seems to prevent pathologic matrix accumulation, the possible role of TGF-beta as an important mediator in the development of glomerulosclerosis may have important implications for prevention of glomerulosclerosis in human renal disease. Extensive study of the in vivo regulation of TGF-beta activity, both in animal models and human diseases, will be necessary for further elucidation of pathogenetic mechanisms and possibilities for therapeutic intervention.

Glomerulonephritis in BALB/c mice following infection with Trypanosoma brucei is characterized by albuminuria and glomerular deposition of immunoglobulins. Electron-dense deposits are present in the mesangium, as well as subendothelially and subepithelially along the glomerular capillary wall. In this study the nature of intracytoplasmic, electron-dense, round structures observed in glomerular endothelial cells was investigated by immunoelectron-microscopy and enzyme histochemistry. The presence of these structures was related in time with the development of proteinuria. Mice from the C57BL10 strain, which upon infection develop glomerular immune complexes without proteinuria, were examined as well. The results demonstrated that the first endothelial changes, occurring 3-4 weeks after infection, were swelling of endothelial cells containing intracytoplasmic, electron-dense, round structures. These changes were seen prior to the onset of proteinuria, and were not present in glomeruli of mice that did not develop proteinuria. The endothelial granules were shown to contain immunoglobulins and typical lysosomal enzymes, providing evidence for phagocytosis by the glomerular endothelial cells. Liver endothelial cells did not show comparable changes. Thus, local phagocytosis by glomerular endothelial cells is shown to be a specific event in the development of glomerular disease.

Extraction of RNA has been described for rat and rabbit glomeruli but not for mouse glomeruli. Due to their small size, mouse glomeruli cannot be isolated by relatively simple sieving techniques. Based on recently reported methods for the isolation of mouse glomeruli, we developed an RNA isolation technique by performing comparative methodological studies. Two standard RNA extraction methods were compared. In addition in separate experiments the influence was studied of protease inhibitors and freezing and thawing of whole kidney prior to glomeruli isolation, on the yield and degradation of RNA. Therefore kidneys were perfused with 10 ml 0.01 M PBS containing 1.25% Fe3O4 through the aorta. Kidneys were decapsulated and passed through a 75-microns metal screen. After pelletting and washing, tubes were placed against a magnet and pelleted glomeruli were washed three times. In a second experiment protease inhibitors were added to the PBS. As a third method, kidneys were frozen before the isolation of glomeruli. From isolated glomeruli RNA was extracted using either caesium chloride or lithium chloride method. The yields of RNA (OD 260) were highest using the lithium chloride method. Hybridization of Northern blots of extracted RNA with cDNA probes showed the best results when RNA was extracted using the lithium chloride method, while the caesium chloride method led to considerable degradation of RNA. Freezing of kidney tissue prior to RNA extraction led to the virtual absence of any signal. We then applied this method successfully in an in-vivo model of experimental lupus nephritis. This is the first description of an optimal protocol for the extraction of RNA from mouse glomeruli.(ABSTRACT TRUNCATED AT 250 WORDS)

Passive Heymann nephritis (PHN) in the rat is induced by the administration of heterologous antibodies to renal tubular epithelium (RTE). The nephritogenic capacity of anti-RTE resides primarily in the antibody fraction directed against a 330-kD glycoprotein (gp330). However, monospecific anti-gp330 antibodies are less nephritogenic than anti-RTE antibodies. This discrepancy led us to study a possible synergizing role for different antibody specificities present within anti-RTE. The enzyme dipeptidyl peptidase type IV (DPP IV) is, like gp330, present in RTE as well as in the glomerulus. Anti-DPP IV antibodies have been shown to induce an acute, transient proteinuria. In this study, we investigated the nephritogenic effect of separate or simultaneous administration of heterologous anti-DPP IV and anti-gp330 antibodies. Injection of anti-DPP IV antibodies resulted in a short-lived glomerular binding, and in a dose-dependent polyuria, albuminuria or proteinuria. Binding of anti-gp330 antibodies was slower, but much more stable. Albuminuria could only be observed in the heterologous phase. Combined injection did not alter antibody-binding kinetics of either antibody nor the albuminuria induced by anti-gp330. However, in the presence of anti-gp330 antibodies, a lower dose of anti-DPP IV was capable to induce transient albuminuria as compared to anti-DPP IV alone. In conclusion, anti-DPP IV and anti-gp330 antibodies bind to different sites in the glomerulus with different kinetics. The presence of anti-gp330 deposits facilitated anti-DPP-IV-induced renal injury.

Background: In murine chronic graft-versus-host disease, an experimental model for lupus nephritis, autoantibodies against renal tubular epithelium can be found. Part of these antibodies are directed against a constituent of renal tubular epithelium, the enzyme Dipeptidyl peptidase IV (DPP IV). DPP IV is present on the cell membrane of glomerular epithelial and endothelial cells, and plays an important role in cell-extracellular matrix interactions. In mice and rats, administration of heterologous anti-DPP IV antibodies can induce proteinuria and podocyte effacement., Experimental Design: In this study, the glomerular DPP IV enzyme activity was investigated in the course of graft-versus-host disease in (C57BL/10 x DBA/2) F1 hybrids both by light- and electron microscopy using a DPP IV-specific substratum., Results: Light microscopical examination revealed an overall reduction of DPP IV activity and an altered distribution pattern in glomeruli as early as 4 weeks after induction of graft-versus-host disease. Immunofluorescence studies using anti-DPP IV antibodies showed actual redistribution, excluding antibody-mediated enzyme inactivation. Enzyme electron microscopy revealed an irregular deposition of reaction product characterized by a patchy, "moth eaten" appearance of the endothelial and epithelial membranes. This process occurred simultaneously with the development of albuminuria and preceded the effacement of epithelial foot processes. In control mice, DPP IV showed a continuous distribution along endothelial and podocyte membranes., Conclusions: In view of these findings we postulate that impairment of the function of DPP IV as a non integrin adhesion molecule may be one of the causative factors underlying the structural and functional lesions observed in this model for lupus nephritis.

Chronic serum sickness was induced in four groups of Wistar rats by immunization with BSA, cationized BSA (cBSA), human IgG (HuIgG), or human IgM (HuIgM), followed by repeated intraperitoneal (i.p.) injection of the antigen used, to study the effect of the characteristics of an antigen on renal immunopathology. Renal tissue sampled 2, 4, 7, and 9 weeks after the start of the i.p. injections was examined by light-, immunofluorescence-, and electron-microscopy. Proteinuria was measured in urine collected over 24 h. All animals given BSA, cBSA, or HuIgG developed progressive renal disease characterized by initial deposition of antigen and rat Ig in the mesangium of rats given BSA or HuIgG, and minimal amounts in those given cBSA, followed by the appearance in the first instance of subendothelial deposits in the animals receiving BSA or HuIgG, and later subepithelial deposits in those given BSA, cBSA, or HuIgG. The appearance of immunoglobulin deposits along the glomerular capillary wall was associated with the onset of massive proteinuria reaching average levels of 450 mg/24 h for rats given BSA or cBSA, and 500 mg/24 h for those given HuIgG. Animals injected with HuIgM showed only mesangial deposits of human IgM and rat Ig without the development of proteinuria. Under light-microscopy, rats given BSA, cBSA, or HuIgM showed minimal abnormalities, whereas those receiving HuIgG showed transient but severe influx of granulocytes in glomeruli with the development of diffuse proliferative glomerulonephritis in association with a long-lasting phase characterized by subendothelially localized immune aggregates.(ABSTRACT TRUNCATED AT 250 WORDS)

Glomerular basement membrane (GBM) antigens used for immunological studies have until now been isolated from human and rat glomeruli but not from mice. However, given the growing awareness of extracellular matrix species-specificity, the need for a purification method of mouse GBM exists. We now report the purification of GBM from isolated mouse glomeruli. 10 ml 0.01 M phosphate buffered saline (PBS) containing 1.25% Fe3O4 was perfused through the aortae of (C57BL10 x DBA/2)F1 mice over 1 min, kidneys were decapsulated and passed through a 0.075 mm mesh metal screen and collected. After pelletting and washing, the tube was placed against one pole of a magnet and the pelleted glomeruli were washed. This procedure was repeated three times. The glomerular suspension was examined by light microscopy, sonicated, and lyophilized. In suspension no free fragments of tubuli or Bowman's capsule were present as observed by light microscopy. Mouse GBM was prepared from the isolated glomeruli using a modification of earlier described methods for human and rat GBM by enzymatic digestion. Elisa studies showed the presence of laminin, type IV collagen, and fibronectin. Monoclonal antibody directed against tubular epithelial antigen gp160 did not react to either mouse or rat GBM. Sera and glomerular eluates from (C57BL10 x DBA/2)F1 mice suffering from chronic graft-versus-host autoimmune disease gave a higher signal against mouse GBM than against rat GBM. Normal rabbit serum, normal mouse serum and normal mouse eluate did not show significant activity against mouse GBM. Immunoblotting showed the presence of both laminin B1 and B2 chains as well as type IV collagen alpha 1 and alpha 2 subunits. In summary, we describe a method by which it is possible to extract mouse GBM with a high purity.

The pathogenesis of renal involvement was studied in murine chronic graft-versus-host disease (GVHD), which is a model for human systemic lupus erythematosus. GVHD was induced by four i.v. injections of lymphocytes from DBA/2 donor mice into (C57BL/10 x DBA/2)F1 hybrids at 3-4-day intervals. Two weeks after the first injection, antibodies were found to have been deposited in the mesangium and along the glomerular basement membrane (GBM) in a linear arrangement, which changed to a granular pattern after 6-8 weeks. In this stage, large electron-dense complexes were present both subepithelially and subendothelially along the GBM. Proteinuria increased up to 11,300 +/- 2140 micrograms/18 h. Indirect immunofluorescence studies and ELISA showed that sera and kidney eluates contained autoantibodies directed against nuclear antigens and GBM component laminin as well as against renal tubular epithelial antigens (RTE). The specificity of the anti-RTE antibodies was further characterized by the use of absorption techniques as well as immunoblotting. The early linear immunofluorescence pattern seems to be associated with glomerular binding of anti-GBM antibodies, while electron-dense complex formation in later stages may be induced by the superimposed deposition of anti-RTE antibodies. Similar phenomena were recently described in Heymann's nephritis in the rat, a model for human membranous nephropathy.