Glomerular immune deposits are associated with increased proteinuria in patients with ANCA-associated crescentic nephritis

Transcription

1 Nephrol Dial Transplant (2003) 18: Original Article Glomerular immune deposits are associated with increased proteinuria in patients with ANCA-associated crescentic nephritis Irmgard Neumann 1, Heinz Regele 2, Renate Kain 2, Rainer Birck 1 and Franz Thomas Meisl 1 1 Department of Nephrology, Wilhelminenspital, Vienna and 2 Department of Clinical Pathology, University Vienna, Vienna, Austria Abstract Background. In small vessel vasculitis and its renallimited form, idiopathic crescentic glomerulonephritis, renal damage is characterized by pauci-immune necrotizing crescentic glomerulonephritis (CGN) without histological evidence of immunoglobulin (Ig) deposition. In some patients, however, significant amounts of immune deposits may be detected. Therefore, we evaluated the clinical significance of these immune deposits in anti-neutrophil cytoplasmic autoantibody (ANCA)-associated pauci-immune CGN. Methods. Renal biopsies of 45 consecutive patients with new onset of Wegener s granulomatosis, microscopic polyangiitis and idiopathic CGN were retrospectively evaluated by light microscopy, immunohistochemistry and electron microscopy and the findings compared with renal function and outcome. Results. Typical pauci-immune CGN was found in 37 patients (group I). In eight patients (18%; group II), however, histopathological examination revealed substantial deposition of Ig in the mesangium anduor along the glomerular basement membrane. Five of these eight patients were canca positive; two initially had panca and developed a canca pattern and one was panca positive. There were no differences between groups in age, gender, renal function or extrarenal organ involvement at the time of biopsy. However, patients in group II had significantly more proteinuria (5.4"3.1 vs 1.3"1.0 gu24 h; Ps0.016). We also observed a trend for a worse outcome with respect to renal function and mortality in group II patients; however, the differences did not reach significance. Conclusions. Our results confirm that in ANCAassociated CGN a substantial percentage of patients have evidence of Ig deposition in renal biopsies. In this subgroup, Ig deposition was associated with a significantly greater degree of proteinuria. Further Correspondence and offprint requests to: Dr Irmgard Neumann, Wilhelminenspital, Department of Nephrology, Montleartstr. 37, A-1160 Vienna, Austria. irmi_neumann@hotmail.com investigations are necessary to define the full clinical impact of immune-complex deposition on the clinical course of renal disease in pauci-immune CGN. Keywords: immune deposits; necrotizing crescentic glomerulonephritis; pauci-immune; renal histology Introduction In the early 1970s, first studies attempted to define the pathogenesis of idiopathic rapid progressive glomerulonephritis using immunofluorescent and electron microscopic techniques. Subsequently, the occurrence of immunopathologic findings that are not characteristic of either immune complex (IC) or anti-glomerular basement membrane (GBM) glomerular disease has been reported and a subset of patients with crescentic glomerulonephritis (CGN) without immune deposits was described [1]. In small vessel vasculitis the characteristic glomerular lesion is a focal necrotizing glomerulonephritis associated with crescents and little or no glomerular staining for immunoglobulins (Ig) in immunohistochemistry [2]. However, there are single reports on positive Ig findings in these patients. Thus, renal biopsy studies of patients with Wegener s granulomatosis resulted in variable results. Immune reactants were identified in some, but not all, renal biopsies of patients with Wegener s granulomatosis [3,4] and circulating ICs were reported in active disease [5]. However, even before anti-neutrophil cytoplasmic autoantibodies (ANCAs) were detected, absence of Ig in necrotizing CGN was strongly suggestive of either angiitis-associated CGN or idiopathic CGN [1]. In 1985, cancas and pancas were reported to be sensitive and specific markers for Wegener s granulomatosis [6], respectively, for microscopic polyangiitis and the renal limited variant, idiopathic CGN [7,8]. In contrast to CGN caused by IC deposition and characterized by granular immunostaining for Ig, necrotizing CGN with no evidence for Ig has been # 2003 European Renal Association European Dialysis and Transplant Association

2 Immune deposition in ANCA-associated crescentic glomerulonephritis shown to be associated with ANCA in about 80% of cases [9]. This observation led to the concept of pauci-immune glomerulonephritis, characterized by the presence of ANCA and the absence of Ig deposition. However, the clinical impact of the deposited ICs in affected kidneys remains to be elucidated. Therefore, we decided to retrospectively analyse patients with Wegener s granulomatosis, microscopic polyangiitis and idiopathic CGN, focusing on renal immunohistology as well as renal functional parameters and outcome. Subjects and methods Patients and study protocol Forty-five consecutive patients with new onset of systemic small vessel vasculitis and CGN or idiopathic CGN were included irrespective of ANCA serology. Patients with anti- GBM disease and IC-mediated diseases, such as systemic lupus erythematosus, post-infectious glomerulonephritis, Henoch Schönlein purpura and membranous and membranoproliferative glomerulonephritis, were excluded. All patients were referred to our renal unit between 1989 and 1999 because of suspected systemic vasculitis, pulmonary renal syndromes or acute renal failure. None of the patients was treated prior to renal biopsy. All 45 renal biopsies were reviewed by light microscopy, immunohistochemistry anduor electron microscopy by two pathologists who were blinded to the clinical history. We collected clinical data regarding renal function and the degree of proteinuria at the time of renal biopsy as well as of organ involvement at presentation and ANCA titre. Twenty-four hour urine protein results were available in 34 patients. Six patients were anuric and no accurate quantification of proteinuria was available in five. Onset of disease was determined by the date of occurrence of the first symptoms attributable to vasculitis and the time period between disease onset and renal biopsy was calculated. All patients were treated with oral cyclophosphamide and prednisolone for remission induction therapy. Renal biopsies Fixation and processing of kidney biopsies. Kidney biopsy specimens were fixed in 4.5% buffered formaldehyde for light microscopy and immunohistochemistry. If available, part of the biopsy was fixed in a 4% buffered paraformaldehyde solution containing 0.1% glutaraldehyde for electron microscopy. Consecutive serial 4 mm sections were used for histological and immunohistochemical staining. Stains employed included haematoxylin and eosin (H&E), periodic acid Schiff (PAS), silver methenamine (Meth) and acid fuchsinu orangeg (AFOG). For immunohistochemistry, sections were deparaffinized in xylene and rehydrated with a graded ethanol series. Sections were incubated with primary antibodies to human Ig (IgG, IgM and IgA), complement (C3 and C1q) and fibrinogen (AF), followed by appropriately biotinylated secondary antibodies and detection by a streptavidin peroxidase complex, according to the manufacturer s instructions (Vectastain ABC kit; Vector Laboratories, Burlingame, CA, USA). Peroxidase reaction product was visualized by diamino benzioline (DAB) and H 2 O 2. Electron microscopy was performed according to standard procedures. After embedding in epon, ultrathin sections were mounted on metal grids and stained with uranyl acetate prior to viewing in a transmission electron microscope. Examination of specimens. Pathological changes of glomeruli, tubules, interstitium and blood vessels were described in light microscopy. The percentages of crescents and sclerotic glomeruli were calculated as the percentage of total number of glomeruli in a biopsy. Interstitial fibrosis was semi-quantitatively scored (quqquqqq). Positive immunohistochemical staining was recorded by localization within the glomerulus (in the mesangium or along GBM) and staining pattern (linear or granular, segmental or diffuse). Intensity of staining was scored as negative (0), mild (1q), moderate (2q) or strong(3q). Pauci-immunity was defined as negative staining or mild positivity for either IgM anduor C1q within the mesangium alone, since such deposits alone do not necessarily indicate fully developed ICs and are quite commonly encountered in sclerosing glomeruli in non-icmediated conditions like diabetic nephropathy [10]. Immunecomplex deposition was defined as the presence of Ig types or complement components other than or in addition to mesangial IgM or C1q. Deposition of ICs was confirmed by the presence of electron-dense deposits in at least two mesangium fields anduor along GBM (either subendothelial, intramembranous or subepithelial) of each glomerulus examined. Autoimmune serology Indirect immunofluorescence (IIF). Blood was drawn into standard serum tubes with separating gel (serum vacuette, 8 ml; Grainer GmbH, Kremsmünster, Austria) and centrifuged within 2 h after collection. On every serum sample, IIF was performed using Hep2 cells as a substrate (Hep-2000; Immuno Concepts Inc., Sacramento, CA, USA), controlling the actual assay performance daily using a titrable IIA control (Immuno Concepts Inc.). In addition, kidney stomach liver slides and ANCA slides (INOVA Diagnostics Inc., San Diego, CA, USA) were used. The slides were routinely fixed with ethanol and formalin, respectively. Once a positive result was obtained by IIF using Hep2 cells (cut-off titre: )1:160), enzyme-linked immunosorbent assay (ELISA)- based assays (detecting antibodies of the IgG subclass) against double-stranded DNA, histones and extractable nuclear antigens were used (INOVA Diagnostics Inc.) to characterize the specificity of the antibodies present further. Enzyme-linked immunosorbent assay (Wieslab, Lund, Sweden) was performed for antibodies to proteinase-3 (PR3) and myeloperoxidase (MPO). Statistical analysis Data are shown as means"sd. Student s t-test or Fisher s exact test were used as appropriate to compare differences between means. Spearman s coefficients were used to test correlation. Results 525 Forty-five patients with new onset of systemic small vessel vasculitis and CGN (ns35) or idiopathic CGN (ns10) underwent renal biopsy. There were 23 females and 22 males (median age: 56.6"15.8 years). At presentation, 24 patients were canca positive, 17

3 526 I. Neumann et al. were panca positive and three were ANCA negative. Determination of ANCA was not available in one patient who was included because of typical Wegener s granulomatosis. Histological analysis confirmed the original diagnosis of CGN in all 45 specimens. The histopathological review demonstrated typical pauci-immune CGN in the majority of patients (ns37) (group I). In eight of 45 (18%) cases (group II), however, there was substantially increased mesangial anduor granular glomerular staining for Ig. The spectrum of organ involvement at presentation was similar in groups I and II (Table 1). Furthermore, there were no differences in age, gender or renal function at the time of biopsy. There was a difference in the diagnostic lag between the first clinical symptoms of vasculitis and renal biopsy for groups I and II which, however, did not achieve statistical significance (23.0"13.5 vs 13.8"14.8 weeks; Ps0.36) (Table 2). None of the Table 1. Organ involvement at presentation in patients with typical pauci-immune NCGN (Group I) and NCGN with immune deposits (Group II) Organ involvement Group I No. of patients (%) Renal 37 (100) 8 (100) Pulmonary 26 (68) 4 (50) Ear, nose 18 (47) 4 (50) Skin 3 (9) 2 (25) Nervous system 8 (21) 0 (0) Joints 24 (63) 5 (63) NCGN, necrotizing crescentic glomerulonephritis. Group II No. of patients (%) Table 2. Clinical data in both groups at the time of presentation Group I (ns37) Group II (ns8) P-value Age (years) 57.5" "21.7 NS S Cr (mgudl) 5.9" "4.7 NS U protein (gu24 h) 1.3" " Delay a (weeks) 13.8" "13.5 NS a Time period between first symptom of vasculitis and renal biopsy. S Cr, serum creatinine; U protein, urinary protein. patients had clinical evidence for another autoimmune disease, such as autoimmune thyreoditis, inflammatory bowel disease, sclerosing cholangitis or for malignancy. Details of the clinical presentation of the eight patients in group II are described in Table 3. Serological testing for ANCA in group I revealed canca in 19 (17 PR3q, two not done) and panca in 14 (12 MPOq, two not done) while three were ANCA negative (both IIF and ELISA) and one patient presented with typical pauci-immune CGN before ANCA testing was available. Of the eight patients who had evidence of positive glomerular immunofluorescence for Ig, five were canca positive and three had a panca pattern (Table 3). Two of the latter patients changed their ANCA pattern later on and became canca positive. One of them (patient 1 in Tables 3 and 4) was tested for antigen specificity revealing positivity in ELISA both for MPO and PR3. An immunoblot to confirm the presence of antibodies against both PR3 and MPO was not performed. Anti- GBM antibodies and cryoglobulins were negative in all cases. Decreased complement levels were found only in patient 3 of group II. Renal biopsy findings of the eight cases (group II) with a substantially increased granular staining for Ig anduor complement in the mesangium anduor capillary loops are listed in Table 4. A P2q staining of Ig or complement was found in the mesangium of all eight patients, IgA being the most frequent Ig in five patients. All except two patients had also positive Ig staining in capillary loops, three of them also in crescents. Representative examples of immunohistochemical findings are given in Figure 1. Compared with patients in group I, those in group II demonstrated a significantly higher 24 h proteinuria (5.4"3.1 vs 1.3"1.0 g; Ps0.016). Moreover, in only one of the patients in group I with classical pauciimmune CGN was proteinuria of )3 gu24 h observed. This was the only patient who received no immunosuppressive therapy because of pronounced chronic lesions in his renal biopsy. No correlation was found between proteinuria, age and renal function. There was also no difference between both groups with respect to the amount of crescent formation, fibrinoid necrosis, glomerulosclerosis or interstitial fibrosis (data not Table 3. Clinical features in patients with positive immunohistochemistry (group II) Patient (no.) Age (years) Gender S Cr (mgudl) U protein (gu24 h) ANAudsDNA ANCA (IIFuELISA) HCV (SerouPCR) Extra-renal manifestations 1 73 M :320udsDNA neg pancaqumpoq Negund Lung, ENT 2 80 M 1.8 Nd Neg cancaqupr3q Negund Lung, ENT, eye, skin 3 47 F :160udsDNA neg cancaqupr3q Posuneg Lung, ENT, joints 4 25 F a Neg cancaqupr3q Negupos None 5 20 F Neg cancaqupr3q Negund Lung, ENT, joints 6 42 M Neg cancaqupr3q Negund ENT, joints 7 59 M Neg pancaqund Negund Skin 8 78 F 4.0 Pos b Neg pancaqumpoq Negund Joints a Nephrotic patient, serum albumin 1.8 gudl. b Positive, no exact quantification available. S Cr, serum creatinine; U protein, urinary protein; neg, negative; pos, positive; nd, not done.

4 Table 4. Renal pathohistological findings in patients of group II Patient (no.) Immunohistochemistry EM deposits Light microscopy Mesangium Capillary loops Crescents Mesangium Capillary loops Crescents (%) Fibrinoid necrosis (%) Scars (%) Tubulointerstitial fibrosis (0 3q) 1 IgA IgM, IgA IgM, C1q Nd q 9 2 IgA Nd q 12 3 IgM, C3 IgM, C3, C1q Nd q 12 C1q 4 C1q IgG (1q), Pos Subendothelial q 9 IgM, C1q 5 IgA, C1q Pos Subendothelial q 9 and subepithelial 6 IgM, C1q, Pos Subendothelial q 10 IgA (1q) and subepithelial 7 IgA, C1q IgM Nd q 9 8 IgA, C1q IgA, IgM IgM, C3 Neg Subendothelial q 23 The table gives any Ig or complement staining P2q unless indicated otherwise. EM, electron microscopy; nd, not done; pos, positive; neg, negative. Glomeruli (total no.) Immune deposition in ANCA-associated crescentic glomerulonephritis 527 Fig. 1. Immunohistochemistry. (A) Image showing only scanty deposition of IgG along the capillary wall as a representative finding for a mild staining (1q) in a patient of group I. In contrast, (B) and (C) show significant mesangial IgA (patient 2) and C3 deposition (patient 3) in group II, respectively.

5 528 I. Neumann et al. shown). Electron microscopy was available in 19 patients of group I and in four of group II, confirming in the latter patients the immunohistological evidence of immune deposition by demonstrating the presence of electron-dense deposits (Figure 2). Focal podocyte retraction with partial loss of foot processes was also Fig. 2. Electron microscopy. Glomerular capillary loop with subendothelial (arrows) and subepithelial (triangles) electron-dense immune deposits. Transmission electron microscopy, found in three patients of group I and in none of group II. Basically, the podocytes were not severely damaged and only minor podocyte changes were detected in nine patients of group I and two patients of group II. Analysis of renal function showed that in group II renal function improved in only one patient, who remained in long-term remission (patient 5), while in the remaining patients (patients 2 and 6) no improvement but stabilization of renal function with serumcreatinine levels of 1.8 and 3.5 mgudl, respectively, was achieved. In contrast, in group I, in all of the 24 patients with renal remission, the disease course was characterized by an improvement of renal function with a mean serum-creatinine of 1.4"0.3 mgudl 2 years after presentation. The latter group also included the five patients who initially required haemodialysis and regained renal function (Figure 3). However, this difference in outcome between the two subgroups did not reach statistical significance. Renal outcome was also evaluated for the whole cohort at 2 years follow-up. Among patients in group II, four of seven required dialysis for end-stage renal disease (ESRD) (patients 1, 3, 4 and 7). Two of them were dialysis-dependent at diagnosis and remained on dialysis and two patients progressed to ESRD 22 and 59 weeks, respectively, after diagnosis (patients 7 and 3). No renal follow-up was available in one patient who died from sudden death after 5 weeks follow-up. In contrast, among the 37 patients in group I, 16 presented with dialysis-dependent renal failure. Five of Fig. 3. Follow-up of renal function in patients with typical pauci-immune CGN (group I) and CGN with immune deposits (group II). Data are expressed as means"sd.

6 Immune deposition in ANCA-associated crescentic glomerulonephritis them regained renal function and seven remained on haemodialysis. One patient experienced a renal relapse within the first year and became dialysis-dependent. Due to death occurring within the first 6 weeks of follow-up, five patients from group I and one from group II were excluded from evaluation of long-term renal outcome. Thus, within the first 2 years, 8u32 patients in group I (25%) required haemodialysis for ESRD compared with 4u7 (57%) in group II (Ps0.171; NS). Follow-up of mortality showed that two deaths occurred in group II and five patients died in group I. Discussion The observation that ANCAs are strongly associated with CGN without immune deposits but not with ICmediated or anti-gbm antibody-mediated glomerulonephritis [2] led to the definition of ANCA-associated CGN as a pauci-immune CGN. Although traces of Ig have been described in these patients [2,9], the term has found worldwide acceptance. While in different forms of crescentic GN the frequency of ANCA 529 has been shown to be inversely proportional to the amount of Ig in the glomeruli, a low number of ANCA-positive patients with GN may have welldefined glomerular Ig-deposits [9]. Our observations document the histological evidence of both granular and mesangial Ig-deposits without the features of any other primary IC-mediated GN in patients with ANCA-associated CGN and small vessel vasculitis. This is consistent with single reports in the literature indicating that not all cases of CGN and positive serology for ANCA, in fact, have negative immunohistological findings (Table 5). In this study we found that immune deposition can be detected in almost 20% of consecutive patients, representing a considerable subset arguing against pauci-immune CGN as the only pathogenetically relevant feature in patients with ANCA-associated CGN. Our data are also in agreement with earlier reports, describing immune deposits in renal biopsies of some, but not all, patients with Wegener s granulomatosis [2 4]. Immunoglobulin A may play a special role in this context. There is evidence that IgA may be found in some renal biopsies of patients with Wegener s granulomatosis (Table 5) and development of de novo Table 5. Reports in the literature of cases with glomerulonephritis with deposition of immunoglobulins and ANCA-associated vasculitisuwegener s granulomatosis (WG) Authoruyear Number of cases ANCA IIFuELISA Renal histology IHuEMulocation Diseaseuorgan manifestation Norton [a] Nd EM deposits WG, NCGN subendothelial Castleman [b] Nd Gammaglobulinu WG, NCGN granular, GBM Aldo [c] Nd EM deposits WG, NCGN subendothelial Horn [4] Nd EM deposits WG, NCGN subepithelial Wolff [d] Nd EM deposits WG subepithelial Balow [e] Nd IgG, C3u WG focal-segmental Ronco [3] Nd IgMucapillary wall, WG IgAumesangial Andrassy [11] cancaupr3q IgA, C3c, WG C3dumesangial, peripheral Komatsuda [f] pancaumpoq IgGumesangial, NCGN capillary wall Allmaras [13] pancaumpoq IgA, C3c, C3u NCGN mesangial, capillary wall McGregor [g] a pancaumpoq IgA, NCGN, P C3umesangial Ramirez [h] a pancaumpoq IgA mesangial NCGN, P, ENT Vrtovsnik [i] PR3q IgAumesangial, WG peripheral Rollino [j] cancaupr3q IgA, C3u WG, NCGN mesangial, GBM Richer [k] pancaumpoq IgAumesangial SV Haas [14] cancaq, 5 PR3q, IgA (IgM, C3)u NCGN, SV 1 pancaq, 2MPOq b, 2 nd mesangial Aasarod [l] cancaupr3q IgA, IgM (IgG, WG C3)umesangial a Child. b One patient with double specificity for MPO and PR3. References [a] [l] can be provided on demand by the first named author. IH, immunohistology; EM, electron microscopy; nd, not done; P, pulmonary manifestation; SV, systemic vasculitis; WG, Wegener s granulomatosis.

7 530 I. Neumann et al. IgA nephropathy during remission of Wegener s granulomatosis has been reported [11]. Although IgA nephropathy is not typically associated with MPOor PR3-specific ANCA [12], the coexistence of IgA nephropathy with anti-mpo antibodies has been described [13] and, recently, mesangial IgA-deposits have been reported in six patients with ANCAassociated CGN [14]. In five of the patients documented here, significant IgA deposition in the mesangium was also detected in the absence of relevant mesangial hypercellularity. Four of these patients, however, additionally had substantial deposition of other Ig anduor complement components in the capillary loop. Interestingly, patient 2, whose biopsy showed a positive staining only for IgA in the mesangium, was also the only one in group II without significant proteinuria, suggesting that the clinical relevance of IC deposition may depend on the type and localization of Ig in the glomerulus. The pathogenetic relevance of glomerular Ig deposition in ANCA-associated CGN still remains unclear. Noteworthy, experimental animal models could not convincingly confirm the concept of pauci-immunity so far. Thus, in rats pre-immunized with human MPO, subsequent kidney perfusion with lysosomal extracts and H 2 O 2 induced CGN, but IgG and C3 were found early in the course of the disease. Moreover, the degree of histologic injury was proportional to the amount of IgG immune deposits. Later in the disease course, however, the immune deposits could not be detected any more [15,16]. Thus, it should be taken into consideration that the presence of IC may be a timedependent phenomenon also in the human situation. Since renal biopsy is a static record of a dynamic process of CGN, it might be possible that Ig deposits were present at an earlier time and had been decreased by phagocytosis and digestion by infiltrating neutrophils. Such mechanisms are known from animal models of vasculitis of the Arthus-reaction type where neutrophils degrade immune deposits within 48 h after deposition [17]. Although the pathological relevance of glomerular immune deposits is evident in many IC-mediated renal diseases [18], their potential influence in pauci-immune nephritis has not been investigated so far. In our study, the patients in group II exhibited significantly higher proteinuria in comparison to the classic pauci-immune group despite comparable glomerular and tubulointerstitial changes in standard light microscopy. Basically, it is known that pauci-immune CGN shows more fibrinoid necrosis, more disruption of Bowman s capsule and less proteinuria when compared with typical immune complex nephritis [19]. In patients with typical IC-GN, such as lupus nephritis, post-streptococcal GN as well as in anti-gbm nephritis, the presence of ANCA has been shown to correlate with more glomerular necrosis and crescent formation, postulating a synergistic effect between ANCA and IC [9]. Furthermore, a more aggressive clinical course of membranous nephropathy has been reported in patients with accompanying vasculitic glomerulonephritis in association with the presence of ANCA [20]. In rat models of CGN, ANCAs also aggravated the nephritogenic response to otherwise subnephritogenic doses of antibodies [21]. Thus, even if the presence of ANCAs per se does not induce glomerular Ig deposition, they might aggravate IC induced pathology or vice versa. In conclusion, classification of ANCA-associated CGN still remains difficult because of overlapping clinical and histopathological features and the largely unknown aetiology and pathogenesis. In some animal models of CGN, cell-mediated immunity with activated T-cell seems to play a major role in immune injury. Also, in patients with ANCA-associated disease, T-cell activation and T-cells reactive with ANCAantigens have been demonstrated [22,23], making T-cell auto-reactivity a possible third pathogenetic pathway besides ANCAs and ICs. Whatever the pathogenetic mechanism, the relatively high frequency of immune deposits in patients having ANCAassociated CGN and its potential influence on renal function and outcome warrant larger studies to elucidate the impact of IC deposition on renal disease in pauci-immune CGN. Acknowledgements. The authors would like to thank Prof. Dr F.J. van der Woude (Fifth Department of Medicine, University Hospital Mannheim, University of Heidelberg, Germany) for critical and helpful remarks concerning the manuscript. References 1. Stilmant MM, Bolton WK, Sturgill BC, Schmitt GW, Couser WG. Crescentic glomerulonephritis without immune deposits: clinicopathologic features. Kidney Int 1979; 15: Jennette JC, Wilkman AS, Falk RJ. Anti-neutrophil cytoplasmic autoantibody-associated glomerulonephritis and vasculitis. Am J Path 1989; 135: Ronco P, Verroust P, Mignon F et al. Immunopathological studies of polyarteritis nodosa and Wegener s granulomatosis: a report of 43 patients with 51 biopsies. Q J Med 1983; 206: Horn RG, Fauci AS, Rosenthal AS, Wolff SM. Renal biopsy pathology in Wegener s granulomatosis. Am J Pathol 1974; 74: Howell SB, Epstein WV. Circulating immunoglobulin complexes in Wegener s granulomatosis. Am J Med 1976; 60: van der Woude FJ, Rasmussen N, Lobatto S et al. Autoantibodies against neutrophils and monocytes: tool for diagnosis and marker of disease activity in Wegener s granulomatosis. Lancet 1985; 1: Falk RJ, Jennette JC. Anti-neutrophil cytoplasmic autoantibodies with specificity formyeloperoxidase in patients with systemic vasculitis and idiopathic necrotizing and crescentic glomerulonephritis. N Engl J Med 1988; 318: Savage COS, Winearls CG, Jones SJ, Marshall PD, Lockwood CM. Prospective study of radioimmunoassay for antibodies against neutrophil cytoplasm in diagnosis of systemic vasculitis. Lancet 1987; 8: Falk RJ, Jennette JC. ANCA small-vessel vasculitis. J Am Soc Nephrol 1997; 8: Ainsworth SK, Hirsch HZ, Brackett NC, Jr et al. Diabetic glomerulonephropathy: histopathologic, immunofluorescent, and ultrastructural studies of 16 cases. Hum Pathol 1982; 13:

8 Immune deposition in ANCA-associated crescentic glomerulonephritis 11. Andrassy K, Waldherr R, Erb A, Ritz E. De novo glomerulonephritis in patients during remission from Wegener s granulomatosis. Clin Nephrol 1992; 38: Lim LCL, Taylor JG, Schmitz JL et al. Diagnostic usefulness of antineutrophil cytoplasmic autoantibody serology. Am J Clin Pathol 1999; 111: Allmaras E, Nowack R, Andrassy K, Waldherr R, van der Woude F, Ritz E. Rapidly progressive IgA nephropathy with anti-myeloperoxidase antibodies benefits from immunosuppression. Clin Nephrol 1997; 48: Haas M, Jafri J, Bartosh SM, Karp S, Adler SG, Meehan SM. ANCA-associated crescentic glomerulonephritis with mesangial IgA deposits. Am J Kidney Dis 2000; 36: Brouwer E, Huitema MG, Klok PA et al. Anti-myeloperoxidaseassociated proliferative glomerulonephritis: an animal model. J Exp Med 1993; 177: Yang JJ, Jennette JC, Falk RJ. Immune complex glomerulonephritis is induced in rats immunized with heterologous myeloperoxidase. Clin Exp Immunol 1994; 97: Cochrane C, Weigle W, Dixon F. The role of polymorphonuclear leucocytes in the initiation and cessation of the Arthus vasculitis. J Exp Med 1959; 110: Ambrus JL Jr, Sridhar NR. Immunologic aspects of renal disease. JAMA 1997; 278: Ferrario F, Tadros MT, Napodano P, Sinico RA, Fellin G, d Amico G. Critical reevaluation of 41 cases of idiopathic crescentic glomerulonephritis. Clin Nephrol 1994; 41: Tse WY, Howie AJ, Adu D et al. Association of vasculitic glomerulonephritis with membranous nephropathy: a report of 10 cases. Nephrol Dial Transplant 1997; 12: Heeringa P, Brouwer E, Klok PA et al. Autoantibodies to myeloperoxidase aggravate mild anti-glomerular-basementmembrane mediated glomerular injury. Am J Pathol 1996; 146: King WJ, Brooks CJ, Holder R, Hughes P, Adu D, Savage OS. T lymphocyte responses to anti-neutrophil cytoplasmic autoantibody (ANCA) antigens are present in patients with ANCA-associated systemic vasculitis and persist during disease remission. Clin Exp Immunol 1998; 112: Schmitt WH, Heesen C, Csernok E, Rautmann A, Gross WL. Elevated serum levels of soluble interleukin-2 receptor in patients with Wegener s granulomatosis. Association with disease activity. Arthritis Rheum 1992; 35: Received for publication: Accepted in revised form: