reperfusion injury contributes to renal damage in experimental anti-myeloperoxidase-associated proliferative glomerulonephritis

Kidney International, Vol. 47 (1995), pp. 1121—1129

Renal ischemia/reperfusion injury contributes to renal damage in experimental anti-myeloperoxidase-associated proliferative glomerulonephritis ELISABETH BROUWER, PIETER A. KLOK, MINIc G. HUITEMA, JAN J. WEENING, and CEES G.M. KALLENBERG Department of Clinical Immunology and Pathology, University of Groningen, Groningen, and Department of Pathology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands

Renal ischemia/reperfusion injury contributes to renal damage in experimental anti-niyeloperoxidase-associated proliferative glomerulone. phritis. The occurrence of focal fibrinoid necrosis of capillary loops in the very early stages of ANCA-associated necrotizing crescentic glomerulonephritis (NCGN) and the increased prevalence of this disease at older

age suggest that renal ischemia may play an additional role in its pathophysiology. In the present study we investigated the contribution of renal ischemia to the induction of anti-myeloperoxidase (MPO) associated NCGN in a previously described rat model of this disease. The develop-

ment of renal lesions is dependent on the presence of an anti-MPO immune response and the localization of a lysosomal extract containing lyric enzymes and MPO in combination with hydrogen peroxide (H202) along the glomerular basement membrane (GBM). The hypothesis tested whether perfusion of hydrogen peroxide (H202) could be replaced by ischemia/reperfusion (I/R) injury, as hR injury activates endothelial cells to produce oxygen metabolites. hR was induced by clamping the renal artery for 20 minutes in kidneys in which the circulation had been restored

tious origin, or may occur in patients without a systemic illness [1]. NCGN can be classified into three immunopathogenic categories:

(i) anti-glomerular basement membrane (GBM) antibody medi-

ated, (ii) immune complex mediated, and (iii) anti-neutrophil cytoplasmic antibody (ANCA) associated [2}. ANCA in patients with ANCA associated NCGN are, generally, either directed to proteinase 3 (Pr3) [3—5] or myeloperoxidase (MPO) [2, 6, 7], both

myeloid lysosomal enzymes. The initial lesion in this form of NCGN is characterized by segmental fibrinoid necrosis of the glomerular basement membrane (GBM) followed by marked infiltration of neutrophils and mononuclear cells, and paucity of IgG deposits [8—12]. Focal fibrinoid necrosis of capillary loops is

present at a very early stage in the development of NCGN, indicating that focal endothelial damage and GBM necrosis,

several minutes after perfusion with the lysosomal extract in MPO possibly aggravated by ischemia, are prerequisites for the develimmunized rats. Rats developed lesions characterized by intra- and opment of ANCA-associated NCGN. Ischemia followed by reperextracapillaiy cell proliferation, periglomerular infiltration, ruptures in fusion (h/R) is known to induce endothelial cell damage as well as Bowman's capsule, ischemic tubuli, and interstitial mononuclear infiltrate. Immune deposits, however, persisted for a longer time along the GBM leukocyte accumulation in reperfused organs. Renal hR injury is after perfusion of lytic enzymes followed by hR injury compared to characterized by reduced renal blood flow, increased vascular previous studies in which H202 in conjunction with lytic enzymes were perfused in MPO-immunized rats. Control groups consisting of control resistance, decreased glomerular filtration rate, erythrocyte trapimmunized rats perfused with the lysosomal extract followed by hR and

MPO immunized rats perfused with the lysosomal extract developed minor lesions only. To further explore the possible mechanisms involved in the contribution of hR to the development of NCGN, we studied the expression of ICAM-1 in renal tissue from the three experimental groups.

Up-regulation of ICAM-1 occurred in MPO-immunized rats perfused with the lysosomal extract followed by hR and also in both control groups, suggesting that up-regulation of ICAM-1 in itself is not sufficient to induce lesions. In conclusion, renal ischemia contributes to the development of experimental anti-MPO associated NCGN possibly by inducing endothehal cells to produce toxic oxygen metabolites.

ping, and tubular obstruction proportional to the duration of the ischemic insult [13]. Activation of endothelial cells in reperfused organs is characterized by elevated intracellular calcium levels, production of platelet activating factor, nitric oxide (NO), and

superoxide (O2) production, and up-regulation of adhesion molecules like selectins and intercellular adhesion molecule-i

(ICAM-1) [14—19]. Evidence has accumulated that neutrophils play a pivotal role in renal reperfusion injury by the production of oxygen radicals and release of lysosomal enzymes [13, 20]. Several recent studies on limb hR associated lung injury and myocardial hR injury have demonstrated that blocking of ICAM-1 on endothelial cells and/or f32-integrins on neutrophils attenuates reperNecrotizing crescentic glomerulonephritis (NCGN) clinically fusion injury [17, 21]. No studies so far have been performed on manifest as rapidly progressive glomerulonephritis, may be asso- the role of adhesion molecules in renal hR. In a recently ciated with multisystem disorders from infectious or non-infec- developed model of anti-MPO associated NCGN, characterized by glomerular intracapillary thrombosis followed by a proliferative glomerulonephritis, we demonstrated that anti-MPO alone is not sufficient for the induction of NCGN but that MPO, lytic enzymes, Received for publication August 8, 1994 and in revised form October 28, 1994 and hydrogen peroxide need to be localized along the GBM [22}. Accepted for publication October 31, 1994 in view of the focal fibrinoid necrosis of capillary loops in human ANCA associated NCGN [23, 24], in which focal ischemia may © 1995 by the International Society of Nephrology

1121

1122

Brouwer et al: Renal ischemia in anti-myeloperoxidase GN

play an additive role, and the need for oxygen radicals in the experimental model for anti-MPO associated NCGN, we tested

Immunization procedure

BN rats were immunized with human MPO in Complete whether ischemia could play a role in the induction of experimenFreunds Adjuvant (CFA) supplemented with H37Ra, 5mg/mi tal NCGN. Ischemia was induced in MPO immunized rats, (Difco, Detroit, MI, USA). Rats received 10 g MPO or control directly after unilateral renal perfusion with a lysosomal extract solution without MPO (= sodium-acetate buffer) subcutaneously containing MPO, HLE, and Pr3, without hydrogen peroxide, by (s.c.) in 0.2 ml at two sites near the tailbase. clamping the left renal artery for 20 minutes starting after the circulation had been reconstituted for several minutes. In addiDetection of anti-MPO antibodies by ELISA

tion, since adhesion molecules have been shown to play an

Labstar (Greiner, Kremsmünster, Austria) microtiter plates were coated overnight with human MPO (isolated as described I/R, we also investigated the presence and up-regulation of before) at a protein concentration of 20 pg/ml in 0.1 M carbonate adhesion molecules in relation to renal hR. buffer, pH 9.6. The plates were incubated with rat plasma, diluted in 0.05 M Tris, 0.05% Tween, 2% BSA, and 0,3 M NaCl, pH 8.0, Methods starting at a dilution of 1:100 in PBS. Antibody binding was detected with sheep anti-rat Ig conjugated to alkaline phosphatase Rats (Serotec, Oxford, UK), followed by p-nitrophenyl phosphate All experiments were performed on conventionally housed, disodium as a substrate. The OD was read at 405 nm, and a three month old Brown-Norway (BN) rats obtained from Harlan, standard curve was prepared from a reference serum. Antibody Bilthoven, The Netherlands, fed ad libitum with standard chow concentrations were computed from the linearized titration curve obtained after log-logit transformation of the absorbances of the (Hope Farms, Woerden, The Netherlands). respective dilutions of the reference serum. The concentration of the reference serum was set at 100 U. Lysosomal enzyme extract important role in limb hR associated lung injury and myocardial

Human polymorphonuclear leukocytes (PMN) were isolated from buffy coats on a ficoll density gradient followed by dextran sedimentation. After sedimentation the remaining erythrocytes were removed by hypotonic lysis. MPO was extracted from PMN by dissolution of the cells in cetyltrimethyl ammonium bromide (CETAB) and sonification. Nuclei and membrane fragments were discarded by ultracentrifugation and the extract was absorbed to a concanavalin A sepharose gel and eluted with a-methyl-D-mannoside. Eluted fractions with a ratio (OD 428/280) greater than

Renal perfusion Unilateral perfusion of the left kidney was performed according

to a modification of the method of Hoyer, Mauer, and Michael

[27]. Rats were anesthesized with halothane, 02, and NO2. Through a midline incision, the aorta and vena cava were exposed

by blunt dissection. After ligation of the tributaries, temporary

clamps were placed on the aorta above the left renal artery, leaving the circulation of the right renal artery undisturbed.

0.5, measured as the ratio between the optical density (OD) Through a puncture hole in the aorta, a needle was inserted and

advanced up to the level of the left renal artery. The left renal vein obtained at 428 nm (showing a specific spectral band for MPO) and the OD obtained at 280 nm, were pooled and extensively was punctured to let blood and perfused fluids escape, which were dialyzed to sodium-acetate buffer pH 6.0 [25]. The resulting collected in cotton wool. PBS (37°C, pH 7.3) was infused until the

lysosomal extract had an OD 428/280 ratio greater than 0.5 and kidney became pale and no visible blood escaped from the contained mainly MPO, proteinase 3, and trace amounts of punctured renal vein. The lysosomal extract containing 40 U of elastase, but no lactoferrin as determined by sandwich ELISA [3]. MPO in PBS was perfused during one minute. After an additional exposure time of two minutes the kidney was again perfused with In accordance with the results of ELISA specific bands for MPO (at 15, 39 and 58 kD) and Pr3 (at 30 kD) were found by gel PBS according to Johnson et al [281. Following perfusion the needle was removed, the aorta was closed with atraumatic sutures, electrophoresis (results not shown). Peroxidase activity of the the clamps were removed and the kidney was allowed to reperextract was assayed spectrophotometrically by measuring the fuse. Total ischemia time was always less than 12 minutes. Only increase in absorbance at 470 nm upon oxidation of guaiacol rats in which the kidney regained normal perfusion and color were (Sigma Chemical Co, St. Louis, MO, USA). In brief, 10 d of the used for ischemia/reperfusion studies. About five minutes after MPO preparation was added to a 1-cm light path cuvette containing 3 ml of the assay solution. The solution consisted of 29.6 ml H2O, 3 ml 0.1 M sodium phosphate buffer, pH 7.0, 0.1 ml 0.1 M H202, and 48 d guaiacol. One unit (U) of peroxidase activity is defined as the amount that consumes 1 mol of H202 per minute [26]. One hundred micrograms of the lysosomal extract contained approximately 78 g MPO with an activity of 30 U.

reperfusion of the left renal artery, the left renal artery was

clamped for 20 minutes in order to induce ischemia followed by reperfusion. After surgery the rats were placed under a heat lamp for two hours and received an injection with buprenorfine (Reckitt and Coleman, UK) s.c. as a pain reliever. Histological examination

At different time intervals after renal perfusion with the lysosomal extract, rats were perfused with PBS at 4°C to remove the For immunization MPO was further purified from the lysoso- blood from both kidneys, and sacrificed. Specimens from both mal extract on a sephadex G150 gel (Pharmacia, Fine Chemicals kidneys were obtained and prepared for light microscopy, immuAB, Uppsala, Sweden). Fractions with an OD 428/280 ratio nofluorescence, and immunohistochemistry. For light microscopy, greater than 0.8 were pooled. Contamination with Pr3 or HLE renal tissue was fixed in 2% paraformaldehyde/PBS and embedwas ruled out by antigen specific ELISA for Pr3 and HLE. ded in plastic. Four micrometer sections were stained with Myeloperoxidase

1123

Brouwer et al: Renal ischemia in anti -myeloperoxidase GN

Table 1. Antibodies used for immunofluorescence and immunohistochemistiy on frozen tissues Code

MPO 4.15 ED1 His 48 R73 1A29

0X42

Antigen recognized human-MPO human-MPO cytoplasmic membrane TCR ICAM-1 CD11b

Ig(sub-)class poly

IgGi IgGi IgGi

IgGi IgGi IgG2a

Specificity

Source

PMN/monocytes PMN/monocytes PMN neutrophils T-cells

Dakopatts

Broad neutrophils,

CLB [40] [41] [42] Seikagaku Oxford

monocytes, NK-cells

Sources are: Dakopatts, Glostrup; Denmark; CLB, Central Laboratory for the Blood Transfusion Service, Amsterdam, The Netherlands; Seikagaku, Kogyo, Japan; Oxford, MRC Cellular Immunology Unit, William Dunn School of Pathology, University of Oxford, England, UK.

antibody for 60 minutes; (c) 0.06% H202 in PBS at RT, to block endogenous peroxidase; (d) rabbit anti-mouse peroxidase labeled immunoglobulin 1:40, supplemented with 1% rat serum for 60 minutes, followed by aminoethylcarbimazole (AEC) 0.2 mg/mi in 0.05 M acetate buffer pH 4.9, containing 5% dimethylformamide supplemented with 0.03% hydrogen peroxide or diaminobenzidine (DAB) 0.5 mg/ml plus imidazole 1 mg/mI, in 50 ml PBS

supplemented with 0.03% hydrogen peroxide for 10 minutes. Sections were counterstained with hematoxylin, mounted in glycerin, and examined. All necessary dilutions were made in PBS. Appendix and spleen sections served as positive control sections for ICAM-1 and CD11b. The intensity of staining was scored on a semi-quantitative scale: — = absent, + = weak, + + = mild,

+ + + = strong, + + + + = very strong. In addition, sections obtained from kidneys of normal rats (N = 2) were also stained for ICAM-1 and CD11b. Plastic sections

hematoxylin-eosin, periodic acid Schiff, or periodic acid silver To investigate the precise localization of ICAM-1 within gbmethamine. For immunofluorescence (IF) and immunoperoxi- meruli and interstitium renal tissue was also embedded in plastic. dase staining (IP), material was snap frozen in freon or iso- In short, biopsies were fixed in 2% paraformaldehyde in PBS pH penthane (isobutane), and 2 or 4 jim sections were cut. Sections 7.4 for one hour at 4°C. After washing overnight in PBS containwere standard fixed with acetone. For detection of MPO, sections ing 6% sucrose, the biopsies were dehydrated with 100% acetone were fixed with acetone-buffered-formalin 9% [29]. After fixation, for 30 minutes at 4°C. After dehydration the Ussues were infilsections were incubated with 10% normal rabbit, goat or swine trated with glycomethacrylate (Technovit 8100 solution A, Kulzer, serum, depending on the conjugate used. Detection of MPO, Ig, GmbH, Wehrheim, Germany) at 4°C for eight hours. Following and complement was done by IF using mouse anti-human MPO infiltration tissues were embedded in a mixture of Technovit (Table 1), rabbit anti-rat IgGIAIM, F(ab')2 goat anti-rat IgG, and solutions A with an initiator for polymerization (30:1) according F(ab')2 goat anti-rat C3, respectively (all from Cappel, West to the manufacturer's recommendation. Polymerization took Chester, PA, USA), followed by FITC labeled conjugates. The place overnight on crushed ice at 4°C. The whole procedure was following conjugates were used: rabbit anti-mouse FITC (Dako- done at 4°C and all necessary incubation steps were performed on patts, Glostrup, Denmark), F(ab')2 goat anti-rabbit FITC, and a rotor. Two micrometer thick sections were cut on a Reichert F(ab')2 rabbit anti-goat FITC (both from Cappel). Granulocytes, Jung Supercut plastic microtome using tungsten carbide tipped monocytes, and T cells were detected by IP using the monoclonal hard metal knives [30]. Sections were caught on 2% paraformalantibodies ED1, His48, and R73, respectively (Table 1), followed dehyde, affixed on glass slides, and dried. by peroxidase (P0) labeled conjugates. We used rabbit antiTo study the localization of adhesion molecules sections were mouse P0 (Dakopatts) with aminoethylcarbimazole as a sub- stained using the following method: (a) preincubation with norstrate. Endogenous peroxidase activity was blocked with 0.05% mal rabbit serum 10% for 30 minutes at 37°C; (b) monoclonal H202 in PBS. Sections were counterstained with hematoxylin. All antibody incubation for two hours at 37°C in the appropriate necessary dilutions were made in PBS pH 7.2. The amount of dilutions; (c) blocking of the endogenous peroxidase activity when

MPO, IgG, and complement deposition was estimated on a necessary with hydrogen peroxide 0.06% in PBS at RT; (d) weak, + mild, ++ peroxidase conjugated rabbit anti-mouse 1:40 (Dakopatts) supsemi-quantitative (3+) scale: — absent, intermediate, + + + strong. To estimate the number of ED1-, plemented with 1% normal rat serum for 60 minutes at 37°C; (e) His48- and R73-positive cells, we calculated the average number DAB 0.5 mg/mi in PBS supplemented with imidazole 1 mg/ml plus of cells per glomerulus by counting all positive cells present within 0.03% H202 for 10 minutes. Sections were counterstained with

50 giomeruli and dividing the total number by 50. Interstitial

hematoxylin or periodic acid Schiff. The median number of

leukocytes were quantified by using a semi-quantitative method: — minimal, + mild, + + moderate, and + + + severe. absent, For immunofluorescence sections were scored in a blinded fashion by two persons.

glomeruli available for counting was 3 and the range varied from 2 to 13 per section.

Expression of adhesion molecules

All values were expressed as the mean SuM. Analysis for statistical differences was done by a nonparametric one way analysis of variance (ANOVA) for unpaired data. Correlations

In an additional set of experiments sequential cryostat sections were stained for ICAM-1 and CD11b (Table 1). The appropriate fixation, dilution, and specificity of the MAbs used was tested on control rat kidney and rat spleen tissues. Briefly, cryostat sections

Statistical analysis

within groups for different values were assessed with Spearman's rank correlation test.

of snap frozen material, cut 4 jim, were dried for at least 10 minutes and subsequently fixed in acetone. Sections were incubated sequentially with: (a) normal rabbit serum 10% in phos-

Experimental design

Three groups of experiments were performed according to the

phate buffered saline (PBS) for 20 minutes; (b) monoclonal following design:

1124

Brouwer et al: Renal ischemia in anti-myeloperoxidase GN

1200

Table 2. Intraglomerular deposition of MPO, IgG, and C3 along the GBM in rats sacrificed 24 hours, 4 days, and 10 days after perfusion

960

MPO S

720

0a-

.

. .. S

480 240

0

SII

MPO Ctrl 1. Anti-MPO antibody titers in control (Ctrl) (N = 6) and MPO Fig. immunized (N = 10) rats. Antibodies to human MPO were measured by ELISA at 5 weeks after immunization. Results are expressed as arbitrary units.

I. Renal perfusion with the lysosomal enzyme extract in MPOimmunized rats followed by hR. The lysosomal enzyme extract was

perfused into the left kidney in MPO-immunized rats, five weeks after immunization. After reperfusion the left kidney was clamped to induce ischemia and reperfused after 20 minutes. Rats were sacrificed at 24 hours (N = 2), 4 days (N = 7), and 10 days (N = 4) after perfusion. II. Renal perfusion with the lysosomal enzyme extract in control immunized rats followed by ER. The same procedure was followed for control immunized rats as described in I. Rats were sacrificed at 24 hours (N = 2), 4 days (N = 2), and 10 days (N = 2) after perfusion. III. Renal perfusion with the lysosomal enzyme extract in MPOimmunized rats. The lysosomal enzyme extract was perfused into the left kidney in MPO-immunized rats, 5 weeks after immunization. Rats were sacrificed at 24 hours (N = 2), 4 days (N = 2), and 10 days (N = 2) after perfusion.

Results Humoral immune response to MPO

BN rats immunized with human MPO developed antibodies

Groups I II III

C3

IgG

24

4

10

hr

days

days

+++

++

—

—

+

—

4

+± — —

10

24 hr days days

24

hr

4

10

days days

+± +± ++± ++±

—

—

—

—

—

++

+

+

++

+

+

Animals were studied in three groups: Group I, MPO immunized and perfused with the lysosomal extract followed by ischemia/reperfusion (N = 13); Group II, control immunized and perfused with the lysosomal extract followed by ischemia/reperfusion (N = 6); and Group III, MPO immu-

nized and perfused with the lysosomal extract (N = 6). The amount of MPO, IgG, and complement deposition was estimated on a semi-quantitative (3+) scale: —, absent; weak; +, mild; ++, intermediate; +++, strong.

(Fig. 2 B, C). IgG and C3 were present also within vessels and in Bowman's space (Fig. 2B). Histopathology

Severe lesions were only found in rats which were immunized

with MPO and perfused with the lysosomal enzyme extract followed by hR. At 24 hours the main abnormalities detected in the perfused left kidney consisted of accumulation of eosinophilic proteinaceous material into the capillary lumen and Bowman's space, sometimes accompanied by necrosis of the GBM (Fig. 3A).

At 4 and even more at 10 days after perfusion proliferative changes could be observed within glomeruli and interstitium. Within glomeruli we found necrosis and wrinkling of the GBM, intra- and extracapillary proliferation, periglomerular infiltration, giant cell formation, and breaks in Bowman's capsule (Fig. 3 B, C). Within the interstitium mononuclear infiltrates were found

surrounding glomeruli, tubuli, and vessels (Fig, 3D). Tubules showed ischemic changes characterized by eosinophilic necrosis and desquamation of tubular epithelial cells. Within the tubular lumina eosinophilic and sometimes cellular casts were found. In some animals vessels were observed with hyalinized walls surrounded by mononuclear cells (Fig. 3 C, D). Within group II rats

(control immunized and perfused with the lysosomal extract

reacting with human MPO (Fig. 1). For perfusion studies, rats five followed by I/R) we observed deposition of eosinophilic material weeks post-immunization were used, that is, at a time when the within the glomeruli at 24 hours after perfusion. Tubules showed humoral response was maximal [22]. ischemic changes at 4 and 10 days after perfusion. In addition Deposition of MPO, IgG, and complement MPO was detected along the GBM together with IgG and C3 at

24 hours, 4 days, and, weakly, 10 days after perfusion in MPO immunized rats perfused with the lysosomal extract followed by hR (group 1) (Table 2). MPO was present in a diffuse granular pattern (Fig. 2A) whereas IgG and C3 were present in a more linear pattern. In control immunized rats (group II) MPO could not be detected along the GBM. MPO was also present along the GBM for a much shorter period, up to 24 hours, in rats from group III (MPO immunized and perfused with the lysosomal

slight interstitial infiltrate was present. However, proliferative changes were not observed. Also, in group III rats (MPO immunized and perfused with the lysosomal extract) no significant lesions were observed, except for a transient cell influx at 24 hours and a slight interstitial infiltrate.

Immunophenotyping In accordance with the findings by light microscopic examination, immunophenotyping of renal tissue from group I rats showed larger numbers of intraglomerular, endo- and extracapillary macrophages and PMN at days 4 and 10 after perfusion (Table 3 and extract) as compared to group I rats (MPO immunized rats Fig. 3 F, H). T cells were observed in small numbers within the perfused with the lysosomal extract followed by I/R) (Table 2). glomeruli and more abundantly within the interstitium at day 10 Despite the short localization of MPO in group III, IgG and C3 (Table 3). A transient glomerular neutrophil and monocyte influx stayed along the GBM up to 10 days both in groups I and III was found at 24 hours in MPO immunized rats perfused with the

Brouwer et al: Renal ischemia in anti-myeloperoxidase GN

1125

Fig. 2. Localization of MPO, lgG, and complement by immunofluoresence.

A, B, and C are photographs from BN rats perfused with the lysosomal extract followed by ischemia/reperfusion injury (group I). (A) Diffuse granular deposits of MPO along the GBM in a rat sacrificed at 24 hours after perfusion (X320). (B) Granular IgG deposits in a rat sacrificed at 10 days after perfusion (X320). (C) Granular deposits of complement along the GBM at day 10 after perfusion (X320).

lysosomal extract (Group III). Within the interstitium mild infil- tended to correlate with the infiltration of CD11b positive cells (r tration of inflammatory cells was seen in groups II and III at 24 = 0.374, P < 0.1). hours and 4 days after perfusion. Discussion Up-regulation of adhesion molecules In the present study we investigated the contribution of renal The intensity of ICAM-1 staining was strong and diffusely ischemia to the induction of anti-myeloperoxidase (MPO) associpresent, with staining of all glomerular cells, in group I (MPO ated necrotizing crescentic glomerulonephritis. In a previous immunized and perfused with the lysosomal extract followed by study we described that the development of NCGN in the rat is IIR) and group III rats (MPO immunized and perfused with the dependent on the presence of an anti-MPO immune response and lysosomal extract). The high intensity of ICAM-1 staining in those the localization of a lysosomal extract containing lytic enzymes groups differed significantly (P < 0.008) from the mild and diffuse and MPO in combination with hydrogen peroxide (H202) along staining as found in group II (control immunized and perfused the glomerular basement membrane (GBM) [22}. It was postuwith the lysosomal extract followed by fIR; Fig. 3 E, G; Table 4). lated that the immune deposits initially present were rapidly However, compared to normal control rats in which glomerular cleared by the active proteolytic enzymes together with reactive ICAM-1 staining was absent or only weakly present (data not oxygen species, resulting in the pauci-immune character of the shown), ICAM-1 was also up-regulated in group II. Proximal lesions [22]. In the present study rats were immunized with MPG tubular epithelial cells were positive for ICAM-l in groups I and and perfused with the lysosomal extract only. It proved that III, but not in group II. Interstitial infiltrates, as present in group perfusion of hydrogen peroxide (H202) could be replaced by

I, were weakly IcAM-1 positive. Cells positive for CDIIb, a ischemialreperfusion (I/R) injury in producing the aforemenligand of ICAM-1, were most abundantly found in group I (MPO tioned lesions although immune deposits were detectable for a immunized and perfused with the lysosomal extract followed by longer period compared to the previous study [22]. The findings I/R) and not in group II (control immunized and perfused with suggest that hR injury, which activates endothelial cells and bind lysosomal extract followed by I/R). The amount of infiltrated neutrophils possibly by the production of oxygen metabolites [14, CD11b positive cells in groups I and III differed significantly from 31], contributes to the development of lesions in NCGN. To study the effect of renal ischemia in this model three that of group II (P < 0.00 1). Intraglomerular ICAM- I expression

1127

Brouwer et al: Renal ischemia in anti-myeloperoxidase GN

Fig. 3. A, B, C, and D. Light microscopy of lesions found in MPO-immunized rats perJbsed with the lysosomal extract followed by ischemia reperfusion injuiy

(group I). Sections were stained with either hematoxylin (HE) or methamine-silver (MS) with aniline red counterstaining. (A) Capillaries occluded with

eosinophilic material and erythrocytes and tuft necrosis at 24 hours after perfusion (MS, X560) (B), (C), and (D) Intra- and extracapillary cell proliferation, crescent formation and periglomerular and perivascular infiltration of inflammatoiy cells, tubular necrosis and interstitial infiltration of inflammatory cells at 10 days after perfusion (MS and HE, X225). E, F, G, and H. Light microscopy of sections stained, by immunohistochemistry, with ED1, a monocyte marker, and ICAM-1, a cell adhesion molecule in MPO-immunized rats perfused with the lysosomal extract followed by ischemia reperfusion injury (group I). Sections were counterstained with hematoxylin. (E) Diffuse staining of ICAM-1 within glomeruli and Bowman's capsule in a rat sacrificed at 4 days after perfusion (X560). (F) Intracapillary infiltration of monocytes in a rat sacrificed at 4 days after perfusion (ED-i, x560). (G) Diffuse staining of ICAM-1 within glomeruli and on vessel walls in the interstitium in a rat sacrificed at 10 days after perfusion (X225). (H) Intraand extracapillary as well as periglomerular and peritubular infiltration of monocytes in a rat sacrificed at 10 days after perfusion (ED-i, x225).

Table 3. Intraglomerular and interstitial presence of neutrophils, macrophages, and T cells, as detected by the mAb's ED1, His 48, and R73, respectively (see Table 1)

Interstitial cells'

Intraglomerular cellsa Infiltrating cells

24 hr

Groups

PMN

T cells

24 hr

4 days

10 days

0.5

+

++

++

0.4 0.1

— —

—

10 days

III

1.7

0.1

1

2.8± 1.1

3.2± 1.0

7.1± 4.3

H

0.3

III

1.5

0.3 1.4

0.6 0.3 0.5

0.2 0.3

0 0 0

1

II III

1.5 0.2

1.6 0.5 0.5

0.5 0.1

II

Mø

4 days 3.1 0.8 0.6

2.6 0.7

1

0.1

0.0 0.1 0.1

0.1

0 0

0 0

++

++±

0.3 0.0

—

0.7

— — —

+

++++ +±

++

— —

—

Rats were sacrificed at 24 hours, 4 and 10 days after perfusion. The three groups included: Group I, MPO immunized and perfused with the lysosomal extract followed by ischemia/reperfusion (N = 13); Group II, control immunized and perfused with lysosomal extract followed by ischemia/reperfusion (N = 6); and Group III, MPO immunized and perfused with the lysosomal extract (N = 6). Numbers represent mean SEM. For estimation of the number of PMN, monocytes and T cells, we calculated the mean number of cells per glomerulus by counting all positive cells present within 50 glomeruli and dividing the total number by 50 Labeled interstitial leukocytes were quantified by using a semiquantitative method: — absent, minimal, + mild, + + moderate, and + + + severe

Table 4. Up-regulation of IcAM-1 and influx of CD11b positive cells in rats sacrificed 24 hours, 4 days, and 10 days after perfusion

IAM-1 Groups I II III

CDIIb

24 hr

4 days

10 days

+++ +±

++± ++ +++

+++ +±

24 hr 3.1 0.1

4 days

10_days

1.7±0.8 3.3±2.2 0.1 0.0

0.5 0.5

0.0 0.0

0.1 0.4

0.0 0.4

no deposition of MPO was observed. Interestingly, MPO remained present along the GBM up to 10 days in group I in contrast to group III. In a previous study in which MPO immunized rats were perfused with the lysosomal extract and H202, MPO was present for a shorter period of time along the GBM [22]. The prolonged localization of MPO in group I rats might be

related to the fact that the renal artery was clamped soon after perfusion with the lysosomal extract in group I and not in group Animals were divided into: Group I, MPO immunized and perfused III. The exact mechanisms underlying removal or persistence of with the lysosomal extract followed by ischemia/reperfusion (N = 13); MPO are at present not clear. In contrast to the short presence of Group II, control immunized and perfused with the lysosomal extract followed by ischemia/reperfusion (N = 6); Group Ill, MPO immunized MPO in group III, IgG and C3 stayed along the GBM up to 10 and perfused with the lysosomal extract (N = 7). The amount of ICAM-1 days after perfusion in both groups I and III, The far longer was estimated on a semi-quantitative (4+) scale: —,absent; +, weak; ++, presence of the antibody in comparison to the antigen, MPO in mild; + + +, strong; + + + +,very strong. For estimation of the number of CD11b positive cells, we calculated the mean number of cells per the present model, has also been reported in other studies in glomerulus by counting all positive cells present within 50 glomeruli and which cationized IgG served as an antigen [32, 33]. The rapid dividing the total number by 50. Numbers represent mean SEM. clearance of the antigen from the GBM might be due to its faster removal or masking of the antigen by the antibody [32]. In a previous study in which we perfused the lysosomal extract in

+++±

++±

0.9

experimental groups were formed: group I in which MPO immunized rats were perfused with the lysosomal extract followed by I/R, group II in which control immunized rats were perfused with

the lysosomal extract followed by I/R, and group III in which MPO immunized rats were perfused with the lysosornal extract only. All rats immunized with MPO developed an anti-MPO response. We chose to clamp the kidney after perfusion and subsequent recovery because clamping before perfusion would result in a poor circulation and distribution of the MPO.

combination with H202, we observed fast removal of IgG after 24 hours [22], in contrast to the present study in which IgG persisted along the GBM up to 10 days. A possible explanation might be the longer presence of MPO in the present study or the more effective

degradation of lgG in the previous study due to the presence of higher concentrations of H202 along the GBM as well as higher numbers of toxic monocytes [22]. In our previously described model [22] in situ immune complex formation, resulting from binding of anti-MPO to MPO localized to the GBM, played an MPO localized along the GBM in a granular pattern, as essential role in the initial phase of the lesions. However, immune detected at 24 hours after perfusion in groups I and III rats that complexes were rapidly cleared from the lesions. In accordance, were MPO immunized. In group II, the control immunized rats, Germuth et al [9] postulated that in human pauci-immune NCGN

1128

Brouwer ci at: Renal ischemia in anti-myeloperoxidase GJV

immune deposits are also initially present but rapidly cleared. in MPO immunized rats also resulted in a strong increase in Various factors may be involved in the rapid clearance of immune deposits as mentioned before. Lytic enzymes and reactive oxygen species derived from activated neutrophils and monocytes/macro-

ICAM-1 expression, but was in itself not sufficient to induce influx of inflammatory cells. Generally, a weak association was observed between intraglomerular expression of ICAM-1 and the presence

phages appear to be involved. In this respect, the neutrophil of cells positive for CD11b (ligand of ICAM-1). Apparently, activating capacity of anti-MPO [34, 35] is of interest. Primed activation of inflammatory cells with up-regulation of their adheneutrophils express their granule constituents, such as MPO and sion molecules is another prerequisite for their influx into the proteinase 3, at their cell membrane [36]. Antibodies to these renal tissue. Models of lung injury induced by hind limb hR injury constituents can bind to their exposed antigens resulting in have demonstrated the important contribution of elevated levels degranulation and the production of reactive oxygen species [34, of cytokines like IL-i and TNF in this respect [17, 21]. These 351. Immunization of rats with human MPO results in the factors might therefore have played an important role in our generation of antibodies that cross react to a variable extent with rat MPO (unpublished observation). The extent of cross reactivity

model by activation of both endothelial cells and PMN, and need further investigation.

In conclusion, hR injury contributes to the development of antibodies and, as such, to the extent of persistence of immune anti-MPO associated NCGN. These data are relevant in view of deposits in the lesions. In this respect it is noteworthy that Yang, the initial histopathological lesions of ANCA associated NCGN Jennette and Falk [1 could not reproduce the pauci-immune and the occurrence of this disease particularly in elderly patients. character of the lesions in a nearly identical rat model of The exact mechanisms involved need further investigation. anti-MPO associated NCGN. Further studies directed, among others, at the induction of autoantibodies to rat MPO, should Acknowledgment elucidate the factors that are precisely involved in the clearance of This study was supported by Grant C88.733 from the Dutch Kidney immune deposits to better understand human anti-MPO associ- Foundation. ated pauci-immune NCGN. By immunohistology the most severe lesions were observed in Reprint requests to C.G.M. Kallenberg, M.D., Ph.D., Department of group I rats. At 24 hours the main abnormalities consisted of Clinical Immunology, University Hospita4 Oostersingel 59, 9713 EZ Gronintracapillary accumulation of eosinophilic proteinaceous mate- ingen, The Netherlands. rial. At 4 days and even more pronouncedly at 10 days, proliferReferences ative changes such as ultra- and extracapillary cell proliferation, periglomerular infiltration, and ruptures of Bowman's capsule 1. COUSER WG: Idiopathic rapidly progressive glomeruloncphritis.Am J could be observed. Within the interstitium tubuli showing ischNephrol 2:57—69, 1982 emic changes were surrounded by a mononuclear infiltrate. Rats 2. FALK Ri: ANCA-associated renal disease (clinical conference). Kidfrom group H only showed slight ischemic changes and some ney mt 38:998—1010, 1990 interstitial infiltration of mononuclear cells, demonstrating that 3. GOLDSCHMEDING R, ScElooT CE VAN DER, BOKKEL-HUININK D TEN, HACK CE, ENDE CE VAN DEN, KALLEN[SERG CGM, BORNE AEGKR hR injury in itself did not induce those lesions. In group III a vON oc: Wegener's granulomatosis autoantibodies identify a novel transient influx of neutrophils and monocytes was observed. The diisopropylfluorophosphate-binding protein in the lysosomes of nor-

may be related to the neutrophil activating capacity of the

transient presence of inflammatory cells, in contrast to the

persistence of IgG and C3, in group III has also been observed in other models of in situ immune complex formation [38, 39].

In the present study we found that renal hR injury in MPO immunized rats perfused with lysosomal extract contributes to the development of proliferative necrotizing glomerulonephritis. hR

injury alone (group II) resulted in minor lesions only. The mechanisms underlying the contribution of ItR to the pathogenesis of anti-MPO associated NCGN might be several-fold. It has been shown that MPO immunized rats perfused with hydrogen peroxide, an oxygen metahohite, in addition to a lysosomal extract developed severe proliferative NCGN [22]. The present findings suggest that renal ischemia induces endothehial cells to produce H202 and, as such, provide the substrate for MPO. The prolonged presence of MPO, as well as IgG and complement, observed after

the induction of hR in MPO immunized rats perfused with the

mal human neutrophils. J Clin Invest 84:1577—1587, 1989 4. NILES JL, MCCLUSKEY RT, Al-IMAD MF, ARNAOUT MA: Wegener's

granulomatosis autoantigen is a novel neutrophil serine protease. Blood 74:1888—1893, 1989

5. LUDEMANN J, UTECHT B, GROSS WL: Anti-neutrophil cytoplasm

antibodies in Wegener's granulomatosis recognize an elastinolytic enzyme.JExp Med 171:357—362, 1990 6. FALK RJ, JENNETTE JC: Anti-neutrophil cytoplasmic antibodies with specificity for myeloperoxidase in patients with systemic vasculitis and

idiopathic necrotizing and crescentic glomerulonephritis. N Engl J Med 318:1651—1657, 1988 7. COHEN TERVAERT JW, GOLDSCHMEDING R, ELEMA JD, GIES5EN M VAN DEE, HUITEMA MG, HEM GK VAN DER, TUE TH, BORNE AEGKR

VON DEM, KALLENBERG CGM: Autoantibodies against myeloid lyso-

somal enzymes in crescentic glomerulonephritis. Kidney mt 37:799— 806, 1990 8. HORN RG, FAuci AS, ROSENTHAL AS, WOLFF SM: Renal biopsy pathology in Wegener's granulomatosis. Am J Pathol 74:423—440, 1974 9. GERMUTH FG JE, VALDES AJ, TAYLOR JJ, WISE O'L, RODRIGUEZ E:

lysosomal extract, might also have influenced the effector phase of NCGN by prolonged attraction of neutrophils and monocytes and by prolonged production of toxic oxygen metabolites. Otherwise,

10. STILMANT MM, BOITON WK, STURGILL BG, SCHMITT GW, COUSER

important for binding, activation, and migration of inflammatory

WG: Crescentic glomerulonephritis without immune deposits: Chinicopathologic features. Kidney In! 15:184—195, 1979

hR might have induced up-regulation of adhesion molecules

Fatal immune complex glomerulonephritis without deposits. John Hopkins Mcdi 136:189—192, 1975

cells. In the present study we observed that renal hR leads to increased expression of ICAM-1 as compared to renal sections

11. RoNco P, VERROUST P, MIGNON F, KOURILSKY 0, VANI-IILLE PH, MEYRIUR A, MERY JPH, MOREL-MAROGER L: Immunopathological

from normal rats. This did not, however, not result in an influx of CD1 lb-positive inflammatory cells. Perfusion of lysosomal extract

report of 43 patients with 51 renal biopsies. Quart JMed 206:212—223,

studies of polyarteritis nodosa and Wegener's granulomatosis: A 1983

Brouwer et al: Renal ischemia in anti-myeloperoxidase GN

12. WEiss MA, CRISSMAN JD: Renal pathologic features of Wegener's granulomatosis: A review. Semin Resp Med 10: 141—I 48, 1989 13. ABRAI-IAMS SL, 1-TAZEN RJ, AYERS KM: Sodium azide protects against

ischemia-induced acute renal failure in rats. Am J Physiol 265:F130— F136, 1993 14. ARNOULD T, MIcmELs C, REMACLE J: Increased PMN adherence on endothelial cells after hypoxia: Involvement of PAF, CDI8!CD11b, and ICAM-I.Am JPhysiol 264:C1102—C11IO, 1993 15. PALLUY 0, MORLIERE L, GEls JC, BONNE C, MODAT G: Hypoxia!

reoxygenation stimulates endothelium to promote neutrophil adhesion. Free Radic Biol Mcd 13:21—30, 1992 16. WEYRICH AS, MA X, LEFER DJ, ALBERTINE KIT, LEFER AM: In vivo

neutralization of P-selectin protects feline heart and endothelium in myocardial ischemia and reperfusion injury. J Clin Invest 91:2620— 2629, 1993 17. YAMAZAKI T, SEKO Y, TAMATANI T, MIYASAKA M, YAGITA H, OKUMURA K, NAGAI R, YAZAKI Y: Expression of intercellular adhe-

sion molecule-I in rat heart with ischemia!reperfusion and limitation

of infarct size by treatment with antibodies against cell adhesion

1129

BAKER P, SCHULZE M, COUSER WG: Morphologic correlates of glomerular oxidant injury induced by the myeloperoxidase-hydrogen peroxide-halide system of the neutrophil. Lab Invest 5:294-301, 1988 29. PRYZWANSKY KB, MARTIN LE, SPITZNAGEL JK: Immunocytochemical

localization of myeloperoxidase, lactoferrin, lysozyme and neutral proteases in human monocytes and neutrophilic granulocytes. J Ret Endo Soc 24:295—310, 1978 30. Gooa. H VAN, HORST MLC VAN DER, FIDLER V, GROND J: Glomerular

macrophage modulation affects mesangial expansion in the rat after renal ablation. Lab Invest 66:564—571, 1992 31. LINAS SL, WHITFENBURG D, REPINE JE: 02 metabolites cause reper-

fusion injury after short but not prolonged renal ischemia. Am J Physiol 153:F685—F691, 1987 32. OITE T, BATSFORD SR, MIHATSCH Mi, TAKAMIYA H, VOGT A: Quantative studies of in situ immune complex glomerulonephritis in the rat induced by planted, cationized antigen. J Exp Med 155:460— 474, 1982 33. FISH Al, MICFIAELAF, VERNIER RL, GOOD RA: Acute serum sickness

nephritis in the rabbit. An immune deposit disease. Am J Pathol

molecules. Am J Pathol 143:410—418, 1993 18. SEEKAMP A, MULLIGAN MS, TILL GO, SMITH CW, MIYASAKA M, TAMATANI T, TODD RF III, WARD PA: Role of J32 integrins and

49:997—1022, 1966 34. FALK Ri, TERRELL RS, CHARLES LA, JENNETrE JC: Anti-neutrophil

ICAM-1 in lung injury following ischemia-reperfusion of rat hind

produce oxygen radicals in vitro. Proc Nati Acad Sci USA 87:4115— 4119, 1990

limbs. Am J Pathol 143:464—472, 1993 19. KUKIELKA GL, HAWKINS HK, MICHAEL L, MANNING AM, YOUKER K,

cytoplasmic autoantibodies induce neutrophils to degranulate and 35. MULDER AHL, HORST 0, LIMBURG PC, KALLENBERG CGM: Activa-

LE C, ENTMAN ML, SMITH CW, ANDERSON DC: Regulation of intercellular adhesion molecule-i (ICAM-i) in ischemic and reperfused canine myocardium. J Gun Invest 92:1504—1516, 1993

36. CHARLES LA, CALDAS MLR, FALK RJ, TERRELL RS, JENNETFE JC:

20. LINAS SL, WI-IITI'ENBURG D, PARSONS PE, REPINE JE: Mild renal

Antibodies against granule proteins activate neutrophils in vitro. J

ischemia activates primed neutrophils to cause acute renal failure.

Leuk Biol 50:539—546, 1991 37. YANG JJ, JENNETTE JC, FALK RJ: Immune complex glomerulonephri-

Kidney fnt 42:610—616, 1992 21. SEEKAMP A, WARREN JS, REMICK DG, TILL GO, WARD PA: Require-

ments for tumor necrosis factor-a and interleukin-1 in limb ischemia! reperfusion injury and associated lung injury. Am J Pathol 143:453— 463, 1993 22. BROUWER E, HUITEMA MG, KLOK PA, WEERD HDE, COHEN TERVAERT JW, WEENING JJ, KALLENBERG CGM: Anti-myeloperoxidase

associated proliferative glomerulonephritis; An animal model. J Exp Med 177:905—914, 1993

23. FIENBERG R: The protracted superficial phenomenon in pathergic (Wegener's) granulomatosis. Hum Pathol 12:458—467, 1981 24. FIENBERG R: Pathergic granulomatosis. Am J Med 19:829—831, 1955 25. MERRILL DP: Purification of human myeloperoxidase by concavalin A-sepharose affinity chromatography. Prep Biochem 10:133—150, 1980 26. KLEBANOFF Si, WALTERSDORPH AM, ROSEN H: Antimicrobial activity

of myeloperoxidase. Meth Enzymol 105:399—403, 1984 27. HOYER JR, MAUER SM, MICHAEL AF: Unilateral renal disease in the

tion of granulocytes by antineutrophil cytoplasmic antibodies is FcRdependent. Clin Exp Immunol 98:270—276, 1994

tis is induced in rats immunized with heterologous myeloperoxidase. C/in Exp Jmmunol 97:466—473, 1994 38. GOLBUS SM, WwsoN CB: Experimental glomerulonephritis induced by in situ formation of immune complexes in glomerular capillary wall. Kidney mt 16:148—157, 1979 39. JOHNSON RJ, KLEBANOFF Si, OCHI RF, ADLER 5, BAKER P, SPARKS L,

COUSER WG: Participation of the myeloperoxidase-H202-halide system in immune complex nephritis. Kidney mt 32:342—349, 1987 40. DIJKSTRA CD, DOpr EA, JOLING P, KRAAL G: The heterogeneity of

mononuclear phagocytes in lymphoid organs: Distinct macrophage subpopulations in the rat recognized by monoclonal antibodies EDI, ED2 and ED3. Immunology 54:589—599, 1985 41. KAMPINGA J, ASPINOLL R: Thymocyte differentiation and thymic microenvironment development in the fetal rat thymus: An immunological approach, in Thymus Update (vol. 3), edited by KENDALL M, RITFERS M, London, Harwood Academic Publishers, 1990, p 149

rat. I. Clinical, morphologic and glomerular mesiangial functional features of the experimental model produced by renal perfusion with

42, HUNIG T, WALLNY Hi, HARTLEY JK, LAWETZKY A, TIEFENTHALER G:

aminonucleoside. J Lab Gun Med 85:756—768, 1975 28. JOHNSON RJ, GUGGENHEIM SJ, KLEBANOFF SJ, Ocn RF, WASS A,

antigen receptor that induces T cell activation. J Exp Med 169:73—86, 1989

A monoclonal antibody to a constant dterminant of the rat T cell