Increased urinary excretion of C5b-9 distinguishes passive Heymann nephritis in the rat

Kidney International, Vol. 35 (1989), pp. 60—68

Increased urinary excretion of C5b-9 distinguishes passive Heymann nephritis in the rat MATTHIAS SCHULZE, PATRICIA J. BAKER, DIANA T. PERKINSON, RICHARD J. JOHNSON,

REX F. OcHI, ROLF A. K. STAHL, and WILLIAM G. COUSER Division of Nephrology, Department of Medicine, University of Washington, Seattle, Washington, USA

Increased urinary excretion of C5b-9 distinguishes passive Heymann nephritis in the rat. Increased urinary excretion of C5b-9 distinguishes passive Heymann nephritis from other forms of experimental glomerulonephritis in the rat. In the passive Heymann nephritis (PHN) model of membranous nephropathy (MN) subepithelial deposits form from anti-

Fx1A antibody reacting with antigen expressed on the glomerular epithelial cell membrane followed by membrane patching and shedding of immune complexes. Immune complex deposits are accompanied by deposits of C5b-9 which is required for the mediation of proteinuria. We tested the hypothesis that C5b-9 assembly on the epithelial cell mem-

in the Heymann rat models (2, 31. Recent preliminary observations on antibody eluted from the kidney of a patient with MN support this view [41. In the passive Heymann nephritis model, recent studies have established that glomerular immune deposit formation results from binding of an antibody to an antigen expressed in coated

pits on the glomerular epithelial cell membrane followed by patching and shedding of immune aggregates from the cell to brane might result in C5b-9 excretion in the urine, which would produce discontinuous subepithelial immune complex deposits distinguish this autoimmune mechanism of MN from other processes [5—81. Moreover, the alterations in glomerular permeability that result in subepithelial immune complex deposits. Using monoclo- which accompany this process appear to acquire assembly of nal antibodies developed to rat C6 and a rat C5b-9 neoantigen, in a sensitive ELISA assay, elevated urinary excretion of rat C5b-9 was the complement C5b-9 (membrane attack) complex [reviewed documented in PHN associated with on-going glomerular immune in 91, and C5b-9 deposits are prominent in both human and experimental MN as well as other glomerular diseases [9—121.

deposit formation. No urinary C5b-9 was detectable in MN induced by an exogenous antigen (cationized IgG) despite equivalent glomerular C5b-9 deposits, or in models of nephrotoxic nephritis, subendothelial immune complex nephritis, anti-mesangial cell membrane antibodyinduced nephritis or two non-immune nephropathies. Infusion of preformed C5b-9 in proteinuric animals excluded glomerular filtration of C5b-9 as a contributing mechanism to urinary C5b-9 excretion. We

With the above observations in mind, we reasoned that epithelial cell membrane insertion of C5b-9 and subsequent patching and capping phenomena might also result in increased

shedding of glomerular C5b-9 complexes into the urine. If

conclude that in the rat, increased urinary excretion of C5b-9 is a marker of MN induced by antibody to a glomerular epithelial cell antigen. Urine C5b-9 excretion reflects active glomerular immune deposit formation and distinguishes MN induced by this mechanism from other forms of MN as well as from other glomerular diseases with equivalent glomerular C5b-9 deposits.

present, urinary C5b-9 might serve to distinguish the anti-cell membrane antibody mechanism of MN from MN induced by other mechanisms not involving cell membrane antigens, as well as from other immune and non-immune glomerular diseases.

Using a monoclonal antibody apparently reactive with a neoantigen on rat C5b-9 in an ELISA assay, we found that

increased urinary C5b-9 excretion is present in MN induced by The diagnosis of membranous nephropathy (MN), the most antibody to rat glomerular epithelial cell antigen, and urinary common cause of idiopathic nephrotic syndrome in adults, can C5b-9 excretion accompanies on-going glomerular immune debe established only by renal biopsy [1]. The immune mecha- posit formation. Increased urinary CSb-9 excretion was not nisms which mediate MN in man are unknown. Therapy is detectable in other forms of complement-dependent or -independent experimental glomerular disease. These observations therefore largely empiric and guided primarily by measurements of the consequences of immune injury such as protein- may have significant implications for the diagnosis and therapy uria and renal function rather than by assessment of activity of of idiopathic MN in man.

the underlying immunopathogenetic mechanism. Because of the striking similarities between the clinical, histologic and immunopathologic findings in idiopathic MN and those in the Heyrnann models of MN in rats, most investigators strongly suspect that the disease mechanisms in man are similar to those

Methods

Preparation of rat C5b-9 complexes and monoclonal antibodies to rat C5b-9

Rat C5b-9 was prepared from rabbit erythrocytes lysed by normal rat serum as described by Perkinson et al [121 using a modification of the method of Biesecker et al [13]. Balb/c mice were immunized by priming with 400

Received for publication November 13, 1987 and in revised form May 16, 1988

g of the isolated C5b-9 in

complete Freund's adjuvant. Three further injections of 20 g, 20 g and 100 g rat C5b-9, respectively, were given during the

© 1989 by the International Society of Nephrology 60

Schulze et a!: Urinary C5b-9 in passive Heymann nephritis

61

next four weeks without adjuvant. Three days after the last injection the spleen cells were prepared and fused with NS- 1 nonproducer myeloma cells using the fusion protocol of Fazekas de St Groth and Scheidegg [14]. The fused cells were

order to avoid proteolytic degradation of urinary C5b-9 complexes by naturally occurring proteases in the urine, urine was

cloned by limiting dilution. The supernatants were screened for antibodies to rat C5b-9 by the presence of positive immunofluorescence staining on cryostat sections from kidney biopsies of rats with PHN, and negative staining on sections from rats with PHN previously complement depleted with cobra venom fac-

(EACA,Sigma), 20 mri EDTA and 100 kallikrein inhibitor units!

tor. Positive IgG clones were recioned [15], propagated and injected i.p. into pristane primed Balb/c mice. Monoclonal antibodies (mabs) purified from ascitic fluid by ammonium sulfate fractionation and DEAE-Sepharose (Pharmacia Fine Chemicals, Piscataway, New Jersey, USA) chromatography were biotinylated at I mg/mi in 100 m carbonate buffer (pH 7.7) using 50 g biotin-X-NHS (Caibiochem, La Jolla, California, USA) per mg of antibody [16].

Electrophoresis and immunoblotting To identify the individual complement components detected by the mabs, reactivity with rat EDTA-plasma and zymosan activated serum (ZAS) separated by sodium dodecyl sulfate gel electrophoresis (SDS—PAGE) was evaluated by immunobiotting. SDS-PAGE was done as described by Hempelmann [17]. Three microliter of EDTA-plasma or ZAS mixed with 7 d of sample buffer containing 0.5% SDS (wt/vol) and 50% glycerol (vol/vol) (Baker, Phillipsburg, New Jersey, USA) were separated by SDS-PAGE and then electrophoretically transferred to nitrocellulose membranes (Schleicher and Schull, Keene, New Hampshire, USA) using the method of Towbin, Staehlin and Gordon [18]. Remaining protein binding sites were blocked with

collected in preservative yielding a final concentration of 10 mM

benzamidine (Sigma), 10 m epsilon-aminocaproic acid ml of aprotinin (Sigma), 0.05% Tween 20 (Sigma) and 0.01% NaN3 in Tris-HC1 buffer, pH 8.6. C5b-9 in urines was stable under these conditions for over one month at 4°C.

ELISA for the detection of rat C5b-9 in plasma and urine The mab 2A 1 directed against a rat CSb-9 neoantigen was adjusted to 30 g/ml in 50 m carbonate buffer, pH 10.6. Three sg (100 jsl) was coated to wells of microtitration plates (Dynatech, Alexandria, Virginia, USA; Falcon, Becton-Dickinson, Oxnard, California, USA) by incubation at 4°C for 16 hours. Protein binding sites were blocked by incubating the wells with PBS-gelatin for one hour at room temperature. One hundred l of urine or plasma samples diluted in PBS-Tween containing the same protease inhibitors as used for the collection of urine were then placed in the wells at 4°C for 16 hours. The plates were washed twice with PBS-Tween and 400 ng/well of the biotinylated mab 3G 11 directed against rat C6 was added for two hours

at room temperature. The plates were washed three times with PBS-Tween and a 1/1000 dilution of horseradish peroxidase

coupled to streptavidin (Amersham) were added. After an incubation for one hour at room temperature, the wells were washed four times with PBS-Tween and 100 sl of solution

containing 0.2 mri 2.2'-azino-di-(3-ethyl-benzthiazolinesulfonate (ABTS, Boehringer Mannheim, Indianapolis, Indiana, USA) and 2.5 mM H2O2 in 100 m acetate buffer at pH 5.0 was placed in the wells. The color was allowed to develop for 30 PBS-gelatin for one hour. After washing with PBS either minutes at room temperature. The extinction at 405 nm was biotin-labeled or unlabeled polyclonal or mabs were incubated read using a Dynatech MR 560 microplate reader (Dynatech). with the membranes. The rat complement components C6 and The amount of CSb-9 in every urine was calculated using a C7 were identified using biotinylated monospecific goat IgG to linear calibration curve constructed with dilutions of standard human C6 and C7 prepared from antisera obtained commer- serum between 1/320 and the 1/5120. Urine samples were run in cially (Genzyme, Boston, Massachusetts, USA). Bound anti- four twofold serial dilutions starting with undiluted urine. Urine body was then detected after incubation of the membranes with concentrations of C5b-9 were calculated from OD values on the the appropriate second antibody (biotinylated sheep anti-mouse linear part of the calibration curve (normally between OD 0.6 IgG, Amersham, Arlington Heights, Illinois, USA) followed by and 0.05) and corrected for the urine dilution. The C5b-9 concentration in the lowest dilution of ZAS streptavidin-peroxidase (Amersham) and the peroxidase substandard which was significantly different from background strate 4-chloro-l-naphtole (Merck, Sharp and Dohme, Rahway, values using Student's t-test was taken as the detection limit of New Jersey, USA). Molecular weight calibration was done with prestained standards obtained commercially (Bethesda Re- the assay. The ability to recovery CSb-9 added to normal and proteinuric urines was determined by addition of zymosansearch Laboratories, Gaithersburg, Maryland, USA). activated serum standard to undiluted urines containing protease inhibitors yielding a concentration of 5 C5b-9 units/mi. After Preparation of a zymosan activated rat serum (ZAS) 30 minutes to 24 hours at room temperature the samples were reference standard diluted in the original Urines and C5b-9 was measured. The Rat serum was incubated at 37°C with 20 mg/mI zymosan amount of CSb-9 detected was divided by the amount added to (zymosan A, Sigma, St. Louis, Missouri, USA) for eight hours give percent recovery. with shaking and stored at —70°C. This serum was arbitrarily The minimum amount of C5b-9 that would have been detectassigned a value of 100 CSb-9 units/ml. Rat C5b-9 was found to able during a 24 hour period was determined in CSb-9 units/24 be stable upon repeated freezing and thawing of the ZAS. hrs using the following formula: Collection of rat plasma and urine samples for C5b-9 measurement Three hundred microliters of blood were drawn from the tail vein into syringes containing sufficient EDTA to result in a final concentration of 60 m, and plasma was frozen at —20°C. In

DL urine = DL ELISA x vol urine/% recovery x 100

where DL urine is detection limit (C5b-9 units/24 hrs), DL ELISA the detection limit of the ELISA (0.01 C5b-9 units/mi),

vol urine is urine volume (m1124 hrs), and % recovery is percentage of C5b-9 recovered after addition of CSb-9 to urine.

62

Schuize et a!: Urinary C5b-9 in passive Heymann nephritis

Inhibition of the rat C5b-9 assay with polyclonal antisera to human terminal complement components In order to identify the complement component on which the antigenic epitope detected by the 2A1 mab to a C5b-9 neoantigen is localized, a 1/160 dilution of the zymosan activated serum standard (0.625 C5b-9 units/mi) was preincubated with twofold serial dilutions of goat antisera to human C9, C7 (Genzyme) and normal goat serum starting at a 1/20 dilution. After one hour at room temperature the samples were added to ELISA plates coated with the 2A1 mab and the C5b-9 concentration was measured as described above. Results are expressed as C5b-9 detected/C5b-9 added x 100.

produced as described in detail elsewhere [27]. Briefly, the right kidney was perfused as described above with 180 jg of concanavalin A (Miles Scientific) in 0.4 ml PBS followed by 20 mg of rabbit anti-con A IgG [28]. The urine was collected during the subsequent 24 hour period, and renal biopsies were obtained 24 hours following antibody injection. e. Anti-mesangial cell antibody glomerulonephritis. To evaluate another form of glomerulonephntis induced by antibody to a resident, non-epithelial glomerular cell antigen, the anti-Thy 1 antibody-induced model of glomerulonephritis was produced in three rats [291. Thymocytes were prepared from 10-month-old

Lewis rats, and i08 cells were used to immunize rabbits three times in complete Freund's adjuvant. Six weeks after the initial

immunization, immune IgG was isolated by ion exchange chromatography, On cryostat sections of normal kidney, the induced in male Sprague Dawley rats (Tyler Laboratories, antibody reacted primarily with mesangial areas by indirect Induction of experimental glomerulonephritis General. All models of experimental glomerulonephritis were

Bellevue, Washington, USA). Urines were collected in metabolic cages with free access to water. Urine protein excretion was measured using a sulfosalicylic acid method and a whole serum standard (Lab Trol, Dade Diagnostics, Aquado, Puerto Rico, USA) [19]. Renal biopsies were obtained under ether anesthesia and processed for immunofluorescence microscopy as described below. a. Passive Heymann nephritis (PHN). PHN was induced in eight rats by i.v. injection of sheep antibody to Fx 1A prepared as described in detail elsewhere [20, 21]. Selected animals underwent renal biopsy, and urine and serum collections to provide data on days zero, one, two, three, four and seven. In most animals, urines were collected from days four to five. In another group of animals, renal biopsies were obtained daily from day one to five and stained for sheep IgG, rat C3 and rat C5b-9. The effect of complement depletion on glomerular C5b-9 deposition and urinary C5b-9 excretion was studied in two PHN rats depleted of complement by administration of cobra venom factor (Cordis Laboratories, Miami, Florida, USA) prior to

anti-Fx lA injection, and daily thereafter using a protocol described previously [22]. b. Cationic IgG induced membranous nephropathy. To study

a model of membranous nephropathy induced by a planted exogenous antigen, human IgG (Miles Scientific, Naperville, Illinois, USA) was cationized as described by Oite et al [23] using a modification of carbodiimide method of Danon et al [24]. Isoelectric points of cationized IgG bands ranged from 8.1 to 9.3

by isoelectric focusing. After removal of the left kidney, the cationized antigen (125 p.g in 500 pA PBS) was perfused ex vivo

immunofluorescence. It stained glomerular mesangial cells in culture and failed to stain glomerular epithelial cells in culture [30]. Mesangial glomerulonephritis was induced by injection of 10, 15 or 20 mg of anti-Thy 1 antibody i.v. Urine was collected from 0 to 24 hours and renal biopsies were obtained 24 hours after antibody injection. f. Aminonucleoside nephrosis (PAN). PAN was induced in six rats by a single intravenous injection of 0.1 mg/kg of PAN (ICN Pharmaceuticals, Inc., Cleveland, Ohio, USA) [31]. Urine protein excretion was measured from days 3 to 4, 10 to 11 and 16 to 17. To study the possible glomerular filtration of circulating C5b-9 complexes as a contributing factor to urinary C5b-9 excretion, five proteinuric PAN rats (113 38 mg/day) were injected with 100 units of C5b-9 as ZAS, and measurements of plasma C5b-9 were carried out before, one minute and 24 hours after injection. Urinary C5b-9 measurements were made in the urine collected for the 24 hours after C5b-9 injection. g. Remnant kidney. A remnant kidney model was induced in five rats by left nephrectomy and subsequent ligation of several branches of the right renal artery until 2/3 to 5/6 of the right kidney was unperfused. Following renal ablation, urines were collected from day seven to eight and 10 to 11. Renal pathology

Fresh renal tissue obtained at open renal biopsy was processed for immunofluorescence microscopy as described previously [32]. Immunofluorescence to detect C5b-9 with biotinylated mab 2A 1 was done by adding 600 ng of antibody in a volume of 20 pA to tissue sections which had been prepared as described above. After an incubation at room temperature for one hour, the slides were washed three times with PBS and 20 p.l of a 1:200 dilution of fluorescein-labeled streptavidin (Amersham Corporation) were added and incubated for 30 minutes at room temperature. Rat C3 and sheep IgG were detected using fluorescein labeled antibodies (goat anti-rat C3, rabbit antisheep IgG, Organon Teknika-Cappel, Westchester, Pennsylvania, USA). The intensity of immunofluorescence was graded

via the superior mesenteric artery into the right kidney of five experimental rats using a perfusion protocol described in detail elsewhere [25]. Thirty minutes after perfusion, 0.9 ml of heatinactivated (56°C, 30 mm) rabbit anti-human IgG antiserum or normal rabbit serum was given i.v. Urine protein excretion was measured from 0 to 24 hours, and renal biopsies were obtained 24 hours after antibody injection. c. Nephrotoxic nephritis. Nephrotoxic (anti-GBM) nephritis was induced in six rats by a single injection of sheep anti-rat GBM antibody, the characteristics of which have been published elsewhere [26]. Urines were collected from 0 to 24 hours, semiquantitatively as follows: trace, barely detectable; 1+, and renal biopsies obtained 24 hours after antibody injection. d. Subendothelial immune complex nephritis. Nephritis due present but faint; 2+, average staining intensity; 3+, strong to subendothelial formation of immune complex deposits was standing intensity; 4+, maximal staining intensity.

63

Schulze et a!: Urinary C5b-9 in passive Heymann nephritis

A

molecular

B C D E F G H weight (kD)

140

Haoo

a —.

-97.4 120

-43 100 —25.7

-18.4 —14.3

>a,

> 0 80 C) a'

Fig. 1. Characterization of monoclonal antibodies to the rat C5b-9 complex. Rat EDTA-plasma (lanes A, B, E, G, H) and zymosan-

activated rat serum standard (lanes C, D, F) were separated on

SDS-PAGE. After transfer to nitrocellulose membranes the lanes were reacted with the following antibodies: A, polyclonal goat anti-human C6; B, C: monoclonal antibody 3G1 1; D, E: monoclonal antibody 2A1; F, G: monoclonal antibody 2H8; H: polyclonal goat anti-human C7. 3G1 I identifies the same protein in EDTA-plasma as antibody to human C6 (lanes A, B) whereas 2H8 identifies rat C7 which is also detected by goat anti-human C7 (lanes G, H). Reacted against ZAS, both monoclonal antibody 3G1 1 and 2H8 also identify a second high molecular weight band (lanes C, F) which presumably represents the C5b-9 complex of rat complement. Monoclonal antibody 2Al reacts intensely with this high molecular weight complex in ZAS but fails to sustain any native complement components in EDTA plasma (lane E), suggesting reactivity with a neoantigen of the C5b-9 complex.

Results

Specificity of the monoclonal antibodies for rat C5b-9 and establishment of the rat C5b-9 ELISA Three monoclonal antibodies (2H8, 2A1, 3011) obtained after screening of the hybridoma cultures were further characterized using SDS-PAGE and immunoblotting. The mab 3G11 identified the same protein in EDTA-plasma as did the antibody to human C6 (Fig. 1, lanes A, B) whereas the mab 2118 identified C7 as seen by comparison with a polyclonal antibody to human

0

0 60

40

20

1/20

1/80

1/320

1/1280

Dilution of antiserum

Fig. 2. Inhibition of the rat C5b-9 ELISA by antisera to human complement components. Rat ZAS was preincubated with polyclonal

goat antiserum to human C9 (circles), goat antiserum to human C7 (squares) and normal goat serum (triangles). Antiserum to human C9 inhibited the detection of rat C5b-9 complexes indicating that the epitope detected by the mab 2Al is located on rat C9.

C7 (Fig. I, lanes 0, H). When tested against ZAS both the anti-rat C6 (mab 3G1 1) and anti-rat C7 (mab 2H8) antibodies also identified a second high molecular weight band (Fig. 1, lanes C, F) which presumably represents the C5b-9 complex of rat complement. The mab 2A1 intensely reacted with this high molecular weight component in ZAS but failed to stain EDTAplasma (Fig. 1, lane D, E), suggesting reactivity with a neoantigen of the C5b-9 complex which is not present on any of the precursor components. Inhibition of the C5b-9 ELISA by anti-human C9 In order to identify the complement component detected by

the 2A1 mab rat zymosan activated serum was preincubated with polyclonal goat antibodies to human C7, C9 and normal goat serum. As indicated in Figure 2, only preincubation with anti-human C9 inhibited the assay in a dose dependent manner. The mab 2A1 mab therefore seems to compete with antibodies

Characteristics of the C5b-9 ELISA

A calibration curve obtained by assaying twofold serial dilutions of zymosan-activated rat serum is depicted in Figure 3 with EDTA-plasma as negative control. The detection limit was 0.01 C5b-9 units/ml. The assay also detected detergent solubi-

lized C5b-9 complexes obtained from rat complement lysed erythrocytes (data not shown). Detection of rat C5b-9 in urine and plasma

When zymosan activated serum was added to normal rat urine (N = 4), 104.9 15.0% (mean SD) of the added C5b-9 was recovered when the assay was performed immediately and 93.2 8.0% was recovered when the assay was performed after

24 hour incubation at room temperature. Less C5b-9 was

recovered when C5b-9 was added to proteinuric urines which did not initially contain measurable CSb-9. Recovery rates of indicating that the reactive epitope is located on the C9 mole- added C5b-9 in these urines were used to calculate the 24-hour cule. detection limits indicated in Table 2.

to human C9 for its binding site on the C5b-9 complex,

64

Schuize et a!: Urinary C5b-9 in passive Heymann nephritis

1.2

1.0 +1

0.8 to

0 d

0

0.6 0.4 0.2

T'

1/40

1/160

*

.,.—

'—' •—--1/2560 1/10240

•':'-

1/640

1/40960

Plasma or serum dilution

Fig. 3. Calibration curve of the rat C5b-9 ELJSA. Four separate samples of rat EDTAplasma (t) or ZAS (•) were run in twofold serial dilutions. Values obtained at 1: 10,000 dilution of ZAS were significantly different from background values (P < 0.01).

Table 1. Glomerular deposition and urinary cxcretion of rat C5b-9 in passive Heymann nephritis Days after anti-FxIA antibody injection Glomerular immunofluorescence Sheep IgG (N = 4)

0

1

2

3

4

5

ND

2-3+ 1+

2-3+

2-4+ 3-4+

2-4+

ND ND

2-4+ 2-3+

2-4+

Rat C3 (N = 4) Rat C5b-9 (N = 4)

0—T+

1—2+

1—3+

2—3+

2—3+

Urine C5b-9 excretion units/24 hr Plasma C5b-9 units/mi Urine protein mg/24 hr a Not done b Mean SD

0

0.03

1.2

Q•3b

5.0

4.2

2l

4.5

0,04b

0.6

0.3

3.0

2.0

2.2

1.0

1.6 1.3

2.8 3.2

0.4 0.3

2.2 9.8

0.6 6.0

1.9

0.5

20.7

15.2

3—4+

3.6

3.3

1.1

0,2 44.6

57.4

Excretion of C5b-9 in urines from rats with passive Hey,nann during the study (Fig. 4C). Treatment of two PHN rats with CVF abolished glomerular C5b-9 and C3 deposition as well as nephritis (PHN) Proteinuria greater than 10 mg/24 hr was induced in PHN rats urinary C5b-9 excretion (Fig. 4B).

on days 3 to 4 after the injection of anti-FxLA (Table 1). However, all rats had detectable levels of urine C5b-9 excretion Plasma levels and glomerular filtration of C5b-9 averaging 0.6 0.5 units/24 hr on the second day after injection of anti-Fx1A. This level increased to 3.0 2.0 units/24 hr on the Plasma levels of CSb-9 were 5.0 units/mi on the first day and third day (P < 0.001 vs. day two) and was 3.6 3.5 units/24 hr decreased to levels seen before the injection of anti-Fx1A over on the fifth day (P < 0.01 vs. day two; Table 1). No excretion the next four days (Table 1). Urinary C5b-9 excretion did not of C5b-9 was detectable in PHN rats depleted of complement by

treatment with cobra venom factor. With the onset of the parallel plasma levels of C5b-9. Because PHN rats were not autologous phase and glomerular deposition of rat anti-sheep proteinuric when C5b-9 levels in plasma were highest, we 6.6 Units on

addressed the question whether elevated plasma levels of CSb-9 could contribute to urinary C5b-9 excretion in proteinuric rats by inducing elevated serum C5b-9 levels in rats already made

Glomerular immunofluorescence Glomerular deposits of sheep IgG were present on the first

proteinuric by prior administration of PAN. Five rats with

deposition of CSb-9 became detectable on the second day

PAN, were injected i.v. with ZAS containing 100 units of

IgG, urinary C5b-9 excretion increased to 8.7 days 7 to 8.

aminonucleoside nephrosis that had heavy proteinuria but no day after injection of anti-FxJA (Table 1). One, 2 + glomerular detectable C5b-9 in the urine 17 days after administration of (Table 1). C5b-9 increased to 3 + on the third day and thereafter C5b-9. C5b-9 plasma levels at one minute were increased to (Table 1, Fig. 4A). Glomerular staining for C3 paralleled stain- 14.1 units/mI, or about three times the maximal levels seen early ing for C5b-9. Rat IgG was not detected in glomeruli until day in PHN (Table 1) and normal values by the end of 24 hours. five. Careful studies failed to detect sheep IgG, rat C3 or rat However, C5b-9 was not detected in the urine of injected rats C5b-9 on the brush border of proximal tubular cells at any point despite a urinary protein excretion of 153 27 mg124 hr.

65

Schuize et a!: Urinary C5b-9 in passive Heymann nephritis

'

Fig. 4. Iu,,,,unofluorccent staining for glomerular deposits of CSb-9 in several models of experimental glo,nerulonephritLc studied. A. Passive Heymann nephritis at five days: B. Passive Heymann nephritis treated with cobra venom factor at five days: C. Proximal tubules in proteinuric rats with passive Heymann nephritis at five days: D. Cationized lgG-anti-lgG model of membranous nephropathy on day one; E: Con A-anti-con A model of subendothelial immune complex glomerulonephritis on day two: F. Glomerulonephritis induced with anti-thy I on day one: C. Normal rat kidney (original magnifications 25O).

Table 2. Urinary C5b-9 excretion in experimental glomerular disease Urinary excretion

Passive Heymann nephritis Cationized IgG anti-IgG Nephrotoxic nephritis Con-A anti-con-A Anti-thy 1 Aminonucleoside nephrosis

Remnant kidney

Glomerular C5b-9 deposits

Protein mg/24 h

u/24 hr

53 l6

3.6

C5b-9

N

Day

8

5

+

subepithelial

5

2

+

subepithelial

165 40

<0.71

0.19"

6

1

linear GBM

423

60

<0.43

0.14"

88

12

<0.82 <0.15

0.24" 0.02"

38

<0.24 0.03"

1

<0.17

0.03"

4

2 2

+ +

6

17

—

5

10

—

3

subendothelial mesangial

4 170

8

I

1.2k

a Mean SEM. b CSb-9 excretion was not detected in this group. The value reported represents and indicates the mean SEM of the maximal amount of C5b-9

that could have been excreted in 24 hrs without being detected by the assay in this group (Methods).

Urinary C5b-9 excretion in other forms of experimental glomerulonephritis

A-anti-con A) and mesangial deposits (anti-Thy 1) also had no detectable urinary CSb-9 despite readily-detectable glomerular C5b-9 deposits (Table 2, Figs. 4E and 4F). Models of immune

Excretion of C5b-9 was not detectable in membranous and non-immune glomerular injury which were not associated nephropathy induced by an exogenous antigen (cationized IgG) with glomerular C5b-9 deposits, including nephrotoxic nephridespite levels of proteinuria and glomerular C5b-9 deposits tis, aminonucleoside necrosis and the remnant kidney model, which exceeded those seen in PHN (Table 2, Fig. 4D). More- also failed to exhibit detectable C5b-9 excretion in urine (Table over, nephritis associated with subendothelial deposits (con 2).

66

Schuize et a!: Urinary C5b-9 in passive Heymann nephritis

Discussion

peak levels in PHN by threefold were produced. However, no detectable urinary C5b-9 excretion occurred. Thus filtration of

C5b-9 membrane attack complexes are assembled from five

the intact C5b-9 complex, even by abnormally permeable

precursor molecules present in the serum in inactive form. Proteolytic cleavage of C5 by the C5 convertase of either the classical or alternative complement pathway generates C5b which initiates assembly of the C5b-9 complex [33]. The last

glomeruli, appears to be minimal and insufficient to account for our findings. A post-glomerular origin of the urinary complexes could be postulated based on the extensive reactivity of the anti-Fx1A

antibody used to induce PHN with antigens expressed on which accompanies insertion of the complex into the lipid proximal tubular brush border as well as glomerular antigens step of C5b-9 complex formation involves polymerization of C9

bilayer of cell membranes [34]. When complement is activated in serum in the absence of cell membranes, C9 polymerization is inhibited by the action of S protein resulting in water soluble complexes [33]. During formation of C5b-8 and C9 polymerization, neoantigens are expressed which are unique to the C5b-9 complex and are not expressed on any of the individual native complement components. Antibodies to such neoantigens have been utilized to demonstrate C5b-9 deposition in a variety of

[20, 21], and the known presence of the putative antigen in this model, GP330, on both glomerular and tubular epithelial cells [6]. However, increased urinary C5b-9 excretion preceded any

body to a C5b-9 neoantigen by SDS-PAGE and immunoblotting studies and was utilized as the capturing antibody in our ELISA

lowed the opposite course. However, other studies in models of

detectable increase in urine protein excretion in PHN, and careful immunofluorescence studies failed to demonstrate tubular brush border deposition of C5b-9 at any point in the course

of disease development. Previous studies have noted faintly positive staining for IgG on proximal tubular brush borders at human and experimental glomerular diseases [10—12]. Our the time of peak circulating antibody levels early on day one, monoclonal antibody 2Al exhibited the characteristics of anti- which usually diminish thereafter [211. C5b-9 excretion folactive Heymann nephritis 1361 which may be mediated by a assay. The bound C5b-9 was then detected using a second similar mechanism [37], and studies of anti-brush border antibodies passively transferred to rats with proteinuria secondary to chronic serum sickness, have noted that IgG, C3 and C5b-9 measure C5b-9 in serum [35], exhibited excellent sensitivity and were components of tubular brush border deposits [7, 38, 39], specificity for detecting rat C5b-9 in serum and urine. although tubular injury is apparently complement independent. Our data document an easily detectable increase in urinary Although we cannot exclude other sources with certainty, we excretion of the C5b-9 complex concomitant with the onset of believe our data, and the studies of others, strongly support the glomerular immune deposit formation in the passive Heymann conclusion that the urinary C5b-9 complexes measured did nephritis (PHN) model of MN in rats. Elevated urinary C5b-9 result from glomerular immune deposit formation. Our data do not define the mechanism by which glomerular excretion preceded the onset of detectable increases in total urinary protein excretion. It parallels the course of glomerular C5b-9 complexes reach the urinary space. Our studies in immune deposit formation as defined by immunofluorescence aminonucleoside nephrosis make it improbable that they do so studies and previous quantitative measurements of glomerular by crossing any component of the glomerular filtration barrier. antibody deposition which document on-going deposit forma- Other possibilities exist. Filtration of antibody IgG may result tion for several days following a single antibody injection [21]. in initial deposit formation in coated pits containing GP330 on Urinary excretion of C5b-9 was present in detectable quantities the urinary space surface of the glomerular podocyte [6]. C5b-9 only in MN induced by an antibody to a glomerular epithelial might be transported into the urine directly or by rearrangement cell membrane antigen and was absent or below detection limits of the cell membrane during the patching and capping process with subepithelial immune complex formation induced by an induced by binding of antibody, as observed in cultured gbexogenous antigen, immune complex formation at other sites, merular epithelial cells [8]. Finally, cellular endocytosis of the or deposition of antibody against mesangial cell membranes immune aggregates containing C5b-9, or of the CSb-9 complex despite extensive glomerular C5b-9 deposits in most of these alone, with transcellular transport and subsequent exocytosis lesions. The levels of urinary C5b-9 excreted in PHN exceeded into the urinary space may be possible. Preliminary immunothe sensitivity levels of the assay in these models by 4- to cytochemical localization studies of the C5b-9 complex and monoclonal antibody reactive with rat C6. This double antibody

ELISA assay, similar in principle to assays used by others to

20-fold.

sequential studies of its movement suggest that the latter

mechanism probably accounts for our findings in PHN [40]. This phenomenon of endocytosis and exocytosis of the CSb-9 complexes by nucleated cells would accord with recent demonstration of a similar phenomenon in neutrophils [41], as well increase in circulating serum C5b-9 complexes was observed on as with the observation that C5b-9 on the glomerular epithelial days 1 to 2 following injection of anti-FxlA antibody to induce cell in vitro appears to be inserted into the cell membrane and PHN, urinary complexes were barely detectable at this point handled by the cell differently from the antigen antibody comand rose to peak levels only on days 3 to 4 when serum C5b-9 plexes which are shed from the cell surface [421. A somewhat surprising finding in our study was that detectlevels had returned to levels seen before the injection of antibody. When soluble C5b-9 complexes, which have a molec- able levels of urinary C5b-9 excretion were seen only in PHN ular weight of about 1000 kd [33], were exogenously adminis- and were not found in other forms of MN or other glomerular tered to rats with a marked and non-selective increase in immune complex diseases despite extensive glomerular C5b-9 glomerular permeability induced by prior injection of aminonu- deposition in other models. MN induced by in situ immune cleoside of puromycin, serum levels of C5b-9 that exceeded complex formation with a cationized exogenous IgG antigen While there is no obvious way to definitively establish that the urinary C5b-9 complexes measured were derived from glomeruli, there are several reasons to believe that they were not of pre- or post-glomerular origin. Thus, although a transient

67

Schu/ze et al: Urinary C5b-9 in passive Heymann nephritis

shares with PHN the characteristics of subepithelial immune complex deposits, glomerular C5b-9 deposition, and heavy proteinuria probably due to a C5b-9 dependent mechanism of injury [43, 44]. However, urinary C5b-9 excretion was below detection limits in this model despite glomerular immune deposits and proteinuria which equalled or exceeded those in PHN. Our data therefore suggest that only immune deposits formed on the glomerular epithelial cell membrane itself, presumably accompanied by cell membrane insertion of C5b-9, result in urinary C5b-9 excretion whereas immune complex deposits formed in a subepithelial distribution by planted exogenous antigens localized on a charge basis and apparently not involving membrane C5b-9 insertion do not cause detectable urinary C5b-9 levels. Similar negative results were found with nephritis induced by subendothelial immune complex deposits

of concanavalin A and anti-concanavalin A, a neutrophil-

mediated model of glomerulonephritis which also has prominent C5b-9 deposits [27, 28]. Induction of glomerulonephritis due to formation of immune deposits on the mesangial cell

membrane resulted in glomerular C5b-9 deposition but no urinary C5b-9 excretion, presumably reflecting interposition of mesangial matrix between these cells and the urinary space. Yamamoto and Wilson have recently characterized this model and suggested that immune injury may be C5b-9 mediated [29]. The lack of urinary C5b-9 excretion in complement-indepen-

dent anti-GBM nephritis and in the non-immune lesions of aminonucleoside necrosis and renal ablation nephropathy further support the specificity of elevated urinary C5b-9 excretion for lesions induced by antibody to the glomerular epithelial cell membrane. While the availability of multiple models of glomerular dis-

ease in the rat which exhibit immunopathologic features of

Ochi. Dr. Stahl was the recipient of a fellowship from the Boehringer Ingelheim Funds, Stuttgart, FRG. The authors are grateful to Ms. Pam Pritzl and Ms. Caryl Campbell for technical assistance and to Ms. Maggie Whitcomb for typing the manuscript. Reprint requests to William G. Couser, M.D., Division of Nephrology, RM-11, University of Washington, Seattle, Washington, 98195, USA.

References 1. COUSER WG: Glomerular diseases, in Textbook of Medicine. A Systemic Approach, edited by STEIN J, Boston, Little Brown and Company, 1986, pp. 834—861 2. COUSER WG: Mechanisms of glomerular injury in immune complex glomerulonephritis (Nephrology Forum). Kidney mt 28:569—583, 1985

3. NEALE TJ, WILSON CB: Glomerular antigens and glomerulonephritis. Springer Semin Immunopathol 5:221—249, 1982 4. COLLINS B, BAIRD L, ERIKs0N M, BRADFORD D, PAN G, HSIUNG CK, SCHEEBERGER E, BHAN A, MCCLUSKEY R: Antibodies reac-

tive with a renal glycoprotein and with deposits in membranous nephritis. (abstract) Kidney mt 3 1:338, 1987 5. KERJASCHKI D, FARQUHAR MG: The pathogenic antigen of Hey-

mann nephritis is a membrane glycoprotein of the renal proximal tubule brush border. Proc Nail Acad Sci USA 79:5557—5561, 1982 6. KERJASCHKI D, FARQUHAR MG: Immunocytochemical localization

of the Heymann nephritis antigen (GP 330) in glomerular epithelial cells of normal Lewis rats. J Exp Med 157:667—685, 1983 7. BIESECKER G, NOBLE B, ANDRES GA, KOFFLER D: Immunopatho-

genesis of Heymann's nephritis. C/in Immuno/ Immunopathol 33: 333—338, 1984 8. CAMuss! G, BRENTJENS JR, NOBLE B, KERJASCHKI D, MALAVASI

F, ROHOLD OA, FARQUHAR MG, ANDRES G: Antibody-induced

redistribution of Heymann's antigen on the surface of cultured glomerular visceral epithelial cells: Possible role in the pathogenesis of Heymann glomerulonephritis. J Immunol 135:2409—2416, 1985

immune mediated renal disease in man allows easy definition of 9. Cousit WG, BAKER PJ, ADLER S: Complement and the direct mediation of immune glomerular injury: A new perspective. Kidney the specificity of elevated urinary C5b-9 levels for various types Int 29:879—890, 1985 and mechanisms of glomerular disease, the utility of a similar 10. FALK RJ, SissoN SP, DALMASSO AP, KIM Y, MICHAEL AF, assay in man will obviously require independent confirmation VERNIER RL: Ultrastructural localization of the membrane attack

using an assay for human C5b-9 that is functional in human urine. The pathogenesis of MN in man remains unknown. If

complex of complement in human renal tissue. Am J Kidney Dis 9:

idiopathic MN is an autoimmune disease involving antibody to glomerular epithelial cell membranes, as preliminary data sug-

11. HINGLAIs N, KAZATCHKINE MD, BHAKDI S, APPAY M-D, MANDET C, GROSSETETE J, BARIETY J: Immunohistochemical study of

gest [4], the nephritogenic antibody does not appear to be

121—128, 1987

the C5b-9 complex of complement in human kidneys. Kidney mt 30:399—410, 1986

directed against an antigen shared with proximal tubular brush 12. PERKINSON DT, BAKER PJ, COUSER WG, JOHNSON RJ, ADLER S: Membrane attack complex deposition in experimental glomerular border as it is in the Heymann rat model [45, 46]. Nevertheless, injury. Am J Pathol 120:121—128, 1985 our findings suggest that measurement of urinary C5b-9 com- 13. BIESECKER G, PODAK ER, HALVERSON CA, MULLER-EBERHARD plex excretion might provide a useful non-invasive diagnostic HJ: C5b-9 dimer: Isolation from complement lysed cells and test for such an entity. Moreover, the availability of a readily ultrastructural identification with complement lysed membrane lesions. J Exp Med 149:448—458, 1979 measured parameter which provides an index of the activity of the disease mechanism against which therapy is directed could 14. FAZEKA5 DE ST GROTH S. SCHEIDEGG D: Production of monoclonal antibodies: Strategy and tactics. J Immunol Meth 32:297—304, be of major clinical significance. The findings in the present 1980 study provide substantial impetus for the pursuit of such a goal. 15. Civir' CI, BANQUENGO ML: Rapid, efficient cloning of murine hybridoma cells in low gelation temperature agarose. J Immunol

Acknowledgments Portions of this work were presented at the 19th Meeting of the American Society of Nephrology, Washington, D.C., 1986 and have been published in abstract form (Kidney In! 31:329, 1987).

Portions of this work were supported by the United States Public Health Service research grants DK 34198, AM 32051, DK 39068, DK 07467 and a research grant from the Northwest Kidney Foundation. Portions of this work were completed during the tenure of National Kidney Foundation Research Fellowship Awards to Drs. Johnson and

Meth6l:l—8, 1983 16. BIEBER F: Biotinylierung monoklonaler Antikorper, in Monoklo-

nale Antikorper. Herstellung und Charakterisierung, edited by PETERS H, Berlin, Springer, 1985, pp. 209—212

17. HEMPELMANN E: Bilden und Auflosen von Proteinstapeln, in Electrophorese Forum, edited by RADOLA, Weinheim, Verlag Chemie, 1982, pp. 111—1 16

18. T0wBIN H, STAEHLIN T, GORDON J: Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: Procedure and some applications. Proc Nat! Acad Sci USA 76:4350— 4354, 1979

68

Schulze et at: Urinary C5b-9 in passive Heymann nephritis

19. BRADLEY GM, BENSON ES: Examination of the urine, in ToddSanford Clinical Diagnosis by Laboratory Methods, edited by DAVIDSON I, HENRY JB, Philadelphia, WB Saunders Company, 1974, p. 74 20. COUSER WG, STEINMULLER DR. STILMANT MM, SALANT DJ, LOWENSTEIN LM: Experimental glomerulonephritis in the isolated perfused rat kidney. J Clin Invest 62:1275—1287, 1978 21. SALANT DJ, DARBY C, COUSER WG: Experimental membranous glomerulonephritis in rats: Quantitative studies of glomerular immune deposit formation in isolated glomeruli and whole animals. J Clin Invest 66:71—81, 1980 22. SALANT DJ, BELOK S, MADAIO MP, COUSER WG: A new role for

complement in experimental membranous nephropathy in rats. J Clin Invest 66:1339—1350, 1980 23. OITE TM, BATSFORD SR, MIHATSCH Mi, TAKAMIYA H, VOGT A:

Quantitative studies of the in situ immune complex glomerulonephritis in the rat induced by planted cationized antigen. J Exp Med 155:460-474, 1982 24. DANON D, GOLDSTEIN L, MARIKOVSKY Y, SKULESKY E: Use of

cationized ferritin as a label of negative charges on cell surfaces. J Ultrastruct Res 38:500—510, 1972

complement components in experimental immune glomerular injury. Kidney In! 26:830—837, 1984 33. MULLER-EBERHARD Hi: The membrane attack complex of complement. Ann Rev Immunol 4:503—528, 1986 34. KOLB WP, MULLER-EBERHARD HJ: Mode of action of human C9:

Adsorption of multiple C9 molecules to cell-bound C8. J Immunol 113:479—488, 1974

35. GAWRYL MS, SIMON MT, EATMAN JL, LINT TF: An enzymelinked immunoabsorbent assay for the quantitation of the terminal complement complex from cell membranes or in activated human sera. J Immunol Meth 9:217—225, 1986 36. MENDRICK D, NOBLE B, BRENTJENS JR. ANDRES GA: Antibody

mediated injury to proximal tubules in Heymann nephritis. Kidney In! 18:328—343, 1980

37. NEALE Ti, COUSER WG, SALANT DJ, LOWENSTEIN LM, WILSON

CB: Specific uptake of Heymann's nephritic kidney eluate by rat kidney, Studies in vivo and in isolated perfused kidneys. Lab Invest

46:450—453, 1982 38. NOBLE B, MENDRICK D, BRENTJENS JR, ANDRES GA: Antibody-

mediated injury to proximal tubules in the rat kidney induced by

passive transfer of homologous anti-brush border serum. Clin Immunol Immunopathol 19:289-301, 1981 39. NOBLE B, ANDRE5 GA, BRENTJENS iR: Passively transferred anti-brush border antibodies induce injury of proximal tubules in the absence of complement. Cliii Exp Immunot 56:281—288, 1983

25. JOHNSON RI, COUSER WG, Cm EY, ADLER S, KLEBANOFF SJ: New mechanism for glomerular injury. Myeloperoxidase-hydrogen peroxide-halide system. J Clin Invest 79:1379—1387, 1987 26. COUSER WG, DARBY C, SALANT DJ, ADLER S, STILMANT MM, LOWENSTEIN LM: Anti-GBM antibody-induced proteinuria in isolated perfused rat kidney. Am J Physiot 249:241—250, 1985 27. JOHNSON Ri, KLEBANOFF SJ, Oci-n RF, ADLER 5, BAKER P,

40. KERJASCHKI D, SCHULZE M, BINDERS, COUSER WG: Localization

SPARKS L, COUSER WG: Participation of the myeloperoxidaseH202-halide system in immune complex nephritis. Kidney mt 32:

trophils from complement attack: Removal of membrane attack

342—349, 1987

28. GOLBUS SM, WILSON CB: Experimental glomerulonephritis induced by in situ formation of immune complexes in glomerular capillary wall. Kidney mt 16:148—157, 1979

and transport by C5b-9 by the glomerular epithelial cell in experimental membranous nephropathy. (abstract) Kidney Int 33:319, 1988

41. MORGAN BP, DANKERT JR, ESSER AF: Recovery of human neu-

complex by endocytosis and exocytosis. J Immunol 138:246—253, 1987 42. CAMUSSI C, SALVIDIO G, BIESECKER G, BRENTJENS i, ANDRES G:

Heymann antibodies induce complement-dependent injury of rat glomerular visceral epithelial cells. J Immuno! 139:2906—2914, 1987

29. YAMAMOTO T, WILSON CB: Complement dependence of antibodyinduced mesangial cell injury in the rat. J Immunol 138:3758—3765, 1987 30. ADLER S, BAKER PJ, JOHNSON Ri, 0cm RF, PRITZL P. COUSER WG: Complement membrane attack complex stimulates production

43. VOGT A, SCHMIDT HU, TAKAMIYA H, BATFORD 5: In situ immune

of reactive oxygen metabolites by cultured rat mesangial cells. J

45. ZAGER RA, COUSER WG, ANDRES LBS, BALTON WK, POHL MN:

Gun Invest 77:762—767, 1986 31. COUSER WG, SALANT DJ, STILMANT MM, ARBEIT LA, DARBY C,

SLIOGERIS VG: The effects of aminonucleoside of puromycin and nephrotoxic serum on subepithelial immune-deposit formation in passive Heymann nephritis. J Lab Clin Med 94:917—932, 1979 32. ADLER S, BAKER PJ, PRITZL P. COUSER WG: Detection of terminal

complex nephritis and basic proteins. Proc EDTA 17:613—620, 1980 44. 0cii RF, JOHNSON RI, BAKER Pi, ADLER 5, COUSER WG: C6

depletion diminishes proteinuria in experimental nephropathy (MN) induced by exogenous antigen. Kidney mt 29:287, 1986 Radioimmunologic search for anti-tubular epithelial antibodies and

circulating immune complexes in patients with membranous nephropathy. Nephron 24:10-16, 1979 46. COLLENS AD, ANDRES GA, MCCLUSKEY RT: Lack of evidence for

a role of renal tubular antigen in human membranous glomerulonephritis. Nephron 27:297—301, 1981