Kidney International, Vol. 45 (1994), pp. 352—359

Role of oxidants and proteases in glomerular injury

RICHARD J. JOHNSON, DAVID LOVETT, ROBERT I. LEHRER, WILLIAM G. COUSER,and SEYMOUR J. KLEBANOFF

Divisions of Nephrology and Infectious Diseases, Department of Medicine, University of Washington Medical Center, Seattle, Washington,the Division of Nephrology, Department of Medicine, Veterans Administration Hospital, San Francisco, and the Department of Medicine,

UCLA School of Medicine, Los Angeles, California, USA

During the last 10 years, there have been many majoradvances in our understanding of the pathogenesis of glomeru-tar injury. Of particular significance has been the recognition ofthe importance of proteolytic enzymes and oxidants in mediat-ing glomerular capillary wall and mesangial damage. The infil-trating polymorphonuclear leukocyte (neutrophil, PMN) andmonocyte have been long recognized as important sources ofproteinases and reactive oxygen species [reviewed in 1—3].However, it is now known that endogenous glomerular cellsmay also release proteinases and oxidants that may also influ-ence the local inflammatory response [reviewed in 4—8]. In thisbrief review, we will highlight some of the studies from ourlaboratory that have dealt with these important mediators. Fora more comprehensive discussion of the role of proteinases andreactive oxygen species in glomerulonephritis (GN), the readeris referred to several excellent reviews [4—8].

Role of proteinases and oxidants derived from infiltrating cellsIn glomerulonephritis (GN)

One of the most critical cells in host defense and inflamma-tion is the neutrophil (PMN). This "professional phagocyte"has an important role in the killing of microorganisms via itsability to release highly toxic reactive oxygen metabolites,proteinases, and cationic proteins [reviewed in 1—3] (Table I).Monocytes, and to a much lesser extent, platelets, are alsocapable of releasing proteinases and oxidants [1, 9—12]. All ofthese cell types may be variably present in the inflamedglomerulus.

Role of the PMN in glomerular injury

An important role for neutrophils (PMN5) in glomerularinjury has been recognized since the l960s [reviewed in 13].PMNs can be demonstrated in glomeruli in a variety of prolif-erative and exudative glomerular diseases in man, includingpost-infectious ON, diffuse proliferative lupus nephritis, cryo-globulinemic GN, Henoch-Schönlein purpura, and variousforms of rapidly progressive glomerulonephritis (RPGN).PMNs can also be demonstrated within glomeruli in experimen-tal models of nephritis, especially in models induced withanti-glomerular basement membrane antibody (nephrotoxic ne-phritis), or in models associated with subendothelial or mesan-

© 1994 by the International Society of Nephrology

gial immune complex deposits [14, 15]. By electron microscopythe PMNs can often be shown to be activated, and may beactively phagocytosing immune complexes or be adherent todenuded glomerular basement membrane (GBM) (Fig. 1). Fur-thermore, studies in several of these experimental models haveshown that the proteinuria can be reduced or eliminated if theanimal is depleted of PMNs by an anti-PMN antibody [15, 16],and that proteinuria can be induced in these leukopenic animalsfollowing a PMN transfusion [17]. This provided strong evi-dence that the PMN can induce glomerular capillary wall injurywith alterations in GBM permeability resulting in proteinuria.

PMN oxidants and glomerular injury

PMNs, when stimulated by immune complexes, activatedcomplement fragments, and a variety of other stimuli, willundergo a respiratory burst in which they release several highlytoxic metabolites derived from the step-wise reduction ofoxygen [reviewed in 1]. Principal among the reactive oxygenproducts are the superoxide anion (02) and hydrogen peroxide(H202). PMNs may also generate hydroxyl radical (OH) via theiron catalyzed interaction of 02 and H202 (Haber-Weissreaction), although this reaction appears to be limited in vivo bythe availability of free iron [18, 19]. Recently the formation of•OH by a myeloperoxidase-dependent mechanism has beenproposed [19]. The formation of singlet oxygen ('02) by intactPMNs remain controversial [reviewed in 1, see 20 for recentstudy in this area].

Some of the most toxic oxygen metabolites produced byPMNs are those generated by the myeloperoxidase-hydrogenperoxide (H202)-halide system [1]. Myeloperoxidase (MPO) is ahighly cationic enzyme present in the azurophil (primary)granules of PMNs which catalyzes the reaction of H202 with ahalide [such as, chloride (Cl)] to generate hypohalous acidsand halogens [1]. These products are capable of killing a varietyof microorganisms as well as mammalian cells [1].

We have examined the ability of the isolated components ofthe MPO-H202-halide system to kill cultured rat mesangial cellsas measured by 5'Cr release. Incubation of mesangial cells withMPO (900 mU/ml) and H202 (l0— M) in Hanks balanced saltsolution (HBSS) as a chloride source resulted in a markedcytotoxicity (Fig. 2). This was not seen if the mesangial cellswere incubated with MPO alone or H202 alone (Fig. 2). Thecytotoxicity could also be prevented if the MPO was inactivated(such as by heat or azide), if the H202 was degraded with

352

Johnson et a!: Oxidants and proteases in injury 353

Table 1. Major neutrophil products potentially involved inglomerular injury

Reactive oxygen metabolitesSuperoxide anion (021Hydrogen peroxide (H202)Products of the myeloperoxidase-H202-chloride systemHydroxyl radical (.OH)aSinglet oxygena

ProteinasesSerine (elastase, cathepsin G, proteinase 3)Metallo (the 92 kD gelatinase, collagenase)

Cationic proteinsDefensinsLysozymeBactericidallpermeability-increasing factor

Phospholipase productsLeukotnene B4Thromboxane B2Platelet activating factora Some controversy exists as to whether PMNs produce 0H [18, 191

or singlet oxygen [1].

Fig. 1. Electron micrograph of a PMN adherent to the glomerularbasement membrane in experimental glomerulonephritis. Arrowed is aPMN within a glomerular capillary of a proteinuric rat with immunecomplex nephritis (the concanavalin Alanti-concanavalin A model [15])48 hours after disease induction. The PMN is adherent to immunedeposits present on the denuded GBM. There is also detachment of theglomerular epithelial cell from the GBM. (x 12,000).

catalase, or if scavengers of the hypohalous acids were present(for example, methionine or taurine; Table 2). The cytotoxicityof the MPO system also requires the presence of a halide (Table2). Mesangial cells incubated with MPO and/or H202 in HBSSin which chloride was replaced by sulfate show minimal cyto-toxicity. With the addition of chloride (10 mM), iodide (0.1 mM),or bromide (0.1 mM), a significant increase in cell death occurs(Table 2).

Varani et al have reported that rat MC could be killed byH202 alone [22]. We were also able to demonstrate that H202could induce 51Cr release from MC, although it required muchhigher concentrations (that is, l0— M or higher) than thoseneeded for MPO-mediated cytotoxicity (data not shown). Incontrast to our data, however, Varani et a! showed that MPOinhibited the H202-mediated killing of mesangial cells [221.However, in their assay system the MPO and H202 werepreincubated for 60 minutes in the presence of bovine serumalbumin prior to the addition of the mesangial cells [221.

Fig. 2. The myeloperoxidase-hydrogen peroxide-halide system is cyto-toxic to cultured rat mesangial cells. Well characterized rat mesangialcells [21] that had been labeled overnight with 51Cr (3 jsCilml) wereincubated in various solutions for 1 hour at 37°C and the percent 51Crrelease measured according to standard methods [21]. Significant 51Crrelease was only observed with mesangial cells that were incubatedwith active myeloperoxidase (MPO) (900 mU/mi), hydrogen peroxide(H202) (10 M), and a halide (that is, C1). All experiments wereperformed using Hanks balanced salt solution (HBSS) as a source ofchloride (Cl). Incubations were 2 hours. Experiments were performedin quadruplicate, and data are given as mean SD. The P value(Student's t-test) for the difference from MPO + H202 and each of theother three groups (HBSS, MPO alone, and H202 alone) was <0.001.

Peroxidase under these conditions would be expected to utilizeH202 for the production of more powerful oxidants (such ashypohalous acids) which are scavenged by albumin before theycan react with the target cell. In vivo, the adherence of the PMNto the target cell may effectively exclude serum proteins andother scavengers from the site of the MPO reaction [23].

The observation that the MPO system can induce cytotoxic-ity in vitro suggests that it may play a role in the cellular injurythat accompanies PMN-dependent GN. Given the highly cat-ionic nature of MPO (isoelectric point> 10), one might hypoth-esize that release of MPO by the activated PMN within theglomerulus could result in its localization to the highly anionicstructures within the basement membrane (such as heparansulfate proteoglycans) and on the glomerular endothelial orepithelial cell (such as sialoproteins). In the presence of suffi-cient local concentrations of H202, the MPO system would thengenerate toxic oxidants on the glomerular capillary wall.

To test this hypothesis, we perfused the renal arteries of ratswith MPO (108 sg) followed by H202 (2 ml of l0— M) inphosphate buffered saline (PBS) [24, 25]. Light and ultrastruc-tural enzyme histochemistry localized the MPO to the GBM[24]. Marked glomerular injury was documented, with thedevelopment of modest proteinuria (31 5 mg/24 hr in MPO-H202 perfused rats vs. 10.0 0.6 mg/24 hr in PBS-perfusedcontrol rats, mean SEM, P < 0.05). Light and electronmicroscopy revealed swelling and lysis of the glomerular endo-thelial and mesangial cells, with fusion of glomerular epithelialcell foot processes. Large areas of endothelium were denudedfrom the GBM, and a pronounced platelet influx was present.Within several days a marked glomerular cell proliferationoccurred, culminating in a hypercellular lesion resemblingproliferative glomerulonephritis [24, 25]. In contrast, infusion ofMPO alone or H202 alone resulted in no proteinuria and only

a)Ca

a)a)

0

100

80

60

40

20

0HBSS MPO + MPO H202

H202

354 Johnson et a!: Oxidants and proteases in injury

Table 2. The myeloperoxidase-hydrogen peroxide-halide system iscytotoxic to cultured rat mesangial cells

Reagents % 51Cr release P value

Experiment 1MPO + 11202 (in HBSS) 75.8 8.3IMP0 + H202 17.0 1.2 <0.001MPO + H202 + azide 16.7 1.6 <0.001MPO + H202 + catalase 15.3 2.4 <0.001MPO + H202 + methionine 16.4 1.4 <0.001MPO + H202 + taurine 16.8 3.2 <0.001

Experiment 2MPO + H202 (chloride-free) 21.4 2.2MPO + H202 + C1 64.9 3.9 <0.001MPO + H202 + 1 69.2 2.2 <0.001MPO + H202 + Br 59.6 2.7 <0.001

Cultured rat mesangial cells [21] were labeled overnight with 51Cr (3CiJml) and were then incubated in various solutions for 1 hour at 37°Cand the percent 51Cr release measured according to standard methods[21]. The concentrations of reagents used in these experiments wereMPO (900 mU/mi), 11202 (1O si), azide (100 MM), catalase (250 iJnil),methionine (5 msi) and taurine (5 mM). MP0 refers to heat-inactivatedMPO. The buffer used in experiment 1 was Hanks balanced SaltSolution (HBSS) as a source for chloride. In experiment 2 the incuba-tion was performed in a solution isoosmolar to HBSS but in which thechloride was replaced by sulfate, The concentrations of the addedhalides were Cl' (10mM), 1 (0.1 mM) and Br (0.1 mM). Experimentswere performed in quadruplicate, and data are given as mean SD. TheP value, determined by Student's f-test, is shown for the differencefrom MPO + H202 (in HESS) in experiment 1 and from MPO + H202(chloride-free) in experiment 2.

minor histologic changes (MPO alone, 8.4 0.8 mg/24 hr; H202alone, 9.2 0.4 mg124 hr) [24, 25].

There is also evidence that the MPO-H202-halide system isactivated in PMN-mediated ON. One of the consequences ofactivation of the MPO system is the halogenation of tissue dueto the oxidation of the halide to a form which binds covalentlyto tissue proteins [1]. We examined whether there was haloge-nation (iodination) of GBM in a PMN-mediated model ofimmune complex GN in the rat [15]. Rats with GN that received125l4abeled iodide had significant incorporation of 125! intoglomeruli and GBM. The iodination of GBM as well as protein-uria were not observed in rats that had been PMN-depleted withan anti-PMN serum. This provides indirect evidence for theinvolvement of MPO in PMN-mediated glomerular disease.

Of great interest has been the recent finding of circulatinganti-MPO antibodies in the majority of patients with idiopathiccrescentic rapidly progressive ON (RPGN) [261. A potentialmechanism by which the anti-MPO antibodies may be patho-genic has been recently provided by Brouwer et a! [27, 28].These investigators reported that the infusion of PMN contentscontaining MPO, and exogenous H2O2 into the renal artery ofrats that had been pre-immunized with MPO develop a severenecrotizing crescentic glomerular lesion indistinguishable fromidiopathic RPGN [27, 28]. In this model of RPGN, the diseaseprocess likely results from two mechanisms: (1) the productionof toxic oxygen metabolites on the capillary wall by theMPO-H202-halide system; and (2) the consequences of thehumoral and cellular immune response to the immune com-plexes of MPO and anti-MPO on the GBM. Anti-MPO antibod-ies may also contribute to the lesion by the stimulation ofcytokine-primed PMNs to release oxidants that subsequently

damage endothelial cells [29, 30]. It seems possible that thesemechanisms may be operative in patients who have autoanti-bodies to MPO and then develop an illness that leads tointravascular and/or intraglomerular PMN activation.

Other reactive oxygen species may also be involved inPMN-mediated glomerular injury. For example, the direct renalartery perfusion in rabbits and/or rats of specific oxidants or ofenzyme systems which have been reported to be capable ofgenerating O2, H2O2, 'OH, or 102 results in severe glomerularinjury with endothelial cell denudation, mesangiolysis, andintraglomerular platelet aggregation [31—33]. Studies using scav-engers of H2O2 [34—37], 02 [37, 38], and OH [39] have alsoimplicated a role for these oxidants in mediating the proteinuriain various PMN-dependent models of GN. However, studiesusing deferoxamine to inhibit 'OH generation [39] must beinterpreted with caution, as this compound also inhibits MPO-mediated reactions [40], and can also accelerate the autooxida-tion of iron with the formation of toxic oxidants [41].

In most of these studies, reactive oxygen species mediateboth cellular injury and an increase in capillary wall permeabil-ity as reflected by proteinuria. Other studies, however, havedemonstrated that continuous infusion of low concentrations ofH202 for one hour induces proteinuria without evidence forhistologic injury [42]. The mechanism by which oxidants induceproteinuria in the absence of cellular damage is unknown, butmay relate to direct oxidant injury to the GBM, to hemody-namic effects, or to the activation of endogenous proteinases.For example, it is known that oxidants such as those generatedby the MPO system may activate latent type IV collagenases(gelatinases) capable of degrading GBM [43].

PMN proteinases and glomerular injury

In vitro studies have demonstrated the ability of PMN andmonocyte proteinases to degrade type IV collagen and otherextracellular matrix components present in GBM and mesangialmatrix [44-48]. Proteinases and GBM degradation products canalso be demonstrated in urine in animals and humans withglomerular diseases in which PMN infiltration is prominent[49—51].

The principal PMN proteinases involved in GBM degradationappear to be the PMN serine proteinases, elastase and cathep-sin G, and a metalloproteinase, the 92 lcD type IV collagenase/gelatinase [45—48]. We have studied the in vivo consequences ofinfusing purified PMN elastase into the renal artery of rats [52].Infusion of small quantities of active enzyme (that is, 50 sg)induced massive but transient proteinuria (Fig. 3). No signifi-cant proteinuria developed if rats were infused with inactivatedelastase, despite the fact that both active and inactive elastaselocalized to the glomerular capillary wall as shown by immu-nofluorescence. Cathepsin 0, but not inactivated cathepsin G,also induced severe proteinuria (Fig. 3). It was of interest thatthe proteinuria induced by these proteinases was not associatedwith any evidence of histologic damage by light or electronmicroscopy. Our interpretation is that these proteinases actprimarily to degrade extracellular matrix components within theGBM as opposed to being directly toxic to the various glomer-ular cell populations. This is also supported by studies ofnephrotoxic nephritis in mice that have PMNs that are deficient

0C)C)XL

,I.

0)C)

Fig. 3. Effect of renal artery perfusion ofvarious PMN components on inducingproteinuria in rats. Shown is the 24-hour urineprotein excretion induced in rats by the renalartery perfusion of phosphate buffered saline(PBS) (N = 15), or by the following PMNproducts in PBS: elastase (N = 13),inactivated elastase (In-elastase) (N = 6),cathepsin G (N = 4), inactivated cathepsin G(In-cathepsin G) (N = 4), proteinase 3 (N = 4at 50 sg, N = 1 at 385 pg), or a mixture ofhuman defensins 1, 2, and 3 (N = 6). Thedoses administered are shown under each bar.All perfusions were administered to rats thathad previously undergone a unilateralnephrectomy. Values shown are mean SEM.

in neutral proteinases [53]. These mice have significant evi-dence of glomerular endothelial cell injury (presumably anoxidant-mediated effect) but little proteinuria [53].

Recently a third neutral serine proteinase has been identifiedin PMNs and has been named proteinase 3 [2, 54]. By ultra-structural immunocytochemistry, proteinase 3 has been shownto reside in the MPO-positive primary granules and on theplasma membrane of the PMN [55]. Proteinase 3 can degradeseveral extracellular matrix components, including type IVcollagen and laminin [56], and can induce emphysema inhamsters upon intratracheal injection [54]. Much attention hasbeen focussed on proteinase 3 following the recent identifica-tion of this proteinase as the antigen to which the anti-neutro-phil cytoplasmic antibody is directed against in Wegener'sgranulomatosis [57].

We have examined the ability of proteinase 3 to induceglomerular injury in vivo (Johnson R, Gray B, and Klebanoff SJ,unpublished data). Rats underwent a unilateral nephrectomyfollowed by renal artery perfusion of the remaining kidney aspreviously described [52]. PBS (0.5 ml) is first injected todisplace the blood from the kidney, followed by infusion of theproteinase. Unlike elastase and cathepsin G, infusion of 50 p.gof purified proteinase 3 did not induce proteinuria (12 2 mg/24hr, normal range 15 3 mg/24 hr; Fig. 3). However, doses of100 g, 150 and 385 jsg given to individual rats resulted in24, 21, and 248 mg proteinuria, respectively. As with theelastase-perfused rats, light microscopy revealed no abnormal-ities. This would again be consistent with the hypothesis thatthe proteinuria induced with high dose proteinase 3 is theconsequence of degradation of the GBM as opposed to glomer-ular cell injury.

PMN cationic proteins and glomerular injuryIn addition to proteinases and oxidants, PMNs can also

release cationic proteins with bactericidal activity, includinglysozyme, bactericidal/permeability increasing factor, and thesmall 3.3 to 3.9 kD defensins [reviewed in 2]. A variety ofstudies have suggested that intraglomerular release of cationicproteins from neutrophils and platelets may result in their

Fig. 4. Localization of human defensins to the glomerular capillaryand mesangium following renal artery perfusion in the rat. Shown is aglomerulus of a rat that has been perfused with defensin (100 sg 5 mmpreviously) and then immunostained with a rabbit anti-defensin anti-body. Defensin is localized to the mesangium and capillary wall.Vehicle-perfused rats were negative when immunostained with theanti-defensin antibody. (Indirect immunofluorescence, x400).

localization to the glomerular capillary wall [58, 59], where theymay have a role in increasing capillary wall permeability byneutralizing the anionic components of the GBM [reviewed in60]. We therefore examined whether the renal artery perfusionof defensins could induce proteinuria in rats (Fig. 3). Infusion of100 pg of purified human defensins (containing a mixture ofdefensins 1, 2, and 3 (1:1:0.5) resulted in minimal proteinuria (21

2 mg124 hr, N = 6) and normal histology. Similarly, largerdoses (that is, 500 sg) given to two rats also resulted in only13.1 and 24.0 mg proteinuria in the first 24 hours. The defensinscould be shown to bind to the glomerular mesangium andcapillary wall by immunofluorescence with a specific antibody(Fig. 4), and in separate experiments, the binding of '25I-labeleddefensins to glomeruli by a double isotope binding assay [61]was documented to be 29.1 1% immediately after perfusion.Thus, it would appear that defensins, while binding to theglomerular capillary wall, may not contribute significantly toPMN-mediated glomerular injury.

Johnson et al: Oxidants and proteases in injury 355

300

250

200

150

100

50

0

356 Johnson et a!: Oxidants and proteases in injury

Role of platelets in PMN-mediated glomerular injuryAn interesting observation is the frequent co-localization of

platelets with PMNs within glomeruli in immune complexnephritis [62, 631. This is particularly evident in a PMN-dependent model of immune complex GN in the rat resemblingdiffuse proliferative lupus nephritis (the concanavalin A-anti-concanavalin A model) in which subendothelial and mesangialdeposits are prominent [63]. To investigate the role of plateletsin this model, we studied rats with GN that had been previouslydepleted of platelets with an anti-platelet antibody [61]. Plateletdepletion significantly reduced proteinuria without affectingPMN localization. Since PMN depletion also decreases theproteinuria in this model [15], this suggests that both plateletsand PMNs are important mediators of the capillary wall injury.

While it is possible that platelets could injure the capillarywall independently of PMNs via the release of vasoactiveamines, proteases, oxidants, thromboxane, and other plateletproducts [reviewed in 64], indirect evidence suggests a potentialPMN-platelet interaction. Numerous in vitro studies have dem-onstrated that platelets can affect various PMN functions [re-viewed in 64, 65]. Some of these functions may be mediated byGMP-140, a recently recognized platelet granule membraneglycoprotein that redistributes to the plasma membrane whenplatelets are activated, and mediates the adherence of theactivated platelets to PMNs [66]. Of particular relevance hasbeen the observation that platelets are also required for thePMN-mediated lung injury that occurs in rats after systemiccomplement activation [67]. It has been proposed that this ismediated by platelet adenine nucleotides (ATP and ADP),which act to enhance 02 and MPO release by the PMN [67,68]. The recent observation that rats with mesangial prolifera-tive GN have an increased glomerular 02 response whenadministered the ATP analog, ATPyS, [69] would be consistentwith this hypothesis.

Role of proteinases and oxidants derived from intrinsicglomerular cells in GN

Mesangial cell proteinases and oxidants in glomerular injuryThe observation that glomerular cells can generate reactive

oxygen species was first demonstrated by Shah in 1981 [70]. Itis now known that mesangial cells produce oxidants in responseto multiple stimuli, including immune complexes, the comple-ment membrane attack complex, and various cytokines [71—74].The precise role of mesangial cell-derived oxidants in theevolution of glomerular injury remains to be defined. As expo-sure of basement membrane or extracellular matrix componentsto reactive oxygen species facilitates degradation by protein-ases [75], it may be that local release of these compounds duringglomerular inflammation augments proteinase-mediated struc-tural damage.

Early studies utilizing mesangial cells described a non-spe-cific neutral protease activity within culture supernatants [76,77]. This activity displayed notable features, including depen-dence on zinc for activity, secretion in a latent, proenzymeform, and activity against isolated glomerular basement mem-branes [76]. Subsequent purification and characterization of theproteolytic activities secreted by cultured human and rat me-sangial cells have defined the structural and molecular featuresof these enzymes [78—81]. Two major matrix metalloproteinases

have been identified in this manner: matrilysin (PUMP-i) andthe 72 kD Type IV collagenase [78—81]. Matrilysin was identi-fied in cultured human mesangial cells by a PCR-based homol-ogy cloning strategy [81]. In addition, matrilysin activity waspurified and identified on the basis of specific substrate activi-ties and by the use of specific anti-peptide antibodies. Theabundance of the matrilysin mRNA was specifically induced bythe inflammatory cytokines, interleukin 1 and tumor necrosisfactor. The matrilysin protein was also detected within themesangial region of clinical biopsy specimens of acute prolifer-ative glomerulonephritis, particularly in those cases character-ized by infiltrating monocytes expressing interleukin 1 [81].

The 72 kD Type IV collagenase activity produced by mesan-gial cells has been rigorously characterized by extensive puri-fication and amino acid sequence analysis [78, 79]. The expres-sion of this enzyme in most studied cells appears to beconstitutive, with little response to inflammatory cytokines orphorbol esters [82]. In contrast, studies with cultured human orrat mesangial cells have demonstrated both cytokine-mediatedand second-messenger inducibility [83]. Mesangial cell synthe-sis of the 72 kD Type IV collagenase is induced by interleukin1, tumor necrosis factor, and transforming growth factor /31(TGF-f31) [83]. In addition, elevation of cyclic AMP inducesmesangial cell synthesis of the enzyme [83]. Recent studieshave defined the presence of multiple AP-2 sites, severaloncogene binding motifs, and a probable TGF-f31 activationelement (Lovett et al, manuscript submitted for publication).Thus, the regulation of this gene in vivo may be subjected tomultiple levels of tissue-specific control by inflammatory cyto-kines.

Recently we have investigated the expression of the 72 kDType IV collagenase in a rat model of mesangial proliferativeGN induced by injecting an antibody directed against the Thy 1antigen expressed on the surface of mesangial cells [80]. Themodel results in an acute complement mediated injury to themesangium, with dissolution of the mesangial matrix and loss ofmesangial cells ("mesangiolysis"), followed by a marked pro-liferative response with matrix expansion [84]. In normal gb-meruli, approximately 10% of the mesangial cell populationexpress the 72 kD Type IV collagenase, usually in areasimmediately adjacent to the basement membrane. Followinginduction of Thy 1 nephritis there is a dramatic increase in theabundance of 72 kD Type IV collagenase-expressing cells,especially within foci of proliferating cells and in regions ofmesangial matrix dissolution. The Type IV collagenase antigenwas also localized to sites of obvious GBM disruption, suggest-ing that augmented secretion of this enzyme may account forthe proteinuria that is variably observed in this model.

Other proteinases may also be involved in mesangial prolif-erative diseases. Recently Tomooka et al have shown thatplastninogen activator activity is reduced and plasminogenactivator inhibitor-i synthesis is increased in diseased glomerulifrom rats with Thy 1 nephritis [85]. The authors suggested thatthis may be one of the mechanisms by which matrix accumu-lates in this disease model.

Glomerular epithelial cell proteinases and oxidants andglomerular injury

Several studies, primarily in vivo, have suggested that oxi-dants mediate glomerular injury in PMN-independent models of

Johnson et a!: Oxidants and proteases in inju,y 357

glomerular disease associated with injury to the glomerularepithelial cell [86—90]. Thus, studies in rats with aminonucleo-side nephrosis [871, passive Heymann nephritis [88, 89], and acomplement-independent model of membranous GN [901 haveall reported a reduction in proteinuna with the administration of0H scavengers. Scavengers of 02, or inhibitors of 02

formation, also reduce proteinuria in the aminonucleosidemodel [86, and reviewed in 6].

Glomerular epithelial cells also produce several proteinases,including cathepsin D [91], urokinase-type plasminogen activa-tor [92], and several variably characterized neutral metallopro-teinases [93, 94]. While several investigators have suggestedthat cultured glomerular epithelial cells secrete the 92 and 72 kDType V collagenases (M. Davies, personal communication), ourlaboratories have focused on the purification and characteriza-tion of a distinct set of Type IV collagen-degrading activities.Gelatin zymographic analysis demonstrates three proteolyticactivites with molecular masses of 220, 150, and 98 kD. Thesemasses distinguish these proteases from the known members ofthe collagenase supergene (matrix metalloproteinase) family.All of these activities are dependent upon zinc for activity.Significantly, these activities are not secreted in latent, pro-enzyme form and are not inhibited by recombinant tissueinhibitor of metalloprotease-2. Ultimate identification of theseenzymes awaits the structural and molecular characterizationcurrently underway in our laboratories, which is utilizing acombined protein purification and expression cloning strategy.

A possible role for these enzymes in glomerular disease issupported by the observation of a nine-fold increase in gelati-nase activity in glomerular homogenates from rats with activeHeymann's nephritis [94]. Neutral proteinases have also beenisolated from the urine of rats with aminonucleoside nephrosisin which the principal lesion is of the glomerular epithelial cell[51].

Finally, there is evidence for other GBM-degrading protein-ases in glomeruli, although the specific cellular source withinthe glomerulus remains unknown. These include a 116 to 120kD membrane associated metalloproteinase [95] as well as thecysteine proteinases, cathepsins B, H, and L [96, 97]. The 116kD membrane-associated enzyme is active against basementmembrane substrates, and recent studies have detected a sim-ilar activity within plasma membrane preparations of culturedglomerular epithelial cells (D. Lovett, unpublished data). Thebiological significance of these various activites is not welldefined, although administration of cysteine proteinase inhibi-tors to rats with a PMN-independent, complement-independentmodel of nephrotoxic nephritis reduces proteinuria by 40 to45% [98].

Conclusion

Proteinases and oxidants are likely to have an important rolein both PMN-dependent and PMN-independent GN. In PMN-mediated injury, the release of reactive oxygen species (andespecially products of the MPO-H202-halide system) by thePMN may result in direct injury of the intrinsic cells within theglomerulus as well as halogenation and oxidation of the GBMthat results in an increase in permeability. In contrast, PMNserine proteinases do not appear to cause cellular toxicity butrather cause degradation of GBM components that results insubstantial proteinuria. PMN defensins, despite localizing to

the GBM and mesangial matrix, induce little or no proteinuriaor morphologic evidence of cellular injury.

Oxidants and proteinases derived from resident glomerularcells may also cause or amplify glomerular injury. Mesangialcells can release oxidants and proteinases in response to avariety of inflammatory mediators, and the up-regulated expres-sion of type IV collagenase in mesangial proliferative nephritismay have an important role in the remodeling and restoration ofthe normal mesangial matrix. Finally, evidence suggests thatglomerular epithelial cells make oxidants and GBM-degradingproteinases, and that these may be involved in proteinuricdiseases characterized by injury to the glomerular epithelialcell.

Acknowledgments

We thank Dr. Beulah Gray (University of Minnesota, Minneapolis,MN) for kindly providing purified proteinase 3. Support for thesestudies was provided by U.S. Public Health Service grants DK-43422,AI-07763, and DK-02l42.

No reprints are available. Correspondence to Richard J. Johnson,M.D., Division of Nephrology, Room 11, Department of Medicine,University of Washington Medical Center, Seattle, Washington 98195,USA.

References

1. KLEBANOFF SJ: Oxygen metabolites from phagocytes, inInflammation: Basic Principles and Clinical Correlates, edited byGALLIN JI, GOLDSTEIN IM, SNYDERMAN R, New York, RavenPress, Ltd. 1992, pp 541—588

2. BAGGI0LINI M, DEwALD B: Exocytosis by neutrophils, in Con-temporary Topics in Immunology, edited by SNYDERMAN R, NewYork, Plenum Press, 1984, pp 221—246

3. LEHRER RI (Moderator): Neutrophils and host defense. Ann InternMed 109:127—142, 1988

4. BAUD L: Reactive oxygen species: production and role in thekidney. Am J Physiol 251 :F765—F776, 1986

5. SHAH SY: Role of reactive oxygen metabolites in experimentalglomerular disease. Kidney mt 35:1093—1106, 1989

6. DIAMOND JR: The role of reactive oxygen species in animal modelsof glomerular disease. Am J Kidney Dis 19:292—300, 1992

7. DAVIES M, COLES GA, THOMAS GJ, MARTIN J, LOVETT DH:Proteinases and the glomerulus: Their role in glomerular diseases.K!in Wochenschr68:1145—1149, 1990

8. BARICOS WH, SHAH SV: Proteolytic enzymes as mediators ofglomerular injury. Kidney mt 40:161—173,1991

9. CAMPBELL EJ, SILVERMAN EK, CAMPBELL MA: Elastase andcathepsin G of human monocytes. Quantification of cellular con-tent, release in response to stimuli, and heterogeneity in elastase-mediated proteolytic activity. J Immunol 143:2961—2968, 1989

10. JAMES HL, WACHTFOGEL YT, ZIMMERMAN JM, C0LMAN RW,COHEN AB: A unique elastase in human blood platelets. J ClinInvest 76:2330—2337, 1985

11. MARCUS AJ, SILK ST, SAFIER LB, ULLMAN HL: Superoxideproduction and reducing activity in human platelets. J Cliii Invest59: 149—158, 1977

12. DEJESUS M JR, FIKRIG S, DETWILER T: Phagocytosis-stimulatednitroblue tetrazolium reduction by platelets. J Lab Clin Med80:117—124, 1972

13. COCHRANE CG: Immunologic tissue injury mediated by neutro-philic leukocytes. Adv Immunol 9:99—162, l%8

14. SALANT DJ, ADLER S, DARBY C, CAPPARELL NJ, GROGGEL GC,FEINTZEIG ID, RENNKE HG, DITFMER JE: Influence of antigendistribution on the mediation of immunological glomerular injury.Kidney mt 27:939—950, 1985

358 Johnson et a!: Oxidants and proteases in injury

15. JOHNSON RJ, KLEBANOFF SJ, OCHI RF, ADLER S, BAKER P,SPARKS L, COUSER WG: Participation of the myeloperoxidase-H202-halide system in immune complex nephritis. Kidney mt32:342—349, 1987

16, COCHRANE CG, UNANUE ER, DIXON FJ: A role of polymorphonu-clear leukocytes and complement in nephrotoxic nephritis. J ExpMed 122:95—110, 1965

17. HENSON PM: Pathologic mechanisms in neutrophil mediated in-jury. Am J Pathol 68:593—612, 1972

18. COHEN MS, BRITIGAN BE, HASSETT DJ, ROSEN GM: Do humansneutrophils form hydroxyl radical? Evaluation of an unresolvedcontroversy. Free Radical Biol Med 5:81—88, 1988

19. RAMOS CL, Pou S, BRITIGAN BE, COHEN MS, ROSEN GM: Spintrapping evidence for myeloperoxidase-dependent hydroxyl radicalformation by human neutrophils and monocytes. J Biol Chem267:8307—8312, 1992

20. STEINBECK SJ, KHAN AU, KARNOVSKY MJ: Intracellular singletoxygen generation by phagocytosing neutrophils in response toparticles coated with a chemical trap. J Biol Chem 267:13425—13433, 1992

21. STAHL RAK, ADLER S, BAKER PJ, JOHNSON Ri, CHEN Y-P,PRITZL P, COUSER WG: Cyclosporin A inhibits prostaglandin E2formation by rat mesangial cells in culture. Kidney mt 35:1161—1167, 1989

22. VARANI J, TAYLOR CG, RISER B, SHUMAKER DK, YEH K-Y, DAMEM, GIBBS DF, TODD RF III, DUMLER F, BROMBERG J, KILLEN PD:Mesangial cell killing by leukocytes: role of leukocyte oxidants andproteolytic enzymes. Kidney ut 42:1169—1177, 1992

23. VISSERS M, DAY A, WINTERBOURN C: Neutrophils adherent to anonphagocytosable surface (glomerular basement membrane) pro-duce oxidants only at the site of attachment. Blood 66:161—166,1985

24. JOHNSON RJ, COUSER WG, CHI EY, ADLER S, KLEBANOFF SJ:New mechanism for glomerular injury. Myeloperoxidase-hydrogenperoxide-halide system. J Cliii Invest 79:1379—1387, 1987

25. JOHNSON RJ, GUGGENHEIM SJ, KLEBANOFF Si, 0cm RF, WASSA, BAKER P, SCHULZE M, COUSER WG: Morphologic correlates ofglomerular oxidant injury induced by the myeloperoxidase-hydro-gen peroxide-halide system of the neutrophil. Lab invest 58:294—301, 1988

26. JENNETTE JC, WILKMAN AS, FALK RJ: Anti-neutrophil cytoplas-mic antibody-associated glomerulonephritis and vasculitis. Am JPathol 135:921—930, 1989

27. BROUWER E, HUITEMA MG, KLOK PA, TERVAERT COHEN JW,WEENING JJ, KALLENBERG CGM: Anti-myeloperoxidase antibod-ies contribute to the damage in MPO-hydrogenperoxide (H202)-halide mediated glomerular injury. (abstract) J Am Soc Nephrol3:577, 1992

28. BROUWER E, HUITEMA MG, KLOK PA, TERVAERT COHEN JW,WEENING JJ, KALLENBERG CGM: Anti-myeloperoxidase (MPO)associated proliferative glomerulonephritis and vasculitis: An ani-mal model. (abstract) JAm Soc Nephrol 3:577, 1992

29. EWERT B, JENNETTE JC, FALK R: Anti-neutrophil cytoplasmicautoantibodies (ANCA) stimulate cytokine-primed neutrophils todamage endothelial cells. (abstract) JAm Soc Nephrol 1:520, 1990

30. SAVAGE COS, POTrINGER BE, GASKIN 0, PUSEY CD, PEARSONJD: Autoantibodies developing to myeloperoxidase and proteinase3 in systemic vasculitis stimulate neutrophil cytotoxicity towardcultured endothelial cells. Am J Pathol 141:335—341, 1992

31. STRATTA P. CANAVESE C, MAzzuccO G, MONGA G, NOVARA R,VoNo P, GATULLO D, BEDINO S, TESTORE G, CONGIN R, FERREROT: Mesangiolysis and endothelial lesions due to peroxidative dam-age in rabbits. Nephron 51:250—256, 1989

32. STRATTA P, CANAVESE C, DOGLIANI M, MONGA G, MAZZUCCO G,NOVARA R, V0N0 P, GATULLO D, BEDINO S, TESTORE G, VEst-CELLONE A: Experimental evidence for mesangiolysis due to hy-drogen peroxide. J Nephrol 1:37—41, 1989

33. ho 5, UEDA Y, SUGISAKI T, IIDAKA K: Induction of glomerularinjury by singlet oxygen. Nephron 60:204—209, 1992

34. REHAN A, JOHNSON KJ, KUNKEL RG, WIGGINS RC: Role ofoxygen radicals in phorbol myristate acetate-induced glomerularinjury. Kidney ut 27:503—511, 1985

35. REHAN A, JOHNSON KJ, WIGGINS RC, KUNKEL RG, WARD PA:

Evidence for the role of oxygen radicals in acute nephrotoxicnephntis. Lab invest 51:396—403, 1984

36, REHAN A, WIGGINS RC, KUNKEL RG, TILL GO, JOHNSON Ri:Glomerular injury and proteinuria in rats after intrarenal injectionof cobra venom factor. Evidence for the role of neutrophil-derivedoxygen free radicals. Am J Pat/so! 123:57—66, 1986

37. BIRTWISTLE RI, MICHAEL J, How AJ, ADU D: Reactive oxygenproducts in heterologous anti-glomerular basement membrane ne-phriti in rats. Br J Exp Pathol 70:207—213, 1989

38. ADACHI T, FUKUTA M, ITo Y, HIRANO K, SUGIURA M, SUGIURAK: Effect of superoxide dismutase on glomerular nephritis. Bio-chem Pharma 35:341—345, 1986

39. BOYCE NW, HOLDSWORTH SR: Hydroxyl radical mediation ofimmune renal injury by desferrioxamine. Kidney mt 30:813—817,1986

40. KLEBANOFF SJ, WALTERSDORPH AM: Inhibition of peroxidase-catalyzed reactions by deferoxamine. Arch Biochem Biophys 264:600—606, 1988

41. KLEBANOFF Si, WALTERSDORPH AM, MICHEL BR, ROSEN H:Oxygen-based free radical generation by ferrous ions and deferox-amine. I Biol Chem 264:19765—19771, 1989

42. YOSHIOKA T, ICHIKAwA I, Fooo A: Reactive oxygen metabolitescause massive, reversible proteinuria and glomerular sieving defectwithout apparent ultrastructural abnormality. J Am Soc Nephrol2:902—912, 1991

43. SHAH SY, BARICOS WH, BASCI A: Degradation of human glomer-ular basement membrane by stimulated neutrophils. I Clin invest

79:25—31, 198744. DAVIES M, BARRTEE AJ, TRAVIS J, SANDERS E, COLES GA: The

degradation of human glomerular basement membrane with purifiedlysosomal proteinases: Evidence for the pathogenic role of thepolymorphonuclear leukocyte in glomerulonephritis. Clin SciMolecMed54:233—240, 1978

45. DAVIES M, COLES GA, HUGHES KT: Glomerular basement mem-brane injury by neutrophil and monocyte neutral proteinases. RenalPhysiol (Base!) 3:106—111, 1980

46. VISSERS MCM, WINTERBOURN CC: Gelatinase contributes to thedegradation of glomerular basement membrane collagen by humanneutrophils. Collagen Re! Res 8:113—122, 1988

47. BARICOS WH, MURPHY G, ZHOU Y, NGUYEN HH, SHAH SV:Degradation of glomerular basement membrane by purified mam-malian metalloproteinases. Biochem J 254:609—612, 1988

48. HECK LW, BLACKBURN WD, IRWIN MH, ABRAHAMSON DR:Degradation of basement membrane laminin by human neutrophilelastase and cathepsin G. Am I Pathol 136:1267—1274, 1990

49. HAWKINS D, COCHRANE CG: Glomerular basement membranedamage in immunological glomerulonephritis. immunol 14:665—680,1968

50. SANDERS E, DAVIES M, COLES GA: On the pathogenesis ofglomerulonephritis. A clinico-pathological study indicating thatneutrophils attack and degrade glomerular basement membrane.Renal Physiol (Base!) 3:355—359, 1980

51. DAVIN i-C, DAVIES M, FOIDART M-C, FOIDART JB, DECHENNECA, MAHIEU PR: Urinary excretion of neutral proteinases innephrotic rats with a glomerular disease. Kidney In: 31:32—40, 1987

52. JOHNSON Ri, COUSER WG, ALPERS CE, VISSERS M, SCHULZE M,KLEBANOFF SJ: The human neutrophil serine proteinases, elastaseand cathepsin 0, can mediate glomerular injury in vivo. J Exp Med168:1169—1174, 1988

53. SCHRIJVER 0, SCHALEWUK J, ROBBEN JCM, ASSMANN KJM,KOENE RAP: Anti glomerular basement membrane nephritis inbeige mice. J Exp Med 169:1435—1448, 1989

54. KAO RC, WEHNER NG, SKUBITZ KM, GRAY BH, HOIDAL iR:Proteinase 3: A distinct human polymorphonuclear leukocyte pro-teinase that produces emphysema in hamsters. I C!in invest 82:

1963—1973, 198855. CSERNOK E, LUDEMANN i, GROSS WL, BAINTON DF: Ultrastruc-

tural localization of proteinase 3, the target antigen of anti-cyto-plasmic antibodies circulating in Wegener's granulomatosis. Am JPathol 137:1113—1120, 1990

56. RAO NV, WEHNER NO, MARSHALL BC, GRAY WR, GRAY BH,HOIDAL JR: Characterization of proteinase-3, a neutrophil serineproteinase. I Bio! Chem 266:9540—9548, 1991

Johnson et a!: Oxidants and proteases in injury 359

57. LUDEMANN J, UTECHT B, GROSS WL: Anti-neutrophil cytoplasmantibodies in Wegener's granulomatosis recognize an elastinolyticenzyme. JExp Med 171:357—362, 1990

58. CAMUSSI G, TErrA C, SEG0L0NI G, CODA R, VERCELLONE A:Localization of neutrophil cationic proteins and loss of anioniccharges in glomeruli of patients with systemic lupus erythematosusglomerulonephritis. Gun immunolImmuno!pathol24:299—314, 1982

59. CAMUSSI G, TETTA C, MERONI M, TORRI-TARELLI L, ROFFINELLOC, ALBERTON A, DERRGIBu5 C, SESSA A: Localization of cationicproteins derived from platelets and polymorphonuclear neutrophilsand local loss of anionic sites in glomeruli of rabbits with experi-mentally-induced acute serum sickness. Lab invest 55:56—62, 1986

60. CAMussi G, TETTA C, VERCELLONE A: Platelets and platelet-derived cationic proteins in human and experimental glomerularpathology. Contr Nephrol 69:27—36, 1989

61. JOHNSON RJ, ALPERS CE, PRITZL P, SCHULZE M, BAKER PJ,PRUCHNO C, Coust WG: Platelets mediate neutrophil-dependentimmune complex nephritis in the rat. J Clin Invest 82:1225—1235,1988

62. GABBIANI G, BADONNEL MC, VASSALLI P: Experimental focalglomerular lesions elicited by insoluble immune complexes. LabInvest 32:33—45, 1975

63. JOHNSON RJ, ALPERS CE, PRUCHNO C, SCHULZE M, BAKER PJ,PRITZL P, COUSER WG: Mechanisms and kinetics for platelet andneutrophil localization in immune complex nephritis. Kidney mt36:780—789, 1989

64. JOHNSON RJ: Platelets in inflammatory glomerular injury. SeminNephrol 11:276—284, 1991

65. KAZURA JW: Platelet-neutrophil interaction: Modulation of theinflammatory response. J Lab Gun Med 114:469—470, 1989

66. HAMBURGER SA, MCEVER RP: GMP-140 mediates adhesion ofstimulated platelets to neutrophils. Blood 75:550—554, 1990

67. WARD PA, MACCON! D, SULAVIK MC, TILL GO, WARREN is,JOHNSON KJ, POWELL J: Rat neutrophil-platelet interactions inoxygen radical-mediated lung injury. UCLA Symposium on Molec-ular and Cellular Biology, New Series 82:83—98, 1988

68. WARD PA, CUNNINGHAM TW, MCCULLOCH KK, PHAN SH,POWELL J, JOHNSON KJ: Platelet enhancement of 02 responses instimulated human neutrophils. Identification of platelet factor asadenine nucleotide, Lab Invest 58:37—47, 1988

69. POELSTRA K, HEYNEN ER, BALLER JF, HARDONK MJ, BAKKERWW: Modulation of anti-Thyl nephritis in the rat by adeninenucleotides. Lab invest 66:555—563, 1992

70, SHAH SV: Light emission by isolated rat glomeruli in response tophorbol myristate acetate. J Lab Gun Med 98:46—57, 1981

71. BAUD L, HAGEGE J, SRAER J, RONDEAU B, PEREZ J, ARDAILLOIJR: Reactive oxygen production by cultured rat glomerular mesan-gial cells during phagocytosis is associated with stimulation oflipoxygenase activity. J Exp Med 158:1836-1852, 1983

72. ADLER S, BAKER J, JOHNSON J, OCR! RF, PRITZL P. COUSER WG:Complement membrane attack complex stimulates production ofreactive oxygen metabolites by cultured rat mesangial cells. J GunInvest 77:762—767, 1986

73. SEDOR LR, CAREY SW, EMANCIPATOR SN: Immune complexesbind to cultured rat glomerular mesangial cells to stimulate super-oxide release, Evidence for a Fe receptor. J Immunol 138:375 1—3757, 1987

74. RADEKE HH, MEIER B, TOPLEY N, FLOEGE J, HABERMEHL GO,RESCH K: Interleukin 1-alpha and tumor necrosis factor-alphainduce oxygen radical production in mesangial cells. Kidney In:37:767—775, 1990

75. Vissits MCM, WINTERBOURN CC: The effect of oxidants onneutrophil-mediated degradation of glomerular basement mem-brane collagen. Biochim Biophys Acta 889:277—286, 1986

76. LOVETT DH, STERZEL RB, KASHGARIAN M, RYAN JL Neutralproteinase activity produced in vitro by cells of the glomerularmesangium. Kidney In: 23:342—349, 1983

77. MARTIN J, LOVETT DH, GEMSA D, STERZEL RB, DAvIES M:Enhancement of glomerular mesangial cell neutral proteinase se-cretion by macrophages: Role of interleukin I. J Immunol 137:525—529, 1986

78. DAVIES M, THOMAS GJ, MARTIN J, LOVETT DH: The purification

and characterization of a glomerular-basement-membrane-degrad-ing neutral proteinase from rat mesangial cells. Biochem J 251:419—425, 1988

79. MARTIN J, DAVIES M, THOMAS G, LOVETT DH: Human mesangialcells secrete a GBM-degrading neutral proteinase and a specificinhibitor. Kidney In: 36:790—801, 1989

80. LOVETT DH, JOHNSON RJ, MART! H-P, MARTIN J, DAVIES M,COUSER WG: Structural characterization of the mesangial cell typeIV collagenase and enhanced expression in a model of immunecomplex-mediated glomerulonephritis. Am J Pathol 141:85—98,1992

81. MARTI H-P, MCNEIL L, THOMAS 0, DAVIES M, LOVETT DH:Molecular characterization of a low-molecular-mass matrix metal-loproteinase secreted by glomerular mesangial cells as PUMP-i.Biochem J 285:899—905, 1992

82. COLLIER IE, WILHELM SM, EISEN AZ, MARMER BL, GRANT GA,SELTZZER JL, KRONBERGER A, HE C, BAUER EA, OLDBERO GI:H-ras oncogene-transformed human bronchial epithelial cells(TBE-1) secrete a single metalloprotease capable of degradingbasement membrane collagen. J Biol Chem 253:6579—6587, 1988

83. MART! H-P, MCNEIL L, DAVIES M, MARTIN J, LOVETT DH:Homology cloning of rat 72 kDa type IV collagenase: Cytokine andsecond-messenger inducibility in glomerular mesangial cells. Bio-chem J (in press)

84. FLOEGE J, JOHNSON RJ, GORDON K, IIDA H, PRITZL P,YOSHIMURA A, CAMPBELL C, ALPERS CE, COUSER WG: Increasedsynthesis of extracellular matrix in mesangial proliferative nephri-tis. Kidney mt 40:477—488, 1991

85. TOMOOKA 5, BORDER W, MARSHALL BC, NOBLE NA: Glomerularmatrix accumulation is linked to inhibition of the plasmin proteasesystem. Kidney mt 42:1462—1469, 1992

86. DIAMOND JR, BONVENTRE JV, KARNOVSKY MJ: A role for oxygenfree radicals in aminonucleoside nephrosis. Kidney mt 29:478—483,1986

87. THAKUR V, WALKER PD, SHAH SV: Evidence suggesting a role forhydroxyl radical in puromycin aminonucleoside-induced protein-uria. Kidney mt 34:494—499, 1988

88. LOTAN D, KAPLAN BS, FONG JSC, 000DYER PR, DE CHADARE-VIAN JP: Reduction of protein excretion by dimethyl sulfoxide inrats with passive Heymann nephritis. Kidney lnt 25:778—788, 1984

89. SHAH SV: Evidence suggesting a role for hydroxyl radical inpassive Heymann nephntis in rats. Am J Physiol 254:F337—F344,1988

90. RAHMAN M, EMANCIPATOR 5, SEDOR J: Hydroxyl radical scaven-gers ameliorate proteinuria in immune complex glomerulonephritis.J Lab Clin Invest 112:619-626, 1988

91. SINGH AK: Proteolytic machinery of glomerular epithelial cellsagainst IgG. Biochem Biophys Res Gommun 186:639—644, 1992

92. RONDEAU E, OcHI 5, LACAVE R, HE C-J, MEDCALF R, DELARUEF, SRAER JD: Urokinase synthesis and binding by glomerularepithelial cells in culture. Kidney Int 36:593—600, 1989

93. JOHNSON RJ, YAMABE H, CHEN YP, CAMPBELL C, GORDON K,BAKER P. LOVETT D, COUSER WG: Glomerular epithelial cellssecrete a glomerular basement membrane-degrading metalloprotein-ase. JAm Soc Nephrol2:1388—1397, 1992

94. WATANABE K, KINOSHITA S, NAKAGAWA H: Gelatinase secretionby glomerular epithelial cells. Nephron 56:405—409, 1990

95. LE QC, SHAH SV, NGUYEN HH, CORTEZ SL, Bxiucos WH: Anovel metalloproteinases present in freshly isolated rat glomeruli.Am J Physiol 260:F555—F561, 1991

96. BARICOS WH, CORTEZ SL, LE QC, ZHOU Y, DICARLO RM,O'CONNOR SE, SHAH SV: Glomerular basement membrane degra-dation by endogenous cysteine proteinases in isolated rat glomeruli.Kidney mt 38:395-401, 1990

97. THOMAS GJ, DAVIES M: The potential role of human kidney cortexcysteine proteinases in glomerular basement membrane degrada-tion. Biochim Biophys Acta 990:246—253, 1989

98. BAItICOS WH, O'CONNOR SE, CORTEZ SL, WV L-T: The cysteineproteinase inhibitor, E-64, reduces proteinurIa in an experimentalmodel of glomerulonephritis. Biochem Biophys Res Comm 155:1318—1323, 1988