The effect of oxidants on neutrophil-mediated degradation of glomerular basement membrane collagen

Bio~imica et Biophvsica Acta 889 (1986) 277-286

277

Elsevier BBA 11842

The effect of oxidants on neutrophU-mediated degradation of glomerular basement membrane collagen Margret C.M. Vissers * and Christine C. Winterbourn Pathology Department, Christchurch School of Medicine, Christchurch Hospital. Christchurch (New Zealand)

(Received 30 June 1986)

Key words: Tissue injury: Oxidant: Myeloperoxidase: Hypochloritc: (Neutrophil) The contribution of activated oxygen species to neutrophil-mediated degradation of basement membrane collagen was investigated. In preliminary experiments, pre-exposure of either albumin or glomerular basement membrane to neutrophil myeloperoxidase with H 202 and chloride increased their susceptibility to proteolysis 2-3-fold. in the basement membrane model, neutrophils are stimulated by trapped immune complexes to adhere, produce oxidants and degranulate. Degradation, measured as the amount of hydroxyproUne solubilised, was due to neutral proteinases, particularly elastase, and depended on cell number and the amount of proteinase released. Experiments with oxidant scavengers and inhibitors and with neutrophils from donors with chronic granulomatous disease or myeloperoxidase deficiency showed that oxidants did not affect degradation of the basement membrane when this was measured on a per cell basis. However, oxidative inactivation of the released granule enzymes occurred. Activities of elastase, fl-glucuronidase and lysozyme were 1.5-2-times higher in the presence of catalase, but were unaffected by superoxide dismutase or hydroxyl radical scavengers. Inactivation did not occur with chronic granulomatous disease or myeloperoxidase deficient neutrophils. When related to the activity of released elastase, or to other degranulation markers, collagen degradation was decreased in the presence of catalase, or with chronic granulomatous disease or myeloperoxidase deficient cells. This implies that the basement membrane was made more digestible by myeloperoxidase-derived oxidants, as occurred in the cell-free experiments. Taken together, the results indicate that neutrophil oxidants have two opposing effects. They increase the susceptibility of the collagen to proteolysis and inactivate the proteinases responsible.

Introduction The production of oxygen free radicals by neutrophils is essential for the microbicidal and cytotoxic function of these cells and has been implicated as a cause of tissue injury in inflamma* Present address: Pathology Department, University of Michigan, Ann Arbor, MI, U.S.A. Abbreviation: IgG, immunoglobulin G. Correspondence: Dr. C.C. Winterbourn Pathology, Department, Christchurch School of Medicine, Christchurch Hospital, Christchurch, New Zealand.

tion [1,2]. Stimulated neutrophils produce superoxide and H 2 0 z. There is less certainty as to whether they produce hydroxyl radicals [3], but the myeloperoxidase-catalysed reaction of H 2 0 2 with chloride can produce hypochlorous acid, a powerful oxidant [4,5]. HOC1 reacts readily with a wide range of biological molecules [6-8]. It can cleave peptide bonds [9], and react with amines to generate chloramines, which are themselves very good oxidants [7,10,11]. Some studies have been carried out to determine the effects of these activated oxygen species on extracellular matrix macromolecules,

0167-4889/86/$03.50 © 1986 Elsevier Science Publishers B.V. (Biomedical Division)

278

and direct effects have been observed. Oxygen radicals generated by xanthine oxidase and hypoxanthine or by stimulated neutrophils can destroy the integrity of collagen supercoils and cause depolymerisation of hyaluronic acid [12,13]. This is probably due to hydroxyl radicals formed in the presence of added ferrous ions. Hydroxyl radicals and ozone have also been shown to degrade collagen [14] and cartilage proteoglycan [15,16], and high concentrations of H202 (80 raM) can degrade mucus glycoproteins [17]. In a recent study [18], myeloperoxidase-derived oxidants caused oxidation of the lysine side chains of elastin. No direct effects of superoxide have been observed. One possible effect of oxidant attack on proteins is to make them more susceptible to proteolysis. This has been observed for collagen by Greenwald and Moy [12] and Curran et al. [14], and, more recently, for albumin [19]. Such a mechanism could allow for synergism between oxidants and proteolytic enzymes in neutrophil-mediated degradation of extracellular matrix. However, it has also been reported that neutrophils can oxidatively inactivate their own granule enzymes [20-22]. If this were to occur with the proteinases, oxidants would have the opposite effect and impair degradation. We have investigated the contribution of oxidants to neutrophil-mediated degradation using a model of inflammatory tissue injury. The extracellular matrix used as a target was glomerular basement membrane. This is an amorphous, highly insoluble matrix of Type IV collagen, the glycoproteins fibronectin and laminin, and glycosaminoglycans [23], and is susceptible to damage in vivo when immune complexes become lodged within its structure [24,25]. In our model, glomerular basement membrane is impregnated with immune complexes or immunoglobulin G (IgG) aggregates. Neutrophils adhere to its surface, produce oxidants and release hydrolytic enzymes, resulting in degradation of the matrix [26 28]. Degradation of the basement membrane collagen is entirely proteinase dependent, with elastase, gelatinase and cathepsin G contributing 60%, 30% and 10%, respectively [46]. To determine the effect of oxidants on the proteolysis, we have investigated the effect of oxidant scavengers and inhibitors and have measured degradation by neu-

trophils from donors with myeloperoxidase deficiency or chronic granulomatous disease, which are deficient in oxidant generation. Materials and Methods

Materia&. Ficoll 400 was from Pharmacia Fine Chemicals, Uppsala, Sweden, and Hypaque was from Sterling Pharmaceuticals, New Zealand. lgG was purified from normal human plasma by ammonium sulphate fractionation and ion exchange chromatography on QAE Sephadex (Pharmacia) [30]. Myeloperoxidase was purified from neutrophil granules by ion exchange chromatography on DEAE-Sephadex and CM-Sephadex [31], after removal of serine proteinases by affinity chromatography on trasylol-Sepharose [32]. All other biochemicals were from Sigma Chemical Co., MO, U.S.A. To minimize contamination with traces of metal ions, all buffers were treated with Chelex 100 resin (BioRad Laboratories, Richmond, CA, U.S.A.), and all glassware was acid-washed and rinsed in deionized water. Neutrophils. Human peripheral blood neutrophils were isolated from healthy volunteers, from one donor with chronic granulomatous disease and one with myeloperoxidase deficiency, by centrifugation through Ficoll-Hypaque, dextran sedimentation and hypotonic lysis of red cells [33]. The cells (95 97% neutrophils, 3-5% eosinophils) were finally suspended in 10 mM phosphatebuffered saline (pH 7.4) supplemented with 1 mM MgCI 2, 1 mM CaCI 2 and 1 m g / m l glucose. Cell-free medium. Neutrophils (107/ml) were stimulated at 37°C with 10 v M fMet-Leu-Phe after a 2 min preincubation with 5 /tg/ml cytochalasin B. After a further 20 min the cells were pelleted by centrifugation at 1 000 × g for 5 rain and the supernatant, containing exocytosed granule contents, was removed. This medium regularly contained 80% of the total neutrophil lysozyme and 45% of the/3-glucuronidase. Lactate dehydrogenase was not released under these conditions. Total enzyme activity was measured on a suspension of cells disrupted by sonication. Glomerular basement membrane. Basement membrane was prepared from glomeruli, isolated from histologically normal post-morten human

279

kidneys, by detergent and DNAase treatment [34]. The basement membrane was impregnated with IgG aggregates by heating a suspension of basement membrane and IgG to 63°C for 20 min, followed by extensive washing, as previously described [26]. Albumin/anti-albumin complexes were formed by reaction at room temperature with rabbit anti-human albumin (Dako Immunoglobulins, Copenhagen, Denmark) after pre-incubation of the basement membrane with human serum albumin [26].

Degradation of glomerular basement membrane containing lgG aggregates by neutrophils and cellfree medium. Neutrophils ( 1 0 7) o r 1 ml cell-free medium from 107 cells were added to 2 mg insoluble basement membrane containing IgG aggregates and incubated at 37°C for 2 h. When the effects of oxidant inhibitors and scavengers were examined, these were pre-incubated with basement membrane in 50/ul total volume for approx. 18 h at room temperature. After incubation at 37°C the cells and basement membrane were pelleted by centrifugation at 1 000 × g for 10 min. Degradation of basement membrane was determined by measuring the amount of hydroxyproline, as a marker for Type IV collagen, released into the supernatant. Hydroxyproline was measured after hydrolysis in 6 M HC1 by reaction with chloramine T and p-dimethylaminobenzaldehyde

I351. Enzyme assays. The extent of neutrophil degranulation was monitored by measuring the release of granule enzymes into the medium, l y s o zyme was measured by monitoring the lysis of Micrococcus lysodeikticus [36], and myeloperoxidase using o-tolidine as the electron donor [37]. 1 unit of lysozyme or myeloperoxidase activity caused an absorbance change of 0.001/min at 20°C. /~-Ghicuronidase was measured using phenolphthalein glucuronic acid [371. 1 unit of enzyme activity liberated 1/~g phenolphthalein in 18 h. These enzymes did not adhere to glomerular basement membrane and were measured in the supernatant at the end of the 2 h incubation. Elastase activity was measured using the specific peptide chromophore substrate, methoxysuccinylAla-Ala-Pro-Val-p-nitroanilide. Elastase released from neutrophils in the presence of basement membrane could not be measured reliably in the

supernatant since approx. 80% of the activity was bound to the glomerular basement membrane, a phenomenon also noted by Davies et al. [38]. Therefore, elastase activity was measured by adding 100 F1 of the whole cell and basement membrane suspension to the substrate (0.32 mM) in phosphate-buffered saline (pH 7.4) and the reaction was monitored continuously at 400 nm. 1 unit of elastase activity caused an absorbance change of 0.001/min. This method measures only extracellular elastase, as no activity was observed in the presence of unstimulated cells. To determine the efficiency of elastase measurement in the presence of 2 m g / m l basement membrane, elastase-containing supernatant from 1 0 7 stimulated neutrophils was added. With the addition of 90 and 300 units of elastase, the assay measured 40% and 60% of the original activity, respectively (n = 3). Elastase release in the intact cell experiments was generally within this range, so errors introduced by variations in the proportion of undetectable enzyme are not likely to be large. Elastase activities are expressed as measured, without attempting to correct for recovery. A time course study showed that elastase release was maximal after 30 min, and that degradation was linear over the 2 h incubation period (Fig. 1). Therefore, the mean value of elastase activities measured at 30 min and 1 h was taken as the amount of elastase available.

~0

200

.;;8

c v

o

120

o

40

Jc

L,U

0

3'o

go

9'o

1~o

Incubation time (rain)

Fig. 1. Time course of elastase release (solid lines) and degradation of glomerular basement membrane (dotted lines). Neutrophils were incubated with basement membrane in the presence ( O ) and absence (O) of catalase. The results shown are from one experiment. Similar findings were obtained on five other occasions.

280

Proteolysis of glomerular basement membrane and bovine serum albumin after exposure to oxidants. Glomerular basement membrane (0.75 rag) was suspended in 0.5 ml phosphate-buffered saline (pH 7.4) and incubated with either 50 mM H 2 0 2 , or myeloperoxidase (6000 units) and H202 (10 additions of 0.1 ~mol at 5 min intervals). After 1 h, or 10 min after the last addition of H 2 0 2 , the basement membrane was washed twice in 2 ml phosphate-buffered saline and 0.8 ml cell-free medium (from 8-106 cells) was added. After a further 1.5 h at 37°C the basement membrane was pelleted and hydroxyproline measured in the supernatant. Albumin (0.5 mg) was dissolved in 0.5 ml phosphate-buffered saline (pH 7.4) and pre-incubated with H 2 0 2 o r myeloperoxidase and H202 under the same conditions for basement membrane. Catalase (400 units) was then added to scavenge any remaining H202. After 15 min the volume was adjusted to 0.9 ml with phosphate-buffered saline and 20 ~g trypsin added. After 1.5 h the protein was precipitated by adding 0.1 ml 100% (w/v) trichloroacetic acid and the soluble peptides measured in the supernatant by the Folin modification of the Lowry method [39]. Inactivation of catalase with 3-amino-l,2,4,triazole. Inactive catalase-aminotriazole complex was formed by overnight dialysis of catalase in 0.1 M aminotriazole against 4 mM H202 [40]. Free

aminotriazole was removed by extensive dialysis against water. Catalase was inhibited by 98% using this method. Results

The effect of oxidants on proteolysis in a cell-free system To determine whether neutrophil-derived oxidants can render proteins more susceptible to proteolysis, glomerular basement membrane and albumin were exposed t o H 2 0 2 alone and to the myeloperoxidase-H2Oz-chloride system. We measured the digestion of albumin by trypsin as a general model for this mechanism, and of basement membrane by neutrophil enzymes because of its application to our experimental system. In both cases, pretreatment with high concentrations of H 2 0 2 had no effect on subsequent proteolysis (Table I). By contrast, pre-exposure to myeloperoxidase and H 2 O 2 resulted in a 3-fold increase in the digestion of basement membrane and doubled the proteolysis of albumin. This effect was largely inhibited by 1 mM azide (Table I). Very little e n h a n c e m e n t of proteolysis was seen if myeloperoxidase was added to the basement membrane without pre-incubation. This is possibly because it is necessary for myeloperoxidase to penetrate the matrix, ensuring that HOCI is formed in close proximity to the target protein. Only

TABLE I THE EFFECT OF MYELOPEROXIDASE-DERIVED OXIDANTS AND H202 ALONE ON THE SUBSEQUENT PROTEOLYSIS OF GLOMERULAR BASEMENT MEMBRANE OR ALBUMIN In controls, 20 /Lg trypsin degraded 38_+ 2 /zg albumin in 1 h, measured as trichloroacetic acid-solublc protein and hydroxyproline solubilised from basement memrbane by neutrophil cell-free granule proteinases was 8.9-+ 2.6 ttg in 1 h. Results are means-+ S.D., with the number of experiments in parentheses. Pre-incubation conditions

Proteolysis (percent of control) albumin

U202 (50 mM) Myeloperoxidase + H 202 ( 10 additions of 200 p~M/l H202) Myeloperoxidase + H 202 + azide (1 mM) Myeloperoxidase Control blank (no trypsin or neutrophil proteinases added) Myeloperoxidase + H 202 blank (no trypsin or neutrophil proteinases added)

basement membranc

90 ± 8 (2)

99 + 5

(4)

207 +31

(6)

296 ±88

(4)

133 ± 8 113 ±24

(4) (2)

106 +55 89 +13

(2) (2)

2.6_+ 0.5(3)

3.74__ 2.2(5)

5.8± 0.5 (4)

2

± 3 (4)

281

slight release of trichloroacetic acid-soluble albumin fragments and no basement membrane solubilization due to myeloperoxidase-derived oxidants alone was seen, and myeloperoxidase without H 2 0 2 did not cause an increase in proteolysis, indicating that the effect was not due to a proteinase contaminant (Table I).

Degradation of glomerular basement membrane by neutrophils In agreement with our previous observations [26-28], neutrophils adhered to basement membrane containing IgG aggregates, degranulated and proteolytically degraded the basement membrane collagen. The amount of degradation was dependent on the number of neutrophils added and reflected the amount of granule enzyme released (Fig. 2). Hydroxyproline solubilization by 10 v neutrophils was 15.7 + 5.3 # g / 2 h per 2 mg basement membrane (n = 57) representing approx. 10% of the total collagen. The extent of neutrophil degranulation and collagen degradation varied between batches of basement membrane and probably reflects the amount of immune stimulus incorporated into the matrix. When albumin-anti-albumin complexes were used instead of IgG aggregates degranulation was approximately the same, but degradation was consistently 60% less. This was an effect of heating to 63°C to aggregate the IgG, since neutrophil supernatant (from 10 v cells) solubilized 20.9 + 5.8 (11 determinations) ~g

600

30

/////J/~

N__ D

Jx "/ ///Jr"

20

//t

if

10

400

/ ~

~ ~o

--

200

.C //

0... I

1

,.,0" ~

~

2

3

cell no x 10 -7

Fig. 2. The effect of increasing cell number on enzyme release and degradation of glomerular basement membrane. Hydroxyproline solubilized (e e); elastase ( e . . . . . . e), Bglucuronidase (© . . . . . . ©); lysozyme (zx. . . . . . ix).

hydroxyproline/2 h from heated basement membrane, compared with 9.8 + 2.9 ( n - - 6 ) from unheated membrane. The presence of IgG in the heated membrane had no effect (21.7 +_ 7.8 ~g hydroxyproline solubilized). To ensure that heating did not introduce artefacts into the oxidant experiments, both IgG aggregates and albuminanti-albumin complexes were used.

The effect of oxidant inhibitors and scavengers on degradation Before incubation with neutrophils, basement membrane was pre-incubated with inhibitors or scavengers to allow these to gain access to the site of oxidant production at the interface between the cell and the target. Catalase, superoxide dismutase, azide to inhibit myeloperoxidase, the hydroxyl radical scavengers mannitol and benzoate, and the iron chelator desferrioxamine, were used. Their effects on basement membrane degradation are shown in Table II. The amount of hydroxyproT A B L E II THE EFFECT OF OXIDANT INHIBITORS AND SCAVENGERS ON THE DEGRADATION OF G L O M E R U L A R BASEMENT M E M B R A N E C O N T A I N I N G IgG A G G R E G A T E S BY N E U T R O P H I L S Basement membrane was pre-incubated overnight at room temperature with inhibitors and scavengers in 50 ~l total volume. The final concentrations were as follows: catalase, 716 ~ g / m l (28,640 units/ml); catalase-aminotriazole complex, 716 ~ g / m l ; superoxide dismutase, 1 6 7 / t g / m l ; azide, 1 raM; mannitol, 25 raM; benzoate, 25 raM; desferrioxamine, 50 ~M. The means_+S.D, are shown with the number of experiments in parentheses. Significance was determined with the Student's t-test using the mean difference of paired observations. Addition to neutrophils and basement membrane

Hydroxyproline solubilised/107 cells (/L g) (percent of control)

Catalase Catalase b Catalase + superoxide dismutase Superoxide dismutase Inactive catalase- amino t riazole complex Azide Mannitol Benzoate Desferrioxamine

93 + 93 + 93 + 100 +

22 (24) 10 (5) 22 (14) 16 (16)

99_+ 10 113 + 24 9 9 + 17 97+ 12 104-+ 12

(6) (12) (6) (6) (5)

a p < 0.02; others not significant. b Results obtained with basement membrane containing albumin/anti-albumin.

282

line solubilized on a per cell basis was slightly lower in the presence of catalase, but was not significantly altered with any of the other inhibitors. A similar result was obtained with catalase when basement membrane contained albuminanti-albumin complexes (Table II). Degradation was linear throughout the incubation time, and at all time points, there were minimal differences in digestion in the presence or absence of catalase (Fig. 1). In the presence of catalase, with or without superoxide dismutase, the activities of the released enzymes were increased (Table III). Activities of /?-glucuronidase and lysozyme were doubled and 50% more elastase was measured A small increase in elastase and lysozyme were seen in the presence of azide, but the enzyme levels were essentially unchanged with the other scavengers. Thus, although hydroxyproline solubilization was almost unaffected by the presence of catalase, the activities of released enzymes were considerably elevated. Basement membrane degradation by neutrophil proteinases under these conditions is proportional to the amount of enzyme present (Fig. 1) [46]. Hence, if the susceptibility of the membrane proteins is unaffected by neutrophil oxidants, the efficiency of degradation should depend on the amount of enzyme available. Since

most of the basement membrane degradation is due to elastase, the amount of hydroxyproline solubilized has been related to the elastase activity released from the cells. However, in view of the approximations in the elastase assay, and because other proteinases (gelatinase and cathepsin G) are also involved, digestion has also been related to neutrophil degranulation and for this purpose the marker enzymes /~-glucuronidase and lysozyme were used as reference. Table IV shows that the efficiency of degradation per unit elastase was decreased by one third in the presence of catalase, with or without superoxide dismutase. Also, when related to/~-glucuronidase or lysozyme activity, degradation when catalase was present was less than one half of control values. Superoxide dismutase alone, azide and the hydroxyl radical scavengers all had no effect. The effect of catalase was dependent on its enzyme activity, as shown by experiments carried out with an equivalent amount of inactive catalase-aminotriazole complex. Tables II, III, and IV show that inactive catalase had a minimal effect on degranulation, causing a small increase in flglucuronidase activity only, and no effect on degradation. Catalase added to basement membrane did not interfere with proteolysis: in three experi-

TABLE III T H E E F F E C T OF I N H I B I T O R S A N D SCAVENGERS OF O X Y G E N R A D I C A L S ON D E ( i R A N U L A T I O N BY NEUTROP HIE S S T I M U L A T E D ON G L O M E R U L A R BASEMENT M E M B R A N E C O N T A I N I N G Ig(i A G G R E G A T E S . Inhibitor and scavenger concentrations are given in Table III. Mean results+ S.D. of (n) experiments are shown. Control values (units/107 cells) were /~-glucuronidase 44_+ 12 (15); lysozyme 67_+28 (13) and elastase 192± 70 (15). Significance was determined using the mean difference of paired observations Addition

Catalase Catalasc* Catalase+superoxidedismutase Superoxide dismutase Inactive catalase-aminotriazole complex Azide Mannitol Benzoate Desferrioxamine

% of Control fl-glucuronidase

lysozyme

elastase

211 + 48 (10) " 160_+ 5 (3)" 213+60 (6) b 123_+ 14 (6) b 131 + 13 (4) d 115+32 (6) 125+ 8 (2) 103_+ 2 (2) 115+ 9 (2)

178 ± 37 170+26 181+27 107 + 26 1 2 0 ! 25 133+ 22 99+ 6 110+ 4 117+ 2

158 + 32 144+11 155+ 42 109+ 7 I01 + 17

(12) ~ (3)b (7) a (5) (4)

(7) b

(6) " (4)

(5) d

117+

(2) (2) (2)

9 4 + 1 0 (3) 110:~31 (3) 1084 8 (3)

" P < 0.001: b p < 0.01; c p < 0.02; d p < 0.05: others not significant. * Results obtained with basement membrane containing a l b u m i n / a n t i - a l b u m i n complexes.

8

(10) ~ (3) "

(7) d

283

TABLE IV T H E EFFECT OF O X I D A N T INHIBITORS A N D SCAVENGERS ON THE D E G R A D A T I O N OF G L O M E R U L A R BASEMENT MEMBRANE PER U N I T ACTIVE ENZYME Degradation of glomerular basement membrane (Table II) was related to the amount of active enzyme released (Table III). The means_+ S.D. are shown with the number of experiments in parentheses. Significance was determined by the Student's t-test using the mean difference of paired observations. Addition

Catalase Catalase* Catalase + superoxide dismutase Superoxide dismutase Inactive catalase-aminotriazole complex Azide Mannitol Benzoate Desferrioxamine

Hydroxyproline solubilised per unit active enzyme, (percent control, ~ g) elastase

/~-glucuronidase

lysozyme

68 + 11 (9) a 65_+ 3(3) " 66 + 12 (6) ~ 98 + 15 (6) 99 -+ 8 (3) 100 + 22 (6) 111 _+ 11 (3) 104+20(4) 104 + 15 (3)

39 + 17 (7) 55_+ 6(3) 40 + 15 (6) 93 + 15 (5) 80 + 13 (4) 104 + 19 (5) 92 + 17 (2) 90+ 3(2) 93 + 17 (2)

54 + 14 (8) a 59_+ 9(3) b 56 + 11 (6) " 100 + 17 (6) 89 _+24 (4) 92 + 17 (5) 96 _+14 (2) 103_+ 2(2) 92 + 10 (2)

" ~ ~ c

P < 0.001' b p < 0.01' c p < 0.05; others not significant. * Results obtained with basement membrane containing albumin/anti-albumin complexes.

ments, cell-free medium solubilized 18.3 _+ 0.4 /~g hydroxyproline/2 h from basement membrane in the presence of 716/~g/ml catalase, and 19.8 _+ 2.0 /~g/2 h in its absence.

Degradation by neutrophils from chronic granulomatous disease and myeloperoxidase-deficient donors On a per cell basis, chronic granulomatous disease neutrophils solubilized 1.34 +_ 0.05 times more hydroxyproline than did control cells when incubated with basement membrane containing albumin/anti-albumin (one experiment) or IgG (three experiments). Enzyme release (measured on one occasion and not repeated because further cells were unavailable) was consistently higher with chronic granulomatous disease neutrophils with 3.2-times as much active elastase, 2.3-times more B-glucuronidase, 2.5-times more lysozyme and 1.9-times more myeloperoxidase. Thus, when related to either the amount of elastase, or to degranulation, degradation was approximately half that with normal neutrophils. Two experiments were carried out with myeloperoxidase-deficient neutrophils, which contained 8-14% of normal myeloperoxidase activity as measured in a sonicated cell sample. After incubating with basement membrane containing IgG ( 1 0 7 cells with 2 mg basement membrane)

myeloperoxidase activity was undetectable in the medium. In both experiments, myeloperoxidasedeficient neutrophils released considerably more active elastase,/~-glucuronidase and lysozyme than did the control cells (Table V). In one experiment, the myeloperoxidase-deficient cells solubilized more hydroxyproline than the control neutrophils, while in the second, degradation by both cell populations was the same (Table V). However, when related to active enzyme release, myeloperoxidase-deficient cells degraded basement membrane less efficiently than did normal neutrophils. The effects of adding myeloperoxidase and catalase to the myeloperoxidase-deficient cells are shown in Table V. In contrast to normal neutrophils, enzyme release by myeloperoxidase-deficient cells was unaffected by catalase. Hence there was no hydrogen peroxide-dependent enzyme inactivation in the absence of myeloperoxidase, and since hydroxyproline solubilization was the same in the presence and absence of catalase, the results imply that H 2 0 2 did not alter the susceptibility of the membrane to proteolysis. When purified myeloperoxidase was added to the deficient neutrophils, the activities of the released enzyme were decreased to approximately one half, and approached the levels of activity released from normal cells. The amount of degradation was unchanged.

284 TABLE V D E G R A N U L A T I O N A N D D E G R A D A T I O N OF G L O M E R U L A R BASEMENT M E M B R A N E BY MYELOPEROXIDASE-DEFICIENT NEUTROPHILS Neutrophils were incubated with basement containing IgG aggregates for 2 h. Catalase (716 p~g/ml) or myeloperoxidase (7 500 units, approx. 70 nmol/1) was pre-incubated overnight with basement membrane, before addition of the cells. The results shown (means _+S.D. of duplicate estimations) are from two separate experiments, n.d., not determined. Conditions

Units released

Hydroxyproline

B-glucuronidase

lysozyme

elastase

( p"g)

Expt. 1 Myeloperoxidase-deficient neutrophils +catalase Normal neutrophils + catalase

n.d. n.d. n.d. n.d.

185_+ 7 180+ 0 120 200

165+20 155+ 7 80 150

10.5 9.0+0.7 10.0 8.5

Expt. 2 Myeloperoxidase-deficient neutrophils + catalase + myeloperoxidase Nornlal neutrophils + catalase

150+17 147 _+ 1 94 + 6 37+ 8 111

180+14 180 + 14 110 + 0 60+ 0 170

515± 9 540 270 ±_56 210 385+ 7

23.3 23.8 ± 0.4 21.5 -+ 0.7 14.5 16.5

Discussion

Our studies of the effects of oxidants on glomerular basement membrane and on albumin agree with those of others [12,14,19] in showing that oxidant treatment can render proteins more susceptible to proteolysis. Previous studies have been with hydroxyl radicals [12,14,19]. Our findings show that a similar effect can be achieved with myeloperoxidase-derived oxidants (HOCI) and that H202 alone is ineffective. These findings, therefore, lend credence to the proposal that neutrophil oxidants could contribute to inflammatory tissue damage by increasing the proteolytic susceptibility of structural proteins. Our findings also support and extend previous observations that neutrophils inactivate their own granule enzymes [20-22]. These studies showed that the oxidant effect is on enzyme activity rather than enzyme release. We have shown that inactivation occurs when the cells are stimulated by surface-bound IgG aggregates or albumin/anti-albumin immune complexes, and that elastase as well as fl-glucuronidase and lysozyme is affected. Protection by catalase and our findings with myeloperoxidase-deficient neutrophils imply that inactivation is myeloperoxidase-dependent, and this is in agreement with the other studies. In a

recent study (unpublished observations), we have observed that the two other proteinases which contribute to basement membrane collagen degradation, gelatinase and cathepsin G, are also inactivated by myeloperoxidase-derived oxidants. Therefore, on the basis of these findings, we might expect oxidants to inhibit membrane degradation. In fact, we found that oxidant scavengers had no effect on hydroxyproline solubilization by a given number of neutrophils. Also, neutrophils from donors with chronic granulomatous disease or myeloperoxidase deficiency degraded basement membrane as efficiently as normal cells. This is in broad agreement with other studies on neutrophilmediated degradation of glomerular basement membrane coated with immune complexes [41], or of subendothelial matrices by phorbol myristate acetate-stimulated neutrophils [42]. These studies, however, did not assess the effect of oxidants on the neutrophil granule enzymes. Since our experiments were carried out under conditions where digestion depended on proteinase activity, our findings support the proposition that neutrophil oxidants have the two opposing effects for which we provide evidence above, and these largely cancel out. They inactivate released proteinases, but because this does not cause a decrease in digestion, they must also increase the

285

digestibility of the protein substrate. This is apparent when digestion is related to active enzyme release, regardless of whether elastase or one of the degranulation markers is used as a reference. On this basis, H202 removal by catalase, with or without superoxide dismutase, consistently decreased digestion to two thirds or less of that seen with neutrophils alone. This, combined with the findings with chronic granulomatous disease neutrophils, implies an essential role for H 2 0 2 in increasing the susceptibility to degradation. There was no evidence for the involvement of superoxide or hydroxyl radicals, since none of the other scavengers, or the iron chelator desferrioxamine, had any effect. Hydroxyl radicals, if formed, would be expected to react with both the basement membrane and granule enzymes, and the lack of effect is probably due to their not being produced under our conditions. The importance of the myeloperoxidase system rather than H202 alone was demonstrated by adding myeloperoxidase or catalase to myeloperoxidase-deficient neutrophils. Catalase did not affect enzyme release or digestion, whereas added myeloperoxidase decreased the released activities of elastase, lysozyme and/3-glucuronidase, with little effect on the amount of degradation. In experiments with normal neutrophils, however, inhibition of myeloperoxidase by azide did not significantly alter the rate of degradation either on a per cell basis or relative to degranulation. The reasons for this are unclear. Myeloperoxidase appeared to be fully inhibited by azide, since it was undetectable in the medium. However, azide can itself form reactive radicals [43] and if these were produced, they could negate the effect of inhibiting myeloperoxidase-derived oxidants. There are two other possible interpretations of the oxidant effect on basement membrane degradation. Neutrophil oxidants can activate the metalloenzyme collagenase [44] and such a mechanism could enhance degradation. However, collagenase does not degrade Type IV collagen [45,46]. Furthermore, collagenase, gelatinase and cathepsin G are more susceptible than elastase to inactivation by neutrophil oxidants (Vissers and Winterbourn, manuscript in preparation). Secondly, if the elastase inhibitor aa-antitrypsin were present, it could inhibit elastolytic digestion, and oxidative inactivation of the al-antitrypsin could

prevent this [26,29]. However, al-antitrypsin was undetectable by immunoelectrophoresis in collagenase-solubilized basement membrane (lower limit of detection = 1 / l g / m g in original sample). Both explanations, therefore, are unlikely. We conclude that myeloperoxidase-derived oxidants produced by neutrophils have two opposing effects on the degradation of extracellular matrix components by these cells. Although they can render proteins more susceptible to proteinases released from the cells, this is a small effect and largely offset by oxidative inactivation of the enzymes themselves. Oxidative damage in vivo, however, may enhance the subsequent degradation by non-neutrophil proteinases. Our results emphasize that neutrophil oxidants are non-discriminatory in their reactions, and also illustrate the potential pitfalls in extrapolating results obtained with oxidant-generating systems and purified proteins to cellular systems.

Acknowledgements We are grateful to Sue Townsend for excellent technical assistance. This work was supported by the Medical Research Council of New Zealand.

References 1 Weiss, S.J. and LoBuglio, A.F. (1982) Lab. Invest. 47, 959-966 2 Babior, B.M. (1984) Blood 64, 959-966 3 Tauber, A.I. and Babior, B.M. (1985) Adv. Free Rad. Biol. Med. 1,265-308 4 Klebanoff, S.J. (1980) Ann. Int. Med. 93,480-489 5 Thomas, E.L. (1979) Infect. Immun. 23, 522-531 6 Zgliczynski, J.M. and Stelmaszynska, T. (1975) Eur. J. Biochem. 56, 157-162 7 Thomas, E.L., Jefferson, M.M. and Grisham, M.B. (1982) Biochemistry 21, 6299-6308 8 Winterbourn, C.C. (1985) Biochim. Biophys. Acta 840, 204-210 9 Thomas, E.L. (1979) Infect. Immun. 25, 110-116 10 Stelmaszynska, T. and Zgliczynski, J.M. (1984) Eur. J. Biochem. 45, 305-312 11 Weiss, S.J., Klein, R., Slivka, A. and Wei, S. (1982) J. Clin. Invest. 70, 598-607 12 Greenwald, R.A. and Moy, W.W. (1979) Arthr. Rheum. 22, 251-259 13 Greenwald, R.A. and Moy, W.W. (1980) Arthr. Rheum. 23, 455-463 14 Curran, S.F., Amoruso, M.A., Goldstein, B.D. and Berg, R.A. (1984) FEBS Lett. 176, 155-160

286 15 Bates, E.J., Harper, G.S., Lowther, D.A. and Preston, B.N. (1984) Biochem. Int. 8, 627-637 16 Dean, R.T., Roberts, C.R. and Forni, L.G. (1984) Biosci. Rep. 4, 1017 1026 17 Creeth, J.M., Cooper, B., Donald, A.S.R. and Clamp, J.R. (1983) Biochem. J. 211,323-332 18 Clark, R.A., Szot, S., Williams, M.A. and Kagan, H.M. (1986) Biochem. Biophys. Res. Commun. 135, 451 457 19 Wolff, S.P. and Dean, R.T. (1986) Biochem. J. 234, 399-403 20 Voetman, A.A., Weening, R.S., Hamers, M.N., Meerhof, LJ.. Bot, A.A.M. and Roos, D. (1981) J. Clin. Invest. 67, 1541 1549 21 Clark, R.A. and Borregaard, N. (1985) Blood 65, 375-381 22 Dri, P., Cramer, R., Menegazzi, R. and Patriarca, P. (1985) Br. J. Haematol. 59, 115-125 23 Timpl, R., Oberb~iumer, I., Furthmayr, H. and Kuehn, K. (1981) in Trends in Basement Membrane Research (Kuehn, K., Schoene, H.-H. and Timpl, R., eds.), pp. 57-67, Raven Press, New York 24 Cochrane, C.G., Unanue, E.R. and Dixon, F.J. (1965) J. Exp. Med. 122, 99 117 25 Henson, P.M., Hollister, J.R., Musson, R.A., Webster, R.O., Spears, P., Henson, J.E. and McCarthy. K.M. (1979) Advan. Inflamm. Res. 1, 34t-352 26 Vissers, M.C.M., Winterbourn, C.C. and Hunt, J.S. (1984) Biochim. Biophys. Acta 804, 154-160 27 Vissers, M.C.M., Day, W.A. and Winterbourn, C.C. (1985) Blood 66, 161-166 28 George, P.M., Vissers, M.C.M., Travis, J., Winterbourn, C.C. and Carrell, R.W. (1984) Lancet ii, 1426-1428 29 Clark, R.A., Stone, P.J., El Hag, A., Calore, J.D. and Franzblau, C. (1981) J. Biol. Chem. 256, 3348-3353 30 Hunt, J.S. and McGiven, A.R. (1980) in Immunological

31 32 33 34 35 36 37 38 39 4(t 41 42 43 44 45 46

Investigation of Renal Disease (McGiven, A.R., ed.), pp 58-59, Churchill Livingston, Edinburgh Matheson, N.R., Wong, P.S. and Travis, J. (1981) Biochem ist~ 20, 325-330 Barrett, A.J. (1981) Methods Enzymol. 80, 581 588 B6yum, A. (1968) Scand. J. Clin. Invest. 21 (Suppl. 97) 77-89 Carlson, E.C., Brendel, K., Hjelle, J.Y. and Meezan, E. (1978) J. Ultrastruct. Res. 62, 26-53 Gordeladze, J.O., Halse, J., Dj6seland, O. and Haugen. H.N. (1978) Biochem. Med. 20, 23 3(/ Worthington Enzyme Manual (1979) Millipore Corporation, Bedford, MA, USA Baggiolini, M., Hirsch, J.G. and De Duve, (7. (1969) J. Cell Biol. 40, 529 541 Davies, M., Hughes, K.T. and Thomas, G.J. (1980) Renal Physiol. (Basel) 3, 112 115 Lowry, O.H.. Rosebrough, N.J., Farr, A.L. and Randall. R.J. (1951) J. Biol. Chem. 193, 265 275 Margoliash, E. and Novogradsky, A. (1958) Biochem. J. 68. 468-475 Bray, J., Hume, D.A. and Robinson, G.B. (1983) Mol. Biol. Med. 1,253-269 Weiss, S.J. and Regiani, S. (1984) J. Clin. Invest. 73. 1297 1303 Butler, J., Land, E.J.. Swallow, A.J. and Prutz, W. (1984) Radiat. Phys. Chem. 23, 265 270 Weiss, S.J., Peppin, G., Ortiz, X., Ragsdale, C. and Test, S.T. (1985) Science 227, 747-749 Murphy, G., Reynolds, J.J., Bretz, U. and Baggiolini, M. (1982) Biochem. J. 203, 209-221 Vissers, M.C.M. and Winterbourne, C.C. (1986) PhD Thesis, University of Otago