Activation of human platelets by C5a-stimulated neutrophils: a role for cathepsin G.

Activation of human platelets by C5a-stimulated neutrophils: a role for cathepsin G PABLO FERRER-LOPEZ, PATRICIA RENESTO, MIRTA SCHATTNER, SYLVIANE BASSOT, PHILIPPE LAURENT, AND MICHEL CHIGNARD Unit& de Pharmacologic Cellulaire, Unit& Associie IP-Institut National de la Santk et de la Recherche Midicale 285, Institut Pasteur, 75724 Paris Cedex 15; and Unit6 de Biopathologie et Toxicologic Pulmonaire et Rkale, Institut National de la Santk et de la Recherche Mkdicale 139, CHU Henri Mondor, 94010 C&e& France FERRER-LOPEZ, PABLO, PATRICIA RENESTO, MIRTA SCHATTNER, SYLVIANE BASSOT, PHILIPPE LAURENT, AND MICHEL CHIGNARD. Activation of human platelets by C5astimulated neutrophils: a role for cathepsin G. Am. J. Physiol. 258 (Cell Physiol. 27): CllOO-C1107, 1990.-Human platelets

can be stimulated by recombinant human fifth component of complement (rhC5a) in the presence of human neutrophils. After challenge with IV-formyl-Met-Leu-Phe or rhC5a, concentrated neutrophils release cathepsin G into the supernatant. The concentrations of cathepsin G recovered by titration of the enzymatic activity correlate with the capability of these supernatants to induce platelet stimulation as measured by serotonin release. Cathepsin G purified from neutrophil granules triggered platelet aggregation and serotonin release independent of arachidonic acid metabolites and platelet-activating factor formation. A concentration of 100 nM of cathepsin G, which was reached in the surrounding space of activated neutrophils, induced a 50% platelet stimulation. Three distinct antiproteinases were tested against cathepsin G-induced platelet activation. Z-Gly-Leu-Phe-CH&l, a specific inhibitor of cathepsin G enzymatic activity, proved to be nonspecific in our biological system. By contrast, aI-antichymotrypsin and aI-antitrypsin displayed specific activities. The physiological specific inhibitor of cathepsin G, aI-antichymotrypsin, was the most potent and was used in the rhC5a-induced neutrophils-mediated platelet activation. A complete inhibition was achieved, showing that release of cathepsin G from neutrophils accounts for platelet activation. Such a chain of events involving C5a, neutrophils, cathepsin G, and platelets may be of relevance in certain inflammatory states, particularly the adult respiratory distress syndrome. cell-cell interaction; serine proteinase; antiproteinase; respiratory distress syndrome

adult

AFTER ADDITION of N-formyl-Met-Leu-Phe (fMLP), human platelets are activated when polymorphonuclear neutrophils (PMN) are present in the surrounding space (5). This is due to a cell-to-cell communication; fMLP stimulates human PMN, which in turn activate nearby platelets. This peculiar in vitro phenomenon where platelets are stimulated through activated PMN is reminiscent of different experimental acute inflammatory states that can be evoked in laboratory animals. Thus platelet deposition at skin sites occurs when fMLP is injected intradermally in rabbits but is absent in PMN-depleted Cl100

platelet-intact animals (14). It has also been observed in a PMN-mediated model of subendothelial immune complex glomerulonephritis that platelets mediate glomerular injury and proteinuria (15). In models of adult respiratory distress syndrome (ARDS), a pathology characterized by loss of integrity of the alveolar-capillary wall, not only are PMN involved (25,29) but platelets are also suspected to participate (27). In humans, however, it is unknown whether platelets are bystanders or culprits in ARDS (12). Since formation of the activated fifth component of complement (C5a) is frequently associated with this disease in humans (11, 22) and that in vivo complement activation (26) or C5a administration lead to acute lung injury in animals (13,24), we looked for an indirect effect of human recombinant C5a (rhC5a) on human platelets due to activation of PMN in vitro. In vitro cooperation between PMN and platelets has been observed in rabbit cells after stimulation with fMLP, and platelet-activating factor (PAF)-acether has been identified as the mediator involved (6,19). However, this is probably not the case in human cells. Indeed, concentrated human PMN, in the presence of fMLP, release in the extracellular medium an aggregating substance the activity of which was not inhibited by a specific PAF-acether antagonist (5). This substance, which resembles PAF-acether in that it induces calcium mobilization and serotonin release, has been shown to be a cationic protein with chymotrypsin-like enzymatic activity. Further studies demonstrated that the aggregating molecule present in the concentrated supernatant fractions of fMLP-activated human PMN was cathepsin G (23), a neutral proteinase stored in their azurophilic granules (7). However, it remains to be proved whether cathepsin G plays a role under conditions when both cell populations are present at physiological concentrations in the same environment. In the present paper, we present evidence that cathepsin G is the mediator involved in the PMN-to-platelets communication as far as human cells are concerned. Furthermore, we demonstrate that the cooperation between PMN and platelets is not confined to fMLP but also extends to rhC5a. MATERIALS

AND

METHODS

Materials. Dr. H. Showell (Pfizer, Groton, CT) kindly provided us with rhCSa, obtained from Escherichia coli

0363-6143/90 $1.50 Copyright 0 1990 the American Physiological

Society

Downloaded from www.physiology.org/journal/ajpcell by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on August 30, 2018. Copyright © 1990 American Physiological Society. All rights reserved.

RHC~A-INDUCED

NEUTROPHIL-MEDIATED

by recombinant DNA techniques. The amino acid composition of rhC5a is indistinguishable from that of authentic human C5a, and there is equal potency between the natural and bacterial source (9). Compound WEB 2086 was kindly given by Dr. H. Heuer (Boehringer Ingelheim, Ingelheim, FRG), prostacyclin by Dr. S. Moncada (Wellcome Research Laboratories, Langley Court, UK), and active site-titrated al-antitrypsin by Drs. C. Boudier and J. Bieth (INSERM U 237, Strasbourg, France). Fibrinogen (grade L) from Kabi (Stockholm, Sweden) was treated with diisopropyl fluorophosphate (DFP) to inactivate coagulant contaminants. The anticoagulant, acid-citrate-dextrose (ACD), was composed of 85 mM trisodium citrate, 66 mM citric acid, and 110 mM glucose. All the other reagents were bought from different manufacturing companies: from Sigma Chemical (St. Louis, MO), N-succinyl-Ala-Ala-Pro-Phe-p-nitroanilide, N-succinyl-Ala-Ala-Val-p-nitroanilide, N-&hydroxyethylpiperazine-N’+ethanesulfonic acid (HEPES), EDTA disodium salt dihydrate, dextran; from Pharmacia (Uppsala, Sweden), Ficoll-Paque, Phast gel gradient lo15, sodium dodecyl sulfate (SDS) buffer strips; from GIBCO (Paisley, Scotland), Hanks’ balanced salt solution; from Merck (Darmstadt, FRG), tris(hydroxymethyl)aminomethane (Tris), 4-nitrophenyl-@-D-glucopyranosiduronic acid; from Amersham (Amersham, UK), 5-hydroxy- [side chain-2-14C] tryptamine creatinine sulfate (55 mCi/mmol), aqueous counting scintillant (ACS) II; from Enzyme Systems Products (Livermore, CA), ZGly-Leu-Phe-CH&l; from Calbiochem Behring Diagnostics (La Jolla, CA), al-antichymotrypsin (99% pure); from Synthelabo (Paris, France), acetylsalicylic acid (ASA), lysine salt; from Choay (Paris, France), heparin sodium. Preparation of washed human platelets. Blood was collected from healthy human volunteers over ACD (1.5 ml for 10 ml of blood) and heparin (20 IU/ml) as anticoagulant. Platelet-rich plasma was prepared by centrifugation of blood at 180 g for 20 min and was then processed at 37*C. Platelet-rich plasma was first incubated with [ 14C]serotonin (1 PM final concentration) for 30 min and then mixed with prostacyclin and heparin (10 pM and 5 IU/ml final concentration, respectively) just before centrifugation at 2,500 g for 15 min. Plasma was discarded, and the resulting pellet was resuspended in an equal volume of Tyrode buffer kept at 37°C (buffer composition in mM: 137 NaCl, 2.68 KCl, 1.0 MgCl, 2.0 CaC12, 11.9 NaHC03, 0.42 NaH2P04, 5.54 glucose, 5.0 HEPES, supplemented with 0.35% bovine serum albumin). Prostacyclin was again added at the same final concentration, and the suspension was centrifuged at 1,600 g for 15 min. The pellet was finally resuspended in the Tyrode buffer to achieve a final cell concentration of 4 X 108/ml. Platelets were kept at 37°C throughout the entire experiment. Preparation of purified human neutrophils. An aliquot of blood from the same donor was mixed with 0.5 vol of 3% dextran in saline, and erythrocytes were allowed to sediment at room temperature for 30 min. The upper phase enriched in leukocytes was collected, and 2 vol gently layered over 1 vol. of Ficoll-Paque. After centrif-

PLATELET

ACTIVATION

Cl101

ugation at 350 g for 45 min at room temperature, the whole aqueous phase was removed and the cell pellet was resuspended in an erythrocyte lysis buffer (composition in mM: 155 NH4Cl, 2.96 KHCOS, and 3.72 EDTA) for 5 min. The suspension was centrifuged at 350 g for 10 min, and the cell pellet was washed in excess Hanks’ balanced salt solution without Mg+ and Ca2+. Cells were centrifuged one more time and resuspended in the same buffer supplemented with 1 mM MgC12 at the final concentration of 107/ml unless otherwise stated. Routinely, 98 & 1% of the cells are PMN which are 97 t 2% viable. Study of neutrophil-platelet interaction. Purified PMN and washed platelets (250 ~1 of each suspension) were incubated together at 37°C in an aggregometer (Dual Aggrometer, Chrono-Log, Havertown, PA) under stirring at 1,100 rpm in presence of 1.3 mM CaC12 and 0.7 mg/ ml DFP-treated fibrinogen. Cytochalasin B was added (5 pg/ml final concentration), and 5 min later, cells were challenged by a specific PMN agonist either fMLP or rhC5a. Aggregation was recorded for 3 min, after which the entire incubate was added to 125 ~1 of a stopping solution composed of 33% formaldehyde, 77 mM EDTA, and 155 mM NaCl (1:9:8 vol/vol/vol). After a 2-min centrifugation at 13,000 g, 400 ~1 of the supernatant were added to scintillation fluid for serotonin radioactivity counting. Aggregation was expressed as percentage of the maximal light transmission through the Tyrode buffer, and serotonin release was expressed as percentage of the total serotonin content. Study of activation of platelets and neutrophils used separately. To study platelets alone, the protocol de-

scribed above was used, except that the 250 ~1 PMN suspension was replaced by 250 ~1 Hanks’ balanced salt solution. The reverse was performed when studying PMN alone, i.e., 250 ~1 Tyrode buffer were used instead of the 250 ~1 platelet suspension. In the latter case, PMN aggregation was monitored for 3 min, and after highspeed centrifugation (13,000 g for 2 min) the supernatant was collected for ,&glucuronidase release determination. In some experiments, PMN were prepared at a final concentration of lO’/ml and l-ml aliquots were preincubated with cytochalasin B (5 pg/ml) at 37°C for 5 min and then challenged for 3 min with maximal concentrations of rhC5a (0.25 PM) or fMLP (1 PM). Supernatants were then collected for the measurement of their content in cathepsin G enzymatic activity and for their ability to trigger platelet activation. Purification and determination of enzymatic activity of cathepsin G. Cathepsin G and elastase were purified

according to the method of Baugh and Travis (1) and Martodam et al. (17) from human PMN obtained by cytopheresis. Briefly, a crude granule extract was prepared and subsequently processed through aprotininSepharose affinity and CM-Sephadex ion-exchange chromatography. The purified cathepsin G migrated on polyacrylamide gel electrophoresis (Phast gel gradient 10-15, Phast System Separation) with an apparent molecular mass of 30 kDa and was devoid of elastase enzymatic activity. The enzymatic activity of cathepsin G and elastase were determined spectrophotometrically according to


Cl102

RHC~A-INDUCED

NEUTROPHIL-MEDIATED

Nakajima et al. (18) using N-succinyl-Ala-Ala-Pro-Phep-nitroanilide and N-succinyl-Ala-Ala-Val-p-nitroanilide as specific substrate, respectively. For the determination of their active site concentration, increasing amounts of the enzymes were reacted with a constant amount of active site titrated cxl-antitrypsin (final concentration, 0.1 PM) in 495 ~1 of 0.1 M Tris. HCl buffer (pH 8) containing 0.1% bovine serum albumin to minimize adsorption of the enzymes on the test tube surface. After 30 min at 25OC, 5 ~1 of a 100 mM solution of the specific synthetic substrate in N-methylpyrrolidone were added to the mixture. Hydrolysis of the substrate (1 mM final concentration) was monitored every 20 s during 10 min by following the release of p-nitroaniline at 410 nm. With knowledge of the active site concentration of the purified cathepsin G, a standard curve was constructed using the same protocol with different amounts of the enzyme without al-antitrypsin. Subsequent assessment of cathepsin G concentrations in PMN supernatants were calculated using this standard curve. The concentrations of elastase were also determined following a similar procedure. Statistical analysis. All results were expressed as means -+ SD. The data were analyzed by unpaired Student’s t test to determine whether differences were statistically significant. Standard linear regression analysis were used to correlate the different parameters.

PLATELET

A

ACTIVATION

50

OJ

C

RESULTS

Platelet activation by rhC5a. PMN (5 X 106/ml) used alone were activated with fMLP or rhC5a to determine the range of active concentrations. ,&Glucuronidase release in the supernatant was chosen as an index of activation. The active concentrations of fMLP ranged from 0.02 to 1 PM. As expected, rhC5a also induced PMN activation. The concentration-activity curve for rhC5a was shifted to the left relative to that for fMLP, with a maximal release obtained with 0.2 PM (Fig. 1A). Although rhC5a was more potent than fMLP, both agonists induced a maximal release of ,&glucuronidase of ~40% in the presence of cytochalasin B. Without cytochalasin B preincubation, neither agonist induced release of P-glucuronidase above control values (not shown). Stimulation was next performed using PMN in the presence of platelets (2 x 108/ml), and activation of the latter was followed by measuring aggregation (Fig. 1B) and serotonin release (Fig. 1C). As already shown (5), fMLP triggered platelet activation. rhC5a was also a potent inducer of platelet aggregation and serotonin release. The effective concentrations were exactly within the range observed for PMN activation. In control experiments in which PMN were omitted, rhC5a was unable to activate platelets. Moreover, in the absence of cytochalasin B treatment, aggregation and serotonin release did not occur upon addition of rhC5a to the mixed cell population assay (not shown). Release of cathepsin G and of a platelet-activating substance by rhC5a-activated neutrophils. Concentrated

PMN (108/ml) were used alone and stimulated by either 1 PM fMLP (n = 8) or 0.25 PM rhC5a (n = 3). Cathepsin G concentration, as measured by hydrolysis of a specific

MOLAR CONCENTRATION 1. Concentration-activity relationship curves for fMLP and rhC5a. Samples were preincubated at 37°C for 5 min in the presence of 5 pg/ml of cytochalasin B before addition of either fMLP (filled squares) or rhC5a (open squares), and reactions were allowed to process for 3 min more. A: PMN used alone at 5 X 106/ml were challenged with different concentrations of agonists, and release of ,&glucuronidase (flGlut) was measured. Values are expressed in % of total P-glucuronidase content of cells. B and C: PMN and platelets (5 X 106/ml and 2 X 108/ ml, respectively) were mixed and challenged as above. Values of aggregation (B) and serotonin release from platelets (C) were expressed in % of maximal light transmission and in % of total serotonin content of cells, respectively. Each point is mean 2 SD of 4 or 5 different experiments. FIG.

synthetic substrate, was determined in the supernatants. All supernatants contained cathepsin G in concentrations ranging from 0.5 to 4.3 PM (Fig. 2) with a mean value of 1.9 t 1.0 PM (n = ll), demonstrating that, like fMLP, rhC5a is able to release cathepsin G from PMN. When added to washed human platelets, cell-free supernatants from rhC5a-activated PMN were able to trigger serotonin release. As shown in Fig. 2, there is a strong


NEUTROPHIL-MEDIATED

RHC~A-INDUCED

PLATELET

ACTIVATION

0.8 0.6

0

1

2

3

CATHEPSIN

G CONCENTRATION

correlation between platelet activation and the concentrations of cathepsin G in the supernatants (r = 0.98, P < 0.001). A correlation was also observed between platelet activation and elastase concentrations in the supernatants (r = 0.93, P < 0.001; not shown). Activation of platelets by purified cathepsin G and its inhibition. Purified elastase up to 20 pg/ml was unable to stimulate washed human platelets (not shown) as observed by others (3, 23). By contrast, purified cathepsin G induced both aggregation and serotonin release. As shown in Fig. 3, the active concentrations ranged between 1.5 and 6 pg/ml, i.e., between 50 and 200 nM. WEB 2086 (1 PM) and ASA (100 PM) did not modify platelet response to 6 pg/ml cathepsin G (data not shown), implying that neither PAF-acether nor arachi-

4

5

Cl103

FIG. 2. Correlation between cathepsin G concentration and serotonin release from platelets. Concentrated PMN (108/ml) were challenged with either 1 FM fMLP (open squares) or 0.25 PM rhC5a (filled squares). Supernatants were collected 3 min later for determination of their cathepsin G content by active site titration and of their plateletactivating property by serotonin release measurement. For the latter, volume of each supernatant releasing 50% of total serotonin content of platelets was determined and assumed to represent 50 arbitrary units. From this value corresponding arbitrary units per ~1 (au/pi) were calculated and plotted as a function of cathepsin G content expressed in PM (r = 0.9&P< 0.001).

(PM) donic acid was involved when a submaximal concentration was used. Z-Gly-Leu-Phe-CH&l, an active site-directed irreversible inhibitor specific for PMN cathepsin G enzymatic activity (21) was then tested. A concentrationdependent inhibition of the enzymatic activity of 4.5 pg/ ml cathepsin G, as measured spectrophotometrically by substrate hydrolysis, was observed with a 50% inhibition at 126.5 t 29.7 PM (n = 4). As expected, Z-Gly-Leu-PheCH&l inhibited 5 pg/ml cathepsin G-induced aggregation and serotonin release in a concentration-dependent manner. Nevertheless, Z-Gly-Leu-Phe-CH&l was more potent in the platelet test (range 20-80 PM) than in the specific substrate test (range SO-400 ELM). In fact, it also displayed an inhibitory effect vis-a-vis collagen-induced aggregation, leading to the conclusion that Z-Gly-Leu-

80 T

60 -

FIG. 3. Platelet activation by purified cathepsin G. Human washed platelets were challenged with different concentrations of cathepsin G, and reaction was followed during 3 min. Maximal light transmission was then noted, and incubate was processed for serotonin release determination. Aggregation (squares) and serotonin release (circles) were expressed as described in Fig. 1 in function of cathepsin G concentrations expressed in p&/ml. Each point is mean k SD of 5 or 6 different experiments.

0

1.5

3

CATHEPSIN

4.5

G CONCENTRATIONS

6

7.5

(pg/ml)


Cl104

NEUTROPHIL-MEDIATED

RHC~A-INDUCED

Phe-CH&l has a direct effect on platelets and cannot be considered specific in our experimental biological system. Consequently it was not further used as a tool for the investigation of cathepsin G participation in the cell-tocell communication. Effect of natural proteinase inhibitors. Two plasma proteinase inhibitors, al-antitrypsin and al-antichymotrypsin, inhibited platelet aggregation and serotonin release triggered by a submaximal concentration of cathepsin G. A comparative study showed that cul-antichymotrypsin was 15 times more potent than cul-antitrypsin. Thus a 50% inhibition of platelet aggregation was obtained with 750 nM arl-antitrypsin compared with 50 nM al-antichymotrypsin (Fig. 4). Contrary to what was observed with Z-Gly-Leu-Phe-CH&l, cul-antichymotrypsin proved to be a specific inhibitor in our experimental conditions. When this inhibitor was tested at a five times higher concentration than the one required for complete inhibition, as expected no effect was observed against platelet aggregation induced by collagen or even by thrombin, which is also a serine proteinase (not shown). Use of inhibitors in neutrophil-platelet incubates. PMN and platelets were mixed and challenged with 200 nM rhC5a (a concentration inducing a full activation of platelets) exactly as described in Fig. 1. WEB 2086 (1 PM) or ASA (100 PM) PreinCUbated with the cells for 1 or 5 min, respectively, before addition of the agonist did not modify the response (Table 1). By contrast, as depicted in Fig. 5, al-antichymotrypsin inhibited both aggregation and serotonin release in a concentration-dependent manner. This effect was not

lmin

PLATELET

ACTIVATION

due to an inhibition of PMN activation, since the release of ,&glucuronidase, which was 40.0 $- 2.5% from controlactivated PMN, was 40.7 t 3.8% when 300 nM cylantichymotrypsin was present (n = 3, P > 0.05). DISCUSSION The first information demonstrated in this study is that the ability of stimulated PMN to activate platelets (5) is not restricted to fMLP as rhC5a is also able to trigger aggregation and the release of serotonin. As previously described with fMLP, activation of platelets requires PMN to be preincubated with cytochalasin B. Since cytochalasin B by its activity on microfilaments allows the secretion of the contents of both specific and azurophilic PMN granules (lo), it could be deduced that platelet activation was due to a material contained in these granules. We thus looked for the presence of cathepsin G in the supernatant of rhC5a-activated PMN. Indeed, ca th epsin G was the only aggregating substance recovered from the supernatant of f’lMLP-activated PMN (23). Three different batches of concentrated PMN ( lo8 cells/ml) suspension were prepared and challenged with rhC5a. As expected, cathepsin G enzymatic activity was recovered from each batch. In the eight other preparatoions stimulated with fMLP, cathepsin G was also detected in confirmation of previous published results (23). The different supernatants also induced serotonin release from platelets, and a strong correlation was found between serotonin release and cathepsin G content. A correlation was also found between the capability of the supernatants to induce serotonin release and their con-

I CONT ROL

CONTROL f*

/

80 CAT.6 5 II g . m I-1

CAT.G 5rg.mP1

FIG. 4. Effect of cul-antichymotrypsin and cul-antitrypsin on platelet activation triggered by cathepsin G. Experiments were conducted as those depicted in Fig. 3 and are presented similarly. Inhibitors, cul-antichymotrypsin (alACT) and cyl-antitrypsin (al-AT), were added to platelets before cathepsin G (Cat G) at concentrations indicated near each tracing and expressed in nM . Results of 1 experiment representative of 3 others.


RHC~A-INDUCED

NEUTROPHIL-MEDIATED

TABLE 1. Effect of WEB 2086 and ASA on platelet activation by rhC5a in the presence of neutrophils Platelet Aggregation, %

Serotonin Release, %

WEB 2086 (1 PM) 101.6zk1.8 118m3.8 ASA (100 PM) 93.2k1.7 86.9k12.2 Values are means k SD of 4 different experiments and are expressed as % of matched control. WEB 2086 and acetylsalicylic acid (ASA) were added 1 and 5 min before rhC5a, respectively. All data, P > 0.05.

tent in elastase. This was not surprising, since cathepsin G and elastase are contained within the same granules (7). Nevertheless, because elastase is devoid of plateletactivating activity (our data and Refs. 3,23), our findings support the hypothesis that cathepsin G released from rhC5a-activated PMN accounts for platelet stimulation by the supernatants. The mean concentration of cathepsin G recovered from lo8 PMN/ml was 1.9 PM. Thus it can be estimated that at physiological concentrations of PMN, i.e., 5 x 106/ml, the concentration of cathepsin G released in the supernatant is ~100 nM. In fact, using purified cathepsin G, we observed a maximal activation of platelets with 200 nM and ~50% activation with 100 nM. It can therefore be inferred that sufficient cathepsin G is released from PMN activated by rhC5a (and fMLP) to account for platelet activation. In addition, PAF-acether and arachidonic acid metabolites are apparently not involved in the platelet response. In our search for a specific inhibitor of cathepsin G that might be useful for investigation of the cell-to-cell interaction, Z-Gly-Leu-Phe-CH&l was first examined. Among a survey of different chloromethyl ketone inhibitors of the enzymatic activity of cathepsin G, this compound was reported as the most effective (21). However, it proved to be nonspecific in our biological system.

PLATELET

ACTIVATION

Cl105

Second, we investigated the effect of natural plasma proteinase inhibitors, i.e., al-antitrypsin and cul-antichymotrypsin, which have previously been reported to com-plex with cathepsin G (2). Although both-were effective, arl-antichymotrypsin was more potent. This is in agreement with biochemical data showing a stronger association of cathepsin G with cxl-antichymotrypsin than with al-antitrypsin, with association rate constants of 5.1 t 0.7 x lo7 and 4.1 t 0.6 X 105, respectively (2). As a consequence, al-antichymotrypsin was the inhibitor of choice to test the interaction between rhC5a-stimulated PMN and platelets. The indirect activation of platelets by rhC5a through PMN stimulation was inhibited by al-antichymotrypsin. A concentration-dependent inhibition was observed with complete blockade at 300 nM. Because this antiproteinase activity cannot be explained by a direct effect on either platelets or PMN, this result demonstrates that an intermediate released proteinase is involved. Because I) cathepsin G is released from rhC5a-activated PMN, 2) cathepsin G triggers platelet activation, and 3) al-antichymotrypsin is recognized as a specific inhibitor of cathepsin G (2), it can be concluded that this proteinase is the intermediate mediator responsible for the communication between activated human PMN and platelets. We can speculate at length about the physiopathological importance of these in vitro observations. PMN and platelets have been observed together in numerous inflammatory states. Thus aggregates composed of PMN and platelets have been found in inflammatory lesions associated with endocarditis, atherosclerosis, and rheumatoid arthritis. If not as aggregates, PMN and platelets are present at the site of injury of various immunologitally mediated pathological states, such as Arthus and Shwartzman reactions, hyperacute transplant rejection,

loo.

FIG. 5. Inhibition of cwl-antichymotrypsin of platelet activation by rhC5a via PMN stimulation. Experiments were performed as described in Fig. 1, B and C, with 200 nM rhC5a, a concentration inducing a submaximal aggregation. Inhibitions of aggregation (squares) and serotonin release (circles) were expressed in % of control and represented in function of al-antichymotrypsin concentration expressed in nM. Each point is mean k SD of 3 or 4 different experiments.

. 601 . 40. 20. O-

0’

.

10’0

.

. 20’0

ALPHA-1-ANTICHYMOTRYPSIN

30’0

(nM)


Cl106

RHC~A-INDUCED

NEUTROPHIL-MEDIATED

some forms of passive cutaneous anaphylaxis, and immune glomerulonephritis. The least speculative pathological situation in which such an interaction may occur is in ARDS, in which PMN and platelets are likely to be involved as mentioned in the introduction. This is now strengthened by our present results showing that rhC5a triggers platelet activation through stimulation of PMN. Indeed in patients with ARDS, complement activation has been detected by the presence of the fragment C5a in the circulating blood (11,22) and in the bronchoalveolar lavage (22). Moreover, in similar lavage fluid of some patients, neutrophil elastase has been detected (16), suggesting that most probably cathepsin G was also present. Thus the clinical data bring evidence that our in vitro findings, although obtained in presence of cytochalasin B, are most probably relevant to the in vivo situation. However, one has to be cautious when setting up an animal model, analogous to that in human, for studying in vivo interaction between PMN and platelets. Identical experiments conducted in vitro with rabbit cells have led to a different conclusion; i.e., PAF-acether but not cathepsin G is the involved mediator (6, 19). In fact this is consistent with another observation that there is no cathepsin G in rabbit PMN (20). In this regard, the rat is most probably a better species than the rabbit, since rat PMN contain cathepsin G (28). It can be speculated that in the human the chain of events starts with the specific activation of PMN by C5a (or other agonists) with, as a consequence, the release of cathepsin G, which in turn stimulates nearby platelets. Activated platelets would then expel the content of their granules, particularly serotonin, and would also form thromboxane, prostaglandins, and PAF-acether. These different substances would then be responsible among others for the change in vasopermeability, a characteristic of inflammatory states, particularly of ARDS. One of the main objections to the in vivo effect of cathepsin G is the presence of the plasma antiproteinases. Nonetheless, although they inhibit cathepsin G and elastase, a significant proteolytic activity was still detectable in the plasma of patients with septicemia, a pathological state for which complication is ARDS (8). Moreover, it has been shown in some experimental conditions that proteinases released from PMN can react with susceptible substrates even in the presence of physiological concentrations of proteinase inhibitors. This may be due to a partial exclusion of the inhibitors from the PMN-substrate interface (4, 30), a situation that could be relevant for cooperation between PMN and platelets. In conclusion, the present investigation describes the activation of washed human platelets by rhC5a via a stimulation of PMN. This previously unidentified mechanism implicates cathepsin G, a relatively ignored proteinase whose involvement in diseases is poorly documented. The present report evidenced an important mechanism operating under in vitro conditions for this proteinase. One can anticipate a similar interaction between PMN and platelets under in vivo conditions, a hvpothesis that remains to be investigated.

PLATELET

ACTIVATION

Prof. B. B. Vargaftig and Dr. M. Pretolani are acknowledged for their advice in the preparation of the manuscript. We are indebted to Drs. M. A. Selak (Temple University, Philadelphia, PA) and A. J. Coyle (King’s College London, London, UK) for their valuable critical review. Present address of M. Schattner: Academia National de Medicina de Buenos Aires, Instituto de Investigaciones Hematologicas, Buenos Aires, Argentina. Address for reprint requests: M. Chignard, Unite de Pharmacologic Cellulaire, Unite Associee IP/INSERM 285, Institut Pasteur, 25, rue du Dr. Roux, 75724 Paris Cedex 15, France. Received 5 September 1989; accepted in final form 1 February 1990. REFERENCES 1. BAUGH, R. J., AND J. TRAVIS. Human leukocyte granule elastase: rapid isolation and characterization. Biochemistry 15: 836-841, 1976. 2. BEATTY, K., J. BIETH, AND J. TRAVIS. Kinetics of association of serine proteinases with native and oxidized a-1-proteinase inhibitor and cw-l-antichymotrypsin. J. Biol. Chem. 255: 3931-3934,198O. 3. BOWER, M. S., R. I. LEVIN, AND K. GARRY. Human neutrophil elastase modulates platelet function by limited proteolysis of membrane glycoproteins. J. Clin. Invest. 75: 657-666, 1985. 4. CAMPBELL, E. J., R. M. SENIOR, J. A. MCDONALD, AND D. L. COX. Proteolysis by neutrophils. Relative importance of cell-substrate contact and oxidative inactivation of proteinase inhibitors in vitro. J. Clin. Invest. 70: 845-852, 1982, 5. CHIGNARD, M., M. A. SELAK, AND J. B. SMITH. Direct evidence for the existence of a neutrophil-derived platelet activator (neutrophilin). Proc. Natl. Acad. Sci. USA 83: 8609-8613,1986. 6. CO~~FFIER, E., D. JOSEPH, M. C. PREVOST, AND B. B. VARGAFTIG. Platelet-leukocyte interaction: activation of rabbit platelets by fMLP-stimulated neutrophils. Br. J. Pharmacol. 92: 393-406,1987. 7. DEWALD, B., R. RINDLER-LUDWIG, U. BRETZ, AND M. BAGGIOLINI. Subcellular localization and heterogeneity of neutral protease in neutrophilic polymorphonuclear leukocytes. J. Exp. Med. 141: 709-723,1975. 8, EGBRING,

R., W. SCHMIDT, G. FUCHS, AND K. HAVEMANN. Demonstration of granulocytic proteases in plasma of patients with acute leukemia and septicemia with coagulation defects. Blood 49:

219-231,1977. 9. FRANKE, A. E., G. C. ANDREWS, GERARD, AND H. J. SHOWELL.

N. P. STIMLER-GERARD, G. J. Human C5a anaphylatoxin: gene synthesis, expression, and recovery of biologically active material from Escherichia coli. Methods Enzymol. 162: 653-668, 1988. 10. GOLDSTEIN, I. M., S. HOFFSTEIN, J. I. GALLIN, AND G. WEISSMANN. Mechanisms of lysosomal enzyme release from human leukocytes: microtubule assembly and membrane fusion induced by a component of complement. Proc. Natl. Acad. Sci. USA 70: 2916-2920,1973. 11. HAMMERSCHMIDT, CRADDOCK, AND

D. E., L. J. WEAVER, L. D. HUDSON, P. R. H. S. JACOB. Association of complement activation and elevated plasma C5a with adult respiratory distress syndrome: pathophysiological relevance and possible prognostic value.

Lancet 1: 947-949,198O. 12. HEFFNER, J. E., S. A. SAHN,

in the adult respiratory

AND J. E. REPINE. The role of platelets distress syndrome. Culprits or bystanders?

Am. Rev. Respir. Dis. 135: 482-492, 13. IRVIN, C. G., N. BEREND, AND P.

1987.

M. HENSON. Airways hyperreacproduced by aerosolization of human C5a

tivity and inflammation des Arg. Am. Rev. Respir. Dis. 134: 777-783,1986. 14. ISSEKUTZ, A. C., M. RIPLEY, AND J. R. JACKSON. Role of neutrophils in the deposition of platelets during acute inflammation. Lab.

Invest. 49: 716-724, 1983. 15. JOHNSON, R. J., C. E. ALPERS, P. PRITZL, M. SCHULZE, P. BAKER, C. PRUCHNO, AND W. G. COUSER. Platelets mediate neutrophildependent immune complex nephritis in the rat. J. Clin. Invest. 82: 1225-1235,1988. 16. LEE, C. T., A. M. FEIN, M. LIPPMANN, H. HOLTZMAN, P. KIMBEL, AND G. WEINBAUM. Elastolytic activity in pulmonary lavage fluid from patients with adult respiratory-distress syndrome. N. Engl. J. Med. 304: 192-196,1981. 17. MARTODAM, R. R., R. J. BAUGH, D. Y. TWUMASI, AND I. E. LIENER. A rapid procedure for the large scale purification of


RHC~A-INDUCED

NEUTROPHIL-MEDIATED

elastase and cathepsin G from human sputum. Prep. Biochem. 9: l&31,1979. 18. NAKAJIMA,

K., J. C. POWERS, B. M. ASHE, AND M. ZIMMERMAN. Mapping the extended substrate binding site of cathepsin G and human leukocyte elastase. Studies with peptide substrates related to the cul-protease inhibitor reactive site. J. &ol. Them. 254: 4027-

4032,1979. 19. ODA, M.,

K, SATOUCHI, K. YASUNAGA, AND K. SAITO. Polymorphonuclear leukocyte-platelet interactions: acetylglyceryl ether phosphocholine-induced platelet activation under stimulation with chemotactic peptide. J. Biochem. 100: 1117-1123,1986. 20. OLSSON, I., AND P. VENGE. The role of the human neutrophils in the inflammatory reaction. Allergy 35: l-13, 1980. 21. POWERS, J. C., B. F. GUPTON, A. D. HARLEY, N. NISHINO, AND R. J. WHITLEY. Specificity of porcine pancreatic elastase, human leukocyte elastase and cathepsin G. Inhibition with peptide chloromethyl ketones. Biochim. Biophys. Acta 485: 156-166, 1977. 22, ROBBINS, R, A., W. D. Russ, J. K. RASMUSSEN, AND M. M. CLAYTON. Activation of the complement system in the adult respiratory distress syndrome. Am. Rev. Respir. Dis. 135: 651-658, 1987. 23.

SELAK, M. A., M. CHIGNARD, AND J. B. SMITH. strong platelet agonist released by neutrophils. 2930299,1988.

Cathepsin G is a Biochem. J. 251:

PLATELET

ACTIVATION

Cl107

24. STIMLER, N. P., T. E. HUGLY, AND C. M. BLOOR. Pulmonary injury induced by C3a and C5a anaphylatoxins. Am. J. Pathol. 100: 327-348,198O.

25. TATE, R. M., AND J. E. REPINE. Neutrophils and the adult respiratory distress syndrome. Am. Rev. Respir. Dis. 128: 552-559,1984. 26. TILL, G. O., M. L. MORGANROTH, R. KUNKEL, AND P. A. WARD. Activation of C5 by cobra venom factor is required in neutrophilmediated lung injury in the rat. Am. J, Pathol. 129: 44-53, 1987. 27 TVEDTEN, H. W., G. 0. TILL, AND P. A. WARD. Mediators of lung ’ injury in mice following systemic activation of complement. Am. J. Pathol. 119: 92-100,1985. 28 VIRCA, G. D., G. METZ, AND H. P. SCHNEBLI. Similarities between human and rat leukocyte elastase and cathepsin G. Eur. J. Biochem. 144: l-9,1984. J. E., W. B. DAVIES, J. F. HOLTER, J. R. MOHAMMED, 2g* WEILAND, P. M. DORINSKY, AND J. E. GADEK. Lung neutrophils in the adult respiratory distress syndrome. Clinical and pathophysiologic significance. Am. Rev. Respir. Dis. 133: 218-225,1986. 3. I., A. J. HUANG, S. L. LANDMAN, S. C. NICHOLSON, 30. WEITZ, AND S. C. SILVERSTEIN. Elastase-mediated fibrinogenolysis by chemoattractant-stimulated neutrophils occurs in the presence of physiologic concentrations of antiproteinases. J. Exp. Med. 166: l

1836-1850,1987.


Eglin C and heparin inhibition of platelet activation induced by cathepsin G or human neutrophils.

Neutrophil collagenase activation: the role of oxidants and cathepsin G.

Cathepsin G binding to human platelets. Evidence for a specific receptor.

Intracellular Ca2+ rise in human platelets induced by polymorphonuclear-leucocyte-derived cathepsin G.

Activation of the human leukocyte proteinases elastase and cathepsin G by various surfactants.

Activation and damage of cultured airway epithelial cells by human elastase and cathepsin G.

G protein activation and mediator release from human neutrophils and platelets after stimulation with sodium fluoride and receptor-mediated stimuli.

In vitro killing of oral Capnocytophaga by granule fractions of human neutrophils is associated with cathepsin G activity.

Proteolytic activation of human cathepsin A.

Local and systemic immunomodulatory mechanisms triggered by Human Papillomavirus transformed cells: a potential role for G-CSF and neutrophils.

Human cathepsin G. Catalytic and immunological properties.

A cytosolic inhibitor of human neutrophil elastase and cathepsin G.

Inhibition by heparin of platelet activation induced by neutrophil-derived cathepsin G.

Activation of permeabilized neutrophils: role of anions.

Cathepsin G, a polymorphonuclear cell protease, affects the fibrinolytic system by releasing PAI-1 from endothelial cells and platelets.

Inhibition of PAF-induced activation of human platelets by verapamil.

Epinephrine potentiates activation of human platelets by low density lipoproteins.

Degradation of human factor X by human polymorphonuclear leucocyte cathepsin G and elastase.

Inhibitors of human neutrophil cathepsin G: structural and biochemical studies.

The role of neutrophils and platelets in a rabbit model of thromboembolic stroke.

In vitro killing of Actinobacillus actinomycetemcomitans and Capnocytophaga spp. by human neutrophil cathepsin G and elastase.

Generation of Gla-domainless FVIIa by cathepsin G-mediated cleavage.

The inhibition of human neutrophil elastase and cathepsin G by peptidyl 1,2-dicarbonyl derivatives.

Degradation of basement membrane laminin by human neutrophil elastase and cathepsin G.