Immunology Letters, 26 (1990)25- 30

Elsevier IMLET 01416

Asbestos fibers induce release of collagenase by human polymorphonuclear leukocytes Mikael H e d e n b o r g 1, T i m o Sorsa 2, A n n e l i L a u h i o 3 a n d M a t t i Klockars ~ 1Institute of Occupational Health, Helsinki, 2Departments of Periodontology and Medical Chemistry, University of HelsinkL and 3Department of Medical Chemistry, University of Helsinki, Helsinki, Finland

(Received 7 March 1990;.accepted 1 May 1990)

1. Summary The asbestos fibers chrysotile and crocidolite cause a dose-dependent release of specific granule collagenase by human polymorphonuclear leukocytes (PMNL). Release of azurophil granule elastase was induced by the asbestos fibers at higher concentrations, suggesting that asbestos fibers primarily cause the release of specific granule contents of human PMNL. Wollastonite, a fibrous silicate mineral, causes a weaker collagenase release and no elastase release. The collagenase was released in inactive, latent form. Carboxymethyl cellulose (CMC), an agent known to blunt chrysotile-induced hemolysis and production of reactive oxygen metabolites by human PMNL, specifically inhibits chrysotileinduced release of collagenase. Chrysotile asbestos was found to bind the P M N L serine proteinase cathepsin G. A role of collagenase release, production of reactive oxygen metabolites and cathepsin G binding by chrysotile for the perpetuation of the asbestos-induced alveolitis is suggested. 2. Introduction

pulmonary fibrosis thought to represent the end stage of a chronic inflammatory process [1]. The asbestos-induced inflammation is characterized by an accumulation of polymorphonuclear leukocytes (PMNL) within the alveolar structures [2]. Although the severity of this PMNL alveolitis has been found to correlate with duration of asbestos exposure [3], the exact role of PMNL in the pathogenesis of asbestosis remains controversial [4]. Active collagenase (matrix metalloproteinase-1, E.C. 3.4.24.7) has been detected in bronchoalveolar lavage fluid (BAL) of individuals with asbestosis [5]. Further, an increased free elastase-like activity has been demonstrated in BAL-fluids of individuals with asbestosis [6] suggesting a role of these proteases in the development of asbestosis. We studied the ability of two asbestos fibers, chrysotile and crocidolite, and a non-asbestos fibrous silicate, wollastonite, to induce the release of collagenase, cathepsin G and elastase by human PMNL. We also studied the effect of carboxymethyl cellulose (CMC), an agent that inhibits chrysotile asbestos induced cytotoxicity [7] and reactive oxygen metabolite (ROM) production by human P M N L [8], on mineral fiber-induced collagenase release.

The pathogenesis of asbestosis in humans is unclear. Asbestosis is characterized by an interstitial

3. Materials and Methods

Key words." Neutrophil collagenase; Activation; Neutrophil

3.1. Cell &olation

elastase; Neutrophil cathepsin G; Asbestos fibres; Carboxymethyl cellulose Correspondence to: Mikael Hedenborg, M.D., Institute of Occupational Health, Topeliuksenkatu41 aA, SF-00250 Helsinki, Finland

PMNL were isolated from buffy coats of healthy blood donors as has been previously described [9]. Briefly, the isolation process involved an initial 1% dextran T-500 sedimentation (Pharmacia) followed

0165-2478 / 90 / $ 3.50 © 1990 ElsevierSciencePublishers B.V.(BiomedicalDivision)

25

by Ficoll (Pharmacia) density centrifugation. Contaminating red cells were lysed with NH4C1 and the white cells were washed twice with Dulbecco's phosphate-buffered saline (PBS) (Gibco) and counted in a Coulter counter. After isolation 9 5 - 9 8 % of the cells were PMNL, and the proportion of viable cells was above 95% (trypan blue dye exclusion). 3.2. Experimental conditions The U I C C (International Union Against Cancer) reference samples of chrysotile and crocidolite and Finnish wollastonite (FW-325, Partek Co.) were suspended in PBS. A total of 10x 106 P M N L and particles (final concentration 100-800 ~g/ml) was incubated in a volume of 1.0 ml for 30 min at 37 °C. After centrifugation (3000 rev./min, 4 °C) the activity of collagenase, cathepsin G, elastase and lactate dehydrogenase (LDH) was determined in the cell free supernatant. In some experiments CMC (Sigma) was included in the reaction mixture as indicated. 3.3. Enzyme assays Collagenase activity was measured using SDSpolyacrylamide gel electrophoresis (SDS-PAGE) for quantitative analysis of the characteristic cleavage products resulting from collagenase action on native soluble type I collagen [10]. P M N L supernatants were incubated with 1.5/~M human skin type I collagen in 50 mM Tris-HC1/0.2 M NaC1/5 mM CaC12 at pH 7.5 for 12 h at 22°C. Collagenase activity in the supernatants was assayed without and with 1 mM phenylmercuric chloride (PMC) treatment to reveal autoactive and total collagenase activity, respectively. The PAGE gels were fixed overnight in a solution containing 10% acetic acid and 40% methanol and stained with Coomassie Brilliant blue, followed by destaining and densitometric quantitation as described [11]. Elastase- and cathepsin G-like activities of the supernatants were measured using 1 mM succinyl-alanyl-alanyl-valine-paranitroanilide (Sigma) or succinyl-alanyl-alanyl-prolyl-phenylalanyl-4-nitranilid (Boehringer) in PBS containing 8% dimethylsulfoxide (DMSO) as substrates [12]. The elastase and cathepsin G activities, measured spectrophotometrically at 405 nm, were expressed as % released of total enzyme activity which was ob26

tained after 0.2% Triton X-100 lysis of the cells. The binding of cathepsin G to the mineral fibers was studied by incubating 100-400/~g/ml of fibers with 0.2% Triton X-100 lysed P M N L supernatants for 30 rain at 37 °C, followed by centrifugation and cathepsin G determination of the supernatant. L D H activity was measured by the rate of change in absorbance at 340 nm due to N A D H oxidation in a mixture containing 0.2 mM N A D H (Sigma), 1.36 mM pyruvate (Boehringer Mannheim) in PBS [13]. Results are expressed as % released of the total enzyme activity obtained after 0.2% Triton X-100 lysis of the cells. 4. Results

The PAGE electrophoresis in Fig. 1 shows that asbestos fibers induced a dose dependent release of collagenase. Chrysotile induced the strongest release followed by crocidolite whereas collagenase release by wollastonite was comparatively lower and showed no clear dose dependency. Collagenase was released in an inactive form. The degradation of collagen calculated from densitometric scans of the gels demonstrates the dose-response effect and is shown in Fig. 2. Fig. 3 demonstrates that CMC inhibits the collagenase release induced by chrysotile asbestos but has no inhibitory effect on the crocidolite in-

5 6 7 8 9 10 !1 12 1314

Fig. 1. Proteolysis of soluble human type ] collagen by collagenase released into the supernatant of mineral fiber exposed PMNL. PMNL (10× 106)were exposed to 100- 800 #g/ml of the mineral fibers for 30 min at 37 °C. A portion of the cell-free supernatant was treated with l mM PMC for l0 min at 37°C and then incubated with 1.5 /~Mhuman skin type I collagenin 50 mM Tris-HC1/0.2 M NaC1/5 mM CaC12at pH 7.5 (25 °C) for 12 h. a indicates undegraded monomers and aA 3/4-degradation products of monomers. Lane l, no stimulus; lane 2, 100 ~g/ml chrysotile; lane 3, 200 #g/ml chrysotile; lane 4, 400/~g/ml chrysotile; lane 5, 800/zg/ml chrysotile; lane 6, 100 ~g/ml crocidolite; lane 7, 200/~g/ml crocidolite; lane 8, 400 #g/ml crocidolite; lane 9, 800 ~g/ml crocidolite; lane 10, 100 ~tg/ml wollastonite; lane II, 200/~g/ml wollastonite; lane 12, 400 p.g/mlwollastonite; lane 13, 800 ~g/ml wollastonite; and lane 14, 0.2% Triton X-100extract 000% collagenase release).

TABLE 1

tO0

The release of elastase and LDH into the supernatant of 10 x 106 PMNL exposed to chrysotile, crocidolite and wollastonite (07o released of total activity obtained after Triton X-100 lysis of cells).

8O

"O

6o 40

Elastase

LDH

20

0

200

400

600

Chrysotile

(400 #g/ml) (800 #g/ml)

10.1 + 10.2 34.3 _+25.3

6.6___2.2 13.0_+2.1

Crocidolite

(400 #g/ml) (800 #g/ml)

2.1 _+ 0.7 18.8 _+11.9

2.1 + 1.6 6.3 _+3.2

Wollastonite (400 #g/ml) (800 #g/ml)

3.6_+ 0.9 5.6_+ 2.6

2.0_+ 1.7 3.3_+2.1

800

Fiber c o n c e n t r a t i o n p g / m l

Fig. 2. Relative collagen degradation (°70)by collagenase released into the supernatants of PMNL exposed to various amounts (100- 800 #g/ml) of chrysotile ( • ), crocidolite ( o ) and wollastonite ( • ). The °7o collagen degradation is based on densitometric analysis of corresponding lanes in Fig. 1.

duced collagenase release. The activities of LDH and elastase released from PMNL after exposure to fibers are shown in Table 1. LDH release into the supernatant was weak in ceils exposed to crocidolite and wollastonite and slightly higher in chrysotile exposed cells. Release of elastase, exceeding the leakage of cytoplasmatic

Values are means _+SD of three different experiments.

LDH due to cytotoxity, was observed in the presence of chrysotile and crocidolite asbestos at a concentration of 800/~g/ml, whereas wollastonite caused no elastase release. No cathepsin G activity was found in supernatants of mineral fiber-exposed PMNL. We subsequently found that cathepsin G was removed from the suspension by the binding to chrysotile. Table 2 shows that chrysotile at concentrations of 100-400 /zg/ml almost completely removed the cathepsin G activity of P M N L supernatants, whereas crocidolite and wollastonite showed only slight or no binding at all of the enzyme. It was shown that cathepsin G activity was not affected by any PBS-soluble fraction of

TABLE 2 Percent cathepsin G activity remaining in PMNL lysates (0.2°7o Triton X-100) after incubation with mineral fibers. i

Chrymotile

Cr ocldollte

80 SO

80 60

Chrysotile

0/zg/ml 100 #g/ml 200/zg/ml 400 #g/ml

100070 12.8°70 7.7°70 1.8 070

Crocidolite

0 #g/ml 100 #g/ml 200/~g/ml 400 #g/ml

100070 81°7o 80070 65°70

Wollastonite

0 #g/ml 100 ~g/ml 200 #g/ml 400 ~g/ml

100% 112% 117% 110070

40

2o

o

o

lOO

Io

CMC (jJg/ml)

1

2o

o

lOO CMC

lO (JJglml)

I

Fig. 3. The effect of CMC (1-100 #g/ml) on collagenase release and corresponding collagen degradation by PMNL exposed to 400/~g/ml chrysotile (A) or crocidolite (B). Lane IA, 400/zg/ml chrysotile; lane 2A, 400/~g/ml chrysotile + 100 #g/ml CMC; lane 3A, 400 ~g/ml chrysotile + 10/~g/ml CMC; lane 4A, 400 ~g/ml chrysotile + 1 ~tg/ml CMC; lane 1B, 400/~g/ml crocidolite; lane 2B, 400~g/ml crocidolite + 100/~g/ml CMC; lane 3B, 400/zg/ml crocidolite + 10/~g/ml CMC; lane 4B, 400 izg/ml crocidolite + 1 /zg/ml CMC.

27

chrysotile. There was no similar binding of PMNL elastase to asbestos fibers. 5. Discussion A role of proteolytic enzymes, including collagenase, released locally by inflammatory cells, has been suggested in the pathogenesis of asbestosis and other chronic lung diseases. Selective release of lysosomal enzymes, in the absence of any detectable cell death, has been observed after exposure of mononuclear phagocytes to asbestos [14]. We report that PMNL after exposure to chrysotile and crocidolite asbestos in vitro release more collagenase (in inactive form) than is seen with the non-asbestos fibrous silicate wollastonite, and that the chrysotile induced collagenase release can be blocked with a particle surface modifying agent CMC. Collagenase is stored in the specific granules of PMNL in a latent, inactive form [15]. It has been shown that different stimuli have different effects on the secretion of enzymes from azurophilic (primary) and specific (secondary) granules [16] and that the extracellular release of specific and azurophilic granules is under separate control [16, 17]. The granules discharge their contents at different rates during phagocytosis [18]. It has been suggested that specific granules serve an external secretory function, whereas azurophilic granules have a predominantly intracellular site of action in the phagolysosome formation [19]. Release of elastase, a serine proteinase located in the azurophilic granules of the PMNL, was also induced by the asbestos fibers. However, significant elastase release occurred only at high concentrations of asbestos fibers suggesting that the fibers primarily cause release of specific granule contents. The PMNL released their collagenase in a latent, inactive form. Using potent stimulators of the respiratory burst, opsonised zymosan and PMA, collagenase autoactivation by hypochlorous acid (HOC1), formed by the MPO-H202-C1 system, has been demonstrated [20]. As different stimuli cause different intensity of the respiratory burst [9], the lack of collagenase activation of asbestos exposed PMNL may depend on the differences in the amount of ROM produced. Thus the intensity of the respiratory burst would determine the degree of PMNL collagenase activation. However, also the PMNL serine 28

proteinase cathepsin G, possibly activated by HOC1, has been implicated as a proteolytic endogenous activator of PMNL collagenase [21]. The binding of cathepsin G to chrysotile asbestos and lack of collagenase activation reported here thus supports the role of serine proteinases, especially cathepsin G, or the cooperative action of ROM and serine proteinase(s) in endogenous activation of PMNL collagenase. We have recently shown that CMC selectively inhibits the chrysotile induced production of ROM by human PMNL, having little effect on the crocidolite-induced production of ROM [8]. CMC also antagonizes the hemolytic and cytotoxic potential of chrysotile by binding to the surface of the mineral [7, 22]. Interestingly CMC also selectively inhibits the chrysotile induced release of PMNL collagenase probably by modifying the cell membranefiber interaction. This is of interest as it is known that specific granule fractions contain a unique cytochrome b believed to be of importance in formation of the complete NADPH oxidase of the plasma membrane responsible for the respiratory burst and production of ROM [19, 23]. The mechanism of the vicious inflammatory circle induced by the deposition of asbestos mineral fibers in the alveolar space is unknown. Related to the release of PMNL collagenase, its activation and the production of ROM it has been shown that collagen degradation products are chemotactic for human PMNL [24], and that soluble collagen and one of its breakdown peptides will trigger superoxide (0 2-) formation by human PMNL [25]. Cathepsin G has been suggested to protect the inflamed lung by degrading collagenous peptides at the site of inflammation [21]. Thus it might be speculated that binding of cathepsin G to the chrysotile fibers plays a role in modifying the inflammatory response. Acknowledgements This work was supported by grants from Finska L~kares~llskapet, Orion Research Foundation, Finnish Cancer Foundation, Duodecim and Magnus Ehrnrooths Foundation.

References [1] Snider, G. L. (1983) Am. Rev. Respir. Dis. 127, 535. [2] Crystal, R. G., Gadek, J. E., Ferrans, V. J., Fulmer, J. D., Line, B. R. and Hunninghake, G. W. (1981) Am. J. Med. 70, 542. [3] Robinson, B. W. S., Rose, A. H., James, A., Whitaker, D. and Musk, A. W. (1986) Chest 90, 396. [4] B6gin, R., Cantin, A., Berthiaume, Y., Boileau, R., Bisson, G., Lamoureux, G., Rola-Pleszczynski, M., Drapeau, G., Mass6, S., Boctor, M., Berault, J., P61oquin, S. and Dalle, D. (1985) Am. J. Industr. Med. 8, 521. [5] Gadek, J., Hunninghake, G., Schoenberger, C., Fells, G. and Crystal, R. (1981) Chest 80 (suppl.), 63. [6] Scharfman, A., Hayem, A., Darvil, M., Marko, D., Hannothiaux, M. H. and Lafitte, J. J. (1989) Eur. Respir. J. 2, 751. [7] Harington, J. S., Miller, K. and Macnab, G. (1971) Environ. Res. 4, 95. [81 Klockars, M., Hedenborg, M. and Vanhala, E. (1990) Arch. Environ. Health, 45, 8. [9] Hedenborg, M. and Klockars, M. (1987) J. Clin. Pathol. 40, 1189. [10] Turto, H., Lindy, S., Uitto, V.-P., Wegelius, O. and Uitto, J. (1977) Anal. Biochem. 83, 557. [11] Sorsa, T. A. (1987) Scand. J. Rheumatol. 16, 167. [12] Bieth, J., Spiess, B. and Wermuth, C. G. (1974) Biochem.

Med. 11, 350. [13] Mold6us, P., H6gberg, J. and Orrenius, S. (1978) Methods Enzymol. 52, 60. [14] Davies, E, Allison, A. C., Ackerman, J., Butterfield, A. and Williams, S. (1974) Nature 251, 423. [15] Weiss, S. J. (1989) New Engl. J. Med. 320, 365. [16] Goldstein, I. M. (1984) Contemp. Top. Immunobiol. 14, 189. [17] Weissman, G., Smolen, J. E. and Korchak, H. M. (1980) New Engl. J. Meal. 303, 27. [18] Stossel, T. E, Pollard, T. D., Mason, R. J. and Vaughan, M. (1971) J. Clin. Invest. 50, 1745. [19] Falloon, J. and Gallin, J. I. (1986) J. Allergy Clin. Immunol. 77, 653. [20] Weiss, S. J., Peppin, G., Ortiz, X., Ragsdale, G. and Test, S. T. (1985) Science 227, 747. [21] Berg, R. A., Capodici, C. and D'arimento, J. (1989) in: Therapeutic Approaches to Inflammatory Diseases (Lewis, A. J., Doherty, N. S. and Ackerman, N. S., Eds.) Elsevier, Amsterdam/New York. [22] Macnab, G. and Harington, J. S. (1967) Nature 207, 522. [23] Borregaard, N., Heiple, J. M., Simons, E. R. and Clark, R. A. (1983) J. Cell Biol. 97, 52. [24] Laskin, D. L., Kimura, T., Sakakibara, S., Riley, D. J. and Berg, R. A. (1986) J. Leuk. Biol. 39, 255. [25] Monboisse, J.-C., Bellon, G., Dufer, J., Randoux, A. and Borel, J.-P. (1987) Biochem. J. 246, 599. [26] Capodici, C. and Berg, R. A. (1989) Inflammation 13, 137.

29