The role of gingival crevicular fluid and salivary interstitial collagenases in human periodontal diseases

Vol. 35, Suppl.,pp. 193%196S, Printedin Great Britain.All rightsreserved

Arrhs oral Bid.

0003-9969/90 $3.00 + 0.00

1990

Copyrightc I990PergamonPressplc

THE ROLE OF GINGIVAL CREVICULAR FLUID AND SALIVARY INTERSTITIAL COLLAGENASES IN HUMAN PERIODONTAL DISEASES T. SORSA'*, K. SUOMALAINEN’ and V.-J. UITTO’ ‘Departments of Periodontology and Medical Chemistry, University of Helsinki, Finland, and ‘Department of Oral Biology, University of British Columbia, Vancouver, BC, Canada Summary-Interstitial collagenases (matrix metalloproteinase-1, EC 3.4.24.7), isolated from extracts of inflamed human gingiva, gingival crevicular fluid and saliva were characterized for their molecular weight, proteolytic and non-proteolytic activation and substrate specificity against soluble collagen types I, II and III. All three collagenases had M, of 70 K. The enzymes existed predominantly in a latent form that could be activated by aminophenylmercuric acetate, gold thioglucose and hypochlorous acid. Among serine proteases tested, trypsin, chymotrypsin, neutrophil cathepsin G and a combination of trypsin and human gingival fibroblast prostromelysin activated gingival and salivary interstitial collagenases. Plasmin and plasma kallikrein, however, were relatively ineffective activators. The collagenases degraded soluble type I and II collagens at apparently equal rates but considerably faster than they did type III collagen. These findings suggest that the characteristics of interstitial collagenases found in inflamed human gingiva, gingival crevicular fluid and saliva are consistent with those of human neutrophil interstitial collagenase rather than the fibroblast-type interstitial collagenase. Thus, neutrophils are suggested to be the main source of such enzymes in inflamed human gingiva, crevicular fluid and saliva during adult periodontitis. Key words: collagenase, neutrophils, activation, periodontal disease.

INTRODUCTION Periodontal diseases are bacterially induced inflammatory conditions affecting the tooth-supporting tissues-gingiva, periodontal ligament, alveolar bone and root cementum (Lindhe, 1989). During the inflammatory process of periodontal disease the tooth-supporting tissues undergo destruction. This destruction is believed to result from increased tissue levels of several types of proteolytic enzymes, including matrix metalloprotease and plaminogen activator/plasmin cascades (Birkedal-Hansen, 1988). Because interstitial collagenases are the only host proteases capable of degrading native collagen triple helix, they may play an essential role in initiation of gingival collagen degradation (Uitto, Tryggvason and Sorsa, 1987). Increased activity of these collagenases in inflamed human gingiva and gingival crevicular fluid was found in comparison with that in healthy tissue and fluid (Uitto et al., 1987; Overall, Weibkin and Thonard, 1987). Also, latent interstitial collagenase is converted to an active form by the inflammatory process of periodontal disease (Uitto and Raeste, 1978; Overall et al., 1987). In addition, interstitial collagenase activity decreases after treatment of periodontal disease (Larivee, Sodek and Ferrier 1986; Hakkarainen, Uitto and Ainamo, 1988). However, the cellular sources of these collagenases and the mechanism(s) of their activation in periodontal collagen degradation remain unclear (Birkedal-Hansen, 1988). We have now addressed *Address correspondence to: Dr Timo Sorsa, Department of Periodontology, University of Helsinki, Mannerheimintie 172, SF-00300 Helsinki, Finland. Abbreviation:APMA, aminophenylmereuric acetate.

this question by studying certain biochemical characteristics of the interstitital collagenases in human gingival tissue, crevicular fluid and saliva. MATERIALAND METHODS Chemicals

Chromatographic materials were purchased from Pharmacia (Uppsala, Sweden). APMA, gold thioglucose, trypsin, soy bean trypsin inhibitor, phenylmercuric chloride, plasmin and plasma kallikrein were obtained from Sigma (St Louis, MO, U.S.A.). Human gingival fibroblast prostromelysin was purified and kindly provided by Dr Gillian Murphy, Strangeways Research Laboratories, Cambridge, England (Murphy et al., 1987). Neutrophil cathepsin G was partially purified by Trasylol-Sepharose and carboxymethyl-cellulose chromatographies from neutrophil extracts (Travis et al., 1980). All other reagents were of the highest analytical grade available. Collection of saliva and gingival crevicular fluid

Stimulated whole saliva was collected from subjects with healthy and diseased periodontium (Uitto, Suomalainen and Sorsa, 1990). Sulcular fluid was collected from adult periodontits patients by the absorption strip technique and eluted from the strips with 50 mM tris, 0.2 M NaCl, 10 mM CaCl,, pH 7.5 (TNC buffer), as described by Sorsa et al. (1988). Extraction of inflamed human gingival tissue

Inflamed human gingival tissue samples were obtained during routine flap surgery at the Department of Periodontology, University of Helsinki, and extracted for interstitial collagenases under optimal conditions (Sorsa et al., 1988). 193s

T. SORSAet al.

194s Gel jiltration chromatography

Two-millimeter samples of saliva, crevicular fluid and gingival tissue extract were applied to a Sephacryl S-200 column (1.5 x 94 cm), equilibrated with 50 mM tris, 0.5 M NaCl, 5 mM CaCl,, I % Triton X-100. A flow rate of 0.5 ml/min was used and 3.5 ml fractions were collected. Aldolase (158 K), bovine serum albumin (68 K), ovalbumin (45 K) and fi-lactoglobulin (35 K) were used in molecular weight calibration (Sorsa et al., 1985). Assay of collagenase activity

Type I, II and III collagens were purified from bovine skin, cartilage and amnion membrane, as described by (Sorsa et al., 1988). Their purity was ascertained by cyanogen bromide-cleavage peptide analysis (Bornstein and Sage, 1980). Samples of saliva, gingival crevicular fluid and gingival tissue extracts, as well as fractions of the gel filtration runs, were incubated with 1.5 PM type I, II and III collagens in TNC buffer, at 22°C for specified intervals (Sorsa et al., 1988). In studies on substrate specificity, 0.5 mM phenylmethylsulphonyl fluoride was present in the incubations to prevent the action of non-specific proteases. To activate latent forms of gingival and salivary interstitial collagenases, APMA (1 mM), gold thioglucose (1.5 mM) and hypochlorous acid (1 mM) were added to the incubations. The proteolytic activation by trypsin, chymotrypsin, neutrophil cathepsin G, plasmin, plasma kallikrein and a combination of trypsin and human gingival fibroblast prostromelysin were carried out as described earlier (Murphy et al., 1987; Okada, Harris Jr and Nagase, 1988). Analysis and densitometric quantitation of the resulting characteristic collagen reaction products was as described by (Sorsa et al., 1985). RESULTS

Gingival tissue extract, gingival crevicular fluid and whole saliva were found to contain collagenolytic activity that degraded native soluble type I collagen by vertebrate-type cleavage. No interstitial collagenase activity could be detected in parotid, sublingual or submandibular fluids. Also, only little activity

was present in saliva of edentulous subjects. The samples of saliva and gingival crevicular fluid collected from patients with periodontitis contained markedly greater amounts of interstitial collagenase than samples obtained from patients with healthy periodontium and the enzyme was mainly in an active form. Interstitial collagenase in saliva of subjects with no oral diseases was predominantly in a latent form. Interstitial collagenases extracted from inflamed gingival tissue, saliva and gingival sulcular fluid degraded preferentially type I and II collagens as compared with type III collagen. Similar to the interstitial collagenase of the human neutrophil, those of saliva and gingival crevicular fluid had an M, of 70 K. Latent forms of gingival and salivary collagenase could be efficiently actived by APMA, phenylmercuric chloride, gold thioglucose and hypochlorous acid and to a lesser extent by trypsin, chymotrypsin and neutrophil cathepsin G. No activation, however, was observed with human plasmin and plasma kallikrein. Activation of the collagenases could also be achieved by a combination of a small amount of trypsin and human gingival fibroblast stromelysin (Table 1). DISCUSSION

We provide evidence that interstitial collagenases in inflamed human gingival tissue, crevicular fluid and saliva are mainly derived ‘from neutrophils. Firstly, their I%!,was about 70 K, /which corresponds to the values for the interstitial collagenase human neutrophi1 (Sorsa et al., 1985; Hasty et al., 1986; Sorsa, 1987; Van Wart, 1989). No collagenase activity could be detected in the 40-60 K region that would correspond to the fibroblast-type interstitial collagenase (Birkedal-Hansen, 1988). Secondly, the collagenases were efficiently activated by gold thioglucose and hypochlorous acid that are potent non-proteolytic activators of the neutrophil collagenase (Weiss et al., 1985; Sorsa et al., 1987a), but are rather ineffective for fibroblast-type collagenase (Spalding, Darby and Heck, 1986). In contrast to the fibroblast-type interstitial collagenase, the neutrophil one is rather ineffectively activated by plasma kallikrein and plasmin, compared to trypsin, chymotrypsin and neutrophil cathepsin G. The two main types of

Table 1. Activation of human gingival tissue, crevicular fluid and salivary collagenases by proteolytic and non-proteolytic agents Source of enzyme Activator APMA PMC GTG HOC1 Trypsin Chymotrypsin Catheosin G STRLSTRL + sTR Plasmin Plasma kallikrein

Ginaival tissue

Saliva

Crevicular fluid

+++ +++ + +(+I ++(+) + +(+) + +(+) ++

APMA = aminoparaphenylmercuric acetate; PMC = phenylmercuric

++ -

chloride; GTG = gold thioglucose; HOC1 = hypochlorous acid; STRL = stromelysin; STRL + sTR = stromelysin and small amount of trypsin.

Collagenases in periodontal diseases

interstitial collagenase may therefore require different enzymatic pathways for their activation. Stromelysin, another member of the matrix metalloprotease family, has been recently found to activate fibroblasttype interstitial collagenase (Murphy et al., 1987). We have found now that it also activates neutrophil collagenase. Prostromelysin is probably not a member of the proteolytic arsenal of the neutrophil (Baricos et al., 1988). It is, however, produced by several connective tissue cells, following their activation (Murphy et al., 1988; Birkedal-Hansen, 1988). This may indicate an important interaction between matrix metalloproteinases from inflammatory and non-inflammatory cellular sources. In addition to the body’s own proteases some proteases of oral bacterial origin can induce the synthesis and activation of interstitial collagenase (Sorsa et al., 1987b; Uitto et al., 1987, 1989). Also substances (organomercurials, thiol-reagents, gold(I) compounds, reactive oxygen metabolites) capable of acting on the S-S or SH-groups of proteins can activate the neutrophil procollagenase. Reactive oxygen metabolites generated by neutrophils during phagocytosis and degranulation may be important and effective in vivo activators of neutrophil interstitial collagenase. Although not proven, it is plausible to assume that these factors may act in concert with proteases in the activation process (Weiss, 1989; Sorsa et al., 1989; Capodici et al., 1989). Moreover, we have recently been able to inhibit the hypochlorous acid- (a myeloperoxidase-derived reactive oxygen metabolite) induced activation of neutrophil and 70 K gingival crevicular fluid interstitial collagenases by ascorbate which is an antioxidant (Sorsa, Suomalainen and Uitto, 1990). Furthermore, we have found that chlorhexidine can inhibit the endogeneous autoactivation of human neutrophil collagenase (Sorsa et al., 1990). The oxidative activation pathway may indeed be a target for drugs capable of scavenging oxidants and/or modifying myeloperoxidase, such as tetracyclines and D-penicillinamine that are used in the therapy of periodontitis and rheumatoid arthritis (Halliwell and Wasil, 1988). Taken together it now seems clear that the body’s own collagenases are mainly responsible for the collagen degradation during periodontal diseases. Our findings favour the view that neutrophils are the main source of collagenase in adult periodontitis (Sorsa et al., 1988; Uitto et al., 1990). This does not exclude the possibility that other cells, e.g. fibroblasts, epithelial cells and macrophages, would contribute to the collagenase production. This may especially be the case where inflammation is not an overwhelming component of the disease process, e.g. in juvenile periodontitis (Suomalainen et al., 1989). More work is required to elucidate the exact roles of different metalloproteinases and their regulation in peridontal diseases. Acknowledgemenrs-This work was supported by the Emil Aaltonen Foundation, the Niilo Helander Foundation, the Finnish Medical Foundation Duodecim, the Finnish Cancer Foundation, the Paulo Foundation and the Orion Research Foundation. Veli-Jukka Uitto is a recipient of a development grant from the MRC, Canada.

195s REFERENCES

Baricos W. H., Murphy G., Zhou Y., Nguyen H. H. and Shah S. V. (1988) Degradation of glomerular basement membrane by purified mammalian metalloproteases. Biothem. J. US, 609-612.

Birkedal-Hansen H. (1988) From tadpole collagenase to family of matrix metalloproteases. .I. oral. Path. 17, 445-451. Bornstein P. and Sage H. (1980) Structurally distinct collagen types. A. Rev. Biochem. 49, 957-982. Capodici C., Muthukumaran G., Amoruso M. A. and Berg R. A. (1989) Activation of human neutrophil collagenase by cathepsin G. Inflammation 12, 245-258. Hakkarainen K., Uitto V.-J. and Ainamo J. (1988) Collagenase activity and protein content of sulcular fluid after scaling and occlusal treatment of teeth with deep periodontal pockets. J. periodont. Res. 23, 203-210. Halliwell B. and Wasil M (1988) Tetracyclines as antioxidants in rheumatoid arthritis: scavenging of hypochlorous acid. J. Rheumatol. 15, 530. Hasty K. A., Hibbs M. S., Kang A. H. and Mainardi C. L. (1986) Secreted forms of human neutrophil collagenase. J. biol. Chem. 261, 2645-2650. Larivee J., Sodek J. and Ferrier J. M. (1986) Collagenase and collagenase inhibitor activity in crevicular fluid receiving treatment for localized juvenile periodontitis. J. periodont. Res. 21, 702-705. Lindhe J. (1989) Textbook of Clinical Periodontology. 2nd edn, pp. 153-159. Murphy G., Cockett M. I., Stephens P. E. Smith B. J. and Cocherty A. J. P. (1987) Stromelysin is an activator of human fibroblast procollagenase. Biochem. J. 248, 265-268. Okada Y., Harris Jr E. D.‘and Nagase H. (1988) The precursor of a metalloprotease from human rheumatoid fibrolasts. Purification and mechanisms of activation by endopeptidases and 4-aminophenylmercuric acetate. Biothem. J. 254, 731-741. Overall C. M., Weibkin 0. W. and Thonard J. C. (1987) Demonstration of tissue collagenase activity in vivo and its relationship inflammation severity in human gingva. J. periodont. Res. 22, 81-87. Sorsa T. (1987) Activation of latent collagenase purified from human leukocytes. Stand. J. Rheumatol. 16,167-l 75. Sorsa T., Suomalainen K., Turto H. and Lindy S. (1985) Partial purification and characterization of latent human neutrophil collagenase. Med. Biol. 63, 6672. Sorsa T., Suomalainen K., Turto H. and Lindy S (1987a) Latent human leukocyte collagenase can be activated gold thioglucose and gold thiomalate, but not by auranofin. Biosci. Rep. 7, 965-968. Sorsa T., Uitto V.-J., Suomalainen K., Turto H. and Lindy S. (1987b) A trypsin-like protease from Bacteroides gingivalis: partial purification and characterization. J. periodont. Res. 22, 375-380. Sorsa T., Uitto V.-J. Suomalainen K., Vauhkonen M. and Lindy S. (1988) Comparison of interstitial collagenases from human gingiva, sulcular fluid and polymorpho- nuclear leukocytes. J. periodont. Res. 23, 3866393. _ Sorsa T., Saari H., Konttinen Y. T., Uitto V.-J. and Lindy S. (1989) Tissue destruction by human neutrophils. A letter. New Engl. J. Med. 321, 327-328. Sorsa T., Suomalainen K., Helenius J., Lindy S., Saari H., Konttinen Y. T. and Uitto V.-J. (1990) Periodontal diseases. New Engl. J. Med. 323, 1331134.’ Sorsa T., Suomalainen K. and Uitto V-J. (1990) Purification, activation and regulation of adult gingival crevicular fluid collagenase J. dent. Res. (Special Issue), in press. Spalding, D. M., Darby W. L. and Heck L. (1986) Alterations in macrophage collagenase secretion induced by gold sodium thiomalate. Arth. Rheum. 29, 75-81.

196s

T. SQRSAef al.

Suomalainen K., Sorsa T., Saxen L., Vauhkonen M., Lindy S. and Uitto V.-J. (1989) Collagenase activity in sulcular fluid of patients with juvenile periodontits. J. dent. Res. 68 (Special Issue), 952. Travis J., Giles P. J., Porcelli L., Reilly C. F., Baugh R. and Powers J. (1980) Human leukocyte elastase and cathepsin G:structural and functional characteristics. Ciba Found. Symp. 75, 5148.

Uitto V.-J. and Raeste A.-M. (1978). Activation of latent collagenase of human leukocytes and gingival fluid by bacterial plaque. J. deni. Res. 57, 844851. Uitto V.-J., Tryggvason K. T. and Sorsa T (1987) Collagenolytic enzymes in periodontal diseases. hoc. Finn. Dent. Sot. 83, 119-129.

Uitto V.-J., Larjava H., Heino J. and Sorsa T. (1989) A protease from Bacteroides gingvalis degrades cell surface

glycoproteins, and induces a secretion of collagenase and plasminogen activator. Infect. Immun. 57, 213-218. Uitto V.-J., Suomalainen K. and Sorsa T. (1990) Salivary collagenase. Origin, characteristics and relationship to periodontal health. J. periodont. Res. 25, 135-142. Van Wart H. (1990) Human neutrophil collagenase. Proc. of 1st Matrix Metalloproteinase Co& San Destin, Florida (Edited by Birkedal-Hansen H., Van Wart H., Welgus H. and Werb Z.) Gustav Fischer, Stuttgart (In press). Weiss S. J. (1989) Tissue destruction by human neutrohils. New Engl. J. Med. 320, 365-376. Weiss S. J., Peppin G., Ortiz X., Ragsdale, C. and Test S. T. (1985) Oxidative autoactivation of latent collagenase by human neutrophils. Science (Wash.) 227, 747-749.