DIFFERENCE IN EXPRESSION OF THE PLASMINOGEN ACTIVATION SYSTEM IN SYNOVIAL TISSUE OF PATIENTS WITH RHEUMATOID ARTHRITIS AND OSTEOARTHRTTIS

British Journal of Rheumatology 1996;35:416-423 DIFFERENCE IN EXPRESSION OF THE PLASMINOGEN ACTIVATION SYSTEM IN SYNOVIAL TISSUE OF PATIENTS WITH RHEUMATOID ARTHRITIS AND OSTERTHRTTIS H. K. RONDAY,*t H. H. SMTTS,t G. N. P. VAN MU1JEN,J M. S. M. PRUSZCZYNSKI,J R. J. E. M. DOLHAIN,* E. J. VAN LANGELAAN, F. C. BREEDVELD* and J. H. VERHEIJENf * Department of Rheumatology, University Hospital Leiden, ^Department of Vasular and Connetive Tissue Researh, Gaubius Laboratory TNO-PG, Leiden, \Departmem of Pathology, University Hospital, Nijmegen and ^Department of Orthopaedis, Rijnland Hospital, Leiderdorp, The Netherlands SUMMARY Proteolyti joint destrution in inflammatory and non-inflammatory arthropathy is believed to be mediated, at least in part, by the plasminogen ativation (PA) system. To further investigate possible involvement of the PA system, we quantified immunoreative urokinase-type plasminogen ativator (u-pa), tissue-type plasminogen ativator (t-pa), both plasminogen ativator inhibitors (PAI-1 and PAI-2) and u-pa-reeptor (u-par) in synovial tissue extrats of 14 s with rheumatoid arthritis () and 12 with osteoarthritis (). u-pa, PAI-1, PAI-2 and u-par onentrations were signifiantly higher in than in s. t-pa antigen levels were signifiantly lower in than in synovial tissue extrats. Immunohistohemistry was performed to ompare the distribution and staining intensity of these omponents in samples of and synovial tissue. Intense immunostaining of u-pa, u-par, PAI-1 and, to a lesser degree, PAI-2 was observed predominantly in the synovial lining of s. In s, u-pa, PAI-1, PAI-2 and u-par were barely detetable. t-pa immunostaining was restrited to the endothelial side of vasular walls in both groups. We onlude that the observed inrease of u-pa, u-par and PAI expression, distributed mainly in the synovial lining area of proliferative and invasively growing synovial tissue in s, supports a pathogeni role for the PA system in destrutive arthritis. Depressed t-pa-mediated plasminogen ativation might ontribute to delayed intra-artiular fibrin removal. KEY WORDS: Urokinase, Plasminogen ativation, Immunohistohemistry, Rheumatoid arthritis, Osteoarthritis. RHEUMATOID arthritis () is a hroni inflammatory disease, haraterized by the destrution of artiular artilage and bone. Proteolyti degradation of the extraellular matrix in inflammatory arthritis is onsidered to be mediated by proteinases like asparti, serine, ysteine and metalloproteinases [1]. Hyperplasti inflamed synovial tissue overgrows and invades the artiular artilage, and may be involved in the destrution of bone, tendons and ligaments by the prodution of proteolyti enzymes [2, 3]. Several studies suggest an important pathogeni role of the plasminogen ativation (PA) system in destrutive joint disease [4-6], but detailed knowledge of the mehanism and omponents involved is laking. The entral enzyme, plasmin, is a broad-spetrum serine protease, involved in fibrinolysis and thrombolysis as well as in the degradation of extraellular matrix that is required for normal and pathologial forms of ellular invasiveness [7]. Plasmin is able to degrade extraellular matrix diretly [8] and by ativation of latent matrix metalloproteinases [9, 10]. It is produed as inative plasminogen whih is onverted into its ative form by limited proteolysis of a single peptide bond by plasminogen ativators. Two types of plasminogen ativators have been haraterized: tissue-type plasminogen ativator (t-pa), generally Submitted 1 August 1995; revised version aepted 30 November 1995. Correspondene to: H. K.. Ronday, Department of Vasular and Connetive Tissue Researh, Gaubius Laboratory TNO-PG, Zemikefreef 9, PO Box 2215, 2301 CE Leiden, The Netherlands. viewed as being important in fibrin dissolution, and urokinase-type plasminogen ativator (u-pa), onsidered to be responsible for plasmin generation in proesses involving tissue remodelling. Other identified proteins of the PA system inlude two plasminogen ativator inhibitors, PAI-1 and PAI-2 [7], and a ell surfae u-pa binding protein, the u-pa reeptor (u-par). This reeptor binds and loalizes pro-u-pa as well as u-pa on the ell surfae [11]. In s with, a positive orrelation between PA and PAI onentrations in synovial fluid and several markers of disease ativity has been demonstrated [12-14]. Inreased levels of u-pa and PAI in synovial fluid of s with inflammatory joint disease, ompared with orresponding plasma levels, are indiative of loal plasminogen ativation [15]. Synovial fibroblasts, hondroytes, and endothelial, mononulear and polymorphonulear ells are apable of synthesizing u-pa, t-pa, PAI-1 and PAI-2. These ells ould be responsible for the loal prodution of PA and PAI in [16-19]. It may be envisaged that loal plasminogen ativation, in the proliferative and invasively growing synovial membrane, promotes the degradation of joint artilage and bone. In the present study, onentrations of several omponents of the PA system, i.e. u-pa, t-pa, PAI-1, PAI-2 and u-par, have been determined in extrats of synovial tissue of s suffering from or requiring surgery. Furthermore, the distribution of these omponents has been investigated by immunohistohemial staining of synovial tissue setions. In 416 1996 British Soiety for Rheumatology

RONDAY ET AL.: PLASMINOGEN ACTIVATION IN RHEUMATOID ARTHRITIS 417 order to investigate a possible relationship between the level of synovial plasminogen ativation and joint destrution, a omparison is made between the findings in destrutive inflammatory arthritis, e.g., with those in non-inflammatory, degenerative joint disease, i.e.. The results of this study support the pathogeni importane of the PA system in destrutive joint disease. MATERIALS AND METHODS Tissue sampling and extration Speimens of synovial tissue were obtained from 14 s and 12 s who required joint surgery for severe disease. All s, who fulfilled the established riteria [20], were operated on at the Orthopaedi Department of the University Hospital, Leiden. Patients with advaned, orresponding to grade 3-4 in the Kellgren lassifiation system [21], were operated on at the Department of Orthopaedis of Rijnland Hospital, Leiderdorp. Speimens of synovial tissue were immediately frozen in liquid nitrogen and stored at 80 C until use. For quantitative u-pa, t-pa, PAI-1, PAI-2 and u-par determination, tissue samples were homogenized in lml 0.1% (v/v) Tween 80, 0.1 M Tris-HCl buffer (ph 7.5) per 60 mg wet tissue, as desribed previously [22]. The homogenates were entrifuged twie at 8 x lo 1^ for 2.5 min, and the supematants olleted and used in the assays. Protein onentrations were determined by the method of Lowry et al. [23]. Quantitative assays u-pa antigen was measured with an enzyme-linked immunoassay that was developed in our laboratory and performed aording to Koolwijk et al. [24]. The monolonal antibodies used in this ELISA reognize all forms of u-pa (pro u-pa, ative u-pa and the u-pa/pai omplex) with omparable effiieny. The detetion limit is around 0.5 ng/ml. To assess u-pa ativity, su-pa and ative tu-pa were measured separately, using a biologial immunoassay as desribed by Dooijewaard et al. [25]. t-pa antigen was determined using the ommerial ELISA Imulyse t-pa (Biopool, Umea, Sweden). This method measures free t-pa antigen and t-pa/pai omplexes with the same sensitivity. The detetion limit is ~ 1.5 ng/ml. PAI-1 antigen was measured with Innotest PAI-1 -ELISA (Innogenetis, Antwerpen, Belgium), using monolonal mouse anti-human PAI-1 antibodies. This assay reognizes all forms of PAI-1 with the same sensitivity, with a detetion limit of ~ 5 ng/ml. PAI-2 antigen was measured with Tintelize PAI-2 ELISA (Biopool, Umea, Sweden), using monolonal mouse anti-human PAI-2 antibodies. This assay reognizes all forms of PAI-2, inluding the low-moleular-weight form and the glyosylated high-moleular-weight form. The detetion limit of this assay is ~ 6 ng/ml. u-par antigen was measured with Imubind u-par ELISA (Amerian Diagnostia In., Greenwih, CT, TABLE I Comparison of u-pa, t-pa, PAI-1, PAI-2 and u-par onentrations in synovial tissue of * with rheumatoid arthritis () and ostoarthritis (). Results are given as median and interquartile range (25-75%). The median onentrations of u-pa, PAI-1, PAI-2 and u-par antigen in synovial tissue are signifiantly higher than those in synovial tissue. The median t-pa antigen onentration is signifiantly lower in than in synovial tissue (Mann-Whitney IZ-test) Parameter (ng/mg protein) u-pa t-pa PAI-1 PAI-2 u-par (n - 14) 6.9 (4.4-7.7) 3.8 (3.0-6.1) 15.5 (6.5-30) 0.1 (0-0.32) 2 (1.1-3.5) 0.03 17 3.3 0 0.8 (n -12) (0-0.77) (10-28.8) (2.4-5.5) (0-0) (0.6-1.4) P 0.0001 0.02 0.002 0.01 0.01 USA), using polylonal rabbit anti-human u-par. This assay reognizes soluble, native u-par as well as u-par/u-pa and u-par/u-pa/pai-l omplexes. The detetion limit is ~0.1 ng/ml. All enzyme immunoassays were performed in dupliate. Antibodies Monolonal antibodies against human u-pa (#3698), polylonal goat anti-human t-pa antibodies (#387), monolonal anti-human PAI-1 (#380) and monolonal anti-human u-par antibodies were purhased from Amerian Diagnostia In. (Greenwih, CT, USA). Goat polylonal antibodies against human PAI-2 were a gift from E. Shuler (Behring Werke AG, Marburg, Germany). The haraterization of these antibodies, inluding their positive and negative ontrols, has been desribed in previous work [26, 37]. Immunohistohemistry Tissue samples of five and five s were frozen in isopentane and stored at 80 C. Cryostat setions (4 fira) were air dried overnight at room temperature and stored at 80 C until use. Setions were fixed for 10 min in aetone at 20 C before inubation with the primary antibody. With monolonal antibodies, a three-step avidinbiotin-peroxidase omplex method was applied (Vetastain Elitekit, Vetor Laboratories, Burlingame, CA, USA), as desribed by de Vries et al. [26]. Polylonal antibodies were applied to the setions, washed and inubated with peroxidase-labelled rabbit anti-goat immunoglobulin. Bound antibodies were TABLE n Immunohistohemial expression of u-pa, t-pa, PAI-1, PAI-2 and u-par in synovial tissue of s with rheumatoid arthritis () and ostoarthritis () u-pa t-pa PAI-1 PAI-2 u-par + + ± + + ± + + = strong expression, + = moderate expression, ± = weak expression, = no expression.

418 BRITISH JOURNAL OF RHEUMATOLOGY VOL. 35 NO. 5 visualized by using 3-amino-9-ethylarbazole as a substrate for peroxidase, as desribed previously [26]. Calulations and statistial analysis Antigen onentrations were expressed as nanograms of antigen per milligram of tissue protein. Differenes between the median of the measured synovial tissue onentrations of u-pa, t-pa, PAI-1 and PAI-2 in the and the group were evaluated with the non-parametri Mann-Whitney [/-test for unpaired parameters, utilizing the standard software pakage 'Solo' (BMDP Statistial Software, Los Angeles, CA, USA). Differenes were onsidered signifiant at P < 0.05. RESULTS Quantitative assays Enzyme-linked immunoassays were performed on homogenates of and synovial tissue samples to measure u-pa, t-pa, PAI-1, PAI-2 and u-par antigen 5 w O Q. O) E en 15 2 12 o O) Q. 12.00 9.60 7.20 4.80 2.40 0.00 p - 0.0001.. 1.. i 1 o o Q. CD E_ ^> < a proi CO E_ PAI-2 ng B 50 40 30 20 10 0 5.00 4.50 1.00 0.75 0.50 0.25 0.00 p - 0.010 p - 0.02 I II 1 1.1.1 rote 7.50 6.00 p - 0.010 d Bui/ 4.50 3.00 u-par 1.50 0.00 JUIIIILJ ilih I.I I, Fio. 1. Antigen onentration in synovial tissue samples of 14 s with rheumatoid arthritis and 12 s with osteoarthritis. Antigen was measured in tissue extrats by enzyme imrmmoassay and expressed as ng/mg tissue protein. (A) u-pa antigen; (B) t-pa antigen; (Q PAI-1 antigen; (D) PAI-2 antigen; (E) u-par antigen. Statistial evaluation is summarized in Table I.

RONDAY ET AL.: PLASMINOGEN ACTIVATION IN RHEUMATOID ARTHRITIS 419 levels. The results in s were ompared with those in s. u-pa antigen ould be deteted in all s and in five out of 12 s. When present, the u-pa onentration was onsiderably higher in than in synovial tissue samples (Fig. 1A). Statistial analysis revealed a signifiantly higher median value in ompared with the median value in (Table I). su-pa and ative tu-pa were determined with a bioimmunoassay. The levels of plasmin-ativatable su-pa and ative tu-pa in samples of synovial tissue of and s were measured separately. Plasmin-ativatable su-pa ould not be deteted in synovial tissue from either or s, although ative tu-pa was found (data not shown). t-pa levels were variable in both groups, but ould be deteted in all available samples (Fig. IB). In number 8, t-pa ould not be measured beause of a shortage of synovial tissue. The median onentration of t-pa antigen in the group was signifiantly lower than that in the group (Table I). In all s, and in 10 out of 12 s, PAI-1 antigen was found in the synovial tissue (Fig. 1Q. The median PAI-1 antigen level was signifiantly higher in the group than in the group (Table I). PAI-2 related antigen was found in synovial tissue of eight out of 14 s and ould hardly be deteted in one out of 12 tissue samples. PAI-2 levels in synovial tissue were highly variable (Fig. ID). The differene between both groups appeared to be statistially signifiant (Table I). In all samples of and s, the presene of u-par antigen was detetable (Fig. IE). The median u-par level was signifiantly higher in the than in the group (Table I). Immwohistohemistry In order to investigate the loalization of PA, PAI and u-par, immunohistohemistry was performed on setions of synovial tissue samples from five and five s. The degree of expression of the various parameters is summarized in Table II and shown in Fig. 2. Marked expression of u-pa was seen in all synovial tissues, espeially in the synovial lining ell area and in giant ells (Fig. 2A), but also in some plasma ells in inflammatory ell infiltrates. Less intense immunostaining was observed in some blood vessels, espeially in the media of arterioles. In the synovial tissue samples, hardly any u-pa ould be deteted (Fig. 2B), but when it was expressed, it was restrited to the synovial lining ell area. t-pa was observed in the endothelial ells of apillaries in both groups (Fig. 2C and D). Modest expression was seen in the synovial lining area. No extravasular t-pa ould be deteted in synovial tissue. Strong PAI-1 immunostaining was seen in all synovial tissue samples. It was onfined to the lining ell area (Fig. 2E) and apillaries. Hardly any PAI-1 was observed in synovial tissue samples (Fig. 2F). In only a few s, PAI-2 was observed in parts of the synovial lining area (Fig. 2G). No PAI-2 ould be deteted in any sample of s (Fig. 2H). Substantial u-par expression, predominantly in the lining ell area, was observed in synovial tissue of all s (Fig. 21). In synovial tissue, no expression or only very weak expression of u-par was seen, mainly in assoiation with mononulear ells in the interstitium (Fig. 2J). DISCUSSION In the present study, we investigated the expression and loalization of several omponents of the PA system in samples of synovium obtained from s with and. As this enzyme system is believed to be involved in extraellular proteolysis leading to joint destrution, a omparison is made between these two groups, to further investigate a relationship between synovial tissue plasminogen ativation and joint destrution. synovial tissue homogenates were found to ontain variable but signifiantly higher onentrations of u-pa, PAI-1, PAI-2 and u-par than synovium homogenates (Figs 1 and 2, Table I). Immunohistohemially, the piture was more homogeneous. u-pa, PAI and u-par appeared to be loated mainly in the synovial lining. t-pa antigen levels in synovial tissue were lower than in synovial tissue. It was loated predominantly in vessel walls and perivasularly. The inrease in u-pa, PAI-1, PAI-2 and u-par in tissue extrats of inflamed synovium presumably finds its origin in an enhaned loal prodution. Inreased levels of u-pa, PAI-1 and, in some severe ases, PAI-2 in synovial fluid of inflamed joints ompared with plasma [13-15] are indiative for generation within the joint. Furthermore, ell and tissue ulture studies have demonstrated that synovial fibroblasts, but also hondroytes, monoytes/marophages and endothelial ells, are apable of synthesizing u-pa and PAI. In vivo, their prodution may be influened by ytokines suh as interleukin-1, tumour nerosis fator-a and granuloyte-marophage olonystimulating fator [16-19]. A loal inrease of u-pa at the expense of t-pa in synovial tissue ompared with synovial tissue is in line with previous observations in synovial fluid [15] and in inflammatory bowel disease [27]. The altered ratio of t-pa (high fibrin affinity) to u-pa (low fibrin affinity) ould reflet a shift from fibrin(ogen) degradation towards extraellular matrix degradation. This might result in protration of fibrin removal on the one hand and enhaned proteolyti degradation of joint bone and artilage on the other hand [15]. The quantitative ELISA data and the immunohistohemial analyses show a lear differene between and synovial tissue. This obvious pattern points towards an inreased expression of all those enzyme system omponents believed to be involved in

420 BRITISH JOURNAL OF RHEUMATOLOGY VOL. 35 NO. 5 fh (B (C <P (E FIG. 2. (A-E).

RONDAY ET AL.: PLASMINOGEN ACTIVATION IN RHEUMATOID ARTHRITIS 421 in (G) (H) (I) 11 (J) FIG. 2. Immunohistohemial staining for omponent! of the plasminogen ativation system in rheumatoid arthritis () and in osteoarthritis (). Immunoreativity for u-pa; (A) distint expression in synovial lining ells and giant ells in ; (B) hardly any expression in. Immunoreativity for t-pa: (C, D) positive staining in endotheliaj ells in synovial blood vessels in and. Immunoreativity for PAH: (E) marked staining of synovial lining ells in ; (F) no staining in. Immunoreativity for PAI-2: (G) foal expression in synovial lining ells in ; (H) no expression in. Immunoreativity for u-par: (I) distint staining of synovial lining ells in ; (J) no staining of synovial ells in. Magnifiation: 250 x.

422 BRITISH JOURNAL OF RHEUMATOLOGY VOL. 35 NO. 5 tissue remodelling. A omparison between neoplasti tissue and invasively growing inflamed synovial tissue has been made before [7]. Indeed, inreased u-pa and PAI-1 levels in tissue extrats or setions have been found in various malignant tumours [26,28-30], and support the involvement of the PA system in extraellular matrix degradation, but how ould an inreased u-pa prodution lead to higher proteolyti ativity when its inhibitors PAI-1 and PAI-2 are elevated? The answer may be found in the upregulation of u-par expression at the ell surfae. First, in vitro studies have shown that several ell types are apable of binding u-pa at speifi sites of the ell surfae, whereas PA1 was found at a different loation [31]. Seond, a differential inhibition of soluble and ell surfae reeptor-bound u-pa has been demonstrated [32], allowing enzymati ativity of reeptor-bound u-pa even in a PAI-rih environment. Third, o-loalization of u-pa/u-par and plasminogen on the ell surfae results in ~ 100-fold more effiient ativation of plasminogen than in the fluid phase [33]. Furthermore, "plasmin bound to the ell surfae is resistant to o2-antiplasmin [34]. Fourth, the interation of u-pa with its ell-bound reeptor has been shown to strongly enhane the degradation of extraellular matrix [35]. Loss of the surfae u-pa ativity by bloking the interation between the reeptor and its ligand has been shown to inhibit invasive growth [36]. These phenomena ould explain net loal proteolyti ativity in the presene of inreased inhibitor. Our idea about the mehanism of plasmin-mediated joint destrution is that inreased u-pa prodution by inflamed hypertrophi and hyperplasti synovial tissue, overgrowing artilage in a 'tumour-like* manner, ould lead to ativation of the readily available plasminogen on the ell surfae at sites oupied by u-par. This loalized formation of ative plasmin, apable of degrading extraellular matrix diretly and by ativation of matrix metalloproteinases, may subsequently result in direted, proteolyti degradation of bone and artilage. In onlusion, the inreased expression of u-pa, u-par, PAI-1 and PAI-2 in arthriti synovium ompared with non-inflamed synovial tissue fits in with the onept of a loalized, u-pa-mediated, plasmindependent degradation of artiular strutures, finding its origin in inflamed synovial tissue. ACKNOWLEDGEMENT The authors wish to thank Dr H. W. Verspaget for providing the PAI-2 antibody. REFERENCES 1. Wrb Z, Alexander CM. Proteinass and matrix degradation. In: Kelley WN, Harris ED, Ruddy S, Sledge CB, eds. Textbook of rheumatology, 4th edn. Philadelphia: W B Saunders Company, 1993:248-69. 2. Yates DB, Sott JT. Rheumatoid synovitis and joint disease. Ann Rheum Dis 1975^4:1-6. 3. Zvaifler NJ, Firestein GS. Pannus and pannovtes, alternative models of joint destrution in rheumatoid arthritis. Arthritis Rheum 1994^7:783-9. 4. Hamilton JA. Plasminogen ativator ativity of rheumatoid and nonrheumatoid synovialfibroblasts.j Rheumatol 1982;9:834-42. 5. Inman RD, Harpel PC. <x2-plasmin inhibitor-plasmin omplexes in synovialfluid.j Rheumatol 1986;13:535 7. 6. Kikuhi H, Tanaka S, Matsuo O. Plasminogen ativator in synovial fluid from s with rheumatoid arthritis. J Rheumatol 1987;14:439-45. 7. Vassali J-D, Sappino A-P, Belin D. The plasminogen ativator/plasmin system. J Clin Invest 1991;88:1067-72. 8. Quax PHA, Van Muijen GNP, Pedersen N et al. The plasminogen ativator system in extraellular matrix degradation. Fibrinolysis 1992;6(suppl. 4):41-4. 9. Werb Z, Mainardi MD, Vater CA, Harris ED. Endogenous ativation of latent ollagenase by rheumatoid synovial ells. N Engl J Med 1977;296: 1017-23. 10. Murphy G, Atkinson S, Ward A, Gavrilovi J, Reynolds JJ. The role of plasminogen ativators in the regulation of onnetive tissue metalloproteinases. In: Brakman P, Kluft C, eds. Plasminogen ativation in fibrinolysis, in tissue remodeling and in development. Ann N Y Aad Sd 1992;667:1-12. 11. Vassali J-D, Baino D, Belin D. A ellular binding site for the Mr 55.000 form of human plasminogen ativator, urokinase. J Cell Biol 1985;10O:86-92. 12. Mohan E, Uhl J. Elevations in synovial fluid plasminogen ativator in s with rheumatoid arthritis. / Rheumatol 1984;ll:123-8. 13. Kummer JA, Abbink JJ, De Boer JP et al. Analysis of intraartiular fibrinolyti pathways in s with inflammatory and noninflammatory joint diseases. Arthritis Rheum 1992;35:884-93. 14. Saxn T, Lander I, Geborek P. Plasminogen ativators and plasminogen ativator inhibitors in synovial fluid. Differene between inflammatory joint disorders and osteoarthritis. J Rheumatol 1993;20.-91-6. 15. Brommer EJP, Dooijewaard G, Dijkmans BAC, Breedveld FC. Depression of tissue-type plasminogen ativator and enhanement of urokinase-type plasminogen ativator as an expression of loal inflammation Thromb Haemostasis 1992;68:180-4. 16. Campbell IK, Piolli DS, Roberts MJ, Muirden KD, Hamilton JA. Effets of tumor nerosis fator a and 0 on resorption of human artiular artilage and prodution of plasminogen ativator by human artiular hondroytes. Arthritis Rheum 1990;33:542-52. 17. Medalf RL, Hamilton JA. Human synovial fibroblasts produe urokinase-type plasminogen ativator. Arthritis Rheum 1986;29:1397-401. 18. Kirhheimer JC, Remold HG, Wanivenhaus A, Binder BR. Inreased proteolyti ativity on the surfae of monoytes from s with rheumatoid arthritis. Arthritis Rheum 1991^4:1430-3. 19. Hamilton JA, Hart PH, Leizer T, Vitti GF, Campbell IK. Regulation of plasminogen ativator ativity in arthriti joints. / Rheumatol 1991;18(suppl. 27):106-9. 20. Amett FC, Edworthy SM, Bloh DA et al. The Amerian Rheumatism Assoiation 1987 revised riteria for the lassifiation of rheumatoid arthritis. Arthritis Rheum 198821:315-24. 21. Kellgren JH, Lawrene JS. Radiologi assessment of osteoarthritis. Ann Rheum Dis 1957;16:494-502. 22. De Bruin PAF, Verspaget HW, Griffioen G, Nap M, Verheijen JH, Lamers CBHW. Plasminogen ativator ativity and omposition in human oloretal arinomas. Fibrinolysis 1989;l:57-62.

RONDAY ET AL.: PLASMINOGEN ACTIVATION IN RHEUMATOID ARTHRITIS 423 23. Lowry OH, Rosebrough NJ, Farr AL, Randall R. Protein measurement with folin phenol reagent. / Biol Chem 1951;193:265-275. 24. Koolwijl P, Miltenburg AMM, Van Erk MGM et al. Ativated gelatinase-b (MMP-9) and urokinase-type plasminogen ativator (u-pa) in synovial fluids of arthritis s. Correlation with linial and experimental parameters. J Rhewnatol 1995^2:385-93. 25. Dooijewaard G, van Iersel JJL, Brommer EJP. Quantitation of pro-uk, UK and UK-inhibitor levels in plasma of s and healthy men. Fibrinolysis 1986; soppl 1:142. 26. De Vries TJ, Quax PHA, Denijn M et al. Plasminogen ativators, their inhibitors, and urokinase reeptor emerge in late stages of melanoyti tumor progression. Am J Pathol 1994; 144:70-82. 27. De Bruin PAF, Crama-Bohbouth G, Verspaghet HW et al. Plasminogen ativators in the intestine of s with inflammatory bowel disease. Thromb Haemostasis 1988;60:262-^. 28. Graeff H, Harbek N, Pahe L, Wilhelm O, Jinike F, Shmitt M. Prognosti impat and linial relevane of tumor-assoiated proteases in breast aner. Fibrinolysis 1992;6(suppl. 4):45-53. 29. Ganesh S, Sier CFM, Griffioen G et al. Prognosti relevane of plasminogen ativators and their inhibitors in oloretal aner tissue. Caner Res 1994;54: 4065-71. 30. Ossowski L. In vivo invasion of modified horioallantoi membrane by tumor ells: the role of ell-surfae bound urokinase. J Cell Biol 1988;107:2437-45. 31. Pollanen J, Saksela O, Salonen EM et al. Distint loalization of urokinase-type plasminogen ativator and its type-1 inhibitor under ultured human fibroblasts and saroma ells. J Cell Biol 1987;104:1085-96. 32. Shwartz BS. Differential inhibition of soluble and ell surfae reeptor-bound single-hain urokinase by plasminogen ativor inhibitor type 2. J Biol Chem 1994;269:8319-23. 33. Mahanda N, Shwartz BS. Single hain urokinase. Augmentation of enzymati ativity upon binding to monoytes. J Biol Chem 1991;266:14580-4. 34. Blasi F. Urokinase and urokinase reeptor: a pararine/ autorine system regulating ell migration and invasiveness. Bioessays 1993; 15:105-10. 35. Quax PHA, Pedersen N, Masui MT et al. Complementation between urokinase-produing and reeptor-produing ells in extraellular matrix degradation. Cell Regul 1991^:793^803. 36. Kobayashi H, Gotoh J, Fujie M, Shinohara H, Moniwa N. Inhibition of metastasis of Lewis lung arinoma by a syntheti peptide within growth fator-like domain of urokinase in the experimental and spontaneous metastasis model. Int J Caner 1994;57:727-33. 37. D Vries TJ, Mooy CM, Van Balken MR et al. Components of the plasminogen ativation system in uveal melanoma. A linio-pathologial study. / Pathol 1995;175:59-67.