Identification of plasma kallikrein as an activator of latent collagenase in rheumatoid synovial fluid

Biochimica et Biophysica A cta, 702 (1982) 133-142

133

Elsevier Biomedical Press BBA31081

IDENTIFICATION OF PLASMA KALLIKREIN AS AN ACTIVATOR OF LATENT COLLAGENASE IN RHEUMATOID SYNOVIAL FLUID HiDEAKI NAGASE a.,, TIM E. CAWSTON b MALCOLM DE SILVA c and ALAN J. BARRETT a

, Biochemistry Department and h Cell Physiology Department, Strangewavs Laborato~, Wort's Causeway, Cambridge CBI 4RN and ' Rheumatology Department, New Addenbrooke's Hospital, Hills Road, Cambridge CB2 2QQ, (U.K.) (Received June 1st, 1981) (Revised manuscript received October 26th, 1981)

Key words: Plasma kallikrein; Collagenase activation; (Rheumatoid synovial fluid)

Rheumatoid synovial fluid contains an activator of latent collagenase from culture medium of pig synovium. The activator was purified by gel chromatography on Ultrogei AcA 44 and affinity chromatography on soybean trypsin inhibitor coupled to Sepharose 4B. The purified material was homogeneous on SDSpolyacrylamide gel electrophoresis with M, 88000. The activator had limited proteolytic activity against azo-casein, but showed amidase activity on Pro-Phe-Arg-NMec, Z-Phe-Arg-NMec, D-Val-Leu-Arg-NPhNO2 and D-Pro-Phe-Arg-NPhNO2, with an optimum at pH 8.0. Activity was completely inhibited by diisopropyl fluorophosphate, soybean trypsin inhibitor, leupeptin and Pro-Phe-Arg-CH2CI, whereas lima bean trypsin inhibitor, Tos-Lys-CH2C!, a specific inhibitor of factor Xlla from maize, EDTA and iodoacetate were not inhibitory. These properties of the activator suggested that it might be plasma kallikrein (EC 3.4.21.34), and the possibility was further examined. The activator was treated with [3H]diisopropyl fluorophosphate, and run in SDS-polyacrylamide gel electrophoresis with reduction; a radioautograph of the gel showed a pair of [3H]diisopropyi phosphoryl-labelled bands ( M r 36000 and 34000) identical to those obtained with authentic plasma kallikrein. Double immunodiffusion with monospecific antiserum against human plasma kallikrein confirmed the identification. This is the first demonstration of collagenase-activating activity of plasma kallikrein, and raises the possibility that activation of prokallikrein in the inflamed joint space may contribute to the disease process not only by the production of bradykinin, but also by activating latent collagenase.

Introduction Collagen breakdown causes irreversible destruction of articular cartilage in various arthritides [1]. This breakdown has been attributed to col-

* To whom correspondence should be addressed at (present address): Connective Tissue Disease Section, Department of Medicine, Dartmouth Medical School, Hanover, N.H. 03755, U.S.A. Abbreviations: -NMec, 4-methyl-7-coumarylamide; -NPhNO 2, 4-nitroanilide; -ONap, -naphthyl ester. 0167-4838/82/0000-0000/$02.75 © 1982 Elsevier Biomedical Press

lagenase, which may arise from the proliferative lesion of rheumatoid arthritis [2,3] or inflammatory cells in the joint space [4,5]. Although earlier reports of collagenase in the culture medium of rheumatoid synovium described the active form of the enzyme [6,7], more recent investigations have shown that the enzyme is secreted into the culture medium in latent form [8-10], so that the full collagenase activity cannot be detected unless the culture medum is first treated with trypsin or mercurial compounds [ 11,14]. Latent forms activatable by these agents have been found in cultures

134

of various tissues [11-15] and cells [16,17], and also in homogenates of rat uterus [ 18], rat Graafian follicles [19] and human uterine cervix [20]. It has not been established, however, how the latent enzyme is activated in vivo, and an understanding of this process seems essential for the assessment of the possible physiological or pathological roles of the enzyme. Kruze and Wojtecka [5] reported that rheumatoid synovial fluid contains a non-dialyzable and heat-labile factor which activates crude polymorphonuclear leucocyte latent collagenase. This finding was confirmed by other investigators [2123], but the activator has not been identified. We have purified an activator of latent collagenase from rheumatoid synovial fluid, and identified it as plasma kallikrein (EC 3.4.21.34). A possible pathological function of plasma kallikrein in tissue damage via the activation of latent collagenase is discussed. Methods

Materials Synovial fluid aspirated from knee joints of 19 patients with rheumatoid arthritis (17 classical and two definite rheumatoid arthritis according to the American Rheumatism Association) was collected in sterile plastic containers and centrifuged at 10000 × g for 20 min at 4°C; the cell-free supernatant fluid was stored - 2 0 ° C until required. Albumin (bovine serum, crystallized), carbonic anhydrase (bovine erythrocyte) cytochrome c (horse heart) type II-A, hyaluronidase (bovine testes) type I-S, ovalbumin (crystallized), phosphorylase a (rabbit muscle, 2 × crystallized), transferrin (human), trypsin (pig pancreas, crystallized) type IX, trypsin inhibitor (soybean) type I-S and Tos-Phe-CH2CI were from Sigma (London) Chemical Co. (Poole, Dorset, U.K.). Aprotinin (as Trasylol) was from Bayer, A.G., (WuppertalElbefeld, F.R.G.). Tos-Lys-CHzC1 and 7-amino-4methylcoumarin were from Bachem (Bubendorf, Switzerland). Pro-Phe-Arg-CH2Cl was the gift of Dr. C. Kettner (Brookhaven National Laboratory, New York, NY). 4-Aminophenylmercuric acetate was from Aldrich Chemical Co. (Gillingham, Kent, U.K.). Iodoacetate was from BDH Chemicals Ltd. (Poole, Dorset, U.K.) and EDTA (disodium) form

Fisons (Loughborough, Leics, U.K.). Sephadex 4B was from Pharmacia (G.B.) Ltd., (Middlesex, U.K.). Ultrogel AcA 44 was from LKB Instruments, (South Croydon, Surrey, U.K.). [1,33H]Diisopropyl fluorophosphate (TRK 207, 3 Ci/mmol) was from the Radiochemical Centre, Amersham (Buckinghamshire, U.K.). Pro-Phe-Arg-NMec, Z-Phe-Arg-NMec, BocVal-Pro-Arg-NMec, Boc-Ile-Glu-Gly-Arg-NMec, glutaryl-Gly-Arg-NMec and leupeptin were from Protein Research Foundation (476 Ina, Minoh, Osaka, Japan). DPro-Phe-Arg-NPhNO2, DVal-Leu -Arg-NPhNO 2 and DVaI-Leu-Lys-NPhNO2 were from Kabi Vitrum (Ealing, London, U.K.). Z-Ala2-ONap was prepared by Dr. C.G. Knight, Strangeways Laboratory (Cambridge, U.K.). Human plasma kallikrein was purified by the method of Nagase and Barret [24]: the enzyme was homogeneous by electrophoretic and immunological criteria. Factor XIIa inhibitor (maize) [25] was a gift from Dr. Y. Hojima (National Heart, Lung and Blood Institute, Bethesda, MD, U.S.A.). Turkey ovomucoid and chicken ovoinhibitor were prepared as described by Lineweaver and Murray [26] and Barrett [27], respectively. Azo-casein was prepared by treating casein with diazotized sodium sulphanilate generally as described by Charney and Tomarelli [28]; the material had an A/6%6lcm of 40. (Soybean trypsin inhibitor)-Sepharose was prepared as described previously [24].

Preparation of pig synovial latent collagenase. Pig synovial latent collagenase was prepared as described by Cawston et al. [29]. Activation of latent collagenase. 4-Aminophenylmercuric acetate was used to obtain maximal activation of latent collagenase. It was added as a 10 mM solution in 0.2 M NaOH (pH 10.0) to latent pig synovial collagenase (0.1-0..2 units) in 25 mM sodium cacodylate buffer, pH 7.2, containing 1 M NaC1, 10 mM CaCI2, 0.05% Brij 35 and 0.02% NaN 3 to give a final concentration of 0.67 mM. These conditions gave maximal activation of latent collagenase assayed by the 'diffuse-fibril' method of Cawston and Barrett [30]. Activation of the latent collagenase with rheumatoid synovial fluid activator was performed by incubating the mixture for 48 h at 37°C, followed by addition of diisopropyl fluorophosphate to a final concentration of 5 mM. Activation with hu-

135 man p.lasma kallikrein was carried out by preincubating for 2h at 37°C at the concentrations shown in Fig. 5, and the reaction was stopped by the addition of 20-fold molar excess of soybean trypsin inhibitor. After activation the collagenase was assayed with and without 0.67 mM 4aminophenylmercuric acetate to determine the total potential activity and the proportion of collagenase activated by plasma kallikrein. Enzyme assays. Activity against peptide 4methyl-7-coumarylamides was measured at 37°C by the microassay method described previously [24]. The relative fluorescence of the reaction product was read with a Locarte fluorometer (Locarte Company, London, U.K.) using excitation filter L F / 2 (340-380 nm) with emission monochromator set at 460 nm. The fluorometer was adjusted to read 500 arbitrary units with 1 #M 7-amino-4-methylcoumarin. The molar concentration of plasma kallikrein was calculated from the specific activity of 9.7 in this assay [24]. Activity against nitroanilide derivatives was measured spectrophotometricaUy. The incubation mixture contained 200 #1 0.1 M Tris-HC1 buffer, pH 7.8, 25 #1 enzyme and 25 /~1 5 mM substrate. After incubation for 15 min at 37°C the reaction was stopped by adding 100 ~1 50% (v/v) acetic acid and the absorbance at 405 nm was read in a 0.3-ml cuvette with 1-cm light path. Activity against Z-Ala-2-ONap was measured at 37°C, pH 7.8, as described by Starkey and Barrett [31]. Activity against azo-casein was measured at 37°C in incubation mixtures (0.4 ml) containing 0.2 ml of 0.1 M Tris-HC1 buffer, pH 7.8, containing 5mM CaC12, and 0.04% NAN3, 0.1 ml of enzyme in the same buffer and 0.1 ml of 6% (w/v) azo-casein. The reaction was stopped by adding 2.5 ml of 3% (w/v) trichloroacetic acid and the mixture filtered. The absorbance at 366 nm of the trichloroacetic acid-soluble reaction products was determined and AA366calculated by subtraction of the blank value. Inhibition studies. Inhibition studies were done by measuring the enzymic activity against ProPhe-Arg-NMec. In general, the enzyme was preincubated with the inhibitor for 15 min at 20°C at the final concentration given in Table IV before addition of the substrate. Pro-Phe-Arg-CH2C1 , Tos-Phe-CH2C1 and Tos-Lys-CH2C1 were allowed

to react for 2h at 20°C before addition of the substrate. Tos-Phe-CH2CI was dissolved in Me2SO at a concentration at least 20-times higher than that of the final reaction mixture.

Preparation of anti-(human plasma kallikrein) antiserum. Rabbit anti-(human plasma kallikrein) antiserum was prepared by intradermal injection of 100/~g of purified plasma kallikrein in complete Freund's adjuvant to New Zealand white rabbits. After 2 and 8 weeks the second and the third injections were given intramuscularly with 100/~g of antigen emulsified in incomplete Freund's adjuvant, and rabbits were killed 10 weeks after the first injection. Antiserum was rendered monospecific for plasma kallikrein by adsorption with proteins that had been bound to (soybean trypsin inhibitor)-Sepharose from plasma (without activation of prokallikrein) and eluted with 5 mM NaOH. SDS-Polyacrylamide gel electrophoresis. This was done as described by Barrett et al. [32] with gels of 7% total acrylamide concentration. Fluoro-radioautography. 'Fluorography' was carfled out as follows. Purified latent collagenase activator was reacted with 0.1 mM [3H]diisopropyl fluorophosphate in 10 mM Tris-HC1, pH 7.8, for 15 min at 20°C and subjected to SDSpolyacrylamide gel electrophoresis. The gel was stained, photographed, impregnated with 2,5diphenyloxazole and dried according to the method of Laskey and Mills [33] before exposure to a Kodak X-OMAT H film for 92 h. The film was developed with a Kodak RPX-OMAT processor. Protein determination. Protein concentrations were determined either by the method of Lowry et al. [34] with crystalline bovine serum albumin as standard, or by measurement of A280. Results

Association of the latent collagenase activator with activity against Pro-Phe-Arg-NMec in gel chromatography. Synovial fluid (2 ml) from a patient with rheumatoid arthritis was treated with 100/~g of hyaluronidase for 40 min at 20°C and applied to an Ultrogel AcA 44 column. Fig. 1 show's a typical chromatographic pattern of rheumatoid synovial fluid. A single peak of collagenase activator was resolved with an apparent M r of 82000. This activity was found to coincide exactly with

136

i

if!

(A)

Vo

BSA Ovalb

Cyt c A

8

44 n

1t "

D 100

0.3

0.5

0.7

0.9

Vel Vt

Fig. 1. Gel chromatography of rheumatoid synovial, fluid on Ultrogel AcA 44. Rheumatoid synovial fluid (2 ml) treated with 100 #g hyaluronidase for 30 min at 20°C was run on an Ultrogel AcA 44 column (1.5 X90 cm). A. Activation of latent collagenase was measured by preincubation of 20 #I of each fraction and 20 #1 of latent pig synovial collagenase (0.06-0.08 units) for 48 h at 37°C, followed by the diffuse fibril collagenase assay. Activity against Pro-Phe-Arg-NMec was measured by incubating 10 #1 of the sample with 10 #1 of 0.5 mM Pre-Phe-Arg-NMec in 20 mM Tris-HC1 buffer, pFI 7.8 at 37°C for 3 min. • • , hydrolysis of Pro-Phe-Arg-NMec; O . . . . . . O, activation of latent pig synovial collagenase; . . . . . . , protein A2so. B. Inhibitory activities against the latent collagenase activator and trypsin were measured by preincubation of samples (20 #1) with either 20/tl of the pooled activator (12 nmol Pro-Phe-Arg-NMec hydrolized/min per ml) from the Ultrogel AcA 44 column, or trypsin (10 ~g/ml) in 10 mM Tris-HCl buffer, pH 7.8, containing 0.2 M NaC1 and 10 mM CaCI 2. The residual activities were measured against Pro-Phe-Arg-NMec. The protein markers used for the calibration of the column were bovine serum albumin (BSA), ovalbumin (Ovalb) and cytochrome c (Cyt c). Ve is elution volume and Vt is the total column bed volume. • . . . . . . • inhibitory activity against the latent collagenase activator; O O, inhibitory activity against trypsin.

the activity against Pro-Phe-Arg-NMec as show in Fig. 1A. Hyaluronidase used in this study to decrease the viscosity of synovial fluid had no action on latent collagenase (activating or destroying) or on Pro-Phe-Arg-NMec and contained no acrosin

detectable in assays with Bz-Arg-NPhNO 2 made as described by Morton [35]. Initial inhibition studies with the partially purified activator showed complete inhibition by diisopropyl fluorophosphate and soybean trypsin inhibitor, but not by iodoacetate or EDTA. This suggests that the activator might belong to the class of serine proteinases [36]. The activity against Pro-Phe-Arg-NMec was also inhibited by diisopropyl fluorophosphate, and because of the close association of activities in the column effluent it seemed possible that both were due to the same serine proteinase. For convenience, most assays were made with the synthetic substrate during the further purification of the enzyme. Purification of the activator. All the purification procedures were carried out at 4°C unless otherwise mentioned. The results of the purification are summarized in Table I. A pool of rheumatoid synovial fluid (30 ml) was treated with 1.5 mg hyaluronidase for 30 rain at 20°C, and centrifuged to remove insoluble material. The supernatant was run on an Ultrogel AcA 44 column (3 X 120 cm) in 20 mM Tris-HC1 buffer, pH 7.8, containing 0.2 M NaC1 and 0.02% NAN3, and the fractions containing activity against Pro-Phe-Arg-NMec were combined. The yield in this step was 320% with 14-fold increase of specific activity. The increase in total activity was attributable to the removal of a high molecular weight inhibitor (apparent M r 160,000 by gel chromatography on Ultrogel AcA 44 (Fig. 1B)). Inhibition of the partially purified activator by soybean trypsin inhibitor led us to employ (soybean trypsin inhibitor)-Sepharose affinity column for the further purification. The pool of active fractions from the previous step (54 ml) was applied to a (soybean trypsin inhibitor)-Sepharose column (1.2X3cm) equilibrated with 20 mM Tris-HC1 buffer, pH 7.8, containing 1 M NaCI and 0.1% Brij 35, at a flow rate of 15-20 ml/h. After the column had been washed with at least 5 bed volumes of the same buffer, the amidase activity was eluted with 5 mM NaOH, pH 11.2, and 1-ml fractions were collected in tubes containing 0.1 ml of 1 M Tris-HC1 buffer, pH 7.8, for the immediate neutralization of the alkaline solution. The affinity

137 TABLE I PURIFICATION OF LATENT COLLAGENASE ACTIVATOR FROM RHEUMATOID SYNOVIAL FLUID 1 unit of activity represents the hydrolysis of I #mol Pro-Phe-Arg-NMec/min at 37°C. Activity Step

Synovial fluid (hyaluronidase-treated) Ultrogel AcA 44 (Soybean trypsin inhibitor)-Sepharose

Volume Protein (ml) ( A280 units) 30 54 7

840 192 0.030

chromatography step gave 2600-fold purification with a recovery of 42%. To assess the purity of the activator, the material was treated with [3H]diisopropyl fluorophosphate and subjected to SDS-polyacrylamide gel electrophoresis (without reduction) and fluorography. The sample showed a protein band of M r 88000 coinciding with the radioactivity (Fig. 2). Enzymic properties of the activator. The purified protein activated latent pig synovial collagenase (Table II), but the activation was completely blocked by 1 m M diisopropyl fluorophosphate. The activator itself did not have any collagenolytic activity. We examined the specificity of the purified activator with coumarylamide and nitroanilide substrates (Table III). The activator showed strong activity against arginyl bonds, particularly when a hydrophobic residue preceded the arginine residue. It is notable that DVal-Leu-Lys-NPhNO 2 was not hydrolyzed, but the corresponding substrate in which the lysine residue was replaced by an arginine was very susceptible to hydrolysis by the activator. The B chain of insulin which contains the sequence -Gly-Glu-Arg-Gly- was not hydrolyzed (1 mg of the substrate was incubated with 3 #g of the activator at p H 7.8, 37°C for 20h, and the result was examined by high-voltage paper electrophoresis and paper chromatography for the second dimension). The activator showed very weak proteolytic activity against azo-casein as compared with trypsin; 1 #g of the activator hydrolyzed only 9.7 ~tg azocasein (15 mg total) in 2 h at 37°C, whereas the

Purification Yield factor (%)

T o t a l Specificactivity (units) (unit/A2s0 unit) 0.09 0.288 0.120

0.00011 0.0015 4.00

I 14 37000

100 320 130

same amount of trypsin hydrolyzed 534 gg in 30 min. The activator produced no activation of plas-

M_r (xlO "3) 150

77-"--

i

a

68-"--" I

9

(a)

(b)

(c)

Fig. 2. SDS-polyacrylamidegel electrophoresis. Purified latent collagenase activator was treated with 0.1 mM [3H]diisopropyl fluorophosphate and subjected to SDS-polyacrylamidegel electrophoresis without reduction. (b) The gel was stained for protein with Coomassieblue R-250 and photographed. (c) The gel was then impregnated with 2,5-diphenyloxazole,dried and fluorographed. (a) Protein standards (immunoglobulin G, transferrin, bovine serum albumin, carbonic anhydrase) were also run without reduction.

138 TABLE II ACTIVATION OF LATENT COLLAGENASE Latent pig synovial collagenase (0.06-0.08 units) was preincubated with the purified activator for 48 h at 37°C and the reaction was stopped by adding diisopropyl fluorophosphate to a final concentration of 5 mM. Collagenase was then assayed in the diffuse fibril assay [30]. Activation is expressed as a percentage of that obtained with 0.5 mM 4-aminophenylmercuric acetate or trypsin at a concentration of 2.5/~g/ml.

Latent Latent Latent Latent Latent

Collagenase activity (c.p.m.)

Activation (%)

0 1860 1810 1260 20

0 100 100 68

collagenase + buffer coltagenase+4-aminophenylmercuricacetate (0.5 mM) collagenase + trypsin (2.5/~ g/ml) collagenase + purified activator (7.4 #g/ml) collagenase + purified activator (7.4 #g/ml) + diisopropyl fluorophosphate (1 mM)

minogen or chymotrypsinogen (data not shown). The results of tests of potential inhibitors of the activator are shown in Table IV. The enzymic activity against Pro-Phe-Arg-NMec was completely inhibited by diisopropyl fluorophosphate, soybean trypsin inhibitor, .leupeptin and Pro-PheArg-CH2C1, and partially inhibited by aprotinin. Lima bean trypsin inhibitor, Tos-Lys-CH2C1 and a specific inhibitor of factor XIIa from maize were

TABLE IV EFFECT OF INHIBITORS ON A LATENT COLLAGENASE ACTIVATOR FROM RHEUMATOID SYNOVIAL FLUID The enzyme was preincubated with each inhibitor for 15 min at 20°C at the final concentration shown, and the residual activity was measured against Pro-Phe-Arg-NMec at pH 7.8. Inhibitor

SUBSTRATE SPECIFICITY OF COLLAGENASE ACTIVATOR Assays were made with methylcoumarylamides and Z-AlaONap at a final concentration of 0.25 mM, and nitroanilides at 0.5 mM.

Z-Phe-Arg-NMec Pro-Phe-Arg-NMec Boc-Val-Pro-Arg-NMec Glt-Gly-Arg-NMec Boc-Ile-Glu-Gly-Arg-NMec DPro-Phe-Arg-NPhNO2 DVaI-Leu-Arg-NPhNO2 DVal-Leu-Lys-NPhNO2 Bz-Arg-NPhNO2 Bz-Tyr-NPhNO2 Z-Ala-ONap

Concentration /zg/ml

TABLE III

Substrate

1

Spec. act. (/zmol/min per mg) 4.7 4.0 0.68 0.04 0.13 19.2 3.2 < 0.01 < 0.01 < 0.0 I < 0.01

None Soybean trypsin inhibitor Aprotinin

Lima-bean trypsin inhibitor Ovoinhibitor (chicken) Ovomucoid (turkey) Factor XIIa inhibitor (maize) Diisopropyl fluorophosphate Pro-Phe-Arg-CH2Cl Tos-Lys-CH2CI Tos-Phe-CH 2C1 Iodoacetate Leupeptin EDTA

/~M

Activity (%) 100 2 12 2 23 74

5 0.5 25 5 0.5 250 100 100

107 93 91

500 2000 1 1 000

1000 10000 I00 5 5000

100 0 0

102 100 110 3 36 105

139 not inhibitory. N o inhibition was observed with iodoacetate or EDTA. We conclude that the substrate specificity and inhibitor sensitivity of the activator are closely similar to those of plasma kallikrein [37].

Identification of the latent collagenase activator as plasma kallikrein. To test the possibility that the latent collagenase activator is plasma kal-

(A)

likrein, both activator and authentic plasma kallikrein were labelled with [3H]diisopropyl fluorophosphate and subjected to SDS-polyacrylamide gel electrophoresis with and without reduction, and fluorographed. We have previously shown that plasma kallikrein exhibits a characteristic doublet of diisopropyl phosphoryl-labelled chains on SDSpolyacrylamide gel electrophoresis with reduction [24]. The fluorograph (Fig. 3A) showed that the electrophoretic pattern of [3H]diisopropyl phosphoryl-labelled activator was identical to that of [3 H]diisopropyl phosphoryl-labelled kallikrein: run with reduction, [3H]diisopropyl p h o s p h o r y l labelled activator showed a characteristic doublet with M r 36000 and 34000. Double immunodiffusion with mono-specific antiserum against human plasma kallikrein showed the activator to be immunologicaly identical to plasma kallikrein (Fig. 3B).

Activation of latent collagenase by plasma kallikrein. The activation of latent collagenase by plasma kallikrein was demonstrated directly in the standard system (see the Methods section). Kallikrein at a concentration of 6 2 . 5 / ~ g / m l at 37°C for 2 h activated about 50% of the latent collagenase, whereas total activation was observed at a concentration of 2 5 0 / x g / m l (Fig. 4). We showed the dependence of activation on the activity of plasma kallikrein by use of the highly specific active-site directed inhibitor Pro-Phe-Arg-CH2C1 [38]. As shown in Fig. 4, Pro-Phe-Arg-CH2C1 (1.19 molar ratio to kallikrein) completely inhibited the

(B)

Fig. 3. Identification of the latent collagenase activator as plasma kallikrein. A. The activator and human plasma kallikrein were each treated with 0.1 mM [3H]diisopropyl fluorophosphate and subjected to SDS-polyacrylamidegel electrophoresis, with and without reduction. After electrophoresis, the gel was impregnated in 2,5-diphenyloxazole,dried and fluorographed. The fluorographs of (a) the latent collagenase activator without reduction and (b) with reduction, and (c) plasma kallikrein without reduction and (d) with reduction showed identical patterns. Molecular weight markers were run in an adjacent lane exactly as in Ref. 24. B. Double immonodiffusion of (a) anti-(human plasma kallikrein) antiserum, against (b) 2/~g latent collagenase activator, (c) 4 ~g human plasma kallikrein and (d) 25/~1 human plasma. The diffusion was carried out for 48 h at 20°C, and the plate was washed in 0.15 M NaCI containing 1% n-butanol overnight at 20°C, dried and stained for protein as described previously.[27].

140

100- ~. . . . . ~o . . . . . . o. . . . . . .

- - ~

4.000

75

3,000

50

2ooo

c = 0o

25

1,000

62.5

125 Plasma

187.5

250

kallikrein (pg/rnl)

Fig. 4. Activation of latent pig synovial collagenase by human plasma kallikrein. Latent pig synovial collagenase (0.06-0.08 units) was treated with plasma kallikrein for 2 h at 37°C at the concentration shown. The reaction was stopped by adding a 20-fold molar excess of soybean trypsin inhibitor and the collagenase activity was assayed in triplicate both with (O . . . . . . ©) and without ( e e ) 4-aminophenylmercuric acetate to determine the proportion of latent enzyme activated. The activation by plasma kallikrein was completely blocked by an affinity label, Pro-Phe-Arg-CH2Cl reacted with kallikrein at the molar ratio of 1.19 (ll II).

activation of latent collagenase. This finding eliminated the possibility that the activation might be due to a small trace of some other proteinase which might have contaminated our plasma kallikrein preparation. The total activatable amount of collagenase remained unchanged even at the highest concentration of kallikrein (250 #g/ml), indicating that kallikrein did not destroy the activated enzyme. Such destruction is seen with trypsin at high concentrations and also with human leukocyte elastase and cathepsin G [29]. Discussion

The nature of latent collagenase is unclear. The activation by a variety of proteinases has suggested that the latent enzyme may be a zymogen requiring activation by limited proteolysis, but this does not explain the activation by mercurial compounds, and the alternative possibility that latent collagenase is an enzyme-inhibitor complex is

coming to be more widely held [10,14,39,40]. It is, of course, perfectly possible that latent collagenases from different sources are chemically different. Proteinases which have previously been shown to activate latent collagenases are trypsin, plasmin [41], pig pancreatic kallikrein [41], cathepsin B [41], cathepsin G [29] and leukocyte elastase [29]. The present report is, as far as we know, the first of activation of collagenase by plasma kallikrein (which is quite a different enzyme from the pancreatic kallikrein: Ref. 37). Activators of collagenase have been detected in tissue cultures of resorbing tadpole tail [42] and alveolar macrophages [43] and in homogenates of rat uterus [18]. For activation in the joint affected by rheumatoid arthritis, it has been suggested that plasmin may be a physiologically significant activator [9]. Also, Kruze et al. [44] have reported the purification from rheumatoid synovial fluid of an activator of crude leukocyte latent collagenase. This substance, with M r 60000, was partially inhibited by diisopropyl fluorophosphate and KCN, but not by soybean trypsin inhibitor. Wize et al. [23] partially purified a factor with M r 35000 from rheumatoid synovial fluid which activated both latent collagenase and gelatinase of polymorphonuclear leukocytes; a similar activator was found in the rheumatoid synovium culture [45]. In our work with a limited number of rheumatoid synovial fluids, we have detected neither plasmin nor the other activators; all of the activation detectable in our system has been attributable to plasma kallikrein. We have not made tests specifically for an activator of leukocyte collagenase, however. Kallikrein is generated from the inactive prokallikrein by the action of activated factor XII (Hageman factor), but then participates in a positive feedback loop by catalyzing the activation of more factor XII [46]. This process is likely to be controlled by the protein inhibitors of kallikrein, including a2-macroglobulin. Prior to our direct detection of plasma kallikrein in synovial fluids, the presence of kallikrein has been inferred from the detection of bradykinin in synovial fluid [47]. Bradykinin is the product of the specific action of plasma kallikrein on high molecular weight kininogen, and is likely to contribute to the

141 i n f l a m m a t i o n b y causing p a i n a n d increasing vasc u l a r p e r m e a b i l i t y [48]. F o r in vitro activation of pig synovial collagenase in a short p e r i o d of i n c u b a t i o n time, relatively high c o n c e n t r a t i o n s of h u m a n p l a s m a kallikrein were required, b u t m o r e recently h u m a n r h e u m a t o i d synovial collagenase was tested a n d f o u n d to be a b e t t e r substrate; a kallikrein conc e n t r a t i o n of 15 / a g / m l a c t i v a t e d a b o u t 50% of h u m a n collagenase (2 U / m l ) in 60 m i n at 37°C (H. N a g a s e a n d C.A. Vater, u n p u b l i s h e d observation). P r o k a l l i k r e i n levels of n o r m a l i n d i v i d u a l s h a v e been r e p o r t e d to be in the range of 9 0 - 1 1 0 / ~ g / m l [49,50]. In i n f l a m e d joints, the level of p l a s m a p r o t e i n s in synovial fluid reach as high as t h a t in p l a s m a [51 ]. However, the c o n c e n t r a t i o n of active kallikrein that we d e t e c t e d in the synovial fluid s a m p l e s [ 0 . 0 1 - 0 . 3 / ~ g / r n l ] was n o t such as to cause r a p i d a c t i v a t i o n of latent coUagenase, a n d a p p r o x i m a t e l y h a l f o f the enzyme was c o m p l e x e d with a 2 - m a c r o g l o b u l i n , as j u d g e d b y its resistance to i n h i b i t i o n b y s o y b e a n trypsin inhibitor. L o c a l c o n c e n t r a t i o n s m a y well be higher t h a n those in the b u l k o f synovial fluid, however. O u r d a t a d o n o t allow us to decide definitely w h e t h e r p l a s m a kallikrein is likely to p l a y a sign i f i c a n t role in the slow, progressive tissue d a m a g e in r h e u m a t o i d arthritis. Nevertheless, the identific a t i o n of this e n z y m e as an activator of latent c o l l a g e n a s e in r h e u m a t o i d synovial fluid m a y be r e l e v a n t to the eventual e x p l a n a t i o n of the disease process.

Acknowledgements W e are grateful to Dr. Y. H o j i m a , N a t i o n a l Lung, H e a r t a n d Blood Institute, Bethesda, M D , U.S.A., for v a l u a b l e discussions a n d the gift of H a g e m a n factor inhibitor; to Dr. E. Philip, Bayer, A . G . , D-5600, W u p p e r t a l 1, F . R . G . , for the gift of Trasylol, and to Dr. C. K e t t n e r , B r o o k h a v e n N a tional L a b o r a t o r y , N e w York, U.S.A., for the p e p t i d y l c h l o r o m e t h a n e i n h i b i t o r of kallikrein. H. N a g a s e is the recipient of a p o s t - d o c t o r a l fellowship from the A r t h r i t i s F o u n d a t i o n (U.S.A.).

References I Harris, E.D., Jr. and Krane, S.M. (1974) New Engl. J. Med. 291,557-563, 605-609, 652-661

2 Harris. E.D., Jr. (1976) Arthritis Rheum. 19, 68-72 " 3 Woolley, D.E., Crossley, M.J. and Evanson, J.M. (1977) Arthritis Rheum. 20, 1231- 1239 4 Lazarus, G.S., Danieis, J.R., Brown, R.S., Bladen, H.A. and Fullmer, H.M. (1968) J. Clin. Invest. 47, 2622-2629 5 Kruze, D. and Wojtecka, E. (1972) Biochim. Biophys. Acta 285,436-446 6 Evanson, J.M., Jeffrey, J.J. and Krane, S.M. (1968) J. Clin. Invest. 47, 2639-2651 7 Bauer, E.A., Eisen, A.Z. and Jeffrey, J.J, (1971) J. Clin. Invest. 50, 2056-2064 8 Dayer, J.-M., Krane, S.M., Russell, R.G.G. and Robinson, D.R. (1976) Proc. Natl. Acad. Sci. U.S.A. 73, 945-949 9 Werb, Z., Mainardi, C.L., Vater, C.A. and Harris, E.D., Jr. (1977) New Engl. J. Med. 296, 1017-1023 10 Vater, C.A., Mainardi, C.L. and Harris, E.D., Jr. (1978) J. Clin. Invest. 62, 987-992 11 Vaes, G. (1972) FEBS Lett. 28, 198-200 12 Hook, R.M., Hook, C.W. and Brown, S.I. (1973) Invest. Ophthalmol. 12, 771-776 13 Bauer, E.A., Stricklin, G.P. and Jeffrey, J.J. (1975) Biochem. Biophys. Res. Commun. 64, 747-754 14 Sellers, A., Cartwright, E., Murphy, G. and Reynolds, J.J. (1977) Biochem. J. 163, 303-307 15 Ehrlich, M.G., Mankin, J., Jones, H., Wright, R., Crispen, C. and Vigliani, G. (1977) J. Clin. Invest. 59, 226-233 16 Birkedal-Hansen, H., Cobb, C.M., Taylor, R.E. and Fullmer, H.M. (1976) Arch. Oral Biol. 21, 21-25 17 Horwitz, A.L. and Crystal, R.G. (1976) Biochem. Biophys. Res. Commun. 69, 296-303 18 Woessner, J.F., Jr. (1977) Biochem. J. 161,535-542 19 Morales, T.I., Woessner, J.F., Jr., Howell, D.S., Marsh, J.M. and LeMaire, W.J. (1978) Biochim. Biophys. Acta 524, 428-434 20 Kitamura, K., Ito, A. and Mori, Y. (1980) J. Biochem. 87, 753-760 21 Oronsky, A.L., Perper, R.J. and Schroder, H.C. (1973) Nature 246, 417-418 22 Sopata, I., Wojtecka-Lukasik, E. and Dancewicz, A.M. (1974) Acta Biochim. Polon. 21,283-289 23 Wize, J., Sopata, I., Wojtecka-Lukasik, E, and Ksi~2ny, S. (1975) Acta Biochim. Polon. 22, 239-250 24 Nagase, H. and Barrett, A.J. (1981) Biochem. J. 193, 187192 25 Hojima, Y., Pierce, J.V. and Pisano, J.J. (.1980) Thromb. Res. 20, 149-t62 26 Lineweaver, H. and Murray, C.W. (1947) J. Biol. Chem. 171,565-581 27 Barrett, A.J. (1974) Biochim. Biophys. Acta 371, 52-62 28 Charney, J. and Tomarelli, R.M. (1947) J. Biol. Chem. 171, 501-505 29 Cawston, T.E., Mercer, E. and Tyler, J.A. (1981) Biochim. Biophys. Acta 657, 73-83 30 Cawston, T.E. and Barrett, A.J. (1979) Anal. Biochem. 99, 340-345 31 Starkey, P.M. and Barrett, A.J. (1976) Biochem. J. 155, 255-263 32 Barrett, A.J., Brown, M.A. and Sayers, C.A. (1979) Biochem. J. 181,401-418

142 33 Laskey, R.A. and Mills, A.D. (1975) Eur. J. Biochem. 56, 335-341 34 Lowry, O.H., Rosebrough, N.J., Farr, A.J. and Randall, R.J. (1951) J. Biol. Chem. 193, 265-275 35 Morton, D.B. (1977) in Proteinases in Mammalian Cells and Tissues (Barrett, A.J., ed.), pp. 445-500, North-Holland Publishing Co., Amsterdam 36 Barrett, A.J. (1980) in Protein Breakdown in Health and Disease (Ciba Foundation Symposium 75) (Evered, D. and Whelan, J., eds.), pp. 1-13, Excerpta Medica, Amsterdam 37 Barrett, A.J. and McDonald, J.K. (1980) Mammalian Proteases. A Glossary and Bibliography. Academic Press, London pp. 56-64 38 Kettner, C. and Shaw, E. (1978) Biochemistry 17, 4778-4784 39 Shinkai, H. and Nagai, Y. (1977) J. Biochem. 81, 1261-1268 40 Sakamoto, S., Sakamoto, M., Matsumoto, A., Goldhaber, P. and Glimcher, M.J. (1978) FEBS Lett. 88, 53-58 41 Eeckhout, Y. and Vaes, G. (1977) Biochem. J. 166, 21-31 42 Harper, E., Block, K.J. and Gross, J. (1971) Biochemistry I0, 3035-3041

43 Horwitz, A.L., Kelman, J.A. and Crystal, R.G. (1976) Nature 264, 772-774 44 Kruze, D., Salgam, P., Cohn, G., Fehr, K. and B6ni, A. (1978) Z. Rheumatol. 37, 355-365 45 Wize, J., Abgarwicz, T., Wojtecka-Lukasik, E., Ksi~ny, S. and Dancewicz, A.M. (1975) Ann. Rheum. Dis. 34, 520-523 46 Heimark, R.L., Kurachi, K., Fujikuwa, K. and Davie, E.W. (1980) Nature 286, 456-460 47 Melmon, K.L., Webster, M.E., Goldfinger, S.E. and Seegmiller, J.E. (1967) Arthritis Rheum. 10, 13-20 48 Elliott, D.F., Horton, E.W. and Lewis, G.P. (1960) J. Physiol. 153, 473-480 49 Bagdasarian, A., Lahiri, B., Talamo, R.C., Wong, P. and Colman, R.W. (1974) J. Clin. Invest. 54, 1444-1454 50 Heber, H., Geiger, R. and Heimburger, H. (1978) HoppeSeyler's Z. Physiol. Chem. 359, 659-669 51 Hamberg, U., Vahtera, E. and Moilanen, L. (1978) Agents Actions 8, 50-56