Collagen ReI. Res. Vol. 611986, pp. 351-364

Characterization of a Specific Antiserum to Rabbit Stromelysin and Demonstration of the Synthesis of Collagenase and Stromelysin by Stimulated Rabbit Articular Chondrocytes

GILLIAN MURPHY, ROSALIND M. HEMBRY and JOHN J. REYNOLDS

Strangeways Research Laboratory, Worts Causeway, Cambridge CBl 4RN.

Abstract

An antiserum to rabbit bone stromelysin (proteoglycanase) was raised in sheep and characterized as specific, recognizing the enzyme from both different tissue sources and different species. This antiserum and a specific antiserum to rabbit bone collagenase were used in the study of metalloproteinase production by rabbit articular chondro­cytes stimulated with either interleukin 1 or mononuclear cell-conditioned medium. It was shown by electroimmunoblotting that the apparently co-ordinate (mole:mole) induction of collagenase and stromelysin activity with time correlated in either case with an increase in enzyme protein. The stimulated production of both enzymes could be modified in parallel by a variety of compounds. Immunohistochemical studies con­firmed that although most cells were producing both metalloproteinases simultaneous­ly, some chondrocytes produced detectable levels of only one. The data are discussed in relation to the mechanisms of breakdown in connective tissues.

Keywords: chondrocyte, collagenase, interleukin 1, stromelysin.

Introduction

The induction of the metalloproteinase, collagenase, by connective tissue cells cul- . tured in the presence of interleukin 1 (lL-1) has been described by many laboratories (Mizel et aI., 1981; Saklatvala et aI., 1984; Trechsel et aI., 1982; Murphy etal., 1985), although the mechanism of action of IL-1-like cytokines is not yet understood. With a number of tissues and stimulated cells in culture it has been observed that a family of metalloproteinases are synthesized and secreted at the same time; these enzymes in­clude collagenase, stromelysin (proteoglycanase) and gelatinase (Sellers et aI., 1978; Cambray et aI., 1981; Murphy et aI., 1981 a; Aggeler et aI., 1984 a). These proteinases are mechanistically related, but their diversity of specificity covers a whole range of connective tissue matrix components (Murphy et aI., 1981 b; Galloway et aI., 1983; Murphy et aI., 1985).

It is of importance to our understanding of the control of matrix turnover to investi­gate in more detail the regulation of the metalloproteinases of connective tissue cells.

24 Collagen 6/4

352 G. Murphy, R. M. Hembry and J. J. Reynolds

Towards this goal we raised an antiserum to collagenase (Hembry et al., 1986) and in this paper we characterize an antiserum to stromelysin. Using these antisera we demon­strate that chondrocytes stimulated by IL-1 (or mononuclear cell-conditioned medium as a source of crude IL-1) produce these metalloproteinases in a co-ordinate fashion.

Methods

Stromelysin Purification. Stromelysin was purified from rabbit bone culture media as described by Galloway et al. (1983).

Antisera. The sheep antiserum to rabbit collagenase has been described (Hembry et al., 1986). Immunoglobulins (lgG) were prepared also as previously described (Hem­bry et al., 1985); 10 mg was labelled with tetra methyl rhodamine isothiocyanate (CL­Rh) by the method of The and Feltkamp (1970).

To prepare the anti$erum to stromelysin, 200 Ilg purified rabbit bone stromelysin was emusified with an equal volume of complete Freund's adjuvant (Difco) and injec­ted into two intramuscular sites of an adult Clun sheep. Three further injections of 100 Ilg stromelysin were given on days 25, 49, and 89, and two injections of 50 Ilg were given on days 882 and 997. Bleeds of 400 ml were taken following all injections: the antisera from all bleeds were found to be essentially similar, but the final bleed, # 16 on day 1011, had the highest titer and was used for the experiments described in this paper.

IgG was prepared from an ammonium sulfate precipitation of whole serum and purified using DEAE-Trisacryl (Corthier et al., 1984). The IgG fraction reacting speci­fically with stromelysin did not bind to the matrix, whereas u2-macroglobulin bound tightly. Purified IgG was coupled to cyanogen bromide activated Sepharose (Pharma­cia) according to the manufacturer's instructions. Anti-stromelysin IgG was labelled with fluorescein isothiocyanate (SL-FITC) by the method of The and Feltkamp (1970). The second antibody for the indirect immunolocalization was a fluorescein-labelled pig Fab to sheep Fab monomer (pig-FITC), prepared as described (Hembry et al., 1985) and diluted in phosphate-buffere.d saline (PBS; 1 :300).

Preparation of Rabbit Chondrocytes and Stimulation with IL-l. Rabbit chondrocy­tes were prepared as described (Trechsel et al., 1982; Hembry et al., 1986). First subcultures were plated either onto glass coverslips or onto 12-well multi well slides (Flow) for immunolocalization studies, and into 15 mm wells for biochemical studies, and grown to confluence in Dulbecco's Modified Eagle's Medium DMEM with 10% fetal calf serum. They were then transferred to DMEM with 0.2% lactalbumin hydro­lysate for 24-48 h in the presence of IL-1 to stimulate collagenase and stromelysin production. IL-1 was obtained from Genzyme Biochemicals, Haverhill, UK; in some experiments pig blood mononuclear cell conditioned medium was used as a source of this activity, prepared as described by Saklatvala et al. (1983). Monensin (5 11M, Sigma) was added to some cultures for the final 3 h of incubation. Other additions were made as described in the Results.

Enzyme Assays. Collagenase and stromelysin activities were routinely assayed in the presence of 0.7 mM 4-aminophenylmercuric acetate (APMA) using [14C]-collagen and e4C]-casein as their respective substrates (Murphy et al., 1981 b). 1 unit of collagenase degrades 1 Ilg of collagen. min.-I 1 unit of stromelysin degrades 1 Ilg casein. min.-I.

354 G. Murphy, R. M. Hembry and ].]. Reynolds

Results

Characterization of Antiserum to Stromelysin

An antiserum raised against purified rabbit bone stromelysin immunoprecipitated a single protein on double diffusion against either crude bone medium or the purified enzyme (Fig. 1). Immunoelectrophoresis gave a similar result (data not shown). Immu­noelectroblotting showed that the antiserum recognized the purified enzyme (Mr 24,500, with traces of Mr 24,000; Galloway et al., 1983) and a protein of Mr 53,000 from crude culture media of various rabbit tissues and cells (Fig. 2 A, B). It has been shown that the latter is the latent form of stromelysin secreted by connective tissue cells, e. g. rabbit synovial fibroblasts (Chin et al., 1985). Activation of the 53,000 form with APMA yields a 44,000 and a 47,000 form, but other lower Mr forms of the enzyme can also be generated on storage or during purification or activation procedu­res, which leads to a complicated picture (see Fig. 2 A, B, C, D, and Chin et al., 1985). It can be shown that most of the immunoreactive forms are active in the degradation of casein, after non-reduced SDS/polyacrylamide gel electrophoresis (Fig. 2, C, D; note that the Mrs are slightly modified by non-reducing conditions, essential to detect activi­ty, and that activity due to gelatinase can be seen at Mrs greater than 70,000; G. Murphy, unpublished work; Murphy et al., 1985).

These activities are all inhibited by 1,10-phenanthroline (data not shown). The variation in lower Mr forms generated by APMA activation of crude culture media from different sources is currently under investigation. Stromelysin purified from bone medium had an Mr of 24,000 and from synovial fibroblast culture medium had an Mr

D

Fig. 1. Ouchterlony double diffusion analysis of sheep anti-rabbit stromelysin antibody. A, antiserum; B, C, E, rabbit bone culture medium, before (B, E) and after (C) APMA activation; D, pure rabbit bone stromelysin.

Metalloproteinase Synthesis by Chondrocytes 353

Stromelysin inhibition studies using purified anti-stromelysin IgG were carried out using [14C]-casein, e4C]-geiatin and proteoglycan beads as substrates (Galloway et aI., 1983). The enzyme was preincubated at 37°C for 1 h with different amounts of IgG prior to the normal assay procedure. Variations upon this incubation time, e. g. 37°C for 2-4 h followed by 4°C for 16 h, had no further effect. Enzyme incubations with either anti-stromelysin IgG bound to Sepharose or unsubstituted Sepharose controls were for 2 h at 25°C followed by 16 h at 4°C. The Sepharose was then removed by centrifugation and the supernatant taken for assays.

Activity gel electrophoresis. Samples of culture media from rabbit chondrocytes and synovial fibroblasts (Aggeler et aI., 1984 a) stimulated with IL-l, rabbit endothelial cells (Herron et aI., 1986) stimulated with phorbol myristate acetate, rabbit bone (Sellers et aI., 1978), were electrophoresed under non-reducing conditions on SDS polyacrylamide (10-12%) gels containing 1 mg/ml casein (Murphy et aI., 1985; Her­ron et aI., 1986).

Immunoblotting. Culture media as above were concentrated by quinine sulphate! SDS precipitation (Herron et aI., 1986) and portions were run on SDS!polyacrylamide gels prior to electrophoretic transfer to nitrocellulose. Detection of antibodies reacting either with anti-collagenase or with anti-stromelysin was carried out using a peroxi­dase labelled second antibody (Hembry et aI., 1986).

Immunoprecipitation. [125Il-Labelled bone culture medium and pure bone stromely­sin (Bolton and Hunter, 1973) were immunoprecipitated with antibodies to stromely­sin, followed by Staphylococcus aureus cell walls as described previously (Hembry et aI., 1985, Murphy et aI. , 1985).

Immunocytochemical Methods . Collagenase and stromelysin were localized in rab­bit chondrocytes stimulated with pig blood mononuclear cell medium essentially as de­scribed for collagenase (Hembry et aI., 1986).

To confirm that the sheep antiserum was specific for stromelysin, absorption experi­ments were carried out as previously described for the absorption of collagenase (Hem­bry et aI., 1986). Stromelysin-Sepharose was prepared by linking purified stromelysin (6.4 units, casein assay) to cyanogen bromide-activated Sepharose 4 B (1 ml; Pharma­cia) using the manufacturer's method. Antibody IgG was then allowed to react with the stromelysin-Sepharose. Stimulated rabbit chondrocytes, grown on multi well slides were incubated with either antibody IgG, antibody IgG absorbed on Sepharose 4 B or antibody IgG absorbed on stromelysin-Sepharose. Intense immunoprecipitation was seen following the first two treatments. No immunoprecipitation was seen around chondrocytes incubated with antibody IgG absorbed on stromelysin-Sepharose, show­ing that the latter had removed all the specific antibody.

To determine whether all chondrocytes were making both collagenase and stromely­sin, cells grown on coverslips were fixed (4% para formaldehyde in PBS, pH 7.4, 5 min), permeabilized (0.1 % Triton X-100 in PBS, 5 min) and incubated for 30 min with a mixture of the sheep antibody to rabbit collagenase labelled with rhodamine (CL-Rh; 1:100 dilution in PBS) and the sheep antibody to rabbit stromelysin-labelled with fluorescein (SL-FITC; 1:30 dilution in PBS). The cells were then washed repeatedly in PBS, mounted in Citifluor (Citifluor Ltd., London) and assessed on a Zeiss Photomi­croscope III using standard FITC and TRITC filters. Photographs were taken on Ko­dak Ektachrome film up rated to 1600 ASA.

A

68-

45-

29-

22-

Metalloproteinase Synthesis by Chondrocytes 355

B

68.., --45-

36-29-

22-

14-

1 2 3 4 5 6 2 3 123456 7 B 123456

Fig. 2. Immunoblotting and activity of different forms of stromelysin from rabbit tissues. A, B, Immunoblotting. Concentrated culture media from the sources listed below were electrophoresed on 10% SOS (reducing) gels and transferred to nitrocellulose as described in the Methods section. A: lane 1, medium from brain endothelial cells; lanes 2 and 3, chondrocyte medium; lanes 4 and 5, synovial fibroblast medium; lanes 6 and 7, bone (calvaria) medium; lane 8, purified active rabbit bone stromelysin. Media were either untreated (lanes 1, 2, 4, 6) or treated (lanes 3, 5, 7) with APMA (0.5 mM, 37 °C, 0.5 h). B: lane 1, purified rabbit stromelysin; lane 2, medium from endothelial cells; lanes 3 and 4, chondrocyte medium; lanes 5 and 6, bone medium. Media were either (untreated lanes 2,3, 5) or treated (lanes 4,6) with APMA (1 mM, 37 °C, 4h). . C, D, Activity gel. Culture media from the sources listed below were electrophoresed on SOS (non-reducing) polyacrylamide gels containing casein, washed and incubated prior to stain­ing with Coomassie Blue, as described in the Methods section. C: lanes 1 and 2, chondrocyte medium; lanes 3 and 4, synovial fibroblast medium; lanes 5 and 6, bone medium. Media were either untreated (lanes 1, 3, 5) or treated (lanes 2, 4, 6) with APMA (0.5 mM 3rC, 0.5 h). 0: lane 1, purified active rabbit bone stromelysin; lanes 2 and 3, medium from chondro­cytes; lanes 4 and 5, synovial fibroblast medium. Media were either untreated (lanes 2, 4) or treated (lanes 3, 5) with APMA (1 mM, 37°C, 1 h). The Mr x 10-3 of standard proteins are indicated to the left of each gel.

of 21,000 (Chin et al., 1985). Immunoprecipitation of [125IJ-labelled crude bone cul­ture medium used as a starting material for the purification, with the anti-stromelysin antiserum showed that the M, 53,000 form predominates at this stage. Traces of Mr 44,000, 35,000 and 24,000 forms can be seen (Fig. 3). Upon treatment with APMA (Galloway et al., 1983) only the 44,000 and 24,000 Mr forms remain.

The anti-stromelysin antiserum did not recognize purified rabbit bone collagenase or gelatinase by immunoblotting or immunoprecipitation. It could detect metallopro­teinase activities degrading casein and proteoglycan found in human fibroblast (phor­bol ester or IL-1 stimulated) culture media and pig synovial culture media. These

4 5

356 G. Murphy, R. M. Hembry and J.J. Reynolds

100-

68-

45-

29-

22-

2 3 4

Fig. 3. [mmunoprecipitation of [125[J-labelled rabbit stromelysin from rabbit bone culture media. Using anti-rabbit stromelysin IgG bound to Sepharose, stromelysin was immunoprecipitated from (1) concentrated [125I]-labelled bone culture medium after APMA treatment; (2) un­treated bone culture medium; (4) ~urified [125I]-labelled active bone stromelysin; lane 3 controls the nonspecific binding of [ 25I]-labelled culture medium to pre-immune IgG bound to Sepharose. The Mr X 10-3 of standard proteins are indicated to the left of the gel.

c: ~ :E

100

:2 50 .5 f!.

o~-----.~~~----------~----------~ 1 10 100 1000

pg anti-stromelysin IgG

Fig. 4. Inhibition of the action of purified stromelysin on different substrates by anti­stromelysin IgG. Purified rabbit bone stromelysin was preincubated with DEAE Trisacryl-purified anti-stro­melysin IgG as described in the Methods, before assay of the activity remaining (expressed as % inhibition) using different substrates; D, 0.12 units with casein as substrate; 0,0.36 units with gelatin as substrate; ., 0.36 units with proteoglycan beads as substrate. Similarly purified non-immune IgG did not inhibit the enzyme activities (data not shown).

Metalloproteinase Synthesis by Chondrocytes 357

A 20

.!!!. Qj 0

"'0 -..... III

.'!:: C :::l

10 Q) III l1S C Q) Ol ~ '0 0

4 8 12 1620 24 48 time (h)

B 20

________________ ...0

4 8 12 16 20 24 48

time (h)

Fig. 5. The co-ordinate stimulation of collagenase and stromelysin production by chondro­cytes in the presence of IL-l. Confluent layers of rabbit articular chondrocytes were cultured for 48 h in DMEM with 0.2% lactalbumin hydrolysate in the absence 0, or presence e, of IL-I. Culture medium was assayed at intervals for the detectable levels of collagenase (A) and stromelysin (B) activities in the presence of APMA.

358 G. Murphy, R. M. Hembry and ].]. Reynolds

A B

68- 68-

4 8 12 16 20 24 48 4 8 12 16 20 24 48

time(h)

Fig. 6. Immunoblotting of culture media from chondrocytes stimulated with IL-l demon­strate collagenase and stromelysin production. Culture media withdrawn from IL-l stimulated chondrocyte cultures from 4-48 h were electrophoresed on SDS polyacrylamide gels, transferred to nitrocellulose and collagenase (A) and stromelysin (B) visualized using their respective antisera and a peroxidase labelled anti-sheep IgG. The Mr x 10-3 of standard proteins are indicated to the left of the gels.

activities had similar Mrs to the rabbit stromelysin and were similarly modified by APMA (data not shown).

Inhibitory Activity of Antiserum

The anti-stromelysin immunoglobulins inhibited the activity of the purified bone enzyme to 50% in the casein assay and up to 85-90% in the assays with either gelatin or proteoglycan beads (Fig. 4). If the antiserum was coupled to agarose, incubated with the enzyme, and complexes removed by centrifugation, 100% inhibition of both gela­tin and proteoglycan-degrading activities could be achieved, and 90% inhibition of the casein-degrading activity. Studies carried out during the purification of bone stromely­sin showed that the three activities were inseparable and were due to the action of a single enzyme (Galloway et al., 1983). We concluded that the antiserum bound the enzyme rather weakly, permitting detectable activity on casein substrate which may also interfere with antibody-antigen binding. The antiserum also inhibited both human and pig stromelysin activities (data not shown).

Strome lysin and Collagenase Synthesis by Chondrocytes

Resting chondrocytes in monolayer culture made negligible amounts of collagenase and stromelysin. After stimulation with mononuclear cell-conditioned medium or par-

Metalloproteinase Synthesis by Chondrocytes 359

Table I. Effectors of IL-1 stimulation of collagenase and stromelysin production by chon­drocytes. Chondrocytes were cultured for 24 h in DMEM, 0.2% lactalbumin hydrolysate in the presence of IL-1 alone, or plus the effectors listed below. Collagenase and stromelysin activities of the culture media were assayed in the presence of APMA and are expressed relative to the activity found in cultures receiving IL-1 alone. These data are assembled from the results of 2-3 experiments where treatments were studied in quadruplicate. Inter-repli­cate and experiment variations were small.

Effector

Retinoic acid, 1O-7M Cycloheximide, 1 !J.g/~l A23187, 2 x 1O-7M Dibutyryl cAMP,S x 1O-7M Trifluoperazine, 10-5 M Colchicine, 1O-6M Cytochalasin B, 0.5 !J.glml Phorbol 12-myristate

13-acetate, 10 ng/ml

% of activity in IL-1-stimulated cell culture medium

Collagenase Stromelysin

o o

44 35 47 83

129 204

o o

40 32 45 80

115 167

tially purified IL-l, enzyme activities could be detected after 6-8 h and activities in­creased linearly for up to 48 h (Fig. 5); the continual presence of the stimulus was required for maximal production of the metalloproteinases (data not shown). The two enzymes appeared to be produced in an approximately mole:mole ratio, as calculated from the specific activity of rabbit bone collagenase of 3,500 units/mg (Hembry et aI., 1986) and of rabbit bone stromelysin of 2,400 units/mg (Galloway et aI., 1983). Both collagenase and stromelysin activities were in a completely latent form, requiring acti­vation with APMA. Immunoblotting of media from various time points after initial stimulation of the chondrocytes was carried out using anti-collagenase and anti-stro­melysin IgG. The blots indicated a comparable rise in the level of enzyme protein over 48 h (Fig. 6). The similarity in Mr of the latent form of stromelysin and of collagenase should be noted, which makes them almost inseparable on SDS gel electrophoresis.

A preliminary study was carried out to define potential activators and inhibitors of the chondrocyte stimulation by IL-l (or mononuclear cell-conditioned medium) in terms of metalloproteinase production (Table I). A range of doses of each effector was analysed but only the most pertinent results are shown. The results clearly demon­strated that collagenase and stromelysin production were modified to the same extent.

Immunolocalization of Collagenase and Stromelysin

Stromelysin secreted by rabbit chondrocytes treated with pig mononuclear cell con­ditioned medium for 24 h was trapped as immunoprecipitates by incubating the cells for 2 h in the presence of the antiserum IgG, exactly as shown for collagenase (Hembry et aI., 1986) . In absorption experiments to confirm again the specificity of the anti­stromelysin antiserum, chondrocytes incubated for 2 h with antibody IgG showed intense immunoprecipitation. In contrast, chondrocytes incubated for 2 h with anti-

360 G. Murphy, R. M. Hembry and ].J. Reynolds

body IgG absorbed with stromelysin-Sepharose 4 B gave no Immunopreclpltation, showing that the stromelysin-Sepharose had removed all the specific antibody. Antibo­dy IgG absorbed on Sepharose 4 B alone showed no diminution of staining.

Monolayers of stimulated rabbit chondrocytes fixed, permeabilized and stained by indirect immunofluorescence for intracellular antigen, showed stromelysin to be pre­sent in most chondrocytes, on a discrete juxtanuclear structure, probably the Golgi

Fig. 7. Dual localization of collagenase and stromelysin in stimulated rabbit chondrocytes. Rabbit chondrocytes stimulated for 24 h with pig blood mononuclear cell-conditioned medium and monensin for the final 3 h of the incubation period were fixed, permeabilized and stained with CL-Rh and SL-FITC. Bars = 10 f!M. (A) Cells viewed through FlTC filters. All have bright fluorescence of a juxtanuclear struc­ture, probably the Golgi apparatus, indicating the presence of stromelysin. (B) Cells viewed through TRITe filters. Most cells have juxtanuclear rhodamine fluores­cence, indicating the presence of collagenase. The two central cells (arrowed) have no collagenase fluorescence: they are secreting stromelysin only.

Metalloproteinase Synthesis by Chondrocytes 361

Table II. Percentage of stimulated rabbit chondrocytes showing immunofluorescence of collagenase and stromelysin. Rabbit chondrocytes treated with pig mononuclear cell conditioned medium for 24 hand monensin for the final 3 h of the incubation period were fixed, permeabilized and stained with CL-Rh and SL-FITC in a single incubation as described in the Methods section. Visible perinuclear fluorescence using their FITC or TRITC filters was scored positive.

Total Number Percentage of cells staining for cells counted Collagenase Collagenase Stromelysin No

and Stromelysin only only Stain

Expt. l. 448 74 3.6 16.7 5.6 Expt. 2. 562 71.3 lOA 8.5 9.6

apparatus, as has been described in detail for collagenase (Hembry et aI., 1986). To determine whether all chondrocytes were making both enzymes simultaneously, dual label immunofluroescent localization was carried out and the results shown in Fig. 7 and Table II. Monensin-treatment of monolayers prior to staining augmented the intracellular immunofluorescence in cells secreting very low levels of enyzmes. The large majority (74%) of cells secreted both enzymes, while a small number (5%) made neither.

Discussion

The development of a specific antiserum to the metalloproteinase stromelysin (for­merly called proteoglycanase), which recognizes the enzyme from a number of different species, will be a useful tool for the study of its localization and production within connective tissue. The antibody has enabled us to demonstrate that the initial high Mr latent form of the enzyme is rapidly broken down to a number of smaller forms in vitro, particularly after treatment with APMA. The precise importance of these obser­vations to events in vivo cannot be ascertained at present, although a physiological activation mechanism involving a fall in stromelysin Mr would appear to be essential for its function. In this paper the anti-stromelysin antiserum and an antiserum to the related metalloproteinase, collagenase, have been used to demonstrate the parallel synthesis and secretion of the enzymes in a model system consisting of monolayers of rabbit articular chondrocytes stimulated with IL-l. By the use of monensin to inhibit the final secretion of the enzymes from stimulated chondrocytes, it was posible to increase the sensitivity of the antisera in immunohistochemical studies and actually pin-point the cells producing collagenase and stromelysin. We found that cells could, and usually did, synthesize both metalloproteinases, but individual cases of detectable levels of only either collagenase or stromelysin occurred at apparently maximal stimu­latory concentration of IL-l. It was concluded that, although closely co-ordinated, the induction of the two metalloproteinases are separable.

Both activity assays and immunoblot analyses showed that collagenase and strome­lysin are detectable in the culture medium 4-6 h after stimulation of the cells. This is similar to the time period for the response of rabbit synovial fibroblasts to phorbol ester stimulation, which was postulated to be the period required for mRNA synthesis (Aggeler et aI., 1984 a). Our data confirm that new protein synthesis is responsible for

362 G. Murphy, R. M. Hembry and J. J. Reynolds

the increases in enzyme activities detectable. Previous work from this laboratory show­ed that the effect of mononuclear cell-conditioned medium (lL-l) on chondrocytes was to inhibit production of collagen, as collagenase synthesis increased (Trechsel et aI., 1982), again similar to the response of activated synovial fibroblasts (Aggeler et aI., 1984 b). Su,ch in vitro manipulations of gene expression by connective tissue cells, causing a switch from production of matrix components to degradative enzyme pro­duction, may be indicative of in vivo mechanisms functioning during destructive and remodelling processes. It is interesting that the two metalloproteinases collagenases and stromelysin (as well as a third activity, gelatinasej unpublished observations) are synthesized and secreted in parallel, not only in response to IL-l, but also to the action of inhibitors and enhancers of IL-l action.

We conducted preliminary investigations into the possible mode of action of IL-1 on chondrocytes, leading to an increase in metalloproteinase expression. Partial inhibition of metalloproteinase production by dibutyryl cAMP suggests that cAMP is unlikely to be directly involved in the cytokine stimulation of chondrocytes. Similar results were obtained with isobutylmethyl xanthine and forskolin (data not shown). The ionophore A2318 7, which raises Ca ++ levels in a receptor-independent fashion, was inhibitory to the effects of II-Ion chondrocytes in terms of metallproteinase production, which may be due to the inhibition of secretion. In contrast to rabbit synovial fibroblasts (Aggeler et aI., 1984), chondrocytes were not stimulated to produce metalloproteinases by either colchicine, cytochalasin B, phorbol 12-myristate 13-acetate or trifluoperazine alone. As observed by Arend et al. (1985), phorbol myristate acetate potentiates the stimula­tion of chondrocytes by mononuclear cell conditional medium but the reason for this is unknown.

The further use of the antisera to the metalloproteinases and TIMP (Hembry et aI., 1985) to analyse their localization in tissues in both specific diseases and during remo­delling and developmental situations should give a good indication of where these enzymes can act and the type of cells producing them. Further studies of the regulatory mechanisms governing the production of metalloproteinases by different cell types are now possible by analysis of the transcription and translation of their respective mRNAs. Such information on the relationship between collagenase and stromelysin and their specific inhibitor TIMP will be important in the elucidation of the mecha­nisms governing tissue resorption (Murphy and Reynolds, 1985).

Acknowledgements

We thank the Medical Research Council for financial support and the Institute of Animal Physiology, Babraham for provision of sheep. We thank Ms. Mary Harrison and Ms. Meenakshi Bagga for expert technical assistance.

References

Aggeler, ] ., Frisch, S. M. and Werb, Z.: Collagenase is a major gene product of induced rabbit synovial fibroblasts. J. Cell Bioi. 98: 1656-1661,.1,984 a.

Aggeler, J., Frisch, S.' M. and Werb, Z.: Changes in cell sh'ape correlate with collagenase gene expression in -rabbit synovial fibroblasts. J. Cell Bioi. 98: 1662-1671, 1984 b.

Arend, W. P., Joslin, F. G. and Massoni, R. J.: Effects of immune complexes on production

Metalloproteinase Synthesis by Chondrocytes 363

by human monocytes of interleukin 1 or an interleukin 1 inhibitor. J. Immunol. 134: 3868-3875, 1985.

Bolton, A. E. and Hunter, W. M.: The labelling of proteins to high specific radioactivities by conjugation to a 125I-containing acylating agent. Application to the radioimmunoassay. Biochem. J. 133: 529-539, 1973.

Cambray, G. ]., Murphy, G., Page-Thomas, D. P. and Reynolds, J. ].: The production in culture of metalloproteinases and an inhibitor by joint tissues from normal rabbits, and from rabbits with a model arthritis. I. Synovium. Rheumatol. Int. 1: 11-16, 1981.

Chin,]. R., Murphy, G. and Werb, Z.: Stromelysin, a connective tissue-degrading metallo­endopeptidase secreted by stimulated rabbit synovial fibroblasts in parallel with collage­nase. J. Bioi. Chem. 260: 12367-12376, 1985.

Galloway, W. A., Murphy, G., Sandy, J. D., Gavrilovic, ]., Cawston, T. A. and Reynolds,] . ].: Purification and characterization of a rabbit bone metalloproteinase that degrades proteoglycan and other connective-tissue components. Biochem. J. 209: 741-752, 1983.

Hembry, R. M., Murphy, G. and Reynolds, J. J.: Immunolocalization of tissue inhibitor of metalloproteinases (TIMP) in human cells. Characterization and use of a specific antiser­um. J. Cell Sci. 73: 105-119, 1985.

Hembry, R. M., Murphy, G. and Reynolds,]. ].: Characterisation of a specific antiserum for mammalian collagenase from several species. J. Cell Sci. 81, 105-123, 1986.

Herron, G. S., Werb, Z., Dwyer, K. and Banda, M. ].: Secretion of metalloproteinases by stimulated capillary endothelial cells. J. Bioi. Chem. 261, 2810-2813, 1986.

Mizel, S. B., Dayer, ].-M., Krane, S. M. and Mergenhagen, S. E. : Stimulation of rheumatoid synovial cell collagenase and prostaglandin production by partially purified lymphocyte­activating factor (interleukin 1). Proc. Nat!' Acad. Sci. USA 78: 2474-2477, 1981.

Murphy, G., Cambray, G. ]., Virani, N., Page-Thomas, D. P. and Reynolds, ]. ]. : The production in culture of metalloproteinases and an inhibitor by joint tissues from normal rabbits, and from rabbits with a model arthritis. II. Articular cartilage. Rheumatol. Int. 1: 17-20, 1981 a.

Murphy, G., Cawston, T. E., Galloway, W. A., Barnes, M. ]., Bunning, R. A. D., Mercer, E., Reynolds,].]. and Burgeson, R. E.: Metalloproteinases from rabbit bone culture medium degrade types IV and V collagens, laminin and fibronectin. Biochem. J. 199: 807-811, 1981 b.

Murphy, G., McAlpine, C. G., Poll, C. T. and Reynolds, J. ]. : Purification and characterisa­tion of a bone metalloproteinase that degrades gelatin and types IV and V collagen. Biochim. Biophys. Acta 831: 49-58, 1985.

Murphy, G. and Reynolds, ]. J. : Current views of collagen degradation. Progress towards understanding the resorption of connective tissues. BioEssays 2: 55-60, 1985.

Saklatvala, ]., Curry, V. A. and Sarsfield, S.].: Purification to homogeneity of pig leucocyte catabolin, a protein that causes cartilage resorption in vitro. Biochem. J. 215: 385-392, 1983.

Saklatvala, ]., Pilsworth, L. M. c., Sarsfield, S. ] ., Gavrilovic, ]. and Heath, ]. K.: Pig catabolin is a form of interleukin 1. Cartilage and bone resorb, fibroblasts make prosta­glandin and collagenase, and thymocyte proliferation is augmented in response to one protein. Biochem. J. 224: 461.-466, 1984.

Sellers, A., Reynolds, ].]. and Meikle, M. c.: Neutral metallo-proteinases of rabbit bone. Separation in laterlt forms of distinct enzymes that when activated degrade collagen, gelatin and proteoglycans. Biochem. J. 171: 493-496, 1978.

Trechsel, U., Dew, G., Murphy, G. and Reynolds, J. J.: Effects of products from macropha­ges, blood mononuclear cells and of retinol on collagenase secretion and collagen synthe­sis in chondrocyte culture. Biochim. Biophys. Acta 720: 364-370, 1982.

Dr. Gillian Murphy, Strangeways Research Laboratory, Worts Causeway, Cambridge CBI4RN.