J Clin Pathol:Mol Pathol 1996;49:M331-M339 M331
Papers Clin Mol Pathol: first published as 10.1136/mp.49.6.M331 on 1 December 1996. Downloaded from
Inhibitors of collagenase but not of gelatinase reduce cartilage explant proteoglycan breakdown despite only low levels of matrix metalloproteinase activity
Carol J Brown, Shamim Rahman, Allison C Morton, Claire L Beauchamp, Helena Bramwell, David J Buttle
Aims-To investigate the level of matrix Aggregating proteoglycan (aggrecan) and type II metalloproteinase activity during the collagen are the major extracellular protein com- time-course of cartilage explant prote- ponents of cartilage. Loss ofproteoglycan and its oglycan breakdown; to determine the associated water can compromise the structure effects of selective small-molecule inhibi- and function ofcartilage by reducing its resilience tors of matrix metalloproteinases on pro- to compressive loading and increasing friction teoglycan degradation. during joint motion.' 2Thus, depletion of articu- Methods-The levels of matrix metallo- lar cartilage proteoglycan (seen on x ray film as proteinase activity in cartilage explant joint space narrowing) may predispose the carti- cultures and conditioned media were lage to further mechanical damage. Preventing monitored by use of a quenched fluores- pathological cartilage proteoglycan breakdown cent substrate. The constants for inhibi- may therefore substantially protect the affected tion of certain matrix metalloproteinases joint and help prevent the development of an by a series of synthetic inhibitors were arthritic lesion. The mechanism by which carti- determined. Bovine and human cartilage lage proteoglycan is degraded is not fully under- explant cultures were treated with stood, although recent evidence would suggest http://mp.bmj.com/ interleukin-1, tumour necrosis factor or the involvement of a number of proteolytic retinoic acid and the amount ofproteogly- enzymes in a cascade reaction that may result in can released into the culture medium in the activation of an as yet undiscovered enzyme, the absence and presence ofthe inhibitors "aggrecanase" (reviewed elsewhere3). was quantified. Control experiments, ex- There is evidence that members of the
amining the inhibition of other protein- matrix metalloproteinase (MMP), or matrixin, on September 29, 2021 by guest. Protected copyright. ases, and investigating possible toxic or family of enzymes are involved in cartilage pro- Institute for Bone and were breakdown. The MMPs, like almost Joint Medicine, non-specific effects of the inhibitors, teoglycan Department of Human carried out. all metalloproteinases that have been charac- Metabolism and Results-The profile of inhibition of pro- terised so far, utilise a zinc atom in the active Clinical Biochemistry, teoglycan release suggested the involve- site.4 Based on their substrate specificity the University of Sheffield ment of interstitial collagenase-like, MMPs can be divided into three groups: the Medical School, interstitial colla- Beech Hill Road, rather than gelatinase- or possibly collagenase group, including Sheffield SlO 2RX stromelysin-like, proteinases. No evidence genase (EC 3.4.24.7); neutrophil collagenase C J Brown was found for toxic or non-specific (EC 3.4.24.34) and collagenase-3'; the strome- S Rahman mechanisms of inhibition. Very low levels lysin group, which includes stromelysin 1 (EC A C Morton metallo- 3.4.24.17), stromelysin 2 (EC 3.4.24.22) and H Bramwell of activity of the known matrix D J Buttle proteinases were present during the time- matrilysin (EC 3.4.24.23); and the gelatinase course of aggrecan breakdown. group consisting of gelatinases A (EC British Biotech Conclusions-A novel collagenase-like 3.4.24.24) and B (EC 3.4.24.35).6 Stromelysin Pharmaceuticals Ltd., proteinase(s) may be involved in cartilage 3 actually represents a distinct group of the Watlington Road, MMPs.4 In in vitro experiments matrilysin Cowley, proteoglycan breakdown. Gelatinase-type Oxford OX4 5LY matrix metalloproteinases do not seem to and stromelysin 1 have been shown to be the C L Beauchamp be involved in this process. Specific colla- most efficient members of this subfamily at genase inhibitors may be therapeutically catalysing aggrecan hydrolysis.8 Correspondence to: Increased MMP concentrations have been Dr D J Buttle. efficacious in the treatment of arthritis. email: ( Clin Pathol: Mol Pathol 1996;49:M33 1-M339) found in both osteoarthritis' and rheumatoid [email protected] arthritis."0 Amounts of the MMPs are raised by Accepted for publication Keywords: cartilage, proteoglycan breakdown, pro- the action of the proinflammatory cytokines, 12 September 1996 teases. interleukin-1 (IL-1) and tumour necrosis M332 Brown, Rahman, Morton, Beauchamp, Bramwell, Buttle
factor (TNF), that mediate catabolic responses Boulder, Colorado, USA. All-trans-retinoic Clin Mol Pathol: first published as 10.1136/mp.49.6.M331 on 1 December 1996. Downloaded from in cartilage."'l4 acid (Ret), gentamicin, cortisol, chondroitin The ability of synthetic inhibitors of the sulfate A, aminophenylmercuric acetate MMPs to suppress cartilage proteoglycan deg- (APMA) and dimethyl sulfoxide were pur- radation in model culture systems is well chased from Sigma Chemical Co., Poole, Dor- documented,'51" and it has been proposed that set, UK. Dulbecco's modification of Eagle's inhibitors of these enzymes might be used for medium (DMEM), minimal essential medium therapeutic intervention in the arthritides." To (MEM), newborn-calf serum (NCS), fetal calf date, all of the inhibitors found to have a chon- serum (FCS), amphotericin B, penicillin, and droprotective effect are broad spectrum inhibi- streptomycin were purchased from Gibco Life tors that do not discriminate between the vari- Technologies Ltd, Paisley, UK. 1,9- ous MMPs. For instance, the most potent Dimethylmethylene blue was from Aldrich, example to date, the hydroxamate BB94, is a Gillingham, Dorset, UK. Cetylpyridinium tight binding inhibitor of all the MMPs investi- chloride, Triton X-100 and OCT embedding gated, with Kj(app) and IC50 values in the low medium were from BDH Chemicals Ltd, nanomolar range and varying by only about a Poole, Dorset, UK. 7-Methoxycoumarin-4- factor of 10 between the different MMPs.'520 ylacetyl (Mca)-Pro-Leu-Gly-Leu-(3-[2,4-dini- Such inhibitors therefore provide us with little trophenyl] - L - 2, 3 - diaminopropionyl) information about which members of this large (Dpa)-Ala-Arg-NH2 was provided by Dr C group of enzymes may be involved in cartilage Graham Knight, Strangeways Research Labo- proteoglycan breakdown. ratory, Cambridge, UK. Human urokinase- We have now attempted to address the role type plasminogen activator (uPa) (EC of MMPs in cartilage proteoglycan breakdown 3.4.21.73) was obtained from Calbiochem- by the use of a series of zinc chelating MMP Novabiochem Ltd, Nottingham, UK. Benzy- inhibitors that demonstrate some discrimina- loxycarbonyl (Z)-Gly-Gly-Arg-aminomethyl- tion for the inhibition of various MMPs, and coumarylamide (NHMec) and Z-Arg-Arg- also by the measurement of levels of MMP NHMec were the products of Bachem activity present in cartilage explant cultures Feinchemikalien AG, Bubendorf, Switzerland. during the time-course of aggrecan depletion. 35SO4 and 35S-methionine were products of Amersham International plc, Little Chalfont, Methods Buckinghamshire, UK. All buffers and salts Human interstitial procollagenase, promatri- were of analytical grade. lysin and prostromelysin 1 were recombinant proteins purified from media conditioned by PROTEINASE ASSAYS CHO cells transfected with cDNA coding for The effects ofone ofthe MMP inhibitors, BB250 the respective enzymes. Progelatinases A and B (10 FM), on the activities ofuPa, cathepsin B and were isolated from media harvested from the cathepsin D, which have been implicated in
malignant melanoma cell line RPMI 7591, and cartilage proteoglycan breakdown,3 25 were http://mp.bmj.com/ the monocytic leukaemia cell line THP-1, tested. In addition, we examined the capacity of respectively. The MMP inhibitors used in this BB250 to inhibit two metallopeptidases that are study, BB87, BB250, BB2001, BB2014 and not MMPs- that is, peptidyl dipeptidase A and BB3003*, were synthesised at British Biotech neprilysin. Pharmaceuticals Ltd., Oxford, UK. Human uPa activity was measured by continuous cathepsin B (EC 3.4.22.1) was purified from rate assays in 50 mM Tris HCl, 0.1% Triton
liver by affinity chromatography21 and human X-100, pH 7.5 at 37°C, with Z-Gly-Gly-Arg- on September 29, 2021 by guest. Protected copyright. cathepsin D (EC 3.4.23.5) from liver was pre- NHMec as substrate (100 jM). The rate of pared by the published method.22 Peptidyl substrate hydrolysis was recorded before and dipeptidase A (angiotensin converting enzyme) after the addition of the inhibitor to the (EC 3.4.15.1) was purified from rabbit lung.23 cuvette. The effect of BB250 on human cathe- Neprilysin (enkephalinase) (EC 3.4.24.11) was psin B activity was determined with Z-Arg- from rabbit kidney, and was obtained in the Arg-NHMec (5 jM) as substrate in 100 mM membrane bound form essentially as Na2H/KH2PO4, pH 6.0, containing 1 mM described.24 Recombinant human IL-i ox (rhIL- Na2EDTA and 2 mM dithiothreitol. The lc) (specific activity 5 x 107 U/mg) was a gift activity of cathepsin D in the presence and from Dr K Ray, Glaxo Group Research Ltd, absence of BB250 was measured at 45°C using Greenford Road, Greenford, Middlesex, UK. 8%(w/v) haemoglobin as substrate in 0.1 M Recombinant human TNFo (rhTNFcx) (spe- formate buffer, pH 3.3. Following precipitation cific activity 8 x 107 U/mg) was kindly provided of undigested protein in 3% (w/v) trichloroace- by Dr E Amento, Genentech Inc. Boulder, tic acid and filtration the change in A280, as Colorado, USA.' compared with a blank without enzyme, was The recombinant human IL-i-receptor an- determined. Peptidyl dipeptidase A activity was tagonist protein was a gift from Synergen Inc., assayed using N-benzoylglycine-His-Leu, as *The structures of the MMP inhibitors are: BB87, N4-hydroxy- described previously.23 Neprilysin was assayed 2R-isobutyl-N1- {2-phenyl-1S- [(pyrid-2-ylmethyl) -carbamoyl]- using the quenched fluorescent substrate ethyl}-succinamide; BB250, N4-hydroxy - 2R - isobutyl - N1 - (1S - methyl -carbamoyl - 2 - phenyl-ethyl)-3S-thiophene-2- sulfonyl- dansyl-D-Ala-Gly-Phe(pNO2)-Gly."2 methyl) succinamide; BB2001, 3R{1S(3 morpholin-4yl- MMP activity in bovine nasal cartilage explant propylcarbamoyl)-2-phenyl-ethyl-carbamoyl} - 6 - phenyl-hex- cultures and conditioned media was assayed by anoic acid; BB2014, 3R-(lS-methylcarbamoyl-2-phenyl - use of the quenched fluorescent substrate ethylcarbamoyl) - 6 - phenylhexanoic acid; BB3003, 3R-(lS- methylcarbamoyl-2-phenyl-ethylcarbamoyl)-nonadecanoic acid, Mca-Pro-Leu-Gly-Leu-Dpa-Ala-Arg-NH2. This dicyclohexylamine salt. substrate is cleaved efficiently by all ofthe MMPs Inhibitors ofcollagenase but not ofgelatinase reduce cartilage explant proteoglycan breakdown M333
collagenase, tested so far, including interstitial shown previously to behave catabolically in a Clin Mol Pathol: first published as 10.1136/mp.49.6.M331 on 1 December 1996. Downloaded from stromelysin 1, matrilysin, gelatinase A,26 gelati- manner similar to the human tissue.3 However, nase B,'5 neutrophil collagenase, and a limited number of experiments were done collagenase-3,27 and can therefore be used to using two samples of non-arthritic human car- assess the combined levels of activity of these tilage. One sample was from the femoral enzymes. The MMP activity in conditioned condyle of an 81 year old man with no history media from explants cultured with or without of arthritic disease. The second sample was rhIL-lx (0.3 nM), rhTNFa (3.5 nM) or Ret from the interphalangeal joint of the left big toe (1 p1M) for up to five days, with and without of a 14 year old boy who underwent fusion of APMA (0.1 rmM; one hour) activation of proM- the joint in response to a non-arthritic foot MPs, was assessed by continuous-rate assays at deformity. Cartilage was sliced, to give an 37°C. The buffer used was 0.1 M Tris HCl, pH explant size of approximately 3 x 3 x 2 mm. 7.5, containing 0.1 M NaCl, 10 mM CaCl2, The slices were pre-incubated in MEM con- 0.2% Triton X-100, and Mca-Pro-Leu-Gly-Leu- taining 10% (v/v) FCS, penicillin (100 u/ml) Dpa-Ala-Arg-NH2 (5 giM) (ex. 328 nm, em. 393 and streptomycin (100 jg/ml) for two days. nm). The fluorimeter was standardised with a They were then transferred individually to known concentration of Mca-Pro-Leu. wells of a 96-well plate containing 195 jl MEM In order to obviate potential problems of containing 1% (v/v) FCS to which the fluorescence quenching, the substrate (10 gM) inhibitors were added from 100-fold concen- was added to explant cultures (two explants of trated stock solutions in methanol. After one approximate dimensions 2 x 2 x 2 mm/culture) hour Ret was added to a final concentration of in 0.4 ml phosphate buffered isotonic saline 1 gM. The cultures were maintained for five (PBS), pH 7.2, containing 4.5 mg/ml D-glu- days after which the media and explants were cose and antibiotics (see later), for periods of stored frozen pending analysis. up to 24 hours. We have previously found that Bovine nasal septum cartilage explants were cartilage explants can be successfully cultured prepared and preincubated as described for at least two days under these conditions, previously.32 The cartilage discs were trans- during which time they behave in a similar ferred individually to wells of a 96-well plate, manner to explants cultured in DMEM, and with 195 gl of serum-free DMEM to which the resorb the proteoglycan component of the metalloproteinase inhibitors were added from extracellular matrix upon treatment with cata- 100-fold stock solutions in methanol. A 1% bolic stimuli. BB94'5 (10 jM) was also added (v/v) concentration of methanol did not affect to some cultures; 0.3 ml of each culture was cartilage catabolism (data not shown). After diluted with an equal part of Tris buffer one hour, rhIL-la, rhTNFx, or Ret was added (earlier), and the quantity of product formed to final concentrations of 250 U (0.3 nM), was determined by comparing the amount of 5000 U (3.5 nM), or 10-6 M, respectively. The fluorescence with that of a known concentra- cultures were maintained for up to five days
tion of standard. with a medium change after two days, the fresh http://mp.bmj.com/ medium containing exactly the same constitu- DETERMINATION OF KINETIC PARAMETERS ents as the original. Conditioned media and Gelatinase A was activated overnight at 40C in discs were stored frozen pending analysis. In lmM APMA. Gelatinase B was activated by studies examining the effects of the MMP the same means at ambient temperature for inhibitors on stimulated proteoglycan release two hours. Interstitial collagenase was activated from all types of cartilage examined, only
with 0.66 mM APMA for 16 hours at 37°C, experiments in which the addition of the on September 29, 2021 by guest. Protected copyright. prior to purification. Purified stromelysin 1 was stimulus resulted in at least a doubling of the activated with trypsin (30 minutes, 370C), and amount of proteoglycan release were included matrilysin was autoactivated by freeze-thawing in the analyses. This helped to ensure that the prior to purification. detected inhibition was of stimulated, as well as For the determination of IC50 values, all basal, release. assays were performed at 370C for 16 hours in Bovine articular cartilage slices were dis- 25 mM Hepes, pH 7.5, containing 5 mM sected from metacarpophalangeal joints and CaC12, 0.05% (w/v) Brij 35 and 0.02% NaN3. preincubated as reported previously,'5 after Collagenase activity was determined in an which the explants were treated as described assay28 utilising '4C-collagen as substrate.29 earlier for the nasal cartilage discs. As shown Matrilysin and stromelysin 1 activities were elsewhere'5 articular cartilage was less respon- measured in assays using '4C-casein as the sive than nasal cartilage to the presence of substrate.30 Gelatinase A and B activities were human cytokines. The concentrations of assayed by a modification of the previously rhIL- 1 ex and rhTNFax were therefore increased described procedure,3' with the addition of 50 to 1800 U (2 nM) and 10 000 U (7 nM), gg/ml gelatin. IC50 values were found by com- respectively. The cultures were maintained for paring enzyme activity in the presence of vary- five days with a medium change on day 2, after ing concentrations of the inhibitors, and fitting which the media and cartilage pieces were a curve to the data thus obtained. The inhibitor stored frozen. concentration that gave 50% inhibition was then acquired by interpolation. DETERMINATION OF PROTEOGLYCAN DEGRADATION CARTILAGE CULTURES The amount of proteoglycan in the culture Most of the work reported here has been done medium was determined as sulfated gly- on normal bovine cartilage, which has been cosaminoglycan by use of dimethylmethylene M334 Brown, Rahman, Morton, Beauchamp, Bramwell, Buttle
1 Table Constants for the inhibition ofsome members of the MMP subfamily by a series inhibitor of interstitial collagenase, gelatinases Clin Mol Pathol: first published as 10.1136/mp.49.6.M331 on 1 December 1996. Downloaded from ofMMP inhibitors A and B, but inhibits stromelysin 1 relatively poorly. BB2001, BB2014 and BB3003 are IC50 values (nM) selective inhibitors of gelatinase A. We also BB250 BB87 BB2001 BB2014 BB3003 examined the inhibition of two non-matrix metallopeptidases, peptidyl dipeptidase A and Interstitial 5 30 12 000 15 000 >100 000 collagenase neprilysin, by the MMP inhibitors. In all cases, Stromelysin 1 20 600 7000 1000 >100 000 IC50 values for the inhibition of these two Gelatinase A 20 30 100 30 30 Gelatinase B 2 10 2000 150 3000 enzymes were greater than 10 000 nM (results Matrilysin 10 80 12 000 >100 000 >100 000 not shown). We could detect no inhibition of human cathepsin B, human cathepsin D or human uPa by BB250 (10 gM) (results not blue by adaptation of the method33 for use with shown). a plate reader.34 The binding of sulfated glycosaminoglycan to the dye was accompa- nied by a metachromatic shift, measured as a STIMULATION OF PROTEOGLYCAN BREAKDOWN decrease in absorbance at 570 nm. Determina- FROM BOVINE CARTILAGE tion of residual sulfated glycosaminoglycan in Although there was some variability between the cartilage explants was by the same method the behaviour of bovine nasal cartilage from following digestion with papain.35 different animals, greater than 50% release of total cartilage proteoglycans was usually HISTOLOGY achieved by two days of culture with 0.3 nM Cartilage explants cultured with or without a rhIL- 1 a. The same level was often not reached catabolic stimulant or MMP inhibitor were until five days in the presence of 1 jM Ret. embedded in OCT, snap frozen in liquid N2 rhTNFax elicited a catabolic response in bovine cooled isopentane and stored at -700C until nasal cartilage, leading to a notable increase immediately before use. Sections (7 im) were in release of proteoglycan, as reported cut on a cryostat, placed on slides and left to previously.37 The tissue was less responsive to dry in air for 30 minutes. The sections were rhTNFox than to rhIL-la, however, such that then fixed in 4% paraformaldehyde in phos- more than a 10-fold higher concentration (3.5 phate buffer, pH 7.2, for 15 minutes and nM instead of 0.3 nM) was required to cause, stained with 1% (w/v) toluidine blue in 3% on average, a 50% release of proteoglycan by (v/v) acetic acid, pH 2.5, for 15 seconds, day 5 (data not shown). The IL-1-receptor destained in 3% (v/v) acetic acid for 30 seconds antagonist protein, when present in a 100-fold and mounted using glycerine. molar excess, completely inhibited rhIL- 1 a stimulated proteoglycan breakdown from nasal LACTATE DETERMINATION The concentration of lactate in explants. TNF stimulated release, however, the medium was not inhibited by this concentration of the was determined a with kit http://mp.bmj.com/ supplied by Sigma receptor antagonist, demonstrating that TNF Chemical Co. which uses the lactate oxidase/ peroxidase method. stimulated proteoglycan release did not occur as a secondary response to TNF stimulation of IL-I synthesis by the chondrocytes (results not PROTEIN SYNTHESIS shown). Bovine nasal cartilage explants were cultured Similar results were found using bovine with or without rhIL-lca and MMP inhibitor metacarpophalangeal cartilage. In terms of for two days. The explants were washed in stimulation of proteoglycan breakdown this on September 29, 2021 by guest. Protected copyright. DMEM and given a two hour pulse with tissue was generally less responsive to the 35S-methionine and then treated as described cytokines, such that the concentrations of previously.32 rhIL- 1 a and rhTNFa had to be increased (see the Methods section) in order to achieve a 50% PROTEOGLYCAN SYNTHSIS release of total proteoglycan by day 5. In The amount of incorporation of 31SO4 into contrast, the response to Ret was not very dif- cetylpyridinium chloride precipitates was as- ferent between the two bovine cartilages. sessed by the previously described method36 with the modification that the explants were cultured in DMEM rather than S04-free EFFECTS OF THE MMP INHIBITORS ON BASAL AND Iscove's MEM. STIMULATED PROTEOGLYCAN RELEASE FROM CARTILAGE EXPLANTS We first examined the effect of a fixed concen- STATISTICAL ANALYSES tration (10 gM) of the MMP inhibitors on The results were analysed using the two-tailed basal and stimulated proteoglycan release from Mann-Whitney U test for non-parametric data. bovine nasal septum cartilage explants. Only two of the five inhibitors tested produced Results statistically significant inhibition of stimulated CONSTANTS FOR MMP INHIBITION release, BB250 and BB87 (table 2). These IC50 values for the inhibition of the MMPs are compounds, both of which are tight binding given in table 1. These results demonstrate that inhibitors of interstitial collagenase (table 1), BB250 (like BB94, already shown to be a were potent inhibitors of proteoglycan release potent inhibitor of cartilage proteoglycan stimulated by rhIL-lat, rhTNFax and Ret. All of breakdown'5) is a tight binding inhibitor of all the inhibitors produced some inhibition of the MMPs tested. BB87 is a tight binding basal (unstimulated) proteoglycan breakdown Inhibitors of collagenase but not ofgelatinase reduce cartilage explant proteoglycan breakdown M335
Table 2 Percentage inhibition ofproteoglycan release from bovine cartilage explants and BB250 on human articular cartilage Clin Mol Pathol: first published as 10.1136/mp.49.6.M331 on 1 December 1996. Downloaded from proteoglycan breakdown. As described in the Stimulant Methods section, we obtained two samples of Inhibitor (101M) Basal rhIL-lax rhTNFax Ret non-arthritic cartilage. Both of these re- sponded catabolically to the presence of 1 ,uM Nasal BB250 48 (3)t 105 (5)t 28 (7)* 78 (4)t Ret, releasing about 40% of total gly- BB87 61 (4)t 108 (2)t 97 (4)t 85 (5)t cosaminoglycan by day 5 in the presence of the BB2001 12 (7) 5 (18) -13 (10) -4 (4) vitamin, and about 10% in its absence. BB250 BB3003 22 (9) -15 (14) -14 (3) -6 (6) BB2014 32 (4)t -10 (17) -68 (12)* -19 (7) and BB87 at a dose of 10 jiM both strongly inhibited the release of proteoglycan from Articular BB87 56 (6)t 137 (10)* 146 (6)t 104 (9)* human cartilage. In one sample, basal levels of BB250 76 (3)t 110 (13)t 163 (4)t 91 (6)* release were inhibited by (mean (SEM)) 32% (12%) (p < 0.001) by BB250 and 41% (6%) (p In the analyses of stimulated release basal levels of proteoglycan release were subtracted. The results are shown as percentage inhibition of proteoglycan release, expressed for purposes of clar- < 0.0001) by BB87. Inhibition of Ret stimu- ity as mean (SEM). *p < 0.005; tP < 0.0005 compared with controls. Values of greater than 100% lated release was even more notable, the two inhibition of stimulated release reflect total inhibition of stimulated release plus some inhibition of inhibitors giving 93% (5%) (p < 0.001) and the basal level. 99% (4%) (p < 0.001), respectively. In a Table 3 Dose response for the inhibition of rhIL-lcxstimulated proteoglycan release from second experiment with cartilage from a differ- bovine nasal cartilage explants by BB87 and BB250 ent patient, BB250 at a concentration of 10 jiM produced 38% (16%) (NS) inhibition of basal Per cent inhibition release and 102% (5%) (p < 0.001) inhibition Inhibitor concentration (,uM) BB250 BB87 of Ret stimulated proteoglycan release. The effects of rhIL-lat and BB87 on bovine 0.001 27 (20) nasal cartilage explants were assessed histologi- 0.01 43 (17) 0.1 26 (7)* 7 (6) cally, by use of toluidine blue, which binds to 1 77 (5)t 35 (6)t sulfated glycosaminoglycans. Cartilage cul- 10 105 (5)t 108 (2)t tured for 24 hours in the absence of catabolic Cartilage explants were cultured for two days with or without rhIL-la and varying concentrations stimulus retained most of its metachromasia of the MMP inhibitors before the culture media and explants were analysed for sulfated and hence its proteoglycan content (fig IA), glycosaminoglycan content. Results are expressed as mean (SEM). *p < 0.005; tp < 0.0005 com- the sections staining darker in the territorial pared with controls. than in the interterritorial matrix. This was also (table 2), though again BB250 and BB87 were the case with cartilage taken from freshly the most potent. Inhibition of basal release of slaughtered cattle and processed for histology proteoglycan by the gelatinase inhibitor without the culture period (not shown). BB20 14 was also statistically significant. Explants cultured in the presence of rhIL- 1 x BB20 14 produced a statistically significant for 24 hours stained less intensely (fig 1 B). The stimulation of proteoglycan release in the pres- presence of BB250 in rhIL-lx treated 24 hour ence of rhTNFax, however. Cells responsive to cultures was chondroprotective, the histologi- http://mp.bmj.com/ TNF are capable of shedding their TNF cal appearance (fig 1 C) being very similar to receptors, a process dependent on a proteolytic that of the unstimulated controls (fig 1A). cleavage.38 The soluble receptors then compete with signalling receptors for the ligand. Our EFFECTS OF BB87 AND BB250 ON THE result could thus be explained by the inhibition METABOLIC ACTIVITY OF NASAL SEPTUM by BB2014 of a proteinase involved in TNF CARTILAGE EXPLANTS Proteoglycan release from cartilage explants on September 29, 2021 by guest. Protected copyright. receptor shedding, or indeed of an enzyme that can be inhibited non-specifically by toxic com- is capable of degrading TNFa. These possibili- pounds such as iodoacetate.32 We therefore ties were not explored further. evaluated the effects of the inhibitory com- We next determined whether BB250 and pounds on some metabolic activities of BB87 were capable of inhibiting proteoglycan chondrocytes, and determined whether the breakdown from bovine metacarpophalangeal inhibitory effect on proteoglycan degradation cartilage. As can be seen in table 2, a 10 jM was reversible. concentration of either inhibitor produced vir- The ability of nasal cartilage explants to tually complete inhibition ofrhIL-1 a, rhTNFcx recover from the effects of BB87 and BB250 and Ret stimulated release, as well as a was investigated. The explants were cultured substantial inhibition of basal proteoglycan for two days with rhIL-1la (0.3 nM) and either breakdown. of the inhibitors (10 gM) after which the In cultures of rhIL-lax stimulated nasal explants were allowed to recover in cytokine- cartilage explants a dose dependent inhibition free DMEM for 24 hours. The explants were of proteoglycan release was seen in the then placed in DMEM containing 0.3 nM presence ofboth BB87 and BB250 (table 3). At rhIL-i1 . During the first incubation with a concentration of 1 jM both inhibitors inhibitor, BB87 and BB250 inhibited rhIL-1la produced statistically significant reductions in stimulated release by 78% and 59%, respec- proteoglycan release from the tissue, though tively. Because of the inhibition ofproteoglycan BB250 seemed to be the more potent. This was breakdown during the first incubation with the confirmed at the lower concentration of 0.1 jiM, inhibitors, these explants retained a larger where BB250, but not BB87, produced statisti- amount of glycosaminoglycan that could be cally significant inhibition. subsequently lost from the tissue. Indeed, when A small number of experiments were carried these explants were thereafter stimulated with out to investigate the inhibitory effects of BB87 rhIL- 1 oC in the absence of the inhibitors they M336 Brown, Rahman, Morton, Beauchamp, Bramwell, Buttle
Chondrocytes are known to respire largely Clin Mol Pathol: first published as 10.1136/mp.49.6.M331 on 1 December 1996. Downloaded from ....E,... ;t ;
7900 (700) (mean (SEM)) counts per minute http://mp.bmj.com/ (cpm)/disc was not significantly altered by the presence of BB87 (7200 (700) cpmldisc). As expected, rhIL-lIx significantly depressed 35SO4 incorporation, to 3000 (200) cpmldisc (p < 0.0001). Again, this value was not significantly changed by the presence of BB87 (2900 (200)
cpmldisc). on September 29, 2021 by guest. Protected copyright. We conclude from these studies on the metabolic activity of cartilage explants that the inhibitory action of the MMP inhibitors on proteoglycan breakdown is likely to be due to a specific inhibition of the activity of a metallo- proteinase, rather than to any toxic or non- specific effect of the compounds.
LEVELS OF MMP ACTIVITY DURING THE TIME-COURSE OF PROTEOGLYCAN BREAKDOWN As described in the Methods section, we used a Figure1 Toluidine blue staining ofsections from bovine nasal cartilage explants cultured quenched fluorescent substrate that is cleaved by in the absence and presence of rhIL-la and BB87. Sections were stained with 1% (w/v) all the MMPs to monitor the appearance of toluidine blue. (A) Section from a cartilage explant culturedfor 24 hours in DMEM alone. MMP activity during the time-course of proteo- (B) Section from a cartilage explant cultured in the presence of rhIL-lafor 24 hours. (C) Section from a cartilage explant cultured in the presence of rhIL-la and BB87 (10jM) glycan breakdown in bovine nasal septum for 24 hours. Original magnification x200. cartilage cultures. Very low levels of activity were found in media conditioned by the explants, even released even more sulfated glycosaminoglycan after five days of culture, irrespective of whether (a mean of 1728 tggdisc following exposure to the explants were cultured in the presence or BB250 and 1548 ig/disc after BB87 treat- absence of agents that stimulate proteoglycan ment) than discs treated with rhIL-1la and no breakdown (table 4), even though the amount of inhibitor for each incubation (404 ig/disc), glycosaminoglycan released into the medium was showing that the effects of the inhibitors were increased by at least threefold in the presence of reversible. the catabolic stimuli (results not shown). Slightly Inhibitors ofcollagenase but not ofgelatinase reduce cartilage explant proteoglycan breakdown M337
Table 4 Quenchedfluorescent substrate hydrolysing activity in media conditioned by Discussion bovine nasal cartilage explants The results presented here are in agreement Clin Mol Pathol: first published as 10.1136/mp.49.6.M331 on 1 December 1996. Downloaded from Days 0-2 Days 3-5 with earlier work demonstrating the efficacy of MMP inhibitors at limiting the breakdown of APMA cartilage proteoglycan in experimental Stimulant - + - + systems."'-"' Our findings also suggest that at least one metalloproteinase may be involved None 0.39 ± 0.12 0.75 ± 0.72 0.48 ± 0.30 2.61 ± 1.50 rhIL-la 0.21 ± 0.21 1.14 ± 0.81 - - irrespective of the catabolic stimulus to carti- rhTNFa 0.57 ± 0.33 1.29 ± 0.78 0.51 ± 0.33 2.04 ± 1.33 lage breakdown (the cytokines IL-1 and TNF, Ret 0.60 ± 0.21 2.28 ± 1.05 0.36 ± 0.27 2.46 ± 1.74 or the vitamin Ret) in both bovine and human Explants were cultured in serum-free DMEM with or without a catabolic stimulant for up to five systems. As far as we are aware, this study is the days with one medium change on day 2. The conditioned media were assayed for Mca-Pro-Leu- first to investigate the effect of MMP inhibitors Gly-Leu-Dpa-Ala-Arg-NH2-hydrolysing activity in continuous rate assays as described in Meth- on TNF stimulated breakdown and the first to ods. The results are expressed as pmol/min of product per explant. demonstrate inhibition of proteoglycan release Table S Quenchedfluorescent substrate hydrolysing activity in bovine nasal explant from human cartilage by MMP inhibitors. cultures Our results clearly show the lack of inhibi- tion of stimulated proteoglycan breakdown by Time (hours) compounds, BB2001, BB2014 and BB3003, Stimulant BB94 1 2 4 6 24 that show a considerable degree of selectivity for the inhibition of gelatinase activity. It there- None - 0.78 ± 0.04 0.66 ± 0.05 0.84 ± 0.06 0.78 ± 0.07 0.54 ± 0.03 None + 0.66 ± 0.05 0.84 ± 0.08 0.60 ± 0.04 0.66 ± 0.05 0.54 ± 0.03 fore seems unlikely that gelatinase(s) is in- rhIL-la - 0.78 ± 0.07 0.72 ± 0.06 0.84 ± 0.07 0.60 ± 0.04 0.48 ± 0.04 volved in proteoglycan breakdown stimulated rhIL-la + 0.72 ± 0.05 0.54 ± 0.06 0.66 ± 0.05 0.60 ± 0.04 0.48 ± 0.01 by IL-1, TNF or Ret in our model systems. Explants were cultured in PBS and glucose, as described in Methods, in the presence of Mca-Pro- We found that inhibitors of collagenase-type Leu-Gly-Leu-Dpa-Ala-Arg-NH2 (10,uM), with or without BB94 (10 p.M). The increase in fluo- MMPs were effective at inhibiting cartilage rescence, as compared with the amount of fluorescence in unconditioned PBS, was then proteoglycan breakdown, implicating the activ- determined. The results are expressed as pmol/min of product per explant. ity of a collagenase-like MMP in proteoglycan more activity was present following treatment loss. We must interpret this result with some with APMA, suggesting the presence of low caution, because inhibitor concentrations well amounts of proMMPs. The low levels of activity in excess of the IC50 values determined in in generated during the time-course of proteo- vitro experiments must be used to obtain inhi- glycan breakdown (< 0.6 pmol/min substrate bition of proteoglycan breakdown in tissue sys- without APMA treatment) tems. An example of this phenomenon was our hydrolysed/explant, earlier work with BB94,'5 an inhibitor with low can be contrasted with the level of over 15 000 nanomolar inhibition constants with all the pmol/min/explant found in week 3 medium from MMPs tested so far. Even so, in terms of the explants cultured in the presence ofIL- lia, when inhibition of proteoglycan breakdown, concen- collagen breakdown is occurring.45 There is thus trations of about 0.1 were to effect ,uM required http://mp.bmj.com/ a difference in MMP activity of over four orders 50% inhibition. The reasons for this apparent ofmagnitude, between the times ofproteoglycan discrepancy are not yet clear, but may be and collagen breakdown. related to problems of accessibility to the target It is possible that the bulk ofMMP activity in enzyme in cartilage, competition between the our cultures remained associated with the inhibitor and very high concentrations of explants rather than being secreted into the substrate within cartilage or, alternatively, may
surrounding medium. It has been reported that suggest that the physiologically important on September 29, 2021 by guest. Protected copyright. active stromelysin 146 and other MMPs bind to enzyme has yet to be identified. Our finding components of the extracellular matrix, and that very low levels of activity of the known indeed the enzymes can be extracted from solid MMPs were present in cultures during the tissues.47 In order to examine this possibility we time-course of proteoglycan breakdown determined the levels of MMP activity in the strengthens the latter possibility. It is worth cultures themselves, as opposed to the activity noting that compounds designed to inhibit the present in isolated conditioned media (earlier). MMPs are not necessarily entirely specific for Once again, the levels of active enzyme were this family of enzymes. Some other metallopro- very small, and there were no differences in the teinases share sufficient structural similarity, amounts of activity detected in cultures with particularly around the active site, to be grouped together with the MMPs as the and without rhIL- 1 a, or any increase of metzincins41 or proteinase clan MB.4 Thus, the enzyme activity with time (table 5). Most structural determinants that govern interac- important was the finding that the activity was tions with small molecule inhibitors may be not inhibited by BB94, a broad spectrum shared features between these families of met- MMP inhibitor that also inhibits proteoglycan alloproteinases. An example is the binding with breakdown.'5 We conclude from these experi- similar affinity of a small molecule inhibitor to ments that MMP activity is very low during the gelatinase A and to atrolysin C (EC 3.4.24.42), first few days in cartilage explants cultured in a snake venom proteinase of the reprolysin the presence of agents, such as IL-1 and TNF, family.49 Mammalian representatives of the that have been reported to promote the expres- reprolysin family have been identified.4 sion of these enzymes"-'' as well as proteogly- Stromelysin 1 has long been considered to be can breakdown. Instead, we found that levels of the most likely MMP to be involved in cartilage active MMPs in cartilage explant cultures cor- breakdown, either by a direct mechanism9 or related with collagen breakdown.45 via activation of other proMMPs, such as M338 Brown, Rahman, Morton, Beauchamp, Bramwell, Buttle
procollagenase.50 However our data might allow us to discriminate between the inhibition Clin Mol Pathol: first published as 10.1136/mp.49.6.M331 on 1 December 1996. Downloaded from argue against a role for stromelysin 1 in that of an enzyme acting directly on aggrecan as BB87, a potent inhibitor of cartilage proteogly- opposed to an enzyme involved in "proaggreca- can breakdown (table 2), is a 20-fold tighter nase" activation. The involvement of members binding inhibitor of interstitial collagenase of the gelatinase group of MMPs in the events than it is of stromelysin 1 (table 1). Although a leading to aggrecan breakdown in our model recent independent study5" has also provided system does now seem unlikely, however. evidence against a role for stromelysin in In any event, the demonstration of potent chondrocyte mediated proteoglycan break- inhibition of chondrocyte mediated aggrecan down, confirmation will await the results of breakdown by inhibitors of collagenase-like experiments using probes that are more MMPs may provide a lead for the development selective for stromelysins. of selective inhibitors as therapeutic agents for The loss of aggrecan from human cartilage, the treatment of arthritic diseases. The same both in various disease states and in tissue cul- inhibitors also inhibit type II collagen break- ture models, predominantly occurs coincident down in cartilage explants45 and discriminating with a specific proteolytic cleavage produced inhibitors are less likely to produce unwanted by the action of an as yet unidentified protein- side effects than those with a broader inhibition ase referred to as "aggrecanase", at a Glu-Ala profile. bond within the interglobular region towards the N-terminus of aggrecan. This results in the This work was supported by grants from the Arthritis & Rheu- loss from the matrix of almost the complete matism Council (UK). We thank Mr D L Douglas, Orthopaed- aggrecan molecule."5 52-55 The Glu-Ala bond is ics Consultant, Royal Hallamshire and Children's Hospitals, not cleaved in vitro by any known MMP other Sheffield, for supplying us with post-operative cartilage samples. than neutrophil collagenase (EC 3.4.24.34),56 1 Kempson GE, Tuke MA, Dingle JT, Barrett AJ, Horsfield perhaps suggesting that neutrophil collagenase PH. The effects of proteolytic enzymes on the mechanical is "aggrecanase" (reviewed elsewhere3). Al- properties of adult human articular cartilage. Biochim Bio- phys Acta 1976;428:741-60. though data presented here, and the fact that 2 Kempson GE, Muir H, Swanson SAV, Freeman MAR. Cor- neutrophil collagenase has been identified relations between stiffness and the chemical constituents of cartilage on the human femoral head. Biochim Biophys Acta recently in chondrocytes,57 would seem to sup- 1970;215:70-7. port this hypothesis, there are observations that 3 Buttle DJ, Bramwell H, Hollander AP. Proteolytic mecha- nisms of cartilage breakdown - a target for arthritis lend credence to the view that the enzyme therapy? Jf Clin Pathol: Mol Pathol 1 995;48:M1 67-77. responsible for aggrecan cleavage is not neu- 4 Rawlings ND, Barrett AJ. Evolutionary families of metal- lopeptidases. Methods Enzymol 1995;248:183-228. trophil collagenase. Neutrophil collagenase 5 Freije JMP, Diez-Itza I, Balbin M, Sanchez LM, Blasco R, cleaves another bond in the interglobular Tolivia J, et al. Molecular cloning and expression of collagenase-3, a novel human matrix metalloproteinase region of aggrecan, a Asn-Phe bond that is the produced by breast carcinomas. Jf Biol Chem 1994;269: favoured cleavage site ofmost MMPs,58-60 more 16766-73. 6 Murphy G, Reynolds JJ. Extracellular matrix degradation. efficiently than it cleaves the Glu-Ala bond.56 In: Royce P, Steinmann B, eds. Connective tissue and its her-
Although the product of cleavage of the itable disorders. New York: Wiley-Liss Inc, 1993:287-316. http://mp.bmj.com/ 7 Murphy GJP, Murphy G, Reynolds JJ. The origin of matrix Asn-Phe bond can be found in human metalloproteinases and their familial relationships. FEBS cartilage,6' the available evidence suggests that Lett 1991 ;289:4-7. 8 Murphy G, Cockett MI, Ward RV, Docherty AJP. Matrix this is only a minor product and that cleavage at metalloproteinase degradation of elastin, type IV collagen the Glu-Ala bond prevails in the human and proteoglycan. A quantitative comparison of the activi- ties of 95 kDa and 72 kDa gelatinases, stromelysins-1 and arthritic diseases54 and in cartilage culture -2 and punctuated metalloproteinase (PUMP). Biochem J systems.55 1991 ;277:277-9. 9 Dean DD, Martel-Pelletier J, Pelletier J-P, Howell DS, Similarly, there is evidence that interstitial col- Woessner JF. Evidence for metalloproteinase and metallo- on September 29, 2021 by guest. Protected copyright. lagenase and collagenase-3 are not responsible proteinase inhibitor imbalance in human osteoarthritic cartilage. J Clin Invest 1989;84:678-85. for aggrecan breakdown. Although interstitial 10 Martel-Pelletier J, Cloutier J-M, Pelletier J-P. In vivo effects collagenase is capable ofslowly cleaving aggrecan of antirheumatic drugs on neutral collagenolytic proteases in human rheumatoid arthritis cartilage and synovium.J in the test tube,62 and cleaves the interglobular Rheumatol 1988;15:1198-204. region at two distinct sites,59 neither of these 11 MacNaul KL, Chartrain N, Lark M, Tocci MJ, Hutchinson NI. Discordinate expression of stromelysin, collagenase, correspond to the "aggrecanase" cleavage site. and tissue inhibitor of metalloproteinases-I in rheumatoid Most importantly we have demonstrated that vir- synovial fibroblasts. Synergistic effect of interleukin-1 and tumor necrosis factor-a on stromelysin expression. J Biol tually no active collagenase is produced by bovine Chem 1990;265: 17238-45. cartilage explants during the first week of 12 Hutchinson NI, Lark MW, MacNaul KL, Harper C, Hoer- rner LA, McDonnell J, et al. In vivo expression of strome- culture.45 It is also worth noting that the natural lysin in synovium and cartilage of rabbits injected protein inhibitor of the MMPs, Timpi, does not intraarticularly with interleukin-1p. Arthritis Rheum 1992; 35:1227-33. inhibit cartilage explant proteoglycan release.'6 13 Dayer JM, Breard J, Chess L, Krane SM. Participation of This observation, together with the very low lev- monocyte-macrophages and lymphocytes in the produc- tion of a factor that stimulates collagenase and prostaglan- els of MMP activity present in our cultures, din release by rheumatoid synovial cells. J7 Clin Invest 1979; virtually precludes a role for any of the known 64:1386-92. 14 Murphy G, Hembry RM, Reynolds JJ. Characterization of a MMPs in aggrecan degradation. Our results do specific antiserum to rabbit stromelysin and demonstration indicate, however, that a collagenase-like metallo- of the synthesis of collagenase and stromelysin by stimulated rabbit articular chondrocytes. Collagen Relat Res proteinase may be important in aggrecan break- 1986;6:351-64. down, and the previous demonstration of 15 Buttle DJ, Handley CJ, Ilic MZ, Saklatvala J, Murata M, Barrett AJ. Inhibition of cartilage proteoglycan release by a "aggrecanase"-like activity by neutrophil specific inactivator of cathepsin B and an inhibitor of collagenase56 may suggest that "aggrecanase" is a matrix metalloproteinases. Evidence for two converging pathways of chondrocyte-mediated proteoglycan degrada- metalloproteinase with some active site features tion. Arthritis Rheum 1993;36:1709-17. similar to those of the neutrophil enzyme. 16 Andrews HJ, Plumpton TA, Harper GP, Cawston TE. A synthetic peptide metalloproteinase inhibitor, but not The cascade nature of the events leading up Timp, prevents the breakdown of proteoglycan within to aggrecan breakdown3 unfortunately do not articular cartilage in vitro. Agents Actions 1992;37:147-54. Inhibitors ofcollagenase but not ofgelatinase reduce cartilage explant proteoglycan breakdown M339
17 Nixon JS, Bottomley KMK, Broadhurst MJ, Brown PA, cytokines on glycolysis and fructose 2,6-bisphosphate in Johnson WH, Lawton G, et al. Potent collagenase inhibitors rheumatoid synovial cells in vitro. Biochem _7 1988;250: Clin Mol Pathol: first published as 10.1136/mp.49.6.M331 on 1 December 1996. Downloaded from prevent interleukin 1-induced cartilage degradation in 111-15. vitro. IntJ Tiss Reac 1991;13:237-43. 42 Buttle DJ, Saklatvala J. Lysosomal cysteine endopeptidases 18 Caputo CB, Sygowski LA, Wolanin DJ, Patton SP, Caccese mediate interleukin 1-stimulated cartilage proteoglycan RG, Shaw A, et al. Effect of synthetic metalloprotease degradation. Biochem37 1992;287:657-61. inhibitors on cartilage autolysis in vitro. 7 Pharmacol Exp 43 Tyler JA. Articular cartilage cultured with catabolin (pig Ther 1987;240:460-5. interleukin 1) synthesizes a decreased number of normal 19 Cawston T. Blocking cartilage destruction with metallopro- proteoglycan molecules. Biochem. 1985;227:869-78. teinase inhibitors: a valid therapeutic target? Ann Rheum 44 Campbell MA, Handley CJ. The effect of retinoic acid on Dis 1993;52:769-70. proteoglycan biosynthesis in bovine articular cartilage cul- 20 Davies B, Brown PD, East N, Crimmin NJ, Balkwill FR. A tures. Arch Biochem Biophys 1987;253:462-74. synthetic matrix metalloproteinase inhibitor decreases 45 Kozaci LD, Buttle DJ, Hollander AP. Degradation of type II tumor burden and prolongs survival of mice bearing collagen, but not proteoglycan, correlates with matrix met- human ovarian carcinoma xenografts. Cancer Res 1993;53: alloproteinase activity in cartilage explant cultures. Arthritis 2087-91. Rheum (in press). 21 Rich DH, Brown MA, Barrett AJ. Purification of cathepsin 46 Allan JA, Hembry RM, Angal S, Reynolds JJ, Murphy G. B by a new form of affinity chromatography. Biochem J Binding of latent and high Mr active forms ofstromelysin to 1986;235:731-4. collagen is mediated by the C-terminal domain. . Cell Sci 22 Barrett AJ. Human cathepsin B 1. Purification and some 1991 ;99:789-95. properties of the enzyme. Biochem3J 1973;131:809-22. 47 Woessner JF. Quantification of matrix metalloproteinases in 23 Piquilloud Y, Reinharz A, Roth M. Studies on the tissue samples. Methods Enzymol 1995;248:510-28. angiotensin converting enzyme with different substrates. 48 Bode W, Gomis-Ruth F-X, Stockler W. Astacins, serralysins, Biochim BiophysActa 1970;206:136-42. snake venom and matrix metalloproteinases exhibit identi- 24 Florentin D, Sassi A, Roques BP. A highly sensitive fluoro- cal zinc-binding environments (HEXXHXXGXXH and metric assay for "enkephalinase", a neutral metalloen- Met-turn) and topologies and should be grouped into a dopeptidase that releases tyrosine-glycine-glycine from common family, the 'metzincins'. FEBS Lett 1993;313: enkephalins. Anal Biochem 1984;141:62-9. 134-40. 25 Poole AR, Hembry RM, Dingle JT. Cathepsin D in 49 Zhang D, Botos I, Gomis-Riith F-X, Doll R, Blood C, cartilage: the immunohistochemical demonstration of Njoroge FG, et al. Structural interaction of natural and extracellular enzyme in normal and pathological condi- synthetic inhibitors with the venom metalloproteinase, tions. J Cell Sci 1974;14:139-61. atrolysin C (form d). Proc Natl Acad Sci USA 1994;91: 26 Knight CG, Willenbrock F, Murphy G. A novel coumarin- 8447-51. labelled peptide for sensitive continuous assays of the 50 Murphy G, Cockett MI, Stephens PE, Smith BJ, Docherty matrix metalloproteinases. FEBS Lett 1992;296:263-6. AJP. Stromelysin is an activator of procollagenase. A study 27 Knauper V, Lopez-Otin C, Smith B, Knight G, Murphy G. with natural and recombinant enzymes. Biochem 7 1987; Biochemical characterization of human collagenase-3. J 248:265-8. Biol Chem 1996;271:1544-50. 51 Yocum SA, Lopresti-Morrow LL, Gabel CA, Milici AJ, Mitch- 28 Cawston TE, Barrett AJ. A rapid and reproducible assay for ell PG. Bafilomycin Al inhibits IL-1-stimulated proteoglycan collagenase using [1-'4C] acetylated collagen. Anal Biochem degradation by chondrocytes without affecting stromelysin 1979;99:340-5. synthesis. Arch Biochem Biophys 1995;316:827-35. 29 Cawston TE, Murphy G. Mammalian collagenases. Methods 52 Sandy JD, Neame PJ, Boynton RE, Flannery CR. Catabo- Enzymol 1981;80:711-22. lism of aggrecan in cartilage explants. _7 Biol Chem 30 Cawston TE, Galloway WA, Mercer E, Murphy G, 1991;266:8683-5. Reynolds JJ. Purification of rabbit bone inhibitor of 53 Sandy JD, Flannery CR, Neame PJ, Lohmander LS. The collagenase. Biochem J 1981 ;195: 159-65. structure of aggrecan fragments in human synovial fluid. 31 Sellers A, Reynolds JJ, Meikle MC. Neutral metalloprotein- Evidence for the involvement in osteoarthritis of a novel ases of rabbit bone. Separation in latent forms of distinct proteinase which cleaves the Glu373-Ala374 bond in the enzymes that when activated degrade collagen, gelatin and interglobular domain. _7 Clin Invest 1992;89:1512-16. proteoglycans. Biochem I 1979;171:493-6. 54 Lohmander LS, Neame PJ, Sandy JD. The structure of 32 Buttle DJ, Saklatvala J, Tamai M, Barrett AJ. Inhibition of aggrecan fragments in human synovial fluid. Evidence that interleukin 1-stimulated cartilage proteoglycan degrada- aggrecanase mediates cartilage degradation in inflamma- tion by a lipophilic inactivator of cysteine endopeptidases. tory joint disease, joint injury, and osteoarthritis. Arthritis Biochem 7 1992;281: 175-7. Rheum 1993;36: 1214-22. 33 Handley CJ, Buttle DJ. Assay of proteoglycan degradation. 55 Ilic MZ, Handley CJ, Robinson HC, Mok MT. Mechanism Methods Enzymol 1995;248:47-58. of catabolism of aggrecan by articular cartilage. Arch 34 Andrews HJ, Edwards TA, Cawston TE, Hazleman BL. Biochem Biophys 1992;294: 115-22. http://mp.bmj.com/ Transforming growth factor-beta causes partial inhibition 56 Fosang AJ, Last K, Neame PJ, Murphy G, Knauper V, of interleukin-l-stimulated cartilage degradation in vitro. Tschesche H, et al. Neutrophil collagenase (MMP-8) Biochem Biophys Res Commun 1989;162:144-50. cleaves at the aggrecanase site E373-A374 in the 35 Farndale RW, Buttle DJ, Barrett AJ. Improved quantitation interglobular domain of cartilage aggrecan. Biochem _7 and discrimination of sulphated glycosaminoglycans by use 1994;304:347-5 1. of dimethylmethylene blue. Biochim Biophys Acta 1986; 57 Cole AA, Chubinskaya S, Schumacher B, Huch K, Cs-Szabo 883:173-7. G, Yao J, et al. Chondrocyte matrix metalloproteinase-8. 36 Buttle DJ, Saklatvala J, Barrett AJ. The inhibition of Human articular chondrocytes express neutrophil colla- interleukin 1-stimulated cartilage proteoglycan degrada- genase.JBiol Chem 1996;271:11023-6. 58 Neame Hamil- tion by cysteine endopeptidase inactivators. Agents Actions Fosang AJ, PJ, Hardingham TE, Murphy G, on September 29, 2021 by guest. Protected copyright. Suppl 1993;39:161-5. ton JA. Cleavage of cartilage proteoglycan between GI and 37 Saklatvala J. Tumour necrosis factor a stimulates resorption G2 domains by stromelysins. _7 Biol Chem 1991 ;266: and inhibits synthesis of proteoglycan in cartilage. Nature 15579-82. 1986;322:547-9. 59 Fosang AJ, Last K, Knauper V, Neame PJ, Murphy G, 38 Gullberg U, Lantz M, Lindvall L, Olsson I, Himmler A. Hardingham TE, et al. Fibroblast and neutrophil colla- Involvement of an Asn/Val cleavage site in the production genases cleave at two sites in the cartilage aggrecan of a soluble form of a human tumor necrosis factor (TNF) interglobular domain. BiochemJ 1993;295:273-6. receptor. Site-directed mutagenesis of a putative cleavage 60 Fosang AJ, Neame PJ, Last K, Hardingham TE, Murphy G, site in the p55 TNF receptor chain. Eur . Cell Biol Hamilton JA. The interglobular domain of cartilage aggre- 1992;58:307-12. can is cleaved by PUMP, gelatinases and cathepsin B. . Biol 39 Stefanovic-Racic M, Stadler J, Georgescu HI, Evans CH. Chem 1992;267: 19470-4. Nitric oxide and energy production in articular chondro- 61 Flannery CR, Lark MW, Sandy JD. Identification of a strome- cytes. J Cell Physiol 1994;159:274-80. lysin cleavage site within the interglobular domain of human 40 Bird TA, Gearing AJH, Saklatvala J. Murine interleukin-l aggrecan. Evidence for proteolysis at this site in vivo in human receptor: differences in binding properties between fibro- articular cartilage. _7 Biol Chem 1992;267:1008-14. blastic and thymoma cells and evidence for a two-chain 62 Hughes C, Murphy G, Hardingham TE. Metalloproteinase receptor model. FEBS Lett 1987;225:21-6. digestion of cartilage proteoglycan. Pattern of cleavage by 41 Taylor DJ, Whitehead RJ, Evanson JM, Westmacott D, stromelysin and susceptibility to collagenase. Biochem _7 Feldmann M, Bertfield H, et al. Effect of recombinant 1991 ;279:733-9.