OBJECTIVE To determine if a new inhibitor, esculetin (EST), can block resorption of cartilage.
METHODS Interleukin 1α (IL1α, 0.04–5 ng/ml) and oncostatin M (OSM, 0.4–50 ng/ml) were used to stimulate the release of proteoglycan and collagen from bovine nasal cartilage and human articular cartilage in explant culture. Proteoglycan and collagen loss were assessed by dimethylmethylene blue and hydroxyproline assays, respectively. Collagenase levels were measured by assay of bioactivity and by enzyme linked immunosorbent assay (ELISA). The effects of EST on the expression of matrix metalloproteinases (MMPs) and tissue inhibitor of metalloproteinase-1 (TIMP-1) in the transformed human chondrocyte cell line T/C28a4 were assessed by northern blot analysis. TIMP-1 protein levels were assayed by ELISA. The effect of EST on the MMP-1 promoter was assessed using a promoter-luciferase construct in transient transfection studies.
RESULTS EST inhibited proteoglycan and collagen resorption in a dose dependent manner with significant decreases seen at 66 μM and 100 μM EST, respectively. Collagenolytic activity was significantly decreased in bovine nasal cartilage cultures. In human articular cartilage, EST also inhibited IL1α + OSM stimulated resorption and decreased MMP-1 levels. TIMP-1 levels were not altered compared with controls. In T/C28a4 chondrocytes the IL1α + OSM induced expression of MMP-1, MMP-3, and MMP-13 mRNA was reduced to control levels by 250 μM EST. TIMP-1 mRNA levels were unaffected by EST treatment. All cytokine stimulation of an MMP-1 luciferase-promoter construct was lost in the presence of the inhibitor.
CONCLUSION EST inhibits degradation of bovine nasal cartilage and human articular cartilage stimulated to resorb with IL1α + OSM.
- interleukin 1
- oncostatin M
Statistics from Altmetric.com
In rheumatoid arthritis (RA) and osteoarthritis (OA), progressive loss of articular cartilage occurs, in particular, degradation of collagen and proteoglycan.1 The integrity of these macromolecules is vital to cartilage and joint function. In response to various stimuli, degradation of proteoglycan is easily achieved relative to collagen, which is lost much less readily.2 However, proteoglycan can be resynthesised, whereas collagen, once degraded, is not replaced.3 4
Matrix metalloproteinases (MMPs) are key enzymes in RA and OA pathogenesis. Collectively they can degrade all components of the extracellular matrix.5-7 MMP-1, MMP-8, and MMP-13 specifically cleave fibrillar collagen to produce characteristic one quarter and three quarter length products.8-10 These MMPs and the collagen cleavage products are found in stimulated explant cultures, in the synovium and serum of rheumatoid patients, and in animal models, providing evidence of MMP involvement in collagen breakdown in vivo.11-15 In addition, a number of MMPs are reported to activate other members of the MMP family—for example, it has been proposed that MMP-3 plays a part in the activation cascade of MMP-1, leading to collagen degradation.16 17 Aggrecan, the principal proteoglycan in cartilage, can be cleaved at two key sites: Asn341-Phe342 and Glu373-Ala374.18 19 Degradation at the Asn341 site can be effected by a number of MMPs, including MMP-1, MMP-3, and MMP-8.19 20 However, a new metalloproteinase family, termed ADAM-TS (a disintegrin and metalloproteinase with thrombospondin motifs), may have a more significant effect than MMPs on aggrecan degradation because they cleave at the Glu373 site that is typical of aggrecan cleavage in vivo.21-24 These enzymes are produced by cartilage in culture but it is not clear if they are identical to the recently described membrane associated aggrecanase.25 In addition, synthesis of proteoglycan may be inhibited by the presence of cytokines, such as interleukin 1α (IL1α), with the overall effect of reducing the levels of proteoglycan present in the tissue.26
The natural inhibitors of MMPs are α2 macroglobulin and tissue inhibitors of metalloproteinases (TIMPs).27 28The former is a general endoproteinase inhibitor, whereas the latter interact with active MMPs, forming 1:1 complexes. The in vivo inhibitors of the aggrecan-degrading ADAM-TS proteins have not yet been identified. Normally, active MMP and TIMP levels are in equilibrium such that controlled turnover of the extracellular matrix occurs. However, in rheumatoid cartilage, this equilibrium moves in favour of the active MMPs and tissue destruction occurs, ultimately resulting in the loss of cartilage.14 29 30 Connective tissue turnover can be prevented by reducing levels of active MMPs or increasing levels of TIMPs, and treatments that act in this way are therefore of therapeutic interest.31-33
Current treatments for RA and OA reduce pain and swelling of the joint but generally do little to block cartilage destruction.34Evidence indicates that MMPs and ADAM-TS proteins degrade cartilage and are therefore potential targets for inhibition. Of the drugs currently in use, there have been conflicting reports of the effects of steroidal and non-steroidal treatments, when tested at therapeutic concentrations, on enzyme profiles.35 36 In addition, synthetic molecules have been developed that bind to the catalytic site of MMPs, inhibiting the enzymes and therefore potentially preventing cartilage degradation.32 36-42 A number of such inhibitors have been investigated in animal models and some are currently being examined in clinical trials. However, to date, these inhibitors are not used in routine clinical practice.
We investigated a new inhibitor of cartilage degradation, esculetin (EST) (fig 1). EST is a coumarin derivative which has been found to protect rabbit articular cartilage from proteoglycan loss in response to IL1α.43 We used established cartilage explant models inducing resorption with the proinflammatory cytokines IL1α and oncostatin M (OSM).44 45 We found that EST was chondroprotective for both collagen and proteoglycan loss, and decreased cytokine stimulated increases in collagenolytic activity and MMP-1 levels. In addition, using a transformed human chondrocyte cell line (T/C28a4), we found that EST decreased the expression of MMP-1, MMP-3, and MMP-13 mRNA, and inhibited MMP-1 promoter activity. These results indicate that EST can protect cartilage from degradation by acting pre-translationally.
Materials and methods
EST was supplied by Kureha Chemical Industry Co Ltd, Tokyo, Japan. It was solubilised in dimethyl sulphoxide (DMSO), stored at −20°C, and subsequently diluted into culture media with a final DMSO concentration less than 0.06% (v/v). Bovine nasal septa were obtained at the time of slaughter. Fresh human articular cartilage was obtained from a local orthopaedic unit. Both tissues were stored at 4°C and used within 24 hours. The transformed human chondrocyte cell line T/C28a4 was a gift from Professor Mary Goldring (Harvard Institute for Medicine, Boston, MA, USA46). Dulbecco's modified Eagle's medium (DMEM) and fetal calf serum (FCS) were purchased from Gibco BRL (Paisley, UK). DMEM was supplemented with penicillin-streptomycin (penicillin 50 U/ml, streptomycin 50 μg/ml), nystatin (20 U/ml), gentamycin (5 μg/ml), and glutamine (2 mmol/l).l-Ascorbic acid was purchased from Wako Pure Chemical Industries (Alpha Laboratories, Eastleigh, UK). ITS+ Premix was from Becton Dickinson (Bedford, MA, USA). Human recombinant IL1α and OSM were gifts from Glaxo-Wellcome (Stevenage, UK) and Professor J Heath (Department of Biochemistry, University of Birmingham, UK), respectively. α[32P]dCTP was purchased from Amersham Pharmacia Biotech (Little Chalfont, UK). All other chemicals were reagent grade. The MMP-1 promoter-luciferase construct (-517/+63 base pairs (bp))47 was supplied by Professor C Brinckerhoff (Dartmouth Medical School, NH, USA).
BOVINE TISSUE EXPERIMENTS
Discs (1 mm diameter) were cut from 1 mm thick slices of bovine nasal cartilage. Three discs per well were placed in 600 μl supplemented DMEM overnight at 37°C, 5% CO2, 95% air. Medium was removed and replaced with DMEM, with or without test reagents, thus: IL1α (0.04–1 ng/ml), OSM (0.4–10 ng/ml), and EST (0–330 μmol/l) to a final volume of 600 μl. Discs were incubated for 14 days. Media were harvested and replenished at day 7 with identical test reagents. At day 14, media were harvested, and the remaining cartilage papain digested overnight at 65°C as described previously.44 Media and digested cartilage were stored at −20°C until assayed.
HUMAN TISSUE EXPERIMENTS
Macroscopically normal human articular cartilage from excised knee joints was cut into 2 mm3 pieces. Articular cartilage from different patients was used in separate experiments. Three pieces were placed in each well in 600 μl supplemented DMEM, and incubated overnight as above. Media were removed and replaced with supplemented DMEM containing IL1α (5 ng/ml), OSM (50 ng/ml), and EST (0–500 μmol/l) to a final volume of 600 μl. At days 7 and 14, media were harvested and replenished as above. At day 21, media were harvested and the remaining cartilage papain digested. Media and digested cartilage were stored at −20°C until assayed.
Viability of cartilage explants was assessed by screening for the production of lactate dehydrogenase (LDH) using the Cytotox 96 assay (Promega, Southampton, UK). No increase in LDH levels with any of the cytokine or inhibitor combinations was found (data not shown). Serum is excluded from cartilage explants because it can increase cartilage metabolism in the absence of exogenous cytokine(s).48 This study aimed at creating a model of cartilage breakdown; therefore the presence of serum was avoided as it contains chondroprotective agents, such as insulin-like growth factor 1.49 50 The absence of serum was also shown not to affect the viability of the tissue (data not shown), and previous studies have shown that cartilage in serum-free culture for eight to nine days can respond to serum and other growth factors.51 In this study, cytokines were added after one day of serum-free culture.
ASSAY OF BIOCHEMICAL MARKERS OF DEGRADATION
Proteoglycan and collagen in media and cartilage samples were assayed as previously described.44 Briefly, glycosaminoglycan (GAG) levels were determined by the amount of polyanionic material reacting with 1,9-dimethylmethylene blue, using shark chondroitin sulphate as standard. Levels of hydroxyproline in samples were assayed as a measure of collagen, using hydroxyproline as standard.44 The proteoglycan and collagen content of media and digested cartilage was expressed as a percentage of the total GAG or total hydroxyproline present in each explant. Collagenolytic activity present in media was assayed by the 3H diffuse collagen fibril assay,52 with modifications according to Koshy et al.53 When appropriate, p-aminophenylmercuric acetate (APMA) was added to 0.7 mmol/l to activate procollagenases. Levels of total MMP-1 (pro-form, active or TIMP-bound) and total TIMP-1 (free and bound to MMPs) in human media samples were assayed using specific enzyme linked immunosorbent assays (ELISAs).11 54
NORTHERN BLOT ANALYSIS
T/C28a4 cells were grown in T25 flasks in supplemented DMEM with 10% FCS, l-ascorbic acid (100 μg/ml). When cells reached 60% confluence the media were replaced with serum-free supplemented DMEM, with ITS+. After overnight culture, cells were washed with phosphate buffered saline and then stimulated with IL1α (1 ng/ml) and OSM (10 ng/ml) ± EST (50–500 μmol/l) in serum-free supplemented DMEM with ascorbate. Total cellular RNA was harvested 24 hours after stimulation using the RNeasy system (Qiagen, Crawley, UK) according to the manufacturer's instructions. Northern blot analysis was carried out essentially as previously described.45 In brief, for each treatment, 15–20 μg of RNA was fractionated on a 1% agarose gel and blotted onto a nitrocellulose membrane. Expression of mRNA was detected using complementary human DNA probes to MMP-1 and TIMP-1,45 MMP-3 (from A Galloway, British BioTechnology, Oxford, UK), MMP-13 (from Dr V Knäuper, University of East Anglia, Norwich, UK), c-fos, c-jun (from Dr I Clark, University of East Anglia, Norwich, UK), and rat glyceraldehyde-3-phosphate dehydrogenase (GAPDH).55 Autoradiographs of blots were scanned and radioactive bands quantified with Alphavision (Flowgen, Lichfield, UK). GAPDH was used to normalise the data.
TRANSIENT TRANSFECTION AND REPORTER GENE ASSAY
T/C28a4 cells at 50–80% confluence in 10 cm diameter tissue culture dishes were transfected with 8 μg of plasmid DNA for the MMP-1 promoter-luciferase construct (-517/+63 bp)47 and 32 μl of lipofectamine reagent (Gibco, BRL) per well according to the manufacturer's instructions. The following day, each dish was passaged into 2 × 6 well plates and incubated overnight in DMEM/1% AT-FCS at 37°C. After which, cells were incubated with IL1α ± OSM ± EST for 24 hours. Lysis buffer (400 μl/well, Promega) was applied at −20°C for two hours to release cytosolic proteins. Supernates were removed and used in the luciferase assay (below). The remaining nuclear pellets were immediately stored at −20°C until used in the Hirt's assay56 to quantify the amount of transfected plasmid. Supernates (10 μl) were mixed with 100 μl of luciferase substrate (beetle luciferin, Promega) in a white walled, 96 well plate (Wallac, Milton Keynes, UK) and read on a TriLux MicroBeta 1450 luminescence and scintillation counter (Wallac). Triplicate samples were performed in three experiments. Results were normalised to the amount of plasmid per well, as determined by Hirt's assay.
In experiments requiring statistical analysis, each treatment was performed in quadruplicate. Student's ttest was performed, assuming two tailed distributions and equal variance. Significance is denoted thus: *0.05>p>0.01; **0.01>p>0.001; ***p<0.001.
EFFECT OF EST ON THE RESORPTION OF BOVINE NASAL CARTILAGE
EST was found to block the release of collagen and proteoglycan loss from bovine nasal cartilage in culture (figs 2 and 3). EST inhibited IL1α + OSM induced collagen resorption in a dose dependent manner, with 66 μM EST significantly inhibiting collagen depletion (p<0.001; fig 2). Similar results were obtained when bovine nasal cartilage was cultured with IL1α alone (1 ng/ml); there was a significant reduction in the stimulated collagen release at 66 μM EST (p<0.05; data not shown).
Although the major emphasis of this study was collagen breakdown, proteoglycan resorption stimulated with IL1α (1 ng/ml) in combination with OSM (10 ng/ml) was also significantly reduced when EST was present at 330 μmol/l (p<0.05) but not at lower concentrations (data not shown). However, when the IL1α concentration was reduced to 0.2 or 0.04 ng/ml, with OSM maintained at 10 ng/ml, EST (100 μmol/l) could significantly inhibit the loss of proteoglycan (fig 3; p<0.05). Similar results were observed when OSM was decreased to 2 or 0.4 ng/ml, with IL1α at 1 ng/ml (fig 3; p<0.05).
PROCOLLAGENOLYTIC AND ACTIVE COLLAGENOLYTIC ACTIVITY IN THE PRESENCE OF EST
Having established that EST can inhibit the resorption of collagen, we investigated its effects on collagenolytic activity in the media harvested from the cultured bovine nasal cartilage (fig 4A). Addition of a metalloproteinase activator, APMA, activated procollagenases present, allowing the measurement of total collagenase levels (fig 4B). It is evident that the levels of collagenolytic activity mirrored those seen for collagen loss under the same conditions (figs 2, 4A and B). By day 14 of culture, 66 μM EST significantly inhibited IL1α + OSM stimulated increases in active and total collagenolytic activity (figs 4A and B). IL1α stimulated effects were also inhibited by EST (data not shown).
TIME COURSE OF EST EFFICACY IN BOVINE NASAL CARTILAGE
To determine if EST can prevent collagen release at various times after the addition of cytokines, the inhibitor was added to bovine nasal cultures up to 72 hours after the addition of IL1α + OSM. EST inhibited the release of collagen when added up to 72 hours after stimulation (fig 5A). If added thereafter, collagen loss was not inhibited (p<0.001). Similarly, levels of total and active collagenolytic activity were suppressed by EST added up to 72 hours after stimulation (figs 5B (p<0.05) and C).
EST EFFECTS ON HUMAN ARTICULAR CARTILAGE RESORPTION
To confirm the results from the bovine studies we examined the effects of EST on resorption of human articular cartilage. Previous studies have shown that the response of human articular cartilage to cytokines is less dramatic than that for animals, but that all human cartilages respond to IL1α + OSM by resorbing proteoglycan, with approximately 50% of these also releasing low levels of collagen.45 In this study with human articular cartilage, proteoglycan loss stimulated by IL1α + OSM, and MMP-1 levels were inhibited by EST (figs 6A and B): 50 μmol/l inhibited control levels (p<0.001 and p<0.01, respectively). Collagen loss was also significantly inhibited by EST at 50 μmol/l in the same explant culture (fig 6C; p<0.01). EST at 50 μmol/l effectively blocked cartilage resorption in four of seven experiments, but in the remaining experiments the effect was seen at higher concentrations of the inhibitor. In contrast with these results, no consistent effect of EST on TIMP-1 production was seen when assayed by ELISA (data not shown).
EFFECT OF EST ON MMP EXPRESSION IN CHONDROCYTES
We then used a human chondrocyte cell line to dissect the mechanism of action of EST. At 24 hours of stimulation, IL1α + OSM up regulated the expression of MMP-1, MMP-3, MMP-13, and TIMP-1 mRNA in the T/C28a4 cell line (fig 7). EST (100–500 μmol/l) inhibited the increased MMP expression in a dose dependent manner. EST at 250 μmol/l was sufficient to decrease mRNA to levels similar to those seen under control conditions (that is, media alone), while a partial effect was seen at 100 μmol/l and 50 μmol/l. Similar effects were seen at 48 and 72 hours of incubation (data not shown). Expression of TIMP-1 was unaffected by EST when normalised to GAPDH levels. GAPDH expression was also unaffected by the concentrations of EST used. No MMP mRNA was expressed in the presence of EST alone (data not shown).
EST INHIBITS CYTOKINE STIMULATED MMP-1 PROMOTER ACTIVITY
We have clearly shown that EST inhibits IL1α + OSM mediated MMP-1 gene expression, and therefore we studied its effects at the transcriptional level. To this end, we used the -517/+63 bp sequence of the human MMP-1 promoter cloned upstream of a luciferase reporter gene.47 The effect of cytokines and EST was assessed after transient transfection into T/C28a4 cells. Treatment of these cells with OSM (10 ng/ml) showed stimulation, whereas IL1α + OSM (1 and 10 ng/ml, respectively) synergistically stimulated the MMP-1 promoter construct (fig 8). IL1α alone was unable to stimulate this construct, as previously described.47 When 100 μM EST was added simultaneously with the cytokines to the transiently transfected cells, all stimulation of the promoter construct by OSM and IL1α + OSM was lost. Following on from this study, we investigated members of the activator protein 1 family that we believe plays a part in MMP transcription: we found that EST does not alter the up regulation of c-fos and c-jun expression induced by IL1α and/or OSM (data not shown). We are currently investigating the mechanism of action of EST further.
We have shown that the coumarin derivative, EST, can inhibit cartilage resorption induced by IL1α + OSM in bovine nasal cartilage and human articular cartilage explant cultures by significantly preventing collagen and proteoglycan degradation in a dose dependent manner. In bovine nasal cartilage, EST decreased the induced total and active collagenolytic activity. These findings corroborate the human articular cartilage data, which showed that total MMP-1 levels were reduced by EST. Further evidence indicates that EST prevents the IL1α + OSM stimulated expression of MMP-1, MMP-3, and MMP-13 mRNA, and thus the production of secreted proenzymes, whereas it has no effect on TIMP-1. In addition, in T/C28a4 cells EST (100 μmol/l) can abolish OSM ± IL1α stimulation of an MMP-1 promoter construct, indicating that EST can inhibit the cytokine stimulated activation of this promoter. Overall, these findings suggest that EST inhibits MMP production, and that it does not act by regulating TIMP-1 or by preventing activation of pro-MMPs. This work also confirms the earlier findings of Watanabe et al that EST does not interact directly with collagenases to inhibit their activity.43
It has been found that IL1α can inhibit proteoglycan synthesis, in addition to the stimulation of metalloproteinase production.26 This results in an overall loss of proteoglycan. Watanabe et al reported that EST does not alter the effects of IL1 on proteoglycan synthesis,43 and in this study we examined the inhibition of proteoglycan resorption by EST. To obtain significant inhibition of proteoglycan release from bovine nasal cartilage stimulated to resorb with IL1α and OSM (1 and 10 ng/ml respectively) required EST at 330 μmol/l. However, we have previously found that IL1α and OSM at these concentrations promote a maximal and rapid loss of proteoglycan by day 7 of culture of bovine nasal cartilage.45 To determine if lower EST concentrations could inhibit proteoglycan resorption, we decreased the concentration of these cytokines, thus decreasing the rate of resorption. Using cytokines at concentrations 5- to 20-fold lower than those typically used, we found that significant inhibition did occur with EST at 100 μmol/l. We conclude that EST can inhibit proteoglycan loss, but where explants are exposed to high levels of cytokines over prolonged periods then this inhibition can be overcome. In addition, it was apparent that the standard deviations increased. We have noted this on a number of occasions when cytokine concentrations are reduced, and propose that not all of the explants respond to the reduced levels of cytokine, resulting in an increased standard deviation.
In response to IL1α + OSM, it is typical for human articular cartilage to respond with rapid proteoglycan loss but low collagen degradation.57 The experiment illustrated had a higher collagen loss relative to many similar experiments. However, the potential inhibitory effect of EST in the human system was also determined by a specific ELISA, which showed that MMP-1 levels were always significantly reduced to control levels in the presence of the inhibitor, corroborating the collagen results in human articular cartilage.
EST can inhibit IL1α + OSM stimulated collagen degradation when added up to 72 hours after the cytokines. Furthermore, both the total and active levels of collagenase are decreased by addition of EST over the same time period. As mentioned above, it has been shown elsewhere that EST does not inhibit collagenases by direct interaction with the enzymes.43 Therefore, these results suggest that in disc culture, three days are required for IL1α + OSM to induce the production and secretion of collagenolytic proenzymes by chondrocytes. EST, which at 325 (mol wt) is relatively small, may be able to diffuse through the cartilage matrix to inhibit the IL1α + OSM effects. It may be that the cytokines' actions are exerted for up to three days of culture and are not required thereafter. In support of this, we have evidence to suggest that collagenolytic proenzymes are present in the matrix by day 3 of culture (data not shown). Furthermore, in culture of T/C28a4 chondrocytes, cytokines encounter no restrictions in reaching the cells and MMP mRNAs are produced within 24 hours of stimulation, with levels of MMP-1 protein rising by 24 hours of stimulation and peaking at 72 hours.45
There are a number of points at which EST could act: it may interfere with transcription factors that bind the promoters for MMP-1, MMP-3, or MMP-13 and so prevent transcription. We investigated members of the activator protein 1 family that are believed to be involved in MMP transcription: we found that EST does not alter the up regulation of c-fos and c-jun expression induced by IL1α and/or OSM,58suggesting that EST does not interfere with cytokine receptor components. These results for EST therefore seem to differ from a study reported by Martel-Pelletier et al,59 which showed decreased IL1 receptor expression on OA and RA synovial fibroblasts in the presence of the antirheumatic drug, Tenidap. EST might bind directly to upstream promoter sites of MMPs or to other associated transcription factor genes. It might also bind to and prevent the binding of known transcription factors to relevant genes. Alternatively, EST may interfere with cytokine signalling pathways, or with the cytokine receptor components.
The concentrations of EST required to block cartilage resorption were not due to cytotoxic effects. If cytotoxicity was the mode of action of EST, one would expect to see a decrease in the mRNA expressed for all proteins examined by northern blot. However, expression of TIMP-1 and GAPDH mRNAs was not affected by the EST concentrations used (0–500 μmol/l), indicating that transcription had not been universally altered, and suggesting that the inhibition by EST was not based on cytotoxic activity. We appreciate that the concentrations of inhibitor required are in the micromolar range and, although relatively high compared with other metalloproteinase inhibitors, the mode of action of EST is different from those inhibitors that directly block the active site of the enzymes.42 This is a lead compound representing a new class of molecule in this field. It is currently undergoing further development and more potent derivatives will be available in the future. The bioavailability of compounds in vivo may also alter their efficacy, such that lower concentrations of more potent agents may be sufficient to block cartilage resorption in patients.
Thus we have shown that EST is an effective inhibitor of cartilage resorption. It significantly reduces proteoglycan and collagen loss, and collagenolytic activity in resorbing cartilage. These effects are not due to changes in the levels of TIMP-1, but can be seen at the mRNA level for MMP-1, MMP-3, and MMP-13. It therefore appears that EST exerts its effects at a pre-transcriptional stage, thereby reducing enzyme levels. The mechanism is new, in that most MMP inhibitors function by direct inhibition of active MMPs or by increasing TIMP levels,33 35 39 42 whereas EST does neither of these. We are currently investigating the precise mechanisms of action of EST.
SE was funded by Kureha Chemical Industry Co Ltd, Tokyo, Japan. ADR, SC, PK, and JBC were funded by the Arthritis Research Campaign, UK. We are indebted to John Matthews, professor of statistics, Newcastle University, UK, for assistance in the statistical analysis of data. We are grateful to Professor Constance Brinckerhoff and Joni Rutter, Dartmouth College Medical School, Hanover, New Hampshire, USA for the MMP-1 promoter construct.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.