Human osteoblasts in culture synthesize collagenase and other matrix metalloproteinases in response to osteotropic hormones and cytokines

Journal of Cell Science, Dec 1992

M.C. Meikle, S. Bord, R.M. Hembry, J. Compston, P.I. Croucher, J.J. Reynolds

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

http://jcs.biologists.org/content/103/4/1093.full.pdf

Human osteoblasts in culture synthesize collagenase and other matrix metalloproteinases in response to osteotropic hormones and cytokines

MURRAY C. MEIKLE 0 2 SHARYN BORD 2 ROSALIND M. HEMBRY 2 JULIET COMPSTON 1 PETER I. CROUCHER 1 JOHN J. REYNOLDS 2 0 Department of Orthodontics, Institute of Dental Surgery, University of London , 256 Gray's Inn Road, London WC1X 8LD , UK 1 Department of Medicine, University of Cambridge, Addenbrooke's Hospital , Cambridge CB2 2QQ , UK 2 Cell and Molecular Biology Department, Strangeways Research Laboratory , Worts Causeway, Cambridge CB1 4RN , UK - Human osteoblasts in culture synthesize collagenase and other matrix Collagenase production by rodent osteoblasts in response to calciotropic hormones has led to the hypothesis that bone cells play a major role in bone resorption by degrading the surface osteoid layer, thereby exposing the underlying mineralized matrix to osteoclastic action. Many studies suggest, however, that this model might not apply to bone resorption in the human. Human osteoblasts have been shown to produce gelatinase-A (72 kDa) and TIMP-1 (tissue inhibitor of metalloproteinases), but previous investigators have been unable to demonstrate the synthesis of collagenase by human osteoblasts either constitutively or in response to bone resorptive agents. In the present study the ability of human osteoblasts to produce the matrix metalloproteinases (MMPs) collagenase, gelatinase and stromelysin, and their specific inhibitors TIMPs-1 and 2, was examined using highly sensitive and specific antisera and by zymography. Semi-quantitative histomorphometric data showed that cells cultured on either glass or a type I collagen substratum constitutively synthesized gelatinase-A and TIMP-1. On type I collagen, however, a small proportion of unstimulated cells produce both collagenase (7%) and gelatinase-B (95 kDa; 3%). Treatment of cells with Over the past decade in vitro studies have provided much information regarding the participation of collagenase and other matrix metalloproteinases (MMPs), such as gelatinase and stromelysin, in bone resorption. It has long been recognized that collagenase production by bone explants cultured with calciotropic hormones is correlated with bone resorption (Walker et al., 1964; Fullmer and Lazarus, 1967; Vaes, 1972; Sellers et al., 1980) but the role of the enzyme in the resorptive process and its cellular source have been controversial. Later works from our own and other laboratories, however, have established osteoblasts as the cell of origin (Heath et al., 1984; Sakamoto and Sakamoto, 1984; Otsuka et al., 1984; Partridge et al., 1987) and led to the either parathyroid hormone (PTH), 1,25-dihydroxyvitamin D3 (1,25(OH)2D3), or partially purified mononuclear cell conditioned medium (MCM), stimulated the synthesis of collagenase, gelatinase-B and stromelysin; MCM was 2- to 3-fold more potent than either PTH or 1,25(OH)2D3. Zymography using SDS/PAGE on conditioned media from cells cultured on type I collagen films revealed the presence of active gelatinase-A and that MCM stimulated progelatinase-B synthesis. Inhibitor gel zymography confirmed that TIMP-1 was constitutively synthesized and that its expression was increased by MCM treatment; TIMP-2 could not be detected. These data confirm the findings of previous investigators that human osteoblasts constitutively produce gelatinase-A and TIMP-1, and show for the first time that they can also be stimulated to synthesize collagenase, gelatinase-B and stromelysin. The working hypothesis that osteoblast MMPs play an important role in bone resorption would therefore appear to be as valid for human as it is for rodent bone. hypothesis that osteoblasts participate in bone resorption by degrading the surface osteoid layer, thereby exposing the underlying mineralized matrix to osteoclastic action (for review see Sakamoto and Sakamoto, 1986). A role for osteoblast-derived collagenase in resorption is further supported by the observation that isolated osteoclasts cultured on calvarial explants depleted of their surface cell layer do not resorb the mineralized matrix unless the osteoid layer is first removed, either by pre-treating the surface with osteoblasts or by including collagenase in the incubation medium (Chambers et al., 1985; Chambers and Fuller, 1985). Recent findings suggest that osteoid removal by osteoblasts is regulated in a complex manner which involves the stimulation of procollagenase, followed by activation of the enzyme extracellularly by proteolytic cleavage requiring plasmin (Thomson et al., 1987; Thomson et al., 1989). All the above mentioned studies involved the use of cells derived from rodent bone. Evidence exists, however, suggesting that this model might not apply to osteoid removal in the human. Human osteoblasts produce gelatinase and TIMP (tissue inhibitor of metalloproteinases) as major cellular products, but previous investigators have been unable to demonstrate the production of collagenase, either constitutively or in response to bone resorptive agents (Gowen et al., 1984; Rifas et al., 1989; Hankey et al., 1991). Since it seemed unlikely that unique cellular and/or molecular mechanisms would have evolved to resorb human bone, we have recently examined the ability of human osteoblasts to produce MMPs using highly sensitive and specific antisera. Our findings show that human osteoblasts in vitro constitutively produce gelatinase-A (72 kDa) and TIMP-1 and provide new evidence that under certain culture conditions they can be stimulated to synthesize collagenase, gelatinaseB (95 kDa) and stromelysin. Materials and methods Isolation of osteoblasts from human bone Samples of human bone were obtained (I) two from road traffic accident victims aged 21 and 47, and (II) core samples of trabecular bone removed from the femoral shaft of 5 patients following hip joint replacement surgery (age range 52-79 years) and transferred to the laboratory in dissecting medium. This is a modified form of BGJ medium containing a reduced bicarbonate level (50 mg/100ml) so as to equilibrate with air, and 10 times the normal concentration of antibiotics (Reynolds, 1976). Bone samples were cleaned of adherent soft tissue and osteoblasts isolated by two methods. (i) The bone was cut into small pieces (2 mm 2 mm), washed in Ca2+- and Mg2+-free Tyrodes solution (10 min), and then sequentially digested (1 mg/ml trypsin, 10 min; 2 mg/ml Dispase, 20 min; and 3 mg/ml collagenase, 2 30 min) in Tyrodes solution at 37C. Cells released by the collagenase digestions were washed and grown to confluence in 25 cm culture flasks (Falcon) in 1:1 (v/v) Hams F12/Dulbeccos modification of Eagles medium (DMEM) supplemented with 10% fetal calf serum (FCS, Globefarm, Esher, Surrey). Incubations were carried out at 37C in an humidified atmosphere of 5% CO2/ 95% air; the medium was changed every 2-3 days. (ii) Bone samples were cut into small pieces and plated into 5 cm Petri dishes containing 4 ml Hams F12/DMEM supplemented with 10% FCS. Bone cells were allowed to grow out from the explants until confluent; the medium was changed every 2-3 days. Characterization of osteoblasts Assay for alkaline phosphatase First passage cells were cultured for 48 h in Hams F12/DMEM plus 10% FCS on coverslips in 3 cm dishes and alkaline phosphatase (ALP) positivity assessed by a semiquantitative cytochemical stain (Sigma; procedure no. 86). The cells were fixed for 30 s in citrate-acetone-formaldehyde (25 ml citrate solution, Sigma 91-5; 65 ml acetone; 8 ml 37% formaldehyde) and then incubated for 15 min in a mixture of napthol AS-B1 alkaline solution and fast red violet LB salts according to the manufacturers instruction; deposits of red pigment reflect the presence of ALP. L-levamisole (4 mM, Sigma), an alkaline phosphatase inhibitor, was added to some cultures. Assay for osteocalcin (bone -carboxyglutamic acid) Cells were cultured in 8-well Labtek slides in Hams F12/DMEM supplemented with 10% FCS, ascorbic acid (50 g/ml), 10-8M Menadione (Vitamin K, Sigma), and stimulated with 1,25-dihydroxyvitamin D3 (1,25(OH)2D3; 10 ng/ml) for 48 h. At the end of the culture period the medium was removed and the cells fixed in 4% paraformaldehyde (5 min) at room temperature. Osteocalcin in culture supernatants was measured using an enzyme immunoassay (EIA) kit (Takara Shuzo Ltd., Kyoto, Japan). Osteocalcin was also immunolocalized by indirect immunofluorescence using a goat polyclonal antibody to human osteocalcin (Biogenesis, Bournemouth, Hampshire; 1:50 dilution in PBS) followed by an anti-goat IgG fluorescein isothiocyanate (FITC) conjugate (Sigma, 1:64 dilution in phosphate buffered saline; PBS). Normal goat serum was used as a control for the immunofluorescence procedure. Human dermal fibroblasts were stained as an osteocalcin negative cell type. Cells were grown to confluence in 3 cm Petri dishes containing 3 ml Hams F12/DMEM supplemented with 10% FCS and ascorbic acid (150 g/ml). Incubations were carried out at 37C in an humidified atmosphere of 5% CO2/95% air for 5 days. The cell layer was pepsin digested (Sigma, 1 mg/ml, 18 h at 4C) and analyzed by electrophoresis on a 5% sodium dodecyl sulphate (SDS)/polyacrylamide gel using non-reduced, reduced and interrupted reduction techniques and silver stained (Sykes et al., 1976). Silver staining bands were compared to native standards of types I, II and III collagen, run under the same conditions. Type I is identifiable by its a 1 and a 2 chains at a 2:1 density, respectively; type II by its single band appearing at a similar position to a 1 (I) chain; type III under non-reduced conditions will not migrate into the resolving gel but upon reduction the bands are cleaved and the released a 1(III) migrate more slowly than a 1(I) chains. Assay for cyclic AMP The cyclic adenosine 3 ,5 -monophosphate (cAMP) response to parathyroid hormone (PTH) was determined from osteoblast monolayers cultured in 24-well plates (Linbro). Cells were cultured in Hams F12/DMEM plus 10% FCS at 37C in an humidified atmosphere of 5% CO2/95% air. After 5 days the cells were washed and maintained in serum-free F12/DMEM for 25 min and then incubated with 100 M 3-isobutyl 1-methylxanthine (IBMX) for 5 min. PTH was added (0.1 unit/ml; 10 l) and incubations continued for a further 6 min. Extraction and determination of cAMP was performed as described in detail elsewhere (Farndale and Murray, 1985). MMPs and TIMP were immunolocalized in osteoblasts by indirect immunofluorescence following two different culture procedures. (i) First passage cells were plated onto 8-well Labtek slides containing Hams F12/DMEM plus 10% FCS as described above. After 48 h the medium was removed, the cells washed with serumfree DMEM and then replaced with 1:1 Hams F12/DMEM plus 2% acid-treated rabbit serum; this contains no detectable proteinase inhibitors or a 2-macroglobulin. Cells were stimulated with either PTH (2 units/ml), 1,25(OH)2D3 (10 ng/ml) or mononuclear cell conditioned medium (MCM, 5%, v/v). The sodium ionophore monensin (Sigma, 5 M), which inhibits the translocation and secretion of newly synthesized proteins, was added for the final 3 h of culture. (ii) First passage cells were cultured as above in 8-well Labtek slides coated with 200 l of type I rat skin collagen (0.5 mg/ml). Monensin was added for the final 18 h of culture. After removal of culture medium the cell layer was fixed for 5 min in 4% paraformaldehyde at room temperature. Cells were permeabilized (0.1% Triton-X100, 5 min) to enable IgG penetration, washed with PBS and incubated with specific polyclonal antibodies to either collagenase (Hembry et al., 1986), gelatinase-A (Hipps et al., 1991), gelatinase-B (Murphy et al., 1989), stromelysin (Allan et al., 1991), TIMP-1 (Hembry et al., 1985), or normal sheep serum (IgGs, 50 g/ml in PBS for 30 min at room temperature). The characterization of these antisera including species specificity, Western blots, inhibition curves and immunoabsorption experiments with purified antigen are detailed in the references. The cells were washed (PBS, 3 5 min) and the second antibody (a pig Fab preparation labelled with FITC; Hembry et al. 1985) was applied for 30 min. After exhaustive washing they were coverslipped with Citifluor (City University, London) and observed by fluorescence microscopy on a Zeiss photomicroscope III with epifluorescence and standard FITC filters. Photographs were taken on Agfachrome RS 1000 film uprated to 2000 ASA. To obtain semi-quantitative histomorphometric data on the number of cells reacting positively for MMPs and TIMP following various treatments, fluorescent cells were counted. When the Golgi complex was visible within the cell it was scored as positive, irrespective of brightness; the number of positive cells was expressed as a percentage of the total cell number counted over five randomly selected fields. For each determination at least 120 cells per well were counted. Gel electrophoresis and zymography First passage cells were plated at a concentration of 105/ml into 24-well plates (Linbro) containing Hams F12/DMEM supplemented with 10% FCS; some wells were coated with 300 l of type I rat skin collagen (0.5 mg/ml). At confluence the cells were washed in serum-free medium and cultured for 72 h in Hams F12/DMEM supplemented with 2% acid-treated rabbit serum with either PTH (2 units/ml), 1,25(OH) 2D3 (10 ng/ml), or MCM (5%, v/v). Gelatin-degrading activity was assessed by electrophoresis on non-reducing SDS-8% polyacrylamide gels (Laemmli and Favre, 1973) containing 0.5 mg/ml gelatin (Heussen and Dowdle, 1980). After electrophoresis the SDS was removed by washing (2 15 min with 2.5% Triton-X100, at room temperature). Degradation of the gel was visualized by overnight incubation at room temperature in TCB (buffer containing 100 mM Tris-HCl, pH 7.9, 30 mM CaCl2, 0.02% sodium azide). Gels were stained with Coomassie blue (30 min), destained and photographed. TIMP activity was detected by inhibitor gel zymography on SDS-10% polyacrylamide gels, using the above method plus an incubation step for 1 h at 37C in a preparation of active rabbit skin MMPs (7 units/ml) prior to the final overnight incubation in TCB buffer at 25C. Molecular mass standards and TIMP-1 and -2 markers were run on the appropriate gels. Bovine PTH (1-84) was supplied by the Division of Biological Standards, National Institute of Medical Research, Mill Hill. 1,25(OH)2D3 was a generous gift from Dr Ian Dickson, Brunel University. MCM from cultured pig leukocytes was partially purified on Ultrogel ACA-54 as described by Saklatvala et al. (1983). DMEM was purchased from Life Technology, Paisley, Scotland, and Hams F12 from Imperial Laboratories, Andover, Hampshire. Characterization of osteoblasts Bone cells were characterized by several criteria that collectively are regarded as being diagnostic for the osteoblast phenotype. Histochemical staining of unstimulated primary cell cultures for ALP was strongly positive; 79.3 8.6% of cells from six separate bone cell preparations exhibited positive staining. Human dermal fibroblasts in contrast showed less than 8% staining. EIA of culture supernatants from 1,25(OH)2D3-treated cells showed a 7-fold increase in osteocalcin levels (3.48 0.96 ng/ml) compared to untreated controls (0.42 0.33 ng/ml), and this was confirmed by immunocytochemical staining in which there was a 3-fold increase in the number of labelled cells over control values; human dermal fibroblasts contained no immunodetectable osteocalcin. SDS-PAGE indicated that the collagen type synthesized by the cells was predominantly type I together with small amounts of type III. In all cases the ratio of a 1(I) and a 2(I) chains was always 2:1 or less, indicating the unlikely presence of type II (data not shown). Finally, the intracellular accumulation of cAMP in response to PTH was determined; treatment with 0.1 unit/ml PTH for 6 min resulted in cAMP levels of 16.0 1.2 pmole/ml, compared with control levels of <0.125 pmole/ml. In some experiments, osteoblasts were prepared by both enzymic digestion and explant culture methods from the same bone sample for a direct comparison of their activities; we found them to be indistinguishable. Immunolocalization of MMPs and TIMP We cultured human osteoblasts on glass surfaces in multichamber slides and immunolocalized MMPs and TIMP using specific polyclonal antisera and the results of a typical experiment are shown in Table 1. In none of the individual bone cell preparations could collagenase, gelatinaseB or stromelysin be immunolocalized: only gelatinase-A and TIMP could be detected in unstimulated cells (Fig. 1A and B). Cells stained with normal sheep serum were negative (not shown). Treatment with PTH for 24 h resulted in a modest increase in the number of collagenase-positive cells (12% of total). Addition of MCM had a dramatic effect, stimulating the number of cells producing collagenase (35%) and stromelysin (37%) as well as increasing the First passage osteoblasts were plated into 8-chamber Labtek slides and cultured in Hams F12/DMEM plus 10% FCS. After 48 h the cells were washed in serum-free DMEM and cultured in Hams F12/DMEM plus 2% acid-treated rabbit serum. Cells were stimulated with either PTH (2 units/ml) or MCM (5%, v/v) for 24 h; monensin (5 M) was added for the final 3 h of culture. MMPs and TIMP were immunolocalized by a 2-stage indirect technique using specific polyclonal antisera. Data are based on cell counts from 5 randomly selected fields; positive cells are expressed as a percentage of the total cells counted in each field. Data are taken from 3 different bone preparations. Variations between individual preparations amount to approximately 5%. Fig. 1. Immunolocalization of gelatinase-A and TIMP-1 in osteoblast monolayers cultured on glass surfaces. Osteoblast monolayers were cultured for 48 h, with monensin (5 M) added for the final 18 h. Cells were fixed, permeabilized, stained by indirect immunofluorescence and observed by fluorescence microscopy. Bars, 10 m. (A) Cells stained with anti-gelatinase-A IgG. Fluorescence of vesiculated juxtanuclear Golgi apparatus indicates gelatinase-A synthesis. (B) Cells stained with antiserum IgG to TIMP-1. Intracellular immunofluorescence demonstrates TIMP-1 synthesis. number reacting positively for gelatinase-A and TIMP. Neither PTH nor MCM had any effect on the synthesis of gelatinase-B. When osteoblasts were cultured on type I collagen films, a small proportion of unstimulated cells showed evidence of both collagenase (7% of total) and gelatinase-B (3%) (Table 2). There was also a slight increase in the number of gelatinase-A and TIMP-positive cells compared to osteoblasts cultured on glass, but unstimulated cells continued to react negatively for stromelysin. When treated with either PTH, 1,25(OH)2D3 or MCM, the synthesis of collagenase and gelatinase-B was stimulated and stromelysin was induced (Table 2), MCM being 2- to 3fold more potent than either PTH or 1,25(OH)2D3 (Fig.2). However, because basal levels of gelatinase-A and TIMP production were already high (70-80%) in unstimulated cells, no clear differences between treated and untreated cells could be detected. These results are typical of the dif First passage osteoblasts were plated into 8-chamber Labtek slides coated with type I rat skin collagen (0.5 mg/ml) and cultured in Hams F12/DMEM plus 10% FCS. After 48 h the cells were washed in serumfree DMEM and cultured in Hams F12/DMEM plus 2% acid-treated rabbit serum. Cells were stimulated with either PTH (2 units/ml), MCM (5%, v/v) or 1, 25(OH)2D3(10 ng/ml) for 24 h; monensin (5 M) was added for the final 18 h of culture. MMPs and TIMP were immunolocalized by a 2-stage indirect technique using specific polyclonal antisera. Data are based on cell counts from 5 randomly selected fields; positive cells are expressed as a percentage of the total cells counted in each field. Data are taken from 3 different bone preparations. Variations between individual preparations amount to approximately 5%. ferent bone cell preparations as a whole, although individual variability in the number of positive cells following stimulation was observed. Zymography To confirm that culture media from human osteoblasts contained MMP activities in addition to collagenase, culture supernatants were electrophoresed on SDS-polyacrylamide gels containing gelatin (Fig. 3). There was clear evidence for the presence of latent gelatinase-A in both unstimulated and stimulated culture media. Active gelatinase-A was barely detectable in culture media from cells cultured on plastic alone but in significant quantities in media from cells cultured on type I collagen films. Progelatinase-B was also evident in media from osteoblasts cultured on type I collagen films and stimulated with MCM. Finally, TIMP-inhibitory activity was detected in osteoblast culture supernatants using inhibitor gel zymography under non-reducing conditions. This showed that TIMP-1 is constitutively synthesized by osteoblasts irrespective of the substrate and that its expression is increased by MCM treatment (Fig. 4). There was no evidence for significant TIMP-2 synthesis. Using sensitive immunocytochemical techniques we have shown for the first time that human osteoblasts cultured under certain conditions are capable of producing collagenase and stromelysin. These findings are contrary to previous reports (Gowen et al., 1984; Rifas et al., 1989; Hankey et al., 1991): in our view there are a number of reasons for this apparent contradiction. Gowen et al. (1984) stimulated human osteoblasts with partially purified human interleukin-1 (IL-1) and measured collagenase and proteoglycanase (stromelysin) in the culture medium. At that time the only methods available for measuring MMPs and TIMP in culture supernatants were activity assays. The disadvantage with such assays is that low levels of MMPs are likely to remain undetected in Fig. 2. Collagenase is synthesized by stimulated human osteoblasts cultured on collagen films. Human osteoblasts were cultured on type I collagen films with 5% MCM for 48 h, and monensin (5 M) added for the final 18 h. Films were fixed, permeabilized and stained by indirect immunofluorescence, then viewed by fluorescence microscopy. Bars, 10 m. (A) Cells stained with antibody to human collagenase. Intracellular fluorescence of the vesiculated Golgi apparatus in the majority of cells indicates synthesis of collagenase. (B) Cells stained with normal sheep serum. No intracellular fluorescence is visible. supernatants containing relatively high levels of TIMP. TIMP does not interact with latent enzyme, but once the enzyme is activated by including APMA or trypsin in the assay, it is avidly bound by free TIMP yielding irreversible MMP-TIMP complexes which cannot be detected in activity assays (Cawston et al., 1983). TIMP production will be similarly underestimated in supernatants where the level of naturally activated enzyme exceeds that of TIMP. Such assays are therefore likely to contain a high proportion of false negatives. In the study of Rifas et al. (1989) cultured human osteoblasts were exposed to a wide variety of agents including phorbol 12-myristate 13-acetate (PMA), PTH, prostaglandin E2, 1,25(OH)2D3 and recombinant human IL1b . Collagenase and TIMP in culture supernatants were measured by enzyme-linked immunosorbant assays (ELISAs) as well as functional assays. They found that collagenase was not secreted in significant quantities under any test condition, and that collagenase mRNA could not be detected in the cells. In contrast TIMP and gelatinase-A were constitutively secreted in abundant quantities. Our results are in agreement with these findings as far as TIMP and gelatinase-A synthesis are concerned. However, we found that critical determinants of collagenase, stromelysin and gelatinase-B production by human osteoblasts were (i) the presence of a type I collagen substratum, and (ii) the particular osteotropic agent(s) used to stimulate the cells. The incorporation of monensin into the culture medium to Fig. 3. Detection of gelatinase-A (72 kDa) and gelatinase-B (95 kDa) by zymography. Samples of culture media from unstimulated and stimulated human osteoblasts grown on glass (lanes a-d) and type I collagen films (lanes e-h) were electrophoresed under non-reducing conditions on an SDS/8% polyacrylamide gel containing gelatin. Positions of molecular mass standards are indicated. Under non-reducing conditions the 72 kDa and 95 kDa gelatinases run, respectively, at the slightly lower molecular masses of 66 kDa and 92 kDa. The samples were (lanes a, e) unstimulated; (lanes b, f) MCM; (lanes c, g) PTH; (lanes d, h) 1, 25(OH)2D3. Arrow denotes active form of gelatinase A. Fig. 4. Detection of TIMP-1 inhibitory activity. TIMP-1 was detected in conditioned media from unstimulated and stimulated human osteoblasts cultured on glass (lanes a-d) and type I collagen films (lanes e-h) by inhibitor gel zymography under nonreducing conditions. Positions of molecular mass standards and TIMP-1 and TIMP-2 standards are indicated. The samples were (lanes a, e) unstimulated; (lanes b, f) MCM; (lanes c, g) PTH; (lanes g, h) 1, 25(OH)2D3. enhance the signal was also essential for good immunocytochemistry. It is not surprising that cells grown on tissue culture plastic or glass surfaces behave differently from those on synthetic substrata such as type I collagen (Gavrilovic et al., 1985) or basement membrane complexes (Kleinman et al., 1986). Structural macromolecules such as collagens, proteoglycans and glycoproteins contribute to the physical characteristics of tissues, as well as providing unique substrata for the attachment, growth and differentiation of cells. Cultured cells function best on their natural substratum or a close approximation to it. Our choice of a collagen substratum was based on previous work, which had shown type I collagen films to be an ideal substratum on which to culture murine calvarial osteoblasts (Thomson et al., 1987, 1989; Meikle et al., 1991), particularly for studies of matrix degradation. While we acknowledge that bone matrix contains many non-collagenous proteins including polypeptide growth factors which individually or synergistically can modulate osteoblast behaviour, to date we have been unable to create a suitable synthetic bone matrix substratum for routine biochemical or immunological assays. Both Gowen et al. (1984) and Rifas et al. (1989) were unable to stimulate collagenase production by human osteoblasts with either partially-purified human IL-1 or recombinant human IL-1b . For that reason, plus the fact that target cells are likely to be exposed to mixtures rather than individual cytokines in vivo, particularly during inflammation, we chose to stimulate the cells with partially purified MCM in addition to the classical osteotropic hormones PTH and 1,25(OH)2D3. MCM contains many cytokines including IL-1, tumour necrosis factor (TNF-a ), lymphotoxin (TNF- b ) and interferon-g , and was 2- to 3-fold more potent than either PTH or 1,25(HO)2D3 in stimulating collagenase synthesis by the cells. Another difference between studies is the origin of the osteoblast cultures used in the various studies. Gowen et al. (1984) cultured cells from femoral/tibial condyles of knee joints after surgical amputation, from femoral heads following femoral neck fracture, or from children undergoing corrective surgery. Rifas et al. (1989) analysed five newborn and seven adult bone cell cultures derived predominantly from rib and vertebral samples. In the present investigation osteoblasts were prepared from normal bone from two road traffic accident victims and from trabecular bone removed from the femoral shaft of 5 patients during hip joint replacement surgery (in which in each case a clinical diagnosis of osteoarthritis had been made). Since the most readily available source of human bone is corrective surgery of various kinds, the extent to which osteoblasts derived from such bone can be regarded as completely normal is uncertain. Moreover, where multiple donors are used, individual qualitative and quantitative differences in the cellular response in vitro can be expected which will reflect both the age of the donor and any pre-existing pathology or treatment. In conclusion, the present study has demonstrated that in addition to producing gelatinase-A and TIMP-1 constitutively, human osteoblasts in culture can also synthesize collagenase, gelatinase-B and stromelysin in response to boneresorbing agents. The working hypothesis that osteoblasts and MMPs play an important role in bone resorption by degrading the non-mineralized osteoid layer would therefore appear to be as valid for human as it is for rodent bone. This study was supported by grants from the Arthritis and Rheumatism Council and the Medical Research Council. We thank Christopher Green and Judith Webdell for preparing the illustrations, and Angela Hedge for secretarial assistance. We also thank the Consultant Orthopaedic Surgeons at Addenbrookes Hospital, Cambridge: Messrs Constant, Matthewson and Meggitt for their invaluable cooperation in providing samples of fresh human bone.


This is a preview of a remote PDF: http://jcs.biologists.org/content/103/4/1093.full.pdf

M.C. Meikle, S. Bord, R.M. Hembry, J. Compston, P.I. Croucher, J.J. Reynolds. Human osteoblasts in culture synthesize collagenase and other matrix metalloproteinases in response to osteotropic hormones and cytokines, Journal of Cell Science, 1992, 1093-1099,