BoneandMinerul.
12(1991)41-S
41
Elsevier BAM 00335
one-derived growth factors modulate collagenase and TIMP (tissue inhibitor of metalloproteinases) activity and type I collagen degradation by mouse calvarial osteoblasts
Murray C. Meikle’q’ , Anne M. McGarrity’, Brian M. Thomson’ and John J. Reynolds’ ‘Cell Physiology Department. Strangeways Research Laboratory, CBI 4RN and ‘Comective
Worrs Causeunay, Ca,nbridge
Tissue Research Unit, Department of Orthodontics. Ittstituteof
Dental Surgery. University of London, 36
Gray’s Inn Road, London WCIX 8LD. Etq$and
(Received 20 April 1990) (Accepted 17 September 1990)
Summary The finding that osteoblasts synthesize collagenase has led to the hypothesis that bone cells play a major role in bone resorption by degrading the surface osteoid layer, thereby preparing the underlying mineralized bone for osteoclastic action. To further understand the mechanisms regulating osteoid removal, mouse calvarial osteoblasts were cultured on ‘JC-labelled type I ccllagen films and the abilities of(i) bovine bone matrix extracts and (ii)purified or recombinant human growth factors. to modify their collagenolytic behaviour were investigated. EDTAITris-HCI extracts of bone matrix containing growth factor activity. exerted a dose-dependent inhibition of type I collagenolysis by osteoblasts stimulated with !.25dihydroxyvitamin D, (I .25(OH),D,, 10 nglml). Inhibition was accompanied by a reduction in collagenase activity and an increase in free TIMP (tissue inhibitor of metalloproteinases) in the culture medium. Transforming growth factor-@. epidermal growth factor, platelet-derived growth factor and the acidic and basic fibroblast growth factors all mimicked these effects. In contrast, insulin-like growth factors-I and -II did not inhibit type I collsgenolysis, only partially inhibited collagenase activity. and did not stimulate TIMP production by either 1.25(OH)zD,-treated or untreated cells. These findings provide additional evidence for the tight control exerted on the proteolytic activity of osteoblasts and the importance of TIMP in its regulation. They suggest strongly that the conversion (coupling) of the initial resorptive phase of the bone remodellinp cycle to one of deposition, may be mediated by polypeptide growth factors either produced iocally by osteoblasts. or released by proteolysis from the bone matrix.
Key words: Growth factor: Osteoblast:
Collagenase:
TIMP: Type I collagenolysis
Correspondence to: Dr M.C. Meikle, Strangeways Research Laboratory, CB14RN, England, U.K. 0169-6009/91/$03.50@
Worts Causeway. Camhridge
1991 Elsevier Science Publishers B.V. (Biomedical Division)
42 Introduction The maintenance of skeletal integrity is dependent upon the continuous remodelling of bone, a co-ordinated sequence of cell mediated events involving an initial period of osteoclastic resorption followed by the refilling of each resorption cavity with new bone by osteoblasts: this process is referred to as coupling [1,2]. Since bone remodelling occurs at anatomically discrete sites throughout the skeleton, it seems likely that it is regulated by factors either produced locally by bone cells or released from the bone matrix during the initial resorptive phase. Non-resorbing bone surfaces are covered by a layer of nonmineralized osteoid [3-S], and recent evidence suggests that one way osteoblasts might regulate bone resorption is through mineral exposure (for review see Ref. 6). Osteoblasts stimulated with bone resorbing agents respond by synthesizing the matrix metalloproteinase (MMP) collagenase [7-91, and isolated osteoclasts in vitro do not resorb bone unless the surface osteoid layer is first removed, either by pre-treatment with collagenase or by a layer of osteoblasts [ 10,111. To understand further the mechanisms regulating osteoid removal by osteoblasts, we have used a model system in which mouse calvarial osteoblasts are grown on type I collagen films, and previously provided evidence for the involvement of a plasminogen-plasmin-MMP activation cascade in type I collagen degradation by osteoblasts, and for its regulation by TIMP (tissue inhibitor of metalloproteinases) and plasminogen activator inhibitor-l (PA&l [12,13]). Type i collagenolysis in this in vitro model, however, is not synonymous with osteoid removal in vivo. Although 90% of total bone protein is type I collagen, bone matrix also contains many noncollagenous proteins including polypeptide growth factors [ 14- 171, which individually or synergistically could modulate osteoblast responsiveness to bone resorbing agents; recent research from several groups has suggested an important role for growth factors and hormones in the regulation of MMP and TIMP synthesis by osteoblasts and the maintenance of skeletal homeostasis [18-211. In the present series of experiments therefore, we tested the abilities of (i) bovine bone matrix extracts and (ii) purified or recombinant human growth factors, to modify the collagenolytic behaviour of mouse calvarial osteoblasts stimulated by 1,25(OH),D,,
Materials and Methods Reagents
1,25(OH),D, was generously supplied by Dr Ian Dickson, Department of Medicine, University of Cambridge. Purified human platelet derived growth factor (PDGF), acidic and basic fibroblast growth factors (aFGF, bFGF) and epidermal growth factor (EGF) were purchased from Biogenesis Ltd, Boumemouth, England. Purified human transforming growth factor-p (TGF-P) was the generous gift of Dr John K. Heath, Department of Biochemistry, University of Oxford, and also purchased from R & D Systems, Minneapolis, Minnesota. Human recombinant insulin-like growth factor-I (IGF-I) was a gift from Dr John T. Dingle, Strangeways Re-
43 search Laboratory, Cambridge; human recombinant insulin-like growth factor-II (IGF-II) was purchased from Bachem Inc., Torrance, CA. Preparation of bone matrix extract
Bovine tarsal bones were cleaned of adherent soft tissue, the epiphyses and marrow removed and cut into 1 cm cylinders. These were quartered, cooled in liquid N, and fragmented prior to being reduced to a fine powder in a Spex freezer mill (Glen Creston Instruments, Stanmore, Middlesex). Bone matrix extract was prepared as described previously [ 143. Briefly, dry bone powder was suspended in 0.02 M TrisHCl, pH 7.5, at 4 “C for 20 min and then centrifuged at 2000 x g for 10 min and the supernatant discarded. The residue was resuspended in 0.5 M ammonium EDTA, 0.02 M Tris-HCl, pH 7.5, with gentle stirring for 24 h at 4 “C. The supernatant was centrifuged (10 000 x g) and filtered. Dialysis at 4 “C in 70 mm dialysis tubing against water for 4 days and 0.02 M NH,HCOs for 2 days with frequent changes was followed by lyophilization and storage at -20 “C. For experimental purposes bone matrix extract was redissolved in 0.1 M NaCl, 10 mM Tris, pH 7.0, dialysed against this buffer and then centrifuged at 2000 x g for 10 min. The protein content of bone matrix extract was estimated at 9.8 mg protein/g using a protein assay kit (Bio-Rad, Watford, Hertfordshire). Growth factor assay
Growth factor activity in bovine bone matrix extract was assayed by its ability to stimulate the proliferation of BALB/c/3T3 cells [22]. BALBlcI3T3 cells were cultured in Dulbecco’s modification of Eagle’s medium (DMEM) supplemented with antibiotics and 10% fetal calf serum (FCS; Globefarm, Esher, Surrey). To test for growth factor activity, cells were plated into 96-well microtitre plates (Cel-Cult; Sterilin, Feltham, Middlesex) at a density of lo4 cells/well in 200 ~1 DMEM plus 10% FCS, and grown to confluence at 37 “C in a humidified atmosphere of 5% CO,/95% air. After a brief wash in DMEM the cells were challenged with serial dilutions of bovine bone marrow extract in 250 ,~l DMEM. On each plate 20 wells were set up with dilutions ranging between 2 and 30 ~1 FCS/250 ~1 DMEM to provide a serum-stimulation curve. One Growth Factor Unit (GFU) was defined as half the maximal stimulation of proliferation in response to FCS [ 141. BALBlcl3T3 cells were incubated with the test factors for 42 h, after which the medium was changed and replaced with 200 ~1 DMEM containing 1 &X/ml [3H]thymidine (Amersham International, Amersham, Buckinghamshire) and incubated for a further 6 h. Acid-insoluble [3H]thymidine in cell DNA was measured ;asfollows: after removing the medium, the cells were washed in DMEM, fixed with 200~15% icecold trichloroacetic acid (TCA) for 20 min, and washed in phosphate-buffered saline (PBS). Cell layers were then dissolved in lOO@ 0.1 M NaOH for 10 min and removed into a scintillation vial; 400~1 of 0.5 M acetic acid and 2 ml scintillant were added to each sample before counting for [3H]thymidine incorporation. Assay for TGF-@
TGF-/3 activity in bone matrix extract was assayed by the ability of TGF-@ in the
44
presence of EGF. to stimulate anchorage-independent growth of normal rat kidney fibroblasts (NRK, clone 49F) in soft agar [23]. A stock solution of 1% purified agar (Oxoid) was prepared in distilled water and sterilized by autoclaving. DMEM was prepared at double strength and both solutions maintained at a temperature of 47 “C. An underlay containing 0.5% agar and 10% FCS in DMEM was prepared from these concentrated solutions and 1 ml pipetted into each well of a 24-well plate (Linbro) and solidified at room temperature. A single cell suspension of NRK-49F cells was prepared in DMEM plus 10% FCS. Once the underlay had solidified, 200 3 X lo4 cells, 10% FCS, 0.25% agar and test factors, in the ,~l DMEM containing 1.A absence or presence of EGF (50 nglml), were pipetted over the underlay in each well and the cells cultured at 37 “C for 7-14 days without further addition. Transformation was assessed in unfixed and unstained cultures by light microscopy. Colonies counted in four random fields (2.27 mm”) in each of three replicate wells were categorized as small (~50 pm in diameter) or large (>50-200 pm in diameter) as previously described [IS]. Purified human TGF-@ (10 nglml) was used as a positive control. Preparation of osteoblastsfrom neonatal mouse calvariae
Calvarial osteoblasts were prepared as previously described [7]. The cells were characterized as osteoblast-like by morphological criteria, and by demonstrating that essentially all the cells stained strongly for alkaline phosphatase, synthesized type I collagen (no type II or III), and accumulated CAMP in response to parathyroid hormone and prostaglandin E, but not to calcitonin or 1,25(OH),D, [7]. Neonatal mouse calvariae (40-50) were dissected free from adherent soft tissue, washed in Modified Biggers’ Medium (Imperial) with 6 mM sodium bicarbonate and I .4 mM glutamine (10 min), and sequentially digested with 1 mg/ml trypsin (10 min), 2 mg/ml dispase (30 min) and 4 mg/ml collagenase (2 x 30 min) in Ca’+ and Mg’+ free Tyrode’s solution. Cells released by the collagenase digestions were washed, and grown to confluence in DMEM containing 10% FCS. Preparation of collagen films
Radiolabelled collagen films were prepared as described previously [24]. Briefly, a!iquots of ‘JC-acetylated collagen (rat skin type I; 150 pg in 300 ~1 of 10 mM phosphate buffer, PI-I 7.4, containing 300 mM NaCl and 0.02% sodium azide) were dispensed into tissue culture wells (Linbro, 16 mm diameter) and dried at 37 “C. The collagen was then washed twice with sterile distilled water and once with DMEM. Osteoblasts and collagen films
Osteoblasts (10’ well) were settled onto the collagen films and cultured in 1 ml DMEM plus 10% FCS. After 16 h the cells were washed with DMEM and cultured for 48 h (72 h in experiments described in Figs. 3 and 4) in 1 ml DMEM plus 2% acid treated rabbit serum plus either 1,25(OH),D, (in ethanol) or vehicle alone. These media were assayed for collagenase inhibitory activity (TIMPs plus cr,-macroglobulin) which was never more than 0.01 units/ml. The cultures were maintained at 37 “C in a humidified atmosphere of 5% CO,/95% air. At the end of the incubation
45 period, the culture medium was centrifuged (15 min, 12 000 x g) to remove any collagen fibrils, and the radioactivity released during collagen degradation quantified by liquid scintillation counting. Residual collagen was digested with bacterial collagenase (50 pug/ml) and assayed for radioactivity. Collagenolysis was expressed as a percentage of the total radioactivity released from the films f standard error of the mean (SEM). Effect of growth factors on type I collagenolysis
Mouse calvarial osteoblasts were cultured on i4C-labelled type I collagen films for 48 h in 16 mm wells containing 1 ml DMEM plus 2% acid treated rabbit serum and stimulated with 1,25(OH)zD, (10 ng/ml). Growth factors were added and collagenolysis determined as above; culture supernatants were assayed for collagenase and TIMP. The effects of growth factors on collagenase and TIMP synthesis by unstimulated osteoblasts were also tested. Assays for collagenase and TIMP
Collagenase activity in culture supernatants was measured by the release of labelled peptides from “C-acetylated rat skin collagen [25]. One unit of collagenase degrades 1 ,ug reconstituted fibrils per min at 35 “C. Latent collagenase was activated by including 4-aminophenylmercuric acetate (APMA) in the fibril assay (0.67 mM final concentration). TIMP was assa,Ied by incubating a standard amount (0.05 unit) of activated rabbit skin collagenase in the fibri! assay with samples of culture medium. One unit of TIMP is defined as the amount of TIMP required to inhibit 2 units of collagenase by 50%.
Results Assays of growth factors in bone matrix extract
We initially examined the effect of various concentrations of bone matrix extract (0.5-40 pg protein/ml) on proliferation and DNA synthesis by BALBlcI3T3 cells. A typical dose-response is shown in Fig. 1. By definition the maximal response possible in this bioassay is 2 GFU; 1 GFU was attained at a dose of approximately 1Opg protein/ml. Since bone matrix is a rich source of TGF-a. we also tested the ability of bone matrix extract to stimulate anchorage-independent growth of NRK-49F cells in soft agar in the presence or absence of EGF (50 @ml). TGF$ alone did not stimulate anchorage-independent growth, but in the presence of EGF resulted in a 3-fold increase in large colony formation over EGF alone (Table 1). At concentrations of 5-2Opg protein/ml, bone matrix extract did not induce anchorage-independent growth and did not enhance the effect of EGF on colony formation. At higher concentrations (40-50 pg protein/ml) however, bone matrix extract alone had a smaii but significant effect on colony formation, and in the presence of EGF stimulated anchorage-independent growth 3-fold and large colony formation 2-3-fold, thereby indicating the presence of both EGF and TGF-/3. Furthermore, although large colony formation by higher doses of bone matrix
46 3r r = 0.97 In C 5
0 2-
& 5 lf r %
l-
G
0
0. 0.1
I 10
1.0 3one
Matrix
Extract
I 100
(pg protein/ml)
Fig. 1. Growth factor activity of bone matrix extract on BALBlcI3T3 fibroblasts. Bone matrix extract was assayed on confluent BALBIcl3T3 cells in 96.well microtitre plates containing 250~1 DMEM. After 48 h the TCA-insoluble incorporation of [3H]thymidine into DNA was quantified as described in Methods. One growth factor unit (GFU) was defined as half the maximal serum-stimulated [3H]thymidine incorporation above background levels. Each point represents the mean of four cultures. The regression iine shown for doses of 1pg protein/ml and above had a correlation coefficient of 0.97.
Table 1 Anchorage-independent growth of NRL49F cells in soft agar in the presence of EGF, TGF-/I and bone matrix extract alone or in combination Addition Factor None TGF-/? Bone matrix extract
Concentration (nglml) 10 5000 10 000 20 000 40 OtM 50 000
% Transformed”
% Large coloniesb
-EGF
+EGF
-EGF
+EGF
0 0 0 0 0 0.1 0.1
1.0 0.7 0.4 1.2 1.3 3.0 3.6
0 0 0 0 0 0 0
6.0 17.6 5.0 6.9 7.9 21.5 14.7
NRK-49F cells (1.2 x l@/well) were &Wed in soft agar in the presence or absence of EGF (50 nglml) \ and the concentrations of TGF-/3 and bone.m;trix extract as indicated. Colonies were measured unfixed and unstained by light microscopy. Small colonies were defined as being ~50 pm in diameter and large colonies >50-200pm in diameter. Data are from four random fields (2.27 mm2) in each of three replicate cultures; each field had approximately 1200plated cells. a Total number of colonies/total number of cells plated. b Number of large colonies/total number of colonies.
..
Fig. 2. EGF-dependent colony formation by NRK-49F cells in soft agar. (a) No addition; (b) EGF (50 ng/ml); (c) TGF-/3 (10 @ml); (d)EGF (50 nglml) plus TGF-B (10 @ml); (e) bone matrix extract (5Opug protein/ml); (I) bone matrix extract (SOpg protein/ml) plus EGF (50 @ml). Bar = 100pm.
100
m ._
-
r = 0.92
80 .
lo h
Z
2
4 1,25(OH)2
6 -vitamin
6
10
12
D6 (ng/ml)
Fig. 3. Dose-responsive degradation of “C-labelled type I collagen films by mouse calvarial osteoblasts in response to 1,25(GH),D,. Cells were cultured for 72 h on type I collagen films in 16 mm wells containing DMEM plus 2% acid treated rabbit serum and the concentrations of 1,25(OH),D, as indicated. Culture supernatants and residual collagen were assayed for radioactivity; collagenolysis is expressed as radioactivity released from each film as a percentage of the total. The regression line had a correlation coefficient of 0.92.
48 extract in the presence of EGF was comparable to that of TGF-P, the percentage of transformation colonies was 4-fold higher (Table 1, Fig. 2). Collagenolysis and the effects of bone matrix extract
The optimal dose of 1,25(OH),D3 on type I collagenolysis by mouse calvarial osteoblasts was found to be approximately 10 ng/ml (Fig. 3). We therefore tested the effect of bone matrix extract on type I collagen degradation by osteoblasts stimulated with this dose of 1,25(OH)zD,. At concentrationc of 2- 4 gg protein/ml, bone matrix extract reduced the degradation of collagen to near control levels (Fig. 4), and it is perhaps noteworthy that inhibition of collagenolysis occurred at lower dose levels than t3H]thymidine incorporation. Assays of the culture medium from this experiment for collagenase and TIMP showed that inhibition of collagenolysis was accompanied by a stimulation in the levels of free TIMP and a concomitant reduction in net collagenase activity (Fig, 5). All the collagenase detected was in the latent form. We also tested the bone matrix extract itself for collagenase and TIMP activity: none could be detected over the range O.Ol-lOpug protein/ml (data not shown). Action of purified and recombinant growth factors
We next determined whether individual growth factors known to be pesent in bone matrix could mimic the effects of the bone matrix extract. First, recovery experiments were performed to demonstrate whether inhibition in collagenolysis by growth factors was due to toxic effects. The results of a typical experiment are
80 r = -0.93 v) ._ v)
:
>r
= 5 01 m = 6
60-
40-
a-0 20 -
27
I
I
0.2
0.5 Bone
Matrix
I
I
1 .o
2.0
Extract
I
5.0
I 10.0
(pg protein/ml)
Fig. 4. Inhibition of type I collagenolysis by bone matrix extract. Mouse calvarial osteoblasts were cultured on L’C-labelledtype I collagen films for 72 h in 16 mm diameter wells containing DMEM plus 2% acid treated rabbit serum and stimulated with 1,25(OH),D, (10 nglml); bone matrix extract was added at the concentrations shown and was present for the 72 h culture period. Collagenolysis is expressed as radioactivity released from each film as a percentage of the total. The regression line had a correlation coefficient of -0.93.
49 0.6
4
0.5 %
1.0
0.4
0.8
-
$ 02
C
5 p c 3
0.3
-
g
Om2
- 0.4
C
0.6
= al : zi zt =
0.1
-
1.0
Bone Matrix Extract
0”
I
I
0 0 ‘.
0.2
10.0”
(pg protein/ml)
Fig. 5. Inhibition of type I collagenolysis by bone matrix extract: effects on collagenase and TIMP activities. Culture supernatants from the experiment described in Fig. 4 were assayed for collagenase and TIMP activities by fibril assays. Each point represents the mean cf quadruplicate cultures and the curves are the best fit second order equations.
shown in Table 2. TGF-/3 was added at concentrations of 1 and 10 nglml to 1.25 (OH),D,-stimulated osteoblast cultures for an initial incubation period of 48 h; this resulted in an inhibition of collagenolysis. TGF-/3 was then removed and the cultures continued for a further 48 h in the presence of 1,25(OH),D,, resulting in lysis of the films and thereby indicating TGF-/? had exerted no toxic effects on the cells. Human TGF$, EGF, PDGF and the FGFs all exerted a dose-responsive inhibition of 1,25(0H),D,-stimulated collagen degradation. Maximal inhibition in this bioassay was 80-90%. The optimal doses of TGF-P, EGF and PDGF were 1. 10 Table 2 Effect of TGF-/3 on 1,25(0H),D,-stimulated calvarial osteoblasts (recovery experiment) -.___
lysis of type I collagen films by mouse
Addition
Initial incubation (O-48 h)
Recovery period (48-96 h)
None 1,25(OH),D, 1,25(OH),D, + 1 nglml TGF-P 1 nglml TGF-P 1,25(OH),D, + 10 nglml TGF-/? 10 ng/ml TGF-/3
8.00 40.00 13.4 2.77 6.95 3.92
85.05 87.50 84.45 59.75
+ f + f + k
0.30 12.00 1.77 0.68 0.27 0.46
+ + + f
0.90 0.62 1.10 3.5
For the initial incubation (O-48 h) TGF-P was added at concentrations of 1 and 10 nglml to mouse calvarial osteoblasts cultured on “C-labelled type I collagen films in the presence or absence of 1.25(OH)LDJ (10 nglml). During the recovery period (48-96 h). TGF-@ was removed and the cultures continued for another 48 h in the presence of 1,25(OH),D,. Collagenolysis is expressed as radioactivity released from the films as a percentage of the total. Data represents the mean k SEM for quadruplicate cultures.
M
TGF-p
d
EGF
M
PDGF
/
l-
I
1
1.0
0.1 Growth
factor
I
1
10
100
(W/ml)
Fig. 6. Effect of growth factors on 1,25(0H)2D3-stimulated collagenolysis. Mouse calvarial osteoblasts
were cultured on 14C-labelledtype I collagen films for 48 h in 16 mm wells containing DMEM plus 2% acid treated rabbit serum and stimulated with 1,25(OHjzD3 (10 @ml). Growth factors were added at the concentrations shown. Data is expressed as percentage inhibition of radioactivity released from the films; collagen degradation in the absence of growth factors was taken as 100%. Each point represents the mean f SEM for triplicate cultures.
loo-
75 C 0 ._ 5
M
aFGF
M
bFGF
50-
z._
I
I
I
I
I
5
10
20
50
100
Growth
factor
(ng/ml)
Fig. 7. Effect of growth factors on 1,25(0H)tD,-stimulated collagenolysis. Mouse calvarial osteoblasts were cultured on “C-labelled type I collagen films for 48 h in 16 mm wells containing DMEM plus 2% acid treated rabbit serum and stimulated with 1,25(OH),D, (10 q/ml). Acidic and basic FGF were added at the concentrations shown. Data is expressed as percentage inhibition of radioactivity released from the films; collagen degradation in the absence of growth factors was taken as 100%. Each point represents the mean + SEM for triplicate cultures.
51
Table 3 Effect of growth factors on collagenase and TIMP activity by 1,25(OH j$,-stimulated mouse calvarial osteoblasts Addition
TGF-@
EGF
PDGF
aFGF
bFGF
Concentration (@ml)
Collagenase (units/ml)
TIMP (units/ml)
0 0.01 0.1 1.0 0 0.1 1.0 10.0 0 0.1 1.0 10.0 50.0 0 1.0 10.0 50.0 0 1.0 10.0 50.0
0.37 0.44 0.27 0.16 0.82 0.80 0.64 0.51 0.42 0.40 0.38 0.24 0.08 1.05 0.92 0.87 0.40 1.05 1.12 0.65 0.39
0 0 0 0.51 0 0 0 0.54 0 0 0 0.41 2.78 0 0 0 1.77 0 0 0.07 0.05
--f + + + + k f + f + k + + + f + + + f + f
0.01 0.13 0.01 0.01 0.09 0.06 0.04 0.01 0.01 0.02 0.01 0.01 0.02 0.14 0.06 0.17 0.04 0.14 0.04 0.06 0.06
+ 0.31
+ 0.25
zk 0.08 + 0.10
+ 0.21
f 0.07 + 0.04
Mouse calvarial osteoblasts were cultured on “C-labelled type I collagen films for 48 h in 16 mm wells containing DMEM plus 2% acid treated rabbit serum and stimulated with 1,25(OH),D, (10 ng!ml). Growth factors were added at the concentrations shown. Culture supernatants were assayed for collagenase and TIMP activity by fibril assay. Each point represents the mean + SEM for triplicate cultures.
Table 4 Effect of insulin-like growth factors on collagenase and TIMP activity by 1,25(OH),D,-stimulated mouse calvarial osteoblasts Addition
IGF-I
IGF-II
(/ml)
Collagenase (units/ml)
TJMP (units/ml)
0 O*lpg I*Oclg 1o.ojcg 0 10.0 ng 50.0 ng 100.0 ng
1.42 1.41 1.36 0.69 1.42 1.14 1.01 1.06
0.02 + 0.01 0 0 0 0.02 + 0.01 0 0 0
Concentration
f k + f f f + +
0.18 0.03 0.14 0.05 0.18 0.09 0.06 0.10
Mouse calvarial osteoblasts were cultured on i4C-labelled type I collagen films for 48 h in 16 mm wells containing DMEM plus 2% acid treated rabbit serum and stimulated with 1,25(OH),D, (10 nglml). The IGFs were added at the concentrations shown. Culture supernatants were assayed for collagenase and TIMP activity by fibril assay. Each point represents the mean + SEM for triplicate cultures.
52 and 50 ng/ml respectively (Fig. 6), and 100 ng/ml for both aFGF and bFGF (Fig. 7). Inhibition of collagenolysis was accompanied by a reduction of collagenase activity and the appearance of free TIMP in the culture medium at the higher doses tested (Table 3). Treatment of unstirnulated osteoblasts with growth factors alone had no effect on collagenase activity, but exerted stimulatory effects on TIMP production (data not shown). Although both IGFs partially inhibited 1,25(OH),D,-stimulated collagenase production by osteoblasts (Table 4), this had no significant effect on type I collagenolysis over the 48 h time-course of the experiment (data not shown); unlike the other growth factors however, the IGFs did not stimulate TIMP synthesis by either 1,25(0H),D,-treated osteoblasts (Table 4) or unstimulated cells (data not shown). From these data we conclude that inhibition of collagenolysis was primarily dependent upon the ability of growth factors to stimulate TIMP synthesis.
Discussion The present series of experiments demonstrate an important role for growth factors in regulating the activity of collagenase and TIMP and type I collagen degradation by mouse calvarial osteoblasts. EDTAITris-HCl extracts of bone matrix exerted a dose-dependent inhibition of type I collagenolysis by osteoblasts stimulated with 1,25(OH),D,. Inhibition was accompanied by a reduction in collagenase activity and an increase in free TIMP in the culture medium. Although growth factor activity in our bone matrix preparation was only partially characterized, EDTAITris-HCl extracts of bovine bone have previously been shown to contain TGF-P, PDGF and FGF activity [14], and we felt it impor?ant to initially test its effect on the proteolytic behaviour of osteoblasts; target cells are likely to be exposed to mixtures rather than individual gro*Nthfactors in vivo. The role of growth factors in the regulation of MMP and TIMP expression by connective tissue cells is complex. The response of the cell type under investigation also appears to be variable. EGF, PDGF and TGF$ have been shown to stimulate collagenase synthesis in human fibroblast cultures [26,27], and Edwards et al. [28] found that exposure of quiescent human MRCJ fibroblasts to bFGF and EGF induced the expression of both collagenase and TIMP mRNA and protein. TGF#l alone had little effect on the expression of these genes, but in the presence of other growth factors had a reciprocal effect on MMP expression; TGF-P repressed the induction of collagenase and interacted with bFGF and EGF to super-induce TIMP expression [28]. Overall et al. also found that TGF-/? suppressed collagenase and increased TIMP and PAI-1 expression by human gingival fibroblasts [29] and rat calvarial osteoblasts [20]. TGF-/3 did, however, stimulate the synthesis of another MMP (72-kDa progelatinase) by both cell types [20,29]. In the present model system, TGF-/3, EGF, PDGF and both acidic and basic FGF, all inhibited collagenase and stimulated TIMP activity by mouse osteoblasts cultured simultaneously with I,25(OH),D,. In contrast the IGFs, which did not inhibit type I collagenolysis, only partially inhibited collagenase activity and did not stimulate TIMP production by
53 either 1,25(0H),D,-treated or untreated cells. These findings suggest that rnhibition of collagenolysis was dependent primarily upon the ability of certain growth factors to stimulate TIMP synthesis. There are disadvantages in relying on the fibril assay for estimating cohagenase and TIMP synthesis by cells in culture: low levels of collagenase are likely to remain undetected in supernatants containing relatively high levels of TIMP. TIMP does not interact with latent enzyme, but once the enzyme is activated by including APMA in the assay it is avidly bound by free TIMP, yielding irreversible collagenase-TIMP complexes which cannot be detected in activity assays [30]. TIMP production will be similarly underestimated in supernatants where the level of naturally activated enzyme exceeds that of TIMP. At no time could active collagenase be detected in culture supernatants from any of the present series of experiments, even though coliagenolysis had occurred. However, we have previously shown that for mouse osteoblasts to degrade type I collagen films, only a minute proportion of the latent collagenase secreted needs to be activated by plasmin [12,13]. Furthermore, although it has not been specifically demonstrated for osteoblasts, some cells possess a specific binding site for urokinase-type plasminogen activator (u-PA) and its proform (for review see Ref. 31), which allows cells to acquire surface bound plasminogen-activating ability. The confinement of u-PA to discrete sites may thus explain how cells which constitutively produce TIMP and PAL1 can still produce focal proteolysis. The finding that rodent osteoblasts synthesize collagenase [7-91, has been responsible for the hypothesis that bone cells play a major role in bone resorption by degrading the surface osteoid layer, thereby preparing the underlying mineralized bone for osteociastic action [6,10,11]. Recent evidence however, suggests that this model might not apply to bone resorption in the human. Human cells of diverse tissue origin have been shown to produce collagenase, and although human osteoblasts secrete gelatinase as a major cellular product they do not produce collagenase in significant quantities, either constitutively or in response to bone resorbing agents [32,33]. While it would be unusual to expect different cellular and/or molecular mechanisms to have evolved for the resorption of human bone, the reasons for these cellular and species differences remain unclear at present, but are the subject of current investigation. In conclusion, the findings of the present series of experiments provide additional evidence for the tight control exerted on the proteolytic activity of osteoblasts and the importance of TIMP in its regulation. They further suggest that the conversion (coupling) of the initial resorptive phase of the bone remodeiling cycle to one of deposition, may be mediated by polypeptide growth factors either produced locally by osteo!riasts, or released by proteolysis from the bone matrix itself.
Acknowledgements Supported by grants from the Arthritis and Rheumatism Council and the Medical Research Council. We thank Christopher Green and Judith Webdell for preparing
54
the illustrations and Elaine Pocock and Angela McMonagle for secretarial assistance.
References 1 Parfitt AM. The coupling of bone formation to bone resorption: a critical analysis of the concept and of its relevance to the pathogenesis of osteoporosis. Metab Bone Dis Rel Res 1982;4:1-6. 2 Frost HM. Dynamics of bone remodelling. In: Frost HM, ed. Boue biodynamics. Boston: Little Brown & Co., 1964. 3 Raina V. Normal osteoid tissue. J Clin Path01 1972;25:229-232. 4 Fornasier VL. Transmission electron microscopy studies of osteoid maturation. Metab Bone Dis Rel Res 1980;25:103-108. 5 Vanderveil CJ. An ultrastructural study of the components which make up the resting surface of bone. Metab Bone Dis Rel Res 1980;28:109-116. 6 Sakamoto M, Sakamoto S. Bone collagenase, osteoblasts and cell-mediated bone resorption. In: Peck WA, ed. Bone and Mineral Research, Vol. 4. Amsterdam: Elsevier, 1986;49-102. 7 Heath JK, Atkinson SJ, Meikle MC, Reynolds JJ. Mouse osteoblasts synthesize collagenase in response to bone resorbing agents. Biochim Biophys Acta 1984;802:151-154. 8 Sakamoto M, Sakamoto S. Immunocytochemical localization of collagenase in isolated mouse bone cells. Biomed Res 1984;5:29-38. 9 Gtsuka K, Sodek J, Limeback H. Synthesis of collagenase and collagenase inhibitors by osteoblastlike cells in culture. Eur J Biochem 1984;145:123-129. 10 Chambers TJ, Darby JA, Fuller K. Mammalian collagenase predisposes bone surfaces to osteoclastic resorption. Cell Tissue Res 1985;241:671-675. 11 Chambers TJ, Fuller K. Bone cells predispose bone surfaces to resorption by exposure of mineral to osteoclastic contact. J Cell Sci 1985;76:155- 165. 12 Thomson BM, Atkinson SJ, Reynolds JJ, Meikle MC. Degradation of type I collagen films by mouse osteoblasts is stimulated by 1,25-dihydroxyvitamin D, and inhibited by human recombinant TIMP (tissue inhibitor of metalloproteinases). Biochem Biophys Res Commun 1987;148:596-602. 13 Thompson BM, Atkinson SJ, McGarrity AM, Hembry RM, Reynolds JJ, Meikle MC. Type I collagen degradation by mouse calvarial osteoblasts stimulated with 1,25_dihydroxyvitamin D-3: evidence for a plasminogen-plasmin-metalloproteinase cascade. Biochim Biophys Acta 1989;1014: 125- 132. 14 Hauschka PV, Mavrakos AE. Iafrati MD, Doleman SE, Klagsburn M. Growth factors in bone matrix, Isolation of multiple types by affinity chromatography on heparin-sepharose. J Biol Chem 1986:261:12665- 12674. 15 Centrella M, Canalis E. Transforming and nontransforming growth factors are present in medium conditioned by fetal rat calvariae. Proc Nat1Acad Sci USA 1985;82:7335-7339. 16 Mohan S, Jennings JC, Linkhart TA, Baylink DJ. Primary structure of human skeletal growth factor: homology with human insulin-like growth factor-II. Biochim Biophys Acta 1988;966:44-55. 17 Frolik CA, Ellis LF, Williams DC. Isolation and characterization of insulin-like growth factor-11 from human bone. Biochem Biophys Res Commun 1988;151:1011-1015. 18 Partridge NC, Jeffrey JJ, Ehlich LS, Teitelbaum SL, Fliszar C, Welgus HG, Kahn AJ. Hormonal regulation of the production of collagenase and a collagenase inhibitor activity by rat osteogenic sarcoma cells. Endocrinology 1987;120:1956-1962. 19 Delaisse JM, Eeckhout Y, Vaes G. Bone-resorbing agents affect the production and distribution of procollagenase as well as the activity of collagenase in bone tissue. Endocrinology 1988;123:264-276. 20 Overall CM, Wrana JL, Sodek J. Transforming growth factor-p regulation of collagenase, 72 kDaprogelatinase, TIMP and PAI- expression in rat bone cell populations and human fibroblasts. Connect Tissue Res 1989;20:289-294. 21 Shen V, Kohler G, Jeffrey JJ, Peck WA. Bone-resorbing agents promote and interferon-y inhibits bone cell collagenase production. J. Bone Mineral Res 1988;3:657-666.
55 22 Gospodarowicz D. Locahsation of a fibroblast growth factor aild its effect alone and with hydrocortisane on 3T3 cell growth. Nature 1974;249:123- 127. 23 Twardzik DR, Ranchalis JE, Todaro GJ. Mouse embryonic transforming growth factors related to those isolated from tumor cells. Cancer Res 1982;42:590-593. 24 Gavrilovic J, Reynolds JJ, Murphy G. Inhibition of type I collagen film degradation by tumor cells using, a specific antibody to collagenase and the specific tissue inhibitor of metalloproteinases (TIMP). Cell Biol Int Rep 1985;9:1097-1107. 25 Cawston TE, Barrett AJ. A rapid and reproducible assay for collagenase using [l-‘4C]-acetylated collagen. Anal Biochem 1979;99:340-345. 26 Bauer EA. Cooper TW, Huang JS, Altman J, Deuel TF, Stimulation of in vitro human skin collagenase expression by platelet-derived growth factor. Proc Nat1 Acad Sci USA 1985;82:4132-4136. 27 Chua CC, Geiman DE, Keller GH, Ladda RL. Induction of collagenase secretion in human fibroblast cultures by growth promoting factors. J Biol Chem 1985;260:5213-5216. 28 Edwards DR, Murphy G, Reynolds JJ, Whitham SE, Docherty AJF, Angel P, Heath JK. Transforming growth factor beta modulates the expression of collagenase and metalloproteinase inhibitor. EMBO J 1987;6:1899- 1904. 29 Overall CM, Wrana JL, Sodek J. Independent regulation of collagenase, 72-kDa-progelotinase, and metalloendoproteinase inhibitor expression in human fibroblasts by transforming growth factor-D. Proc Nat1 Acad Sci USA 1989;264:1860-1869. 30 Cawston TE, Murphy G, Mercer E, Galloway WS, Hazelman BL, Reynolds JJ. The interaction of purified rabbit bone collagenase with purified rabbit bone metalloproteinase inhibitor. Biochem J 1983;211:313-318. 31 Blasi F, Vassalli JD, Dano K. Urokinase-type plasminogen activator: proenzyme, receptor, and inhibitors. J Cell Biol1987;104:801-804. 32 Gowen M, Wood DD, Ihrie EJ. Meats JE, Russell RGG. Stimulation by human interleukin 1 of cartilage breakdown and production of collagenase and proteoglycanase by human chondrocytes but not human osteoblasts in vitro. Biochim Biophys Acta 1984;797:186-193. 33 Rifas L. Halstead LR, Peck WA, Avioli LV, Welgus HG. Human osteobiasts in vitro secrete tissue inhibitor of metalloproteinases and gelatinase but not intersiitial coilagenase as major cellular products. J Clin Invest 1989;84:686-694.
12(1991)41-S
41
Elsevier BAM 00335
one-derived growth factors modulate collagenase and TIMP (tissue inhibitor of metalloproteinases) activity and type I collagen degradation by mouse calvarial osteoblasts
Murray C. Meikle’q’ , Anne M. McGarrity’, Brian M. Thomson’ and John J. Reynolds’ ‘Cell Physiology Department. Strangeways Research Laboratory, CBI 4RN and ‘Comective
Worrs Causeunay, Ca,nbridge
Tissue Research Unit, Department of Orthodontics. Ittstituteof
Dental Surgery. University of London, 36
Gray’s Inn Road, London WCIX 8LD. Etq$and
(Received 20 April 1990) (Accepted 17 September 1990)
Summary The finding that osteoblasts synthesize collagenase has led to the hypothesis that bone cells play a major role in bone resorption by degrading the surface osteoid layer, thereby preparing the underlying mineralized bone for osteoclastic action. To further understand the mechanisms regulating osteoid removal, mouse calvarial osteoblasts were cultured on ‘JC-labelled type I ccllagen films and the abilities of(i) bovine bone matrix extracts and (ii)purified or recombinant human growth factors. to modify their collagenolytic behaviour were investigated. EDTAITris-HCI extracts of bone matrix containing growth factor activity. exerted a dose-dependent inhibition of type I collagenolysis by osteoblasts stimulated with !.25dihydroxyvitamin D, (I .25(OH),D,, 10 nglml). Inhibition was accompanied by a reduction in collagenase activity and an increase in free TIMP (tissue inhibitor of metalloproteinases) in the culture medium. Transforming growth factor-@. epidermal growth factor, platelet-derived growth factor and the acidic and basic fibroblast growth factors all mimicked these effects. In contrast, insulin-like growth factors-I and -II did not inhibit type I collsgenolysis, only partially inhibited collagenase activity. and did not stimulate TIMP production by either 1.25(OH)zD,-treated or untreated cells. These findings provide additional evidence for the tight control exerted on the proteolytic activity of osteoblasts and the importance of TIMP in its regulation. They suggest strongly that the conversion (coupling) of the initial resorptive phase of the bone remodellinp cycle to one of deposition, may be mediated by polypeptide growth factors either produced iocally by osteoblasts. or released by proteolysis from the bone matrix.
Key words: Growth factor: Osteoblast:
Collagenase:
TIMP: Type I collagenolysis
Correspondence to: Dr M.C. Meikle, Strangeways Research Laboratory, CB14RN, England, U.K. 0169-6009/91/$03.50@
Worts Causeway. Camhridge
1991 Elsevier Science Publishers B.V. (Biomedical Division)
42 Introduction The maintenance of skeletal integrity is dependent upon the continuous remodelling of bone, a co-ordinated sequence of cell mediated events involving an initial period of osteoclastic resorption followed by the refilling of each resorption cavity with new bone by osteoblasts: this process is referred to as coupling [1,2]. Since bone remodelling occurs at anatomically discrete sites throughout the skeleton, it seems likely that it is regulated by factors either produced locally by bone cells or released from the bone matrix during the initial resorptive phase. Non-resorbing bone surfaces are covered by a layer of nonmineralized osteoid [3-S], and recent evidence suggests that one way osteoblasts might regulate bone resorption is through mineral exposure (for review see Ref. 6). Osteoblasts stimulated with bone resorbing agents respond by synthesizing the matrix metalloproteinase (MMP) collagenase [7-91, and isolated osteoclasts in vitro do not resorb bone unless the surface osteoid layer is first removed, either by pre-treatment with collagenase or by a layer of osteoblasts [ 10,111. To understand further the mechanisms regulating osteoid removal by osteoblasts, we have used a model system in which mouse calvarial osteoblasts are grown on type I collagen films, and previously provided evidence for the involvement of a plasminogen-plasmin-MMP activation cascade in type I collagen degradation by osteoblasts, and for its regulation by TIMP (tissue inhibitor of metalloproteinases) and plasminogen activator inhibitor-l (PA&l [12,13]). Type i collagenolysis in this in vitro model, however, is not synonymous with osteoid removal in vivo. Although 90% of total bone protein is type I collagen, bone matrix also contains many noncollagenous proteins including polypeptide growth factors [ 14- 171, which individually or synergistically could modulate osteoblast responsiveness to bone resorbing agents; recent research from several groups has suggested an important role for growth factors and hormones in the regulation of MMP and TIMP synthesis by osteoblasts and the maintenance of skeletal homeostasis [18-211. In the present series of experiments therefore, we tested the abilities of (i) bovine bone matrix extracts and (ii) purified or recombinant human growth factors, to modify the collagenolytic behaviour of mouse calvarial osteoblasts stimulated by 1,25(OH),D,,
Materials and Methods Reagents
1,25(OH),D, was generously supplied by Dr Ian Dickson, Department of Medicine, University of Cambridge. Purified human platelet derived growth factor (PDGF), acidic and basic fibroblast growth factors (aFGF, bFGF) and epidermal growth factor (EGF) were purchased from Biogenesis Ltd, Boumemouth, England. Purified human transforming growth factor-p (TGF-P) was the generous gift of Dr John K. Heath, Department of Biochemistry, University of Oxford, and also purchased from R & D Systems, Minneapolis, Minnesota. Human recombinant insulin-like growth factor-I (IGF-I) was a gift from Dr John T. Dingle, Strangeways Re-
43 search Laboratory, Cambridge; human recombinant insulin-like growth factor-II (IGF-II) was purchased from Bachem Inc., Torrance, CA. Preparation of bone matrix extract
Bovine tarsal bones were cleaned of adherent soft tissue, the epiphyses and marrow removed and cut into 1 cm cylinders. These were quartered, cooled in liquid N, and fragmented prior to being reduced to a fine powder in a Spex freezer mill (Glen Creston Instruments, Stanmore, Middlesex). Bone matrix extract was prepared as described previously [ 143. Briefly, dry bone powder was suspended in 0.02 M TrisHCl, pH 7.5, at 4 “C for 20 min and then centrifuged at 2000 x g for 10 min and the supernatant discarded. The residue was resuspended in 0.5 M ammonium EDTA, 0.02 M Tris-HCl, pH 7.5, with gentle stirring for 24 h at 4 “C. The supernatant was centrifuged (10 000 x g) and filtered. Dialysis at 4 “C in 70 mm dialysis tubing against water for 4 days and 0.02 M NH,HCOs for 2 days with frequent changes was followed by lyophilization and storage at -20 “C. For experimental purposes bone matrix extract was redissolved in 0.1 M NaCl, 10 mM Tris, pH 7.0, dialysed against this buffer and then centrifuged at 2000 x g for 10 min. The protein content of bone matrix extract was estimated at 9.8 mg protein/g using a protein assay kit (Bio-Rad, Watford, Hertfordshire). Growth factor assay
Growth factor activity in bovine bone matrix extract was assayed by its ability to stimulate the proliferation of BALB/c/3T3 cells [22]. BALBlcI3T3 cells were cultured in Dulbecco’s modification of Eagle’s medium (DMEM) supplemented with antibiotics and 10% fetal calf serum (FCS; Globefarm, Esher, Surrey). To test for growth factor activity, cells were plated into 96-well microtitre plates (Cel-Cult; Sterilin, Feltham, Middlesex) at a density of lo4 cells/well in 200 ~1 DMEM plus 10% FCS, and grown to confluence at 37 “C in a humidified atmosphere of 5% CO,/95% air. After a brief wash in DMEM the cells were challenged with serial dilutions of bovine bone marrow extract in 250 ,~l DMEM. On each plate 20 wells were set up with dilutions ranging between 2 and 30 ~1 FCS/250 ~1 DMEM to provide a serum-stimulation curve. One Growth Factor Unit (GFU) was defined as half the maximal stimulation of proliferation in response to FCS [ 141. BALBlcl3T3 cells were incubated with the test factors for 42 h, after which the medium was changed and replaced with 200 ~1 DMEM containing 1 &X/ml [3H]thymidine (Amersham International, Amersham, Buckinghamshire) and incubated for a further 6 h. Acid-insoluble [3H]thymidine in cell DNA was measured ;asfollows: after removing the medium, the cells were washed in DMEM, fixed with 200~15% icecold trichloroacetic acid (TCA) for 20 min, and washed in phosphate-buffered saline (PBS). Cell layers were then dissolved in lOO@ 0.1 M NaOH for 10 min and removed into a scintillation vial; 400~1 of 0.5 M acetic acid and 2 ml scintillant were added to each sample before counting for [3H]thymidine incorporation. Assay for TGF-@
TGF-/3 activity in bone matrix extract was assayed by the ability of TGF-@ in the
44
presence of EGF. to stimulate anchorage-independent growth of normal rat kidney fibroblasts (NRK, clone 49F) in soft agar [23]. A stock solution of 1% purified agar (Oxoid) was prepared in distilled water and sterilized by autoclaving. DMEM was prepared at double strength and both solutions maintained at a temperature of 47 “C. An underlay containing 0.5% agar and 10% FCS in DMEM was prepared from these concentrated solutions and 1 ml pipetted into each well of a 24-well plate (Linbro) and solidified at room temperature. A single cell suspension of NRK-49F cells was prepared in DMEM plus 10% FCS. Once the underlay had solidified, 200 3 X lo4 cells, 10% FCS, 0.25% agar and test factors, in the ,~l DMEM containing 1.A absence or presence of EGF (50 nglml), were pipetted over the underlay in each well and the cells cultured at 37 “C for 7-14 days without further addition. Transformation was assessed in unfixed and unstained cultures by light microscopy. Colonies counted in four random fields (2.27 mm”) in each of three replicate wells were categorized as small (~50 pm in diameter) or large (>50-200 pm in diameter) as previously described [IS]. Purified human TGF-@ (10 nglml) was used as a positive control. Preparation of osteoblastsfrom neonatal mouse calvariae
Calvarial osteoblasts were prepared as previously described [7]. The cells were characterized as osteoblast-like by morphological criteria, and by demonstrating that essentially all the cells stained strongly for alkaline phosphatase, synthesized type I collagen (no type II or III), and accumulated CAMP in response to parathyroid hormone and prostaglandin E, but not to calcitonin or 1,25(OH),D, [7]. Neonatal mouse calvariae (40-50) were dissected free from adherent soft tissue, washed in Modified Biggers’ Medium (Imperial) with 6 mM sodium bicarbonate and I .4 mM glutamine (10 min), and sequentially digested with 1 mg/ml trypsin (10 min), 2 mg/ml dispase (30 min) and 4 mg/ml collagenase (2 x 30 min) in Ca’+ and Mg’+ free Tyrode’s solution. Cells released by the collagenase digestions were washed, and grown to confluence in DMEM containing 10% FCS. Preparation of collagen films
Radiolabelled collagen films were prepared as described previously [24]. Briefly, a!iquots of ‘JC-acetylated collagen (rat skin type I; 150 pg in 300 ~1 of 10 mM phosphate buffer, PI-I 7.4, containing 300 mM NaCl and 0.02% sodium azide) were dispensed into tissue culture wells (Linbro, 16 mm diameter) and dried at 37 “C. The collagen was then washed twice with sterile distilled water and once with DMEM. Osteoblasts and collagen films
Osteoblasts (10’ well) were settled onto the collagen films and cultured in 1 ml DMEM plus 10% FCS. After 16 h the cells were washed with DMEM and cultured for 48 h (72 h in experiments described in Figs. 3 and 4) in 1 ml DMEM plus 2% acid treated rabbit serum plus either 1,25(OH),D, (in ethanol) or vehicle alone. These media were assayed for collagenase inhibitory activity (TIMPs plus cr,-macroglobulin) which was never more than 0.01 units/ml. The cultures were maintained at 37 “C in a humidified atmosphere of 5% CO,/95% air. At the end of the incubation
45 period, the culture medium was centrifuged (15 min, 12 000 x g) to remove any collagen fibrils, and the radioactivity released during collagen degradation quantified by liquid scintillation counting. Residual collagen was digested with bacterial collagenase (50 pug/ml) and assayed for radioactivity. Collagenolysis was expressed as a percentage of the total radioactivity released from the films f standard error of the mean (SEM). Effect of growth factors on type I collagenolysis
Mouse calvarial osteoblasts were cultured on i4C-labelled type I collagen films for 48 h in 16 mm wells containing 1 ml DMEM plus 2% acid treated rabbit serum and stimulated with 1,25(OH)zD, (10 ng/ml). Growth factors were added and collagenolysis determined as above; culture supernatants were assayed for collagenase and TIMP. The effects of growth factors on collagenase and TIMP synthesis by unstimulated osteoblasts were also tested. Assays for collagenase and TIMP
Collagenase activity in culture supernatants was measured by the release of labelled peptides from “C-acetylated rat skin collagen [25]. One unit of collagenase degrades 1 ,ug reconstituted fibrils per min at 35 “C. Latent collagenase was activated by including 4-aminophenylmercuric acetate (APMA) in the fibril assay (0.67 mM final concentration). TIMP was assa,Ied by incubating a standard amount (0.05 unit) of activated rabbit skin collagenase in the fibri! assay with samples of culture medium. One unit of TIMP is defined as the amount of TIMP required to inhibit 2 units of collagenase by 50%.
Results Assays of growth factors in bone matrix extract
We initially examined the effect of various concentrations of bone matrix extract (0.5-40 pg protein/ml) on proliferation and DNA synthesis by BALBlcI3T3 cells. A typical dose-response is shown in Fig. 1. By definition the maximal response possible in this bioassay is 2 GFU; 1 GFU was attained at a dose of approximately 1Opg protein/ml. Since bone matrix is a rich source of TGF-a. we also tested the ability of bone matrix extract to stimulate anchorage-independent growth of NRK-49F cells in soft agar in the presence or absence of EGF (50 @ml). TGF$ alone did not stimulate anchorage-independent growth, but in the presence of EGF resulted in a 3-fold increase in large colony formation over EGF alone (Table 1). At concentrations of 5-2Opg protein/ml, bone matrix extract did not induce anchorage-independent growth and did not enhance the effect of EGF on colony formation. At higher concentrations (40-50 pg protein/ml) however, bone matrix extract alone had a smaii but significant effect on colony formation, and in the presence of EGF stimulated anchorage-independent growth 3-fold and large colony formation 2-3-fold, thereby indicating the presence of both EGF and TGF-/3. Furthermore, although large colony formation by higher doses of bone matrix
46 3r r = 0.97 In C 5
0 2-
& 5 lf r %
l-
G
0
0. 0.1
I 10
1.0 3one
Matrix
Extract
I 100
(pg protein/ml)
Fig. 1. Growth factor activity of bone matrix extract on BALBlcI3T3 fibroblasts. Bone matrix extract was assayed on confluent BALBIcl3T3 cells in 96.well microtitre plates containing 250~1 DMEM. After 48 h the TCA-insoluble incorporation of [3H]thymidine into DNA was quantified as described in Methods. One growth factor unit (GFU) was defined as half the maximal serum-stimulated [3H]thymidine incorporation above background levels. Each point represents the mean of four cultures. The regression iine shown for doses of 1pg protein/ml and above had a correlation coefficient of 0.97.
Table 1 Anchorage-independent growth of NRL49F cells in soft agar in the presence of EGF, TGF-/I and bone matrix extract alone or in combination Addition Factor None TGF-/? Bone matrix extract
Concentration (nglml) 10 5000 10 000 20 000 40 OtM 50 000
% Transformed”
% Large coloniesb
-EGF
+EGF
-EGF
+EGF
0 0 0 0 0 0.1 0.1
1.0 0.7 0.4 1.2 1.3 3.0 3.6
0 0 0 0 0 0 0
6.0 17.6 5.0 6.9 7.9 21.5 14.7
NRK-49F cells (1.2 x l@/well) were &Wed in soft agar in the presence or absence of EGF (50 nglml) \ and the concentrations of TGF-/3 and bone.m;trix extract as indicated. Colonies were measured unfixed and unstained by light microscopy. Small colonies were defined as being ~50 pm in diameter and large colonies >50-200pm in diameter. Data are from four random fields (2.27 mm2) in each of three replicate cultures; each field had approximately 1200plated cells. a Total number of colonies/total number of cells plated. b Number of large colonies/total number of colonies.
..
Fig. 2. EGF-dependent colony formation by NRK-49F cells in soft agar. (a) No addition; (b) EGF (50 ng/ml); (c) TGF-/3 (10 @ml); (d)EGF (50 nglml) plus TGF-B (10 @ml); (e) bone matrix extract (5Opug protein/ml); (I) bone matrix extract (SOpg protein/ml) plus EGF (50 @ml). Bar = 100pm.
100
m ._
-
r = 0.92
80 .
lo h
Z
2
4 1,25(OH)2
6 -vitamin
6
10
12
D6 (ng/ml)
Fig. 3. Dose-responsive degradation of “C-labelled type I collagen films by mouse calvarial osteoblasts in response to 1,25(GH),D,. Cells were cultured for 72 h on type I collagen films in 16 mm wells containing DMEM plus 2% acid treated rabbit serum and the concentrations of 1,25(OH),D, as indicated. Culture supernatants and residual collagen were assayed for radioactivity; collagenolysis is expressed as radioactivity released from each film as a percentage of the total. The regression line had a correlation coefficient of 0.92.
48 extract in the presence of EGF was comparable to that of TGF-P, the percentage of transformation colonies was 4-fold higher (Table 1, Fig. 2). Collagenolysis and the effects of bone matrix extract
The optimal dose of 1,25(OH),D3 on type I collagenolysis by mouse calvarial osteoblasts was found to be approximately 10 ng/ml (Fig. 3). We therefore tested the effect of bone matrix extract on type I collagen degradation by osteoblasts stimulated with this dose of 1,25(OH)zD,. At concentrationc of 2- 4 gg protein/ml, bone matrix extract reduced the degradation of collagen to near control levels (Fig. 4), and it is perhaps noteworthy that inhibition of collagenolysis occurred at lower dose levels than t3H]thymidine incorporation. Assays of the culture medium from this experiment for collagenase and TIMP showed that inhibition of collagenolysis was accompanied by a stimulation in the levels of free TIMP and a concomitant reduction in net collagenase activity (Fig, 5). All the collagenase detected was in the latent form. We also tested the bone matrix extract itself for collagenase and TIMP activity: none could be detected over the range O.Ol-lOpug protein/ml (data not shown). Action of purified and recombinant growth factors
We next determined whether individual growth factors known to be pesent in bone matrix could mimic the effects of the bone matrix extract. First, recovery experiments were performed to demonstrate whether inhibition in collagenolysis by growth factors was due to toxic effects. The results of a typical experiment are
80 r = -0.93 v) ._ v)
:
>r
= 5 01 m = 6
60-
40-
a-0 20 -
27
I
I
0.2
0.5 Bone
Matrix
I
I
1 .o
2.0
Extract
I
5.0
I 10.0
(pg protein/ml)
Fig. 4. Inhibition of type I collagenolysis by bone matrix extract. Mouse calvarial osteoblasts were cultured on L’C-labelledtype I collagen films for 72 h in 16 mm diameter wells containing DMEM plus 2% acid treated rabbit serum and stimulated with 1,25(OH),D, (10 nglml); bone matrix extract was added at the concentrations shown and was present for the 72 h culture period. Collagenolysis is expressed as radioactivity released from each film as a percentage of the total. The regression line had a correlation coefficient of -0.93.
49 0.6
4
0.5 %
1.0
0.4
0.8
-
$ 02
C
5 p c 3
0.3
-
g
Om2
- 0.4
C
0.6
= al : zi zt =
0.1
-
1.0
Bone Matrix Extract
0”
I
I
0 0 ‘.
0.2
10.0”
(pg protein/ml)
Fig. 5. Inhibition of type I collagenolysis by bone matrix extract: effects on collagenase and TIMP activities. Culture supernatants from the experiment described in Fig. 4 were assayed for collagenase and TIMP activities by fibril assays. Each point represents the mean cf quadruplicate cultures and the curves are the best fit second order equations.
shown in Table 2. TGF-/3 was added at concentrations of 1 and 10 nglml to 1.25 (OH),D,-stimulated osteoblast cultures for an initial incubation period of 48 h; this resulted in an inhibition of collagenolysis. TGF-/3 was then removed and the cultures continued for a further 48 h in the presence of 1,25(OH),D,, resulting in lysis of the films and thereby indicating TGF-/? had exerted no toxic effects on the cells. Human TGF$, EGF, PDGF and the FGFs all exerted a dose-responsive inhibition of 1,25(0H),D,-stimulated collagen degradation. Maximal inhibition in this bioassay was 80-90%. The optimal doses of TGF-P, EGF and PDGF were 1. 10 Table 2 Effect of TGF-/3 on 1,25(0H),D,-stimulated calvarial osteoblasts (recovery experiment) -.___
lysis of type I collagen films by mouse
Addition
Initial incubation (O-48 h)
Recovery period (48-96 h)
None 1,25(OH),D, 1,25(OH),D, + 1 nglml TGF-P 1 nglml TGF-P 1,25(OH),D, + 10 nglml TGF-/? 10 ng/ml TGF-/3
8.00 40.00 13.4 2.77 6.95 3.92
85.05 87.50 84.45 59.75
+ f + f + k
0.30 12.00 1.77 0.68 0.27 0.46
+ + + f
0.90 0.62 1.10 3.5
For the initial incubation (O-48 h) TGF-P was added at concentrations of 1 and 10 nglml to mouse calvarial osteoblasts cultured on “C-labelled type I collagen films in the presence or absence of 1.25(OH)LDJ (10 nglml). During the recovery period (48-96 h). TGF-@ was removed and the cultures continued for another 48 h in the presence of 1,25(OH),D,. Collagenolysis is expressed as radioactivity released from the films as a percentage of the total. Data represents the mean k SEM for quadruplicate cultures.
M
TGF-p
d
EGF
M
PDGF
/
l-
I
1
1.0
0.1 Growth
factor
I
1
10
100
(W/ml)
Fig. 6. Effect of growth factors on 1,25(0H)2D3-stimulated collagenolysis. Mouse calvarial osteoblasts
were cultured on 14C-labelledtype I collagen films for 48 h in 16 mm wells containing DMEM plus 2% acid treated rabbit serum and stimulated with 1,25(OHjzD3 (10 @ml). Growth factors were added at the concentrations shown. Data is expressed as percentage inhibition of radioactivity released from the films; collagen degradation in the absence of growth factors was taken as 100%. Each point represents the mean f SEM for triplicate cultures.
loo-
75 C 0 ._ 5
M
aFGF
M
bFGF
50-
z._
I
I
I
I
I
5
10
20
50
100
Growth
factor
(ng/ml)
Fig. 7. Effect of growth factors on 1,25(0H)tD,-stimulated collagenolysis. Mouse calvarial osteoblasts were cultured on “C-labelled type I collagen films for 48 h in 16 mm wells containing DMEM plus 2% acid treated rabbit serum and stimulated with 1,25(OH),D, (10 q/ml). Acidic and basic FGF were added at the concentrations shown. Data is expressed as percentage inhibition of radioactivity released from the films; collagen degradation in the absence of growth factors was taken as 100%. Each point represents the mean + SEM for triplicate cultures.
51
Table 3 Effect of growth factors on collagenase and TIMP activity by 1,25(OH j$,-stimulated mouse calvarial osteoblasts Addition
TGF-@
EGF
PDGF
aFGF
bFGF
Concentration (@ml)
Collagenase (units/ml)
TIMP (units/ml)
0 0.01 0.1 1.0 0 0.1 1.0 10.0 0 0.1 1.0 10.0 50.0 0 1.0 10.0 50.0 0 1.0 10.0 50.0
0.37 0.44 0.27 0.16 0.82 0.80 0.64 0.51 0.42 0.40 0.38 0.24 0.08 1.05 0.92 0.87 0.40 1.05 1.12 0.65 0.39
0 0 0 0.51 0 0 0 0.54 0 0 0 0.41 2.78 0 0 0 1.77 0 0 0.07 0.05
--f + + + + k f + f + k + + + f + + + f + f
0.01 0.13 0.01 0.01 0.09 0.06 0.04 0.01 0.01 0.02 0.01 0.01 0.02 0.14 0.06 0.17 0.04 0.14 0.04 0.06 0.06
+ 0.31
+ 0.25
zk 0.08 + 0.10
+ 0.21
f 0.07 + 0.04
Mouse calvarial osteoblasts were cultured on “C-labelled type I collagen films for 48 h in 16 mm wells containing DMEM plus 2% acid treated rabbit serum and stimulated with 1,25(OH),D, (10 ng!ml). Growth factors were added at the concentrations shown. Culture supernatants were assayed for collagenase and TIMP activity by fibril assay. Each point represents the mean + SEM for triplicate cultures.
Table 4 Effect of insulin-like growth factors on collagenase and TIMP activity by 1,25(OH),D,-stimulated mouse calvarial osteoblasts Addition
IGF-I
IGF-II
(/ml)
Collagenase (units/ml)
TJMP (units/ml)
0 O*lpg I*Oclg 1o.ojcg 0 10.0 ng 50.0 ng 100.0 ng
1.42 1.41 1.36 0.69 1.42 1.14 1.01 1.06
0.02 + 0.01 0 0 0 0.02 + 0.01 0 0 0
Concentration
f k + f f f + +
0.18 0.03 0.14 0.05 0.18 0.09 0.06 0.10
Mouse calvarial osteoblasts were cultured on i4C-labelled type I collagen films for 48 h in 16 mm wells containing DMEM plus 2% acid treated rabbit serum and stimulated with 1,25(OH),D, (10 nglml). The IGFs were added at the concentrations shown. Culture supernatants were assayed for collagenase and TIMP activity by fibril assay. Each point represents the mean + SEM for triplicate cultures.
52 and 50 ng/ml respectively (Fig. 6), and 100 ng/ml for both aFGF and bFGF (Fig. 7). Inhibition of collagenolysis was accompanied by a reduction of collagenase activity and the appearance of free TIMP in the culture medium at the higher doses tested (Table 3). Treatment of unstirnulated osteoblasts with growth factors alone had no effect on collagenase activity, but exerted stimulatory effects on TIMP production (data not shown). Although both IGFs partially inhibited 1,25(OH),D,-stimulated collagenase production by osteoblasts (Table 4), this had no significant effect on type I collagenolysis over the 48 h time-course of the experiment (data not shown); unlike the other growth factors however, the IGFs did not stimulate TIMP synthesis by either 1,25(0H),D,-treated osteoblasts (Table 4) or unstimulated cells (data not shown). From these data we conclude that inhibition of collagenolysis was primarily dependent upon the ability of growth factors to stimulate TIMP synthesis.
Discussion The present series of experiments demonstrate an important role for growth factors in regulating the activity of collagenase and TIMP and type I collagen degradation by mouse calvarial osteoblasts. EDTAITris-HCl extracts of bone matrix exerted a dose-dependent inhibition of type I collagenolysis by osteoblasts stimulated with 1,25(OH),D,. Inhibition was accompanied by a reduction in collagenase activity and an increase in free TIMP in the culture medium. Although growth factor activity in our bone matrix preparation was only partially characterized, EDTAITris-HCl extracts of bovine bone have previously been shown to contain TGF-P, PDGF and FGF activity [14], and we felt it impor?ant to initially test its effect on the proteolytic behaviour of osteoblasts; target cells are likely to be exposed to mixtures rather than individual gro*Nthfactors in vivo. The role of growth factors in the regulation of MMP and TIMP expression by connective tissue cells is complex. The response of the cell type under investigation also appears to be variable. EGF, PDGF and TGF$ have been shown to stimulate collagenase synthesis in human fibroblast cultures [26,27], and Edwards et al. [28] found that exposure of quiescent human MRCJ fibroblasts to bFGF and EGF induced the expression of both collagenase and TIMP mRNA and protein. TGF#l alone had little effect on the expression of these genes, but in the presence of other growth factors had a reciprocal effect on MMP expression; TGF-P repressed the induction of collagenase and interacted with bFGF and EGF to super-induce TIMP expression [28]. Overall et al. also found that TGF-/? suppressed collagenase and increased TIMP and PAI-1 expression by human gingival fibroblasts [29] and rat calvarial osteoblasts [20]. TGF-/3 did, however, stimulate the synthesis of another MMP (72-kDa progelatinase) by both cell types [20,29]. In the present model system, TGF-/3, EGF, PDGF and both acidic and basic FGF, all inhibited collagenase and stimulated TIMP activity by mouse osteoblasts cultured simultaneously with I,25(OH),D,. In contrast the IGFs, which did not inhibit type I collagenolysis, only partially inhibited collagenase activity and did not stimulate TIMP production by
53 either 1,25(0H),D,-treated or untreated cells. These findings suggest that rnhibition of collagenolysis was dependent primarily upon the ability of certain growth factors to stimulate TIMP synthesis. There are disadvantages in relying on the fibril assay for estimating cohagenase and TIMP synthesis by cells in culture: low levels of collagenase are likely to remain undetected in supernatants containing relatively high levels of TIMP. TIMP does not interact with latent enzyme, but once the enzyme is activated by including APMA in the assay it is avidly bound by free TIMP, yielding irreversible collagenase-TIMP complexes which cannot be detected in activity assays [30]. TIMP production will be similarly underestimated in supernatants where the level of naturally activated enzyme exceeds that of TIMP. At no time could active collagenase be detected in culture supernatants from any of the present series of experiments, even though coliagenolysis had occurred. However, we have previously shown that for mouse osteoblasts to degrade type I collagen films, only a minute proportion of the latent collagenase secreted needs to be activated by plasmin [12,13]. Furthermore, although it has not been specifically demonstrated for osteoblasts, some cells possess a specific binding site for urokinase-type plasminogen activator (u-PA) and its proform (for review see Ref. 31), which allows cells to acquire surface bound plasminogen-activating ability. The confinement of u-PA to discrete sites may thus explain how cells which constitutively produce TIMP and PAL1 can still produce focal proteolysis. The finding that rodent osteoblasts synthesize collagenase [7-91, has been responsible for the hypothesis that bone cells play a major role in bone resorption by degrading the surface osteoid layer, thereby preparing the underlying mineralized bone for osteociastic action [6,10,11]. Recent evidence however, suggests that this model might not apply to bone resorption in the human. Human cells of diverse tissue origin have been shown to produce collagenase, and although human osteoblasts secrete gelatinase as a major cellular product they do not produce collagenase in significant quantities, either constitutively or in response to bone resorbing agents [32,33]. While it would be unusual to expect different cellular and/or molecular mechanisms to have evolved for the resorption of human bone, the reasons for these cellular and species differences remain unclear at present, but are the subject of current investigation. In conclusion, the findings of the present series of experiments provide additional evidence for the tight control exerted on the proteolytic activity of osteoblasts and the importance of TIMP in its regulation. They further suggest that the conversion (coupling) of the initial resorptive phase of the bone remodeiling cycle to one of deposition, may be mediated by polypeptide growth factors either produced locally by osteo!riasts, or released by proteolysis from the bone matrix itself.
Acknowledgements Supported by grants from the Arthritis and Rheumatism Council and the Medical Research Council. We thank Christopher Green and Judith Webdell for preparing
54
the illustrations and Elaine Pocock and Angela McMonagle for secretarial assistance.
References 1 Parfitt AM. The coupling of bone formation to bone resorption: a critical analysis of the concept and of its relevance to the pathogenesis of osteoporosis. Metab Bone Dis Rel Res 1982;4:1-6. 2 Frost HM. Dynamics of bone remodelling. In: Frost HM, ed. Boue biodynamics. Boston: Little Brown & Co., 1964. 3 Raina V. Normal osteoid tissue. J Clin Path01 1972;25:229-232. 4 Fornasier VL. Transmission electron microscopy studies of osteoid maturation. Metab Bone Dis Rel Res 1980;25:103-108. 5 Vanderveil CJ. An ultrastructural study of the components which make up the resting surface of bone. Metab Bone Dis Rel Res 1980;28:109-116. 6 Sakamoto M, Sakamoto S. Bone collagenase, osteoblasts and cell-mediated bone resorption. In: Peck WA, ed. Bone and Mineral Research, Vol. 4. Amsterdam: Elsevier, 1986;49-102. 7 Heath JK, Atkinson SJ, Meikle MC, Reynolds JJ. Mouse osteoblasts synthesize collagenase in response to bone resorbing agents. Biochim Biophys Acta 1984;802:151-154. 8 Sakamoto M, Sakamoto S. Immunocytochemical localization of collagenase in isolated mouse bone cells. Biomed Res 1984;5:29-38. 9 Gtsuka K, Sodek J, Limeback H. Synthesis of collagenase and collagenase inhibitors by osteoblastlike cells in culture. Eur J Biochem 1984;145:123-129. 10 Chambers TJ, Darby JA, Fuller K. Mammalian collagenase predisposes bone surfaces to osteoclastic resorption. Cell Tissue Res 1985;241:671-675. 11 Chambers TJ, Fuller K. Bone cells predispose bone surfaces to resorption by exposure of mineral to osteoclastic contact. J Cell Sci 1985;76:155- 165. 12 Thomson BM, Atkinson SJ, Reynolds JJ, Meikle MC. Degradation of type I collagen films by mouse osteoblasts is stimulated by 1,25-dihydroxyvitamin D, and inhibited by human recombinant TIMP (tissue inhibitor of metalloproteinases). Biochem Biophys Res Commun 1987;148:596-602. 13 Thompson BM, Atkinson SJ, McGarrity AM, Hembry RM, Reynolds JJ, Meikle MC. Type I collagen degradation by mouse calvarial osteoblasts stimulated with 1,25_dihydroxyvitamin D-3: evidence for a plasminogen-plasmin-metalloproteinase cascade. Biochim Biophys Acta 1989;1014: 125- 132. 14 Hauschka PV, Mavrakos AE. Iafrati MD, Doleman SE, Klagsburn M. Growth factors in bone matrix, Isolation of multiple types by affinity chromatography on heparin-sepharose. J Biol Chem 1986:261:12665- 12674. 15 Centrella M, Canalis E. Transforming and nontransforming growth factors are present in medium conditioned by fetal rat calvariae. Proc Nat1Acad Sci USA 1985;82:7335-7339. 16 Mohan S, Jennings JC, Linkhart TA, Baylink DJ. Primary structure of human skeletal growth factor: homology with human insulin-like growth factor-II. Biochim Biophys Acta 1988;966:44-55. 17 Frolik CA, Ellis LF, Williams DC. Isolation and characterization of insulin-like growth factor-11 from human bone. Biochem Biophys Res Commun 1988;151:1011-1015. 18 Partridge NC, Jeffrey JJ, Ehlich LS, Teitelbaum SL, Fliszar C, Welgus HG, Kahn AJ. Hormonal regulation of the production of collagenase and a collagenase inhibitor activity by rat osteogenic sarcoma cells. Endocrinology 1987;120:1956-1962. 19 Delaisse JM, Eeckhout Y, Vaes G. Bone-resorbing agents affect the production and distribution of procollagenase as well as the activity of collagenase in bone tissue. Endocrinology 1988;123:264-276. 20 Overall CM, Wrana JL, Sodek J. Transforming growth factor-p regulation of collagenase, 72 kDaprogelatinase, TIMP and PAI- expression in rat bone cell populations and human fibroblasts. Connect Tissue Res 1989;20:289-294. 21 Shen V, Kohler G, Jeffrey JJ, Peck WA. Bone-resorbing agents promote and interferon-y inhibits bone cell collagenase production. J. Bone Mineral Res 1988;3:657-666.
55 22 Gospodarowicz D. Locahsation of a fibroblast growth factor aild its effect alone and with hydrocortisane on 3T3 cell growth. Nature 1974;249:123- 127. 23 Twardzik DR, Ranchalis JE, Todaro GJ. Mouse embryonic transforming growth factors related to those isolated from tumor cells. Cancer Res 1982;42:590-593. 24 Gavrilovic J, Reynolds JJ, Murphy G. Inhibition of type I collagen film degradation by tumor cells using, a specific antibody to collagenase and the specific tissue inhibitor of metalloproteinases (TIMP). Cell Biol Int Rep 1985;9:1097-1107. 25 Cawston TE, Barrett AJ. A rapid and reproducible assay for collagenase using [l-‘4C]-acetylated collagen. Anal Biochem 1979;99:340-345. 26 Bauer EA. Cooper TW, Huang JS, Altman J, Deuel TF, Stimulation of in vitro human skin collagenase expression by platelet-derived growth factor. Proc Nat1 Acad Sci USA 1985;82:4132-4136. 27 Chua CC, Geiman DE, Keller GH, Ladda RL. Induction of collagenase secretion in human fibroblast cultures by growth promoting factors. J Biol Chem 1985;260:5213-5216. 28 Edwards DR, Murphy G, Reynolds JJ, Whitham SE, Docherty AJF, Angel P, Heath JK. Transforming growth factor beta modulates the expression of collagenase and metalloproteinase inhibitor. EMBO J 1987;6:1899- 1904. 29 Overall CM, Wrana JL, Sodek J. Independent regulation of collagenase, 72-kDa-progelotinase, and metalloendoproteinase inhibitor expression in human fibroblasts by transforming growth factor-D. Proc Nat1 Acad Sci USA 1989;264:1860-1869. 30 Cawston TE, Murphy G, Mercer E, Galloway WS, Hazelman BL, Reynolds JJ. The interaction of purified rabbit bone collagenase with purified rabbit bone metalloproteinase inhibitor. Biochem J 1983;211:313-318. 31 Blasi F, Vassalli JD, Dano K. Urokinase-type plasminogen activator: proenzyme, receptor, and inhibitors. J Cell Biol1987;104:801-804. 32 Gowen M, Wood DD, Ihrie EJ. Meats JE, Russell RGG. Stimulation by human interleukin 1 of cartilage breakdown and production of collagenase and proteoglycanase by human chondrocytes but not human osteoblasts in vitro. Biochim Biophys Acta 1984;797:186-193. 33 Rifas L. Halstead LR, Peck WA, Avioli LV, Welgus HG. Human osteobiasts in vitro secrete tissue inhibitor of metalloproteinases and gelatinase but not intersiitial coilagenase as major cellular products. J Clin Invest 1989;84:686-694.
