Mouse osteoblasts synthesize collagenase in response to bone resorbing agents

Biochimica et Biophysica Acta, 802 (1984) 151-154

151

Elsevier

BBA Report BBA 20084

MOUSE OSTEOBLASTS SYNTHESIZE COLLAGENASE IN RESPONSE TO BONE RESORBING AGENTS JOAN K. HEATH a SUSAN J. A T K I N S O N a, M U R R A Y C. MEIKLE a,b and JOHN J. REYNOLDS a..

a Cell Physiology Department, Strangeways Research Laboratory, Worts Causeway, Cambridge, CB1 4RN, and b Department of Orthodontics, Eastman Dental Hospital and Institute of Dental Surgery, University of London, London

tU.K.) (Received June 14th, 1984)

Key words: Collagenase synthesis," Bone resorbing agent; Parathyroid hormone; cAMP; (Mouse osteoblast)

Bone cells isolated from mouse calvariae by a sequential digestion procedure have many osteoblast characteristics: they respond to PTH and prostaglandin E z by activation of adenylate cyclase hut not to calcitonin, they stain for alkaline phosphatase and they make only type I collagen. In confluent monolayer culture, they do not secrete collagenase in appreciable quantities, unless stimulated with resorptive substances such as PTH, prostaglandin E2, 1,25(OH)2 vitamin D-3 and monocyte-conditioned medium. This suggests they play a direct role in bone resorption.

Much recent evidence suggests that osteoblasts, as well as being important in bone matrix synthesis, play an indirect role in bone resorption. Studies with monolayer cultures have shown that osteoblasts possess functional receptors for the bone resorbing agents parathyroid hormone (PTH) [1,2], vitamin D [3,4] and prostaglandins [5]. Also, the ability of various prostaglandin analogues to stimulate cyclic-AMP accumulation in osteoblastlike osteosarcoma-derived cells was shown to be closely correlated with the resorptive activity of these analogues in tissue culture [5]. These findings indicate that osteoblasts play a role in the resorptive process by mediating the signal of local and systemic factors to osteoclasts, and it has been suggested that agents like PTH might work by inducing a shape change in osteoblasts, thereby enabling access of osteoclasts to the bone surface where they can initiate demineralization [6]. We now present evidence that osteoblasts may have a direct role in the process of resorption. * To whom correspondence should be addressed. 0304-4165/84/$03.00 © 1984 Elsevier Science Publishers B.V.

Osteoblast-enriched populations of bone cells prepared from neonatal mouse calvariae and treated with various resorbing agents respond by secreting mammalian collagenase. Unstirnulated cells release low amounts of collagenase, but do secrete tissue inhibitor of metalloproteinases (TIMP) [7], which is thought to modulate metalloproteinase activity in vivo. Since collagenase is specific in its ability to degrade collagen fibrils at neutral pH, its synthesis by osteoblasts strongy implicates these cells in the extracellular resorption of bone matrix. Our first indication that bone resorption was not the specific responsibility of the osteoclast came with experiments showing that the release of active collagenase from mouse calvarial explants in response to vitamin A was not abolished if salmon calcitonin was present in the culture medium [8]. Since then, we have also observed that vitamin A is able tostimulate the release of active and latent collagenase from calvarial explants taken from osteopetrotic (grey-lethal) mice [9] that do not resorb. These studies indicated that cells other than osteoclasts contribute to collagenase produc-

152 tion. To clarify the role played by osteoblasts in resorption, we prepared monolayer cultures of osteoblast-enriched bone cells from neonatal mouse calvariae by a sequential digestion procedure. About 50 calvariae were carefully stripped of periostea, blood vessels and soft connective tissue under a dissecting microscope and minced with fine scissors prior to a prewash in dissecting medium (a modified form of BGJ medium containing a lower bicarbonate level so as to equilibrate with air [10]). The dissecting medium was then aspirated and the bone initially digested at 37°C in 5 ml of Ca 2+and Mg2+-free Tyrode's solution containing 1 m g / m l trypsin (Sigma) for 10 rain with gentle stirring. The supernatant was discarded, the bones were washed with Tyrode's solution, and the digestion was continued in 5 ml of 2 m g / m l dispase (Boehringer) for 30 min. The supernatant was again discarded, and the residue was further digested in 2 m g / m l bacterial collagenase (Worthington Type II) containing 4 mM E D T A for two periods of 30 min. Cells released from the bone matrix during each of these two final incubations were harvested and pooled, washed with Tyrode's solution and plated out at a density of 5 . 1 0 5 cells per 30-mm petri dish in 1.5 ml of Dulbecco's modification of Eagle's medium containing 10% foetal calf serum. The cells were identified as osteoblasts by morphological cri.'teria and by the fact that essentially all the cells stained strongly for alkaline phosphatase (Fig. 1), synthesised type I collagen (no type II or III), and accumulated cyclic A M P in response to PTH and prostaglandin E z (Fig. 2), but not to salmon calcitonin or 1,25(OH)2 vitamin D-3. In order to assay for collagenase and tissue inhibitor of metalloproteinases production the bone cells (primary or first passage) were allowed to reach confluence (4-5 days) in 30-mm petri dishes with 1.5 ml Dulbecco's modification of Eagle's medium containing 10% FCS. The serum-containing medium was aspirated, the cells were washed carefully with serum-free medium and then incubated with 1.5 ml medium containing 0.5% bovine serum albumin (Sigma) either with or without the various resorbing agents. The medium was harvested and replaced with fresh medium every 2 days. The synthesis of collagenase and tissue inhibitor of metalloproteinases was measured for each prep-

~J4

Fig. 1. Photomicrographsof osteoblast-enriched populations of mouse calvarial bone-cells. 5-105 cells were seeded into 3-cm dishes containing glass coverslips. (A) Near-confluent cells fixed in methanol and stained with haematoxylin and eosin (x 300). (B) Confluent cells fixed in 60% citrate-buffered acetone and stained with methylene blue (X 600). (C) Alkaline phosphatase staining of mouse osteoblasts using naphthol ASMX phosphate (Histozymekit No. 85L-1R, Sigma). Confluent cells were fixed in 60% citrate-buffered acetone and incubated in a solution containing naphthol AS-MX phosphate. All'aline phosphatase activity is indicated by the clark spots which represent sites of naphthol AS-MX liberation. This has become coupled with a diazonium salt (Fast Violet B salt), forming discrete loci of insoluble, visible pigment ( x 3000). The plump, cuboid cells are like the rabbit osteoblasts demonstrated by Yee [211. aration of bone cells, and the levels of activity in medium containing PTH (1-4 U / m l ) , 1,25(OH) 2 vitamin D-3 (between 1 0 - 8 - 1 0 6 M), prostaglan-

153 1.4-

6O

•~

1.2

-

++

tk 4C

1.0-

< o

q~

o

++

o.+

-6

0.6

o.

20

+> -.=

m 0.4 ..i.

C

PIH 2 4

PGE 2

2

D3

~

MGM

mmm CT

5

Fig. 2. Intracellular accumulation of cAMP by osteoblasts in response to parathyroid hormone (PTH). 1.105 cells were established in miniwells of area 2 cm e (Linbro, Flow Labs) and allowed to reach confluence (4-5 days). The cells were washed and exposed to 100 p,M IBMX (3-isobutylmethyl-l-xanthineSigma) in fresh modified Eagle's medium for 10 min. They were then incubated for 6 min with fresh medium containing 1 0 0 / t M IBMX plus any additions: [] 2 and 4 U / m l PTH, [] 2 and 5 # g / m l prostaglandin E 2 (PGE2), • 1.10 -7 M 1,25(OH)2 vitamin D-3 (D3) , [] 100/~l/ml monocyte-conditioned medium (MCM) or [] 1 m U / m l salmon calcitonin (CT). Intracellular c A M P was then extracted with acidified ethanol overnight. The extracts were evaporated in a dessicator and the residues redissolved in 100 m M T r i s / 8 m M theophylline/6 m M 2mercaptoethanol/20 m M magnesium s u l p h a t e / 4 m M E D T A (pH 7.4). The c A M P was assayed by the competitive binding method of Brown et al. [22], using bovine adrenal glands as source of specific binding protein. C, control cells. Values are m e a n s for six wells_ S.E. of the mean.

din E 2 (between 1-5 /~g/ml) were compared to those of control (untreated) samples. In addition the effect of medium conditioned by human peripheral monocytes, which is known to contain a non-dialysable resorbing factor [11], was tested (100/~l/ml). Unstimulated osteoblasts were found to produce low levels of collagenase, but consistently released relatively large amounts of tissue inhibitor of metalloproteinases (0.5-2.0 U/ml). However, treatment of these cells with known resorbing agents induced the production of much higher levels of collagenase in the latent form (Fig. 3). The reaction of activated enzyme [7] with collagen was analysed by SDS-polyacrylamide gel electrophoresis according to the method of Laemmli and Favre [12]. The gel (not shown) demonstrated that rat skin type I collagen was cleaved

E 0.2-

--'e

r 0-2

[ 2-4 D a y s in culture

4-16

Fig. 3. S t i m u l a t i o n o f collagenase release b y mouse osteoblasts

in response to various resorbing agents. Collagenase activity was measured by the release of []4C]acetylated peptides from thermally reconstituted, trypsin-resistant fibrils of []4C]acetylated rat skin collagen either in the presence or absence of 0.67 m M 4-aminophenylmercurie acetate. In the absence of 4aminophenylmercuric acetate, only already-active collagenase is detectable, whereas in its presence, latent collagenase activity is unmasked. 1 U of collagenase is that activity which degrades 1 ~g of reconstituted collagen fibrils in 1 rain at 35°C. Tissue inhibitor of metalloproteinases (TIMP) activity was assayed by adding samples of medium to a known a m o u n t of already-active collagenase (0.05 U) in the assay system outlined above. 1 U of tissue inhibitor activity is defined as that activity which inhibits 2 U of collagenase by 50% [23]. Cumulative collagenase activity is expressed as the mean for six dishes _+S.E. of mean. All values in the presence of resorbing agent are significantly different from controls ( P < 0.05 at least). • •, 4 U/ml PTH; • • , 1 - 10 -7 M 1,25(OH)2 vitamin D-3; [] [], 5 /.tg/ml prostaglandin E2; O O, unstimulated; • • , 100 /.d/ml monocyte-conditioned medium; Lx ,% control for monocyte-conditioned medium.

into 3 / 4 and 1 / 4 pieces (TCa and TCb fragments), confirming the identity of the osteoblast product as mammalian collagenase [13]. Predictably, the presence of 2 mM 1,10-phenanthroline or 5 mM EDTA in the assay mixture completely prevented the generation of 3 / 4 and 1 / 4 pieces, whereas diisopropylphosphorofluoridate, a serine proteinase inhibitor, had no effect. The response of osteoblasts to resorbing agents in terms of tissue inhibitor of metalloproteinases production was less

154

dramatic, though PTH consistently raised the tissue inhibitor levels. Though the response of osteoblasts to resorbing agents in many ways mimics the responses observed for intact bone, one difference is that the collagenase released by osteoblasts is in the latent form. We favour the idea that latency is best explained in terms of collagenase being released in a form requiring activation before it can degrade collagen [14]. Though the mechanism of activation is unknown, enzymes such as plasmin [15] and kallikrein [16] are capable of activating latent enzyme. The failure of a pure population of osteoblasts to generate active enzyme in the culture medium implies the involvement of other cells in bone, possibly osteoclasts. Once activated, collagenase becomes susceptible to irreversible interaction with tissue inhibitor of metalloproteinases and we propose that this would provide the necessary restraint to excessive extracellular matrix degradation in normal turnover conditions [17]. Recently, a study utilising osteoblasts originally grown out from explants of human bone trabeculae, failed to detect the synthesis of collagenase by these cells in response to a partially-purified preparation of human interleukin-1 [18] known to have bone resorptive activity [19]. This result was interpreted as evidence against osteoblasts playing a direct role in bone resorption. However, other resorptive substances, like PTH and prostaglandin E 2, were not tried in this system and interestingly, we have found that pure catabolin [20] does not stimulate collagenase release from our population of osteoblasts either (Atkinson, S.J., Heath, J.K. and Saklatvala, J., unpublished data). Since catabolin and interleukin-1 may be closely-related cytokines, the results of the two studies are not contradictory. Our data clearly suggest that it is no longer appropriate to consider osteoblasts as only bonesynthesising cells, because they can secrete collagenase in response to resorbing agents. We propose that osteoblasts are the target cells for resorbing agents, capable of responding directly to the catabolic signal. There may be a complex interaction between osteoblasts and osteoclasts, and studies utilising mixed populations of purified bone cells should contribute to our understanding of the resorptive process including the role of collagenase and the mechanism of its activation.

We would like to acknowledge the generous gift of 1,25(OH)2 vitamin D-3 from Dr. David Fraser, Dunn Nutritional Laboratory, University of Cambridge and Medical Research Council, Milton Road, Cambridge. These studies were funded by the Medical Research Council. References 1 Luben, R.A., Wong, G.L. and Cohn, D.V. (1976) Endocrinology 99, 526-534 2 Peck, W.A., Burke, J.K., Wilkins, J., Rodan, S.B. and Rodan, G.A. (1977) Endocrinology 100, 1357-1364 3 Manolagas, S.C., Haussler, M.R. and Deftos, L.J. (1980) J. Biol. Chem. 255, 4414-4417 4 Partridge, N.C., Frampton, R.J., Eisman, J.A., Michelangeli, V.P., Elms, E., Bradley, T.R. and Martin, T.J. (1980) FEBS Lett. 115, 139-142 5 Martin, T.J. and Partridge, N.C. (1981) in Hormonal Control of Calcium Metabolism (Cohn, D.V., Talmage, R.V. and Matthews, J.L., eds.), pp. 147-156, Excerpta Medica, Amsterdam 6 Rodan, G.A. and Martin, T.J. (1981) Calcif. Tissue Int. 33, 349-351 7 Sellers, A., Murphy, G., Meikle, M.C. and Reynolds, J.J. (1979) Biochem. Biophys. Res. Commun. 87, 581-587 8 Sellers, A., Meikle, M.C. and Reynolds, J.J. (1980) Calcif. Tissue Int. 31, 35-43 9 Heath, J.K., Sellers, A., Meikle, M.C. and Reynolds, J.J. (1982) J. Dent. Res. 61,544 10 Reynolds, J.J. (1976) in Organ Culture in Biomedical Research (Balls, M. and Monnickendam, H., eds.), pp. 355-366, Cambridge University Press, Cambridge 11 Gowen, M., Meikle, M.C. and Reynolds, J.J. (1983) Biochim. Biophys. Acta 762, 471-474 12 Laemmli, U.K. and Favre, M. (1973) J. Mol. Biol. 80, 575-599 13 Werb, Z. and Reynolds, J.J. (1975) Biochem. J. 151,645-653 14 Nagase, H., Jackson, R.C., Brinkeroff, C.G., Vater, C.A. and Harris, E.D., Jr. (1981) J. Biol. Chem. 256, 11951-11954 15 Eeckhout, Y. and Vaes, G. (1977) Biochem. J. 166, 21 31 16 Nagase, H., Cawston, T.E., De Silva, M.D. and Barrett, A.J. (1982) Biochim. Biophys. Acta 702, 133-142 17 Cawston, T.E., Murphy, G., Mercer, E., Galloway, W.A., Hazleman, B.L. and Reynolds, J.J. (1983) Biochem. J. 211, 313-318 18 Gowen, M., Wood, D.D., Ihrie, E.J., Meats, J.E. and Russell, R.G.G. (1984) Biochim. Biophys. Acta 797, 186-193 19 Gowen, M., Wood, D.D., Ihrie, E.J., McGuire, M.K.B. and Russell, R.G.G. (1983) Nature 306, 378-380 20 Saklatvala, J., Curry, V.A. and Sarsfield, S.J. (1983) Biochem. J. 215, 383-392 21 Yee, J.A. (1983) Calcif. Tissue Int. 35, 571-577 22 Brown, B.L., Albano, J.D.M., Ekins, R.P. and Sgherzi, A.M. (1971) Biochem. J. 121, 561-562 23 Murphy, G., Cartwright, E.C., Sellers, A. and Reynolds, J.J. (1977) Biochim. Biophys. Acta. 483, 394-498