- Email: [email protected]
“Matrigenin” activity from bovine bone—I. Partial purification of activity
Comp. Blochem. Physiol. Vol. 88B, No. 2, pp. 529-534, 1987 Printed in Great Britain
0305-0491/87 $3.00+ 0.00 © 1987 Pergamon Journals Ltd
"MATRIGENIN" ACTIVITY FROM BOVINE BONE--I. PARTIAL PURIFICATION OF ACTIVITY* D. S. IRWIN and T. P. ANASTASSIAOES Department of Medicine and Biochemistry and the Rheumatic Diseases Unit, Queen's University, Etherington Hall, Stuart Street, Kingston, Ontario, Canada K7L 3N6 (Tel.: 613-545-2971) (Received 16 December 1986)
Abstract--1. Bovine bone contains an extractable activity which stimulated the synthesis of glycosaminoglycans by bovine synovial, human synovial and mouse 3T3 fibroblastic cells in culture. Human cells were used to develop an assay for purification of the stimulatory activity ("matrigenin" activity) from bovine bone. 2. Partial purification of "matrigenin" activity was achieved by precipitation of the EDTA extract at pH 3.5 and Sepharose CL-6B chromatography in 4 M guanidinium HC1. Dissociative conditions were necessary to prevent aggregation. 3. On SDS-polyacrylamide gel electrophoresis the activity ran with a mobility equivalent to a Mr = 27,500 and could be recovered from the SDS gels.
In this paper we report on a relatively rapid assay for the stimulation of glycosaminoglycan synthesis by cultured fibroblastic cells. The conditions of the assay were designed to minimize effects of the bone extracts on fibroblastic proliferation. We refer to this stimulatory activity on extracellular glycosaminoglycan synthesis as "matrigenin" activity (Anastassiades and Irwin, 1985), which is an operational term based on the tissue culture assay used for purification of the activity from bovine bone.
INTRODUCTION
It has been known for a long time that implantation of bone fragments into the muscle or skin of mammals leads to the formation of new cartilage and bone in the host (Barth, 1893; Bonome, 1886). This type of experiment can be done with xenogeneic mammalian bone and has been studied extensively in rodents (Urist, 1965; Huggins et al., 1970), where the early biochemical events are characterized by an increase in hyaluronic acid synthesis, followed by increased cartilage-type proteoglycan synthesis (Reddi, 1981). We had reported (Anastassiades et al., 1984) that an activity that stimulated glycosaminoglycan synthesis by cultured mammalian fibroblastic cells could be extracted with a solution of 0.34 M EDTA in 2.5 M NaC1 from bovine but not human bone matrix. Subsequent extraction of the residues with 4 M guanidinium HC1 yielded activity from both matrices. In addition, synovial fibroblastic cells from man and rabbit were evaluated for their ability to respond to the stimulatory effects of the extracts by increased glycosaminoglycan synthesis. The human synovial cells constituted a sensitive assay for the glycosaminoglycan stimulatory activity (Anastassiades et al., 1984) and we have used these cells routinely as they are easily available to us. We had previously reported (Anastassiades et al., 1978b, 1982; Anastassiades and Starkey, 1983) that a similar stimulatory activity could be extracted from rat bone using 4 M guanidinium HC1 solutions. Fibroblastic cells from neonatal rat muscle were used as the assay system in that work. The bovine bone provided advantages as a source for purification of this activity that included ready availability of large quantities of bone and the relative ease of extraction of the activity compared to rat or human bone.
MATERIALS AND METHODS
Preparation of bone powder Bovine femurs and tibias obtained from a local slaughterhouse were stored at -20°C. They were cleaned and the cortical bone was powdered, using a Wiley Mill (A. H. Thomas, Philadelphia), to a size less than 1 nun (20 mesh screen).
*Supported by Grants from the Arthritis Society, the Medical Research Council of Canada and an Arthritis Society Fellowship to Dr D. S. Irwin.
EDTA extraction of bone powder Extraction was done by a modification of the method of Herring (Herring et al., 1974) but with the addition of protease inhibitors. Bone powder (0.7 g/ml) was placed in Spectrapor 1 dialysis bags (Spectrum Medical Industries, Los Angeles) in 2.5 M NaC1, 100raM 6-aminohexanoic acid, 5mM benzamidine, l mM phenylmethylsulfonyl fluoride and dialyz_~dagainst a solution of 0.34 M EDTA adjusted to pH 7.4 with NaOH and containing the same concentrations of inhibitors as in the bag (38 ml of solution per g of bone powder). The bags were intermittently agitated. After 4 days the liquid in the bags was removed, saved and replaced with fresh solution of 2.5 M NaC1 containing the inhibitors and the bags again placed in a fresh EDTA solution plus the inhibitors for a further extraction period of 4 days. At the end of the second extraction period the combined supernatants were filtered. The clear, yellow filtrate was then dialyzed against three changes of 25 mM 6-aminohexanoic acid, 0.2 mM benzamidine in water and two changes of plain water. All procedures were done at 4°C.
529
530
D . S . IRWIN and T. P. ANASTASSIADES
Acid precipitation
Routine assay for "matrigenin " activity
The EDTA extract was made 100 mM in glycine HC1 buffer, pH 3.5 at 0°C. In some cases the pH was adjusted to 3.0 with HC1 to complete the precipitation. The suspension was centrifuged at 0°C, the supernatant discarded and the precipitate was resuspended in a small volume of Dulbecco's phosphate buffered saline (PBS). The suspension was dialyzed against water and freeze-dried. This is termed the "acid precipitate".
Human synoviai fibroblasts were obtained from synovial explants, subcultured and grown to confluency in 35 mm dishes, containing CMRL 1969 medium supplemented with 20% fetal calf serum (FCS), by methods we have previously described (Anastassiades et al., 1978a; Anastassiades and Wood, 1981). Freeze dried samples of the fractions to be tested (and "control" fractions) were dissolved in fresh medium with 10% fetal calf serum, containing also 1.25 #Ci/ml of D-{1,6-3H(N)} glucosamine (43.2 Ci/mmol) and 2/~Ci/ml of (35S) Na2SO4 (New England Nuclear Co). The old medium was removed and the fresh, labeled media were used to feed the cultures (1.5 ml medium per dish). After 38-40 hr of incubation, the cell surface-associated material was obtained by trypsinization (Hronowski and Anastassiades, 1980) and cells were removed by centrifugation. To the supernatant 100 #l of solution containing the "carrier" giycosaminoglycan consisting of 1 mg/ml of hyaluronic acid (human umbilical cord, Grade III-S, Sigma Chemical Co.) and 1 mg/ml of ehondroitin sulfate (whale or shark cartilage, grade III, Sigma Chemical Co.) were added. Sixteen millilitres of absolute ethanol were added with thorough mixing and the tube left several hours at 4°C. The precipitate was centrifuged and the supernatant discarded. The precipitate was then dissolved in 2.0ml, 0.075 M Na2EDTA-0.005M cysteine HC1 buffer, pH6.2. Undissolved material was removed by centrifugation and the supernatant was poured into 20 ml counting vials. Two hundred microlitres of 5% cetylpyridinium chloride (CPC) in 0.2M Na2SO 4 was added then 5.0ml H20 and the contents were left overnight at room temperature. The precipitated glycosaminoglycans and proteogiycans were centrifuged at 10,000g for 10 min at 20°C. The supernatant was poured off, l0 ml of Atomiight (New England Nuclear Co.) added and the radioactivity measured by liquid scintillation counting. The results for 3H or 35S incorporation for the test fractions are expressed in terms of units per #g protein added per ml of medium. A unit = [(dpm incorporated by cultures to which test fractions were added/dpm incorporated by control cultures) - 1 x 100]. This simplified assay procedure was validated against the longer procedure, which included extensive papain digestion (Hronowski and Anastassiades, 1980) and essentially identical values were observed for the incorporation of radioactivity into the glycosaminoglycan CPC precipitate from the cell surface and for the glycosaminoglycan isolated by the longer procedure. The period of time chosen for labelling the cultures was based on maximal labelling of the cell surface glycosaminoglycan (Anastassiades et al., 1984) and on convenience.
Sephacryl S-200 gel fihration For non-dissociative chromatography a 90 x 1.6 cm column of Sephacryl S-200 was equilibrated with Dulbecco's PBS, pH 7.2, with 0.02% sodium azide. Three milligrams of the acid precipitate was dissolved in 1 ml of this buffer and this was run on the column and eluted at l0 ml/hr. Five ml fractions were collected, dialyzed and aliquots analyzed for protein and assayed for "matrigenin" activity as described below. For dissociative chromatography, the same column was equilibrated with 4 M guanidinium HC1, 50 mM Tris buffer, pH 7.4. Material that had been extracted with 4 M guanidinium HC1, 0.34 M NaEDTA, 50 mM Tris, pH 7.2, was concentrated to 6 mi with Amicon PM-10 and applied to the column.
Sepharose Cl-6B gel filtration under dissociative conditions Two 95 x 1.6 cm columns, packed with Sepharose C1-6B, were connected in series, equilibrated with a solution of 4 M guanidinium HC1, 50raM Tris buffer, pH 7.2, 100ram 6-aminohexanoic acid, 5 mM benzamidine HCI and calibrated with protein standards. The acid precipitate fraction (20-100 rag) was dissolved in 2 ml of the eluting buffer and applied to the column. Flow rates were 10-15 mi/hr, 4.8 ml fractions were collected and dialyzed and assayed for protein concentration (Lowry, 1951), absorbance at 280 um and "matrigenin" activity. Tubes containing high "matrigenin" activity were pooled and are referred to as the "j-fraction'.
Sepharose CI-4B-Con .4 a~nity chromatography A 1.6 x 10 cm column was equilibrated in 100 mM Na acetate buffer, pH 6, in 1 M NaC1 containing I mM of each of MgC12, MnC12 and CaC12. Elution was carried out with a 0-250 mM gradient of methyl D-mannoside in the same buffer/salt solution. Fractions (5 ml) were analyzed for absorbance at 230 nm and protein concentration, dialyzed and assayed for "matrigenin" activity.
Gelatin a~nity chromatography A 10 x 2 cm affinity column was prepared with calf skin collagen and cyanogen bromide-activated Sepharose C1-4B and equilibrated in 50 mM Tris buffer, pH 7.4, in 90 mM NaC1. The samples were applied in the same buffer and unbound material eluted with 50ml of this buffer. A solution of 8 M urea, 50 mM Tris, 90 mM NaCl, pH 7.4, was then used to elute bound material. Fractions were dialyzed, freeze-dried and then protein concentration and "matrigenin" activity were estimated.
Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) This was carried out using the method of Laemmli with 18% separating gels. No mercaptoethanol was added to the samples. The gels were stained by the silver stain method. For assessment of "matrigenin" activity the samples were run on SDS-PAGE as duplicates in adjacent lanes. One of the duplicate lanes was not stained, but was sliced in 1 cm segments, extracted overnight with 0.375M Tris-HCl, pH 7.2, 0.1% SDS and dialyzed extensively for 4 days against water to remove the SDS. The extracts were lyophilized and tested for "matrigenin" activity, as described above. Traces of SDS did not interfere with the "matrigenin" assay.
Bovine synovial cell and 3 T3 cell cultures Bovine synovial fibroblastic cells were obtained from bovine ankle joint synovium and grown to confluency, as described above for the human cells. The 3T3 cells were a mouse fibroblast line, originally obtained from the National Institutes of Health, Bethesda, Md. and donated to us by Dr R. Kerbel, National Cancer Institute Group, Dept. of Pathology, Queen's University. These cells were maintained in RPMI medium in 10% fetal calf serum. The comparative studies on the human synovial, bovine synovial and 3T3 mouse fibroblastic cells were done utilizing the fraction ("j-peak") from the Sepharose CL-6B that contained high specific activity material in the "matrigenin" assay.
RESULTS T h e " m a t r i g e n i n " assay, using h u m a n synovial fibroblasts, gave a linear response to the a d d i t i o n o f increasing c o n c e n t r a t i o n s o f the E D T A extract f r o m b o v i n e b o n e matrix, in the c o n c e n t r a t i o n range o f
Matrigcnin purification
500--
c~....
'.~~.g~.
z
/
~///
/
/
~
"
3Hperdish
/
/
'
531
~
.~
per cell
to._ .....
.~
. / ~ S per dish 0
'~ , ; ~ / f ~E ~
~'-- ~" 3SS per cell
/
o
I
0
I
I00 200 BONE EXTRACT CONCENTRATION(jug/rnl)
Fig. l. The effect of varying the protein concentration in the EDTA extract on "matrigenin" activity. The acid precipitate of the EDTA extract of bovine bone was diluted into medium at various concentrations and added to cultures of synovial fibroblastic cells. "Matrigenin" activity was determined as described in the Methods. "Matrigenin" activity was measured using both 3H and 35Sincorporation and is expressed both on a per cell and on a per dish basis, as shown. 0-60 # g protein/ml of medium (Fig. 1). The slope of the line indicated in the concentration-activity plot varied with the strain of the human fibroblast used. A standard sample of the acid precipitate fraction was used routinely to monitor the sensitivity of the assay. Figure 2 shows a comparison of responses of bovine fibroblasts, human fibroblasts and 3T3 cells, tested simultaneously in the "matrigenin" assay. As shown in Table 1, the acid precipitation step, at pH 3.5, resulted in a 2-fold enrichment of the biological activity in the precipitate and full recovery of the activity from the EDTA extract. This was found to be a convenient first step in the preparation of large amounts of material containing stable "matrigenin" activity. This preparation could be stored for several months at - 20°C without significant loss of activity. Initially we attempted to further purify the acid
o 6oo o o
._= 500
3H
~ .......
/ ,C -o . . . . . . .
~e,.
precipitate by gel filtration on Sephacryl S-200 in Dulbecco's PBS, pH 7.2. In Fig. 3, it can be seen that almost all of the biological activity eluted with the bulk of the protein at and just after the void volume, suggesting that under these conditions the "matrigenin" activity behaved as large mol. wt material. Since with this low ionic strength buffer the activity coeluted with the bulk of protein in the extract, we considered carrying out the gel filtration step under dissociative conditions, as we had evidence (see below) that the active material could associate with collagen and self aggregation was also a possibility. Because of the distortion of apparent mol. wt in 4 M guanidinium HCI we considered using a gel filtration matrix with a larger fractionation range and lengthening the column. Gel filtration was then done in 4 M guanidinium HC1 in Sepharose C1-6B and the
eov
~
3s S
FIB
o 3T3
400 ..~
0 _~ 3 0 0
~s
/
Z
~,.I 200
/
--.e
BOVFIB
/
~ " ......~o,
3T3
HUM
3E
7_ I00 0
I 5
jug/ml j Fraction
I I0
I/A" 0
I
I 5
I IO
jug/ml j Fraction
Fig. 2. The effect of varying concentrations of the "j-fraction" from bovine bone extract on the stimulation of incorporation of (3H) glucosamine and (3sS) SO4 into pericellular glycosaminoglycans of fibroblastic cells from three mammalian species. Bovine synovial (BOV FIB), human synovial (HUM FIB) and 3T3 fibroblasts were grown to confluency, labelled with radioisotopes and the incorporation of the radioactivity into pericellular glycosaminoglycans (GAG) was assayed using the "matrigenin" assay, as described under Methods. The stimulation of the incorporation is expressed as a percentage over control.
532
D. S. IRWINand T. P. ANASTASSIADES Table 1. Recovery of "matrig©nin" activity following acid precipitation of EDTA extract pH
7.2
Specific activity (units//zg per ml) Total activity (units)
5.0 Supernatant
4.7
5.0 Precipitate
6.0
15,900
3.5 Supernatant
6.0
14,700
5,900
3.5 Precipitate
0.5
8.1
600
16,900
Three aliquots of the EDTA extract from bovine bone were adjusted to the indicated pH. Where a precipitate was present it was centrifuged out. Supernatants and precipitates were analysed for protein and "matrigenin" activity.
Vo
I
l
\
25K
>F--
50
z-
-
i
~25
I
z fr o_
~~ 5
{0''0 OxO~. x . \o---o
~ oS/.
I IO
. . .
l I 15 20 FRACTION NUMBER
o\\\
." O.
~
~ .--~' O~
I 25
Fig. 3. Sephacryl S-200 gel filtration of the acid precipitate. The gel filtration was carried out under non-dissociative conditions. See Methods. length of the column was increased as described in the Methods. It was observed that the acid precipitate dissolved much better in 4 M guanidinium HC1 buffer than in PBS, where the solubility was only about 3 mg/ml. The results are shown in Fig. 4. The elution range of the active material was confined to one area on the downslope of the broad second peak. By comparison with protein standards chromatographed on the same columns and in the same dissociative buffer, the biologically active material eluted in the Mr range between 10,000 and 25,000.
o
g
2~
0.10 I
II
,
E
'
L
,oo ~ ?
Z o.o~
I
I
i
J
20
30
40
50
60
70
80
FRACTION NUMBER
Fig. 4. Sepharose C1-6B gel filtration under dissociative conditions. Chromatography was carried out in tandem columns, as described in the Methods. Twenty milligrams of the acid precipitate from the EDTA extract was applied in the 4 M guanidinium HCI buffer.
In order to estimate the molecular size of the activity, material containing the "matrigenin" activity peak from the Sepharose CI-6B (Fig. 4) was run on SDS--PAGE, as described under Methods. Preliminary experiments indicated that a large proportion of the "matrigenin" activity could be recovered from the SDS gels, following extraction of the gel and removal of the SDS by dialysis. Figure 5 shows the results of the electrophoresis. Nearly all of the activity migrated in the region of Mr = 27,500 and was distinct from the major protein bands. The material containing the activity constituted only a small portion of the total protein in the high specific activity fraction from Sepharose C1-6B chromatography. When similar S D S - P A G E gels were run in the presence of mercaptoethanol no activity could be recovered. The peak containing the activity from Sepharose CI-6B ("j-peak") was also run on acidurea gels. When sections of these gels were extracted and assayed for "matrigenin" activity no distinct area of activity could be identified. Affinity chromatography was assessed for further purification of the pooled fractions that contained the biological activity. Gelatin affinity chromatography showed that only about 0.73% of the applied material (by protein estimation) and approx. 4% of the "matrigenin" activity bound to the column. The specific activity (units//~g per ml medium) of the material that bound and could be eluted off with 8 M urea was about 5.7-fold greater than that of the unbound material. Con A affinity chromatography was also evaluated. Some of the activity was detected in the breakthrough volume, which also contained
Matrigenin purification
I L9
~"
"
~//////~///#/....... L,,,~.-~ , , I I
W
.-
, ,i',i
]
od
Fig. 5. SDS--PAGE of the "matrigenin" activity peak from Sepharose CI-6B chromatography. The left side of the figure shows a lane of the gel that has been stained with the silver reagent and the positions of the molecular weight standards. The bar graph indicates "matrigvnin" activity. The percentag¢ stimulation of incorporation of radioactivity into glycosaminoglycans in the assay is represented by the total height of the bar for (3H) glucosamine and by the cross hatched portion of the bar for (35S) SO4.
most of the protein. A large part of the "matrigenin" activity and a very small amount of protein was very tightly bound to the Con A and eluted off only with > 1 M methyl D-mannoside. The affinity chromatography procedures did not yield significant purification or resulted in low recoveries. DISCUSSION
533
staining. We evaluated several other purification methods. Gelatin aifmity chromatography gave good purification but low recovery. Experiments with DEAE-Sephacel have suggested that detergents as well as 8 M urea may be necessary to prevent aggregation on these columns since even with high pH buffers a significant portion of the activity did not stick to the column. Results from Con A affinity chromatography suggested that there was carbohydrate on the active material, since most of the activity bound to the column in 1 M NaCl/acetate buffer, but this interpretation is not certain because of the tendency for the activity to aggregate and stick. We were successful in recovering "matrigenin" activity from samples subjected to SDS-PAGE. With this approach we were able to recover "matrigenin" activity at a Mr of about 27,500 and obtained significant purification from other proteins contained in the activity peak from Sepharose CI-6B (Fig. 4), as evidenced by the silver staining of the gel (Fig. 5). It is not clear at this stage whether "matrigenin" activity is the same as the cartilage and bone inducing activity that bone matrices may express when implanted in intact animals. Purification work on the latter activity has been done by using a test system where purified material in the milligram range has been implanted in whole animals or in bony defects of whole animals and bone formation has been assessed morphologically (Urist et al., 1984). This should be contrasted with our tissue culture assay which measures a linear biochemical response in the low microgram range. Another "cartilage inducing factor" has been isolated from bovine bone using proteoglycan production by embryonic rat fibroblasts embedded in agar as the assay (Seyedin et al., 1985) and this has been recently shown to be identical with "transforming growth factor-beta" (Seyedin et al., 1986). However, this factor appears to have significant growth stimulatory activity in that system (rat muscle fibroblasts embedded in agar) and the activity could not be recovered from SDS gels, while "matrigenin" activity could be readily recovered from SDS gels. The precise role of "matrigenin" activity in repair of cartilage and bone has not been determined. We speculated that this activity may be part of a system of matrix-associated proteins found in bones of mammals that would be advantageous in initiating repair, following injury or repeated mechanical stress (Anastassiades, 1986). It is also possible that this type of activity may play a role in the repair of articular cartilage in the regions abutting to the subchondral bone.
The features of the assay used in the purification of "matrigenin" activity are linearity (Fig. 1) and relative speed of performance, compared to the full procedure for isolating surface glycosaminoglycans. The assay is sensitive to 1 #g of protein/ml of partially purified material. There appears to be no species specificity of the fibroblastic cells that were tested from the different mammalian sources in the "matrigenin" assay (Fig. 2). Any of the three types of cells could be used in a similar assay. The extraction procedure used was based on observations (Anastassiades et al., 1984) that the EDTA extract from bovine bone contained a proportion of the stimulatory activity but some remained bound to the collagenous residue and could be extracted with 4 M guanidinium HC1. During the extraction procedure and subsequent purification steps consideration was given to possible susceptibility of noncollagenous bone proteins to protease activity and we have included several protease inhibitors. Acknowledgements--The authors wish to thank Mr James The bulk of the activity eluted in a relatively high Kosir for his excellent technical assistance. mol. wt range under non-dissociative gel filtration conditions (Fig. 3), while under dissociative conditions the activity eluted at a much lower mol. wt REFERENCES range, about Mr 20,000 (Fig. 4). This suggests that the molecules containing the activity tend to associate Anastassiades T. (1986) Control of repair in articular tissues and the development of osteoarthritis. Med. Hypotheses under non-dissociative conditions. The material from 19, 89-92. Sepharose CI-6B still required further purification, Anastassiades T. and Irwin D. (1985) Stimulation of glycossince SDS-electrophoresis of the fraction containing aminoglycan synthesis in culture by factors from mamthe activity showed at least four bands by Coomassie malian bones (matrigenin activity). In Degenerative Joints (Edited by Verhruggen G. and Veys E. M.) Vol 2, blue staining and several additional bands by silver
534
D.S. Igw~ and T. P. A~ASTASSIADV.S
pp. 167-180. Elsevier Science Publishers Biomedical Division, New York. Anastassiades T., Irwin D., Wood A. and Robertson W. (1984) The effect of solubllized bone matrix fractions from different msmmalian species on glycosaminoglycan synthesis by cultured fibroblasts. Comp. Biochem. Physiol. 7911, 623-63 I. Anastassiades T., Ley J. and Wood A. (1978a) The growth kinetics of synovial fibroblastic ceils from inflammatory and non-inflammatory arthropathies. Arthritis Rheum. 21, 461--466. Anastassiades T., Puzic O. and Puzic R. (1978b) Effect of solubilized bone matrix components on cultured flbroblasts derived from neonatal rat tissues. Calcif. Tiss. Res. 26, 173-179. Anastassiades T. and Starkey P. (1983) Incorporation of radioactive precursors into glycosaminoglycans by rat muscle fibroblasts exposed to a solubilized bone matrix fraction. Cell. Differ. 13, 341-348, Anastassiades T., Starkey P. and Puzic O. (1982) Glycosaminoglycan and surface glycoprotein synthesis by rat muscle fibroblast monolayers: response to solubilized bone matrix and effect of serum. Cell. Differ. 11, 81-90. Anastassiades T. and Wood A. (1981) Effect of soluble products from lectin-stimulated lymphocytes on the growth, adhesiveness and glycosaminoglycansynthesis of cultured synovial fibroblastic cells. J. clin. Invest. 68, 792-802. Barth A. (1893) Ueber histoiogische befunde nach knochenimplantationen. Archs kiln. Chit. 46, 409-417. Bonome A. (1886) Intorno alia rigenerazione del tessuto osseo. Archs. Perle Sci. Med. 9, 131-190. Herring G. M., Ashton B. A. and Chipperfield A. R. (1974)
The isolation of soluble proteins, glycoproteins, and proteoglycans from bone. Prepar. Biochem. 4, 179-200. Hronowski L. and Anastassiades T. (1980) Rates of glycosaminoglycan synthesis and rates of incorporation of radioactive precursors into newly synthesized glycosaminoglycans by confluent rat muscle flbroblasts. J. biol. Chem. 255, 9210-9217. Huggins C., Wiseman S. and Reddi A. H. (1970) Transformation of fibroblasts by allogeneic and xenogeneic transplants of demineralized tooth and bone. J. exp. Med. 132, 1250-1258. Lowry O. H., Rosebrough N., Farr A. L. and Randall R. J. (1951) Protein measurement with the Folin phenol reagent. J. biol. Chem. 193, 265-275. Reddi A. (1981) Cell biology and biochemistry of endochondral bone development. Coll. Res. 1, 209-226. Seyedin S. M., Thompson A. Y., Bentz H., Rosen D. M., McPherson J. M., Conti A., Siegel N. R., Galluppi G. R. and Piez K. A. (1986) Cartilage-inducing factor A. Apparent identity to transforming growth factor-beta. J. biol. Chem. 261, 5693-5695. Seyedin S. M., Thomas T. C., Thompson A. Y., Rosen D. M. and Piez K. A. (1985) Purification and characterization of two cartilage-inducing factors from bovine demineralized bone. Proc. hath. Acad. Sci. USA 82, 2267-2271. Urist M. (1965) Bone formation by autoinduction. Science 150, 893-899. Urist M. R., Huo Y. K., Brownell A. G., Hohl W. M., Buyske J., Lietze A., Tempst P., Hunkapiller M. and DeLange R. J. (1984) Purification of bovine morphogenetic protein by hydroxyapatite chromatography. Proc. natn. Acad. Sci. 81, 371-375.
0305-0491/87 $3.00+ 0.00 © 1987 Pergamon Journals Ltd
"MATRIGENIN" ACTIVITY FROM BOVINE BONE--I. PARTIAL PURIFICATION OF ACTIVITY* D. S. IRWIN and T. P. ANASTASSIAOES Department of Medicine and Biochemistry and the Rheumatic Diseases Unit, Queen's University, Etherington Hall, Stuart Street, Kingston, Ontario, Canada K7L 3N6 (Tel.: 613-545-2971) (Received 16 December 1986)
Abstract--1. Bovine bone contains an extractable activity which stimulated the synthesis of glycosaminoglycans by bovine synovial, human synovial and mouse 3T3 fibroblastic cells in culture. Human cells were used to develop an assay for purification of the stimulatory activity ("matrigenin" activity) from bovine bone. 2. Partial purification of "matrigenin" activity was achieved by precipitation of the EDTA extract at pH 3.5 and Sepharose CL-6B chromatography in 4 M guanidinium HC1. Dissociative conditions were necessary to prevent aggregation. 3. On SDS-polyacrylamide gel electrophoresis the activity ran with a mobility equivalent to a Mr = 27,500 and could be recovered from the SDS gels.
In this paper we report on a relatively rapid assay for the stimulation of glycosaminoglycan synthesis by cultured fibroblastic cells. The conditions of the assay were designed to minimize effects of the bone extracts on fibroblastic proliferation. We refer to this stimulatory activity on extracellular glycosaminoglycan synthesis as "matrigenin" activity (Anastassiades and Irwin, 1985), which is an operational term based on the tissue culture assay used for purification of the activity from bovine bone.
INTRODUCTION
It has been known for a long time that implantation of bone fragments into the muscle or skin of mammals leads to the formation of new cartilage and bone in the host (Barth, 1893; Bonome, 1886). This type of experiment can be done with xenogeneic mammalian bone and has been studied extensively in rodents (Urist, 1965; Huggins et al., 1970), where the early biochemical events are characterized by an increase in hyaluronic acid synthesis, followed by increased cartilage-type proteoglycan synthesis (Reddi, 1981). We had reported (Anastassiades et al., 1984) that an activity that stimulated glycosaminoglycan synthesis by cultured mammalian fibroblastic cells could be extracted with a solution of 0.34 M EDTA in 2.5 M NaC1 from bovine but not human bone matrix. Subsequent extraction of the residues with 4 M guanidinium HC1 yielded activity from both matrices. In addition, synovial fibroblastic cells from man and rabbit were evaluated for their ability to respond to the stimulatory effects of the extracts by increased glycosaminoglycan synthesis. The human synovial cells constituted a sensitive assay for the glycosaminoglycan stimulatory activity (Anastassiades et al., 1984) and we have used these cells routinely as they are easily available to us. We had previously reported (Anastassiades et al., 1978b, 1982; Anastassiades and Starkey, 1983) that a similar stimulatory activity could be extracted from rat bone using 4 M guanidinium HC1 solutions. Fibroblastic cells from neonatal rat muscle were used as the assay system in that work. The bovine bone provided advantages as a source for purification of this activity that included ready availability of large quantities of bone and the relative ease of extraction of the activity compared to rat or human bone.
MATERIALS AND METHODS
Preparation of bone powder Bovine femurs and tibias obtained from a local slaughterhouse were stored at -20°C. They were cleaned and the cortical bone was powdered, using a Wiley Mill (A. H. Thomas, Philadelphia), to a size less than 1 nun (20 mesh screen).
*Supported by Grants from the Arthritis Society, the Medical Research Council of Canada and an Arthritis Society Fellowship to Dr D. S. Irwin.
EDTA extraction of bone powder Extraction was done by a modification of the method of Herring (Herring et al., 1974) but with the addition of protease inhibitors. Bone powder (0.7 g/ml) was placed in Spectrapor 1 dialysis bags (Spectrum Medical Industries, Los Angeles) in 2.5 M NaC1, 100raM 6-aminohexanoic acid, 5mM benzamidine, l mM phenylmethylsulfonyl fluoride and dialyz_~dagainst a solution of 0.34 M EDTA adjusted to pH 7.4 with NaOH and containing the same concentrations of inhibitors as in the bag (38 ml of solution per g of bone powder). The bags were intermittently agitated. After 4 days the liquid in the bags was removed, saved and replaced with fresh solution of 2.5 M NaC1 containing the inhibitors and the bags again placed in a fresh EDTA solution plus the inhibitors for a further extraction period of 4 days. At the end of the second extraction period the combined supernatants were filtered. The clear, yellow filtrate was then dialyzed against three changes of 25 mM 6-aminohexanoic acid, 0.2 mM benzamidine in water and two changes of plain water. All procedures were done at 4°C.
529
530
D . S . IRWIN and T. P. ANASTASSIADES
Acid precipitation
Routine assay for "matrigenin " activity
The EDTA extract was made 100 mM in glycine HC1 buffer, pH 3.5 at 0°C. In some cases the pH was adjusted to 3.0 with HC1 to complete the precipitation. The suspension was centrifuged at 0°C, the supernatant discarded and the precipitate was resuspended in a small volume of Dulbecco's phosphate buffered saline (PBS). The suspension was dialyzed against water and freeze-dried. This is termed the "acid precipitate".
Human synoviai fibroblasts were obtained from synovial explants, subcultured and grown to confluency in 35 mm dishes, containing CMRL 1969 medium supplemented with 20% fetal calf serum (FCS), by methods we have previously described (Anastassiades et al., 1978a; Anastassiades and Wood, 1981). Freeze dried samples of the fractions to be tested (and "control" fractions) were dissolved in fresh medium with 10% fetal calf serum, containing also 1.25 #Ci/ml of D-{1,6-3H(N)} glucosamine (43.2 Ci/mmol) and 2/~Ci/ml of (35S) Na2SO4 (New England Nuclear Co). The old medium was removed and the fresh, labeled media were used to feed the cultures (1.5 ml medium per dish). After 38-40 hr of incubation, the cell surface-associated material was obtained by trypsinization (Hronowski and Anastassiades, 1980) and cells were removed by centrifugation. To the supernatant 100 #l of solution containing the "carrier" giycosaminoglycan consisting of 1 mg/ml of hyaluronic acid (human umbilical cord, Grade III-S, Sigma Chemical Co.) and 1 mg/ml of ehondroitin sulfate (whale or shark cartilage, grade III, Sigma Chemical Co.) were added. Sixteen millilitres of absolute ethanol were added with thorough mixing and the tube left several hours at 4°C. The precipitate was centrifuged and the supernatant discarded. The precipitate was then dissolved in 2.0ml, 0.075 M Na2EDTA-0.005M cysteine HC1 buffer, pH6.2. Undissolved material was removed by centrifugation and the supernatant was poured into 20 ml counting vials. Two hundred microlitres of 5% cetylpyridinium chloride (CPC) in 0.2M Na2SO 4 was added then 5.0ml H20 and the contents were left overnight at room temperature. The precipitated glycosaminoglycans and proteogiycans were centrifuged at 10,000g for 10 min at 20°C. The supernatant was poured off, l0 ml of Atomiight (New England Nuclear Co.) added and the radioactivity measured by liquid scintillation counting. The results for 3H or 35S incorporation for the test fractions are expressed in terms of units per #g protein added per ml of medium. A unit = [(dpm incorporated by cultures to which test fractions were added/dpm incorporated by control cultures) - 1 x 100]. This simplified assay procedure was validated against the longer procedure, which included extensive papain digestion (Hronowski and Anastassiades, 1980) and essentially identical values were observed for the incorporation of radioactivity into the glycosaminoglycan CPC precipitate from the cell surface and for the glycosaminoglycan isolated by the longer procedure. The period of time chosen for labelling the cultures was based on maximal labelling of the cell surface glycosaminoglycan (Anastassiades et al., 1984) and on convenience.
Sephacryl S-200 gel fihration For non-dissociative chromatography a 90 x 1.6 cm column of Sephacryl S-200 was equilibrated with Dulbecco's PBS, pH 7.2, with 0.02% sodium azide. Three milligrams of the acid precipitate was dissolved in 1 ml of this buffer and this was run on the column and eluted at l0 ml/hr. Five ml fractions were collected, dialyzed and aliquots analyzed for protein and assayed for "matrigenin" activity as described below. For dissociative chromatography, the same column was equilibrated with 4 M guanidinium HC1, 50 mM Tris buffer, pH 7.4. Material that had been extracted with 4 M guanidinium HC1, 0.34 M NaEDTA, 50 mM Tris, pH 7.2, was concentrated to 6 mi with Amicon PM-10 and applied to the column.
Sepharose Cl-6B gel filtration under dissociative conditions Two 95 x 1.6 cm columns, packed with Sepharose C1-6B, were connected in series, equilibrated with a solution of 4 M guanidinium HC1, 50raM Tris buffer, pH 7.2, 100ram 6-aminohexanoic acid, 5 mM benzamidine HCI and calibrated with protein standards. The acid precipitate fraction (20-100 rag) was dissolved in 2 ml of the eluting buffer and applied to the column. Flow rates were 10-15 mi/hr, 4.8 ml fractions were collected and dialyzed and assayed for protein concentration (Lowry, 1951), absorbance at 280 um and "matrigenin" activity. Tubes containing high "matrigenin" activity were pooled and are referred to as the "j-fraction'.
Sepharose CI-4B-Con .4 a~nity chromatography A 1.6 x 10 cm column was equilibrated in 100 mM Na acetate buffer, pH 6, in 1 M NaC1 containing I mM of each of MgC12, MnC12 and CaC12. Elution was carried out with a 0-250 mM gradient of methyl D-mannoside in the same buffer/salt solution. Fractions (5 ml) were analyzed for absorbance at 230 nm and protein concentration, dialyzed and assayed for "matrigenin" activity.
Gelatin a~nity chromatography A 10 x 2 cm affinity column was prepared with calf skin collagen and cyanogen bromide-activated Sepharose C1-4B and equilibrated in 50 mM Tris buffer, pH 7.4, in 90 mM NaC1. The samples were applied in the same buffer and unbound material eluted with 50ml of this buffer. A solution of 8 M urea, 50 mM Tris, 90 mM NaCl, pH 7.4, was then used to elute bound material. Fractions were dialyzed, freeze-dried and then protein concentration and "matrigenin" activity were estimated.
Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) This was carried out using the method of Laemmli with 18% separating gels. No mercaptoethanol was added to the samples. The gels were stained by the silver stain method. For assessment of "matrigenin" activity the samples were run on SDS-PAGE as duplicates in adjacent lanes. One of the duplicate lanes was not stained, but was sliced in 1 cm segments, extracted overnight with 0.375M Tris-HCl, pH 7.2, 0.1% SDS and dialyzed extensively for 4 days against water to remove the SDS. The extracts were lyophilized and tested for "matrigenin" activity, as described above. Traces of SDS did not interfere with the "matrigenin" assay.
Bovine synovial cell and 3 T3 cell cultures Bovine synovial fibroblastic cells were obtained from bovine ankle joint synovium and grown to confluency, as described above for the human cells. The 3T3 cells were a mouse fibroblast line, originally obtained from the National Institutes of Health, Bethesda, Md. and donated to us by Dr R. Kerbel, National Cancer Institute Group, Dept. of Pathology, Queen's University. These cells were maintained in RPMI medium in 10% fetal calf serum. The comparative studies on the human synovial, bovine synovial and 3T3 mouse fibroblastic cells were done utilizing the fraction ("j-peak") from the Sepharose CL-6B that contained high specific activity material in the "matrigenin" assay.
RESULTS T h e " m a t r i g e n i n " assay, using h u m a n synovial fibroblasts, gave a linear response to the a d d i t i o n o f increasing c o n c e n t r a t i o n s o f the E D T A extract f r o m b o v i n e b o n e matrix, in the c o n c e n t r a t i o n range o f
Matrigcnin purification
500--
c~....
'.~~.g~.
z
/
~///
/
/
~
"
3Hperdish
/
/
'
531
~
.~
per cell
to._ .....
.~
. / ~ S per dish 0
'~ , ; ~ / f ~E ~
~'-- ~" 3SS per cell
/
o
I
0
I
I00 200 BONE EXTRACT CONCENTRATION(jug/rnl)
Fig. l. The effect of varying the protein concentration in the EDTA extract on "matrigenin" activity. The acid precipitate of the EDTA extract of bovine bone was diluted into medium at various concentrations and added to cultures of synovial fibroblastic cells. "Matrigenin" activity was determined as described in the Methods. "Matrigenin" activity was measured using both 3H and 35Sincorporation and is expressed both on a per cell and on a per dish basis, as shown. 0-60 # g protein/ml of medium (Fig. 1). The slope of the line indicated in the concentration-activity plot varied with the strain of the human fibroblast used. A standard sample of the acid precipitate fraction was used routinely to monitor the sensitivity of the assay. Figure 2 shows a comparison of responses of bovine fibroblasts, human fibroblasts and 3T3 cells, tested simultaneously in the "matrigenin" assay. As shown in Table 1, the acid precipitation step, at pH 3.5, resulted in a 2-fold enrichment of the biological activity in the precipitate and full recovery of the activity from the EDTA extract. This was found to be a convenient first step in the preparation of large amounts of material containing stable "matrigenin" activity. This preparation could be stored for several months at - 20°C without significant loss of activity. Initially we attempted to further purify the acid
o 6oo o o
._= 500
3H
~ .......
/ ,C -o . . . . . . .
~e,.
precipitate by gel filtration on Sephacryl S-200 in Dulbecco's PBS, pH 7.2. In Fig. 3, it can be seen that almost all of the biological activity eluted with the bulk of the protein at and just after the void volume, suggesting that under these conditions the "matrigenin" activity behaved as large mol. wt material. Since with this low ionic strength buffer the activity coeluted with the bulk of protein in the extract, we considered carrying out the gel filtration step under dissociative conditions, as we had evidence (see below) that the active material could associate with collagen and self aggregation was also a possibility. Because of the distortion of apparent mol. wt in 4 M guanidinium HCI we considered using a gel filtration matrix with a larger fractionation range and lengthening the column. Gel filtration was then done in 4 M guanidinium HC1 in Sepharose C1-6B and the
eov
~
3s S
FIB
o 3T3
400 ..~
0 _~ 3 0 0
~s
/
Z
~,.I 200
/
--.e
BOVFIB
/
~ " ......~o,
3T3
HUM
3E
7_ I00 0
I 5
jug/ml j Fraction
I I0
I/A" 0
I
I 5
I IO
jug/ml j Fraction
Fig. 2. The effect of varying concentrations of the "j-fraction" from bovine bone extract on the stimulation of incorporation of (3H) glucosamine and (3sS) SO4 into pericellular glycosaminoglycans of fibroblastic cells from three mammalian species. Bovine synovial (BOV FIB), human synovial (HUM FIB) and 3T3 fibroblasts were grown to confluency, labelled with radioisotopes and the incorporation of the radioactivity into pericellular glycosaminoglycans (GAG) was assayed using the "matrigenin" assay, as described under Methods. The stimulation of the incorporation is expressed as a percentage over control.
532
D. S. IRWINand T. P. ANASTASSIADES Table 1. Recovery of "matrig©nin" activity following acid precipitation of EDTA extract pH
7.2
Specific activity (units//zg per ml) Total activity (units)
5.0 Supernatant
4.7
5.0 Precipitate
6.0
15,900
3.5 Supernatant
6.0
14,700
5,900
3.5 Precipitate
0.5
8.1
600
16,900
Three aliquots of the EDTA extract from bovine bone were adjusted to the indicated pH. Where a precipitate was present it was centrifuged out. Supernatants and precipitates were analysed for protein and "matrigenin" activity.
Vo
I
l
\
25K
>F--
50
z-
-
i
~25
I
z fr o_
~~ 5
{0''0 OxO~. x . \o---o
~ oS/.
I IO
. . .
l I 15 20 FRACTION NUMBER
o\\\
." O.
~
~ .--~' O~
I 25
Fig. 3. Sephacryl S-200 gel filtration of the acid precipitate. The gel filtration was carried out under non-dissociative conditions. See Methods. length of the column was increased as described in the Methods. It was observed that the acid precipitate dissolved much better in 4 M guanidinium HC1 buffer than in PBS, where the solubility was only about 3 mg/ml. The results are shown in Fig. 4. The elution range of the active material was confined to one area on the downslope of the broad second peak. By comparison with protein standards chromatographed on the same columns and in the same dissociative buffer, the biologically active material eluted in the Mr range between 10,000 and 25,000.
o
g
2~
0.10 I
II
,
E
'
L
,oo ~ ?
Z o.o~
I
I
i
J
20
30
40
50
60
70
80
FRACTION NUMBER
Fig. 4. Sepharose C1-6B gel filtration under dissociative conditions. Chromatography was carried out in tandem columns, as described in the Methods. Twenty milligrams of the acid precipitate from the EDTA extract was applied in the 4 M guanidinium HCI buffer.
In order to estimate the molecular size of the activity, material containing the "matrigenin" activity peak from the Sepharose CI-6B (Fig. 4) was run on SDS--PAGE, as described under Methods. Preliminary experiments indicated that a large proportion of the "matrigenin" activity could be recovered from the SDS gels, following extraction of the gel and removal of the SDS by dialysis. Figure 5 shows the results of the electrophoresis. Nearly all of the activity migrated in the region of Mr = 27,500 and was distinct from the major protein bands. The material containing the activity constituted only a small portion of the total protein in the high specific activity fraction from Sepharose C1-6B chromatography. When similar S D S - P A G E gels were run in the presence of mercaptoethanol no activity could be recovered. The peak containing the activity from Sepharose CI-6B ("j-peak") was also run on acidurea gels. When sections of these gels were extracted and assayed for "matrigenin" activity no distinct area of activity could be identified. Affinity chromatography was assessed for further purification of the pooled fractions that contained the biological activity. Gelatin affinity chromatography showed that only about 0.73% of the applied material (by protein estimation) and approx. 4% of the "matrigenin" activity bound to the column. The specific activity (units//~g per ml medium) of the material that bound and could be eluted off with 8 M urea was about 5.7-fold greater than that of the unbound material. Con A affinity chromatography was also evaluated. Some of the activity was detected in the breakthrough volume, which also contained
Matrigenin purification
I L9
~"
"
~//////~///#/....... L,,,~.-~ , , I I
W
.-
, ,i',i
]
od
Fig. 5. SDS--PAGE of the "matrigenin" activity peak from Sepharose CI-6B chromatography. The left side of the figure shows a lane of the gel that has been stained with the silver reagent and the positions of the molecular weight standards. The bar graph indicates "matrigvnin" activity. The percentag¢ stimulation of incorporation of radioactivity into glycosaminoglycans in the assay is represented by the total height of the bar for (3H) glucosamine and by the cross hatched portion of the bar for (35S) SO4.
most of the protein. A large part of the "matrigenin" activity and a very small amount of protein was very tightly bound to the Con A and eluted off only with > 1 M methyl D-mannoside. The affinity chromatography procedures did not yield significant purification or resulted in low recoveries. DISCUSSION
533
staining. We evaluated several other purification methods. Gelatin aifmity chromatography gave good purification but low recovery. Experiments with DEAE-Sephacel have suggested that detergents as well as 8 M urea may be necessary to prevent aggregation on these columns since even with high pH buffers a significant portion of the activity did not stick to the column. Results from Con A affinity chromatography suggested that there was carbohydrate on the active material, since most of the activity bound to the column in 1 M NaCl/acetate buffer, but this interpretation is not certain because of the tendency for the activity to aggregate and stick. We were successful in recovering "matrigenin" activity from samples subjected to SDS-PAGE. With this approach we were able to recover "matrigenin" activity at a Mr of about 27,500 and obtained significant purification from other proteins contained in the activity peak from Sepharose CI-6B (Fig. 4), as evidenced by the silver staining of the gel (Fig. 5). It is not clear at this stage whether "matrigenin" activity is the same as the cartilage and bone inducing activity that bone matrices may express when implanted in intact animals. Purification work on the latter activity has been done by using a test system where purified material in the milligram range has been implanted in whole animals or in bony defects of whole animals and bone formation has been assessed morphologically (Urist et al., 1984). This should be contrasted with our tissue culture assay which measures a linear biochemical response in the low microgram range. Another "cartilage inducing factor" has been isolated from bovine bone using proteoglycan production by embryonic rat fibroblasts embedded in agar as the assay (Seyedin et al., 1985) and this has been recently shown to be identical with "transforming growth factor-beta" (Seyedin et al., 1986). However, this factor appears to have significant growth stimulatory activity in that system (rat muscle fibroblasts embedded in agar) and the activity could not be recovered from SDS gels, while "matrigenin" activity could be readily recovered from SDS gels. The precise role of "matrigenin" activity in repair of cartilage and bone has not been determined. We speculated that this activity may be part of a system of matrix-associated proteins found in bones of mammals that would be advantageous in initiating repair, following injury or repeated mechanical stress (Anastassiades, 1986). It is also possible that this type of activity may play a role in the repair of articular cartilage in the regions abutting to the subchondral bone.
The features of the assay used in the purification of "matrigenin" activity are linearity (Fig. 1) and relative speed of performance, compared to the full procedure for isolating surface glycosaminoglycans. The assay is sensitive to 1 #g of protein/ml of partially purified material. There appears to be no species specificity of the fibroblastic cells that were tested from the different mammalian sources in the "matrigenin" assay (Fig. 2). Any of the three types of cells could be used in a similar assay. The extraction procedure used was based on observations (Anastassiades et al., 1984) that the EDTA extract from bovine bone contained a proportion of the stimulatory activity but some remained bound to the collagenous residue and could be extracted with 4 M guanidinium HC1. During the extraction procedure and subsequent purification steps consideration was given to possible susceptibility of noncollagenous bone proteins to protease activity and we have included several protease inhibitors. Acknowledgements--The authors wish to thank Mr James The bulk of the activity eluted in a relatively high Kosir for his excellent technical assistance. mol. wt range under non-dissociative gel filtration conditions (Fig. 3), while under dissociative conditions the activity eluted at a much lower mol. wt REFERENCES range, about Mr 20,000 (Fig. 4). This suggests that the molecules containing the activity tend to associate Anastassiades T. (1986) Control of repair in articular tissues and the development of osteoarthritis. Med. Hypotheses under non-dissociative conditions. The material from 19, 89-92. Sepharose CI-6B still required further purification, Anastassiades T. and Irwin D. (1985) Stimulation of glycossince SDS-electrophoresis of the fraction containing aminoglycan synthesis in culture by factors from mamthe activity showed at least four bands by Coomassie malian bones (matrigenin activity). In Degenerative Joints (Edited by Verhruggen G. and Veys E. M.) Vol 2, blue staining and several additional bands by silver
534
D.S. Igw~ and T. P. A~ASTASSIADV.S
pp. 167-180. Elsevier Science Publishers Biomedical Division, New York. Anastassiades T., Irwin D., Wood A. and Robertson W. (1984) The effect of solubllized bone matrix fractions from different msmmalian species on glycosaminoglycan synthesis by cultured fibroblasts. Comp. Biochem. Physiol. 7911, 623-63 I. Anastassiades T., Ley J. and Wood A. (1978a) The growth kinetics of synovial fibroblastic ceils from inflammatory and non-inflammatory arthropathies. Arthritis Rheum. 21, 461--466. Anastassiades T., Puzic O. and Puzic R. (1978b) Effect of solubilized bone matrix components on cultured flbroblasts derived from neonatal rat tissues. Calcif. Tiss. Res. 26, 173-179. Anastassiades T. and Starkey P. (1983) Incorporation of radioactive precursors into glycosaminoglycans by rat muscle fibroblasts exposed to a solubilized bone matrix fraction. Cell. Differ. 13, 341-348, Anastassiades T., Starkey P. and Puzic O. (1982) Glycosaminoglycan and surface glycoprotein synthesis by rat muscle fibroblast monolayers: response to solubilized bone matrix and effect of serum. Cell. Differ. 11, 81-90. Anastassiades T. and Wood A. (1981) Effect of soluble products from lectin-stimulated lymphocytes on the growth, adhesiveness and glycosaminoglycansynthesis of cultured synovial fibroblastic cells. J. clin. Invest. 68, 792-802. Barth A. (1893) Ueber histoiogische befunde nach knochenimplantationen. Archs kiln. Chit. 46, 409-417. Bonome A. (1886) Intorno alia rigenerazione del tessuto osseo. Archs. Perle Sci. Med. 9, 131-190. Herring G. M., Ashton B. A. and Chipperfield A. R. (1974)
The isolation of soluble proteins, glycoproteins, and proteoglycans from bone. Prepar. Biochem. 4, 179-200. Hronowski L. and Anastassiades T. (1980) Rates of glycosaminoglycan synthesis and rates of incorporation of radioactive precursors into newly synthesized glycosaminoglycans by confluent rat muscle flbroblasts. J. biol. Chem. 255, 9210-9217. Huggins C., Wiseman S. and Reddi A. H. (1970) Transformation of fibroblasts by allogeneic and xenogeneic transplants of demineralized tooth and bone. J. exp. Med. 132, 1250-1258. Lowry O. H., Rosebrough N., Farr A. L. and Randall R. J. (1951) Protein measurement with the Folin phenol reagent. J. biol. Chem. 193, 265-275. Reddi A. (1981) Cell biology and biochemistry of endochondral bone development. Coll. Res. 1, 209-226. Seyedin S. M., Thompson A. Y., Bentz H., Rosen D. M., McPherson J. M., Conti A., Siegel N. R., Galluppi G. R. and Piez K. A. (1986) Cartilage-inducing factor A. Apparent identity to transforming growth factor-beta. J. biol. Chem. 261, 5693-5695. Seyedin S. M., Thomas T. C., Thompson A. Y., Rosen D. M. and Piez K. A. (1985) Purification and characterization of two cartilage-inducing factors from bovine demineralized bone. Proc. hath. Acad. Sci. USA 82, 2267-2271. Urist M. (1965) Bone formation by autoinduction. Science 150, 893-899. Urist M. R., Huo Y. K., Brownell A. G., Hohl W. M., Buyske J., Lietze A., Tempst P., Hunkapiller M. and DeLange R. J. (1984) Purification of bovine morphogenetic protein by hydroxyapatite chromatography. Proc. natn. Acad. Sci. 81, 371-375.