acin (1 ,uM) did not affect the inhibition of MNC formation by. OIF, although OIF did induce prostaglandin synthesis in these cultures. We have shown (28) that ...
Proc. Nail. Acad. Sci. USA Vol. 87, pp. 3023-3026, April 1990 Medical Sciences
Osteoinductive factor inhibits formation of human osteoclast-like cells A. KUKITA*, L. BONEWALD*, D. ROSENt, S. SEYEDINt, G. R. MUNDY*, AND G. D. ROODMAN*t *Research Service and Geriatric Research, Education and Clinical Center, Audie Murphy Veterans Administration Hospital, and University of Texas Health Science Center, San Antonio, TX 78284; and tCollagen Corp., Palo Alto, CA 94303
Communicated by George J. Todaro, January 29, 1990
ABSTRACT Osteoinductive factor (OWF) is a glycoprotein in bone that induces ectopic bone formation. Implantation of OIF plus transforming growth factor (3 (TGF-(8) type 1 or 2 into subcutaneous tissues of rats induces formation of bone at the implantation site. Since TGF-(3 is also present in bone matrix and inhibits formation of multinucleated cells that express an osteoclast phenotype in long-term human marrow cultures, we tested the effects of OIF on formation of these osteoclast-like cells to determine the effects of OIF on cells in the osteoclast lineage. We found that OIF inhibited total multinucleated cell (MNC) formation in a dose-dependent fashion and preferentially inhibited formation of MNCs that react with monoclonal antibody 23c6 (23c6-positive MNCs), an antibody that identifies osteoclasts. In addition, low concentrations of OWF in combination with low concentrations of TGF-(3 acted synergistically to inhibit 23c6-positive MNC formation. The inhibition of 23c6-positive MNC formation by OWF was not mediated by prostaglandin synthesis. These data suggest that regulatory growth factors, such as OWF or TGF-(3, that are stored within the bone matrix and released when bone is resorbed can serve as natural inhibitors of osteoclast activity by inhibiting osteoclast formation. An osteoinductive factor (GIF) has been isolated and characterized from bovine bone (1). OIF is a 22- to 28-kDa heavily glycosylated protein that possesses a distinct amino acid sequence unrelated to any other known protein (2). Implantation of OIF on a soluble bovine dermal collagen and calcium phosphate ceramic carrier plus transforming growth factor ( (TGF-f3) type 1 or 2 into subcutaneous tissue of rats in vivo induces formation of ectopic bone at the implantation site (1). OIF and TGF-p are both present in the bone matrix. TGF-f3 is more abundant than OIF in bone although TGF-j3 is present in a latent form that must be activated. TGF-f8 is also present in a variety of other cell types (3-7) and is multifunctional (8-13). In contrast, OIF appears to affect only bone cells. OIF can stimulate proliferation and alkaline phosphate activity of osteoblastic cells (14) and bone marrow stromal cells (14), but to date no other activity has been reported for OIF. Since OIF and TGF-/3 have potentiating effects on ectopic bone formation and TGF-,f inhibits formation of multinucleated cells (MNCs) with osteoclast characteristics, we tested the effects of OIF on cells of the osteoclast lineage. We used a long-term human marrow culture system in which MNCs with osteoclast characteristics form (15, 16, 29-31). Approximately 50%o of the total MNCs formed in these cultures fulfill the functional characteristics of osteoclasts. These MNCs (i) contain a tartrate-resistant acid phosphatase, which is a marker enzyme for osteoclasts, (ii) react strongly with monoclonal antibody 23c6, which preferentially identifies osteoclasts (17), (iii) contract in response to calcitonin (18), and (iv) form resorption lacunae when cultured on sperm whale The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. 3023
dentine (18). In addition, MNC formation is appropriately regulated by osteotropic hormones (16). In this report we demonstrate that OIF preferentially inhibits, in a dose-dependent fashion, the formation of those MNCs that cross-react with monoclonal antibody 23c6 (23c6positive MNCs). In addition, OIF in combination with TGF,B1 acts synergistically to inhibit 23c6-positive MNC formation. OIF also inhibited formation of 23c6-negative MNC, but to a lesser extent. These data suggest that factors within the bone matrix may serve as natural inhibitors of osteoclast activity by inhibiting osteoclast formation.
MATERIALS AND METHODS Materials. 1,25-Dihydroxyvitamin D3 [1,25(0H)2D3] was generously provided by M. R. Uskokovic (Hoffmann-La Roche, Nutley, NJ). Bovine parathyroid hormone (PTH) peptide (1-34) was obtained from Sigma. Salmon calcitonin was obtained from Rorer Pharmaceuticals (Fort Washington, PA). Monoclonal antibody 23c6, which preferentially binds to osteoclasts, was kindly provided by M. Horton (Saint Bartholomew's Hospital, London). Glutaraldehyde, Histopaque-1077, indomethacin, and methyl green were obtained from Sigma. Vectastain ABC-AP kits were purchased from Vector Laboratories. The a modification of minimal essential medium (aMEM) was purchased from GIBCO. Horse serum was obtained from Flow Laboratories. Twenty-four-well multiwell plates were obtained from Coming. Labtek slides were obtained from Nunc. OIF and TGF-f31 were purified from bone as described (2, 6). OIF was free of any TGF-/3 activity. Culture of Human Marrow Cells. Bone marrow cells were aspirated from the posterior iliac crest of healthy normal donors. All donors gave informed written consent before bone marrow aspiration. Marrow mononuclear cells were obtained by density gradient centrifugation on Hypaque/ Ficoll as described (19). The cells were washed twice with aMEM and cultured in aMEM containing 20%6 (vol/vol) horse serum at 1 x 106 cells per ml in 24-well plates (0.5 ml per well). All cultures were maintained in a humidified atmosphere of 4% C02/96% air at 370C. Cultures were fed weekly by removing one-half of the medium and replacing the spent medium with an equal volume of fresh medium. The cultures were maintained for 3 weeks. Fixation and Staining of Cultures. After 3 weeks of culture, the cells were fixed in 1% formaldehyde and then tested for reactivity with monoclonal antibody 23c6, by using a Vectastain ABC-AP kit as described (17). The nuclei were counterstained with 1% methyl green. In some experiments the cells Abbreviations: OIF, osteoinductive factor; TGF-(3, transforming growth factor (8; MNC, multinucleated cell; 1,25(OH)2D3, 1,25dihydroxyvitamin D3; PGE1, prostaglandin E1; PGE2, prostaglandin E2; PTH, parathyroid hormone. tTo whom reprint requests should be addressed at: Research Service (151), Audie Murphy VA Hospital, 7400 Merton Minter Boulevard, San Antonio, TX 78284.
3024
Proc. Natl. Acad. Sci. USA 87 (1990)
Medical Sciences: Kukita et al.
were fixed with 5% (vol/vol) glutaraldehyde and stained with Wright's stain to count total MNCs. Cells containing three or more nuclei that reacted with 23c6 were counted as osteoclast-like MNCs. Prostaglandin Determination. Supernatants from OIFtreated and control human bone marrow cultures at 1, 2, and 3 weeks were assayed for prostaglandins El and E2 (PGE1 and PGE2, respectively) by using the commercially available radioimmunoassay (Seragen). Samples were tested undiluted in duplicate and expressed as pg/100,41 by using B/Bo (where B is the number of cpm bound minus nonspecific binding and Bo is the total cpm). These assays can detect PGE, at 1.78 pg/100 p.l and PGE2 at 2.56 pg/100 Al. Statistical Analyses. Statistical significance was tested using two-way analysis of variance for repeated measures. Results are reported as the mean ± SEM for four determinations. Dunnett's T test was used to determine if OIF preferentially suppressed formation of MNCs reactive with monoclonal antibody 23c6.
RESULTS OIF inhibited total MNC formation in a dose-dependent manner in marrow cultures treated with 1,25(OH)2D3, a potent stimulator of MNC formation. The lowest concentration of OIF that significantly inhibited total MNC formation ranged from 0.1 to 10 ng/ml, depending on the particular bone marrow culture. OIF at 10 ng/ml significantly inhibited total MNC formation regardless of the source of the marrow tested. Results from one such experiment demonstrating inhibition of total MNC formation by OIF are shown in Fig. 1. We have shown (17) that 40-60% of total MNCs formed in long-term human marrow cultures react with monoclonal antibody 23c6, which preferentially binds osteoclasts in bone slices. This was confirmed in these experiments (Fig. 1; see Figs. 5 and 6). Therefore, we examined the effects of OIF on the formation of MNCs reacting with this antibody. OIF inhibited formation of 23c6-positive MNCs in 1,25(OH)2D3treated cultures in a dose-dependent manner (Fig. 1). OIF significantly decreased the percentage of MNCs reacting with monoclonal antibody 23c6 in these cultures. In cultures treated with 1,25(OH)2D3 alone, 48% of total MNCs reacted with monoclonal antibody 23c6 (Fig. 1). When the cultures were treated with OIF at 100 ng/ml, the percentage of MNCs reacting with 23c6 significantly decreased to 23% of total MNCs (Fig. 1). OIF (1-100 ng/ml) inhibited formation of
250 200 -
I
0)
1502
z
100-
50 -
0-
PTH
Control
PTH+OIF
FIG. 2. Effect of OIF on PTH-stimulated MNC formation. Bone marrow mononuclear cells were cultured for 3 weeks with PTH (50 ng/ml) in the presence or absence of OIF (40 ng/ml). Results are shown as the mean ± SEM for three or four cultures. **P < 0.01 compared to cultures treated with PTH alone.
23c6-positive MNCs to a significantly greater extent than formation of 23c6-negative MNCs (P < 0.05). OIF did not decrease the total number of mononuclear cells present in these cultures (659 vs. 669 mononuclear cells per 10 microscopic fields in control vs. OIF-treated cultures, respectively), suggesting that it did not inhibit total cell growth. A similar inhibition of MNC formation occurred when OIF was added to marrow cultures treated with PTH which we have shown (16) stimulates MNC formation (Fig. 2). These data demonstrate that OIF can affect MNC formation in cultures not treated with 1,25(OH)2D3. OIF significantly inhibited 23c6-positive MNC formation in marrow cultures stimulated with 1,25(OH)2D3 regardless of whether it was present for only the first, second, or third week of culture, and maximally inhibited 23c6-positive MNC formation when it was present for the entire 3 weeks of culture (Fig. 3). OIF also significantly decreased the number of nuclei per 23c6-positive MNC in cultures treated with 1,25(OH)2D3. In cultures with 10 nM 1,25(OH)2D3, there were 4.43 ± 0.2 nuclei per 23c6-positive MNC. In contrast, in cultures treated with 1,25(OH)2D3 and OIF (100 ng/ml), there were 3.23 ± 0.07 nuclei per 23c6-positive MNC (P < 0.01). We then measured the effects of OIF on PGE1 and PGE2 production in long-term cultures and tested the effects of indomethacin, which inhibits prostaglandin production, on 800-
2000-
%000D 1000-
= Total MNC - 23C6 Positive MNC
A1
i
LO.
400-
I
I
z
0
~200-
0*
500-
0.1
10
100
OIF (ng/ml)
-1W 2W
3W
FIG. 1. Effect of OIF on formation of total MNCs and MNC reactive with the osteoclast-specific monoclonal antibody 23c6 in cultures stimulated with 1,25D3. Bone marrow mononuclear cells were cultured with 1,25(OH)2D3 (10 nM) in the presence or absence of various concentrations of OIF for 3 weeks. Results are shown as mean ± SEM for four cultures. *P < 0.05 compared to cultures treated with 1,25(OH)2D3 alone. **P < 0.01 compared to cultures treated with 1,25(OH)2D3 alone.
Moda
moum
U
D3 D3 D3
D3+01F D3+01F D3+01F D3+01F
D3
D3 D3+0F
D3
D3
I1
D3
D3
D3 +OF
FIG. 3. Time course for the effect of OIF on MNC formation by cells stimulated with 1,25(OH)2D3. Bone marrow mononuclear cells were cultured with 1,25(OH)2D3 (10 nM; D3) for 3 weeks. OIF (50 ng/ml) was added only at the first, second, or third week of culture or for the entire 3 weeks of culture. Results are reported as the mean ± SEM for three cultures. **P < 0.01 compared to cultures treated with 1,25(OH)2D3 alone for 3 weeks.
Medical Sciences: Kukita et A
Proc. Nati. Acad. Sci. USA 87 (1990)
z 00-
~50 0
Contro
D3D3 +D3 OF
OIF
(ng/ml)
FIG. 4. Effect of indomethacin on the inhibition of total and
236-positive OIF-induced MNC formation in marrow cultures stimulated with 1,25(OH)2D3. Bone marrow mononuclear cells were
cultured for 3 weeks in the presence or absence of various concentrations of OIF. 1,25(OH)2D3 (10 nM) and indomethacin (1 AM) were added to all cultures and were present for the entire culture period. Results are shown as mean ± SEM for four cultures. *P < 0.05 compared to cultures without OIF. **P < 0.01 compared to cultures without OIF.
the inhibition of MNC formation by OIF. Several investigators have suggested that the effects of many osteotropic
factors are mediated by prostaglandin production (20-22), and we have shown (23) that prostaglandins inhibit MNC formation in human marrow cultures. Indomethacin (1 ,uM) did not block the inhibition by OIF of either total or 23c6positive MNC formation (Fig. 4). In cultures treated with 300and OIF, PGE, and PGE2 levels were 97 ± 5 and 1,25(OH)2D3 94 ± 3 pg/100 ,ul, respectively, compared to control cultures treated with 1,25(OH)AD3alone, which had 0 and 7 ± 4 pg/100 ,ul of PGE, and PGE2, respectively. This increase in PGE, and PGE2 levels induced by OIF was 10 times below the minimal concentration of PGE2 necessary to inhibit MNC formation in these cultures (24). Addition of 1AM indomethacin to these cultures treated with 1,25(OH)2D3 and OIF PGEo and PGE2levels to 31 2 and 13 ± 4 pg/100 decreased z
,ul, respectively. Since in vivo OIF acts synergistically with TGF-P1 to induce bone formation (1) and TGF-P1 inhibits MNC forma-
5001 400-
Total MNC IM 23CO Positive MNC
200X
Contrdl
63
D3
D3 +
D3+a1F
T(fg TGF + OOF
FIG. 5. Effect of OIF in combination with TGFion on MNC formation. Bone marrow mononuclear cells were cultured with 1,25(cH)2D3 (10 nM; D3) in the presence or absence ofTGF-vi (0.01 ng/ml), OIF (1 ng/ml), or both for 3 weeks. Results are shown as mean SEM for four cultures. *P < 0.05 compared to cultures treated with 1,25(OH)2D3, 1,25(OH)2D3/TGF-,B, or 1,25(OH)2D3/ cOF. **P < 0.01 compared to cultures treated with 1,25(OH)2D3
alone, 1,25(oH)2D3/TGF-nM
, or
1,25(oH)2D3/OIF.
3025
+
D3+OF+
Cabtonm Cactorin
FIG. 6. Effect of QIF in combination with calcitonin on MNC formation. Bone marrow mononuclear cells were cultured with 1,25(OH)2D3 (10 nM; D3) in the presence or absence of OIF (1 ng/ml) or salmon calcitonin (1 unit/ml) for 3 weeks. Results are shown as the mean ± SEM for four cultures. OIF, calcitonin, or both significantly decreased MNC formation compared to 1,25(OH)2D3 alone (P < 0.05).
tion in this culture system (25), we determined if TGF-81 could enhance the inhibitory effects of OIF on 23c6-positive MNC formation. Low concentrations of either OIF (1 ng/ml) or TGF-/3 (0.01 ng/ml) did not significantly inhibit 23c6positive MNC formation stimulated by 1,25(OH)2D3 in the marrow cultures shown in Fig. 5. However, when these low concentrations of OIF and TGF-/31 were added to the cultures simultaneously, MNC formation was almost inhibited completely, suggesting that OIF and TGF-/3 can act synergistically to inhibit both total and 23c6-positive MNC formation (Fig. 5). In contrast, when low concentrations of OIF, which did not significantly inhibit MNC formation in these cultures, were added simultaneously with calcitonin, OIF did not further enhance the calcitonin-induced inhibition of MNC formation (Fig. 6).
DISCUSSION In this report we demonstrate that OIF is a potent inhibitor of total and 23c6-positive MNC formation. OIF preferentially inhibited formation of MNCs reactive with monoclonal antibody 23c6, which identifies osteoclasts, and significantly decreased the percentage of these MNCs in the total MNC population. Monoclonal antibody 23c6 reacts only with osteoclasts in bone sections, reacts strongly with freshly isolated bone-derived osteoclasts, does not react with inflammatory polykaryons, and does not react with monocytes, macrophages, or macrophage polykaryons formed in vitro (refs. 26 and 27; N. Kurihara and G.D.R., unpublished work). Thus, monoclonal antibody 23c6 appears to preferentially react with cells in the osteoclast lineage. These data suggest that OIF preferentially inhibits formation of MNCs expressing the osteoclast phenotype but also inhibits to a lesser extent formation of MNCs that do not react with monoclonal 23c6. The relationship of MNCs that do not react with monoclonal antibody 23c6 to cells in the osteoclast lineage is currently unknown, but since it is likely that these cells are not in the osteoclast lineage, these data suggest that OIF may also affect formation of other types of polykaryons, albeit to a lesser extent. OIF shares several properties with TGF-,B1 in terms of its effects on MNC formation. Both TGF-j1 and OIF inhibit MNC formation stimulated by PTH. Addition of indomethacin (1 ,uM) did not affect the inhibition of MNC formation by OIF, although OIF did induce prostaglandin synthesis in these cultures. We have shown (28) that the effects of TGF-,B1 on MNC formation are not blocked by indometh-
3026
Medical Sciences: Kukita et al.
acin. These data suggest that the OIF and TGF-,B1 do not mediate their effects in this system by prostaglandin synthesis. However, the effects ofTGF-(31 and OIF differ in several respects. Treatment of the cultures with OIF in the first week or in the third week of the cultures inhibited MNC formation stimulated by 1,25(OH)2D3. We have shown (25) that TGF-f31 inhibited MNC formation predominantly, during the first week of culture. TGF-P1 mainly affected the precursor pool for MNCs and only had modest effects on the fusion of mononuclear precursor cells. In contrast, our results suggest that OIF inhibits all stages of 23c6-positive MNC formation, including fusion of the mononuclear precursor cells since it decreases the number of nuclei per MNC. The number of nuclei per MNC should reflect the relative rate of fusion of mononuclear precursors to form MNCs. Therefore, it was not surprising that TGF-,/1 and OIF in combination showed enhanced inhibition of 23c6-positive MNC formation compared to OIF or TGF-P1 alone. The lack of synergy between calcitonin and OIF on the inhibition of MNC formation suggests that these factors may act at similar steps or sites of osteoclast formation. Bone matrix contains both TGF-p and OF. During bone resorption both of these factors may be released and act together to inhibit further osteoclast formation. Pfeilschifter an Mundy (28) have shown that active TGF-f3 is generated in organ cultures during the bone resorptive process. When combined with TGF-P, OIF is effective in inhibiting 23c6positive MNC formation at concentrations 100 times lower than those required for OIF by itself to maximally inhibit 23c6-positive MNC formation. The small amounts of active TGF-,j and OIF released during bone resorption may be sufficient to inhibit osteoclast activity even when the concentration of OIF or TGF-f by itself is too low to affect osteoclast activity. Thus these data suggest that OIF may be an important regulator of osteoclast activity in vivo by inhibiting osteoclast formation. We thank J. Laderer for secretarial assistance. This work was supported by research funds from the Veterans Administration, Grant AM35188 from the National Institute of Diabetes and Digestive and Kidney Diseases, and Grant CA-40035 from the National Cancer Institute. G.D.R. is a recipient of a Clinical Investigator Award from the Veterans Administration Research Service.
1. Bentz, H., Nathan, R., Rosen, D., Armstrong, R., Thompson, A., Segarini, P., Mathews, M., Dasch, J., Piez, K. & Seyedin, S. (1989) Am. Soc. Cell Biol. 107, 162 (abstr.). 2. Bentz, H., Nathan, R., Rosen, D., Armstrong, R., Thompson, A., Segarini, P., Dasch, J., Piez, K. & Seyedin, S. (1990) J. Biol. Chem. 264, 20805-20811. 3. Assoian, R. K., Komoriya, A., Meyers, C. A., Miller, D. M. & Sporn, M. B. (1983) J. Biol. Chem. 258, 7155-7160. 4. Frolik, C. A., Dart, L. L., Meyers, C. A., Smith, D. M. & Sporn, M. B. (1983) Proc. Natl. Acad. Sci. USA 80, 3676-3680. 5. Roberts, A. B., Anzano, M. A., Meyers, C. A., Wideman, J., Blancher, R., Pan, Y. C. E., Stein, S., Lehrman, R., Smith, J. M., Lamb, L. C. & Sporn, M. B. (1983) Biochemistry 22, 5692-5698. 6. Seyedin, S. M., Thomas, T. C., Thompson, A. Y., Rosen,
Proc. Natl. Acad. Sci. USA 87 (1990)
7. 8.
9.
10. 11. 12. 13. 14.
15. 16. 17. 18.
19. 20. 21. 22.
23. 24.
25.
26. 27. 28.
29. 30. 31.
D. M. & Piez, K. A. (1985) Proc. Natl. Acad. Sci. USA 82, 2267-2271. Wrann, M., Bodmer, S., deMartin, R., Siepl, C., HoferWarbinek, R., Frei, K., Hofer, E. & Fontana, A. (1987) EMBO J. 6, 1633-1636. Masni, T. L., Wakefield, L., Lechner, J. F., LaVeck, M. A., Sporn, M. B. & Harris, C. C. (1986) Proc. Natl. Acad. Sci. USA 83, 2438-2442. Sporn, M. B., Roberts, A. B., Shull, J. H., Smith, J. M., Ward, J. M. & Sodek, J. (1983) Science 219, 1329-1331. Seyedin, S. M., Segarini, P. R., Rosen, D. M., Thompson, A. Y., Bentz, H. & Graycar, J. (1987) J. Biol. Chem. 262, 1946-1949. Laiho, M., Saksela, O., Andreasen, P. A. & Keski-Oja, J. (1986) J. Cell Biol. 103, 2403-2410. Ignotz, R. A., Endo, T. & Massague, J. (1987) J. Biol. Chem. 262, 6443-6446. Robey, P. G., Young, M. F., Flanders, K. C., Roche, N. S., Kondaiah, P., Reddi, A. H., Termine, J. D., Sporn, M. B. & Roberts, A. B. (1987) J. Cell Biol. 105, 321-325. Rosen, D. M., Stempien, S. A., Bonewald, L., Mundy, G., Bentz, H. & Seyedin, S. M. (1989) Am. Soc. Cell Biol. 107, 162 (abstr.). Pfeilschifter, J., Chenu, C., Bird, A., Mundy, G. R. & Roodman, G. D. (1989) J. Bone Miner. Res. 4, 113-118. MacDonald, B. R., Takahashi, N., McManus, L., Mundy, G. R. & Roodman, G. D. (1987) Endocrinology 120,2326-2333. Kukita, T., McManus, L. M., Miller, M., Civin, C. & Roodman, G. D. (1989) Lab. Invest. 60, 532-538. Takahashi, N., Kukita, T., MacDonald, B. R., Bird, A., Mundy, G. R. & Roodman, G. D. (1989) J. Clin. Invest. 83, 534-550. Roodman, G. D., Ibbotson, K. J., MacDonald, B. R., Kuehl, T. J. & Mundy, G. R. (1985) Proc. Natl. Acad. Sci. USA 82, 8213-8217. Tashjian, A. H. & Levine, L. (1978) Biochem. Biophys. Res. Commun. 85, 966-975. Tashjian, A. H., Voelkel, E. F., Lazzaro, M., Singer, F. R., Roberts, A. B., Derynck, R., Winkler, M. E. & Levine, L. (1985) Proc. Natl. Acad. Sci. USA 82, 4535-4538. Tashjian, A. H., Voelkel, A. H., Lazzaro, M., Goad, D., Bosma, T. & Levine, L. (1987) Endocrinology 120, 2029-2036. Chenu, C., Mundy, G. D. & Roodman, G. R. (1988) J. Bone Miner. Res. 3, 409 (abstr.). Chenu, C., Kukita, T., Mundy, G. R. & Roodman, G. D. (1990) J. Bone Miner. Res., in press. Chenu, C., Pfeilschifter, J., Mundy, G. R. & Roodman, G. D. (1988) Proc. Natl. Acad. Sci. USA 85, 5683-5687. Horton, M. A., Lewis, D., McNulty, K., Pringle, J. A. S. & Chambers, T. J. (1985) Cancer Res. 45, 5663-5669. Horton, M. A., Lewis, D., McNulty, K., Pringle, J. A. S. & Chambers, T. J. (1985) Br. J. Exp. Pathol. 66, 103-108. Pfeilschifter, J. & Mundy, G. (1987) Proc. Natl. Acad. Sci. USA 84, 2024-2028. Takahashi, N., Yamana, H., Yoshiki, S., Roodman, G. D., Mundy, G. R., Jones, S. J., Boyde, A. & Suda, T. (1988) Endocrinology 122, 1373-1382. Takahashi, N., MacDonald, B. R., Hon, J., Winkler, M. E., Derynck, R., Mundy, G. R. & Roodman, G. D. (1986) J. Clin. Invest. 78, 894-898, MacDonald, B. R., Mundy, G. R., Clark, S., Kuehl, T. J., Stanley, E. R. & Roodman, G. D. (1985) J. Bone Miner. Res. 1, 227-233.
