0013-7227/90/1263-1416$02.00/0 Endocrinology Copyright © 1990 by The Endocrine Society
Vol. 126, No. 3 Printed in U.S.A.
Leukemia Inhibitory Factor: A Novel Bone-Active Cytokine* I. R. REID, C. LOWE, J. CORNISH, S. J. M. SKINNER, D. J. HILTON, T. A. WILLSON, D. P. GEARING, AND T. J. MARTIN Departments of Medicine (I.R.R., J.C.) and Surgery (C.L., S.J.M.S.), University of Auckland, New Zealand; and the Walter and Eliza Hall Institute of Medical Research (D.J.H., T.A.W., D.P.G.), and St. Vincents Institute of Medical Research (T.J.M.), Melbourne, Australia M indomethacin. hLIF also stimulated incorporation of [3H] thymidine into calvaria, but the dose-response relationship was distinct from that for bone resorption, and this effect was not blocked by indomethacin. Similarly, hLIF increased [3H]phenylalanine incorporation into calvaria, and this was also not inhibited by indomethacin. It is concluded that LIF stimulates bone resorption by a mechanism involving prostaglandin production, but that a distinct mechanism is responsible for its stimulation of DNA and protein synthesis. The primary structure of LIF differs from that of other fully characterized, boneactive cytokines, and it, thus, represents a novel factor which may be involved in the normal regulation of bone cell function {Endocrinology 126: 1416-1420, 1990)
ABSTRACT. A number of cytokines have been found to be potent regulators of bone resorption and to share the properties originally attributed to osteoclast-activating factor. One such activity, differentiation-inducing factor (DIF, D-factor) from mouse spleen cells, shares a number of biological and biochemical properties with the recently characterized and cloned leukemia inhibitory factor (LIF). We have assessed the effects of recombinant LIF on bone resorption and other parameters in neonatal mouse calvaria. Both recombinant murine and human (h) LIFs stimulated 45Ca release from prelabeled calvaria in a dose-dependent manner. The increase in bone resorption was associated with an increase in the number of osteoclasts per mm2 bone. The osteolytic effects of hLIF were blocked by 10~7
I
N RECENT years it has become clear that cytokines play a key role in the regulation of bone cell function (1). One of the first bone-active cytokines to be described was osteoclast-activating factor (OAF), a potent stimulator of bone resorption which is secreted by leukocytes or spleen cells exposed to mitogens (2). Since its original description, it has become apparent that OAF is not a single entity, and both interleukin-1 (3, 4) and the tumor necrosis factors (5) have been found to have OAF-like properties. More recently, the OAF activity in conditioned media from mitogen-stimulated mouse spleen cells has been found to copurify with a factor that induces differentiation of the murine myeloid leukemic cell line, Ml (6). This factor, differentiation-inducing factor (DIF, D-factor); is also found in conditioned media from Erhlich ascites tumor cells and mouse fibroblast L929 cells, and purified preparations are associated with a glycoprotein of 62,000 mol wt (7). We have described another factor, leukemia inhibitory factor (LIF), which was isolated from medium condiReceived August 11,1989. Address all correspondence and request for reprints to: Dr. I. R. Reid, Department of Medicine, University of Auckland, Auckland, New Zealand. * This work was supported by the Medical Research Council of New Zealand and Amrad Corp. Ltd. (Melbourne, Australia).
tioned by murine Krebs II ascites tumor cells on the basis of its capacity to induce differentiation and suppress proliferation of Ml cells (8, 9). Purified LIF is a single chain glycoprotein of 58,000 mol wt (10). A murine LIF (mLIF) cDNA clone was isolated on the basis of partial amino acid sequence (8), and a human LIF (hLIF) genomic clone was subsequently isolated (11). The shared biochemical and biological characteristics of LIF and DIF led to the present assessment of the activity of LIF on bone in vitro. We have found that LIF not only stimulates bone resorption, but also increases the rates of DNA and protein synthesis in neonatal mouse calvaria. Materials and Methods Recombinant mLIF and hLIF were produced in yeast cells as previously described (8, 11) and purified to homogeneity according to the following strategy: 1) concentration and buffer exchange using a hollow fiber concentrator equipped with a 50,000 mol wt cut-off cartridge; 2) carboxymethyl-Sepharose chromatography employing a salt gradient in 100 mM acetate, pH 5.0; 3) DEAE-Sepharose chromatography employing a salt gradient in 10 mM Tris-HCl, pH 8.0; 4) gel filtration chromatography using a Sephacryl S-300 matrix in 50 mM Na2 HPO4 at pH 8.0. Fifty units per ml of LIF activity is defined as the concentration that induces 50% of Ml colonies to exhibit
1416 The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 06 July 2015. at 01:49 For personal use only. No other uses without permission. . All rights reserved.
1417
LIF ACTIONS ON BONE differentiation (9). The specific activity of the purified native peptide is 1-2 x 108 U/mg, thus 1000 U/ml is equivalent to 24 X 10"10 M. Recombinant tranforming growth factor-/?! (TGF/3i) was provided by Genentech, Inc. (South San Francisco, CA). Purified bovine PTH was the generous gift of Dr. Peter Barling, University of Auckland, New Zealand. Bone resorption studies were carried out in neonatal mouse calvaria, as previously described (12). Mice were injected sc with 5 /uCi 45Ca at 2 days of age, and hemicalvaria were dissected out 4 days later. Bones were preincubated for 24 h in Minimum Essential Medium with 1% charcoal-stripped heat-inactivated serum, then changed to fresh medium containing the experimental compounds. Incubation continued for a further 48 h (or 96 h; see text) before media and bones were harvested and their 45 Ca content determined by liquid scintillation counting. 45Ca was extracted from bones using 5% trichloroacetic acid. 45Ca release is expressed as a percentage of the total bone 45Ca content, except where data from more than one experiment have been combined, in which case treatment/control ratios are used. [3H]Thymidine incorporation was measured during the last 4 h of the 48-h incubation, using the method of Canalis (13). [3H]Phenylalanine incorporation was assessed using the same incubation and extraction techniques. Functional osteoclast numbers were determined by photography of each calvarium after supravital staining with neutral red (14). Each photograph was projected onto a grid, the number of osteolcasts taking up the dye counted, and the bone area determined. Data are presented as the mean ± SEM and have been analyzed by two-tailed t tests.
**
a
1.6
1
1.4
/ 1.2
o 1.0
•r 1000
2000
3000
4000
[mLIF] (units/ml)
Results Both recombinant mLIF and hLIF stimulated 45Ca release from prelabeled calvaria in a dose-dependent manner (Fig. 1). mLIF appeared to be more active than its human counterpart, but the observed differences are within the error of the bioassay used to determine LIF activity. The maximal stimulation of 45Ca released by hLIF was comparable to that produced by TGF/3 in this assay system (treatment/control ratio, 1.6 ± 0.2 with 50 ng/ml TGF/3) but substantially less than that elicited by PTH (treatment/control ratio, 3.2 ± 0.3 with 10~7 M bovine PTH). The stimulation of 45Ca release by hLIF was associated with an increase in the number of functional osteoclasts per mm2 calvarium, as determined by the uptake of neutral red (Fig. 2). A number of cytokines that stimulate bone resorption do so via stimulation of local prostaglandin production. To determine whether this was the case for hLIF, its effects on 45Ca release in the presence and absence of indomethancin 10"7 M were assessed (Fig. 3). hLIF did not significantly influence bone resorption in the presence of this inhibitor of prostaglandin synthesis. The effect of mLIF on [3H]thymidine incorporation
100
1000
10000
[hLIF] (units/ml)
FlG. 1. Dose-response relationship for 46Ca release from neonatal mouse calvaria during a 48-h incubation with either recombinant mouse LIF (a) or recombinant hLIF (b). Results are expressed as treatment/ control ratios, and the significance of differences from control values are as follows: a) *, P < 0.01; **, P < 0.005; b) *, P < 0.05; **, P < 0.01; ***, P < 0.001. n = 4-6 in each group. Each data point was confirmed in at least two experiments.
into calvaria is shown in Fig. 4. mLIF produced a dosedependent stimulation of [3H]thymidine incorporation, but the dose-response relationship was clearly distinct from that for 45Ca release, determined in the same bones (Fig. la). The effects of lower concentrations of hLIF on [3H]thymidine incorporation are shown in Table 1 and contrasted with the changes in 45Ca release measured in the same bones. [3H]Thymidine incorporation is stimulated by concentrations of hLIF from 100-1000 U/ml,
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 06 July 2015. at 01:49 For personal use only. No other uses without permission. . All rights reserved.
Endo • 1990 Voll26«No3
LIF ACTIONS ON BONE
1418
TABLE 1. Effect of low concentrations of hLIF on [3H]thymidine incorporation and 46Ca release in neonatal mouse calvaria
- E « p
o 0)
Control
hLIF 7500 u/ml
hLIF (U/ml)
[3H]Thymidine incorporation (treatment/control)
100 300 1000
1.31 ± 0.12° 1.17 ± 0.07° 1.38 ± 0.086
46 Ca release (treatment/control)
1.01 ± 0.07 1.04 ± 0.06 1.05 ± 0.07 Data are the mean ± SEM and are given as treatment/control ratios for ease of comparison of the two indices. Significant differences from control are shown, n = 3-4 in each group. Similar results were obtained in two other experiments. 0 P < 0.05. 6 P
Vol. 126, No. 3 Printed in U.S.A.
Leukemia Inhibitory Factor: A Novel Bone-Active Cytokine* I. R. REID, C. LOWE, J. CORNISH, S. J. M. SKINNER, D. J. HILTON, T. A. WILLSON, D. P. GEARING, AND T. J. MARTIN Departments of Medicine (I.R.R., J.C.) and Surgery (C.L., S.J.M.S.), University of Auckland, New Zealand; and the Walter and Eliza Hall Institute of Medical Research (D.J.H., T.A.W., D.P.G.), and St. Vincents Institute of Medical Research (T.J.M.), Melbourne, Australia M indomethacin. hLIF also stimulated incorporation of [3H] thymidine into calvaria, but the dose-response relationship was distinct from that for bone resorption, and this effect was not blocked by indomethacin. Similarly, hLIF increased [3H]phenylalanine incorporation into calvaria, and this was also not inhibited by indomethacin. It is concluded that LIF stimulates bone resorption by a mechanism involving prostaglandin production, but that a distinct mechanism is responsible for its stimulation of DNA and protein synthesis. The primary structure of LIF differs from that of other fully characterized, boneactive cytokines, and it, thus, represents a novel factor which may be involved in the normal regulation of bone cell function {Endocrinology 126: 1416-1420, 1990)
ABSTRACT. A number of cytokines have been found to be potent regulators of bone resorption and to share the properties originally attributed to osteoclast-activating factor. One such activity, differentiation-inducing factor (DIF, D-factor) from mouse spleen cells, shares a number of biological and biochemical properties with the recently characterized and cloned leukemia inhibitory factor (LIF). We have assessed the effects of recombinant LIF on bone resorption and other parameters in neonatal mouse calvaria. Both recombinant murine and human (h) LIFs stimulated 45Ca release from prelabeled calvaria in a dose-dependent manner. The increase in bone resorption was associated with an increase in the number of osteoclasts per mm2 bone. The osteolytic effects of hLIF were blocked by 10~7
I
N RECENT years it has become clear that cytokines play a key role in the regulation of bone cell function (1). One of the first bone-active cytokines to be described was osteoclast-activating factor (OAF), a potent stimulator of bone resorption which is secreted by leukocytes or spleen cells exposed to mitogens (2). Since its original description, it has become apparent that OAF is not a single entity, and both interleukin-1 (3, 4) and the tumor necrosis factors (5) have been found to have OAF-like properties. More recently, the OAF activity in conditioned media from mitogen-stimulated mouse spleen cells has been found to copurify with a factor that induces differentiation of the murine myeloid leukemic cell line, Ml (6). This factor, differentiation-inducing factor (DIF, D-factor); is also found in conditioned media from Erhlich ascites tumor cells and mouse fibroblast L929 cells, and purified preparations are associated with a glycoprotein of 62,000 mol wt (7). We have described another factor, leukemia inhibitory factor (LIF), which was isolated from medium condiReceived August 11,1989. Address all correspondence and request for reprints to: Dr. I. R. Reid, Department of Medicine, University of Auckland, Auckland, New Zealand. * This work was supported by the Medical Research Council of New Zealand and Amrad Corp. Ltd. (Melbourne, Australia).
tioned by murine Krebs II ascites tumor cells on the basis of its capacity to induce differentiation and suppress proliferation of Ml cells (8, 9). Purified LIF is a single chain glycoprotein of 58,000 mol wt (10). A murine LIF (mLIF) cDNA clone was isolated on the basis of partial amino acid sequence (8), and a human LIF (hLIF) genomic clone was subsequently isolated (11). The shared biochemical and biological characteristics of LIF and DIF led to the present assessment of the activity of LIF on bone in vitro. We have found that LIF not only stimulates bone resorption, but also increases the rates of DNA and protein synthesis in neonatal mouse calvaria. Materials and Methods Recombinant mLIF and hLIF were produced in yeast cells as previously described (8, 11) and purified to homogeneity according to the following strategy: 1) concentration and buffer exchange using a hollow fiber concentrator equipped with a 50,000 mol wt cut-off cartridge; 2) carboxymethyl-Sepharose chromatography employing a salt gradient in 100 mM acetate, pH 5.0; 3) DEAE-Sepharose chromatography employing a salt gradient in 10 mM Tris-HCl, pH 8.0; 4) gel filtration chromatography using a Sephacryl S-300 matrix in 50 mM Na2 HPO4 at pH 8.0. Fifty units per ml of LIF activity is defined as the concentration that induces 50% of Ml colonies to exhibit
1416 The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 06 July 2015. at 01:49 For personal use only. No other uses without permission. . All rights reserved.
1417
LIF ACTIONS ON BONE differentiation (9). The specific activity of the purified native peptide is 1-2 x 108 U/mg, thus 1000 U/ml is equivalent to 24 X 10"10 M. Recombinant tranforming growth factor-/?! (TGF/3i) was provided by Genentech, Inc. (South San Francisco, CA). Purified bovine PTH was the generous gift of Dr. Peter Barling, University of Auckland, New Zealand. Bone resorption studies were carried out in neonatal mouse calvaria, as previously described (12). Mice were injected sc with 5 /uCi 45Ca at 2 days of age, and hemicalvaria were dissected out 4 days later. Bones were preincubated for 24 h in Minimum Essential Medium with 1% charcoal-stripped heat-inactivated serum, then changed to fresh medium containing the experimental compounds. Incubation continued for a further 48 h (or 96 h; see text) before media and bones were harvested and their 45 Ca content determined by liquid scintillation counting. 45Ca was extracted from bones using 5% trichloroacetic acid. 45Ca release is expressed as a percentage of the total bone 45Ca content, except where data from more than one experiment have been combined, in which case treatment/control ratios are used. [3H]Thymidine incorporation was measured during the last 4 h of the 48-h incubation, using the method of Canalis (13). [3H]Phenylalanine incorporation was assessed using the same incubation and extraction techniques. Functional osteoclast numbers were determined by photography of each calvarium after supravital staining with neutral red (14). Each photograph was projected onto a grid, the number of osteolcasts taking up the dye counted, and the bone area determined. Data are presented as the mean ± SEM and have been analyzed by two-tailed t tests.
**
a
1.6
1
1.4
/ 1.2
o 1.0
•r 1000
2000
3000
4000
[mLIF] (units/ml)
Results Both recombinant mLIF and hLIF stimulated 45Ca release from prelabeled calvaria in a dose-dependent manner (Fig. 1). mLIF appeared to be more active than its human counterpart, but the observed differences are within the error of the bioassay used to determine LIF activity. The maximal stimulation of 45Ca released by hLIF was comparable to that produced by TGF/3 in this assay system (treatment/control ratio, 1.6 ± 0.2 with 50 ng/ml TGF/3) but substantially less than that elicited by PTH (treatment/control ratio, 3.2 ± 0.3 with 10~7 M bovine PTH). The stimulation of 45Ca release by hLIF was associated with an increase in the number of functional osteoclasts per mm2 calvarium, as determined by the uptake of neutral red (Fig. 2). A number of cytokines that stimulate bone resorption do so via stimulation of local prostaglandin production. To determine whether this was the case for hLIF, its effects on 45Ca release in the presence and absence of indomethancin 10"7 M were assessed (Fig. 3). hLIF did not significantly influence bone resorption in the presence of this inhibitor of prostaglandin synthesis. The effect of mLIF on [3H]thymidine incorporation
100
1000
10000
[hLIF] (units/ml)
FlG. 1. Dose-response relationship for 46Ca release from neonatal mouse calvaria during a 48-h incubation with either recombinant mouse LIF (a) or recombinant hLIF (b). Results are expressed as treatment/ control ratios, and the significance of differences from control values are as follows: a) *, P < 0.01; **, P < 0.005; b) *, P < 0.05; **, P < 0.01; ***, P < 0.001. n = 4-6 in each group. Each data point was confirmed in at least two experiments.
into calvaria is shown in Fig. 4. mLIF produced a dosedependent stimulation of [3H]thymidine incorporation, but the dose-response relationship was clearly distinct from that for 45Ca release, determined in the same bones (Fig. la). The effects of lower concentrations of hLIF on [3H]thymidine incorporation are shown in Table 1 and contrasted with the changes in 45Ca release measured in the same bones. [3H]Thymidine incorporation is stimulated by concentrations of hLIF from 100-1000 U/ml,
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 06 July 2015. at 01:49 For personal use only. No other uses without permission. . All rights reserved.
Endo • 1990 Voll26«No3
LIF ACTIONS ON BONE
1418
TABLE 1. Effect of low concentrations of hLIF on [3H]thymidine incorporation and 46Ca release in neonatal mouse calvaria
- E « p
o 0)
Control
hLIF 7500 u/ml
hLIF (U/ml)
[3H]Thymidine incorporation (treatment/control)
100 300 1000
1.31 ± 0.12° 1.17 ± 0.07° 1.38 ± 0.086
46 Ca release (treatment/control)
1.01 ± 0.07 1.04 ± 0.06 1.05 ± 0.07 Data are the mean ± SEM and are given as treatment/control ratios for ease of comparison of the two indices. Significant differences from control are shown, n = 3-4 in each group. Similar results were obtained in two other experiments. 0 P < 0.05. 6 P
