0013-7227/90/1266-3069$02.00/0 Endocrinology Copyright © 1990 by The Endocrine Society
Vol. 126, No. 6 Printed in U.S.A.
Inhibitory Effects of the Bone-Derived Growth Factors Osteoinductive Factor and Transforming Growth Factor-/? on Isolated Osteoclasts* R. O. C. OREFFO, L. BONEWALD, A. KUKITA, I. R. GARRETT, S. M. SEYEDIN, D. ROSEN, AND G. R. MUNDY University of Texas Health Science Center, Department of Medicine, Division of Endocrinology and Metabolism, San Antonio, Texas 78284-7877; and Collagen Corp. (S.M.S., D.R.), Palo Alto, California 94303
ABSTRACT. Demineralized bone matrix contains a number of growth factors for osteoblast-like cells. Two of these, the novel glycoprotein osteoinductive factor (OIF) and transforming growth factor-/3 (TGF/3), act together to cause ectopic bone formation in vivo. Since OIF, like TGF/3, is likely released from bone when the matrix is resorbed, we examined the effects of homogeneous OIF and TGF0 on osteoclast function. Osteoclast function was tested in isolated avian osteoclasts and was measured in terms of tartrate-resistant acid phosphatase (TRAP) activity, oxygen-derived free radical production, and formation of characteristic resorption lacunae on slices of sperm whale dentine. OIF (50-100 ng/ml) inhibited the capacity of these osteoclasts to form lacunae whether assessed by the number of
excavations per slice or by the total area resorbed. OIF (10-100 ng/ml) or TGFjS (10-20 ng/ml) caused a decrease in TRAP activity as well as a reduction in oxygen-derived free radical generation detected by nitroblue tetrazolium staining. TGF/3 had no effect on the resorption capacity of isolated osteoclasts in concentrations that inhibited TRAP activity and nitroblue tetrazolium staining. These results suggest that growth regulatory factors, such as OIF and TGF/3, released during the resorption of bone may be endogenous inhibitors of continued osteoclastic activity. This cessation of osteoclast activity may be an essential preliminary step to the new bone formation that occurs at resorption sites during bone remodeling. {Endocrinology 126: 3069-3075, 1990)
r I THERE are a number of biologically active growth JL factors sequestered in bone that may play important regulatory roles in normal bone remodeling (1-3). Transforming growth factor-/? (TGF/3), which is the most abundant of these, has potent effects on the proliferation and differentiated function of cells of both the osteoblastic and osteoclastic phenotype (4-6). Moreover, we have found that resorbing bone releases active TGF/3 (7), and that stimulated osteoclasts cause the direct release of active TGF/3 from the bone-derived latent TGF/3 complex (8). A unique 25-kDa glycoprotein called osteoinductive factor (OIF) has recently been isolated from bovine bone (9, 10). This glycoprotein has the properties of a bone morphogenetic protein. When OIF is implanted together with TGF/3I or -II into sc tissue of rats in vivo with a collagen carrier, the formation of ectopic bone is induced. Since both of these factors are likely to be released from Received December 14, 1989. Address requests for reprints to: Dr. Gregory R. Mundy, Department of Medicine, University of Texas Health Science Center, 7703 Floyd Curl Drive, San Antonio, Texas 78284. * This work was supported in part by NIH Grants AR-28149, AR39357, DEO-856901, CA-40035, and DE-08569 and Welch Foundation Grant AQ-1123.
the bone matrix during bone remodeling, we examined the effects of OIF and TGF/3 on isolated osteoclasts in culture. We found that both OIF and TGF0 inhibited osteoclast activity, as assessed by their effects on acid phosphatase activity and oxygen-radical production, detected by staining with nitroblue tetrazolium (NBT). OIF also inhibited the capacity of isolated osteoclasts to form resorption lacunae on sperm whale dentine, whereas TGF/3 had no such effect. These results suggest that regulatory growth factors sequestered in the bone matrix, such as OIF and TGF/3, may not only affect osteoblast activity, but may also inhibit osteoclast activity. This latter effect may be an important preliminary step to the coupling of new bone formation to previous bone resorption that occurs during the normal bone remodeling process.
Materials and Methods Materials TGFjSI and -II and OIF were purified from bovine demineralized bone, as previously described (10, 11). Dulbecco's Minimum Essential Medium (DMEM), Minimum Essential Medium a-modification («MEM), medium 199, fetal calf serum (FCS), penicillin-streptomycin solution, and tissue culture re-
3069
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 November 2015. at 12:00 For personal use only. No other uses without permission. . All rights reserved.
EFFECTS OF OIF AND TGF-/3 ON OSTEOCLASTS
3070
agents were purchased from Gibco (Grand Island, NY). Vitamin A, NBT (grade III), and all other reagents were purchased from Sigma (St. Louis, MO). All stock solutions of reagents were diluted 1:1000 in medium and made fresh immediately before use. Isolation of avian osteoclasts
Avian osteoclasts were isolated from medullary bone of laying white leghorn hens (Gallus domesticus; Kingswheel, OH), as previously described (12). In brief, bone marrow suspensions from the medullary bone of femora were filtered through Nytex cloth (110 nm; Tetko, Elmsford, NY) and centrifuged for 5 min at 1200 rpm, and the cell pellet was resuspended in 0.2% NaCl for 2 min to remove red blood cells. After layering the cells on 75% FCS for 1-1.5 h, sedimented cells were further filtered through Nytex filters (55 jim). Cells were harvested, resuspended in aMEM (Gibco) containing 10% FCS with penicillin (100 U/ml), streptomycin (100 /ug/ml), and arabinose-/3-D-cytosine furanoside (ara-C; 3 iig/ra\) to inhibit proliferation of nonosteoclastic cells. Since the presence of ara-C inhibited lacunar formation by osteoclasts, ara-C was not included in those experiments in which resorption was being assessed. Cells were plated in 12-well plates (Costar, Cambridge, MA) at 1 X 104 cells/well and incubated at 37 C in 5% CO2 humidified air for 48 h, which allows time for osteoclasts to adhere. After this, they were washed to remove nonadherent cells. The cells were then incubated with or without retinol (10~6 M) for 24 h. Factor was then added, and experiments were terminated after a further 48 h. Quantitation of oxygen radical production by NBT staining NBT (Sigma grade III) was dissolved in PBS, pH 7.4 (2 mg/ ml). Isolated avian osteoclasts in 12-well plates were incubated with 600 MI of this NBT solution at 37 C in a 5% CO2 humidified atmosphere for 30 min, after which this solution was aspirated, and the plates were fixed in 10% phosphate-buffered formalin. The plates were then rinsed twice with distilled, deionized H2O and allowed to dry. Three hundred microliters of 2 M KOH were added to each well to dissolve the cells after intense mixing. Approximately 10 n\ were removed for quantitation of protein before adding 300 fi\ dimethylsulfoxide to completely dissolve the NBT crystals. Two hundred microliters of the dissolved sample were added to a 96-well plate and read at 600 nm using a Flow Multiscan II plate reader (Flow Laboratories, Rockville, MD). Specific activity was determined as follows. The protein standard consists of 1-4 ;ug human immunoglobulin G (Bio-Rad, Richmond, CA)/160 n\ dH20/well in triplicate in a 96-well plate with percent KOH taken into account. BioRad protein reagent (40 ^I/well) was then added and immediately mixed (14). The plate was read at 600 nm on a Flow Multiscan II plate reader within 5-30 min after mixing. The standard curve for reduced NBT consisted of 0.2-2.0 ^g/well in 2 M KOH-dimethylsulfoxide (1:1) in 96-well plates. The plate was read at 600 mn. Specific activity was calculated as micromolar concentrations of NBT per ng protein.
Endo-1990 Vol 126 «No 6
Quantitation of tartrate- resistant acid phosphatase (TRAP) by fluorescence spectroscopy Osteoclast TRAP activity was measured using fluorescence spectroscopy, as described previously (15) with minor modifications. In brief, media from osteoclast cultures were harvested and stored at —20 C until ready for assay. The cells were washed twice with PBS and harvested by scraping with a rubber policeman in 0.5 ml Triton X-100 (0.1%, wt/vol). Aliquots of media or lysate (30 /xl) were incubated with 170 n\ 2 mM methylumbelliferyl phosphate, pH 5.0, in 0.48 M acetate buffer (0.48 M sodium acetate and 0.48 M acetic acid, pH 5.0) and 20 mM tartaric acid. Samples were incubated for 30 min at 37 C, and the reaction was terminated with 1.0 ml stop solution containing 50 mM glycine-5 mM EDTA, pH 10.4. Fluorescence was measured at emission 448 and excitation 360 nm using a fluorimeter (Fluoroskan, Flow Instruments). Enzyme activity was expressed as micromoles of methylumbelliferyl phosphate hydrolyzed per ng protein/min. The protein content was measured as described above for NBT specific activity. Effects of isolated osteoclasts on calcified matrices
Quantitation of the effects of isolated osteoclasts on calcified matrices was performed using minor modifications of the disaggregated osteoclast resorption assay, as described by Boyde et al. (16). Avian osteoclasts, isolated as described above, were dispersed on prewetted (aMEM plus 5% fetal calf serum) slices of sperm whale dentine for 48 h before washing and addition of factor. Slices of sperm whale dentine (0.25 x 4 x 6 mm) were prepared using a Buehler low speed diamond saw (Buehler, Lake Bluff, IL), followed by sonication (15 min) in several changes of distilled water. Slices were sonicated in ethanol and dipped in acetone to air dry before storing. Before all experiments, slices were transferred through ethanol to culture medium (aMEM plus 5% FCS) for 1-2 h. Cultures were performed in 48-well microtiter plates in humidified air (10% CO2) at 37 C in aMEM medium (1 slice/well in 400 /x\ medium). A single experiment used 16-20 slices, with a minimum of 3-4 slices/ treatment. Experiments were repeated at least 3-4 times. After 72-h incubation with factor the experiments were stopped by placing the slices in 2% glutaraldehyde-0.1 M sodium cacodylate buffer, pH 7.3. Osteoclasts were examined by toluidine blue (0.1%, wt/vol) staining and light microscopy. For examination of resorption lacunae, the osteoclasts were removed by sonication in distilled H2O, followed by ethanol for 10 min before restaining with toluidine blue. Lacunae were counted using light microscopy, and the plan area of matrix resorbed was quantitated using a computer-assisted morphometric program on a Quantex image and process analysis system (Quantex Instruments, Sunnybrook, CA). Data were analyzed by pattern analysis of variance, described below. Fetal long bone assay
Fetal rat long bone assays were performed as previously described (17). Pregnant rats were injected with 200 ^tCi 45Ca on the 18th day of gestation. The following day the mothers were killed, and fetuses were removed. The mineralized shafts of the radii and ulnae were dissected free of cartilaginous tissue
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 November 2015. at 12:00 For personal use only. No other uses without permission. . All rights reserved.
EFFECTS OF OIF AND TGF-0 ON OSTEOCLASTS
3071
and incubated in BGJb medium (Sigma) for 24 h at 37 C in a humidified atmosphere of 5% CO2 and 95% air to allow for the exchange of loosely bound 45Ca with stable calcium in the medium. Bones were cultured for a further 48-120 h in BGJb medium supplemented with 1 mg/ml BSA with penicillin and streptomycin (RIA grade, Sigma) in the presence of test or control substances. Bone-resorbing activity was expressed as the percentage of the total 45Ca released into the medium. Statistical analysis Data were analyzed by pattern analysis of variance with the Student-Newman-Keuls test, using a statistical package for the IBM PC (SAS Institute, Inc., Cary, NC). Due to the nonnormality of the data in studies on the formation of resorption lacunae avian osteoclasts and the inequality of the variances of the groups, a log transformation was performed on all data before analysis. In all cases an analysis of variance was performed to determine the level of significance for the groups being tested. If a difference between groups was indicated by the analysis, a Student-Newman-Keuls test was performed to determine where the differences occurred.
Control
10
25
SO
100
OIF (ng/ml)
120
tr
z
100 80
Q_
60
Results The capacity of OIF to affect pit formation by disaggregated osteoclasts was assessed using avian osteoclasts cultured on sperm whale dentine. OIF at 50 and 100 ng/ ml consistently reduced the plan area of matrix resorbed by avian osteoclasts compared with that of the controls. Figure 1 shows results from all of the experiments we performed. Partition analysis on data from all experiments (8) showed significant decreases in area resorbed compared with that of the controls. OIF (50-100 ng/ml) caused a similar decrease in the number of excavations per slice. When the total plan area resorbed per slice was plotted against the number of excavations per slice (Fig. 2), a positive correlation was shown, suggesting that the decrease in total area resorbed was due to a decrease in pit numbers rather than a decrease in lacunae area per individual pit. We also tested the effects of OIF on osteoclast activity, as assessed by two other markers, the activity of the enzyme TRAP and staining with the dye NBT, which is reduced to a blue-insoluble formazan in the presence of oxygen-derived free radicals. NBT staining of osteoclasts occurs in response to factors that stimulate osteoclast activity, such as PTH, 1,25-dihydroxyvitamin D, and interleukin-1, and is decreased by factors that inhibit osteoclasts, such as calcitonin (18). OIF at a concentration of 100 ng/ml inhibited TRAP activity in both osteoclast lysates as well as osteoclast-conditioned medium (Table 1). This effect was seen in both unstimulated cells and cells activated by incubation with vitamin A. Vitamin A is the only well characterized factor known to activate isolated osteoclasts directly (19). OIF at concen-
Vol. 126, No. 6 Printed in U.S.A.
Inhibitory Effects of the Bone-Derived Growth Factors Osteoinductive Factor and Transforming Growth Factor-/? on Isolated Osteoclasts* R. O. C. OREFFO, L. BONEWALD, A. KUKITA, I. R. GARRETT, S. M. SEYEDIN, D. ROSEN, AND G. R. MUNDY University of Texas Health Science Center, Department of Medicine, Division of Endocrinology and Metabolism, San Antonio, Texas 78284-7877; and Collagen Corp. (S.M.S., D.R.), Palo Alto, California 94303
ABSTRACT. Demineralized bone matrix contains a number of growth factors for osteoblast-like cells. Two of these, the novel glycoprotein osteoinductive factor (OIF) and transforming growth factor-/3 (TGF/3), act together to cause ectopic bone formation in vivo. Since OIF, like TGF/3, is likely released from bone when the matrix is resorbed, we examined the effects of homogeneous OIF and TGF0 on osteoclast function. Osteoclast function was tested in isolated avian osteoclasts and was measured in terms of tartrate-resistant acid phosphatase (TRAP) activity, oxygen-derived free radical production, and formation of characteristic resorption lacunae on slices of sperm whale dentine. OIF (50-100 ng/ml) inhibited the capacity of these osteoclasts to form lacunae whether assessed by the number of
excavations per slice or by the total area resorbed. OIF (10-100 ng/ml) or TGFjS (10-20 ng/ml) caused a decrease in TRAP activity as well as a reduction in oxygen-derived free radical generation detected by nitroblue tetrazolium staining. TGF/3 had no effect on the resorption capacity of isolated osteoclasts in concentrations that inhibited TRAP activity and nitroblue tetrazolium staining. These results suggest that growth regulatory factors, such as OIF and TGF/3, released during the resorption of bone may be endogenous inhibitors of continued osteoclastic activity. This cessation of osteoclast activity may be an essential preliminary step to the new bone formation that occurs at resorption sites during bone remodeling. {Endocrinology 126: 3069-3075, 1990)
r I THERE are a number of biologically active growth JL factors sequestered in bone that may play important regulatory roles in normal bone remodeling (1-3). Transforming growth factor-/? (TGF/3), which is the most abundant of these, has potent effects on the proliferation and differentiated function of cells of both the osteoblastic and osteoclastic phenotype (4-6). Moreover, we have found that resorbing bone releases active TGF/3 (7), and that stimulated osteoclasts cause the direct release of active TGF/3 from the bone-derived latent TGF/3 complex (8). A unique 25-kDa glycoprotein called osteoinductive factor (OIF) has recently been isolated from bovine bone (9, 10). This glycoprotein has the properties of a bone morphogenetic protein. When OIF is implanted together with TGF/3I or -II into sc tissue of rats in vivo with a collagen carrier, the formation of ectopic bone is induced. Since both of these factors are likely to be released from Received December 14, 1989. Address requests for reprints to: Dr. Gregory R. Mundy, Department of Medicine, University of Texas Health Science Center, 7703 Floyd Curl Drive, San Antonio, Texas 78284. * This work was supported in part by NIH Grants AR-28149, AR39357, DEO-856901, CA-40035, and DE-08569 and Welch Foundation Grant AQ-1123.
the bone matrix during bone remodeling, we examined the effects of OIF and TGF/3 on isolated osteoclasts in culture. We found that both OIF and TGF0 inhibited osteoclast activity, as assessed by their effects on acid phosphatase activity and oxygen-radical production, detected by staining with nitroblue tetrazolium (NBT). OIF also inhibited the capacity of isolated osteoclasts to form resorption lacunae on sperm whale dentine, whereas TGF/3 had no such effect. These results suggest that regulatory growth factors sequestered in the bone matrix, such as OIF and TGF/3, may not only affect osteoblast activity, but may also inhibit osteoclast activity. This latter effect may be an important preliminary step to the coupling of new bone formation to previous bone resorption that occurs during the normal bone remodeling process.
Materials and Methods Materials TGFjSI and -II and OIF were purified from bovine demineralized bone, as previously described (10, 11). Dulbecco's Minimum Essential Medium (DMEM), Minimum Essential Medium a-modification («MEM), medium 199, fetal calf serum (FCS), penicillin-streptomycin solution, and tissue culture re-
3069
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 November 2015. at 12:00 For personal use only. No other uses without permission. . All rights reserved.
EFFECTS OF OIF AND TGF-/3 ON OSTEOCLASTS
3070
agents were purchased from Gibco (Grand Island, NY). Vitamin A, NBT (grade III), and all other reagents were purchased from Sigma (St. Louis, MO). All stock solutions of reagents were diluted 1:1000 in medium and made fresh immediately before use. Isolation of avian osteoclasts
Avian osteoclasts were isolated from medullary bone of laying white leghorn hens (Gallus domesticus; Kingswheel, OH), as previously described (12). In brief, bone marrow suspensions from the medullary bone of femora were filtered through Nytex cloth (110 nm; Tetko, Elmsford, NY) and centrifuged for 5 min at 1200 rpm, and the cell pellet was resuspended in 0.2% NaCl for 2 min to remove red blood cells. After layering the cells on 75% FCS for 1-1.5 h, sedimented cells were further filtered through Nytex filters (55 jim). Cells were harvested, resuspended in aMEM (Gibco) containing 10% FCS with penicillin (100 U/ml), streptomycin (100 /ug/ml), and arabinose-/3-D-cytosine furanoside (ara-C; 3 iig/ra\) to inhibit proliferation of nonosteoclastic cells. Since the presence of ara-C inhibited lacunar formation by osteoclasts, ara-C was not included in those experiments in which resorption was being assessed. Cells were plated in 12-well plates (Costar, Cambridge, MA) at 1 X 104 cells/well and incubated at 37 C in 5% CO2 humidified air for 48 h, which allows time for osteoclasts to adhere. After this, they were washed to remove nonadherent cells. The cells were then incubated with or without retinol (10~6 M) for 24 h. Factor was then added, and experiments were terminated after a further 48 h. Quantitation of oxygen radical production by NBT staining NBT (Sigma grade III) was dissolved in PBS, pH 7.4 (2 mg/ ml). Isolated avian osteoclasts in 12-well plates were incubated with 600 MI of this NBT solution at 37 C in a 5% CO2 humidified atmosphere for 30 min, after which this solution was aspirated, and the plates were fixed in 10% phosphate-buffered formalin. The plates were then rinsed twice with distilled, deionized H2O and allowed to dry. Three hundred microliters of 2 M KOH were added to each well to dissolve the cells after intense mixing. Approximately 10 n\ were removed for quantitation of protein before adding 300 fi\ dimethylsulfoxide to completely dissolve the NBT crystals. Two hundred microliters of the dissolved sample were added to a 96-well plate and read at 600 nm using a Flow Multiscan II plate reader (Flow Laboratories, Rockville, MD). Specific activity was determined as follows. The protein standard consists of 1-4 ;ug human immunoglobulin G (Bio-Rad, Richmond, CA)/160 n\ dH20/well in triplicate in a 96-well plate with percent KOH taken into account. BioRad protein reagent (40 ^I/well) was then added and immediately mixed (14). The plate was read at 600 nm on a Flow Multiscan II plate reader within 5-30 min after mixing. The standard curve for reduced NBT consisted of 0.2-2.0 ^g/well in 2 M KOH-dimethylsulfoxide (1:1) in 96-well plates. The plate was read at 600 mn. Specific activity was calculated as micromolar concentrations of NBT per ng protein.
Endo-1990 Vol 126 «No 6
Quantitation of tartrate- resistant acid phosphatase (TRAP) by fluorescence spectroscopy Osteoclast TRAP activity was measured using fluorescence spectroscopy, as described previously (15) with minor modifications. In brief, media from osteoclast cultures were harvested and stored at —20 C until ready for assay. The cells were washed twice with PBS and harvested by scraping with a rubber policeman in 0.5 ml Triton X-100 (0.1%, wt/vol). Aliquots of media or lysate (30 /xl) were incubated with 170 n\ 2 mM methylumbelliferyl phosphate, pH 5.0, in 0.48 M acetate buffer (0.48 M sodium acetate and 0.48 M acetic acid, pH 5.0) and 20 mM tartaric acid. Samples were incubated for 30 min at 37 C, and the reaction was terminated with 1.0 ml stop solution containing 50 mM glycine-5 mM EDTA, pH 10.4. Fluorescence was measured at emission 448 and excitation 360 nm using a fluorimeter (Fluoroskan, Flow Instruments). Enzyme activity was expressed as micromoles of methylumbelliferyl phosphate hydrolyzed per ng protein/min. The protein content was measured as described above for NBT specific activity. Effects of isolated osteoclasts on calcified matrices
Quantitation of the effects of isolated osteoclasts on calcified matrices was performed using minor modifications of the disaggregated osteoclast resorption assay, as described by Boyde et al. (16). Avian osteoclasts, isolated as described above, were dispersed on prewetted (aMEM plus 5% fetal calf serum) slices of sperm whale dentine for 48 h before washing and addition of factor. Slices of sperm whale dentine (0.25 x 4 x 6 mm) were prepared using a Buehler low speed diamond saw (Buehler, Lake Bluff, IL), followed by sonication (15 min) in several changes of distilled water. Slices were sonicated in ethanol and dipped in acetone to air dry before storing. Before all experiments, slices were transferred through ethanol to culture medium (aMEM plus 5% FCS) for 1-2 h. Cultures were performed in 48-well microtiter plates in humidified air (10% CO2) at 37 C in aMEM medium (1 slice/well in 400 /x\ medium). A single experiment used 16-20 slices, with a minimum of 3-4 slices/ treatment. Experiments were repeated at least 3-4 times. After 72-h incubation with factor the experiments were stopped by placing the slices in 2% glutaraldehyde-0.1 M sodium cacodylate buffer, pH 7.3. Osteoclasts were examined by toluidine blue (0.1%, wt/vol) staining and light microscopy. For examination of resorption lacunae, the osteoclasts were removed by sonication in distilled H2O, followed by ethanol for 10 min before restaining with toluidine blue. Lacunae were counted using light microscopy, and the plan area of matrix resorbed was quantitated using a computer-assisted morphometric program on a Quantex image and process analysis system (Quantex Instruments, Sunnybrook, CA). Data were analyzed by pattern analysis of variance, described below. Fetal long bone assay
Fetal rat long bone assays were performed as previously described (17). Pregnant rats were injected with 200 ^tCi 45Ca on the 18th day of gestation. The following day the mothers were killed, and fetuses were removed. The mineralized shafts of the radii and ulnae were dissected free of cartilaginous tissue
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 November 2015. at 12:00 For personal use only. No other uses without permission. . All rights reserved.
EFFECTS OF OIF AND TGF-0 ON OSTEOCLASTS
3071
and incubated in BGJb medium (Sigma) for 24 h at 37 C in a humidified atmosphere of 5% CO2 and 95% air to allow for the exchange of loosely bound 45Ca with stable calcium in the medium. Bones were cultured for a further 48-120 h in BGJb medium supplemented with 1 mg/ml BSA with penicillin and streptomycin (RIA grade, Sigma) in the presence of test or control substances. Bone-resorbing activity was expressed as the percentage of the total 45Ca released into the medium. Statistical analysis Data were analyzed by pattern analysis of variance with the Student-Newman-Keuls test, using a statistical package for the IBM PC (SAS Institute, Inc., Cary, NC). Due to the nonnormality of the data in studies on the formation of resorption lacunae avian osteoclasts and the inequality of the variances of the groups, a log transformation was performed on all data before analysis. In all cases an analysis of variance was performed to determine the level of significance for the groups being tested. If a difference between groups was indicated by the analysis, a Student-Newman-Keuls test was performed to determine where the differences occurred.
Control
10
25
SO
100
OIF (ng/ml)
120
tr
z
100 80
Q_
60
Results The capacity of OIF to affect pit formation by disaggregated osteoclasts was assessed using avian osteoclasts cultured on sperm whale dentine. OIF at 50 and 100 ng/ ml consistently reduced the plan area of matrix resorbed by avian osteoclasts compared with that of the controls. Figure 1 shows results from all of the experiments we performed. Partition analysis on data from all experiments (8) showed significant decreases in area resorbed compared with that of the controls. OIF (50-100 ng/ml) caused a similar decrease in the number of excavations per slice. When the total plan area resorbed per slice was plotted against the number of excavations per slice (Fig. 2), a positive correlation was shown, suggesting that the decrease in total area resorbed was due to a decrease in pit numbers rather than a decrease in lacunae area per individual pit. We also tested the effects of OIF on osteoclast activity, as assessed by two other markers, the activity of the enzyme TRAP and staining with the dye NBT, which is reduced to a blue-insoluble formazan in the presence of oxygen-derived free radicals. NBT staining of osteoclasts occurs in response to factors that stimulate osteoclast activity, such as PTH, 1,25-dihydroxyvitamin D, and interleukin-1, and is decreased by factors that inhibit osteoclasts, such as calcitonin (18). OIF at a concentration of 100 ng/ml inhibited TRAP activity in both osteoclast lysates as well as osteoclast-conditioned medium (Table 1). This effect was seen in both unstimulated cells and cells activated by incubation with vitamin A. Vitamin A is the only well characterized factor known to activate isolated osteoclasts directly (19). OIF at concen-
