Calcif Tissue Int (1992) 51:305-311
Calcified Tissue International 9 1992 Springer-Verlag New York Inc.
Mechanism of Action of I Cell Mineralization
-Glycerophosphate on Bone
Chun-Hsi Chung, 1 Ellis E. Golub, 2 Elizabeth Forbes, 2 Toshikazu Tokuoka, 2 and Irving M. Shapiro 2 Department of 1Orthodontics and 2Biochemistry, School of Dental Medicine, University of Pennsylvania, 4001 Spruce Street, Philadelphia, Pennsylvania 19104-6003, USA Received August 13, 1991, and in revised form March 17, 1992
Summary. Experiments were performed to determine whether 13-glycerophosphate (13-GP) promoted mineralization in vitro by modulating bone cell metabolic activity and/ or serving as a local source of inorganic phosphate ions (Pi). Using MC3T3-E1, ROS 17/2.8, and chick osteoblast-like cells in the presence of 13-GP or Pi, we examined mineral formation, lactate generation, alkaline phosphatase (AP) activity, and protein and phospholipid synthesis. Neither 13-GP nor Pi modulated any of the major biosynthetic activities of the bone cells. Thus, we found no change in the levels of phospholipids, and the total protein concentration remained constant. Measurement of lactate synthesis showed that 13-GP did not effect the rate of anaerobic glycolysis. Evaluation of medium Pi levels clearly indicated that 13-GP was hydrolyzed by bone cells; within 24 hours, almost 80% of 10 mM 13-GPwas hydrolyzed. It is likely that this local increase in medium Pi concentration promoted rapid mineral deposition. Chemical, energy dispersive X-ray, and Fourier transform infrared analysis of the mineral formed in the presence of 13-GP showed that it was nonapatitic; moreover, mineral particles were also seen in the culture medium itself. Experiments performed with a cell-free system indicated that mineral particles formed spontaneously in the presence of AP and 13-GP and were deposited into a collagen matrix. We conclude that medium supplementation with 13-GP or Pi should not exceed 2 mM. If this value is exceeded, then there will be nonphysiological mineral deposition in the bone cell culture. Key words: Mineralization - Phosphate - 13-glycerophosphate - Bone cells - Alkaline phosphatase.
In his classic investigation of the ossification mechanism, Robison [1] reported that hexose phosphate esters induced mineral formation. In a later study, Robison and Rosenheim [2] demonstrated that ester phosphates such as a- and 13glycerophosphate and fructose diphosphate increased the rate of mineralization of calcifying cartilage. Robison proposed that organic phosphates served as a source of inorganic phosphate (Pi) for the mineralizing tissue. More recently, Tenenbaum and Heersche [3, 4] reported that addition of 13-glycerophosphate (13-GP) to chick periosteal cultures induced osteogenesis and promoted apatite deposition around osteoblasts. In a separate study, Ecarot-Charrier et al. [5] confirmed that 13-GP caused mineralization of osOffprint requests to: I. M. Shapiro
teoblast cultures. Based on these studies, 13-GP has been used routinely by many workers to study different aspects of bone cell function [6-13]. This agent has also been used to induce mineralization of transformed bone cells, cultured chondrocytes, and other cell types [14-17]. There is some disagreement on the mechanism by which 13-GP modulates osteogenesis and/or mineral deposition. For example, Tenenbaum [18] reported that the addition of 13-GP to cultured periostea increased the rate of bone formation, inhibited alkaline phosphatase (AP) activity, and decreased thymidine incorporation into bone cells. Likewise, Kodama et al. [14], using MC3T3-E1 cells, reported that 13-GP decreased the activity of AP, lowered the rate of protein synthesis, and induced matrix vesicle formation. In contrast, Gerstenfeld et al. [9] could not detect any effect of 13-GP on cell proliferation or transcription and although collagen synthesis remained unaffected by 13-GP, the protein content of the treated culture actually decreased. In addition, these workers reported that for the first 18 days, the AP values were elevated in the treated cultures, after which there was a rapid decline in enzyme activity. The mechanism by which 13-GPinduces mineralization is closely linked to the high AP activity of bone cell cultures. This enzyme is found in all mineralizing tissues and can release Pi from ester phosphates at neutral pH [19]. APinduced hydrolysis of organic phosphates can elevate the local Pi concentration and thereby provide the chemical potential for mineral deposition. In support of this notion was the observation that treatment of cells with levamisole, an inhibitor of AP activity, blocked mineral deposition [18]. In addition to the direct effect of Pi on mineral deposition, Pi may modulate osteoblast function by increasing bone remodeling and collagen synthesis [20]. Kodama et al. [14] suggested that Pi may increase matrix vesicle biogenesis. Thus, the possibility exists that organic phosphates may stimulate cellular metabolic processes that were required for the mineralization of bone. Indeed, our own studies of chondrocyte function have shown that a change in energy metabolism, in terms of levels of phosphorylated metabolites of ATP, signaled the initiation of cartilage mineralization [21, 22]. The purpose of this investigation was to determine the mechanism by which 13-GP initiated mineralization by cultured bone cells. We asked the following questions: Does 13-GP induce mineralization of bone cell cultures by stimulating cellular metabolic activity, and alternatively, does the release of Pi from 13-GP serve as a metabolic stimulus for apatite formation? Results of the investigation clearly showed that both 13-GP and Pi had a minimal effect on cellular biosynthetic activity. We noted that 13-GPserved as a Pi donor for calcium phosphate mineral formation and that this
306 e v e n t can lead to the formation of nonbiological mineral particles in the culture system. These particles can be incorporated into the matrix and thereby obscure the normal rate of mineral accretion by bone cells.
C-H Chung et al.: [3-Glycerophosphate and Bone Cell Mineralization T
10 :~ 8
1
Mr
[~
0sLeoblasts
~
ROS 1 7 / 2 8
~
Medium
\\
\\
\\
o
Materials and Methods
9o
6
Cell Culture Experiments were performed utilizing ROS 17/2.8 cells, MC3T3-E1 cells, rat embryonic skin fibroblasts (FR, American Type Culture Collection), and chick osteoblasts. Osteoblasts were isolated from day 19 calvaria using a previously published technique [5]. Cells were maintained in growth medium consisting of Dulbecco's Modified Eagle's Medium (DMEM, Gibco, Grand Island, NY) supplemented with 10% fetal calf serum (Hyclone), L-glutamine (2 mM), and antibiotics. The Pi concentration in the medium was 0.9 mM and the calcium level was 1.8 mM. Ascorbate (50 txg/ml) was added to bone cell cultures. Cells were plated at a density of 70 cells/mm2 for ROS, MC3T3-E1, and chick osteoblasts or 30 cells/mm2 for fibroblasts. Cultures were maintained at 37~ in 5% CO2-95% air and the medium was changed every 2 days. For scanning electron microscopy (SEM) studies, ROS cells were grown on glass coverslips coated with 50 p~g/mlpoly-L-lysine (Sigma, St Louis, MO). To study spontaneous mineral deposition, cell-free collagen discs (Colla-Tec, NJ) were adhered to the bottom of culture dishes using poly-Llysine. The discs were treated with 0.07 U AP (Sigma, chick intestine) suspended in 1 ml growth medium. The growth medium was supplemented with 1-10 mM [3-GP (13-glycerophosphate disodium salt, Sigma), 1-10 mM Na2HPO4 (Pi), or 1-3 mM glycerol.
\\
\\ \\
\\
\ \ \\ \\ \\ \\
\\ \\
\\ \\
0
24
48
(Hours) Fig. 1. [3-GP hydrolysis by MC3T3-E1, chick osteoblast-like cells, and ROS cells. Postconfluent cultures of ROS cells (day 8), MC3T3E1 cells, and osteoblast-like cells (day 16) were maintained in DMEM supplemented with 10 mM [3-GP. The Pi concentration was determined at 0, 24, and 48 hours after replenishing the medium. The Pi content of the 13-GP-supplemented medium, in the absence of ceils, was also determined. Values shown are means -+- SD of 3--4 determinations.
0.1% LaC13. Calcium was visualized in cell cultures by staining with 2% alizarin red S (pH 4.2, 5 minutes).
Biochemical Assays Protein Biosynthesis. Cultures were labeled with 35[S] methionine (25 ixCi/ml, 1320 Ci/mmol) for 20 minutes or 34[p] (50 ~Ci/ml, carrierfree orthophosphate) for 24 hours using a method described earlier [23]. Cells were harvested, washed with HBSS, treated with 10% ice-cold TCA and 70% ethyl alcohol, and solubilized in 2% sodium dodecyl sulfate (SDS). Aliquots were counted in a liquid scintillation spectrometer. For SDS polyacrylamide gel electrophoresis (SDSPAGE), cells were washed with phosphate-buffered saline (PBS) and collected by centrifugation. Samples were solubilized in running buffer and the proteins were separated by SDS-PAGE. The gels were stained with Coomassie blue. Radioactive proteins were visualized by autoradiography. Total protein was measured by the method of Bradford [24].
Phospholipid Determination. Cells were labeled with 32[p] orthophosphate (50 IxCi/ml). After 24 hours the cells were harvested, washed with PBS, and collected by centrifugation. The labeled cells were homogenized in chloroform/methanol (2:1; v/v); aliquots of the lipid extracts were removed, and the 32[p] content was determined by counting. Phospholipids were separated by two-dimensional, high efficiency, thin-layer chromatography on silica gel H plated impregnated with 1% potassium oxalate/2 mM EDTA [25]. The solvent systems consisted of chloroform/methanol/4 M ammonia (9:7: 2; v/v) and butanol/acetic acid/water (6:1:1; v/v). The lipids were identified by comparison with phospholipid standards and evaluated by autoradiography and scintillation counting.
Alkaline Phosphatase (AP). AP activity was measured at 410 nm
Scanning Electron Microscopy (SEM) and Energy Dispersive X-Ray Microanalysis (EDX) Cell cultures were dehydrated, fixed with ethanol, critical point dried, and coated with carbon. Postconfluent culture medium was filtered through a 0.4 ~m filter (Nuclepore, Pleasanton, CA) and coated with carbon or gold. Samples were viewed in a JEOL JSMT330A scanning electron microscope. Elemental microanalysis was performed using a KEVEX EDX system, and background counts were subtracted from each defined peak. Calcium and phosphorus standards for these studies were powdered chicken bone and synthetic hydroxyapatite.
Fourier Transform Infrared (FTIR) Analysis FTIR analysis was performed on freeze-dried samples using a Nicolet 5DXC spectrometer. Samples were analyzed as 10% mixtures in KBr in the diffuse reflectance mode. The range studied was 4004000 cm 1. FTIR spectra were produced from 100 scans. Spectra of mineralized matrices (10 mM 13-GP treated cultures) were compared with spectra of nonmineralized controls using a computerized interactive subtraction program. Synthetic hydroxyapatite and octacalcium phosphate were used as standards.
Results
using Sigma kit 104; p-nitrophenyl phosphate was the substrate [26].
Inorganic Phosphate (Pi). The Pi concentration of the medium was measured by the method of Chen et al. [27].
Lactate. The lactate concentration of the medium was determined using Sigma kit 826-uv.
Calcium. For total calcium determination, samples were treated with 1.0 N HC1. The calcium concentration of the extract was assessed by atomic absorption spectrophotometry in the presence of
The objective of the first series of experiments was to determine the extent of [3-GP phosphorolysis by b o n e cells. Cells were treated with 10 m M 13-GP, and the Pi c o n c e n t r a t i o n of the medium was then measured. Figure 1 shows that after 24 hours, about 80% of 10 m M 13-GP was h y d r o l y z e d by R O S cells and chick osteoblasts and o v e r 60% by MC3T3-E1 cells. By 48 hours, the Pi c o n c e n t r a t i o n of the m e d i u m was o v e r 9 mM. To ascertain w h e t h e r [3-GP hydrolysis was cell mediated, we also m e a s u r e d the Pi concentration of the me-
C-H Chung et al.: 13-Glycerophosphate and Bone Cell Mineralization I ~p
15
E223
307
MC3T3-E1 Osteoblasts
Ros iv/zs e2
~
o
~k~.0 9 ~ ~ E ;m ~ "~ o.5
\\\
~5 O0
Con[rol
10 mM ~-C,P
10 mM Pi
Fig. 2. AP activity of bone cells treated with 13-GPor Pi. MC3T3-E1 cells, chick osteoblast-like cells, and ROS cells were treated with 13-GP(10 mM). In addition, ROS cells were treated with Pi (10 raM). After 8 days (ROS cells) and 16 days (MC3T3-E1 cells, chick osteoblast-like cells), cultures were extracted with 1% triton and AP activity was determined. Values shown are means -+ SD of 3-4 determinations. Control cultures represent AP values of cells in the absence of [3-GP or Pi.
Fig. 3. Calcium accumulation in ROS cell cultures treated with 13-GP, Pi, or glycerol. Preconfluent culture of ROS cells were treated with [3-GP (3 mM), Pi (3 mM), Pi + glycerol (3 mM), or glycerol (3 mM). After 7 days, cultures were washed with normal saline and then stained with alizarin red. (a) Control, (b) glycerol, (e) glycerol + Pi, (d) 13-GP, (e) Pi. Control (a) and glycerol-treated (b) cultures did not stain with the alizarin.
dium in the absence of cells. We found that there was limited breakdown of 13-GP in the medium. Thus, the Pi content was about 1.4 mM after 24 hours and about 1.9 mM after 48 hours (Fig. 1). As [3-GP phosphorolysis is probably mediated by AP, we also determined the activity of this enzyme in the bone cell cultures; as might be expected, it was high. However, treatment of ROS, MC3T3-E1 cells, and osteoblasts with 13-GP resulted in a decrease in AP activity, and a similar decrease was noted when ROS cells were treated with Pi (Fig. 2). This latter finding confirmed earlier studies of AP inhibition by Pi [28]. The goal of the second series of investigations was to ascertain whether 13-GP modulates bone cell metabolic activity. As 13-GP could undergo phosphorolysis to form glycerol and Pi, experiments were conducted to learn if these compounds could modify bone cell metabolism and mineralization. F o r these studies, ROS cells were utilized to assess the effect of glycerol on bone cell metabolism. Cells were treated with glycerol, glycerol and Pi, or [3-GP; mineralization was assessed histochemically. We found that there was no calcium accumulation in glycerol-treated cultures. On the other hand, cultures supplemented with glycerol and Pi or t3-GP resulted in the presence of alizarin-positive material (Fig. 3). As glycerol had no effect on calcium accumulation, we examined the effects of [3-GP and Pi on ROS cell metabolism. We first measured lactate synthesis to learn whether [3-GP serves as a metabolic fuel for priming energy requiring reactions associated with mineral deposition. We found that the level of lactate synthesis was high; after 2 days, the culture medium contained 39 --- 2 mM (n = 4) lactate. However, neither 13-GP nor Pi increased the rate of lactate synthesis by the cells. We also used measurements of the rate and type of phosphatides synthesized by bone cells in culture to serve as an indicator of cellular metabolism. In terms of mineral formation, synthesis of selected phospholipids are closely associated with the initial development of mineralization [29-31]. We examined phospholipid synthesis by 13-GP supplemented cultures and noted that 32[p] was avidly incorporated into the phospholipids of the cultured bone cells. However, no significant differences were noted in the rate of 32[p] incorporation into phospholipids. Thus, in the presence of 10 mM
13-GP, 631 -+ 81 cpm/104 (n = 3) cells was incorporated into phosphatides; the uptake into control cells was 658 -+ 41 cpm/104 cells. In terms of type of phospholipids synthesized by these cells, those identified included lecithin, phosphatidylethanolamine, phosphatidylinositol, sphingomyelin, cardiolipin, and phosphatidylserine. The chromatographic distribution of phosphatides indicated that 13-GP had no effect on the types or quantities of phospholipids synthesized by the ROS cell cultures (Fig. 4). The effects of [3-GP and Pi on proteins synthesized by ROS cells were examined. We found minimal stimulation of protein synthesis by these agents. The general pattern of labeling (and the extent of labeling) was similar in treated ([3-GP and Pi) and control cultures (not shown). 35[S] and 32[p] incorporation into proteins in control cells was 350 --- 56 and 1277 -+ 156 cpm/104 cells (n = 3), respectively. No increase in labeling was observed when cultures were treated with [3-GP or Pi. As far as cell number was concerned, there was no significant change after 7 days in culture. The aim of the final series of experiments was to examine the mineral formed by the cultured ceils. Figure 5 shows that by 7 days, treatment of cells with both [3-GP and Pi caused a dose-dependent increase in calcium accumulation by the cultured cells. The addition of Pi to the ROS culture resulted in high levels of calcium accumulation. At both of the concentrations used (3 and 10 mM), the matrix calcium concentration in the presence of Pi was greater than [3-GP. In addition, although the concentration of calcium increased with time, the ratio of calcium:phosphorus at 7 days was low (10 mM [3-GP 0.53 -+ 0.2, n = 4; 10 mM Pi 0.85 - 0.3, n = 4). A similar type of study was performed with MC3T3-E1 cells. After 18 days, cultures treated with 10 mM [3-GP contained 5-7 times as much calcium as control cultures. However, the calcium:phosphorus ratio of mineral present in the cell layer was very low (0.67 - 0.1; n = 4). Selected areas of day 14 cultures were evaluated using SEM and EDX. Figure 6A-C shows the morphological appearance of ROS cells treated with 13-GP or Pi. The presence of 13-GP caused the appearance of mineral deposits in the ROS cell layer; the effect of Pi was somewhat different in that some areas of the culture were covered with mineral even in the absence of cells. EDX analysis of selected areas
308
C-H Chung et al.: [~-Glycerophosphate and Bone Cell Mineralization
Fig. 4. Autoradiogram of high efficiency, thin layer chromatogram of ROS cell culture phospholipids. Preconfluent ROS cell cultures (day 7) were pretreated with 32[p] orthophosphate for 24 hours. Phosphatides were extracted from the cell layer and separated by two-dimensional chromatography (directions of the 1st and 2nd solvent systems are shown in the figure). The patterns of labeling of phospholipids of the treated and control cultures were similar. (A) Control, (B) culture treated with 10 mM [3-GP. o = origin, a = diphosphatidylinositol, b = sphingomyelin, c = lecithin, d = phosphatidylinositol, e = phosphatidylethanolamine, f = phosphatidylserine, g = cardiolipin.
16
12
Fo co--.. 8
r
t-,el,ill
Con
3 mM
i0 rnM
3 mM
10 mM
fi-GP Pi Pi f{ GP Fig. 5. Accumulation of calcium by ROS cells treated with 13-GPor Pi. Preconfluent cultures of ROS cells were treated with 13-GP(3 or 10 mM) or Pi (3 or 10 mM). After 7 days, cells were harvested and calcium was released by treatment with 1.0 N HC1. Calcium was determined by atomic absorption spectrophotometry. Values shown are means -+ SEM of 3--4 determinations. Con = control.
showed that the calcium:phosphorus molar ratio was about 1.0 ([3-GP range 0.9-1.1; Pi range 0.8-1.0). Similar results were obtained using 3 mM [3-GP and Pi; here the amount of mineral deposited was less than that of the 10 mM cultures. However, mineral was only seen in a limited number of areas in the culture, and EDX analysis indicated that the calcium:phosphorus ratio was low (below 1.1). Mineral formed by day 14 bone cell cultures treated with 10 mM 13-GP was evaluated by FTIR. Comparison of the difference spectrum of the mineralized matrix with the spectrum of an apatite standard indicated that it was nonapatitic (Fig. 7A,B). However, the mineral appeared to share some similarities with octacalcium phosphate. Thus, absorption bands were seen at 854, 878, 910, 958, 1020, 1081, and 1125 c m - 1 (Fig. 7B). In view of the high level of Pi in the medium of [3-GPtreated cultures (see Fig. 1), the possibility existed that an elevation in the calcium Pi ion product could cause spontaneous (nonbiological) mineral formation. To learn whether
culture conditions promoted mineral formation, medium from postconfluent cultures was collected after 24 and 48 hours and filtered through a 0.4 tam filter. We found that mineral particles were present in medium obtained from 3 and 10 mM [3-GP and Pi-treated ROS, MC3T3-E1, and osteoblast cell cultures at both time periods. No mineral was present in control media. EDX analysis of the particles indicated that they contained mainly calcium and phosphorus (Fig. 6D). Little difference was seen in the composition of particles obtained from medium of [3-GP or Pi-treated osteoblasts, MC3T3-E1, and ROS cells. To further investigate the possibility that mineral could be precipitated in [3-GP-treated bone cell cultures, experiments were performed with fibroblasts that have a low endogenous AP activity and do not mineralize. In this study, fibroblasts were treated with [3-GP or Pi and calcium accumulation in the culture was examined by alizarin staining. After 7 days, there was no evidence of calcium accumulation by control cells and cells treated with [3-GP; in contrast, when the medium Pi supplementation was greater than 2 mM, alizarin staining indicated that calcium was precipitated into the cell layer (Fig. 8). In the presence of 10 mM Pi, a considerable amount of alizarin-positive material was seen in the fibroblast culture. At the latter Pi concentration, mineral deposition was associated with cell death. Finally, to ascertain whether AP alone could cause spontaneous mineral deposition, collagen discs were adhered to the bottom of culture dishes that contained medium supplemented with purified AP and 10 mM [3-GP. In this cell-free system, [3-GP was hydrolyzed by AP to release Pi. After 24 hours, the Pi concentration of the medium was about 9 mM; after 48 hours, analysis of the discs indicated that 70.3 +- 5.9% (n = 4) of the total medium calcium concentration was deposited.
Discussion
The goal of this investigation was to examine the mechanism by which [3-GP promoted mineral deposition by bone cells in culture. Analysis of a number of metabolic parameters indicated that neither [3-GP nor Pi modulated any of the major biosynthetic activities of ROS cells. Thus, we found no change in the levels of acidic phospholipids, such as phos-
C-H Chung et al.: 13-Glycerophosphate and Bone Cell Mineralization
309
Fig. 6. SEM appearance of selected areas of 14-day ROS cultures and medium mineral particles. (A) Control culture showing the presence of ROS cells; (B) culture treated with 10 mM 13-GP;(C) culture treated with 10 mM Pi; (D) mineral particles in medium of cultures
treated with 10 mM [3-GP and filtered through a 0.4 p.m Nuclepore membrane. Insets show EDX spectrum and the calcium and phosphorus peaks. Bar = 10 IxM (A, B, C); bar = 1 I~M (D).
phatidylserine, a phosphatide that is required for bone and cartilage mineralization; and the cell number and total protein concentration remained constant. Measurement of lactate synthesis showed that f3-GP did not effect the rate of anaerobic glycolysis. In addition, the design of the investigation permitted us to evaluate the concept that hydrolysis of 13-GP provides components (glycerol and Pi) that modify cellular metabolism and thereby promotes mineralization of the extracellular matrix. We found that glycerol had no effect on mineral formation; moreover, the addition of Pi to the medium resulted in only minor changes in metabolic and biosynthetic activities of bone cells. Based on these findings, we reject the notion that 13-GP, a compound that can serve as an intermediate in carbohydrate and lipid metabolism, regulates bone cell activity and thereby fosters the formation of a calcified matrix. In contrast to these metabolic studies, we noted that when the medium was supplemented with Pi or [3-GP, there was rapid enhancement in the rate of mineral deposition. This event is most likely due to a rise in the Pi content of the medium. Indeed, the results of the study clearly showed that 13-GP was hydrolyzed by bone cells to produce high concentration of Pi. Thus, within 24 hours, almost 80% of 10 mM f3-GP was hydrolyzed by the bone cells. The rate of mineral formation with [3-GP was lower than that of Pi. These results were expected because (1) 13-GP has to undergo hydrolysis
to liberate Pi and as this is a time-dependent process, the Pi level of [3-GP supplemented cultures is always less than that of cultures supplemented directly with Pi; (2) the AP activity of the nonconfluent ceils is low initially and 5-7 days are required for A P activity to become maximum. Experiments with fibroblasts that have a low endogenous AP activity, and do not normally mineralize, confirmed that this enzyme was required for [3-GP-mediated mineralization. As mineral was found in fibroblast cultures treated with Pi, it must be concluded that the mineral deposited into these cultures was a precipitation artifact. We used a number of cell types to investigate the importance of t3-GP as a mediator of mineral formation by cells in culture. The cells we chose were those that have been used by previous investigators to characterize aspects of the mineralization process. Of these, the ROS 17/2.8 cells have been utilized for many studies [32-34]. This cell exhibits high AP activity and rapidly forms mineral in the presence of [3-GP [15, 35]. However, as the amount of type I collagen secreted by this cell is low, it could be argued that it is not an optimum cell type for mineralization studies. F o r this reason, we also utilized another cell line (MC3T3-E1), as well as calvarial osteoblast-like cells. Though some differences were observed between these cell types, in general, the data were consistent with the notion that the mineralization was dependent on AP-mediated release of Pi from [3-GP.
C-H Chung et al.: ~3-Glycerophosphate and Bone Cell Mineralization
310
E3
A
4000
2800
1600
Wavenumber (cm-1)
Fig. 7. FTIR difference spectrum of mineralized matrix present in day 14 ROS cell culture treated with 10 mM 13-GP. FTIR analysis was performed on freeze-dried samples using a Nicolet 5DXC spectrometer. Samples were analyzed as 10% mixtures in KBr in the
Fig. 8. Calcium accumulation in fibroblast cell cultures treated with Pi. Preconfluent cultures of rat skin fibroblasts were treated with Pi (total Pi concentration: (a) 0.9 mM, control; (b) 1.9 raM, (r 2.9 mM, (d) 3.9 mM, (e) 5.9 mM, (f) 10.9 mM). After 7 days, cultures were washed and then stained with alizarin red. Ceils stained positive with alizarin when the Pi supplementation was above 2 mM (d, e, f).
We investigated the mineral formed in bone cell cultures and spontaneous mineral deposition in the presence of 10 mM 13-GP. With respect to cell-mediated mineralization, chemical, elemental, and the F T I R analyses of the ROS cultures and chemical analysis of the MC3T3-EI cells showed that the mineral was abnormal. Thus, the calcium:phosphorus ratios were depressed and the mineral was nonapatitic. These results confirmed an earlier study that showed that 10 mM 13-GP caused the formation of nonbiological (ectopic) mineralization in cultured calvaria [36]. Moreover, we found that calcium phosphate mineral particles were present in the medium of postconfluent ROS cells, MC3T3-E1, and calva-
400
1200
1000
800
Wavenurnber (cm-1) diffuse reflectance mode. (A) Shows the spectra from 400 to 4000
cm-1. Hydroxyapatite and octacalcium phosphate were used as standards. (B) Shows details of the region 800-1200 cm-1.
rial bone cell cultures treated with 10 mM [3-GP. Experiments performed with the cell-free system in which the culture medium contained AP and 10 mM ~-GP indicated that mineral particles formed spontaneously and were deposited into the collagen matrix. The amount of mineral that accumulated in the collagen discs was very high. Thus, after 48 hours, over 70% of the total medium calcium was present as mineral in the discs. In our experiments, the concentrations of [3-GP (10 raM) was similar to that used by other investigators [3-17]. It should be noted that according to Marshall in Nordin's book and others, in serum or plasma, the level of inorganic phosphate is 1.0-1.5 mM, whereas the organic phosphate level (which is mostly phospholipid) is about 3 mM [37, 38]. However, the concentration of hydrolyzable ester phosphate is 1.1 mM or less [39, 40]. Thus, a medium concentration of 10 mM 13-GP is high when compared with physiological levels, and as we have shown here, its hydrolysis results in medium Pi levels that are supraphysiological. A consequence of the elevated Pi concentration of the medium, in the presence of 1.8 mM calcium, is the rapid formation of mineral. To minimize spontaneous mineral deposition, supplementation of the culture medium with Pi or [~-GP should be limited. Resuits of experiments performed with fibroblasts indicated that medium supplementation with Pi or 13-GP should not exceed 2 mM. If this value is exceeded, nonphysiological mineral formation will occur and this would serve to obscure the relatively low rate of mineral accretion by osteoblasts in culture.
Acknowledgments. This work was supported by NIH Grants AR34411, DE-08239, and DE-06533. C-H.C. was supported by a NIH Postdoctoral Fellowship (AR-07481). The help of Dr. S. Radin, Mr. A. El-Ghannam, Mr. J. Mayro, Mr. G. Harrison, and Ms. K. DeBott is gratefully acknowledged.
C-H Chung et al.: 13-Glycerophosphate and Bone Cell Mineralization
References 1. Robison R (1923) The possible significance of hexosephosphoric esters in ossification. Biochem J 17:286--293 2. Robison R, Rosenheim AH (1934) Calcification of hypertrophic cartilage in vitro. Biochem J 28:684-4598 3. Tenenbaum HC (1981) Role of organic phosphate in mineralization of bone in vitro. J Dent Res 60(C):1586-1589 4. Tenenbaum HC, Heersche JNM 0982) Differentiation of osteoblasts and formation of mineralized bone in vitro. Calcif Tissue Int 34:76-79 5. Ecarot-Charrier B, Glorieux FH, Van Der Rest M, Pereira G (1983) Osteoblasts isolated from mouse calvaria initiate matrix mineralization in culture. J Cell Biol 96:639-643 6. Whitson SW, Harrison W, Dunlap MK, Bowers DE, Fisher LW, Gehron Robey P, Termine JD (1984) Fetal bovine bone " cells synthesize bone-specific matrix proteins. J Cell Biol 99: 607-614 7. Gehron Robey P, Termine JD (1985) Human bone cells in vitro. Calcif Tissue Int 37:453-460 8. Bellows CG, Aubin JE, Heersche JNM, Antosz ME (1986) Mineralized bone nodules formed in vitro from enzymatically released rat calvaria cell population. Calcif Tissue Int 38:143-154 9. Gerstenfeld LC, Chipman SD, Glowacki J, Lian JB (1987) Expression of differentiated function by mineralizing cultures of chicken osteoblasts. Dev Biol 122:49-60 10. Bhargava U, Bar-Lev M, Bellows CG, Aubin JE (1988) Ultrastructural analysis of bone nodules formed in vitro by isolated fetal rat calvaria cells. Bone 9:155-163 11. Berry L, Shuttleworth CA (1989) Expression of the chondrogenie phenotype by mineralizing cultures of embryonic chick calvarial bone cells. Bone Miner 7:31-45 12. Masquelier D, Herbert B, Hauser N, Mermillod P, Schonne E, Remacle C (1990) Morphological characterization of o steoblastlike cell cultures isolated from newborn rat calvaria. Calcif Tissue Int 47:92-104 13. Bellows CG, Aubin JE, Heersehe JNM (1991) Initiation and progression of mineralization of bone nodules formed in vitro: the role of alkaline phosphatase and organic phosphate. Bone Miner 14:27-40 14. Kodama H-A, Amagai Y, Sudo H, Ohno T, Iijima K-I (1986) Culture conditions affecting differentiation and calcification in the MC3T3-E1 osteogenic cell line. In: Ali SY (ed) Cellmediated calcification and matrix vesicles. Elsevier Science Publishers B.V., pp 297-302 15. Nishimoto SK, Stryker WF, Nimni ME (1987) Calcification of osteoblastlike rat osteosarcoma cells in agarose suspension cultures. Caleif Tissue Int 41:274-280 16. Thomas JT, Boot-Handford RP, Grant ME (1990) Modulation of type X collagen gene expression by calcium [3-glycerophosphate and levamisole: implications for a possible role for type X collagen in endochondral bone formation. J Cell Sci 95:639-648 17. Luria EA, Owen ME, Friedenstein AJ, Morris JF, Kuznetsow SA (1987) Bone formation in organ cultures of bone marrow. Calcif Tissue Int 248:449-454 18. Tenenbaum HC (1987) Levamisole and inorganic pyrophosphate inhibit 13-glycerophosphate-induced mineralization of bone formed in vitro. Bone Miner 3:13-26 19. Fortuna R, Anderson HC, Carty RP, Sajdera SW (1980) Enzymatic characterization of the matrix vesicle alkaline phosphatase isolated from bovine fetal epiphyseal cartilage. Calcif Tissue Int 30:217-225 20. Raisz LG, Dietrich JW, Canalis EM (1976) Factors influencing bone formation in organ cultures. Isr J Med Sci 12:108--114
311 21. Kakuta S, Golub EE, Shapiro IM (1985) Morphochemical analysis of phosphorus pools in calcifying cartilage. Calcif Tissue lnt 37:293-299 22. Kakuta S, Golub EE, Haselgrove JC, Chance B, Frasca P, Shapiro IM (1986) Redox studies of the epiphyseal growth cartilage: pyridine nucleotide metabolism and the development of mineralization. J Bone Miner Res 1(5):433--440 23. Pacifici M, Oshima O, Fisher LW, Young MF, Shapiro IM, Leboy PS (1990) Changes in osteonectin distribution and levels are associated with mineralization of the chicken tibial growth cartilage. Calcif Tissue Int 47:51--61 24. Bradford MM (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72:248-254 25. Sasaki T, Hasegawa-Sasaki H (1985) Breakdown of phosphatidylinositol 4,5-biphosphate in a T-cell leukaemia line stimulated by phytohaemagglutinin is not dependent on Ca2+ mobilization. Biochem J 227:971-979 26. Golub EE, Schattschneider SC, Berthold P, Burke A, Shapiro IM (1983) Induction of chondrocyte vesiculation in vitro. J Biol Chem 258:616--621 27. Chen PS, Toribara TY, Warner H (1956) Microdetermination of phosphorus. Anal Chem 28:1756-1758 28. Jung K, Pergande M (1980) Influence of inorganic phosphate on the activity determination of isoenzymes of alkaline phosphatase in various buffer systems. Clin Chim Acta 102:215-219 29. Boskey AL, Posner AS (1977) The role of synthetic and boneextracted Ca-phospholipid-PO4 complexes in hydroxyapatite formation. Calcif Tissue Res 23:251-258 30. Wuthier RE, Gore ST (1977) Partition of inorganic ions and phospholipids in isolated cell, membrane and matrix vesicle fractions: evidence for Ca-Pi-Acidic phospholipid complexes. Calcif Tissue Res 24:163-17l 31. Boyan-Salyers BD, Boskey AL (1980) Relationship between proteolipids and calcium-phospholipid-phosphate complexes in bacterionema matruchotii calcification. Calcif Tissue Int 30: 16%174 32. Rodan GA, Majeska RJ (1982) Phenotypic maturation of osteoblastic osteosarcoma cells in culture. Clin Biol Res 110(B):249259 33. Rodan GA, Rodan SB (1983) Expression of the osteoblastic phenotype. In: Peck WA (ed) Bone and mineral research, Annual 2. Elsevier Science Publishers B.V., pp 244-285 34. Rodan GA, Heath JK, Yoon K, Noda M, Rodan SB (1988) Diversity of the osteoblastic phenotype. Ciba Foundation Symposium 136:78-91 35. Yoon K, Golub E, Rodan GA (1989) Alkaline phosphatase eDNA transfected cells promote calcium and phosphate deposition. Connect Tissue Res 22:17-25 36. Gronowiez G, Woodiel FN, McCarthy M-B, Raisz LG (1989) In vitro mineralization of fetal rat parietal bones in defined serumfree medium: effect of 13-glycerolphosphate. J Bone Miner Res 3:313-324 37. Nordin BEC (1976) Calcium, phosphate and magnesium metabolism. Churchill Livingston, Edinburgh, London, New York 38. Scully RE (1980) Normal reference values. N Engl J Med 302:3% 48 39. Hoffman WS (1970) The biochemistry of clinical medicine, 4th ed. Year Book Medical Publishers 40. Geigy Scientific Tables, Lentner C (ed) (1984), vol 3, 8th ed. Ciba-Geigy
Calcified Tissue International 9 1992 Springer-Verlag New York Inc.
Mechanism of Action of I Cell Mineralization
-Glycerophosphate on Bone
Chun-Hsi Chung, 1 Ellis E. Golub, 2 Elizabeth Forbes, 2 Toshikazu Tokuoka, 2 and Irving M. Shapiro 2 Department of 1Orthodontics and 2Biochemistry, School of Dental Medicine, University of Pennsylvania, 4001 Spruce Street, Philadelphia, Pennsylvania 19104-6003, USA Received August 13, 1991, and in revised form March 17, 1992
Summary. Experiments were performed to determine whether 13-glycerophosphate (13-GP) promoted mineralization in vitro by modulating bone cell metabolic activity and/ or serving as a local source of inorganic phosphate ions (Pi). Using MC3T3-E1, ROS 17/2.8, and chick osteoblast-like cells in the presence of 13-GP or Pi, we examined mineral formation, lactate generation, alkaline phosphatase (AP) activity, and protein and phospholipid synthesis. Neither 13-GP nor Pi modulated any of the major biosynthetic activities of the bone cells. Thus, we found no change in the levels of phospholipids, and the total protein concentration remained constant. Measurement of lactate synthesis showed that 13-GP did not effect the rate of anaerobic glycolysis. Evaluation of medium Pi levels clearly indicated that 13-GP was hydrolyzed by bone cells; within 24 hours, almost 80% of 10 mM 13-GPwas hydrolyzed. It is likely that this local increase in medium Pi concentration promoted rapid mineral deposition. Chemical, energy dispersive X-ray, and Fourier transform infrared analysis of the mineral formed in the presence of 13-GP showed that it was nonapatitic; moreover, mineral particles were also seen in the culture medium itself. Experiments performed with a cell-free system indicated that mineral particles formed spontaneously in the presence of AP and 13-GP and were deposited into a collagen matrix. We conclude that medium supplementation with 13-GP or Pi should not exceed 2 mM. If this value is exceeded, then there will be nonphysiological mineral deposition in the bone cell culture. Key words: Mineralization - Phosphate - 13-glycerophosphate - Bone cells - Alkaline phosphatase.
In his classic investigation of the ossification mechanism, Robison [1] reported that hexose phosphate esters induced mineral formation. In a later study, Robison and Rosenheim [2] demonstrated that ester phosphates such as a- and 13glycerophosphate and fructose diphosphate increased the rate of mineralization of calcifying cartilage. Robison proposed that organic phosphates served as a source of inorganic phosphate (Pi) for the mineralizing tissue. More recently, Tenenbaum and Heersche [3, 4] reported that addition of 13-glycerophosphate (13-GP) to chick periosteal cultures induced osteogenesis and promoted apatite deposition around osteoblasts. In a separate study, Ecarot-Charrier et al. [5] confirmed that 13-GP caused mineralization of osOffprint requests to: I. M. Shapiro
teoblast cultures. Based on these studies, 13-GP has been used routinely by many workers to study different aspects of bone cell function [6-13]. This agent has also been used to induce mineralization of transformed bone cells, cultured chondrocytes, and other cell types [14-17]. There is some disagreement on the mechanism by which 13-GP modulates osteogenesis and/or mineral deposition. For example, Tenenbaum [18] reported that the addition of 13-GP to cultured periostea increased the rate of bone formation, inhibited alkaline phosphatase (AP) activity, and decreased thymidine incorporation into bone cells. Likewise, Kodama et al. [14], using MC3T3-E1 cells, reported that 13-GP decreased the activity of AP, lowered the rate of protein synthesis, and induced matrix vesicle formation. In contrast, Gerstenfeld et al. [9] could not detect any effect of 13-GP on cell proliferation or transcription and although collagen synthesis remained unaffected by 13-GP, the protein content of the treated culture actually decreased. In addition, these workers reported that for the first 18 days, the AP values were elevated in the treated cultures, after which there was a rapid decline in enzyme activity. The mechanism by which 13-GPinduces mineralization is closely linked to the high AP activity of bone cell cultures. This enzyme is found in all mineralizing tissues and can release Pi from ester phosphates at neutral pH [19]. APinduced hydrolysis of organic phosphates can elevate the local Pi concentration and thereby provide the chemical potential for mineral deposition. In support of this notion was the observation that treatment of cells with levamisole, an inhibitor of AP activity, blocked mineral deposition [18]. In addition to the direct effect of Pi on mineral deposition, Pi may modulate osteoblast function by increasing bone remodeling and collagen synthesis [20]. Kodama et al. [14] suggested that Pi may increase matrix vesicle biogenesis. Thus, the possibility exists that organic phosphates may stimulate cellular metabolic processes that were required for the mineralization of bone. Indeed, our own studies of chondrocyte function have shown that a change in energy metabolism, in terms of levels of phosphorylated metabolites of ATP, signaled the initiation of cartilage mineralization [21, 22]. The purpose of this investigation was to determine the mechanism by which 13-GP initiated mineralization by cultured bone cells. We asked the following questions: Does 13-GP induce mineralization of bone cell cultures by stimulating cellular metabolic activity, and alternatively, does the release of Pi from 13-GP serve as a metabolic stimulus for apatite formation? Results of the investigation clearly showed that both 13-GP and Pi had a minimal effect on cellular biosynthetic activity. We noted that 13-GPserved as a Pi donor for calcium phosphate mineral formation and that this
306 e v e n t can lead to the formation of nonbiological mineral particles in the culture system. These particles can be incorporated into the matrix and thereby obscure the normal rate of mineral accretion by bone cells.
C-H Chung et al.: [3-Glycerophosphate and Bone Cell Mineralization T
10 :~ 8
1
Mr
[~
0sLeoblasts
~
ROS 1 7 / 2 8
~
Medium
\\
\\
\\
o
Materials and Methods
9o
6
Cell Culture Experiments were performed utilizing ROS 17/2.8 cells, MC3T3-E1 cells, rat embryonic skin fibroblasts (FR, American Type Culture Collection), and chick osteoblasts. Osteoblasts were isolated from day 19 calvaria using a previously published technique [5]. Cells were maintained in growth medium consisting of Dulbecco's Modified Eagle's Medium (DMEM, Gibco, Grand Island, NY) supplemented with 10% fetal calf serum (Hyclone), L-glutamine (2 mM), and antibiotics. The Pi concentration in the medium was 0.9 mM and the calcium level was 1.8 mM. Ascorbate (50 txg/ml) was added to bone cell cultures. Cells were plated at a density of 70 cells/mm2 for ROS, MC3T3-E1, and chick osteoblasts or 30 cells/mm2 for fibroblasts. Cultures were maintained at 37~ in 5% CO2-95% air and the medium was changed every 2 days. For scanning electron microscopy (SEM) studies, ROS cells were grown on glass coverslips coated with 50 p~g/mlpoly-L-lysine (Sigma, St Louis, MO). To study spontaneous mineral deposition, cell-free collagen discs (Colla-Tec, NJ) were adhered to the bottom of culture dishes using poly-Llysine. The discs were treated with 0.07 U AP (Sigma, chick intestine) suspended in 1 ml growth medium. The growth medium was supplemented with 1-10 mM [3-GP (13-glycerophosphate disodium salt, Sigma), 1-10 mM Na2HPO4 (Pi), or 1-3 mM glycerol.
\\
\\ \\
\\
\ \ \\ \\ \\ \\
\\ \\
\\ \\
0
24
48
(Hours) Fig. 1. [3-GP hydrolysis by MC3T3-E1, chick osteoblast-like cells, and ROS cells. Postconfluent cultures of ROS cells (day 8), MC3T3E1 cells, and osteoblast-like cells (day 16) were maintained in DMEM supplemented with 10 mM [3-GP. The Pi concentration was determined at 0, 24, and 48 hours after replenishing the medium. The Pi content of the 13-GP-supplemented medium, in the absence of ceils, was also determined. Values shown are means -+- SD of 3--4 determinations.
0.1% LaC13. Calcium was visualized in cell cultures by staining with 2% alizarin red S (pH 4.2, 5 minutes).
Biochemical Assays Protein Biosynthesis. Cultures were labeled with 35[S] methionine (25 ixCi/ml, 1320 Ci/mmol) for 20 minutes or 34[p] (50 ~Ci/ml, carrierfree orthophosphate) for 24 hours using a method described earlier [23]. Cells were harvested, washed with HBSS, treated with 10% ice-cold TCA and 70% ethyl alcohol, and solubilized in 2% sodium dodecyl sulfate (SDS). Aliquots were counted in a liquid scintillation spectrometer. For SDS polyacrylamide gel electrophoresis (SDSPAGE), cells were washed with phosphate-buffered saline (PBS) and collected by centrifugation. Samples were solubilized in running buffer and the proteins were separated by SDS-PAGE. The gels were stained with Coomassie blue. Radioactive proteins were visualized by autoradiography. Total protein was measured by the method of Bradford [24].
Phospholipid Determination. Cells were labeled with 32[p] orthophosphate (50 IxCi/ml). After 24 hours the cells were harvested, washed with PBS, and collected by centrifugation. The labeled cells were homogenized in chloroform/methanol (2:1; v/v); aliquots of the lipid extracts were removed, and the 32[p] content was determined by counting. Phospholipids were separated by two-dimensional, high efficiency, thin-layer chromatography on silica gel H plated impregnated with 1% potassium oxalate/2 mM EDTA [25]. The solvent systems consisted of chloroform/methanol/4 M ammonia (9:7: 2; v/v) and butanol/acetic acid/water (6:1:1; v/v). The lipids were identified by comparison with phospholipid standards and evaluated by autoradiography and scintillation counting.
Alkaline Phosphatase (AP). AP activity was measured at 410 nm
Scanning Electron Microscopy (SEM) and Energy Dispersive X-Ray Microanalysis (EDX) Cell cultures were dehydrated, fixed with ethanol, critical point dried, and coated with carbon. Postconfluent culture medium was filtered through a 0.4 ~m filter (Nuclepore, Pleasanton, CA) and coated with carbon or gold. Samples were viewed in a JEOL JSMT330A scanning electron microscope. Elemental microanalysis was performed using a KEVEX EDX system, and background counts were subtracted from each defined peak. Calcium and phosphorus standards for these studies were powdered chicken bone and synthetic hydroxyapatite.
Fourier Transform Infrared (FTIR) Analysis FTIR analysis was performed on freeze-dried samples using a Nicolet 5DXC spectrometer. Samples were analyzed as 10% mixtures in KBr in the diffuse reflectance mode. The range studied was 4004000 cm 1. FTIR spectra were produced from 100 scans. Spectra of mineralized matrices (10 mM 13-GP treated cultures) were compared with spectra of nonmineralized controls using a computerized interactive subtraction program. Synthetic hydroxyapatite and octacalcium phosphate were used as standards.
Results
using Sigma kit 104; p-nitrophenyl phosphate was the substrate [26].
Inorganic Phosphate (Pi). The Pi concentration of the medium was measured by the method of Chen et al. [27].
Lactate. The lactate concentration of the medium was determined using Sigma kit 826-uv.
Calcium. For total calcium determination, samples were treated with 1.0 N HC1. The calcium concentration of the extract was assessed by atomic absorption spectrophotometry in the presence of
The objective of the first series of experiments was to determine the extent of [3-GP phosphorolysis by b o n e cells. Cells were treated with 10 m M 13-GP, and the Pi c o n c e n t r a t i o n of the medium was then measured. Figure 1 shows that after 24 hours, about 80% of 10 m M 13-GP was h y d r o l y z e d by R O S cells and chick osteoblasts and o v e r 60% by MC3T3-E1 cells. By 48 hours, the Pi c o n c e n t r a t i o n of the m e d i u m was o v e r 9 mM. To ascertain w h e t h e r [3-GP hydrolysis was cell mediated, we also m e a s u r e d the Pi concentration of the me-
C-H Chung et al.: 13-Glycerophosphate and Bone Cell Mineralization I ~p
15
E223
307
MC3T3-E1 Osteoblasts
Ros iv/zs e2
~
o
~k~.0 9 ~ ~ E ;m ~ "~ o.5
\\\
~5 O0
Con[rol
10 mM ~-C,P
10 mM Pi
Fig. 2. AP activity of bone cells treated with 13-GPor Pi. MC3T3-E1 cells, chick osteoblast-like cells, and ROS cells were treated with 13-GP(10 mM). In addition, ROS cells were treated with Pi (10 raM). After 8 days (ROS cells) and 16 days (MC3T3-E1 cells, chick osteoblast-like cells), cultures were extracted with 1% triton and AP activity was determined. Values shown are means -+ SD of 3-4 determinations. Control cultures represent AP values of cells in the absence of [3-GP or Pi.
Fig. 3. Calcium accumulation in ROS cell cultures treated with 13-GP, Pi, or glycerol. Preconfluent culture of ROS cells were treated with [3-GP (3 mM), Pi (3 mM), Pi + glycerol (3 mM), or glycerol (3 mM). After 7 days, cultures were washed with normal saline and then stained with alizarin red. (a) Control, (b) glycerol, (e) glycerol + Pi, (d) 13-GP, (e) Pi. Control (a) and glycerol-treated (b) cultures did not stain with the alizarin.
dium in the absence of cells. We found that there was limited breakdown of 13-GP in the medium. Thus, the Pi content was about 1.4 mM after 24 hours and about 1.9 mM after 48 hours (Fig. 1). As [3-GP phosphorolysis is probably mediated by AP, we also determined the activity of this enzyme in the bone cell cultures; as might be expected, it was high. However, treatment of ROS, MC3T3-E1 cells, and osteoblasts with 13-GP resulted in a decrease in AP activity, and a similar decrease was noted when ROS cells were treated with Pi (Fig. 2). This latter finding confirmed earlier studies of AP inhibition by Pi [28]. The goal of the second series of investigations was to ascertain whether 13-GP modulates bone cell metabolic activity. As 13-GP could undergo phosphorolysis to form glycerol and Pi, experiments were conducted to learn if these compounds could modify bone cell metabolism and mineralization. F o r these studies, ROS cells were utilized to assess the effect of glycerol on bone cell metabolism. Cells were treated with glycerol, glycerol and Pi, or [3-GP; mineralization was assessed histochemically. We found that there was no calcium accumulation in glycerol-treated cultures. On the other hand, cultures supplemented with glycerol and Pi or t3-GP resulted in the presence of alizarin-positive material (Fig. 3). As glycerol had no effect on calcium accumulation, we examined the effects of [3-GP and Pi on ROS cell metabolism. We first measured lactate synthesis to learn whether [3-GP serves as a metabolic fuel for priming energy requiring reactions associated with mineral deposition. We found that the level of lactate synthesis was high; after 2 days, the culture medium contained 39 --- 2 mM (n = 4) lactate. However, neither 13-GP nor Pi increased the rate of lactate synthesis by the cells. We also used measurements of the rate and type of phosphatides synthesized by bone cells in culture to serve as an indicator of cellular metabolism. In terms of mineral formation, synthesis of selected phospholipids are closely associated with the initial development of mineralization [29-31]. We examined phospholipid synthesis by 13-GP supplemented cultures and noted that 32[p] was avidly incorporated into the phospholipids of the cultured bone cells. However, no significant differences were noted in the rate of 32[p] incorporation into phospholipids. Thus, in the presence of 10 mM
13-GP, 631 -+ 81 cpm/104 (n = 3) cells was incorporated into phosphatides; the uptake into control cells was 658 -+ 41 cpm/104 cells. In terms of type of phospholipids synthesized by these cells, those identified included lecithin, phosphatidylethanolamine, phosphatidylinositol, sphingomyelin, cardiolipin, and phosphatidylserine. The chromatographic distribution of phosphatides indicated that 13-GP had no effect on the types or quantities of phospholipids synthesized by the ROS cell cultures (Fig. 4). The effects of [3-GP and Pi on proteins synthesized by ROS cells were examined. We found minimal stimulation of protein synthesis by these agents. The general pattern of labeling (and the extent of labeling) was similar in treated ([3-GP and Pi) and control cultures (not shown). 35[S] and 32[p] incorporation into proteins in control cells was 350 --- 56 and 1277 -+ 156 cpm/104 cells (n = 3), respectively. No increase in labeling was observed when cultures were treated with [3-GP or Pi. As far as cell number was concerned, there was no significant change after 7 days in culture. The aim of the final series of experiments was to examine the mineral formed by the cultured ceils. Figure 5 shows that by 7 days, treatment of cells with both [3-GP and Pi caused a dose-dependent increase in calcium accumulation by the cultured cells. The addition of Pi to the ROS culture resulted in high levels of calcium accumulation. At both of the concentrations used (3 and 10 mM), the matrix calcium concentration in the presence of Pi was greater than [3-GP. In addition, although the concentration of calcium increased with time, the ratio of calcium:phosphorus at 7 days was low (10 mM [3-GP 0.53 -+ 0.2, n = 4; 10 mM Pi 0.85 - 0.3, n = 4). A similar type of study was performed with MC3T3-E1 cells. After 18 days, cultures treated with 10 mM [3-GP contained 5-7 times as much calcium as control cultures. However, the calcium:phosphorus ratio of mineral present in the cell layer was very low (0.67 - 0.1; n = 4). Selected areas of day 14 cultures were evaluated using SEM and EDX. Figure 6A-C shows the morphological appearance of ROS cells treated with 13-GP or Pi. The presence of 13-GP caused the appearance of mineral deposits in the ROS cell layer; the effect of Pi was somewhat different in that some areas of the culture were covered with mineral even in the absence of cells. EDX analysis of selected areas
308
C-H Chung et al.: [~-Glycerophosphate and Bone Cell Mineralization
Fig. 4. Autoradiogram of high efficiency, thin layer chromatogram of ROS cell culture phospholipids. Preconfluent ROS cell cultures (day 7) were pretreated with 32[p] orthophosphate for 24 hours. Phosphatides were extracted from the cell layer and separated by two-dimensional chromatography (directions of the 1st and 2nd solvent systems are shown in the figure). The patterns of labeling of phospholipids of the treated and control cultures were similar. (A) Control, (B) culture treated with 10 mM [3-GP. o = origin, a = diphosphatidylinositol, b = sphingomyelin, c = lecithin, d = phosphatidylinositol, e = phosphatidylethanolamine, f = phosphatidylserine, g = cardiolipin.
16
12
Fo co--.. 8
r
t-,el,ill
Con
3 mM
i0 rnM
3 mM
10 mM
fi-GP Pi Pi f{ GP Fig. 5. Accumulation of calcium by ROS cells treated with 13-GPor Pi. Preconfluent cultures of ROS cells were treated with 13-GP(3 or 10 mM) or Pi (3 or 10 mM). After 7 days, cells were harvested and calcium was released by treatment with 1.0 N HC1. Calcium was determined by atomic absorption spectrophotometry. Values shown are means -+ SEM of 3--4 determinations. Con = control.
showed that the calcium:phosphorus molar ratio was about 1.0 ([3-GP range 0.9-1.1; Pi range 0.8-1.0). Similar results were obtained using 3 mM [3-GP and Pi; here the amount of mineral deposited was less than that of the 10 mM cultures. However, mineral was only seen in a limited number of areas in the culture, and EDX analysis indicated that the calcium:phosphorus ratio was low (below 1.1). Mineral formed by day 14 bone cell cultures treated with 10 mM 13-GP was evaluated by FTIR. Comparison of the difference spectrum of the mineralized matrix with the spectrum of an apatite standard indicated that it was nonapatitic (Fig. 7A,B). However, the mineral appeared to share some similarities with octacalcium phosphate. Thus, absorption bands were seen at 854, 878, 910, 958, 1020, 1081, and 1125 c m - 1 (Fig. 7B). In view of the high level of Pi in the medium of [3-GPtreated cultures (see Fig. 1), the possibility existed that an elevation in the calcium Pi ion product could cause spontaneous (nonbiological) mineral formation. To learn whether
culture conditions promoted mineral formation, medium from postconfluent cultures was collected after 24 and 48 hours and filtered through a 0.4 tam filter. We found that mineral particles were present in medium obtained from 3 and 10 mM [3-GP and Pi-treated ROS, MC3T3-E1, and osteoblast cell cultures at both time periods. No mineral was present in control media. EDX analysis of the particles indicated that they contained mainly calcium and phosphorus (Fig. 6D). Little difference was seen in the composition of particles obtained from medium of [3-GP or Pi-treated osteoblasts, MC3T3-E1, and ROS cells. To further investigate the possibility that mineral could be precipitated in [3-GP-treated bone cell cultures, experiments were performed with fibroblasts that have a low endogenous AP activity and do not mineralize. In this study, fibroblasts were treated with [3-GP or Pi and calcium accumulation in the culture was examined by alizarin staining. After 7 days, there was no evidence of calcium accumulation by control cells and cells treated with [3-GP; in contrast, when the medium Pi supplementation was greater than 2 mM, alizarin staining indicated that calcium was precipitated into the cell layer (Fig. 8). In the presence of 10 mM Pi, a considerable amount of alizarin-positive material was seen in the fibroblast culture. At the latter Pi concentration, mineral deposition was associated with cell death. Finally, to ascertain whether AP alone could cause spontaneous mineral deposition, collagen discs were adhered to the bottom of culture dishes that contained medium supplemented with purified AP and 10 mM [3-GP. In this cell-free system, [3-GP was hydrolyzed by AP to release Pi. After 24 hours, the Pi concentration of the medium was about 9 mM; after 48 hours, analysis of the discs indicated that 70.3 +- 5.9% (n = 4) of the total medium calcium concentration was deposited.
Discussion
The goal of this investigation was to examine the mechanism by which [3-GP promoted mineral deposition by bone cells in culture. Analysis of a number of metabolic parameters indicated that neither [3-GP nor Pi modulated any of the major biosynthetic activities of ROS cells. Thus, we found no change in the levels of acidic phospholipids, such as phos-
C-H Chung et al.: 13-Glycerophosphate and Bone Cell Mineralization
309
Fig. 6. SEM appearance of selected areas of 14-day ROS cultures and medium mineral particles. (A) Control culture showing the presence of ROS cells; (B) culture treated with 10 mM 13-GP;(C) culture treated with 10 mM Pi; (D) mineral particles in medium of cultures
treated with 10 mM [3-GP and filtered through a 0.4 p.m Nuclepore membrane. Insets show EDX spectrum and the calcium and phosphorus peaks. Bar = 10 IxM (A, B, C); bar = 1 I~M (D).
phatidylserine, a phosphatide that is required for bone and cartilage mineralization; and the cell number and total protein concentration remained constant. Measurement of lactate synthesis showed that f3-GP did not effect the rate of anaerobic glycolysis. In addition, the design of the investigation permitted us to evaluate the concept that hydrolysis of 13-GP provides components (glycerol and Pi) that modify cellular metabolism and thereby promotes mineralization of the extracellular matrix. We found that glycerol had no effect on mineral formation; moreover, the addition of Pi to the medium resulted in only minor changes in metabolic and biosynthetic activities of bone cells. Based on these findings, we reject the notion that 13-GP, a compound that can serve as an intermediate in carbohydrate and lipid metabolism, regulates bone cell activity and thereby fosters the formation of a calcified matrix. In contrast to these metabolic studies, we noted that when the medium was supplemented with Pi or [3-GP, there was rapid enhancement in the rate of mineral deposition. This event is most likely due to a rise in the Pi content of the medium. Indeed, the results of the study clearly showed that 13-GP was hydrolyzed by bone cells to produce high concentration of Pi. Thus, within 24 hours, almost 80% of 10 mM f3-GP was hydrolyzed by the bone cells. The rate of mineral formation with [3-GP was lower than that of Pi. These results were expected because (1) 13-GP has to undergo hydrolysis
to liberate Pi and as this is a time-dependent process, the Pi level of [3-GP supplemented cultures is always less than that of cultures supplemented directly with Pi; (2) the AP activity of the nonconfluent ceils is low initially and 5-7 days are required for A P activity to become maximum. Experiments with fibroblasts that have a low endogenous AP activity, and do not normally mineralize, confirmed that this enzyme was required for [3-GP-mediated mineralization. As mineral was found in fibroblast cultures treated with Pi, it must be concluded that the mineral deposited into these cultures was a precipitation artifact. We used a number of cell types to investigate the importance of t3-GP as a mediator of mineral formation by cells in culture. The cells we chose were those that have been used by previous investigators to characterize aspects of the mineralization process. Of these, the ROS 17/2.8 cells have been utilized for many studies [32-34]. This cell exhibits high AP activity and rapidly forms mineral in the presence of [3-GP [15, 35]. However, as the amount of type I collagen secreted by this cell is low, it could be argued that it is not an optimum cell type for mineralization studies. F o r this reason, we also utilized another cell line (MC3T3-E1), as well as calvarial osteoblast-like cells. Though some differences were observed between these cell types, in general, the data were consistent with the notion that the mineralization was dependent on AP-mediated release of Pi from [3-GP.
C-H Chung et al.: ~3-Glycerophosphate and Bone Cell Mineralization
310
E3
A
4000
2800
1600
Wavenumber (cm-1)
Fig. 7. FTIR difference spectrum of mineralized matrix present in day 14 ROS cell culture treated with 10 mM 13-GP. FTIR analysis was performed on freeze-dried samples using a Nicolet 5DXC spectrometer. Samples were analyzed as 10% mixtures in KBr in the
Fig. 8. Calcium accumulation in fibroblast cell cultures treated with Pi. Preconfluent cultures of rat skin fibroblasts were treated with Pi (total Pi concentration: (a) 0.9 mM, control; (b) 1.9 raM, (r 2.9 mM, (d) 3.9 mM, (e) 5.9 mM, (f) 10.9 mM). After 7 days, cultures were washed and then stained with alizarin red. Ceils stained positive with alizarin when the Pi supplementation was above 2 mM (d, e, f).
We investigated the mineral formed in bone cell cultures and spontaneous mineral deposition in the presence of 10 mM 13-GP. With respect to cell-mediated mineralization, chemical, elemental, and the F T I R analyses of the ROS cultures and chemical analysis of the MC3T3-EI cells showed that the mineral was abnormal. Thus, the calcium:phosphorus ratios were depressed and the mineral was nonapatitic. These results confirmed an earlier study that showed that 10 mM 13-GP caused the formation of nonbiological (ectopic) mineralization in cultured calvaria [36]. Moreover, we found that calcium phosphate mineral particles were present in the medium of postconfluent ROS cells, MC3T3-E1, and calva-
400
1200
1000
800
Wavenurnber (cm-1) diffuse reflectance mode. (A) Shows the spectra from 400 to 4000
cm-1. Hydroxyapatite and octacalcium phosphate were used as standards. (B) Shows details of the region 800-1200 cm-1.
rial bone cell cultures treated with 10 mM [3-GP. Experiments performed with the cell-free system in which the culture medium contained AP and 10 mM ~-GP indicated that mineral particles formed spontaneously and were deposited into the collagen matrix. The amount of mineral that accumulated in the collagen discs was very high. Thus, after 48 hours, over 70% of the total medium calcium was present as mineral in the discs. In our experiments, the concentrations of [3-GP (10 raM) was similar to that used by other investigators [3-17]. It should be noted that according to Marshall in Nordin's book and others, in serum or plasma, the level of inorganic phosphate is 1.0-1.5 mM, whereas the organic phosphate level (which is mostly phospholipid) is about 3 mM [37, 38]. However, the concentration of hydrolyzable ester phosphate is 1.1 mM or less [39, 40]. Thus, a medium concentration of 10 mM 13-GP is high when compared with physiological levels, and as we have shown here, its hydrolysis results in medium Pi levels that are supraphysiological. A consequence of the elevated Pi concentration of the medium, in the presence of 1.8 mM calcium, is the rapid formation of mineral. To minimize spontaneous mineral deposition, supplementation of the culture medium with Pi or [~-GP should be limited. Resuits of experiments performed with fibroblasts indicated that medium supplementation with Pi or 13-GP should not exceed 2 mM. If this value is exceeded, nonphysiological mineral formation will occur and this would serve to obscure the relatively low rate of mineral accretion by osteoblasts in culture.
Acknowledgments. This work was supported by NIH Grants AR34411, DE-08239, and DE-06533. C-H.C. was supported by a NIH Postdoctoral Fellowship (AR-07481). The help of Dr. S. Radin, Mr. A. El-Ghannam, Mr. J. Mayro, Mr. G. Harrison, and Ms. K. DeBott is gratefully acknowledged.
C-H Chung et al.: 13-Glycerophosphate and Bone Cell Mineralization
References 1. Robison R (1923) The possible significance of hexosephosphoric esters in ossification. Biochem J 17:286--293 2. Robison R, Rosenheim AH (1934) Calcification of hypertrophic cartilage in vitro. Biochem J 28:684-4598 3. Tenenbaum HC (1981) Role of organic phosphate in mineralization of bone in vitro. J Dent Res 60(C):1586-1589 4. Tenenbaum HC, Heersche JNM 0982) Differentiation of osteoblasts and formation of mineralized bone in vitro. Calcif Tissue Int 34:76-79 5. Ecarot-Charrier B, Glorieux FH, Van Der Rest M, Pereira G (1983) Osteoblasts isolated from mouse calvaria initiate matrix mineralization in culture. J Cell Biol 96:639-643 6. Whitson SW, Harrison W, Dunlap MK, Bowers DE, Fisher LW, Gehron Robey P, Termine JD (1984) Fetal bovine bone " cells synthesize bone-specific matrix proteins. J Cell Biol 99: 607-614 7. Gehron Robey P, Termine JD (1985) Human bone cells in vitro. Calcif Tissue Int 37:453-460 8. Bellows CG, Aubin JE, Heersche JNM, Antosz ME (1986) Mineralized bone nodules formed in vitro from enzymatically released rat calvaria cell population. Calcif Tissue Int 38:143-154 9. Gerstenfeld LC, Chipman SD, Glowacki J, Lian JB (1987) Expression of differentiated function by mineralizing cultures of chicken osteoblasts. Dev Biol 122:49-60 10. Bhargava U, Bar-Lev M, Bellows CG, Aubin JE (1988) Ultrastructural analysis of bone nodules formed in vitro by isolated fetal rat calvaria cells. Bone 9:155-163 11. Berry L, Shuttleworth CA (1989) Expression of the chondrogenie phenotype by mineralizing cultures of embryonic chick calvarial bone cells. Bone Miner 7:31-45 12. Masquelier D, Herbert B, Hauser N, Mermillod P, Schonne E, Remacle C (1990) Morphological characterization of o steoblastlike cell cultures isolated from newborn rat calvaria. Calcif Tissue Int 47:92-104 13. Bellows CG, Aubin JE, Heersehe JNM (1991) Initiation and progression of mineralization of bone nodules formed in vitro: the role of alkaline phosphatase and organic phosphate. Bone Miner 14:27-40 14. Kodama H-A, Amagai Y, Sudo H, Ohno T, Iijima K-I (1986) Culture conditions affecting differentiation and calcification in the MC3T3-E1 osteogenic cell line. In: Ali SY (ed) Cellmediated calcification and matrix vesicles. Elsevier Science Publishers B.V., pp 297-302 15. Nishimoto SK, Stryker WF, Nimni ME (1987) Calcification of osteoblastlike rat osteosarcoma cells in agarose suspension cultures. Caleif Tissue Int 41:274-280 16. Thomas JT, Boot-Handford RP, Grant ME (1990) Modulation of type X collagen gene expression by calcium [3-glycerophosphate and levamisole: implications for a possible role for type X collagen in endochondral bone formation. J Cell Sci 95:639-648 17. Luria EA, Owen ME, Friedenstein AJ, Morris JF, Kuznetsow SA (1987) Bone formation in organ cultures of bone marrow. Calcif Tissue Int 248:449-454 18. Tenenbaum HC (1987) Levamisole and inorganic pyrophosphate inhibit 13-glycerophosphate-induced mineralization of bone formed in vitro. Bone Miner 3:13-26 19. Fortuna R, Anderson HC, Carty RP, Sajdera SW (1980) Enzymatic characterization of the matrix vesicle alkaline phosphatase isolated from bovine fetal epiphyseal cartilage. Calcif Tissue Int 30:217-225 20. Raisz LG, Dietrich JW, Canalis EM (1976) Factors influencing bone formation in organ cultures. Isr J Med Sci 12:108--114
311 21. Kakuta S, Golub EE, Shapiro IM (1985) Morphochemical analysis of phosphorus pools in calcifying cartilage. Calcif Tissue lnt 37:293-299 22. Kakuta S, Golub EE, Haselgrove JC, Chance B, Frasca P, Shapiro IM (1986) Redox studies of the epiphyseal growth cartilage: pyridine nucleotide metabolism and the development of mineralization. J Bone Miner Res 1(5):433--440 23. Pacifici M, Oshima O, Fisher LW, Young MF, Shapiro IM, Leboy PS (1990) Changes in osteonectin distribution and levels are associated with mineralization of the chicken tibial growth cartilage. Calcif Tissue Int 47:51--61 24. Bradford MM (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72:248-254 25. Sasaki T, Hasegawa-Sasaki H (1985) Breakdown of phosphatidylinositol 4,5-biphosphate in a T-cell leukaemia line stimulated by phytohaemagglutinin is not dependent on Ca2+ mobilization. Biochem J 227:971-979 26. Golub EE, Schattschneider SC, Berthold P, Burke A, Shapiro IM (1983) Induction of chondrocyte vesiculation in vitro. J Biol Chem 258:616--621 27. Chen PS, Toribara TY, Warner H (1956) Microdetermination of phosphorus. Anal Chem 28:1756-1758 28. Jung K, Pergande M (1980) Influence of inorganic phosphate on the activity determination of isoenzymes of alkaline phosphatase in various buffer systems. Clin Chim Acta 102:215-219 29. Boskey AL, Posner AS (1977) The role of synthetic and boneextracted Ca-phospholipid-PO4 complexes in hydroxyapatite formation. Calcif Tissue Res 23:251-258 30. Wuthier RE, Gore ST (1977) Partition of inorganic ions and phospholipids in isolated cell, membrane and matrix vesicle fractions: evidence for Ca-Pi-Acidic phospholipid complexes. Calcif Tissue Res 24:163-17l 31. Boyan-Salyers BD, Boskey AL (1980) Relationship between proteolipids and calcium-phospholipid-phosphate complexes in bacterionema matruchotii calcification. Calcif Tissue Int 30: 16%174 32. Rodan GA, Majeska RJ (1982) Phenotypic maturation of osteoblastic osteosarcoma cells in culture. Clin Biol Res 110(B):249259 33. Rodan GA, Rodan SB (1983) Expression of the osteoblastic phenotype. In: Peck WA (ed) Bone and mineral research, Annual 2. Elsevier Science Publishers B.V., pp 244-285 34. Rodan GA, Heath JK, Yoon K, Noda M, Rodan SB (1988) Diversity of the osteoblastic phenotype. Ciba Foundation Symposium 136:78-91 35. Yoon K, Golub E, Rodan GA (1989) Alkaline phosphatase eDNA transfected cells promote calcium and phosphate deposition. Connect Tissue Res 22:17-25 36. Gronowiez G, Woodiel FN, McCarthy M-B, Raisz LG (1989) In vitro mineralization of fetal rat parietal bones in defined serumfree medium: effect of 13-glycerolphosphate. J Bone Miner Res 3:313-324 37. Nordin BEC (1976) Calcium, phosphate and magnesium metabolism. Churchill Livingston, Edinburgh, London, New York 38. Scully RE (1980) Normal reference values. N Engl J Med 302:3% 48 39. Hoffman WS (1970) The biochemistry of clinical medicine, 4th ed. Year Book Medical Publishers 40. Geigy Scientific Tables, Lentner C (ed) (1984), vol 3, 8th ed. Ciba-Geigy
