Reversible Suppression of in Vitro Biomineralization by Activation of Protein Kinase A*

Parathyroid hormone (PTH-(1–34)) potently suppresses apatite deposition in osteoblastic cultures. These inhibitory effects are mediated through signaling events following PTH receptor binding. Using both selective inhibitors and activators of protein kinase A (PKA), this study shows that a transient activation of PKA is sufficient to account for PTH’s inhibition of apatite deposition. This inhibition is not a result of reduced cell proliferation, reduced alkaline phosphatase activity, increased collagenase production, or lowering medium pH. Rather, data suggest a functional relation-ship between matrix assembly and apatite deposition in vitro . Bone sialoprotein (BSP) and apatite co-localize in the extracellular matrix of mineralizing cultures, with matrix deposition of BSP temporally preceding that of apatite. Transient activation of PKA by either PTH-(1– 34) or short term cAMP analog treatment blocks the deposition of BSP in the extracellular matrix without a significant reduction in the total amount of BSP synthesized and secreted. This effect is reversible after allow-ing the cultures to recover in the absence of PKA activators for several days. Thus, a transient activation of PKA may suppress mineral deposition in vitro as a con-sequence of altering the assembly of an extracellular matrix permissive for apatite formation. The need to better understand the biological nature of bone is by the fact

The need to better understand the biological nature of bone mineralization is illustrated by the fact that a properly mineralized endoskeleton is crucial for vertebrate survival (1,2). To date, several biomineralization theories have been formulated, and each advances a mechanistic model to explain bone formation (3). Lately, these models have emphasized the role of several bone-specific, noncollagenous proteins (4) as promoting the tightly regulated processes of apatite crystal nucleation and growth in an extracellular environment. One such candidate is bone sialoprotein (BSP), 1 a sulfated, phosphorylated matrix glycoprotein (5-8) with a restricted pattern of expres-sion within mineralizing tissues (9 -16), which co-localizes with the smallest detectable foci of newly forming mineralized matrix in osteoid (17).
Parathyroid hormone (PTH) is a calciotropic hormone that can stimulate either formation or resorption of bone in vivo depending on such factors as dosage, delivery modality, and treatment frequency (18 -21). In vitro, a continuous exposure to PTH results in the suppression of biomineralization in primary cultures of calvarial osteoblasts (22), calvarial explants (23), or growth plate chondrocytes (24,25). However, the molecular mechanisms mediating PTH's inhibitory effect on in vitro biomineralization reactions have not been elucidated yet. The first 34 amino acid residues in PTH contain most of the biological activity of the full-length, natural hormone including ligand-receptor binding and signal transduction functions (26,27). Two major plasma membrane reactions occur as a result of PTH's binding to its receptor that then initiate three intracellular signaling cascades. First, there is a G protein-coupled stimulation of adenylate cyclase that produces cAMP and thereby activates protein kinase A (PKA) (28). Second, there is a stimulation of phospholipase C activity, which produces inositol triphosphate and diacyl glycerol, thereby activating intracellular calcium release and protein kinase C, respectively (29).
UMR 106-01 BSP osteosarcoma cells model a number of mature osteoblast phenotypic properties. These include their morphological appearance (30), normal receptor-mediated responses to calciotropic agents such as parathyroid hormone (31) and 1,25-(OH) 2 vitamin D 3 (32), and a relatively high expression of cell surface alkaline phosphatase activity (30). UMR cells also produce large quantities of BSP (6) and deposit a bioapatitic mineral phase in their extracellular matrix when exposed to phosphate supplements (33). This biomineralization process is an active metabolic process requiring ongoing protein synthesis and secretion (33). Accordingly, this current study aimed to (a) determine whether the biomineralization process exhibited by UMR osteoblastic cells, like primary osteoblasts, is inhibited by PTH, (b) identify the predominant signal transduction process that leads to this inhibitory effect, and (c) identify whether BSP is involved in these PTH-regulated biomineralization reactions.
Cell Culture-UMR 106 -01 BSP cells were routinely passaged in T-75 culture flasks and cultured in Eagle's minimum essential medium plus nonessential amino acids, 20 mM HEPES (pH 7.2), and 10% fetal bovine serum (growth medium) (6). The normal calcium and P i concentrations for Eagle's minimum essential medium are 1.8 and 1 mM, respectively. Cultures were incubated at 37°C in a humidified 5% CO 2 atmosphere with routine passage every 3 days.
Experiments were set up by briefly washing the confluent T-75 cell layer with Hanks' balanced saline solution without Ca 2ϩ or Mg 2ϩ followed by trypsinization of the cells (10 ml of 0.05% trypsin plus 0.53 mM EDTA in Hanks' solution at 37°C for 10 min). Cells were counted by hemacytometer and plated at 2000 cells/mm 2 into six-well (960 mm 2 , 3 ml of growth medium) or 12-well (380 mm 2 , 1.2 ml of growth medium) cluster dishes. Cultures were incubated for 4 -6 h to allow the attachment and spreading of UMR cells on the culture surface. Peptide hormones or chemical agents were added to the cultures after this initial attachment period, and these treatments were continued throughout the entire culture period unless otherwise stated. Plating efficiency of UMR cells was not affected by any of the reported treatments. At 64 h of incubation, the medium was replaced with 2 ml of fresh growth medium for an additional 24 h (biomineralization period) supplemented with or without either ␤-glycerophosphate (␤-GP) or phosphoserine, each at a final concentration of 7 mM (33). UMR biomineralization capacity is stable and reproducible from passages 5-50; all experiments were done between passage 10 and 30. All organophosphate supplements were prepared as sterile 0.5 M stocks in Nanopure water titrated to a final pH of 7.0. Aliquots of these agents were added directly to fresh medium immediately before initiating biomineralization. All peptides and reagents were stored as prealiquoted, sterile stock solutions at Ϫ70°C and diluted to working concentrations in growth medium on the day of the experiment. Unless otherwise indicated, experimental data are reported as mean Ϯ S.D.; n ϭ 3/trial; data are representative of at least two trials.
Alkaline Phosphatase Assay-Organophosphate hydrolytic activity on the surface of viable UMR cells was determined in the presence or absence of PTH-(1-34) via a modification of an existing in situ assay (34). Briefly, p-nitrophenyl phosphate (final concentration of 10 mM) was added to living cultures bathed in phenol red-free medium and placed in a rotary shaking incubator at 37°C. During this incubation, aliquots of the medium were removed at 15, 30, and 60 min and added to a solution of 0.1 M NaOH. Released p-nitrophenol was quantified in each sample by its absorbance at 400 nm using a multiwell spectrophotometer (SpectraMAX 250, Molecular Devices) and a p-nitrophenol standard curve. The conversion rate of p-nitrophenyl phosphate to p-nitrophenol was calculated from the slope of the regression lines generated for each set of time points (nmol of p-nitrophenol produced per min). Specific activity was calculated by normalizing the enzyme conversion rate to the total DNA content of each culture. Phosphate release from ␤-GP was assayed from medium samples as described previously (33) to confirm the results of the above described alkaline phosphatase assay.
Alizarin Red-S Assay to Quantify Apatite Content-Alizarin red-S (AR-S) is a dye that binds selectively to calcium salts binding ϳ2 mol of Ca 2ϩ per mole of dye and is used widely for calcium mineral histochemistry (33). At the end of each experiment, cultures were briefly rinsed with PBS followed by fixation (ice-cold 70% ethanol, 1 h). Cultures were rinsed with Nanopure water and stained for 10 min with 40 mM AR-S, pH 4.2, at room temperature with rotation (1 ml/35-mm dish). Cultures were then rinsed five times with water followed by a 15-min wash with PBS (with rotation) to reduce nonspecific AR-S stain. Stained cultures were photographed followed by a quantitative destaining procedure using 10% (w/v) cetylpyridinium chloride in 10 mM sodium phosphate (pH 7.0) for 15 min at room temperature. Aliquots of these AR-S extracts were diluted 10-fold in PBS, and the AR-S concentration was determined by absorbance measurement at 552 nm on a multiwell spectrophotometer using an AR-S standard curve in the same solution. Values were normalized to total DNA content per culture.
Atomic Absorption-Cultures were washed twice with PBS, and cell layer minerals were extracted with a 24-h exposure to 0.6 N HCl in 0.02 M PBS at room temperature. Aliquots were then added to a solution of 5 mM lanthanum oxide in 40 mM HCl followed by atomic absorption analysis of calcium content using a Perkin-Elmer model 2380 atomic absorption spectrophotometer optimized to 422.7 nm (35). The atomic absorption lamp (0.125 M detection limit) was optimized with a calcium standard curve (2.5-1200 M) using a serial dilution of a calcium atomic absorption standard (Sigma, C-5649). Values were normalized to total DNA as described below.
DNA Assay-Ethanol-fixed cultures were rinsed with Nanopure water and then solubilized with a solution of 10 M formamide, 1% (w/v) SDS, 50 mM sodium acetate, pH 6.0 (60°C, 1.5 h). After cooling, lysates were sonicated briefly on ice (200 watts, 20 s) to reduce the viscosity of the samples. This solubilization procedure did not affect the doublestranded nature of the DNA necessary for fluorometric detection. DNA content was determined from replicate cultures using the high salt (2 M NaCl) Tris-NaCl-EDTA fluorometric approach with the Hoechst 33258 dye binding assay (Hoefer Scientific) in an SLT Fluorostar ϩ multiwell fluorometer (Tecan).
Northern Blot Analysis-RNA was extracted and purified from PBSwashed cultures using the Tri-Reagent method (36,37). Equal amounts of RNA were denatured by the method of Gong (38) and fractionated on 1.2% agarose (Life Technologies, Inc.) gels containing 2.2 M formaldehyde, 1 mM EDTA, 5 mM sodium acetate, and 20 mM MOPS, pH 7.0. rRNA bands were visualized by ethidium bromide/UV transillumination, transferred by capillary blotting to Hybond-N nylon membranes in 20ϫ standard saline citrate (SSC) according to the method of Southern (39) and cross-linked to the membrane by UV irradiation (600 microwatts/cm 2 at 250 nm for 3 min). Membranes were washed with 2ϫ SSC, air-dried, and baked at 80°C under vacuum for 2 h. cDNA probes are labeled with [␣-32 P]dCTP by random primer method using a random primer DNA labeling kit (Roche Molecular Biochemicals). Membranes were prehybridized for 15 min at 65°C in Quik Hyb hybridization solution (Stratagene). Hybridization was carried out at 65°C for 1 h in Quik Hyb solution containing 500 g/ml heat-denatured salmon sperm DNA and 32 P-labeled cDNA probes. After hybridization, blots were washed twice for 15 min at 42°C in 2ϫ SSC, 0.1% SDS; once for 30 min in 1ϫ SSC, 0.1% SDS at 42°C; and finally for 30 min in 0.1ϫ SSC, 0.1% SDS at 55°C. mRNA signals were detected by x-ray film (Eastman Kodak Co.), and rRNA bands were photographed using Polaroid film. All films were digitized at 12-bit depth using a ScanMaker 4 flatbed scanner (Microtek), and relative band intensities were quantified using Gel-Pro Analyzer, version 3.0 (Media Cybernetics).
Western Blot Analyses-Proteins were extracted from cell/matrix and medium samples using a 2-h treatment at 60°C with a solution of 10 M formamide, 1% SDS, and 50 mM sodium acetate pH 6.0 (with or without 10 mM EDTA). Total protein content was determined using the Mi-croBCA assay (Pierce) as per the manufacturer's instructions. Extracted proteins were separated on SDS-polyacrylamide gel electrophoresis using linear 4 -20% minigels and the buffer system of Laemmli (40). Proteins were transferred onto polyvinylidene difluoride membranes at 100 V for 1 h in a minitransblot apparatus using a transfer buffer of 25 mM Tris, 192 mM glycine, 20% methanol, pH 8.3. Membranes were blocked with 3% (w/v) bovine serum albumin (BSA) in Tris-buffered saline (TBS) to prevent nonspecific binding of antibodies. Primary antibodies were diluted (1:500) in 1% BSA in TBS and applied for 1 h to the transfer membranes. After extensive washing, secondary antibody was applied (1:1000 dilution with 1% BSA in TBS) for 30 min and detected using the ECL Western blotting analysis sytem (Amersham Pharmacia Biotech) as per the manufacturer's instructions. Western blot signals were detected by BioMax x-ray film (Eastman Kodak Co.). All films were digitized at 12-bit depth using a ScanMaker 4 flatbed scanner (Microtek), and relative band intensities were quantified using the Gel-Pro Analyzer, version 3.0 (Media Cybernetics).
Immunohistochemistry-Samples were washed with TBS and then fixed with 70% ethanol (4°C) for 1 h. Fixed samples were washed three times with TBS followed by a 30-min incubation with 1% BSA in TBS to block nonspecific binding sites. Primary antibodies were diluted 200fold in 1% BSA in TBS and applied to blocked samples for 1 h. After primary antibody exposure, the cultures were washed three times with TBS and then incubated for 1 h with a FITC-conjugated secondary antibody (1:1000 dilution with 1% BSA in TBS), and finally the samples were washed five times with TBS to remove the secondary antibody. Specimens were washed briefly with Nanopure water, stained with either 400 M alizarin red-S, pH 4.2, or 10 M alizarin complexon, pH 4.0 (both from Sigma) for 10 min, washed five times with Nanopure water, and finally washed with TBS for 10 min. Stained samples were mounted in VectaShield (Vector Laboratories), and the edges of the coverslips were sealed with clear nail polish. Confocal images of the samples were obtained using a Leica TCS-NT laser-scanning, confocal microscope equipped with krypton and argon lasers. Digital images reported in this study are maximum intensity projections. Primary antibody specificity for BSP was confirmed by Western blot analysis using protein extracts from UMR cultures. Secondary antibody specificity was confirmed by omitting the primary antibody during the staining protocol. Alizarin specificity for staining apatite deposits was confirmed by decalcifying UMR cultures (10 mM EDTA in TBS for 30 min at 21°C or 5% trichloroacetic acid for 5 min at 4°C) prior to staining. Decalcified cultures bound primary and secondary antibodies in a manner identical to that for fully mineralized cultures, indicating that these antibodies were not binding to apatite crystals.

PTH-(1-34) Potently Suppresses Apatite Deposition in UMR
Cultures-UMR cultures rapidly form an apatite crystalline phase in their extracellular matrix when supplemented with an exogenous organophosphate source (33). Cultures treated for 24 h with either 7 mM ␤-GP or serine phosphate (SP) bind 12-13 nmol of AR-S/g of DNA in the cell/matrix layer, while unsupplemented cultures bind ϳ1 nmol of AR-S/g of DNA ( Fig. 1B, left panel). This difference, 11-12 nmol of AR-S/g of DNA, equating to 17-18 nmol of Ca 2ϩ /g of DNA (33), is a measure of the amount of apatite deposited in the cell/matrix layer. Hereafter, we set this value at 100%, and the value for unsupplemented cultures is set at 0%, representing a relative scale we refer to as the percentage of apatite content of UMR cultures.
Cultures exposed to as little as 0.1 nM PTH-(1-34) exhibited marked reductions in their AR-S staining ( Fig. 1A) with the percentage of apatite content assayed at ϳ20 or ϳ10% of the untreated control value when supplemented with either ␤-GP or SP, respectively (Fig. 1B, right panel). PTH-(1-34) concentrations of 1 nM or greater completely suppressed apatite deposition in cultures supplemented with either ␤-GP or SP. This inhibition of apatite deposition in UMR cultures, including the effective PTH dose range, is remarkably similar to PTH's potent suppression of bone nodule formation in primary osteoblast cultures exposed to 10 mM ␤-GP (22). This concordance with the response of normal osteoblasts to PTH substantiates the use of this osteoblastic cell line to study the mechanism(s) of this hormone's inhibition of in vitro biomineralization.
PTH-(1-34) Suppresses Apatite Deposition via an Activation of PKA-PTH induces both adenylate cyclase and phospholipase C activities in UMR cells (41)(42)(43)(44)(45). Several strategies were employed to assess which, if any, of these signal cascades induced by PTH-(1-34) is involved in the suppression of apatite production in UMR cultures. One approach employed a PTH peptide antagonist, Nle 8,18 -Tyr 34 -PTH-(3-34) amide (PTH-(3-34)), which binds to the PTH receptor with high affinity (46, 47) but does not produce a cAMP response needed to activate PKA at doses up to 10 nM (43)(44)(45)(46)(47)(48)(49). When added to cultures at concentrations up to 10 nM, PTH-(3-34) did not suppress apatite deposition in UMR cultures (Fig. 1, A and B). The 10 -20% suppression in apatite content of UMR cultures treated with a 100 nM concentration of the antagonist peptide may be the result of a slight stimulation of PKA at this high dose (43,48), a low affinity stimulation of other signaling mechanisms (44,45,49), 2 or a minor agonist contaminant in the peptide prepa-ration (46,47). These data indicate that mere binding of ligand to the PTH receptor is not sufficient to achieve a significant inhibition of apatite deposition in UMR cultures. Rather, they indicate that a signal transduction response is required and suggest that a cAMP-mediated activation of PKA is likely to be a major factor that transduces PTH's suppressive effects on osteoblast-mediated biomineralization.
Another approach attempted to block the inhibitory effect of PTH with a selective competitive inhibitor of PKA known as H89 (51). H89 alone, at doses as high as 0.5 M, did not affect the apatite content (Fig. 2) or cell density (not shown) of UMR cultures. A small reduction in apatite content and cell density occurred at H89 doses above 0.5 M. Cultures exposed to 0.2 nM PTH-(1-34) exhibited only 8% of the apatite content of untreated cultures, while similar cultures treated with increasing concentrations of H89 yielded dose-dependent increases in their apatite contents (Fig. 2). An optimal reversal of the in-2 Treatment with protein kinase C activators such as phorbol esters or mezerein (50)     amide in the presence of 7 mM SP. Duplicate negative (Ϫ7 mM SP) and positive (ϩ7 mM SP) control wells are shown. Apatite was detected in ethanol-fixed cultures using AR-S staining. B, AR-S extracted from stained cultures and normalized to DNA content; data represent the mean Ϯ S.D. from three independent trials (n ϭ 6 total cultures per value). Percentage of apatite content is defined under "Results." The 100% value is 11.9 Ϯ 0.7 nmol of AR-S/g of DNA. Identical results were obtained using natural sequence PTH peptides. hibitory effects of 0.2 nM PTH-(1-34) was achieved with 0.5 M H89, yielding an apatite content 71% of that measured for untreated control cultures. Concentrations of H89 greater than 0.5 M did not increase further the apatite content of PTH-(1-34)-treated cultures, although they did continue to decrease the difference between treated and control samples (Fig. 2). The efficacy of H89 at reversing the effects of PTH decreased with increasing doses of PTH used in the experiment. For example, 0.5 M H89 increased the apatite content of cultures treated with 0.2 nM PTH-(1-34) to 71% of control levels, while that for cultures treated with 2 nM PTH-(1-34) increased to only 42% of untreated controls. This presumably reflects H89's reduced capabilities to inhibit PKA activity in intact cells (51), especially when competing against a more effective production of cAMP at higher PTH doses. Thus, a selective inhibitor of PKA substantially reversed the inhibitory effects of PTH-(1-34) on apatite deposition in UMR cultures. This observation suggests that PKA activation plays a major role in PTH's suppression of biomineralization in vitro.
UMR cells exhibit a proportional increase in PKA activity when intracellular cAMP levels are increased (41). Thus, a third approach to delineate the underlying signal transduction mechanism involved replacing PTH-(1-34) with cell-permeable, nonhydrolyzable cAMP analogs that would substitute for the natural second messenger and specifically activate PKA in UMR cells. 3 Exposure of UMR cultures to increasing concentrations of various cAMP analogs yielded a dose-dependent inhibition of apatite production (Fig. 3) as effective as that induced by PTH- . Treatment with 0.1 mM cAMP analog resulted in an apatite content only 13% that of untreated controls, and a 0.2 mM dose nearly eliminated the apatite content of UMR cultures. The inhibitory effect of cAMP analogs (Ͻ0.2 mM) on the apatite content of UMR cultures was not the result of a greatly decreased cell density or reduction in alkaline phosphatase activity (see below). This effect was specific to cAMP analogs, since cGMP analogs did not reduce the apatite content of UMR cultures (Fig. 3), although these cells can produce and respond to cGMP (52). In addition, H89 substantially reversed the inhibitory effects of 0.1 mM cAMP analog, yielding an apatite content that was 82% of the untreated control value (Fig. 3). Thus, a direct activation of PKA is sufficient to initiate a substantial inhibition of apatite deposition in UMR cultures. All of these findings indicate that PTH-(1-34) exerts much of its potent inhibitory effect on apatite deposition in UMR cultures through a cAMP-mediated signal transduction mechanism, leading to an activation of PKA.
PKA Activation Does Not Significantly Reduce Cell Density-Exposure to 0.1 nM PTH-(1-34) did not alter the DNA content of cultures compared with untreated controls, while 1 or 10 nM doses only reduced the DNA content at 72-h incubation by 11 or 18% (Table I). Similar data were obtained using 0.1-0.5 mM cAMP analog (data not shown). Thus, a substantial decrease in cell density cannot account for the suppression of apatite deposition by 0.1-2 nM PTH-(1-34) or 0.1-0.2 mM cAMP analog treatments.
PKA Activation Does Not Significantly Reduce Alkaline Phosphatase Activity-Treatment with up to 10 nM PTH-(1-34) did not significantly reduce the rate of organophosphate hydrolysis by alkaline phosphatase compared with untreated controls (Table I). These data indicate that the suppression of apatite deposition by PTH-(1-34) at concentrations below 10 nM appears to be a metabolic effect not involving an inhibition of alkaline phosphatase activity. Thus, PTH's inhibition of apatite deposition in UMR cultures cannot be explained simply as an inadequate availability of phosphate ions to promote hydroxyapatite formation.  2. Effects of a protein kinase A inhibitor, H89, on PTH-(1-34) inhibition of apatite content in UMR cultures. Cells were treated for a total of 84 h starting at 4 h after plating; H89 was added 30 min before the PTH peptide. Mineralization was initiated by adding 7 mM ␤-GP during the final 24 h of culture. Data represent the mean Ϯ S.D. from two independent trials (n ϭ 6 total cultures per value). The 100% value is 11.3 Ϯ 0.5 nmol of AR-S/g of DNA.
FIG. 3. Effects of cyclic nucleotide analogs on the apatite content of UMR cultures. Data are from experiments using 8-bromocyclic nucleotide analogs; similar data were obtained with dibutyryl analogs. Cells were treated for a total of 84 h starting at 4 h after plating; H89 was added 30 min before the cAMP analog. Mineralization was initiated by adding 7 mM ␤-GP during the final 24 h of culture. Data represent the mean Ϯ S.D. from two independent trials (n ϭ 6 cultures per value). The 100% value is 11.1 Ϯ 0.9 nmol of AR-S/g of DNA.
Other Biological Responses Elicited by PKA Activation-PTH-(1-34) at 0.1 nM inhibits 80 -90% of mineral deposition in UMR cultures (Fig. 1). This same concentration only lowers intracellular pH by 0.01-0.02 units (53) and does not stimulate Na ϩ influx/Hϩ efflux (54) in UMR cells. No significant change in medium pH was detected up to 2 nM PTH-(1-34) with the pH for all media ranging from 6.9 to 7.2. Furthermore, apatite crystals deposited in UMR (33) and chondrocyte (25) cultures are detectable by AR-S staining, which is done at a pH of 4.2.
Thus, the mineral deposits are likely to be stable during any small pH drop caused by PTH. Taken together, data indicate that it is highly unlikely that a lowering of intracellular or extracellular pH by PTH could account for PTH's potent suppression of mineral deposition by PTH.
UMR cells have been reported to secrete neutral collagenases when treated with 10 -100 nM PTH, and such protease activity might alter extracellular matrix deposition in these cultures (55). However, lower doses of PTH (0.1-1 nM) do not stimulate collagenase activity and would not be expected to cause alterations in collagenous matrix deposition. Furthermore, exogenous treatment of UMR cultures with bacterial collagenase does not significantly inhibit their ability to deposit apatite (data not shown). Thus, a secretion of neutral collagenases by UMR cells is unlikely to account for the potent suppression of mineral deposition by low doses of PTH.
UMR cells can secrete and activate TGF-␤ when treated with PTH (56), and such an autocrine response might affect osteoblastic expression and matrix deposition (57)(58)(59). However, the concentration of PTH needed to generate significant amounts of active TGF-␤ is rather high, 20 nM. This concentration is 100-fold greater than the amount of PTH needed to completely suppress UMR mineralization (0.2 nM). Furthermore, TGF-␤ inhibits alkaline phosphatase activity in osteoblastic cultures (57)(58)(59), while PTH treatment of UMR cells does not reduce alkaline phosphatase activity. Thus, a secretion of active TGF-␤ in response to PTH treatment is unlikely to explain this hormone's potent suppression of mineral deposition.
Spatial-Temporal Patterns of BSP and Apatite Deposition in Mineralizing UMR Cultures-As discussed above, many of the possible biological responses of UMR cells to PTH treatment could not account for the potent suppression of mineral deposition by this hormone. Thus, we sought to better understand the mineralization process in an attempt to identify matrix alterations that better correlate with the loss of mineral content caused by PTH treatment. Fig. 4 depicts the immunostaining results for localizing BSP versus apatite deposits in fixed UMR cell/matrix layers at the end of the biomineralization period. The results indicate that unmineralized cultures do not deposit any apatite and only a relatively small amount of BSP (Fig. 4, A-D), while mineralized cultures exhibit larger amounts of both apatite and BSP deposited in close spatial correspondence (Fig. 4, E-H).
In order to determine whether BSP participates in apatite crystal formation or merely binds to preformed crystals, time course experiments were done to assess the order, location, and quantities of BSP as compared with apatite in the extracellular matrix of mineralizing UMR cultures. Using confocal imaging, BSP is first detected in the extracellular matrix 4 h after ␤-GP addition (Fig. 5A), whereas no evidence of apatite deposition is observed at this time (Fig. 5D). A significant increase in the amount of BSP deposited in the matrix is observed by 8 h after ␤-GP addition (Fig. 5B), while evidence of apatite deposition is barely detectable at this time (Fig. 5E). An additional 4 h of treatment (12 h total) shows a continued increase in BSP deposition (Fig. 5C) and clear evidence for the presence of apatite in the matrix (Fig. 5F). At this time, nearly all of the nascent mineral is spatially located in close proximity to the BSP deposits (Fig. 5I).
While the above results establish the location and kinetics of accumulation, the temporal order of BSP or apatite deposition in UMR cultures is best addressed by quantifying the amounts of calcium and BSP in the cell/matrix layers over the mineralization period. Table II shows that calcium levels in the cell/ matrix layers of ␤-GP-treated cultures are not significantly different from those of controls at 4-and 8-h time points but increase linearly thereafter to a difference nearly 20-fold higher than control levels at the 16-h time point. These values are virtually identical to data previously reported for this model system (33), and the temporal profile is consistent with the timing of when apatite crystals first appear as revealed by confocal imaging (Fig. 5). Thus, there does not seem to be any significant amounts of apatite deposited by 8 h after adding ␤-GP to UMR cultures.
In contrast, substantial amounts of BSP were detected in the cell/matrix layer of UMR cultures after 8 h of exposure to ␤-GP.
Using the same BSP antibody as that used for the immunohistochemistry shown in Figs. 4 and 5, Western blot analysis indicated that roughly one-third of the total BSP produced during the first 8 h of ␤-GP exposure was deposited in the cell/matrix layer of UMR cultures (Fig. 6). Taken together, the morphologic and biochemical data indicate that BSP is deposited in significant amounts in the extracellular matrix of UMR cultures a few hours prior to the detection of the first apatite crystals, which subsequently appear in the very same locations as for BSP. These observations meet the necessary spatialtemporal criteria consistent with the rationale that some of the newly synthesized BSP is involved in the deposition of apatite in UMR cultures and not just binding to preformed mineral deposits. Thus, BSP represents a convenient reporter of matrix assembly required for apatite deposition in UMR cultures. Further experiments were undertaken to ascertain the effects of PTH on BSP deposition in UMR cultures under mineralizing conditions. As expected, cultures treated with 2 nM PTH-(1-34) did not deposit any detectable apatite in the presence of 7 mM ␤-GP (results identical to that in Fig. 4B). Similarly, only small amounts of BSP accumulated in their cell/ matrix layers (results similar to that in Fig. 4C). Thus, PTH coordinately suppresses both BSP and apatite deposition in the cell/matrix layers of UMR cultures under mineralizing conditions. Transient Reduction of bsp mRNA Levels by PKA Activation-Northern blot analysis of UMR cultures after a 12-h exposure to 2 nM PTH-(1-34) or 0.2 mM cAMP analog revealed large reductions in the steady-state level of bsp mRNA, respectively (Fig. 7). These reductions were not the result of a general suppression of transcription by these agents, since similar 12-h treatments did not affect significantly col1a1 mRNA levels (Fig. 7).
A complete inhibition of bsp mRNA expression was observed for a long term, continuous treatment with cAMP analog but not with PTH- (1-34), which exhibited no effect on bsp mRNA levels after 64 h of incubation (Fig. 7). This recovery of bsp mRNA levels with PTH-(1-34) treatment could reflect a combination of two metabolic processes. First, PTH-(1-34) only transiently elevates intracellular cAMP levels in UMR cells (60) due, in part, to the hydrolytic activity of cAMP phosphodiesterases, which degrades this potent second messenger molecule, thereby diminishing overall PKA activity. In contrast, cAMP analogs are poorly hydrolyzed by these phosphodiesterases, resulting in a specific and constitutive activation of PKA. Second, PTH-(1-34) directly stimulates several other signaling cascades not activated by cAMP analogs, and these reactions TABLE II Calcium per DNA levels for cultures treated with ␤-glycerophosphate UMR cells were cultured in 9.6-cm 2 wells for 64 h. Mineralization was initiated at this time by adding fresh media without (control) or with 7 mM ␤-glycerophosphate (␤-GP) during the final 24 h of culture. Cultures were terminated at the indicated times and washed with Tris-buffered saline several times, and the cell/matrix layers were processed for calcium content via atomic absorption or DNA content by fluorometric assay as described under "Experimental Procedures." Data represent the mean Ϯ S.D. from two independent trials (n ϭ 6 cultures/value  (Fig. 1), which is substantially reversed by simultaneous treatment with PKA inhibitors (Fig. 2). These data indicate that an activation of PKA by PTH is an essential element in initiating the suppressive effects of PTH. However, PTH does not reduce bsp mRNA expression by UMR cells after long term treatment (Fig. 7), and Western blot analyses confirm that BSP is produced (Fig. 8).
Some striking results were observed when PTH-treated cultures were analyzed in the presence of 7 mM ␤-GP to stimulate mineralization. Under these conditions, control cultures retained most of their BSP in the cell/matrix layer with minimal secretion into the culture medium (Fig. 8), thus confirming the BSP immunostaining results shown in Fig. 4G. However, UMR cultures treated with 2 nM PTH-(1-34) and then exposed to 7 mM ␤-GP failed to retain any of the synthesized BSP in their cell/matrix layers; rather, they continued to secrete all of the BSP into the culture medium. This inability to retain BSP in the matrix of PTH-treated cultures is not due to an insufficient supply of P i from the hydrolysis of ␤-GP by alkaline phosphatase (Table I).
Recovery of BSP and Apatite Deposition after Transient Activation of PKA-A transient activation of PKA by PTH can be mimicked in vitro by shortening the duration of cAMP analog exposure. UMR cultures treated with 0.2 mM cAMP analog from 4 -16 h of incubation (i.e. a 12-h exposure) exhibited a complete recovery of bsp mRNA levels (Fig. 9A) and secreted nearly control amounts of BSP protein into the culture medium during the final 24 h of an 88-h incubation (Fig. 9B). In the presence of 7 mM ␤-GP, these 12-h cAMP-treated UMR cultures exhibited a partial recovery (23% of control) in apatite content deposited during the last 24 h of incubation (Fig. 10A), and a partial recovery of BSP deposition in their extracellular matrix (Fig. 10B). Identical recovery results for both apatite content and BSP deposition in the cell/matrix layer were observed when 2 nM PTH-(1-34) was substituted in this 12-h treatment protocol (data not shown).
Increasingly longer recovery times resulted in progressively greater amounts of apatite content and BSP deposition in the extracellular matrix of these osteoblastic cultures. For example, UMR cultures treated with 0.2 mM cAMP analog from 4 -16 h of incubation and allowed to recover an additional 24 h beyond the 88-h time point deposited 78% of the control culture's apatite content (9.4 Ϯ 1.2 versus 12.0 Ϯ 0.8 nmol of AR-S/g of DNA, respectively). Thus, the negative effects of a transient PKA activation on the biomineralization process in UMR cultures are reversible, requiring several days for a nearly complete recovery. In addition, a recovery of BSP accumulation in the extracellular matrix of UMR cultures occurs in proportion to the extent of apatite deposition in these cultures. DISCUSSION The results indicate that PTH reversibly inhibits the biomineralization process of UMR cultures even at doses approaching physiological significance. These findings are remarkably similar to those previously reported for primary cultures of osteoblasts (22) and growth chondrocytes (24,25). Bellows et al. (22) reported that PTH-(1-84) suppressed bone nodule formation in primary osteoblast cultures derived from newborn rat calvaria. PTH's potency (ED 50 ϳ 0.05 nM), reversibility, and efficacy at a late stage in the differentiation of osteoprogenitor cells are nearly identical with those obtained for UMR cells in this study. Kato et al. (24) reported that either PTH-(1-84) or PTH-(1-34) suppressed the calcification of primary chondrocyte cultures derived from rabbit costal cartilage. Again, PTH's potency (ED 50 ϳ 1-2 nM), reversibility, and need for long term treatment to maximize the inhibitory effects are consistent with the same parameters in the current study. Thus, PTH's suppression of the biomineralization capacity of UMR cells is a natural response and not one specific to highly differentiated osteosarcoma cells in vitro.
Since the effects of PTH can be blocked by selective PKA inhibitors and mimicked by selective PKA activators, it is concluded that PTH initiates its effects on UMR biomineralization via its activation of PKA. While Kato et al. (24) did not address potential signaling mechanisms, Bellows et al. (22) reported that steady-state cAMP levels in primary osteoblast cultures were only marginally elevated at a PTH-(1-84) dose that completely inhibited bone nodule formation (1 nM). These authors raised the possibility that cAMP may not be critically involved in transducing the inhibitory effects of the hormone. However, given the asynchrony of differentiation in these cultures, only a portion of the total cell population would at any time point respond to PTH with a large increase in intracellular cAMP levels. In addition, the transient nature of cAMP as a second messenger places a temporal constraint on the ability to detect high steady-state levels of cAMP in heterogenous cell popula- tions. Thus, calvarial osteoblasts might have produced enough cAMP at lower PTH doses to activate PKA sufficiently to initiate this hormone's suppression of biomineralization. It has yet to be reported whether PTH's activation of PKA is involved in the suppression of calcification of chondrocyte cultures (24,25).
The UMR cell line consists of a homogenous population of osteoblastic cells that yields a uniform temporal response to PTH exposure in vitro. The efficacy of specific PKA inhibitors at blocking PTH's suppression of biomineralization in UMR cultures provides clear proof that PKA is at least involved in the initiation of the inhibitory response, while other signaling processes activated by PTH do not seem to initiate the suppressive effects. The use of selective agents to constitutively activate PKA in UMR cultures and a substantial reversal using simultaneous exposure to PKA inhibitors mimic the suppression of biomineralization by PTH, while activation of protein kinase C with phorbol esters or protein kinase G with cGMP does not. Altogether, these data argue that the mechanism underlying PTH's suppression of the biomineralization response of UMR cells is initiated by PKA activation. Given the similarities between UMR and primary cells in response to PTH, it is likely that PKA activation is involved also in the suppression of biomineralization by normal osteoblast and chondrocyte cultures.
An in vivo disorder known as McCune-Albright Syndrome is the result of a constitutively activated stimulatory G protein (G s ␣) within somatic cells leading to abnormally high steadystate levels of intracellular cAMP (61). The histopathological evaluation of this disease in bone shows a fibrous dysplasia appearing as ample osteoid-like matrix deposition but poorly mineralized (62). Furthermore, radiographic analysis of fibrous dysplasia patients reveals a lower than normal bone mineral content and a higher incidence of bone fractures (13,62). Thus, a constitutive hyperactivation of PKA in the bone tissues of McCune-Albright Syndrome patients correlates with a reduced bone mineral content in vivo, consistent with our observations of osteoblastic cells in vitro.
A quantitative, spatial co-localization exists between BSP and apatite deposits in the UMR model system. Kinetic analysis of the biomineralization process reveals that a substantial amount of BSP is deposited in the extracellular matrix of UMR cultures treated with ␤-GP a few hours before the first apatite deposits are detected, which subsequently appear in close proximity to the BSP deposits. These findings suggest that a prior accumulation of BSP in the extracellular matrix of UMR cultures is a diagnostic reporter of subsequent apatite deposition and may be involved in the mineralization process itself.
In direct proportion to its negative effects on apatite deposition, PTH also reversibly blocks the deposition of BSP in the extracellular matrix of UMR cultures. Cultures treated with PTH or cAMP analogs did not deposit either BSP or apatite in their extracellular matrix. Yet, when either agent is removed and the cultures are allowed to recover, they begin to deposit both BSP and apatite in their matrix in proportional amounts. Since BSP normally accumulates in the extracellular matrix of UMR cultures prior to apatite deposition, we hypothesize that PTH's transient activation of PKA early in the culture period leads to subsequent alterations in the assembly of an extracellular matrix, which does not allow BSP retention even in the presence of elevated phosphate levels.
In conclusion, low doses of PTH reversibly suppress apatite and BSP deposition in UMR cultures via an activation of PKA. Two major points can be drawn from this conclusion. First, any external signal that ultimately hyperactivates PKA in osteoblasts should manifest an impact on bone matrix production and subsequent mineralization. Second, bone matrix proteins such as BSP may need to be assembled in specific conformations in osteoid in order to promote mineralization reactions. Experiments are ongoing to determine whether this effect is more a function of a change in BSP's post-translational modifications (6,7,31) or an alteration in the structure of the surrounding extracellular matrix.