Physiological changes in extracellular calcium concentration directly control osteoblast function in the absence of calciotropic hormones

Abstract

We investigated the direct effects of changes in free ionized extracellular calcium concentrations ([Ca²⁺]_o) on osteoblast function and the involvement of the calcium-sensing receptor (CaR) in mediating these responses. CaR mRNA and protein were detected in osteoblast models, freshly isolated fetal rat calvarial cells and murine clonal osteoblastic 2T3 cells, and in freshly frozen, undecalcified preparations of human mandible and rat femur. In fetal rat calvarial cells, elevating [Ca²⁺]_o and treatment with gadolinium, a nonpermeant CaR agonist, resulted in phosphorylation of the extracellular signal-regulated kinases 1 and 2, Akt, and glycogensynthase kinase 3β, consistent with signals of cell survival and proliferation. In agreement, cell number was increased under these conditions. Expression of the osteoblast differentiation markers core binding factor α1, osteocalcin, osteopontin, and collagen I mRNAs was increased by high [Ca²⁺]_o, as was mineralized nodule formation. Alkaline phosphatase activity was maximal for [Ca²⁺]_o between 1.2 and 1.8 mM. Inhibition of CaR by NPS 89636 blocked responses to the CaR agonists. In conclusion, we show that small deviations of [Ca²⁺]_o from physiological values have a profound impact on bone cell fate, by means of the CaR and independently of systemic calciotropic peptides.

Fluctuations in extracellular free ionized calcium concentrations ([Ca²⁺]_o) accompany bone remodeling and contribute to systemic free ionized calcium (Ca²⁺_o) homeostasis (1). In tissue culture models, elevations in [Ca²⁺]_o increase osteoblast chemotaxis and proliferation (2, 3) and alter the levels of expression of some differentiation markers (4, 5). During mineralization, decreases in [Ca²⁺]_o are also likely to occur (6), but the effect of lowering [Ca²⁺]_o in bone cells has not been extensively addressed.

The mechanism of [Ca²⁺]_o-sensing by osteoblasts is unclear. The parathyroid extracellular calcium-sensing receptor (CaR) is a key player in the maintenance of a constant systemic [Ca²⁺]_o, predominantly through regulation of parathyroid hormone (PTH) secretion and urinary calcium excretion (7, 8). CaR is also present in osteoblasts (9 and references therein), where a functional role is currently debated. Recently two studies have shown that CaR-deficient mice exhibit an essentially normal skeletal phenotype when the hyperparathyroidism resulting from the lack of the parathyroid CaR is prevented (10, 11). Thus, it remains unclear whether the osteoblast CaR is a true regulator of bone function or whether its expression is vestigial (12).

In this study, we investigated the effects of both decreasing and increasing [Ca²⁺]_o on osteoblast proliferation and intracellular signaling events, the expression of several osteoblast differentiation markers [core binding factor α1 (Cbfa1, also termed Runx2 and Osf2), osteocalcin (OC), osteopontin (OP), and type I collagen (collaI)], the activity of alkaline phosphatase (AlP), and mineralized nodule formation in the absence of systemic calciotropic factors, namely PTH and vitamin D. We further investigated the role played by the CaR in these events using an alternative, nonpermeant CaR agonist, gadolinium (Gd³⁺) and a CaR inhibitor, NPS 89636 (a “calcilytic”). We used well characterized osteo-blast models, freshly isolated fetal rat calvarial cells (FRC) (13) and the clonal murine osteoblast cell line, 2T3 cells (14). The expression of CaR in freshly frozen sections of rat and human bone was also determined.

Materials and Methods

Animals. Sprague-Dawley rats (Charles River Breeding Laboratories) were killed by cervical dislocation and used in accordance to the U.K. Animals Scientific Procedures Act of 1986.

Cell Culture. FRC cells were isolated and cultured as described (15). The treatment media contained increasing concentrations of Ca²⁺_o (CaCl₂, 0.5-3 mM) or Gd³⁺ (GdCl₃, 10-100 μM; both from Sigma-Aldrich) with or without NPS 89636 (NPS Pharmaceuticals, Toronto). Treatments were initiated 24 h after seeding and lasted for the duration of the experiments, in the presence of 100 μg/ml ascorbic acid and 3 mM β-glycerophosphate (both from Sigma) from confluence.

Mouse osteoblast 2T3 cells (a kind gift from Stephen Harris, University of Missouri, Kansas City) have been characterized (14). 2T3 cells stably transfected with a 2.3-kb portion of the collaI promoter driving GFP expression (also a gift from Stephen Harris) were cultured as described (15).

RNA Extraction. Total RNA from FRC cells was extracted and shown to be pure and intact as described (16).

Expression of CaR Transcripts in FRC Cells. Random hexamer-primed first-strand cDNAs were prepared from 500 ng of FRC, 2T3, and rat kidney total RNAs and reverse-transcribed by using avian myeloblastosis virus-reverse transcription enzyme (Roche Molecular Biochemicals). Genomic DNA contamination was removed by using a RQ1 DNase kit (Promega). PCR amplification was performed by using oligonucleotide primers designed to amplify a 480-bp product from the mouse 2T3 cells (sense, 5′-AGAAGTTCCGAGAGGAAGCC-3′; antisense, 5′-ACCTGTTGCCGCCTTCTTCG-3′) and two different sets of primers to amplify 383-bp (sense, 5′-ACCTTTACCTGTCCCCTGAA-3′; antisense, 5′-GGGCAACAAAACTCAAGGTG-3′) and 545-bp (sense, 5′-ACCTGCTTACCCGGAAGAGG-3′; antisense, 5′-GTGAGAGCGATTCCAAAGGG-3′) fragments in the extracellular region of the CaR, designed to span two and four introns, respectively, in the FRC cells. The products were sequenced with an ABI Prism BigDye terminator sequencing-ready kit (Applied Biosystems), by using Amplitaq DNA polymerase and an ABI 373 DNA Sequencer.

Northern Blotting. Northern blotting of total RNA (15 μg) was performed at high stringency as described (16). The following probes used in this study were produced by RT-PCR and were rat OC, mouse collaI, mouse Cbfa1, human OP, and the housekeeping gene rat β-actin (17, 18).

CaR Immunoblotting and Immunofluorescence. Crude membrane fractions were isolated from FRC cells, 2T3 cells, and rat kidney as described (19), and SDS/PAGE immunoblotting was performed as described (20). For immunofluorescence, the cells were grown on glass coverslips and stained as described (19).

CaR Immunocytochemistry. The rat femurs and kidneys were aseptically dissected and snap frozen for 60 sec at -170°C in isopentane. The samples of human mandibles were neck resections from patients with squamous cell carcinoma invading bone and were taken after ethical approval and signed informed consent.

Frozen samples were embedded and sectioned as described (21). The sections were fixed in 10% buffered formalin (Sigma-Aldrich; 3-5 min), blocked with Seablock (Perbio Science, Tattenhall, U.K.; 30 min), and incubated with affinity-purified rabbit anti-CaR polyclonal antibodies (1:100 in 0.2% BSA in PBS) or primary antibodies preabsorbed with an excess of antigenic peptide. Secondary antibodies [FITC-conjugated swine anti-rabbit immunoglobulins 1:50 in PBS (Molecular Probes)] were applied for 30 min. Between all steps, the sections were washed in PBS only or 0.2% BSA in PBS. They were mounted in an antifade mounting medium (Perbio Science).

For EDTA-decalcified paraffin wax histology, the femurs were prepared, and the immunoperoxidase experiments were performed as described (22) by using affinity-purified CaR polyclonal primary antibodies (1:100) alongside a peptide block negative control as above. The sections were counterstained in hematoxylin. All experiments were repeated on sections from at least three tissue samples. The images were prepared by using a Zeiss Axioplan microscope with a manual photographic camera and Agfa RS50 color film or acquired digitally by using a Zeiss Axiocam system. Confocal microscopy was performed as described (19) by using undecalcified sections (30 μm) permeabilized in 0.075% saponin in PBS.

Intracellular Signaling in FRC and 2T3 Cells. The experiments were performed as described (19) with minor modifications. Briefly, the cells were equilibrated in the experimental buffer (20 min, 0.5 mM Ca²⁺), treated for 10 min (or 5-60 min for the time course experiments) in experimental buffer alone or with various CaR agonists (5 mM Ca²⁺_o or various concentrations of the nonpermeant CaR agonist Gd³⁺) and lysed. The effects of antagonists were ascertained by pretreating the cells before the stimulation step (5 min for mitogen-activated protein kinase kinase inhibitor PD98059 and phosphatidylinositol 3-kinase inhibitors wortmannin and LY294002, 3 min for CaR inhibitor NPS 89636) followed by the cotreatment with the agonist.

Relative changes in phosphorylation of extracellular signal-regulated kinases 1 and 2 (ERK1/2) mitogen-activated protein kinases, Akt, and GSK3β were measured by semiquantitative Western blotting by using affinity-purified antibodies raised against the specific dual phosphopeptide sequences of ERK1/2 (Promega) and antibodies raised against the Akt and GSK3β phosphopeptide sequences (Cell Signaling, Beverly, MA). As a control, a non-phospho-specific antibody recognizing ERK1/2 mitogen-activated protein kinase was used to assess total ERK levels.

Proliferation Assay. The cells were plated in 24-well plates (5,000 cells per cm²) and, after treatment, trypsinized and counted three times with a Coulter Counter (Beckman Coulter). In addition they were assayed in 96-well plates for proliferating activity by the BrdUrd incorporation assay according to the manufacturer's instructions (Roche Diagnostics).

collaI Promoter Studies. 2T3 cells stably transfected with the collaI promoter driving GFP expression were treated for 24 h with increasing [Ca²⁺]_o and [Gd³⁺]. The GFP levels were assessed in 0.05% Triton X-100 freeze-thawed lysates by measuring fluorescence emission levels at excitation of 488 nm by using a 520-nm long-pass emission filter (Wallac Victor² 1420 multilabel counter, Perkin-Elmer).

AlP Activity. The AlP activity was assayed by colorimetry as described (23).

Von Kossa Staining. Von Kossa staining was performed as described (24). The images of mineralized nodules were captured on a flatbed scanner and image analysis software (scion, Frederick, MD) was used to count the number of and total area of the mineralized nodules.

Statistical Analysis. The statistical significance was assessed by one-way ANOVA with Tukey's post hoc test or Student's t test. Observations were considered to be significantly different for P < 0.05.

Results and Discussion

Bone formation is characterized by a distinctive sequence of events beginning with the commitment of mesenchymal cells to osteoblast lineage, followed by osteoblast proliferation, growth, and differentiation, culminating in the formation of a mineralized extracellular matrix by terminally differentiated osteoblasts (25). Fluctuations in systemic [Ca²⁺]_o may influence these processes by inducing changes in circulating levels of PTH and 1,25(OH)₂-vitamin D₃. However, it is currently unclear whether changes in [Ca²⁺]_o can directly affect bone cell function independently of their effects on circulating calciotropic factors. The aim of this study was to investigate the direct effects of acute (minutes), medium-term (hours), and chronic (days to weeks) changes in [Ca²⁺]_o on osteoblast models of freshly isolated FRC cells and the clonal osteoblast cell line 2T3 and, furthermore, to investigate the involvement of the CaR in these responses.

CaR Is Expressed in Osteoblasts in Vitro and in Vivo. The expression of CaR transcripts (Fig. 1A) and protein (Fig. 1 B and C) in rat (FRC) and murine osteoblast clonal (2T3) cell lines was confirmed by RT-PCR using intron-spanning primers and by immunoblotting, respectively. The PCR products were sequenced and shown to be identical to the rat and mouse kidney CaRs, respectively.

Fig. 1.

CaR expression in osteoblastic cells. (A) Two fragments of the extracellular domain of the CaR were amplified by PCR from the FRC cells by using specific, intron-spanning primers. A 480-bp region of the mouse CaR was also amplified from the 2T3 cells. The fragments were sequenced and shown to be homologous to the rat and mouse sequences. (B) CaR immunoblot analysis of 25 μg of crude membrane homogenates from FRC and 2T3 cells showed the expected immunoreactive species for high-mannose monomeric (150-160 kDa) and nonglycosylated monomeric (120 kDa) forms of the receptor (rat kidney, positive control). The bands were ablated when the primary antibody was preabsorbed with the immunizing peptide. (C) High-power immunofluorescence photomicrograph of FRC and 2T3 cells showing peptide-ablatable CaR immunostaining (×400).

We confirmed the presence of CaR-specific (i.e., peptide-protectable) immunofluorescence in the cells of osteoblastic lineage in decalcified rat femur preparations (Fig. 2A). We also provide evidence for CaR expression in freshly frozen undecalcified rat (Fig. 2 B and C) and human bone (Fig. 2D). Osteoblast and osteocyte staining is evident in both decalcified (Fig. 2A) and undecalcified freshly frozen (Fig. 2 B and C) rat femur preparations. At higher magnifications we observed differential immunoreactivity in the osteocyte populations, possibly related to various maturation stages of the osteocytes (Fig. 2C). Confocal microscopy confirmed a plasma membrane distribution (Fig. 2C Inset). This distribution pattern was confirmed in the human mandible (Fig. 2D).

Fig. 2.

CaR expression in rat femur and human mandible. (A) Bright-field photomicrograph of CaR immunoperoxidase staining in EDTA-decalcified paraffin sections of rat femur show both osteoblast (arrows) and osteocyte (arrowheads) CaR immunoperoxidase staining. (B-D) Photomicrographs of CaR immunofluorescence in cryosections of rat femur show that osteoblasts (B, arrows) and osteocytes (C, arrows) are positive for CaR immunofluorescence. A subpopulation of osteocytes displays canaliculi staining (C, arrowhead). Confocal microscopy reveals a predominantly plasma membrane expression of the CaR (C Inset). (D) Photomicrograph of CaR immunofluorescence in cryosections of human mandible. The CaR immunofluorescence is peptide blockable. (D Inset) Arrow shows the osteocyte staining. (Bars, 20 μm.)

Osteoblast Models Respond to [Ca²⁺]_o by Intracellular Signaling Events. Increasing [Ca²⁺]_o stimulates proliferation and migration of osteoblasts through the mitogen-activated protein kinase/ERK pathway (26, 27). In our models, both high Ca²⁺_o and Gd³⁺ induced ERK1/2 phosphorylation in FRC cells (Fig. 3A) and 2T3 cells (not shown). Elevated [Ca²⁺]_o or Gd³⁺ treatment stimulated the serine/threonine kinase Akt, as evident from the phosphorylation of its activation sites (Thr-308 and Ser-473), as well as phosphorylation of GSK3β on its inhibitory site (Ser-9; Fig. 3A). These signals are associated with proliferative (ERK) and antiapoptotic (Akt, GSK3β) responses. Gd³⁺-induced ERK1/2 activation was both time-(Fig. 3Bi) and concentration-(Fig. 3Bii) dependent, was abolished by inhibition of mitogen-activated protein kinase kinase 1 (with PD98059), and was in part attenuated by inhibition of phosphatidylinositol 3-kinase (with two structurally unrelated compounds, wortmannin and LY294002; Fig. 3C). The Ca²⁺_o and Gd³⁺ responses were significantly reduced by the CaR antagonist NPS 89636 (Fig. 3D).

Fig. 3.

Intracellular signaling in FRC and 2T3 cells. (A) Western blot analysis for ERK phosphorylation/activation was performed by using lysates of cells previously treated with 5 mM Ca²⁺_o or 50 μMGd³⁺ for 5 min in FRC cells. The treatments did not affect total ERK content. The same treatments elicited Akt phosphorylation at both threonine-308 and serine-473 and GSK3β phosphorylation at serine-9. (B) In FRC cells, Gd³⁺-induced ERK activation was time- and concentration-dependent (i and ii, respectively). (C) Gd³⁺-induced ERK phosphorylation was prevented by the mitogen-activated protein kinase inhibitor PD98059 (PD) and was partially inhibited by the phosphatidylinositol 3-kinase kinase inhibitors wortmannin (W) and LY294002 (LY). (D) Both Ca²⁺_o- and Gd³⁺-induced ERK phosphorylation (i and ii, respectively) are inhibited by the CaR inhibitor NPS 89636. The significant differences are marked by asterisks (^*, P < 0.05; ^**, P < 0.01).

FRC Cell Proliferation Is [Ca²⁺]_o-Dependent. The proliferative intracellular signaling in FRC cells in response to high Ca²⁺_o or Gd³⁺ was associated with a concentration- and time-dependent mitogenic response (Fig. 4). Elevations in [Ca²⁺]_o from 1.2 to 1.8 or 2.5 mM were sufficient to elicit proliferation after chronic treatments of at least 7 days. Lowering the [Ca²⁺]_o to 0.5 mM decreased FRC cell proliferation (Fig. 4A Upper). The reduced cell number resulted in part from increased cell death (approximately 9 times more trypan blue-positive cells at day 10, not shown). Conversely, high [Ca²⁺]_o treatments did not affect cell death (not shown). Chronic treatment with 25 μM Gd³⁺ induced FRC cell proliferation after 14 days of treatment, whereas the 50 μM dose increased proliferation after 5 days of treatment (Fig. 4B Upper). In addition, after 10 days of treatment, BrdUrd incorporation in FRC cells was greater in the presence of 2.5 mM Ca²⁺_o (+17 ± 1.8%) or 50 μM Gd³⁺ (+17 ± 1%) in comparison to 1.2 mM Ca²⁺_o control (n = 6; P < 0.05 for both). The increases in FRC proliferation in response to both 2.5 mM Ca²⁺_o and 50 μM Gd³⁺ were reduced by the CaR antagonist NPS 89636 as observed by cell counting (Fig. 4 Lower) and BrdUrd incorporation assay (decrease of Ca²⁺_o- and Gd³⁺-stimulated proliferation by -7 ± 0.8% and -11 ± 0.7% respectively; n = 6; P < 0.05 and 0.01, respectively) after 10 days of treatment.

Fig. 4.

Effects of Ca²⁺_o (A) and Gd³⁺ (B) on the FRC cell proliferation. (A Upper) FRC cells were grown in increasing concentrations of Ca²⁺_o (0.5, 1.2, 1.8, and 2.5 mM) in triplicate wells and counted three times at different time points. After 1, 3, and 5 days, there were no significant differences between the treatments. Both the 1.8 and the 2.5 mM treatment increased proliferation after 7 (P < 0.05) and 10 (P < 0.05) days. The 0.5 mM Ca²⁺_o treatment decreased proliferation after 7 (P < 0.05) and 10 (P < 0.001) days. The Ca²⁺_o-induced proliferation was inhibited by the CaR antagonist NPS 89636 after 10 days of treatment (n = 3; P < 0.05; Lower). (B Upper) FRC cells chronically treated with 25 μMGd³⁺ increased proliferation after 14 days of treatment (P < 0.05). The 50 μM dose significantly increased proliferation after 5 (P < 0.05), 7 (P < 0.05), 10 (P < 0.01), and 14 (P < 0.01) days of treatment. This response is inhibited by the CaR antagonist NPS 89636 after 10 days of treatment (n = 3; P < 0.05; Lower).

Osteoblast Gene Expression in FRC Cells Is [Ca²⁺]_o-Dependent. Small elevations of [Ca²⁺]_o promote differentiation of FRC cells, as evident from the up-regulation in the mRNA expression levels of osteoblast markers at early and late stages of differentiation (Fig. 5). Doubling of the normal [Ca²⁺]_o to 2.5 mM increased the promoter activity (Fig. 5A) as well as mRNA expression levels (by 10%; Fig. 5B) of collaI, the major structural component of the organic matrix. This increase was maximal 24 h after the Ca²⁺_o treatment and subsided at later time points (not shown), consistent with an early role of collaI in the osteoblast maturation sequence.

Fig. 5.

Effects of Ca²⁺_o and Gd³⁺ on osteoblast promoter activity and gene expression. (A) 2T3 cells stably expressing the collaI promoter GFP constructs were grown in increasing concentrations of Ca²⁺_o (Left) and Gd³⁺ (Right, baseline Ca²⁺_o is 1.2 mM) for 24 h (n = 4). The significant differences with regard to the 1.2 mM Ca²⁺_o control are marked by asterisks (^*, P < 0.05; ^**, P < 0.01; ^***, P < 0.001). (B and C) Total RNA was extracted from the cells grown in 0.5, 1.2, 1.8, and 2.5 mM Ca²⁺_o and 1.2 mM and 50 μM Gd³⁺ at different time points (days 1, 3, 7, 12, and 18) and subjected to Northern analysis with the appropriate cDNA probes for collaI, Cbfa1, OC, OP, and β-actin (control).

A notable finding is the up-regulation of mRNA levels for Cbfa1, the master osteoblast transcription factor responsible for osteoblast differentiation and mineralization (ref. 28 and Fig. 5C). Transcription of OC and OP, the major osteoblast proteins downstream of Cbfa1, were also induced by increasing [Ca²⁺]_o. This effect was differentiation-stage specific, beginning after 7 days and at its most marked after 12 days of treatment, which coincides with the commencement of nodule formation. This is evident for both the 1.8 mM treatment (+87% for Cbfa1, +146% for OC, and +115% for OP) and the 2.5 mM treatment (+75% for Cbfa1, +168% for OC, and +82% for OP; Fig. 5C). This effect was sustained during the formation of mineralized nodules (day 18) but subsided during mineral accruement (day 23, data not shown). Gd³⁺ mimicked the effects of high [Ca²⁺]_o on Cbfa1, OC, and OP (+82% for Cbfa1 and +145% for OC after 12 days and +115% for OP after 7 days; Fig. 5C), but not collaI mRNA expression levels. Treatments with various [Ca²⁺]_o or [Gd³⁺] did not alter the mRNA expression of bone morphogenetic protein 2 (data not shown) or of the housekeeping gene β-actin (Fig. 5C).

AIP Activity in FRC Cells Is [Ca²⁺]_o-Dependent. AlP has important functions during crystal formation (29). We found that its optimal enzymatic activity depends on maintaining [Ca²⁺]_o within a narrow (1.2-1.8 mM) range (Fig. 6A), and that small deviations inhibit its activity. A recent study identified a Ca²⁺ binding site in the tissue nonspecific AlP protein (the predominantly expressed isoform in bone), the mutation of which results in a severe phenotype, indicating that calcium is fundamental to the AlP activity (29). Our experiments support these observations and raise the question of the interplay between AlP function and Ca²⁺_o sequestration during mineral formation. The effects of high Ca²⁺_o were not mimicked by Gd³⁺ (Fig. 6B), probably resulting from direct functional interaction between Ca²⁺ and AlP.

Fig. 6.

AlP activity in FRC cells treated with Ca²⁺_o (A)orGd³⁺ (B). (A) FRC cells were grown in increasing concentrations of [Ca²⁺]_o and assayed for AlP activity. Raising the [Ca²⁺]_o from 1.2 mM to 1.8 mM decreased the activity of AlP after 7 days of treatment (n = 6; P < 0.01). Increasing the [Ca²⁺]_o to 2.5 mM and to 3 mM decreased AlP activity further. Lowering the [Ca²⁺]_o from 1.2 mM to 0.8 mM and furthermore to 0.5 mM also decreased the activity of AlP significantly across the time points measured. (B) In the presence of 1.2 mM Ca²⁺_o, Gd³⁺ (25-100 μM) increased AlP activity in FRC cells after 2 days of treatment (n = 6; P < 0.05 for 25 and 100 μM and P < 0.001 for 50 μM). However, this effect was transient, and, after 3 days, 100 μM Gd³⁺ decreased the enzyme activity. No differences were observed after 18 days of treatment. The significant differences with regard to the control are marked by asterisks (^*, P < 0.05; ^**, P < 0.01; ^***, P < 0.001).

Elevations in [Ca²⁺]_o Increase Mineralized Nodule Formation by FRC Cells. Increasing [Ca²⁺]_o directly stimulated the production of mineralized bone nodules by FRC cells, in the absence of systemic calciotropic factors (Fig. 7A). The steepness of the sigmoidal curve that fits the relationship between mineralized nodule formation (area and number of nodules) and [Ca²⁺]_o and its EC₅₀ of ≈1.1 mM (Fig. 7A) indicates that relatively small deviations in Ca²⁺_o from the physiological range exert a significant impact on mineralization. This effect is reminiscent of the CaR-regulated PTH secretion from the parathyroid glands, where the [Ca²⁺]_o necessary to produce 50% inhibition of PTH secretion is 1.2 mM (30). The increased mineralization in the presence of 1.8 mM Ca²⁺_o compared with the 1.2 mM baseline was not simply due to the increased availability of Ca²⁺_o, because it was also observed in the presence of Gd³⁺ (Fig. 7B) and was saturated when [Ca²⁺]_o was ≥2.5 mM (data not shown). Hence, the increase in mineralized nodule formation in FRC cells is the result of an increase in both proliferation and maturation of osteoblasts in response to the elevations in [Ca²⁺]_o. The increase in mineralization in response to 2.5 mM Ca²⁺_o was reduced by the CaR antagonist NPS 89636 (number of nodules by -27 ± 7% and area of nodules by -27 ± 3.5%; n = 3, P < 0.01 for both; Fig. 7C).

Fig. 7.

Effects of extracellular Ca²⁺_o (A)orGd³⁺ (B) on the FRC cell mineralized nodule formation (von Kossa staining). (A) FRC cells grown in the presence of ascorbic acid and β-glycerophosphate for 21 days showed a [Ca²⁺]_o-dependent increase in the formation mineralized nodule number and area (n = 6; P < 0.01). The Ca²⁺_o EC₅₀ for both the number and the area occupied by the mineralized nodules was ≈1.1 mM. (B) Twenty-one-day treatment of FRC cells with 50 μMGd³⁺ (baseline Ca²⁺, 1.2 mM) significantly increased the production of mineralized nodules (n = 3; P < 0.05). (C) The Ca²⁺_o-induced mineralization is inhibited by the CaR inhibitor NPS 89636.

Inhibitory effects of high [Ca²⁺]_o on osteoblast AlP activity and release have been reported (31, 32). Furthermore, the loss of AlP activity has been correlated with the accumulation of calcium in matrix vesicles, which are involved in the process of mineralization (23). Nevertheless, the concomitant increase in mineralization with a decrease in AlP activity in the presence of elevated [Ca²⁺]_o has not been previously observed. We speculate that in hypercalcemic conditions AlP may be involved in inhibition of excessive and/or dystrophic mineralization.

Mineralization is regulated by a complex arrangement of stimulatory and inhibitory components (33, 34). We observed a concentration-dependent increase in collaI, Cbfa1, OC, and OP mRNAs and a decrease in the activity of AlP in response to relatively small elevations in [Ca²⁺]_o. We hypothesize that Ca²⁺_o regulates a variety of processes in osteoblasts, resulting in a balance between stimulatory and inhibitory factors, producing the saturable sigmoidal curve that describes the relationship between [Ca²⁺]_o and mineralization, thereby regulating crystal growth and preventing excessive mineralization.

Effects of Lowering [Ca²⁺]_o on FRC Cells. We observed profound effects of lowering [Ca²⁺]_o from 1.2 mM to 0.5 mM on osteoblasts, mostly opposite of those of elevating [Ca²⁺]_o. However, lowering the [Ca²⁺]_o from 1.2 to 0.5 mM increases OC mRNA expression (+45% and +31% at 12 and 18 days, respectively; Fig. 5C). The significance of these findings requires further investigations.

Is the CaR Involved in Ca²⁺_o-Sensing by Osteoblasts? The CaR expression in osteoblasts has been reported (9). We have confirmed these findings in decalcified sections, established the presence of CaR transcripts in freshly isolated osteoblasts and in an osteoblast-derived clonal cell line, and demonstrated the presence of the CaR in freshly frozen bone. It has been reported that the osteoblast CaR exhibits a different pharmacological profile from that of parathyroid cells, e.g., in its response to Mg²⁺, and that therefore the osteoblast and parathyroid CaRs are distinct (35). The portions of the FRC and 2T3 CaR amplified, sequenced, and confirmed to be homologous to the kidney CaR in this study constitute the Ca²⁺ binding site (36). Although sequence differences could arise in the transmembrane region of the osteoblast CaR, which also binds triand polyvalent cations, the functional discrepancies between the parathyroid and the osteoblast CaRs cannot be explained simply on the basis of putative structural differences. G protein-coupled receptors exhibit “conditional efficacy,” the ability to change their behavior toward the host cell in a way that ligand structure manipulates both the extent and the quality of efficacy of an agonist (37). This ability could account for the observed pharmacological differences of the CaR in bone and in other cell types.

Our most striking finding is the narrow range of osteoblast sensitivity to [Ca²⁺]_o. During bone resorption, in the sealed compartment between the osteoclast resorbing surface and the mineral surface, up to 40 mM Ca²⁺_o can accumulate (38). Nevertheless, a resorbing osteoclast releases <2 mM Ca²⁺ from the nonresorbing surface, a concentration that plateaus after 3 h of activity (39). These findings indicate that the [Ca²⁺]_o that osteoblasts are exposed to in vivo are likely to be an order of magnitude smaller than the commonly thought 8-40 mM. This range of sensitivity in osteoblasts overlaps with the activation range for the parathyroid CaR and the one observed here.

To dissociate the extracellular and intracellular effects of elevating the [Ca²⁺]_o on osteoblast function, we confirmed that an alternative CaR agonist, Gd³⁺, mimics, whereas the CaR antagonist, NPS 89636, blocks most of the effects of increasing [Ca²⁺]_o. Non-CaR osteoblast mechanisms have been suggested (35), for instance, certain metabotropic glutamate receptors that are present in bone (40) and can be activated both by Ca²⁺_o and by Gd³⁺ (41). Nevertheless, the Ca²⁺_o- and Gd³⁺-sensitive subtype metabotropic glutamate receptor 1α is not expressed in osteoblasts (42). In addition, the concentrations of Ca²⁺_o and Gd³⁺ required to evoke metabotropic glutamate receptor 1α responses are considerably higher than those used here (41). Taking these results together, the simplest interpretation of the current data is that osteoblasts express a functional CaR that plays a fundamental role in various aspects of their function.

In conclusion, we show that osteoblasts sense and respond to Ca²⁺_o independently of systemic calciotropic factors in a time- and concentration-dependent manner. The local fluctuations in [Ca²⁺]_o may therefore regulate osteoblast activity during both Ca²⁺ load buffering by bone, where the cells are exposed to large increases in [Ca²⁺]_o, and remodeling, when the local elevations in [Ca²⁺]_o are more moderate. These findings, together with the demonstration that the effects of Ca²⁺_o are mimicked by Gd³⁺ and are blocked by a calcilytic compound, indicate that Ca²⁺_o, most likely acting through the CaR, is a key regulator of osteoblast cell fate.