Bone remodeling is crucial for the maintenance of skeletal integrity and calcium homeostasis and is regulated by systemic hormones and local signaling factors. During bone resorption, Ca2+ is released into the remodeling microenvironment and may act as a paracrine signal to produce anabolic effects on osteoblasts. These include chemotaxis, proliferation, differentiation, matrix synthesis, and mineralization (reviewed in Dvorak and Riccardi).1 Characterization of the molecular mechanisms for sensing the changes in extracellular Ca2+ ([Ca2+]e) by osteoblasts is essential for understanding bone remodeling and may represent a novel target for treatment of bone diseases, such as osteoporosis and osteomalacia.
Parathyroid and kidney cells employ the extracellular Ca2+-sensing receptor (CaR), a G protein-coupled receptor, to control parathyroid hormone (PTH) secretion, urinary Ca2+ excretion, and thereby, systemic Ca2+ homeostasis.2 We and others have used multiple approaches to show CaR expression in bone sections and activation of a Ca2+-sensing mechanism in cells of the osteoblastic lineage in vitro.3–8 These studies suggest that the changes in the [Ca2+]e affect various steps in the remodeling cycle via the CaR, including osteoblast proliferation, differentiation, and mineralization.3, 7, 8 Although the CaR has long been a candidate for mediating Ca2+-sensing by osteoblasts, its role in vivo has been controversial because of the lack of definitive animal models.9–11
Defining the function of the skeletal CaR in vivo using global CaR knockouts has proven difficult because these animals manifest severe hypercalcemia and hyperparathyroidism,12, 13 disturbances that have secondary effects on bone and other tissues. In addition, a CaR splice variant is expressed by several cell types in these animals,14, 15 suggesting that the knockout may be incomplete. To definitively address the role of CaR signaling in skeletal tissues, we generated mice in which loxP sites flank exon 7 of the CaR gene (Flox-CaR+/+), thereby precluding synthesis of the transmembrane and signaling domains of the receptor.16 Using cre-lox technology, we generated tissue-specific CaR null animal models (as outlined in Fig. 1A) and showed that parathyroid-, osteoblast-, and chondrocyte-specific knockouts of the receptor lead to defective skeletal development.16 These studies lend initial support to the role of the CaR in mediating important functions of osteoblasts in vivo. The current study was designed to further assess the roles of the osteoblast CaR, specifically in regulation of bone remodeling and control of mineralization, and examine the cellular and molecular mechanisms responsible for these processes.
We knocked out the CaR across a broad population of osteoblasts by mating Flox-CaR+/+ mice with transgenic mice expressing Cre recombinase under control of the 3.6-kb fragment of the rat α1(I) collagen promoter (Col 3.6-cre), which is expressed throughout cells of the osteoblastic lineage.17 Col 3.6-cre+/− // Flox-CaR+/+ CaR conditional osteoblast knockouts (CaR obKOs) are, therefore, an appropriate model for study of bone-specific CaR physiology. We used these mice to address the hypothesis that the osteoblast CaR plays a pivotal role in the control of bone mineralization, as well as remodeling via the modulation of osteoclastogenesis. We describe an in-depth analysis of skeletal phenotype development in the absence of the osteoblast CaR and, furthermore, elucidate some of the cellular and molecular mechanisms in vitro that may be responsible for observed changes.
Materials and Methods
Generation of conditional CaR knockout in osteoblasts
Inactivation of the CaR in osteoblastic cell populations was achieved by mating Flox-CaR+/+ mice to transgenic mice expressing Cre-recombinase, under control of the 3.6-kb fragment of the rat a1(I) collagen promoter (Col 3.6-cre mice; gift from Dr. Barbara Kream, University of Connecticut, Farmington, CT, USA), according to the breeding strategy detailed in Figure 1B. The genotypes of the mice were determined by PCR (Fig. 1A, B). Littermates of both sexes were weighed every 1 to 2 days up to 21 days and euthanized by isofluorane for experiments. All protocols were approved by the Animal Care Committee of the San Francisco Department of Veterans Affairs Medical Center. The mouse backgrounds were mixed: SVJ129/C57 Black (Floxed CaR mice) and CD1 (3.6 Col1-cre mice).
Detection of the CaR by immunofluorescence
Undecalcified humeri of 21-day-old mice were isolated, fixed, frozen-embedded, and sectioned at 4 µm using a Leica Jung CM1800 Cryostat (Leica, Inc., Nussloch, Germany) equipped with CryoJane Frozen Sectioning Kit (Instrumedics, Inc., Hackensack, NJ, USA), as described previously.18 After incubation in PBS (15 minutes), sections were briefly decalcified in 10% EDTA (2 minutes), incubated in 1% hyaluronidase (10 minutes, 37°C), then in blocking buffer (4% goat serum, 0.2% bovine serum albumin in PBS; 30 minutes). Rabbit primary polyclonal CaR antibodies (321113B),19 raised against an intracellular epitope of the bovine parathyroid CaR, were applied to sections overnight (4°C, 6.3 ng/mL in blocking buffer). For negative controls, primary antibodies were omitted. After a series of washes (PBS, 2 × 5 minutes; blocking buffer, 2 × 5 minutes), secondary goat anti-rabbit antibodies conjugated to fluorescein isothiocyanate (FITC) were applied for 30 minutes (1:50; Invitrogen, Carlsbad, CA, USA). Sections were washed (blocking buffer, 2 × 5 minutes; PBS, 2 × 5 minutes), mounted (ProLong Gold antifade reagent with 4,6-diamidino-2-phenylindole [DAPI]; Invitrogen), and visualized (Carl Zeiss microscope, Carl Zeiss MicroImaging, LLC, Thornwood, NY, USA). Unless stated otherwise, all steps were performed at room temperature, and all reagents were from Sigma (St. Louis, MO, USA).
Sera were collected from 21-day-old mice and assayed for intact PTH (Mouse Intact PTH ELISA; Immutopics, San Clemente, CA, USA), total calcium, phosphorus, magnesium, sodium, chloride, creatinine, total protein, albumin, alanine aminotransferase, aspartate aminotransferase, and alkaline phosphatase (Antech Diagnostics, Morrisville, NC, USA).
Analysis of skeletal morphology
Distal femurs were isolated from 1-, 7-, 14-, and 21-day-old mice, fixed in 10% phosphate-buffered formalin, dehydrated, defatted, and embedded in plastic (methyl methacrylate [MMA]). For general histological evaluation, 4-µm sections were stained according to the Goldner Trichrome protocol (1% picric acid, 4°C, overnight; 1% fuchsine solution, 15 minutes; 6% phosphomolybdic acid, 5 minutes; 4% light green solution, 15 minutes; with washes in tap water between steps). For osteoblast quantification, 4-µm plastic sections of 7-day-old femurs were mounted on gelatinized slides, deplasticized, stained according to the von Kossa method, and counterstained with tetrachrome (Polyscience, Warrington, PA, USA).20 Osteoblasts lining trabeculae of the secondary spongiosa were counted and normalized for total trabecular bone perimeter in bone sections from 7-day-old mice. Osteoclasts were identified as tartrate-resistant acid phosphatase (TRACP)-positive cells, after TRACP activation by naphthol-1-phosphate sodium and counterstaining with fast violet, in sections of bone from 21-day-old mice. Only cells in direct contact with a resorption pit (indentation in the mineralized matrix) were quantified. Erosion surfaces below osteoclasts and total trabecular bone perimeter were delineated and quantified. For visualization of the mineralization front, 19-day-old mice were injected with calcein (15 mg/kg) subcutaneously, 48 hours prior to euthanasia. Tibias were isolated, embedded in plastic as above, and sectioned transversely (10 µm). Fluorochrome-labeled surfaces were photographed under UV illumination. Images were captured (Carl Zeiss microscope, Carl Zeiss MicroImaging, LLC) and analyzed (Bioquant, Nashville, TN, USA) in a manner blinded to the genotype.
Measurement of bone quality and quantity by microcomputed tomography (µCT)
Femurs and L4 vertebrae for in vitro µCT analysis were isolated, fixed, and maintained in 70% ethanol. For trabecular bone, 160 serial cross-sectional scans (1.05 mm) of the secondary spongiosa of the left distal femoral metaphysis were obtained from the end of the growth plate extending proximally to the shaft. The isotropic voxel (volumic pixel) size was 10.5 µm and X-ray energy 55 kV. For cortical bone, 41 serial cross sections (0.21 mm) of the femoral midshaft were scanned. The voxel had an isotropic size of 21 µm and X-ray energy 55 kV. For analysis of µCT images, a global threshold was applied to segment mineralized from soft tissue (marrow). The threshold of 22% and 35% of the gray scale was set for analysis of trabecular and cortical bone, respectively. These values were previously determined to detect only mineralized tissue in control animals. Linear attenuation was calibrated using hydroxyapatite. The µCT measurements were made using a SCANCO VivaCT 40 scanner (SCANCO Medical AG, Bassersdorf, Switzerland). Image analysis was performed using software provided by SCANCO.
Osteoblast culture and osteoblast-osteoclast coculture
Calvarial osteoblasts were isolated from 7- to 9-day-old littermates by sequential trypsin/collagenase digestions and cultured as previously described.21 For mineralization assays, medium containing 1.8 mM Ca was supplemented with ascorbic acid (284 µM) and β-glycerophosphate (3 mM) from time of confluence. Mineralized nodules were visualized in cells that were 21 days postconfluence by von Kossa staining and analyzed as described previously.22 The area of mineralized nodules was normalized to the total number of cells, estimated by incorporation of a nuclear stain, crystal violet. After staining of the cultures, crystal violet was eluted and quantified by spectrophotometry at 595 nm. Confluent calvarial osteoblasts were also cocultured with nonadherent spleen cells on 18-mm coverslips (3 coverslips/group) for 9 days, and the cultures stained with TRACP as described.23 Using the Zeiss AxioVision 4.6 software together with the Zeiss microscope (Carl Zeiss MicroImaging, LLC) equipped with a motorized stage, images of each coverslip were subdivided into 156 fields and TRACP-positive (mononucleated and multinucleated) cells were counted in every second field (78 fields total) at 10× magnification. Each experiment was repeated three times.
Assessment of genotype and efficiency of recombination in calvarial osteoblasts by PCR
DNA was isolated from confluent calvarial osteoblasts and purified using the DNeasy Blood & Tissue Kit (Qiagen, Valencia, CA, USA). Genotyping and efficiency of recombination were assessed by PCR, using the following primers (Fig. 1A): lox7 up (5′-GTG ACG GAA AAC ATA CTG C-3′), low (5′-CGA GTA CAG GCT TTG ATG C-3′), and MCRI6 (5′-CCT CGA ACA TGA ACA ACT TAA TTC GG-3′).
Determination of gene expression by quantitative real-time RT-PCR
RNA was isolated from humeral cortices (marrow flushed out) of 1-, 7-, and 21-day-old littermates, as well as confluent cultured calvarial osteoblasts. Levels of mRNA encoding osteoblast differentiation markers were determined as previously described.24 Primers and probes for receptor activator of NF-κB ligand (RANKL) and osteoprotegerin (OPG) were previously described,23 and the remaining primers and probes are shown in Table 1.
Table 1. Primers and Probes for qRT-PCR
Taqman primers/probes; Applied Biosystems (Foster City, CA, USA).
Custom-made primers and probes; Integrated DNA Technologies (Skokie, IL, USA).
SYBR Green primers; Elim Biopharmaceuticals (Hayward, CA, USA).
Data are represented as mean ± SEM. All experiments were performed at least three times. Data from two groups were compared using unpaired Student's t test. Data from three or more groups were compared using ANOVA and Tukey post hoc analysis. Significance was assigned for p < 0.05.
Generation of CaR obKO mice
Flox-CaR+/+ mice were generated as outlined in Figure 1A and described previously.16 The breeding strategy detailed in Figure 1B provided homozygous conditional knockouts, CaR obKOs (Col 3.6-Cre+/−//Flox-CaR+/+), heterozygotes (Col 3.6-Cre+/−//Flox-CaR+/−), and flox-CaR controls (Col 3.6-Cre−/−//Flox-CaR+/+). Previous studies confirmed successful excision of exon 7 by Cre recombinase and that the truncated protein lacking exon 7 does not signal nor interfere with signaling of the full-length CaR.16 The current study confirmed successful recombination in CaR obKO mice using tail DNA (Fig. 1B). Both male and female CaR obKO mice and their heterozygous and control littermates were used for the experiments.
Cre expression and activity in Col 3.6-Cre+/− transgenic mice are primarily detectable in calvarial and long bone preosteoblasts, osteoblasts, and osteocytes, and, to a much lesser degree, in kidney, liver, skin, testes, and ovaries.17 We confirmed the skeletal predominance of this expression pattern in the conditional knockouts (Fig. 1D) and noted low levels of expression of Cre in the lungs, heart, and kidneys. Necropsy of 5-day-old knockouts revealed immaturity of the lungs and kidneys (not shown), possibly resulting from the knockout of the CaR in the lung and renal fibroblasts. Nevertheless, there was no evidence of renal failure or substantial renal dysfunction in these mice by blood chemistries (Table 2). CaR obKO mice exhibited significantly reduced expression (by ∼70%) of CaR transcripts in their humeral cortices, whereas heterozygotes expressed ∼50% of the CaR transcripts compared with control mice (Fig. 1E). Residual expression of the CaR mRNA in bones from CaR obKOs may be attributed to the presence of nonosteoblastic cells in the tissue preparations that did not express Col 3.6-Cre, necessary for gene excision. Expression of the CaR protein in osteoblasts and osteocytes is absent, as shown by immunofluorescence studies of frozen sections of tibias from the CaR obKO animals (Fig. 1F). When primary antibodies were omitted, no immunofluorescence was visible (not shown).
ALT = alanine aminotransferase; AST = aspartate aminotransferase; ALP = alkaline phosphatase.
Data shown as mean (± SEM).
Normative data for adult mice shown in parentheses if available.
9, c6, or d8 samples from 21-day-old mice were pooled into b4, c3, or d2 samples, respectively, for all measurements except calcium and PTH. Calcium values are from 16b, 11c, and 7d samples. PTH values are from 9b, 8c, and 6d samples.
p < 0.05 for heterozygous or homozygous knockout versus control.
At birth, the weights of CaR obKO mice were significantly reduced (by 18%, p < 0.001) compared with their littermates (Fig. 1G). From the second week of life, the CaR obKO mice suffered multiple fractures, particularly in the tibias, which impaired ambulation. CaR obKO mice died in either the third or fourth week of life, at which point their weights were ∼70% less than their littermates. They lived longer in smaller litters, presumably because of reduced competition for nursing. There were no significant differences in serum levels of phosphorus, magnesium, sodium, chloride, creatinine, liver enzymes (alanine aminotransferase, aspartate aminotransferase), and total protein in knockouts and heterozygotes at 21 days of age compared with controls (Table 2). We noted mildly elevated serum total calcium in CaR obKO mice (12.5 ± 0.35 mg/dL) compared with control (11.5 ± 0.15 mg/dL) and heterozygous CaR obKO (11.4 ± 0.22 mg/dL) littermates (p < 0.05). Serum intact PTH values were slightly elevated in heterozygous CaR obKO (113 ± 20 pg/mL) and homozygous CaR obKO (91 ± 10 pg/mL) versus controls (64 ± 8 pg/mL) (p < 0.05). There was, however, no gene dosage effect. The mildly elevated serum calcium may reflect the ongoing spontaneous fractures, fracture repair, failure to mineralize bone normally, and/or simple hemoconcentration in the knockout mice. The lack of a progressive dosage effect of CaR knockout on PTH levels in heterozygous versus homozygous mice does not support the idea that targeted CaR deletion via the 3.6 Col promoter Cre mice construct is consistently modifying CaR expression in the parathyroid gland, causing a “hyperparathyroid state.” Serum alkaline phosphatase levels were also elevated by ∼60% in the CaR obKO mice (Table 2), potentially the result of fracture repair rather than abnormal liver function because liver transaminases were normal. Reduced levels of albumin in the knockouts (Table 2) could reflect the poor nutritional status or reduced hepatic synthesis in these animals.25
Sections of 1-, 7-, 14-, and 21-day-old femurs were examined after Goldner staining (Fig. 2). By gross examination, femurs from CaR obKO were comparable in size to controls at day 1 of life (Fig. 2A), but cortical architecture was already altered in knockouts with widening and less tightly arranged mineral. At day 7, delayed skeletal growth and development were obvious in the knockout animals because the long bones were smaller with expanded growth plates (Fig. 2B). Delayed skeletal maturation was also reflected in poorly developed secondary spongiosa in the CaR obKO mice (Fig. 2B). By 14 days of age, changes in the growth plate were reversed (Fig. 2C). At 21 days of age, the prevailing histological features included decreased mineralization in both trabecular bone (secondary ossification centers, secondary spongiosa; Fig. 2D) and cortical bone (Fig. 2D), with excessive accumulation of osteoid in both compartments.
Examination of the sections at high power revealed increased cortical porosity in CaR obKO bones, evident from day 1 (Fig. 2E). Cortical porosity increased with age and was accompanied by pronounced accumulation of unmineralized osteoid in the CaR obKO mice from 7 to 21 days of age (Fig. 2F–H). Pronounced thinning of the mineralized cortex in the knockouts was obvious from 14 days of age (Fig. 2F) and persisted through 21 days of age (Fig. 2G). At this time point, the cortices of CaR obKO mice also appeared disorganized, with areas of trabecularization (Fig. 2G). Femurs from heterozygotes were indistinguishable from the controls at all time points examined (data not shown).
We quantitatively assessed three-dimensional structural characteristics of trabecular and cortical bone of control, heterozygote, and knockout mice at 21 days of age using in vitro µCT (Figs. 3 and 4). Both trabecular (Fig. 3A) and cortical (Fig. 4A) compartments showed marked reductions in bone mass. There were statistically significant reductions in the trabecular bone volume fraction (BV/TV; Fig. 3B; by∼75%), tissue bone mineral density (BMD of TV; by >90%; not shown), mineral content (BMD of BV; Fig. 3C; by ∼20%), trabecular number (Fig. 3D; by ∼40%), trabecular thickness (Fig. 3E; by ∼30%), and connectivity density (Fig. 3G; by ∼80%) in the CaR obKO mice compared with these parameters in control mice. This was accompanied by a ∼50% increase in trabecular spacing (Fig. 3F). The changes were comparable, although less pronounced, in the L4 vertebra (not shown). The cortical compartment of the femur showed marked decreases in mineral content (Fig. 4A), which are quantified in the density histogram (Fig. 4B). The density distribution of voxels from the images of control cortices was composed of two peaks: a large peak at ∼400-mg equivalents of hydroxyapatite/cm3, representing soft tissue (ie, marrow elements and osteoid), and a smaller peak at ∼2000-mg equivalents of hydroxyapatite/cm3, representing mineralized bone (Fig. 4B, solid line). Cortices from the knockout animals exhibited a marked leftward shift in the voxel density distribution and no ∼2000-mg hydroxyapatite/cm3 peak (Fig. 4B, dashed line), further confirming marked hypomineralization of the bones. We observed statistically significant differences in the following parameters in the CaR obKO mice compared with the control mice: reductions in BV (Fig. 4C; by ∼60%), cortical thickness (Fig. 4D; by ∼70%), and BMD (Fig. 4E; by ∼45%) and a marked increase in cortical porosity (Fig. 4F; of ∼70%). Compared with the control animals, no significant differences were observed in heterozygotes in any parameters tested.
We next assessed expression levels of well-established markers of osteoblast differentiation in humeral cortices, from which the marrow had been removed, by qRT-PCR. mRNAs encoding collagen I, an early osteoblast marker, osteocalcin, a marker of mature osteoblasts, as well as sclerostin, a protein expressed predominantly by osteocytes, were all significantly reduced (Fig. 5A–C). Reductions in osteocalcin were observable at days 1 and 7 (Fig. 5B). Mineralization, an important function of differentiated osteoblasts, is regulated by multiple proteins including NPP1, ANK, and alkaline phosphatase. NPP1, which generates inorganic pyrophosphate (PPi) from nucleoside triphosphates,26, 27 and the transmembrane protein ANK, which mediates intracellular to extracellular transport of PPi,28 act in concert to inhibit mineralization.29 Alkaline phosphatase, on the other hand, hydrolyzes PPi and other substrates to generate inorganic phosphate that is incorporated into calcium crystals, thereby supporting mineral formation.30 Levels of alkaline phosphatase mRNA in the bones of CaR obKO animals were elevated, but these changes did not reach statistical significance (not shown). mRNAs encoding osteopontin (OPN), ANK, and NPP1 were increased in knockouts, in a time-dependent fashion, with the most pronounced increases in expression at 21 days of age (Fig. 5D–F). It has been proposed that the balance of the concentrations of RANKL and osteoprotegerin (OPG) is the crucial modulator of osteoclastogenesis and resorption.31 We found that RANKL mRNA was significantly elevated in bones from the CaR obKO mice (Fig. 5G). Although there was a trend toward increased mRNA levels of the soluble inhibitor and decoy receptor for RANKL, OPG, it was not significant (Fig. 5H).
To address the cellular mechanisms underlying the phenotype, we performed histomorphometry of the distal femur and femoral midshaft (Fig. 6). Compared with the cuboidal shape of osteoblasts in the bone sections from control mice, osteoblasts in cancellous bone from CaR obKO animals had more irregular polygonal shapes. Osteoblast numbers/bone surface (Fig. 6A) were reduced in the trabecular bone of the CaR obKO mice. Impaired osteoblastic activity was evident from the significantly reduced calcein incorporation into the secondary spongiosa of CaR obKO mice compared with the controls (Fig. 6B). Cortical bone also exhibited abnormal mineralization, manifested as disorganized endosteal calcein incorporation and absent periosteal labeling (Fig. 6C). Quantification of TRACP-stained osteoclasts revealed a doubling of osteoclast numbers (Fig. 6D) and erosion surfaces (Fig. 6E) in the trabecular bone of the knockout animals.
Analysis of cultured osteoblasts and osteoclasts
To assess the function of bone cells from CaR obKO mice, we cultured calvarial osteoblasts and examined their gene expression, proliferation, and mineralized nodule formation. PCR confirmed efficient excision of the CaR gene in the cultured calvarial osteoblasts from the knockouts (Fig. 7A). Analysis of gene expression in confluent calvarial osteoblasts (Fig. 7B) showed reduced levels of genes associated with osteoblast differentiation, collagen I and osteocalcin, as well as the osteocyte marker, dentin matrix protein (DMP1), in the CaR obKO cells. Levels of OPN were significantly increased, whereas elevations in expression of alkaline phosphatase mRNA did not reach statistical significance (not shown).
Proliferation of CaR obKO calvarial osteoblasts, assessed by cleavage of the tetrazolium salt WST-1 by mitochondrial dehydrogenases (Roche, Mannheim, Germany), was equivalent to cells isolated from the control animals (not shown). On the other hand, production of mineralized nodules, an activity present in terminally differentiated osteoblasts, was impaired in calvarial osteoblasts from CaR obKO mice, which showed a ∼50% reduction in mineralized nodule area compared with the control osteoblasts (Fig. 7C).
Since osteoclastic precursors require cell-cell interaction with osteoblasts/stromal cells for commitment, differentiation, and maturation,32 we investigated the ability of CaR obKO osteoblasts to facilitate osteoclastogenic differentiation in vitro. The numbers of TRACP-positive osteoclasts formed in the coculture system of CaR obKO calvarial osteoblasts and wild-type spleen osteoclast progenitors were significantly elevated (5-fold) compared with the controls, indicating that the increased numbers of osteoclasts seen in vivo is at least in part a function of the loss of CaR in the osteoblasts (Fig. 7D).
These studies examined the role of the osteoblast CaR in the regulation of bone development and remodeling in conditional CaR knockout mice, which avoid the severe metabolic complications of prior global CaR knockout models. In our model of conditional CaR knockout, the receptor was deleted broadly across the osteoblast lineage, including preosteoblasts and proliferating, differentiating, and mature osteoblasts, as well as osteocytes—all cells that express the Col 3.6 promoter.17 Markedly decreased CaR protein expression was evident in osteoblasts and osteocytes in CaR obKO mice. The most striking characteristic of the CaR obKO mouse phenotype was its early lethality. Considering that prevention of hyperparathyroidism in “global knockouts” of the CaR by two different strategies rescues their phenotype,9, 33 this was unexpected. Osteoblast-specific knockouts of the CaR (current study and Chang et al.)16 further support the idea that the skeletal phenotype in the “global CaR knockouts” is primarily due to severe hyperparathyroidism. However, in the “global knockout” model studied previously, insight into the functionality of the osteoblast CaR is masked by the compensatory actions of the alternatively spliced CaR, which in the skeleton, if not in the parathyroid gland, appears sufficient to mediate calcium responsiveness. As previously observed by our group using other promoter-driven Cre transgenic mice (2.3 Col 1 promoter),16 the expression of the full-length CaR in cells expressing type 1 collagen is essential for postnatal survival and normal skeletal development.
We studied the development of the skeletal phenotype in the first 3 weeks of life in the CaR obKO mice. The Col 3.6-Cre recombinase construct acts from embryonic day E18.17 We observed slightly but significantly reduced body weight and expression of osteocalcin mRNA at 1 day of age. The phenotype progressively worsened and was hallmarked by retarded skeletal development and growth. From the second week of life, knockout mice suffered long bone fractures. Difficulties with ambulation may have contributed to their undernourishment, although the subtle changes observed in the lungs and kidneys may also have restricted development, leading to their early deaths. However, renal function, as reflected by serum creatinine and electrolyte levels, was comparable to control mice and did not point to overt kidney pathology. We did note that serum calcium and/or PTH values in both heterozygous and homozygous CaR obKO mice were slightly higher than control mice. Although some of the differences were statistically significant, the pattern of changes in serum calcium and PTH did not support primary hyperparathyroidism or a gene dosage effect of CaR deletion on parathyroid function. Serum calcium and PTH alterations could be contributing to skeletal abnormalities in CaR obKO mice, but these alterations are unlikely to explain the bone phenotype of the CaR obKO. The heterozygous conditional parathyroid CaR KOs previously reported16 have PTH levels twofold higher than control mice and similar to the PTH levels in heterozygous and homozygous obCaR KO. Yet bone of heterozygous conditional parathyroid CaR KOs was only mildly osteopenic during the first 6 months of life and showed none of the dramatic alterations in mineralization that we saw in the homozygous obCaR KOs reported herein.
These observations, together with a comparable skeletal phenotype development in mice in which the osteoblast CaR knockout is directed by the Col 2.3 promoter,16 lead us to conclude the CaR obKO phenotype is primarily due to an osteoblast defect, although it could have been further exacerbated by a reduction of the CaR expression in lungs, kidneys, and even parathyroid glands. Further detailed examination of extraskeletal tissues, which is outside the scope of the current study, will offer more insight into the effects of CaR knockout in such tissues.
The most striking skeletal finding beyond growth retardation was impaired mineralization in CaR obKO mice, and it was evident early. Histological evaluation identified excess osteoid accumulation, delayed development of secondary ossification centers, and impaired mineralization of the extracellular matrix. µCT studies showed reduced mineral content of trabecular and cortical bone, decreased bone volume, and a deterioration in trabecular architecture, reflected in decreased trabecular connectivity density, leading to severe osteoporosis. Similar changes were also evident in the vertebrae. Further studies of fluorescently labeled mineralizing surfaces showed delayed and disorganized mineralization of trabecular and cortical bone, confirming problems in skeletal development in CaR obKOs. Finally, lack of periosteal labeling of cortical bone is consistent with delayed cortical bone growth.34 These changes confirm and substantially extend the observations we reported for mice in which the CaR was knocked out in cells expressing Col 2.3 kb promoter.16
Our findings in vitro support the involvement of the CaR in the regulation of bone formation.3 Immature osteoblasts are unable to form a mineralized extracellular matrix, a phenomenon we observed when we inhibited CaR activity in vitro.3 These studies also showed that the functional CaR is essential for osteoblast differentiation and production of mineralized matrix in vivo. Observations that support this include the following: 1) Expression levels of markers of osteoblastic differentiation (collagen I, osteocalcin, sclerostin) are significantly reduced in skeletal tissues of CaR obKO mice; 2) Absence of the CaR resulted in changes in both the morphology of osteoblasts and a reduction in the percentage of trabecular bone surfaces lined by osteoblasts; 3) Cultured calvarial osteoblasts recapitulated our findings in vivo. Cells isolated from the knockouts had an immature differentiation profile, evident by both reduced expression of osteoblast and osteocyte markers and an impaired ability to form mineralized nodules. Proliferation in these cultures was unaffected. Taken together, we propose that the inability of osteoblasts to reach full maturity results in their defective mineralizing capacity and that CaR is required for this differentiation to occur.
Although expression of osteocyte molecules sclerostin and DMP1 is decreased in CaR obKOs, it is unclear whether this is secondary to the retarded osteoblast differentiation resulting in reduced numbers of osteocytes, or a direct consequence of absent CaR signaling in osteocytes. Osteocyte-specific CaR knockouts will directly address this question in the future studies.
We investigated the molecular mechanism responsible for the mineralization defect in CaR obKO mice. In the bones of CaR obKO mice, we detected elevated expression of NPP1 and ANK, membrane proteins that regulate mineralization in a complex but highly integrated fashion.29 On the other hand, expression of alkaline phosphatase, an enzyme thought to stimulate mineralization, was not significantly affected. Such directional alterations in these key mediators are predicted to result in an overall inhibition of mineralization, which was observed in the CaR obKO mice. Alkaline phosphatase, NPP1, and ANK coordinately regulate levels of OPN, the levels of which are decreased in alkaline phosphatase knockouts and increased in mice with inactivated NPP1 and ANK genes.29 Among its many functions, OPN is thought to inhibit mineral growth directly.35 We confirmed elevated OPN expression in the CaR obKO model, both in vivo and in cultures of calvarial osteoblasts in vitro, and propose that the reduced mineralization, at least in part, results from its inhibitory effects on crystal growth. Further work, using double CaR/OPN, CaR/ANK, and CaR/NPP1 knockouts, will be required to dissect out the contribution of each of these molecules to the mineralization defects in these mice. As well, direct measurements of alkaline phosphatase enzymatic activity in tissue and cultured osteoblasts will be informative as to the meaning of the PCR results.
Another mechanism that may fuel the development of osteopenia in CaR obKOs is altered control of bone resorption. We observed increased osteoclast numbers and activity, reflected in the percent of erosion surface in CaR obKO mice. Resorption is regulated by the interplay of a large number of local autocrine and paracrine factors. The most critical local factors are RANKL, a stimulator of osteoclast formation and activity, and OPG, its antagonist. Both are produced by osteoblasts.31, 36 In femoral cortices from these mice, RANKL mRNA expression was increased, indicating that lack of CaR signaling in osteoblasts affects this important local stimulator of remodeling. Over time, the increase in RANKL/OPG ratio would promote bone loss and osteopenia.37 Furthermore, such changes could be exacerbated by the increased levels of OPN that we observed in bone tissue and in cultured calvarial osteoblasts isolated from the knockout animals because OPN has been shown to stimulate osteoclast maturation and bone resorption.38 The increase in cortical porosity shown in CaR obKO mice may be caused or exacerbated by the elevated osteoclast activity, in concert with the mineralization defect.
To confirm that absence of the osteoblast CaR is responsible for changes in osteoclast numbers and activity that we observed in vivo, we explored the osteogenic potential of the CaR obKO osteoblasts in vitro. Cultured osteoblasts isolated from the knockouts exhibited an elevated osteoclastogenic potential, stimulating the development of up to 5-fold more TRACP-positive osteoclasts in the coculture experiments compared with cells from control mice. High [Ca2+]e has been shown to directly inhibit osteoclast function, via the osteoclast CaR, in vitro.39 This is the first study to our knowledge that provides evidence for an inhibitory role of the osteoblast CaR in regulation of osteoclastogenesis and bone resorption in vivo.
Why is bone resorption in the CaR obKO mice increased? The matrix formed by the osteoblasts in these mice is highly undermineralized, evident by osteoid accumulation. We hypothesize that such abnormalities may trigger matrix remodeling. Alternatively, these findings may support the hypothesis that, during resorption, Ca2+ is released from the mineral and acts as a coupling factor to inhibit osteoclast function and stimulate bone formation, via the CaR. When the CaR is deleted in the CaR obKO mice, resorption increases and formation decreases. Our prior studies with a transgenic mouse model, in which constitutively active CaRs were targeted only to mature osteoblasts and osteocytes under the control of the osteocalcin promoter (Act-CaR mice), showed increased bone resorption and elevated RANKL levels,24 findings that are seemingly contradictory to the current study. However, combining the findings from these animal models, we propose that CaR signaling in a broad population of young to mature osteoblasts promotes mineralized bone formation and inhibits resorption by regulating RANKL expression, thereby aiding accretion of newly formed bone. On the other hand, activation of the CaR exclusively in mature osteoblasts and osteocytes (which express constitutively active CaRs under control of the osteocalcin promoter in Act-CaR mice) contributes to remodeling of old bone by stimulating osteoclast differentiation and function.24
We present evidence for the essential role of the osteoblast CaR in postnatal survival, as well as osteoblast differentiation and function, including regulation of bone formation and resorption. Elucidation of the roles of the osteoclast and osteocyte CaRs awaits generation and study of osteoclast- and osteocyte-specific CaR knockout models, a task in which Flox-CaR mice will prove a valuable tool.
All the authors state that they have no conflicts of interest.
We gratefully acknowledge Dr Barbara Kream for the gift of the Col 3.6-Cre mice, as well as the technical help of Hashem Elalieh and Michelle Buttler. This work was supported by a Veterans Affairs Merit Review (DS), a Department of Veterans Research Enhancement Award Program in Bone Disease (DS, DB), NIH RO1 AG21353 (WC), NIH RO1 AR055888 (DS), and NIH RO1 DK054793 (DB).
Authors' roles: MMD-E performed the majority of the experiments reported and drafted all versions of the manuscript. T-HC, CG, NL, and BL assisted with technical aspects of animal breeding, genotyping, PCR, and biochemical analyses. CT contributed to the development of the animal model. DPB, WC, and DS directed the experiments and contributed substantially to the manuscript development.