Mouse Osteoblastic Cell Line (MC3T3-E1) ExpressesExtracellular Calcium (Ca21

o)–Sensing Receptor and ItsAgonists Stimulate Chemotaxis and Proliferation of

MC3T3-E1 Cells

TORU YAMAGUCHI,1 NAIBEDYA CHATTOPADHYAY,1 OLGA KIFOR,1 ROBERT R. BUTTERS, JR.,1

TOSHITSUGU SUGIMOTO,2 and EDWARD M. BROWN1

ABSTRACT

The calcium-sensing receptor (CaR) is a G protein–coupled receptor that plays key roles in extracellular calciumion (Ca21

o) homeostasis in parathyroid gland and kidney. Osteoblasts appear at sites of osteoclastic boneresorption during bone remodeling in the “reversal” phase following osteoclastic resorption and preceding boneformation. Bone resorption produces substantial local increases in Ca21

o that could provide a signal forosteoblasts in the vicinity, leading us to determine whether such osteoblasts express the CaR. In this study, we usedthe mouse osteoblastic, clonal cell line MC3T3-E1. Both immunocytochemistry and Western blot analysis, usingan antiserum specific for the CaR, detected CaR protein in MC3T3-E1 cells. We also identified CaR transcriptsin MC3T3-E1 cells by Northern analysis using a CaR-specific riboprobe and by reverse transcription-polymerasechain reaction with CaR-specific primers, followed by nucleotide sequencing of the amplified products. Exposureof MC3T3-E1 cells to high Ca21

o (up to 4.8 mM) or the polycationic CaR agonists, neomycin and gadolinium(Gd31), stimulated both chemotaxis and DNA synthesis in MC3T3-E1 cells. Therefore, taken together, our datastrongly suggest that the osteoblastic cell line MC3T3-E1 possesses both CaR protein and mRNA very similar, ifnot identical, to those in parathyroid and kidney. Furthermore, the CaR in these osteoblasts could play a key rolein regulating bone turnover by stimulating the proliferation and migration of such cells to sites of bone resorptionas a result of local release of Ca21

o. (J Bone Miner Res 1998;13:1530–1538)

INTRODUCTION

BONE FORMATION DURING remodeling of the skeleton isinitiated by the migration of preosteoblasts to sites of

osteoclastic bone resorption during the “reversal” phasethat precedes the laying down of new bone. These cellssubsequently differentiate into mature osteoblasts andeventually deposit and mineralize osteoid protein.(1) Boneresorption induces local increases in the extracellular cal-cium concentration (Ca21

o) within the immediate vicinityof osteoclasts that are known to reach levels as high as40 mM.(2) The latter could therefore provide preosteoblasts

with a signal that modulates their subsequent physiologicalresponses, such as migration and proliferation. In fact,some studies have shown that high Ca21

o induces chemo-taxis(3,4) and DNA synthesis(5,6) of the mouse osteoblasticclonal cell line MC3T3-E1,(7) which differentiates frompreosteoblasts to mature osteoblasts as a function of time inculture.(8–10)

Although the mechanism by which MC3T3-E1 cells re-spond to changes in Ca21

o is unclear, one possibility is thatthis occurs via the Ca21

o-sensing receptor (CaR), which hasrecently been cloned from bovine and human parathyroidgland,(11,12) rat kidney,(13) and thyroid C cells.(14) The phys-

1Endocrine-Hypertension Division, Department of Medicine, Brigham and Women’s Hospital, and Harvard Medical School, Boston,Massachussetts, U.S.A.

2Third Division, Department of Medicine, Kobe Universeity School of Medicine, Kobe, Japan.

JOURNAL OF BONE AND MINERAL RESEARCHVolume 13, Number 10, 1998Blackwell Science, Inc.© 1998 American Society for Bone and Mineral Research

1530

iological relevance of the CaR has been documented inhumans by showing that inactivating and activating muta-tions of the CaR gene cause inherited hyper- and hypocal-cemic disorders,(15,16) respectively, rendering affectedfamily members inappropriately “resistant” or “sensitive,”respectively, to the usual effects of Ca21

o on parathyroidand renal functions. The CaR binds its cationic ligands,such as Ca21, gadolinium (Gd31), and neomycin, with EC50

values of 3 mM, 20 mM, and 60 mM, respectively. Theseagonists produce a G protein–dependent activation ofphospholipase C (PLC), leading to elevations in the levelsof inositol trisphosphate (IP3) and the cytosolic calciumconcentration.(11) Indeed, Quarles et al. showed that thesethree agonists stimulate DNA synthesis in MC3T3-E1 cellswith EC50 values close to those described above,(6) suggest-ing that these cells have a Ca21

o-sensing mechanism func-tionally similar to the CaR. However, they failed to detectexpression of the CaR by Northern analysis or reversetranscription-polymerase chain reaction (RT-PCR) inMC3T3-E1 cells,(17) and thus the precise molecular mech-anism(s) by which these osteoblastic cells detect changes inCa21

o still remains unknown.In a previous study, using immunohistochemistry with

CaR-specific antisera, we showed expression of this recep-tor in diverse cell types in human bone marrow, includingalkaline phosphatase (ALP)-positive, putative osteoblastprecursors, nonspecific esterase–positive mononuclearcells, erythroid precursors, and megakaryocytes.(18) Thesefindings suggested that the CaR might be involved in theCa21

o-sensing mechanism of these bone marrow-derivedcells. In this study, we used MC3T3-E1 cells as a model ofosteoblasts in bone marrow and further examined the pres-ence and role of the CaR in these cells. We demonstrateclear expression of the CaR in MC3T3-E1 cells as assessedby immunocytochemical staining and Western blot analysisusing an anti-CaR antiserum as well as Northern analysiswith a CaR-specific probe and RT-PCR with CaR-specificprimers. We also confirm that CaR agonists stimulate bothchemotaxis and DNA synthesis in the CaR-expressingMC3T3-E1 cells examined in this study. These results showthat both CaR protein and mRNA are expressed in theMC3T3-E1 cell line; in addition, they suggest that the re-ceptor could potentially play a pivotal role in regulating thefunction of osteoblasts present within the marrow by sens-ing local changes in Ca21

o related to bone remodeling.

MATERIALS AND METHODS

Materials

All routine culture media were obtained from GIBCOBRL (Grand Island, NY, U.S.A.). Neomycin sulfate andanhydrous calcium chloride (CaCl2) were purchased fromSigma Chemical Co. (St. Louis, MO, U.S.A.), and Gd31

(III) chloride hexahydrate was from Aldrich Chemical Co.(Milwaukee, WI, U.S.A.). [3H]methylthymidine was pur-chased from DuPont New England Nuclear (Boston, MA,U.S.A.).

Cell culture

MC3T3-E1 cells, established as an osteoblastic cell linefrom normal mouse calvaria,(7) were the generous gift fromDr. H. Kodama (Ohu University, Koriyama, Japan). Adifferent batch of MC3T3-E1 cells was provided by NPSPharmaceuticals, Inc. (Salt Lake City, UT, U.S.A.).MC3T3-E1 cells were grown in alpha-modified minimalessential medium (a-MEM; Ca21, 1.8 mM; Mg21, 0.81mM; H2PO4, 1.0 mM) supplemented with 10% fetal bovineserum (Hyclone, Logan, UT, U.S.A.) and 1% penicillin/streptomycin in 5% CO2 at 37°C. The medium was changedtwice weekly, and the cells were subcultured into 25 cm2

culture flasks by detaching them gently with a cell scraperafter reaching subconfluency. For morphological evalua-tion, MC3T3-E1 cells were plated onto 12-mm circular glasscoverslips in 24-well (2.0 cm2) plates. After 24 h of culture,the medium was discarded, and each coverslip with adher-ent cells was washed once with phosphate-buffered saline(PBS), fixed with 4% formaldehyde in PBS for 5 minutes,and washed with PBS once again. Each coverslip was storedat 4°C until assessment for the presence of the CaR asdescribed below.

Immunocytochemistry for CaR in MC3T3-E1 cells

A CaR-specific polyclonal antiserum (4637) was gener-ously provided by Drs. Forrest Fuller and Karen Krapcho ofNPS Pharmaceuticals, Inc. This antiserum was raisedagainst a peptide (FF-7) corresponding to amino acids 345–359 of the bovine CaR, which resides within the predictedamino-terminal extracellular domain of the CaR. The anti-serum was subjected to further purification using an affinitycolumn conjugated with the FF-7 peptide, and the affinity-purified antiserum was used for immunocytochemistry andWestern blot analysis as described below. As describedpreviously,(18) while there is some similarity in the region ofthe metabotropic glutamate receptors corresponding to theFF-7 peptide (40–53%), the 4637 antiserum and severalother anti-CaR antisera that we have employed in otherstudies (M. Bai et al., manuscript in preparation) failed toreact on Western analysis with cellular protein isolatedfrom HEK293 cells transfected with the metabotropic glu-tamate receptor (mGluR1) cDNA. In contrast, these pro-teins were readily detected on Western blot using an anti-mGluR antiserum. Fixed MC3T3-E1 cells were treated withperoxidase blocking reagent (DAKO Corp., Carpenteria,CA, U.S.A.) for 10 minutes to inhibit endogenous peroxi-dases, followed by treatment with protein block serum-freesolution (DAKO Corp.) for 1 h and then incubated over-night at 4°C with the affinity-purified, anti-CaR antiserumat a concentration of 5 mg/ml in blocking solution. Negativecontrols were carried out by incubating the cells with affin-ity-purified antiserum 4637 that had been preabsorbed with10 mg/ml of the FF-7 peptide. After washing the cells threetimes with 0.5% bovine serum albumin (BSA) in PBS for10 minutes each, peroxidase-coupled, goat anti-rabbit im-munoglobulin G (IgG, 1:200; Sigma Chemical Co.) wasadded and incubated for 1 h at room temperature. The cellswere then washed with PBS three times for 10 minutes

CaR AND MOUSE OSTEOBLASTSIC CELL LINE 1531

each, and the color reaction was developed using theDAKO AEC substrate system (DAKO Corp.) for about10 minutes. The color reaction was stopped by washingthree times in water.

Western analysis of CaR in MC3T3-E1 cells

Monolayers of MC3T3-E1 cells in 75 cm2 culture flasksthat had been cultured for 5, 13, or 20 days were rinsedtwice with 1 mM EDTA in PBS and lysed with 1.0 ml of alysis solution (1% sodium dodecyl sulfate [SDS], 10 mMTris-HCl, pH 7.4) heated to 65°C. The cells were scrapedfrom the flasks, transferred to microcentrifuge tubes, andheated for an additional 5 minutes at 65°C. The viscosity ofthe sample was reduced by brief sonication, and insolublematerial was removed by centrifugation for 5 minutes. Theresultant whole cell lysate in the supernatant was stored at–20°C until Western blot analysis was carried out.

Aliquots of 150 mg of protein were dissolved in SDS-Laemmli gel loading buffer containing 100 mM dithiothre-itol, incubated at 37°C for 15 minutes, and resolved elec-trophoretically on 6.5% SDS-polyacrylamide gels. Proteinswere electrophoretically transferred to nitrocellulose at 240mA for 40 minutes in transfer buffer containing 19 mMTris-HCl, 150 mM glycine, 0.015% SDS, and 20% metha-nol. The blots were blocked for 2 h with 1% BSA in PBScontaining 0.25% Triton X-100 (blocking solution) andthen incubated overnight at 4°C with the affinity-purifiedantiserum (4637) or with peptide-blocked antiserum (thesame amount of antiserum preincubated at room tempera-ture for 60 minutes with twice the amount of FF-7 peptide)at a concentration of 1 mg/ml in the blocking solution. Theblots were washed three times with PBS containing 0.25%Triton X-100 (washing solution) at room temperature for10 minutes each. The blots were further incubated with a1:2000 dilution of horseradish peroxidase-coupled, goat an-ti-rabbit IgG (Sigma Chemical Co.) in the blocking solutionfor 1 h at room temperature. The blots were then washedthree times with the washing solution at room temperaturefor 40 minutes each, and specific protein bands were de-tected using an enhanced chemiluminescence system (Am-ersham, Arlington Heights, IL, U.S.A.).

Western analysis of the mGluR in MC3T3-E1 cells

Western analysis of mGluR1 in MC3T3-E1 cells wasperformed using an anti-mGluR1 antiserum (Pharmingen,San Diego, CA, U.S.A.) and horseradish peroxidase-cou-pled, goat anti-mouse IgG (Sigma Chemical Co.) essentiallyas described above for Western blots carried out with theanti-CaR antiserum.

Detection of CaR transcripts by Northern blot analysis

For the purpose of determining the sizes of the CaRtranscripts in MC3T3-E1 cells, Northern blot analysis wasemployed on aliquots of 5 mg of poly(A1) RNA obtainedusing oligo-dT cellulose chromatography of total RNA. Ratkidney poly(A1)-enriched RNA, 5.0 mg, was used as posi-tive control employing the same cRNA probe in a different

blot. RNA samples were denatured and electrophoresed in2.2 M formaldehyde-1% agarose gels along with an 0.24–9.5 kb RNA ladder (GIBCO BRL) and transferred over-night to nylon membranes (Duralon; Stratagene, La Jolla,CA, U.S.A.). A 577 bp XhoI-SacI fragment correspondingto nucleotides 721–1298 of the RaKCaR cDNA was sub-cloned into the pBluescript(SK1) vector. The plasmid wasthen linearized with BglII, and a 32P-labeled riboprobe wassynthesized with the MAXIscript T3 kit (Pharmacia Bio-tech, Piscataway, NJ, U.S.A.) using T3 polymerase and32P-UTP. Nylon membranes were prehybridized for 2 h at52°C in a solution consisting of 50% formamide, 43 Den-hardt’s solution (503 Denhardt’s 5 5 g of Ficoll, 5 g ofpolyvinylpyrrolidone, and 5 g of BSA), 53 SSPE (203SSPE 5 2.98 M NaCl and 0.02 M EDTA in 0.2 M phos-phate buffer, pH 7.0), 0.5% SDS, 10% dextran sulfate,250 mg/ml yeast tRNA, and 200 mg/ml calf thymus DNA.Labeled cRNA probe (2 3 106 cpm/ml) was then added,and the membranes were hybridized overnight at 52°C forthe MC3T3-E1 sample and at 65°C for kidney RNA sample.Washing was carried out for 20 minutes at moderate strin-gency (0.33 SSC [203 SSC 5 3 M NaCl and 0.3 M Na3-citrate z 2H2O], 0.5% SDS at 55°C) for the RNA fromMC3T3-E1 cells, while for mouse kidney RNA, the blot waswashed at 65°C for 20 minutes. Moderate stringency wash-ing was employed for MC3T3-E1 RNA samples because ofthe relatively low level of expression of CaR transcripts inthese cells and because of the use of a heterologous probe(e.g., derived from the rat kidney CaR). In control studiescomparing the use of moderate versus high stringencywashes for Northern blot analysis of RNA from kidney,although there was some increase in the intensity of thebands after washing at lower stringency, no additionalbands were observed. Membranes were then exposed toX-ray film (Kodak XAR-5; Eastman Kodak, Rochester,NY, U.S.A.) for 4 days at –70°C.

PCR amplification of CaR in MC3T3-E1 cells

Total RNA was prepared from MC3T3-E1 cells using theTRIzol Reagent (GIBCO BRL). One microgram of totalRNA was used for the synthesis of single-stranded cDNA(cDNA synthesis kit; GIBCO BRL). The resultant first-stranded cDNA was used for the PCR procedure. PCR wasperformed at a final concentration of 20 mM Tris-HCl(pH 8.4), 50 mM KCl, 1.8 mM MgCl2, 0.2 mM dNTP,0.4 mM of forward primer, 0.4 mM of reverse primer and1 ml of ELONGASE enzyme mix (a Taq/Pyrococcus speciesGB-D DNA polymerase mixture; GIBCO BRL). Theprimer sequences were: 59-ATGGTTTGGCTACTGTT-TGG-39, sense; 59-CAGAGCCTTGGAGACGGTGT-39,antisense, which encode nucleotides 40–59 and 339–358,respectively, of the partially cloned extracellular domain ofthe mouse AtT-20 pituitary cell CaR.(19) This primer setwas designed to span one intron of the CaR gene, todistinguish products amplified from cDNA and genomicDNA. To perform “hot start” PCR, the enzyme was addedduring the initial 3-minute denaturation and was followedby 35 cycles of amplification (30-s denaturation at 94°C,30-s annealing at 47°C and 1-minute extension at 72°C).

1532 YAMAGUCHI ET AL.

The reaction was completed with an additional 10-minuteincubation at 72°C to allow completion of extension. PCRproducts were fractionated on 1.2% agarose gels. The pres-ence of a 319 nucleotide-long amplified product was indic-ative of a positive PCR reaction. The PCR product in thereaction mixture was purified using the QIAquick PCRpurification kit (Qiagen, Santa Clarita, CA, U.S.A.) andsubjected to direct, bidirectional sequencing employing thesame primer pairs used for PCR by means of an automatedsequencer (AB377; Applied Biosystems, Foster City, CA,U.S.A.) in the DNA Sequence Faculty of the University ofMaine (Orono, ME, U.S.A.), using dideoxy terminator Taqtechnology.

Chemotaxis assay of MC3T3-E1 cells

Chemotaxis was evaluated using a blindwell chamber(BW200S; Neuro Probe Inc., Gaithersburg, MD, U.S.A.) aspreviously described.(3,20) CaCl2, 1.8 or 4.8 mM, 300 mMneomycin sulfate, or 100 mM GdCl3 z 6H2O in serum-freeDulbecco’s MEM were loaded in the lower chamber, whichis separated from the upper well by a 5 mm membrane with5-mm pores. MC3T3-E1 cells (1 3 105 cells/ml) were dis-sociated with trypsin-EDTA solution (GIBCO BRL),washed twice, suspended in serum-free a-MEM, and addedto the upper chamber. After a 12-h incubation at 37°C, cellson the upper surface of the membrane that had not mi-grated were scraped from the membrane, and cells that hadmigrated to the opposite side of the membrane were fixedwith methanol and stained with Giemsa. The cells that hadmigrated to the lower surface of the filter were counted insix high-power fields (3400) using a light microscope. Forthe purpose of comparison between multiple assays, thedata were normalized as the fold increase in cellular che-motaxis relative to that of the control.

DNA synthesis in MC3T3-E1 cells

We assessed DNA synthesis in MC3T3-E1 cells using[3H]thymidine incorporation. MC3T3-E1 cells were disso-ciated by gentle scraping and repeated pipeting and seededin 24-well (2.0 cm2) plates at a density of 1000 cells/well in500 ml of a-MEM containing 10% fetal bovine serum aswell as various concentrations of CaCl2, neomycin sulfate,or GdCl3 z 6H2O as indicated in Fig. 5. After a 48 h incu-bation at 37°C, cells were pulsed with [3H]thymidine (1 mCi/well). Incubations were terminated after overnight incuba-tion by removal of the medium and addition of 5%trichloroacetic acid. Cells were scraped and transferred tomicrocentrifuge tubes. After centrifugation at 15,000g andremoval of the supernatant, the precipitate was washed with75% ethanol and desiccated at room temperature. Theresidual pellet was dissolved in 20 mM NaOH and 1% SDS,and a scintillation cocktail was added. Samples werecounted in a liquid scintillation counter.

Statistics

Results are expressed as the mean 6 SEM. Statisticalevaluation for differences between groups was done using

one-way analysis of variance followed by Fisher’s protectedleast significant difference. For all statistical tests, values ofp , 0.05 were considered significant.

RESULTS

Immunoreactivity of CaR protein in MC3T3-E1 cellsusing CaR-specific antiserum

To clarify whether the CaR is expressed by osteoblasts,we investigated the presence of the receptor in the mouseclonal osteoblastic cell line, MC3T3-E1. Immunocytochem-istry of MC3T3-E1 cells with a CaR-specific polyclonalantiserum revealed strong CaR staining (Fig. 1A), whichwas eliminated by preincubating the primary antiserum withthe peptide against which it was raised (Fig. 1B).

We also performed Western analysis of CaR on proteinsisolated from MC3T3-E1 cells. Bands were stained in pro-teins obtained from the two different batches of the cells inthe presence of specific antiserum (Fig. 2, lanes 1 and 2).The molecular weight of the band at ;150–160 kDa in eachlane was of a size consistent with those of the intact, glyco-sylated human CaR (140–200 kDa),(21) bovine parathyroidCaR (;130–150 kDa),(22) and mouse kidney CaR (;140

FIG. 1. Immunocytochemistry of MC3T3-E1 cells carriedout as described in the Materials and Methods using aCaR-specific antiserum (4637). Immunocytochemistry re-vealed strong CaR staining (A), which was eliminated bypreincubating the primary antiserum with the peptideagainst which it was raised (B). The photomicrographs weretaken at a magnification of 31000.

CaR AND MOUSE OSTEOBLASTSIC CELL LINE 1533

kDa).(23) The specificity of the anti-CaR antiserum used inthis study was confirmed by the abolition of the bandsfollowing preabsorption of the anti-CaR antiserum with thepeptide against which it was raised (Fig. 2, lanes 3 and 4).No immunoreactivity for mGluR1 was observed in proteinsextracted from MC3T3-E1 cells by Western analysis usingan anti-mGluR1 antiserum (Fig. 2, lanes 5 and 6).

When MC3T3-E1 cells were cultured up to 20 days,they began to generate osmiophilic nodular regions withstrong ALP staining (data not shown), confirming theircapacity to differentiate to mature osteoblasts as re-ported previously.(8) Western analysis using anti-CaRantiserum on proteins isolated from the cells cultured for5, 13, or 20 days (lanes 7–9, respectively) showed that thecells expressed CaR protein over the entire range ofculture periods during which osteoblastic differentiationtook place.

Detection of CaR mRNA in MC3T3-E1 cells byNorthern blot analysis and RT-PCR

In Fig. 3, lane 1 shows Northern blot analysis performedusing a riboprobe derived from the rat kidney CaR onpoly(A1) RNA isolated from MC3T3-E1 cells, which re-vealed three transcripts with sizes of 9.5, 4.5, and 1.5 kb;these sizes were consistent with those of the mouse CaRtranscripts from AtT-20 pituitary cell line reported previ-ously.(19) However, a 7.5 kb band observed in AtT-20 cellswas not visible in MC3T3-E1 cells even after exposing theblot for 24 h in a phosphorimager cassette. Northern blotanalysis of mouse kidney revealed a predominant CaRmRNA transcript of 7.5 kb and two minor transcripts of9.5 kb and 4.1 kb (Fig. 3, lane 2).

RT-PCR with mouse CaR-specific primers, which wereintron-spanning to preclude amplification of a similar

sized product from any contaminating genomic DNA,amplified a product of the expected size, 319 bp for aCaR-derived product (Fig. 4, lane 1). No products wereobserved when the RT was omitted during synthesis ofcDNA (Fig. 4, lane 2). The 319 bp PCR product was alsoamplified from cDNA prepared from a different batch ofMC3T3-E1 cells (Fig. 4, lane 3). No products were de-tected in the control PCR reaction without RT (Fig. 4,lane 4). DNA sequence analysis of the 319 bp PCR prod-ucts revealed a 100% sequence identity with the mouseAtT-20 cell CaR sequence reported previously(19) (datanot shown). These results show that the RT-PCR prod-ucts corresponded to an authentic CaR sequence, indi-cating the presence of bona fide CaR transcripts in thesecells.

FIG. 2. Western analysis of whole cell lysates fromMC3T3-E1 cells. Western analyses of CaR and mGluR1were each performed on the two different batches of theMC3T3-E1 cells as described in the Materials and Methods.Bands at ;150–160 kDa that were stained in the presenceof specific antiserum were consistent with the intact glyco-sylated CaR(21–23) (lanes 1 and 2). The specificity of thelabeling by the antiserum used in this study to detect theCaR was confirmed by abolition of the bands in extracts ofcells incubated with peptide-preabsorbed CaR-antiserum(lanes 3 and 4). No mGluR1 immunoreactivity was ob-served in the cells using the anti-mGluR1 antiserum (lanes5 and 6). The cells cultured for 5, 13, or 20 days (lanes 7–9,respectively) showed immunoreactivity for the CaR overthe entire range of culture periods.

FIG. 3. Northern blot analysis of CaR transcripts inMC3T3-E1 cells (lane 1) and mouse kidney (lane 2) per-formed as described in the Materials and Methods. Arrowsshow the sizes of the CaR transcripts.

FIG. 4. Identification of CaR transcripts in MC3T3-E1cells using RT-PCR with intron-spanning mouse CaR-spe-cific primers, performed as described in the Materials andMethods. A product was amplified from reverse transcribedRNA isolated from MC3T3-E1 cells, which was of theexpected size, 319 bp (lane 1) for a CaR-derived product.The PCR reaction without RT showed no products (lane 2).A 319-bp PCR product was also amplified from cDNAisolated from the different batch of MC3T3-E1 cells (lane3), while no PCR product was apparent in the RT-minuscontrol reaction (lane 4). DNA sequence analysis of the319 bp PCR products revealed a 100% sequence identitywith the mouse AtT-20 cell CaR sequence reported previ-ously(19) (data not shown).

1534 YAMAGUCHI ET AL.

Chemotactic activity of MC3T3-E1 cells toward highCa21

o, neomycin sulfate, or Gd31o

A chemotaxis assay was performed to determine thecapacity of MC3T3-E1 cells to migrate toward CaR ago-nists. As shown in Fig. 5A, 4.8 mM Ca21

o induced a che-motactic response of MC3T3-E1 cells over control values at1.8 mM Ca21

o ( p , 0.05). Both neomycin sulfate (300 mM)and Gd31

o (100 mM) also induced significant chemotacticresponses over the control ( p , 0.05).

DNA synthesis of MC3T3-E1 cells stimulated by highCa21

o, neomycin sulfate, or Gd31o

We found that treatment of MC3T3-E1 cells with in-creasing levels of Ca21

o up to 4.8 mM resulted in a dose-dependent stimulation of DNA synthesis over control val-ues at 1.8 mM Ca21

o ( p , 0.05) (Fig. 5B). Neomycin

sulfate (100 and 300 mM) and Gd31o (25 and 100 mM) also

induced significant stimulations of DNA synthesis over thecontrol ( p , 0.05).

DISCUSSION

In a previous study, we examined the expression of theCaR in primary cultures of human bone marrow andshowed that low-density mononuclear cells isolated fromwhole human bone marrow, which are putatively enrichedin marrow progenitor cells, including bone cell precursors,expressed the CaR by Northern analysis, RT-PCR, andimmunocytochemistry.(18) About one third of these adher-ent, CaR-immunoreactive cells were also positive for ALP,suggesting the expression of the CaR in putative osteoblastprecursors or osteoblasts. Since these osteoblast-like cellswere taken directly from bone marrow and may be similarto those present in marrow in vivo, this finding suggestedthat the CaR might be expressed by preosteoblasts or os-teoblasts in vivo. However, we observed that these prepa-rations also contained other types of cells, including non-specific esterase–positive, monocyte-macrophage–like cells.Thus, such marrow-derived preparations might not be idealfor addressing specifically the existence and role of the CaRin osteoblasts. To circumvent this problem, we used mouseMC3T3-E1 cells possessing an osteoblastic phenotype inthis study.(7) Our results showed that MC3T3-E1 cellsclearly expressed CaR protein by immunocytochemistry(i.e., Fig. 1) as well as by Western blot analysis (e.g., Fig. 2),which revealed a specific band at a molecular weight con-sistent with that of the intact, glycosylated CaR (;150–160 kDa).(21–23) In addition, both Northern analysis per-formed on poly(A1) RNA from MC3T3-E1 cells and RT-PCR performed on total RNA from these cells followed bysequence analysis of the PCR products indicated the pres-ence of bona fide CaR transcripts (Figs. 3 and 4). Thus, thepresent study shows that this osteoblast cell line expressesboth CaR protein and mRNA.

It is known that MC3T3-E1 cells exhibit properties ofosteoprogenitor cells and preosteoblasts in their activelygrowing stage; following growth arrest, they differentiate todevelop markers of mature osteoblasts, including the ex-pression of high levels of ALP and the capacity to formmineralized bone matrix.(8–10) Thus, this cell line seems tobe a useful model for examining the process of osteoblasticdevelopment in vitro. In this study, MC3T3-E1 cells ex-pressed the CaR by Western blot analysis over the entirerange of culture periods examined (5, 13, and 20 days)(Fig. 2), which may provide an in vitro model of the devel-opmental stages from preosteoblast to mature osteo-blast.(9,10) This result suggests that MC3T3-E1 cells (and,perhaps, osteoblastic cells in vivo) express the CaRthroughout their differentiation.

Osteoblasts are known to play a crucial role in the for-mation phase of bone remodeling, by laying down the struc-tural components of bone (matrix and mineral) and secret-ing various cytokines and growth factors that influence bothbone formation and resorption.(24) Bone formation is initi-ated by the migration of preosteoblasts into resorption pits

FIG. 5. (A) Chemotactic activity of MC3T3-E1 cells to-ward high Ca21

o, neomycin sulfate or Gd31o. The number

of MC3T3-E1 cells that migrated to the side of the mem-brane to which CaCl2, neomycin sulfate, or GdCl3 z 6H2Ohad been added 12 h previously were counted as describedin the Materials and Methods. Each bar expresses themean 6 SEM for six determinations. *p , 0.05 comparedwith cells exposed to 1.8 mM Ca21

o. (B) Stimulation ofDNA synthesis in MC3T3-E1 cells by high Ca21

o, neomycinsulfate, or Gd31

o. MC3T3-E1 cells were treated with highCa21

o, neomycin sulfate, or Gd31o for 3 days. Values are

expressed as percentage of control (cells treated with1.8 mM Ca21

o). Each bar represents the mean 6 SEM forsix determinations. *p , 0.05 compared with cells treatedwith 1.8 mM Ca21

o.

CaR AND MOUSE OSTEOBLASTSIC CELL LINE 1535

at the end of osteoclastic bone resorption.(1) Substantialamounts of Ca21

o are released from mineralized bone ma-trix during osteoclastic resorption,(2) raising the level ofCa21

o in the vicinity of resorption sites. It is possible thatthe CaR senses these high levels of Ca21

o, thereby provid-ing a signal for preosteoblasts that induces their migrationinto sites where new bone formation is required. In fact, wepreviously showed that high Ca21

o induced both a chemo-tactic response and DNA synthesis of MC3T3-E1 cells,(3,5)

and Quarles et al. showed that high Ca21o and other CaR

agonists, such as Gd31o and neomycin, stimulated DNA

synthesis in MC3T3-E1 cells.(6) These findings suggest thata calcium-sensing mechanism is present in these osteoblas-tic cells and is involved in their migration and proliferation.In this study, we confirmed these previous observations andfurther showed that the CaR agonists, Gd31 and neomycin,as well as high Ca21

o per se could induce chemotaxis ofMC3T3-E1 cells (Fig. 5). Although Quarles et al. failed todetect CaR expression by either RT-PCR or Northern anal-ysis in MC3T3-E1 cells,(17) we have documented using mul-tiple detection methods that MC3T3-E1 cells express bothCaR protein and mRNA that are similar if not identical tothose in the parathyroid gland and kidney (Figs. 1–4), sug-gesting that the CaR might also potentially be involved inthe migration and proliferation of MC3T3-E1 cells.

Although there are other reports failing to show expres-sion of the CaR in human peripheral blood monocytes(25)

and human osteoblast-like SAOS-2 cells,(26) these resultswere based on the single finding of negative results ofRT-PCR using CaR-specific primers, but not on methodsdetecting CaR protein expression, such as immunocyto-chemistry and Western analysis. We found that detection ofCaR transcripts by RT-PCR or Northern blot analysis inMC3T3-E1 cells was, in general, more difficult than iden-tification of receptor protein using immunocytochemistryand Western analysis. Therefore, we believe that negativeresults obtained using the former two methods requiremore cautious interpretation. Since the complete DNAsequence of the mouse CaR has not been determined,Quarles et al. performed RT-PCR of mouse CaR withprimers designed from human CaR. They failed to amplifyCaR-derived products at an annealing temperature of 60°Cin MC3T3-E1 cells.(17) However, this temperature might betoo high for effective priming of the mouse CaR inMC3T3-E1 cells in view of the species mismatch betweenprimers and their DNA template as well as the probablelower expression of CaR mRNA in the osteoblastic cellsrelative to those in the parathyroid and kidney. Indeed, wealso failed to amplify any PCR products even with mouseCaR-derived primers in MC3T3-E1 cells when annealingtemperatures above 50°C were used and found that a tem-perature of 47°C was optimal for successful amplification ofthe CaR-derived product by PCR. Moreover, we found thatordinary Taq DNA polymerase was relatively ineffective foramplifying CaR mRNA from MC3T3-E1 cells, and thatonly “hot start” PCR using a mixture of Taq/Pyrococcusspecies GB-D DNA polymerases, which have proofreadingactivity, resulted in successful amplification. Therefore, RT-PCR of the mouse CaR may require careful selection of theexperimental conditions that are ideal for PCR. In this

study, the possibility of amplification of CaR from genomicDNA was totally eliminated because the primer pair wasdesigned to span one intron of the CaR gene, the primerpair amplified a product of the size expected for a CaR-derived product, and no PCR product was amplified with-out RT.

Similarly, we only successfully detected CaR transcriptsin MC3T3-E1 cells using Northern blot analysis under mod-erately stringent hybridization and washing conditions, uti-lizing a rat CaR-specific cRNA probe and 5 mg of poly(A1)RNA instead of the cDNA probe and 2 mg of poly(A1)RNA that were employed by Quarles et al.(17) The principalbands observed in MC3T3-E1 cells were similar in size totwo minor bands observed on Northern analysis of mousekidney performed on poly(A1) under high stringency con-ditions. We employed two different stringencies withMC3T3-E1 cells and mouse kidney because of the use of aheterologous (e.g., rat) probe and since the relative abun-dance of the CaR transcripts in MC3T3-E1 cells are muchlower than in mouse kidney. The use of high stringencyconditions for hybridization and washing would have greatlyreduced the signal intensity observed in the MC3T3-E1cells relative to that in kidney. The 9.5 and 4.5 kb transcriptspresent in MC3T3-E1 cells are of sizes similar to transcriptsexpressed in both mouse kidney and murine AtT-20cells.(19) However, MC3T3-E1 cells did not express any ofthe 7.5 kb transcript observed in AtT-20 cells(19) or mousekidney, even after exposing the blot for 24 h using phosphor-imager analysis. There are previous instances where therelative ratios of the abundance of CaR transcripts variesfrom organ to organ.(27) Because 4.5 kb transcript encodesthe entire functional CaR protein, the significance of thelarger 9.5 and 7.5 kb transcripts remain uncertain. How-ever, the possibility of organ/cell type specific, post-tran-scriptional regulation of CaR expression as a result ofvariations in the stabilities of the various CaR transcriptscannot be ruled out.

Another explanation for the difference between our re-sults and those of Quarles et al. could be variations in theclones of MC3T3-E1 cells that were employed in the twostudies, and that some clones show little or no expression ofthe CaR. However, this explanation appears less likely be-cause we have confirmed the existence of the CaR in thetwo different batches of MC3T3-E1 cells, both of whichhave the capacity to differentiate to mature osteoblastsafter prolonged culture, as assessed by Western blot anal-ysis and RT-PCR using intron-spanning primers.

The CaR is known to share amino acid sequence homol-ogy and topological similarity with the mGluRs,(28) whichare G protein–coupled receptors present in the centralnervous system that respond to glutamate, the major exci-tatory neurotransmitter in the brain. Thus, it was importantto consider the possibility of our anti-CaR antiserum cross-reacting with mGluRs in MC3T3-E1 cells. We comparedthe amino acid sequences of the CaR-derived FF-7 peptide,to which the antiserum, 4637, that was used in this study wasraised, and the corresponding regions of the mGluRs, andfound that they shared 40–53% homologies to one another.These values may well not represent sufficient degrees ofsimilarity to produce cross-reactivity of our anti-CaR anti-

1536 YAMAGUCHI ET AL.

serum with the mGluRs. We failed to find any evidence inthe published literature describing the expression ofmGluRs in MC3T3-E1 cells or other osteoblast-like cells.Moreover, Western analysis in this study showed that ananti-mGluR1 antiserum failed to detect mGluR protein inMC3T3-E1 cells. Moreover, we have observed thatHEK293 cells transfected with the cDNA for mGluR1 werenegative on Western analysis using anti-CaR antiserum,4637, despite being positive on Western analysis using ananti-mGluR1 antiserum(18) (M. Bai and E. M. Brown,manuscript in preparation), indicating that it is unlikely thatthe anti-CaR antiserum could detect a mGluR even if thecells expressed it. Thus, we consider it unlikely that theprotein in MC3T3-E1 cells recognized by our anti-CaRantiserum represents mGluR(s) instead of CaR.

Recently, Hinson et al. have reported that nucleotidesequences of putative CaR-related receptors (Casr-rs) wereidentified in mouse genomic libraries by PCR.(29) The de-duced protein sequence of one of these putative receptors(Casr-rs1) was 63% similar and 40% identical to the CaRover the available transmembrane region. However, as withthe mGluRs, these levels of identity in predicted proteinsequences may well not be large enough to result in cross-reactivity of our anti-CaR antiserum with these relatedprotein. Moreover, although this CaR-related nucleotidesequence was initially identified in MC3T3-E1 cells by RT-PCR and was used as a probe to screen mouse genomiclibraries to identify other related sequences, it could not beidentified in subsequent analyses of mouse tissues, includ-ing MC3T3-E1 cells, by RT-PCR.(29) Additional studies arenecessary to determine whether these CaR-related nucleo-tide sequences are actually expressed as mature proteins inMC3T3-E1 cells using specific antisera raised to their pre-dicted protein sequences.

The role of the CaR in the control of cellular prolifera-tion has not been clear until recently, when the receptor hasbeen conclusively shown to be involved in the stimulation ofcell proliferation by CaR agonists. Mailland et al. reportedthat CaR agonists stimulate the proliferation of CCL39hamster fibroblasts transfected with the CaR,(30) and Hoffet al. reported that transfection of NIH-3T3 cells with ahuman CaR cDNA harboring an activating mutation in-duced cellular transformation and proliferation.(31) In ad-dition, McNail et al. have recently demonstrated that rat-1fibroblasts express an endogenous CaR and that high Ca21

o

stimulates the proliferation of these cells through a prolif-erative pathway involving CaR-mediated activation of SRCkinase and ERK1.(32) Moreover, several recent studies haveclarified the signaling mechanisms underlying Ca21

o-stim-ulated chemotaxis and proliferation of MC3T3-E1 cells.Godwin and Soltoff demonstrated that inhibition of theactivation of either G protein or PLC blocked nearly all ofthe chemotaxis in MC3T3-E1 cells, whereas the inhibitionof protein kinase C (PKC) or phosphoinositide-3-kinase didnot.(4) These results suggested that Ca21

o-stimulated che-motaxis of this cell line is linked to the activation of Gprotein and PLC. Quarles et al. showed that a variety ofpolyvalent cations, including Al31, Gd31, Ca21 and neomy-cin, stimulated DNA synthesis in MC3T3-E1 cells througha mechanism coupled to the activation of G protein and

PKC.(6) Since the CaR activates PLC in a G protein–dependent manner, thereby raising the levels of IP3 and thecytosolic calcium concentration and activating PKC,(11,33)

these findings do not contradict the hypothesis that the CaRis involved in the proliferation and chemotaxis ofMC3T3-E1 cells induced by high Ca21

o and other CaRagonists. However, one study reported that neither Gd31

o

nor Ca21o increased either inositol phosphate formation or

the cytosolic calcium concentration in MC3T3-E1 cells.(34)

Thus, additional studies are needed to document furthercausal relationships between expression of the CaR, itssignal transduction pathways and the control of chemotaxisand cell proliferation by CaR agonists in this osteoblasticcell line.

Although the CaR was first cloned from the parathyroidand kidney,(11–13) bone is also intimately involved in sys-temic calcium ion homeostasis.(35) In this study, usingMC3T3-E1 cells as a model, we show that osteoblasts,which play an important role in bone remodeling within theskeleton, express both CaR protein and mRNA. Our resultssuggest that this receptor may be involved in importantphysiological responses of these cells, such as chemotaxisand proliferation after stimulation by Ca21

o. These eventsare observed at the beginning of the bone formation phaseof skeletal remodeling in vivo, suggesting that the CaRcould potentially play a key role in the functions of bonecells within the bone/bone marrow microenvironment.

ACKNOWLEDGMENTS

The authors gratefully acknowledge generous grant sup-port from the following sources: The Mochida MemorialFoundation Grant for Medical and Pharmaceutical Re-search (to T.Y.), The Yamanouchi Foundation Grant forResearch on Metabolic Disorders (to T.Y.), NPS Pharma-ceuticals, Inc. (to E.M.B.), The St. Giles Foundation (toE.M.B.), USPHS Grants DK-41415, DK-48330, and DK-52005 (to E.M.B.), and the National Space Bioscience Re-search Institute (NSBRI) (to E.M.B.).

REFERENCES

1. Baron R 1996 Anatomy and ultrastructure of bone. In: FavusMJ (ed.) Primer on the Metabolic Bone Diseases and Disor-ders of Mineral Metabolism, 3rd Ed. Lippincott-Raven, Phil-adelphia, PA, U.S.A., pp. 3–10.

2. Silver IA, Murrills RJ, Etherington DJ 1988 Microelectrodestudies on the acid microenvironment beneath adherent mac-rophages and osteoclasts. Exp Cell Res 175:266–276.

3. Sugimoto T, Kanatani M, Kano J, Kaji H, Tsukamoto T,Yamaguchi T, Fukase M, Chihara K 1993 Effects of highcalcium concentration on the functions and interactions ofosteoblastic cells and monocytes and on the formation of os-teoclast-like cells. J Bone Miner Res 8:1445–1452.

4. Godwin SL, Soltoff SP 1997 Extracellular calcium and platelet-derived growth factor promote receptor-mediated chemotaxisin osteoblasts through different signaling pathways. J BiolChem 272:11307–11312.

5. Sugimoto T, Kanatani M, Kano J, Kobayashi T, Yamaguchi T,Fukase M, Chihara K 1994 IGF-I mediates the stimulatoryeffect of high calcium concentration on osteoblastic cell pro-liferation. Am J Physiol 266:E709–E716.

CaR AND MOUSE OSTEOBLASTSIC CELL LINE 1537

6. Quarles LD, Hartle JE II, Middleton JP, Zhang J, Arthur JM,Raymond JR 1994 Aluminum-induced DNA synthesis in os-teoblasts: Mediation by a G-protein coupled cation sensingmechanism. J Cell Biochem 56:106–117.

7. Kodama H, Amagai Y, Ando H, Yamamoto S 1981 Establish-ment of a clonal osteogenic cell line from newborn mousecalvaria. Jpn J Oral Biol 23:899–901.

8. Sudo H, Kodama H, Amagai Y, Yamamoto S, Kasai S 1983 Invitro differentiation and calcification in a new clonal osteo-genic cell line derived from newborn mouse calvaria. J CellBiol 96:191–198.

9. Quarles LD, Yohay DA, Lever LW, Canton R, Wenstrup RJ1992 Distinct proliferative and differentiated stages of murineMC3T3-E1 in culture: An in vitro model of osteoblast devel-opment. J Bone Miner Res 7:683–692.

10. Yohay DA, Zhang J, Thrailkill KM, Arthur JM, Quarles LD1994 Role of serum in the developmental expression of alka-line phosphatase in MC3T3-E1 osteoblasts. J Cell Physiol 158:467–475.

11. Brown EM, Gamba G, Riccardi D, Lombardi M, Butters R,Kifor O, Sun A, Hediger MA, Lytton J, Hebert SC 1993Cloning and characterization of an extracellular Ca21-sensingreceptor from bovine parathyroid. Nature 366:575–580.

12. Garrett JE, Capuano IV, Hammerland LG, Hung BCP, BrownEM, Hebert SC, Nemeth EF, Fuller F 1995 Molecular cloningand characterization of the human parathyroid calcium recep-tor. J Biol Chem 270:12919–12925.

13. Riccardi D, Park J, Lee WS, Gamba G, Brown EM, Hebert SC1995 Cloning and functional expression of a rat kidney extra-cellular calcium/polyvalent cation-sensing receptor. Proc NatlAcad Sci USA 92:131–135.

14. Garrett JE, Tamir H, Kifor O, Simin RT, Rogers KV, MithalA, Gagel RF, Brown EM 1995 Calcitonin-secreting cells of thethyroid gland express an extracellular calcium-sensing receptorgene. Endocrinology 136:5202–5211.

15. Pollak MR, Brown EM, Chou YH, Hebert SC, Marx SJ, Stein-mann B, Levi T, Seidman CE, Seidman JG 1993 Mutations inthe human Ca21-sensing receptor gene cause familial hypocal-ciuric hypercalcemia and neonatal severe hyperparathyroidism[see comments]. Cell 75:1297–1303.

16. Pollak MR, Brown EM, Estep HL, McLaine PN, Kifor O, ParkJ, Hebert SC, Seidman CE, Seidman JG 1994 Autosomaldominant hypocalcaemia caused by a Ca21-sensing receptorgene mutation. Nat Genet 8:303–307.

17. Quarles LD, Hartle JE II, Siddhanti SR, Guo R, Hinson TK1997 A distinct cation-sensing mechanism in MC3T3-E1 osteo-blasts functionally related to the calcium receptor. J BoneMiner Res 12:393–402.

18. House MG, Kohlmeier L, Chattopadhyay N, Kifor O,Yamaguchi T, LeBoff MS, Glowacki J, Brown EM 1997 Ex-pression of an extracellular calcium-sensing receptor in humanand mouse bone marrow cells. J Bone Miner Res 12:1959–1970.

19. Emanuel RL, Adler GK, Kifor O, Quinn SJ, Fuller F, KrapchoK, Brown EM 1996 Calcium-sensing receptor expression andregulation by extracellular calcium in the AtT-20 pituitary cellline. Mol Endocrinol 10:555–565.

20. Maeda K, Nakai M, Maeda S, Kawamata T, Yamaguchi T,Tanaka C 1997 Possible different mechanism between amy-loid-beta (25–35)- and substance P-induced chemotaxis of mu-rine microglia. Gerontology 43 (Suppl 1):11–15.

21. Bai M, Quinn S, Trivedi S, Kifor O, Pearce SHS, Pollak MR,Krapcho K, Hebert SC, Brown EM 1996 Expression and char-acterization of inactivating and activating mutations in thehuman Ca21

o-sensing receptor. J Biol Chem 271:19537–19545.

22. Kifor O, Moore FD, Wang P, Goldstein M, Vassilev P, Kifor I,Hebert SC, Brown EM 1996 Reduced immunostaining for theextracellular Ca21

o-sensing receptor in primary and uremicsecondary hyperparathyroidism. J Clin Endocrinol Metab 81:1598–1606.

23. Ho C, Conner DA, Pollak MR, Ladd DJ, Kifor O, Warren HB,Brown EM, Seidman JG, Seidman CE 1995 A mouse model ofhuman familial hypocalciuric hypercalcemia and neonatal se-vere hyperparathyroidism. Nat Genet 11:389–394.

24. Puzas JE 1996 Osteoblast cell biology-lineage and functions.In: Favus MJ (ed.) Primer on the Metabolic Bone Diseases andDisorders of Mineral Metabolism, 3rd Ed. Lippincott-Raven,Philadelphia, PA, U.S.A., pp. 11–16.

25. Bornefalk E, Ljunghall S, Lindh E, Bengtson O, JohanssonAG, Ljunggren O 1997 Regulation of interleukin-6 secretionfrom mononuclear blood cells by extracellular calcium. J BoneMiner Res 12:228–233.

26. Gunnet JW, Zhou L, Demarest KT 1997 Comparison of thecalcium-sensing receptor in human SK-N-MC, TT, HEK 293,and SAOS-2 osteosarcoma cells. J Bone Miner Res 12 (Suppl1):S327.

27. Chattopadhyay N, Cheng I, Rogers K, Riccardi D, Hall A, DiazR, Hebert SC, Soybel DI, Brown EM 1998 Identification andlocalization of extracellular Ca21-sensing receptor in rat intes-tine. Am J Physiol 274:G122–G130.

28. Nakanishi S 1992 Molecular diversity of glutamate receptorsand implications for brain function. Science 258:597–603.

29. Hinson TK, Damodaran TV, Chen J, Zhang X, Qumsiyeh MB,Seldin MF, Quarles LD 1997 Identification of putative trans-membrane receptor sequences homologous to the calcium-sensing G-protein–coupled receptor. Genomics 45:279–289.

30. Mailland M, Waelchli R, Ruat M, Boddeke HGWM, SeuwenK 1997 Stimulation of cell proliferation by calcium and acalcimimetic compound. Endocrinology 138:3601–3605.

31. Hoff AO, Cote GJ, Fritsche HA Jr, Schultz P, Khorana S,Parthasarathy R, Gagel RF 1997 An activating mutation of thecalcium-sensing receptor creates a new oncogene. J BoneMiner Res 12 (Suppl 1):S143.

32. McNeil SE, Hobson SA, Nipper V, Rodland KD 1998 Func-tional calcium-sensing receptors in rat fibroblasts are requiredfor activation of SRC kinase and mitogen-activated proteinkinase in response to extracellular calcium. J Biol Chem 273:1114–1120.

33. Chattopadhyay N, Mithal A, Brown EM 1996 The calcium-sensing receptor: A window into the physiology and pathologyof mineral ion metabolism. Endocr Rev 17:289–307.

34. Hartle JE II, Prpic V, Siddhanti SR, Spurney RF, Quarles LD1996 Differential regulation of receptor-stimulated cyclic aden-osine monophosphate production by polyvalent cations inMC3T3-E1 osteoblasts. J Bone Miner Res 11:789–799.

35. Brown EM 1991 Extracellular Ca21 sensing, regulation ofparathyroid cell function, and role of Ca21 and other ions asextracellular (first) messengers. Physiol Rev 71:371–411.

Address reprint requests to:Toru Yamaguchi, M.D.

Endocrine-Hypertension DivisionBrigham and Women’s Hospital

221 Longwood AvenueBoston, MA 02115 U.S.A.

Received in original form December 9, 1997; in revised form April20, 1998; accepted May 27, 1998.

1538 YAMAGUCHI ET AL.