Utilization of Bone Morphogenetic Protein Receptors During Chondrocyte Maturation

Utilization of Bone Morphogenetic Protein ReceptorsDuring Chondrocyte Maturation

SUSAN W. VOLK, MARINA D’ANGELO, DAVID DIEFENDERFER, and PHOEBE S. LEBOY

ABSTRACT

Cartilage from the upper, cephalic portion of embryonic chick sternums undergoes hypertrophy, while thelower, caudal portion of the sternum remains as cartilage. Bone morphogenetic proteins (BMPs) induce typeX collagen (colX) in cultured upper but not lower sternal chondrocytes (LSCs). We have examined theutilization of BMP receptors (BMPRs) by upper sternal chondrocytes (USCs) and LSCs both by analyzingreceptor expression and by overexpressing mutant BMPRs. Reverse-transcription polymerase chain reaction(RT-PCR) analyses indicate that both upper and lower chondrocytes produce messenger RNA (mRNA) for allthree receptors: BMPR type IA (BMPR-IA), BMPR type IB (BMPR-IB), and BMPR type II (BMPR-II).Infection of USC with retroviral vectors expressing constitutively active (CA) BMPRs showed that CA-BMPR-IB, like exogenous BMP-4, induced both colX mRNA and elevated alkaline phosphatase (AP), while CA-BMPR-IA was markedly less potent. However, expression of activated receptors in LSC cultures resulted inonly minimal induction of hypertrophic markers. Consistent with the results seen for CA receptors, dominantnegative (DN) BMPR-IB blocked BMP-induced hypertrophy in USCs more effectively than DN-BMPR-IA.These results imply that the major BMPR required for BMP induction of chondrocyte hypertrophy isBMPR-IB, and that difference between permanent and prehypertrophic chondrocytes is not caused byabsence of receptors required for BMP signaling. (J Bone Miner Res 2000;15:1630–1639)

Key words: hypertrophic, bone morphogenetic protein, receptors, chondrocytes, cartilage

INTRODUCTION

CHONDROCYTES IN regions destined for endochondral os-sification undergo a series of changes including hyper-

trophy, release of extracellular matrix vesicles, and alter-ations in production of extracellular matrix proteins,followed by mineralization. Chondrocyte maturation can bepromoted by bone morphogenetic proteins (BMPs), thecomponent of demineralized bone that has long been knownto induce ectopic bone formation.(1) In the last decade,several BMP cell surface receptors have been identified andcloned.(2,3) Like other members of the transforming growthfactor b (TGF-b) superfamily, functional BMP receptors(BMPRs) are comprised of heterodimers or heterotetramersof type I and type II receptors.(4–6) Ligand binding to the

cell surface receptor complex induces phosphorylation oftype I receptor by type II receptor. BMP is thought to haveone type II receptor (BMPR-II) and two type I receptors(BMPR-IA and BMPR-IB), although BMPs also may beable to signal through activin receptors.(7,8) Activated BMPand other TGF-b superfamily type I receptors have beenshown to phosphorylate members of a family of intracellu-lar signaling molecules known as Smads.(9,10) Recent stud-ies suggest that Smads exert transcriptional effects in com-bination with other transcription factors, and that thesecomplexes bind to several discrete response elements.(11–14)

Recently, we have shown that BMP-2, -4, and -7 are allcapable of inducing chondrocyte maturation in chondro-cytes from chick embryos that are destined for hypertro-phy.(15) Cultures of prehypertrophic upper (cephalic) ster-

Department of Biochemistry, School of Dental Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, U.S.A.

JOURNAL OF BONE AND MINERAL RESEARCHVolume 15, Number 8, 2000© 2000 American Society for Bone and Mineral Research

1630

nal chondrocytes (USC) derived from day 15 embryos(15dUSC) that were treated with BMPs expressed hyper-trophic markers such as type X collagen (colX) messengerRNA (mRNA) and high levels of alkaline phosphatase (AP)while BMP type II collagen (colII) mRNA was down-regulated.(15,16) In contrast, chondrocytes derived from theimmature, lower (caudal) region of the sternum (LSC) wereunaffected by BMP exposure, expressing neither colX norAP mRNA. In this article, we examine the role of the twotype I receptors, BMPR-IA (also known as ALK-3) andBMPR-IB (ALK-6), in the two chondrocyte populationsusing the avian retrovirus RCAS to express both constitu-tively active (CA) and dominant negative (DN) forms of thereceptors. We provide evidence that, while either CA-BMPR-IA or -IB induces chondrocyte hypertrophy inUSCs, the BMPR-IB is consistently more potent thanBMPR-IA. Similarly, while DN forms of both type I recep-tors inhibit induction of hypertrophic markers by exogenousBMP, the DN-IB receptor blocks BMP effects more effec-tively than DN-IA. Immature LSCs, which do not maturewith exogenous BMPs, express relatively high levels of theBMP-binding protein, noggin, and show a small but detect-able response if provided with a CA type I receptor. We alsoprovide evidence that both caudal and cephalic sternal chon-drocytes possess mRNA for BMP-2 as well as the threeBMPRs, BMP-II, -IA, and -IB.

MATERIALS AND METHODS

Cell culture

Cells were isolated from the upper (cephalic) and lower(caudal) one-third portions of sternums from 15-day and17-day chick embryos (B & E Eggs, Stevens, PA, U.S.A.)by digestion for 3.5 h at 37°C, 5% CO2 in calcium andmagnesium-free Hanks’ balanced salt solution (CMF-HBSS) containing 0.6 mg/ml collagenase and 0.04% tryp-sin. Cartilage isolated from the 17-day-old upper sternumswas divided into core and periphery before digestion; thecore region of day 17 sternums contains hypertrophic chon-drocytes whereas the peripheral region contains primarilyprehypertrophic chondrocytes.(17) Cells were initially resus-pended in a complete medium composed of high glucoseDulbecco’s modified Eagle’s medium (DMEM) with 10%NuSerum IV (NuS; Collaborative Biomedical Products/Becton-Dickinson, Bedford, MA, U.S.A.) and 100 U/mlpenicillin/streptomycin (Pen/Strep) and cultured for 5 daysas described previously.(15) The floating chondrocytes werethen replated at 2.4 to 4.83 104 cells/cm2 in completemedium supplemented with 4 U/ml hyaluronidase (HA) topromote attachment.(18) After 24 h, cultures were trans-ferred to serum-free (SF) conditions to assess more accu-rately the direct effects of BMP. Cultures were washedtwice in CMF-HBSS before switching to SF conditions. SFmedium contained DMEM (with Pen/Strep and HA) sup-plemented with 10 PM. triiodothyronine (Sigma, St. Louis,MO, U.S.A.), 60 ng/ml insulin (bovine pancreatic; Sigma),and 1mM cysteine. Recombinant human BMP-4 (kindlyprovided by Genetics Institute, Cambridge, MA, U.S.A.)was added to cultures where appropriate at a final concen-

tration of 30 ng/ml. Ascorbate-supplemented cultures con-tained 0.15 mM ascorbate phosphate (Wako Chemicals,Richmond, VA, U.S.A.). Cells were harvested 48 h afterswitching to SF medium for mRNA or AP assays.

Retroviral (RCAS) infection of chondrocytes withmutated BMPRs

CA and DN type I BMPRs were expressed in chondro-cytes using recombinant retroviral (RCAS) constructs pre-pared by Dr. Lee Niswander, Sloan-Kettering Institute(New York, NY, U.S.A.).(19,20) Viral stocks were preparedin chick embryo fibroblasts. The fibroblasts were cultured toconfluence in M199 medium (Gibco BRL, Grand Island,NY, U.S.A.) supplemented with 5% fetal bovine serum(FBS; Atlanta Biological, Norcross, GA, U.S.A.), 1% chickserum (Sigma), 2.95 mg/ml tryptose phosphate broth, and100 U/ml penicillin/streptomycin (Pen/Strep). Twelve hoursbefore transfection, fibroblasts were split 1:4 and 1 hourbefore transfection, fibroblasts were refed with DMEM con-taining 10% fetal calf serum (FCS) and 100 U/ml Pen/Strep.12.5 mg plasmid was transfected into each 100mm culturedish using the calcium-phosphate precipitation method. Ap-proximately 6 h later, cells were washed twice with CMF-HBSS before refeeding with fresh media supplemented with15% fibroblast-conditioned medium. Medium from the fi-broblast cultures was collected on days 6, 8, and 10 andused to infect cultured chondrocytes.

The 15-day embryonic sternal chondrocytes used for ret-roviral infections were plated at 2.43 104 cells/cm2 inDMEM supplemented with 10% NuS, HA, and Pen/Strep asdescribed above. The 17-day USC (17dUSC) were plated at4.8 3 104 cells/cm2 to accommodate for their reducedproliferation. Medium containing unconcentrated virus wasadded to cultures at the time of plating (10% total volume).The infection was allowed to spread throughout the chon-drocyte cultures for 6 days before switching the cultures toSF medium with or without BMP-4 and ascorbate phos-phate for an additional 48 h.

Immunofluorescence

To determine efficiency of virus infection, chondrocytecultures were harvested by trypsinization and replated onpoly-L-lysine–coated dishes at a density of 1.03 106 cellsper 35 mm well. After 10 minutes, cells were washed twiceand fixed with 70% ethanol. Cells were preincubated withblocking serum (10% normal goat serum in phosphate-buffered saline [PBS]) for 1 h before incubation with AMV-3C2 (2.0mg immunoglobulin/ml). This mouse monoclonalantibody, obtained from the Developmental Studies Hybrid-oma Bank (Iowa City, IA, U.S.A.), reacts with the gagprotein of avian myeloblastosis virus and therefore recog-nizes cells harboring RCAS virus. A rhodamine-conjugatedgoat anti-mouse antibody (Southern Biotechnology Associ-ates, Inc., Birmingham, AL, U.S.A.) was used to visualizeRCAS virus positive cells under an Olympus fluorescentmicroscope (Olympus America, Melville, NY, U.S.A.). Thepercentage of infected cells was determined by counting thenumber of brightly stained cells per 100 cells.

1631BMP RECEPTORS FOR CHONDROCYTE HYPERTROPHY

Isolation of total RNA and Northern blots

RNA from cultured cells was prepared by guanidineisothiocyanate and guanidine extraction as described previ-ously.(18) RNA samples (8mg) were denatured by glyoxy-lation, electrophoresed on agarose gels for 2 h at 7V/cm2,and transferred onto nylon membrane (Gene Screen Plus;Du Pont, Wilmington, DE, U.S.A.). Relative amounts ofspecific mRNAs were determined by hybridizing Northernblots to [32P]-labeled chick riboprobes for chick colX andcolII.(18) The extent of hybridization to blots was quanti-tated with a Molecular Dynamics FluorImager (MolecularDynamics, Sunnyvale, CA, U.S.A.). Northern blots werestained with 0.02% methylene blue to ensure equal loadingof RNA.

Reverse-transcription polymerase chain reaction

RNA from cultured cells was prepared as describedabove, except for RNA isolated directly from 17-day em-bryonic sternums. Three regions were isolated from thesesternums: the inner core of the upper sternum (USCc),which have already matured and express type X mRNA, theperiphery of the upper sternum containing proliferating cellsdestined for hypertrophy (USCp), and immature cells of thelower sternum (LSC). Total RNA was reverse-transcribedby Superscript reverse transcriptase (Gibco BRL) with anoligo-dT primer. Subsequent amplification was carried outusing Ready-To-Go PCR Beads (Pharmacia Biotech, Pisca-taway, NJ, U.S.A.) and gene-specific primers for 30 cyclesunder the following conditions: 95°C for 10 s, 60°C for 1minute, and 72°C for 2 minutes. Primer sequences foramplification of BMPs and BMPRs were obtained fromEnomoto-Iwomoto et al.(21) and are as follows: 59-CAGCTT CCA CCA CGA AGA AG-39 and 59-TAG GTT GTCCGT GTG CAA TC-39 for chick BMP-2, generating a341-base pair (bp) fragment; 59-CAC CAG GCA CAGACT CAT CA-39 and 59-AGG TAG AGC ATG GAG ATGGC-39 for chick BMP-4, generating a 425-bp fragment;59-ACC TTG GCC TGC AGC TAT CT-39 and 59-CTCTGG AGC GAT GAT CCA GT-39 for chick BMP-7, gen-erating a 332-bp fragment; 59-AGC GAT TGC TTG GAGCCT ATC T-39 and 59-AGC TGG CTT CTT CTG TGGTGA A-39 for chick BMPR-IA, generating a 792-bp frag-ment; 59-ATG CTG CAC AGG CCA AGA TTA C-39 and59-CCA AGA GCC TGT GCC TTT AAT G-39 for chickBMPR-IB, generating a 536-bp fragment; and 59-GGTCGA TAC GGA GCA GTG TAC A-39 and 59- CTG CTCCTT CAA GCA CTT CTG G-39 for chick BMPR-II, gen-erating a 541-bp fragment. A 118-bp RT-PCR product forchick noggin was obtained using the primers 59-ATC TAATCG AGC ACC CGG AC-39 and 59-GGC AGG GAA ATAGCC GTA AAG-39. A 292-bp product was obtained forhuman BMPR-IA using the following gene-specific prim-ers: 59-GCATAACTAATGGACATTGCT-39 and 59-GCAGCT GGAGAAGATGATCATAGC-39. RT-PCR us-ing the chick B-actin primers 59-CCCTGAACCCC-AAAGCCAAC-39 and 59-CCTGCTCGAAATCCAGTG-CG-39 produced a 351-bp product. All RT-PCR reactionswere carried out with two independent isolates of RNA, and

PCR products were analyzed after 22, 24, 26, 28, and 30cycles. Levels of PCR product were analyzed using a KodakImage Station 440 (Eastman Kodak, Rochester, NY,U.S.A.).

AP assays

Cultures were rinsed twice with CMF-HBSS and ex-tracted with 0.15 M Tris, pH 9.0, 0.1 mM ZnCl2, 0.1 mMMgCl2, and 1% Triton X-100 for 30 minutes at 37°C. A 10-to 100-ml aliquot of the solubilized cell layer extract wasreacted with 7.5 mMp-nitrophenol phosphate (Sigma 104)in 1.5 M Tris (pH 9.0), 1 mM ZnCl2, and 1 mM MgCl2 atroom temperature and the change in absorbance (A410nm)was measured over a 6-minute time period. Enzyme levelswere expressed as nanomoles ofp-nitrophenol per minuteper well assuming 1 A410 5 64 nmol of product. Chondro-cytes for AP assays were plated in triplicate for each con-dition examined.

RESULTS

Expression of BMPs, BMPRs, and noggin inembryonic chicken sternal chondrocytes

We have previously reported that BMP-2, -4, and -7 areall capable of inducing maturation of the prehypertrophicchondrocytes (15dUSC) from the upper (cephalic) sternumof day 15 chick embryos, and that BMP-4 is more effectivethan the other two BMPs.(16) In contrast, none of theseBMPs were able to induce maturation of chondrocytes fromthe lower portion of the sternum of day 15 chick embryosalteration (15dLSC), which does not undergo hypertrophyand endochondral ossification during development. To ex-plore whether the lack of response to exogenous BMPs byLSC was caused by absence of BMPRs, we assayed recep-tor mRNA by RT-PCR. Cartilage was isolated from threeregions of the 17-day chick embryonic sternum, as de-scribed in the Materials and Methods section. These repre-sent three different stages of chondrocyte development:hypertrophic cells from the ossifying core of the uppersternum (17dUSCc), prehypertrophic cells from the periph-ery of day 17 upper sternums, (17dUSCp), and immatureLSCs from day 17 chick embryos alteration (17dLSC).Total RNA was extracted and subjected to RT-PCR withsequence-specific primers for chick BMPR-II and the twoBMPR-I receptors. After 28 cycles of amplification, it wasevident that mRNAs for all three receptor types werepresent in both USCs and LSCs; however, levels ofBMPR-II were consistently lower in RNA extracted fromthe hypertrophic USCc (Fig. 1A).

We also examined whether BMP-2 and -4 were producedby the three regions of day 17 sternal cartilage. Althoughcorrectly sized PCR products were detected for each of thethree BMP species, only BMP-2 was expressed consistentlyin all three locations, with hypertrophic d17USCc showeddiminished BMP-2 mRNA (Fig. 1A). BMP-4 mRNA levelswere lower than those for BMP-2, and variable results wereobserved with different RNA preparations (data not shown).Because our previous studies(16) showed that prehypertro-

1632 VOLK ET AL.

phic chondrocytes were most responsive to BMP-4, it isplausible that expression of this BMP is highly regulatedand varies depending on the state of chondrocyte matura-tion.

Noggin, a BMP-binding protein capable of inhibitingBMP-2 and -4 signaling, was strongly expressed in thenonhypertrophying lower sternal region; however nogginmRNA was barely detectable in the prehypertrophic carti-lage at the periphery of upper sternal cartilage (Fig. 1B).These results are consistent with the hypothesis that highlevels of noggin suppress BMP-induced hypertrophy inlower sternal cartilage, and diminished noggin in prehyper-trophic USCs permits BMP-induced hypertrophy.

Retroviral expression of DN and CA BMPRs in chicksternal chondrocytes

Enomoto-Iwamoto et al.(21) reported that day 17 chickembryonic sternal chondrocytes infected with retrovirus ex-pressing a truncated DN BMPR-II showed dedifferentiationof the chondrocyte phenotype, increased proliferation, andshowed a loss of response to BMP-2 after 5–7 days. Lesssevere effects on the chondrocyte phenotype were seenwith overexpression of truncated DN-BMPR-IB, whereasBMPR-IA produced little effect. These results suggestedthat BMP signaling may be required for maintaining thechondrocyte phenotype, but the data were not capable ofdefining which BMPRs are involved in hypertrophy. Ourprevious studies have shown that, under SF conditions,d15USCs show expression of colX and AP within 48 h inthe presence of BMPs.(15) We therefore analyzed BMP typeI receptor function under these conditions, by expressingCA-BMPR-IA and -IB in 15dUSCs in the presence andabsence of exogenous BMP. Similarly, DN-BMPR-IA and-IB were expressed to test whether they were capable ofblocking BMP-induced hypertrophy. The CA-BMPR-IAand -IB also were used to determine whether overexpressionof these mediators could induce hypertrophy in immature15dLSCs, which are resistant to the maturation effects ofexogenous BMPs.

The retroviral expression vector RCAS-A was used tooverexpress both CA and DN mutants of the BMPR-IA and-IB.(19) The DN mutant receptors contained a single aminoacid substitution (K3R) within the ATP binding site of thetype I receptor,(19) whereas the CA mutants contained asingle amino acid substitution (Q3D) within the GS box ofthe type I receptor.(20,22) These single amino acid substitu-tions either disrupt or enhance the ability of the type Ireceptors to signal without destroying their ability to inter-act with either the type II receptor or downstream media-tors.(19,20) Chondrocytes were incubated with medium con-taining virus for 6 days before switching the cultures to SFconditions with or without BMP-4 and ascorbate phosphate,as described in the Materials and Methods section. Cellswere harvested 48 h later for Northern analysis of colII andcolX mRNAs and for assays of AP activity. Infection withthe RCAS retrovirus was monitored by immunofluores-cence using monoclonal antibody directed against the ret-roviral gag protein. Infection with the four different RCASconstructs containing the mutant BMPRs and the control(empty) RCAS vector all showed 50–80% infectivity. Thevariation in percent of infected cells reflected differencesbetween experiments, with no consistent differences notedbetween different BMPR mutants; however, LSCs consis-tently showed a slightly higher rate of RCAS infection thanthe USC. This cell-specific difference presumably reflectsthe requirement of RCAS virus for proliferating cells andthe increased proliferative rate of the LSCs as comparedwith the more mature USC.(23)

The relative overexpression of each subtype of mutantBMPR in chondrocytes was examined by RT-PCR of RNAfrom virus-infected cultures (Fig. 2). Primers designed torecognize the chick BMPR-IA amplified both the endoge-nous chick sequences and the mutant human IA sequences,

FIG. 1. RT-PCR for BMP receptors, BMP-2, and nogginin three chick sternal chondrocyte populations. Total RNAwas prepared from hypertrophic chondrocytes isolated fromthe core of the upper sternums of 17-day embryos(17dUSC-c); prehypertrophic chondrocytes isolated fromeither the periphery of 17-days upper sternums (17dUSCp);and immature chondrocytes of the 17-day lower sternums(17dLSC). RNA was subjected to RT-PCR as described inthe Materials and Methods section, using 22–30 cycles ofamplification. Amplified products were analyzed by 1.5%agarose gel electrophoresis. (A) PCR products after 28cycles of amplification. (B) Quantitation of levels of nogginPCR product with increasing cycles of amplification.

1633BMP RECEPTORS FOR CHONDROCYTE HYPERTROPHY

whereas primers designed for human IA recognized themutated human IA receptor but did not amplify the endog-enous chick IA. As seen in Fig. 2, infection with viruscontaining mutated human IA receptor led to increasedlevels of total CA-IA and DN-IA sequences, whereas infec-tion with virus expressing the mutated chick IB receptorscaused increased expression of BMPR-IB mRNA. The levelof expression of mutant receptor mRNAs was 1- to 2-foldthose of the endogenous BMPR, whereas BMPR-II wasdetected in all samples at similar amounts. No consistentchanges in cellular morphology were observed in chondro-cyte cultures infected either with control RCAS virus orRCAS encoding mutated BMPRs, suggesting that viral in-fection had not induced chondrocyte dedifferentiation.

Effects of retroviral expression of CA- and DN-BMPRson AP induction in sternal chondrocytes

AP assays were performed on cultures of 15dUSC in-fected with RCAS viruses expressing the four mutant BMPtype I receptors as well as control virus. AP activity incultures infected with the control RCAS virus under SFconditions, with or without BMP, was similar to that seenwith comparably treated uninfected cultures (p . 0.05; datanot shown). Our laboratory has previously shown the syn-ergistic action of ascorbate and BMP-2 on AP induction inUSC cultured under SF conditions.(15) As shown in Fig. 3,when 15dUSCs were infected with control RCAS and

treated with BMP-4, AP activity increased approximately20-fold and ascorbate induced AP activity approximately2-fold compared with SF values, whereas the combinationof BMP and ascorbate lead to a dramatic increase of over100-fold. Expression of either CA-IA or CA-IB BMPR(Fig. 3A) was able to induce AP activity in 15dUSCscultured in SF medium in the absence of exogenous BMP,but the IB receptor was more effective. This response to theconstitutive active receptors could be augmented further bythe addition of BMP, implying that the 20–50% of cellswhich had not been infected with virus were still capable ofresponding to exogenous BMP. The CA receptors showedthe same synergistic effect with ascorbate as seen withexogenous BMP. In the presence of BMP1 ascorbate, bothIA and IB receptor expression increased AP levels over thatseen with virus control, but overexpression of CA-IB was

FIG. 2. RT-PCR for BMP receptors in RCAS-infected15-day upper and LSCs. 15dUSC and LSC were infectedwith empty RCAS vector (RCAS) or with RCAS expressingCA- or DN-BMPRs for 6 days. Total RNA was preparedand RT-PCR was performed (30 cycles) using gene-specificprimers for chicken 95-actin, chicken type II BMPR(BMPR-II), chicken (Ch) type IA BMPR (recognizes bothendogenous and RCAS-expressed mutant IA receptor), hu-man (Hu) type IA BMPR (recognizes only RCAS-expressedCA-IA or DN-IA), and the chicken type IB receptor (whichrecognizes both endogenous and RCAS-expressed CA-IBor DN-IB). FIG. 3. AP levels in 15-days USC cultures expressing (A)

CA-IA and -IB receptors and (B) DN-IA and -IB receptors.The 15-day USCs infected with control (empty) RCASvirus or RCAS containing mutant BMPR were cultured for6 days before switching the cultures to SF medium 95 30ng/ml BMP-4 or 0.15 mM ascorbate phosphate (Asc) for anadditional 48 h. Results are expressed as the mean 95 SD forthree independent experiments performed with triplicatewells. A Student’st-test was used to determine statisticalsignificance. Difference compared with control virus is sig-nificant at **p 5 0.015, *p ,0.001, or #p , 0.005. Differ-ence between CA-IA and CA-IB or DN-IA and DN-IB issignificant at11p 5 0.05 or1p , 0.001.

1634 VOLK ET AL.

more effective at inducing high-level AP activity than over-expression of CA-IA. Like the nonascorbate cultures, addi-tion of exogenous BMP to ascorbate-treated cultures in-creased AP activity in cells containing CA-BMP type Ireceptors.

As a further test of the relative importance of IA and IBreceptors, USCs were infected with the DN forms of eachreceptor. Both DN-IA and DN-IB significantly reduced APinduction by exogenous BMP-4 (Fig. 3B). Neither one ofthe DN-BMPRs altered the small amount of induction seenwith ascorbate alone but both clearly reversed the dramaticinduction of AP resulting from exposure to BMP and ascor-bate. Although significant differences between DN-IA andDN-IB expression were not apparent in the presence ofBMP alone, DN-IB was more effective than DN-IA inblocking AP induction in the presence of BMP and ascor-bate (p 5 0.05).

Alterations in colX and colII mRNA levels in 15dUSCsby mutant type IB BMPRs

Four independent Northern analyses were performed inorder to assess the effect of DN- and CA-IB receptors onchanges in collagen mRNA associated with chondrocytehypertrophy. A representative Northern blot is seen in Fig.4. As expected, treatment of USCs with BMP-4 inducedhigh levels of colX and modestly reduced levels of colIImRNA. The DN form of BMPR-IB was able to block theeffects of exogenously added BMP-4 (Fig. 4) as well asBMP-2 and BMP-7 (not shown). Expression of the CAmutant of the type IB receptor (CA-IB) in 15dUSC mim-icked the effects of BMP addition, inducing high levels ofcolX mRNA. CA-IA also was able to induce colX mRNA inthese cells, although the degree of induction was morevariable (not shown). None of the virally infected culturesshowed a marked decrease in colII mRNA, which would beindicative of chondrocyte dedifferentiation.

Use of mutated BMPRs to analyze BMP signaling inthree different chondrocyte populations

Because the mutant type I receptors altered maturation ofprehypertrophic chondrocytes (15dUSC) in culture, wetested if they also would affect expression of hypertrophicmarkers in cells that are at a later stage of maturation. Byembryonic day 17, the central core of the cephalic sternalregion contains hypertrophic chondrocytes, which are ex-pressing both colX and AP.(17) Chondrocytes derived fromthis region (17dUSCc) were placed in culture, infected withRCAS constructs, and after 6 days were transferred to SFmedium with inducers. Not surprisingly, AP levels in17dUSCc were approximately 2-fold higher than seen with15dUSC under all conditions (Figs. 3A and 5A). However,as seen with the 15dUSC, overexpression of either CA-IAor CA-IB in 17dUSCc effectively increased AP levels underall other conditions. Like the 15dUSC, CA-IB was signifi-

FIG. 4. The colII and colX mRNA levels in sternal chon-drocytes expressing mutated BMP-IB receptors. USCs andLSCs from day 15 chick embryos (15dUSC and LSC,respectively) were cultured for 6 days in the absence ofRCAS(2) or with RCAS encoding DN- and CA-IB-BMPRs. Cultures were then placed under SF conditionswith or without 30 ng/ml BMP-4 (B4) for the last 48 h ofculture. A representative Northern blot, hybridized to ribo-probes for chick colX and colII is shown.

FIG. 5. AP levels in 17-day USC cultures expressing (A)CA-IA and CA-IB receptors and (B) DN-IA and -IB recep-tors. Cultures of USCs from day 17 embryos were treated asdescribed in Fig. 4. Results are expressed as the mean6 SDfor at least two independent experiments performed withtriplicate wells. A Student’st-test was used to determinestatistical significance. Different compared with control vi-rus is significant at *p , 0.001 or #p , 0.005; differencebetween CA-IA and CA-IB or between DN-IA and DN-IBis significant at1p , 0.001.

1635BMP RECEPTORS FOR CHONDROCYTE HYPERTROPHY

cantly better than CA-IA at inducing higher levels of APactivity (p , 0.005), and addition of BMP-4 was able toincrease AP activity further (p , 0.05) in cells infected witheither CA receptor. Expression of the DN-IA receptor (Fig.5B) slightly decreased AP activity in the presence ofBMP-4, with or without ascorbate (p , 0.05), whereasDN-IB expression caused a much greater reduction ofBMP-induced AP levels (p , 0.0001).

Cultured LSCs do not undergo hypertrophy in response toexogenous BMPs, with or without ascorbate.(15) If elevatednoggin expression in LSC inhibited BMP effects by bindingexogenously added BMPs, it was plausible that CA-BMPR-IB receptor might induce maturation in these cells.We therefore infected 15dLSC with CA-BMPRs and ana-lyzed the effects on collagen mRNA (Fig. 4) and AP activity(Fig. 6). Northern blot analysis of LSC overexpressingCA-IB indicated barely detectable levels of colX mRNA;however, the amounts are too low to be visible in Fig. 4.Consistent with a modest induction of colX, retroviral ex-pression of CA-IB in LSC led to slightly diminished levelsof colII mRNA compared with LSCs with or without BMPs(Fig. 4). Consistent with the RNA analyses, overexpressionof CA-IB led to a 3-fold elevation in AP activity (p # 0.001compared with control RCAS-infected LSCs). As had beenseen with upper sternal cells, CA-IB in lower sternal cellswas more effective than CA-IA, which significantly in-creased AP only in the presence of Asc. It therefore appearsthat LSCs can respond to BMP signaling when providedwith activated BMPR, but their response is only one-tenththat seen with prehypertrophic chondrocytes (15dUSC) orhypertrophic chondrocytes (17dUSCc) expressing CA-IB.

DISCUSSION

We have previously reported that 15dUSCs but not15dLSCs are induced to mature by addition of BMP, with or

without ascorbate. Data presented here suggest that thedifference in BMP responsiveness between the two popu-lations of chondrocytes is not attributable to absence ofBMPRs in the LSCs. RT-PCR of RNA from LSCs, prehy-pertrophic USCs, and hypertrophic USCs showed no detect-able differences in either type IA, IB, or type II chickBMPR expression (Fig. 1). It is therefore unlikely that alack of BMPRs in 15dLSCs accounts for their lack ofresponse to exogenous BMPs. These results generally areconsistent with immunohistochemical studies of BMPRlocalization in mammalian embryos.(24,25) AlthoughBMPR-IA expression is widespread and BMPR-IB is morerestricted, both are found throughout cartilage and develop-ing endochondral bone. Within mouse embryo long bones,both BMPR-IA and -IB are found in immature, mature, andhypertrophic regions of cartilage, along with BMPR-II.(24,25) In situ hybridization studies with chick embryossuggest that BMPR-IB and -II are expressed early in chon-drogenesis and persist throughout later stages of chondro-cyte differentiation, whereas BMPR-IA expression is lowduring early chondrocyte differentiation and increases inprehypertrophic chondrocytes.(20,26)

Because LSCs were found to possess all three BMPRs,we examined the alternative possiblility that LSCs cannotundergo BMP-induced hypertrophy because they produceelevated noggin levels. Noggin is a BMP-binding proteinthat functionally inactivates BMP-2, -4, and -7. It is widelyexpressed throughout the forelimb cartilage of 16.5-dpcmouse embryos, but its expression was reported to becomeextinguished in maturing chondrocytes.(27) Furthermore,noggin-infected chick limbs show inhibition of chondrocytehypertrophy.(28) Therefore, if noggin expression was re-sponsible for the differences seen in BMP responsivenessbetween the two cell types, noggin transcripts should beabundant in mRNA from LSC but not USC. RT-PCR resultssuggest that LSCs do express noggin at high levels, and thatlowest noggin expression is seen in prehypertrophic chon-drocytes, which undergo hypertrophy in response to BMP.However, CA-BMPRs, which should bypass noggin effects,show only modest induction of hypertrophy in LSC; expres-sion of colX mRNA was barely detectable (Fig. 4) and APactivity was an order of magnitude lower than that seen with15dUSCs (Fig. 6). The detectable response of LSC to CA-BMPRs implies that these cells possess some downstreammediators of BMP signaling. The limited response to CABMPRs in LSC suggests that factors in addition to nogginexpression may be involved in preventing LSC hypertro-phy, and that such factors are likely to act downstream ofreceptor activation.

Three downstream mediators, Smad1, -5, and -8, havebeen identified as intracellular proteins phosphorylated byactivated BMPRs, and an additional two inhibitory Smads,Smad6 and Smad7, suppress both BMP and TGF-b signal-ing.(9,10,29)Either high levels of inhibitory Smad6 or Smad7or a lack of stimulatory Smads in LSCs might make thesecells resistant to the effects of both BMPs and CA-BMPRs.A recent immunohistochemical study by Sakou et al.(30) hasexamined Smad localization in rat epiphyseal plate. Stain-ing for Smad6 and Smad7 shows no obvious accumulationin cells of the resting zone, but these Smads were increased

FIG. 6. Effects of expressing CA-IA and CA-IB in 15-dayLSC cultures. Cultures of LSCs from day 15 embryos weretreated as described in Fig. 4. Results are expressed as themean6 SD for three independent experiments performedwith triplicate wells. Different compared with control virusis significant at**p , 0.05 and *p , 0.001 ; differencebetween CA-IA and CA-IB is significant at11p , 0.05 or1p , 0.001.

1636 VOLK ET AL.

in hypertrophic chondrocytes, suggesting that elevated lev-els of inhibitory Smads subsequent to hypertrophy maydown-regulate BMP effects in these cells. Less surprisingly,Smad1 and Smad5 levels appeared highest in late prolifer-ating and early hypertrophic chondrocytes, consistent with arequirement for these Smads in BMP-stimulated hypertro-phy. Smad1 and Smad5 expression was weak in chondro-cytes near the zone of resting cartilage, and relatively fewchondrocytes within the resting zone stained positively,suggesting that cells not destined for hypertrophy mightlack the Smads required for BMP signaling. However, fur-ther analyses of Smad expression in permanent cartilagewill be required to confirm whether low levels of BMP-specific Smads limit the ability of the chondrocytes toundergo BMP-induced hypertrophy.

Expression of mutated BMPRs has previously been usedto analyze the role of BMP signaling in embryonic pattern-ing, neurogenesis, and osteogenesis as well as chondrocytedifferentiation and maturation.(20,21,26,31–36)Our comparisonof both DN and CA type I BMPRs suggests that the majortype I receptor used during hypertrophy is BMPR-IB. Ex-pression of DN-IB in 15dUSC cultures blocked the effectsof exogenous BMPs on both induction of AP activity (Fig.3B) and colX mRNA (Fig. 4). Whereas overexpression ofDN-IA also reduced AP levels in BMP-treated USC, thereduction was significantly different from virus control onlyin the presence of BMP1 ascorbate (Fig. 3B) and wasmarkedly less than that seen with DN-BMPR-IB. The ob-servation that the DN-IB receptor can reduce type X mRNAto levels equal to or below that seen in control cultures, inwhich BMP had not been added (Fig. 4), suggests that a lowlevel of endogenous BMP activity may be produced in theSF chondrocyte cultures. This is consistent with our obser-vation that BMP-2, -4, and -7 are all detectable in USCscultured under SF conditions (Fig. 1).

The results with CA-BMPRs also indicated that the IBreceptor was more effective than the IA receptor. CA-IBreceptor consistently induced elevated levels of colXmRNA in the absence of exogenous BMPs (Fig. 4), whereasresults with CA-IA were variable. Except for cells treatedwith BMP alone, cultures infected with the CA-IA constructproduced AP at significantly lower levels than culturesinfected with CA-IB (Fig. 3A). Both in the absence and inthe presence of ascorbate, higher AP levels (Fig. 3A) wereseen with cultures expressing CA-IB receptor than thosereceiving exogenous BMP (p , 0.001); this was not truewith CA-IA. Therefore, our observations, using both CAand DN constructs, strongly suggest that the major type IBMPR involved in chondrocyte hypertrophy is BMPR-IB.

Zou et al.(20) have provided additional evidence thatBMPR-IA is not the receptor normally involved in chon-drocyte hypertrophy. These workers analyzed the relativeroles of BMPR-IA and -IB in chondrocyte differentiation byin ovo infection of developing chick embryos with the sameRCAS constructs used in our experiments. In situ hybrid-ization of uninfected embryos showed that BMPR-IB wasexpressed in all cartilage condensations and regions ofendochondral bone formation. BMPR-IA was found in pre-hypertrophic chondrocytes, and overexpression of CA-IA inthe chick limb bud delayed chondrocyte hypertrophy.

Furthermore, CA-IA induced expression of parathyroidhormone–related protein (PTHrP), which is known to blocktransition to hypertrophy.(37–39)Zou et al.(20) therefore con-cluded that activated BMPR-IA does not promote but doesdelay hypertrophy by functioning as an upstream mediatorof PTHrP. Their results with BMPR-IB mutants showedthat the IB receptor was important for early stages of chon-drogenesis, but studies of the effect of DN-IB on hypertro-phy were not possible because the DN-IB receptor com-pletely blocked cartilage formation. However, the CA-IBreceptor, unlike CA-IA, did permit colX expression, andtype X mRNA appeared more broadly distributed. Theseresults are therefore consistent with our data suggesting thatthe major type I receptor mediating BMP action duringhypertrophy is the IB receptor.

Like the present study, the work of Enomoto-Iwamoto etal.(21) also examined the effect of DN-BMPRs on culturedchick sternal chondrocytes. Although complicated by de-differentiation of the chondrocyte phenotype after infectionwith retrovirus expressing DN-BMPR constructs, thesestudies also implicated BMPR-IB, along with BMPR-II, inchondrocyte maturation. However, the effect of the DN-IBreceptor on either AP or colX mRNA was relatively modestcompared with the results reported here. Because the studiesof Enomoto-Iwamoto et al.(21) employed sternal chondro-cytes derived from day 17 embryos, it was possible thatthese more mature, hypertrophic chondrocytes might showaltered sensitivity to mutated BMPR-IB. We therefore com-pared the effects of mutant BMPRs in cultured chondro-cytes from day 15 and day 17 sternums. Although day17–derived cells, as expected from their more mature status,showed higher AP levels than day 15 cells under all condi-tions, the pattern of CA- and DN-BMPR effects was essen-tially similar with both 15dUSC (Fig. 3) and 17dUSC (Fig.5). However, the effects of CA-IB on 17dUSC were only60–75% of those seen with day 15 prehypertrophic chon-drocytes. More pronounced differences between cells from15-day and 17-day sternums were seen with DN-BMPR-IB;inhibition of BMP-induced AP was over 70% in 15dUSC,but only 20% in 17dUSC. Similarly, the effect of mutant IAreceptors was less severe in day 17 chondrocytes than in day15 chondrocytes. Therefore, cultures of chondrocytes thatare already hypertrophic appear less sensitive to the effectsof exogenous BMP type I receptors. Although Enomoto-Iwamoto et al.(21) reported that DN-BMPR constructscaused altered culture morphology and marked dedifferen-tiation, we did not detect any morphological changes inchondrocyte cultures infected with RCAS encoding eitherDN- or CA-BMPRs. This may reflect differences in cultureconditions; the experiments of Enomoto-Iwamoto et al.involved subculturing chondrocytes for 5–6 days beforeassay, while our protocol did not. It is known that subcultureof primary chondrocytes in monolayers causes an increasedtendency to dedifferentiate, particularly with environmentalperturbation.(40)

Analysis of BMP expression in the embryonic mousehumerus has indicated that BMP-2 and -6 are expressed byhypertrophic chondrocytes, BMP-4 in cells of the transitionzone immediately adjacent to the mature hypertrophic chon-drocytes as well as in the perichondrium, and BMP-7 in

1637BMP RECEPTORS FOR CHONDROCYTE HYPERTROPHY

both proliferating chondrocytes and perichondrium.(41) OurRT-PCR analysis of mRNA suggests that various chicksternal chondrocyte populations also differ in expression ofBMPs (Fig. 1). BMP-2 mRNA can be amplified to compa-rable levels in cultured prehypertrophic 15dUSC, culturedimmature 17dLSC, hypertrophic chondrocytes from uppersternal core, and prehypertrophic chondrocytes from uppersternal periphery. In contrast, BMP-4 and -7 appear high-est in the prehypertrophic populations (15dUSC and17dUSCp). Furthermore, only trace amounts of RT-PCRproduct for BMP-7 are found in immature chondrocytesfrom the caudal sternum after 30 cycles, while it is readilyamplified in both prehypertrophic and hypertrophic popula-tions. Because our in vitro data with prehypertrophic chon-drocytes indicate that BMP-4 is a more effective inducer ofhypertrophy than BMP-2 or -7(16) and BMP-4 levels areelevated in both the transition (pre-hypertrophic) zone ofmouse long bones(41) and our avian prehypertrophic chon-drocytes, it is plausible that BMP-4 is the primary stimulatorof hypertrophy. However, Grimsrud et al.(42) have recentlypresented data suggesting that BMP-6 plays this role. Littleinformation is available on expression patterns for BMP-6during development, and we have been unable to testBMP-6 in our system. Therefore, this issue remains unre-solved.

The results reported here suggest that, whether it isBMP-4 or -6, the ligand activates a receptor complex con-taining BMPR-IB to initiate signaling. Given the broadexpression pattern for the IB receptor during chondrocytedifferentiation, it is unlikely that the transition to hypertro-phy is triggered by increased levels of the receptor. Further-more, there is a large body of evidence showing that regu-lation of hypertrophy in the growth plate involves a balancebetween stimulatory and inhibitory factors.(28,38,43) Thus,any model for the control of chondrocyte maturation andsubsequent endochondral bone formation must include notonly BMPs but also regulatory factors that can suppresshypertrophy, such as PTHrP and Indian Hedgehog.

ACKNOWLEDGMENTS

We gratefully acknowledge the gift of RCAS vectorscontaining mutant BMPRs from Dr. Lee Niswander at theSloan-Kettering Cancer Institute. We also thank GeneticsInstitute for its gift of BMP-4 and Kyle Mansfield for actinprimers. This work was supported by the National Institutesof Health (NIH) grant AR40075 from the Institute of Ar-thritis and Musculoskeletal Diseases.

REFERENCES

1. Urist MR 1965 Bone: Formation by autoinduction. Science150:893–899.

2. Massague´ J 1996 TGFb signalling: receptors, transducters andmad proteins. Cell85:479–487.

3. Sakou T 1998 Bone morphogenetic proteins: From basic stud-ies to clinical approaches. Bone22:591–603.

4. Koenig BB, Cook JS, Wolsing DH, Ting J, Tiesman JP, CorreaPE, Olson CA, Pecquet AL, Ventura F, Grant, Chen G-X,Wrana JL, Massague´ J, Rosenbaum JS 1994 Characterization

and cloning of a receptor for BMP-2 and BMP-4 from NIH3T3 cells. Mol Cell Biol14:5961–5974.

5. Nohno T, Ishikawa T, Saito T, Hosokawa K, Noji S, WolsingDH, Rosenbaum JS 1995 Identification of a human type IIreceptor for bone morphogenetic protein-4 that forms differ-ential heteromeric complexes with bone morphogenetic pro-tein type I receptors. J Biol Chem270:22522–22526.

6. Liu F, Ventura F, Doody J, Massague´ J 1995 Human type IIreceptor for bone morphogenic proteins (BMPs): Extension ofthe two-kinase receptor model to the BMPs. Mol Cell Biol15:3479–3486.

7. Ten Dijke P, Yamashita H, Sampath TK, Reddi AH, EstevezM, Riddle DL, Ichijo H, Heldin C-H, Miyazono K 1994Identification of type I receptors for osteogenic protein-1 andbone morphogenetic protein-4. J Biol Chem269:16985–16988.

8. Yamashita H, Ten Dijke P, Huylebroeck D, Sampath TK,Andries M, Smith JC, Heldin C-H, Miyazono K 1995 Osteo-genic protein-1 binds to activin type II receptors and inducescertain activin-like effects. J Cell Biol130:217–226.

9. Whitman M 1998 Smads and early developmental signaling bythe TGFb superfamily. Genes Dev12:2445–2462.

10. Derynck R, Zhang Y, Feng XH 1998 Smads: Transcriptionalactivators of TGF-b responses. Cell95:737–740.

11. Yanagisawa J, Yanagi Y, Masuhiro Y, Suzawa M, WatanabeM, Kashiwagi K, Toriyabe T, Kawabata M, Miyazono K, KatoS 1999 Convergence of transforming growth factor-b andvitamin D signaling pathways on SMAD transcriptional coac-tivators. Science283:1317–1321.

12. Wotton D, Lo RS, Lee S, Massague´ J 1999 A Smad transcrip-tional corepressor. Cell97:29–39.

13. Liberati NT, Datto MB, Frederick JP, Shen X, Wong C,Rougier-Chapman EM, Wang XF 1999 Smads bind directly tothe Jun family of AP-1 transcription factors. Proc Natl AcadSci U S A 96:4844–4849.

14. Shi XM, Yang XL, Chen D, Chang ZJ, Cao X 1999 Smad1interacts with homeobox DNA-binding proteins in bone mor-phogenetic protein signaling. J Biol Chem274:13711–13717.

15. Leboy PS, Sullivan TA, Nooreyazdan M, Venezian RA 1997Rapid chondrocyte maturation by serum-free culture withBMP-2 and ascorbic acid. J Cell Biochem66:394–403.

16. Volk SW, LuValle P, Leask T, Leboy PS 1998 A BMPresponsive transcriptional region in the chicken type X colla-gen gene. J Bone Miner Res13:1521–1529.

17. D’Angelo M, Pacifici M 1997 Articular chondrocytes producefactors that inhibit maturation of sternal chondrocytes inserum-free agarose cultures: a TGF-b independent process.J Bone Miner Res12:1368–1377.

18. Leboy PS, Vaias L, Uschmann B, Golub EE, Adams SL,Pacifici M 1989 Ascorbic acid induces AP, type X collagenand calcium deposition in cultured chick chondrocytes. J BiolChem264:17281–17286.

19. Zou HY, Niswander L 1996 Requirement for BMP signalingin interdigital apoptosis and scale formation. Science272:738–741.

20. Zou H, Wieser R, Massague J, Niswander L 1997 Distinctroles of type I BMP receptors in the formation and differen-tiation of cartilage. Genes Dev11:2191–2.

21. Enomoto-Iwamoto M, Iwamoto M, Mukudai Y, Kawakami Y,Nohno T, Higuchi Y, Takemoto S, Ohuchi H, Noji S, KurisuK 1998 Bone morphogenetic protein signaling is required formaintenance of differentiated phenotype, control of prolifera-tion, and hypertrophy in chondrocytes. J Cell Biol140:409–418.

22. Wieser R, Wrana JL, Massague J 208 1995 GS domain mu-tations that constitutively activate Tb R-I, the downstreamsignaling component in the TGF-b receptor complex. EMBOJ. 14:2199–2208.

1638 VOLK ET AL.

23. Morgan BA, Fekete DM 1996 Manipulating gene expressionwith replication-competent retroviruses. In: Bronner-Fraser M(ed.) Methods in Avian Embryology. Academic Press, SanDiego, CA, U.S.A., pp. 185–218.

24. Ishidou Y, Kitajima I, Obama H, Maruyama I, Murata F,Imamura T, Yamada N, Ten Dijke P, Miyazono K, Sakou T1995 Enhanced expression of type I receptors for bone mor-phogenetic proteins during bone formation. J Bone Miner Res10:1651–1659.

25. Onishi T, Ishidou Y, Nagamine T, Yone K, Imamura T, KatoM, Sampath TK, Ten D, Sakou T 1998 Distinct and overlap-ping patterns of localization of bone morphogenetic protein(BMP) family members and a BMP type II receptor duringfracture healing in rats. Bone22:605–612.

26. Kawakami Y, Ishikawa T, Shimabara M, Tanda N, Enomoto-Iwamoto M, Iwamoto M, Kuwana T, Ueki A, Noji S, NohnoT 1996 BMP signaling during bone pattern determination inthe developing limb. Development122:3557–3566.

27. Brunet LJ, McMahon JA, McMahon AP, Harland RM 1998Noggin, cartilage morphogenesis, and joint formation in themammalian skeleton. Science280:1455–1457.

28. Pathi S, Rutenberg JB, Johnson RL, Vortkamp A 1999 Inter-action of Ihh and BMP/Noggin signaling during cartilagedifferentiation. Dev Biol209:239–253.

29. Massague´ J 1998 TGF-b signal transduction. Annu Rev Bio-chem67:753–791.

30. Sakou T, Onishi T, Yamamoto T, Nagamine T, Sampath TK,ten Dijke P 1999 Localization of Smads, the TGF-b familyintracellular signaling components during endochondral boneossification. J Bone Miner Res14:1145–1152.

31. Suzuki A, Thies RS, Yamaji N, Song JJ, Wozney JM, Mu-rakami K, Ueno N 1994 A truncated bone morphogeneticprotein receptor affects dorsal-ventral patterning in the earlyXenopusembryo. Proc Natl Acad Sci U S A 91:10255–10259.

32. Ishikawa T, Yoshioka H, Ohuchi H, Noji S, Nohno T 1995Truncated type II receptor for BMP-4 induces secondary axialstructures inXenopusembryos. Biochem Biophys Res Com-mun 216:26–33.

33. Akiyama S, Katagiri T, Namiki M, Yamaji N, Yamamoto N,Miyama K, Shibuya H, Ueno N, Wozney JM, Suda T 1997Constitutively active BMP type I receptors transduce BMP-2signals without the ligand in C2C12 myoblasts. Exp Cell Res235:362–369.

34. Varley JE, McPherson CE, Zou H, Niswander L, Maxwell GD1998 Expression of a constitutively active type I BMP receptorusing a retroviral vector promotes the development of adren-ergic cells in neural crest cultures. Dev Biol196:107–118.

35. Chen D, Ji X, Harris MA, Feng JQ, Karsenty G, Celeste AJ,Rosen V, Mundy GR, Harris SE 1998 Differential roles for

bone morphogenetic protein (BMP) receptor type IB and IA indifferentiation and specification of mesenchymal precursorcells to osteoblast and adipocyte lineages. J Cell Biol142:295–305.

36. Kaps C, Lauber J, Ju W, Czichos S, Gross G 1998 Therecombinant expression of bone morphogenetic protein typeIA receptor (Alk3) in mesenchymal progenitors C3H10T1/2 Issufficient for osteo-/chondrogenic development. Biochem SocTrans26:27–32.

37. Vortkamp A, Lee K, Lanske BMK, Segre GV, KronenbergHM, Tabin CJ 1996 Regulation of rate of cartilage differenti-ation by Indian hedgehog and PTH-related protein. Science273:613–622.

38. Lee K, Lanske BMK, Karaplis AC, Deeds JD, Kohno H,Nissenson RA, Kronenberg HM, Segre GV 1996 Parathyroidhormone-related peptide delays terminal differentiation ofchondrocytes during endochondral bone development. Endo-crinology 137:5109–5118.

39. Kronenberg HM, Lee K, Lanske BMK, Segre GV 1997 Para-thyroid hormone-related protein and Indian hedgehog controlthe pace of cartilage differentiation. J Endocrinol154:S39–S45.

40. Cancedda R, Cancedda FD, Castagnola P 1995 Chondrocytedifferentiation. Int Rev Cytol159:265–358.

41. Solloway MJ, Dudley AT, Bikoff EK, Lyons KM, HoganBLM, Robertson EJ 1998 Mice lackingBmp6 function. DevGenet22:321–339.

42. Grimsrud CD, Romano PR, D’Souza M, Puzas JE, ReynoldsPR, Rosier RN, O’Keefe RJ 1999 BMP-6 is an autocrinestimulator of chondrocyte differentiation. J Bone Miner Res14:475–482.

43. Volk SW, Leboy PS 1999 Regulating the regulators of chon-drocyte hypertrophy. J Bone Miner Res14:483–486.

Address reprint requests to:Dr. Phoebe S. Leboy

Department of BiochemistrySchool of Dental MedicineUniversity of Pennsylvania

4001 Spruce StreePhiladelphia, PA; 19104-6003, U.S.A.

Received in original form July 2, 1999; in revised form January 28,2000; accepted March 20, 2000.

1639BMP RECEPTORS FOR CHONDROCYTE HYPERTROPHY