rhBMP2/7 Heterodimer: An Osteoblastogenesis Inducerof Not Higher Potency but Lower Effective Concentration

Compared with rhBMP2 and rhBMP7 Homodimers

Yuanna Zheng, D.D.S., Ph.D., Gang Wu, D.D.S., Ph.D., Juan Zhao, D.D.S., Ph.D.,Linhong Wang, D.D.S., Ping Sun, D.D.S., M.S., and Zhiyuan Gu, D.D.S., Ph.D.

Heterodimeric bone morphogenetic proteins (BMPs) were exhibited to be more potent than and thus potentialsubstitutes for homodimeric BMPs whose clinical application is limited for the drawbacks resulted from theirhigher effective doses. This study aims to delineate the biofunctional characteristics of recombinant humanBMP2=7 (rhBMP2=7) heterodimer in inducing osteoblastogenesis of MC3T3-E1 through in vitro time–course anddose–response studies. rhBMP2=7 heterodimer induced cell migration with a significantly lower optimal con-centration and higher peak effect than the respective homodimers. rhBMP2=7 heterodimer induced cell differ-entiation with significantly lower threshold concentrations but similar maximum effects. On day 28, the area ofcalcium depositions induced by 50 ng=mL rhBMP2=7 was 12- or 38-fold more than that of 50 ng=mL rhBMP2or 50 ng=mL rhBMP7, respectively. The results indicated that rhBMP2=7 heterodimer was an osteoblastogene-sis inducer of not higher potency but lower effective concentration compared with rhBMP2 and rhBMP7homodimers.

Introduction

The osseous restoration of voluminous bone defectremains a challenge in orthopedics, maxillofacial surgery,

and implantology.1,2 Bone tissue engineering including gene,cell, and cytokine therapies has been reported to substituteautografting and allografting whose clinical applicationsare hampered for their varieties of limitations.3,4 Among thethree therapies, cytokine therapy is advantageous in safety,feasibility, and potential for nearest clinical application overthe other two.1,5

Bone morphogenetic proteins (BMPs), a group of dimericdisulfide-linked polypeptide growth factors under trans-forming growth factor-b superfamily, are one of the para-mount cytokines in promoting bone regeneration.6 However,the clinical effective doses of the homodimeric BMPs to inducebone formation are extremely high (e.g., up to milligrams),7,8

which results in not only a substantial economic burden topatients but also a series of potential side effects, such as over-stimulation of osteoclastic activity and ectopic bone forma-tion in unintended area.9,10

One alternative approach to solve the problem is to adoptmore potent forms of BMPs. Previous studies showed thatBMP heterodimers achieved several- or dozen-fold more ef-fects than the respective homodimers in inducing alkaline

phosphatase (ALP) activity and bone formation.11–14 Thesefindings suggested that BMP heterodimers were feasibleand promising substitutes for homodimers. However, thebiofunctional characteristics and mechanisms of BMP heter-odimers on osteogenesis remain ambiguous. How BMPheterodimers affect the sequential osteoblastogenetic events,such as chemotactic migration, proliferation, differentiation,and mineralization, hereto, remains uncovered. Earlier stud-ies confined the parameters to ALP activity, which is inad-equate to reveal the characteristics of their biofunctionalactivities.11,12 Recently, some conclusions were based on theconditioned medium of gene-transfected BMP producers in-stead of purified BMP heterodimers, which could not excludethe influence of other factors in the medium on cellular ac-tivities.13,14 These studies mostly focused on the in vivo genetherapy that could only suggest their clinical efficacy insteadof the functional characteristics of BMP heterodimers.

The parameters such as cell migration index, DNA content,ALP activity as early marker of differentiation, osteocalcin(OCN) as terminal marker of differentiation, and calciumdeposition of cell culture were adopted to evaluate the bio-logical activities of BMPs in previous studies.13,15–17 Murinecalvaria-derived MC3T3-E1 cell, a precursor of functionalizedosteoblast, was also shown to be able to examine the potencyof osteoinductive agents.13,17

This abstract has appeared in the IADR=AADR 87th General Session (April 1–4, 2009).School=Hospital of Stomatology, Zhejiang University, Hangzhou, P.R. China.

TISSUE ENGINEERING: Part AVolume 16, Number 3, 2010ª Mary Ann Liebert, Inc.DOI: 10.1089=ten.tea.2009.0312

879

In this article, we wished to unveil the functional character-istics of BMP heterodimers by comparing osteoblastogenesis-inducing effects of recombinant human BMP2=7 (rhBMP2=7)heterodimer with rhBMP2 and rhBMP7 homodimers.

Materials and Methods

Cell culture

Preosteoblasts, MC3T3-E1 cell line (ATCC; Chinese Acad-emy of Sciences, Shanghai, China), were cultured in a-Minimum essential medium (a-MEM) containing 10% fetalbovine serum (FBS) (Gibco� Invitrogen, Grand Island, NY).The medium was changed every 3 days. Exponentiallygrowing cells were plated at 1.5�104 cells=well in 48-wellplates for cell proliferation, at 4�104 cells=cm2 in 6-well platesfor ALP activity assay, in 24-well plates for OCN detection, orin 48-well plates for alizarin red staining. After 24-h incuba-tion, cells were subjected to a low-serum medium (2% FBS) foranother 24 h and then followed by rhBMPs (R&D Systems,Inc., Minneapolis, MN) treatment.

rhBMP concentration selection test

To identify the concentration range for examining the bi-ological activities of rhBMP2=7 heterodimer, a dose–responsetest of ALP assay was performed the same way as described in‘‘ALP activity assay’’ section. The results showed that the ALPactivity induced by rhBMP2=7 reached the maximum at 150–200 ng=mL and significantly decreased at 250 ng=mL. Besides,ALP activity induced by rhBMP2 or rhBMP7 reached a pla-teau stage at 200–400 ng=mL, respectively, with similarhighest level. Consequently, the maximum concentration inthis study was determined at 200 ng=mL.

Real-time cell migration monitoring

Cell migration of MC3T3-E1 induced by rhBMPs wascontinuously monitored using Real-Time Cell ElectronicSensing=Cell Invasion=Migration System (RT-CES=CIM�System; ACEA Biosciences, San Diego, CA)18,19 according tothe manufacturer’s instructions. Briefly, the bottoms of theupper chambers, the both sides, were coated with 0.01% bo-vine plasma fibronectin (Invitrogen, Carlsbad, CA), and 20 mLof 0.1% bovine serum albumin (BSA)–a-MEM, and 150 mL of0.1% BSA serum-free medium containing various amounts ofrhBMP2=7, rhBMP2, rhBMP7, or their 1:1 mixture (0 as neg-ative control, 0.01, 0.1, 1, and 10 ng=mL)15,16 were added intothe upper and lower chambers, respectively. ConfluentMC3T3-E1 cells were subjected to serum-free starvation for4 h and then suspended in 0.1% BSA–a-MEM at a density of12.5�104 cells=mL followed by adding 80 mL of cell suspen-sion into the RT-CIM upper chambers. Cell migration wascontinuously monitored for a period of 0–8 h, and data pre-sented as cell index (CI) were collected every 15 min, chieflyindicating the numbers of cells that migrated to the lower sideof the upper chambers.

Cell proliferation assay

To investigate cell proliferation of MC3T3-E1 cells in re-sponse to various concentrations (0 as control, 5, 10, 50, 100,150, and 200 ng=mL) of different rhBMPs, the number of cellswas determined by the fluorometric quantification of the

amount of cellular DNA after stimulation for 1 day, 4 days,and 7 days using Quant-iT� PicoGreen� dsDNA Reagentand kits (Invitrogen�, Molecular Probes�, Eugene, OR) aspreviously described.20 The fluorescent intensity of the celllysate and dye mixed solution was measured using a fluo-rescence spectrometer (SPECTRAmax M2, Molecular Devices,Sunnyvale, CA, Ex 480 nm=Em 520 nm).

ALP activity assay

To determine the early differentiation of preosteoblastsstimulated by rhBMPs at different concentrations (0 as con-trol, 5, 10, 50, 100, 150, and 200 ng=mL), ALP activity andprotein content were measured after rhBMPs treatment ondays 2, 4, and 7. ALP activity in the cell lysates (Sigma-Aldrich, St. Louis, MO) was determined using LabAssay�ALP colorimetric assay kit (Wako Pure Chemicals, Osaka,Japan).11 The cell number was estimated by determining totalprotein content that was measured at 570 nm using a com-mercial BCA Protein Assay kit (Beyotime, Shanghai, China).The values that presented ALP activity were expressed asmmol p-NP=mg total protein.

OCN expression assay

To assess the terminal differentiation of preosteoblastsstimulated by rhBMPs (concentrations set at 0 as control, 5, 10,50, 100, 150, and 200 ng=mL), OCN secreted into the cell cul-ture medium was determined. The cell supernatants werecollected on days 4 and 7 and were centrifuged (10,000 rpm,48C, 5 min) before detection. The OCN concentrations of thesupernatants were determined by ELISA using a mouse OCNEIA kit (Biomedical Technologies, Stoughton, MA).13,20

Alizarin red staining

According to the results of above cellular events, we com-pared the mineralization possibility of MC3T3-E1 cells stim-ulated by 50 ng=mL rhBMPs. Quadruplicate cell cultures wereprepared in the same way as described previously and thentreated with mineralizing medium (10% FBS, 50mg=mL L-ascorbic acid, and 5 mM b-glycerophosphate; Sigma-Aldrich)containing 50 ng=mL rhBMPs. The medium was replacedevery 3 days. After 3 and 4 weeks, mineralized nodules weredetermined by alizarin red staining (Sigma-Aldrich). Cultureplates were photographed by NIS-Elements F2.20 (NikonEclipse 80i, Tokyo, Japan), and the calcified area was quanti-fied using Image-Pro Plus 6.0 analysis.

Statistical analysis

Statistical comparisons between results obtained with thevarious concentrations of rhBMPs and different kinds ofrhBMPs were made by one-way analysis of variance. Post hoccomparisons were made using Bonferroni corrections. Thelevel of significance was set at p< 0.05. SPSS software (ver-sion 15, SPSS Inc., Chicago, IL) for a Windows computersystem was employed for the statistical analysis.

Results

Chemotactic migration

CI of cell chemotactic migration for each rhBMP was in thesame pattern which contained two stages: the initial steady-

880 ZHENG ET AL.

increase stage in the first 4–5 h and the following plateaustage (Fig. 1A–D). Cell started to migrate immediately afterbeing installed in the system regardless of the presence ofrhBMPs. The speeds of cell migration were enhanced tohigher levels by different rhBMPs of certain concentrationsthan negative controls especially after 1–2 h. The chemotacticeffects of rhBMPs lasted for 5 h after which relative cell mi-grations tapered. Despite the same pattern of CI, the optimalconcentration (the lowest concentration to get the maximumeffect) and the maximum value of chemotactic effects dif-

fered according to the rhBMP types. The chemotactic effect ofrhBMP2=7 showed a bell-shaped, dose-dependent curve. Theoptimal concentration was 0.1 ng=mL, followed by 1 ng=mL(Fig. 1A). The effect dramatically decreased when the concen-tration decreased to 0.01 ng=mL or increased to 10 ng=mL.The maximum values of rhBMP2 and rhBMP7 were obtainedaround 1 ng=mL (Fig. 1B, C). Their decreasing trends weredemonstrated when the concentration went either higher orlower. The chemotactic effect of 1:1 mixture of rhBMP2 andrhBMP7 increased slowly in the selected concentration range

FIG. 1. (A–D) Comparison of cell migration under different rhBMPs treatment. Real-time cell migration was continuouslymonitored using RT-CES=CIM� System in a period of 8 h. The data presented as CI were collected every 15 min, chieflyindicating the numbers of cells that migrated. (E) Fold of CI at 5 h normalized by CI of control group (**p< 0.01, *p< 0.05,indicated significant increase vs. control group, a,b,cp< 0.05, indicated significant increase vs. rhBMP2, rhBMP7, or their 1:1mixture group, respectively, n¼ 3). CI, cell index; rhBMPs, recombinant human bone morphogenetic proteins.

ACTION OF rhBMP2/7 HETERODIMER IN OSTEOBLASTOGENESIS 881

(Fig. 1D). The maximum fold of CI (CI normalized by CI ofnegative control group) of rhBMP2=7 (1.93 at 0.1 ng=mL) wassignificantly higher than that of rhBMP2 (1.48 at 1 ng=mL),rhBMP7 (1.31 at 1 ng=mL), or their 1:1 mixture (1.43 at10 ng=mL) (Fig. 1E).

Cell proliferation

After stimulation for 1 day, there is no significant differ-ence in cell proliferation among different BMPs of the se-lected concentrations.

On day 4 (Fig. 2), compared with the control group, theDNA contents increased significantly under the stimulationof rhBMP2=7 and rhBMP2 at the concentrations of 5–50 or200 ng=mL. In contrast, rhBMP7 showed a stimulative effecton the proliferation of preosteoblasts in a broader range ofconcentration, from 5 to 200 ng=mL. The highest values ofDNA content were obtained at 5 ng=mL for rhBMP2 orrhBMP2=7, at 200 ng=mL for rhBMP7, and at 50 ng=mL forthe mixture of rhBMP2 and rhBMP7.

The DNA contents in all the rhBMP groups further in-creased on day 7. Compared with the respective controlgroups, the DNA contents were significantly increased byrhBMP2=7 at the concentrations of 5–10 ng=mL. For rhBMP2,the effective concentration range was 5–50 ng=mL. In ac-cordance with the result on day 4, the DNA contents wereincreased by rhBMP7 in the selected concentration rangefrom 5 to 200 ng=mL, with the peak value occurring at100 ng=mL. The highest values of stimulative effects on cellproliferation were obtained at 5 ng=mL for rhBMP2 orrhBMP2=7, at 100 ng=mL for rhBMP7, and at 50 ng=mL forthe mixture of rhBMP2 and rhBMP7. The mixture of rhBMP2and rhBMP7 showed no synergistic effect. Besides, nosignificance was found among the highest values of theDNA contents in different rhBMP groups on day 4 or 7( p> 0.05).

ALP activity

The minimum concentration for rhBMP2=7 heterodimer toexhibit a significant promoting effect on ALP was as low as5 ng=mL, at which rhBMP2, rhBMP7, or their mixture re-mained ineffective (Fig. 3). The rhBMP homodimers startedto exert their promotive effect on the ALP activities at theconcentration of 50 ng=mL, at which the effect of rhBMP2=7on stimulating the ALP activity has already reached a pla-teau (Fig. 3A–C). Therefore, the threshold concentration ofrhBMP2=7 heterodimer was almost 1=10 of that of thehomodimers. In addition, the levels of ALP activity inducedby rhBMP2=7 on day 2 were about 1.4- to 2.4-fold of thatof rhBMP2 or rhBMP7 at all the selected concentrations(Fig. 3A). After 4-day stimulation, the ALP activities furtherincreased under the stimulation of rhBMP2=7 heterodimer atthe concentrations of 5 and 10 ng=mL at which rhBMP2 andrhBMP7, remained ineffective (Fig. 3B). At the concentrationof 50 ng=mL, the ALP activity stimulated by rhBMP2=7 wasstill significantly higher than rhBMP2, rhBMP7, and their 1:1mixture. The ALP activities induced by rhBMP2=7 were stablein the concentration range of 50–200 ng=mL, at which the ef-fects of rhBMP2, rhBMP7, and their 1:1 mixture showeda dose-dependent increasing curves. The advantages ofrhBMP2=7 over the homodimers, thus, tapered with the in-creased concentrations. At the concentration of 100–200 ng=mL, rhBMP2=7 was only advantageous over rhBMP7. On day7, the ALP levels induced by all rhBMPs were about 1.5-foldmore than those of day 4 (Fig. 3C). The effect pattern ofrhBMPs at the concentrations of 5 and 10 ng=mL was similarto that of day 4. The effect of rhBMP2 reached the same level ofrhBMP2=7 at concentrations of 50–200 ng=mL. In contrast,rhBMP7 showed a weaker effect on inducing ALP activity ofpreosteoblast than rhBMP2=7 or rhBMP2 in the span of thegiven concentrations. The effect of the mixtures of rhBMP2and rhBMP7 lied between the two homodimers.

FIG. 2. rhBMPs showed different optimal concentration with similar peak values on stimulating MC3T3-E1 proliferation.The cellular DNA contents were measured on days 1, 4, and 7. Values shown are means� SD for quadruplicate cultures(*p< 0.05 indicated significant increase vs. control group). SD, standard deviation.

882 ZHENG ET AL.

OCN expression

In accordance with its effect on the ALP activity, thethreshold dose of rhBMP2=7 heterodimer to promote OCNexpression (5 ng=mL) was also significantly lower than thoseof the rhBMP homodimers (50 ng=mL) (Fig. 4).

After 4-day culture, the OCN expression level induced byrhBMP2=7 (5 ng=mL) increased with concentration until50 ng=mL, after which its effect maintained stable (Fig. 4A).OCN expression showed a dose-dependent increasing curvefor rhBMP2 and rhBMP7 from 50 ng=mL at which the effectof rhBMP2=7 was sixfold higher than the respective homo-dimers. The advantage of rhBMP2=7 decreased with con-centration increase but maintained significant over thehomodimers until the concentration of 200 ng=mL, at whichthere was no significant difference between rhBMP2=7 het-erodimer and the respective homodimers ( p> 0.05).

After culturing with rhBMPs for 7 days, the OCN ex-pression increased to much higher levels, about twofold

higher than those measured on day 4 (Fig. 4B). The effect ofrhBMP2=7 on stimulating OCN was still significantly ad-vantageous over rhBMP2 or rhBMP7 at the concentrationrange of 5–150 ng=mL. rhBMP2 increased the OCN expres-sion significantly from 50 ng=mL and reached a plateau atthe concentrations of 100–200 ng=mL. At 200 ng=mL, nosignificant difference could be found among the OCN levelsinduced by different rhBMPs.

Cell matrix mineralization

After 21-day culture, matrix mineralization of MC3T3-E1was only found under the stimulation of 50 ng=mL rhBMP2=7 (Fig. 5). On day 28, calcium depositions were detected in allgroups, with the significantly maximal level occurring inrhBMP2=7 group, which was 12-fold higher than that ofrhBMP2, 38-fold higher than that of rhBMP7, or 26-foldhigher than that of the mixture group. Besides, comparingthe two time points, the area of calcified nodules was over

FIG. 3. ALP activities of MC3T3-E1 under different rhBMP treatments were determined by colorimetric assay on day 2 (A),day 4 (B), and day 7 (C). Values shown are means� SD for triplicate wells in duplicate experiments normalized by totalcellular protein (*p< 0.05 indicated significant increase vs. control group; a,b,cp< 0.05 indicated significant increase vs.rhBMP2, rhBMP7, or their 1:1 mixture group, respectively). ALP, alkaline phosphatase.

FIG. 4. OCN expressions of MC3T3-E1 under different rhBMP treatments were determined by ELISA on day 4 (A) and day7 (B). Values shown are means� SD for quadruplicate wells (*p< 0.05 indicated significant increase vs. control group;a,b,cp< 0.05 indicated significant increase vs. rhBMP2, rhBMP7, or their 1:1 mixture group, respectively). OCN, osteocalcin.

ACTION OF rhBMP2/7 HETERODIMER IN OSTEOBLASTOGENESIS 883

three times at day 28 than that at day 21 for rhBMP2=7 group(Fig. 5B).

Discussion

BMPs play pivotal roles in many body morphogenic eventsvarying from the earliest embryonic patterning to organ de-velopment.21 The actions and activities of BMPs are depen-dent on the targeting cells and BMPs’ resouces.22,23 In thisstudy, we adopted commercially available purified recombi-nant human BMP heterodimer and homodimers to comparetheir actions on a committed osteogenic cell type MC3T3-E1by gauging a series of cellular events during osteoblasto-genesis.

Infiltration of bone forming cells is an important eventduring the restoration of bone defect. Exogenous BMPs maybe one of the dominant modulators on cell migration, espe-cially when they are applied in an unphysiologically highamount for bone regeneration. Different BMP homodimersdiffered in their chemotactic effect to different targeting celltypes.15,16 BMP heterodimers can be even distinct fromhomodimers because they could exclusively assemble het-eromeric type I receptors and thus pattern the dorsoventralaxis of drosophila embryo while the homodimers could not.24

It is, therefore, worthwhile monitoring the chemotactic effectof BMP heterodimers. Our results showed that the optimalconcentration of rhBMP2=7 for chemotactic effect was signif-icantly lower (1=10), and the maximum effect of the hetero-dimer was significantly higher than those of the respectivehomodimers. These results suggested that rhBMP2=7 het-erodimer might play a more important role than homodimersin enhancing chemotactic migration and patterning the spatialinterrelationship of osteogenic cells.

Consistent with the previous findings, we found that thethreshold and the optimal concentrations of rhBMP2=7 het-erodimer for each cellular event were significantly lowerthan the respective homodimers. However, our resultsshowed that the maximum effects of rhBMP2=7 heterodimeron stimulating proliferation and differentiation were similar

to instead of higher than those of rhBMP2 or rhBMP7homodimers. In addition, the gene expression results of ALPand OCN on days 1, 4, and 7 determined by real-time reversetranscription (RT)–polymerase chain reaction also indicatedthat the final maximum levels stimulated by rhBMP2=7heterodimer were similar to that by rhBMP2 homodimer(Supplemental Fig. S1, available online at www.liebertonline.com=ten). Although the novel findings were not inconflict with the previous definition higher potency forrhBMP2=7 heterodimer than the respective homodimers in thelower concentration range (0–50 ng=mL), the advantage ofheterodimer in osteoinductive potency tapered in the higherconcentration range (50–200 ng=mL). BMP2=7 heterodimerproduced by transfected CHO cells was shown to be morepotent than the BMP2 homodimer in stimulating osteogenicdifferentiation.25 The inconsistency with our results may bedue to the use of mammalian producer, conditioned mediuminstead of purified protein, and the single selected concentra-tion. And this needs to be further clarified. Moreover, our re-sults also exhibited that for stimulating cell migration (Fig. 1)and proliferation (Fig. 2), dose-dependent effects of rhBMPs, nomatter heterodimer or homodimer, unanimously showed adose-dependent, bell-shaped curve in which rhBMPs of eitherlower or higher concentrations than the optimal one woulddownregulate their effects. In the concentration selection test,the results also partially indicated that the ALP activities alsoanalogously exhibited a bell-shaped, dose-dependent curve.These findings strongly suggest that the specificity ofrhBMP2=7 heterodimer, as an osteoblastogenesis-inducer, islower effective concentration instead of higher potency incomparison with the homodimers. In fact, the phenomenonthat different BMPs affect in different concentration range(usually shown as different potencies) may generally occur todifferent heterodimers and homodimers.14,26

By now, the molecular mechanisms accounting for thespecificities of BMP remained unclear. One mechanism wouldbe the various receptor complex created by the pairing com-binations from three type I and three type II receptors.27 Onthe other hand, this hypothesis is challenged by the phe-

FIG. 5. (A) The calcium depositions of cell culture were stained using alizarin red staining. (B) Quantitative calculation ofcalcified area in each rhBMP group. Values shown are means� SD for quadruplicate wells (**p< 0.01, *p< 0.05 indicatedsignificant increase vs. control group; a,b,cp< 0.05 indicated significant increase vs. rhBMP2, rhBMP7, or their 1:1 mixturegroup, respectively). Color images available online at www.liebertonline.com=ten.

884 ZHENG ET AL.

nomenon that several BMPs can use one receptor, and theabsence of one receptor does not totally block signaling acti-vated by specific BMPs.28 A recent study that revealed dif-ferent reliance of BMPs on different receptors,29 made themechanism more complete and reasonable. The source of thespecificity of BMP2=7 heterodimer could also be its highercombined affinity for both the type I and type II receptors, andthus facilitating the formation of active signaling complexesendowed by the presence of both BMP2 and BMP7 moieties inthe same ligand.24 Another mechanism may be the compli-cated BMP antagonism network. BMP2=7 heterodimer wasshown to induce a significant lower level Noggin, a BMP-induced competitive antagonist, and lower affinity to Nogginthan the respective homodimers.30 Another antagonist Cas-interacting zinc finger protein behaves on BMP2=7 hetero-dimer similarly to Noggin.23,31

Similarly, the biological significance of the low-effectiveconcentration specificity of heterodimer also remains unclear.Because the coexpression of two different BMP genes has beendetected in developing limb and during fracture healing,32,33

heterodimers might be a pivotal effecter on the site of lessBMP producers where they could only provide a lower con-centration of BMP homodimers, such as the beginning phaseof fracture healing and embryonic development. For this hy-pothesis, our results provided several positive clues such asthe significantly higher maximum level of rhBMP2=7 forchemotactic effect (the initial step of osteogenesis) or the sig-nificantly lower threshold of rhBMP2=7 for promoting pro-liferation and differentiation than the respective homodimers.Besides, it is also important to note that the ALP activitiesinduced by rhBMP2=7 heterodimer were significantly higherthan those of homodimers at concentrations of 100–200 ng=mL on day 2 (Fig. 3A) and the advantages tapered with time.On day 7, rhBMP2=7 was only advantageous over rhBMP7 ininducing the ALP activities. The differences of OCN expres-sion between rhBMP2=7 and the respective homodimers atconcentrations of 50–150 ng=mL were more significant on day4 than on day 7. Mineralization induced by rhBMP2=7 wasobviously observed on day 21, but not by the homodimers.These findings indicated that rhBMP2=7 heterodimer showedeffects more rapidly and accelerated the process of differen-tiation and mineralization, and therefore suggested a pivotalrole of BMP heterodimer in the beginning stage of osteogenicactivities.

As we showed, although rhBMP2=7 heterodimer was alow-effective concentration inducer, its action also showed aconcentration-dependent, bell-shaped curve for the cellularevents that sequentially occurred with increasing concen-trations. One event always booms with the former one ta-pering, although there were broad overlaps. Only thespecifically appropriate concentration of rhBMP2=7 hetero-dimer could exert an optimal promoting effect on certaincellular events. This finding is significant for designing thein vivo application pattern of rhBMPs that the appropriatelevel of BMPs should fit a certain stage of osteoblastogenesis.Overdose of BMPs in the initial stage may lead to the down-regulation of cell migration, and thus suppress the accumu-lation of osteogenic cells as well as their proliferation. Thedelivering mode of BMPs was shown to influence the osteo-conductivity of implant surfaces during the early phase ofosseointegration.34 Slow-released and low level of BMPs (lowlocal concentration of BMP in initial stage of bone regenera-

tion) is superior to burstly released ones in recruiting cells. Wealso observed fewer osteogenic activities in the initial stagewhen applying a higher amount of BMPs in an in vivo bonysite (unpublished data). Although the overstimulation of os-teoclastic activity was also believed to be one of the mecha-nisms for the less bone volume,35 our findings stronglysuggested that the suppression of osteoblastic activities by thelocally transient high concentration of BMPs might also ac-count for the attenuated bone formation. Therefore, cautionmust be especially taken to the application pattern ofrhBMP2=7. However low its effective concentration is, theBMP heterodimer should also be delivered in a physiological-like manner. The principle was suitable not only for proteintherapy but also for gene therapy. Controlling the level andduration of heterodimeric BMPs using the lower titer or re-duced numbers of transduced cells or regulated expressionsystem was also recommended to avoid bone overgrowth orimmune overreaction in combinatorial gene therapy withBMP2=7.36

Significantly lower effective dose of BMP heterodimerswas also expected to bring fewer side effects than thehomodimers. However, because the maximum capability ofrhBMP2=7 in inducing osteogenic activities which is ob-tained in a relatively lower effective concentration range wassimilar to the homodimers, the capability of the heterodimerin inducing the side effects might also be similar to that of thehomodimers in the same range. Consequently, further studyshould clarify this point before clinical application of BMPheterodimers.

One limitation in this study was that we only adopted onekind of osteogenic cell line, which might behave differentlyfrom the others such as bone marrow stromal cells. Moreover,a broader concentration range could be further studied tocompletely exhibit the functional characteristics of both het-erodimers and homodimers. In addition, the bioactivities ofthe rhBMPs generated by Escherichia coli might be differentfrom the ones generated by mammalian cells.

The specificity of lower effective concentration rangeconferred rhBMP2=7 heterodimer a promising clinical ap-plication potential over its respective homodimers. However,further studies must be performed to examine the potentialside effects of rhBMP2=7 heterodimer before its in vivoapplication.

Acknowledgments

We thank Professor Ke, Y. for providing RT-CES=CIMsystem biosensor and also Fu, H. for technical assistance inLaboratory of Cell Signaling and Molecular Genetics, Zhe-jiang University. We also want to thank the Laboratory ofPathology for supplying ABI PRISM 7900HT Sequence De-tection System and Song, K. for technical assistance. Thiswork was supported by the International Team of Im-plantology Research Foundation (no. 547-2007).

Disclosure Statement

No competing financial interests exist.

References

1. Saito, N., and Takaoka, K. New synthetic biodegradablepolymers as BMP carriers for bone tissue engineering. Bio-materials 24, 2287, 2003.

ACTION OF rhBMP2/7 HETERODIMER IN OSTEOBLASTOGENESIS 885

2. Claffey, N., Clarke, E., Polyzois, I., and Renvert, S. Surgicaltreatment of peri-implantitis. J Clin Periodontol 35, 316,2008.

3. Sokolsky-Papkov, M., Agashi, K., Olaye, A., Shakesheff, K.,and Domb, A.J. Polymer carriers for drug delivery in tissueengineering. Adv Drug Deliv Rev 59, 187, 2007.

4. Carson, J.S., and Bostrom, M.P. Synthetic bone scaffolds andfracture repair. Injury 38 Suppl 1, S33, 2007.

5. Haman, J.D., Scripa, R.N., Rigsbee, J.M., and Lucas, L.C.Production of thin calcium phosphate coatings from glasssource materials. J Mater Sci Mater Med 13, 175, 2002.

6. Linkhart, T.A., Mohan, S., and Baylink, D.J. Growth factorsfor bone growth and repair: IGF, TGF beta and BMP. Bone19, 1S, 1996.

7. Boden, S.D., Kang, J., Sandhu, H., and Heller, J.G. Use ofrecombinant human bone morphogenetic protein-2 toachieve posterolateral lumbar spine fusion in humans: aprospective, randomized clinical pilot trial: 2002 VolvoAward in clinical studies. Spine 27, 2662, 2002.

8. Govender, S., Csimma, C., Genant, H.K., Valentin-Opran,A., Amit, Y., Arbel, R., Aro, H., Atar, D., Bishay, M., Borner,M.G., Chiron, P., Choong, P., Cinats, J., Courtenay, B., Fei-bel, R., Geulette, B., Gravel, C., Haas, N., Raschke, M.,Hammacher, E., van der Velde, D., Hardy, P., Holt, M.,Josten, C., Ketterl, R.L., Lindeque, B., Lob, G., Mathevon, H.,McCoy, G., Marsh, D., Miller, R., Munting, E., Oevre, S.,Nordsletten, L., Patel, A., Pohl, A., Rennie, W., Reynders, P.,Rommens, P.M., Rondia, J., Rossouw, W.C., Daneel, P.J.,Ruff, S., Ruter, A., Santavirta, S., Schildhauer, T.A., Gekle,C., Schnettler, R., Segal, D., Seiler, H., Snowdowne, R.B.,Stapert, J., Taglang, G., Verdonk, R., Vogels, L., Weckbach,A., Wentzensen, A., and Wisniewski, T. Recombinanthuman bone morphogenetic protein-2 for treatment of opentibial fractures: a prospective, controlled, randomized studyof four hundred and fifty patients. J Bone Joint Surg Am84-A, 2123, 2002.

9. Kaneko, H., Arakawa, T., Mano, H., Kaneda, T., Ogasawara,A., Nakagawa, M., Toyama, Y., Yabe, Y., Kumegawa, M.,and Hakeda, Y. Direct stimulation of osteoclastic bone re-sorption by bone morphogenetic protein (BMP)-2 and ex-pression of BMP receptors in mature osteoclasts. Bone 27,

479, 2000.10. Uludag, H., D’Augusta, D., Palmer, R., Timony, G., and

Wozney, J. Characterization of rhBMP-2 pharmacokineticsimplanted with biomaterial carriers in the rat ectopic model.J Biomed Mater Res 46, 193, 1999.

11. Aono, A., Hazama, M., Notoya, K., Taketomi, S., Yamasaki,H., Tsukuda, R., Sasaki, S., and Fujisawa, Y. Potent ectopicbone-inducing activity of bone morphogenetic protein-4=7heterodimer. Biochem Biophys Res Commun 210, 670, 1995.

12. Israel, D.I., Nove, J., Kerns, K.M., Kaufman, R.J., Rosen, V.,Cox, K.A., and Wozney, J.M. Heterodimeric bone morpho-genetic proteins show enhanced activity in vitro and in vivo.Growth Factors 13, 291, 1996.

13. Zhu, W., Rawlins, B.A., Boachie-Adjei, O., Myers, E.R.,Arimizu, J., Choi, E., Lieberman, J.R., Crystal, R.G., andHidaka, C. Combined bone morphogenetic protein-2 and -7gene transfer enhances osteoblastic differentiation and spinefusion in a rodent model. J Bone Miner Res 19, 2021, 2004.

14. Zhao, M., Zhao, Z., Koh, J.T., Jin, T.C., and Franceschi, R.T.Combinatorial gene therapy for bone regeneration: cooper-ative interactions between adenovirus vectors expressingbone morphogenetic proteins 2, 4, and 7. J Cell Biochem 95,

1, 2005.

15. Lind, M., Eriksen, E.F., and Bunger, C. Bone morphogeneticprotein-2 but not bone morphogenetic protein-4 and -6 stim-ulates chemotactic migration of human osteoblasts, humanmarrow osteoblasts, and U2-OS cells. Bone 18, 53, 1996.

16. Fiedler, J., Roderer, G., Gunther, K.P., and Brenner, R.E.BMP-2, BMP-4, and PDGF-bb stimulate chemotactic migra-tion of primary human mesenchymal progenitor cells. J CellBiochem 87, 305, 2002.

17. Suzuki, A., Ghayor, C., Guicheux, J., Magne, D., Quillard, S.,Kakita, A., Ono, Y., Miura, Y., Oiso, Y., Itoh, M., and Ca-verzasio, J. Enhanced expression of the inorganic phosphatetransporter Pit-1 is involved in BMP-2-induced matrix min-eralization in osteoblast-like cells. J Bone Miner Res 21, 674,2006.

18. Atienza, J.M., Yu, N., Kirstein, S.L., Xi, B., Wang, X., Xu, X.,and Abassi, Y.A. Dynamic and label-free cell-based assaysusing the real-time cell electronic sensing system. AssayDrug Dev Technol 4, 597, 2006.

19. Ke, Y., Wu, D., Princen, F., Nguyen, T., Pang, Y., Lesperance,J., Muller, W.J., Oshima, R.G., and Feng, G.S. Role of Gab2 inmammary tumorigenesis and metastasis. Oncogene 26,

4951, 2007.20. Horii, A., Wang, X., Gelain, F., and Zhang, S. Biological

designer self-assembling peptide nanofiber scaffolds signif-icantly enhance osteoblast proliferation, differentiation and3-D migration. PLoS ONE 2, e190, 2007.

21. O’Connor, M.B., Umulis, D., Othmer, H.G., and Blair, S.S.Shaping BMP morphogen gradients in the Drosophila em-bryo and pupal wing. Development 133, 183, 2006.

22. Leboy, P.S. Regulating bone growth and developmentwith bone morphogenetic proteins. Ann N Y Acad Sci 1068,

14, 2006.23. Pan, Q., Yang, S., Dong, Q., and Sun, F. Relationship be-

tween the bioactivity of BMP2=7 heterodimers and its reg-ulation of CIZ expression. China Biotechnol 27, 14, 2007.

24. Little, S.C., and Mullins, M.C. Bone morphogenetic proteinheterodimers assemble heteromeric type I receptor complexesto pattern the dorsoventral axis. Nat Cell Biol 11, 637, 2009.

25. Pan, Q., Yang, S., and Sun, F. [Expression of BMP-2=7 het-erodimers with potent ability to stimulate osteogenic dif-ferentiation in CHO cells]. Sheng Wu Gong Cheng Xue Bao24, 473, 2008.

26. Cheng, H., Jiang, W., Phillips, F.M., Haydon, R.C., Peng, Y.,Zhou, L., Luu, H.H., An, N., Breyer, B., Vanichakarn, P.,Szatkowski, J.P., Park, J.Y., and He, T.C. Osteogenic activityof the fourteen types of human bone morphogenetic proteins(BMPs). J Bone Joint Surg Am 85-A, 1544, 2003.

27. Nohe, A., Hassel, S., Ehrlich, M., Neubauer, F., Sebald, W.,Henis, Y.I., and Knaus, P. The mode of bone morphogeneticprotein (BMP) receptor oligomerization determines differentBMP-2 signaling pathways. J Biol Chem 277, 5330, 2002.

28. Yu, P.B., Beppu, H., Kawai, N., Li, E., and Bloch, K.D. Bonemorphogenetic protein (BMP) type II receptor deletionreveals BMP ligand-specific gain of signaling in pulmonaryartery smooth muscle cells. J Biol Chem 280, 24443, 2005.

29. Lavery, K., Swain, P., Falb, D., and Alaoui-Ismaili, M.H.BMP-2=4 and BMP-6=7 differentially utilize cell surface re-ceptors to induce osteoblastic differentiation of human bonemarrow-derived mesenchymal stem cells. J Biol Chem 283,

20948, 2008.30. Zhu, W., Kim, J., Cheng, C., Rawlins, B.A., Boachie-Adjei,

O., Crystal, R.G., and Hidaka, C. Noggin regulation of bonemorphogenetic protein (BMP) 2=7 heterodimer activityin vitro. Bone 39, 61, 2006.

886 ZHENG ET AL.

31. Shen, Z.J., Nakamoto, T., Tsuji, K., Nifuji, A., Miyazono, K.,Komori, T., Hirai, H., and Noda, M. Negative regulationof bone morphogenetic protein=Smad signaling by Cas-interacting zinc finger protein in osteoblasts. J Biol Chem277, 29840, 2002.

32. Lyons, K.M., Hogan, B.L., and Robertson, E.J. Colocalizationof BMP 7 and BMP 2 RNAs suggests that these factorscooperatively mediate tissue interactions during murinedevelopment. Mech Dev 50, 71, 1995.

33. Katagiri, T., Boorla, S., Frendo, J.L., Hogan, B.L.M., andKarsenty, G. Skeletal abnormalities in doubly heterozygousBmp4 and Bmp7 mice. Dev Genet 22, 340, 1998.

34. Liu, Y., Enggist, L., Kuffer, A.F., Buser, D., and Hunziker,E.B. The influence of BMP-2 and its mode of delivery on theosteoconductivity of implant surfaces during the early phaseof osseointegration. Biomaterials 28, 2677, 2007.

35. Liu, Y., Huse, R.O., de Groot, K., Buser, D., and Hunziker,E.B. Delivery mode and efficacy of BMP-2 in associationwith implants. J Dent Res 86, 84, 2007.

36. Koh, J.T., Zhao, Z., Wang, Z., Lewis, I.S., Krebsbach, P.H.,and Franceschi, R.T. Combinatorial gene therapy withBMP2=7 enhances cranial bone regeneration. J Dent Res 87,

845, 2008.

Address correspondence to:Zhiyuan Gu, D.D.S., Ph.D.

School=Hospital of StomatologyZhejiang University

395 Yan’an RoadHangzhou

Zhejiang 310006P.R. China

E-mail: gzy@zju.edu.cn

Received: May 5, 2009Accepted: October 7, 2009

Online Publication Date: January 27, 2010

ACTION OF rhBMP2/7 HETERODIMER IN OSTEOBLASTOGENESIS 887