The Osteoconductivity of Biomaterials Is Regulatedby Bone Morphogenetic Protein 2 Autocrine Loop

Involving a2b1 Integrin and Mitogen-ActivatedProtein Kinase/Extracellular Related

Kinase Signaling Pathways

ZuFu Lu, Ph.D., M.D., and Hala Zreiqat, Ph.D.

It is critical to understand the complex interactions between cells and scaffolds for a successful tissue engineeringapproach for bone regeneration. Beyond providing structural support for the cells, synthetic scaffolds act to-gether with some soluble biofactors through intracellular signaling pathways to provide the appropriate cluesfor cells to form bone tissue. The aim of this study was to investigate the mechanism by which beta-tricalciumphosphate (b-TCP), a clinically used bone graft substitute, exerts its osteoconductivity on primary humanosteoblasts. Culturing human osteoblasts on b-TCP scaffold for 1 and 7 days induced gene expression of bonemorphogenetic protein 2 (BMP2) and its receptors and activated its downstream Smad1/5 signaling pathway,which were orchastrated with induced osteoblastic differentiation. Blocking BMP2 activity by its inhibitor(Noggin) led to the abrogation of osteoblastic differentiation and partially inhibited Smad1/5 signaling pathway.Finally, blocking a2b1 integrin or inhibiting mitogen-activated protein kinase/extracellular related kinase sig-naling pathway attenuated the induction of gene expression of BMP2 and its receptors and the activation ofSmad1/5 signaling pathway. We concluded that b-TCP scaffold promotes osteoblastic differentiation by a BMP2autocrine loop, a process involving a2b1 integrin and mitogen-activated protein kinase/extracellular relatedkinase signaling pathways. The findings of this study might provide a useful principle for fabricating or de-signing an ideal scaffold for bone tissue engineering.

Introduction

Bone regeneration is critical to a broad range of ap-plications, including nonunion fracture, maxillofacial

reconstruction, spinal fusion, and repair of defects caused byinjury, cancer, and congenital abnormalities.1 Clinically, re-placing extensive local bone defects presents a significantchallenge. Synthetic bone graft substitutes (scaffolds) can beused to replace, repair, or regenerate the lost, injured, ordiseased bone. However, the synthetic materials currently inuse are far from optimal. The interaction between cells andscaffolds plays a critical role in determining the fate of see-ded cells on the scaffolds.2–4 Although numerous studieshave so far investigated the responses of osteoblasts toscaffolds, little is known about the intracellular events thatoccur after the initial adhesion of osteoblasts to biomaterials,and those that facilitate endosseous integration.4 Synthetic

scaffolds for use in bone tissue regeneration should elicitcertain cellular responses and provide the appropriate cluesfor cells to form bone tissue.

Tissues in the musculoskeletal system differ greatly infunction and phenotype due to the exposure to distinct tissuemicroenvironment, which is a mixture of soluble chemo-kines, cytokines, and growth factors, as well as insolubletransmembrane receptor ligands and extracellular matrix(ECM) molecules. The roles of the tissue microenvironmentin determining cell fate have been extensively exploredduring the last decade, including biological, chemical, andphysical signals.5–7 Mesemchymal stem cells co-culturedwith different primary cells such as osteoblasts or chon-drocytes can be directed into osteogenic8,9 and chondrogenicdifferentiation,10,11 respectively; in addition, the ECM mole-cules from bone or cartilage tissue provide the instructivesignals for osteogenic12 and chondrogenic differentiation,13

Biomaterials and Tissue Engineering Research Unit, School of AMME, The University of Sydney, Sydney, Australia.

TISSUE ENGINEERING: Part AVolume 16, Number 10, 2010ª Mary Ann Liebert, Inc.DOI: 10.1089/ten.tea.2010.0204

3075

respectively. However, apart from understanding the rolesfrom single factor, it is of great importance to elucidate theinteractions between the different factors in the tissue niches,since there is a sophisticated and/or synergistic crosstalkbetween these factors.14,15

Tissue ECM molecules have many functions beyond pro-viding structural support to the cells. They transduce signalsinto cells through binding cellular adhesion receptor such asintegrins, and integrin-mediated adhesion to extracellularproteins has been shown to transduce signals from the ECMto the interior of the cell, thus regulating differentiation andmorphology in different cell types, including bone cells.16,17

Bone tissue is composed mainly of type I collagen (90%) andnoncollagenous proteins (10%, including osteopontin, bonesialoprotein [BSP], osteocalcin, osteonectin, fibronectin, andlaminin18) that harbor various growth factors such as bonemorphogenetic proteins (BMPs) and determine the bone-specific microenvironment regulating the bone regenerationand degeneration process.19 A line of studies showed thattissue ECM molecules can bind some soluble growth factorssuch as BMPs and regulate their distribution, activation, andpresentation to the cells.20,21 It comes with no surprise thatthere is a close cross-talk between bone tissue ECM mole-cules and the embedded growth factors such as BMPs. Su-zawa and his colleagues showed that type I collagenpromotes osteogenic differentiation by modulating the BMPsignaling pathway.22 In the present study, we hypothesizedthat synthetic scaffold for bone tissue engineering should actin a similar way to that for natural bone ECM molecules bymodulating some cellular signaling pathways mediated bysoluble factors such as BMPs.

This study was aimed at investigating the mechanisms bywhich beta-tricalcium phosphate (b-TCP), a clinically usedbone graft substitute, exerts its in vitro osteoconductivity inprimary human osteoblasts (HOBs). Our results showed thatseeding HOBs on b-TCP scaffold induced the mRNA ex-pression levels of BMP2 and its receptors (BMPR-1a, BMPR-1b, and BMPR2) and activated BMP2 downstream Smad1signaling pathway, which were orchestrated with the pro-motion of osteoblastic differentiation. In addition, weshowed that the treatment of HOBs with the BMP2 inhibitor(Noggin) led to the abrogation of osteoblastic differentiationand partially inhibited Smad1/5 signaling pathway. Finally,either blocking a2b1 integrin by neutralizing antibodies orinhibiting mitogen-activated protein kinase (MAPK)/extracellular related kinase (ERK) signaling pathway by in-hibitor (PD98059) attenuated the induction of gene ex-pression of BMP2 and its receptors and the activation ofSmad1/5 signaling pathway.

Materials and Methods

Isolation and culture of primary HOBs

Permission to use discarded human tissue was granted bythe Human Ethics Committee of the University of Sydney,and informed consent was obtained. HOBs were isolatedfrom normal human trabecular bone as previously de-scribed.23 Briefly, bone was divided into 1 mm3 pieces, wa-shed several times in phosphate-buffered saline (PBS), anddigested for 90 min at 378C with 0.02% (w/v) trypsin(Sigma–Aldrich) in PBS. Digested cells were cultured in acomplete medium containing a-minimal essential medium

(Gibco Laboratories), supplemented with 10% (v/v) heat-inactivated fetal calf serum (Gibco Laboratories), 2 mMl-glutamine (Gibco Laboratories), 25 mM Hepes Buffer(Gibco Laboratories), 2 mM sodium pyruvate, 30 mg/mLpenicillin, 100 mg/mL streptomycin (Gibco Laboratories),and 1 mM l-ascorbic acid phosphate magnesium salt (WakoPure Chemicals). The cells were cultured at 378C with 5%CO2, and the medium was refreshed every 3 days untilconfluence when cells were passaged.

HOB attachment, BMPs, a2b1 integrin,and MAPK/ERK signaling blocking

All HOBs used in the experiments were at passage 3. At80%–90% confluence, cells were trypsinized with trypsina-zer/ethylenediaminetetraacetic acid (Invitrogen) and subse-quently suspended in the complete medium. HOBs wereseeded on b-TCP scaffolds (cylindrical shape with diameterof 5 mm and height of 3 mm, pore diameters range from 1 to100 mm; Vitoss) at a density of 200,000 cells per scaffold. Acell suspension of 100 mL was dropped gently into the scaf-folds (n¼ 4) to ensure that all cells adhered to the scaffoldand not to the surrounding tissue culture plastic, and incu-bated for 90 min at 378C to allow cells to attach to the scaf-folds only, before flooding with the cell culture medium.

HOBs were cultured on the b-TCP scaffolds for 24 h wherethe cells were rinsed three times in PBS and fixed in 1.25%glutaraldehyde, and postfixed in osmium tetroxide followedby sequential dehydration in graded ethanol. The sampleswere then dried in hexamethyldisilizane and coated withgold for scanning electron microscopy (SEM) analysis usingFE-SEM (Zeiss Ultra).

BMP2 inhibition, and a2b1 integrin and MAPK/ERKsignaling blocking were performed by replacing the culturemedium with a medium containing Noggin (500 ng/mL;Sigma), neutralizing a2b1 integrin antibody (10mg/mL;Chemicon), or PD98059 (20mM; Sigma), respectively, andcultured for a designated timeframe. The medium containingnormal immunoglobulin G (10mg/mL; Chemicon) or di-methyl sulfoxide (Sigma) was used as a control for a2b1 in-tegrin blocking or MAPK/ERK signaling blocking.

Quantitative real-time polymerase chain reaction

Total RNA was isolated from HOBs on the b-TCP scaf-folds by directly adding Trizol reagent (Sigma) after themedium were removed and purified by RNeasy Mini Kitfrom Qiagen (Valencia) according to the manufacturer’s in-structions. First-strand cDNA was synthesized from 0.7 mgtotal RNA using the Omniscript RT Kit (Qiagen) according tothe manufacturer’s instructions. cDNA was analyzed for theosteoblast-related genes (runt-related transcription factor 2[Runx2], osteopontin, osteocalcin, and BSP), BMP2, BMPR-1a, BMPR-1b, and BMPR-2. Their relative gene expressionlevels were obtained by normalizing to the house-keepinggene (glyceraldehyde 3-phosphate dehydrogenase). Primersfor the selected genes are listed in Table 1.

Western blotting

Western blotting was performed on the HOBs seeded onthe b-TCP and tissue culture plastic for 6 and 24 h. At thepredetermined times point, cells were washed with ice-cold

3076 LU AND ZREIQAT

PBS and lysed for 30 min in ice-cold Radio ImmunoPrecipitation Assay (RIPA) lysis buffer (20 mM Tris–HCl [pH7.5], 1 mM ethylenediaminetetraacetic acid, 1 mM ethyleneglycol tetraacetic acid (EGTA), 150 mM NaCl, and 1%Triton X-100, protease inhibitor cocktail [Sigma], and phos-phatase inhibitor cocktail 2 [Roche]). Protein concentrationwas measured using the Bicinchoninic Acid Assay (BCA)protein assay kit (Bio-Rad). Equal aliquots of protein (10mg)were heated at 708C for 10 min in 4�sample buffer(WesternBreeze; Invitrogen) and 10�reducing buffer,and separated on 8%–12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis gels (WesternBreeze;Invitrogen). Proteins were transferred to polyvinylidenefluoride membranes, washed with 1�TBS-T (20 mM Tris–HCl [pH 7.6] and 137 mM NaCl, containing 0.1% Tween 20),and blocked for 1 h at room temperature in 1�TBS-T with 1%bovine serum albumin. The membranes were washed threetimes followed by incubation with primary antibody (antiSmad1, 1:1000; anti-phospho-Smad1/5, 1:500; Cell Signaling)in TBS-T containing 1% bovine serum albumin overnight at48C. After three washes, the membranes were incubated withsecondary antibody (WesternBreeze; Invitrogen) for 60 min,followed by another three washes before protein bands wereobserved with chemiluminescent reagents (WesternBreeze;Invitrogen) in Alpha Innotech Digital Imaging System(Alpha Innotech).

Statistical analysis

Data in this study were obtained from four independentexperiments and represented as mean� standard error. Forstatistical analysis, first, Levene’s test was performed to de-termine the homogeneity of variance for all the data, andthen independent t-test tests were performed for analyzingthe data from two groups; otherwise, Tukey Honestly Sig-nificant Difference (HSD) post hoc tests were used. SPSS 17.0

program was employed for all statistical analysis and dif-ferences were considered significant if p< 0.05.

Results

HOBs attached on b-TCP scaffold and showedincreased osteoblastic, BMP2, and its receptors’gene expression and the activation of Smad1/5signaling pathway

SEM showed that HOBs have attached to b-TCP scaffoldand spread well after 24 h of culturing (Fig. 1A). To validatethe osteoconductivity of b-TCP scaffolds, HOBs were seededon b-TCP scaffolds for 1 and 7 days, and the mRNA expres-sion levels for Runx2, osteopontin, BSP, and osteocalcin wereanalyzed using quantitative real-time polymerase chain re-action by comparing to the HOBs before seeding (day 0).Results showed that osteogenic gene expression in HOBsseeded on b-TCP scaffolds significantly increased over thetime from day 0 to day 7, whereby after 7 days of culture, therewere 7-fold increase in Runx2, 10-fold increase in osteopontin,and 5-fold increase in BSP and osteocalcin (Fig. 1B).

The gene expression levels of BMP2 and its receptors andthe activation of Smad1 signaling in HOBs were investigatedin an attempt to reveal the possible mechanisms that deter-mine the osteogenic behavior of HOBs when interacting withb-TCP scaffold. After HOBs were seeded in b-TCP scaffoldsfor 6 and 24 hours, gene expression of BMP2 and its receptors(BMPR-1a, BMPR-1b, and BMPR2) were measured by quan-titative real-time–polymerase chain reaction. Orchestratedwith the induction of osteogenic gene expression, we showedthat HOBs seeded on b-TCP scaffolds showed significantlyhigher gene expression for BMP2, BMPR-1a, BMPR-1b, andBMPR2 after 6 and/or 24 hours comparing to levels obtainedat day 0 (Fig. 1C, D). In addition, b-TCP scaffolds activatedSmad1/5 signaling pathway in HOBs, as demonstrated by thesteady increase in the amount of phosphorylated Smad1/5protein from day 0 to day 7 (Fig. 1E).

Autocrine BMP2 is involved in the activation of Smad1/5signaling pathway and osteoblastic differentiation

Since b-TCP scaffolds can stimulate the synthesis of en-dogenous BMP2 and Smad1/5 signaling pathway in HOBs,we asked whether the autocrine BMP2 contributes to theactivation of Smad1/5 signaling pathway and to the osteo-conductivity of b-TCP scaffolds. The inhibitor of BMPsignaling pathway (Noggin) was utilized for this purpose.HOBs were cultured on b-TCP scaffolds in the mediumsupplemented with or without Noggin, and the resultsshowed that the treatment with Noggin abrogated expres-sion of early HOBs’ differentiation-related gene expression(Runx2 and osteopontin) at days 1 and 7, but no effect wasseen for the later differentiation markers (BSP and osteo-calcin) (Fig. 2A–D). In addition, Noggin partially attenuatedthe activation of Smad1/5 signaling pathway with about20% decrease in phosphorylated Smad1/5 protein at day 7,with no apparent difference at day 1 (Fig. 2E).

a2b1 integrin mediates the induction of autocrineBMP2 and the activation of Smad1 signaling pathway

To understand how the autocrine BMP2 is triggered bythe interactions between HOBs and b-TCP scaffolds, we

Table 1. Primers Used for Real-Time Polymerase

Chain Reaction

Gene Sequence (50–30)

Meltingtemperature

(8C)

GAPDH F ACCCAGAAGACTGTGGATGG 60R CAGTGAGCTTCCCGTTCAG

Runx2 F ATGCTTCATTCGCCTCAC 60R ACTGCTTGCAGCCTTAAAT

Osteopontin F TTCCAAGTAAGTCCAACGAAAG 60R GTGACCAGTTCATCAGATTCAT

Osteocalcin F ATGAGAGCCCTCACACTCCTCG 60R GTCAGCCAACTCGTCACAGTCC

BSP F ATGGCCTGTGCTTTCTCAATG 60R GGATAAAAGTAGGCATGCTTG

BMP2 F AGTTGCGGCTGCTCAGCATGTT 60R CCGGGTTGTTTTCCCACT

BMPR-1a F TTTATGGCACCCAAGGAAAG 60R TGGTATTCAAGGGCACATCA

BMPR-1b F AAAGGTCGCTATGGGGAAGT 60R GCAGCAATGAAACCCAAAAT

BMPR-2 F CATCCGAACCCTCTCTTGAT 60R TGCATAAAGATCCATTGGGA

GAPDH, glyceraldehyde 3-phosphate dehydrogenase; BSP, bonesialoprotein; BMP2, bone morphogenetic protein 2; BMPR, BMP receptor;Runx2, runt-related transcription factor 2.

OSTEOCONDUCTIVITY IS REGULATED BY BMP2 AUTOCRINE LOOP 3077

investigated the modulation of a2b1 integrin, which plays acrucial role in mediating osteoblastic differentiation con-ducted by b-TCP scaffold.24 To this end, HOBs cultured in b-TCP scaffolds were grown in the medium with or withoutthe supplementation of a2b1 integrin-neutralizing antibodiesfor 6 and 24 h. Results showed that a2b1 integrin blockinginhibited BMP2 gene expression but did not affect BMP2receptors’ gene expression at 6 h of incubation (Fig. 3A–D).At 24 h, BMP2, BMPR-1a, BMPR-1b, and BMPR-2 mRNAlevels decreased by three- to fourfolds when the cultureswere treated with a2b1 integrin-neutralizing antibody (Fig.3A–D). Blocking a2b1 integrin inhibited phosphorynatedSmad1/5 protein expression by about 20% after 7 days ofculturing HOBs on the b-TCP scaffolds (Fig. 3E).

MAPK/ERK signaling pathway participatesin the induction of autocrine BMP2 and the activationof Smad1 signaling pathway

We investigated the role of downstream signaling ofMAPK/ERK signaling pathway for the HOBs cultured on b-TCP scaffolds. To address this, HOBs cultured on b-TCP

scaffolds were grown in the medium with or without thesupplement of MAPK/ERK signaling inhibitor (PD98059) for6 and 24 h. We showed that the treatment of PD98059 sig-nificantly inhibited both BMP2 and BMPR-1a gene expres-sion after 6 h of incubation. This trend was maintained after24 h, though to a lesser extent than the effect seen with a2b1integrin inhibition, where only 30% decrease in BMP recep-tors was found by the treatment of the culture medium withPD98059 (Fig. 4A–D). Moreover, phosphorynated Smad1/5protein level was markedly lower (>50%) in PD98059-treated group than in the control group at day 7 (Fig. 4E).

Discussion

Providing an instructive and/or conductive signal for di-recting stem cells/progenitor cells into a specific lineagedifferentiation is one of the criteria for designing the idealscaffold for bone tissue engineering. Over the last decade,efforts have been made in an attempt to decipher the mys-terious mechanisms by which a specific scaffold or ECMmolecule drives stem cells/progenitor cells into a specificlineage differentiation.25 Despite gaining some insight into

FIG. 1. Attachment of HOBs to b-TCP scaffold and the effects of b-TCP scaffoldon osteoblastic, BMP2, BMP receptor gene expression, and Smad1/5 signalingpathway in HOBs cultured on b-TCP scaffold. HOBs attached to b-TCP scaffoldand spread well after 24 h of culturing (A). Seeding HOBs on b-TCP scaffoldsignificantly increased osteoblastic gene expression (Runx2, Osteopontin, BSP, andOsteocalcin) after 1 and/or 7 days of culturing (B), and significantly boosted thegene expression levels of BMP2 and BMP receptors after 6 and/or 24 h of culturing(C, D) and dramatically increased phosphorylated Smad1/5 protein levels after 1and 7 days of culturing (E). *p< 0.05. HOBs, human osteoblasts; b-TCP, beta-tricalcium phosphate; BMP2, bone morphogenetic protein 2; BSP, bone sialopro-tein; GAPDH, glyceraldehyde 3-phosphate dehydrogenase.

3078 LU AND ZREIQAT

these mechanisms2,7,25,26 over the years, a great deal of re-search is required to gain a better understanding of themechanisms that determine cell interaction with syntheticbiomaterials. We now know that the physical property of thescaffolds or ECM molecules play crucial role in controllingcells’ differentiation fate by modulating their shape.27 Resultspresented in this study demonstrate that b-TCP scaffold, awidely used bone graft substitute, promotes osteoblasticdifferentiation through triggering a BMP2 autocrine loopinvolving a2b1 integrin and MAPK/ERK signaling path-ways.

In vitro, different ECM molecules or growth factors play arole in modulating osteoblasts phenotype.28,29 The capabilityof promoting or preserving osteoblastic phenotype by scaf-folds is defined as osteoconductivity. The osteoconductive

property of scaffolds plays a key role for a successful bonegraft implantation in vivo since bone growth on the implantsurface largely depends on the interactions between thescaffolds and preosteoblasts or osteoblasts. b-TCP scaffold isa clinically used synthetic graft because of its osteo-conductivity and osteoinductivity.30–34 In this study, b-TCPscaffold was utilized to explore the molecular mechanismsby which the scaffold provides the osteoconductive signalsto osteoblasts. First, we confirmed that b-TCP scaffold in-deed is able to promote or preserve the osteoblastic pheno-type of HOBs. This was evident by the upregulation of theosteogenic gene expression for Runx2, osteopontin, BSP, andosteocalcin in HOBs when cultured on b-TCP scaffold,compared to the downregulation seen when HOBs werecultured on the tissue plastic.28 Recently, studies have

FIG. 2. Effects of BMP inhibition on osteoblastic gene expression andSmad1/5 signaling pathway in HOBs cultured on b-TCP scaffold.HOBs on b-TCP scaffolds were cultured in the medium supplementedwith or without BMP2 inhibitor (Noggin) for 1 and 7 days. Comparedto the control (without Noggin), (A–D) Noggin significantly decreasedthe gene expression levels of Runx2 and osteopontin after 1 and 7 days,but did not affect mRNA levels of osteocalcin and BSP. (E) Noggindid not affect the phosphorylated Smad1/5 protein levels at day 1, butled to about 20% decrease of phosphorylated Smad1/5 protein atday 7. *p< 0.05.

OSTEOCONDUCTIVITY IS REGULATED BY BMP2 AUTOCRINE LOOP 3079

investigated the underlying mechanisms of the osteo-conductivity of b-TCP30–34; however, little is known aboutthe intracellular events that occur after the initial adhesion ofosteoblasts to synthetic scaffolds, and those that facilitateendosseous integration.

BMPs, members of transforming growth factor family,play a critical role in the process of bone formation and re-modeling.35 The action of BMPs is mediated through het-erotetrameric serine/threonin kinase receptor and thedownstream Smad1/5 signaling pathway; once the signalingpathway is activated, a series of osteogenesis-relative geneexpression is stimulated.36 Many groups have attempted totether BMPs in ECM molecules or synthetic material forboosting bone regeneration.37,38 On the basis of the evidencethat some antiosteoporosis agents can enhance bone forma-tion by increasing endogenous BMP2 synthesis in osteo-

blasts,39–42 we proposed that synthetic scaffolds may providethe osteoinstructive and/or osteoconductive signals by trig-gering the endogenous BMP2 production and its down-stream signaling pathway. In this study, we showed thatb-TCP scaffolds significantly induced BMP2, BMPs recep-tors gene expression levels, and Smad1 signaling pathwayin HOBs. Our results suggest that BMP2 and its down-stream Smad1 signaling pathway might mediate the os-teoconductive functionality of b-TCP scaffold. Thisspeculation was consolidated by our finding indicating thatBMP2 inhibitor (Noggin) inhibited the HOBs differentia-tion when cultured on b-TCP scaffolds. However, phos-phorylation of Smad1/5 protein was only partially (20%)abrogated by Noggin, suggesting that the signaling fromBMPs is not the only contributor of the activation of Smad1signaling pathway. Jadlowiec and his colleagues41 also

FIG. 3. Effects of a2b1 integrin blocking on BMP2, BMP receptors geneexpression, and Smad1/5 signaling pathway in HOBs on b-TCP scaffold.HOBs on b-TCP scaffold were cultured in the medium supplemented withor without a2b1 integrin-neutralizing antibodies for 1 and 7 days. Com-pared to the control (nonspecific immunoglobulin G), (A–D) a2b1 integrinblocking significantly decreased the gene expression levels of BMP2, BMPR-1A, BMPR-1B, and BMPR-2 after 6 and/or 24 hours. (E) a2b1 integrinblocking did not affect the phosphorylated Smad1/5 protein levels at day 1,but resulted in about 20% decrease of phosphorylated Smad1/5 proteinlevels at day 7. *p< 0.05.

3080 LU AND ZREIQAT

found that dentin phosphophoryn activates the Smadpathway via the pathways other than BMP-mediated sig-naling pathway.

As BMP2 autocrine loop plays an important role in oste-oblastic differentiation mediated by b-TCP scaffold, we wereinterested in investigating the mechanisms by which b-TCPscaffold stimulate the synthesis of endogenous BMP2, whichin turn promotes or preserves the osteoblastic phenotype. Itis well known that there is a close crosstalk between thedifferent intracellular signals triggered by soluble factors andsome extracellular ligands such as integrin subunits.43–46 Aline of studies, on one hand, showed that BMPs modulatecell attachment, growth, and differentiation by upregulatinga certain integrin unit and its downstream signaling path-way.47,48 On the other hand, ECM was found to induce ex-pression of some cytockines or growth factors thorugh anintegrin-mediated signaling pathway, which then acts syn-ergistically with these soluble factors in determining the fate

of the cells.22,44,49 a2b1 integrin, one of the main cell receptorsin osteoblasts to interact with the surrounding ECM mole-cules, was found to play a critical role in some substratesand/or scaffold-mediated osteogenic differentiation by en-gaging MAPK/ERK signaling pathway.50,51 We previouslyalso demonstrated that a2b1 integrin and MAPK/ERK sig-naling pathways are involved in the osteoconductivitly of b-TCP scaffold.24 Herein we postulated that a2b1 integrin andMAPK/ERK signaling pathway may play a role in contribut-ing to the osteoconductivitly of b-TCP scaffold by triggering aBMP2 autocrine loop. Our results showed that the induction ofBMP2 and its receptors genes expression and the activation ofits downstream Smad1 signaling pathway were inhibited byblocking a2b1 integrin, suggesting that a2b1 integrin indeed isinvolved in the process of triggering the BMP2 autocrine loop.In addition, we observed that inhibiting MAPK/ERK signalingpathway abrogated the production of BMP2 and its receptorsand more substantially (>50%) abolished the phosphorylated

FIG. 4. Effects of MAPK/ERK inhibition on BMP2, BMP receptors’ geneexpression, and Smad1/5 signaling pathway in HOBs on b-TCP scaffold.HOBs on b-TCP scaffold were cultured in the medium supplemented withor without MAPK/ERK inhibitor (PD98059) for 1 or 7 days. Compared tothe control (DMSO), (A–D) PD98059 significantly decreased the gene ex-pression levels for BMP2, BMPR-1A, BMPR-1B, and BMPR-2 after 6 and/or24 hours. (E) PD98059 did not affect the phosphorylated Smad1/5 proteinlevels at day 1 but decreased phosphorylated Smad1/5 protein levels byabout 50% at day 7. *p< 0.05. MAPK, mitogen-activated protein kinase;ERK, extracellular related kinase; DMSO, dimethyl sulfoxide.

OSTEOCONDUCTIVITY IS REGULATED BY BMP2 AUTOCRINE LOOP 3081

Smad1/5 protein than that seen when Noggin or blockinga2b1 integrin were used, implying that there are other cross-talks between the MAPK/ERK signaling pathway and theSmad1/5 signaling pathway rather than solely through theautocrine BMP2 pathway. A model of osteoblastic differenti-ation conducted by scaffolds is proposed in Figure 5, sug-gesting that scaffolds elicit the BMP2 autocrine loop througha2b1 integrin and MAPK/ERK signaling pathway.

In summary, the current study demonstrates that b-TCPscaffold conducts osteoblastic differentiation by triggering aBMP2 autocrine loop and a2b1 integrin and MAPK/ERKsignaling pathways are involved in this process. To the au-thors’ knowledge, this is the first report indicating thatsynthetic scaffold is able to elicit the endogenous BMP2production through the interactions between osteoblasts andscaffolds via a2b1 integrin and MAPK/ERK signalingpathway, in turn leading to the enhancement of bone re-generation process. The findings of this study might providea useful principle for fabricating or designing an ideal scaf-fold for bone tissue engineering.

Acknowledgments

The authors acknowledge the Australia National Healthand Medical Research Council and the Australian ResearchCouncil for funding this research. The authors also thankStryker Australia Pty. Ltd. for its kind provision of the b-TCPscaffolds.

Disclosure Statement

No competing financial interests exist.

References

1. Redman, S.N., Oldfield, S.F., and Archer, C.W. Currentstrategies for articular cartilage repair. Eur Cell Mater 9,

23, 2005.2. Hynes, R.O. The extracellular matrix: not just pretty fibrils.

Science 326, 1216, 2009.

3. Bettinger, C.J., Langer, R., and Borenstein, J.T. Engineeringsubstrate topography at the micro- and nanoscale to controlcell function. Angew Chem Int Ed Engl 48, 5406, 2009.

4. Marquis, M.E., Lord, E., Bergeron, E., Drevelle, O., Park, H.,Cabana, F., Senta, H., and Faucheux, N. Bone cells-biomaterials interactions. Front Biosci 14, 1023, 2009.

5. Bratt-Leal, A.M., Carpenedo, R.L., and McDevitt, T.C. En-gineering the embryoid body microenvironment to direct em-bryonic stem cell differentiation. Biotechnol Prog 25, 43, 2009.

6. Spiegel, A., Kalinkovich, A., Shivtiel, S., Kollet, O., and La-pidot, T. Stem cell regulation via dynamic interactions of thenervous and immune systems with the microenvironment.Cell Stem Cell 3, 484, 2008.

7. Tenney, R.M., and Discher, D.E. Stem cells, microenviron-ment mechanics, and growth factor activation. Curr OpinCell Biol 21, 630, 2009.

8. Ilmer, M., Karow, M., Geissler, C., Jochum, M., and Neth, P.Human osteoblast-derived factors induce early osteogenicmarkers in human mesenchymal stem cells. Tissue Eng PartA 15, 2397, 2009.

9. Csaki, C., Matis, U., Mobasheri, A., and Shakibaei, M. Co-culture of canine mesenchymal stem cells with primarybone-derived osteoblasts promotes osteogenic differentia-tion. Histochem Cell Biol 131, 251, 2009.

10. Vats, A., Bielby, R.C., Tolley, N., Dickinson, S.C., Boccaccini,A.R., Hollander, A.P., Bishop, A.E., and Polak, J.M. Chon-drogenic differentiation of human embryonic stem cells:the effect of the micro-environment. Tissue Eng 12, 1687, 2006.

11. Lu, Z.F., Doulabi, B.Z., Wuisman, P.I., Bank, R.A., andHelder, M.N. Influence of collagen type II and nucleuspulposus cells on aggregation and differentiation of adiposetissue-derived stem cells. J Cell Mol Med 12, 2812, 2008.

12. Thibault, R.A., Baggett, L.S., Mikos, A.G., and Kasper, F.K.Osteogenic differentiation of mesenchymal stem cells on pre-generated extracellular matrix scaffolds in the absence ofosteogenic cell culture supplements. Tissue Eng Part A, 2009[Epub ahead of print].

13. Lu, Z., Doulabi, B.Z., Huang, C., Bank, R.A., and Helder,M.N. Collagen type II enhances chondrogenesis in adiposetissue-derived stem cells by affecting cell shape. Tissue EngPart A 16, 81, 2010.

14. LaBarge, M.A., Nelson, C.M., Villadsen, R., Fridriksdottir,A., Ruth, J.R., Stampfer, M.R., Petersen, O.W., and Bissell,M.J. Human mammary progenitor cell fate decisions areproducts of interactions with combinatorial microenviron-ments. Integr Biol (Camb) 1, 70, 2009.

15. Discher, D.E., Mooney, D.J., and Zandstra, P.W. Growthfactors, matrices, and forces combine and control stem cells.Science 324, 1673, 2009.

16. Alam, N., Goel, H.L., Zarif, M.J., Butterfield, J.E., Perkins,H.M., Sansoucy, B.G., Sawyer, T.K., and Languino, L.R. Theintegrin-growth factor receptor duet. J Cell Physiol 213,

649, 2007.17. Yamashita, H., Tripathi, M., Harris, M.P., Liu, S., Weidow, B.,

Zent, R., and Quaranta, V. The role of a recombinant fragmentof laminin-332 in integrin alpha3beta1-dependent cell binding,spreading and migration. Biomaterials 31, 5110, 2010.

18. Allori, A.C., Sailon, A.M., and Warren, S.M. Biological basisof bone formation, remodeling, and repair-part II: extracel-lular matrix. Tissue Eng Part B Rev 14, 275, 2008.

19. Hoshiba, T., Kawazoe, N., Tateishi, T., and Chen, G. De-velopment of stepwise osteogenesis-mimicking matrices forthe regulation of mesenchymal stem cell functions. J BiolChem 284, 31164, 2009.

FIG. 5. A schematic illustration of osteoblastic differentia-tion triggered by scaffold/ECM through a BMP2 autocrineloop. b-TCP scaffold exerts osteoconductive functionality byeliciting the production of endogenous BMP2 and activatingSmad1/5 signaling pathway, which involve a2b1 integrinand its downstream MAPK/ERK signaling pathways. ECM,extracellular matrix.

3082 LU AND ZREIQAT

20. Wang, X., Harris, R.E., Bayston, L.J., and Ashe, H.L. Type IVcollagens regulate BMP signalling in Drosophila. Nature455, 72, 2008.

21. Zhu, Y., Oganesian, A., Keene, D.R., and Sandell, L.J. TypeIIA procollagen containing the cysteine-rich amino propep-tide is deposited in the extracellular matrix of pre-chondrogenic tissue and binds to TGF-beta1 and BMP-2.J Cell Biol 144, 1069, 1999.

22. Suzawa, M., Tamura, Y., Fukumoto, S., Miyazono, K., Fujita,T., Kato, S., and Takeuchi, Y. Stimulation of Smad1 tran-scriptional activity by Ras-extracellular signal-regulatedkinase pathway: a possible mechanism for collagen-dependent osteoblastic differentiation. J Bone Miner Res 17,

240, 2002.23. Zreiqat, H., Valenzuela, S.M., Nissan, B.B., Roest, R., Knabe,

C., Radlanski, R.J., Renz, H., and Evans, P.J. The effect ofsurface chemistry modification of titanium alloy on signal-ling pathways in human osteoblasts. Biomaterials 26, 7579,2005.

24. Lu, Z., and Zreiqat, H. Beta-tricalcium phosphate exertsosteoconductivity through alpha2beta1 integrin and down-stream MAPK/ERK signaling pathway. Biochem BiophysRes Commun 394, 323, 2010.

25. Lutolf, M.P., Gilbert, P.M., and Blau, H.M. Designing ma-terials to direct stem-cell fate. Nature 462, 433, 2009.

26. Czyz, J., and Wobus, A. Embryonic stem cell differentia-tion: the role of extracellular factors. Differentiation 68, 167,2001.

27. Engler, A.J., Sen, S., Sweeney, H.L., and Discher, D.E. Matrixelasticity directs stem cell lineage specification. Cell 126, 677,2006.

28. Chung, C.Y., Iida-Klein, A., Wyatt, L.E., Rudkin, G.H., Ish-ida, K., Yamaguchi, D.T., and Miller, T.A. Serial passage ofMC3T3-E1 cells alters osteoblastic function and responsive-ness to transforming growth factor-beta1 and bone mor-phogenetic protein-2. Biochem Biophys Res Commun 265,

246, 1999.29. Calvert, J.W., Chua, W.C., Gharibjanian, N.A., Dhar, S., and

Evans, G.R. Osteoblastic phenotype expression of MC3T3-E1 cells cultured on polymer surfaces. Plast Reconstr Surg116, 567, 2005.

30. Kasten, P., Luginbuhl, R., van Griensven, M., Barkhausen,T., Krettek, C., Bohner, M., and Bosch, U. Comparison ofhuman bone marrow stromal cells seeded on calcium-deficient hydroxyapatite, beta-tricalcium phosphate anddemineralized bone matrix. Biomaterials 24, 2593, 2003.

31. Marino, G., Rosso, F., Cafiero, G., Tortora, C., Moraci, M.,Barbarisi, M., and Barbarisi, A. Beta-tricalcium phosphate3D scaffold promote alone osteogenic differentiation of hu-man adipose stem cells: in vitro study. J Mater Sci Mater Med21, 353, 2010.

32. Matsushima, A., Kotobuki, N., Tadokoro, M., Kawate, K.,Yajima, H., Takakura, Y., and Ohgushi, H. In vivo osteogeniccapability of human mesenchymal cells cultured on hy-droxyapatite and on beta-tricalcium phosphate. Artif Organs33, 474, 2009.

33. Neamat, A., Gawish, A., and Gamal-Eldeen, A.M. Beta-tricalcium phosphate promotes cell proliferation, osteogen-esis and bone regeneration in intrabony defects in dogs.Arch Oral Biol 54, 1083, 2009.

34. Takahashi, Y., Yamamoto, M., and Tabata, Y. Osteogenicdifferentiation of mesenchymal stem cells in biodegradablesponges composed of gelatin and beta-tricalcium phosphate.Biomaterials 26, 3587, 2005.

35. Rosen, V. BMP2 signaling in bone development and repair.Cytokine Growth Factor Rev 20, 475, 2009.

36. Javed, A., Bae, J.S., Afzal, F., Gutierrez, S., Pratap, J., Zaidi,S.K., Lou, Y., van Wijnen, A.J., Stein, J.L., Stein, G.S., andLian, J.B. Structural coupling of Smad and Runx2 for exe-cution of the BMP2 osteogenic signal. J Biol Chem 283, 8412,2008.

37. Haidar, Z.S., Hamdy, R.C., and Tabrizian, M. Delivery ofrecombinant bone morphogenetic proteins for bone regen-eration and repair. Part B: delivery systems for BMPs inorthopaedic and craniofacial tissue engineering. BiotechnolLett 31, 1825, 2009.

38. Haidar, Z.S., Hamdy, R.C., and Tabrizian, M. Delivery ofrecombinant bone morphogenetic proteins for bone regen-eration and repair. Part A: current challenges in BMP de-livery. Biotechnol Lett 31, 1817, 2009.

39. Hsu, Y.L., Chang, J.K., Tsai, C.H., Chien, T.T., and Kuo,P.L. Myricetin induces human osteoblast differentiationthrough bone morphogenetic protein-2/p38 mitogen-activated protein kinase pathway. Biochem Pharmacol 73,

504, 2007.40. Ikegame, M., Ishibashi, O., Yoshizawa, T., Shimomura, J.,

Komori, T., Ozawa, H., and Kawashima, H. Tensile stressinduces bone morphogenetic protein 4 in preosteoblasticand fibroblastic cells, which later differentiate into osteo-blasts leading to osteogenesis in the mouse calvariae in or-gan culture. J Bone Miner Res 16, 24, 2001.

41. Jadlowiec, J.A., Zhang, X., Li, J., Campbell, P.G., and Sfeir, C.Extracellular matrix-mediated signaling by dentin phos-phophoryn involves activation of the Smad pathway inde-pendent of bone morphogenetic protein. J Biol Chem 281,

5341, 2006.42. Tang, C.H., Yang, R.S., Chien, M.Y., Chen, C.C., and Fu,

W.M. Enhancement of bone morphogenetic protein-2 ex-pression and bone formation by coumarin derivatives viap38 and ERK-dependent pathway in osteoblasts. Eur JPharmacol 579, 40, 2008.

43. Jiang, F.X., and Harrison, L.C. Convergence of bone mor-phogenetic protein and laminin-1 signaling pathways pro-motes proliferation and colony formation by fetal mousepancreatic cells. Exp Cell Res 308, 114, 2005.

44. Jikko, A., Harris, S.E., Chen, D., Mendrick, D.L., andDamsky, C.H. Collagen integrin receptors regulate earlyosteoblast differentiation induced by BMP-2. J Bone MinerRes 14, 1075, 1999.

45. Tamura, Y., Takeuchi, Y., Suzawa, M., Fukumoto, S., Kato,M., Miyazono, K., and Fujita, T. Focal adhesion kinase ac-tivity is required for bone morphogenetic protein—Smad1signaling and osteoblastic differentiation in murine MC3T3-E1 cells. J Bone Miner Res 16, 1772, 2001.

46. Xiao, G., Gopalakrishnan, R., Jiang, D., Reith, E., Benson,M.D., and Franceschi, R.T. Bone morphogenetic proteins,extracellular matrix, and mitogen-activated protein kinasesignaling pathways are required for osteoblast-specific geneexpression and differentiation in MC3T3-E1 cells. J BoneMiner Res 17, 101, 2002.

47. Fong, Y.C., Li, T.M., Wu, C.M., Hsu, S.F., Kao, S.T., Chen,R.J., Lin, C.C., Liu, S.C., Wu, C.L., and Tang, C.H. BMP-2increases migration of human chondrosarcoma cells viaPI3K/Akt pathway. J Cell Physiol 217, 846, 2008.

48. Shah, A.K., Lazatin, J., Sinha, R.K., Lennox, T., Hickok,N.J., and Tuan, R.S. Mechanism of BMP-2 stimulated ad-hesion of osteoblastic cells to titanium alloy. Biol Cell 91,

131, 1999.

OSTEOCONDUCTIVITY IS REGULATED BY BMP2 AUTOCRINE LOOP 3083

49. Maeda, T., Matsunuma, A., Kurahashi, I., Yanagawa, T.,Yoshida, H., and Horiuchi, N. Induction of osteoblast dif-ferentiation indices by statins in MC3T3-E1 cells. J Cell Bio-chem 92, 458, 2004.

50. Kundu, A.K., Khatiwala, C.B., and Putnam, A.J. Extracellularmatrix remodeling, integrin expression, and downstreamsignaling pathways influence the osteogenic differentiationof mesenchymal stem cells on poly(lactide-co-glycolide)substrates. Tissue Eng Part A 15, 273, 2009.

51. Mauney, J., and Volloch, V. Collagen I matrix contributes todetermination of adult human stem cell lineage via differ-ential, structural conformation-specific elicitation of cellularstress response. Matrix Biol 28, 251, 2009.

Address correspondence to:Hala Zreiqat, Ph.D.

Biomaterials and Tissue Engineering Research UnitSchool of AMME

The University of SydneySydney 2006

Australia

E-mail: hala.zreiqat@sydney.edu.au

Received: April 01, 2010Accepted: May 10, 2010

Online Publication Date: June 14, 2010

3084 LU AND ZREIQAT