1281

INTRODUCTION

COMPLICATIONS ASSOCIATED with the use of autograftsand allografts in current bone repair strategies dem-

onstrate the need for alternative implants that are capa-ble of facilitating functional bone formation. Ideally,these implants would induce a low inflammatory re-

sponse (biocompatible), be safely degraded in the body(biodegradable), support ostoblast-mediated bone depo-sition (osteoinductive), and support the ingrowth of vas-culature to facilitate the newly formed tissue (osteocon-ductive). Numerous biocompatible and biodegradablea-polyhydroxy acids have been examined for their effec-tiveness as replacements for autologous tissue grafts, on

TISSUE ENGINEERINGVolume 11, Number 7/8, 2005© Mary Ann Liebert, Inc.

Poly(�-hydroxybutyrate-co-�-hydroxyvalerate) Supports in Vitro Osteogenesis

A. KUMARASURIYAR,1,2 R.A. JACKSON,1,2 L. GRØNDAHL,3 M. TRAU,4V. NURCOMBE,2 and S.M. COOL2

ABSTRACT

Studies have demonstrated that polymeric biomaterials have the potential to support osteoblast growthand development for bone tissue repair. Poly(�-hydroxybutyrate-co-�-hydroxyvalerate) (PHBV), abioabsorbable, biocompatible polyhydroxy acid polymer, is an excellent candidate that, as yet, hasnot been extensively investigated for this purpose. As such, we examined the attachment charac-teristics, self-renewal capacity, and osteogenic potential of osteoblast-like cells (MC3T3-E1 S14) whencultured on PHBV films compared with tissue culture polystyrene (TCP). Cells were assayed over2 weeks and examined for changes in morphology, attachment, number and proliferation status, al-kaline phosphatase (ALP) activity, calcium accumulation, nodule formation, and the expression ofosteogenic genes. We found that these spindle-shaped MC3T3-E1 S14 cells made cell–cell andcell–substrate contact. Time-dependent cell attachment was shown to be accelerated on PHBV com-pared with collagen and laminin, but delayed compared with TCP and fibronectin. Cell numberand the expression of ALP, osteopontin, and pro-collagen �1(I) mRNA were comparable for cellsgrown on PHBV and TCP, with all these markers increasing over time. This demonstrates the abil-ity of PHBV to support osteoblast cell function. However, a lag was observed for cells on PHBV incomparison with those on TCP for proliferation, ALP activity, and cbfa-1 mRNA expression. In ad-dition, we observed a reduction in total calcium accumulation, nodule formation, and osteocalcinmRNA expression. It is possible that this cellular response is a consequence of the contrasting sur-face properties of PHBV and TCP. The PHBV substrate used was rougher and more hydrophobicthan TCP. Although further substrate analysis is required, we conclude that this polymer is a suit-able candidate for the continued development as a biomaterial for bone tissue engineering.

1School of Biomedical Sciences, University of Queensland, St. Lucia, Australia.2Institute of Molecular and Cell Biology, Singapore.3School of Molecular and Microbial Sciences, University of Queensland, St. Lucia, Australia.4Centre for Nanotechnology and Biomaterials, University of Queensland, St. Lucia, Australia.

the basis of their ability to support osteoblast cell devel-opment. Of these, polylactic acid (PLA),1–4 polyglycolicacid (PGA),5 and copolymers of PGA and PLA(PGA/PLA)6,7 have been shown to support osteoblast at-tachment, growth, and differentiation in vitro. In addi-tion, these polymeric materials have also been shown tosupport bone formation in vivo.8,9

An alternative polyhydroxy acid for bone tissue engi-neering is poly(�-hydroxybutyrate-co-�-hydroxyvaler-ate) (PHBV), a biodegradable polymer derived from Al-caligenes eutrophus.10 PHBV has been shown to be morestable in vivo than the more commonly used �-polyhy-droxy acids, as its degradation can be reduced11–13 by al-tering its molecular weight14,15 and valerate con-tent.11,14–16 Therefore, the accumulation of degradationproducts at the site of implantation is less problematicfor PHBV than for the faster degrading �-polyhydroxyacids. In addition, the principal degradation product ofPHBV, �-hydroxybutyric acid,17–19 is a normal con-stituent of blood that is safely metabolized by theliver.17,20 PHBV has also been shown to be biocompati-ble, as minimal inflammatory responses were observedin long-term studies of subcutaneous PHBV implants inmice12 and rats.21 PHBV can also be engineered to mimicthe structure and mechanical properties of bone by com-bining it with reinforcing phases such as hydroxyap-atite.15,22 Finally, PHBV can be manipulated to facilitateosteoconduction at the implant site, as the thermoplasticproperties of PHBV allow for the appropriate shape,form, and porosity of the implant to be created.15 Takentogether, PHBV may be a suitable candidate to supportlong-term bone regeneration in vivo.

Despite satisfying the criteria for an ideal bone regen-erative implant, in vivo studies examining PHBV for thispurpose are limited. A related polymer, poly(hydroxy-butyrate) (PHB), has been shown to support bone tissueformation in vivo with minimal inflammation and fibroustissue formation.23,24 Similar results have also been re-ported for PHB composites containing hydroxyap-atite.25–27 More recently, in vitro growth of osteoblastcells on macroporous three-dimensional PHBV matriceshas been reported,28 and increases in osteocalcin expres-sion and alkaline phosphatase (ALP) activity over a 60-day growth period have also been shown for cells grownon PHBV.29 Although these findings support the use ofPHBV for in vivo bone regeneration, a comprehensivecharacterization of osteoblast growth and differentiationon nonporous PHBV has yet to be conducted. This is par-ticularly important, as it has been proposed that initialcell attachment and proliferation are dependent on theoutermost functional groups of an implant surface ratherthan its bulk composition.30 Furthermore, an under-standing of the initial cell response to an implant mater-ial, such as time-dependent cell attachment, is critical fordetermining the potential of an implant material to sup-

KUMARASURIYAR ET AL.

port accelerated wound repair. Such studies would alsofacilitate the determination of strategies to improve theimplant surface if necessary.

To examine the interaction between the PHBV surfaceand osteoblast cells, this study assessed the ability ofPHBV to support preosteoblast-like MC3T3-E1 S14 cellattachment, proliferation, and differentiation on the sur-face of solvent-cast PHBV films in vitro. We show thatPHBV supports these aspects of cell behavior and there-fore the osteogenic progression of MC3T3-E1 S14 cells.

MATERIALS AND METHODS

Solvent casting of PHBV

Amounts of 0.277 g of PHBV (8% valerate, Monsanto,St. Louis, MO) and 15 mL of chloroform (Biolab, Mul-grave, Australia) were stirred at 40°C for 15 min. The re-sulting polymer solution was poured into a 70-mm glassPetri dish, covered, and cooled to ambient temperature,and the solvent was allowed to evaporate slowly over a3-day period. Samples of these casts were then subjectedto X-ray photoelectron microscopy (PHI model 560XPS/SAM/SIMS I multitechnique surface analysis sys-tem; PerkinElmer, Wellesley, MA) to confirm completechloroform evaporation from the polymer. Disks 3 cm indiameter (herein referred to as “film”) were punched fromthe larger casts, washed three times in ultrapure water,dried, and sterilized in an autoclave.

Cell culture

MC3T3-E1 S14 cells (American Type Culture Col-lection [ATCC], Manassas, VA) were cultured in 25-cm2

flasks (Nunc, Roskilde, Denmark) in �-minimum essen-tial medium (�-MEM; CSL, Parkville, Australia) sup-plemented with 10% fetal bovine serum (FCS, batch097059401; JRH Biosciences, Brooklyn, Australia), 2mM L-glutamine (CSL), 100 mM sodium pyruvate (TraceBiosciences, Castle Hill, Australia), gentamicin (25�g/mL; Trace Biosciences), and penicillin (100units/mL); Trace Biosciences) (herein referred to as“standard medium”) at 37°C, 5% CO2. Cells were sub-cultured every 3 days, using 0.25% trypsin–1 mM EDTA(Invitrogen Life Technologies, Mount Waverly, Aus-tralia) for 5 min at 37°C, 5% CO2.

Cell culture on tissue culture polystyrene

Before seeding, cells were serum starved in standardmedium replaced with 0.25% FCS for 24 h at 37°C, 5%CO2. Cells were then trypsinized and seeded into 6-wellplates (Nunc) at a density of 5 � 104 cells/cm2. Cultureswere maintained until 1, 7, 14, or 21 days postseeding(herein referred to as the “endpoint”) in �-MEM supple-mented with 10% FCS, 2 mM L-glutamine, 0.05 mM

1282

ascorbic acid 2-phosphate (Sigma, St. Louis, MO), and25 mM �-glycerophosphate (Sigma) (herein referred toas “osteogenic medium”) with medium changes twiceweekly.

Cell culture on PHBV films

To prevent cell attachment onto underlying tissue cul-ture plastic (TCP), six-well plates (Nunc) were precoatedwith 150 �L of 12% poly(2-hydroxyethylmethacrylate)(poly-HEME; Sigma) in 95% ethanol and allowed to dry.PHBV films were placed into poly-HEME-coated six-wellplates (Nunc) and cells were seeded as described for TCP.At assay endpoint, films were transferred to poly-HEME-free conditions when conducting cell lysis procedures. Allother aspects of cell culture were as for cells on TCP.

Cell attachment and morphology

To examine cell attachment to the films and TCP,28.5 � 103 cells/cm2 were seeded and grown for 7 daysin osteogenic medium as described previously. The sub-strates were then washed with phosphate-buffered saline(PBS; GIBCO, Grand Island, NY) and stained with acri-dine orange (0.5 �g/mL; USB, Cleveland, OH) accordingto the manufacturer’s instructions and viewed by confocalmicroscopy. For scanning electron microscopy (SEM),cells were grown as described for fluorescence microscopyand prepared for SEM by fixation in 3% glutaraldehydein cacodylate buffer (ProSciTech, Thuringowa, Australia).Fixed cells were then incubated in 1% OsO4 (ProSciTech)and dehydrated with ethanol. Substrates were then em-bedded in hexamethyldisilazane (HMDS; ProSciTech) andplatinum coated with a sputter coater (Eiko, Ibaraki,Japan). SEM was conducted with the assistance of the Cen-tre for Microscopy and Microanalysis (CMM, Universityof Queensland, St. Lucia, Australia), using a JEOL 6300SEM (JEOL USA, Peabody, MA).

Time-dependent cell attachment

To compare cell attachment to PHBV with cell matrixproteins known to improve osteoblast attachment, cellswere serum starved for 48 h and then seeded at a densityof 28.5 � 103 cells/cm2 in osteogenic medium ontoPHBV films, TCP (positive control), TCP preblockedwith bovine serum albumin (BSA) (negative control), andBD BioCoat plates (BD Biosciences Discovery Labware,Bedford, MA) coated in matrix proteins known to influ-ence cell attachment: fibronectin, collagen, and lamininat �10 �g/mL. To inhibit nonspecific binding, the BDBioCoat plates and negative control TCP were washedthree times with PBS and then incubated with 2% BSAfor 2 h at 37°C and washed again with PBS. Cells werethen seeded and, after 5 min, the medium was collectedand attached cells were trypsinized. Cells from both frac-tions were then counted on a Z series Coulter Counter

PHBV SUPPORT OF IN VITRO OSTEOGENESIS

(Beckman Coulter, Fullerton, CA) and the number of at-tached cells was expressed as a percentage of the totalnumber of cells from both fractions (i.e., the total num-ber of cells seeded). Triplicate readings were taken ofeach sample, and the assay was repeated six times forseparate wells at intervals of 30, 45, 60, 90, and 120 minafter cell seeding.

Cell proliferation: [3H]thymidine incorporationassay

To assess cell proliferation on PHBV and TCP,[3H]thymidine ([3H]Td) incorporation into replicatingDNA was measured. Cells were incubated with 4 �Ci of[3H] Td (Amersham Biosciences/GE Healthcare, LittleChalfont, UK) for 24 h before assay endpoint. They werethen washed in PBS, trypsinized for 1 h, and scraped andthe content of each well was transferred to separate fil-ter papers (Whatman, Middlesex, UK), using a Milliporemanifold system (Millipore, Bedford, MA). The filter pa-pers were thoroughly washed on the manifold systemwith ultrapure water to completely lyse the cells and toremove nonspecific [3H]Td. Filters were then transferredto scintillation vials with 5 mL of Starscint scintillationfluid (PerkinElmer Life and Analytical Sciences, Boston,MA) and counted in a Beta counter (PerkinElmer Lifeand Analytical Sciences).

Total DNA content

DNA content was measured by protocols based on pre-viously described methods.31 At assay endpoint, growthmedium was removed and cells were washed in PBS andstored at �20°C until required. Films were transferred tonew poly-HEME-free plates for freezing. Later, 500 �Lof lysis buffer (25 mM Tris, 0.5% Triton X-100, pH 8)was added to each well and the plates were freeze–thawedfour times. The matrix was disrupted with a cell scraperand the contents were transferred to 5-mL tubes. Three20-�L samples were removed for the ALP activity assaydescribed below. To the tubes, 1 mL of 10 mM EDTA(pH 12.3) was added to the samples and sonicated on ice,using a Sonifer cell disruptor B-30 (Branson Ultrasonics,Danbury, CT) set on a 60% pulse cycle for 10 min. 1 MKH2PO4 (200 �L) was added to neutralize the pH. APicoGreen dsDNA quantification kit (Molecular Probes,Eugene, OR) was used to fluorescently label the DNA asper the manufacturer’s instructions. The samples were vi-sualized at 700 nm with a dual-line laser system (Fluo-rimager 595; Molecular Dynamics, Sunnyvale, CA) andDNA was quantified with reference to a standard curve(in accordance with the manufacturer’s guidelines), us-ing NIH Image software version 1.62 (http://rsb.info.nih.gov/nih-image/). Total cell number was calculatedfrom measurements of DNA content on a basis of 8 pgof DNA per MC3T3-E1 cell.32

1283

ALP activity

Samples (20 �L) obtained during DNA extractionwere combined with 20 �L of p-nitrophenyl phosphate(pNPP) substrate buffer (1% pNPP [Zymed, South SanFrancisco, CA], 10% substrate buffer [Zymed], 89% ul-trapure water) and 20 �L of alkaline buffer solution(Sigma) in a 96-well plate and incubated for 20 min at37°C. The reaction was stopped by the addition of 180�L of 0.05 M NaOH and read at 405 nm (Bio-Rad mi-croplate reader benchmark 10892; Bio-Rad, Hercules,CA). A second reading was taken after the addition of 8�L of concentrated HCl (37%) to bleach the reaction.The second reading was subtracted from the first and con-verted into ALP activity units (1 unit of enzyme hydro-lyzes 1 �mol of pNPP per minute at 37°C) as per thesupplier’s instructions, using a standard curve. Calf in-testinal ALP (Roche, Mannheim, Germany) was used asa positive control.

Matrix/intracellular calcium accumulation

To measure calcium accumulation, a 96-well plate as-say was adapted from the Calcium C test kit (Wako PureChemical, Kyoto, Japan), in accordance with the manu-facturer’s recommendations. At endpoint, cells werewashed with PBS and incubated with 1 mL of 1%trichloroacetic acid for 5 min. A cell scraper was used todissociate the monolayer and the lysate was transferredto a 1.5-mL Eppendorf, vortexed, and centrifuged for 20min at 1600 � g at 25°C. Supernatant (100 �L) was com-bined with 10 �L of color solution and 100 �L of buffersolution in each well and incubated at room temperaturefor 5 min. The absorbance were read at 595 nm (Bio-Radplate reader) and compared with a standard curve pre-pared with standard solutions provided in the kit.

KUMARASURIYAR ET AL.

von Kossa nodule staining

von Kossa staining was used to determine bone nod-ule formation in vitro. Cells were seeded onto films andTCP at a density of 1 � 104 cells/cm2 and cultured for 3weeks in osteogenic medium. The monolayer was thenwashed with PBS, fixed with 4% paraformaldehyde(Sigma), and washed with ultrapure water. To stain nod-ules, the monolayer was treated with 1% silver nitrate(Sigma) for 45 min under ultraviolet (UV) and washedwith ultrapure water. Monolayers were then treated with5% sodium thiosulfate (Sigma) and photographed with adissection microscope (Zeiss, Jena, Germany) equippedwith a digital camera (AxioCam; Zeiss) using AxioVi-sion software version 3.1 (Zeiss).

RNA extraction and cDNA synthesis

RNA was extracted for use in real-time polymerasechain reaction (PCR) analysis. Cells were washed in PBSand RNA was extracted with a NucleoSpin RNA II kit(Machery-Nagel, Düren, Germany) as per the manufac-turer’s instructions. RNA concentrations were measuredby spectrophotometer readings at 260/280 nm (Gene-Quant pro, Amersham Biosciences/GE Healthcare).RNA (1 �g) was reverse transcribed with SuperScript II(Invitrogen, Carlsbad, CA) and oligo(dT) (Invitrogen)according to the manufacturer’s instructions.

Real-time PCR

The expression of selected osteogenic genes was quan-tified by real-time PCR, using an ABI PRISM 7000 se-quence detection system and SYBR Green master mix(Applied Biosystems, Foster City, CA) and specificprimers (Table 1) synthesized by GeneWorks (Hind-

1284

TABLE 1. PRIMERS FOR REAL-TIME PCR

Tm Sizea

Gene Primers (°C) Accession no. (bp)

HPRT F: 5�-CAGTCCCAGCGTCGTGATTAG-3� 56 BC004686.1 97R: 5�-AAACACTTTTTCCAAATCCTCGG-3� 52

Alkaline F: 5�-GATAACGAGATGCCACCAGAGG-3� 57 X13409 140phosphatase R: 5�-TCCACGTCGGTTCTGTTCTTC-3� 54

Osteopontin F: 5�-CCAGGTTTCTGATGAACAGTATCC-3� 56 BC057858 163R: 5�-ACTTGACTCATGGCTGCCCTTT-3� 55

cbfa-1 F: 5�-ACAAACAACCACAGAACCACAAGT-3� 54 NM_009820 111R: 5�-GTCTCGGTGGCTGGTAGTGA-3� 56

Osteocalcin F: 5�-GAGGGCAATAAGGTAGTGAACAGA-3� 56 U11542 134R: 5�-AAGCCATACTGGTTTGATAGCTCG-3� 56

Procollagen F: 5�-GGTCCTCGTGGTGCTGCT-3� 55 BC050014 138�1(I) R: 5�-ACCTTTGCCCCCTTCTTTG-3� 51

Abbreviations: F, forward; R, reverse; Tm, annealing temperature; accession no., GenBank accession number of primer target.aPCR product size.

marsh, Australia). The reaction cycle consisted of a firststage for 10 min at 95°C followed by 45 cycles of com-bined annealing and extension for 15 s at 95°C and for1 min at 60°C. Results are expressed as a relative ex-pression of hypoxanthine-guanine phosphoribosyltrans-ferase (HPRT).

Water contact angle

Water contact angle measurements were performed ona custom-built apparatus consisting of a table-mountedWild Heerbrugg/Leica longitudinal telescopic lens at-tached to a Cohu (San Diego, CA) high-performancecharge-coupled device (CCD) camera. Samples weremounted on an adjustable Teflon mount and a drop of ul-trapure water was delivered by a Hamilton syringe. Thewater drop on the material was observed using real-timevideo pictures captured via an LG-3 Scion frame wrap-per operated by a PC. Contact angles were calculated withNIH Image software. Advancing contact angles were ob-tained by adding one water drop onto the material with

PHBV SUPPORT OF IN VITRO OSTEOGENESIS

the end of the needle above the water drop. A minimumof five repeats was conducted for each sample.

Scanning probe microscopy

An NT-MDT Solver P47 scanning probe microscope(SPM) (NT-MDT, Moscow, Russia) was used in contactmode (constant force) to obtain information on surfaceroughness of the samples. Contact “Golden” silicon tips(NT-MDT) with nominal tip curvature of about 10 nmwere used for the scans. The contact cantilever had a forceconstant of 0.1 N/m. Scans were done over a nominalarea of 7.5 � 7.5 �m. The scan speed was 0.8–1 Hz. Aminimum of five areas on the surface was scanned.

Statistical analyses

All assays were conducted in triplicate for three sepa-rate repeats. The attachment assay was conducted six in-dependent times, using triplicates for each sample Two-way analysis of variance (ANOVA) with Bonferroni post

1285

FIG. 1. Cell morphology of MC3T3-E1 S14 cells on PHBV (A and B) and TCP (C and D) was similar after 7 days of culturein osteogenic medium. Original magnification for confocal microscopy (A and C) was �40 and cells were stained with acridineorange. Original magnification for SEM (B and D) is indicated in the figure.

hoc testing was used to determine the effect of time andsubstrate on cell proliferation, cell attachment, ALP ac-tivity, calcium accumulation, and the expression of os-teogenic genes. Statistical analysis for water contact an-gle and surface roughness was performed with anindependent samples t test for a minimum of five inde-pendent repeats. All statistical analysis was conductedwith SPSS software version 11 (SPSS, Chicago, IL).

RESULTS

Cell morphology and attachment

Both confocal microscopy (Fig. 1A and C) and SEM(Fig. 1B and D) demonstrated the morphology of cellscultured on PHBV and TCP to be the same. Irrespectiveof substrate, cells appeared flattened with clear substrateattachments and cellular processes extending to makecontact with neighboring cells. Time-dependent cell at-tachment (Fig. 2) as determined by Coulter Counter mea-surements indicated that cell attachment to PHBV wascomparable to other matrix proteins. After 30 min, cellattachment to TCP was significantly higher than attach-

KUMARASURIYAR ET AL.

ment to all other substrates, with greater than 60% at-tachment achieved (p � 0.001 over all substrates). By 45min, TCP and fibronectin had attained approximately70% attachment, significantly higher than cell attachmenton the other substrates (p � 0.001). After 60 min, cell at-tachment to PHBV was the same as to TCP and fi-bronectin, all three showing significantly more cell at-tachment than collagen and laminin. Minimal changeswere observed after 60 min, except for fibronectin, whichdisplayed maximal cell attachment (90% attachment) af-ter 90 min over all other substrates. Cell attachment tolaminin was the lowest, being comparable to the BSA-coated negative control wells, both of which showed min-imal cell attachment of 10–15%.

Cell number

DNA content was the same for PHBV and TCP, indi-cating equal cell numbers across all time points (Fig. 3).However, irrespective of substrate, there was a time-de-pendent increase in cell number (p � 0.001), post hoctests showing that this was due to a 38% increase in cellnumber from day 1 to day 7 (p � 0.001). No differencein cell numbers was observed in the last 7 days of growth.

1286

BSA block TCP

5 mins

PHBV CollagenSubstrates

Fibronectin Laminin

Per

cen

tag

e ce

ll at

tach

men

t

0

10

20

30

40

50

60

70

80

90

100

30 mins 45 mins 60 mins 90 mins 120 mins

FIG. 2. Time-dependent cell attachment of MC3T3-E1 S14 cells to BSA-blocked TCP, TCP, PHBV, collagen, fibronectin, andlaminin. Cells were seeded at a density of 28.5 � 103 cells/cm2 and the percentage of cell attachment was examined after 5, 30,45, 60, 90, and 120 min. Cell attachment on PHBV was accelerated compared with attachment to collagen (p � 0.001 at 60 min)and laminin (p � 0.001 at 5 min), but delayed in comparison with attachment to TCP (p � 0.001 at 30 min) and fibronectin (p �0.001 at 45 min).

Proliferation

Irrespective of the substrate, proliferation rates de-creased over time (p � 0.001) (Fig. 4), with the great-est decreases observed between days 7 and 14 for cellson PHBV (p � 0.01) and between days 1 and 7 forcells on TCP (p � 0.001). The highest rate of prolif-eration was observed on day 1 for cells on TCP (6 �105 dpm per 1 � 105 cells) whereas the highest forcells on PHBV was on day 7 (2.5 � 105 dpm per 1 �105 cells).

Alkaline phosphatase activity

Regardless of substrate, alkaline phosphatase (ALP)activity increased over time (p � 0.001) (Fig. 5), withsignificant increases observed between days 1 and 7 forcells grown on TCP (p � 0.01) and between days 7 and14 for cells on PHBV (p � 0.001). ALP activity for cellsgrown on PHBV was 62% lower on day 7 (p � 0.01) but50% higher on day 14 (p � 0.01), indicating a delay inALP activity for cells grown on PHBV compared withTCP.

Mineralization

Mineralization was measured by calcium accumulationand von Kossa staining. As observed with ALP activity,we saw a time-dependent increase in calcium accumula-tion regardless of substrate (p � 0.001) (Fig. 6). A signif-icant increase in calcium accumulation was observed be-tween days 1 and 14 for cells on PHBV (p � 0.01) whereassignificant increases in calcium were detected between

PHBV SUPPORT OF IN VITRO OSTEOGENESIS

days 1 and 7 (p � 0.001) and days 7 and 14 (p � 0.001)for cells grown on TCP. Calcium accumulation was 52%lower for cells grown on PHBV than for those on TCP atday 14 (p � 0.001). von Kossa staining indicated fewer pos-itively stained bone nodules on PHBV compared with TCPafter 3 weeks of culture in osteogenic medium (Fig. 7).

Real-time PCR

All osteogenic markers analyzed by real-time PCRwere expressed by cells when grown on either substrate.Across all time points, osteopontin was the most abun-dantly expressed gene. Expression levels of ALP (Fig.8A), osteopontin (Fig. 8B) and procollagen �1(I) (Fig.8C) were comparable for cells grown on TCP and PHBV,with expression of all three factors increasing signifi-cantly over time (p � 0.001, p � 0.001, and p � 0.05 re-spectively). cbfa-1 mRNA expression levels (Fig. 8D)also increased over time, irrespective of substrate, withincreasing levels between days 1 and 7 (p � 0.001) andbetween days 7 and 14 (p � 0.001). On day 1, cbfa-1 ex-pression was 73% lower for cells grown on PHBV thanfor those on TCP (p � 0.05) whereas by day 14 cbfa-1expression was increased by 68% for cells growing onPHBV compared with TCP (p � 0.01). OsteocalcinmRNA expression (Fig. 8E) also increased over time, ir-respective of substrate; however, its expression was sig-nificantly reduced for cells grown on PHBV than forthose grown on TCP (p � 0.001], especially on day 14(p � 0.001). These findings for cbfa-1 and osteocalcinalso show a delay in osteogenic differentiation on PHBVcompared with TCP.

1287

FIG. 3. DNA content measurements were used as an indirect measure of cell number on PHBV (solid columns) and TCP(shaded columns) on days 1, 7, and 14. There was no substrate-dependent difference in cell numbers, which increased equallyon both substrates between day 1 and day 7. *1p � 0.001.

PHBV and TCP surface hydrophobicity

PHBV solvent cast films were found to be more hy-drophobic than TCP control samples. PHBV samples dis-played an advancing contact angle of �A � 84°, indicat-ing a hydrophobic surface, whereas TCP samples werefound to be less hydrophobic, with an advancing contactangle of �A � 69° (p � 0.001).

KUMARASURIYAR ET AL.

PHBV and TCP surface roughness

An average surface roughness of Ra � 308 nm was de-termined for PHBV whereas TCP yielded an average sur-face roughness of Ra � 6.4 nm, making it the smoothersurface (p � 0.001). The topography of the PHBV sam-ples showed a uniform surface punctuated with randompits and grooves (Fig. 9A and B). In contrast, TCP sam-

1288

FIG. 4. Cell proliferation as measured by [3H]Td incorporation into replicating DNA standardized per 1 � 105 cells. Prolifer-ation on PHBV (solid columns) was delayed compared with TCP (shaded columns), with lower proliferation on day 1 and higherproliferation on day 7. *1p � 0.001; *2p � 0.01.

FIG. 5. Alkaline phosphatase (ALP) activity, standardized per 1 � 105 cells, was measured in a colorimetric assay. Activity waslower on PHBV (solid columns) than on TCP (shaded columns) on day 7; the opposite was true on day 14. *1p � 0.01; *2p � 0.01.

ples had random scratches that can be observed at anglesto the scan direction (Fig. 9C and D).

DISCUSSION

Nonunion fracture repair depends on the ability of animplant to temporarily bear load at the site of injury whilefacilitating bone formation. Although it is possible to en-gineer PHBV to degrade at a rate that allows for thesmooth transfer of the mechanical load to the developingtissue, little is known about the ability of the polymer

PHBV SUPPORT OF IN VITRO OSTEOGENESIS

surface to facilitate the cellular attachment, proliferation,and differentiation necessary for accelerated bone for-mation.

To address this, we evaluated these parameters by cul-turing MC3T3-E1 S14 preosteoblast-like cells on solventcast PHBV films. We found that the attachment of cellsto PHBV was comparable to various matrix proteinsknown to aid osteoblast attachment. Long-term cultureover 2 weeks indicated cells on PHBV were comparablein cell number, morphology, and gene expression pro-files for ALP, osteopontin, and procollagen �1(I) to thosecultured on TCP; an in vitro surface optimized for non-

1289

FIG. 6. Calcium accumulation (per 1 � 105 cells), as determined by colorimetric assay, increased over time for both substrates,with levels lower on PHBV (solid columns) than on TCP (shaded columns) on day 14. *1p � 0.001.

FIG. 7. Silver nitrate staining of cell nodule formation on (A) PHBV and (B) TCP after 3 weeks of culture in osteogenicmedium. Fewer nodules (black staining, indicated by white arrows) are evident on PHBV in comparison with TCP. Original mag-nification: �20.

specific cell growth. Increases in both cell number andexpression of these osteogenic factors over time indicatethat PHBV is able to support osteoblast-like cell attach-ment, growth, and maturation.

However, differences in cell proliferation were observed between cells on PHBV and TCP. The lowerproliferation rate of cells on PHBV on day 1 may be ex-plained by the findings of the time-dependent cell at-tachment assays, which showed that cell attachment waslower on PHBV than on TCP in the first 120 min of cul-ture. However, this cannot completely account for the ob-served trends in proliferation, as DNA content measure-ments indicated that cell numbers were equal on bothsubstrates after 24 h of culture.

KUMARASURIYAR ET AL.

It is possible that cells cultured on PHBV remained ina longer lag phase of growth than cells grown on TCP.The lag phase is a period of cell adjustment to its substrate,and because it is often characterized by low levels of pro-liferation,33 it may have been the overriding factor con-tributing to the lower proliferative rates on PHBV on day1. Furthermore, a longer lag phase would also explain theincreased proliferation of cells on PHBV on day 7, whilecorresponding cells on TCP were in a reduced prolifera-tive state. Presumably, the initially high proliferative ratesof cells on TCP resulted in their early confluence and theassociated contact inhibition of proliferation.

Given the differences in proliferation of cells on eachsubstrate, cell numbers still remained equal on both sub-

1290

FIG. 8. Real-time PCR was used to quantify gene expression of osteogenically related factors. Gene expression is expressedrelative to HPRT mRNA expression (relative expression units). All factors increased with respect to time, with no significant dif-ference between PHBV (solid columns) and TCP (shaded columns) for ALP (A), osteopontin (B), and procollagen �1(I) (C).cbfa-1 (D) and osteocalcin (E) expression on PHBV was delayed compared with TCP. *1p � 0.01; *2p � 0.001; *3p � 0.001;*4p � 0.001; *5p � 0.05; *6p � 0.05; *7p � 0.001; *8p � 0.001.

strates at all time points. It is possible that sustained pro-liferation of cells on PHBV until day 7, contrasting witha decrease in proliferation of cells on TCP during thesame time period, would allow cell numbers on both sub-strates to be equal by day 7. A measurement of prolifer-ation rates at earlier time points is needed to confirm this.

The MC3T3-E1 cell line follows a time-dependent developmental progression along the osteogenic lin-eage,34,35 and thus a longer lag phase would consequentlycause a delay in the onset of cell differentiation whencompared to a control culture. This hypothesis is sup-ported by cbfa-1 mRNA expression. cbfa-1 is a tran-scription factor known to control the expression of os-teogenic factors such as osteocalcin.36,37 Studies haveshown that it is down-regulated during proliferation37 andincreased during the onset of differentiation.36 Given thatproliferation rates of cells on PHBV were lower than

PHBV SUPPORT OF IN VITRO OSTEOGENESIS

those on TCP at day 1, the lower levels of cbfa-1 mRNAin cells on PHBV at the same time point could be ex-plained by a delay in the time-dependent expression ofthis factor. Although statistical analysis detected a sig-nificant difference in cbfa-1 mRNA levels between cellson each substrate on day 14, this may not necessarily bea biological effect, given that statistical analysis indicatedthat mRNA levels did not change over time for eithersubstrate between day 7 and day 14. A lag in differenti-ation is also supported by ALP activity. ALP, an enzymeindicative of osteoblast differentiation,38 peak activitywas delayed for cells on PHBV compared with those onTCP.

The delays observed in the expression of late stagebone markers further corroborate a lag in osteogenesis onPHBV. Osteocalcin, a protein expressed highly by ma-ture cells during the later stages of osteogenesis,39,40

1291

FIG. 9. Scanning probe micrographs of areas of 7.5 � 7.5 �m. (A) Three-dimensional image of PHBV solvent cast film dis-playing an average surface roughness of 308 nm (scale on y axis, 0–2000 nm). (B) Area scan of PHBV solvent cast film. (C)Three-dimensional image of TCP surface with an average surface roughness of 6.4 nm (scale on y axis, 0–80 nm), (D) Area scanof TCP surface. PHBV was rougher than TCP (p � 0.001).

demonstrated significantly reduced mRNA expression incells on PHBV compared with those on TCP on day 14,indicating a less mature phenotype on PHBV. In accor-dance with this finding, calcium accumulation and invitro bone nodule formation after 3 weeks in culture werealso delayed for the cells grown on PHBV. Analysis oflater time points is necessary to confirm lags in the ex-pression of these late stage osteogenic markers.

A lag was not observed for the gene expression of ALP,osteopontin, and procollagen �1(I). However, given thatthese are all midstage markers of bone development andare expressed in both immature and mature cells,38–43 itis likely that mRNA expression levels may vary little overthe course of differentiation. Alternatively, it may be thatthese factors are post-translationally controlled, whichcould explain why ALP gene expression did not corre-late with protein activity. Similar findings have been re-ported by other studies that analyzed both gene and pro-tein expression patterns for the same factor.41 To ourknowledge, this is the first study to use real-time PCR asa tool to analyze cell responses to an implant surface. Thediscrepancy with ALP activity and mRNA expression il-lustrates the need to corroborate gene expression withprotein expression to achieve a conclusive understandingof the cell phenotype.

The PHBV substrate used in this study was also eval-uated by scanning probe micrographs and water contactangle measurements. Substrate surface properties affectthe cellular response and therefore the overall perfor-mance of the implant in vivo. In particular, it has beenshown that the chemical composition, surface morphol-ogy, surface energy, and hydrophobicity are key con-tributors to cell phenotype.30,44 It has been proposed thatsurface roughness affects both cell adhesion and cell–cellinteractions.30 On rough surfaces, cells must realignthemselves on the substrate30,45 and form focal adhesionsthat allow them to span across surface troughs. The re-sulting geometry of these adhesions affects the shape ofthe cell through cytoskeletal rearrangements, which inturn influence gene expression through integrins.30

We showed that solvent cast films of PHBV had amuch rougher surface topography than TCP. Other stud-ies of MC3T3-E1 cell growth on titanium demonstratedthat smooth surfaces enhance cell proliferation, whereasthe expression of adhesion proteins such as fibronectinincreases with increasing surface roughness.46 Increasingthe surface roughness of titanium has also been shown tocause a decrease in MG-63 human osteosarcoma cellnumber and an increase in cellular osteocalcin, TGF-�1

and PGE2 production.47 Other studies using MG-63 cellshave reported that titanium surface roughness preventscell proliferation and ALP expression while increasingcellular RNA, protein, and matrix synthesis.48,49 Thesefindings imply that rough titanium surfaces tend to selectcell differentiation over proliferation. However, results of

KUMARASURIYAR ET AL.

other studies suggest that cell response to surface rough-ness may be substrate specific. For example, studies ex-amining cell response to TCP surface morphology foundthat rat bone marrow cells were sensitive to grooves onTCP, whereas MC3T3-E1 cells were not.50 Similarly, astudy examining rat bone marrow cell growth on variouspolyhydroxy acids found that cell morphology, ALP, andcell proliferation rates were not different between cellson PLA and PHB surfaces, despite large differences insubstrate surface morphology.51 As cell response to sur-face roughness appears to differ between substrate andcell types, it is likely that it is not the sole factor affect-ing MC3T3-E1 S14 cell growth on PHBV.

As mentioned, another important surface property of abiomaterial is hydrophobicity, measured by determiningwater contact angle, where a high contact angle indicatesa hydrophobic surface. Hydrophobicity is related to thesurface energy of a substrate, which in turn influencesthe type and orientation of proteins adsorbed onto the sur-face of a biomaterial. Protein adsorption directly affectscell attachment,45,52 and hence substrates with higher sur-face energy (more hydrophilic) are better at adsorbingcell adhesion proteins30 and inducing cell attachment.Supporting previous findings, we found PHBV to bemore hydrophobic than TCP.53 As a result, the adsorp-tion of adhesion proteins to PHBV from the culturemedium may have been inferior to their adsorption onTCP. This in turn may explain the observed differencesbetween the substrates in cell attachment. This notion isconfirmed by previous studies, where the adsorption ofcollagen and fibronectin increased on PHBV after it washydrolyzed (i.e., hydrophilicity increased).44 It was alsodemonstrated that PHBV substrates, treated with radiofrequency-oxygen plasma to increase the hydrophilicityof the surface, were more effective in supporting rat stro-mal osteoblast cell growth, ALP activity, and osteocal-cin expression than nontreated surfaces.29

In summary, on the basis of other studies detailedabove, it is possible that the hydrophobic PHBV surfacedid not adsorb proteins appropriate for cell attachment aswell as TCP. In addition, the rougher surface morphol-ogy may have challenged cell spread and proliferation.The cells may have responded to these conditions by se-creting proteins and matrix molecules to aid their at-tachment and growth. Previous studies showed an in-crease in RNA expression, protein synthesis (notablyfibronectin), and proteoglycan synthesis for cells onrough titanium surfaces,46,48,49 which may reflect thisphenomenon. Studies have also suggested that this read-justment of the microenvironment precedes the cells’commitment to proliferation and developmental progres-sion.45 As these processes take time, a lag may result, af-fecting the continued time-dependent development of thecells. Measurements of cell phenotypic changes overmore frequent time points than presented here would con-

1292

firm this lag in cell development. In addition, it wouldbe valuable to compare cell behavior between PHBV andTCP, particularly actin polymerization and matrix pro-tein expression, at early time points when many of thefactors involved in microenvironment readjustment maybe expressed.

Although this hypothesis may explain the results ob-tained in this study, it is important to consider that theMC3T3-E1 cell response to PHBV may not be mediatedby surface roughness and hydrophobicity alone. The dis-crepancy in the literature surrounding the effects of sur-face properties on cell phenotype is most likely a reflec-tion of the complex nature of a biomaterial surface andthe synergistic influence that various surface parametershave on cell behavior. Surface parameters not measuredin this study could also be influential and further exper-imentation is needed to elucidate this. Use of techniquesto carefully alter a single surface property alone will al-low better understanding of the influence of that partic-ular property on cell phenotype.

Despite the lag observed for some phenotypic param-eters, cell development was otherwise normal on PHBV,especially given that it was compared with a substrateoptimized for in vitro cell growth. Considering the engi-neering advantages and bioabsorbable properties ofPHBV, we conclude that it is an appropriate substrate toaugment bone regeneration. This study supports previousfindings, and has further illustrated the necessity of acomprehensive evaluation of the cell response to a bio-material surface. It has led to an improved understand-ing of the complex interactions that occur at this interface, crucial to the success of the implant in os-seointegration and wound healing. More importantly,these in vitro data will allow improvements to be madeto the implant system to ensure greater success in vivo.One such improvement would be to change the surfaceproperties of PHBV or couple factors to it such as ma-trix proteins that would increase the rate of cell attach-ment and decrease any lag in cell growth.

ACKNOWLEDGMENTS

The authors are grateful to the Wesley Research Institute, Brisbane, Australia (grant 2000100294), and the Australian Research Council (DP0209873 andDP0343547) for support of this project. The authors alsothank the following people for assistance with data col-lection and analysis: Ms. Shannon Beasley, Dr. GwenLawrie (Nanotechnology and Biomaterials Centre, Uni-versity of Queensland [UQ]), Mr. Bob Simpson (Schoolof Molecular and Microbial Sciences, UQ), Ms. MelindaKastanya, Dr. Thor Bostrom (Analytical EM Facility, andSchool of Physical and Chemical Sciences, QueenslandUniversity of Technology), Dr. Barbara Rolfe (School of

PHBV SUPPORT OF IN VITRO OSTEOGENESIS

Biomedical Sciences, UQ), Dr. Barry Wood (BrisbaneSurface Analysis Centre, UQ), Mr. Michael Stapleberg,Dr. Rick Webb (Centre for Microscopy and Microanaly-sis [CMM], UQ), Assoc. Prof. Lesley Lluka (School ofBiomedical Sciences, UQ), and Ms. M. Bates (Nan-otechnology and Biomaterials Centre, UQ).

REFERENCES

1. Liao, S.S., Cui, F.Z., and Zhang, W. [Mineralized collagenbased composite for bone tissue engineering]. ZhongguoYi Xue Ke Xue Yuan Xue Bao 25, 36, 2003.

2. Hu, Y., Winn, S.R., Krajbich, I., and Hollinger, J.O. Porouspolymer scaffolds surface-modified with arginine-glycine-aspartic acid enhance bone cell attachment and differenti-ation in vitro. J. Biomed. Mater. Res. 64A, 583, 2003.

3. Matsuzaka, K., Walboomers, F., de Ruijter, A., and Jansen,J.A. Effect of microgrooved poly-L-lactic (PLA) surfaceson proliferation, cytoskeletal organization, and mineralizedmatrix formation of rat bone marrow cells. Clin. Oral Im-plants Res. 11, 325, 2000.

4. Malekzadeh, R., Hollinger, J.O., Buck, D., Adams, D.F.,and McAllister, B.S. Isolation of human osteoblast-likecells and in vitro amplification for tissue engineering. J.Periodontol. 69, 1256, 1998.

5. Hatton, P.V., Walsh, J., and Brook, I.M. The response ofcultured bone cells to resorbable polyglycolic acid and sil-icone membranes for use in orbital floor fracture repair.Clin. Mater. 17, 71, 1994.

6. Ishaug-Riley, S.L., Crane-Kruger, G.M., Yaszemski, M.J.,and Mikos, A.G. Three-dimensional culture of rat calvar-ial osteoblasts in porous biodegradable polymers. Bioma-terials 19, 1405, 1998.

7. Di Toro, R., Betti, V., and Spampinato, S. Biocompatibil-ity and integrin-mediated adhesion of human osteoblasts topoly(DL-lactide-co-glycolide) copolymers. Eur. J. Pharm.Sci. 21, 161, 2004.

8. Liao, S.S., Guan, K., Cui, F.Z., Shi, S.S., and Sun, T.S.Lumbar spinal fusion with a mineralized collagen matrixand rhBMP-2 in a rabbit model. Spine 28, 1954, 2003.

9. Howard, D., Partridge, K., Yang, X., Clarke, N.M., Okubo,Y., Bessho, K., Howdle, S.M., Shakesheff, K.M., and Or-effo, R.O. Immunoselection and adenoviral genetic modu-lation of human osteoprogenitors: In vivo bone formationon PLA scaffold. Biochem. Biophys. Res. Commun. 299,208, 2002.

10. Luzier, W.D. Materials derived from biomass biodegrad-able materials. Proc. Natl. Acad. Sci. U.S.A. 89, 839, 1992.

11. Taylor, M.S., Daniels, A.U., Andriano, K.P., and Heller, J.Bioabsorbable polymers: In vitro acute toxicity of accu-mulated degradation products. J. Appl. BioMater. 5, 151,1994.

12. Gogolewski, S., Jovanovic, M., Perren, S.M., Dillon, J.G.,and Hughes, M.K. Tissue response and in vivo degradationof selected polyhydroxyacids: Polylactides (PLA), poly(3-hydroxybutyrate) (PHB), and poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHB/VA). J. Biomed. Mater. Res. 27,1135, 1993.

1293

13. Gilding, D.K. Biodegradable polymers. In: Biocompatibil-ity of Clinical Implant Materials. D.F. Williams, ed. BocaRaton, FL: CRC Press, 1981, p. 209.

14. Gogolewski, S. Resorbable polymers for internal fixation.Clin. Mater. 10, 13, 1992.

15. Hammond, T., and Liggat, J.J. Properties and applicationsof bacterially derived polyhydroxyalkanoates. In: Degrad-able Polymers. G. Scott and D. Gilead, eds. London: Chap-man & Hall: 1995, p. 87.

16. Lin, T.C. Effects of crystallinity on polylactide degrada-tion. In: The 11th Annual Meeting of the Society for Bio-materials, 1985, San Diego, CA.

17. An, Y.H., Woolf, S.K., and Friedman, R.J. Pre-clinical invivo evaluation of orthopaedic bioabsorbable devices. Bio-materials 21, 2635, 2000.

18. Hankermeyer, C.R., and Tjeerdema, R.S. Polyhydroxybu-tyrate: Plastic made and degraded by microorganisms. Rev.Environ. Contam. Toxicol. 159, 1, 1999.

19. Holland, S.J., Jolly, A.M., Yasin, M., and Tighe, B.J. Poly-mers for biodegradable medical devices. 2. Hydroxybu-tyrate–hydroxyvalerate copolymers: Hydrolytic degrada-tion studies. Biomaterials 8, 289, 1987.

20. Yasin, M., Holland, S.J., Jolly, A.M., and Tighe, B.J. Poly-mers for biodegradable medical devices. 6. Hydroxybu-tyrate–hydroxyvalerate copolymers: Accelerated degrada-tion of blends with polysaccharides. Biomaterials 10, 400,1989.

21. Galgut, P., Pitrola, R., Waite, I., Doyle, C., and Smith, R.Histological evaluation of biodegradable and nondegrad-able membranes placed transcutaneously in rats. J. Clin.Periodontol. 18, 581, 1991.

22. Lutton, C., Read, J., and Trau, M. Nanostructured bioma-terials: A novel appraoch to artificial bone implants. Aust.J. Chem. 54, 621, 2001.

23. Kostopoulos, L., and Karring, T. Guided bone regenerationin mandibular defects in rats using a bioresorbable poly-mer. Clin. Oral Implant. Res. 5, 66, 1994.

24. Kostopoulos, L., and Karring, T. Augmentation of the ratmandible using guided tissue regeneration. Clin. Oral Im-plant. Res. 5, 78, 1994.

25. Wang, M., Chen, L.J., Ni, J., Weng, J., and Yue, C.Y. Man-ufacture and evaluation of bioactive and biodegradable ma-terials and scaffolds for tissue engineering. J. Mater. Sci.Mater. Med. 12, 855, 2001.

26. Doyle, C., Tanner, E.T., and Bonfield, W. In vitro and invivo evaluation of polyhydroxybutyrate and of polyhy-droxybutyrate reinforced with hydroxyapatite. Biomateri-als 12, 841, 1991.

27. Knowles, J.C., Hastings, G.W., Ohta, H., Niwa, S., andBoeree, N. Development of a degradable composite for or-thopaedic use: In vivo biomechanical and histological eval-uation of two bioactive degradable composites based on thepolyhydroxybutyrate polymer. Biomaterials 13, 491, 1992.

28. Kose, G.T., Kenar, H., Hasirci, N., and Hasirci, V. Macro-porous poly(3-hydroxybutyrate-co-3-hydroxyvalerate) ma-trices for bone tissue engineering. Biomaterials 24, 1949,2003.

29. Kose, G.T., Korkusuz, F., Korkusuz, P., Purali, N., Ozkul,A., and Hasirci, V. Bone generation on PHBV matrices:An in vitro study. Biomaterials 24, 4999, 2003.

KUMARASURIYAR ET AL.

30. Boyan, B.D., Hummert, T.W., Kieswetter, K., Schraub, D.,Dean, D.D., and Schwartz, Z. Effect of titanium surfacecharacteristics on chondrocytes and osteoblasts in vitro.Cell Mater. 5, 323, 1995.

31. Payne, R.G., Yaszemski, M.J., Yasko, A.W., and Mikos,A.G. Development of an injectable, in situ crosslinkable,degradable polymeric carrier for osteogenic cell popula-tions. 1. Encapsulation of marrow stromal osteoblasts insurface crosslinked gelatin microparticles. Biomaterials 23,4359, 2002.

32. Shea, L.D., Wang, D., Franceschi, R.T., and Mooney, D.J.Engineered bone development from a pre-osteoblast cellline on three-dimensional scaffolds. Tissue Eng. 6, 605,2000.

33. Freshney, R.I. Culture of Animal Cells: A Manual of Ba-sic Technique, 4th ed. Brisbane: Wiley-Liss, 2000.

34. Quarles, L.D., Yohay, D.A., Lever, L.W., Caton, R., andWenstrup, R.J. Distinct proliferative and differentiatedstages of murine MC3T3-E1 cells in culture: An in vitromodel of osteoblast development. J. Bone Miner. Res. 7,683, 1992.

35. Sudo, H., Kodama, H.A., Amagai, Y., Yamamoto, S., andKasai, S. In vitro differentiation and calcification in a newclonal osteogenic cell line derived from newborn mousecalvaria. J. Cell Biol. 96, 191, 1983.

36. Xiao, G., Jiang, D., Gopalakrishnan, R., and Franceschi,R.T. Fibroblast growth factor 2 induction of the osteocal-cin gene requiRes. MAPK activity and phosphorylation ofthe osteoblast transcription factor, Cbfa1/Runx2. J. Biol.Chem. 277, 36181, 2002.

37. Pratap, J., Galindo, M., Zaidi, S.K., Vradii, D., Bhat, B.M.,Robinson, J.A., Choi, J.Y., Komori, T., Stein, J.L., Lian,J.B., Stein, G.S., and van Wijnen, A.J. Cell growth regu-latory role of Runx2 during proliferative expansion of pre-osteoblasts. Cancer Res. 63, 5357, 2003.

38. Wang, D., Christensen, K., Chawla, K., Xiao, G., Krebs-bach, P.H., and Franceschi, R.T. Isolation and characteri-zation of MC3T3-E1 preosteoblast subclones with distinctin vitro and in vivo differentiation/mineralization potential.J. Bone Miner. Res. 14, 893, 1999.

39. Stein, G.S., Lian, J.B., and Owen, T.A. Relationship of cellgrowth to the regulation of tissue-specific gene expressionduring osteoblast differentiation. FASEB J. 4, 3111, 1990.

40. Ibaraki, K., Termine, J.D., Whitson, S.W., and Young, M.F.Bone matrix mRNA expression in differentiating fetalbovine osteoblasts. J. Bone Miner. Res. 7, 743, 1992.

41. Frank, O., Heim, M., Jakob, M., Barbero, A., Schafer, D.,Bendik, I., Dick, W., Heberer, M., and Martin, I. Real-timequantitative RT-PCR analysis of human bone marrow stro-mal cells during osteogenic differentiation in vitro. J. Cell.Biochem. 85, 737, 2002.

42. Shibata, Y., Abiko, Y., Moriya, Y., Yoshida, W., andTakiguchi, H. Effects of transforming growth factor-� oncollagen gene expression and collagen synthesis level inmineralizing cultuRes. of osteoblast-like cell line, MC3T3-E1. Int. J. Biochem. 25, 239, 1993.

43. Diefenderfer, D.L., Osyczka, A.M., Garino, J.P., andLeboy, P.S. Regulation of BMP-induced transcription incultured human bone marrow stromal cells. J. Bone JointSurg. Am. 85A (Suppl. 3), 19, 2003.

1294

44. Rouxhet, L., Legras, R., and Schneider, Y.J. Interactionsbetween a biodegradable polymer, poly(hydroxybutyrate-hydroxyvalerate), proteins and macrophages. Macromol.Symp. 130, 347, 1998.

45. Tamada, Y., and Ikada, Y. Fibroblast growth on polymersurfaces and biosynthesis of collagen. J. Biomed. Mater.Res. 28, 783, 1994.

46. Linez-Bataillon, P., Monchau, F., Bigerelle, M., and Hilde-brand, H.F. In vitro MC3T3 osteoblast adhesion with re-spect to surface roughness of Ti6Al4V substrates. Biomol.Eng. 19, 133, 2002.

47. Lohmann, C.H., Bonewald, L.F., Sisk, M.A., Sylvia, V.L.,Cochran, D.L., Dean, D.D., Boyan, B.D., and Schwartz, Z.Maturation state determines the response of osteogeniccells to surface roughness and 1,25-dihydroxyvitamin D3.J. Bone Miner. Res. 15, 1169, 2000.

48. Lincks, J., Boyan, B.D., Blanchard, C.R., Lohmann, C.H.,Liu, Y., Cochran, D.L., Dean, D.D., and Schwartz, Z. Re-sponse of MG63 osteoblast-like cells to titanium and tita-nium alloy is dependent on surface roughness and compo-sition. Biomaterials 19, 2219, 1998.

49. Martin, J.Y., Schwartz, Z., Hummert, T.W., Schraub, D.M.,Simpson, J., Lankford, J., Dean, D.D., Cochran, D.L., andBoyan, B.D. Effect of titanium surface roughness on pro-liferation, differentiation, and protein synthesis of human

PHBV SUPPORT OF IN VITRO OSTEOGENESIS

osteoblast-like cells (Mg63). J. Biomed. Mater. Res. 29,389, 1995.

50. Bruinink, A., and Wintermantel, E. Grooves affect primarybone marrow but not osteoblastic MC3T3-E1 cell cultures.Biomaterials 22, 2465, 2001.

51. Wang, Y.W., Wu, Q., and Chen, G.Q. Attachment, prolif-eration and differentiation of osteoblasts on randombiopolyester poly(3-hydroxybutyrate-co-3-hydroxyhexa-noate) scaffolds. Biomaterials 25, 669, 2004.

52. Hayman, E.G., Pierschbacher, M.D., Ohgren, Y., and Ruo-slahti, E. Serum spreading factor (vitronectin) is present atthe cell surface and in tissues. Proc. Natl. Acad. Sci. U.S.A.80, 4003, 1983.

53. Mas, A. Modification de durface de poly(hydroxybutyrate-co-9%-hydroxyvalerate) par polymerisation par plasmad’alcool allylique. Eur. Polym. J. 33, 331, 1997.

Address reprint requests to:Simon Cool, Ph.D.

Stem Cell and Tissue Repair LaboratoryInstitute of Molecular and Cell Biology

61 Biopolis Drive (Proteos)Singapore 138673

E-mail: scool@imcb.a-star.edu.sg

1295