Polymer Edition g Journal of Biomaterials Science,

This article was downloaded by: [UQ Library]On: 24 May 2015, At: 18:13Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Biomaterials Science,Polymer EditionPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/tbsp20

Investigations into poly(3-hydroxybutyrate-co-3-hydroxyvalerate) surface propertiescausing delayed osteoblast growthImelda Keen a , Liza J. Raggatt b , Simon M. Cool c , VictorNurcombe d , Peter Fredericks e , Matt Trau f & LisbethGrøndahl ga Nanotechnology and Biomaterials Centre, The University ofQueensland, Brisbane, Queensland, 4072 Australiab Institute for Molecular Bioscience, The University ofQueensland, Brisbane, Queensland, 4072 Australiac Department of Orthopaedic Surgery, National University ofSingapore, Singapore 117597; Institute of Molecular and CellBiology, 61 Biopolis Drive, Singapore 138673d Institute of Molecular and Cell Biology, 61 Biopolis Drive,Singapore 138673e School of Physical and Chemical Sciences, QueenslandUniversity of Technology, Brisbane, Queensland, Australiaf Nanotechnology and Biomaterials Centre, The University ofQueensland, Brisbane, Queensland, 4072 Australiag School of Molecular and Microbial Sciences, The University ofQueensland, Brisbane, Queensland, 4072 AustraliaPublished online: 02 Apr 2012.

To cite this article: Imelda Keen , Liza J. Raggatt , Simon M. Cool , Victor Nurcombe , PeterFredericks , Matt Trau & Lisbeth Grøndahl (2007) Investigations into poly(3-hydroxybutyrate-co-3-hydroxyvalerate) surface properties causing delayed osteoblast growth, Journal of BiomaterialsScience, Polymer Edition, 18:9, 1101-1123, DOI: 10.1163/156856207781554046

To link to this article: http://dx.doi.org/10.1163/156856207781554046

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the“Content”) contained in the publications on our platform. However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoever as tothe accuracy, completeness, or suitability for any purpose of the Content. Any opinionsand views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Contentshould not be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions, claims,proceedings, demands, costs, expenses, damages, and other liabilities whatsoeveror howsoever caused arising directly or indirectly in connection with, in relation to orarising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden. Terms& Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

J. Biomater. Sci. Polymer Edn, Vol. 18, No. 9, pp. 1101–1123 (2007) VSP 2007.Also available online - www.brill.nl/jbs

Investigations into poly(3-hydroxybutyrate-co-3-hydroxyvalerate) surface properties causing delayedosteoblast growth

IMELDA KEEN 1, LIZA J. RAGGATT 2, SIMON M. COOL 3,4,VICTOR NURCOMBE 4, PETER FREDERICKS 5, MATT TRAU 1 andLISBETH GRØNDAHL 6,∗1 Nanotechnology and Biomaterials Centre, The University of Queensland,

Brisbane, Queensland, 4072 Australia2 Institute for Molecular Bioscience, The University of Queensland,

Brisbane, Queensland, 4072 Australia3 Department of Orthopaedic Surgery, National University of Singapore, Singapore 1175974 Institute of Molecular and Cell Biology, 61 Biopolis Drive, Singapore 1386735 School of Physical and Chemical Sciences, Queensland University of Technology,

Brisbane, Queensland, Australia6 School of Molecular and Microbial Sciences, The University of Queensland,

Brisbane, Queensland, 4072 Australia

Received 5 September 2006; accepted 3 April 2007

Abstract—Osteoblast proliferation is sensitive to material surface properties. In this study, theproliferation of MC3T3 E1-S14 osteoblastic cells on poly(3-hydroxybutyrate-co-3-hydroxyvalerate)(PHBV) films with different surface characteristics was investigated with the aim of evaluating thecause of a lag in cell growth previously observed. The solvent-cast films were prepared using threedifferent solvents/solvent mixtures which produced PHBV films with both a rough (at the air interface)and smooth (at the glass interface) surface. Investigation of the surface roughness by scanning electronand scanning probe microscopy revealed that the surfaces had features that were different in bothaverage lateral size and average amplitude (Ra 20–200 nm). Water contact angles showed that allsurfaces were hydrophobic in nature (θA in the range 69–82◦). The lateral distribution of surfacecrystallinity of the films was evaluated by use of micro-attenuated total reflectance Fourier transforminfrared (ATR–FT-IR) by determining the surface crystallinity index (CI) which was found to differbetween samples. MC3T3-E1-S14 osteoblasts were cultured on the six surfaces and proliferation wasdetermined. After 2 days, cell proliferation on all surfaces was significantly less than on the controlsubstrate; however, after 4 days cell proliferation was optimal on three surfaces. It was concludedthat the initial lag on all substrates was due to the hydrophobic nature of the substrates. The ability ofthe cells to recover on the materials was attributed to the degree of heterogeneity of the crystallinity

∗To whom correspondence should be addressed. Tel.: (61-7) 3365-3671; Fax: (61-7) 3365-4299;e-mail: [email protected]

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

1102 I. Keen et al.

and surface roughness: samples with a roughness of �80 nm were found to support cell proliferation.In addition, the lateral surface features influenced the proliferation of osteoblasts on the PHBV filmsurface.

Key words: Roughness; surface crystallinity; wettability; osteoblast proliferation; PHBV.

INTRODUCTION

Artificial bone scaffolds that have properties akin to native bone and that canbe made on demand and shaped as required during surgery would be a valuableresource for orthopaedic surgery [1]. The current options available to replace boneinclude allograph and autograph bone; however, there is a limit to the amount ofautograph bone that can be harvested, and allograph bone must be irradiated priorto use, a process that severely compromises the quality of the bone. Thus, thefabrication of artificial bone scaffolds that can be used to replace and repair injuredor diseased bone are a clinical necessity that can be achieved by using biodegradablepolymers [1].

The biodegradable polymer poly(3-hydroxybutyrate-co-3-hydroxyvalerate)(PHBV), which is derived form various micro-organisms, including the energy stor-age granules of Gram-negative bacteria, has a number of intrinsic features suitablefor making an artificial bone biomaterial. Specifically, PHBV has a slow degrada-tion rate which allows sufficient time for the bone to repair itself and forms degra-dation products that are non-toxic and are metabolised via beta-oxidation and thetricarboxylic acid cycle (TCA cycle) [2]. In addition, PHBV has mechanical prop-erties which are superior to those of cancellous bone (Young’s modulus of 1.0 GPaand tensile strength of 13 MPa) [3] and which can be improved further by the gener-ation of composite biomaterials made from PHBV and hydroxyapatite [4, 5]. Recentstudies have shown that PHBV supports the growth and proliferation of osteoblasts[6, 7]. However, these studies reported that the growth and differentiation of os-teoblasts on PHBV was affected by a lag period [7] and the surface roughness andthe wettability of PHBV were the properties proposed to hinder the initial growof osteoblasts on this material [7]. A number of excellent reviews have describedhow the proliferation and differentiation of osteoblasts on biomaterial substrates areaffected by the unique surface properties of the particular material used [8–11].Specifically, surface wettability [12], roughness [13–15] and crystallinity [16] havebeen suggested to influence the behaviour of osteoblasts grown on polymer-basedbiomaterials.

Low surface wettability (i.e., high hydrophobicity) is a common surface char-acteristic of many polyesters, including PHBV. Generally for thermoplastics, hy-drophobic surfaces (polystyrene bacteriological culture plastic, water contact an-gle 75◦) are known to inhibit proliferation and increase the rate of apoptosis ofanchorage-dependent osteoblastic cells compared to cells grown on hydrophilic sur-faces (tissue-culture-grade polystyrene produced by chemical modification, water

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

Investigations into PHBV surface properties causing delayed osteoblast growth 1103

contact angle 56◦) [12]. Hydrophobicity reflects the surface energy of a substrateand it is generally accepted that this influences the adsorption of proteins onto ma-terial surfaces [17, 18]. This in turn has been suggested to influence the behaviourof cells grown on the substrate. It is important, however, to appreciate that wet-tability is linked directly to other surface properties such as surface chemistry androughness and as such it cannot be studied in isolation.

Informative studies on patterned surfaces have given insight into cell responses tosurface roughness. Specifically, the influence that the average amplitude of featureson the surface of a material (Ra) has on osteoblast proliferation has been investigatedfor several polymer substrates [13–15]. In these studies the surface features and thebulk polymer had the same chemical composition; however, wettability is likelyto have been affected by the roughness. Wan et al. reported that the addition ofsurface features with both micro- (2.5 µm) and nano- (40 nm) scale amplitude topoly(L-lactide) (PLLA) and polystyrene (PS) polymers improved the attachment ofrat osteoblast-like cells to both materials at a 6 h time point, whereas no differencein cell proliferation was seen at 2 days [14]. Liao et al. found that rat calvaria bonecells had higher alkaline phosphatase activity at day 9 on pyramid patterned siliconesurfaces (square base 33 µm, height 23 µm) compared to flat featureless surfacesand proposed that the grooves between the surface features created a specificbiochemical micro-environment around each cell which promoted proliferation[15]. Hatano et al. have also demonstrated that the amplitude of surface featuresinfluences osteoblastic proliferation of rat calvarial osteoblastic cells, by changingthe roughness (0.37–2.9 µm) of PS surfaces with different coarseness of grindingpaper and that the cell performance was optimal on surfaces with features of 0.81 µm[13]. In addition to the alignment of the surface pattern influencing cell behaviour,the sizes of the topographical features (grooves or pits) are also influential and canlimit the interaction of rat calvaria bone cells with the surface as cell filopodia havebeen observed to anchor on top of a surface feature rather than on the surface inbetween them [19].

The effect of material crystallinity on cell growth has been studied for crystalsurfaces and metal oxides and it has been found that cells can indeed discriminatebetween surfaces that differ only in crystallinity [20] and variations in surfaceorganization at the Ångstrom scale [21, 22]. For polymeric materials a recent studyutilised annealing to modify the degree of polymer crystallinity however in turnsurface roughness was also modified [16]. The rate of proliferation of MC3T3-E1osteoblastic cells was found to be greater on the smooth regions of the films thanon the rougher regions, an observation that contradicts the general dogma thatosteoblasts prefer rough surfaces. However, as the changes in surface crystallinitywere not assessed in this study it is not possible to definitively identify which surfaceparameter, roughness or crystallinity, was the most influential.

Identification of the specific surface characteristics responsible for influencing cellbehaviour is challenging because the modification of one surface property oftenresults in simultaneous changes to other surface properties [11]. Moreover, it has

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

1104 I. Keen et al.

been shown in a study by Chesmel et al. [23] that surface morphology and chemistryhad a synergistic effect on bone cells grown on micro-machined grooved PS culturedishes with 5-µm-wide grooves which were either 0.5 or 5 µm deep and that thiseffect could not be predicted from the responses to the individual features alone.This emphasises the complexity associated with studying cell response to surfacefeatures.

While the generation of patterned polymer surfaces can provide important infor-mation regarding cellular responses to surface topography and chemistry on bioma-terial films, these fabrication techniques are not always relevant to biomaterials andcannot be applied to the generation of surface features within 3D scaffolds. Solventcasting produces surface features that mimic those formed in 3D scaffolds fabricatedvia solvent-based techniques. Polymers with differing degrees of crystallinity, wet-tability and roughness can be produced by using different solvents during solventcasting [24–26] and the surface characteristics of these solvent-cast films can alsobe influenced by the type of casting substrates [27, 28].

The aim of this study is to identify surface feature(s) of solvent-cast PHBV filmsthat affect osteoblast proliferation, in particular, to assess the features responsiblefor the lag in osteoblast growth that occurs on these substrates. Solvent cast filmswere produced in this study from a variety of solvents. Surface characteristics of thedifferent films were determined using scanning probe microscopy (SPM), scanningelectron microscopy (SEM), attenuated total reflectance Fourier transform infrared(ATR–FT-IR) spectroscopy and contact-angle measurements. The differences insurface properties were then correlated to the proliferation and morphology ofosteoblastic cells cultured on the different films.

MATERIAL AND METHODS

Preparation of solvent-cast films

Poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV) with 8.8 mol% 3-hydroxy-valerate was purchased from Sigma-Aldrich (St. Louis, MO, USA). Films wereproduced by dissolving at 50◦C 0.30 g PHBV in 15 ml solvent; either chloroform(99.4% purity, Biolab, Clayton, Australia), dichloromethane (DCM) (99.8% purity,Labscan, Bangkok, Thailand), or a chloroform/acetone mixture (50:50, v/v: acetone99.5% purity, Ajax, Seven Hills, Australia). When the chloroform/acetone solventmixture was used, the powder was dissolved first in chloroform before acetone wasadded. A covered glass Petri dish (70 mm i.d.) was used as casting substrate.Solvents were allowed to evaporate at room temperature (25◦C) over several days.Two types of surfaces resulted; a ‘rough’ surface was produced at the air interfaceand a ‘smooth’ surface was produced at the glass interface. Sample surfaceswere labelled as follows: CHCl3-R, rough surface of the chloroform-cast PHBVfilm; CHCl3-S, smooth surface of the chloroform-cast PHBV film; DCM-R, roughsurface of the DCM-cast PHBV film; DCM-S, smooth surface of the DCM-cast

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

Investigations into PHBV surface properties causing delayed osteoblast growth 1105

PHBV film; MIX-R, rough surface of the chloroform/acetone-cast PHBV film;MIX-S, smooth surface of the chloroform/acetone-cast PHBV film.

Characterisation of solvent-cast films

Scanning electron microscopy (SEM). Samples were mounted on aluminiumbases using carbon double-sided tape and then sputter coated with platinum (Eiko,Japan) to prevent the sample from charging during image acquisition. SEM imageswere obtained using a Jeol-6400F scanning electron microscope (Jeol, Tokyo,Japan) at an accelerating voltage of 5–10 kV. Digital images were captured andsaved using an image slave software program (used to digitise SEM images,Dindima Group, Melbourne, Australia).

Scanning probe microscopy (SPM). An NT-MDT Solver P47 SPM (NT-MDT,Moscow, Russia) was used in a semi-contact (“tapping”) mode to obtain informationon surface roughness of the samples. Measurements were performed using non-contact “Golden” Si cantilevers (type NSG11 from NT-MDT) with a nominal tipradius of about 10 nm and a scan speed of approximately 1 Hz. The cantilever hada force constant of 5 N/m. A nominal area of either 7.5 µm × 7.5 µm or 18.0 µm ×18.0 µm was scanned. Four to eight random areas on each surface was scanned. Theerrors in the Ra and Rz values are the standard deviations.

Contact-angle measurements. Water contact angle measurements were acquiredusing the sessile drop method. Both advancing (θA) and receding (θR) contactangles were obtained using drops of Milli-Q water on the film surfaces. The θA

was measured on a 5-µl drop and subsequently after each addition of 5 µl until atotal 20 µl volume was added. For θR, 5-µl volumes were withdrawn from a 25-µlwater drop. Using the equation 2h/� = tan θ/2, contact angles (θ ) were calculated(� is the base diameter of the drop and h is the height of the drop) [29]. A minimumof three repeats were carried out for each sample and the errors given in θA and θR

are the standard derivation of the 12 or more contact-angle values.

Attenuated total reflectance Fourier transform infrared (ATR–FT-IR) spectros-copy. ATR–FT-IR spectra (64 scans, 8 cm−1 resolution, wavenumber range 4000–525 cm−1) were acquired using a Nicolet Nexus 870 (Thermo-Nicolet, Madison,WI, USA) with a Smart Endurance diamond ATR accessory where the penetrationdepth was 0.91 µm at 1550 cm−1 when using a value of 1.5 for the refractive index ofthe polymer. Micro ATR–FT-IR spectra (64 scans, 8 cm−1 resolution, wavenumberrange 4000–700 cm−1) were collected from a Nicolet Continuµm microscopeequipped with a liquid-nitrogen-cooled mercury-cadmium-telluride (MCT) detectorand a silicon ATR objective where the penetration depth was 0.65 µm at 1550 cm−1.Contact between the crystal and sample was automated and monitored by a pressuregauge. An area of 1000 µm × 1000 µm was mapped where each spectrum was

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

1106 I. Keen et al.

obtained from an area defined by an aperture of 40 µm × 40 µm. ATR spectrawere corrected for wavelength dependence. The step size for each spectrum was40 µm. Spectral information was extracted by means of spectral analysis software(GRAMS/32, Galactic Industries, Salem, NH, USA) and maps were illustratedusing the Origin graphics software package (OriginLab, Northampton, MA, USA).The error in the average crystallinity index (CI) value is based on the standarddeviation of 8 repeats of the CHCl3-R film.

X-ray diffraction (XRD). XRD spectra were recorded on a Bruker D8 AdvanceX-ray diffractometer equipped with Cu Kα (λ = 0.1542 nm) source and Göbelmirrors to achieve a parallel X-ray beam. Each scan was recorded in the range of2θ = 6–36◦ at a scan step of 0.01◦/10 s at 40 kV and 30 mA. Traces were processedusing the Diffracplus Evaluation Package Release 2004 and PDF (Powder DiffractionFile)-2 Release 2004 (Bruker AXS, Madison, WI, USA). Percent crystallinity wascalculated from the crystalline and total area of the diffractogram and carriedout using Peak Fit Version 5 (Systat Software, Richmond, CA, USA) and Excelsoftware.

Differential scanning calorimetry (DSC). A Perkin Elmer DSC 7 (Perkin Elmer,Wellesley, MA, USA) was calibrated using the melting temperatures of indium(429.4 K) and zinc (692.5 K) and their heats of fusion. About 5 mg specimenswere heated from 20 to 200◦C at a rate of 10◦C/min. Melting enthalpy (�Hm)was determined from the area under a peak of the DSC trace run using a dataacquisition program (PYRIS Version 3.5 Thermal software, Perkin Elmer). Percentbulk crystallinity was determined by using the following equation:

% Bulk crystallinity = (�Hm)/(�H ∗m) × 100%,

where �H ∗m, the enthalpy of fusion for 100% crystalline PHB, was taken to be equal

to 146 J/g [30].

Osteoblast studies

Cell morphology assessment and cell proliferation assay. Cell morphologyand proliferation assays involved culturing MC3T3-E1-S14 (MC3T3) cells ontissue culture plastic, PHBV or glass substrates for 2 and 4 days. The MC3T3cells were grown in Minimal Essential Medium (Invitrogen, Carlsbad, CA, USA)supplemented with 10% fetal bovine serum, 50 units/ml penicillin G sodium,50 µg/ml of streptomycin sulfate and 2 mM L-Glutamax (Invitrogen). To ensure thatthe cells only adhered to the glass or PHBV substrates and did not migrate onto thetissue culture plastic wells into which the glass and PHBV substrates were placed,the tissue-culture plastic was pre-coated with 30 µl/cm2 12% poly(2-hydroxyethylmethacrylate) (PHEMA) in 95% ethanol (Sigma). PHEMA was allowed to hardenovernight, the different PHBV substrates were placed into wells and the MC3T3

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

Investigations into PHBV surface properties causing delayed osteoblast growth 1107

cells were seeded the following day at 2000 cells/well in a 96-well plate and grownfor 2 and 4 days at 37◦C in 5% CO2. For comparative studies between PHBV and areference substrate, glass was used to allow for PHEMA treatment in all wells andremoval of all substrates for microscopic analysis.

To assess cell morphology at the end of the culture period, the cells were incubatedfor 30 min at 37◦C in 4% paraformaldehyde. The cells were then washed 3 timesin de-ionized water, incubated in Harris haematoxylin for 5 min and then washed3 times in tap water. The PHBV and glass cover slips where then mounted inaqueous mounting media. Cells were visualized and imaged on an Olympus IX-70microscope (Olympus Australia, Mount Waverley, Australia) using a SPOT RTcamera and SPOTTM 3.2.6 software (Diagnostic Instruments, Sterling Heights, MI,USA).

To allow for detailed assessment of cell morphology by SEM the cells grownon the different PHBV substrates were fixed and dehydrated prior to platinumcoating. Briefly cells were fixed for 30 min at 4◦C with 3% (v/v) glutaldehydein cacodylate buffer (0.1 M sodium cacodylate), then washed 4 times for 10 minin cacodylate buffer at 4◦C. The samples were then exposed to 1% (v/v) osmiumtetroxide in cacodylate buffer for 1 h at 4◦C and washed twice in cacodylate bufferfollowed by pure water for 10 min at 4◦C. The samples were then dehydratedstep wise by washing for 10 min at room temperature in 60, 70, 80, 90 and100% ethanol. The samples were then washed for 10 min at room temperaturein 100% ethanol/hexamethyldisilazane (HMDS) (2:1), 100% ethanol/HMDS (1:1)and 100% ethanol/HMDS (1:2). Finally the samples were washed twice for 10 minat room temperature in 100% HMDS before being allowed to air-dry over night.For SEM analysis the samples were then mounted on aluminium stubs and sputtercoated with platinum (Eiko, Ibaraki, Japan). SEM images were obtained using aJeol-6400F scanning electron microscope at an accelerating voltage of 5 to 10 kV.Digital images were captured and saved using an image slaver software program.

Cell proliferation was assessed indirectly using the CellTiter® 96 AqueousOne Solution Cell Proliferation Assay, a methylthiazol tetrazolium (MTT) assay(Promega, Madison, WI, USA), which measures the activity of the subcellularmitochondrial enzyme succinate dehydrogenase spectrophotometrically. At the endof the culture period a final concentration of 200 µg/ml of MTT solution was addedto all wells and the cells were incubated for a further 2 h at 37◦C in 5% CO2 beforebeing read on a Powerwave XS spectrophotometer using KCjunior software (Bio-TEK Instruments, Winooski, VT, USA) at 490 nm using a reference wavelength of600 nm.

To assess the effect of conditioned media, solvent and lipopolysaccharide (LPS)on MC3T3 cell viability, the cells were cultured on tissue culture plastic and treatedwith these respective agents. Cells were allowed to adhere and then treated for 2 or4 days with media that had been conditioned by soaking in either, CHCl3, DCM orMIX cast PHBV for either 2 or 4 days. For LPS and solvent treatment, the cells weretreated with 0.01–10 ng/ml of LPS or 1–0.0001% of the respective solvents from the

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

1108 I. Keen et al.

time of cell seeding for either 2 or 4 days, and cell viability was then assessed usingthe MTT assay described above. LPS, derived from Salmonella minesata whichhad been chromatographically purified by gel filtration, was purchased from Sigma.Bioactivity of LPS-at the concentrations used was confirmed by measuring cytokineproduction from LPS-treated macrophages (data not shown).

Statistical analysis

Statistically significant differences in the proliferation assays were determined byone-way analysis of variance (ANOVA) followed by a Tukey’s Multiple Compar-ison Test using the GraphPad software program PRISM (GraphPad Software, SanDiego, CA, USA).

RESULTS

Roughness

Chloroform and a mixture of chloroform/acetone solvents have been used previ-ously to change the surface features of biodegradable polymers other than PHBV[26]. In the current study, DCM was used in addition to these two solvent sys-tems. The solvent casting of the films produced two surfaces, a rough side (at theair interface) and a smooth side (from the glass interface). A combination of thevisual information from the SEM images (Fig. 1), the Ra (the mean amplitude ofsurface features relative to the centre plane) and Rz (the maximum height betweenpits and groves) values as measured using SPM (Table 1) provided a full descriptionof the surface roughness. The Rz values paralleled the Ra values and, therefore, thedescription below is limited to include Ra values only.

The rough surfaces of all three films displayed relatively high Ra values (150–210 nm), which were not significantly different between the films (Table 1).However, comparison of the SEM micrographs of the three rough surfaces revealedstriking differences in the lateral patterns on the polymer surfaces (Fig. 1). TheCHCl3-R and MIX-R surfaces were highly textured, displaying extensive pits andridges compared to the DCM-R surface. Specifically, the CHCl3-R surface had pitsof 2–5 µm in diameter and up to 2 µm apart; while larger pits in the range of 5–20 µmand up to 8 µm in width were generated on the PHBV surface when a mixture ofchloroform and acetone was used to produce the film (i.e., MIX-R).

From the SEM images it can be seen that the smooth surfaces were uniformand featureless compared to their respective rough sides and their Ra values wereall significantly different from their respective rough partner. In addition, the Ra

values of the smooth material surfaces were significantly different from each other,with the MIX-S surface displaying the smallest Ra value (20 ± 10 nm) and afeatureless surface similar to that found previously for melt processed PHBV [31].Interestingly, the lateral surface features of the DCM-R, CHCl3-S and DCM-S filmswere very similar, as evident from the SEM images.

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

Investigations into PHBV surface properties causing delayed osteoblast growth 1109

Figure 1. Scanning electron microscopy images of the solvent-cast PHBV films obtained at 1400×magnification. Scale bar represents 10 µm.

Surface hydrophobicity

Advancing contact angles of the different PHBV film surfaces are tabulated inTable 1. There was no significant difference between the θA values of the roughside compared to the smooth side of each of the films; however, the tendency wasfor the rough side to have a slightly higher advancing contact angle. The advancingcontact angles for the CHCl3-R and DCM-R surfaces were similar but 8–10◦ higherthan the MIX-R surface. The smooth side of the films followed a similar trend

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

1110 I. Keen et al.

Table 1.Surface properties of PHBV substrates

Samplea Ra Rz θA θA–θR CI index Average Bulk(nm) (nm) (◦) (◦) min–max (range) CI index crystallinityb

CHCl3-R 210 ± 40 1700 ± 550 80 ± 3 18 ± 9 0.9–1.1 (0.2) 1.0 ± 0.1 35%CHCl3-S 80 ± 10 750 ± 180 80 ± 3 20 ± 6 0.7–1.0 (0.3) 0.8 ± 0.1DCM-R 160 ± 40 1300 ± 400 82 ± 4 15 ± 8 0.9–1.2 (0.3) 1.1 ± 0.1 38%DCM-S 40 ± 10 350 ± 150 75 ± 2 12 ± 4 1.0–1.6 (0.6) 1.1 ± 0.1MIX-R 150 ± 50 1300 ± 550 72 ± 5 13 ± 6 0.8–1.2 (0.4) 1.1 ± 0.1 37%MIX-S 20 ± 10 200 ± 70 69 ± 2 12 ± 3 0.8–1.2 (0.4) 0.9 ± 0.1

a See Materials and Methods for an explanation of the acronyms used.b Determined by XRD.

with the CHCl3-S and DCM-S surfaces having higher advancing contact anglesthan the MIX-S surface. Contact angle hysteresis (θA–θR) was low for all samplesand the surfaces with the lowest roughness values (DCM-S and MIX-S) displayedthe lowest hysteresis as expected.

Crystallinity

The bulk crystallinity of the PHBV powder as received was measured by XRD tobe 51%, whereas that of the three films fell within the range of 35–38%, showingthat the crystallinity of these films was not significantly different (Table 1). Ingeneral, samples with differing crystallinity are measured to obtain intrinsic valuesto be utilised for the calculation of percent crystallinity by indirect methods such asDSC and vibrational spectroscopy [32–34]. In the current study, the crystallinitycalculated from the DSC measurement agreed with that obtained by XRD for eachsample when the value of 146 J/g for the enthalpy of fusion for 100% crystallinePHB was used [30].

Changes in infrared band intensity, band shape or position during heating canreveal bands in the infrared spectra that are sensitive to a change in crystallinity[32–34]. Bloembergen et al. observed from the infrared spectrum of PHBV thatthe intensity of the band at 1185 cm−1 displayed the largest difference betweencrystalline and amorphous states. A crystallinity index, CI, was determined bynormalising the 1185 cm−1 band to that of the 1382 cm−1 band, which was found tobe insensitive to the degree of crystallinity [34] (Fig. 2). PHBV powder as receivedyielded a CI value of 1.2 ± 0.1, while the rough side of a solvent-cast CHCl3 filmyielded a CI value of 0.9 ± 0.1. These CI values were obtained using a diamondATR accessory which measures the infrared spectrum from an area of 0.75 mm2

and to a depth of approximately 1 µm. Comparing these values with the absolutecrystallinity obtained by XRD it can be seen that a higher CI value correspondsto a higher crystallinity, thus verifying the use of the CI values to assess relativemagnitudes of crystallinity. Since different FT-IR techniques can result in differentCI values it is not possible to compare the results obtained here with those obtained

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

Investigations into PHBV surface properties causing delayed osteoblast growth 1111

Figure 2. Diamond ATR–FT-IR spectra obtained from a CHCl3-cast PHBV film (top spectrum) andPHBV powder as received (bottom spectrum). When the absorbance of the 1382 cm−1 band wasdivided by that of the 1185 cm−1 band the top spectrum exhibited a lower CI value compared to thebottom spectrum.

by other groups [34, 35].The fraction and location of crystalline and non-crystalline domains within a

polymer is affected by the processing method [16, 24–26]. By acquiring aninfrared map by the point illumination method using an ATR objective, the lateraldistribution of components across a sample surface in heterogeneous samples canbe explored [36]. The micro ATR–FT-IR technique probes the outmost 0.7 µm ofthe surface using evanescent infrared light [37] and can be used to map a surface bybuilding up a mosaic of infrared spectra recorded at discrete points in a grid pattern,where each point has an area of 40 µm × 40 µm. The micro ATR–FT-IR map of CIvalues for the DCM-S surface (Fig. 3A) display a large range (difference betweenthe highest and the lowest CI value) of CI values of 0.6 (Table 1). The CI mapof the CHCl3-R surface showed a significantly smaller CI range, thus presenting aless heterogeneous surface (Fig. 3B). Surface CI maps of all surfaces were obtainedand their CI range values are tabulated in Table 1. The DCM-S surface showed thehighest CI value (i.e., 1.6) and largest CI range while the CHCl3-S surface showedthe lowest surface CI value (i.e., 0.7). The average CI values (obtained by averagingall the points in the grid; Table 1) are the same within experimental error, whichhighlights the necessity of evaluating the lateral distribution of this surface feature.

Osteoblast morphology

To investigate the influence of the PHBV material surface properties on themorphology of osteoblasts, in vitro culture of MC3T3 cells on the different PHBV

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

1112 I. Keen et al.

(A)

(B)

Figure 3. Micro ATR–FT-IR CI maps obtained from the (A) smooth side of a DCM-cast PHBV and(B) rough side of a CHCl3-cast PHBV film. In the CI gradient the red colouration indicates the highestCI value. This figure is published in colour at http://www.ingenta.com

substrates was performed for 4 days, after which the cells were fixed and stainedwith Harris hematoxylin (Fig. 4). On the CHCl3-R, CHCl3-S, DCM-R, DCM-S andMIX-S surfaces the morphology of the cells was similar to that of cells grown onglass. The majority of cells on these surfaces appeared flat and highly spread with

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

Investigations into PHBV surface properties causing delayed osteoblast growth 1113

Figure 4. Morphology of MC3T3 osteoblast cells grown on various substrates for 4 days. Theseimages are representative of two independent experiments. Magnification 10×. This figure ispublished in colour at http://www.ingenta.com

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

1114 I. Keen et al.

Figure 5. Scanning electron microscopy images of MC3T3 cells grown on DCM-R and MIX-RPHBV films. These images are representative of two independent experiments. Scale bar represents10 µm for 500× and 1 µm for 5000× magnification.

significant cytoplasmic processes extending onto the material (Fig. 4). However, themorphology of the cells grown on the MIX-R surfaces was dramatically differentto that of cells grown on glass. These cells were less spread and exhibited reducednumbers of cellular processes extending on to the MIX-R surface (Fig. 4). Tofurther visualize the differences in cell morphology due to the different PHBVsubstrates SEM analysis was performed on cells grown on the substrates for 4days. The SEM analysis confirmed the flat spread morphology of cells grown onmost surfaces; that of cells on the DCM-R surface is shown in Fig. 5 (DCM-R500×). In comparison, cells growing on the MIX-R surface appeared smaller andless spread on the material surface (Fig. 5, MIX-R 500×). High-magnificationimages of these cells revealed that the cells growing on the DCM-R surface hadorganised pseudopod interactions with the material (white arrows), while the cellson the MIX-R surface had less pseudopods extending on the material surface andthose that were present appeared highly disorganised (Fig. 5, DCM-R 5000× andMIX-R 5000×). There was also a striking reduction in the number of cells seen onthe MIX-R surface compared to all the other substrates tested, suggesting that theproliferation of cells was significantly influenced by the different PHBV substrates(Fig. 4).

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

Investigations into PHBV surface properties causing delayed osteoblast growth 1115

Osteoblast proliferation

To identify differences in the proliferation of osteoblasts grown on the differentPHBV substrates an indirect assessment of proliferation was used. A MTT assaywas performed on cells grown on different PHBV substrates to directly measurethe metabolic activity of the cells. As the metabolic activity of the MC3T3 cellsis consistent regardless of the growth substrate, a change in metabolic activityrefects a change in cell number and hence the proliferation of cells over time onthe different PHBV substrates and as such MTT analysis was used to indirectlymeasure cell proliferation. There was significantly less proliferation of osteoblastsgrown on all PHBV substrates at day 2 compared to the cells grown on glass(Fig. 6A). For some substrates, this lag in cell proliferation was temporary as by day4 there were no significant differences between the growth of cells seeded on glass,CHCl3-R, CHCl3-S or the DCM-R substrates, indicating that all these substratescan effectively support osteoblast proliferation. However, for three of the PHBVsubstrates, DCM-S, MIX-R and MIX-S, the lag in proliferation persisted, as cellsgrown on these films showed reduced proliferation compared to all other substratesat day 4 (Fig. 6B).

Casting of the PHBV films involves the use of solvents that can be toxic to cellsand two alternative approaches were taken to investigate if the solvent cast films hadresidual solvent contaminants that were influencing the growth of the osteoblasts.Firstly, pieces of PHBV were initially bathed in media prior to the media beingadded to the cells for the indicated times. Conditioning the media by pre soakingit with different PHBV substrates did not affect proliferation of the osteoblasts(Table 2). The second approach involved adding the two different solvents and thesolvent mixture to osteoblasts grown on tissue-culture plastic. Only at extremelyhigh solvent concentrations of 1% (v/v) was the proliferation of the osteoblastsaffected (Table 3). Considering that the weight of the PHBV samples used in thecell assays were approximately 2 mg and that no solvent could be detected by FT-IR (detection limit estimated to 5%), the maximum amount of solvent in a PHBVsample would be 0.05 µl. Taking into account that the cell assay used a volumeof 200 µl of media, a maximum concentration of solvent produced by elution fromthe PHBV film would be equivalent to adding 0.025% (v/v) of solvent to the cellsin culture and this level of contamination had no effect on the proliferation of theMC3T3 cells (Table 3).

PHBV is produced by the fermentation of Gram-negative bacteria and it has beenshown that it contains the pyrogen LPS [38]. Osteoblasts express the receptor forLPS, Toll-like receptor 4, and LPS has been shown to induce expression of theessential osteoclast regulating cytokine RANKL in osteoblasts [39]. LPS is alsoknown to influence the proliferation of other cells and thus, to investigate if LPSwas influencing the proliferation of osteoblasts MC3T3 cells were treated withdifferent concentrations of LPS and their proliferation assessed. No difference incell proliferation was seen at any of the LPS concentrations investigated (Table 4).

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

1116 I. Keen et al.

Table 2.Conditioned media from solvent-cast PHBV films applied to MC3T3 cells foreither 2 or 4 days assayed for proliferation

Days in culture CHCl3 DCM MIX (acetone/chloroform)

2 93 ± 5 102 ± 2 100 ± 14 102 ± 3 107 ± 3 103 ± 2

Data are expressed as the average percent in absorbance compared to cellsgrown in standard media ± standard error of the mean and are representative ofthree independent experiments.

Table 3.Influence of solvent treatment on the proliferation of MC3T3 cells

Solvent Solvent/media (%, v/v) Cell proliferationa

2 days 4 days

CHCl3 1 39 ± 4b 55 ± 1b

0.1 106 ± 2 99.6 ± 0.70.01 101 ± 4 99 ± 20.001 106 ± 3 98 ± 10.0001 96 ± 2 98 ± 1

DCM 1 59 ± 9b 47 ± 16b

0.1 104 ± 3 99 ± 30.01 105 ± 3 102 ± 30.001 104 ± 1 99 ± 40.0001 104 ± 1 102 ± 2

MIX 1 27.5 ± 0.2b 25.6 ± 0.3b

0.1 97.1 ± 0.3 102 ± 20.01 96 ± 2 100.3 ± 0.40.001 94 ± 4 98.7 ± 0.80.0001 98 ± 7 100 ± 3

a Data are expressed as a percent of untreated cells ± standard error of mean, and arerepresentative of three independent experiments.

b Differences between control and treated samples are statistically significant (P < 0.05).

Table 4.Effect of LPS on MC3T3 cell proliferation

LPS Cell proliferation

(ng/ml) 2 days 4 days

0.01 101 ± 7 110 ± 50.1 110 ± 4 97 ± 21 100 ± 5 99.6 ± 0.5

10 105.1 ± 0.5 98 ± 7

Data are expressed as a percent of untreated cells ± standard error of mean and arerepresentative of three independent experiments.

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

Investigations into PHBV surface properties causing delayed osteoblast growth 1117

(A)

(B)

Figure 6. Proliferation of MC3T3 osteoblast cells grown on various substrates: cells grown for 2 (A)or 4 (B) days on the indicated substrates. Data are expressed as average ± standard error of the meanand are representative of three independent experiments. *Significant differences from cells grown onglass (P < 0.05).

DISCUSSION

Stimulation of bone growth on a biomaterial can be improved by modifying theinterface between the biomaterial and its host environment. The material surface

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

1118 I. Keen et al.

properties, roughness, chemistry (including atomic composition and crystallinity)and wettability may all affect the biological response that a given material produces.Identification of the specific surface characteristics responsible for influencing cellbehaviour is challenging because the modification of one surface property resultsin simultaneous changes to other surface properties. Hence, the relationshipof one particular surface property to cell behaviour is not well understood andneeds more rigorous study. Few studies exploring osteoblast responses to surfaceproperties have thoroughly investigated all surface parameters of the biomaterialbeing investigated. In the present study, there is no variation in surface compositionbetween samples (as evident from X-ray photoelectron spectroscopy, data notshown) and all other surface properties have been characterised. This has allowedthe surface characteristics (roughness, crystallinity and wettability) to be evaluatedfor their importance in influencing the growth of osteoblasts on the biomaterialPHBV.

Surface properties

Variations in surface features can be obtained using different solvents during thecasting of PHBV films. The pit patterns produced at the air interphase (the roughsurfaces) were produced by water droplets condensing on the surface as the solventevaporated from the surface of the film [31, 40]. The Ra values obtained for the threerough surfaces, CHCl3-R, DCM-R and MIX-R, are similar within experimentalerror, however, the surfaces showed vastly different physical features. The CHCl3-Rand DCM-R surfaces showed pits with the diameter sizes being smaller on theDCM-R surface. The size differences of the pit diameters are caused by thedifferent evaporation rate of the solvents. The MIX-R surface had very differentlateral surface features compared to the other rough surfaces. During the solventevaporation process of this film, it was observed that it gelled after about 2 days andtook longer to form a film compared to films made with a single solvent.

The surface of the smooth side of the films showed very little texture as the surfacecharacteristics of the casting substrate were retained as an imprint on the polymersurface [28]. In this case, the surface of the glass Petri dish controlled the surfacefeatures created on the smooth side. The Ra values for these surfaces were all verylow (20–80 nm) with the DCM-S and MIX-S surfaces comparing well with melt-processed PHBV substrates (Ra = 30 nm) [41].

All surfaces were hydrophobic in nature with the surfaces produced using themixed solvents displaying the lowest contact angles. There was a trend for thesmooth surface of each film to display a lower θA value compared to the corre-sponding rough side in general agreement with the Wenzel’s area correction theory.Similar observation has been made by Lampin et al. where a direct relationshipbetween roughness and surface energy for PMMA materials was found [42]. Thelateral distribution of CI values of the different surfaces illustrates their differentlevels of heterogeneity with respect to crystallinity. It was found that the group ofsubstrates with the most hydrophobic surfaces (CHCl3-R, CHCl3-S, DCM-R) dis-

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

Investigations into PHBV surface properties causing delayed osteoblast growth 1119

play the least heterogeneous surface crystallinities (Table 1). The contact angle hys-teresis was generally low but no direct correlation was found between contact anglehysteresis and either Ra values or CI range (surface heterogeneity). Both surfaceroughness and heterogeneity in surface chemistry are features often associated withincreased contact angle hysteresis [43]. For the samples of this study, the smoothsurfaces have higher or similar CI ranges to their rough counterpart; thus, all sam-ples display features that are expected to contribute to the contact angle hysteresis.

Cell viability on PHBV surfaces

To evaluate appropriately the influence of different PHBV surface properties onthe growth of osteoblasts it was essential to eliminate the possibility that potentialsurface contaminants were influencing cell proliferation. LPS is known to be acontaminant of PHBV following its purification from Gram-negative bacteria. Inaddition, it is possible that the casting solvents used to produce the PHBV filmscould leave a residual contaminant on the surface of these substrates and as suchthis was also investigated. Although it cannot be eliminated that surface trappedsolvent molecules at very low concentrations could affect the cell response, this isnot a likely explanation in this study since of the two surfaces produced by castingfrom DCM solution, only one resulted in a lag in cell proliferation beyond 2 days.Addition of LPS, the solvents CHCl3, DCM or a mixture of CHCl3 and acetone, orPHBV-conditioned media to osteoblasts had no affect on their proliferation, whichindicates that the lag in osteoblast proliferation seen with all PHBV surfaces onday 2 (Fig. 5A) was not due to any contaminating factors and, therefore, must be aconsequence of the PHBV substrate.

Surface wettability is a material property known to influence the growth ofosteoblasts [12] and it is widely accepted that cell adhesion and subsequent activityare superior on hydrophilic surfaces [9, 44–46] such as the glass substrate usedin this study. All the PHBV substrates generated in the current study had watercontact angles greater than 65◦ and as such are classified as hydrophobic [47],which is likely to account for the lag phase in cell growth that was seen with allPHBV substrates after 2 days. It is interesting that by day 4, osteoblast proliferationwas not different on the two chloroform and the DCM-R surfaces compared tothe control surfaces, despite the fact that these were the more hydrophobic PHBVsurfaces investigated. This would suggest that, as the advancing contact angle valuesfor all the PHBV substrates fell within the hydrophobic classification, the 12◦ ofvariation between the different PHBV substrates investigated here is not responsiblefor the differences in osteoblast viability seen after 4 days of growth on these PHBVsurfaces. Thus, in order to understand the differences in cell proliferation on thesubstrates observed after 4 days, other surface properties need to be considered.

Studies have shown that cells can respond to variations in crystallinity [20, 21].In the current study, surface cystallinity was measured and its lateral distributionwas seen to vary between the different PHBV substrates. It was found that the threesubstrates which best support cell growth (CHCl3-R, CHCl3-S, DCM-R) have the

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

1120 I. Keen et al.

least heterogeneous surface crystallinity (i.e., smallest CI range). This magnitudein CI range is, however, only slightly different to that of the MIX-S surface and,therefore, it is unlikely that surface crystallinity is the only parameter influencingcell growth.

Surface roughness and gross morphology are other critical parameters that influ-ence cell growth on a given substrate [10, 48, 49]. Many studies have investigatedthe influence of surface roughness on cell proliferation; however, in the processof altering the roughness of a given substrate, surface chemistry is often changedconcurrently, making it challenging to identify accurately the true effect of rough-ness alone [50]. It was pointed out in a recent review that surface roughness ap-pears to promote osteointegration in orthopaedic and oral/maxillofacial applications[9]. There is no consensus on the effect of surface roughness on osteoblast attach-ment and growth on polymer based materials. Although there is precedence forhigher cell proliferation [13] and alkaline phosphatase activity [15] on rough sur-faces at long time points (week or more) other studies find only higher attachmentbut not proliferation on rough surfaces [14, 19]. The PHBV surfaces generated inthe present study have Ra values in the nanometre to sub-micrometre range (20–210 nm). The smoothest surfaces of this study (DCMS and MIX-S with Ra valuesof 20 and 40 nm, respectively) inhibited the proliferation of osteoblasts possibly asa consequence of smaller surface area. All other PHBV substrates (Ra � 80 nm),except the MIX-R surface, supported cell proliferation. Thus, for our substrates itis generally found that cell proliferation is enhanced on rough surfaces.

The finding that the MIX-R surface was the least effective at supporting osteoblastgrowth (observed both with respect to cell number and cell morphology) suggeststhat the lateral surface features are indeed influential. The distance between thesurface ridges on the MIX-R surface was 10× greater than the distances on theDCM-R material (5–20 and 1–2 µm, respectively) and are approaching the actualsize of the MC3T3 cells (30–40 µm). It is likely that this inhibits the ability ofthe cells to spread on these surfaces. As pointed out in a review by Singhvi et al.[11] a number of independent studies have found that surface morphology modulatethe extent of cell spreading and cell shape; and since it has been establishedthat cell shape is a major determinant of cell growth and function, these subtledifferences in the topographical features translate into significant differences incell behaviour. Furthermore, it was pointed out that interaction of cells withsubstrata with nonplanar surfaces (such as PHBV) depends on the dimensions ofthe underlying topography. Thus, it is not surprising that the different topographicalfeatures observed in our samples (e.g., MIX-R and DCM-R) are causing vastdifferences in cell growth.

In our previous study [7] the lag in cell growth was attributed to the highhydophobicity, as well as the topography of the CHCl3-R films used. We can nowspecify that it is solely the hydrophobic nature that causes this initial lag observed onthe CHCl3-R films. Longer-term lag is not observed for this sample as this alreadyhas optimal roughness for MC3T3 cell growth.

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

Investigations into PHBV surface properties causing delayed osteoblast growth 1121

CONCLUSIONS

The reason for an observed lag in cell development on solvent-cast PHBV filmswas investigated in this study. It was established that organic solvents or LPScontaminations were not the cause of this lag. Through thorough characterisation ofthe surface features of a set of PHBV films a greater understanding of the parameterswhich affect osteoblast cell growth on this biomaterial has been obtained. The initiallag in osteoblast proliferation on all substrates is ascribed to the hydrophobic natureof the material. However, the ability of the cells to recover longer term on thesubstrates is affected by both the degree of heterogeneity of the crystallinity and thesurface roughness. From the data presented it would appear that for PHBV surfaceroughness has the greatest influence on osteoblast proliferation and that a roughnessof 80 nm or more is desired. In addition, for surfaces of sufficient roughness,the lateral distribution of topographical features appears to be an additional criticalfactor influencing osteoblast spreading and growth.

Acknowledgements

I. K. and L. J. R. contributed equally to this study. The authors are grateful tothe Australian Research Council (Grant No. DP0343547) for their support of thisproject and the National Health and Medical Research Council (Grant ID 252934)for their Peter Doherty Fellowship to Liza Raggatt. The authors would like tothank Professor Traian V. Chirila for consultation and useful discussions. Theauthors would also like to thank the following people for their assistance with datacollection and analysis: Ms Anya Yago and Dr Graeme Auchterlonie (Centre forMicroanalysis and Microspectroscopy, The University of Queensland) for obtainingthe XRD traces and assisting with the calculation of bulk crystallinity, and DrThor Bostrom (Analytical Electron Microscopy Facility, Queensland University ofTechnology, Brisbane) for SPM measurements.

REFERENCES

1. K. J. Burg, S. Porter and J. F. Kellam, Biomaterials 21, 2347 (2000).2. S. Gogolewski, M. Jovanovic, S. M. Perren, J. G. Dillon and M. K. Hughes, J. Biomed. Mater.

Res. 27, 1135 (1993).3. T. Hammond and J. J. Liggat, in: Degradable Polymers: Principles and Applications, p. 88,

G. Scott and D. Gilead (Eds). Chapman & Hall, London (1995).4. C. Doyle, E. T. Tanner and W. Bonfield, Biomaterials 12, 841 (1991).5. C. Lutton, J. Read and M. Trau, Aust. J. Chem. 54, 621 (2001).6. G. T. Kose, S. Ber, F. Korkusuz and V. Hasirci, J. Mater. Sci. Mater. Med. 14, 121 (2003).7. A. Kumarasuriyar, R. A. Jackson, L. Grøndahl, M. Trau, V. Nurcombe and S. M. Cool, Tissue

Eng. 11, 1281 (2005).8. U. Meyer, A. Büchter, H. P. Wiesmann, U. Joss and D. B. Jones, Eur. Cell Mater. 9, 39 (2005).9. C. J. Wilson, R. E. Clegg, D. I. Leavesley and M. J. Pearcy, Tissue Eng. 11, 1 (2005).

10. R. G. Flemming, C. J. Murphy, G. A. Abrams, S. L. Goodman and P. F. Nealey, Biomaterials 20,573 (1999).

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

1122 I. Keen et al.

11. R. Singhvi, G. Stephanopoulos and D. I. C. Wang, Biotechnol. Bioeng. 43, 764 (1994).12. E. J. Chang, H. H. Kim, J. E. Huh, I. A. Kim, J. S. Ko, C. P. Chung and H. M. Kim, Exp. Cell.

Res. 303, 197 (2005).13. K. Hatano, H. Inoue, T. Kojo, T. Matsunaga, T. Tsujisawa, C. Uchiyama and Y. Uchida, Bone

25, 439 (1999).14. Y. Wan, Y. Wang, Z. Liu, X. Qu, B. Han, J. Bei and S. Wang, Biomaterials 26, 4453 (2005).15. H. Liao, A. S. Andersson, D. Sutherland, S. Petronis, B. Kasemo and P. Thomsen, Biomaterials

24, 649 (2003).16. N. R. Washburn, K. M. Yamada, C. G. Simon Jr., S. B. Kennedy and E. J. Amis, Biomaterials

25, 1215 (2004).17. Y. Ikada, Biomaterials 15, 725 (1994).18. Y. Tamada and Y. Ikada, J. Biomed. Mater. Res. 28, 783 (1994).19. M. Riehle, M. Dalby, H. Johnstone, A. Mackintosh and S. Afforssman, Mater. Sci. Eng. C23,

337 (2003).20. B. D. Boyan, T. W. Hummert, D. D. Dean and Z. Schwartz, Biomaterials 17, 137 (1996).21. D. Hanein, B. Geiger and L. Addadi, Science 263, 1413 (1994).22. S. Faghihi, M. R. Batene, F. Azari, J. A. Szpunar, H. Vali and M. Tabrizian, Mater. Sci. Forum

495–497, 705 (2005).23. K. D. Chesmel, C. C. Clark, C. T. Brighton and J. Black, J. Biomed. Mater. Res. 29, 1101 (1995).24. W. S. Koegler and L. G. Griffith, Biomaterials 25, 2819 (2004).25. M. Kiremitci, M. Pulat, C. Senvar, A. I. Serbetci and E. Piskin, Clin. Mater. 6, 227 (1990).26. J. Zeng, X. Chen, X. Xu, Q. Liang, X. Bian, L. Yang and X. Jing, J. Appl. Polym. Sci. 89, 1085

(2003).27. K. Matsuzaka, X. F. Walboomers, J. E. de Ruijter and J. A. Jansen, Biomaterials 20, 1293 (1999).28. R. Smith and R. Pitrola, J. Appl. Polym. Sci. 83, 997 (2002).29. H. Y. Erbil, in: Handbook of Surface and Colloid Chemistry, p. 5, K. S. Birdi (Ed.). CRC Press,

Boca Raton, FL (2003).30. M. Avella, G. La Rota, E. Martuscelli, M. Raimo, P. Sadocco, G. Elegir and R. Riva, J. Mater.

Sci. 35, 829 (2000).31. L. Grøndahl, A. Chandler-Temple and M. Trau, Biomacromolecules 6, 2197 (2005).32. J. P. Runt, in: Encyclopedia of Polymer Science and Engineering, Vol. 4, p. 482, H. F. Mark,

N. M. Bikales, C. G. Overberger, G. Menges and J. Kroschwitz (Eds). Wiley, New York, NY(1986).

33. D. L. Bower and W. F. Maddams, in: The Vibrational Spectroscopy of Polymers, p. 264, D. Bowerand W. Maddams (Eds). Cambridge University Press, Cambridge (1988).

34. S. Bloembergen, D. Holden, G. Hamer, T. L. Bluhm and R. H. Marchessault, Macromolecules19, 2865 (1986).

35. N. Galego, C. Rozsa, R. Sanshez, J. Fung, A. Vazquez and J. Tomas, Polym. Test. 19, 485 (2000).36. I. Keen, L. Rintoul and P. Fredericks, Appl. Spectrosc. 55, 984 (2001).37. F. M. Mirabella Jr. (Ed.), in: Internal Reflection Spectroscopy: Theory and Applications,

Practical Spectroscopy Series, Vol. 15, p. 17. Marcel Dekker, New York, NY (1993).38. A. C. K. Wu, L. Grøndahl, K. S. Jack, M. X. Foo, M. Trau, D. A. Hume and A. I. Cassady,

Biomaterials 27, 4715 (2006).39. T. Kikuchi, T. Matsuguchi, N. Tsuboi, A. Mitani, S. Tanaka, M. Matsuoka, G. Yamamoto,

T. Hishikawa, T. Noguchi and Y. Yoshikai, J. Immunol. 166, 3574 (2001).40. N. Marayuma, T. Koito, J. Nishida, T. Sawadaishi, X. Cieren, K. Ijiro, O. Karthasus and

M. Shimomura, Thin Solid Films 327–329, 854 (1998).41. B. Kenny, Synthesis of a novel biomimetic scaffold for bone implantation. Honours Thesis,

University of Queensland, Brisbane (2002).42. M. Lampin, C. Warocquier, C. Legris, M. Desgrange and M. F. Sigot-Luizard, J. Biomed. Mater.

Res. 36, 99 (1997).

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015

Investigations into PHBV surface properties causing delayed osteoblast growth 1123

43. D. Kwok and A. Neumann, Adv. Colloid Interface Sci. 81, 167 (1999).44. G. Altankov and T. Groth, J. Biomater. Sci. Polymer Edn 8, 299 (1996).45. T. Groth and G. Altankov, Biomaterials 17, 1227 (1996).46. S. A. Redey, M. Nardin, D. Bernache-Assolant, C. Rey, P. Delannoy, L. Sedel and P. J. Marie,

J. Biomed. Mater. Res. 50, 353 (2000).47. E. A. Vogler, Adv. Colloid Interface Sci. 74, 69 (1998).48. K. Anselme, M. Bigerelle, B. Noel, A. Iost and P. Hardouin, J. Biomed. Mater. Res. 60, 529

(2002).49. M. J. Dalby, L. Di Silvio, G. W. Davies and W. Bonfield, J. Mater. Sci. Mater. Med. 11, 805

(2000).50. J. Lincks, B. D. Boyan, C. R. Blanchard, C. H. Lohmann, Y. Liu, D. L. Cochran, D. D. Dean and

Z. Schwartz, Biomaterials 19, 2219 (1998).

Dow

nloa

ded

by [

UQ

Lib

rary

] at

18:

13 2

4 M

ay 2

015