A new approach to graft bioactive polymer on titanium implants: Improvement of MG 63 cell differentiation onto this coating

Available online at www.sciencedirect.com

Acta Biomaterialia 5 (2009) 124–133

www.elsevier.com/locate/actabiomat

A new approach to graft bioactive polymer on titanium implants:Improvement of MG 63 cell differentiation onto this coating

Gerard Helary *, Flavie Noirclere, Josselin Mayingi, Veronique Migonney

Laboratoire des Biomateriaux et Polymeres de Specialite, UMR 7032, Universite Paris 13, Avenue Jean Baptiste Clement, 93430 Villetaneuse, France

Received 3 March 2008; received in revised form 28 July 2008; accepted 31 July 2008Available online 28 August 2008

Abstract

Integration of titanium implants into bone is only passive and the resulting fixation is mainly mechanical in nature, with anchoragefailure. Our objective, to increase the biointegration of the implant and the bone tissue, could be obtained by grafting a bioactive ionicpolymer to the surface of the titanium by a covalent bond. In this paper, we report the grafting of an ionic polymer model poly(sodiumstyrene sulfonate) (polyNaSS), in a two-step reaction procedure. Treatment of the titanium surface by a mixture of sulfuric acid andhydrogen peroxide allows the formation of titanium hydroxide and titanium peroxide. In the second reaction step, heating of a metalimplant, placed in a concentrated solution of sodium styrene sulfonate monomer (NaSS), induces the decomposition of titaniumperoxides with the formation of radicals capable of initiating the polymerization of NaSS. Various parameters, such as temperatureof polymerization and time of polymerization, were studied in order to optimize the yield of polyNaSS grafting. Colorimetry,Fourier-transformed infrared spectra recorded in an attenuated total reflection, X-ray photoelectron spectroscopy techniques andcontact angle measurements were applied to characterize the surfaces. MG63 osteoblastic cell response was studied on polished, oxidizedand grafted titanium samples. Cell adhesion, alkaline phosphatase activity and calcium nodules formation were significantly enhanced ongrafted titanium samples compared to unmodified surfaces.� 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Keywords: Titanium; Surface grafting; Cell differentiation; Bioactive polymer

1. Introduction

Although titanium implants are widely used in ortho-paedic fields, insufficient integration into surrounding boneoften occurs. An uncontrolled inflammation process induc-ing fibrous capsule formation prevents the generation of astable implant to host tissue binding and consequentlyimplants can fail under shear stress, requiring revision sur-gery [1].

Over the last 10 years, many studies have been devotedto increasing the osteointegration by modifying surfaceproperties (roughness, topography, surface charges, passiv-ation and wettability) using different methods, such asmechanical treatment [2,3], chemical treatment [4], thermal

1742-7061/$ - see front matter � 2008 Acta Materialia Inc. Published by Else

doi:10.1016/j.actbio.2008.07.037

* Corresponding author. Tel.: +33 1 49 40 36 80.E-mail address: [email protected] (G. Helary).

treatment [5,6] and electrochemical methods [7,8]. How-ever, even if modified surfaces have a higher early level ofcell attachment than untreated titanium surface, the suc-cessful implantation rate is not satisfying.

Biochemical methods of surface modification are prom-ising approaches, at least at the laboratory level [9–11]. Theaim is to control the tissue–implant interface by the immo-bilization of proteins, enzymes or peptides for the purposeof inducing specific cell responses. The most investigatedpeptide sequence immobilized to titanium or polymer sur-faces is RGD (arginine–glycine–aspartic acid) derived fromfibronectin and recognized by almost all a/b integrins [12].The main difficulty is to ensure the stability of the biomol-ecules binding to the surface of the implant and its accessi-bility to active sites of cells. Physical adsorption is notsuccessful for long-term implantation mainly due to thedesorption of biomolecules. The covalent attachment of

vier Ltd. All rights reserved.

G. Helary et al. / Acta Biomaterialia 5 (2009) 124–133 125

biomolecules to the titanium surface allows the problem ofdisruption under physiological medium or mechanicalstress to be resolved. Covalent attachment requires theuse of different chemical reactions which can be aggressivetowards the biomolecules, reducing their potential bioac-tivity [13]. Other problems are the cost of the biomolecules,which can be a limitation to industrial applications, as wellas the possible in vivo enzymatic degradation.

In our laboratory, we have shown that polymers bearingappropriate chemical functions can modulate the cellattachment and spreading onto these bioactive polymersand the cell activity [14,15]. The distribution of these ionicgroups along the macromolecular chains creates active siteswhich can interact with extracellular proteins, such asfibronectin, implicated in cell response.

In this paper, we describe the grafting of an ionic groupsuch as sulfonate groups by radical polymerization ofmonomers bearing these chemical functions (sodium sty-rene sulfonate NaSS). Oxidation of the titanium createstitanium peroxides at the surface which under heating pro-duce radicals initiating the polymerization of ionic mono-mer NaSS. Various parameters to optimize the graftingof polyNaSS on titanium samples were studied. The influ-ence of the grafted surfaces on MG63 cells adhesion andmineralization is discussed.

2. Materials and methods

2.1. Materials

Titanium foil from Alfa Aesar (purity: 99.7%; 0.5 mm inthickness) was cut into squares 1 cm2 in size. Sodium sty-rene sulfonate NaSS, hydrogen peroxide (30 wt.%) werepurchased from Fluka. Sodium styrene sulfonate was puri-fied by recrystallization in a mixture of water/ethanol 10/90 vol.%.

2.2. Preparation of titanium samples

Titanium samples were mechanically polished with wet-ted metallographic polishing (grade 800, 1000 and 1200).These samples were referred to as Ti1200. After polishing,Ti1200 samples were cleaned in an ultrasonic bath: 10 minin acetone followed by 20 min in distilled water. Prior tocell culture, all samples were washed with 1.5 M NaCl,0.15 M NaCl, pure water and phosphate-buffered saline(PBS), three times each, and sterilized by ultraviolet (UV)irradiation under a germicide UV lamp with a power of30 W applied for 15 min on each titanium face, which isnot enough to degrade the polymeric coating.

2.3. Grafting of bioactive polymers on titanium

Ti1200 samples were ultrasonically cleaned in the follow-ing solvents: hexane, acetone and water, for 15 min each.After drying under vacuum for a few hours, there werekept under argon. Oxidation was performed by immersion

in sulphuric acid for 1 min before an identical volume ofhydrogen peroxide solution was added. (The details of thisprocedure will be presented in the Sections 3 and 4.) Thensamples were rinsed extensively with water. After drying,samples were immersed in an aqueous solution containingmonomers (NaSS) at a concentration of 0.7 M. The solu-tion was heated at a temperature of 70 �C for 15 h. Sampleswere rinsed three times under stirring with water to removenon-reacted monomers.

2.4. Surface characterization

2.4.1. ATR–FTIR

Fourier-transformed infrared (FTIR) spectra, recordedin an attenuated total reflection (ATR), were obtained usinga Thermo Nicolet Avatar 370 Spectrometer. Spectra wereobtained with a 4 cm�1 resolution using a 45� Ge crystal.Samples were pressed against the crystal using a smart Omnisampler. Data presented are averaged from 128 spectra.

2.4.2. Colorimetric method

Complexation of quaternary ammonium groups of tolu-idine blue with sulfonate groups allows the amount ofNaSS grafted to titanium samples to be determined.According to Ikada et al. [16], the toluidine blue is capableof making a complex through its N+(CH3)2 with the sam-ple’s carboxylate. This concept was transposed to thesodium styrene sulfonate by preparing a network of polyN-aSS which was immersed in a solution of toluidine blue.After rinsing, each sulfonate group was complexed withthe N+(CH3)2 of the marker. Grafted samples wereimmersed in an aqueous solution of toluidine blue(5 � 10�4 M) for 6 h at 30 �C. Then samples were rinsedwith an aqueous solution of NaOH (5 � 10�3 M) in orderto remove uncomplexed dye. A standard series was donewith seven different concentrations between 4 � 10�6 and5 � 10�5 M, which allows us to determine the concentra-tion of decomplexed toluidine blue thanks to the molarextinction coefficient. Decomplexation of toluidine blueoccurs by immersing titanium samples in an aqueous solu-tion of acetic acid (50 vol.%) for 24 h. Concentration ofdecomplexed toluidine blue is measured by visible spectros-copy at 633 nm using a Perkin-Elmer spectrometer lambda25.

2.4.3. XPSX-ray photoelectron spectroscopy (XPS) analyses were

conducted using an Escalab VG 220i-XL spectrometer(VG instruments). X-rays were produced by a monochro-matic Al Ka source of 1486.6 eV. The photoelectron takeoffangle (the angle between the sample normal and the inputaxis of the energy analyzer) for all XPS experiments was45�. This takeoff angle corresponds to a sampling depthof approximately 10 nm. The energy resolution was0.1 eV. Four spots were analyzed for each sample.

The Ellipse program from VG was used to determinepeak areas, calculate the elemental compositions from

126 G. Helary et al. / Acta Biomaterialia 5 (2009) 124–133

those peak areas and peak-fit the high-resolution spectra.The binding energy scale was calibrated by assigning thehydrocarbon peak in the C 1 s high-resolution spectra toa binding energy of 284.8 eV.

2.4.4. Contact angle measurementStatic solvent contact angles were measured using a

DSA10 contact angle measuring system from KRUSSGmbH. A 0.5 ll droplet of solvent was suspended fromthe tip of a microliter syringe supported above the samplestage. The image of the droplet was captured and the con-tact angle was measured using DSA drop shape analysisprogram from KRUSS. The dispersive (cD) and polar(cP) components of the surface energy were calculated fromthe contact angle measurements using the Owens–Wendtmethod. Surface energy c was calculated by addition ofcD and cP. In order to increase the accuracy of our results,four solvents with increasing polarity were used: diiodo-methane, ethylene glycol, formamide and water.

2.4.5. Roughness measurement

Roughness was measured using a tactile profilometer(Surfascan III S from Semicronic firm). The sensor wasequipped with a diamond stylus probe with a pyramidalshape and a vertex angle of 60�, which induces a pressureof 0.2 mN on the sample. Scanning operations were per-formed by a step engine which moves the samples in onedirection at a speed of 0.5 mm s–1. Among the roughnessparameters, we have selected the Ra parameter, averageroughness.

2.5. Biological tests

2.5.1. Cell seeding

MG63 osteoblast-like cells used in the present studywere obtained from the American Type Culture Collection(ATCC No CRL 1427). These cells exhibit alkaline phos-phatase specific activity. The cells were cultured in Dul-becco’s modified Eagle’s medium (DMEM) containing10% foetal bovine serum (FBS) and 1% penicillin/strepto-mycin at 37 �C in an atmosphere of 5% CO2 and 100%humidity. The media were exchanged every 48 h. Onlyearly passage cells were used for these experiments.

A 2 ml aliquot containing 105 cells was added to tita-nium samples placed on the bottom of a 24-well cultureplate. Cells were cultured at 37 �C under a humidifiedatmosphere in a 5% CO2 incubator for different periodsof time, according to the study underway. After incuba-tion, cells were submitted to a shear stress of 8 dyn cm–2

for 15 min. After removing the solution containing non-adhered cells, adhered cells were detached by trypsin diges-tion. The number of adherent cells on each titanium samplewas quantified using a cell counter (Multisizer III—CoulterCounter from Beckman). The resulting cell number wasexpressed as a percentage of the cells originally seeded onthe assessed surface. Experiments were performed in tripli-cate in three separate experiments.

Adherent cells were stained with 2.5% phalloidin (Inter-chim) in PBS buffer 4 h after seeding onto titanium sam-ples. Pictures of stained cells observed under an invertedfluorescence microscope (Axiolab, Zeiss, Germany) weretaken with a digital camera (Diagnostic Instruments,USA). The cell areas were calculated with Image Pro Plussoftware.

2.5.2. ALP biochemical activity

Alkaline phosphatase (ALP) is an early marker of osteo-blast differentiation and relates to the production of a min-eralized matrix. After 14 days, ALP activity was evaluatedon the transformation of p-nitrophenylphosphate intop-nitrophenol at 37 �C and pH 10.2. Briefly, enzyme wasisolated from the cell membrane by reaction with cold Tri-ton X-100 for 1 h under agitation. Then 50 ll of the solu-tion was added to 50 ll of p-nitrophenylphosphatesubstrate in 2-amino-2-methyl-1-propanol buffer (pH10.2). After 2 h incubation at 37 �C, the p-nitrophenol pro-duced was measured by absorbance measurements at405 nm. Enzymatic activity, expressed in nmol of p-nitro-phenol produced per min, was standardized to the massof protein (expressed in mg). Protein content was measuredusing a commercially available colorimetric assay (BCAassay, Interchim). Thus, enzymatic activity is expressed innmol of p-nitrophenol produced per min per mg of protein.

2.5.3. Calcium incorporation assays

After 28 days of culture of MG63 osteoblast-like cellson, respectively, Ti1200, Tiox and Tigrafted, the calcium con-tent was determined by dissolving calcium phosphate with400 ll of trichloroacetic acid for 60 min. A 5 ll quantity ofthe solution was added to 300 ll of arsenazo III-containingcalcium reagent (Diagnostic Services Ltd.). The absorbanceof the resulting samples was measured at 650 nm and com-pared to a linear standard curve of CaCl2 from 50 to 1000in 1000 lg l–1 in trichloroacetic acid (5% w/v).

Calcein (Sigma) labelling was performed after 4 weeksof cell culture. Briefly, the cells were incubated with calceinat a concentration of 25 mg ml–1 for 4 h at 37 �C. After thisperiod, cells were washed twice with PBS, fixed in 3.7%paraformaldehyde in 0.1 M Na-cacodylate buffer, pH 7.4.After rinsing three times with PBS, pictures of stained cellsobserved under an inverted fluorescence microscope (Axio-lab, Zeiss, Germany) were taken with a digital camera(Diagnostic Instruments, USA).

2.5.4. Statistical analysis

Each experiment was repeated at least three times withsimilar results. Results was expressed as mean ± standarddeviation, with n = 9. Errors bars in the figures representstandard deviations. Statistical analysis was performedusing SPSS software. The significance of the obtained datawas performed using one-way analysis of variance, fol-lowed by a Tuckey test comparing grafted and non-graftedsamples. Differences at p P 0.05 were not considered statis-tically significant.

1 3 50

1

2

3

4

5

6

Am

ount

of g

rafte

d po

lyN

aSS

(µg/

cm²)

Time in min

Kinetic of oxidation - Grafting at 70°C

2 4

Fig. 1. Amounts of grafted polymers measured as a function of time ofoxidation through the concentration of decomplexed toluidine blue.

30ºC 50ºC 70ºC0

1

2

3

4

5

6

Am

ount

of g

rafte

d po

lym

ers

(µg/

cm²)

Temperature

Fig. 2. Amounts of grafted polymers measured as a function oftemperature through the concentration of decomplexed toluidine blue.

G. Helary et al. / Acta Biomaterialia 5 (2009) 124–133 127

3. Results

3.1. Grafting of bioactive polymers on titanium

Sulfonate groups were covalently attached to the tita-nium surface by radical polymerization of sodium styrenesulfonate initiated by radicals issued from titanium perox-ide. Immersion of titanium samples in a mixture of puresulphuric acid and hydrogen peroxide (30%) produces onits surface titanium hydroxide and titanium peroxide. Theexistence of Ti-peroxide radical was suggested by differentresearch groups [17,18] by the presence of bands in therange of 890–900 cm�1 on the IR spectra of H2O2-oxidizedTi powder. Takemoto et al. have shown that the radicalspecies decreased from titanium oxide by heat treatment[18]. The heating of an aqueous solution of NaSS in whichwere immersed titanium samples induces decomposition oftitanium peroxide in radicals, thereby initiating the poly-merization of NaSS monomer. Amounts of grafted poly-mers were measured by colorimetric method as describedin Materials and methods, and were considered as the yieldof grafting reaction. Titanium samples Ti1200 and titaniumoxidized samples Tiox were used as control. No colorationwas observed with control samples, confirming that color-ation is only due to the decomplexation of the dye frommonomer units of the bioactive polymer and consequentlyto the presence of polyNaSS at the surface of the titaniumsurface. However, the coloration could be due to the com-plexation of NaSS monomer or polyNaSS merely adsorbedto the surface of titanium with toluidine blue. To rule outthis hypothesis, polyNaSS and NaSS monomer at a con-centration of 0.7 M were put in contact with Tiox for1 day. After this period of time, no reaction with the dyewas observed.

Different parameters, such as time of oxidation, temper-ature and time of polymerization, and monomer concen-tration, influence the yields of the grafting, so theseparameters were studied in order to determine the optimalconditions.

The amount of grafted polymers measured as a functionof time of oxidation (Fig. 1) at the polymerization temper-ature of 70 �C shows that the highest value was obtainedafter 3 min of immersion in the oxidation solution, so thistime was chosen for the study of other parameters. A tem-perature of 70 �C allows 5 lg cm–2 of grafted polymer to beobtained, whereas lower or higher temperatures producelower yields (Fig. 2). At 30 �C, no polymer was grafted.The kinetics of polymerization (Fig. 3) shows that a plateauis reached after 15 h of reaction, indicating the disappear-ance of hydroxide peroxide and thus the ability to produceradicals. It is noteworthy that polymerization kinetics car-ried out at temperatures higher than 70 �C are not accuratedue some evaporation of the solvent. Polymerizations wereperformed in dimethylsulfoxide (boiling point 189 �C) attwo temperatures: 80 and 100 �C. At the low temperature,the amount of grafted polyNaSS was less than 0.3 lg cm–2

and no grafting was observed at 100 �C. The evaluation of

the distribution of polymer grafted at the titanium surfaceis quite difficult because the concentration of radical of Ti-peroxide radical at the surface is not known. Another prob-lem is to determine the efficiency of the radical created toinitiate the polymerization of NaSS. Finally, a third prob-lem is to know the degree of polymerization.

Polymerization kinetics is quite similar to that presentedin Fig. 3, whatever the concentrations of the monomersstudied. This is quite normal because monomer concentra-tions are in a large excess and can be considered as con-stant. In order to obtain high grafting yields, allpolymerizations were performed under inert gas at 70 �Cfor 15 h, with a monomer concentration of 0.7 M.

3.2. Characterization of grafted titanium surfaces with

polyNaSS

Different techniques were used to check the presence ofthe bioactive polymers at the surface of titanium samples:

1h 3h 5h 8h 15h 24h0

1

2

3

4

5

6

Am

ount

of g

rafte

d po

lym

ers

(µg/

cm²)

polymerization time

Fig. 3. Kinetics of radical polymerization of NaSS on a titanium surface(temperature 70 �C) through the concentration of decomplexed toluidineblue.

128 G. Helary et al. / Acta Biomaterialia 5 (2009) 124–133

� IR/ATR spectra of grafted titanium Tigrafted surfacesshow the presence of the bands characteristic of the sul-fonate groups at 1009 cm�1 1039 cm�1 (Fig. 4). No peakwas detected on control samples Ti and Tiox

� XPS analysis

XPS analyses were carried out on samples at the differ-ent steps of grafting bioactive polymers on titanium sur-face: Ti1200, Tiox and Tigrafted. The atomic compositiondetermined from XPS measurements are summarized inTable 1. By considering the Ti1200 sample, the detectedvalue is seen to be in good agreement with the literatureon titanium surfaces [19]. The surface of titanium is cov-ered by a thin oxide layer of TiO2 (459.3 eV) with contam-inants, which introduces a significant amount of carbonand further decreases the concentration of Ti. However,the value of 14.7% is indicative of a comparatively cleansurface.

Oxidation with a mixture of pure sulphuric acid andhydrogen peroxide induces a decrease in the amount of

92

93

94

95

96

97

98

99

100

11150120012501300wavenu

trans

mis

sion

Fig. 4. FTIR/ATR spectra of polyN

titanium due to the presence of sulphured contaminantsfrom the oxidation solution.

The presence of polyNaSS grafted at the surface of tita-nium induces an increase in carbon (by a factor of 1.6) anda decrease in oxygen (by the same factor). The polymercoating contains mainly carbon and less oxygen, explainingthe variation observed. The polymer hides the titaniumoxide indicated by a decrease of the titanium content(5%). Further proof of the presence of polyNaSS is thedetection of atomic sodium (4%) linked to the sulfonategroups. It is noteworthy that atomic S is observed on Tiox

due to the addition of sulphuric acid for the chemical oxi-dation step (contaminants). In fact, the deconvolution ofthe S2p3/2 spectrum of Tigrafted shows the appearance of apeak at 168.1 eV due to the sulfonate group according tothe work of Brion [20].

� Contact angle analysis

The surface wettability and surface energy were deter-mined by measuring the contact angles of polar (water)and non-polar (diiodomethane) liquids on Ti1200, Tiox

and Tigrafted. The total surface energy is the sum of thepolar and dispersive components. The surface energy cal-culation was based on the geometric mean analyses of fourliquid probes: water, formamide, ethylene glycol anddiiodomethane. The contact angles and surface energy(mN m–1) of the substrates are shown on Table 2. The sur-face energy increases from Ti1200 to Tigrafted. This result ismainly due to the polar surface tension component, whichincreases by a factor 2 between Ti1200 and Tiox and by afactor 3 between Ti1200 and Tigrafted. Oxidation and graft-ing of polyNaSS increase the hydrophilic character of thetitanium surface by formation of TiOH in a first step thenby the presence of the ionic polymer polyNaSS. It is note-worthy that in some points of Tigrafted samples the mea-surement of the water contact angle was not possible dueto the absorption of the solvent. This observation couldindicate a heterogeneous distribution of the polymergrafted at the surface of the titanium.

90095010001050100mber (cm-1)

10391009

aSS grafted on titanium surface.

Table 1Surface composition (at.%) of the different samples as detected by XPS analysis

C O Ti S Na Contaminants

Ti1200 33.2 ± 4.4 46.4 ± 2.1 14.7 ± 0.4 5.7 ± 2.2Tiox 33.2 ± 2.6 47.9 ± 1.6 7.4 ± 0.8 5.1 ± 1.7 5.4 ± 0.6Tigrafted 53.8 ± 1.3 30.3 ± 1.9 5.1 ± 1.4 5.8 ± 0.9 3.8 ± 0.8 1.2 ± 0.4

Table 2Surface free energy of Ti1200, Tiox and Tigrafted calculated with the Owens–Wendt model

Surface energy Total surface energy c (mN/m) Dispersive components cd (mN/m) Polar components cP (mN/m) cP/c (%)

Ti1200 44.3 ± 2.3 27.7 ± 1.1 16.6 ± 1.2 37Tiox 59.2 ± 2.3 26.1 ± 1.0 33.1 ± 1.3 56Tigrafted 64.9 ± 1.9 16.9 ± 1.9 48.0 ± 1.2 74

Fig. 5. Cell attachment of MG63 osteoblast-like cells on Ti1200, Tiox andTigrafted. N2 and N1 are the number of adhered cell submitted or non-submitted to the shear stress of 8 dyn cm–2. Nsc is the number of seeded cells.

G. Helary et al. / Acta Biomaterialia 5 (2009) 124–133 129

� Roughness determination

The average roughnesses (Ra) determined for all the tita-nium samples are reported in Table 3. Ti1200 is the smooth-est sample, with Ra = 0.12 lm. The roughnesses of Tiox andTigrafted are comparable, but slightly higher than that ofTi1200 (Table 3). Etching with strong acids such as sulphuricacid is a way to roughen titanium implants [2].

3.3. Biological tests

3.3.1. Cell attachment

The in vitro cellular response to the grafted titanium wasassessed in terms of cell attachment. Some 105 cells wereincubated on Ti1200, Tiox and Tigrafted for 30 min. For shearstresses higher and lower than 8 dyn cm–2, differences in thenumber of remaining adhered cells on the various supportsare not significant. For this reason, a shear stress of8 dyn cm–2 was applied for 15 min. Remaining adheredcells after the shear stress application were detached fromthe supports by trypsin digestion. To define the variationof cell adhesion on Ti1200, Tiox and Tigrafted, the followingexpression was used for each titanium sample: N1–N2/Nsc, where N2 and N1 are the number of adhered cells sub-mitted or non-submitted to the shear stress of 8 dyn cm–2.Nsc is the number of seeded cells. The lower this ratio,the higher the cell adhesion on the support. Analysis ofthe values reported in Fig. 5 shows that the cell adhesionon Tigrafted is higher than on oxidized titanium.

3.3.2. Cell morphology

After 4 h of incubation, the cells are quite spread outwhatever the titanium samples, as reported in Fig. 6. How-ever, cell areas are slightly higher when cells are cultured on

Table 3Average surface roughness (Ra) values for different titanium surfaces bymeans of a tactile profilometer

Average roughness Ra (lm)±SD

Ti1200 0.12 ± 0.02Tiox 0.15 ± 0.03Tigrafted 0.15 ± 0.04

Tigrafted (1325 lm2) compared to Tiox (1150 lm2). Thisresult is in agreement with the lower detachment of cellscultured on Tigrafted under shear force.

3.3.3. ALP activity

The ALP activity of osteoblasts seeded on Ti1200 mea-sured after 14 days is 1.4 nmol min–1 mg–1, a value slightlyhigher than those reported in the literature [21]. The impor-tant point is that ALP activity (Fig. 7) on Tigrafted was sig-nificantly higher than on Tiox and Ti1200 (+19%). Thespreading of cells correlated with the cell adhesion strengthfavours their differentiation.

3.3.4. Calcification

The second step of cell differentiation is the formation ofmineralized bone cell by apparition of calcium nodules.Calcium content of four weeks cultured MG63 cells on tita-nium samples is significantly higher on Tigrafted comparedto Ti1200 as reported in Fig. 8. According to some authors[22], precipitation of calcium could be formed indepen-

Fig. 6. Cell morphology of MG63 osteoblast-like cells after 4 h ofincubation.

Fig. 7. ALP activity of MG63 osteoblast-like cells after 14 days ofproliferation.

Fig. 8. Calcium content of cells matrix at 4 weeks of proliferation.

Fig. 9. Coloration of calcium nodules by calcein at 4 weeks of prolifer-ation (magnification 100�, bar = 200 lm).

130 G. Helary et al. / Acta Biomaterialia 5 (2009) 124–133

dently of the cell activity. To eliminate this hypothesis,DMEM containing 10% FBS was deposited on titaniumsamples and the calcium content was determined after 4weeks. No calcium nodule was detected, indicating thatprecipitation of calcium is due to MG63 cells. Whateverthe parameter studied – adhesion strength, ALP or calcifi-cation – polyNaSS grafted to titanium surface induces anMG63 differentiation. Coloration of calcium nodules bycalcein (Fig. 9) shows differences in the calcification processof cells proliferating on Tigrafted compared to Ti1200.

Indeed, as shown in Fig. 9, the amount and pattern of cal-cification nodules are undefined and very diffuse in the caseof Ti1200, whereas nodules are very distinct, big in size andhighly concentrated in calcium in the case of Tigrafted sam-ples, as shown by the intensity of calcein (see Fig. 9). Thebehaviour of cells on Tiox is intermediate: the amountand pattern of calcification nodules are similar to thoseon Tigrafted, but the size of the nodules is much less impor-tant. These results show that the grafting of the titaniumsample by polyNaSS stimulates mineralized of thematrix.

G. Helary et al. / Acta Biomaterialia 5 (2009) 124–133 131

4. Discussion

In the last decade, many studies have been devoted toimproving the bone cell–implant interaction, with the goalof improving the osseointegration of devices for dentaland orthopaedic applications [23,24]. There are variousways to modify surface implants, and our strategy is to graftpolymer model at their surface. Macromolecular synthesisallows the preparation of well-defined polymers which canbe biologically tested. By this method, we have preparedvarious ionic polymers models such as poly(sodium styrenesulfonate), poly(methacrylic acid) and copolymers contain-ing different compositions of these monomers. These ionicpolymers can modulate the proliferation of cells [14] as wellas their differentiation [15]. For instance, the ALP activityof MG 63 cells cultured on a terpolymer poly(sodium sty-rene sulfonate–methacrylic acid–methyl methacrylate) isthree times higher than on PMMA, considered as the con-trol. The level of ALP activity depends on the chemicalcomposition of the ionic groups, and the highest value isobtained for R = [COO�]/[COO� + NaSS] = 0.7. Graftingthese ionic groups at the titanium surface represents a dou-ble challenge: first, the determination of parameters toobtain the highest grafting yield, and secondly, the controlof the ratio between carboxylic and sulfonate groups. Forthis reason, we have decided in a first step to control thegrafting of the sulfonate group alone by assuming that itis possible to produce radicals at the surface of implantsthat can initiate the polymerization of this monomer. Thisattempt was successful and various surface characterizationtechniques have shown that grafting of polyNaSS at thesurface of titanium reaches 5 lg cm–2, a higher value thanthose reported in the literature for grafted biomolecules[11,13,25]. Our approach is interesting from an industrialpoint of view because two steps are necessary to obtain ahigh yield of polymer grafting and could be easily carriedout for implants with a convoluted shape. The silanizationreaction, which is a common way to graft polymers at thesurface of titanium implants, involves three steps [11].Recently, it has been shown that the electropolymerizationof pyrrole on titanium and titanium alloy substrates is aone-step reaction that enables a coating able to stimulatepositive interactions with bone tissue to be obtained[26,27]. However, the advantage of our method is that thecoating is linked by a covalent bond to the titanium sub-strate whereas the polypyrrole is electrodeposited.

The greatest differentiation of MG63 cells is observed onTigrafted compared to Ti1200 and Tiox, as indicated by thevalues of cell adhesion, ALP activity and mineralizationobtained on Tigrafted cellular behaviour is greatly influencedby surface properties, including hydrophilicity and rough-ness. Many authors have shown that the surface energyhas an influence on the cellular adhesion and proliferationof fibroblasts and osteoblasts [28]. Cell attachment andspreading was higher on hydrophilic surfaces than onhydrophobic surfaces. According to Scharenraad et al.[29], a surface energy of 57 mN m–1 is the threshold

between surfaces which favor adhesion and those whichprevent adhesion. Tiox and Tigrafted have surface energiesof 64.9 and 59.2, respectively, both higher than the thresh-old of 57 mN m–1. These values are in agreement with thecell adhesion strength, which is higher on Tigrafted thanon Tiox. However, cell adhesion strengths are of the sameorder of magnitude on Ti1200 and on Tiox, whereas theenergy surfaces are on each side of the surface energythreshold of 57 mN m–1. According to Hallab and co-workers [30], cellular adhesion is maximal at a fractionalpolarity, FP (cp/(cp + cd)), equal to 0.3. This was confirmedby Martelet et al. [31], who showed that a low FP is themajor prerequisite for good cell proliferation. In our case,FP is equal to 0.3 and 0.7 for Ti1200 and Tigrafted, respec-tively, indicating that the polar force plays an importantrole at least with the osteoblast cell MG63. These observa-tions mean that the hydrophilic character of a surface isnot sufficient to explain our results.

The fact that roughness is quite similar for the samplesstudied, Tiox and Tigrafted, means that this factor asreported by many authors is not the key point to explainingthe different behaviour of cell adhesion [32,33].

The explanation could be an ionic interaction betweenthe surface of titanium and the cell membrane. Someauthors [34] have shown that a surface with a positivecharge (quaternary ammonium groups) favors celladhesion. However, the cell membrane is constitutedof negatively charged phospholipids and glycoproteins.PolyNaSS, which bears negative charges, could drive anelectrostatic repulsion with the cell membrane. In fact,Kowalczynska et al. [35] have shown that sulfonic groupspresent on polymer surfaces increase cell adhesion. Theexplanation for these results is the presence of adhesiveproteins, such as fibronectin, that interact with the cellmembrane and the polyNaSS grafted to the titanium sur-face. This hypothesis was recently demonstrated by hidingthe a5b1 integrins capable of recognizing fibronectin in thepresence of polymer with or without sulfonic groups [36].The cell adhesion was strongly reduced, followed by theinhibition of cell proliferation. On the other hand, blockinga5b1 integrins has no effect on cell adhesion cultured onpolymer without sulfonic groups, indicating that otherintegrin receptors are implicated in the recognition process.

Proteins such as extracellular matrix proteins, cytoskel-etal proteins and membrane receptors (integrins) areinvolved in cell-substrate interactions. Cell adhesion,growth and differentiation depend on interaction betweenthese proteins and their specific receptors. Fibronectin isknown to mediate cell interaction in contact with biomate-rials. The interaction of this protein with polyNaSS graftedto titanium sample could induce a modification of its con-formation to favor the differentiation of MG63 cells.

5. Conclusion

We have succeeded in grafting a bioactive polymer to atitanium surface with high yields by a method that can be

132 G. Helary et al. / Acta Biomaterialia 5 (2009) 124–133

easily adapted to modify implants for medical applications.This finding might be adapted to graft polymers onto othermetallic devices, such as titanium alloys implants, whichare widely used in medical applications. Determination ofthe cell attachment strength, ALP activity and calcificationof MG63 cells seeded onto the grafted titanium surfaceshowed the improved differentiation of these cells, indicat-ing the possibility of obtaining a higher anchorage ofgrafted implants. However, higher activity could beobtained with a mixture of carboxylic acid and sulfonateionic groups, as previously obtained with a polymer model.The second step consists of polymerizing monomers bear-ing these ionic groups at the appropriate ratio. The finalstep consists of achieving a better understanding of themechanisms involved in the cell-grafted titanium at theatomic and molecular levels.

References

[1] Klabunde Windler M. Titanium for hip and knee prostheses. In:Brunette DM T, Tengvall P, Textor M, Thomsen P, editors. Titaniumin medicine. Berlin: Springer; 2001. p. 703–46.

[2] Ban S, Iwaya Y, Kono H, Sato H. Surface modification of titaniumby etching in concentrated sulphuric acid. Dental Mater2006;22:1115–20.

[3] Bigelle M, Anselme K, Noel B, Ruderman I, Hardouin P, Iost A.Improvement in the morphology of Ti-based surfaces: a new processto increase in vitro human osteoblast response. Biomaterials2002;23:1563–77.

[4] Wang J, Layrolle P, Stigter M, De Groot K. Biomimetic andelectrolytic calcium phosphate coatings on titanium alloy: physicalcharacteristics and cell attachment. Biomaterials 2004;25:583–92.

[5] Park YJ, Song HJ, Kim I, Yang HS. Surface characteristics andbioactivity of oxide film on titanium metal formed by thermaloxidation. J Mater Sci Mater Med 2007;18:565–75.

[6] Saldana L, Barranco V, Gonzalez-Carrasco JL, Rodriguez M,Munuera L, Vilaboa N. Thermal oxidation enhances early interac-tions between human osteoblasts and alumina blasted Ti6Al4 V alloy.J Biomed Mater Res 2007;A81:334–46.

[7] Sul YT et al. Characteristics of the surface oxides on turned andelectrochemically oxidized pure titanium implants up to dielectricbreakdown: the oxide thickness, micropore, surface roughness, crystalstructure and chemical composition. Biomaterials 2002;23:491–501.

[8] Le Guehennec L, Soueidan A, Layrolle P, Amouriq Y. Surfacetreatments of titanium dental implants for rapid osseointegration.Dental Mater 2007;23:844–54.

[9] Reyes CD, Petrie TA, Schwartz Z, Garcia AJ. Biomolecular surfacecoating to enhance orthopaedic tissue healing and integration.Biomaterials 2007;28:3228–35.

[10] Pallu S et al. The effect of cyclo-DfKRG peptide immobilization ontitanium on the adhesion and differentiation of human osteoprogen-itor cells. Biomaterials 2005;26:6932–40.

[11] Porte-Durrieu MC et al. Cyclo-(DfKRG) peptide grafting onto Ti–6Al–4 V: physical characterization and interest towards humanosteoprogenitor cells adhesion. Biomaterials 2004;25:4837–46.

[12] Zhao B, Tian W, Feng H, Lee IS, Cui F. Effects of RGD peptidesgrafting to titanium dental implants on the adhesion of humangingival fibroblasts and epithelial cells. Curr Appl Phys2005;5:407–10.

[13] Muller R et al. Influence of surface pretreatment of titanium andcobalt based biomaterials on covalent immobilization of fibrillarcollagen. Biomaterials 2006;27:4059–68.

[14] El Khadali F, Helary G, Pavon-Djavid G, Migonney V. Modulatingfibroblast cell proliferation with functionalized poly(methyl methac-

rylate) based copolymers: chemical composition and monomerdistribution effect. Biomacromolecules 2002;3:51–6.

[15] Anagnostou F, Debet A, Pavon-Djavid G, Goudaby Z, Helary G,Migonney V. Osteoblast functions on functionalized PMMA-basedpolymers exhibiting Staphylococcus aureus adhesion inhibition. Bio-materials 2006;27:3912–9.

[16] Sano S, Kato K, Ikada Y. Introduction of functional groups onto thesurface of polyethylene for protein immobilization. Biomaterials1993;14:817–22.

[17] Tengvall P, Walivaara B, Westerling J, Lundstrom I. Stable titaniumsuperoxide radicals in aqueous Ti-peroxy gels and Ti-peroxidesolutions. J Coll Interf Sci 1991;143:589–92.

[18] Takemoto S, Yamamoto T, Tsuru K, Hayakawa S, Osaka A,Takashima S. Platelet adhesion on titanium oxide gels: effect ofsurface oxidation. Biomaterials 2004;25:3485–92.

[19] Barber TA, Golledge SL, Castner DG, Healy KE. Peptide-modified p(AAm-co-EG/AAc) IPNs grafted to bulk titaniummodulate osteoblast behavior in vitro. J Biomed Mater Res A2003;64:38–47.

[20] Brion D. Etude par spectroscopie de photoelectrons de la degradationsuperficielle de FeS2, CuFeS2, ZnS et PbS a l’air et dans l’eau. ApplSurf Sci 1980;5:133–52.

[21] Massaro C, Baker MA, Cosentino F, Ramires PA, Klose S, Milella E.Surface and biological evaluation of hydroxyapatite-based coatingson titanium deposited by different techniques. J Biomed Mater Res2001;58:651–7.

[22] Webster TJ, Massa-Schlueter Elizabeth A, Smith Jennifer L,Slamovich Elliot B. Osteoblast response to hydroxyapatite dopedwith divalent and trivalent cations. Biomaterials 2004;25:2111–21.

[23] Bagno A et al. Improvement of Anselme’s adhesion model forevaluating human osteoblast response to peptide-grafted titaniumsurfaces. Bone 2007;41:704–12.

[24] Gotz HE, Muller M, Emmel A, Holzwarth U, Erben RG, Stangl R.Effect of surface finish on the osseointegration of laser-treatedtitanium alloy implants. Biomaterials 2004;25:4057–64.

[25] Chollet C, Chanseau C, Brouillaud B, Durrieu MC. RGD peptidesgrafting onto poly(ethylene terephthalate) with well controlleddensities. Biomol Eng 2007;24:477–82.

[26] De Giglio E, Guascito MR, Sabbatini L, Zambonin G. Electrop-olymerization of pyrrole on titanium substrate for the futuredevelopment of new biocompatible surfaces. Biomaterials2001;22:2609–16.

[27] Earley ST, Dowling DP, Lowry JP, Breslin CB. Formation ofadherent polypyrrole coatings on Ti and Ti–6Al–4 V alloy. Synth Met2005;148:111–8.

[28] Redey SA, Razzouk S, Rey C, Bernache-Assollant D, Leroy G,Nardin M, et al. Osteoclast adhesion and activity on synthetichydroxyapatite, carbonated hydroxyapatite and natural calciumcarbonate: relationship to surface energies. J Biomed Mater Res1999;45:140–7.

[29] Schakenraad JM, Busscher HJ, Wilevuur CRH, Arends J. Theinfluence of substratum surface free energy on growth and spreadingof human fibroblasts in the presence and absence of serum proteins.J Biomed Mater Res 1986;20:773–84.

[30] Hallab N, Bundy K, O’Connor V, Clark R, Moses R. Cell adhesionto biomaterials: correlations between surface charge, surface rough-ness, adsorbed protein and cell morphology. J Long-Term Eff MedImplants 1995;53:209–31.

[31] Ponsonnet L, Reybier K, Jaffrezic N, Comte V, Lagneau C, Lissac M,et al. Relationship between surface properties (roughness, wettability)of titanium and titanium alloys and cell behaviour. Mater Sci Eng2003;23:551–60.

[32] Nicula R, Luthen F, Stir M, Nebe B, Burkel E. Spark plasmasintering synthesis of porous nanocrystalline titanium alloys forbiomedical applications. Biomol Eng 2007;24:564–7.

[33] Luthen F, Lange R, Becker P, Rychly J, Beck U, Nebe B. Theinfluence of surface roughness of titanium on b1- and b3-integrin

G. Helary et al. / Acta Biomaterialia 5 (2009) 124–133 133

adhesion and the organization of fibronectin in human osteoblasticcells. Biomaterials 2005;26:2423–40.

[34] Webb K, Hlady V, Tresco PA. Relative importance of surfacewettability and charge functional groups on NIH 3T3 fibroblastattachment, spreading and cytoskeletal organization. J Biomed MaterRes 1998;41:422–30.

[35] Kowalczynska HM, Nowak-Wyrzykowska M. Adhesion of L1210cells to sulfonated styrene copolymer surfaces: imaging of F-actin a-actinin. Cell Biol Int 1999;23:359–72.

[36] Kowalczynska HM, Nowak-Wyrzykowska. Modulation of adhesion,spreading and cytoskeleton organization of 3T3 fibroblasts by sulfonicgroups present on polymer surfaces. Cell Biol Int 2003;27:101–14.