Abstract

Many methods are currently under investigation to improve the integration of dental implants to surrounding bones. Among these methods, peptide-modified surfaces have been highlighted as one of the most promising. Our study, aimed at the cellular response to RGD-immobilized surface in vitro, investigated the basis for designing a bone-active surface coating with RGD-containing peptide. Gold-coated titanium surfaces were used as indicative control surfaces for peptide immobilization. Using self-assembly monolayer techniques, 2 types of peptides, RGDC (Arg-Gly-Asp-Cys) and RDGC (Arg-Asp-Gly-Cys), were immobilized onto the gold surfaces. Surface justification was realized through X-ray photoelectron spectroscopy and Fourier transform infrared spectra. Primary calvarial osteoblasts were cultured on RGDC, RDGC, and non–peptide-coated surfaces. Cell attachment, morphology, proliferation, and expression of osteocalcin (OC) messenger RNA (mRNA) were assessed using cell counting, immunolabeling fluorescence microscopy, and Northern blot assay. Four and 8 hours after culture, cell attachment was enhanced on RGDC surfaces. Correspondingly, increased cell spreading and significantly greater cell proliferation were also observed in cells grown on the RGDC-coated surfaces. More importantly, osteoblasts on RGDC surfaces showed earlier and significant OC mRNA expression at day 15 compared with controls having the similar expression at day 21. These results provided evidence of the enhanced functions of osteoblasts cultured on the RGDC-modified surfaces, which might be effective in improving osseointegration for dental implants.

ORAL IMPLANTOLOGYTuesday Mar 11 2003 02:01 PMAllen Press x DTPro SystemENHANCED OSTEOBLAST FUNCTIONS ON RGDIMMOBILIZED SURFACERESEARCHHui Huang, DDS, PhDYimin Zhao, DDS, PhDZhiguo Liu, MD, PhDYumei Zhang, DDS, PhDHui Zhang, DDS, MScTao Fu, MScXuanxiang Ma, DDS, PhDKEY WORDSRGDCell functionOsteoblastHui Huang, DDS, PhD, Yimin Zhao, DDS,PhD, Yumei Zhang, DDS, PhD, Hui Zhang,DDS, MSc, and Xuanxiang Ma, DDS, PhD,are at the Stomatological College, FourthMilitary Medical University, Xi'an, China.Address correspondence to Prof. Ma atDepartment of Prosthodontics, StomatologicalCollege, Fourth Military Medical University, 1,Kang Fu Road, Xi'an City 710032, China(e-mail: maxx@fmmu.edu.cn).Zhiguo Liu, MD, PhD, is at the XijingHospital, Fourth Military Medical University,Xi'an, China.Tao Fu, MSc, is at the Institute of MaterialScience and Engineering, Xi'an JiaotongUniversity, Xi'an, China.INTRODUCTIONccessful clinical use of dentalimplants requires thefunctional integration of implantswith surrounding tissues.Regarding biologicalconsequences caused by theimplantation, the type and magnitudeof tissue-implant interactions are determinedby cellular and molecularorim 29_204 Mp_73File # 04emMany methods are currently under investigation to improve the integration ofdental implants to surrounding bones. Among these methods, peptide-modifiedsurfaces have been highlighted as one of the most promising. Our study, aimedat the cellular response to RGD-immobilized surface in vitro, investigated thebasis for designing a bone-active surface coating with RGD-containing peptide.Gold-coated titanium surfaces were used as indicative control surfaces forpeptide immobilization. Using self-assembly monolayer techniques, 2 types ofpeptides, RGDC (Arg-Gly-Asp-Cys) and RDGC (Arg-Asp-Gly-Cys), wereimmobilized onto the gold surfaces. Surface justification was realized throughX-ray photoelectron spectroscopy and Fourier transform infrared spectra.Primary calvarial osteoblasts were cultured on RGDC, RDGC, and non-peptidecoatedsurfaces. Cell attachment, morphology, proliferation, and expression ofosteocalcin (OC) messenger RNA (mRNA) were assessed using cell counting,immunolabeling fluorescence microscopy, and Northern blot assay. Four and 8hours after culture, cell attachment was enhanced on RGDC surfaces.Correspondingly, increased cell spreading and significantly greater cellproliferation were also observed in cells grown on the RGDC-coated surfaces.More importantly, osteoblasts on RGDC surfaces showed earlier and significantOC mRNA expression at day 15 compared with controls having the similarexpression at day 21. These results provided evidence of the enhanced functionsof osteoblasts cultured on the RGDC-modified surfaces, which might be effectivein improving osseointegration for dental implants.events at the tissue-implant interface.The biocompatibility of biomaterials ispredominantly reflected by cell functionsin contact with them. Althoughindicated as being biocompatible,many implant materials in practicaluse do not interact with the host tissueactively. So-called biocompatibility isstill a concept of ''bioinert'' propertiesJournal of Oral Implantology 73ORAL IMPLANTOLOGYTuesday Mar 11 2003 02:01 PMAllen Press x DTPro SystemENHANCED OSTEOBLAST FUNCTIONSthat implies limited adverse disturbanceon the surrounding tissue formation.To accelerate and enhance tissueintegration and to maintain desiredlong-term stability, a type of specific,active, and guided tissue-implantinteraction capable of inducing rapidand strong anchorage has been stronglyexpected.1Surface characteristics of materials,such as topography, chemical composition,and surface energy, play essentialroles in the osteoblast's reactions tobiomaterials. Treatment outcomes indental implantology critically dependon implant surface designs that can optimizethe biologic response during thewhole integration mechanism.2,3 Manyongoing investigations are focused onthe modification of the material surface.Among these, a distinctive trendis to associate bioactive agents withimplant materials to change their osteo-conductibility into osteo-inductibility,and to offer more bioactive featuresto these biocompatible materials.RGD was first identified as the attachmentsite for fibronectin and thenas a ubiquitous adhesive motif in proteinsthroughout the body, includingmany bone extracellular matrix proteinssuch as thrombospondin, osteopontin,type II collagen, osteonectin,and bone sialoprotein. It is recognizedby integrins, a heterodimeric cell membranereceptor family that uses multipleintracellular signaling pathways.4 Itwas reported that mimicking of RGDpeptides adsorbed on substrate couldenhance subsequent cell adhesion.5 Ithas also been shown that RGD-relatedpeptides influence osteoblast mineralization,cytoskeleton reorganization,and migration in vitro.6,7 These investigationsprovide evidence that an implantcoated with tailor-made RGDpeptides is an attractive strategy fordesigning a new generation of proactivedental implants. This type of implantis expected to be capable of elicitingselective responses from targetcell species, performing tissue regeneration,and leaving implant-tissue interactionunder control. This strategy74 Vol. XXIX/No. Two/2003might serve as a useful key technologyfor enhanced osseointegration.In the present study, to investigatethe general cellular response to RGDcoatedsurfaces, the strategy of self-assemblymonolayers was used to covalentlyimmobilize RGD-containingpeptide RGDC and the control peptideRDGC onto gold-coated titanium surfaces.Titanium as a bulk material forbone implants is much better for theformation of osseointegration, but it ismuch easier to immobilize some peptideson the gold surface in a simpleand cost-efficient way.8 Gold coatingwas only as a type of indicative experimentalmodel surface for these in vitroexperiments. To exclude the possibleinterference factors from the varietiesof the substrate type, every titaniumspecimen was coated with gold, andonly effects of the peptides on thosesurfaces were studied. Both X-ray photoelectronspectroscopy (XPS) and Fouriertransform infrared spectra (FTIR)were used to confirm the immobilization.Primary calvarial osteoblasts werecultured on RGDC, RDGC, and non-peptide-coated surfaces. Cell countswere taken to analyze the cell attachmentand proliferation. Fluorescein isothiocyanateconjugated (FITC) antibodiesdetecting the vaculin assemblywere used to visualize cell spreading.Total RNA of the cultured cells wasisolated, and Northern blot assay wasused to analyze the osteocalcin (OC)messenger RNA (mRNA) expression.MATERIALS AND METHODSSubstrate preparationSubstrates used in the present studywere precleaned, 18- and 30-mm,round, commercial, pure titaniumdisks (NMRI, Xi'an, China). The 18-mm disks were prepared for use in cellattachment, spreading, and proliferationexperiments, and the 30-mm diskswere used in detecting OC mRNA expression.Twelve-nanometer-thick goldlayers were prepared by electronicbeam evaporation over 2 nm of titaniumfilm. Peptides RGDC and RDGCorim 29_204 Mp_74File # 04em(Auka, Beijing, China) were dissolvedin distilled, deionized water and ethanol(1:1) and adjusted to a 2 mM concentration.Disks were immersed in thesolution overnight at room temperaturein closed containers with gentlemixing. Gold-coated implants immersedin non-peptide-containing solutionwere used as controls (nonpeptide).Surface identificationXPS (ESCALAB 220i XL, Fisons, Germany)was used to analyze the elementalcomposition and the bondingproperties of the RGD-containing peptidecoating. Power of nonmonochromatizedMgKa (hn 5 1256.6 eV) was100 kw. Change compensation was attainedby using an electron flood gunsetting at 6 eV. The vacuum in the analysischamber was 2.5 3 1027 Pa. C1S,N1S, O1S, and S2P signals were acquiredat constant pass energy of 20 eV. Fittingwas finished with built-in software,and the binding energy of themain lines was determined with eachspectrum being referenced to carbonpollution at 284.7 eV.We performed FTIR on an EP1760spectrometer (Hitachi, Tokyo, Japan) todetect the change of organic chemicalgroups on the substrate before and afterimmobilization. Gain, velocity, aperture,and resolution were set at 1, 1.9,28, and 4 cm21, respectively. Apodizationwas set as the Happ-Genzel mode.Data were collected over a spectralrange of 4000 to 1000 cm21. In both theXPS and FTIR experiments, only thenonpeptide surface was used as thecontrol to RGDC surface.Cell cultureThe phenotype of neonatal rat calvarialosteoblasts, isolated via sequential enzymaticdigestion, was confirmed bypresence of alkaline phosphatase, formationof calcium phosphate mineraldeposits, and production of OC. Then,osteoblasts were cultured in Dulbeccomodified Eagle medium (DMEM, Gibco,Carlsbad, Calif) containing 10% fetalbovine serum (Gibco) in a humidi-ORAL IMPLANTOLOGYTuesday Mar 11 2003 02:01 PMAllen Press x DTPro Systemfied atmosphere containing 5% carbondioxide at 378C. Osteoblasts at 2 to 3passages were used in the experiments.Cell attachmentCells used for attachment assay wereplated at a density of 1 3 104 cells/cm2.After 4, 8, 12, 16, and 20 hours ofgrowth in culture, the disks werewashed with phosphate-buffered saline(PBS) to remove nonadherent cellsand then transferred to new wells. Theattached cells were trypsinized with0.25% trypsin and 0.5 mM EDTA andcounted in a Coulter counter (BeckmanCoulter, Fullerton, Calif). Complete removalof the attached cells from thedisks with trypsin-EDTA was con-firmed by immunofluorescencemicroscopy.Results were expressed as meanswith SDs (n 5 5). Significance wascompared by analysis of variance (ANOVA).Indirect immunofluorescencemicroscopyIndirect immunofluorescence microscopyfor vinculin assembly was performedat the initial stage of cell adhesionto detect the discrepancy in cellspreading. Incubation in PBS containing1% (wt/vol) bovine serum albuminand 10% (vol/vol) normal goat serum(Sigma Chemical Company, St Louis,Mo) was performed to reduce nonspecificbackground staining. After cellseeding for 4 hours on the differentsubstrates, the medium was removedand cells were fixed in methanol at2208C for 6 minutes after a washingstep in PBS. Two-hour incubationswith monoclonal anti-vinculin hVIN-1(Sigma) as primary and FITC anti-ratIgG (Sigma) as secondary antibodieswere performed at 378C with 3 washesin PBS between each incubation.Stained specimens were photographedwith a Zeiss photomicroscope (CarlZeiss, Oberkochen, Germany) aftermounting coverslips.Cell proliferationOsteoblasts in DMEM supplementedwith 10% fetal bovine serum wereseeded onto different substrates at adensity of 0.5 3 104 cells/cm2. The cellswere then cultured under standardconditions for 1, 4, 7, 10, and 15 days.At the end of each of the described periods,osteoblasts were detached withtrypsin-EDTA as mentioned previously.Cell numbers were counted in aCoulter counter, and the results wereexpressed as means with SDs (n 5 5).Significance was compared by ANOVA.Northern blot analysisThe total RNA of the 10-, 15-, and 21-day cultures was extracted using guanidinethiocyanate and phenol/chloroform.9 In particular, cells on differentsurfaces were scraped before usingguanidine thiocyanate to prevent extractedRNA from binding to the peptide-coated disks. After the spectrophotometricmeasurement of the RNA,10 mg of each total RNA was fractionatedon 1% (wt/vol) agarose gels,stained with ethidium bromide, andtransferred to immobilon-N membranes(Millipore, Bedford, Mass) usingthe capillary elution method. Themembranes were hybridized withcomplementary DNA probes that werelabeled with peroxidase using ECLsystems (Amersham Pharmacia, Piscataway,NJ). All the hybridization andwashing steps were performed accordingto the instructions for the ECL systems.Hybridized membranes were exposedto Kodak X-OMAT film. The OCcomplementary DNA was generouslyprovided by Dr J. C. Wang (Departmentof Biochemistry, Fourth MilitaryMedical University). The signal intensityof the autoradiograph was normalizedto the GAPDH housekeepinggene.RESULTSThe XPS wide spectra allowed characterizationof the chemical environmentof the different elements. Peaks foundin peptide-coated surfaces at 285.9,288.5, 287.6, and 286.5 eV correspondedto the C1S signal. Among these . .05).peaks, the major component was locatorim29_204 Mp_75File # 04emHui Huang et aled around 285.9 eV, with what mighthave been some other forms of carbon(C) pollution with oxygen. Nitrogen(N) clearly appeared on the RGD-containingsurfaces, exhibiting a majorcontribution at 399.2 eV, which wastypical of C-NH2 bonding. The N/Cratio was approximately 0.25. This wasmuch lower than the theoretical calculationof 2 from the RGDC moleculeand might be explained by the highlevel of carbon contamination. Sulfur(S) bonding with gold was verifiedwith S2P signal spectra. A total of 80%sulfur was present as a bound thiolatewith the rest in the oxidized form. TheN/S ratio was approximately 7.57,which was similar to the theoreticalvalue of 7. The O1S signal appeared at533.2 eV. Based on these major data acquiredfrom XPS spectra, it can be con-firmed that the RGDC peptide hadbeen firmly immobilized on the substratesurface.In FTIR spectra, characterizedbroad amide I stretch peak found inthe 1675- to 1640-cm21 range differentiatedthe RGDC immobilized surfacesfrom controls. Moreover, on the RGDCsurface, there were also some sharpand symmetric absorbance peaksfound at 1735, 1281, and 964 cm21.These peaks can be attributed to theester group (C 5 O) and CH2-CH2-COOH as well as the COOH group inAsp. All these findings confirmed thesuccessful grafting of RGD-containingpeptide onto the substrates from anotheraspect.Cell attachment kinetic spotted at4, 8, 12, 16, and 20-hour time points isshown in Figure 1 as changes of celldensity (cell numbers/surface area insquare centimeters). At 4 and 8 hours,obviously enhanced cell attachmentwas observed on RGDC-coated surfaces(P , .01, P , .05), whereas no significantdifference could be found betweenRGDC and uncoated surfaces atmean time (P . .05). After 12 hours,cell attachment appeared similar on allsurfaces and kept stable to the end (PAssembly of vinculin during theJournal of Oral Implantology 75ORAL IMPLANTOLOGYTuesday Mar 11 2003 02:01 PMAllen Press x DTPro SystemENHANCED OSTEOBLAST FUNCTIONS76 Vol. XXIX/No. Two/2003orim 29_204 Mp_76File # 04emFIGURES 1-4. FIGURE 1. Osteoblast attachment on 3 different surfaces. Rat calvarial osteoblasts in Dulbeccomodified Eaglemedium (DMEM)supplemented with 10% fetal bovine serum were seeded on RGDC, RDGC, and nonpeptide surfaces at a density of 1 3 104 cells/cm2 ina humidified atmosphere containing 5% carbon dioxide at 378C. Cells were counted at 4, 8, 12, 16, and 20 hours. Values were mean 6SEM; n 5 5; Significance was compared with analysis of variance (ANOVA). **P , .01, *P , .05.FIGURE 2. Spreading pattern of osteoblasts on 3 different surfaces after 4-hour culture. Cells were labeled with monoclonal vinculinantibody hVIN-1 as primary and fluorescein isothiocyanate conjugated-labeled anti-rat IgG as secondary antibody. (a) On the RGDCsurface, most osteoblasts displayed fattened pattern with nuclei clearly bounded by cytoplasm. In peripheral filopodia-like extensions,scattered vinculin-containing focal adhesions led to intimate contact between the surface and the cell. The cell spreading pattern wasmature. (b) Cells on the RDGC surfaces. (c) Cells on nonpeptide surfaces. Most cells on both the RDGC and nonpeptide surfaces displayeda small and solid shape with obscure nuclei outline. Multiple rodlike projections evenly protruded outward, with vinculin spots in apunctuate circular pattern. The cell spreading pattern was immature.FIGURE 3. Osteoblast proliferation on 3 different surfaces. Rat calvarial osteoblasts in DMEM supplemented with 10% fetal bovine serumwere seeded on RGDC, RDGC, and nonpeptide surfaces at a density of 0.5 3 104 cells/cm2 in a humidified atmosphere containing 5%carbon dioxide at 378C. Cells were counted at 1, 4, 7, 10, and 15 days. Values were mean 6 SEM; n 5 5; Significance was compared withANOVA. *P , .05.FIGURE 4. Northern hybridization analysis of osteocalcin (OC) gene expression of osteoblasts cultured on RGDC, RDGC, and nonpeptidesurfaces. The OC messenger RNA (mRNA) levels of rat calvarial osteoblasts cultured on RGDC, RDGC, and nonpeptide (gold) surfacesfor 10, 15, and 21 days were determined by Northern blot assay (n 5 6) as described in the ''Materials and Methods'' section. Equalsample loading was confirmed by normalization to the density of the GAPDH band. Cells on RGDC surfaces displayed significant earlyOC mRNA expression at 15 days.ORAL IMPLANTOLOGYTuesday Mar 11 2003 02:01 PMAllen Press x DTPro Systeminitial stage of cell-surface interactionindicates great discrepancy of the osteoblastspreading pattern on differentsurfaces. With methanol fixation, osteoblastsexhibited a diffuse cytoplasmicimmunofluorescence background staining.After 4-hour culture, cells on theRGDC surfaces demonstrated distinguishedflattened pattern with the nucleiclearly bounded by cytoplasm(Figure2a). Most nuclei in these cells werelightly stained, peripherally situated,and ovally shaped, indicating a relativelymature appearance. Vinculincontainingfocal adhesions (FAs) wereclustered in peripheral filopodia-likecell extensions in an apparent attemptto spread their cell bodies over the substrates.On the contrary, cells on theRDGC and nonpeptide surfaces displayedcontracted and solid patternswithout apparent nuclei contour. Beingcovered under undiffused cytoplasm,nuclei shape was still unable to see.Cell outlines were generally small andspheroid, with multiple rodlike projectionsevenly protruding outward. Theends of these pseudopodia were enlargedin the vinculin spot in a punctatecircular pattern inside their margins.Cells on RDGC and nonpeptidesurfaces displayed similar immaturecell spreading patterns (Figure 2b andc).Figure 3 illustrates the change ofcell quantities at 1, 4, 7, 10, and 15days. Osteoblast proliferation onRGDC surfaces increased significantlyat 4 and 7 days compared with thoseon RDGC and nonpeptide surfaces (P, .05). Cell growth at 1, 10, and 15days showed no significant differenceon different surfaces. Meanwhile, osteoblastproliferation on RDGC anduncoated surfaces did not differ fromeach other in the whole culture process(P . .05).Northern blot analysis showed anearlier OC mRNA expression onRGDC surfaces. At day 15, intense OCmRNA bands appeared and no signaloccurred in the other 2 control groups(Figure 4). On day 21, OC mRNA expressionstayed the same on RGDCsurfaces but began to show distinct expressionin the other control groups. Atthis time, the mRNA levels in allgroups were similar. Equal sampleloading was confirmed by the unchangeddensity of the housekeepinggene GAPDH.DISCUSSIONSelf-assembly monolayer is a techniqueby which molecules under thermodynamicequilibrium conditions canspontaneously align on a surface into2-dimensional, quasi-crystalline domainsthat are stable and ordered.10,11When gold and silver were exposed tosolution or vapors of an alkanethiol(RSH), these metals were able to bondwith sulfur to form RSH self-assemblymonolayer. The gold-sulfur bond wasbelieved to be a covalent thiolate bond.8In this study, titanium substrates wereevaporated with a thin layer of titaniumto promote the bonding with thethin film of gold. Gold surfaces are nottoxic to living cells and are consideredbiocompatible with conditions used forcell culture.12 This rationalized our useof gold as an experimental model surface.Of course, the most ideal approachis to immobilize the peptideson a titanium surface, but this requiresmuch more complicated chemical approachesand instruments. The reactionis much more expensive and cannotbe completed in the normal roomtemperature or in an efficient manner.Although, compared with titanium,gold does not possess the good biocompatibilityfor bone integration, thisdisadvantage did not affect our use ofgold as in vitro model surfaces to evaluatethe biological impact of peptideson osteoblasts. Peptides (RGDC,RDGC) with terminal cysteine (C)groups, which contain free thiol residues,were exposed to gold surfaces.The thiol groups, serving as anchors,covalently bonded with gold, allowingthe oligopeptides to attach onto surfacesto form monolayers. Functionalgroup RGD and RDG formed a ligandlayer to cell membrane receptors. TheXPS and FTIR results confirmed theorim 29_204 Mp_77File # 04emHui Huang et alimmobilization of RGD peptide ontoexperimental surfaces. Meanwhile,some of the symmetric peak shapes ofthe FTIR spectra also suggested thatthe molecular organization on the surfacesis highly ordered. Our procedurecorresponded with some previouslydescribed similar surface modificationmethods using RGD grafts,13 and thesurface identification results verifiedthe availability of this procedure for invitro experiments.Endosseous integration can be brokeninto 3 distinct bony healing phases:osteoconduction, de novo bone formation,and bone remodeling. Osteoconductionrelies on the migration andcolonization of differentiating osteogeniccells on the implant surface. Denovo bone formation primarily requiresthe recruitment of potentiallyosteogenic cells to the site of futurematrix formation.14 When implants aresurgically placed in bone bed, they areexposed to blood and extracellular fluids.Various proteins are adsorbed ontoimplant surfaces and then mediate differentcell types to attach onto thesesurfaces. As anchorage-dependentcells, osteoblasts, the major osteogeniccells, can only begin their cell functionsafter fully attaching and spreading.15,16So cell attachment and spreading areconsidered important indicators of biocompatibilityand bioactivity of implantmaterials.In the present study, cell attachmenton RGDC surfaces was signifi-cantly enhanced within 8 hours afterplating. At the 4-hour point, attachmentrate (attachment density/seedingdensity) on the RGDC surface was approximately60% rather than the 35%that appears on the RDGC and nonpeptidesurfaces. At the 8-hour point,the attachment rate was 85% on RGDCsurfaces rather than the 70% seen onRDGC and nonpeptide surfaces. After12 hours, most cells in different treatmentgroups finished the attachmentprocess and entered the G1 phase, sono difference could be found amongthe 3 groups. Generally, these resultssuggested a promotional effect of RGDJournal of Oral Implantology 77ORAL IMPLANTOLOGYTuesday Mar 11 2003 02:01 PMAllen Press x DTPro SystemENHANCED OSTEOBLAST FUNCTIONSpeptide on osteoblast attachment, especiallyat the initial phase of tissueimplantcontact.Cell attachment is a complex processthat includes several steps. Theinitial attachment is followed by theformation of FA sites and subsequentcell spreading. Focal adhesions are thelargest and tightest matrix adhesions,which are highly dynamic and heterogeneousstructures and represent amorphologically prominent associationbetween integrins and the cytoskeleton.17 Vinculin is one of many prominentcytoplasmic constituents of FAsand is believed to play important rolesin stabilizing cell adhesion and regulatingcell shape, morphology, and mobility.18 In our experiments, the resultsof the cell attachment assay were paralleledwith the results of cell spreadingmicroscopy. Four hours after seeding,most osteoblasts on RGDC surfacesdisplayed flattened, maturely spreadmorphology with scattered FA, leadingto intimate contact between the surfaceand the cell. On RDGC and nonpeptidesurfaces, diffused FA could not befound and cells were not fully spread.This discrepancy in spreading dynamicsprovides more evidence of enhancedcell attachment on the RGDCsurfaces.The cell proliferation on the RGDCsurfaces was increased at the initialstage of cell growth. After 10 days, cellproliferation on different surfaces becamesimilar. This could be attributedto contact inhibition when most of thesurfaces were occupied by cells. Enhancedproliferation of osteoblastsmight advance extracellular matrix secretion,thus favoring early stabilizationof implant in supporting tissue.This will undoubtedly benefit clinicaloutcomes of the implantation.Osteocalcin is the marker of differentiatedosteoblasts. The secretion ofOC is believed to be related to the coordinationof bone resorption and newbone deposition.19 It signals the beginningof the bone remodeling by promotingthe maturation of the osteoclastand appears to be the only protein as-78 Vol. XXIX/No. Two/2003sociated with the mineralizationfront.20 The earlier expression of OC onRGDC surfaces provided evidence ofan earlier differentiation of the osteoblastson those surfaces. This enhancedosteoblast turnover may be significantlyuseful in accelerating tissue integrationand inducing early anchorage.The difference between RGDC andRDGC is the order of constitutive peptides.Although both of them had identicalchemical composition and couldbe immobilized onto gold surfacesthrough Cys residues, their effects onthe osteoblast functions were totallydifferent in our experiments. This suggesteda structure-dependent effect ofRGD peptide rather than a composition-dependent effect. These ligandbindingproperties are important foradhesion receptors such as integrins.Integrins link extracellular matrix proteinson the extracellular face of the cellmembrane to cytoskeletal proteins andactin filaments on the cytoplasmic face.The short cytoplasmic domains of thea and b integrin subunits appear tofunction by coupling with cytoplasmicproteins that nucleate the formation oflarge protein complexes containingboth cytoskeletal and catalytic signalingproteins, thus transmitting biochemicalsignals and mechanical forceacross the plasma membrane.4 The cytoskeletallinkages enable integrins tomediate cell adhesion and regulate cellshape and gene expression.21 Surfacesmodified with RGD may trigger the integrinactivation, advance or intensifysubsequent signal transduction, and alterthe behavior of osteoblasts. Thismight be the reason for the enhancedcell function on RGD surfaces in ourexperiments. Although no clear mechanismconcerning the regulation of celldifferentiation and proliferation by integrinshas been identified, our resultsindicate the possible influence of RGDon those aspects.Regarding the practicality of thissurface modification model in in vivostudies, some of our pilot experimentswere not in favor of this model, because,due to loose bonding betweenorim 29_204 Mp_78File # 04emgold film and titanium substrate, mostof the coating shed as debris whenspecimens are prepared. The ultimategoal of this research is to firmly immobilizethose bioactive peptides on titaniumor titanium alloy surfaces andexplore their in vivo effects. Fortunately,there appear to be some applicableways to achieve this, such as the covalentcoupling technique. Various methodsare currently under investigation.22Before any practical application of theproactive modification strategy can beachieved, a valid and cost-effective approachfor persistent and effective deliveryof bioactive agents in implanttissueinterface must be found.CONCLUSIONWe used self-assembly monolayer techniquesto immobilize RGD-containingpeptides on substrate surfaces. In vitrostudies have indicated that the immobilizedadhesive motif RGD not onlypromotes attachment and spreading ofosteoblasts but also enhances the proliferationand differentiation of thosecells on the test surfaces. Immobilizationof RGD peptides onto implant surfacesmight be an effective strategy forthe development of new proactive dentalimplants.ACKNOWLEDGMENTThis work was supported by the NationalScience Foundation of China.REFERENCES1. Hanns PJ. Prosthesis-bone interface.J Biomed Mater Res (Appl Biomater).1998;43:350-355.2. Baier RE. Surface properties in-fluencing biological adhesion. In: ManlyRS, ed. Adhesion in Biological Systems.New York, NY: Academic; 1970:15-48.3. Brunette DM. The effects of implantsurface topography on the behaviorof cells. Int J Oral Maxillofac Implants.1988;3:231-246.4. Edwin AC, Joan SB. Integrinsand signal transduction pathways: theroad taken. Science. 1995;268:233-239.5. Puleo D, Bizios R. RGDS tetrapeptidebinds to osteoblasts and inhib-ORAL IMPLANTOLOGYTuesday Mar 11 2003 02:01 PMAllen Press x DTPro Systemits fibronectin-mediated adhesion.Bone. 1991;12:271-276.6. Dee KC, Andersen TT, Bizios R.Osteoblast population migration characteristicson substrates modified withimmobilized adhesive peptides. Biomaterials.1999;20:221-227.7. Dee KC, David C, Rueger A,Thomas T. Conditions which promotemineralization at the bone-implant interface:a model in vitro study. Biomaterials.1996;17:209-215.8. Mrksich M, Chen CS, Xia Y,Dike LE, Whitesides GM. Controllingcell attachment on contoured surfaceswith self-assembled monolayers of alkanethiolateson gold. Proc Natl Acad US A. 1996;93:10775-10778.9. Chomczynski P, Sacchi N. Single-step method of RNA isolation byacid guanidinium thiocyanate-phenolchloroformextraction. Ann Biochem.1987;162:156-159.10. Whitesides GM, Mathias JP,Seto CT. Molecular self-assembly andnanochemistry: a chemical strategy forthe synthesis of nano-structures. Science.1991;254:1312-1319.11. Mrksich M, Whitesides GM.Using self-assembled monolayers tounderstand the interactions of manmadesurfaces with proteins and cells.Ann Rev Biophys Biomol Struct. 1996;25:55-78.12. Mrksich M, Dike LE, Tien J,Ingber DE, Whitesides GM. Using microcontactprinting to pattern the attachmentof mammalian cells to selfassembledmonolayers of alkanethiolateson transparent films of gold andsilver. Exp Cell Res. 1997;235:305-313.13. Zhang S, Lin Y, Michael A. Biologicalsurface engineering: a simplesystem for cell pattern formation. Biomaterials.1998;115:1213-1220.14. Davies JE. Mechanisms of endosseousintegration. Int J Prosthodon.1998;11:391-401.15. Puleo DA, Holleran LH, DoremusRH, Bizios R. Osteoblast responsesto orthopedic implant materials invitro. J Biomed Mater Res. 1991;25:711-723.16. Anselme K. Osteoblast adhesionon biomaterials. Biomaterials. 2000;21:667-681.17. Burridge K, Magdalena CW.Focal adhesions, contractility, and sigorim29_204 Mp_79File # 04emHui Huang et alnaling. Annu Rev Cell Dev Biol. 1996;12:463-519.18. Bubeck P, Pistor S, Wehland J,Jockusch BM. Ligand recruitment byvinculin domains in transfected cells. JCell Sci. 1997;110:1361-1371.19. Heersche JNM, Reimers MS,Wrana JL, Waye MMY, Gupta AK.Changes in expression of alpha I typeI collagen and osteocalcin mRNA inosteoblasts and odontoblasts at differentstages of maturity as shown by insitu hybridization. Proc Finn Dent Soc.1992;88:173-182.20. Nefussi JR, Brami G,ModrowskiD, Oboeuf M, Forest N. Sequentialexpression of bone matrix proteinsduring rat calvaria osteoblast differentiationand bone nodule formation invitro. J Histochem Cytochem. 1997;45:493-503.21. Choquet D, Felsenfed DP,Sheetz MP, et al. Extracellular matrixrigidity causes strengthening of integrin-cytoskeleton linkages. Cell. 1997;88:39-48.22. Dee KC, Bizios R. Mini-review:proactive biomaterials and bone tissueengineering. Biotechnol Bioeng. 1996;50:438-442.Journal of Oral Implantology 79