Attachment of Human Marrow Stromal Cells to Titanium Surfaces

ORAL IMPLANTOLOGYTuesday Mar 11 2003 01:58 PMAllen Press x DTPro SystemATTACHMENT OF HUMAN MARROW STROMALCELLS TO TITANIUM SURFACESRESEARCHMamle Mante, DMD, MSBina Daniels, BSEleanor Golden, BSDavid Diefenderfer, VMD, PhDGwen Reilly, PhDPhoebe S. Leboy, PhDKEY WORDSTitaniumMesenchymal stromal cellsCell attachmentCell proliferationSurface rougheningNitric acid passivationMamle Mante, DMD, MS, is with theDepartment of Restorative Dentistry at theSchool of Dental Medicine at the University ofPennsylvania, 4001 Spruce Street, Philadelphia,PA 19104. Correspondence should be addressedto Dr Mante.Bina Daniels, BS; Eleanor Golden, BS;Gwen Reilly, PhD; and Phoebe S. Leboy,PhD, are with the Department of Biochemistryin the School of Dental Medicine at theUniversity of Pennsylvania in Philadelphia, PA19104.David Diefenderfer, VMD, PhD, is with theSchool of Veterinary Medicine at theUniversity of Pennsylvania in Philadelphia, PA19104.66 Vol. XXIX/No. Two/2003INTRODUCTIONntal implants provide ameans of anchoring variousoral prostheses in theThe attachment of human bone marrow stromal cells to titanium alloy (Ti6Al4V)surfaces was investigated. Titanium disks were polished and modified by surfaceroughening and by passivation in nitric acid. Cell attachment to titanium surfacesand tissue culture plastic (TCP) was determined by tetrazolium bromide (MTT)assay at 2, 6, 24, and 48 hours after seeding. Cell proliferation was determinedby thymidine incorporation. Attachment on titanium surfaces was 75.6% to 94.9%of attachment on TCP control. The difference between cell attachment on the TCPcompared with smooth or rough titanium was statistically significant (P , .05).However, no statistically significant difference was found between attachment toTCP and passivated titanium. Cell proliferation on titanium surfaces after 24hours was approximately 70% of proliferation on TCP. There was a statisticallysignificant difference (P , .05) between proliferation on tissue culture andsmooth and passivated titanium but not on rough titanium. These results indicatethat titanium provides a surface that is conducive to cell attachment and thatpassivating titanium improves cell attachment, approaching levels seen with TCP,a surface specifically developed to enhance cell attachment. Increasing surfaceroughness results in improved cell proliferation on titanium.orim 29_202 Mp_66File # 02emmaterials, although bioactive in natureand able to stimulate osteogenesis, areof poor mechanical strength, whichprecludes their use as load-bearing implants.Titanium is not an osteoinductivematerial but has superior mechanicalproperties and the ability to''bond'' to bone or osseointegrate, in aprocess similar to fracture healing,which has contributed to its clinicalsuccess.2-4 When titanium is exposedto atmospheric conditions, or in vivofluids, a thin tenacious film of titaniumoxide spontaneously forms on the implantsurface. This film renders themandible and maxilla.The success of titaniumand titanium alloy rootform implants has resulted in theirroutine use in dentistry today. Implantsuccess depends on the nature of thesurrounding bone. The ideal bone conditionis found where there is goodcancellous bone surrounded by corticalbone of adequate thickness on bothsides.1Most polymer and ceramic implantORAL IMPLANTOLOGYTuesday Mar 11 2003 01:58 PMAllen Press x DTPro Systemotherwise reactive material bioinertand biocompatible by preventing itfrom further reactions, only minimalreaction products leach out into surroundingfluids, and toxicity to surroundingcells is very low.5 An increasein the thickness of the oxide film resultsin improved bone formation atthe implant surface6-8; implant manufacturersroutinely passivate titaniumimplants by immersion in nitric acid toincrease the thickness of the oxide layer.Studies have shown that the surfaceroughness of the implant influencesnew bone formation at the implantsurface regardless of implant composition,with rougher surfaces (at the micronlevel) showing better bonegrowth.6-8 In vitro studies have demonstratedthat surface roughness influencescell attachment, proliferation, differentiation,and matrix production.9-12Rougher implant surfaces provide anincreased surface area for cell attachmentand subsequent bone formation.11When cells were cultured on titanium,there was an increase in the productionpromoters of osteoblast differentiationand bone growth, with increasing titaniumsurface roughness.12-15Marrow stromal cells, also calledmesenchymal stem cells, are precursorcells that can differentiate into bone,cartilage, and fat cells. Because of theirability to differentiate into osteoblasts(bone-forming cells), they play an importantrole in normal bone16,17 repairand remodeling, as well as in osteogenesisaround implant materials.18-20Several investigators have studiedthe effects of implant surface treatmenton the attachment of osteoblast or osteoblast-like cells7,13,14,21-24 of vertebrateand human origin. These are cells thatare already committed to osteogenesis.There is less extensive literature on theattachment patterns of marrow stromalcells (MSCs)11,25 of nonhuman originand the behavior of specifically humanMSCs on implant materials.10,19Greater knowledge of the attachmentpatterns of human MSCs is im-Smooth, rough, and passivated Ti surofProstaglandin E2 and TGF-b1, known St. Joseph, Mich), with 600-grit silicon faces without cells were examined usatedcells at the interface. The fractionwas collected, washed once withaMEM media (Gibco), and suspendedin aMEM with 15% fetal calf serumand antibiotics. Primary cultures ofthese cells were established at 5 3 105cells/cm2, and nonadherent cells wereremoved after 3 days. First- and sec-The aim of this project was to ex- ond-passage cells were used for the examinethe attachment of human mar- periment and were plated on tissuerow stromal cells on a titanium alloy culture plastic (TCP) or titanium alloy(Ti6Al4V) routinely used for dental im- (Ti6Al4V) in 24 well plates at a densityof 10 000 cells/well.orim 29_202 Mp_67File # 02emMamle Mante et alThe cells were seeded in a smallvolume of media on the respective surfacesand allowed to attach for 1 hourprior to the addition of aMEM mediumcontaining 15% fetal calf serum(Atlanta Biologicals, Norcross, Ga).Cells were then harvested at 2, 6, 24,and 48 hours after seeding. The titaniumdisks with attached cells were removedfrom the plates they were culturedin and placed in clean wells priorto assay.Scanning electron microscopying a JEOL T300 scanning electron microscope(JEOL, Tokyo, Japan) at an acceleratedvoltage of 20 kV. Cells attachedto Ti and tissue culture surfaceswere prepared for microscopy by fixationin 2% formaldehyde followed bydehydration in graded alcohol. Thesamples were then carbon-sputtercoatedand viewed by scanning electronmicroscopy.Tetrazolium bromide (MTT) assay forviable cell numberTitanium samples were transferred toclean wells after an incubation period.Cell cultures were incubated in 0.5mg/mL solution of tetrazolium bromidein Hanks buffered saline solution(HBSS, GIBCO, Grand Island, NY) dilutedin 5 parts of clear media for 1hour. After the incubation period, themedia was aspirated, and 200 mL of dimethylsulfoxide (DMSO) was placedin each well. Next, 190 mL of the solutionwas transferred into a microplateJournal of Oral Implantology 67portant in order to understand the processesthat influence their proliferation,maturation, and differentiation at theimplant surface at the cellular as wellas the molecular level. Modification ofimplant surfaces to optimize the osteogeniccapability of MSCs could leadto greater clinical implant success.plants and evaluate the effects of increasingsurface roughness of titaniumsamples or passivation in nitric acid oncell attachment.MATERIALS AND METHODSPreparation of titanium samplesTitanium alloy (Ti6Al4V) rods, 12.5 mmin diameter, were sectioned into discs2 mm in thickness. To obtain differencesin surface characteristics, the titaniumdiscs were subjected to 1 of 3 differentsurface treatments. The firstgroup, designated smooth Ti, werepolished on a LECO polisher (LECO,carbide paper. The average particle sizeof the 600 grit used for final polish was15.3 mm. The second group, designatedrough Ti, was polished to final polishwith 240-grit paper, average particlesize 58.5 mm. The third group, passivatedTi, was polished to 600 grit andthen passivated in 30% nitric acid for60 minutes. The discs were then ultrasonicallycleaned in acetone for 15 minutesand sterilized by immersion in70% ethanol for 1 hour.Cell culturePrimary human MSCs were obtainedfrom human bone marrow followinghip replacement surgery. The cellswere obtained from a pool of randomlyselected patients consisting of 5women and 2 men ranging in age from24 to 48 years. The samples were layeredon Ficoll-Paque (Amersham PharmaciaBiotech, Piscataway, NJ) andcentrifuged 30 minutes at 1900 relativecentrifugal force to concentrate nucle-FIGURES 1-3. FIGURE 1. (a) SEM micrograph of titanium surface showing the surface morphology of a titanium disc polished to 600 griton silicon carbide paper (smooth and passivated titanium). (b) SEM micrograph of titanium surface showing the surface morphology ofa titanium disc polished to 240 grit on silicon carbide paper (rough titanium). FIGURE 2. (a) SEM micrograph (original magnification3500) of titanium surface with attached marrow stromal cells on smooth titanium. (a) SEM micrograph (original magnification 35000)of titanium surface with attached marrow stromal cells on smooth titanium. FIGURE 3. (a) SEM micrograph (original magnification 3500)of titanium surface with attached marrow stromal cells on rough titanium. (b) SEM micrograph (original magnification 35000) of titaniumsurface with attached marrow stromal cells on rough titanium. reg;orim 29_202 Mp_68File # 02emTABLEPaired t-test human MSC cellnumber (optical density at 570 nm)First 48 hours in culture P value.04.003.14.92.59.33control vs smoothcontrol vs roughcontrol vs passivatedsmooth vs roughsmooth vs passivatedrough vs passivatedDISCUSSIONIn this study, we used in vitro cell cultureto characterize the attachment patternsof human marrow stromal cellson a titanium alloy (Ti6Al4V) routinelyused for dental and orthopedic implants.The surface of the titanium wasmodified by increasing surface roughnessand by passivation in nitric acidto increase the thickness of the surfaceoxide layer.Results of the MTT assay show thatduring the initial 24 hours after seeding,cell attachment to the 3 titaniumsurfaces ranged from 75.6% to 94.9%of the attachment to the TCP control.Increasing surface roughness resultedin a higher cell count at 6 and 24 hours;however, the differences at these 2 timepoints and over the initial 24-hour periodwere not statistically significantfrom attachment to smooth Ti. The increasein attachment of cells to roughertitanium could be a direct effect of thetitanium surface or an indirect effect ofaltered levels of adhesion proteins. Fibronectinand vitronectin are extracellularproteins that promote cell attachmentand are present in the serumused in tissue culture. It has beenshown that the serum proteins albumin,fibronectin, and vitronectin binddifferently to various implant and bonesubstitute materials.10,19 Changes in ti-ORAL IMPLANTOLOGYTuesday Mar 11 2003 01:58 PMAllen Press x DTPro SystemATTACHMENT OF MARROW STROMAL CELLS TO TITANIUMand the absorbance was measured at570 nm. Each experiment was done intriplicate.Thymidine incorporation assay forproliferationThe cell cultures were treated with 3Hthymidine (1 mCi/mL) for 18 hours beforeharvesting. They were thenwashed in HBSS and lifted in 0.25%trypsin in 1 mM EDTA for 15 minutes.The cell suspensions were pipettedonto glass fiber filters prewetted with5% trichloroacetic acid (TCA). The filterswere allowed to dry and thenwashed for 5 minutes with ice cold10% TCA followed by four 5-minutewashes with ice cold 5% TCA and afinal wash with ice cold 100% ethanol.The filters were allowed to dry andthen transferred to vials, covered with3 mL of scintillation fluid (ICN), andthen counted. The average decay perminute (dpm) on the different titaniumsurfaces was expressed relative to controlsamples cultured on the TCP.RESULTSSEM of titanium surfacesSmooth and passivated titanium samplesshowed shallow residual grooves3-5 mm apart, whereas rough titaniumspecimens had deeper irregulargrooves, which were approximately 1m apart (Figure 1). After 24 hours, cellsattached to plastic and titanium hadnot yet reached confluence and exhibiteda typical fibroblastic phenotypewith centrally placed nuclei and elongatedprocesses. There was no noticeabledifference in cell morphology onthe different titanium surfaces after 24hours. Cells attached to a smooth titaniumsurface appeared to spread outacross the residual polishing grooves68 Vol. XXIX/No. Two/2003(Figure 2). On rough titanium, the attachedcells were attached in the samedirection as the deeper and wider residualpolishing grooves (Figure 3).The attachment of human MSC totitanium was compared with the attachmentto TCP (control) at 2, 6, 24,and 48 hours (Figure 4). The opticaldensity of the MTT suspension read at570 nm was proportional to the viablecell number. Although all surfacesshowed an increase in cell numberwith time, the number of cells attachedto the 3 titanium surfaces during thefirst 24 hours was approximately 80%of the number attached to the TCP control(Figure 5). When cell numberswere compared over the entire 48-hourperiod, by combining data from the 2,6, 24, and 48 hour groups, cell numbersfrom TCP were significantly higherthan smooth and rough titanium (P ,.05). However, the smaller differencebetween passivated titanium and TCP(P 5 .14). was not statistically signifi-cant (Table). There was no statisticallysignificant difference between cell attachmentto the 3 different titaniumsurfaces at any of the time points investigatedor over the 48-hour period.Cell proliferationThe proliferation of human MSC on Tisurface was examined by measuringthe incorporation of labeled thymidineduring a 24-hour interval betweendays 1 and 2. The relative rates of proliferationon smooth, rough, and passivatedtitanium when compared withTCP were 68%, 74%, and 71%, respectively(Figure 6). The differences werestatistically significant for smooth andpassivated titanium, but not for roughtitanium.ORAL IMPLANTOLOGYTuesday Mar 11 2003 01:58 PMAllen Press x DTPro Systemorim 29_202 Mp_69File # 02emMamle Mante et alJournal of Oral Implantology 69In this experiment, marrow stromalcells grown on smooth and roughtitanium formed a monolayer of cellsof typical fibroblastic appearance withcell processes. Cells cultured on roughtitanium appeared to spread out in thesame direction as the residual groovesresulting from surface roughening.Cells on smooth titanium were orientatedin different directions. Differences26in the amount and type of adhe-ORAL IMPLANTOLOGYTuesday Mar 11 2003 01:58 PMAllen Press x DTPro SystemATTACHMENT OF MARROW STROMAL CELLS TO TITANIUMtanium surface roughness have beenshown to alter the adsorption of theseproteins,10 with rougher surfaces bindinga higher amount of total serumprotein and fibronectin than smoothertitanium surfaces.increase in the thickness of the oxidelayer from passivation would presumablydecrease the potential for toxicity,and therefore result in improved cellattachment. Increasing the thickness ofthe titanium oxide layer results in animprovement in bone formation at theimplant surface in vivo.6,2770 Vol. XXIX/No. Two/2003FIGURES 4-6. FIGURE 4. Human MSC change in number of human MSC attached to smooth rough and passivated titanium and tissueculture plastic (TCP) from 2 to 48 hours. The number of cells attached is expressed as the optical density of the MTT-cell suspension.FIGURE 5. Human MSC cell attachment expressed as a percentage of cells attached to TCP at 2, 6, and 24 hours. There was no statisticallysignificant difference with TCP at each of the time points. FIGURE 6. Human MSC proliferation expressed as relative thymidine incorporationafter 24 hours in culture. The asterisk (*) indicates a statistically significant difference (P , .05) with TCP control.Cell attachment to passivated titaniumwas not statistically significantfrom the TCP control. There are a fewreports in the literature of the possibletoxic effects of Al ions leaching out ofTi6Al4V alloys into the solution.26 An sion proteins adsorbed onto the deeper proteins, and consequently influenceorim 29_202 Mp_70File # 02emfluids containing proteins and polysaccharides,as well as cells of mesenchymalorigin. Adhesion proteins such asfibronectin and vitronectin, present intissue fluids, as well as integrins expressedby cells, adhere to the implantsurface, leading to complex and dynamicbiochemical reactions occurringat the implant surface.28-32 The surfacecharacteristics of the implant surfaceaffect the adsorption19,20 of adhesioncellular attachment, proliferation, anddifferentiation around the implant.Using different methods and awider range of surface roughnesses,Deligianni et al10 have reported that increasingthe surface roughness ofTi6Al4V results in a significant increasein the attachment of human MSC at 0.5and rougher residual grooves could explainthis difference in orientation ofattaching cells. There was no noticeabledifference in the overall cell shape betweencells cultured on the 2 surfaces.Protein coating presumably occursin vivo after surgical placement; theimplant surface is exposed to tissueORAL IMPLANTOLOGYTuesday Mar 11 2003 01:58 PMAllen Press x DTPro Systemand 2 hours. In the present study, therewas an increase in cell attachment at 6and 24 hours over a narrower range ofsurface roughness of titanium, indicatingthat even smaller increases in surfaceroughness improve cell attachment.There was a statistically significantdifference between cell attachment toTCP and attachment to smooth andrough titanium, indicating that thesesurfaces were less favorable for cell attachment.This is not surprising giventhat TCP is an idealized cell culturesubstrate. However, cell attachment totitanium was approximately 80% of attachmentto TCP, and thus the differencein attachment is not very large.This confirms the biocompatibility oftitanium. Presumably, high cell attachmentto titanium will result in betterbone formation on the implant surface,fixation or osseointegration of the implant,and a reduced chance of the formationof a fibrous capsule around theimplant, which results in looseningand clinical failure. Due to its high mechanicalstrength, a well-anchored titaniumimplant is able to withstandthe high forces encountered in dentaland orthopedic applications.All 3 titanium surfaces showedlower cell proliferation rates comparedwith TCP. As well as increasing attachment,compared with smooth nonpassivatedsurfaces, increasing surfaceroughness also resulted in an increasein cell proliferation when comparedwith all types of smooth titanium.Clinical reports indicate that increasingimplant surface roughness results inimproved bone formation and osseointegrationof titanium implants in laboratoryanimals as well as humans.33-35The results of this study suggest thatthis may be because proliferation of osteoprogenitorcells are increased onrougher surfaces, which would resultin the formation of more osteoblastsand therefore increased bone formation.CONCLUSIONTitanium provides a surface that isconducive to the attachment of humanmarrow stromal cells. The rate of cellproliferation on titanium was lowerthan the rate on TCP; however, increasingsurface roughness of titanium improvedthe proliferation rate, approachingthe rate of TCP.ACKNOWLEDGMENTThis study was funded by NIH grantR01DE13800-1.REFERENCES1. Ellingsen JE. Surface configurationsof dental implants. Periodontology.2000;17:36-46.2. Branemark PI, Adell R, Breine U,Hansson BO, Lindstrom J, Ohlsson A.Intra-osseous anchorage of dentalprostheses. I. Experimental studies.Scand J Plast Reconstr Surg. 1969;3:81-100.3. Listgarten MA, Lang NP,Schroeder HE, Schroeder A. Periodontaltissues and their counterpartsaround endosseous implants. Clin OralImplants Res. 1991;2:1-19.4. Schroeder A, van der Zypen E,Stich H, Sutter F. The reactions of bone,connective tissue, and epithelium toendosteal implants with titaniumsprayedsurfaces. J Maxillofac Surg.1981;9:15-25.5. Williams D. Titanium and Its Alloys.Boca Raton, FL: CRC Press; 1981.6. Larsson C, Thomsen P, AronssonBO, et al. Bone response to surfacemodifiedtitanium implants: studies onthe early tissue response to machinedand electropolished implants with differentoxide thicknesses. Biomaterials.1996;17:605-616.7. Bigerelle M, Anselme K, Noel B,Ruderman I, Hardouin P, Iost A. Improvementin the morphology of Tibasedsurfaces: a new process to increasein vitro human osteoblast response.Biomaterials. 2002;23:1563-1577.8. Mustafa K, Wennerberg A,Wroblewski J, Hultenby K, Lopez BS,Arvidson K. Determining optimal surfaceroughness of TiO(2) blasted titaniumimplant material for attachment,proliferation and differentiation of cellsderived from human mandibular alorim29_202 Mp_71File # 02emMamle Mante et alveolar bone. Clin Oral Implants Res.2001;12:515-525.9. Lincks J, Boyan BD, BlanchardCR, et al. Response of MG63 osteoblast-like cells to titanium and titaniumalloy is dependent on surfaceroughness and composition. Biomaterials.1998;19:2219-2232.10. Deligianni DD, Katsala N, LadasS, Sotiropoulou D, Amedee J, MissirlisYF. Effect of surface roughness ofthe titanium alloy Ti6Al4V on humanbone marrow cell response and on proteinadsorption. Biomaterials. 2001;22:1241-1251.11. Knabe CKF, Fitzner R, RadlanskiRJ, Gross U. In vitro investigationof titanium and hydroxyapatite dentalimplant surfaces using a rat bone marrowstromal cell culture system. Biomaterials.2002;23:3235-3245.12. Kieswetter KS, Hummert TW.Surface roughness modulates the localproduction of growth factors and cytokinesby osteoblast-like MG-63 cells.J Biomed Mat Res. 1996;32:55-63.13. Lohmann CH, Schwartz Z,Koster G, et al. Phagocytosis of weardebris by osteoblasts affects differentiationand local factor production in amanner dependent on particle composition.Biomaterials. 2000;21:551-561.14. Boyan BD, Sylvia VL, Liu Y, etal. Surface roughness mediates its effectson osteoblasts via protein kinaseA and phospholipase A2. Biomaterials.1999;20:2305-2310.15. Boyan BD, Lincks J, LohmannCH, et al. Effect of surface roughnessand composition on costochondralchondrocytes is dependent on cell maturationstate. J Orthopaed Res. 1999;17:446-457.16. Zohar R, Sodek J, McCullochCA. Characterization of stromal progenitorcells enriched by flow cytometry.Blood. 1997;90:3471-3481.17. Ghilzon R MC, Zohar R. Stromalmesenchymal progenitor cells.Leuk Lymphoma. 1999;32:211-221.18. Lee YM SY, Lim YT, Kim S, etal. Tissue-engineered growth of bonemarrow cell transplantation using po-Journal of Oral Implantology 71ORAL IMPLANTOLOGYTuesday Mar 11 2003 01:58 PMAllen Press x DTPro SystemATTACHMENT OF MARROW STROMAL CELLS TO TITANIUMrous calcium metaphosphate matrices.J Biomed Mat Res. 2001;54:216-223.19. Kilpadi KL, Bellis SL. Hydroxyapatitebinds more serum proteins,purified integrins, and osteoblastprecursor cells than titanium or steel. JBiomed Mater Res. 2001;57:258-267.20. Lohmann CHS, Sylvia R, CochranVL, et al. Surface roughness modulatesthe response of MG63 osteoblast-like cells to 1,25-(OH)2D3through regulation of phospholipaseA2 activity and activation of protein kinaseA. J Biomed Mater Res. 1999;47:139-151.21. Matsuura T HR, Okamoto K,Kimoto T, Akagawa Y. Diverse mechanismsof osteoblast spreading on hydroxyapatiteand titanium. Biomaterials.2000;21:1121-1127.22. Mayr-Wohlfart U, Fiedler J,Gunther KP, Puhl W, Kessler S. Proliferationand differentiation rates of ahuman osteoblast-like cell line (SaOS-2) in contact with different bone substitutematerials. J Biomed Mat Res.2001;57:132-139.23. Yang YJT, Deng L, Ong JL.Morphological behavior of osteoblastlikecells on surface-modified titanium72 Vol. XXIX/No. Two/2003in vitro. Biomaterials. 2002;23:1383-1389.24. Ku CH, Pioletti DP, Browne M,surface treatments on osteoblasts behaviour.Biomaterials. 2002;23:1447-1454.25. Rocca MFM, Giavaresi G, AldiniNN, Giardino R. Osteointegrationof hydroxyapatite-coated and uncoatedtitanium screws in long-term ovariectomizedsheep. Biomaterials. 2002;23:1017-1023.26. Ku CHB, Gregson M, CorbeilPJ, Pioletti JDP. Large-scale gene expressionanalysis of osteoblasts culturedon three different Ti6Al4V surfacetreatments. Biomaterials. 2001;23:4193-4202.27. Larsson C, Thomsen P, LausmaaJ, Rodahl M, Kasemo B, EricsonLE. Bone response to surface modifiedtitanium implants: studies on electropolishedimplants with different oxidethicknesses and morphology. Biomaterials.1994;15:1062-1074.28. Brunette DM. The effects of implantsurface topography on the behaviorof cells. Int J Oral Maxillofac Implants.1988;3:231-246.Gregson PJ. Effect of different Ti6Al4V 448.30. Clark P, Connolly P, Curtis AS,orim 29_202 Mp_72File # 02emDow JA, Wilkinson CD. Topographicalcontrol of cell behaviour. I. Simple stepcues. Integrat Ann Index. 1987;99:439-Dow JA, Wilkinson CD. Topographicalcontrol of cell behaviour: II. Multiplegrooved substrata. Integrat Ann Index.1990;108:635-644.31. Hay DI, Moreno EC. Differentialadsorption and chemical affinitiesof proteins for apatitic surfaces. J DentRes. 1979;58:930-942.32. Andrade J. Interfacial phenomenaand biomaterials. Med Instrum.1973;7:110-119.33. Goldberg VM, Jinno T. Thebone-implant interface: a dynamic surface.J Long-Term Effects Med Implants.1999;9:11-21.34. Klokkevold PR, Nishimura RD,Adachi M, Caputo A. Osseointegrationenhanced by chemical etching of the titaniumsurface. A torque removalstudy in the rabbit. Clin Oral ImplantsRes. 1997;8:442-447.35. Trisi P, Rao W, Rebaudi A. Ahistometric comparison of smooth andrough titanium implants in humanlow-density jawbone. Int J Oral MaxillofacImplants. 1999;14:689-698. 29. Clark P, Connolly P, Curtis AS,