poly(lactic acid) hybrid

Electrospun microfiber meshes of silicon-doped vaterite/poly(lactic acid)hybrid for guided bone regenerationAkiko Obata a,*,Toshiki Hotta a ,Takashi Wakita a,b ,Yoshio Ota c ,Toshihiro Kasuga aaGraduate School of Engineering,Nagoya Institute of Technology,Gokiso Cho,Showa ku,Nagoya 466-8555,Japan bYamahachi Dental Manufacturing Co.,54-1Ochigara,Nishiura Cho,Gamagori,Aichi 443-0105,Japan cYabashi Industries Co.Ltd.,188-1Akasaka Cho,Ogaki,Gifu 503-2213,Japana r t i c l e i n f o Article history:Received 29May 2009Received in revised form 23October 2009Accepted 9November 2009Available online 11November 2009Keywords:Polylactic acidCalcium carbonate Silicon Membrane Osteoblasta b s t r a c tSilicon-releasable microfiber meshes consisting of silicon-doped vaterite (SiV)particles andpoly(lactic acid)(PLA)hybrids were prepared by electrospinning.Due to their flexibility and porosity they formed ideal membranes or scaffolds for guided bone regeneration.In addition,a trace amount of silicon species has been reported to stimulate osteogenic cells to mineralize and enhance bone formation.We propose a new method of preparation of silicon-releasing microfiber meshes by electrospinning.Their structure and hydroxyapatite (HA)-forming abilities in simulated body fluid were examined.In addition,we studied their stimulatory effects on osteoblast-like cells in vitro and bone-forming ability in vivo,with a special emphasis on their ability to release silicon.The meshes consisted of a hybrid of carboxy groups in PLA and amino groups in siloxane,derived from aminopropyltriethoxysilane or calcium ions on the SiV surface.This hybrid exhibited an enhanced ability to form HA.The meshes coated with HA released 0.2–0.7mg l À1silicon species into the culture medium over 7days.Enhanced proliferation of osteo-blast-like cells was observed using the meshes and new bone formed on the meshes when implanted into the calvaria of rabbits.These meshes,therefore,provide an excellent substrate for bone regeneration and exhibit enhanced bone-forming ability under both in vitro and in vivo conditions.Ó2009Acta Materialia Inc.Published by Elsevier Ltd.All rights reserved.1.IntroductionBiodegradable polymer scaffolds have been developed for use in bone reconstruction,tissue engineering and drug delivery systems [1,2].In addition,three-dimensional scaffolds are common in bio-materials.Electrospinning is a process that can generate polymer fiber meshes with high porosity and flexibility,both essential com-ponents of tissue engineering [3–5].In general,the process of elec-trospinning is affected by both system parameters and process parameters.System parameters include polymer molecular weight (Mw)and polymer solution properties such as viscosity and con-ductivity.Process parameters,on the other hand,involve flow rate of the polymer solution,electric potential and the distance be-tween the capillary and collector,among others [3].It is possible to control the fiber diameter and the thickness of fiber mesh scaf-folds by optimizing these parameters.To use electrospun fiber meshes in bone tissue engineering applications,their various prop-erties,including material,fiber orientation,porosity and surface modification,must be considered.To optimize the mechanical and biomimetic properties of a fiber mesh it is necessary to choose the appropriate materials,since theorientation and diameter of the fibers in electrospun scaffolds can influence cellular compatibility [6–8].The fiber diameter in elec-trospun scaffolds has,for example,been reported to influence osteoblast proliferation and mineralization.Electrospun fiber meshes have been fabricated to contain DNA and proteins,which enhance cellular compatibility and bone formation upon their re-lease [8–11].The DNA and proteins are released by degradation of polymer matrices such as poly(lactic acid)(PLA),poly(lactide-co -glycolide)(PLGA)and poly(glycolide acid)(PGA).Despite their stimulatory effects on osteoblasts and bone mar-row-derived stem cells,electrospun fiber scaffolds consisting of polymer/ceramic composites have not been widely researched as biomaterials.Polymer and ceramic composites have been developed for bone reconstruction and showed good mechanical and bone-forming properties [12–14].We hypothesize that elec-trospun fiber scaffolds consisting of PLA/calcium carbonate com-showed excellent hydroxyapatite (HA)-forming ability,cellular compatibility and biocompatibility in in vitro and in vivo tests [17,18].1742-7061/$-see front matter Ó2009Acta Materialia Inc.Published by Elsevier Ltd.All rights reserved.doi:10.1016/j.actbio.2009.11.013*Corresponding author.Tel./fax:+81527357250.E-mail address:obata.akiko@nitech.ac.jp (A.Obata).Acta Biomaterialia 6(2010)1248–1257Contents lists available at ScienceDirectActa Biomaterialiajournal homepage:/locate/ac tabiomatGuided bone regeneration(GBR)is a useful method of bone reconstruction in vivo.GBR membranes requireflexibility to adapt to the shape of the bone defect and strength to maintain the space for bone formation and to connect with the soft tissues.An electro-spunfiber mesh is a good candidate for a GBR membrane due to theflexibility inherent in the electrospinning process.Release of calcium and silicon species from an electrospunfiber mesh is an additional attribute which promotes bone regeneration.Calcium species and a trace amount of silicon species have been reported to enhance bone formation by osteoblasts[19–23].Previously we developed silicon-doped vaterite(SiV)powders and their PLA com-posites and tested their cellular compatibility using mouse osteo-blast-like cells[24].Afilm of the composites was prepared by dip coating the composite solution on a coverglass and drying at room temperature.Thefilm showed release of calcium and silicon into the culture medium through degradation of the PLA and SiV matrix.Cell proliferation and mineralization on the compositefilm was enhanced compared with an undoped vaterite and PLA com-positefilm,which showed no silicon release.Our preliminary experiments demonstrated that aflexiblefiber mesh consisting of PLA composites with a large amount of SiV could be successfully prepared by the electrospinning method, while afiber mesh consisting of a PLA composite with vaterite(un-doped type)was brittle.We propose to further characterize the SiV and PLA hybrid(Si–PVH)and compare its chemical structure with the undoped vaterite and PLA hybrid(PVH).To enhance cellular adhesion the compositefiber meshes were coated with HA using simulated bodyfluid(SBF).The cellular and tissue compatibilities of both preparedfiber meshes were evaluated by in vitro tests using mouse osteoblast-like cells(MC3T3-E1cells)and in vivo tests using rabbit calvaria.2.Materials and methods2.1.Preparation of SiV and vaterite powdersSilicon-doped vaterite(SiV)powders were prepared by a car-bonation process with methanol.A sample of150g Ca(OH)2was mixed with60ml aminopropyltriethoxysilane(APTES)and 2000ml methanol with the addition of CO2gas for75min at a rate of2000ml minÀ1.The resulting slurry was dried at110°C,result-ing in the SiV powder.The content of silicon in SiV was estimated to be approximately3wt.%by X-rayfluorescence analysis (RIX3000,Rigaku,Japan)(50kV,60mA).A calibration curve(X-ray intensity vs.SiO2content)was established by measuring lime-stone containing various amounts of SiO2.Vaterite powders were also prepared by the carbonation process without APTES.SiV and vaterite powders were characterized by X-ray diffractometry (XRD)(RAD-B,Rigaku,Japan)(CuK a,40kV,20mA).The specific surface areas of SiV and vaterite were measured according to a BET theory using nitrogen gas adsorption(NOVA3000,Quanta-chrome Corp.,USA).The powders were deaerated at150°C for 1h before measurement.The structure of silicon in SiV was exam-ined by29Si magic angle spinning nuclear magnetic resonance(29Si MAS-NMR)(UNITY plus,Varian,USA),operating at pulses of6.3l s and at recycle delays of15s.The sample spinning speed was 3.0kHz.Chemical shifts were measured relative to polydime-thylsiloxane.2.2.Preparation of microfiber meshesThe Si–PVH and PVH microfiber meshes were prepared by an electrospinning method.SiV or vaterite powder was mixed with PLA(PurasorbÒ;molecular weight260kDa,PURAC,The Nether-lands)at200°C for10min with a kneader,resulting in the formation of PLA composites containing60wt.%of the powder. The composites were dissolved in chloroform at10wt.%to prepare the solution for electrospinning.In our preliminary experiments, this ratio was found to be optimal for preparing microfiber meshes of approximately10l m diameter.The PLA microfiber mesh was also prepared by an electrospinning method.The PLA was dis-solved in chloroform at9wt.%to prepare the solution for electros-pinning.The prepared solutions were loaded into a syringe pump (FP-W-100,Melquest,Japan)with a syringe needle(22gauge for Si–PVH and PVH,19gauge for PLA),which was set at 0.05ml minÀ1.The distance between the needle and an aluminum collector was150mm.Electrospinning was carried out at20kV of impressed voltage and$100l A of applied current(Power Supply HARb-40P0.75,Matsusada Precision Inc.,Japan)at room tempera-ture and approximately55%relative humidity.The SiV powders and resultant electrospunfibers were coated with amorphous osmium using plasma chemical vapor deposition (CVD)equipment and then morphologically observed byfield emission scanning electron microscopy(SEM)(JSM-6301F,JEOL, Japan)(accelerating voltage5keV).Structural analyses of Si–PVH were performed using13C cross-polarization magic angle spinning nuclear magnetic resonance(13C CP/MAS-NMR)(UNITY plus,Var-ian,USA)and Fourier transform infrared(FTIR)spectroscopy(FT-IR-460Plus,JASCO,Japan).13C CP/MAS-NMR spectra were taken with4.7l s and8.0s recycle delays.The sample spinning speed was 3.0kHz.Chemical shifts were measured relative to adamantane.The molecular weights of PLA matrix in the prepared meshes were measured by gel permeation chromatography(GPC)(Shima-dzu,Japan)using two columns(KF-604,Shodex,Japan).The meshes were dissolved in chloroform at a concentration of10mg PLA per5ml chloroform and the solutionsfiltered.The measure-ments were carried out at40°C at aflow rate of0.6ml minÀ1.Poly-styrene standards(SM-105,Shodex,Japan)were used to establish a calibration curve.2.3.Coating of thefibers with HAThe prepared microfiber mesh was immersed in10ml of an aqueous solution containing ion concentrations1.5times those of SBF(1.5SBF)at37°C for24h.The1.5SBF solution consisting of 3.75mM Ca2+,213.0mM Na+, 2.25mM Mg2+,7.5mM K+, 222.45mM ClÀ,6.3mM HCO3À,1.5mM HPO42À,0.75mM SO42À, 50mM(CH2OH)3CNH2and45.0mM HCl was prepared using re-agent grade NaCl,NaHCO3,KCl,K2PO4Á3H2O,MgCl2Á6H2O,HCl, CaCl2,NaSO4and NH2C(CH2OH)3.After immersion in1.5SBF solu-tion the mesh was washed with distilled water and then dried at room temperature.The dried samples were coated with amor-phous osmium using plasma CVD equipment and observed by SEM.2.4.Measurement of silicon release from a Si–PVH microfiber mesh before and after HA coatingThe amount of silicon released from the Si–PVH microfiber meshes before and after HA coating was evaluated by immersing them in4ml of a culture medium(alpha minimum essential med-ium,a-MEM)containing10%fetal bovine serum(FBS)and incu-bated at37°C in a humidified atmosphere of95%air,5%CO2for 5days.The sample coated with HA is denoted Si–PVH/HA mesh. The size of the sample used was10Â10mm.The culture medium was changed after either1or3days soaking,according to the pro-cess of the cell culture test.The sample was taken out of the culture medium and the mediumfiltered.The level of silicon species in the resulting medium was measured by inductively coupled plasma atomic emission spectroscopy(ICP-AES)(ICPS-500,Shimadzu,Ja-A.Obata et al./Acta Biomaterialia6(2010)1248–12571249pan).A calibration curve was established using an aqueous solu-tion of Si4+ions at concentrations of1,10and20ppm.2.5.Cell culture testsThe microfiber mesh used for the cell culture tests was steril-ized using ethylene oxide gas.Mouse osteoblast-like cells (MC3T3-E1cells)were seeded onto the prepared mesh(Si–PVH/ HA or PLA mesh)in24-well plates at a density of30,000 cells wellÀ1.A TermanoxÒplastic plate,which has often been used as a control sample,was used as a control sample and a-MEM con-taining10%FBS was used as the culture medium.The medium was changed after1day of culture and then changed every other day. The number of MC3T3-E1cells was evaluated after treatment with Cell Counting Kit-8(Dojindo,Japan).One reagent in the kit is a water soluble tetrazolium salt which is reduced by dehydrogen-ases in live cells to generate water soluble formazan.Water soluble formazan has an absorption maximum at about460nm.The num-bers of live cells in the samples were counted by measuring the absorbance of the resulting medium at450nm.Student’s t-test was used for the cell counts.Differences between the samples were determined by Student’s t-test,with p<0.05considered to be sta-tistically significant.The cells cultured on the samples for3days werefixed in2.5%glutaraldehyde for40min at4°C,dehydrated through a series of increasing concentrations of ethanol and,final-ly,dried with hexamethyldisilazane.The dried samples were coated with amorphous osmium using plasma CVD equipment and observed by SEM.2.6.Preparation of a bilayered microfiber mesh consisting of Si–PVH/ HA and PLA meshesA microfiber mesh having large spaces is not an appropriate candidate for a GBR membrane because the mesh may allow fibrous cells and soft tissues to permeate through them[25].To address this issue we prepared a bilayered microfiber mesh using Si–PVH/HA mesh and the dense PLA mesh,which was not expected to allow any cell permeation.In addition,the strength of the Si–PVH/HA mesh was improved by combination with the dense PLA mesh.Details of the preparation method have been reported by Wakita et al.[26].It was found that the combination induced no fracture or fragmentation of Si–PVH/HA and PLA.The PLA mesh was pressed at15MPa for1min to increase its density.Adherence between the Si–PVH mesh and the dense PLA mesh was accomplished by pressing at15MPa using a stainless mesh(No.40)at150°C and soaking in1.5SBF for1day.The result was a bilayered mesh consisting of Si–PVH/HA mesh and the dense PLA mesh(Fig.1A).2.7.Evaluation of in vivo bone-forming abilityThe in vivo study was carried out at the laboratory of Hamri Co. Ltd.(Japan).Fifteen male New Zealand rabbits(Kbl:NZW), 14weeks of age and between2.50and3.20kg in weight,were pur-chased from Kitayama Labes Co.Ltd.(Japan).The animals were housed in vivaria at23±3°C,50±20%humidity and a12h on/ off light cycle and ventilation was cycled approximately12timesFig.1.Schematic images of(A)preparation of the bilayered mesh and(B)protocol of the in vivo test.Adherence between the Si–PVH mesh and the dense PLA mesh was accomplished by pressing with a stainless steel mesh at150°C and soaking in1.5SBF for1day.The result was a bilayered mesh consisting of Si–PVH/HA mesh and the dense PLA mesh.An8mm diameter hole was drilled into the front midline of the animal’s calvaria and then covered with the prepared bilayered mesh.1250 A.Obata et al./Acta Biomaterialia6(2010)1248–1257per hour.The animals were maintained according to the National Institutes of Health(NIH)guide for care and use of laboratory animals.After allowing the animals to adapt to the new environment for 12days they were stratified into three groups offive animals each. After anesthetizing the animals an8mm diameter hole was drilled into the front midline of the calvaria using a bone cutter(BL-302A, Osada Electric Co.Ltd.,Japan)and then covered with the prepared bilayered meshes(Fig.1B).The bilayered mesh wasfixed by being placed between the skin and bone.The meshes were implanted with the Si–PVH/HA mesh in contact with the bone and the dense PLA mesh in contact with the skin.The implanted meshes were harvested with the surrounding tissues for histological examina-tion after4,8and12weeks.The samples were dehydrated, embedded in methyl methacrylate and then sliced,resulting in slides of approximately20l m thickness.The slides were stained with Villanueva-Goldner and observed by optical microscopy (CH-2,Olympus,Japan).Villanueva-Goldner stains mineralized tis-sues green.3.Results3.1.Materials characterizationFig.2shows SEM micrographs of the SiV and vaterite particles. The SiV and vaterite particles had diameters of$1l m and $500nm,respectively.These two types of particles consisted of aggregates of nano-sized primary particles.The sizes of the SiV particles were larger than those of the vaterite particles.The re-sults of the nitrogen gas adsorption method revealed that the spe-cific surfaces of SiV and vaterite were67.2and30.7m2gÀ1, respectively.These results suggest that the conditions for aggrega-tion of the primary particles in the suspension were influenced by the addition of APTES.Fig.3shows the XRD patterns of SiV and vaterite.A weak peak corresponding to calcite appeared at approximately29°in the pat-tern of the vaterite sample.A trace amount of calcite was found with a predominant vaterite phase,because the calcite phase is more stable than the other polymorphs of calcium carbonate,vate-rite and aragonite.In the present work the sample is denoted ‘‘vaterite”.The peaks of the SiV pattern are broader than those of the vaterite pattern.SiV was found to consist predominantly of vaterite.The structure of APTES in SiV was further characterized by29Si MAS-NMR spectroscopy.Fig.4shows the29Si MAS-NMR spectrum of SiV.A peak atÀ50toÀ80ppm can be seen.The peak may be superimposed on those corresponding to T2[NH2(CH2)3Si(OSi)2 (OH,OC2H5)]and T3[NH2(CH2)3Si(OSi)3][27].This indicates that no APTES in SiV existed as the monomer and that siloxane was formed by condensation of APTES during the preparation of SiV.Meshflexibility was felt to be a reflection of its structural prop-erties,such as the molecular weight of the PLA matrix and struc-tural changes in the interface between the matrix and the vaterite or SiV powders.Table1shows the weight average molec-ular weight(Mw),number average molecular weight(Mn)and polydispersity(Mw/Mn)of PLA and the PLA matrix in PVH and Si–PVH.Although the molecular weight of PLA decreased slightly after kneading with the SiV powders,it decreased significantly after kneading with vaterite powder.The prepared Si–PVH mesh showedflexibility without breaking,whereas the PVH mesh was brittle and was easily broken by hand.The ratios Mw/Mn for PVH and Si–PVH were higher than that for PLA.This indicates that the molecular weights of these compos-ites were disperate.Fig.5shows various peaks corresponding to the carboxy groups in the13C CP/MAS-NMR spectra of the prepared composites.The band of the carboxy peak shifted slightly and a new peak appeared in the low magneticfield area of the band, where carboxy groups form coordinate bonds with bivalent ions [28].The peaks further separated into two peaks,peak A andB. Fig.2.SEM photographs of(A)vaterite and(B)SiV.The SiV and vaterite particles have diameters of$1l m and$500nm,respectively.The SiV particles are larger than the vaterite particles.This suggests that the conditions for aggregation of the primary particles in the suspension were influenced by addingAPTES.Fig.3.XRD patterns of(A)vaterite and(B)SiV.A weak peak corresponding to calcite appears at approximately29°in the pattern of the vaterite sample,but the vaterite sample and SiV were found to consist predominantly of vaterite phase.The peaks for SiV were broader than those of vaterite.A.Obata et al./Acta Biomaterialia6(2010)1248–12571251Peak A was defined at one magnetic field in the band and peak B as one magnetic field lower in the band.The intensity of peak B in the spectrum of PVH was larger than in the spectrum of Si–PVH.This indicates that more carboxy groups formed coordinate bonds with calcium ions in PVH than Si–PVH and that the formation of coordinate bonds caused the decrease in molecu-lar weight.Fig.6shows the FTIR spectra of Si–PVH and PVH.In the spectrum of Si–PVH a new weak peak appears at around 1650cm À1,as well as the peaks corresponding to PLA and CO 32À.This peak appears to originate from the amide bond,which indi-cates that carboxy groups on PLA form chemical bonds with amino groups on APTES.Fig.7A demonstrates the greater flexibility of the Si–PVH mesh,which could be folded without fracture or fragmentation using a pair of tweezers.Microfibers approximately 10l m in diameter were observed on the Si–PVH mesh surface and were found to twine around one another.The size of the gaps between the microfibers varied between 10l m and several hundred microme-ters,as shown in Fig.7B.The SiV particles observed on the micro-fiber surface were coated with a thin film of the PLA matrix,as shown in Fig.7C.The fracture face of the microfibers showed that the particles were closely packed in the microfibers (Fig.7D).The particles were embedded in the PLA matrix and after immersion in 1.5SBF solution for 1day the surfaces of the microf-ibers were coated with leaf-shaped deposits,as shown in Fig.7E.No significant change was observed in the gap sizes between the microfibers before and after immersion (Fig.7F).After HA coating further testing demonstrated several cracks,but the coating did not peel off.The deposits were confirmed by XRD to be HA of approximately 1l m thickness (Fig.7G).On the other hand,the PVH mesh consisted of short microfibers approximately 7l m in diameter (Fig.7H).3.2.Ease of silicon releaseFig.8shows the dynamics of the release of silicon ions from both the Si–PVH mesh and Si–PVH/HA mesh into the culture med-ium over 5days.Closed bars indicate the ion amount in the culture medium and open bars indicates the total amount of ions released up to each measurement point.After 1day of immersion a concen-tration of 8.2mg l À1ionic silicon species had been released from Si–PVH,while this amount was reduced to 1.2mg l À1on day 3and 0.59mg l À1on day 5.The Si–PVH mesh was then soaked in cul-ture medium after being incubated in 1.5SBF for 1day and coated with HA.The amount of the silicon species released from it was significantly reduced,with only 0.68mg l À1released from the mesh after 1day of soaking in culture medium.The amount re-leased was even smaller on days 3and 5of soaking,at 0.39mg l À1on day 3and 0.41mg l À1on day 5.3.3.Proliferation of MC3T3-E1cells on the prepared meshesThe cellular compatibility of the Si–PVH/HA mesh compared with the PLA mesh was evaluated in culture using MC3T3-E1cells.Fig.9shows the numbers of cells after culture on Si–PVH/HA mesh,PLA mesh and Thermanox Òfor 2weeks.The number of cells after culture on the Si–PVH/HA mesh was higher than that on thePLAFig. 4.29Si MAS-NMR spectrum of SiV.The peak at À50to À80ppm may be superimposed on those corresponding to T 2[NH 2(CH 2)3Si(OSi)2(OH,OC 2H 5)]and T 3[NH 2(CH 2)3Si(OSi)3].This indicates that no APTES in SiV exists as the monomer and siloxane was formed.Table 1Weight average molecular weight (Mw),number average molecular weight (Mn)and polydispersity (Mw/Mn)of PLA,PVH and Si–PVH.Mw (kDa)Mn (kDa)Mw/Mn PLA 259168 1.5PVH 177 2.4Si–PVH185862.1Fig.5.13C CP/MAS-NMR spectra of (A)PVH and (B)Si–PVH.The peaks further separated into two peaks,A and B.The intensity of peak B in the spectrum of PVH is larger than that of Si–PVH.This indicates that more carboxy groups formed coordinate bonds with calcium ions in PVH than Si–PVH.1252 A.Obata et al./Acta Biomaterialia 6(2010)1248–1257mesh at all time points.The PVH mesh could not be used as a con-trol mesh due to its extreme brittleness.Fig.10shows SEM micrograph of cells cultured on the Si–PVH/HA mesh for 3days.The cells elongated on the fiber surfaces and extended over several fibers.Some of them entered the fiber mesh and extended within it.3.4.Bone formation in the prepared meshesAn 8mm diameter hole in bone is generally not regenerated if nothing is done to stimulate it.In our preliminary experiment a PLGA mesh was implanted in the same position as that in this study for 12weeks and no new bone was formed in/on the mesh at the center of the hole (data not shown).Bone growth into the defect from the margin was observed in all of the samples.There-fore,we tested the effects of the bilayered meshes on bone forma-tion by observing new bone formation at the center of the 8mm holes.This animal test was a fundamental study.Although this ani-mal test was not a load-bearing model,we believe that it was use-ful in evaluating the bone-forming ability of the bilayered meshes.Fig.11shows the in vivo response to the bilayered meshes and new bone formation at the center of the holes.Villanueva-Goldner stain showed that new bone formed in the mesh 4weeks after implantation.Bone began to form around the pressed points formed during layering of the two types of mesh,the Si–PVH/HA mesh and the dense PLA mesh.The bone area expanded in the Si–PVH/HA mesh after 8weeks.After 12weeks new bone forma-tion was observed over almost the entire area of implanted Si–PVH/HA mesh.There was no tissue inflammation observed histologically.4.DiscussionEarlier work showed that composite films which release silicon can stimulate MC3T3-E1cells to proliferate and mineralize [17,24].These composite films were prepared by dip coating a solution of SiV and PLA composites on a coverglass,drying at room tempera-ture and pealing the dried composite from the glass.The prepared composite film had no pores to allow cells to enter,therefore,cells proliferated only on the film surface.When applying this technique to a GBR membrane the composites should have porosity in order to induce adhesion and proliferation of cells and to provide nutri-tion to the cells.To address this issue we developed a fiber mesh using PLA and SiV formed by electrospinning.We anticipated that the electrospun fiber mesh would contain a large volume of pores and would show flexibility,both essential components of a GBR membrane.The Si–PVH mesh showed greater flexibility and a higher weight averaged molecular weight PLA than the PVH mesh,in the case of SiV due to siloxane on the vaterite surface,which resulted in the formation of a flexible mesh containing a large amount of SiV powder (60wt.%).There has been increasing interest in the use of electrospun fi-brous meshes as scaffolds in medicine.In particular,biocompatible polymers such as PLGA and poly(caprolactone)have been applied as materials for meshes,incorporating various growth factors or DNA [3].On the other hand,very few studies have been performed on fibrous meshes consisting of polymer/ceramic composites.Pre-viously,PLGA and HA composite scaffolds fabricated by electros-pinning contained only 5–10wt.%HA nanocrystals [29].We believe that a larger amount of calcium is needed to enhance the bone-forming ability of electrospun fibrous meshes and vaterite powders are effective materials for providing calcium.In the pres-ent work a fibrous mesh containing a large amount of vaterite (up to 60wt.%)was successfully prepared by kneading and subsequent electrospinning.Vaterite or SiV particles were embedded in a PLA matrix,with very thin PLA layers on the particle surfaces,as shown in Fig.7C and D.Calcium ions are supposed to be released through degradation of the thin PLA layer.The PVH mesh,however,was brittle,without flexibility,due to a drastic decrease in the PLA molecular weight after kneading with vaterite powders,as shown in Table 1.Maeda et al.reported that in vaterite and PLA composites cal-cium ions on vaterite particles formed coordinate bonds with car-boxy groups in the PLA structure [30].The polymer chains in the Si–PVH mesh,however,were not broken by the formation of coor-dinate bonds between calcium ions and carboxy groups (Fig.5)due to siloxane derived from APTES on the vaterite surface,which pre-vents the formation of coordinate bonds.This resulted in the for-mation of a flexible mesh containing a large amount of SiV powder (60wt.%).The specific surface areas of SiV and vaterite were estimated to be 67.2and 30.7m 2g À1,respectively,by nitrogen gas adsorption.The peaks corresponding to vaterite in the XRD pattern of SiV were broader than that of vaterite (Fig.3).This result indicates that the diameter of the primary particles of SiV is smaller than that of vaterite.Crystal growth of the primary particles of vateriteap-Fig.6.FTIR spectra of (A)PVH and (B)Si–PVH.A weak peak appears at around 1650cm À1in the spectrum of Si–PVH.This peak appears to originate from the amide bond between carboxy groups in PLA and amino groups in APTES.A.Obata et al./Acta Biomaterialia 6(2010)1248–12571253。