Hydroxyapatite/SiO 2 –CaO–P 2 O 5 glass materials: In vitro bioactivity and biocompatibility S. Padilla, J. Roma ´n, S. Sa ´nchez-Salcedo, M. Vallet-Regı ´ * Departamento de Quı ´mica Inorga ´ nica y Bioinorga ´ nica, Facultad de Farmacia, Universidad Complutense de Madrid, 28040 Madrid, Spain Received 6 September 2005; received in revised form 22 December 2005; accepted 25 January 2006 Abstract Materials obtained by the heat treatment of mixtures of hydroxyapatite (HA) and a silicate-based glass of the system SiO 2 –CaO–P 2 O 5 have been investigated. The influence of the glass content on the porosity, microstructure and on the constituent phases of the final mate- rials was studied. The influence of these factors on the in vitro bioactive behaviour of the obtained materials was also investigated. In addition, an in vitro biocompatibility assay with osteoblastic-like cells was carried out. The addition of the glass to HA induced different solid-state reactions that yield the transformation of HA into a- and b-tricalcium phosphate as well as the formation of silicon-containing phases (silicocarnotite or pseudowollastonite). In these mixtures an enhancement in the porosity, pore size and a heterogeneous micro- structure was observed, compared with the precursors. As the sol gel glass content increased, the previous effects were higher. The mate- rials showed the formation of an apatite-like layer on their surface when soaked in simulated body fluid, being faster in the sample with a higher content of glass. The formation of the new layer began in preferential zones in both samples, depending on the different reactivity of the crystalline phases formed. A synergistic effect between HA and glass was observed, showing in the mixtures a faster bioactive behaviour than in HA and glass themselves. The obtained materials allow a good attachment, spread and proliferation of the osteoblas- tic-like cells and no cytotoxic effect was observed. Ó 2006 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. 1. Introduction Calcium phosphate materials and hydroxyapatite (HA), in particular, seem to be ideally suited to be bone implants, because of their similarity to the chemical composition of osseous tissues, their biocompatibility, lack of inflamma- tory response, and osteoconductive capacity. HA allows a specific biological response in the tissue–implant interface, which leads to the formation of bonds between the bone and the materials (bioactivity) [1]. However, although HA is bioactive it shows a limited in vitro reactivity [2] and in vivo assays have shown slow formation of osseous tissue [3]. Other materials such as bioactive silicate-based glasses [4] show a higher bioactive behaviour than calcium phos- phate materials [1,5]. For instance, the bioactivity reactions in silica-based glasses occur in a few minutes [6], whereas those in HA take several days [5]. A very important char- acteristic of the silica-based glasses is that they show a genetic control of the cellular response of osteoblasts. It has been observed that different genes are up regulated within 48 h of exposure of primary human osteoblasts to the ionic dissolution products of bioactive glasses [7]. The rapid bioactive behaviour of these silica-based glasses has been related to the role of SiO 2 or silicon in their surface reactions and therefore on their in vivo and in vitro behav- iour. The surface reactions that occur in the bioactive glasses allow the subsequent crystallization of apatite crys- tals, cell adhesion and collagen formation [8,9]. Silicon is believed to be essential in skeletal development; the first indications of a physiological role for silicon were reported by Carlisle [10], who observed that silicon was involved in the early stage of bone calcification. Schwarz and Milne 1742-7061/$ - see front matter Ó 2006 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2006.01.006 * Corresponding author. Tel.: +34 91 3941861; fax: +34 91 3941786. E-mail address: [email protected](M. Vallet-Regı ´). Acta Biomaterialia 2 (2006) 331–342 www.actamat-journals.com
Transcript
Acta Biomaterialia 2 (2006) 331–342
www.actamat-journals.com
Hydroxyapatite/SiO2–CaO–P2O5 glass materials: In vitrobioactivity and biocompatibility
S. Padilla, J. Roman, S. Sanchez-Salcedo, M. Vallet-Regı *
Departamento de Quımica Inorganica y Bioinorganica, Facultad de Farmacia, Universidad Complutense de Madrid, 28040 Madrid, Spain
Received 6 September 2005; received in revised form 22 December 2005; accepted 25 January 2006
Abstract
Materials obtained by the heat treatment of mixtures of hydroxyapatite (HA) and a silicate-based glass of the system SiO2–CaO–P2O5
have been investigated. The influence of the glass content on the porosity, microstructure and on the constituent phases of the final mate-rials was studied. The influence of these factors on the in vitro bioactive behaviour of the obtained materials was also investigated. Inaddition, an in vitro biocompatibility assay with osteoblastic-like cells was carried out. The addition of the glass to HA induced differentsolid-state reactions that yield the transformation of HA into a- and b-tricalcium phosphate as well as the formation of silicon-containingphases (silicocarnotite or pseudowollastonite). In these mixtures an enhancement in the porosity, pore size and a heterogeneous micro-structure was observed, compared with the precursors. As the sol gel glass content increased, the previous effects were higher. The mate-rials showed the formation of an apatite-like layer on their surface when soaked in simulated body fluid, being faster in the sample with ahigher content of glass. The formation of the new layer began in preferential zones in both samples, depending on the different reactivityof the crystalline phases formed. A synergistic effect between HA and glass was observed, showing in the mixtures a faster bioactivebehaviour than in HA and glass themselves. The obtained materials allow a good attachment, spread and proliferation of the osteoblas-tic-like cells and no cytotoxic effect was observed.� 2006 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction
Calcium phosphate materials and hydroxyapatite (HA),in particular, seem to be ideally suited to be bone implants,because of their similarity to the chemical composition ofosseous tissues, their biocompatibility, lack of inflamma-tory response, and osteoconductive capacity. HA allows aspecific biological response in the tissue–implant interface,which leads to the formation of bonds between the boneand the materials (bioactivity) [1]. However, althoughHA is bioactive it shows a limited in vitro reactivity [2]and in vivo assays have shown slow formation of osseoustissue [3].
Other materials such as bioactive silicate-based glasses[4] show a higher bioactive behaviour than calcium phos-
1742-7061/$ - see front matter � 2006 Acta Materialia Inc. Published by Else
phate materials [1,5]. For instance, the bioactivity reactionsin silica-based glasses occur in a few minutes [6], whereasthose in HA take several days [5]. A very important char-acteristic of the silica-based glasses is that they show agenetic control of the cellular response of osteoblasts. Ithas been observed that different genes are up regulatedwithin 48 h of exposure of primary human osteoblasts tothe ionic dissolution products of bioactive glasses [7]. Therapid bioactive behaviour of these silica-based glasses hasbeen related to the role of SiO2 or silicon in their surfacereactions and therefore on their in vivo and in vitro behav-iour. The surface reactions that occur in the bioactiveglasses allow the subsequent crystallization of apatite crys-tals, cell adhesion and collagen formation [8,9]. Silicon isbelieved to be essential in skeletal development; the firstindications of a physiological role for silicon were reportedby Carlisle [10], who observed that silicon was involved inthe early stage of bone calcification. Schwarz and Milne
332 S. Padilla et al. / Acta Biomaterialia 2 (2006) 331–342
[11] performed similar studies and showed that silicon defi-ciency in rats resulted in skull deformations.
Some calcium phosphates such as a- and b-tricalciumphosphate (TCP), unlike HA, are usually classified as‘‘resorbable materials’’ [12,9] and they are designed to begradually eliminated over a period of time, being simulta-neously replaced by the natural host tissue. Several authorshave shown that composites formed by sintering of HAwith a small addition of glass (phosphate or silica-based),produce the transformation of HA into a- and b-TCP; thisprocess depends on the glass composition and the thermaltreatment. In addition, phosphate-based composites mayexhibit greatly improved mechanical properties and bioac-tive behaviour compared to pure HA [13–15].
Because of the importance of silicon in bone formationand mineralisation, and the possibility of producing solublecalcium phosphate phases, it is very interesting to study thebehaviour of mixtures of HA and silicate-based glasses.However, few articles have been published about this typeof material [16,17].
In this work, mixtures of HA and a silicate-based glass(55%SiO2–41%CaO–4%P2O5; 55S), prepared by the sol–gel method, were sintered at 1300 �C. The influence ofthe glass content (5 and 20 wt.%) on the porosity, micro-structure and the constituent phases of the final materials,was studied. The influence of these factors on the in vitrobioactive behaviour of the obtained materials was alsoinvestigated. In addition, the in vitro biocompatibility ofthese materials was evaluated.
2. Experimental
2.1. Samples preparation and characterisation
The 55S powder was prepared by the sol–gel method.Tetraethyl orthosilicate, calcium nitrate tetrahydrate andtriethyl phosphate were used as oxide precursors. The syn-thesis was carried out as described elsewhere [18]. The driedgel was milled and sieved and the fraction of particles withsizes ranging between 32 and 63 lm was selected. Afterthat, the gel was treated at 700 �C for 3 h (55S-700).
The HA was obtained by reaction between a Ca(OH)2
suspension and a H3PO4 solution, as previously reported[19]. The resulting HA was calcined at 1200 �C for 1 hand afterwards it was dry milled for 20 h (HA-1200).
The HA-1200 and 55S-700 powders were dry mixed in apowder mixer for 24 h. Mixtures containing 5 and 20 wt.%of 55S-700 (M95-5 and M80-20 samples, respectively) wereprepared. Pellets (13 mm of diameter and 2 mm in height)were obtained from 0.5 g of the mixture powder by meansof 55 MPa of uniaxial and 150 MPa of isostatic pressure.After that, the pellets were heat treated at 700 �C (5 �C/min) for 3 h and subsequently heated at 1300 �C (5 �C/min) for 24 h.
The raw powders and the pieces obtained were charac-terised by X-ray diffraction (XRD) and scanning electronmicroscopy coupled with energy dispersive spectroscopy
(SEM-EDS). The XRD study was carried out using aPhilips X’Pert MPD diffractometer using CuKa radiationin the range of 5–120� 2h with a step size of 0.02� and atime per step of 10 s. The phase quantification was doneby the Rietveld method [20] using the software X’Pert Plus(Philips). The SEM study used a JEOL 6400 Microscope-Oxford Pentafet super ATW system. The porosity studyinvolved Hg intrusion porosimetry in a MicromeriticsASAP2010 porosimeter.
2.2. In vitro bioactivity assay in simulated body fluid
In vitro assay was performed by soaking the disks, ver-tically supported in a platinum scaffold, in simulated bodyfluid (SBF) [21] (pH = 7.3) at 37 �C. The geometric surfaceto solution volume ratio was 0.075 cm�1. The samples weresoaked in SBF for 3 h, 1, 3, 5 and 7 days. Both the changesin the ionic concentration of the SBF and the surface of thesamples were studied. The Ca2+ concentration and pHwere measured in an Ilyte Na+, K+, Ca2+, pH system.The formation of apatite-like layer on the pellets surfacewas investigated by XRD, SEM-EDS analyses, asdescribed above, and by Fourier transform infrared spec-troscopy (FTIR) using a Nicolet Nexus spectrometer.
2.3. In vitro biocompatibility assay
Cell spread, proliferation and cytotoxicity assays werecarried out in order to investigate the biocompatibility ofthe obtained materials. A human osteoblastic-like cell line,denoted as HOS (ECACC no. 87070202), was used. Fourpieces were used in each assay and they were sterilised bydry heat at 180 �C for 24 h.
The cells were cultured in a complete medium, whichcontained 2 mM glutamine, 100 U/mL penicillin, 100 lg/mL streptomycin and 10 vol.% foetal calf serum in Dul-becco’s modified Eagle medium. Culturing of the cellswas carried out at 37 �C in a humidified atmosphere of95% air and 5% CO2 (standard conditions).
Cells were seeded onto the materials surface in 24-wellculture plates at a density of 5 · 103 cells/cm2 per well inthe proliferation assay and at 2.5 · 103 cells/cm2 in thespreading and cytotoxicity assays. These assays were car-ried out in the complete medium supplemented with freshlyprepared ascorbic acid solution (50 lg/mL) and b-glycero-phosphate (10 mM) [22,23] and incubated in standard con-ditions. The culture medium was renewed after 3 days.Tissue culture plastic was used as the control surface.
In order to study the behaviour (attachment and spread-ing) of the osteoblasts onto the surface of the samples, themorphology of the cells was analysed by SEM at 1 h, 6 h,1 day and 3 days.
The reduction of the MTT reagent (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) was taken as anindicator of the cell proliferation [24].
Lactate dehydrogenase (LDH) released from the osteo-blastic-like cells to the culture medium was taken as an
Fig. 1. XRD patterns of: (a) HA calcined at 1200 �C and the glass treatedat 700 �C and (b) HA, 55S, M95-5 and M80-20 after sintering at 1300 �Cfor 24 h.
Table 1Crystal phase analysis of HA, 55S, M95-5 and M80-20 samples treated at1300 �C for 24 h (wt.%)
S. Padilla et al. / Acta Biomaterialia 2 (2006) 331–342 333
indicator of the cytotoxicity of the materials. The measure-ments were made at 3 days of seeding. The determinationof the LDH was made using a commercially available kit(LDH, Pyruvate, Kinetic UV. DGKC. Spinreact SA,Spain).
In order to estimate the amount of LDH per lysed cells,the amount of LDH released when all seeded cells werelysed was determined. For this purpose the seeded cellswere destroyed by three freeze–defrost cycles and theLDH determined as described above.
The statistical study was carried out by analysis of var-iance. The Scheffe test was used for evaluations of statisti-cal differences among groups. In all statistical evaluations,p < 0.05 was considered as statistically significant.
3. Results
3.1. Powder and pieces characterization
The XRD pattern of HA powder treated at 1200 �C andthe one of the glass stabilized at 700 �C are shown inFig. 1(a). The HA-1200 showed the diffraction maximacorresponding to a hydroxyapatite phase [25] (JCPDS #9-0432) and, in addition, two diffraction maxima assignedto CaO (JCPDS # 43-1001) were observed. The CaO for-mation was due to the decomposition of the raw carbonatehydroxyapatite [26]. The 55S powder treated at 700 �Cshowed a XRD pattern characteristic of the amorphousmaterials, but the broad maximum centred at 32� (2h) indi-cates the crystallization of a small nucleus of an apatitephase [27].
Fig. 1(b) shows the XRD patterns of HA, 55S and themixtures after the thermal treatment at 1300 �C for 24 h(HA-1300, 55S-1300, M95-5, M80-20). In Table 1 theresults of crystalline phase quantification are collected.
HA powder calcined at 1300 �C shows diffraction max-ima corresponding to HA and tetracalcium phosphate(JCPDS # 25-1137) (Ca4P2O9, TTCP, 16 wt.%) whereasthe CaO maxima were not observed. However, in themixtures of HA/glass, TTCP phase was not detected buta-TCP (JCPDS #70-0364) and b-TCP (JCPDS # 70-2065) (Ca3(PO4)2) appeared. In M80-20 sample a notice-able increase of these phases took place, b-TCP being themain one (Table 1). In addition, silicocarnotite (JCPDS #73-1181) (Ca5(PO4)2SiO4; SiCA) crystallized in M95-5 sam-ple whereas this phase was not observed in theM80-20 sample in which pseudowollastonite (JCPDS #74-0874) (a-CaSiO3, PsW) was present instead. In 55S-1300, PsW, b-TCP and cristobalite (JCPDS # 76-0939)(SiO2, Ct) were identified as crystalline phases.
Fig. 2 shows the surface micrographs of mixtures heatedat 1300 �C as well as the distribution of elements, deter-mined by EDS. In general, the elements are quite homoge-neously distributed, although small and isolated zones withdifferent silicon and phosphorus contents are observed. InFigs. 2–4 it can be observed that samples present a surfacecommonly found in sintered materials. Generally, the
morphology was similar in all material surfaces, but someparticular zones with a different morphology and composi-tion were observed. In areas of the M95-5 sample (Fig. 3, 0d-zone 2) the Si content was much higher than in most ofthe remaining surface (Fig. 3, 0 d-zone 1). The proportionsof Si, Ca and P and the XRD results suggest that theseminority zones are rich in SiCA. In the M80-20 sample,in addition to the one of the majority zones, two additionalmorphologies were observed (Fig. 4, 0 d). One of them
Fig. 2. SEM micrographs and element distribution of M95-5 and M80-20samples treated at 1300 �C for 24 h.
334 S. Padilla et al. / Acta Biomaterialia 2 (2006) 331–342
(zone 3), showing a rough aspect, only contained Si andCa, whereas in the other one, with a smooth aspect (zone2), only Ca and P were present. According to the XRDresults, these zones must correspond to PsW and calciumphosphate phases, respectively.
Fig. 5 compares the fracture surface obtained by SEMand the curves of Hg intrusion and pore size distributionof the pieces of HA-1300 and 55S-1300 with those of theM95-5 and M80-20 samples. As can be observed, HA-1300 shows a homogeneous and dense microstructurewhereas the one corresponding to 55S-1300 is heteroge-neous and porous. The M95-5 sample shows dense zonessimilar to the ones observed in HA-1300 but other areaswith higher porosity could be observed. The M80-20 sam-ple shows a very porous microstructure quite different tothe one observed in HA-1300 and 55S-1300. The pore vol-ume and pore size increased in both mixtures with respectto HA-1300 pieces being significantly higher in the M80-20sample. This increase was higher than that theoreticallyexpected by a simple mixture of the HA and the sol–gelglass.
3.2. In vitro bioactivity assay in SBF
Fig. 6 shows the changes in the Ca2+ concentration andpH as a function of the soaking time in SBF. Both samplesshow a sharp increase of the Ca2+ concentration and pHuntil 1 day of immersion. As the soaking time increasedthe Ca2+ concentration decreased and the pH remainedconstant. The changes in Ca2+ concentration and pH weresignificantly higher in M80-20 sample than in M95-5.
The surface micrographs of M95-5 and M80-20 samplesbefore and after immersion in SBF are shown in Figs. 3and 4, respectively. A new layer covered the surface of bothsamples after soaking in SBF, but some differences in thenew layer growth could be observed between them. In theM95-5 sample no changes were observed until 3 days ofsoaking. At this time, almost the entire surface was coveredby a new layer formed by spherical particles constituted byhundreds of small needle-like aggregates. The EDS analysisshowed that this layer was composed of Ca, P, Mg, Cl andNa, whereas in the uncovered zones only Ca, P and Si weredetected. At longer times the layer grew but, even at 7 daysof immersion, small-uncovered zones, with the same com-position as after 3 days, were observed.
In the M80-20 sample the layer formation occurred in adifferent way and faster than in the M95-5 sample. In thissample, the formation of small crystals was observed after3 h of soaking. The EDS showed that these crystals wereformed on the zones with the highest silicon content. After1 day of soaking, a layer composed of spheres of crystallineaggregates was observed, but this layer did not cover theentire surface. The uncovered zones contained Ca, P andSi, but the Si content was lower than in the covered zone.The layer and the size of the crystalline aggregates grewwith the immersion time. After 7 days, small zonesremained uncovered but the composition was different tothe similar ones observed in the M95-5 sample. In this casesilicon was not present, whereas in the M95-5 sample theseareas showed the highest Si content.
Fig. 7 shows the cross sections of M95-5 and M80-20samples after soaking in SBF for 7 days. The layer thick-ness was similar for both samples, around 3 lm, in spiteof the faster layer formation observed in the M80-20sample.
The FTIR spectra of both samples after 7 days soaking(Fig. 8) were similar to the initial ones but new bands at1415 and 1500 cm�1, attributed to carbonate groups, weredetected. In the M80-20 sample a decrease of the intensityof the bands corresponding to PsW (938, 720 and472 cm�1) was also observed.
The XRD patterns of the M95-5 and M80-20 samples,before and after immersion in SBF, are shown in Figs. 9and 10, respectively. No significant changes are observedin the XRD patterns of the M95-5 sample as the immersiontime progresses due to the similar composition of the newlayer formed and the HA present in the starting material.The bulk XRD pattern after 7 days showed the samephases as the initial one, whereas in the glancing angle
Fig. 3. SEM micrographs and EDS spectra of the surface of M95-5 sample treated at 1300 �C for 24 h before (0 d) and after soaking in SBF for differenttimes.
S. Padilla et al. / Acta Biomaterialia 2 (2006) 331–342 335
XRD a widening of the apatite maxima were observed andonly maxima corresponding to this phase were observed.
In the M80-20 sample, significant changes were observedin the XRD patterns after immersion in SBF. Glancingangle and bulk XRD patterns showed a great decrease ofthe maxima corresponding to b-TCP and PsW after 1 dayof soaking, whereas the maxima corresponding to a-TCPdid not show changes. The glancing angle XRD patternafter 7 days only showed broad maxima corresponding toHA indicating that a new apatite layer with low crystal size
was formed. An increase of the intensity of the maximumat 26.2�(h) (002) was observed with the soaking time inSBF, which is characteristic of the apatite crystallized fromsolutions simulating the human plasma.
3.3. In vitro biocompatibility assay
Fig. 11 shows the micrograph of the samples’ surfacesafter different times of seeding. After 1 h attached and par-tially spread out cells were observed. A noticeable increase
Fig. 4. SEM micrographs and EDS spectra of the surface of M80-20 sample sintered at 1300 �C for 24 h before (0 d) and after soaking in SBF for differenttimes.
336 S. Padilla et al. / Acta Biomaterialia 2 (2006) 331–342
of the cell spreading is observed with the assay time. At 6 hthe osteoblast-like cells exhibit a fusiform morphology withnumerous filopodia. After 1 day of seeding the cells show apolygonal morphology, which indicate a mitosis stage. Thisfact is more evident in the M80-20 sample. After 3 days theosteoblast-like cells are grown to confluence forming a
cellular multilayer. This multilayer is more compact andthicker in the M95-5 sample than in the M80-20 sample.
The results of the proliferation assay (Fig. 12(a)) showedthat the proliferation of the osteoblasts increased with theincubation time. The behaviour of both samples was statis-tically similar to the control.
Fig. 5. SEM micrographs of fracture surface and curves of cumulative Hg intrusion and pore size distribution of mixtures pieces, HA and 55S treated at1300 �C for 24 h.
S. Padilla et al. / Acta Biomaterialia 2 (2006) 331–342 337
The results of the cytotoxicity assay after 3 days(Fig. 12(b)) showed that the LDH activity was statisticallylower in the M95-5 sample than in the M80-20 sample andthe control. The test (the amount of LDH released when allseeded cells were lysed) was 5–6 times higher than the con-trol, M95-5 and M80-20 samples. This result shows thenon-cytotoxic effect for the osteoblastic-like cells.
4. Discussion
The XRD results showed that when the initial HA istreated at 1300 �C, TTCP is formed as a result of the reac-tion between HA and CaO present in the initial powder [19]but neither CaO nor TTCP were observed when the glasswas added. These results indicate that other reactionsinvolving the CaO must occur when the glass is added. Inthe mixtures, the transformation of HA into a- andb-TCP took place, and this reaction was higher as the glasscontent increased. In addition, depending on the glasscontent, different crystalline silicon-containing phases wereformed. So, in the M95-5 sample SiCA was formed,whereas in the M80-20 sample, the crystallised phase wasPsW. On the basis of these results it is possible to deducethat a reaction between HA and the glass took place.
In the mixture with low glass content, the SiO4�4 ions
could be incorporated into the HA structure and, as a con-sequence of the charge difference between the silicate and
phosphate groups, the OH� ions would be eliminated fromthe structure as water vapour. This reaction would producethe formation of SiCA and TCP phases, which have alower Ca/P ratio than HA (Eq. (1))
4Ca5ðPO4Þ3OHðsÞ þ SiO2ðsÞ ! Ca5ðPO4Þ2SiO3ðsÞ
þ 5Ca3ðPO4Þ2ðsÞ þ 2H2OðgÞ ð1Þ
The theoretical calculations for this reaction indicate thatsilicon oxide does not react completely because in that case,a higher content of TCP phases than the experimentally ob-tained value should be observed. Therefore, some siliconoxide must remain in a non-detectable form by XRD anal-ysis, i.e., as an amorphous phase, as a nanocrystallinephase, or even as a crystalline phase in a very lowproportion.
In the M80-20 sample the silicon oxide content is highenough to produce a reaction with HA giving PsW, asshown in Eq. (2)
In this sample the SiO2 content is higher than the theoret-ical one needed to produce the complete transformation ofthe HA; therefore, as previously discussed for the M95-5sample, some SiO2 does not react and could remain as anamorphous phase as observed in 55S-1300 (Table 1).
Fig. 6. Changes in Ca2+ concentration and pH as a function of the time ofimmersion in SBF in M95-5 and M80-20 samples.
Fig. 7. SEM micrographs and EDS spectra of the cross section of M95-5and M80-20 samples after soaking in SBF for 7 days.
338 S. Padilla et al. / Acta Biomaterialia 2 (2006) 331–342
The CaO present in the initial HA must react with theP2O5 of the glass yielding calcium phosphate phases, asin 55S-1300 (Fig. 2), because these phases are the first thatcrystallise in the glass [28].
The increase of the porosity observed in the mixturescompared with isolated HA-1300 or 55S-1300 could beattributed to the H2O formed in the reactions previously
500 1000 1500 2000 2500 3000 3500 4000
7 days
M95-5
% T
rans
mitt
ance
Initial
CO32-
Wavenumbers (cm-1)
Fig. 8. FTIR spectra of M95-5 and M80-20 sampl
indicated (Eqs. (1) and (2)). In addition, the presence ofthe glass seems to hinder the sintering of the mixtures com-pared to pure glass or HA. This agrees with the resultsobserved by other authors for composites of HA with sili-cate-based glasses in which a lower degree of densificationwas achieved compared with HA/phosphate-based glasscomposites [29].
Santos et al. [29] reported that, in the compositeobtained by sintering of HA with a 5 wt.% of Bioglass�
at 1350 �C, a complete transformation of HA into a-TCPand SiCA took place. Goller et al. [17] also observed thetotal transformation of HA into SiCA, Na2HPO4 Æ 7 H2Oand Ca2P2O7 Æ 4H2O in composites of HA with a 10 wt.%of Bioglass�, sintered at 1200 �C. These results do notagree with the ones obtained in the present work, becausewe have only observed a small transformation of HA(30%) in the sample with 5 wt.% of glass (M95-5) and inthe M80-20 sample 25% of HA is present. This differencecould be attributed to the presence of Na2O in the glasses
Fig. 9. XRD patterns of M95-5 sample, both bulk and glancing-angle incidences after different soaking times in SBF.
Fig. 10. XRD patterns of M80-20 sample, both bulk and glancing-angle incidences after different soaking times in SBF.
S. Padilla et al. / Acta Biomaterialia 2 (2006) 331–342 339
used in the above-mentioned work, as observed in compos-ites of HA/phosphate-based glasses. In this sense Santoset al. [29] observed a detrimental effect of the Na2O onthe stability of HA, giving rise to a high transformationof hydroxyapatite into either a- or b-TCP in comparisonwith two-oxide glass addition. Similar behaviour wasreported by Knowles [30], who studied the effect of two-oxide phosphate–glass additions on the phase stability ofsintered HA.
Regarding the in vitro bioactivity study, the SEM-EDSand XRD results evidenced that the different crystallisedphases which correspond to areas with different morphol-ogy, played an important role in the bioactive behaviour.
In this regard, in the M80-20 sample a higher increase inthe Ca2+ concentration and pH than in the M95-5 samplewas observed when soaking in SBF. According to XRDand FTIR results this behaviour can be attributed to thedissolution of PsW and b-TCP phases. In addition, the
Fig. 11. SEM micrographs of M95-5 and M80-20 samples after 1 h, 6 h, 1 day and 3 days of seeding.
340 S. Padilla et al. / Acta Biomaterialia 2 (2006) 331–342
highest porosity of this sample would favour the dissolu-tion of these phases and therefore the highest ionic changesobserved.
The samples studied in the present work showed anin vitro bioactive behaviour i.e., an apatite-like layer wasformed on their surface when soaked in SBF. The resultsof SEM, EDS, FTIR and XRD analyses showed that thelayer formed on the samples surface is a poorly crystallisedapatite with ionic substitution, such as CO2�
3 , Na, Mg andCl, similar to the biological apatites [31].
In these materials, the layer began to grow on preferen-tial zones and small zones remained uncovered even after 7days of immersion. This behaviour was different to thatpreviously observed in similar mixtures [32,33] treated at700 �C whose surface was entirely covered since the initialperiod of immersion in SBF. These differences could be dueto the presence of crystalline phases with different reactivity
in the M80-20 and M95-5 samples, whereas the mixturestreated at 700 �C are a simple and homogeneous biphasicmixture of HA and glass.
In the M80-20 sample the layer began to grow in theareas with high silicon content and then the growth contin-ued on other zones of the surface that also contain silicon.However, after 7 days soaking some zones that only con-tain Ca and P remained uncovered.
The nucleation and growth mechanism of the apatite-likelayer on bioactive materials containing CaO–SiO2, as firstproposed by Kokubo et al. [34], propose that an inter-change takes place between the Ca2+ ions of the materialand the H3O+ of solution, giving rise to the formation ofSi–OH groups on the material surface that induce the apa-tite nucleation. The nuclei thus formed later grow at theexpense of the ions in the solution that has been saturatedwith respect to the apatite. Therefore, the silicon seems to
Fig. 12. Results of the: (a) cell proliferation and (b) cytotoxicity assays at3 days of seeding (the test is the amount of LDH released when all seededcells were lysed).
S. Padilla et al. / Acta Biomaterialia 2 (2006) 331–342 341
be very important to increase the in vitro bioactivebehaviour.
As previously mentioned with regard to the M80-20sample, the layer began to grow in areas containing silicon,rather than another soluble phase (b-TCP) present in thissample, which is known to be a bioactive phase. Therefore,as has been mentioned in the introduction section, the sili-con played a very important role in the bioactive behaviourof the materials.
According to the above mentioned mechanism, in M95-5 sample the layer must begin to grow on areas rich in sil-icon; however, this is not observed and, in addition, theseareas remained uncovered even after 7 days soaking. Inthese zones, unlike the M80-20 sample, the silicon is pres-ent as SiCA, which is a less soluble and reactive phase thanPsW, as evidenced by the XRD and SEM-EDS, and there-fore the nucleus for the layer growth could not be formedon this phase. For this reason, in the M95-5 sample, thelayer began to grow in the areas with low silicon contentbut present as a more soluble phase.
The lower silicon content and lower solubility of thephases present in the M95-5 sample caused a delay in the
layer formation when compared with the M80-20 sample.However, once the first apatite nuclei were formed, thelayer grew quickly and after 7 days the layer thicknesswas similar in both samples (Fig. 9).
It is worth noting that similar studies of bioactivity per-formed with 55S-1300 [28] and HA [33] materials showedthe formation of the apatite layer after 3 days in 55S-1300 and even at 60 days no signal of bioactive behaviourwas observed in HA.
According to the results obtaining in this work, thein vitro bioactive behaviour of HA can be significantlyincreased by the addition of a silicate-based glass and asynergistic effect between HA and glass occurs in the mix-tures. This fact is directly related to the silicon-containingphases present in the obtained materials that initiate themechanism of the apatite layer formation and HA actingas the nucleus for the apatite grow. In addition, the heattreatment of mixtures of HA and a silicon–phosphate glassgive rise the formation of a- and b-TCP phases, which aremore soluble than HA.
In addition the biocompatibility assay showed that thematerials are biocompatible with osteoblastic-like cells.The attachment of cells, spread and proliferation occur inboth materials in a similar way to the control. The cyto-toxic effect was also similar to the control. These resultsindicate that the material does not affect the osteoblastic-like cells development.
The M80-20 sample shows a faster cellular developmentthan material M95-5; however, at longer times M95-5 hadan even better cellular development (higher cell spread,proliferation and lower LDH activity than the M80-20sample and control). This result agrees with the bioactivestudy in SBF as previously mentioned.
5. Conclusions
Highly in vitro bioactive and biocompatible materials,composed of HA, a-TCP, b-TCP and SiCA or PsW, havebeen obtained by sintering at 1300 �C of mixtures of HAand 5 or 20 wt.% of a silicate-based glass.
Reactions between the glass and HA took place,producing TCP phases and Si-containing phases as wellas increases in the porosity, pore size and a heteroge-neous microstructure, which was related to the glasscontent.
A synergistic effect between HA and glass took place inthe mixtures, in that their in vitro bioactive behaviour wasfaster than in HA and glass themselves.
The layer formation rate was different in each sampleand was related to the constituent phases present in thematerials obtained. In both samples the layer began togrow in areas containing the most soluble phase of silicon,which was different in each sample, with PsW the mostfavourable phase.
By changing the amounts of glass, the porosity, solubi-lity and the in vitro bioactive behaviour of the materialscan be tailored to reflect the clinical requirement.
342 S. Padilla et al. / Acta Biomaterialia 2 (2006) 331–342
Acknowledgements
The financial support of CICYT, Spain, through re-search projects MAT02-0025 is acknowledged. Authorsalso thank CAI electron microscopy and CAI X-ray dif-fraction, UCM, for valuable technical and professionalassistance.
References
[1] Hulbert SF, Hench LL, Forbers D, Bowman LS. History ofbioceramics. Ceram Int 1982;8:131–40.
[2] Ducheyne P, Radin S, King L. The effect of calcium phosphateceramic composition and structure on in vitro behavior. I. Dissolu-tion. J Biomed Mater Res 1993;27:25–34.
[3] Schepers E, de Clercq M, Ducheyne P, Kempeneers R. Bioactive glassparticulate material as a filler for bone lesions. J Oral Rehab1991;18:439–52.
[4] Vallet-Regı M, Ragel CV, Salinas AJ. Glasses with medical applica-tions. Eur J Inorg Chem 2003;6:1029–42.
[5] Oonishi H et al. Quantitative comparison of bone growth behavior ingranules of bioglass�, A-W glass–ceramic, and hydroxyapatite. JBiomed Mater Res 2000;51:37–46.
[6] Zhong J, Greenspan DC. Processing and properties of sol–gel bioactiveglasses. J Biomed Mater Res (Appl Biomat) 2000;53:694–701.
[7] Xynos D, Edgar AJ, Buttery LDK, Hench LL, Polak JM. Gene-expression profiling of human osteoblasts following treatment withthe ionic products of Bioglass� 45S5 dissolution. J Biomed Mater Res2001;55:151–7.
[8] Hench LL, Clark AE. Adhesion to bone. In: Williams DF, editor.Biocompatibility of orthopaedics implants, vol. II. Boca Raton,FL: CRC Press; 1982. p. 129–70.
[9] Hench LL. Bioceramics: from concept to clinic. J Am Ceram Soc1991;74:1487–510.
[10] Carlisle EM. Silicon: a possible factor in bone calcification. Science1970;167:179–80.
[11] Schwarz K, Milne DB. Growth promoting effects of silicon in rats.Nature 1972;239:333–4.
[12] Ducheyne P, Radin S, King L, Ishikawa K, Kim CS. Bioceramics. In:Bonfield W, Hasting GW, Tanner KE, editors. Proceedings of the 4thInternational Symposium on Ceramics in Medicine. London: But-terworth-Heinemann; 1991. p. 135–44.
[13] Oktar FN, Goller G. Sintering effects on mechanical properties of glass-reinforced hydroxyapatite composites. Ceram Int 2002;28:617–21.
[14] Lopes MA, Santos JD, Monteiro FJ, Ohtsuki C, Osaka A, Kaneko S,et al. Push-out testing and histological evaluation of glass reinforcedhydroxyapatite composites implanted in the tibia of rabbits. J BiomedMater Res 2001;54:463–9.
[15] Kalita SJ, Bose S, Hosick HL, Bandyopadhyay A. CaO–P2O5–Na2Obased sintering additives for hydroxyapatite (HAp) ceramics. Bioma-terials 2004;25:2331–9.
[16] Tancred DC, Carr AJ, McCormack BAO. The sintering andmechanical behaviour of hydroxyapatite with bioglass additions. JMater Sci Mater Med 2001;12:81–93.
[17] Goller G, Demirkiran H, Oktar FN, Demirkesen E. Processing andcharacterization of Bioglass reinforced hydroxyapatite composites.Ceram Int 2003;29:721–4.
[18] Vallet-Regı M, Ramila A. New bioactive glass and changes inporosity during the growth of a carbonate hydroxyapatite layer onglass surfaces. Chem Mater 2000;12:961–5.
[19] Padilla S, Garcıa-Carrodeguas R, Vallet-Regı M. Hydroxyapatitesuspensions as precursors of pieces obtained by gelcasting method. JEur Ceram Soc 2004;24:2223–32.
[20] Hill RJ, Howard CJ. Quantitative phase analysis from neutronpowder diffraction data using the Rietveld method. J Appl Cryst1987;20:467–74.
[21] Kokubo T, Kushitani H, Sakka S, Kitsugi T, Yamamuro T. Solutionsable to reproduce in vivo surface-structure changes in bioactive glass-ceramic A-W3. J Biomed Mater Res 1990;24:721–34.
[22] Buttery LD, Bourne S, Xynos JD, Wood H, Hughes FJ, Hughes SP,et al. Differentiation of osteoblasts and in vitro bone formation frommurine embryonic stem cells. Tissue Eng 2001;7:89–99.
[23] Skillington J, Choy L, Derynck R. Bone morphogenetic protein andretinol acid signalling cooperate to induce osteoblast differentiation ofpreadipocytes. J Cell Biol 2002;159:135–46.
[24] Slater TF, Sawyer B, Strauli V. Studies on succinate–tetrazoliumreductace system III. Point of coupling of four different tetrazoliumsalts. Biochim Biophys Acta 1963;77:383–93.
[25] JCPDS # 9-0432 Database sets 1-49 plus 70-86. International Centerfor Diffraction Data, Newton Square, PA, 1999.
[26] Padilla S, Vallet-Regı M, Ginebra MP, Gil FJ. Processing andmechanical properties of hydroxyapatite pieces obtained by thegelcasting method. J Eur Ceram Soc 2005;25:375–83.
[27] Perez-Pariente J, Balas F, Vallet-Regı M. Surface and chemical studyof SiO2–P2O5–CaO–(MgO). Chem Mater 2000;12:750–5.
[28] Roman J, Padilla S, Vallet-Regı M. Sol–gel glasses as precursors ofbioactive glass ceramics. Chem Mater 2003;15:798–806.
[29] Santos JD, Knowles JC, Reis RL, Monteiro FJ, Hastings GW.Microstructural characterization of glass-reinforced hydroxyapatitecomposites. Biomaterials 1994;15:5–10.
[30] Knowles JC. Development of hydroxyapatite with enhanced mechan-ical properties: effect of high glass additions n mechanical propertiesand phase stability of sintered hydroxypatite. Brit Ceram Trans1994;93:100–3.
