Available online at www.sciencedirect.com

Acta Biomaterialia 5 (2009) 134–143

www.elsevier.com/locate/actabiomat

Fabrication and cellular biocompatibility of porous carbonatedbiphasic calcium phosphate ceramics with a nanostructure

Bo Li a,*, Xuening Chen a, Bo Guo b, Xinlong Wang c, Hongsong Fan a,*, Xingdong Zhang a

a National Engineering Research Center for Biomaterials, Sichuan University, Chengdu 610064, Chinab West China Eye Center of Huaxi Hospital, Sichuan University, Chengdu 610064, China

c College of Chemical Engineering, Sichuan University, Chengdu 610064, China

Received 4 March 2008; received in revised form 25 July 2008; accepted 31 July 2008Available online 22 August 2008

Abstract

Microwave heating was applied to fabricate interconnective porous structured bodies by foaming as-synthesized calcium-deficienthydroxyapatite (Ca-deficient HA) precipitate containing H2O2. The porous bodies were sintered by a microwave process with activatedcarbon as the embedding material to prepare nano- and submicron-structured ceramics. By comparison, conventional sintering was usedto produce microstructured ceramics. The precursor particles and bulk ceramics were characterized by transmission electron microscopy(TEM), dynamic light scattering, scanning electron microscopy (SEM), X-ray diffraction (XRD), Fourier-transformed infrared spectros-copy (FTIR) and mechanical testing. TEM micrographs and assessment of the size distribution showed that the needle-like precursorparticles are on the nanoscale. SEM observation indicated that the ceramics formed by microwave sintering presented a structure ofinterconnective pores, with average grain sizes of �86 and �167 nm. XRD patterns and FTIR spectra confirmed the presence of carbon-ated biphasic calcium phosphate (BCP), and the mechanical tests showed that the ceramics formed by microwave sintering had a com-pressive strength comparable to that obtained by conventional methods. Rat osteoblasts were cultured on the three kinds of BCPceramics to evaluate their biocompatibility. Compared with the microscale group formed by conventional sintering, MTT assay andALP assay showed that nanophase scaffolds promoted cell proliferation and differentiation respectively, and SEM observation showedthat the nanoscale group clearly promoted cell adhesion. The results from this study suggest that porous carbonated biphasic calciumphosphate ceramics with a nanostructure promote osteoblast adhesion, proliferation and differentiation. In conclusion, porous carbon-ated BCP ceramics with a nanostructure are simple and quick to prepare using microwaves and compared with those produced by con-ventional sintering, may be better bone graft materials.� 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Keywords: Nanostructure; Porous; Carbonated; Biphasic calcium phosphate; Microwave

1. Introduction

With the development of nanotechnology, bone apatitehas been proved to consist of nanosized carbonated cal-cium phosphate crystals [1]. Many studies have shown thatbone-forming cells are accustomed to interacting withnanoscale surfaces of biomaterials, and this nanoscale fea-ture of implanted materials is critical to keep the body from

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

doi:10.1016/j.actbio.2008.07.035

* Corresponding author. Tel.: +86 28 85410703; fax: +86 28 85410246.E-mail addresses: leewave@126.com (B. Li), hsfan68@hotmail.com

(H. Fan).

rejecting artificial parts [2,3]. It has gradually been recog-nized that a nanoscale surface could promote the adhesion,proliferation and differentiation of osteoblasts [4,5]. Inaddition, the bony apatite contains a significant amountof carbonate ions (3–8 wt.%) [6]. According to the previousstudies preparing carbonated hydroxyapatite (HA,Ca10(PO4)6(OH)2), the carbonate groups would increaseits bioactivities [7,8]. Therefore, porous apatite ceramicsmimic bony apatite in chemical and structural compositionmay be better bone grafts.

Porous BCP ceramics containing HA and tricalciumphosphate (TCP, Ca3(PO4)2) phases as an ideal bone graft

vier Ltd. All rights reserved.

Fig. 1. A flow chart of fabricating porous BCP bioceramics.

B. Li et al. / Acta Biomaterialia 5 (2009) 134–143 135

substitute have recently attracted great attention due to thesimilarity of their chemical composition and porous struc-ture with bone mineral [9]. The conventional foamingmethod of fabricating an interconnective porous structureis complex and time consuming [10], which limits the appli-cation of porous BCP ceramics. On the other hand, the sin-tering method of porous biphasic calcium phosphate (BCP)ceramics is generally based on a conventional heating pro-cess with a long time and a high temperature, which pro-duces a grain size greater than that of bone apatitecrystal and reduces the bioactivity [11].

A microwave process has been applied to fabricateceramics, especially nanosized or sub-microsized ceram-ics, in past decade [12,13]. This novel method has severaladvantages over conventional sintering, such as a lowersintering temperature (100 �C lower) and a shorter sinter-ing time (only a few minutes). However, CaP ceramicslack the ability to absorb the microwave power under800 �C. Therefore, it is necessary to choose a properembedding material to help to elevate temperature under800 �C.

Many previous studies have focused on the fabricationof nanosized CaP powder and dense CaP ceramics andthe evaluation of their cellular biocompatibility [14,15].To the best of our knowledge, few studies have preparedporous nanosized BCP ceramics. In this study, we used awet precipitation method to synthesize nanosized Ca-defi-cient HA as a precursor. After aging and washing, theas-synthesized precipitate was directly foamed with H2O2,assisted by microwave heating. Subsequently, anothermicrowave sintering process with activated carbon as theembedding material was conducted to produce porous car-bonated BCP ceramics with nanosized or sub-micron-sizedgrains. Conventional sintering was used to produce micro-structured BCP as control. The current study aimed to pre-pare the novel bone graft and compare its cellularbiocompatibility with that of conventional porousceramics.

2. Materials and methods

2.1. Preparation of nanosized Ca-deficient HA precursor

The flow chart shown in Fig. 1 outlines the experimentalprocedure used to process porous BCP ceramics in thisstudy.

Ca-deficient HA was synthesized as the precursor by thewet precipitation method. Analytically pure calcium nitrate(Ca(NO3)2�4H2O) and diammonium phosphate((NH4)2HPO4) were dissolved into deionized water withconcentrations of 0.5 and 0.3 M, respectively, and the ini-tial Ca/P ratio was 1.5. Next, 3 wt.% citric acid (CA) wasadded to Ca(NO3)2 solution as a dispersant, and the pHvalue was kept at 9.5 with ammonia solution. (NH4)2HPO4

solution was dropped into the solution of Ca(NO3)2 withcontinuous stirring at 55 �C. After adding (NH4)2HPO4,the suspension was stirred for another hour then aged for

36 h at 25 �C. The white precipitation was centrifugedand washed with deionized water for six times. Subse-quently, 0.2 wt.% analytically pure methylcellulose (MC)and 0.4 wt.% polyethylene glycol with a molecular weightof 6000 (PEG6000) were added to the precipitate as disper-sant and viscous agent. With 6 vol.% analytically pureH2O2 as a vesicant, the slurry was heated by microwave(domestic microwave oven, Galanz, China) for 30–60 sand the foaming slurry was poured immediately into themolds with good permeability. After drying at 48 �C for12 h and removing the organic addition at 550 �C for 6 h,porous green bodies were produced.

2.2. Preparation of porous ceramics by microwave sintering

and conventional sintering

The as-prepared porous green bodies were transferredinto a microwave furnace for sintering. The set-up of themicrowave-sintering furnace is shown in Fig. 2. The micro-wave sintering furnace was refitted from the domesticmicrowave oven (2.45 GHz/750 W, Sanyo, Japan). Sam-ples were embedded by activated carbon filled in a hollowporous mullite cube. The temperature of specimens inmicrowave furnace was measured by an optical fiberpyrometer in the range of 700–1400 �C. The BCP greenbodies were sintered by microwave at 950 and 1050 �Cfor 1 min to obtain nanosized BCP ceramics (NBCP) and

Fig. 2. Set-up of the microwave sintering furnace.

136 B. Li et al. / Acta Biomaterialia 5 (2009) 134–143

submicron-sized BCP ceramics (MBCP), respectively. Bycomparison, conventional sintering with the same greenbodies was conducted at 1100 �C with a heating rate of5 �C min�1 and a holding time of 2 h to obtain micron-sized BCP ceramics (CBCP). The three kinds of cylinder-like samples were cut into disks with a diameter of 9 mmand a height of 2 mm. Subsequently, disks were thoroughlywashed with ultrapure water several times. Finally, diskswere dried and autoclaved prior to use.

2.3. Material characterization

The microstructures of precursor particles were exam-ined by a transmission electron microscope (TEM; JEM-100cx, Japan) operated at 200 kV. The particle size distri-bution and the mean particle size of the precipitate weremeasured by the dynamic light scattering technique (Mas-tersizer nanoplus, Malvern, UK) using deionized water asthe dispersant. The compressive strengths were measuredwith a testing machine (SLP-5 Biomechanical TestingMachine, Chaoyang Instrument, Changchun, China) ata loading rate of 0.5 mm min�1, and seven specimenswere fractured for each test condition. The porosity ofthe sintered cylinder was determined by the Archimedesmethod, with deionized water as the immersion medium.Microstructural observations of the sintered ceramicswere conducted using a scanning electron microscope(SEM; JSM-5900 LV, JEOL, Japan). The precursor par-ticle sizes and grain sizes of sintered cylinders were calcu-lated using image analysis of TEM and SEMmicrographs, respectively. The width and length of 50crystals or grains were measured from five randomlyselected areas of each sample, and the average widthand length were calculated from measurements of those50 grains or crystals [16]. The phase composition of sin-tered BCP was determined by X-ray diffraction (XRD;X’pert Pro MPD, Philips) with Cu Ka radiation overthe 2h range of 20–50�. FTIR (FTIR 1750, Perkin-Elmer,USA) was performed to determine the presence of car-bonate anions partially substituting for PO4

3� and/orOH� groups in BCP ceramics.

2.4. Cell culture

2.4.1. Cell morphological studies

Rat osteoblast cells (ROS17/2.8) were cultured in com-plete Dulbecco’s modified Eagle’s medium (DMEM) con-taining 10% fetal bovine serum (FBS), 1% antibiotics(penicillin, streptomycin, amphotericin) and 0.85 mMascorbic acid–2 phosphate at 37 �C in a humidified incuba-tor with 5% CO2. Cells harvested with 0.25% trypsin–EDTA solution in phosphate-buffered saline (PBS, pH7.4) from cell culture flasks were resuspended in culturemedium and seeded at a concentration of 40,000 cells cm�2

onto the disks of NBCP, MBCP and CBCP. Cells werethen incubated in 24-well plates for 1, 3 and 7 days. Atthe pre-determined time points, samples were rinsed inPBS to remove any nonadherent cells and culture medium.The remaining cells were fixed in 1.5% glutaraldehyde in0.1 M phosphate buffer for 30 min at 4 �C then dehydratedthrough a series of ethanol concentrations (50, 70, 90 and100%) and dried using a critical point drier (HCP-2, Hit-achi, Japan). Once dried, the samples were coated withgold and examined using SEM.

2.4.2. Cell proliferation

MTT assay was used to determine cell proliferation.Cells were seeded at a concentration of 40,000 cells cm�2

onto the disks of NBCP, MBCP and CBCP. Cells weregrown on the disks for 1, 3 and 7 days in a 37 �C incubatorwith 5% CO2. At the pre-determined time points, each diskwas transferred to wells of a new 24-well plate and 1.5 ml ofmedium was added to each disk. One hundred and fiftymicroliters of freshly prepared 5 mg ml�1 MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-dipheyltetrazolium bromide) wasadded to each well containing the disks. The plates wereplaced in an incubator at 37 �C for 3 h. The supernatantof each well was removed and acidified isopropanol(0.04 M HCl in isopropanol) was added to all wells andmixed thoroughly to dissolve the dark-blue crystals. Afterall crystals were dissolved, the plates were read on a Micro-plate reader (Ascent) with a test wavelength of 570 nmagainst a reference wavelength of 490 nm.

2.4.3. Total intracellular protein synthesis

Osteoblasts (40,000 cells cm�2) were seeded onto thedisks and cultured in DMEM supplemented with 10%FBS, 1% P/S, 50 lg ml–1

L-ascorbate (Sigma) and 10 mMb-glycerophosphate (Sigma) under standard cell cultureconditions for 7 days. Medium was replaced every 2 days.At the end of the prescribed time period, supernatant med-ium was removed and the remaining osteoblasts were lysedusing distilled water and three freeze–thaw cycles. This pro-tocol only removes intracellular and cell membrane-boundproteins [17,18]. Sodium azide (Sigma) was used to preventdegradation of proteins. Total protein content in the celllysates was determined spectrophotometrically using acommercially available kit (BCA, Pierce Chemical Co.)according to the manufacturer’s instructions. The total

B. Li et al. / Acta Biomaterialia 5 (2009) 134–143 137

intracellular protein synthesized by osteoblasts cultured onthe disks was determined from a standard curve of absor-bance vs. known concentrations of albumin run in parallelwith the experimental samples [2].

2.4.4. Alkaline phosphatase activityOsteoblasts were seeded (40,000 cells cm�2) onto the

disks and cultured in DMEM supplemented with 10%FBS, 1% P/S, 50 lg ml�1

L-ascorbate and 10 mmol b-glyc-erophosphate under standard cell culture conditions for 1,3 and 7 days. The medium was replaced every 2 days. ALPsynthesized by osteoblasts cultured on the disks was deter-mined from a standard curve of absorbance vs. known con-centrations of p-nitrophenol run in parallel withexperimental samples. Results were expressed as ALPactivity (U g�1 protein) normalized by the total proteincontent.

3. Results

3.1. Material characterization

3.1.1. Characterization of ceramics precursor

The morphology of Ca-deficient HA particles and theirselected area diffraction (SAD) patterns are shown inFig. 3A. The TEM micrograph shows that the observedparticles were needle-like. The length and width are123 ± 25 and 26 ± 5 nm, respectively, and the ratiobetween the length and width is 4.7 ± 0.8. As seen fromthe TEM micrograph, the multispot rings of the SAD pat-tern indicates that the HA exhibits nanocrystallinity. Parti-cle size distribution of the as-obtained precipitate is shownin Fig. 3B using dynamic light scattering technique. Thismethod is generally used to determine the agglomerate sizeof particles, and the particles present a monomodal distri-bution with an average particle size of 160 nm. The nano-sized Ca-deficient hydroxyapatite particles are well suitedfor use as precursors for fine ceramics produced followingsintering.

Fig. 3. TEM image with SAD pattern (A) and particle s

3.1.2. Characterization of ceramics

Following aging and washing, the as-prepared precipi-tate was directly foamed with H2O2, assisted by microwaveheating. The porous green bodies were sintered in either themicrowave furnace or the conventional furnace, and the as-obtained cylinder-like ceramics had similar linear shrink-age rates. Fig. 4A–C shows the typical morphologies ofthe ceramics thus formed. The SEM micrographs showthe three kinds of porous BCP ceramics, possessing similarsurface topography and porous structures, which mimicthe morphology of the spongy bone [19]. The Archimedesmethod showed that the porosity of NBCP, MBCP andCBCP was 75, 73 and 74%, respectively. The compressivestrength of NBCP, MBCP and CBCP is about 2.4, 2.65and 2.53 MPa, respectively. A compressive strength similarto that of conventionally prepared ceramics can beachieved by only a short period of microwave sintering.

SEM micrographs of fracture sections from both themicrowave and conventionally sintered porous BCP samplesare shown in Fig. 4D–F. Porous ceramics consisting ofunequiaxed grains with average grain sizes of 86 and167 nm were obtained by the microwave method for 1 minat 950 and 1050 �C, respectively (shown in Fig. 4D and E),while the equiaxed grain size is about 768 nm when sinteredvia the conventional method at 1100 �C (Fig. 4F). It is clearthat it is the rapid heating that is responsible for the finergrain size, since grain growth in ceramics is a function ofboth the peak sintering temperature and its soak time. Theresults of physico-chemical characterization of the ceramicssintered by the different methods are summarized at Table 1.

The XRD patterns of samples sintered by microwaveand conventional sintering are shown in Fig. 5. The threekinds of samples exhibited similar XRD patterns, and thecharacteristic peak of �31� is assigned to the presence ofb-TCP. The XRD patterns of the samples produced bymicrowave sintering show slightly wider peaks and aslightly more diffuse background, which means they havea lower crystallinity than the samples produced by conven-tional sintering. This lower crystallinity may be due to the

ize distribution (B) of Ca-deficient HA as precursor.

Fig. 4. Typical scanning electron micrographs of the porous NBCP (A,D), MBCP (B,E) and CBCP (D,F). Note the presence of macropores (A–C, 50�)and ceramics grains (D–F, 20,000�).

Table 1Physico-chemical characterization results of CaP ceramics sintered bydifferent methods

NBCP MBCP CBCP

Compressivestrength

2.4 ± 0.21 MPa 2.65 ± 0.16 MPa 2.53 ± 0.23 MPa

Porosity 75 ± 5% 73 ± 4% 74 ± 3%Macropore size 348 ± 90 lm 334 ± 106 lm 323 ± 95 lmGrain size 86 ± 20 nm 167 ± 31 nm 768 ± 321 nm

138 B. Li et al. / Acta Biomaterialia 5 (2009) 134–143

fast sintering speed in microwave sintering and the shortholding time, and indicates greater bioactivity [20].

Fig. 6 shows the typical FTIR spectra of ceramics pro-duced by microwave and conventional sintering. FTIRspectra indicate that all the samples exhibit the characteris-tic bands of phosphate groups of the apatitic structure atabout 550 and 600 cm�1 (m4), 960 cm�1 (m1) and 1020–1120 cm�1 (m3). The bands at 630 and 3540 cm�1 areassigned to apatitic OH groups. The broad bands in theregions 1600–1700 and 3200–3600 cm�1 correspond toadsorbed water. It is particularly clearly visible on the IRspectra of BCP samples that there are two regions ofcarbonate vibrations in apatites: 850–890 cm�1, associatedwith m2 vibrations of carbonate groups, and

Fig. 5. XRD patterns of BCP ceramics, note NBCP (A), MBCP (B) withmicrowave sintering method at 950 and 1050 �C for 1 min respectively andCBCP (C) at 1100 �C for 2 h with the conventional sintering method.

B. Li et al. / Acta Biomaterialia 5 (2009) 134–143 139

1400–1650 cm�1, associated with m3 vibrations of carbon-ate groups (Fig. 6A and B). The formation of carbonategroups is likely caused by the reaction of HA and activatedcarbon at high temperature [7], and the carbonated BCP ismore similar to natural bone [6].

3.2. Cell culture

3.2.1. Cell morphological studies

Fig. 7 shows the morphologies of rat osteoblast cells cul-tured on NBCP, MBCP and CBCP after 3 and 7 days.From direct SEM observation, there are more cellsattached on the NBCP surface than on the CBCP surface

Fig. 6. FTIR spectra of BCP ceramics: NBCP (A) and MBCP (B) produced byCBCP (C) produced by the conventional sintering method at 1100 �C for 2 h.

at day 3 (Fig. 7A–C). In the high-magnification image inFig. 8, different morphologies can be observed on the var-ious surfaces. Abundant filapodia are visible at the edge ofcells on both the NBCP and MBCP surfaces, in particularon the NBCP surface. Furthermore, a large number ofmini-filopodia is spread out, surrounding the cell, like hun-dreds of protrusions radiating from the cellular body orfrom the main protrusion to probe the surrounding topog-raphy. The different morphologies of the cells cultured onBCP surfaces may be attributed to different grain sizes.At day 7, all the surfaces of the three kinds of porousBCP were covered by cells (Fig. 7D–F).

3.2.2. Cell proliferation assayFig. 9 presents the absorbency value (OD value) in the

proliferation tests of the three kinds of porous BCP ceram-ics at 1, 3 and 7 days. At days 1 and 3, the OD values showlittle difference. However, at day 7, the OD value of NBCPis higher than that of MBCP and NBCP. Student’s t-testindicated that there was a significant difference betweenthe nano- and micron groups (p < 0.05). The statisticalanalysis was performed with SPSS10.0 software for Win-dows. It shows that osteoblasts on porous nanosizedBCP ceramics proliferated faster than those on conven-tional microsized BCP ceramics.

3.2.3. Alkaline phosphatase activity

In contrast to the MTT assay, there was a detectable dif-ference in ALP activity by osteoblasts cultured on all sub-strates tested in the present study after 1 day (Fig. 10).Specifically, there were obvious differences when osteo-blasts were cultured with NBCP and CBCP. Interestingly,

the microwave sintering method at 950 and 1050 �C for 1 min, respectively;

Fig. 7. SEM micrographs of rat osteoblasts cultured for 3 and 7 days on the surface of NBCP (A,D), MBCP (B,E) and CBCP (C,F). Arrow indicates cells.

140 B. Li et al. / Acta Biomaterialia 5 (2009) 134–143

the ALP activity of all osteoblasts decreased by day 3, butincreased greatly by day 7. The t-test indicated that therewas a significant difference between the nano- and microngroups at day 7 (p < 0.05). ALP is a marker that can dis-cern between non-calcium depositing and calcium deposit-ing cells [21,22]. The results show that nanosized BCPceramics promote osteoblast differentiation.

4. Discussion

Apatite is the major inorganic constituent in bone;apatite crystals are on the nanometer scale and, partlysubstituted by carbonate groups, are embedded in the

collagen matrix [6]. In this study, nanosized Ca-deficientHA particles were produced, presenting needle-like mor-phology with nanoscale dimensions, as revealed by TEMand particle size distribution examination. With thenanosized Ca-deficient HA precipitate as precursor, por-ous green bodies can be easily produced with H2O2,heated by microwaves. This is an efficient and easy foam-ing method, which prevents hard agglomeration whenthe precipitate is dried. In the sintering process, BCPceramics with homogeneous b-TCP and HA are obtainedvia the following chemical reaction at high temperature[23]. (The microwave process would accelerate the trans-formation from Ca-deficient HA to HA and b-TCP due

Fig. 8. High-magnification SEM micrographs of the surface of NBCP (A), MBCP (B) and CBCP (C) cultured with rat osteoblasts at day 3.

B. Li et al. / Acta Biomaterialia 5 (2009) 134–143 141

to its special heating model compared with conventionalsintering.)

Ca10-zðHPO4Þ2zðPO4Þ6-2zðOHÞ2 ! Ca10-zðP2O7Þz-sðPO4Þ6-2zþ2sðOHÞ2ð1-sÞ þ ðzþ sÞH2O ð1ÞCa10-zðP2O7Þz-sðPO4Þ6-2zþ2sðOHÞ2ð1-sÞ þ ðzþ sÞH2O! ð1� zÞCa10ðPO4Þ6ðOHÞ2 þ Ca3ðPO4Þ2 þ ðz� sÞH2O ð2Þ

In the microwave process, the heat is generated inter-nally within the material instead of originating from exter-nal sources [24], hence, there is an inverse and fast heatingprofile. By the aid of microwave-assisted H2O2 foaming, anabundant and interconnective porous structure was easilyproduced (Fig. 4A–C). However, CaP ceramics lack theability to absorb microwave power under 800 �C. Wedesigned a novel porous mullite insulation box filled withactivated carbon, which is favorable for elevating tempera-tures below 800 �C. Above 800 �C, the temperature wouldincrease quickly under the dual effect of calcium phosphatewith microwave power. In our study, the temperaturecould easily increase to 800 �C in just a few minutes. Thefast sintering speed by microwave sintering inhibits thegrowth of ceramic grains compared with conventional sin-tering. The microwave sintering method produces a muchfiner and more homogeneous size of grain (Fig. 4D and

E), which can result in a similar and even higher mechani-cal strength compared with that obtained by the conven-

tional method (Table 1). During the microwave process,only the major HA and b-TCP reflection peaks are pre-sented in the XRD patterns of these ceramics and no otherphase, such as calcium oxide, was observed (Fig. 5). To seeif there was any reaction between HA and activated carbonat high temperature [7], carbonate groups were introducedto the BCP crystal. FTIR spectra confirm the presence ofcarbonate groups in the samples by microwave sintering,whereas there were no carbonate groups in the samplesby conventional sintering (Fig. 6). The results proved thatporous carbonated BCP with a nanostructure could be pro-duced easily by microwave sintering combined with acti-vated carbon as an embedding material.

From a bioanalytical point of view, cells in contact witha surface will first attach, adhere and spread. This firstphase of cell adhesion and spreading is known to affectthe long-term phenotype of anchorage-dependent cells[25]. The different cellular morphologies and multiplication

Fig. 9. Rat osteoblasts proliferation on the surface of NBCP, MBCP andCBCP after 1, 3 and 7 days in culture. *Significantlt different from CBCPand NBCP (p < 0.05, n = 5).

Fig. 10. ALP activity of rat osteoblasts cultured on ceramics on thesurface of NBCP, MBCP and CBCP after 1, 3 and 7 days. *Significantfrom CBCP and NBCP (p < 0.05, n = 5).

142 B. Li et al. / Acta Biomaterialia 5 (2009) 134–143

are more likely to be caused by the microtopography ratherthan the chemical differences as observed in a previousstudy using calcium phosphate ceramics [26]. In particular,when the microtopography of the ceramics is decreased tothe nanoscale, osteoblast adhesion was independent of theceramic surface chemistry and material phase, but depen-dent on the topography of the nanophase ceramics[27,28]. One reasons for this may be the difference betweenthe specific surface areas of the three groups. The NBCPhas the smallest average grain size, under 100 nm, whichcould provide the greatest specific surface area. Nanophasematerials possess increased grain boundaries at the surface(due to the smaller grain size), which benefits cell adhesionand proliferation. It has been reported that osteoblastadhesion appears primarily at the grain boundaries [29].In other words, the contractile forces necessary for cellmigration could not overcome the strength of cell contact

points formed on highly adhesive substrate surfaces [30].Cell migration was inhibited on the substrate surfaces ofnanograins that promote cell adhesion. ALP, which isone kind of membrane-binding protein, is highly expressedin the early stage of bone maturation [31]. The high expres-sion of ALP on NBCP likely implies that porous NBCPhas the potential to improve the osteogenic rate in vivo.Compared with conventional BCP ceramics sintered at1100 �C, which are fine osteoinductive materials [20],NBCP ceramics may have better osteoinductive potential.These results showed that nanograin ceramics promote bet-ter cell adhesion, proliferation and differentiation com-pared with CBCP, which is in accord with previousstudies about dense nanosized HA [14,15,32].

5. Conclusions

The as-prepared precipitate is easily foamed under theexistence of H2O2 with the assistance of microwave heat-ing, and this method can be used to prepare scaffolds withan abundant and interconnective porous structure. Car-bonated BCP ceramics with nanograins are easily obtaineddue to the fast sintering speed of microwave sintering usingactivated carbon as an embedding material, and a compres-sive strength comparable to that obtained by the conven-tional method could be achieved in such a short time.Compared with conventional porous BCP ceramics,in vitro cell culture shows that porous carbonated biphasicphosphate ceramics with a nanostructure promote celladhesion, proliferation and differentiation. The new mate-rial may well be a better material for grafting bone andpossesses promising osteoinductive potential.

Acknowledgements

This work was financially supported by National BasicResearch Program of China (Contract No.: G2005cb623901).We thank Mrs. Hui Wang for assistance with the SEManalysis in the manuscript.

References

[1] Rubin MA, Jasiuk L, Taylor J, Rubin J, Ganey T, Apkarian RP.TEM analysis of the nanostructure of normal and osteoporotichuman trabecular bone. Bone 2003;33:270–82.

[2] Price RL, Gutwein LG, Kaledin L, Tepper F, Webster TJ.Osteoblast function on nanophase alumina materials: influence ofchemistry, phase, and topography. J Biomed Mater Res A2003;67:1284–93.

[3] Price RL, Ellison K, Haberstroh KM, Webster TJ. Nanometersurface roughness increases select osteoblast adhesion on carbonnanofiber compacts. J Biomed Mater Res A 2004;70:129–38.

[4] Washburn NR, Yamada KM, Simon CG, Kennedy SB, Amis EJ.High-throughput investigation of osteoblast response to polymercrystallinity: influence of nanometer-scale roughness on proliferation.Biomaterials 2004;25:1215–24.

[5] Wan YQ, Wang Y, Liu ZM, Qu X, Han BX, Bei JZ, Wang SG.Adhesion and proliferation of OCT-1 osteoblast-like cells on micro-and nano-scale topography structured poly(L-lactide). Biomaterials2005;26:4453–9.

B. Li et al. / Acta Biomaterialia 5 (2009) 134–143 143

[6] Elliott JC, Holcomb DW, Young RA. Infrared determination of thedegree of substitution of hydroxyl by carbonate ions in human dentalenamel. Calcif Tissue Int 1985;37:372–5.

[7] Lafon JP, Champion E, Bernache-Assollant D. Processing of AB-typecarbonated hydroxyapatite Ca10�x(PO4)(6�x)(CO3)(x)(OH)(2�x�2y)-(CO3)(y) ceramics with controlled composition. J Eur Ceram Soc2008;28:139–47.

[8] Murugan R, Ramakrishna S. Production of ultra-fine bioresorbablecarbonated hydroxyapatite. Acta Biomater 2006;2(2):201–6.

[9] Hench LL. Bioceramics. J Am Ceram Soc 1998;81:1705–28.[10] Sopyan I, Mel M, Ramesh S, Khalid KA. Porous hydroxyapatite for

artificial bone applications. Sci Technol Advan Mater 2007;8:116–23.[11] Yuan HP, Yang ZJ, Li YB, Zhang XD, De Bruijn JD, De Groot K.

Osteoinduction by calcium phosphate biomaterials. J Mater Sci-Mater Med 1998;9:723–6.

[12] Siddharthan A, Seshadri SK, Kumar TSS. Influence of microwavepower on nanosized hydroxyapatite particles. Scripta Mater2006;55:175–8.

[13] Wang XL, Fan HS, Xiao YM, Zhang XD. Fabrication andcharacterization of porous hydroxyapatite/beta-tricalcium phosphateceramics by microwave sintering. Mater Lett 2006;60:455–8.

[14] Guo XY, Gough JE, Xiao P, Liu J, Shen ZJ. Fabrication ofnanostructured hydroxyapatite and analysis of human osteoblasticcellular response. J Biomed Mater Res A 2007;82:1022–32.

[15] Zhao YT, Zhang YM, Ning F, Guo DG, Xu ZX. Synthesis andcellular biocompatibility of two kinds of HAP with differentnanocrystal morphology. J Biomed Mater Res B 2007;83:121–6.

[16] Shen ZJ, Zhao Z, Peng H, Nygren M. Formation of toughinterlocking microstructures in silicon nitride ceramics by dynamicripening. Nature 2002;417:266–9.

[17] Price RL, Haberstroh KM, Webster TJ. Enhanced functions ofosteoblasts on nanostructured surfaces of carbon and alumina. MedBiol Eng Comput 2003;41:372–5.

[18] Lowry OH, Roberts NR, Wu ML, Hixon WS, Crawford EJ. Thequantitative histochemistry of brain: 2. Enzyme measurements. J BiolChem 1954;207:19–37.

[19] Landi E, Celotti G, Logroscino G, Tampieri A. Carbonatedhydroxyapatite as bone substitute. J Eur Ceram Soc 2003;23:2931–7.

[20] Habibovic P, Yuan HP, van der Valk CM, Meijer G, van BlitterswijkCA, de Groot K. 3D microenvironment as essential element forosteoinduction by biomaterials. Biomaterials 2005;26:3565–75.

[21] Parfitt AM, Simon LS, Villanueva AR, Krane SM. Procollagen type-Icarboxy-terminal extension peptide in serum as a marker of collagenbiosynthesis in bone – correlation with iliac bone-formation rates andcomparison with total alkaline-phosphatase. J Bone Miner R1987;2:427–36.

[22] Tuckermann JP, Pittois K, Partridge NC, Merregaert J, Angel P.Collagenase-3 (MMP-13) and integral membrane protein 2a (Itm2a)are marker genes of chondrogenic/osteoblastic cells in bone forma-tion: sequential temporal, and spatial expression of Itm2a, alkalinephosphatase, MMP-13, and osteocalcin in the mouse. J Bone MinerR 2000;15:1257–65.

[23] Tamai M, Nakamura M, Isshiki T, Nishio K, Endoh H, Nakahira A.A metastable phase in thermal decomposition of Ca-deficienthydroxyapatite. J Mater Sci-Mater Med 2003;14:617–22.

[24] Brosnan KH, Messing GL, Agrawal DK. Microwave sintering ofalumina at 2.45 GHz. J Am Ceram Soc 2003;86:1307–12.

[25] Gough JE, Notingher I, Hench LL. Osteoblast attachment andmineralized nodule formation on rough and smooth 45S5bioactive glass monoliths. J Biomed Mater Res A 2004;68:640–50.

[26] Lohmann CH et al. Response of normal female human osteoblasts(NHOst) to 17 beta-estradiol is modulated by implant surfacemorphology. J Biomed Mater Res 2002;62:204–13.

[27] Webster TJ, Ergun C, Doremus RH, Siegel RW, Bizios R. Specificproteins mediate enhanced osteoblast adhesion on nanophase ceram-ics. J Biomed Mater Res 2000;51:475–83.

[28] Webster TJ, Schadler LS, Siegel RW, Bizios R. Mechanisms ofenhanced osteoblast adhesion on nanophase alumina involve vitro-nectin. Tissue Eng 2001;7:291–301.

[29] Webster TJ, Ejiofor JU. Increased osteoblast adhesion on nanophasemetals: Ti, Ti6Al4V, and CoCrMo. Biomaterials 2004;25:4731–9.

[30] Dickinson RB, Tranquillo RT. A stochastic model for adhesion-mediated cell random motility and haptotaxis. J Math Biol1993;31:563–600.

[31] Wang CY, Duan YR, Markovic B, Barbara J, Howlett CR, ZhangXD, Zreiqat H. Phenotypic expression of bone-related genes inosteoblasts grown on calcium phosphate ceramics with different phasecompositions. Biomaterials 2004;25:2507–14.

[32] Webster TJ, Ergun C, Doremus RH, Siegel RW, Bizios R. Enhancedfunctions of osteoblasts on nanophase ceramics. Biomaterials2000;21:1803–10.