lable at ScienceDirect

Biomaterials 30 (2009) 1512–1523

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomater ia ls

In vitro and in vivo evaluation of the surface bioactivity of a calcium phosphatecoated magnesium alloy

Liping Xu a,b, Feng Pan c, Guoning Yu d, Lei Yang a,b, Erlin Zhang a,e,*, Ke Yang a

a Institute of Metal Research, Chinese Academy of Sciences, 72 Wenhua Road, Shenyang 110016, Chinab Graduate University of Chinese Academy of Sciences, Beijing 100049, Chinac China Medical University, Shenyang 110001, Chinad The people’s Hospital of Liaoning Province, Shenyang 110016, Chinae Jiamusi University, Jiamusi 154007, China

a r t i c l e i n f o

Article history:Received 17 September 2008Accepted 1 December 2008Available online 27 December 2008

Keywords:MagnesiumCalcium phosphate coatingBioactivityBiocompatilityBonein vitro test

* Corresponding author. Institute of Metal ReseSciences, 72 Wenhua Road, Shenyang 110016, China.

E-mail address: erlin.zhang@imr.ac.cn (E. Zhang).

0142-9612/$ – see front matter � 2008 Elsevier Ltd.doi:10.1016/j.biomaterials.2008.12.001

a b s t r a c t

Magnesium has shown potential application as a bio-absorbable biomaterial, such as for bone screws andplates. In order to improve the surface bioactivity, a calcium phosphate was coated on a magnesium alloyby a phosphating process (Ca–P coating). The surface characterization showed that a porous and netlikeCaHPO4$2H2O layer with small amounts of Mg2þ and Zn2þ was formed on the surface of the Mg alloy.Cells L929 showed significantly good adherence and significantly high growth rate and proliferationcharacteristics on the Ca–P coated magnesium alloy (p< 0.05) in in-vitro cell experiments, demon-strating that the surface cytocompatibility of magnesium was significantly improved by the Ca–P coating.In vivo implantations of the Ca–P coated and the naked alloy rods were carried out to investigate thebone response at the early stage. Both routine pathological examination and immunohistochemicalanalysis demonstrated that the Ca–P coating provided magnesium with a significantly good surfacebioactivity (p< 0.05) and promoted early bone growth at the implant/bone interface. It was suggestedthat the Ca–P coating might be an effective method to improve the surface bioactivity of magnesiumalloy.

� 2008 Elsevier Ltd. All rights reserved.

1. Introduction

Recently, magnesium alloys have attracted much attention aspotential biodegradable bone implant materials due to theirbiodegradability in the bioenvironment [1–3], and their excellentmechanical properties such as high strength and possessing anelastic modulus close to that of bone [4]. However, the rapidcorrosion of magnesium alloys in chloride containing solutionsincluding human body fluid or blood plasma has limited theirclinical applications [4,5]. Therefore, it is very important to improvethe corrosion resistance of magnesium alloys in order for them tobe applied clinically.

Element alloying has been studied for developing biodegradablemagnesium alloys with good mechanical and corrosion properties.Mn and Zn were selected as the alloying elements to develop Mg–Mn–Zn alloys due to the good biocompatibility of Mn and Zn [6,7].The addition of Mn and Zn improves both the mechanical proper-ties and the corrosion resistance of magnesium alloys. It has been

arch, Chinese Academy ofTel./fax: þ86 24 23971605.

All rights reserved.

shown that the corrosion resistance of an Mg–1.0 Mn–1.0 Zn alloyin simulated body fluid (SBF) is slightly better than that of theWE43 alloy (containing 3.78 wt.% Y, 2.13 wt.% Nd and 0.46 wt.% Zr),which is one of the best commercially available magnesium alloys[8]. In vivo experiments indicated that new bone tissue formedaround the alloy implant after 6 weeks postimplantation, and about10–17% of the implant degraded after 9 weeks implantation, andmore than 55% degraded after 26 weeks implantation [3,9]. Nodisorders of liver, kidneys and blood composition were caused bythe degradation of the Mg–Mn–Zn alloy after 6 weeks, 15weeks, 26weeks and 52 weeks postimplantion, respectively [9]. Ca was alsochosen to produce a binary Mg–Ca alloy [10–12] and Ca–containingAZ91 magnesium alloy (AZ91Ca) [13] for the bone implant due tothe beneficial effect of Ca ion on bone growth. It was shown that theaddition of Ca increased the corrosion resistance of AZ91 alloy [13].Results on binary Mg–Ca alloys indicated that a Mg–Ca binary alloywith 0.6–1.0 wt.% Ca provides good overall mechanical propertiesand corrosion resistance [10,11]. Further increase in the Ca contentwould lead to a deterioration in the mechanical properties andcorrosion resistance [14].

In vivo studies have shown good bone attachment to magne-sium implants after 9 and 18 weeks postimplantion and goodbiocompatibility of magnesium in long-term service [1,3,14]. In vivo

Fig. 1. Implantation of magnesium rod samples with one naked sample and one Ca–Pcoated sample in one femora of a rabbit.

L. Xu et al. / Biomaterials 30 (2009) 1512–1523 1513

research on AZ31 [15] reported that osteotylus surrounds themagnesium alloy implant after 2 weeks postimplantation, whichwas turned into ripe bone tissue after 8 weeks. In contrast, ourresearch on the early bone response to magnesium bone implantsfound that only 50% of magnesium implants were fixed after 5weeks postimplantation [16]. Therefore, it is necessary to speed upthe early bone response to magnesium.

Although the element alloying can increase the bio-corrosionresistance of magnesium alloys to some extent, it is obvious thatalloying cannot significantly improve the bone response tomagnesium alloys, especially the bone response at the earlystage. Surface modification such as Al coating [17,18], Ti coating[19] and heat treatment [20] have been applied to magnesiumalloys to improve the corrosion resistance. However, surfacebiocompatibility was not enhanced. Bioactive coatings such asvarious calcium–phosphate compounds (Ca–P) are of importancefor modifying the surface of implanted devices, and have beensuccessfully applied to the surface modification of Ti and itsalloys in order to promote direct attachment of the surroundinghard tissue and to suppress the release of corrosion productsinto the human body [21]. Lately, calcium phosphate coatingshave been used to protect magnesium alloys from fast corrosion.For example, Cui reported that a Ca–P coating, includinghydroxyapatite, octacalcium phosphate and dicalcium phosphatedihydrate protected AZ31Mg alloy from rapid degradation in3.0% NaCl solution [22]. A previous study has reported thesuccessful preparation of a Ca–P layer on a Mg–Mn–Zn alloy,which effectively improved the corrosion resistance [23]. It wasalso reported that an electrodeposited HA coating on AZ91Dmagnesium alloy can obviously slow down the biodegradationrate of AZ91D magnesium alloy in simulated body fluid [24].However, no bioactivity of the Ca–P coating has been reportedso far.

In this paper, a calcium phosphate was coated on an Mg–Mn–Znalloy by a phosphating treatment in order to improve the surfacebioactivity of the magnesium alloy. Cell experiments were used toevaluate the cytocompatibility. Routine pathological examinationanalysis and growth factor expressions at the interface betweenmagnesium alloy implants and bone were utilized to assess in vivothe surface bioactivity during the early 1–4 weeks postoperation.

2. Experimental

2.1. Sample preparation

Extruded Mg–Mn–Zn (Mg–1.2 Mn–1.0 Zn, in wt.%) bars were prepared in ourlaboratory [7]. For the calcium phosphating treatment, plate samples witha dimension of 10 mm in diameter and 2 mm in thickness were cut from theextruded bar and moulded into epoxy resin with only one side exposed. Platesamples for the in vitro cell experiments with a dimension of 10 mm in diameter and2 mm in thickness, and rod samples for in vivo test with 2.8 mm in diameter and10 mm in length were cut from the extruded bar, respectively. All samples wereground with SiC emery papers of up to 1000 grits, and then ultrasonically cleaned inalcohol for 5 min and dried in warm air. In the cell experiment and the in vivo test,pure titanium samples of the same dimension were also used for comparison afterthe same surface treatment.

2.2. Calcium phosphate (Ca–P) coating treatment

The samples were first immersed in an alkaline solution at 63 �C for 15 min fordegreasing and subsequently immersed in a mixed acid solution (2%) of H3PO4 andH2SO4 at room temperature for 5–10 s for surface activation. Then, the samples weretreated in a phosphating bath for 6 min for calcium phosphate (Ca–P) treatment in

Table 1Chemical composition of the phosphating bath.

Composition H3PO4 Ca(H2PO4)2$H2O Zn(H2PO4)2$2H2O NaNO3 NaNO2

Concentration 7 mL/L 7.92 g/L 1.55 g/L 1.5 g/L 3.6 g/L

order to obtain a Ca–P coating on the surface. Table 1 listed the chemical compo-sition of the phosphating bath.

2.3. In vitro cell experiment

2.3.1. Cell cultureCell lines L929 were cultured in a PRMI1640 medium containing 2 mM

L-glutamine, 100 IU/mL penicillin, 100 mg/mL streptomycin sulphate and 10% (vol/vol) FBS. Then, the cells were maintained at 37 �C with 5% CO2 in air ina humidified incubator. Cells L929 were seeded onto the naked Mg–Mn–Znsamples (without a Ca–P coating) and the Ca–P coated Mg–Mn–Zn samples aswell as the pure titanium samples (control sample) at a cell density of 1�105 cell/mL. Cultures were incubated for 1 day, 3 days and 5 days at 37 �C with 5% CO2 inair in a humidified incubator in 48 well plates (Corning, NY, USA), respectively.Three parallel samples were used for each experimental condition. After differentincubation periods, the samples were washed three times with PBS to removenon-adherent cells.

2.3.2. Cell countFor the cell count, the samples were fixed in 4.0% formaldehyde solution for

24 h, and then used for immunofluorescence staining. Cell counts were conducted atfive different locations (four peripherals and one central) of the surface of eachsample under a TCS-SP2 laser scanning confocal microscope (LSCM). Each area wasabout 187.5 mm� 187.5 mm. A mean and a standard deviation were obtained fromthe five measurements.

Fig. 2. XRD diffraction patterns on the surface of the Ca–P coated Mg–Mn–Zn alloy.The insert figure shows the diffraction peaks of CaHPO4$2H2O in a high magnification.

Fig. 3. XPS spectra of the Ca–P coated Mg alloy (a) P2p, (b) Ca2p, (c) O1s, (d) Zn2p and (e) Mg2p. 1: top surface; 2: after ion etching for 300 s; 3: after ion etching for 600 s and 4: afteretching for 900 s.

L. Xu et al. / Biomaterials 30 (2009) 1512–15231514

2.3.3. MorphologyTo determine the cellular morphology, the samples were processed for

scanning electron microscopy (SEM) by a 24 h fixation in 2.5% (w/v)glutaraldehyde, gradual dehydration in 50–100 vol.% alcohol and coatedwith Au.

2.4. In vivo study

2.4.1. Surgical procedureAll animal experiments were conducted according to the ISO 10993-2:1992

animal welfare requirements. 18 healthy Japanese big-ear rabbits were used. All

Fig. 4. Surface structure and EDS result of the Ca–P coated sample. (a) SEM surfacemicrograph, (b) EDS spectrum of the framed area in (a).

L. Xu et al. / Biomaterials 30 (2009) 1512–1523 1515

rabbits were anaesthetized with 0.5% pentobarbital sodium solution for surgery. Onenaked Mg–Mn–Zn and one Ca–P coated Mg–Mn–Zn rod sample were implanted intothe left femoral shaft of a rabbit after predrilling with a 2.8 mm hand-operated drill,as shown in Fig. 1. After the operation, all rabbits received a subcutaneous injectionof gentamycin as an antibiotic prophylaxis. Intravital staining was performed weeklyusing subcutaneous injections of 1% water solution of calcein (0.3 mg/kg) to observethe newly formed bone. Postoperatively, the rabbits were allowed to move freely intheir cages without external support. Three rabbits were sacrificed at 1, 2 and 3weeks after surgery, respectively. Six rabbits (three for fluorescent observation,another three for routine pathological examination and immunohistochemistry)were sacrificed at 4 weeks after surgery.

2.4.2. Fluorescent observationFor microstructure analysis, the bone samples with magnesium implants were

fixed in 2.5% glutaraldehyde solution and then embedded in epoxy resin. Then, thesamples were ground by emery paper to get a ground Mg implant cross-section withbone. The cross-section was observed on an Olympus BX61 microscope.

2.4.3. Routine pathological examinationThe bone specimens including implants were taken out to observe osseointe-

gration and evaluate the surface bioactivity of the Ca–P coated samples. Followingfixation in 4.0% formaldehyde solution, the specimens were decalcified in 15% EDTAsolution for 3 weeks and embedded in paraffin. During this process, the residualmagnesium alloy implant was corroded by the solution completely, leaving a hole inthe implantation site of the bone sample. The paraffin-embedded bone specimenswere used for routine pathological examination and immunohistochemistry. Eachbone specimen was consecutively cut into 5-mm-thick slices. Deparaffinized sliceswere stained with hematoxylin and eosin (HE) stains, and microscopically examined.

2.4.4. ImmunohistochemistryImmunohistochemistry staining was performed using antibodies for BMP-2,

TGF-b1 and PDGF respectively. This was developed using a Strept Actividin-Biotin

Complex (SABC) method. Deparaffinized sections were pretreated with 0.3% H2O2

for 30 min at room temperature to block endogenous peroxidase activity. Afterwashing three times with distilled water, the sections were blocked with a 10%normal goat serum solution for 40 min at room temperature, and then the redun-dant solution was discarded. The primary antibodies (Boster, China) diluted in 0.1 M

PBS (1:200) were applied onto the sections, and the sections were maintained at4 �C for 48 h in a humidified chamber. After washing three times with 0.1 M PBS fora total of 6 min, the sections were reacted with the secondary antibody (biotin-goatanti-rabbit IgG, Boster, China) diluted in 0.1 M PBS (1:200) for 20 min at 37 �C. Afterrinsing three times with PBS for a total 6 min, a strept avdin-biotin complex (Boster,China) was applied onto the sections, and the sections were maintained 30 min at37 �C. After rinsing the sections four times with PBS for a total 20 min, visualizationof the antibody was accomplished by incubating sections in 0.03% 3,3’-dia-minobenzidine (DAB; Boster, China) for 5 min, and the sections were counterstainedin hematoxylin. Primary antibody replacement with PBS from the same animalsamples was used as the controls. Positive signals were stained brown. The routinepathological examination and immunohistochemisty were observed on an OlympusBX50 microscope. The immunohistochemistry was assessed quantitatively bycalculating the mean optical density (MOD) value of the implant/bone interface. Foreach specimen and factor, six fields (original magnification �400) at the implant/bone interface of each section were randomly picked for quantitative assessment,and the MOD values were calculated using image-analysis software (IPP5.1).

2.4.5. MicrostructureThe microstructure of the Ca–P coated samples and the cellular morphology was

observed on a SSX-550 scanning electronic microscope (SEM). The chemicalcomposition of the Ca–P layer was determined by energy dispersive spectrum (EDS).The surface structure of the Ca–P coated samples was characterized by X-rayphotoelectron spectroscopy (XPS, Escalab250, Thermo Corp.). The XPS measurementswere performed using an X-ray source of Mg Ka (1253.6 eV). Measured bindingenergies were corrected by referring to the binding energy of C1s of methylene groupsof the hydrocarbon (284.6 eV) absorbed on the surface of substrate. In order toidentify the phase constituents of the Ca–P coating on the magnesium sample, thesurface was examined with an X-ray diffractometer (XRD, D/MAX-RB, Rigaku) usingCu Ka radiation.

2.5. Statistical analysis

All statistics were performed using SPSS 16.0. Cell counting and immunohis-tochemisty analysis were all expressed as the mean� SD. Differences between twogroups were analyzed by the paired-samples t-test. Differences of more than twogroups were analyzed by an analysis of variance (ANOVA). Statistical significancewas defined as p< 0.05.

3. Results

3.1. Microstructure of Ca–P coating

Fig. 2 illustrates the XRD pattern on the Ca–P coated Mg–Mn–Znalloy. Besides the diffraction peaks from the magnesium matrix, thediffraction peaks of CaHPO4$2H2O were also detected by XRD,indicating that CaHPO4$2H2O was produced as the main newproduct on the surface of the Mg–Mn–Zn alloy after the calciumphosphate treatment.

Fig. 3 shows P2p, Ca2p, O1s, Zn2p and Mg2p XPS spectra for the Ca–P coated Mg alloy. It can be seen that the P2p3/2 spectrum wasdetected as a single peak and the Ca2p spectra were detected asdoublet peaks. From the binding energies of P2p3/2 and Ca2p, it canbe concluded that the element P exists in the layer in a form ofa phosphate. O1s was detected as a single peak. Zn2p spectra weredetected as doublet peaks. Small amounts of Mg were also detectedby XPS. According to the binding energies, it is deduced that theelements Zn and Mg are in the form of Zn2þ and Mg2þ. After 300 s,600 s and 900 s ion etching, respectively, no difference wasobserved in the P2p, Ca2p, O1s, Zn2p and Mg2p spectra. Combinedwith the XRD result shown in Fig. 2, it can be concluded that thesurface layer is mainly composed of CaHPO4$2H2O with a smallamount of Zn and/or Mg containing phosphate compoundsprecipitated from the phosphating solution. However, the amountof the Zn and/or Mg containing compounds is too small to bedetected and identified by XRD.

Fig. 4 shows the surface microstructure and the EDS result of theCa–P coated Mg–Mn–Zn alloy. A porous and netlike surface

Fig. 5. Cell morphology at low magnification after the incubation of 1, 3, and 5 days on different samples.

Fig. 6. Cell morphology at a high magnification after 1, 3, and 5 days incubation on different samples.

L. Xu et al. / Biomaterials 30 (2009) 1512–15231516

Fig. 7. Growth of L929 cells versus culturing time on the naked Mg sample, the Ca–Pcoated sample and the pure Ti sample. (*p< 0.05).

Fig. 8. Fluoroscopic images of the cross-section of (a) a naked magnesium implant and(b) a Ca–P coated magnesium implant after 4 weeks implantation. (The dark area is Mgimplant).

L. Xu et al. / Biomaterials 30 (2009) 1512–1523 1517

structure was observed clearly on the surface of the phosphatedmagnesium alloy, as shown in Fig. 4(a). Small cracks were alsofound, which were due to dehydration during the SEM samplepreparation. EDS analysis on a small square area of the phosphatedmagnesium, as shown in Fig. 4(b), indicates that the layer wasmainly composed of O, P, Ca, Mg and a small amount of Zn, sup-porting the XPS results in Fig. 3.

3.2. In vitro cell experiments

3.2.1. Cell morphologyFig. 5 presents the morphology of the cells cultured for 1 day, 3

days and 5 days on the surfaces of the naked Mg alloys, the Ca–Pcoated Mg alloys and the pure Ti. Differences in the response of thecells to the different surfaces were obvious. For the naked Mg alloy,only a few cells were observed on the surface after a 1-day culture.After 3 days and 5 days, no obvious change in the cell morphologyand the cell number was observed in the cultures. For the Ca–Pcoated Mg alloy, many cells were observed and the cells seemed tobe connected with each other after only a 1-day culture. After a 3-day culture, more cells were found on the surface. The cells havespread and connected together and covered most of the sample.After a 5-day incubation, a nearly dense and continuous cell layercovered the whole surface. For the pure Ti, some cells wereobserved on the surface after a 1-day culture and more cells wereseen after a 3-day incubation. After a 5-day culture, a cell layercovered the surface completely. By comparison, more cells wereobserved on the surface of the Ca–P coated Mg alloy than on thesurface of the naked Mg alloy during the whole incubation period,indicating a better cell response to the Ca–P coated Mg alloy. Fromthe point of view of cell morphology, the Ca–P coated Mg alloyshows better cell response than the pure Ti in the first 1 day culturedue to the fact that the cells on the surface of the Ca–P coated Mgalloy have been spreading and connecting with each other.However, in the further incubation periods, no obvious differencein the cell morphology was observed at a low magnificationbetween the Ca–P coated Mg sample and the pure Ti sample.

In the higher magnification morphology, the differences in thecells response to different surfaces were distinct, as shown in Fig. 6.The cells on the surface of the naked Mg alloys maintained a roundor spindle-like morphology during the whole incubation period.For the Ca–P coated Mg alloy, the cells were sail-like, elongated andthicker in the central area of the nucleus and nucleolus and flat-tened in the peripheral regions after a 1-day incubation. Some cellsspread across the surface and contacted with each other. After a 3-day culture, the cells and the excreted matrixes were connectedtogether and it is hard to distinguish between the cells and thematrix. After a 5-day culture, no difference in the morphology ofthe cells was found in comparison with that incubated for 3 days.For the pure Ti, the cells were spindle-like, and displayed a strongadhesion to the substrate after a 1-day culture. After a 3-dayculture, the cells displayed a flattened morphology with numerousfilamentous extensions. Further culture did not bring change to thecells morphology. From the morphology of the cell, both the Ca–Pcoated Mg sample and the pure Ti samples show significantlybetter cell response than the naked Mg sample during the wholeincubation period. In addition, the cells on the surface of the Ca–Pcoated Mg alloy excreted more matrix than the cells on the surfaceof the pure Ti did.

3.2.2. Cell proliferationFig. 7 illustrates the cell proliferation on the different samples.

The cell activity represented by the cell number was found toincrease with culture time, indicating that a cell could attach andproliferate on the surface of the samples. For the naked Mgalloys, there is no evident increase in the cell number at all time

Fig. 9. Pathological photographs of the implant/bone interfaces after 1, 2, 3 and 4 weeks postimplantation (HE stained). (a)–(d) the naked alloy, (e)–(h) the Ca–P coated alloy.I: implant; F: fibroblast band; N: newly formed osteoid tissue; C: connective tissue.

Table 2Summary of the routine pathological examination of the implant/bone interface.

Periods(weeks)

Naked Mg alloy Ca–P coated Mg alloy

1 Lymphocytic infiltration Connective tissueNo lymphocytic infiltrationNo plasmablastic infiltration

2 Continuous fibroblast band Thinner connective tissueNo lymphocytic infiltration OsteoblastsNo plasmablastic infiltration Newly formed bone matrix

Newly formed osteoid with osteocyte

3 Thinner fibroblast band No connective tissueSmall amount of newlyformed osteoid

OsteoblastsBone matrixConnecting trabecular

4 Newbone More new bonesTrabecular New bone trabecular aligned compactly

and regularlyOsteoblast

Fig. 11. MOD values of BMP-2 expression at the interface between the implant and thebone tissue after different periods of implantation. *p< 0.05.

L. Xu et al. / Biomaterials 30 (2009) 1512–1523 1519

intervals (p> 0.05), indicating that the naked Mg alloy does notpromote cell growth and proliferation. For the Ca–P coated Mgalloy and the pure Ti, there is a significant increase (p< 0.05) inthe cell number between day 1 and day 3, but no significantdifference (p> 0.05) between day 3 and day 5. In comparisonwith the naked Mg alloy, the cell number on the surfaces of theCa–P coated Mg alloy and the pure Ti showed a statisticallysignificant increase at all time intervals, indicating that both theCa–P coated Mg alloy and the pure titanium have a significantlybetter surface bioactivity than the naked Mg alloy. However, nostatistically significant difference in the cell number between theCa–P coated Mg alloys and the pure Ti was noted at all timeintervals (p> 0.05).

3.3. In vivo study

3.3.1. Fluorescent observationThe fluoroscopic images of the cross-section of bone and

magnesium implants after 4 weeks implantation are shown inFig. 8. It can be seen that newly formed osteoid tissue was observedaround both the naked Mg alloy implant and the Ca–P coated Mgalloy implant. Compared with the naked Mg alloy implant, thenewly formed osteoid tissue around the Ca–P coated Mg alloyimplant was compact and uniform. In addition, the outline shape ofthe magnesium implants was slightly changed, indicating theimplants were corroded by the body fluid, or the implant degraded

Fig. 10. Photomicrographs of BMP-2 expression at the interfaces between the

in the body. However, it is hard to distinguish the difference in thedegradation between the naked Mg alloy implant and the Ca–Pcoated Mg alloy implant after 4 weeks implantation because theduration is not long enough to evaluate the in vivo degradation.

3.3.2. Routine pathological examinationFig. 9 shows the optical microstructure of the interfaces

between the magnesium implants and new bone stained by HEafter 1–4 weeks postimplantation. For the naked Mg alloy,lymphocytic infiltration at the interface was observed after 1 weekpostimplantation, as indicated by ‘‘L’’ in Fig. 9(a). After 2 weeksimplantation, there was a continuous fibroblast band between theimplant and the bone, as indicated by ‘‘F’’ in Fig. 9(b). However,lymphocytic infiltration and plasmablastic infiltration were notnoted. Three weeks later, fibroblast band became thinner, as shownby ‘‘F’’ in Fig. 9(c), and small amount of newly formed osteoid tissuewas found, as indicated by ‘‘N’’ in Fig. 9(c). At week 4, the surface ofthe implant was taken up by newborn bone and bone trabecular, asindicated by ‘‘N’’ in Fig. 9(d). Crowded osteoblasts and bone matrixwere observed. For the Ca–P coated Mg alloy, connective tissue wasseen at the interface after 1 week postimplantation, as indicated by‘‘C’’ in Fig. 9(e), but lymphocytic infiltration and plasmablasticinfiltration was not seen. At week 2, the connective tissue becamethinner, and osteoblasts and bone matrix were noted. Newlyformed osteoid tissue with embedded osteocyte formed in some

implants and the bone tissue after 1, 2, 3 and 4 weeks postimplantation.

Fig. 12. Photomicrographs of TGF-b1 expression at the interfaces between the implants and bones after 1, 2, 3 and 4 weeks of implantation.

L. Xu et al. / Biomaterials 30 (2009) 1512–15231520

areas, as indicated by ‘‘N’’ in Fig. 9(f). Three weeks later, theconnective tissue was replaced by osteoblasts and bone matrix, asindicated by ‘‘N’’ in Fig. 9(g). Bone trabeculaes connected together,but the alignment of the connected bone trabeculae was disorga-nized. Four weeks later, more newborn bone was observed, asindicated by ‘‘N’’ in Fig. 9(h). Osteoid tissue and newborn bonetrabeculae almost covered the implant surface completely. Osteoidtissues connected together, and bone trabeculae aligned compactlyand regularly, in which mature osteocyte was embedded. Ossifi-cation was manifest. Table 2 summaries the routine pathologicalexamination results.

3.3.3. Immunohistochemistry3.3.3.1. BMP-2 expression at the implant/bone interface. Fig. 10shows BMP-2 staining photomicrographs at the interfaces betweenthe implants and bones after different periods of implantation.Fig. 11 shows the MOD values in the BMP-2 expression at theimplant/bone interface after different periods of implantation. Atweek 1, strong positive activities were observed at the interfaces forboth the naked alloy implant group and the Ca–P coated alloyimplant group. At week 2, positive activities increased and peaked.At weeks 3 and 4, osteoid tissues were clearly observed at theinterface, and the BMP-2 expressions gradually reduced. At all timeintervals, the Ca–P coated Mg implant group shows a higher MODvalue in the BMP-2 expression than the naked implant. A significant

Fig. 13. MOD values of TGF-b1 expression at the interface between the implant and thebone after different periods of implantation. *p< 0.05.

positive activity was found at the first 3 weeks as found for the Ca–Pcoated implant (p< 0.05).

3.3.3.2. TGF-b1 expression at the implant/bone interface. Fig. 12shows TGF-b1 staining photomicrographs at the interfacesbetween the implants and bones after different periods ofimplantation. Fig. 13 shows the MOD values from the TGF-b1expression at the implant/bone interface after different periods ofimplantation. The TGF-b1 expression has a tendency to givea similar result as that from the BMP-2 expression with theimplantation duration. The expression peaked at week 2, and thendecreased markedly. At week 4, the TGF-b1 expression becameweaker and few positive areas were detectable. The Ca–P coatedimplant exhibited a higher MOD value in the TGF-b1 expressionthan the naked implant at all durations and a significant differencein the TGF-b1 positive expression was observed between the nakedimplant and the coated implant group (p< 0.05) at weeks 2, 3 and4, respectively.

3.3.3.3. PDGF expression at the implant/bone interface. Fig. 14 showsPDGF staining photomicrographs at the interfaces between theimplants and bones after different periods of implantation. Fig. 15shows the MOD values of the PDGF expression at the implant/boneinterface after different periods of implantation. A similar tendencyof PDGF expression to the BMP-2 and TGF-b1 expressions wasobserved with implantation duration. Strong positive expressionswere observed at week 1 for both the naked and the coatedimplant. The expressions peaked at 2 weeks, and then the positivereaction diminished at week 3 and became weaker at week 4. Thecoated implant shows a higher MOD value than the naked implantat nearly all durations, and a significantly high MOD value at weeks1, 2 and 4, respectively (p< 0.05).

4. Discussion

Recently, research on magnesium alloys as biodegradablematerials has been conducted extensively. However, the hydrogenrelease and the alkalization caused by the in vivo corrosion ofmagnesium alloys are the most critical obstacles to using magne-sium alloys as biodegradable implant materials. One of the effectivemeasurements to reduce the corrosion of magnesium alloys issurface modification. For biomaterials, a surface coating is also aneffective way to improve the surface bioactivity. Therefore, it ispossible to reduce the corrosion of magnesium alloy and improvethe surface bioactivity by selecting a proper surface treatment.

Fig. 14. Photomicrographs of PDGF expression at the interfaces between the implants and bones after 1, 2, 3 and 4 weeks of implantation.

L. Xu et al. / Biomaterials 30 (2009) 1512–1523 1521

The microstructure of the Ca–P coated Mg alloy shows that thecoating layer was porous with a netlike surface structure, as shownin Fig. 4(a). EDS and SAXS results indicate that the layer was mainlycomposed of O,P,Ca,Zn and a little Mg, and the main phaseconstituent was CaHPO4$2H2O. Besides Ca2þ and PO4

4�, Mg2þ andZn2þ were also detected by XPS on the surface. However, the Mg2þ

and/or Zn2þ containing compounds were not detected by XRD dueto their small amounts. Magnesium alloy is a very active alloy.When a Mg alloy is immersed in a phosphating bath, magnesiumdissolves and turns into Mg2þ and releases H2. The NO3

� and NO2�

in the phosphating bath can react with Hþ and consume it quickly,so the local pH at the metal-solution interface increases and facil-itates the precipitation of insoluble phosphate [25]. At the sametime, Ca(H2PO4)2 has the potential to hydrolyse and the hydrolysisproduct brushite (CaHPO4$2H2O) will precipitate on the surface ofthe magnesium alloy. During this process, Zn2þ in the phosphatingbath and Mg2þ released from the magnesium alloy could react withany negative ions in the phosphating bath, such as PO4

3� to formsmall amounts of Zn and/or Mg containing phosphate compounds.

Lots of adhered, sail-like and conjoint cells after a 1-day culture,the connected cells and excreted matrix layer after 3 days and 5days cultures and the significant increase (p< 0.05) in the cellnumber during the whole incubation on the surface of the Ca–Pcoated Mg alloy in comparison with the naked Mg alloy indicate the

Fig. 15. MOD values of PDGF expression at the implant/bone interface after differentperiods of implantation. *p< 0.05.

Ca–P coating improves the surface bioactivity of magnesium alloysignificantly as proposed.

According to the fluorescent observation, more new osteoidtissues, which are compact and uniform, were observed around theCa–P coated Mg alloy implant than around the naked Mg alloyimplant after 4 weeks implantation, indicating that the Ca–P coatedMg alloy implant is more compatible for bone growth at the earlyhealing process.

Compared with the naked Mg alloy, routine pathologicalexamination analysis results such as no lymphocytic infiltration(inflammation) at week 1, thinner connective tissue and theformation of bone matrix at week 2, more bone matrix and inter-connected bone trabecular at week 3, and more newborn bones atweek 4, as shown in Fig. 9, reveal clearly that the Ca–P coated Mgalloy exhibits better surface biocompatibility than the naked alloyat the first 4 weeks postoperation.

Numerous growth factors have been implicated in the repair offracture healing, so expressions of bone growth factors can be usedto evaluate osteogenesis at the interface between implant andbone. Among those growth factors, the BMP plays crucial roles innormal skeletal development as well as bone healing, and is able toactivate transcription of genes involved in cellular migration,proliferation and differentiation [26]. It was reported that endog-enous BMP-2 is an indispensable osteogenic stimulus for initiationof fracture healing in mice [27]. In addition, it is believed that TGF-b stimulates osteogenesis, angiogenesis, fibroblast migration, anddeposition of matrix [28] and has osteoinductive properties [29–32]. PDGF stimulates osteoblasts proliferation, collagen synthesis,and may play a regulatory role in fracture repair [28,33]. Immu-nohistochemical analysis of our experiment results demonstratedthat the Ca–P coated implant provided a high BMP-2 expressionduring the first 4 weeks postimplantation, especially statisticallysignificant differences in BMP-2 expression between the Ca–Pcoated Mg alloy and the naked Mg alloy after 1, 2 and 3 weekspostoperation. The Ca–P coated implant also exhibited a high TGF-b1 expression during the first 4 weeks postimplantation, andstatistically significant differences in the TGF-b1 expressionbetween the Ca–P coated Mg alloy and the naked Mg alloy after 2, 3and 4 weeks postoperation. Similarly, a statistically significantdifference was also observed in PDGF expression between the Ca–Pcoated Mg alloy and the naked Mg alloy after 1, 2 and 4 weekspostoperation. All the above in vivo results demonstrated signifi-cantly good osteoconductivity of the Ca–P coated Mg alloy at theearly osseous integration stage.

For biomaterials application, the surface bioactivity is mainlycontrolled by the physical properties and the chemical properties ofthe surface. A porous surface at the microscale or nanoscale level

L. Xu et al. / Biomaterials 30 (2009) 1512–15231522

would contribute greatly to the faster adhesion and growth of cells,and a porous coating helps in bone cells growth and proliferationon the surface of the implant, resulting in a significantly strongerbond to the parent tissue [34]. In the surface biomodification ofpure titanium and titanium alloys, alkaline treatment has beenused to produce a porous surface structure at the micro- or nano-scale level to accelerate the deposition of hydroxyapatite [35,36].Moreover, the porous structure increases the particle boundaries,creates surface roughness, and may help toward numerous proteininteractions and thereby aids cell adhesion, alignment and finallytissue integration [37–42]. Further, these microscale features alsomimic the extra cellular matrix (ECM) present in the body tosupport cells [40,42]. In the present experiments, a porous surfacestructure was successfully prepared on a magnesium sample by thephosphating treatment, as shown in Fig. 4, which would definitelycontribute to the good surface bioactivity.

On the other hand, the surface chemical properties of thebiomaterials also play a very important role in good surfacebioactivity. Ti and Ti alloys are inert materials to cells. In order toimprove the surface biocompatibility, various calcium phosphatecoatings including brushite, octacalcium phosphate and hydroxy-apatite (HA) have been successfully applied to titanium-based andother alloys [43–45] because these compounds contain the samechemical composition or structure as the mineral composition ofnatural bone and the release of Ca2þ and PO4

3� during hydrolysiscan be utilized in the course of forming new bone [46]. Further-more, Ca2þ is also essential in chemical signaling with cells [47].Research has shown that the osteoblast amount and activity onthe calcium containing titanium surface are higher than those onthe titanium surface with solely phosphate ions and the surface ofpure Ti. And Ca2þ sites on the material surfaces favor proteinabsorption, such as fibronectin and vitronectin which are impor-tant cell attachment-promoting proteins and influence over cellattachment and spreading, onto the surface due to positive elec-tricity, chemical and biological function [48]. Mg2þ is known to beactive in cell adhesion mechanisms [34,49], and magnesium isnecessary for calcium incorporation [50]. Research has shown thatmodifying biomaterials with Mg2þ results in an increase in oste-oblasts adhesion [51], and Mg substituted tricalcium phosphatehas been shown to stimulate cell proliferation and the synthesisand secretion of collagenase [52]. Zn2þ also plays important rolesin controlling the function of osteoblasts and increasing osteoblastadhesion and the alkaline phosphatase activity of bone cells [53].The cytocompatibility of calcium phosphate/collagen compositesdoped with Zn was acceptable and the composites can beconsidered as a promising biomaterial that has the effect ofpromoting bone formation [54]. In addition, Zn and Mg ions in thecoating may promote osteogenesis due to their stimulatory effectson osteoblastic cell proliferation and bone formation, and theirinhibitory effect on osteoclastic bone resorption and decreasingand regulating the inflammatory reaction [53,55,56]. Willi Paulet al. [57] reported that osteoblast-like MG63 cells are attachedand spread completely on a calcium zinc magnesium phosphate(CaZnMgP) matrix and this matrix appears to be comparable tothe control group (hydroxyapatite matrix). Therefore, divalentcations such as Ca2þ, Zn2þ and Mg2þ in the Ca–P layer alsopromote cell growth and differentiation.

In addition, it has to be pointed out that the alkalization effectcaused by the fast corrosion of magnesium, which resulted in a fastincrease in pH value of the SBF to over 9 in about 2 h and finally torelatively stable value of 10.5, is undesirable at all for cell adhesion,growth and proliferation [58]. Previous studies have shown that theCa–P coating can protect effectively the magnesium alloy from fastcorrosion and reduce the increasing of the pH value of the SBFsolution [23]. Thus, the Ca–P coating would also benefit celladhesion and growth to some extent.

5. Conclusion

A porous and netlike Ca–P coating was successfully prepared onmagnesium alloy by a phosphating treatment to improve thesurface bioactivity of the magnesium substrate. Microstructurecharacterization showed the Ca–P coating was mainly composed ofCaHPO4$2H2O with small amounts of Zn and Mg ion. In vitro celltests demonstrated that the Ca–P coating provided the magnesiumalloy with a significantly better surface cytocompatibility and invivo results also confirmed that the Ca–P coating exhibited signif-icantly improved osteoconductivity and osteogenesis in the earlyfirst 4 weeks postoperation period.

Acknowledgements

One of the authors (Erlin Zhang) would like to acknowledge thefinancial support from the Institute of Metal Research (IMR),Chinese Academy of Sciences (CAS), Shenyang Science and Tech-nology Institute (Program No. 1062109-1-00), and HeilongjiangProvince Natural Funding (No. E2007-18).

References

[1] Witte F, Kaese V, Haferkamp H, Switzer E, Meyer-Lindenberg A, Wirth CJ, et al.In vivo corrosion of four magnesium alloys and the associated bone response.Biomaterials 2005;26(17):3557–63.

[2] Witte F, Fischer J, Nellesen J, Crostack H-A, Kaese V, Pisch A, et al. In vitro andin vivo corrosion measurements of magnesium alloys. Biomaterials 2006;27(7):1013–8.

[3] Xu L, Yu G, Zhang E, Pan F, Yang K. In vivo corrosion behavior of Mg–Mn–Znalloy for bone implant application. J Biomed Mater Res 2007;83A(3):703–11.

[4] Staiger MP, Pietak AM, Huadmai J, Dias G. Magnesium and its alloys asorthopedic biomaterials: a review. Biomaterials 2006;27(9):1728–34.

[5] Song G. Control of biodegradation of biocompatable magnesium alloys. CorrosSci 2007;49(4):1696–701.

[6] Yin D, Zhang E, Zeng S. Effect of Zn on mechanical property and corrosionproperty of extruded Mg–Zn–Mn alloy. T Nonferrous Metal Soc 2008;18(4):763–8.

[7] Zhang E, Yin D, Xu L, Yang L, Yang K. Microstructure, mechanical and corrosionproperties and biocompatibility of Mg–Zn–Mn alloys for biomedical applica-tion. Mater Sci Eng C – Bio S, in press, doi:10.1016/j.msec.2008.08.024.

[8] Xu L, Zhang E, Yin D, Zeng S, Yang K. In vitro corrosion behaviour of Mg alloysin a phosphate buffered solution for bone implant application. J Mater Sci –Mater M 2008;19(3):1017–25.

[9] Zhang E, Xu L, Yu G, Pan F, Yang K. In vivo evaluation of biodegradablemagnesium alloy bone implant in the first 6 months implantation. J BiomedMater Res, in press, doi:10.1002/jbm.a.32132.

[10] Wan Y, Xiong G, Luo H, He F, Huang Y, Zhou X. Preparation and character-ization of a new biomedical magnesium–calcium alloy. Mater Design2008;29(10):2034–7.

[11] Li Z, Gu X, Lou S, Zheng Y. The development of binary Mg–Ca alloys for use asbiodegradable materials within bone. Biomaterials 2008;29(10):1329–44.

[12] Kim W-C, Kim J-G, Lee J-Y, Seok H-K. Influence of Ca on the corrosion prop-erties of magnesium for biomaterials. Mater Lett 2008;62(25):4146–8.

[13] Kannan MB, Raman RKS. In vitro degradation and mechanical integrity ofcalcium-containing magnesium alloys in modified-simulated body fluid.Biomaterials 2008;29:2306–14.

[14] Vojtech D, Ciova H, Volenec K. Investigation of magnesium-based alloys forbiomedical applications. Kovove Mater 2006;44:211–23.

[15] Zhang G, Huang J, Yang K, Zhang B, Ai H. Experimental study of in vivoimplantation of a magnesium alloy at early stage. Acta Metallu Sin2007;43(11):1186–90.

[16] Zhang E, Xu L, Yang K. Investigation on new biomedical degradable magne-sium alloys. In: 2006 Beijing International Materials Week, June 25–30, Bei-jing, China; 2006. p. 317.

[17] Chiu LH, Lin HA, Chen CC, Yang CF, Chang CH, Wu JC. Effect of aluminumcoatings on corrosion properties of AZ31 magnesium alloy mater. Sci Forum2003;419–422:909–14.

[18] Chiu L-H, Chen C-C, Yang C-F. Improvement of corrosion properties in analuminum-sprayed AZ31 magnesium alloy by a post-hot pressing and anod-izing treatment. Surf Coat Tech 2005;191(2–3):181–7.

[19] Zhang E, Xu L, Yang K. Formation by ion plating of Ti-coating on pure Mg forbiomedical applications. Scripta Mater 2005;53(5):523–7.

[20] Liu C, Xin Y, Tang G, Chu PK. Influence of heat treatment on degradationbehavior of bio-degradable die-cast AZ63 magnesium alloy in simulated bodyfluid. Mater Sci Eng A – Struct 2007;456(1–2):350–7.

[21] Zheng CY, Li SJ, Tao XJ, Hao YL, Yang R, Zhang L. Calcium phosphate coating ofTi–Nb–Zr–Sn titanium alloy. Mat Sci Eng C – Bio S 2007;27(4):824–31.

L. Xu et al. / Biomaterials 30 (2009) 1512–1523 1523

[22] Cui F, Yang J, Jiao Y, Yin Q, Zhang Y, Lee IS. Calcium phosphate coating onmagnesium alloy for modification of degradation behavior. Fron Mater SciChina 2008;2(2):143–8.

[23] Xu L, Zhang E, Yang K. Phosphating treatment and corrosion properties of Mg–Mn–Zn alloy for biomedical application. J Mater Sci – Mater M, in press, doi:10.1007/s10856-008-3648-2.

[24] Song YW, Shan DY, Han EH. Electrodeposition of hydroxyapatite coating onAZ91D magnesium alloy for biomaterial application. Mater Lett 2008;62(17–18):3276–9.

[25] Kouisni L, Azzi M, Zertoubi M, Dalard F, Maximovitch S. Phosphate coatings onmagnesium alloy AM60 part 1: study of the formation and the growth of zincphosphate films. Surf Coat Tech 2004;185(1):58–67.

[26] Derynck R, Zhang Y. Intracellular signalling: the mad way to do it. Curr Biol1996;6(10):1226–9.

[27] Tsuji K, Bandyopadhyay A, Harfe BD, Cox K, Kakar S, Gerstenfeld L, et al. BMP2activity, although dispensable for bone formation, is required for the initiationof fracture healing. Nat Genet 2006;38(12):1424–9.

[28] Helm GA, Dayoub H, Jane JA. Gene-based therapies for the induction of spinalfusion. Neurosurgical Focus 2001;10(4). Article 5.

[29] Noda M. In vivo stimulation of bone formation by transforming growth factor-beta. Endocrinology 1989;124(6):2991–4.

[30] Beck LS, Amento EP, Xu Y, Deguzman L, Lee WP, Nguyen T, et al. TGF-beta 1induces bone closure of skull defects: temporal dynamics of bone formation indefects exposed to rhTGF-beta 1. J Bone Miner Res 1993;8(6):753–61.

[31] Nielsen HM, Andreassen TT, Ledet T, Oxlund H. Local injection of TGF-b increases the strength of tibial fractures in the rat. Acta Orthop 1994;65(1):37–41.

[32] Zhang H, Ahmad M, Gronowicz G. Effects of transforming growth factor-beta 1(TGF-b1) on in vitro mineralization of human osteoblasts on implant mate-rials. Biomaterials 2003;24(12):2013–20.

[33] Helm GA, Sheehan JM, Sheehan JP, Jane Jr JA, diPierro CG, Simmons NE, et al.Utilization of type I collagen gel, demineralized bone matrix, and bonemorphogenetic protein-2 to enhance autologous bone lumbar spinal fusion.J Neurosurg 1997;86(1):93–100.

[34] Paul W, Sharma CP. Nanoceramic matrices: biomedical applications. Am JBiochem Biotechnol 2006;2(2):41–8.

[35] Fujibayashi S, Neo M, Kim HM, Kokubo T, Nakamura T. Osteoinduction ofporous bioactive titanium metal. Biomaterials 2004;25(3):443–50.

[36] Kim HM, Miyaji F, Kokubo T, Nakamura T. Preparation of bioactive Ti and itsalloys via simple chemical surface treatment. J Biomed Mater Res 1996;32(3):409–17.

[37] Kasemo B. Biological surface science. Surf Sci 2002;500(1-3):656–77.[38] Assender H, Bliznyuk V, Porfyrakis K. How surface topography relates to

materials’ properties. Science 2002;297(5583):973.[39] Webster TJ, Ejiofor JU. Increased osteoblast adhesion on nanophase metals: Ti,

Ti6Al4V, and CoCrMo. Biomaterials 2004;25(19):4731–9.[40] Stevens MM, George JH. Exploring and engineering the cell surface interface.

Science 2005;310:1135–8.

[41] Lutolf MP, Hubbell JA. Synthetic biomaterials as instructive extracellularmicroenvironments for morphogenesis in tissue engineering. Nat Biotechnol2005;23(1):47–55.

[42] Paital SR, Dahotre NB. Laser surface treatment for porous and textured Ca–Pbio-ceramic coating on Ti-6Al-4V. Biomed Mater 2007;2:274–81.

[43] Shirkhanzadeh M. Direct formation of nanophase hydroxyapatite on cathod-ically polarized electrodes. J Mater Sci – Mater M 1998;9(2):67–72.

[44] Vijayaraghavan TV, Bensalem A. Electrodeposition of apatite coating on puretitanium and titanium alloys. J Mater Sci Lett 1994;13(24):1782–5.

[45] Redepenning J, McIsaac JP. Electrocrystallization of brushite coatings onprosthetic alloys. Chem Mater 1990;2(6):625–7.

[46] Higashi S, Yamamuro T, Nakamura T, Ikada Y, Hyon SH, Jamshidi K. Polymer-hydroxyapatite composites for biodegradable bone fillers. Biomaterials1986;7(3):183–7.

[47] Ilich JZ, Kerstetter JE. Nutrition in bone health revisited: a story beyondcalcium. J Am Coll Nutr 2000;19(6):715–37.

[48] Feng B, Weng J, Yang BC, Qu SX, Zhang XD. Characterization of titaniumsurfaces with calcium and phosphate and osteoblast adhesion. Biomaterials2004;25(17):3421–8.

[49] Hynes RO. Integrins: versatility, modulation, and signaling in cell adhesion.Cell 1992;69(1):11–25.

[50] Serre CM, Papillard M, Chavassieux P, Voegel JC, Boivin G. Influence ofmagnesium substitution on a collagen-apatite biomaterial on the productionof a calcifying matrix by human osteoblasts. J Biomed Mater Res 1998;42(4):626–33.

[51] Zreiqat H, Howlett CR, Zannettino A, Evans P, Schulze-Tanzil G, Knabe C, et al.Mechanisms of magnesium-stimulated adhesion of osteoblastic cells tocommonly used orthopaedic implants. J Biomed Mater Res 2002;62(2):175–84.

[52] LeGeros RZ, Gatti AM, Kijkowska R, Mijares DQ, LeGeros JP. Mg-substitutedt ricalcium phosphates: formation and properties. Key Eng Mater 2004;254–256:127–30.

[53] Yamaguchi M, Oishi H, Suketa Y. Stimulatory effect of zinc on bone formationin tissue culture. Biochem Pharmacol 1987;36(22):4007–12.

[54] Santos MH, Heneine LGD, Mansur HS. Synthesis and characterization ofcalcium phosphate/collagen biocomposites doped with Zn2þ. Mater Sci Eng C –Bio S 2008;28(4):563–71.

[55] Grandjean-Laquerriere A, Laquerriere P, Jallot E, Nedelec J-M, Guenounou M,Laurent-Maquin D, et al. Influence of the zinc concentration of sol-gel derivedzinc substituted hydroxyapatite on cytokine production by human monocytesin vitro. Biomaterials 2006;27(17):3195–200.

[56] Moonga BS, Dempster DW. Zinc is a potent inhibitor of osteoclastic boneresorption in vitro. J Bone Miner Res 1995;10(3):453–7.

[57] Paul W, Sharma C. Effect of calcium, zinc and magnesium on the attachmentand spreading of osteoblast like cells onto ceramic matrices. J Mater Sci –Mater M 2007;18(5):699–703.

[58] Song G, Song S. A possible biodegradable magnesium implant material. AdvEng Mater 2007;9(4):298–302.