Materials Science and Engineering C 41 (2014) 309–319

Contents lists available at ScienceDirect

Materials Science and Engineering C

j ourna l homepage: www.e lsev ie r .com/ locate /msec

Short-term in vivo evaluation of zinc-containing calcium phosphateusing a normalized procedure

Monica Calasans-Maia a,⁎, José Calasans-Maia a, Silvia Santos b, Elena Mavropoulos b, Marcos Farina c,Inayá Lima d, Ricardo Tadeu Lopes d, Alexandre Rossi b, José Mauro Granjeiro a,e

a Dental Clinical Research Center, Dentistry School, Fluminense Federal University, Niteroi, Rio de Janeiro, Brazilb LABIOMAT, Brazilian Center for Physics Research, CBPF, Rio de Janeiro, Brazilc Institute of Biomedical Sciences, Federal University of Rio de Janeiro, Rio de Janeiro, Brazild Nuclear Instrumentation Laboratory, Nuclear Engineering Program, COPPE, Federal University of Rio de Janeiro, Rio de Janeiro, Brazile Bioengineering Division, National Institute of Metrology, Quality and Technology, Duque de Caxias, Rio de Janeiro, Brazil

⁎ Corresponding author at: Av. Rui Barbosa, 582/501, F22250-020, Brazil. Tel.: +55 21 81535884; fax: +55 21 2

E-mail addresses: monicacalasansmaia@gmail.com (Mjosecalasans@gmail.com (J. Calasans-Maia), silviaquimicaelena@cbpf.br (E. Mavropoulos), mfarina@anato.ufrj.br (Minayacorrea@gmail.com (I. Lima), rossi@cbpf.br (A. Rossi)(J.M. Granjeiro).

http://dx.doi.org/10.1016/j.msec.2014.04.0540928-4931/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 6 September 2013Received in revised form 4 April 2014Accepted 23 April 2014Available online 4 May 2014

Keywords:HydroxyapatiteZincSynthesisBiocompatibilityPhysico-chemical characterization

The effect of zinc-substituted calcium phosphate (CaP) on bone osteogenesis was evaluated using an in vivo nor-malized ISO10993-6protocol. Zinc-containing hydroxyapatite (ZnHA) powderwith 0.3%bywt zinc (experimen-tal group) and stoichiometric hydroxyapatite (control group) were shaped into cylindrical implants (2 × 6 mm)andwere sintered at 1000 °C. Thermal treatment transformed the ZnHA cylinder into a biphasic implant thatwascomposed of Zn-substituted HA and Zn-substitutedβ-tricalcium phosphate (ZnHA/βZnTCP); the hydroxyapatitecylinder was a highly crystalline and poorly soluble HA implant. In vivo tests were performed in New ZealandWhite rabbits by implanting two cylinders of ZnHA/βZnTCP in the left tibia and two cylinders of HA in theright tibia for 7, 14 and 28 days. Incorporation of 0.3% by wt zinc into CaP increased the rate of Zn release tothe biological medium. Microfluorescence analyses (μXRF-SR) using synchrotron radiation suggested thatsome of the Zn released from the biomaterial was incorporated into new bone near the implanted region. Incontrast with previous studies, histomorphometric analysis did not show significant differences between thenewly formed bone around ZnHA/βZnTCP and HA due to the dissolution profile of Zn-doped CaP. Despite thegreat potential of Zn-containing CaP matrices for future use in bone regeneration, additional in vivo studiesmust be conducted to explain themobility of zinc at theCaP surface and its interactionswith a biologicalmedium.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

There is a critical demand for bone substitute materials in medicineand dentistry. Bone defects that result from congenital, infectious, trau-matic or neoplastic processes represent the most important challengesin reconstructive treatment. Autologous bone grafting is considered tobe the gold standard in reconstruction, but the supply of autologousbone is limited, and the harvesting of the graft is associated withmorbidity [1].

Hydroxyapatite (HA) is widely used as a biomaterial to fill bonedefects and to coat the metal parts of prostheses. Synthetic HA is wellknown as an implant material and has excellent biocompatibilitycharacteristics, including non-toxicity, low biodegradability and bone

lamengo, Rio de Janeiro, RJ CEP:5566074.. Calasans-Maia),@gmail.com (S. Santos),. Farina),, jmgranjeiro@gmail.com

affinity [2,3]. HA is considered to be osteoconductive due to its abilityto strongly bond with natural bone tissue [4,5].

Despite having these optimal properties, synthetic HA differs frombiological apatite. Several research studies have attempted to mitigatethese differences by doping synthetic HA with small amounts of impu-rities. These ionic substitutions can alter the properties of HA, includingits crystallinity, morphology, lattice parameters, stability, solubility andmechanical character [6–8].

The inorganic component of the bone tissue is a nonstoichiometriccarbonated apatite containing substitutions of Na+, K+, Mg2+, Sr2+,Cl−, F−, HPO4 and Zn2+ [8]. Studies have been carried out to investigatethe effects of changes in the composition of HA to better understand andimprove the tissue response after HA implantation. The presence of traceelements affects bone formation and resorption through direct or indirecteffects on bone cells or bone mineral [9]. The gradual release of divalentcations (Mg2+ and Zn2+) incorporated into HA may favor bone repairby improving the cytocompatibility conditions for osteoblast adhesionor may serve as an in vivo model of heavy metal toxicity [7]. These mod-ified forms of HA, known as proactive bioceramics, stimulate the desiredresponses from the surrounding cells and tissues necessary to promotebonding of orthopedic and dental implants to the bone [6].

310 M. Calasans-Maia et al. / Materials Science and Engineering C 41 (2014) 309–319

Zinc is an essential trace element formany animals, including humans,and it is known to play a role in growth and development [10]. In vitrostudies revealed that granules consisting of 5% mol Zn2+ incorporatedinto hydroxyapatite are cytocompatible [11]. Recent in vitro studiesshowed that zinc-containing hydroxyapatite decreased inflammationand increased chemotaxis [2] and demonstrated in vitro antimicrobial ef-fects towards Streptococcus mutans and Staphylococcus aureus [12,13]. Inaddition, doping HA with 2% mol Zn2+ significantly increased osteoblastadhesion compared to undopedHA [6], while doping HAwith concentra-tions of Zn2+ between 0.6 and 1.2 wt.% enhanced the proliferation ofmouse osteoblast-like cells in composite ceramic [14]. A recent studyshowed that ZnHA possessed enhanced bioactivity because an increasein the growth of human adipose-derived mesenchymal stem cells, alongwith the bone cell differentiation markers, was observed [13].

In vivo evaluations of Ca substitution for Zn in CaP have been per-formed using implant materials with different compositions (βZnTCP,αZnHA, ZnHA/βZnTCP, βZnTCP/HA and ZnHA), variable Zn2+ concen-trations (from0.01 to 2.7 wt.%), different preparation routes and animalmodels (rats, rabbits). Although these works demonstrated that Zn2+

had stimulatory effects on bone formation and an inhibitory effecton osteoclastic bone resorption [4,5,15–20], many doubts remainconcerning the optimum amount of zinc in implants and the role ofzinc in osteogenesis involving Zn-containing CaP biomaterials.

Based on the background discussed above, the purpose of the pres-ent work was to compare the in vivo efficacy of a biphasic CaP (ZnHA/βZnTCP) with 0.3 wt.% Zn and sintered HA using the normalized ISO10993-6 protocols. According to this standard procedure, CaP cylindersthat were 2 mm in diameter and 6 mm in height were implanted intorabbit tibias. The efficacy of the cylinders in promoting bone formationwas evaluated using histological and histomorphometric techniques.In this work, the ZnHA/βZnTCP cylinderswere sintered at 1000 °C to re-duce the influence of porosity and to induce a partial decomposition ofZnHA as well as to form a biphasic system containing a more solubleβZnTCP phase in addition to ZnHA.

2. Materials and methods

2.1. Synthesis of HA and ZnHA powders

Colloidal hydroxyapatite (HA) was precipitated by the drop-wiseaddition of an (NH4)2HPO4 solution (99% pure, Merck®, Darmstadt,Germany) to an aqueous solution of Ca(NO3)2 (99% pure, Merck) at90 °C. The original solutions were adjusted to pH 9.0 by addingNH4OH. The precipitate was separated by filtration, repeatedly washedwith boiling deionized water at pH 7.0 and dried at 100 °C for 24 h. Thedried powder wasmanually ground, and particle aggregates measuringless than 210 μm were separated with a sieve.

The synthesis of HA doped with Zn (nominal concentration of1% mol) followed the same procedure as described above, except thatsolutions of both Zn(NO3)2 and Ca(NO3)2 were used.

2.2. Cylinder (implant) preparation for the ISO 10993-6 protocol

The ZnHA and HA powders were compacted into cylinders under200 N in an isostatic press for 2 min. For each cylinder, 0.70 g of ZnHAor HA powder was used, resulting in 2 mm × 6 mm cylinders thatwere then sintered at 1000 °C for 2 h.

2.3. Characterization of the powders and cylinders

The crystalline mineral phases present in the samples, their crystal-linity and the proportion of the hydroxyapatite and β-tricalcium phos-phate phases were examined by X-ray diffraction (XRD). The XRDpatterns were obtained with a HZG4 diffractometer operating at 30 kVand 15mA,with CuKα radiation (λ= 1.542 Å). The datawere collectedin the 2θ range of 10°–100° with a step of 0.02° point per second. The

contents of the hydroxyapatite and β-tricalcium phosphate phases inthe cylinders were evaluated by the relative intensities of specificpeaks of βTCP and HA XRD patterns in the sample as described byBalmain et al. [21]. Mixtures of powders containing known proportionsof HA and βTCP sintered at 1000 °C were used to construct a standardcurve, X = (IβTCP / (IβTCP + IHA) versus the βTCP content (wt.%),where IβTCP and IHA are the integrated intensities of the (0210) βTCPand (300) HA peaks, respectively. The relative intensity of the (003)and (0210) peaks in the XRD patterns of the standard HA and βTCPstructures (60% and 100%) were also considered for the determinationof X. The β-tricalcium phosphate content of the cylinder was estimatedby comparing the X value of the ZnHA sample after calcination at1000 °C with the X values from the standard curve.

X-rayfluorescence (XRF) and atomic absorption spectroscopy (AAS)were performed to determine the HA stoichiometry, particularly theZn/Ca molar ratio and calcium substitutions. The XRF spectrometer(XRF — Phillips PW 2400) operated at 40 kV and 50 mA with a Ge(111) crystal, a collimator of 550 μm and a flow detector for the Kαlines of phosphorus, calcium and zinc. The incorporation of Zn2+ intoHA was estimated with an AA-6800 Spectrophotometer (Shimadzu)operating with an air–acetylene flame atomizer. The zinc absorbancewasmeasured at 213.9 nm. Vibrationalmodes of phosphate andhydroxylgroups in ZnHA samples were analyzed by Fourier transform infraredspectroscopy. The spectra were obtained in a FTIR-IR—Prestige 21(Schimadzu) operating in transmission mode from 400 to 4000 cm−1.The dissolution of the ZnHA cylinders was evaluated by determining cal-cium and zinc released by inductively coupled plasma-optical emissionspectrometry (ICP-OES) in 0.088 M MES (4-morpholineethanesulfonicacid hydrate) with pH 5.9 and 0.088 M Hepes (hydroxyethylpiperazine-N-2-ethane sulfonic acid) with pH 7.4 after 1 and 7 days at 37 °C. Themorphology of the implants was examined by scanning electron micros-copy (SEM— JEOL JSM 5310).

2.4. Animals

Fifteen skeletally mature male and female New Zealand Whiterabbits weighing between 2500 and 3000 g were used. The animalexperiments and breeding were performed under conditions thatwere approved by the Institutional Review Board (CEP/UFF) no. 195,in compliance with the NIH Guide for Care and Use of Laboratory Ani-mals and with Brazilian legislation on animal experimentation.

2.5. Surgical procedures

All animals were pre-anesthetized with ketamine (20 mg/kg) andxylazine (1 mg/kg) and anesthetized with spinal anesthesia (lidocaine,0.3 mg/kg andmorphine, 0.1 mg/kg). Following anesthesia and trichot-omy, a 4-cm incision wasmade in the epithelial lining of the animal leg.After exposure of the tibia bone surface and under constant saline irriga-tion, two holes separated by a 1-cm margin were drilled; each holemeasured 2 mm in diameter and penetrated both tibia cortices in a di-rection perpendicular to the bone axis. Two ZnHA cylinders and twoHA cylinders (control) were implanted in the left and right sets oftibia holes, respectively, of each rabbit. The tissue flap was then placedin its original position, and the incision was closed with interrupted#5-0 nylon sutures (Fig. 1). After surgery, the rabbits were allowed tomove freely in their cages. The rabbits were injected intramuscularlywith one dose of Pentabiótico® (1 mL/kg; Penicillin, antibiotic) andMaxicam® (1 mL/kg; Meloxicam, an anti-inflammatory). A group offive animals was euthanized at the end of each experimental period(7, 14 and 28 days after surgery).

2.6. Histological and histomorphometric analyses

Two fragments, each containing one of the two cylinders from eachtibia, were collected. One fragment was fixed in 10% buffered formalin

Fig. 1. Surgical procedures for bone implantation. A. Surgical bone defect, B. Cylinders implanted and C. Micro-CT shows the direction of implantation.

Table 1XRF analysis.

Sample Ca (mol) P (mol) Zn (mol%) Ma/P

HA 1.001 0.604 1.66ZnHA 0.382 0.250 0.464 1.54

a M = Metal, where M = Ca in the HA and M = Ca + Zn in the ZnHA.

311M. Calasans-Maia et al. / Materials Science and Engineering C 41 (2014) 309–319

for 48 h and demineralized for 24 h (Allkimia®, a bone demineralizingsolution) prior to histological processing for paraffin embedding; paraf-fin sectionsmeasuring 5 μmthickwere stainedwith hematoxylin–eosin(HE). The other fragment was fixed in 70% ethanol prior to dehydrationin successive ethanol solutions and then impregnated and embedded inmethylmethacrylate to prepare undecalcified sections. We obtainedthree histological sections from each undecalcified block that were cutat thicknesses of 30–50 μm and 100 μm to acquire a transverse sectionof the tibia along the central axis of the cylinder. The unstained, polished30–50-μmsectionswere analyzed by bright field and polarized lightmi-croscopy (Zeiss Axioplan). Elemental mapping and analyses were per-formed, respectively, by X-ray microfluorescence using synchrotronradiation on the D09-XRF beam line (μXRF-SR; LNLS, Campinas, Brazil)and by energy dispersive spectrometry with a scanning electron micro-scope (SEM — JEOL JSM 5310).

The decalcified sections were observed under a light microscope(Nikon Eclipse E400) with a 10× magnification objective lens. Eightnon-consecutive images of each specimen section were capturedusing a CCD camera (Evolution MP color 5.0 Media Cybernetics). Theimages were used for histomorphometrical analysis with Image Pro-Plus 6.0® (Media Cybernetics, Inc.) to calculate the volume density ofthe newly formed bone (NFB) and to evaluate connective tissue forma-tion (CTF) by measuring the Area of Interest (AOI).

2.7. Statistical analysis

A statistical analysis of all the histomorphometric data was per-formed using GraphPad Instat®, a commercially available softwarepackage. The paired Student's t test was used to compare the meandifferences between the two groups. For all measures, a p value lessthan 0.05 was considered to be statistically significant.

2.8. Synchrotron radiation X-ray microfluorescence

μXRF-SR elemental mapping was carried out on the D09-XRF beamline at Brazilian Synchrotron Light Laboratory, Campinas, São Paulo,Brazil. A white beam (4 keV min–23 keVmax) was used for sample ex-citation. The measurements were performed on a computer-controlledXYZ table, which allowed us to choose the region of interest (ROI) ineach bone sample. The samples were affixed to the resin of the speci-men holder using a single piece of tape. The specimen holder did notcontribute to the μXRF-SR sample spectra because the measurementswere performed away from the holder, at the center of the bonesamples.

The choice of ROIwasmade based on the interface between the boneand the implant (HA or ZnHA/ßZnTCP) using a microscope and a CCDcamera. To obtain the μXRF-SR data, several spectra were recorded atdifferent points on each sample (t = 50 s/point). The μXRF-SR elemen-tal mapping was performed in an area equal to 1.2 mm2 with an imageresolution of 30 μmtoproduce 1681 spectra. Themeasurement timeperpoint was 10 s; it took approximately 4 h to complete one single scan.

All XRF spectra were analyzed using the free QXAS/AXIL softwarefrom IAEA. The resulting X-ray fluorescence mapping scans of eachbone samplewasproduced at an image resolution of 30 μm.Themethodto calculate the elemental concentration, as well as the calibrationprocedure and absorption term used, was the same as that describedpreviously [22,23]. Finally, each element was plotted separately bydistribution mapping, which was performed by the authors with thehelp of a commercial MATLAB® program.

To standardize our experimental method, measurements weretaken using an NIST Standard Reference Material (SRM 1400-BoneAsh). Five spectra were acquired at five different points on the SRMpellet, each one taking 100 s to complete. Approximately 180 mg ofthis standard was pressed (12 tons) into a tablet that was 1 inch(±0.02 in.) in diameter.

3. Results

The process of synthesis used in this study yielded a stoichiometricHA powder with a Ca/P ratio of 1.66 and a ZnHA powder that was0.29% by wt zinc (% molar), with a (Ca + Zn)/P ratio of 1.54, as deter-mined by XRF (Table 1). Atomic absorption spectroscopy (AAS) analy-sis, which is a more accurate test, confirmed that the ZnHA contained0.3% by wt zinc.

TheXRDpatterns for theHAand ZnHApowders before sintering andHA and ZnHA/ßZnTCP after sintering are shown in Fig. 2. The tempera-ture of calcination (1000 °C) was found to increase the crystallinity ofboth materials. The new peaks (arrow in Fig. 2) indicated that thermaltreatments induced the partial decomposition of ZnHA and the forma-tion of a β-tricalcium phosphate phase (βZnTCP). The β-tricalciumphosphate content in the cylinder was estimated by the methoddescribed inMaterials andmethods section. The use of the standard cal-ibration curve X versus the β-tricalcium phosphate content (result notshown) revealed that the X value of ZnHA sample after thermal treat-ment at 1000 °C corresponded to a β-tricalcium phosphate content of16.0 wt.%.

FTIR spectroscopy analysis of the ZnHA and HA powders aftersintering showed bands ranging from 1100 to 950 cm−1, correspondingto the vibrational modes of (PO4)3−, and a band at 3570 cm−1, corre-sponding to the vibration mode for (OH−1) (Fig. 3). The low intensityof the OH−1 band indicated the partial decomposition of ZnHA intoβZnTCP.

Macroscopic analysis of both cylinders showed that after sintering,the HA cylinder had a smaller diameter (2.12 mm) than that of the

Fig. 2. X ray diffraction pattern (XRD) of 0.3 wt.% ZnHA samples in their as-synthesized condition before sinterization (black) and after sinterization (red). Observe the peak representingCaZnTCP (black arrow). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

312 M. Calasans-Maia et al. / Materials Science and Engineering C 41 (2014) 309–319

ZnHA cylinder (2.32 mm), suggesting a higher densification of HA.This effect was due to the partial decomposition of ZnHA into βZnTCP,which inhibited the densification of the ZnHA/βZnTCP cylinder. At themicroscopic level, the surface of both cylinders showed micrometricroughness. However, the HA cylinder showed a surface topographywith more pronounced roughness, cracks and microporosity that werenot observed on the ZnHA surface (Fig. 4).

Incubation of the ZnHA/βZnTCP sintered cylinder in 0.088 M MES(pH 5.9) and 0.088 M Tris–HCl (pH 7.4) for 1 and 7 days showed that

Fig. 3. Fourier transformed infrared (FTIR) spectra of the 0.3 wt.% ZnHA (red) and HA (black) sathis figure legend, the reader is referred to the web version of this article.)

the release of calcium increased with time, from 10 to 19 mg/L and 82to 150 mg/L, respectively, and the release of calcium was stronglyaffected by the pH of the medium (Fig. 5). The amount of zinc was notdetected because the released Zn was most likely below the limit ofdetection for our experimental conditions. Fig. 5 also shows that thesintered ZnHA/βZnTCP cylinder released more calcium than the ZnHApowder before thermal treatment. This result confirms the high dissolu-tion rate of βZnTCP and its importance in the release of zinc from thezinc-containing implant.

mples after thermal treatment at 1000 °C. (For interpretation of the references to color in

Fig. 4. Scanning electronmicroscopy (SEM)micrographsof a cross section of HA (A) and ZnHA (C) cylinders at 35×magnification (scale bar=500 μm)and the surface ofHA (B) and ZnHA(D) cylinders at 5000× magnification (scale bar = 5 μm).

313M. Calasans-Maia et al. / Materials Science and Engineering C 41 (2014) 309–319

3.1. Evaluation of ZnHA/ßZnTCP after implantation

The μXRF-SR spectra for the HA and ZnHA implants showed that theonly difference between the samples was the intensity of the Ca and Zncomponents; Ca was more abundant in the HA sample, while Zn waspresent only in the ZnHA sample. The absence of Zn and the presenceof high Ca levels were found in the μXRF-SRmappings of HA. High levelsof Ca were observed in the area corresponding to the pre-existing bone.These levels appeared to be reduced in the areas of newly formed bone,with the exception of the HA group at 28 days after implantation, inwhich the newly formed bone showed high levels of Ca. In some areasof pre-existing bone and newly formed bone, low levels of Zn could beobserved in both of the studied samples. μXRF-SR elemental mappingof Ca and Zn showed the presence of Zn and Ca in the ZnHA cylindersat all the studied time points (Fig. 6). The μXRF-SR 2D mappings of Ca

Fig. 5. Calcium release from ZnHA/ßZnTCP cylinders on dissolution experiment in buffersolutions, MES pH 5.9 and HEPES pH 7.4, for 1 and 7 days. P: Powder and Cy: Cylinder.

and Zn showed that the level of bone mineralization was positivelycorrelated with the length of the time segment being studied (7, 14and 28 days post-surgery), suggesting that the newly formed bone isprogressively mineralized.

Table 2 shows the results of the analysis of the NIST Standard Refer-ence Material (SRM 1400-Bone Ash), and differences between thevalues can be observed based on the absorption method used to calcu-late the elemental concentrations. The observed differences in thevalues in Table 2 can be assigned to the semi-empirical method usedto calculate the absorption factor. Importantly, no gold standard existsfor the calculation of the absorption factor in XRF, and various method-ologies can be used.

3.2. Histological analysis of the sections

Our histological findings varied according to the length of the im-plantation period due to the healing and remodeling processes in therabbit bone. New bone formation was evident at 2 and 4 weeks post-implantation. Earlier, at the 1-week time point, both HA and ZnHAgroups showed a combination of connective tissue with mononuclearcell infiltrate (lymphocytes and plasmocytes) and newly formed boneprojecting from the pre-existing bone to the cylinder, a characteristicof osteoconduction. A difference between theprimary bone (trabecular)and the native bone (lamellar) was evident. The newly formed bonebonded directly to the pre-existing surface from the implant boneafter 4 weeks (Fig. 7).

Polarized light microscopy showed that the collagen fibrils of thenew bone projecting from the pre-existing bone to the implants had adifferent organization. The histological findings were similar in the 2groups at 2 weeks after implantation, as well as after 4 weeks, whenthe gap between the cylinder and the pre-existing bonewas completelyfilled by new bone in both groups (Fig. 8). This technique wasperformed to further characterize the different regions around thecylinders.

Fig. 6. Examples of the μXRF-SR spatial distributions of Ca and Zn. ZnHA/ßZnTCP implants are shown after A. 7 days, B. 14 days and C. 28 days.

314 M. Calasans-Maia et al. / Materials Science and Engineering C 41 (2014) 309–319

Fig. 6. (continued).

315M. Calasans-Maia et al. / Materials Science and Engineering C 41 (2014) 309–319

3.3. Histomorphometric examination

Histomorphometric analysis confirmed the histological findings.Compared with the ZnHA group, the control group showed morenewly formed bone around the cylinders at the first week after implan-tation (p = 0.039), but no difference was found between these groupsat 14 and 28 days (p N 0.05). In both groups, the differences in theareas of newly formed bone observed at 7 and 14 days were not statis-tically significant (p N 0.05), but the differences between 7 and 28 dayswere highly significant (p b 0.01). Between the 14-day and 28-day timepoints, the difference in the area of new bone was statistically signifi-cant for the HA group (p b 0.05) but not for the ZnHA group(p N 0.05), as shown in Fig. 9.

4. Discussion

The possibility of incorporating zinc into calcium phosphate ce-ramics for biomedical applications has been discussed in the literaturefor over a decade [2,4–7,10–20]. The substitution of Zn2+ for Ca2+ re-sults in changes in the physico-chemical properties of the CaP structure,such as solubility, thermal stability, particle morphology and size[24–27] and biological behavior.

One challenge of the in vitro and in vivo studies conducted with CaPcompounds with incorporated zinc is the evaluation of the effect of zincon bone regeneration. The in vivo data suggest that the action of zinc de-pends on its release from the biomaterial surface and its interactionwith the mineralizing cells, ions and molecules of the physiologicalmedium [16–18]. The assessment of the ideal zinc concentration for a

CaP implant depends on the chemical phase in which the zinc is incor-porated as well as the biomaterial dissolution rate, its porosity andsurface topography. Furthermore, the animal model used in thein vivo tests may also interfere with the final result.

In this study, we used a standard in vivo protocol (ISO 10993-6) toallow for comparative analysis, even when the tests were performedin different centers [28,29]. The rabbit was an adequate and viablemodel for experimental studies because of its low cost, ease of handling,availability in sufficient numbers to permit statistical analysis and simi-larity to humans in terms of bone composition, bone mineral densityand fracture resistance [30,31]. The cylindrical implants used in theISO 10993-6 protocol were adequate because of their large axial surfacethat contacts the bone. In addition, the flat surface of the implant led toeasier histological and histomorphometric analyses of the bone/implantinterface.

The CaP cylinder used as a control for the ISO 10993-6 test was pro-duced from a single phase, it was highly crystalline and had nearly stoi-chiometric HA. The cylinderwas subjected to thermal treatment at 1000°C to decrease themicro-roughness of the implant surface and to reducethe Ca and P dissolution rate, as shown in Figs. 4 and 5, respectively. TheZn-containing CaP implant was produced using the same procedure asthe control (HA). The amount of Zn2+ associated with CaP was 0.3%by wt, which allowed for an easy comparison with previous in vivostudies conducted with Zn2+ concentrations in the range of 0.06 to2.7% by wt [15–20]. As shown in Fig. 2, thermal treatment at 1000 °Cdecomposed the Zn-containing cylinder into a biphasic system contain-ing ZnHA (major phase) and βZnTCP. The phase composition of theZnHA/βZnTCP implant was similar to that used in the in vivo evalua-tions of Zn-containing CaP conducted by Kawamura et al. [16] and Luo

Table 2Values from the μXRF-SR analysis of NIST standard reference material (SRM 1400-BoneAsh).

Elements Standard values Measured values

Ca (%) 38.18 ± 0.13 20.52 ± 1.8Zn (μg·g−1) 181 ± 3 176 ± 1.06

316 M. Calasans-Maia et al. / Materials Science and Engineering C 41 (2014) 309–319

et al. [32]. The thermal treatment and ZnHA decomposition into βZnTCPalso induced changes in the mechanical properties of the implant.Cracks or fissures on the surfaces of the HA cylinders were observed inan analysis of the cylinders by scanning electron microscopy before im-plantation, suggesting that the cylinders aremost likelymore brittle andless elastic than the ZnHA/βZnTCP cylinders, which had a more homo-geneous and smooth surface (see Fig. 4). From a clinical point of view,the HA cylinder was harder and less soluble in the blood medium thanthe ZnHA/βZnTCP cylinder. These observations are supported by thefact that photomicrographs from 5-μm-thick sections were obtainablefor the ZnHA/βZnTCP sample. In contrast, these sections were impossi-ble to obtain for theHA sample, whichwas extremely brittle during sec-tioning on the ultramicrotome. In histological sections obtained fromthe ZnHA/βZnTCP group, we observed HE residues in the ZnHA/βZnTCPmaterial itself. This result was most likely due to the presence of

Fig. 7.Micrographs after 7, 14 and 28 days of implantation. A, C and E. The HA group and B, D anMagnification: 10×.

proteins in the biological fluids that permeated the ZnHA/βZnTCP cylin-der and/or to the nature of the calcium phosphate material (see Fig. 7B).

Dissolution experiments performed inMES (pH 5.9) and HEPES (pH7.4) buffers showed that the amount of Ca2+ released from the ZnHA/βZnTCP cylinder increased with time (Fig. 5) and was significantlyhigher than the release of Ca2+ from the ZnHA powder treated usingthe same conditions. These data confirmed the decisive contributionof βZnTCP to the dissolution rate (Ca2+, P2+ and Zn2+) of the ZnHA/βZnTCP cylinder. The strong increase in calcium release as the pHchanged from 7.4 to 5.9 (from 19 mg/L to 150 mg/L) may signal asharp increase in Zn2+ release when the local pH decreased during aninflammatory reaction in the first days after implantation.

Table 3 summarizes the results of the in vivo evaluations performedwith Zn-containing CaP implants [16,18–20,31]. The 7 day release ofCa2+ from ZnHA/βZnTCP in buffers at pH 7.4 (19 mg/L) and pH 5.9(150 mg/L) was significantly higher than the Ca2+ dissolution ofZnTCP/HA with 0.3% by wt Zn for 60 days in 0.9% by wt NaCl (14ppm) and sodium acetate/acetic acid solutions (42 mg/L) [16]. Theseresults suggest that the release of zinc by the ZnHA/βZnTCP implant issuperior to ZnTCP/HA because Ca and Zn are elements of the CaPstructure.

The histological analysis of the ZnHA/βZnTCP implant sections from7 days after implantation revealed typical events that occur in the first

d F. ZnHA group. PEB: Pre-existing bone; NFB: Newly formed bone; CT: connective tissue.

317M. Calasans-Maia et al. / Materials Science and Engineering C 41 (2014) 309–319

stage of bone remodeling: a moderate inflammatory reaction; connec-tive tissue around the implant; and formation of new bone linked tothe pre-existing bone without a direct contact to the biomaterial. TheμXRF-SR 2D mapping of Ca (Fig. 6) revealed new evidence concerningthe calcium distribution in the region between the implant and thepre-existing bone in this early stage of mineralization. Although thehistological analyses demonstrated that new bone was not present inthat area, μXRF-SR 2D mapping confirmed the existence of a calciumdistribution occupying the space between the old bone and the implant.In that region, the calcium signal was less intense than in the region ofthe implant and the pre-existing bone, suggesting the formation of a

Fig. 8. Image obtained using polarization light microscopy after 28 days for the ZnHAgroup. Note that the region corresponding to the newly formed bone (diffuse birefrin-gence) does not present organized collagen fibrils as in the pre-existing bone (parallelbright lines).

calcium complex with a lower density than the bone that is likely dueto an amorphous calcium phosphate. μXRF-SR 2D mapping of Zn2+

showed that themetal was concentrated near the implant region. How-ever, a continuous decrease in the amount of zinc was observed fromthe cylinder interface in the direction of the interior of the implantedcavity. This effect may be attributed to Zn dissolution from the cylindersurface.

At 14 days after implantation, histological analyses showed that thenewbone interconnected the implantwith the pre-existing bone. μXRF-SR 2Dmapping of calcium confirmed this conclusion because the inten-sity of the calcium signal in the space between the bone and the implantwas comparable to the signal of the pre-existing bone and implant. Thestrong μXRF-SR Zn signal around the cylinder (Fig. 6) revealed that partof the released zinc that remained in the neighborhood of the implantwas most likely associated with the newly formed bone. Kawamuraet al. [16] had a similar conclusion from the indirect finding that theZn2+ content in the rabbit plasma did not change during the period inwhich Zn2+ was released from the implant.

The histomorphometric analysis conducted on the rabbit's tibia ac-cording to ISO 10993-6 showed that the ZnHA/βZnTCP implants wereefficient in stimulating bone regeneration because the new bone areaincreased from 10.5% to 84.9% between the first and the fourth weekafter implantation, respectively. However, the Zn-containing implantwas not more efficient in promoting new bone formation than thesintered HA implants as shown in Fig. 9. This finding agreed with thework of Resende et al. [20] but contradicted previous in vivo works byKawamura et al. [16,19], Xia Li et al. [18] and Xiaoman Luo et al. [32],as shown in Table 3. Although the Zn2+ concentrations used in thesestudies ranged from0.03 to 2.7% bywt, the optimal zinc content to stim-ulate bone formationwas estimated as 0.3% bywt [16] and 0.013% bywt[18] in bone tissue and 2.7% by wt [31] in intramuscular tissue. Accord-ing to Kawamura H. et al. [16], for Zn2+ contents higher than the opti-mal value (0.3% by wt), the bone formation area per medullary cavityarea decreased and achieved the same value as the control (implantwithout zinc) for a Zn2+ concentration of 0.6% by wt.

The in vivo data in the present work using 0.3% by wt Zn2+ in theZnHA/βZnTCP implants did not confirm the conclusion of the previousstudies. Therefore, the explanation for this controversy should not be fo-cused on the zinc content of the Zn-doped CaP implants but on its disso-lution profile in the biological medium. In this regard, Table 3 showsthat βZnTCP/ZnHA with 0.3% by wt Zn released 10 mg/L and 19 mg/Lafter 1 and 7 days at pH 7.1, respectively, while βZnTCP/HA with 0.6%by wt Zn and 0.136% by wt Zn released 14 mg/L and 42 mg/L, respec-tively, after 60 days in NaCl medium. Considering the high dissolution

Fig. 9. Relative area of newly formed bone for theHAandZnHAgroups at 7, 14 and28dayspost-implantation. The areas of newly formed bone increased for both tested materials.

Table 3Zinc contents, In vitro release of Zn, P, Ca and in vivo evaluation of zinc-containing calcium phosphate.

Material Zn content (Wt%) In vitro release of Ca, P, Zn and experimentalconditions

Effect on Zn on bone regeneration References

Ion release (ppm) Medium, pH, time

ZnTCP and ZnTCP/HA (cylinders) 0.063 Ca (15), P(12), Zn(0) NaCl 60 d Stimulate bone formation (rabbits, 4 weeks). Theoptimum zinc content was 0.316 wt.%.

Kawamura H. et al. [16]0.316 Ca(14), P(10), Zn(0.03)0.633 Ca(42), P(23), Zn(2)

ZnTCP/HA (cylindrical rods) 0.316 Ca(14), P(10), Zn(0.03) NaCl 60 d Favorable effects in short term (rabbits, 6 weeks).Unfavorable effects (bone resorption) in the longterm (60 weeks)

Kawamura H. et al. [19]

ZnTCP (granules) 0.3 Ca(45), P(45), Zn(0.3) Cult. Med 4 d Stimulate intramuscular bone formation. Highzinc content in TCP led to osteoinductivity

Xiaoman Luo et al. [32]0.9 Ca(40), P(40), Zn(0.3)2.7 Ca(40), P(40), Zn(1.8)

a-ZnTCP (cement) 0.03 Stimulate bone formation (rabbits, 4 weeks). Theoptimum zinc content was 0.013 wt.% to avoidinflammatory reactions.

Xia Li et al. [18]0.060.100.18

ZnHA/ZnTCP (spheres) 0.3 Zn did not stimulate bone formation. Themorphometric analysis indicates that there was agreater absorption in the Zn group (rabbits, 26, 52and 78 weeks)

Resende et al. [20]

ZnHA/ZnTCP (cylinders) 0.3 Ca(82), P(−), Zn(−) MES pH = 5.9 1 d Zn did not stimulate bone formation (rabbits, 1, 2and 4 weeks). No statistical difference wasobserved between ZnHA/ZnTCP and HA.

This paperCa(150), P(−), Zn(−) MES pH = 5.9 7 dCa(10), P(−), Zn(−) HEPES pH = 7.1 1 dCa(19), P(−), Zn(−) HEPES pH = 7.1 7 d

318 M. Calasans-Maia et al. / Materials Science and Engineering C 41 (2014) 309–319

rate of the βZnTCP/ZnHA implant and the μXRF-SR result that showedthat part of the Zn2+ remained in the region near the implant, we canspeculate that the dissolution behavior of the ZnHA/βZnTCP implantwith 0.3% by wt Zn (this work) was more similar to βZnTCP/HA with0.6% by wt Zn than 0.136% by wt Zn [16].

Based on the above discussion, we conclude that ZnHA/βZnTCPwith0.3% by wt Zn most likely had a dissolution behavior and zinc releaserate above the level required to stimulate bone remodeling. This factmay explain the apparent contradiction between this and previousworks. Hence, the future of Zn-containing CaP biomaterials for applica-tions in bone regeneration will depend on the knowledge of thephysico-chemical and morphological properties of the CaP implant atthe micro- and nanoscales and more directly on the mechanisms ofzinc release from the biomaterial.

5. Conclusion

In this study, the effect of zinc released from the zinc-doped CaPimplant on bone remodeling was evaluated using a normalized ISO10993-6 protocol. We demonstrated that sintered biphasic ZnHA/βZnTCP with 0.3% by wt Zn and sintered HA showed equivalent levelsof bone formation in rabbit tibias. The ZnHA/βZnTCP implant showeda high dissolution rate for a sintered biomaterial but did not induce astrong inflammatory response. μXRF-SR analyses from thin sections ofimplanted material using synchrotron radiation suggested that part ofthe zinc released from the biomaterial remained close to the implantedregion and was likely linked to new bone. The lack of an observedstimulatory effect of zinc on bone remodeling was attributed to thedissolution behavior of the implant and the excess zinc released to thebiological medium.

Acknowledgments

The authors acknowledge the support from LNLS and COPPE and fi-nancial support from Faperj (E-26/102.729/2008), CNPq (303770/2007-4) and Finep (01.11.0090.00).

References

[1] K. Eid, S. Zelicof, B.P. Perona, C.B. Sledge, J. Glowacki, Tissue reactions to particles ofbone-substitute materials in intraosseous and heterotopic sites in rats: discrimination

of osteoinduction, osteocompatibility, and inflammation, J. Orthop. Res. 19 (2001)962–969.

[2] A. Grandjean-Laquerriere, P. Laquerriere, E. Jallot, J.M. Nedelec, M. Guenounou, D.Laurent-Maquin, T.M. Phillips, Influence of the zinc concentration of sol–gel derivedzinc substituted hydroxyapatite on cytokine production by human monocytesin vitro, Biomaterials 27 (2006) 3195–3200.

[3] K.A. Hing, L.F. Wilson, T. Buckland, Comparative performance of three ceramic bonegraft substitutes, Spine J. 7 (2007) 475–490.

[4] M.D. Calasans-Maia, G.V.O. Fernandes, A.M. Rossi, E.P. Dias, G.D.S. Almeida, F.F. Mitri,J.M. Granjeiro, Effect of hydroxyapatite and zinc-containing hydroxyapatite on osse-ous repair of critical size defect in the rat calvaria, Key Eng. Mater. 361–363 (2008)1273–1276.

[5] M.D. Calasans-Maia, A.M. Rossi, E.P. Dias, S.R.A. Santos, F.O. Ascoli, J.M. Granjeiro,Stimulatory Effect on Osseous Repair of Zinc-substituted Hydroxyapatite: Histolog-ical Study in Rabbit's Tibia, Key Eng. Mater. 361–363 (2008) 1269–1272.

[6] T.J. Webster, C. Ergun, R.H. Doremus, R. Bizios, Hydroxyapatite with substitutedmagnesium, zinc, cadmium, and yttrium. II: Mechanisms of osteoblast adhesion, J.Biomed. Mater. Res. 59 (2002) 312–317.

[7] T.J. Webster, E.A. Massa-Schlueter, J.L. Smith, E.B. Slamovich, Osteoblast response tohydroxyapatite doped with divalent and trivalent cations, Biomaterials 25 (2004)2111–2121.

[8] A. Bigi, E. Foresti, M. Gandolfi, M. Gazzano, N. Roveri, Inhibiting effect of zinc onhydroxyapatite crystallization, J. Inorg. Biochem. 58 (1995) 49–58.

[9] E. Landi, G. Logroscino, L. Proietti, A. Tampieri, M. Sandri, S. Sprio, Biomimetic Mg-substituted hydroxyapatite: from synthesis to in vivo behaviour, J. Mater. Sci.Mater. Med. 19 (2008) 239–247.

[10] J. Brandão-Neto, V. Stefan, B.B. Mendonça,W. Bloise, A.V.B. Castro, The essencial roleof zinc in growth, Nutr. Res. 15 (1995) 335–358.

[11] I.R. de Lima, G.G. Alves, C.A. Soriano, A.P. Campaneli, T.H. Gasparoto, E.S. Jr Ramos, L.A. de Sena, A.M. Rossi, J.M. Granjeiro, Understanding the impact of divalent cationsubstitution on hydroxyapatite: An in vitromultiparametric study on biocompatibil-ity, J. Biomed. Mater. Res. 98 (2011) 351–358.

[12] X.C. Shi, H.Z. Jiang, J. Xue, Y.M. Liao, L.Y. Xiao, C.Y. Bao, W. Li, Antibacterial effects ofsol–gel derived zinc-containing hydroxyapatite, Key Eng. Mater. 361–363 (2008)139–142.

[13] E.S. Thian, T. Konishi, Y. Kawanobe, P.N. Lim, C. Choong, B. Ho, Zinc-substitutedhydroxyapatite: a biomaterial with enhanced bioactivity and antibacterial proper-ties, J. Mater. Sci. Mater. Med. 24 (2013) 437–445.

[14] A. Ito, K. Ojima, H. Naito, N. Ichinose, T. Tateishi, Preparation, solubility, andcytocompatibility of zinc-releasing calcium phosphate ceramics, J. Biomed. Mater.Res. 50 (2000) 178–183.

[15] L. Nascimento, M. Medeiros, J. Calasans-Maia, A. Alves, A.M. Rossi, G. Alves, J.M.Granjeiro, M.D. Calasans-Maia, Osseoinduction evaluation of hydroxyapatite andzinc containing hydroxyapatite granules in rabbits, Key Eng. Mater. 493–494(2012) 252–257.

[16] H. Kawamura, A. Ito, S. Miyakawa, P. Layrolle, K. Ojima, N. Ichinose, T. Tateishi,Stimulatory effect of zinc-releasing calcium phosphate implant on bone formationin rabbit femora, J. Biomed. Mater. Res. 50 (2000) 184–190.

[17] A. Ito, H. Kawamura, M. Otsuka, M. Ikeuchi, H. Ohgushi, K. Ishikawa, K. Onumo, N.Kanzaki, Y. Sogo, N. Ichinose, Zinc-releasing calcium phosphate for stimulatingbone formation, Mater. Sci. Eng. C 22 (2002) 21–25.

[18] X. Li, Y. Sogo, A. Ito, H. Mutsuzaki, N. Ochial, T. Kobayashi, S. Nakamura, K.Yamashita, R.Z. Legeros, The optimum zinc content in set calcium phosphate ce-ment for promoting bone formation in vivo, Mater. Sci. Eng. C 29 (2009) 969–975.

319M. Calasans-Maia et al. / Materials Science and Engineering C 41 (2014) 309–319

[19] H. Kawamura, A. Ito, T. Muramatsu, S. Miyakawa, N. Ochial, T. Tateishi, Long-termimplantation of zinc-releasing calcium phosphate ceramics in rabbit femora, J.Biomed. Mater. Res. A 65 (2003) 468–474.

[20] R.F.B. Resende, G.V.O. Fernandes, S.R.A. Santos, A.M. Rossi, I. Lima, J.M. Granjeiro,M.D. Calasans-Maia, Long-term biocompatibility evaluation of 0.5 % zinc containinghydroxyapatite in rabbits, J. Mater. Sci. Mater. Med. 24 (2013) 1455–1463.

[21] N. Balmain, R. Legros, Bonel, Calcif. Tissue Int. 34 (1982) S93–S98.[22] T. Shiraiwa, N. Fujino, Theoretical Calculation of Fluorescent X-Ray Intensities in

Fluorescent X-Ray Spectrochemical Analysis, Jpn. J. Appl. Phys. 10 (1966) 886–899.[23] D.K.G. De Boer, Calculation of x-ray fluorescence intensities from bulk and multilay-

er samples, X-Ray Spectrom. 19 (1990) 145–154.[24] A.M. Costa, R. Calixto, A.M. Rossi, G.A. Soares, Preparation and Properties of Zinc

Containing Bioceramics, Key Eng. Mater. 254–256 (2004) 119–122.[25] D. Shepherd, S.M. Best, Production and characterization of zinc substituted hydroxy-

apatite, Key Eng. Mater. 361–363 (2008) 159–162.

[26] Y. Tang, H.F. Chappell, M.T. Dove, R.J. Reeder, Y.J. Lee, Zinc incorporation intohydroxyapatite, Biomaterials 30 (2009) 2864–2872.

[27] H. Chappell, D. Shepherd, S. Best, Zinc Substituted Hydroxyapatite — A Comparisonof Modelling and Experimental Data, Key Eng. Mater. 396–398 (2009) 729–732.

[28] International Organization for Standardization, Biological Evaluation of MedicalDevices — Part 6: Tests for local effects after implantation (ISO 10993-6), 1992.

[29] P.J. Upman, ISO 10993-6: Tests for Local Effects After Implantation, BONEZone,Spring 2006. 50–52.

[30] M.D. Calasans-Maia, M.L. Monteiro, F.O. Ascoli, J.M. Granjeiro, The rabbit as ananimal model for experimental surgery, Acta Cir. Bras. 24 (2009) 325–328.

[31] A.L. Pearce, R.G. Richards, S. Milz, E. Schneider, S.G. Pearce, Animal models forimplant biomaterial research in bone: a review, Eur. Cell Mater. 2 (2007) 1–10.

[32] X. Luo, D. Barbier, N. Davison, Y. Yan, J.D. Brujin, H. Yuan, Zinc in calcium phosphatemediates bone induction: In vitro and in vivo model, Acta Biomater. (2013), http://dx.doi.org/10.1016/j.actbio.2013.10.011.