Available online at www.sciencedirect.com

Acta Biomaterialia 5 (2009) 419–428

www.elsevier.com/locate/actabiomat

In vitro and in vivo behaviour of zinc-doped phosphosilicate glasses

Gigliola Lusvardi a, Davide Zaffe b, Ledi Menabue a,*, Carlo Bertoldi c,Gianluca Malavasi a, Ugo Consolo c

a Department of Chemistry, University of Modena and Reggio Emilia, Via G.Campi 183, 41100 Modena, Italyb Department of Anatomy and Histology, Section of Human Anatomy, University of Modena and Reggio Emilia, 41100 Modena, Italy

c Department of Neurosciences, Head-Neck and Rehabilitation, Section of Dentistry and Maxillofacial Surgery,

University of Modena and Reggio Emilia, 41100 Modena, Italy

Received 6 November 2007; received in revised form 2 May 2008; accepted 7 July 2008Available online 23 July 2008

Abstract

The aim of this work was to study the behaviour of zinc-doped phosphosilicate glasses based on Bioglass� 45S5. In vitro (in simulatedbody fluid), the reactivity was analysed by means of inductively coupled plasma spectrometry, environmental scanning electron micros-copy–energy-dispersive spectroscopy (ESEM–EDS) and X-ray diffraction. In vivo (a rat implanted with glass), the reactivity and the tis-sue behaviour were analysed by conventional histology, histochemistry, microradiography and ESEM–EDS. The in vivo behaviourmatches that in vitro perfectly; they show comparable glass degradation processes and rates, ruled by the amount of zinc in the glass.The reaction mechanism for the formation of a polymerized silica layer superimposed with a peripheral calcium phosphate layer is clearlysubstantiated by ESEM–EDS investigations. The crystallization of a biologically active hydroxyapatite (HA) layer is observed in bothcases; the in vitro experiment shows the presence of HA after 4 days.� 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Keywords: Bioactive glasses; Zinc; In vitro test; In vivo test; Environmental scanning electron microscopy (ESEM)

1. Introduction

In the early 1970s, the concept of bioactive material wasformulated by Hench et al. [1]. Since then, the field of bio-active ceramics has expanded to include a great variety ofcompositions [1–4], and various kind of glasses and glass-ceramics have been found to bond to living bone [5]. Thecommon characteristic of bioactive ceramics is a time-dependent, kinetic modification on the surface that occursupon implantation. When exposed to physiological solu-tions, they form an amorphous calcium phosphate layeron their surface; this layer crystallizes into a biologicallyactive hydroxycarbonate apatite layer [1–3] within a fewdays. This phase is chemically and structurally equivalentto the mineral phase in bone and is responsible for theinterfacial bonding.

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

doi:10.1016/j.actbio.2008.07.007

* Corresponding author. Tel.: +39 0592055042; fax: +39 059373543.E-mail address: ledi.menabue@unimore.it (L. Menabue).

The first and best-studied bioactive glass is Bioglass�

45S5 [1], of molar composition 1Na2O–1.1CaO–0.1P2O5–1.9SiO2. The development of Bioglass 45S5 enabled therequirements for a silicate glass to be considered a bioglassto be defined: optimized surface reactivity and the forma-tion of a stable chemical bond between the tissues andthe glass. These requirements can be summarized as a con-tent of SiO2 less than 60% (molar composition), a high Ca/P ratio and a high percentage of alkaline/alkaline earthoxides.

Glasses of very different compositions have been studiedto evaluate their mechanical strength up to a level compa-rable to cortical bone, thus overcoming the most severelimiting factor in medical use [6–8]. Kokubo et al. [9,10]also revealed that the same kind of biologically active apa-tite layer can be formed on the surfaces of bioactive glassesand glass-ceramics even in an acellular simulated bodyfluid (SBF), with ion concentrations and pH nearly equalto those of human blood plasma [11,12].

vier Ltd. All rights reserved.

Table 1Batch composition (mol.%) from the general glass formula 1Na2O–1.1CaO–0.1P2O5–1.9SiO2–xZnO

Glass x SiO2 Na2O CaO P2O5 ZnO

H 0 46.2 24.3 26.9 2.6 –HZ5 0.16 44.4 23.4 25.9 2.5 3.8HZ10 0.32 42.5 22.5 24.8 2.4 7.8HZ20 0.78 38.8 20.5 22.6 2.2 15.9

420 G. Lusvardi et al. / Acta Biomaterialia 5 (2009) 419–428

By varying the chemical nature and the concentration ofthe glass constituents [13], new important biological prop-erties (bacteriostatic, cariostatic, etc.) can also be addedand the glass can be tailored to specific clinical applica-tions. For example, it is known [13] that the addition ofzinc to silicate and borosilicate glasses improves the ther-mal and mechanical properties, and in the case of phos-phate glasses improves the chemical durability in water.It has been also demonstrated that zinc, an essential traceelement, manifests stimulatory effects on bone formationboth in vitro and in vivo; in fact, the slow release of zincincorporated into an implanted material promotes boneformation around the implant and accelerates the patient’srecovery [14,15]. The partial replacement of sodium withzinc in bioactive glass materials has been proposed to stim-ulate bone cells proliferation and differentiation, and toimprove the bone-bonding ability of bioglass [16].

Our previous studies [13], based on experimental andcomputational approaches (molecular dynamics simula-tions), revealed that, in phosphosilicate glasses, zinc adoptsa tetrahedral coordination irrespective of its concentrationand copolymerizes with the silicon tetrahedral, causing anoverall increment of glass reticulation with respect to Bio-glass 45S5. This is expressed by the Qn species distribution,which is defined as the number of bridging oxygens sur-rounding a network formed of ions such as silicon or phos-phorous. This behaviour explains the increment in thechemical durability (tests in water) of zinc-bearing glasses.

Moreover, our studies based on cell culture tests (MC-3T3osteoblast cells) and related cytotoxicity tests [17] allowedthe development of a zinc-bearing glass (hereafter HZ5) withan optimal Zn/P ratio for the maintenance of cell adhesionand cell growth, comparable to the Bioglass 45S5.

Conventionally, the in vivo bioactivity of materials isevaluated using animal experiments, including the bondingstrength of bone to the material or the bone ingrowth ratioamong the material particles [18]. In the present work, weinvestigated the reactivity of phosphosilicate glasses basedon Bioglass 45S5 and doped with different amount of zincoxide by means of in vitro and in vivo tests. In the firstcase, the glasses were tested in simulated body fluid(SBF) and their reactivity analysed by means of inductivelycoupled plasma spectrometry (ICP), environmental scan-ning electron microscopy–energy-dispersive spectroscopy(ESEM–EDS) and X-ray diffraction (XRD). In the secondcase, glasses were implanted into the dorsal muscles orinside a calvaria pocket in rats (without press-fit fixationto bone); their reactivity and the response of host tissueswere analysed by means of conventional histology, histo-chemistry, microradiography and ESEM–EDS.

2. Materials and methods

2.1. Specimen preparation

Three different types (Table 1) of doped glasses wereprepared with the composition 1Na2O–1.1CaO–0.1P2O5–

1.9SiO2–xZnO and, as a reference, a glass with the compo-sition corresponding to that of Bioglass� 45S5 (hereafter Hglass) was also prepared.

About 100 g of batch was obtained by mixing reagentgrade SiO2, Na2CO3, CaCO3, Na3PO4�12H2O and ZnCO3

raw materials in a sealed polyethylene bottle for 1 h. Pre-mixed batches were put into a 50 ml platinum crucibleand melted in an electric oven for 2 h at 1550 �C.

The melts were poured into graphite moulds of differentdimensions in order to obtain rectangular specimens(10 � 10 � 1 mm3 in size) and cylinders (Ø 3 mm, length6 mm) for in vitro and in vivo study, respectively; someglasses were also ball milled in agate jars and sieved to pro-duce particles in the size range of 600–1000 lm for in vivotesting.

2.2. In vitro bioactivity assay in SBF

The rectangular glass specimens were washed with pureacetone and immersed in 24 ml of SBF, prepared accordingto Kokubo et al. [12]. A constant solid/liquid ratio wasmaintained in all cases; the SBF soaking was carried outat 37 �C and for various times (1, 2 and 10 h, and 4, 15,30 and 60 days).

2.3. In vivo experiments

Experiments were performed in compliance with theItalian laws on animal experimentation. The animalresearch protocol was approved by the Ethical Committeeof the University of Modena and Reggio Emilia as requiredby Italian law.

Adult male Sprague–Dawley rats (n = 16, Charles RiverItalia, Italy), with an average body weight of 600 g (range550–650 g, 180 days of age), were used for the glassimplants. The rats were kept in stainless-steel cages sup-plied with a self-wash system, air conditioning and lighting,in agreement with Italian guidelines on housing of labora-tory animals. They were given at a daily ‘‘good laboratorypractice” diet (4RF21, Charles River Italia) and water adlibitum, and checked with regard to health status and bodyweight. The glass insertion was performed under generalanesthesia induced by an intramuscular injection of a mix-ture of 30 mg per kg body wt. ketamine (Ketavet 100,Intervet prod. S.r.l., Italy), 0.75 mg kg�1 acepromazine(Prequillan, Fatro, Italy) and 4 mg kg�1 xylazine (Rom-pun, Bayer, Germany).

G. Lusvardi et al. / Acta Biomaterialia 5 (2009) 419–428 421

Two glasses, HZ5 and HZ20 (Table 1), were testedsingly by inserting a cylinder (Ø 3 mm, l 6 mm) and300–400 mg of powdered glass (600 � 1000 lm) of thesame glass type (the corresponding micrograph isreported as Supplementary Material) in the rat dorsalmuscles (five rats for each type of glass). Antibiotic treat-ment (Rubrocillina, Intervet) was performed after surgeryto reduce the risk of infection. Rats were sacrificed after30 days by ether euthanasia. Flat specimens

Fig. 1. Element concentrations in the SBF solution vs. soaking

(4 � 3 � 1.5 mm) of H and HZ5 glasses were singlytested by inserting them into intrabony pockets, pro-duced by piezosurgery in the external surface of the ratcalvaria, exactly matching the specimens (a micrographof the implanted glasses is reported as SupplementaryMaterial, Fig. 1). Glasses were maintained in situ by fas-cial and tegument covering and repositioning. Six ratswere used for each glass type: three were sacrificed after1 day and the others after 30 days.

time of: (a) H, (b) HZ5 and (c) HZ10 and HZ20 glasses.

422 G. Lusvardi et al. / Acta Biomaterialia 5 (2009) 419–428

2.4. Characterization

Before and after soaking in SBF, the surface morphol-ogy of the specimens was examined by means of ESEM(FEI Quanta 200, Fei Company, The Netherlands) andEDS (INCA 350, Oxford Instruments, UK). These speci-mens were also analysed by means of XRD (PanalyticalX0PertPro, The Netherlands). Changes in the concentra-tions of sodium, calcium, silicon, zinc and phosphorous,after SBF soaking, were measured by means of ICP (PerkinElmer Optima 4200 DV, USA). Each measurement wasderived and elaborated from four measurements [19].

During the SBF testing, changes in the pH were mea-sured by a pH meter (Orion 900A, UK) equipped with aprinter and with automatic temperature compensation.

2.5. Histology

All excised specimens containing the glasses were firstfixed in 4% paraformaldehyde in 0.1 M phosphate buffer(pH 7.2) for 2 h at room temperature, then dehydratedand embedded in methyl methacrylate (MMA) at 4 �C,as reported elsewhere [19].

Thick (500 lm) and semi-thick (150 lm) sections wereobtained from embedded samples using a diamond sawmicrotome (1600 Leica Microsystems GmbH, Germany).Sections were perfectly polished by alumina and X-raymicroradiographed (semi-thick sections) using a Compact3K5 X-ray generator (Italstructures, Italy). Thick sectionswere fastened on a stub and characterized by means ofESEM–EDS. The glass was manually removed from somesections containing cylinders or calvaria specimens and thesection re-embedded before thin sectioning. Thin (5 lm)sections were obtained from PMMA specimens using anAutocut 1150 bone microtome (Reichert-Jung GmbH,Germany); these sections were stained with toluidine blue,trichrome Gomori stain, alkaline phosphatase (ALP) andtartrate resistant acid phosphatase (TRAP), as reportedelsewhere [20]. Microradiographs and thin sections wereanalysed by means of Axiophot microscope (Carl Zeiss,Germany) under ordinary or polarized light.

3. Results

3.1. In vitro bioactivity assay in SBF

In agreement with the literature [1–3,9,10], H glass degra-dation takes place after a few hours of immersion in SBF: theconcentrations of sodium and silicon increase, while those ofcalcium and phosphorus decrease (Fig. 1a). In the case ofzinc-doped glasses, some differences were observed, depen-dent on the zinc content: the glass with the lower zinc content(HZ5) showed behaviour (Fig. 1b) quite similar to that ofzinc-free glass (H), especially in terms of the decrease inphosphate concentration, which reached a constant valueafter 14 days of immersion. At a higher zinc content(HZ10, Fig. 1c), the phosphate concentration reached an

equilibrium value after 30 days of immersion; for the HZ20glass, it was possible to observe a slight variation in the phos-phate concentration after 60 days (Fig. 1d).

The sodium concentration increased with time for alldoped glasses, but the equilibrium value was generallylower with respect to that of H and was reached after alonger time; a similar trend was seen for the silicon concen-tration. A continuous calcium concentration increase wasobserved for HZ5, HZ10 and HZ20. These results confirmthe chemical durability (water tests) behaviour observedpreviously [13]; the addition of zinc causes a global reduc-tion in ion leaching. The presence of zinc in the solutionwas clearly established only for the HZ20 glass, with itsconcentration reaching a value of 4.6 ppm after 60 days;this value is significantly smaller with respect to the averagefound in human blood plasma (6.4 ppm) [17].

The pH increased especially for H and HZ5, with DpHof 1.10 and 0.85, respectively; for HZ10 and HZ20, theincrease was lower, at 0.15 and 0.10, respectively (corre-sponding graphs are reported as Supplementary Material,Fig. 2).

The morphology of the surface of undoped and dopedglasses was investigated by ESEM. Some differences wereobserved dependent on the zinc content. For H and HZ5the morphology changed with respect to that of as-quenched glasses after a few hours of contact and somecracks were observed on the surface (Fig. 2a and b); anincreasing amount of zinc slowed down these modificationsand the surface presented only the initial corrosion. ForHZ20 glass, uniform glass corrosion was clearly observedafter only 4 days of soaking.

The presence of zinc retards the glass degradation (espe-cially at high zinc content); however, in all cases, the degra-dation increases with time. For the H and HZ5 glasses,different superimposed layers and spherical particles becamedistinguishable on the outermost layer. The morphology andcomposition of these layers and spherical particles becameclearer with time; during the reaction, the above-mentionedparticles covered the glass surface. Fig. 2c and d shows thedifferent particle distributions of H and HZ5 after 14 days;in the latter case the distribution is not perfectly homoge-neous. In the case of HZ10 and HZ20, irregular layers weredetected only after 60 days; there was no evidence of the for-mation of spherical particles.

EDS spectra, after SBF soaking, revealed that theglasses became poorer in sodium compared to the as-quenched glasses. The internal layer was constituted essen-tially of silicon, while high concentrations of calcium, phos-phorus and zinc were observed in the most peripheral layer.Generally, in the zinc-rich glasses, zinc was identified bothinternally and in the peripheral layer, with the higher con-centration in the peripheral layer.

The spherical particles were essentially constituted ofcalcium and phosphorus, with the Ca/P molar ratio (calcu-lated by peak area) of 1.63 being close to the stoichiometricvalue (1.67) of hydroxyapatite (HA, Ca10(PO4)6(OH)2); inthe case of HZ5, it was also possible to observe aggregates

Fig. 2. ESEM micrographs of H (a) and HZ5 (b) glasses after 10 h ofsoaking in SBF solution. ESEM micrograph of H (c) and HZ5 (d) glassesafter 14 days of soaking in the SBF solution. Bars: a = 0.02 mm,b = 0.05 mm, c = 0.005 mm, d = 0.002 mm.

Fig. 3. HZ5 (a) and XRD patterns of H (b) glasses after 14 days ofsoaking in the SBF solution, and H (c) and HZ5 (d) sections of anembedded specimen removed after 30 days from the rat. Characteristicpeaks of phases are indicated with symbols: Ca10(PO4)6(OH)2 (;),Zn3(PO4)2�4H2O (.) and SiO2 (O).

G. Lusvardi et al. / Acta Biomaterialia 5 (2009) 419–428 423

with a Ca/P ratio of �1 and a Zn/P ratio of �0.4. Some-times, silicon was also detected on the external layer ofthe zinc glasses; this is related to the non-uniform forma-tion of the layers.

The XRD patterns for all as-quenched glasses werecharacteristic of soda-lime phosphosilicate glasses (thesepatterns are reported as Supplementary Material, Fig. 3);the presence of zinc did not affect XRD patterns at highzinc content, and in all cases a broad hump was observedcentred at about 32–33� (2h range), corresponding to aninterplanar distance (d value) of 2.71–2.79 A. Generally,the characteristic peak of vitreous silica is centred at21.30� (2h range, d = 4.17 A); this shift to a lower d valueis due to the presence of metal oxides, such as sodium, cal-cium and zinc oxides, in the glass.

After the SBF soaking, the XRD patterns changed,especially with regard to the broad hump, the intensity ofwhich increased and became more resolved. By matchingthe JCPDS files [21], the XRD spectra for H and HZ5 after4 days revealed a number of peaks, identified as the mostintense peaks of HA. All the peaks became better resolvedwith time; this was especially evident for H and HZ5.

The patterns of the H and HZ5 glasses after 14 days ofSBF soaking (Fig. 3a and b) revealed characteristic peaksof HA at 25.90, 29.10 and 31.74� (2h range, d = 3.44,3.08 and 2.81 A, respectively); the peaks identified in theH glass spectrum were better resolved than those of HZ5,indicating better and more rapid crystallization of HA.

In the HZ5 pattern (Fig. 3a), new peaks were recognizedat 9.72 and 19.32� (2h range, d = 9.10 and 4.57 A, respec-tively) and identified as those of a secondary phase, thatis, a zinc-containing phase of formula Zn3(PO4)2�4H2O.Also identified, in agreement with the EDS results, was acharacteristic peak of SiO2, in the form of silica gel, centredat about 21.22� (2h range, d = 4.18 A); the presence of thispeak prevented the identification of other characteristicpeaks of crystalline phases.

The XRD spectra of HZ10 and HZ20 were poorlyresolved and gave no clear evidence for the presence ofany crystalline phase.

424 G. Lusvardi et al. / Acta Biomaterialia 5 (2009) 419–428

3.2. In vivo results

No histological differences were observed in the tissuesurrounding HZ5 or HZ20 glasses. A fibrous layer rich incollagen fibers and fibroblasts surrounded the glassesimplanted in muscles. The thickness of the fibrous layerranged from 50 to 250 lm around cylinders and decreasedto 10–200 lm around the powdered glasses (Fig. 4a–c). Nogranulation tissue was found around glasses after 30 days.Only fibers and fibroblasts were found in the thinnest layers(Fig. 4b), whereas capillaries and large venous vessels orseveral arterioles and venules were detectable inside thethickest layers (Fig. 4c). No multinucleated cells or multi-nuclear TRAP-positive cells (Fig. 4d) were detected on

Fig. 4. Optical microscope images of (a) HZ5 cylinder and (b–e) granules impGomori, (d) TRAP and (e) ALP methods. (a) Cylinder enveloped by the musclefrom the HZ5 granules surface; (c) several vessels (arrows) located close to the(up). (d) Sole (mononucleate) TRAP-positive cell (arrow) found in the fibrous tlocated close to the HZ5 granule surface. Bars: a = 2 mm; b, c, e = 50 lm; d =the reader is referred to the web version of this article.)

the glass surface or inside the layer surrounding the glasses.On the contrary, a great number of cells positive to ALPwere found inside the fibrous layer, in close vicinity tothe HZ5 glasses (Fig. 4e).

No morphological or compositional differences weredetected between cylinders and powders. After implanta-tion, HZ20 glass degradation was very poor compared tothe as-quenched glass (Fig. 5a–c) and the maximum valueof thickness was about 25 lm; the HZ5 glass showed amore marked degradation of layers, with thickness rangingfrom 20 to 150 lm (Fig. 5a and b). The layer closest to theintact glass was constituted only of silicon (Fig. 6), while anincreased amount of phosphorus and calcium was observedin the peripheral layer with respect to the intact glass. The

lanted into muscles for 30 days. Thin sections stained with (a–c) trichrome. (b) A very thin (about 10 lm) fibrous layer separates the muscle cells (red)HZ5 granules where fibrous offshoots propagate inside its degraded layerissue separating two HZ5 granules. (e) Several ALP-positive cells (arrows)25 lm. (For interpretation of the references to color in this figure legend,

Fig. 5. SEM micrographs of glass granules: (a) HZ5 as-quenched; (b) HZ5and (c) HZ20 implanted into muscles for 30 days. (b) Irregular thickness ofthe HZ5 degraded (peripheral) layer; (c) the very thin HZ20 degraded(peripheral) layer. Bar: a = b = c = 100 lm.

Fig. 6. SEM micrograph and topographic EDS analyses for Na, Si, P, K,Ca and Zn of a thick section of the peripheral part of a HZ5 granule,implanted into muscles for 30 days. The dark points of the degraded layerof Na (1.04 keV) are not due to the Na presence, but to the overlapping ofZn-L peak (1.01 keV; see Fig. 7). P and Ca form a thick layer overlappingthe peripheral part of the thin Si-rich layer. K and Zn have an almostuniform distribution inside the degraded layer. Bar = 50 lm.

G. Lusvardi et al. / Acta Biomaterialia 5 (2009) 419–428 425

amount of zinc was similar to that of the intact glass, and asmall amount (�4–5 wt.% by EDS) of potassium was alsoobserved (Figs. 6 and 7); this element is absent in the intactglass.

In the H glass, the flat specimens implanted for 1 dayshowed a dark degraded layer with a thickness of about10–20 lm; silicon was detected in all layers, whereas onlythe peripheral layer contained also calcium and phospho-rus (Fig. 8). In contrast, the HZ5 glass showed an unevendegraded layer with a thickness of about 1–5 lm; siliconwas not homogenously distributed throughout the layers,with the peripheral layer essentially constituted of calciumand phosphorus (Fig. 8).

The flat specimens implanted for 30 days were com-pletely enveloped by a fibrous layer, which was moreordered around the HZ5 glass (Fig. 9), the thickness of

which could reach 500–1000 lm laterally and towards thetegument. For the H glass, the degraded layer showed adark appearance, with a thickness of about 200–250 lm(Fig. 9). The calvaria holding the specimens showed rem-nants of pre-existing bone fragments and very disorderedbone formations, most likely newly formed, both erodedby osteoclasts (Fig. 9). For the HZ5 glass, the degradedlayer had a multilaminar appearance with an uneven thick-ness of about 20–100 lm (Fig. 9). Part of the retaining cal-varia was absent and the residual pre-existing bonerevealed previous erosion activity. Osteoblasts was foundto have formed new bone in apposition to the pre-existingbone supporting the HZ5 flat glass (Fig. 9).

XRD patterns obtained from sections of PMMA-embedded specimens of the H and HZ5 glasses, removedfrom the rats after 30 days, revealed the characteristic

Fig. 7. EDS graphs of the as-quenched HZ5 glass (gray) and its degradedlayer (black) (star in Fig. 6). Graphs are plotted on the same scale. Note, inthe black graph, the presence of Zn (L peak at 1.01 keV), the presence of K(arrow), the reduced amount of Si (but still present) and the increasedamounts of P and Ca.

Fig. 8. SEM micrographs and topographic EDS analyses for Si, P and Caof thick sections of the corner of the H (left) and HZ5 (right) flat glassesimplanted into calvaria for 1 day. The two glasses show a degraded(peripheral) layer of different thickness and Si amount. A thin Ca–P richlayer is already present after 1 day (arrows). Bar = 25 lm.

426 G. Lusvardi et al. / Acta Biomaterialia 5 (2009) 419–428

peaks of HA (Fig. 3c and d); the peaks’ resolution and con-sequently the degree of crystallinity of HA in the twoglasses were very similar, indicating a delay in HA crystal-lization due to the interaction between the glass and the liv-ing tissues. The broad hump at about 15� (2h range) is dueto material used for the sample preparation (PMMA-embedded specimens; see Section 2.5).

4. Discussion

The in vitro results are in agreement with our previousstudies on zinc-containing glasses [13]; zinc improves thechemical durability of phosphosilicate glasses, and in factacts as a chaperon for the insertion of phosphorus into thethree-dimensional glass network and its progressive incorpo-ration into the network as a function of zinc content. Theeffect of zinc on glass reticulation is mainly evident at highzinc content (ZnO P 10%), and this is consistent with theincrease in chemical durability both in vitro and in vivo,and with the low degree of surface modification of theHZ20 glass. The small differences between H and HZ5 interms of chemical durability and surface reactivity is in linewith the similarity of their glass structures, because 5% ofZnO can increase the Qn value of silicate groups but is notable to significantly increase the incorporation of phosphategroups into the glass network [13].

It is worth noting that the zinc addition does not preventthe formation of HA on the glass surface of HZ5 withinjust a few days, and after 15 days of immersion HA isaccompanied by a second phosphate-containing crystallinephase, Zn3(PO4)2�4H2O. The presence of two metal phos-phates could explain the presence of phosphate-rich parti-cles with a Ca/P ratio of �1. The formation of thepoorly soluble zinc phosphate could be responsible forthe absence of zinc in the solution.

The similarity between the behaviour of H and HZ5in vivo is also demonstrated by the XRD patterns(Fig. 3c and d) in terms of the phase identified, the resolu-tion of the peaks and the degree of crystallinity of HA; as aconsequence, we can state that the addition of zinc to Hglass also does not inhibit HA formation in vivo and wecan define HZ5 as a bioactive glass.

The in vivo results confirmed the influence of the zinc con-tent on the formation of degraded layers, the thickness ofwhich was inversely dependent on zinc content. In the caseof the HZ5 glass, a peripheral calcium–phosphorus layer,in both the muscle and calvaria implantations, was detectedalready after 1 day, and was best defined after 30 days. Incomparison to the as-quenched glass, EDS analysis revealeda similar amount of zinc, an increased amount of calciumand phosphorus in the peripheral layer, an increased amountof silicon in the internal layer and the presence of potassium,absent in the intact glass. The almost constant amount ofzinc found inside the degraded layer of the HZ5 glass couldprove that this element is released slowly into the biologicalenvironment, thus stimulating osteogenic processes[15,16,22]. The presence of potassium was due to ion

Fig. 9. SEM micrographs (a and c) and toluidine blue-stained (b and d) thin sections (after glass removal and re-embedding) of H (a and b) and HZ5 (cand d) flat specimens implanted into calvaria for 30 days. (a) Several bone remnants, most showing erosion surfaces (indented contours). (b) A great(multinucleated) osteoclast (arrow) eroding the bone beneath of the glass (H). (c) New bone (arrow) formed in apposition to the bone spared byosteoclasts; note also that much bone has been resorbed (below and on the left of the arrow). (d) Active osteoblasts (arrows) forming new bone inapposition to the pre-existing bone inside the fibrous tissue separating the bone from the HZ5 glass. Compare in (a) and (c) the aspect and the thickness ofthe degraded layer of the two glasses. Bars: a, c = 100 lm; b, d = 50 lm. (For interpretation of the references to color in this figure legend, the reader isreferred to the web version of this article.)

G. Lusvardi et al. / Acta Biomaterialia 5 (2009) 419–428 427

exchange between the glass and the biological fluids [1–3].Histology showed no adverse reaction of the tissues aroundnot only the H glass but also the HZ5 glass. Contrary toVogel et al. [23], no granulation tissue, multinucleated mac-rophages or lymphocyte aggregates were found around theglasses. Moreover, the presence of several vessels andenhanced ALP activity expressed by the peri-implant cellsfound in the muscle implants indicated a environment favor-able for osteogenic processes.

The muscle implant was chosen as the first step of thein vivo evaluation [24]. This kind of implantation maysometimes be limited, but it clearly allows the rejection ofnon-biologically acceptable materials. The thin connectivelayer we sometimes observed between the muscle and theHZ5 glass revealed how well accepted this glass is in thebiological environment.

The calvaria implant was a very critical implant condi-tion. Glasses were located into precisely matched excava-tions produced on the external half of rat calvaria. In

this area, only very good biomaterials allow healing pro-cesses, as any perturbing effect will enhance the unavoid-able reaction to surgery. The surgery was performedusing a piezosurgical instrument, the best device for pro-ducing minimally injurious osteotomies [25], but much ero-sion of the remaining part of calvaria occurred anyway.After 30 days, this erosion was sometimes active aroundthe H glass, while osteogenic processes started aroundHZ5; at least this evidence highlights the good bioaccept-ability of the HZ5 glass. The absence of bone formationon the glasses was exclusively due to mechanical reasonssince the glasses were not press-fit fixed.

5. Conclusions

The in vivo behaviour of zinc-containing phosphosili-cate glasses agrees well with their in vitro behaviour; theglass degradation processes and its rate are similar in bothcases, and strictly dependent on zinc content. Also, with

428 G. Lusvardi et al. / Acta Biomaterialia 5 (2009) 419–428

high zinc content in the glass, a very low zinc concentrationis detected after 60 days of immersion in SBF and it is stilllower than the human blood plasma value.

From in vivo studies, the almost constant amount ofzinc found inside the degraded layer of HZ5 glass couldprove the slow release of this element into the biologicalenvironment, thus stimulating osteogenic processes.

In the case of HZ5 glass, in both cases, there is evidenceof HA formation and in vivo it shows a goodbioacceptability.

These reassuring results prompt us to continue ourstudy on zinc-containing glasses in the future, performingbone (press-fit fixed) implants and comparing the resultswith those of well-known glasses.

Acknowledgements

The authors wish to thank Prof. M. Del Bue, S. Parisoli,S. Guazzetti, Department of Animal Health, University ofParma, for animal surgery and treatment assistance, Mr.Vincenzo Molino, Centro Servizi Stabulario Interdiparti-mentale, for his remarkable technical assistance with theanimal care and treatment, and the ‘‘Centro Interdiparti-mentale Grandi Strumenti” (CIGS) of the University ofModena e Reggio Emilia for making the instruments avail-able. The Research University Fund (FAR) of the Depart-ments of Anatomy and Histology, University of Modenaand Reggio Emilia, supported this investigation. Theauthors thank MIUR-Ministero dell’Universita e dellaRicerca of Italy for financial support (Cofin06, Prot.2006032335_004).

Supplementary data

Supplementary data associated with this article can befound, in the online version, at doi:10.1016/j.actbio.2008.07.007.

References

[1] Hench LL, Splinter RJ, Allen WC, Greenlee Jr TK. Bondingmechanism at the interface of ceramic prosthetic materials. J BiomedMater Res 1971;5(6):117–41.

[2] Hench LL. Bioceramics: from concept to clinic. J Am Ceram Soc1991;74:1487–510.

[3] Hench LL. The story of bioglass. J Mater Sci: Mater Med2006;17:967–78.

[4] Gross U, Kinne R, Schmitz HJ, Strunz V. The response of bone tosurface active glass/glass–ceramics. CRC Crit Rev Biocompat1988;4(2):155–79.

[5] Kim HM, Miyaji F, Kokubo T. Bioactivity of Na2O–CaO–SiO2

glasses. J Am Ceram Soc 1995;78:2405–11.

[6] Kokubo T, Shigematsu M, Hagashima Y, Tashiro T, Nakaura T,Yamamuro T, et al. Apatite and wollastonite containing glass-ceramics for prosthetic application. Bull Inst Chem Res Kyoto Univ1982;60(3–4):260–8.

[7] Beall GH, Chyung K, Watkins HJ. Mica glass-ceramics. US PatentNo. 3801295; 1974.

[8] Vogel W, Holand W, Naumann K, Gummel J. Development ofmachineable bioactive glass ceramics for medical uses. J Non-CrystSolids 1986;80(1-3):34–51.

[9] Ohtsuki C, Kokubo T, Yamamuro T. Mechanism of apatiteformation on CaO–SiO2–P2O5 glasses in a simulated body fluid. JNon-Cryst Solids 1992;143:84–92.

[10] Filgueiras MR, Torre GR, Hench LL. Solution effects on the surfacereactions of a bioactive glass. J Biomed Mater Res 1993;27(4):445–53.

[11] Gamble J. Chemical anatomy, physiology and pathology of extra-cellular fluid: a lecture syllabus. 6th ed. Cambridge, MA: HarvardUniversity Press; 1967.

[12] Kokubo T, Kushitani H, Sakka S, Kitsugi T, Yamamuro T. Solutionsable to reproduce in vivo surface-structure changes in bioactive glass-ceramic A-W. J Biomed Mater Res 1990;24:721–34.

[13] Linati L, Lusvardi G, Malavasi G, Menabue L, Menziani MC,Mustarelli P, et al. Qualitative and quantitative structure–propertyrelationships analysis of multicomponent potential bioglasses. J PhysChem B 2005;109:4989–98.

[14] Ito A, Kawamura H, Otsuka M, Ikeuchi M, Ohgushi H, Ishikawa K,et al. Zinc-releasing calcium phosphate for stimulating bone forma-tion. Mater Sci Eng C 2002;22:21–5.

[15] Yamaguchi M, Igarashi A, Ychiyama S. Bioavailability of zinc yeastin rats: stimulatory effect on bone calcification in vivo. Journal ofHealth Science 2004;50(1):75–81.

[16] Oki A, Parveen B, Hossain S, Adeniji S, Donahue H. Preparation andin vitro bioactivity of zinc containing sol–gel-derived bioglassmaterials. J Biomed Mater Res A 2004;69(2):216–21.

[17] Lusvardi G, Malavasi G, Menabue L, Menziani MC, Pedone A,Segre U, et al. Properties of zinc releasing surfaces for clinicalapplications. J Biomater Appl 2008;22(6):505–26.

[18] Fujibayashi S, Neo M, Kim H-M, Kokubo T, Nakamura T. Acomparative study between in vivo bone growth and in vitro apatiteformation on Na2O–CaO–SiO2 glasses. Biomaterials2003;24:1349–56.

[19] Taylor JR. An introduction to error analysis, the study of uncertain-ties in physical measurements. Sausalito, CA: University ScienceBooks; 1982.

[20] Zaffe D, D’Avenia F. A novel bone scraper for intraoral harvesting: adevice for filling small bone defects. Clin Oral Implant Res2007;18:525–33.

[21] International Centre for Diffraction Data. The Powder DiffractionFile 2002 [CD-ROOM].Version 2.3.

[22] Bougle D, Sabatier JP, Guaydier-Souquieres G, Guillon-Metz F,Laroche D, Jauzacd P, et al. Zinc status and bone mineralization inadolescent girls. J Trace Elem Med Biol 2004;18(1):17–21.

[23] Vogel M, Voigt C, Gross UM, Muller-Mai CM. In vivo compar-ison of bioactive glass particles in rabbits. Biomaterials 2001;22(4):357–62.

[24] Xin R, Leng Y, Chen J, Zhang Q. A comparative study of calciumphosphate formation on bioceramics in vitro and in vivo. Biomate-rials 2005;26:6477–86.

[25] Stubinger S, Kuttenberger J, Filippi A, Sader R, Zeilhofer HF.Intraoral piezosurgery: preliminary results of a new technique. J OralMaxillofac Surg 2005;63:1283–7.