Composition–structure–property (Zn2+ and Ca2+ ion release) evaluation of Si–Na–Ca–Zn–Ce glasses: Potential components for nerve guidance conduits

Materials Science and Engineering C 31 (2011) 669–676

Contents lists available at ScienceDirect

Materials Science and Engineering C j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / m s e c

Composition–structure–property (Zn2+ and Ca2+ ion release) evaluation of Si–Na–Ca–Zn–Ce glasses: Potential components for nerve guidance conduits X.F. Zhang a, S. Kehoe a, S.K. Adhi b, T.G. Ajithkumar c, S. Moane d, H. O'Shea a, D. Boyd e,⁎ a

Cork Institute of Technology, Cork, Ireland Department of Instrumentation Science, University of Pune, Pune 411007, India Central NMR Facility, National Chemical Laboratory, Pune 411008, India d Shannon ABC Research Group, Limerick Institute of Technology, Limerick, Ireland e Department of Applied Oral Sciences, Biomaterials, and Oral Biology, Dalhousie University, Halifax, NS, B3H 3J5, Canada b c

a r t i c l e

i n f o

Article history: Received 31 August 2010 Received in revised form 23 November 2010 Accepted 29 December 2010 Available online 8 January 2011 Keywords: Bioglass Dissolution Modelling

a b s t r a c t Bioactive glasses have demonstrated tailored therapeutic ion release, primarily with respect to the augmentation of hard tissues. However, controlled degradation and release of therapeutic ions from biomaterials may also play an important role in soft tissue regeneration such as repair of peripheral nerve discontinuities. In this study, three silica based glasses (0.5SiO2–0.2CaO–0.13ZnO–XNa2O–(0.17–X) CeO2) where, (0.04 b X b 0.14) were synthesised and characterised. The local environment of the 29Si isotope was probed for each glass using 29Si MAS–NMR, whilst the thermal characteristics of each glass were examined using DTA. Following these analyses, ion release profiles for Ca2+ and Zn2+ were evaluated; an equivalent specific surface area of 1 m2 of each glass powder was incubated (37 °C) in 10 ml of citric acid buffer and TRIS–HCI buffer solution (pH 3.0 and pH 7.4 respectively) for incubation periods of up to 30 days. The Zn2+ concentration of each filtrate was analysed using flame Atomic Absorption Spectroscopy (Varian AA240FS Fast Sequential AAS) and the Ca2+ concentration of each filtrate was determined using Inductively Coupled Plasma–Mass Spectrometer (Varian 820 ICP–MS). Results obtained from the 29Si MAS–NMR spectra indicated Q2 structures pervading the network. An analytical model was proposed to analyse the ion release profiles for each glass, and indicated heterogeneous dissolution of glass networks. The ion release data demonstrates that ion release in the range (19.26–3130 ppm) for Ca2+ and in the range (5.97–4904 ppm) for Zn2+ occurred. Release of such elements, at appropriate levels, from peripheral nerve guidance conduits may be advantageous with respect to the repair of peripheral nerve discontinuities. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Over the last 20 years many multidisciplinary approaches have been investigated with respect to reconstructing peripheral nerve injuries (PNI) [1]. Much research and development has focused on artificial nerve guidance conduits (NGCs), fabricated from natural or synthetic materials to enclose (entubulate) opposing nerve stumps (of the severed nerve) within a protective microenvironment favoring peripheral nerve regeneration [1]. In clinical use, the most popular artificial NGC is based upon the natural material Type I collagen (e.g. NeuroGen™, Integra LifeSciences Corporation, Plainsboro, NJ) [2]. Collagen is the principle extracellular protein of connective tissue which consists of a family of 28 proteins [3–6]. Type I collagen is typically expressed in vertebrates [7,8] and forms a unique triplehelical structure, which permits the construction of a tubular matrix ⁎ Corresponding author. Department of Applied Oral Sciences, Biomaterials, and Oral Biology, Dalhousie University, 5981 University Ave., Halifax, NS, B3H 3J5, Canada. Tel.: + 1 902 494 6347. E-mail address: [email protected] (D. Boyd). 0928-4931/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.msec.2010.12.016

that provides for mechanical strength and defined permeability [2,9,10]. Type I collagen molecules also provide a biomechanical scaffold for cellular attachment and anchorage of macromolecules [10]; thus, promoting cellular proliferation and tissue healing [4,11]. Consequently, such grafts have proved successful in repairing short gap defects [11]. However, their suitability to longer discontinuities is compromised due to inadequate stability [21]. Current efforts to mitigate against this problem focus on increased cross-linking, however it has been observed that increased cross-linking decreases the likelihood of nerve regeneration and also decreases the number of regenerated fibers in vivo [12]. As an alternative to natural materials like collagen, synthetic NGCs, based on polyphosphoesters, such as: polylactic acid (PLA), polyglycolic acid (PGA), and copolymers of lactide/glycolide, poly(lactic-co-glycolic acid) (PLGA) [1] and poly (DL-lactide-ε-caprolactone) have been considered [1,13,14]. These materials have received much attention due to their degradability [15], ease in processing, and well established safety and efficacy as suture materials [1]. PGA conduits (e.g. Neurotube™; Neuroregen LLC, Bel Air, MD) were amongst the first synthetic NGCs to be investigated in clinical studies for peripheral nerve

670

X.F. Zhang et al. / Materials Science and Engineering C 31 (2011) 669–676

regeneration [16]. PGA has proven to be successful in the clinical repair of digital nerves with defects of up to 3 cm [1,17]. However, PGA degrades slowly and material may persist of up to 4 months after surgery [18]. To improve the degradation rate of PGA, PLGA (copolymer of PLA and PGA) has been synthesised [19]. PLGA may be formed by varying the lactide to glycolide monomer contents, (85:15, 75:25 and 50:50 monomer ratios) for which the degradation rate accordingly adjusts from slowest to fastest respectively; with the 50:50 monomer ratio possessing the fastest degradation rate of 0.2 month of half-life [18,20,21]. Thus, the degradation rates for these materials can be easily modified by tailoring the monomer ratio of lactide to glycolide [22]. It must be noted however, that the sudden and rapid degradation of these materials compromises their ability to repair nerve discontinuities greater than 4 cm [16,19,23]. An additional polyphosphoester based biodegradable material that has demonstrated appropriate support in the peripheral nerve regeneration is poly (DL-lactide-ε-caprolactone) (DLLA-ε-CL) guides (e.g. Neurolac®, Polyganics BV, The Netherlands) [14,24]. Neurolac® is an artificial NGC of a semi-crystalline copolymer of L-lactide and ε-caprolactone (50/50 mol/mol) [25]. However, its properties impact on its biocompatibility; upon implantation it is prone to fragmentation and swelling resulting in compromised regeneration of the nerve [25]. The issue of fragmentation arises in the context of the presence of amorphous and crystalline phases within the material [26,27]. When subjected to physiological conditions, the amorphous component undergoes hydrolytic degradation, once degraded, the scaffold collapses and releases crystalline particles which results in sterile inflammation [26]. Succinctly, present state of the art materials approaches have yielded excellent progress, however much research and development is still required and new biomaterials are being actively investigated to overcome present challenges [16]. Composite materials combining bioactive glasses (BG) and polymers have been explored [28–32]. PLGA does not possess sufficient mechanical strength and stability to support tissue growth. However, adding BG to a PLGA base in the form of a BG/PLGA composite material enhances mechanical and biological properties without compromising degradability [33,34]. BG/polyhydroxybutyrate (PHB) composites [35], also demonstrate in vitro biocompatibility for use in augmentation of peripheral nervous tissue. Furthermore, the in vitro degradation rate of this composite is decreased, thus allowing for improved stability of mechanical properties over time [32,36]. Such composites are found to exhibit a strong interaction with soft tissue [37,38]. The literature also describes improved cell adhesion and proliferation (in vitro) for BG doped polymers [33,38], with in vivo results showing increased vascularisation versus control materials. In summation, initial studies which have examined composites incorporating BG into a biodegradable polymer at varying degrees exhibit [39]: (i) improved biocompatibility, (ii) improved degradability, and (iii) improved mechanical properties. However, the bioactive glasses under consideration have been limited to phosphate-based materials relating to the original bioglass composition [37,40]. An emerging philosophy in tissue engineering is to develop materials that establish key interactions with cells in ways that unlock the body's inherent ability to selfrepair. In this respect the investigation of new compositions of glasses may enhance current state of the art approaches and build knowledge in the area of composition–structure–property relationships where BG composites are leveraged for soft tissue repair. Thus, the investigation of new bioactive glass compositions designed to mediate specific host responses in specific applications is valuable, and may provide for improved outcomes in respect of peripheral nerve regeneration. Consequently, the Si–Na–Ca–Zn–Ce glass system may be of interest. This multi-component system comprises ions that are known to have therapeutic effects on peripheral nerve regeneration; as such the release of degradation products from such glasses in vivo may facilitate a beneficial host response not seen in conventional BG composites for this indication [41–43]. Specifically, achieving

optimal level of intracellular calcium within a NGC may promote normal neurite elongation and growth cone motility [44–51]. Experiments conducted to alter the intracellular calcium level by addition and reduction of the calcium ionphone A23187of cultured chick dorsal root ganglia [52] demonstrated that calcium can act as a regulator of actin filaments in living cell. Further investigations pertaining to the influence of calcium on the growth cone behaviour in xenopus embryos have been carried out by Gamez and Spitzer and demonstrated the importance of Ca2+ on growth cone guidance [48]. The literature demonstrates that calcium transients in growth cones in vivo are a major determinant of axonal extension rates and provide for growth cone guidance [47–49]. Thus, release of Ca2+ at levels appropriate to mediate such responses may be beneficial with respect to design and performance of a NGC. Synergistically, the Zn2+ ion also has specific applicability to peripheral nerve regeneration. There is a growing awareness that zinc plays a role as a signaling substance through the body [53], in particular zinc ions are liberated to the synaptic space during normal neuronal activity. The physiological significance of zinc as a modulator of neuronal excitability in toad olfactory epithelia has been evaluated by Felip et al. [54]. Felip found that at low concentrations (0.065 to 3.27 ppm) Zn2+ impacts on the opening kinetics of sodium channels (the channel is voltage gated: it opens in response to a small depolarisation of the cell [55]) without altering the closing kinetics and induced a concentration dependent increase in the neuronal firing rate. In contrast, at higher concentrations of Zn2+ (6.5 to 32.7 ppm), inhibited sodium and delayed rectifier-type potassium currents and decreased firing rates were observed [54]. Zn2+ may also act to inhibit bacterial growth and may facilitate multiple therapeutic effects, such as: (i) neural growth [56], (ii) inhibited bacterial growth at the surgical site [57], and (iii) improve wound healing [58]. Therefore its inclusion in the BG component of a NGC composite may be beneficial. The impact of cerium in respect of neural tissue has received little attention in the literature, however, it has been shown that nano-particles composed of cerium protect nerve cells from oxidative stresses and act as neuroprotectants [59]. Consequently, this study, the first component of multi-part evaluation of a new composite for use as an NGC, aims to evaluate the composition–structure–property relationships, (Ca2+ and Zn2+ ion release profiles) of these multi-component glasses with respect to their potential use as fillers in polymeric (PLGA) composite constructs for peripheral nerve regeneration.

2. Material & methods 2.1. Glass synthesis Three glass compositions (Table 1) were synthesised. Appropriate amounts of analytical grade reagents; silicon dioxide, calcium carbonate, zinc oxide, sodium carbonate and cerium oxide (Sigma Aldrich, Wicklow, Ireland) were weighed out as appropriate and thoroughly mixed by shaking (30 min) in a plastic container. Each batch of powder was then fired (1520 °C, 1 h) in a platinum crucible and shock quenched into water. The resulting glass frit was dried in an oven (120 °C, 1 day), ground and sieved (b45 μm aperture) to retrieve glass powder for subsequent analysis.

Table 1 Composition of the bioglass (mol. fraction). Glass ID

SiO2

Na2O

CaO

ZnO

CeO2

CNG1 CNG2 CNG3

0.5 0.5 0.5

0.04 0.09 0.14

0.2 0.2 0.2

0.13 0.13 0.13

0.13 0.08 0.03

X.F. Zhang et al. / Materials Science and Engineering C 31 (2011) 669–676

2.2. Differential Thermal Analysis (DTA) DTA was used to determine the onset of the glass transition temperature (Tg) for each glass using a combined thermal analyser– thermal gravimetric analyser (DTA-TGA) (Stanton Redcroft STA1640, Rheometric Scientific, Epsom, UK). A heating rate of 10 °C min−1 (up to 950 °C) was used in an air atmosphere with alumina as a reference in a matched platinum crucible. Sample measurements were performed every 6 s, the Tg onset was determined graphically based on the onset point of the glass transition endotherm. The tolerance of the DTA used in this work is 2%. 2.3. Specific surface area determination N2 adsorption/desorption was determined by Brunauer–Emmett– Teller (BET) measurements using an ASAP-2010 surface area analyser (Micrometrics Instrument Corporation, Norcross, USA) to determine the specific surface area (SSA) and porosimetry of the prepared glass powders. Glass samples (~ 0.15 g ± 0.05 g) were placed under a nitrogen atmosphere at 77.35 K with an equilibration interval of 10 s. 2.4. X-ray Diffraction (XRD) XRD was employed to validate the amorphous nature of the glasses. Powdered samples of each glass were pressed to form discs (Ø32 mm × 3 mm). Diffraction patterns were collected using a Philips Xpert MPD Pro 3040/60 X-ray Diffraction Unit (Philips, Eindhoven, Netherlands), with monochromated CuKα (λ = 1.54060 A) radiation at 40 KV and 35 mA. The scanning angle range (2θ) was performed from 10° to 70° with a step size of 0.033423° and step time of 59.69 s. 2.5. Network Connectivity The Network Connectivity (NC) of the glasses was calculated with Eq. (1) [41] using the molar compositions of the glass (see Table 1) NC =

No: BOs−No: NBOs Total No: Bridging species

ð1Þ

where: NC BO NBO

Network Connectivity Bridging Oxygens Non-Bridging Oxygens

2.6. Magic Angle Spinning–Nuclear Magnetic Resonance (MAS-NMR) spectroscopy 29 Si MAS–NMR was used to analyse the different glass samples at 7.05 T using a Bruker AV300 NMR spectrometer (Bruker BioSpin, Germany) equipped with a BLMAS probe (4 mm Ø) and wide bore superconducting magnet operating at a resonating frequency of 59.6 MHz. Powdered samples were packed into a 4 mm zirconia rotor and spun at the magic angle (54.74°) to remove anisotropy effects. The 29Si MAS–NMR spectra of the glasses were recorded at a spinning frequency of 5 kHz using a high power pulse (P1) acquisition of 1.5 μs for silicon, where a total number of signals of 20,000 pulses were accumulated. The 29Si MAS–NMR samples were spun with a recycle/ delay time set to ~ 2 s. 29Si NMR chemical shifts are reported in ppm and recorded at an ambient probe temperature with 29Si referenced externally relative to 2,2, dimethyl-2 silapentane-5-sulfonate sodium salt (DSS). For solid state NMR, shifts recorded using MAS are independent of the isotropic bulk magnetic susceptibility of the sample, and therefore external referencing should be quite accurate. All NMR spectra were recorded in a room for exclusive use of NMR,

671

where the room temperature was kept at 300 K by means of an airconditioner. The estimated error of chemical shift for data is ca. 0.1 ppm for an ideal material; small variations above this value were observed. 2.7. Quantification of degradation products (Zn2+ and Ca2+) TRIS–HCI buffer and citric acid buffer solutions with a pH of 7.4 ± 0.1 and pH of 3.0 ± 0.2 respectively, were prepared to simulate normal and extreme physiological conditions (according to ISO10993-14 [60]). Both buffer solutions were used for the quantification of degradation products (Ca2+ and Zn2+). An equivalent specific surface area of 1 m2 of each glass powder was immersed in 10 ml of each solution (n = 3) in polypropylene tubes maintained at 37 °C in a shaking waterbath (Stuart Sb40, Reagecon, Shannon, Ireland), agitated at 2 Hz (longitudinal movement). Specimens were stored for 1, 3, 7, and 30 days. After each time period, specimens were removed and filtered through Grade 5 Whatman filter paper, the filtrate was retained for ionic content analysis. The Zn2+ concentration of each filtrate was analysed using flame Atomic Absorption Spectroscopy (Varian AA240FS Fast Sequential AAS) using a Zn hollow cathode lamp at a wave length of 213.9 nm. The Ca2+ concentration of each filtrate was determined using Inductively Coupled Plasma–Mass Spectrometer (Varian 820 ICP–MS). Analysis of each each extract was performed in triplicate (n = 3 (extracts per condition), 3 analyses performed on each extract). 2.8. Analysis of ion release profiles The Zn2+ and Ca2+ ion release profiles were fitted in a nonlinear fashion to obtain the analytical function in order to examine degradation. Based on the literature, the first part of the ion release profile was fitted with the function y = a⁎xb, [61], whilst the second part of the ion release profile was fitted with a association exponential function y = y0 + ymax(1 − exp−kx) [62]. Where: ‘y’ and ‘x’ were the ion release concentration in ppm and incubation time in days, respectively; ‘a’ and ‘b’ were two parameters; ‘y0’ was the ion release concentration (ppm) when it started to release; ‘ymax’ was the ion release concentration at infinite times (ppm); ‘tau’ is denoted as the time necessary for ion release to reach 63% of the estimated ‘ymax’ (ppm); ‘k’ is the rate constant, expressed in reciprocal of the ‘tau’ incubation time and units is inverse days; ‘t1/2’ also refer as half-life is the time to reach 50% of final ‘ymax’ value, ‘t1/2 = tau⁎LN (2)’; ‘ts’ was the incubation time at which the results fit both function with the best ‘R2’; ‘R2’ was computed from the sum of the squares of the distances of the points from the best-fit the exponential nonlinear regression determined by the Prism5 (GraphPad Software). The value of R2 is a fraction between 0.0 and 1.0, and the best-fit line with a R2 equals to 1.0.

3. Results 3.1. Glass characterisation Tg and SSA results obtained for each glass are presented in Table 2. The mole fraction of SiO2, ZnO and CaO remains constant for

672

X.F. Zhang et al. / Materials Science and Engineering C 31 (2011) 669–676

Table 2 Tg, SSA and Na2O:CeO2 ratio for each glass. Glass ID

Na2O:CeO2 ratio

Tg (°C)

Specific surface area (m2/g)

CNG1 CNG2 CNG3

0.308 1.125 4.667

655 601 568

8.3299 10.7441 16.0334

all glass compositions, the only variant being the ratio of Na2O to CeO2. The Tg of the three glasses follows the diminishing order of CNG1 N CNG2 N CNG3 within the range of 655 °C (CNG1) to 568 °C (CNG3). The XRD patterns of the prepared glass compositions (CNG1 to CNG3) are shown in Fig. 1, and confirm the amorphous nature of the glass powders. 3.2. Network Connectivity and MAS–NMR spectroscopy Fig. 2 illustrates the 29Si MAS–NMR spectra collected from each glass with chemical shift data collated in Table 3. The peak chemical shift for each glass was −84 ppm, −82 ppm, and −80 ppm for CNG1, CNG2, and CNG3 respectively indicating a predominantly Q2 structure in each glass. The Network Connectivity of each glass was determined to be 2. 3.3. Quantification of degradation products (Zn2+ and Ca2+)

Intensity (arbitrary units)

3.3.1. Zinc ion release profiles The Zn2+ release profiles of each glass composition incubated at pH 7.4 are shown in Fig. 3(a). CNG2 and CNG3 exhibited the highest levels of Zn2+ release after a 30 day period (38.71 ppm and 35.21 ppm, respectively) in comparison to CNG1 which released 21.21 ppm of Zn2+. After 30 days incubation CNG3 released nearly twice as much Zn2+ as CNG1. The Zn2+ release profiles of each glass composition incubated at pH 3 are shown in Fig. 3(b). CNG2 displays the highest level of Zn2+ release after 30 days incubation reaching levels of up to 4904 ppm in contrast to 3438 ppm for CNG3 and 2552 ppm for CNG1. The lowest level of Zn2+ release identified was observed for CNG1 after 1 day (1423 ppm) about half of the Zn2+ release observed for CNG3 (2994 ppm) at the same time period. Table 4(a) and (b) indicate best fit parameters for Zn2+ release as derived from the nonlinear two phase analysis. At pH7.4, CNG1 displayed the longest half life (3.799 days), whilst the half life for CNG2 and CNG3 was 2.026 and 1.446 days respectively. Table 4(b) indicates the best fit parameters for Zn2+ release from the nonlinear two phase plot. At pH3 CNG1 again displayed the longest half life (3.171 days), whilst the half life for CNG2 and CNG3 was 2.627 and 0.5152 days respectively. The shortest half life recorded was for CNG3 at pH3 with t1/2 = 0.5152 days and the longest half life was CNG1 under normal condition with t1/2 = 3.799 days. The value of ts, found by separating the two phase function for the curve fit of the ion release 2000 1500 CNG1 CNG2 CNG3

1000 500 0 10.0

20.0

30.0

40.0

50.0

60.0

Fig. 2.

29

Si MAS–NMR spectra for glasses CNG1–CNG3.

concentration, correlated the half life for three glasses under normal physiological condition; on the other hand, CNG2 had the highest ts value of 4.297 days under extreme physiological condition, while ts = 2.86 days for CNG1, and CNG3 had the lowest of ts = 1.425 days. The R2 obtained for CNG2 and CNG3 at pH7.4 was very high (R2 N 0.95) and R2 obtained for all three glasses under both normal and extreme condition were between 0 and 1. 3.3.2. Calcium ion release profiles The Ca2+ release profiles of each glass composition incubated at pH 7.4 are shown in Fig. 4(a). CNG2 and CNG3 exhibited the highest levels of Ca2+ release after a 30 day period (102.34 ppm and 133.72 ppm, respectively) in comparison to CNG1 which released 47.42 ppm of Ca2+. After a 30 day incubation period CNG3 released over twice as much Ca2+ as CNG1. The Ca2+ release profiles of each glass composition incubated at pH 3 are shown in Fig. 4(b). CNG3 displays the highest Ca2+ release after 30 days incubation reaching levels of up to 3000 ppm in contrast to 2422 ppm for CNG2 and 1539 ppm for CNG1. The lowest level of Ca2+ release identified was observed for CNG1 after 1 day (875 ppm) with about one third of the Ca2+ release observed for CNG3 (2726 ppm) at the same time period. Table 5(a) and (b) indicate the best fit parameters for Ca2+ release as derived from the nonlinear two phase plot. The associated half life of each glass follows the order of CNG2 N CNG1 N CNG3 under normal physiological condition. Under extreme physiological condition, the half life of CNG1 and CNG2 follows the reverse order of that observed at normal physiological condition (CNG1 N CNG2). The dissolution associated with CNG3 does not fit the model outlined, thus did not allow for evaluation of its half life. The shortest half life observed occurred for CNG2 incubated at pH3 where t1/2 = 1.698 days. The longest half life observed occurred for CNG1 incubated at pH3 where t1/2 = 3.0895 days. The value of ts was obtained to separate the two functions for the best fit; whereby the ts value was in the same order as the half life for three glasses under extreme physiological condition.

Table 3 Network Connectivity (NC), predominant Qn structures, 29Si MAS–NMR peak positions and Na2O:CeO2 ratios for each glass. Glass designation

NC

Predominant Qn structure

29 Si peak (ppm)

Na2O:CeO2 ratio

CNG1 CNG2 CNG3

2.0 2.0 2.0

Q2 Q2 Q2

−84 −82 −80

0.308 1.125 4.667

70.0

Position of [2θ° θ°] Fig. 1. XRD pattern for each glass composition (CNG1 to CNG3).

X.F. Zhang et al. / Materials Science and Engineering C 31 (2011) 669–676

a

Zn2+ Concentration (ppm)

50

40 CNG1

30

CNG2

20

CNG3

10

Ca2+ Concentration (ppm)

a

673

150

100

CNG1 CNG2 CNG3

50

0

0 0

10

20

0

30

Incubation Time (Days)

10

20

30

b 6000

b

Zn2+ Concentration (ppm)

Ca2+ Concentration (ppm)

Incubation Time (days)

4000 CNG1 CNG2 CNG3

2000

4000

3000 CNG1 CNG2

2000

CNG3

1000

0

0 0

10

20

30

0

Incubation Time (Days)

10

20

30

Incubation Time (days)

Fig. 3. (a): Zn2+ release profiles under normal physiological condition (pH 7.4) at 1, 3, 7, and 30 days. (b): Zn2+ release profiles under extreme physiological condition (pH 3) at 1, 3, 7, and 30 days.

Fig. 4. (a): Ca2+ release exponential profile under normal physiological condition (pH 7) at 1, 3, 7, and 30 days. (b): Ca2+ release exponential profile under extreme physiological condition (pH 3) at 1, 3, 7, and 30 days.

The R2 obtained for CNG1 and CNG2 under normal physiological condition was very high (R2 N 0.90) in comparison to the R2 value obtained at extreme physiological condition (0.40 b R2 b 0.66) (refer to Table 5(a) and (b)).

capacity of shielding to the nucleus [64]. However, specific resonances within the range are subject to some debate. Elgayar et al. [66,67] and Oliveira et al. [68] associated resonances between −60 ppm and −83 ppm with Q0/1 structural unit, while, Matsuya et al. have identified peak maxima of −81.6 to −104 ppm for commercial ionomer glasses to Q3 to Q4 structures, respectively [67]. Shrikhande et al. [69], on the other hand report values of chemical shift corresponding to Q1, Q2, Q3 and Q4 structural configurations at about −78, −85, −95, and −105, respectively, which is also in good agreement with values reported by Fayon et al. [70]. In addition, it must be noted that Lockyer et al. [71] have identified precedence for peak merging as a result of two resonances corresponding to different Q structures overlapping and merging into one peak. The 29 Si MAS–NMR results (Table 3 and Fig. 2) indicate that the glasses studied are structurally similar, with the exception of the decreased peak width evident with CNG3. The decreased peak width appears to be attributable to a tightening of the peak from a more negative position toward a more positive position about the peak maxima. This feature is likely attributable to the destruction of Q4 species in the network as a result of an increasing Na2O:CeO2. An additional feature associated with the CNG3 spectra is the emergence of a doubleshoulder at a chemical shift between −60 ppm and −20 ppm, indicating the increased prevalence of Q1/Q0 within CNG3 as compared with the CNG1 and CNG2. All spectra obtained exhibit large broad symmetrical peaks with resonating peak maxima at a chemical shift ranging between −84 and −80 ppm: a chemical shift which the authors have previously established as being associated with Q2 units in similar silicate glasses [72]. An increase in the Na2O: CeO2 ratio from 0.308 to 4.667 causes an increase in the chemical shift of the peak maxima in the glass network from −84 to −80 ppm indicating increased disruption of the glass due this altered ratio. In

4. Discussion 4.1.

29

Si MAS–NMR

29 Si MAS–NMR spectra for glasses CNG1 to CNG3 (Fig. 2) indicate chemical shifts between −60 and −120 ppm; typical of Si in tetrahedral configuration [63–65]. The chemical shift is dependent upon the ratio of BOs to NBOs present in the glass, where an increase in the number of NBOs is found to result in peak maxima moving in a positive direction due to increased shielding of the 29Si nucleus. Alternatively, an increase in the number of BOs is associated with a movement of the peak in a negative direction due to a decreased

Table 4 The best fit parameters for Zn conditions. Glass

t1/2 (days)

(a) Nonlinear fit of Zn CNG1 3.799 CNG2 2.026 CNG3 1.446

2+

2+

release under normal and extreme physiological

tau (days)

ts (days)

ymax (ppm)

release under normal physiological condition 5.48 3.102 21.78 2.923 2.4 39.21 2.086 2.05 35.45

(b) Nonlinear fit of Zn2+ release under extreme physiological condition CNG1 3.171 4.574 2.86 2534 CNG2 2.627 4.232 4.297 4934 CNG3 0.5152 0.7433 1.425 3415

R2 0.8877 0.9632 0.9529

0.9116 0.5064 0.396

674

X.F. Zhang et al. / Materials Science and Engineering C 31 (2011) 669–676

Table 5 The best fit parameters for Ca2+ release under normal and extreme physiological conditions. Glass

t1/2 (days)

tau (days)

ts (days)

ymax (ppm)

R2

(a) Nonlinear fit of Ca2+ release under normal physiological condition CNG1 3.766 5.433 3.15 48.99 CNG2 3.955 5.706 3.406 103.7 CNG3 2.077 2.997 2.418 133.8

0.9702 0.939 0.8957

(b) Nonlinear fit of Ca2+ release under extreme physiological condition CNG1 3.895 5.619 3.416 1444 CNG2 1.698 2.45 2.811 2576 CNG3 Intersect NA 1.308 NA

0.6613 0.1586 NA

summary, the peaks broaden towards a more negative direction for lower Na2O:CeO2 ratios. In accordance to literature, it is accepted that the 29Si peaks for each glass implicate the presence of a wide array of Qn structures in the glass network, such that, it may be assumed, based on the NC calculations and the chemical shift data that the Q2 structure is pervasive to glasses CNG1 to CNG3. Furthermore the replacement of Na2O with CeO2 in the glass formulation affects the resultant glass network: such that, a subsequent increase in the Na2O:CeO2 ratio may result in a tightening of the curve about a peak maxima which moves, in a limited way, toward a more positive chemical shift. Similar findings may also be interpreted when comparing the glass structure to the Tg values for the glasses. Finally, the data herein indicates that the Zn2+ ions are predominantly network-modifying and as such is in agreement with work previously reported by the authors [64]. 4.2. Analysis of zinc and calcium ion release profiles The study of glass dissolution has been challenging scientists for more than 40 years [62]; with the heterogeneous and homogenous models preferred in the literature [62,73]. In respect of the heterogeneous model, two stages are recognized during the reaction between glass and an aqueous solution [74]. During the first stage, extraction of alkali metals varies with the square root of time, and for the second stage the extraction is logarithmic with time [62]. In order to examine the dissolution of each glass, and examine the burst effect of ion release observed in the results (Section 3.3), the ion release profiles were fitted, as per the literature, with an exponential equation and a logarithm equation. The first part of the ion release profile was fitted with the function y = a⁎tk, (in Section 3.3, the grey dash line indicated the exponential function); the second part of the ion release profile was fitted with function y = y0 + ymax (1 − exp−kx). (In Section 3.3, the coloured line indicated the logarithm function). ts was defined as the period of time up to which the first stage of dissolution occurs. Therefore, for x b ts, the curve fitting function employed is required to be y = a⁎tk (the ion release concentration is related by the power–law to incubation time). When x = ts, the plot fits both curves. To enable the continuity of the ion release profile, for x N ts, the curve fitting function is y = y0 + ymax (1 − exp−kx), the ion release function is the original concentration y0 (that ion release at x = ts) plus the decay exponential relate to the incubation time. ts was obtained for different glasses under both sets of conditions. The smallest value for ts recorded related to Ca2+ release (CNG3) exposed to the extreme physiological pH and was observed to be 1.3078 days. The largest value of ts recorded related to Zn2+ release (CNG2) exposed to extreme physiological pH and was observed to be 4.297. Two time points were used to estimate the ion release of y ppm at x = ts, with the result parameters being highly dependent on the first set of data (1 day ion release), this means that the observed results of t1/2 and tau may be higher than the original value, consequently t1/2 and tau should be considered as indicators. The t1/2 and tau were obtained for each CNG glass under different physiological conditions

form the Prism5 (GraphPad Software) in Tables 4 and 5. As one would anticipate the data indicates that glass dissolution proceeds at an accelerated pace under extreme physiological conditions (pH = 3) compared to normal physiological conditions. CNG1 had the longest half life followed by CNG2 then CNG3; a result explained by increased sodium in the glass composition leading to increased dissolution [75]. Synergistically, the role of cerium in a glass network as reported by Mansour et al. [76], observed that cerium could act as a network former or a network modifier, dependent on the concentration of cerium present [76,77]. In the current study, the cerium content was reduced from 0.13 mol to 0.03 mol for glasses CNG1 to CNG3, respectively. The lower concentration of cerium was found to limit the cerium environment in the form of Ce3+ (which acts as a glass network former). This observation may indicate that competing process of glass stabilisation and disruption are occurring up to a critical concentration of Na in the glass, hence the observed chemical shift response in Fig. 2 and Table 3. The data observed, and subsequent functional analysis of the degradation profiles indicate that, in general, upon exposure to normal physiological pH each glass undergoes heterogenous dissolution (R2 values range of 0.88–0.97 for heterogenous dissolution model). Interestingly, the heterogenous model of dissolution breaks down when the glasses are exposed to the extreme physiological pH (3); with the exception of CNG1 (R2 = 0.91 for heterogenous dissolution model), all other glass show poor fitting to the heterogenous dissolution model proposed by the literature (R2 as low as 0.39). It is proposed that the poor fit is a function of rapid dissolution and subsequent release of Ca2+, Ce3+, and Zn2+ into the citric acid solution [75]. The ions released decrease the overall chemical durability, with the cation exchange between the glass surface and the aqueous environment (Ca2+, Na+ with H+/H3O+), which results in a pH excursion from de-alkalization in the early periods of immersion–glass dissolution [78]. In addition, the H+ in the acidic environment can promote the breaking of the Si–O–Si bonds and the chain structure of the glass matrix may become destroyed, thus increasing the number of Si and may alter the surrounding pH environment. The Ca2+ ions initially released by the glass may be exchanged with H+ or H3O+ in the solution and precipitate onto the entire surface of the sample to form a hydrogel like layer on the glass surface and further impact the rate of ions released from the glass [79–81]. This gel layer formation is dependent on the glass composition; with different compositions having the capacity to form various multi-layered structures. Further dissolution is found to result in the mass transport of alkali ions from the center of the particulate glass through the hydrogel rich layer into the aqueous media, causing a reduction in the rate at which dissolution of the particulate occurs. The Ca2+ release for CNG3 (at pH 3.2) followed the order: 1 days N 7 days N 3 days which may be attributed to the Ca2+ originally acquired in the hydrogel formation being faster than the process of Ca2+ transfer from the inner phase to the reactive surface layer of glass over 3 days [78,82]. As immersion time increases, the release of glass constituents depends on the penetration of water molecules into the subsurface layer, thus the alkali ions may be mobilized again with continued degradation of the glass [82]. This effect is more pronounced for larger particle sizes [78,83]. On the basis of these results, further in vitro and in vivo investigations are required to assess various composition–particle sizes and their corresponding ion release profiles (for all the alkali ions) to assess their optimum therapeutic efficacy. The Zn2+ release concentrations for all three glasses varied from 5.97 to 38.91 ppm under normal physiological conditions and from 1229 to 4904 ppm under extreme physiological conditions (see Fig. 3(a) and (b)). Zn2+ levels previously reported to have clinically beneficial effects on nerve growth were reported in the range of 0.065 ppm to 3.27 ppm in toad olfactory epithelia, with higher levels appearing to inhibit neural firing rates [84]. Similarly, early

X.F. Zhang et al. / Materials Science and Engineering C 31 (2011) 669–676

research has also demonstrated that Zn2+ may be required for neuroprotection following sub-lethal ischemia in both in vivo and in vitro models [85]. More recently, it has been suggested that Zn2+ may also be valuable in specific therapies, such as uremic polyneuropathy, with the preferred concentration of Zn2+ release in this regard between 12 and 16 uM (0.784–1.046 ppm) and the optimal Zn2+ release for muscle cells of about 40–50 uM (2.615– 3.27 ppm) [86]. Moreover, Zn2+ has been suggested to act as a type of neurotransmitter. The concentration of Zn2+ released during transmission is found to be approximately 300 uM (19.614 ppm) which is more than sufficient to be neurotoxic in neuronal cell culture [87]. In addition, the antibacterial effects of Zn2+ have been evaluated by broth dilution methods [88], in which bacterial growth was inhibited in the most concentrated Zn2+ oxide (0.719 ppm and 1.046 ppm) suspension under the same conditions [89]. For these applications, the appropriate Zn2+ concentration was found to range between 0.065 ppm and 19.614 ppm. These rapid therapeutic effects for zinc supplementation on neuronal cell may involve some aspects of peripheral nerve regeneration (PNR) function. Moreover, there may be greater benefit of zinc supplementation for infections in implantation sites specific to nerve (motor and sensory) recovery. Unfortunately, there is limited information concerning the effects of Zn2+ release in clinical PNR. While the Zn2+ release from the glasses in this study (CNG1–CNG3) was observed to cover a wide range of Zn2+ ion release under normal physiological conditions, the highest Zn2+ release value seen was approximately double that reported as an appropriate range in literature. Adversely, Zn2+ release under extreme physiological conditions exceeded the reported optimal ranges, which may produce undesirable cytotoxic effects [90]. It is well published that the structural role of zinc in sodium zinc silicate glasses is dependent on the initial concentration of the sodium present [91]. This phenomena are based upon the assumption that the greater the sodium concentration present the more likely that zinc will act as a network former [92]. This correlates well with the results obtained comparing CNG1 (0.04 sodium mole fraction) to CNG2 (0.09 sodium mole fraction), whereby the reduced sodium content results in the zinc acting in the role of network modifier. This suggests that the final Zn2+ release concentrations of the bioactive glasses produced herein, may be controlled via alteration of the composition, thus facilitating the degradation rate and hence ion release concentration of the proposed nerve guidance conduits to promote peripheral nerve regeneration post trauma. The Ca2+ release concentrations for all three glasses ranged from 19.16 ppm up to 143.07 ppm under normal physiological conditions and from 857 up to 3130 ppm under extreme physiological condition (see Fig. 4(a) and (b)). A central role of Ca2+ is the regulation of the nerve growth cone motility [45]. It has been reported that there are grades affecting Ca2+ ion concentration on the growth cone behaviour, with maximal activity occurring at an optimal level within this permissive range. This expanded model resembles other graded calcium-dependent processes; whereby many of these processes also display a similar dependence on optimal calcium [93]. The intracellular Ca2+ levels may regulate outgrowth by affecting the stability of the peripheral cytoskeleton, as investigated by Janet (about 0.008 ppm) who claims that the extension of growth cones appears to be greatest at an optimal Ca2+ level [94]. The dorsal root ganglia (DRG) from white chick embryos have been used to evaluate the optimal Ca2+ level for neuritis outgrowth, (2.5 μM calcium ionophores A23187 (27.196 ppm)) as used in culture medium to examine optimal levels of Ca2+ [52]. In addition, the same group observed how the calcium ions influence the modification polymerization/depolymerization and enzymatic degradation of many different cytoskeletal components [52]. The role of Ca2+ in secretion has been studied by Care et al. [95] and the concentration required to depolarize chromaffin cells was found to range from 0.16⁎10−3 up to 0.21⁎10−3 ppm [96]. Hence, it can be

675

perceived that any glasses releasing Ca2+ levels within these ranges will have similar beneficial effects. The Ca2+ release from the three glasses synthesised under normal physiological conditions therefore may be deemed to suit the intended medical application (PNR). However, the Ca2+ release under extreme physiological conditions is much higher than the optimal ranges reported in literature. This sufficient amount of Ca2+ release provides suitable conditions for further development of the proposed PLGA/BG composites which may be further investigated for its therapeutic effect regarding the ions released from the glass housed in its polymeric structure in final composite form. In effect, this will not limit the function of the glasses designed herein, since the ion release concentration may be tailored by the pore size and the Na:Ce ratio of the glass synthesised prior to fabrication of the final composite material. In addition, the degradation of PLGA/BG composite is controlled by the weight ratio of PLGA, the poloxF127 and BG, once the glass is combined with PLGA. This suggests further investigation of the therapeutic ion delivery and biocompatibility is necessary for the PLGA/F127 and glass composite material. 5. Conclusion This research hypothesizes the use of unique glass compositions for use as a component of composite nerve guidance conduits. This initial study has shown that the structure of the glasses indicates a predominate Q2 speciation (29Si chemical shift of −84 to −80 ppm), which may be controlled by modification of composition with respect to controlling the Na2O:CeO2. The results indicate heterogeneous dissolution of the glasses upon exposure to physiological pH. The ion release data demonstrates that ion release in the range (19.26– 3130 ppm) for Ca2+ and in the range (5.97–4904 ppm) for Zn2+ occurred. Release of such elements, at appropriate levels, from peripheral nerve guidance conduits may be advantageous with respect to the repair of peripheral nerve discontinuities. On the basis of these results, these glasses will be incorporated into degradable polymeric matrices for further in vitro and in vivo investigation on biocompatibility and mechanical properties in order to assess the composition and corresponding behavior for an optimum peripheral NGCs. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22]

S. Ichihara, Y. Inada, T. Nakamura, Injury 39 (2008) 29. W. Colin, R.B. Donoff, Journal of Dental Research 63 (1984) 987. W. Friess, Advanced Drug Delivery Reviews 55 (2003) 1529. S. Yoshii, M. Oka, Journal of Biomedical Materials Research 56 (2001) 400. C.H. Lee, A. Singla, Y. Lee, International Journal of Pharmaceutics 221 (2001) 1. Z. Ruszczak, W. Friess, Advanced Drug Delivery Reviews 55 (2003) 1679. T. Waitayawinyu, D.M. Parisi, B. Miller, S. Luria, H.J. Morton, S.H. Chin, T.E. Trumble, The Journal of Hand Surgery 32 (2007) 1521. G. Keilhoff, F. Stang, G. Wolf, H. Fansa, Biomaterials 24 (2003) 2779. H.S. Dresner, T.A. King, H.B. Clark, S.K. Juhn, S.C. Levine, Annals of Otology Rhinology and Laryngology 115 (2006) 631. L. Chung, D. Dinakarpandian, N. Yoshida, J.L. Lauer-Fields, G.B. Fields, R. Visse, H. Nagase, The EMBO Journal 23 (2004) 3020. S.J. Archibald, J. Shefner, C. Krarup, R.D. Madison, Journal of Neuroscience 15 (1995) 4109. D. Ceballos, X. Navarro, N. Dubey, G. Wendelschafer-Crabb, W.R. Kennedy, R.T. Tranquillo, Experimental Neurology 158 (1999) 290. C.C.C., Chein-Chi Lin, Chia-Chi-Ho, Biomaterials 26 (2005). W.F.A.D.D., Marcel F. Meek, Jeff M. Schakenraad, Peter H. Robinson, Microsurgery 17 (1996) 555. B.L. Seal, T.C. Otero, A. Panitch, Materials Science and Engineering: R: Reports 34 (2001) 147. E.O. Johnson, P.N. Soucacos, Injury 39 (2008) 30. J.S. Belkas, M.S. Shoichet, R. Midha, Operative Techniques in Orthopaedics 14 (2004) 190. W.M. Saltzman, in, Oxford, London, 2004, pp. 538. L. Singh, V. Kumar, B.D. Ratner, Biomaterials 25 (2004) 2611. L. Corvetti, F. Rossi, J Neurosci 25 (2005) 7150. A. Ferndez-Carballido, P. Pastoriza, E. Barcia, C. Montejo, S. Negro, International Journal of Pharmaceutics 352 (2008) 50. E.E.S., E.J. Frazza, Journal of Biomedical Materials Research 5 (1971) 43.

676

X.F. Zhang et al. / Materials Science and Engineering C 31 (2011) 669–676

[23] S.C. Baker, G. Rohman, J. Southgate, N.R. Cameron, Biomaterials 30 (2009) 1321. [24] A.J. Wang, Q. Ao, W.L. Cao, M.Z. Yu, Q. He, L.J. Kong, L. Zhang, Y.D. Gong, X.F. Zhang, Journal of Biomedical Materials Research Part A 79A (2006) 36. [25] W.d.D., M.F. Meek, 2.J.M. Schakenraad, 3.P.H. Robinson, A Thin-walled Biodegradable p(DLLA-ε-CL) Nerve Guide, Groningen, 1997. [26] J.C. Knowles, K. Franks, I. Abrahams, Biomaterials 22 (2001) 3091. [27] M. Therin, P. Christel, S. Li, H. Garreau, M. Vert, Biomaterials 13 (1992) 594. [28] H.H. Lu, S.F. El-Amin, K.D. Scott, C.T. Laurencin, Journal of Biomedical Materials Research 64A (2003) 465. [29] Q.-Q. Qiu, P. Ducheyne, P.S. Ayyaswamy, Journal of Biomedical Materials Research 52 (2000) 66. [30] M. Marcolongo, P. Ducheyne, J. Garino, E. Schepers, Journal of Biomedical Materials Research 39 (1998) 161. [31] J.A. Roether, A.R. Boccaccini, L.L. Hench, V. Maquet, S. Gautier, R. J e, Biomaterials 23 (2002) 3871. [32] K. Zhang, Y. Wang, M.A. Hillmyer, L.F. Francis, Biomaterials 25 (2004) 2489. [33] S. Verrier, J.J. Blaker, V. Maquet, L.L. Hench, A.R. Boccaccini, Biomaterials 25 (2004) 3013. [34] Orthomed, in: r.t.b.f.n. regeneration (Ed.) Product Web Site, 1980, pp. Collagen Tube. [35] M.G.D., André Oliveira Paivaa, Maria Helena Vaz Fernandesb, Maria Helena Gil, Necésio Gomes Costa, Material Research 9 (2006) 7. [36] S.G. Kazarian, K.L. Andrew Chan, V. Maquet, A.R. Boccaccini, Biomaterials 25 (2004) 3931. [37] A.R. Boccaccini, V. Maquet, Composites Science and Technology 63 (2003) 2417. [38] R.M. Day, A.R. Boccaccini, S. Shurey, J.A. Roether, A. Forbes, L.L. Hench, S.M. Gabe, Biomaterials 25 (2004) 5857. [39] G.W.a.P.X. Ma, Scaffolding for tissue engineering, 1 edCRC Press, 2005. [40] S. Ramakrishna, J. Mayer, E. Wintermantel, K.W. Leong, Composites Science and Technology 61 (2001) 1189. [41] J.E. Shelby, Introduction to glass science and technology, , 2005. [42] A. Balamurugan, G. Sockalingum, J. Michel, J. Faur, V. Banchet, L. Wortham, S. Bouthors, D. Laurent-Maquin, G. Balossier, Materials Letters 60 (2006) 3752. [43] A. Balamurugan, G. Balossier, D. Laurent-Maquin, S. Pina, A.H.S. Rebelo, J. Faure, J. M.F. Ferreira, Dental Materials 24 (2008) 1343. [44] S.R. Bolsover, Cell Calcium 37 (2005) 395. [45] S. Konur, A. Ghosh, Neuron 46 (2005) 401. [46] S. Bunting, L. Di Silvio, S. Deb, S. Hall, The Journal of Hand Surgery: British & European 30 (2005) 242. [47] G. Gallo, P.C. Letourneau, Current Biology 9 (1999) R490. [48] T.M. Gomez, N.C. Spitzer, Nature 399 (1999) 84. [49] P.C. Letourneau, D.M. Snow, T.M. Gomez, Progress in Brain Research 102 (1994) 35. [50] H.M. Liu, Journal of Peripheral Nerve System 1 (1996) 97. [51] M. Mattson, S. Kater, Journal of Neuroscience 7 (1987) 4034. [52] K. Lankford, P. Letourneau, The Journal of Cell Biology 109 (1989) 1229. [53] C.J. Frederickson, J.Y. Koh, A.I. Bush, Nature Reviews Neuroscience 6 (2005) 449. [54] F. Aedo, R. Delgado, D. Wolff, C. Vergara, Neurochemistry International 50 (2007) 591. [55] Webster's New World™ Medical Dictionary in, Wiley, 2008. [56] Y. Yin, Q. Cui, Y. Li, N. Irwin, D. Fischer, A.R. Harvey, L.I. Benowitz, Journal of Neuroscience 23 (2003) 2284. [57] Y.-H. Cho, S.-J. Lee, J.Y. Lee, S.W. Kim, C.B. Lee, W.Y. Lee, M.S. Yoon, International Journal of Antimicrobial Agents 19 (2002) 576. [58] A.B. Lansdown, U. Mirastschijski, N. Stubbs, E. Scanlon, M.S. Agren, Wound Repair Regen 15 (2007) 2. [59] D. Schubert, R. Dargusch, J. Raitano, S.-W. Chan, Biochemical and Biophysical Research Communications 342 (2006) 86.

[60] B. Standard, Part 14 : Identification and Quantification of Degradation Products from Ceramics (19903-14), 2009. [61] V. Dondur, D. Fidler, B. Adnadjević, V. Rakić, Journal of Thermal Analysis and Calorimetry 32 (1987) 613. [62] M.G. Cerruti, D. Greenspan, K. Powers, Biomaterials 26 (2005) 4903. [63] A. Stamboulis, R.V. Law, R.G. Hill, Biomaterials 25 (2004) 3907. [64] M.R.T.D. Boyd, R.V. Law, R.G. Hill, Journal of Materials Science: Materials in Medicine 17 (2006) 6. [65] A. Stamboulis, S. Matsuya, R.G. Hill, R.V. Law, K. Udoh, M. Nakagawa, Y. Matsuya, Journal of Dentistry 34 (2006) 574. [66] H. Aguiar, E.L. Solla, J. Serra, P. Gonz ez, B. Le , F. Malz, C. J er, Journal of NonCrystalline Solids 354 (2008) 5004. [67] S. Matsuya, A. Stamboulis, R.G. Hill, R.V. Law, Journal of Non-Crystalline Solids 353 (2007) 237. [68] J.M. Oliveira, R.N. Correia, M.H. Fernandes, J. Rocha, Journal of Non-Crystalline Solids 265 (2000) 221. [69] V.K. Shrikhande, V. Sudarsan, G.P. Kothiyal, S.K. Kulshreshtha, Journal of NonCrystalline Solids 283 (2001) 18. [70] F. Fayon, C. Bessada, D. Massiot, I. Farnan, J.P. Coutures, Journal of Non-Crystalline Solids 232–234 (1998) 403. [71] M.W.G. Lockyer, D. Holland, R. Dupree, Journal of Non-Crystalline Solids 188 (1995) 207. [72] D. Boyd, S. Murphy, M.R. Towler, A.W. Wren, S. Hayakawa, Journal of NonCrystalline Solids 355 (2009) 2285. [73] J.J. Mazer, J.V. Walther, Journal of Non-Crystalline Solids 170 (1994) 32. [74] C.R.D., R.W. Das, Physics and Chemistry of Glasses, October 1967. [75] L.M.E., Julian R. Jones, Larry L. Hench, Journal of Materials Science: Materials in Medicine 17 (2006) 8. [76] E. Mansour, K. El-Egili, G. El-Damrawi, Physica B: Condensed Matter 392 (2007) 221. [77] J.L. Rygel, C.G. Pantano, Journal of Non-Crystalline Solids 355 (2009) 2622. [78] C. Pritsos, E. Kontonasaki, X. Chatzistavrou, L. Papadopoulou, F. Pappas, P. Koidis, K.M. Paraskevopoulos, Journal of the European Ceramic Society 25 (2005) 891. [79] K. Lin, J. Chang, Z. Liu, Y. Zeng, R. Shen, Journal of the European Ceramic Society 29 (2009) 2937. [80] I. Ahmed, A.J. Parsons, G. Palmer, J.C. Knowles, G.S. Walker, C.D. Rudd, Acta Biomaterialia 4 (2008) 1307. [81] M. Cerruti, D. Greenspan, K. Powers, Biomaterials 26 (2005) 1665. [82] B. Grambow, in: D.E.Z., B.K. Clerk (Eds.), Corrosion of Glass, Ceramics and Ceramics Superconductor, Principles, Testing, Characterisation and Application, 1992. [83] L.L.H., R.S. Pryce, Key Engineering Materials 240 (2003) 5. [84] H.H. Sandstead, The Journal of Trace Elements in Experimental Medicine 16 (2003) 165. [85] E.J. Pawson, B. Duran-Jimenez, R. Surosky, H.E. Brooke, S.K. Spratt, D.R. Tomlinson, N.J. Gardiner, Diabetes, 59, (2010) 509–518. [86] R.H.N., J.E. McCusker, S.B. Broussard, K.W. Kelley, in, 2003. [87] A.I. Bush, Current Opinion in Chemical Biology 4 (2000) 184. [88] T.A.S.B.S.S.H.T.E.G.H.S. Sjoumlberg, Plastic & Esthetic Surgery 24 (1990) 5. [89] S.A.K.G.R., Tr. ÇIÇEK2, Journal of Medical Sciences 28 (1998) 3. [90] e.a., S. Murphy, Material in Medicine 20 (2009) 8. [91] T.I. Pesina, V.A. Zakreuskii, O.P. Pukh, Physics and Chemistry of Glasses (1980). [92] D. Boyd, Department of Materials Science & Technology, University of Limerick, Limerick, 2005. [93] T.B. Seo, I.S. Han, J.H. Yoon, I.C. Seol, Y.S. Kim, H.K. Jo, J.J. An, K.E. Hong, Y.B. Seo, D.H. Kim, S.K. Park, D.C. Yang, U. Namgung, Acta Pharmacol Sin 27 (2006) 50. [94] J. Holliday, N.C. Spitzer, Developmental Biology 141 (1990) 13. [95] A.D. Care, R.F.L. Bates, General and Comparative Endocrinology 3 (1972) 448. [96] E.N. Reinhold Penner, Journal of Experimental Biology 139 (1988) 16.