In vivo bone tissue response to a canasite glass-ceramic

Biomaterials 23 (2002) 2895–2900

In vivo bone tissue response to a canasite glass-ceramic V.M. da Rocha Barrosa, L.A. Salataa, C.E. Sverzuta, S.P. Xaviera, R. van Noortb, A. Johnsonb, P.V. Hattonb,* b

a Department of Oral Surgery, University of Sao Paulo, Ribeirao Preto, Brazil Centre for Biomaterials and Tissue Engineering, School of Clinical Dentistry, University of Sheffield, Sheffield S10 2TA, UK

Received 20 May 1999; accepted 28 November 2001

Abstract The aim of this study was to determine the biocompatibility and osteoconductive potential of a high-strength canasite glass ceramic. Glass-ceramic rods were produced using the lost-wax casting technique and implanted in the mid-shafts rabbit femurs. Implants were harvested at 4, 13 and 22 weeks and prepared for light and electron microscopy. Hydroxyapatite was used as a control material. Hydroxyapatite implants were surrounded by new mineralised bone tissue after 4 weeks of implantation. The amount of bone surrounding the implant increased slightly at 13 weeks. In contrast, canasite glass and glass ceramic implants were almost entirely surrounded by soft tissue during all the time periods. Close contact between bone and canasite glass-ceramic implant without the intervening fibrous tissue was observed in only a few regions. The canasite formulation evaluated was not osteoconductive and appeared to degrade in the biological environment. It was therefore concluded that the canasite formulation used was unsuitable for use as implant. Further work is required to improve the biocompatibility of these materials with bone tissue. It is possible that this could be achieved by reducing the solubility of the glass and glass ceramic. r 2002 Elsevier Science Ltd. All rights reserved. Keywords: Canasite glass ceramic; Bone-implant interface; Osteoconductivity; Biocompatibility

1. Introduction Biomaterials for bone-implant applications may be classified as being either ‘‘bioinert’’ or ‘‘bioactive’’ [1,2]. In the case of the former, the implant plays a passive role in the biological environment. The local biological reaction is often characterised by fibrous tissue encapsulation of the material. In bioactive or osteoconductive materials, new tissue bone is formed in direct or intimate contact with the implant surface. Hydroxyapatite (HAp) is an example of an osteoconductive bioceramic [3]. However, an inherent shortcoming of HAp is its brittle nature and poor mechanical strength. This limits the application of such materials to implant coatings or situations of low mechanical load [4]. Thus, there is a need for new biomaterials that combine the osteoconductive characteristics of bioactive ceramics with sufficient strength and toughness for load-bearing applications [5].

*Corresponding author. E-mail address: p.v.hatton@sheffield.ac.uk (P.V. Hatton).

The glass-ceramic route has the potential for development of high strength, osteoconductive bioceramics for mineralised tissue repair and augmentation [6–8]. Glass ceramics are produced by the nucleation and controlled crystallisation of glasses [9]. It is the fine dispersion of the crystalline phase that results in improved strength and toughness of such materials. It is therefore not surprising that several bioceramics based on glass-ceramic technology have been produced to date. Kokubo et al. have developed an apatite wollastonite (A-W) glass ceramic that showed bone bonding following implantation, as well as having adequate mechanical strength [6,10]. Ceravitals is another osteoconductive glass ceramic, which contains only apatite as the crystalline phase [11]. However, these and related glass-ceramic materials are unsuitable for casting to shape and are expensive to machine. Their use has therefore been limited to date. Ideally, complex or custom shapes could be produced more easily by using the lost-wax casting technique. Chain silicates consisting of canasite, fluororichterite and enstatite crystalline phases in a glassy matrix that may be processed by this route have been developed

0142-9612/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 1 4 2 - 9 6 1 2 ( 0 1 ) 0 0 4 1 7 - 3

2896

V.M. da Rocha Barros et al. / Biomaterials 23 (2002) 2895–2900

[12]. The canasite glass ceramic has been shown to have a high strength and fracture toughness due to a crystalline microstructure of interpenetrating blades [13]. However, no in vivo work has yet been carried out on this material to establish its biocompatibility or osteoconductive potential for medical and dental applications. The existing bioceramics are either tough or osteoconductive, but little success has been had in producing materials with both properties. Thus, the aim of this study was to determine the osteoconductive potential of a high-strength canasite glass ceramic.

2. Materials and methods Materials were prepared as described in previous studies [14,15]. To summarise, a canasite glass (0.47K2O 0.94Na2O 1.42CaO 5.67SiO2 1.5CaF2) was produced using a conventional melt route. Implants (2 mm  4 mm rods) were produced using the lost-wax casting technique, melting the glass at 12001C and casting into a refractory mould (Whip-mix gypsum-bonded investment, Whip-mix Corp., Louisville NY, USA) at 5901C [14,15]. After devesting and grit blasting the glass surface with 50 mm alumina, half of the rods were devitrified for 1 h at 5201C and 1 h at 8601C [15]. Dense HAp were used as a control material. Samples were cleaned with 10% Decon in an ultrasonic bath at 601C and rinsed three times in distilled water. The samples were then washed with Analar grade acetone (BDH Laboratory Supplies, Poole, UK) in an ultrasonic bath for 5 min and finally rinsed with sterile distilled water at 601C (15 min). Surface roughness was determined using a Surftest 301 (Mitutoyo Corp., Kawasaki, Japan) and solubility studies were performed according to ISO6872: 1995). Samples were stored in glass vials and air-dried at 371C for 3 h before hot-air sterilisation for 4 h at 1601C. 2.1. Animal implantation procedures and sample harvesting A total of 27 3-month old rabbits weighing between 2.0 and 2.5 kg were used in the study with two implants placed bilaterally in the femur of each animal. In cases where only cast and cerammed glass were compared, the implants were inserted in the right femur. 2.2. Surgical procedures The animals were anaesthetised using 40 mg/kg of sodium pentobarbital (Sigmat, USA) intravenously. Using an aseptic technique, the right or left femur was exposed and two holes of E2.2 mm in diameter were drilled through the cortical bone using a surgical burr (Branemark implant Systemt, SDIB 091, Sweden)

under controlled rotation and profused irrigation with sterile saline. HAp, cast glass or cerammed glass were then inserted into the implant sites and the reflected tissues were closed in layers using Vicrylt sutures 4-0 (Ethicon Ltd., Brazil). The skin was closed with interrupted sutures with Mononylont 3-0 (Ethicon Ltd., Brazil) and the wounds were moistened with iodine solution (Riodeinet, Rioquimica Ltd., Brazil). Post-operatively, wounds were inspected daily for clinical signs of complications or adverse reactions and to monitor healing. The rabbits were maintained in cages with free access to food and water. Four weeks after surgery, 10 animals were sacrificed by intraperitoneal administration of an overdose of thiopental. Eleven animals were sacrificed 13 weeks after surgery and the remaining three after 22 weeks. 2.3. Preparation for microscopy The bone and implants were immediately fixed in 3% glutaraldehyde (0.1 m sodium cacodylate buffer) for 3 h. These specimens were then dehydrated through a series of graded ethanols and embedded in resin (LR White Hard Grade, London, UK) over a period of up to 28 days. Each specimen was placed in a polyethylene cup and filled with fresh resin which was polymerised in an oven at 601C for 24 h. The block was cut using a band saw until the implant was exposed. X-ray radiographs were taken of all femurs to locate the implants to help with gross reduction of the blocks. The surface with the capsuled implant was polished to a 0.5 mm finish using silicon carbide paper and diamond polishing pastes. The polished block surfaces were stained and processed as described by Maniatopoulos et al. with Stevenel’s blue and Alizarin red [16]. The stained blocks were mounted on slides using glass bond (Loctite UK Ltd.) and were cut using a diamond band saw (Ekakt Apparatebau, Nordersted, Germany) to leave a section E1 mm thick. The cut surfaces of the sections were then polished and thinned down to 30–40 mm using silicon carbide paper without disruption of the implant tissue interface. Cover slips were used to protect the sections and the slides were examined under a standard light microscope. 2.4. Scanning electron microscopy Backscattered electron imaging (BSI) in association with energy dispersive X-ray microanalysis (EDS) was performed on unstained resin blocks that had been sectioned and polished to reveal the specimen using a Philips 501b SEM (Eindhoven, The Netherlands) with an accelerating voltage of 30 kV. The blocks were cleaned with 70% ethanol, placed on specimen holders and coated with a layer of carbon. Micrographs of representative areas were taken.

V.M. da Rocha Barros et al. / Biomaterials 23 (2002) 2895–2900

2897

3. Results 3.1. Surface roughness and solubility The canasite glass rods had a roughness (Ra) of 2.17 mm (standard deviation 70.34), and the devitrified material had an Ra of 2.46 mm (70.77). Surface roughness determination showed no significant difference between the canasite in the glassy and crystallised states. Solubility of the crystallised material was determined as 1244 mg cm 2 (7397). 3.2. Clinical observations Three animals died from causes unrelated to the surgery or implants during the 13 weeks follow-up period. No complications were noted with the remaining animals. Sites of implantation appeared to have healed with no visible signs of inflammation or adverse tissue reaction. The radiographs confirmed that the implants had been placed successfully in the burr hole penetrating the cortex into the marrow of the femur. 3.3. Histological evaluation Control hydroxyapatite implants were surrounded by a new mineralised bone tissue after 4 weeks of implantation (Fig. 1). The findings at 13 weeks were quite similar, although the amount of bone surrounding the implants had increased slightly. Scanning electron microscopy (SEM) confirmed that the bone-implant interface was intimate and not interspersed with the fibrous tissue (Fig. 2). In contrast, canasite glass and glass-ceramic implants were almost entirely surrounded by soft tissue at 4 and 13 weeks implantation. The soft tissue extended back into the compact bone of the cortex where the surgical

Fig. 1. Ground section of femur with hydroxyapatite implant (HA) at 4 weeks after implantation showing a new mineralised bone tissue (B) around the implant; field width=2 mm.

Fig. 2. Backscattered SEM photomicrographs of hydroxyapatite implant (HA) in midshaft of femur showing direct contact with new mineralised bone tissue (B) at low (a) and high (b) magnification: (a) field width=800 mm and (b) field width=200 mm.

site had otherwise healed. While the soft tissue around the canasite glass appeared stable, areas of inflammation or granulation tissue were observed in sites where the glass-ceramic surface appeared to be degrading (Fig. 3). This was more evident in the 22-week specimens where debris was observed in the fibrous tissue surrounding the implant (Fig. 4). The canasite glass-ceramic showed evidence of gradual surface dissolution. This was suggested by the presence of an electron-lucent layer at the surface of the implant when viewed using the backscattered electron detector in the SEM (Fig. 5). An intimate relationship between bone and canasite glass ceramic without soft-tissue intervention was observed in only a few regions (Fig. 6). However, even in these unusual sites, there did not appear to be direct contact. Overall, little difference in the bone response to canasite glass ceramic was observed at all time periods. Further evidence for the presence of glass debris was provided by SEM study. Fig. 7 shows irregular, electron dense particles up to 30 mm in diameter in the soft tissue

2898

V.M. da Rocha Barros et al. / Biomaterials 23 (2002) 2895–2900

Fig. 3. Ground section of femur with canasite glass-ceramic implant (C) at 4 weeks after implantation showing fibrous tissue and inflammation (arrowed) between the implant surface and local bone tissue (B); field width=1 mm.

Fig. 4. Ground section showing implant debris (arrowed) in a dense fibrous tissue between the implant (C) and marrow space (M) after 22 weeks implantation; field width=1 mm.

Fig. 6. High-power SEM photomicrograph showing area where bone tissue (B) is in close proximity to a defect in the canasite implant (C). Note that a narrow (o10 mm) electron lucent region still separates the two regions; field width=800 mm.

Fig. 7. SEM photomicrograph showing glass-ceramic debris (arrowed) in the soft tissue between the canasite glass-ceramic implant (C) and lamellar bone tissue (B); field width=400 mm.

between the bone and implant surfaces. EDS confirmed that the composition of the debris was similar to that of the glass-ceramic implant.

4. Discussion

Fig. 5. Low-power SEM photomicrograph taken using the backscattered electron detector showing lack of contact between mineralised bone tissue (B) and canasite implant (C). Identity of fibrous tissue (F) was confirmed by light microscopy in Figs. 3 and 4. Note the evidence of surface dissolution; field width=1500 mm.

The observations of HAp implants confirmed its status as a biocompatible, osteoconductive implant material and provided evidence for the suitability of the rabbit femur model to evaluate biocompatibility and osteoconductivity. The results also showed that this canasite composition was not osteoconductive in this animal model. The irregular appearance of the canasite implant surface in all micrographs suggested that the composition tested was unstable in the biological environment due to surface dissolution. The gradual

V.M. da Rocha Barros et al. / Biomaterials 23 (2002) 2895–2900

and sustained surface dissolution of the canasite implants, as well as the canasite particles released at the implant surface, appeared to have caused fibrosis and the recruitment of inflammatory cells. Numerous canasite lathes are visible at the glass-ceramic/tissue interface, and it is highly likely that these too are contributions to the poor host response. The material therefore actively interfered with local bone healing and osteogenesis via promotion of inflammation. Inflammation was most likely further stimulated by the presence of particulate debris released from the surface of the material. The tendency of this glassceramic material to degrade in vivo may have been in turn actively encouraged by local inflammation (Fig. 3). Macrophage activity in the fibrous tissue adjacent to the implant usually results in the implant being exposed to body fluid with an acidic pH that may have increased the rate of dissolution [17]. There is published evidence that CaF2 present in the canasite glass ceramic could have interfered with bone formation. The addition of CaF2 in place of CaO in a bioactive glass resulted in a composition with slower reactivity and reduced bonebonding ability [18]. Schepers et al. (1993) also demonstrated that the partial substitution of SiO2, CaO and Na2O by CaF2 could retard the bone formation at the surface of the glass implants [19]. The addition of F2 to silica results in a gradual breakdown of the silicon network by increasing the number of non-bridging oxygens [2]. This affects glass structure and could therefore inhibit the bioactivity of these glass ceramics. It is likely that the residual glass phase exerts a marked influence on a number of important properties of glass ceramics including their chemical stability and durability, which are in close connection with bioactivity [2]. Equally important may have been the contribution of surface nucleation and crystallisation to the rapid dissolution in vivo and release of irritating particulate debris. SEM suggested that surface degradation was accompanied by infiltration into defects that are perpendicular to the implant surface (Figs. 5 and 6). Surface nucleation is most likely to have occurred due to interaction of the glass and investment material during casting. The interactions of bioactive materials in a physiological environment comprise biodissolution, apatite deposition and bone formation at the implant surface [20]. This sequence did not occur with canasite and this may partly explain the poor results reported here. The essential condition for glass and glass ceramic to bond with bone is the formation of a silica-rich layer and apatite layer in vivo [21], but it is not essential to contain apatite within the material [22]. While the absence of phosphate from the canasite composition was not therefore responsible for poor biocompatibility, the composition was important as a determinant of residual glass solubility and hence stability in the biological

2899

environment. In this study, it was likely that the poor stability in the biological environment was a direct result of the low cross-link density of the residual glass phase in the canasite glass ceramic. Further work is therefore required to reduce both the solubility of these materials and their interaction with the investment materials before they can find applications as bioceramics for bone tissue repair. Substitution of sodium with calcium, or incorporation of phosphate, as suggested by Miller et al., might be suitable routes to produce a more biocompatible canasite [23].

5. Conclusions The presence of new lamellar bone around the hydroxyapatite implants confirmed the suitability of the rabbit femur for the in vivo evaluation of osteoconduction. The canasite formulation evaluated was not osteoconductive and appeared to degrade in the biological environment. The canasite formulation tested was therefore considered unsuitable for implant applications. Further work is required to improve the biocompatibility of these materials with bone tissue. It is possible that this could be achieved by reducing the solubility of the residual glass in the glass ceramic.

Acknowledgements This work was supported in part by CNPq (Grant No. 200794/97-3) and FAPESP (Grant No. 96/04 106-0), Brasil.

References [1] Hench LL. Bioactive ceramics. In: Ducheyene, Lemons J, editors. Bioceramic material characterisation versus in vivo behaviour, vol. 523. New York: Annals of New York Academy of Sciences, 1986. [2] Strnad Z. Role of glass phase in bioactive glass-ceramics. Biomaterials 1992;13:317–21. [3] Tracy BM, Doremus RH. Direct electron microscopy studies of the bone–hydroxyapatite interface. J Biomed Mater Res 1984;18:719–26. [4] Yoshimari M, Ohtsuka Y, Derand T. Thin hydroxyapatite coating produced by the ion beam dynamic mixing method. Biomaterials 1994;15:529–35. [5] Bonfield W. New trends in implant materials. In: Pizzoferrato L, editor. Biomaterials and clinical applications. Amsterdam: Elsevier, 1987. [6] Kokubo T, Ito S, Sakka S, Yamamuro T. Formation of a high strength bioactive glass-ceramic in the system MgO–CaO–SiO2– P2O5. J Mater Sci 1986;21:536–40. [7] Vogel W, Holland W, Naumman K, Gummel J. Development of machinable bioactive glass-ceramics for medical use. J Non Cryst Solids 1986;80:34–51.

2900

V.M. da Rocha Barros et al. / Biomaterials 23 (2002) 2895–2900

[8] Wood D, Hill R. A glass-ceramic approach to controlling the properties of a glass-ionomer bone cement. Biomaterials 1991;12:164–70. [9] Strnad Z. Glass-ceramic materials. Amsterdam: Elsevier, 1986. [10] Kokubo T, Shigematsu Y, Nagoshima M, Tashiro M, Nakamura T, Yamamuro T, Higashi S. Apatite and wollastonite containing glass-ceramics for prosthetic application. Bull Inst Chem Res, Kyoto Univ 1982;60:260–8. [11] Gross U, Strunz V. The interface of various glasses and glassceramics with a bony implantation bed. J Biomed Mater Res 1985;19:251–71. [12] Beal G. Chain-silicate glass ceramics. J Non Cryst Solids 1991;129:163–73. [13] Johnson A, Shareef MY, van Noort, Walsh JM. The effect of furnace type and ceramming conditions on the biaxial flexural strength of a canasite glass-ceramic. Dent Mater 2000;14:280–4. [14] Beale GH. Design and properties of glass-ceramics. Annu Rev Mater Sci 1992;22:91–119. [15] Johnson A, Shareef MY, Walsh JM, Hatton PV, van Noort R, Hill RG. The effect of casting conditions on the biaxial flexural strength of glass-ceramic materials. Dent Mater 1998;14:412–6. [16] Maniatopoulos C, Deporter DA, Melcher AH. An improved method of preparing histological sections of metallic implants. Int J Oral Maxillofac Implants 1986;1:31–7.

[17] Hench LL, Paschall HA. Direct chemical bond bioactive glassceramic materials to bone and muscles. J Biomed Mater Res Symposium 1973;4:25–42. [18] Clark AE, Hench LL, Paschall HA. The influence of surface chemistry on implant interface histology: a theoretical basis for implant materials selection. J Biomed Mater Res 1976;10: 161–74. [19] Schepers E, De Clercq M, Ducheyne P. Interfacial tissue response to bioactive glasses with reduced surface reactivity. J Mater Sci: Mater Med 1993;4:300–4. [20] Daculsi G, Le Geros RZ, Deudon C. Scanning and transmission electron microscopy and electron probe analysis of the interface between implants and host bone. Osseo-coalescence versus osseointegration. Scanning Microsc 1990;4:309–14. [21] Hench LL.The compositional dependence of bioactive glass, glass-ceramics. In: Seventh CIMTEC World Ceramic Congress. Italy: Montecatini Terme, 1990. [22] Kokubo T, It S, Hvang ZT, Hayashi T, Sakka S, Kitsugi T, Yamamuro T. Ca, P-rich layer formed on high strength bioactive glass-ceramic A-W. J Biomed Mater Res 1990;24: 331–43. [23] Miller CA, Kobuko T, Reaney IM, Hatton PV, James PF. Formation of apatite layers on modified glass-ceramics in simulated body fluid. J Biomed Mater Res 2002;59:473–80.