Repairing a critical-sized bone defect with highly porous modified and unmodified baghdadite scaffolds

Acta Biomaterialia 8 (2012) 4162–4172

Contents lists available at SciVerse ScienceDirect

Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat

Repairing a critical-sized bone defect with highly porous modified and unmodified baghdadite scaffolds S.I. Roohani-Esfahani a, C.R. Dunstan a, B. Davies a, S. Pearce b, R. Williams c, H. Zreiqat a,⇑ a

Biomaterials and Tissue Engineering Research Unit, School of AMME, The University of Sydney, Sydney 2006, Australia University of Ballarat B, Victoria 3350, Australia c Adelaide Microscopy, The University of Adelaide, South Australia 5005, Australia b

a r t i c l e

i n f o

Article history: Received 1 April 2012 Received in revised form 15 June 2012 Accepted 19 July 2012 Available online 27 July 2012 Keywords: Bone regeneration In vivo Critical-sized defect Scaffold

a b s t r a c t This is the first reported study to prepare highly porous baghdadite (Ca3ZrSi2O9) scaffolds with and without surface modification and investigate their ability to repair critical-sized bone defects in a rabbit radius under normal load. The modification was carried out to improve the mechanical properties of the baghdadite scaffolds (particularly to address their brittleness) by coating their surfaces with a thin layer (400 nm) of polycaprolactone (PCL)/bioactive glass nanoparticles (nBGs). The b-tricalcium phosphate/hydroxyapatite (TCP/HA) scaffolds with and without modification were used as the control groups. All of the tested scaffolds had an open and interconnected porous structure with a porosity of 85% and average pore size of 500 lm. The scaffolds (six per scaffold type and size of 4 mm  4 mm  15 mm) were implanted (press-fit) into the rabbit radial segmental defects for 12 weeks. Micro-computed tomography and histological evaluations were used to determine bone ingrowth, bone quality, and implant integration after 12 weeks of healing. Extensive new bone formation with complete bridging of the radial defect was evident with the baghdadite scaffolds (modified/unmodified) at the periphery and in close proximity to the ceramics within the pores, in contrast to TCP/HA scaffolds (modified/unmodified), where bone tended to grow between the ulna adjacent to the implant edge. Although the modification of the baghdadite scaffolds significantly improved their mechanical properties, it did not show any significant effect on in vivo bone formation. Our findings suggest that baghdadite scaffolds with and without modification can serve as a potential material to repair critical sized bone defects. Ó 2012 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction Repair and regeneration of critical size bone defects are challenging, and are hampered by frequent suboptimal outcomes, resulting in a significant increase in morbidity at a high economical cost to society [1]. The gold standard for bone defect repair – bone grafting with autologous bone – has significant drawbacks such as limited availability, second site surgery and donor site morbidity, leading to prolonged hospitalization [2]. Allografting also has several disadvantages which limit its use, including reduced bioactivity and increased risk of disease transmission. Scaffold-based solutions offer an alternative way for promoting the bone growth in large bone defects by using a porous material [2–4]. Significant efforts have been made to develop an ideal synthetic scaffold that reproduces bone’s structural properties combined with the necessary porosity, interconnectivity, bioactivity and mechanical strength [5,6]. While bioactive ceramics such as TCP (beta-tricalcium phosphate), HA (hydroxyapatite), TCP/HA (b-tricalcium ⇑ Corresponding author. Tel.: +61 2 93512392; fax: +61 2 93517060. E-mail address: [email protected] (H. Zreiqat).

phosphate/hydroxyapatite) and bioactive glasses bond with hard (and in some cases soft) tissues, they are brittle and are difficult to form into complex shapes in highly porous form (porosity >80%, pore size >300 lm and interconnectivity between pores 100) while providing an adequate mechanical stability for the defect site [7–13]. Calcium silicate ceramics have been proposed as potential biomaterials for bone tissue regeneration due to their good bioactivity and improved mechanical properties. However, a major drawback of the calcium silicate (CaSiO3) bioceramics is their chemical instability. Our strategy for developing new biomaterials for use as bone substitutes is to select CaSiO3 as the base material and to modify it through the incorporation of elements in order to enhance its physical and biological properties. Based on our previous investigations we identified zirconium as a candidate for inclusion with calcium silicate due to its ability to enhance bioactivity of calcium phosphate based materials [14]. We generated a calcium silicate ceramic containing zirconium and determined the optimal phase composition to be chemically identical to the previously identified mineral ‘‘baghdadite’’ [15] (Provisional Patent Application # 2007905843) and demonstrated its in vitro biocompatibility with

1742-7061/$ - see front matter Ó 2012 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.actbio.2012.07.036

4163

S.I. Roohani-Esfahani et al. / Acta Biomaterialia 8 (2012) 4162–4172

primary human osteoblasts, osteoclasts and with human endothelial cells [16,17]. In this study we prepared the baghdadite ceramic scaffold in a highly porous form with and without surface modification; the modified scaffolds more closely duplicated the mechanical strength and toughness of cancellous bone. For surface modification, a thin nanocomposite layer (polycaprolactone (PCL) as polymer matrix and bioactive glass nanoparticles (nBGs) as ceramic filler) was coated over the surface of scaffolds. PCL is a semi-crystalline and FDA-approved polymer that has a slower degradation rate and higher fracture energy than most biocompatible polymers. It shows 7% elongation and yield and 80% elongation at break point [18,19], which makes it the best candidate to address the brittleness of the ceramic scaffolds. The clinically relevant TCP/HA scaffolds (with and without modification) with similar pore structure to baghdadite scaffolds were used as the control materials. We applied the rabbit radial critical size defect model to determine the in vivo performance of these scaffolds. This model has been extensively used to validate the efficacy of scaffolds for bone regeneration using, for example, rhBMP-2-loaded collagen/chitosan, bone graft seeded with mesenchymal stem cells or with nHA/PCL/PLA scaffolds [20–23]. The aims of the present study were to investigate: (1) the in vivo performance and osteoconductive capacity of the modified and unmodified baghdadite scaffolds in the repair of large bone defects in the absence of added growth factors or stem cells; and (2) the effect of surface modification on new bone formation.

2. Materials and methods 2.1. Fabrication of baghdadite and TCP/HA scaffolds Baghdadite powders were synthesized by the sol-gel method using zirconia oxide nitrate (ZrO(NO3)2), calcium nitrate tetrahydrate (Ca(NO3)2
4H2O) and tetraethyl orthosilicate (TEOS, (C2H5O)4Si) as raw materials. Briefly, TEOS was mixed with ethanol and 2 MHNO3 (molar ratio: TEOS/ethanol/HNO3 = 1:8:0.16) and hydrolysed for 30 min under stirring. Then the ZrO(NO3)2 and Ca(NO3)2
4H2O were added into the mixture (molar ratio: ZrO(NO3)2/ Ca(NO3)2
4H2O/TEOS = 1:3:2) respectively, and reactants were stirred for 5 h at room temperature. After the reaction, the clear solution was maintained at 60 °C for 1 day and dried at 100 °C for 2 days to obtain the dry gel. The dry gel was calcined at 1150 °C for 3 h. Calcium phosphate deficient apatite powder was prepared by an aqueous precipitation reaction. Briefly, Ca(NO3)2
4H2O (0.92 M) and (NH4)2HPO4 (0.58 M) solutions were mixed gradually at room temperature and pH 11. The precipitated powder was thermally treated at 900 °C for 1 h, to produce TCP/HA powder composed of 40% HA and 60% b-TCP. After calcination, the powders were loosely agglomerated together with a size range of 1–5 mm. The powders were ground using a mortar/pestle and ball mill to an average size of 10 lm. The standard stainless steel sieve (75 lm) was used to make sure that all of the powders were smaller than 75 lm. Scanning electron microscopy (SEM) and image analysis method (Image J) were used to measure the particle size distribution. Based on our experimental findings, the optimum average size of the powder for making scaffold with this method is between 4 and 15 lm. Fully reticulated polyurethane foam was used as a sacrificial template for scaffold replication via the polymer sponge method. The optimum concentration range of PVA solution for preparation of the slurry is from 6 to 9 wt.%. In this study we used 6 wt.% PVA solution in water. The ceramic slurry was prepared by adding powders to polyvinyl alcohol (PVA) solution to prepare a 30 wt.% suspension. Foam templates were cut to appropriate dimensions and treated in NaOH solution for 30 min to improve surface hydrophilicity.

After cleaning and drying, foams were immersed in the slurry and compressed slightly to facilitate slurry penetration. Excessive slurry was squeezed out and the foam was subsequently blown with compressed air to ensure uniform ceramic coating on the foam surface. The weight of polyurethane foams increased approximately five times after coating with the slurry. After drying at 37 °C for 48 h, coated foams were fired in air in an electric furnace using a four-stage schedule: (i) heating from 25 to 600 °C at a heating rate of 1 °C min 1; (ii) further heating from 600 to 1200 °C at 2 °C min 1 for TCP/HA and from 600 to 1350 °C at 2 °C min 1 for baghdadite; (iii) holding the temperature at 1200 °C for 2 h for TCP/HA and at 1350 °C for 3 h for baghdadite; and (iv) cooling to 25 °C at a cooling rate of 5 °C min 1. 2.2. Modification of TCP/HA and baghdadite scaffolds To prepare the modified scaffolds, PCL pellets (–[(CH2)5COO]n–, Mw = 80,000, Sigma-Aldrich) were dissolved in chloroform (SigmaAldrich) at a concentration of 10% (w/v) under stirring. The spherical bioactive glass (58S) nanoparticles (nBGs) with an average size of 40 nm were prepared by the sol-gel method based on a previous report [24]. A preset amount of the nBGs (10 wt.%) was dispersed in 10 cm3 chloroform by sonication for 15 min and then added to PCL solution. The resultant mixture was stirred for 24 h at room temperature for increasing homogenization and then sonicated for 10 min just before coating the scaffolds. The surface of TCP/HA and baghdadite scaffolds were cleaned with ethanol and acetone, then dipped into the composite solutions for 1 min and afterwards excess solution was removed to form a uniform coating. The coated scaffolds were dried for 7 days in an oven at 37 °C and subsequently dried in a fume hood for another 24 h. Designation and composition of the prepared scaffolds for in vivo and in vitro examinations are shown at Table 1. The microstructure and fracture surface of the scaffolds were characterized by field emission scanning electron microscopy (FE-SEM) (Zeiss; Carl Zeiss, Germany) and all of the samples were coated with gold prior to analysis. Internal structure, porosity and interconnectivity of the scaffolds were evaluated by micro-computed tomography (l-CT) (Skyscan 1076, Micro-Computed Tomography). A spatial resolution of 18 lm was used for scanning. An optimized threshold was used to isolate the ceramic component from the background (void) for the evaluation of different morphological parameters. 30 image slices were sampled at regular intervals for porosity analysis. 2.3. Degradation in simulated body fluid In vitro degradation of the scaffolds was investigated by soaking the scaffolds in simulated body fluid (SBF). The SBF solution was prepared according to the procedure described by Kokubo and Takadama [25]. Cubic scaffolds (8 mm  8 mm  8 mm) were immersed in SBF solution at 37 °C for 1, 7, 14, 21 and 28 days at a solid/liquid ratio of 150 mg l 1. All scaffolds were held in plastic flasks and sealed. At each time point, the scaffolds were removed, rinsed with Milli-Q water and dried at 40 °C for 4 days, after which the final weight of each scaffold was measured. The concentration of the ions in the SBF after soaking was tested using inductive Table 1 Designation and composition of the prepared scaffolds for in vivo and in vitro examinations. Designation

Baghdadite Bag-PCL-nBG TCP/HA TCP/HA-PCL-nBG

Composition Scaffold substrate

Coating layer

Baghdadite Baghdadite B-TCP/HA B-TCP/HA

– PCL + nBG – PCL + nBG

4164

S.I. Roohani-Esfahani et al. / Acta Biomaterialia 8 (2012) 4162–4172

coupled plasma atomic emission spectroscopy (ICP-AES; Perkin Elmer, Optima 3000DV, USA). The weight loss (calculated according to the percentage of initial weight before soaking in SBF) and pH changes results were expressed as mean ± standard deviation (SD). 2.4. Mechanical properties of the scaffolds Mechanical properties of the scaffolds were determined in dry and wet conditions. For wet conditions, the scaffolds were first soaked in SBF for 24 h and extra liquid was removed carefully with filter paper just before testing. The compressive strength was determined by crushing cubic scaffolds (7 mm  7 mm  7 mm) between two flat plates using a computer-controlled universal testing machine (Instron 8874, UK) with a ramp rate of 0.5 mm min 1. Five identical specimens from each sample group were used for compressive testing in dry and wet conditions. 2.5. In vivo evaluation of the baghdadite, Bag-PCL-nBG, and TCP/HA and TCP/HA-PCL-nBG 2.5.1. Implanting the scaffolds in a radial segmental defect in rabbits In vivo osteoinductivity and/or osteoconductivity of the developed baghdadite and Bag-PCL-nBG were assessed in comparison to TCP/HA and TCP/HA-PCL-nBG by their implantation (press-fit) into segmental defects in the radius of rabbits. A total of 12 male rabbits (New Zealand White, 20 weeks of age) were used. The ulna provides significant mechanical support in this model, hence internal fixation is not required; however, partial mechanical loading is transmitted to the scaffold through weight bearing and muscle activity. Animal experiments were performed according to a protocol reviewed and approved by the Flinders University Animal Welfare Committee. This is a well described model for evaluating selected biomaterials to predict their performance in an orthopedic load-bearing environment [26]. This rabbit model is unilateral, and a consistent location for the ostectomy in each animal was achieved by ensuring that the ostectomy was positioned immediately proximal to the insertion of the fibrous insertion of the pronator teres. The rabbits received 0.4 mg kg 1 diazepam subcutaneously 15 min prior to mask induction of anesthesia with isoflurane 4% in 100% oxygen, delivered by a T-piece circuit with a fresh gas flow rate of 3.5 l min 1. Upon induction of stable anesthesia, 2.5% Isoflurane was delivered at the same fresh gas flow rate. The rabbit was placed in lateral recumbency with the operated limb dependent. The limb was prepared for aseptic surgery and draped to isolate the surgical site. A 2.5 cm surgical incision was made over the dorso-medial aspect of the forearm, and the underlying tissues separated to reveal the radial bone. Two parallel cuts were made 15 mm apart, and the excised radial bone was removed from the defect. The ulna was left intact. Animals were randomized to either implantation with the baghdadite, Bag-PCL-nBG, TCP/HA or TCP/HA-PCL-nBG. The implant (4 mm  4 mm  15 mm) was placed into the defect, and the wound closed in three layers using an absorbable suture material (3-0 vicryl). Six types of scaffolds (n = 6/scaffold type) were implanted: one implant per rabbit. Following placement of the implant, the wound was closed. The rabbits were recovered and maintained for 12 weeks with ample space for exercise, allowing normal weight bearing of the implanted limb. Perioperative antimicrobial therapy was provided in the form of Cephalosporin at 30 mg kg 1 by intramuscular injection and analgesia was provided by the administration of meloxicam, 0.2 mg kg 1 by intramuscular injection prior to surgery and then every 24 h for 3 days (four doses altogether). The post-operative explantation period for all groups was 12 weeks, when the animals were radiographed (dorso-palmar and latero-medial projections) at the time of sacrifice using a Kodak CR-500 digital acquisition system. Images obtained were saved as DICOM (Digital

Table 2 Scoring system for semiquantitative evaluation of radiographs. Criterion

Score

No obvious bone regeneration Less than 50% bone regeneration More than 50% bone regeneration Almost fused Fused – not full thickness Fused – full thickness

0 1 2 3 4 5

Imaging and Communications in Medicine) files. These radiographs provided qualitative and semiquantitative measurement of defect healing (new bone formation/defect fusion). 12 weeks after the surgery, the animals were sacrificed by an intravenous overdose injection of pentobarbital. At euthanasia, for each time point, the region of the radius containing the defect and the associated intact ulna were removed en bloc for peripheral quantitative l-CT evaluation and for histological evaluation. 2.5.2. New bone ingrowth and resorbability Bone growth was assessed radiographically, histologically and by l-CT. Radiographs were scored by two observers using a semiquantitative scoring system [27] outlined in Table 2. 2.5.3. Peripheral quantitative Micro-CT l-CT was employed to evaluate volumetric new bone formation before embedding in plastic for undecalcified histology for the qualitative and quantitative analysis of new bone formation in the defect. Each explanted sample was analysed at 19 lm resolution, and a volume of interest was defined with a length of 15 mm centered on the regenerating tissue. Each sample was wrapped in ‘‘cling wrap’’ to prevent drying and associated movement artefact, then placed horizontally in the Skyscan 1076 and orientated uniformly to achieve equivalent geometrics of all samples. The same settings were applied to every sample for consistency. The outline of the new tissue was traced manually to exclude the ulna, and thresholding was performed to differentiate mineralized bone from graft material and soft tissue. From this sub-volume of the dataset, a region of interest (selected region) was drawn freehand (using the Skyscan CT Analyser ROI tool), to exclude the ulna. The grey-level threshold value was determined manually after viewing a number of the sample datasets and comparing the pre- and post-threshold images. Once selected, the threshold level was applied to each sample in the study. The volumetric percentage of bone formation was measured over the defect zone. 2.5.4. Histological assessment and scaffold resorption 2.5.4.1. Embedding. Specimens were dehydrated through graded ethanol, cleared in xylene at room temperature and then infiltrated with, and embedded in, poly(methyl methacrylate) resin using a standard processing infiltration schedule as previously described [28]. Upon polymerization, multiple transverse sections through the defect region were prepared from the embedded resin block using a water-cooled slow-speed Beuhler isomet saw (Beuhler, Germany) and subsequently polished. Sections from the midpoint of the defect were used for histological analysis. 2.5.4.2. Histological staining. Sections were surface stained with toluidine blue and histomorphometry conducted using image analysis software (OsteoMeasure, Osteometrics, Georgia). Bone ingrowth was evaluated within the defect boundaries, by measuring the bone volume and ceramic volume. Bone and ceramic were individually identified and colour thresholded. The area of each component relative to the defined region of the defect was calculated using image J software (NIH).

S.I. Roohani-Esfahani et al. / Acta Biomaterialia 8 (2012) 4162–4172

2.5.4.3. Enzyme histochemistry. The extent of bone and ceramic surfaces lined with bone cells was identified in sections following toluidine blue staining or after enzyme histochemistry for tartrate-resistant acid phosphatase (TRAP) activity using naphtholAS-BI phosphate (Sigma, St. Louis, MO) as a substrate and fast red violet LB salt (Sigma) as a detection agent. [16]. Osteoblasts were identified on the basis of their cuboidal morphology in toluidine blue stained sections. Osteoclasts were identified based on positive staining for TRAP activity. Resorbability was assessed by comparing final ceramic volume for the different scaffolds and by the noting the presence of osteoclasts and resorption lacunae. 2.6. Statistical analysis For statistical analysis, t-tests and Mann–Whitney Tests (when data were not normally distributed) were performed to test for significance. All data are expressed as the mean and standard error unless otherwise stated. All statistical analysis was performed in Microsoft Excel and differences were considered significant if p < 0.05. 3. Results 3.1. Macro- and microstructure of the scaffolds SEM examination revealed that the baghdadite and Bag-PCLnBG scaffolds maintained a highly porous structure with 100% pore interconnectivity (Fig. 1a and d). The struts of the baghdadite scaffolds contained numerous cracks and defects (Fig. 1b); however, Bag-PCL-nBG scaffolds displayed a more uniform and continuous strut microstructure with no obvious cracks or pores (Fig. 1e). Sintered surfaces of baghdadite contained rectangular-prismshaped grains with an average size of 1.2 lm (Fig. 1c). Surfaces of Bag-PCL-nBG scaffolds were covered by the nanocomposite layer, consisting of a PCL layer as the matrix and nBGs as the dispersed particles (Fig. 1f). The nanoparticles were homogenously distributed in the polymer matrix with an average size of 40 nm. The baghdadite and Bag-PCL-nBG scaffolds had porosities of 88% and 85%, respectively, with an average pore size of 500 lm. Baghdadite scaffolds displayed an average reduction in porosity of 3% and a weight increase of 22% after coating with nanocomposite layer. The TCP/HA and TCP/HA-PCL-nBG showed similar structural morphology and physical properties to the baghdadite scaffolds. 3.2. Mechanical properties and fracture behaviour of the scaffolds Compression tests were performed on the TCP/HA, baghdadite, TCP/HA-PCL-nBG and Bag-PCL-nBG. In general, the compressive strength, elastic modulus and toughness of the scaffolds were strongly influenced by the addition of PCL and nBGs (Fig. 2a and b). Baghdadite scaffolds were able to withstand a maximum compressive stress of 0.27 MPa, compared to 0.12 MPa for TCP/HA scaffolds. Bag-PCL-nBG and TCP/HA-PCL-nBG scaffolds had a compressive strength of 1.1 and 1.2 MPa, respectively. A similar trend was observed for the modulus of elasticity of the modified scaffolds. As shown in Fig. 2a, the incorporation of the nBGs resulted in a marked increase in modulus of elasticity of Bag-PCLnBG (40.1 MPa) and TCP/HA-PCL-nBG scaffolds (37.1 MPa) compared to the unmodified baghdadite (15.3 MPa) and TCP/HA (10.5 MPa) scaffolds. As indicated in Fig. 2b, baghdadite scaffolds showed very brittle behaviour, with low failure strain of 0.5%, which was significantly increased to 7% for Bag-PCL-nBG scaffolds. A similar trend in the failure strain was found for baghdadite scaffolds coated only with PCL or with PCL-nBG; however, their

4165

compressive strength and modulus were compromised. As shown in Fig. 2b, the area under stress–strain curve of Bag-PCL-nBG scaffolds is dramatically greater than that for baghdadite scaffolds, which highlights the improvement in toughness of the modified scaffolds. Under wet conditions (1 and 28 days of soaking in SBF), the compressive strength of Bag-PCL-nBG scaffolds was significantly higher than that of baghdadite scaffolds. After 28 days of soaking, these values for nanocomposite coated scaffolds and for the uncoated baghdadite were 0.2 MPa and 0.03 MPa, respectively. Fig. 2c and d shows the fracture surface morphology of baghdadite and Bag-PCL-nBG scaffolds, respectively. A flat fracture surface formed for baghdadite scaffold at breaking point with complete disconnected struts, indicating the brittle nature of the scaffold (Fig. 2c). In contrast, the Bag-PCL-nBG scaffolds exhibited the formation of thin fibrils prior to the strut rupture in an attempt to hold back the broken parts together (Fig. 2b). At this point, severe plastic deformation in the nanocomposite layer was found. A similar trend of improvement was found for TCP/HA-PCL-nBG scaffolds, as we reported in a previous study [24]. 3.3. Degradation characteristics of the scaffolds in SBF Weight loss of all the scaffolds increased as a function of incubation time (Fig. 3a). After 28 days of soaking, the weight loss of the baghdadite and Bag-PCL-nBG scaffolds was 10% and 4%, respectively, relative to initial weight. For baghdadite and BagPCL-nBG scaffolds the pH of the solution increased continuously, from the initial value of 7.4 to 8. The Ca, Si, P and Zr ions release profile after soaking baghdadite and Bag-PCL-nBG scaffolds in SBF for 28 days are shown in Fig. 3c–f. The concentration of Zr and P ions in solution was low (<0.2 ppm) for baghdadite and Bag-PCL-nBG scaffolds, even after 28 days of soaking. For both scaffolds, Si and Ca ions were released relatively quickly within a short period, which slowed down as the soaking time increased (Fig. 3c and d). Such a phenomenon is similar to that seen for drug delivery systems, known as an initial burst in microspheres and films [29]. Fig. 4 shows SEM images of baghdadite and Bag-PCL-nBG strut’s scaffolds after soaking in SBF for 1 day, 14 and 28 days. After 1 day of soaking, the surface morphology of both scaffolds remained intact (Fig. 4a and b), and by 2 weeks of soaking, some scattered deposits (100 nm in size) were found to cover the grain surfaces of baghdadite scaffolds. The matrix of the Bag-PCL-nBG scaffolds appeared to be disrupted, with signs of degradation in the coated layer and the formation of apatite deposits underneath the coating. After 28 days, apatite granules were evenly covering the surface of both scaffolds. In contrast, the matrix of the BagPCL-nBG scaffolds was completely disrupted and the size and uniformity of the apatite deposits were similar to those found in the baghdadite scaffolds. The weight reduction value for TCP/HA and TCP/HA-PCL-nBG was 30% and 20%, respectively and the pH of the solution was 7.4 after 28 days of soaking for both groups of scaffolds. After 28 days, the surface of the TCP/HA scaffolds was covered by an apatite layer with a flower-like structure. For TCP/ HA-PCL-nBG scaffolds a large number of small apatite cluster was found on the surface and also with a similar degradation of its nanocomposite layer was similar to Bag-PCL-nBG scaffolds [24]. 3.4. In vivo osteoconductivity of the baghdadite and Bag-PCL-nBG, and TCP/HA and TCP/HA-PCL-nBG scaffolds in a radial critical size defect rabbit model Critical to the assessment of this material for implantation is a confirmation of in vivo biocompatibility and osteoconductivity. We assessed the ability of baghdadite and Bag-PCL-nBG scaffolds to promote bone regeneration in an orthopedically relevant model of a critical sized defect in the radius of rabbits, in the absence of

4166

S.I. Roohani-Esfahani et al. / Acta Biomaterialia 8 (2012) 4162–4172

Fig. 1. Macro and microstructure of the baghdadite (a–c) and Bag-PCL-nBG scaffolds (d–f). The highly porous structure of the scaffolds is shown, as well as the difference in the struts microstructure of modified and unmodified scaffolds, where the modified strut has a smooth surface with no presence of crack or pore.

Fig. 2. Fracture surfaces of (a) baghdadite and (b) Bag-PCL-nBG scaffolds. Arrows feature severe plastic deformation of the nanocomposite layer. (c) Compression strength and compressive modulus of coated and uncoated scaffolds. (d) Selected stress–strain curves of baghdadite, Bag-PCL-nBG and baghdadite scaffolds modified only with PCL.

S.I. Roohani-Esfahani et al. / Acta Biomaterialia 8 (2012) 4162–4172

4167

Fig. 3. (a) Weight loss of the scaffolds in SBF solution after different soaking times. (b) pH value of the SBF and concentrations of (c) Ca, (d) Si, (e) P and (f) Zr elements in the SBF solution after different soaking times.

added growth factors or cells. In this model, a critical sized 15 mm defect was produced in the radius and filled with either baghdadite, Bag-PCL-nBG or TCP/HA, or the modified TCP/HA-PCL-nBG scaffolds. Rabbits were ambulant immediately following recovery from anesthesia and were able to bear weight. While much of the weight bearing load in this model is supported by the ulna, load is transmitted though the implant in the radius during ambulation and from the activity of muscles. This model is considered by convention to be a critical sized defect, although there is some evidence in the literature that 15 mm defects may regenerate to a limited extent over a 12 week period [27].

defect (arrows) can be seen to be translucent, with TCP/HA and the TCP/HA-PCL-nBG indicating no bridging of the defect. In contrast, the bone defect for baghdadite and Bag-PCL-nBG is seen to contain new bone bridging the entire defect (Fig. 5b and c). Expert blinded radiographic assessment of the repair status of defects using a standard scoring system [27] (Table 2) demonstrated a statistically significant improved bridging of the defects with baghdadite and Bag-PCL-nBG compared to TCP/HA and TCP/ HA-PCL-nBG (Fig. 6a).

3.6. l-CT analysis 3.5. Radiographic assessment The radiographs obtained confirmed the quite consistent location of the defect created in the radius; also, the dimensions were approximately as defined. The implanted baghdadite and Bag-PCL-nBG scaffolds produced extensive new bone formation, with complete bridging of the radial defect at 12 weeks in all the samples tested (six per scaffold type), in contrast to TCP/HA and TCP/HA-PCL-nBG (Fig. 5a–d). In representative radiographs, the

l-CT analysis of the radial defects showed trends for improved bone formation with the baghdadite and the Bag-PCL-nBG scaffolds relative to the TCP/HA and the TCP/HA-PCL-nBG, although statistical significance was not achieved (Fig. 6b). Setting thresholds for discriminating TCP/HA and bone was challenging given the similar radio-opacity of these two materials (TCP/HA and bone). This difficulty was not apparent with baghdadite due to its higher radio-opacity.

4168

S.I. Roohani-Esfahani et al. / Acta Biomaterialia 8 (2012) 4162–4172

Fig. 4. Strut surface morphology of baghdadite (a–c) and Bag-PCL-nBG (d–f) scaffolds after soaking in SBF for different time periods.

3.7. Histological evaluation

Fig. 5. Radiographical images demonstrated that at 12 weeks there was superior bridging of the defect with baghdadite (c), Bag-PCL-nBG (d), compared to TCP/HA (a) and the TCP/HA PCL-nBG (b).

3.7.1. Histology Undecalcified histological assessment was conducted to assess the presence of bone bridging the radial defects containing either TCP/HA, baghdadite or their modifications with PCL and nBG. Transverse sections were taken within the defect spanning from the edge of the defect and through its midpoint. Toluidine blue staining revealed that all scaffolds were well tolerated with no signs of rejection, necrosis, or infection (Fig. 7a–h). When baghdadite and Bag-PCl-nBG scaffolds were implanted for 12 weeks, extensive new bone formation was evident both at the periphery and in close proximity to the ceramics within the pores of all the baghdadite (Fig. 7e) and Bag-PCL-nBG (Fig. 7g) scaffolds tested. Bone filled in the pores within the plane of the cortical bone of all these scaffolds. In contrast the response to TCP/HA (Fig. 7a) and TCP/HA-PCL-nBG (Fig. 7c) scaffolds, while also well tolerated, differed from baghdadite and Bag-PCL-nBG, where ceramics tended to collapse and bone tended to grow between the scaffold and ulna adjacent to the implant edge rather than through the ceramic pores. In a few sections from the baghdadite and BagPCL-nBG, open spaces within the newly formed bone were seen to be consistent with the possible recovery of radial architecture through the formation of an endocortical space (Fig. 7e and g). Extensive osteoblast lined osteoid seams were observed on many of the bone surfaces within the defects, indicating ongoing bone formation with all the materials studied (Fig. 7b, d, f and h). The presence of intracellular ceramic particles internalized in large multinucleated cells was observed for baghdadite (with and with-

S.I. Roohani-Esfahani et al. / Acta Biomaterialia 8 (2012) 4162–4172

4169

maintained after 12 weeks for all materials (with no significant difference between ceramics).

4. Discussion

Fig. 6. (a) Radiological scores demonstrating superior bridging of the defect for baghdadite and Bag-PCL-nBG based scaffolds compared to TCP/HA and TCP/HAPCL-nBG (ap < 0.05 TCP/HA vs. baghdadite; bp < 0.01 TCP/HA-PCL-nBG vs. Bag-PCLnBG), (b) l-CT analysis for measuring the volume of new bone.

out modification) but not TCP/HA (Fig. 7f and h). Enzyme histochemistry for TRAP activity was applied to the sections adjacent to those used for the toluidine blue staining as a histochemical marker for the presence of osteoclasts. Osteoclasts were seen lining both the bone and implant surfaces (Fig. 8a–d) for all materials, indicating that remodeling of the newly formed bone was occurring and that the ceramic materials are potentially being gradually removed by osteoclast actions (Fig. 8).

3.7.2. Histomorphometric assessment Quantitative analysis of the histological sections taken at the midpoint of the defect showed a strong trend for increased new bone formation in baghdadite (3.0 ± 3.1 mm2, mean ± SE) ceramics compared to TCP/HA (1.3 ± 1.0), and in Bag-PCL-nBG (3.0 ± 4.1) compared to TCP/HA-PCL-nBG (0.6 ± 0.2). Quantitative analysis of the histological sections taken at the midpoint of the defect showed a strong trend for increased new bone formation in baghdadite ceramics compared to TCP/HA, and a significant increase in the amount of new bone formed with Bag-PCL-nBG compared to TCP/HA-PCL-nBG, particularly at the defect midpoint. The ceramic area was also measured in the histological sections at the midpoint of the defect and this demonstrated that significant ceramic was

Structural properties for a three-dimensional bone scaffold, such as mechanical strength, pore size, porosity and interconnectivity between the pores, play a key role in bone regeneration both in vitro and in vivo [30–33]. However, mechanical properties of a scaffold significantly depend on the pore architecture of the scaffold as increasing porosity, pore size and interconnectivity between the pores result in the loss of the strength of construct. Ceramic scaffolds are an excellent choice for bone tissue engineering as they provide the appropriate chemical cues for the cells [34]. However, the major problem of highly porous ceramic scaffolds is their brittleness and inadequate mechanical properties, compared to native bone [35]. Coating the struts of the ceramic scaffolds with a thin bioactive and flexible layer is one approach to address these problems. We previously developed baghdadite ceramics by incorporating zirconium in the Ca–Si system and demonstrated their in vitro bioactivity using primary human osteoblasts, osteoclasts and microendothelial cells [17], suggesting the potential use of this ceramic in bone regeneration. In this study, we fabricated the baghdadite in porous form and further improved its mechanical properties by coating its surface with a thin nanocomposite layer consisting of nBGs and PCL matrix. In the present study we demonstrated that the developed nanocomposite coated scaffolds (Bag-PCL-nBG) are of high interconnectivity, porosity with large pore size, combined with a remarkable toughness and compressive strength. These properties are essential for enhancing bone in-growth in load-bearing applications. Modifying the baghdadite scaffolds with the nanocomposite layer produced a negligible decrease in porosity (3%), but resulted in a significant improvement in the mechanical properties; whereby the compressive strength, modulus and failure strain increased 9, 4 and 14 times respectively compared to TCP/HA scaffolds [36]. The struts of the baghdadite scaffolds had crack-like defects which were filled with the nanocomposite layer, possibly contributing to the improved mechanical properties of nanocomposite ceramic scaffolds. The fracture surface of Bag-PCL-nBG scaffolds showed the formation of fibrils with morphologies similar to those for collagen fibrils in the fracture surface of human bone [37]. The combination of PCL and nBG is inspired by the structure of bone, being composed of nearly 60 wt.% of an inorganic phase (the remaining being the organic phase (collagen) and water). The fracture behavior of mineralized tissues such as bone (and dentin) is influenced by the optimal interaction between the inorganic and organic phases, and the toughening mechanisms induced by the presence of collagen fibrils in bone [38]. Therefore, the addition of a nanocomposite phase to a porous ceramic scaffold is expected to enhance the toughness of the composite while the addition of bioactive nanoparticles would improve its bioactivity as well as increasing the mechanical properties of the matrix [39]. An overview of the literature on the development of tissue engineering scaffolds shows that the mean measured compressive strength value for the available scaffolds is indeed low (for load-bearing applications) and is close to the lower limit for spongy bone (0.2–4 MPa) [40,41]. In the present study, we demonstrated that by coating a nanocomposite layer consisting of nBG and PCL over the struts of baghdadite scaffolds, the compressive strength of the resultant nanocomposite scaffold increased to 1.1 MPa. Besides the uniform dispersion of the nanoparticles, the interfacial property between nanoparticles and the polymer is of serious concern to material scientists. Strong interactions would contribute to the enhanced mechanical behavior, whereas the poor

4170

S.I. Roohani-Esfahani et al. / Acta Biomaterialia 8 (2012) 4162–4172

Fig. 7. Toluidine blue staining of undecalcified plastic sections through the midpoint of the defect revealed that the baghdadite (e and f) and Bag-PCL-nBG (g and h) scaffolds were well tolerated, with no signs of rejection, necrosis, or infection. At 12 weeks post implantation, bone formation was evident in close proximity to all the ceramic scaffolds. We clearly see superior new bone formation with the baghdadite (e and f) and Bag-PCL-nBG (g and h) compared to the TCP/HA (a and b) and TCP/HA-PCL-nBG (c and d). In a few sections from the baghdadite and Bag-PCL-nBG, open endocortical spaces (g, arrows) were noticeable and this was not altered by the presence of the nBG. Baghdadite (with and without modification), but not TCP/HA, showed the presence of intracellular ceramic particles internalized in large multinucleated cells (Fig. 7f and h, arrows).

interaction leads to a decreased mechanical behavior compared to the neat matrix. The nBGs were evidently well-bonded to the polymer substrate, as no loose particles were observed in the scaffold surfaces before and after mechanical testing, contributing to the increase seen in the mechanical strength of the developed scaffolds. The total area under the stress–strain curve is equivalent to the work that must be done per unit volume on the specimen before it breaks and also is an indication for toughness and strain energy per unit volume of scaffold. As shown in Fig. 2b, the area under the stress–strain curve of Bag-PCL-nBG scaffolds is dramatically greater than that for baghdadite scaffolds. This indicated that the toughness of Bag-PCL-nBG scaffolds was markedly improved compared to others. The interesting fact is that there is not any

visible breaking point for these scaffolds as they start packing instead of showing a catastrophic failure. While nanocomposite coating slightly decreased the degradability of baghdadite scaffolds (6%, after 28 days of soaking the scaffolds in SBF), the differences in ion release patterns, pH changes of the solution and apatite formation for baghdadite and Bag-PCL-nBG scaffolds were not significant. We hypothesize that the presence of nBGs in the coating layer compensated for the barrier effect of the PCL layer limiting the release of Si, Ca and P ions from the baghdadite substrate. PCL is a biodegradable aliphatic (poly (a-hydroxyacid)) and its degradation is highly related to chemical hydrolysis. PCL degradation starts by the hydrolysis of ester linkages in its backbone. By adding nBG into PCL, the hydrophilicity of the PCL layer increases and the

S.I. Roohani-Esfahani et al. / Acta Biomaterialia 8 (2012) 4162–4172

4171

Fig. 8. Osteoclasts highlighted by TRAP staining (red arrows) were present on both bone and ceramic surfaces for all scaffolds: (a) TCP/HA, (b) TCP/HA-PCL-nBG, (c) baghdadite and (d) Bag-PCL-nBG.

SBF diffuses easier through the thin layer on the surface, forming microcracks on the surface of the coating, followed by increased degradation in the bulk of the material. Once SBF diffuses through the microcrack and the thin layer, nBGs are attacked and the surface eventually starts to decay and partially dissolve, while exhibiting apatite-forming ability as per the mechanisms discussed by Hench [42]. The small amount and size of the deposited apatites can hardly affect the degradation process. The radiological and histological analyses of the radial defects at the completion of the present study clearly confirm the biocompatibility and bioactivity of the baghdadite scaffolds (with and without modification) by quantifying their ability to induce new bone growth bridging a critical size bone defect. In vivo implantation of baghdadite scaffolds (with and without modification) in a radial critical size defect rabbit model for 12 weeks showed superior bridging of the defect compared to TCP/HA and the modified TCP/HA. Osteoinductivity is the ability to drive the formation of bone from mesenchymal precursors in the absence of mature bone cells [43]. The ability of baghdadite to induce bone formation in a critical size bone defect without adequate functionalization of the synthetic material with growth factor or osteogenic cells suggests that this material has osteoinductive properties consistent with its in vitro osteogenic interactions with human osteoblasts as previously described [17]. It could also be argued that bone formation far from bone surfaces in a critical bone defect is indicative of the scaffolds’ osteoinductivity. To monitor the process of the bone tissue formation, radiographical and l-CT measurements, histology and histomorphometry studies on the specimens were performed after 12 weeks of scaffolds implantation. Active bone regeneration in the radial defects was found in all scaffolds used; however, the bone regeneration process was different between baghdadite and Bag-PCL-nBG, compared to the TCP/HA and the TCP/HA-PCL-nBG, although both demonstrated good biocompatibility. A thin nanocomposite layer that coated over the surface of the scaffolds did not significantly affect the new bone formation in two groups of scaffolds. In this study, implantation of baghdadite scaffolds (with and without modification) resulted in increased new bone

formation, leading to defect bridging. Qualitative histological assessment demonstrated close association of new bone with the surface of the ceramic scaffolds (Fig. 7). It may be assumed that baghdadite ceramic, just like the TCP/HA, provides a biological scaffold for bone formation. The growth of new bone in direct contact with the baghdadite implants and associated ongoing bone remodelling, as indicated by osteoblast and osteoclast lined bone and ceramic surfaces, are also indications of excellent biocompatibility, as we previously demonstrated in vitro [17]. While mechanical properties constitute an important aspect of scaffolds destined for bone regeneration, the ability to remodel with bone over time is critical. This is partly governed by the scaffold’s material resorbability by osteoclasts and subsequent replacement by newly formed bone through osteoblast activity. Ideally, bioactive scaffolds for bone substitution should be able to be resorbed by osteoclasts [44] and replaced with new bone by osteoblasts, before the stability of the biomaterials is compromised. The presence of TRAP positive osteoclasts on ceramic surfaces, and the presence of baghdadite (but not TCP/HA) particles internalized in cells (Fig. 3f and h), are consistent with the occurrence of cell mediated ceramic degradation, though confirmation of full degradation and disappearance of the ceramic implants will be the subject of future studies. The baghdadite scaffolds, during the formation of new bone, maintained their structural integrity, while TCP/HA scaffolds collapsed, reinforcing the limitations imposed by its poor mechanical strength. The structural integrity of both scaffolds was maintained upon their modification with PCL and nBG. This is an important consideration for an ideal scaffold for use in clinical orthopedics.

5. Conclusion In conclusion, this study is the first to demonstrate that baghdadite scaffolds, with or without modification, are able to support bridging of large bone defects in the rabbit. This study shows that the baghdadite scaffolds have the potential for use in translational

4172

S.I. Roohani-Esfahani et al. / Acta Biomaterialia 8 (2012) 4162–4172

applications for maximizing bone formation in mechanically unfavourable environments. Acknowledgements The authors acknowledge the Australia National Health and Medical Research Council, Australian Research Council, Australian Orthopedic Association and the Rebecca Cooper Foundation. We acknowledge the Australian Center for Microscopy and Microanalysis at the University of Sydney for their help with microscopic analysis. Appendix A. Figures with essential colour discrimination Certain figures in this article, particularly Figs. 5, 7 and 8, are difficult to interpret in black and white. The full colour images can be found in the on-line version, at http://dx.doi.org/10.1016/ j.actbio.2012.07.036. References [1] Willie BM, Petersen A, Schmidt-Bleek K, Cipitria A, Mehta M, Strube P, et al. Designing biomimetic scaffolds for bone regeneration: why aim for a copy of mature tissue properties if nature uses a different approach? Soft Matter 2010;6:4976–87. [2] Seiler III JG, Johnson J. Iliac crest autogenous bone grafting: donor site complications. J South Orthop Assoc 2000;9:91–7. [3] Hutmacher DW, Schantz JT, Lam CX, Tan KC, Lim TC. State of the art and future directions of scaffold-based bone engineering from a biomaterials perspective. J Bone Joint Surg Am 2007;1:245–60. [4] Tabata Y. Recent progress in tissue engineering. Drug Discov Today 2001;6:483–7. [5] Greenwald AS, Boden SD, Goldberg VM, Khan Y, Laurencin CT, Rosier RN. Bonegraft substitutes: facts, fictions, and applications. J Bone Joint Surg – Series A 2001;83:98–103. [6] Ikada Y. Challenges in tissue engineering. J R Soc Interf 2006;3:589–601. [7] Boccaccini AR, Blaker JJ. Bioactive composite materials for tissue engineering scaffolds. Expert Rev Med Dev 2005;2:303–17. [8] Salgado AJ, Coutinho OP, Reis RL. Bone tissue engineering: state of the art and future trends. Macromol Biosci 2004;4:743–65. [9] Gentleman E, Fredholm YC, Jell G, Lotfibakhshaiesh N, O’Donnell MD, Hill RG, et al. The effects of strontium-substituted bioactive glasses on osteoblasts and osteoclasts in vitro. Biomaterials 2010;31:3949–56. [10] Arinzeh TL, Tran T, McAlary J, Daculsi G. A comparative study of biphasic calcium phosphate ceramics for human mesenchymal stem-cell-induced bone formation. Biomaterials 2005;26:3631–8. [11] Kasten P, Luginbühl R, Van Griensven M, Barkhausen T, Krettek C, Bohner M, et al. Comparison of human bone marrow stromal cells seeded on calciumdeficient hydroxyapatite, b-tricalcium phosphate and demineralized bone matrix. Biomaterials 2003;24:2593–603. [12] Yuan H, Van Den Doel M, Li S, Van Blitterswijk CA, De Groot K, De Bruijn JD. A comparison of the osteoinductive potential of two calcium phosphate ceramics implanted intramuscularly in goats. J Mater Sci Mater Med 2002;13:1271–5. [13] Khan SN, Tomin E, Lane JM. Clinical applications of bone graft substitutes. Orthop Clin North Am 2000;31:389–98. [14] Knabe C, Berger G, Gildenhaar R, Klar F, Zreiqat H. The modulation of osteogenesis in vitro by calcium titanium phosphate coatings. Biomaterials 2004;25:4911–9. [15] Ramaswamy Y, Wu C, Van Hummel A, Combes V, Grau G, Zreiqat H. The responses of osteoblasts, osteoclasts and endothelial cells to zirconium modified calcium-silicate-based ceramic. Biomaterials 2008;29:4392–402. [16] Zreiqat H, Ramaswamy Y, Wu C, Paschalidis A, Lu Z, James B, et al. The incorporation of strontium and zinc into a calcium–silicon ceramic for bone tissue engineering. Biomaterials 2010;31:3175–84. [17] Ramaswamy Y, Wu C, Zhou H, Zreiqat H. Biological response of human bone cells to zinc-modified Ca-Si-based ceramics. Acta Biomater 2008;4:1487–97.

[18] Shor L, Güçeri S, Wen X, Gandhi M, Sun W. Fabrication of three-dimensional polycaprolactone/hydroxyapatite tissue scaffolds and osteoblast-scaffold interactions in vitro. Biomaterials 2007;28:5291–7. [19] Agrawal CM, Ray RB. Biodegradable polymeric scaffolds for musculoskeletal tissue engineering. J Biomed Mater Res 2001;55:141–50. [20] Shao H, Shao J, Bian H, Zhao Q, Zhou Q, Qi J, et al. Loaded true bone ceramic heals rabbit critical-sized segmental radial defect. Die Pharmazie 2012;67:247–52. [21] Zhang W, Li G, Deng R, Deng L, Qiu S. New bone formation in a true bone ceramic scaffold loaded with desferrioxamine in the treatment of segmental bone defect: a preliminary study. J Orthop Sci 2012;17:1–10. [22] Guo X, Zheng Q, Kulbatski I, Yuan Q, Yang S, Shao Z, et al. Bone regeneration with active angiogenesis by basic fibroblast growth factor gene transfected mesenchymal stem cells seeded on porous b-TCP ceramic scaffolds. Biomed Mater 2006;1:93–9. [23] Hao W, Dong J, Jiang M, Wu J, Cui F, Zhou D. Enhanced bone formation in large segmental radial defects by combining adipose-derived stem cells expressing bone morphogenetic protein 2 with nHA/RHLC/PLA scaffold. Int Orthop 2010;34:1341–9. [24] Roohani-Esfahani SI, Nouri-Khorasani S, Lu ZF, Appleyard RC, Zreiqat H. Effects of bioactive glass nanoparticles on the mechanical and biological behavior of composite coated scaffolds. Acta Biomater 2011;7:1307–18. [25] Kokubo T, Takadama H. How useful is SBF in predicting in vivo bone bioactivity? Biomaterials 2006;27:2907–15. [26] An YH, Friedman RJ. Animal models of orthopedic implant infection. J Invest Surg 1998;11:139–46. [27] Bodde EWH, Spauwen PHM, Mikos AG, Jansen JA. Closing capacity of segmental radius defects in rabbits. J Biomed Mater Res A 2008;85:206–17. [28] Knabe C, Kraska B, Koch C, Gross U, Zreiqat H, Stiller M. A method for immunohistochemical detection of osteogenic markers in undecalcified bone sections. Biotech Histochem 2006;81:31–9. [29] Wang J, Wang BM, Schwendeman SP. Characterization of the initial burst release of a model peptide from poly(D,L-lactide-co-glycolide) microspheres. J Control Release 2002;82:289–307. [30] Karageorgiou V, Kaplan D. Porosity of 3D biomaterial scaffolds and osteogenesis. Biomaterials 2005;26:5474–91. [31] Mastrogiacomo M, Scaglione S, Martinetti R, Dolcini L, Beltrame F, Cancedda R, et al. Role of scaffold internal structure on in vivo bone formation in macroporous calcium phosphate bioceramics. Biomaterials 2006;27:3230–7. [32] Silva MMCG, Cyster LA, Barry JJA, Yang XB, Oreffo ROC, Grant DM, et al. The effect of anisotropic architecture on cell and tissue infiltration into tissue engineering scaffolds. Biomaterials 2006;27:5909–17. [33] Melchels FPW, Barradas AMC, Van Blitterswijk CA, De Boer J, Feijen J, Grijpma DW. Effects of the architecture of tissue engineering scaffolds on cell seeding and culturing. Acta Biomater 2010;6:4208–17. [34] Guo D, Xu K, Han Y. The in situ synthesis of biphasic calcium phosphate scaffolds with controllable compositions, structures, and adjustable properties. J Biomed Mater Res A 2009;88:43–52. [35] Rezwan K, Chen QZ, Blaker JJ, Boccaccini AR. Biodegradable and bioactive porous polymer/inorganic composite scaffolds for bone tissue engineering. Biomaterials 2006;27:3413–31. [36] Roohani-Esfahani SI, Nouri-Khorasani S, Lu Z, Appleyard R, Zreiqat H. The influence hydroxyapatite nanoparticle shape and size on the properties of biphasic calcium phosphate scaffolds coated with hydroxyapatite-PCL composites. Biomaterials 2010;31:5498–509. [37] Nalla RK, Kinney JH, Ritchie RO. Mechanistic fracture criteria for the failure of human cortical bone. Nat Mater 2003;2:164–8. [38] Nalla RK, Kinney JH, Ritchie RO. Effect of orientation on the in vitro fracture toughness of dentin: the role of toughening mechanisms. Biomaterials 2003;24:3955–68. [39] Mohamad Yunos D, Bretcanu O, Boccaccini AR. Polymer-bioceramic composites for tissue engineering scaffolds. J Mater Sci 2008;43:4433–42. [40] Chen QZ, Thompson ID, Boccaccini AR. 45S5 BioglassÒ-derived glass-ceramic scaffolds for bone tissue engineering. Biomaterials 2006;27:2414–25. [41] Wagoner Johnson AJ, Herschler BA. A review of the mechanical behavior of CaP and CaP/polymer composites for applications in bone replacement and repair. Acta Biomater 2011;7:16–30. [42] Hench LL. Bioactive ceramics. Ann N Y Acad Sci 1988;523:54–71. [43] Kawai T, Mieki A, Ohno Y, Umemura M, Kataoka H, Kurita S, et al. Osteoinductive activity of composites of bone morphogenetic protein and pure titanium. Clin Orthop Relat Res 1993;290:296–305. [44] Teitelbaum SL. Bone resorption by osteoclasts. Science 2000;289:1504–8.