- Email: [email protected]
Monolithic and assembled polymer–ceramic composites for bone regeneration
Acta Biomaterialia 9 (2013) 5708–5717
Contents lists available at SciVerse ScienceDirect
Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat
Monolithic and assembled polymer–ceramic composites for bone regeneration Anandkumar Nandakumar a, Célia Cruz a,b, Anouk Mentink a, Zeinab Tahmasebi Birgani a, Lorenzo Moroni a, Clemens van Blitterswijk a, Pamela Habibovic a,⇑ a b
Department of Tissue Regeneration, MIRA Institute for Biomedical Technology and Technical Medicine, University of Twente, The Netherlands Faculty of Engineering, University of Porto, Rua Dr. Roberto Frias, Porto, Portugal
a r t i c l e
i n f o
Article history: Received 31 July 2012 Received in revised form 23 October 2012 Accepted 30 October 2012 Available online 7 November 2012 Keywords: Polymer/ceramic composite Rapid prototyping Conventional Assembled Bone regeneration
a b s t r a c t The rationale for the use of polymer–ceramic composites for bone regeneration stems from the natural composition of bone, with collagen type I and biological apatite as the main organic and inorganic constituents, respectively. In the present study composite materials of PolyActive™ (PA), a poly(ethylene oxide terephthalate)/poly(butylene terephtalate) co-polymer, and hydroxyapatite (HA) at a weight ratio of 85:15 were prepared by rapid prototyping (RP) using two routes. In the first approach pre-extruded composite filaments of PA–HA were processed using three-dimensional fibre deposition (3DF) (conventional composite scaffolds). In the second approach PA scaffolds were fabricated using 3DF and combined with HA pillars produced inside stereolithographic moulds that fitted inside the pores of the PA threedimensional structure (assembled composite scaffolds). Analysis of calcium and phosphate release in a simulated physiological solution, not containing calcium or phosphate, revealed significantly higher values for the HA pillars compared with other scaffolds. Release in simulated body fluid saturated with respect to HA did not show significant differences among the different scaffolds. Human mesenchymal stromal cells were cultured on polymer (3DF), conventional composite (3DF-HA) and assembled composite (HA assembled in 3DF) scaffolds and assessed for morphology, metabolic activity, DNA amount and gene expression of osteogenic markers using real time quantitative polymerase chain reaction (PCR). Scanning electron microscopy images showed that the cells attached to and infiltrated all the scaffolds. Assembled composites had a higher metabolic activity compared with 3DF-HA scaffolds while no significant differences were observed in DNA amounts. Gene expression of osteopontin in the assembled composite was significantly higher compared with the conventional composites. The strategy of composite fabrication by assembly appears to be a promising alternative to the conventional composite fabrication route for scaffolds for bone regeneration. Ó 2012 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction Bone is a naturally occurring biocomposite composed of organic and inorganic components and water. The organic part of bone is mainly composed of collagen type I among other proteins and is responsible for the tensile strength and toughness of bone tissue, while the inorganic mineral component is calcium phosphate, essentially hydroxyapatite (HA), that provides stiffness and strength to bone [1,2]. Besides this mechanical role, the mineral part of bone is of great importance in calcium homeostasis [3,4]. While autografts are generally considered the gold standard in bone repair and regeneration, problems with donor site morbidity and availability exist [5], while allografts suffer from a risk of disease transmission [6,7]. To overcome these issues recent developments have focused on alternative regenerative strategies, such as ⇑ Corresponding author. Tel.: +31 53 489 3400. E-mail address: [email protected] (P. Habibovic).
synthetic bone graft substitutes and tissue engineered constructs [8,9]. Whether a biomaterial is used as a stand-alone bone graft substitute or as a scaffold that acts as a carrier of cells and/or growth factors in a tissue engineered construct, a number of requirements regarding mechanical strength, porosity, pore interconnectivity and bioactivity need to be met for successful use in a clinical setting. In order to meet these various requirements combinations of different types of materials and different fabrication processes are often required. While with polymers a wide range of mechanical properties can be achieved, calcium phosphate ceramics suffer from intrinsic brittleness. On the other hand, the bioactivity of calcium phosphates, in terms of osteoconduction, osteoinduction and bone bonding capacity, is in most cases superior to that of polymers, owing to the chemical composition of calcium phosphate ceramics and events occurring on their surface in a physiological environment [8]. With the rationale of recreating the composition of bone, polymer–ceramic composite scaffolds have been explored as potential bone graft substitutes. Synthetic
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.2012.10.044
A. Nandakumar et al. / Acta Biomaterialia 9 (2013) 5708–5717
polymers like poly(L-lactic acid) (PLLA) [10–12], poly(e-caprolactone) (PCL) [13,14], poly(lactic-co-glycolic acid) (PLGA) [10,15,16] and naturally occurring collagen [17,18] have been combined with hydroxyapatite (HA) to fabricate composite materials. Use of conventional fabrication methods like salt leaching–phase inversion [19], thermally induced phase separation (TIPS) [20] and other similar techniques do not offer strict control over the porosity and interconnectivity and invariably use organic solvents. Pore interconnectivity is essential for the supply of nutrients and metabolites while residuals of organic solvents can be cytotoxic [21]. The need to control the architecture, porosity and mechanical properties of scaffolds led to the advent of rapid prototyping (RP) technologies. Several RP techniques, such as selective laser sintering (SLS) [22,23], fused deposition modelling (FDM) [13] and precision extrusion deposition (PED) [24], have been used to fabricate composites by combining polymers like poly(hydroxybutyrate-co-hydroxyvalerate), PLLA, and PCL with different calcium phosphate phases, such as carbonated hydroxyapatite, HA and b-tricalcium phosphate. These techniques also offer the advantage of being integrated to imaging modalities (e.g. magnetic resonance imaging and computed tomography) in order to fabricate – scaffolds and implants customized for a particular patient. In the present study we have used three-dimensional (3-D) fibre deposition (3DF) to fabricate two types of scaffolds consisting of PolyActive™ (PA), a poly(ethylene oxide terephthalate)/ poly(butylene terephtalate) (PEOT/PBT) co-polymer, and an HA ceramic to assemble a monolithic composite. In the first, conventional approach, filaments of polymer ceramic composite at desired weight ratios were extruded and used as the starting material to prepare the scaffolds by 3DF. In the other approach, previously developed by our group [25], the polymer scaffolds were fabricated by 3DF while the ceramic particles were fabricated in the form of pillars by sintering a ceramic slurry poured into negative molds prepared by stereolithography. The two components were then assembled by manually press fitting the HA pillars into the pores of the 3DF scaffold to create a composite. 3DF is an extrusion-based RP technique in which a polymer is melted and extruded through a nozzle and the material deposition is controlled by a robotic arm. This method allows the fabrication of scaffolds with closely controlled porosity, pore size, interconnectivity and fibre orientation between successive layers. 3DF has been used to fabricate polymeric scaffolds for tissue engineering of cartilage [26] and osteochondral [25] defects. While 3DF is a very versatile technique for fabricating polymeric scaffolds, the processing of composites is more difficult owing to the high viscosity of the polymer–ceramic melt resulting in clogging of the nozzles. In addition to affecting fabrication and processability, these phenomena also limit the amount of ceramic that can be incorporated into the scaffold, while the amount of ceramic largely determines the osteoconductivity and osteoinductivity of composite materials. For example, Barbieri et al. recently demonstrated osteoinductivity of a PLLA/HA composite, but only when the ceramic component was as high as 40 wt.% [27]. Furthermore, in a conventional monolithic composite the majority of the ceramic particles are covered by polymer, while accessibility of the ceramic surface to the physiological environment, which is associated with dissolution/reprecipitation processes, is believed to be the origin of the bioactivity of calcium phosphates [28–30]. Our study aimed at increasing the amount of ceramic component in the composite and improving the exposure of its surface, while employing 3DF as a RP technique to closely control the scaffold architecture. hMSCs were seeded on the monolithic and assembled composite scaffolds, as well as on a polymer control, and cell proliferation and osteogenic differentiation were studied.
5709
2. Materials and methods PA, a PEOT/PBT co-polymer with tunable physico-chemical properties that has previously been used in different forms for tissue engineering applications [26,31,32], was provided by PolyVation BV (The Netherlands). Following the aPEOTbPBTc nomenclature the composition used in this study was 300PEOT55PBT45, where a is the molecular weight in g mol1 of the starting poly(ethylene glycol) (PEG) blocks used in the co-polymerization, while b and c are the weight ratios of the PEOT and PBT blocks, respectively. The HA powder for ceramic preparation was provided by Progentix BV (The Netherlands). It was synthesized using a wet precipitation method. The powder was not sintered and was phase pure HA with a particle size below 63 lm. 2.1. Production of composite filaments and fabrication of 3DF scaffolds by RP Composite filaments of PA and HA were extruded using a screw extruder (Artecs). Briefly, the desired amounts of PA pellets and HA powder were added to the extruder and mixed for 5 min at a temperature of 180 °C. The composite filaments were then drawn out and cut into small pellets for processing by RP. For scaffold fabrication composite filaments with a maximum of 15% HA by weight were used. 3-D scaffolds were fabricated by PR using a Bioplotter (Envisiontec GmbH, Germany), which is an xyz robotic device as previously described [33]. Briefly, polymer or composite granules (PA– HA) were loaded onto a stainless steel syringe set at a temperature between 195 and 210 °C (higher temperatures for composites) using a thermostat controlled cartridge unit, fixed on the x mobile arm of the apparatus. A pressure of 5 bars (nitrogen) for composites (4 bars for polymer) was applied to the syringe through a pressurized cap once the polymer had melted. Rectangular block models were loaded in the Bioplotter CAM (PrimCAM, Switzerland) software and deposited layer by layer as the polymer was extruded through a nozzle (0.7 mm OD) onto a stage. The deposition speed was varied between 150 and 250 mm min1 (lower speeds for composites). The spacing between fibres in the same layer was set at 0.8 mm, the layer thickness was 0.225 mm and the scaffold height was set at 4 mm. A 0-90° configuration was used for the scaffold architecture where fibres were deposited with 90° orientation steps between successive layers. 2.2. Preparation of the ceramic slurry and moulds and assembly of HA into 3DF scaffolds Pillars of HA were fabricated in a two step process. First, a negative mask was designed using Rhinoceros and fabricated by stereolithography (Envisiontec, Germany). The masks were then filled with an HA slurry prepared by vigorously mixing 8.2 g of calcined HA (20 h in an oven at 1000 °C), 3.55 g of non-calcined HA, 5 g of demineralized water, 0.3 g of methylcellulose solution (2% w/v methylcellulose in demineralized water), 0.55 g of ammonia, and 3.525 g of sieved paraffin particles (600–1000 lm), resulting in 30% porosity of the HA particles. The HA pillars were obtained by sintering in a furnace (Nabertherm, Germany) at 1150 °C, resulting in a fully crystalline phase pure HA ceramic [34]. The HA particles were sterilized by autoclaving and the pillars were press fitted into the pores of the 3DF scaffolds. 2.3. Characterization of the scaffolds The architecture and composition of the different types of scaffolds were characterized by scanning electron microscopy (SEM)
5710
A. Nandakumar et al. / Acta Biomaterialia 9 (2013) 5708–5717
(XL30 ESEM-FEG, Philips). The porosity of the RP scaffolds was calculated from SEM measurements of fibre diameter, layer thickness and fibre spacing using the formula [35]:
Porosity ¼ 1
p 4
1
1
Fibre spacing Layer thickness Fibre diameter Fibre diameter
Compression tests were performed on scaffolds with a diameter of 4 mm and a length of 4 mm using a mechanical test apparatus (Zwick Z050, Germany) to obtain the modulus of elasticity under compression. A 0.5 kN load cell was used and the samples were compressed at a rate of 1 mm min1 up to a maximum of 30% strain. The modulus was calculated from the stress–strain curves as the slope of the initial linear region (5–8% strain) neglecting the toe region due to initial settling of the specimen. A total of four samples per scaffold group were evaluated. A representative stress–strain curve for each scaffold type is shown in Supplementary Fig. 1. 2.4. Calcium and phosphate release Release of calcium and phosphate was studied in a simulated physiological solution (SPS) containing 137 mM Na+, 177 mM Cl and 50 mM HEPES (N-[2-hydroxyethyl]piperazine-N0 -(2-ethanesulfonic acid)) buffered at pH 7.4 and in a simplified simulated body fluid (SBF JL-2) containing 142 mM Na+, 109.90 mM Cl, 2+ 34.88 mM HCO and 1.39 mM HPO2 3 , 2.31 mM Ca 4 , pH 7.4, according to Bohner and Lemaitre [36]. Sodium azide was used as an antimicrobial agent. Triplicates of each sample (PA, 3DF-HA and HA pillars) weighing 250 mg each were soaked in 30 ml of the solution at 37 °C under shaking for 3 weeks, after which samples were removed and the concentrations of calcium and phosphate measured using QuantiChrom™ Calcium Assay (DICA-500) and QuantiChrom™ Phosphate Assay (DIPA-500) kits. The optical density of samples after adding reagents from the kits was read with a microplate spectrophotometer (Thermo Scientific Multiskan GO) at 612 and 620 nm for calcium and phosphate release, respectively. HA pillars were used in the release study in order to determine the maximum amounts of calcium and phosphate ions that could be released. After immersion in SBF for 21 days one sample of each type was used for SEM analysis. 2.5. Human mesenchymal stromal cell (hMSC) isolation and culture Bone marrow aspirates were obtained after written informed consent and hMSCs were isolated and proliferated as described previously [37,38]. Briefly, aspirates were resuspended using 20 gage needles, plated at a density of 5 105 cells cm2 and cultured in hMSC proliferation medium containing a-minimal essential medium (Life Technologies), 10% foetal bovine serum (Lonza), 0.2 mM ascorbic acid (Life Technologies), 2 mM L-glutamine (Life Technologies), 100 units ml1 penicillin (Life Technologies), 10 lg ml1 streptomycin (Life Technologies), and 1 ng ml1 basic fibroblast growth factor (FGF) (Instruchemie). Cells were grown at 37 °C in a humid atmosphere with 5% CO2. The medium was refreshed twice a week, and cells were used for further sub-culturing or cryopreservation. The hMSC basic medium was composed of hMSC proliferation medium without basic FGF. Mineralization medium was composed of hMSC basic medium supplemented with 108 M dexamethasone (Sigma, St. Louis, MO) and 0.01 M b-glycerophosphate (BGP) (Sigma). Cells were trypsinized prior to seeding on scaffolds. The in vitro experiments were performed with cells from one donor at passage 3. Scaffolds (10 10 4 mm3) were soaked in 70% ethanol for 30 min and dried overnight in a laminar flow cabinet. The scaffolds were washed twice with sterile phosphate-buffered saline (PBS), transferred to a 25-well non-treated polystyrene
plate (Greiner Bio One) and incubated at 37 °C in a humid atmosphere with 5% CO2 for 4 h in basic cell culture medium. After removing the medium each scaffold was seeded with 800,000 cells in approximately 105 ll of basic medium. The cell–scaffold constructs were incubated for 3 h to allow cell attachment and topped up to 2 ml with the appropriate medium. Scaffolds were cultured in either basic medium or mineralization medium and the medium was refreshed twice a week. As a control T25 flasks were seeded at 5000 cells cm2 in basic and mineralization medium. 2.6. Cell attachment, metabolism and amount of DNA Cell metabolism was assessed using the alamarBlueÒ assay according to the manufacturer’s protocol. Briefly, culture medium was replaced with medium containing 10 vol.% alamarBlueÒ solution (Biosource, Camarillo, CA) and the cell–scaffold constructs (n = 3) were incubated at 37 °C for 4 h. Fluorescence was measured at 590 nm in a Perkin Elmer Victor3 1420 Multilabel plate reader. Cell metabolism was analysed on days 11 and 21 and the readout from the scaffolds was corrected with a blank from each group. After 21 days culture seeded scaffolds (n = 3) were washed with PBS and frozen at 80 °C overnight. The constructs were then digested at 56 °C in a Tris–EDTA buffered solution containing 1 mg ml1 proteinase K, 18.5 lg ml1 pepstatin A and 1 lg ml1 iodoacetamide (Sigma–Aldrich) for 18 h. DNA amounts were determined using the CyQuantÒ DNA quantification kit (Invitrogen) with 50 ll of cell lysate according to the manufacturer’s protocol. Fluorescence at an excitation wavelength of 480 nm and an emission wavelength of 520 nm was measured using a Perkin Elmer Victor3 1420 Multilabel plate reader. All data for metabolic activity and DNA assay were normalized to the dry weight of the corresponding scaffold. 2.7. Cell morphology After 21 days culture SEM analysis was performed to assess the cell response to different scaffolds. The medium was removed and the scaffolds washed twice with PBS and fixed in 10% formalin for 1 h. After rinsing with PBS the scaffolds were dehydrated in a series of increasing ethanol concentrations (70%, 80%, 90%, 96%, 100% 2), 15 min in each concentration, before drying in a critical point dryer (Balzers CPD-030). The samples were then sputter coated with gold (Cressington) for observation by SEM. 2.8. Cell differentiation To analyse the expression of osteogenic markers by hMSCs total RNA was isolated using a combination of the TRIzolÒ method with the NucleoSpinÒ RNA II isolation kit (Bioké). Briefly, scaffolds (n = 3) were washed once with PBS and 1 ml of TRIzol reagent (Invitrogen) was added to the samples. After 5 min the samples were stored at 80 °C for RNA isolation. After chloroform addition and phase separation by centrifugation the aqueous phase containing the RNA was collected, mixed with an equal volume of 75% ethanol and loaded onto the RNA binding column of the kit. Subsequent steps were in accordance with the manufacturer’s protocol. RNA was collected in RNase-free water. The quality and quantity of RNA was analysed by gel electrophoresis and spectrophotometry. 750 ng of RNA were used for first strand cDNA synthesis using iScript (Bio-Rad) according to the manufacturer’s protocol. 1 ll of undiluted cDNA was used for subsequent analysis. PCR was performed in an iQ5 real time PCR machine (Bio-Rad) using SYBR Green supermix (Bio-Rad). Expression of osteogenic marker genes was normalized to glyceraldehyde-3-phosphate dehydrogenase (50 -CGCTCTCTGCTCCTCCTGTT-30 and 50 CCATGGTGTCTGAGCGATGT-30 ) levels and fold induction was
A. Nandakumar et al. / Acta Biomaterialia 9 (2013) 5708–5717
calculated using the comparative DCt method [39]. The following primer sequences were used for the osteogenic marker genes: bone sialoprotein (BSP) (50 -TGCCTTGAGCCTGCTTCC-30 and 50 -CAAAATTAAAGCAGTCTTCATTTTG-30 ); osteopontin (OP) (50 -CCAAGTAAGTCCAACGAAAG-30 and 50 -GGTGATGTCCTCGTCTGTA-30 ); alkaline phosphatase (ALP) (50 -GACCCTTGACCCCCACAAT-30 and 50 -GCTCGTACTGCATGTCCCCT-30 ). For amplification of bone morphogenetic protein-2 (BMP-2) a gene-specific primer mix was used (SA Biosciences) according to the manufacturer’s protocol.
2.9. Statistical analysis One-way ANOVA with Tukey’s multiple comparison post hoc test was performed. The level of significance was set at 0.05. All data presented are expressed as means ± standard deviation. Only significant differences between different scaffold groups in the same medium or the same scaffold type in different media are shown.
3. Results 3.1. Fabrication and characterization of scaffolds RP using 3DF was shown to be a suitable technique for fabricating polymeric and composite 3-D scaffolds. An SEM image of rapid prototyped 3DF and 3DF-HA scaffolds in top view and cross-section is presented in Fig. 1. Strands of the scaffolds appeared to decrease in thickness while being deposited, possibly due to the large fibre spacing. The fibre diameter, layer thickness and fibre spacing were calculated to be 0.315 ± 0.1, 0.221 ± 0.02 and 1.96 ± 0.13 mm, and 0.34 ± 0.08, 0.238 ± 0.03 and 1.84 ± 0.03 mm for the 3DF and 3DF-HA scaffolds, respectively. The porosities of the 3DF and 3DF-HA scaffolds were determined to be 81.9% and 79.3%, respectively.
5711
The assembled composites were fabricated by inserting the ceramic pillars into the 3DF scaffolds. Fig. 2A shows optical images of various configurations in which HA pillars can be assembled into 3DF scaffolds. For the experiments the weight of the incorporated HA pillars was kept to 15% of the total weight (a configuration similar to the image on the left in Fig. 2A) of the construct to ensure that equal amounts of HA were incorporated into the assembled and conventional composite scaffolds. Fig. 2B shows an SEM micrograph of the assembled composite with a ceramic pillar in the middle. The structure of the HA particles is shown in Fig. 2C and D. The grain size of the HA ceramic was between 0.8 and 1.5 lm, as measured on the SEM images. The results of the mechanical tests are shown in Fig. 3. The incorporation of HA in the scaffold bulk had an impact on the scaffold stiffness. The modulus of elasticity under compression of the 3DF-HA scaffolds (39 ± 15 MPa) was significantly higher than that of the other two scaffold groups. The addition of HA as pillars increased the modulus of the assembled construct (9 ± 4.5 MPa) compared with the 3DF scaffold (8.1 ± 2.4 MPa), but this increase was minimal. Release of calcium and phosphate was measured in two physiological solutions after 21 days immersion, SPS, with an ionic strength similar to that of human blood plasma but without calcium or phosphate, and SBF saturated with respect to HA, (Table 1). In SPS calcium was only detected in the solutions containing HA pillars. The phosphate concentration of the SPS containing HA pillars was about 10 times higher than that in which the 3DF composite was immersed, while only background levels were found in the 3DF controls. Immersion of all samples in SBF for 21 days resulted in a milky solution with precipitates visible. The same was the case for the SBF solution alone. Concentrations were measured after the precipitate was filtered through a 0.45 lm filter, and no significant differences were observed among different scaffolds in terms of either calcium or phosphate concentration (Table 1). SEM analysis revealed the deposition of globular precipitates on all scaffold types immersed in SBF, although their amount seemed to be great-
Fig. 1. Conventional rapid prototyping (RP) for fabrication of polymeric and composite scaffolds. (A) SEM image of a 3DF scaffold fabricated by RP. (B) Cross-section of a 3DF scaffold. (C) SEM image of a monolithic polymer–ceramic composite (3DF-HA). (D) Cross-section of a 3DF-HA composite. Scale bars: (A and C) 2 mm; (B and D) 1 mm.
5712
A. Nandakumar et al. / Acta Biomaterialia 9 (2013) 5708–5717
Fig. 2. Assembled polymer–ceramic composite fabricated using RP and stereolithography. (A) Photographs of assembled composites with different amounts of HA, fabricated using a negative stereolithography mould, loaded on a 3DF polymer scaffold fabricated by RP (scale bar 5 mm). (B) SEM image of an assembled scaffold with an HA pillar in one of the pores (scale bar 1 mm). (C and D) Microstructure of HA at different magnifications (scale bars 50 and 10 lm, respectively).
Fig. 3. Modulus of elasticity under compression. The modulus of 3DF-HA scaffolds was significantly higher (P < 0.01) compared with the 3DF and assembled composites.
Table 1 Calcium and phosphate concentrations in SPS and SBF after 21 days immersion of 3DF, 3DF-HA and HA pillars.
3DF 3DF-HA HA pillars
Concentration in SPS (lM)
Concentration in SBF (lM)
Ca2+
PO3 4
Ca2+
PO3 4
0 0 30.2 ± 9.4
0 6.6 ± 0.6 95.1 ± 23.6
126.0 ± 27.0 131.0 ± 49.3 156.1 ± 10.8
344.9 ± 25.9 422.6 ± 39.1 389.7 ± 31.6
er on 3DF-HA and HA pillars compared with 3DF control samples (Fig. 4A–C). Further, deposits on 3DF-HA composites and HA pillars consisted of sharp crystals that were oriented perpendicular to the scaffold surface, while the deposits found on the 3DF samples had a smoother morphology (Fig. 4D–F). Energy-dispersive X-ray analysis demonstrated the presence of calcium and phosphate in all precipitates (data not shown).
3.2. Cell morphology, metabolic activity and proliferation Fig. 5 shows SEM images of scaffolds after 21 days cell culture. Qualitative SEM analyses suggested a lower number of cells in the 3DF-HA scaffolds compared with the other two groups. Cells were observed throughout the scaffolds, confirming their infiltration into the pores, and they were also observed on the HA pillars in the assembled composite. Extracellular matrix (ECM) production was observed on all scaffold types. alamarBlueÒ assays were performed on days 11 and 21 to assess cell metabolism in the various scaffolds (Fig. 6A). On day 11 no significant differences in cell metabolism were observed among the different scaffold groups, irrespective of the culture medium. On day 21 in basic medium cells cultured on the 3DF-HA composites had a significantly lower metabolic activity compared with the 3DF and assembled composites. The same trend was observed in mineralization medium, although statistically significant differences were only observed between the 3DF and 3DF-HA groups.
A. Nandakumar et al. / Acta Biomaterialia 9 (2013) 5708–5717
5713
Fig. 4. SEM images showing mineral formation upon immersion in SBF for 21 days: (A and D) 3DF; (B and E) 3DF-HA composite; (C and F) HA pillars. Scale bars: (A–C) 200 lm; (D–F) 5 lm.
Fig. 5. SEM images of cells on (A and D) 3DF, (B and E) 3DF + HA and (C and F) assembled scaffolds after 21 days culture in (A–C) basic and (D–F) mineralization medium. (Inset C) Cells on a HA pillar inside the assembled construct. Scale bar 500 lm. (Inset in F) hMSCs along with the ECM produced. Scale bar 10 lm.
A quantitative biochemical assay to determine the amount of DNA in different scaffolds was performed after 21 days as an indirect measure of cell numbers (Fig. 6B). No statistically significant differences due to effect of culture medium or scaffold type were observed, although a trend similar to that for metabolic activity was seen, with the 3DF-HA scaffolds having the lowest amount of DNA in both media among the different scaffold groups. 3.3. Gene expression analysis of osteogenic markers The gene expression of different osteogenic markers was analysed using quantitative real time PCR, and the results are shown in Fig. 7. Only statistically significant differences among cells grown on different scaffolds in the same medium are indicated in the figure. Mineralization medium had a positive effect on the expression of BSP and ALP compared with basic medium, and this effect was independent of the scaffold type. In contrast, BMP-2 expression in mineralization medium was lower than in basic medium on all scaffold types. Cells on 3DF scaffolds in basic medium showed a basal level of BSP expression which was significantly higher compared with the 3DF-HA scaffolds. In mineralization medium the 3DF-HA scaffolds showed basal expression of BSP,
which was up-regulated in the 3DF scaffolds and assembled composites, although not significantly. ALP expression by cells on the 3DF scaffolds in basic medium was also at basal levels, while it was down-regulated in both composites (significantly for the assembled composite). No differences among the scaffolds were observed in mineralization medium. Cells on the assembled composite scaffolds showed significant up-regulation of OP in basic medium compared with the 3DF and 3DF-HA scaffolds. Up-regulation of gene expression was also observed in mineralization medium, although no statistically significant differences were observed. In the case of BMP-2 the trend in expression levels was 3DF < 3DF-HA < assembled composite, although no significant differences were observed. 4. Discussion 3-D composite scaffolds containing PA and HA were fabricated using two different approaches. The conventional approach to fabricate composite scaffolds using polymer–HA composite filaments was successful only up to a 15 wt.% loading of HA. Above this amount it was impossible to fabricate 3-D scaffolds with the current choice of nozzles and equipment due to clogging by the cera-
5714
A. Nandakumar et al. / Acta Biomaterialia 9 (2013) 5708–5717
Fig. 6. (A) Metabolic activity of hMSCs cultured on different scaffolds in basic and mineralization medium on days 11 and 21, measured by alamarBlueÒ assay. (B) Amount of DNA after 21 days on different scaffolds in basic and mineralization medium as measured by CyQuant assay. Data are represented as means ± standard deviation. ⁄Statistically significant differences (P < 0.05).
mic particles and a consequent increase in the viscosity of the molten blend. The amount of ceramic that could be loaded in the present study was comparable with the amount in a study by Lam et al. [40], who succeeded in preparing a b-tricalcium phosphate (bTCP)–PCL composite with a 20 wt.% ceramic content by fused deposition modelling. Using a precision extrusion process Shor et al. [24] obtained a maximum loading of 25 wt.% HA in PCL. Higher ceramic loadings are possible using solvent casting methods [41], but these involve the use of possibly cytotoxic organic solvents. Besides, the scaffolds produced using solvent casting often lack close control over the porosity and architecture. In this regard, the assembly approach is innovative and versatile. As shown in Fig. 2A, different amounts of HA can be loaded onto scaffolds without the problems associated with processing of high amounts of ceramic and the need for complex extrusion systems. In addition, in the assembled constructs the ceramic surface is exposed to the
biological environment, resulting in different surface dynamics of dissolution/reprecipitation and protein adsorption and, hence, an effect on cell behaviour compared with bulk composites where the majority of the ceramic particles are covered by polymer and exposure is dependent on polymer degradation. Compression tests performed to evaluate the stiffness of the scaffolds showed that the addition of HA in the bulk increased the stiffness by almost 5 times compared with the 3DF polymer only scaffolds. This result was to be expected since in an in vivo setting the mineral component of bone imparts stiffness. In a previous study [25] assembled composites with different shapes of biphasic calcium phosphate (BCP) (75.4% HA, 24.6% b-TCP) particles were inserted into 3DF scaffolds and an increase in bending storage modulus and compressive storage modulus were observed compared with the 3DF scaffolds during dynamic mechanical analysis. In the current study no significant increase in modulus was
A. Nandakumar et al. / Acta Biomaterialia 9 (2013) 5708–5717
5715
Fig. 7. qPCR analysis for a panel of osteogenic genes after 21 days culture. Data are represented as means ± standard deviation. ⁄Statistically significant differences (P < 0.05).
observed due to the insertion of HA pillars. One possible reason could be the different loading of the ceramic particles. In the earlier study [25] BCP particles formed about 60% of the scaffold weight, while only 15% by weight of HA was used in the present study. In a review on the mechanical properties of calcium phosphates and composites [42] it was observed that the bulk properties of HA and BCP were spread across two orders of magnitude (compressive strength 5–500 MPa). In general, b-TCP possesses lower strength and the other mechanical properties are also poorer. Therefore, it can be reasonably stated that if the same amount of HA was used instead of BCP the resulting scaffolds would have had comparable or better mechanical properties. While the conventional composites (40 MPa) had a modulus above the minimum range stated for cancellous bone (12 MPa) [43], the polymeric and assembled scaffolds fell slightly short (8–9 MPa). However, it is possible to further increase the modulus of the assembled composite scaffolds by increasing the loading of HA particles as insertion of more ceramic particles has been shown to lead to higher values [25]. This adds another interesting facet to the assembly-based fabrication approach, the differential loading of ceramic particles to tune scaffold stiffness. Control over the mechanical properties, as well as over implant shape and architecture, increases clinical applicability of the assembled constructs to different sites, ranging from, for example, unloaded skull defects to load-bearing sites of long bones. On the other hand, it should not be ignored that an increase in the ceramic content, which can be advantageous in terms of construct bioactivity and mechanical properties, will decrease the porosity and pore interconnectivity of the assembled construct, which are of great importance for nutrient and oxygen supply, as well as cell and tissue infiltration and proliferation. Nevertheless, different components of the assembled constructs, such as polymer choice, ceramic choice and loading regime, can all be used as tools to design optimal scaffolds for the intended application. As mentioned before, the origin of the bioactivity of calcium phosphate-based bone graft substitutes is suggested to lie in dissolution/reprecipitation processes occurring on the surface of the
material in physiological environments [28–30,44]. However, as recently reviewed, it is difficult to pinpoint a single parameter that is responsible for the biological responses to a calcium phosphate material in vitro or in vivo [45]. While elevated concentrations of calcium have been suggested to positively affect osteogenic proliferation and differentiation of, for example, hMSCs and human periosteal derived stem cells (hPDCs) [46,47], release of calcium from a calcium phosphate ceramic is accompanied by other events, such as phosphate release, reprecipitation of a biological apatite layer, possibly containing endogenous proteins and other factors, and changes in surface topography. In the present study we have shown that in a medium that initially did not contain calcium or phosphate significantly more calcium and phosphate was released from the HA pillars than from 3DF-HA composites, confirming the better accessibility of the ceramic surface in the assembled than in classical monolithic composites. Although the release experiments were performed on HA pillars, and not on assembled composites, with the goal of determining the maximum numbers of ions that could be released, the accessibility of the ceramic surface in an assembled composite with 15 wt.% HA, in which a part of the surface of the ceramic pillars is ‘‘covered’’ by the surrounding polymer, is still expected to be higher than in 3DF-HA composite scaffolds. In contrast to release in SPS, 21 days immersion in SBF did not result in significant differences in calcium or phosphate concentration among different materials, suggesting that in an environment that is saturated with HA the ion release dynamics are different. The question remains, however, which medium is most relevant in mimicking the in vivo situation. Although SBF better represents the mineral composition of human blood plasma than SPS and other often used release media, such as PBS, the fluid replenishment regimes in vivo cannot be compared with the static release conditions in vitro. The presence and possible effects of endogenous biomolecules in vivo also cannot be ignored. The method used to measure calcium and phosphate concentrations in the present study could not be reliably applied to cell culture medium. Nevertheless, we did not observe mineral deposits on any of the scaffolds by SEM analysis (Fig. 5) or in methylene
5716
A. Nandakumar et al. / Acta Biomaterialia 9 (2013) 5708–5717
blue/basic fuchsin stained histological sections (data not shown), even after 21 days culture, in contrast to the scaffolds immersed in SBF, emphasizing the importance of medium composition when discussing release dynamics and associated events. For these reasons it is difficult to relate cell proliferation and osteogenic differentiation to the data on ion release and mineral deposition, although control over the release dynamics was one of the rationales for developing the assembled composites as an alternative to classical monolithic ones. Cells attached and proliferated on all scaffold types in both basic and mineralization medium. An increase in metabolic activity between days 11 and 21 was observed on all scaffolds, although less pronounced on the 3DF-HA composite compared with the polymer control and assembled composite. While no differences in metabolic activity were observed among the three scaffold types after 11 days culture, after 21 days the values for 3DF-HA were lower than the values for 3DF and HA assembled by 3DF. These data for metabolic activity correlated with the trend seen in the amount of DNA after 21 days, with the PLA–HA composite having lower amounts than both the 3DF control and assembled composite, although these differences were not statistically significant. In the classical 3DF-HA composite the presence of homogeneously distributed ceramic particles will change the properties of the surface throughout the scaffold, including the local concentration of calcium and phosphate ions, hydrophilicity and surface roughness, all of which could affect cell attachment and proliferation. Furthermore, in comparison with assembled composites, it is plausible that the release of calcium and phosphate ions from the ceramic component will occur later, which may explain why differences in metabolic activity were only significant after 21 days. In a study by Wilmowsky et al. [48] on different polyetheretherketone-based composites fabricated by SLS it was observed that human foetal osteoblasts cultured on composites containing 10 wt.% b-TCP showed the least pronounced proliferation and viability. They suggested that calcium phosphate compounds around the chosen concentration do not have a favourable effect on osteoblast proliferation in vitro. Zhou et al. [49] showed more pronounced attachment of human alveolar osteoblasts on PCL–TCP RP composites compared with PCL scaffolds and attributed this observation to increased hydrophilicity of the composite scaffolds, along with the ability of TCP to adsorb more proteins. In the assembled composites two distinct phases can be distinguished and therefore two different surface types are available for cells to attach to and proliferate on, which may explain why the effect of the ceramic was smaller in the assembled than in the monolithic composite. Among the four markers of osteogenic differentiation analysed a positive effect of the scaffolds containing ceramic was only observed on the expression of OP. OP, a phosphorylated acidic glycoprotein, is synthesized by osteoblasts in vitro and is found in mineralized tissue. OP regulates the attachment of osteoblasts and osteoclasts to the ECM [50,51]. Assembled composites in basic and mineralization medium showed the highest level of OP expression (>5-fold) amongst the scaffolds tested. Previous studies using sintered calcium phosphate ceramics with different degradation properties [52] and calcium phosphate-coated 3-D scaffolds prepared from PA (unpublished results) also showed high expression of OP compared with tissue culture plastic and uncoated scaffolds, respectively, suggesting that the presence of calcium phosphates enhances OP gene expression. Using MC3T3-E1 cells Beck et al. [50] observed that OP gene expression was induced by the presence of phosphate, which is a result of the hydrolysis of b-glycerol phosphate, an additive used in cell culture media to induce mineralization. A significant up-regulation of OP expression in hPDCs was also observed after addition of calcium, phosphate or both to the medium [53]. However, as discussed before, although plausible, whether the up-regulation of OP observed in our study is
caused by elevated levels of phosphate or the combination of both calcium and phosphate in the assembled composites can only be proven by precise measurement of ionic concentrations in the scaffold vicinity during cell culture. A similar discussion holds for the expression of BMP-2 and BSP, which were shown to be up-regulated by various cell types in the presence of elevated calcium ion levels in a concentration-dependent manner [46,47,54,55], suggesting that the levels in the present study may have been below the necessary threshold to observe an effect. Regarding the down-regulation of ALP expression on assembled composites in basic medium, in a previous study we observed lower ALP levels on polymeric fibres coated with calcium phosphate compared with the uncoated controls [56]. However, ectopic bone formation following subcutaneous implantation in mice only occurred on hMSCs/coated fibre constructs, in contrast to constructs consisting of hMSCs and uncoated fibres, emphasizing the difficulty of assessing the bioactivity of biomaterials in general and calcium phosphate-based ceramics in particular in vitro. Although, clearly, additional studies are needed to understand the biological mechanisms by which the scaffolds tested in this study can affect processes related to bone regeneration, and to further explore the possibilities they offer to control these processes, we have presented an alternative way of assembling polymer/ceramic composites for bone regeneration. By keeping the two material types as separate components of the composite material, the versatility of RP can be employed to a greater extent, offering more freedom in the design of scaffold properties than is the case with conventional composites. 5. Conclusions Two different strategies for the fabrication of 3-D polymer– ceramic composite scaffolds for bone regeneration using RP have been studied. While the conventional approach, in which the ceramic particles were dispersed within the polymeric component, resulted in a mechanically stiffer construct, assembled composites, consisting of distinct polymer and ceramic phases, allowed the incorporation of larger amounts of ceramic. The two methods of producing composite scaffolds further expand the possibilities for controlling exposure of the individual components to the biological environment. Acknowledgement Ms. Anita Podt is acknowledged for preparation of the molds by stereolithography. Appendix A. Figures with essential colour discrimination Certain figures in this article, particularly Figs. 3, 6 and 7, 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.2012.10.044. Appendix B. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.actbio.2012. 10.044. References [1] Augat P, Schorlemmer S. The role of cortical bone and its microstructure in bone strength. Age Ageing 2006;35(Suppl. 2):ii27–31.
A. Nandakumar et al. / Acta Biomaterialia 9 (2013) 5708–5717 [2] Fratzl P, Gupta HS, Paschalis EP, Roschger P. Structure and mechanical quality of the collagen-mineral nano-composite in bone. J Mater Chem 2004;14:2115–23. [3] Zaidi M. Skeletal remodeling in health and disease. Nat Med 2007;13:791–801. [4] Erlebacher A, Filvaroff EH, Gitelman SE, Derynck R. Toward a molecular understanding of skeletal development. Cell 1995;80:371–8. [5] Laurencin C, Khan Y, El-Amin SF. Bone graft substitutes. Expert Rev Med Devices 2006;3:49–57. [6] Kainer MA, Linden JV, Whaley DN, Holmes HT, Jarvis WR, Jernigan DB, et al. Clostridium infections associated with musculoskeletal-tissue allografts. New Engl J Med 2004;350:2564–71. [7] Mroz TE, Joyce MJ, Steinmetz MP, Lieberman IH, Wang JC. Musculoskeletal allograft risks and recalls in the United States. J Am Acad Orthop Surg 2008;16:559–65. [8] Damien CJ, Parsons JR. Bone graft and bone graft substitutes: a review of current technology and applications. J Appl Biomater 1991;2:187–208. [9] Szpalski C, Wetterau M, Barr J, Warren SM. Bone tissue engineering: current strategies and techniques - Part I: Scaffold. Tissue Eng B Rev 2012;18:246–57. [10] Zhang P, Hong Z, Yu T, Chen X, Jing X. In vivo mineralization and osteogenesis of nanocomposite scaffold of poly(lactide-co-glycolide) and hydroxyapatite surface-grafted with poly(L-lactide). Biomaterials 2009;30:58–70. [11] Kim HW, Lee HH, Knowles JC. Electrospinning biomedical nanocomposite fibers of hydroxyapatite/poly(lactic acid) for bone regeneration. J Biomed Mater Res A 2006;79:643–9. [12] Montjovent MO, Mathieu L, Hinz B, Applegate LL, Bourban PE, Zambelli PY, et al. Biocompatibility of bioresorbable poly(L-lactic acid) composite scaffolds obtained by supercritical gas foaming with human fetal bone cells. Tissue Eng 2005;11:1640–9. [13] Yefang Z, Hutmacher DW, Varawan SL, Meng LT. Comparison of human alveolar osteoblasts cultured on polymer–ceramic composite scaffolds and tissue culture plates. Int J Oral Max Surg 2007;36:137–45. [14] Yang F, Both SK, Yang X, Walboomers XF, Jansen JA. Development of an electrospun nano-apatite/PCL composite membrane for GTR/GBR application. Acta Biomater 2009;5:3295–304. [15] Aboudzadeh N, Imani M, Shokrgozar MA, Khavandi A, Javadpour J, Shafieyan Y, et al. Fabrication and characterization of poly(D,L-lactide-co-glycolide)/ hydroxyapatite nanocomposite scaffolds for bone tissue regeneration. J Biomed Mater Res A 2010;94:137–45. [16] Jose MV, Thomas V, Johnson KT, Dean DR, Nyairo E. Aligned PLGA/HA nanofibrous nanocomposite scaffolds for bone tissue engineering. Acta Biomater 2009;5:305–15. [17] Song JH, Kim HE, Kim HW. Electrospun fibrous web of collagen–apatite precipitated nanocomposite for bone regeneration. J Mater Sci Mater Med 2008;19:2925–32. [18] Thomas V, Dean DR, Jose MV, Mathew B, Chowdhury S, Vohra YK. Nanostructured biocomposite scaffolds based on collagen coelectrospun with nanohydroxyapatite. Biomacromolecules 2007;8:631–7. [19] Laschke MW, Strohe A, Menger MD, Alini M, Eglin D. In vitro and in vivo evaluation of a novel nanosize hydroxyapatite particles/poly(ester-urethane) composite scaffold for bone tissue engineering. Acta Biomater 2010;6:2020–7. [20] Huang YX, Ren J, Chen C, Ren TB, Zhou XY. Preparation and properties of poly(lactide-co-glycolide) (PLGA)/ nano-hydroxyapatite (NHA) scaffolds by thermally induced phase separation and rabbit MSCs culture on scaffolds. J Biomater Appl 2008;22:409–32. [21] Peltola SM, Melchels FP, Grijpma DW, Kellomaki M. A review of rapid prototyping techniques for tissue engineering purposes. Ann Med 2008;40:268–80. [22] Duan B, Wang M, Zhou WY, Cheung WL, Li ZY, Lu WW. Three-dimensional nanocomposite scaffolds fabricated via selective laser sintering for bone tissue engineering. Acta Biomater 2010;6:4495–505. [23] Tan KH, Chua CK, Leong KF, Cheah CM, Cheang P, Abu Bakar MS, et al. Scaffold development using selective laser sintering of polyetheretherketone– hydroxyapatite biocomposite blends. Biomaterials 2003;24:3115–23. [24] Shor L, Guceri S, Wen XJ, Gandhi M, Sun W. Fabrication of three-dimensional polycaprolactone/hydroxyapatite tissue scaffolds and osteoblast–scaffold interactions in vitro. Biomaterials 2007;28:5291–7. [25] Moroni L, Hamann D, Paoluzzi L, Pieper J, de Wijn JR, van Blitterswijk CA. Regenerating articular tissue by converging technologies. PLoS One 2008;3:e3032. [26] Woodfield TBF, Malda J, de Wijn J, Peters F, Riesle J, van Blitterswijk CA. Design of porous scaffolds for cartilage tissue engineering using a three-dimensional fiber-deposition technique. Biomaterials 2004;25:4149–61. [27] Barbieri D, Renard AJ, de Bruijn JD, Yuan H. Heterotopic bone formation by nano-apatite containing poly(D,L-lactide) composites. Eur Cell Mater 2010;19:252–61. [28] Geesink RG, de Groot K, Klein CP. Bonding of bone to apatite-coated implants. J Bone Joint Surg Br 1988;70:17–22. [29] Hanawa T, Kamiura Y, Yamamoto S, Kohgo T, Amemiya A, Ukai H, et al. Early bone formation around calcium-ion-implanted titanium inserted into rat tibia. J Biomed Mater Res A 1997;36:131–6.
5717
[30] Barradas AM, Yuan H, van Blitterswijk CA, Habibovic P. Osteoinductive biomaterials: current knowledge of properties, experimental models and biological mechanisms. Eur Cell Mater 2011;21:407–29. [31] Du C, Meijer GJ, van de Valk C, Haan RE, Bezemer JM, Hesseling SC, et al. Bone growth in biomimetic apatite coated porous Polyactive((R)) 1000PEGT70PBT30 implants. Biomaterials 2002;23:4649–56. [32] Nandakumar A, Fernandes H, de Boer J, Moroni L, Habibovic P, van Blitterswijk CA. Fabrication of bioactive composite scaffolds by electrospinning for bone regeneration. Macromol Biosci 2010;10:1365–73. [33] Moroni L, de Wijn JR, van Blitterswijk CA. 3D fiber-deposited scaffolds for tissue engineering: influence of pores geometry and architecture on dynamic mechanical properties. Biomaterials 2006;27:974–85. [34] Habibovic P, Yuan H, van der Valk CM, Meijer G, van Blitterswijk CA, de Groot K. 3D microenvironment as essential element for osteoinduction by biomaterials. Biomaterials 2005;26:3565–75. [35] Landers R, Pfister A, Hubner U, John H, Schmelzeisen R, Mulhaupt R. Fabrication of soft tissue engineering scaffolds by means of rapid prototyping techniques. J Mater Sci 2002;37:3107–16. [36] Bohner M, Lemaitre J. Can bioactivity be tested in vitro with SBF solution? Biomaterials 2009;30:2175–9. [37] Both SK, Van der Muijsenberg AJC, Van Blitterswijk CA, De Boer J, De Bruijn JD. A rapid and efficient method for expansion of human mesenchymal stem cells. Tissue Eng 2007;13:3–9. [38] Fernandes H, Mentink A, Bank R, Stoop R, van Blitterswijk C, de Boer J. Endogenous collagen influences differentiation of human multipotent mesenchymal stromal cells. Tissue Eng A 2010;16:1693–702. [39] Livak KJ, Schmittgen TD. Analysis of relative gene expression data using realtime quantitative PCR and the 2(-Delta Delta C(T)) method. Methods 2001;25:402–8. [40] Lam CXF, Teoh SH, Hutmacher DW. Comparison of the degradation of polycaprolactone and polycaprolactone-(beta-tricalcium phosphate) scaffolds in alkaline medium. Polym Int 2007;56:718–28. [41] Cao H, Kuboyama N. A biodegradable porous composite scaffold of PGA/betaTCP for bone tissue engineering. Bone 2010;46:386–95. [42] 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. [43] Athanasiou KA, Zhu CF, Lanctot DR, Agrawal CM, Wang X. Fundamentals of biomechanics in tissue engineering of bone. Tissue Eng 2000;6:361–81. [44] Ducheyne P, Beight J, Cuckler J, Evans B, Radin S. Effect of calcium phosphate coating characteristics on early post-operative bone tissue ingrowth. Biomaterials 1990;11:531–40. [45] Habibovic P, Barralet JE. Bioinorganics and biomaterials: bone repair. Acta Biomater 2011;7:3013–26. [46] Barradas AM, Fernandes HA, Groen N, Chai YC, Schrooten J, van de Peppel J, et al. A calcium-induced signaling cascade leading to osteogenic differentiation of human bone marrow-derived mesenchymal stromal cells. Biomaterials 2012;33:3205–15. [47] Chai YC, Roberts SJ, Schrooten J, Luyten FP. Probing the osteoinductive effect of calcium phosphate by using an in vitro biomimetic model. Tissue Eng A 2011;17:1083–97. [48] von Wilmowsky C, Vairaktaris E, Pohle D, Rechtenwald T, Lutz R, Munstedt H, et al. Effects of bioactive glass and beta-TCP containing three-dimensional laser sintered polyetheretherketone composites on osteoblasts in vitro. J Biomed Mater Res A 2008;87A:896–902. [49] Zhou YF, Hutmacher DW, Varawan SL, Lim TM. In vitro bone engineering based on polycaprolactone and polycaprolactone–tricalcium phosphate composites. Polym Int 2007;56:333–42. [50] Beck GR, Zerler B, Moran E. Phosphate is a specific signal for induction of osteopontin gene expression. Proc Natl Acad Sci USA 2000;97:8352–7. [51] Roach HI. Why does bone-matrix contain noncollagenous proteins – the possible roles of osteocalcin, osteonectin, osteopontin and bone sialoprotein in bone mineralization and resorption. Cell Biol Int 1994;18:617–28. [52] Yuan H, Fernandes H, Habibovic P, de Boer J, Barradas AM, de Ruiter A, et al. Osteoinductive ceramics as a synthetic alternative to autologous bone grafting. Proc Natl Acad Sci USA 2010;107:13614–9. [53] Chai YC, Roberts SJ, Desmet E, Kerckhofs G, van Gastel N, Geris L, et al. Mechanisms of ectopic bone formation by human osteoprogenitor cells on CaP biomaterial carriers. Biomaterials 2012;33:3127–42. [54] Tada H, Nemoto E, Kanaya S, Hamaji N, Sato H, Shimauchi H. Elevated extracellular calcium increases expression of bone morphogenetic protein-2 gene via a calcium channel and ERK pathway in human dental pulp cells. Biochem Biophys Res Commun 2010;394:1093–7. [55] Honda Y, Anada T, Kamakura S, Nakamura M, Sugawara S, Suzuki O. Elevated extracellular calcium stimulates secretion of bone morphogenetic protein 2 by a macrophage cell line. Biochem Biophys Res Commun 2006;345:1155–60. [56] Nandakumar A, Yang L, Habibovic P, van Blitterswijk C. Calcium phosphate coated electrospun fiber matrices as scaffolds for bone tissue engineering. Langmuir 2010;26:7380–7.
Contents lists available at SciVerse ScienceDirect
Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat
Monolithic and assembled polymer–ceramic composites for bone regeneration Anandkumar Nandakumar a, Célia Cruz a,b, Anouk Mentink a, Zeinab Tahmasebi Birgani a, Lorenzo Moroni a, Clemens van Blitterswijk a, Pamela Habibovic a,⇑ a b
Department of Tissue Regeneration, MIRA Institute for Biomedical Technology and Technical Medicine, University of Twente, The Netherlands Faculty of Engineering, University of Porto, Rua Dr. Roberto Frias, Porto, Portugal
a r t i c l e
i n f o
Article history: Received 31 July 2012 Received in revised form 23 October 2012 Accepted 30 October 2012 Available online 7 November 2012 Keywords: Polymer/ceramic composite Rapid prototyping Conventional Assembled Bone regeneration
a b s t r a c t The rationale for the use of polymer–ceramic composites for bone regeneration stems from the natural composition of bone, with collagen type I and biological apatite as the main organic and inorganic constituents, respectively. In the present study composite materials of PolyActive™ (PA), a poly(ethylene oxide terephthalate)/poly(butylene terephtalate) co-polymer, and hydroxyapatite (HA) at a weight ratio of 85:15 were prepared by rapid prototyping (RP) using two routes. In the first approach pre-extruded composite filaments of PA–HA were processed using three-dimensional fibre deposition (3DF) (conventional composite scaffolds). In the second approach PA scaffolds were fabricated using 3DF and combined with HA pillars produced inside stereolithographic moulds that fitted inside the pores of the PA threedimensional structure (assembled composite scaffolds). Analysis of calcium and phosphate release in a simulated physiological solution, not containing calcium or phosphate, revealed significantly higher values for the HA pillars compared with other scaffolds. Release in simulated body fluid saturated with respect to HA did not show significant differences among the different scaffolds. Human mesenchymal stromal cells were cultured on polymer (3DF), conventional composite (3DF-HA) and assembled composite (HA assembled in 3DF) scaffolds and assessed for morphology, metabolic activity, DNA amount and gene expression of osteogenic markers using real time quantitative polymerase chain reaction (PCR). Scanning electron microscopy images showed that the cells attached to and infiltrated all the scaffolds. Assembled composites had a higher metabolic activity compared with 3DF-HA scaffolds while no significant differences were observed in DNA amounts. Gene expression of osteopontin in the assembled composite was significantly higher compared with the conventional composites. The strategy of composite fabrication by assembly appears to be a promising alternative to the conventional composite fabrication route for scaffolds for bone regeneration. Ó 2012 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction Bone is a naturally occurring biocomposite composed of organic and inorganic components and water. The organic part of bone is mainly composed of collagen type I among other proteins and is responsible for the tensile strength and toughness of bone tissue, while the inorganic mineral component is calcium phosphate, essentially hydroxyapatite (HA), that provides stiffness and strength to bone [1,2]. Besides this mechanical role, the mineral part of bone is of great importance in calcium homeostasis [3,4]. While autografts are generally considered the gold standard in bone repair and regeneration, problems with donor site morbidity and availability exist [5], while allografts suffer from a risk of disease transmission [6,7]. To overcome these issues recent developments have focused on alternative regenerative strategies, such as ⇑ Corresponding author. Tel.: +31 53 489 3400. E-mail address: [email protected] (P. Habibovic).
synthetic bone graft substitutes and tissue engineered constructs [8,9]. Whether a biomaterial is used as a stand-alone bone graft substitute or as a scaffold that acts as a carrier of cells and/or growth factors in a tissue engineered construct, a number of requirements regarding mechanical strength, porosity, pore interconnectivity and bioactivity need to be met for successful use in a clinical setting. In order to meet these various requirements combinations of different types of materials and different fabrication processes are often required. While with polymers a wide range of mechanical properties can be achieved, calcium phosphate ceramics suffer from intrinsic brittleness. On the other hand, the bioactivity of calcium phosphates, in terms of osteoconduction, osteoinduction and bone bonding capacity, is in most cases superior to that of polymers, owing to the chemical composition of calcium phosphate ceramics and events occurring on their surface in a physiological environment [8]. With the rationale of recreating the composition of bone, polymer–ceramic composite scaffolds have been explored as potential bone graft substitutes. Synthetic
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.2012.10.044
A. Nandakumar et al. / Acta Biomaterialia 9 (2013) 5708–5717
polymers like poly(L-lactic acid) (PLLA) [10–12], poly(e-caprolactone) (PCL) [13,14], poly(lactic-co-glycolic acid) (PLGA) [10,15,16] and naturally occurring collagen [17,18] have been combined with hydroxyapatite (HA) to fabricate composite materials. Use of conventional fabrication methods like salt leaching–phase inversion [19], thermally induced phase separation (TIPS) [20] and other similar techniques do not offer strict control over the porosity and interconnectivity and invariably use organic solvents. Pore interconnectivity is essential for the supply of nutrients and metabolites while residuals of organic solvents can be cytotoxic [21]. The need to control the architecture, porosity and mechanical properties of scaffolds led to the advent of rapid prototyping (RP) technologies. Several RP techniques, such as selective laser sintering (SLS) [22,23], fused deposition modelling (FDM) [13] and precision extrusion deposition (PED) [24], have been used to fabricate composites by combining polymers like poly(hydroxybutyrate-co-hydroxyvalerate), PLLA, and PCL with different calcium phosphate phases, such as carbonated hydroxyapatite, HA and b-tricalcium phosphate. These techniques also offer the advantage of being integrated to imaging modalities (e.g. magnetic resonance imaging and computed tomography) in order to fabricate – scaffolds and implants customized for a particular patient. In the present study we have used three-dimensional (3-D) fibre deposition (3DF) to fabricate two types of scaffolds consisting of PolyActive™ (PA), a poly(ethylene oxide terephthalate)/ poly(butylene terephtalate) (PEOT/PBT) co-polymer, and an HA ceramic to assemble a monolithic composite. In the first, conventional approach, filaments of polymer ceramic composite at desired weight ratios were extruded and used as the starting material to prepare the scaffolds by 3DF. In the other approach, previously developed by our group [25], the polymer scaffolds were fabricated by 3DF while the ceramic particles were fabricated in the form of pillars by sintering a ceramic slurry poured into negative molds prepared by stereolithography. The two components were then assembled by manually press fitting the HA pillars into the pores of the 3DF scaffold to create a composite. 3DF is an extrusion-based RP technique in which a polymer is melted and extruded through a nozzle and the material deposition is controlled by a robotic arm. This method allows the fabrication of scaffolds with closely controlled porosity, pore size, interconnectivity and fibre orientation between successive layers. 3DF has been used to fabricate polymeric scaffolds for tissue engineering of cartilage [26] and osteochondral [25] defects. While 3DF is a very versatile technique for fabricating polymeric scaffolds, the processing of composites is more difficult owing to the high viscosity of the polymer–ceramic melt resulting in clogging of the nozzles. In addition to affecting fabrication and processability, these phenomena also limit the amount of ceramic that can be incorporated into the scaffold, while the amount of ceramic largely determines the osteoconductivity and osteoinductivity of composite materials. For example, Barbieri et al. recently demonstrated osteoinductivity of a PLLA/HA composite, but only when the ceramic component was as high as 40 wt.% [27]. Furthermore, in a conventional monolithic composite the majority of the ceramic particles are covered by polymer, while accessibility of the ceramic surface to the physiological environment, which is associated with dissolution/reprecipitation processes, is believed to be the origin of the bioactivity of calcium phosphates [28–30]. Our study aimed at increasing the amount of ceramic component in the composite and improving the exposure of its surface, while employing 3DF as a RP technique to closely control the scaffold architecture. hMSCs were seeded on the monolithic and assembled composite scaffolds, as well as on a polymer control, and cell proliferation and osteogenic differentiation were studied.
5709
2. Materials and methods PA, a PEOT/PBT co-polymer with tunable physico-chemical properties that has previously been used in different forms for tissue engineering applications [26,31,32], was provided by PolyVation BV (The Netherlands). Following the aPEOTbPBTc nomenclature the composition used in this study was 300PEOT55PBT45, where a is the molecular weight in g mol1 of the starting poly(ethylene glycol) (PEG) blocks used in the co-polymerization, while b and c are the weight ratios of the PEOT and PBT blocks, respectively. The HA powder for ceramic preparation was provided by Progentix BV (The Netherlands). It was synthesized using a wet precipitation method. The powder was not sintered and was phase pure HA with a particle size below 63 lm. 2.1. Production of composite filaments and fabrication of 3DF scaffolds by RP Composite filaments of PA and HA were extruded using a screw extruder (Artecs). Briefly, the desired amounts of PA pellets and HA powder were added to the extruder and mixed for 5 min at a temperature of 180 °C. The composite filaments were then drawn out and cut into small pellets for processing by RP. For scaffold fabrication composite filaments with a maximum of 15% HA by weight were used. 3-D scaffolds were fabricated by PR using a Bioplotter (Envisiontec GmbH, Germany), which is an xyz robotic device as previously described [33]. Briefly, polymer or composite granules (PA– HA) were loaded onto a stainless steel syringe set at a temperature between 195 and 210 °C (higher temperatures for composites) using a thermostat controlled cartridge unit, fixed on the x mobile arm of the apparatus. A pressure of 5 bars (nitrogen) for composites (4 bars for polymer) was applied to the syringe through a pressurized cap once the polymer had melted. Rectangular block models were loaded in the Bioplotter CAM (PrimCAM, Switzerland) software and deposited layer by layer as the polymer was extruded through a nozzle (0.7 mm OD) onto a stage. The deposition speed was varied between 150 and 250 mm min1 (lower speeds for composites). The spacing between fibres in the same layer was set at 0.8 mm, the layer thickness was 0.225 mm and the scaffold height was set at 4 mm. A 0-90° configuration was used for the scaffold architecture where fibres were deposited with 90° orientation steps between successive layers. 2.2. Preparation of the ceramic slurry and moulds and assembly of HA into 3DF scaffolds Pillars of HA were fabricated in a two step process. First, a negative mask was designed using Rhinoceros and fabricated by stereolithography (Envisiontec, Germany). The masks were then filled with an HA slurry prepared by vigorously mixing 8.2 g of calcined HA (20 h in an oven at 1000 °C), 3.55 g of non-calcined HA, 5 g of demineralized water, 0.3 g of methylcellulose solution (2% w/v methylcellulose in demineralized water), 0.55 g of ammonia, and 3.525 g of sieved paraffin particles (600–1000 lm), resulting in 30% porosity of the HA particles. The HA pillars were obtained by sintering in a furnace (Nabertherm, Germany) at 1150 °C, resulting in a fully crystalline phase pure HA ceramic [34]. The HA particles were sterilized by autoclaving and the pillars were press fitted into the pores of the 3DF scaffolds. 2.3. Characterization of the scaffolds The architecture and composition of the different types of scaffolds were characterized by scanning electron microscopy (SEM)
5710
A. Nandakumar et al. / Acta Biomaterialia 9 (2013) 5708–5717
(XL30 ESEM-FEG, Philips). The porosity of the RP scaffolds was calculated from SEM measurements of fibre diameter, layer thickness and fibre spacing using the formula [35]:
Porosity ¼ 1
p 4
1
1
Fibre spacing Layer thickness Fibre diameter Fibre diameter
Compression tests were performed on scaffolds with a diameter of 4 mm and a length of 4 mm using a mechanical test apparatus (Zwick Z050, Germany) to obtain the modulus of elasticity under compression. A 0.5 kN load cell was used and the samples were compressed at a rate of 1 mm min1 up to a maximum of 30% strain. The modulus was calculated from the stress–strain curves as the slope of the initial linear region (5–8% strain) neglecting the toe region due to initial settling of the specimen. A total of four samples per scaffold group were evaluated. A representative stress–strain curve for each scaffold type is shown in Supplementary Fig. 1. 2.4. Calcium and phosphate release Release of calcium and phosphate was studied in a simulated physiological solution (SPS) containing 137 mM Na+, 177 mM Cl and 50 mM HEPES (N-[2-hydroxyethyl]piperazine-N0 -(2-ethanesulfonic acid)) buffered at pH 7.4 and in a simplified simulated body fluid (SBF JL-2) containing 142 mM Na+, 109.90 mM Cl, 2+ 34.88 mM HCO and 1.39 mM HPO2 3 , 2.31 mM Ca 4 , pH 7.4, according to Bohner and Lemaitre [36]. Sodium azide was used as an antimicrobial agent. Triplicates of each sample (PA, 3DF-HA and HA pillars) weighing 250 mg each were soaked in 30 ml of the solution at 37 °C under shaking for 3 weeks, after which samples were removed and the concentrations of calcium and phosphate measured using QuantiChrom™ Calcium Assay (DICA-500) and QuantiChrom™ Phosphate Assay (DIPA-500) kits. The optical density of samples after adding reagents from the kits was read with a microplate spectrophotometer (Thermo Scientific Multiskan GO) at 612 and 620 nm for calcium and phosphate release, respectively. HA pillars were used in the release study in order to determine the maximum amounts of calcium and phosphate ions that could be released. After immersion in SBF for 21 days one sample of each type was used for SEM analysis. 2.5. Human mesenchymal stromal cell (hMSC) isolation and culture Bone marrow aspirates were obtained after written informed consent and hMSCs were isolated and proliferated as described previously [37,38]. Briefly, aspirates were resuspended using 20 gage needles, plated at a density of 5 105 cells cm2 and cultured in hMSC proliferation medium containing a-minimal essential medium (Life Technologies), 10% foetal bovine serum (Lonza), 0.2 mM ascorbic acid (Life Technologies), 2 mM L-glutamine (Life Technologies), 100 units ml1 penicillin (Life Technologies), 10 lg ml1 streptomycin (Life Technologies), and 1 ng ml1 basic fibroblast growth factor (FGF) (Instruchemie). Cells were grown at 37 °C in a humid atmosphere with 5% CO2. The medium was refreshed twice a week, and cells were used for further sub-culturing or cryopreservation. The hMSC basic medium was composed of hMSC proliferation medium without basic FGF. Mineralization medium was composed of hMSC basic medium supplemented with 108 M dexamethasone (Sigma, St. Louis, MO) and 0.01 M b-glycerophosphate (BGP) (Sigma). Cells were trypsinized prior to seeding on scaffolds. The in vitro experiments were performed with cells from one donor at passage 3. Scaffolds (10 10 4 mm3) were soaked in 70% ethanol for 30 min and dried overnight in a laminar flow cabinet. The scaffolds were washed twice with sterile phosphate-buffered saline (PBS), transferred to a 25-well non-treated polystyrene
plate (Greiner Bio One) and incubated at 37 °C in a humid atmosphere with 5% CO2 for 4 h in basic cell culture medium. After removing the medium each scaffold was seeded with 800,000 cells in approximately 105 ll of basic medium. The cell–scaffold constructs were incubated for 3 h to allow cell attachment and topped up to 2 ml with the appropriate medium. Scaffolds were cultured in either basic medium or mineralization medium and the medium was refreshed twice a week. As a control T25 flasks were seeded at 5000 cells cm2 in basic and mineralization medium. 2.6. Cell attachment, metabolism and amount of DNA Cell metabolism was assessed using the alamarBlueÒ assay according to the manufacturer’s protocol. Briefly, culture medium was replaced with medium containing 10 vol.% alamarBlueÒ solution (Biosource, Camarillo, CA) and the cell–scaffold constructs (n = 3) were incubated at 37 °C for 4 h. Fluorescence was measured at 590 nm in a Perkin Elmer Victor3 1420 Multilabel plate reader. Cell metabolism was analysed on days 11 and 21 and the readout from the scaffolds was corrected with a blank from each group. After 21 days culture seeded scaffolds (n = 3) were washed with PBS and frozen at 80 °C overnight. The constructs were then digested at 56 °C in a Tris–EDTA buffered solution containing 1 mg ml1 proteinase K, 18.5 lg ml1 pepstatin A and 1 lg ml1 iodoacetamide (Sigma–Aldrich) for 18 h. DNA amounts were determined using the CyQuantÒ DNA quantification kit (Invitrogen) with 50 ll of cell lysate according to the manufacturer’s protocol. Fluorescence at an excitation wavelength of 480 nm and an emission wavelength of 520 nm was measured using a Perkin Elmer Victor3 1420 Multilabel plate reader. All data for metabolic activity and DNA assay were normalized to the dry weight of the corresponding scaffold. 2.7. Cell morphology After 21 days culture SEM analysis was performed to assess the cell response to different scaffolds. The medium was removed and the scaffolds washed twice with PBS and fixed in 10% formalin for 1 h. After rinsing with PBS the scaffolds were dehydrated in a series of increasing ethanol concentrations (70%, 80%, 90%, 96%, 100% 2), 15 min in each concentration, before drying in a critical point dryer (Balzers CPD-030). The samples were then sputter coated with gold (Cressington) for observation by SEM. 2.8. Cell differentiation To analyse the expression of osteogenic markers by hMSCs total RNA was isolated using a combination of the TRIzolÒ method with the NucleoSpinÒ RNA II isolation kit (Bioké). Briefly, scaffolds (n = 3) were washed once with PBS and 1 ml of TRIzol reagent (Invitrogen) was added to the samples. After 5 min the samples were stored at 80 °C for RNA isolation. After chloroform addition and phase separation by centrifugation the aqueous phase containing the RNA was collected, mixed with an equal volume of 75% ethanol and loaded onto the RNA binding column of the kit. Subsequent steps were in accordance with the manufacturer’s protocol. RNA was collected in RNase-free water. The quality and quantity of RNA was analysed by gel electrophoresis and spectrophotometry. 750 ng of RNA were used for first strand cDNA synthesis using iScript (Bio-Rad) according to the manufacturer’s protocol. 1 ll of undiluted cDNA was used for subsequent analysis. PCR was performed in an iQ5 real time PCR machine (Bio-Rad) using SYBR Green supermix (Bio-Rad). Expression of osteogenic marker genes was normalized to glyceraldehyde-3-phosphate dehydrogenase (50 -CGCTCTCTGCTCCTCCTGTT-30 and 50 CCATGGTGTCTGAGCGATGT-30 ) levels and fold induction was
A. Nandakumar et al. / Acta Biomaterialia 9 (2013) 5708–5717
calculated using the comparative DCt method [39]. The following primer sequences were used for the osteogenic marker genes: bone sialoprotein (BSP) (50 -TGCCTTGAGCCTGCTTCC-30 and 50 -CAAAATTAAAGCAGTCTTCATTTTG-30 ); osteopontin (OP) (50 -CCAAGTAAGTCCAACGAAAG-30 and 50 -GGTGATGTCCTCGTCTGTA-30 ); alkaline phosphatase (ALP) (50 -GACCCTTGACCCCCACAAT-30 and 50 -GCTCGTACTGCATGTCCCCT-30 ). For amplification of bone morphogenetic protein-2 (BMP-2) a gene-specific primer mix was used (SA Biosciences) according to the manufacturer’s protocol.
2.9. Statistical analysis One-way ANOVA with Tukey’s multiple comparison post hoc test was performed. The level of significance was set at 0.05. All data presented are expressed as means ± standard deviation. Only significant differences between different scaffold groups in the same medium or the same scaffold type in different media are shown.
3. Results 3.1. Fabrication and characterization of scaffolds RP using 3DF was shown to be a suitable technique for fabricating polymeric and composite 3-D scaffolds. An SEM image of rapid prototyped 3DF and 3DF-HA scaffolds in top view and cross-section is presented in Fig. 1. Strands of the scaffolds appeared to decrease in thickness while being deposited, possibly due to the large fibre spacing. The fibre diameter, layer thickness and fibre spacing were calculated to be 0.315 ± 0.1, 0.221 ± 0.02 and 1.96 ± 0.13 mm, and 0.34 ± 0.08, 0.238 ± 0.03 and 1.84 ± 0.03 mm for the 3DF and 3DF-HA scaffolds, respectively. The porosities of the 3DF and 3DF-HA scaffolds were determined to be 81.9% and 79.3%, respectively.
5711
The assembled composites were fabricated by inserting the ceramic pillars into the 3DF scaffolds. Fig. 2A shows optical images of various configurations in which HA pillars can be assembled into 3DF scaffolds. For the experiments the weight of the incorporated HA pillars was kept to 15% of the total weight (a configuration similar to the image on the left in Fig. 2A) of the construct to ensure that equal amounts of HA were incorporated into the assembled and conventional composite scaffolds. Fig. 2B shows an SEM micrograph of the assembled composite with a ceramic pillar in the middle. The structure of the HA particles is shown in Fig. 2C and D. The grain size of the HA ceramic was between 0.8 and 1.5 lm, as measured on the SEM images. The results of the mechanical tests are shown in Fig. 3. The incorporation of HA in the scaffold bulk had an impact on the scaffold stiffness. The modulus of elasticity under compression of the 3DF-HA scaffolds (39 ± 15 MPa) was significantly higher than that of the other two scaffold groups. The addition of HA as pillars increased the modulus of the assembled construct (9 ± 4.5 MPa) compared with the 3DF scaffold (8.1 ± 2.4 MPa), but this increase was minimal. Release of calcium and phosphate was measured in two physiological solutions after 21 days immersion, SPS, with an ionic strength similar to that of human blood plasma but without calcium or phosphate, and SBF saturated with respect to HA, (Table 1). In SPS calcium was only detected in the solutions containing HA pillars. The phosphate concentration of the SPS containing HA pillars was about 10 times higher than that in which the 3DF composite was immersed, while only background levels were found in the 3DF controls. Immersion of all samples in SBF for 21 days resulted in a milky solution with precipitates visible. The same was the case for the SBF solution alone. Concentrations were measured after the precipitate was filtered through a 0.45 lm filter, and no significant differences were observed among different scaffolds in terms of either calcium or phosphate concentration (Table 1). SEM analysis revealed the deposition of globular precipitates on all scaffold types immersed in SBF, although their amount seemed to be great-
Fig. 1. Conventional rapid prototyping (RP) for fabrication of polymeric and composite scaffolds. (A) SEM image of a 3DF scaffold fabricated by RP. (B) Cross-section of a 3DF scaffold. (C) SEM image of a monolithic polymer–ceramic composite (3DF-HA). (D) Cross-section of a 3DF-HA composite. Scale bars: (A and C) 2 mm; (B and D) 1 mm.
5712
A. Nandakumar et al. / Acta Biomaterialia 9 (2013) 5708–5717
Fig. 2. Assembled polymer–ceramic composite fabricated using RP and stereolithography. (A) Photographs of assembled composites with different amounts of HA, fabricated using a negative stereolithography mould, loaded on a 3DF polymer scaffold fabricated by RP (scale bar 5 mm). (B) SEM image of an assembled scaffold with an HA pillar in one of the pores (scale bar 1 mm). (C and D) Microstructure of HA at different magnifications (scale bars 50 and 10 lm, respectively).
Fig. 3. Modulus of elasticity under compression. The modulus of 3DF-HA scaffolds was significantly higher (P < 0.01) compared with the 3DF and assembled composites.
Table 1 Calcium and phosphate concentrations in SPS and SBF after 21 days immersion of 3DF, 3DF-HA and HA pillars.
3DF 3DF-HA HA pillars
Concentration in SPS (lM)
Concentration in SBF (lM)
Ca2+
PO3 4
Ca2+
PO3 4
0 0 30.2 ± 9.4
0 6.6 ± 0.6 95.1 ± 23.6
126.0 ± 27.0 131.0 ± 49.3 156.1 ± 10.8
344.9 ± 25.9 422.6 ± 39.1 389.7 ± 31.6
er on 3DF-HA and HA pillars compared with 3DF control samples (Fig. 4A–C). Further, deposits on 3DF-HA composites and HA pillars consisted of sharp crystals that were oriented perpendicular to the scaffold surface, while the deposits found on the 3DF samples had a smoother morphology (Fig. 4D–F). Energy-dispersive X-ray analysis demonstrated the presence of calcium and phosphate in all precipitates (data not shown).
3.2. Cell morphology, metabolic activity and proliferation Fig. 5 shows SEM images of scaffolds after 21 days cell culture. Qualitative SEM analyses suggested a lower number of cells in the 3DF-HA scaffolds compared with the other two groups. Cells were observed throughout the scaffolds, confirming their infiltration into the pores, and they were also observed on the HA pillars in the assembled composite. Extracellular matrix (ECM) production was observed on all scaffold types. alamarBlueÒ assays were performed on days 11 and 21 to assess cell metabolism in the various scaffolds (Fig. 6A). On day 11 no significant differences in cell metabolism were observed among the different scaffold groups, irrespective of the culture medium. On day 21 in basic medium cells cultured on the 3DF-HA composites had a significantly lower metabolic activity compared with the 3DF and assembled composites. The same trend was observed in mineralization medium, although statistically significant differences were only observed between the 3DF and 3DF-HA groups.
A. Nandakumar et al. / Acta Biomaterialia 9 (2013) 5708–5717
5713
Fig. 4. SEM images showing mineral formation upon immersion in SBF for 21 days: (A and D) 3DF; (B and E) 3DF-HA composite; (C and F) HA pillars. Scale bars: (A–C) 200 lm; (D–F) 5 lm.
Fig. 5. SEM images of cells on (A and D) 3DF, (B and E) 3DF + HA and (C and F) assembled scaffolds after 21 days culture in (A–C) basic and (D–F) mineralization medium. (Inset C) Cells on a HA pillar inside the assembled construct. Scale bar 500 lm. (Inset in F) hMSCs along with the ECM produced. Scale bar 10 lm.
A quantitative biochemical assay to determine the amount of DNA in different scaffolds was performed after 21 days as an indirect measure of cell numbers (Fig. 6B). No statistically significant differences due to effect of culture medium or scaffold type were observed, although a trend similar to that for metabolic activity was seen, with the 3DF-HA scaffolds having the lowest amount of DNA in both media among the different scaffold groups. 3.3. Gene expression analysis of osteogenic markers The gene expression of different osteogenic markers was analysed using quantitative real time PCR, and the results are shown in Fig. 7. Only statistically significant differences among cells grown on different scaffolds in the same medium are indicated in the figure. Mineralization medium had a positive effect on the expression of BSP and ALP compared with basic medium, and this effect was independent of the scaffold type. In contrast, BMP-2 expression in mineralization medium was lower than in basic medium on all scaffold types. Cells on 3DF scaffolds in basic medium showed a basal level of BSP expression which was significantly higher compared with the 3DF-HA scaffolds. In mineralization medium the 3DF-HA scaffolds showed basal expression of BSP,
which was up-regulated in the 3DF scaffolds and assembled composites, although not significantly. ALP expression by cells on the 3DF scaffolds in basic medium was also at basal levels, while it was down-regulated in both composites (significantly for the assembled composite). No differences among the scaffolds were observed in mineralization medium. Cells on the assembled composite scaffolds showed significant up-regulation of OP in basic medium compared with the 3DF and 3DF-HA scaffolds. Up-regulation of gene expression was also observed in mineralization medium, although no statistically significant differences were observed. In the case of BMP-2 the trend in expression levels was 3DF < 3DF-HA < assembled composite, although no significant differences were observed. 4. Discussion 3-D composite scaffolds containing PA and HA were fabricated using two different approaches. The conventional approach to fabricate composite scaffolds using polymer–HA composite filaments was successful only up to a 15 wt.% loading of HA. Above this amount it was impossible to fabricate 3-D scaffolds with the current choice of nozzles and equipment due to clogging by the cera-
5714
A. Nandakumar et al. / Acta Biomaterialia 9 (2013) 5708–5717
Fig. 6. (A) Metabolic activity of hMSCs cultured on different scaffolds in basic and mineralization medium on days 11 and 21, measured by alamarBlueÒ assay. (B) Amount of DNA after 21 days on different scaffolds in basic and mineralization medium as measured by CyQuant assay. Data are represented as means ± standard deviation. ⁄Statistically significant differences (P < 0.05).
mic particles and a consequent increase in the viscosity of the molten blend. The amount of ceramic that could be loaded in the present study was comparable with the amount in a study by Lam et al. [40], who succeeded in preparing a b-tricalcium phosphate (bTCP)–PCL composite with a 20 wt.% ceramic content by fused deposition modelling. Using a precision extrusion process Shor et al. [24] obtained a maximum loading of 25 wt.% HA in PCL. Higher ceramic loadings are possible using solvent casting methods [41], but these involve the use of possibly cytotoxic organic solvents. Besides, the scaffolds produced using solvent casting often lack close control over the porosity and architecture. In this regard, the assembly approach is innovative and versatile. As shown in Fig. 2A, different amounts of HA can be loaded onto scaffolds without the problems associated with processing of high amounts of ceramic and the need for complex extrusion systems. In addition, in the assembled constructs the ceramic surface is exposed to the
biological environment, resulting in different surface dynamics of dissolution/reprecipitation and protein adsorption and, hence, an effect on cell behaviour compared with bulk composites where the majority of the ceramic particles are covered by polymer and exposure is dependent on polymer degradation. Compression tests performed to evaluate the stiffness of the scaffolds showed that the addition of HA in the bulk increased the stiffness by almost 5 times compared with the 3DF polymer only scaffolds. This result was to be expected since in an in vivo setting the mineral component of bone imparts stiffness. In a previous study [25] assembled composites with different shapes of biphasic calcium phosphate (BCP) (75.4% HA, 24.6% b-TCP) particles were inserted into 3DF scaffolds and an increase in bending storage modulus and compressive storage modulus were observed compared with the 3DF scaffolds during dynamic mechanical analysis. In the current study no significant increase in modulus was
A. Nandakumar et al. / Acta Biomaterialia 9 (2013) 5708–5717
5715
Fig. 7. qPCR analysis for a panel of osteogenic genes after 21 days culture. Data are represented as means ± standard deviation. ⁄Statistically significant differences (P < 0.05).
observed due to the insertion of HA pillars. One possible reason could be the different loading of the ceramic particles. In the earlier study [25] BCP particles formed about 60% of the scaffold weight, while only 15% by weight of HA was used in the present study. In a review on the mechanical properties of calcium phosphates and composites [42] it was observed that the bulk properties of HA and BCP were spread across two orders of magnitude (compressive strength 5–500 MPa). In general, b-TCP possesses lower strength and the other mechanical properties are also poorer. Therefore, it can be reasonably stated that if the same amount of HA was used instead of BCP the resulting scaffolds would have had comparable or better mechanical properties. While the conventional composites (40 MPa) had a modulus above the minimum range stated for cancellous bone (12 MPa) [43], the polymeric and assembled scaffolds fell slightly short (8–9 MPa). However, it is possible to further increase the modulus of the assembled composite scaffolds by increasing the loading of HA particles as insertion of more ceramic particles has been shown to lead to higher values [25]. This adds another interesting facet to the assembly-based fabrication approach, the differential loading of ceramic particles to tune scaffold stiffness. Control over the mechanical properties, as well as over implant shape and architecture, increases clinical applicability of the assembled constructs to different sites, ranging from, for example, unloaded skull defects to load-bearing sites of long bones. On the other hand, it should not be ignored that an increase in the ceramic content, which can be advantageous in terms of construct bioactivity and mechanical properties, will decrease the porosity and pore interconnectivity of the assembled construct, which are of great importance for nutrient and oxygen supply, as well as cell and tissue infiltration and proliferation. Nevertheless, different components of the assembled constructs, such as polymer choice, ceramic choice and loading regime, can all be used as tools to design optimal scaffolds for the intended application. As mentioned before, the origin of the bioactivity of calcium phosphate-based bone graft substitutes is suggested to lie in dissolution/reprecipitation processes occurring on the surface of the
material in physiological environments [28–30,44]. However, as recently reviewed, it is difficult to pinpoint a single parameter that is responsible for the biological responses to a calcium phosphate material in vitro or in vivo [45]. While elevated concentrations of calcium have been suggested to positively affect osteogenic proliferation and differentiation of, for example, hMSCs and human periosteal derived stem cells (hPDCs) [46,47], release of calcium from a calcium phosphate ceramic is accompanied by other events, such as phosphate release, reprecipitation of a biological apatite layer, possibly containing endogenous proteins and other factors, and changes in surface topography. In the present study we have shown that in a medium that initially did not contain calcium or phosphate significantly more calcium and phosphate was released from the HA pillars than from 3DF-HA composites, confirming the better accessibility of the ceramic surface in the assembled than in classical monolithic composites. Although the release experiments were performed on HA pillars, and not on assembled composites, with the goal of determining the maximum numbers of ions that could be released, the accessibility of the ceramic surface in an assembled composite with 15 wt.% HA, in which a part of the surface of the ceramic pillars is ‘‘covered’’ by the surrounding polymer, is still expected to be higher than in 3DF-HA composite scaffolds. In contrast to release in SPS, 21 days immersion in SBF did not result in significant differences in calcium or phosphate concentration among different materials, suggesting that in an environment that is saturated with HA the ion release dynamics are different. The question remains, however, which medium is most relevant in mimicking the in vivo situation. Although SBF better represents the mineral composition of human blood plasma than SPS and other often used release media, such as PBS, the fluid replenishment regimes in vivo cannot be compared with the static release conditions in vitro. The presence and possible effects of endogenous biomolecules in vivo also cannot be ignored. The method used to measure calcium and phosphate concentrations in the present study could not be reliably applied to cell culture medium. Nevertheless, we did not observe mineral deposits on any of the scaffolds by SEM analysis (Fig. 5) or in methylene
5716
A. Nandakumar et al. / Acta Biomaterialia 9 (2013) 5708–5717
blue/basic fuchsin stained histological sections (data not shown), even after 21 days culture, in contrast to the scaffolds immersed in SBF, emphasizing the importance of medium composition when discussing release dynamics and associated events. For these reasons it is difficult to relate cell proliferation and osteogenic differentiation to the data on ion release and mineral deposition, although control over the release dynamics was one of the rationales for developing the assembled composites as an alternative to classical monolithic ones. Cells attached and proliferated on all scaffold types in both basic and mineralization medium. An increase in metabolic activity between days 11 and 21 was observed on all scaffolds, although less pronounced on the 3DF-HA composite compared with the polymer control and assembled composite. While no differences in metabolic activity were observed among the three scaffold types after 11 days culture, after 21 days the values for 3DF-HA were lower than the values for 3DF and HA assembled by 3DF. These data for metabolic activity correlated with the trend seen in the amount of DNA after 21 days, with the PLA–HA composite having lower amounts than both the 3DF control and assembled composite, although these differences were not statistically significant. In the classical 3DF-HA composite the presence of homogeneously distributed ceramic particles will change the properties of the surface throughout the scaffold, including the local concentration of calcium and phosphate ions, hydrophilicity and surface roughness, all of which could affect cell attachment and proliferation. Furthermore, in comparison with assembled composites, it is plausible that the release of calcium and phosphate ions from the ceramic component will occur later, which may explain why differences in metabolic activity were only significant after 21 days. In a study by Wilmowsky et al. [48] on different polyetheretherketone-based composites fabricated by SLS it was observed that human foetal osteoblasts cultured on composites containing 10 wt.% b-TCP showed the least pronounced proliferation and viability. They suggested that calcium phosphate compounds around the chosen concentration do not have a favourable effect on osteoblast proliferation in vitro. Zhou et al. [49] showed more pronounced attachment of human alveolar osteoblasts on PCL–TCP RP composites compared with PCL scaffolds and attributed this observation to increased hydrophilicity of the composite scaffolds, along with the ability of TCP to adsorb more proteins. In the assembled composites two distinct phases can be distinguished and therefore two different surface types are available for cells to attach to and proliferate on, which may explain why the effect of the ceramic was smaller in the assembled than in the monolithic composite. Among the four markers of osteogenic differentiation analysed a positive effect of the scaffolds containing ceramic was only observed on the expression of OP. OP, a phosphorylated acidic glycoprotein, is synthesized by osteoblasts in vitro and is found in mineralized tissue. OP regulates the attachment of osteoblasts and osteoclasts to the ECM [50,51]. Assembled composites in basic and mineralization medium showed the highest level of OP expression (>5-fold) amongst the scaffolds tested. Previous studies using sintered calcium phosphate ceramics with different degradation properties [52] and calcium phosphate-coated 3-D scaffolds prepared from PA (unpublished results) also showed high expression of OP compared with tissue culture plastic and uncoated scaffolds, respectively, suggesting that the presence of calcium phosphates enhances OP gene expression. Using MC3T3-E1 cells Beck et al. [50] observed that OP gene expression was induced by the presence of phosphate, which is a result of the hydrolysis of b-glycerol phosphate, an additive used in cell culture media to induce mineralization. A significant up-regulation of OP expression in hPDCs was also observed after addition of calcium, phosphate or both to the medium [53]. However, as discussed before, although plausible, whether the up-regulation of OP observed in our study is
caused by elevated levels of phosphate or the combination of both calcium and phosphate in the assembled composites can only be proven by precise measurement of ionic concentrations in the scaffold vicinity during cell culture. A similar discussion holds for the expression of BMP-2 and BSP, which were shown to be up-regulated by various cell types in the presence of elevated calcium ion levels in a concentration-dependent manner [46,47,54,55], suggesting that the levels in the present study may have been below the necessary threshold to observe an effect. Regarding the down-regulation of ALP expression on assembled composites in basic medium, in a previous study we observed lower ALP levels on polymeric fibres coated with calcium phosphate compared with the uncoated controls [56]. However, ectopic bone formation following subcutaneous implantation in mice only occurred on hMSCs/coated fibre constructs, in contrast to constructs consisting of hMSCs and uncoated fibres, emphasizing the difficulty of assessing the bioactivity of biomaterials in general and calcium phosphate-based ceramics in particular in vitro. Although, clearly, additional studies are needed to understand the biological mechanisms by which the scaffolds tested in this study can affect processes related to bone regeneration, and to further explore the possibilities they offer to control these processes, we have presented an alternative way of assembling polymer/ceramic composites for bone regeneration. By keeping the two material types as separate components of the composite material, the versatility of RP can be employed to a greater extent, offering more freedom in the design of scaffold properties than is the case with conventional composites. 5. Conclusions Two different strategies for the fabrication of 3-D polymer– ceramic composite scaffolds for bone regeneration using RP have been studied. While the conventional approach, in which the ceramic particles were dispersed within the polymeric component, resulted in a mechanically stiffer construct, assembled composites, consisting of distinct polymer and ceramic phases, allowed the incorporation of larger amounts of ceramic. The two methods of producing composite scaffolds further expand the possibilities for controlling exposure of the individual components to the biological environment. Acknowledgement Ms. Anita Podt is acknowledged for preparation of the molds by stereolithography. Appendix A. Figures with essential colour discrimination Certain figures in this article, particularly Figs. 3, 6 and 7, 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.2012.10.044. Appendix B. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.actbio.2012. 10.044. References [1] Augat P, Schorlemmer S. The role of cortical bone and its microstructure in bone strength. Age Ageing 2006;35(Suppl. 2):ii27–31.
A. Nandakumar et al. / Acta Biomaterialia 9 (2013) 5708–5717 [2] Fratzl P, Gupta HS, Paschalis EP, Roschger P. Structure and mechanical quality of the collagen-mineral nano-composite in bone. J Mater Chem 2004;14:2115–23. [3] Zaidi M. Skeletal remodeling in health and disease. Nat Med 2007;13:791–801. [4] Erlebacher A, Filvaroff EH, Gitelman SE, Derynck R. Toward a molecular understanding of skeletal development. Cell 1995;80:371–8. [5] Laurencin C, Khan Y, El-Amin SF. Bone graft substitutes. Expert Rev Med Devices 2006;3:49–57. [6] Kainer MA, Linden JV, Whaley DN, Holmes HT, Jarvis WR, Jernigan DB, et al. Clostridium infections associated with musculoskeletal-tissue allografts. New Engl J Med 2004;350:2564–71. [7] Mroz TE, Joyce MJ, Steinmetz MP, Lieberman IH, Wang JC. Musculoskeletal allograft risks and recalls in the United States. J Am Acad Orthop Surg 2008;16:559–65. [8] Damien CJ, Parsons JR. Bone graft and bone graft substitutes: a review of current technology and applications. J Appl Biomater 1991;2:187–208. [9] Szpalski C, Wetterau M, Barr J, Warren SM. Bone tissue engineering: current strategies and techniques - Part I: Scaffold. Tissue Eng B Rev 2012;18:246–57. [10] Zhang P, Hong Z, Yu T, Chen X, Jing X. In vivo mineralization and osteogenesis of nanocomposite scaffold of poly(lactide-co-glycolide) and hydroxyapatite surface-grafted with poly(L-lactide). Biomaterials 2009;30:58–70. [11] Kim HW, Lee HH, Knowles JC. Electrospinning biomedical nanocomposite fibers of hydroxyapatite/poly(lactic acid) for bone regeneration. J Biomed Mater Res A 2006;79:643–9. [12] Montjovent MO, Mathieu L, Hinz B, Applegate LL, Bourban PE, Zambelli PY, et al. Biocompatibility of bioresorbable poly(L-lactic acid) composite scaffolds obtained by supercritical gas foaming with human fetal bone cells. Tissue Eng 2005;11:1640–9. [13] Yefang Z, Hutmacher DW, Varawan SL, Meng LT. Comparison of human alveolar osteoblasts cultured on polymer–ceramic composite scaffolds and tissue culture plates. Int J Oral Max Surg 2007;36:137–45. [14] Yang F, Both SK, Yang X, Walboomers XF, Jansen JA. Development of an electrospun nano-apatite/PCL composite membrane for GTR/GBR application. Acta Biomater 2009;5:3295–304. [15] Aboudzadeh N, Imani M, Shokrgozar MA, Khavandi A, Javadpour J, Shafieyan Y, et al. Fabrication and characterization of poly(D,L-lactide-co-glycolide)/ hydroxyapatite nanocomposite scaffolds for bone tissue regeneration. J Biomed Mater Res A 2010;94:137–45. [16] Jose MV, Thomas V, Johnson KT, Dean DR, Nyairo E. Aligned PLGA/HA nanofibrous nanocomposite scaffolds for bone tissue engineering. Acta Biomater 2009;5:305–15. [17] Song JH, Kim HE, Kim HW. Electrospun fibrous web of collagen–apatite precipitated nanocomposite for bone regeneration. J Mater Sci Mater Med 2008;19:2925–32. [18] Thomas V, Dean DR, Jose MV, Mathew B, Chowdhury S, Vohra YK. Nanostructured biocomposite scaffolds based on collagen coelectrospun with nanohydroxyapatite. Biomacromolecules 2007;8:631–7. [19] Laschke MW, Strohe A, Menger MD, Alini M, Eglin D. In vitro and in vivo evaluation of a novel nanosize hydroxyapatite particles/poly(ester-urethane) composite scaffold for bone tissue engineering. Acta Biomater 2010;6:2020–7. [20] Huang YX, Ren J, Chen C, Ren TB, Zhou XY. Preparation and properties of poly(lactide-co-glycolide) (PLGA)/ nano-hydroxyapatite (NHA) scaffolds by thermally induced phase separation and rabbit MSCs culture on scaffolds. J Biomater Appl 2008;22:409–32. [21] Peltola SM, Melchels FP, Grijpma DW, Kellomaki M. A review of rapid prototyping techniques for tissue engineering purposes. Ann Med 2008;40:268–80. [22] Duan B, Wang M, Zhou WY, Cheung WL, Li ZY, Lu WW. Three-dimensional nanocomposite scaffolds fabricated via selective laser sintering for bone tissue engineering. Acta Biomater 2010;6:4495–505. [23] Tan KH, Chua CK, Leong KF, Cheah CM, Cheang P, Abu Bakar MS, et al. Scaffold development using selective laser sintering of polyetheretherketone– hydroxyapatite biocomposite blends. Biomaterials 2003;24:3115–23. [24] Shor L, Guceri S, Wen XJ, Gandhi M, Sun W. Fabrication of three-dimensional polycaprolactone/hydroxyapatite tissue scaffolds and osteoblast–scaffold interactions in vitro. Biomaterials 2007;28:5291–7. [25] Moroni L, Hamann D, Paoluzzi L, Pieper J, de Wijn JR, van Blitterswijk CA. Regenerating articular tissue by converging technologies. PLoS One 2008;3:e3032. [26] Woodfield TBF, Malda J, de Wijn J, Peters F, Riesle J, van Blitterswijk CA. Design of porous scaffolds for cartilage tissue engineering using a three-dimensional fiber-deposition technique. Biomaterials 2004;25:4149–61. [27] Barbieri D, Renard AJ, de Bruijn JD, Yuan H. Heterotopic bone formation by nano-apatite containing poly(D,L-lactide) composites. Eur Cell Mater 2010;19:252–61. [28] Geesink RG, de Groot K, Klein CP. Bonding of bone to apatite-coated implants. J Bone Joint Surg Br 1988;70:17–22. [29] Hanawa T, Kamiura Y, Yamamoto S, Kohgo T, Amemiya A, Ukai H, et al. Early bone formation around calcium-ion-implanted titanium inserted into rat tibia. J Biomed Mater Res A 1997;36:131–6.
5717
[30] Barradas AM, Yuan H, van Blitterswijk CA, Habibovic P. Osteoinductive biomaterials: current knowledge of properties, experimental models and biological mechanisms. Eur Cell Mater 2011;21:407–29. [31] Du C, Meijer GJ, van de Valk C, Haan RE, Bezemer JM, Hesseling SC, et al. Bone growth in biomimetic apatite coated porous Polyactive((R)) 1000PEGT70PBT30 implants. Biomaterials 2002;23:4649–56. [32] Nandakumar A, Fernandes H, de Boer J, Moroni L, Habibovic P, van Blitterswijk CA. Fabrication of bioactive composite scaffolds by electrospinning for bone regeneration. Macromol Biosci 2010;10:1365–73. [33] Moroni L, de Wijn JR, van Blitterswijk CA. 3D fiber-deposited scaffolds for tissue engineering: influence of pores geometry and architecture on dynamic mechanical properties. Biomaterials 2006;27:974–85. [34] Habibovic P, Yuan H, van der Valk CM, Meijer G, van Blitterswijk CA, de Groot K. 3D microenvironment as essential element for osteoinduction by biomaterials. Biomaterials 2005;26:3565–75. [35] Landers R, Pfister A, Hubner U, John H, Schmelzeisen R, Mulhaupt R. Fabrication of soft tissue engineering scaffolds by means of rapid prototyping techniques. J Mater Sci 2002;37:3107–16. [36] Bohner M, Lemaitre J. Can bioactivity be tested in vitro with SBF solution? Biomaterials 2009;30:2175–9. [37] Both SK, Van der Muijsenberg AJC, Van Blitterswijk CA, De Boer J, De Bruijn JD. A rapid and efficient method for expansion of human mesenchymal stem cells. Tissue Eng 2007;13:3–9. [38] Fernandes H, Mentink A, Bank R, Stoop R, van Blitterswijk C, de Boer J. Endogenous collagen influences differentiation of human multipotent mesenchymal stromal cells. Tissue Eng A 2010;16:1693–702. [39] Livak KJ, Schmittgen TD. Analysis of relative gene expression data using realtime quantitative PCR and the 2(-Delta Delta C(T)) method. Methods 2001;25:402–8. [40] Lam CXF, Teoh SH, Hutmacher DW. Comparison of the degradation of polycaprolactone and polycaprolactone-(beta-tricalcium phosphate) scaffolds in alkaline medium. Polym Int 2007;56:718–28. [41] Cao H, Kuboyama N. A biodegradable porous composite scaffold of PGA/betaTCP for bone tissue engineering. Bone 2010;46:386–95. [42] 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. [43] Athanasiou KA, Zhu CF, Lanctot DR, Agrawal CM, Wang X. Fundamentals of biomechanics in tissue engineering of bone. Tissue Eng 2000;6:361–81. [44] Ducheyne P, Beight J, Cuckler J, Evans B, Radin S. Effect of calcium phosphate coating characteristics on early post-operative bone tissue ingrowth. Biomaterials 1990;11:531–40. [45] Habibovic P, Barralet JE. Bioinorganics and biomaterials: bone repair. Acta Biomater 2011;7:3013–26. [46] Barradas AM, Fernandes HA, Groen N, Chai YC, Schrooten J, van de Peppel J, et al. A calcium-induced signaling cascade leading to osteogenic differentiation of human bone marrow-derived mesenchymal stromal cells. Biomaterials 2012;33:3205–15. [47] Chai YC, Roberts SJ, Schrooten J, Luyten FP. Probing the osteoinductive effect of calcium phosphate by using an in vitro biomimetic model. Tissue Eng A 2011;17:1083–97. [48] von Wilmowsky C, Vairaktaris E, Pohle D, Rechtenwald T, Lutz R, Munstedt H, et al. Effects of bioactive glass and beta-TCP containing three-dimensional laser sintered polyetheretherketone composites on osteoblasts in vitro. J Biomed Mater Res A 2008;87A:896–902. [49] Zhou YF, Hutmacher DW, Varawan SL, Lim TM. In vitro bone engineering based on polycaprolactone and polycaprolactone–tricalcium phosphate composites. Polym Int 2007;56:333–42. [50] Beck GR, Zerler B, Moran E. Phosphate is a specific signal for induction of osteopontin gene expression. Proc Natl Acad Sci USA 2000;97:8352–7. [51] Roach HI. Why does bone-matrix contain noncollagenous proteins – the possible roles of osteocalcin, osteonectin, osteopontin and bone sialoprotein in bone mineralization and resorption. Cell Biol Int 1994;18:617–28. [52] Yuan H, Fernandes H, Habibovic P, de Boer J, Barradas AM, de Ruiter A, et al. Osteoinductive ceramics as a synthetic alternative to autologous bone grafting. Proc Natl Acad Sci USA 2010;107:13614–9. [53] Chai YC, Roberts SJ, Desmet E, Kerckhofs G, van Gastel N, Geris L, et al. Mechanisms of ectopic bone formation by human osteoprogenitor cells on CaP biomaterial carriers. Biomaterials 2012;33:3127–42. [54] Tada H, Nemoto E, Kanaya S, Hamaji N, Sato H, Shimauchi H. Elevated extracellular calcium increases expression of bone morphogenetic protein-2 gene via a calcium channel and ERK pathway in human dental pulp cells. Biochem Biophys Res Commun 2010;394:1093–7. [55] Honda Y, Anada T, Kamakura S, Nakamura M, Sugawara S, Suzuki O. Elevated extracellular calcium stimulates secretion of bone morphogenetic protein 2 by a macrophage cell line. Biochem Biophys Res Commun 2006;345:1155–60. [56] Nandakumar A, Yang L, Habibovic P, van Blitterswijk C. Calcium phosphate coated electrospun fiber matrices as scaffolds for bone tissue engineering. Langmuir 2010;26:7380–7.