Preparation, Characterization and Biological Studies of Β-TCP and Β-TCP/Al2O3 Scaffolds Obtained by Gel-Casting of Foams

Replacement tissues for tissue engineering can be produced by seeding human cells onto scaffolds. In order to guarantee adequate bio-compatibility, porosity and mechanical resistance for promoting cellular growth, proliferation and differentiation within scaffold structures, it is necessary to investigate and improve materials and processing routes. β-tricalcium phosphate can be considered a very suitable bio-ceramic material for bone therapy because of its biocompatibility, osteo-conductivity and neo-vascularization potential. Alumina is commonly used as a sintering additive. In this study, β-TCP and β-TCP/Al2O3 scaffolds were obtained by gel-casting method. The scaffolds showed high porosity (86-88%) and pore sizes ranging from 200 to 500 µm. Even though alumina did not promote improvement in β-TCP/Al2O3 scaffolds in terms of mechanical performance, they showed great cytocompatibility as there was no cytotoxic and genotoxic effect. Therefore, β-TCP and β-TCP/Al2O3 scaffolds are good candidates for application in tissue engineering.


Introduction
Tissue engineering is a scientific field that integrates biological components, such as cells and growth factors, with engineering principles and synthetic materials.Therefore, as an interdisciplinary field, tissue engineering concerns the development of biological substitutes capable of replacing or repairing diseased or damaged tissue and organs in humans [1][2][3][4] .Since the 80's, and especially in the last decade, tissue engineering and regenerative medicine have been proving its importance and potential to revolutionize important areas of medicine [1][2][3][4] , such as cardiovascular [5][6] , skin regeneration 7 and bone treatments [8][9][10][11] .
Replacement tissues can be produced by first seeding human cells onto scaffolds [1][2][3][4] .Scaffolds are temporary support structures that provide adequate conditions for cell proliferation, migration, and differentiation in 3D, allowing formation of specific tissues with appropriate functions.Some common properties of an ideal scaffold for tissue engineering are biocompatibility, porosity with interconnected pores and adequate mechanical strength, depending mainly on the tissue to be repaired [1][2]4,[10][11] .
Gel-casting method consists in adding a foaming agent to a ceramic suspension composed of ceramic powder, organic monomers and dispersants, thus providing a mechanical action.After gelation of the foam due to in situ polymerization of water-soluble monomers (previously added to the ceramic suspension) and its sintering, some of the scaffold properties produced were: ideal porosity (40-90%) and porous morphology with spherical-shaped pores measuring between 50 and 800 mm, and thick walls with homogeneous micro-structure, thus improving the mechanical properties of these materials, which can help cellular growth.Furthermore, this method is cheaper than others and does not require atmospheric control [20][21][22][23] .
Bone scaffolds can have an optimal performance so long as they can be bio-degradable, bio-resorbable and capable of providing mechanical support for repair and regeneration of diseased or damaged bone while the growing tissue replaces the scaffold [1][2][3][4][9][10] .
Unfortunately, β-TCP undergoes a β→α polymorphic transformation while being processed, which takes place at temperatures as low as 1150 ºC 28,32 .As a straightforward consequence, a poor mechanical performance (e.g.compressive strength) attributed to porous structures is more pronounced, limiting its application.Also, the porosity increases the complexity for manufacturing reproducible scaffolds.
One way to improve such mechanical properties would be to use sintering additives, which could improve the thermal stability of β-TCP or form a liquid phase during sintering, either process favoring material densification [33][34][35][36] .Among the various additives (e.g.alumina, titania, calcium pyrophosphate, zirconium, magnesium and others), alumina stands out due to the combination of its excellent corrosion resistance, good bio-compatibility, high wear resistance, high compressive strength and high hardness 37 .For this reason, alumina is commonly used to improve β-TCP densification during sintering 35 .
In the literature, there are few reports about bio-ceramic materials produced by gel-casting method 17,[38][39] .Thus, this manuscript makes a scientific contribution to the processing of bio-ceramic materials with high porosity, adequate sphericity and high interconnectivity all desirable characteristics for a good scaffold.
Based on this context, the current work reports the preparation of β-TCP and β-TCP/Al 2 O 3 scaffolds by using gelcasting of foams with appropriate porosity, interconnectivity and bio-active behavior for application in tissue engineering.

Fabrication of β-TCP and β-TCP/Al 2 O 3 scaffolds
β-TCP powder was synthesized by means of solid-state reaction at 1050 ºC from an appropriate mixture of calcium carbonate (CaCO 3 -Synth) and calcium phosphate (CaHPO 4 -Synth), as already published elsewhere 40 .The resulting powder was ballmilled for 48 hours (alumina grinding media with 10 mm spheres at ball/powder ratio of 10:1) and presented a mean size of 3.4 µm and particle size distribution of 0.3-26 µm.
To produce the foam, non-ionic surfactant (Lutensol ON-110, BASF) was added to the slurry before being stirred for 2 minutes with a mixer.Ammonium persulfate (APS, Vetec) and N,N,N,N´-tetramethylethylenediamine (TEMED, Sigma-Aldrich) were used as initiator and catalyst for gelation reaction, respectively.The foams were dried at 50 ºC for 24 hours and then the samples were core-drilled (cylinders, d = 7 mm and h = 14 mm).Heat treatment was performed as follows: samples were fired at 500 ºC (heating rate of 1 ºC/min and dwell time of 1 hour) to eliminate all organic components; next, sintering was made at temperature of 1200 ºC (heating rate of 5 ºC/min and dwell time of 2 hours).

Characterization
β-TCP and β-TCP/Al 2 O 3 scaffolds were analyzed by X-ray diffraction (Shimadzu XRD7000, CuKα radiation, 2θ = 20-40º, 30 mA, 40 k).The JCPDS database files used to identify the crystalline phases formed were 09-0169 for β-TCP, 09-0348 for α-TCP and 10-0173 for Al 2 O 3 .The fracture surface of the sintered scaffolds was observed by using a scanning electron microscope (JEOL JSM6360-LV).The total porosity of the scaffolds was determined according to Equation (1) 41 .The geometric density (ρ scaffold ) of the cylindrical samples was calculated by using their dimensions (height and diameter) and weight.The theoretical density (ρ He ) was calculated for each formulation according to Equation (2), where ρ ß-TCP and ρ Al2O3 are the theoretical densities of β-TCP (3.07 g/cm 3 ) and Al 2 O 3 (3.98 g/cm 3 ), respectively.X β-TCP and X Al 2 O 3 are the mass fractions of each phase added.
X-ray microtomographic images of β-TCP and β-TCP/ Al 2 O 3 scaffolds were obtained by using a high resolution X-ray microtomograph (SkyScan 1272 Bruker microCT, Kontich, Belgium) operating at 30 kV and 160 µA.The nominal resolution of 6.5 µm was obtained by using X-ray detector of 1224 x 820 pixels.NRecon software (Software Version 1.6.9.8, Bruker MicroCT ® ) was used to reconstruct cross-section images from microtomographic projections into 3D images based on the Feldkamp algorithm.The same electronic density contrast was selected for all samples in order to allow comparisons.The 3D images were viewed with CTVOX software (Version 2.7.0,Bruker MicroCT ® ), whereas 2D images per slices were obtained with Dataviewer software (Version 1.5.1.2,Bruker MicroCT ® ).Total porosity and pore size distribution were quantified by using the CT analyzer software (Version 1.14.4.1, Bruker MicroCT ® ).For acquisition of porosity parameters, the volume object of interest (VOI) was previously set to 4 mm 3 .
The compressive strength of at least ten cylindrical samples (h/d = 2) was determined by using a universal testing machine (EMIC DL2000) at cross-head speed of 0.5 mm.min -1 .

Biological tests
In this study, osteoblast-like cells (MG 63) were obtained from the Rio de Janeiro Cell Bank (Rio de Janeiro, RJ, Brazil).The cells were cultured in DMEM (Cultilab Curitiba, Brazil), supplemented with 10% fetal bovine serum (FBS), penicillin (100 UI/mL), streptomycin (100 μg/mL) and then maintained at 37 °C and 5% of CO 2 .Cell culture flasks of 75 mL and 250 cm 2 were used.The culture medium was changed every two days and the cell growth was assessed by using a reverse phase microscope (Microlimaging GmbH -Axiovert 40C, Carl Zeiss Microscope, Jena, Germany).

Cell attachment
For assessment of cell attachment and morphology of the cells in contact with the scaffolds, these were placed in 24-well plates and the cells plated on them before being maintained at 37 °C and 5% CO 2 .After 3-day growth of MG63 on the scaffolds, the cells were investigated by using SEM (Inspect S50, FEI Company, Brno, Czech Republic).For this, the culture medium was removed and the scaffolds were twice rinsed with PBS.The cells adhered in the scaffolds were fixed with 0.25% glutaraldehyde and 4% paraformaldehyde diluted in PBS before being dehydrated by a series of ascending ethanol concentrations (70%, 80%, 90% and 100% for 10 minutes each).Next, the scaffolds were air-dried at room temperature for 24 hours.For SEM examination, the scaffolds were sputter-coated with palladium-gold alloy (Polaron SC 7620 Sputter Coater, Quorum Technologies, Newhaven, UK) at a thickness of 7-10 nm (10-15 mA, under a vacuum of 130 mTorr), with scanning electron microscope operating between 15-30 kV and spot 7.

Cytotoxicity
For the cytotoxicity test, 2x10 4 cells were plated in 24well plates and maintained at 37 °C and 5% CO 2 .After 24 hours, the scaffolds were placed on the cells and the plates maintained at 37 °C and 5% CO 2 for further 24 hours.Next, the viability of the cells put in contact with the scaffolds was determined by MTT assay (3-(4,5 dimethylthiazol-2yl)-2,5-diphenyltetrazolium bromide, Sigma, St Louis, MO, USA), which is a test that evaluates not only the number of cells, but also the level of their metabolic activity as it is based on the activity of enzymes, such as succicil dehydrogenase, present in viable cells 42 .MTT solution was added to each well and the cells were incubated for 1 hour.Next, DMSO (dimethyl sulfoxide solvent, Sigma-Aldrich, St Louis, MO, USA) was also added to each well and the plates were shaken at room temperature for 10 minutes.The resulting optical density was measured in a spectrophotometer (Biotek-EL808IU, BioTek Instruments, Inc., Winooski, VT, USA) at 570 nm.The cytotoxicity was expressed as a percentage of the control group (100%) and the differences between values were statistically analyzed by ANOVA (P ≤ 0.05).

Genotoxicity
For assessment of the genotoxicity, 2x10 4 cells were plated in 24-well plates and maintained at 37 °C and 5% CO 2 for 24 hours.After this period, the scaffolds were placed on the cells and the plates were maintained at 37 °C and 5% CO 2 for further 24 hours.Next, the scaffolds were removed from the plates and the cells were rinsed with PBS and fixed with 4% paraformaldehyde diluted in PBS for 10 minutes.PBS and Fluroshield with DAPI solution (Sigma, St Louis, MO, USA) were added to each well and the nuclei of the cells were observed by means of immunofluorescence and then photographed with a digital camera (Sony F828 Digital, CyberShot, 8.0 megapixels) coupled to an inverted light microscope (Carl Zeiss Microscope Micro lmaging GmbH -Axiovert 40C, Germany).The micronucleus test is based on the loss of chromosomes or their fragments during cell mitosis.Entire chromosomes or their fragments not reincorporated by the nucleus after cell division lead to micronuclei formation 43 .A cell counter (Image J software) was used to assist in the counting of micronuclei.The micronuclei were determined microscopically at 1,000 cells/well and the differences between the values were statistically analyzed by using Mann-Whitney U test (P ≤ 0.05).Figure 2 presents SEM micrographs of the fracture surface of the β-TCP and β-TCP/Al 2 O 3 scaffolds.All scaffolds (Figure 2-B1, C1 and D1) presented homogeneous macro-structure with spherical and interconnected pores.The mean pore size varied from 200 to 500 μm, which is considered adequate for bone tissue growth 11,27 .For β-TCP scaffolds, it is possible to observe an inter-granular porosity.With regard to β-TCP/Al 2 O 3 scaffolds, alumina particles filled the micro-porosity of the scaffolds, as can be seen in Figure 2 (B2-B3, C2-C3 and D2-D3).It can also be observed that the micro-structure of the scaffolds consists of two phases: calcium phosphate and small grains related to alumina.These results are in agreement with those reported by Ayed and Bouaziz 35 , who studied the effect of alumina addition on densification of tricalcium phosphate-26.52wt.-% fluorapatite composites.

Results and Discussion
Furthermore, in another study, high-density tricalcium phosphate-26.52wt.-% fluorapatite-20 wt.-% zirconia composites were produced after sintering at various temperatures 44 .The authors observed that high temperatures (above 1300 ºC) or mass percentages of zirconia (near 20 wt.-%) are not the best conditions for composite densification.This is probably due to the new compound formation (CZ and TTCP).Also, it was observed that partial decomposition of TCP increases during the sintering process by the addition of zirconia (near 20 wt.-%) at 1400 °C.
Table 1 shows the porosity values for β-TCP and β-TCP/ Al 2 O 3 scaffolds, whose geometric porosities varied from 86 to 88%, values higher than those reported in the literature.Therefore, the processing route influences the properties of the scaffolds.The comparison carried out here is between scaffolds made of other bioactive ceramic and high-porosity glass materials (~70%) and prepared by gel-casting method.Therefore, few reports were found.For example, Rehman and Zhang prepared hydroxyapatite scaffolds with porosities ranging from 70 to 77% by combining gel-casting and replica methods 45 .Wu and collaborators produced bioglass scaffolds with 74-84% porosity by using a modified gel-casting of foams 46 .
The porosity of β-TCP and β-TCP/Al 2 O 3 scaffolds determined by microtomography varied from 70 to 77%, values lower than those obtained by geometric porosity for all scaffolds.This result was already expected as meso-and nano-pores can be calculated by this method.
Moreover, it is necessary to point out that the presence of alumina does not influence the scaffold porosity.
X-ray microtomographic images were also used to obtain 3D images of the scaffolds, as shown in Figure 3.The color bar is proportional to the attenuation of X-rays passing through the scaffolds: colors on the left side show lower attenuation than those on the right side.The attenuation is influenced by changes in the sample, such as thickness, density and electronic density (basically, the atomic number of the involved samples).No significant changes in the micro-structures were observed in the scaffolds.
Different cross-sections of these images (Figure 4) provide a two-dimensional visualization of the interior of the scaffolds, showing a highly porous material.
Figure 5 presents the 3D model obtained (Figure 3) and its pore size distribution.All samples show a wide pore size distribution, as observed in the SEM images, as well as similar porosity, confirming that alumina acts mainly as a filler for β-TCP grains.
Table 2 shows the compressive strength of β-TCP and β-TCP/Al 2 O 3 scaffolds.By comparing between β-TCP and β-TCP/Al 2 O 3 scaffolds, alumina did not improve the mechanical resistance of the former.This fact can be associated to the great difference in the sintering temperature of both powders.Another possibility could be the size of alumina particles, which are still in micrometer scale.In the literature, it is known that alumina particles are commonly used in micro-nanometric scale to improve the β-TCP densification during the sintering step.In the case of micrometric particles, these should be used in small amounts (≤ 2.5 wt.-%).This was confirmed by Bouslama et al., who used alumina as a promising material for reinforcement of calcium phosphate 47 .In their study, the mechanical properties of the β-TCP-fluorapatite composite was improved by adding 2.5 wt.-% micro-sized alumina powder as additive.
In recent years, the use of nano-sized ceramic additives for strengthening bone substitutes has been studied.In the study by Ghzanfari et al. 48, the effect of nano-alumina content on phase transformation and micro-structural and mechanical properties of HA/nAl 2 O 3 scaffolds, fabricated by the freeze-casting method, was evaluated.Their results showed that the decomposition of HA to TCP was accelerated by increasing the nAl 2 O 3 content at sintering temperature.In addition, the increase in the nAl 2 O 3 content increased the pore size and the compressive strength of the scaffolds.In another study, Hesaraki et al. 49 studied the addition of alumina and silica nano-particles to produce beta-tricalcium phosphate scaffolds, resulting in improved mechanical and biological properties.Therefore, we intend to use nano-sized particles in further works to obtain bio-ceramic scaffolds by gel-casting method, since this method is still little explored in the literature and allows the production of highly porous scaffolds with highly interconnected spherical pores, which would promote cell growth within the structure.
As it is known, porosity and mechanical strength are inversely proportional properties.Therefore, the low values found for the mechanical strength of β-TCP and β-TCP/ Al 2 O 3 scaffolds were already expected.However, our results indicated that the compressive strength of ~0.2-0.4MPa is sufficient for the scaffolds to be handled and manipulated during all tests.The compressive strength of spongy bone is in the range of 0.2-4.0MPa when the relative density is 0.1 50 .Hence, the compressive strength of the present scaffolds falls in this range, but lies closer to the lower bound.It has also been speculated that it might not be necessary to fabricate a scaffold with mechanical strength equal to that of bone because cells cultured on scaffolds and in vitro formation of new tissues will create a bio-composite and will increase the time-dependent strength of the scaffold significantly 25 .
Cell attachment was observed in all scaffolds.There was an intimate contact between the cells as they were scattered all over the surface (Figure 6).
Cell viability was analyzed by MTT assay and the average percentages of absorbance values measured for each concentration are shown in Figure 7.The data obtained in this test were statistically analyzed by using ANOVA, which showed that there was no statistical difference between control (100%) and experimental groups (P > 0.05).Thus, none of the concentrations tested was found to be cytotoxic to the cells.Moreover, there was no statistical difference between the groups regarding the amount of alumina used (P > 0.05).These results indicate that different amounts of alumina within β-TCP did not affect the cell viability, which was already expected as β-TCP and Al 2 O 3 have been investigated and exhibited high biocompatibility [51][52][53] .
Genotoxicity was evaluated in this study through the micronucleus test.The DNA structures contained in the cytoplasm are clearly separated from the main nuclei as their area is 1/3 smaller, thus being considered as micro-nuclei (Figure 8).Only cells containing less than five micronuclei were counted, whereas mitotic and apoptotic cells were not.
The mean number of micronuclei observed in this study (1,000 cells/well) can be observed in Figure 9.There was no statistical difference between control and experimental groups regarding the number of micronuclei (P > 0.05), although β-TCP scaffolds have presented the highest number of micronuclei.Furthermore, the groups were not statistically different from each other (P > 0.05).These results were  already expected as no degree of cytotoxicity had been previously observed in this study and β-TCP scaffold is known to be biocompatible 47 .According to Tsaousi et al. 54 , alumina particles are weakly genotoxic only to human cells in vitro and the induction of micronuclei is not affected by the size or shape of the particles.However, in our study, different amounts of alumina added to β-TCP scaffold were not found to be genotoxic to the cells.
Based on the results above, β-TCP and β-TCP/Al 2 O 3 scaffolds produced by gel-casting method showed good characteristics, such as high porosity, interconnected pores with spherical geometry and pore size in the range of 200-500 µm, which make these scaffolds good candidates for use in tissue engineering.

Conclusions
It was possible to obtain β-TCP and β-TCP/Al 2 O 3 scaffolds with high porosity (> 86%) and spherical and interconnected pores by using the gel-casting method.All scaffolds showed cytocompatibility with no cytotoxic and genotoxic effects.These findings indicate a possible application of this β-TCP/Al 2 O 3 scaffold for bone tissue engineering.
Preparation, Characterization and Biological Studies of Β-TCP and Β-TCP/Al 2 O 3 Scaffolds Obtained by Gel-Casting of Foams

Figure 7 .
Figure 7. Graph showing mean percentage of absorbance, obtained with MTT assay, after contact of the cells with β-TCP/Al 2 O 3 scaffolds for 24 hours.

Figure 8 .
Figure 8. Photomicrography of MG-63 nuclei after contact of the cells with β-TCP/Al 2 O 3 -85/15 for 24 hours.Note the presence of a micronucleus (arrow) in this site (Fluroshield with DAPI Immunofluorescence staining; original magnification x20).

Figure 9 .
Figure 9. Graph showing number of micronuclei in 1,000 cells/well after contact of the cells with β-TCP/Al 2 O 3 scaffolds for 24 hours.