Novel synthesis method combining a foaming agent with freeze-drying to obtain hybrid highly macroporous bone scaffolds

1. Introduction

Restoration of severe bone defects, in which the normal process of bone regeneration is hindered, still poses a significant problem. Currently, there are several therapeutic strategies for bone fractures treatment in order to accelerate the regeneration process, inter alia autografts, allografts or implantation of bone-substituting biomaterials. The limitations of the use of autogenic and allogenic bone grafts—including donor site morbidities, restricted donor source, pain, potential immunogenic reaction, infection, or disease transmission - gave rise to intense development of bone tissue engineering [1,2].

Typical tissue engineered products include a three-dimensional (3D) porous scaffold, which is often loaded with osteoprogenitor cells or stem cells and/or growth factors [3]. The porous scaffold provides mechanical support and space for migration and proliferation of cells and facilitates tissue ingrowth as well as implant vascularisation [4]. In the field of engineering of biomaterials, porosity is determined as the percentage of void space in a solid material. Total porosity is a sum of closed (surrounded by a solid material, not accessible to fluids) and open (having connection to the space outside the solid material and enabling fluids flow through the material) pores [5]. Importantly, open and interconnected porosity possess the most significant biomedical importance because they facilitate diffusion of nutrients and metabolites within the implant as well as provide good oxygenation, promoting bone tissue ingrowth and new blood vessel formation [6]. Dependent on the pore size, porosity of the biomaterials may be classified as macroporosity (pore diameter > 50 μm), mesoporosity (pore diameter in the range 2-50 μm) and microporosity (pore diameter < 2 μm) [4,7]. It was also estimated that minimum diameter of pores enabling effective regeneration and restoration of mineralized bone tissue is approx. 100 μm [4].

The aim of this work was to synthesise novel highly macroporous bone scaffolds made of polysaccharide chitosan/agarose cryogel matrix reinforced with hydroxyapatite nanopowder (nanoHA). Chitosan was selected for the production of novel biomaterials due to its non-toxicity, high biocompatibility, rapid biodegradation, and inherit antibacterial properties [8]. Whereas hydroxyl groups of agarose component are expected to form chemical interactions with NH groups of chitosan molecules [9], leading to the formation of hybrid chitosan/agarose matrix characterized by improved stability and better mechanical properties compared to the pure chitosan or agarose material. It should be noted that according to the available literature, there are no scientific reports describing tri-component biomaterial composed of chitosan, agarose, and nanoHA. However, there are some papers describing materials for regenerative medicine applications comprising chitosan and agarose, e.g. chitosan/agarose hydrogel wound dressings or cartilage scaffolds [[9], [10], [11]], chitosan/HA bone scaffolds [[12], [13], [14]] or agarose/HA biomaterials [[15], [16], [17]].

To obtain macroporous scaffold with a network of interconnected pores, two simple and cost-effective methods were combined: freeze-drying and gas-foaming. Very often, complex methods [[18], [19], [20], [21]] were employed to obtain scaffolds for tissue engineering. Implementation of simple scaffold fabrication methods may be a key feature to allow translation of the biomaterials to the biomedical sector. It is worth noting that not only chemical composition of fabricated biomaterials has characteristics of novelty but also the method for the scaffold production itself is innovative as that has been shown in the Polish patent application No. P.426788 [22]. The secondary goal of this research was to comprehensively assess biomedical potential of the novel macroporous chitosan/agarose cryogel-based biomaterials via determination of their mechanical properties, porosity, physicochemical properties, biodegradation, bioactivity (scaffold biomineralization in vitro), and cytotoxicity in vitro.

2. Experimental

2.1. Biomaterials fabrication

Bone scaffolds were prepared by mixing the blend of 2 wt.% chitosan (75 %-85 % deacetylation degree, viscosity ≤ 300 cP, 50-190 kDa molecular weight, Sigma-Aldrich Chemicals) and 5 wt.% agarose (low EEO, gel point 36 ± 1.5 °C, Sigma-Aldrich Chemicals) in 2 % acetic acid solution (Avantor Performance Materials) with nanoHA (particle size < 200 nm, Sigma-Aldrich Chemicals) at low concentration of 40 wt.% (sample marked as chit/aga/HA_L) or at high concentration of 70 wt.% (material marked as chit/aga/HA_H). Then, sodium bicarbonate (NaHCO₃, Sigma-Aldrich Chemicals) was added as a foaming agent and the obtained paste was transferred into cylinder-shaped moulds, which were immersed for 15 min. in a water bath at temperature of 95 °C. Then, the samples were cooled and frozen in a liquid nitrogen vapour phase. Frozen samples were lyophilised (LYO GT2-Basic) at a medium vacuum (6 × 10² mbar) for a period of 20 h and the resultant scaffolds were neutralized in 1 % NaOH solution (Avantor Performance Materials), rinsed with deionised water, and left to dry in air for 24 h. Porosity of the biomaterials was achieved due to the solvent sublimation and carbon dioxide production as a result of the NaHCO₃ reaction with acetic acid:

NaHCO₃ + CH₃COOH → CH₃COONa + H₂O + CO₂↑ (1)

and also during heat treatment (95 °C):

NaHCO₃ → Na₂CO₃ + H₂O + CO₂↑ (2)

In the case of Fourier Transform Infrared spectroscopy (FTIR) analysis, additionally cryogel chitosan/agarose matrix (without nanoHA) as well as pure chitosan and agarose matrices were prepared in an analogous manner to the chit/aga/HA biomaterial production. Briefly, polysaccharides in acetic acid were mixed with NaHCO₃, incubated at 95 °C in a water bath, frozen, lyophilised, neutralized in NaOH, washed with deionised water, and air-dried. Prior to all experiments, the samples were sterilized by ethylene oxide.

2.2. Microstructure characterization

The specific surface area (SSA) of the scaffolds was evaluated by Nitrogen adsorption using the Brunauer-Emmett-Teller (BET) theory using an ASAP 2020 (Micromeritics). Mercury Intrusion Porosimetry (MIP, AutoPore IV, Micromeritics) was performed to determine the pore entrance size distribution (PESD) within the materials. Moreover, the open macroporosity was determined as the integral of the MIP PESD for pore diameters greater than 10 μm. Four to five cylindrical samples were introduced in the sample holder for the measurement, and a single measurement was performed for each composition. The porosity of the scaffolds was also evaluated by micro-computed tomography—microCT (Skyscan 1174, Bruker microCT) with the resolution of 12 μm. The set of images (a total of 360 pieces) was reconstructed into cross-section images of the scaffolds using NRecon software (Bruker microCT) to determine pores diameter as well as total, open and closed porosity using CTAnalyser software (Bruker microCT). Surface microstructure and topography of the biomaterials were visualized by scanning electron microscopy (SEM, FEI Nova NanoSem 450) and stereoscopic microscope (Olympus SZ61TR). Specimens for SEM imaging were sputtered under high vacuum conditions with gold layer of thickness about 20 nm.

2.3. Compression test

Three separate cylinder-shaped composites of 8 mm in diameter and 8 mm in length were subjected to compression testing. The compression behaviour was determined using an Autograph AG-X plus (Shimadzu) testing machine with pre-load value of 1 N, crosshead moving speed 10 mm/min followed by basic load rate 0.5 mm/min. The maximum compression stress was calculated at 30 % of strain. The obtained data allowed the determination of compressive strength values and Young's modulus.

2.4. Evaluation of chemical properties

FTIR evaluation was done on a Tensor 27 (Bruker Optics) medium infrared spectrometer, operating in the mode of attenuated total reflectance (diamond crystal ATR, MIRacle single reflection, PIKE Technologies). ATR-FTIR was applied to find out whether there are any chemical interactions between the two polysaccharides and/or nanoHA. In this method, the spectra are recorder from as deep as 2 μm from the surface of the sample. The spectra were collected in the range of 4000-550 cm^-1, with a resolution of 4 cm^-1 and total of 64 scans was recorded and averaged for each measurement. Prior to experiments, the samples were desiccated at 40 °C for 24 h in vacuum drier (Vacucell 55 comfort). Afterwards, at least three separate specimens of each sample were tested and the presented results are representative of each material. Post measurements, the spectra were smoothed and the baseline was manually corrected using Opus 7.2 software. For better comparison, the spectra were normalized in the following range: 3900-2370 cm^-1 (region where OH bands are present). Furthermore, to reveal any possible differences between the spectra, operation of spectra subtraction was performed. For the matrices without the nanoHA, this was done by subtracting the spectrum of aga matrix from the spectrum of chit/aga matrix and the resultant difference was then compared to the spectrum of pure chitosan matrix. For the nanoHA-reinforced samples, this was done by subtracting the spectrum of HA from chit/aga/HA_H spectrum and comparing this to the spectra of chit/aga matrix and nanoHA.

2.5. Apatite-forming ability

The apatite-forming ability of the biomaterials was estimated in accordance with ISO 23317 procedure by soaking of the samples in a simulated body fluid (SBF) [23]. Scaffold discs were placed in polypropylene conical tubes, immersed in SBF (pH 7.4) with ion concentrations similar to those of human blood plasma, and incubated at 37 °C for 14 and 28 days. At the end of determined intervals, the chit/aga/HA_L and chit/aga/HA_H specimens were removed from the SBF, gently washed with deionised water and dried in a desiccator for subsequent analysis. The apatite-forming ability of specimens was examined using SEM equipped with Octane Pro EDS detector (EDAX). The Ca/P atomic ratio was calculated based on EDS data to prove the presence of apatite crystals. Additionally, SBF samples after incubation with biomaterials were collected on the 7th, 14th, 21 st, and 28th day to evaluate the concentrations of Ca²⁺ and HPO₄^2- ions by colorimetric methods using calcium and phosphorus detection kit, respectively (BioMaxima).

2.6. In vitro biodegradation test

The in vitro biodegradation test was performed in an enzymatic solution according to a procedure reported in the literature with small modifications [24] as well as in accordance with the international procedure for ceramic materials described in ISO 10993-14 [25]. Enzymatic biodegradation test was performed using the enzymatic solution composed of 10 μg/mL lysozyme (Sigma-Aldrich Chemicals) and 160 μg/mL collagenase (Gibco) prepared in a phosphate buffered saline (PBS, Pan-Biotech GmbH). In parallel, PBS without the addition of enzymes was used to investigate non-enzymatic degradation of the scaffolds (control). Briefly, the samples were placed in polypropylene conical tubes and immersed in the enzymatic solution or PBS (maintaining the ratio 20 mg sample/10 mL solution). The tubes were incubated at 37 °C with mild shaking (60 rpm) for 2, 4, and 6 weeks. In the case of experiments performed according to ISO 10993-14, biomaterials were granulated to the grains size higher than 300 μm but lower than 400 μm (ISO 10993-14 recommendations). Granulated specimens were placed in polypropylene conical tubes and immersed in the citric acid buffer (C.A., pH 3, Avantor Performance Materials) - extreme solution test and Tris-HCl buffer, pH 7.4 (Sigma-Aldrich Chemicals) - simulation solution test (maintaining the ratio 250 mg sample/5 mL solution). The tubes were incubated at 37 °C with shaking (120 rpm) for 5 days. When incubation of biomaterials in enzymatic solution, PBS, C.A., and Tris-HCl was completed, the samples were removed by filtration and biodegradation behaviour of the scaffolds was assessed by detection of their degradation products in the filtrate as well as by elemental analysis of the dry retentate using a carbon, hydrogen, and nitrogen (CHN) analyzer (PerkinElmer CHN 2400). The biodegradation of polysaccharide components of the scaffolds was evaluated by CHN analysis of retentate as well as by estimation of reducing sugars in filtrate using 3,5-dinitrosalicylic acid (DNS)-based (Fluka Chemika) colorimetric method according to the procedure described in the literature [26]. The biodegradation of hydroxyapatite component of the scaffolds was assessed by determination of the concentrations of Ca²⁺ and HPO₄^2- ions in the filtrate. Ca²⁺ concentrations were determined by an inductively coupled plasma optical emission spectrometer (ICP-OES, 720 ES, Varian). The sample introduction system consisted of a Conikal^® nebulizer, cyclonic spray chamber and three-channel peristaltic pump. The optimal operating parameters used for determination of the studied elements by the ICP-OES technique were as follows: power 1200 W, plasma argon gas flow rate 15 dm³/min, auxiliary argon gas flow rate 2.25 dm³/min, nebulizer argon gas flow rate 0.2 dm³/min, pump rate 12 rpm as well as analytical wavelength 396.847 for Ca²⁺. The ICP-OES instrument was calibrated using the ICP standards for Ca²⁺ ions. For the preparation of all standards and blank samples, the ultrapure nitric acid was used to avoid any matrix interference. HPO₄^2- concentrations were evaluated by colorimetric method using commercially available kit for the determination of phosphate ions (BioMaxima).

2.7. Liquid absorption ability

The liquid absorption ability of the biomaterials was conducted in accordance with the procedure described by Prabaharan et al. with own modifications [5,27]. The dry chit/aga/HA_L and chit/aga/HA_H samples were weighed and subsequently immersed in phosphate buffered saline (PBS) solution and human blood plasma (obtained from a healthy volunteer) at room temperature. At predetermined time intervals, the samples were removed from the solutions, wiped with filter paper and weighed. The liquid absorption ability of the biomaterials was demonstrated as the percentage of the weight increase (W_i) with time.

2.8. In vitro cell culture experiments

Mouse calvarial preosteoblast cell line (MC3T3-E1 Subclone 4, ATCC) and normal human foetal osteoblast cell line (hFOB 1.19, ATCC) were used in the experiments. MC3T3-E1 cells were cultured in alpha MEM medium (Gibco) with 10 % foetal bovine serum (FBS, Pan-Biotech GmbH), 100 U/mL penicillin, 0.1 mg/mL streptomycin (Sigma-Aldrich Chemicals), and incubated at 37 °C in 5 % CO₂ in air atmosphere. The hFOB 1.19 cells were cultured in a 1:1 mixture of DMEM/Ham's F12 medium without phenol red (Sigma-Aldrich Chemicals), with 10 % FBS, 100 U/mL penicillin, 0.1 mg/mL streptomycin, 0.3 mg/mL G418 (Sigma-Aldrich Chemicals) and incubated at 34 °C in 5 % CO₂ in air atmosphere (ATCC recommendations).

2.8.1. Cytotoxicity assessment according to ISO 10993-5

The cytotoxicity of the investigated biomaterials was determined according to ISO 10993-5 [28] by indirect methods: (1) agar diffusion test and (2) test on 24-h extracts of the scaffolds prepared as it was described earlier [29]. In the case of agar diffusion test, MC3T3-E1 and hFOB 1.19 osteoblasts were seeded into 24-well plate in 500 μL of a culture medium at a concentration of 3 × 10⁵ cells/mL and 2 × 10⁵ cells/mL, respectively. After 24 h incubation, culture media were replaced with the mixture of agar/culture medium, and then specimens of tested scaffolds (with surface area = 19.6 mm²—approx. 1/10 of the cell growth area) and control samples (latex material, positive control, polypropylene, negative control of cytotoxicity) were placed on the solidified agar. After 48-h incubation, the cells were observed under inverted optical microscope (Olympus CKX53) to determine changes in cell morphology, detachment or cell lysis under and around the tested materials. In the case of test on extracts, MC3T3-E1 and hFOB 1.19 cells were seeded into 96-well plates in 100 μl of a culture medium at a concentration of 2 × 10⁵ cells/mL and 1.5 × 10⁵ cells/mL, respectively. After 24-h incubation, the culture media were replaced with appropriate extracts (latex extract served as a positive control and polypropylene extract was a negative control of cytotoxicity) and cells were incubated for further 24 h and 48 h. Afterwards, MTT (Sigma-Aldrich Chemical) colorimetric assay was performed to determine cell viability as it was described previously [29]. The results were shown as the percentage of absorbance value obtained with the negative control revealing 100 % viability.

2.8.2. Cytotoxicity assessment in direct contact with scaffolds

Before the experiment, cylinder-shaped scaffold discs with approximately 2 mm in thick and 8 mm in diameter were placed in 48-well plates and preincubated overnight in appropriate complete culture medium. Subsequently, osteoblasts were seeded directly on the samples in 500 μL of culture medium at a concentration of 4 × 10⁵ cells/mL (MC3T3-E1) and 3 × 10⁵ cells/mL (hFOB 1.19). After 48 h of culture, cells growing on the surface of the scaffolds were stained using Live/Dead Double Staining Kit (Sigma-Aldrich Chemical) according to the manufacturer procedure. Viability of stained cells was assessed using confocal laser scanning microscope (CLSM, Olympus Fluoview equipped with FV1000).

2.9. Statistical analysis

Statistical analysis of the data was carried out using GraphPad Prism 8.0.0 Software. The results were expressed as mean values ± standard deviation (SD) and were representative of at least three independent experiments. An unpaired t-test or One-way ANOVA followed by Tukey’s multiple comparison test were used to determine statistical differences (P < 0.05) among groups.

3. Results

3.1. Microstructure characterization

Both kind of fabricated scaffolds were characterized by high SSA (approx. 30-31 m²/g) and a total open porosity of approx. 70 % based on MIP results (Table 1). Increasing percentage of nanoHA content in the scaffolds did not alter the specific surface area of the resultant biomaterials. Only a minor decrease in total porosity was recorded for chit/aga/HA_H sample compared to its counterpart with lower amount of nanoHA. Since microCT technique is not as precise as MIP method, the open porosity values obtained with microCT were lower in comparison to porosity values obtained with the MIP. However, microCT technique allowed also to characterize closed and total porosity of the scaffolds and to compare percentage of each type of porosity within the samples. MicroCT analysis revealed that chit/aga/HA_L was characterized by almost 2-fold higher open porosity compared to chit/aga/HA_H and slightly lower closed porosity.

Fig. 1(a) shows that both scaffolds display a bimodal pore entrance size distribution, with a peak centred around 50 nm, and a distribution of pores within a wide range of diameters, ranging between 6 and 300 μm. The chit/aga/HA_L sample displayed higher volume of pores in the micrometric range (pore entrance size diameter > 10 μm) than its counterpart with higher content of nanoHA. The cross-section images of chit/aga/HA_L and chit/aga/HA_H obtained with microCT scanning demonstrated their highly macroporous structure (Fig. 1(b)). The chit/aga/HA_L not only revealed an extensive interconnected pore network, but also showed more uniform distribution of pores (black areas on the microCT images) compared to chit/aga/HA_H. The sample with lower content of nanoHA had also greater share of pores with diameter in the range of 50-230 μm compared to the biomaterial containing 70 % nanoHA (Fig. 1(c)). Whereas the chit/aga/HA_H scaffold was characterized by the presence of higher amount of large pores with diameter > 500 μm compared to chit/aga/HA_L.

The stereoscopic microscope images and SEM surface micrographs of chit/aga/HA_L and chit/aga/HA_H scaffolds demonstrated differences in their microstructure and topography (Fig. 1(d)). Stereoscopic microscope observation showed the chit/aga/HA_L possessed more rough and ragged structure compared to the chit/aga/HA_H, which had more smooth and compact surface. Nevertheless, both scaffolds were characterized by high surface roughness and macroporosity.

3.2. Mechanical properties

Compressive strength values determined for fabricated biomaterials were as follows: 1.40 ± 0.18 MPa for chit/aga/HA_L and 1.09 ± 0.14 MPa for chit/aga/HA_H. Young’s modulus values estimated for the scaffolds were low: 13.98 ± 1.80 MPa for chit/aga/HA_L and 11.79 ± 1.77 MPa for chit/aga/HA_H, indicating their great elasticity.

3.3. Evaluation of chemical properties

ATR-FTIR investigation (Fig. 2(a)) revealed spectra typical of polysaccharides: a wide composite band between 3700 and 3000 cm^-1 (OH groups), a triplet with maxima at 2956, 2920 and 2850 cm^-1 (CH₃ and CH₂) and a variety of spectral modes apparent below 1750 cm^-1, indicating the presence of various double and single bonds between carbon, oxygen, nitrogen and hydrogen. The spectrum of chitosan had the same characteristics as reported previously in our study and had already been extensively evaluated, proving presence of both N-acetyl glucosamine and N-glucosamine units [30]. Spectrum of agarose was typical for this material: lack of characteristic bands attributed to NH functional groups in chitosan (most notably, two maxima at 3358 and 3294 cm^-1) and presence of some of the distinctive bands, namely: a shoulder at 2700 cm^-1, indicative of CHO group (most likely in 3,6-anhydrogalactose), a doublet at 1065 and 1041, attributed to C—O—H and C—O—C (glycosoid linkage) [31], respectively, and bands at 930 and 889 cm^-1 indicative of galactose units in agarose [32].

The spectrum of chit/aga matrix was majorly the sum of the chit matrix and aga matrix spectra with some exceptions. The bands attributed to CH₃ and CH₂ were sharpened, the shoulder at 2700 cm^-1 in agarose disappeared, the shape of 1563 cm^-1 was altered (pyranose ring in agarose, NH bending in chitosan [33]). Interestingly, no shift towards higher frequencies of the NH₂ and OH bands was observed, which was detected by Felfel et al. [33] and was said to be indicative for the formation of hydrogen bonds between the NH₂/OH groups of chitosan and the OH groups of agarose. The above-mentioned findings were furthermore confirmed by spectral subtraction (Fig. 2(b)), where the obtained spectrum of subtract (spectrum of chitosan in the chit/aga matrix) had apparent differences from the original spectrum of chitosan. Subtraction revealed a significant blue shift of the 1563 cm^-1 band towards higher wavenumbers.

Spectra of HA revealed bands typical of carbonated hydroxyapatite, with bands attributed to phosphate (1086, 1023, 961, 601 and 568 cm^-1), OH groups (3574 and 630 cm^-1) and CO₂^3- doublet at 1459 and 1414 cm^-1 and band at 876 cm^-1) [34]. ATR-FTIR investigation of chit/aga/HA_L and chit/aga/HSA_H scaffolds resulted in the obtainment of spectra that were mainly HA (Supplementary material 1), with increased intensity of 601 and 568 cm^-1 bands, attributed to phosphate. To better visualize the changes in the spectra induced by mixing of the compounds, spectral subtraction was performed (spectrum of HA was subtracted from the chit/aga/HA_H spectrum). The obtained spectrum was then compared against chit/aga matrix and pure HA (Fig. 2(c)). It was clearly seen that HA attributed bands were not entirely eliminated from the subtract spectrum. Still, upon subtraction changes in the triplet with maxima at 2956, 2920 and 2850 cm^-1 (CH₃ and CH₂) and in the shape and relative intensity of the bands at 1041 (C—O—C) [31] and 995 cm^-1 (C—O, C—C) [30] could be found.

3.4. Apatite-forming ability

Both biomaterials had the ability to induce apatite formation with globular and hemispherical morphology already after 14-day incubation in SBF (Fig. 3(a)). EDS analysis confirmed that observed crystals were calcium phosphate apatite since Ca/P atomic ratio of the precipitates on the surface of chit/aga/HA_L and chit/aga/HA_H was equal to 2.23 ± 0.10 and 2.19 ± 0.96, respectively. On the 28th day of the experiment, surfaces of biomaterials were covered by numerous apatite clusters with Ca/P atomic ratio equal to 1.68 ± 0.11 for chit/aga/HA_H scaffold, which was close to the stoichiometric value of 1.67 for natural HA [35]. However, the Ca/P atomic ratio of apatite crystals formed on the surface of chit/aga/HA_L remained unchanged. Fig. 3(b) shows time-dependent changes in Ca²⁺ and HPO₄^2- concentrations in the SBF during incubation with the biomaterials. Typical fluctuations in the concentration of calcium ions as well as complete uptake of phosphate ions were observed. These changes in ionic concentrations of SBF were well correlated with the formation of apatite crystals on the scaffolds.

3.5. In vitro biodegradation test

The degree of biodegradation of polysaccharide matrices of the scaffolds is shown in Table 2. After 2-week incubation of biomaterials in enzymatic solution, the content (%) of CHN elements in the chit/aga/HA_L and chit/aga/HA_H samples was significantly lower compared to the corresponding untreated controls, indicating matrix degradation. However, CHN content in enzymes-treated samples did not decrease with increasing incubation time and remained similar throughout the whole duration of the experiment. CHN analysis revealed also non-enzymatic degradation of the scaffolds since CHN content in the biomaterials after incubation in PBS (without enzymes) was slightly lower compared to the untreated controls. The analysis of retentate showed also higher polysaccharides loss rate for chit/aga/HA_L compared to the chit/aga/HA_H sample.

Detection of the reducing sugars in filtrates revealed progressive polysaccharide matrix degradation with time at similar rate for both biomaterials. The experiment demonstrated also non-enzymatic matrix degradation in PBS and simulation solution (Tris-HCl) at significantly lower rate when compared to the enzymatic solution or the citric acid extreme solution (C.A.). The highest concentration of reducing sugars was detected in filtrates collected after incubation of biomaterials in C.A. Moreover, chit/aga/HA_L biomaterial showed significantly faster biodegradation of the polysaccharide matrix in acidic solution compared to the scaffold with higher content of bioceramics.

The biodegradation of the nanoHA of the scaffolds was assessed by examining the concentrations of Ca²⁺ and HPO₄^2- ions in the solutions after incubation with biomaterials (Fig. 4). The results showed uptake of Ca²⁺ and HPO₄^2- ions by the samples incubated in the PBS-based enzymatic solution and in the control non-enzymatic PBS solution during long-term experiment (Fig. 4(a)). Uptake of these ions was most likely caused by spontaneous apatite formation on the surfaces of the scaffolds (which was demonstrated in apatite-forming ability test, Fig. 3), making it impossible to determine nanoHA degradation process in the PBS-based solutions. Nevertheless, 5-day incubation of the scaffolds in Tris-HCl - performed according to ISO 10993-14 - resulted in a slight release of HPO₄^2- and in a significant release of Ca²⁺ ions, indicating degradation of the biomaterials at physiological pH of 7.4 (Fig. 4(b)). Incubation of the samples in the C.A. extreme solution caused the highest dissolution of the ceramics and thus significant release of HPO₄^2- and Ca²⁺ ions (Fig. 4(b)). Apart from HPO₄^2- release after incubation in Tris-HCl, there were no significant differences between the chit/aga/HA_L and chit/aga/HA_H samples regarding degradation rate of the bioceramics component.

3.6. Liquid absorption ability

Liquid absorption ability test was conducted in two different physiological solutions: PBS and human blood plasma (Fig. 5). The results obtained showed that the immersion solution did not influence liquid retention ability of the biomaterials. The chit/aga/HA_L scaffold reached absorption equilibrium already after 12 s of soaking in PBS (W_i = 189.45 % ± 22.74) and plasma (W_i = 168.60 % ± 30.56). The sorption tendency of chit/aga/HA_H was much lower than chit/aga/HA_L. Scaffold with higher content of nanoHA not only absorbed significantly less liquid than chit/aga/HA_L, but also reached absorption equilibrium after 360 s of soaking in PBS (W_i = 111.49 % ± 10.67) and plasma (W_i = 117.89 % ± 12.24).

3.7. In vitro cell culture experiments

Indirect agar diffusion test showed that biomaterials were non-toxic to the eukaryotic cells. MC3T3-E1 and hFOB 1.19 osteoblasts cultured around and under the specimens had normal morphology and were well adhered to the polystyrene surface. Moreover, no signs of cell lysis were observed (Fig. 6(a)). Results obtained with MTT test confirmed non-toxicity of the scaffolds against MC3T3-E1 and hFOB 1.19 cells (Fig. 6(b)). Viability of osteoblasts exposed to the extracts of the biomaterials for 24 h was near 100 %, whereas 48 h exposure time resulted in slight reduction in cell viability to approx. 80 %-90 %. Nevertheless, according to ISO 10993-5, 100 % extracts of the biomaterials causing reduction in cell viability by less than 30 % should be considered as non-toxic. Results obtained with live/dead staining of MC3T3-E1 and hFOB 1.19 cells were consistent with indirect cytotoxicity tests. CLSM images showed a great number of viable cells (green fluorescence) and only several dead cells (red fluorescence) on the surface of the scaffolds (Fig. 6(c)). Osteoblasts were well flattened indicating their good adhesion to the surface of chit/aga/HA_L and chit/aga/HA_H.

4. Discussion

In this study, we demonstrated that simultaneous application of relatively simple methods - freeze-drying (F-D) and gas-foaming (G-F) - allows introduction of pores in the material structure to obtain macroporous bone scaffolds (diameter of pores > 50 μm, Fig. 1(c)) characterized by high open and interconnected porosity (Table 1, Fig. 1(b)). It is worth to emphasize that we are the first who combined a foaming agent with freeze-drying technique to synthesise highly macroporous biomaterials. Most researchers apply mentioned methods separately for the fabrication of porous biomaterials. However, although the resultant materials (produced by solely F-D or G-F technique) reveal high total porosity, it is very often closed porosity which does not have much biomedical importance [33]. However, Catanzano et al. [36] developed macroporous alginate foams crosslinked with strontium via combination of internal gelation with gas foaming agent (sodium bicarbonate) followed by freeze-drying technique. Resultant materials were characterized by interconnected pore structure and porosity in the range of 35 %-50 % depending on the calcium and strontium molecular ratio used during the production. Unfortunately, there is no information about the type (open, closed) of studied porosity. In our preliminary research, solely the F-D method was applied for chit/aga/HA_L production and the obtained scaffold had high total porosity of 60 %, but it was mostly closed porosity (50 %). Moreover, the biomaterial exhibited poor interconnected porosity (Supplementary Materials 2 and 3).

Open and interconnected porosity are very important parameters from the biomedical application point of view since they provide high growth surface area for osteoblasts and mesenchymal stem cells as well as enable good material vascularisation and bone ingrowth deep into the implant, accelerating osseointegration after scaffold implantation. Fabricated biomaterials showed high degree of total open porosity—approx. 70 % (Table 1), which was similar to the porosity range (30 %-95 %) occurring in natural human cancellous bone [37]. Importantly, the increase in the content of nanoHA ceramics in the scaffolds significantly reduced their open (from 39 % to 19 %) and interconnected porosity (Table 1, Fig. 1(b)). This was probably related to the thicker paste formed after mixing of the liquid phase (blend of chitosan and agarose) with nanoHA, which may have hindered gas release and pore formation through G-F technique. It should be also noted that biomaterials fabricated within this study exhibited the highest share of pores with diameter in the range 100-410 μm, which were demonstrated to be essential for good implant osseointegration [38,39], as it is suitable size for cell colonization of the scaffold. In contrast, pores with diameter <75 μm favour penetration of fibrous tissue, causing implant encapsulation [4,7].

Simultaneous application of F-D and G-F methods also allowed to obtain pores with entrance size diameters ranged from 10 to 100 nm (Fig. 1(a)), which are known to influence the bioactivity of the materials. It was demonstrated that pores with diameters below 1 μm provide higher surface roughness of the biomaterial, enhancing its apatite-forming and protein adsorption ability [38,39]. Although both fabricated scaffolds had similar pore share with entrance size diameter in the range of 10-100 nm (Fig. 1(a)), microstructure visualization revealed more rough and ragged surface of chit/aga/HA_L biomaterial (Fig. 1(d)). Apart from surface topography specific surface area is also known to have great impact on biological response [40]. Both fabricated bone scaffolds showed similar and relatively high SSA (approx. 30 m²/g). For comparison, BioOss^® material made of 100 % cancellous bone and OsteoBiol^® implant containing up to 80 % cancellous bone were demonstrated to possess SSA equal to 59.7 m²/g and 42.4 m²/g, respectively [41].

Biomaterials for application in bone tissue engineering should possess appropriate mechanical strength to provide sufficient support for the formation of new bone tissue at the site of implantation. Young’s modulus values of the scaffolds (approx. 12-14 MPa) were significantly lower compared to human cancellous bone (0.1-0.5 GPa) [37], whereas compressive strength values determined for produced biomaterials (approx. 1-1.4 MPa) were in the lower range of the compressive strength value of cancellous bone (1.9-12 MPa). The low Young’s modulus values obtained for the developed scaffolds were related to their high elasticity resulting from the polysaccharide content. It should be noted that elastic materials may exert physical pressure at the implantation site, generating mechanical signal inducing bone formation. However, prolonged exposure of elastic material to the mechanical loads may cause excessive bone formation resulting in biomaterial failure [5,42]. Thus, the present bone scaffolds should rather be used in non-load bearing implantation areas or in combination with screws, plates, wires, or pins to provide better stiffness and stability [43].

Engineering of biomaterials distinguishes 2 types of materials: composites and hybrids. Materials are classified as composites when two or more phases (often organic and inorganic) are simply mixed. Hybrid materials reveal chemical interactions between the components. Hybrids showing weak bonds (e.g. hydrogen, van der Waals) are classified as class I hybrids, whereas materials exhibiting strong chemical interactions (e.g. covalent) between the phases are defined as class II hybrids [7]. Hybrids have an advantage over composites since they might reveal better stability, mechanical properties, and biodegradation rate. ATR-FTIR evaluation revealed that there were chemical interactions between the two polysaccharides and between the chit/aga matrix and nanoHA (Fig. 2). The polysaccharides most likely interacted via their side groups: C—O—C, C—H—O and C—OH of agarose supposedly bonded with NH and -CONH of chitosan. As one of the most distinguishable change in the spectra was disappearance of the C—H—O band from the spectrum of agarose (2700 cm^-1), it is suggested that mixing of the two polysaccharides resulted in breaking of the bond between C and O in the 3,6-anhydrogalactose unit. This in turn should result in creation of negative charge on oxygen and positive charge on CH₂ unit, which should readily react with positively charged NH and negatively charged OH groups, respectively, via covalent bonds. Another important observation was shifting of the 1563 cm^-1 band upon mixing towards higher frequencies (in chitosan this band originated mainly from the C—N stretching vibrations and N—H in-plane bands and a blue-shift indicated formation of weak hydrogen bonds between those and functional groups of agarose - most likely OH, leading to the shortening of the N—H bond [44]). Thus, it can be implied that agarose and chitosan interacted with each other with both covalent and hydrogen bonds, suggesting that they were able to form class II hybrid. When it comes to interactions with nanoHA, an important observation was significantly increased intensity of the bands attributed solely to phosphates (601 and 568 cm^-1). Such observation suggests increased polarity of these functional groups, indicating presence of either covalent of ionic interactions between this group and the groups in the polysaccharides mixture. There were also possible interactions between the C—O, C—O—C and C—OH groups in polysaccharide and OH groups in nanoHA, via hydrogen bonds. All these observations suggest that the chit/aga/HA scaffolds might be classified as class II hybrid.

The ability of bone scaffolds to form a bone-like apatite layer on their surfaces is an important issue since it provides good osseointegration of the implant with host bone after its implantation into a living organism. Bioactivity of biomaterials can be easily predicted in vitro by detection of apatite crystals formed on their surfaces upon long-term incubation in SBF solution, which has similar chemical composition of inorganic ions to human blood plasma [45]. SEM and EDS analysis demonstrated that both fabricated biomaterials induced apatite formation on their surfaces after incubation in SBF for 14 and 28 days (Fig. 3(a)). Formed apatite precipitates had globular and hemispherical morphology, which was similar to the morphology of granular apatite layer observed on the surface of silica by other authors [46]. It is known that the stoichiometric value of Ca/P for HA is equal to 1.67 [35]. After 14-day incubation in SBF, Ca/P atomic ratio of precipitates detected on the surface of chit/aga/HA_L and chit/aga/HA_H was in the range of 2.19-2.23, indicating that Ca-rich amorphous calcium phosphate (ACP) was formed. According to the available literature after short incubation period in SBF, Ca-rich ACP is formed on the ceramics-based biomaterials as a result of interactions between the Ca²⁺ ions in the SBF and the negatively charged hydroxyapatite surface. Nevertheless, prolonged incubation in SBF results in the conversion of ACP into bone-like apatite by incorporation of phosphate ions [47]. In our studies this phenomenon was observed only for chit/aga/HA_H scaffold as EDS analysis revealed Ca/P atomic ratio of apatite precipitates equal to 1.68 after 28-day soaking in SBF. However, Ca/P atomic ratio of precipitates on the surface of chit/aga/HA_L remained unchanged, indicating reduced potency to adsorb phosphate ions from the SBF.

Bioactivity of the developed scaffolds was also confirmed by analysis of changes in the ionic composition of SBF with respect to Ca²⁺ and HPO₄^2- ions (Fig. 3(b)). The experiment demonstrated fluctuations in the concentrations of Ca²⁺ ions which are typical of bioceramics-based materials. Release of Ca²⁺ ions from the scaffolds probably resulted from the dissolution of nanoHA or from the ionic substitution of calcium by other ions occurring in SBF, whereas Ca²⁺ uptake from SBF was most likely associated with apatite formation [35,48,49]. During bioactivity test, SBF was almost completely deprived of HPO₄^2- ions what possibly resulted from either ion exchange events or electrostatic interactions between the phosphate ions in the SBF and the charged surfaces related to apatite nucleation on the surface of biomaterials [50].

In the present work, biodegradation of the developed scaffolds was also assessed under differing in vitro conditions: (1) enzymatic degradation in the presence of lysozyme and collagenase, (2) non-enzymatic degradation in PBS and Tris-HCl at physiological pH of 7.4, and (3) degradation in acidic environment (citric acid, pH 3) that may occur during osteoclast-mediated implant resorption. The optimal biodegradation rate is essential for good implant osseointegration since it provides space for growth of newly formed bone and contributes to enhanced ion-exchange at the implantation site [7]. The chit/aga/HA_L revealed significantly higher degree of matrix degradation compared to the chit/aga/HA_H material only in the acidic environment. This could be most likely related to the higher content of chitosan and agarose in the chit/aga/HA_L sample (Table 2). Importantly, the developed materials had the ability to biodegrade in non-enzymatic solutions (PBS and Tris-HCl) under physiological conditions (pH 7.4) (Table 2, Fig. 4). However, addition of enzymes to PBS significantly enhanced biodegradation of polysaccharide components of the biomaterials (Table 2). Ability of scaffolds to degrade under physiological conditions and in the presence of enzymes (including matrix metalloproteinases) is very important from the clinical point of view since it enables the correct course of bone remodeling process and gradual replacement of the implant with newly formed bone tissue. Bone resorption process during bone remodeling is induced by osteoclasts that acidify the surrounding microenvironment and dissolve the organic matrix as well as mineral component of the bone [51]. Osteoclast-mediated biomaterial resorption is a crucial step for good implant osseointegration. In this work, it was proved that acidic environment significantly accelerated degradation of the biomaterials, indicating that developed scaffolds are prone to osteoclast-mediated resorption.

The liquid absorption/retention ability of the biomaterials is also an important parameter which influences physiological fluids absorption and flow of nutrients and metabolites within the implanted biomaterial [52]. Both chit/aga/HA_L and chit/aga/HA_H were highly absorbent and required short time of soaking to reach absorption equilibrium (Fig. 5). However, due to the greater share of smaller pores (with diameter ranged from 15 to 150 μm) in the structure of chit/aga/HA_L compared to chit/aga/HA_H, its sorption tendency was much higher than that of chit/aga/HA_H sample. Interestingly, no significant differences between PBS and blood plasma absorption ability were observed. Soaking of biomaterials in blood plasma is a typical step required in the pre-operative procedure that should be relatively short to avoid implant contamination and potential post-surgery infections. Great PBS absorption ability of the scaffolds may suggest that they are promising candidates to be used as effective drug delivery carriers.

Cell culture in vitro experiments revealed great potential of chit/aga/HA_L and chit/aga/HA_H as biomaterials for regenerative medicine applications. In vitro tests performed according to ISO10993-5 proved non-cytotoxicity of the scaffolds against murine (MC3T3-E1) and human (hFOB 1.19) osteoblasts. Moreover, the microstructure and composition of the biomaterials were favourable to cell attachment and growth (Fig. 6), indicating that scaffolds would support osteoblasts-mediated bone formation. It is worth emphasising that in our previous work [53] we proved that chit/aga/HA_L and chit/aga/HA_H scaffolds were characterized by high biocompatibility and osteoconductivity since surface of the biomaterials was prone to protein adsorption (primarily fibronectin), favouring osteoblast adhesion, spreading and proliferation. Moreover, we demonstrated that novel scaffolds had osteoinductive properties because they revealed ability to induce differentiation of mesenchymal stem cells derived from bone marrow (BMDSCs) and adipose tissue (ADSCs) into osteoblastic cell lineage, supporting bone formation process.

It should be also noted that according to available literature, pores interconnectivity and large pore size (>100 μm) are essential factors for new blood vessels formation and thus implant vascularisation [[54], [55], [56]]. Xiao et al. [54] and Somo et al. [55] demonstrated that interconnected structure of biomaterials with pore size of approx. 150 μm significantly improved scaffold vascularization in vitro and in vivo. In this study, we fabricated biomaterials which are highly macroporous and possess open and interconnected porosity. Therefore, it may be hypothesized that the scaffolds would support formation of new blood vessel network and neovascularization after their implantation. However, further in vitro studies with the use of endothelial cells and in vivo studies on animal model are needed to be performed to confirm this assumption.

5. Conclusion

The present work demonstrated that combination of cost-effective and relatively simple methods (gas-foaming and freeze-drying) allows production of hybrid highly macroporous biomaterials characterized by open and interconnected porosity. Bone scaffolds fabricated by simultaneous application of a gas foaming agent and lyophilisation process show high share of macropores, which are crucial for implant vascularisation and new bone ingrowth. Novel scaffolds are non-toxic to the cells, favour osteoblasts adhesion and growth and induce apatite formation on their surfaces, indicating their high bioactivity that is essential for good implant osseointegration. The novel biomaterials are biodegradable, being prone to enzymatic degradation, osteoclast-mediated degradation in acidified microenvironment as well as slow degradation under physiological pH of 7.4. Moreover, the scaffolds reveal microstructure (70 % open porosity, SSA approx. 30 m²/g, high share of macropores with diameter in the range 100-410 μm) and compressive strength (approx. 1-1.4 MPa) comparable to cancellous bone. All the mentioned features make these biomaterials promising implants for cancellous bone regeneration and reconstruction. Nevertheless, increasing the content of nanoHA in the structure of the scaffold reduces its open and interconnected porosity, slightly worsens its biodegradation, and significantly decreases liquid absorption ability. Thus, chit/aga/HA scaffold with lower content (40 %) of bioceramics is more appropriate for biomedical applications, suggesting that finding the optimal concentration of the nanoHA is a necessity in producing this new class of advanced biomaterials.

Acknowledgements

The work was financially supported by the National Science Centre (NCN) in Poland within OPUS 16 (No. UMO-2018/31/B/ST8/00945). Analysis (ATR-FTIR) performed by Aleksandra Benko was supported by the National Science Centre (NCN) in Poland (No. UMO-2017/24/C/ST8/00400). The paper was developed using the equipment purchased within agreement No. POPW.01.03.00-06-010/09-00 Operational Program Development of Eastern Poland 2007-2013, Priority Axis I, Modern Economy, Operations 1.3. Innovations Promotion. Authors acknowledge the Spanish Government for financial support through Project PCIN-2017-128/AEI and Ramon y Cajal fellowship of CC. Authors also acknowledge M. Molmeneu for her technical support.

---