Preparation, physicochemical properties and in vitro bioactivity of hierarchically porous bioactive glass scaffolds

Abstract

Exploring bioactive scaffolds with the desired architecture and high osteogenic capacity has been a challenging issue for biomaterials and bone tissue engineering. Herein, a hierarchically porous bioactive glass scaffold (noted as YMBG) has been prepared by using P123, yeast cells and polyurethane sponges as templates. This synthesis strategy created high porosity scaffolds with interconnected macropores (300–500 μm), well-ordered mesopores (∼5 nm) and mid-level pores (20 nm to 2 μm). The utilization of yeast cells as a midpore biotemplate in the scaffolds significantly promoted the crystallization and improved their mechanical strengths. In addition, the prepared scaffolds could support attachment and proliferation of BMSCs, and more importantly, significantly improved the expression of bone-related genes (Runx2, OCN, Col I and BMP2). The results suggest that the novel bioactive glass scaffolds with widely-distributed porous structures, which can be employed as an effective carrier for a variety of therapeutic molecules and tissue cells, are expected to be a promising biomaterial for bone regeneration and tissue engineering.

---

1. Introduction

The effective treatment of bone defects, especially large bone defects resulting from congenital malformation or acquired deformity (infection, trauma, tumor, etc.), has been a tough problem clinically.¹ An alternative and promising approach to solve these issues is implantation of bioactive scaffolds in situ for bone tissue regeneration. Generally, bioactive scaffolds with three-dimensionally interconnected porous structures provide a useful platform for cell adhesion, migration and proliferation, which could promote cell ingrowth, nutrient delivery and be ultimately replaced by regenerating tissues.^1,2 An ideal scaffold for bone repairing should possess multifunction, such as angiogenic, osteogenic stimulation, and even drug delivery with certain therapeutic properties.³ However, how to develop such multifunctional scaffolds still remains challenging. Previous studies have revealed that bioactive composition, porous architecture and surface property are essential for tissue engineering scaffolds.^1,4,5 Therefore, tailoring of composition, microstructure and surface chemistry makes it possible to design multifunctional scaffolds for bone tissue engineering.

Bioactive glass (BG) has been an attractive material for bone regeneration owning to its intrinsic biocompatibility,⁶ osteoconductivity⁷ and osteostimulation.⁸ Its texture and chemical composition are easily tuned over a wide range by changing components^4,9–11 and processing techniques^10,12–17 for enhancing the bioactivity. Mesoporous bioactive glass (MBG) scaffolds were previously prepared by using polyurethane sponges and P123 as co-templates through an evaporation-induced self-assembly (EISA) process, which exhibited a hierarchical structure with interconnected macropores (300–500 μm) and uniform mesopores (∼5 nm).^10,13 Further studies have demonstrated that the macropores could support tissue ingrowth⁹ while the mesopores provide high specific surface area and large pore volume, which could not only enhance bioactivity via releasing of ionic products,⁴ but also serve as a carrier for growth factors and therapeutic drugs to the defect sites.^3,18 As proved in many previous studies, the pore size of biomaterials has a great influence on the enzyme adsorption and the suitable pore size was found to be 2–5 times that of the protein molecules.^19,20 However, the size of most osteogenic growth factors is larger than 5 nm. For example, collagens (type I, II, III, etc.) has been used to promote osteoblast adhesion, proliferation and differentiation,²¹ whose molecules consist of an continuous triple helix of approximately 300 nm in length and 1.5 nm in diameter.²² For better delivery of macromolecule drugs, a specific pore-size range is needed to physically encapsulate them inside the scaffolds.

Biotemplates from microbial cells (e.g., yeast, cyanobacteria and Lactobacillus bulgaricus) have been a novel source for designing and synthesizing porous nanostructures.^23–25 The negatively charged functional groups on their cell wall, such as carboxyl, hydroxyl, amide and phosphate groups, can be combined with positively charged metal irons. The self-assembly behavior induced the formation of metal cation layers and these network structure could be reserved after calcination.^23,26–28 Recently, He et al. synthesized hierarchically porous bioactive phosphosilicate glasses with open mesopores (2–40 nm) and the highly curved macropores (200–500 nm) by using yeast cells as a biotemplate.²⁸ Since the surface proteins and polysaccharides are negatively charged, the positively charged colloidal particle can spontaneously assemble on the surface of yeast cells by electrostatic interaction when mixed with the yeast cell solution. The biotemplates were then removed via heat treatment, leaving the nanovoids (size: 200–500 nm) in the matrix of materials. These proper-sized macropores significantly enhanced enzyme immobilization efficiency and its catalytic activity. Compared with other fabrication processes, the self-assembly behavior of yeast cells provides an easier, greener and more economical way to create macropore structures.

Bioactive glass scaffolds, which possess the highly-porous feature with a pore size in the range of 5 nanometers to 500 micrometers and 3D interconnected pore structures, may have the potential to satisfy the requirement of multifunction in bone regeneration applications. There have been enormous reports showing that the morphology, pore structure and pore size of scaffolds have great influence in their physicochemical properties and cell behaviors such as migration, proliferation and differentiation.^{7,14,15,29,30} However, to the best of our knowledge, there is no report about the fabrication of hierarchically porous BG scaffolds with pore sizes continuously distributed in the range of 5 nm to 500 μm, and the mechanical properties and bioactivity of these novel porous scaffolds are unknown. Herein, a hierarchically bioactive glass scaffold with macropores (300–500 μm), midpores (20 nm to 2 μm) and mesopores (∼5 nm) was successfully prepared by using tri-template (P123, yeast cells and polyurethane sponges) method. The porous structures and physicochemical properties of the prepared scaffolds were systematically analyzed and the in vitro osteogenic activity was further investigated.

2. Materials and methods

2.1 Fabrication of hierarchically porous YMBG scaffolds

Hierarchically porous YMBG scaffolds were prepared using tri-templates of non-ionic block polymer P123 (EO20-PO70-EO20), polyurethane sponges and yeast cells (Scheme 1). P123 as a soft template for the formation of mesoporous structures (pore size: ∼5 nm) and polyurethane sponges as a hard template for macroporous structures (pore size: 300–500 μm) were reported in our previous publication.³ The yeast cells were employed in the sol–gel process to produce mid-level pores (20 nm to 2 μm). For instance, to prepare MBG scaffolds containing 5 wt% yeast cells (yeast/TEOS mass ratio = 5%), 4.0 g of P123 (M_w = 5800, Sigma-Aldrich, Germany), 6.7 g of tetraethyl orthosilicate (TEOS, 98%), 1.4 g of Ca(NO₃)₂, 0.73 g of triethyl phosphate (TEP, 99.8%) and 2 g of 0.5 M HCl were dissolved in 60 g of ethanol (molar ratio Ca/P/Si = 15/5/80) at room temperature while 0.17 g of yeast cells (Angel Yeast Co., Ltd.) were nurtured in 2.5 g deionized water at 35 °C for 25 min. The two liquids were mixed and stirred at room temperature for 24 h. Polyurethane sponges (25 ppi) were cleaned and dried firstly, and then completely immersed into the aforementioned liquid mixture and compressed to force the sol migrate into all pores, and finally transferred to a smooth dish to undergo a solvent evaporation process at room temperature for 12 h. This procedure was repeated for 3 times. After completely drying, the raw porous bodies were calcined at 650 °C for 5 h with a heating rate of 1 °C min⁻¹ to remove the templates. The obtained scaffolds were named as 5YMBG. MBG scaffolds without yeast cells (named as 0YMBG), 15 wt% yeast cells (named as 15YMBG) and 30 wt% yeast cells (named as 30YMBG) were prepared by the same method with that of 5YMBG.

Scheme 1 Schematic illustration of synthetic process of YMBG scaffolds. The scaffolds with interconnected macropores (300–500 μm), widely distributed midpores (20 nm to 2 μm) and well-ordered mesopores (∼5 nm) were prepared using P123 (denoted in purple), yeast cells (denoted in green) and polyurethane sponges (denoted in yellow) as templates, respectively.

2.2 Characterization of hierarchically porous YMBG scaffolds

Differential thermal analyses (DTA, Thermoplus EVO II, Rigaku, Japan) were conducted to explore the physicochemical property of raw YMBG scaffolds. The macropores, midporous and mesoporous structure of the calcined YMBG scaffolds were characterized by scanning electron microscopy (SEM, S-4800, Hitachi, Japan), X-ray diffraction (XRD, G-670, Huber, Germany) and transmission electron microscopy (TEM, JEM-2100F, JEOL, Japan). N₂ adsorption–desorption isotherms were recorded on a surface area analyzer (Quadrasob SI, Quantachrome, USA). BET and BJH models were utilized to determine the specific surface area, pore size distribution and pore volume.

2.3 Mechanical properties of porous YMBG scaffolds

To investigate the effect of yeast cells on the mechanical properties of the scaffolds, the compressive strengths of all scaffolds (Ø 10 × h 10 mm) were tested using a computer controlled universal testing machine (AG-I, Shimadzu, Japan) at a speed of 0.5 mm min⁻¹ and about 50% compressive deformation was obtained at room temperature. Five samples were measured for each type.

2.4 Degradation behavior of porous YMBG scaffolds

To evaluate the degradation behavior, YMBG scaffolds (0.05 g) for each group (0YMBG, 5YMBG, 15YMBG, 30YMBG) were immersed in 10 mL of Tris–HCl buffered solution (pH = 7.40) in a shaker at 37 °C for 3, 7 and 14 days. The solution was collected and refreshed at each time point and the concentrations of Ca²⁺, PO₄³⁻ and SiO₄⁴⁻ ions were analyzed by inductively coupled plasma atomic emission spectrometry (ICP-AES, Varian Co., USA).

2.5 Cell attachment of BMSCs on YMBG scaffolds

For assessment of the interaction of cells with YMBG scaffolds with hierarchical porous structures, rabbit BMSCs were seeded on obtained scaffolds placed in 24-well plates at an initial seeding density of 1.0 × 10⁴ cells per scaffold. The cells were incubated in a humidified atmosphere of 5% CO₂ at 37 °C for 1, 3 and 7 days in growth medium supplemented with low glucose Dulbecco's Modified Eagle Medium (DMEM), 10% (v/v) fetal bovine serum (FBS) and 1% (v/v) penicillin/streptomycin. After being fixed in 2.5% glutaraldehyde/phosphate buffered saline (PBS) for 1 h, the cell/scaffold samples were washed 3 times with 4% (w/v) sucrose/PBS and post fixed with 1% osmium tetroxide/PBS. Then the samples were dehydrated through a graded series of ethanol (50, 70, 90, and 100% (v/v)) and hexamethyldisilazane (HMDS). Finally, the morphological characteristics of BMSCs on the scaffolds (coated with gold beforehand) were determined by SEM (Quanta™ 200). Pure MBG scaffolds were used as blank control.

2.6 Proliferation of BMSCs on YMBG scaffolds

To investigate the influence of YMBG scaffolds on proliferation of BMSCs, 3-(4,5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) assay was performed in triplicate according to our previous publication.³ 40 μL 0.5 mg mL⁻¹ of MTT solution and 360 μL of growth medium were added into each scaffold and cultured at 37 °C at each time point. 24 hours later, the media were removed and the formazan was dissolved in 100 mL of dimethyl sulfoxide (DMSO). The absorbance of formazan/DMSO solution was measured at λ = 495 nm on a SpectraMax Microplate Reader (Biotek instruments, USA).

2.7 Bone-related gene expression of BMSCs on YMBG scaffolds

Total RNA of the cells on YMBG scaffolds was extracted using TRIzol® Reagent. Complementary DNA was synthesized from 1 mg of total RNA using a DyNAmo™ cDNA Synthesis Kit following the manufacturer's instructions. Bone-related gene expression, including osteocalcin (OCN), runt-related transcription factor 2 (Runx2), collagen type I (Col I) and Bone Morphogenetic Protein-2 (BMP2), was determined by RT-qPCR analysis on the ABI Prism 7300 Sequence Detection System using SYBR® Green detection reagent. All the RT-qPCR primers were designed on the basis of cDNA sequences from the NCBI sequence database and the primer specificity was verified by BLASTN searches. Reactions were performed in triplicate.

2.8 Statistical analysis

All data were collected from 3–6 parallel groups and are presented as means ± standard deviations (SD). The non-parametric Kruskale–Wallis test was used to determine the level of significance and a P-value < 0.05 was considered statistically significant.

3. Results

3.1 Characterization of hierarchically porous YMBG scaffolds

SEM images show that the five types of YMBG scaffolds had a similar highly porous structure with an interconnected macroporous structures (pore size: 300–500 μm) (Fig. 1A₁–E₁), suggesting that mixing with yeast cells had no effect on macroporous structures. High magnification images indicate that there were highly curved mid-level pores (pore size: 1–2 μm) on the strut surface of macropores in 5, 10, 15 and 30YMBG scaffolds (Fig. 1A₂–E₂ referred to the strut surface morphology of 0, 5, 10, 15, 30YMBG scaffolds, respectively). It is obvious that the amount of midpores increased with the yeast cells added in preparation. TEM images demonstrate that the obtained scaffolds possessed a typical well-ordered mesoporous channels (pore size: ∼5 nm) and midpores (pore size: 20–200 nm) in the inner of scaffold struts (Fig. 2).

Fig. 1 SEM images of 0YMBG (A₁), 5YMBG (B₁), 10YMBG (C₁), 15YMBG (D₁) and 30YMBG (E₁) scaffolds with interconnected macropores (300–500 μm) in the scaffolds and midpores (1–2 μm) (A₂–E₂) on the struts surface of scaffolds.

Fig. 2 TEM images of 0YMBG (A), 5YMBG (B), 15YMBG (C) and 30YMBG (D) scaffolds with ordered mesoporous channels (∼5 nm) and midpores (20–200 nm, shown by arrow) in the struts of scaffolds.

N₂ adsorption–desorption isotherm of the YMBG scaffolds, which could be identified as type IV isotherms, further confirmed the presence of mesoporous structure (Fig. 3A). Apparently, all samples show a very narrow pore size distribution, mainly in the vicinity of 5 nm (Fig. 3B). It is worth mentioning that with the increasing additional amount of yeast cells used in scaffold preparation, there were more pores distributed over about several hundred of nanometers, suggesting an increase amount of the midpores (insert of Fig. 3B). The specific surface area, pore volume and pore size of YMBG scaffolds are shown respectively in Table 1. It can be seen that with the increase of addition amount of yeast cells, the specific surface area, pore volume and pore size of YMBG scaffolds showed a downward trend. However, the prepared YMBG scaffolds maintained high surface area (>224 m² g⁻¹) and pore volume (>0.27 cm³ g⁻¹). The average pore size of YMBG scaffolds was in the range of 2.0–5.5 nm (Table 1).

Fig. 3 N₂ adsorption–desorption isotherm (A), pore size distribution (B), small-angle XRD (C), wide-angle XRD (D), DTA analysis (E) and mechanical strength (F) of YMBG scaffolds (*P < 0.05). There are some weak characteristic peaks of CaO and SiO₂ appearing on the broad bump when yeast cells content reaches 10 wt%, and they become even sharper when yeast cell contents reached 15 and 30 wt% (D). The position of two exothermic peaks shifts to higher temperature with increasing amount of yeast cells, suggesting that the combustion of the templates may be put off and release more heat at higher temperature during calcination ((E), shown by arrow). The compressive strength significantly increases with yeast cells amount used in the preparation process, whereas 15YMBG possesses the highest strength.

Table 1 Specific surface area, mesopore volume and mesopore size of YMBG (Y represents yeast) scaffolds by N₂ adsorption–desorption analysis. The specific surface area, pore volume and pore size of YMBG scaffolds show a downward trend with the increasing content of yeast cells

Scaffolds	Surface area (m^2 g^−1)	Pore volume (cm^3 g^−1)	Pore size (nm)
0YMBG	438.4	0.5936	5.597
5YMBG	364.4	0.4899	4.877
15YMBG	224.5	0.2738	3.427
30YMBG	254.9	0.4878	2.011

---

XRD analysis revealed that all YMBG scaffolds with different yeast cells contents (0, 5, 10, 15 and 30 wt%) had no obvious crystal peaks and their diffraction patterns had a wide SiO₂ peak with low intensity (Fig. 3C and D). Generally, small-angle XRD patterns of five samples (Fig. 3C) exhibited three low-angle diffraction peaks in the 2θ range of 1.1–1.6°, which indicated that the uniform mesoporous structures of calcined samples might undergo shrinkage with increasing yeast cells contents, in accordance with TEM and N₂ adsorption–desorption analysis. More importantly, wide-angle XRD patterns(Fig. 3D) showed some weak characteristic peaks of CaO and SiO₂ appearing on the broad bump when yeast cells content reached 10 wt%, and they became even sharper when yeast cells content reached 15 and 30 wt%.

DTA analysis for raw YMBG scaffolds is shown in Fig. 3E. As the temperature increased, there were two obvious exothermic peaks, appearing around 250–500 °C, which might be attributed to the thermal decomposition of inorganic precursors, polyurethane sponges, block copolymer templates and yeast cells. It is important to note that the position of exothermic peaks shifted to higher temperature as contents of yeast cells increased.

3.2 Mechanical properties of YMBG scaffolds

The average compressive strengths of 0, 5, 15 and 30YMBG scaffolds were 44.1 ± 4.0, 92.2 ± 13.9, 140.9 ± 18.7 and 118.8 ± 11.6 kPa, respectively (Fig. 3F). The compressive strength significantly increased with the amount of yeast cells used in the preparation process, whereas 15YMBG scaffolds possessed the highest compressive strength.

3.3 Degradation behavior of YMBG scaffolds

The results of degradation experiments of YMBG scaffolds show that the concentrations of SiO₄⁴⁻, Ca²⁺ and PO₄³⁻ presented a sustained increase with soaking time in Tris–HCl solution (Fig. 4). Both the concentrations of SiO₄⁴⁻ and Ca²⁺ of the 5, 10 and 30YMBG scaffolds had a slightly decreased degradation profile compared with that of 0YMBG scaffolds (Fig. 4A and B).

Fig. 4 Degradation behavior of YMBG scaffolds in Tris–HCl: release of Si (A), Ca (B) and phosphate ions (C). The YMBG scaffolds (5, 10 and 30 wt%) have a decreased degradation profile compared with 0YMBG scaffolds.

3.4 Attachment and proliferation of BMSCs on YMBG scaffolds

The rabbit BMSCs adhesion and morphology on the 5 and 15MBG scaffolds were examined by SEM (Fig. 5A and B). The images verified that after 24 hour culture, both scaffolds could support the adhesion of BMSCs, which spread well and were closely attached to both types of scaffolds. The proliferation activity of rabbit BMSCs on each experimental scaffold represents an increasing trend of cell growth in a time dependent manner, as shown in Fig. 5C. There was no statistically significant difference between the blank control and the prepared YMBG groups after incubation for 1, 3 and 7 days.

Fig. 5 Adhesion of rabbit BMSCs on 5YMBG (A) and 15YMBG (B) scaffolds after cultured for 24 hours. Both scaffolds could support the adhesion of rabbit BMSCs, which spread well and are closely attached to both types of scaffolds. Proliferation activity of BMSCs on YMBG scaffolds (C) after cultured for 1, 3 and 7 days shows no significant difference between the blank control and the YMBG groups.

3.5 Bone-related gene expression of BMSCs on YMBG scaffolds

After cultured for 7 days, bone-related gene expression of BMSCs, including Runx2, OCN, Col I and BMP2, was detected by quantitative RT-PCR. Compared with blank control, all YMBG groups showed improved osteogenic gene expression in vitro, especially for 15 and 30YMBG scaffolds (Fig. 6). The BMSCs cultured on 15 and 30YMBG had a higher Runx2, OCN and BMP2 gene expression than those on blank control and 0YMBG groups. The 30YMBG samples represented significantly improved Col I gene expression as compared to other groups.

Fig. 6 Bone-related gene expression of BMSCs after cultured for 7 day: (A) Runx2, (B) OCN, (C) Col I and (D) BMP2 (*P < 0.05, **P < 0.01, ***P < 0.001). 15 and 30YMBG shows significantly improved osteogenic gene expression compared with other groups and blank control.

4. Discussion

In this study, a hierarchically porous bioactive glass scaffold with interconnected macropores (300–500 μm), widely-distributed midpores (20 nm to 2 μm) and well-ordered mesopores (∼5 nm) was successfully prepared by using polyurethane sponges, yeast cells and P123 as templates. The effects of yeast cells, as the porogen of midpores, on physicochemical properties of the MBG scaffolds and cell proliferation, differentiation, bone-related gene expression of rabbit BMSCs were systematically studied. It is found that the addition of yeast cells promoted the crystallization of bioactive glass scaffolds, significantly improved the compressive strength, slightly decreased the degradation rate, and more importantly, enhanced osteogenic gene expression of rabbit BMSCs cultured in the scaffolds. The current results imply that the hierarchically porous YMBG scaffolds could be a promising carrier for various biomolecules and tissue cells to endow the biomaterial with multifunctional therapeutic effects for bone tissue engineering application.

Herein, different from the preparation of conventional MBG scaffolds, in which usually two templates (i.e., polyurethane sponges and P123) were applied, YMBG scaffolds were prepared by using a tri-template method. Yeast cells were utilized to produce mid-level pores (20 nm to 2 μm) in these scaffolds. This natural biotemplate displays many advantages, including operating simplicity, economic feasibility, environmental friendliness and good biosafety. By varying the addition amount of yeast cells in preparation process (as illustrated in Scheme 1), the porous scale range of YMBG scaffolds could be effectively achieved. The interconnected macroporous structures (300–500 μm), which was previously reported to greatly benefit tissue ingrowth and nutrient transportation,^1,2 were successfully maintained in all groups of YMBG scaffolds. Highly curved midpores distributed widely on the surface of the macroporous strut (Fig. 1A₂–E₂), of which the size is in consistent with that of yeast cells. It is interesting to note that the size distribution of midpores has a wide range (i.e., 20 nm to 2 μm). Previous studies revealed that the pore size in the materials varied by using yeast as a porogen. For instance, the mesoporous LiFePO₄ materials consisted of a nonuniform open porous structure (∼3.5 nm) and a homogeneous porous structure (∼28.5 nm),²⁶ but hierarchically structured porous enzyme-protecting materials had larger pore size (200–500 nm).²⁸ These results indicate that the physiological state of yeast, the inorganic precursors and addition manner of yeast, as well as the experiment parameters (stirring, time, temperature, etc.), have important effect on the final pore size. For application in protein delivery, the suitable pore size was found to be 2–5 times that of the protein molecules^19,20 and the size of most biomolecules is larger than 5 nm. In this respect, the mid-level porous structures are capable of physically encapsulating a variety of biomolecules inside the obtained scaffolds, which could enhance immobilization efficiency and therapeutic effects. In addition, the well-ordered mesopores (∼5 nm) and high specific surface areas (>224.5 m² g⁻¹) could also be an important factor benefiting sustainable small-molecule drug delivery as previously reported.^3,18,31 Hence, it is speculated that this bioactive glass scaffold with widely-distributed porous structures (5 nm to 500 μm) could be efficiently loading with various guest molecules (small molecules and proteins) and tissue cells to satisfy the requirement of multifunction for bone regeneration.

In bone tissue engineering applications, the mechanical strength of scaffolds is of great importance for in vitro cell activity. It is well-known that increasing the sintering temperature is an effective way of changing the textural porosity and mechanical strength of the foam scaffolds. Previous publications have demonstrated that the crystallization onset temperature of bioactive glasses was approximately 700 °C.^32,33 With the further increase of the calcination temperature, the average pore size decreased due to the collapse of the mesoporous structure with the nucleation and growth of crystalline phase. Consequently, the compressive strength increased owing to a thickening of the pore walls and a reduction in the textural porosity.^34,35 In the present work, it is interesting to find that four YMBG scaffolds generally maintain amorphous state; however, the increasing amount of yeast cells (e.g. 15YMBG and 30YMBG) significantly promoted the crystallization of scaffolds with appearing some weak CaO peaks (Fig. 3D). DTA curves (Fig. 3E) show that there were two obvious exothermic peaks from 250 °C to 500 °C, corresponding to decomposition of triblock copolymer P123 ³² and yeast cells. Importantly, the exothermal reaction was postponed and the exothermic peaks shifted to higher temperature, indicating that the additive exothermal effect of burning yeast has a non-negligible influence in the thermodynamic behavior of sponge scaffolds during the calcination process (650 °C). It is inferred that the released heat from burning yeast raised the local temperature and thereby promoted phase transformation behaviour in certain region of the YMBG scaffolds. Therefore, the formation of microcrystals resulted in the decrease of the mesopore size and specific surface area along with the increase of mechanical strength. All above revealed that the existence of microcrystals inside the scaffolds may play a key role in reinforcing the scaffolds with significantly improved compressive strength. As shown in Fig. 5F, the compressive strength was significantly increased with the increasing amount of yeast cells and 15YMBG scaffolds have the highest strength of 140.9 kPa, which is more than twice of conventional MBG scaffolds (<60 kPa).^36,37 Therefore, one of distinct advantages for the prepared YMBG with three-level pores by incorporating yeast cells as porogen, is their significantly improved mechanical strength as compared to conventional MBG scaffolds with two-level pores, which makes them easier to handle and utilize for cell culture.

The other distinct advantage for the prepared YMBG scaffolds is their significantly stimulatory effect on the osteogenic differentiation of BMSCs. As shown in Fig. 6, the bone-related gene expression of BMSCs (Runx2, OCN, Col I and BMP2) in 15Y and 30YMBG scaffolds was significantly enhanced as compared to 0YMBG scaffolds. BMPs are members of the transforming growth factor β (TGF-β) superfamily with important effects on bone cell differentiation. Runx2 is not only a common target of TGF-β and BMP2 but also a transcription factor required for bone formation.³⁸ Collagen and OCN are both important markers for osteogenic differentiation.^3,21 In this study, there may be two main factors to positively stimulate the osteogenic differentiation of BMSCs. One is the significantly improved mechanical strength due to the existence of microcrystal in the scaffolds. The other is the enhanced roughness of the macroporous strut surface. Previous studies suggested that the differentiation of BMSCs can be manipulated according to the mechanical properties of the culture substrate.²⁹ Soft scaffolds (∼1 kPa) favored neuronal differentiation²⁹ while substrates with the stiffness close to bone (25–40 kPa) enhanced osteogenic differentiation of human mesenchymal stem cells.³⁰ The average compressive strength of spongy bone is in the range of 0.2–4 MPa.³⁹ Here we demonstrated that the measured compressive strength of the 15YMBG scaffolds (140.9 kPa) is closer to that of cancellous bone than conventional MBG scaffolds (<60 kPa). Therefore, it is reasonable to speculate that the improved mechanical strength in our study may play a major role in enhancing gene expression of BMSCs. In addition, the microscale and nanoscale surface topography can influence cell interactions with material surfaces⁴⁰ and appropriate surface roughness can facilitate in vivo bioactivity and osseointegration compared with smooth surfaces.⁴¹ In present study, the widely-distributed midpores on the macroporous strut surface (Fig. 1A₂–E₂) enhanced the interface roughness between cells and materials, which might have positive effect on the enhanced bone-related gene expression of BMSCs (Fig. 6). Besides, proper degradation rate is also an important factor in the cellular behaviors.¹ The YMBG scaffolds maintain continuous release of calcium, silicate and phosphate ions; however, the release kinetics of three ions in four YMBG scaffolds shows no significant difference, indicating that the concentration difference of released ions is not the major factor contributing to the improved osteogenic differentiation of BMSCs.

5. Conclusion

A hierarchically porous bioglass scaffold (noted as YMBG) with interconnected macropores (300–500 μm), widely-distributed midpores (20 nm to 2 μm) and well-ordered mesopores (∼5 nm), have been successfully prepared by using tri-template (P123, yeast cells and polyurethane sponges) method. The utilization of yeast biotemplate has significantly influenced the physicochemical properties and osteogenic activity of rabbit bone marrow stromal cells (BMSCs). The incorporation of yeast cells as mid-level pore (20 nm to 2 μm) biotemplate in the scaffolds significantly promoted the crystallization and improved their compressive strengths. In addition, the prepared scaffolds could support attachment and proliferation of rabbit BMSCs, and more importantly, significantly enhanced bone-related gene expression (Runx2, OCN, Col I and BMP2). Among four types of YMBG scaffolds, 15YMBG scaffolds have the highest mechanical strength of 140.9 kPa (more than twice of conventional MBG scaffolds) and best stimulatory effect on the osteogenic differentiation. Considering their special hierarchical structures and bioactivity, the scaffolds have great potential for effectively loading with various biomolecules and tissue cells to enhance the therapeutic effects for bone tissue engineering application.

Acknowledgements

Funding for this study was provided by the Recruitment Program of Global Young Talent, China (Dr Wu), the National High Technology Research and Development Program of China (863 Program, SS2015AA020302) and Program of Shanghai Outstanding Academic Leaders (15XD1503900).

References

1. L. L. Hench and I. Thompson, J. R. Soc., Interface, 2010, 7, S379–S391.
2. D. Arcos, I. Izquierdo-Barba and M. Vallet-Regi, J. Mater. Sci.: Mater. Med., 2009, 20, 447–455.
3. C. T. Wu, Y. H. Zhou, M. C. Xu, P. P. Han, L. Chen, J. Chang and Y. Xiao, Biomaterials, 2013, 34, 422–433.
4. A. Hoppe, N. S. Guldal and A. R. Boccaccini, Biomaterials, 2011, 32, 2757–2774.
5. M. Xu, Y. Zhang, D. Zhai, J. Chang and C. Wu, Biomater. Sci., 2013, 1, 933–941.
6. M. Alcaide, P. Portoles, A. Lopez-Noriega, D. Arcos, M. Vallet-Regi and M. T. Portoles, Acta Biomater., 2010, 6, 892–899.
7. M. Kobayashi, T. Nakamura, Y. Okada, A. Fukumoto, T. Furukawa, H. Kato, T. Kokubo and T. Kikutani, J. Biomed. Mater. Res., 1998, 42, 223–237.
8. H. W. Kim, H. E. Kim and J. C. Knowles, Adv. Funct. Mater., 2006, 16, 1529–1535.
9. Y. F. Zhu, C. T. Wu, Y. Ramaswamy, E. Kockrick, P. Simon, S. Kaskel and H. Zrelqat, Microporous Mesoporous Mater., 2008, 112, 494–503.
10. X. X. Yan, C. Z. Yu, X. F. Zhou, J. W. Tang and D. Y. Zhao, Angew. Chem., Int. Ed., 2004, 43, 5980–5984.
11. X. Li, X. P. Wang, D. N. He and J. L. Shi, J. Mater. Chem., 2008, 18, 4103–4109.
12. J. Fan, C. Z. Yu, T. Gao, J. Lei, B. Z. Tian, L. M. Wang, Q. Luo, B. Tu, W. Z. Zhou and D. Y. Zhao, Angew. Chem., Int. Ed., 2003, 42, 3146–3150.
13. X. Li, X. P. Wang, H. R. Chen, P. Jiang, X. P. Dong and J. L. Shi, Chem. Mater., 2007, 19, 4322–4326.
14. H. S. Yun, S. E. Kim and Y. T. Hyeon, Chem. Commun., 2007, 2139–2141.
15. C. T. Wu, Y. X. Luo, G. Cuniberti, Y. Xiao and M. Gelinsky, Acta Biomater., 2011, 7, 2644–2650.
16. Y. Wu, S. A. Zhu, C. T. Wu, P. Lu, C. C. Hu, S. Xiong, J. Chang, B. C. Heng, Y. Xiao and H. W. Ouyang, Adv. Funct. Mater., 2014, 24, 4473–4483.
17. Y. L. Hong, X. S. Chen, X. B. Jing, H. S. Fan, B. Guo, Z. W. Gu and X. D. Zhang, Adv. Mater., 2010, 22, 754–758.
18. C. T. Wu and J. Chang, Interface Focus, 2012, 2, 292–306.
19. Y. Kang, J. He, X. D. Guo, X. Guo and Z. H. Song, Ind. Eng. Chem. Res., 2007, 46, 4474–4479.
20. S. L. Gao, Y. J. Wang, X. Diao, G. S. Luo and Y. Y. Dai, Bioresour. Technol., 2010, 101, 3830–3837.
21. T. Douglas, S. Heinemann, C. Mietrach, U. Hempel, S. Bierbaum, D. Scharnweber and H. Worch, Biomacromolecules, 2007, 8, 1085–1092.
22. K. E. Kadler, D. F. Holmes, J. A. Trotter and J. A. Chapman, Biochem. J., 1996, 316, 1–11.
23. R. Selvakumar, N. Seethalakshmi, P. Thavamani, R. Naidu and M. Megharaj, RSC Adv., 2014, 4, 52156–52169.
24. B. Zhang, S. Yang, Y. Zhang, Q. X. Wang and T. R. Ren, Colloids Surf., B, 2015, 131, 129–135.
25. J. J. Cui, W. He, H. T. Liu, S. J. Liao and Y. Z. Yue, Colloids Surf., B, 2009, 74, 274–278.
26. X. D. Zhang, X. G. Zhang, W. He, C. Y. Sun, J. Y. Ma, J. L. Yuan and X. Y. Du, Colloids Surf., B, 2013, 103, 114–120.
27. M. J. Huang and Y. J. Wang, J. Mater. Chem., 2012, 22, 626–630.
28. W. He, D. D. Min, X. D. Zhang, Y. Zhang, Z. Y. Bi and Y. Z. Yue, Adv. Funct. Mater., 2014, 24, 2206–2215.
29. G. J. Her, H. C. Wu, M. H. Chen, M. Y. Chen, S. C. Chang and T. W. Wang, Acta Biomater., 2013, 9, 5170–5180.
30. Z. Li, Y. W. Gong, S. J. Sun, Y. Du, D. Y. Lu, X. F. Liu and M. Long, Biomaterials, 2013, 34, 7616–7625.
31. S. Bose, M. Roy and A. Bandyopadhyay, Trends Biotechnol., 2012, 30, 546–554.
32. W. Xia and J. Chang, Microporous Mesoporous Mater., 2008, 108, 345–351.
33. X. Chatzistavrou, T. Zorba, E. Kontonasaki, K. Chrissafis, P. Koidis and K. M. Paraskevopoulos, Phys. Status Solidi A, 2004, 201, 944–951.
34. J. R. Jones, L. M. Ehrenfried and L. L. Hench, Biomaterials, 2006, 27, 964–973.
35. X. X. Yan, H. X. Deng, X. H. Huang, G. Q. Lu, S. Z. Qiao, D. Y. Zhao and C. Z. Yu, J. Non-Cryst. Solids, 2005, 351, 3209–3217.
36. Y. F. Zhu, Y. F. Zhang, C. T. Wu, Y. Fang, J. H. Yang and S. L. Wang, Microporous Mesoporous Mater., 2011, 143, 311–319.
37. C. T. Wu, Y. F. Zhang, Y. F. Zhu, T. Friis and Y. Xiao, Biomaterials, 2010, 31, 3429–3438.
38. K. S. Lee, H. J. Kim, Q. L. Li, X. Z. Chi, C. Ueta, T. Komori, J. M. Wozney, E. G. Kim, J. Y. Choi, H. M. Ryoo and S. C. Bae, Mol. Cell. Biol., 2000, 20, 8783–8792.
39. M. F. A. L. J. Gibson, Cellular solids: structure and properties, Pergamon Press, Oxford, 1999, pp. 429–452.
40. C. Luo, L. Li, J. R. Li, G. Yang, S. Ding, W. Zhi, J. Weng and S. B. Zhou, J. Mater. Chem., 2012, 22, 15654–15664.
41. Y. Deng, X. C. Liu, A. X. Xu, L. X. Wang, Z. Y. Luo, Y. F. Zheng, F. Deng, J. Wei, Z. H. Tang and S. C. Wei, Int. J. Nanomed., 2015, 10, 1425–1447.

---

Footnote

† X. W and M. S contribute equally.

---