Introduction

As a fibrous protein, silk fiber from Bombyx mori has been used as biomedical material due to its unique porous structure, biocompatibility, permeability, biodegradability, and minimal inflammatory reaction (Altman et al. 2003; Jin et al. 2004). However, it still has problematic characteristics for further clinical applications and use in industrial processes, such as its low stiffness and high production cost (Freed et al. 1994; Zhou et al. 2009). For this reason, several attempts have been made to improve the properties of silk biomaterials, and this field has become a rapidly growing area of tissue engineering research. Research has demonstrated that the main inorganic component of human bone tissues, hydroxyapatite (HA), enhances the compressive strength of silk biomaterials and also induces new bone formation (Na et al. 2007). Therefore, the HA/silk scaffold is an attractive candidate for bone tissue engineering and has thus been widely studied in the last decade (Kim et al. 2008; Bhumiratana et al. 2011).

Recently, we developed a novel method to convert silk fiber into regenerated silk fibroin (RSF) solutions that can be subsequently reconstituted into macroporous architecture (Cao et al. 2007). Research has demonstrated that RSF is stronger, tougher, and more extensible than natural silk fiber (Chen et al. 2009; Zhou et al. 2011).

Here, we combined the advantages of both RSF and HA to prepare a novel HA/RSF scaffold. After first characterizing the scaffold it was then investigated for its potential in osteoregenerative applications. Both HA/RSF scaffold osteoconductivity and osteoinductivity were assessed by culturing rat bone marrow-derived mesenchymal stromal cells (BMSCs) on the scaffold surface and observing cell morphology, cell viability, DNA content, gene expression, alkaline phosphatase (ALP) activity, etc.

Materials and methods

Preparation of the RSF scaffold

The RSF scaffold was prepared according to previous methods (Cao et al. 2007). Briefly, B. mori silk was degummed in 0.5 % (w/v) Na2CO3 at 95 °C for 1 h, washed with copious distilled water, and then dried. The degummed silk was dissolved in 9.5 mM LiBr. After filtration, the fibroin solution was dialyzed against deionized water for four days at room temperature with a 12 kDa cut-off, semi-permeable membrane for salt removal. The dialyzed silk fibroin solution was centrifuged at ~5,000×g for 10 min. Next, 10 % (w/w) ethanol was added to the RSF solution under gentle stirring at 25 °C over 2 min, and then held at −80 °C for 24 h. The sample was defrosted by freeze-drying. Dry particles of RSF were obtained by lyophilization with a freezing dryer after stable RSF microspheres were collected by centrifugation at ~10,000×g for 30 min.

HA/RSF scaffold preparation

HA was deposited onto the RSF scaffold using an alternate soaking process (Wang et al. 2007). In brief, the RSF scaffold was immersed in 200 mM CaCl2 in a Petri dish and shaken at 150 rpm for 1 h at 37 °C. The scaffold was then blotted onto filter paper to remove excess water and then immersed in 120 mM Na2HPO4 for 1 h. After soaking alternately for three cycles, HA/RSF scaffolds were washed in distilled water and air-dried at room temperature for 24 h. The morphology of the RSF and HA/RSF scaffolds was observed using scanning electron microscopy (SEM).

Scaffold weight change after HA coating

The dry weights of the RSF and HA/RSF scaffolds were measured and the weight percentage of HA deposited onto the silk scaffolds was calculated as:

$$ {\text{Percentage}}\;{\text{deposition}}\;{\text{of}}\;{\text{HA}}\; ( {\text{w/w}},\% ) = (W_{t} - W_{r} )/W_{t} \times 100\,\% , $$

where W r is weight of RSF scaffold, W t is weight of HA/RSF scaffold.

Cell isolation, seeding, and culturing onto scaffolds

BMSCs were isolated from five New Zealand white rabbits (average weight, 2.5 kg) according to procedures described previously (Shao et al. 2006), and cultured in Dulbecco’s modified Eagle’s medium at 37 °C with 5 % CO2. Scaffolds were sterilized by autoclaving and seeded with BMSCs at 1.5 × 105 cells per scaffold.

MTT assay

The MTT assay was carried out as previously described (Wang et al. 2007). Briefly, culture medium was removed and 100 μl MTT was added to each well. After incubation for 4 h, the medium was discarded and the precipitated formazan was dissolved in dimethyl sulfoxide. The A570 was then measured in a microplate reader.

DNA content analysis

The RSF and HA/RSF scaffolds from various time intervals were washed in phosphate-buffered saline and lysed in 0.2 % (w/v) Triton X-100 and 5 mM MgCl2. A Pico Green assay was used to measure the total amount of DNA.

ALP activity assessment

The ALP activity of cells cultured on RSF and HA/RSF scaffolds was detected at several time intervals using a phosphatase substrate kit. ALP activity was expressed as p-nitrophenol produced in nmol/min/mg protein.

Gene expression analysis by quantitative real-time PCR (RT-PCR)

Gene expression of various proteins involved in osteogenesis, including Runt-related transcription factor 2 (Runx2), collagen type I, osteopontin (OPN), and osteonectin (ON) were analyzed in BMSCs seeded with RSF and HA/RSF scaffolds. Primer sequences of the genes were obtained from published literature Fiedler et al. (2002). Total RNA was extracted from each cell sample using TRIzol and the cDNA was synthesized from total RNA. Quantitative PCR (QPCR) was carried out using SyberGreen I in a Rotor-Gene 3000. To correct for differences in RNA quality and quantity among samples, the data were normalized to that of GAPDH.

Statistical analysis

Results are expressed as the mean ± SD. Multiple comparisons involved single factor analysis of variance (ANOVA) using SPSS Statistics 17.0 statistical software package. P values <0.05 were considered statistically significant.

Results and discussion

HA/RSF scaffold morphology

SEM images revealed a porous, spongy morphology of the HA/RSF scaffold, while HA nanoparticles were deposited uniformly on the surface of the silk sponge without any apparent effect on scaffold pore architecture (Fig. 1a). As shown in Fig. 1b, high-magnification SEM images confirmed the presence of dense flake-like HA crystals emanating from the surface and throughout the pore walls of the silk framework. Approx. 12.4 ± 0.06 % (w/w) of HA was deposited on the silk scaffold after HA deposition by weight quantification analysis. Energy dispersive X-ray (EDX) analysis (Fig. 1c) revealed that the ratio of Ca and P was 1.52, similar to the composite of mature bone.

Fig. 1
figure 1

(a) SEM image and (b) high-magnification SEM image of a prepared HA/RSF scaffold. (d) Energy dispersive X-ray (EDX) analysis of a prepared HA/RSF scaffold. SEM images of BMSCs seeded onto a HA/RSF scaffold after (c) 1 day, (e) 7 days, and (f) 14 days. The blank arrow in (f) indicates typical osteoblast-like cells

Full size image

SEM images of BMSC-seeded scaffolds after 1 day in culture revealed that HA/RSF scaffolds support the adhesion and proliferation of BMSCs (Fig. 1d). After 7 days, BMSCs adhered to the surface of the HA/RSF scaffold and exhibited a spreading-out phenotype (Fig. 1e). After 14 days, it was obvious that typical osteoblast-like cells appeared in the scaffolds (Fig. 1f).

Cell viability and proliferation

As shown in Fig. 2, after 14 days the cell viability and DNA content of BMSCs increased significantly (18.5 ± 0.6 and 33.2 ± 1.2 %, respectively) on the HA/RSF scaffold compared to the HA scaffold. Both the MTT assay and DNA content analysis suggested that BMSCs seeded on the HA/RSF scaffold increased and remained viable over 2 weeks in culture, indicating that the scaffold supported the proliferation of BMSCs.

Fig. 2
figure 2

(a) MTT assay and (b) DNA content analysis of BMSCs seeded onto a HA/RSF scaffold at different time intervals. Results are shown as the mean ± SD (n = 5)

Full size image

ALP activity and gene expression analysis by quantitative RT-PCR

ALP has been implicated as a marker of osteogenic differentiation that is expressed early in the differentiation process. As shown in Fig. 3, ALP activity increased by 41.3 ± 2.5 % after 14 days in HA/RSF scaffolds compared with that in RSF scaffolds. Furthermore, as shown in Fig. 4, after 2 weeks in culture, gene expression analysis of bone-related genes, such as Runx2, collagen type I, OPN, and ON, were upregulated in the BMSC-seeded HA/RSF scaffolds compared with BMSC-seeded RSF scaffolds. These observations suggest that prepared HA/RSF scaffolds may enhance the differentiation of BMSCs toward osteoblast cells.

Fig. 3
figure 3

Detection of ALP activity for RSF and HA/RSF scaffolds at different time intervals. Results are shown as the mean ± SD (n = 5)

Full size image
Fig. 4
figure 4

Gene expression analysis of BMSCs seeded onto both RSF and HA/RSF scaffolds at different time intervals by quantitative RT-PCR. Results are shown as the mean ± SD (n = 5)

Full size image

Conclusions

We successfully fabricated HA/RSF hybrid scaffolds that combine the advantages of both RSF and HA. The feasibility of these HA/RSF hybrid scaffolds as implants for bone damage was evaluated in vitro. Our results suggest that the incorporation of HA not only increases the surface roughness of scaffolds, but also enhances the viability and proliferation of BMSCs on the scaffold surface. Furthermore, these scaffolds improve the differentiation of BMSCs toward osteoblast cells in vitro. In conclusion, our work demonstrates that HA/RSF scaffolds may be a promising biomaterial for bone tissue engineering in the future.