Characterization of SDF-1ɑ and its role in rBMSC recruitment

As shown in Fig. 1a, pure SDF-1α protein was obtained by Ni-NTA affinity chromatography and desalting chromatography, and the molecular weight was consistent with the theoretical molecular weight of 9.0 kDa. As shown in Fig. 1b, the WB results showed a single band at 9.0 kDa, which further indicated that the purified protein was SDF-1α. The SDF-1α/CXC signalling pathway plays an important role in stem cell migration and is the main reason that SDF-1α can recruit stem cells. Therefore, we identified the presence of the CXCR4 receptor on the surface of our extracted rBMSCs by immunofluorescence staining. As shown in Fig. 1c, the immunofluorescence results showed that there were a large amount of CXCR4 protein on the surface of rBMSCs; thus, CXCR4, as a target on the cell surface, could be recruited by SDF-1α. Cell migration is the slow directional movement of cells through cell body deformation after receiving endogenous or exogenous signals. Many studies have found that SDF-1α has a chemotactic effect on BMSCs [34]. Therefore, we determined the concentration of SDF-1α with the strongest chemotactic effect by measuring the number of cells on the Transwell membrane. The Transwell chemotaxis assay results showed that SDF-1α promoted rBMSC migration in a dose-dependent manner (Fig. 1d and e) compared to the control. SDF-1α significantly promoted rBMSC migration starting at 0.5 µmol. The in vitro data showed that with increasing SDF-1α concentration, the ability of the chemokine SDF-1α to promote rBMSC migration gradually improved. According to the experimental results, a concentration of 2 µmol/mL SDF-1α was used in the subsequent experiments.

Fig. 1

a SDF-1α SDS‒PAGE electrophoretic diagram. b WB analysis of SDF-1α at different concentrations. c AF488-labelled CXCR4 immunofluorescence identification of rBMSCs; scale bars represent 40 μm. d Transwell cell migration assay. The cells were stained with crystal violet; the scale bars represent 500 μm. e The corresponding quantitative evaluation of cell migration in different concentrations of SDF-1α, P < 0.05; error bars represent standard deviation for n = 3

Characterization of the GelMA/SDF-1α@PCL/BG scaffold

The appearance of the scaffold was white, and the scaffold is presented in Fig. 2a as a 10 × 10 × 3 mm cube. The PCL and PCL@BG20 scaffolds showed an obvious porous structure. The porous structures of PCL@BG20-GelMA and PCL@BG20-GelMA/SDF-1α were covered by GelMA. As shown in Fig. 2b, the porosity of the PCL scaffold was 49.34%, and that of the scaffolds containing 20% BG ranged between 48.57% and 48.88%. As shown in Fig. 2c, SEM revealed the surface and cross-section of the PCL scaffold to be relatively smooth, with no GelMA between the pores. The cross-sectional SEM images showed BG particles uniformly dispersed in PCL@BG20, PCL@BG20-GelMA and PCL@BG20-GelMA/SDF-1α. After GelMA and GelMA/SDF-1α hydrogels were added to the scaffold, hydrogels with a highly porous structure formed between the scaffold pores. As shown in Fig. 2d, SDF-1α was released continuously during GelMA hydrogel degradation for up to 15 days, indicating a sustained cytokine delivery function of PCL@BG20-GelMA/SDF-1α. In our previous research, we explored the effects of different proportions (5–20%) of BG in PCL on osteogenic differentiation, and the results showed that PCL scaffolds containing 20% BG had the strongest ability to promote osteogenesis; thus, 20% BG was used in this study [35].

Fig. 2

a Macroscopic view and (b) porosity of the scaffold. c SEM images of different scaffolds. The red arrows indicate BG particles. d SDF-1α cumulative release curve from the GelMA/SDF-1α@PCL/BG scaffold, n = 3

The results of the surface element analyses of different scaffolds are shown in Fig. 3. The EDS maps showed Si (blue) exposure on the surface of the PCL@BG, PCL@BG20-GelMA and GelMA/SDF-1α@PCL/BG scaffolds. Compared with the PCL/BG scaffold, new elements (N) could be seen on the surface of the PCL@BG20-GelMA/SDF-1α scaffold, indicating successful immobilization of the hydrogel on both scaffolds. Subsequently, FTIR was used to further study the chemical composition of the different scaffold materials. As shown in Fig. 4a, FTIR revealed the characteristic peaks of PCL at 2850–3000 cm−1 (representing - (CH2)4 -), 1750 cm−1 (representing -C = O), and 1150–1250 cm−1 (representing -(- C-O-)). Compared with the PCL scaffold, the PCL@BG20 scaffold showed no visible alteration in the FTIR spectrum, indicating that the BG was dispersed into the PCL only by physical mixing instead of by chemical reaction. The FTIR spectrum of GelMA showed a newly formed vibration peak of the amide bond at 1680 cm−1, indicating that the methacrylate group had been successfully grafted onto the gelatine skeleton. On the other hand, a new band at approximately 3400 cm−1, which represented the stretching vibration of the alcoholic hydroxyl group, appeared on the PCL@BG20-GelMA and PCL@BG20-GelMA/SDF-1α scaffolds. Furthermore, another band at 1550–1660 cm−1 was observed in the spectra of PCL@BG20-GelMA@ and PCL@BG20-GelMA/SDF-1α, which represented –NH2. There were no –NH2 groups in the PCL or PCL@BG20 scaffolds. We also tested the mass ratios of various components in the scaffolds using TGA. As shown in Fig. 4b, with increasing temperature, PCL, GelMA and SDF-1α in the scaffold decomposed successively, and the remaining substance in the scaffold was BG. The TGA results showed that the mass percentage of BG in the PCL@BG20, PCL@BG20-GelMA and PCL@BG20-GelMA/SDF-1α scaffolds was 22.987%, 22.889% and 24.784%, respectively.

Fig. 3

Contents of Si and N in different scaffolds analysed by EDX mapping. The scale bar represents 1 mm

Fig. 4

a FTIR spectra and (b) TGA curves of different scaffold materials

In the treatment of bone defects, scaffolds must not only fill the bone defect site but also provide mechanical support. Therefore, the mechanical strength of scaffolds is very important for bone repair, especially at load-bearing sites. In this study, the compressive and tensile strengths of different scaffold materials were analysed by mechanical tests. As shown in Fig. 5a and b, the compressive and tensile strength of the pure PCL scaffold was 34.73 ± 3.77 MPa and 5.68 ± 0.33 MPa, respectively. After the addition of BG, the compressive strength of the scaffolds increased to 45.24 ± 2.63 MPa, but the tensile strength of the scaffolds decreased to 4.52 ± 0.46 MPa, which indicates that BG can effectively improve the stiffness of the materials but decreases their flexibility. After the addition of GelMA and SDF-1α, the compressive and tensile strength of the scaffold did not change significantly, so the addition of GelMA and SDF-1α had little effect on the mechanical properties of the scaffold. Chen et al. [36] found that the compression modulus of GelMA was somewhat related to the degree of methacryloyl substitution and that GelMA with a degree of substitution below 65% had little effect on the compression modulus of scaffolds. According to the above results, the compressive strength of the PCL@BG20-GelMA/SDF-1α scaffold was determined to be similar to that of human cancellous bone, making this scaffold an ideal bone replacement material.

Fig. 5

a compression strength and (b) tensile strength of different scaffolds. *P < 0.05, error bars represent standard deviation for n = 3

In vivo bone repair efficacy of the PCL@BG20-GelMA/SDF-1α scaffold

To assess bone formation in different groups, micro-CT images were taken at 8 weeks after surgery (Fig. 6a). Compared with the PCL group, the other groups showed more obvious new bone formation at the defect site, and the density of new bone was also significantly increased. Among all experimental groups, the PCL@BG20-GelMA/SDF-1α group showed the largest area of newly formed bone and large amounts of bone tissue ingrowth into the pores of the scaffold. Sagittal images of the bone tissue also showed that in the PCL@BG20-GelMA/SDF-1α groups, more new bone had formed around the scaffold at the bone defect site. We performed a quantitative analysis of the new bone tissue by 3D reconstruction (Fig. 6b). The BV/TV in the PCL@BG20 group was 15.59 ± 1.88%, which was significantly higher than that in the PCL group (10.76 ± 1.39%). There was no significant difference in the BV/TV between the PCL@BG20 group and the PCL@BG20-GelMA group. The BV/TV was significantly higher in the PCL@BG20-GelMA/SDF-1α scaffold group than in the other groups. As shown in Fig. 6c, the Tb.N in the PCL@BG20 group was obviously higher than that in the PCL group. There was no significant difference in the Tb.N between the PCL@BG20 group and the PCL@BG20-GelMA group. PCL@BG20-GelMA/SDF-1α showed a higher Tb.N than the other groups. As shown in Fig. S2, macroscopic images of each rabbit tibia 8 weeks after implantation of the different scaffolds showed that the bone surface was smooth, the scaffold was completely within the defect, and there was no obvious inflammatory reaction. The above results prove that the combination of BG + GelMA + SDF-1α can significantly improve the biological activity of scaffold materials and has a very good effect on bone tissue regeneration after injury.

Fig. 6

a Representative 3D micro-CT images of the region of the proximal radius defect at 8 weeks postsurgery. Quantitative analysis of (b) BV/TV and (c) Tb.N in new trabecular bone at 8 weeks after surgery. *P < 0.05; error bars represent the standard deviation for n = 3

Histological analysis

Eight weeks after surgery, the rabbits were sacrificed, and bone tissue was removed. Subsequently, we analysed bone formation and collagen deposition at the defect site by H&E, Masson, and sirius red staining. As shown in Fig. 7, the scaffolds in each group were located at the bone defect site, and the scaffolds were fixed well, providing a good framework for the growth of bone tissue. In the PCL group, H&E and Masson staining showed only a small amount of new bone formation in the pores of the scaffold. However, there was a significant increase in the volume and thickness of new bone in the PCL@BG20 and PCL@BG-GelMA groups. Among all the groups, the PCL@BG20-GelMA/SDF-1α group showed the largest amount and greatest thickness of new bone, which was consistent with the results of micro-CT analysis. These results indicate that the addition of SDF-1α can promote the aggregation of rBMSCs, ensure that bone defect sites have a sufficient cell supply, and promote the rapid formation of new bone tissue, which was also consistent with the findings of Wang et al. [37]. Wang et al. [37] found that scaffolds carrying SDF-1α exhibited better bone conduction and induction ability. Scaffold materials can be used to repair large bone defects in a shorter time when carrying SDF-1α, with a repair effect similar to that of autogenous bone. To observe the distribution of collagen fibres in the new bone tissue, the sections were stained with sirius red. As shown in Fig. 7 (sirius red), the collagen fibre area in the PCL@BG20, PCL@BG20-GelMA and PCL@BG20-GelMA/SDF-1α groups was larger than that in the PCL group. Among all groups, type I collagen fibres were best distributed in the PCL@BG20-GelMA/SDF-1α group and arranged in an orderly manner, which indicates good bone tissue formation. However, there were fewer and more disordered collagen fibres in the other groups, indicating inferior bone formation compared to that in the PCL@BG20-GelMA/SDF-1α group and suggesting that the addition of SDF-1α further promoted bone tissue healing. These results further indicate that the recruitment of rBMSCs is important for bone regeneration.

Fig. 7

Histological analysis using H&E, Masson trichrome and sirius red staining

Scaffold biosafety assessment

Bone tissue imaging and histological experiments confirmed that BG, GelMA and SDF-1α can significantly improve the bone repair ability of PCL scaffolds. However, such nanomaterials can enter other parts of the body through the circulation of body fluids. Therefore, the biosafety of scaffold materials also needs to be studied. The heart, liver, spleen, kidneys and lungs of the experimental animals were observed histopathologically 8 weeks after the operation. As shown in Fig. 8, the results of H&E staining of the heart tissue of healthy rabbits showed that the myocardial fibre tissue arrangement was normal, with no obvious changes to the myocardial interstitium. The structure of the liver lobule was intact, and there was no degeneration or sclerosis of the liver cells. Furthermore, no obvious pathological changes or inflammatory cell infiltration were observed in the spleen, kidneys or lungs. At 8 weeks, no significant differences were found in the heart, liver, spleen, lungs or kidneys of the experimental animals in each group compared with those of healthy rabbits. Therefore, when the PCL@BG20-GelMA/SDF-1α scaffold is applied locally to a bone defect site, it will not have any adverse effects on the important organs of rabbits. The above experimental results indicate the excellent biocompatibility and bone repair ability of the PCL@BG20-GelMA/SDF-1α scaffold we prepared.

Fig. 8

At 8 weeks after surgery, H&E-stained sections of important organs (heart, liver, spleen, lungs and kidneys) were obtained from experimental animals. Scale bars represent 200 μm

3D scaffolds play an important role in the treatment of bone defects by providing suitable mechanical support and architectural clues for bone regeneration. In recent years, PCL scaffolds prepared by 3D printing technology have attracted widespread attention due to their excellent mechanical properties, biocompatibility, and porous structure, which promote bone growth [7]. However, the hydrophobicity and lack of necessary osteogenic activity limit the further application of PCL [38]. In addition to suitable mechanical strength and porosity, materials applied in bone defect repair also need to provide a good regenerative microenvironment to promote cell proliferation, cell differentiation and extracellular matrix synthesis. Therefore, another type of material that is widely used in bone repair is hydrogels. Hydrogel materials exhibit good biocompatibility and encapsulation, allowing them to simulate the structure of the natural extracellular matrix and promote cell growth and nutrient exchange [39]. However, hydrogels also have the disadvantage of low mechanical properties, which are not conducive to their application in load-bearing areas. In this study, the versatility of 3D printing technology enabled us to combine the mechanically strong polymer PCL with the hydrogel GelMA in one structure, fully combining the advantages of both polymeric and hydrogel scaffolds. SEM observation showed that the prepared scaffolds had an appropriate pore size and could effectively promote the diffusion of nutrients and oxygen. Pore structures of 300 μm or larger have been reported to favour cell migration, bone formation, and angiogenesis. Furthermore, the scaffold we prepared had good mechanical strength, similar to that of natural cancellous bone. Thus, the hydrogel could effectively improve the hydrophilicity of the scaffold without negatively affecting its mechanical properties.

Although the combination of hydrogel and 3D-printed PCL scaffolds can effectively overcome the shortcomings of the original materials, the bone induction ability of the materials itself is still insufficient. To further improve the bone repair ability of scaffolds, some bioactive molecules should also be added. In this study, BG and SDF-1α were added to PCL and GelMA hydrogel, respectively, to jointly improve the osteogenic activity of the scaffold. BG is an important bioceramic material with a strong ability to bond with bone. BG can provide temporary support for tissue regeneration during bone healing. Furthermore, BG can ameliorate the hydrophobic properties and poor cell adhesion of polymeric materials and improve their mechanical properties [40]. During degradation, BG releases Ca2+, PO43− and Si4+, which contribute to bone regeneration. Our experimental results showed that the mechanical properties of the scaffold material were effectively improved after BG was added. Animal experiments further demonstrated that the addition of BG improved the effect of bone repair.

When tissue-engineered scaffolds are used to treat bone defects, the lack of stem cells in the repair site is often caused by poor integration of the grafts and host tissues. To solve this problem, in this study, SDF-1α was loaded into the scaffold material. SDF-1α is a very important chemokine of stem cells that can bind with CXCR4 on the surface of stem cells and then guide stem cells to migrate to a defect area [29]. Some studies have also found that SDF-1α can protect newborn cells from hypoxic damage, reducing the apoptosis rate and promoting the formation of new blood vessels, thereby creating a more favourable microenvironment for bone tissue regeneration [41, 42]. Our results show that SDF-1α can effectively recruit stem cells. Furthermore, when SDF-1α and BG were used together to modify the scaffold material, the scaffold showed the best effect on bone repair. The PCL@BG20-GelMA/SDF-1α group showed the largest area of new bone and large amounts of bone tissue ingrowth into the pores of the scaffold; additionally, collagen deposition at the defect site was also effectively improved. Furthermore, the scaffold showed high biocompatibility and no obvious toxicity to or side effects on the important organs (liver, spleen, kidneys, lungs and heart) of rabbits. Therefore, we believe that the PCL@BG20-GelMA/SDF-1α scaffold we prepared has good clinical development prospects and may play an essential role in bone tissue regeneration medicine.