In this study, a total of 120 healthy and clean adult female SD rats, aged 4 weeks and weighing 200–250 g, were used and provided by the Experimental Animal Center of West China Clinical Medical College of Sichuan University. Ninety-six of them were only used to construct bone defect models, while the others were used for BMSC isolation and culture. All animal experiments conducted in this study were approved by the animal management and use committee of the West China Clinical Medical College of Sichuan University (approval number: SCXK20150012). The rats were put into a cage 1 week before the experiment to adapt to the environment. Three rats/cage were served with sufficient conventional animal feed, maintaining the room temperature at 21°C, 60% air humidity, and 12-h circadian rhythm.

The GelMA was synthesized following the procedure described. Briefly, 10 g of gelatin derived from porcine skin was dissolved in 100 ml of PBS in a cleaned Erlenmeyer flask with magnet fish. Then 5 ml of methacrylic anhydride was added very slowly and dropwisely with a syringe pump, and the emulsion was rotated (240 rpm) at 50°C for 2 h and covered with an aluminum foil. Dialysis membrane (Pectro/Por molecular porous membrane tubing, Fisher Scientific, USA) was prepared by cutting them in proper sizes and immersed them into distilled water to soften them. One side was closed by twisting the membrane end and making a knot. The GelMA was transferred with a funnel into the membranes. The second end of the membrane was closed the same way as the first. Membranes were placed into distilled water in a 5-L plastic beaker, and the dialysis was ran at 40°C for 7 days with a magnetic stirrer and covered. The GelMA solution was then quickly and successively filtered with a coffee filter and sterile vacuum Express Plus (0.22 µm) Milipore filtration cup. The sterilized polymer was transferred into 50-ml Falcons and horizontally stored at −80°C for 2 days. The frozen GelMA was lyophilized for 3 days and stored in the dark until use.

The GelMA was dissolved in D₂O for analysis using 400-Hz nuclear magnetic resonance (Bruker AVANCE AV Ⅱ-400 MHz). The degree of methacrylate substitution was determined by the formula: 1 − (lysine integration signal of GelMA/lysine integration signal of unsubstituted Gelatin) × 100% (Brinkman et al., 2003; Nichol et al., 2010; Loessner et al., 2016). The morphology of the GelMA hydrogel was observed by scanning electron microscope (SEM). Dynamic Mechanical Analyzer (TA Instruments, Q-800, USA) was used to test the storage modulus and loss modulus of the GelMA hydrogel. The rheological properties of the GelMA hydrogel were analyzed by rheometer (MCR302, Anton Paar) (Zhou, 2021).

The GelMA solution with a concentration of 5% was prepared by using deionized water, and a photocrosslinking agent (Irgacure 500, BASF Corporation, Germany) with a dosage of 0.25% of GelMA solution (w/v) was added. After mixing the GelMA solution with the photocrosslinking agent evenly, a mixed solution was obtained. The mixed solution was filtered through a 0.22-µm filter membrane and mixed with BMSCs to make the cell suspension. The cell density in the suspension was 2 × 10/ml (Stanovici et al., 2016). The bone defect model was constructed according to the needs, and the PDMS mold of the corresponding size (4 × 4 × 5 cm) was prepared according to the bone defect model. The suspension of the GelMA–BMSCs was added into the PDMS mold, and UV irradiation (λ = 365 nm, 40 s) was given. After crosslinking, the GelMA hydrogel bone repair scaffold containing BMSCs was obtained.

The previously prepared BMSC–GelMA hydrogel scaffolds were cultured in normal medium. LIVE/DEAD assay was applied to evaluate the cell viability at 1, 3, 7, and 14 days after culture. The hydrogel scaffolds were washed with PBS and stained with Calcein AM (0.5 μl ml⁻¹) and ethidium homodimer-1 (EthD-1, 2 μl/ml) for 2 h at 37°C. The samples were observed under an inverted fluorescence microscope (Nikon, Japan). The number of living and dead cells in the scaffold was counted, and the percentage of living cells in the total number of cells was calculated. Cell Counting Kit-8 (CCK-8, Dojindo, Japan) was used to detect the influence of scaffolds on cell activity on days 1, 4, 7, and 14. A 10-µl CCK8 solution was add to each well of the 96-well plate, incubated at 37°C for 2 h, and the absorbance at 450 nm was measured with a microplate analyzer (Thermo Scientific, Shanghai, China). Each experiment was repeated at least three times.

SD rats were randomly divided into four groups (16 rats in each group): group A was the model control group, group B was the GelMA hydrogel scaffold group (control group), BMSCs were in group C (control group), and group D was the GelMA hydrogel scaffold containing BMSCs (experimental group). The rats were anesthetized by intraperitoneal injection of pentobarbital, the hair of the left hind limb was cleaned, sterilized with alcohol, and was covered with a sterile dressing. A longitudinal incision was taken from the posterior middle posterior tibia, and the subcutaneous and muscular layers were incised. A 5-mm-long segmental bone defect was created with a bone saw. In group A, the intramedullary nail was used for retrograde fixation. In group B, the GelMA hydrogel scaffold was implanted and fixed with intramedullary nails. BMSC suspension was injected into the bone defect of group C. In group D, BMSC-loaded GelMA hydrogel scaffolds were implanted and fixed with intramedullary nails. The muscle, subcutaneous tissue, and skin were sewn up step by step. At weeks 4 and 8 after surgery, the bone tissue in the bone defect area was taken for histomorphological test, biomechanical property test, and micro-CT test.

Bone tissue was taken from the bone defect area at weeks 4 and 8 after surgery, and hematoxylin–eosin (HE) staining was performed. The isolated specimens were decalcified through ethylene diamine tetraacetic acid (EDTA, Sigma, USA), dehydrated by 80, 90, and 100% ethanol, and embedded in paraffin. The specimens were then cut into 5-μm sections and stained with hematoxylin–eosin (HE), and observed under a BX53 microscope (Olympus, Japan). The new bone and new blood vessels were quantitatively analyzed. Experimental data were expressed as mean ± SD.

On the fourth and eighth weeks after surgery, the tibia of the rats was taken for the biomechanical test. The residual soft tissue was removed, and the tibial tip was trimmed to an appropriate length so that the bone defect was located in the middle of the sample. The three-point bending test (Ruige Technology, China) was performed on the biomechanical tester to measure the bending stiffness and ultimate load to evaluate the biomechanical properties. Experimental data were expressed as mean ± SD.

Bone tissue was taken from the bone defect area on the eighth week after surgery for micro-CT detection, 3D reconstruction of the bone defect area, and quantitative analysis of bone mass and bone density. Specimens were collected and fixed with 4% paraformaldehyde for micro-CT analysis. Micro-CT scanning was performed by a Guantum GX microCT imaging system (Perkin Elmer, USA) with the following settings: acquisition, 36; voxel, 50 µm; reconnaissance, 25. The Guantum GX software was used for 3D reconstruction. Experimental data were expressed as mean ± SD.

Statistical analyses were performed using Statistical Package for the Social Sciences (SPSS 19.0, IBM, New York, NY, USA). All data were expressed as the mean value ± standard deviation (SD). Statistical comparisons were conducted using analysis of variance (ANOVA) in which a p-value of less than 0.05 was considered statistically significant.

The overall process of this study is shown in Figure 1. As shown in Figure 2A, the GelMA can be rapidly dissolved in deionized water. After freeze drying of the GelMA photocured hydrogel, the porous structure of the hydrogel was observed by scanning electron microscope (SEM), which can be seen in Figure 2B. As can be seen from Figure 2C, the proton peak of methacrylic acid can be observed in the range of 5–6 ppm, which indicates that the GelMA has been successfully synthesized. In addition, in the range of 2.8–2.95 ppm, it was found that compared with the lysine proton peak of gelatin, the lysine proton peak of the GelMA was significantly weakened, indicating that the target reaction amino acid was consumed. The methacrylate substitution degree of the GelMA was calculated to be 87.6%. As shown in Figure 2D, the storage modulus of the GelMA hydrogel reaches 14.22 kPa at 10 Hz, and this mechanical strength can maintain the stability of the scaffold. In Figure 2E, as the shear rate increases from 0 to 100 s⁻¹, the viscosity of the GelMA solution decreases gradually and remains stable, exhibiting shear thinning properties, which indicate that the GelMA is injectable. In Figure 2F, when the temperature decreases from 40°C to 10°C, the viscosity of the GelMA solution increases rapidly below 22°C to the point of physical gel formation.

The staining results of BMSCs with live/dead cells in the GelMA hydrogel bone repair scaffolds containing BMSCs are shown in Figure 3A (fluorescence staining of live/dead cells) and Figure 3B (percentage of the number of live cells in the total number of cells). Within 1–3 days of culture, the cells were spherical under the influence of a low adherent matrix. Over time, the cells increased and stretched, compared with the first day. Three-dimensional (3D) thermal imaging showed that fluorescence intensity increased over time. The staining results of live/dead cells showed that the survival rate of BMSCs in BMSC-loaded GelMA hydrogel bone repair scaffolds was high. After 1, 3, 7, and 14 days of culture, the number of living cells in the scaffold gradually increased, and the percentage of living cells exceeded 90%. In cell activity and toxicity tests performed on days 1, 4, 7, and 14, the results showed that the OD value increased with time (Figure 3C). The biocompatibility testing results showed that the GelMA hydrogel scaffold had good cell compatibility, and BMSCs could proliferate well in the scaffold.

To investigate whether the BMSC-loaded GelMA hydrogel bone repair scaffold can promote bone regeneration in the defect area, histological analysis was performed at weeks 4 and 8, respectively. Figure 4 shows the HE staining results of bone tissue in the bone defect area. As can be seen from Figure 4A, bone growth in the bone defect area in the BMSC group and the BMSC-carrying GelMA hydrogel scaffold group was vigorous at the fourth and eighth weeks after surgery, and new blood vessels were observed in the regenerated bone area. At the same time, compared with the BMSC group, the BMSCs-containing GelMA hydrogel scaffold group had more new bone formation and more mature tissue structure. In contrast, in the model control group and the GelMA hydrogel scaffold group, only a small amount of new bone formation occurred in the bone defect area, accompanied by more fibrous connective tissue formation.

Image Pro-Plus 6.0 software was used for quantitative analysis of the new bone and new blood vessels. The percentage of the new bone area was calculated according to the new bone area/total defect area × 100%, and the density of new blood vessels was measured according to the number of new blood vessels/bone defect area. Figures 4B , C show that the number of new bones and the density of new blood vessels increased in each group from weeks 4 to 8. The number of new bones and density of new blood vessels in the BMSC group was significantly higher than those in the model control group and GelMA hydrogel scaffold group at each time point (p < 0.01). However, the number of new bones and density of new blood vessels in the BMSC-carrying GelMA hydrogel scaffold group were significantly higher than those in the BMSC group (p < 0.05). The results show that the GelMA scaffold has the ability to promote the growth of new bone, and new blood vessels, that is, it has a good ability to promote bone regeneration.

The biomechanical property test results are shown in Figure 5. According to Figures 5A , B, at the fourth week after surgery, the bending stiffness and ultimate load of the BMSC-loaded GelMA hydrogel scaffold group were significantly higher than those of the BMSC group (p < 0.05), the model control group, and the GelMA hydrogel scaffold group (p < 0.01). There was no statistically significant difference between the model control group and the GelMA hydrogel scaffold group (p > 0.05). The biomechanical performance at week 8 was similar to that at week 4, and the difference in bending stiffness and ultimate load between the BMSC-containing GelMA hydrogel scaffold group and the BMSCs group was more significant (p < 0.01). The experimental results show that the BMSC-loaded GelMA hydrogel bone repair scaffold can improve the mechanical strength of the defect tibia.

Micro-CT detection was performed on the bone defect area at the eighth week after surgery, and three-dimensional reconstruction was conducted. Figure 6A (micro-CT 3D reconstruction model) shows that the bone bridge and callus formation in BMSC-carrying GelMA hydrogel scaffold was significantly better than that in the BMSCs group, GelMA hydrogel scaffold group, and model control group. The quantitative results of bone mass (Figure 6B) and bone mineral density (Figure 6C) were consistent with the above: the mean bone mass and bone mineral density of the BMSC-loaded GelMA hydrogel scaffold group were significantly higher than that of the BMSC group alone, the GelMA hydrogel scaffold group, and the control group (p < 0.01). There was no statistically significant difference between the GelMA hydrogel scaffold group and the control group (p > 0.05).