The mice primary BMSCs were purchased from the Cyagen Biosciences Co. (MUBMX-01001). At stem cell state, the BMSCs were cultured with αMEM with 10%FBS and 1% penicillin-streptomycin antibiotics. As for osteogenic induction, 0.1uM dexamethasone (ST1254, Beyotime Biotechnology), 50ug/ml vitamin C (ST1434, Beyotime Biotechnology) and 10 mM β-glycerophosphate (G9140-5, Solaribio) were contained in αMEM complete medium. As for chondrogenic induction, 1% ITS (C0341, Beyotime Biotechnology), 0.1uM dexamethasone (ST1254, Beyotime Biotechnology), 37.5ug/ml vitamin C (ST1434, Beyotime Biotechnology), 1 mmol/L sodium pyruvate (ST1661, Beyotime Biotechnology) and 10 ng/ml TGF-β1 (PRP1017, Abbkine Scientific Co.) were contained in high glucose DMEM complete medium. The cells were cultured in the cell incubator (37 °C. 5% CO₂) and changed medium every 2 days.

The Lenti-virus was constructed by Syngen Technology. With the Fgfr2 overexpression Lenti-virus (Lenti-Fgfr2), a control virus was also constructed (Lenti-control). Both Lenti-Fgfr2 and Lenti-control were transfected into BMSCs and the cells were subsequently named as the Fgfr2 overexpression group and the Con group. After screening the optimum multiplicity of infection (MOI), we chose MOI = 50 for the transfection. The virus supernatant was used with 10ug/ml Polybrene (C0351, Beyotime Biotechnology) to transfect BMSCs for 8 h. After 48 h, the total RNA and protein were extracted to test the transfection efficiency.

The total RNA of BMSCs was extracted by RNAeasy™ Animal RNA Extraction Kit (R0026, Beyotime Biotechnology) according to the manufactures recommended procedure. After determining the concentration, reverse transcription was performed using the Evo M-MLV RT Kit (AG11728, Accurate Biology). The real-time qPCR process was done on the ABI ViiA 7(Life Technologies) with the SYBR Green Pro Taq HS Kit (AG11701, Accurate Biology) following the recommended protocol.

Total proteins of BMSCs were extracted using a total protein extraction buffer (P0013, Beyotime Biotechnology). BCA protein quantification kit (P0010S, Beyotime Biotechnology) was used to quantify the protein contents. The obtained protein was denatured with SDS loading buffer (D0071, Beyotime Biotechnology). Protein was fractionated by SDS polyacrylamide gel electrophoresis and incubated overnight at 4 °C with antibodies against FGFR2 (1:2000, ab10648, Abcam), RUNX2 (1:1000, ab236639, Abcam), COLII (1:1000, 28459-1-AP, Proteintech) and GAPDH (1:50000, AC033, Abclonal), respectively. After washing 3 times with TBST, the membranes were incubated with a secondary antibody. Finally, the proteins were detected by ECL luminous fluid (P0018S, Beyotime Biotechnology).

The CCK-8 assay was carried out followed by the protocol provided by the manufacturer (C0037, Beyotime Biotechnology). In brief, 1000 cells were plated in 96-well plates for cell proliferation detection. At 0 h, 24 h, 48 h, and 72 h, the CCK-8 kit was added and plates were cultured in the cell incubator for 2 h. The absorbance was detected with a microplate reader at 405 nm wavelength.

The vertical migration ability of BMSCs was performed by a transwell system (14,141, 8um, Polycabonate membrane, Labselect). BMSCs were seeded at a density of 1 × 10⁵ per well. After 24 h, collected the chamber and removed the cells on the upper side of the Transwell filter. The cells on the downside of the filter were stained with crystal violet (C0121, Beyotime Biotechnology) and photographed. Later, the cells were incubated with 200ul decoloring solution (33% glacial acetic acid) and detected the quantification of transwell results with a microplate reader at 570 nm wavelength.

The scratch healing experiment was performed using the scratch plug-in method. The scratch wound was made with a fixed scratch 500 μm wide. Suspensions were prepared with cells at a density of 3 × 10⁵/mL, and a 70 μl cell suspension was seeded into each plug-in. When the degree of fusion of the cells reached 85%, the plug-ins were removed. Cells were washed with PBS and cultured in αMEM complete medium. The width of the scratch was observed and imaged under a microscope (0h, 24h, 48h, 72 h).

For the CD200 and CD105 staining, the cells were digested and stained by the CD200 (12-5200-82, ThermoFisher Scientific) and CD105 (17-1051-82,ThermoFisher Scientific) antibodies for 30min. After 2 times washing, the mean fluorescence intensities of CD200 and CD105 were detected by BD FACS Aria and the results were analyzed by FlowJo 10.5.3.

The tunel staining was applied to BMSCs after 72 h serum shock following the instruction of Tunel staining kit (A112-02, Vazyme). After staining, images were taken under a confocal microscopy and the percentage of Tunel positive cells was counted by Photoshop.

After osteogenic induction for 7 days, BMSCs were fixed with 4% PFA and incubated with BCIP/NBT Alkaline phosphatase color developing kit (C3206, Beyotime Biotechnology) for 30 min. After washing by PBS, the staining cells were photographed by a stereoscope and a microscope.

After osteogenic induction for 14 days, BMSCs were fixed with 4% PFA and incubated with 0.2% Alizarin Red S Solution (G1450, Solarbio) for 30 min. After washing with PBS, the staining cells were photographed by a stereoscope and a microscope.

After chondrogenic induction for 14 days, BMSCs were fixed with 4% PFA and incubated with an Alcian blue Staining Kit (C0153S, Beyotime Biotechnology) for 1 h. After washing with PBS, the staining cells were photographed by a stereoscope and a microscope.

The light-induced GelMA was purchased from Engineering For Life Co. (5%, EFL-DM-60). BMSCs were mixed with the GelMA after it was configured and filtered according to the instructions. The mixed GelMA was then cured under 405 nm blue light (EFL-LS-1600-405). For in vitro experiments, the density of BMSCs was 5 × 10⁴/ml, and for in vivo experiments, the density of BMSCs was 1 × 10⁷/ml.

The live and dead assay was carried out using a Calcein/PI cell viability assay kit (C2015S, Beyotime Biotechnology). GelMA mixed with BMSCs in each well was added with 1 ml Calcein AM/PI solution, and incubated for 1 h at 37°C avoided light. Then, the live/dead cells were observed under a fluorescence microscope and a confocal microscope.

The 6-week-male C57BL/6 mice (21–23 g) were from Shanghai Bikai Keyi Biotechnology Co. (Certificate No. 20180006044831). A total of 24 mice were used (Figure 1). The mice were anesthetized by gas anesthesia using isoflurane. A round defect of 2 mm in diameter on the parietal bone was made by a trephine drill and then the GelMA with or without BMSCs was implanted into the defect. A blank group was also made as a control. After 6 weeks of repairation, the mice were sacrificed by cervical dislocation following anesthesia. All animal experiments were approved by the Animal Ethics Committee of Nanjing University (Permit No. IACUC——D2303101).

The skulls of mice in each group were fixed in 4% PFA and scanned by SCANCO vivaCT 80 (MedicalAG). The X-ray voltage was 57 kV and the current was 184μA, with a resolution of 10 μm per pixel. The bone volume/tissue volume (BV/TV) of the defected area was calculated and the 3D images of the defect area were also reconstructed. The Photoshop software was used to measure the area of the bone defect to obtain quantitative data.

The decalcified skulls were dehydrated and then embedded in paraffin. The skulls were then cut into sections for H&E and Masson Staining. For immunofluorescence staining, the sections were incubated with anti-OCN (1:200, A6205, Abclonal) and then incubated with secondary antibodies for inmmunohistochemistry. Slices were sealed with neutral gum and photographed under a microscope.

Quantitative data were presented as the mean ± standard error. The statistics were analyzed with an independent t-test (two-tailed) by SPSS 15.0. *p < 0.05, **p < 0.01, **p < 0.001.

We overexpressed Fgfr2 in BMSCs by Lenti-virus carrying a green fluorescent protein sequence so that the cells that were successfully transfected would show green fluorescence. First, we screened the optimum multiplicity of infection (MOI) of transfection in vitro (Figure 2B) and found that with an increase in MOI, the number of fluorescent cells and fluorescence intensity improved significantly. Considering the potential risk for cells survival condition with high MOI, we selected MOI = 50 as the transfection concentration for BMSCs (the transfection efficiency achieved at 80%).

To identify overexpression efficiency of Fgfr2, RNA and protein were also extracted (Figures 2C,D). The results showed that after Lenti-Fgfr2 transfection, the RNA level of Fgfr2 was significantly up-regulated (9.31-fold higher, p = 0.000), and the protein level was also raised (1.83-fold, p = 0.005), suggesting that the overexpression of Fgfr2 by Lenti-virus was indeed effective.

In order to understand the effect of Fgfr2 on the function of BMSCs itself, we first detected the influence of Fgfr2 overexpression in cell proliferation, migration, and surface markers. CCK8 assay (Figure 3A) showed that the Fgfr2 overexpression group had a faster proliferation rate compared to the control group at 24 h (1.04-fold, p = 0.000), and the raise was more significant at 48 h and 72 h (1.29-fold and 1.34-fold respectively, p = 0.000). To test cell migration ability, transwell and scratch assays were both performed as well (Figures 3B,C). These results showed that more BMSCs went across the transwell membranes in the Fgfr2 overexpression group (1.20-fold in transwell, p = 0.002) compared to the control group, and the scratch width was also narrower, indicating the stronger migration ability of BMSCs after Fgfr2 overexpression. In addition to its effect on proliferation and migration, Fgfr2 overexpression also enhanced the resistance of BMSCs to apoptosis. 72 h after serum shock, the percentage of tunel positive cells in Fgfr2-overexpressing BMSCs was significantly fewer compared with the control group (Figure 3E).

We further performed flow cytometry of CD105 and CD200 to test the changes in surface markers after Fgfr2 overexpression (Figure 3D). CD105 is a positive marker that is commonly used for BMSC identification, and CD200 can indicate colony-forming ability and osteogenic potential (Lin et al., 2013; Pontikoglou et al., 2016). We observed that the mean fluorescence intensities of CD105 and CD200 were significantly increased after Fgfr2 overexpression, especially CD105 (3.75-fold, p = 0.000). Considering this enhanced proliferation, migration and anti-apoptosis ability, we hypothesize that Fgfr2 overexpression also promoted the cell stemness of the BMSCs.

Fgfr2 has been reported to be highly expressed during the differentiation of BMSCs into osteoblasts (Su et al., 2014). Therefore, we investigated the osteogenic differentiation ability of BMSCs after Fgfr2 overexpression. BMSCs in both the control group and the Fgfr2 overexpression group were subjected to osteogenic induction for 0, 3, 7, and 10 days, and their RNA was then extracted for detection (Figure 4A). The results showed that compared to the control group, the osteogenic-related genes such as Runx2, Alp, Bmp4, and Col1a1 were obviously raised after osteogenic induction in the Fgfr2 overexpression group. On day 3, the up-regulation levels of Runx2, Alp, Bmp4, and Col1a1 were similar (1.43 to 1.55-fold), and at day 7, the up-regulation levels of Alp (2.19-fold, p = 0.031) and Bmp4 (1.97-fold, p = 0.000) were the most obviously. When it came to day 10, the up-regulation levels of Alp (2.78-fold, p = 0.001) and Col1a1 (1.78-fold, p = 0.026) were the highest.

After detecting the osteogenic potential in RNA levels, we also extracted proteins from the BMSCs at day 0 and day 10 (Figures 4B-D). Consistent with the above, RUNX2 expression was not up-regulated without osteogenic induction, but significantly increased at day 10 in the Fgfr2 overexpression group after osteogenic induction (1.29-fold, p = 0.002). These results suggest that Fgfr2 is a potential regulator of osteogenesis in BMSCs. Thus, we find that although overexpression of Fgfr2 can enhance the osteogenic potential, it still requires environmental stimulation to drive this process.

To examine the mineralization ability of BMSCs further, ALP and alizarin red stainings were performed. Obvious mineralized nodules could be easily observed under a light microscope in the Fgfr2 overexpression group at day 5 with osteogenic induction, but the control group only had small mineralized nodules (Figure 4E). Compared to the control group, the Fgfr2 overexpression group all showed better alkaline phosphatase activity (Figure 4F) and more mineralized nodules with alizarin red staining, demonstrating the superior osteogenic potential of BMSCs after Fgfr2 overexpression. Therefore, we find that overexpression of Fgfr2 can significantly promote osteogenic differentiation and mineralization in BMSCs.

Fgfr2 is involved not only in intramembrane osteogenesis but also in endochondral osteogenesis. In diseases resulting from abnormal Fgfr2 mutations, such as Apert syndrome (Fgfr2 P253R mutation), an obvious delay in the intracoronal ossification center has been observed that affects the structure of the skull base and long bones (Yin et al., 2008). Therefore, we speculated that Fgfr2 overexpression might also promote the differentiation of BMSCs into chondrocytes. To test this, we applied a chondrogenic induction environment to the BMSCs and extracted RNA at day 0, day 3, day 7, and day 10 for chondrogenic marker detection (Figure 5A). The results showed that Col2a1, Col10a1, and Aggrecan were all up-regulated with chondrogenic induction and that Col10a1 was the most dramatically up-regulated gene (3.88-fold, p = 0.004, day 3; 2.07-fold, p = 0.001, day 7; 2.70-fold, p = 0.000, day 10). Subsequently, the protein of BMSCs at 0 and 10 days after chondrogenic induction was extracted, and the expression of COLII was detected (Figure 5C). Consistent with the RNA results, when compared to the control group, COLII showed significant up-regulation in the Fgfr2 overexpression group 10 days after induction (1.65-fold, p = 0.001).

Alcian blue staining further demonstrated the effect of Fgfr2 overexpression on the chondrogenic ability of BMSCs (Figure 5B). 14 days after chondrogenic induction, the control group did not show the formation of chondrogenic pellets, but the Fgfr2 overexpression group had a scattered distribution of chondrogenic pellets, and the cells were more deeply stained with Alcian blue. These results indicated that Fgfr2 not only participated in the osteogenic differentiation of BMSCs but also promoted chondrogenic differentiation. We find that Fgfr2 is therefore important for both intramembrane osteogenesis and endochondral osteogenesis in BMSCs.

After demonstrating that Fgfr2 overexpression could significantly promote the proliferation and osteogenic and chondrogenic differentiation of BMSCs, we wondered whether Fgfr2 could be used as a novel gene editing site for the treatment of critical cranial bone defects. To investigate this, we selected light-induced porous gelatin methacryloyl hydrogel (5% GelMA) as a scaffold for the BMSCs since it has been shown to be a good three-dimensional scaffold material for cell cultures by supporting the proliferation and multi-lineage differentiation of mesenchymal stem cells (Bashir et al., 2023). We first reconfirmed in vitro whether the BMSCs could survive in this GelMA material by calcein/PI staining assay and found that the BMSCs survived quite well in GelMA and proliferated steadily over time (Figure 6). Additionally, many BMSCs began to spread around and connect with each other in the GelMA, suggesting that the BMSCs had begun to form a natural reticular structure. These results indicated that light-induced GelMA did indeed have good biocompatibility with the BMSCs.

To evaluate the effect of the BMSCs with Fgfr2 overexpression on the bone regeneration process in vivo, we used GelMA and/or BMSCs from the control group or the Fgfr2 overexpression group as appropriate to treat bone defects on mouse skulls (Figure 7A). By visual observation and micro-CT scan, the Fgfr2 overexpression BMSCs carried by GelMA greatly enhanced both the osteogenesis and regeneration in the area of the cranial defect (Figures 7B,C), and single GelMA or control BMSCs carried by GelMA also had therapeutic effects to a certain extent. Compared to the blank group, the defect area (Figure 7D) in the GelMA group decreased by 22.24% (p = 0.006), the control BMSCs carried by GelMA decreased it by 44.91% (p = 0.000), and the Fgfr2 overexpression BMSCs carried by GelMA decreased it by 79.99% (p = 0.000). Consistent with the above, the BV/TV showed an upward trend as well. Compared to the blank group, the BV/TV in the GelMA group increased by 24.07% (p = 0.010), the control BMSCs carried by GelMA increased it by 88.65% (p = 0.000), and the Fgfr2 overexpression BMSCs carried by GelMA increased it by 140.41% (p = 0.000).

Histological staining further demonstrated the great repair abilities of the Fgfr2-overexpressed BMSCs combined with GelMA on cranial bone defects (Figure 7F). Using H&E staining, obvious new bone formations could be observed in the Fgfr2 overexpression group. The defect areas were entirely filled with fibrous soft tissues as well as new bone. Furthermore, under Masson’s trichrome staining, collagen fibers that were relatively loose were colored blue, and new generated bone was colored red. With the treatment of the control BMSCs carried by GelMA, obvious collagen fibers also appeared in the defect area. However, compared to the control BMSCs, more new bone formation could be observed after the treatment of Fgfr2 overexpression BMSCs carried by GelMA, indicating the promotion of osteogenesis after Fgfr2 overexpression. Corresponding to the above, the Fgfr2 overexpression group had the highest OCN expression in the defect area as well. These results suggest that Fgfr2 overexpression in BMSCs fulfilled significant osteogenic and potential chondrogenic ossification functions in bone regeneration. Compared to the blank group, the GelMA group and the control BMSCs group, the Fgfr2-overexpressed BMSC material significantly advanced ossification and regeneration in bone defect areas. In conclusion, we find that Fgfr2 can be used as an effective gene editing target for the treatment of critical cranial bone defects in mice.