The neurofibromatosis type I gene promotes autophagy via mTORC1 signalling pathway to enhance new bone formation after fracture

Abstract Bone fracture is one of the most common injuries. Despite the high regenerative capacity of bones, failure of healing still occurs to near 10% of the patients. Herein, we aim to investigate the modulatory role of neurofibromatosis type I gene (NF1) to osteogenic differentiation of bone marrow–derived mesenchymal stem cells (BMSCs) and new bone formation after fracture in a rat model. We studied the NF1 gene expression in normal and non‐union bone fracture models. Then, we evaluated how NF1 overexpression modulated osteogenic differentiation of BMSCs, autophagy activity, mTORC1 signalling and osteoclastic bone resorption by qRT‐PCR, Western blot and immunostaining assays. Finally, we injected lentivirus‐NF1 (Lv‐NF1) to rat non‐union bone fracture model and analysed the bone formation process. The NF1 gene expression was significantly down‐regulated in non‐union bone fracture group, indicating NF1 is critical in bone healing process. In the NF1 overexpressing BMSCs, autophagy activity and osteogenic differentiation were significantly enhanced. Meanwhile, the NF1 overexpression inhibited mTORC1 signalling and osteoclastic bone resorption. In rat non‐union bone fracture model, the NF1 overexpression significantly promoted bone formation during fracture healing. In summary, we proved the NF1 gene is critical in non‐union bone healing, and NF1 overexpression promoted new bone formation after fracture by enhancing autophagy and inhibiting mTORC1 signalling. Our results may provide a novel therapeutic clue of promoting bone fracture healing.


| INTRODUC TI ON
Bone fracture healing cascades involve a series of complex biophysiological and pathological processes, which need the participation of various cellular and biochemical cues. Depending on the nature of the injury and the post-injury care, it can take several months for bone fracture to repair. Although the bone tissue has strong regenerative capacity among the human organs, for certain type of fractures such as tibia diaphysis, or for patients with limited healing capabilities, delayed repair or facture non-union may happen, 1 leading to impaired rehabilitation and poor repair quality. Current therapies to address failed fracture repair include non-surgical and surgical treatments. The most common non-surgical treatment is to use electrical stimulator; however, with different devices used, such as direct current, inductive coupling and capacitive coupling, the effectiveness may be inconsistent for treating non-union. 2 Surgical treatments on the other hand will usually introduce secondary trauma and raise the risk of infection. Therefore, new therapies to enhance bone formation after fracture are in need.

Neurofibromatosis type I (NF1) is an inherited autosomal dominant disorder caused by mutation in the tumour suppressor NF1
gene. This disease affects about 1:3000 world population, and the symptoms include pigmentary lesions, dermal neurofibromas, and in some cases, skeletal abnormalities. 3 The NF1 gene encodes neurofibromin 1, which is a Ras-GTPase activating protein (Ras-GAP).
The neurofibromin has been shown to be essential in osteoblast functioning; therefore, mutation in the NF1 gene results in skeletal abnormalities such as scoliosis, unilateral growth and congenital pseudoarthroses of long bones. 4 Moreover, the NF1 gene expression is also found in adult osteoblasts and osteoclasts, as well as hypertrophic chondrocytes, which presents during endochondral ossification. 5 The involvement of the NF1 gene and skeletal development provides a potential therapeutic target to modulate bone fracture healing.
Bone marrow mesenchymal stem cells (BMSCs) are multipotent stem cells that bestow the high regenerative potential of human skeleton. During bone fracture repair, BMSCs are activated and differentiated into osteoblasts, and bone marrow-derived macrophages (BMDM) differentiated into osteoclasts. The balance between bone resorption and formation is critical to fracture repair.
Increased activity of osteoclasts will result in higher rate of bone resorption that compromises the repair efficiency. Autophagy is the body's self-cleaning process that is essential for cell viability, differentiation and metabolism. 6 It is reported that autophagy plays an important role in bone cell function and pathology. 7 Specifically, for BMSCs differentiation, emerging evidences proved that autophagy was involved in osteogenic differentiation and bone mineralization. 8,9 Qiao et al 10 reported that the NF1 overexpression induced autophagy and inhibited mTOR/P70S6K signalling in ovarian carcinoma cells and in turn inhibited ovarian carcinoma cell proliferation and invasion. Mammalian target of rapamycin (mTOR) functions critically in the metabolic signalling, proliferation, autophagy and cellular fate. 11 Park et al 12 reported that modulating mTOR signalling resulted in different cell fate such as autophagy vs proliferation balance via binding and phosphorylating autophagy-related gene 14 (ATG14). However, the role of NF1/autophagy functioning axis is yet to be investigated in osteogenic differentiation and bone formation.
In our current study, we evaluated the modulatory role of the NF1 gene on bone formation during fracture repair, and we investigated in depth of the possible mechanisms. We found the NF1 gene was able to enhance autophagy via inhibiting mTOR complex 1 (mTORC1) signalling, leading to pro-osteogenic and anti-osteoclastic differentiation. Our study demonstrated NF1 may be one of the critical therapeutic targets to promote bone fracture healing.

| BMSCs isolation and culture
Bone marrow mesenchymal stem cells were isolated from Sprague-Dawley (SD) rat femur. In brief, the bone marrow was flushed using a disposable aseptic syringe with antibiotic supplemented phosphatebuffered saline (PBS) (Gibco). The cell suspension was centrifuged at   then, BMSCs were cultured in osteogenic medium, and BMDMs were cultured in osteoclast differentiation medium. For the in vivo study, the GFP or NF1 lentivirus vectors were injected to the fracture site with a microsyringe (Hamilton).

| Quantitative real-time polymerase chain reaction (qRT-PCR)
Total RNA was extracted using TRIzol reagent (Invitrogen) and purified with RNeasy mini kit (Qiagen) following the manufacturer's protocol. Total quantity of 1 µg of RNA was reverse transcript to cDNA with High Capacity cDNA kit (Thermo Fisher Scientific). qRT-PCR was performed using SYBR Green master mix (Thermo Fisher Scientific), and the relative expression of the NF1 gene was calculated using the 2 −∆∆Ct method and normalized to GAPDH. The sequence information of primers used in this study was as listed below:

| Western blot
Total proteins from BMSCs or bone tissues were extracted using Radio-Immunoprecipitation Assay Lysis Buffer (Beyotime). Same

| Flow cytometry
Bone marrow mesenchymal stem cells were cultured in growth medium, and cells from passage 3 were harvested and stained with CD90, CD44, CD34 and CD31. Briefly, cells were digested with 0.05% trypsin and washed with cold PBS. Cell pellets were resuspended in blocking buffer at concentration of 1 × 10 6 cell/ mL. Then, 100 µL of the cell suspension was transferred to a 5 mL culture tube and stained with CD90, CD44, CD34 and CD31 antibodies at concentration of 1:500. All antibodies were purchased from Biolegend. After 2 washes with cold PBS, the stained cells were resuspended in 500 µL PBS and then analysed by flow cytometry.

| ALP and Alizarin red staining
Bone marrow mesenchymal stem cells that underwent osteogenic induction were fixed with 4% paraformaldehyde for 10 minutes.
The ALP staining was conducted with the BCIP/NBT regent kit (Beyotime), and Alizarin Red staining, cells were stained with Alizarin Red solution (Cyagen Biosciences). All procedures were carried according to the manufacture's protocols.

| Immunofluorescence staining
The BMSCs layers were washed with PBS and fixed in 4% paraformaldehyde (PFA) for 15 minutes. The cells were blocked in 1% BSA in PBS for 30 minutes at room temperature and then stained in anti-p-mTORC1 (1:250) antibody at 4°C overnight. After staining with secondary antibody for 1 hour at room temperature, the cell layers were washed for 3 times and counterstained DAPI for 10 minutes. Then, samples were imaged with fluorescence microscopy (Zeiss).

| Osteoclast differentiation
Bone marrow (BM) cells collected from SD rat femur were cultured in α-Minimum Essential Medium (α-MEM) for 3 hours to initiate adherence; then, the non-adherent cells in the medium were collected and re-seeded at a density of 1 × 10 6 cells/mL. The cells were cultured in complete medium with macrophage colony-stimulating factor (M-CSF) (20 ng/mL; Sigma Aldrich) for 3 days; then, the medium was refreshed every 3 days with complete medium supplemented with M-CSF (20 ng/mL) and receptor activator of nuclear factor kappa-Β ligand (RANKL) (20 ng/mL; Sigma Aldrich).

| Tartrate-resistant acid phosphatase (TRAP) staining and bone resorption
The differentiated osteoclast cells were fixed with 4% PFA and stained with TRAP-kit (Sigma) according to the manufacturer's instructions.
Mature osteoclasts were counterstained with methylene blue. For bone resorption assay, the differentiated osteoclast cells were cultured on the pre-made dentin discs for 12 days. The resorption pits of the bone slices were imaged with bright field microscope (Zeiss).

| Enzyme-linked immunosorbent assay (ELISA)
The medium of the differentiated osteoclasts was refreshed with serum-free α-MEM 24 hours prior to harvest. At harvest, the supernatant was collected and analysed using C-telopeptide of type I collagen ELISA kit (Biocalvin) for CTX and N-telopeptide of type I collagen ELISA kit (Biocalvin) for NTX.

| Rat model of bone fracture
Healthy male SD rats (around 2 months old) were purchase from SLAC. All the animal experimental procedures were approved by the Animal Experiment Committee of Hunan Children's Hospital (Changsha, China) and performed in accordance with the guidelines and regulations. The rats were anesthetized with an intraperitoneal injection of ketamine, and the middle femoral fracture surgery was performed on the right femur. In brief, we made a transverse incision and carefully removed the skin and subcutaneous tissue to expose the right femoral bone shaft. Fracture was created by a metal-wire saw. To fix the fracture site, a Kirschner wire was inserted to align the fracture site. Then, the subcutaneous incision was sutured close.
In the non-union group, the periosteum of femur was removed as much as possible. The determinations were performed at 8 weeks post-surgery.

| Histology
The femurs of each group of SD rats were decalcified in 10% ethylenediaminetetraacetic acid (EDTA) (pH 7.0) and embedded in paraffin according to standard protocol. The sectioned slices were stained with haematoxylin and eosin, and Masson's trichrome according to the manufacture's protocols.

| Statistics
All experiments were repeated independently for at least three times with n = 6 in each group. Statistical analyses were performed using Prism software (GraphPad Prism 8, USA). One-way analysis of variance (ANOVA) or Student's t test was used to analyse the data, and P < .05 was considered statistically significant.

| The NF1 gene expression is inhibited in nonunion model
We established rat bone fracture model to evaluated the difference of normal bone fracture healing from non-union. Western blot data demonstrated osteogenic protein expressions in the normal fracture and non-union groups at 4 weeks post-surgery. In comparison with normal fracture group, non-union rats showed significantly less osteogenic protein Osterix, Runx2, ALP and OCN expressions, indicating poor osteogenesis ( Figure 1A). We found significant difference in the NF1 gene expression and protein production in normal fracture and non-union groups ( Figure 1B), as non-union rat expressed significantly lower NF1 in comparison with normal fracture rats on both transcript and protein levels. Elevated ERK activity was also observed in non-union bones ( Figure S1A). These results demonstrate osteogenesis and the NF1 gene are both inhibited in non-union models, indicating potential correlation between NF1 and osteogenesis.

| The NF1 gene enhances osteogenic differentiation via promoting autophagy
We next evaluated how NF1 influenced osteogenic differentiation of BMSCs. Flow cytometry graph presented the isolated cells stained positively in CD90 and CD44, whereas negatively in CD34 and CD31, exhibiting the standard phenotype of BMSCs (Figure 2A).

| The NF1 gene promotes autophagy via inhibiting mTORC1 signalling
To investigate the impact of the NF1 gene on mTORC1 signalling pathway, we evaluated mTORC1-related protein expressions by Western blotting and immunostaining. Western blotting data showed overexpression of NF1 inhibited the phosphorylation of mTORC1, S6K1 and 4EBP1 ( Figure 3A). Moreover, immunostaining of p-mTORC1 confirmed that NF1 overexpression suppressed phosphorylation of mTORC1 ( Figure 3B). Therefore, the possible mechanism of NF1 promoting autophagy is via mTORC1 signalling inhibition.

| mTORC1 signalling impacts NF1-mediated BMSCs osteogenic differentiation
We next evaluated the role of mTORC1 signalling in NF1-mediated BMSCs differentiation. Western blot showed Lv-NF1 group inhibited p-mTORC1, S6K1 and 4EBP1 expressions, while activating mTOR by 3BDO significantly enhanced the p-mTORC1, S6K1 and 4EBP1 expressions. When treating BMSCs with both NF1 overexpression and 3BDO, the suppressive effect on mTORC1 signalling by NF1 was weakened ( Figure 4A). The autophagy activity was also impacted by mTOR activation and NF1 overexpression. The NF1 overexpression promoted autophagy activity by enhancing LC3 II/I and Beclin1 expressions and inhibiting p62 expression. Similar effect was achieved by using MEK inhibitor PD0325901 on BMSCs ( Figure S1C). Lv-GFP + 3BDO group inhibited autophagy activity by inhibiting LC3II/I and Beclin1 and enhancing p62. Treating NF1 overexpressing cells with 3BDO attenuated Lv-NF1-activated autophagy ( Figure 4B). ALP activity and cell mineralization were enhanced by NF1 overexpression and suppressed by 3BDO ( Figure 4C and D). Similar trend was seen in Western blotting data detecting osteogenic protein levels of Osterix, Runx2, ALP and OCN ( Figure 4E). These results indicate mTORC1 signalling is involved in NF1-mediated BMSCs osteogenic differentiation. Activation of mTORC1 pathway inhibits the osteogenic effect induced by NF1 overexpression.

| NF1 overexpression inhibits osteoclast differentiation
The efficacy of the NF1 gene overexpression was validated by qRT-PCR. Lv-NF1 group demonstrated significantly higher NF1 gene expression in BM cells, indicating a successful transfection ( Figure 5A). TRAP staining and bone resorption assay demonstrated inhibited osteoclast differentiation and less bone resorption in Lv-NF1 group in comparison with control and Lv-GFP groups ( Figure 5B and C).
Bone metabolism markers CTX and NTX were also significantly reduced in Lv-NF1 group ( Figure 5D). Western blotting data further demonstrated osteoclast-related markers ITGβ3, CALCR and CTSK were reduced in Lv-NF1 group ( Figure 5E). Autophagy activity was promoted by Lv-NF1, as LC3II/I and Beclin1 expressions were enhanced, while p62 was inhibited in Lv-NF1 group ( Figure 5F). Similar with Lv-NF1, MEK inhibitor treatment also enhanced autophagy activity ( Figure S1D). These results indicated NF1 overexpression inhibits osteoclast differentiation.

| NF1 promotes bone formation during fracture healing by enhancing autophagy via mTORC1 signalling
With rat non-union bone fracture model, we examined the effect of NF1 overexpression to bone formation. H&E and Masson's trichrome staining presented histological evidences of newly formed bones. As shown in Figure 6A and B, more bony formation was observed in Lv-NF1 group, while control groups have less bony formation in the fracture region. Western blot data of osteogenic proteins Osterix, Runx2, ALP and OCN proved Lv-NF1 enhanced osteogenesis in rat non-union bone fracture model ( Figure 6C).

F I G U R E 1
The NF1 gene expression in normal and non-union bone fracture model. A, Western blot of osteogenic markers Osterix, Runx2, ALP and OCN and the intensity analysis of normal and non-union bone fracture model. B, The NF1 transcript and NF1 protein level in normal and non-union bone fracture model. All data were presented as mean ± SD. *P < .05; **P < .01. n = 6 in each group | 11529 TAN eT Al.
Autophagy-related marker LC3 II/I and Beclin1 was also enhanced by Lv-NF1 treatment, while p62 was suppressed in the meantime ( Figure 6D). Western blotting data demonstrated p-mTORC1, S6K1 and 4EBP1 were inhibited by Lv-NF1 treatment ( Figure 6E), indicating mTORC1 inhibition played a role in NF1-mediated osteogenesis.
ERK activity was also inhibited by NF1 overexpression, as p-ERK F I G U R E 2 The NF1 gene activates autophagy and promotes osteogenic differentiation of BMSCs. A, Flow cytometry analysis of CD90, CD44, CD34 and CD31. B, The NF1 gene expression of BMSCs after transfection. C, Autophagy-related protein levels of control and transfected BMSCs. D, ALP and (E) Alizarin Red staining of control and transfected BMSCs. F, Western blot of osteogenic markers Osterix, Runx2, ALP and OCN in control and transfected BMSCs with GAPDH as reference gene. All data were presented as mean ± SD. *P < .05; **P < .01; ***P < .001. n = 6 in each group expression was inhibited in Lv-NF1 group ( Figure S1B). The data presented above suggest NF1 overexpression enhances autophagy via inhibiting mTORC1 signalling, to promote bone formation during fracture healing.

| D ISCUSS I ON
Bone fracture healing is a complex biological process that requires the participation of complicated biochemical markers and signalling pathways. Despite the bone is one of the few tissues with high regenerative capacity, the healing process sometimes fails with the development of non-union. 13 In the present study, we evaluated the NF1 gene expression in normal healing bones and non-union bones, and found down-regulated NF1 expression in the non-union bones. Overexpressing the NF1 gene enhanced autophagy and osteogenic differentiation in BMSCs, and suppressed osteoclastic differentiation in BM cells via mediating mTORC1 signalling. Out results for the first time reveal the potential mechanism of the modulatory role of NF1 in the bone formation during fracture healing process.
The NF1 gene encodes neurofibromin 1 that regulates RAS/ MAPK pathway. It has relatively high mutation rate, and the mutation results in alteration of cell growth and neural development.

Previous study conducted by Kuorilehto et al associated the NF1
gene expression with mouse fracture healing and rat pseudarthrosis, 14 and reported normal NF1 function was needed to achieve a normal bone fracture healing. When NF1 function is impaired, reduced osteogenic activity is seen in NF1-deficient model. 15 In our current study, we found reduced NF1 expression in non-union bones as compared to normal fracture bones, supporting the theory that NF1 is critical in bone fracture healing process. Autophagy is the process that allows cells to recycle damaged cellular components, and it plays an essential role in multiple tissue functions such as liver 16 and bone, 17 as well as in diseases such as cancers. 18 In skeletal tissue regeneration process, studies have shown autophagy promotes osteogenic differentiation in BMSCs, 19 and impairment of autophagy may cause abnormal ageing in skeletal tissues. 20 Although the NF1 gene The results from the current study proved as the key regulator of Ras/MAPK pathway, NF1 overexpression inhibited ERK and mTOR signalling, leading to an enhanced autophagy activity to promote osteogenesis. Our study for the first time associated the NF1 gene expression with mTORC1 signalling and autophagy activity, revealing the potential molecular mechanism of the NF1-regulated osteogenesis. The detailed regulatory mechanism will be investigated in our future study to establish the signalling axis.
Osteoclastic bone resorption is essential in bone remodelling.
However, overactivated osteoclastic bone resorption may cause excessive bone loss and lead to delayed new bone formation and osteoporosis, 25 which is one of the symptoms of NF1 mutation-related disease neurofibromatosis type 1. 26

F I G U R E 5
The NF1 gene inhibits osteoclastic bone resorption. A, The NF1 gene expression in control, Lv-GFP and Lv-NF1 groups. B, TRAP staining to detect osteoclast differentiation in control, Lv-GFP and Lv-NF1 groups. C, Bone resorption in control, Lv-GFP and Lv-NF1 groups. D, ELISA of osteoclastic markers CTX and NTX levels in control, Lv-GFP and Lv-NF1 groups. E, Western blot of osteoclastic markers ITGβ3, CALCR and CTSK in control, Lv-GFP and Lv-NF1 groups. F, Western blot of autophagy-related proteins in CTSK in control, Lv-GFP and Lv-NF1 groups. All data were presented as mean ± SD. *P < .05; **P < .01. n = 6 in each group