This study adhered to guidelines by the animal care and use committee of Nanjing Medical University (Approval ID 1811055). Experimental murine orthodontic tooth movement was conducted as described previously (Hirt and Liton, 2017). Before bonding the nickel-titanium coil spring with a flowable repair resin (3 M ESPE) between the right maxillary first molar and the maxillary incisor, anesthesia of mice was done by intraperitoneal administration of chloral hydrate (5%; 0.1 ml/10 g). Force value was approximately 30 g each (Tompkins, 2016). The left upper jaw was unloaded and used as a contralateral control. Sample sizes were at least n = 3 to allow for statistical analysis.

After 7 days of OTM, mouse maxillary bones were collected and fixed in 4% PFA for 12 h. After removing the nickel-titanium coil spring, micro CT scan was performed at a resolution of 15.6 μm, 55 kvp voltage, and 145 μA current. Next, DataViewer was used to adjust the head position so that for all skulls, brain tissue faced down and the occipital bones faced back. CT Vox software was used for image reconstruction. CTAn software was used to make 2 shots in the coronal direction of the sagittal suture on all skulls. Volume measurement based on volume of interest (VOI) was done as described before (Hirt and Liton, 2017). In the sagittal plane, the alveolar bone between the mesial and distal buccal roots of the first molar was selected. Ten consecutive images, starting from disappearance of the distal buccal root were evaluated.

For frozen sections, samples fixation was done in 4% PFA, decalcified with EDTA (10%), embedded in Tissue-Tek, and sectioned at 5 μm. Section staining was done using a TRAP/ALP staining kit (WAKO) following manufacturer instructions. We analyzed alveolar bone around the tensioned side of maxillary first molar mesiobuccal root was microscopically (Olympus Optical, Japan) analyzed. The unit area of positive staining in each section was calculated. Fixing of cultured cells was done in PFA (4%) for 10 min 7 days after osteogenic induction and stained with 1-Step™ NBT/BCIP substrate solution (Thermo Scientific). Images were taken randomly, and ALP positive area of each group calculated on ImageJ.

Immunostaining was performed following a standard protocol. The protocol used to stain sections and cells is practically very similar. Fixation of samples was done in PFA (4%) for 10 min, permeabilized using TritonX-100 (0.5%) for 10 min, blocked with goat serum (BOSTER, AR0009) for 30 min at room temperature, and incubated in the presence of primary antibodies against LC3B at 4°C overnight. After washing thrice with PBS, they were incubated in the presence of a Cy3-labeled secondary antibody (Proteintech, SA00009-2) at 37°C for 1 h. They were then rinsed thrice and counter-stained with DAPI (VECTASHIELD, H-1500) for 30 s. They were then examined by laser confocal microscopy (OLYMPUS, FV3000) and the number of LC3B-positive osteocytes determined.

MLO-Y4, a murine osteocyte cell line was grown on mouse tail collagen type I (BD Biosciences, Bedford, MA) coated dishes and maintained in α-minimum essential medium (α-MEM, Gibco) supplemented with L-glutamine, nucleosides, 5% FBS (Sciencell), calf serum (5%; CS, Gibco), and 1% penicillin/streptomycin (Hyclone). The pre-osteoblast cell line MC3T3-E1 was purchased from the Chinese Academy of Sciences. Pre-osteoblast isolation from skull bones of newborn mice was as follows: eight newborn mice (purchased by the Experimental Animal Center of Nanjing Medical University) were sacrificed and disinfected in 75% alcohol 5–10 min. The top skin was incised to take out the skull, then remove the periosteum and surrounding connective tissue. The cartilage between the sutures was removed as well.Next, soak and rinse the bones several times with sterile PBS until there is no blood on the bones, then move to another sterile plate. Ophthalmic scissors were then used to cut the bone block into 1 mm³ pieces or smaller. 0.25% trypsin+0.02% EDTA compound enzyme digestion solution was then added for digestion for 10 min at 37°C. Sloughed cells were pipetted to enhance digestion, and the digestion solution containing cells carefully transferred into a centrifuge tube. Supernatants were discarded while the pellet was evenly resuspended in complete culture media before seeding in tissue culture dishes. Cells were incubated in a humidified 5% CO₂ atmosphere at 37°C. Experiments were repeated 3 times.

MLO-Y4 osteocyte-like cells were allocated into a control group (no loading to cultured cells/0 h) and four treatment groups and seeded on 6-well flexible bottom plates at 1 × 10⁴ cells/well. They were then cultured overnight to near-confluence and then serum-starved for 8 h before being tensioned. A Flexcell strain unit (Flexcell FX-5000T; Flexcell Corp., Burlington, NC, United States) was used to cyclically stretch the treatment group for 15, 30, 60, and 120 min in the cell culture medium at a frequency of 1Hz. Culture of the control group cells was done on similar plates and kept in the same incubator without cyclic stretching. Three biological replicates were cultivated and measured for each analytical method. After exposure to mechanical tensile stress, The MLO-Y4 cells were seeded in fresh α-MEM medium for 12 h. Conditioned media was then collected from each group and stored in a refrigerator at -80°C for use within 1 week.

MC3T3-E1 cells and murine primary osteoblast precursor cells were harvested using trypsin and ethylenediaminetetraacetic acid. They were then lysed in 1% Triton for 30 min. Then the ALP activity was measured by a colorimetric assay of enzyme activity using a commercially available assay kit (Nanjing Jiancheng Bioengineering Institute, China, A059-2). 5 μL of cell lysate was then mixed with 50 μL of the buffer, followed by incubation in a 96-well plate at 37°C for 15 min without light. 150µL of spectrophotometric developer was then added and ALP-associated absorbance read on a plate reader at 520 nm. ALP activity (U/gprot) was quantified with the equation:

ELISA was used to quantify FGF23 secretion by MLO-Y4 cells. Cells were inoculated on 6-well collagen type I-coated BioFlex culture plates and incubated as described above. Culture media was then collected, centrifuged, and the supernatant subjected to FGF23 quantitative ELISA using elabscience Mouse FGF23(Fibroblast Growth Factor 23) ELISA Kit immediately. Experimental procedures were as described by the manufacturer.

Cells were harvested and total RNA extracted at the indicated timepoints using Takara Minibest Universal RNA extraction kit (Catalog number: 9767), then transformed to cDNA using Takara Primescript RT Master Mix kit (Takara, RR036A). The mRNA levels were then measured on an ABI (QuantStudio 7) RT-PCR system using SYBR-Green (Roche) and following manufacturer instructions. Primer sequences are shown on Table 1.

Cells were harvested, washed, and lysed using whole cell lysis assay kit (KeyGEN BioTECH, China, KGP250). Protein concentration was assessed using a Bradford protein analysis kit (Beyotime, China, P0006). Sixty micrograms of protein were loaded into each well and resolved by 12% SDS-PAGE before transfer onto PVDF membranes (Millipore, Billerica, MA). Membrane blocking was done for 2 h using 5% skim milk followed by overnight incubation in the presence of various primary antibodies at 4°C. Details of the primary antibodies used are listed on Table 2. They were then washed with 0.1% TBST and incubated in the presence of a suitable secondary antibody for 1 h at room temperature. Signal development was done using a Tanon High-sig ECL Western Blotting Substrate (180–501) followed by imaging on a Tanon 5200 chemiluminescence imaging system. GAPDH was the loading control.

Data are shown as either mean ± s.e.m. or mean ± s.d for a minimum of three independent experiments, unless otherwise stated. Comparisons of means between and among groups was done by the student’s t-test or one-way ANOVA, respectively. P=<0.05 was considered significant.

To study if orthodontic tooth movement process involves osteocyte autophagy, a previously reported OTM mouse model was used (Li et al., 2020). Seven days later, the first upper right molar was moved mesially (Figure 1A) and bone volume around root furcation determined by Micro-CT. It was found that Bone Volume (BV)/Total Volume (TV)% as well as Trabecular Thickness (Tb⋅Th) were markedly increased, although Trabecular separation (Tb.Sp) differences were not statistically significant (Figures 1B–D). To assess osteoblasts formation and changes in osteocyte autophagy, maxillary bone adjacent frozen sections were cut. ALP staining showed that the area of the alveolar bone on the tension side of the distobuccal root of the maxillary first molar increased significantly (Figure 1E). Immunofluorescence (IF) showed a high number of LC3B-positive osteocytes, which were near these osteoblasts (Figure 1F), implying enhanced autophagy in bone cells during the OTM process. Therefore, OTM increases autophagic levels in osteocytes.

To determine if mechanical tension induces osteocyte autophagy in vitro, we used a Flexcell^® FX tensile stress loading system to establish a tensile stress loading model using MLO-Y4, a bone cell line and evaluated mRNA as well as protein levels of various autophagy markers. Tension significantly elevated mRNA expression levels of ATG5, ATG4, ATG7, LC3B as well as ULK1 (Figure 2A). Western blot analysis confirmed elevated levels of the autophagy markers ATG7 and LC3BII, while the autophagy substrate, P62 gradually decreased (Figure 2B). P62 is also called SQSTM1. It is expressed in a variety of cells and tissues and it can participate in a variety of signal transduction processes. P62 can connect LC3 and the ubiquitinated substrate, P62 and its bound polyubiquitinated proteins are integrated into autophagosomes and are degraded in autophagic lysosomes, so P62 is the substrate of autophagy.When autophagy is activated, autophagosomes fuse with lysosomes, proteins such as P62 or organelles in autophagic vesicles are degraded by lysosomes, and P62 level decreases. When autophagy is inhibited, autophagosomes accumulate and P62 level increases. Therefore, P62 can be used as an indicator of autophagy ability, and the reduction of P62 expression by Western blotting technique can reflect the degree of autophagy activity. To assess autophagic flux, we futher used IF to visualize autophagosomes (indicated by LC3-positive puncta) and observed significant increase in autophagosomes after tension, beginning after 15 min and peaking at 60 min before gradual decrease, indicating that mechanical tension induces osteocytic autophagy in osteocytic-like MLO-Y4 cells in vitro. In addition, The magnification in the IF image analysis is ×400.

Given that osteocytes influence osteogenesis (Bays et al., 2017), we examined if tension affects the ability of osteocytes to regulate bone remodeling. To this end, MC3T3-E1, an osteoblast precursor cell line was seeded in conditioned medium (CM) derived from MLO-Y4 cells exposed to tension and subjected them to alkaline phosphatase (ALP) staining. ImageJ analysis revealed that relative to the control group, the CM group had stronger ALP levels, indicating that the osteocyte conditioned media from MLO-Y4 cells subjected to tension significantly enhanced early osteogenesis in MC3T3-E1 osteoblasts. Analysis of ALP activity in the 2 groups revealed significantly higher ALP activity in CM group relative to the control group, which is consistent with RT-qPCR results of the expression of the osteoblast markers ALP, OPN and RUNX2. Consistently, western blot analysis revealed significantly elevated levels of the osteogenic related proteins, ALP and RUNX2 (Figure 3E). To verify these findings, we cultivated murine primary osteoblast precursor cells in CM. ALP staining as well as ALP activity assays revealed that osteoblast precursor cells cultured in CM had enhanced osteogenesis ability (Figures 3F–I), which is consistent with RT-qPCR and western blot results (Figure 3J).

To investigate if autophagy is involved in osteocyte-associated osteogenesis, the MLO-Y4 cells were treated with rapamycin to activate autophagy before the application of tension. Western blot analysis revealed that rapamycin enhances mechanically-initiated autophagy (Figure 4A). ALP staining revealed that relative to controls, MC3T3-E1 cells cultivated in CM media from tension-exposed MLO-Y4 (Figures 4B–D), enhances the osteogenic function of osteoblasts. Results from similar experiments using murine osteoblast precursor cells yielded identical results (Figures 4E–G). Indicating that both mechanical and chemical induction of autophagy promotes osteoblasts development.

Autophagy has been implicated in secretion of proteins (Manjithaya et al., 2010; Dupont et al., 2011; Kraya et al., 2015; Wang et al., 2019). FGF23 is recognized as an important regulator of phosphate and calcium homeostasis and is mainly secreted by osteocytes (Feng et al., 2006; Edmonston and Wolf, 2020)Therefore, to determine if autophagy affects osteoblast function via protein secretion, we evaluated FGF23 expression in MLO-Y4 cells. RT-qPCR analysis showed that tensile stress increases FGF23 mRNA levels, which is further enhanced by rapamycin (Figure 5A). Similar observations were made using ELISA (Figure 5B). In addition, Western Blot results showed that the protein expression of FGF23 increased when tension or rapamycin was added (Figure 5C).

AMPK signaling is reported to promote autophagy, which is a key modulator of metabolism and energy homeostasis (Shin et al., 2016). Because mechanical stimulation activates AMPK, which stimulates actomyosin contractility, glucose uptake, and ATP production, and that the increased energy strengthens adhesion complexes and actin cytoskeleton (Zhao et al., 2002), we evaluated if AMPK is involved in tension-mediated autophagy. Mechanical tension suppressed AMPK protein expression as well as phosphorylation time-dependently, but increased CARM1 expression (Figure 6A). Next, we used dorsomorphin (compound C) to inhibit AMPK in MLO-Y4 cells before applying tension and observed reduced AMPK expression and phosphorylation, while ATG7 and LC3BII expression levels were increased (Figure 6B). It should be noted that in the experimental results shown in this part, the tension “+” marked in the second and fourth columns in Figure 6B refers to the results obtained by collecting the cell sample after the tension is loaded for 60 min. Together, these results indicate that mechanical tension increases bone cell autophagy via AMPK signaling involving AMPK and CARM1.