All research protocols were approved by the Ethics Committee of Peking University School of Stomatology, Beijing, China (PKUSSIRB-202385020). We obtained informed consent from all donors and their legal guardians prior to collecting samples. All animal experiments conducted in this study followed standard procedures and adhered to ethical principles for animal welfare. The animal experimental protocols were approved by the Institutional Animal Care and Use Committee of Peking University (LA2022288).

After being isolated from extracted teeth, human PDLCs (hPDLCs) were cultured according to the established protocols (Seo et al., 2004; Xie et al., 2022). These cells, obtained from 3 individuals aged 12–25 years, were combined and maintained in α-MEM (Gibco, Grand Island, NY, United States) supplemented with 15% fetal bovine serum (FBS; Gibco), 0.292 mg/mL of glutamine (Gibco), 100 U/mL of penicillin and 100 μg/mL of streptomycin (Solarbio, Tongzhou, Beijing, China) and incubated at 37°C under a humidified atmosphere containing 5% CO₂.

Flow cytometry (Accuri C6, BD, Franklin Lakes, NJ, United States) was used to detect surface markers of cells, including CD31 (Anti-CD31, ab9498, Abcam, Boston, MA, United States), CD34 (Anti-CD34, ab8536, Abcam), CD45 (Anti-CD45, ab40763, Abcam), CD29 (Anti-CD29, ab30394, Abcam), CD90 (Anti-CD90, EPR28145-53, Abcam) and PE-CD105 (anti-CD105, 303205, Biolegend, San Diego, CA, United States) (Chang et al., 2022). FITC-conjugated rabbit IgG (ab6717, Abcam) and PE-conjugated rabbit IgG (ab7003, Abcam) were used as secondary antibodies.

Subsequent experiments utilized cells at passages 3–6.

hPDLCs (1 × 10⁵/mL) were seeded into 6/24-well plates and cultured. Until the cell confluence reached 70%–80%, the osteogenic-inducing medium containing 100 μM ascorbic acid (Sigma-Aldrich, St. Louis, MO, United States), 2 mM β-glycerophosphate (Sigma), and 10 nM dexamethasone (Sigma) (Zhang et al., 2022) were added and the medium was changed every 2 days. Alkaline Phosphatase (ALP) staining (Beyotime, Shanghai, China) was performed at 7 days after osteogenic induction, and Alizarin Red staining (Sigma) was performed at 21 days after osteogenic induction. According to our previous procedure (Zhan et al., 2020), 10% (w/v) cetyl-pyridinium chloride (CPC, HB11089, Macklin, Shanghai, China) was added to dissolve the red matrix sediment to detect the degree of mineralization quantitatively. The absorbance of the solution was measured at 490 nm OD using an ELx808™ plate reader (Bio Tek, Vermont, United States). The samples for quantitative real-time polymerase chain reaction (qPCR) and Western blot analyses were collected at 7 days after osteogenic induction. Each experiment was repeated three times.

Static compressive force was applied to hPDLCs following a previously established method (Jin et al., 2020). In brief, a glass layer with added metal balls was placed on top of a cell layer cultured in 6/24-well plates, which had reached 70%–80% confluence (Supplementary Figure S1A). We exposed hPDLCs to static compressive force at 1 g/cm² for a duration of 24 h (Wang et al., 2018; Jin et al., 2020). Then, we removed the loading device and proceeded with subsequent processing.

The LIPUS exposure device manufactured by the Institute of Acoustics (Chinese Academy of Sciences, China) consisted of an array of 6 big transducers (each with a diameter of 34 mm, designed for use in a 6-well culture plate) and 6 small transducers (each with a diameter of 15 mm, designed for use in a 24-well culture plate and SD rats). The LIPUS conditions used in this study were consistent with those commonly used in previous studies, with a frequency of 1.5 MHz, a pulse duty cycle of 1:4, a pulse repetition frequency of 1.0 kHz, and an average intensity of 30 mW/cm² (El-Bialy et al., 2020).

The in vitro experiments were carried out at 37°C, and LIPUS was applied to hPDLCs after static compressive force was removed for 20 min each day before harvesting. To carry out these experiments, the culture plates were placed on the ultrasound transducer with a thin layer of standard ultrasound gel. The sham controls were handled in the same way as the treated samples, but the ultrasound generator was not turned on (Supplementary Figure S1B–D).

The transfection experiments with small interfering RNA (siRNA) were conducted once the cells achieved 80% confluence, using a riboFECT™ CP Transfection Kit (RiboBio, Guangzhou, China) and following the manufacturer’s instructions. To knockdown Piezo1, hPDLSCs were transfected with siPIEZO1 (siG150327130626-1-5, RiboBio, Guangzhou, China). The transfection reagent mixture was added to the medium for 48 h prior to the compressive force stimulation.

According to the manufacturer’s instructions, Yoda1 (S6678, Selleck, Houston, TX, United States) was dissolved in dimethyl sulfoxide (DMSO) to prepare a stock solution of 50 mM. To activate Piezo1, hPDLCs were treated with the pharmacological agonist Yoda1 at a final concentration of 10 μM (diluted in the culture medium) for 24 h prior to osteogenic induction (Sugimoto et al., 2017). For the control group, an equal volume of DMSO was added.

The RAW264.7 cell lines were purchased from Procell Life Science and Technology Co., Ltd. (Procell, Jiangxia, Wuhan, China). Cells were cultured in DMEM high glucose medium (Gibco) supplemented with 10% FBS (Gibco), and incubated at 37°C in a humidified atmosphere containing 5% CO₂. Subsequent experiments utilized cells within passages 10.

The culture medium of hPDLCs in each group was collected after removing static pressure and applying 20 min of LIPUS stimulation. After centrifugation, the supernatant was retained and stored at −20°C. Raw264.7 cells (5 × 10⁴/mL) were seeded into 6/24-well plates and cultured overnight. And on the second day, the conditioned medium consisting of 50% supernatant, 50% DMEM and 10 ng/mL RANKL (R&D, Emeryville, MN, United States) were used to induce osteoclastic differentiation (Xin et al., 2020). The conditioned medium was changed every 3 days. After 5 days induction, the cells were collected for qPCR and TRAP staining. TRAP staining was performed using an acid phosphatase kit (387A-1KT; Sigma) according to the manufacturer’s instructions. TRAP-positive, multinucleated (≥3 nuclei) cells were counted in each well. Each experiment was repeated three times.

As previously described (Chen et al., 2022a), we extracted RNA from hPDLCs and RAW264.7 cells using TRIzol reagent (Invitrogen, Carlsbad, CA, United States) as per the manufacturer’s guidelines. The purity of the isolated RNA was assessed by measuring the A260/A280 ratio using a spectrophotometer (Nanodrop 8000, Thermo, Waltham, MA, United States), and only samples with ratios between 1.8 and 2.0 were used for downstream experiments. RNA was reverse transcribed into complementary first-strand cDNA using cDNA synthesis kits (Takara Bio, Tokyo, Japan), followed by amplification using qPCR with SYBR Green Master Mix (Roche Applied Science, Basel, Switzerland). The qPCR was performed on the ABI Prism 7500 Real-Time PCR System (Applied Biosystems, Foster City, CA, United States). PCR reaction conditions were as follows: 95°C for 30 s, then 55 cycles of 95°C for 10 s, 60 C for 30 s. The 2^−ΔΔCq method was used for comparative quantitation. The expression levels of target genes were determined relative to glyceraldehyde-3-phosphate dehydrogenase (GAPDH/Gapdh) using primers listed in Supplementary Table S1, which were designed using Primer Premier 5.0 software. Specifically, expression levels of human genes [collagen type I alpha 1 (COL1A1), runt-related transcription factor 2 (RUNX2), and Sp7 Transcription Factor (SP7/OSX)] and mouse genes (acid phosphatase 5, tartrate resistant (Acp5), matrix metalloproteinase 9 (Mmp9), cathepain K (Ctsk) and nuclear factor 1 of activated T cells (Nfatc1)] were assessed. All PCR reactions were performed in triplicate.

We utilized radioimmunoprecipitation assay lysis buffer (Solarbio), which was supplemented with a protease inhibitor mixture (Solarbio), to extract total protein from the cultured cells. Total protein (25 μg) was separated using 10% SDS–polyacrylamide gel, and proteins were transferred to polyvinyli-dene difluoride (PVDF) membranes (Millipore, Massachusetts, United States) (Xu et al., 2022). Then, we employed a solution of 0.1% Tween-20% and 5% non-fat milk to obstruct the membranes for a duration of 1 h. The membranes were incubated overnight at 4°C with the primary antibodies according to previous studies as follows (Ye et al., 2021; Chen et al., 2022b; He et al., 2022; Shen et al., 2023): anti-Piezo1 (1:1,000, 15939-1-AP; Proteintech, Cook, Illinois, United States), anti-COL1A1 (1:1,000, ab260043; Abcam), anti-Runx2 (1:1,000, ab236639; Abcam), anti-Osx (1:1,000, ab209484; Abcam), anti-GAPDH (1:3,000, 60004-1-Ig, ProteinTech), and anti-β actin antibodies (1:3,000, 81115-1-RR, ProteinTech). Then, the membranes were subjected to incubation with a horseradish peroxidase-conjugated mouse or rabbit IgG (1:5,000; Zhongshan Golden Bridge Biotechnology, Beijing, China). Super Signal West Pico Chemiluminescent Substrate (Thermo) and BioMax film (Kodak, Rochester, New York, United States) were used to detect the immunoreactive proteins. We quantified the relative density of three comparable results by utilizing the ImageJ software (National Institutes of Health, Bethesda, MD, United States). The analysis of immunoreactive bands employed GAPDH or β-actin as the control reference.

hPDLCs seeded on chamber slides were subjected to force and/or LIPUS, followed by osteogenic induction for 7 days. The cells from different groups underwent fixation at 4°C for 15 min using 4% polyformaldehyde (PFA, Solarbio), followed by permeabilization for 5 min using 0.1% Triton X-100 (Solarbio). After a PBS wash, the cells were subjected to blocking with 5% bovine serum albumin (BSA, Solarbio) at room temperature for 60 min. Subsequently, the cells were subjected to immunostaining, employing a primary antibody, anti-Piezo1 (1:200, 15939-1-AP; Proteintech), overnight at 4°C. Following this, the cells were incubated with a Cy3 fluorescent secondary antibody (1:200, SA00009-2, Proteintech) at room temperature for 1 h. Subsequently, after PBS washing, the cells were sealed using a mounting medium containing 4′,6-diamidino-2-phenylindole (DAPI, Solarbio). Images were captured using an inverted confocal microscope (FV3000, Olympus, Tokyo, Japan).

Six-week-old male Sprague-Dawley (SD) rats weighing 200 ± 20 g were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). The rats were anesthetized using pentobarbital sodium (40 mg/kg of body weight) for the application of orthodontic devices. A rat OTM model was established as described before (Xin et al., 2023), with a closed-coil spring (Tomy International, Inc. ,Tokyo, Japan) ligated to the right maxillary first molar and incisor neck via a 0.025 mm diameter stainless-steel ligation wire (Tomy International, Inc.). The spring applied a force of 50 g to move the upper first molar mesially, and we ensured that the force applied to each rat was consistent using a force gauge (3M Unitek, United States) (Jin et al., 2020; Jiang et al., 2021). The experimenters involved in constructing the OTM model were blinded to the group allocation, ensuring the precise application of force. To enhance the retaining force and prevent the device from falling off, a 0.5 mm deep groove was made using a slow-speed mill near the gingival margin of the maxillary incisor to accommodate the ligature wire and filled with flowing resin (3M Unitek) (Figure 1B). After all rats regained consciousness, they were fed soft food daily.

One hundred and eight SD rats were involved in this study. 60 SD rats were enrolled in the LIPUS function assay and were divided into 10 groups (Control; LIPUS, Force and Force + LIPUS at day 3, 5, and 7) using random drawing of lots, and each group consisting of 6 rats.

Twenty-four SD rats were enrolled in the Piezo1 functional inhibition assay and were evenly divided into 4 groups using random drawing of lots. Based on the established experimental method (Jiang et al., 2021), the Force + GsMTx4 and Force + GsMTx4 + LIPUS groups received a subcutaneous injection behind the neck of the Piezo1 inhibitor, Grammostola spatulata mechanotoxin 4 (GsMTx4, P1205, Selleck, Houston, TX, United States), at a dosage of 100 μL/kg with a concentration of 10 μM every other day, while the other groups received only the vehicle (PBS).

Twenty-four SD rats were enrolled in the Piezo1 functional activation assay and were evenly divided into 4 groups using random drawing of lots. The animals in the Yoda and Yoda + LIPUS groups received an intraperitoneal injection of Piezo1 agonist Yoda1 (S6678, Selleck, Houston, TX, United States) at a dosage of 100 μL/kg with a concentration of 25 μM (diluted in PBS) every other day according to the instruction of the manufacturer and the effective concentration for in vitro experiments, whereas the other groups received only the vehicle (DMSO + PBS) (Zeng et al., 2022; Zhang et al., 2023a).

LIPUS was applied after mechanical loading had been exerted on the right maxillary first molars for 20 min each day before harvesting. After general anesthesia, the skin overlying the first molar sites was placed on the ultrasound transducer with a thin layer of standard ultrasound gel (Figure 1C). The sham controls underwent the same handling procedures as the treated rats, but the ultrasound generator was not turned on. After the experiment was completed, the experimenters waited for all animals to fully regain consciousness and restore their normal activity before concluding the experiment and returning them to their appropriate housing environments.

After 7 days of treatment, the rats (n ≥ 3) were sacrificed via overdose of pentobarbital sodium, and the maxillae were harvested and fixed in 4% PFA. Subsequently, we acquired micro-CT images of the maxillae at an 8.82 μm resolution, utilizing a tube voltage of 80 kV, a tube current of 500 μA, and an exposure time of 1,500 ms. Three-dimensional (3D) reconstructions were performed using a multimodal 3D visualization software (Inveon Research Workplace; Siemens, Munich, Germany) (Xin et al., 2023). The measurement of tooth movement resulting from mechanical loading adopted a modified technique outlined in an earlier work (Xin et al., 2023). In brief, two proficient investigators, both unaware of the group allocation, measured the distance between two easily identifiable points: the midpoint of the distal marginal ridge of the first molar and the midpoint of the mesial marginal ridge of the second molar. The mean of these dual measurements was used to calculate the distance of tooth movement.

To assess changes in alveolar bone density on the pressure side, a square-shaped region of interest (ROI) measuring 0.3 mm × 1 mm × 1 mm in volume was selected at the mesial aspect of the distobuccal root of the maxillary first molar after three-dimensional reconstruction (Supplementary Figure S3). The bone volume fraction (BV/TV) and trabecular morphology parameters, including trabecular number (Tb.N), trabecular thickness (Tb.Th), and trabecular separation (Tb.Sp), were analyzed within this ROI. Each sample was analyzed three times by two proficient investigators, who were blinded to the group allocation. The mean of these measurements was considered the final measurement for each sample.

After fixation, the maxillae underwent demineralization in 15% ethylenediaminetetraacetic acid and were then embedded in paraffin blocks. Subsequently, 4 μm thick slices perpendicular to the root axis were obtained from their respective experimental groups; the slides were subjected to the staining procedures involving both hematoxylin and eosin (H&E) as well as tartrate-resistant acid phosphatase (TRAP). The observed area was the periodontal ligament (PDL) on the pressure side of the distal buccal root of the maxillary first molar. For TRAP staining, an acid phosphatase kit (387A-1KT; Sigma) was employed, and the guidelines provided by the manufacturer were strictly followed. We conducted quantification of TRAP-positive multi-nucleated cells (≥3 nuclei) along the entire surface of the alveolar bone’s pressure side of each slice.

The slices were subjected to antigen retrieval using 0.125% trypsin (Gibco) and 20 μg/mL of proteinase K solution (Invitrogen, Carlsbad, CA, United States) at 37°C for 1 h. The slides were blocked with 5% BSA for 30 min at room temperature, followed by overnight incubation at 4°C with the primary antibodies, including anti-Piezo1 (1:100, 15939-1-AP; Proteintech), anti-COL1A1 (1:200, ab260043; Abcam), anti-osteocalcin (anti-OCN; 1:100, 23418-1-AP; Proteintech) antibodies. After thorough PBS washing, the slides were exposed to HRP-conjugated secondary antibodies and visualized using diaminobenzidine (Zhongshan Golden Bridge Biotechnology) as per the manufacturer’s instructions. Subsequently, the slides were counterstained with H&E and examined using a light microscope (BX53, Olympus, Tokyo, Japan). The Image-Pro Plus 6.0 Software (Media Cybernetics, Bethesda, MD, United States) was employed for the analysis of positive cells after IHC staining.

The analysis of statistical data was executed using the Statistical Package for the Social Sciences 19.0 software. The data were presented as mean values with 95% confidence interval. The Shapiro-Wilk test was employed to verify the normal distribution of the original dataset, while the evaluation of significance relied on both two-tailed independent Student’s t-test and one-way analysis of variance (ANOVA). To further compare the ANOVA results, Tukey’s multiple comparison test was adopted. Statistical significance was attributed to differences where p-value was less than 0.05.

Following the establishment of the OTM model, the rats in the Force + LIPUS group received daily LIPUS stimulation for 20 min (Figures 1A–C). LIPUS accelerated OTM, evident from the third day of treatment (Figure 1D). The H&E staining results indicated that compared to the control group, the fiber arrangement was disordered in the Force group and the Force + LIPUS group (Supplementary Figure S2). TRAP staining results indicated an increase in the number of TRAP-positive osteoclasts accumulated in the periodontal tissues on the compression side during both mechanical loading and LIPUS application. On the Day 3 and Day 5, the Force + LIPUS group exhibited a higher quantity of osteoclasts on the pressure side compared to the Force group. However, by the seventh day, there was no significant difference observed between the Force and Force + LIPUS groups (Figure 1E).

Given the absence of a substantial variance in the number of osteoclasts differentiated on the pressure side between the Force group and the Force + LIPUS group on the seventh day, we seek to delve deeper into the potential influence of LIPUS on the bone deposition. Micro-CT analysis indicated that LIPUS reduced alveolar bone resorption after force application (Figures 2A, B). LIPUS enhanced the expression of COL1A1 and OCN in the PDL on the pressure side. But there was no significant difference on the number of COL1A1-positive cells and OCN-positive cells between Control group and LIPUS group (Figures 2C–E).

According to the results of the flow cytometry analysis, the surface markers CD31, CD45 and CD34 were negative, while CD29, CD90 and CD105 were positive (Supplementary Figure S4).

To elucidate the impact of LIPUS on the osteogenic differentiation of force-treated hPDLCs, we conducted ALP staining and Alizarin Red staining at 7 and 21 days after osteogenic induction, respectively. Compared to the control group, mechanical force hindered the osteogenic differentiation of hPDLCs. However, the Force + LIPUS group exhibited a higher level of ALP activity (Figures 3A, B) and an increased number of mineralized nodules compared to the Force group (Figures 3C, D). In the Force + LIPUS group, the mRNA levels of osteogenic markers, including SP7, RUNX2 and COL1A1, were significantly higher than those in the Force group (Figure 3E). The Western blot analysis further revealed a significant increase in the protein expression levels of OSX, RUNX2 and COL1A1 in force-treated hPDLCs following LIPUS stimulation (Figures 3F, G). These findings confirmed that LIPUS promoted the osteogenesis of force-treated hPDLCs.

To investigate the influence of LIPUS on in vitro osteoclast differentiation, supernatants from hPDLCs subjected to distinct treatments were collected. Subsequently, conditioned medium was prepared to observe its effects on osteoclast differentiation of RAW264.7 cells. Semi-quantitative analysis indicates that both Force and Force + LIPUS treatments enhance osteoclast differentiation. However, there is no significant difference in the number of osteoclasts between the CM-Force group and CM-Force + LIPUS group (Supplementary Figures S5A, B). The mRNA expression patterns of Acp5, Mmp9, Ctsk and Nfatc1 align with the findings from TRAP staining (Supplementary Figure S5C).

Since Piezo1 can respond to mechanical or ultrasonic stimulation (Zhang et al., 2023b; Du and Yang, 2023; Schröder et al., 2023), we identified changes in Piezo1 expression in hPDLCs after force and LIPUS treatment in vitro and in the pressure side PDL of rats. The Western blot analysis revealed a significant increase in Piezo1 expression in force-treated hPDLCs, but LIPUS downregulated the Piezo1 (Figures 4A, B). The results of the immunofluorescence staining analysis were consistent with the Western blot analysis (Figures 4C, D). The immunohistochemical staining of Piezo1 in the periodontal tissues following force and LIPUS application revealed that LIPUS downregulated the number of Piezo1-positive cells in the force-treated periodontal tissues (Figures 4E, F). By comparing the results of the control group and the LIPUS group, we found that LIPUS stimulation alone did not alter piezo1 expression, which was consistent with the changes in osteogenic phenotype.

After the knockdown of Piezo1 in hPDLCs, osteogenic differentiation was promoted in force-treated hPDLCs. The Force + siPiezo1 group exhibited a higher level of ALP activity (Figures 5A, B) and an increased number of mineralized nodules compared to the Force group (Figures 5C, D). The mRNA levels of the osteogenic markers SP7, RUNX2 and COL1A1 were significantly higher in the Force + siPiezo1 group than in the Force group (Figure 5E). The Western blot analysis also showed that the protein expression levels of OSX, RUNX2 and COL1A1 were significantly increased in the Force + siPiezo1 group after the reduction in Piezo1 in force-treated hPDLCs (Figures 5F, G). No significant difference was observed in the levels of mRNA and protein related to osteogenesis between the Force + siPiezo1 group and the Force + siPiezo1+LIPUS group (Figures 5E–G). These results suggested that the reduction in piezo1 partially attenuated the effects of LIPUS.

To explore whether Piezo1 plays an active role in the process of OTM and the effectiveness of LIPUS, we conducted a functional inactivation trial (Supplementary Figure S7). GsMTx4 has been reported in previous research to be capable of inhibiting mechanical-sensitive cation channels (Bae et al., 2011).

The micro-CT analysis indicated that GsMTx4 effectively inhibited OTM and alleviated bone resorption induced by orthodontic force. And the effect of LIPUS on accelerating OTM was also blocked (Figures 6A–D). Immunohistochemistry staining confirmed that Piezo1 was inhibited by the injection of GsMTx4 (Figures 6E, F). An increased number of COL1A1-and OCN-positive cells was observed in the PDL on the compression side of both the Force + GsMTx4 group and the Force + GsMTx4+LIPUS group compared to the Force group, but there was no significant difference between the Force + GsMTx4 group and the Force + GsMTx4+LIPUS group (Figures 6G–I).

Yoda1 has been reported to be capable of activating purified Piezo1 channel (Botello-Smith et al., 2019). To further explore the role of Piezo1 in the LIPUS application process, hPDLCs were pre-stimulated with Yoda1 before osteogenic induction. The osteogenic differentiation of Yoda1-treated hPDLCs was enhanced following LIPUS treatment, as evidenced by increased ALP activity (Figures 7A, B) and augmented mineralization nodules (Figures 7C, D). The alterations in the mRNA levels of SP7, RUNX2, and COL1A1 in the Yoda group and the Yoda + LIPUS group also proved this (Figure 7E). The protein level of Piezo1 in the Yoda group was upregulated, while it was downregulated in the Yoda + LIPUS group. Simultaneously, the western blot analyses demonstrated a significant decrease in the expression levels of OSX, RUNX2 and COL1A1in the Yoda group, with the LIPUS intervention reversing this effect (Figures 7F, G). These changes closely resembled those observed in hPDLCs after mechanical force treatment and LIPUS stimulation.

Yoda1 was administered to SD rats to investigate the role of Piezo1 in the context of LIPUS (Supplementary Figure S8). Micro-CT revealed that Yoda1 injections for 1 week did not affect alveolar bone mineral density. There was no difference in bone density between the Yoda and Yoda + LIPUS group (Figures 8A, B). The immunohistochemical staining confirmed the activation of Piezo1 following Yoda1 injection. Furthermore, LIPUS stimulation downregulated Piezo1 expression (Figures 8C, D). The Yoda + LIPUS group exhibited a higher number of COL1A1-and OCN-positive cells in the PDL on the “compression side” compared to the Yoda group (Figures 8E–G). These observed changes closely resembled those seen in PDLCs after the force treatment and LIPUS stimulation in vivo.