Piezo1 helps prevent bone loss from inflammation through a special calcium-independent process in bone-breaking cells

To generate Piezo1^fl/fl LysM-Cre (Piezo1^ΔLysM) mice, Piezo1^fl/fl mice (Jackson Laboratories, stock no. 029213) mice were crossed with LysM-Cre mice (Jackson Laboratories, stock no. 004781). Piezo1^fl/fl littermates were used as controls. Genotypes were confirmed by PCR from tail DNA. Wild-type C57BL/6J mice (Jackson Laboratories) were used in experiments not requiring genetic manipulation. 6- to 8-week-old males were used. The experimental procedures employed in this study were approved by the NSU IACUC (Protocol #TK7).

Bone marrow-derived mononuclear cells (BMMCs) were collected from tibias and femurs of mice and were cultured with minimum essential medium-α (MEM-α) supplemented with 10% fetal bovine serum (FBS), streptomycin (100 µg/mL), penicillin (100 U/mL), Amphotericin B (0.25 µg/mL), L-glutamine (292 μg/mL) and M-CSF (25 ng/mL; BioLegend, San Diego, CA, USA) at 37 °C in humidified air with 5% CO₂ for 3 days to obtain pre-OCs. Subsequently, pre-OCs were differentiated to OCs using RANKL (10 ng/mL; BioLegend). Human Peripheral Blood Mononuclear Cells (PBMCs) were purchased from STEMCELL Technologies (Vancouver, Canada). PBMCs were stimulated with M-CSF (30 ng/mL; BioLegend) at 37 °C in humidified air with 5% CO₂ for 3 days. Human pre-OCs were differentiated to OCs using RANKL (100 ng/mL; BioLegend). Yoda1 (Tocris), GsMTx4 (Tocris), LY294002 (Cell Signaling Technology), FK506 (Tocris), Ocadaic acid (R&D Systems) and DT-061 (Tocris) were administered with validated concentration. Aa vehicle control, the same concentration of Dimethylsulfoxide (DMSO, Sigma) was used.

Shear stress was generated by a rocker (Thermo Fisher Scientific) or ibidi pump system (ibidi, Fitchburg, WI). Pre-OCs were seeded in wells of a 24-well plate (1x10⁶ cells/well) and then stimulated with shear stress by rocking (15°, 30 rpm). On the other hand, for the ibidi pump system, pre-OCs were cultured in the µ-Slide I 0.4 Luer (3x10⁵ cells/well). Cells were cultured with shear stress at 5 dyn/cm² or 20 dyn/cm². For hydrostatic pressure (HP) loading, cell culture dishes were positioned at the bottom of a beaker (1000 ml or 100 ml), and medium was added to heights of 5.5 cm, generating consistent HP. Control cells were cultured with a 1.2 cm medium height under atmospheric pressure (46).

TRAP staining was performed with a TRAP staining kit (Sigma-Aldrich) according to the manufacturer’s protocol. Briefly, OCs were fixed by a citrate (0.38 mol/L)/acetone solution for 30 seconds at room temperature. Cells were stained with a TRAP staining solution (L(+)-tartrate buffer, 0.67 mol/L; acetate buffer, 2.5 mol/L; Naphthol AS-BI phosphoric acid, 12.5 mg/mL; and Fast Garnet GBC base, 7.0 mg/mL) for 10 min at 37°C in the dark. OCs with ≥ 3 nuclei were determined to be multinucleated OCs.

A plate coated with calcium phosphate was prepared according to previous reports (47, 48). Briefly, 0.12 M Na₂HPO₄ and 0.2 M CaCl₂ (50 mM Tris-HCl, pH 7.4) were mixed at 37°C. The calcium phosphate slurry was washed with sterile water and then applied into wells of a culture plate and dried at 37°C overnight. BMMCs (3.0 × 10⁵ cells/well for a 96-well plate or 1× 10⁶ cells/well for a 24-well plate) were then seeded in wells of calcium phosphate-coated plates, respectively. After treating cells for 6 days, the plates were washed with 10% sodium hypochlorite for 10 min and then dried overnight. The pit areas were microscopically imaged (Evos Cell Imaging System, Thermo Fisher Scientific). Images were analyzed using ImageJ software (version 1.50).

Cells (4×10⁴ cells/well) were seeded in wells of a black wall/clear bottom plate. Fluo-8 No Wash Calcium Assay Kit (AAT Bioquest, Pleasanton, CA) was employed to measure Ca²⁺ influx in accordance with the manufacturer’s protocol. Briefly, Fluo-8 NW and 0.04% Pluronic™ F-127 in HHBS buffer were added for 30 min at 37°C, followed by 30 min at room temperature. After Yoda1 treatment, fluorescence alternately excited at 490 nm and emission at 525 nm was measured every 10 sec using a FilterMax F5 Microplate Reader (Molecular Devices, San Jose, CA).

Ca²⁺ influx was also analyzed under flow conditions with the BioFlux One system (Fluxion Biosciences, Oakland, Ca) or ibidi pump system (ibidi). For the BioFlux One system, a 48-well microfluidic plate (Fluxion Biosciences) was first coated for 1 h at room temperature with rat tail type 1 collagen (50 μg/ml; Thermo Fisher Scientific) in 0.2% acetic acid. Before using the plate, microfluidic channels were washed with PBS, followed by the introduction of cells in the channels. After 24 h, the Fluo-8 No Wash Calcium Assay Kit was employed to image Ca²⁺ influx. The assay was performed with a wall shear stress at 20 dyn/cm². Fluorescence intensity was measured and analyzed by the BioFlux One system (Fluxion Biosciences). For ibidi pump system, pre-OCs were seeded in µ-Slide I 0.4 Luer (ibidi) at 3x10⁵ cells/plate. The assay was performed with a wall shear stress at 20 dyn/cm². Fluorescence intensity was measured and analyzed by the EVOS (Thermo Fisher Scientific).

BMMCs were seeded in wells of a 96-well plate (3×10⁵ cells/well) or 24-well plate (1×10⁶ cells/well). After 24 h, BMMCs were maintained in Opti-MEM™ Reduced Serum Medium (Thermo Fisher Scientific) and transfected with 50 nM Piezo1-specific siRNA (107969, Thermo Fisher Scientific; siPiezo1), PP2A-specific siRNA (152168, Thermo Fisher Scientific; siPP2A), PP2B-specific siRNA (162266, Thermo Fisher Scientific; siPP2B) or negative control siRNA (AM4611, Thermo Fisher Scientific; siCTL) using Lipofectamine™ RNAiMAX Transfection Reagent (Thermo Fisher Scientific) according to the manufacturer’s instructions. 48 h post- transfection, BMMCs were used for indicated experiments.

Total RNA was extracted from pre-OCs using the PureLink RNA Mini Kit (Thermo Fisher Scientific), following the manufacturer’s protocol. The first strand cDNA was assembled from 100 ng of sample RNA using a Verso cDNA Synthesis Kit (Thermo Fisher Scientific). Amplification reactions were performed by Taqman Fast Advanced Master Mix (Thermo Fisher Scientific) or SYBR Green Master Mix (Thermo Fisher Scientific). The resultant cDNA was amplified by specific probes (Thermo Fisher Scientific) for Gapdh (Mm99999915_g1), Piezo1 (Mm01241549_m1), Piezo2 (Mm01265861_m1), Trpa1 (Mm01227437_m1), Trpv4 (Mm00499025_m1), Stoml3 (Mm01289590_m1), Kcnk10 (Hs01026663_m1), Kcnk4 (Mm00434626_m1), Kcnk1 (Mm00434624_m1), Ocstamp (Mm00512445_m1), Mmp9 (Mm00442991_m1), Ctsk (Mm00484039_m1), Acp5 (Mm00475698_m1), Oscar (Mm01338227_g1), Nfatc1 (Mm00438670_m1), Tnfsf11 (Mm00441906_m1) and Tnfsf11b (Mm00435454_m1) on a QuantStudio™ 3 (Thermo Fisher Scientific). The ratios of mRNA levels to those of the control gene were calculated using the ΔCt method (2^−ΔΔCt).

After extraction of total RNA from cells and synthesis of cDNA described above, the resultant cDNA was tested using Taqman Fast Advanced Master Mix (Thermo Fisher Scientific) and TaqMan^® Array Mouse Osteogenesis (Thermo Fisher Scientific) according to the manufacturer’s instruction. An integrated web-based software package was used for data analysis (https://www.thermofisher.com/account-center/simplified-username.html↗).

The Phospho Explorer Antibody Array was used according to the manufacturer’s instruction (Full Moon BioSystems, Sunnyvale, CA) to profile the levels of phosphorylated proteins. Briefly, cell lysate from BMMCs was collected and quantified by BCA Protein Assay Kit (Thermo Fisher Scientific). Microarray slides were blocked, and proteins were labeled using biotin and then coupled to slides. Slides were washed, and Cy3-streptavidin was added to bind biotin. Fluorescence intensity was measured and analyzed by the manufacturer (Full Moon BioSystems). The clustering of target proteins and signaling pathways was assessed using Kyoto Encyclopedia of Genes and Genomes (KEGG) and Ingenuity Pathway Analysis (IPA) (Qiagen, Germantown, MD).

After incubation for various times, pre-OCs were lysed by incubation on ice for 30 min with RIPA buffer (Thermo Fisher Scientific) supplemented with a protease inhibitor cocktail (Sigma-Aldrich). Protein concentration of the resultant lysates was measured with the BCA Protein Assay Kit (Thermo Fisher Scientific). Fifteen μg of sample per lane were loaded onto a 4–12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) gel (Thermo Fisher Scientific). Proteins separated in the SDS-PAGE gel were electrotransferred to a polyvinylidene difluoride membrane. The detection of specific proteins in pre-OCs was assessed using anti-phospho-Akt rabbit mAb (193H12, 1:1000; Cell Signaling Technology, Danvers, MA), anti-Akt rabbit mAb (C67E7, 1:1000; Cell Signaling Technology), anti-phospho-p38 MAPKs rabbit mAb (3D7, 1:1000; Cell Signaling Technology), anti-p38 MAPKs rabbit mAb (D13E1, 1:1000; Cell Signaling Technology), anti-phospho-ERK rabbit mAb (D13.14.4E, 1:2000; Cell Signaling Technology), anti-ERK rabbit mAb (137F5, 1:2000; Cell Signaling Technology), anti-phospho-JNK rabbit mAb (81E11​, 1:1000; Cell Signaling Technology), anti-JNK rabbit mAb (9258, 1:1000; Cell Signaling Technology), anti-IkBa rabbit mAb (L35A5, 1:1000; Cell Signaling Technology), anti-phospho-PP2A mouse mAb (F-8, 1:1000; Santa Cruz Biotechnology, Dallas, TX), anti-PP2A mouse mAb (6F9, 1:1000; Santa Cruz Biotechnology) or an anti-GAPDH rabbit mAb (14C10, Cell Signaling Technology). Protein bands that reacted with the respective antibody were visualized by incubation with an HRP-conjugated rabbit or mouse secondary antibody (Cell Signaling Technology), followed by detection using ECL Western Blotting Substrate (Thermo Fisher Scientific). Densitometric analysis was performed using ImageJ software (Version 1.50).

To examine the interaction between RANK and TRAF6 or Piezo1 and PP2A, Co-IP was performed. Pre-OCs were lysed in ice-cold RIPA buffer (Thermo Fisher Scientific) supplemented with a protease inhibitor cocktail (Sigma-Aldrich). The lysates were cleared by centrifugation and incubated overnight at 4°C with an anti-RANK antibody (sc-374360, 10 ug/ml/sample, Santa Cruz Biotechnology). Or anti-Piezo1 antibody (2-10, 10 ug/ml/sample, Thermo Fisher Scientific). Immune complexes were captured using Protein A/G magnetic beads (Thermo Fisher Scientific) for 1 h at RT with gentle rotation. Beads were washed three times with Tris-Buffered Saline with Tween 20 (TBST), and bound proteins were eluted in LDS sample buffer, followed by immunoblotting with anti-TRAF6 antibody (E2K9D, 1:1000; Cell Signaling Technology) or anti-PP2A antibody (F-8, 1:1000; Santa Cruz Biotechnology).

Pre-OCs cultured on Millicell EZ SLIDE (Sigma) were fixed with 4% paraformaldehyde at room temperature for 10 min, permeabilized with 2% Triton X-100, and blocked with 5% BSA in PBS for 1 h. Cells were incubated with anti-Piezo1 antibody conjugated with Alexa Fluor^® 594 (CL9714​, 1 ug/ml, Novus, Centennial, CO), isotype control antibody conjugated with Alexa 594 (141945, 1 ug/ml, Novus), anti-NFATc1 antibody (7A6, 1 ug/ml, Santa Cruz Biotechnology) or isotype control antibody (MOPC-21​, 1 ug/ml, Bio X Cell) at 4°C overnight. Cy3-conjugated goat anti-mouse IgG secondary antibody (1:100, Jackson ImmunoResearch, West Grove, PA) was applied for 1 hour at room temperature, followed with CellMask™ Green Actin Tracking Stain (Thermo Fisher Scientific) and Fluoromount-G containing DAPI (Thermo Fisher Scientific). Immunofluorescence was imaged with a Zeiss LSM880 confocal microscope (Carl Zeiss, Jena, Germany).

To induce periodontal disease in mice (6–8 weeks old; male or female), the maxillary second molar was attached with a 5–0 silk ligature, following the previously published protocol (49, 50). Yoda1 (0.4 mg/kg), DT-061 (0.4 mg/kg) or DMSO (0.86%) was diluted in PBS with 5% ethanol, followed by injection through the intraperitoneal route every 2 days (day 0, 2, 4 and 6). After 7 days, mice were euthanized for postmortem analyses.

Real-time BPU was monitored by Laser Doppler Flowmetry (OxyFlo pro, Oxford Optronix, UK). A fine needle-type blood flow probe (Diameter: 0.5 mm, Oxford Optronix) was attached to the mesial or distal surface of palatal gingiva, respectively (Figure 1). Real-time data were captured and analyzed by Labchart 8 software (ADInstruments, Colorado Springs, CO).

Murine maxillary bones were fixed in 4% paraformaldehyde overnight at 4°C before decalcification in 10% EDTA at 4°C for 2 weeks. Tissues were embedded in an OCT compound (Sakura Finetek USA, Torrance, CA, USA) overnight at −20°C and cut into 8 μm sections with a cryostat (Leica Biosystems, Deer Park, IL). TRAP staining of decalcified periodontal tissue was performed using an Acid Phosphatase Leukocyte (TRAP) Kit (Sigma-Aldrich), as described above, followed by nuclear counterstaining with methyl green. Sections were imaged with an EVOS XL Core microscope (Thermo Fisher Scientific). For immunofluorescence-based detection of OC-STAMP and phospho-Akt, the sections were reacted with anti-OC-STAMP rabbit pAb (HPA031116, 1:200; Sigma-Aldrich) or anti-phospho-Akt rabbit mAb (D9E, 1:200; Cell Signaling Technology) as the primary antibody at 4°C overnight. Cy3-conjugated anti-rabbit IgG FC goat pAb (1:200; Jackson ImmunoResearch) was used as a secondary antibody. The stained sections were mounted with Fluoromount-G containing DAPI (Thermo Fisher Scientific). Immunofluorescence was observed with a Zeiss LSM880 confocal microscope (Carl Zeiss, Jena, Germany).

Mouse maxillary alveolar bone was fixed in 4% phosphate-buffered paraformaldehyde and stored at 4˚C for 16 hours. Micro-CT images were obtained with the Microfocus X-ray CT scanning system (Skyscan 1176, Bruker, Billerica, MA), using the following settings: acceleration voltage, 50 kV; current, 500 µA; voxel size, 18 µm/pixel; matrix size, 2,000 × 1,336. Images were reconstructed with NRecon software, version 1.7.0.3 (Bruker), and images of both ligature side and control untreated side were acquired. As regions of interest (ROI), 50 sliced images coronally from the contact point between the maxillary first molar and maxillary second molar were evaluated. Bone volume (BV) of the whole palatal alveolar bone, including the ipsilateral hard palate, was measured using three-dimensional (3D) analysis CTAn software, version v.1.18 (Bruker). 3-dimensional images were obtained using CTVox software, version 3.2.0 (Bruker). To evaluate periodontal bone resorption, distances from the cement–enamel junction to the alveolar bone crest on the palatal side of root were measured for the maxillary second molar.

Statistical analyses were performed by one-way ANOVA and Tukey’s Honestly Significant Difference (HSD) test to compare differences among multiple groups and Student’s t-test for comparisons between two groups. All statistical analyses were performed using GraphPad Prism, version 10.0.1 (GraphPad Software, Inc., La Jolla, California, USA). Statistical significance was considered to be at p < 0.05. All data were expressed as the mean ± SD.

Although, as noted above, the blood flow rate in the vasculature of human periodontitis lesions is significantly diminished (30); however, it is unknown whether blood flow in the periodontitis induced in mice is also reduced. Therefore, to induce murine periodontitis, a silk ligature was attached to the second maxillary molar for 7 days. To assess the impact of periodontitis on the local blood flow, the blood perfusion unit (BPU) of gingival tissue was measured using Laser Doppler Flowmetry. Irrespective of inflammation induced in the periodontal tissue, the BPU in periodontally diseased tissue was significantly reduced compared to that in healthy tissue (Figures 1A, B). Since interstitial pressure is generally considered to be proportional to the local blood flow rate (51, 52), it is assumed that the mechanical stress to the cells in the laminar propria of periodontal tissue is also reduced.

According to a previous report, the blood flow rate in the microcapillaries of healthy periodontal tissue is approximately 20 dyn/cm², whereas that in diseased periodontal tissue is diminished to approximately 5 dyn/cm^{2 30}. To compare the effect of flow rates, a microfluidics system was employed to evaluate the effects of fluid shear stress on RANKL-induced OC-genesis. OC-genesis, as well as OC-related gene expression (Ocstamp, Mmp9, Acp5, Oscar, Dcstamp and Nfatc1), were significantly suppressed by a high flow rate (20 dyn/cm²) compared to a low flow rate (5 dyn/cm²) or static condition (Figures 1C, D; Supplementary Figure S1a). However, Piezo1 expression did not significantly differ between high- or low-flow conditions and static control (Supplementary Figure S1a). Moreover, shear flow generated in the tissue culture plate by a rocker (15°, 30 rpm) also suppressed RANKL-induced OC-genesis (Supplementary Figures S1b, c). Hydrostatic pressure (HP) via the tissue interstitial fluid has an important role in providing mechano-stimulation to cells (53). To examine the effect of HP on OC differentiation, the OCs were cultured in two different culture flasks following the protocol published by another group (46, 54). A 5.5 cm deep beaker with 100 ml of medium gives approximately HP of 3.7 mmHg, compared to a 1.2 cm deep flask with 100 ml of medium (0 mmHg). Such a difference of HP promoted Piezo1 stimulation, which, in turn, significantly suppressed OC differentiation (Supplementary Figure S1d).

Out of 8 major mechanoreceptors (Piezo1, Piezo2, Trpv1, Trpv4, Stoml3, Kcnk10, Kcnk4 and Kcnk1), Piezo1 mRNA was expressed at the highest level (Figure 1E). The protein expression of Piezo1 in pre-OCs was confirmed by both immunofluorescence staining and flow cytometry (58.5%) (Figures 1F, G). GsMTx4, a spider venom that selectively inhibits Piezo1 (55–57), suppressed Ca²⁺ influx induced in pre-OCs via microfluidics-generated shear flow (Figure 1H), indicating that pre-OCs appeared to sense shear flow-generated mechanical force via Piezo1.

RANK-positive mononuclear pre-OCs, which are derived from monocyte lineage cells, circulate in the vasculature and migrate to bone (58). Fluid shear stress influences the local migration of circulating immune cells, such as CD4 T cells, neutrophils, and monocytes (59–61). As demonstrated by the above-noted result (Figure 1E) and a previous report (62), Piezo1, but little, or no, Piezo2, is expressed by pre-OCs. The functionality of Piezo1 in pre-OCs was examined by siRNA-based loss-of-function assay. The silencing of Piezo1 mRNA by Piezo1-specific siRNA (siPiezo1) in pre-OCs (Figure 1I) downmodulated Ca²⁺ influx induced by Yoda1, a chemical agonist of Piezo1, which was not observed in the pre-OCs treated with siControl (Figure 2J). The shear flow (20 dyn/cm²) created in a microfluidics system caused an influx of Ca²⁺ in pre-OCs. Such shear flow-induced Ca²⁺ influx was, however, abrogated by treating pre-OCs with siPiezo1 (Figure 2K). Also, mechanical stress generated by the rocker suppressed OC-genesis-related genes, including the expression of OCSTAMP, MMP9 and ACP5, while mature TRAP+ OC formation and pit formation were downregulated by treating pre-OCs with siPiezo1 (Figure 2L; Supplementary Figure S1g). These data indicated that Piezo1 expressed in pre-OCs acts as a mechanosensor that downregulates RANKL-induced OC-genesis.

Yoda1 is a chemical agonist that can selectively open Piezo1 and promote intercellular Ca²⁺ to initiate a variety of biological events (63–66). We therefore used Yoda1 to determine if RANKL-induced OC-genesis was regulated by Piezo1. We found that TRAP-positive multinucleated OC formation, as well as bone resorptive activity, were both significantly diminished by Yoda1 administration (Figure 2A). In addition, GsMTx4 inhibited Yoda1-induced Ca²⁺ influx in pre-OCs (Supplementary Figure S1e). To test the effect of Yoda1 on human OC-genesis, peripheral blood mononuclear cells (PBMC)-derived pre-OCs were employed. Yoda1-mediated Piezo1 activation also inhibited RANKL-mediated human OC-genesis (Supplementary Figure S2a). PCR array was used to screen for RANKL-stimulated genes that were impaired by Yoda (Figure 2B). Yoda1-mediated suppression of OC-genesis-associated genes (Ocstamp, Ctsk, Mmp9, Acp5 and Oscar) was also confirmed by qPCR (Supplementary Figure S2b). Upon stimulation with RANKL, NFATc1, a master TF of OC-genesis (10, 67, 68), translocates from cytoplasm to nucleus, and induces the transcription of genes required for OC-genesis and fusion (69, 70). Yoda1 inhibited the expression of Nfatc1 gene and its protein during RANKL-induced OC-genesis (Figures 2C–E and Supplementary Figure S5a). Furthermore, Yoda1 inhibited NFATc1 nuclear localization and disrupted the redistribution of β-actin from the inner to the outer regions, suggesting that Yoda1 impairs osteoclast mobility, preventing fusion with adjacent OCs (Figure 2E). These results suggested that the pharmacological activation of Piezo1 caused the downregulation of RANKL-induced OC-genesis in conjunction with the suppression of both Nfatc1 expression and NFATc1 nuclear translocation.

Piezo1 is reported to provoke intracellular signaling activation in numerous cells (37, 71, 72). However, Piezo1-related signaling in OCs is still unknown. To address this question, a phospho antibody array was performed to discover the specific signaling of Piezo1 in pre-OCs in vitro. A search of the Kyoto Encyclopedia of Genes and Genomes (KEGG) and Ingenuity^® Pathway Analysis (IPA^®) found PI3K/Akt signaling to be the most likely Piezo1 signaling pathway (Figure 3A). Akt is a serine/threonine kinase that plays a critical role in cell survival, growth, and metabolism. It is well known that Akt is closely involved in OC-genesis (73, 74). Akt phosphorylation activates GSK3β, then evokes NFATc1 nuclear translocation during OC formation (75–77). As further confirmation, the inhibition of PI3K/Akt signaling by LY294002, a morpholine-containing chemical compound, resulted in suppressed OC formation and Nfatc1 expression, suggesting that PI3K/Akt signaling is, indeed, involved in OC-genesis (Supplementary Figure S1f). In addition, Yoda1 application significantly reduced Akt phosphorylation and moderately inhibited ERK phosphorylation (Figure 3B). The inhibitory effect of Yoda1 on Akt phosphorylation in OCs was observed in a time-dependent manner, but it was diminished in Piezo1-deficient OCs (Figure 3C; Supplementary Figure S5b)4f Moreover, shear flow-mediated Akt dephosphorylation was regulated via Piezo1 (Figure 3D; Supplementary Figure S5b). Therefore, Akt dephosphorylation induced by Piezo1 activation is associated with OC formation. RANKL/RANK signaling is known to involve the TRAF6/PI3K/Akt pathway in OCs (78). To further clarify the interaction between Piezo1 activation and RANK/TRAF6/PI3K signaling, we examined the effects of Yoda1 on RANK/TRAF6 binding and PI3K phosphorylation. Piezo1 activation by Yoda1 suppressed RANK/TRAF6 binding, which was accompanied by a reduction in PI3K phosphorylation (Figure 3B, E). However, unlike the strong suppression observed in Akt phosphorylation (Figure 3B), PI3K phosphorylation was only partially inhibited by Yoda1. These findings indicate that Piezo1 activation partially regulates OC-genesis by interfering with RANK/TRAF6/PI3K signaling.

In the absence of extracellular Ca²⁺, Yoda1 was not able to induce Ca²⁺ influx in pre-OCs, whereas Yoda1 induced Ca²⁺ influx in pre-OCs suspended with Ca²⁺ at 2 mM (normal Ca²⁺ concentration in medium) and 40 mM (high Ca²⁺ concentration representing on the bone surface) (Figure 4A; Supplementary Figure S5c). BAPTA-AM is mostly used for cell-permeable intercellular Ca²⁺ chelating agent (79, 80). Whereas BAPTA-AM could block Yoda1-induced intercellular Ca²⁺ influx, Yoda1 could induce Akt-dephosphorylation in the presence of BAPTA-AM, suggesting that a Ca²⁺-independent pathway mediates Yoda1-induced Akt dephosphorylation in OCs (Figure 4D; Supplementary Figure S5d). Akt is regulated by the protein phosphatase family, such as protein phosphatase 2A (PP2A) or protein phosphatase 2B (PP2B), known as calcineurin (81). We demonstrated that okadaic acid, a PP2A inhibitor, but not FK506, a calcineurin inhibitor, could counteract Yoda1-induced dephosphorylation of Akt (Figures 4C, F; Supplementary Figures S5e, d). Subsequently, as a gain-of-function approach, DT-061, a PP2A activator (82), was employed to elucidate PP2A’s functional role in OC-genesis. We demonstrated that DT-061, through its activation of PP2A, significantly suppressed TRAP+ OC formation, resorption pit formation, and Akt phosphorylation in RANKL-stimulated pre-OCs (Figure 5A). RANKL-induced NFATc1 expression was also downregulated by treatment with the PP2A activator DT-061 (Figure 5C). We found that Yoda 1 downregulated the induction of PP2A phosphorylation at Tyr³⁰⁷ in RANKL-stimulated pre-OCs (Figure 4D; Supplementary Figure S5g). It is noteworthy that the PP2A catalytic subunit (PP2Ac) is inactivated by single phosphorylation at Tyr³⁰⁷ residue (83), whereas phosphorylation of Tyr¹²⁷ and Tyr²⁸⁴ can activate PP2Ac (84). These findings indicate that the activation of Piezo1 can suppress phosphorylation of PP2A at Tyr³⁰⁷ to increase phosphatase activity by PP2A which, in turn, suppresses the phosphorylation of Akt, i.e., Akt dephosphorylation, as well as NFATc1. RNAi-based silencing of PP2A mRNA expression (siPP2A), but not calcineurin (siCalcineurin), resulted in increasing Akt phosphorylation in Yoda1-treated OCs (Figure 4E and Supplementary Figure S5h). Furthermore, treatment with siPP2A, but not siCalcineurin, prevented shear stress-dependent suppression of RANKL-stimulated OC-genesis, otherwise activated by PP2A, including TRAP-positive OC formation, pit formation, and OC-genesis-related gene expression (Figures 4G, H). Collectively, these results suggested that Piezo1-mediated mechanosensing by pre-OCs suppresses RANKL-induced OC-genesis through cell signaling that involves the PP2A/Akt-dephosphorylation pathway toward the suppression of NFATc1, the master TF controlling OC-genesis. To confirm the correlation between Piezo1 and PP2A, a Co-IP assay was performed using an anti-Piezo1 antibody. The results demonstrated that PP2A directly binds to Piezo1, indicating that Piezo1 activation in osteoclasts may directly regulate PP2A-mediated Akt dephosphorylation (Figure 4F; Supplementary Figure S5i).

Given the decreased shear stress and increased osteoclastic bone resorption observed in periodontitis, we examined the effect of systemic (i.p.) injection of Yoda1 in the mouse ligature-induced periodontitis model. Murine periodontitis was induced by the attachment of a silk ligature at the upper second molar, following previous reports (49, 50, 85). Systemically administered Yoda1 significantly suppressed bone resorption and ligature-induced TRAP-positive OC formation in alveolar bone compared to vehicle control (Figure 6A). It also inhibited the mRNA expression of OCSTAMP, ACP5 and MMP9, but not RANKL mRNA (Tnfsf11) or osteoprotegerin mRNA (Tnfsf11b) (Figure 6B). Furthermore, the number of phosphorylated Akt-positive OCs increased in mouse alveolar bone (Figure 6C).

These results revealed that Piezo1 activation by Yoda1 directly reversed bone resorption in periodontitis by restoring the mechanical stress-signaling in pre-OCs which was attenuated in an inflammation-dependent fashion.