A motion-activated skeletal “Pacemaker” for systemic bone remodeling via oral piezoelectric microspheres Authors: Wang Han, Jing Luo, Yingkun Xu, Xu He, Xinhe Li, Jindong Tan, Zijie Wang, Yi He, Peng Zou, Yi Peng, Xiaoyu Han, Dingqun Bai, Jieliang Shen Published: 2026 ABSTRACT AbstractThe loss of endogenous bioelectrical rhythms in bone is a critical driver of osteoporosis initiation and progression. Reconstructing bioelectrical rhythms synchronized with mechanical stimulation is critical for restoring bone homeostasis. Here, we propose a motion-activated skeletal “bioelectrical pacemaker” strategy that microfluidically encapsulates alendronate-modified BaTiO3 piezoelectric nanoparticles (BTO@ALN) within chitosan/alginate core–shell hydrogel microspheres, enabling gastric protection, pH-responsive intestinal release, and noninvasive oral delivery with systemic skeletal enrichment. Upon skeletal enrichment, the piezoelectric units compensate for impaired electrophysiological responsiveness in osteoporotic bone and are activated by physiological mechanical stimuli during daily motion to generate movement-synchronized dynamic bioelectrical signals. This process reconstructs endogenous bone bioelectrical rhythms systemically, thereby restoring bone homeostasis. Experimentally, BTO@ALN exhibited 6.8-fold higher skeletal accumulation than controls; oral administration led to a 52% increase in bone piezoelectric performance within 24 h, accompanied by concurrent suppression of osteoclast activity and promotion of osteogenesis. When combined with low-intensity exercise, motion-triggered piezoelectric output further amplified therapeutic efficacy, resulting in a 2.4-fold increase in osteogenesis and a 1.5-fold stronger suppression of bone resorption. Collectively, this closed-loop strategy integrating oral delivery, bone-targeted enrichment, and motion-synchronized electrical activation systematically re-establishes endogenous bone bioelectrical rhythms, introducing a strategy for osteoporosis based on electrophysiological regulation of bone.Keywords: Oral hydrogel microspheres, Bone regeneration, Piezoelectric nanoparticles, Osteoporosis, Bone targeting Introduction Osteoporosis is a systemic skeletal disorder characterized by progressive bone loss and deterioration of bone microarchitecture, with a pathological hallmark of persistent disruption of bone homeostasis [1]. Current clinical therapies primarily modulate bone metabolism through pharmacological suppression of bone resorption or stimulation of bone formation. Although these strategies can improve parameters such as bone mineral density in the short term, their long-term efficacy is often limited by single-target intervention, safety concerns, and poor patient adherence[a), b), c), d)[2]]. More importantly, bone homeostasis is not governed merely by the unidirectional “enhancement” or “suppression” of osteoclast and osteoblast activity, but also depends on the continuous sensing of and responsiveness to the physiological mechanical environment by bone tissue [3]. Bone is not a passive mechanical scaffold but a highly dynamic, metabolically active, and electroactive organ[[4], a), b)]. During daily movement, cyclic mechanical loading induces time-varying electrical signals within bone through matrix piezoelectricity, streaming potentials, and related electromechanical coupling processes. These signals fluctuate with loading frequency and temporal pattern while remaining synchronized with mechanical input [5]. Here, we define this physiologically driven, motion-synchronized pattern of intraosseous electrical signals as the endogenous bone bioelectrical rhythm. In this context, piezoelectricity refers to the physical conversion of mechanical deformation into electrical potential or charge, whereas motion-synchronized bioelectrical signals describe the tissue-level dynamic electrical responses coupled to cyclic loading[[6], a), b)]. This rhythm depends on the integrity of the collagen–mineral architecture, trabecular continuity, mineralized matrix orientation, and loading state of bone tissue[[7], a), b), c)]. During osteoporotic progression, bone mass loss, trabecular disruption, cortical degeneration, and deterioration of collagen–mineral organization may weaken the ability of bone to convert cyclic mechanical loading into synchronized electrical cues, manifesting as reduced motion-coupled skeletal electromechanical responsiveness a), [8]. Thus, osteoporosis can be viewed not only as a cellular imbalance between bone resorption and formation, but also as an attenuation of bone bioelectrical rhythm in an electromechanically active organ. Restoring this motion-coupled electromechanical response may therefore offer a biophysical therapeutic strategy distinct from conventional pharmacological regulation. Electrostimulation has long been recognized as a promising modality for modulating bone remodeling [9]. Both in vitro and in vivo evidence indicates that microcurrent stimulation can upregulate osteogenic markers such as runt-related transcription factor 2 (RUNX2), alkaline phosphatase (ALP), and osteocalcin (OCN) [10]. However, conventional electrical-stimulation technologies typically rely on surface electrodes or implantable devices to deliver exogenous cues and thus suffer from limited penetration depth and insufficient spatial coverage; they may also incur surgical trauma, face power-supply constraints, and make it difficult to achieve synchronized stimulation across the entire skeleton [11]. More importantly, these exogenous electrical signals are typically governed by preset device-defined parameters, and their output timing is not inherently synchronized with the cyclic mechanical inputs experienced by bone during daily movement. As a result, they fail to reproduce the native load-responsive electromechanical signaling mode of bone [12]. Against this backdrop, piezoelectric biomaterials provide an alternative route toward bone electroregulation that better aligns with physiological principles. Representative piezoelectric materials such as barium titanate (BaTiO3) can generate electrical signals upon mechanical deformation, converting motion-related mechanical inputs into electrical outputs without requiring an external power source [13]. For example, Wu et al. incorporated polydopamine-modified BaTiO3 nanoparticles into a chitosan/gelatin porous hydrogel scaffold and leveraged walking- or micromotion-induced mechanical stimuli to activate the piezoelectric effect, thereby promoting osteogenic differentiation, enhancing mineral deposition, and improving the quality of newly formed bone, suggesting potential utility for bone-defect repair [14]. The key advantage of such materials lies in the intrinsic synchronization between electrical-signal generation and mechanical loading, which more closely recapitulates the physiological paradigm of “motion-induced electrification” in bone. Under pathological conditions such as osteoporosis, if piezoelectric units can effectively enrich within bone tissue and generate local electrical outputs in response to daily movement, they may compensate for the insufficient motion-synchronized electrical signaling caused by attenuated matrix electromechanical responsiveness[a), b), [15]]. In this context, piezoelectric materials can be endowed with a skeletal “pacemaker” like function: rather than autonomously generating electrical rhythms, they convert mechanical inputs into motion-synchronized electrical outputs when loading occurs, thereby enhancing the endogenous bioelectrical rhythm of bone and recoupling mechanical activity with bone remodeling. Although piezoelectric materials have been extensively investigated for bone repair, most existing strategies rely on local surgical implantation, which is inherently unsuitable for systemic skeletal disorders such as osteoporosis that involve the entire skeleton. Moreover, the invasive nature of these approaches substantially limits their applicability for long-term, whole-body therapy[d), [16], a), b), c)]. Bone-targeted nanomedicine offers a potential solution to achieve systemic skeletal enrichment. Bisphosphonate molecules, such as alendronate sodium (ALN), exhibit a high affinity for hydroxyapatite in bone tissue. ALN functionalization of nanoparticles, liposomes, or polymeric carriers markedly enhances their selective deposition and prolonged retention in bone while preserving ALN's intrinsic anti-resorptive activity [17]. For instance, Sun et al. designed an amphiphilic molecule (ALN–Pt(IV)–Lipo) by covalently conjugating ALN to a Pt(IV) prodrug and a lipid tail; the resulting conjugate self-assembled into nanoparticles capable of targeting bone tissue following intravenous administration for osteosarcoma chemotherapy and suppression of bone destruction [18]. However, intravenous delivery is associated with poor patient compliance, particularly for chronic disease management, such as osteoporosis [19]. In contrast, oral administration is noninvasive and more convenient; however, oral delivery of piezoelectric nanoparticles is challenged by degradation in gastric acid and digestive enzymes, limited intestinal absorption efficiency, and potential gastrointestinal irritation[b), c), [20], a)]. Recently, microfluidics-fabricated hydrogel microspheres have emerged as a promising platform for oral delivery [21]. These systems can provide structural encapsulation to protect cargo in the gastric environment and undergo pH-responsive dissociation, enabling sustained release in the intestinal tract[[22], a), b)]. For example, Liu et al. developed ionically and covalently crosslinked alginate hydrogel microspheres with stress-relaxation behavior and acid resistance, enabling efficient encapsulation of probiotics and promoting the formation of multicellular microcolonies within the microspheres. This design significantly enhanced intestinal colonization efficiency following oral administration while maintaining favorable biocompatibility and delivery robustness [23]. Therefore, incorporating ALN-functionalized piezoelectric nanomaterials into such oral hydrogel microsphere systems is expected to simultaneously address gastrointestinal stability and absorption barriers, providing a safe, noninvasive, and systemically efficacious strategy for whole-skeleton delivery. In summary, this study focuses on the attenuation of motion-coupled bone bioelectrical rhythms as a key pathological dimension of osteoporosis and proposes a motion-activated oral skeletal “bioelectrical pacemaker” strategy. Specifically, BaTiO3 nanoparticles were employed as the core piezoelectric units and chemically functionalized with alendronate sodium (ALN) to enable selective skeletal accumulation while retaining anti-resorptive activity. These bone-targeting piezoelectric nanoparticles were subsequently encapsulated into alginate hydrogel microspheres via microfluidic fabrication, followed by deposition of a chitosan outer shell, yielding a structurally stable, orally administrable delivery system with intestinal targeting and sustained release capability. The gastrointestinal stability, intestinal targeting, and skeletal enrichment of the oral piezoelectric microspheres were systematically evaluated through simulated gastrointestinal release assays, in vivo fluorescence imaging of intestinal and bone tissues, and inductively coupled plasma–optical emission spectroscopy (ICP-OES) analysis. In vitro, osteoblast and osteoclast models were established to assess bone-cell responses and the biosafety of the microspheres. In vivo, an osteoporotic mouse model was employed, and the systemic therapeutic efficacy of the microsphere system was comprehensively evaluated using imaging, mechanical testing, skeletal electrophysiology, and biochemical analyses. Furthermore, transcriptomic profiling was performed to elucidate the molecular basis by which the microspheres regulate bone homeostasis. By integrating a closed-loop strategy encompassing oral delivery, bone-targeted enrichment, and motion-synchronized electrical activation, this work establishes a systemic intervention paradigm for osteoporosis based on the reconstruction of endogenous skeletal bioelectrical rhythms, providing a noninvasive therapeutic avenue. Results and discussion Fabrication and characterization of orally piezoelectric microspheres As an oral biomaterial intended for systemic skeletal modulation, Orally Piezoelectric Microspheres must not only exhibit good biocompatibility but also maintain structural integrity in the complex gastrointestinal environment and enable controlled release under intestinal conditions[[24], a), b)]. To this end, sodium alginate (Alg) and chitosan (CHI), two naturally derived polysaccharides, were selected to construct pH-responsive core–shell hydrogel microspheres [25]. Through electrostatic interactions between anionic and cationic chains, Alg and CHI form a polyelectrolyte complex network that preserves structural integrity under strongly acidic gastric conditions while gradually dissociating under near-neutral intestinal conditions, thereby enabling site-specific release of functional payloads. Using a gas-assisted microfluidic technique, highly monodisperse alginate droplets were generated and subsequently crosslinked in CaCl2 solution via rapid Ca2+-induced ionic gelation to yield size-uniform alginate hydrogel microspheres (HMs). A conformal CHI outer layer was then assembled on the microsphere surface through electrostatic adsorption between CHI and Alg, producing structurally stable core–shell hydrogel microspheres (Fig. 1B). Schematic illustration of the orally piezoelectric microsphere system (BTO@Aln@HMs/CHI). (A) Synthesis route of BTO@ALN via silane coupling and subsequent amidation. (B) Fabrication process of orally piezoelectric microspheres encapsulating BTO@ALN using gas-assisted microfluidic technology. (C) Schematic illustration of the in vivo behavior of orally piezoelectric microspheres. (D) Proposed mechanism of action of the orally piezoelectric microsphere system. Optical microscopy images showed that Orally Piezoelectric Microspheres after loading BTO@ALN exhibited a regular spherical morphology with a narrow size distribution, and that the piezoelectric nanoparticles were successfully encapsulated within the microspheres (Fig. 2O and P). Scanning electron microscopy further confirmed a well-defined hierarchical architecture, featuring a dense and uniform core with a continuous CHI shell conformally coating the surface (Fig. 2Q and R). The wrinkled surface observed in the SEM image of BTO@ALN@HMs/CHI mainly resulted from dehydration-induced shrinkage of the hydrogel network during sample preparation and vacuum drying, rather than irreversible collapse of the hydrated microspheres. Consistently, optical microscopy showed that the microspheres maintained a regular spherical morphology in the hydrated state. Meanwhile, the average particle size increased from 196.71 to 220.19 μm, and the surface ζ potential reversed from −15.63 to +37.42 mV (Fig. 2S and T), further verifying successful formation of the CHI shell. This core–shell architecture endows the microspheres with resistance to disintegration under gastric conditions and provides a structural basis for pH-responsive degradation and sustained release under intestinal conditions. Characterization of the orally piezoelectric microspheres. (A, B) TEM and crystal diffraction images of pristine barium titanate nanoparticles (BaTiO3) and alendronate-grafted barium titanate nanoparticles (BTO@ALN). (C, D) Particle size and zeta potential distributions of BaTiO3 and BTO@ALN. (E, F) SEM images of BTO@ALN. (G) Thermogravimetric analysis (TGA) of BTO@ALN. (H) X-ray diffraction (XRD) pattern of BaTiO3. (I) Energy-dispersive X-ray spectroscopy (EDS) elemental analysis of BTO@ALN. (J, K) Atomic force microscopy (AFM) phase and amplitude images of BTO@ALN. (L–N) Butterfly loop, polarization–electric field hysteresis loop, and d33 curves of BTO@ALN. (O) Optical microscopy image of alginate hydrogel microspheres (HMs). (P) Optical microscopy image of BTO@ALN-loaded alginate/chitosan hydrogel microspheres; white arrows indicate BTO@ALN nanoparticles. (Q, R) SEM images of HMs and BTO@ALN@HMs/CHI; white arrows indicate the alginate core and chitosan shell, respectively. (S, T) Particle size and zeta potential distributions of HMs and BTO@ALN@HMs/CHI. Regarding the piezoelectric functional core, barium titanate (BaTiO3, BTO) has been widely used in bone-related studies owing to its excellent piezoelectric properties and favorable biosafety profile [26]. More importantly, BaTiO3 can continuously convert mechanical energy into electrical signals under low-intensity mechanical stimulation, closely aligning with physiological movement [27]. Therefore, BTO nanoparticles were selected as the core piezoelectric units of the “skeletal pacemaker” and were further endowed with bone-targeting and antiresorptive capabilities via surface functionalization (Fig. 1A). BaTiO3 nanoparticles were synthesized via a hydrothermal method, and their XRD pattern matched the standard PDF card of tetragonal BaTiO3; notably, characteristic peak splitting was observed within the 2θ range of 42–50°, indicating a piezoelectrically active tetragonal crystal structure (Fig. 2H) [28]. Subsequently, alendronate (ALN) was grafted onto the BaTiO3 surface via silane coupling and amidation to construct piezoelectric nanoparticles (BTO@ALN) that integrate bone-targeted enrichment with mechanical-to-electrical transduction. TEM images indicated that the lattice fringes, crystal diffraction characteristics and overall morphology of BaTiO3 were largely preserved after modification, whereas BTO@ALN exhibited a continuous low-contrast surface coating accompanied by a slight increase in particle size, suggesting successful introduction of organic moieties (Fig. 2A–D). Meanwhile, the ζ potential shifted markedly toward more negative values, attributable to the strongly anionic phosphonate groups of ALN (Fig. 2C and D). EDS further confirmed a homogeneous distribution of phosphorus on the particle surface, and thermogravimetric analysis revealed a continuous mass loss of approximately 7.1% for BTO@ALN between 150 and 600 °C, consistent with thermal decomposition of the surface organic layer (Fig. 2G). Collectively, these results spanning morphology, surface charge, elemental composition, thermal behavior, and FTIR spectra verified successful grafting of ALN onto BTO. To further quantify the ALN conjugation/loading content, we established a ninhydrin-based colorimetric assay. ALN reacts with ninhydrin in pyridine solution to form a colored product with a characteristic absorption peak at 565 nm. Fig. S1 shows the characteristic absorption spectrum of the ALN–ninhydrin reaction product and the standard curve established at 565 nm. Based on the residual free ALN content in the supernatant and washing solutions after the reaction, the ALN loading content in BTO@ALN was calculated to be 6.83 wt%. To distinguish the biological contribution of tetragonal BTO-mediated piezoelectric activity from that of ALN functionalization, non-piezoelectric cubic-phase BaTiO3 nanoparticles were prepared as a control material and similarly functionalized with ALN. XRD and Scanning electron microscopy (SEM) confirmed the cubic-phase structure and morphology of nBTO nanoparticles (Fig. S2). This nBTO@ALN control allowed comparison between ALN-functionalized nanoparticles with and without effective piezoelectric activity under the same mechanical stimulation condition. To further validate surface grafting and core–shell assembly at the chemical-bond level, Fourier transform infrared (FTIR) spectroscopy was performed for the chitosan-coated BTO@ALN-loaded alginate hydrogel microspheres (BTO@ALN@HMs/CHI) system (Fig. 2U). A broad and intense absorption band centered at approximately 541 cm−1 was assigned to Ti–O stretching vibrations, indicating that the inorganic piezoelectric framework remained intact [29].A strong and broad band at 1024 cm−1, characteristic of siloxane network formation, was mainly attributed to asymmetric Si–O–Si stretching vibrations and further supported formation of Si–O–Ti bonds via condensation between silane groups and surface Ti–OH on BTO [30]. In addition, the appearance of a characteristic amide II band at approximately 1543 cm−1 suggested successful ALN grafting [31]. For the carrier components, the symmetric stretching vibration of –COO- at approximately 1432 cm−1 was characteristic of alginate; meanwhile, splitting/shifting of the –COO- bands around 1600 and 1410 cm−1, accompanied by changes in their intensity ratio, indicated formation of a polyelectrolyte complex layer between chitosan and alginate, consistent with the core–shell assembly mechanism[a), b), [32]]. These FTIR results, together with morphological, surface charge, elemental and thermal evidence, collectively confirm surface functionalization of BTO@ALN and its successful assembly within Alg/CHI core–shell microspheres. Finally, atomic force microscopy (AFM) was used to characterize the piezoelectric performance of BTO@ALN. Under an alternating electric field, the amplitude and phase images exhibited consistent bright/dark domain distributions and showed characteristic butterfly and hysteresis loops, indicating stable polarization switching behavior of BTO@ALN (Fig. 2J–M). Quantitative fitting of the d33 curve revealed a piezoelectric coefficient of 8.29 p.m. V−1, comparable to that of native human bone (7.66–8.72 p.m. V−1) (Fig. 2N) [33]. These results suggest that the engineered BTO@ALN nanoparticles can recapitulate the endogenous piezoelectric properties of bone, thereby providing a foundation for their function as a “skeletal pacemaker” to reconstruct motion-driven bioelectrical rhythms. Intestinal-Targeted Delivery of Orally Piezoelectric Microspheres and Evaluation of the bone-targeting capability of BTO@ALN To determine whether Orally Piezoelectric Microspheres can effectively protect the piezoelectric functional units during gastrointestinal transit, enable intestinal pH–responsive sustained release, and promote bone-targeted delivery and enrichment of BTO@ALN, we performed a systematic evaluation encompassing gastrointestinal stability and responsiveness, release kinetics, intestinal adhesion/retention, and in vivo skeletal distribution. First, to simulate sequential gastrointestinal transit after oral administration, simulated gastric fluid (SGF), simulated intestinal fluid (SIF), and simulated colonic fluid (SCF) were prepared, and the incubation protocol was designed according to adult physiological emptying times: 2 h in SGF and 4 h in SIF, followed by incubation in SCF until the microspheres were fully degraded (Fig. 3A)[a), b), [34]]. BTO@ALN@HMs/CHI and control microspheres lacking the CHI shell (BTO@ALN@HMs) were sequentially incubated in these media to evaluate structural stability and release behavior. Given the governing effect of the CHI shell on dissociation, release kinetics were further screened using chitosan with different molecular weights[a), b), [35]]. The results showed that CHI with a molecular weight of approximately 100 kDa enabled the microspheres to release approximately 95% of the piezoelectric cargo within 4 h in SIF, closely matching the physiological transit time of the small intestine; this formulation was therefore selected for subsequent studies (Fig. 3B). The slight shrinkage of BTO@ALN@HMs/CHI in SGF was mainly caused by acid-induced densification of the Alg/CHI hydrogel network (Fig. 3A). Under low-pH conditions, protonation of alginate carboxylate groups reduces chain hydration and electrostatic repulsion, resulting in contraction of the Ca2+-alginate network. Meanwhile, protonated chitosan strengthens the Alg/CHI polyelectrolyte complex shell, further restricting water penetration and premature cargo leakage. In SIF, increased pH and weakening of Ca2+-alginate crosslinking promote microsphere swelling and sustained release, explaining the transition from gastric shrinkage/stability to intestinal release b), [36]. Upon transfer to SIF, the microspheres gradually swelled and the shell slowly dissociated, resulting in sustained release; in SCF, the microspheres rapidly disintegrated and fully degraded. In contrast, BTO@ALN@HMs exhibited structural instability in SGF and released rapidly in SIF, resulting in a markedly earlier overall release profile (Fig. S3). Quantitative analysis further demonstrated that BTO@ALN@HMs/CHI reduced mass loss, particle-size collapse, and cargo leakage during the SGF stage, and extended the main release window in SIF from 0.75 h to 4 h (Fig. 3C–F), thereby providing a sufficient time window for subsequent absorption and systemic delivery. Intestinal-targeted delivery of orally piezoelectric microspheres and verification of the bone-targeting capability of BTO@ALN. (A) Optical microscopy images showing the morphology of orally piezoelectric microspheres at different time points. (B) Drug release profiles of microspheres fabricated with chitosan of different molecular weights. (C–F) Changes in mass, particle size, and cumulative drug release of the orally piezoelectric microspheres in simulated gastric and intestinal fluids at different time points. (G) In vivo biodistribution of fluorescently labeled BTO@ALN nanoparticles in mice. (H) Images of mechanical piezoelectric output electric field curves recorded at different time points in mice after a single oral administration of piezoelectric microspheres. (I, J) Ex vivo fluorescence images of fluorescently labeled BTO@ALN nanoparticles in mouse femora and corresponding quantitative analysis of fluorescence intensity. Intestinal adhesion and retention were then evaluated. Ex vivo imaging revealed that BTO@ALN@HMs/CHI exhibited the strongest fluorescence signal in the small intestine, with retention extended to approximately 4 h; in contrast, microspheres lacking the CHI shell were retained for <2 h, whereas free nanoparticles persisted only transiently for approximately 1 h (Fig. S4). This difference was likely attributable to multivalent interactions between chitosan and intestinal mucin, thereby enhancing mucoadhesion. On this basis, we further evaluated the in vivo bone-targeted distribution of BTO@ALN. Fluorescently labeled BTO@ALN or nonfunctionalized BTO was loaded into HMs/CHI microspheres; following oral gavage, biodistribution was tracked by in vivo fluorescence imaging with X-ray imaging as an anatomical reference. The fluorescence signal of BTO@ALN@HMs/CHI peaked at 8 h post-administration and showed substantial co-localization with skeletal regions over 4–24 h, whereas the signal of BTO@HMs/CHI decayed more rapidly and was predominantly localized to the renal region (Fig. 3G). To further verify whether bone-targeted BTO@ALN could generate functional electrical outputs after intraosseous accumulation, we measured the mechanically induced output voltage of hydrated ex vivo femurs under a 2 Hz, 10 N sinusoidal load after a single oral administration of BTO@ALN@HMs/CHI. The results showed that the output voltage of OVX femurs was markedly lower than that of the Sham group, suggesting impaired motion-coupled electromechanical responsiveness of bone tissue under osteoporotic conditions. After a single oral administration of BTO@ALN@HMs/CHI, the femoral output voltage was significantly enhanced at 24 h and remained at a relatively high level from day 1 to day 5. Although the signal attenuated by day 7, it remained higher than that of the OVX control group; by day 14, it further declined and approached the control level (Fig. 3H). These results indicate that intraosseously accumulated BTO@ALN can enhance the mechanoelectrical conversion response of osteoporotic bone under cyclic mechanical loading, thereby providing functional support for the once-weekly dosing regimen. Ex vivo fluorescence quantification of femurs showed that BTO@ALN@HMs/CHI remained higher than the control at multiple time points; although the signal decreased by day 7, residual fluorescence remained detectable at day 14 (Fig. 3I and J). To more quantitatively elucidate the bone tissue distribution kinetics, we employed ICP-OES to measure the Ba content in the femur at different time points after a single oral dose. Fig. S5A shows that femoral Ba content increased after a single oral dose, remained at detectable levels throughout the one-week dosing interval, and then gradually declined. Further regional analysis revealed that at 24 h post-dosing, Ba content in the proximal and distal femur was higher than that in the femoral shaft, suggesting a regional enrichment of BTO@ALN within the femur (Fig. S5B). This may be attributable to the abundance of cancellous bone in the proximal and distal regions, the larger mineralized surface area of trabecular bone, and more hydroxyapatite binding sites, which facilitate ALN-mediated binding and retention of BTO@ALN through its bisphosphonate structure. In contrast to the gradual decline of fluorescence signal after a single dose, ICP-OES results after four consecutive weekly doses showed that femoral Ba content remained elevated above the control level, supporting the formation of stable intraosseous enrichment under repeated dosing conditions (Fig. S6). To further verify whether the retained BTO@ALN within bone could provide sustained electromechanical output during the one-week dosing interval, we measured the mechanically induced electrical output of hydrated ex vivo femurs at different time points after a single oral administration. Specifically, a 2 Hz, 10 N sinusoidal mechanical load was applied to the distal femur, and the mechanically induced output voltage was recorded using fine wires connected to a bioamplifier–oscilloscope system (Fig. S7). After a single oral administration of BTO@ALN@HMs/CHI, the femoral output voltage was markedly enhanced at 24 h and remained at a relatively high level from day 1 to day 5. Although the output signal declined by day 7, it was still higher than that of the control group; by day 14, the signal was substantially reduced and approached the control level (Fig. 3H). This temporal profile provides functional support for the once-weekly dosing regimen: a single administration covers the major functional window of the one-week dosing cycle, whereas the attenuation after day 7 supports the rationale for replenishment at approximately weekly intervals. Furthermore, Fig. S5C shows a positive correlation between femoral Ba content and mechanically induced output voltage, indicating that intraosseous Ba accumulation was not merely a material biodistribution phenomenon, but was directly associated with functional electromechanical output. In other words, bone-targeted BTO@ALN accumulation could enhance the mechanoelectrical conversion response of femurs under cyclic mechanical loading. We further evaluated the in vivo excretion behavior of Ba after a single oral administration. The results showed that, within 168 h, the cumulative fecal Ba excretion rate was 62.37%, the cumulative urinary Ba excretion rate was 6.44%, and the total cumulative excretion rate was 68.81% (Tables S1 and S2). Feces represented the predominant excretion route, suggesting that Ba-containing components entering the body after oral administration could undergo substantial clearance within one week. Combined with the temporal kinetics of femoral Ba content and the mechanically induced output voltage results, this excretion profile supports that the system can achieve short-term functional intraosseous accumulation while also undergoing a certain degree of in vivo clearance. In summary, BTO@ALN@HMs/CHI maintained structural stability in the gastric fluid environment and achieved sustained release under simulated intestinal conditions. After oral administration, BTO@ALN achieved bone-targeted accumulation and generated enhanced mechanically induced output voltage under cyclic mechanical loading. After a single administration, this output remained at a relatively high level from day 1 to day 5, was still higher than the control level at day 7, and approached the control level by day 14. Together, these results provide experimental support for the once-weekly dosing regimen and subsequent motion-triggered piezoelectric activation from the perspectives of bone-targeted distribution, Ba quantification, functional electrical output, and excretion kinetics. Biocompatibility evaluation of orally piezoelectric microspheres As an oral delivery system for systemic skeletal modulation, Orally Piezoelectric Microspheres should exhibit favorable biocompatibility under complex physiological conditions. Accordingly, we systematically assessed biosafety using both in vitro cellular assays and in vivo tissue-level analyses. Human intestinal epithelial Caco-2 cells and primary mouse bone marrow–derived mesenchymal stem cells (BMSCs) were selected to model the intestinal barrier exposed during gastrointestinal transit and the downstream bone-related cells, respectively. The experimental groups included Control, BTO@ALN, BTO@ALN + Stretch, BTO@ALN@HMs, and BTO@ALN@HMs/CHI; cells were directly co-cultured with the corresponding formulations. Cell morphology, viability, and proliferation were evaluated on days 1, 3, and 5. Cytoskeletal fluorescence staining showed that Caco-2 cells in all groups spread well with intact, clearly defined F-actin structures, without evident contraction, cytoskeletal disorganization, or stress-associated morphologies (Fig. 4A). A subset of cells adhered to and grew along the surface of the hydrogel microspheres, suggesting favorable cytocompatibility at the material interface. Live/dead staining further confirmed that viable cells predominated in all groups, with no obvious cell death or localized cytotoxicity; incorporation of piezoelectric nanoparticles, microspheres, and the CHI shell did not measurably reduce cell viability (Fig. 4B). Under cyclic mechanical stretching (BTO@ALN + Stretch), cell morphology and viability remained comparable, indicating that piezoelectric activation did not introduce detectable additional cytotoxicity under these conditions. To quantitatively assess proliferation, CCK-8 assays were performed for Caco-2 cells and BMSCs after 1 and 3 days of co-culture. CCK-8 assays showed that the metabolic activity of Caco-2 cells and BMSCs remained comparable to that of the Control group after short-term co-culture with the tested formulations (Fig. 4C and D). Notably, BTO@ALN@HMs/CHI slightly increased Caco-2 cell metabolic activity on day 1, and BTO@ALN + Stretch showed a mild upward trend in both Caco-2 cells and BMSCs on day 3, although these changes did not reach statistical significance. This mild trend may be associated with the cell-friendly chitosan shell and stretch-triggered piezoelectric activation. Consistently, lactate dehydrogenase (LDH) release remained at a low level in all material-treated groups, indicating that the formulations did not induce evident membrane damage (Fig. 4E and F). Overall, these results indicate that the tested formulations maintained favorable cytocompatibility. Biocompatibility of the orally piezoelectric microspheres and in vitro verification of the anti-osteoclastogenic activity of BTO@ALN. (A, B) Cytoskeletal staining and live/dead fluorescence images of Caco-2 cells co-cultured with microspheres and related samples. (C–F) Cytotoxicity evaluation and CCK-8 assay of BMSCs and Caco-2 cells co-cultured with microspheres and related samples. (G) Schematic illustration of the cell-stretching system used to apply mechanical stimulation. (H) Immunofluorescence staining of osteoclasts differentiated from primary mouse bone marrow–derived macrophages (BMDMs) (red: actin; green: CTSK and TRAP). (I, J) Quantitative analysis of fluorescence intensity. (K, L) TRAP staining images of osteoclasts and corresponding quantification of TRAP-positive areas. (M) Expression of the osteoclast-related gene NFATc1 determined by RT–PCR. Considering that this system is administered orally, we further supplemented the evaluation of blood compatibility and intestinal barrier safety. The in vitro hemolysis assay showed that the microsphere-related groups did not induce obvious hemolytic responses (Fig. S8A and B). Meanwhile, routine blood analysis showed no obvious abnormal changes in RBC, HGB, or HCT among the groups, suggesting that no erythrocyte damage or marked hematological abnormality was observed under the current dosing regimen (Fig. S8F–H). Regarding intestinal safety, jejunal H&E staining showed intact small intestinal villus structures in all groups, without obvious inflammatory cell infiltration, epithelial disruption, or mucosal architectural disorder (Fig. S8C). ZO-1 immunofluorescence further showed continuous tight-junction structures without an obvious decrease in fluorescence intensity, suggesting that BTO@ALN@HMs/CHI did not cause detectable disruption of the intestinal tight-junction barrier (Fig. S8D and E). To further determine an appropriate therapeutic dose, we supplemented dose-screening and soft-tissue Ba exposure analyses. As shown in Fig. S9, the low-dose group showed limited improvement in trabecular microarchitecture and BMD, indicating insufficient bone-protective efficacy. The medium-dose group exhibited clear restoration of trabecular architecture and increased BMD. Although the high-dose group produced only a slight additional improvement in bone mass compared with the medium-dose group, Ba contents in the liver, kidney, and spleen were significantly increased and markedly higher than background levels, suggesting an increased risk of soft-tissue exposure. Therefore, 0.1 mg BTO@ALN nanoparticle equivalent per 10 g body weight was selected as the therapeutic dose for subsequent OVX treatment experiments, in order to balance effective bone protection with controlled soft-tissue Ba exposure. This dose-screening result provided dose-related and safety support for the once-weekly administration regimen over 8 consecutive weeks. To further evaluate the long-term systemic safety after repeated oral administration, we examined the histology of major organs, Ba contents in the liver/kidney/spleen, and serum ALT, AST, and CREA levels after treatment. H&E staining showed intact structures of the heart, liver, spleen, lung, and kidney, without obvious inflammation, necrosis, foreign-body reaction, or vascular obstruction-like changes (Fig. S10A). Meanwhile, no abnormal Ba accumulation was observed in the liver, kidney, or spleen, and ALT, AST, and CREA showed no obvious elevation, suggesting that no detectable major-organ toxicity or hepatic/renal functional impairment occurred under the current dosing regimen (Fig. S10B and C). Taken together, the in vitro and in vivo safety evaluations indicate that BTO@ALN@HMs/CHI exhibits favorable preliminary biocompatibility and systemic safety within the current experimental dose and observation window. Combined with the dose-screening results, this system can achieve effective intraosseous accumulation and bone-protective effects while avoiding obvious soft-tissue Ba exposure and major-organ toxicity, thereby providing a safety and dosing basis for subsequent in vivo therapeutic experiments. In vitro evaluation of the anti-osteoclastogenic activity of BTO@ALN Osteoporosis is closely associated with dysregulation of the bone remodeling network [37]. Mechanical–electrophysiological cues mediated by the bone matrix constitute an important physical input for maintaining the dynamic balance between bone resorption and bone formation, and attenuation of these cues is considered a key contributor to impaired bone homeostasis [5,38]. Under pathological conditions such as osteoporosis, insufficient mechanical stimulation, weakened piezoelectric properties of the bone matrix, and disturbed local bioelectrical rhythms collectively impair osteoblast–osteoclast coupling and promote osteoclast overactivation a), b). Therefore, reconstructing effective mechanical-to-electrical signal inputs under physiologically relevant mechanical stimulation may suppress aberrant bone resorption and help restore remodeling homeostasis. Based on this rationale, a RANKL-induced osteoclastogenesis model was established using primary mouse bone marrow–derived macrophages (BMDMs), and a cyclic stretching system was introduced to mimic physiological mechanical stress via low-intensity periodic stretching (1% strain, 1 Hz, sinusoidal waveform, 2 h day−1). This stimulation was also used to trigger the piezoelectric effect of BTO@ALN nanoparticles, aiming to emulate motion-associated endogenous bioelectrical inputs (Fig. 4G). The in vitro experimental groups included Non-RANKL, Control, Stretch, BTO@ALN, nBTO@ALN + Stretch, and BTO@ALN + Stretch. The nBTO@ALN + Stretch group was treated with ALN-functionalized non-piezoelectric cubic-phase BaTiO3 nanoparticles under the same cyclic stretching conditions, and was used to distinguish the relative contributions of ALN functionalization, mechanical stretching, and tetragonal BTO-mediated piezoelectric activation. Immunofluorescence co-staining showed that RANKL induction markedly increased TRAP and CTSK expression, accompanied by cytoskeletal remodeling and cell fusion into multinucleated osteoclast-like cells; TRAP staining likewise revealed abundant TRAP-positive multinucleated cells, supporting successful model establishment (Fig. 4H–K). For intervention evaluation, periodic stretching alone (Stretch group) slightly reduced TRAP/CTSK signals without obvious changes in osteoclast morphology or number, suggesting limited inhibition by low-intensity mechanical stimulation per se. In contrast, BTO@ALN treatment markedly decreased TRAP and CTSK signals and reduced the number of large, multinucleated osteoclasts (Fig. 4H–J), consistent with suppressed osteoclast differentiation and maturation. Notably, BTO@ALN + Stretch further inhibited the formation of TRAP-positive multinucleated osteoclasts compared with BTO@ALN or Stretch treatment alone, and reduced osteoclast-related markers, including TRAP and CTSK. More importantly, under the same stretching conditions, the anti-osteoclastic effect of the nBTO@ALN + Stretch group was weaker than that of the BTO@ALN + Stretch group, indicating that ALN modification and mechanical stretching alone could not fully account for the enhanced effect of BTO@ALN + Stretch. These results support that stretch-triggered piezoelectric output mediated by tetragonal BTO provides an additional contribution to the inhibition of osteoclast differentiation (Fig. 4H–L). To corroborate these effects at the molecular level, NFATc1 mRNA expression, a key transcriptional regulator of osteoclastogenesis, was analyzed. qPCR results showed that BTO@ALN significantly downregulated NFATc1, and this inhibitory effect was further enhanced under mechanical stimulation (Fig. 4M), suggesting that piezoelectrically triggered bioelectrical cues may modulate osteoclast differentiation programs at the transcriptional level. Overall, these results demonstrate that, under physiologically relevant mechanical stress, BTO@ALN nanoparticles can generate rhythmic bioelectrical signals via their piezoelectric effect. This piezoelectrically mediated output may act synergistically with the intrinsic antiresorptive activity of ALN to markedly suppress osteoclast differentiation and function, thereby providing mechanistic support for restoring bone remodeling balance. In vitro evaluation of the osteogenic-promoting activity of BTO@ALN Bone remodeling is a highly dynamic process, and homeostasis relies on finely tuned coupling between bone formation and bone resorption under the integration of multiple cues [39]. Endogenous bioelectrical signals represent an intrinsic physical regulatory dimension of bone and are characterized by continuity and rhythmicity, contributing to the osteogenic microenvironment b), [40]. Previous studies have shown that osteoblast-lineage cells are highly sensitive to electrical cues, and electrical stimulation or electroactive materials can modulate their differentiation and mineralization [41]. Unlike exogenous intermittent electrical stimulation, physiological electrical signals within bone more closely resemble dynamic rhythmic inputs coupled to daily movement and loading. Therefore, establishing a piezoelectric system that converts physiologically relevant mechanical stimuli into synchronized electrical outputs is important for mimicking and reconstructing the endogenous electrophysiological microenvironment of bone. On this basis, we evaluated the effects of BTO@ALN on osteogenic differentiation in an in vitro model under mechanically induced piezoelectric activation. BMSCs were cultured in osteogenic induction medium and directly co-cultured with BTO@ALN nanoparticles. Meanwhile, low-intensity cyclic mechanical stimulation (1% strain, 1 Hz, sinusoidal waveform, 2 h day−1) was applied using a cell-stretching system to trigger the piezoelectric effect and aim to emulate motion-associated rhythmic bioelectrical inputs (Fig. 4G). The experimental groups included Non-OIC, Control, Stretch, BTO@ALN, nBTO@ALN + Stretch, and BTO@ALN + Stretch. The effectiveness of osteogenic induction was first validated. Immunofluorescence results showed that, compared with the Non-OIC group, the Control group exhibited markedly enhanced nuclear RUNX2 and ATF4 signals, accompanied by a morphological transition from spindle-like cells toward osteoblast-like polygonal and triangular phenotypes (Fig. 5A and B). Functional staining further showed that ALP-positive signals increased after 7 days of osteogenic induction, and ARS-positive mineralized nodules appeared by day 21, supporting successful osteogenic induction (Fig. 5E). In intervention assessments, cyclic stretching alone (Stretch) or BTO@ALN alone induced only modest increases in RUNX2/ATF4; ALP and ARS staining showed similar trends, with no significant differences overall (Fig. 5A–H). This may be attributable to the low-intensity stretching parameters, which were selected to minimize the effect of mechanical stimulation per se and better resolve the contribution of piezoelectrically generated electrical cues. Notably, BTO@ALN + Stretch significantly enhanced RUNX2 and ATF4 expression and increased the intensity of ALP and ARS staining, indicating its ability to promote osteogenic differentiation and mineralization. Compared with nBTO@ALN + Stretch, the BTO@ALN + Stretch group exhibited stronger RUNX2/ATF4 activation and more pronounced ALP/ARS-positive staining. These results indicate that, under the same ALN modification and mechanical stretching conditions, the piezoelectric response of tetragonal BTO can further enhance osteogenesis-related biological effects (Fig. 5A–H). To further assess osteogenic programs at the molecular level, transcriptional levels of COL1A1 and Osterix were analyzed. The Stretch and BTO@ALN groups showed only slight upregulation, whereas the BTO@ALN + Stretch group exhibited a pronounced increase (Fig. S11), suggesting that piezoelectrically generated bioelectrical cues may modulate osteogenic differentiation at the transcriptional level. In vitro evaluation of the osteogenic-promoting capability of BTO@ALN. (A–D) Immunofluorescence staining of osteoblasts differentiated from primary mouse bone marrow–derived mesenchymal stem cells (BMSCs) (red: actin; green: RUNX2 and ATF4) and corresponding quantitative analysis of fluorescence intensity. (E–H) Alkaline phosphatase (ALP) and Alizarin Red S (ARS) staining of osteoblasts derived from primary mouse BMSCs and quantitative analysis of staining intensity. Overall, these results show that BTO@ALN generates motion-synchronized electrical outputs under physiologically relevant mechanical stimulation, thereby promoting osteogenic differentiation and mineralization of BMSCs. This provides cellular-level support for the “skeletal pacemaker” strategy to reconstruct bone homeostasis. In vivo therapeutic efficacy of orally piezoelectric microspheres for osteoporosis To evaluate this strategy under an estrogen-deficiency osteoporotic condition, we employed an ovariectomized (OVX) mouse model, which is widely used to reproduce postmenopausal bone loss characterized by enhanced bone resorption, trabecular deterioration, and disrupted remodeling balance. In the present study, the OVX model was used to establish an osteoporotic pathological background and to assess whether skeletal electromechanical responsiveness was impaired and could be enhanced by bone-targeted piezoelectric microspheres. The core pathology of osteoporosis is not simply a unidirectional change in either arm of remodeling, but rather a sustained disruption of the remodeling network under prolonged loss of effective physiological cues, manifested as a persistent shift in osteoclast–osteoblast coupling [42]. Our in vitro data suggest that BTO@ALN, under mechanical stimulation, can generate movement-synchronized rhythmic bioelectrical signals and is associated with concurrent suppression of osteoclast differentiation and promotion of osteogenic differentiation. To determine whether this “motion-synchronized electrical activation” strategy can re-establish remodeling homeostasis in vivo and mitigate osteoporotic phenotypes, we evaluated the therapeutic efficacy of Orally Piezoelectric Microspheres in an ovariectomized (OVX) mouse model. To evaluate the therapeutic efficacy of orally administered piezoelectric microspheres in the OVX osteoporotic model, interventions were initiated at 2 weeks after OVX surgery. Mice received BTO@ALN@HMs/CHI by oral gavage once weekly for 8 consecutive weeks, with low-intensity treadmill exercise combined in the corresponding groups. Low-intensity exercise was defined as treadmill running at a 0° incline and 5 m min−1 for 20 min per session, three times per week, for 8 consecutive weeks. This regimen was designed to provide regular, mild, and reproducible cyclic mechanical inputs to trigger mechanically induced outputs from the piezoelectric units accumulated in bone, while minimizing the effect of exercise itself as a strong therapeutic factor.The in vivo experimental groups included Sham, OVX, Sports, Aln, BTO@ALN@HMs/CHI, nBTO@ALN@HMs/CHI + Sports, and BTO@ALN@HMs/CHI + Sports. The Aln group was included to compare the pharmacological effect of clinically used oral alendronate. The nBTO@ALN@HMs/CHI + Sports group was used to control for ALN functionalization, microsphere delivery, and exercise stimulation, but without effective piezoelectric activity (Fig. 6A) [43]. At week 10, femora were harvested for micro-CT analysis, three-point bending testing, histological evaluation, d33 measurement, and serum bone turnover marker assays. Radiological evaluation of OVX mice treated with orally piezoelectric microspheres. (A) Schematic illustration of the animal experimental design. (B) Representative micro–computed tomography (micro-CT) reconstructions of cortical and trabecular bone in mouse femora from different groups at 12 weeks. (C–F) Quantitative analysis of bone mineral density, bone volume fraction, trabecular number, and trabecular thickness. (G, H) Statistical analysis and illustrations of ultimate fracture stress of mouse femur in three-point bending test. (I, J) Serum levels of CTX-1 and PINP in mice. Micro-CT 3D reconstructions showed that the OVX group exhibited markedly sparse, thinned, and discontinuous trabeculae in the distal femur with a severely disrupted trabecular network, together with reduced cortical thickness—features consistent with an osteoporotic phenotype (Fig. 6B). Quantitatively, BMD in the OVX group decreased to about 42% of that in the Sham group, supporting successful model establishment (Fig. 6C). Low-intensity exercise alone (Sports group) produced only a mild, nonsignificant improvement, consistent with the intended sub-therapeutic regimen. In contrast, BTO@ALN@HMs/CHI treatment significantly improved bone mass and microarchitecture, restoring BMD to about 55% of the Sham level (Fig. 6C–F). Further, the effective piezoelectric coefficient (d33) of femurs from OVX mice was lower than that of the Sham group, suggesting impaired mechanoelectrical conversion capacity in osteoporotic bone. At 24 h after oral administration of BTO@ALN@HMs/CHI, femoral d33 increased by approximately 52%, indicating that the bone-targeted piezoelectric units rapidly enhanced the mechanical-to-electrical responsiveness of osteoporotic femurs. Together with the multi-time-point measurements of mechanically induced output voltage shown in Fig. 3H, these data support that intraosseous accumulation of BTO@ALN can provide functional electromechanical outputs within the one-week dosing window (Fig. S12). Therapeutic outcomes were further amplified when combined with low-intensity exercise: the BTO@ALN@HMs/CHI + Sports group exhibited a more pronounced increase in d33, with BMD restored to about 72%, and key parameters including Tb.N, BV/TV, and Ct.Th were significantly improved relative to BTO@ALN@HMs/CHI alone (Fig. 6C–F). In addition, the Aln group partially ameliorated OVX-induced bone loss, particularly as reflected by improvements in bone mass and selected bone microarchitectural parameters, consistent with the expected antiresorptive effect of alendronate. However, the BTO@ALN@HMs/CHI + Sports group showed more comprehensive improvements in BMD, BV/TV, Tb.N, and Ct.Th. More importantly, under the same exercise conditions, the BTO@ALN@HMs/CHI + Sports group significantly outperformed the nBTO@ALN@HMs/CHI + Sports group, indicating that ALN, microsphere delivery, and exercise stimulation alone could not fully account for the overall therapeutic efficacy. These results support that motion-triggered piezoelectric activation mediated by tetragonal BTO provided an additional therapeutic contribution. These findings support that periodic exercise can serve as an effective mechanical trigger to potentiate the in vivo piezoelectric “pacemaker” effect. To determine whether the structural improvements translated into functional benefits, we performed three-point bending tests [44]. Three-point bending results showed that the femoral mechanical properties in the Aln group were partially restored compared with those in the OVX group, suggesting that antiresorptive therapy alone could improve bone quality to some extent. However, the BTO@ALN@HMs/CHI + Sports group exhibited more pronounced improvements in mechanical parameters, including maximum load and bending strength, indicating that this combined strategy not only improved bone mass but also further enhanced the overall mechanical integrity of bone tissue (Fig. 6G and H). Serological analysis showed that the OVX group exhibited a marked increase in the bone resorption marker CTX-1 and a pronounced decrease in the bone formation marker PINP. BTO@ALN@HMs/CHI treatment reduced CTX-1 and partially restored PINP; when combined with low-intensity exercise, CTX-1 further decreased and PINP increased significantly, approaching Sham levels (Fig. 6I and J), consistent with the micro-CT and mechanical findings. Serum bone turnover markers further revealed the distinct effects of different therapeutic strategies. The Aln group markedly reduced CTX-1, consistent with its classical antiresorptive activity, but showed a relatively limited effect on the restoration of PINP. In contrast, the BTO@ALN@HMs/CHI + Sports group not only further reduced CTX-1, but also more effectively restored PINP levels, suggesting that motion-triggered piezoelectric activation further promoted bone formation-related remodeling regulation on the basis of ALN-mediated antiresorptive effects. Histologically, H&E staining of the trabecular region beneath the growth plate showed that the OVX group exhibited reduced trabecular area, expanded marrow cavities, and disrupted architecture; these changes were substantially alleviated by BTO@ALN@HMs/CHI treatment, and trabeculae became denser and more continuous with exercise (Fig. 7A). TRAP staining revealed numerous TRAP-positive osteoclasts lining trabecular surfaces in the OVX group, whereas BTO@ALN@HMs/CHI significantly reduced osteoclast numbers; with additional exercise, both the number and coverage area of TRAP-positive cells further decreased (Fig. 7C–F). Notably, in the above histological evaluation of femoral tissues, no obvious diffuse inflammatory cell infiltration, bone tissue necrosis, granuloma-like foreign-body reaction, or abnormal fibrous encapsulation was observed in either the BTO@ALN@HMs/CHI group or the BTO@ALN@HMs/CHI + Sports group. Together with the TRAP, Masson, and osteogenesis-/osteoclastogenesis-related immunostaining results, these findings indicate that, within the current dosing regimen and observation window, intraosseously accumulated BTO@ALN did not induce detectable local bone inflammation or foreign-body reactions, while still regulating the bone remodeling process. Consistently, quantitative immunofluorescence analyses of CTSK and MMP-9 demonstrated suppression of resorption-related protein expression, which was further enhanced by exercise (Fig. 7B–D,E). Histological results further supported this interpretation. The Aln group reduced the number of TRAP-positive osteoclasts and decreased the expression of bone resorption-related markers, including CTSK and MMP-9, indicating the expected antiresorptive effect of ALN. The BTO@ALN@HMs/CHI + Sports group exhibited a stronger inhibitory effect on osteoclast activity. Compared with the nBTO@ALN@HMs/CHI + Sports group, the BTO@ALN@HMs/CHI + Sports group showed a further reduction in osteoclast-related staining, suggesting that piezoelectric activation of tetragonal BTO enhanced the antiresorptive effect under exercise conditions. In parallel, Masson's trichrome staining showed decreased collagen deposition and a discontinuous bone matrix in the OVX group; BTO@ALN@HMs/CHI increased collagen deposition, and the matrix became denser and more organized with exercise (Fig. 8A). COL1A1 expression was upregulated after treatment and was further enhanced in the combined group (Fig. 8C–F). Moreover, OPN and OCN co-staining revealed substantially increased expression around trabeculae after treatment, with further expansion in distribution and intensity in the combined group (Fig. 8B–D,E), supporting enhanced osteogenic activity in the trabecular compartment. Unlike the Aln group, which mainly exhibited an antiresorptive effect, the BTO@ALN@HMs/CHI + Sports group showed more pronounced collagen deposition in Masson staining and significantly enhanced the expression of osteogenesis-related markers, including COL1A1, OPN, and OCN. Compared with the nBTO@ALN@HMs/CHI + Sports group, the BTO@ALN@HMs/CHI + Sports group displayed stronger osteogenic signals, indicating that motion-triggered piezoelectric activation of tetragonal BTO further promoted bone formation beyond the pharmacological effect of ALN and exercise-induced mechanical stimulation. Osteoclast-related histological analyses of OVX mice treated with orally piezoelectric microspheres. (A) Representative H&E-stained images of mouse femora with magnified views of the trabecular region. (B) Immunofluorescence staining of osteoclast-associated proteins in mouse femora (red: CTSK; green: MMP-9). (C) Representative high-magnification images of TRAP staining in mouse femora. (D, E) Quantitative analysis of immunofluorescence intensity. (F) Quantification of TRAP-positive osteoclasts in mouse femora. Osteogenesis-related histological analyses of OVX mice treated with orally piezoelectric microspheres. (A) Representative Masson's trichrome–stained images of mouse femora with magnified views of the trabecular region. (B) Immunofluorescence staining of osteogenesis-associated proteins in mouse femora (red: OCN; green: OPN). (C) Representative high-magnification immunohistochemical images of the bone formation–related marker COL-1 in mouse femora. (D, E) Quantitative analysis of immunofluorescence intensity. (F) Quantification of COL-1 expression in mouse femora. Overall, orally delivered BTO@ALN@HMs/CHI significantly improves bone microarchitecture and mechanical performance in OVX mice, while concurrently suppressing bone resorption and promoting bone formation at the cellular and molecular levels. Low-intensity periodic exercise, serving as a physiological mechanical trigger, further amplifies the piezoelectric “pacemaker” efficacy, supporting improved restoration of bone remodeling homeostasis. Mechanistic insights into bone homeostasis restoration by orally piezoelectric microspheres To elucidate the molecular basis by which Orally Piezoelectric Microspheres restore bone homeostasis in vivo, RNA-seq was performed on mouse femoral tissue. The results suggested that this system concurrently modulates osteoclast- and osteogenic-related biological processes and induces broad transcriptomic remodeling. Compared with the OVX group, differential expression analysis in the BTO@ALN@HMs/CHI group identified 21 downregulated genes and 15 upregulated genes (Q value < 0.05, |log2FC| > 1), among which Gpnmb, a key gene associated with osteoclast differentiation, maturation, and resorptive function, was significantly downregulated (Fig. 9A). Given that Gpnmb is implicated in osteoclast cytoskeletal remodeling, adhesion/migration, and resorptive activity, its downregulation supports suppression of osteoclast maturation and function [45]. This finding is consistent with the in vivo phenotype of reduced bone resorption, as reflected by TRAP staining, decreased CTSK/MMP-9 expression, and serum bone turnover markers. Transcriptomic analysis of femora from OVX mice treated with orally piezoelectric microspheres. (A) Volcano plot comparing the BTO@Aln@HMs/CHI group with the OVX group. (B) GeneMANIA network analysis of the key differentially expressed gene GPNMB. (C) Chord diagram of Gene Ontology (GO) functional enrichment analysis of differentially expressed genes. (D) Gene set enrichment analysis (GSEA). (E) Volcano plot comparing the BTO@Aln@HMs/CHI + Sports group with the Sports group. (F) GeneMANIA network analysis of the key differentially expressed gene TRPA1. (G) Chord diagram of GO functional enrichment analysis of differentially expressed genes. (H) GSEA. To further interpret the biological significance of Gpnmb downregulation, we constructed its functional interaction network, which highlighted associations with lipid metabolism, membrane organization/maintenance, and ERK1/2 signaling (Fig. 9B). GO enrichment analysis showed that the differentially expressed genes were significantly enriched in bone resorption–related processes (Fig. 9C). Consistently, GSEA supported overall suppression of the osteoclast differentiation pathway (Fig. 9D). Together with the in vitro osteoclast assays and the in vivo evidence from TRAP staining and reduced CTSK/MMP-9 expression, the RNA-seq results support a molecular profile consistent with suppressed osteoclast differentiation and function. This profile is compatible with the canonical pharmacological mechanism of ALN: as a nitrogen-containing bisphosphonate, ALN inhibits farnesyl pyrophosphate synthase (FPPS) in the mevalonate pathway, thereby blocking isoprenylation of small GTPases and preventing stable formation of osteoclast cytoskeletal and resorptive structures (sealing zone/ruffled border), ultimately attenuating bone resorption. Meanwhile, suppression of ERK1/2 signaling and NFATc1-centered transcriptional programs has been implicated in reduced expression of resorption-related genes such as Gpnmb, CTSK, and MMP-9[[46], a), b)]. Accordingly, we propose that the primary osteoclast-side action cascade of BTO@ALN@HMs/CHI follows “ALN-mediated FPPS/mevalonate inhibition impaired structural assembly and signal transduction downregulation of resorption programs,” and that the Gpnmb-centered changes together with the network and enrichment patterns provide transcriptomic support for this model (Fig. 9A–D). Regarding mechanistic insights associated with motion-activated piezoelectric stimulation, the BTO@ALN@HMs/CHI + Sports group exhibited a more extensive transcriptomic shift than the Sports group, with 595 genes upregulated and 319 genes downregulated (Q value < 0.05, |log2FC| > 1) (Fig. 9E). Notably, Trpa1, an ion-channel gene implicated in osteogenesis and mechanosensation, was significantly upregulated [47]. GeneMANIA analysis showed that the TRPA1-associated network was enriched in Ca2+ transmembrane transport and osteogenesis-relevant pathways such as PI3K/AKT signaling (Fig. 9F); GO enrichment highlighted metal ion binding and cytosolic signaling processes (Fig. 9G); and GSEA indicated significant upregulation of osteogenic differentiation pathways (Fig. 9H). Taken together, these data are consistent with the possibility that motion-triggered piezoelectric outputs may promote osteogenic programs through a TRPA1–Ca2+–PI3K/Akt–RUNX2–linked axis. We propose that, under low-intensity exercise, BTO@ALN enriched on bone surfaces undergoes motion-synchronized microdeformation and generates rhythmic electrical cues, which could activate TRPA1 on osteoprogenitor cell membranes to induce Ca2+ influx; elevated cytosolic Ca2+ may subsequently facilitate PI3K/Akt activation, enhance nuclear accumulation and transcriptional activity of RUNX2, and thereby upregulate osteogenesis-related genes such as Col1a1, Alp, and Ocn to promote matrix synthesis and mineral deposition[47,48]. In addition, comparative analysis between the OVX and Sham groups identified 2023 differentially expressed genes, and both GO and GSEA enrichment analyses converged on osteogenic and osteoclast-related functional pathways (Fig. S13). Overall, this study supports a dual-pathway model whereby “motion-synchronized electrical activation” restores bone homeostasis: (i) ALN suppresses bone resorption by inhibiting the FPPS/mevalonate pathway, thereby impairing osteoclast structural assembly and signal transduction; and (ii) motion-triggered piezoelectric outputs may promote osteogenesis via TRPA1-mediated Ca2+ signaling and downstream activation of the PI3K/Akt–RUNX2 axis, ultimately rebalancing osteoclast and osteoblast activities. Mechanistically, the bone-enriched piezoelectric units may interact with bone remodeling biology by converting cyclic mechanical loading into local electrical cues, which could modulate membrane potential, Ca2+ influx, and downstream mechanosensitive signaling in osteogenic and osteoclastic cells, thereby influencing intercellular remodeling crosstalk within the bone microenvironment. From a translational perspective, this oral delivery plus motion-triggered activation strategy may be particularly suitable for osteopenic or osteoporotic patients who retain the ability to perform regular low-intensity physical activity. Nevertheless, its clinical translation will require further optimization of long-term oral safety, bone-retention biocompatibility, patient-specific activity dependence, and standardized exercise prescriptions. Conclusion We developed motion-driven Orally Piezoelectric Microspheres that aim to re-establish the motion-coupled endogenous bioelectrical rhythms attenuated in osteoporosis via a “skeletal pacemaker” strategy. The platform uses tetragonal BaTiO3 as the piezoelectric core and grafts ALN to enable bone-targeted enrichment with antiresorptive synergy; microfluidic encapsulation within chitosan/alginate core–shell hydrogel microspheres affords gastric protection and pH-responsive release in the small intestine, enabling noninvasive delivery at the whole-skeleton level. After oral administration, skeletal enrichment of the piezoelectric nanoparticles increased by 6.8-fold versus controls, and under mechanical loading during daily activity or low-intensity exercise, the piezoelectric units were activated to generate rhythmic bioelectrical signals synchronized with motion, thereby partially compensating for impaired electrophysiological responsiveness and increasing bone piezoelectric performance by 52% at 24 h. Accordingly, the treatment achieved bidirectional in vivo modulation by suppressing bone resorption while promoting bone formation. The addition of low-intensity exercise further amplified the efficacy of BTO@ALN@HMs/CHI through motion-coupled piezoelectric activation; compared with BTO@ALN@HMs/CHI alone, the combined BTO@ALN@HMs/CHI + Sports treatment increased osteogenesis-related outcomes by 2.4-fold and enhanced bone resorption inhibition by 1.5-fold, thereby improving bone microarchitecture and mechanical properties in OVX mice. Mechanistically, the coupled inputs of rhythmic bioelectrical cues and drug-mediated bone targeting are expected to remodel the bone remodeling network and promote coordinated osteoclast suppression and osteogenic activation toward homeostasis. Overall, this work establishes a closed-loop strategy integrating oral delivery, bone-targeted enrichment, and motion-synchronized electrical activation, advancing a biomaterials-based intervention paradigm for osteoporosis centered on the reconstruction of endogenous bioelectrical rhythms. Experimental section Preparation and Characterization of BTO@ALN Nanoparticles: Tetragonal barium titanate (BaTiO3) piezoelectric nanoparticles were synthesized via a solvothermal route. Briefly, anhydrous ethanol (20 mL) was used as the reaction medium, into which Ti(OC4H9)4 (1 mL, Sigma-Aldrich, USA) and Ba(OH)2·8H2O (1 g, Alfa Aesar) were sequentially added, followed by dropwise addition of H2O2 (0.4 mL, Fisher Chemical) as an oxidizing agent. A pre-stirring step (70 °C, 0.5 h) was conducted to homogenize the dispersion of the precursors. The homogenized mixture was then sealed in a Teflon-lined stainless-steel autoclave and subjected to a hydrothermal reaction at 180 °C for 24 h to induce crystallization. Subsequently, the precipitate was isolated via vacuum filtration, rinsed alternately with ethanol and deionized water to remove residual ions and unreacted components, and dried to obtain BaTiO3 nanoparticles. For hydroxylation, 100 mg of BaTiO3 nanoparticles were mixed with 30% H2O2 in a conical flask and stirred at 85 °C for 5 h. The resulting material was then centrifuged and rinsed with 95% ethanol to obtain hydroxylated BaTiO3 nanoparticles (BaTiO3–OH). Next, COOH–PEG–silane (10 mg, Aladdin, China) and BaTiO3–OH (10 mg) were dispersed in 95% ethanol and subjected to probe sonication (Qsonica) for 2 h (amplitude 30; on/off = 5/30 s), followed by sealing and reaction overnight. The final product was isolated by centrifugation and subjected to washing cycles, yielding BaTiO3–PEG–COOH. Finally, BaTiO3–PEG–COOH was suspended in 0.1 M MES buffer at 1 mg mL−1 and sonicated for 10 min. EDC·HCl (10 mg, Aladdin, China) and NHS (5 mg, Macklin, China) were added and the mixture was activated under shaking at 37 °C for 0.5 h. After centrifugation, the solid was washed once using MES buffer and then re-suspended in PBS (pH 7.4). And then reacted with alendronate sodium (SJ-MX2276, Sparkjade) under stirring overnight. BaTiO3–PEG–ALN (BTO@ALN) was obtained after centrifugation, triple washing with deionized water to remove residual ALN, and lyophilization. The crystal structure and phase composition were analyzed by X-ray diffraction (XRD; Bruker, Germany). Morphology and surface microstructure were observed using field-emission scanning electron microscopy (FE-SEM; Hitachi, Japan). Internal structure and elemental distribution were further characterized by transmission electron microscopy coupled with energy-dispersive spectroscopy (TEM/EDS; JEOL, Japan). Hydrodynamic diameter and size distribution were determined using dynamic light scattering (Zetasizer Nano; Malvern Panalytical, UK). Thermal behavior was assessed via thermogravimetric analysis (TGA; TA Instruments, USA) with samples heated from ambient temperature to 800 °C at a rate of 10 °C min−1. Functional groups were characterized by Fourier transform infrared spectroscopy (FTIR; VERTEX, Bruker, Germany). Piezoelectric properties were measured by atomic force microscopy (AFM; Park Systems, South Korea). in piezoresponse force microscopy (PFM) mode. Lyophilized BTO@ALN nanoparticles were pressed into a pellet, mounted on a metallic sample holder, and scanned using a conductive probe in contact mode to acquire amplitude and phase images. Subsequently, a DC bias sweep with a superimposed AC excitation was applied at a selected location to record the phase hysteresis and amplitude butterfly loops, from which the d33V curve was derived. Ex vivo measurement of mechanically induced femoral output voltage:At predetermined time points after a single oral administration of BTO@ALN@HMs/CHI, mice were euthanized and femora were harvested. After removal of surrounding soft tissues, the femora were maintained in phosphate-buffered saline (PBS) to preserve hydration before testing. Each wet femur was placed on a mechanical loading platform, and a sinusoidal load of 2 Hz and 10 N was applied to the distal femur. Two thin-wire electrodes were attached to the proximal and distal regions of the femur and connected to a differential bioelectric amplifier (Model 1700, A-M Systems, USA) and a digital oscilloscope (UPO1000S-E, Uni-Trend Technology,China) to record mechanically induced output voltage signals. For each femur, at least three stable loading cycles were recorded, and the peak-to-peak output voltage was averaged for quantitative analysis. Repeated loading cycles from the same femur were treated as technical replicates, whereas each femur from an independent mouse was considered one biological replicate. The experimental setup is shown in Fig. S7, and representative output voltage curves are shown in Fig. 3H. Preparation and Characterization of Orally Piezoelectric Microspheres: Microspheres were fabricated using a gas-assisted microfluidic technique. A 1% w/v sodium alginate solution (Aladdin, China) was prepared, and BTO@ALN nanoparticles were added to obtain a 10% (w/v) suspension, which was stirred until homogeneous. A CaCl2 crosslinking bath was prepared at 100 mM (0.1 mol L−1; Macklin, China). In addition, a 1% (w/v) chitosan solution (Aladdin, China) was prepared by dissolving chitosan in 1% HCl until clear. A gas-assisted coaxial needle microfluidic setup was used to generate droplets: the alginate suspension was loaded into a syringe and delivered through the inner channel of a vertically fixed coaxial needle, while nitrogen served as the shearing gas in the outer phase. Droplets formed at the needle tip and fell into the CaCl2 bath positioned below to allow immediate ionic crosslinking of alginate droplets into microspheres. After gelation and curing in the CaCl2 solution, the calcium alginate microspheres were harvested and rinsed with deionized water, then transferred into the 1% w/v chitosan solution and coated by shaking at 100 rpm at 37 °C for 0.5h. The microspheres were subsequently collected and rinsed with deionized water to eliminate residual surface solution, yielding chitosan-coated Orally Piezoelectric Microspheres (BTO@ALN@HMs/CHI). Microsphere morphology was analyzed using SEM (Hitachi, Japan), and hydrodynamic size distribution and zeta potential were measured using a particle size/zeta potential analyzer (Malvern Panalytical, UK). Quantification of ALN loading content. The conjugation/loading content of alendronate sodium (ALN) on BTO nanoparticles was quantified using a ninhydrin-based colorimetric assay. Briefly, ALN standard solutions with different concentrations were prepared in deionized water. Each ALN standard solution was mixed with ninhydrin reagent and pyridine solution, followed by heating at 100 °C for 2 h to generate a colored reaction product. After cooling to room temperature, the absorbance was measured at 565 nm using a UV spectrophotometer (Shimadzu, UV-1900i) and an ALN standard calibration curve was established. To quantify unconjugated ALN, the supernatant and washing solutions collected after the ALN grafting reaction were reacted with ninhydrin under the same conditions. The amount of free ALN was calculated according to the standard curve. The ALN conjugation/loading content was calculated using the following equation:ALN loading content (%) = [(initial ALN amount − free ALN amount in supernatant and washing solutions)/mass of BTO@ALN nanoparticles] × 100%.The ALN loading content of BTO@ALN was determined to be 6.83 wt%, which was used for calculating the ALN-equivalent dose in subsequent in vivo experiments. In Vitro Degradation of Orally Piezoelectric Microspheres under Simulated Gastrointestinal Conditions:Microspheres from different groups were removed from PBS buffer (Biosharp, China) and sequentially incubated in SGF (Mimic Fluid, China) for 2 h, SIF (Mimic Fluid, China) for 4 h, and then transferred to simulated colonic fluid (SCF; Mimic Fluid, China) for further incubation until complete degradation. At each stage, microspheres were imaged under an optical microscope (Olympus, Japan) to record changes in particle size and morphology. At 15 min and 1 h, microspheres were collected, lyophilized, and weighed to determine mass changes. The corresponding incubation media were also collected, and released BTO@ALN nanoparticles were recovered by centrifugation, lyophilized, and weighed. Lyophilization was performed using a freeze dryer (Labconco, USA). The cumulative release fraction was calculated as the mass of recovered, lyophilized BTO@ALN divided by the initial BTO@ALN loading amount. Intestinal-Targeted Delivery of Orally Piezoelectric Microspheres and Evaluation of the Bone-Targeting Capability of BTO@ALN: For fluorescence labeling during BTO@ALN preparation, a fluorescent silane reagent was introduced during the silanization step on the BaTiO3–OH surface. Specifically, during grafting of COOH–PEG–silane (Aladdin, China) onto BaTiO3–OH, Cy5–PEG–silane (Kaixin Bio, Xian, China) was co-added at a molar ratio of COOH–PEG–silane:Cy5–PEG–silane = 9:1; the remaining procedures were identical to those described in Section 4.1, yielding Cy5-labeled BaTiO3–PEG–ALN (denoted BTO@ALN–Cy5). BTO@ALN–Cy5 was then encapsulated into Orally Piezoelectric Microspheres following the established protocol for subsequent in vivo imaging. OVX mice were randomly assigned to groups and administered Orally Piezoelectric Microspheres by oral gavage at 0.1 mg BTO@ALN nanoparticle equivalent/10 g body weight. The gavage suspension was prepared in sterile PBS, and the administration volume was calculated according to body weight (10 mL kg−1). In vivo fluorescence imaging was performed after post-gavage (PerkinElmer, USA). Imaging parameters were set for the Cy5 channel and kept constant across all time points. For ex vivo femoral imaging, another cohort of OVX mice was euthanized at 1, 3, 5, 7, and 14 d after gavage. Femora were harvested, briefly rinsed with PBS to remove residual surface fluid, and imaged ex vivo using the same imaging system with identical channel settings. For intestinal distribution studies, fluorescein isothiocyanate (FITC; Beyotime, China) was added to the alginate solution at 1% (w/v) during microsphere fabrication, with all other procedures unchanged, to obtain FITC-labeled Orally Piezoelectric Microspheres. Mice were fasted for 12 h prior to gavage and then administered FITC-labeled microspheres at 10 mL kg−1. Mice were euthanized at 1, 2, 4, and 8 h after gavage, and the intact gastrointestinal tract was collected and imaged ex vivo using the imaging system(Luciferin-9500, Yuyanbio). Imaging parameters were set for the FITC channel and kept constant across samples. Cell Isolation and Culture:C57BL/6 mice were euthanized and all subsequent procedures were performed aseptically in a biosafety cabinet. After whole-body sterilization with 75% ethanol, the skin was incised and bilateral femora and tibiae were harvested. Muscle and connective tissues were stripped, and then transferred the bones to pre-chilled sterile PBS containing 1% penicillin–streptomycin. The bone marrow cavity was lavaged repeatedly with prewarmed culture medium via suitable syringe. The collected marrow suspension was followed by centrifugation at 300 × g for 5 min. The supernatant was discarded and the pellet resuspended in fresh medium for cell counting. The resulting bone marrow–derived single-cell suspension was used for the isolation of BMSCs and BMDMs. BMSCs were obtained by adherence–based selection. Bone marrow cells were cultured in BMSC culture medium, Fetal Bovine Serum, FBS (BaiDi Biotechnology Co., Ltd.(BDBIO)). After 24 h, the medium was refreshed, and then every 2-3 days. When confluence reached 85%, cells were detached using 0.25% Trypsin (Gibco, USA). BMDMs were generated by M-CSF induced differentiation. Bone marrow cells from the same preparation were cultured in BMDM induction medium (M-CSF at 25 ng mL−1, novoprotein, Jiangsu, China). Fresh medium containing M-CSF was provided on day 3. After 5–7 d, cells exhibited a uniform adherent macrophage-like morphology and were designated as BMDMs. Cells were gently rinsed with cold PBS and harvested using a cell scraper for subsequent polarization or functional assays. Prior to formal experiments, cells were characterized by flow cytometry for surface marker expression to confirm cell identity and purity. The Caco-2 cell line (ATCC HTB-37) was provided by Chongqing Medical University and stored in liquid nitrogen using Cryotubes (EIKON HEALTH). Direct contact co-culture with microspheres: For direct-contact co-culture experiments, Caco-2 cells or BMSCs were first seeded into culture plates at the indicated density and allowed to adhere overnight. Sterilized BTO@ALN@HMs/CHI microspheres were then gently resuspended in complete culture medium and added to the adherent cells at the predetermined amount. The microspheres naturally settled onto the cell layer, allowing direct contact between the particulate microspheres and adherent cells.During medium replacement, the microspheres were collected using a sterile 100 μm cell strainer (Biosharp, BS-100-XBS), gently rinsed with fresh culture medium, and then resuspended in new complete medium before being returned to the corresponding wells. This procedure minimized microsphere loss during medium changes and helped maintain a microsphere amount comparable to the initial seeding amount throughout the co-culture period. Biocompatibility Assays: To evaluate the cytocompatibility and potential cytotoxicity of Orally Piezoelectric Microspheres, direct co-culture experiments were conducted using Caco-2 cells and primary BMSCs. Cells were seeded into plates at standard densities and incubated with sterilized Orally Piezoelectric Microspheres at 100 μg mL−1. After 1, 3, and 5 d of co-culture, live/dead working solution (Calcein-AM/PI kit; Beyotime, China) was added and incubated for approximately 15 min in the dark. Confocal laser scanning microscope (Leica SP8, Germany) was used to record green (live) and red (dead) signals. CCK-8: Cells were co-cultured with Orally Piezoelectric Microspheres for 24 and 72 h. And then CCK-8 reagent (Beyotime, China) was added incubated. Absorbance at 450 nm was measured via a microplate reader (Thermo Multiskan, USA) to quantify metabolic activity and cell proliferation. Cytotoxicity (LDH): After 24 and 72 h of co-culture, culture supernatants were collected and analyzed via an LDH cytotoxicity assay kit (Beyotime, China). Cytoskeleton and nucleus staining: To assess cell morphology, adhesion, and spreading in the presence of the materials, cells were fixed with 4% paraformaldehyde for approximately 30 min, and permeabilized with 0.1% Triton X-100 for 15 min after 1, 3, and 5 d of co-culture. After PBS washing, F-actin was stained using fluorescent phalloidin (Beyotime, China), and DAPI staining (Beyotime, China) was applied to visualize nuclei. Fluorescence signal was collected by confocal microscopy for comparison of cell morphology, stress fiber organization, and spreading area among groups. In Vitro Osteoclastogenic and Osteogenic Assays: Osteoclastogenesis was induced by supplementing BMDM culture medium with RANKL (50 μg mL−1; R&D Systems, USA). Osteogenic differentiation was induced by culturing BMSCs in osteogenic induction medium (Procell, Wuhan, China). Culture medium was refreshed every 2 d. Osteoclast induction was continued for 5–7 d until multinucleated osteoclast-like cells were observed, followed by TRAP staining (Solarbio, Beijing, China). For osteogenesis, cells were induced for 7–14 d for early osteogenic assessments (ALP staining; SCR004, Sigma-Aldrich, USA), and for approximately 21 d for Alizarin Red S staining (C0148S, Beyotime, China) to evaluate mineralized nodule formation. In cell-based assays, BTO@ALN nanoparticles were co-cultured with cells at 50 μg mL−1, and cyclic stretching was applied using a cell stretching system (Cell Tank, Surface Power, Hangzhou, China) (1% strain, 1 Hz, sinusoidal waveform, 2 h day−1) to trigger piezoelectric activation. After the indicated treatments, protein expression/localization and gene transcription were analyzed by immunofluorescence staining and qPCR, respectively. Immunofluorescence staining: Cells were fixed with 4% paraformaldehyde and then permeabilized with 0.1% Triton X-100 (Beyotime, China) for 15 min. After blocking with 1% BSA (Solarbio, Beijing, China) for 0.5h, cells were incubated with primary antibodies against CTSK (ab187647, Abcam), TRAP (ab2721, Abcam), RUNX2 (ab192256, Abcam), and ATF4 (ab216839, Abcam) at 4 °C overnight. And then cells were incubated with the corresponding fluorophore-conjugated secondary antibodies (Proteintech, Wuhan, China) for 60 min. Cells were then stained with fluorescent phalloidin (C220S, Beyotime, China), followed by nuclear counterstaining with DAPI (1:500, Beyotime, China). qPCR: Total RNA was extracted using the RNeasy Plus Mini Kit (QIAGEN; Cat. No. 74134). RNA concentration and purity were determined by spectrophotometry (A260/A280 approximately 1.8–2.1). Reverse transcription was performed using PrimeScript RT reagent Kit (Takara, Japan). Equal amounts of RNA were treated with gDNA Eraser and then reverse-transcribed to cDNA. Real-time PCR was conducted using TB Green® Premix Ex Taq™ II (Takara, Japan). Reaction mixtures were prepared according to the manufacturer's protocol (TB Green mix, forward/reverse primers, and cDNA template). Relative expression was normalized to GAPDH and calculated using the 2^−ΔΔCt method. Animal Model Establishment: All animal experiments were conducted in accordance with the guidelines of the Animal Ethics Committee of the First Affiliated Hospital of Chongqing Medical University (Approval No. 2025-0343). Eight-week-old female C57BL/6 mice were randomly assigned to experimental groups. Mice were anesthetized with pentobarbital sodium (50 mg kg−1). After anesthesia was confirmed, mice were placed in the prone position, and the dorsolumbar region was shaved and disinfected. Small bilateral dorsolumbar incisions were made, and the abdominal cavity was accessed by blunt dissection. The ovaries and oviducts were exposed, the ovarian vessels and oviducts were ligated and transected, and bilateral ovariectomy was performed to establish the OVX osteoporosis model. In sham-operated mice, the ovaries were exposed but not removed. Incisions were closed in layers, and mice were kept warm until recovery from anesthesia and received standard postoperative care. Interventions were initiated 2 weeks post-surgery. Mice in the treatment groups were administered Orally Piezoelectric Microspheres via oral gavage weekly at 0.1 mg BTO@ALN nanoparticle equivalent/10 g body weight. The gavage suspension was prepared in sterile PBS, and the administration volume was set at 10 mL kg−1 according to body weight. For the exercise intervention, mice were subjected to treadmill running (SA101, SansBio, Jiangsu, China) at 0° incline and 5 m min−1 for 20 min per session, three times per week. At the end of the intervention period, mice were euthanized and samples were collected for subsequent analyses. Micro-CT Analysis: At week 10, mice were euthanized and femora were harvested. The knee joint was opened using ophthalmic scissors, the tibia was removed, and the femur was retained. Surrounding soft tissues, including muscle and connective tissue, were carefully removed. Femora were scanned via micro-CT (SkyScan, Bruker, Belgium) under standardized acquisition and reconstruction settings (voxel size, tube voltage/current, filter, exposure time, rotation step, and reconstruction algorithm) applied consistently across groups. Three-dimensional reconstructions were generated, and quantitative analysis was performed for the distal femoral trabecular region using a predefined region of interest (ROI). The primary morphometric parameters included BV/TV, BMD, Tb.N, and Ct.Th. Mechanical Testing and Piezoelectric Characterization: After euthanasia, femora were harvested and surrounding soft tissues were carefully removed. Samples were fixed in 70% ethanol at 4 °C for 48 h, followed by graded ethanol dehydration. After dehydration, specimens were infiltrated and embedded in epoxy resin (Sigma-Aldrich, USA) according to the manufacturer's protocol, followed by polymerization and curing. The embedded femora were sectioned using a low-speed diamond saw and machined into regular blocks, followed by sequential grinding and fine polishing to obtain parallel, flat end surfaces. Electrodes were subsequently applied uniformly to both end faces of each specimen and allowed to dry/cure prior to measurement. Piezoelectric coefficients (d33) were measured by placing each specimen between the probes of a Berlincourt-type d33 m (ZJ-3AN; Institute of Acoustics, China) under a constant preload. Measurements were repeated at multiple locations per specimen and across multiple specimens per group, and the mean d33 value was calculated. Specimen dimensions, surface conditions, electrode preparation, and testing parameters were kept consistent across groups.Three-point bending tests were performed (Instron 5943; Instron, USA). Femoral length was measured prior to testing, and specimens were kept hydrated throughout the test. Each femur was positioned on two lower supports with the loading nose centered at the midshaft, and compression was applied at a constant crosshead speed until fracture. Load–displacement curves were recorded for subsequent calculation of mechanical parameters. ELISA for Serum Bone Turnover Markers: Blood was collected prior to euthanasia under anesthesia. After adequate anesthesia and appropriate positioning, blood (approximately 0.3 mL per mouse) was collected from the retro-orbital venous plexus using a non-anticoagulant capillary tube inserted at the medial canthus and allowed to clot for 1 h. After clotting, samples were centrifuged at 4 °C (3000 rpm, 15 min). The supernatant serum was aspirated and aliquoted without disturbing the clot. Serum CTX-1 (EELM3023; Elabscience, Wuhan, China) and PINP (EELM023; Elabscience, Wuhan, China) levels were quantified using ELISA kits. Briefly, standards, blanks, and samples were incubated with the detection antibody/enzyme conjugate. Absorbance was measured using a microplate reader (Molecular Devices, USA) at the wavelength specified by the kit, and analyte concentrations were calculated from standard curves. All measurements were performed using the same kit lot with standardized operating conditions throughout. Histological and Immunostaining Analyses: Femora were harvested, and surrounding soft tissues were removed, followed by fixation in 4% paraformaldehyde (PFA; Sigma-Aldrich, USA) for 48 h. After rinsing with PBS, samples were decalcified in 10% EDTA(Solarbio, Beijing, China) at 37 °C, and the decalcification solution was refreshed every 2 d until complete decalcification was confirmed by X-ray. Specimens were then processed by routine dehydration, xylene clearing, and paraffin embedding (Paraplast; Leica, Germany). Serial sections (5 μm) were cut, dried, and used for subsequent staining and immunostaining analyses. H&E staining: paraffin sections were deparaffinized and rehydrated, stained with hematoxylin, differentiated and blued, counterstained with eosin, dehydrated, cleared, and mounted using neutral resin (Solarbio, Beijing, China), following standard procedures. Masson's trichrome staining: Masson staining was performed using a commercial kit (Solarbio, Beijing, China) to assess collagen fibers and bone matrix morphology and distribution. TRAP staining: osteoclast activity was evaluated using a TRAP staining kit (Sigma-Aldrich, USA). Color development was performed in a tartrate-containing reaction system, followed by brief hematoxylin counterstaining and mounting. Immunohistochemistry (IHC): after deparaffinization and rehydration, antigen retrieval was performed in sodium citrate buffer (Beyotime, China). After cooling, endogenous peroxidase activity was quenched with 3% H2O2 (solarbio, Beijing, China). Sections were blocked with 5% BSA (Solarbio, Beijing, China), then incubated with a primary antibody against COL1A1 (ab316222, Abcam). And then sections were incubated with secondary antibody (ZSGB-BIO, China), developed with DAB substrate, counterstained with hematoxylin, dehydrated, cleared, and mounted. Double immunofluorescence staining: after deparaffinization, rehydration, and antigen retrieval, sections were blocked and incubated with primary antibodies raised in different host species, including CTSK (ab187647, Abcam), MMP-9 (ab58803, Abcam), OPN (sc-21742, Santa Cruz), and OCN (ab93876, Abcam). And then sections were incubated in the dark with the corresponding fluorophore-conjugated secondary antibodies. Nuclei were counterstained with DAPI (Beyotime, Shanghai, China), and sections were mounted. All Fluorescence signal was acquired using a confocal microscope (Olympus, Japan), and imaging parameters were kept identical across groups. Sample size and replicate definition: For all animal experiments, n represents the number of independent animals and was considered the number of biological replicates. The in vivo therapeutic efficacy study included n = 5 independent mice per group. The mechanically induced femoral output voltage measurement in Fig. 3H included n = 3 independent mice per group at each time point. Femoral Ba quantification and Ba–output correlation analyses in Figs. S5 and S6 included n = 3 independent mice per group at each time point. Hemocompatibility, intestinal safety, dose-screening, long-term biosafety, and femoral d33 measurements shown in Fig. S8–S10 and S12 included n = 3 independent mice per group unless otherwise stated. For histological, immunohistochemical, and immunofluorescence analyses, each animal was considered one biological replicate. For each animal, three randomly selected microscopic fields from comparable anatomical regions were analyzed as technical replicates and averaged to obtain one animal-level value. Repeated ROI measurements, repeated d33 readings, repeated mechanically induced output voltage recordings, and repeated loading cycles from the same femur were treated as technical replicates. Unless otherwise stated, technical measurements were performed in triplicate and averaged before statistical analysis. Technical replicates were not used as independent biological samples. RNA-seq and Bioinformatics Analysis: Femoral tissues were harvested, and total RNA was immediately extracted and purified using TRIzol reagent (Thermo Fisher Scientific, USA). After assessing RNA integrity and purity, and transcriptome sequencing was performed by a commercial service provider (OEbiotech, Shanghai, China) on an Illumina sequencing platform. Raw sequencing data were processed through quality control, read alignment, and normalization to generate a gene expression matrix. DEGs were identified using the criteria of false discovery rate (FDR) ≤ 0.05 and |log2 fold change| > 1. In addition, gene–gene interaction networks were constructed using GeneMANIA to explore functional associations among genes. GSEA was performed to evaluate pathway-level transcriptomic alterations between groups. Data Analysis and Statistics:Data are presented as the mean ± standard deviation (SD). Statistical analyses were conducted using GraphPad Prism 9.5.1 and Origin 2021. Comparisons between two groups were performed using a two-tailed Student's t-test. For multiple-group comparisons, ANOVA was applied. Differences were considered statistically significant at P < 0.05. CRediT authorship contribution statement Wang Han: Writing – original draft, Validation, Formal analysis, Data curation, Conceptualization. Jing Luo: Formal analysis, Data curation, Conceptualization. Yingkun Xu: Formal analysis, Data curation, Conceptualization. Xu He: Investigation. Xinhe Li: Methodology. Jindong Tan: Software. Zijie Wang: Visualization. Yi He: Methodology. Peng Zou: Software. Yi Peng: Data curation. Xiaoyu Han: Resources, Project administration, Funding acquisition. Dingqun Bai: Supervision, Resources, Project administration, Funding acquisition. Jieliang Shen: Supervision, Resources, Project administration, Funding acquisition. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Ethics approval and consent to participate All animal experiments were conducted in accordance with the guidelines of the Animal Ethics Committee of the First Affiliated Hospital of Chongqing Medical University (Approval No. 2025-0343). Acknowledgements This work was supported in part by the National Natural Science Foundation of China (No. 82372572 and 82572940), the Natural Science Foundation of Chongqing Science and Technology Commission (No. CSTB2025NSCQ-GPX1126), Chongqing Municipal Education Commission Scientific and Technological Research Program for Young Scholars (KJQN202400449), Chongqing Medical Scientific Research Project (Joint Project of Chongqing Health Commission and Science and Technology Bureau) (2026KFXM044), Research and Innovation Leading Team Program of The Affiliated Bishan Hospital, Chongqing Medical University (BYKY-CX2024002), Chongqing Municipal Science and Health Joint Youth (Physician) Scientist Program (No. 2026QNKX003), The Postdoctoral Fellowship Program and China Postdoctoral Science Foundation (No.BX20240454, 2024M763902), Chongqing Natural Science Foundation Innovation and Development Joint Fund(key point) (No.CSTB2023NSCQ-LZX0016), Chongqing Science and Health Joint Medical Research Key Project (No. 2024ZDXM005), Doctoral Research Innovation Project of the First Clinical College of Chongqing Medical University(CYYY-BSCX202512). Footnotes Peer review under the responsibility of editorial board of Bioactive Materials. Supplementary data to this article can be found online at https://doi.org/10.1016/j.bioactmat.2026.05.051. Contributor Information Wang Han, Email: 2495791320@qq.com. Jing Luo, Email: LuoJing_08@126.com. Yingkun Xu, Email: 202462000632@email.sdu.edu.cn. Xu He, Email: hexu1009@163.com. Xinhe Li, Email: lixinhehard@163.com. Jindong Tan, Email: 1019685681@qq.com. Zijie Wang, Email: 18090851129@163.com. Yi He, Email: briona@163.com. Peng Zou, Email: rsqzpzz@163.com. Yi Peng, Email: nczxyypy@163.com. Xiaoyu Han, Email: xiaoyuhan@hospital.cqmu.edu.cn. Dingqun Bai, Email: baidingqun@hospital.cqmu.edu.cn. Jieliang Shen, Email: 204498@hospital.cqmu.edu.cn. Supplementary data The following is the Supplementary data to this article. Multimedia component 1 References Associated Data Supplementary Materials Multimedia component 1