Membrane-coated Prussian blue nanocomplex with enhanced enzyme-mimic activity alleviates senile osteoporosis by reducing osteoprogenitor cell senescence and macrophage inflammation

Abstract

Senile osteoporosis (SOP) is a chronic age-related skeletal disorder characterized by progressive bone loss and high fracture risk, with pathogenesis linked to excessive reactive oxygen species (ROS), osteoprogenitor cell senescence, and macrophage inflammation. Prussian Blue (PB) nanozymes show antioxidative potential for SOP but are limited by poor targeting, short circulation time, and unclear mechanisms, restricting clinical translation. Herein, we fabricated macrophage membrane-camouflaged citrate-modified PB (M@CPB) nanozymes for targeted SOP therapy and elucidated the underlying mechanism. Macrophage membrane coating prolonged the blood half-life of M@CPB by 1.47-fold and enhanced its bone tissue accumulation by 3.09-fold relative to CPB. In vitro, M@CPB inhibited O₂⁻ by 42.17% and scavenged H₂O₂ by 22.38%, reduced osteoprogenitor cell senescence, promoted M1-to-M2 macrophage polarization, and decreased the secretion of receptor activator of nuclear factor-κB ligand (RANKL) and tumor necrosis factor-α (TNF-α) to 47.41% and 53.10% of the model group, respectively. In vivo, M@CPB increased bone volume/tissue volume ratio by 3.83-fold relative to SOP mice, effectively ameliorating bone loss, motor dysfunction, and depressive-like behaviors. Collectively, M@CPB mitigated SOP by intervening in the ROS-senescence-inflammation axis and balancing bone remodeling, providing an innovative therapeutic strategy for SOP and potentially other age-related disorders.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12951-026-04771-x.

Introduction

With the rapid aging of the global population, senile osteoporosis (SOP) has become a major age-related skeletal disease worldwide characterized by bone loss and microstructural destruction. The high disability rate of SOP substantially affects the quality of life and increases the medical burden of the elderly population [1, 2]. Current clinical treatments for SOP primarily comprise anti-resorptive agents (e.g., bisphosphonates and denosumab) and osteoanabolic therapies (e.g., teriparatide). However, the efficacy of these treatments is often limited, and their long-term administration is associated with considerable adverse effects [1]. Recently, numerous studies have demonstrated the pathogenesis of SOP is linked to excessive reactive oxygen species (ROS) accumulation at osteoporotic sites [3]. As individuals age, a significant reduction in superoxide dismutase (SOD) and catalase (CAT) activities in senescent cells exacerbates the imbalance between oxidative and antioxidative processes, increasing ROS accumulation [4]. As a critical element in cellular senescence and inflammatory progression, ROS facilitate bone loss in the elderly population. Moreover, the extracellular matrix remodeling driven by senescent osteoprogenitor cells can reduce newly formed bone tissue and halt osteogenesis [5]. Additionally, senescent macrophages exhibiting increased sensitivity and responsiveness to inflammatory signals facilitate chronic inflammation and the persistent high expression of senescence-associated secretory phenotype (SASP) factors, exacerbating bone resorption [4, 6]. Consequently, targeting this ROS-senescence-inflammation axis to reverse cellular senescence and mitigate chronic inflammation is an essential objective for SOP treatment.

Antioxidants are regarded as the most direct and safe option for effectively resisting oxidative stress and treating SOP among the elderly [7–9]. However, natural antioxidant enzymes suffer from poor stability, and short‑term therapeutic effects. In contrast, nanozyme systems have emerged as a promising alternative to natural enzymes due to their stability and sustained catalytic activity. Among these nanozymes, Prussian Blue (PB)-based nanozymes, derived from the Food and Drug Administration (FDA)-approved PB material, exhibit robust SOD-and CAT-like activities [10]. Their superior ROS-scavenging capacity effectively mitigated osteoblast senescence and restored osteogenic potential [11, 12]. Notably, citrate-modified Prussian Blue nanozymes (CPB), characterized by small particle size and high specific surface area, offer more catalytically active sites, enhancing catalytic activity [13]. However, free Prussian Blue nanozymes are limited by inadequate targeting capability and limited ability to eliminate SASP factors. Furthermore, the underlying molecular mechanisms of PB nanozymes for SOP are poorly understood. Therefore, developing a safe and effective strategy for the targeted delivery of PB nanozymes to SOP sites while elucidating their mechanisms remains important.

Coating nanoparticles with cell membranes has recently been widely utilized to enhance targeting performance, prolong circulation half-life, and improve the clearance of cytokines by utilizing the intrinsic functions of source cells [14]. The recruitment of macrophages to the inflamed site is a highly specific process dependent on surface proteins, including integrin α4, integrin β1, and CCR2 [15, 16]. Additionally, cell membranes provide diverse binding sites that enable the scavenging of multiple hazardous substances, making it superior to single-target antibodies. For instance, macrophage-derived cell membranes can bind to and neutralize RANKL as well as various proinflammatory cytokines [6], thereby alleviating inflammation. Moreover, the “don’t eat me” proteins on the macrophage surface, such as CD47, can effectively evade immune phagocytosis and prolong the circulating half-life of nanomaterials [17]. Among these strategies, macrophage membrane coating offers a promising approach to target PB nanozymes to SOP sites and enhance its ability to alleviate SOP.

To improve the therapeutic efficacy of PB nanozyme for SOP and elucidate its underlying mechanisms, we developed a novel targeted nanosystem called M@CPB nanozymes, which consists of citrate-modified Prussian Blue nanozymes coated with macrophage membranes. This design achieves extended nanodrug half-life and enrichment of PB NPs at the osteoporotic sites, thereby exerting significant antioxidant activities. Modulation of redox balance effectively reverses osteoprogenitor cell senescence and enhances bone formation. Furthermore, attenuation of ROS also suppresses inflammation in senescent macrophages and limits bone resorption [6]. Overall, M@CPB can effectively inhibit SOP formation and progression, address a key pathological driver of age-related bone loss, and therefore provide a promising strategy for SOP treatment (Fig. 1).

Fig. 1.

Schematic illustration of the preparation of M@CPB and its therapeutic strategy for senile osteoporosis (created with Figdraw.com; permission No. TWRYAb891b). (A) Fabrication of M@CPB via RAW264.7 membrane extraction, CPB nanoparticle synthesis, and membrane coating by stirring and centrifugation. (B) Therapeutic effects of M@CPB on senile osteoporosis via regulation of osteoclastogenesis and osteogenesis, restoring bone homeostasis. (C) Anti-oxidative effects via alleviation of mitochondrial dysfunction and ROS reduction. (D) Anti-senescence effects via modulation of p16/p21 expression and release from cell cycle arrest, alleviating osteoprogenitor senescence and restoring function. (E) Anti-inflammatory effects via macrophage polarization from M1 to M2 and regulation of TNF-α, IL-6, and IL-10.

Materials and methods

Cells and animals

Osteoprogenitor cells (MC3T3-E1) and mouse macrophage cells (RAW264.7) were procured from Wuhan Procell Life Science & Technology Co., Ltd. Four-month-old male senescence-accelerated mouse prone 6 (SAMP6) mice and age-matched regular senescence-accelerated mouse resistant 1 (SAMR1) mice, which served as the control strain for senescence-accelerated mice, were obtained from Peking University Health Science Center. The animal experimental protocol for this study was approved by the Institutional Animal Care and Use Committee (IACUC) of the Second Xiangya Hospital of Central South University (Ethical Approval No. 20240989), and all animal experiments were conducted in accordance with the institutional guidelines for animal care and use.

Preparation and characterization of M@CPB NPs

Isolation of macrophage membranes (MΦm)

MΦm were isolated from RAW264.7 cells by suspending the cells in Tris-Mg buffer (pH 7.4; 10 mM Tris, 1 mM MgCl₂), lysing overnight at 4 °C, adding sucrose to a concentration of 0.25 M, and subjecting the mixture to three freeze-thaw cycles (–80 °C/37 °C). After centrifugation at 2000 g for 10 min, the supernatant was centrifuged at 13,000 rpm for 30 min to pellet membranes, which were subsequently washed twice with cold 0.25 M sucrose, collected, centrifuged, and stored at − 80 °C.

Preparation of citrate-modified Prussian blue nanozymes (CPB)

CPB were synthesized according to established methods reported in the literature [12] with slight modifications: (1) Solution A was prepared by dissolving 54.1 mg of ferric chloride hexahydrate (FeCl₃·6H₂O) (Aladdin Biochemical Technology Co., Ltd., Shanghai, China) in 2 mL of deionized water, followed by dilution to 1 mM with 25 mM citric acid solution (Shanghai Reagent, Shanghai, China). The solution was heated to 60 °C. (2) Solution B was prepared by dissolving 85 mg of potassium ferrocyanide in 25 mM citric acid solution and diluting it to 1 mM, after which the solution was also heated to 60 °C. (3) Subsequently, solution B was added dropwise to solution A under continuous stirring at 60 °C for 30 min. The resulting product was collected by centrifugation, washed, and freeze-dried to obtain CPB NPs.

Preparation of macrophage membrane-coated Prussian Blue nanozymes (M@CPB)

We mixed 1 mL of CPB NPs suspension (1 mg/mL) with 0.1 mL of MΦm suspension (1 mg/mL). The mixture was ultrasonicated for 20 s and then stirred at 37 °C and 800 rpm for 1 h. The resulting precipitate was collected by centrifugation and washed twice with deionized water to obtain macrophage membrane-coated nanocomplexes (M@CPB).

Characterization of M@CPB

Transmission electron microscopy (TEM) images and particle size distributions of CPB and M@CPB were obtained using a JEM-2100 microscope (JEOL, 200 kV). Macrophage membranes (MΦm) and CPB were labeled with DiO (green fluorescence) and Ce6 (red fluorescence), respectively. The membrane-coating status of the nanocomplexes was visualized using confocal laser scanning microscopy (CLSM). The hydrodynamic particle size and zeta potential were measured using a Zetasizer Nano ZSP (Malvern Instruments). The stability of the nanoparticles in deionized water and DMEM containing 10% FBS was evaluated over 14 days. X-ray photoelectron spectroscopy (XPS) was performed by Puchuan Testing Co., Ltd. (Guangdong, China) to analyze CPB and M@CPB elemental compositions. Fourier-transform infrared spectroscopy (FT-IR) was conducted to identify the chemical structures of CPB and M@CPB. Membrane coating efficiency, protein copy number per nanoparticle, and membrane protein density were estimated based on bicinchoninic acid (BCA) assay and nanoparticle characterization data; batch-to-batch consistency was evaluated by coefficient of variation (CV).

Coomassie brilliant blue staining, western blot analysis, and quantitative membrane protein retention analysis

MΦm and M@CPB were lysed using protein lysis buffer and centrifuged at 12,000 rpm at 4 °C for 30 min. The supernatant was collected, and the protein concentration was measured using a BCA assay. Proteins from MΦm and M@CPB were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE). The gel was subsequently stained with Coomassie Brilliant Blue and then destained to visualize the total protein profile. For Western blot analysis, proteins were transferred to PVDF membranes and immunoblotted with antibodies against CCR2, CD47, RANK, TNFR2, and the membrane protein loading reference caveolin-1 to verify membrane protein retention after coating.

To quantitatively evaluate the retention of functional membrane proteins after coating, Western blot bands were analyzed by densitometry using ImageJ software. Band intensities were corrected by subtracting the local background signal. The expression level of each functional membrane protein was normalized to caveolin-1 in the same lane. Native macrophage membrane (MΦm) was set as 100%, and the relative retention rate of each protein in M@CPB was calculated using the following formula:

Three independently prepared batches were analyzed, and the data are presented as mean ± SD.

Assay for cytokine scavenging

Receptor activator of nuclear factor-κB ligand (RANKL, 200 pg/mL) or tumor necrosis factor-α (TNF-α, 500 pg/mL) was incubated with CPB (500 µg/mL), liposome membrane-coated CPB (Lip@CPB, 500 μg/mL), MΦm (50 µg/mL), or M@CPB (500 µg/mL) at 37 °C for 2 h. For CPB, Lip@CPB, and M@CPB, concentrations were normalized according to the mass of the CPB core, and all groups were adjusted to contain an equivalent CPB concentration of 500 µg/mL. After incubation, the mixtures were centrifuged at 14,000 rpm for 10 min, and the concentrations of residual RANKL and TNF-α in the supernatants were quantified using enzyme-linked immunosorbent assay (ELISA) kits.

Characterization of antioxidant enzyme-like activity of M@CPB

Assay for SOD activity: SOD-like activity of CPB at 0–400 µg/mL was measured using a total SOD activity kit (WST-8 method). The SOD activities of CPB (100 µg/mL), MΦm (10 µg/mL), and M@CPB (100 µg/mL) were compared. Absorbance (A) was recorded at 650 nm, and color changes were documented. The SOD-like activity was expressed as the inhibition rate relative to the control. Inhibition rate of O₂^– was calculated using the following formula:

Assay for CAT activity: To evaluate sustained CAT activity, CPB (100 µg/mL), MΦm (10 µg/mL), and M@CPB (100 µg/mL) were incubated with H₂O₂ (5 mM) at 37 °C for 30 min, and dissolved oxygen (DO) was monitored. The cycle was repeated four times. The samples were mixed with H₂O₂ (500 mM), and images were captured at 0 min and 2 h. The samples were also mixed with H₂O₂ (5 mM), and DO changes were monitored at 10 min using a portable DO meter. Decomposition rate of H₂O₂ was calculated using the following formula:

Assay for POD activity: CPB (2 µg/mL), MΦm (0.2 µg/mL), and M@CPB (2 µg/mL) were mixed with 100 mM H₂O₂ and 3,3′,5,5′-tetramethylbenzidine (TMB) substrate at 37 °C. The absorbance at 650 nm was continuously monitored for 10 min. For M@CPB, H₂O₂ concentrations ranging from 0 to 2.5 mM were tested, and after 10 min, absorbance was measured. The catalytic parameters were calculated using the Michaelis–Menten equation:

Biodistribution, pharmacokinetics, inflammatory targeting, and immune escape of M@CPB

Biodistribution assay

We intravenously injected 6-month-old male SAMP6 mice with CPB^Ce6 and M@CPB^Ce6 (Ce6, 2.5 mg/kg). After 2 h, the depilated mice were imaged using a small-animal IVIS in vivo imaging system. After 36 h, the mice were euthanized, and the upper and lower limb bones, lumbar spine, and major organs were collected. Fluorescence signals from these tissues were analyzed using the same imaging system. Bone-to-organ targeting indices were calculated as the fluorescence intensity ratio of bone to each organs.

Pharmacokinetic study

We collected retro-orbital blood from adult male Kunming white mice at 0, 1, 2, 4, 6, 8, 12, and 24 h post-injection after intravenous injection of CPB^Ce6 and M@CPB^Ce6 (Ce6, 2.5 mg/kg). Blood fluorescence intensity was measured utilizing a small-animal in vivo imaging system, and the circulation half-life (t_1/2) was calculated.

In vitro CCR2–CCL2 transendothelial transport, cellular uptake, and immune escape assays

CCR2–CCL2 transendothelial transport assay: Transwell chambers (0.4 μm pores) were seeded with mouse vascular endothelial cells (VECs) to form a confluent monolayer. CPB^Ce6, M@CPB^Ce6, CCR2-blocked M@CPB^Ce6 (CCR2i-M@CPB^Ce6), and Lip@CPB^Ce6 were added to the upper chamber at 10 µg/mL. Recombinant mouse CCL2 (50 ng/mL) was added to the lower chamber; medium without CCL2 served as the control. After incubation for 6 h at 37 °C, fluorescence intensities of Ce6 in the lower chambers were measured to quantify the efficiency of nanoparticle transport. CCR2i-M@CPB^Ce6 was prepared by incubating M@CPB^Ce6 with the selective CCR2 antagonist RS102895 (5 µM) at 37 °C in the dark for 30 min to functionally block CCR2-mediated CCL2 binding.

Cellular uptake assay: RAW264.7 and MC3T3-E1 cells were seeded into 24-well plates preloaded with cell-climbing slides and divided into Normal, H₂O₂-stimulated (200 µM, 24 h), and CCL2-treated (50 ng/mL, 24 h) groups. M@CPB^Ce6, CCR2i-M@CPB^Ce6, and Lip@CPB^Ce6 were added at a final concentration of 10 µg/mL and incubated for 4 h. The medium was then discarded, and the cells were washed thrice with PBS, fixed with 4% paraformaldehyde (PFA) for 30 min, and stained with DAPI. Subsequently, CLSM was utilized to detect fluorescence signals and assess nanoparticle cellular uptake efficiency.

Immune escape: RAW264.7 cells were seeded into 24-well plates preloaded with cell-climbing slides. After adhesion, cells were separately incubated with CPB^Ce6, Lip@CPB^Ce6, or M@CPB^Ce6 at a final concentration of 10 µg/mL for 0, 2, 4, and 6 h. The medium was discarded, the cells were washed thrice with PBS, fixed with 4% PFA for 30 min, and stained with DAPI. Subsequently, CLSM was utilized to detect intracellular fluorescence signals.

In vitro anti-senescence and osteogenic differentiation-promoting assays

Cellular senescence inhibition assays

Immunofluorescence staining: MC3T3-E1 and RAW264.7 cells were seeded into 24-well plates preloaded with cell-climbing slides at 1 × 10⁵ cells/well, divided into control, senescence, and drug treatment groups, including CPB (10 µg/mL), MΦm (1 µg/mL), and M@CPB (10 µg/mL). After incubation with nanoparticles for 2 h, cells in the senescence and treatment groups were stimulated with 400 µM H₂O₂ for 24 h, followed by fixation in 4% PFA for 30 min. Cells were incubated overnight at 4 °C with primary antibodies against γ-H2AX (1:200), p21 (1:200), and p16INK4a (1:200), and subsequently incubated with fluorescence-labeled secondary antibodies for 2 h. After DAPI counterstaining for 10 min, fluorescence images were photographed using a Nikon confocal microscope (Ti-E+A1MP, Japan). Additionally, a 5 mM N-acetylcysteine (NAC) positive control was included exclusively for p21 immunofluorescence in MC3T3-E1 cells to verify the involvement of ROS scavenging in the anti-senescence effect.

Senescence-associated β-Galactosidase (β-gal) staining: MC3T3-E1 and RAW264.7 cells were seeded and treated as described above. β-gal activity was evaluated utilizing a Cellular Senescence β-Galactosidase Staining Kit (C0602, Beyotime). Blue-stained cells, indicative of senescence, were counted for quantitative analysis.

Cell cycle analysis: We seeded MC3T3-E1 cells at 4 × 10⁵ cells/well in 6-well plates and treated them as described above. A cell cycle and apoptosis kit (C1052, Beyotime) was employed to analyze cell cycle distribution in the control, senescence, and drug-treated groups. The percentages of cells in G1, S, and G2 phases were calculated using FlowJo software.

Osteogenesis assay

Alkaline phosphatase (ALP) staining and activity assay: We seeded MC3T3-E1 cells at 1 × 10⁵ cells/well in 24-well plates and cultured them in osteogenic induction medium (10 mM β-glycerophosphate, 50 µg/mL ascorbic acid, 10 nM dexamethasone). Grouping and treatment protocols were the same as described above. The medium containing H₂O₂ and the corresponding treatments was replaced every 2–3 days. After 3 or 7 days, the cells were fixed with 4% PFA for 30 min and stained with BCIP/NBT (C3206, Beyotime) for 30 min, followed by microscopic observation. MC3T3-E1 cells were seeded at 4 × 10⁵ cells/well in 6-well plates and treated as described above. Cells were lysed using 0.5% Triton X-100 on ice. The supernatant was collected by centrifugation. The ALP activity was measured using an ALP assay kit (Nanjing Jiancheng). Total protein concentration was determined using a BCA assay, and ALP activity was normalized to the protein content. ALP activity was expressed as King units/gprot according to the manufacturer’s definition.

Alizarin red S (ARS) staining: We seeded MC3T3-E1 cells at 4 × 10⁵ cells per well in 6-well plates and treated them as above. After 14 or 21 days of osteogenic induction, the cells were fixed with 4% PFA for 30 min and stained with 0.1% ARS (OriCell^®, ALIR-10001) at room temperature for 30 min. After three washes with PBS, calcium nodules (orange-red) were observed and photographed under a microscope.

In vitro mitochondrial protection and oxidative stress resistance assays

Mitochondrial membrane potential detection assay

We evaluated the mitochondrial membrane potential using 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethyl-imidacarbocyanine (JC-1) dye and observed it under a fluorescence microscope (Leica, Germany). MC3T3-E1 and RAW264.7 cells were seeded onto cell slides in 24-well plates at 1 × 10⁵ cells per well, cultured overnight, and treated as described above. The cells were stained with JC-1 at 37 °C for 20 min after treatment. The supernatant was discarded, and the cells were washed twice with PBS. Fluorescence images were captured, and intensity was quantified with ImageJ software.

Mitochondrial analysis

TEM: To observe mitochondrial morphology, we seeded MC3T3-E1 and RAW264.7 cells at 1 × 10⁶ cells/well in 6-well plates, cultured overnight, and treated as described above. After treatments, the cells were harvested with a scraper and centrifuged at 1500 rpm for 10 min. The supernatant was discarded, and cells were fixed with 2.5% glutaraldehyde overnight at 4 °C. Samples were subsequently analyzed using Bio-TEM at the Scientific Compass Service Platform (Hangzhou Yanqu Information Technology Co., Ltd.).

ATP production assay: MC3T3-E1 and RAW264.7 cells at 1 × 10⁶ cells/well were seeded in 6-well plates, cultured overnight, and treated as above. Cells were lysed with buffer and centrifuged at 12,000 rpm at 4 °C for 5 min. The supernatant was collected, and ATP levels were measured using an ATP assay kit (Beyotime, S0026).

ROS scavenging assay

MC3T3-E1 and RAW264.7 cells were seeded at 1 × 10⁵ cells/well in 24-well plates and grouped as described above. After 2 h of drug pre-treatment, H₂O₂ was added to the model and treatment groups at 400 μM for 4 h. Subsequently, 10 µM DCFH-DA was added and incubated in the dark for 30 min. Fluorescence intensity was measured using CLSM. Cells were also seeded at 4 × 10⁵ cells per well in 6-well plates, treated similarly, fixed with 2% PFA, and analyzed using flow cytometry.

Cell viability assay

MTT assay: MC3T3-E1 cells were seeded in 96-well plates. With or without pretreatment with M@CPB (10 µg/mL) for 2 h, the cells were co-incubated with H₂O₂ at 0, 100, 200, and 400 µM for 24 or 72 h. Cell viability was evaluated using the MTT method to assess cellular resistance to oxidative stress in vitro. Cell viability was calculated using the following formula:

Cytotoxicity staining: MC3T3-E1 cells were seeded in 24-well plates. With or without 2-h pretreatment with M@CPB (10 µg/mL), the cells were co-incubated with H₂O₂ at 0, 100, 200, and 400 µM for 24 h. Calcein-AM and propidium iodide (PI) were subsequently added for live/dead staining.

In vitro anti-inflammatory and bone resorption inhibition assays

Evaluation of anti-inflammatory efficacy in vitro

Determination of macrophage polarization: RAW264.7 cells were seeded onto cell slides in 24-well plates at 1 × 10⁵ cells per well, following the same grouping and treatment protocols as above. Cells were incubated overnight at 4 °C with primary antibodies against CD80 (1:200) and CD206 (1:200), followed by incubation with fluorescent secondary antibodies for 2 h. After 10 min of DAPI staining, fluorescence images were captured using a Nikon confocal microscope.

Detection of cytokines: RANKL, TNF-α, interleukin (IL)-6, and IL-10 levels in the supernatants were measured with ELISA kits (Neobioscience Biotechnology, Shenzhen, China).

Cell co-culture and bone resorption detection assay

TRAP staining: MC3T3-E1 cells were seeded at 4 × 10⁵ per well in 6-well plates. Following 2 h of NP pre-incubation for the treatment group, both model and treatment cells were stimulated with 400 µM H₂O₂ for 24 h, whereas the control group received no intervention. The medium was subsequently replaced with α-MEM containing 10% FBS for 12 h. The supernatant was collected, centrifuged at 3000 rpm at 4 °C for 10 min, and utilized as conditioned medium. RAW264.7 cells were seeded at 1 × 10⁵ per well in 24-well plates. The conditioned medium was mixed 1:1 with DMEM containing 10% FBS and employed to treat RAW264.7 cells, which were refreshed daily. Tartrate-resistant acid phosphatase (TRAP) staining was performed after 5 days to assess osteoclast differentiation.

Flow cytometry assay for apoptosis: RAW264.7 cells were seeded at 4 × 10⁵ cells/well in 6-well plates. Following 2 h of NP pre-incubation for the treatment groups, both model and treatment cells were stimulated with 400 µM H₂O₂ for 24 h, whereas the control group received no intervention. The medium was replaced with 10% DMEM and incubated for 12 h. The medium was collected, centrifuged at 3000 rpm at 4 °C for 10 min, and the supernatant was utilized as RAW264.7-conditioned medium. The MC3T3-E1 cells were seeded similarly; the conditioned medium was mixed 1:1 with 10% MEMα and introduced to the cells for 24 h. Cells were harvested, stained with Annexin V-FITC and PI, incubated in the dark for 15 min, and assessed for apoptosis using flow cytometry.

In vivo anti-osteoporotic efficacy assay

Animal grouping and drug administration

Senescence-accelerated mice (SAM) are an inbred strain selected during the breeding of AKR/J mice at Kyoto University, Japan [18]. The SAMP6 mice exhibit decreased bone mineral density, reduced bone mass, and vascular calcification, making them a suitable model for studying SOP [19, 20]. SAMR1 mice served as the control group, and SAMP6 mice were randomly stratified into four groups: Model, CPB (2.5 mg/kg), MΦm (0.25 mg/kg), and M@CPB (2.5 mg/kg), with six mice per group. After 1 week of acclimation, drugs were administered via tail vein injection every two days. The mice were euthanized after 8 weeks.

Micro-CT analysis

The right femurs of each group were collected, fixed in 4% PFA for 24 h, and preserved in 75% ethanol. We performed micro-CT to assess changes in bone microstructure.

Bone tissue analysis and immunohistochemistry staining

We harvested the left femurs of the mice, fixed them in 4% PFA for 24 h, decalcified them in 10% ethylenediaminetetraacetic acid for 30 days, embedded them in paraffin, and sectioned them at 5 μm. Subsequently, we performed hematoxylin and eosin (H&E), Masson, and TRAP staining. Immunohistochemical staining was performed with anti-RANKL (1:200) and anti-TNF-α (1:200) antibodies. After drying and sealing with neutral resin, sections were observed and photographed under an inverted microscope, and subsequently analyzed using ImageJ software. For ROS detection, bone sections were stained with DHE. For immunofluorescence staining, bone sections were incubated overnight at 4 °C with primary antibodies against CD80 (1:200), CD206 (1:200), p16INK4a (1:200), p21 (1:200), γ-H2AX (1:200), RUNX2 (1:200), NGF (1:200), and NF-H/NF200 (1:200), and subsequently incubated with fluorescently labeled secondary antibodies for 2 h. After DAPI staining for 10 min, the sections were imaged using a Nikon confocal laser scanning microscope (Ti-E+A1MP, Japan).

Behavioral tests

Open-field test: The apparatus comprised gray acrylate walls measuring 40 × 40 × 40 cm, with one open side for video recording. A 15-s background was recorded first, and subsequently, the mouse was placed in the center. Its movement was recorded for 6 min, with the mouse kept in the same position for each trial.

Elevated plus maze test: The apparatus, made of polymethyl methacrylate, had two open (30 cm long, 5 cm wide) and two closed arms (30 cm long, 5 cm wide, 15 cm high) connected to a 5 × 5 cm central platform and was elevated 60 cm above ground. During the experiment, a 15-s background was recorded. Subsequently, a mouse was placed facing an open arm on the platform and video-recorded for 6 min, with the placement consistent for each mouse.

Pole climbing test: The apparatus was a 40 cm wooden pole inclined at 45°. During the test, the mouse was placed at the top end with its back facing downward, and the time taken to turn around and climb down was recorded. Mice were trained to climb until they could complete a single movement. During the experiment, a 15-s background was recorded, followed by a video of the climbing behavior. If the mouse remained still, recording stopped after 2 min.

Tail suspension test: The test apparatus had blue acrylic walls measuring 20 × 40 × 60 cm, with one open side for video recording. Two animals could be tested simultaneously, separated by a central light-shielding plate. During the experiment, each mouse’s tail was fixed with tape 1 cm from the tip and held steady 60 cm above the floor. Behavioral changes were recorded for 6 min.

Biomechanical test

We assessed the biomechanical properties of the left tibiae using a three-point bending test. We placed the left tibiae from each group on a material mechanical performance tester (uTS-100 N/1.6 kN) with a 5-mm span. The loading rate was set to 0.155 mm/s, and the load was gradually increased until the tibia fractured. The maximum load and break load were calculated from the instrument’s load-deflection curve.

Proteomics and metabolomics analysis

We collected the right upper limb bones from mice in the control, model, and M@CPB groups, snap-frozen them in liquid nitrogen for 15 min, and stored them at − 80 °C. The samples were subsequently sent to Hangzhou Jingjie Biotechnology Co., Ltd. for proteomic analysis. Serum samples from mice in the same groups were collected, snap-frozen in liquid nitrogen for 15 min, and stored at − 80 °C. The serum samples were subsequently sent to Hangzhou Jingjie Biotechnology Co., Ltd. for metabolomic detection.

In vitro biocompatibility and in vivo safety assays

HepG2, HK-2 cells, HUVECs, and MC3T3-E1 cells were seeded separately into 96-well plates. The cells were treated with M@CPB (10 µg/mL), with or without 400 µM H₂O₂-induced oxidative stress for 2 h, and subsequently incubated for 24 h. Cell viability was assessed using the MTT assay, with the formula:

Mouse plasma samples were mixed with PBS, thrombin, CPB (10 µg/mL), MΦm (1 µg/mL), or M@CPB (10 µg/mL) and incubated at 37 °C for 6 h. Absorbance was measured at 540 nm.

Red blood cells were co-cultured with different CPB and M@CPB concentrations at 37 °C for 6 h. After centrifugation, red blood cell morphology was observed under a microscope, and the absorbance of the supernatant at 540 nm was measured to assess hemolysis.

Zebrafish embryos were co-cultured with CPB (10 µg/mL), MΦm (1 µg/mL), and M@CPB (10 µg/mL) for 4 days. Hatching, heart rate, and body length were monitored daily to assess developmental toxicity.

H&E staining was employed to examine the histological morphology of the main organs. Blood samples were collected from mice for biochemical and hematological analyses.

Statistical analysis

GraphPad Prism software (version 9) was utilized for statistical analysis. All quantitative data are presented as mean ± standard deviation (SD). The exact sample size or number of independent experiments is indicated in the corresponding figure legends. Data normality was assessed using the Shapiro–Wilk test, and homogeneity of variance was assessed using the Brown–Forsythe test. For comparisons between two groups, a two-tailed unpaired Student’s t-test was used when data were normally distributed and variances were equal; Welch’s correction was applied when variances were unequal. For comparisons among three or more groups, ordinary one-way analysis of variance (ANOVA) followed by Tukey’s multiple-comparisons test was used when the assumptions of normality and equal variance were satisfied. When multiple treatment groups were compared with a single control or model group, Dunnett’s multiple-comparisons test was used. For datasets with unequal variances, Welch’s ANOVA followed by Dunnett’s T3 multiple-comparisons test was applied. For datasets that did not satisfy normality, the Kruskal–Wallis test followed by Dunn’s multiple-comparisons test was used. P < 0.05 was considered statistically significant. Significance was defined as follows: ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; and ****P < 0.0001.

Results and discussion

Preparation and characterization of M@CPB

CPB nanozymes were synthesized as previously reported [12] and subsequently coated with MΦm to construct M@CPB. TEM images showed successful synthesis of uniform cubic CPB (Fig. 2A). A distinct membrane structure was observed on M@CPB after membrane coating (Fig. 2B).

Fig. 2.

Preparation and characterization of M@CPB. (A, B) TEM images of CPB and M@CPB. (C) Particle size, zeta potential and PDI of CPB and M@CPB. (D) Stability of M@CPB in ultrapure water over 14 days. (E) SDS–PAGE analysis of membrane proteins in MΦm and M@CPB. (F) Western blot analysis of characteristic membrane proteins in MΦm and M@CPB. (G) Residual RANKL levels in the supernatants after incubation with CPB, Lip@CPB, MΦm and M@CPB (initial concentration: 200 pg/mL). (H) Residual TNF-α levels in the supernatants after incubation with CPB, Lip@CPB, MΦm and M@CPB (initial concentration: 500 pg/mL). (I) Schematic illustration of the SOD-mimetic activity of M@CPB. (J) O₂⁻ scavenging activity of CPB at different concentrations. (K) Representative absorbance spectra of O₂⁻ scavenging by CPB, MΦm and M@CPB. (L) Quantitative O₂⁻ inhibition rates of CPB, MΦm and M@CPB. (M) Schematic illustration of the CAT-mimetic activity of M@CPB. (N) Catalytic performance of CPB, MΦm and M@CPB upon repeated H₂O₂ addition. (O) Representative images of oxygen bubble generation by different samples at 0 and 120 min. (P) H₂O₂ decomposition rate of CPB, MΦm, and M@CPB within 10 min. (Q) Schematic illustration of the POD-mimetic activity of M@CPB. (R) Absorbance of oxidized TMB in CPB, MΦm and M@CPB groups. (S) Time-dependent absorbance changes of oxidized TMB over 10 min. (T) Michaelis–Menten curves of M@CPB at different H₂O₂ concentrations. (U) Schematic illustration of ROS scavenging by M@CPB through enzyme-mimetic activities (created using Figdraw.com; permission No. ATYURfe6fe). All quantitative data are presented as mean ± SD (n = 3)

The physicochemical properties of M@CPB were further characterized. EDS analysis confirmed the presence of the characteristic Fe element in both CPB and M@CPB. Notably, the nitrogen content increased from 14.89% in CPB to 23.30% following membrane coating, reflecting the successful incorporation of membrane-derived biomolecules onto the nanoparticle surface (Figs. S1A and B). Consistently, FTIR spectra revealed the characteristic C ≡ N (2085 cm^− 1) and Fe–C ≡ N (503 cm^− 1) absorption peaks of Prussian Blue nanoparticles, whereas an additional P = O absorption peak at 1080 cm⁻¹ emerged after membrane coating, further confirming the presence of MΦm on M@CPB (Fig. S1C). Fluorescence colocalization analysis using separately labeled CPB and MΦm showed extensive overlap between the two fluorescence signals, demonstrating efficient membrane encapsulation (Fig. S1D). Dynamic light scattering analysis revealed that M@CPB exhibited an average hydrodynamic diameter of 115.20 ± 2.24 nm, a zeta potential of − 31.50 ± 0.46 mV, and a PDI of 0.26 ± 0.05 (Fig. 2C). The colloidal stability of M@CPB was subsequently evaluated in ultrapure water and DMEM containing 10% FBS. Throughout the observation period, the nanoparticles generally maintained acceptable particle size distribution, negative surface charge, and PDI values, indicating satisfactory colloidal stability under storage and physiologically simulated conditions (Fig. 2D and Fig. S1E). To quantitatively evaluate membrane coating performance, coating efficiency and membrane protein loading were determined. The membrane coating efficiency reached 99.13 ± 0.10%, and each nanoparticle carried approximately 4777.72 ± 275.48 membrane protein molecules, corresponding to a protein density of 0.06 ± 0.002 copies/nm². Moreover, three independently prepared batches exhibited low coefficients of variation for particle size, zeta potential, and protein loading efficiency, all below 10%, demonstrating excellent fabrication reproducibility (Table S1). Together, these findings confirm the efficient, uniform, and highly reproducible coating of MΦm onto CPB nanozymes.

Because the biological functions of biomimetic nanoparticles largely depend on the preservation of membrane-associated proteins, the protein composition of M@CPB was further analyzed. SDS-PAGE showed that M@CPB retained a protein profile highly similar to that of native MΦm, suggesting successful transfer of membrane proteins during the coating process (Fig. 2E). Western blot analysis further confirmed the preservation of several functionally important proteins, including CCR2, CD47, RANK, and TNFR2 (Fig. 2F), which are associated with inflammatory homing, immune evasion, and cytokine sequestration [6, 16, 17]. To quantitatively assess the retention efficiency of these key membrane proteins after coating, densitometric analysis was performed using caveolin-1 as the membrane protein loading control. Compared with native MΦm, M@CPB retained 92.96 ± 6.71% of CCR2 and 93.35 ± 8.67% of CD47, supporting the preservation of proteins associated with inflammatory tropism and immune evasion. In addition, RANK and TNFR2 were retained at 92.46 ± 2.10% and 92.83 ± 1.98% (Fig. S1F and Table S2). These quantitative results demonstrate that M@CPB preserves key functional membrane proteins after coating, thereby supporting the proposed mechanisms of CCR2-mediated inflammatory tropism, CD47-mediated immune evasion, and cytokine scavenging. To further determine whether the retained membrane proteins remained biologically active, the cytokine-scavenging capability of M@CPB was evaluated using RANKL and TNF-α, two critical mediators of SOP progression. As expected, both MΦm and M@CPB efficiently depleted RANKL and TNF-α from solution, whereas CPB and liposome-coated CPB (Lip@CPB) only exhibited weak scavenging activity (Fig. 2G, H). Notably, M@CPB reduced residual RANKL and TNF-α levels to 20.87 ± 1.27% and 24.80 ± 1.34%, respectively, demonstrating strong cytokine-neutralizing capability. Collectively, these findings suggest that M@CPB inherits the cytokine-scavenging function of MΦm through the preservation of membrane-associated proteins.

The enzyme-mimicking activities of M@CPB were subsequently investigated. SOD-mimetic activity was first assessed by measuring superoxide anion scavenging (Fig. 2I). CPB exhibited concentration-dependent O₂⁻ scavenging activity (Fig. 2J), and M@CPB retained this catalytic capability after membrane coating (Fig. 2K). At an equivalent concentration of 100 µg/mL, CPB inhibited O₂⁻ by 37.70 ± 0.06%, whereas M@CPB achieved a significantly higher inhibition rate of 42.17 ± 0.20%. In contrast, MΦm alone exhibited only limited scavenging activity, with an inhibition rate of 11.49 ± 0.16% (Fig. 2L). These results indicate that membrane coating does not compromise the intrinsic SOD-mimetic activity of the nanozyme and may even slightly enhance its superoxide-scavenging capability.

Catalase (CAT)-mimetic activity was assessed by quantifying O₂-generating capability (Fig. 2M). CPB and M@CPB rapidly degraded H₂O₂ over 120 min, demonstrating sustained CAT-mimetic activity. Conversely, MΦm exhibited rapid H₂O₂ decomposition within the first 30 min, suggesting transient early-phase activity (Fig. 2N), which may be attributed to the co-isolation of intracellular peroxisomal membrane components during membrane extraction [21]. O₂ bubble visualization served as a direct readout of CAT-mimetic activity [7]. Figure 2O shows that CPB and M@CPB generated significantly more bubbles than MΦm at 120 min, with no significant difference between them. Figure 2P shows that membrane coating contributed to the early CAT-like activity of M@CPB. The H₂O₂ decomposition rates of M@CPB, MΦm, and CPB were 22.38 ± 0.88%, 34.13 ± 1.84%, and 9.79 ± 3.69%, respectively.

Additionally, TMB was utilized to evaluate the POD-mimetic activity of M@CPB. Figure 2Q depicts that CPB, MΦm, and M@CPB all exhibited an absorption peak at 650 nm, with POD-mimetic activity predominantly derived from CPB (Figs. 2R, S). The initial reaction rates at varying H₂O₂ concentrations were calculated based on the absorbance at 650 nm and fitted to the Michaelis–Menten model, yielding a V_max of 0.01885 ΔA₆₅₀/min and a K_m value of 0.3008 mM (Fig. 2T). Unlike some iron-based Fenton catalysts that generate ·OH, CPB can mitigates ·OH levels by competitively consuming H₂O₂, blocking ·OH generation through redox cycling, and directly scavenging existing ·OH [10, 22]. These findings collectively demonstrated that M@CPB exhibits robust SOD-, CAT- and POD-mimetic activities, conferring efficient ROS-scavenging capability (Fig. 2U).

Given its cytokine-neutralizing and ROS-scavenging functions, as well as its anti-senescence potential, M@CPB was selected as a candidate therapeutic nanozyme for SOP. The biocompatibility of CPB was first evaluated in osteoblast precursor cells, M0 macrophages, and M1 macrophages. Cell viability remained above 80% at concentrations up to 30 µg/mL, indicating acceptable cytocompatibility (Figs. S2A–C). Furthermore, flow cytometric analysis demonstrated that 10 µg/mL CPB produced the most effective intracellular ROS scavenging in MC3T3-E1 cells exposed to H₂O₂ (Figs. S2D and E). Accordingly, 10 µg/mL was selected as the optimal concentration for all subsequent in vitro experiments.

Inflammatory tropism and immune-evasive properties facilitate the accumulation of M@CPB in osteoporotic lesions

MΦm possessed intrinsic inflammatory tropism and immune-evasive properties [14]. To investigate the effect of MΦm coating on the in vivo inflammatory tropism of nanozymes, the biodistribution and circulation profile of biomimetic nanozymes were first evaluated in SOP mice using Ce6-labeled CPB (CPB^Ce6). At 2 h after intravenous injection, M@CPB^Ce6 exhibited stronger fluorescence signals in the limbs than CPB^Ce6 (Fig. 3A). Fluorescence imaging of harvested tissues at 36 h further confirmed increased fluorescence retention in limb bones, lumbar vertebrae, and several major organs in the M@CPB^Ce6 group (Figs. 3B, C). Quantitative analysis showed that fluorescence intensities in the upper and lower limbs were 2.55- and 3.09-fold higher than those of CPB^Ce6, respectively. Enhanced fluorescence was also detected in the liver, spleen, and lung (Fig. 3D), which may reflect both the systemic low-grade inflammatory status characteristic of SOP mice and uptake by the reticuloendothelial system (RES). Consistently, M@CPB^Ce6 displayed significantly higher bone/liver and bone/lung targeting indices than CPB^Ce6 (Fig. 3E). However, the absolute values of these indices remained relatively low, indicating that the increased skeletal accumulation could not be simply interpreted as specific bone targeting.

Fig. 3.

Inflammatory tropism and immune-evasive properties facilitate the accumulation of M@CPB in osteoporotic lesions. (A) In vivo fluorescence imaging of SOP mice 2 h after intravenous injection with CPB^Ce6 or M@CPB^Ce6. (B, C) Ex vivo fluorescence imaging of bones and major organs (heart, liver, spleen, lung, kidney, and brain) at 36 h post-injection. (D) Fluorescence intensity of bones (upper limb, lower limb, lumbar spine) and major organs (heart, liver, spleen, lung, kidney, brain). (E) Bone-to-organ targeting indices. (F) Fluorescence images of plasma samples collected at different time points within 24 h after intravenous injection with CPB^Ce6 or M@CPB^Ce6. (G) Pharmacokinetic profiles and circulation half-lives of CPB^Ce6 and M@CPB^Ce6. (H) Schematic illustration of the Transwell assay (created using Figdraw.com; permission No. YUUUT435d8). (I) Quantification of nanoparticle transendothelial transport. (J) Cellular uptake of M@CPB^Ce6, CCR2i-M@CPB^Ce6 and Lip@CPB^Ce6 by MC3T3-E1 and RAW264.7 cells under different conditions. (K) Time-dependent uptake of CPB^Ce6, Lip@CPB^Ce6 and M@CPB^Ce6 by resting RAW264.7 cells. All quantitative data are presented as mean ± SD (n = 3)

Pharmacokinetic analysis further revealed that M@CPB^Ce6 exhibited a prolonged circulation time compared with CPB^Ce6, with a 1.47-fold increase in half-life (1.22 h vs. 0.83 h; Figs. 3F, G), consistent with previous reports [16, 23, 24]. While prolonged circulation may increase overall tissue exposure, it is insufficient to fully explain the preferential accumulation observed in osteoporotic skeletal tissues. Considering the elevated expression of inflammatory chemokines in SOP lesions, particularly the CCL2-rich microenvironment [4, 25], the enhanced skeletal distribution of M@CPB^Ce6 is more likely associated with inflammatory tropism that facilitates preferential transport and retention within inflamed tissues. Nevertheless, because increased accumulation was also observed in RES-associated organs, the relative contributions of inflammatory tropism and nonspecific RES sequestration could not be distinguished from biodistribution data alone and therefore required further investigation.

To address this question, in vitro CCL2-directed chemotaxis, cellular uptake, and immune escape assays were conducted. CCL2 is a key chemokine involved in the recruitment of inflammatory cells and is highly expressed by senescent cells within SOP lesions, thereby establishing a chemotactic gradient in the diseased bone microenvironment [4, 25]. To mimic the transendothelial transport of M@CPB^Ce6 from the circulation toward inflamed skeletal tissues, a Transwell model with an endothelial monolayer was established (Fig. 3H). VECs were seeded in the upper chamber to simulate the vascular barrier, while CCL2-containing medium was added to the lower chamber to mimic the inflammatory microenvironment of SOP lesions. 24 h later, CPB^Ce6, M@CPB^Ce6, liposome-coated CPB^Ce6 (Lip@CPB^Ce6), or CCR2-blocked M@CPB^Ce6 (CCR2i-M@CPB^Ce6) were added to the upper chamber. By measuring the fluorescence intensity of Ce6 in the lower chamber, it was found that both CPB^Ce6 and M@CPB^Ce6 exhibited minimal transendothelial transport in the absence of CCL2. In contrast, the presence of CCL2 markedly promoted the transport of M@CPB^Ce6 across the endothelial monolayer, whereas this effect was largely abolished following CCR2 blockade with RS102895. Similarly, Lip@CPB^Ce6 displayed a transport efficiency comparable to that of CCR2i-M@CPB^Ce6 despite the presence of CCL2 (Fig. 3I). These findings demonstrate that the enhanced transendothelial transport of M@CPB^Ce6 is predominantly mediated by the CCL2–CCR2 axis, providing direct evidence for its inflammatory tropism toward CCL2-enriched SOP lesions.

Cellular uptake experiments were further conducted using H₂O₂-treated MC3T3-E1 and RAW264.7 cells as inflammatory cell models, while untreated cells and CCL2-supplemented cells served as the controls. In MC3T3-E1 cells, nanoparticle uptake increased after oxidative stimulation but was only marginally affected by CCR2 inhibition or CCL2 supplementation, indicating that internalization mainly occurred through non-specific endocytosis (Fig. 3J). In contrast, RAW264.7 cells exhibited markedly enhanced uptake of M@CPB^Ce6 following H₂O₂ stimulation, which was significantly reduced by CCR2 blockade. Exogenous CCL2 further promoted M@CPB^Ce6 uptake but had little effect on CCR2i-M@CPB^Ce6 or Lip@CPB^Ce6. These findings suggest that inflammatory macrophages preferentially internalize M@CPB^Ce6 through CCR2-mediated recognition, whereas uptake by osteoblastic cells is largely independent of CCR2-mediated interactions. These results further confirmed the function of membrane-associated CCR2.

Given that the liver, spleen, and lung are major RES organs, immune escape was assessed using resting RAW264.7 cells. As shown in Fig. 3K, intracellular fluorescence of CPB^Ce6 and Lip@CPB^Ce6 increased progressively over time, whereas M@CPB^Ce6 maintained minimal cellular uptake throughout the incubation period. Consistent with previous studies [16, 17], these results indicated that MΦm coating reduced macrophage recognition and phagocytic clearance, thereby prolonging systemic circulation.

Collectively, these findings demonstrated that the enhanced skeletal accumulation of M@CPB was mainly attributed to prolonged circulation-mediated passive accumulation, CCR2-mediated inflammatory tropism, and local cellular retention within inflammatory lesions, whereas nonspecific RES uptake also contributed to its distribution in major organs. Specifically, MΦm coating reduced macrophage recognition and phagocytic clearance, thereby extending systemic circulation, while the CCL2–CCR2 axis facilitated preferential transendothelial transport toward CCL2-enriched osteoporotic lesions. Following lesion localization, enhanced uptake by inflammatory macrophages further promoted nanoparticle retention within the diseased microenvironment. Together, these mechanisms cooperatively contributed to the increased accumulation and prolonged residence of M@CPB at osteoporotic sites.

M@CPB ameliorated oxidative stress in vitro

The antioxidative effect of M@CPB was initially evaluated in vitro. H₂O₂ was employed to mimic the pathological oxidative microenvironment associated with osteoporosis. Treatment of MC3T3-E1 cells with H₂O₂, led to a concentration‑ and time‑dependent decline in cell viability. MTT assay demonstrated that after 72 h of incubation, cell survival rates in 200 µM and 400 µM H₂O₂‑treated groups decreased to 77.79 ± 0.98% and 69.04 ± 1.10%, respectively. In contrast, M@CPB intervention significantly improved cell survival rates of the two groups to 95.07 ± 2.41% and 93.49 ± 2.55%, respectively (Figs. 4A, B). Live/dead cell staining results were consistent with those of the MTT assay. Following M@CPB intervention, the number of PI‑labeled dead cells was significantly decreased (Figs. 4C, D), confirming that M@CPB effectively protects MC3T3-E1 cells against oxidative stress‑induced injury. Subsequently, we measured ROS levels in MC3T3-E1 cells. Compared to the control group, intracellular ROS levels were significantly increased after H₂O₂ exposure. M@CPB treatment significantly attenuated the intracellular green fluorescence signal, and the intracellular ROS fluorescence intensity was reduced. Furthermore, although the CPB- and MΦm-treated groups lowered ROS levels, their effects were weaker than that of M@CPB (Figs. 4E–G), confirming that M@CPB can efficiently scavenge excessive intracellular ROS.

Fig. 4.

M@CPB alleviates oxidative stress and ameliorates mitochondrial dysfunction in MC3T3-E1 cells. (A) Cell viability of MC3T3-E1 cells after treatment with different concentrations of H₂O₂. (B) Cell viability of MC3T3-E1 cells treated with different concentrations of H₂O₂ in the presence of M@CPB. (C) Representative confocal images of Calcein-AM/PI-stained MC3T3-E1 cells treated with different concentrations of H₂O₂. (D) Representative confocal images of Calcein-AM/PI-stained MC3T3-E1 cells treated with different concentrations of H₂O₂ in the presence of M@CPB. (E, F) Flow cytometric analysis and quantification of intracellular ROS in MC3T3-E1 cells using the DCFH-DA probe. (G) Representative confocal images of intracellular ROS in MC3T3-E1 cells detected using the DCFH-DA probe. (H) Intracellular ATP levels in MC3T3-E1 cells. (I, J) Representative confocal images and quantification of mitochondrial membrane potential (ΔΨm) in MC3T3-E1 cells assessed by JC-1 staining. (K) Representative TEM images showing mitochondrial morphology in MC3T3-E1 cells. All quantitative data are presented as mean ± SD (n = 3)

Given the central role of mitochondria in oxidative stress, we investigated the protective effect of M@CPB on mitochondrial function under 400 µM H₂O₂ treatment. The ATP assay revealed that the relative intracellular ATP level in the model group dropped to 53.67 ± 12.34% of the control group, whereas all treatment groups exhibited significant restoration. Notably, the M@CPB group exhibited a relative ATP level of 95.19 ± 2.44%, comparable to the control (Fig. 4H). JC-1 staining revealed that the elevated green-to-red fluorescence ratio in the model group was reversed by M@CPB, indicating restoration of mitochondrial membrane potential (ΔΨₘ). Additionally, a similar trend in this ratio was observed in CPB‑ and MΦm‑treated groups, though to a lesser extent (Figs. 4I, J). TEM images revealed that H₂O₂ stimulation induced marked mitochondrial damage, including cristae disruption and matrix swelling, which was markedly alleviated after M@CPB treatment (Fig. 4K). These results indicate that M@CPB effectively preserves mitochondrial structural and functional integrity under oxidative stress. Additionally, analogous protective effects were observed in macrophages. M@CPB also scavenged H₂O₂‑induced intracellular ROS and alleviated mitochondrial dysfunction, with better efficacy than CPB and MΦm. These results were consistent with those in MC3T3‑E1 cells (Figs. S3, S4).

M@CPB reversed osteoprogenitor cell senescence and promoted bone formation

ROS are key drivers of cellular senescence [26, 27], a stable cell-cycle arrest induced by diverse cellular stresses or damage, accompanied by distinct phenotypic alterations. Cell-cycle arrest markers (p16INK4a and p21), the DNA damage marker (γ-H2AX), and senescence-associated β-galactosidase (β-gal) activity are canonical biomarkers of senescent cells [28]. Based on previous studies [7] and Fig. S5, we employed H₂O₂ simulation (400 µM) to induce a senescent MC3T3-E1 cell model. H₂O₂ exposure significantly increased the proportion of β-gal-positive cells, accompanied by upregulated γ-H2AX, p16INK4a, and p21 levels (Figs. 5A–E). However, M@CPB treatment significantly reversed the H₂O₂-induced alterations in these markers. A similar but weaker reversal of these markers was observed in MC3T3-E1 cells treated with CPB. Furthermore, the group treated with MΦm alone exhibited the weakest capacity to reverse these H₂O₂-induced changes. These results revealed that M@CPB exerts potent anti-senescence effects, which are primarily attributed to CPB activity. Notably, MΦm alone exhibited modest anti-senescence activity. Although no direct studies have demonstrated the anti-senescence effects of macrophage membranes, previous studies have demonstrated that macrophage membrane-coated nanomedicines exhibit favorable anti-senescence activity, attributed to the ability of macrophage membranes to neutralize SASP factors and inhibit senescence spread [6, 23, 29]. Notably, M@CPB showed stronger inhibition of p21 than the classic antioxidant NAC, suggesting that ROS scavenging substantially contributes to its anti-senescence effect, with the macrophage membrane potentially providing additional benefits through SASP modulation and suppression of senescence spread (Fig. S6). M@CPB also exhibited equally favorable anti-senescence effects in macrophages (Fig. S7).

Fig. 5.

M@CPB reverses cellular senescence and promotes osteogenic differentiation in MC3T3-E1 cells. (A) Schematic illustration of M@CPB-mediated reversal of cellular senescence in MC3T3-E1 cells (created using Figdraw.com; permission No. AWYPTa2aa4). (B) Representative β-gal staining images of MC3T3-E1 cells following different treatments. (C) Representative confocal images of γ-H2AX in MC3T3-E1 cells following different treatments. (D) Representative confocal images of p16INK4a in MC3T3-E1 cells following different treatments. (E) Representative confocal images of p21 in MC3T3-E1 cells following different treatments. (F–I) Flow cytometric cell-cycle analysis of MC3T3-E1 cells and quantification of cell-cycle distribution, G1-phase proportion, and G2-phase proportion following different treatments. (J) Quantification of ALP activity in MC3T3-E1 cells following different treatments. (K, L) Evaluation of osteogenic differentiation in MC3T3-E1 cells by ALP and ARS staining following different treatments. All quantitative data are presented as mean ± SD (n = 3)

Subsequently, flow cytometry was utilized to examine the effects of M@CPB on the cell cycle of senescent osteoprogenitor cells. As presented in Fig. 5F and quantified in Figs. 5G–I, compared with the control group (G1 phase: 39.23 ± 3.63%, G2 phase: 19.47 ± 0.38%), H₂O₂-induced senescent cells exhibited a remarkable G1 phase arrest (55.07 ± 0.75%), accompanied by reduced cell proliferative capacity, with the G2 phase proportion being 17.10 ± 0.53%. M@CPB treatment significantly alleviated cell cycle arrest and enhanced proliferative potential (G1 phase: 46.20 ± 1.00%, G2 phase: 24.83 ± 1.93%). The CPB alone group exhibited a moderate effect (G1 phase: 49.13 ± 0.99%, G2 phase: 21.30 ± 0.36%), whereas the MΦm alone group exhibited the weakest effect (G1 phase: 51.97 ± 0.93%, G2 phase: 20.73 ± 1.29%). These findings confirm the anti-senescence potential of CPB and the supportive role of macrophage membrane coating in facilitating its biological function.

Subsequently, we assessed the effects of M@CPB on the osteogenic differentiation capacity of senescent osteoprogenitor cells. After 7 days of osteogenic induction, ALP activity assays revealed that ALP levels in the control group and M@CPB group were 44.44 ± 1.02 King units/gprot and 31.71 ± 0.46 King units/gprot, respectively, which were significantly higher than those of the model group (15.14 ± 0.12 King units/gprot). Furthermore, CPB and MΦm modestly increased ALP activity (Fig. 5J). ALP staining revealed that the M@CPB group exhibited significantly deeper staining and a larger positive staining area compared with the model group (Fig. 5K), indicating enhanced early-stage osteogenic differentiation. Meanwhile, a similar changing trend with a weaker effect was observed in the CPB and MΦm groups. Additionally, ARS staining was used to identify osteoblast-produced calcium nodules in the advanced stage of osteogenic differentiation, namely bone mineralization. The staining results revealed that compared with the Model group, the number and area of orange-red calcium nodules in CPB and MΦm groups increased slightly, whereas the two parameters in the M@CPB group were significantly higher, suggesting a higher overall level of bone mineralization in the M@CPB group, similar to the control group (Fig. 5L). These findings indicate that M@CPB can effectively restore the impaired osteogenic differentiation ability of senescent osteoprogenitor cells.

M@CPB suppressed inflammation in senescent macrophages and reduced bone resorption

Senescent macrophages maintain a sustained M1-like pro-inflammatory phenotype, perpetuating chronic inflammation in the osteoporotic microenvironment and supporting persistent secretion of SASP factors [9]. Previous studies have demonstrated that nanomedicines capable of reprogramming macrophage polarization [30] and neutralizing SASP factors [6] exhibit favorable therapeutic effects in osteoarthritis and osteoporosis treatment. We subsequently investigated the effect of M@CPB on phenotypic polarization of senescent macrophages. Senescent RAW264.7 cells exhibited a prominent M1 phenotype with upregulated CD80 and downregulated CD206 relative to the control, whereas M@CPB reversed this profile and shifted macrophages from an M1-like toward an M2-like phenotype. Additionally, a similar but weaker reversal of CD80 and CD206 expression was observed in the senescent macrophages treated with CPB and MΦm (Figs. 6A and B). ELISA assays revealed that M@CPB reduced secreted RANKL levels in MC3T3-E1 cells to 47.41 ± 2.92% of the model group, decreased TNF-α and IL-6 levels in RAW264.7 cells to 53.10 ± 2.80% and 45.10 ± 1.85% of the model group, respectively, and restored IL-10 expression to 83.86 ± 1.61% of the control group. Moreover, a similar but weaker trend was observed in CPB- and MΦm-treated groups (Figs. 6C–F). These findings revealed that M@CPB effectively inhibits the pro-inflammatory state of macrophages and the secretion of SASP factors in senescent cells.

Fig. 6.

M@CPB inhibits inflammation in senescent macrophages and reduces bone resorption. (A, B) Representative confocal images of CD80 and CD206 in RAW264.7 cells following different treatments. (C–F) Quantification of secreted RANKL levels in MC3T3-E1 cell culture supernatants and secreted TNF-α, IL-6, and IL-10 levels in RAW264.7 cell culture supernatants by ELISA. (G) Schematic illustration of the MC3T3-E1–RAW264.7 cell co-culture system (created using Figdraw.com; permission No. SYAUUcd900). (H) Schematic illustration of the RAW264.7–MC3T3-E1 cell co-culture system (created using Figdraw.com; permission No. SYAUUcd900). (I) Representative TRAP staining images of osteoclasts following incubation with MC3T3-E1 cell culture supernatants. (J) Quantification of TRAP-positive multinucleated osteoclasts. (K) Flow cytometric analysis of apoptosis in MC3T3-E1 cells following incubation with RAW264.7 cell culture supernatants. (L) Quantification of the total apoptotic rate of MC3T3-E1 cells (early and late apoptosis). All quantitative data are presented as mean ± SD (n = 3)

M@CPB restored the osteogenic potential of senescent osteoprogenitor cells and suppressed the inflammation in senescent macrophages. However, osteoprogenitor cells and macrophages engage in intimate crosstalk in vivo through shared cytokine signaling and growth factors. Notably, RANKL, highly expressed in senescent osteoblasts, promotes the differentiation of monocyte precursors into mature osteoclasts, whereas TNF-α, a key pro-inflammatory cytokine overexpressed in senescent macrophages, exacerbates osteoblast apoptosis [6]. The two pathological processes synergistically accelerate bone resorption. To investigate whether M@CPB could interrupt this pro-resorption crosstalk between these two cell types, osteoprogenitor cells and macrophages were co-cultured in a Transwell system. First, RAW264.7 cells were incubated with conditioned medium from treated MC3T3-E1 cells, followed by TRAP staining (Figs. 6G and I) and quantitative analysis (Fig. 6J). Compared with the control, the model group exhibited abundant, typical multinucleated TRAP-positive osteoclasts, whereas osteoclast differentiation was markedly inhibited in CPB, MΦm, and M@CPB groups, with the M@CPB group exhibiting phenotypic levels comparable to those of the control. We subsequently incubated MC3T3-E1 cells with conditioned medium from treated macrophages, and apoptosis was assessed by Annexin V-FITC/PI staining and flow cytometric analysis (Figs. 6H and K), accompanied by quantitative analysis of the apoptotic rate (Fig. 6L). The model group exhibited a significantly higher apoptotic rate than the control group, whereas apoptotic cells were decreased in CPB, MΦm and M@CPB groups. Among these groups, M@CPB produced the greatest reduction, with the apoptotic rate approaching that of the control group. Collectively, these findings demonstrate that M@CPB can disrupt the pathogenic cytokine crosstalk between senescent osteoprogenitors and senescent macrophages, thereby mitigating bone destruction.

M@CPB ameliorated SOP-related bone loss, motor dysfunction and depressive-like behaviors

Based on the findings above, we evaluated the therapeutic efficacy of M@CPB against SOP. Four-month-old male SAMP6 mice were utilized as the SOP model, with age-matched SAMR1 mice serving as the control group [18].SAMP6 mice exhibit typical SOP phenotypes [31, 32], human-like age-related bone loss [33], and are widely accepted for SOP research [12, 34]. Male mice were selected to avoid the confounding effects of estrogen. The detailed experimental protocol is shown in Fig. 7A. Micro-CT scanning of the femurs was performed to evaluate the effect of M@CPB on bone microstructure in osteoporotic mice. Three-dimensional (3D) reconstruction revealed marked bone loss in the model group compared to the control group. In contrast, all treatment groups effectively reversed this osteoporotic phenotype, with M@CPB exhibiting the strongest effect (Fig. 7B). Quantitative analysis of trabecular bone parameters, including bone volume/tissue volume (BV/TV), trabecular number (Tb. N), trabecular thickness (Tb. Th), trabecular separation (Tb. Sp), and structure model index (SMI), was performed as reliable indicators of bone mass changes [7]. Compared with the control group, BV/TV, Tb. N, and Tb. Th were significantly reduced in the model group. Specifically, BV/TV dropped from 28.57 ± 5.44% to 9.15 ± 2.82%, Tb. N from 0.0037 ± 0.0006 μm⁻¹ to 0.0016 ± 0.0005 μm⁻¹, and Tb. Th from 76.91 ± 9.20 μm to 56.08 ± 3.95 μm. In contrast, Tb. Sp and SMI were markedly increased, from 181.70 ± 17.31 μm to 270.10 ± 61.16 μm and from 1.48 ± 0.23 to 2.38 ± 0.15, respectively. Following M@CPB treatment, these parameters were substantially restored toward control levels, with BV/TV reaching 35.06 ± 4.77%, Tb. N 0.0040 ± 0.0004 μm⁻¹, Tb. Th 87.79 ± 9.68 μm, Tb. Sp 167.90 ± 24.25 μm, and SMI 1.31 ± 0.19. Notably, the therapeutic efficacy of M@CPB was superior to that of CPB (BV/TV: 23.38 ± 4.13%; Tb. N: 0.0035 ± 0.0006 μm⁻¹; Tb. Th: 78.31 ± 5.67 μm; Tb. Sp: 211.90 ± 42.22 μm; SMI: 1.51 ± 0.21) and MΦm (BV/TV: 22.64 ± 2.20%; Tb. N: 0.0030 ± 0.0006 μm⁻¹; Tb. Th: 77.37 ± 9.03 μm; Tb. Sp: 232.80 ± 44.65 μm; SMI: 1.78 ± 0.17) (Figs. 7E–I). These results indicate that M@CPB markedly enhances trabecular bone mass and effectively restores bone microarchitecture in SOP mice. H&E staining revealed that trabecular bone beneath the femoral growth plate was markedly reduced in the model group compared with the control group (Fig. 7C). The M@CPB treated group exhibited a denser trabecular structure, indicating effective reversal of bone loss. In contrast, CPB- and MΦm-treated groups exerted similar but weaker effects. Masson staining confirmed the bone-protective effect of M@CPB (Fig. 7D), indicating improved collagen deposition and bone matrix organization after M@CPB treatment, whereas CPB- and MΦm-treated groups exerted comparable yet milder effects.

Fig. 7.

M@CPB improves bone mass in SOP mice. (A) Schematic illustration of the animal experimental design. (B) Representative micro-CT images of the distal femur. (C) Representative H&E staining images of femoral sections. (D) Representative Masson’s trichrome staining images of femoral sections. (E–I) Quantification of micro-CT parameters of the distal femur, including BV/TV, Tb. N, Tb. Th, Tb. Sp, and SMI. All quantitative data are presented as mean ± SD (n = 6)

To evaluate the effects of M@CPB on motor function and depressive-like behaviors in SOP mice, we conducted a series of behavioral tests. In the open field test (OFT), the model group exhibited decreased total moving distance and time spent in the central zone compared with the control group (Figs. S8A, B). Similarly, in the elevated plus maze test (EPMT), the model group exhibited a marked reduction in total moving distance and percentage of open‑arm time (OT%) relative to the control group (Figs. S8C, D). Following treatment, total moving distance and OT% were elevated in all treatment groups (Figs. S8E–H), with the M@CPB‑treated group exhibiting the optimal therapeutic effect. These findings indicate that SOP mice exhibit impaired motor function and depressive‑like behaviors, which are effectively ameliorated by M@CPB administration.

To confirm the motor function-enhancing effect of M@CPB, the pole test (PT) was performed, and the total traversal time (TT) was recorded. The model group exhibited significantly prolonged TT compared with the control group (76.72 ± 30.59 s versus 11.50 ± 1.74 s). M@CPB treatment (12.22 ± 3.31 s) restored TT to a level almost identical to that of the control group, which was significantly shorter than that in the CPB (16.50 ± 3.02 s) and MΦm (21.61 ± 3.66 s) treatments. These findings further confirmed that M@CPB ameliorated motor function deficits in SOP mice (Figs. S8I, J). Additionally, the tail suspension test (TST) was conducted to evaluate depressive-like behaviors. As anticipated, all treatment groups shortened immobility time relative to the model group, with M@CPB producing the most significant reduction (Figs. S8K, L). Additionally, a tibial three-point bending test confirmed that M@CPB outperformed CPB and MΦm in enhancing bone mechanical resistance, exerting the most potent anti-osteoporotic effect (Figs. S8M–P).

Collectively, these findings indicated that M@CPB ameliorates bone loss, motor dysfunction, and depressive-like behaviors in mice with SOP.

Molecular mechanisms underlying M@CPB-mediated reversal of SOP

We utilized data-independent acquisition (DIA) proteomics to identify differentially expressed proteins (DEPs) in the bone tissues from the upper limbs of mice in the control, SOP, and SOP + M@CPB groups to explore the protein-level mechanisms underlying the anti-osteoporotic effects of M@CPB. Principal component analysis (PCA) revealed clear separation among the control, SOP, and SOP + M@CPB groups (Fig. 8A), with 4899 proteins commonly identified across all three groups (Fig. 8B). We subsequently performed gene set enrichment analysis (GSEA) for the SOP versus control and SOP + M@CPB versus SOP comparisons. GSEA of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway revealed that the osteoclast differentiation pathway was significantly upregulated in the SOP group compared with the control group. In contrast, cell cycle and DNA replication-related pathways were significantly downregulated. In the SOP + M@CPB group, DNA replication and neurodegenerative disease-related signaling pathways were significantly upregulated compared with the SOP group (Fig. 8C). Notably, the DNA replication gene set exhibited significant differences between the two comparison groups (Fig. 8D). Cellular senescence is generally defined as a state of stable cell cycle arrest [28], and DNA replication occurs during normal cell cycle progression [35]. Neurodegenerative diseases are closely associated with aging [36]. These findings suggest that SOP progression is associated with cellular senescence and alterations in aging- and nerve-related pathways, which may contribute to bone loss and behavioral dysfunction.

Fig. 8.

M@CPB regulates the expression of proteins associated with cellular senescence, nerve-related markers, and osteogenesis in the bone tissues of SOP mice. (A) PCA score plot of bone proteomics data from the Control, SOP, and SOP + M@CPB groups. (B) Venn diagram showing commonly identified proteins among the Control, SOP, and SOP + M@CPB groups. (C) GSEA of KEGG pathways in the SOP versus Control and SOP + M@CPB versus SOP comparisons. (D) GSEA enrichment plot of the DNA replication gene set. (E) Heatmap of proteins associated with the cell cycle. (F) Heatmap of proteins associated with bone metabolism. (G) Representative immunofluorescence images of p16INK4a in femoral sections. (H) Representative immunofluorescence images of p21 in femoral sections. (I) Representative immunofluorescence images of γ-H2AX in femoral sections. (J) Representative double immunofluorescence images of NGF and NF-H/NF200 in femoral sections. (K) Representative immunofluorescence images of RUNX2 in femoral sections. (L–P) Quantification of p16INK4a, p21, γ-H2AX, NGF, and RUNX2 fluorescence intensity in femoral sections. All quantitative data are presented as mean ± SD (n = 3)

We subsequently identified DEPs associated with the cell cycle and bone metabolism. Compared with the SOP model group, the SOP + M@CPB group exhibited increased levels of cell-cycle-promoting regulators, including Ccnd3, Cdk1, Cdk2, and Rb1, and a decreased levels of cell-cycle inhibitory regulators, including Rbl1, Bccip, and Trp53i11 (Fig. 8E). Ccnd3 is a key cyclin involved in the G1/S phase transition. CDKs interact with cyclins to orchestrate cell cycle progression, with the CDK2-cyclin E complex being essential for S-phase entry, and the CDK1-cyclin A/B complex inducing S/G2 and G2/M transitions to facilitate DNA replication [37]. Rb1 controls G1/S phase progression and serves as a CDK4/6 target [38], whereas Rbl1 (p107) interacts with E2F to inhibit the G1/S transition [39]. These protein changes indicated that M@CPB may restore cell-cycle progression in the bone tissues of SOP mice. Among bone metabolism-related proteins, the SOP + M@CPB group showed elevated levels of osteogenesis-related proteins (P4ha1, Omd, Spp1, Plod2, Alpl, Bglap3, and Col10a1) and reduced levels of the osteoclast-related protein Ctsk (Fig. 8F). Spp1 (osteopontin) facilitates bone mineralization and calcium deposition [40], Alpl, the early marker of osteogenesis, regulates mineralization [41], and Bglap3 (osteocalcin) is involved in bone development and osteoblast differentiation [42]. Col10a1, P4ha1, Omd, and Plod2 [43] are associated with collagen formation and extracellular matrix synthesis, whereas Ctsk degrades collagen and elastin to promote bone resorption [44]. These findings, consistent with the in vitro results, support that M@CPB facilitates osteogenic differentiation and inhibits osteoclastogenesis.

Therefore, we assessed the levels of cell-cycle inhibitors (p16INK4a and p21), the DNA damage marker γ-H2AX, neural markers (NGF and NF-H/NF200), and the osteogenic marker RUNX2 [45] in mouse femurs. Immunofluorescence staining demonstrated that M@CPB markedly reduced the levels of senescence-associated proteins, including p16INK4a, p21, and γ-H2AX, in the femurs of SOP mice, indicating inhibition of cellular senescence; CPB and MΦm exerted similar but milder inhibitory effects (Figs. 8G–I). Meanwhile, NGF, NF-H/NF200, and RUNX2 levels were significantly increased compared with the model group, with stronger effects observed in the M@CPB group than in the CPB and MΦm groups (Figs. 8J and K). Quantitative analysis confirmed that these changes were statistically significant (Figs. 8L-P). Collectively, these results demonstrated that M@CPB may preserve nerve-related marker expression and facilitate osteogenesis, at least partly through the suppression of cellular senescence.

Gene ontology (GO) biological process enrichment analysis indicated that, relative to the control group, the SOP group exhibited marked upregulation of the acute-phase response, ROS metabolic and biosynthetic processes, the classical NF-κB signaling pathway, and the MAPK cascade (Fig. 9A). Conversely, compared with the SOP group, the SOP + M@CPB group showed enrichment of biological processes related to bone remodeling, with reduced enrichment in the acute-phase response and apoptosis-related mitochondrial changes. Cellular senescence is associated with systemic low-grade chronic inflammation [4], in which ROS activates NF-κB signaling to drive SASP secretion and bone resorption. These findings suggest that cellular senescence in osteoporotic bone may be associated with elevate ROS levels, inflammatory cascades, and bone deterioration.

Fig. 9.

M@CPB regulates the expression of proteins associated with inflammation and bone resorption in the bone tissues of SOP mice. (A) GO enrichment analysis of differentially expressed proteins. (B) Heatmap of proteins associated with energy metabolism and antioxidant defense. (C) Heatmap of proteins associated with the NF-κB signaling pathway. (D) Representative fluorescence images of DHE staining in femoral sections. (E) Representative immunofluorescence images of CD80 in femoral sections. (F) Representative immunofluorescence images of CD206 in femoral sections. (G) Quantification of serum RANKL, TNF-α, IL-6, and IL-10 levels by ELISA. (H) Representative immunohistochemical images of RANKL in femoral sections. (I) Representative immunohistochemical images of TNF-α in femoral sections. (J) Representative TRAP staining images of femoral sections. (K–O) Quantification of DHE fluorescence intensity and CD80, CD206, RANKL, and TNF-α staining intensities in femoral sections. All quantitative data are presented as mean ± SD (n = 3)

To further investigate the underlying mechanism, DEPs associated with energy metabolism, antioxidant defense, and the NF-κB signaling pathway were analyzed. The SOP + M@CPB group exhibited increased expression of Keap1, Ndufa4l2, Gpx8, Cyp2f2, Ndufv3, Pycr2, Gstm5, and Mtatp6 compared with the model group (Fig. 9B). Keap1 is a key regulator of the Nrf2 antioxidant pathway [46]. Ndufv3, Ndufa4l2, and Mtatp6 are essential components of the mitochondrial respiratory chain, contributing to electron transport and energy production [47–49]. GPx8 scavenges hydrogen peroxide and organic peroxides, thereby protecting cells from oxidative damage [50]. These results collectively support the involvement of oxidative stress regulation in the therapeutic effects of M@CPB. Additionally, compared with the SOP group, the SOP + M@CPB group exhibited reduced levels of NF-κB signaling-related pro-inflammatory proteins, including Irak1, Nfkb1, Tnfaip2, Ikbkb, Cxcl5, Myd88, and Card9 (Fig. 9C), suggesting suppression of NF-κB pathway activation.

DHE fluorescence staining revealed that M@CPB markedly attenuated ROS accumulation in SOP mouse femurs, whereas CPB and MΦm exhibited similar but milder effects (Fig. 9D). M@CPB treatment downregulated the M1 marker CD80 and upregulated the M2 marker CD206 relative to the model group (Figs. 9E and F), promoting a shift from an M1-like toward an M2-like phenotype, and this regulatory effect was significantly stronger than that of CPB and MΦm. ELISA assays revealed that M@CPB decreased RANKL, TNF-α, and IL-6 levels and increased IL-10 levels compared with the model group (Fig. 9G), indicating suppression of SASP factor secretion and inflammatory responses, with effects more pronounced in the M@CPB group than in the CPB and MΦm groups. Immunohistochemical staining further confirmed reduced RANKL and TNF-α levels in M@CPB-treated mouse femurs, with weaker reductions detected in the CPB and MΦm groups (Figs. 9H and I). Moreover, TRAP staining indicated that M@CPB inhibited osteoclast formation and mitigated bone resorption more effectively than CPB and MΦm (Fig. 9J). Additionally, quantitative analysis confirmed that these marker changes were statistically significant following M@CPB treatment (Figs. 9K–O). Collectively, these findings demonstrate that M@CPB can mitigate bone resorption by alleviating oxidative stress, suppressing inflammatory responses, and reducing SASP factor secretion in the osteoporotic microenvironment.

Untargeted metabolomics was performed to identify differentially abundant metabolites (DAMs) in the serum of control, SOP, and SOP + M@CPB mice. PCA revealed a clear separation among the three groups (Fig. S9A), with 1911 metabolites commonly detected across the three groups (Fig. S9B). Using screening criteria of fold change (FC) > 1.2 and P < 0.05, we identified 74 upregulated and 44 downregulated metabolites in control versus SOP, and 50 upregulated and 39 downregulated metabolites in SOP versus SOP + M@CPB (Fig. S9C). KEGG pathway annotation showed that the two comparisons were mainly enriched in pyrimidine metabolism, linoleic acid metabolism, arachidonic acid metabolism, and arginine/proline metabolism (Figs. S9D, E). To clarify the therapeutic mechanism of M@CPB against senile osteoporosis, we focused on metabolites associated with oxidative stress, inflammation, and bone matrix metabolism. The SOP + M@CPB group exhibited altered levels of energy metabolism- and antioxidant-related metabolites including 3-dehydro-D-glucose-6-phosphate, fructose 6-phosphate, L-glutamine acid, indole-3-pyruvic acid, and L-glutamic acid, suggesting modulation of oxidative stress-related metabolism (Fig. S9F). For inflammation-related metabolites, M@CPB treatment remodeled lipid inflammatory mediators and related metabolites, includingprostaglandin D3, prostaglandin E3, tetranor-PGFM, leukotriene B4, and D-homocysteine, suggesting regulation of inflammation-associated metabolic pathways (Fig. S9G). In addition, M@CPB regulated bone matrix-associated metabolites including L-proline, pyrophosphate, 19-hydroxy testosterone, L-alanine, docosahexaenoic acid glycine, and dioleoyl phosphatidylcholine (Fig. S9H). Collectively, these metabolomic results suggest that M@CPB alleviates senile osteoporosis by modulating metabolic pathways associated with oxidative stress, inflammation, and bone matrix metabolism, which is consistent with the in vitro observations.

Safety evaluation of M@CPB

Good biocompatibility is essential for the clinical translation of nanomedicines. The cytocompatibility of M@CPB was first evaluated via MTT assays. HepG2 cells, HK-2 cells, HUVECs, and MC3T3-E1 cells, with or without oxidative stress, were incubated with M@CPB for 24 h. Notably, the cell viability of all M@CPB-treated groups remained above 80% compared with the control groups, irrespective of whether oxidative stress was induced (Figs. S10A, B). These results indicated that M@CPB exhibits favorable cytocompatibility toward these cells. Subsequently, the hemocompatibility of M@CPB was further evaluated using hemolysis and coagulation assays. After 6 h of co-incubation, the hemolysis rate of red blood cells exposed to CPB and M@CPB at 10, 20, and 40 µg/mL remained below 8%, using water as the positive control and PBS as the negative control (Fig. S10C), and red blood cell morphology was preserved (Fig. S10D). Coagulation assays revealed no significant difference in the coagulation rate between M@CPB and PBS groups (Fig. S10F). Collectively, these results revealed that M@CPB exhibits low cytotoxicity and favorable biocompatibility in vitro.

Nanomedicine biosafety is often assessed using zebrafish embryos because of their simple reproductive system, rapid developmental cycle, and high genetic similarity to human genes. The data indicated no significant differences in hatching process, morphology, survival rate, or embryonic heart rate among the control, CPB, and M@CPB groups during zebrafish development (Figs. S10E, G, and H). These findings indicate that M@CPB has favorable developmental biosafety, which is essential for in vivo treatment.

The in vivo safety of CPB, MΦm, and M@CPB was assessed in mice after 8 weeks of administration. H&E staining of the heart, liver, spleen, lungs, and kidneys revealed no obvious morphological abnormalities or signs of inflammation after in vivo administration of CPB, MΦm, or M@CPB. Specifically, the major organs exhibited normal histological structures: well-organized myocardial fibers in the heart; intact hepatic lobular structure with regularly arranged hepatocytes and distinct hepatic sinusoids in the liver; normal red-white pulp in the spleen; well-defined alveolar structures in the lungs; and intact glomerular and tubular structures in the kidneys. (Fig. S10I). Serum biochemical analysis (Figs. S10J–M) and routine peripheral examination (Figs. S10N–Q) showed no obvious treatment-related abnormalities in the M@CPB group, suggesting that M@CPB did not induce evident hepatic, renal, or hematological toxicity under the tested conditions [51]. The results indicate that M@CPB exhibits favorable biological safety, supporting its potential for further therapeutic development.

Conclusions

In this study, we developed a novel nanozyme (M@CPB) and investigated its therapeutic effects and underlying mechanisms in SOP. The rationally designed M@CPB exhibited potent oxidative stress-ameliorating activity, effectively reducing osteoprogenitor cell senescence and promoting bone formation while suppressing inflammation in senescent macrophages and inhibiting bone resorption. Notably, M@CPB effectively ameliorated SOP-related pathological changes with favorable safety. Collectively, as a novel multifunctional nanodrug, M@CPB combines therapeutic efficacy and favorable biosafety, highlighting its potential for SOP treatment and, potentially, other aging-related diseases.

Author contributions

L. L., B. L. and Y. L. conceived and designed the study. Y. L. and Y. T. designed and synthesized the materials, performed the structural characterizations and catalytic experiments, and analyzed the data. Y. L. and L. D. performed the in vitro and in vivo study, and analyzed the data. H. L. and J. Y. assisted with the in vitro and in vivo study. X. H., S. Z. and X. X. provided constructive suggestions for the investigation. L. L., B. L. and Y. L. analyzed the results and wrote the paper. All the authors discussed the results and commented on the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (grant number 82171581); the National Undergraduate Innovation and Entrepreneurship Training Program (grant number S202410533139); the Innovation Training Program for College Students, Central South University (grant number X202510533658); the Hunan Provincial Basic Education Teaching Reform Research Project (grant number Y2024898); the Research Project on Education and Teaching Reform, Central South University (grant number 2023jy100); the Hunan Provincial Natural Science Foundation (grant number 2022JJ40698); and the Hunan Provincial College Students’ Innovation and Entrepreneurship Training Program (grant number S202510533421).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

The animal use protocol has been approved by the Institutional Animal Care and Use Committee, The Second Xiangya Hospital, Central South University, China (No. 20240989), and none of the animal experiments violated the aims of the Declaration of Helsinki of the Ministry of Health.

Consent for publication

Written informed consent was obtained from all participants.

Competing interests

The authors declare no competing interests.