Spatiotemporally adaptive metal-gas nanotherapy: Acid-triggered cascade release for osteoporosis treatment

Abstract

The treatment of osteoporosis remains a critical challenge due to the limitations of current therapies in simultaneously restoring homeostasis of bone metabolism and modulating the chronic inflammation of bone microenvironment. While the acidic microenvironment greatly exacerbates bone resorption, it also lays the foundation for the application of responsive therapeutic nanoplatforms. Herein, inspired by the rocket-like responsive cascade release principle-defined by a “booster-payload” mechanism, in which a metal ion-doped mesoporous silica shell serves as an early-release booster and a calcium sulfide core acts as a sustained-release payload-a bone-targeted nanoplatform (CSM₃P) was developed. This platform integrates calcium sulfide nanoparticles with magnesium/manganese-doped mesoporous silica for bone targeting. It leverages the pathological acidic microenvironment of osteoporosis to enable stimuli-responsive and stepwise release of multiple mineral ions (Mg²⁺/Mn²⁺/Ca²⁺) and hydrogen sulfide (H₂S), coordinates bone regeneration and regulation of the osteoimmunological microenvironment and thereby achieves metal-gas targeted therapy for osteoporosis. This acidic responsive and sequential ion/gas release model establishes a self-reinforcing therapeutic cycle, where the pathological acidity itself drives a synergistic and adaptive treatment regimen. Both in vitro and in vivo evaluations demonstrate the superior efficacy of CSM₃P in ameliorating the inflammatory bone microenvironment and rebalancing bone remodeling. This work offers a novel paradigm for intelligent, feedback-driven nanotherapy against osteoporosis and other microenvironment-associated diseases.

Graphical abstract

Highlights

• •A bone-targeting nanoplatform (CSM3P) triggers cascade release of H₂S and therapeutic metal ions in acidic osteoporotic niches.
• •CSM₃P integrates gas therapy and metal ion supplementation via a spatiotemporally adaptive mechanism.
• •CSM₃P attenuates inflammation by suppressing cytokines and driving M2 macrophage polarization.
• •CSM₃P rebalances bone remodeling by inhibiting osteoclastogenesis and enhancing osteogenesis.

1. Introduction

Osteoporosis is one of the most prevalent bone metabolic disorders, characterized by an imbalance between osteoclastic bone resorption and osteoblastic bone formation [1]. Persistent imbalance in bone remodeling leads to progressive bone loss, deterioration of bone microarchitecture, and a chronic low-grade inflammatory state, rendering the skeleton highly susceptible to fragility fractures [2]. With a rapidly aging global population, the prevalence of osteoporosis continues to rise, posing a significant public health challenge worldwide [3,4]. Yet, although several drugs, including antiresorptive and osteoanabolic agents, have been approved for clinical use, they often fail to achieve precise, cell-type-specific control over diverse cellular populations amidst the disease’ s multifactorial pathology, resulting in suboptimal therapeutic outcomes [5].

A defining attribute of the osteoporotic niche is its abnormally acidic extracellular milieu, wherein overactivated osteoclasts secrete excessive protons and proteases, creating an acidic bone microenvironment [6]. This pathological acidosis exacerbates osteoclast activation and formation, unleashes pro-inflammatory cytokines and reactive oxygen species (ROS), and cripples neighboring osteoblasts [7]. Notably, this acidic environment also presents a unique opportunity for designing smart stimulus-responsive nanotherapy, which was aligned with the design of leveraging the acidic microenvironment for targeted drug release [6,8]. However, most current pH-sensitive platforms are limited to single-drug delivery or lack the ability to dynamically adapt to disease severity. Therefore, there is a pressing need for nanotherapeutic strategies capable of multi-cellular regulation and spatiotemporal adaptability [9,10]. This necessity arises from the complex cellular interplay underlying osteoporosis. The acidic microenvironment polarizes macrophages toward a pro-inflammatory M1 phenotype, leading to the secretion of TNF-α and IL-6, which promote osteoclast differentiation while inhibiting osteoblast function [11,12]. Overactivated osteoclasts further acidify the niche, perpetuating a vicious cycle that impairs osteoblast activity. Notably, macrophages, osteoclasts, and osteoblasts operate in a temporally sequential cascade: early M1 macrophage-driven immune dysregulation initiates osteoclast activation and subsequent osteoblast dysfunction, ultimately resulting in bone loss. Thus, an ideal therapy must sequentially repolarize macrophages, inhibit osteoclasts, and restore osteoblast function with precise temporal coordination [9,13].

Hydrogen sulfide (H₂S), an endogenous gasotransmitter, has recently garnered attention for its pleiotropic roles in combating inflammation, inhibiting osteoclastogenesis, and promoting osteogenesis. At physiological concentrations, H₂S scavenges excess ROS and inflammatory cytokines, thereby exerting anti-inflammatory effects [14]. It also suppresses osteoclast differentiation by inhibiting NF-κB and RANKL signaling while enhancing osteoblast function via signaling pathways such as Wnt/β-catenin. [15,16]. Despite its therapeutic potential, the clinical translation of H₂S is hampered by its short half-life, poor tissue targeting, and reliance on carbonyl sulfide (COS) as an intermediate-a strategy with intrinsic limitations [17]. COS-dependent donors require a two-step mechanism: triggered release of COS followed by carbonic anhydrase (CA)-catalyzed hydrolysis to H₂S [18]. This enzyme dependence leads to unpredictable kinetics, as CA expression varies across tissues. Moreover, COS itself is toxic, and many systems generate electrophilic quinone methide byproducts that alkylate cellular nucleophiles [19]. Most photoactivated variants rely on UV light, limiting tissue penetration and causing phototoxicity [20,21]. These limitations highlight the urgent need for direct, COS-independent H₂S donors. Furthermore, H₂S alone is insufficient to reverse the severe mineral deficiency in osteoporotic bone [22], as mineral ions (e.g., Ca²⁺, Mg²⁺, Zn²⁺, Mn²⁺) play indispensable roles in bone matrix mineralization and bone homeostasis. Calcium is a core component for bone mineralization, while magnesium and manganese are vital trace elements that regulate the immunoregulation and the balance of osteoblast and osteoclast homeostasis [[23], [24], [25], [26]]. We posit that H₂S creates a favorable “soil” for bone regeneration, whereas mineral ions act as “fertilizers” to promote the robust growth of the “seed”. Thus, a therapeutic strategy that seamlessly integrates H₂S with mineral ion delivery represents an ideal approach for treating osteoporosis. However, the integration of acid responsiveness, bone targeting, and synergistic delivery of multiple therapeutic agents (ions and gas) into a single nanoplatform remains a major challenge, as existing systems often suffer from uncontrollable release kinetics, insufficient targeting efficiency, or lack of multi-mechanism synergy.

Inspired by the pathological characteristics of the osteoporotic acidic microenvironment and the rocket-inspired responsive cascade release principle, we engineered a pH-responsive nanotherapeutic platform (CSM₃P) for the spatiotemporally controlled co-delivery of H₂S and mineral ions (Mg²⁺/Mn²⁺/Ca²⁺) in a metal-gas targeted therapy for osteoporosis. The construct features a core-shell structure: a magnesium/manganese-doped mesoporous silica shell (MSN-MgMn) serving as a “booster” for rapid ion release, and a calcium sulfide (CaS) core acting as the “payload” for sustained H₂S and Ca²⁺ release [27,28]. In contrast to COS-dependent H₂S donors, which require enzymatic conversion and generate potentially toxic byproducts, CSM₃P nanoplatform generates H₂S directly from acid-triggered decomposition of CaS, bypassing COS intermediates and enzymatic conversion. This design ensures predictable, self-regulating H₂S release coupled to disease severity, and produces only beneficial mineral ions (Ca²⁺, Mg²⁺, Mn²⁺) as byproducts. Furthermore, modification with alendronate provides these “weapons” with “guidance”, endowing it with excellent bone-targeting capability to enhance local bioavailability while minimizing off-target effects [29].

Crucially, this unique responsiveness establishes a self-reinforcing therapeutic cycle: the pathological acidity that drives bone loss simultaneously triggers a sequential, programmed release of therapeutic agents. Specifically, under acidic conditions, the MSN-MgMn shell first degrades rapidly, releasing osteogenic Mg²⁺ and Mn²⁺ to initiate bone formation and suppress osteoclast activation [30,31]. Subsequently, the exposed CaS core decomposes, concurrently generating Ca²⁺ for bone mineralization and H₂S for anti-inflammatory and anti-osteoclastogenic effects. This spatiotemporally sequential release ensures optimal therapeutic impact at different stages of bone metabolism. The synergistic effects of the multi-component therapy are pivotal: Mg²⁺ and Mn²⁺ cooperatively regulate NF-κB, MAPK, and Wnt signaling pathways to dually inhibit osteoclasts and promote osteoblasts; H₂S scavenges reactive oxygen species (ROS) and modulates macrophage polarization; and Ca²⁺ supplements bone mineral components [32]. Moreover, the acid-responsive degradation neutralizes the pathological acidic microenvironment, breaking the vicious cycle of 'acidification-osteoclast activation-bone loss. By integrating multi-ion supplementation with gaseous signaling within a single targeting platform, this work introduces a spatiotemporally cascaded treatment paradigm and provides a targeted, effective solution to a critical gap in osteoporosis therapy (Scheme 1).

Scheme 1.

Schematic illustration of the CSM₃P preparation. CSM₃P precisely targets bone tissue and, in response to the acidic microenvironment of osteoporosis, cascades the release of hydrogen sulfide and metallic mineral particles in a spatiotemporally adaptive manner to treat osteoporosis by regulating the inflammatory microenvironment and reversing bone metabolic imbalance.

2. Methods

Materials. Manganese acetate tetrahydrate (Mn (CH₃COO)₂·4H₂O), Magnesium acetate tetrahydrate (Mg (CH₃COO)₂·4H₂O), oleylamine (OM) and oleic acid (OA) were purchased from Sigma. Tetraethyl orthosilicate (TEOS) were purchased from Alfa. Poly (acrylic acid) (PAA, Mw ≈ 2000) was purchased from Acros. 4-(4,6-Dimethoxy-1,3,5-triazin-2-yl)-4-methylmorpholinium chloride (DMTMM) was purchased from J&K Scientific. Alendronate sodium trihydrate (Ald) was purchased from Macklin. Methanol, ethanol, chloroform and cyclohexane were purchased from Xilong Science. Calcium acetate monohydrate (Ca (CH₃COO)₂·H₂O), N, N′-diphenylthiourea (DPTU), triethanolamine (TEA), trioctylamine (TOA), hexadecyl trimethylammonium bromide (CTAB), Aminopropyltriethoxysilane (APTES) were purchased from Aladdin. All the chemical reagents used were analytical grade without any further purification.

Characterization. The morphology of the as-synthesized nanoparticles was observed using field emission transmission electron microscopes (TEM, Tecnai G2 20, Thermo Fisher) and field-emission scanning electron microscopy (SEM, ZEISS Sigma 300). The X-ray diffraction (XRD) pattern was recorded by a D8A A25 X (Bruker). Nitrogen adsorption−desorption isotherms were obtained by an Autosor-iQ (Quantanchrome Instruments). Fourier transform infrared (FT-IR) spectroscopy was performed on a Nicolet iS20 (Thermo Fisher). The Brunauer-Emmett-Teller (BET) method was utilized to calculate the specific surface area by using adsorption data in the range of the relative pressures from 0.01 to 1.00. The pore-size distributions were calculated using the Barrett-Joyner-Halanda (BJH) method. Zeta potentials of samples were determined by a Zetasizer Nano ZS90 (Malvern). The thermogravimetric analysis was performed on a TGA 4000 (PE). The X-ray photoelectron spectroscopy (XPS) was recorded by ESCALAB250Xi (Thermo Fisher). ESR spectrum was recorded by spectrometer (JEOL-FA200). Ultraviolet-visible (UV-vis) spectra were collected on a UV-2600 spectrophotometer (Shimadzu). The elements contained in the materials were determined by an inductively coupled plasma (ICP) emission spectrometer (VARIAN 715-ES, USA). The stained cells were analyzed using a microscope (Olympus) and inverted fluorescence microscopy (Olympus). The absorbance of Cell Counting Kit 8 (CCK8) was measured by the microplate reader of VICTOR Nivo 3S (PerkinElmer).

Synthesis of CaS Nanoparticles. First, 1 mmol of Ca (CH₃COO)₂·H₂O was ultrasonically mixed and dispersed in a three-neck flask containing 24 mL of OM, 4 mL of OA, and 12 mL of TOA. Under nitrogen (N₂) protection, the resulting mixture was heated to 120 °C and continuously stirred for 30 min to form a clear and transparent solution. After cooling to room temperature, 20 mL of ethanol containing 3 mmol of DPTU was added, and the solution was stirred at 80 °C for 30 min to evaporate the ethanol. After that, the resulting solution was heated to 320 °C under N₂ with vigorous stirring for 60 min and then cooled to room temperature to obtain calcium sulfide (CaS) nanoparticles (CaS NPs). The CaS NPs were washed three times with anhydrous ethanol and finally redispersed in 10 mL of cyclohexane.

Synthesis of Amino-Functionalized CaS@mSiO₂ Nanoparticles. First, 1 mL of CaS NPs was washed and redissolved in 10 mL of chloroform. Under a dark and inert environment, a solution of CTAB in anhydrous ethanol was slowly added and vigorously stirred for 30 min. Then, the chloroform was removed to obtain CTAB-modified CaS (CaS@CTAB). The CaS@CTAB was washed and redispersed in anhydrous ethanol. Next, a small amount of ultrapure water was added, and the pH was adjusted to 10–11 using TEA dropwise. Under vigorous stirring, 500 μL of a prepared TEOS ethanol solution (10% v/v) was added dropwise. A mixed aqueous solution (0.1 g/mL of each) of Mn (CH₃COO)₃·4H₂O and Mg (CH₃COO)₂·4H₂O was added dropwise, simultaneously. After stirring for 1 h at room temperature and inert environment, the magnesium ion (Mg²⁺) and manganese ion (Mn²⁺) doped mesoporous silica (MSN-Mg/Mn) was coated to obtain the metal ion doped mesoporous silica coated CaS (CaS@MSN-MgMn, CSM₃). Finally, an ethanol solution containing TEOS and APTES (molar ratio 1:1, 1% v/v) was added and stirred for another 1 h for surface aminoation of CSM₃ (CSM₃-NH₂). The final product was repeatedly washed with anhydrous ethanol and refluxed in a methanol solution containing sodium chloride (1% w/v) for 6 h remove the CATB.

Synthesis of Bone-Targeting Ligand. Following our previous methodology, the bone-targeting ligand was prepared [2]. 0.002 g of PAA and 0.065 g of Ald were dissolved in 100 mL of borate buffer (0.1 M, pH 8.5). After stirring for 10 min, borate buffer solution (0.1 M, pH 8.5) containing 0.055 g of DMTMM was added. The final pH was adjusted to 7.5 using 2 M NaOH. The reaction proceeded under stirring for 24 h. The product was purified by dialysis and lyophilized to obtain the bone-targeting ligand PAA-Ald (PA) as a white foam.

Surface Modification with Bone-Targeting Ligands on CSM₃-NH₂. For the modification of the bone-targeting ligand, CSM3-NH₂ was redispersed in a 45% ethanol-water solution. An aqueous solution of PA (0.1 g/mL) was added dropwise under vigorous stirring. After the solution became clear, the final product of PA decorated CSM₃-NH₂ (CSM₃P) obtained. The final product was centrifuged and washed repeatedly with deionized water, and then CSM₃P was obtained through vacuum drying.

The pH Responsiveness of CSM₃P. To verify the acid-neutralizing capacity of the material, CSM₃P was dissolved in a pH 4.5 solution at a concentration of 10 mg/mL under continuous stirring. The pH variation of the solution was continuously monitored with a pH meter and recorded every 30 s until the solution turned neutral. The pH of the solution was then re-adjusted to 4.5, and the corresponding pH change was recorded again; this cycle was repeated three times.

In Vitro Co-Culture Experiment with Bone Slices. To simulate the pH-responsive properties of the material in vivo, CSM₃P was co-cultured with fresh bone slices in complete cell culture medium for 5 days in vitro. Briefly, bone slices were first sterilized in 75% ethanol for 2 h and subsequently dried. The bone slices were then immersed in a solution containing the FITC-labeled substance (hereinafter abbreviated as FITC-CSM₃P) and incubated at 37 °C in the dark for 12 h. After being washed with physiological saline three times, the treated bone slices were placed in 96-well plates containing complete DMEM medium, with the medium replaced every two days. The fluorescence signal of FITC-CSM₃P was observed using an inverted fluorescence microscope throughout the culture period, and fluorescence images were acquired on days 1, 3, and 5.

Establishment of ovariectomized (OVX) osteoporosis model. 12-week-old female C57BL/6J mice were ovariectomized to establish an osteoporosis model. Intraperitoneal injections of 1% pentobarbital sodium (0.1 mL per 20 g) were administered to anesthetize the mice. The sham procedure, in which the ovaries were only exteriorized but not removed, was performed on mice in the sham group. Four weeks after surgery, the aforementioned animals were randomly divided into 8 groups (n = 5 each), namely the Sham group, OVX group, CaS group, CSM group, CSM-Mg group, CSM-Mn group, CSM₃ group, and CSM₃P group. Mice in the Sham group and OVX group were intraperitoneally injected with normal saline (0.9% sodium chloride) as controls, while mice in the remaining groups were intraperitoneally injected with the corresponding prepared solutions (1 mL of 50 μg/mL) once every 2 days for a continuous period of 4 weeks [33,34]. All the mice were humanely sacrificed after 4 weeks of administration, and the femurs of the mice were collected for micro-CT scans to assess the efficacy. Mice were kept in a specific pathogen-free (SPF) environment with temperature (23 ± 2 °C) and fed with standard pellet diet and pure water. All experimental procedures were approved by the Ethics Committee and Institutional Review Board of Nanchang University (Approval No. NCULAE-20221228021).

Live/Dead staining assay. Live/Dead assay was performed to assess the viability of MC3T3-E1 or RAW264.7 cells following coculture with Cas, CSM, CSM-Mg, CSM-Mn, CSM₃, and CSM₃P. The cells were first seeded statically at a density of 2 × 10⁴ cells per well in 24-well plates and incubated at 37 °C under a 5% CO₂ atmosphere for 24 h. Following this, the culture medium was aspirated, and cells were rinsed with sterile phosphate-buffered saline (PBS) before being incubated with 2 μM calcein AM for 45 min. Subsequently, cells were rinsed again with PBS and imaged using an inverted fluorescence microscope.

Intracellular Reactive oxygen species (ROS) determination. ROS detection kit (Beyotime, China) was used to detect intracellular ROS levels. Cells exposed to different treatments were collected, and incubated with DCFH-DA at 37 °C for 20 min. Then, the cells were washed with serum-free medium. Finally, DCF fluorescence of cells was analyzed by fluorescence microscope (ZISSS, Germany) at 488 nm.

Flow cytometry. To assess cell apoptosis, MC3T3-E1 cells were harvested and centrifuged at RT for 5 min. Cells were then resuspended in 195 μL of buffer solution and incubated with 5 μL Annexin V-FITC and 10 μL propidium iodide (PI) at RT for 15 min according to the manufacturer's instructions (Beyotime, China). Subsequently, the stained cells were analyzed using the FACSCanto II system (BD Biosciences) to determine the proportion of apoptotic cells. To assess ROS levels, RAW264.7 cells were harvested and centrifuged at RT for 5 min. The cell pellet was resuspended in 1 mL serum-free culture medium, followed by the addition of 10 μM 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA; Beyotime, China) according to the manufacturer's instructions. Cells were then incubated at 37 °C in the dark for 20 min, after which they were centrifuged again, washed with phosphate-buffered saline (PBS), and resuspended in 500 μL PBS. Finally, the fluorescence intensity of the stained cells was analyzed on a FACSCanto II flow cytometer (BD Biosciences) to quantify intracellular ROS levels. Flow cytometry data were analyzed using FlowJo software (version 10.8.1). A sequential gating strategy was applied to ensure data quality. Initially, the main cell population was identified based on FSC/SSC characteristics to exclude debris and non-cellular events. Doublets were subsequently removed by SSC-A versus SSC-H gating, allowing for the selection of single cells. All subsequent analyses were performed on the resulting single-cell population.

Tartrate-resistant acid phosphatase (TRAP) staining. RAW 264.7 cells were seeded into a 24-well plate (10,000 cells per well), followed by the addition of 500 μL of culture medium containing 10% FBS and receptor activator of RANKL at a concentration of 50 ng/mL. PBS, Cas, CSM, CSM-Mg, CSM-Mn, CSM₃, and CSM₃P were separately added to each well. The cell culture medium was refreshed every three days. On day 5, the differentiation of RAW 264.7 cells into osteoclast-like cells were evaluated through TRAP staining dye liquor (Sigma, USA), and the stained cells were observed and photographed using a microscope.

Alizarin red S staining. MC3T3-E1 cells were seeded into a 24-well plate (10,000 cells per well), followed by the addition of 500 μL of osteogenic differentiation induction medium. Subsequently, PBS, Cas, CSM, CSM-Mg, CSM-Mn, CSM₃, and CSM₃P were separately added to each well. The osteogenic differentiation induction medium was refreshed every three days for a total duration of 21 days. After that, the cells were washed with PBS and fixed with 4% paraformaldehyde. Cellular populations underwent systematic staining following paraformaldehyde fixation according to the manufacturer's instructions (Beyotime, China). The stained cells were then observed and photographed under an optical microscope. The quantitative analysis involved dissolving stained calcium deposition utilizing 1 mL of 10% cetylpyridinium chloride and precise absorbance measurements at 405 nm wavelength.

Alkaline phosphatase (ALP) staining. MC3T3-E1 cells were seeded into a 24-well plate (10,000 cells per well), followed by the addition of 500 μL of osteogenic differentiation induction medium. Subsequently, PBS, Cas, CSM, CSM-Mg, CSM-Mn, CSM₃, and CSM₃P were separately added to each well. The osteogenic differentiation induction medium was refreshed every three days for a total duration of 7 days. ALP activity analysis was performed using the ALP staining kit according to the manufacturer's instructions (Beyotime, China). Then the samples were observed, stained, and photographed using an inverted fluorescence microscope. The percentage of ALP activity in each group was analyzed using ImageJ software.

Quantitative real-time PCR Analysis. Total RNA was isolated from the cells using Trizol Reagent. First-strand cDNA was synthesized by incubating 1 μl of total RNA with oligo dT and reverse transcriptase (Takara, Japan), according to the manufacturer's protocol. The primers are listed in Supplementary Table S1. All qRT-PCR was performed using the SYBR Green PCR superMix (Novoprotein, China) on the Bio-Rad 5-Color System (Bio-Rad, USA). The expression of GAPDH was employed as an internal control to normalization. Relative changes in expression levels were calculated using the 2 ^{-(ΔCT,Tg−ΔCT,control)} method. All analyses were performed in biological triplicates for each sample.

Western blotting. Cells were lysed in RIPA buffer supplemented with protease and phosphatase inhibitor cocktail. The homogenates were then centrifuged at 12,000 g for 20 min at 4 °C. The clear supernatant was boiled in SDS loading buffer. Proteins samples were normalized according to BCA results, separated by SDS-PAGE, and transferred to PVDF membrane. The membrane was blocked with non-fat milk at room temperature (RT) for 2 h, then incubated with the corresponding primary antibodies overnight at 4 °C, followed by incubation with the secondary antibodies at RT for 1 h, with subsequent detection using ECL substrate. The antibodies used in this study are listed in Table S2. Finally, the membrane was scanned using the Quantity One Imaging system (Bio-Rad, USA). For the image quantifications, band densities were assessed using ImageJ software. β-Actin was used as loading control, and the data were presented as fold change to the control group.

Immunofluorescence. The induced cells were seeded onto coverslips and fixed in 4% PFA for 15 min. After overnight incubation, the samples were incubated with primary antibodies at 4 °C overnight and Alexa Fluor® 488-, or 594-conjugated secondary antibodies at RT for 1 h, counterstained with DAPI, then images were acquired using a fluorescence microscope. The immunofluorescence intensity was quantified as the mean signal intensity using ImageJ software.

Transcriptomics analysis. For RNA sequencing library preparation, 1 μg of total RNA was used per sample, following the manufacturer's recommended protocols for the NEBNext UltraTM RNA Library Prep Kit. Index-coded samples were subjected to clustering using the TruSeq PE Cluster Kit v3-cBot-HS (Illumina), after which sequencing was performed on an Illumina Novaseq platform to generate 150 bp paired-end reads. Raw read counts were quantified using featureCounts v1.5.0-p3, and fragments per kilobase of transcript per million mapped reads (FPKM) values were calculated for each gene. Differentially expressed genes (DEGs) were identified using the criteria of |log₂ fold change| ≥ 1 and adjusted p-value <0.05. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses of DEGs were conducted using the g: Profiler platform (accessed in 2025). The background gene set was defined as all expressed genes detected in the RNA-seq dataset after preprocessing. Additionally, Gene Set Enrichment Analysis (GSEA) was employed to compare pathway enrichment patterns across different experimental groups.

Micro-CT scanning and analysis. After sacrificing the mice, the femurs were collected and fixed in 4% PFA. Then, the femurs were analyzed by micro-CT (Skyscan1176, USA) and three-dimensional (3D) reconstruction was performed. The bone morphometric parameters of BMD, BV/TV, Tb.Th, Tb.Sp, Tb.Pf and Tb.N were determined by analyzing the volume of interest (VOI).

Histological evaluation. The fixed bone tissue was gently agitated and decalcified in 10% EDTA solution for 4 weeks. Following decalcification, conventional paraffin embedding was performed, with all specimens maintained in as consistent a position as possible. Sections of 4-μm thickness were cut from these specimens. For histological analysis, H&E and Masson staining were conducted using H&E and Masson staining kits, respectively.

Statistical analysis. Data were analyzed via GraphPad Prism (version 9.0) and represented as the mean ± SD from three or more independent experiments. Statistical analyses were performed using either Student's t-test (two-group comparison) or one-way analysis of variance (ANOVA) followed by Tukey's or least significant difference (LSD) for post hoc test (more than two groups), and differences with P values less than 0.05 were considered to be statistically significant.

3. Results and discussions

3.1. Preparation and characterization of CSM₃P

As shown in Fig. 1a, the pH responsive core-shell nanoplatform of CSM₃P with bone-targeting property was prepared and characterized, based on the CaS core and magnesium/manganese ion (Mg²⁺/Mn²⁺) doped mesoporous silica (mSiO₂) shell. First of all, the CaS core was prepared by a thermal decomposition method with slight modification [35]. The as prepared CaS nanoparticles presented monodisperse and uniform spherical structure with size about 36 nm (Fig. 1b). In order to load the mineral elements that coordinate bone regeneration, the Mg²⁺/Mn²⁺ doped mSiO₂ shell was coated [36]. The as obtained mSiO₂ coated CaS (CaS@MSN-MgMn, CSM₃) showed a core-shell structure with size about 51 nm (Fig. 1c). The mesoporous structure was further confirmed by the N₂ adsorption/desorption measurement (Fig. 1i). Such a structure not only enables the nanoplatform the property of loading therapeutic elements, but also the therapeutic drugs [36]. During the synthesis process, the modification of amino group was conducted to make its zeta potential positive (Fig. S1), which was confirmed by the ninhydrin test (Fig. S2). To endow it bone-targeting property, the bone-targeting ligands of alendronate anchored polyacrylic acid (PA) were adsorbed onto the surface of CSM₃ by electrostatic interaction to form the pH responsive core-shell nanoplatform of CSM₃P. It could be observed from the transmission electron microscope (TEM) that the amination and the PA decoration did not affect the morphology of CSM₃, which could ensure the effective release of therapeutic elements (Fig. 1d and e). In addition, the results of elemental mapping showed that uniform distribution of the major elements of CSM₃P, including the calcium (Ca) and sulfur (S) elements of CaS core, the Mg, Mn and silicon (Si) element of metal ion doped mSiO₂ shell, the specific phosphorus (P) element of bone-targeting ligand (Fig. 1f and S3).

Fig. 1.

Synthesis and characterization of CSM₃P nanoparticles. (a) Schematic illustration of the synthesis of CSM₃P nanoparticles. (b-e) TEM images of CaS, CSM₃, CSM₃-NH₂ and CSM₃P, respectively. Scale bar = 50 nm. (f) Elemental mapping images corresponding to CSM₃P. Scale bar = 100 nm. (g) XRD patterns of CSM₃, CSM₃-NH₂ and CSM₃P, respectively. (h) TGA curves of CSM₃, CSM₃-NH₂ and CSM₃P. (i) N₂ adsorption-desorption isotherms of CSM₃, CSM₃-NH₂ and CSM₃P. (j-N) XPS spectrums of CSM₃P.

Then, the relative characterization was performed. The zeta potentials were measured first. Notably, after the decoration of amino group, the negatively charged CSM₃ has become positively charged. However, by conjugating with PA, the potential of CSM₃P became negative again, further confirming the successful surface modification of CSM₃P in each step (Fig. S1). Furthermore, as shown in Fig. 1g, the crystal structure of nanomaterials obtained in each step was investigated using X-ray diffraction (XRD), the characteristic peaks of CaS were observed in each step, along with a broad peak at 2θ = 22° corresponding to amorphous SiO₂ in ion doped MSN in CSM₃, CSM₃-NH₂ and CSM₃P. In addition, thermogravimetric analysis (TGA) (Fig. 1h) and fourier-transform infrared spectroscopy (FT-IR) (Fig. S4) further confirmed the successful modification of CSM₃P in each step. The nitrogen adsorption measurements (Fig. 1i) revealed the mesoporous structure of CSM₃, CSM₃-NH₂ and CSM₃P, with pores diameter about 4 nm (Fig. S5). The x-ray photoelectron spectroscopy (XPS) analysis demonstrated the successful synthesis and modification of CaS, which was confirmed by the characteristic peak of Ca 2p and S 2p from CaS, Si 2p, Mg 1s and Mn 1s peaks from MSN, and P 2p peaks from PA (Fig. 1j–n).

3.2. Assessment of the mineral ions and H₂S release property of CSM₃P

A hallmark feature of osteoporotic bone is its acidic microenvironment, which not only stimulates osteoclast activity and aggravates the bone microenvironment, but also impairs osteoblast function, which would ultimately create a vicious cycle that accelerates osteoporosis progression. However, this acidic microenvironment offers a novel idea for the application of pH-responsive nanomedicine. Both CaS and metal-ion-doped mSiO₂ could degrade gradually in acidic environments, showing excellent smart responsiveness and self-adaptability [36,37]. Under acidic conditions, the Si-O-Mn and Si-O-Mg bonds on the surface of CSM₃P gradually break, releasing Mg²⁺ and Mn²⁺ ions, followed by hydrolysis of the internal CaS to release H₂S and Ca²⁺, ultimately leading to the full degradation of CSM₃P and the release of therapeutic mineral ion and H₂S gas (Fig. 2a). Then, the pH-responsive ability of CSM₃P was determined.

Fig. 2.

Acidic responsive degradation of CSM₃P. (a) Scheme illustration of pH -responsive sequential degradation of CSM₃P. (b-d) Release profiles of H₂S at different conditions. (e-g) Release profiles of Mg²⁺, Mn²⁺ and Ca²⁺ from CSM₃P at different conditions. (h) Representative TEM images of CSM₃P after different times of incubation under different conditions. Scale bar = 100 nm. (i-k) The ROS scavenging efficiency of CSM₃P for H₂O₂, ·OH and O₂^·-.

First of all, the H₂S release kinetics was explored. As shown in Fig. 2b–d, the H₂S release kinetics was detected firstly. It could find that in alkaline (pH 8.5) and neutral (pH 7.4) environments within 6 h, only 0.168 mM and 0.311 mM of H₂S released from CaS, whereas the concentration increased to 0.467 mM when pH was set at 6.5 (Fig. 2b). A similar release profile could be found in CSM₃P (Fig. 2c and S6). Furthermore, the H₂S release kinetics was detected among CaS, CSM, CSM₃, and CSM₃P when pH was set at 7.4, it could be found that H₂S release curves varied significantly (Fig. 2d). CaS exhibited a burst release within 6 h, while coating with Mg²⁺/Mn²⁺-doped mSiO₂ (CSM₃ and CSM₃P) resulted in slower and more sustained H₂S release due to the gradual degradation of the shell, with the concentration within the safe range [[38], [39], [40]]. In contrast, undoped CSM showed the slowest H₂S release, attributed to the low degradability of pure mSiO₂ hindering CaS hydrolysis.

To quantitatively characterize the cascade release behavior, we analyzed the cumulative release profiles of Mg²⁺, Mn²⁺, and Ca²⁺ under different pH conditions (Fig. 2e–g). When pH was set at 6.5, 80.05% of Mg²⁺ was released within 96 h, in a burst manner, while release was much slower at pH 7.4 and 8.5. Similar pH-dependent kinetics were observed for Mn²⁺ and Ca²⁺. Based on these profiles, three temporal stages can be identified: (i) an early phase (0-6 h) with rapid release of Mg²⁺ and Mn²⁺ from the doped silica shell; (ii) an intermediate phase (6-24 h) where Ca²⁺ and H₂S release becomes detectable as the CaS core is exposed; and (iii) a sustained phase (24-96 h) characterized by continuous release of all therapeutic species due to gradual hydrolysis of both the remaining shell and the core. Notably, at pH 6.5, the cumulative release of Mg²⁺ and Mn²⁺ exceeded 50% within the first 24 h, whereas Ca²⁺ release remained below 15% during the same period. This clear temporal offset between shell-derived and core-derived ions strongly supports the sequential degradation model rather than a simultaneous burst release.

To correlate degradation behavior with release kinetics, CSM₃P nanoparticles were collected at different time points under varying pH conditions and examined by TEM (Fig. 2h). At pH 6.5, the shell rapidly disintegrated initially, followed by gradual core dissolution and eventual structural collapse. In contrast, at pH 7.4, the core-shell structure still remained over 24 h, consistent with the slow-release kinetics observed under neutral conditions. In addition, the adaptive responsiveness of CSM₃P was further evaluated under fluctuating pH conditions. As shown in Fig. S7a, incubation at pH 4.5 led to gradual pH increase over 24 h; however, upon re-acidification to pH 4.5, pH increased again, demonstrating reproducible pH responsive adaptability of CSM₃P. Furthermore, bone slices were used to mimic the in vivo bone microenvironment. It could be found that the FITC fluorescence signals on the bone slice diminished slowly under neutral and alkaline conditions but decreased rapidly after the addition of an acidic buffer, indicating rapid degradation of the CSM₃P and its acid responsiveness (Fig. S7b). Importantly, the acidic osteoporotic microenvironment is sustained by continuous osteoclastic proton secretion and localized resorption lacunae (pH lower than 4.5), which persist independently of inflammatory status [41,42]. Since H₂S modulates intracellular signaling rather than directly neutralizing extracellular pH, its anti-inflammatory action does not eliminate local acidity. Thus, sufficient acidity remains to sustain nanoparticle degradation throughout active bone remodeling, establishing a self-adaptive, pH modulated release system consistent with the proposed self-reinforcing therapeutic cycle.

The occurrence and progression of osteoporosis are often accompanied by the generation of high levels of reactive oxygen species (ROS) and the release of inflammatory cytokines. Therefore, we further tested the ability of CSM₃P in eliminating ROS including H₂O₂, ·OH and O₂⁻. It could be found that the scavenging rates increased in a concentration-dependent manner (Fig. 2i–k). In addition, electron spin resonance (ESR) spectroscopy was used and confirmed it ROS scavenging ability (Fig. S8). The above results demonstrated the excellent scavenging ability of CSM₃P for various ROS, which could provide a favorable microenvironment foundation for the treatment of osteoporosis.

3.3. CSM₃P drives M2 Polarization of Macrophages and Inhibits Osteoclastogenesis in Vitro

The occurrence of osteoporosis is often accompanied by inflammatory cells infiltration and the secretion of numerous inflammatory factors, leading to the gradual deterioration of the bone microenvironment and enhancing osteoclast formation and bone resorptive activity. It was found that macrophages sense damage-associated molecular patterns and transition into a proinflammatory state, exacerbating bone resorption; Therefore, the shift between M1 and M2 macrophages and their homeostasis are crucial for efficient bone tissue repair processes and provide a favorable 'soil' for bone regeneration [11,43].

Initially, the biocompatibility of CSM₃P was evaluated using CCK-8 cell viability and live/dead staining assays. After 72 h of incubation, no significant cytotoxicity was observed in MC3T3-E1 osteoblastic cells across all tested concentrations or material groups (Fig. S9a and b). A slight decrease in cell viability was observed at concentrations above 50 μg/mL; however, cell viability remained above 80% across all tested concentrations, indicating acceptable cytocompatibility (Fig. S9b). Consistently, live/dead staining further confirmed high cell viability in both RAW264.7 macrophages and MC3T3-E1 osteoblasts across all treatment groups (Fig. S9c). Besides, results from the Annexin V-PI apoptosis assay also revealed that treatment with CSM₃P hardly induced cell apoptosis (Fig. S10). To accurately assess the effects of CSM₃P on macrophage polarization, RAW 264.7 were then induced with lipopolysaccharide (LPS) and co-cultured with the nanoparticles obtained at different steps (Fig. 3a). Next, the homeostasis and reactivity of macrophages affected by nanoparticles was investigated by measuring phenotypic changes associated with morphological activation and the levels of inflammatory factors. CD86 staining identified pro-inflammatory macrophages, while CD206 staining marked their reparative counterparts, which is of crucial importance for regulating bone regeneration [10]. As determined by immunofluorescence staining, compared with other groups, the group of CSM₃ not only induced a significant downregulation of the M1 macrophage marker CD86 but also a marked upregulation of the M2 marker CD206, as well as the CSM₃P group, indicating that CSM₃ and CSM₃P could effectively activate the phenotypic transformation of macrophages from M1 to M2 (Fig. 3b, c, and S11a, b). These results were also confirmed by the protein expression and corresponding quantification of pro-inflammatory and anti-inflammatory factors. CSM₃P treatment significantly reduced the protein expression of pro-inflammatory cytokines in LPS-induced macrophages, while promoting the protein expression of anti-inflammatory cytokines associated with M2 polarization (Figs. S12 and 13). Furthermore, oxidative stress and reactive oxygen species (ROS) levels are involved in regulating inflammation and immunity, and persistent ROS production induces redox imbalance, impeding the phenotypic transformation of M1 to M2 anti-inflammatory macrophages and thereby prolonging the inflammatory process [44,45]. Thus, the ROS scavenging capacity were evaluated and the 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) staining results revealed that the treatment with CaS, CSM-Mg, CSM-Mn, CSM₃, and CSM₃P could markedly reduce LPS-induced excessive intracellular ROS accumulation in macrophages, which was consistent with the results from precise quantification of positive cells via flow cytometry (Fig. 3d, e, and S14).

Fig. 3.

CSM₃P Drives M2 Polarization of Macrophages and Inhibits Osteoclastogenesis in Vitro. (a) Schematic representation illustrating the regulatory effect of CSM₃P on macrophage polarization and osteoclastic inhibition. (b) Representative immunofluorescence staining of F4/80 (red), CD86 (green), and DAPI (blue) in RAW 264.7 macrophages treated with Control, LPS, CaS, CSM, CSM-Mg, CSM-Mn, CSM₃, or CSM₃P. Scale bar = 50 μm. (c) Shown in corresponding expression heatmaps of CD86. (d) DCFH-DA staining to detect intracellular ROS levels in macrophages under different treatments. Scale bar = 100 μm. (e) Quantitative analysis of ROS production (DCF-positive percentage). (f) Representative TRAP staining of osteoclasts differentiated from RAW 264.7 cells stimulated with RANKL, under various treatments. Scale bar = 200 μm. (g, h) Western blotting results showed the protein expression levels and quantification of TRAF6, CTSK, and β-actin (loading control) in osteoclasts. (n = 3 per group; Values represented mean ± SD, ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001).

Notably, as important precursors of osteoclasts, macrophages would differentiate and then fuse into multinucleated mature osteoclasts in the adverse bone microenvironment of osteoporosis [46]. By disrupting the acidic and inflammatory microenvironment that favors osteoclast growth, along with inhibiting osteoclast activity, it would effectively reduce bone resorption in osteoporotic conditions. To this end, the inhibitory effect of CSM₃P nanoparticles on the formation and activity of osteoclasts derived from macrophage differentiation was evaluated via tartrate-resistant acid phosphatase (TRAP) staining. While macrophages were first induced to form mature multinucleated osteoclasts upon RANKL cytokines stimulation, CSM₃P nanoparticles obviously suppressed osteoclastogenesis, characterized by fewer TRAP-positive osteoclast numbers with smaller average volume (Fig. 3f–S15a). Moreover, CSM₃P nanoparticles treatment also inhibited RANKL-induced expression of TRAF and CTSK (Fig. 3g, h, and S15b), the critical marker of osteoclastogenesis. Notably, although CSM alone lacked anti-osteoclastogenic activity, Mg²⁺ or Mn²⁺ doping enables CSM to suppress this process. Among them, the effect of Mn²⁺ was more significant, indicating that Mn²⁺ can effectively enhance the ability of the nanoplatform to interfere with osteoclast differentiation [47]. Altogether, the functionalized “nanobomb” not only inhibited osteoclast activity but also regulated the surrounding cellular state by releasing metal ion and H₂S, enabling timely switching towards an anti-inflammatory/reparative microenvironment that favors bone regeneration.

3.4. CSM₃P Enhances Osteogenic Differentiation in Vitro

Osteocytes are involved in intricate signaling transduction to facilitate communication with a range of cell types, such as osteoblasts, osteoclasts, and immune cells, thereby enabling the regulation of the growth, repair, and remodeling of the skeletal system [48]. Hence, to investigate the ability of CSM₃P to promote osteogenic differentiation, we conducted further analysis by closely monitoring the osteogenic differentiation status of MC3T3-E1 cells after culturing and exposure to osteogenic differentiation medium (Fig. 4a). The alkaline phosphatase (ALP) activity serves as an early indicator for the commencement of osteoblast differentiation, while the level of mineralized nodule formation within the extracellular matrix directly manifests the extent of this differentiation process. As shown in Fig. 4b and d, MC3T3-E1 cells treated with CSM₃P nanoparticles exhibited significantly higher ALP activity than the control group, with a similar trend observed in the formation of extracellular matrix mineralized nodules in MC3T3-E1 cells (Fig. 4c–e). The enhanced mineralization observed in the CSM₃P group is likely due to the combined effects of Ca²⁺/Mg²⁺/Mn²⁺ release and the formation of H₂S, which create a more favorable environment for extracellular matrix maturation and mineral deposition, indicating the significance of microenvironment in the bone formation [49]. Subsequently, Western Blotting (WB) was utilized to evaluate the expression of Runt-related transcription factor 2 (RUNX2) and Osteopontin (OPN). The expression trends of these two pivotal regulating osteogenic differentiation-related proteins were consistent with the ALP and ARS staining results (Fig. 4f, g, and S16a, b). The enhanced osteoinductive microenvironment created by this synergistic effect was also reflected in the elevated expression of other osteogenesis-related genes, as demonstrated by mRNA analysis (Fig. S16c–e).

Fig. 4.

CSM₃P Enhances Osteogenic Differentiation in Vitro. (a) Illustration of CSM₃P treatment of MC3T3-E1 cells for quantitative polymerase chain reaction (q-PCR), Western blot analysis, cell staining and Immunofluorescence. (b) Quantification of ALP activity and (c) mineralization. (d) Alkaline phosphatase (ALP) staining of MC3T3-E1 cells treated with Control, CaS, CSM, CSM-Mg, CSM-Mn, CSM₃, or CSM₃P. Top panel: whole-well view; bottom panel: magnified view. Scale bars = 1 mm (top), 200 μm (bottom). (e) Alizarin red staining to assess mineralized nodule formation in MC3T3-E1 cells under different treatments. Top panel: whole-well view; bottom panel: magnified view. Scale bars = 1 mm (top), 200 μm (bottom). (f, g) Western blotting results showed the protein expression levels and quantification of RUNX2, OPN, and β-actin (loading control) in MC3T3-E1 cells. (h) Representative immunofluorescence staining images of OPN (green) and DAPI (blue) in MC3T3-E1 cells. Scale bar = 50 μm. (i) Quantitative analysis of OPN fluorescence intensity from immunofluorescence staining. (j) Schematic diagram of the effect of CSM₃P in promoting osteogenesis and inhibiting osteoclasts. (n = 3 per group; all data shown as mean ± SD, ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001).

Furthermore, the expressions of OPN and Collagen type I alpha 1 (COL1A1) in MC3T3-E1 cells following 14 days of osteogenic induction were assessed via immunofluorescence. OPN is a phosphorylated glycoprotein with high affinity for calcium, and is recognized as a structural component of the bone matrix, which is thought to promote or regulate collagen fiber formation by controlling the adhesion of osteoclasts and osteoblasts to bone surfaces during remodeling and plays a key role in the resorption and formation of collagen matrices [[50], [51], [52]]. COL1A1 is a predominant component of the extracellular matrix in bone, which constitutes over 90% of the organic matrix, which could synergistically regulate the formation and homeostasis maintenance of bone tissue during osteoblast differentiation and bone mineralization [53]. As shown in Fig. 4h and S17a, compared with other groups, the cells treated with CSM₃P nanoplatform dispalyed enhanced OPN and COL1A1 fluorescence intensity relatives, and the corresponding quantitative analysis of fluorescence intensity further confirmed these results, indicating superior performance in promoting osteoblast differentiation (Fig. 4i–S17b). These findings provide compelling evidence that CSM₃P inhibits osteoclast activity while concurrently facilitating osteogenic differentiation, offering insights into its potential therapeutic applications in bone regeneration (Fig. 4j). These results indicate that mineral elements and hydrogen sulfide play an important role in bone regeneration, as well as integrating multiple strategies, which would lay a solid foundation for their further in vivo applications [10,30,54].

3.5. Transcriptional Profiling Revealed Differentially Expressed Genes in CSM₃P-Treated cells

Although CSM₃P has excellent anti-inflammatory properties, the ability to inhibit osteoclasts and the function of coordinating bone formation, the specific molecular mechanism by which it regulates these functions remains unclear. Therefore, we further conducted transcriptomic analysis to deeply explore the underlying mechanism of its action at the gene expression level, providing a more precise molecular basis for comprehensive understanding of its biological effects. First, the anti-inflammatory properties were investigated. From the heatmap of the macrophage group, distinct clusters of gene expression differences were observed between the CSM₃P-treated group and the control group (Fig. S18a). A total of 10878 transcripts were detected in the RNA-seq dataset after preprocessing. Compared with the LPS-treated group, 1633 genes were upregulated and 2239 were downregulated in the CSM3P-treated group, representing 36 % of all detected transcripts (Fig. S18b). To further characterize the biological functions of these DEGs, Gene Ontology (GO) enrichment analysis revealed significant enrichment in inflammation-related processes, including inflammatory response and canonical nuclear factor κB (NF-κB) signaling (Fig. S18c). Consistently, Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis demonstrated that DEGs were significantly enriched in NF-κB and TNF signaling pathways, which are key regulators of inflammatory responses(Fig. S18d). Notably, among the total 3872 DEGs, 254 genes (∼6.56%) were annotated as inflammation-related based on functional classification, of which 250 genes (98.42%) were significantly differentially expressed. Moreover, the majority of these inflammation-related genes were downregulated (67.72%), indicating a potential suppression of inflammatory responses by CSM₃P. Among these, the NF-κB signaling cascade represents a central axis in immune regulation, playing critical roles in inflammation, cell proliferation, differentiation, and apoptosis [55]. For visualization purposes, representative genes were selected from significantly enriched pathways (e.g., NF-κB signaling and macrophage polarization) based on their statistical significance and biological relevance. These genes do not represent the complete set of DEGs but were chosen to illustrate key regulatory patterns within the identified pathways (Fig. S18e). Given that enrichment analysis highlighted pathways related to macrophage inflammation, polarization, and NF-κB signaling, we further performed gene set enrichment analysis (GSEA) using a ranked gene list. The results showed that the NF-κB signaling pathway was negatively enriched in the CSM₃P-treated group compared with the LPS-treated group, as indicated by the running enrichment score, which increased in the upregulated region and subsequently decreased across the ranked gene list. Leading-edge analysis further showed that core genes driving the enrichment were predominantly proinflammatory genes. These findings support that CSM₃P suppresses NF-κB pathway activity, thereby inhibiting M1 macrophage polarization and inflammatory responses (Fig. S18f and g). To validate RNA-seq results and explore the expression of inflammation-related genes at the mRNA levels, macrophages from each group were examined by qRT-PCR. Compared with the control group, the mRNA expression levels of three most representative pro-inflammatory cytokines, namely TNF-α, IL-1β, and IL-6, were significantly decreased in CSM₃P-treated macrophages; likewise, the expression of the anti-inflammatory cytokines IL-4 and IL-10 was significantly increased, highlighting their substantial inhibitory effects on inflammatory responses (Fig. S18h and i).

Subsequently, we interrogated the underlying mechanisms through which CSM₃P nanoparticles modulates osteoclast differentiation of RAW 264.7 cells. The heatmap highlights DEGs between the CSM₃P-treated and the control groups, with 4718 significantly altered genes in the CSM₃P-exposed cohort (Fig. S19a and b). GO enrichment analysis indicated that genes associated with the MAPK cascade and canonical NF-κB signaling were significantly downregulated in the CSM₃P-treated group, including representative regulators such as TRAF family members, ACP5, and MMP9 (Fig. S19c). KEGG enrichment analysis further showed that DEGs were significantly enriched in the MAPK signaling pathway, which is closely linked to osteoclast differentiation and maturation, as well as in the inflammation-associated TNF signaling pathway (Fig. S19d). Radar plot visualization and GSEA consistently demonstrated that genes involved in MAPK signaling-particularly those regulating osteoclast differentiation and inflammatory amplification-were predominantly downregulated in response to CSM₃P treatment (Fig. S19e–g). At the mRNA level, CSM₃P treatment robustly suppressed the expression of CTSK and ACP5-canonical marker enzymes of osteoclasts-and MMP9, a matrix metalloproteinase functionally implicated in osteoclast migration and invasion (Fig. S19h–j). These findings demonstrate that CSM₃P nanoparticles treatment effectively suppresses osteoclast activity and attenuates inflammatory signaling pathways, thereby creating a microenvironment conducive to bone remodeling through this dual regulatory mechanism.

Mechanistically, Mg²⁺ and Mn²⁺ in CSM₃P exert synergistic effects on bone regeneration. Mg²⁺ promotes osteoblast adhesion, proliferation, and angiogenesis via integrin/PI3K/Akt/eNOS signaling and stimulates pro-osteogenic exosome miR-196a-5p targeting Hoxa7/MAPK axis [56,57]. Mn²⁺ activates Wnt/β catenin while inhibiting NF-κB, scavenges ROS via Nrf2, and promotes M2 macrophage polarization through Nrf2/HO-1 and TGF-β/Smad pathways [47]. Transcriptomic data confirmed Wnt/BMP upregulation and NF-κB/MAPK suppression, indicating cooperative anabolic and anti-catabolic actions. In addition, CSM₃P establishes a causal link between immunomodulation and osteogenesis. H₂S inhibits NF-κB/MAPK, drives M2 polarization, and together with Mg²⁺/Mn²⁺, creates an anti-inflammatory milieu. M2 secreted IL-10 and TGF-β amplify osteoblast differentiation, while Mg²⁺/Mn²⁺ directly activate Wnt/BMP pathways. Thus, initial anti-inflammatory action sensitizes bone cells to osteogenic cues, forming a self-reinforcing cycle consistent with osteoimmune paradigms.

Finally, and most importantly, we investigated the factors governing the regulation of MC3T3-E1 osteoblast activation by CSM₃P nanoparticles. Hierarchical clustering analysis showed a clear distinction between the two groups (Fig. 5a). Through volcano plot analysis, we identified 1078 significantly upregulated genes and 1091 downregulated genes in the CSM₃P-treated group (Fig. 5b). GO enrichment analysis demonstrated significant enrichment in biological processes related to Wnt-mediated cell–cell signaling and cell activation in the CSM₃P-treated group, suggesting enhanced osteoblast-related activity (Fig. 5c). Further analysis showed that representative genes involved in osteoblast differentiation and bone formation, such as Wnt family member 7a (Wnt7a) and bone morphogenetic protein 8a (BMP8a), were significantly upregulated. KEGG pathway analysis identified the Wnt signaling pathway and the PI3K-Akt signaling pathway as key enriched pathways (Fig. 5d). Consistently, genes associated with osteoblast function and bone remodeling, including Wnt3a, BMP8a, Bgn, Lgr6, and Wnt7b, were upregulated, whereas inflammation-related genes such as Cx3cl1, Fes, and IL-1β were downregulated (Fig. 5e). GSEA further confirmed the enrichment of the Wnt signaling pathway in the CSM₃P-treated group, indicating a global upregulation of Wnt-related gene expression (Fig. 5f). Notably, the upregulation of Wnt7b and activation of the canonical Wnt signaling pathway were further highlighted for subsequent mechanistic investigation (Fig. 5g). In the osteoporotic microenvironment, Wnt7b promotes osteoblast differentiation and bone formation by regulating the canonical Wnt/β-catenin pathway, while enhancing the expression of key osteogenic markers including Runx2 and OPN [58]. Furthermore, it can directly downregulate RANKL transcription in osteocytes via noncanonical Wnt pathways, thereby inhibiting osteoclastogenic signaling [59]. Assessment of mRNA levels of genes associated with osteoblast differentiation and maturation via PCR revealed consistency with transcriptomic results, further supporting that CSM₃P promotes the bone remodeling process by enhancing osteogenic activation (Fig. 5h–j).

Fig. 5.

Transcriptional Profiling Revealed Differentially Expressed Genes in CSM₃P-Treated Cells. (a) Heatmap of differentially expressed genes (DEGs) in MC3T3-E1 cells treated with Control or CSM₃P. Red indicates upregulation, and blue indicates downregulation. (b) Volcano plot depicted DEGs between Control and CSM₃P groups. Red dots represent significantly upregulated genes, blue dots represent significantly downregulated genes, and gray dots represent non-significant genes. (c) Circular chord diagram showed the functional classification of DEGs in various biological pathways. The width of the chords corresponds to the number of genes associated with each pathway. (d) Enrichment bubble plot of KEGG pathways, with bubble size indicating gene count and color representing Q-value. (e) Differential gene radar map of transcriptomic data in the Control and CSM₃P groups. (f) GSEA plot for the Wnt signaling pathway, showing significant upregulation in CSM₃P-treated cells. (g) Schematic diagram illustrates the proposed mechanism of CSM₃P in activating the Wnt7b and Wnt pathways. (h-j) Quantitative real-time PCR analysis of RUNX2 (h), OPN (i), and Wnt7b (j) mRNA expression in MC3T3-E1 cells under different treatments. Data are presented as relative mRNA expression. (n = 3 per group; all data shown as mean ± SD, ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001).

Obviously, bone-targeted and acid-responsive CSM₃P delivers a superior multi-target therapy by integrating acid neutralization, anti-resorption, osteogenesis promotion, and anti-inflammatory effects, which was rooted in its core innovation of the rocket-inspired "booster-payload" design coupled with pathology-guided functional synergy. This design integrates three key methodological optimizations: COS-independent H₂S release (acid-triggered CaS decomposition), kinetically matched cascade release (aligned with disease progression), and targeting-function compatibility (high bone targeting without compromising pH responsiveness), resolving inherent limitations of individual components. This combined approach effectively overcomes the limitations of single-mechanism anti-resorption agents such as bisphosphonates (which have limited efficacy and may cause severe side effects like osteonecrosis of the jaw) and denosumab (which carries risks of hypocalcemia and rapid bone loss after discontinuation), while achieving enhanced cost-effectiveness through low-cost materials and reduced dosing frequency [5].

3.6. CSM₃P Administration Effectively Attenuates Bone Loss and Osteoporosis in OVX mice

As estrogen deficiency disrupt the dynamic balance between osteoclasts and osteoblasts, we performed bilateral ovariectomy (OVX) in female mice to simulate the estrogen loss in postmenopausal women (Fig. 6a). Before in vivo therapeutic efficacy test, the bone-targeting ability of CSM₃P was evaluated, to ensure its effect of the clinical application. Due to the ability of chelating calcium ions which was rich in the bone microenvironment, the hydroxyapatite (HAP) and mice was chosen for in vitro and in vivo experiment model, respectively [2]. It could be seen in Fig. S20a, the CSM₃P targeted HAP at different time points was taken out for scanning electron microscopy (SEM) characterization, and the results showed that the number of CSM₃P nanoparticles on the surface of HAP increased over time, and the targeting rate at 6 h was up to 80.14% (Fig. S20b). However, osteoporosis is a systematic bone disorder, and the in vivo bone targeting property is of vital importance. After the intraperitoneal injection of cy5 labeled nanomaterials for a certain time, the femurs were collected for fluorescence imaging. It could be found that hardly could see the fluorescence in the groups without Ald modification (Sham, CaS, CSM, CSM-Mg, CSM-Mn and CSM₃), but significant fluorescence was observed in the CSM₃P group (Fig. 6b–S21a). In addition, for CSM₃P alone, the fluorescence intensity increased over a 3-h period and then decreased, this may be due to the degradation of CSM₃P in bone tissue (Fig. 6c–S21b). At the same time, fluorescence images confirmed that the main accumulation organs outside bone tissue were the liver and kidney (Fig. S22), which were the primary organs of metabolism in vivo. These results demonstrated that the modification of Ald could endow the ion/gas nanoplatform bone-targeting property and reduce the side effects, which would be beneficial for the treatment of osteoporosis. In addition, the in vivo biocompatibility was conducted. The results of hemolysis test were less than 5% (Fig. S23a). After a one-month treatment period, histological examination of major organs revealed preserved tissue integrity without evident pathological abnormalities compared with normal tissues (Fig. S23b). Consistently, no significant alterations were observed in serum biochemical indicators, including alanine aminotransferase, aspartate aminotransferase, and blood urea nitrogen (Fig. S23c–f), indicating favorable biosafety of CSM₃P during prolonged administration. These favorable acute safety results suggested satisfactory biocompatibility of CSM₃P nanoparticles in vivo and providing a solid guarantee for its in vivo application.

Fig. 6.

CSM₃P Administration Effectively Attenuates Bone Loss and Osteoporosis in OVX Mice. (a) Schematic timeline of the in vivo study, including mouse acclimatization, ovariectomy (OVX), intraperitoneal (i.p.) injection of formulations, and subsequent sacrifice with organ harvesting and analysis. (b) In vivo fluorescence imaging of femurs from OVX mice at 4 weeks post-treatment with different groups, indicating the distribution and retention of CSM₃P. (c) Time-dependent in vivo fluorescence imaging of CSM₃P in OVX mouse femurs, demonstrating its spatiotemporal accumulation in the bone microenvironment. (d) Representative micro-computed tomography (μCT) images of femurs from Sham, OVX, Cas, CSM, CSM-Mg, CSM-Mn, CSM, or CSM₃P-treated OVX mice. Top panel: 3D reconstructions of whole femurs; bottom panel: 3D reconstructions of trabecular bone regions. (e-j) Quantitative μCT analysis of bone mineral density (BMD, e), bone volume over total volume (BV/TV, f), trabecular number (Tb.N, g), trabecular thickness (Tb.Th, h), trabecular separation (Tb.Sp, i), and trabecular pattern factor (Tb.Pf, j). (n = 5 mice per group; all data shown as mean ± SD, ∗∗∗P < 0.001).

Upon completion of the treatment, the bone microstructure was first evaluated accurately and intuitively via microcomputed tomography (micro-CT), followed by three-dimensional (3D) bone reconstruction analysis. As illustrated in Fig. 6d, the reconstructed 3D models of osteoporotic bone demonstrated that the trabecular bone volume was reduced in mice from the OVX group compared with the Sham group, whereas bone loss and osteopenic phenotypes in trabecular bones were significantly alleviated in mice from the CSM₃P group. Apparently, the bone microarchitecture parameters including bone mineral density (BMD), bone volume fraction (BV/TV), trabecular number (Tb. N), trabecular thickness (Tb. Th), trabecular separation (Tb. Sp), and trabecular pattern factor (Tb. Pf) were largely improved after CSM₃P treatment (Fig. 6e–j). Interestingly, CSM₃ did not exhibit a significant osteoporotic bone-healing effect, which may be attributed to the absence of the bone-targeting capacity. Most importantly, the key parameters mentioned above were altered to levels comparable to those in sham-operated mice following CSM₃P treatment in OVX mice, highlighting its superior efficacy in reversing bone loss.

Apart from micro-CT analysis, various staining was carried out to confirm its excellent anti-osteoporosis ability. As shown in Fig. 7a, compared with the Sham group, the regular reticular structure of the trabecular bone of the OVX group was disrupted, characterized by sparse and irregular arrangement. Similar results could be found in the CaS, CSM, CSM-Mg, CSM-Mn and CSM₃ group, due to the lack of bone-targeting property. However, benefiting from its bone-targeting property, potent immunomodulatory properties and spatiotemporally adaptive feedback-driven capabilities, the CSM₃P treatment maintained nearly intact trabecular bone mass and an improved bone microstructure, with trabeculae arranged in an orderly manner. Masson's trichrome staining further confirmed that CSM₃P significantly increased collagen content and promoted more new bone formation (Fig. 7b). Besides, it should be noted that the rebalancing of osteoblasts and osteoclasts in the osteoporotic microenvironment also merits attention, as the sustained overactivation of osteoclasts is detrimental to new bone formation [60]. Then, the TRAP immunohistochemical staining was conducted and the results confirmed that CSM₃P could maximally suppress the activity and number of osteoclasts, returning them to nearly normal levels (Fig. 7c–e). In addition, immunofluorescence staining of CSTK showed a decreased expression of the femoral tissues of OVX mice treated with CSM₃P (Fig. 7h and S24). These findings suggested that CSM₃P nanoparticles could effectively disrupt osteoclast maturation in the inflammatory microenvironment. Moreover, histological staining of femur sections from OVX mice for OPN and COL1A1 revealed a marked upregulation of these two osteogenic proteins following CSM₃P treatment, which was consistent with the observations of new bone formation in the Sham group, demonstrating the excellent osteogenic capability of CSM₃P in vivo (Fig. 7d–f, and g). In contrast, in CaS, CSM, CSM-Mg, CSM-Mn and CSM₃ group, there was an increase in osteoclastogenesis and enhanced osteoclast activity, along with a weakened osteoblast activity, demonstrating that the delivery of CSM₃P shapes a more favorable niche for the rebalancing of osteogenesis and osteoclastogenesis, thereby restoring osteogenic potential under osteoporotic conditions.

Fig. 7.

CSM₃P Administration Effectively Attenuates Bone Loss and Osteoporosis in OVX Mice. (a) Representative hematoxylin and eosin (H&E) staining of femur sections showed bone tissue morphology. Top panel: low-magnification view; bottom panel: magnified view of the boxed region. Scale bar = 200 μm. (b) Representative Masson's trichrome staining of femur sections visualized collagen deposition (blue) and bone matrix. Top panel: low-magnification view; bottom panel: magnified view of the boxed region. Scale bar = 200 μm. (c) Representative TRAP staining of femur sections, with red arrows indicating TRAP-positive osteoclasts. Scale bar = 200 μm. (d) Immunohistochemical staining of osteopontin (OPN, top) and type I collagen (COL1A1, bottom) in femur sections, marking osteogenic activity. Scale bar = 200 μm. (e-g) Quantitative analysis of TRAP-positive osteoclast surface over bone surface (e), OPN-positive fraction (f), and COL1A1-positive fraction (g) in respective stained sections. (h) Representative immunofluorescence staining of CTSK, red) and DAPI (blue) in femur sections. Merge panel showed co-localization. Scale bar = 200 μm. (n = 5 mice per group; all data shown as mean ± SD).

4. Conclusion

In this work, inspired by the concept of explosive earth-penetrator weapons, a novel pH-responsive bone-targeting nanoplatform (CSM₃P) was successfully developed and prepared for the treatment of osteoporosis, with core innovation lying in pathology-guided functional synergy. CSM₃P nanoplatform leverages the acidic microenvironment of osteoporotic bone to trigger a controlled release of mineral ions (Mg²⁺, Mn²⁺, Ca²⁺) and hydrogen sulfide (H₂S), modulating the inflammatory bone microenvironment and rebalancing the metabolic balance between osteogenesis and osteolysis. It bears out that CSM₃P could effectively polarize macrophages towards an M2 phenotype, inhibits osteoclastogenesis, and enhances osteoblast differentiation. Transcriptomic analysis and validation experiments revealed that these beneficial effects are mediated by the regulation of key signaling pathways, including the NF-κB, MAPK, and Wnt pathways. Furthermore, the in vivo experiments using an OVX model confirmed the therapeutic efficacy of CSM₃P in attenuating bone loss and improving bone microarchitecture. The alendronate modification endowed the nanoplatform bone-targeting property and facilitated preferential accumulation of the nanoplatform in bone tissue, minimizing off-target effects, as well as maximizing therapeutic impact. By integrating ion/gas therapy and bone-targeting ability, CSM₃P offers a self-reinforcing therapeutic cycle that adaptively responds to the pathological acidity of the osteoporotic niche. CSM₃P rebalances bone metabolism by coordinately modulating the inflammatory microenvironment (via H₂S-driven M2 polarization and NF-κB inhibition) and directly promoting osteogenesis (via Mg²⁺/Mn²⁺-activated Wnt/BMP pathways), thereby establishing a self-reinforcing therapeutic cycle. This work provides a compelling paradigm for intelligent, spatiotemporally adaptive nanotherapy, with potential applications extending beyond osteoporosis to other microenvironment-associated diseases. Future research should focus on optimizing the nanoplatform's design, exploring its long-term safety profile, and investigating its efficacy in combination with existing osteoporosis treatments. It should be noted that further exploration in larger animal models and ultimately human clinical trials will be necessary to verify long-term biocompatibility and effectiveness, as well as dose-dependent relationships. Future research should focus on optimizing the nanoplatform's design, exploring its long-term safety profile in chronic models, and investigating its efficacy in combination with existing osteoporosis treatments.

CRediT authorship contribution statement

Wei Xiong: Writing – review & editing, Writing – original draft, Project administration, Methodology, Funding acquisition, Data curation, Conceptualization. Yiming Hu: Validation, Software, Methodology, Investigation, Data curation. Changxiong Cai: Writing – original draft, Validation, Project administration, Methodology, Investigation, Formal analysis, Data curation. Xingyun Fang: Validation, Software, Methodology, Investigation, Data curation. Ye Liu: Methodology, Investigation, Conceptualization. Xiangyu Tian: Validation, Software, Project administration. Qingsong Ye: Writing – review & editing, Methodology, Data curation, Conceptualization. Xigao Cheng: Project administration, Funding acquisition, Data curation. Jing Ye: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Project administration, Funding acquisition, Conceptualization.

Ethics approval and consent to participate

All animal experiments were reviewed and approved by the Ethics Committee and Institutional Review Board of Nanchang University (Approval No. NCULAE-20221228021).

Declaration of competing interest

Qingsong Ye is an editorial board member for Bioactive Materials and was not involved in the editorial review or the decision to publish this article. All authors declare that there are no competing interests.

Appendix A. Supplementary data

The following are the Supplementary data to this article.