Ethical statement

This study strictly adheres to international guidelines and principles for animal experimentation to ensure that experiments are carried out with minimal animal suffering. Prior to the start of the experiments, we obtained approval from the animal ethics committee at Shanghai Fourth People’s Hospital, School of Medicine, Tongji University, on 27 February 2023 (approval number: TJBH12523101). All experimental animals were housed in specific pathogen-free (SPF) environments, provided with adequate food and water, and given appropriate resting conditions. Humane methods were employed for animal disposal after the experiments were concluded. And the animal ethics committee at Shanghai Fourth People’s Hospital, School of Medicine, Tongji University follows the rules of Basel Declaration.

Experimental animals

To obtain P2X7f/f; LysM-cre mice, we performed crosses between P2X7-flox/flox (NM-CKO-220373, Nanmo Biotech) mice and LysM-cre (NM-KI-215037, Nanmo Biotech) mice using the Cre-loxP system. All mice were bred under SPF conditions and fed with specialized feed and water provided by Beijing Weitonglihua Experimental Technology Company. Additionally, temperature and lighting conditions were strictly controlled to ensure consistency throughout the experiment.

Establishment of the OP model

Ovariectomy (OVX) was performed to induce osteoporosis in female C57BL/6 mice aged 8–10 weeks and P2X7f/f; LysM-cre mice. The procedure involved the following steps: mice were anesthetized with pentobarbital sodium (328,510-1G, Merck) at a dose of 50 mg/kg. A small incision was made in the abdominal cavity of the mice to locate and expose the ovaries. Using surgical forceps, the fallopian tubes, ovaries, and surrounding tissues were separated to ensure complete ovary removal, followed by wound closure. The success rate of the osteoporosis model was determined by calculating the ratio of mice with successful modeling to the total number of mice operated on. At week 8, micro computed tomography (micro-CT) was employed to evaluate bone density, fracture rate, and other bone parameters in the mice, comparing them with the control group to assess the success of the model establishment. Throughout the experiment, 90% of the mice were retained in the successful modeling group for further research.

In the first part, the mice were divided into two groups, each comprising three mice: (1) wild type (WT) + OVX group, in which OVX was performed on C57BL/6 mice; (2) KO + OVX group, where OVX was conducted on P2X7f/f; LysM-cre mice.

In the second part, as depicted in Fig. S1, the modeling process entailed categorizing the mice into four groups, with each group consisting of six mice: (1) WT + Sham group, involving sham surgery on WT mice; specifically, they were anesthetized akin to the OVX procedure, a simulated incision was made without affecting ovarian tissues, followed by disinfection of the surgical site. Sampling occurred at week 8; (2) WT + OVX group, where WT mice underwent OVX and were sampled at week 8; (3) KO + OVX group, where P2X7f/f; LysM-cre mice underwent OVX and were sampled at week 8; (4) KO + OVX + Recilisib group, where P2X7f/f; LysM-cre mice underwent OVX and initiated daily intraperitoneal injections of recilisib (HY-101625, MCE) the day after surgery, with sampling at week 8.

For Recilisib, the concentration administered was 10 mg/kg, with an injection volume of 100 μL. The formulation was prepared according to the following ratio: 10% dimethylsulfoxide (DMSO), 40% polyethylene glycol 300 (PEG300; HY-Y0873, MCE), 5% Tween 80 (HY-Y1891, MCE), and 45% saline [33].

RNA extraction and transcriptome sequencing

For the partial transcriptome sequencing of animals, we selected the femur tissue of mice, with three mice per group. The tissue (20–50 mg) was cut on a dry ice platform for sequencing. Total RNA was extracted from the tissue using TRIzol (catalog number: 15596026, ThermoFisher, USA), and the purity and concentration of the extracted RNA were assessed using a nanodrop2000 spectrophotometer (ThermoFisher, USA). Following the instructions of the PrimeScript RT reagent Kit (catalog number: RR047A, Takara, Japan), the RNA was reverse transcribed into cDNA for transcriptome sequencing. Differential analysis was performed using the “limma” package in R, with a cutoff of |log2(FoldChange)|> 2 and a significance threshold of P < 0.05 for selecting differentially expressed genes [34].

LC–MS/MS analysis of metabolites

Mouse fecal samples were collected from the established C57BL/6 and P2X7f/f; LysM-cre mice OP model, with ten samples per group. All samples were stored at −80 °C to maintain stability until further analysis. After collecting fecal samples from each cage, they were freeze-dried and ground using a centrifugal grinding mill (Retsch, Haan, Germany). The samples were then added to precooled 50% methanol and thoroughly mixed by vortexing. After incubating on ice for 5 min, the supernatant was collected by centrifugation at 4 °C and 15,000 g for 15 min for subsequent analysis. Metabolite analysis was performed using liquid chromatography-tandem mass spectrometry (LC–MS/MS). Protein precipitation and metabolite extraction of the samples were carried out using methanol (HPLC grade, Merck, Germany). The samples were chromatographically separated using a Vanquish UHPLC system (100 × 2.1 mm, 1.9 μm) at a constant temperature of 40 °C, and the eluted metabolites were detected using an Orbitrap Q Exactive series mass spectrometer (Thermo Fisher). A C18 column was used in the UHPLC-MS/MS analysis. The sample injection volume was 5 mL, and the column flow rate was maintained at 0.2 mL/min. The mobile phase consisted of two eluents. In positive mode, eluent A was 0.1% formic acid (FA) in water, and eluent B was methanol. In negative mode, eluent A was a 5 mM ammonium acetate solution at pH 9.0, and eluent B was methanol. The gradient elution was as follows: 2% B for 1.5 min, 2–100% B for 12.0 min, 100% B for 14.0 min, 100–2% B for 14.1 min, and finally 2% B for 17 min. The spray voltage of the mass spectrometer was set at 3.2 kV, the capillary temperature was 320 °C, the gas flow rate was set at 35 arb, and the auxiliary gas flow rate was set at 10 arb.

The collected data were initially processed using MassHunter Workstation Software (Agilent, USA) to extract the mass spectrometric features of the metabolites. Subsequently, SIMCA software (Umetrics, Sweden) was used for multivariate statistical analysis, including principal component analysis (PCA) and orthogonal projections to latent structure discriminant analysis (OPLS-DA), to identify differential lipid metabolite markers and their distribution [35].

The correlation between differential metabolites and differential genes was analyzed using the Pearson correlation method, and the correlation relationship was visualized using the “ggplot” package in R to generate a heatmap.

Functional enrichment analysis

The core gene set and metabolites were subjected to pathway analysis using the Kyoto Encyclopedia of Genes and Genomes (KEGG) and Reactome databases. Firstly, the core gene set and metabolites obtained from transcriptomics and metabolomics data were uploaded to the database’s analysis tools. We specifically focused on information related to the PI3K-Akt-GSK3β signaling pathway to better understand its role and importance in the study [35]. All the analyses were performed using R language and Bioconductor packages.

Isolation of osteoclast precursor cells from C57BL/6 mouse

Bone marrow macrophages (BMM) were extracted from C57BL/6 mice for subsequent investigation of osteoclast differentiation and absorption. After ethical handling, mouse femurs were extracted, and bone marrow cells were separated from osteoclast precursor cells by density gradient centrifugation. The bone marrow was placed in a culture dish containing trypsin to release the bone marrow cells, and the bone marrow was digested and separated using the enzyme solution in the culture dish to release the osteoclast precursor cells. The bone marrow cells were purified and separated through multiple centrifugation steps. In the first centrifugation, cells were centrifuged at 300 g for 5 min to remove most cell debris and tissue residues, and the cell-containing supernatant was collected. The supernatant was transferred to a new centrifuge tube, and larger cells were precipitated by increasing the centrifugation speed and time to 800 g for 10 min. Finally, centrifugation at 1200 g for 15 min was performed to precipitate and separate the BMM cells. The cells were cultured in Roswell Park Memorial Institute (RPMI)-1640 medium containing 10% fetal bovine serum (16,140,089, Gibco, USA) and 1% antibiotic mixture (15,140,122, Invitrogen, USA) for subsequent experiments [36].

Flow cytometry was used to identify the isolated primary BMM cells. The isolated cells were collected and washed twice with phosphate buffered saline (PBS) containing 1% fetal bovine serum, each time at 300 g for 5 min. Then, 1 × 106 cells were suspended in a 96-well plate with 200 μl per well. The cells were labeled with diluted antibodies in the dark at 4 °C for 30 min. The antibodies used were f4/80-PE (ab237335, Abcam) and CD11b-FITC (ab24874, Abcam). After labeling, the cells were washed twice with PBS containing 1% fetal bovine serum to remove excess antibodies. Finally, cell analysis was performed using the FACS Calibur flow cytometer (BD Biosciences, USA). FlowJo software was used for data analysis to calculate the proportion of cells labeled as f4/80+ CD11b+.

Constructing lentiviruses for gene silencing and overexpression

The lentivirus packaging service was provided by GenBio Co., Ltd. (Shanghai, China). pHAGE-puro plasmids and auxiliary plasmids pSPAX2, pMD2.G, and pSuper-retro-puro plasmids and auxiliary plasmids gag/pol, VSVG were cotransfected into 293T cells (CRL-3216, ATCC, USA). After 48 h of cell culture, the supernatant was collected, and the filtered supernatant, after centrifugation through a 0.45 μm filter, was collected as the virus. After 72 h, the supernatant was harvested again and centrifuged for concentration. The two rounds of virus were mixed, and the titer was determined. The sequences of the siRNAs used for lentiviral silencing are shown in Table S1 and were validated for silencing efficiency in bone marrow-derived mesenchymal stem cells (BMSCs) cell lines (PCS-500–012, ATCC, USA).

As shown in Fig. S2, the sequence (Sh-P2X7-2) with the most optimal silencing effect was selected for subsequent experiments.

Cell treatment and grouping

In vitro experiments were performed using bone marrow-derived osteoclast precursors (BMM) obtained from C57BL/6 mice. For lentivirus-mediated cell transfection, 5 × 105 cells were seeded into 6-well plates. When the confluence of BMSCs reached 70–90%, the culture medium containing an appropriate amount of packaged lentivirus [multiplicity of infection (MOI) = 10, working titer approximately 5 × 106 TU/mL] and 5 μg/mL polybrene (TR-1003, Merck, USA) was added for transfection. After 4 h of transfection, an equal amount of culture medium was added to dilute polybrene. The medium was replaced with a fresh culture medium after 24 h of transfection, and the transfection was observed by the luciferase reporter gene after 48 h. G418 (A1113803, Gibco, Grand Island, NY, USA) was used for resistance selection to obtain stable cell lines. Cells were collected when they no longer died in the presence of G418, and silencing efficiency was confirmed by reverse transcription-quantitative polymerase chain reaction (RT-qPCR) [37].

In the first part, cells were divided into four groups: (1) Sh-NC group: transfected with lentivirus carrying empty vector for silencing; (2) Sh-P2X7 group: transfected with lentivirus carrying P2X7 vector for silencing; (3) OE-NC group: transfected with lentivirus carrying empty vector for overexpression; and (4) OE-P2X7 group: transfected with lentivirus carrying P2X7 vector for overexpression. The working titer of lentivirus for all groups was 5 × 106 TU/mL.

In the second part, cells were divided into four groups: (1) Sh-P2X7 + DMSO group: transfected with lentivirus carrying P2X7 vector for silencing, treated with 10 μL DMSO; (2) Sh-P2X7 + recilisib group: transfected with lentivirus carrying P2X7 vector for silencing, treated with 10 μM recilisib (HY-101625, MCE); (3) OE-P2X7 + DMSO group: transfected with lentivirus carrying P2X7 vector for overexpression, treated with 10 μL DMSO; (4) OE-P2X7 + LY294002 group: transfected with lentivirus carrying P2X7 vector for overexpression, treated with 10 μM LY294002 (HY-10108, MCE).

After the respective treatments, all cells were induced for osteoclast differentiation using M-CSF (50 ng/mL) and RANKL (50 ng/mL; 462-TEC-010, R&D, USA). Cells were cultured in the differentiation medium for 7 days for subsequent experiments [38].

Cell viability and proliferation assay

Cell viability and proliferation were assessed using the Cell Counting Kit 8 (CCK-8) assay kit (CK04, Dojindo, Japan). After pretreating the cells according to the aforementioned grouping method, a single-cell suspension was prepared by diluting it in a complete culture medium to a concentration of 5 × 104 cells/mL. Subsequently, 100 μL of the cell suspension was seeded into each well of a 96-well plate. The outer wells of the plate were filled with PBS solution, and the plate was then incubated at 37 °C with 5% CO2 for 24 h. The supernatant was discarded, and the plate was washed twice with PBS. Subsequently, six wells were randomly assigned for each group and incubated for an additional 24 and 48 h, respectively. To each well, 10 μL of CCK-8 solution was added, mixed well, and incubated at 37 °C with 5% CO2. Finally, the absorbance at 450 nm was measured [39].

Tartrate-resistant acid phosphatase (TRAP) staining

The processed cells were fixed with 10% formaldehyde at room temperature for 10 min. Subsequently, they were stained using the TRAP acid phosphatase assay kit (P0332, Beyotime) following the manufacturer’s protocol. After staining, the cells were transferred to a new culture dish and treated with an equal volume of TRAP solution [containing 4.93 mg p-nitrophenyl phosphate (PNPP) in 0.5 M acetic acid solution (750 ml) and 150 ml tartaric acid solution] for 30 min at 37 °C. The detection buffer from the kit served as the negative control. The reaction was stopped by adding 0.5 M sodium hydroxide (NaOH), and the absorbance was measured at a wavelength of 405 nm using a Multiskan™ FC microplate reader (1,410,101, ThermoFisher).

For tissues, paraffin-embedded sections were deparaffinized in a gradient of ethanol to expose the tissue. After deparaffinization, the tissue sections were incubated in TRAP staining solution at 37 °C for 30 min, followed by washing off the excess staining solution with PBS. The sections were then mounted, observed, and captured under a microscope. Three slices were selected from each animal, and three fields were randomly chosen from each slice for image capture. The number of TRAP-positive cells was counted [40].

Bone resorption pit detection

The bone resorption assay kit (BRA-24KIT, Cosmo Bio, Japan) was used. Following the grouping of BMM cells as described above, they were added to the wells of the kit’s plate along with the assay buffer. Detection was carried out at excitation and emission wavelengths of 485 nm and 535 nm, respectively. Subsequently, cells were removed after treatment with 5% sodium hypochlorite, washed with PBS, and observed under a microscope for the assessment of their resorption activity. The area of resorption pits was measured, calculated, and subjected to statistical analysis [40].

F-actin staining

To assess the formation of F-actin rings, osteoclasts were first fixed in a 4% paraformaldehyde solution at room temperature for 15 min and then washed with PBS. After treatment with 0.3% Triton X-100, the osteoclasts were stained with rhodamine-labeled phalloidin (ab176753, Abcam) at room temperature for 60 min, followed by staining with 4′,6-diamidino-2-phenylindole (DAPI; 10,236,276,001, Sigma-Aldrich, USA) for 5 min at room temperature. Formation of F-actin rings was observed using a fluorescence microscope (IX71, Olympus, Japan) and statistically analyzed using ImageJ software [41].

Micro-CT analysis

In this study, we utilized the mCT-40 micro-CT system from Scanco Medical (Switzerland) to scan the region of interest in the femoral bone tissue and analyze its growth. The scanning parameters employed were as follows: power current (385 μA), power voltage (65 kV), pixel size (9 μm), filter (AI 1.0 mm), and rotation step (0.4°). The reconstructed images were generated using Bruker’s NRecon software, and the data were analyzed using the CTAn program. Specifically, we determined the volume of interest (VOI) at 0.5 mm above and 0.25 mm in height from the growth plate of the femoral head. Within this volume, the region of interest (ROI) was manually defined as the subchondral area of cartilage, with a constant threshold (50–255) applied for trabecular bone binary segmentation. The micro-CT parameters assessed included: (a) bone volume fraction (BV/TV), which represents the ratio of bone surface area to tissue volume; (b) trabecular separation (Tb/Sp), indicating the distance between trabeculae within the bone; (c) trabecular thickness (Tb.Th), reflecting the average thickness of trabeculae and describing their structural changes; (d) trabecular number (Tb.N), which counts the intersections between bone and nonbone tissue within a given length; (e) bone mineral density (BMD), a measure of bone quantity and distribution density [42, 43].

Hematoxylin and Eosin (HE) staining

After sample collection, decalcification was performed followed by paraffin embedding. Histological examination of tissue morphology was conducted using 4 μm thick sections stained with hematoxylin and eosin (HE). Paraffin sections were deparaffinized in xylene and different concentrations of ethanol (100%, 95%, 80%, 70%), followed by immersion in tap water for 15 min. The sections were then stained with hematoxylin staining solution (H8070, Solarbio, Beijing, China) for 5–10 min at room temperature. Subsequently, the slides were rinsed with distilled water, dehydrated in 95% ethanol, and placed in eosin staining solution (G1100, Solarbio, Beijing, China) for 5–10 min at room temperature, followed by standard dehydration, clearing, and mounting procedures [44].

Von Kossa staining

Mouse femoral tissue sections were stained using the calcium salt staining kit (G3282, Solarbio) for Von Kossa staining. The sections were immersed in 1% silver nitrate and exposed to bright light for 30 min, followed by three washes with deionized water. Subsequently, 5% sodium thiosulfate was added and incubated for 5 min to remove unreacted silver. Calcium phosphate deposits were visualized in black (room temperature). The stained sections were observed and captured under an optical microscope [45]. The Von Kossa control solution from the kit was used as the negative control reagent.

Immunohistochemical staining

Mouse femoral tissue sections were dehydrated in 100% ethanol, 95% ethanol, and 70% ethanol, followed by rinsing with water. Antigen retrieval was conducted in a high-temperature, high-pressure environment. The activity of endogenous peroxidases was reduced using a 3% hydrogen peroxide solution. Subsequently, cover slips were incubated overnight at 4 °C with specified primary antibodies: rabbit anti-p-PI3K (ab182651, 1:1000, Abcam), rabbit anti-p-AKT (ab38449, 1:1000), and rabbit anti-p-GSK3β (ab68476, 1:1000, Abcam). PBS solution was used as the negative control reagent. Following three washes with PBS, 100 μL of enzyme-labeled goat anti-rabbit immunoglobulin G (IgG) polymer (pv6000, Zhongshanjinqiao, China) was added and incubated for 30 min. Subsequent color development was carried out using the DAB chromogenic kit (ZLI-9018, Zhongshanjinqiao, China) at room temperature, and the staining was observed under a microscope. Images were captured post-staining. In total, there were six animals per group, with three sections stained per animal and three fields selected for imaging. After imaging, the positive area percentage was calculated using ImageJ 1.48u software (V1.48, National Institutes of Health, USA) [46].

Enzyme-linked immunosorbent assay (ELISA)

We used a mouse Cross-linked Type I Collagen C-Telopeptide (CTX) ELISA kit (M3023, Elabscience, Wuhan, China) and a mouse Cross-linked Type I Collagen N-Telopeptide (NTX) ELISA kit (E-EL-M3022, Elabscience, Wuhan, China) for mice. Following the instructions, we analyzed the levels of bone turnover markers CTX and NTX in the serum of the model mice. Firstly, standards and test samples were added to each well and incubated at 37 °C for 90 min. After that, the liquid was removed, and 100 μL of biotinylated antibody working solution was added and incubated at 37 °C for another 60 min. After three washes, 100 μL of horseradish peroxidase (HRP) conjugate working solution was added and incubated at 37 °C for 30 min. The liquid was removed, and the wells were washed five times. Then, 90 μL of substrate solution was added to each well and incubated for approximately 15 min at 37 °C. Finally, 50 μL of stop solution was added to each well, and the absorbance (OD value) was measured at 450 nm wavelength using a microplate reader to calculate the sample concentration. The blank well was set as zero [40].

Western blot

The mouse femoral tissue and cells were ground and homogenized in radioimmunoprecipitation assay (RIPA) lysis buffer (P0013B, Beyotime Biotechnology, China) and a protease inhibitor (P1005, Bi Yun Tian, China). Digestion of the cells was performed on ice. After grinding, the mixture was left to stand at 4 °C for 1 h, followed by centrifugation at 12,000 rpm for 15 min at 4 °C to extract the supernatant, which was then stored at −80 °C. The protein concentration was determined using a BCA assay kit (A53226, Thermo Fisher Scientific, Rockford, IL, USA). The proteins were transferred from the polyacrylamide gel to a polyvinylidene fluoride (PVDF) membrane (IPVH85R, Millipore, Darmstadt, Germany) using a wet transfer method. The membrane was then blocked with 5% bovine serum albumin (BSA) at room temperature for 1 h, followed by an incubation with the following primary antibodies: rabbit anti-P2X7 (ab307718, 1 µg/ml, Abcam), rabbit anti-MMP9 (ab283575, 1 µg/ml, Abcam), rabbit anti-CK (ab53280, 0.1 µg/ml, Abcam), rabbit anti-NFATc1 (A303-508A-T, 1 µg/ml, ThermoFisher), rabbit anti-PI3K (ab302958, 1 µg/ml, Abcam), rabbit anti-p-PI3K (ab182651, 2 µg/ml, Abcam), rabbit anti-AKT (ab8805, 1 µg/ml), rabbit anti-p-AKT (ab38449, 1 µg/ml), rabbit anti-GSK3β (ab185141, 1 µg/ml), rabbit anti-p-GSK3β (ab68476, 1 µg/ml, Abcam), mouse anti-β-actin (3700, 1 µg/ml, Cell Signaling Technology, USA), rabbit anti-Tub (ab131034, 1 µg/ml, Abcam), and mouse anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH; ab8245, 1 µg/ml, Abcam). After washing, the membrane was incubated with an HRP-conjugated anti-rabbit IgG secondary antibody (ab6721, 0.2 µg/ml, Abcam) or an HRP-conjugated anti-mouse IgG secondary antibody (ab205719, 0.2 µg/ml, Abcam) for 1 h. Detection was carried out using a chemiluminescence imager. Protein quantification analysis was performed with ImageJ 1.48u software, calculating the ratio of grayscale values of each protein to a reference grayscale value [47]. The experiment was repeated three times.

RT-qPCR

Total RNA from the cells was extracted using TRIzol (catalog number: 15596026, ThermoFisher, USA). The purity and concentration of the extracted RNA were assessed using a nanodrop2000 spectrophotometer (ThermoFisher, USA). The RNA was reverse transcribed into cDNA using the PrimeScript RT reagent Kit (RR047A, Takara, Japan), according to the manufacturer’s instructions. RT-qPCR was performed using the Fast SYBR Green PCR kit (11,736,059, Thermo Fisher Scientific, China) with three replicates per well, using GAPDH as an internal reference. The relative expression level was calculated using the 2−ΔΔCt method. The experiment was repeated three times. The primer sequences used in this study are provided in Table S2 [48].

Statistical analysis

All data were analyzed using GraphPad Prism 8.0. Continuous data are presented as mean ± standard deviation (Mean ± SD). Unpaired t-tests were used to compare data between two groups, while a one-way analysis of variance (ANOVA) was used to compare multiple groups.

The homogeneity of variances was tested using Levene’s test, and if the variances were homogenous, Dunnett’s t-test and LSD-t test were used for pairwise comparisons. If the variances were not homogenous, Dunnett’s T3 test was used. Pearson’s analysis was used to evaluate the correlation between genes and immune cell content. P < 0.05 was considered statistically significant for the differences between the two groups [49].