Targeting integrin αVβ3–Ptgs2–mTOR signaling rescues bone formation in osteoporosis: from molecular mechanism toward therapy

Abstract

Background

Integrin αVβ3, a key ECM receptor, is essential for bone metabolism, yet its role in postmenopausal osteoporosis (PMOP) remains unclear. This study investigates the molecular mechanisms by which integrin αVβ3 regulates osteoblast function and bone homeostasis in PMOP.

Methods

Using clinical samples, OVX mice, and in vitro models, we analyzed integrin αVβ3 expression and its impact on osteogenesis. Clustered regularly interspaced short palindromic repeats (CRISPR)/Cas9-mediated knockout, multi-omics profiling, and protein interaction assays (Co-IP, BLI, and structural modeling) were employed to dissect the underlying pathway. An AAV9-based in vivo overexpression system was developed to evaluate therapeutic potential.

Results

Integrin αVβ3 was downregulated in PMOP patients and OVX mice, correlating with osteoblast dysfunction and reduced bone formation. Mechanistically, integrin αVβ3 deficiency upregulated Ptgs2, which directly bound to mammalian target of rapamycin (mTOR) via a hydrogen bond between Ptgs2-Glu52 and mTOR-Ser2159, inhibiting mTOR phosphorylation. This suppression disrupted mTORC1-S6K/4EBP1 signaling, impairing osteoblast proliferation and survival. Notably, AAV9-mediated integrin αVβ3 overexpression rescued bone loss in OVX mice.

Conclusions

Our findings unveil a novel integrin αVβ3–Ptgs2–mTOR axis in PMOP pathogenesis: estrogen deficiency reduces integrin αVβ3, enabling Ptgs2-mediated mTOR inhibition and osteogenic decline. This study identifies integrin αVβ3 as a potential therapeutic target to restore bone formation in osteoporosis.

Supplementary Information

The online version contains supplementary material available at 10.1186/s11658-025-00842-3.

Introduction

Osteoporosis (OP) is a systemic skeletal disease characterized by loss of bone mass and destruction of bone microstructure, among which postmenopausal osteoporosis (PMOP) is the most common type, and its pathological mechanism is mainly closely related to bone metabolic imbalance caused by estrogen deficiency [1–3]. Estrogen exerts its biological effects mainly through two classical nuclear receptors: estrogen receptor alpha (ERα) and estrogen receptor beta (ERβ), which are widely expressed in bone tissue, including osteoblasts, osteocytes, and osteoclasts [4, 5]. These receptors regulate gene transcription and various signaling pathways critical for bone remodeling. Notably, ERα is considered the primary mediator of estrogen's protective effects on bone, particularly in regulating osteoclast apoptosis and bone resorption. ERβ appears to have more modulatory and context-dependent roles, which may include fine-tuning the actions of ERα and influencing bone formation [6, 7]. The loss of estrogen signaling through ERα and ERβ disrupts the balance between bone resorption and formation, contributing to the pathogenesis of PMOP [6, 7]. Previous studies primarily focused on the role of excessive activation of osteoclasts in the pathogenesis of PMOP, and it was believed that PMOP was mainly caused by excessive activation of osteoclasts caused by estrogen deficiency, leading to systemic bone mass loss and bone microstructure destruction [8]. However, the therapeutic strategy of inhibiting bone resorption alone has limited effects in some patients with PMOP [2, 3], which suggests that bone formation disorders may play a nonnegligible pathological role in the process of bone mass loss in PMOP [2, 3].

Integrin αvβ3 is a heterodimeric transmembrane protein composed of αV and β3 subunits. As an important receptor of extracellular matrix (ECM), it is widely expressed on the surface of osteoblasts, osteoclasts, and osteocytes, and plays multiple roles in the regulation of bone metabolism [9–11]. Estrogen, primarily acting through its nuclear receptors ERα and ERβ, is known to regulate the expression and activity of various integrins in bone cells via both genomic and nongenomic pathways [4, 5]. This regulatory crosstalk may represent a key mechanism by which estrogen maintains bone homeostasis. Previous studies have shown that integrin αVβ3 is involved in regulating of bone remodeling by mediating cell-ECM bidirectional signal transduction, and its abnormal expression is closely related to the occurrence and development of osteoporosis [9, 10, 12, 13]. However, current studies on integrin αVβ3 in bone metabolism have primarily focused on its regulatory mechanisms on osteoclast activity. For example, Li et al. [14] demonstrated that integrin αVβ3 promotes osteoclast migration and bone resorption by activating the Pyk2/Src signaling axis. Inhibition of integrin αvβ3 prevents osteoclast-mediated bone resorption by reducing Th17 activation and receptor activator of nuclear factor-κB ligand (RANKL) levels [15]. In recent years, some scholars have started to focus on the relationship between integrin αVβ3 and bone formation. For example, Voisin et al. [16] found that the expression of integrin αVβ3 in osteocytes was significantly decreased in the estrogen-deficient postmenopausal osteoporosis rat model. Geoghegan's team further demonstrated that estrogen depletion disrupts the structural stability of αVβ3 within osteocyte focal adhesions, impairing its downstream signaling [17, 18]. However, significant knowledge gaps remain regarding how integrin αVβ3 regulates osteoblast function and its precise mechanistic role in PMOP. Based on these findings, this study employed a multidimensional experimental approach (including clinical sample analysis, OVX animal models, and cellular/molecular experiments) to demonstrate a definitive correlation between downregulated integrin αVβ3 expression in bone tissue and osteoblast-mediated bone formation impairment under PMOP conditions.

Further investigation revealed that integrin αVβ3 modulates osteoblast function through regulating the “Ptgs2–mTOR” signaling axis. This study for the first time confirmed the direct physical interaction between Ptgs2 and mTOR, and structural simulation demonstrating that Glu52 (Ptgs2) and Ser2159 (mTOR) can form stable hydrogen bonds, thereby inhibiting mTOR phosphorylation. Moreover, we demonstrated that integrin αVβ3 alleviates Ptgs2-mediated suppression of mTOR by downregulating Ptgs2 expression, consequently activating the mTORC1-S6K/4EBP1 pathway to promote osteoblast proliferation. In vivo experiments showed that AAV9-mediated integrin αVβ3 overexpression improved bone microstructure in OVX mice, accompanied by decreased Ptgs2 levels and enhanced mTOR activity. These findings not only elucidate the functional of integrin αVβ3 in osteoblasts but also establish a novel theoretical framework for PMOP pathogenesis: estrogen deficiency may impair osteoblast function by suppressing the “integrin αVβ3–Ptgs2–mTOR” axis, ultimately leading to bone loss. Furthermore, this study provides a potential strategy for developing pro-osteogenic therapies targeting integrin αVβ3, with a particular emphasis that specific interventions at the Ptgs2–mTOR interaction interface may represent a promising future research direction.

Materials and methods

Collection of human specimens

Human trabecular bone samples were obtained from female patients undergoing surgical treatment for femoral neck or knee joint fracture. Patients with secondary causes of osteoporosis (e.g., endocrine/metabolic disorders, chronic kidney disease, malignancies, immunosuppression, infection, etc.) were excluded. A total of 12 samples were collected, including the postmenopausal osteoporosis group (n = 6, age 55–75 years, bone mineral density (BMD) T-score ≤ −2.5) and premenopausal non-osteoporosis group (control, n = 6, age 45–60 years, BMD T-score ≥ −1.0). Written informed consent was obtained from all participants. The study protocol was approved by the Medical Ethics Committee of the Second Hospital of Lanzhou University (approval no. 2024 A-1265). Detailed patient information is provided in Supplementary Tables S1 and S2.

Establishment of the OVX animal model and intervention

All female C57BL/6 J mice (6–8 weeks old, 18 ± 2 g) were purchased from the Lanzhou Veterinary Research Institute of the Chinese Academy of Agricultural Sciences and maintained under standard specific pathogen-free (SPF) laboratory conditions. All animal procedures were performed in compliance with institutional ethical guidelines and were approved by the Animal Care Committee of the Second Hospital of Lanzhou University (approval no. D2024-920).

The OVX mouse model was established following previously published protocols [19, 20]. Briefly, anesthetized mice were placed in prone position on a surgical platform. Bilateral ovaries were carefully exposed and excised, and the surgical incision was closed using standard suturing techniques. For sham-operated controls (Sham group), bilateral ovaries were similarly exposed with removal of surrounding adipose tissue while preserving ovarian integrity. After an 8-week postoperative period, bone tissues were harvested from both OVX-induced postmenopausal osteoporotic mice and Sham controls (n = 5 per group) for comprehensive analyses, including micro-computed tomography (micro-CT), immunofluorescence (IF) staining, western blotting, and histopathological examination (Masson, tartrate-resistant acid phosphatase (TRAP), and hematoxylin and eosin (H&E) staining).

To investigate the biological functions of integrin αVβ3 in vivo, we established a mouse model overexpressing integrin αVβ3 using osteospecific adeno-associated virus serotype 9 (AAV9) constructed by GenePharma (Shanghai, China). Thirty mice were randomly allocated into five experimental groups (n = 6 per group): Sham-operated controls (Sham), OVX + AAV9-negative control (AAV9-NC), OVX + AAV9-ITGAV (αV subunit overexpression), OVX + AAV9-ITGB3 (β3 subunit overexpression), and OVX + AAV9-ITGAV&ITGB3 (combined αVβ3 overexpression). Following established protocols [20, 21] and manufacturer’s guidelines, viral delivery was achieved through localized bilateral injections targeting the distal femur region, administered weekly post-surgery. After an 8-week experimental period, all mice were euthanized. Femoral specimens were collected for comprehensive analysis including micro-CT, IF staining, and H&E and Masson staining.

Micro-CT analysis

The distal femora from each experimental group were subjected to ex vivo microcomputed tomography (μCT) analysis using an NMC-200 system (NEMO, China) with the following acquisition parameters: 80 kV X-ray source voltage, 120 μA current, and 13 μm isotropic voxel resolution. Three-dimensional volume reconstruction and quantitative analysis were performed using Recon 1.6 software. Trabecular bone microarchitecture was evaluated with the following morphometric parameters quantified: bone mineral density (BMD), bone volume fraction (BV/TV), trabecular separation/spacing (Tb.Sp), trabecular thickness (Tb.Th), and trabecular number (Tb.N).

H&E, Masson, and TRAP staining

Femur specimens were postfixed with 4% paraformaldehyde (PFA) for 24 h, decalcified with 10% ethylenediaminetetraacetic acid (EDTA) for 21 days, embedded in paraffin, and sectioned at a thickness of 4 μm. H&E, Masson, and TRAP staining were performed according to the manufacturer's instructions (Solarbio, China). The mouse femora morphology was observed using a TissueFAXS whole-slide scanning system (TissueGnostics GmbH, Austria).

Dynamic bone histomorphometry

Mice were injected intraperitoneally with calcein (20 mg/kg; Sigma-Aldrich) at 9 and 2 days before euthanasia. After collection, undecalcified femoral shafts were embedded in methyl methacrylate. Sections (~5 µm thickness) were cut using a microtome (Leica SM2500) and visualized under a fluorescence microscope. The mineral apposition rate (MAR) was calculated as the distance between the two calcein labels divided by the time interval (7 days) and is expressed in µm/day. Measurements were performed blinded to the experimental groups.

Cell extraction and culture

The MC3T3-E1((RRID: CVCL_0409)) cell line was obtained from the National Stem Cell and Translational Biomedical Resource Center (NSTI-BMCR), Beijing, China (cat. no. 1101MOU-PUMC000012). The cell line was authenticated by short tandem repeat (STR) profiling and tested negative for mycoplasma contamination using the polymerase chain reaction (PCR)-based method upon distribution by the resource center. The cells were confirmed to be free of mycoplasma contamination prior to the commencement of the experiments described in this study. The cell line is publicly available from the NSTI-BMCR upon request. Culture was carried out in α-MEM medium (Gibco) supplemented with 10% fetal bovine serum (FBS, Gibco) at 37 °C and 5% CO₂. Primary osteoblasts were isolated from the calvaria of C57BL/6 J neonatal mice using the tissue explant method and cultured in Dulbecco’s modified Eagle medium (DMEM, Gibco) supplemented with 10% fetal bovine serum (FBS) at 37 °C and 5% CO₂.

Cell model of integrin αVβ3-knockout was established using CRISPR-Cas9 technology

To investigate the regulatory role of integrin αVβ3 in osteoblast function, we established an integrin αVβ3-deficient cellular model by knocking out the ITGAV and ITGB3 genes in MC3T3-E1 and primary osteoblasts using CRISPR-Cas9 gene editing technology. ITGAV and ITGB3 specific single-guide RNA (sgRNA) were designed using the online tool CRISPR DESIGN (http://crispor.tefor.net/) and synthesized and cloned into the px459-puro plasmid vector. The sgRNA sequences are presented in Supplementary Table S3. Plasmid vectors were then transfected into target cells using a BEX CUY21 EDIT II cell electroporation instrument (Beihui, Beijing, China). Monoclonal cell line was selected by puromycin and limiting dilution analysis (LDA). Monoclonal cell lines were selected by puromycin and limiting dilution analysis (LDA). The expression level of integrin αVβ3 was verified by reverse-transcription quantitative PCR (RT-qPCR) detection, Western blot analysis, and immunofluorescence (IF) staining.

Cell transfection

For the overexpression of integrin αVβ3, we used specific overexpression plasmids custom-synthesized by Genecarer Co., Ltd. (Xian, China) to transiently transfect MC3T3-E1 cells and primary osteoblasts to achieve integrin αVβ3 overexpression. To knockdown Ptgs2 in integrin αVβ3 knockout cells (ITGAV^−/− and ITGB3^−/− cells), we used siRNA designed and synthesized by Generalbiol Co., Ltd. (Anhui, China). Knockdown and overexpression were performed by transient transfection of siRNA and plasmid after 6 h of cell starvation. The transient transfection of overexpression plasmids was performed using a BEX CUY21 EDIT II cell electroporation instrument (Beihui, Beijing, China), whereas transfection of Ptgs2-siRNA was conducted with the Transmate transfection reagent. The transfected cells were harvested 48 to 72 h post-transfection. The siRNA sequences for Ptgs2 can be found in Supplementary Table S4.

Estrogen withdrawal (EW) cell model

To further investigate the impact of estrogen deficiency on integrin αVβ3 expression in osteoblasts, an estrogen withdrawal model was established using the MC3T3-E1 cell line and primary osteoblasts, employing 17β-estradiol (HY-B0141, MCE, USA). The specific experimental procedures were as follows: (1) The MC3T3-E1 cell line and primary osteoblasts were cultured in standard α-MEM and DMEM medium, respectively. (2) The estrogen-treated group (Estrogen, E) received continuous treatment with 10 nM 17β-estradiol for 5 days. (3) The estrogen withdrawal model group (Estrogen withdrawal, EW) was pretreated with 10 nM 17β-estradiol for 3 days, followed by withdrawal of 17β-estradiol for 2 days. Following completion of the modeling process, total cellular protein and RNA were extracted for subsequent experimental validation.

Proteome analysis

To elucidate the regulatory mechanism of integrin αVβ3 in osteoblasts, we performed Asral DIA proteomic quantitative sequencing on integrin αVβ3-knockout osteoblasts (ITGAV^−/− and ITGB3^−/− cells) and control osteoblasts, which was conducted by BIOPROFILE TECHNOLOGY CO., LTD (Shanghai, China). Differentially expressed proteins (DEPs) were screened using Student’s t-test combined with fold change (FC), with the criteria set at |log₂FC|≥ 1 and P < 0.05. The overlapping DEPs between the two experimental groups (ITGAV^−/− and ITGB3^−/−) and the control group were identified as common differentially expressed proteins (CO-DEPs). Subsequently, enrichment analysis of the identified CO-DEPs was performed using the Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) databases.

Multi-omics analysis

To further elucidate the molecular mechanism of integrin αVβ3 regulating osteoblast function in the development of PMOP, we obtained transcriptome sequencing data (GSE222752) of OVX-induced osteoporotic mouse models and Sham mice from the Gene Expression Omnibus (GEO) database. Differentially expressed genes (DEGs) were screened using DESeq2 software with the criteria of |FC|≥ 1.5 and P < 0.05. Subsequently, a conjoint analysis was performed between the common differentially expressed proteins (CO-DEPs) identified from proteomic sequencing and the DEGs obtained from transcriptomic sequencing. This approach enabled us to identify the most relevant target proteins associated with integrin αVβ3, osteoblasts, and PMOP.

Cell proliferation assay

Cells in each group were washed with phosphate-buffered saline (PBS) buffer; CFDA SE storage solution (5 mM) was added according to the instructions of the CFSE Cell Division Tracker Kit, and incubated in a cell incubator in the dark for 5 min. The labeling reaction was terminated by addition of complete medium and continued with two washes using complete medium. An appropriate amount of complete medium was added, and culture was continued in a 5% CO₂ incubator at 37 °C. Cells were harvested after 72 h for flow cytometry (NovoCyte, ACEA BIO, China) to analyze the proliferation of cells in each group.

Cell apoptosis assay

Cells in each group were washed with PBS buffer; Annexin V–fluorescein isothiocyanate (FITC) binding solution was added following the protocol of the Annexin V-FITC Cell Apoptosis Detection Kit (Yeasen, Shanghai, China), followed by the addition of Annexin V-FITC and propidium iodide staining solution. The mixture was gently mixed and then incubated in a light-avoiding environment at room temperature for 15 min. The samples were used to detect cell apoptosis via flow cytometry (NovoCyte, ACEA BIO, China).

Bio-layer interferometry (BLI)

We employed the FortéBio Octet^® R2 biomolecular interaction analysis system (Danaher, USA) to evaluate the interaction between Ptgs2 and mTOR. Before the experiment, Ptgs2, mTOR, and control IgG antibodies were diluted to 10 μg/mL, and cell lysates were prepared from each group (NC group, ITGB3^−/− group, and ITGAV^−/− group). The procedure was performed according to the manufacturer's instructions by first prewetting the Bio-Pro A sensor in PBS with Tween-20 (PBST) buffer for 10 min. Subsequently, Ptgs2 and IgG antibodies were immobilized onto separate Bio-Pro A sensors, followed by blocking, associating with cell lysates, binding with mTOR antibody, and dissociation, yielding real-time kinetic curves for each group. Following data acquisition, Octet Analysis Studio 13 software was used for real-time curve fitting, with binding kinetic parameters calculated using a 1:1 binding model to assess the interaction capacity between Ptgs2 and mTOR.

Co-immunoprecipitation (Co-IP)

After washing cells three times with PBS, cells were lysed using radioimmunoprecipitation assay (RIPA) lysis buffer (Beyotime, Shanghai, China) and protease/phosphatase inhibitors (Beyotime, Shanghai, China) for 30 min at 4 °C according to the manufacturer’s instructions, and the cell lysates were collected. Subsequently, protein samples were incubated with IgG of the same species or corresponding primary antibody protein A + G agarose beads (P2012, Beyotime Biotechnology, Shanghai, China) overnight at 4 °C on a rotary shaker. The next day, the magnetic bead antibody complex was washed with washing solution, sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) electrophoresis loading buffer was added, the beads were resuspended, and samples were boiled for 10 min for Western blot analysis.

Molecular docking

Molecular docking analysis of mTOR and Ptgs2 proteins was conducted. The protein structures of mTOR (Q9JLN9-F1-model_v4) and Ptgs2 (AF-Q05769-F1-model_v4) were obtained from the AlphaFold database, and docking studies were performed based on their predicted models. Firstly, the proteins’ structure was dewatered, hydrogenated, energy minimized, and completed with missing residues. Subsequently, semiflexible docking between the two proteins was performed using GRAMM (http://gramm.compbio.ku.edu/) with an root mean square deviation (RMSD) threshold set at 0.2 Å to obtain the optimal low-energy conformation. Following docking, interaction analysis was conducted using LigPlot software. Finally, Pymol (version 2.6) visualization software was employed to generate three-dimensional (3D) structural models of the protein–protein interactions. This computational approach effectively predicted the interprotein interactions and provided crucial theoretical guidance for experimental investigations.

RT-qPCR

Total RNA was extracted from cells using TRIzol (Accurate Biology, Hunan, China). RNA purity and concentration were determined using an ultraviolet–visible spectrophotometer (Denovix, ND-1000, USA). Complementary DNA (cDNA) was synthesized using Oligo (dT) 18 primer and the Evo M-MLV RT PreMix kit with genomic DNA (gDNA) removal (Accurate Biology, Hunan, China). RT-qPCR was performed on an LC480 system (Roche) using SYBR Green Pro Taq HS (Accurate Biology, Hunan, China). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as an internal control. Data from three biological replicates were analysed following the relative quantification method (2^−ΔΔCT).

Western blot analysis

Total protein was extracted from bone tissue specimens of human and mouse and cells using RIPA lysis buffer (Beyotime, Shanghai, China) with protease/phosphatase inhibitor cocktail (Beyotime, Shanghai, China). Protein concentration was measured using a bicinchoninic acid (BCA) protein assay kit (Beyotime, Shanghai, China). Subsequently, proteins were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE). The membranes were blocked with 5% bovine serum albumin at 4 °C for 1 h, and incubated overnight at 4 °C with primary antibodies. An informative list of the antibodies is presented in Supplementary Table S5.

Immunofluorescence (IF) staining

Cells were washed with PBS, fixed with 4% paraformaldehyde at room temperature for 30 min, permeabilized with 0.1% Triton X-100 at room temperature for 30 min, and blocked with 10% goat serum at 37 °C for 1 h. Tissue sections were blocked for 1 h using 10% bovine serum albumin (BSA) after baking, dewaxing, hydration, and antigen retrieval steps. The sections and cells were then incubated with primary antibodies overnight at 4 °C. The next day, after washing, the cells and tissue sections were treated with fluorescent secondary antibody for 1 h at room temperature. Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI). Imaging was performed using a two-photon laser confocal microscope (Carl Zeiss, Zeiss LSM880). Information on antibodies is presented in Supplementary Table S5.

Statistical analysis

The cellular and molecular experiments in this study were repeated independently at least three times, while the animal assays were repeated with at least five mice. Results are presented as the mean ± standard deviation (SD) of triplicate experiments. GraphPad Prism 8.0 (GraphPad Software, Inc., San Diego, CA, USA) was used for statistical analysis. Differences between two groups were determined using an unpaired t-test, and analysis of variance (ANOVA) was conducted for comparing multiple groups, with statistical significance set at P < 0.05.

Results

Integrin αVβ3 expression is decreased in PMOP

To investigate the role of integrin αVβ3 in PMOP, we first examined the expression of integrin αVβ3 in bone tissues of patients with PMOP. The IF staining of human trabecular bone sections revealed that the expression of integrin αVβ3 was significantly reduced in patients with PMOP compared with premenopausal controls (P < 0.001, Fig. 1A). Western blot analysis further confirmed this finding, demonstrating synchronous downregulation of both αV and β3 subunit proteins with statistically significant differences (Fig. 1B, C). These results suggest that decreased integrin αVβ3 expression may be associated with PMOP pathogenesis.

Fig. 1.

The expression level of integrin αVβ3 is decreased in PMOP. A IF staining of human trabecular bone sections showed that the fluorescence intensity of integrin αVβ3 was significantly decreased in patients with PMOP (n = 6). B, C Western blot analysis confirmed the simultaneous downregulation of αV and β3 subunit proteins in bone tissues of patients with PMOP (n = 6). D The two-dimensional (2D)/3D images and quantitative analysis of micro-CT scans indicated significant bone loss in OVX mice (n = 3). E H&E staining showed that the bone microstructure of OVX mice was significantly damaged, accompanied by proliferation of bone marrow adipose tissue, consistent with the typical characteristics of postmenopausal osteoporosis (n = 3). F IF staining of the trabecular bone region in mouse distal femora clearly showed that the fluorescence intensity of integrin αVβ3 was significantly decreased in OVX mice (n = 4). G, H Western blot analysis of mouse bone tissue confirmed that the protein expression levels of integrin αV and β3 subunits were significantly decreased in OVX mice (n = 4). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Values presented as mean ± SD of at least three independent experiments

To further validate these findings, we established an OVX mouse model to mimic the pathological changes of PMOP. Micro-CT analysis revealed significant bone loss in OVX mice, characterized by markedly reduced bone mineral density (BMD), bone volume fraction (BV/TV), and trabecular number (Tb.N), along with increased trabecular separation (Tb.Sp) (P < 0.001, Fig. 1D). In addition, H&E staining demonstrated that the damage of femoral bone microstructure in OVX mice was aggravated, accompanied by hyperplasia of bone marrow adipose tissue, which was consistent with the typical characteristics of postmenopausal osteoporosis (Fig. 1E). Consistent with human findings, IF staining of the trabecular bone region in mouse distal femora clearly showed that integrin αVβ3 signal was predominantly localized to the osteoblast layer on the bone surface. This signal was significantly decreased in OVX mice compared with Sham controls (Fig. 1F). Western blot analysis further confirmed the downregulation of αV and β3 subunit proteins in OVX mouse bone tissue (Fig. 1G, H). These results suggest that downregulation of integrin αVβ3 may represent a key contributor to bone loss in postmenopausal osteoporosis.

Estrogen deficiency downregulates the expression of integrin αVβ3 in osteoblasts

To investigate how integrin αVβ3 influences the progression of PMOP by regulating bone metabolism, we first performed Masson and TRAP staining on the femur tissues of OVX and Sham mice. Masson staining revealed a significant reduction in collagen fiber deposition in the distal femur of OVX mice compared with Sham controls (Fig. 2A). TRAP staining showed no significant difference in osteoclast number between the two groups (Fig. 2B). To obtain direct, dynamic evidence of bone formation, we performed calcein labeling. This analysis demonstrated a significant reduction in the mineral apposition rate (MAR) in the cortical bone of OVX mice compared to Sham controls (P < 0.001; Supplementary Fig. S1). Collectively, the reduction in collagen deposition, coupled with a decreased MAR and the absence of increased osteoclast activity, strongly suggests that impaired osteoblast-mediated bone formation is a primary contributor to the early bone loss observed in our OVX model.

Fig. 2.

Estrogen deficiency downregulates integrin αVβ3 in osteoblasts and impairs bone formation. A Masson staining revealed significantly reduced collagen fiber deposition in OVX mice, indicating impaired osteoblast-mediated bone formation (n = 3). B TRAP staining showed no significant difference in osteoclast activity between the two groups (n = 3). C RT-qPCR demonstrated marked downregulation of both ITGAV and ITGB3 mRNA in osteoblasts following estrogen withdrawal. D, E Western blot analysis confirmed that the protein expression levels of integrin αV and β3 subunits were significantly decreased in osteoblasts after estrogen withdrawal. F IF staining further confirmed that the expression of integrin αVβ3 protein in osteoblasts decreased significantly after estrogen withdrawal (bar 10 μm). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Values presented as mean ± SD of at least three independent experiments

To further validate this hypothesis, we established an estrogen-withdrawal cell model using the MC3T3-E1 cell line and primary osteoblasts to simulate the pathological state of PMOP [22, 23], and to investigate changes in the expression level of integrin αVβ3. RT-qPCR analysis revealed significantly reduced mRNA levels of integrin subunits αV (ITGAV) and β3 (ITGB3) in osteoblasts following estrogen withdrawal (P < 0.0001, Fig. 2C). This suppression was particularly pronounced in primary osteoblasts compared with the MC3T3-E1 cell line, underscoring the greater sensitivity of primary cells to estrogen deprivation. Western blot (Fig. 2D, E) and cellular immunofluorescence staining (Fig. 2F) further confirmed that both integrin αVβ3 and its αV/β3 subunit proteins were markedly downregulated in osteoblasts following estrogen deprivation. These results align with our observations in patient with PMOP and OVX mouse bone tissues. These results suggest that estrogen deficiency may play a key role in the pathogenesis of PMOP by downregulating the expression of integrin αVβ3 in osteoblasts and impairing bone formation.

Integrin αVβ3 regulates osteoblast proliferation and apoptosis

Studies have demonstrated that integrin αVβ3 is a critical cell surface receptor widely expressed on endothelial cells, osteoblasts, and osteoclasts that plays essential roles in regulating cell proliferation, differentiation, adhesion, migration, and signal transduction [24–26]. To investigate the functional of integrin αVβ3 in osteoblasts, we established integrin αVβ3 knockout monoclonal cell lines (ITGAV^−/− and ITGB3^−/−) using CRISPR/Cas9 gene editing technology in both the MC3T3-E1 cell line and primary osteoblasts. RT-qPCR and Western blot analyses confirmed the successful knockout of integrin αV and β3 subunits at both transcriptional and protein levels using ITGAV-sgRNA 2/3 and ITGB3-sgRNA 1/3 (Supplementary Fig. S2A–J). Subsequently, Western blot analysis was performed to validate the knockout efficiency in monoclonal cells generated using ITGAV-sgRNA 3 and ITGB3-sgRNA 3. The results demonstrated complete absence of the corresponding αV or β3 subunit proteins, confirming successful establishment of ITGAV^−/− and ITGB3^−/− monoclonal cell lines (Fig. 3A–C). Interestingly, knockout of either the αV or β3 subunit led to reduced expression of its counterpart, suggesting interdependent regulation between the two subunits, consistent with the classical heterodimeric nature of integrin αVβ3. In addition, IF staining confirmed complete loss of integrin αVβ3 protein in both ITGAV^−/− and ITGB3^−/− cells (Supplementary Fig. S2K, L). To elucidate the regulatory mechanisms of integrin αVβ3 in osteoblasts, we performed Asral DIA proteomic sequencing on knockout integrin αVβ3 (ITGAV^−/− and ITGB3^−/−) and control osteoblasts. Differentially expressed proteins (DEPs) were identified using thresholds of |log₂(FC)|≥ 1 and P < 0.05. Proteomic analysis revealed significant alterations in protein expression profiles, visualized by heatmaps and volcano plots (P < 0.05, Supplementary Fig. S3A–D). Intersection analysis of DEPs from both comparison groups (ITGAV^−/− versus NC and ITGB3^−/− versus NC) identified 259 common DEPs (CO-DEPs) (Supplementary Fig. S3E). Subsequently, GO analysis of these 259 CO-DEPs demonstrated their significant enrichment in biological processes related to bone formation (Fig. 3D), suggesting that integrin αVβ3 is involved in regulating osteoblast-mediated bone formation processes.

Fig. 3.

Impaired expression of integrin αVβ3 in osteoblasts suppresses proliferation and promotes apoptosis. A-C Western blot detected the absence of integrin αV and β3 subunit protein expression in ITGAV^−/− and ITGB3^−/− monoclonal cell lines. D GO analysis of CO-DEPs revealed their predominant role in regulating bone formation rather than bone resorption during bone homeostasis. E, F Western blot demonstrated that integrin αVβ3 deficiency significantly downregulated proliferation-specific proteins (PCNA, Cyclin D, and CDK4) in osteoblasts. G, H Western blot analysis showed increased expression of pro-apoptotic proteins (Bax and Clv.cas 3) and decreased levels of the anti-apoptotic protein Bcl-2 in αVβ3-deficient osteoblasts. I–L Flow cytometry confirmed that the loss of integrin αVβ3 inhibited osteoblast proliferation and promoted apoptosis. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Values presented as mean ± SD of at least three independent experiments

Building upon these findings, we further evaluated the proliferative capacity and apoptotic status of both MC3T3-E1 cells and primary osteoblasts following integrin αVβ3 knockout. Western blot results showed that the levels of proliferation-specific proteins PCNA, Cyclin D1, and CDK4 were significantly decreased in osteoblasts lacking integrin αVβ3 expression (ITGAV^−/− and ITGB3^−/−) (P < 0.0001, Fig. 3E, F). Conversely, we observed markedly elevated levels of pro-apoptotic proteins Bax and Cleaved Caspase-3 (Clv.cas 3) (P < 0.001), coupled with decreased expression of the anti-apoptotic protein Bcl-2 (P < 0.05, Fig. 3G, H). In addition, flow cytometry analysis showed that, compared with the control group, the proliferation ability of integrin αVβ3 knockout cells was significantly decreased (P < 0.0001, Fig. 3I, J), while the apoptosis level was significantly increased (P < 0.0001, Fig. 3K, L). These results indicate that integrin αVβ3 knockdown leads to decreased proliferation and increased apoptosis of osteoblasts.

To further validate the effects of integrin αVβ3 on osteoblast proliferation and apoptosis, we established integrin αVβ3-overexpressing MC3T3-E1 cells and primary osteoblasts. Initial validation using RT-qPCR (Supplementary Fig. S4A) and Western blot analysis (Fig. 4A, B) demonstrated significant overexpression of both αV and β3 subunits at transcriptional and protein levels (P < 0.0001), confirming successful model establishment. Subsequent IF staining further verified markedly enhanced integrin αVβ3 protein expression (P < 0.0001, Fig. 4C, D). Notably, similar to the knockout experiments, overexpression of either ITGAV or ITGB3 led to concurrent upregulation of both integrin αVβ3 and its αV/β3 subunits, reinforcing the coordinated regulatory relationship between these two subunits. Western blot results showed that integrin αVβ3 overexpression significantly elevated expression levels of proliferation-specific proteins PCNA, Cyclin D, and CDK4 (P < 0.01, Fig. 4E, F). Conversely, pro-apoptotic proteins Bax and Clv.cas 3 were significantly downregulated while anti-apoptotic protein Bcl-2 was markedly upregulated (Fig. 4I, J). Flow cytometry analysis corroborated these findings, demonstrating enhanced proliferative capacity (Fig. 4G, H) and reduced apoptosis rates (Fig. 4K, L) in αVβ3-overexpressing osteoblasts. Importantly, these results showed consistent trends in both MC3T3-E1 cells and primary osteoblasts. Furthermore, no significant differences (P > 0.05) were observed between ITGAV^−/− and ITGB3^−/− cells, as well as between OE-ITGAV and OE-ITGB3 cells, indicating that αV and β3 subunits primarily function as the αVβ3 heterodimer in osteoblasts. These in vitro findings demonstrate that integrin αVβ3 is a critical regulator of osteoblast proliferation and apoptosis, suggesting its potential role in affecting osteoporosis progression, a conclusion further supported by our subsequent in vivo validation (Sect. “Overexpression of integrin αVβ3 alleviates OVX-induced postmenopausal osteoporosis in vivo”).

Fig. 4.

Overexpression of integrin αVβ3 enhances osteoblast proliferation and inhibits apoptosis. A, B Western blot analysis confirmed successful overexpression of integrin αV and β3 subunits. C, D Significant overexpression of integrin αVβ3 protein was confirmed by IF staining (bar 10 μm). E, F Western blot analysis revealed that αVβ3 overexpression significantly increased the expression of proliferation-specific proteins (PCNA, Cyclin D1, and CDK4) in osteoblasts. G, H Flow cytometry analysis verified enhanced proliferative capacity in αVβ3-overexpressing osteoblasts. I, J Western blot showed reduced pro-apoptotic proteins (Bax and Clv.cas 3) and increased anti-apoptotic Bcl-2 expression in αVβ3-overexpressing osteoblasts. K, L Flow cytometry analysis confirmed that overexpression of integrin αVβ3 inhibited cell apoptosis. *p < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Values presented as mean ± SD of at least three independent experiments

Integrin αVβ3 regulates Ptgs2 expression and mTOR activation

To further elucidate the molecular mechanisms by which integrin αVβ3 regulates osteoblast proliferation and apoptosis in PMOP progression, we analyzed transcriptome sequencing data from OVX versus Sham mouse models obtained from the Gene Expression Omnibus (GEO) database. The differentially expressed genes (DEGs) were extracted and intersected with the CO-DEPs obtained by proteomic sequencing, and we identified three key target molecules: Ptgs2, Acad12, and Ifit1 (Fig. 5A). Volcano plot analysis of the proteomic data revealed that, among these targets, Ptgs2 exhibited the most significant upregulation in integrin αVβ3-knockout osteoblasts (Fig. 5B). Ptgs2 (COX-2), a key enzyme in prostaglandin synthesis, plays an important role in inflammation and bone metabolism. Recent studies have highlighted its dual roles in diseases such as osteoarthritis (OA), rheumatoid arthritis (RA), and osteoporosis (OP), where it is involved in both inflammatory responses and the regulation of bone remodeling [27–29]. Notably, KEGG pathway analysis of CO-DEPs highlighted significant enrichment in the NOD-like receptor signaling pathway, mTOR signaling pathway, and calcium signaling pathway (P < 0.05, Fig. 5C). The mammalian target of rapamycin (mTOR) signaling pathway is the central pathway regulating cell growth, metabolism, and autophagy. It is involved in various pathological processes by forming two distinct complexes (mTORC1 and mTORC2) and plays a dual regulatory role in bone and joint diseases [30–32].

Fig. 5.

Integrin αVβ3 regulates mTOR activation via Ptgs2. A Intersection analysis of DEGs and CO-DEPs identified three key target molecules (Ptgs2, Acad12, and Ifit1). B Proteomic volcano plot revealed significant upregulation of Ptgs2 in osteoblasts following integrin αVβ3 knockout. C KEGG pathway analysis of CO-DEPs highlighted significant enrichment in NOD-like receptor signaling pathway, mTOR signaling pathway, and calcium signaling pathway. D, E Western blot showed that the loss of integrin αVβ3 expression upregulated Ptgs2 expression and inhibited mTOR phosphorylation. F–H Western blot confirmed that integrin αVβ3 overexpression inhibited Ptgs2 expression and promoted mTOR phosphorylation. I–K Western blot assays confirmed that Ptgs2 mediates the regulatory effect of integrin αVβ3 on mTOR activation. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Values presented as mean ± SD of at least three independent experiments

To further explore whether integrin αVβ3 targeted regulation of Ptgs2 expression and mTOR activation in osteoblasts, we performed analyses using integrin αVβ3 knockout and overexpression cells. Firstly, Western blot analysis showed that the Ptgs2 protein expression level was significantly increased in integrin αVβ3 knockout cells (P < 0.001); however, the phosphorylation level of mTOR protein (p-mtor protein) was significantly decreased (P < 0.05, Fig. 5D, E). Conversely, in integrin αVβ3 overexpressing cells, the expression level of Ptgs2 was significantly decreased (P < 0.05), while the protein expression level of P-mTOR was notably increased (P < 0.001), with statistically significant differences. In addition, it is noteworthy that the phosphorylation levels of S6K and 4EBP, two key effectors downstream of mTORC1, were also significantly increased after the overexpression of integrin αVβ3 (Fig. 5F–H). These results suggest that integrin αVβ3 may promote osteoblast proliferation and inhibit apoptosis by targeting the inhibition of Ptgs2 and/or activating the mTOR signaling pathway, thereby promoting phosphorylation of S6K and 4EBP1 to enhance protein synthesis.

Ptgs2 mediates the regulatory effect of integrin αVβ3 on mTOR

To further elucidate the regulatory relationship among integrin αVβ3, Ptgs2, and mTOR, we established Ptgs2 knockdown models (ITGAV^−/− + siPtgs2 and ITGB3^−/− + siPtgs2) using primary osteoblasts with integrin αVβ3 knockout (ITGAV^−/− and ITGB3^−/− cells). First, RT-qPCR analysis confirmed the success of knockdown of Ptgs2, and the difference was statistically significant (P < 0.05, Supplementary Fig. S4B–D). Subsequent western blot analysis revealed changes in key proteins of the Ptgs2 and mTOR signaling pathways across different cell groups. Compared with normal control cells, the protein expression levels of Ptgs2 in Ptgs2 knockdown cells were reduced by more than 70% (P < 0.05), demonstrating statistically significant differences and further validating the successful construction of the Ptgs2 knockdown model (Fig. 5I, K). Similarly, Ptgs2 expression was significantly decreased in Ptgs2 knockdown cells compared with integrin αVβ3 knockout cells (P < 0.001), which reversed the increase of Ptgs2 caused by integrin αVβ3 knockout. More importantly, the phosphorylation levels of mTOR and its key downstream effectors, S6K and 4EBP, were significantly increased in Ptgs2 knockdown cells, with highly statistically significant differences (Fig. 5I, J). These findings suggest that integrin αVβ3 in osteoblasts can activate the mTOR signaling pathway by inhibiting Ptgs2 protein expression.

Ptgs2 directly binds physically with mTOR to inhibit its phosphorylation

Previous studies have suggested a potential complex bidirectional regulatory relationship between Ptgs2 and the mTOR signaling pathway, forming a positive feedback loop between inflammation and metabolism [33]. To elucidate the specific molecular mechanism between Ptgs2 and mTOR, we first investigated their subcellular localization using immunofluorescence (IF) staining. The results revealed colocalization of Ptgs2 and mTOR (Fig. 6A), suggesting a potential direct interaction between them. To validate this hypothesis, we first confirmed the direct physical interaction between Ptgs2 and mTOR using endogenous co-immunoprecipitation (Co-IP) (Fig. 6B). Notably, this interaction was observed not only in integrin αVβ3-knockout cells but also in normal MC3T3-E1 cells and primary osteoblasts. Furthermore, compared with normal cells, the expression of Ptgs2 and mTOR (IP: Ptgs2 group) increased after knocking out integrin αVβ3. Intriguingly, the expression of Ptgs2 (IP: mTOR group) was significantly higher in the case of no significant difference in mTOR (IP: mTOR group) protein. These results indicating that the knockout of integrin αVβ3 enhanced the interaction capacity between Ptgs2 and mTOR. To further corroborate these findings, we employed bio-layer interferometry (BLI) to quantitatively analyze Ptgs2-mTOR binding affinity across different cell groups. Kinetic curves demonstrated specific binding between Ptgs2 and mTOR in all tested groups (NC, ITGB3^−/−, and ITGAV^−/−), with the lowest binding response (~0.6 nm thickness) in normal cells. The ITGB3^−/− group showed stronger binding (~ 1.1 nm), while ITGAV^−/− cells exhibited the strongest interaction (~1.4 nm) (Fig. 6C). These results indicated that knockdown of integrin αVβ3 not only upregulated Ptgs2 expression but also enhanced the ability of Ptgs2 to directly interact with mTOR.

Fig. 6.

Ptgs2 inhibits mTOR phosphorylation through direct interaction. A Two-photon laser confocal microscopy images demonstrate colocalization of mTOR (red) and Ptgs2 (green) in osteoblasts. B Co-IP confirmed the direct interaction between Ptgs2 and mTOR, which was enhanced by integrin αVβ3 deletion. C BLI analysis verified the binding capacity of Ptgs2 and mTOR in each group of cells, which confirmed that integrin αVβ3 knockout significantly enhance their direct interaction. D, E Molecular docking revealed stable binding affinity between Ptgs2 and mTOR (ΔG = −10.3 kcal/mol), with particular importance of the hydrogen bond formed between Ser2159 (mTOR) and Glu52 (Ptgs2). F, G IF staining combined with two-photon laser confocal microscopy confirmed that the Ptgs2–mTOR interaction inhibited mTOR protein phosphorylation. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Values presented as mean ± SD from at least three independent experiments

In addition, as shown in Fig. 5, the expression level of Ptgs2 protein was negatively correlated with the phosphorylation level of mTOR in osteoblasts with knockout of integrin αVβ3. To further investigate whether Ptgs2 suppresses mTOR phosphorylation through direct interaction, we performed semiflexible protein docking between Ptgs2 and mTOR using GRAMM. The results demonstrated that the binding energy was ΔG = −10.3 kcal/mol, which indicated that there was a stable binding potency between Ptgs2 and mTOR (Fig. 6D). In addition, the docking analysis revealed extensive hydrophobic interactions and hydrogen bonds (Fig. 6E), which are critical for the protein–protein interaction. Notably, a stable hydrogen bond formed between Ser2159 (mTOR) and Glu52 (Ptgs2) (Fig. 6D, E), potentially serving as the key mechanism by which Ptgs2 binding inhibits mTOR phosphorylation. Ser-2159 is one of the most crucial phosphorylation sites of mTOR [34–36], and its phosphorylation can promote autophosphorylation [36] and mTORC1 activity [35], thereby regulating cell proliferation, metabolism, and growth. We further performed IF staining for Ptgs2 and P-mTOR, and the images were observed by two-photon laser confocal microscopy. The results revealed that, although integrin αVβ3-deficient cells (ITGAV^−/− and ITGB3^−/− groups) showed significantly stronger Ptgs2 fluorescence intensity compared with normal cells, their activated P-mTOR levels on lysosomal membranes and in the cytoplasm were markedly lower (Fig. 6F, G). These findings are consistent with the results in Fig. 5. These results demonstrate that integrin αVβ3 regulates mTOR pathway activation by modulating both Ptgs2 expression and its direct interaction capacity with mTOR, ultimately influencing osteoblast proliferation and apoptosis.

Integrin αVβ3 regulates osteoblast proliferation and apoptosis via the Ptgs2–mTOR signaling axis

To further validate whether integrin αVβ3 regulates osteoblast proliferation and apoptosis through the Ptgs2–mTOR signaling axis, we conducted experiments using integrin αVβ3 knockout cells (ITGAV^−/− and ITGB3^−/− cells) and corresponding Ptgs2-knockdown cells (ITGAV^−/− + siPtgs2 and ITGB3^−/− + siPtgs2). Firstly, western blot analysis revealed that, compared with integrin αVβ3 knockout cells, Ptgs2 knockdown significantly increased the expression of proliferation-related proteins PCNA, Cyclin D, and CDK4 (Fig. 7A, B), while markedly decreasing the expression of pro-apoptotic proteins Bax and Clv.cas 3 and increasing anti-apoptotic protein Bcl-2 levels (Fig. 7C, D). Furthermore, flow cytometry analysis of osteoblast proliferation and apoptosis demonstrated that Ptgs2 knockdown reversed the proliferation inhibition and enhanced apoptosis caused by integrin αVβ3 knockout. Compared with integrin αVβ3-deficient cells, Ptgs2-knockdown cells exhibited significantly enhanced proliferative capacity (Fig. 7E, F) and reduced apoptosis (Fig. 7G, H). Combined with the above confirmed functional of integrin αVβ3 regulating osteoblast and molecular mechanism of integrin αVβ3 regulating Ptgs2–mTOR interaction, these results confirm that integrin αVβ3 regulates osteoblast proliferation and apoptosis through the Ptgs2–mTOR signaling axis.

Fig. 7.

Integrin αVβ3 regulates osteoblast proliferation and apoptosis via the Ptgs2-mTOR signaling axis. A–D Western blot and quantitative analysis demonstrates that Ptgs2 knockdown reversed the regulatory effects of integrin αVβ3 knockout on proliferation- and apoptosis-related proteins in osteoblasts. E–H Flow cytometry and quantitative analysis confirmed that low expression of Ptgs2 reversed osteoblast dysfunction caused by integrin αVβ3 deficiency, thereby promoting proliferation (E, F) and inhibiting apoptosis (G, H). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Values presented as mean ± SD of at least three independent experiments

Overexpression of integrin αVβ3 alleviates OVX-induced postmenopausal osteoporosis in vivo

Given the critical role of integrin αVβ3 in osteoblast proliferation and apoptosis, targeting integrin αVβ3 overexpression in vivo may provide a promising therapeutic strategy for PMOP. To evaluate the therapeutic potential of integrin αVβ3 in promoting bone formation and mitigating bone mass loss in vivo, we designed an animal experimental protocol to alleviate OVX-induced postmenopausal osteoporosis by localized overexpression of integrin αVβ3 in the distal femur of mice (Fig. 8A). Then, IF staining of integrin αVβ3 was performed on the femur tissue of mice in each group to detect its expression level. The IF results confirmed successful AAV9-mediated integrin αVβ3 overexpression in vivo (Fig. 8B, C). The results showed that the expression level of integrin αVβ3 in the femoral tissue of OVX + AAV9-NC group was significantly lower than in the Sham group (P < 0.05), which was consistent with our findings in Sect. “Integrin αVβ3 expression is decreased in PMOP.” Notably, mice with targeted integrin αVβ3 overexpression (OVX + AAV9-ITGAV, OVX + AAV9-ITGB3, and OVX + AAV9-ITGAV & ITGB3 groups) showed markedly higher integrin αVβ3 levels versus the OVX + AAV9-NC group (P < 0.0001), demonstrating the efficacy of AAV9-mediated in vivo delivery.

Fig. 8.

Overexpression of integrin αVβ3 alleviates OVX-induced postmenopausal osteoporosis in vivo. A Schematic diagram of the in vivo experimental protocol. B, C IF staining of integrin αVβ3 in the trabecular bone region of mouse distal femora from different groups confirms the effectiveness of AAV9-mediated integrin αVβ3 overexpression in vivo. D, E Micro-CT analysis demonstrated that integrin αVβ3 overexpression in vivo significantly reduced OVX-induced bone loss and increased trabecular bone volume. F H&E and Masson staining showed that integrin αVβ3 overexpression in vivo effectively reduced trabecular degeneration while enhancing collagen fiber deposition and osteogenesis. G, H IF staining of Ptgs2 (green) and P-mTOR (red) in the trabecular bone region confirmed that overexpression of integrin αVβ3 in vivo reduced Ptgs2 expression and enhanced mTOR phosphorylation. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Values presented as mean ± SD of at least three independent experiments

To further validate that targeted overexpression of integrin αVβ3 alleviates OVX-induced postmenopausal osteoporosis, we first analyzed femoral bone mass in experimental mice using micro-CT. The results showed that, compared with the Sham group, the OVX + AAV9-NC group exhibited significant bone loss in the femur, with markedly reduced bone mineral density (BMD), bone volume fraction (BV/TV), and trabecular number (Tb.N), along with significantly increased trabecular separation (Tb.Sp) (P < 0.0001). These results indicated that the OVX-induced postmenopausal osteoporosis mouse model was successfully constructed (Fig. 8D, E). Notably, targeted integrin αVβ3 overexpression in the distal femur (OVX + AAV9-ITGAV, OVX + AAV9-ITGB3, and OVX + AAV9-ITGAV&ITGB3 groups) significantly reversed these effects versus OVX + AAV9-NC controls. All treatment groups showed elevated BMD, BV/TV, and Tb.N, with reduced Tb.Sp (P < 0.05) (Fig. 8D, E). These results suggest that overexpression of integrin αVβ3 can significantly enhance trabecular bone volume and promote the osteogenesis process. Histological examinations (H&E and Masson staining) corroborated these findings. The results showed that, compared with OVX + AAV9-NC, the mice overexpressing integrin αVβ3 had significantly reduced trabecular degeneration, increased collagen deposition, and enhanced osteoblast activity (Fig. 8F). To quantitatively assess dynamic bone formation and osteoclast activity in response to integrin αVβ3 overexpression, we performed calcein labeling and TRAP staining analyses. Calcein labeling revealed that the mineral apposition rate (MAR) was significantly reduced in OVX + AAV9-NC mice compared with Sham controls (P < 0.0001; Supplementary Fig. S5A, B). Notably, AAV9-mediated overexpression of integrin αVβ3, particularly in the OVX + AAV9-ITGAV and OVX + AAV9-ITGAV&ITGB3 groups, markedly rescued the MAR (P < 0.0001; Supplementary Fig. S5A, B), providing direct evidence of enhanced osteoblast-mediated bone formation. Analysis of osteoclast parameters by TRAP staining showed that the number of osteoclasts was not significantly different between Sham, OVX + AAV9-NC, and OVX + AAV9-ITGAV groups. However, a significant increase in osteoclast number was observed in the OVX + AAV9-ITGB3 and OVX + AAV9-ITGAV&ITGB3 groups (Supplementary Fig. S5C). Despite this context-dependent increase in osteoclasts in groups overexpressing the β3 subunit, the net outcome across all treatment groups was a significant increase in bone mass (Fig. 8D, E) and improved microstructure (Fig. 8F), indicating that the potent pro-osteogenic effect dominated the bone remodeling process.

Additionally, IF staining was performed to assess Ptgs2 expression and mTOR phosphorylation levels following targeted integrin αVβ3 overexpression in mice. As shown in Fig. 8G, H, compared with the OVX + AAV9-NC group, all treatment groups (OVX + AAV9-ITGAV, OVX + AAV9-ITGB3, and OVX + AAV9-ITGAV&ITGB3) exhibited significantly reduced Ptgs2 expression and elevated P-mTOR levels in femoral tissues. Interestingly, the combined analysis of the results in Fig. 8B, C, G, H showed that the OVX + AAV9-NC group revealed significantly decreased integrin αVβ3, increased Ptgs2, and reduced mTOR phosphorylation versus Sham controls. Conversely, integrin αVβ3-overexpressing OVX mice demonstrated upregulated integrin αVβ3, downregulated Ptgs2, and enhanced mTOR phosphorylation relative to OVX + AAV9-NC. These in vivo findings precisely mirror our in vitro osteoblast results, confirming the conserved regulatory mechanism of the integrin αVβ3–Ptgs2–mTOR axis.

In summary, our findings demonstrate that estrogen deficiency impaired osteoblast function by inhibiting the “integrin αVβ3–Ptgs2–mTOR” signaling axis, ultimately leading to bone loss and the progression of PMOP. As illustrated in Fig. 9, normal estrogen expression promotes integrin αVβ3 expression, which downregulates Ptgs2 to relieve its inhibitory effect on mTOR. This subsequently activates the mTORC1–S6K/4EBP1 pathway, enhancing osteoblast proliferation and suppressing apoptosis. Consequently, these mechanisms promote osteogenesis and mitigate PMOP progression.

Fig. 9.

Schematic illustration of the regulation of osteoblast proliferation and apoptosis by estrogen through the “integrin αVβ3–Ptgs2–mTOR (mTORC1–S6K/4EBP1)” axis

Discussion

Integrin αVβ3, as an extracellular matrix receptor involved in signal transduction between cells and the bone matrix, is one of the key molecules regulating bone metabolism [9, 10, 12, 37]. The expression and function of integrins can be modulated by hormonal signals, including estrogen. Estrogen receptors (ERα and ERβ) are known to transcriptionally regulate genes involved in cell adhesion and cytoskeletal organization [4, 5]. It is plausible that estrogen deficiency in PMOP leads to diminished transcriptional activation or altered post-translational modification of integrin αVβ3, thereby impairing its membrane localization, stability, or affinity for ECM ligands. This disruption in estrogen–integrin signaling likely constitutes an initial critical event in the pathogenesis of PMOP, compromising osteoblast–ECM communication and subsequent survival/proliferation signals. Our previous study demonstrated that the loss of integrin αVβ3 severely impairs the proliferation, differentiation, migration, and adhesion of osteoblasts [38]. Building on this foundation, the current research systematically revealed the critical role of integrin αVβ3 in postmenopausal osteoporosis (PMOP) through clinical samples, animal models, and cellular experiments. The results showed that the expression level of integrin αVβ3 was significantly downregulated in bone tissues of patients with PMOP and OVX mice, which was closely related to the impaired function of osteoblasts. Furthermore, this study is the first to reveal the precise regulation mechanism of the integrin αVβ3–Ptgs2–mTOR signaling axis in osteoblasts. Our findings not only confirmed the direct physical interaction between Ptgs2 and mTOR but, more importantly, elucidated the molecular mechanism by which integrin αVβ3 regulates mTOR activity through this interaction. Additionally, by constructing an in vivo overexpression model of integrin αVβ3 in mouse femurs, we demonstrated that overexpression of integrin αVβ3 effectively alleviates OVX-induced bone loss. These findings provide a new perspective to understand the molecular mechanism of PMOP; that is, estrogen deficiency may inhibit bone formation by suppressing the integrin αVβ3–Ptgs2–mTOR axis, ultimately leading to bone mass loss and accelerated PMOP progression. Moreover, our findings highlight integrin αVβ3 as a potential therapeutic target.

Previous studies have shown that the expression of integrin αVβ3 is closely related to the balance of bone metabolism and the development of osteoporosis [12, 13]. For example, Voisin et al. [16] found that, in an estrogen-deficient postmenopausal osteoporosis rat model, the number of osteocytes containing integrin β3 in cortical bone was lower than in the control group. Our study further demonstrated that the reduced integrin αVβ3 expression may be one of the key factors contributing to bone loss in patients with PMOP and OVX mice (Fig. 1). Immunofluorescence (IF) staining and Western blot analysis of bone tissues from patients with PMOP revealed significantly lower expression levels of integrin αVβ3 and its subunits (αV and β3) compared with premenopausal non-osteoporosis patients. We further verified this conclusion by using the OVX mouse model and found that the expression of integrin αVβ3 was significantly decreased in the femoral tissue of OVX mice, which was closely related to the typical features of PMOP, such as reduction of the trabecular number and bone structure destruction.

Regarding integrin αVβ3, as a crucial receptor for the extracellular matrix (ECM), previous studies have focused on its role in osteoclast-mediated bone resorption [14, 39, 40]. For instance, Li et al. [14] demonstrated that, under acidic conditions, nuclear factor of activated T cells 1 (NFATc1) is activated and upregulates integrin αvβ3 expression, promoting osteoclast migration, adhesion, and bone resorption via the integrin αvβ3/Pyk2/Src axis. However, our study further revealed its potential role in the regulation of osteoblast function. Our study found that estrogen deficiency not only reduced the expression of integrin αVβ3 in osteoblasts but also inhibited the process of bone formation. In estrogen-deficient MC3T3-E1 cell and primary osteoblast models, both the transcriptional and protein expression levels of integrin αVβ3 were significantly decreased (Fig. 2C–F). Meanwhile, TRAP staining and Masson staining analyses of femoral tissues from OVX mice showed that the activity of osteoclasts and bone resorption process were not significantly enhanced, while the deposition of collagen fibers was significantly reduced and the function of osteoblast-mediated bone formation was significantly impaired in OVX mice. Our dynamic histomorphometric analysis revealing a decreased MAR, along with static evidence of reduced collagen formation and unaltered osteoclast numbers, supports the conclusion that impaired bone formation is a key pathogenic feature in the early stage (8 weeks) of our OVX model (Fig. 2A, B; Supplementary Fig. S1). This finding aligns with the results of Verbruggen et al. [41], who observed that, after short-term estrogen deficiency (5 weeks), osteoporotic osteocytes experienced higher maximal strain compared with healthy osteocytes, with a significantly greater proportion exceeding the osteogenic strain threshold (10,000 με) (15.74% versus 5.37%), accompanied by markedly weakened bone formation capacity. This finding may explain why bisphosphonate treatment, which primarily targets osteoclasts, has limited efficacy in some patients with PMOP.

Previous studies have demonstrated that integrin αVβ3 affects the adhesion, migration, and growth of osteoblasts by interacting with extracellular matrix, which is essential for bone repair and the balance of bone metabolism [24, 42, 43]. In this study, we employed CRISPR/Cas9 gene editing technology and proteomic analysis to investigate its effects on osteoblast proliferation and apoptosis at the molecular level, providing novel insights into the mechanisms of osteoporosis. Our experiments revealed that knockout of integrin αVβ3 led to decreased proliferation and increased apoptosis in both MC3T3-E1 cells and primary osteoblasts (Fig. 3). Further proteomic data analysis provided direct molecular evidence for the role of integrin αVβ3 in the regulation of osteoblast function. Additionally, overexpression experiments demonstrated that elevated integrin αVβ3 expression effectively enhanced osteoblast proliferation and suppressed apoptosis, further underscoring its functional importance (Fig. 4). Notably, our findings also revealed a synergistic regulatory relationship between the αV and β3 subunits, consistent with the classic heterodimeric characteristics of integrins [44, 45].

Through multidimensional experimental verification, this study reveals for the first time the precise regulation mechanism of the integrin αVβ3–Ptgs2–mTOR signaling axis in osteoblasts. First, by integrating proteomic and transcriptomic analyses with experimental verification, we identified Ptgs2 as a key downstream effector molecule of integrin αVβ3 (Fig. 5). Our results demonstrated that knockout of integrin αVβ3 significantly upregulates Ptgs2 expression, while this effect can be reversed by integrin αVβ3 overexpression. This finding aligns with Kim et al.’s study, which showed that treatment of interleukin (IL)−1β-induced inflamed chondrocytes (Chon-001 cells) with integrin αvβ3 markedly reduced the expression of Ptgs2. Furthermore, we confirmed that integrin αVβ3 promotes protein synthesis and cell proliferation by activating the mTOR (mTORC1–S6K/4EBP1) signaling pathway. This result is consistent with Lee et al.’s research, which demonstrated that mechanical stress triggers sustained activation of the PI3K/Akt/mTOR/S6K signaling cascade through the integrin αVβ3–FAK pathway, thereby enhancing osteoblast proliferation [46]. Notably, our study revealed that Ptgs2 knockdown can rescue mTOR signaling suppression and osteoblast dysfunction caused by integrin αVβ3 deficiency, suggesting that Ptgs2 may serve as a “molecular switch” in the integrin αVβ3–mTOR signaling axis (Fig. 6). Our findings further highlighted the specific role of integrin αVβ3 in maintaining mTOR signaling homeostasis: on the one hand, it suppresses Ptgs2 to relieve its negative regulation on mTOR; on the other hand, it may directly activate mTOR through the FAK/PI3K/AKT pathway. This dual regulatory mechanism ensures the precision and stability of signal transduction.

Secondly, the most breakthrough finding of this study is that the direct binding between Ptgs2 and mTOR was confirmed by Co-IP and BLI techniques. Through structural simulation, it was found that Ptgs2 formed a stable hydrogen bond (Ser2159–Glu52) with Ser2159 of mTOR through its Glu52, thereby directly interfering with mTOR phosphorylation. Ser-2159 is one of the most crucial phosphorylation sites of mTOR [34–36], and its phosphorylation can promote autophosphorylation [36] and mTORC1 activity [35], thereby regulating cell proliferation, metabolism, and growth. Furthermore, our two-photon confocal microscopy results demonstrated that Ptgs2 overexpression reduces P-mTOR localization on lysosomal membranes—a finding consistent with the known requirement of lysosomal localization for mTOR activation [47–49]. Particularly noteworthily, our BLI experiments revealed that integrin αVβ3 knockout significantly strengthens the Ptgs2–mTOR binding affinity, with binding thickness increasing from 0.6 nm to 1.4 nm. These quantitative data provide crucial evidence for understanding the kinetic characteristics of signal transduction (Fig. 7). Collectively, these findings support a novel regulatory model: deficiency of integrin αVβ3 upregulates Ptgs2 expression, enhances the direct binding interaction of Ptgs2 with mTOR, obstructs Ser2159 phosphorylation of mTOR, and ultimately attenuates downstream signaling.

This study systematically demonstrated the therapeutic potential of AAV9-mediated integrin αVβ3 overexpression for postmenopausal osteoporosis (PMOP) through in vivo experiments. The results showed that targeted overexpression of integrin αVβ3 in the distal femur significantly increases trabecular bone volume, improves bone mineral density and trabecular number, reduces trabecular degeneration, enhances osteoblast activity, and promotes collagen deposition (Fig. 8A–F). These findings suggest the potential therapeutic effect of integrin αVβ3 in bone formation processes. Consistent with our observations, Qin et al. [11] found that deletion of β3 integrin in osteocytes of adult mice significantly reduced the osteoblast-mediated bone formation rate and impaired osteogenic differentiation of bone marrow stromal cells in the bone microenvironment, leading to decreased bone mass and compromised biomechanical properties of weight-bearing long bones. However, it should be noted that other studies [26, 40] have shown that integrin αVβ3 antagonists can alleviate bone loss and improve osteoporosis symptoms by inhibiting osteoclast formation and differentiation. Our findings in osteoblasts expand the potential applications of integrin αVβ3 in osteoporosis treatment.

This study also observed a phenomenon that is worthy of in-depth exploration: AAV9-mediated in vivo overexpression of integrin αVβ3 significantly promoted bone formation while concurrently increasing osteoclast numbers (Supplementary Fig. S5). However, the final bone metabolic phenotype demonstrated a net increase in bone mass (Fig. 8D–F), indicating that the pro-osteogenic effect of αVβ3 overexpression predominated. This phenomenon suggests a dual role for integrin αVβ3 in bone remodeling. On the one hand, as elucidated in this study, signaling through the αV subunit robustly enhances osteoblast function via the Ptgs2–mTOR axis. On the other hand, the β3 subunit, a key regulator of osteoclast function, may be directly or indirectly modulated by its overexpression. We hypothesize that this aligns with the theory of “osteoblast-mediated bone resorption coupling,” whereby enhanced osteoblast activity may indirectly stimulate osteoclastogenesis as an adaptive response through coupling mechanisms such as an increased RANKL/OPG ratio [50, 51]. Furthermore, the established role of αVβ3 in osteoclast precursor adhesion and migration [52–54] may also contribute to this observation. Thus, integrin αVβ3 functions as a complex in which the αV and β3 subunits may finely regulate bone formation and resorption, respectively, working in concert to maintain bone remodeling balance. Our therapeutic strategy, utilizing localized, AAV-mediated overexpression, ultimately achieved a significant anabolic advantage, providing a new perspective on targeting αVβ3 for osteoporosis treatment—namely, that its efficacy may stem from a potent stimulation of bone formation rather than solely from inhibiting bone resorption.

This study further validated the close correlation between the integrin αVβ3–Ptgs2–mTOR signaling axis and osteoporosis in animal models (Fig. 8G, H). Our results demonstrated that AAV9-mediated integrin αVβ3 overexpression significantly reduces Ptgs2 levels (by approximately 50%) and enhances mTOR phosphorylation (approximately 2.5-fold increase), with these changes showing significant positive correlation with improved bone microstructure. In addition, from a pathophysiological perspective, our findings reveal a molecular bridge connecting inflammation and metabolic disorders in osteoporosis. Our results showed a negative feedback regulation between increased Ptgs2 expression and decreased mTOR activity, which is consistent with the clinically observed chronic low-grade inflammatory state in bone tissues of osteoporosis patients [55].

Although this study has confirmed the crucial role of the “integrin αVβ3–Ptgs2–mTOR” signaling axis in regulating osteoblast-mediated bone formation, several limitations remain. First, the direct regulatory relationship between estrogen receptors (ERα and ERβ) and integrin αVβ3 has not been fully elucidated. Estrogen is known to modulate integrin expression in various cell types through genomic and nongenomic pathways mediated by ERα and ERβ. Future studies should investigate whether estrogen receptors directly transcriptionally regulate ITGAV or ITGB3 gene expression, or whether they influence integrin αVβ3 stability and signaling through post-translational modifications or interaction with focal adhesion complexes. Understanding the crosstalk between ERα/ERβ and integrin αVβ3 will provide deeper insights into the molecular basis of estrogen-deficient osteoporosis. Furthermore, while our molecular docking predicted a critical hydrogen bond between Ptgs2-Glu52 and mTOR-Ser2159, site-directed mutagenesis experiments are required to definitively confirm its functional relevance. Second, although our data primarily support mTORC1 involvement based on downstream S6K/4EBP1 phosphorylation, we cannot fully exclude potential contributions from mTORC2. Future studies should examine Raptor/Rictor binding and mTORC2-specific substrates to clarify this distinction.

Conclusions

This study establishes the critical role of the integrin αVβ3–Ptgs2–mTOR signaling axis in PMOP, providing both a theoretical foundation and potential therapeutic targets for developing novel bone-forming therapies. For the first time, our research confirmed that the direct interaction between Ptgs2 and mTOR inhibits mTOR activation, proposing a new regulatory model in PMOP progression: estrogen deficiency downregulates integrin αVβ3 expression, promotes Ptgs2 expression, and enhances its direct binding with mTOR, thereby impeding mTOR phosphorylation and subsequent activation of the mTORC1–S6K/4EBP1 pathway. This process inhibits osteoblast proliferation while promoting apoptosis, ultimately accelerating osteoporosis progression.

Abbreviations

OP

Osteoporosis

PMOP

Postmenopausal osteoporosis

OVX

Ovariectomy

ECM

Crucial extracellular matrix

RANKL

Receptor activator of nuclear factor-κB ligand

Ptgs2

Prostaglandin-endoperoxide synthase 2

Co-IP

Co-immunoprecipitation

BLI

Bio-layer interferometry

IF

Immunofluorescence

KEGG

Kyoto Encyclopedia of Genes and Genomes

GO

Gene Ontology

DEPs

Differentially expressed proteins

Author contributions

C.C.: Writing—original draft, conceptualization, visualization, methodology, investigation. J.G. and C.Y.: Writing—original draft, visualization, methodology, investigation. F.Y. and Z.L.: Formal analysis, investigation, methodology. L.W. and R.C.: investigation, methodology. B.G. and Y.X.: Writing—review and editing, supervision, project administration, funding acquisition, conceptualization. All authors have read and agreed to the published version of the manuscript.

Funding

This study was supported by National Natural Science Foundation of China (82,060,405 and 82,360,436), Lanzhou Science and Technology Plan Program (2021-RC-102), Natural Science Foundation of Gansu Province (24YFFA043, and 23JRRA1500), and Cuiying Scientific and Technological Innovation Program of Lanzhou University Second Hospital (CY2022-MS-A19), Key Talents Program of Gansu Province (2025RCXM082), Talent Innovation and Entrepreneurship Project in Chengguan District, Lanzhou City (2025-rc-8).

Data availability

The datasets utilized and/or analyzed during the present study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The study protocol was performed in accordance with the Declaration of Helsinki, and written informed consent was obtained from all participants. The study protocol was approved by the Medical Ethics Committee of the Second Hospital of Lanzhou University (approval no. 2024 A-1265), approved on 10 October 2024.

All animal procedures were conducted following the recommendations of the International Council for Laboratory Animal Science (ICLAS) guidelines. All experimental designs and protocols involving animals were approved by the Animal Care Committee of the Second Hospital of Lanzhou University (approval no. D2024-920) on 8 October 2024.

Consent for publication

All authors gave their consent for publication.

Competing interests

The authors declare no competing interests.