Ethics statement

This study was approved by the Ethics Committee of the Eighth Affiliated Hospital Sun Yat-sen University, Shenzhen, China (Approval No. 2022-084-01). All patients and volunteers were made aware of the study procedures and potential risks, and informed consent was obtained. Animal experiments were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC), Sun Yat-Sen University (SYSU-IACUC-2022-002487). All the experiments performed in this study were conducted in accordance with the committee’s regulations and guidelines.

Isolation and culture of BMSCs

All the participants signed informed consent forms. The experimental protocol was approved by the institutional ethics committee of The Eighth Affiliated Hospital of Sun Yat-sen University. Bone marrow aspirates (10–20 ml) from each donor were seeded in 5-cm2 flasks containing medium (DMEM (HyClone, USA), 10% foetal bovine serum (Gibco, USA), 100 U/ml penicillin G and 100 mg/ml streptomycin [HyClone, USA]) in a cell incubator at 37 °C. After 3–5 days of culture, the supernatant and non-adherent cells were removed. The attached cells were detached with 0.25% trypsin (Gibco, USA) and reseeded into new 75 cm2 flasks. BMSCs at passages 3–5 were used for the experiments on differentiation and proliferation.

Collection of human bone samples

Osteoporosis samples from patients with aging osteoporosis who underwent surgery were collected. Osteopenia group as well as normal control samples from patients who underwent surgery after an accident were acquired. T-score evaluated after a dual energy X-ray absorptiometry (DXA) examination on the lumbar spine was taken as criterion to classify the populations into three groups: young control group (healthy donor, n = 11, mean age = 22.5 ± 3.8 y), osteopenia group (n = 9, mean age = 55.4 ± 2.9 y, mean T-score = -1.70 ± 0.50) and osteoporotic group (n = 10, mean age = 76.0 ± 7.8 y, mean T-score = -2.59 ± 0.30). After obtaining informed consent, we acquired the bone tissues. Sections and protein of bone tissue of different groups were acquired. BMSCs of different groups were isolated.,

Adipogenic and osteogenic differentiation of BMSCs

BMSCs were seeded at 0.5–0.8 × 105 cells/well in 12-well plates, and the osteogenic medium (DMEM supplemented with 10% FBS, 100 IU/ml penicillin, 100 IU/ml streptomycin, 0.1 mM dexamethasone, 10 mM β-glycerol phosphate, and 50 mM ascorbic acid (Sigma-Aldrich, St. Louis)) was replaced every 3 days. For adipogenic induction, BMSCs were induced in adipogenic medium (high-glucose DMEM supplemented with 10% FBS, 1 mM dexamethasone (Sigma-Aldrich), 10 mg/ml insulin (Sigma-Aldrich), 0.5 mM 3-isobutyl-1-methylxanthine (Sigma-Aldrich), and 0.2 mM indomethacin (Sigma-Aldrich)), with the medium replaced every 3 days.

ALP, ARS and ORO staining and quantification

For alkaline phosphatase (ALP) staining, BMSCs, osteogenically differentiated for 7 days, were fixed in 4% formaldehyde, stained using a BCIP/NBT alkaline phosphatase kit (Beyotime Institute of Biotechnology) in accordance with the manufacturer’s instructions, and then imaged using a Nikon TS-2 microscope. For the ALP activity assay, ALP activity kits (Nanjing Jiancheng Biotech) were used in accordance with the manufacturer’s instructions. Briefly, BMSCs osteogenically differentiated for 7 days were lysed using RIPA buffer, the supernatants were incubated with reaction buffer for 15 min, and then stop solution was added. The absorbance was measured at 405 nm. The total protein concentration was detected by a bicinchoninic acid (BCA) protein assay kit (Thermo Fisher Scientific). ALP activity was normalized to the total protein content and reported as units per gram of protein per 15 min (U/g pro/15 min).

For Alizarin Red S (ARS) staining, BMSCs cultured in osteogenic medium for 14 days were first fixed in 4% paraformaldehyde for 30 min and then stained with 1% ARS (A5533, Sigma-Aldrich, pH 4.3) for 30 min at room temperature. After being washed at least three times with phosphate-buffered saline (PBS), the stained cells were observed and photographed under a Nikon TS-2 microscope. Then, the cells were destained with 10% cetylpyridinium chloride monohydrate (Sigma-Aldrich) for 1 h at room temperature; 200 µL of the extracted liquid was transferred to a 96-well plate, and the absorbance was measured at 562 nm.

For Oil Red O (ORO) staining, BMSCs subjected to adipogenic induction for 14 days were fixed with 4% paraformaldehyde for 30 min. Then, ORO working solution (O0625, Sigma-Aldrich, 10 mg/ml) was added and incubated with the fixed cells for 20 min at room temperature. The stained cells were observed under a microscope. Quantitative analysis was performed by extraction with isopropanol and subsequent measurement of the absorbance at 520 nm.

ASOs transfection in vitro

In the circSTX12 ASO group, 1.0 × 105 BMSCs were transfected with circSTX12 ASO (50 nM) or control ASO (NC) when cell confluence reached 60 to 80%. The circSTX12 ASO sequence (3’-GAAGUCUUGUCGACGAGUCU-5’) and the NC sequence were designed and synthesized by Guangzhou Ruibo Biological Co., Ltd.

RNA interference

Small interfering RNAs (siRNAs) targeting CBL and LMO7 and negative controls were purchased from IGEbio (Guangzhou, China). When cell confluence reached 60 to 80%, BMSCs were transfected with Lipofectamine™ RNAiMAX (Invitrogen) and siRNA cocktail (1 OD/1.8 × 106 cells) in Opti-MEM reduced serum medium according to the manufacturer’s instructions. qPCR was detected to test the level of the targeted RNAs 48 h after transfection. The siRNA with the highest knockout rate was used in the followed experiments. The sequences of these siRNAs are shown in Supplementary Table 1.

Plasmid construction and transfection

Expression plasmid pLC5-ciR-circSTX12 was constructed by Geneseed (Guangzhou, China). pcDNA3.4(+)-MST1 were constructed by IGE Biotechnology (Guangzhou, China). Transfection was performed using a Lipofectamine 3000 Transfection Kit (Invitrogen, MA, USA) according to the manufacturer’s instructions with minor modifications. Briefly, P4 BMSCs were seeded in 12-well plates at a density of 0.8 × 105 cells/well. A total of 0.5 µg plasmid supplemented with 1 µg Lipofectamine 3000 (Thermo Scientifc™) and 1 µg p3000 (Thermo Scientifc™) mixed in Opti-MEM (Gibco) was added to each well. 8 h after transfection, the culture medium was replaced.

YAP inhibitor treatment

BMSCs of passge 4 were seeded in 12-well plates at a density of 0.5–0.8 × 105 and treated with different concentrations (1.25, 2.5, and 5nM) of YAP Inhibitor (MCE, Italy) dissolved in the osteogenic or adipogenic medium for 14 days. After 14 days, the BMSCs was observed under a microscope and ARS or ORO staining assays were performed followed the above intruduction.

RNA extraction, reverse transcription, and quantitative realtime PCR

Total RNAs were extracted from tissues or cells using TRIzol reagent (Thermo Fisher, USA). RNAs from nucleus and cytoplasm of BMSCs were separated by the PARIS™ Kit (Life Technologies, USA) following the manufacturer’s instructions. The purity and concentration of the RNA were spectrophotometrically analysed using a NanoDrop One (Thermo Fisher, USA). To determine the expression levels of circRNAs, 4 U/µg of RNase R (Beyotime, China) was added to digest linear transcripts. cDNA was transcribed using the PrimeScript RT reagent kit (TaKaRa, Japan). To determine mRNA expression, cDNA was transcribed using a First Strand cDNA Synthesis Kit (Beyotime, China). PCR was performed on an ABI 7500 Real-Time PCR system (Thermo Fisher, USA) using TB Green Premix Ex Taq II (Taraka, Japan). GAPDH was used as the reference gene for mRNAs or circRNAs in cell extracts. The primers shown in Supplementary Table 2 were utilized for qPCR.

Protein extraction

The cells were washed three times with PBS and then lysed in RIPA buffer (CWBIO, Jiangsu, China) supplemented with protease inhibitors (CWBIO) and phosphatase inhibitors (CWBIO) for 30 min on ice. Total protein was acquired by centrifuging the lysates at 14,000 rpm for 10 min at 4 °C. The soluble material was collected, and the protein concentration was measured using a Pierce BCA protein assay kit (Thermo Scientific™). Nuclear and cytoplasmic proteins were separated with NE-PER™ Nuclear and Cytoplasmic Extraction Reagents (Thermo Scientific™) according to the manufacturer’s instructions, and protein concentrations were determined as described above.

Western blot

Equal amounts of protein samples were diluted with 5 × SDS-PAGE Sample Loading Buffer (Beyotime, P0015) and subsequently transferred to polyvinylidene difluoride (PVDF) membranes (Merck Millipore, IPVH00010). Then, the membranes were transferred to Tris-buffered saline with Tween (TBST) (10 mM Tris-HCl, 15 mM NaCl, 0.05% Tween-20, pH 7.5) solution and washed three times. Afterwards, the membranes were blocked with 5% skim milk (Wako, 190-12865), incubated for 60 min at room temperature, and then incubated with primary antibodies against GAPDH (CST, 5174 S, 1:1000), CBL (Abcam, ab32027, 1:1000) osteocalcin (OCN; Abcam, ab133612, 1:1000), RUNX2 (CST, 12556 S, 1:1000), OSX (Abcam, ab209484, 1:1000), YAP (Affinity Biosciences, DF3182-200, 1:1000), phospho-YAP (Biorigin, BN42106R, 1:1000), LATS1 (SAB, C36187, 1:1000), BMP2 (Abcam, ab284387, 1:1000), Wnt1 (Abcam, Wnt1, #63934, 1:1000), Gli (SAB, 49399-1, 1:1000), Gli2 (SAB, 29217, 1:1000), MST1 (CST, bs-8477R1:1000), β-Catenin (ABclonal, A19657, 1:1000), p-Smad1/Smad5/Smad8 (SAB, 21684, 1:1000), OPN (SAB, 42036, 1:1000), Osteocalcin (Abcam, ab93876, 1:1000), FABP4 (Abcam, ab92501, 1:1000), CEBPA (Abcam, ab40764, 1:1000), PPARG (Abcam, ab178860, 1:1000), and Ubiquitin (Abcam, 134,953, 1:1000) overnight in a 4 °C refrigerator. After washing three times with TBST, horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG (ComWinBiotech, cw0102s, 1:3000) or goat anti-rabbit IgG (ComWin Biotech, cw0103s, 1:3000) secondary antibodies were added and incubated for 60 min at room temperature on a shaker. Finally, the membranes were washed three times with PBS, and the immunoreactive protein bands were detected with Immobilon Western Chemiluminescent HRP Substrate (Merck Millipore, WBKLS0500). Relative quantitative analysis was performed by ImageJ software.

RNA pulldown and mass spectrometry

To identify the proteins that interact with circSTX12, an RNA pull-down assay was performed. Biotin-labelled oligonucleotide probes (CTTTATCTGAGCAGCTGTTC-/3bio/) and antisense probes (GAACAGCTGCTCAGATAAAG-/3bio/) targeting circSTX12 ligation sites for RNA pulldown and mass spectrometry were synthesized by RiboBio (Guangzhou, China). RNA pull-down assays were performed with the PureBindingTM RNA-protein pull-down kit (Geneseed) according to the manufacturer’s protocol. Briefly, streptavidin-coated magnetic beads were incubated with biotinylated probes for 1 h at 4 °C. Afterwards, the beads were separated using magnets and incubated with the lysate from 1.0 × 107 BMSCs (in lysis buffer) for an additional 2 h at 4 °C. The probe-RNA-protein complex was pulled down and separated using a magnet. Proteins in the complexes were separated from the magnetic beads by boiling for 10 min and then subjected to Western blot or mass spectrometry (Medical Research Center, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, China).

RNA Immunoprecipitation (RIP) assay

Interactions between CBL, LMO7 and circSTX12 were detected with an RNA immunoprecipitation kit (Geneseed, Guangzhou, China) applied according to the manufacturer’s instructions. Briefly, 2 × 107 cells were collected, lysed with RIP lysis buffer, and then incubated with IgG-conjugated magnetic beads (Cell Signaling Technology, Beverly, MA, USA) or antibodies against CBL or LMO7. Coprecipitated RNA was detected by RT-qPCR. The sequences of the primers used are listed in Supplementary Table 1.

Coimmunoprecipitation

BMSCs or HEK293T cells were quickly harvested and homogenized on ice in modified RIPA buffer (Beyotime biotechnology). The cell extracts (approximately 200 µg of total protein) were incubated with antibodies against CBL (Abcam, ab32027, 1:1000), MST1 (CST, bs-8477R, 1:1000), or their IgG controls (Cell Signaling Technology, 3452 or 37988) at 4 °C overnight. Then, protein-G agarose beads (40 L, Beyotime Biotechnology) were added, and the mixture was incubated at 4 °C for 3 h. The agarose beads were collected, washed, and resuspended in 60 µL of sample buffer containing 50 mM Tris-HCl, pH 7.6, 2% (wt/vol) SDS, 10% (vol/vol) glycerol, 10 mM DTT, and 0.2% bromophenol blue. Then, the samples were boiled for 10 min. SDS-PAGE was used to separate the samples. The Western blot protocols used are described above. A special secondary antibody (1:1000, Abcam, ab131366) that recognizes only native (nonreduced) antibodies to minimize the detection of heavy and light chains was used to test the IP samples to avoid the influence of the IP antibodies.

RNA fluorescence in situ hybridization (FISH)

RNA FISH was performed with the Fluorescent in Situ Hybridization Kit (RiboBio, Guangzhou, China) according to the manufacturer’s instructions. First, BMSCs were seeded in 12-well plates, covered with sterile glass, and cultured overnight. Then, the cells were fixed with 4% paraformaldehyde for 10 min and permeabilized with cold 0.5% Triton for another 5 min. After being rinsed 3 times, the BMSCs were prehybridized with prehybridization buffer for 30 min at 37 °C. A Cy3-labelled circSTX12 probe (TATCTGAGCAGCTGTTCTGAAGTAG) was constructed by Geneseed (Guangzhou, China). The probe was dissolved in hybridization buffer at 20 µM. Then, the slides were hybridized in hybridization buffer overnight. After sequential washes to remove the unconjugated probe, the cells were stained with 4′,6-diamidino-2-phenylindole (DAPI) and observed under an LSM 880 laser scanning confocal microscope.

H2O2induced senescence

BMSCs of passage 4 were seeded in 12-well plates at a density of 0.7 × 105 cells/well. Then, the cells were treated with 600 µM H2O2 for 2 h to induce senescence. Subsequent experiments were performed on day 7 after induction.

5-Ethyl-2’-deoxyuridine (EdU) incorporation assay

BMSCs of passage 4 were seeded in 24-well plates at a density of 0.4 × 105 cells/well. BMSCs were transfected with siRNAs, ASOs or plasmids and then evaluated for proliferation. Two days after transfection, the cells were observed by fluorescence microscopy. A BeyoClick™ EdU Cell Proliferation Kit (Beyotime) was used according to the manufacturer’s protocol. EdU-positive cells were identified and counted using ImageJ software 1.4.

Clonal expansion assay

BMSCs of passage 4 were seeded at a density of 0.4 × 104 cells per well in 12-well plates. Cultures were continued for 10 days with ASOs, siRNAs or plasmids transfection. The cells were then fixed with 4% paraformaldehyde for 10 min and stained with 0.2% crystal violet (Beyotime) for 1 h at room temperature. Clonal expansion was scored using ImageJ 1.4 software.

CUT&tag assay

A NovoNGS® CUT&Tag 2.0 High-Sensitivity Kit (for Illumina®) (Novoprotein Scientific, Inc., Cat# N259-YH01-01 A) was used to perform the CUT&Tag assay. Briefly, cells were harvested and enriched with ConA magnetic beads. A total of 50 000 cells were resuspended and washed twice with 100 µL of Dig-Wash Buffer. The samples were incubated with anti-LMO7 primary antibody (1:100, 4 °C, 18 h) and secondary antibody (1:200, 25 °C, 1 h). After incubation, the beads were washed three times in Dig-Hisalt Buffer. The cells were incubated with protein A-Tn5 transposome at 25 °C for 1 h and washed three times in Dig-Hisalt buffer. The cells were resuspended in 50 µL of tagmentation buffer and incubated at 37 °C for 1 h, and then the reaction was terminated with 1 µL of 10% SDS at 55 °C for 10 min. Phenol chloroform was used to extract the DNA fragments [82].

CUT&Tag sequencing and analysis

The libraries were sequenced on an Illumina NovaSeq 6000 platform at Novogene Science and Technology Co., Ltd. (Beijing, China), which generated PE150 sequencing data. TrimGalore (v0.6.6) was used to filter the sequencing adaptors and low-quality reads with the following parameters: -q 20 --phred33 --stringency 3. Bowtie2 (v2.5.1)52 was used to map clean reads to the genome with the default parameters. MACS2 (v.2.1.1)55 was used to call peaks with the following parameters: -q 0.05 --call-summits --nomodel–shift − 100 --extsize 200 --keep-dup all. The computeMatrix and plotHeatmap commands in deepTools (v2.3.6.0)56 were used to plot heatmaps of the CUT&Tag data [82].

Animal models

Male C57BL/6J mice were obtained from the Animal Experimental Center of Sun Yat-sen University. 24-month-old (aged) mice and 6-month-old (young) mice were prepared and kept in the experimental environment for at least 1 week prior to additional treatment. The mice were housed under controlled environmental conditions (23 ± 1 °C; 50–60% humidity) with a 12: 12 h light: dark cycle and free access to food and water.

ASOs in vivo treatment experiment

For the ASOs in vivo treatment experiment, we first conducted a preliminary experiment. Aged mice were randomized into 4 groups (3 mice in each group). ASOs (0, 5, 10, and 15 nM) in 100 µl of saline were delivered through tail vein injection at weeks 0, 1, 2, 4, and 6 for a total of 5 injections. Femur samples were harvested at week 8. The group of 15 nM of circStx12 ASO use in mice was excluded from the followed analysis, for one mouse died one day after the first injection and two died one day after the second injection, indicating the toxicity of the 15 nM dose in vivo. Tail-intravenous injection of 10 nM of ASOs caused 60-70% down-regulation of circStx12 even after 2 weeks in femur without toxicity in vivo. While 5 nM of ASOs caused only ~ 10% down-regulation of circStx12 in vivo. Then the dose of 10 nM ASO was chosen for the subsequent expanded sample size experiment (n = 7 mice per group). Four groups of mice were assessed in this study: the saline-injected 6-month-old (Young) group, the saline-injected 24-month-old group (Aged), the control ASO-injected 24-month-old group (Aged-NC), and the circSTX12-ASO-injected 24-month-old group (Aged-ASO). Ten and three days before sample harvesting, calcein (10 µg/g) was intraperitoneally injected, and at week 8, the bone tissue, primary organ and BMSCs samples were harvested, and bone static and dynamic histomorphometry were assessed [83]. The circStx12 ASOs sequence was designed and synthesized by Guangzhou Ruibo Biological Co., Ltd.

Bone tissue staining

Mice femurs were fixed in 4% paraformaldehyde for 36 h and decalcified with EDTA-decalcifying fluid (BOSTER, AR1071) for 3 weeks. Afterwards, paraffin embedding was performed, and 5-µm sections were stained with H&E. The sections were dewaxed in xylene, rehydrated in gradient ethanol, and then rinsed in distilled water. For H&E staining, sections were stained in haematoxylin for 8 min. The samples were rinsed with running water and dehydrated in 70% and 90% alcohol for 10 min each. Subsequently, the femur sections were stained with alcohol eosin for 3 min.

For OCN immunohistochemical staining, after rehydration, the slides were immersed in 10 mM citrate buffer (pH 7.5) and microwaved at 750 W for 30 min for antigen retrieval. Then, the samples were treated with 3% H2O2 for 20 min, blocked with 5% normal goat serum for 1 h at room temperature, and stained with rabbit anti-OCN primary antibody (Abcam, ab133612, 1:200) and the fat vacuoles were stained with a Perilipin-1 antibody (Cell Signaling Technology Cat# 9349,1:100) in blocking buffer overnight at 4 °C. Pure donkey anti-rabbit IgG antibody (Jackson ImmunoResearch, USA, Cat# 711-225-152) was used as a secondary antibody (1:200). After being extensively washed with PBS, the slides were stained with standard DAB for chromogenic detection via immunohistochemistry. After staining, all sections were dehydrated with increasing concentrations of ethanol and xylene. Images were taken and analysed using a Leica DMi8 microscope.

For immunofluorescence staining, the BMSCS was seeded seeded in 12-well plates, covered with sterile glass, and cultured overnight. Then, the cells were fixed in 4% paraformaldehyde for 30 min, permeabilized by 0.1% Triton X-100 for 15 min at room temperature, and then blocked by 5% normal goat serum for 30 min. Then the BMSCs or the sections were incubated with primary antibodies against YAP (Affinity Biosciences, DF3182-200, 1:1000) overnight at 4℃, respectively. Then the BMSCs or sections were stained with fluorophore-conjugated secondary antibodies and mounted with DAPI in anti-fade mounting medium (Thermo Fisher Scientific, Cat# S36963). Images were taken and analysed using a Leica DMi8 microscope.

For Toluidine blue staining, soak the slices in 0.5% toluidine blue staining solution for 30 min. Then rinse with water slightly to remove excess dye. Subsequently, differentiation is carried out using glacial acetic acid solution, which can be prepared in a ratio of 0.5-1:100 with distilled water. The differentiation time depends on the staining effect, usually ranging from a few seconds to a few minutes, until the nuclei and particles are clear. After slightly washing with water, dry the slices with cold air. Finally, use xylene for transparency treatment and seal with neutral gum. Images were taken and analysed using a Leica DMi8 microscope. The number of Osteoblasts (N.Obs) were evaluated on Osteomeasure Software (OsteoMetrics, Decatur, USA).

For tartrate-resistant acid phosphatase (TRAP) staining, after the sections were dewaxed to water, TRAP staining solution was added and incubated at 37 ℃ for 30–60 min. Rinse with distilled water, lightly stain the nucleus with hematoxylin for 3–5 min, and then rinse with distilled water to return to blue. The tissue sections were sealed with a neutral gum after dehydration and transparency. Images were taken and analysed using a Leica DMi8 microscope. The number of osteoclasts (N.Ocs) were evaluated on Osteomeasure Software (OsteoMetrics, Decatur, USA).

Microcomputed tomography (micro-CT) and histomorphometric analyses

Mice femurs were dissected, fixed in 4% paraformaldehyde for 36 h, and stored in 70% ethanol. The samples were scanned using a Siemens mCT40 scanner (Siemens, Germany) at 55 kV/70 mA and a Siemens 1275 mCT imaging system (Siemens, Germany) following the guidelines for the assessment of bone microstructure in rodents. For trabecular bone analysis, the regions of interest were defined as the areas 0.5 mm proximal to the growth plate in the distal femurs. Two- and three-dimensional bone structure images were reconstructed. To examine the rate of bone formation, mice were injected intraperitoneally with 10 mg/kg calcein (Sigma‒Aldrich, Cat# T3383) dissolved in PBS on day 10 and day 3 before harvesting the samples. Femurs were collected and fixed in 4% paraformaldehyde for 36 h, and undecalcified bone slices (8 mm) were obtained. The mineralized surface/bone surface area (MS/BS), mineral apposition rate (MAR) and bone formation rate (BFR) were measured as previously described [84].

Statistical analysis

Each experiment was repeated three times. All the data are expressed as the mean ± SEM. Statistical analysis was performed using GraphPad Prism 7.0. Student’s t test was used to test the difference between two groups. One-way analysis of variance (ANOVA) was used to test differences between three or more groups. Receiver operating characteristic (ROC) analysis was performed to evaluate the diagnostic (disease prediction) ability of gene expression. Differences were considered statistically significant when p < 0.05.