Mouse studies
All animal experiments were approved by the Animal Care and Use Committee of the First Affiliated Hospital of USTC (ethical approval no. 2022-N(A)-078). All the procedures followed the guidelines set forth by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996) on the protection of animals used for scientific purposes. Wild-type C57/BL6 mice were purchased from the Shanghai Slake Experimental Animal Company (Shanghai, China). All mouse strains were on a pure C57BL/6 genetic background and were housed in the animal facility of the Fist Affiliated Hospital of USTC in a pathogen-free, temperature-controlled environment under a 12:12 h light–dark cycle.
Bone marrow macrophage (BMM) isolation
BMMs were isolated from 6-week-old mice as previously described18. Briefly, bone marrow cells were flushed from the femur and tibia and then resuspended in red blood lysis buffer (Beyotime, C3702-120 ml) for 2 min to remove red blood cells. The cells were cultured in complete α-MEM (Gibco, 22561-021) with 30 ng/ml M-CSF (R&D Systems, 416-ML-050) in a suspension culture dish (Corning, 430591) at 37 °C for 3 days. After that, nonadherent cells were removed by washing, and the attached BMMs were released using 0.25% trypsin-EDTA (Thermo Fisher, 25200056) for the indicated experiments.
Osteoclast differentiation
For induction of osteoclast differentiation, BMMs were seeded in plates at a density of 25,000 cells/cm2. The cells were cultured in complete α-MEM with 30 ng/ml M-CSF and 10 ng/ml RANKL (R&D Systems, 462-TEC-010) for 5 days to obtain mature osteoclasts as previously described18,19. For circular RNA sequencing, BMMs were treated with RANKL and collected at the indicated time points corresponding to different osteoclast differentiation stages: D0, undifferentiated (without RANKL treatment); D1, early-stage (with RANKL treatment for 1 day); D3, middle-stage (with RANKL treatment for 3 days); and D5, late-stage (with RANKL treatment for 5 days).
Circular RNA sequencing
For circular RNA sequencing, total RNA was extracted from BMMs treated with RANKL for the indicated time points (D0, D1, D3, D5) using TRIzol reagent (Invitrogen Thermo Fisher, 15596026) following the manufacturer’s instructions. The quantity and quality of the RNA were evaluated using a NanoDrop spectrophotometer and an Agilent 2100 bioanalyzer. Subsequently, DNase I was used to degrade double-stranded and single-stranded DNA. The Ribo-of rRNA Depletion kit (Vazyme Biotech Co., N406) was used to remove ribosomal RNA, and RNase R (Epicentre Illumina) was utilized to eliminate linear RNA. Purification was carried out using Agencourt RNAClean XP magnetic beads (Invitrogen; Thermo Fisher Scientific, Inc.). The libraries were sequenced on the BGISEQ-500 (BGI Group). Raw reads were filtered to remove those with low quality, linker contamination and excessively high levels of unknown base N using SOAPnuke software v1.5.2 (https://github.com/BGI-flexlab/SOAPnuke). The remaining clean reads were aligned to the reference genome (Mus_musculus, UCSC_mm9; ftp://hgdownload.soe.ucsc.edu/goldenPath/mm9/). Circular RNAs were detected and identified using CIRI20 and Find_circ6,21. The results from those two software tools were integrated based on the circRNA start and stop positions (combining circRNAs with start and stop positions within the first and last 10 bases into one class). The expression of circRNAs was calculated based on the number of junction reads aligned to the two ends of the circRNA. As both CIRI (BWA‑MEM genome alignment algorithm) and Find_circ (Bowtie2 genome alignment algorithm) software were used in the profiling, the final numbers of junction reads were the mean of the two results. CircRNAs exhibiting fold changes ≥2.0 with P values ≤0.05 were classified as significantly differentially expressed circRNAs. The differentially expressed genes (DEGs) obtained from all groups were analyzed by bidirectional clustering using the Pheatmap package (v1.0.12; rdocumentation.org/packages/pheatmap/versions/1.0.12) and are presented as heatmaps.
Transcriptome analysis
For transcriptome RNA-seq analysis, BMMs transfected with ASO-ctrl or ASO-1 were cultured in the presence of M-CSF and RANKL for 5 days. Afterward, cells were collected, and total RNA was isolated using TRIzol reagent. RNA sequencing was performed by Novogene Company (Beijing, China). Subsequent data analysis was conducted by Rstudio and GSEA software.
RNA preparation, RNase R treatment, qRT‒PCR and actinomycin D treatment
Total RNA was isolated from cells or tissues using TRIzol reagent according to the manufacturer’s instructions. For verification of the backspliced junction point of circRNAs, rRNA and linear RNA were removed using the rRNA Depletion kit (Vazyme Biotech Co., N406) and RNase R (Epicentre Technologies, Madison, WI). cDNA was synthesized using a random primer (TaKaRa, Dalian, China). For quantification of the mRNA and circRNA levels, cDNA was synthesized using PrimeScript RT Master Mix (TaKaRa, Dalian, China). Real-time PCR analyses were performed using SYBR Premix Ex Taq II (TaKaRa). Notably, divergent primers annealing at the distal ends of circRNA were used to determine the abundance of circRNA. Amplification was performed using the StepOnePlus Real-Time PCR System (Applied Biosystems, Foster City, CA), and Ct thresholds were determined using the software. All the primers used in this study are listed in Supplementary Table 1.
For RNase R treatment, 1 μg of total RNA was incubated for 15 min at 37 °C with or without 3 U of RNase R (Epicentre Technologies, Madison, WI). For actinomycin D treatment, cells were treated with 2 μg/ml actinomycin D or DMSO and collected after 24 h.
Tartrate-resistant acid phosphatase (TRAP) staining
BMMs with the indicated treatment were seeded in a 96-well plate and cultured in the presence of M-CSF and RANKL for 5 days. Afterward, TRAP staining was performed using a TRAP staining kit (Sigma‒Aldrich, 387A) following the manufacturer’s protocol. TRAP-positive cells with two or more nuclei were considered osteoclasts and were counted.
Bone resorption pit assay
BMMs with the indicated treatment were cultured in the presence of M-CSF and RANKL for 5 days to allow the formation of mature osteoclasts. Then, the cells were dissociated using cell dissociation buffer (Gibco, 13151014), reseeded on dentin slices in a 96-well plate, and cultured together with M-CSF and RANKL for another 2 days. Then, the dentin slice was collected, and pit lesion staining was performed. The brown area representing the bone resorption area was measured using ImageJ software (http://imagej.nih.gov/ij/).
Antisense oligonucleotides, siRNAs and lentivirus
For knockdown of target circular RNAs (circFam190a, circAtrnl1 and circMkln1), locked nucleic acid enhanced antisense oligonucleotides (LNA-ASOs) were custom-designed and purchased from Qiagen (Hilden, Germany). Three different sequences were designed for each target circRNA, and a negative control sequence was included. The sequence information is listed in Supplementary Table 1. The knockdown efficiency was validated via RT‒qPCR.
For the knockdown of HSP90β and FUS, siRNA was designed and purchased from General Biol Company (Chuzhou, China). The sequence is shown in Supplementary Table 1.
For the overexpression of circFam190a, Fus and AKT1, pCDH (acting as a control), pCDH-Fus, pCDH-circFam190a, and pCDH-AKT1 plasmids were constructed and purchased from General Biol Company (Chuzhou, China). Lentiviral particles were produced by cotransfecting HEK-293T cells with these plasmids and helper plasmids (psPAX2 and pMD2.G) using Lipofectamine 3000 reagent (Thermo Fisher Scientific, Waltham, USA). The harvested supernatant was filtered through a 0.45-μm filter and used as a rich viral source for later cell transduction.
ASO treatment in vitro and in vivo
For in vitro cell transfections, 2 × 105 cells (BMMs or BMSCs) were seeded in a 6-well plate and transfected by using Lipofectamine 3000 (Thermo Fisher Scientific, Waltham, USA) following the manufacturer’s protocol. A mix of the indicated ASOs was used at a final concentration of 30 nM, and the cells were incubated for 24 h.
For in vivo animal experiments, eight-week-old female mice underwent either OVX or sham surgery. One week later, ASO-ctrl (negative control) or ASO-1 (targeting circFam190a) was injected into the lateral tail vein of the mice at a dose of 50 mg/kg twice a week as previously described22. Five weeks after the first injection, the mice were harvested. BMMs were collected for the validation of the knockdown efficiency. The left femur was sent for micro-CT analysis, the right femur was sent for bone histomorphometric analysis, and serum samples were collected for CTX1 and P1NP analysis.
AAV and MK-2206 treatment
In vivo overexpression of circFam190a was achieved using AAV9 vectors. Murine circFam190a-overexpression AAV9 (AAV-circFam190a) and the corresponding control AAV9 (AAV-ctrl) were manufactured by GeneChem (Shanghai, China). Eight-week-old mice received a tail vein injection of 1 × 1011 viral particles of AAV-circFam190a or AAV-ctrl. After one week, the mice were intravenously treated with vehicle (PBS) or MK-2206 (120 mg/kg) three times per week for five weeks. The mice were then harvested, and BMMs, serum and femurs were collected for further analysis.
Human sample collection
All protocols involving human samples were approved by the Research Ethics Committees of the First Affiliated Hospital of USTC (ethical approval no. 2000ky19). The procedures complied with the principles of the Declaration of Helsinki. Fifty postmenopausal women were randomly selected from the physical examination center at the First Affiliated Hospital of USTC. The inclusion criteria were postmenopausal women aged between 50 and 65 years with informed consent. Exclusion criteria included a history of alcohol abuse, severe renal or hepatic diseases, hyperthyroidism, chronic use of corticosteroids, severe gastrointestinal diseases, malignant tumors, diabetes mellitus, severe cardiac diseases, rheumatoid arthritis, other rheumatic-inflammatory diseases, and treatment for osteoporosis or osteopenia such as selective estrogen receptor modulators, bisphosphonate, and calcitonin.
In our study, all subjects were divided into three groups based on different BMD manifested by T score23. The “osteoporosis” group refers arbitrarily to those with T values below −2.5, the “osteopenia” group has values between −1.0 and −2.5, and the “normal” group has values above −1.0. Blood samples from each patient were taken in the morning (after an overnight fast) between 8 a.m. and noon using precooled tubes. Peripheral blood mononuclear cells (PBMCs) were isolated freshly by standard Ficoll (GE Healthcare, Little Chalfont, United Kingdom) density gradient centrifugation. Total RNA from PBMCs was isolated and stored at −80 °C for further analysis of circFAM190A expression. Detailed information about the patients is presented in Supplementary Table 2.
Micro-CT analysis
For micro-CT analysis, a high-resolution μCT35 desktop microtomographic imaging system (Scanco Medical AG, Brüttisellen, Switzerland) was used. The left femurs from mice were collected for μCT scanning. Scans were acquired with a 12 µm3 isotropic voxel size, 55 kVP, 145 μA and a 600 ms integration time. The images were subjected to Gaussian filtration and segmented using a fixed threshold of 700 mgHA/cm3 as previously described18. The image acquisition and analysis protocols followed the JBMR guidelines24. The region of interest (ROI) was set as 100 μCT slices 1200 μm under the growth plate. The following variables were computed: bone mineral density (BMD), bone volume/total volume (BV/TV), trabecular bone number (Tb.N), trabecular bone thickness (Tb.Th), trabecular bone separation (Tb.Sp), and connectivity density (Conn.D). μCT images were reconstructed using built-in software.
Bone histomorphometric analysis
For bone histomorphometric analysis, mice were injected with calcein and Alizarin red 4 days and 2 days prior to sacrifice, respectively. The right femurs were collected, fixed in 70% ethanol, and embedded as previously described25,26. Then, 5 μm standard undecalcified sections were cut using a microtome (Leica RM2255). von Kossa staining, toluidine blue staining and TRAP staining were performed. Quantitative bone histomorphometric measurements were performed using the OsteoMeasure system. The following variables were analyzed: bone volume/total volume (BV/TV), trabecular bone thickness (Tb.Th), trabecular bone space (Tb.Sp), trabecular bone number (Tb.N), osteoclast surface/bone surface (Oc.S/BS), osteoclast number/bone surface (Oc.N/BS), mineral apposition rate (MAR), and bone formation rate/bone surface (BFR/BS).
CTX1 and P1NP assay
Mouse serum was collected after a 6-hour fasting period. The serum levels of the bone resorption marker collagen C-terminal telopeptide (CTX-1) and bone formation marker procollagen 1 N-terminal peptide (P1NP) were measured using ELISA kits (AC-06F1 and AC-33F1, Immunodiagnostics Systems, Maryland, USA) following the manufacturer’s instructions.
Western blot (WB) and immunoprecipitation (IP)
For WB analysis, whole cell lysates were collected using RIPA buffer containing protease inhibitors (P1008, Beyotime, Shanghai, China). After boiling, the supernatants were subjected to SDS‒PAGE and transferred to membranes. The bands were visualized using enhanced chemiluminescence (E412–01, Vazyme, Nanjing, China). For IP analysis, cells were lysed in co-IP buffer with a protease inhibitor cocktail for 40 minutes on ice. Cell lysates were incubated with the indicated antibodies adsorbed to protein A/G Agarose (20421, Thermo Fisher Scientific, Waltham, USA) for 4 h at 4 °C, followed by three washes in co-IP buffer and elution at 95 °C for 10 min. The antibodies used are listed in Supplementary Table 3.
RNA immunoprecipitation (RIP)
RIP experiments were performed using the Magna RIP RNA-Binding Protein Immunoprecipitation Kit (17-700, Millipore, Bedford, USA). Approximately 107 cells were collected with RIP lysis buffer (approximately 100 μl) containing protease and RNase inhibitors. The cell lysates were then incubated with 5 μg of AGO2, the indicated antibodies, or control rabbit IgG (31235, Thermo Fisher Scientific, Waltham, USA)-coated beads with rotation at 4 °C overnight. After treatment with proteinase K, the immunoprecipitated RNAs were extracted using the RNeasy MinElute Cleanup Kit (Qiagen) and reverse transcribed using PrimeScript RT Master Mix (TaKaRa).
Fluorescence in situ hybridization (FISH)
Cy3-labeled circFam190a probes (Supplementary Table 1) were synthesized by TSINGKE (Wuhan, China). The circFam190a FISH was performed as described previously27. Briefly, cells were fixed with the fixative solution, followed by permeabilization. Hybridization was performed at 37 °C overnight in a dark moist chamber. After being washed three times in 2×SSC (Solarbio, Beijing, China) for 10 min, the coverslips were sealed with parafilm containing DAPI. The images were acquired using a confocal laser scanning microscope (LSM 780, Carl Zeiss).
RNA pulldown assays and spectrometry analysis
Biotin-labeled circFam190a (antisense) and control (sense) probes (Supplementary Table 1) were synthesized by TSINGKE (Wuhan, China). RNA pulldown assays were performed as previously described. Briefly, 107 cells were washed in ice-cold phosphate-buffered saline, lysed in 500 μl of co-IP buffer, and incubated with 3 μg biotinylated DNA oligo probes at room temperature for 2 h. A total of 50 μl of washed streptavidin C1 magnetic beads (Invitrogen) was added to each binding reaction and further incubated at room temperature for another hour. The beads were washed briefly with co-IP buffer five times. The bound proteins in the pulldown materials were analyzed by mass spectrometry or Western blotting. Mass spectrometry analysis was performed by Oebiotech (Shanghai, China).
Two-step RNA-binding protein immunoprecipitation
Two-step IP analysis was performed as previously described28. Briefly, cell lysates were collected using lysis buffer (20 mM HEPES (pH 7.8), 400 mM KCl, 5% glycerol, 5 mM EDTA, 1% NP40, protease inhibitor cocktail and RNase inhibitor). The lysates were first immunoprecipitated with anti-Flag antibodies before elution with Flag peptides. Ten percent of the sample was collected for WB and RT‒PCR analysis, and the rest of the eluate was further incubated with control IgG or anti-HA antibodies for second-phase immunoprecipitation.
Luciferase reporter and mammalian two-hybrid assays
Mammalian two-hybrid assays were performed as previously described28. Assays were performed according to the manufacturer’s instructions (E2440, Promega, Madison, USA) after transfection of the indicated plasmids. For the mammalian two-hybrid assay, complementary DNAs for HSP90β and AKT1 were cloned into the pBIND and pACT vectors and transfected along with firefly (pG5luc) luciferase vectors. Reporter activities were measured 48 h later for the luciferase reporter and mammalian two-hybrid assays using the Dual-Luciferase Reporter Assay Kit (Promega). Renilla measurements were used to normalize changes in firefly luciferase activity.
Sequencing data accession
The raw sequence data reported in this paper have been deposited in the Genome Sequence Archive (Genomics, Proteomics & Bioinformatics 2021) in the National Genomics Data Center (Nucleic Acids Res 2022), China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA: CRA009230) and are publicly accessible at https://ngdc.cncb.ac.cn/gsa29,30.
Statistical analysis
Data are expressed as the mean ± SEM. Statistical analysis was conducted using unpaired two-tailed Student’s t test. For comparison of three or more groups, two-way ANOVA followed by Tukey’s multiple comparisons test for all groups was used. GraphPad Prism 9 was used for statistical analysis. A P value <0.05 was considered statistically significant.