Human BMSCs isolation and culture
The human BMSCs used in the present study were isolated from bone marrow tissues collected from healthy volunteers as previously described [23, 24, 25]. Briefly, 8 ml of bone marrow tissue from each volunteer were first washed with 12 ml of pre-chilled phosphate-buffered saline (PBS) and separated by density gradient centrifugation (500 g; 20 min) using Lymphoprep medium (Axis-Shield, Oslo, Norway). The interface liquid layer containing the mononuclear cells was carefully collected, washed with PBS, and resuspended in low-glucose Dulbecco’s modified Eagle’s medium (DMEM; Thermo Fisher Scientific) containing 10% fetal bovine serum (FBS; Gibco). After culturing at 37°C in a humidified chamber supplied with 5% CO2 for 48 h, non-adherent cells were removed by changing the culture medium. The remaining cells were cultured at 37°C under normal cell culture conditions with the renewal of DMEM medium every 3 days and passaged when cell confluence reached over 80%. After three successive passages, the isolated BMSCs were used for various assays. The research was approved in advance by the Ethical Committee of Sun Yat-sen University. Written informed consent was obtained from each volunteer.
Osteogenic differentiation induction and treatment
Human BMSCs were cultured in six-well plates at 37°C until cell confluency reached approximately 85%. The cells were then subjected to induction of osteogenic differentiation by culturing in adult BMSCs osteogenic differentiation medium (#HUXMA-90021; Cyagen Biosciences, Guangzhou, China) according to the manufacturer’s instructions. The culture medium was renewed every 3 days. The human BMSCs were then incubated with 100 μmol/L MEL (#M5250-1G; Sigma Aldrich), which was added to the osteogenic differentiation medium. The 293T human embryonic kidney cell line (#SCSP-502) was purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China) and cultured at 37°C in DMEM containing 10% FBS and supplied with 5% CO2.
Cell transfections and infection
The sequences for the siRNA targeting circ_0003865 were sense: 5′-AGUUACACAGAUUCAGAUCCAUU-3′, antisense: 5′-AAUGGAUCUGAAUCUGUGUAACU-3′; its negative control, sense: 5′-UUCUCCGAACGUGUCACGUTT-3′, antisense: 5′-ACGUGACACGUUCGGAGAATT-3′; the miR-3653-3p mimics, sense: 5′-CUAAGAAGUUGACUGAAG-3′, antisense: 5′-UCAGUCAACUUCUUAGUU-3′; and mimics NC(the same as negative control); the miR-3653 inhibitors : 5′-CUUCAGUCAACUUCUAG-3′ and inhibitor NC : 5′-CAGUACUUUUGUGUAGUACAA-3′. All oligos were synthesized by the GenePharma Biotech company (Shanghai, China). As designated, the above sequences were transfected into human BMSCs using the Lipofectamine TM 2000 transfection kit (Invitrogen, USA) according to the manufacturer’s instructions. For overexpression of circ_0003865, the sequences of circ_0003865 were synthesized and ligated into the LV5 plasmid to construct the recombinant LV5-circ_0003865 vector. LV5 and LV5-circ_0003865 lentiviruses were packaged by the Vigene Biosciences company (Jinan, Shandong, China) and infected into BMSCs according to the manufacturer’s instructions.
qRT-PCR and RT-PCR
Total RNA samples from cultured BMSCs or mouse tissues were prepared using the Total RNA Extraction Kit (#17200; AmyJet Scientific, Wuhan, China) according to the manufacturer’s instructions. RNA concentrations were determined by the Nanodrop 2000 instrument (Thermo Fisher Scientific). The cDNA for qRT-PCR was subsequently synthesized from 3 µg total RNA using the Omniscript RT Kit (#205111; Qiagen) as instructed by the manufacturer. Subsequently, the real-time quantitative PCR assay was then conducted using the TransStart® Green qPCR SuperMix kit (#AQ101-01; TransGen Biotech, Beijing, China) according to the manufacturer’s instructions. The final expressional levels of circRNAs, microRNAs, or mRNAs were calculated by the standard 2−△△Ct method based on at least three biological replicates. The covalently closed-loop structures of circRNAs were measured by PCR using divergent primers, whereas opposite-directed primers targeting linear genomic DNA were used as controls. The sequences of all primers used for quantitation are listed in Table 1.
Western blotting
Total protein samples were extracted from cultured BMSCs using the Total Protein Extraction kit (#C006225-0050; Sangon Biotech, Shanghai, China) according to the manufacturer’s instructions. The protein concentrations were determined by the bicinchoninic acid method. Approximately 30 µg of protein was then boiled at 100°C for 5 min, separated by 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis, and blotted onto polyvinylidene fluoride (PVDF) membranes The membranes were subsequently blocked in 5% lipid-free milk solution, incubated with diluted primary antibodies, washed three times with TBST solution, and incubated with diluted secondary antibodies. Finally, the PVDF membranes were developed with the EasyBlot ECL (enhanced chemiluminescence) kit (#D601039-0050; Sangon Biotech, Shanghai, China) according to the manufacturer’s instructions. At least three biological replicates were conducted for protein quantitation, and GAPDH was used as an internal standard. The antibodies used in the present study included anti-ALP (#ab229126; Abcam), anti-osteopontin (OPN) (#ab8448; Abcam), anti-runt-related transcription factor 2 (RUNX2) (#12556; CST), anti-GAPDH (#ab181602; Abcam), and anti-growth arrest-specific gene 1 (GAS1) (#ab236618; Abcam).
CircRNA and mRNA profiling
The differential expression of circRNAs and mRNA profiles for human BMSCs between, BMSCs osteogenic differentiation induction group and BMSCs osteogenic differentiation induction with 100 µM MEL treatment at the same time group were determined by the deep RNA sequencing method. Briefly, total RNA samples from each group were prepared using TRIzol reagent (Thermo Fisher Scientific) according to the manufacturer’s instructions, and the concentration of the RNA was measured using the Nanodrop 2000 instrument. RNA integrity was evaluated using the Agilent 2100 Bioanalyzer instrument. Then, the rRNA components were removed from the RNA samples using the RiboMinus™ Eukaryote Kit for RNA-Seq (#A1083708; Thermo Fisher Scientific) according to the manufacturer’s instructions. For circRNA sequencing, the digestion of linear RNAs was done using RNase R. For mRNA sequencing, the eukaryotic mRNA samples were enriched using Oligo-dT beads (Thermo Fisher Scientific) according to the manufacturer’s instructions. An RNA library was generated from the collected RNA samples using the NEBNext Ultra II RNA Library Prep Kit (#E7770S; NEB) according to the manufacturer’s instructions. Subsequently, the RNA library was sequenced using the HiSeq 2000 sequencing system (Illumina, USA).
Bioinformatics
The clean reads obtained from the RNA sequencing were then aligned with the human reference genome database using Bowtie2 software. The back-splice algorithm was subsequently conducted to select for read junctions. The CIRI software was used to predict and annotate the circRNAs from all mapped reads. The differential expression levels of circRNAs in human BMSCs induced by MEL treatment after osteogenic differentiation induction were determined by calculating the RPKM (mapped back-splicing junction reads per kilobase per million mapped reads) value, which was calibrated to the total read numbers. The significantly differential expression of circRNAs or mRNAs in MEL-treated BMSCs compared with BMSCs was determined using a log2Ratio of >1 combined with a false discovery rate of <0.05. The differential expression of circRNAs and mRNAs were further analyzed by hierarchical clustering and scatter plots, which were established using the R software package (Version 0.2.3). Subsequently, the functional categorization of differentially expressed mRNAs based on Gene Ontology (GO) terms was done by searching the Database for Annotation, Visualization, and Integrated Discovery. The enrichment of differentially expressed mRNAs in signaling pathways was done by searching the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (http://www.genome.jp/kegg/). The interaction between circRNA, microRNA, and mRNAs was predicted using String V10 software.
Cell Counting Kit-8 (CCK-8) assay
After treatment, the proliferation of BMSCs were detected with the CCK-8 kit (#C0037, Beyotime Biotech, Shanghai, China) following the manufacturer’s instructions. Briefly, BMSCs were digested by trypsin (Thermo Fisher Scientific) and then were seeded into 96-well plate at a density of 4×103 cells pre well. After 24 h, 48 h, and 72 h, Cell Counting Kit-8 (CCK-8) was added to detect the proliferation of BMSCs via Elx800 (BioTek, Winooski, Vermont, USA) at 450 nm spectrophotometric.
Alkaline phosphatase and Alizarin Red staining
The expression of alkaline phosphatase (ALP) in human BMSCs was detected using the BCIP/NBT Alkaline Phosphatase Color Development Kit (#C3206; Beyotime Biotech) according to the manufacturer’s instructions. Briefly, cell slides were washed three times with PBS solution, fixed in 4% paraformaldehyde, and washed again three times with PBS solution. The slides were then incubated with BCIP/NBT staining solution for 1–2 h at room temperature in darkness, washed twice with distilled water for 3 min. Finally, ALP expression in human BMSCs was observed by gross scanning (Bio-rad, GS-800, Hercules, CA, USA) and light microscopy (OPTEC, TP510, Chongqing, China). At least three biological replicates were conducted for comparison. To evaluate osteogenic differentiation, BMSCs were also stained with Alizarin Red S reagent (#ST1078-25g; Beyotime Biotech) for 1–5 min at room temperature, as outlined in the manufacturer’s instructions. Quantification of ALP and Alizarin Red S were conducted using Image J software (National Institutes of Health, Bethesda, MD, USA)
RNA pull-down assay
An RNA pull-down assay was conducted to validate the interaction between circ_0003865 and miR-3653-3p using the biotin-LNA-3865 probe (5′-TGGATCTGAATCTGTGTAACT-3′) synthesized by the GENEray company (Shanghai, China) as previously described [26]. Briefly, cultured 293T cells overexpressing circ_0003865 were lysed and incubated with the above-mentioned probes for 2 h and “pulled down” with streptavidin-coated magnetic beads (#08014; Sigma Aldrich). After washing three times with PBS, the eluates were analyzed by quantitative PCR method to detect both circ_0003865 and miR-3653-3p. At least three biological replicates were conducted.
Dual-luciferase reporter assay
The interactions between miR-3653-3p and circ_0003865 or the 3′ UTR region of the GAS1 gene were verified using the Dual-Luciferase Reporter Assay System (Promega) according to the manufacturer’s instructions. Briefly, the circ_0003865 sequences and its mutant version, as well as the GAS1 3′ UTR sequences and its mutant version, were amplified by RT-PCR and were ligated into PmirGLO plasmids. The plasmids were transfected into 293T cells using the Lipofectamine TM 2000 transfection kit (Invitrogen, USA) according to the manufacturer’s instructions, along with miR-3653-3p mimics or a negative control. Subsequently, the cells were lysed using the Passive Lysis Buffer and analyzed using the GloMax-20/20 luminometer (#E5311; Promega). At least three biological replicates were conducted for comparison of luciferase activities between groups.
Mouse osteoporosis model and evaluation
The mouse osteoporosis model was established by ovariectomy surgery as previously described [27] and approved by the Experimental Animal Care and Usage Committee of the Sun Yat-sen University (Guangdong, China). The LV5-sh_NC and LV5-sh-hsa_circ_0003865 lentiviruses were packaged and prepared by the Vigene Biosciences company (Jinan, Shandong, China). Thirty female SPF-grade C57BL/6J mice between 6–8 weeks old were purchased from Vital River Laboratories (Beijing, China) and randomly divided into an sh-NC group (n = 10), an OP + sh-NC group (n = 10) and an OP + sh-hsa_circ_0003865 group (n = 10). The sh-NC group was injected with 10 µl empty vector at the distal femur (1013 copies/mouse) before the sham operation. Mice in the OP + sh-NC group were injected with 10 µl empty vector (1013 copies/mouse) in the distal femur, followed by ovariectomy surgery. Mice in the OP + sh-hsa_circ_0003865 group were injected with 10 µl of sh-hsa_circ_0003865 (1013 copies/mouse) in the distal femur and then subjected to ovariectomy. Bone microstructure analysis by micro-CT examination was conducted as previously described [16]. Quantitative RT-PCR was used to detect the relative expression of circ-0003865, ALP, RUNX2, OPN, GAS1, and miR-3865-3P in bone tissues. The relative abundance of RFP (abcam, Cambridge, MA, USA), ALP, RUNX2 OPN, and GAS1 proteins in bone tissues was measured by immunofluorescence using the Immunol Fluorescence Staining Kit (#P0179; Beyotime Biotech) according to the manufacturer’s instructions.
Statistical analysis
All quantitative data obtained from at least three biological replicates are presented as the mean ± standard deviation in the present study and analyzed using SPSS 20.0 software for evaluating statistical significances. The differences between 2 and >2 groups were determined by the Student’s T-test and analysis of variance methods, respectively. Significant differences between groups were defined by a P value of <0.05.