MC3T3-E1 (American Type Culture Collection) and MDPC23 (Cellcook) cells were plated on 6-well plate culture dishes at density of 1×10⁵ cells/well. In addition, 2 ml Dulbecco's modified Eagle's medium (Gibco; Thermo Fisher Scientific, Inc.) containing 10% fetal bovine serum (Gibco; Thermo Fisher Scientific, Inc.) was added to each well. Cells were maintained in a humidified atmosphere of 95% air and 5% CO₂ at 37°C.

The siRNAs, mimics and inhibitors were designed and synthesized by Guangzhou RiboBio Co., Ltd. (siRNA: Sense: 5′-GCUAGAACAGCAUGGUCCAdTdT-3′, antisense: 3′-dTdTCGAUCUUGUCGUACCAGGU-5′; inhibitors: 5′-UCCUCCCUCCCCUACCCGGUUCAAG-3′; mimic sense: 5′-AGGAGGGAGGGGAUGGGCCAAGUUC-3′, antisense: 5′-UCCUCCCUCCCCUACCCGGUUCAAG-3′) (Table I). MC3T3-E1 and MDPC23 cells were seeded in 6-well plates at a density of 1×10⁵/well. The cells were transfected with 20 nM siRNA mimics or inhibitors using GenMute™ siRNA Transfection kit and Transfection Buffer (SignaGen Laboratories) according to the manufacturer's protocol. To ensure a high silencing efficiency, the transfection medium with siRNAs was changed every 72 h. Experiments were performed 24–72 h post transfection.

The MDPC23 cells were divided into two groups and treated as before. In the subject group, 1×10⁵ cells were treated with the induction medium containing 20 nM siRNA of mmu_circ_003795 for 7 days. In the control groups, the cells (1×10⁵) were treated with the induction medium containing 20 nM negative control siRNA for 7 days. The sequences of siRNA were as follows: (Sense 5′-GCUAGAACAGCAUGGUCCAdTdT-3′ and antisense 3′-dTdTCGAUCUUGUCGUACCAGGU-5′). These two groups were used to detect the expression of mRNAs and miRNAs following silencing of mmu_circRNA_003795. Total RNA from each sample was quantified using the NanoDrop ND-2000 (Thermo Fisher Scientific, Inc.). The sample preparation and microarray hybridization were performed based on the standard protocols of the BGI group (10).

Cells were seeded into 48-well plates at a density of 5×10³ cells/well. The MDPC23 and MC3T3-E1 cells were divided into two groups: The control group and the subject group. In the subject group, the cells (5×10³) were treated with the induction medium containing 20 nm siRNA of mmu_circ_003795 for 14 days. In the control group, the cells (5×10³) were treated with the induction medium containing 20 nM negative control siRNA for 14 days. The medium was changed every 3 days. An Alkaline Phosphatase Staining Kit (BestBio, Inc.) was used to detect the ALP activity, according to the manufacturer's protocol.

Mineralization in MC3T3-E1 and MDPC23 cells was measured by staining with ARS. Cells were seeded in 48-well plates at a density of 5×10³ cells/well. As previously the MDPC23 and MC3T3-E1 cells were divided into two groups. In the subject group, the cells were treated with the induction medium containing 20 nm siRNA of mmu_circ_003795. In the control group, the cells were treated with the induction medium containing 20 nM negative control siRNA. The medium was changed every 3 days. Following culture for 21 days, cells were fixed with 4% paraformaldehyde for 10 min after washing three times with PBS at room temperature. Subsequently, the ARS solution (BestBio, Inc.) was used to stain the cells at 37°C for 30 min. The stained cells were observed by a stereo light microscope (Leica Microsystems GmbH). For quantification analysis, 10% hexadecyl pyridinium chloride monohydrate (CPC; Sigma-Aldrich; Merck KGaA) was used to dissolve the mineralized nodules and then the absorbance was measured at 562 nm using a Multiskan™ FC spectrophotometer (Thermo Fisher Scientific, Inc.).

The total RNA of BMSCs was extracted using TRIzol reagent (Thermo Fisher Scientific, Inc.). The quantity and quality of RNA were determined using a NanoDrop 2000. Following RNA extraction, SuperScript III reverse transcriptase (Invitrogen; Thermo Fisher Scientific, Inc.) was used to synthesize complementary DNA (cDNA), according to the manufacturer's protocol. In the reverse transcription reaction the temperature protocol was as follows: 65°C for 5 min and 4°C for 1 min in the first step, followed by 37°C for 1 min, 50°C for 60 min and 70°C for 15 min in the second step. Subsequently, qPCR was performed using 12.5 µl SYBR Premix EX Taq (Takara Bio, Inc.), 0.5 µl forward primer (10 pmol/l), 0.5 µl reverse primer (10 pmol/l;), 1 µl cDNA and ddH₂O to create a final volume of 25 µl. The primers used are listed in Table I. The reaction conditions were as follows: 95°C for 30 sec; and 95°C for 5 sec, 55°C for 30 sec and 72°C for 30 sec for 40 cycles. The instrument automatically generated the cycle quantification (Cq) value of mRNA and the internal reference gene GAPDH. Cq value difference (ΔCq) indicated the relative expression of each mRNA investigated (11).

Quantile normalization of the original data and subsequent data processing was performed using the R software package (version 3.1.2) (12). After quantile normalization of the original data was performed, low-intensity filtering was carried out and the mRNAs with P (present) or M (marginal) (‘all target values’) tagged in at least one or two samples were retained for further analysis. When comparing the two sets of contours, the ‘fold-change’ (namely, the ratio of the group mean) between groups of mRNAs was calculated. The statistical significance of the differences between groups was determined using a t-test. mRNAs exhibiting a significant difference in expression exhibited a fold-change >2 and P<0.05. The sorting and filtering function of Microsoft Excel 2019 (Microsoft Corporation) was used to filter data, analyze output and sort the mRNAs with differential expression according to the fold-changes and P-values. In addition, gene ontology (GO) analysis was performed to classify the related mRNAs (13). miRDB (http://www.mirdb.org/), TargetScan (http://www.targetscan.org/) and mirbases (http://www.mirbase.org/) were used to determine the corresponding target genes. All differentially expressed miRNAs were annotated in detail with information on the interaction between mmu_circ_003795 and miRNAs. Cytoscape (version 3.6.0; http://cytoscape.org) software was used to generate a network map.

Radioimmunoprecipitation assay buffer [50 mM Tris HCl (pH 7.4), 150 mM NaCl, 1% Nonidet P40 and 0.1% sodium dodecyl sulfate] and phenylmethanesulfonyl fluoride were used at 4°C to extract protein from the cells according to the standard procedure. The concentration was determined using a bicinchoninic acid protein assay kit (BestBio, Inc.). Total protein (10 µg/well) was separated using 10% SDS-PAGE and then transferred to a polyvinylidene difluoride membrane. Skimmed milk was used to block membranes for 1 h at room temperature, then the blots were incubated with primary antibodies against OPN (1:1,000; catalog no. ab124830; Abcam), COL15A1 (1:1,000; catalog no. PA5-69518; Thermo Fisher Scientific, Inc.) and β-actin (1:5,000; catalog no. ab8226; Abcam) overnight at 4°C. Following washing with 0.1% Tween PBS-T, the blots were incubated with goat anti-mouse horseradish peroxidase-conjugated secondary antibody (1:5,000; catalog no. ab6799; Abcam) for 1 h at room temperature. An enhanced chemiluminescent (ECL) kit (Beijing Dingguo Changsheng Biotechnology Co., Ltd.) was used to detect the proteins. ImageJ (version k1.45; National Institutes of Health) was used to measure densitometry.

All data were analyzed using the SPSS 11.5 software package (SPSS, Inc.). The data were expressed as the mean ± standard deviation and the comparison of the data between groups was performed using paired t-test or two-way analysis of variance according to the nature of the data. Each experiment was repeated three times. The data of three groups were evaluated using analysis of variance and followed by Fisher's least significant difference post hoc test when the variance was normal, otherwise Dunnett's T3 test was used. P<0.05 was considered to indicate a statistically significant difference.

MC3T3-E1 and MDPC23 cells were cultured with induction medium for 3 days. RT-qPCR was used to analyze the expression levels of mmu_circRNA_003795 and the osteogenic marker OPN, ALP, Sp7 transcription factor (OSX), osteocalcin (OCN) and runt-related transcription factor 2 (RUNX2). The results demonstrated that mmu_circRNA_003795, OPN, ALP, OSX, OCN and RUNX2 were increased (Fig. 1A). The ALP and ARS staining were performed to verify the successful induction (Fig. 1B and C). An siRNA sequence was designed to silence mmu_circRNA_003795. MDPC23 cells were transfected with negative control RNAs or siRNAs for 7 days. In the miRNA microarray, 1,729 miRNAs were different between the control and subject groups (Fig. 2A). The present study identified that mmu_circRNA_003795 is most likely bound to mmu-miR-1249-5p, mmu-miR-504-3p, mmu-miR-6399, mmu-miR-7054-5p and mmu-miR-667-5p (Fig. 2B). By comparing the aforementioned data, the present study selected mmu-miR-1249-5p as the target miRNA. The seed sequences of the circRNA were compared with miRNA sequences to identify the miRNAs that potentially bind to circRNAs (Fig. 2C).

GO analysis was performed to classify the related mRNAs. The analysis focused on ‘Biological Process’, ‘Cellular Component’ and ‘Molecular Function’ terms that may be associated with osteogenesis differentiation. The top GO terms were ranked by number of genes (Fig. 2D). miRDB (http://www.mirdb.org/), TargetScan (http://www.targetscan.org/) and mirbases (http://www.mirbase.org/) were used to determine the corresponding target genes of miR1249-5p. The above data was compared with the current mRNA microarray data. According to the analysis, miR1249-5p and COL15A1 may have regulated osteoblast differentiation and mineralization in MDPC23 cells.

siRNA was used to knockdown mmu_circRNA_003795 in MC3T3-E1 and MDPC23 cells. Compared with the negative control group, the expression of mmu_circRNA_003795 in the siRNA group was decreased in both MC3T3-E1 and MDPC23 cells (Fig. 3A and B). Following successful silencing of mmu_circRNA_003795, the expression of miR-1249-5p was increased both in MC3T3-E1 and MDPC23 cells (Fig. 3C and D). Compared with the negative control group, the expression of OPN was significantly decreased in the mRNA microarray. It is known that OPN is one of the early marker proteins involved in osteogenic differentiation (14). The present study speculates that OPN may be involved in regulating osteoblast differentiation and mineralization in MDPC23 cells. Therefore, the current study also tried to verify the changes in OPN when interfering with circ003795 and miR-1249-5p. The expression levels of COL15A1 and OPN were significantly decreased in both cell lines (Fig. 3E and F). Western blotting revealed the same result; the western blot analysis demonstrated that silencing of mmu_circRNA_003795 reduced the protein expression levels of COL15A1 and OPN (Fig. 3G and H).

An inhibitor and mimic were used to knockdown and overexpress miR1249-5p, respectively, in MC3T3-E1 and MDPC23 cells. Compared with the negative control group, the expression of miR1249-5p was decreased in the inhibitor group and increased in the mimic group in both MC3T3-E1 and MDPC23 cells (Fig. 4A and B). Following successful knockdown of miR1249-5p, the expression levels of COL15A1 and OPN were increased in MC3T3-E1 cells, whereas the expression levels of COL15A1 and OPN were decreased when miR1249-5p was overexpressed (Fig. 4C). This result was also demonstrated in MDPC23 cells (Fig. 4D). Western blotting revealed the same result. When miR1249-5p was upregulated or downregulated, the protein expression levels of COL15A1 and OPN were downregulated and upregulated, respectively (Fig. 4E and F). Compared with the negative control group, the staining and activity of ALP in the mmu_circRNA_003795 inhibition group was significantly reduced at 7 days (Fig. 5A). In addition, the ARS staining at 21 days revealed the same result. Following osteogenic induction for 21 days, the intensity of ARS staining in the mmu_circRNA_003795 inhibition group was significantly reduced. Compared with the negative control group, the matrix mineralization of the mmu_circRNA_003795 inhibition group was decreased (Fig. 5B and C).