Bone marrow (5–7 ml) was obtained from five male healthy voluntary donors (22–32 years old) at Department of Orthopedics, Qinghai Provincial People's Hospital (Quinghai, China) from January 2017 to March 2018. The use of human samples was approved by the Ethical Committee of the Qinghai Provincial People's Hospital and informed consent was obtained from each patient.

A total of 10 male C57BL6J mice (4–6 weeks old; 18–20 g) were obtained for use in the current study (Charles River Laboratories, Inc.). All mice were housed at 25±5°C with 40–50% humidity and a 12 h dark/light cycle, and access to food and water ad libitum. All animal experiments were performed following the guidelines provided by the National Institutes of Health for the Care and Use of Laboratory Animals. The current study was approved by the Ethics Committee of Qinghai Provincial People's Hospital (Quinghai, China).

Primary mBMMSCs were isolated from C57BL/6 mice and hBMMSCs were isolated from human bone marrow as previously described (30). BMMSCs were cultured in DMEM supplemented with 10% fetal bovine serum (FBS) and antibiotics [100 units/ml penicillin G and 100 µg/ml streptomycin (all, Gibco; Thermo Fisher Scientific, Inc.)]. Cells from passages 3–5 were used in the subsequent experiments.

The mouse preosteoblast cell line, MC3T3-E1, was obtained from the American Type Culture Collection (cat. no. CRL-2594). Cells were cultured in α-minimum essential medium (α-MEM; Gibco; Thermo Fisher Scientific, Inc.) containing 10% FBS and 1% penicillin-streptomycin mix. Cells were then incubated at 37°C with 5% CO₂.

To induce osteogenic differentiation, mBMMSC, hBMMSC and MC3T3-E1 cells (4×10⁴ cells) were seeded into six-well plates. When 80% confluence was achieved, cells were incubated in osteogenic-inducing medium [10% FBS; 5 mM L-glycerophosphate; 100 nM dexamethasone (Sigma-Aldrich; Merck KGaA) and 50 mg/ml ascorbic acid (Sigma-Aldrich; Merck KGaA)] at 37°C for 14 days. Control group cells were cultured in α-MEM supplemented with 10% FBS at 37°C. To assess the osteoblast phenotype, osteoblast marker genes were examined by RT-qPCR and alkaline phosphatase (ALP) activity was determined.

MC3T3-E1 cells were seeded into six-well plates at a density of 1×10⁶ cells per well and incubated at 37°C for 24 h. Cells were then transfected with 100 nM miR-98-5p mimic (sense 5′-UGAGGUAGUAAGUUGUAUUGUU-3′, antisense 5′-CAAUACAACUUACUACCUCAUU-3′), 100 nM mimic control (sense 5′-UUCUCCGAACGUGUCACGUTT-3′ and antisense, 5′-ACGUGACACGUUCGGAGAATT-3′; both Guangzhou RiboBio Co., Ltd.), 2 µg Control CRISPR activation plasmid (control-plasmid; cat. no. sc-437275), 2 µg HMGI-C CRISPR Activation Plasmid (HMGA2-plasmid; cat. no. sc-420880-ACT; Santa Cruz Biotechnology, Inc.) or 100 nM miR-98-5p mimic with 2 µg HMGA2-plasmid using Lipofectamine 2000^® (cat. no. 11668-027; Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. Control cells received no treatment. 48 h after treatment, transfection efficiency was assessed using RT-qPCR.

To detect extracellular ALP activity, the spectrophotometric method (31) was performed with an Alkaline Phosphatase Assay kit (cat. no. P0321; Beyotime Institute of Biotechnology) was used according to the manufacturer's protocol. Absorbance was detected at a wavelength of at 405 nm using a microplate reader (BD Biosciences) following the manufacturer's protocol.

Transfected MC3T3-E1 cells were seeded into 96-well plates (5×10³ cells per well). After transfection for 48 h with miR-98-5p mimics, mimic controls, or miR-98-5p mimics in combination with the HMGA2-plasmid, 100 µl α-MEM medium mixed with 10 µl cell counting kit-8 (CCK-8) solution (cat. no. C0041; Beyotime Institute of Biotechnology) was added to each well. Cells were subsequently incubated at 37°C for 30 min. Cell viability was then evaluated by detecting absorbance at 450 nm using a FLUOstar Omega Microplate Reader (BMG Labtech GmbH). Cells without treatment were used as controls. The aforementioned experiment was repeated three times.

To assess the effect of miR-98-5p on MC3T3-E1 cell apoptosis, a FITC/propidium iodide (PI) apoptosis detection kit (cat. no. 70-AP101-100; MultiSciences Biotech Co., Ltd.) was utilized. Subsequent to 48 h transfection with miR-98-5p mimics, mimic controls, or miR-98-5p mimics in combination with the HMGA2-plasmid, MC3T3-E1 cells were digested with 0.25% Trypsin, washed with PBS and subsequently stained with 5 µl Annexin V-FITC and 5 µl PI for 30 min at room temperature. Finally, to analyze cell apoptosis, the FACSCanto II flow cytometer (BD Biosciences) was used according to the manufacturer's protocol, and data was analyzed by Cell Quest software (version 5.1; BD Biosciences). Cells without treatment were used as the control.

Proteins were extracted from MC3T3-E1 cells using RIPA buffer (cat. no. P0013E; Beyotime Institute of Biotechnology) following the manufacturer's protocol. A Bicinchoninic Acid Assay kit (cat. no. BCA1-1KT; Sigma-Aldrich; Merck KGaA) was used to quantify protein samples. Equal quantities of protein (30 µg per lane) were separated on 10% SDS-PAGE and transferred onto PVDF membranes. Membranes were then blocked with 5% non-fat milk at room temperature for 1 h and incubated with the following primary antibodies (all 1:1,000) overnight at 4°C: HMGA2 (cat. no. 8179; Cell Signaling Technology, Inc.), ALP (cat. no. sc-365765; Santa Cruz Biotechnology), runt related transcription factor 2 (Runx2; cat. no. 12556; Cell Signaling Technology, Inc.), transcription factor sp7 (Osterix; cat. no. sc-393060; Santa Cruz Biotechnology, Inc.) and β-actin (cat. no. 4970; Cell Signaling Technology, Inc.). Samples were then incubated with a horseradish peroxidase-conjugated anti-rabbit immunoglobulin G secondary antibody (cat. no. 7074; 1:1,000; Cell Signaling Technology Inc.) at room temperature for 2 h. Corresponding protein bands were subsequently visualized using Western Blotting Luminol reagent (cat. no. sc-2048; Santa Cruz Biotechnology, Inc.) following the manufacturers' protocol.

Total RNA was extracted from cells using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc.) following the manufacturer's protocol. Total RNA was then reverse transcribed into cDNA using the TaqMan MicroRNA Reverse Transcription kit (cat. no. 4366596; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. Synthesized cDNA was subsequently analyzed via qPCR using a SYBR Premix Ex TaqTM II (TliRNaseH Plus) kit (cat. no. RR820a; Takara Bio, Inc.). Primer sequences were as following: miR-98-5p forward, 5′-ACACTCCCUAUACAACUUAC-3′ and reverse, 5′-GGGAAAGUAGUGAGGCCTCAGA-3; U6 forward, 5′-GCTTCGGCAGCACATATACTAAAAT-3′ and reverse, 5′-CGCTTCACGAATTTGCGTGTCAT-3; GAPDH forward, 5′-CTTTGGTATCGTGGAAGGACTC-3′ and reverse, 5′-GTAGAGGCAGGGATGATGTTCT-3; RUNX2 forward, 5′-GATGATGACACTGCCACCTCT-3 and reverse, 5′-AGGGCCCAGTTCTGAAGC-3; ALP forward, 5′-CCGAATTCATGTTGGCCTGTTCAACT-3′ and reverse, 5′-ATGTCGACTTAGTTATTTTCATAATACCAAATTCC-3; Osterix forward, 5′-AGAGATCTGAGCTGGGTAG-3; and reverse, 5′-AAGAGAGCCTGGCAAGAGG-3; HMGA2 forward, 5′-TCCCTCTAAAGCAGCTCAAAA-3′ and reverse, 5′-ACTTGTTGTGGCCATTTCCT-3. The following thermocycling conditions were used for the qPCR: Initial denaturation at 95°C for 10 min, and 35 cycles of 95°C for 15 sec and 55°C for 40 sec. U6 or GAPDH was used as an internal control for miRNA and mRNA expression, respectively. The 2^−ΔΔCq method was used to calculate relative gene expression (32).

miR-98-5p targets were predicted using TargetScan bioinformatics software (www.targetscan.org/vert_71) and binding sites between HMGA2 and miR-98-5p were observed. The wild type (WT-HMGA2) and mutant (MUT-HMGA2) 3′-UTR of HMGA2 were cloned into a pmiR-RB-ReportTM dual luciferase reporter gene plasmid vector (Guangzhou RiboBio Co., Ltd.). MC3T3-E1 cells were co-transfected with WT-HMGA2, MUT-HMGA2 and miR-98-5p mimics or mimic control using Lipofectamine^® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. Subsequent to 48 h of cell transfection, luciferase activity was measured using the dual-luciferase assay system (Promega Corporation) and normalized to Renilla luciferase activity.

All experiments were performed three times. SPSS software version 17.0 was used to analyze data (SPSS, Inc.). Experimental data were expressed as the mean ± standard deviation. Differences between multiple groups were identified using one-way ANOVA followed by a Tukey's post-hoc. A Student's t-test was used to compare the difference between two groups. P<0.05 was considered to indicate a statistically significant difference.

To assess whether miR-98-5p served a role in bone formation, miR-98-5p expression during osteogenic differentiation was detected. Three osteoblast cell models (primary hBMMSC, primary mBMMSC and MC3T3-E1 cells) were used to achieve this. The osteoblast phenotype was confirmed by increased extracellular ALP activity (Fig. 1A) and by the enhanced expression of osteoblast makers including ALP, Runx2 and Osterix (Fig. 1B-D). Compared with day 0, levels of miR-98-5p in hBMMSC, mBMMSC and MC3T3-E1 cells were all significantly reduced at day 7 and day 14 after osteogenic differentiation incubation (Fig. 1E). The results indicate that miR-98-5p may therefore be involved in bone formation.

TargetScan (http://www.targetscan.org) was utilized to predict potential targets of miR-98-5p. Binding sites between HMGA2 and miR-98-5p were revealed (Fig. 2A). To confirm these, potential binding sites a luciferase reporter assay was performed. As presented in Fig. 2B, compared with MUT-HMGA2 and miR-98-5p mimic transfected cells, luciferase activity was significantly reduced following WT-HMGA2 and miR-98-5p mimic transfection. The results demonstrate that miR-98-5p may directly target HMGA2.

MC3T3-E1 cells were transfected with a miR-98-5p mimic, a mimic control, a control-plasmid, a HMGA2-plasmid, or a miR-98-5p mimic in combination with a HMGA2-plasmid for 48 h and subsequently subjected to osteogenic differentiation induction. Transfection efficiency was measured using RT-qPCR. As presented in Fig. 3A-C, the miR-98-5p mimic significantly upregulated miR-98-5p in MC3T3-E1 cells. Furthermore, protein and mRNA levels of HMGA2 in MC3T3-E1 cells were enhanced following treatment with the HMGA2-plasmid. The data demonstrated that the miR-98-5p mimic significantly reduced the mRNA level of HMGA2. The protein level of HMGA2 was markedly decreased by miR-98-5p mimic. However, these decreases were reversed by HMGA2-plasmid co-transfection (Fig. 3D and E).

On day 14 after induction, extracellular ALP activity increased and was subsequently inhibited by treatment with the miR-98-5p mimic. However, this inhibition was eliminated following HMGA2-plasmid treatment (Fig. 4A). Furthermore, treatment with the miR-98-5p mimic markedly decreased the protein levels, and significantly decreased the mRNA levels of osteoblast makers including ALP, Runx2 and Osterix. However, these decreases were reversed following HMGA2-plasmid transfection (Fig. 4B-E). The results demonstrate that miR-98-5p may prevent osteogenic differentiation.

MC3T3-E1 cells were transfected with a miR-98-5p mimic, a mimic control, or a miR-98-5p mimic in combination with a HMGA2-plasmid for 48 h. A CCK-8 assay and flow cytometry were subsequently performed to detect MC3T3-E1 cell viability and apoptosis, respectively. The results revealed that the miR-98-5p mimic significantly reduced MC3T3-E1 cell viability and induced MC3T3-E1 cell apoptosis and these changes were reversed by HMGA2 upregulation (Fig. 5A-C).