CircRUNX2 through has‐miR‐203 regulates RUNX2 to prevent osteoporosis

Abstract Objective We aimed to discover the molecular mechanism of hsa_circ_0076694 (circRUNX2) on osteogenic differentiation. We also explored the interaction between circRUNX2, miR‐203 and RUNX2. Methods Clinical samples obtained from femoral neck fracture patients’ bone tissues were used to collect circRUNX2, miR‐203, and RUNX2 expression data, while their expression changes were observed in human bone mesenchymal stem cells (hBMSCs) during osteogenic differentiation. QRT‐PCR and Western blot were used to analyse levels of RNAs and proteins. Biotin pull down, RIP, RNA FISH, and Dual‐Luciferase Reporter assays demonstrated the relationship between circRUNX2, miR‐203, and RUNX2. ALP and ARS staining were used to measure the degree of osteogenic differentiation under the control of circRUNX2, miR‐203. Results CircRUNX2 were down‐regulated in osteoporotic patients’ bone tissues. CircRUNX2 could inhibit miR‐203 expression by sponging miR‐203. MiR‐203 inhibited osteogenic differentiation by targeting the 3′‐UTR of RUNX2 and down‐regulate RUNX2 expression. Overexpression of circRUNX2 promoted the expression of osteogenic differentiation‐related proteins such as RUNX2, OCN, OPN, BSP, and prevented osteoporosis. Conclusion circRUNX2 could sponge miR‐203 and enhance RUNX2 expression, thus circRUNX2 prevents osteoporosis and may provide a novel therapeutic strategy for it.

Recently much attention has been paid to Circular RNAs (circRNAs), which is one of a class of noncoding RNAs, with circular structure by joining the 3′end of the RNA to the 5′end. 4 They were important in regulating the expression of other genes. For example, Liu et al found 2:27713879|27755789 and 2:240822115| 240867796 participated in regulation network in osteogenic differentiation. 5 Also, they were able to affect the expression of some essential carcinogenic mRNAs by targeting miRNAs and changing the miRNAs expression which would subsequently change the miRNAs level. 6 It was predicted that circRNAs would combine with miRNA and compete for the binding sites on the miRNA with mRNAs. 7 Cir-cRNAs expression would alter the mRNA expression by targeting miRNAs that were binding to the mRNAs chosen. For instance, it is reported that hsa_circRNA_103801, a novel circRNA, targeted genes which were mainly involved in biological processes and cellular components. 8 CircUBAP2 inhibited the expression of microRNA-143 (miR-143) giving rise to promote osteosarcoma growth and inhibit apoptosis both in vitro and in vivo. 9 Also, circRNA_0009910 was found to be overexpressed in osteosarcoma cells (OS cells), and knocking down of which inhibited cell proliferation, and induced cell cycle arrest as well as apoptosis in OS cells. It worked by targeting miR-449a, which had the opposite effect on OS cells. 10 These researches provide us great examples to deem that miRNA-circRNA interaction is considered to one of better direction for therapy. Runt related transcription factor, known as RUNX factors, which proteins located in sub-nuclear domains and integrated cell signals during formation of gene promoter regulatory complexes. 11 RUNX2 is a member of the human RUNT related transcription factor family essential during embryogenesis for skeletal development. It was recognized on account of its oncogenic properties and many studies showed that a de-regulating of RUNX2 function leads to progression and invasion of different tumours. 12 Also, RUNX2 was reported to closely relate to bone formation and hypertrophic chondrocyte differentiation. 13 Some evidence showed that the down-regulated RUNX2 protein expression inhibited bone formation, and decreased bone mass. 14 Although there were already some researches on RUNX2-related regulation in osteogenic differentiation, the understanding towards to the mechanism of it was far from being comprehensive. Herein, we wanted to investigate the molecule mechanism of RUNX2 in the osteogenic differentiation.
In this study, we tried to find the expression levels of miR-203, circRUNX2, and RUNX2 in osteoporotic patients and how miR-203, circRUNX2, and RUNX2 affect each other expression levels to influence the development of the osteoporosis.

| Clinical samples
Trabecular bone samples from the trochanteric region of the femur at a site far from the periarticular bone were obtained from 20 patients undergoing hip replacement because of femoral neck fracture. A total of 10 osteoporotic specimens were collected from women aged 62-89 (average age: 75). Ten nonosteoporotic specimens derived from women patients with suffered external traumatic fracture. To keeping samples fresh, frozen environment was needed.
None of the patients underwent any medical treatment that affected bone or mineral metabolism before enrolling in this study. The study was authorized by the Ethics Committee of Wuxi No. 9 People's Hospital Affiliated to Soochow University, and all patients were offered informed consents and signed for those confirmations.

| Cell culture and induction of osteogenic differentiation
Human bone mesenchymal stem cells (hBMSCs) were bought from the BeNa Culture Collection (BNCC, Beijing, China). HBMSCs were cultured in minimum essential medium Alpha Medium (α-MEM) supplemented containing 10% FBS, 2Mm L-Glutamine, 100 U/mL penicillin and 100 μg/mL streptomycin and stored in a humidified atmosphere of 5% CO 2 at 37°C. Osteogenic (OS) differentiation of hBMSCs was induced according to a previously published protocol. 15 The OS medium was supplemented with osteogenic inducers:

| QRT-PCR
Before extract of total RNA, the bone is rapidly placed in the mortar that is precooled with liquid nitrogen, and then repeatedly ground to the powder. Total RNA was extracted using TRIzol reagent (Thermo-Fisher, China) and cDNA was synthesized using iScript ™ cDNA Synthesis Kit (Biorad, China) following the kit's protocol. iQ ™ SYBR Green supermix (Biorad, China) was used to perform QRT-PCR. 7500 HT Fast Real-Time PCR System (Applied Biosystems) was used to cycle and quantify reactions. Relative gene expression levels of circRUNX2, miR-203 and RUNX2 were evaluated. U6 and GADPH were used as miRNA and long RNA endogenous normalization controls. Primers for RT-PCR were supplied in Table 1.

| Biotin pull down assay
The pull-down assay with biotinylated RNA was performed as described. 16,17 In brief, for circRUNX2 pulled down miRNAs, the

| RNA fluorescence in situ hybridization (FISH)
The RNA fluorescence in situ hybridization assay was performed by

| Alizarin red staining
Alizarin red staining (ARS) was performed to detect the osteoblast calcification. For ARS, hBMSCs were seeded at a density of

| Statistical analysis
All experiments were repeated at least three times. Data were analysed using Graph Pad Prism 6.0 software. All data were presented as mean ± SD. Differences between two groups were identified by unpaired t test, and differences between more than two groups were analysed using one-way analysis of variance (ANOVA).

| CircRUNX2 and RUNX2 was lowly expressed in osteoporotic patients
CircRUNX2 and RUNX2 mRNA expression levels was found to be lower in osteoporotic patients' bone tissues than nonosteoporotic bone tissues ( Figure 1A,B). Spearman correlation analysis analysed correlation between circRUNX2 and RUNX2 which showed a positive relationship ( Figure 1C). In addition, proteins of RUNX2, ALP, BSP, OCN, and OPN were detected by western blotting in osteoporotic and nonosteoporotic bone tissues. Results showed protein levels of RUNX2, BSP, OCN, and OPN was lower in osteoporotic bone tissues than that in nonosteoporotic bone tissues and ALP protein level was no obvious difference between osteoporotic and nonosteoporotic bone tissues ( Figure 1D).

| MiR-203 contained the targets of circRUNX2 and RUNX2
MiRNAs targeting circRUNX2 were predicted by Home database.

RNA.org and in
Following, dual-Luciferase Reporter assay was performed to verify the target relationships of miR-203 and circRUNX2/RUNX2 (Figure 3A,B). The results implied the target relationships with the phenomenon that miR-203 reduced the luciferase activity in the group of wild-type circular RNA (WT-circ) and wild-type RUNX2 3′ UTR (WT-RUNX2). MiR-203 level was also detected in osteoporotic patients, and osteoporotic patients were with a higher miR-203 level compared to nonosteoporotic patients ( Figure 3C). Furthermore, the negative correlation relationship between RUNX2 mRNA and miR-203 were observed by Spearman correlation analysis in 10 clinical samples ( Figure 3D).

| CircRUNX2 and miR-203 affected osteogenic differentiation through regulating RUNX2
The expression levels of circRUNX2, miR-203, and RUNX2 were detected at the 0 day, 7th days, 14th days, and 21th days of osteogenic differentiation. During hBMSCs osteogenic differentiation, cir-cRUNX2 and RUNX2 expressions showed an increasing trend, but miR-203 expression was gradually decreased ( Figure 4A,B,C). To further investigate the regulation links between circRUNX2, miR-203, and RUNX2, qRT-PCR assays were performed after regulating cir-cRNUX2, miR-203, or RUNX2. Results showed that, in hBMSCs, cir-cRUNX2 overexpression lead to the reduction in miR-203 and the increase in RUNX2 mRNA, circRUNX2 down-regulation caused the opposite results ( Figure 4D). MiR-203 up-regulation inhibited RUNX2 expression, and miR-203 down-regulation promoted RUNX2 expression. In Addition, circRUNX2 was not regulated by miR-203 (Figure 4E). Similarly, circRUNX2 and miR-203 were not regulated by RUNX2 ( Figure S1A). Furthermore, the differentiation level was measured by ALP and ARS staining assays in these four kinds of hBMSCs ( Figure 6). After several days (7 16 CircHIPK3 regulated cells growth by combining to miR-124 and reducing its activity. 21 Therefore, we continued to explore molecular mechanism during the progress of osteogenic differentiation with circRNA-miRNA-mRNA axis models. In the past, some circRNAs, like circ19142 and circ5846, are reported to be involved in BMP2-reduced osteoblast differentiation, through a circ19142/circ5846-targeted miRNA-mRNA axis. 6 However, the role of circRNAs in this process, for instance, hsa-circ-0076694 (circRUNX2), remained to be researched. In this research, we focused on how circRUNX2 influenced osteoporosis and revealed the mechanism of osteogenic differentiation related to cir-cRUNX2. We linked the correlation of circRUNX2 and RUNX2 and predicted their common targeting miRNA. What's more, we used, RNA pull down assay, RIP assay, RNA FISH assay, dual-Luciferase Reporter assay, and functional experiments to verify our prediction.
The present research revealed the molecular regulation mechanism of circRUNX2/miR-203/RUNX2 axis during the progress of osteogenic differentiation, which will provide a novel therapy target for osteoporosis.
Although we tried to take all the factors into consideration in our experiments, some aspects could still be improved. For example, because of the difficulties in obtaining clinical samples, we only applied 10 osteoporosis specimens for our investigation. The conclusions would be more reliable on condition that there were more samples.
Furthermore, there are several other bone differentiation related proteins detected in our experiments, like BSP, OCN, and OPN. Thus, next experiment might focus on how circRUNX2 regulated these proteins and enrich the mechanism of osteoblast differentiation.

| CONCLUSION
This study mainly researches the effects of the circRUNX2 and its relations with other close genes. CircRUNX2 might prevent osteoporosis through regulating has-miR-203 and RUNX2. In order to get the more safe and effective therapies for osteoporosis, more cir-cRNAs, mRNAs, and miRNAs should be researched to find out their effect and the links among them.

CONFLI CT OF INTEREST
The authors confirm that there are no conflicts of interest.