Emerging roles of circular RNAs in osteoporosis

Abstract Osteoporosis is one bone disease characterized with skeletal impairment, bone strength reduced and fracture risk enhanced. The regulation processes of bone metabolism are associated with several factors such as mechanical stimulation, epigenetic regulation and hormones. However, the mechanism of osteoporosis remains unsatisfactory. Increasing high‐throughput RNA sequencing and circular RNAs (circRNAs) microarray studies indicated that circRNAs are differentially expressed in osteoporosis. Growing functional studies further pinpointed specific deregulated expressed circRNAs (e.g., circ_28313, circ_0016624, circ_0006393, circ_0076906 and circ_0048211) for their functions involved in bone metabolism, including bone marrow stromal cells (BMSCs) differentiation, proliferation and apoptosis. Moreover, CircRNAs (circ_0002060, Circ_0001275 and Circ_0001445) may be acted as diagnostic biomarkers for osteoporosis. This review discussed recent progresses in the circRNAs expression profiling analyses and their potential functions in regulating BMSCs differentiation, proliferation and apoptosis.

In the manuscript, we reviewed expression profiling of circRNA studies in osteoporosis to offer the datasets for choosing specific osteoporosis-associated circRNAs for next studies in the future. The potential diagnostic and therapeutic roles of circRNAs in the clinical application for osteoporosis are also discussed.

| CIRCRNA E XPRE SS ION PROFILING AND INTEG R ATIVE ANALYS IS IN SCI
Huang et al. 45 used circRNA chip analysis to study circRNA expression profiling in the plasma or serum from 40 osteoporosis patients and 40 control adults. A total of 237 circRNAs were found to be differently expressed and 95 circRNAs were downregulated and 162 circRNAs were upregulated. Moreover, they performed qRT-PCR analysis to prove that four circRNAs such as circ_0081047, circ_0068459, circ_0006873 and circ_0002060 were upregulated and circ_0017736 and circ_0062582 were downregulated in osteoporosis patients compared to control adults.
Zhao et al. 46 performed circRNA microarray to analysis circRNA expression profiling in PBMCs (peripheral blood mononuclear cells) from three postmenopausal osteoporosis patients and three controls. There were 381 circRNAs were differently expressed in postmenopausal osteoporosis compared to control. Among these, 203 circRNAs were overexpressed and 178 circRNAs were downregulated. They performed qRT-PCR assay to confirm that circ_0028882, circ_0001275, circ_0006766, circ_0007788 and circ_0003391 was upregulated and circ_0006801 was downregulated.
Jin et al. 47 used RNA sequencing to study circRNA expression profiling in three cases with postmenopausal osteoporosis and three healthy controls. A total of 260 circRNAs were differentially regulated in postmenopausal osteoporosis group. Among these, 154 circRNAs were downregulated and 106 circRNAs were overexpressed in osteoporosis group. The top five downregulated circRNAs were circ_0021739, circ _0011269, circ_0019693, circ_0005245 and circ_0010349, and the top five upregulated circRNAs were circ_0010452, circ_0022348, circ_0015566, circ_0003323 and circ_0013121.
Chen et al. 48 performed microarray profiling analyses to detect the cirRNA expression pattern in the bone marrow monocyte/ macrophage-derived osteoclasts and undifferentiated BMM cells. A total of 6259 circRNAs were decreased, and 5449 circRNAs were overexpressed after BMM cells induction. Moreover, a total of 81 circRNAs were remarkably different by hierarchical clustering and 52 circRNAs were downregulated and 29 circRNAs were upregulated. They confirmed that the expression of circ_012460, circ_8313, circ_28312, circ_28309, circ_40206 and circ_28236 were overexpressed in induction group.
Yu et al. 49 used RNA sequencing and bioinformatics assay to screen for circRNAs expression profiling in six postmenopausal osteoporosis patients and healthy controls. A total of 387 circRNAs were found to be differentially expressed in the osteoporosis compared to control including 176 decreased circRNAs and 211 overexpressed circRNAs.
Moreover, they showed that circ_0057340 and circ_0134944 were overexpressed and circ_0005692, circ_0088422 and circ_0076906 were decreased in osteoporosis group by qRT-PCR analysis.
Lin et al. 50 have studied the circRNAs expression profiling in the osteoclast differentiation without and with alendronate treatment.
There were 1394 circRNAs were upregulated and 214 circRNAs were downregulated in the osteoclast (OC) precursors (OPCSs) groups compared to OC group. GO assay showed that differentially expressed circRNAs were distributed into three groups: molecular function, biological process and cellular component. In addition, a total of 110 circRNAs were deregulated expressed among OC +alendronate, OC and OPCS groups and 15 circRNAs were downregulated and 95 circRNAs were overexpressed after alendronate treatment.
Wang et al. 51 used RNA sequencing to analyse aberrantly expressed circRNAs in BMSCs from ovariectomy mice and controls.
There are 45 circRNAs were found to be differentially expressed and 21 circRNAs were decreased and 24 circRNAs were overexpressed in the ovariectomy mice compared to controls. Moreover, they confirmed that circ-0020 expression was overexpressed and circ-3832 level was downregulated in ovariectomy mice compared to controls.
Xu et al. 52 used circRNAs microarray to study circRNAs expression profiling in the plasma or serum from osteoporosis and controls.
A total of 69 circRNAs were differentially expressed in osteoporosis group compared to control groups, and among these, 35 circRNAs were decreased and 34 circRNAs were overexpressed. They performed qRT-PCR assay to prove that circ_0019693, circ_0011269, circ_0028958, circ_0005245, circ_0006487 and circ_0010452 expression was consistent with the microarray data.
Shen et al. 53 used circRNA microarray assay to study circRNAs expression profiling in bone tissues from osteoporosis group and noosteoporosis group. A total of 4972 circRNAs were differentially expressed in osteoporosis group compared to no-osteoporosis group, and among these, 2645 circRNAs were overexpressed and 2327 cir-cRNAs were decreased.
Yao et al. 54 performed RNA sequencing to identify circRNAs expression profiling in peripheral blood from three patients from senile osteoporotic vertebral compression fracture (OVCF) and three healthy controls. They discovered that there are 884 circRNAs were differentially expressed and 330 circRNAs were downregulated and 554 circRNAs were overexpressed in OVCF groups compared to healthy controls.
Liu et al. 55 performed RNA sequencing to investigate circRNAs expression profiling in five postmenopausal osteoporosis patients with five normal controls. There were 250 circRNAs were deregulated in osteoporosis compared to normal controls. Among these, 186 circRNAs were upregulated and 64 circRNAs were downregulated.
Wang et al. 56 explored circRNAs expression profiling in the BMSCs after treated with melatonin or not by using RNA sequencing. They found that there were 209 circRNAs were differentially expressed in human BMSCs after treated with melatonin.
Zhi et al. 57 performed the circRNA microarray to detect the differentially expressed circRNAs in serum of three controls and three osteoporosis patients. There were 589 circRNAs were differentially expressed in the serum in osteoporosis patients compared to controls. Among these, 213 circRNAs were decreased and 376 circRNAs were overexpressed in the osteoporosis patients.

| circ_0002060 and circ_0006873
Huang et al. 45 performed circRNA chip method to study circRNA expression profiling in the plasma or serum from 40 osteoporosis patients and 40 control adults, and their data indicated that circ_0006873 and circ_0002060 were upregulated in osteoporosis patients compared to control adults. The level of circ_0002060 and circ_0006873 was correlated with bone mineral density (BMD) and T-score. The circ_0002060 levels have diagnostic values for osteoporosis patients (AUC = 0.746, p < 0.05), and the specificity and sensitivity for circ_0002060 were 69% and 78%.

| circ_28313
Chen et al. 48 showed that circ_28313 was overexpressed in the

| circ_0016624
Yu et al. 60 demonstrated that circ_0016624 was downregulated in the osteoporotic patient samples compared to healthy controls.
They indicated that miR-98 may be one target of circ_0016624 and miR-98 was upregulated in osteoporosis group compared to controls. Moreover, they found that BMP2 was decreased in osteoporosis group compared to controls. The expression of BMP2 and circ_0016624 was overexpressed while miR-98 level was downregulated during osteogenic differentiation. Ectopic expression of circ_0016624 can suppress miR-98 expression and enhance BMP2 expression. Furthermore, circ_0016624 overexpression induced osteogenesis differentiation but decreased through miR-98 mimics. These data suggested that circ_0016624 inhibited osteoporosis through sponging miR-98 and promoting BMP2 expression.

| circ_0011269
Xu et al. 52 demonstrated that circ_0011269 expression was downregulated in the osteoporosis group compared to control group.
They found that miR-122 was one target gene of circ_0011269 and

| circ_0026827
Ji et al. 56 showed that circ_0026827 was upregulated in the dental

| CircFOXP1
Shen et al. 44 showed that circFOXP1 was decreased in osteoporosis bone tissues compared to control group and circFOXP1 acted as a sponge to regulated miR-33a-5p expression and induced its target gene FOXP1 expression.

| circ_0006859
Zhi et al. 70  These data suggested that circ_0006215 act a critical role in the osteogenesis and may be one therapy target for senile osteoporosis.

| circ_0021739
Guan et al. 71 demonstrated that circ_0021739 was downregulated in the postmenopausal osteoporosis patients compared to controls.
The expression of circ_0021739 was associated with femur, forearm and vertebra T-scores. The AUC of ability of circ_0021739 expression was 0.849, with a specificity of 42.9% and sensitivity of 100%.
Ectopic expression of circ_0021739 suppressed miR-502-5p expression and suppressed the osteoclasts differentiation. In summary, they demonstrated that circ_0021739 may be one potential biomarker for postmenopausal osteoporosis and circ_0021739 modulated osteoclasts differentiation through regulating miR-502-5p.

| Circ_0062582
Li et al. 46 showed that the expression of circ_0062582 was upregulated in BMSCs during osteogenic differentiation. Overexpression of circ_0062582 induced osteogenic differentiation and promoted the expression of osteogenic differentiation-associated genes such as COL1, OCN and OSX. Furthermore, they found that circ_0062582 sponged the expression of miR-145 and CBFB was a direct target gene of the miR-145. Ectopic expression of circ_0062582 regulated BMSCs osteogenic differentiation through modulating miR-145/CBFB axis.

| circ_0001275
Xu et al. 57 demonstrated that dexamethasone-induced the hFOB1.19 cell proliferation inhibition was reversed through silencing circ_0001275. Dexamethasone promoted the calcium nodules and ALP activity in the hFOB1.19 cell, while this function was also reversed by circ_0001275 siRNA. circ_0001275 sponged miR-377 expression in the hFOB1.19 cells and CDKN1B was one direct target gene of miR-377. Furthermore, they showed that knockdown of circ_0001275 reverses dexamethasone-induced cell proliferation inhibition through promoting miR-377/CDKN1B axis.
Their data suggested that inhibition expression of circ_0001275 can reverse dexamethasone-induced osteoblast proliferation inhibition through promoting miR-377/CDKN1B axis (Table 3 and Figures 3 and 4).

| CON CLUS I ON S AND FUTURE PER S PEC TIVE S
Growing data of high-throughput RNA sequencing and circRNAs microarray studies indicated that circRNAs are differentially expressed in osteoporosis. CircRNAs may be acted as diagnostic biomarkers for osteoporosis. Increasing studies suggested that cir-cRNAs played important roles in the BMSCs proliferation, differentiation and apoptosis. However, the expression levels and function roles of these differentially expressed circRNAs in osteoporosis remain uncharacterized. Further studies may be required to confirm the expression level of these differentially expressed circRNAs in more samples of human osteoporosis and controls. Moreover, further functional works on these differentially expressed circR-NAs are needed to build their potential therapeutic or pathogenic significance.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

CO N S E NT TO PA RTI CI PATE
Not applicable.

CO N S E NT TO PU B LI S H
Not applicable.

DATA AVA I L A B I L I T Y S TAT E M E N T
Research data are not shared.