MiR‐664‐3p suppresses osteoblast differentiation and impairs bone formation via targeting Smad4 and Osterix

Abstract Osteoporosis is a metabolic disorder characterized by low bone mass and deteriorated microarchitecture, with an increased risk of fracture. Some miRNAs have been confirmed as potential modulators of osteoblast differentiation to maintain bone mass. Our miRNA sequencing results showed that miR‐664‐3p was significantly down‐regulated during the osteogenic differentiation of the preosteoblast MC3T3‐E1 cells. However, whether miR‐664‐3p has an impact on bone homeostasis remains unknown. In this study, we identified overexpression of miR‐664‐3p inhibited the osteoblast activity and matrix mineralization in vitro. Osteoblastic miR‐664‐3p transgenic mice exhibited reduced bone mass due to suppressed osteoblast function. Target prediction analysis and experimental validation confirmed Smad4 and Osterix (Osx) are the direct targets of miR‐664‐3p. Furthermore, specific inhibition of miR‐664‐3p by subperiosteal injection with miR‐664‐3p antagomir protected against ovariectomy‐induced bone loss. In addition, miR‐664‐3p expression was markedly higher in the serum from patients with osteoporosis compared to that from normal subjects. Taken together, this study revealed that miR‐664‐3p suppressed osteogenesis and bone formation via targeting Smad4 and Osx. It also highlights the potential of miR‐664‐3p as a novel diagnostic and therapeutic target for osteoporotic patients.


| INTRODUC TI ON
Bone homeostasis is a dynamic balance that includes bone formation by osteoblasts and bone resorption by osteoclasts. 1 Once the orchestrated balance is destroyed, it causes various destructive bone diseases, including osteoporosis. 2 Osteoporosis has become a global health problem, especially for postmenopausal women and ageing population. 3 MiRNAs are a class of small non-coding RNAs, around 22 nucleotides in length, which bind via incomplete or complete base pairing to specific sequences in the 3′-untranslated region (3′UTR) or coding sequence (CDS) of mRNAs, inducing either translational repression or mRNA degradation. 4,5 MiRNAs have been reported to play diverse biological roles in regulating multiple physiological and pathological processes, including cellular differentiation, proliferation, glucose metabolism, cholesterol biosynthesis and cancer development. [6][7][8][9][10] Increasing numbers of miRNAs have been implicated as regulators of different aspects of bone development.
Certain miRNAs, such as miR-139-3p, miR-34a and miR-143, inhibit osteogenic differentiation, [11][12][13] and conversely, miR-200c, miR-149 and miR-196a promote osteogenic differentiation. [14][15][16] Besides, several miRNAs have emerged as critical regulators of osteoclast biology. 17 Although several studies have revealed that miRNAs regulate bone formation and bone resorption, thus contributing to bone homeostasis maintenance, most of these miRNAs have only been identified in vitro, and their functional roles in the pathophysiological mechanisms responsible for reduced bone formation in skeletal disorders remain to be established before they can be applied in a clinical setting. MiRNAs are also recognized as attractive therapeutic targets due to their size, known sequence and the fact that they can target multiple genes to subsequently alter cellular pathways and networks. In fact, a number of Phase I/II human clinical trials are underway towards testing the effects of miRNA mimics or antagomirs in vivo to treat specific diseases including cancers. For example, a locked-nucleic-acid-modified anti-miR-122 drug Miravirsen was successfully tested in phase II clinical trials in anti-HCV therapy. 18 TargomiRs loaded with miR-16 mimic were tested in phase I clinical trials in malignant pleural mesothelioma. 19 Icariin (ICA), a prenylated flavonol glycoside isolated from the Epimedium pubescens, exerts beneficial effects on preventing postmenopausal bone loss and treating osteoporosis. 20 Previously published data and our recent study indicated that ICA promoted osteogenic differentiation in vitro and alleviated osteoporosis in vivo. [21][22][23] To identify novel miRNAs involved in bone formation, differential miRNA expression profiles between ICA-treated and untreated MC3T3-E1 cells were evaluated. MiR-664-3p was one of the most significantly down-regulated miRNAs during osteogenic differentiation. Previous studies demonstrated that miR-664-3p was closely related to the occurrence and development of a variety of tumours. [24][25][26] However, there have been no studies on the potential role of miR-664-3p in bone metabolism. In addition, two key regulators of osteogenesis, including Smad4 and Osterix (Osx), are the potential targets of miR-664-3p. On this basis, we chose miR-664-3p as a research object to explore its negative effect on osteoblast differentiation and bone formation.
Our data reveal new regulatory mechanisms of bone metabolism and also suggest a potential diagnostic and therapeutic target for osteoporosis.

| Cell culture and osteoblast differentiation
Murine preosteoblast MC3T3-E1 cells were purchased from Cell Bank of the Chinese Academy of Sciences (Shanghai, China) and cultured in α-MEM (Gibco, Carlsbad, CA, USA). HEK 293T cells and murine embryonic mesenchymal stem cell line C3H10T1/2 were obtained from American Type Culture Collection (Manassas, VA, USA) and grown in DMEM (Gibco), respectively. All culture media were supplemented with 10% fetal bovine serum and 1% penicillin/ streptomycin (Invitrogen, Carlsbad, CA, USA). Cells were incubated in a humidified atmosphere with 5% CO 2 at 37°C. In vitro osteoblast differentiation of MC3T3-E1 and C3H10T1/2 cells was carried out using osteogenic induction medium containing standard growth medium supplemented with 50 mg/L ascorbic acid, 10 mmol/L βglycerophosphate and 10 nmol/L dexamethasone (Sigma-Aldrich, Louis, MO, USA).

| Cell transfection
MiR-664-3p mimic (Mimic-664) and its negative control (Mimic-NC) were purchased from RiboBio (Guangzhou, China) and transfected into cells at a concentration of 100 nmol/L. For plasmid transfection, 2 μg DNA was used for each plasmid when cotransfected with miR-NAs into cells in six-well plates. All transfections were performed using Lipofectamine 2000 (Invitrogen), according to the manufacturer's instructions.

| RNA isolation and quantitative real-time reverse transcription PCR (qRT-PCR)
Total RNA was isolated from cells using TRIzol reagent (Invitrogen) and reverse transcribed into cDNA using the PrimeScript RT Reagent Kit (Takara, Otsu, Japan). Real-time PCR was performed using FastStart Universal SYBR Green Master (Roche, Indianapolis, IN, USA) on a LightCycler 96 Real-Time PCR System. The amplification conditions were as follows: 95°C for 10 minutes, followed by 40 cycles of 95°C for 10 seconds and 60°C for 30 seconds. β-actin was used as an endogenous control. The expression level of mature miR-664-3p was determined by stem-loop qRT-PCR as previously described. 27 U6 was used for normalization of miRNA qRT-PCR data.
The primers used are listed in Table S1 and S2.

| Dual-luciferase reporter assay
The 3′UTR of mouse and human Smad4, the CDS region of mouse Osx, as well as the 3′UTR of human Osx, including the predicted miR-664-3p-binding site, was amplified and cloned into pGL3-Promoter vector (Promega, Madison, WI, USA), respectively. Mutations and deletions were generated by site-directed mutagenesis, by replacing or deleting the ribonucleotides of the miR-664-3p complementary sequence. All constructions were confirmed by sequencing.
The primers used are listed in Table S3. The wide-type, deleted or mutated luciferase reporter constructs were cotransfected into HEK 293T cells with Mimic-664 or Mimic-NC using Lipofectamine 2000. Luciferase activities were measured using the Dual-luciferase Reporter Assay Kit (Promega) following the manufacturer's protocol.
Renilla luciferase was used as an internal control.

| Enzyme-linked immunosorbent assay (ELISA)
Blood was collected intraorbitally from mice, and sera were stored at −80°C. Five samples were obtained for each group. Serum osteocalcin (OCN) level was detected by ELISA (Amersham Pharmacia Biotech, Piscataway, NJ, USA).

| Therapeutic inhibition of miR-664 in OVX mice
All the female C57BL/6J mice used were maintained under standard animal housing conditions (12 hours light, 12 hours dark cycles and free access to food and water). The mice were ovariectomized or sham-operated at 8 weeks of age. At 8 weeks after surgery, all mice were killed by carbon dioxide asphyxiation and the bilateral femurs were collected for micro-computed tomography (μCT). For the therapeutic model, mice were ovariectomized or sham-operated at 8 weeks of age and left untreated until they were 16 weeks old.
The miR-664-3p antagomir (antagomir-664) and its negative control (antagomir-NC) obtained from RiboBio were dissolved in PBS and delivered by subperiosteal injection into the femur metaphysis at 1 nmol/mouse. Control groups were injected with same volume of PBS. All animal experiments were performed in accordance with the National Institutes of Health guidelines and were approved by the Ethics Committee on Animal Care of Nanjing Medical University (Nanjing, China).

| Bone mineral density (BMD) and morphometry
The distal femurs were dissected free of soft tissue, fixed in 4% paraformaldehyde for 24 hours and scanned using μCT (SkyScan 1176; Bruker, Germany). Image acquisition was performed at 100 kV and 98 μA with a 0.98-degree rotation between frames.
The resolution of the μCT images was 18.2 μm. During scanning, the samples were enclosed in tightly fitting plastic wrap to prevent movement and dehydration. For visualization, the segmented data were imported and reconstructed as three-dimensional renderings displayed in the 3D software that comes with the instrument.

| Statistical analyses
Adequate sample size was determined according to the previous studies that performed analogous experiments. Experimental data are presented as mean ± SD from at least three independent experiments. Two sets of data were analysed by two-sided Student's t test. Two-sided P values < 0.05 were considered statistically significant.

| MiR-664-3p expression is down-regulated during osteogenic differentiation
Our previous study has proved that ICA, at 5 μmol/L concentration, exhibited the most prominent stimulatory effects on osteogenic differentiation of MC3T3-E1 cells. 21 To identify novel miRNAs that are associated with osteogenic differentiation, high-throughput sequencing was performed to detect the differential miRNA expression after treatment with 5 μmol/L ICA in MC3T3-E1 cells. The expression of 346 miRNAs was significantly altered (fold-change > 2; P < 0.05), with 251 down-regulated and 95 up-regulated in ICA-treated cells compared with the control cells ( Figure 1A and Table S5). Among them, twenty miRNAs, with potential roles in osteoblast differentiation based on target and pathway prediction analysis, were selected for further validation by stem-loop qRT-PCR, and thirteen of them were validated, including miR-20a-5p, miR-204-5p, miR-224-5p, Figure 1B). Considering that miR-664-3p is at the top of the downregulated miRNAs during ICA-induced osteoblast differentiation, and no report is available on the involvement of miR-664-3p in bone metabolism, we selected miR-664-3p for further study.
To confirm the decreased expression of miR-664-3p during osteogenesis, we examined the expression level of miR-664-3p in MC3T3-E1 and C3H10T1/2 cells after incubation in osteogenic induction medium for 12 days. As expected, miR-664-3p was downregulated gradually in a time-dependent manner in both cells. As expected, the early osteoblastogenic differentiation markers, such as Runx2, Osx and Alp, were remarkably increased at the early stages of culture (0-7 days) and decreased at the later stages of culture (12 days); while the late osteoblastogenic differentiation marker Ocn was gradually increased during osteogenic differentiation ( Figure 1C,D). In addition, we also found miR-664-3p was lowly expressed in the bone tissues derived from 8-week-old mice ( Figure   S1). These data indicated that miR-664-3p might have a negative effect on osteogenic differentiation and bone formation.

| MiR-664-3p inhibits osteoblast activity and matrix mineralization
To explore the effects of miR-664-3p on osteoblast differentiation, Collectively, these data indicated that miR-664-3p was able to inhibit osteogenic differentiation and matrix mineralization in vitro.

| Inhibition of osteogenic differentiation by miR-664-3p is partially mediated by Smad4 and Osx
To determine whether miR-664 functionally targets Smad4 and Osx in regulating osteoblast activity, we used the WT Smad4 3′UTR and

| Inhibition of miR-664-3p partially counteracts the decreased bone phenotype in OVXinduced osteoporotic mice
Numerous clinical studies have shown that the miRNAs associated with bone metabolism are implicated in osteoporosis. [31][32][33] We chose an oestrogen-deficient ovariectomized model to mimic postmenopausal osteoporosis. 34 Figure S6A,B).
To investigate the therapeutic effects of inhibition of miR-664-3p on OVX-induced osteoporosis, antagomir-664 was injected into the femur metaphysis of oestrogen-depleted mice 8 weeks after OVX (OVX+Antagomir-664) ( Figure 6A). μCT analysis revealed that F I G U R E 3 Characterization of bone phenotypes in osteoblast-specific TG664 mice. A, Schematic representation of the generation of TG664 mice. B, qRT-PCR analysis of miR-664-3p levels in bones and other tissues from two different TG664 mouse lines (TG664-1: cre, miR-664-3p +/− and TG664-2: cre, miR-664-3p +/+ ) and control mice (miR-664-3p +/+ ). C, Representative μCT reconstructive images of the femoral metaphysis collected from indicated groups of mice. D, Bone morphometric analysis of trabecular bone of the distal femurs isolated from TG664 and control mice. BMD, bone mineral density; BV/TV, bone volume/tissue volume; Tb.Th, trabecular thickness; Tb.N, trabecular number; Tb.Sp, trabecular separation. E, qRT-PCR analysis of Alp, Ocn, Bsp and Col1a1 in bone tissues collected from TG664 and control mice at 9 weeks of age. F, ELISA assay to determine the serum concentrations of OCN in TG664 and control mice. Data are mean ± SD, n = 5 mice in each group. * P < 0.05, ** P < 0.01 and *** P < 0.001. ns, no significant difference | 5033 XU et al.

| MiR-664-3p is up-regulated in osteoporotic individuals
Given that miR-664-3p silencing prevents OVX-induced osteoporosis, we further evaluated the expression level of miR-664-3p in serum collected from osteoporotic patients and healthy donors.
The stem-loop qRT-PCR results showed that miR-664-3p was highly enriched in the serum of osteoporotic patients as compared with healthy donors ( Figure 7A).

| D ISCUSS I ON
In this study, we identified the differentially expressed miRNAs To determine whether miR-664-3p is a physiologically relevant regulator of bone formation, we also generated osteoblastic miR-664-3p transgenic mice. The TG664 mice exhibited osteoporotic bone phenotypes and lower levels of SMAD4 and OSX proteins compared to control mice. In addition, the expression levels of Alp, Ocn, Bsp and Col1a1 were significantly lower in the femurs of TG664 mice; and similarly, serum OCN level was obviously decreased in TG664 mice. These data suggest that miR-664-3p impaired bone formation, which was partially due to suppressed osteoblast function. Furthermore, we constructed an ovariectomized mouse model to assess the efficacy of therapeutic targeting of miR-664-3p to prevent osteoporosis. μCT revealed an obvious decrease in bone mass in the OVX group, indicating successful establishment of the model. Treatment with antagomir-664 markedly counteracted the decreased bone phenotype in OVX-induced osteoporotic mice. These data indicate that inhibition of miR-664-3p might be a new strategy to treat age-related bone loss and senile osteoporosis.
Most clinical investigations have been performed using bone samples from patients; however, it is not feasible for non-invasive, early diagnosis of osteoporosis. In this study, we observed high enrichment of miR-664-3p in the serum from patients with osteoporosis, suggesting that serum miR-664-3p alone or in combined with other biomarkers may be used for osteoporosis diagnosis.
However, to better explore the diagnostic roles of miR-664-3p, In summary, we demonstrated that miR-664-3p suppressed osteoblast differentiation and impaired bone formation partially via regulation of Smad4 and Osx expression. Silencing of miR-664-3p reversed OVX-induced osteoporosis in vivo ( Figure 7B). Moreover, we observed a significant enrichment of miR-664-3p in the serum of osteoporotic patients. Hence, this study is the first to reveal the role of miR-664-3p in osteogenic differentiation and bone formation, and also highlights the significance of miR-664-3p in the diagnosis and therapy of osteoporosis. F I G U R E 5 Inhibition of osteogenic differentiation by miR-664-3p partially depends on Smad4 and Osx. A and B, WT Smad4 3′UTR or Osx CDS was transfected into MC3T3-E1 and C3H10T1/2 cells, and then cultured in osteogenic induction medium for 48 h, respectively. The mRNA expressions of Alp, Ocn, Bsp and Col1a1 were determined by qRT-PCR (A). SMAD4 and OSX proteins were quantified by western blotting (B). C and D, Representative images of ALP and ARS staining in MC3T3-E1 (C) and C3H10T1/2 (D) cells transfected with Mimic-664 and Smad4 or Osx overexpression plasmid in osteogenic induction medium for 7 or 21 d, respectively. E, qRT-PCR analysis of Alp, Ocn, Bsp and Col1a1 mRNA levels in MC3T3-E1 and C3H10T1/2 cells transfected with Mimic-664 and Smad4 or Osx overexpression plasmid in osteogenic induction medium for 48h. Data are mean ± SD, n = 3. * P < 0.05 and ** P < 0.01 F I G U R E 7 MiR-664-3p is upregulated in the serum of patients with osteoporosis. A, qRT-PCR analysis of miR-664-3p levels in serum samples from female patients with osteoporosis and healthy controls. B, Schematic diagram representing the role of miR-664-3p in osteoblast differentiation and bone formation. *** P < 0.001 F I G U R E 6 Therapeutic inhibition of miR-664-3p counteracts decreased bone phenotype in OVX-induced osteoporotic mice. A, A schematic diagram illustrating the experimental design for the timeline of subperiosteal injection of antisense oligonucleotides specific to miR-664-3p and its negative control. B, Representative μCT reconstructive images of the femoral metaphysis collected from indicated groups of mice. C, Bone morphometric analysis of trabecular bone of the distal femurs isolated from each group. BMD, bone mineral density; BV/TV, bone volume/tissue volume; Tb.Th, trabecular thickness; Tb.N, trabecular number; Tb.Sp, trabecular separation. Data are mean ± SD, n = 5 mice in each group. * P < 0.05 and ** P < 0.01