MicroRNA‐21 affects mechanical force–induced midpalatal suture remodelling

Abstract Objectives miR‐21 can promote osteoblast differentiation of periodontal ligament stem cells. However, the effect of miR‐21 on bone remodelling in the midpalatal suture is unclear. This study aimed to elucidate the effects of miR‐21 on the midpalatal suture bone remodelling by expanding the palatal sutures. Materials and methods miR‐21 deficient (miR‐21−/−) and wild‐type (WT) mice were used to establish animal models by expanding the palatal sutures. Micro‐CT, haematoxylin‐eosin (HE) staining, tartrate‐resistant acid phosphatase (TRAP) staining, fluorescence labelling and immunohistochemistry were used to investigate the function of miR‐21 in midpalatal suture bone remodelling. Besides, bone mesenchymal stem cells (BMSCs) derived from both miR‐21−/− and WT mice were cultured. The MTT, CCK8, EdU analysis, transwell and wound healing test were used to assess the effects of miR‐21 on the characteristics of cells. Results The expression of ALP was suppressed in miR‐21‐/‐ mice after expansion except 28 days. The expression of Ocn in WT mice was much higher than that of miR‐21‐/‐ mice. Besides, with mechanical force, miR‐21 deficiency downregulated the expression of Opg, upregulated the expression of Rankl, and induced more osteoclasts as TRAP staining showed. After injecting agomir‐21 to miR‐21‐/‐ mice, the expression of Alp, Ocn and Opg/Rankl were rescued. In vitro, the experiments suggested that miR‐21 deficiency reduced proliferation and migration ability of BMSCs. Conclusions The results showed that miR‐21 deficiency reduced the rate of bone formation and prolonged the process of bone formation. miR‐21 regulated the bone resorption and osteoclastogenesis by affecting the cell abilities of proliferation and migration.

At present, numerous studies on promoting bone remodelling during RME have been focused on growth factors, drugs and physical stimulation. [12][13][14][15] Whether there are factors regulating RME and mediating palatal bone remodelling at post-transcriptional level remains unknown. microRNA (miRNA), a type of small non-coding RNA, has been reported as a critical post-transcriptional modulator in bone remodelling. 16 Furthermore, many studies have revealed that some miR-NAs, including miR-21, can regulate osteogenic differentiation as a response to the mechanical stimuli. 17 Our previous study found that the expression of miR-21 in human periodontal ligament stem cells (PDLSC) with mechanical stimuli was significantly different from that in non-stressed ones. 18 Based on bioinformatics prediction, the target genes of miR-21 are significantly enriched in Jak-STAT signalling pathways and MAPK signalling pathways, which are related with osteogenic differentiation. 18 In addition, our previous work confirmed that miR-21 can promote stretch-induced PDLSC osteogenesis by acting on activin receptor 2B (ACVR 2B) in vitro. 19 Besides, miR-21 can also promote the differentiation of osteoclasts induced by the receptor activator of nuclear factor κB ligand (RANKL) in vitro. 20,21 In vivo, Chen and his colleagues reported that miR-21 plays an important role in alveolar bone remodelling, including osteogenic differentiation of periodontal ligament stem cells and osteoclast differentiation. 22 Nevertheless, the in vivo function of miR-21 in the regulation of palatal bone remodelling at the post-transcriptional level, particularly in response to orthopaedic force, remains elusive.
The present study, therefore, aimed to detect the influence of miR-21 on palatal bone remodelling in mice by analysing the physical and metabolic changes of midpalatal suture in both normal and mechanical stimuli environments.

| MATERIAL S AND ME THODS
We have already conformed to the ARRIVE guidelines.

| Animals and genotyping
Wild-type (WT) C57BL/6 mice and miR-21 −/− mice were provided by Animal Experimental Center Shandong University and Jackson Laboratory, respectively. Animal experiments were performed based on the guidelines of the Animal Use and Care Committee of Shandong University. Mice were housed in a well-ventilated room with 12h/12h light-dark cycle. Food and water were offered ad libitum.
Standard polymerase chain reaction (PCR) genotyping for WT and miR-21 −/− mice was performed as previously described. 23 Briefly, genomic DNA was extracted from the tail of mice. The primer sequences of PCR were directly sourced from the Jackson Laboratory: WT (5′-TTG CTT TAA ACC CTG CCT GAG CAC-3′) and mutant miR-21 (5′-ACT TCC ATT TGT CAC GTC CTG CAC-3′). The PCR products were evaluated by agarose gel electrophoresis. A single band obtained at 262 base pair (bp) was identified as WT mice, while a band at 500 bp was identified as miR-21 −/− mice ( Figure 1A).
In addition, total RNA was also extracted from mouse heart, liver, spleen and palate. miR-21 expression of miR-21 −/− mice was much lower than that of WT mice ( Figure 1A).

| Animals and treatments
Six-week-old male WT and miR-21 −/− mice (weight: 19 to 21g) were used to establish a model of RME as previously described. 24 Briefly, an opening loop made from the stainless steel of the orthodontic wire (0.014 inch wire size) (Tomy, Japan) was bonded with a light cured adhesive (3M Unitek, CA) to the first and second maxillary molars on both sides ( Figure 1B). It provided an initial force of 0.49 N. Mice without operation served as non-expansion control.
WT and miR-21 −/− mice were randomly divided into 4 groups, including the control WT group, WT group with expansion force (WT + RME) group, control miR-21 −/− group, and miR-21 −/− group with expansion force (miR-21 −/− +RME) group. Mice were euthanized at different time points: 1, 3, 7, 14 or 28 days, with 3 mice in each group at each time point. Animals were weighed at the beginning and the end of the experimental period. The scheme for the animal experiment was shown in Figure S1.

| Fluorescence labelling and sample processing
Mice were labelled by fluorescence according to previous methods. 25 On day 3 and day 13 after the opening loops were applied, alizarin complexone (Sigma, USA) and calcein (Sigma, USA) were injected intraperitoneally at 60 and 20 mg/kg body weight, respectively ( Figure S1).
Mice were euthanized on the second day after the second dye injection. Maxillae were harvested and fixed with 4% paraformaldehyde.
Then, they were embedded in methyl methacrylate without decalcification for slicing hard tissue. Sections of 140 μm were cut with a hard tissue-slicing machine (Leica, Germany). Double-labelled sections were viewed by a fluorescence microscope (Leica, Germany).

| Histology and histochemistry
Detailed methods were described in the Appendix S1.

| Immunohistochemistry
Detailed methods and all antibodies used in this study were described in the Appendix S1.

| Cell culture
Detailed methods were described in the Appendix S1.

| EdU labelling
Detailed methods were described in the Appendix S1.

| Migration assays and Wound Healing Test
Detailed methods were described in the Appendix S1.

| Statistical Analysis
The differences between different groups of mice were analysed with Student's t test. P < .05 was viewed as statistically significant.

| Changes in Body Weight, Palatal Bone Volume and Suture Morphology
During the experiment, the mice that the opening loops had fallen off were not counted in the statistics. The body weights of the mice with activated (expansion) or without opening loops decreased on days 1 and 3 ( Figure S2). The loops bonded to maxillary molars may disturb food intake at initial stages, but the mice recovered quickly.
At later time points, body weights of these operated mice increased gradually ( Figure S2). Therefore, the changes observed following midpalatal suture expansion were unlikely to be caused by systemic physiological responses to the procedure.
Histologically, as Figure 1C and Figure S3 showed, the midpalatal suture of non-operated groups consisted mainly of cartilage, that is two masses of chondrocytes (arrows showed) covering the edges of palatal bones. During the experimental period, the suture in control animals underwent minor changes related to normal growth with a decrease in the number of chondrocytes ( Figure 1D). In the expansion groups, the midpalatal suture was expanded, chondrocytes decreased in numbers ( Figure 1D), and the collagen fibres were reoriented across the suture. At the same time, periosteal cells started to migrate into the suture. Periosteal cells in miR-21 −/− mice migrating to the palatal suture need more time than WT mice. As Figure 1C showed, the number of periosteal cells in miR-21 −/− mice is much smaller than WT mice on day 7 after expansion. However, on day 28, the number of migrated cells in miR-21 −/− mice is more than that of WT ones, while WT mice started to reconstruct the normal palatal suture. The suture of miR-21 −/− mice on day 28 ( Figure 1C) is similar to that of WT mice on day 14 ( Figure S3R). It indicated periosteal cells in miR-21 −/− mice migrating to the palatal suture need more time than WT mice. Bone formation was initially observed at the edges of the palatal bones at day 7 in WT mice. While in miR-21 −/− mice, newly formed bone was observed at day 28, which was later than WT mice. On the oral side, several layers of chondrocytes with a structure similar to the cartilage layers of the original suture can only be detected at day 28 in WT mice.

| miR-21 Deficiency Reduced the Rate of Bone Formation
To better assess the function of miR-21 in new bone formation, we labelled bone surfaces with alizarin complexone and calcein during expansion period ( Figure S1). As shown in Figure 2A After force application, the expressions of Alp in both miR-21 −/− and WT mice were reduced during the early days, compared with control groups ( Figure S4A). But there were significant elevations of Alp in WT mice after 7 days of expansion and in miR-21 −/− mice after 28 days compared with control groups (Figure 2B,D). Moreover, the expression of Alp was suppressed in miR-21 −/− mice after expansion except 28 days ( Figure 2B,D, Figure S4A). Since the expression of Ocn usually occurs in the late osteogenesis process, 26 a little difference was detected in the expression of Ocn. The expression of Ocn in WT mice after 28 days of expansion was much higher than that of control group ( Figure 2C,E). But no significant change was found in miR-21 −/− mice in these groups. The expression of Ocn was significantly suppressed in miR-21 −/− mice ( Figure 2C,E, Figure S4B). These findings collectively suggested that miR-21 deficiency extended the time of bone formation and reduced the rate of bone formation after RME.

| miR-21 Regulated the Bone Resorption and Osteoclastogenesis
We examined the potential effects of miR-21 in bone resorption.
Under the physiologic condition, miR-21 deficiency inhibited bone resorption as shown by tartrate-resistant acid phosphatase (TRAP) staining ( Figure 3A,D, Figure S5A). This result was further supported by increased osteoprotegerin (Opg) expression ( Figure 3B,E, Figure   S5B,D) and decreased Rankl expression ( Figure 3C,F, Figure S5C Figure   S5C,E), which was consistent with TRAP staining. The function that miR-21 regulated bone resorption and osteoclastogenesis was realized by regulating Opg and Rankl.

| Agomir-21 Injection Rescued Decreased Bone Formation and Ratio of Opg/Rankl During Midpalatal Expansion in miR-21 Deficiency Mice
As mentioned before, miR-21 −/− mice presented decreased bone formation and ratio of Opg/Rankl. To further prove the function of miR-21, miR-21 −/− mice received agomir-21 through intraperitoneal injection, which could increase the expression of miR-21. The result of PCR showed that the expression of miR-21 in miR-21 −/− mice after injecting agomir-21 was much higher than that of miR-21 −/− mice ( Figure   S6). Figure 4B showed periosteal cells in the expanded palatal suture of agomir-21 injected mice is more than that of miR-21 -/mice ( Figure   S3M). Calcein-positive bone surface of agomir-21-injected mice was stronger than that of miR-21 −/− mice after midpalatal expansion ( Figure 4A). In order to confirm the finding, the expressions of Alp and Ocn were both detected with immunohistochemical staining and immunofluorescence. After intraperitoneal injection of agomir-21, the expressions of Alp and Ocn were significantly rescued compared to miR-21 −/− mice ( Figure 4D-G), which suggested that miR-21 was required for bone formation after midpalatal expansion.
Furthermore, the regulation of osteoclastogenesis was confirmed by both TRAP staining and the ratio of Opg/Rankl ( Figure 4L). As Figure 4C presented, fewer osteoclastogenesis formed after injecting agomir-21. The expressions of Opg and Rankl were lower in mice after injecting agomir-21 than miR-21 −/− and WT mice ( Figure 4H-K). However, the ratio of Opg/ Rankl was higher than that of miR-21 −/− mice, which indicated less bone resorption in agomir-21-injected mice. These findings indicated that miR-21 participated in the process of osteoclastogenesis in midpalatal expansion.  A). The miR-21 expression of WT mice was much higher than that of miR-21 −/− mice (Figure S7 B). The proliferation of miR-21-deficient BMSCs was poorer compared to wild ones through EdU analysis ( Figure 5C,D). The finding was further supported by MTT ( Figure 5A) and CCK8 ( Figure 5B). As Figure 5A-D showed, in terms of proliferation, there was a significant difference between these two kinds of BMSCs in 48 h. Moreover, we examined the migration ability of two kinds of BMSCs by transwell migration assays and wound healing test ( Figure 5E-G). miR-21 deficiency led to significantly decreased migrated cells ( Figure 5E,F). In wound healing test, the distance of cell migration in wild cells was longer than miR-21deficient ones, which can be obviously detected in 24 h ( Figure 5G). These findings suggested that miR-21 regulated proliferation and migration ability of BMSCs to regulate bone formation during RME.

| D ISCUSS I ON
Our previous study showed that miR-21 was related with mechanical force-induced osteogenesis. 19  increased. Previous studies have reported that with the action of mechanical force, many factors such as smad5 can affect the osteoclast differentiation. 32 Li et al 33 reported miR-21 could increase the expression of smad5, which will downregulate the expression of osteoclast serum markers. 32 Therefore, with mechanical force, smad5 may be one of the factors regulated by miR-21 to affect bone resorption. In our study, bone resorption mainly started from bone surface where osteoclasts were mainly distributed.
We further observed that the number of periosteal cells, which can promote osteogenic differentiation, 24,34 in WT mice was much more than that of miR-21 −/− mice, suggesting that miR-21 deficiency might affect the biological characteristics of cells. In vitro, it revealed that proliferation and migration ability of miR-21deficient cells were poorer, which was consistent with previous studies. 35,36 Since decreased proliferation and migration ability of periosteal cells lead to decreased osteogenesis, less osteoblasts and fibroblasts in miR-21 −/− expansion group can be exactly explained in this study. In addition, many researches have proved that miR-21 in BMSCs is benefit for osteogenesis both in vitro and in vivo. 37,38 For example, miR-21 overexpressing can promote osteogenic differentiation and accelerate fracture healing. Our study is consistent with previous studies.
In summary, our findings showed that miR-21 is related with the changes in biological characteristics of cells and the maturation of newly formed bone in the process of RME.

ACK N OWLED G EM ENTS
This study was supported by grants from National Natural Science Foundation of China (81771030), and Construction Engineering Special Fund of "TaishanScholars" (tsqn20161068 and ts201511106).

CO N FLI C T O F I NTE R E S T
All authors declare that they have no potential conflicts of interest.

AUTH O R CO NTR I B UTI O N S
M. Li contributed to design, data acquisition, analysis and interpretation, and drafted and critically revised the manuscript; Z. Zhang and X.
Gu contributed to data acquisition and analysis, and critically revised the manuscript; Y. Jin, C. Feng and S. Yang contributed to data analysis, and critically revised the manuscript; F. Wei contributed to conception, and critically revised the manuscript. All authors gave final approval and agreed to be accountable for all aspects of the work.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data used to support the findings of this study are available from the corresponding author upon request.