CircRNA AFF4 promotes osteoblast cells proliferation and inhibits apoptosis via the Mir-7223-5p/PIK3R1 axis

Fracture healing is a complex process involving various cell types, cytokines, and mRNAs. Here, we report the roles of the circRNA AFF4/miR-7223-5p/PIK3R1 axis during fracture healing. We found that increased expression of PIK3R1 during fracture healing is directly associated with augmented proliferation and decreased apoptosis of MC3T3-E1 cells. Furthermore, miR-7223-5p targeted PI3KR1 and inhibited MC3T3-E1 proliferation while promoting apoptosis. CircRNA AFF4 acted as a sponge of miR-7223-5p, thereby promoting MC3T3-E1 cell proliferation and inhibiting apoptosis. Local injection of circRNA AFF4 into femoral fracture sites promoted fracture healing in vivo while the injection of miR-7223-5p delayed healing. These findings suggest that CircRNA AFF4 promotes fracture healing by targeting the miR-7223-5p/PIK3R1 axis, and suggests miR-7223-5p, CircRNA AFF4, and the miR-7223-5p/PIK3R1 axis are potential therapeutic targets for improving fracture healing.


INTRODUCTION
Despite the advancement in medical technology, regeneration and restoration of bone defects remain challenging [1]. It is estimated that 10% of all the fractures result in non-union, which likely lead to a poor quality of life for the affected patients and an economical burden to society [2][3][4]. The fracturehealing process is highly complex, with an active involvement of various signaling molecules [4]. Under the regulation of these molecules, bones undergo remolding through either intramembranous ossification or endochondral ossification. Osteoblasts have been identified as key players that enhance bone repair [5]. Therefore, finding a way to promote osteoblast proliferation and suppress apoptosis during fracture healing could greatly improve fracture care.
miRNAs are small non-coding RNA molecules that act as post-transcriptional regulators of gene expression. By interacting with the 3'untranslated region of the target mRNAs, they regulate protein production, thereby modulating a series of physiological processes [6]. Circular RNAs (circRNAs) are a novel class of regulatory RNAs that are generated from precursor mRNA by back-splicing of exon(s). They can be located in the cytoplasm or the nucleus, and modulate gene expression either directly or indirectly, through regulation of miRNAs or RNA binding proteins, respectively [7]. Recently, cytoplasmic circRNAs have been well characterized as an endogenous "sponge" for miRNAs [8]. While circRNAs increase the expression of their target genes through inhibition of miRNAs, the expression of both of these types of RNA during fracture healing is poorly understood.
Recent microarray analyses have implicated many miRNAs in various diseases [9]. In the present study, we hypothesized that miRNA levels are altered during fracture healing compared to controls. To test this hypothesis, we analyzed publicly available data from the open-source gene expression omnibus (GEO) database. We indeed identified new miRNAs that are highly associated with fracture healing and predicted their downstream target genes using other open-source databases such as Targetscan (http://www.targetscan.org/mamm_31/) and miRDB (http://mirdb.org/). In addition, we investigated their upstream regulation by circRNAs using the CircBase (http://www.circbase.org), and explored their roles using in vitro and in vivo femoral fracture models. We successfully elucidated the circRNAs-miRNAs-mRNAs interactions during fracture healing.

miR-7223-5p inhibits fracture healing
We explored whether miRNAs are differentially expressed during fracture healing using data obtained from the Gene Expression Omnibus database (GSE76197). We found six miRNAs that were downregulated (fold change ≤-2.0, p<0.05) during fracture healing ( Figure 1A), and selected miR-7223-5p for further analyses given its large changes in expression during fracture healing ( Figure 1B). We will explore the effects of the other miRNAs during fracture healing in future studies.
To investigate the roles of miR-7223-5p during fracture healing, equal amounts of PBS (control) and agomiR-7223-5p were injected locally into fracture sites in mice and measured various parameters for each group during fraction healing. The levels of miR-7223-5p in the calluses were higher in the agomiR-7223-5p-treated group until Day 21 ( Figure 1C). We performed micro computed tomography (microCT) and found that bone volume (BV), total volume (TV), and bone mineral density (BMD) were lower in the agomiR-7223-5p group compared to the control group, both on day 14 and day 21 following the fracture ( Figure 1D, 1E). AgomiR-7223-5p treatment also inhibited the expression of some genes related with osteoblast differentiation (e.g., BMP2 and Runx2) ( Figure 1F-1H). Overall, the results of our in vivo study thus indicate that upregulation of miR-7223-5p delays fracture healing.
Luciferase reporter assay showed that WT circRNA AFF4 signals were suppressed by miR-7223-5p treatment while circRNA AFF4 mutations abolished the inhibitory effect of miR-7223-5p ( Figure 4B). CircRNA AFF4 was markedly increased during the first several days of fracture healing ( Figure 4C). The predicted splice junction of circRNA AFF4 was validated in MC3T3-E1 cells, and the product was amplified using divergent primers ( Figure 4D, 4E). FISH analysis demonstrated that circRNA AFF4 was mainly located in the cytoplasm of MC3T3-E1 cells ( Figure 4F). CircRNA AFF4 was more resistant to RNase R treatment compared to mRNA AFF4 ( Figure  4G). Pull-down assays revealed a higher level of circRNA AFF4 in the miR-7223-5p-captured fraction compared to the miR-NC-capture fraction ( Figure 4H). AGO2 immunoprecipitation showed that after the transfection of cells with miR-7223-5p, the endogenous circRNA AFF4 was pulled down by AGO2 ( Figure 4I). RNA FISH showed co-localization of circRNA AFF4 and miR-7223-5p in the cytoplasm of MC3T3-E1 cells ( Figure 4J) and the cells in the uninjured bone and calluses ( Figure 4K). Taken together, our data indicate that circRNA AFF4 can directly bind to miR-7223-5p.

CircRNA AFF4 promotes proliferation and inhibits apoptosis of MC3T3-E1 cells in vitro
To further investigate the roles of circRNA AFF4 during fracture healing, we transfected MC3T3-E1 AGING cells with circRNA AFF4 or si-circRNA AFF4. The transfection efficiency was assessed by qRT-PCR, and the result confirmed that circRNA AFF levels were increased in the circRNA AFF4-transfected cells ( Figure 5A). CCK8 assay revealed that circRNA AFF promotes cell proliferation, a result that was further confirmed by Edu staining (Figure 5B, 5C). CircRNA AFF4 cells exhibited highly levels of Cyclin D1 and Cyclin D3, proteins that promote cell cycle progression ( Figure 5D). Regard apoptosis, the proapoptotic protein Bax was suppressed while the antiapoptotic protein (Bcl-2) was induced by circRNA  AFF4 transfection ( Figure 5E). Moreover, circRNA AFF4 decreased the percentage of apoptotic cells ( Figure 5F). In addition, overexpression of circRNA AFF4 resulted in higher levels of PIK3R1 ( Figure 5G, 5H). Collectively, these data indicated that circRNA AFF4 promotes MC3T3-E1 cell proliferation and inhibits apoptosis through targeting miR-7223-5p and PIK3R1.

Therapeutic use of circRNA AFF4 in a mouse model of femoral fracture
To explore the role of circRNA AFF4 in facture healing in vivo, we locally injected equal amounts of PBS, circRNA AFF4, and si-circRNA AFF4 into the fracture sites. Our microCT results showed that the BV, TV, and BMD of the calluses were increased by the injection of circRNA AFF4 ( Figure 6A, 6B). Taken together, these results indicate a potential therapeutic effect of circRNA AFF4 in promoting fracture healing.

DISCUSSION
Despite recent improvements in our knowledge of miRNAs and circRNAs, our understanding of their roles during fracture healing is still limited. Dysregulation of miRNAs has been widely reported in various biological processes, such as cell migration, cell proliferation, and cell metabolism, among others [10]. Several miRNAs have been reported to be involved in bone formation, exhibiting pro-or antiosteogenesis effects during fracture healing [11,12]. In order to further characterize the miRNAs that regulate fracture healing, we used microarray data to identify the miRNAs that are dysregulated during fracture healing and found that miR-7223-5p was downregulated in calluses during fracture healing compared normal bones. We confirmed that miR-7223-5p levels were downregulated in our in vivo fracture mouse model during the first several days of fracture healing and gradually increased with bone formation. In order to further verify the effects of miR-7223-5p on fracture healing, we injected agomiR-7223-5p into the fracture sites. Although we injected miR-7223-5p on days 0, 4, and 7, miR-7223-5p levels were still higher in the agomiR-7223-5p group than in the control group on day 21. Overall, our in vivo data showed that fracture healing was suppressed by injection of agomiR-7223-5p. Results from our in vitro analyses were consistent with those of from our in vivo experiments,  showing that miR-7223-5p overexpression inhibited proliferation and promoted apoptosis of MC3T3-E1 cells.
CircRNAs can be generated from exons, intronic or intergenic regions. Their ability to regulate mRNA expression depends on their location within the cell. In the present study, we found that circRNA AFF4 was mainly located in the cytoplasm of MC3T3-E1 cells. Previous studies reported that the cytoplasmic circRNAs act as an endogenous miRNA sponge of miRNA, thereby affecting the functions of the target miRNAs at the post-transcriptional level [17]. The circular form of circRNAs protect them from degradation by exonucleases [18], thus allowing longer exertion of their effects in cells. We observed increased levels of circRNA AFF4 during the first several days following fracture in vivo. Its expression pattern was the opposite of that of miR-7223-5p, suggesting that circRNA AFF4 may promote fracture healing by suppressing miR-7223-5p. RIP, FISH, and luciferase assays in vitro revealed that miR-7223-5p was a downstream target of circRNA AFF4. CircRNA AFF4 promotes proliferation and inhibits apoptosis of MC3T3-E1 cells. In addition, PI3KR1 expression was enhanced by circRNA AFF4. Overall, our results indicate that circRNA AFFT, miR-7223-5p, and PI3KR1 compose an axis that regulates fracture healing.
Previous studies reported that apoptosis and senescence may occur after RNA transfection into cells [19], both of which inhibit cell proliferation and migration. However, it normally takes a longer time (more than 48 h) to trigger cell senescence than apoptosis [20,21]. In the present study, we measured cell proliferation and apoptosis 24 h after miR-7223-5p and circRNA AFF4 transfection. Impaired proliferation of MC3T3-E1 cells may have been the main contributing factor inducing cell apoptosis.
The results from our in vitro and in vivo experiments using an animal model warrant further studies with calluses collected from patients. In addition, fracture healing is highly complex; therefore, there may exist other cell types (preosteoblasts, osteoblasts, fibroblasts, stromal cells, stem cells, etc.), non-coding RNAs, and mRNAs involved in fracture healing. Further studies are needed to understand the roles of circRNAs, miRNAs, and mRNAs on various cell types during fracture healing.
In conclusion, our study showed that circRNA AFF4 could promote proliferation and inhibit apoptosis of MC3T3-E1 cells through the miR-7223-5p/PIK3R1 axis (Figure 7), which may be exploited for therapeutic benefits to improve fracture healing.

Cell counting kit-8 (CCK8) assay
1×10 4 transfected cells were seeded in a 96-well plate. After 0, 24, 48 and 72 h following the seeding, 10 μL of CCK8 reagent was added to each well and quantified using a microplate reader at 450 nm.

Edu staining
Transfected MC3T3-E1 cells were seeded into 48well plates. 24 h after the seeding, Edu regent (Sigma-Aldrich) was added to each well. After 2h, the cells were fixed with 4% formaldehyde for 15 min and treated with 0.5% Triton X-100 for permeabilization. Then, Apollo reaction cocktail was added to each well, followed by cell staining with Hoechst 33258.

Cell apoptosis assay
Cell apoptosis was assessed by staining cells with Annexin V-fluorescein isothiocyanate (FITC) and propidium iodide (PI) according to the manufacturer's instructions. AGING

Western blot
Total proteins were isolated from cells or calluses. Equal amounts of protein were loaded onto the wells, separated in 10% sodium dodecyl sulfate-polyacrylamide (SDS-PAGE) gel, and transferred to PVDF membranes by electrophoresis. The membranes were incubated with the primary antibodies at 4 °C overnight, then with HRPconjugated secondary antibodies for 2 h at room temperature. The primary antibodies used are as follows

qRT-PCR
Total RNAs were isolated from cells or calluses using TRIzol reagent (Invitrogen, USA). qRT-PCR was performed on a real-time PCR system. The samples were treated with RNase R (3 u/mg) for 15 min. GAPDH and U6 snRNA were selected as controls for normalizing mRNA or miRNA, respectively. The results were analyzed using the comparative Ct method (2 -ΔΔCt ). The primers used are listed in Supplementary Table 1.

Luciferase reporter assay
To study the interaction of miR-7223-5p with circRNA AFF4 MC3T3-E1 cells were co-transfected with wildtype (WT) or mutated circRNA AFF4 reporter plasmids and miR-7223-5p or miR-NC. Similarly, MC3T3-E1 cells were co-transfected with either the WT or mutated FGFR3 reporter plasmids and miR-7223-5p or miR-NC. The luciferase signal was evaluated using the Dual Luciferase Reporter Assay according to the manufacturer's instructions.

RNA fluorescence in situ hybridization (FISH)
To investigate the intracellular localization circAFF4, we seeded cells into 6-well plates. After 24 h, we fixed cells by incubating them with 4% paraformaldehyde for 20 min. The cells were then incubated in a hybridization buffer with Cy3-labeled circRNA AFF4 probes at 37 °C overnight. The tissue sections were deparaffinized and permeabilized by proteinase K treatment at 37 °C for 30 min. Then, the slices were incubated in the hybridization buffer with Cy3-labeled circRNA AFF4 probes and FAM-labeled miR-7223-5p at 37 °C overnight. The nuclei were counterstained with 4,6diamidino-2-phenylindole (DAPI). The images were obtained using a confocal microscope (Nikon, Tokyo, Japan). The sequences of the probes were as follows: circRNA AFF4: 5′-CY3-GCTGCTCTTCCTCATCTC TGCTGTA CCTC-3′; miR-7223-5p: 5′-FAM-CCGA CACCTCAGCAAAGCTA-3′. Briefly, a transverse osteotomy on the femur was created, and the bones were fixed with a 23-gauge needle intramedullary. Equal amounts (30 μL) of PBS, AgomiR-7223-5p, and Plasmid CircRNA AFF4 were injected into the fracture site on days 0, 4, and 7. The mice were sacrificed at a specific timepoint, and samples of calluses were harvested for WB, qRT-PCR, and microCT analysis.

Micro-CT and radiography examination
The femoral fracture sites were scanned ex vivo and three-dimensional images were reconstructed using a microCT system (BRUKER, Germany). The BV, TV, BV/TV and BMD values were calculated to assess the quality of calluses.

CircRNAs in vivo precipitation (circRIP)
CircAFF4-overexpressing MC3T3-E1 cells were fixed by incubation with 1% formaldehyde for 10 min and lysed. After centrifugation at 4 °C and 10,000 g for 10 min, 50 μL of the supernatant was retained. The remaining supernatant was incubated with streptavidin dyna-beads (M-280; Invitrogen) overnight. On the following day, the streptavidin dyna-beads-circAFF4 probes mixture was incubated with 200 μL of lysis buffer for 2 h to reverse the formaldehyde cross-linking. Finally, TRIzol (Invitrogen, USA) was used to isolated RNA from the mixture for qRT-PCR.

Statistical analysis
All the data are shown as mean ± SD. The results were analyzed with Graph Prism 8.0 (GraphPad Software, AGING