Circular RNA hsa_circ_0000073 contributes to osteosarcoma cell proliferation, migration, invasion and methotrexate resistance by sponging miR-145-5p and miR-151-3p and upregulating NRAS

An increasing number of studies have demonstrated that circular RNAs (circRNAs), as promising therapeutic targets, are essential for diverse human diseases, especially cancer. However, the potential functions and complex mechanisms of most circRNAs in osteosarcoma (OS) are still not fully elucidated. In the present study, we obtained the expression profile of circRNAs from a GEO database (GSE96964) and identified hsa_circ_0000073 as a highly expressed candidate in OS. Overexpression of hsa_circ_0000073 accelerated the proliferation, migration, invasion and MTX resistance of OS cells, and knockdown of hsa_circ_0000073 resulted in the opposite effects. Mechanistically, hsa_circ_0000073 upregulated NRAS expression by targeting miR-145-5p and miR-151-3p in OS cells. In addition, the promotion of OS progression by hsa_circ_000007 was blocked by miR-145-5p and miR-151-3p-mediated NRAS inhibition. In conclusion, hsa_circ_0000073 facilitated the proliferation, migration, invasion and MTX resistance of OS cells through the inhibition of miR-145-5p and miR-151-3p-mediated downregulation of NRAS.


INTRODUCTION
Osteosarcoma (OS), a highly invasive and metastatic disease, is the most universal primary malignant bone tumor; it has an incidence of approximately 4.4/100,000 [1]. With the development of adjuvant chemotherapy drugs and technologies, these treatments can be applied not only for the removal of cancer cells but also for the evaluation of surgical procedures and for the prevention of early metastasis of OS cells [2]. Statistics reveal that the 5-year survival rate of patients with nonmetastatic OS can reach 65-70%, and only 10-20% of patients require surgical intervention [3]. Despite this, the prognosis for patients with metastatic OS is not encouraging, and the 5-year survival rate for patients with pulmonary metastasis is only 20% [4]. Therefore, an in-depth understanding of the mechanism of metastasis is of great significance for improving the therapeutic effect of OS.
AGING have been reported to serve as sponges on miRNAs, thereby affecting gene transcription or RNA-binding protein interactions [6]. Many studies have also revealed that circRNAs are specifically expressed in tumor tissues and have significant effects on cancer progression [7][8][9]. At present, there have been roles reported in OS for certain circRNAs, such as hsa_circ_0081001 [10], hsa_circ_0001564 [11], hsa_circ_001569 [12], hsa_circ_0002052 [13] and hsa_circ_0051079 [14]. However, research on the function and mechanism of circRNAs in osteosarcoma is still just beginning.
In the present study, we screened the expression profile of circRNAs in methotrexate (MTX)-resistant or nonresistant OS cells from a Gene Expression Omnibus (GEO) database (GSE96964). We identified a novel circRNA, hsa_circ_0000073 (hsa_circ_001069), and we explored its biological roles in the progression of OS. In addition, we also found that hsa_circ_0000073 could positively regulate NRAS expression in OS by competitively binding with miR-145-5p and miR-151-3p. Our results revealed that the hsa_circ_0000073/(miR-145-5p, miR-151-3p)/NRAS axis in OS might provide potential biomarkers and therapeutic targets for OS.

The circRNA hsa_circ_0000073 was identified in a screen because of its high expression in OS
We preliminarily explored the circRNA expression profile of a publicly available OS dataset. The expression profiles of circRNAs from GSE96964 were determined by hierarchical clustering of OS cells (MG63, 143B, U2OS and HOS), human osteoblast cells (hFOB 1.19) and OS MTX-resistant cells (ZOS, ZOSM, and U2OS_MTX) ( Figure 1A). The fold changes were determined for circRNAs that were differentially expressed between U2OS, HOS, MG-63, 143B and hFOB 1.19 cells (Group 1), between U2OS_MTX, ZOS, ZOSM and hFOB 1.19 cells (Group 2), and between U2OS_MTX, ZOS, ZOSM and U2OS, HOS, MG63, 143B cells (Group 3) ( Figure  1B). Through Venn diagrams, we discovered that 8 circRNAs were identified in all three groups. Then, the fold change levels of the 8 circRNAs were displayed: hsa_circ_0000073 and hsa_circ_0084582 were upregulated in OS or MTX-resistant OS among the three groups, while hsa_circ_0003271, hsa_circ_0006422 and hsa_circ_0001449 were downregulated ( Figure 1C). Hsa_circ_0000073 (Supplementary Figure 1) was chosen for the research target because it had the most significant increase.
First, the circular form of hsa_circ_0000073 was assessed by using divergent primers and the cDNA from U2OS and MG-63 cells ( Figure 1D). We also found that the expression of hsa_circ_0000073 was not different between an RNase R treatment group and a mock treatment group, suggesting the strong stability of hsa_circ_0000073 in U2OS and MG-63 cells (P<0.01, Figure 1E). In addition, our results showed that hsa_circ_0000073 was dramatically upregulated in OS cells and tissues (P<0.05, Figure 1F and 1G). Moreover, our data revealed that high expression of hsa_circ_0000073 was associated with poor stage and poor survival of OS (P<0.05, Figure 1H and 1I). In summary, we found that hsa_circ_0000073 was successfully identified from a screen and was highly expressed in OS.

Hsa_circ_0000073 induced proliferation, migration and invasion of OS cells in vitro
To investigate whether hsa_circ_0000073 could be involved in OS progression, two shRNAs and an overexpression plasmid of hsa_circ_0000073 were constructed ( Figure 2A). Meanwhile, the transfection effects of the two shRNAs were verified in U2OS and MG-63 cells, and shRNA01 was selected for further experiments (Supplementary Figure 2A, 2B). Our results first showed that hsa_circ_0000073 was observably upregulated in the overexpression group, and hsa_circ_0000073 was significantly downregulated in the shRNA transfection group (P<0.05, Figure 2B). Next, the results from Edu staining revealed that knockdown of hsa_circ_0000073 notably reduced the proliferation abilities of U2OS and MG-63 cells, and overexpression of hsa_circ_0000073 resulted in a proliferation effect that was opposite to that of the hsa_circ_0000073 knockdown (P<0.05, Figure 2C and 2D). We also revealed that OS cell proliferation results revealed by the CCK-8 assay were similar to those of the Edu staining assay (P<0.05, Figure 2C and 2D). Moreover, our results certified that increased migration and invasion abilities were observed in hsa_circ_0000073-overexpressing U2OS and MG-63 cells, while sharp reductions in migration and invasion were discovered in hsa_circ_0000073-silenced U2OS and MG-63 cells (P<0.05, Figure 2E and 2F). Taken together, our data demonstrated that hsa_circ_0000073 can significantly promote the proliferation, migration and invasion of OS cells.

Hsa_circ_0000073 accelerated MTX resistance and proliferation of MTX-resistant OS cells in vitro
To investigate the impacts of hsa_circ_0000073 on the resistance and cytotoxicity of MTX-resistant OS cells, the IC50 of MTX was evaluated by CCK-8 assay. We discovered that the IC50 of MTX was dramatically increased in MG-63/MTX and U2OS/MTX cells relative to MG-63 and U2OS cells, suggesting that the AGING AGING inhibition rates of MTX on MG-63 and U2OS cells were significantly enhanced compared with those of MG-63/MTX and U2OS/MTX cells (P<0.05, Figure  3A). Meanwhile, we revealed that hsa_circ_0000073 expression was also markedly elevated in MTXresistant MG-63 and U2OS cells with respect to MG-63 and U2OS cells (P<0.05, Figure 3B). Additionally, hsa_circ_0000073 expression was higher in OS patients with an inadequate response to MTX than it was in those with a good response (P<0.05, Figure  3C). Moreover, our results revealed that overexpression of has_circ_0000073 prominently raised the IC50 of MTX in MG-63 and U2OS cells (P<0.05, Figure 3D), and knockdown of hsa_circ_0000073 notably reduced the IC50 of MTX in MG-63/MTX and U2OS/MTX cells (P<0.05, Figure   3E). The results of the colony formation assay also showed that knockdown of hsa_circ_0000073 significantly decreased the proliferation of MG-63/MTX and U2OS/MTX cells (P<0.05, Figure 3F). In summary, our results showed that MG-63/MTX and U2OS/MTX cells were resistant to MTX, and hsa_circ_0000073 could further facilitate MTX resistance in MTX-resistant OS cells.

Knockdown of NRAS notably suppressed hsa_circ_0000073-mediated proliferation, migration and invasion of OS cells
Subsequently, the results of western blot analysis found that NRAS was observably upregulated in OS tissues when compared to paired para-tumor tissues ( Figure 4A). AGING Meanwhile, we revealed that NRAS was memorably increased in OS cells with respect to hFOB 1.19 cells ( Figure 4B). There was a good correlation between the expression of NRAS and hsa_circ_0000073 (r=0.4219, P=0.0357, Figure 4C). In cell-based experiments, we first revealed that hsa_circ_0000073 knockdown significantly downregulated the level of NRAS; additionally, hsa_circ_0000073 overexpression dramatically upregulated the level of NRAS (P<0.05, Figure 4D and 4E). Functional experiments further showed that silencing NRAS prominently suppressed the proliferation of MG-63 and U2OS cells, which was induced by hsa_circ_0000073 overexpression (P<0.05, Figure 4F and 4G). Simultaneously, Transwell assays showed that silencing NRAS could also significantly reduce the migration and invasion of MG-63 and U2OS cells, which were promoted by hsa_circ_0000073 overexpression (P<0.05, Figure 4H and 4I). Consequently, we concluded that NRAS was closely related to the regulation of hsa_circ_0000073 on OS functions.

Knockdown of hsa_circ_0000073 prevented tumor growth of OS by regulating the miR-145-5p or miR-151-3p/NRAS axis in vivo
Furthermore, we also validated the potential significance of hsa_circ_0000073 knockdown on OS tumor growth in vivo, using a xenograft model. Nude mice were subcutaneously injected with MG-63 cells transfected with hsa_circ_0000073 shRNA or a NC shRNA. Throughout the growth of the tumor, we observed that the tumors grew more slowly in the hsa_circ_0000073 knockdown group than in the NC shRNA (control) group (P<0.05, Figure 7A). Meanwhile, the results of IHC and western blot assays revealed that knockdown of hsa_circ_0000073 markedly reduced the expression of NRAS in OS tumors (P<0.05, Figure 7B). In addition, qRT-PCR analysis revealed that knockdown of hsa_circ_0000073 prominently upregulated miR-145-5p and miR-151-3p in OS tumors (P<0.05, Figure 7C). Overall, our study proved that hsa_circ_0000073 could accelerate the proliferation, migration, invasion and MTX resistance of OS by regulating NRAS mediation by miR-145-5p or miR-151-3p ( Figure 7D).

DISCUSSION
Currently, treatment for OS is a combination of local surgery and adjuvant chemotherapy [15]. The development of adjuvant chemotherapy not only can enhance the prognosis of patients with OS but also can improve the rate of limb salvage [16]. MTX is an antitumor drug with a structure similar to folic acid, and it is widely used in a variety of tumor chemotherapies and immunosuppressive therapies [17]. Studies have revealed that MTX can dramatically improve the therapeutic effect of OS treatments [18,19]. However, chemotherapy resistance notably affects the outcome of chemotherapy on the tumor, especially in patients with highly malignant OS. Therefore, a better understanding of the MTX resistance mechanism is of great importance for OS treatment. In our study, we used a GEO database (GSE96964), and we identified as an OS-related circular RNA hsa_circ_0000073 (hsa_circ_001069), which is located in the host gene (OMA1) and is found on chr1:58992932-59002413. We also discovered that the identified hsa_circ_0000073 was markedly upregulated in OS cells and tissues, and  AGING high expression of hsa_circ_0000073 resulted in poor OS survival. Functionally, we revealed for the first time that hsa_circ_0000073 could accelerate the proliferation, migration, invasion and MTX resistance of OS cells.
MicroRNAs, as a class of noncoding small RNA molecules, can regulate gene expression at the posttranscriptional level by binding to the 3' UTR of target mRNAs [20,21]. Currently, it is generally believed that miRNAs are not only connected with the physiological processes of proliferation, apoptosis, proliferation, and embryonic development but can also be involved in cancer progression by acting as a carcinogen or tumor suppressor gene [21,22]. In our study, to ascertain whether hsa_circ_0000073 participates in the ceRNA model, miR-145-5p and miR-151-3p were predicted and were shown to bind directly to hsa_circ_0000073. Meanwhile, miR-145-5p and miR-151-3p were negatively correlated with hsa_circ_0000073. Functionally, we also revealed that the inhibition of hsa_circ_0000073 on the proliferation, migration and invasion of OS cells occurred via downregulation of miR-145-5p and miR-151-3p. In previous studies, miR-145-5p and miR-151-3p have also been proven to prevent tumorigenesis [23][24][25]. These solid pieces of evidence support the idea that the circRNA/miRNA axis plays important roles in OS progression.
Previous studies have also shown that circRNAs can suppress mRNA expression by acting as miRNA sponges [26]. As previous studies have reported, miR-145-5p could repress the progression of bladder cancer by regulating SOX11 [27]; further, miR-145-5p could suppress the progression of laryngeal squamous cell carcinoma via FSCN1 [23]. Research has also certified that miR-151-3p could prevent the migration of breast cancer by targeting TWIST1 [25]. In our study, we found that NRAS was a direct regulatory target of miR-145-5p and miR-151-3p. NRAS could be upregulated by hsa_circ_0000073 by the inhibition of miR-145-5p and miR-151-3p in OS cells. Functionally, we also demonstrated that knockdown of NRAS also inhibited the proliferation, migration and invasion of OS cells, which were mediated by hsa_circ_0000073.

CONCLUSION
Our study suggested that hsa_circ_0000073 could facilitate the proliferation, migration and invasion of OS cells by directing miR-145-5p or miR-151-3p regulation of NRAS. We also speculated that the hsa_circ_0000073/miR-145-5p and miR-151-3p/NRAS axes were closely connected with MTX resistance in OS. Therefore, hsa_circ_0000073, miR-145-5p, miR-151-3p and NRAS might be possible markers in OS, providing potential targets for the therapy of clinical OS patients.

Clinical samples
Twenty-five pairs of OS and paired paratumor tissues were harvested at the Affiliated Huai'an Hospital of Xuzhou Medical University and the Second People's Hospital of Huai'an. OS was confirmed in patients through pathological examination. All 25 patients received initial treatment and did not receive any antitumor therapy. We also obtained informed consent from all patients. Our research has been approved by the Ethics Committee of the Affiliated Huai'an Hospital of Xuzhou Medical University and the Second People's Hospital of Huai'an. All tissues were extracted, labeled and immediately stored at -80 °C.

Identification of differentially expressed circRNAs
Data on circRNAs were collected from the websites of GEO (http://www.ncbi.nlm.nih.gov/geo/). GEO2R was adopted to obtain the top 250 differentially expressed circRNAs in OS cells, hFOB 1.19 and OS MTXresistant cells.  Table 1.

Western blot analysis
Total protein was extracted using RIPA lysis buffer (Beyotime; Cat. no. P0013B) with protease inhibitor, and the concentration was determined using a BCA method (Cwbio, Beijing, China; Cat. no. CW0014). Forty micrograms of protein from each group was separated using 10% SDS-PAGE (Solarbio, China), and the proteins were transferred to PVDF membranes (Millipore; cat. no. ISEQ00010). After blocking the membranes for 2 hr, primary antibodies were incubated with the membranes overnight at 4 °C. After treatment with a secondary antibody (Abcam) for 2 hr, an ECL substrate kit (Thermo Scientific) was utilized to examine the results.

Ago2-RNA immunoprecipitation (RIP) assay
RIP assays were carried out according to the manufacturer's instructions using a Magna RIP™ RNAbinding protein immunoprecipitation kit (Millipore); HEK293 cells were used after transfection with miR-145-5p mimics, miR-151-3p mimics or NC-mimics for 48 hr. Briefly, cells were lysed and incubated with RIP buffer including magnetic beads and anti-Argonaute2 (AGO2) antibody (Millipore) or IgG (Millipore). After incubation with Proteinase K, the immunoprecipitated RNA was extracted. The results were analyzed by qRT-PCR assay.

Statistical analysis
All measurement data are displayed as the mean ± SD. The experimental data were statistically analyzed using 20.0 SPSS software (PSS, Inc., Chicago, USA). The results were calculated via t-test or one-way analysis of variance. P <0.05 indicated that the results were statistically significant.

CONFLICTS OF INTEREST
The authors declare no conflicts of interest.

FUNDING
This study was funded by the Youth Science and Technology Project of Suzhou (grant no. KJXW2018023), the National Natural Science AGING SUPPLEMENTARY MATERIALS

Edu staining
The treated OS cells (1×10 4 cells/well) during logarithmic growth were inoculated into 24-well plates and cultured overnight at 37 °C. Cells were added with 500 μL EdU solution (50 μM; Thermo Fisher Scientific) for 2 hrs at 37 °C. After washing, cells were fixed using 500 μL 4% paraformaldehyde solution (Sigma Aldrich; cat. no. 158127-100G) for 30 mins, and treated with 500 μL 2 mg/mL glycine for 5 mins. After washing, the cells were treated with 500 μL 0.5% TritonX-100 (Sigma-Aldrich; cat. no. T8787) for 10 mins. After treatment with 1 × Apollo staining solution for 30 mins in dark, cells were processed with the penetrating agent and washed using methanol and PBS, respectively. Finally, the DNA staining was performed using 1×Hoechst 33342 (Sigma). The results were obtained using a fluorescence microscopy.

Transwell assay
Cell invasion and migration were monitored using Transwell chambers (BD Pharmingen) with a pore diameter of 8 μmol/L. For cell migration, the U2OS and MG-63 cells in each group were suspended using serum-free medium. The suspended cells (200 μL, 2×10 5 cells/well) were administered to the upper of the Transwell chamber, the complete medium (600 μL) was placed into the lower chamber. After 24 hrs of culture, cells were fixed with 4% paraformaldehyde (Merckmillipore) and dyed with 4% crystal violet (Solarbio, China). The migrated cells were confirmed using a microscope. Before the cell invasion experiment, the cells were humidified with Matrigel, and the other steps were the same as the migration experiment.

Colony formation assay
The transfected U2OS and MG-63 cells (300 cells/well) were seeded into 6-well plates and cultured for 10 days under standard conditions. Cells were washed in PBS, fixed in 4% paraformaldehyde (Merck-millipore) for 15 mins and dyed with 5% crystal violet (Solarbio, China) for 3 mins.