Circusp34 Promotes Osteosarcoma Proliferation and Invasion by Sponging Mir-16-5p: An in Vivo and in Vitro Analysis

Osteosarcoma (OS) is a primary and highly malignant mesenchymal tissue tumor. Although it has achieved a signicant breakthrough in therapeutic administration, the specic pathological mechanism remains unclear. CircRNAs are a type of covalently circular RNA with a head-to-tail junction site. Several studies have reported that abnormal expression of circRNAs is closely associated with tumorigenesis and progression. CircRNAs promote or inhibit the progression of OS by sponging miRNAs. In this study, we aimed to investigate the endogenous competition between circRNAs and miRNAs in OS. CircUSP34 expression was detected by qRT-PCR in KHOS and 143B cells. To conrm the effect of circRNAs and miRNAs, function assays including CCK8, EdU, wound healing, transwell, and western blot were performed in OS cells. FISH and dual-luciferase reporter assays were performed to conrm the sponging mechanism of circUSP34 and miR-16-5p. Additionally, qRT-PCR, immunohistochemistry, HE staining, and western blot were used to conrm the expression of proteins and genes in tumor tissues from mice.


Results
In the present study, we found that circUSP34 promoted OS proliferation, migration, and invasion in vitro and in vivo. circUSP34 and miR-16-5p were upregulated and downregulated, respectively, in OS cells. OS cells' biological behaviors, such as proliferation, migration, and invasion, were inhibited after circUSP34 knockdown. Western blot results also showed that Vimentin and Ki-67 expression levels were decreased.
Similarly, miR-16-5p mimic compromised OS cell proliferation and migration. FISH assay indicated that circUSP34 and miR-16-5p were colocalized in the cytoplasm. The dual-luciferase reporter assay showed that miR-16-5p was sponged by circUSP34. Interestingly, the miR-16-5p mimic partly reversed the inhibitory effect of sh-circUSP34 on the malignancy of OS cells. Finally, animal experiment results about IHC indicated that Vimentin, N-cadherin, and Ki-67 protein expression decreased, but E-cadherin protein expression increased.

Conclusion
Collectively, circUSP34 promoted OS proliferation, migration, and invasion by sponging miR-16-5p. It can serve as a potential therapeutic target and biomarker.
Background Page 3/19 Osteosarcoma (OS) is a primary mesenchymal malignant tumor that commonly occurs in adolescents and young adults [1][2][3][4][5][6]. Present studies show that a combination of surgery and chemotherapy signi cantly improves patient prognosis [5]. However, due to its high malignancy and frequent distant metastases, most patients with advanced-stage disease suffer from a terrible prognosis [6]. The initial cause of osteosarcoma remains unclear: its diagnosis and early treatment are thus hindered. Therefore, investigating its pathophysiologic mechanisms and exploring novel therapeutic targets are urgently needed.
Circular RNA(circRNA) is a closed circular RNA formed by back-splicing pre-mRNA and is usually over 200 nucleotides long [7,8]. Increasing studies have reported that abnormally expressed circRNAs are associated with the clinical characteristics of tumorigenesis, proliferation, metastasis, and invasion [9,10]. It is often aberrantly expressed in a variety of tumors, such as non-small cell lung cancer, colorectal cancer, and hepatocellular cancer [11][12][13]. Some circRNAs have oncogenic function, whereas others have anti-tumoral effects [14]. For example, circ_0000285 promotes the proliferation, migration, invasion, but inhibits apoptosis of osteosarcoma by sponging miR-409-3p [15]. Upregulated circ_0028171 in cells competes endogenously with miR-218-5p to promote osteosarcoma progression [16]. Based on the above studies, we speculate that circUSP34 plays an oncogenic role in OS by sponging miR-16-5p.

Materials And Methods
Cell culture Normal human umbilical vein endothelial cells (HUVEC) and OS cells (KHOS and 143B cell lines) were purchased from the American Type Culture Collection (ATCC, USA). All of them were stored in CELLSAVING solution (New Cell & Molecular Biotech, Suzhou, China) at -80 °C. KHOS cells were cultured in RPMI-1640 (Gibco, Grand Island, NY, USA) and 143B cells were cultured in DMEM (Gibco), and both were supplemented with 10% fetal bovine serum (Gibco), and 1% penicillin and streptomycin (Gibco).
Both cell lines were cultured in a humidi ed incubator with 5% CO 2 at 37 °C.

RNA extraction and quantitative RT-PCR (qRT-PCR)
Total RNA was isolated from OS cells according to the manufacturer's protocol using RNeasy Plus Mini (Qiagen, Germany) and dissolved in 15 μL of RNase free water. The concentration and puri cation of RNA were assessed using a NanoPhotometer Pearl (IMPLEN, Germany) by measuring the A260/280 absorbance. Complementary DNA (cDNA) was synthesized using PrimeScript RT Master Mix (TaKaRa Biotechnology, Japan). CircUSP34 and miR-16-5p were quanti ed using iTaq™ Universal SYBR® Green (Bio-Rad, Hercules, CA, USA) with GAPDH and U6 as internal references, respectively. The sequences of primers are listed in Table 1. Wound-healing assay KHOS and 143B cells transfected with sh-circUSP34 and sh-NC were seeded in 6-plate plates at 8 × 10 5 cells per well. When the cell con uence became approximately 80%, a wound line was scratched using a P-200 pipette tip. Next, images were photographed using a microscope (Leica, Germany) at 0 h and after 24 h. The distance between the injury lines was measured and analyzed.
Transwell assay Brie y, 200 μL serum-free medium containing 5×10 5 cells was added to the upper chamber (Corning, NY, USA) in triplicate. The bottom chamber was lled with 600 μL of medium with 20% FBS. Next, all chambers were incubated at 37 °C for 24 h. The next day, the chambers were xed with 4% paraformaldehyde and stained with 0.1% crystal violet. Cells in the upper membrane of the chamber were slightly wiped using a swab. Finally, the migrating cells in ve different areas were counted under a microscope.

Protein extraction and western blot
The cells were lysed with radioimmunoprecipitation assay lysis buffer (Beyotime, Shanghai, China).

Results
The character and expression level of circUSP34 in OS cells To better understand the features of circUSP34, we depicted its location on the chromosome. Data from circBank (http://www.circbank.cn/index.html) and UCSC (http://genome.ucsc.edu/) showed that circUSP34, which is derived from chr.2p14, exhibited a circular structure via post-transcriptional head-totail splicing (Fig.1a). It has been reported that circRNA, including circUSP34 which is relatively stable, was resistant to RNase R digestion [17]. To verify this, RNase R treatment was initially performed. Subsequently, we designed divergent and convergent primers to verify the differentiation of circUSP34 with and without RNase R treatment using qRT-PCR. The results showed that circUSP34 was not digested, but USP34 mRNA was (Fig. 1b). Agarose gel electrophoresis was also performed to examine the product of qRT-PCR and con rmed that circUSP34 was resistant to RNase R digestion (Fig. 1c), as opposed to USP34 mRNA. qRT-PCR revealed circUSP34 had upregulated expression in both 143B and KHOS cells (Fig. 1d). Therefore, we veri ed the circular structure and stability of circUSP34. Meanwhile it was upregulated in OS cells CircUSP34 knockdown inhibits OS cell malignancy, especially proliferation and migration The above observations prompted us to further investigate the potential role of circUSP34 in OS cells.
Loss-of-function assays were also performed. Then, 143B and KHOS cells were transfected with sh-circUSP34 to knock down circUSP34 expression, which was con rmed by qRT-PCR. The results showed that its expression level signi cantly decreased in both cell lines (Fig. 2a). To determine the role of sh-circUSP34 in proliferation viability in vitro, both CCK8 and EdU assays were performed. All results revealed that interfering circUSP34 compromised the proliferation ability of 143B and KHOS cells (Fig. 2b  and c). Moreover, the colony formation assay also indicated that silencing circUSP34 suppressed the proliferation potential of OS cells, which was consistent with the above results (Fig. 2d). Overall, these results suggested that silencing circUSP34 inhibited the proliferation of OS cells.
Transwell migration with invasion and wound healing assays were carried out to further explore the effect of sh-circUSP34 on OS cell migration or aggression. A wound-healing assay showed that migration capacity was inhibited in 143B and KHOS cell lines transfected with sh-circUSP34 compared to the NC group (Fig. 2e). Transwell migration and invasion assays demonstrated that migration and invasion were inhibited (Fig. 2f). Therefore, these results demonstrate that circUSP34 knockdown suppressed both migration and invasion.
Furthermore, western blot results indicated that Vimentin expression decreased after circUSP34 knockdown. Meanwhile, Ki-67 protein expression levels were also downregulated (Fig. 2g). Taken together, our data indicate that silencing circUSP34 can inhibit the progression of OS in vitro.
CircUSP34 sponges miR-16-5p to regulate OS cell malignancy Given the interaction of both, rst, the target site was predicted by Starbase (http://starbase.sysu.edu.cn/starbase2) as shown in Fig. 3a. Second, we performed a FISH assay to colocalize circUSP34 (green) and miR-16-5p (red) in OS cells: both co-localized in the cytoplasm (Fig. 3b). Third, miR-16-5p relative expression level was detected by qRT-PCR in OS cells transfected with sh-circUSP34, which obviously increased (Fig. 3c). Finally, a dual-luciferase reporter assay was performed on the OS cells. The results showed that the luciferase activity of the circUSP34-WT group with miR-16-5p mimic group obviously decreased compared with miR-16-5p mimic NC group (Fig. 3d). As evidenced by the above results, we hypothesized that miR-16-5p was sponged by circUSP34 to promote OS.

MiR-16-5p inhibits malignancy of OS in vitro
We next explored miR-16-5p expression in KHOS and 143B cells. qRT-PCR data showed that miR-16-5p expression was signi cantly downregulated in OS cells compared with normal HUVECs (Fig. 4a). To con rm the possible anti-tumor effect of miR-16-5p, we designed a miRNA mimic to elevate miR-16-5p expression levels. Subsequently, the inhibitory effect of miR-16-5p mimic on proliferation, migration, and invasion was studied using several assays. CCK-8 assay results showed that miR-16-5p suppressed the proliferation of KHOS and 143B cells (Fig. 4b).
To perform double veri cation, an EdU assay was adopted, which exhibited increased inhibition in OS cells in vitro (Fig. 4c). Similarly, the proliferation ability was evaluated in vitro using a colony-forming assay (Fig. 4f). The miR-16-5p mimic comprised the migration ability of KHOS and 143B cells, as con rmed by the wound healing assay (Fig. 4d). Transwell assay results revealed that the migration and invasion abilities were inhibited by upregulated miR-16-5p (Fig. 4e). Our results show evidence that miR-16-5p plays a suppressive role in malignancy progression in vitro.

MiR-16-5p inhibitor reversed the anti-tumor function of sh-circUSP34
To clarify the speci c relationship between miR-16-5p and circUSP34, OS cells treated with sh-circUSP34 were transfected with miR-16-5p inhibitor. The e cacy of transfection was quanti ed using qRT-PCR. The results showed that miR-16-5p expression level decreased, but that of circUSP34 increased (Fig. 5a). EdU and CCK8 results indicated that the miR-16-5p inhibitor rescued the inhibitory effect of sh-circUSP34 on OS cell viability (Fig. 5b and c). The results showed that miR-16-5p activates the promotion effect of circUSP34 on proliferation. Wound healing assay results indicated that the inhibitory function of sh-circUSP34 was rescued by the miR-16-5p inhibitor (Fig. 5d). Collectively, these results showed that circUSP34 served as a sponge for miR-16-5p to regulate OS malignancy.

Silencing circUSP34 inhibits proliferation and metastasis of OS cells in vivo
To con rm the oncogene function of circUSP34 in vivo, we constructed six BALB/c nude mice models. Every week, the tumor volumes in both groups were measured, showing that tumor weight and volume in the sh-circUSP34 group was inhibited signi cantly comparing with the sh-NC group (Fig. 6a-d). We performed qRT-PCR to detect miR-16-5p expression levels in tumor tissues. The results showed an increased expression level in the sh-circUSP34 group (Fig. 6e). Subsequently, the H&E staining results indicated that it was consistent with the pathological characteristics of OS (Fig. 6f). Additionally, immunocytochemistry results showed that Vimentin, N-cadherin, and Ki-67 protein expression decreased, but E-cadherin protein expression increased (Fig. 6g). Simultaneously, western blot results also indicated that OS malignancy was markedly compromised. Vimentin, N-cadherin, and Ki-67 expression levels were elevated signi cantly; however, E-cadherin expression levels increased (Fig. 6h). These results suggest that circUSP34 may act as an oncogene to promote OS development by sponging miR-16-5p.

Discussion
Although osteosarcoma is a relatively rare malignant tumor derived from mesenchymal tissue, its highgrade malignancy, especially distant metastasis, cannot be neglected. The combination of surgery and chemotherapy markedly improves patients' prognosis [18]. However, local invasion, metastasis, and drug resistance are still to be improved [18,19]. Exploring an effective therapeutic target is an urgent and critical issue. An increasing number of studies has shown that circRNAs are aberrantly expressed in a variety of tumors involving OS [20,21]. For instance, high expression of circRNA can promote the invasion ability of gastrointestinal cancer, lung cancer, and liver cancer [22][23][24]. Thus, we predicted circUSP34 by Starbase as an upstream "sponge" of miR-16-5p, which was reported to inhibit OS [25]. In the present study, the expression level of circUSP34 was con rmed by qRT-PCR, which showed high expression compared to the normal control. Epithelial-mesenchymal transition (EMT) is considered to play a signi cant role in malignancy progression, such as metastasis, in OS. To con rm the effect of circUSP34 on EMT-related proteins of OS, vimentin, E-cadherin, N-cadherin protein expression was quanti ed by western blot. Loss-of-function assays verifying the oncogene circUSP34 in KHOS and 143B cells showed that knocking out circUSP34 compromised the malignant biological behavior of OS cells in vitro. Low expression of circUSP34 inhibits tumor growth in vivo by establishing subcutaneous tumors in mice.
It is widely accepted that miRNA function is modulated by circRNA sponging in various pathological and pathophysiological activities [26,27]. Circ_NOTCH3 competes with miR-205-5p to promote the progression of basal-like breast carcinoma [28]. Circular RNA MYLK, a competing endogenous RNA, promotes bladder cancer progression [29]. Moreover, miR-622 is sponged by circ_0119872 to promote uveal melanoma development [30]. It is worth noting that not every circRNA can sponge miRNAs to modulate their biological functions, as their activity has several requirements, such as a consistent subcellular location. Although miR-16-5p has been previously reported to impede tumor growth [25], in our study, we con rmed its inhibitory function. Low expression level was quanti ed by performing qRT-PCR, but it signi cantly increased after circUSP34 knockdown. This result further supported the hypothesis of interaction between circUSP34 and miR-16-5p. Similarly, gain-of-function assays results showed that overexpression of miR-16-5p impeded OS development. Therefore, it is increasingly possible credible for circUSP34 to sponge miR-16-5p. To better understand circUSP34 as a sponge of miR-16-5p, a rescue experiment was performed. Most importantly, Zhang et al reported that Smad3 acts as a target of miR-16-5p in chordoma cells [31]. Therefore, we speculate that miR-16-5p inhibits OS malignancy by targeting Smad3. Further studies are needed to explore the underlying mechanism.
In conclusion, our results revealed that circUSP34 promotes OS proliferation, migration, and invasion rather than apoptosis by sponging miR-16-5p. However, it was negatively correlated with tumor size in vivo. Moreover, circUSP34 could serve as a potential therapeutic target and a predictable biomarker.

Conclusion
In summary, our ndings revealed that highly regulated circUSP34 plays a role in oncogenes and underlying biomarkers in OS. Furthermore, circUSP34 may be a potential therapeutic target for OS. We rst demonstrated that circUSP34 promotes OS proliferation, migration, and invasion by sponging miR-16-5p. However, our study needed further exploration of the mechanism of circUSP34 and its clinicopathological correlation with OS patients.  The character and expression level of circUSP34 in OS cells. a. We depicted the head-to-tail back-splicing form by UCSC and circBank. b and c. The stability and existence of circUSP34 in 143B cell were detected by qRT-PCR and nucleic acid electrophoresis. d. CircUSP34 expression level in OS cells was con rmed by qRT-PCR. Data were showed as mean ± SD. ns indicated no signi cance, *P < 0.05, **P < 0.01.

Figure 2
CircUSP34 knockdown inhibits OS cell malignancy, especially proliferation and migration. a. Scenes of both KHOS and 143B cells transfected by sh-circUSP34 and sh-NC vector were photographed. And knockout e cacy was quanti ed by qRT-PCR. b,c and d. EdU,CCK8 and colony formation assays results showed that proliferation was inhibited in KHOS and 143B cells transfected by sh-circUSP34 or sh-NC. e and f. OS cells' migration and invasion ability were evaluated by wound healing as well as transwell assays. g. Vimentin and Ki-67 protein expression was assessed by western blot. Data were showed as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001. CircUSP34 sponges miR-16-5p to regulate OS cell malignancy. a. The possible target site between circUSP34 and miR-16-5p was predicted by Starbase. b. The cellular location of circUSP34 (green) and miR-16-5p (red) in cells was observed by performing FISH (magni cation, 400×, scale bar, 100 μm). c. miR-16-5p expression level was con rmed by qRT-PCR after transfecting sh-circUSP34. d. The relative luciferase activities were detected in 293T cells after co-transfection with circUSP34-WT or circUSP34-suppressed in vitro. f. Colony formation results showed that proliferation ability was suppressed. Data were showed as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001. Figure 5 miR-16-5p inhibitor reversed the anti-tumor function of sh-circUSP34. a. The expression level miR-16-5p and circUSP34 with inhibitor and sh-circUSP34 was con rmed by qRT-PCR respectively. b and c. CCK8 and EdU assays showed that the inhibitory function of sh-circUSP34 was rescued by miR-16-5p inhibitor.

Figure 6
Sliencing circUSP34 inhibits proliferation and metastasis of OS cells in vivo. a and b. Subcutaneous tumors of circUSP34-knockout and NC. c. Tumor weight was measured in sh-circUSP34 and sh-NC group which was analyzed. d. Tumor volume was measured every week. e. Relative expression level of miR-16-