Circular RNA circFOXO3 regulates KDM2A by targeting miR‐214 to promote tumor growth and metastasis in oral squamous cell carcinoma

Abstract Oral squamous cell carcinoma (OSCC) is a pathological type of oral cancer, which accounts for over 90% of oral cancers. It has been widely shown that circRNA is involved in the regulation of multiple malignant oral diseases including OSCC. However, the mechanism underlying how circRNA regulates OSCC is still not clearly elucidated. In this article, we report circFOXO3 promotes tumor growth and invasion of OSCC by targeting miR‐214 which specifically degrades the lysine demethylase 2A (KDM2A). CircRNA sequencing was conducted in OSCC tumor and tumor‐side tissues, and the expression of circFOXO3 is found to be markedly increased in tumor tissues. CircFOXO3 is also highly expressed in several OSCC cell lines compared with human oral keratinocytes. Transwell assay and colony formation showed that knockdown of circFOXO3 prevents the invasion and proliferation of oral cancer cells. Via bioinformatic research, miR‐214 was found to be the target of circFOXO3 and correlate well with circFOXO3 both in vitro and in vivo. KDM2A was then validated by database analysis and luciferase assay to be the direct target of miR‐214. KDM2A helps to promote tumor invasiveness and proliferation of OSCC. Collectively, our results proved that circFOXO3 sponges miR‐214 to up‐regulate the expression of KDM2A, thus promotes tumor progression in OSCC.

MicroRNAs (miRNAs) were found to play crucial roles in major cancers including OSCC. In addition, RNA sequencing (RNAseq) has identified multiple families of non-coding RNAs, such as long intergenic non-coding RNAs (lncRNA) and circular RNAs (circRNAs). [4][5][6] Compared with linear RNAs, circRNAs are more stable due to the non-canonical splicing without a free 3' or 5' end. 6,7 CircRNA can function as miRNA sponges, whose sequences can competitively bind miRNAs to regulate the expression of target genes. 6 Our investigations focused on circular RNA circFOXO3 (also termed as Hsa_ circ_0006404), which is markedly increased in OSCC tumor tissues and has been reported to be related to tumor progression. 8,9 The lysine demethylase KDM2A, also known as JHDM1A or FBXL11, was the first JmjC-domain containing demethylase to be identified. 10,11 KDM2A demethylates mono-and di-methylated lysine K36 of histone H3(H3K36me1/2), 12 thus functions as a transcriptional repressor of targeted gene promoters. 13,14 KDM2A has recently been shown to promote breast cancer cell proliferation by activating rRNA genes. 15 In this study, we demonstrated that circFOXO3 up-regulates KDM2A by targeting miR-214 to promote OSCC cell invasion and progression. Our results manifested that regulation of KDM2A by circFOXO3-miR-214 may take part as a novel therapeutic strategy for OSCC treatment.

| Clinical tissues
Thirty-six paired OSCC and adjacent normal tissues were obtained from patients who received surgery in Foshan Stomatology Hospital.
None of them received chemoradiotherapy prior to surgery. The study protocol was approved by the Ethics Committee of Foshan Stomatological Hospital, School of Stomatology and Medicine, with written informed consent obtained from all patients. The relevant clinical information was provided in Table S1.

| Western blotting
Western blot was conducted as previously described. 6 In brief, the cells were harvested with a scraper and then washed once with cold PBS. The cells were then lysed in lysis buffer containing 50 mmol/L Tris-HCl, 250 mmol/L NaCl, 5 mmol/L EDTA, 50 mmol/L NaF, 0.1% NP-40 and 1% protease inhibitor cocktail.
Equal amounts of proteins were size-fractionated by 7.5%-15%

| RNA immunoprecipitation assay
RIP assay was conducted using Magna RIP Kit (EMD Millipore).
Cells were lysed in RIP lysis buffer, and the cell lysate was treated with magnetic beads conjugated to human anti-Ago2 antibody (Millipore) or control antibody (normal mouse IgG; Millipore). qRT-PCR was performed to detect circFOXO3 expression.

| Colony formation assay
Cells were plated and transfected with indicated siRNA/plasmid and then cultured for 48 hours. Cells were trypsinized and plated into 60mm plates with the concentration of 5000 cells/plate. After 2 weeks, methanol fixation and staining with methylene blue was undertaken to identify visible colonies. Plating efficiencies were calculated as follows: number of colonies formed/number of cells plated.

| Reporter vectors constructs and luciferase reporter assay
The fragments of KDM2A (containing predicted binding sites) were cloned into the pMIR-REPORT Vector (Promega) to form the reporter vector KDM2A wild type (KDM2A-wt). The sequence replacing the putative binding site was named KDM2A mutant (KDM2A-mut).
The vector and the miR-214 were cotransfected into HEK 293T cells to test the luciferase activity by Dual-Luciferase Reporter Assay System (Promega).

| Bioinformatics
We used Circinteractome database to predict circFOXO3 potential miRNA. We applied TargetScan to search miRNA-targeted genes.

| Tumorigenesis in nude mice
Four-week-old BALB/c nude mice were purchased from the Experimental Animal Centre of Peking University Health Science Centre and housed in a pathogen-free environment. DLD1 cells (~1 × 10 7 ) were delivered into animals via hypodermic injection.
After 2 weeks, all mice were randomly divided into four groups (n = 6) and administered adenovirus via tail intravenous injection. At the end of the experiment (5 weeks after tumor implantation), the mice were sacrificed and the weight of each tumor was determined.

| Statistical analysis
For all statistical tests, three or more independent experiments were performed, data are shown as means ± SD P < .05, by unpaired Student's t test, was considered statistically significant. Data were analyzed by Graphpad version 8.0 (Graph Pad Software).

| Study approval
This study was approved by the Ethics Committee of Foshan Stomatology Hospital Affiliated to Foshan University, Foshan, Guangdong, China. Informed consent was obtained from each patient before the use of their tumor tissues.

| CircFOXO3 is significantly increased in oral squamous tumors and OSCC cell lines
In order to explore the potential role of circFOXO3 in oral cancer, we firstly detected the expression of circFOXO3 in oral squamous tumor and tumor-side tissue samples from OSCC patients.
As shown in Figure 1A, the expression of circFOXO3 was significantly increased in tumor tissues compared with tumor-side tissues.

| CircFOXO3 is required for the invasiveness and proliferation of OSCC
The increased expression of circFOXO3 in OSCC indicated that circFOXO3 may promote tumor growth. To test this hypothesis, we firstly measured the invasive abilities of two OSCC cell lines specifically treated with circFOXO3 knockdown ( and SCC-9 ( Figure 2E). In accordance with this, colony formation assay also manifested reduced clone numbers upon treatment with circFOXO3 siRNA (Figure 2F-G). Collectively, these results indicated that circFOXO3 might be essential for the progression and invasion of OSCC malignant tumors.

| CircFOXO3 functions as a miRNA sponge for miR-214 in OSCC
Previous studies have reported that circRNAs can function as miRNA sponges. By sharing one or more microRNA response elements, circRNAs can bind to miRNAs and in turn arrest miRNA functions. To examine whether circFOXO3 can function as miRNA sponges, bioinformatics software (RegRNA) was used to predict the potential circRNA/miRNA interactions and found circFOXO3 might interact with several miRNAs ( Figure 3A). We selected miR-214, which was shown to be the top hit, for further analysis. Luciferase In conclusion, these results demonstrated that circFOXO3 act as a sponge for miR-214 and promote OSCC tumor progression and invasion by targeting miR-214.

| CircFOXO3 up-regulates KDM2A expression by targeting miR-214
To further explore the potential target of miR-214, we searched for the putative gene target of miR-214 via bioinformatic analysis.

CO N FLI C T O F I NTE R E S T
The authors have declared that no competing interest exists.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analyzed during this study are included in this article. Further details are available on request. F I G U R E 4 circFOXO3 up-regulates KDM2A expression by targeting miR-214. A, Binding sites between miR-214 and KDM2A was shown. B, Luciferase activity was measured in miR-214 and KDM2A WT, Mut cDNA plasmids infected 293T. Data are presented as means ± SD (n = 3). **P < .01. C, Relative expression of KDM2A in SCC-4 cells with indicated treatment. Data are presented as means ± SD (n = 3). *P < .05, **P < .01. D, KDM2A protein level was measured by western blot of SCC-4 with indicated treatment. E, The effects of KDM2A on both SCC-4 cell invasion. A total of 1 × 10 5 transfected cells were seeded onto a permeable membrane in a Boyden chamber to allow the cells to invade the opposite layer of the membrane. F, SCC-4 cells were transfected as indicated and cultured for different times. Cell viability was measured by CCK-8 assay. Data represent the means ± SD (n = 3). *P < .05, **P < .01. G, Proliferation of SCC-4 cells transfected with indicated siRNA/plasmid, as detected by colony formation assay. H, The correlation of miR-214/KDM2A, circFOXO3/ KDM2A expression level in OSCC tumor tissues analyzed by Pearson. I, Four-week-old nude mice were engrafted with 1 × 10 7 SCC-4 cells and randomly divided into four groups (n = 6). After two weeks, the tumor-bearing mice were treated with adenovirus expressing different siRNA or inhibitor as indicated. Tumor volumes were calculated by measuring the length and width using Vernier calipers every two days. Data represent the means ± SD (n = 6). J, Images of the tumors for Figure 4I