Circ_RPL23A acts as a miR-1233 sponge to suppress the progression of clear cell renal cell carcinoma by promoting ACAT2

Clear cell renal cell carcinoma (ccRCC) is a prevalent urological carcinoma with high metastatic risk. Circular RNAs (circRNAs) have been identified as effective diagnostic and therapeutic biomarkers for ccRCC. This research aims to disclose the effect and regulatory mechanism of circRNA ribosomal protein L23a (circ_RPL23A) in ccRCC. We performed quantitative real-time polymerase chain reaction (qRT-PCR) to examine circ_RPL23A, microRNA-1233 (miR-1233) and acetyl-coenzyme A acetyltransferase 2 (ACAT2). Cell cycle progression, apoptosis, cell viability, invasion and migration, which were respectively conducted by using flow cytometry, 3-(4, 5-dimethylthiazol-2-y1)-2, 5-diphenyl tetrazolium bromide (MTT), transwell assays. The levels of ACAT2 protein and cell cycle proteins, proliferation-associated protein, and epithelial-mesenchymal transition (EMT) associated proteins were measured by western blot. Target relationship was analyzed via dual-luciferase reporter assay and RNA pull down assay. The animal model was used to study how circ_RPL23A affects in vivo. Circ_RPL23A was lower expressed in ccRCC tissues and cells. The elevated circ_RPL23A suppressed cell cycle progression, proliferation, migration and invasion but promoted apoptosis in ccRCC cells. MiR-1233 was a target of circ_RPL23A and direct targeted to ACAT2. Besides, circ_RPL23A exerted its anti-tumor effect by sponging miR-1233, and then relieved the inhibition effect of miR-1233 on ACAT2. Overexpression of circ_RPL23A also curbed ccRCC tumor growth in vivo. Circ_RPL23A inhibited ccRCC progression by upregulating ACAT2 expression by competitively binding miR-1233, which might provide an in-depth cognition for ccRCC pathogenesis and circ_RPL23A might be a promising biomarker in ccRCC diagnosis and treatment.


Introduction
Renal cell carcinoma (RCC) is a dominant form of cancers originated from renal tubular epithelial cells, which ranks third among malignant tumors in the genitourinary system (Siegel et al. 2018). It can be categorized into several subtypes, mainly including clear cell RCC (ccRCC, 75%), papillary RCC (15%), chromophobe RCC (5%) and other types less than 1% (Hsieh et al. 2018;Li et al. 2019). As the most common type of RCC, the incidence and mortality of ccRCC have rapidly increased in the past few decades on account of its high metastasis manner (Siegel et al. 2017). Despite the recent advances in therapeutic approaches including targeted therapy have improved the prognosis of ccRCC (Ljungberg et al. 2011;Barata et al. 2016;Lee et al. 2018), the overall survival and prognosis of ccRCC patients remains poor due to the distant metastasis and local recurrence (Jiang et al. 2016). Hence, identification of novel targets and exploration of the underlying mechanism possess clinical importance for the treatment and diagnosis of ccRCC.
Circular RNAs (circRNAs) are covalent closed non-coding RNAs (ncRNAs) without 5′ cap and 3′ tail in eukaryotes (Qu et al. 2015). Dysregulated circRNAs have been recognized to participate in the regulation of oncogenesis and malignance of Liang Cheng and Huifeng Cao These authors contributed equally to this work. multiple human cancers (Zhao and Shen 2017;Chen and Huang 2018). A newly discovered circRNA, circRNA ribosomal protein L23a (circ_RPL23A), also termed hsa_circ_0092360, is derived from the gene RPL23A and located in chr17: 27047048-27,047,688, was aberrantly downregulated in ccRCC tissues . However, its biological function and underlying mechanism in ccRCC are still unknown.
CircRNAs could serve as "sponges" for microRNAs (miRNAs) to mediate miRNA activity by competing for miRNAs binding, thereby participating in cancer progression (Yin et al. 2019). MiRNAs are a class of small ncRNAs with a length of 22 nucleotides (nt) that complementary bind to the 3′-untranslated regions (3'-UTRs) of target messenger RNAs (mRNAs), thereby affecting corresponding gene expression and regulating multiple normal cellular function and disease pathogenesis in many cancers (Hammond 2015). Interestingly, the circRNA/miRNA/mRNA regulatory network has been found in several cancers, including osteosarcoma (Qiu et al. 2020), hepatocellular carcinoma (Sun et al. 2020), papillary thyroid carcinoma , and ccRCC . MiR-1233, also termed miR-1233-3p, was first reported to act as a RCC-related miRNA in 2011, which was upregulated in serum of RCC patients (Wulfken et al. 2011). Besides, Dias et al. also confirmed that miR-1233 could be as a prognostic biomarker for RCC, as miR-1233 level is closely associated with the risk of specific death by RCC and cancer-specific survival (Dias et al. 2017). Furthermore, miR-1233 was negatively regulated by hsacirc-0007766 and directly targeted growth and differentiation factor 15 (GDF15) to mediate gastric cancer progression ). These researches have disclosed the vital role of miR-1233 in cancer diagnosis and treatment. However, the function and regulatory mechanism of miR-1233 in ccRCC has not been reported.
The reprogramming of cellular metabolism is considered as a vital hallmark of cancer, which is necessary for cancer initiation and progression (Pavlova and Thompson 2016). Acetylcoenzyme A acetyltransferase 2 (ACAT2), belongs to the thiolase family, is an enzyme involved in lipid metabolism (Shintani et al. 1999). Ibrahim et al. indicated that the Mikania micrantha extract could inhibit the activity of enzyme HMG-CoA reductase (HMGCR) and ACAT2 to reduce the lipid peroxidation and accumulation, thereby improving hypercholesterolemia in high cholesterol-fed Rats (Ibrahim et al. 2020). Furthermore, ACAT2 was reported to be inversely correlated with ccRCC stage, tumor size, and cancer-specific survival, which could be used as potential prognostic and therapeutic target for ccRCC (Zhao et al. 2016). However, the underlying mechanism of ACAT2 in ccRCC still needs further investigation.
In the present research, we investigated the expression and function of circ_RPL23A in ccRCC. Besides, we also explored the regulatory mechanism of circ_RPL23A/miR-1233/ACAT2 axis in ccRCC.

Dataset analysis
The gene expression microarray GSE100186 was searched from Gene Expression Omnibus (GEO) datasets and analyzed using GEO2R, a powerful tool to screen the differentially expressed genes (DEGs) between ccRCC and normal samples.

Tissue collection and ethical statement
Sixty pairs of ccRCC tissues and matched para-carcinoma normal tissues were acquired from ccRCC patients who underwent surgical resection at The First Affiliated Hospital of Jiamusi University, and preserved in liquid nitrogen immediately. The patients were diagnosed by pathology examination. Tissue collection was performed according to the Declaration of Helsinki, and obtained written informed consent from all specimens. Besides, our study was approved by the Ethical Committee of The First Affiliated Hospital of Jiamusi University.
To establish the stable 786-O cell line over-expressing circ_RPL23A, circ_RPL23A sequence was synthesized and cloned into pLV-Luci (2A) puro vector (inovogen, Chongqing, China). Then the pLV-circ_RPL23A-Luci (2A) puro vector and pLV-Luci (2A) puro empty vector were transfected into the HEK-293 T cells to generate packaging plasmids. Subsequently, 786-O cells were infected with the lentivirus of pLV-circ_RPL23A-Luci (2A) puro vector or pLV-Luci (2A) puro empty vector for 48 h and then the stable cell line was established after 2 weeks of 4 μg/ml puromycin selection.

Stability analysis
Actinomycin D and Ribonuclease R (RNase R) assays were used to detect the stability of circ_RPL23A. For the Actinomycin D assay, 786-O and 769-P cells were incubated with 2 mg/mL Actinomycin D (Millipore) for 0 h, 8 h, 16 h and 24 h. Total RNA was isolated at the specified point in time, and the expression of circ_RPL23A and its linear transcript RPL23A was analyzed by qRT-PCR.
For RNase R treatment, 4 μg of total RNA was incubated with or without 12 U Ribonuclease R (RNase R; Epicentre Technologies, Madison, WI, USA) at 37°C for 1 h. Then, circ_RPL23A and RPL23A expression levels were determined via qRT-PCR.

Cell cycle detection
Cell cycle distribution was analyzed by using a flow cytometer (BD Biosciences, San Diego, CA, USA) as previously described (Yu et al. 2015;Wang et al. 2017b). In brief, 1 × 10 6 786-O and 769-P cells were harvested and stained with 5 μL propidium iodide (PI; 5 μL; BD Biosciences) and 1 U/mL RNase for 1 h in the dark. Then stained cells were analyzed by a flow cytometer.

Apoptosis detection
Cell apoptosis was evaluated using fluorescein isothiocyanate (FITC) Annexin V Apoptosis Detection Kit (BD Biosciences) through the flow cytometer (BD Biosciences). 1 × 10 6 cells were harvested and stained with 5 μL FITC Annexin V and 5 μL PI for 15 min in the dark. The apoptosis cells (Annexin V+/PI-and Annexin V+/PI+ cells) were analyzed using a flow cytometer, and the apoptosis rate was calculated (the apoptotic cells/total cells × 100%).

Transwell assay
Cell migration and invasion were determined using 24-well transwell chambers (Corning Inc.) uncoated or pre-coated with matrigel (Corning Inc.). 5 × 10 4 transfected 786-O and 769-P cells in serum-free DMEM medium were plated into the upper chamber, while the lower chamber was supplemented with DMEM medium containing 10% FBS. Following a 24 h-incubation, cells that invaded or migrated to the bottom of chambers were fixed with methanol and stained with crystal violet (Sangon). The migratory or invasive cells were photographed and counted using an inverted microscope (Olympus, Tokyo, Japan).

Dual-luciferase reporter assay
Circ_RPL23A or ACTA2 3' UTR sequences contained the wild-type (wt) or mutant-type (mut, CUUUAGA) miR-1233 binding sites were cloned into the pGL3 luciferase basic vector (Promega, Madison, WI, USA) to generate circ_RPL23Awt, circ_RPL23A-mut, ACAT2-wt or ACAT2-mut reporter plasmids. Prior to transfection, 786-O and 769-P cells were cultured in 24-well plates for 24 h at a density of 3 × 10 5 cells/ well. Then the reporter plasmids (to a total amount of 500 ng of pGL3 luciferase vector per well along with 20 ng pRL-TK luciferase vector) and miR-1233 or miR-NC were transfected into 786-O and 769-P cells for 48 h. Then, luciferase activities were measured by dual-luciferase assay system (Promega) according to the manufacture's instruction. The Renilla luciferase activity was used as the internal normalization reference for firefly luciferase activity.

RNA pull down assay
The biotinylated circ_RPL23A-WT or circ_RPL23A-MUT probe was purchased from Geneseed (Guangzhou, China). In brief, 786-O and 769-P cells were harvested and lysed. The C-1 magnetic beads (Life Technologies) were incubated with biotinylated probes at 25°C for 2 h. Then the cell lysates were incubated with biotinylated NC, circ_RPL23A-WT or circ_RPL23A-MUT probe at 4°C overnight, and then washed in wash buffer. The RNA complexes were purified using RNeasy Mini Kit (QIAGEN) for qRT-PCR analysis.

Xenograft tumor assay
A total of 10 BALB/c nude mice (six-week-old, 20-25 g) were bought from Animal Center of Chinese Academy of Medical Sciences (Beijing, China) and fed in a specific pathogen-free (SPF) cage with 70% humidity at 26°C. Food and water were provided ad libitum. Mice were subcutaneously injected with 2 × 10 6 786-O cells stable over-expressing circ_RPL23A or Vector, respectively (Five mice for each group). After injection for 10 days, the length and width of tumors were measured by a digital caliper every 5 days and tumor volume was calculated with the following formula: length × width 2 × 0.5. No mice died during the 30 dobservation period. After euthanizing all of the mice, the tumor tissues dissected from mice were weighed and stored at −80°C for further detection. All animal experiments were conducted in accordance with the National Institutes of Health (NIH) guidelines, and approved by the Institutional and Local Animal Ethics Committee of The First Affiliated Hospital of Jiamusi University.
Statistical analysis SPSS 22.0 was utilized to conduct statistical analysis. All experiments were administrated in three biological replicates and data were displayed as mean ± standard deviation (SD). Spearman's correlation coefficient was used to analyze the correlation between RNAs and mRNA. Difference analysis was analyzed using Student's t test and one-way analysis of variance (ANOVA) followed by Tukey's test. P value <0.05 was considered as a statistically significant difference.

Circ_RPL23A was lower expressed in ccRCC tissues and cells
According to the gene expression data that obtained from GEO (GSE100186) dataset, circ_RPL23A was the most down-regulated circRNA in ccRCC tissues compared with matched non-tumor tissues (Fig. 1a). And it was further confirmed by qRT-PCR analysis, as the expression of circ_RPL23A in ccRCC tissues (n = 60) and cells (786-O: P = 0.0005, 769-P: P = 0.0002, CAKI-1: P = 0.0028 and A498: P = 0.0105) was lower than that in matched paracarcinoma normal tissues and human normal kidney epithelial cellHK-2 cells (P = 0.0050) (Fig. 1b and c). Besides, the linear mRNA level of the host gene RPL23A was also downregulated in ccRCC tissue in contrast with the normal tissues ( Supplementary Fig. 1). Next, we explored the circular characteristics of circ_RPL23A by RNase R digestion and Actinomycin D assays. The data of Actinomycin D assay disclosed that the half-life of circ_RPL23A (more than 24 h) was longer than the linear transcript RPL23A (about 16 h), implying the high stability of circ_RPL23A ( Fig. 1d and e). Moreover, RNase R digestion decreased RPL23A linear mRNA level, but not the level of circ_RPL23A ( Fig. 1f and  g). Thus, circ_RPL23A harbored a loop structure and was lower expressed in ccRCC.

Circ_RPL23A regulated ccRCC progression by mediating cell proliferation, apoptosis, migration and invasion
To investigate the potential role of circ_RPL23A in ccRCC, gain-of-function and loss-of-function experiments were conducted by transfecting circ_RPL23A overexpression vector or si-circ_RPL23A into 786-O and 769-P cells. As shown in Fig. 2a, circ_RPL23A expression elevated by 7-fold while its host linear transcript RPL23A changed little in cells with circ_RPL23A overexpression vector transfection. Flow cytometry assay indicated that elevated expression of circ_RPL23A increased the cell populations at the G0/G1 phase but decreased the cell number in S phase, indicating circ_RPL23A overexpression induced G0/G1 cell cycle arrest in ccRCC cells (Fig. 2b and c). In consistent with the data in flow cytometry, the protein levels of cell cycle proteins CyclinD1 and CDK4 were significantly inhibited by upregulation of circ_RPL23A ( Fig. 2d and e). Besides, the elevated circ_RPL23A also suppressed the proliferation of ccRCC cells, as identified by the decreasing cell viability (Fig. 2F and G), as well as the declined protein levels of Ki67 (a proliferation marker for human tumor cells) and proliferating cell nuclear antigen (PCNA) in circ_RPL23A overexpression group ( Fig. 2h and i). Furthermore, over-expression of circ_RPL23A significantly promoted cell apoptosis (Fig. 2j), but suppressed the migration (Fig. 2k) and invasion (Fig. 2l) abilities in ccRCC cells. Moreover, the protein level of snail (an epithelial-mesenchymal transition (EMT)-promoting marker) decreased while E-cadherin (an EMT-inhibitory marker) protein level upregulated in 786-O and 769-P cells transfected with circ_RPL23A overexpression vector demonstrated that overexpressed circ_RPL23A inhibited EMT in ccRCC cells (Fig. 2m and n). These results uncovered that circ_RPL23A acted as a tumor suppressor in ccRCC cells.  Fig. 2A and B, transfection of si-circ_RPL23A obviously suppressed the expression of circ_RPL23A while it showed no effect on the level of linear RPL23A. Besides, circ_RPL23A knockdown elevated the cell number at S phase cells but decreased the cell number at G0/ G1 phase (Supplementary Fig. 2C and D), and it also increased the protein levels of cell cycle proteins CyclinD1 and CDK4 (Supplementary Fig. 2E and F), indicating that si-circ_RPL23A advanced the cell cycle progression. Besides, circ_RPL23A depletion also increased the cell viability (Supplemental Fig. 2G and H) and the protein levels of cell proliferation marker Ki67 and PCNA (Supplemental Fig. 2I and J), as well as inhibited the apoptosis of cells (Supplemental Fig. 2K). Furthermore, cell migration and invasion abilities were also promoted by the downregulation of circ_RPL23A (Supplemental Fig. 2L-O). Collectively, circ_RPL23A played a suppressive role in ccRCC progression.
Circ_RPL23A acted as a miR-1233 sponge in ccRCC Given that circRNAs function as miRNA sponge to participate in the regulation of tumor initiation and progression, we predicted the target miRNA of circ_RPL23A by Circinteractome (Livak and Schmittgen 2001). The results showed that circ_RPL23A (AGGGCUC) sequence contained the complementary binding sites of miR-1233 (UCCCGAG) (Fig. 3a). Furthermore, dual-luciferase assay uncovered that miR-1233 transfection considerably inhibited the luciferase activity of circ_RPL23A-wt group in contrast with the cell with miR-NC transfection, whereas it had little effect on the luciferase activity of circ_RPL23A-mut group (Fig. 3b and c). Furthermore, RNA pull down assay indicated that miR-1233 was obviously enriched in the pellet pulled down by bio-circ_RPL23A-wt compared with bio-NC and bio-circ_RPL23A-mut ( Supplementary Fig. 3), implying the presence of miR-1233 binding sites in circ_RPL23A. Besides, overexpression of circ_RPL23A decreased the level of miR-1233, suggesting that circ_RPL23A competitively binding miR-1233 to decrease its level in ccRCC cells (Fig. 3d). Furthermore, we found that miR-1233 was upregulated in ccRCC tissues compared with paired normal tissues (Fig. 3e), and a negative correlation (r = −0.6939, P < 0.001) was found between the levels of miR-1233 and circ_RPL23A in 60 cases of ccRCC tissues (Fig. 3f). The miR-1233 level in 786-O and 769-P cells was also higher than that in HK-2 cells (Fig. 3g). Taken together, circ_RPL23A negatively regulated miR-1233 by sponging miR-1233.

Circ_RPL23A inhibited the development of ccRCC by sponging miR-1233
To investigate if circ_RPL23A meditate ccRCC progression by interacting with miR-1233, rescue experiments were conducted by transfecting circ_RPL23A overexpression vector or circ_RPL23A + miR-1233 into the ccRCC cells. The expression of miR-1233 was suppressed by circ_RPL23A overexpression, whereas miR-1233 mimic partly reversed the suppression effect of circ_RPL23A on miR-1233 level (Fig. 4a). Besides, miR-1233 mimic significantly relieved the promotion effect of circ_RPL23A overexpression on cell cycle progression ( Fig. 4b and c), as well as the inhibition effects on the protein levels of CyclinD1 and CDK4 (Fig. 4d and e). Furthermore, the proliferation inhibition ( Fig. 4f-i) and apoptosis promotion (Fig. 4j) in 786-O and 769-P cells induced by elevated circ_RPL23A were partly overturned by the cotransfection of miR-1233 mimic. Moreover, miR-1233 also reversed the suppression effects of circ_RPL23A overexpression on ccRCC cell migration (Fig. 4k), invasion (Fig. 4l) and EMT ( Fig. 4m and n). These data suggested that circ_RPL23A restrained ccRCC progression by sponging miR-1233.

Circ_RPL23A enhanced ACAT2 expression by sponging miR-1233 in ccRCC
Interestingly, miRDB online software predicted that ACAT2 3'UTR sequences harbored the complementary binding sites of miR-1233 (Fig. 5a), and it was further confirmed by dualluciferase reporter assay. As displayed in Fig. 5b and c, miR-1233 overexpression significantly suppressed the luciferase activity of ACAT2-wt group, but not ACAT2-mut group, indicating the combination between ACAT2 and miR-1233. Next, the transfection efficiencies of miR-1233 mimic or inhibitor were evaluated, and the results indicated that miR-1233 level was upregulated nearly 60-fold by miR-1233 mimic, but was depressed more than three-quarters by miR-1233 inhibitor (Fig. 5d). As displayed in Fig. 5e and f, ACAT2 mRNA and protein levels were inhibited by miR-1233 overexpression but were upregulated by miR-1233 inhibitor, implying that ACAT 2 was negatively regulated by miR-1233. Besides, ACAT2 mRNA and protein levels were distinctly decreased in The determination of snail and E-cadherin protein levels was carried out via western blot. Data are reported as mean ± SD, n = 3.*P < 0.05 vs Vector group (Student's t test) ccRCC cells (Fig. 5g and h). Similarly, ACAT2 mRNA and protein levels were downregulated in ccRCC tissues in contrast with the normal tissues, and ACAT2 mRNA was negatively correlated (r = −0.7151, P < 0.001) to miR-1233 expression in 60 cases of ccRCC tissues (Fig. 5i-k). Additionally, circ_RPL23A overexpression elevated ACAT2 mRNA and protein levels, while this promotion effects were overturned by miR-1233 upregulation ( Fig. 5l and m). Therefore, miR-1233 could target ACAT2, and circ_RPL23A promoted the expression of ACAT2 by sponging miR-1233 in ccRCC.

MiR-1233 played a tumor-promoting role by targeting ACAT2 in ccRCC
We next explored the role and functional mechanism of miR-1233 in ccRCC by transfecting anti-miR-1233 or anti-miR-1233 + si-ACAT2 into 786-O and 769-P cells. As shown in Fig. 6a and b, si-ACAT2 transfection overturned the anti-miR-1233-induced promotion effects on ACAT2 mRNA and protein levels. Besides, we found that miR-1233 inhibitor hampered cellcycle conversion from G0/G1 to S phase, but this phenomenon was mitigated by ACAT2 knockdown (Fig. 6c and d). In addition, ACAT2 knockdown also attenuated the suppression effects of anti-miR-1233 on CyclinD1 and CDK4 protein levels ( Fig. 6e  and f). Furthermore, ACAT2 knockdown overturned the suppression effects of miR-1233 inhibitor on cells proliferation (Fig. 6g-j), and the promotion effect on cell apoptosis (Fig. 6k). Moreover, miR-1233 down-regulation exhibited an inhibitory impacts on cell migration (Fig. 6l), invasion (Fig. 6m) and EMT (snail inhibition and E-cadherin promotion in protein expression) (Fig. 6n-o), whereas these effects were partly restored by ACAT2 knockdown. Altogether, anti-miR-1233 retarded ccRCC progression by up-regulating ACAT2, which indicated the promotion role of miR-1233 in ccRCC via targeting ACAT2.

Circ_RPL23A curbed ccRCC tumor growth in vivo
A xenograft model was established to investigate the effect of circ_RPL23A in vivo. In comparison with the control group, the tumor volume (Fig. 7a) and weight (Fig. 7b) of circ_RPL23A overexpression group were significantly decreased. Besides, we also detected the levels of circ_RPL23A, miR-1233 and ACAT2 in xenograft tumor tissues by qRT-PCR and western blot. The results uncovered that circ_RPL23A and ACAT2 levels were elevated, while miR-1233 level was reduced in xenograft tumor tissues with circ_RPL23A overexpression (Fig. 7c-f). Thus, circ_RPL23A acted as a tumor suppressor in ccRCC by regulating miR-1233/ACAT2 axis in vivo.

Discussion
Recent studies have reported the dysregulation of circRNAs in ccRCC and their potential clinical significance (Franz et al. 2019). For instance, hsa_circ_001895 is an overexpressed circRNA in ccRCC, and promoted ccRCC progression by regulating miR-296-5p/sex determining region Y (SRY)- box 12 (SOX12) axis (Chen et al. 2020b). CircRNA hippocampus abundant transcript 1 (circHIAT1) was downregulated and functioned as a metastatic inhibitor and tumor suppressor in ccRCC (Wang et al. 2017a). Wang et al. disclosed that hsa_circ_0001451 was lowly expressed in ccRCC cells and could be used as a promising biomarker for the diagnosis and treatment of ccRCC patients (Wang et al. 2018). In this research, we found a novel downregulated circRNA, circ_RPL23A, in ccRCC tissues according to the gene expression data of GEO (GSE100186) dataset and differential expression analysis in ccRCC tissues and cells. With the closeloop structures, circRNAs are highly conserved and stable compared to the linear RNAs (Cai et al. 2019). We found that circ_RPL23A showed resistance when suffering Actinomycin D and RNase treatment, which indicating the loop structure of circ_RPL23A. The expression levels of ACAT2 in 60 ccRCC and normal tissues was analyzed via qRT-PCR. Data are reported as mean ± SD. *P < 0.05 vs normal tissues. (J) The association between miR-1233 and ACAT2 was conducted by Spearman's correlation coefficient analysis (N = 60, r = −0.7151, P < 0.0001). (K) Western blot assay was conducted to analyze ACAT2 protein level in three ccRCC tissues. Data are reported as mean ± SD. *P < 0.05 vs normal tissues (Student's t test). (L-M) 786-O and 769-P cells were transfected with Vector, circ_RPL23A, circ_RPL23A + miR-NC or circ_RPL23A + miR-1233, separately. And ACAT2 mRNA (L) and protein (M) levels were determined by qRT-PCR and western blot. Data are reported as mean ± SD. *P < 0.05 vs Vector group (ANOVA) (circ_ABCB10) enhanced cell proliferation and reduced cell apoptosis to facilitate ccRCC progression (Huang et al. 2019). Differentially, circRNA threonine-specific protein kinase 3 (circ-AKT3) overexpression was demonstrated to decrease migrated and invaded ccRCC cells (Xue et al. 2019). CircRNA Rap guanine nucleotide exchange factor 5 (cRAPGEF5) (Chen et al. 2020a) and has_circ_0072309 ) also impeded tumor growth and metastasis in ccRCC. In present research, circ_RPL23A overexpression induced cell cycle arrest and apoptosis, but inhibited proliferation, migration and invasion in vitro, whereas circ_RPL23A knockdown promoted ccRCC progression. Consistent with the in vitro results of circ_RPL23A in ccRCC, the in vivo xenotransplanted tumor model uncovered that circ_RPL23A overexpression inhibit tumor growth of ccRCC in vivo, indicating the tumor repressor role of Given that circRNAs function as a "sponge" for miRNAs to mediate miRNA level and stability, thereby regulating ccRCC development, we further explored the downstream target of circ_RPL23A. And we found that miR-1233 is a target of circ_RPL23A and was negatively regulated by circ_RPL23A. A number of researches have disclosed the vital biological functions and clinical implication values of miRNAs in multiple cancers (He et al. 2018). MiR-1233 was significantly downregulated in Aβ-treated human platelets and their precursor megakaryocytes, and mediated the expression of P-selectin and cell adhesion to fibronectin in patients with Alzheimer's pathologic change, which could be used as a pathologic marker for Alzheimer's pathologic change with mild cognitive impairment (Lee et al. 2020). Besides, miR-1233 was reported to have high clinical value in the early diagnosis of acute exacerbation of chronic obstructive pulmonary disease with acute pulmonary embolism (Peng et al. 2020). Moreover, miR-1233 was significantly increased in the renal tissue and serum samples, and exhibited a high sensitivity and specificity for the diagnosis of ccRCC (Yadav et al. 2017). In accordance with previous research, miR-1233 was lower expressed in ccRCC tissues and cells. Besides, miR-1233 knockdown inhibited the progression of ccRCC, and miR-1233 overexpression partly reversed the suppression effects of circ_RPL23A overexpression on ccRCC progression. Thus, circ_RPL23A negatively regulated miR-1233 expression by sponging miR-1233 to mediate the progression of ccRCC.
Furthermore, we found that miR-1233 directly targeted ACAT2 and negatively regulated ACAT2 expression. And circ_RPL23A could upregulate ACAT2 expression by sponging miR-1233 in ccRCC. ACAT2 is an enzyme involved in lipid metabolism, which has been reported to be inversely correlated with the prognosis of ccRCC patients, whereas the underlying mechanism of ACAT2 in ccRCC remains unclear (Zhao et al. 2016). Our research showed that ACAT2 exerted a tumor suppressor role in ccRCC. And ACAT2 knockdown partly reversed the suppression effect of miR-1233 knockdown on ccRCC progression. Thus, circ_RPL23A/miR-1233/ACAT2 axis regulated ccRCC progression in vitro and in vivo.
However, there are still some deficiencies in the current research work. ACAT2 is involved in lipid metabolism, and might participate in the regulation of reprogramming of cellular metabolism. In present research, ACAT2 is lower expressed in ccRCC, implying that ACAT2 may affect reprogramming of cellular metabolism to mediate the initiation and progression of ccRCC. Thus, it is necessary to investigate whether ACAT2 mediates cellular metabolic reprogramming to participate in the regulation of ccRCC progression.
In conclusion, our results indicated that circ_RPL23A functions as a tumor suppressor, which inhibited ccRCC progression by regulating cell cycle progression, proliferation, EMT, migration, invasion, and apoptosis through sponging miR-1233 and upregulating ACAT2 (Fig. 8). We verified for the first time that circ_RPL23A/miR-1233/ACAT2 axis regulated ccRCC progression, which suggests that circ_RPL23A may be a promising therapeutic biomarker for ccRCC.
Availability of data and materials The data that support the findings of this study are available on request from the corresponding author.

Declarations
Competing interests The authors declare that they have no competing interest.
Ethics approval and consent to participate The design of this protocol follows the tenets of the Declaration of Helsinki, approved by the Ethics Committee of The First Affiliated Hospital of Jiamusi University.