circAMOTL1L Suppresses Renal Cell Carcinoma Growth by Modulating the miR-92a-2-5p/KLLN Pathway

Accumulating evidence indicates that the dysregulation of circular RNAs (circRNAs) contributes to tumor progression; however, the regulatory functions of circRNAs in renal cell carcinoma (RCC) remain largely unknown. In this study, the function and underlying mechanism of circAMOTL1L in RCC progression were explored. qRT-PCR showed the downregulation of circAMOTL1L in RCC tissues and cell lines. The decrease in circAMOTL1L expression correlated with the tumor stage, metastasis, and poor prognosis in patients with RCC. Functional experiments revealed that circAMOTL1L inhibited cell proliferation and increased apoptosis in RCC cells. Subcutaneous implantation with circAMOTL1L-overexpressing cells in nude mice decreased the growth ability of the xenograft tumors. Mechanistically, circAMOTL1L served as a sponge for miR-92a-2-5p in upregulating KLLN (killin, p53-regulated DNA replication inhibitor) expression validated by bioinformatics analysis, oligo pull-down, and luciferase assays. Further, reinforcing the circAMOTL1L–miR-92a-2-5p–KLLN axis greatly reduced the growth of RCC in vivo. Conclusively, our findings demonstrate that circAMOTL1L has an antioncogenic role in RCC growth by modulating the miR-92a-2-5p–KLLN pathway. Thus, targeting the novel circAMOTL1L–miR-92a-2-5p–KLLN regulatory axis might provide a therapeutic strategy for RCC.


Introduction
Renal cell carcinoma (RCC) is a common malignant tumor of the human urinary system and the second leading cause of death among patients with urologic tumors, accounting for approximately 3% of all adult malignancies [1][2][3]. Although partial and radical nephrectomy is the most effective treatment for early-stage or localized RCC, approximately 30% of patients with RCC experience recurrence and metastasis [4,5]. Owing to the insensitivity to radiotherapy and chemotherapy, the prognosis of patients with RCC is poor and the 5-year survival rate of advanced RCC is <10% [6,7]. However, the pathogenesis of RCC remains to be clarified. Thus, novel molecular mechanisms underlying RCC progression need to be further explained to explore more effective therapeutic strategies for RCC. Circular RNAs (circRNAs) are a novel class of singlestranded noncoding RNA characterized by covalently closed loops mostly generated from back-splicing of exons [8]. Compared with linear RNAs, circRNAs are structurally more stable and not easily degraded by RNases owing to their circular structure [9] and are widely identified in various eukaryotic cells [10,11]. Accumulating studies have confirmed that circRNAs perform different critical regulatory activities in gene expression, including sponging miRNA [12], modulating alternative splicing [8], interacting with RNA binding proteins [13], and regulating gene transcription and protein-encoding ability [14]. Thus, circRNAs have been implicated in the pathogenesis of several human diseases, such as cancer, Alzheimer's disease, and cardiac diseases [15][16][17][18]. Particularly, accumulating evidence suggests that circRNAs play critical roles in tumor progression by regulating the proliferation, apoptosis, invasion, and metastasis of tumor cells [19][20][21] and are gradually being expected to become valuable prognostic biomarkers and therapeutic targets for malignant tumors. Recently, circRNAs have been increasingly reported to participate in RCC tumorigenesis. circSDHCs promote the proliferation and metastasis of RCC via the CDKN3/E2F1 axis [22]. circPVT1 promotes progression of ccRCC by sponging miR-145-5p and regulating TBX15 expression [23]. circPRRC2A promotes TRPM3induced epithelial-mesenchymal transition of RCC through sponging miR-514a-5p and miR-6776-5p [24]. However, the exact function and underlying mechanism of circRNAs in RCC remain largely unclear.
circAMOTL1L (circBase ID: hsa_circ_0000350) is derived from back-splicing of the Amotl1 mRNA (from exon-2 to exon-3) and is located on chr11:94,528,176-94,533,477; it has been reported to exhibit a tumor suppressor characteristic in prostate cancer [25]. Here, we identified that circAMOTL1L is downregulated in RCC tissues and cell lines and is negatively associated with the growth of RCC cells. Mechanistically, circAMOTL1L served as a sponge of miR-92a-2-5p, thereby upregulating KLLN expression and inducing RCC growth inhibition. Conclusively, circA-MOTL1L may serve as an antioncogene in RCC progression, and a novel regulatory network involving the circA-MOTL1L-miR-92a-2-5p-KLLN axis might serve as a therapeutic strategy for RCC.

Clinical Sample Collection.
All 51 pairs of RCC tissues and matched adjacent noncancerous tissues were collected from patients with RCC who had undergone surgical resection at the Kaifeng Central Hospital from 2015 to 2018. According to the histological classification, all RCC tissues were of clear cell RCC (ccRCC). All patients with RCC were pathologically diagnosed and received no preoperative treatment. This study protocol was permitted by the Ethics Committee of the Kaifeng Central Hospital, and all patients provided their informed consent.
2.4. qRT-PCR Analysis. Total RNA was extracted from clinical samples by treating cultured cells with the TRIzol reagent (Invitrogen). NanoDrop 2000 was used to determine RNA quality. To determine circAMOTL1L expression, the M-MLV First-Strand Kit (Life Technologies) and Platinum SYBR Green qPCR Super Mix UDG Kit (Invitrogen) were used on the ABI 7500 FAST system (Life Technologies). The miScripIIRT kit (Qiagen) and miScript SYBR Green PCR kit (Qiagen) were used to quantify miRNA expression. GAPDH was used as an internal control of circAMOTL1L, and U6 was used as an internal control of microRNAs. Relative transcript expression levels were calculated using the 2-ΔΔCt method. Supplementary Table 1 lists the sequences of primers.
2.5. Cell Proliferation Assay. Cell proliferation was detected by Cell Counting Kit-8 (CCK-8) (Dojindo, Japan). Transfected 786-O cells were cultured in a 96-well plate (2 × 10 3 per well) for 24, 48, 72, and 96 h. Next, 10 ml CCK-8 solution was added to each well and incubated for 2 h. The optical density value of each well was measured using a microplate reader (Bio-Rad, USA).

Apoptosis
Assay. An apoptosis assay was performed using the TUNEL Apoptosis Detection kit (Yisheng, Shanghai, China) according to the specified instructions. DAPI (5 mg/ml) was used to stain the cell nuclei. Samples were detected using a Leica DM6000B microscope and digitized using LAS V.4.4 software. The TUNEL-positive rate reflecting the degree of apoptosis was calculated by enumerating the total number of TUNEL-stained nucleus divided by total DAPI-positive nucleus.

Biotin-Oligo Pull-Down Analysis.
To find the interactions between circAMOTL1L and miRNAs, biotin-oligo pull-down analysis was performed as previously described. 2.11. Statistical Analyses. The applicable statistical methods were used depending on the type of data. Student's t-test was performed for comparisons between two groups. Linear correlations were evaluated using the Pearson correlation analysis. A chi-squared test was used to analyze the correlation among the clinicopathological characteristics of patients with RCC and circAMOTL1L expression. The Kaplan-Meier method was used to assess the overall survival and disease-free survival curves. Pearson correlation analysis was used to analyze the correlations between the groups. All data from independent repeated trials are presented as means ± SEM. A p value of <0.05 was considered statistically significant. Statistical analysis was performed using Graph-Pad Prism 7.0 software.

circAMOTL1L Expression Is Downregulated in Human RCC Tissues and Cell Lines and Correlates with Poor
Prognosis. Fifty-one pairs of RCC tissues and adjacent noncancerous tissues were collected to detect circAMOTL1L expression using qRT-PCR. The circAMOTL1L expression level significantly decreased in RCC tissues compared with adjacent noncancerous tissues (Figures 1(a) and 1(b)). Moreover, the circAMOTL1L expression level in RCC samples negatively correlated with tumor size, tumor stage, and metastasis, but not with age and sex (Figures 1(c) and 1(d), Table 1). In addition, low circAMOTL1L expression in patients with RCC was associated with a lower overall survival rate than high circAMOTL1L expression ( Figure 1(e)). Next, we evaluated circAMOTL1L expression in different RCC cell lines. Results of qRT-PCR showed that the circAMOTL1L expression level was lower in RCC cell lines, particularly in 786-O and Caki-1 cell lines, than in normal kidney cells (HK-2) (Figure 1(f)). These 3 Oxidative Medicine and Cellular Longevity findings indicate that circAMOTL1L might act as an antioncogene in RCC.     Figure 3a). Furthermore, PCNA, P53, Bax, and bcl-2 expression levels were examined in RCC clinical samples. Correlation analysis revealed that the expression level of PCNA and bcl-2 was negatively correlated with that of circAMOTL1L, but the expression level of P53 and Bax was positively correlated with that of circAMOTL1L (Figures 2(g) and 2(h), Supplementary Figures 3b and 3c). These results suggested that circAMOTL1L prompted RCC cells to gain more apoptotic activity in vivo. Overall, circAMOTL1L may be responsible for RCC progression by regulating the proliferation and apoptosis of RCC cells.

miR-92a-2-5p Mediates the Regulation of circAMOTL1L in the Proliferation and Apoptosis in RCC Cells.
Furthermore, we tested the involvement of miR-92a-2-5p in cell proliferation and apoptosis regulated by circAMOTL1L. From Figure 4(a), transfection with the miR-92a-2-5p mimic strikingly increased, whereas transfection with anti-miR-92a-2-5p decreased the miR-92a-2-5p expression level in 786-O cells compared with their negative control. Next, 786-O cells were transfected with pcDNA-circAMOTL1L or miR-92a-2-5p mimic or cotransfected with both. CCK-8 and TUNEL data showed that the miR-92a-2-5p overexpression facilitated cell proliferation and suppressed apoptosis of 786-O cells. Moreover, the pcDNA-circAMOTL1L-induced inhibition of proliferation and promotion of apoptosis of cells were both abrogated by miR-92a-2-5p upregulation (Figures 4(b) and 4(d)). Next, we transfected 786-O cells with pcDNA-circAMOTL1L or anti-miR-92a-2-5p or both. Depletion of miR-92a-2-5p by its inhibitor decreased the cell proliferation and increased the apoptosis of 786-O cells (Figures 4(c) and 4(e)). However, the inhibition of proliferation and promotion of apoptosis of cells further improved after the knockdown of miR-92a-2-5p combined with circA-MOTL1L overexpression. These data show that miR-92a-2-5p mediates the regulatory action of circAMOTL1L on the proliferation and apoptosis of RCC cells.

10
Oxidative Medicine and Cellular Longevity mediated by wild-type KLLN 3 ′ UTR, but not by the mutant one ( Figure 5(b)). Further, we transfected 786-O cells with anti-miR-92a-2-5p, miR-92a-2-5p mimic, or their comparable control RNA. Western blotting showed that miR-92a-2-5p knockdown clearly increased KLLN protein expression, whereas miR-92a-2-5p overexpression significantly decreased KLLN protein expression ( Figure 5(c)). These results suggest that miR-92a-2-5p inhibits the translation of KLLN by binding to 3 ′ UTR of KLLN. Subsequently, we used KLLN siRNAs to investigate the effect of KLLN on proliferation and apoptosis of cells and whether KLLN mediates the regulatory effect induced by miR-92a-2-5p. First, the transfection of si-KLLN in 786-O cells significantly reduced the KLLN protein expression compared with that in the control group (Figure 5(d)). Next, 786-O cells were transfected with anti-miR-92a-2-5p or si-KLLN or both. CCK-8 data display that KLLN depletion increased cell proliferation and further reversed the suppression of cell proliferation induced by miR-92a-2-5p knockdown ( Figure 5(e)). Consistently, apoptosis rates were significantly decreased after KLLN depletion. However, the apoptosis facilitation after miR-92a-2-5p knockdown could be neutralized by cotransfection with si-KLLN ( Figure 5(f)). These results indicate that miR-92a-2-5p exerts the promotion of cell proliferation and antiapoptotic effect by targeting KLLN. Correspondingly, the tumor weights of nude mice implanted with 786-O cells infected with LV-circAMOTL1L plus LVanti-miR-92a-2-5p were further reduced compared with each of them alone (Figure 6(f)). Furthermore, circA-MOTL1L overexpression or miR-92a-2-5p depletion decreased the expression of PCNA and Bcl-2 but increased the expression of P53 and Bax in the xenograft tumor, and this situation was further exacerbated by the superposition of the two (Figures 6(g) and 6(h), Supplementary  Figures 4a and 4b). Next, we detected the KLLN protein levels in xenograft tissues by western blotting, and the results showed that KLLN protein levels significantly increased in xenograft tumors when circAMOTL1L overexpression was combined with miR-92a-2-5p depletion ( Figure 6(i)). Overall, these results demonstrate that the circAMOTL1L-miR-92a-2-5p-KLLN axis is involved in RCC growth through regulation of the proliferation and apoptosis of RCC cells.

Discussion
Accumulating studies have confirmed that the dysregulation of circRNAs contributes to tumor progression [27,28].
Recently, some circRNAs have been reported to participate in RCC tumorigenesis. However, the function and underlying mechanism of circRNAs in RCC remain largely unknown. In this study, a novel mechanism that circA-MOTL1L can go through the circAMOTL1L-miR-92a-2-5p-KLLN axis to suppress RCC growth was discovered. circAMOTL1L, derived from the circularization of the exon-2 and exon-3 of the Amotl1 gene, was reported to exhibit a tumor suppressor characteristic in prostate cancer by modulating the miR-193a-5p/Pcdha axis to suppress the migration and invasion of PCa cells [25]. The role of circA-MOTL1L in RCC development was investigated in this study. We first indicated that circAMOTL1L was notably downregulated in RCC tissues and cell lines, and its level was associated with clinicopathological features and overall survival ( Figure 1, Table 1), implying that it may be involved in RCC progression and probably used as a prognostic biomarker for RCC.
The imbalance between cell proliferation and apoptosis is one of the mechanisms underlying tumorigenesis and is associated with the metastasis and drug resistance of tumor cells [29][30][31]. To explore whether circAMOTL1L exerts regulatory effects on RCC growth, cell function assays were performed both in vitro and in vivo with a focus on cell proliferation and apoptosis. Our results showed that circA-MOTL1L inhibited proliferation and promoted apoptosis in RCC cells. Furthermore, subcutaneous implantation of circAMOTL1L-overexpressing tumor cells decreased the growth ability of the xenograft tumors in nude mice. Moreover, these circAMOTL1L-overexpressing tumor cells exhibit more of an apoptotic phenotype than a proliferative phenotype in vivo (Figure 2, Supplementary Figures 1-3). These results suggested that circAMOTL1L acts as an antioncogene in RCC by regulating RCC cell proliferation and apoptosis. In addition, circAMOTL1L may regulate RCC cell migration and invasion. Yang et al. found that decreased circAMOTL1L level facilitates PCa cell migration and invasion, through downregulating E-cadherin and increasing vimentin expression, and leads to EMT and PCa progression [25]. The major limitation of the present study 11 Oxidative Medicine and Cellular Longevity 13 Oxidative Medicine and Cellular Longevity is the lack of evaluation of the effect of circAMOTL1L on the migration and invasion of RCC cells, which needs further investigation. Another limitation is that the rationale underlying the downregulation of circAMOTL1L in RCC is unclear. Yang et al. found that P53 regulated circAMOTL1L expression via RBM25, a RNA binding protein, which directly binds to circAMOTL1L and induces its biogenesis [25]. However, further investigation is also needed to demonstrate the upstream regulatory action of circAMOTL1L in RCC cells. 14 Oxidative Medicine and Cellular Longevity circRNAs were mostly reported to serve as a miRNA sponge to regulate the function and activity of miRNAs [12,32]. Particularly, studies have indicated that circRNAs can sponge miRNA to modulate miRNA target gene expression in various cancers. Specifically, Guan et al. found that circPUM1 promoted tumorigenesis and progression of ovarian cancer by sponging miR-615-5p and miR-6753-5p to upregulate NF-κB and MMP2 expression [33]. Chen et al. reported that circPRMT5 promoted tumor metastasis by sponging miR-30c to regulate the expression of snail-1 and E-cadherin in urothelial carcinoma of the bladder [34]. Moreover, circRNA 100146 promoted the proliferation and invasion of cells and inhibited apoptosis by sponging miR-361-3p and miR-615-5p in non-small-cell lung cancer [20]. In this study, oligo pull-down and luciferase reporter assays confirmed that circAMOTL1L could directly bind to miR-92a-2-5p, but circAMOTL1L overexpression could not affect miR-92a-2-5p expression in RCC cells. Notably, whether cir-cRNAs participate in the posttranscriptional regulation of miRNAs remains conflicting. Specifically, circRNA CDR1a harbors 63 conserved miR-7 target sites to serve as the miR-7 sponge but could not affect miR-7 expression [11]. circNCX1 promotes ischemic myocardial injury by sponging miR-133a-3p to inhibit its activity without regulating the miR-133a-3p expression levels [35]. Moreover, accumulating evidence has demonstrated that circRNAs could not affect miRNA degradation but still act as a miRNA sponge and compete for miRNA target sites to suppress its activity. Therefore, our study suggests that circAMOTL1L acts as a "miRNA sponge" of miR-92a-2-5p and is probably uninvolved in miR-92a-2-5p decaying activity ( Figure 3).
However, the role of miR-92a-2-5p in RCC has never been studied. We found that miR-92a-2-5p facilitated cell proliferation and suppressed apoptosis in RCC cells and mediated the regulation of circAMOTL1L in the proliferation and apoptosis of RCC cells (Figure 4), indicating that the circAMOTL1L-miR-92a-2-5p pathway plays a regulatory role in RCC development.
circAMOTL1L served as a competing endogenous RNA (ceRNA) for miR-92a-2-5p. miR-92a-2-5p bound by circA-MOTL1L could not bind to its target mRNA and lost its ability to inhibit target gene expression. In our study, bioinformatics analysis revealed that KLLN is a potential target of miR-92a-2-5p. Further, luciferase reporter assays and western blotting demonstrated that miR-92a-2-5p could directly bind to the 3 ′ UTR of KLLN and inhibit its protein expression. At the function level, downregulated KLLN promoted cell proliferation and attenuated apoptosis, which reversed the antiproliferative and proapoptotic effects induced by miR-92a-2-5p knockdown. Alternatively, miR-92a-2-5 improved RCC cell proliferation and reduced apoptosis by targeting KLLN (Figure 5).
KLLN (also called KILLIN), first reported in 2008 which is localized to 10q23 and shares the same transcription start site with PTEN, was indicated as a tumor suppressor necessary and sufficient for p53-mediated apoptosis [26]. Studies have demonstrated that KLLN function is associated with the regulation of cell growth and death. Particularly, KLLN inhibits breast cancer growth by activating p53/p73-medi-ated apoptosis [36]. In prostate carcinomas, KLLN inhibits tumor cell proliferation and invasiveness by transcriptionally regulating the expression of AR, TP53, and TP73 [3]. Moreover, miR-143-3p promoted proliferation and migration and suppressed apoptosis in vascular smooth muscle cells by targeting KLLN [37]. Observably, methylation significantly downregulated KLLN expression, which led to increased risks of RCC [38]. However, the involvement of KLLN in RCC progression has not yet been explored. Here, we revealed the novel upstream circAMOTL1L-miR-92a-2-5p regulatory pathway for the posttranslational regulation of KLLN. Furthermore, depleted KLLN abolished the antiproliferative and proapoptotic effects induced by circAMOTL1L overexpression in RCC cells. Reinforcing the circA-MOTL1L-miR-92a-2-5p-KLLN axis greatly reduced RCC growth in vivo ( Figure 6, Supplementary Figure 4).

Data Availability
The raw data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors declare that they have no conflict of interest. expression level of Bax was positively correlated with that of circAMOTL1L but expression level of bcl-2 was negatively correlated with that of circAMOTL1L in RCC clinical samples. Supplementary Figure 4: circAMOTL1L overexpression or miR-92a-2-5p depletion increased Bax expression and decreased Bcl-2 expression in the xenograft tumor. Supplementary Table 1: oligos used in this study. Supplementary Table 2: circAMOTL1L harbors 6 supposed binding sites for miR-92a-2-5p. (Supplementary Materials)