The lncRNA KIF9-AS1 Accelerates Hepatocellular Carcinoma Growth by Recruiting DNMT1 to Promote RAI2 DNA Methylation

Background Hepatocellular carcinoma (HCC) is a very common malignant tumor. Long noncoding RNAs (lncRNAs) enable discoveries of new therapeutic tumor targets. We aimed to study the role and potential regulatory mechanisms of the lncRNA KIF9-AS1 in HCC. Methods CCK-8, scratch assay, and flow cytometry were used to detect cell proliferation, migration, and apoptosis, respectively. Bax, Bcl-2, ERK, and pERK expression were measured by western blotting. StarBase predicted KIF9-AS1 expression in HCC and paracancerous tissues. RPISeq predicted the interaction score of KIF9-AS1 and DNMT1, and MethyPrimer revealed the CpG island distribution in the RAI2 promoter. MSP was performed to measure RAI2 methylation. RIP and ChIP were performed to examine lncRNA KIF9-AS1, DNMT1, and RAI2 interactions. Finally, the effect of KIF9-AS1 knockdown on HCC was verified with nude mice. Results We found that KIF9-AS1 expression was increased in HCC tissues. KIF9-AS1 knockdown inhibited the proliferation and migration, and facilitated the apoptosis of HCC cells. lncRNA KIF9-AS1-mediated RAI2 expression led to DNMT1 recruitment and regulated RAI2 DNA methylation. RAI2 overexpression inhibited the proliferation and migration and promoted the apoptosis of HCC cells. KIF9-AS1 knockdown inhibited subcutaneous tumor formation in vivo. Conclusion This study shows that KIF9-AS1 accelerates HCC growth by inducing DNMT1 promotion of RAI2 DNA methylation.


Introduction
Hepatocellular carcinoma (HCC), the most common primary liver cancer, is an invasive disease that develops in patients with chronic liver disease [1,2]. HCC is characterized by a high degree of molecular phenotypic heterogeneity and is a major cause of cancer-related death [3,4]. In recent years, the incidence and mortality of HCC have continued to rise, especially due to the obesity pandemic leading to nonalcoholic fatty liver disease, and the global HCC mortality is expected to increase by another 41% by 2040 [5,6]. The main causes of HCC development are cirrhosis, alcoholic liver disease, diabetes, and obesity [7]. However, effective treatments are lacking. Drugs targeting specific epigenetic mechanisms are suitable tools for effective clinical treatment of various diseases [8]. Wilms' tumor 1-associated protein (WTAP) has been reported to promote HCC progression through m6A-HuR-dependent epigenetic silencing of ETS1 [9]. Circulating tumor DNA methylation markers also play vital roles in HCC diagnosis and prognosis [10]. However, the epigenetic changes in HCC and possible use of DNA methylation markers as prognostic biomarkers remain unclear.
Long noncoding RNAs (lncRNAs) play essential roles in epigenetic and gene expression regulation and have been reported to be critically involved in HCC progression [11]. lncRNA KIF9-AS1 has been reported and studied in cancers not related to HCC. An analysis with the lncATLAS website revealed that lncRNA KIF9-AS1 is located mainly in the nucleus, indicating that KIF9-AS1 may play epigenetic roles. In different diseases, KIF9-AS1 expression levels and functions differ, suggesting that KIF9-AS1 may regulate microRNA (miRNA) expression. For example, lncRNA KIF9-AS1 might promote nasopharyngeal carcinoma progression by targeting miR-16 [12]. lncRNA KIF9-AS1 has also been shown to enhance chemotherapy resistance in renal cell carcinoma mediated by microRNA-497-5p [13]. In addition, we previously found that lncRNA KIF9-AS1 expression was upregulated in liver hepatocellular carcinoma (LIHC) patients through StarBase database (https:// starbase.sysu.edu.cn/index.php). It can be inferred that lncRNA KIF9-AS1 may promote HCC development. Moreover, the analysis of lncATLAS database revealed that the lncRNA KIF9-AS1 is located mainly in the nucleus, suggesting that it may play a regulatory role through the epigenetic modification of target genes. However, the function and potential molecular mechanisms of lncRNA KIF9-AS1 action in HCC remain unclear.
DNA methylation is an epigenetic process, and DNA methyltransferase 1 (DNMT1) can maintain the methylation of newly copied DNA [14]. A previous report revealed oncogenic roles played by DNMT1 in various malignancies [15]. lncRNAs might influence DNMT1 action, and the dysregulation of lncRNAs has been shown to lead to abnormal DNA methylation patterns [16]. For example, DNMT1mediated MEG3 methylation facilitated the endothelialmesenchymal transformation (EMT) mediated through the PI3K/Akt/mTOR pathway in diabetic retinopathy [17]. Moreover, the lncRNA MIR210HG promoted the proliferation and invasion of non-small-cell lung cancer cells by binding DNMT1 and thus upregulating CACNA2D2 promoter methylation [18]. Li et al. found that the lncRNA DDX11-AS1 epigenetically suppressed LATS2 action by interacting with EZH2 and DNMT1 in HCC [19]. Xu et al. showed that the lncRNA Linc-GALH promoted tumorigenesis via the AKT pathway by regulating gankyrin promoter methylation in HCC [20]. Retinoic acid-induced 2 (RAI2) is a newly discovered tumor suppressor, and RAI2 promoter region methylation has been reported to indicate poor prognosis in colorectal cancer, RAI2 downregulated expression or hypermethylation has been identified in colorectal cancer [21]. In the present study, we predicted that there was CpG island distribution in the RAI2 promoter by MethyPrimer database. Meanwhile, we also predicted that lncRNA KIF9-AS1 could interact with DNMT1. Taking prediction results into account, we speculated that lncRNA KIF9-AS1 may be involved in HCC development by recruiting DNMT1 to promote DNA methylation of RAI2. However, regulatory mechanism was not reported in the other literatures, which needs to be explored in our study.
In addition, the upregulation of the circular RNA (circRNA), circRBPMS inhibited bladder cancer cell proliferation and metastasis by targeting the miR-330-3p/RAI2 axis and inhibiting ERK pathway activation [22]. It has been reported that anlotinib induced HCC cell apoptosis and repressed HCC cell proliferation via the ERK and Akt pathways [23]. Chen et al. found that loss of RAD52 motif 1 accelerated the progression of HCC mediated through the p53 and Ras/Raf/ERK pathways [24]. However, the relationship of RAI2 and the ERK pathway, and the role they play in HCC need further study.
Considering the aforementioned studies, we mainly wanted to explore the role played by the lncRNA KIF9-AS1 in HCC and identify the potential regulatory mechanism. We found that the lncRNA KIF9-AS1 promoted HCC cell proliferation and migration and inhibited HCC cell apoptosis. Furthermore, the lncRNA KIF9-AS1 promoted RAI2 DNA methylation by recruiting DNMT1 and repressing its expression, thus activating the ERK pathway. This finding might lead to new diagnostic and therapeutic targets for HCC treatment. Then, an antibody against Ago2 (CST, Boston, USA) or antirabbit IgG (the negative control, CST, Boston, USA) were added to magnetic beads. The samples and inputs were processed with proteinase K to extract RNA. Then, qRT-PCR was performed to identify precipitated KIF9-AS1. Total RNA was the input control.

Chromatin Immunoprecipitation (ChIP). A ChIP Kit
(Sigma, USA) was used to detect the interaction between DNMT1 and the RAI2 promoter. Cells were immobilized with 1% formalin for 10 min, and then DNA was randomly segmented by ultrasound into 200-800-bp fragments. This DNA was immunoprecipitated with an antibody against the target protein RAI2 (ab247100, 0.2 μg/mL, Abcam, UK). Finally, ChIP DNA was purified and eluted with 100 μL of H 2 O, and 2.5 μL of ChIP-DNA was analyzed by qPCR.

Xenograft Mice Model.
Twenty-four specific-pathogenfree (SPF)-grade and four-week-old mice were randomly assigned to a control group, an sh-NC group and an sh-KIF9-AS1 group, with eight mice in each group. The animal studies were approved by the Animal Ethics Committee of The Fifth Medical Center, Chinese PLA General Hospital. The subcutaneous tumor-forming nude mouse model was established by injecting Huh-7 cells transfected with sh-NC or sh-KIF9-AS1. The tumor volume in each group was measured every seven days. The nude mice were sacrificed 28 days after the injection. The tumor tissues were removed and weighed. Then, the tumor tissues of six mice were embedded into paraffin and cut into sections. The tissue from the other two mice in each group were prepared as tissue homogenates.
2.11. Immunohistochemistry (IHC). The paraffin sections were dewaxed and dehydrated following standard procedures. A Ki67 primary antibody (ab15580, Abcam, UK) was incubated with the sections overnight at 4°C, and then, the sections were rinsed 3 times with PBS for 5 min each time. A secondary antibody was incubated with the sections at 37°C for 30 min. DAB was used for color development.
Hematoxylin stain was also incubated with the sections for 5-10 min. Then, the sections were washed with distilled water. PBS turned blue. An alcohol series (60-100%) was used for dehydration with the sections incubated for 5 min at each gradient level. The sections were removed from the gradient and placed twice in xylene for 10 min each time, sealed with neutral gum, and observed under a microscope.
2.12. Nuclear/Cytosol Fractionation Assay. A PARIS™ kit (Invitrogen) was used to isolate nuclear and cytoplasmic RNA from cells according to the manufacturer's instructions. The RNA levels of GAPDH, U6, and KIF9-AS1 were detected by qRT-PCR.
2.13. Quantitative Real-Time PCR (qRT-PCR). Total RNA was extracted by the TRIzol method (Invitrogen, USA), and a cDNA reverse transcription kit (Invitrogen) was used to reverse transcribe RNA into cDNA. SYBR Green qPCR Mix (Invitrogen) was used to test the relative gene expression with an ABI 7900 system (PRISM® 7900HT, ABI, Massachusetts, USA). GAPDH was the reference gene, and the 2 −ΔΔCt method was used to calculate the relative gene expression level. The following primer sequences were used in this study: the lncRNA KIF9-AS1 forward: 5′-ACCCTCAGCCC TTCCACTAA -3 ′ and the lncRNA KIF9-AS1 reverse: 5 ′ -TGGTTTACTTC CACATAGCTGACT-3 ′ .
DNMT1 forward: 5 ′ -ATGCTTACAACCGGGAAGTG -3 ′ and DNMT1 reverse: 5 ′ -TGAACGCTTAGCCTCTCCAT -3 ′ .   Journal of Oncology  , and the experimental data are expressed as the means ± standard deviation (SD). Each group included three replicates in cell experiments. Student's t-test was performed to compare two groups. One-way analysis of variance (ANOVA) followed by Tukey's post-hoc test was performed to compare multiple groups. P < 0:05 was considered to be statistically significant.

Results
3.1. Knocking down lncRNA KIF9-AS1 Expression Inhibited the Proliferation and Migration and Promoted the Apoptosis of HCC Cells. lncRNA KIF9-AS1 had been previously speculated to promote nasopharyngeal carcinoma progression by targeting miR-16 [12]. In this study, we found that KIF9-AS1 expression was upregulated in LIHC patients, as determined through an analysis of the StarBase database (Figure 1(a)). However, no relevant literature on KIF9-AS1 expression in HCC had been reported. Therefore, we detected KIF9-AS1 expression in HCC cells and normal hepatic cells by performing qRT-PCR. Compared with that in normal hepatic cells (HHL-5 cells), KIF9-AS1 expression was upregulated in HCC cells (Huh-7, BEL-7405, SNU-398, SNU-387, and Li-7 cells), and the expression of KIF9-AS1 was most significantly upregulated in the Huh-7 cells (Figure 1(b)). Therefore, Huh-7 cells were selected for further study. To study the effects of the lncRNA KIF9-AS1 on Huh-7 cells, we knocked down KIF9-AS1 expression in the Huh-7 cells (Figure 1(b)). The results showed that after knocking down lncRNA KIF9-AS expression, cell proliferative and migratory capacities were significantly reduced, while the cell apoptosis rate was significantly     (Figures 1(c)-1(e)). Moreover, the expression of the proapoptotic protein Bax was elevated, and the expression of the antiapoptotic protein Bcl-2 was reduced after KIF9-AS knockdown (Figure 1(f)). Moreover, after KIF9-AS knockdown, ERK pathway activation was inhibited, and the phosphorylation level of ERK was decreased (Figure 1(f)). These results indicated that lncRNA KIF9-AS1 knockdown repressed HCC cell proliferation and migration and promoted HCC cell apoptosis.

The lncRNA KIF9-AS1 Regulated RAI2 Expression by
Recruiting DNMT1, Inducing the Regulation of RAI2 DNA Methylation. The subcellular localization website lncATLAS was used to predict that the lncRNA KIF9-AS1 is mainly located in the nucleus (Figure 2(a)). As shown in Figure 2(b), the qRT-PCR data verified that the lncRNA KIF9-AS1 is mainly expressed in the nucleus. These results suggested that KIF9-AS1 might play a regulatory role through the epigenetic modification of target genes. An analysis of the website RPISeq (http://pridb.gdcb.iastate.edu/ RPISeq/) revealed that lncRNA KIF9-AS1 bound DNMT1 (the random forest [RF] classifier was 0.6, and the support vector machine [SVM] classifier was 0.99). RIP assays confirmed interactions between the lncRNA KIF9-AS1 and DNMT1, DNMT3a, and DNMT3b. The results further revealed that only DNMT1 interacted with KIF9-AS1 (Figure 2(c)). DNMT1 expression was significantly decreased after KIF9-AS knockdown (Figures 2(d) and 2(e)). These results indicated that the lncRNA KIF9-AS1 interacted with DNMT1.
The online MethyPrimer tool was used to predict CpG island distribution in the RAI2 promoter, and we found that the CpG islands in the RAI2 promoter region might be methylated. The results of promoter methylation of RAI2 assay showed that RAI2 was hypermethylated in Huh-7 cells that after the addition of the demethylation reagent 5-aza-dc or DNMT1 knockdown, RAI2 was unmethylated, and that after DNMT1 was overexpressed, RAI2 was hypermethylated (Figure 3(a)). Furthermore, after lncRNA KIF9-AS1 knockdown, the RAI2 promoter was unmethylated, but it was hypermethylated after lncRNA KIF9-AS1 overexpression (Figure 3(b)). After DNMT1 was overexpressed, the degree of DNMT1 binding to RAI2 was decreased, and ChIP revealed an interaction between DNMT1 and the RAI2 promoter (Figure 3(c)). The expression of RAI2 was significantly downregulated after DNMT1 overexpression (Figures 3(d) and 3(e)). These findings suggested that DNMT1 regulated RAI2 expression by regulating RAI2 DNA methylation. They also revealed that the lncRNA KIF9-AS1 regulated RAI2 expression by recruiting DNMT1, which regulated RAI2 DNA methylation.

Overexpression of RAI2 Inhibited HCC Cell Proliferation and Migration and Promoted HCC Cell Apoptosis.
A previous study showed that the expression of RAI2 was downregulated and that it was hypermethylated in colorectal cancer [21]. To explore the role played by RAI2 in HCC, we measured the expression of RAI2 by qRT-PCR. Compared with normal hepatic cells (HHL-5 cells), RAI2 expression was downregulated in HCC cells (Huh-7, BEL-7405, SNU-398, SNU-387, and Li-7 cells), and the expression of RAI2 was found to be the most significantly upregulated in the Huh-7 cells. Subsequently, to further investigate the role played by RAI2 in HCC, RAI2 was overexpressed in Huh-7 cells. The qRT-PCR and western blot results indicated that the RAI2 was successfully overexpressed (Figures 4(a) and  4(b)). After RAI2 overexpression, cell proliferation and migration were significantly inhibited, and cell apoptosis was significantly increased (Figures 4(c)-4(e)). Moreover, after RAI2 was overexpressed, the expression of the proapoptotic protein Bax was elevated, and the expression of the antiapoptotic protein Bcl-2 was reduced. In addition, the ERK pathway was inhibited when RAI2 was overexpressed (Figure 4(f)). These results indicated that RAI2 overexpression repressed HCC cell proliferation and migration and promoted HCC cell apoptosis. The results showed that RAI2 expression was downregulated when only RAI2 expression was knocked down, but lncRNA KIF9-AS1 expression remained unchanged. After simultaneously knocking down lncRNA KIF9-AS1 and RAI2, RAI2 expression was decreased, and lncRNA KIF9-AS1 expression remained unchanged (Figures 5(a) and 5(b)). After only RAI2 expression was knocked down, the HCC cell proliferation and migration rates were significantly increased, and HCC cell apoptosis was suppressed. In addition, after lncRNA KIF9-AS1 and RAI2 expression was knocked down, the proliferative and migratory abilities of the HCC cells were significantly enhanced, and the apoptosis rate of these cells was significantly decreased, which weakened the inhibitory effect of KIF9-AS1-only expression knockdown (Figures 5(c)-5(e)). Moreover, Bax expression was downregulated and Bcl-2 expression was upregulated after only RAI2 expression was knocked down. Moreover, the ERK pathway was activated after only RAI2 expressed was knocked down. However, the effect of KIF9-AS1 knockdown on apoptosis and the ERK pathway was weakened after KIF9-AS1 and RAI2 expressions were knocked down ( Figure 5(f)). These results suggested that knocking down RAI2 expression reversed lncRNA KIF9-AS1-knockdown effects on the proliferation, migration, and apoptosis of HCC cells.    The Huh-7 cell apoptosis rate was analyzed by flow cytometry. (f) WB was performed to detect Bax, Bcl-2, ERK, and pERK expression. The data are expressed as the means ± SD. * P < 0:05, * * P < 0:01, and * * * P < 0:001.

13
Journal of Oncology HCC cells by regulating miR-122-5p and RPL4 [31]. The lncRNA MYCNOS accelerated the proliferation and invasion of liver cancer by regulating miR-340 [32]. lncRNA KIF9-AS1 is one of the most widely studied lncRNAs. It has been reported that lncRNA KIF9-AS1 regulated TGF-β and autophagy signaling through miR-497-5p to enhance chemoresistance in renal cell carcinoma [33]. However, the role that lncRNA KIF9-AS1 plays in HCC has been unclear. In this study, we found that KIF9-AS1 expression was increased in HCC tissues and cells. Moreover, knocking down lncRNA KIF9-AS1 expression repressed HCC cell proliferation and migration, promoted HCC cell apoptosis in vitro and inhibited subcutaneous HCC tumor formation in nude mice. This was the first report of lncRNA KIF9- expression. The data are expressed as the means ± SD, n = 8. * P < 0:05, * * P < 0:01, and * * * P < 0:001.