XIST sponges miR-320d to promote chordoma progression by regulating ARF6

Highlights • XIST was highly expressed in chordoma tissues.• XIST knockdown inhibited chordoma progression by downregulating ARF6.• MiR-320d inhibited the malignant behaviors of chordoma cells.• XIST positively upregulated ARF6 expression via sponging miR-320d in chordoma cells.


Introduction
Chordoma is a rare primary malignant bone tumor with aggressive and high recurrence [1]. Chordoma is caused by leftover notochordal cells during fetal development and is most common in the sacrum, skull base, and vertebral column [2]. Chordoma is resistant to chemotherapy and poorly sensitive to conventional radiotherapy, so radical surgery combined with adjuvant radiotherapy is the most commonly used treatment [3,4]. In spite of this, many chordoma patients still suffer local recurrence, with metastasis occurring in 30 % to 40 % of patients, and the overall 5-year postoperative survival rate is about 50 % [5]. Hence, it is critical to clarify the molecular mechanisms that affect the progression of chordoma, which might provide a new theoretical basis for the treatment of chordoma.
Non-coding RNAs (ncRNAs) have gradually become one of the research hotspots in the tumor field in the past few decades [6]. Long non-coding RNAs (lncRNAs; >200 nucleotides) are an important type of ncRNAs without protein-coding capacity [7][8][9]. LncRNAs have been demonstrated to play crucial roles in the physiological and pathological processes of various diseases [10,11]. Moreover, previous studies have shown that lncRNAs are implicated in chordoma tumorigenesis via acting as tumor promoters or suppressors. For example, LOC554202 may play an anti-tumor role in chordoma via suppressing proliferation and invasion of chordoma cells [12]. LncRNA KRT8P41 acted as a tumor promoter via accelerating chordoma cell proliferation and invasion [13]. In addition, the former report indicated that lncRNA X-inactive specific transcript (XIST) was upregulated in chordoma, which might promote the development of chordoma [14]. However, more roles and regulatory mechanisms of XIST in chordoma are still largely unknown.
A large body of evidence has indicated that lncRNAs can upregulate the expression levels of target genes via acting as sponges for microRNAs (miRNAs), thereby affecting cell fate decisions [15,16]. Dysregulated miRNAs are strongly linked with the development of many diseases, including chordoma [17]. Accumulating evidence has revealed that miR-320d has an anti-tumor role in some tumors, including chordoma [18][19][20]. ADP-ribosylation factor 6 (ARF6) is widely expressed in mammalian cells and is used as a tumor-promoting gene in chordoma [21]. By searching the bioinformatics tools, we found that both XIST and ARF6 3 ′ UTR had the complementary binding sequence for miR-320d. According to these findings, we hypothesized that ARF6 might be involved in chordoma development via regulating miR-320d and ARF6.
In our work, we explored XIST expression in chordoma samples. Meanwhile, the role of XIST in chordoma was explored. In addition, the relationships among XIST, miR-320d, and ARF6 in chordoma cells were also investigated using bioinformatic databases and a series of experiments. We aimed to provide a possible therapeutic target for chordoma patients.

Specimen collection
In the current study, chordoma tissue specimens (n = 45) and adjacent normal tissue specimens (n = 45) were obtained from chordoma patients who had undergone surgery at Renmin Hospital of Wuhan University. The median age at diagnosis was 65 years old (range: 49-76 years old). All patients had not received any treatment before the operation. The tissues were frozen in liquid nitrogen after resection. Informed consent was obtained from each patient. The detailed clinical characteristics of patients are described in Table 1. This research was approved by the ethics committee of Renmin Hospital of Wuhan University.

Real-time quantitative polymerase chain reaction (RT-qPCR)
The total RNA was extracted from the tissues (chordoma and normal) and cells (U-CH1 and JHC7) using TRIzol reagent (Invitrogen). Next, cDNA was synthesized using a TIANScript RT Kit (for lncRNA and mRNA; Tiangen Biotech, Beijing, China) or miScript II RT kit (for miRNA; Invitrogen). Thereafter, RT-qPCR reaction was manipulated on a CFX96 Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) with SYBR GreenMaster Mix kit (Vazyme, Nanjing, China). 2-ΔΔCt method was used to determine RNA levels. GAPDH and U6 were utilized as the reference controls for lncRNA/mRNA and miRNA, respectively. The primers used in experiments were presented in Table 2.

Subcellular localization
PARIS™ Kit (Invitrogen) was utilized for isolating cytoplasm and nucleus fractions. The expression levels of XIST, U6 (a nucleus control), and GAPDH (a cytoplasm control) were determined by RT-qPCR.
For MTT assay, we seeded U-CH1 and JHC7 cells into a 96-well plate.
Each well was added with 5 mg/mL MTT (20 μL) for 4 h at the pointed times. After removing the cultured medium, dimethyl sulfoxide (DMSO; 150 μL; Beyotime) solution was added to each well. At last, a microplate reader (Bio-Rad) was applied to determine the absorbance at 490 nm. For EdU assay, we seeded U-CH1 and JHC7 cells into 24-well plates.
After transfection for 48 h, EdU solution (20 μM) was placed into each well and then incubated for 2 h. Next, cells were fixed using paraformaldehyde (4 %), followed by treatment with Triton-X-100 (0.5 %) to permeabilize the cells. The nucleic acids were stained using DAPI after staining with Click Additive Solution for 0.5 h. Finally, EdU-positive cells were counted and photographed with a fluorescence microscope (Leica, Wetzlar, Germany). For colony formation assay, we seeded U-CH1 and JHC7 cells into a 6-well plate at a density of 5 × 10 3 cells/well following transfection for 48 h. Thereafter, these cells were grown in a complete growth medium for 14 days and the culture medium was updated every 2-3 days. Next,  cells were subsequently fixed with methanol (Sangon Biotech, Shanghai, China) and stained with 0.4 g/L Giemsa (Beyotime). Lastly, a microscope (Leica) was applied to count the number of colonies (containing ≥ 50 cells per colony).

Transwell assay
Transwell chambers (Costar, Corning, NY, USA) were used for detecting cell migration and invasion. U-CH1 and JHC7 cells were suspended in a serum-free medium and plated in the top chamber precoated with (5 × 10 4 cells for detecting cell invasion) or without (1 × 10 5 cells for detecting cell migration) Matrigel. Meanwhile, the culture medium with 10 % FBS was placed into the bottom chamber. 24 h later, the non-migrated and non-invaded cells were wiped out with a cotton swab, and 4 % paraformaldehyde (Beyotime) was used to fix the migrated and invaded cells. Next, 0.1 % crystal violet (Beyotime) was used to stain the migrated and invaded cells. Lastly, these cells were photographed by a microscope (Leica).

Extracellular acidification rate (ECAR) assay
To detect ECAR, Seahorse Bioscience XF96 extracellular flux analyzer (Seahorse Bioscience, North Billerica, MA, USA) was used. Briefly, U-CH1 and JHC7 cells were placed in an XF microplate, followed by a baseline measurement. Glucose, Oligomycin (OM; oxidative phosphorylation inhibitor), and 2glycolytic inhibitor (2-DG) were then injected into each well according to the specified time point. At last, the results were analyzed using Seahorse XF96 Wave software (Seahorse Bioscience).

Measurement of lactate production
Based on the manufacturer's protocol, lactate production was examined using Lactate Assay Kit (BioVision, Milpitas, CA, USA) in U-CH1 and JHC7 cells.

Western blot assay
Total protein was extracted by lysing cells in RIPA lysis buffer (KeyGene, Nanjing, China). After measurement of protein concentration, equal amounts of protein (30 μg per lane) were subjected to 10 % SDS-PAGE. After transferring onto the PVDF membranes (Bio-Rad), the membranes were blocked in 5 % skim milk (Beyotime), followed by immunoblotting with primary antibodies at 4℃for 10-14 h. Next, the corresponding secondary antibody (ab205718; 1:4000; Abcam, Cambridge, UK) was applied in combination with primary antibodies. Lastly, the combined signals were visualized with an enhanced chemiluminescence kit (KeyGene). In this study, the primary antibodies were

Dual-luciferase reporter assay
Target prediction was administrated through Starbase 3.0 (https://st arbase.sysu.edu.cn/). The fragments of XIST and 3 ′ UTR of ARF6 containing the miR-320d binding site region and the corresponding mutated region were synthesized and individually cloned into the psiCHECK2 vector (Promega, Madison, WI, USA), generating XIST-WT/MUT or ARF6 3 ′ UTR WT/MUT. Thereafter, U-CH1 and JHC7 cells were cointroduced with the corresponding reporter vector together with miR-320d/miR-NC for 48 h, followed by detection of the luciferase activity using Lastly, Dual-Luciferase Reporter Gene Assay Kit (Yeasen, Shanghai, China).

RNA pull-down assay
Biotinylated miR-320d (Bio-miR-320d) and biotinylated negative control (Bio-NC) were purchased from RiboBio and introduced into U-CH1 and JHC7 cells for 24 h. After lysing by RIPA buffer (Beyotime), cell extracts were then incubated with M− 280 streptavidin magnetic beads (Invitrogen). The RNAs attached to the beads were purified using TRIzol reagent (Invitrogen). Finally, the enrichment of XIST was tested via RT-qPCR after RNA isolation from magnetic beads.

Tumor formation assay
To construct a tumor xenograft model, we purchased BALB/c nude mice from Vital River (Beijing, China). A total of mice (five-week-old; female; n = 10) were segmented into 2 groups (5 mice per group). Approximately 4 × 10 6 U-CH1 cells (control or XIST knockdown) were inoculated into mice by subcutaneous injection. A caliper was utilized to examine tumor volume every 5 days for 25 days. And we calculated tumor volume using the formula: length × width 2 × 1/2. After 25 days, these mice were sacrificed and tumor weight was collected for weighing and further analysis. Animal experiments were conducted with the permission of the Animal Care and Use Committee of Renmin Hospital of Wuhan University. This assay was carried out at least three times.

Statistical analysis
The data were analyzed using GraphPad Prism 7. All data were obtained from at least three repeated experiments and displayed as mean ± standard deviation. The data were analyzed using a Student's ttest or a one-way analysis of variance (ANOVA). The survival curve was analyzed by Kaplan-Meier method. The correlations among XIST, miR-320d and ARF6 were analyzed with Pearson's correlation coefficient. P < 0.05 suggested statistically significant.

XIST expression was increased in chordoma tissues
To determine XIST expression in chordoma tissues and corresponding adjacent tissues, RT-qPCR was performed. We found that the XIST level was enhanced in chordoma tissues in contrast to normal tissues (Fig. 1A). Moreover, chordoma patients with high XIST expression had a poor overall survival than patients with low XIST expression (Fig. 1B). Meanwhile, our data suggested that XIST expression was associated with Tumor size and Invasion condition (p < 0.05) ( Table 1). Next, the localization of XIST was determined in chordoma cells (U-CH1 and JHC7). As shown in Fig. 1C and 1D, XIST was predominantly located in the cytoplasm. These data suggested that XIST might be associated with chordoma progression.

Knockdown of XIST inhibited chordoma cell proliferation, migration, invasion, and glycolysis
To investigate the biological role of XIST in chordoma cells, we performed loss-of-function experiments by transfection of shRNA in U-CH1 and JHC7 cells. Transfection of sh-XIST resulted in a significant downregulation of XIST expression in U-CH1 and JHC7 cells ( Fig. 2A). MTT assay, EdU assay, and colony formation assay showed that XIST silencing repressed cell viability, DNA synthesis, and colony formation ability in U-CH1 and JH7C cells, indicating that XIST knockdown inhibited chordoma cell proliferation (Fig. 2B-2E). Transwell assay indicated that U-CH1 and JHC7 cell migration and invasion were suppressed after downregulating XIST (Fig. 2F and 2G). Next, we explored whether XIST played a critical role in glycolysis in chordoma cells. The ECAR approximated the glycolysis fux rate. We observed a decrease in ECAR in U-CH1 and JHC7 cells transfected with sh-XIST ( Fig. 2H and  2I). Moreover, XIST silencing significantly decreased lactate production (Fig. 2J). Western blot assay indicated that E-cad (an epithelial marker) protein level was increased and N-cad (a mesenchymal marker), GLUT1 and LDHA (glycolysis markers) protein levels were reduced after knockdown of XIST (Fig. 2K). All these data indicated that XIST knockdown inhibited the progression of chordoma cells.

MiR-320d was a direct target of XIST
LncRNAs have been increasingly reported to regulate the expression of target genes via sponging certain miRNAs. Next, we used an online bioinformatics database starBase to predict the potential targets of XIST. As presented in Fig. 3A, miR-320d had putative binding sites for XIST. Transfection of miR-320d increased the expression of miR-320d in U-CH1 and JHC7 cells (Fig. 3B), indicating a high transfection efficiency. Dual-luciferase reporter and RNA pull-down assays were performed to  further confirm whether XIST could directly bind to miR-320d. The results demonstrated that miR-320d overexpression strikingly reduced the luciferase activity of XIST-WT but not XIST-MUT ( Fig. 3C and 3D). Meanwhile, the enrichment of XIST was markedly increased in the Bio-miR-320d group (Fig. 3E). In addition, miR-320d was found to be downregulated in chordoma tissues compared to normal tissues (Fig. 3F). A negative correlation between miR-320d and XIST expression was observed in chordoma tissues (Fig. 3G). Next, the effect of XIST on miR-320d expression was explored. Knockdown of XIST increased miR-320d expression (Fig. 3H). Taken together, these findings demonstrated that miR-320d was targeted by XIST.

Overexpression of miR-320d suppressed the proliferation, migration, invasion, and glycolysis in chordoma cells
To detect the role of miR-320d in chordoma cells, we overexpressed miR-320d expression in chordoma cells. We found that miR-320d upregulation suppressed cell viability, DNA synthesis, colony-forming ability, migration, and invasion in U-CH1 and JHC7 cells (Fig. 4A-4F). Moreover, ECAR and lactate production was decreased by overexpression of miR-320d in U-CH1 and JHC7 cells (Fig. 4G-4I). Further, the upregulation of miR-320d increased E-cad protein expression and reduced the protein levels of N-cad, GLUT1, and LDHA in U-CH1 and JHC7 cells (Fig. 4J). Collectively, miR-320d might play an anti-tumor role in chordoma.

ARF6 was targeted by miR-320d in chordoma cells
As most miRNAs can modulate the expression of their downstream target genes to exert their biological functions, we used starBase to predict the target of miR-320d. We observed that 3 ′ UTR of ARF6 shared binding sites for miR-320d (Fig. 5A), suggesting that ARF6 might be a target for miR-320d. The results of the dual-luciferase reporter assay showed that miR-320d overexpression markedly reduced the luciferase activity of ARF6 3 ′ UTR WT, but did not affect the luciferase activity of ARF6 3 ′ UTR MUT in U-CH1 and JHC7 cells (Fig. 5B and 5C). Then, we explored ARF6 mRNA expression in chordoma tissues, and the data suggested that ARF6 mRNA expression was higher in chordoma tissues than that in normal tissues, and its expression was negatively correlated with miR-320d expression and positively correlated with XIST expression in chordoma tissues (Fig. 5D-5F). ARF6 protein expression was also increased in chordoma tissues (Fig. 5G). Transfection of in-miR-320d markedly reduced miR-320d expression in U-CH1 and JHC7 cells (Fig. 5H), suggesting that miR-320d was successfully inhibited. Next, we explored whether XIST regulated ARF6 expression by sponging miR-320d. The data showed that XIST silencing reduced the protein expression of ARF6, which was reversed by downregulating miR-320d (Fig. 5I), indicating that XIST sponged miR-320d to positively regulate ARF6 expression. The putative binding sites between miR-320d and ARF6 3 ′ UTR were predicted by starBase. (B and C) The interaction between miR-320d and ARF6 was confirmed using dual-luciferase reporter assay in U-CH1 and JHC7 cells. (D) The expression of ARF6 mRNA was detect using RT-qPCR in chordoma tissues and normal tissues. (E and F) The correlation between ARF6 mRNA expression and miR-320d or XIST expression in chordoma tissues was analyzed. (G) The expression of ARF6 protein was examined by western blot assay in chordoma tissues and normal tissues. (H) The expression of miR-320d was tested using RT-qPCR in U-CH1 and JHC7 cells transfected with in-miR-NC or in-miR-320d. (I) ARF6 protein expression was measured by western blot assay in U-CH1 and JHC7 cells transfected with sh-NC, sh-XIST, sh-XIST + in-miR-NC, or sh-XIST + in-miR-320d. **P < 0.01, ***P < 0.001.

XIST silencing repressed malignancy of chordoma cells by regulating ARF6
Transfection of ARF6 increased ARF6 protein expression in U-CH1 and JHC7 cells (Fig. 6A), suggesting a high transfection efficiency. To determine whether XIST regulated chordoma cell behaviors by targeting ARF6, we performed rescue experiments in U-CH1 and JHC7 cells. The inhibitory effects of XIST knockdown on cell viability, DNA synthesis, colony-forming ability, migration, and invasion were overturned by the addition of ARF6 (Fig. 6B-6G). Moreover, ARF6 overexpression counteracted the suppressive effects of XIST downregulation on ECAR and lactate production in U-CH1 and JHC7 cells (Fig. 6H-6 J). In addition, the upregulation of E-cad protein expression and the downregulation of N-cad, GLUT1, and LDHA protein expression caused by XIST silencing was reversed by increasing ARF6 expression ( Fig. 6K and 6L). Overall, these results suggested that ARF6 overexpression abated inhibitory effects of XIST downregulation on chordoma cell proliferation, metastasis, and glycolysis.

XIST downregulation inhibited chordoma tumorigenicity in vivo
Next, we explored whether XIST acted as a tumor promoter in vivo. Significant inhibition of volume and weight was observed after XIST knockdown ( Fig. 7A and 7B). The results of RT-qPCR showed that XIST knockdown reduced the level of XIST and enhanced miR-320d abundance in tumor tissues (Fig. 7C). XIST knockdown inhibited ARF6, Ncad, GLUT1, and LDHA protein expression and increased E-cad protein expression in tumor tissues (Fig. 7D). IHC analysis showed that XIST interference suppressed Ki67 (a proliferation marker) and cytokeratin (a chordoma marker) expression in tumor tissues (Fig. 7E). Collectively, these in vivo data indicated that XIST downregulation could inhibit the tumorigenicity of chordoma.

Discussion
The previous study showed that some lncRNAs are dysregulated in human chordoma, and they might help improve the diagnosis and treatment of chordoma [22,23]. In this work, we demonstrated that XIST was highly expressed in chordoma, and its silencing limited chordoma cell proliferation, metastasis, and glycolysis via regulating the miR-320d/ARF6 axis.
Many studies revealed that abnormal expression of XIST was related to the progression and development of diverse tumors. For example, XIST expression was upregulated in colorectal cancer, and its downregulation repressed colorectal cancer cell proliferation, migration, and invasion [24]. Moreover, XIST acted as an upregulated lncRNA in liver (J) Lactate Assay kit was used to measure lactate production. (K and L) E-cad, N-cad, GLUT1, and LDHA protein expression were detected using western blot assay. **P < 0.01, ***P < 0.001. cancer to promote liver cancer cell growth and metastasis [25]. In addition, XIST could also facilitate the migration and invasion of papillary thyroid cancer cells [26]. By contrast, downregulation of XIST was observed in ovarian cancer cells and tissues, and XIST suppressed ovarian cancer cell proliferation and inversely promoted cell apoptosis [27]. Further, XIST overexpression hampered breast cancer cell progression in vitro and in vivo [28]. These results indicated that XIST played a different role in different cancers. In chordoma, XIST expression is closely correlated with the poor prognosis of the patient, and XIST promoted chordoma cell proliferation and inhibited apoptosis [14]. In our research, XIST upregulation was also observed in chordoma tissues. XIST silencing repressed chordoma cell proliferation, migration, invasion, and glycolysis. Importantly, XIST downregulation inhibited the tumorigenesis of chordoma in vivo. These results disclosed that XIST might act as a tumor facilitator in chordoma.
In recent years, the interaction between miRNAs and lncRNAs has become one of the research hotspots in the study of tumor pathological mechanisms. Given that lncRNAs in the cytoplasm can function as miRNA sponges via interacting with miRNAs to positively regulate the expression and function of downstream target mRNAs [29]. Here, XIST was predominantly located in the cytoplasm. Hence, we assumed that XIST might absorb miRNA. As expected, XIST was a sponge for miR-320d. Previous studies have indicated that miR-320d served as a tumor suppression miRNA in some cancers, and its downregulation is closely related to the poor prognosis of patients [30,31]. In chordoma, miR-320d was lowly expressed, and miR-320d inhibition increased the growth, invasion, and migration of chordoma cells [20]. Consistently, we also uncovered that miR-320d content was declined in chordoma tissues. Functional experiments showed that miR-320d upregulation could restrict the growth, metastasis, and glycolysis of chordoma cells, suggesting that miR-320d has an anti-tumor role in chordoma.
MiRNAs can bind to target mRNAs to regulate many biological and pathological processes [32]. Next, we analyzed the downstream target of miR-320d. Our data proved that ARF6 could be targeted by miR-320d. ARF6 is a member of the ARF protein family of small GTPases, which plays a critical role in pathological and physiological processes [33,34]. In addition, the available evidence indicated that ARF6 acted as an oncogene in many cancers to promote the proliferation invasion, and migration of cancer cells [35,36]. ARF6 is also believed to be implicated in the glycolysis of cancer cells [37]. Importantly, an increase in ARF6 expression was found in chordoma tissues, and ARF6 overexpression reversed the suppressive effect of MDFIC-7 silencing on proliferation and aerobic glycolysis of chordoma cells [21]. Here, we also observed upregulation of the ARF6 expression in chordoma tissue samples. Overexpression of ARF6 could abate the suppressive influence of XIST deficiency on chordoma cell growth, metastasis, and glycolysis. Mechanistically, XIST positively modulated the ARF6 level via sponging miR-320d. Compared with previous articles [14], the novelty of this study lies in the confirmation that XIST downregulation might inhibit tumor migration, invasion, and glycolysis, and the construction of a novel network mechanism of the XIST-miR-320d-ARF6 axis. These data indicated that XIST served as a sponge of miRNAs to regulate the more comprehensive lncRNA-miRNA-mRNA co-expression network.
In conclusion, our results proved that XIST knockdown suppressed chordoma cell growth, metastasis, and glycolysis as well as inhibited tumor growth in vivo by regulating the miR-320d/ARF6 axis. This is the first report to demonstrate that the XIST miR-320d/ARF6 axis contributed to the tumorigenesis of human chordoma. These results might offer potential targets for chordoma therapy and help understand the mechanisms of chordoma progression.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.