LncRNA MALAT1/miR‐129 axis promotes glioma tumorigenesis by targeting SOX2

Abstract We aimed to explore the interaction among lncRNA MALAT1, miR‐129 and SOX2. Besides, we would investigate the effect of MALAT1 on the proliferation of glioma stem cells and glioma tumorigenesis. Differentially expressed lncRNAs in glioma cells and glioma stem cells were screened out with microarray analysis. The targeting relationship between miR‐129 and MALAT1 or SOX2 was validated by dual‐luciferase reporter assay. The expressions of MALAT1, miR‐129 and SOX2 mRNA in both glioma non‐stem cells and glioma stem cells were examined by qRT‐PCR assay. The impact of MALAT1 and miR‐129 on glioma stem cell proliferation was observed by CCK‐8 assay, EdU assay and sphere formation assay. The protein expression of SOX2 was determined by western blot. The effects of MALAT1 and miR‐129 on glioma tumour growth were further confirmed using xenograft mouse model. The mRNA expression of MALAT1 was significantly up‐regulated in glioma stem cells compared with non‐stem cells, while miR‐129 was significantly down‐regulated in glioma stem cells. MALAT1 knockdown inhibited glioma stem cell proliferation via miR‐129 enhancement. Meanwhile, miR‐129 directly targeted at SOX2 and suppressed cell viability and proliferation of glioma stem cells by suppressing SOX2 expression. The down‐regulation of MALAT1 and miR‐129 overexpression both suppressed glioma tumour growth via SOX2 expression promotion in vivo. MALAT1 enhanced glioma stem cell viability and proliferation abilities and promoted glioma tumorigenesis through suppressing miR‐129 and facilitating SOX2 expressions.

Long non-coding RNAs (lncRNAs), a group of non-protein coding transcripts longer than 200 nucleotides, play a key role in many types of malignant tumours, including glioma. 8 LncRNAs may function as competing endogenous RNAs (ceRNAs) to modulate miRNA expressions. 9 Metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) is an lncRNA with a length of approximately 8000-nt. 8 Wang et al revealed that MALAT1 was aberrantly expressed in carcinoma cells. 10 Ma et al 11 found that MALAT1 was up-regulated in glioma tissues and correlated with the progression of glioma. However, the molecular mechanism of MALAT1 in glioma cells remains unknown.
MicroRNAs (miRNAs) are short single-stranded RNA molecules which function as a critical modulator in gene expression by wholly or partially binding to corresponding mRNAs. 12 MiRNAs play a crucial role in the progression of various cancers, such as osteosarcoma 12 and glioma. 13 MiR-129 is a miRNA family mainly composed of miR-129-3p, miR-129-2-3p and miR-129-5p. 14 Previous researches have already demonstrated that miR-129-5p acted as a tumour inhibitor in multiple cancers such as ovarian cancer, 15 breast cancer 16 and also glioma. 17 Chen et al 18 found that Notch-1/ E2F7/Beclin-1 axis was regulated and further the viability of malignant glioma cells could be impaired by up-regulation of miR-129. However, the role of miR-129 in glioma stem cells has not been elucidated so far. SRY (sex determining region Y)-box 2 (SOX2) is revealed as a stemness marker in glioma 19 and a transcription factor highly associated with pluripotency. 20 Current study investigated the relationship between SOX2 and miR-129 to further elucidate the molecule network of miR-129 in glioma stem cells progression. Activated SOX2 expression maintained stemness and self-renewal of GSCs. 21

SOX2
is associated with GBM stem-like phenotype which is more resistant to c-radiation. 22 Therefore, SOX2 expression changes are significant for GSC stemness maintainment.
This study hypothesized the promoter role of lnc MALAT1 in glioma and validated its effect in both in vivo and in vitro experiments.
MALAT1 was suspected to bind to miR-129 which target at SOX2, an oncogene, and their interaction was depicted in this study. By illustrating the underlying mechanism that facilitated glioma progression, this study may contribute to the application of glioma target therapy. College, Huazhong University of Science and Technology. Inclusion criteria were the following: (i) WHO graded gliomas confirmed by histopathology. Exclusion criteria were as follows: (i) an unknown IDH1-mutation status, (ii) patients with a previous history of brain tumours, (iii) patients younger than 18 years of age. All glioma tissues were first preserved in liquid nitrogen and stored at À80°C for the subsequent experiment. This project was ratified by the Ethics Committee of the Union Hospital, Tongji Medical College, Huazhong University of Science and Technology and written informed consents were collected from all patients.

| Glioma cell culture and stem cell isolation
Glioma cells were cultured in Dulbecco's modified Eagle medium (DMEM)/high glucose with 10% fetal bovine serum (FBS, Gibco, Carlsbad, CA, USA) and were maintained in a humidified incubator at 37°C with 5% CO 2 . GSCs were separated as described previously from tissue cells. 23,24 Primary cells were detached with trypsin, washed once in FACs buffer (PBS containing 1%-2% BSA and 5 mM EDTA), then stained with anti-CD24-FITC (Invitrogen, Carlsbad, CA, USA) and anti-CD133-PE (Invitrogen) using 10 ll of antibody per 10 6 cells, and incubated at 4°C for 15 minutes. Following incubation, cells were washed once with FACs buffer. The CD44 + CD24 À cells were considered to be GSCs.

| Microarray analysis
GSE23806 microarray data obtained from Gene Expression Omnibus database was applied to the filtration of aberrantly expressed lncRNAs. "Match matrix" file from GEO website (https://www.ncbi. nlm.nih.gov/geo/query/acc.cgi?acc=GSE23806) were downloaded and processed for samples and corresponding probes. Corresponding Human Genome U133 Plus 2.0 Array) was downloaded and processed as well for matching probes and corresponding gene symbols. 14 conventional glioma cell lines (G121,   G84, G120, G61, G112, G130, U118MG, SF268, G124, G118, G142, G44, SW1783 and G22) and 5 glioma stem-like cell lines (GS01, GS02, GS03, GS04 and GS05) were analysed using unpaired T-test method in the LIMMA package and the obtained P values were adjusted with Benjamini-Hochberg method. A volcano plot filtering (fold change > 4, adjusted P-value < .01) was drawn and all the differentially expressed lncRNAs were listed in the heat map.

| Western blot
The isolation of total protein from tissues and protein concentration measurement were performed by using RIPA lysis buffer and BCA protein concentration kit (Beyotime, Shanghai, China). Proteins were segregated by SDS-PAGE and then transferred onto PVDF membranes (Invitrogen) in accordance with the instructions, followed by blocking in 5% nonfat milk for 60 minutes and incubation with primary antibodies against SOX2 (ab137385, 1:1000, Abcam corporation) at 4°C overnight. After that, the membranes were washed with Tris Buffered Saline Tween (TBST) every 5 minutes for four times and then incubated in HRP-conjugated goat anti-rabbit IgG secondary antibody (1:2000) for 2 hours. After being rinsed twice in TBST, the immunoreactive bands were developed using an enhanced chemiluminescence detection system (Amersham Pharmacia Biotech, Buckinghamshire, England).

| Cell transfection
Glioma stem cells in logarithmic growth were first seeded in the culture dish, and then placed onto the 6-well plate when cell growth reached 80%-90% confluence. MALAT1 siRNA, miR-129 mimics and miR-129 inhibitor were synthesized by Genepharma Company  and pmir-GLO-SOX2 was constructed. Glioma stem cells were seeded into the 24-well culture plate until 90% confluency. Then Recombinant vector (wild-type or mutated type) was transfected into the cells together with miR-129 mimics or mimics control by using Lipofectamine 2000 reagent, followed by incubation for 48 hours. After that, the fluorescence intensity of transfected cells was examined by luciferase reporter assay kit (Promega, Madison, WI, USA).

| EdU assay
Transfected glioma stem cells were cultured in 96-well plates.
Briefly, glioma stem cells were incubated with EdU labelling medium at moderate concentration for 2 hours. The cells were then fixed with 0.5% TritonX-100 in PBS (100 lL) for 25 minutes, and stained with 100 lL Apollo dye solution (Ribobio) for 30 minutes at room temperature. The cells were subsequently stained using DAPI (Invitrogen) and incubated for half an hour. The percentage of EdU positive cells was calculated using ImageJ software.

| Sphere formation assay
Glioma stem cells were seeded into 6-well plate at 5 9 10 3 cells/ mL in the culture medium containing 20 ng/mL basic Fibroblast Growth Factor (bFGF) and Epidermal Growth Factor (EGF), 5 L g/ mL insulin (Sigma-Aldrich,St. Louis, MO, USA), 0.4% BSA (Invitrogen, USA) medium and 0.02% B27 (Invitrogen). After incubation for 7 days, cells were fixed using 10% formalin and photographed under a conventional microscope. Sphere Formation Efficiency (SFE) representing the ability of sphere formation (diameter > 75 lm) was calculated. The formula of SFE was: the numbers of cell sphere in each well / the total number of cells originally seeded in each well.

| Statistical analysis
Statistical analyses were performed through GraphPad Prism 6.0 (GraphPad Software, San Diego, California, USA). All quantitative data are presented as means AE standard deviation (SD). Differences between groups were compared using Student's t-test while comparison in multiple groups was practiced with one-way ANOVA. Benjamini-Hochberg was used to adjust multiple testing for selection of differentially expressed lncRNAs. Statistical significance was based on P value < .05.
The results of fold-change and P value of differential genes  Figure 1A). Microarray analysis results revealed that MALAT1 was significantly up-regulated in glioma stem-like cells. To validate the prediction, we used 14 patient tissue samples to isolate GBCs. Those CD133 + CD24 À cells were regarded as GSCs as reported previously. 25,26 Characteristics of patients were listed in Table 3. After we obtained glioma stem cells from glioma cells, qRT-PCR was utilized to measure the expression level of MALAT1.
As shown in Figure 1B,C, MALAT1 expression in glioma stem cells was remarkably higher than compared with non-stem cells (P < .0001).

| MiR-129 restrains the propagation of glioma stem cells
The results of qRT-PCR indicated that miR-129 expression in glioma stem cells was remarkably lower than that in non-stem cells ( Figure 4A,

| SOX2 is a direct target of miR-129
Previous studies ever identified that SOX2 was required for cancer cell line progression in lung and esophageal squamous carcinoma, highlighting the importance of SOX2 as a lineage-survival oncogene. 27 High SOX2 expression has been associated with several human solid tumours, including glioma. 28 The present study evaluated that expression of SOX2 mRNA in glioma stem cells was significantly higher than that in glioma cells detected by qRT-PCR ( Figure 5A, P < .001). Tar-getScan predicted the direct targeting relationship between miR-129 and SOX2 and the dual-luciferase reporter assay validated their direct targeting relationship since the luciferase activity in SOX2-WT + miR-129 mimics group was remarkably weaker compared with that in NC group ( Figure 5B,C, P < .001). Hence, SOX2 was a direct target of miR-129. Furthermore, miR-129 mimic suppressed SOX2 expression, while miR-129 inhibitor enhanced SOX2 expression ( Figure 5D, P < .01). In addition, after being transfected with MALAT1 siRNA (si-MALAT1#1 or si-MALAT1#2), the mRNA and protein expressions of SOX2 were significantly down-regulated ( Figure 5E, P < .01). We thus confirmed negative regulation between miR-129 and SOX2 and positive regulation between MALAT1 and SOX2.

| The down-regulation of MALAT1 expression and miR-129 overexpression suppresses glioma tumour growth in vivo
Xenograft mice models were established by subcutaneously injecting glioma stem cells with different expression of MALAT1 and miR-129. The tumour size was measured every 7 days. At the 28th day, mice were sacrificed and tumours were obtained and weighed. The results disclosed that the tumour volume and weight in sh-MALAT1 and AgomiR-129 groups significantly reduced in comparison with control group (Figure 6A Earlier studies have reported that miR-129-2 suppressed glioma by targeting HMGB1 in glioma in an DNA methylation way. 33 MiR-129 also impair human malignant glioma progression via autophagic flux enhancement by regulating a novel Notch-1/ E2F7/Beclin-1 axis. 18 Kang et al 34 revealed that miR-129-2 suppressed proliferation of esophageal carcinoma cells through down-regulation of SOX4 expression. Elevated SOX2 enforces glioblastoma stem cell identity. 35 High miR-21/low SOX2 axis was capable of classifying patients with longer survival. 36 We also diagnosed miR-129 as a tumour suppressor via SOX2 modulation firstly in glioma tumorigenesis. Besides, the MALAT1/miR-129/SOX4 was of great significance in present molecular system investigation.
As an important pluripotent marker of stem cells, SOX2 has been recognized as serving a crucial role in maintaining the properties of cancer stem cells. 37 According to the study of Wang et al 38 SOX2 was considered as a predictor of survival in gastric cancer to inhibit cell proliferation and metastasis. Chen et al also revealed that silencing of SOX2 could regulate the apoptosis rate of human lung cancer cells. 39 In our research, we speculated that the interaction between miR-129 and SOX2 was associated with the suppression of glioma tumorigene- F I G U R E 4 MiR-129 suppresses the proliferation of glioma stem cells. A, The expression of miR-129 in glioma stem cells was significantly lower than that in the glioma tissue cells detected by qRT-PCR. B, The expression of miR-129 in miR-129 mimics group dramatically increased, while that in miR-129 inhibitor group significantly decreased observed by qRT-PCR. C-E, CCK-8 assay and EdU assay showed that cell proliferation of glioma stem cells were significantly attenuated in miR-129 overexpression group but remarkably enhanced in miR-129 inhibitor group compared with NC group. F-G, The sphere formation assay results also revealed that the sphere formation efficiency of glioma stem cells in miR-129 mimics group was significantly lower than that in NC group, while the cells transfected with miR-129 inhibitor presented higher sphere formation efficiency in comparison with NC group determined by sphere formation assay. Scale bar, 100 lm. **P < .01. ***P < .001, ****P < .0001, compared with NC group F I G U R E 5 SOX2 is a direct target of miR-129. A, The expression of SOX2 mRNA in glioma stem cells was significantly higher than that in glioma tissue cells detected by qRT-PCR. ***P < .001, compared with cancer cells (glioma tissue cells). B, SOX2 was a candidate target of miR-129 predicted by TargetScan. C, The targeted relationship between miR-129 and SOX2 was validated by dual luciferase reporter gene assay. D, The expression of SOX2 in glioma stem cells transfected with miR-129 mimics remarkably decreased, while glioma stem cells transfected with miR-129 inhibitor presented a significant increase in the expression of SOX2. E, Compared with NC group, SOX2 mRNA and protein expression drastically decreased after transfected with si-MALAT1#1 and si-MALAT1#2 detected by qRT-PCR and Western blot respectively. **P < .01, ***P < .001, all compared with NC group F I G U R E 6 Either MALAT1 down-regulation or miR-129 overexpression suppresses glioma tumour growth in vivo. A-C, The tumour volume and weight of mice transfected with sh-MALAT1 and AgomiR-129 significantly reduced compared with control group. D, The expression of SOX2 protein in mice transfected with sh-MALAT1 or AgomiR-129 was significantly down-regulated. **P < .01, compared with control group

| CONCLUSION
In summary, our research showed that MALAT1 was up-regulated in glioma stem cells and acted as a tumour promoter in glioma progression. Silencing of MALAT1 suppressed the proliferation and stemness of GSCs and in vivo tumour growth via up-regulating miR-129 and further inhibiting SOX2, providing a promising therapy target for the treatment of glioma.

CONF LICT OF I NTEREST
The authors confirm that there are no conflicts of interest.

AUTHOR CONTRIBU TI ON
Zhiyong Xiong and Luyang Wang contributed to research conception and design as well as manuscript drafting; Ye Yuan analysed and interpreted data; Luyang Wang made statistical analysis; Qiangping Wang revised the manuscript and took the role of funding collectors.
In addition, all authors approved final manuscript.