LncRNA MIR4435‐2HG potentiates the proliferation and invasion of glioblastoma cells via modulating miR‐1224‐5p/TGFBR2 axis

Abstract Glioblastoma (GBM) belongs to the high‐grade (IV) gliomas with extremely poor prognosis. Accumulating evidence uncovered the key roles of long non‐coding RNAs (lncRNAs) in GBM development. This study aimed to determine the biological actions and the clinical relevance of lncRNA MIR4435‐2 Host Gene (MIR4435‐2HG) in GBM. Data from GEPIA database showed that MIR4435‐2HG was up‐regulated in GBM tissues and high expression of MIR4435‐2HG correlated with shorter overall survival of GBM patients. Further experimental assays verified the up‐regulation of MIR4435‐2HG in GBM tissues and cell lines. In vitro cell studies and in vivo animal studies showed that knockdown of MIR4435‐2HG resulted in the inhibition of GBM cell proliferation and invasion and in vivo tumour growth, while MIR4435‐2HG overexpression driven GBM progression. Furthermore, MIR44435‐2HG was found to sponge miR‐1224‐5p and suppress miR‐1224‐5p expression; overexpression of miR‐1224‐5p attenuated the enhancement in GBM cell proliferation and invasion induced by MIR4435‐2HG overexpression. In a subsequent study, miR‐1224‐5p was found to target transforming growth factor‐beta receptor type 2 (TGFBR2) and repressed TGFBR2 expression, and in vitro assays showed that miR‐1224‐5p exerted tumour‐suppressive effects via targeting TGFBR2. More importantly, TGFRB2 knockdown antagonized hyper‐proliferation and invasion of GBM cells with MIR4435‐2HG overexpression. Clinically, the down‐regulation of miR‐1224‐5p and up‐regulation of TGFBR2 were verified in the GBM clinical samples. Taken together, the present study suggests the oncogenic role of MIR4435‐2HG in GBM and underlies the key function of MIR4435‐2HG‐driven GBM progression via targeting miR‐1224‐5p/TGFBR2 axis.


| INTRODUC TI ON
Glioma is one of the most common types of brain tumours and belongs to astrocytic tumours. 1 Glioma can be divided into four grades (I, II, III and IV) based on the World Health Organization (WHO) classification. 2 Low-grade glioma belongs to WHO grades I and II, while high-grade glioma belongs to WHO grades III and IV. 2 Glioblastoma belongs to the WHO IV glioma. 2 Patients with glioblastoma without developing from low-grade glioma were diagnosed as primary glioblastoma; while secondary glioblastoma was developed from low-grade glioma. 3,4 Though various therapeutic strategies including surgery, chemotherapy and radiotherapy have been developed, the heterogeneity of glioblastoma cells makes the tumour cells be less sensitive to chemo/radiotherapy. 5,6 Thus, the outcomes of the current treatments for glioblastoma are un-satisfying and the prognosis is extremely poor in the patients with glioblastomas. 7 Therefore, there is an urgent need for the scientific community to develop novel therapies for the treatment of glioblastoma.
Long non-coding RNAs (lncRNAs) belong to a class of the non-coding RNAs, and lncRNAs are longer than 200 nucleotide base pairs in length and are lack of capacity for coding proteins. 8 According to the genomic studies, the human genome has more than 50,000 lnRNA genes, and the functional roles of these lncRNAs are poorly studied. 9 Up to date, lncRNAs have been elucidated for their roles in various aspects of cellular functions including apoptosis, proliferation, metastasis, and stemness, which are closely related to pathogenesis of various diseases including glioblastoma. [10][11][12] For examples, Voce et al, found that lncRNA metastasis-associated lung adenocarcinoma transcript-1 (MALAT1) was a target for GBM chemo-sensitization to temozolomide, though the up-regulation of MALAT1 in GBM tissues was not a prognostic factor for the overall survival of GBM patients. 13 Wu et al, showed that lncRNA MIR155 host gene/miR-185/annexin A2 loop had a regulatory role in the GBM progression, which underlies the importance of lncRNA MIR155 host gene in the GBM pathophysiology. 14 Liu et al, found that LINC00470 epigenetically regulated the expression of extracellular leucine rich repeat and fibronectin type III domain containing 2 to distract cell autophagy in GBM. 15 Recent studies suggest the lncRNA MIR4435-2 Host Gene (MIR4435-2HG) may involve in the regulation of brain tumour progression. Specifically, MIR4435-2HG encodes a nearly identical hairpin of LINC00152, while LINC00152 was associated with aggressive tumours and the enhanced proliferation of GBM cells. 16 In addition, MIR4435-2HG exerted the oncogenic functions in other types of malignant tumours including hepatocellular carcinoma, 17 colorectal cancer, 18 gastric cancer 19 and lung cancer. 20 To date, the underlying mechanisms of MIR4435-2HG in GBM progression has yet to be explored.
Here, we showed that MIR4435-2HG is clinically relevant in GBM, as MIR4435-2HG is up-regulated in GBM tissues and high expression of MIR4435-2HG is associated with poor overall survival rate of GBM patients. Furthermore, MIR4435-2HG knockdown leads to the inhibition of GBM cell proliferation and invasion; while MIR4435-2HG overexpression drives the GBM progression. The mechanistic studies revealed that MIR4435-2HG exerts its actions via targeting miR-1224-5p/ transforming growth factor-beta receptor type 2 (TGFBR2) axis in GBM. Together, the present study underlies the key function of MIR4435-2HG-driven GBM progression and brings forth MIR4435-2HG as a therapeutic target for this malignant tumour.

| Clinical sample collection
A total of 40 patients diagnosed with GBM were recruited in this study, and this study was approved by the Ethics Committee of Shenzhen People's Hospital, and all the patients signed the written informed consent. The GBM tissues as well as the adjacent normal tissues were collected from these patients who underwent surgical resection for brain tumours between January 2016 and December 2018 in the Shenzhen People's Hospital. The GBM patients had not received any chemotherapy or radiotherapy before surgery. All the clinical samples were collected into liquid nitrogen and stored in −80°C for further examination.

| Cell culture
The normal human astrocytes (NHAs) were obtained from Lonza and were cultured in ABM growth medium (ABM ™ , Lonza). The GBM cell lines (LN229, U87MG, U87 and U251) were all obtained from ATCC and were cultured in DMEM (ATCC ® 30-2002 ™ ) supplemented with 10% antagonized hyper-proliferation and invasion of GBM cells with MIR4435-2HG overexpression. Clinically, the down-regulation of miR-1224-5p and up-regulation of TGFBR2 were verified in the GBM clinical samples. Taken together, the present study suggests the oncogenic role of MIR4435-2HG in GBM and underlies the key function of MIR4435-2HG-driven GBM progression via targeting miR-1224-5p/TGFBR2 axis.
The miRNAs including miR-1224-5p mimics and the NC for mimics (mimics NC) were purchased from Ribobio. The short-hairpin RNAs (shRNAs) that target MIR4435-2HG as well as the scrambled shRNA were packaged in lentivirus and were infected into GBM cells to generate the stable GBM cell lines for MIR4435-2HG knockdown.
All the GBM cell transfections were conducted using Lipofectamine 2000 reagent (ThermoFisher Scientific).

| Cell proliferation assay
The cell proliferative ability of U87 and U251 cells were determined by the Cell Counting Kit-8 (CCK-8; Beyotime) assay. In brief, the transfected U87 and U251 cells were seeded onto the 96-well plates, and after a further culture for indicated time durations (0, 24, 48 and 72 hours), the transfected U87 and U251 cells were subjected to incubation with 10 μL CCK-8 solution for 2 hours at 37°C. The optical density was examined at 450 nm wavelength using a Microplate Reader (Bio-Rad) to determine cell proliferative index.

| Colony formation assay
The cell growth of U87 and U251 cells were assessed using colony formation assay. In brief, the transfected U87 and U251 cells were collected and inoculated into the 6-well plates, and GBM cells were then cultured for 10 days. The cultured medium was refreshed every 3 days. At 10 d after inoculation, the colonies were visualized by staining the GBM cells with 0.1% crystal violet, and the number of colonies was counted with a light microscope.

| Cell invasion assay
Cell invasive ability was determined using the transwell invasion assay. In brief, the transwell inserts with 8 µm pore size filters (Millipore) were coated with Matrigel (Sigma), and the transfected U87 and U251 cells were seeded onto the upper chamber with transwell inserts. The upper chamber was filled with serum-free medium, while lower chamber was filled with medium containing 20% FBS. The transfected U87 and U251 cells were cultured for 24 hours, after that, the cells invaded into the lower filter membranes were fixed with 100% methanol and stained with 0.1% crystal violet for the quantification of the invasive cell number.

| Cell apoptosis assays
Cell apoptotic rates of the transfected U87 and U251 cells were determined on a flow cytometer using the Annexin V-FITC/propidium iodide Cell Apoptosis Detection kit (Thermo Fisher Scientific) by following the manufacturer's protocol.
Capase-3 activities of the transfected U87 and U251 cells were evaluated using the Caspase-3 Activity Assay kit (Thermo Fisher Scientific) by following the manufacturer's protocol.

| In vivo tumorigenicity assay
The male athymic BALB/c mice (4-5 weeks old) were purchased from Vital River Laboratories. All the animal experimental procedures were approved by the Animal Ethics Committee of Shenzhen People's Hospital. U87 cells with stable MIR4435-2HG overexpression or knockdown were injected subcutaneously into the right flank of the animals (each group had 5 mice). After injection, the tumour volume was measured from day 7 to day 35 with a 7-day internal.
The tumour volume was measured using the following formula: volume = length × width × width/2. The mice were euthanized at 35 days after cell injection, and the tumours were taken out for further determinations.

| Western blot assay
Proteins from transfected cells were extracted using RIPA buffer (Sigma) supplemented with Protease Inhibitor Cocktail (Roche). The protein concentrations of the lysed samples were measured using the bicinchoninic acid assay (Bio-Rad). The proteins were electrophoresed on a 10% SDS-PAGE and were then transferred to the polyvinylidene difluoride (PVDF) membranes. After incubating with 5% skimmed milk for 1 hour at room temperature, the PVDF membranes were probed with primary antibodies against TGFBR2 (Cell Signaling Technology) and β-actin (Cell Signaling Technology) by a further overnight incubation at 4°C. Afterwards, the PVDF membranes were washed with phosphate buffered saline with Tween-20 for 3 times × 5 mins and were then incubated with horseradish peroxidase-labelled secondary antibodies (Cell Signaling Technology). The Western blot signals were visualized using the ELC Substrates (ThermoFisher Scientific).

| Statistical analysis
SPSS software V20.0 (IBM, Armonk, USA) and GraphPad Prims V6.0 (GraphPad Software) were used to perform data analysis. In each experiment, triplicate samples were used for experimental assays.
The data were shown as mean ± standard deviation. Statistical comparison between groups were analysed using unpaired Student's t test or one-way ANOVA followed with Bonferroni's multiple comparison tests. Correlation between two variables were determined using Pearson's Correlation analysis. P < .05 was accepted as statistical significance.

| MIR4435-2HG is up-regulated in GBM tissues
The expression of MIR4435-2HG was first analysed from the GEPIA database (http://gepia.cance r-pku.cn/); based on the database, the expression of MIR4435-2HG was compared between 207 normal tissues and 163 GBM tissues. As shown in Figure 1A, MIR4435-2HG expression was significantly up-regulated in the GBM group compared to that in the normal group. The overall survival data from GEPIA database showed that GBM patients with high expression of MIR4435-2HG had shorter overall survival rate than F I G U R E 1 MIR4435-2HG is up-regulated in GBM tissues. A, MIR4435-2HG was up-regulated in GBM tissues compared to normal tissues as determined from GEPIA database. B, GBM patients with high expression of MIR4435-2HG has shorter overall survival rate than that with lower expression of MIR4435-2HG as determined from GEPIA database. C, MIR4435-2HG in clinical samples was determined by qRT-PCR, and MIR4435-2HG was up-regulated in GBM tissues from 40 patients compared to normal adjacent tissues. *P < .05 that with low expression of MIR4435-2HG ( Figure 1B). To confirm the results from GEPIA database, we analysed the expression of MIR4435-2HG in GBM tissues and adjacent normal tissues from 40 patients using qRT-PCR, and MIR4435-2HG was up-regulated in GBM tissues from 40 patients compared to normal adjacent tissues ( Figure 1C).

| Knockdown of MIR4435-2HG inhibited GBM cell proliferation and invasion and in vivo tumour growth
The

| MIR4435-2HG acts as a sponge for miR-1224-5p
The starBase tool was utilized to predict the potential miRNAs for MIR4435-2HG and the prediction results showed that miR-1224-5p had a binding site for MIR4435-2HG ( Figure 4A). The results from qRT-PCR assay showed that miR-1224-5p was down-regulated in LN229, U87MG, U87, and U251 cells compared to NHA cells ( Figure 4B). The findings from the luciferase report assay showed that the luciferase activity of MIR4435-2HG-WT was suppressed by transfecting with miR-1224-5p mimics in U87 cells ( Figure 4C,D), while MIR4435-2HG-Mut

| TGFBR2 is a direct target of miR-1224-5p
To further predicted the targets for miR-1224-5p, we employed the starBase tool to predict the potential targets and found that TGFBR2 3'UTR had a binding site for miR-1224-5p ( Figure 5A).
The results from qRT-PCR assay showed that TGFBR2 mRNA was up-regulated in LN229, U87MG, U87, and U251 cells compared to NHA cells ( Figure 5B

| MIR4435-2HG regulated proliferation, growth and invasion of U87 cells via regulating TGFBR2
To determine if MIR4435-2HG regulated GBM progression via modulating TGFBR2, we further performed the rescue experiments. The down-regulation of TGFBR2 were detected in U87 cells with TGFBR2 siRNA transfection ( Figure 6A,B). The CCK-8 assay revealed that TGFBR2 knockdown antagonized MIR4435-2HG overexpression-induced an enhancement in U87 cell proliferation and growth ( Figure 6C,D). Additionally, TGFBR2 inhibition reversed the increased cell invasive number induced by MIR4435-2HG overexpression in U87 cells ( Figure 6E).

| MIR-1224-5p and TGFBR2 mRNA expression in clinical tissues
The expression of miR-1224-5p and TGFBR2 in clinical tissues was verified with qRT-PCR. As shown in Figure 7A,B, miR-1224-5p was down-regulated in GBM tissues from 40 patients compared to normal adjacent tissues, while TGFBR2 mRNA was up-regulated in GBM tissues from 40 patients compared to normal adjacent tissues ( Figure 7A,B).Moreover, the MIR4435-2HG expression level was inversely correlated with miR-1224-5p expression level, but was positively correlated with TGFBR2 mRNA expression level in the GBM tissues ( Figure 7C,D).

| D ISCUSS I ON
In the present study, to explore the clinical relevance of MIR4435-2HG in GBM, we firstly derived the MIR4435-2HG data for GBM from GEPIA database and found that MIR4435-2HG was up-regu- The correlation between MIR4435-2HG and miR-1224-5p expression levels and (D) the correlation between MIR4435-2HG and TGFBR2 mRNA expression levels in GBM tissues were analysed by Pearson's correlation test. *P < .05 and **P < .01 protein kinases and vascular endothelial growth factor pathways. 18 MIR4435-2HG was highly expressed in lung cancer tissues and driven the progression of lung cancer via enhancing the β-catenin signalling. 20 In hepatocellular carcinoma, MIR4435-2HG was up-regulated in the tumour tissues and exerted the oncogenic actions in hepatocellular carcinoma cells via modulating miR-487a expression. 17 In our study, we consistently showed that MIR4435-2HG was up-regulated in GBM tissues and cell lines; knockdown of MIR4435-2HG lead to a decreased activity in GBM cell proliferation and invasion and in vivo tumour growth, while MIR4435-2HG overexpression caused the opposite effects, suggesting oncogenic role of MIR4435-2HG in GBM.
The lncRNAs have been well-studied for mechanistic actions by interacting with different miRNAs, and this action usually leads to miRNA expression repression. 22 In this study, starBase online prediction tool was utilized and revealed that miR-1224-5p had a binding site for MIR4435-2HG, which was verified by luciferase reporter assay. Additionally, MIR4435-2HG suppressed the miR-1224-5p expression in U87 cells. MiR-1224-5p was identified as tumour suppressor in several types of cancers. MiRNA profiling results showed that miR-1224-5p was down-regulated in lung cancer tissues. 23 Down-regulation of miR-1224-5p was associated with poor overall survival of patients with metastatic colorectal cancer. 24 In the aspect of lncRNA-miRNA interaction, the miR-1224-5p suppressed melanoma progression and was repressed by lncRNA zinc finger E-boxbinding homeobox 2 antisense RNA 1. 25 Additionally, miR-1224-5p was tumour-suppressive in bladder cancer and was negatively regulated by circular RNA hsa_circ_0075828. 26 In gliomas, miR-1224-5p level was decreased in GBM when compared to low-grade gliomas, and miR-1224-5p overexpression suppressed GBM cell proliferation and induced apoptosis via targeting cAMP responsive element binding protein 1. 27 Consistently, we showed that miR-1224-5p overexpression inhibited U87 cell proliferation and invasion and counteracted the GBM progression potentiated by MIR4435-2HG, suggesting the MIR4435-2HG-mediated GBM progression involved in the regulation of miR-1224-5p.
To have a deeper understanding into the regulatory network of MIR4435-2HG in GBM, we utilized starBase tool to further predict the targets of miR-1224-5p, and identified TGFBR2 as the target of miR-1224-5p. The luciferase reporter assay confirmed the interaction between miR-1224-5p and TGFBR2 3'UTR, while in vitro assays showed that TGFBR2 expression was repressed by miR-1224-5p overexpression while being promoted by MIR4435-2HG. TGFBR2 is a key receptor in mediating the tumour growth factor-beta signal propagation, and knockdown of TGFBR2 was effective in suppressing GBM invasion via a tumour growth factor-beta-dependent manner. 28 In addition, TGFBR2 has also been identified as a novel regulator of GBM stemness, 29 which may be related to the platelet-derived growth factor receptor inhibitor resistance in GBM treatment.

AUTH O R CO NTR I B UTI O N S
HX and JM participated in the design of the study and performed the experiments. ZL, BZ and PZ performed the data analysis. HX, PZ, WW, and HZ participated in the development of the experimental design and drafted the manuscript. HX and JM were major contributors to the design of this study and also revised the manuscript. All authors read and approved the final manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets used and/or analysed in this study are available from the corresponding author upon reasonable request.