Marsdenia tenacissima extract prevents the malignant progression of glioma through upregulating lncRNA MEG3 and SFRP1‐dependent inhibition of Wnt/β‐catenin pathway

Abstract Background/Aim Recent studies have highlighted the tumor‐suppressive effect of Marsdenia tenacissima extract (MTE) on human cancers. This research unveils the potential impact of MTE on glioma and ascertains the relevant molecular mechanisms. Methods Glioma cells were treated with MTE, with normal human astrocytes (NHAs) as controls. A battery of function experiments, including the CCK‐8 viability test, colony formation assay, scratch migration assay, and Transwell invasion assay, was executed to address the responses of glioma cells to MTE treatment and gain or loss of function of lncMEG3, miR‐542‐3p, and SFRP1. FISH, RIP, and dual‐luciferase reporter assays were adopted for assessing gene interactions. U251‐GFP‐Luc cells were delivered into nude mice through intracranial injection to develop an orthotopic glioma model for in vivo validation. Results 200 mg/mL MTE could suppress the proliferating, migrating, and invading properties of glioma cells but not affect those of NHAs. MTE treatment enhanced the expression of lncMEG3, which competes with SFRP1 for binding miR‐542‐3p. SFRP1 could inactivate the Wnt/β‐catenin pathway. Animal experimentation substantiated the antitumor activity and mechanism of MTE in nude mice. Conclusions MTE suppresses glioma via the lncMEG3/miR‐542‐3p/SFRP1/Wnt/β‐catenin axis. These findings contribute to a theoretical basis for the use of MTE for glioma patients.

gliomas and effective medicines penetrating the blood-brain barrier. 4 Traditional Chinese medicine (TCM) has attracted great attention as an alternative therapeutic strategy against human cancers, 5,6 including glioblastoma (GBM). 7 Marsdenia tenacissima has been defined to be a promising antitumor TCM herb. 8 Marsdenia tenacissima extract (MTE) shows an in vitro strong antilymphoma potential, which is potentially related to its antiangiogenic potential in tumors. 9 Furthermore, a prior study has highlighted the proapoptotic effects of ethanolic MTE on hematologic neoplasm cells, contributing to its tumor-suppressive impact on hematological malignancy. 10 More recent studies have uncovered the antitumor effect of MTE on human ovarian cancer 11 and lung cancer. 12 However, its functional significance and mechanism of action in glioma are largely undetermined.
Long noncoding RNAs (lncRNAs) are noncoding transcripts with their length surpassing 200 nucleotides, which are involved in malignant transformation. 13 LncRNA MEG3 has shown low expression in glioma cells as well as antiproliferative and antimigratory impacts on glioma cells through modulation of a miRNA-mRNA interaction (miR-6088/SMARCB1). 14 Since our in silico analysis by starBase database showed a binding relationship between lncMEG3 and miR-542-3p, we speculated whether lncMEG3 could mediate miR-542-3p to affect the progression of glioma. A latest study has documented that miR-542-3p is abundantly expressed in high-grade gliomas relative to low-grade ones and that its high expression correlates with worse outcomes. 15 Recently, miR-542-3p has been suggested to target secreted frizzled-related protein 1 (SFRP1) via pairing to its 3′UTR. 16 SFRP1 belongs to the group of SFRPs and acts as a crucial antagonist of the canonical Wnt pathway. 17,18 Their interaction has been extensively illustrated to be mediated by miRNAs to exert regulatory roles in the progression of tumors. For instance, miR-27a-induced inhibition of SFRP1 can induce the Wnt/β-catenin pathway activation, resulting in enhanced proliferation and invasiveness of osteosarcoma cells. 19 Additionally, miR-1301-3p-dependent inhibition of SFRP1 activates the Wnt pathway to attenuate the prostate cancer stem cell expansion. 20 Furthermore, SFRP1 exerts an antitumor role in glioma through suppressing the Wnt/β-catenin pathway. 21 Following the aforementioned findings, we surmised that the lncMEG3/miR-542-3p/SFRP1/Wnt axis plays a modulatory effect of MTE in glioma. Here, both in vitro and in vivo models were utilized to testify this assumption to offer a basis for the in-depth mechanism underlying the TCM treatments of glioma.

| Treatment of MTE
The MTE was extracted referring to a previous study. 12 In short, the powder of the stem of M. tenacissima (1 kg) was immersed into water for three times (1.5, 1, and 0.8 h) for lixiviation. Next, the extracts were pooled, filtered, and concentrated, followed by precipitation with 8 times of 85% ethanol (v/w) at 4°C for 24 h. The ethanol in the extract was recovered and 85% ethanol was supplemented again for further precipitation. Subsequently, the ethanol was recovered thoroughly, and then the insoluble precipitate was removed by filtration.
Lastly, the extract was concentrated to 200 mL, diluted with water for injection, and supplemented with 0.3% polysorbate 80 (pH: 5.5-6.0). The MTE extract was identified to possess more than 95% of purity.

| Quantitative reverse transcriptionpolymerase chain reaction (qRT-PCR)
TRIzol reagent (Invitrogen) was implemented for the extraction of total RNA, and the reverse transcription was employed with a reverse transcription kit (TaKaRa). Expression levels of genes were detected using a fluorescence qPCR instrument (LightCycler 480, Roche). The reaction conditions of PCR were implemented with the fluorescence qPCR kit (SYBR Green Mix, Roche Diagnostics). The mRNA levels were normalized by glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and the miRNA level was normalized by U6.
Data were quantified by the 2 −ΔΔCt method. The primer sequences for target genes in this study are listed in Table 1.

| The cancer genome atlas (TCGA) analysis
Differentially expressed genes in GBM were analyzed by TCGA combined with Genotype-Tissue Expression (GTEx). The transcriptome data of GBM were collected from the TCGA database (https://xenab rowser.net/datap ages/), and the transcriptome data of normal tissues were collected from TCGA and GTEx. The limma package was used to compare and analyze the gene expression data of the GBM tissue samples and normal tissue samples obtained in the expression matrix. The screening threshold was set at |logFC| > 2 and adjusted p value <0.05. A volcano plot was created in the SangerBox website (http: https://sange rbox.com/Tool) to display the differentially expressed genes.

| Actinomycin D test
Cells were seeded in 6-well plates and cultured overnight with a low level of serum. Total RNA was extracted from the cells after they were treated with 2 g/L actinomycin D for 0, 12, and 24 h, followed by qRT-PCR.

| Colony formation assay
Cells were harvested, trypsinized, centrifuged for 5 min (25°C, 1500 rpm), and added with the complete medium for resuspension.
Each well was seeded with 500 cells in a 6-well plate containing complete culture medium (2 mL), and the cells were subsequently cultured for 2-3 weeks. The culture was terminated when the cell colonies in the 6-well plate were visible to the naked eye. Next, the culture medium was aspirated and the cells were subjected to two times of rinsing with phosphate-buffered saline (PBS). Then, each well was appended with methanol (1.5 mL) and fixed for 15 min.
After removing the methanol, the cells were slowly added with 1 mL of Giemsa staining solution along the well wall, and dyed in the dark for 20 min. Lastly, the Giemsa staining solution was washed away with running water, and the 6-well plate was dried by placing upside down on clean absorbent paper, followed by counting the number of colonies.

| Dual-luciferase reporter gene assay
The binding site of miR-542-3p to SFRP1, together with miR-542-3p to lncMEG3 was predicted by the online websites starBase (http://starb ase.sysu.edu.cn/) and miRDB (http://www.mirdb.org/). According to the predicted results, the wild type and mutant type sequences Firefly luciferase activity was standardized with the Renilla luciferase activity. The ratio of Firefly luciferase activity to Renilla luciferase activity was considered as the relative activity of luciferase.
Analysis was evaluated using a FISH kit. Nuclei were stained with 4′-6-diamidino-2-phenylindole for 20 min, and the images were finally analyzed by a fluorescence microscopy.

| RNA immunoprecipitation (RIP)
Sample preparation: cells were harvested, washed twice with precooled PBS, and centrifuged for 5 min, followed by adding an equal volume of RIP lysis buffer.
Magnetic bead preparation: the magnetic beads were resuspended, and 50 μL of the magnetic bead suspension was appended to each centrifuge tube, followed by adding with 500 μL RIP Wash RNA purification: Each sample was appended with 150 μL Proteinase K Buffer to suspend the magnetic bead-antibody complex, followed by incubation at 55°C for 30 min, which was then placed on a magnetic stand for the removal of the supernatant. After RNA extraction, lncMEG3, miR-542-3p, and SFRP1 expression was tested via the qRT-PCR.
All padding, drinking water, pellet feed, and other items were autoclaved. These animals were reared in a sterile laminar flow room of SPF grade under (22-26°C) temperature and (55 ± 5%) humidity.

| In vivo imaging system
The mice's late reaction was regularly observed. The mice were treated with luciferase substrate (15 mg/mL, 10 μL/g) via the intraperitoneal injected on the 0, 7th, 14th, 21st, and 28th day after U251-GFP-Luc cell transplantation. The luminescence signal changes were viewed by a small in vivo imager (IVIS Spectrum, Caliper, USA), and the tumor formation and growth state in the brain were assessed.

| Immunohistochemistry
The tumor tissues were fixed for 48 h with 4% paraformaldehyde, followed by the preposition of paraffin sections (4 μm). In brief, the paraffin sections were heated for 30 min, dewaxed with conventional xylene, and incubated with Ki-67 rabbit monoclonal antibody (ab16667, 1:200, Abcam) and the secondary antibody. Following washing three times in PBS, the sections were developed with diaminobenzidine, which was terminated 1-3 min later. Subsequently, the sections (nucleus) were stained with hematoxylin for 3 min, followed by dehydration, permeabilization, and blocking. Ki-67 showing yellow or brown granules in the nucleus was regarded to be positive.
The percentage of positive cells was counted in the selected 5 highpower fields.

| Statistical analysis
Data were processed with GraphPad prism8 software, and all data were reported as mean ± standard deviation. The normality of the data was analyzed using the Shapiro-Wilk test, and the homogeneity of variance was analyzed using the Bartlett's test. For data with normal distribution and homogeneity of variance, two-group comparisons were performed using the t-test, and multiple-group comparisons were performed using the one-way analysis of variance (ANOVA) and Tukey's post hoc test. p less than 0.05 indicates a significant difference.
Based on this, we could speculate that the MTE treatment suppressed the malignant behaviors of glioma cells through upregulating lncMEG3, reducing miR-542-3p expression, and finally upregulating SFRP1 expression. To confirm this conjecture, we first performed qRT-PCR for the detection of miR-542-3p and SFRP1 expression in NHAs and U87 and U251 glioma cells, and a high expression of miR-542-3p and a low expression of SFRP1 were observed in U87 and U251 cells in comparison to NHAs ( Figure 4D). Next, the expression levels of lncMEG3, miR-542-3p and SFRP1 in U87 and U251 cells with the addition of MTE were determined, and the results disclosed that the MTE treatment led to an increased expression of lncMEG3 and SFRP1 and a decreased expression of miR-542-3p ( Figure 4E).
According to the FISH results, we found that miR-542-3p colocalized with lncMEG3 and SFRP1 in the cytoplasm in U87, U251, and NHAs ( Figure 5A). RIP assay was then performed, which suggested that a large amount of lncMEG3, miR-542-3p and SFRP1 were witnessed in cells with the addition of Ago2 antibody, while with the addition of IgG, lncMEG3, miR-542-3p and SFRP1 were barely detected ( Figure 5B). Subsequently, a dual-luciferase reporter gene assay was  group, with no change in miR-542-3p expression ( Figure 5F). It validated that SFRP1 was a target gene of miR-542-3p and miR-542-3p negatively modulated SFRP1 expression. All these results imply that lncMEG3 and SFRP1 can competitively bind to miR-542-3p, and lncMEG3 upregulates SFRP1 expression by enhancing miR-542-3p expression.

| Overexpression of SFRP1 blocks the activation of the Wnt/β -catenin pathway, thereby suppressing the malignant phenotype of glioma cells
Emerging evidence has shown that SFRP1 can inhibit the Wnt/βcatenin pathway and further impact cell proliferation and other abilities. 26 in GSK-3β expression ( Figure 6A,B). In response to pcDNA3.1-SFRP1 treatment, the capabilities of proliferation ( Figure 6C,D), migration ( Figure 6E) and invasion ( Figure 6F) were revealed to be diminished.
To conclude, restoration of SFRP1 can restrict the malignant phenotype of glioma cells by restraining the activation of the Wnt/βcatenin pathway.

| MTE inhibits in-situ glioma formation in nude mice
At the animal level, we finally verified the effect of MTE on glioma in situ in nude mice. The nude mice were treated with lncMEG3 shRNA-or NC shRNA-transfected U251-GFP-Luc cells by transplantation, followed by the intraperitoneal injection of 200 μg/ F I G U R E 5 LncMEG3 and SFRP1 competitively bind to miR-542-3p. A. The localization of lncMEG3, miR-542-3p, and SFRP1 was detected by FISH assay. B. The binding of miR-542-3p to lncMEG3 and SFRP1 was examined by RIP assay. C-D. Dual-luciferase reporter gene assay was carried out to further verify the relationships among lncMEG3, miR-542-3p, and SFRP1. On the basis of MTE treatment, U87 cells were continued to be treated with lncMEG3 shRNA alone or cotransfected with miR-542-3p Antagomir: E. LncMEG3 and miR-542-3p expression was tested by qRT-PCR. On the basis of MTE treatment, U87 cells were further transfected with miR-542-3p Agomir alone or cotransfected with pcDNA3.1-SFRP1: F. miR-542-3p and SFRP1 expression was tested by qRT-PCR. n.s refers to no significance; *p < 0.05. N = 3.

| DISCUSS ION
As a TCM herb, Marsdenia tenacissima not only impairs the growth of lung cancer cells (A549 and LLC) and motivate their apoptosis but also shows an antitumor activity in vivo. 28 Its extract MTE also exerts a potent potential of inducing apoptosis through upregulating proapoptotic Bax, caspase-9, and caspase-3 and downregulating cyclin D1 and antiapoptotic Bcl-2 in hematologic neoplasm cells. 10 Also, MTE can attenuate ovarian cancer cell (SKOV3) growth and their resistance to apoptosis. 11 Given the antitumor effect of MTE F I G U R E 7 MTE suppresses the malignant behaviors of glioma cells by decreasing miR-542-3p, enhancing SFRP1 and blocking the Wnt/β-catenin pathway. U87 and U251 cells with MTE treatment were then transfected with miR-542-3p Agomir and pcDNA3.1-SFRP1: A. miR-542-3p and SFRP1 expression was tested by qRT-PCR. B. The Wnt/β-catenin pathway-associated factors (GSK-3β, p-GSK-3β and βcatenin) and EMT-related factors (E-cadherin, N-cadherin, Vimentin and Snail) were determined by western blot analysis. C. Cell viability was measured by CCK-8 assay. D. Cell proliferation ability was detected by colony formation assay. E. Cell migration ability was determined by wound healing assay. F. Cell invasion ability was tested by Transwell assay. n.s refers to no significance; *p < 0.05. N = 3.
in these cancers, we aimed at assessing its function in glioma. Our main findings evidenced that 100 mg/mL and 200 mg/mL of MTE appreciably impeded the in vitro growth, colony-forming, migratory, and invasive functions as well as in vivo oncogenicity of glioma cells (U87 and U251), accompanied with reversal of the EMT process.
Furthermore, we probed into the possible mechanism by which MTE exerted its antitumor impact on glioma cells.
First of all, our study unveiled that lncMEG3 could be elevated by MTE in the glioma cells. LncMEG3 is regarded as an underexpressed lncRNA in glioma tissues/cells while reexpression of lncMEG3 leads to impaired proliferation, migration, and invasiveness in glioma cells (GSC11 and D54). 29 A study has illustrated that deficiency of lnc-MEG3 shares a strong association with advanced WHO grade of glioma, tumor relapse, short overall survival, etc. 30 Reversely, enhancement of its expression considerably restrains the glioma cell proliferation while stimulating their apoptosis and autophagy. 30 Another study further demonstrates that lncMEG3 knockdown can augment glioma U118 cell growth through binding to miR-377. 31 In F I G U R E 8 MTE upregulates SFRP1 and blocks the Wnt/β-catenin pathway activation via the lncMEG3/miR-542-3p axis, thereby restricting glioma cell malignant behaviors. U87 and U251 cells in response to MTE treatment were then transfected with lncMEG3 shRNA and miR-542-3p Antagomir: A. LncMEG3, miR-542-3p and SFRP1 expression was determined by qRT-PCR. B. The Wnt/β-catenin pathwayassociated factors (GSK-3β, p-GSK-3β and β-catenin) and EMT-related factors (E-cadherin, N-cadherin, Vimentin and Snail) were tested by western blot analysis. C. Cell viability was analyzed by CCK-8 assay. D. Cell proliferation ability was determined by colony formation assay. E. Cell migration ability was detected by wound healing assay. F. Cell invasion ability was tested by Transwell assay. n.s refers to no significance; *p < 0.05. N = 3. All in all, this study unraveled that MTE treatment contributed to prevention against the growth and progression of glioma, highlighting its promise as a potential alternative therapy for glioma patients. Exploration on the molecular mechanisms suggests that activation of the lncMEG3/miR-542-3p/SFRP1 axis contributed F I G U R E 11 MTE inhibits in-situ glioma formation via lncMEG3/miR-542-3p/SFRP1. A. Ki-67 immunohistochemistry was implemented on the brain tumor tissues in the nude mice. B. LncMEG3, p53, miR-542-3p, and SFRP1 expression in the transplanted tumors was determined by qRT-PCR. C. The protein levels of p53 and SFRP1 were determined by western blot analysis. *p < 0.05. n = 6.

F I G U R E 9 MTE inhibits in
to the antitumor action of MTE, which gives insight into the crucial significance of lncRNA/miRNA/mRNA and pathways in the TCM treatment of patients with cancers. Neovascularization is a major factor in tumor progression and metastasis. More future in vivo studies are planned to investigate the MTE treatment of brain glioma in respect of tumor angiogenesis. Molecular magnetic resonance imaging 40 and quantitative ultramicroscopy 41 can be used to analyze various parameters of vessel structures (i.e., vessel length, radius, tortuosity, etc.) to characterize in-situ glioma formation and its response to MTE. Thus far, clinical delivery of the majority of antiglioma drugs is restricted at the blood-brain barrier by ABCB1 and ABCG2. 42 A previous study has shown that intraperitoneal administration of MTE can suppress the activity of ABCG2 in lung cancer xenografts. 43 Moreover, polyoxypregnane compounds isolated from MTE have been reported to overcome ABCB1-or ABCG2-mediated multidrug resistance in cancer cells. 44,45 Therefore, the efficiency of MTE brain delivery may not be restricted by ABCB1 and ABCG2, since itself can suppress the efflux activities of ABCB1 and ABCG2. The use of MTE alone or its use in combinational chemotherapy is worth more in vivo studies to improve the treatment of glioma.

AUTH O R CO NTR I B UTI O N S
CL and HMY conceived the ideas. CL and HMY designed the experiments. CL and GX performed the experiments. CL and GX analyzed the data. CL and GX provided critical materials. CL and GX wrote the manuscript. HMY supervised the study. All the authors have read and approved the final version for publication.

ACK N OWLED G M ENT
Thanks to all the contributors.

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors have nothing to be declared.

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets used or analyzed during the current study are available from the corresponding author on reasonable request.