Plasmid construction was performed as described in our previous studies (37–39). pLenti-CMV-MDM2-green fluorescent protein (GFP)-puro was constructed based on a plenti-CMV-puro backbone. The present study first PCR-amplified the MDM2 sequence lacking a termination codon using the plasmid pcDNA3.1-MDM2 as a template using the following primer pair and the Prime Star DNA Polymerase kit (Takara Bio, Inc.): Forward, 5′-CGCGCCACCATGGTGAGGAGCAGGCAAAT-3′ and reverse, 5′-ACGCGGGGAAATAAG-3′. The PCR conditions used were as follows: 94°C for 5 min; 94°C for 30 sec; 56°C for 60 sec; 72°C for 1 min; 72°C for 10 min; 4°C for 10 min; cycle number, 35. PCR fragments were purified on a 1.5% agarose gel with ethidium bromide using the E.Z.N.A.^® DNA Kit (Omega Bio-Tek, Inc.). A purified GFP cDNA fragment that lacked an initiation codon was PCR-amplified using plenti-CMV-GFP as a template by the following primer pair and the Prime Star DNA Polymerase kit (Takara Bio, Inc.): Forward, 5′-CGTCAGATCCGCTAGCGCTACCGGTCG-3′ and reverse, 5′-GCTTACTTGTACAGCTCGTCCATGCCG-3′. The PCR conditions used were as follows: 94°C for 5 min; 94°C for 30 sec; 55°C for 30 sec; 72°C for 30 sec; 72°C for 10 min; 4°C for 10 min; cycle number, 32. PCR fragments were purified on a 1.5% agarose gel with ethidium bromide using the E.Z.N.A.^® DNA Kit (Omega Bio-Tek, Inc.). Overlap PCR was used to connect the purified MDM2 and GFP fragments using the following primer pair and the Prime Star DNA Polymerase kit (Takara Bio, Inc.): Forward, 5′-CGCGCCACCATGGTGAGGAGCAGGCAAAT-3′ and reverse, 5′-GCTTACTTGTACAGCTCGTCCATGCCG-3′. The PCR conditions used were as follows: 94°C for 5 min; 94°C for 30 sec; 56°C for 60 sec; 72°C for 1.5 min; 72°C for 10 min; 4°C for 10 min; cycle number, 35. PCR fragments were purified on a 1.5% agarose gel with ethidium bromide using the E.Z.N.A.^® DNA Kit (Omega Bio-Tek, Inc.). Subsequently, Kozak sequences were added to the purified MDM2-GFP fragment and restriction enzyme sites for BamHI and SalI were also inserted. The recombinant fragment and pLenti-CMV-puro were digested with BamHI and SalI (Takara Bio, Inc.), respectively, and then T4 ligase (Takara Bio, Inc.), was used to form the recombinant plasmid pLenti-CMV-MDM2-GFP-puro plasmid (Fig. S1A and B). In the present study, fragments and plasmids were digested with Takara QuickCut enzymes (Takara Bio, Inc.) and fragments and vectors were ligated with T4 ligase (Takara Bio, Inc.). In the present study, the pLenti-CMV-MDM2-GFP-puro plasmid sequencing process was conducted at Sangon Biotech Co., Ltd. Stbl3 competent cells were used for plasmid transformation, and plasmid was extracted from Stbl3 using a TIANGEN Kit. The pLenti-CMV-MDM2-GFP-puro plasmid sequencing result and original sequence were aligned using APE software (version 2.0; https://jorgensen.biology.utah.edu/wayned/ape/; Fig. S1B). Flag-14-3-3 and pcDNA3.1-p53 were purchased from Addgene, Inc.

U251MG, A549, MCF-7 and 293T cells (The Cell Bank of Type Culture Collection of The Chinese Academy of Sciences) were used in the present study. All cell lines were cultured at 37°C with 5% CO₂. Ham's F-12K (Kaighn's) Medium (Thermo Fisher Scientific, Inc.) was used to culture A549 cells. U251MG, 293T and MCF-7 cells were cultured in DMEM with high glucose (HyClone; Cytiva). All media were supplemented with 10% FBS (Shanghai ExCell Biology, Inc.) and 100 µg/ml penicillin and streptomycin to make complete medium. Media were changed every other day. The inhibitors and activators used in the present study included an ERK inhibitor (FR180204; 10 µM, 24 h; cat. no. S7524; Selleck Chemicals), a 90 KDa ribosomal protein S6 kinase (RSK) inhibitor (SL0101; 50 µM, 24 h; cat. no. 77307-50-7; Sigma-Aldrich; Merck KGaA), vemurafenib (2 µM, 24 h; cat. no. A3004; APeXBIO Technology LLC), a MEK inhibitor U0126 (2 µM, 24 h; cat. no. 9903; Cell Signaling Technology, Inc.), TGF-β (10 ng/ml, 48 h; cat. no. 8915; Cell Signaling Technology, Inc.) and 20 mM Citrate pH 3.0 (Sterile) (cat. no. 9871; Cell Signaling Technology, Inc.). To examine TGF-β inducing EMT, U251 cells were incubated with 10 ng/ml TGF-β, and the control group was incubated with an equal volume of 20 mM citrate solution buffer, which was the buffer solution for TGF-β.

pcDNA6B, pcDNA6B-Flag-14-3-3, pcDNA3.1 and pcDNA3.1-p53 were purchased from Addgene, Inc. Before transfection, MDM2-GFP stable cells were digested by tryptase that contained EDTA when the cells were 80% confluent. Then, 2×10⁵ cells/well were seeded into six-well plates (Suzhou Beaver Biomedical Engineering Co., Ltd.). pcDNA6B-Flag-14-3-3 or pcDNA3.1-p53 plasmid (2 µg) were transfected using PloyFect Effectene Transfection Reagent (Qiagen, Inc.) according to the manufacturer's protocols when cells were 60–80% confluent. Related empty vector (2 µg) was also transfected into MDM2-GFP stable cells using PloyFect as a control group. Cells were cultured at 37°C with 5% CO₂ for 8 h, and then the media were replaced with fresh media. After transfection, cells were cultured at 37°C with 5% CO₂ for 36 h. When the transfection was finished, cells were harvested in 1.5-ml microcentrifuge tubes for determination of gene overexpression by quantitative PCR (qPCR). Samples successfully transfected with plasmids were further tested by western blotting.

According to the manufacturer's protocols of the PloyFect Effectene Transfection Reagent (Qiagen, Inc.) The second generation lentiviral system plasmid plenti-CMV-MDM2-GFP-puro, pCMV–VSV-G (Addgene, Inc.) and pCMV-dR8.2 dvpr (Addgene, Inc.) (at a ratio of 4:1:3; 2 µg total plasmids) were co-transfected into 293T packaging cells. Additionally, plenti-CMV-puro empty vector (Addgene, Inc.), pCMV–VSV-G and pCMV-dR8.2 dvpr (at a ratio of 4:1:3; 2 µg total plasmids) were co-transfected into 293T packaging cells as a control group for control stable cell lines. All cells were cultured at 37°C with 5% CO₂ for 8 h, and the media were replaced by fresh media. After 24 and 36 h, media were collected in centrifuge tubes. Subsequently, media were concentrated by ultracentrifugation at 16,100 × g for 30 min at 4°C and filtered with a 4-µm filter. Viral titers were measured, and then recombinant lentiviral particles (10⁷ TU/ml) were added to U251, A549 and MCF7 cells. The multiplicity of infection for U251, A549 and MCF7 cells was 3, 40 and 30, respectively. Equal volumes of complete media that contained 8 µg/ml Polybrene (Sigma-Aldrich; Merck KGaA) were then added. The media were replaced with fresh media containing 2 µg/ml Puromycin (Puro; Sigma-Aldrich; Merck KGaA) for selection after 24 h for stable cell line enrichment. The duration of the selection process was 7 days. Subsequently, complete media with 1 µg/ml Puro for maintenance were used to culture this purified stable cell line. Then, 2×10⁵ cells per well were seeded into six-well plates using 0.25% Trypsin-EDTA solution (Sigma-Aldrich; Merck KGaA) for 5 min at 37°C, and then the stable cells were collected for qPCR and western blotting.

Cells were harvested in 1.5-ml microcentrifuge tubes and washed with PBS. Subsequently, Nonidet P (NP)-40 lysis buffer (2% NP-40, 80 mM NaCl, 100 mM Tris-HCl and 0.1% SDS) was added to the cells, and PMSF (1 mM; Thermo Fisher Scientific, Inc.) was supplemented. Cell lysate was incubated on ice for 5 min and then the lysate was centrifuged at 16,100 × g for 10 min at 4°C. The sample protein concentration was determined using a Bio-Rad Protein DC Assay kit (Bio-Rad Laboratories, Inc.). Protein samples (50 µg per lane) from the control group and treatment group were subjected to SDS-PAGE (15%). Next, proteins were transferred from gels to a 0.45-µm PVDF membrane (EMD Millipore) at 200 mA for 70 min. The membranes were blocked with 5% BSA (Sigma-Aldrich; Merck KGaA) in PBS with 0.02% Tween 20 for 1 h at room temperature. Membranes were incubated with the primary antibodies (diluted to recommended concentration with 5% BSA) overnight at 4°C. The antibodies used were as follows: Anti-GAPDH (rabbit; dilution, 1:1,000 for western blotting; cat no. 10494-1-AP; ProteinTech Group, Inc.), anti-GFP (mouse; dilution, 1:1,000 for western blotting; cat no. 66002-1-Ig; ProteinTech Group, Inc.), anti-p53 (mouse; dilution, 1:1,000 for western blotting; cat no. TA808657; OriGene Technologies, Inc.), phosphorylated (p-)B-Raf (dilution, 1:1,000 for western blotting; cat. no. 2696; Cell Signaling Technology, Inc.), B-Raf (dilution, 1:1,000 for western blotting; cat. no. 14814; Cell Signaling Technology, Inc.), p-MEK (dilution, 1:1,000 for western blotting; cat. no. 3958; Cell Signaling Technology, Inc.), MEK (dilution, 1:1,000 for western blotting; cat. no. 4694; Cell Signaling Technology, Inc.), p-ERK (T202/Y204) (dilution, 1:1,000 for western blotting; cat. no. 4370; Cell Signaling Technology, Inc.), ERK (dilution, 1:1,000 for western blotting; cat. no. 4695; Cell Signaling Technology, Inc.), p-ribosomal protein S6 kinase (p70S6K) (T389) (dilution, 1:1,000 for western blotting; cat. no. 9206S; Cell Signaling Technology, Inc.), p70S6K (dilution, 1:1,000 for western blotting; cat. no. 9202S; Cell Signaling Technology, Inc.), p-eukaryotic translation initiation factor 4E binding protein 1 (4EBP1) (T37/46) (dilution, 1:1,000 for western blotting; cat. no. 9451T; Cell Signaling Technology, Inc.), 4EBP1(dilution, 1:1,000 for western blotting; cat. no. 9644S; Cell Signaling Technology, Inc.), 14-3-3 (pan) (dilution, 1:1,000 for western blotting; cat. no. 8312; Cell Signaling Technology, Inc.), E-CAD (dilution, 1:1,000 for western blotting; cat. no. 20874-1-AP; ProteinTech Group, Inc.), N-cadherin (N-CAD; dilution, 1:1,000 for western blotting; cat. no. GTX-127345; GeneTex, Inc.), vimentin (VIM; dilution, 1:1,000 for western blotting; cat. no. 5741T; Cell Signaling Technology, Inc.), zinc finger E-box binding homeobox 1 (ZEB1; dilution, 1:1,000 for western blotting; cat. no. 3396T; Cell Signaling Technology, Inc.), Snail (dilution, 1:1,000 for western blotting; cat. no. 3879T; Cell Signaling Technology, Inc.), Snail family transcriptional repressor 2 (Slug; dilution, 1:1,000 for western blotting; cat. no. 9585T; Cell Signaling Technology, Inc.), GFP (dilution, 1:1,000 for western blotting; cat. no. GTX-113617; GeneTex, Inc.) and Flag (dilution, 1:1,000 for western blotting; cat. no. F3165; Sigma-Aldrich; Merck KGaA). GAPDH served as a control. Then, appropriate anti-rabbit and anti-mouse secondary antibodies (cat nos. SA00001-1 and SA00001-2; ProteinTech Group, Inc.; dilution, 1:3,000 with 5% BSA) were added to the membranes for 1 h at room temperature. Subsequently, membranes were exposed to ECL solution (cat no. WBKLS0500; EMD Millipore) using an Amersham Imager 600 machine (GE Healthcare). For protein expression intensity measurement and quantification, densitometry of the bands was performed using ImageJ version 1.49 software (National Institutes of Health).

A RNeasy Mini kit (cat. no. 74104; QIAGEN Inc.) was used to extract total RNA of U251, MCF7 and A549 cells. The RNA products isolated using this kit were loaded on a 1% agarose gel and separated by electrophoresis. Subsequently, bands for 28S RNA, 18S RNA and 5S RNA were inspected to ensure that RNA was extracted correctly. Next, the RNA quality and concentration were assayed using a Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific, Inc.). A total of 1 µg RNA was used for cDNA synthesis using the iScript™ cDNA Synthesis Kit (cat no. 170-8890; Bio-Rad Laboratories, Inc.). The reverse transcription (RT) conditions were as follows: 25°C for 5 min, 46°C for 20 min, 95°C for 60 sec, hold at 4°C. Primers used for qPCR are listed in Table I. PerfeCTa^® SYBR^® Green SuperMix (cat no. 95054-100; Quantabio) was used for qPCR and the conditions were as follows: 95°C for 3 min, 95°C for 15 sec, 55°C for 45 sec, 72°C for 30 sec, total of 35 cycles. A CFX96 Touch Real-Time PCR Detection System (Bio-Rad Laboratories, Inc.) was used to obtain and analyze the data. Actin was used to normalize gene expression levels. For quantification, the 2^−ΔΔCq method was used for this experiment (40).

Tissue sections were fixed in 10% neutral buffered formalin at room temperature for 48 h and then embedded in paraffin. The thickness of tissue sections was 4 µm. For immunohistochemical staining, slides were placed in a 65°C incubator for 30 min. Tissue slides were deparaffinized in xylene for 5 min, then in fresh xylene for another 5 min, 100% ethanol for 5 min, 95% ethanol for 5 min and 75% ethanol for 5 min, and then washed in water for 5 min. Endogenous peroxidase was blocked with 3% hydrogen peroxide. Antigen retrieval was achieved using a hot water bath (100°C) and 10 mM citric sodium buffer (pH 6.0) for 15 min. Sections were then blocked with 5% BSA for 1 h at room temperature and incubated overnight at 4°C with the indicated primary antibodies in 5% BSA: E-CAD (dilution, 1:100; cat. no. 20874-1-AP; ProteinTech Group, Inc.) and N-CAD (dilution, 1:50; cat. no. GTX-127345; GeneTex, Inc.). Antibody binding was detected using the EnVision™ Dual Link System-HRP DAB kit (undiluted; cat. no. K4010; Dako; Agilent Technologies, Inc.). Anti-Rabbit HRP labeled polymer (cat. no. K4010; Dako; Agilent Technologies, Inc.) was used to cover tissues, followed by incubation for 30 min at room temperature. Subsequently, 20 µl DAB (~1 drop) in 1 ml DAB substrate buffer was applied to each slide and incubated for 5 min. Sections were then counterstained with hematoxylin for 5 min at room temperature, followed by washing of the slides at room temperature. After staining, the slides were washed with running tap water for 5 min. Subsequently, tissue slides were placed in 75% ethanol for 5 min, 95% ethanol for 5 min, 100% ethanol for 5 min, xylene for 5 min and then in fresh xylene for another 5 min. For negative controls, primary antibodies were excluded. The mitotic index was quantified by viewing and capturing images of 10 random high-power fields for each tissue section on a Keyence All-In-One Fluorescence Microscope (Keyence Corporation), using a 40× or 20× objective. For evaluation and quantification of immunohistochemical data, 10 fields within the tumor area under high power magnification (40×) were randomly selected for evaluation using ImageJ version 1.49 software (National Institutes of Health). The investigators performed blind counting for all quantifications.

All clinical patient glioma samples were obtained from The Affiliated Hospital of Southwest Medical University (Luzhou, China). Recruitment was between March 2017 and May 2018. And the recruitment patient mean age is 33 years old. And all patient ages are among 18–59 years old. The recruitment patient proportion is 50 percent male and 50 percent female. The inclusion criteria were: i) Signature of informed consent voluntarily; ii) pathological diagnosis of glioma; iii) aged 18–59 years old, male and female; iv) normal blood-routine test; and v) normal liver function. The exclusion criteria were: i) Patients unwilling to participate in the experiment or had poor compliance; ii) other clinical trials being conducted; iii) previous history of brain tumor and radiotherapy conducted; iv) multiple lesions with extensive radiation volume; and v) with history of radiotherapy for brain tumors.

All samples were obtained from resection surgery and handled following clinical guidelines and in compliance with ethical standards (approval no. Y2019045; Southwest Medical University Ethics Committee, Luzhou, China). An informed consent document was signed by the patients to agree to the use of their samples in scientific research. Clinical samples were graded according to tumor cell density, mitotic index, degree of necrosis from the WHO Classification of Tumors (3rd Edition) (41). Samples were frozen at −196°C in liquid nitrogen for qPCR, western blotting and paraffin sectioning for subsequent histological examination.

After treatment (24 h exposure), cells were trypsinized, counted, and 800 cells per well were re-plated in six-well plates for colony formation assays. After incubation for 10–14 days, colonies were fixed with methanol:acetic acid (3:1; pre-cooled at 4°C) at 4°C for 10 min, and stained with 1% crystal violet at room temperature for 15 min, and then counted. A light microscope (Nikon Corporation) was used to observe and count colonies, and GraphPad Prism 6.0 software (GraphPad Software, Inc.) was used for quantification and analysis. Plating efficiency (PE) was determined, and the surviving fraction was calculated based on the number of colonies that arose after treatment and is expressed in terms of PE. Each experiment was repeated three times.

RNA was prepared using the Qiagen RNeasy Kit (cat. no. 74104; Qiagen, Inc.) according to the manufacturer's protocols. Strand-specific libraries were generated using the KAPA Stranded mRNA-Seq Kit with KAPA mRNA Capture Beads (cat no. KK8421; Kapa Biosystems; Roche Diagnostics). Sequencing libraries were quantified using the Qubit 2.0 Fluorometer (Invitrogen; Thermo Fisher Scientific, Inc.) and validated using High-Performance Capillary Electrophoresis and the Agilent Tapestation 4200 (Agilent Technologies, Inc.). RNA-seq was performed by Genewiz, Inc. The sequencing libraries were multiplexed and clustered on one lane of a Flow Cell. After clustering, the Flow Cell was loaded on the Illumina HiSeq instrument (Illumina, Inc.) according to the manufacturer's protocol. The samples were sequenced using a 2×150 Paired End configuration. Image analysis and base-calling were conducted by the HiSeq Control Software (version 3.3.20; Illumina, Inc.). Raw sequence data (.bcl files) generated from Illumina HiSeq were converted into Fastq files and de-multiplexed using Illumina's bcl2fastq software (version 2.17; Illumina, Inc.). One mismatch was allowed for index sequence identification. The quality of Fastq files was checked with FastQC (version 0.11.8; Baraham Bioinformatics group; http://www.bioinformatics.babraham.ac.uk/). To be referred to as ‘differentially expressed’, a gene had to have a false discovery rate adjusted P<0.05 in Differential Expression Analysis of RNA-seq Data (DEseq) as well as in null model of hypothesis. DesEQ2 performs a likelihood ratio test that compares how well a gene count data fits a full model (with independent variable time) compared with a reduced model without those variables. The null model of hypothesis takes the average expression of groups into consideration. The gene list was further ranked using fold change criteria.

GSEA was used to study the enrichment of genes in different pathways. Non-parametric GSEA was performed using GSEA 3.0 (Broad Institute, Inc.; http://www.gsea-msigdb.org/gsea/index.jsp). The gene sets that significantly out-performed random-class permutations were considered significant.

Cells were digested and then seeded into 24-well culture plates at a density of 1×10⁴ cells/well. Subsequently, cells were treated with Silibinin at the indicated concentration for 24 and 48 h. MTT reagent (5 mg/ml; HyClone; Cytiva) was then added to each well, and the cells were incubated at 37°C for 4 h. After incubation, DMSO was used to replace the MTT reagent, and cells were further incubated at room temperature for 30 min. The absorbance of formazan at 570 nm was detected using a spectrophotometer (Bio-Rad Laboratories, Inc.).

U251 and A549 cells were first seeded into a 12-well cell culture plate (2×10⁵ cells/well). The monolayer was scratched using a 200-µl pipet tip when the cell confluence reached 75–80%. Subsequently, cells were cultured in an incomplete medium that contained 0.5% serum. Images were captured at 0, 24 and 48 h after wounding using a light microscope (Olympus Corporation). ImageJ (version 1.49; National Institutes of Health) was used to analyze these images, and the cell migration ratio was calculated as follows: N=[1-(D_n/D₀)] ×100%, where N represents the ratio of cell migration, D_n represents the closure distance at the sampling time and D₀ represents the original distance.

NOD/SCID mice (stock no. 001303) were purchased from Chengdu Dossy Experimental Animals Co., Ltd. All animal experiments were performed following the guidelines that were reviewed and approved by Southwest Medical University Ethics Committee (Luzhou, China) before conduction (approval no. 201903-34). For subcutaneous xenografts, U251 cells and stable MDM2 and 14-3-3-expressing cells (5×10⁶) in PBS were injected subcutaneously into the lower flank of 6-week-old female NOD/SCID mice (weight, ~20 g; 6 mice per group; total of 24 mice). All mice were housed in a SPF mice room under standard conditions (temperature, 20-26°C; humidity, 40–70%; ammonia concentration below 14 mg/m³; light/dark cycle was 12/12 h; free access to food and water). Animal health and behavior were monitored daily. Mice were monitored for tumor development by measurements of tumor weight, tumor length and width. Tumor volume was calculated according to the following formula: (Length × width²)/2. Related mice were treated daily with 100 mg/kg (body weight) FR180204 via intraperitoneal injection using a 27-gauge needle. The duration of the experiment was 28 days. After a 4-week period (post-inoculation), once tumors grew to a palpable size (~500 mm³), the mice were sacrificed (no mice died before the endpoint), and tumors were dissected (no other tumors were observed). To minimize the struggle of the animals and reduce the pain of the animal, the cervical dislocation method was performed after the mice were anesthetized. When mice were sacrificed, mice were placed in the induction chamber with 2% isoflurane (induction dose) in oxygen, and then cervical dislocation was used for euthanasia immediately. Half of each tumor was frozen at −196°C in liquid nitrogen, and half was fixed in 4% paraformaldehyde for subsequent histological examinations.

During the 4-week observation period, if there was evidence of weight loss >20% from the baseline, partial rigor of the limbs, inability to regain normal posture if placed on back, labored breathing, inability to eat or drink, severe dyspnea or any other serious illness, the mice would have been euthanized. Consequently, tumor growth of 1.0 cm³ will be a humane endpoint to euthanize mice. Mice would be euthanized if tumor growth plateaus for more than five measurements in which case controls would be concurrently euthanized with experimental populations. Animals were checked daily for body condition scoring and signs of discomfort. Those animals in pain or extreme discomfort would have been euthanized. The final decision to perform euthanasia was at the discretion of the clinical veterinarian.

Images were acquired using Olympus Stream software (version 2.4; Olympus Corporation) under a fluorescence microscope (Olympus Corporation), and ImageJ (version 1.49; National Institutes of Health) was used to analyze the images. Band densitometry and data quantification was also conducted using ImageJ. All images in the present study were grouped using Adobe Illustrator software (version 24.0; Adobe Systems, Inc.).

Data were normalized using control markers. Data from each group were obtained from at least three replicates. Mean values and SD are presented. In this study, statistical analysis was performed using GraphPad Prism 6.0 software (GraphPad Software, Inc.). Statistical significance was analyzed using an unpaired Student's t-test or one-way ANOVA with Tukey's post hoc test to correct for multiple comparisons. P<0.05 was considered to indicate a statistically significant difference, P<0.01 was considered to indicate a statistically very significant difference and P<0.001 was considered to indicate a statistically extremely significant difference.

According to previous studies, MDM2 is an important factor related to cell invasion, metastasis and EMT (7,8). Therefore, the present study aimed to determine the MDM2 levels in cells undergoing EMT. TGF-β can induce EMT (6). U251 cells were incubated with 10 ng/ml TGF-β or an equal volume of 20 mM Citrate solution buffer which was the buffer solution for TGF-β. Subsequently, cells were collected and total RNA was extracted. As shown in Fig. 1A, compared with the control group, full length MDM2 was upregulated ~2.5 times after incubation based on semi-quantitative analysis of band intensity from PCR results. Similarly, MDM2 gene expression in U251 cells treated in the same manner was increased ~3 times after incubation according to qPCR (Fig. 1B). Subsequently, the protein expression levels of MDM2, p53, E-CAD, N-CAD and VIM in TGF-β-treated U251 cells were analyzed by western blotting (Fig. 1C). As shown in the quantitative analyses, U251 cells treated with TGF-β exhibited decreased E-CAD protein expression and increased N-CAD and VIM protein expression, which suggested that EMT had been induced. In the same cells, RT-qPCR indicated that the MDM2 expression was significantly increased after incubation with TGF-β peptide (Fig. 1D). Furthermore, since MDM2 is an upstream gene of p53, the present study also determined p53 protein expression (2–4). Consistent with other publications, it was observed that the accumulation of MDM2 was associated with the degradation of p53 (Fig. 1C). To summarize, MDM2 expression was upregulated in EMT induced by TGF-β.

The aforementioned results demonstrated that MDM2 expression was upregulated by TGF-β-induced EMT. Furthermore, a previous study has reported that MDM2 can promote EMT in breast cancer cells (42). Therefore, clinical glioma samples were collected from patients with different pathological gradings. These tumor samples were then processed for immunohistochemistry. According to the immunohistochemistry results and positive cell quantification data, it was identified that the higher the glioma grading, the higher the protein expression levels of MDM2 (Fig. 2A). Additionally, the present study used qPCR and western blotting to analyze MDM2 gene and protein expression, respectively, in tumor samples. Similar to the immunohistochemistry findings, it was observed that the expression levels of MDM2 were positively associated with pathological grade. MDM2 expression in G3 glioma samples was the highest, and it was the lowest in G1 glioma samples (Fig. 2B-D). It is well known that higher tumor pathological grades are strongly associated with higher malignancy and lower tumor differentiation, which results in higher metastatic potential (6,7). EMT is an important mechanism enabling tumor metastasis (6–8). The present data suggested that MDM2 may serve a key role in inducing EMT in glioma. Therefore, the present study aimed to determine whether MDM2 overexpression could induce EMT in U251 cells. pLenti-CMV-MDM2-GFP-puro (Fig. S1A) was a flexible and convenient tool for studying MDM2 expression by tracking green fluorescence by microscopy. Recombinant lentiviral plasmid (pLenti-CMV-MDM2-GFP-puro) was co-transfected into 293T packaging cells to generate MDM2-GFP lentivirus. In the control group, empty vector pLenti-CMV-Puro was used. After 36 h, green fluorescence was observed (Fig. S1C), and the virus was then collected and concentrated by ultracentrifugation. Viral titers were measured, and then the virus was added to U251 cells. After purification, green fluorescence was observed in stable puro positive U251 cells (Fig. S1D). Stable control U251 cells and stable U251 cells overexpressing MDM2-GFP were collected, and total RNA was extracted. The gene expression levels of MDM2, p53, E-CAD, N-CAD and VIM were analyzed (Fig. 2E). Subsequently, band intensity was quantified by SqRT-PCR (Fig. 2E). As shown by the quantification data, MDM2 overexpression significantly decreased the gene expression levels of E-CAD and significantly increased the gene expression levels of N-CAD and VIM at the same time. This suggested that EMT had been successfully induced compared with the control group (Fig. 2E and F). qPCR was also used to verify these in stable control U251 cells and stable U251 cells overexpressing MDM2-GFP (Fig. 2F). p53 gene expression was also decreased significantly compared with the control group due to the degradation of p53 by MDM2 overexpression (Fig. 2E and F). MDM2, p53, E-CAD, N-CAD and VIM protein expression in stable control U251 cells and stable U251 cells overexpressing MDM2-GFP was then analyzed by western blotting. Consistent with the gene expression data, MDM2 overexpression significantly decreased the protein expression levels of E-CAD and p53, and markedly increased the protein expression levels of N-CAD and VIM (Fig. 2G). MDM2 overexpression and control U251 cell lines were then used for RNA-Seq. GSEA revealed that the EMT pathway gene set was closely associated with the expression levels of MDM2 compared with the control group (Fig. 2H). Overall, MDM2 overexpression could induce EMT in U251 cells.

The aforementioned results demonstrated that overexpression of MDM2 could promote EMT in the U251 glioma cell line. Subsequently, the present study examined whether MDM2 overexpression induced EMT in other cancer cell lines. Lentivirus from 293T cells was added to the A549 lung cancer cell line to yield control A549 cells and A549 cells stably overexpressing MDM2-GFP. Green fluorescence was observed in stable A549 cells after purification (Fig. S1E), and the gene and protein expression levels of MDM2-GFP were assessed by qPCR (Fig. S2A) and western blotting (Fig. S2B). Subsequently, the gene expression levels of E-CAD, N-CAD and VIM were determined by SqRT-PCR and qPCR in these two stable A549 cell lines (Fig. 3A and B). As shown in the quantification analysis, MDM2 overexpression significantly decreased the gene expression levels of E-CAD and significantly increased the gene expression levels of N-CAD and VIM. Furthermore, the migration ability of these two stable cell lines was determined. Scratch wound assays were used to evaluate their migration abilities and the percentage closure of the scratch wound was quantified at 0, 24 and 48 h post injury in these two cell lines (Fig. 3C). As shown by the images of migration and quantification data, the percentage closure in the MDM2 overexpression group was higher than that in the control group at 24 and 48 h post injury (Fig. 3C). These data demonstrated that MDM2 overexpression increased the cell migration abilities of A549 cells. Additionally, E-CAD, N-CAD and VIM protein expression was analyzed by western blotting (Fig. 3D and E). Consistently, MDM2 overexpression significantly decreased the protein expression levels of E-CAD and significantly increased the protein expression levels of N-CAD and VIM (Fig. 3D and E). Therefore, the present data demonstrated that MDM2 overexpression induced EMT in the A549 lung cancer cell line. Furthermore, this was also tested in the MCF7 breast cancer cell line. Green fluorescence was observed in stable MCF7 cells after purification (Fig. S1F), and gene and protein expression levels of MDM2-GFP were assessed by qPCR (Fig. S2C) and western blotting (Fig. S2D). According to SqRT-PCR, qPCR and western blotting data, MDM2 overexpression also induced EMT in MCF7 cells (Fig. 3F-H). Therefore, overexpression of MDM2 promoted EMT in a number of cancer cell lines.

Previously, it was demonstrated that MDM2 promoted EMT in glioma cells (U251), lung cancer cells (A549) and breast cancer cells (MCF7). However, the mechanism by which MDM2 induced EMT was not clear. Therefore, the present study aimed to determine the mechanism driving this EMT. It was previously identified that the protein expression levels of E-CAD, N-CAD and VIM were altered, and the gene levels were also altered. This suggested that the transcription factors that control EMT may have been upregulated when MDM2 was overexpressed. Snail, Slug and ZEB1 are the most important transcriptional factors of EMT (Fig. 4A) (6,43,44). Therefore, Snail, Slug and ZEB1 protein expression was analyzed by western blotting in stable MDM2-overexpressing U251 cells and control U251 cells. As shown in the semi-quantification data, MDM2 overexpression significantly upregulated the protein expression levels of Snail, Slug and ZEB1 (Fig. 4A). Since these transcription factors were activated, the present study aimed to determine the upstream signaling pathway driving their upregulation. Therefore, GSEA was used to analyze these two stable cell lines, and revealed that the gene set for the B-Raf signaling pathway was closely associated with MDM2 expression compared with the control group (Fig. 4B). Additionally, GSEA was used to analyze whether the mTOR signaling pathway was associated with the expression levels of MDM2. However, according to the present data, no significant association was observed between the mTOR signaling pathway and MDM2 overexpression in U251 cells (data not shown). Subsequently, western blotting was used to analyze the levels of p-B-Raf, p-MEK and p-ERK to verify whether the B-Raf signaling pathway was activated by MDM2 overexpression. Consistent with the present GSEA data, western blotting demonstrated that the B-Raf signaling pathway was activated (Fig. 4C). Furthermore, the levels of p-P70S6K and p-4EBP1, downstream target genes of mTOR, were determined by western blotting (45–47). No significant changes in the levels of p-P70S6K and p-4EBP1 were detected, indicating that this process was independent of the mTOR signaling pathway (Fig. 4C). E-CAD levels were also analyzed by western blotting in stable control U251 cells and stable U251 cells overexpressing MDM2-GFP following treatment with DMSO, vemurafenib (B-Raf inhibitor) (48,49), U0126 (MEK inhibitor) (50,51) and FR180204 (ERK inhibitor) (52,53). As shown in Fig. 4, vemurafenib, U0126 and FR180204 could prevent the decrease in the protein expression of E-CAD (Fig. 4D and F). However, RSK inhibitor SL0101 did not affect the protein expression of E-CAD (Fig. 4D and F), which suggested that inhibition of B-Raf signaling could prevent MDM2 from inducing EMT. In other words, MDM2 promoted EMT by activating the B-Raf signaling pathway. Subsequently, the protein expression levels of Snail, Slug and ZEB1 were analyzed by western blotting in MDM2-overexpressing U251 cells which were treated with FR180204 and in control U251 cells (U251 cell line with stable overexpression of plenti-CMV-puro empty vector). No significant changes in the expression levels of Snail, Slug or ZEB1 were detected between the two groups by western blotting, which suggested that ERK inhibition could inhibit the increase of these EMT transcription factors (Fig. 4E). These data verified that MDM2 induced EMT by upregulating EMT transcription factors via the B-Raf signaling pathway. Additionally, the present data demonstrated that B-Raf inhibitors, MEK inhibitors and ERK inhibitors could all rescue the protein expression levels of E-CAD. RSK is a downstream protein kinase of ERK (54,55). To determine whether components downstream of ERK also participated in this process, western blotting was used to analyze E-CAD protein expression in control U251 cells and MDM2-overexpressing U251 cells treated with SL0101 (RSK inhibitor). As shown in the western blot images and semi-quantification data, only the RSK inhibitor had no effect on the protein expression levels of E-CAD (Fig. 4F), while all other drugs did have an effect. These data demonstrated that MDM2 induced EMT by upregulating EMT transcription factors via activation of ERK rather than through the ERK signaling pathway. Additionally, this mechanism was assessed in other cell lines, and E-CAD levels were analyzed by western blotting in control A549 cells and MDM2-overexpressing A549 cells treated with DMSO, vemurafenib, U0126 or FR180204. Similar to the results in U251 cells, B-Raf, MEK and ERK inhibitors could all rescue a decrease in E-CAD protein expression (Fig. 4G). Furthermore, the same experiment was repeated in MCF7 cells (Fig. 4H). To summarize, MDM2 induced EMT by upregulating EMT transcription factors via activation of ERK regulated by B-Raf.

As demonstrated in the aforementioned experiments, MDM2 induced EMT by upregulating EMT transcription factors through ERK. Furthermore, it was revealed that MDM2 expression was associated with the degradation of p53, thus decreasing the protein expression levels of p53 (Fig. 2E-G). Therefore, the present study aimed to explore whether the mechanism of MDM2-induced EMT depended on p53. qPCR and western blotting were used to analyze p53, E-CAD, N-CAD and VIM protein expression in control U251 cells and U251 cells overexpressing p53 together with MDM2-GFP (Figs. 5A and S2E). As shown in Fig. 5, overexpression of p53 rescued the decrease of E-CAD and inhibited the increase of N-CAD and VIM (Fig. 5A). These results indicated that the overexpression of p53 inhibited EMT induced by MDM2. In other words, MDM2-induced EMT was p53-dependent. Previous studies have reported that 14-3-3 is one of the most important p53 transcriptional targets (56–59). Notably, 14-3-3 can deactivate Raf directly and prevent its activation (24–27). The present study used western blot analysis to determine the expression levels of 14-3-3 in stable MDM2-overexpressing U251 cells and control U251 cells. As shown by the semi-quantitative analysis data, 14-3-3 expression was markedly decreased in stable MDM2-overexpressing U251 cells compared with the control group (Fig. 5B). Furthermore, as the most important downstream autophagy transcription factor of 14-3-3, transcription factor EB (TFEB) expression was decreased following MDM2 overexpression in the U251 cell line (data not shown). Previously, the relationship of 14-3-3 with B-Raf has been discussed, and it has been revealed that lower 14-3-3 expression is associated with an activated B-Raf signaling pathway (39), which was consistent with the present B-Raf signaling pathway data. Therefore, the present study aimed to verify whether MDM2 activated the B-Raf signaling pathway via 14-3-3. 14-3-3 overexpression was checked by qPCR (Fig. S2F) and MDM2, 14-3-3, E-CAD, N-CAD, VIM and p-B-Raf protein levels were analyzed by western blotting in control cells and cells stably overexpressing MDM2 together with Flag-14-3-3 (Fig. 5C). Overexpression of 14-3-3 inhibited the activation of the B-Raf signaling pathway. Notably, overexpression of 14-3-3 rescued the decrease in E-CAD and inhibited the increase in N-CAD and VIM, which meant that overexpression of 14-3-3 could prevent MDM2-induced EMT (Fig. 5C). The colony formation abilities of control cells and MDM2-overexpressing U251 cells with or without FR180204 were determined. As shown in the quantification data, MDM2 overexpression significantly increased the colony formation rate. However, this increased colony formation rate induced by MDM2 could be inhibited by treatment with an ERK inhibitor (Fig. 5D). Scratch wound assays were used to detect the migration abilities in control cells and MDM2-overexpressing U251 cells with or without FR180204 treatment. The percentage closure of the scratch was quantified at 0, 24 and 48 h after wounding. As shown in the images of the migration and the quantification data, the percentage closure of the scratch in the MDM2 overexpression group was higher than that in the control group at 24 and 48 h post wounding (Fig. 5E). However, there was no significant difference between MDM2 U251 cells treated with FR180204 and the control group, which meant that the migration ability was inhibited by ERK inhibition in MDM2 U251 cells (Fig. 5E). To summarize, MDM2 induced EMT by activating the B-Raf signaling pathway through 14-3-3, which depended on p53.

To assess whether MDM2 activated B-Raf signaling function in tumor progression, a xenograft model was established in NOD/SCID mice with U251 tumors. In agreement with the colony formation data (Fig. 5D), MDM2 expression accelerated tumor growth. By contrast, tumor growth was significantly inhibited when MDM2 was overexpressed together with 14-3-3 or when treated with an ERK inhibitor (Fig. 6A-C). These results indicated that the MDM2-activated B-Raf signaling pathway promoted clonogenicity and glioma xenograft tumor growth.

Immunohistochemical analyses also demonstrated increased N-CAD expression and decreased E-CAD expression in U251 ×enografts expressing MDM2, which represented the occurrence of an EMT process (Fig. 6D). Loss of E-cadherin has been associated with accelerated tumor progression and metastasis (6,7). By contrast, EMT was inhibited in xenografts overexpressing MDM2 together with 14-3-3 or those treated with an ERK inhibitor (Fig. 6D). This demonstrated a partial EMT-like process in MDM2 tumors, which was reversed in tumors which overexpressed Flag 14-3-3 or were treated with FR180204. Collectively, these results indicated that the B-Raf signaling pathway serves as a downstream effector of MDM2, contributing to tumor progression and higher metastatic potential.

It has been reported that EMT contributes to cancer cell resistance to chemotherapy agents (30–32). According to previous studies, numerous chemicals have little or no effect on cells that have undergone EMT in a number of cancer types, for example breast cancer cells and glioma (30–32). Therefore, the present study aimed to explore whether MDM2 could promote cell insensitivity to drug treatment by inducing EMT. Silibinin, also known as silybin, is a promising cancer treatment drug (60,61). The present study aimed to determine whether overexpression of MDM2 could affect drug sensitivities. First, stable overexpression MDM2 U251 cells and control cells were treated with indicated doses of silibinin (0, 10, 20, 40, 60, 80, 100 and 200 µM) and cell viability was determined using an MTT assay after 24 h. As shown in the quantification data, control cells were more sensitive to 40, 60, 80 and 100 µM silibinin treatment than MDM2-U251 cells at 24 h (Fig. 7A). Subsequently, an MTT assay was used again after 48 h in these two cell lines. Similar to the data from 24 h, control cells were more sensitive to silibinin treatment than MDM2-U251 cells at 48 h (Fig. 7B). This suggested that MDM2 promoted cell insensitivity to silibinin by inducing EMT in the U251 glioma cell line. Furthermore, this was verified in an A549 cell model. As shown in the quantification data, control cells were more sensitive to 20, 40, 60 and 80 µM silibinin treatment than MDM2-GFP cells at 48 h in A549 cell model (Fig. 7C). Control cells were more sensitive to 20–200 µM silibinin treatment than MDM2-GFP cells in an MCF7 cell model (Fig. 7D). According to previous literature reports, silibinin inhibits the relevant pathways for angiogenesis and decreases the expression of VEGF and other related genes (35,36). Therefore, total mRNA was subjected to SqRT-PCR to analyze changes in VEGF expression in U251 cells (Fig. 7E). MDM2, β-arrestin1, β-arrestin2, MMP-2, MMP-9 and VEGF expression was analyzed by qPCR (Fig. 7F). According to SqRT-PCR and qPCR, MDM2 overexpression rescued the decreased expression of β-arrestin1, β-arrestin2, MMP-2, MMP-9 and VEGF, which are downstream genes of silibinin. Subsequently, the changes in VEGF expression were also verified by SqRT-PCR in A549 and MCF7 cell models (Fig. 7G and H). Therefore, MDM2 induced EMT by upregulating EMT transcription factors via activation of the B-Raf signaling pathway through 14-3-3, and this depended on p53. Subsequently, the EMT induced by MDM2 could promote cell insensitivity to drug treatment (Fig. 8).