Notch promotes tumor growth and invasion by directly targeting Twist1 in cancer stem cells derived from primary and recurrent glioblastomas

Formation of neurosphere-like colonies derived from primary and recurrent GBM patients

First, neurosphere-like colonies were obtained from GBM patients. Ten GBM specimens from 5 patients with primary and recurrent GBMs were collected in this study. Tumor tissues were mechanically dissociated into single cell suspensions and cultured in serum-free DMEM/F12 medium supplemented with EGF, bFGF, and B-27 supplement. Results from microscopic observations confirmed that three primary GBMs (GSC2, GSC3, and GSC5) and four recurrent GBMs (GSC1R, GSC2R, GSC3R, and GSC5R) showed proliferating tumor spheres, and the typical morphology of neurospheres, as shown in Figure 1A (indicated by cyan arrows). Therefore, we studied the proliferative potential and in vitro tumorigenicity experiments on these sever GSC lines. Neurospheres derived from recurrent GBMs were noted within 24 h of culturing, which was significantly more rapid than neurospheres derived from primary GBMs (within 72 h; Figure 1B); owever, all cell lines showed similar exponential growth patterns within 7 days, while GSC3 cells exhibited a slightly slower growth rate (data not shown).

Figure 1. Patient glioma-derived stem cells formed neurosphere-like colonies and demonstrated characteristics of NSCs. (A) Morphology of typical primary neuorspheres derived from GBM tissues. The scale bar in high-magnification images (cyan arrow pointed region in low-magnification image) represent 50 μm. (B) The neurosphere formation time on different GSCs. Five primary and paired recurrent GBM tissues were obtained and cultured for at least 7 d. Neurospheres from 4 recurrent and 3 primary GBMs were formed. The recurrent GSCs showed higher formation ability compared with the primary GSCs (*: P < 0.05 and **: P < 0.01). (C) Immunofluorescent staining of primary tumor spheres with stem cell markers (CD133 [green]) and Nestin [red]) confirmed stemness of neurospheres. (D) The neurospheres were cultured in differentiation medium for 7 d, and differentiation markers (GFAP [green] and beta-tubulin III [red]) were immunofluorescence-stained. The obtained neurospheres maintained the typical multipotency ability. The scale bar in high-magnification images (white arrow in low-magnification images) represents 20 μm.

To further confirm whether the tumor spheres-maintained neuro stem cell (NSC) characteristics, the NSC specific markers, CD133 and Nestin, were measured using immunofluorescence. As shown in Figure 1C, the primary tumor spheres had a high level of CD 133 (green) and Nestin (red) expression, which suggested that these cells were NSCs. Then, a differentiation assay was performed to determine the multipotency of these GBM-derived neurospheres. After treatment with differentiation-inducing medium for 7 d, different molecular markers on neurospheres were verified. The astrocyte-specific marker, GFAP, as well as the neuron-specific marker, beta-tubulin III, were observed under fluorescence microscopy (Figure 1D), demonstrating that neurospheres differentiated into astrocytes and neurons under the differentiation conditions. Because only a few neuronal markers were studied, further investigations with additional neuronal markers or glioblastoma cells as controls to compare with neurospheres are warranted. It should be noted that these tumor spheres obtained from mechanical dissociation could be cultured for at least 5 generations (approximately 5 weeks), and the tumor spheres will maintain stemness and multipotency in serum-free medium in the presence of growth factors (data not shown).

Upregulation of Notch and Twist1 genes in neurospheres from primary or recurrent GBM

After verifying the stemness and multipotency of GBM-derived neurospheres, the global gene expression on these GSCs was evaluated using an Affymetrix 133 plus 2.0 microarray. In general, there were 1527 genes with significant transcriptional changes in primary and recurrent GSCs, of which 673 genes were upregulated and 854 were downregulated. Specifically, the genes associated with glioma development, including Notch, NF-kB, TGF-β, STAT3, and HIF1-α, were prominently up- or down-regulated in GSCs (Supplementary Information 1 and 2). Notch signaling-associated genes, such as Notch1, Notch3, Notch4, DLL3, Jagged1, Hes1, and RBPJ, were significantly enriched. In addition, the Twist1 gene, an important member in the HLH family related to glioma invasion, was also upregulated (Figure 2A). It is notable that the levels of Notch1, Jagged1, Twist1, and Hes1 mRNA expression in recurrent GSCs were significantly higher than primary GSCs (Figure 2B–2E). The other Notch signaling molecules investigated were not statistically different between primary and recurrent GSCs.

Figure 2. Global gene expression profiles in GSCs from primary and recurrent gliomas. (A) Global gene expression was assayed using the Affymetrix 133 plus 2.0 microarray. The Notch signaling-related and Twist1 genes were notably upregulated. (B–E) Notch signaling-related genes and Twist1 had significant differences in mRNA levels between primary and recurrent GSC lines. Data are presented as the mean ± SD. *: P < 0.05 and **: P < 0.01.

Twist1 is in positive correlation with Notch activation in GSCs

Because Notch and Twist1 genes were significantly upregulated in GSCs, we then investigated the relationship between Notch and Twist by modulating the level of Notch intracellular domain (NICD) expression in GSCs. GSCs were transduced with NICD-lentivirus to induce NICD expression or treated with the Notch antagonist, DAPT, to block Notch signaling. DAPT is a typical Notch signaling pathway inhibitor that has been used in many cell types, including glioblastomas [34]. Twist1 expression was then determined by RT-PCR and Western blot analysis. As shown in Figure 3A and 3B, and Supplementary Figure 1, Twist1 protein and mRNA levels were increased in GSCs on day 2 after NICD transduction, and maintained a high level of expression up to day 7. In contrast, when GSCs were treated with the Notch antagonist, DAPT, for 2 days, Twist1 expression was significantly decreased because the Notch signal pathway was significantly blocked by DAPT (Figure 3C and 3D, and Supplementary Figure 2). Interestingly, the level of Twist1 protein expression in GSCs maintained the original level of expression on day 1 of NICD transduction or DAPT treatment, then increased (NICD transduction) or decreased (DAPT treatment) expression (Figure 3A and 3C). This finding may be due to a delayed response of Twist upon Notch signaling regulation. Moreover, the data indicated that Twist1 was positively correlated with activation of Notch signaling in GSCs.

Figure 3. Regulatory relationship of Notch and Twist1 in GSCs. (A, B) Twist1 protein and mRNA levels were significantly increased following NICD lentiviral transduction from day 2-7, while NICD expression level was induced from day 1-7. (C, D) A significant decrease in the levels of NICD and Twist1 protein and mRNA expression in GSCs after treatment with DAPT. Data are presented as the mean ± SD. *: P < 0.05 and **: P < 0.01.

Notch/Twist1 promotes GSCs proliferation and migration

Having verified that Twist1 was positively correlated with activation of Notch in GSCs, the specific effects of Notch/Twist1 on GSCs biological functions were further studied. Because Notch signaling is associated with cell proliferation and migration, cell proliferation, colony formation, and cell migration assays were performed on GSCs with NICD transduction or DAPT treatment. The results showed that the GSC growth and neurosphere formation rates were significantly enhanced after NICD transduction (Figure 4A and 4B), while there was notable suppression after DAPT treatment (Figure 4D and 4E). Moreover, inducing NICD expression in GSCs significantly increased the number of GSCs invading the chamber inserts (Figure 4C), while blocking Notch signaling in GSCs strongly inhibited cell migration (Figure 4F), which was consistent with the cell proliferation data.

Figure 4. Effect of up- or down-regulation of NICD on cell proliferation and migration ability in GSCs. (A–C) Upregulated NICD promoted cell proliferation, neurosphere formation, and invasion ability of GSCs. The * represents a significant difference from vector-transduced GSCs to NICD-transduced GSCs. (D–F) DAPT suppressed the cell proliferation, neurosphere formation, and invasion ability of GSCs by the transwell assay. Data are presented as the mean ± SD. *: P < 0.05 and **: P < 0.01.

The Notch/Twist1 effects on GSC biological functions were also investigated using Twist1-downregulated or -upregulated recurrent GSCs. After transducing GSCs with Twist1-shRNA (shTwist1), the Twist levels of mRNA and protein expression were significantly decreased in recurrent GSCs (Figure 5A and 5B, and Supplementary Figure 3). After confirming Twist1 silencing in these cells, the transduced cells were re-plated and counted every day for 7 d with renewed NSC medium. As expected, the GSC growth rate was significantly inhibited after transduction of shTwist1 in GSCs (Figure 5D). Interestingly, the promoting effect of NICD on the cell growth rate by NICD-lentivirus transduction completely yielded to shTwist1 inhibition because there was no significant difference in the level of Twist1 expression in cells between shTwist1-alone transduction, and NICD and shTwist1 co-transduction (Figure 5C). Similarly, colony formation and cell migration exhibited a notable decrease in shTwist1-treated GSCs (Figure 5E and 5F), as well as cells co-transduced with NICD and shTwist1. Cell proliferation and migration were significantly increased after overexpression of Twist1 on GSCs by lentivirus transduction; however, these promoting effects of Twist1 were completely suppressed by Notch antagonist DAPT treatment (Supplementary Figure 4). These findings further indicated that regulation of Notch/Twist1 strongly mediates GCA cell biological abilities, such as cell proliferation and migration.

Figure 5. Suppression of Twist1 in GSCs inhibited cell proliferation and migration ability in GSCs. (A) The levels of mRNA expression and (B) protein expression in shTwist1 treated-GSCs. (C) NICD and Twist1 mRNA expression, (D) cell proliferation ability, (E) neurosphere formation, and (F) cell migration ability in GSCs after transduction with NICD lentivirus or/and shTwist1. Data are presented as the mean ± SD. *: P < 0.05 and **: P < 0.01. NS represents no significance.

Notch/Twist1 suppresses the orthotopic glioma growth in xenografted mice

Because in vitro findings confirmed the Notch/Twist1 effects on GSC proliferation, colony formation, and cell migration, Notch/Twist1 were further evaluated in a glioma mouse model. Nude mice were injected intracranially with NICD-transduced GSC3R or NICD-shTwist1-transduced GSC3R, and tumor size was monitored by bioluminescence imaging. As shown in Figure 6A, the luminescence intensity was much lower in NICD-shTwist1-treated group when compared with the NICD alone-treated group, which demonstrated that downregulation of Twist1 inhibited glioma growth. A significantly prolonged survival was observed in mice with NICD-shTwist1 GSCs injection when compared with mice injected with NICD GSCs (Figure 6B). This finding was potentially attributed to the lower Twist1 mRNA level of mouse tumors in the NICD-shTwist1 group compared to the NICD group (Figure 6C). These in vivo data clearly demonstrated that regulation of Notch/Twist1 can significantly mediate tumor growth in tumor-bearing mice, which supports the in vitro data shown above.

Figure 6. Effect of Notch/Twist1 on orthotopic glioma tumor-bearing mice. (A) Representative bioluminescent images showed tumor growth 8 d after tumor implantation in mice. (B) The survival of mice in different groups. The curve was quantified with a Kaplan-Meier survival curve using Graphpad Prism 8 software. (C) The tumors were collected after mice were sacrificed, and Twist1 mRNA levels in the different groups were measured by RT-qPCR. Data are presented as the mean ± SD. ***: P < 0.001.

Twist1 is directly regulated by Notch signaling

To confirm whether Twist1 serves as a potential direct downstream target of Notch signaling in GSCs, a luciferase reporter driven by a fragment containing one potential NICD/RBPJK-binding-site sequence in the Twist1 promoter as wild type was constructed (Figure 7A). The mutant type with two nucleotide mutations in the binding site was used as a control. Then, the GSCs were co-transfected with NICD and wild or mutant type. The cells were then harvested for luciferase activity measurement. As shown in Figure 7B, a 7-fold increase in luciferase activity in cells with a wild type NICD/RBPJK binding site after NICD transfection when compared with vector transfection; however, there was no significant difference in cells with a mutant type NICD/RBPJK binding site between NICD transfection and vector transfection, suggesting that the NICD/RBPJK binding site is necessary for inducing Twist1 promoter activity.

Figure 7. RBPJK binding site in Twist1 promoter is required for molecular interaction with NICD. (A) One possible evolutionary-conserved NICD/RBPJK binding site in Twist1 promotor was revealed using bioinformatic analyses, and mutations of the binding site which are highlighted in red (GG→TT). (B) Measurement of co-transfected firefly luciferase activity in GSCs after co-transfected NICD and wild- or mutant-type Twist1 promoter by bioluminescence assay. The luciferase activity was calculated with an absolute value of the bioluminescent intensity. (C) CHIP assay showed NICD/RBPJK-precipitated fragment binding site. IgG was used as a negative control. Data are shown as the mean ± SD. *: P < 0.05; NS represents no significance.

To further determine whether NICD directly binds to the putative RBPJK binding site in Twist1 promoter, a chromatin immunoprecipitation (CHIP) assay was performed. Soluble chromatin from GSC cells transduced with or without NICD lentivirus was extracted and anti-NICD antibody was used to immunoprecipitate chromatin. Then, the chromatin complex was analyzed by semi-quantitative PCR using primers flanking the RBPJK binding site in Twist1 promoter. In the NICD-upregulated cells, the RBPJK mRNA was significantly enhanced when compared with vector transduction (Figure 7C), indicating a sequence-specific association of NICD with RBPJK in the Twist1 promoter of GSCs. These results suggested that NICD directly bound to the Twist1 promoter, and thus regulated GSC function.

Patient specimens and cell culture

This study was approved by the Ethics Committee of the Sixth Medical Center (Chinese PLA General Hospital, city, China) and conducted with informed consent from all patients. The specimens were obtained from five patients with primary and recurrent gliomas and undergoing surgical operations in the Department of Neurosurgery at the Sixth Medical Center. The glioma stem cells (GSCs) were derived from glioma tissues and cultured, as described previously [7, 13] Briefly, the glioma tissues were mechanically-dissociated with a 200-μL pipette tip. The neurosphere-forming cells were then obtained and cultured in serum-free Dulbecco’s modified Eagle’s medium (DMEM)/F12 medium (Invitrogen, Carlsbad, CA, USA) supplemented with 20 ng/ml of epidermal growth factor (EGF; Peprotech, Rocky Hill, NJ, USA), 20 ng/ml of basic fibroblast growth factor (bFGF; Peprotech), and B-27 supplement (1:50; Gibco, Gaithersburg, MD, USA). The GSCs were designated GSC1, GSC2, GSC3, GSC4, and GSC5 when derived from primary glioma tissues, and GSC1R, GSC2R, GSC3R, GSC4R, and GSC5R when derived from recurrent glioma tissues.

Neurosphere formation and immunofluorescent analysis

The dissociated cells were seeded at 1000 cells/well in 96-well plates in 100 μL of stem cell medium and cultured for 7 days. The number of neurospheres was manually counted under a dissection microscope with 4×4 objective magnification. The expression of stem cell markers (CD133 and Nestin) and differentiation markers (the astrocyte-specific marker, glial fibrillary acidic protein (GFAP), and the neuron-specific marker, beta-tubulin III) on the GSC neurospheres were analyzed using an immunofluorescence assay. Briefly, neurospheres were placed into cell inserts (Millipore, city, state, USA) and fixed with 0.4% paraformaldehyde (PFA; Solarbio, city, China) at 4 °C for 10 min. After thrice-rinsing with PBS, the neurospheres were incubated with mouse anti-CD133 (1:500; Abcam, city, UK) or mouse anti-Nestin (1:500; Abcam) overnight at 4 °C. For detection of cell differentiation, neurospheres were cultured in DMEM/F12 medium containing 10% fetal bovine serum (FBS) for 7 d. Then, the neurospheres were collected and fixed with 0.4% PFA. After rinsing with PBS, the neurospheres were incubated with mouse anti-GFAP (1:500; Sigma, St. Louis, MO, USA) or rabbit anti-beta-tubulin III (1:1000; Sigma) overnight at 4 °C, and Cy3-conjugated goat anti-rabbit IgG (1:500; Jackson ImmunoResearch, West Grove, PA, USA) or FITC-conjugated goat anti-mouse IgG (1:200; Chemicon, Temecula, CA, USA) as secondary antibodies. Then, the neurospheres were stained with DAPI solution (Solarbio) and fluorescently-imaged with a fluorescence microscope (BX-41; Olympus, city, Japan).

Gene-expression mcroarray analysis

Affymetrix 133 Plus 2.0 CEL files were processed using the MAS5 algorithm and probe sets were converted to a refseq transcript ID using a custom chip description file (CDF). To identify differentially-expressed genes between primary and recurrent GBM-derived GSCs, gene set enrichment analysis (GSEA) was performed. The resulting P values were corrected using the method of Benjamini and Hochberg for conversion to a false discovery rate.

RNA isolation and real-time quantitative polymerase chain reaction (RT-qPCR)

Total RNA was extracted using TRIzol reagent (Invitrogen, city, state, USA), and reverse-transcribed into cDNA using the Reverse Transcription cDNA Kit (Osaka, city, Japan) according to the manufacturer’s protocol. The RT-qPCR kit (SYBR Premix EX Taq; Takara Dalian, Dalian, China) was performed on an ABI Prism 7500 real time PCR system (Applied Biosystems, Life Technologies, Carlsbad, CA, USA). Human GAPDH was used as a reference control. The PCR primers are listed in Supplementary Table 1. After the reaction, the collected PCR products were analyzed by 1% agarose gel electrophoresis. The images of Twist1 and GAPDH DNA running on the gel after the gel electrophoresis assay are shown in Supplementary Figure 6. Twist1 DNA (length = 225 bp) and GAPDH (length = 200 bp) were successfully separated by gel electrophoresis with good resolution. Basically, only one band was shown in each lane after PCR processing, which demonstrated the high reliability of qPCR assay performance and choosing GAPDH as the housekeeping gene in this study. In addition, our data are also consistent with other report results.

Notch antagonist assay

The dissociated cells were seeded at a density of 5 × 10⁴ cell/well in 24-well plates, and subsequently cultured for 7 d by renewing fresh medium supplemented with the Notch antagonist, N-[N-(3,5-difluorophenacetyl)-1-alanyl]-S-phenylglycine t-butyl ester (DAPT, 10 μM [a γ-secretase inhibitor], Selleck Chemicals, Houston, TX, USA) every 3 d. Because DAPT was dissolved in dimethyl sulfoxide (DMSO; Sigma), the cells treated with the indicated concentration of DMSO were set as controls. Then, the cells were collected for Western blot analysis.

Lentivirus transduction of cells with the NICD gene

The cDNAs encoding the notch intracellular domain (NICD) gene was cloned into the EF.v-CMV lentiviral vector according to the manufacturer’s instructions. The primer sequences for gene amplification are listed in Supplementary Table 2. After the construction, a total of 10,000 GSC cells were seeded in 12-well plates and incubated for 2 h at 37 °C prior to introduction with NICD-lentivirus in the medium containing 5 μg/mL of polybrene. Lentivirus vector-treated or non-treated cells were used as controls.

Lentivirus transduction of cells with Twist1 gene

The human Twist1 cDNA clone (GenBank ID BC036704) was purchased from the ATCC (city, state, USA). The open reading frame (ORF) was amplified with PCR primers: 5‘-cgccaccatgatgcaggacgtgtccagctcg-3’; and 5’-actagtgggacgcggacatggaccag-3’. The coding region of the full-length human Twist1 was inserted into pLenti-GFP to construct pLenti-Twist1. Viral packaging service was provided by Genechem (Shanghai, China). Lentivirus infection was performed using a kit (Genechem) according to the manufacturer’s instructions. Then, GSCs were co-cultured with viral suspensions for 48 h and cultured in normal medium for further analysis.

Cell transduction with Twist1-shRNA

Transduction of Twist1 short hairpin RNA (shRNA) sequences were designed using online RNAi design software. The non-coding RNA (ncRNA) with the same bases served as controls. The shRNA and ncRNA oligos are listed in Supplementary Table 3. Construction of the lentiviral vector, pLKO.1, with Twist1-shRNA or ncRNA was performed according to the manufacturer’s instructions. Human embryonic kidney cells (HEK293FT) were co-transfected with the shRNA lentiviral plasmid (pLKO.1-puro) and Vira Power Lentiviral packaging mix (pLP1, pLP2, pLP-VSV-G; Life Technologies) using Lipofectamine 2000 (Invitrogen). After transfection for 48 h, the lentiviruses were harvested, tittered, and stored at −80°C. GSC2, GSC2R, GSC3, GSC3R, GSC4, and GSC4R were transduced with Twist1-shRNA or ncRNA lentiviruses. Tumor spheres were dissociated into single cells to increase the transduction efficiency and uniformity. After transduction with lentivirus for 24 h, cells were washed and cultured for an additional 2 d with NSC medium. The resulting cells were collected and analyzed for Twist expression by RT-PCR analysis and Western blotting assays.

Cell proliferation assay

Primary or treated neurospheres were harvested and mechanically dissociated into single cells. The Twist1 lentivirus-transduced GSCs were treated with DAPT for 72 h before performing the cell proliferation assay. Briefly, the cells were seeded at 5 × 10⁴ cells/well in 24-well plates and cultured for a total of 7 d. The cell number was counted every day using a hemocytometer. The experiment was performed in triplicate. The cell doubling time was calculated using GraphPad Prism 8 software with exponential equations and exponential growth patterns, and the data are shown in the supporting information (Supplementary Figures 7 and 8).

Transwell invasion assay

The transwell invasion assay was performed using 24-well transwell inserts (diameter 8 mm; Corning, NY, USA) precoated with Matrigel (60-80 μL; Corning) on the upper surface of the polycarbonic membrane (pore size, 8 μm). Briefly, the dissociated cells were seeded at a density of 5 × 10⁴ cells/insert in the upper chambers in 200 μL of EC medium. The lower compartments were filled with 500 μL of NSC medium, as indicated. After incubation for 48 h at 37°C, the cells migrating to the lower chambers were harvested and counted under a microscope (Nikon, Tokyo, Japan).

Western blot analysis

Cells were collected and lysed in RIPA buffer on ice for 30 min. The total protein concentration was determined using the BCA Protein Assay Kit (Solarbio). Equal proteins were loaded and separated by gel electrophoresis, and transferred to PVDF membranes (Millipore), The membranes were incubated with the indicated primary antibodies overnight at 4 °C. The following primary antibodies were used: anti-NICD1 (NOTCH intracellular domain 1; Cell Signaling and Technology, Boston, MA, USA); anti-HES1 (Abcam, Cambridge, MA, USA); anti-Twist1 (Abcam); and anti-β-Actin (Zhongshan Golden Bridge Biotechnology, city, China). Then, horseradish peroxidase (HRP)-conjugated goat anti-mouse or anti-rabbit IgG antibodies (1:2000; Zhongshan Golden Bridge Biotechnology) were added and incubated for additional 2 h. The immunocomplex on the membrane was visualized under an Image Quant LAS 4000 mini (GE, city, state, USA) with HRP substrate luminol reagent and peroxide solution.

Luciferase reporter construction

The full-length Twist1 promoter was amplified from the genomic Twist1 plasmid using the forward primer, 5′-TGCCACGTTTTGTTTCCAAG-3′, and reverse primer, 5′-AGGTGTCTGAGAGTTGGGC-3′. Then, the PCR product was digested with enzymes (NheI and HindIII) and constructed into the pGL-3 vector with firefly luciferase reporter gene (Promega, city, state, country). The positive clone was designated as the Twist1-luc reporter plasmid. To generate the Twist1 −1432/−1426 mutant, the forward primer, 5′-AAGGTCCTCAGCTTG TTAACAGGGCAGGAGTTCGAGACGC-3′, and reverse primer, 5′-GCGTCTCGAACTCCTGCCCTGTTAACAAGCTGAGGACCTT-3′, were used. Site-directed mutagenesis was performed using the QuickChange Site-directed Mutagenesis Kit (Stratagene, La Jolla, CA, USA). All constructed plasmids were subjected to DNA sequencing to confirm that no PCR artifacts were generated.

Plasmid transfection and luciferase assay

The GSC3R cells were prepared at a density of 5×10³ cells/well in 12-well plates and co-transfected with Twist1-luc reporter plasmids (500 ng/well) and 3XFlag NICD expression plasmid by Lipofectamine 2000. After 48 h of transfection, the cells were lysed, washed twice with ice-cold PBS, and resuspended with a lysis buffer (Promega). The firefly luciferase activity in the lysed solution was determined by luminescence measurement with luciferin, and obtained as an absolute value of the bioluminescent intensity.

Chromatin immunoprecipitation (CHIP) assay

The CHIP assay was performed using a kit (Millipore) according to the manufacturer’s protocol. The primer sequences for Twist1 promoter were as follows: forward, 5′-ACCCAGAAAGGGGCACTTG-3′; and reverse, 5′-GGCAATTCTGGGTCC GGA-3′. The 180-bp PCR product was quantified by real time PCR, as previously described [51].

Intracranially glioma-xenografted mouse model

Four-week old female immunocompromised nude mice were maintained in the Experimental Animal Center at the Sixth Medical Center (Chinese PLA General Hospital) in a specific pathogen-free environment. All of the animal experiments reported herein were carried out in accordance with the approved guidelines and approved by the Committee on the Animal Care and Use of Sixth Medical Center (Chinese PLA General Hospital). Mice were anaesthetized, placed in a stereotactic frame (RWD Life Science, city, China). Then, 5 ×10⁵ luciferase-NICD- or luciferase-NICD-shTwist1-transfected GSC3R in 10 μL of PBS were injected intracranially into the caudate nucleus of mice at a rate of 0.5 μL/min using a guide-screw system (n = 10 for each group). Follow the injection, the needle remained at the injection site for at least 5 min before removal. Bioluminescence monitoring of tumor growth in mice was performed 8 d after tumor implantation by intraperitoneal injection of 15 mg/mL of D-luciferin into anesthetized mice and imaging on the IVIS Spectrum Live Imaging System (Tianjin Medical University, city, China). Images were analyzed using Live Image 4.4 software. After the mice were sacrificed, the tumors were collected and the Twist1 mRNA levels were detected using RT-qPCR.

Statistical analysis

All experiments were carried out at least in triplicate and data in this study are presented as the mean ± SD. Independent two-tailed Student’s t-test or a two-way analysis of variance (ANOVA) was used to assess statistical significance of the results (*: P < 0.05, **: P < 0.01, and ***: P < 0.001).