Cullin-7 (CUL7) is overexpressed in glioma cells and promotes tumorigenesis via NF-κB activation

The data for the mRNA expression microarrays and the attendant clinical information for the samples were downloaded from The Cancer Genome Atlas Research Network (n = 603; TCGA, http://​cancergenome.​nih.​gov) and were used for the analysis. In addition, the Chinese Glioma Genome Atlas (n = 301; CGGA, http://​www.​cgga.​org.​cn), an external independent glioma database, was also mined. Archived paraffin embedded glioma tissues (WHO grades I–IV) were gathered from patients (n = 38) who underwent surgery in the Department of Neurosurgery, Qilu Hospital of Shandong University. Normal brain tissue samples (n = 4) were collected from severe traumatic brain injury patients who experienced partial resection of the normal brain as decompression treatment.

Sections were obtained from formalin-fixed, paraffin-embedded tissues of different grades of human gliomas and normal brains. Sections were heated, deparaffinized, rehydrated and placed in sodium citrate buffer (pH 6.0) for antigen retrieval, and endogenous HRP activity was blocked with 3% hydrogen peroxide (H₂O₂). The slides were blocked with 10% normal goat serum and incubated with primary antibodies (mouse anti-CUL7 antibody, Santa Cruz, USA; rabbit anti-Ki67 antibody, Cell Signaling Technology, USA) at 4 °C overnight. The signal was visualized using standard protocols with horseradish-peroxidase-conjugated secondary antibodies and 3, 3′-diaminobenzidine (DAB) as the substrate. The sections were incubated with normal mouse serum as negative controls. Then, the slides were counterstained with haematoxylin, and typical images were obtained using a Leica DM 2500 microscope. The protein expression were scored by Image Pro Plusversion (IPP) image processing software.

To gain insight into the biological processes and signalling pathways associated with CUL7 expression in gliomas, GSEA was performed using the Broad Institute GSEA version 4.0 software. The TCGA database was downloaded. The gene sets used for the enrichment analysis were downloaded from the Molecular Signatures Database (MsigDB, http://​software.​broadinstitute.​org/​gsea/​index.​jsp).

U87MG, U251 and A172 human glioma cell lines and normal human astrocytes (NHA) were obtained from the Culture Collection of the Chinese Academy of Sciences (Shanghai, China) and cultured in Dulbecco’s modified Eagle’s medium (DMEM; Thermo Fisher Scientific; USA) that was supplemented with 10% foetal bovine serum (FBS; Gibco; USA). GSC 267 were kindly provided by Dr. Krishna P.L. Bhat (The University of Texas, M.D. Anderson Cancer Center, Houston, TX) and cultured in DMEM/F12 media supplemented with B27 (Invitrogen, USA), EGF (R&D Systems, USA), and bFGF (R&D Systems, USA). These cells were maintained at 37 °C in a humidified chamber containing 5% CO2.

Small interfering RNA (siRNA) targeting CUL7 were synthesized (GenePharma; Shanghai, China). siRNAs were transfected with Lipofectamine™ 3000 reagent (Thermo Fisher Scientific; USA) according to the manufacturer’s protocol. Stable knockdown of CUL7 in cells was generated using lentiviral transduction of shCUL7 (Genepharma; Shanghai, China). Knockdown efficiency was assessed 48 h after transfection by western blotting. Negative control (NC) sequences are as follows: 5′-UUCUCCGAACGUGUCACGUTT-3′; siRNA sequences that generated efficient knockdown are as follows: si-CUL7#1: 5′-UGAGAUCCUAGCUGAACUGTT-3′; and si-CUL7#2: 5′-AGAACUCCGCUACAGGGAAUU − 3′. Plasmid construction of GV141-CUL7 was performed by GeneChem (Shanghai, China). Cells were transfected with GV141-CUL7 to induce the overexpression of CUL7 and with empty GV141 vector (GV141) as a control.

Total RNA was isolated from glioma cells using TRIzol reagent (Invitrogen, Life Technologies). Reverse transcription was performed using High Capacity cDNA Reverse Transcription Kits (Applied Bio-systems) according to the manufacturer’s protocols. The cDNA was subjected to real-time PCR using the quantitative PCR System Mx-3000P (Stratagene). The sequences of the PCR primers are as follows: hsa-miR-3940-5p F primer, 5′-TAAAAGTGGGTTGGGGCGG-3′, R primer, 5′-GTGCAGGGTCCGAGGT-3′; U6 F primer, 5′-CAGCACATATACTAAAATTGGAACG-3′, R primer, 5′-ACGAATTTGCGTGTCATCC-3′; CUL7 F primer, 5′ -ACCTGAAGGCGGTCTCTGT-3′, R primer, 5′-CCTTGCTGCCATCTCGAATC-3′; MST1 F primer, 5′-GGCCTTCCACTACAACGTGA-3′, R primer, 5′-GCAGGTCCGTACGTAGTCTTT-3′; GADPH F primer, 5′-GCACCGTCAAGGCTGAGAAC-3′, R primer: 5′-TGGTGAAGACGCCAGTGGA-3′. The relative miRNA/mRNA expression normalized to U6/GAPDH was analysed using GraphPad Prism 6 software.

Harvested cells were lysed with heat denaturation in RIPA cell lysis buffer. Protein lysates (20 μg) were loaded and separated on SDS-PAGE, and the proteins were transferred to polyvinylidene difluoride (PVDF) membranes. The blots were incubated with primary antibodies against CUL7, cleaved-PARP, Ub (Santa Cruz, USA); p21, p27, Cyclin D1, CDK4, Cyclin E1, CDK2, MST1, Phospho-IKK, IKKα, Phospho-IκBα, IκBα, Phospho-NF-κB p65, NF-κB p65, Bcl-2, Bax, cleaved-caspase3, N-cadherin, Vimentin, Slug, Twist, MMP2, GAPDH (Cell Signaling Technology, USA); and E-cadherin, MMP9 (Proteintech, China). Specific proteins were visualized with enhanced chemiluminescence (ECL, Millipore, Bredford, USA). The intensity of the protein bands was measured (ImageJ software) and normalized to GAPDH.

Cells were lysed in IP buffer (Pierce, Rockford, USA) including protease inhibitor cocktail (Sigma). The lysates were incubated with 5 μg appropriate primary antibodies or IgG and Protein A/G agarose beads (Pierce) overnight at 4 °C with gentle shaking. The immunoprecipitated complexes were then washed with lysis buffer three or four times and eluted from the beads with protein loading buffer. Western blotting analysis was then performed for detection of proteins.

Cells were infected with siRNA targeting CUL7. After 48 h, CHX (25 μg/mL) was added to the culture medium to inhibit translation, and cells were incubated for 0, 2, 4, or 6 h. Cell lysates were prepared, and protein was examined using western blot analysis. Experiments were performed in triplicate.

The Cell Counting Kit-8 (CCK-8) was used to evaluate the cell viability according to the manufacturer’s instructions (Dojindo, Japan). U87MG or U251 cells (5 × 10 ³ cells/well) were incubated in 96-well plates for 24, 48, and 72 h. The CCK-8 solution (10 μL) was added to each well and the plates were incubated for 1 h at 37 °C, and then the absorbance at 450 nm wavelength (OD450) was measured in a Microplate Reader (Bio-Rad).

Cell proliferation rates were measured by an EdU cell proliferation assay kit (RiboBio, #C10310–1; China). Cells were incubated with 200 μL of 5-ethynyl-20-deoxyuridine (EdU) for 2 h at 37 °C. Cells were fixed in 4% paraformaldehyde for 20 min, permeabilized with 0.4% Triton X-100 for 10 min, and incubated with Apollo® reagent (100 μL) for 30 min. The cells were stained with Hoechst for 30 mins, and representative images were obtained using a Nikon inverted fluorescence microscope. The cell proliferation rate was assessed using the ratio of EdU positive cells (Red) to the total Hoechst positive cells (Blue).

Cell cycle analysis was performed by determining the DNA content with propidium iodide (PI) staining (BD Biosciences; San Jose, CA, USA). Briefly, U87MG and U251 glioma cells were harvested, re-suspended and stained with propidium iodide (PI; BD Biosciences) in the presence of RNase A for 20 min. Apoptosis was evaluated in the U87MG and U251 cells with Annexin V-FITC and PI staining (15 min; BD Biosciences). Cells were analysed using a flow cytometer (BD Biosciences) according to the manufacturer’s instructions.

For the colony formation assay, cells were seeded in 6-well plates at a density of 2000 cells/well. The DMEM containing 10% FBS was changed every third day. After 15 days, the colonies were fixed with 4% paraformaldehyde for 30 mins and stained with crystal violet for 15 mins, and representative colonies were imaged and quantified.

Glioma spheroids were generated by incubating cells in the spheroid formation matrix for 72 h in a 3D culture qualified 96-well spheroid formation plate. Spheroids with a diameter of > 200 mm were embedded into the invasion matrix (Trevigen, Gaithersburg, USA) composed of basement membrane proteins in the 96-well plate. Glioma spheroids were photographed every 24 h under Nikon microscopy. The spheroids at 0 h were used as a reference point for measurement of the area invaded by sprouting cells.

To further assess invasiveness, the filters were pre-coated with Matrigel. Glioma cells were added to the top chamber in serum-free media. The bottom chamber was filled with 10% FBS DMEM. After 24–48 h of incubation, the top chamber cells were removed using a cotton swab, and the membrane was fixed in 4% paraformaldehyde for 15 min and stained with crystal violet for 15 min. Five fields of adherent cells in each well were photographed randomly. To measure migration, the filters were not pre-coated with Matrigel.

GSCs were plated in 48-well plates at a density of 2000 cells per well with 250 μL of the GSC culture media mentioned above, and the transfections were performed. Three to four days later, an additional 70 μL of the aforementioned treatment was added. Tumorsphere numbers were calculated on the seventh day after cell placement.

To establish intracranial gliomas, U87MG luciferase cells (3 × 10⁵) were transfected with Lenti-sh-CUL7 or Lenti-Control virus and then stereotactically implanted into the brains of 4-week-old nude mice (SLAC Laboratory Animal Center; Shanghai, China). Bioluminescence imaging was used to detect intracranial tumour growth on days 5, 10, 15, 20 and 25. Kaplan-Meier survival curves were plotted to determine the survival time and weight. Tumour tissues were harvested at 15 days after implantation, fixed in formalin, embedded in paraffin, sectioned and incubated with antibodies against CUL7 (Santa Cruz, USA), MST1(Abcam, UK), Ki-67 (Cell Signaling Technology, USA) and N-cadherin (Cell Signaling Technology, USA).

HEK293 cells were cotransfected with firefly luciferase reporters and the indicated plasmids using Lipofectamine 3000 (Invitrogen/Thermo Fisher Scientific), and luciferase assays were performed 24 h later using the Dual Luciferase Reporter Assay Kit (Promega). Renilla activity was used to normalize the luciferase reporter activity. The reporter genes containing GV272-CUL7 and GV272-mutCUL7 were synthesized by Genechem (Shanghai, China).

To understand the function of CUL7 in glioma development, we first analysed the mRNA levels of CUL7 in glioblastoma (GBM) tissues and normal brain tissues from the TCGA dataset. Compared to normal brain tissues, the gene expression levels of CUL7 were significantly increased in GBM tissues (Fig. 1a). We also validated these findings in the CGGA dataset (Fig. S1a).

Gliomas have been classified into three distinct molecular subtypes: proneural, mesenchymal and classical. Compared to the proneural subtypes, the mesenchymal subtype has been associated with a poor prognosis in patients [22]. Therefore, we examined the mRNA level of CUL7 in the TCGA and CGGA on the basis of the molecular subtype. The expression level of CUL7 was significantly higher (> median value) in the mesenchymal subtype than in the proneural subtype (Fig. 1b; Fig. S1b; Table 1), and the ROC curve further showed the sensitivity of CUL7 as a marker to discriminate between the mesenchymal subtype and the non-mesenchymal subtype glioma patients (Fig. 1c). Moreover, the protein levels of CUL7 were examined by immunohistochemical staining in an unconnected cohort of glioma and normal brain tissue samples from Qilu Hospital. Correspondingly, the expression of the CUL7 protein was higher in high-grade gliomas (WHO III-IV; n = 21) than in normal brain tissues (n = 4) and low-grade gliomas (WHO I-II; n = 17; Fig. 1d; Fig. S1c; Table 2). These findings were further confirmed by western blot analysis (WHO II-IV; n = 11; normal brain tissues; n = 3; Fig. 1e; Fig. S1d). Thus, high CUL7 expression was correlated with an increased tumour grade in glioma patients.

The expression levels of CUL7 were also associated with the clinicopathological characteristics of patients in the TCGA. Age (≥ 45 years) and KPS (< 80; Table 1) were statistically associated with high CUL7 expression when comparing the median values. In addition, we analysed whether CUL7 expression was correlated with some molecular genetic characteristics, including IDH1 mutations, MGMT promotor methylation, 1p/19q codeletion, low TERT expression level, and ATRX mutations, because these factors have been found to be associated with a beneficial prognosis in gliomas [23, 24]. A significantly higher number of patients with low CUL7 expression levels possessed these advantageous characteristics.

We also examined the prognostic value of CUL7 using data from LGG (n = 457) and GBM (n = 141) patients in the TCGA. Kaplan-Meier survival analysis revealed that participants with high expression levels of CUL7 had a shorter overall survival (OS) than those with low expression levels of CUL7 (63.51 vs 95.51 months, p < 0.001 in LGG, 11.73 vs 15.11 months, p < 0.05 in GBM) (Fig. 1f). These results were validated in CGGA (n = 285) (Fig. S1c). Furthermore, CUL7 expression was confirmed as an independent indicator of OS in glioma after multivariate Cox regression analysis (HR = 3.084, 95% CI = 1.918 to 4.959, P < 0.0001; Table 3). Therefore, CUL7 might be a novel prognostic biomarker in gliomas.

GSEA based on CUL7 expression in the TCGA database was performed to explore the potential biological functions of CUL7 in gliomas. Interestingly, high expression of CUL7 was associated with cell proliferation, apoptosis and epithelial-mesenchymal transition (EMT) (Fig. S2a-d). GSEA results further confirmed that CUL7 positively correlated with the Phillips-queried MES gene set but negatively correlated with the PN gene set (Fig. S2e) [25]. Therefore, CUL7 is implicated in the malignant properties of CUL7 in gliomagenesis.

Next, a series of experiments were performed to assess the role of CUL7 in several cellular processes of human glioma cells. We designed two independent siRNA sequences against CUL7. Both demonstrated efficient silencing of CUL7 in U87MG and U251 cells, two malignant glioma cell lines (Fig. 2h). To directly assess the role of CUL7 in glioma cell survival and proliferation, we performed EdU and CCK-8 assays in cells that were or were not transfected with siRNA. Downregulation of CUL7 resulted in a statistically significant decrease in the percentage of EdU-positive cells as well as changes in OD450 values in both U87MG and U251 cells 48 h after transfection (Fig. 2a-c). In addition, these findings were further confirmed by colony forming assays (Fig. S3a-b). Cell cycle analysis by flow cytometry also demonstrated that silencing CUL7 with a single siRNA increased the proportion of U87MG and U251 cells in the G0/G1 phase (Fig. 2d-e). Furthermore, knockdown of CUL7 increased the percentage of apoptotic cells in the U87MG and U251 cell lines (Fig. 2f-g). Moreover, the prominent effect of CUL7 silencing on the self-renewal abilities of GSCs was verified by a tumoursphere formation assay (Fig. S3c-d).

Next, we used western blotting to thoroughly investigate the downstream targets of CUL7 that influenced the cell cycle and apoptosis. A previous study reported that CUL7 could degrade tumour suppressors, including p21 and p27 [13, 26]. In our study, p21 and p27, which were identified as cyclin-dependent kinase inhibitors [27, 28], were present in higher amounts in the si-CUL7 group in gliomas than in the controls (Fig. 2h). In contrast, the expression of cyclin-dependent kinase 2 (CDK2)/cyclin E1 and cyclin-dependent kinase 4 (CDK4)/cyclin D1 was decreased after CUL7 silencing (Fig. 2h). Taken together, these results suggest that loss of CUL7 suppressed cell cycle progression and caused apoptosis in human glioma cells.

To assess the role of CUL7 in cell invasion and migration, 3D collagen spheroid invasion assays and Transwell chamber assays were performed. In the 3D collagen spheroid invasion assay, the area invaded by the U87MG and U251 spheroids in the CUL7 knockdown groups was reduced relative to that in the controls (Fig. 3a, c). Similarly, the results of the Transwell invasion assays also revealed that CUL7 silencing reduced the number of glioma cells invading the Matrigel-coated membrane compared with that in the control groups (Fig. 3b, d). Moreover, the same results were found with the Transwell migration assay (Fig. 3b, d).

In addition to GSEA analysis, a number of reports have demonstrated that CUL7 induces epithelial-mesenchymal transition (EMT) [15, 29], which plays a significant role in the invasiveness and metastasis of several cancers [30‐32]. The results of western blotting revealed that knockdown of CUL7 inhibited EMT by decreasing several mesenchymal markers, such as TWIST1, N-cadherin, Vimentin and Slug, and increased the expression of epithelial factors (E-cadherin) (Fig. 3e). Furthermore, CUL7 silencing induced the downregulation of the expression of MMP2 and MMP9, matrix metalloproteinase (MMP) family proteins that play important roles in ECM degradation and in the migration and invasion of tumour cells [33] (Fig. 3e). These results demonstrated that CUL7 promoted the migration and invasion of human glioma cells.

To further study the effect of CUL7 overexpression on the proliferation, invasion and migration of glioma cells, cells were transfected with the GV141-CUL7 plasmid to overexpress CUL7. EdU and CCK-8 assays were performed. Upregulation of CUL7 resulted in an increase in the percentage of EdU-positive cells and the OD450 values in both U87MG and U251 cells (Fig. 4a-c). Cell cycle analysis also demonstrated that overexpression of CUL7 decreased the population of U87MG and U251 cells in G0/G1 phase (Fig. 4d, f). Furthermore, overexpression of CUL7 decreased the percentage of apoptotic cells in the U87MG and U251 cell lines (Fig. 4e, g). In the Transwell assays, the number of migrated or invaded glioma cells in the CUL7 overexpression groups was increased compared with that in the control groups (Fig. 4h). Moreover, western blot analysis verified that the protein expression levels of cyclin E1, CDK2, N-cadherin, Slug, MMP2 and MMP9 increased while those of p27, p21 and E-cadherin decreased in the CUL7 overexpression group (Fig. S3e).

To investigate the regulatory mechanism of CUL7 in glioma progression, we used GSEA to explore the potential pathways of CUL7 and found that CUL7 was positively associated with the NF-κB pathway (Fig. 5a). Interestingly, gene silencing of CUL7 in U87MG, U251 and GSC267 cells significantly inhibited the phosphorylation of IKK-α and IκBα, the phosphorylation of NF-κB (p65), and the translocation of p65 to the nucleus, inhibiting the activity of the NF-κB pathway (Fig. 5b). In contrast, overexpression of CUL7 increased the activity of the NF-κB pathway. (Fig. S3f). These results demonstrated that CUL7 promoted the activation of the NF-κB pathway.

In 2018, Zou et al. found that CUL7 mediates the ubiquitination and degradation of MST1 [34]. Mammalian sterile 20-like 1 (MST1) functions as a suppressor in glioma [35, 36]. In addition, we detected MST1 protein levels in glioma tissues of different grades by IHC. Interestingly, the results demonstrated that MST1 protein levels were downregulated in high-grade glioma tissues (Fig. S3h). In addition, we determined if MST1 expression was correlated with expression of CD44, a representative mesenchymal marker [25, 37, 38]. As shown in Fig. S3h, expression of MST1 and CD44 were negatively correlated, which demonstrated that MST1 protein levels were downregulated in mesenchymal subtype glioma tissues. Lee et al. recently established that MST1 negatively regulates the NF-κB pathway [17]. In our study, MST1 protein levels were higher in the si-CUL7 group than in the control group (Fig. 5c). In contrast, overexpression of CUL7 decreased the protein level of MST1 (Fig. S3f). To determine the mechanism by which CUL7 promotes the degradation of MST1, we examined whether CUL7 regulates MST1 transcriptionally. However, MST1 mRNA levels were not significantly different in the si-CUL7 groups compared with the control groups (Fig. 5d). Therefore, we tested whether CUL7 modulates the degradation of MST protein. To do this, we performed a series of immunoprecipitations (IPs) to pulldown complexes by using antibodies against CUL7 or MST1. Both CUL7 and MST1 were found in complexes with antibodies against CUL7 or MST1 (Fig. 5e). In addition, the half-life of MST1 protein was altered in the si-CUL7 group. In the presence of the protein synthesis inhibitor cycloheximide (CHX), we found that the half-life of MST1 was prolonged in U87MG-, U251- and GSC267-si-CUL7 cells (Fig. 5f; Fig. S2f). Finally, endogenous MST1 ubiquitination was decreased in the si-CUL7 group (Fig. 5g). Therefore, we found that CUL7 physically associated with MST1 and furthermore led to ubiquitin-mediated MST1 protein degradation, which promoted activation of the NF-κB signalling pathway [17].

Furthermore, modified U87MG luciferase cells were implanted in nude mice orthotopically to verify the function of CUL7 in gliomas in vivo. The expression level of CUL7 was downregulated in U87MG luciferase cells transfected with Lenti-sh-CUL7 (Fig. S3g). The tumour sizes examined by bioluminescence imaging were reduced in animals bearing sh-CUL7 cells (Fig. 6a-b), and the weight loss of the sh-CUL7 group was much slower than that of the control group (Fig. 6c). HE staining of tumours collected 15 days after implantation in the two groups displayed a smaller size and more defined borders in the sh-CUL7 group (Fig. 6d). Moreover, the survival rate of animals implanted with U87MG-sh-CUL7 was longer than that of the controls (Fig. 6e). Immunohistochemistry also certified the low expression level of CUL7 in sh-CUL7 cells (Fig. 6f-g). In addition, the positive rate of Ki-67 was decreased in sh-CUL7 xenografts (Fig. 6f-g). Interestingly, MST1 expression was significantly increased in the sh-CUL7 group compared with the control group. In contrast, the expression of N-cadherin was significantly decreased in the sh-CUL7 group (Fig. 6f-g). These results demonstrated that CUL7 silencing led to reduced proliferation and invasion of glioma cells in vivo.

To explore the potential miRNAs regulating CUL7, we first used the online miRNA target prediction algorithm TargetScan. We found that miR-3940-5p was highly likely to directly target CUL7. QRT-PCR demonstrated that miR-3940-5p was significantly downregulated in glioma cell lines compared with NHA (Fig. S4a), which predicted that miR-3940-5p was a tumour suppressor. Moreover, the miR-3940-5p expression level was significantly lower in glioma than in normal brain, which was opposite to the trend in CUL7 expression level (Fig. 7a-b). Interestingly, qRT-PCR and western blotting demonstrated that overexpression of miR-3940-5p (via miR-3940-5p mimics transfected into cell lines after 48 h) induced a strong decrease in CUL7 expression in glioma cell lines (Fig. S4b-c, Fig. 7c). To further validate the direct interactions between CUL7 and miR-3940-5p, we performed luciferase reporter assays. A vector encoding the wild-type (WT) sequence of the 3′-UTR of CUL7 mRNA or a vector encoding the mutant (MUT) sequence of the 3′-UTR of CUL7 mRNA lacking the predicted miR-3940-5p target site were transfected into 293 T cells. The putative target site of miR-3940-5p in the 3′-UTR of CUL7 is illustrated in Fig. 7d. The relative luciferase activity of the WT construct of the CUL7 3′-UTR in miR-3940-5p-transfected cells decreased to approximately 40% compared with that of the control miRNA, but the MUT construct of the CUL7 3′-UTR abolished the suppressive effect of miR-3940-5p (Fig. 7e). The above results showed that CUL7 is a direct target of miR-3940-5p.

To further investigate the biological roles of miR-3940-5p in gliomas, EdU assays, CCK-8 assays, flow cytometry analysis, colony forming assays and Transwell chamber assays were conducted with the U87MG and U251 cell lines that were transfected with miR-3940-5p mimics or miR-NC. The results showed that the proliferation, migration and invasion of glioma cells were notably suppressed in the miR-3940-5p mimic group (Fig. 8a-h, Fig. S4d-e). The prominent effect of miR-3940-5p overexpression on the self-renewal abilities of GSCs was verified by a tumoursphere formation assay (Fig. S4f-g). Moreover, the results of western blotting revealed that the changes in the expression levels of MST1, NF-κB signalling pathway markers, cyclin-dependent kinase inhibitors, cyclin-dependent kinase, and EMT markers in the cells transfected with miR-3940-5p mimics were the same as those in the CUL7-silenced cells (Fig. 7f-g).

To further confirm that miR-3940-5p inhibits cell proliferation, migration and invasion of glioma by targeting CUL7, we restored CUL7 expression by transfecting U87MG and U251 glioma cells with full-length CUL7 plasmids. The EdU assay showed that the effects of miR-3940-5p on glioma proliferation were counteracted by CUL7 overexpression (Fig. S5a-b). These findings were further confirmed by CCK-8 assays (Fig. S5c). Cell cycle analysis also demonstrated that overexpression of CUL7 offset the effects of miR-3940-5p on the population of U87MG and U251 cells in G0/G1 phase (Fig. S5d-e). In addition, overexpression of CUL7 inhibited the apoptosis of the U87MG and U251 cell lines induced by miR-3940-5p (Fig. S5f-g). Furthermore, Transwell assays showed that the effects of miR-3940-5p on glioma migration and invasion were counteracted by CUL7 overexpression (Fig. S6a-b). Moreover, western blot analysis confirmed that CUL7 overexpression attenuated the effects of miR-3940-5p (Fig. S6c-d). These results indicated that miR-3940-5p inhibits glioma tumorigenesis by downregulating CUL7.