Wnt/β-catenin signaling inhibitor (ICG-001) was purchased from Selleck Chemicals. All other reagents were of the commercially available grade.

Human GBM cell line (U251MG) was obtained from the Japanese Collection of Research Bio Resources Cell Bank. Cells were cultured in the Dulbecco's modified Eagle's medium (DMEM) with 10% heat-inactivated fetal bovine serum (FBS; both from Thermo Fisher Scientific, Inc.) at 5% CO₂ and 37°C.

For siRNA transfection, U215MG cells were incubated in a six-well plate (2×10⁵ cells/well) or 24-well plate (1×10⁴ cells/well) for 24 h and transiently transfected with siRNA (20 nM) using Lipofectamine^® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) at 37°C for 48 h, according to the manufacturer's protocol. Following transfection of the cells and incubation for 48 h, the experiments were performed. Silencer Negative Control siRNA (cat. no. AM4636; Ambion/Applied Biosystems; Thermo Fisher Scientific, Inc.) was used as a control (siN, https://genesdev.cshlp.org/content/suppl/2019/06/04/gad.324814.119.DC1/Supplemental_methods.pdf) (31). Sequences of the siRNAs targeting CYLD (siCYLD) were sense, 5′-GAUUGUUACUUCUAUCAAAtt-3′ and antisense, 5′-UUUGAUAGAAGUAACAAUCtt-3′.

For plasmid DNA transfection, U251MG cells were incubated in a 12-well plate (1.6×10⁵ cells/well; Corning, Inc.) at 37°C for 24 h and transiently transfected at 37°C for 48 h with the control vector (pcDNA3) or wild-type CYLD expression plasmid (pcDNA3) (500 ng/well) (15,32) using Lipofectamine^® 2000, according to the manufacturers' protocol. The efficacy of transfections was confirmed in the present study (Fig. S1).

After 72–120 h of transfection, 50 µl/well of the Cell Counting Kit-8 solution (Dojindo Laboratories, Inc.) was added to the cells and incubated at 37°C for 2 h, and the absorbance at 450 nm was measured using EMax SOFTmaxPRO (Molecular Devices, LLC). To evaluate the effect of ICG-001 on cell viability, cells were treated with ICG-001 (0–50 µM) and control reagent [dimethyl sulfoxide (DMSO)] in serum-free DMEM after transfection at 37°C, and the absorbance 72 h after treatment was measured.

Following transfection in a six-well plate, U251MG cells were recovered by trypsin addition and suspended in serum-free DMEM, and reseeded (1.0×10⁴ cells/well) in the upper part of an 8-µm Transwell insert (Corning, Inc.) with a total volume of 200 µl/well. Cell migration was induced using DMEM containing 10% heat-inactivated FBS at a total volume of 600 µl/well. To evaluate the effect of ICG-001, ICG-001 (50 µM) or control (DMSO) with serum-free DMEM was added to the upper Transwell plate. After incubation at 37°C for 24 h, the migrating cells were stained with crystal violet at room temperature for 20 min, and the ratio of the stained area observed by light microscope was quantified using the ImageJ software (version 1.51j8; National Institutes of Health).

Following transfection in a six-well plate, U251MG cells were recovered by trypsin addition and suspended in serum-free neural stem cell (NSC) medium, and reseeded (1.0×10⁴ cells/well) in an ultra-low adhesion 96-well plate (Corning, Inc.). NSC medium including serum-free DMEM/F12, human leukemia thinner (Sigma Aldrich; Merck KGaA), human basic fibroblast growth factor (human FGF-basic; PeproTech, Inc.), human epithelial cell growth factor (human EGF; PeproTech.Inc.), heparin (Sigma Aldrich; Merck KGaA), insulin (Sigma Aldrich; Merck KGaA), N2 (Gibco), B27 (Gibco), GlutaMax (Gibco), and penicillin/streptomycin was used, as previously described (33,34). Cells with masses ≥50 µm observed by light microscope were counted as spheres.

Following transfection in a six-well plate, U251MG cells were recovered by trypsin addition, suspended in the NSC medium, and reseeded (1,500-2,000 cells/well) in the ultra-low adhesion 96-well plate (Corning, Inc.) using the serial dilution method. To evaluate the effect of ICG-001, after 24 h of incubation at 37°C in NSC medium, ICG-001 (10 µM) or control (DMSO) was added to the cells, cell masses ≥50 µm were counted as spheres, and the proportion of wells in which the spheres were formed was calculated to be the percentage/4 wells after 72 h.

Patient clinical data (RNA-sequencing data of 270 samples) were obtained from a previous study (35), and RNA sequencing (RNA-seq) data were obtained from the Ivy Glioblastoma Atlas Project (GAP) database made publicly available by the Allen Institute (2015 Allen Institute for Brain Science, Ivy Glioblastoma Atlas Project 2; http://glioblastoma.alleninstitute.org/). The Ivy Glioblastoma Atlas Project is a collaborative partnership between the Ben and Catherine Ivy Foundation, Allen Institute for Brain Science, and Ben and Catherine Ivy Center for Advanced Brain Tumor Treatment. The aforementioned study provides a foundation for GBM research (35). The Ivy Glioblastoma Atlas Project provided RNA-seq data to study the gene expression patterns of anatomical structures (identified by hematoxylin and eosin staining) and cancer stem cell clusters (identified by ISH investigation) in GBM with 270 samples in both analyses combined. The expression levels of CYLD were compared with each marker or signaling-activating factor (Ki-67, fibronectin, nestin, CK2β, and CKIε) in these samples. Correlation coefficient (r) was calculated, and | r | ≥0.2 indicated a correlation.

Whole cell lysates of U251MG cells were prepared using the phase transfer surfactant (PTS) method, as previously described (36,37). Sodium deoxycholate (SDC), sodium N-lauroylsarcosinate (SLS), ammonium bicarbonate, dithiothreitol, iodoacetamide, mass spectrometry grade lysyl endoprotease, ethyl acetate, acetonitrile, acetic acid, methanol, trifluoroacetic acid (FUJIFILM Wako Pure Chemical Corporation), modified trypsin (Promega Corporation), and 4-(2-aminoethyl) benzenesulfonyl fluoride hydrochloride (Nacalai Tesque, Inc.) were used in this study. Proteins were extracted using the PTS solution [12 mM SDC, 12 mM SLS, and 100 mM Tris-HCl (Ph 9.0)] and ultrasonically crushed for 20 min. After incubating for 5 min at 95°C, proteins in the supernatant solution were quantified using the BCA method with the BCA Protein Assay Kit (Thermo Fisher Scientific, Inc.). The proteins were reduced with 10 mM dithiothreitol for 30 min and alkylated with 50 mM iodoacetamide in the dark for 30 min at room temperature. The protein mixture was 2-fold diluted with 50 mM ammonium bicarbonate, digested with Lys-C (1/50 sample weight) at room temperature for 3 h prior to the addition of trypsin (1/50 sample weight), and incubated at 37°C for 20 h. An equal volume of ethyl acetate was added to the sample solution, and the mixture was acidified with 0.5% trifluoroacetic acid (final concentration). The mixture was shaken for 2 min and centrifuged at 15,600 × g for 2 min at room temperature. The upper layer was removed using a pipette and dried using a vacuum evaporator. The sample was then suspended in 100 µl buffer A (5% acetonitrile, 0.1% TFA) and desalted with GL-Tip SDB (GL Sciences, Inc.) and ODS-A-HG AAG12S50 (YMC CO., LTD.), as previously described (38,39). Peptides were eluted with buffer B (80% acetonitrile and 0.1% TFA). To concentrate the phosphorylated proteins, the Titasphere Phos-TiO Kit (GL Sciences, Inc.), was used according to the manufacturer's protocol. Subsequently, a 5% pyrrolidine solution was used for elution, as previously described (40). The eluted fraction was acidified with TFA, desalted using GL-Tip SDB, and concentrated in a vacuum evaporator. TripleTOF 5600 (SCIEX) equipped with Dionex Ultimate 3000 RSLS (Thermo Fisher Scientific, Inc.) was used for nano LC-MS/MS measurements. The injection volume was 5 µl, and the flow rate was 300 ml/min. A nanotrap column (100 µm ID, 2-cm length, packed with 5 µm Acclaim PepMap100C18; Thermo Fisher Scientific, Inc.) and an analytical nanocolumn (75 µm ID, 25-cm length, packed with 2 µm Acclaim PepMap C18; Thermo Fisher Scientific, Inc.) were used. MS data were acquired using Analyst Software TF 1.7 (SCIEX). For peptide identification, data were acquired in a data-dependent acquisition mode and analyzed using ProteinPilot 4.5 (SCIEX) connected to the UniProt human reference proteome database (Release 2017_11). Phosphorylation was set as the ‘Special Factor’ in the sample description. The protein identification confidence for the dataset was evaluated based on the false-discovery rate. For protein and peptide quantification, data were acquired in a data-independent acquisition mode (sequential window acquisition of all theoretical fragment ion spectra; SWATH-MS) with a variable window for the precursor ions. The proteomic datasets have been submitted to jPOSTrepo [https://repository.jpostdb.org/preview/187456970364aca64292d78; Access key: 4555; Accession no. JPST002236 (PXD043537)].

The results were analyzed using Kinase Enrichment Analysis 2 (https://www.maayanlab.net/KEA2/). By setting the library as ‘Literature Based Kinase-Substrate Library with Phosphosites’ and analyzing it using Kinase Enrichment Analysis 2, the combination of protein and phosphorylation sites was clarified. In addition, the kinases that interacted predominantly with these proteins were inferred and mapped. The results were visualized based on the Wnt/β-catenin signaling pathway-Homo sapiens (humans) in the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway.

Unpaired Student's t-test was used to evaluate the differences between the two groups. Statistical analysis was performed using Statcel (ver. 4; OMS Publishing Co., Ltd.). Data are presented as the mean ± standard deviation. P<0.05 was considered to indicate a statistically significant difference.

To develop clinically effective treatments for CYLD-negative patients with GBM, first the involvement of CYLD downregulation in the malignant characteristics of GBM cells was determined. As revealed in Fig. 1A, GBM cell proliferation was significantly promoted in CYLD-silenced GBM cells. Consistently, RNA-seq data obtained from the Ivy GAP database further revealed that, in the tissues of patients with GBM, CYLD expression was negatively correlated with the expression of Ki-67, a proliferation marker in GBM tissues (Fig. 1B). Next, a Transwell migration assay was performed to evaluate GBM cell migration, an epithelial-mesenchymal transition (EMT)-like change. CYLD knockdown significantly enhanced GBM cell migration (Fig. 1C), whereas CYLD overexpression markedly suppressed this effect (Fig. 1D). The Ivy GAP database further revealed a significant negative correlation between the expression levels of CYLD and fibronectin, a mesenchymal marker, in the infiltrating area of GBM tissues (Fig. 1E), indicating that CYLD downregulation enhanced the EMT-like changes and infiltration in GBM. These results indicated the promotion of cell proliferation and migration in CYLD-silenced GBM cells.

As cancer stem cell-like cells critically contribute to malignancy in patients with GBM (41), the effects of CYLD knockdown on stem cell-like characteristics were determined in GBM cells. The sphere-forming ability, a characteristic of cancer stem cell-like cells, was evaluated. Notably, CYLD knockdown significantly increased the sphere-forming ability of GBM cells (Fig. 2A), whereas CYLD overexpression markedly suppressed this effect (Fig. 2B).

In the limiting dilution assay, CYLD-knockdown GBM cells were able to form spheres with lower cell numbers than the control GBM cells (Fig. 2C), whereas CYLD-overexpressed GBM cells required more cells (Fig. 2D). Moreover, RNA-seq data obtained from the Ivy GAP database further revealed that CYLD expression was negatively correlated with the expression of nestin, a cancer stem cell marker, in GBM tissues (Fig. 2E), suggesting that GBM cells acquire stem-like characteristics via CYLD knockdown.

To identify novel therapeutic target signals in CYLD-knockdown cells, a proteomic analysis was performed to comprehensively identify the intracellular signaling pathways involved in the malignant characteristics caused by CYLD downregulation. In total 2,914 phosphorylated proteins were identified, among which, 337 proteins were phosphorylated more than double due to CYLD downregulation (Fig. 3A and B). Kinase Enrichment Analysis 2 was used to analyze the 337 proteins. It revealed 44 combinations of proteins and phosphorylation sites, and the identified proteins were likely to interact with 20 kinases (Table I and Fig. 4A). Among the 20 kinases, 9 kinases exhibited significant differences, and the majority of them (CKIε, CK2, MAPK9, CAMK2A, and CKIα) were involved in Wnt/β-catenin signaling (Fig. 4B). RNA-seq data from the Ivy GAP database further confirmed that CYLD expression was negatively correlated with the expression of CK2β and CKIε, Wnt/β-catenin signaling activators, in GBM tissues (Fig. 4C and D). These results indicated that Wnt/β-catenin signaling plays key roles in the malignancy of cells via CYLD knockdown.

Finally, the therapeutic effects of inhibiting Wnt/β-catenin signaling on the malignant characteristics of CYLD-knockdown cells were determined. As shown in Fig. 5A and B, the promotion of cell proliferation and migration by CYLD silencing was significantly suppressed by treatment with ICG-001, a Wnt/β-catenin signaling inhibitor. Furthermore, in the limiting dilution assay, ICG-001 treatment significantly inhibited the sphere-forming ability of CYLD-knockdown GBM cells (Fig. 5C), suggesting that the inhibition of Wnt/β-catenin signaling may suppress the stem cell-like characteristics caused by CYLD knockdown. Taken together, the results of the present study suggest targeting Wnt/β-catenin signaling as a potential therapeutic strategy for CYLD-negative patients with GBM.