Gene expression profile data of glioma (GBM) GSE50161 (11) and GSE108474 (12) were downloaded from the Gene Expression Omnibus (GEO) database (13) using the R package GeoQuery and were divided into the tumor and normal tissue groups based on gene expression. The R package TCGAbiolinks (14) was used to obtain the gene expression profile and clinical data of patients with glioma from TCGA database. Detailed data on patients with GBM and the control groups are shown in Table 1 .

The R package limma (15) was used to analyze differentially expressed genes (DEGs) in GBM tumor tissues relative to normal and adjacent tissues. In the three datasets, genes with adj p-value < 0.05 were deemed as DEGs, genes with log FC > 1 and adj p-value < 0.01 were considered upregulated genes, and genes with log FC < -1 and adj p-value < 0.01 were considered downregulated genes. We screened the intersection of DEGs of the three datasets to obtain the common DEGs.

In this study, two functional enrichment analyses were used. The Gene Ontology (GO) function annotation analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of the common DEGs were performed using the R software package clusterProfiler.

GSEA (16) was performed using the GSEA package in R. Analysis was performed using default parameters. Functional and pathway enrichment analyses were performed on the DEGs in the high and low TRAF3IP3 expression groups.

Using the Human Protein Reference Database (HPRD) to obtain the PPI network related to DEGs, we extracted the interaction subnets directly connected to at least 25 DEGs and visualized the obtained PPI network model with Cytoscape (17), using cytoHubba (18) to extract key subnets. Closely connected local regions in the single gene-related PPI network may represent molecular complexes and have specific biological functions, and ClueGO (19) was used for functional annotation.

Patients were divided into high- and low-expression groups according to the TRAF3IP3 expression value in their tumor tissue. The difference between the TRAF3IP3 high-expression and low-expression groups was analyzed, and the prognosis of patients was evaluated at the gene expression level.

Total cellular RNA was isolated from HMC3, U87, HS683, U251, and U118 cells using the TRIZOL reagent (15596-018, Invitrogen, USA), dissolved in RNase-free H₂O, and stored at -80°C. cDNA synthesis was performed using M-MLV Reverse Transcriptase (AT341, TransGen, China) following the manufacturer’s instructions. The PCR system was used to amplify all transcripts using the AceQTM Universal SYBR Qpcr Master Mix (Q511-03, vazyme). A reaction mixture (total volume, 20 µL) was prepared using cDNA (8 μL), SYBR^® Green Master Mix (10 μL), upstream and downstream primers (0.5 μL), and ddH₂O (1 μL). The qPCR was performed under the following conditions: pre-denaturation at 95°C for 5 min, denaturation at 95°C for 10 s, and annealing at 60°C for 30 s. Steps 2–3 were repeated for 40 cycles, and thereafter at 95°C for 15 s, at 60°C for 60 s, at 95°C for 15 s, and at 4°C. Target gene mRNA expression was verified and analyzed using the Illumina eco real-time PCR system using GAPDH as an internal standard. Quantification was performed by normalizing the expression levels of the target genes to the GAPDH expression levels; the expression levels of the genes were calculated using the 2^-ΔΔCt method. All experiments were performed in triplicate.

For western blot analysis, cell lysates were prepared from cell lines or tissues with a RIPA lysis buffer kit (P0013B, Beyotime, China), and the protein concentrations were quantified using a Beyotime protein assay (P0010, Beyotime, China). Whole-cell proteins were separated using 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene difluoride membranes (Millipore, USA). The blots were blocked with 5% skim milk powder in TBST for 1 h at room temperature and incubated overnight with the following primary antibodies: TRAF3IP3 (ER65081, Huaan Bio, China, 1:1000), ERK (#4695, cst, 1:1000), p-ERK (#4370, cst, 1:2000), and GAPDH (#5174, 1:20000). Following three washes with TBST buffer, the membranes were incubated with secondary goat anti-rabbit (#7074, cst, 1:5000) or goat anti-mouse antibodies (#7076, cst, 1:5000) for 1 h at room temperature, followed by three washes with TBST buffer. Signals were detected using an enhanced chemiluminescence kit (Thermo, USA) and exposed to a film (4741023953, Guangxi Juxing Medical Instrument Co., Ltd). The protein density was normalized to that of GAPDH.

Human glioma cell lines U251, U87, U118, Hs683, and human microglial cell line HMC3 were obtained from Shanghai Fuheng Biotechnology Co., Ltd. Cells were cultured in DMEM (Gibco) supplemented with 10% fetal bovine serum (Gibco). HEB cells were cultured in microglial medium (Gibco). All cell lines were cultured at 37°C and under 5% CO₂ conditions.

A lentivirus system encoding shRNA targeting a scrambled sequence and TRAF3IP3 mRNA was used for the knockdown. Sequences targeting TRAF3IP3-sh1 5′-CCATCAAGAAGCCACCCAA-3′ and TRAF3IP3-sh2 5′-CCACGTGCTTCAGTCCAAA-3′ were cloned into the PGIPZ-puro. The coding sequence of human TRAF3IP3 was synthesized and subcloned into a PCDH vector to construct overexpression plasmids. TRAF3IP3-F: AGCGAATTCGCCACCatgatcagcccagacccca; TRAF3IP3-R: GGATCCGATTTAAATtcagatcatcaggttgtctt. The integrity of the plasmid constructs was confirmed by DNA sequencing. Together with the packaged plasmids, the constructed plasmids were transfected into U118 or U251 cells. After 48 h, the supernatant containing the target lentivirus was collected. Lentivirus infection was performed on cells at 80% confluency with a multiplicity of infection (MOI) of 50. Cells were used for downstream assays or transplantation 72 h after infection.

The Cell Counting Kit-8 (CCK-8, BBI Life Sciences, E606335-0500) was used to assess the cell proliferation rate. For the cell proliferation assay, U251 cells were infected with lentivirus containing TRAF3IP3 shRNA or scramble shRNA and seeded into 24-well plates. U118 cells were infected with lentivirus containing the PCDH vector or an empty vector. After 96 h, U251 or U118 cells were digested with 0.25% trypsin (Gibco, 25200-056). Cell suspensions were seeded at a density of 3×10³ cells/well in 96-well culture plates with three identical wells as duplicate wells, followed by overnight incubation in a 5% CO₂ humidified incubator at 37°C. After removing the medium, 10 μL of CCK-8 solution was added, and cells were incubated for 1 h at 37°C under 5% CO₂ conditions. Absorbance was measured at 450 nm using an ELISA plate reader (EPOCH2, BioTek, USA). The OD450 value is inversely proportional to the degree of cell proliferation. Each group comprised three duplicate wells, and the assays were independently performed three times.

Cells (1 ×10⁵/600 μL/well) were seeded in 12-well plates and cultured overnight at 37°C to form a confluent monolayer. Then, an artificial wound was made in the monolayer with a 1 mL pipette tip and washed three times with PBS. Cell activity was recorded every 24 h using an inverted microscope. The test was completed after 48 h. The wound was analyzed using the ImageJ software (version 1.80, National Institute of Health, Bethesda, MD, USA).

Transwell invasion assays were performed using Transwell chambers with Matrigel (354480, CORNING). Briefly, 600 μL of 10% FBS-containing medium was placed in the lower chamber, and 2×10⁴ cells suspended in 100 μL serum-free medium were seeded into the upper chamber for 24 h. After 24 h, cells migrated through the Matrigel were fixed with 4% paraformaldehyde (A500684, BBI Life Sciences, China), stained with 0.1% crystal violet (E607309, BBI Life Sciences, China), and counted under a microscope.

Twelve female BALB/c nude mice at 6 weeks of age, weighing 18–22 g, were obtained from Shanghai Jihui Experimental Animal Feeding Co., LTD (Shanghai, China) and bred in a special pathogen-free (SPF) grade laboratory at Tongji University. The mice were housed under a 12:12 h light:dark cycle environment, with ad libitum access to food and water. All experiments were approved by the Animal Care and Use Committee of Tongji University. A total of 5×10⁴ U118-GFP-expressing cells or U118 TRAF3IP3-GFP-expressing cells were injected in situ into the right caudate nucleus of BALB/c nude mice (six mice per group). Images were captured 1 week after injection and 2 weeks later using a Caliper Life Sciences camera (Ivis Lumina Xr).

Paraffin-embedded sections of tumor tissues from nude mice were de-paraffinized using different concentrations of ethanol (100, 95, 85, and 75%) and rehydrated with deionized water. These slices were then immersed in EDTA solution (pH 9.0, G1203, Servicebio) and heated, maintaining the temperature between 95°C and 100°C for 30 min. After washing with PBS, the sections were incubated with a 3% hydrogen peroxide solution for 25 min. The sections were then incubated with primary antibodies targeting Ki67 (1:1000, GB111141, Servicebio) and PCNA (1:500, 10205-2-AP, Proteintech) overnight at 4°C and then stained using HRP-labeled goat anti-rabbit IgG or anti-mouse IgG (Servicebio GB23303).

All data are expressed as the means ± standard deviations (SDs). Comparisons between two groups were analyzed using the Student’s t-test. Comparisons among multiple groups were made using one-way analysis of variance (ANOVA), followed by Dunnett’s test. Statistical significance was set at p < 0.05, p < 0.01, or p < 0.001.

Using the limma package, we analyzed the DEGs in gliomas in the three datasets and obtained a set of genes whose expressions were significantly upregulated or downregulated in glioma tissues ( Figures S1A–C ). In addition, we performed a cluster analysis of the expression values of DEGs using the three datasets. The results indicated that the DEGs in each group could distinguish tumor tissues from normal tissues ( Figures S1D–F ).

Subsequently, we screened the intersection of DEGs in the three datasets ( Figure 1A ), obtaining 5383 DEGs shared by the three datasets. To verify the impact of DEGs on clinical diagnosis, we used DEGs and data groupings to plot a classification heat map ( Figures 1B–D ) and determined that DEGs can distinguish diseased samples from normal specimens.

To identify potential genes that could serve as biomarkers of glioma prognosis, we first performed a GO analysis of the DEGs. DEGs were mainly enriched in the modulation of chemical synaptic transmission and regulation of transsynaptic signaling, synaptic membrane, presynaptic, tubulin binding, cell adhesion molecule binding, among other biological processes, cell components, and molecular functions ( Figures 1E–H , Table 2 ). Next, KEGG analysis of the DEGs was performed. The results showed that DEGs were enriched in the oxytocin signaling pathway, protein processing in the endoplasmic reticulum, human papillomavirus infection, cell cycle, and other biological pathways ( Figures 1I, J, L ). In addition, the GO analysis results showed that DEGs were significantly enriched in epilepsy syndrome, retinoblastoma, retinal cell carcinoma, hereditary breast cancer, and other diseases ( Figure 1K ).

To identify the impact of genes that could serve as glioma prognosis predictors at various stages of glioma development, we downloaded the PPI information from the HPRD database, removing the isolation data nodes, loop interactions, repeated interactions, etc., and extracted the PPI network related to DEGs in the network ( Figure 2A ). To determine PPI networks closely related to DEGs in gliomas, we extracted at least 25 DEGs that interacted with non-DEGs as subnets ( Figure 2B ). We visualized the network with Cytoscape using the cytoHubba plug-in roots to extract the interaction subnet of the top 15 genes. Genes in the PPI subnet were significantly enriched in cancer pathways, MAPK signaling pathways, hepatitis B, pancreatic cancer, colorectal cancer, and other pathways ( Figure 2C ).

By analyzing the glioma-related DEGs and PPI subnets, we obtained five key genes: TRAF3IP3, WIPI1, ARHGAP11A, ABCB8, and RHOQ. To accurately analyze the mechanism of each key gene, we extracted the PPI subnet of each key gene from the STRING database. We used Cytoscape to visualize the critical gene subnets and used the ClueGO plug-in for functional enrichment analysis ( Figures 2D–H ).

To determine the value of the key genes, we analyzed their differential expression in normal and tumor tissues ( Figures 3A–E ) and classified the expression levels ( Figure 3F ). The results showed that TRAF3IP3, WIPI1, ARHGAP11A, and RHOQ were significantly differentially expressed in normal and tumor tissue, displaying good classification performance. As a result, TRAF3IP3 may serve as an important indicator for glioma diagnosis. We analyzed the impact of key genes on the overall and disease-free survival time ( Figures 3G–K ). The results showed that TRAF3IP3 and ABCB8 expression significantly impacted the OS of patients, whereas WIPI1 expression had a significant impact on the disease-free survival of patients. Therefore, TRAF3IP3 may be an important prognostic factor in gliomas.

To further analyze the influence of TRAF3IP3 expression in glioma, we identified DEGs between the TRAF3IP3 high- and TRAF3IP3 low-expression groups ( Figures S2A, B ). We performed functional annotations of the DEGs between the two groups ( Table 3 ). The results showed that the DEGs between the TRAF3IP3 high- and low-expression groups were significantly enriched in autophagy, autophagy mechanisms, proteasome protein catabolism, proteasome-mediated ubiquitin-dependent protein catabolism, mitochondrial matrix, cell-substrate connection, interneuron synapse, cell adhesion molecule binding, protein serine/threonine kinase activity, and ubiquitin-like protein ligase binding as GO terms ( Figures 4A–D ). In addition, annotations in the DEGs between the TRAF3IP3 high-and low-expression groups were significantly enriched in herpes simplex virus type 1 infection, Salmonella infection, ubiquitin-mediated proteolysis, T cell leukemia virus type 1 infection, and neurodegeneration changes in a variety of diseases, small cell lung cancer, cell cycle, and other biological pathways ( Figures 4G, H ). Salmonella infection is a major public health concern. Infection in humans can be chronic and increase the risk of cancers (20). Ubiquitin-mediated proteolysis is reportedly associated with pancreatic cancer metastasis (21). Human T cell leukemia virus type 1 (HTLV-1) mainly infects CD4⁺ T cells and induces chronic, persistent infection in infected individuals, with some developing adult T cell leukemia/lymphoma (ATL) (22). Glioblastoma is one of the most aggressive and frequent primary brain tumors. This type of glioma expands and infiltrates into the brain, causing neuronal degeneration and neurological decay (23). A previous study has shown that high TUBA1C expression is associated with poor prognosis in patients with glioma. TUBA1C inhibition reduced glioma cell proliferation through cell cycle arrest (24). Thus, TRAF3IP3 may be associated with cancer progression-related pathways. Furthermore, we analyzed the disease gene network (DGN) and the cancer gene network (CGN), wherein the DEGs between the TRAF3IP3 high- and low-expression groups were significantly enriched ( Figures 4E, F ). These DEGs were significantly enriched in glioblastoma, invasive breast cancer, bladder cancer, and other tumors. TRAF3IP3 is involved in marginal zone B lymphocyte development and survival. TRAF3IP3 contributes to MZ B cell survival by upregulating autophagy, thereby promoting the TI-II immune response (25). TRAF3IP3 was associated with the OS of patients with T1-4N0-2M0 in TCGA dataset. TRAF3IP3 can be used as a prognostic signature for OS in patients with lung adenocarcinoma (26). Gastric cancer is one of the leading causes of cancer-associated mortality. Recent studies show that TRAF3IP3 positively correlated with pathological tumor and lymph node stages (27). Therefore, TRAF3IP3 may be involved in glioma progression.

To further identify the functions related to the DEGs in the TRAF3IP3 high-and low-expression groups, we performed enrichment analysis on different GSEA datasets and found that these genes were significantly enriched in neuroactive ligand-receptor interactions, olfactory transmission, proteasome, cytokine–cytokine receptor interaction, and calcium signaling pathway in KEGG biological pathways ( Figure 5A ). They were also significantly enriched in the detection of stimuli in sensory perception, the adhesion of homophilic cells through plasma membrane adhesion molecules, mitochondrial membrane tissue keratinization, sensory perception, and other biological processes ( Figure 5B ); cation channel complex, cytoplasmic ribosomal large subunit, endopeptidase complex, extracellular matrix, mitochondrial ribosomal large subunits, and other cellular components ( Figure 5C ); and in cation channel activity, cytokine activity, gated channel activity, G protein-coupled receptor activity, metal ion transmembrane transport protein activity, and other molecular functions ( Figure 5D ). A biological information analysis suggested that neuroactive ligand-receptor interactions may be associated with glioma (28), which is consistent with our results. Despite all other cells that have the potential to prevent cancer development and metastasis through tumor-suppressor proteins, cancer cells can promote the activity of the ubiquitin-proteasome system (UPS) to degrade tumor-suppressor proteins and avoid apoptosis. As less invasive chemotherapy drugs, UPS inhibitors are increasingly used to alleviate the symptoms of various malignant cancers. Ubiquitin-proteasome is associated with cancer development (29). A previous study demonstrated that the DEGs in glioma were significantly enriched in the cytokine–cytokine receptor interaction, which is consistent with our study (30). We also carried out enrichment analysis on the cancer characteristic gene set and immune characteristic-related gene set ( Figures 5E, F ). The cancer characteristic gene set included PTEN, IL21, KRAS, etc. Loss of the tumor-suppressor molecule PTEN is among the most common molecular dysfunctions associated with glioma malignancy (31). Besides, interleukin-21 (IL21) increases the reactivity of allogeneic human Vγ9Vδ2 T cells against primary glioblastoma tumors. IL-21 increased intracellular granzyme B levels and cytotoxicity of allogeneic human Vγ9Vδ2 T lymphocytes in vitro (32). Kirsten rat sarcoma viral oncogene (KRAS) expression exhibited a positive correlation with the ERK pathway. KRAS overexpression promoted glioma proliferation and invasion (33). The immune characteristic-related gene set included IL2, NOD2, IL21, etc. Interleukin-2 and histamine jointly inhibit tumor growth and angiogenesis in malignant glioma (34). A study showed a significant correlation between nucleotide oligomerization domain 2 (NOD2) variants and a higher risk of glioblastoma (35). Therefore, TRAF3IP3 may be associated with glioma progression-related pathways.

To investigate the role of TRAF3IP3 in glioma development, TRAF3IP3 expression in glioma cell lines was detected using qRT-PCR and western blot ( Figure 6A ). We showed that TRAF3IP3 expression was upregulated in glioma cell lines compared to human glial cell line HEB. These results indicate that TRAF3IP3 may be involved in glioma development.

To explore the biological function of TRAF3IP3 in tumorigenesis, qRT-PCR and western blot were used to verify TRAF3IP3 expression after its overexpression in U118 cells ( Figure 6B ). Similarly, we detected TRAF3IP3 expression in glioma cells U251 after TRAF3IP3 knockdown. The western blot showed partial knockdown (50%) of TRAF3IP3 ( Figure 6C ).

To explore the potential role of TRAF3IP3 in tumorigenesis, we evaluated its effects on the growth of glioma cells in vitro. CCK-8 assays revealed that TRAF3IP3 overexpression significantly increased the proliferation of U118 cells compared to the NC and control group ( Figure 6D ). However, partial knockdown of TRAF3IP3 in U251 glioma cells significantly suppressed U251 cell proliferation compared to controls ( Figure 6E ). Collectively, these data suggest that TRAF3IP3 is involved in the proliferative ability of gliomas. In addition, we used the EdU assay to verify that TRAF3IP3 overexpression in U118 glioma cells significantly promoted cell proliferation ( Figure 6F ). In contrast, TRAF3IP3 knockdown in U251 glioma cells significantly reduced cell proliferation ( Figure 6G ).

To evaluate the effect of TRAF3IP3 on glioma cell migration, we performed a wound healing assay. The wound healing assay results showed that ectopic overexpression of TRAF3IP3 enhanced the migration abilities of U118 cells ( Figure 6H ). Conversely, TRAF3IP3 knockdown in U251 cells had the opposite effect ( Figure 6I ). Taken together, these results indicated that TRAF3IP3 promotes glioma cell migration.

Transwell assays were performed to validate the effect of TRAF3IP3 overexpression and silencing on the migration and invasion abilities of glioma U118 and U251 cells. As shown in Figure 6J , ectopic expression of TRAF3IP3 significantly enhanced the invasion ability of glioma cells, whereas silencing of TRAF3IP3 expression decreased the number of invading glioma cells ( Figure 6K ). These results revealed that TRAF3IP3 promotes glioma cell invasion.

To further verify the role of TRAF3IP3 in glioma cell growth, cells expressing ectopic TRAF3IP3 and control cells were injected in situ into the caudate nucleus of nude mice to generate a xenotransplantation model. Fluorescence imaging experiments in nude mice showed no statistically significant difference in tumor growth in mice 7 days after the injection of TRAF3IP3-overexpressing cells. However, tumor growth significantly differed in mice 21 days after injection of TRAF3IP3-overexpressing glioma cells ( Figure 6L ). Moreover, we detected that the expression of the proliferation-related biomarkers, Ki67 and PCNA, was increased in TRAF3IP3-overexpressing tumors ( Figure 6M ). Taken together, these results indicated that TRAF3IP3 plays a critical role in glioma cell growth.

Bioinformatics analysis showed that DEGs were enriched in the MAPK signaling pathway in GBM. To explore the downstream signaling pathway of TRAF3IP3, we detected the changes in downstream signaling pathways in U118 and U251 cells after TRAF3IP3 overexpression or knockdown. TRAF3IP3 knockdown in U251 cells decreased ERK phosphorylation, whereas TRAF3IP3 overexpression increased ERK phosphorylation ( Figure 7A ). TRAF3IP3 overexpression in U118 cells increased ERK activation ( Figure 7B ). Thus, we speculated that the ERK signaling pathway might be associated with the oncogenic function of TRAF3IP3 in glioma cells.

To confirm the association between TRAF3IP3 and the ERK pathway and to identify whether the ERK pathway is involved in TRAF3IP3-mediated glioma proliferation, migration, and invasion in vitro, U118 cells transfected with the PCDH/TRAF3IP3 vector were treated with or without the ERK pathway inhibitor U0126. We found that U0126 treatment significantly inhibited cell proliferation, migration, and invasion ( Figures 7C–F ) in PCDH/TRAF3IP3-transfected glioma cells. Quantitative western blot analysis revealed that the ERK pathway was inhibited when TRAF3IP3 was knocked down alone, whereas TRAF3IP3 rescue by overexpression resulted in ERK pathway activation ( Figure 7G ). In TRAF3IP3-overexpressing cells, the addition of U0126 led to the reversal of ERK pathway activation. This change was also found to be statistically significant when compared to the U0126 control group. These results indicated that TRAF3IP3 functions as an oncogene, possibly via ERK pathway activation in glioma cells.