This study utilized a retrospective cohort design for bioinformatic analysis. Publicly available RNA-sequencing data and corresponding clinical information for 1,027 glioma patients were downloaded from The Cancer Genome Atlas (TCGA, https://portal.gdc.cancer.gov/) and the Chinese Glioma Genome Atlas (CGGA, http://www.cgga.org.cn/) databases (Zhao et al., 2021; Brennan et al., 2013). Key variables extracted included FTO expression levels, overall survival (OS), WHO grade, histological type, IDH mutation status, and 1p/19q co-deletion status. The study protocol for using public data was exempt from institutional review board approval.

Additionally, a total of 25 fresh-frozen tissue samples, comprising 8 low-grade gliomas (2 Grade I, 6 Grade II), 14 high-grade gliomas (6 Grade III, 8 Grade IV), and 3 normal brain tissues (obtained during surgical decompression for trauma), were collected from the Department of Neurosurgery of Xiangyang No. 1 People’s Hospital, Hubei University of Medicine, between October 2021 and June 2023. All participants provided written informed consent. The study was approved by the Ethics Committee on Scientific Research of Xiangyang No. 1 People’s Hospital (Approval No. 2021KYLX05) and conducted in accordance with the Declaration of Helsinki.

Human glioma cell lines U251 (Cat# CL-0237) and U87MG (Cat# CL-0238), and the human microglial cell line HMC3 (Cat# CL-0620), were purchased from Procell Life Science & Technology Co., Ltd. (Wuhan, China). Cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM, Cat# PM150210; Procell) supplemented with 10% Fetal Bovine Serum (FBS, Cat# SA211.01; CellMax) and 1% Penicillin-Streptomycin solution (Procell). All cells were maintained in a humidified incubator at 37 °C with 5% CO₂. Mycoplasma contamination was routinely tested using a PCR-based kit. The PI3K inhibitor LY294002 (Selleck Chemicals, United States of America), FTO inhibitor FB23-2 (Huang et al., 2019), and transcription inhibitor Actinomycin D (Sigma-Aldrich, United States) were used as indicated.

Patients from the TCGA and CGGA cohorts were stratified into high- and low-FTO expression groups based on the median expression value. Differentially expressed genes (DEGs) were identified using the DESeq2 R package with criteria of |log2(Fold Change)| > 1 and adjusted p-value <0.05. The identified DEGs were uploaded to the Database for Annotation, Visualization, and Integrated Discovery (DAVID, v6.8, https://david.ncifcrf.gov/) for Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses (Sherman et al., 2022). The top enriched terms (p < 0.05) are presented (Supplementary Tables S1–S4).

Lentiviral vectors for FTO overexpression (oeFTO) were constructed by cloning the human full-length FTO cDNA sequence into the GV492 vector. A corresponding empty vector served as a negative control (oeNC). For FTO knockdown, two short hairpin RNA (shRNA) sequences targeting FTO (shFTO-1, shFTO-2) and a non-targeting scramble control (shNC) were cloned into the pLKD vector. All lentiviral constructs were synthesized and packaged by GenePharma (Shanghai, China). Target sequences are listed in Supplementary Table S5.

U251 and U87MG cells were seeded in 6-well plates and transduced with lentiviruses at approximately 50% confluency in the presence of polybrene (8 μg/mL). After 48 h, the medium was replaced with fresh medium containing puromycin (4 μg/mL; Thermo Fisher Scientific) for 2 weeks to select for stably transduced cells. Transfection efficiency was confirmed by qRT-PCR and Western blot.

Total RNA was extracted from cells using TRIzol reagent (Cat# 10606ES60; YEASEN). The global m6A level in 200 ng of total RNA was measured using the EpiQuik m6A RNA Methylation Quantification Kit (Colorimetric) (Cat# P-9005-48; EpiGentek) according to the manufacturer’s protocol (Shi et al., 2021). Absorbance was read at 450 nm, and the relative m6A level was calculated based on a standard curve.

For the 5-ethynyl-2′-deoxyuridine (EdU) assay, cells were seeded in 24-well plates (5 × 10⁴ cells/well). After 24 h, cell proliferation was assessed using a BeyoClick™ EdU Cell Proliferation Kit with Alexa Fluor 488 (Cat# C0071S; Beyotime) following the manufacturer’s instructions. EdU-positive (green) and DAPI-stained (blue) nuclei were imaged using a fluorescence microscope, and the percentage of EdU-positive cells was quantified from at least five random fields.

For the colony formation assay, 300 cells per well were seeded into 6-well plates and cultured for 10–14 days until visible colonies formed. The colonies were fixed with 4% paraformaldehyde and stained with 0.1% crystal violet solution. Colonies containing more than 50 cells were counted.

Cells were harvested, washed with ice-cold PBS, and fixed in 70% ethanol at −20 °C overnight. After fixation, cells were washed again and resuspended in PBS containing propidium iodide (PI, 50 μg/mL) and RNase A (100 μg/mL) from a Cell Cycle Staining Kit (Cat# BL114A; Biosharp). After incubation for 30 min in the dark, the DNA content was analyzed using a flow cytometer (Beckman Coulter). The percentages of cells in G1, S, and G2/M phases were quantified using ModFit LT software.

Total RNA was isolated using TRIzol, and 1 µg was reverse-transcribed into cDNA using the RevertAid First Strand cDNA Synthesis Kit (Takara, Japan). qRT-PCR was performed on a Bio-Rad CFX96 Real-Time PCR System using SYBR Green Master Mix (Takara, Japan). The relative expression of target genes was calculated using the 2^−ΔΔCT method, with GAPDH serving as the endogenous control. Primer sequences are provided in Supplementary Table S5.

Total protein was extracted using RIPA lysis buffer (Cat# BL651A; Biosharp) supplemented with protease and phosphatase inhibitor cocktails. Protein concentrations were determined using a BCA assay kit. Equal amounts of protein (20–30 µg) were separated by SDS-PAGE and transferred to PVDF membranes. The membranes were blocked with 5% non-fat milk and incubated overnight at 4 °C with the following primary antibodies: FTO (1:1000, Cat# R24361; ZENBIO), EREG (1:1000, Cat# 12048; Cell Signaling Technology), PI3K (1:1000, Cat# 60225-1-Ig; Proteintech), p-PI3K (1:1000, Cat# ab38449; Abcam), Akt (1:2000, Cat# 60203-2-Ig; Proteintech), p-Akt (Ser473) (1:2000, Cat# 80455-1-RR; Proteintech), p53 (1:1000, Cat# abs130605; Absin), p21 (1:1000, Cat# YT3497; Immunoway), and GAPDH (1:10000, Cat# ANT324s; AntGene). After incubation with appropriate HRP-conjugated secondary antibodies (1:5000; Proteintech), protein bands were visualized using an enhanced chemiluminescence (ECL) kit (Beyotime, China) and quantified using ImageJ software.

Paraffin-embedded tissue sections (5 μm) from clinical samples and subcutaneous xenograft tumors were deparaffinized and rehydrated. Antigen retrieval was performed by heating in citrate buffer (pH 6.0). The sections were then incubated with primary antibodies against FTO, EREG, or p53 overnight at 4 °C, followed by incubation with an HRP-conjugated secondary antibody. Staining was developed with DAB chromogen and counterstained with hematoxylin. Staining intensity was semi-quantitatively scored based on the proportion of positive cells and staining intensity, calculated as the integrated optical density (IOD) per area using Image-Pro Plus software.

All animal experiments were conducted in strict accordance with the ARRIVE guidelines and were approved by the Animal Use and Care Committee of Xiangyang No. 1 People’s Hospital (Approval No. 2021KYLX05). Male BALB/c nude mice (4–6 weeks old) were purchased from Hunan Silaikejingda Experimental Animal Co., Ltd. (License No. SCXK (Hunan) 2019-0004) and housed under specific pathogen-free conditions with a 12-h light/dark cycle and ad libitum access to food and water.

After 1 week of acclimatization, a subcutaneous xenograft model was established. A total of 28 mice were randomly allocated into four groups (n = 7 per group) using a simple randomization method. Prior to injection, mice were briefly anesthetized using inhaled isoflurane (2%–3%). Subsequently, U251 cells stably expressing oeNC, oeFTO, shNC, or shFTO-1 (5 × 10⁶ cells in 100 µL PBS) were injected subcutaneously into the left flank of each mouse. Tumor growth was monitored every 3 days by measuring the length (L) and width (W) with calipers. Tumor volume was calculated using the formula: Volume = (L × W²)/2. After 28 days, mice were euthanized by CO₂ inhalation using a gradual displacement rate of 20% of the chamber volume per minute, followed by cervical dislocation to ensure death, and the tumors were excised, weighed, and processed for IHC analysis. No animals were excluded from the analysis.

All experiments were performed in triplicate unless otherwise specified. Data are presented as mean ± standard deviation (SD). Statistical analyses were performed using GraphPad Prism 8.0 (GraphPad Software, United States) and R software (v4.1.2). Comparisons between two groups were analyzed using an unpaired two-tailed Student’s t-test. For comparisons among three or more groups, one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test was used. Kaplan-Meier survival curves were generated, and the log-rank test was used to compare survival differences. A p-value <0.05 was considered statistically significant. Significance levels are denoted as *p < 0.05, **p < 0.01, and ***p < 0.001.

To investigate the clinical relevance of FTO in glioma, we analyzed transcriptomic and clinical data from the CGGA and TCGA cohorts. Heatmaps of patient characteristics stratified by FTO expression revealed that low FTO levels were associated with adverse features, including higher WHO grade, wild-type IDH status, and non-codel 1p/19q status (Figures 1A,B). Quantitative analysis confirmed that FTO expression was significantly lower in high-grade gliomas (WHO grades III-IV) compared to low-grade gliomas (WHO grades I-II) (Figures 1C,G). FTO levels were also markedly reduced in the most aggressive histological subtype, glioblastoma (GBM), and in patients with wild-type IDH and non-codel 1p/19q status (Figures 1D–F,H–J).

Kaplan-Meier survival analysis demonstrated that patients with low FTO expression had significantly shorter overall survival (OS) in both the CGGA and TCGA cohorts (Figures 2A,B). To validate these bioinformatic findings, we examined FTO expression in our own clinical samples. IHC and qRT-PCR analyses showed a progressive decrease in FTO protein and mRNA levels with increasing glioma grade, consistent with the database results (Figures 2C,E,F). Furthermore, FTO expression was significantly lower in U251 and U87MG glioma cell lines compared to the HMC3 normal microglial cell line (Figures 2D,G). Collectively, these data strongly indicate that FTO is downregulated in malignant glioma and that its low expression is a marker of poor prognosis, suggesting a tumor-suppressive role.

To explore the molecular functions regulated by FTO, we performed GO and KEGG pathway analyses on the DEGs between high- and low-FTO expression groups. In both the CGGA and TCGA datasets, GO analysis revealed that genes co-expressed with FTO were significantly enriched in biological processes related to cell proliferation, apoptosis, and immune response (Figures 3A–H). Notably, KEGG pathway analysis consistently highlighted the PI3K/Akt signaling pathway (hsa04151) as one of the most significantly enriched pathways associated with FTO expression (Figures 3I,J). These bioinformatic predictions suggest that FTO may exert its tumor-suppressive effects by modulating cell proliferation via the PI3K/Akt pathway.

To directly test the tumor-suppressive function of FTO, we generated U251 and U87MG cell lines stably overexpressing FTO (oeFTO). Successful overexpression was confirmed at both mRNA and protein levels (Figures 4A,B). As expected, FTO overexpression significantly decreased the global m6A RNA levels, confirming its demethylase activity (Figure 4C). Functional assays showed that the oeFTO group had a significantly lower proportion of EdU-positive cells compared to the control (oeNC) group, indicating inhibited proliferation (Figure 4D). Similarly, FTO overexpression markedly reduced the number and size of colonies in the colony formation assay (Figure 4G). Flow cytometry analysis revealed that FTO overexpression caused a significant accumulation of cells in the G1 phase, with a corresponding decrease in the S phase population, indicating G1 phase arrest (Figures 4E,F). These results demonstrate that FTO suppresses the malignant growth of glioma cells in vitro.

Conversely, to mimic the low FTO expression observed in high-grade glioma, we knocked down FTO using two independent shRNAs (shFTO-1, shFTO-2). Both shRNAs effectively reduced FTO mRNA and protein levels (Figures 5A,B) and led to a significant increase in global m6A levels (Figure 5C). FTO knockdown robustly enhanced cell proliferation, as shown by an increased percentage of EdU-positive cells (Figure 5D) and a greater number of colonies formed (Figure 5F). Cell cycle analysis showed that FTO knockdown resulted in a decreased G1 phase population and an increased S phase population, indicating accelerated cell cycle progression (Figure 5E). These findings corroborate our overexpression results and confirm that loss of FTO promotes the proliferative capacity of glioma cells.

To validate the tumor-suppressive role of FTO in vivo, we established a subcutaneous xenograft model using U251 cells with stable FTO overexpression (oeFTO) or knockdown (shFTO-1). The results were striking: tumors in the oeFTO group grew significantly slower and were substantially smaller and lighter at the end of the experiment compared to the oeNC group (Figures 6A–C). In contrast, tumors in the shFTO group grew much faster and were significantly larger and heavier than those in the shNC group. IHC analysis of the excised tumors revealed that oeFTO tumors had lower EREG expression and higher p53 expression, whereas shFTO tumors showed the opposite trend (Figure 6D). These in vivo data provide compelling evidence that FTO acts as a potent suppressor of glioma tumor growth.

To identify the downstream effector of FTO, we searched for DEGs associated with cell proliferation and found that Epiregulin (EREG), a member of the epidermal growth factor family, was a top candidate consistently regulated by FTO in both public datasets (Figure 7A). We confirmed this at the molecular level: FTO overexpression decreased EREG mRNA and protein levels, while FTO knockdown increased them (Figures 7B,C,G,H). To determine if this regulation was dependent on FTO’s demethylase activity, we treated oeFTO cells with the FTO inhibitor FB23-2. The inhibitor reversed the FTO-induced decrease in global m6A levels and restored EREG protein expression to control levels (Figures 7D,E). This suggests FTO regulates EREG in a demethylase-dependent manner.

Next, we investigated whether FTO affects EREG mRNA stability. Cells were treated with the transcription inhibitor Actinomycin D, and EREG mRNA levels were measured over time. The half-life of EREG mRNA was significantly shorter in FTO-overexpressing cells and markedly longer in FTO-knockdown cells compared to their respective controls (Figure 7F). This indicates that FTO promotes the degradation of EREG mRNA. Taken together, these results suggest that FTO suppresses EREG expression by removing its m6A modification, thereby reducing its mRNA stability.

Having established the FTO-EREG link and recalling our KEGG analysis, we hypothesized that FTO exerts its function through the PI3K/Akt pathway, a known downstream target of EREG (Huang et al., 2019). Western blot analysis revealed that FTO knockdown significantly increased the phosphorylation of PI3K and Akt, whereas FTO overexpression had the opposite effect (Figures 7G,H). Furthermore, the expression of key downstream effectors of the PI3K/Akt pathway, the tumor suppressor p53 and the cell cycle inhibitor p21, was decreased upon FTO knockdown and increased upon FTO overexpression (Figures 7G,H).

To confirm the causal role of this pathway, we treated FTO-knockdown cells with the PI3K inhibitor LY294002 (LY) (Figure 8A). Treatment with LY successfully reversed the accelerated cell proliferation and G1-S transition induced by FTO knockdown, as demonstrated by colony formation and cell cycle analyses (Figures 8B–E). These rescue experiments confirm that FTO suppresses glioma cell proliferation and cell cycle progression by inhibiting the PI3K/Akt signaling axis.