Prognostic value of palmitoylation-regulated mechanisms in glioblastoma: integrated multi-omics analysis via least absolute shrinkage and selection operator (LASSO) regression and single-cell sequencing

Introduction

Glioblastoma multiforme (GBM) exhibits intratumoral and intertumoral heterogeneity, leading to the emergence of drug-resistant subclones that evade interventions. Its diffuse invasiveness hinders complete surgical resection, accelerating recurrence (1-3). Meanwhile, the blood-brain barrier severely hinders the delivery of chemotherapeutic drugs to the tumor microenvironment, thereby diminishing drug bioavailability at the lesion site (4,5). These interconnected challenges collectively culminate in profoundly poor clinical outcomes for GBM patients, highlighting the urgency of developing innovative strategies to overcome current therapeutic limitations.

Recent studies have revealed that post-translational modifications play an indispensable role in cancer initiation and progression (6). Among them, palmitoylation, a dynamically reversible lipid modification, has gained widespread attention for its profound biological significance (7). This modification precisely regulates protein subcellular localization, conformational changes, and molecular interactions by adding palmitic acid chains to cysteine residues, thereby influencing cellular signal transduction (8-10). At the molecular level, the dynamic balance of palmitoylation is maintained by two enzyme families: the zDHHC family of palmitoyltransferases catalyzes the attachment of palmitic acid chains, while the APT family of depalmitoylases mediates their removal (11). Disruption of this balance is strongly linked to the malignant progression of various tumors. For example, in GBM, abnormal palmitoylation of the oncoprotein EGFR can inhibit p53, triggering aberrant activation of the G1/S transition and consequent dysregulation of cell proliferation (12). Aberrant palmitoylation or dysregulation of palmitoylation-related molecules may further augment the migratory and invasive potentials of glioma cells (13-15). However, there are significant gaps in the research on palmitoylation-related genes in GBM. Therefore, exploring the functional diversity of these genes in GBM will lay a crucial foundation for advancing precise diagnosis and targeted therapy of GBM.

In this study, we identified palmitoylation-related differentially expressed genes (DEGs) with prognostic value, explored their biological functions in GBM, and established correlations with previously reported palmitoylation-related genes in GBM.

GABRB2, the gene encoding for GABA_A receptors β2 subunit, regulates synaptic function by participating in GABA-mediated inhibitory neurotransmission (16). Neutrophil cytosolic factor 2 (NCF2) is a key component of the nicotinamide adenine dinucleotide phosphate (NADPH) oxidase complex (17). Studies have demonstrated that NCF2 is upregulated in clear cell renal cell carcinoma and hepatocellular carcinoma, with a correlation to poor prognosis (18,19). GRIN2A, which encodes the NMDA receptors in cerebral neurons, is an ionotropic glutamate receptor subunit, and has been identified as a frequently mutated gene in tumor. Studies have confirmed that GRIN2A mutations enhance the efficacy of immune checkpoint inhibitors in various cancers (20). Further exploration of the specific roles and regulatory mechanisms of GABRB2, NCF2, and GRIN2A in GBM pathogenesis holds important clinical value for the development of novel therapeutic targets. We present this article in accordance with the TRIPOD and MDAR reporting checklists (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1953/rc).

Methods

Data acquisition

Transcriptomic profiles and clinical metadata of GBM patients were acquired from The Cancer Genome Atlas (TCGA) (https://portal.gdc.cancer.gov). RNA sequencing datasets were retrieved from the Gene Expression Omnibus (GEO; http://www.ncbi.nlm.nih.gov/geo/) repository. Normal brain tissue mRNA transcriptomic data, used as controls, were obtained from Genotype-Tissue Expression (GTEx) (https://gtexportal.org/home/). Immunochemistry was analyzed through the Human Protein Atlas (HPA; https://www.proteinatlas.org/). Palmitoylation-associated genes were sourced from the GeneCards database (https://www.genecards.org/), with detailed entries provided in https://cdn.amegroups.cn/static/public/tcr-2025-1953-1.xls. A total of 343 palmitoylation-related genes were ultimately included in the subsequent analyses. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Identification of palmitoylation-related DEGs

DEGs between tumor tissues and adjacent normal tissues were identified using the “limma” package in RStudio, with significance thresholds set at P<0.05 and |log₂fold change (FC)| ≥2. Volcano plots were generated using the “ggrepel” package for visualization. Palmitoylation-related DEGs were determined by intersecting 3,256 palmitoylation-associated genes with the 1,592 identified DEGs.

Survival analysis

To assess the prognostic significance of these 343 genes, both ROC curve analyses and survival analyses were performed. Kaplan-Meier (K-M) survival curves were generated using the R packages “survival” and “survminer” to compare overall survival (OS) between high- and low-risk cohorts, with differences measured via the log-rank test (P<0.05). Univariate Cox regression analysis was performed by incorporating z-score-normalized expression data of differentially expressed palmitoylation-related genes along with OS duration and status. This analysis quantifies the independent prognostic impact of individual genes on survival outcomes, as reflected by hazard ratios (HRs) and 95% confidence intervals (CIs), while excluding genes without prognostic relevance. Ultimately, palmitoylation-related DEGs with a P value <0.05 were identified as potential prognostic biomarkers, serving as candidates for subsequent validation and functional experiments.

Functional enrichment analysis and protein-protein interaction (PPI) network construction

The R package “pheatmap” served to illustrate the variability present in the dataset. Functional characterization of the 343 palmitoylation-associated DEGs was conducted via Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses, with the findings visualized using the R package “ggplot”. Construction of the PPI network relied on the search tool for the retrieval of interacting genes (STRING) database (https://string-db.org/). The target genes were uploaded to STRING to generate the corresponding PPI network.

Single-cell RNA sequencing (scRNA-seq) data processing

Raw expression matrices were processed using the Seurat package. Initially, quality control was performed to retain genes expressed in no less than 5 cells and cells with a minimum of 40 detected genes. Low-quality cells were further excluded through secondary filtering, specifically those with mitochondrial gene content exceeding 20% or fewer than 300 detected genes. After quality control, the expression matrix was normalized using the LogNormalize method with a scale factor of 10,000. The top 2,500 highly variable genes, identified by the FindVariableFeatures function, were selected for subsequent dimensionality reduction and clustering. Cell clustering was carried out using a K-nearest neighbor graph with 25 principal components and a resolution of 0.6, and the clusters were visualized via uniform manifold approximation and projection (UMAP).

Differential expression analysis was implemented using Seurat’s FindAllMarkers function, with thresholds set as log₂FC >1 and P value <0.05. Cell types were annotated by comparison with human annotations.

Trajectory inference was performed using monocle3 (v1.3+) to construct cell trajectories, integrating dimensionality reduction and pseudotime analysis in the UMAP space. For target genes, cells were divided into high- and low-expression groups based on the median expression level. The distributions and densities of these groups in the trajectory space were visualized, and the pseudotime distribution characteristics across clusters and cell types were subjected to statistical analysis.

Cell culture and transfection

Human glioma cell lines LN229 and U251 were obtained from the Cell Resource Center of the Chinese Academy of Medical Sciences. Cells were cultured in DMEM medium (Hyclone, Logan, UT, USA) supplemented with 10% fetal bovine serum (FBS; Hyclone) and 1% penicillin/streptomycin (Gibco, Waltham, MA, USA) under a humidified atmosphere of 5% CO₂ at 37 ℃.

At 70% confluence, cells were transfected with small-interfering RNA targeting NCF2 (si-NCF2) or its negative control (si-NC), overexpression plasmids for GABRB2 (oe-GABRB2) or GRIN2A (oe-GRIN2A), and their respective negative control (oe-NC) using Lipofectamine 3000 (Thermo Fisher Scientific, Waltham, MA, USA). Oligonucleotides were all purchased from GenePharma (Shanghai, China). The si-NCF2, its negative control, as well as the oe-GABRB2, oe-GRIN2A, and control overexpression plasmids were all purchased from GenePharma (Shanghai, China). Specifically, the oe-GABRB2, oe-GRIN2A, and control overexpression plasmids were constructed by the company using the pcDNA3.1 (+) vector (NTCC, China). Transfection efficiency was verified by reverse transcription-quantitative polymerase chain reaction (RT-qPCR). Transfected cells were harvested for subsequent experiments.

RT-qPCR analyses

Total RNA was isolated from cellular samples with Trizol reagent (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) following the manufacturer’s recommended protocol, and reverse transcription was conducted immediately afterward. The sequences of gene-specific primers were as follows: GABRB2 (forward 5'-AGGCAAGGCACTATGATCGG-3', reverse 5'-TGCCTTCGATTTCTCTGCGT-3'), NCF2 (forward 5'-TGGCGATCTCAGCAAAAGGTGG-3', reverse 5'-GTACTGTCCCACCTCCATCTTG-3'), GRIN2A (forward 5'-CGACCCCGGCAGCTTTGGAA-3', reverse 5'-GCGAGTGGGTCCGATTCTCTGC-3'), and GAPDH (forward 5'-GCGACACCCACTCCTCCAC-3', reverse 5'-ACCACCCTGTTGCTGTAGCC-3'). Relative expression levels of VAMP5 were quantified via the 2^−ΔΔCt approach, with GAPDH serving as the endogenous control. All reactions were run in triplicate.

Western blotting

To extract proteins from cultured cells, RadioImmunoPrecipitation Assay (RIPA) lysis buffer (Beyotime, Haimen, China) supplemented with protease and phosphatase inhibitors was used. Protein concentrations were quantified using a BCA protein assay kit (Thermo Scientific, Waltham, MA, USA). Equal amounts of protein (30–50 µg) were subjected to 10–12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) separation and subsequently transferred onto polyvinylidene fluoride (PVDF) membranes (Roche Applied Sciences, Indianapolis, IN, USA). Following blocking with 5% non-fat milk in tris-buffered saline with Tween 20 (TBST) for 1 hour at room temperature, membranes were incubated overnight at 4 ℃ with the following primary antibodies: anti-GABRB2 (1:1,000, Affinity Biosciences, Changzhou, China; DF6671), anti-NCF2 (1:1,000, Proteintech, Rosemont, IL, USA; 15551-1-AP), anti-GRIN2A (1:1,000, Proteintech, 28525-1-AP), and anti-GAPDH (1:1,000, Proteintech, 10491-1-AP). After washing, membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies for 1 hour at room temperature. Blotted protein bands were visualized by incubating the membranes with chemiluminescence HRP substrate (Millipore, Burlington, MA, USA) and imaged using a Bio-Rad Gel Doc EZ imager (Bio-Rad Laboratories, Inc., Hercules, CA, USA). All experiments were performed at least three times with biological replicates.

Colony formation assay

Digested cells were seeded at 1,500 cells/well in 6-well plates, shaken crosswise for uniform distribution, and cultured at 37  °C with 5% CO₂ for 2 weeks after confirming no clumps microscopically. The medium was discarded, cells were washed 3 times with PBS, fixed in paraformaldehyde for 30 min, and stained with crystal violet for 30 min. Unbound dye was washed off with PBS. Plates were air-dried, photographed, and colonies were counted and analyzed.

Transwell migration assay

Transfected cells were seeded into the upper chamber of Transwell inserts. The lower chamber was filled with 700 µL of DMEM containing 20% FBS. After 24 hours of incubation at 37 ℃, non-migrated cells in the upper chamber were removed with a moist cotton swab. Migrated cells adhering to the lower surface were fixed with 4% paraformaldehyde for 15 min, then stained with 0.1% crystal violet for 20 min and visualized under a microscope. Each experiment was independently repeated three times.

Wound healing assay

Transfected cells were seeded at 1.8×10⁶ cells/well in 6-well plates and cultured at 37 ℃ with 5% CO₂ until 90–100% confluent. In a biosafety cabinet, uniform scratches were created across the cell monolayer using a sterile 200 µL pipette tip. Detached cells and debris were removed by rinsing 3 times, and serum-free DMEM was added for further incubation. Images of the scratched areas were captured at the same marked fields via inverted microscope 0 and 24 h post-scratching. Migration rate (%) = [(initial scratch width at 0 h − width at 24 h) / initial scratch width at 0 h] × 100%. Each experiment was independently repeated three times.

Statistical analysis

All statistical analyses were performed using R v4.2.2 (https://www.r-project.org/) or GraphPad Prism (version 10.4.1). A P value <0.05 was considered statistically significant. For continuous variables that exhibited a normal distribution and homogeneous variance, an independent samples t-test was used for analysis; otherwise, the Wilcoxon rank-sum test was used. Data are presented as mean ± standard deviation from at least 3 independent experiments.

Results

Identification of 343 palmitoylation-related DEGs

We created a flowchart based on our research (Figure 1A). Differential gene analysis of TCGA and GTEx databases identified 1592 DEGs in gliomas (P<0.05, |log₂ FC| ≥2, Figure 1B). A total of 3256 palmitoylation-related genes were obtained from the GeneCards database. Finally, intersection analysis of the 1,592 DEGs and 3,256 palmitoylation-related genes via a Venn diagram identified 343 candidate palmitoylation-related DEGs (Figure 1C).

Figure 1 Identification of palmitoylation-related DEGs (n=343). (A) Schematic workflow of the study. (B) Volcano plot comparing differential expression between GBM tumor tissues from TCGA and normal tissues from GTEx. |log₂FC| >1 and adjusted P<0.05 were set as thresholds for DEGs. (C) Venn diagram identifying palmitoylation-related DEGs. DEGs, differentially expressed genes; FC, fold change; GBM, glioblastoma multiforme; GTEx, Genotype-Tissue Expression; PPI, protein-protein interaction; TCGA, The Cancer Genome Atlas.

Functional enrichment analysis of DEGs and construction of PPI network

We used a heatmap to visualize the expression of the 343 candidate palmitoylation-related DEGs (Figure 2A). Correlation analysis of these genes revealed significant associations among DDR1, ANXA5, and MSN (Figure 2B). GO term and KEGG pathway analyses were performed (Figure 2C,2D). The primary biological processes (BPs) such as regulation of monoatomic ion transport, wound healing, and regulation of membrane potential. The cellular components (CCs) are primarily concentrated in synaptic membrane, endocytic vesicles, and transmembrane transporter complex. The most prevalent molecular functions (MFs) include monoatomic ion channel activity, metal ion transmembrane transport activity, and gated channel activity. KEGG pathway enrichment results showed that these genes are associated with neuroactive ligand-receptor interaction, neuroactive ligand signaling, and calcium signaling pathways. A PPI network was constructed using Cytoscape, with the top 80 hub genes were identified. The STRING database was employed to construct the PPI network for these 80 candidate genes (Figure 2E).

Figure 2 Functional enrichment analysis and PPI network of the 343 DEGs. (A) Heatmap of gene expression levels in tumor vs. normal tissues. (B) Correlation analysis among the 343 DEGs. (C) GO enrichment items. (D) KEGG pathway enrichment. (E) The PPI network of the top 80 core genes constructed based on STRING, highlighting hub genes that play pivotal roles in regulating the malignant progression of GBM through PPI. BP, biological process; CC, cellular component; DEGs, differentially expressed genes; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function; PPI, protein-protein interaction; STRING, search tool for the retrieval of interacting genes.

Prognostic significance and predictive efficacy of palmitoylation-related genes

To investigate the impact of palmitoylation-related genes on GBM prognosis, we identified 21 genes associated with patients’ OS using univariate Cox regression analysis (P<0.05, Figure 3A). Through least absolute shrinkage and selection operator (LASSO) regression and stepwise Cox regression analyses, excellent performance with the fewest variables was achieved when Logλ=−3.29, and 11 OS-related palmitoylation-related genes, including EEF1A1, ITGB1, GABRB2, GRIN2A, FCER1G, MYC, NCF2, ITGB3, MPO, CXCL10, and WNT1, were selected for model development (Figure 3B). Patients were divided into low-risk and high-risk groups based on the median risk score for prognostic analysis. Scatter plots of survival status and distribution of risk scores demonstrated that the survival time of glioma patients decreases with increasing risk scores, with high-risk patients having shorter survival times than low-risk patients (Figure 3C,3D). A heatmap of the expression of the 12 palmitoylation-related genes in the model, combined with clinicopathological variables, showed significant differences in gene expression between the high-risk and low-risk groups (Figure 3E). Receiver operating characteristic (ROC) curves indicated that this predictive model has strong predictive power for survival outcomes (Figure 3F).

Figure 3 Construction of a LASSO regression-based risk model. (A,B) Displays the coefficient of the selected feature using the lambda parameter. The horizontal axis represents the lambda value, and the vertical axis represents the coefficient of the independent variable. (C) Risk score, survival time and survival status scatter plot: shows the distribution of risk scores from low to high, with different colors representing different groups. The scatter plot distribution shows the corresponding relationship between risk scores, survival time and survival status of different samples. (D) Diagnostic efficacy according to the ROC curve for the risk model in the dataset. (E) Heatmap of expression levels for feature genes, reflecting the differential expression of prognostic genes in GBM and their potential as therapeutic targets for inhibiting tumor progression. (F) K-M curves combined with log-rank tests were utilized to assess OS disparities among palmitoylation metabolism-related clusters. Additionally, time-dependent ROC curve analyses were performed to calculate AUC values at key follow-up time points (1, 3, and 5 years), which systematically quantifies the dynamic survival discriminative ability and prognostic performance of the clustering model. AUC, area under the curve; CI, confidence interval; CV, cross-validation; GBM, glioblastoma multiforme; LASSO, least absolute shrinkage and selection operator; HR, hazard ratio; K-M, Kaplan-Meier; OS, overall survival; ROC, receiver operating characteristic.

Single-cell sequencing analysis of GABRB2, NCF2, and GRIN2A characteristics in GBM

Analysis of GSE273274 identified 25 cell clusters and 8 medium cell types (Figure 4A). Gene set variation analysis (GSVA) heatmaps showed that GABRB2, NCF2 and GRIN2A are significantly enriched in pathways related to MYC targets, KRAS signaling, and DNA repair (Figure 4B). Cellular trajectory and pseudo-time distribution of GBM cells revealed two cellular states during the development of oligodendrocyte precursor cells (OPCs) (Figure 4C-4I).

Figure 4 ScRNA-seq analysis in GBM. (A) Single cells from three GBM tumor samples and three adjacent normal samples were classified by cell type in the scRNA-seq database. (B) GSVA heatmap shows specific signature pathways. (C) Distribution of cells in each cluster in pseudo-time trajectories. (D-I) Pseudo-time trajectories of single-cell transcriptome data colored by cluster, with cells distributed along the main branches, elucidating the developmental lineage of GBM cells and key transition stages that promote tumor progression. *, P<0.05; **, P<0.01; ***, P<0.001; ****, P<0.0001. GBM, glioblastoma multiforme; GSVA, gene set variation analysis; scRNA-seq, single-cell RNA sequencing; TPM, transcript per million; UMAP, uniform manifold approximation and projection.

Association of GABRB2, NCF2 and GRIN2A with GBM progression

Analysis via the GEPIA database showed significant differences in the expression of GABRB2, NCF2, and GRIN2A in GBM (Figure 5A-5C). Correlation analysis between these three candidate genes and previously reported palmitoylation-related genes in GBM showed negative correlations between GABRB2, GRIN2A and zDHHC15, and a positive correlation between NCF2 and zDHHC15 (Figure 5D-5F). Further analysis using the HPA database revealed that the expression levels of GABRB2 and GRIN2A in GBM were considerably lower than those in normal samples, while the expression level of NCF2 in GBM was higher than that in normal samples (Figure 5G-5I). To explore the roles of these three candidate genes in GBM cells, we transfected oeGABRB2, oeGRIN2A and siNCF2 into two GBM cell lines. The results showed that overexpression of GABRB2 and GRIN2A reduced GBM proliferation and migration abilities, while inhibition of NCF2 reduced GBM proliferation and migration abilities (Figures 6,7).

Figure 5 Expression of GABRB2, NCF2, and GRIN2A in GBM. (A-C) Differential expression of the three candidate genes between GBM and normal tissues. (D-F) Correlation analyses between the three candidate genes and palmitoylation-related genes. (G-I) IHC staining for the protein expression levels of GABRB2, NCF2, and GRIN2A in the human cerebral cortex and glioma derived from HPA. The original data are available from HPA v25 (https://v25.proteinatlas.org/) with the following specific URLs: GABRB2 (https://www.proteinatlas.org/ENSG00000145864-GABRB2/brain/cerebral+cortex; https://www.proteinatlas.org/ENSG00000145864-GABRB2/cancer/glioma), NCF2 (https://www.proteinatlas.org/ENSG00000116701-NCF2/brain/cerebral+cortex; https://www.proteinatlas.org/ENSG00000116701-NCF2/cancer/glioma), GRIN2A (https://www.proteinatlas.org/ENSG00000183454-GRIN2A/brain/cerebral+cortex; https://www.proteinatlas.org/ENSG00000183454-GRIN2A/cancer/glioma). Scale bar, 100 μm. *, P<0.05. GBM, glioblastoma multiforme; HPA, Human Protein Atlas; IHC, immunohistochemical; TPM, transcript per million.

Figure 6 GABRB2, NCF2, and GRIN2A regulate the proliferation and migration of glioblastoma cells via colony formation assay and transwell migration assay. (A-F) The colony formation assay was performed in LN229 and U251 cells when GABRB2 or GRIN2A was overexpressed, or NCF2 was knocked down (n=3). Overexpression of GABRB2, GRIN2A decreases colony formation ability in LN229 and U251 cells. Knockdown of NCF2 decreases colony formation ability in LN229 and U251 cells. (G-L) Transwell assay was performed to assess the migration ability of LN229 and U251 cells when GABRB2 or GRIN2A was overexpressed, or NCF2 was knocked down (n=3). Overexpression of GABRB2, GRIN2A decreases cell migration in LN229 and U251 at 24 h. Knockdown of NCF2 reduces cell migration in LN229 and U251. Stained with crystal violet. Scale bar, 100 μm. **, P<0.01; ***, P<0.001; ****, P<0.0001. oe-NC, overexpression plasmid negative control; si-NC, small interfering RNA negative control.

Figure 7 GABRB2, NCF2 and GRIN2A regulate glioblastoma cell migration via wound healing assay. (A-L) Cell migration potential of LN229 and U251 cells was assessed via wound healing assay (n=3). At 24 h, overexpression of GABRB2 or GRIN2A was found to decrease the migration rate of LN229 and U251 cells, whereas knockdown of NCF2 consistently attenuated the migratory capacity of both cell lines, indicating their functional involvement in GBM cell migration. Scale bar, 100 μm. ***, P<0.001; ****, P<0.0001. GBM, glioblastoma multiforme; oe-NC, overexpression plasmid negative control; si-NC, small interfering RNA negative control.

Discussion

GBM presents a huge challenge in clinical treatment due to its high recurrence rate. At present, the prognosis of GBM patients has not improved significantly (16). Recent research progress has revealed that post-translational modifications play a pivotal role in regulating multiple carcinogenic pathways (6). As one of them, palmitoylation has been widely studied in various solid tumors such as lung cancer and colorectal cancer (17,18). Its mechanisms involve the regulation of oncoprotein functions, metabolic reprogramming, and remodeling of the immune microenvironment (18-20). However, studies on palmitoylation-related genes in GBM remain relatively limited. This study aims to clarify the impact of palmitoylation-related genes on GBM prognosis and provides novel therapeutic strategies for GBM treatment.

GABRB2, NCF2, and GRIN2A are all key molecules involved in neural signal transmission and immune regulation. Existing studies have shown that abnormal expression of GABRB2 is closely related to neurological diseases such as epilepsy (21); NCF2 modulates inflammatory responses by regulating NADPH oxidase activity (22); and GRIN2A mutations may lead to neurodevelopmental disorders (23). However, the specific molecular mechanisms of these three genes in GBM remain unclear. This study is the first to systematically analyze the effects of GABRB2, NCF2, and GRIN2A on GBM progression. We screened a panel of palmitoylation-related characteristic genes, including these three genes, and constructed a reliable risk model for predicting the prognosis of GBM patients. Through in vitro experiments, tumor cells were transfected with siRNA targeting NCF2 (si-NCF2), plasmids overexpressing GABRB2 (oe-GABRB2) and plasmids overexpressing GRIN2A (oe-GRIN2A). We demonstrated that NCF2 exerts a tumor-promoting effect in GBM, while GABRB2 and GRIN2A function as suppressors of tumor progression in GBM.

Previous studies have confirmed that the arrangement of COL1A1 is closely associated with the invasive spread of GBM, and invasive aligned structures and collagen tissue trajectories can directly regulate GBM progression (24). Given that extracellular matrix (ECM)-guided collective migration is a core feature of GBM invasiveness, it is speculated that the regulation of GBM migration by GABRB2, NCF2, and GRIN2A is likely mediated through the key pathway of interacting with ECM or remodeling tissue trajectories, providing a clear target for mechanistic exploration.

Regarding other candidate genes, including EEF1A1, ITGB1, FCER1G, MYC, ITGB3, MPO, CXCL10, and WNT1, existing studies have confirmed that they have potential roles in GBM and affect the prognosis of GBM patients. EEF1A1 promotes glioma cell proliferation and invasion through oxidative stress (25); ITGB1 affects the progression of GBM by enhancing the stability of tumor HIF1α protein (26); FCER1G may regulate myeloid cells to affect the immune microenvironment by activating the SYK kinase pathway (27); MYC is a key regulatory gene in glioma. Its expression level is affected by m6A RNA modification. Overactivation of MYC can promote the proliferation of tumor cells (28); ITGB1 transcriptional activation can activate its target proteins AKT and focal adhesion kinase (FAK), thereby promoting the growth of glioma cells (29); CXCL10 promotes the enhancement of anti-tumor immune response by activating T cells, thereby inhibiting glioma (30); as an important member of the Wnt signaling pathway; Wnt1 promotes the proliferation of tumor cells by activating the Wnt signaling pathway (31); MPO mainly promotes anti-tumor responses by regulating immune responses and altering the tumor microenvironment (32).

The advancement of single-cell sequencing technology has brought major breakthroughs to tumor research (33). Single-cell sequencing can perform high-throughput analysis of individual cells to reveal the heterogeneity and subpopulation structure within tumors (34,35). Our research results show that at the single-cell level, GABRB2, NCF2, and GRIN2A exhibit significantly heterogeneous expression patterns in GBM.

In conclusion, our research results highlight the importance of palmitoylation in GBM, provide a framework for elucidating the role of palmitoylation modification in tumor progression, and support targeting palmitoylation-related processes to improve patient prognosis.

This study has several limitations. First, univariate Cox regression linked the target gene to survival, while the independent prognostic significance of the target gene remains to be further validated in large-scale prospective cohorts. Second, the role of GABRB2, NCF2, and GRIN2A in regulating extracellular matrix and collagen fiber alignment is unassessed. Third, direct validation of these genes’ regulation by palmitoylation is lacking, and with no available clinical sample data. Furthermore, GABRB2, NCF2, and GRIN2A’s involvement in extracellular matrix and collagen fiber regulation is unassessed, which is essential to confirm cell migration modulation. Most available external cohorts lack complete survival information (follow-up time, event status) required for model validation, an objective data limitation that precludes us from performing the requested external validation to date. Additionally, most accessible external cohorts (including multiple GEO series) lack complete survival data, a constraint precluding external validation for now. Future studies need to further investigate the functional roles of these genes, palmitoylation regulation, and biomarker potential.

Conclusions

This study systematically reveals the prognostic potential of palmitoylation-related genes in the development and progression of GBM, providing new directions for future precision therapeutic strategies based on protein post-translational modifications.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD and MDAR reporting checklists. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1953/rc

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1953/prf

Funding: This research was supported by the “Taihu Talent Plan” Healthcare High-end Talent (No. THRC0009).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1953/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Zhou S, Huang Y, Chen Y, et al. Reprogramming systemic and local immune function to empower immunotherapy against glioblastoma. Nat Commun 2023;14:435. [Crossref] [PubMed]
2. Rong L, Li N, Zhang Z. Emerging therapies for glioblastoma: current state and future directions. J Exp Clin Cancer Res 2022;41:142. [Crossref] [PubMed]
3. Obrador E, Moreno-Murciano P, Oriol-Caballo M, et al. Glioblastoma Therapy: Past, Present and Future. Int J Mol Sci 2024;25:2529. [Crossref] [PubMed]
4. Gan HK, van den Bent M, Lassman AB, et al. Antibody-drug conjugates in glioblastoma therapy: the right drugs to the right cells. Nat Rev Clin Oncol 2017;14:695-707. [Crossref] [PubMed]
5. Tylawsky DE, Kiguchi H, Vaynshteyn J, et al. P-selectin-targeted nanocarriers induce active crossing of the blood-brain barrier via caveolin-1-dependent transcytosis. Nat Mater 2023;22:391-9. [Crossref] [PubMed]
6. Pan S, Chen R. Pathological implication of protein post-translational modifications in cancer. Mol Aspects Med 2022;86:101097. [Crossref] [PubMed]
7. S Mesquita F. Mechanisms and functions of protein S-acylation. Nat Rev Mol Cell Biol 2024;25:488-509. [Crossref] [PubMed]
8. Solis GP, Kazemzadeh A, Abrami L, et al. Local and substrate-specific S-palmitoylation determines subcellular localization of Gαo. Nat Commun 2022;13:2072. [Crossref] [PubMed]
9. Reilly L, Howie J, Wypijewski K, et al. Palmitoylation of the Na/Ca exchanger cytoplasmic loop controls its inactivation and internalization during stress signaling. FASEB J 2015;29:4532-43. [Crossref] [PubMed]
10. Jeyifous O, Lin EI, Chen X, et al. Palmitoylation regulates glutamate receptor distributions in postsynaptic densities through control of PSD95 conformation and orientation. Proc Natl Acad Sci U S A 2016;113:E8482-91. [Crossref] [PubMed]
11. Kwon H, Choi M, Ahn Y, et al. Flotillin-1 palmitoylation turnover by APT-1 and ZDHHC-19 promotes cervical cancer progression by suppressing IGF-1 receptor desensitization and proteostasis. Cancer Gene Ther 2023;30:302-12. [Crossref] [PubMed]
12. Fan X, Sun S, Yang H, et al. SETD2 Palmitoylation Mediated by ZDHHC16 in Epidermal Growth Factor Receptor-Mutated Glioblastoma Promotes Ionizing Radiation-Induced DNA Damage. Int J Radiat Oncol Biol Phys 2022;113:648-60. [Crossref] [PubMed]
13. Wang Y, Shen N, Yang Y, et al. ZDHHC5-mediated S-palmitoylation of FAK promotes its membrane localization and epithelial-mesenchymal transition in glioma. Cell Commun Signal 2024;22:46. [Crossref] [PubMed]
14. Tang F, Liu Z, Chen X, et al. Current knowledge of protein palmitoylation in gliomas. Mol Biol Rep 2022;49:10949-59. [Crossref] [PubMed]
15. Lu F, Shen SH, Wu S, et al. Hypomethylation-induced prognostic marker zinc finger DHHC-type palmitoyltransferase 12 contributes to glioblastoma progression. Ann Transl Med 2022;10:334. [Crossref] [PubMed]
16. Louis DN, Perry A, Reifenberger G, et al. The 2016 World Health Organization Classification of Tumors of the Central Nervous System: a summary. Acta Neuropathol 2016;131:803-20. [Crossref] [PubMed]
17. Zhang Y, Li F, Fu K, et al. Potential Role of S-Palmitoylation in Cancer Stem Cells of Lung Adenocarcinoma. Front Cell Dev Biol 2021;9:734897. [Crossref] [PubMed]
18. Zhang Q, Yang X, Wu J, et al. Reprogramming of palmitic acid induced by dephosphorylation of ACOX1 promotes β-catenin palmitoylation to drive colorectal cancer progression. Cell Discov 2023;9:26. [Crossref] [PubMed]
19. Li M, Zhang L, Chen CW. Diverse Roles of Protein Palmitoylation in Cancer Progression, Immunity, Stemness, and Beyond. Cells 2023;12:2209. [Crossref] [PubMed]
20. Chen Y, Li Y, Wu L. Protein S-palmitoylation modification: implications in tumor and tumor immune microenvironment. Front Immunol 2024;15:1337478. [Crossref] [PubMed]
21. Mohammadi NA, Ahring PK, Yu Liao VW, et al. Distinct neurodevelopmental and epileptic phenotypes associated with gain- and loss-of-function GABRB2 variants. EBioMedicine 2024;106:105236. [Crossref] [PubMed]
22. Tan W, Hildebrandt MA, Pu X, et al. Role of inflammatory related gene expression in clear cell renal cell carcinoma development and clinical outcomes. J Urol 2011;186:2071-7. [Crossref] [PubMed]
23. Strehlow V, Heyne HO, Vlaskamp DRM, et al. GRIN2A-related disorders: genotype and functional consequence predict phenotype. Brain 2019;142:80-92. [Crossref] [PubMed]
24. Faisal SM, Clewner JE, Stack B, et al. Spatiotemporal Insights into Glioma Oncostream Dynamics: Unraveling Formation, Stability, and Disassembly Pathways. Adv Sci (Weinh) 2024;11:e2309796. [Crossref] [PubMed]
25. Anguraj Vadivel AK, Pajovic S, Siddaway R, et al. The proteomic landscape of diffuse midline glioma highlights the therapeutic potential of non-histone protein methyltransferases. Neuro Oncol 2025;27:1829-46. [Crossref] [PubMed]
26. Zhu C, Chen X, Liu TQ, et al. Hexosaminidase B-driven cancer cell-macrophage co-dependency promotes glycolysis addiction and tumorigenesis in glioblastoma. Nat Commun 2024;15:8506. [Crossref] [PubMed]
27. Xu H, Zhu Q, Tang L, et al. Prognostic and predictive value of FCER1G in glioma outcomes and response to immunotherapy. Cancer Cell Int 2021;21:103. [Crossref] [PubMed]
28. Su R, Dong L, Li C, et al. R-2HG Exhibits Anti-tumor Activity by Targeting FTO/m(6)A/MYC/CEBPA Signaling. Cell 2018;172:90-105.e23. [Crossref] [PubMed]
29. Wang J, Cai C, Nie D, et al. FRK suppresses human glioma growth by inhibiting ITGB1/FAK signaling. Biochem Biophys Res Commun 2019;517:588-95. [Crossref] [PubMed]
30. Mao J, Li J, Chen J, et al. CXCL10 and Nrf2-upregulated mesenchymal stem cells reinvigorate T lymphocytes for combating glioblastoma. J Immunother Cancer 2023;11:e007481. [Crossref] [PubMed]
31. Lan X, Gui Z, Chen T, et al. Genistin Represses the Proliferation and Angiogenesis While Accelerating the Apoptosis of Glioma Cells by Modulating the FOXC1-Mediated Wnt Signaling Pathway. Discov Med 2024;36:332-42. [Crossref] [PubMed]
32. Jalali Motlagh N, Wang C, Kuellenberg EG, et al. D-Mannose Slows Glioma Growth by Modulating Myeloperoxidase Activity. Cancers (Basel) 2021;13:6360. [Crossref] [PubMed]
33. Zhang Y, Wang D, Peng M, et al. Single-cell RNA sequencing in cancer research. J Exp Clin Cancer Res 2021;40:81. [Crossref] [PubMed]
34. Wu H, Guo C, Wang C, et al. Single-cell RNA sequencing reveals tumor heterogeneity, microenvironment, and drug-resistance mechanisms of recurrent glioblastoma. Cancer Sci 2023;114:2609-21. [Crossref] [PubMed]
35. Xu Z, Xi B, Huang J, et al. Integration of Single-Cell RNA and Bulk RNA Sequencing Reveals Cellular Heterogeneity and Identifies Survival-Associated Regulatory Networks in Glioblastoma. IET Syst Biol 2025;19:e70025. [Crossref] [PubMed]