Introduction

Gliomas constitute the most prevalent and aggressive primary brain tumors, representing approximately 30% of all central nervous system (CNS) malignancies and up to 80% of all malignant brain tumors1,2. Glioblastoma represents one of the most aggressive and fatal brain tumors, with patients typically surviving only 15–18 months, and fewer than 3% maintaining progression-free survival (PFS) beyond five years3,4,5. Although recent progress in imaging modalities, neurosurgical techniques, radiotherapy, and the development of novel chemotherapeutic agents has led to noticeable improvements in the management of gliomas6, malignant forms such as GBM remain highly invasive, and an effective curative treatment is still lacking7. The limited therapeutic success is largely due to the intricate and strongly immunosuppressive tumor microenvironment (TME) characteristic of glioblastoma, which includes glioma cells, glioma stem-like cells (GSCs), diverse immune cell populations, neural components, brain vasculature, and extracellular matrix (ECM) elements8,9. Consequently, there is an urgent need to identify novel molecular biomarkers and therapeutic targets to enhance diagnostic accuracy, prognostic assessments, and treatment efficacy in glioma patients.

Engrailed-1 (EN1) is a homeobox-containing transcription factor traditionally recognized for its pivotal role in embryonic neural development, including the patterning of the midbrain and the regulation of dopaminergic neuron differentiation10,11. Beyond its developmental functions, emerging evidence has implicated EN1 in the pathogenesis of various cancers. Studies have demonstrated that EN1 is aberrantly expressed in several tumor types, where it contributes to tumor progression, metastasis, and resistance to apoptosis12,13. For instance, EN1 overexpression has been associated with poor prognosis in breast cancer and esophageal squamous cell carcinoma, suggesting its potential as an oncogenic driver14,15. However, the specific roles and underlying mechanisms of EN1 in glioma, particularly in relation to the tumor immune microenvironment and chemoresistance, remain largely unexplored.

This study aims to comprehensively explore the role of EN1 in glioma progression and immune evasion. We integrate transcriptomic analyses from bulk RNA-seq, single-cell RNA-seq, and ceRNA network predictions to investigate EN1’s expression patterns, its relationship with immune cell infiltration, and its potential regulatory mechanisms. Additionally, we perform in vitro validation of EN1’s functional roles in glioblastoma cells, focusing on its effects on proliferation and invasion. Ultimately, this work aims to identify EN1 as a potential biomarker and therapeutic target for improving glioma treatment outcomes.

Materials and methods

Data source

Transcriptomic and clinical data for 33 cancer types, including 698 glioma samples (GBM + LGG), were obtained from The Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/). An independent validation cohort consisting of 693 glioma samples (CGGA_693) was downloaded from the Chinese Glioma Genome Atlas (CGGA) database (http://www.cgga.org.cn/). Normal brain controls (1,153 samples) were retrieved from the Genotype-Tissue Expression (GTEx) project (https://gtexportal.org). Single-cell RNA-seq data from 44 glioma specimens (GSE182109) were obtained from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). All datasets were publicly available, fully de-identified, and therefore exempt from institutional ethical approval.

EN1 expression analysis

We obtained RNA-seq data for 33 tumor types and their corresponding normal tissues from The Cancer Genome Atlas (TCGA) and the Genotype Tissue Expression (GTEx) databases. Additionally, glioma RNA-seq data were downloaded from the Chinese Glioma Genome Atlas (CGGA). The differences in EN1 expression between unpaired normal and tumor tissues, as well as paired normal and tumor samples were evaluated using the Wilcoxon rank-sum test. P value < 0.05 was considered statistically significant. All expression data were log2-transformed [log2​(value+1)].

Diagnostic value analysis

To determine the diagnostic performance of EN1, we performed receiver operating characteristic (ROC) curve analysis, calculating the area under the ROC curve (AUC). Tumor data from TCGA and normal tissue data from GTEx were used. An AUC > 0.9 indicated excellent performance, AUC > 0.8 was considered good, and AUC > 0.7 suggested useful discrimination.

Survival prognosis analysis

Overall survival (OS) was evaluated using Kaplan–Meier analysis, and subgroup analyses were conducted across different clinical categories of glioma. Patients were stratified into high- and low-EN1 expression groups based on the median EN1 level. Survival curves and hazard ratios were generated using standard survival analysis procedures implemented in R, with statistical significance defined as P < 0.05.

Construction and validation of the nomogram

Variables identified as significant in univariate Cox regression were further analyzed using multivariate Cox models. A prognostic nomogram was constructed to predict 1-, 3-, and 5-year OS, followed by calibration and ROC analyses to assess model performance. Decision curve analysis (DCA) was used to evaluate clinical utility. All procedures were performed using established survival modeling frameworks in R.

Immune microenvironment assessment

Immune cell infiltration was quantified using the CIBERSORT algorithm, while stromal, immune, and ESTIMATE scores were calculated to assess tumor purity. Weighted gene co-expression network analysis (WGCNA) was performed to identify EN1-associated immune modules using an appropriate soft-thresholding power selected according to scale-free topology criteria. Functional enrichment analyses, including GO, KEGG, GSEA, and GSVA, were conducted on genes within key modules using standard enrichment analysis pipelines.

scRNA-seq data processing and analysis

Single-cell RNA-seq data (GSE182109) were processed using Seurat. Quality control was performed by excluding cells with the following thresholds: gene count per cell > 200 and < 6,000; mitochondrial gene percentage < 15%; and UMI count > 1,000. Highly variable genes were identified, and principal component analysis (PCA) and uniform manifold approximation and projection (UMAP) were used for dimensionality reduction and clustering. Cell types were annotated using SingleR. Marker genes were identified using differential expression analysis, and pseudotime trajectories were reconstructed using the Monocle3 package.

Construction of the loop CeRNA network

We employed miRBD (https://mirdb.org/) and starBase (https://rnasysu.com/encori/) to predict miRNAs targeting EN1. The parameters were set to Clade = mammal, Genome = human, Assembly = hg19, and a minimum of 1 program support. Results from both databases were intersected. Potential lncRNAs targeting has-miR-9-5p and has-miR-128-3p were predicted via miRNet (https://www.mirnet.ca/) and starBase; intersections were then taken. The R package “ggalluvial” was used to visualize the ceRNA network. Lastly, JASPAR (jaspar.elixir.no) was employed to predict NEAT1 promoter sites potentially bound by EN1.

Cell culture and transfection

Normal human astrocyte cell line (HA1800) and glioblastoma (GBM) cell lines (U251, U118, and U87) were obtained from the Shanghai Institute of Biochemistry and Cell Biology (Shanghai, China). All cells were cultured in standard DMEM supplemented with 10% FBS and 1% penicillin-streptomycin. Cells were transfected with EN1 shRNA or scrambled control shRNA using Lipofectamine 3000 following the manufacturer’s guidelines. All experiments were performed using three independent biological replicates.

Western blotting

Protein extraction, SDS-PAGE separation, and PVDF membrane transfer were performed using standard procedures. After incubation with primary and secondary antibodies, protein bands were visualized using enhanced chemiluminescence.

CCK-8 assay

Cell proliferation was evaluated using the CCK-8 assay according to the manufacturer’s instructions. Cell viability was calculated as the ratio of the OD value at the indicated time/the OD value at 0 h of the input cells.

Transwell assay

Invasion assays were conducted using Matrigel-coated chambers. Cells transfected with shEN1 or control shRNA were seeded in serum-free medium in the upper chamber, and medium containing 10% FBS was placed in the lower chamber as a chemoattractant. After 24 h, invaded cells were fixed, stained, and counted.

Statistical analysis

All computational analyses were performed in R (version 4.4.0) using the following packages: survival 3.3-1.3, survminer 0.4.9, rms 6.3-0.3, ggplot2 3.4.4, WGCNA 1.73, clusterProfiler 4.4.4, GSVA 1.52.3, estimate 1.0.13, and Seurat 4.4.0. P < 0.05 was considered statistically significant unless otherwise specified. Multiple-testing correction (FDR) was applied for high-throughput analyses including enrichment, GSVA, differential expression, and immune profiling. Biological experiments were analyzed using two-tailed Student’s t-tests or ANOVA where appropriate.

Results

Expression of EN1 in Pan-Cancer

Comparisons of EN1 expression between tumors (TCGA) and normal tissues (GTEx) revealed significantly elevated EN1 levels in adrenocortical carcinoma (ACC), bladder urothelial carcinoma (BLCA), cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC), cholangiocarcinoma (CHOL), colon adenocarcinoma (COAD), diffuse large B-cell lymphoma (DLBC), esophageal carcinoma (ESCA), glioblastoma (GBM), head and neck squamous cell carcinoma (HNSC), kidney chromophobe (KICH), kidney renal clear cell carcinoma (KIRC), kidney renal papillary cell carcinoma (KIRP), lower grade glioma (LGG), liver hepatocellular carcinoma (LIHC), lung squamous cell carcinoma (LUSC), ovarian serous cystadenocarcinoma (OV), pancreatic adenocarcinoma (PAAD), rectum adenocarcinoma (READ), stomach adenocarcinoma (STAD), testicular germ cell tumors (TGCT), thymoma (THYM), and uterine carcinosarcoma (UCS). Conversely, EN1 was downregulated in breast invasive carcinoma (BRCA), acute myeloid leukemia (LAML), prostate adenocarcinoma (PRAD), and skin cutaneous melanoma (SKCM) (Fig. 1A). In paired normal–tumor comparisons within TCGA, EN1 was significantly higher in BLCA, COAD, HNSC, KICH, KIRC, KIRP, LIHC, lung adenocarcinoma (LUAD), LUSC, STAD, and thyroid carcinoma (THCA) (Fig. 1B). These findings suggest that EN1 may play a role in broad oncogenic processes.

Fig. 1
figure 1

EN1 Expression in Pan-Cancer. (A) EN1 expression was significantly upregulated in multiple tumor types (e.g., ACC, BLCA, CESC, COAD, GBM, HNSC, LGG, LIHC, LUSC, PAAD) and downregulated in BRCA, LAML, PRAD, and SKCM. (B) Paired tumor-normal comparisons within TCGA showed higher EN1 expression in BLCA, COAD, HNSC, KICH, KIRC, KIRP, LIHC, LUAD, LUSC, STAD, and THCA.

Full size image

Diagnostic value of EN1 in Pan-Cancer

ROC curve analyses indicated that EN1 had strong diagnostic capabilities (AUC > 0.7) in 13 cancer types: BLCA, CHOL, ESCA, esophageal squamous cell carcinoma (ESCC), glioma, HNSC, KIRC, LUSC, oral squamous cell carcinoma (OSCC), sarcoma (SARC), SKCM, THCA, and THYM (Fig. 2).

Fig. 2
figure 2

Diagnostic Value of EN1 in Pan-Cancer. ROC curve analyses demonstrated that EN1 exhibits strong diagnostic performance (AUC > 0.7) in 13 cancer types, including BLCA, CHOL, ESCA, ESCC, glioma, HNSC, KIRC, LUSC, OSCC, SARC, SKCM, THCA, and THYM.

Full size image

Prognostic analysis of EN1 in cancers

High EN1 expression was significantly associated with poor OS in ACC, BLCA, glioma, KIRC, and KIRP (Fig. 3). Across multiple clinical subgroups—stratified by age, grade, IDH status, and 1p/19q codeletion—elevated EN1 expression consistently predicted unfavorable outcomes in glioma (Fig. 4). These results underscore the prognostic relevance of EN1 in glioma progression.

Fig. 3
figure 3

Prognostic Value of EN1 in Cancers. High EN1 expression was significantly associated with poorer overall survival (OS) in ACC, BLCA, glioma, KIRC, and KIRP.

Full size image
Fig. 4
figure 4

Subgroup Prognostic Analysis of EN1 in Glioma. Elevated EN1 expression was correlated with shorter OS across multiple clinical subgroups in glioma.

Full size image

Construction and validation of nomogram

We identified potential prognostic variables using univariate and multivariate Cox analyses in TCGA glioma cohorts. Age, WHO grade, 1p/19q codeletion, and EN1 were confirmed as independent prognostic factors for glioma (Fig. 5A). Using these factors, we developed a nomogram to predict 1-year, 3-year, and 5-year OS in glioma (Fig. 5B). Internal validation within the TCGA cohort showed strong concordance between predicted and observed survival (Fig. 5C), with ROC AUCs of 0.864, 0.907, and 0.869 for 1-year, 3-year, and 5-year OS, respectively (Fig. 5D). Decision curve analysis (DCA) demonstrated that the nomogram provided higher net benefits than the “treat-all” or “treat-none” strategies across a wide range of threshold probabilities (Fig. 5E-G).

Fig. 5
figure 5

Nomogram Construction and Internal Validation. (A) Age, WHO grade, 1p/19q codeletion, and EN1 were identified as independent prognostic factors for glioma. (B) A nomogram was constructed to predict 1-year, 3-year, and 5-year OS. (C, D) Internal validation in TCGA showed good calibration and high AUCs (0.864, 0.907, and 0.869). (E-G) DCA confirmed higher net benefits of the nomogram.

Full size image

External validation was conducted using CGGA data (Fig. 6A). Similar calibration curve results indicated robust predictive performance (Fig. 6B). The ROC analysis produced AUCs of 0.794, 0.840, and 0.836 for 1-year, 3-year, and 5-year OS, respectively (Fig. 6C). DCA further confirmed the nomogram’s clinical utility in an independent cohort (Fig. 6D-F).

Fig. 6
figure 6

External Validation of the Nomogram. (A) External validation using the CGGA cohort confirmed the robustness of the nomogram. (B) Calibration curves showed excellent agreement between predicted and observed survival in the CGGA cohort. (C) ROC analysis yielded AUCs of 0.794, 0.840, and 0.836. (D-F) DCA demonstrated strong clinical utility in the CGGA cohort.

Full size image

Immune microenvironment analysis

Immune cell composition varied significantly between high- and low-EN1 groups, such as M2 macrophages (27.35%) in high EN1 group, compared with the low-EN1 group (20.37%) (Fig. 7A). High EN1 expression was linked to elevated levels of immunosuppressive cells, such as macrophages and Tregs (Fig. 7B). Patients with high EN1 expression also had higher stromal, ESTIMATE, and immune scores but lower tumor purity (Fig. 7C–F). Kaplan-Meier curves showed that low EN1 expression paired with higher levels of M2 macrophages or resting NK cells (Fig. 7G, H) was associated with poor survival. Likewise, high EN1 expression combined with elevated infiltration of M0 macrophages, M2 macrophages, or neutrophils (Fig. 7I–K) correlated with significantly worse outcomes. Correlation analysis revealed a strong positive association between EN1 and immune checkpoints (PDCD1/PD-1, CD274/PD-L1, CTLA4, LAG3, HAVCR2/TIM-3, and LGALS9/Galectin-9) (Fig. 7L, M).

Fig. 7
figure 7

Immune Microenvironment Analysis of EN1. (A, B) High EN1 expression was associated with increased immunosuppressive cells, such as M2 macrophages, Tregs. (C-F) High EN1 expression correlated with higher stromal, ESTIMATE, and immune scores, but lower tumor purity. (G, H) Low EN1 expression combined with higher levels of M2 macrophages or resting NK cells was linked to poor survival. (I-K) High EN1 expression paired with elevated infiltration of M0 macrophages, M2 macrophages, or neutrophils was associated with worse outcomes. (L, M) EN1 showed strong positive correlations with immune checkpoints.

Full size image

WGCNA analysis and functional enrichment

We performed WGCNA on the CIBERSORT results to identify gene modules related to immune cell infiltration. When the soft threshold was set to 9, the network achieved near scale-free topology (Fig. 8A). After clustering, distinct gene modules were generated (Fig. 8B), each showing unique correlations with various immune cells (Fig. 8C). EN1 showed the strongest positive correlation with the MEblue module (R = 0.724, P < 0.001) (Fig. 8D). Functional enrichment of the MEblue module indicated enrichment in pathways related to oxidative stress, the actin cytoskeleton, and MAPK signaling16 (Fig. 8E). GSEA further revealed significant pathways such as Collagen Formation and TP53-related cell cycle regulation (Fig. 8F), while GSVA showed enrichment of Wnt/β-catenin, TGF-β, and IL-6/JAK/STAT3 signaling in the high-EN1 group (Fig. 8G). These data indicate that EN1 may regulate key oncogenic and immunomodulatory pathways.

Fig. 8
figure 8

WGCNA Analysis and Functional Enrichment. (A) WGCNA identified immune-related gene modules with a soft threshold of 9. (B, C) Gene modules were clustered and correlated with immune cell infiltration. (D) EN1 showed the strongest correlation with the MEblue module (R = 0.724, P < 0.001). (E) Functional enrichment of the MEblue module revealed pathways related to oxidative stress, actin cytoskeleton, and MAPK signaling. (F) GSEA indicated significant enrichment in Collagen Formation and TP53-related cell cycle regulation pathways. (G) GSVA showed enrichment of Wnt/β-catenin, TGF-β, and IL-6/JAK/STAT3 signaling pathways in the high-EN1 group.

Full size image

Single-Cell analysis and drug sensitivity

We analyzed nine GBM samples from GSE182109. PCA followed by UMAP partitioned these cells into 21 clusters (Fig. 9A). Annotation via “SingleR” identified astrocytes, macrophages, monocytes, NK cells, T cells, and tissue stem cells (Fig. 9B). EN1 was predominantly expressed in astrocytes (Fig. 9C, D). Pseudotime analysis revealed an upward trend in EN1 expression from early to late cell states (Fig. 9E, F), suggesting its role in advanced tumor progression. Moreover, drug sensitivity tests showed that the high-EN1 group showed higher predicted IC50 values, indicating potentially reduced sensitivity (i.e., lower sensitivity) for temozolomide, carmustine, vincristine, and cisplatin compared to the low-EN1 group (P < 0.001) (Fig. 9G–J).

Fig. 9
figure 9

Single-Cell Analysis and Drug Sensitivity. (A, B) UMAP and cell type annotation identified 21 clusters, including astrocytes, macrophages, monocytes, NK cells, T cells, and tissue stem cells. (C, D) EN1 was predominantly expressed in astrocytes. (E, F) Pseudotime analysis showed increasing EN1 expression in advanced cell states. (G-J) High-EN1 group exhibited lower sensitivity to temozolomide, carmustine, vincristine, and cisplatin (higher IC50 values).

Full size image

Construction of the loop CeRNA network

Using starBase and miRBD, we identified 20 candidate miRNAs targeting EN1 (Fig. 10A, B). Survival analysis indicated that high expression of has-miR-9-5p and has-miR-128-3p correlated with better prognosis, whereas has-miR-216a-3p, has-miR-369-3p, has-miR-381-3p, and has-miR-944 were associated with poorer outcomes (Fig. 10C–H). Notably, has-miR-9-5p and has-miR-128-3p showed negative correlations with EN1 (Fig. 10I–N).

Fig. 10
figure 10

Identification of miRNAs Targeting EN1. (A, B) Twenty candidate miRNAs targeting EN1 were identified. (C-H) High miR-9-5p and miR-128-3p expression correlated with better prognosis, while miR-216a-3p, miR-369-3p, miR-381-3p, and miR-944 were linked to poorer outcomes. (I-N) miR-9-5p and miR-128-3p showed negative correlations with EN1.

Full size image

Subsequent miRNet and starBase analyses identified four lncRNAs (XIST, SNHG7, NEAT1, and LINC00921) that could bind has-miR-9-5p (Fig. 11A, B), and ten lncRNAs (NEAT1, MAGI2-AS3, SNHG16, etc.) that could bind has-miR-128-3p (Fig. 11C, D). Among these, NEAT1 was common to both miR-9-5p and miR-128-3p. Elevated NEAT1 expression was associated with poor glioma prognosis (Fig. 11E), and correlation analysis showed negative correlations between NEAT1 and both miR-9-5p and miR-128-3p (Fig. 11F, G). EN1 binding motifs were predicted in the promoter region of NEAT1 according to JASPAR (Fig. 11H). Integrating these findings, we constructed a NEAT1/miR-9-5p/miR-128-3p/EN1 loop ceRNA network (Fig. 11I).

Fig. 11
figure 11

Construction of the Loop ceRNA Network. (A-D) NEAT1, shared by miR-9-5p and miR-128-3p, was identified among binding lncRNAs. (E) NEAT1 expression was associated with poor prognosis. (F, G) NEAT1 negatively correlated with miR-9-5p and miR-128-3p. (H) JASPAR predicted EN1 binding motifs in the NEAT1 promoter region. (I) The NEAT1/miR-9-5p/miR-128-3p/EN1 ceRNA network was constructed.

Full size image

EN1 expression and functional validation in glioblastoma cells

Based on the bioinformatic evidence suggesting an oncogenic role of EN1 in glioma, we next performed in-vitro experiments to validate its biological functions. Western bloting analysis indicated markedly higher EN1 expression in glioblastoma cell lines (U87 and U118) compared to normal astrocyte (HA1800) (P < 0.05) (Fig. 12A). After EN1 knockdown via shRNA (shEN1), both U87 and U118 cells exhibited significantly reduced proliferation at 24 h, with more pronounced effects at 72 and 120 h, as shown by CCK-8 assays (Fig. 12B, C). Transwell invasion assay further demonstrated that EN1 knockdown substantially inhibited the invasive capacity of these glioblastoma cells (P < 0.05) (Fig. 12D). These in vitro findings support a pro-tumorigenic role of EN1 in glioblastoma cells and are consistent with the bioinformatic associations observed in patient cohorts.

Fig. 12
figure 12

EN1 Expression and Functional Validation in Glioblastoma Cells. (A) EN1 expression was significantly higher in glioblastoma cell lines (U87, U118) than in normal astrocytes (P < 0.05). (B, C) EN1 knockdown reduced proliferation of U87 and U118 cells (P < 0.05). (D) EN1 knockdown significantly impaired the invasive ability of glioblastoma cells (P < 0.05).

Full size image

Discussion

In this study, we systematically evaluated the expression patterns and clinical relevance of Engrailed-1 (EN1) across multiple malignancies, with a particular focus on glioma. Our analyses across TCGA, GTEx, and CGGA databases demonstrated that elevated EN1 expression correlates with poor prognosis in a subset of solid tumors, including ACC, BLCA, glioma, KIRC, and KIRP. Notably, in glioma cohorts, EN1 showed strong diagnostic and prognostic potential, suggesting that EN1 may serve as a biomarker and a potential contributor to glioma progression.

Several studies have highlighted the significance of homeobox-containing transcription factors in tumorigenesis, typically through their ability to modulate cell fate, proliferation, and differentiation programs17,18,19,20. Our findings underscore this notion, as high EN1 expression was strongly associated with worse OS in glioma. By integrating EN1 into a prognostic nomogram with age, WHO grade, and 1p/19q codeletion status, we achieved robust predictive performance. This multidimensional approach aligns with the contemporary trend of employing multi-parameter models rather than relying on single-gene markers, thereby improving the accuracy of patient stratification. Given that clinical management of glioma varies significantly based on tumor grade and molecular profiles, the addition of EN1 into existing prognostic frameworks could refine therapeutic decision-making.

Our immune infiltration analyses revealed a pronounced correlation between elevated EN1 levels and increased infiltration of immunosuppressive cells, notably M2 macrophages. M2-polarized macrophages are known to secrete anti-inflammatory cytokines (e.g., IL-10, TGF-β) and support tumor immune evasion, angiogenesis, and metastasis21,22,23. We also observed that high EN1 expression was associated with lower tumor purity and higher ESTIMATE and immune scores, suggesting a more complex stromal-tumor interface in glioma. These findings align with recent investigations demonstrating that transcription factors can modulate tumor-associated macrophage (TAM) polarization24,25. Specifically, these observations raise the possibility that EN1-related transcriptional programs may be linked to macrophage recruitment or polarization, thus fostering an immunosuppressive microenvironment conducive to tumor progression. However, direct experimental validation is required.

Furthermore, our correlation analyses showed a robust positive association between EN1 and various immune checkpoint molecules (e.g., PD-1/PDCD1, PD-L1/CD274, CTLA4), echoing literature indicating that high levels of immune checkpoints often coexist with immunosuppressive microenvironment26,27,28.

Future mechanistic studies, such as chromatin immunoprecipitation (ChIP) and reporter gene assays, could clarify whether EN1 directly controls the transcription of immunomodulatory genes or interacts with other transcription factors regulating checkpoint pathways.

By leveraging single-cell RNA-seq data, we identified that EN1 is predominantly localized to astrocyte glioma cells and exhibits progressive upregulation over pseudotime. This suggests that EN1 may play a critical role in maintaining a more advanced or stem-like phenotype in glioma cells, consistent with reports linking transcription factors to glioma stem cell compartments29,30. Intriguingly, our drug sensitivity analyses showed that EN1-high glioma cells were more resistant to chemotherapeutic agents, including temozolomide—considered a cornerstone in glioma therapy. Potential mechanisms underlying EN1-associated resistance could include (i) regulation of DNA damage repair pathways, as has been noted in other homeoproteins31,32; (ii) direct upregulation of multidrug efflux transporters (e.g., ABC family transporters)33,34; or (iii) enhanced pro-survival signaling (e.g., PI3K/AKT, MAPK) that reduces apoptosis in response to chemotherapeutic stress12,13.

One of the notable aspects of our study is the establishment of a ceRNA-based regulatory loop involving EN1, NEAT1, and miR-9-5p/miR-128-3p. ceRNA networks have emerged as a crucial layer of post-transcriptional regulation in cancer, where lncRNAs act as molecular sponges for tumor-suppressive miRNAs, thereby preventing them from binding to their mRNA targets35,36. NEAT1, in particular, has been described as an oncogenic lncRNA in multiple tumor types, capable of modulating cell proliferation, migration, and even therapy resistance37,38,39.

We observed that NEAT1 could bind both miR-9-5p and miR-128-3p, while these miRNAs showed a strong negative correlation with EN1 expression. This arrangement is consistent with a putative ceRNA model in which NEAT1 may act as a molecular sponge for miR-9-5p and miR-128-3p, potentially attenuating their inhibitory effects on EN1 expression. Moreover, predictive binding site analysis (via JASPAR) indicated that EN1 may, in turn, bind to the NEAT1 promoter region, implying a possible positive feedback loop. Such reciprocal “activation” loops can significantly amplify oncogenic signals and contribute to tumor aggressiveness.

Our functional assays confirmed that EN1 knockdown significantly attenuates glioblastoma cell proliferation and invasion, reinforcing the oncogenic role of this transcription factor. Given EN1’s links to immunosuppressive signals and chemoresistance, a multifaceted therapeutic strategy could be envisioned. On one hand, combining EN1 inhibitors (or knockdown approaches) with immune checkpoint blockade might help overcome the limitations of monotherapy in glioma patients, as both intrinsic and extrinsic mechanisms of tumor immune evasion could be diminished.

On the other hand, modulating the NEAT1/miR-9-5p/miR-128-3p/EN1 axis—perhaps through antisense oligonucleotides targeting NEAT1 or miRNA mimics—may further sensitize tumor cells to conventional therapies.

Despite our comprehensive bioinformatic analyses of EN1 across multiple datasets and the functional validation in glioblastoma cell lines, this study still lacks in-depth mechanistic investigations. While our results strongly suggest that EN1 exerts oncogenic effects and modulates the tumor microenvironment, the precise molecular pathways and direct interactions remain unclear. Future work involving detailed mechanistic experiments (e.g., chromatin immunoprecipitation, co-immunoprecipitation, or proteomics) will be essential to elucidate how EN1 regulates downstream signaling networks and contributes to tumor progression.

Conclusion

In summary, our study demonstrates that EN1 is consistently associated with glioma progression, immune microenvironment–related features, and predicted therapeutic response across multiple datasets. Integrative analyses suggest a putative NEAT1/miR-9-5p/miR-128-3p/EN1 ceRNA regulatory axis that may contribute to EN1 dysregulation in glioma. Together, these findings highlight EN1 as a potential biomarker and candidate target, warranting further mechanistic and translational investigations.