SMPD1 as a Potential Prognostic Biomarker in Glioma Is Associated With an Immunosuppressive Microenvironment
Yanan Xu, Xing Liu, Boya Xu, Qiuling Li, Luofei Zhang, Cao Li, Zhigang Zhao
TL;DR
SMPD1 is a potential biomarker in glioma linked to poor prognosis and an immunosuppressive environment, suggesting a role in tumor progression and immunotherapy.
Contribution
This study identifies SMPD1 as a novel prognostic biomarker and therapeutic target in glioma associated with immune cell infiltration and tumor progression.
Findings
High SMPD1 levels correlate with poor prognosis and immunosuppressive tumor microenvironment in gliomas.
ASM inhibition promotes M1 macrophage polarization and suppresses tumor growth in vivo.
SMPD1 expression is elevated in high-grade, IDH-wildtype, and MGMT-unmethylated gliomas.
Abstract
Acid sphingomyelinase (ASM), encoded by SMPD1, regulates sphingolipid metabolism and has been implicated in tumor progression and immune modulation. However, its role in glioma remains poorly defined. We performed a comprehensive analysis of SMPD1 in gliomas using TCGA and CGGA datasets, evaluating its expression patterns, prognostic significance, immune correlations, pathway enrichment, and copy number variation. Using qRT–PCR, we validated in vitro the effect of SMPD1 expression on macrophage polarization. Immunofluorescence staining was used to assess the levels of ASM of clinical samples and its correlation with tumor‐associated macrophages. The functional role of SMPD1 was further validated in vivo. SMPD1 expression was significantly elevated in high‐grade, IDH‐wildtype, and MGMT‐unmethylated gliomas. High SMPD1 levels were associated with poor prognosis and served as an…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
FIGURE 1
FIGURE 2
FIGURE 3
FIGURE 4
FIGURE 5
FIGURE 6
FIGURE 7| Variable | HR | 95% CI |
|
|---|---|---|---|
|
| 2.448 | 1.774–3.379 | < 0.001 |
| Gender | 1.001 | 0.743–1.347 | 0.997 |
| Age | 1.075 | 1.063–1.088 | < 0.001 |
| WHO grade | 4.912 | 3.830–6.298 | < 0.001 |
| IDH status | 11.071 | 7.772–15.771 | < 0.001 |
| 1p19q status | 4.541 | 2.671–7.719 | < 0.001 |
| MGMT status | 3.207 | 2.312–4.447 | < 0.001 |
| Variable | HR | 95% CI |
|
|---|---|---|---|
|
| 1.027 | 1.018–1.036 | < 0.001 |
| Gender | 0.941 | 0.716–1.236 | 0.660 |
| Age | 1.033 | 1.020–1.046 | < 0.001 |
| WHO grade | 2.912 | 2.417–3.508 | < 0.001 |
| IDH status | 2.821 | 2.137–3.724 | < 0.001 |
| 1p19q status | 5.890 | 3.610–9.609 | < 0.001 |
| MGMT status | 1.206 | 0.919–1.582 | 0.177 |
- —Young elite scientist sponsorship program
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsGlioma Diagnosis and Treatment · Caveolin-1 and cellular processes · Clusterin in disease pathology
Background
1
Gliomas constitute the most prevalent primary tumors of the adult central nervous system. According to the 2021 WHO classification, gliomas are categorized into astrocytomas, oligodendrogliomas, and glioblastomas (GBM), among which GBM exhibits the most aggressive subtype [1]. These tumors are characterized by poor prognosis and a high likelihood of recurrence. For GBM in particular, the median overall survival (OS) rarely exceeds 15 months, and the 5‐year survival rate remains below 10% [2]. Although astrocytomas and oligodendrogliomas are associated with relatively prolonged survival, many ultimately progress to higher‐grade disease or relapse, which substantially compromises prognosis [3]. These challenges underscore an urgent need for novel therapeutic strategies to improve outcomes for glioma patients.
Immunotherapy has emerged as a transformative modality in cancer treatment, demonstrating substantial benefits in malignancies such as non‐small‐cell lung cancer [4] and melanoma [5]. However, advances in glioma immunotherapy have been comparatively modest. The restrictive nature of the blood–brain barrier, the presence of an immunosuppressive tumor microenvironment (TME), and tumor heterogeneity limit robust immune activation in these tumors [6]. Therefore, strategies aimed at remodeling the immunosuppressive TME may be crucial for converting the “cold” immune landscape into a more responsive state, ultimately enhancing the therapeutic potential of immunotherapy in neuro‐oncology.
One molecule implicated in shaping immune responses is acid sphingomyelinase (ASM). ASM hydrolyses sphingomyelin to generate ceramide, interacts with multiple immune cell subsets to orchestrate the balance between pro‐tumor and anti‐tumor immunity [7]. In melanoma, ASM has been implicated in shaping the immune milieu by modulating immune cell recruitment and cytokine secretion [8]. Moreover, ASM influences the functional states of CD8^+^ T cells [9, 10], CD4^+^ T cells [11], macrophages [12], and natural killer cells [13, 14], highlighting its importance as a regulator of tumor‐associated immune responses.
Despite these insights, the role of SMPD1—the gene encoding ASM—in glioma remains insufficiently characterized. In this study, we carried out integrated bioinformatic analyses using publicly available datasets from TCGA and CGGA to delineate SMPD1 expression patterns across gliomas and to evaluate its relationship with clinical outcomes and tumor immunity. Furthermore, analyses of human glioma specimens and experiments in vitro and in vivo were conducted to validate the function of ASM and to assess its potential as a prognostic biomarker and therapeutic target in glioma.
Materials and Methods
2
Data Acquisition
2.1
RNA‐ seq data, clinical characteristics, and copy number variation data for gliomas, including both GBM and lower‐grade gliomas (LGG), were obtained from the Cancer Genome Atlas Program (TCGA) database [15]. In addition, RNA‐seq data and clinical information for glioma cases were retrieved from the Chinese Glioma Genome Atlas (CGGA) [16]. In total, 702 cases from TCGA and 325 cases from CGGA were included in the bioinformatic analyses.
Patients and Samples
2.2
Tumor specimens were obtained from glioma patients who underwent surgical resection at Beijing Tiantan Hospital. The collection of clinical specimens and corresponding patient information used in this study was approved by the Ethics Committee of Beijing Tiantan Hospital. Disease diagnosis and histopathological evaluations were performed by the Beijing Tiantan Hospital and Beijing Neurosurgical Institute.
Gene and Phenotype Analyses
2.3
Clinical phenotypes included 2021 WHO classification grade, IDH mutation status, 1p/19q codeletion status, and MGMT promoter methylation status. Associations between gene expression and clinical phenotypes were visualized as heatmaps using the “pheatmap” package in R (version 4.3.3, hereafter). Plots illustrating differences in gene expression across clinical subgroups were generated using GraphPad Prism (version 10).
Prognostic Analysis
2.4
OS was calculated from the date of diagnosis to death from any cause or last follow‐up. Glioma cases were classified as high and low SMPD1 expression groups using the median expression level as the cutoff within each dataset. Kaplan–Meier survival curves were generated using the “survival” and “survminer” R packages. The differences between groups were assessed using the log‐rank test. Time‐dependent receiver operating characteristic (ROC) curves were constructed to evaluate the prognostic performance of SMPD1 expression at 1, 3, and 5‐year OS.
Univariate and multivariate Cox proportional hazards models were fitted to identify the independent prognostic value of SMPD1. Results are presented as hazard ratios (HRs) with 95% confidence intervals (CIs) and p values. The results are presented as summary tables and forest plots.
Establishment and Validation of the Nomogram Model
2.5
A prognostic nomogram model was constructed to predict 1, 3, and 5‐year OS in glioma patients using the “rms” and “survival” R packages [17]. The nomograms were developed based on multivariate Cox proportional hazards regression analyses. To evaluate generalizability, we performed reciprocal external validation between the TCGA and CGGA cohorts. Specifically, a nomogram was first trained in the TCGA cohort and externally validated in the CGGA cohort using the same model coefficients. Conversely, a second nomogram was trained in the CGGA cohort and externally validated in the TCGA cohort. Calibration curves at 1, 3, and 5 years were generated to compare predicted and observed survival probabilities. Model discrimination was quantified by the concordance index (C‐index). Time‐dependent Brier scores at 1, 3, and 5 year OS were estimated using the “pec” package to evaluate the overall predictive accuracy. This reciprocal validation strategy enables assessment of model performance across independent patient cohorts and helps reduce overfitting [18, 19].
Immune Infiltration Analysis
2.6
Immune cell infiltration was estimated using the CIBERSORT algorithm [20], implemented via the “e1071” and “preprocessCore” R packages. Pearson correlation analyses were then performed to explore associations between SMPD1 expression and immune‐related markers, including T‐cell markers, macrophage markers, and immune checkpoint molecules. Correlation coefficients (r) and corresponding p values were calculated, and results with |r| > 0.2 and p < 0.05 were considered statistically significant.
Functional Enrichment Analysis
2.7
Genes were ranked based on their Pearson correlation with SMPD1 expression, and the top 500 most positively correlated genes (p < 0.05) were uploaded to the Database for Annotation, Visualization, and Integrated Discovery (DAVID, version 6.8, https://david.ncifcrf.gov/). Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses [21] were conducted to identify enriched signaling pathways. A false discovery rate (FDR) of < 0.1 was used as the threshold for statistical significance. Data visualizations were carried out using the “ggplot2” R package and the SangerBox platform [22].
Copy Number Variation (CNV) Analysis
2.8
CNV data for glioma patients were downloaded from the TCGA database by using the “TCGAbiolinks” package in R. Significant genomic amplifications and deletions were identified with GISTIC 2.0, based on segmented copy‐number data aligned to the hg19 reference genome [23]. Key parameters included an amplification/deletion threshold of ±0.3 and a confidence level of 0.95 [24]. All other parameters were kept at their default settings. The deletion group was further divided into loss (−0.8 < segment mean (seg.mean) < −0.3) and homozygous deletion (seg.mean ≤ −0.8), whereas the amplification group was subdivided into gain (0.3 < seg.mean < 0.8) and high‐level amplification (seg.mean ≥ 0.8) [25]. For downstream survival analysis, CNV status was summarized into deletion, wildtype, using ±0.3 as the operational thresholds, consistent with GISTIC 2.0 calling conventions.
The “maftools” R package was employed to visualize the mutation landscape of gliomas [26]. For gene‐specific analyses, Kaplan–Meier curves were generated to compare outcomes across SMPD1 CNV status.
Cell Isolation, Culture, and Polarization
2.9
Bone‐marrow‐derived macrophages (BMDMs) were generated from ASM wildtype (WT) and full‐knockout (KO) C57BL/6 mice (6–8 weeks). Briefly, mice were euthanized, and femurs and tibias were aseptically dissected. Bone marrow cells were flushed with cold complete medium and passed through a 70 μm strainer. Cells were washed and seeded in DMEM supplemented with 10% fetal bovine serum (FBS), 1% penicillin/streptomycin, and 25 ng/mL macrophage colony‐stimulating factor (315‐02, PeproTech). Cells were cultured at 37°C in a humidified atmosphere with 5% CO_2_, and fresh M‐CSF–containing medium was replaced on day 3 and day 5. After 7 days of differentiation, adherent cells were considered mature BMDMs and used for subsequent experiments.
For polarization, BMDMs were gently scraped, reseeded at 2 × 10^6^ cells/well in 6‐well plates. To induce M1‐like macrophages, cells were stimulated with 100 ng/mL LPS (L4391, Sigma‐Aldrich) and 20 ng/mL IFN‐γ (AF‐300‐02, PeproTech) for 24 h. To induce M2‐like macrophages, cells were treated with 20 ng/mL IL‐4 (214‐14, PeproTech) for 24 h. To inhibit ASM during polarization, amitriptyline (20 μM, A8404, Sigma‐Aldrich) was added simultaneously [27, 28].
Quantitative Real‐Time PCR
2.10
Total RNA was extracted from polarized BMDMs using TRIzol reagent (15596018, Invitrogen) according to the manufacturer's instructions. RNA concentration and purity were determined using a NanoDrop One spectrophotometer, and samples with an A260/280 ratio between 1.8 and 2.0 were used for further experiments. For cDNA synthesis, 1 μg of total RNA was reverse‐transcribed using SweScript All‐in‐One RT SuperMix for qPCR kit (G3329, Servicebio) following the manufacturer's protocol. Quantitative real‐time PCR (qRT‐PCR) was performed using the SYBR Green qPCR kit (G3326, Servicebio). The thermocycling conditions were as follows: initial denaturation at 95°C for 30 s, followed by 40 cycles of 95°C for 15 s and 60°C for 30 s. Melting curve analysis was performed to confirm the specificity of amplification. Hprt was used as the internal control. Relative mRNA abundance was quantified with the 2−ΔΔCt method. Primer sequences are available in Table S1.
Immunofluorescence Staining
2.11
Frozen tumor tissue sections were incubated overnight at 4°C in a humidified chamber with primary antibodies (ASM, 1:50 dilution, ab315810, Abcam; CD163, 1:500 dilution, GB15340, Servicebio). After incubation, appropriate secondary antibodies (1:800 dilution, GB21303 and GB25303, Servicebio) were applied for 50 min at room temperature. Sections were then washed and stained with DAPI for 10 min at room temperature in the dark. After a final rinse, slides were mounted and imaged using an upright fluorescence microscope.
In Vivo Tumor Model in C57BL/6J Mice
2.12
ASM‐full KO mice were generated by breeding heterozygous ASM‐deficient C57BL/6J mice. Genotypes were confirmed by PCR analysis prior to experimental use. Five‐week‐old ASM‐KO and wild‐type mice (n = 7 per group) were subcutaneously injected with the murineglioma cell line GL261 (5 × 10^6^ cells per mouse). Tumor volumes were measured every 3 days using a digital caliper, and body weight was recorded throughout the study. Tumor growth curves were analyzed using two‐way analysis of variance (ANOVA) with repeated measures, and endpoint tumor weights were compared using an unpaired two‐tailed Student's t‐test. All animal procedures were approved by the Institutional Animal Care and Use Committee of the Beijing Neurosurgical Institute.
Statistical Analysis
2.13
Statistical computations and graphical outputs were generated in R and GraphPad Prism. Normality of continuous variables was assessed using the Shapiro–Wilk test or Kolmogorov–Smirnov test, together with histogram and Q–Q plot inspection. Normally distributed data were analyzed with Student's t‐test (two groups) or one‐way ANOVA (three or more groups), whereas non‐normal data were evaluated using the Mann–Whitney U test or Kruskal–Wallis test. All tests were two‐sided, and statistical significance was set at p < 0.05.
Results
3
High Expression of
SMPD1 Relates to Malignant Clinical Characteristic in Gliomas
3.1
Patients with different levels of SMPD1 expression exhibited distinct clinicopathological characteristics. In both the TCGA and CGGA datasets, increasing SMPD1 expression was associated with a non‐random distribution of O6‐methylguanine‐DNA methyltransferase (MGMT) promoter methylation status, 1p/19q codeletion status, IDH mutation status, WHO grade, and histopathological diagnosis (Figure 1A,B). Specifically, SMPD1 expression was significantly elevated in WHO grade 4 gliomas in both datasets (Figure 1C,G). It was also markedly higher in IDH‐wildtype gliomas (Figure 1D,H) and in tumors with unmethylated MGMT (Figure 1F,J). In the CGGA dataset, SMPD1 expression was also significantly increased in tumors without 1p/19q codeletion (Figure 1I), although no significant difference was observed in the TCGA cohort (Figure 1E). Collectively, these findings suggest a strong association between SMPD1 expression and key clinical features of glioma.
*Association between SMPD1 expression and clinical characteristics of gliomas in TCGA (A, C–F) and CGGA (B, G, H) datasets. Landscape of SMPD1 expression and clinical characteristics in each dataset (A, B). Association with WHO grade (C, G). Association with IDH mutation status (D, H). Association with 1p19q codeletion status (E, I). Association with MGMT promoter methylation status (F, J). *, p < 0.05; **, p < 0.01; ***, p < 0.001; ***, p < 0.0001.
High Expression of
SMPD1 Indicates Poor Prognosis in Gliomas
3.2
Survival analyses from both the TCGA (Figure 2A) and CGGA (Figure 2D) datasets revealed that patients with high SMPD1 expression had significantly poorer OS compared with those with low SMPD1 expression. Specifically, in high‐grade gliomas, elevated SMPD1 expression was associated with a significantly reduced survival or a trend toward poorer survival (Figure S1). ROC curve analysis showed that the area under the curve (AUC) values for 1, 3, and 5‐year OS in the TCGA cohort were 0.701, 0.672, and 0.619, respectively (Figure 2B), while the corresponding AUCs in the CGGA cohort were 0.654, 0.692, and 0.690 (Figure 2E), indicating a fair prognostic ability of SMPD1 as a single gene predictor [29].
Prognostic significance of SMPD1 expression in gliomas. Kaplan–Meier curves for overall survival in TCGA (A) and CGGA (D) datasets. Time‐dependent ROC curves for overall survival prediction in TCGA (B) and CGGA (E) datasets. Multivariate Cox regression analysis of overall survival in TCGA (C) and CGGA (F) datasets. Nomogram models for prognostic prediction in TCGA (G) and CGGA (I) datasets, with corresponding calibration plots (H, J).
Cox models were fitted in univariable and multivariable analyses to assess prognostic factors. In the TCGA cohort, univariable analysis identified SMPD1 expression, age, WHO grade, IDH mutation status, 1p/19q codeletion, and MGMT promoter methylation status as significant prognostic variables (p < 0.001) (Table 1). Similar findings were observed in the CGGA cohort, with the exception of MGMT promoter methylation status (Table 2). Multivariate Cox regression further confirmed that SMPD1 expression remained an independent prognostic factor in both datasets (Figure 2C,F; TCGA: p = 0.022; CGGA: p = 0.033).
Establishment of Clinical Nomogram in Gliomas
3.3
To further investigate the potential prognostic utility of SMPD1 in gliomas, nomogram models were constructed using the TCGA (Figure 2G) and CGGA (Figure 2I) datasets.
In the TCGA training cohort, the nomogram showed excellent discrimination with a C‐index of 0.847. Calibration curves with 95% CIs demonstrated agreement between predicted and observed survival probabilities across the entire risk spectrum. (Figure 2H). The time‐dependent Brier scores at 1, 3, and 5 years were 0.101, 0.104, and 0.134, indicating good overall prediction accuracy (Figure S2A).
In the independent CGGA training cohort, the model retained favorable performance. Calibration curves with 95% CIs indicated good concordance between predicted and observed survival outcomes (Figure 2J). The model achieved a C‐index of 0.762, and the time‐dependent Brier scores at 1, 3, and 5 years were 0.162, 0.138, and 0.118, respectively (Figure S2B).
External validation of the TCGA‐derived nomogram in the independent CGGA cohort showed limited predictive performance. Calibration curves at 1, 3, and 5 years revealed noticeable deviations from the ideal 45° line, indicating some overestimation of survival probabilities in high‐risk patients (Figure S2C). The C‐index was 0.551, indicating only slightly better discrimination than random chance. Time‐dependent Brier scores were 0.239, 0.278, and 0.245 at 1, 3, and 5 years, respectively (Figure S2E), suggesting limited overall predictive accuracy.
When the prognostic model trained on the CGGA cohort was applied to the TCGA cohort for external validation, the calibration plots showed a tendency to lie above the ideal 45° reference line, suggesting underestimation of survival probabilities (Figure S2D). The C‐index in the TCGA dataset was 0.638, and the corresponding Brier scores at 1, 3, and 5 years were 0.210, 0.221, and 0.163, respectively (Figure S2F). Overall, the model exhibited moderately better generalizability and calibration properties than the nomogram constructed using the TCGA cohort [30].
Higher
SMPD1 Expression Is Associated With Increased Infiltration of Immunosuppressive Cells
3.4
To determine whether SMPD1 influences the immune microenvironment in glioma, immune cell infiltration analysis was performed using the CIBERSORT algorithm. In the TCGA dataset (Figure 3A), SMPD1 expression was positively correlated with multiple immune cell types, including regulatory T cells (Tregs), resting natural killer (NK) cells, M0 macrophages, and M2 macrophages. Similarly, in the CGGA dataset (Figure 3B), SMPD1 showed strong positive correlations with M0 macrophages, M2 macrophages, γδ T cells, and Tregs. Notably, regulatory T cells and M0/M2 macrophages were consistently positively correlated with SMPD1 expression across both datasets.
Association between SMPD1 and immune microenvironment. Correlation between SMPD1 expression and immune cell infiltration in TCGA (A) and CGGA (B). Correlation between SMPD1 expression and immune cell markers in both datasets (C). The upper panel shows results from the TCGA dataset, and the lower panel shows results from the CGGA dataset. Correlation between SMPD1 expression and immune checkpoint molecules (D).
We next focused on T cells and macrophages to explore correlations between SMPD1 and canonical cell markers. As shown in Figure 3C, in both datasets, SMPD1 expression was positively associated with the T‐cell exhaustion marker PD‐1 [31], M2‐associated macrophage markers CD163, CD204, and TGF‐β [32, 33], and the antigen‐presenting cell marker CD40 [34]. SMPD1 expression also correlated with the M1 transcription factor IRF5 and with pan‐macrophage markers, including CD68 [35], CD16, CD32, CD64 [36], and CCL2 [37]. The detailed expression heatmaps are provided in Figures S3 and S4. SMPD1‐high tumors appear to recruit more macrophages overall, and many of these macrophages may be immunosuppressive (M2‐polarized), as indicated by the strong correlations with CD163, CD204, TGF‐β, among others.
Additionally, the relationship between SMPD1 and immune checkpoint molecules was assessed using both TCGA and CGGA datasets (Figure 3D). SMPD1 expression showed positive correlations with the majority of immune checkpoint genes, whereas KLRK1 and VTCN1 were significantly negatively correlated.
SMPD1
Is Enriched in Immune and Inflammatory Regulatory Pathways
3.5
To explore the potential biological functions of SMPD1, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed based on the top 200 genes most strongly correlated with SMPD1 expression.
In the TCGA dataset (Figure 4A), GO enrichment analysis revealed that SMPD1‐related genes were significantly enriched in biological processes such as regulation of inflammatory response and NF‐κB signaling pathway activation. The most enriched cellular components included the Golgi apparatus and lysosome, while molecular function terms were primarily associated with various binding activities, including identical protein binding, misfolded protein binding, and phosphatidylinositol‐3‐phosphate binding. In the CGGA dataset (Figure 4B), NF‐κB signaling, innate immune response, interleukin‐6 (IL‐6) and interleukin‐1β (IL‐1β) production, and the toll‐like receptor 4 signaling pathway were significantly enriched. Enriched cellular components and molecular functions were broadly consistent with those identified in the TCGA dataset.
Enrichment analysis based on SMPD1‐associated genes. Biological processes, cellular components, and molecular functions related to SMPD1 in TCGA (A) and CGGA (B). KEGG pathway analysis of SMPD1 in TCGA (C) and CGGA (D).
KEGG pathway analysis indicated that SMPD1 may be involved in pathways such as lysosome function and metabolic regulation, as well as immune‐related pathways, including the toll‐like receptor, NOD‐like receptor, and IL‐17 signaling pathway (Figure 4C,D).
SMPD1
Deletion Predicts Favorable Early Survival, While High Expression Associates With Adverse Mutations
3.6
CNV and mutation analyses were conducted to assess the genomic alterations of SMPD1 in glioma. Homozygous (seg.mean ≤ −0.8) or heterozygous (−0.8 < seg.mean ≤ −0.3) deletions of SMPD1 were more common than amplifications in glioma samples (Figure 5A). The prognostic outcomes of patients with mutant and wild‐type SMPD1 are shown in Figure S5. Notably, SMPD1 deletion correlated with a significant improvement in short‐term survival (1–4 years) among patients with glioma (Figure 5B). Moreover, SMPD1 deletions led to lower mRNA expression compared with copy‐number neutral tumors (Figure 5C), confirming the CNV effect.
*CNV analysis of SMPD1. CNV status and proportion of SMPD1 in the TCGA dataset (A). Kaplan–Meier curves of overall survival for patients with SMPD1 deletion and wild‐type (B). SMPD1 expression across different CNV statuses (C). Waterfall plot of the top 15 mutated genes based on SMPD1 expression in low‐grade gliomas (D) and in glioblastoma (E). *, p < 0.05; **, p < 0.01; ***, p < 0.001; ***, p < 0.0001.
In patients with lower‐grade glioma, high SMPD1 expression was associated with a higher frequency of TP53 and ATRX mutations (Figure 5D). In GBM, high SMPD1 expression was more frequently accompanied by PTEN mutations (Figure 5E), a genetic alteration known to be linked to poor prognosis [38].
ASM Is Associated With M2 Macrophage Polarization in Polarized BMDMs
3.7
To experimentally validate the role of ASM in macrophage polarization, qPCR analysis was performed in murine BMDMs under M1 and M2 polarizing conditions. Compared with WT BMDMs, both ASM‐deficient BMDMs and amitriptyline‐treated BMDMs exhibited significantly increased expression of M1 markers (Nos2 and Il‐1b) and decreased expression of M2 markers (Arg1 and Cd163) (Figure 6A). These findings indicate that inhibition of ASM skews macrophage polarization away from an M2‐like phenotype toward a more pro‐inflammatory M1‐like phenotype.
*qPCR analysis of macrophage polarization markers and immunofluorescence analysis of human glioma tissues. qPCR quantification of M1‐associated (Nos2, Il‐1β) and M2‐associated (Arg1, Cd163) macrophage markers (A). Representative immunofluorescence staining of ASM and CD163 in human glioma tissues (B). Quantification of ASM immunofluorescence intensity in gliomas stratified by WHO grades (C), IDH type (D), and 1p19q type (E). Pearson correlation between ASM and CD163 (F). *, p < 0.05; **, p < 0.01; ***, p < 0.001; ***, p < 0.0001.
ASM Expression Varies Across Human Glioma Samples and Positively Correlates With M2 Macrophage Polarization
3.8
Thirty‐four clinical glioma specimens were collected to examine intratumoral expression of ASM and its association with CD163. The results revealed that ASM expression increased with the glioma grade (Figure 6B,C). Elevated ASM levels were also observed in gliomas with IDH wild‐type status and absence of 1p/19q codeletion (Figure 6D,E). Moreover, ASM expression was positively correlated with the M2 macrophage marker CD163 (Figure 6F), consistent with findings from database analyses and in vitro experiments, suggesting that ASM may contribute to an immunosuppressive tumor microenvironment.
Smpd1
KO Suppresses Glioma Growth In Vivo
3.9
To evaluate the functional role of SMPD1 in glioma progression in vivo, GL261 glioma cells were subcutaneously injected into Smpd1‐KO C57BL/6J mice. Compared with wild‐type controls, tumors in Smpd1‐KO mice were significantly smaller (Figure 7A,C), exhibited markedly slower growth rates (Figure 7B), and had significantly lower tumor weights at endpoint (Figure 7D).
*Subcutaneous tumor model for evaluation of SMPD1 in vivo. Photographs of the dissected tumor (A). Glioma growth curves of SMPD1 knocked‐out and wild‐type mice (B). Glioma volumes in SMPD1 knocked‐out and wild‐type mice (C). Glioma weights in SMPD1 knocked‐out and wild‐type mice (D). *, p < 0.05; **, p < 0.01; ***, p < 0.001; ***, p < 0.0001.
Discussion
4
Major Findings and Comparison With Previous Studies
4.1
Gliomas are the most common and aggressive primary tumors of the central nervous system. Among them, GBM represents the most malignant form. Despite advances in clinical management, the prognosis for glioblastoma (GBM) remains poor. The current standard of care includes maximal safe surgical resection, followed by radiotherapy and temozolomide chemotherapy [39, 40]. However, these interventions offer only limited benefit. The median survival for patients with GBM is approximately 12–15 months [41]. The limited efficacy of conventional treatments and resistance to therapy underscore the urgent need to identify novel molecular targets and therapeutic strategies.
ASM, encoded by SMPD1, plays a crucial role in sphingolipid metabolism [42]. Ceramide functions as a bioactive lipid mediator involved in regulating cell proliferation, apoptosis, autophagy [43], and immune responses [44]. Aberrant ceramide signaling has been implicated in tumor progression and treatment resistance in various malignancies, including gliomas [45], breast [46], prostate [47], and colorectal cancers [48]. Recent studies have suggested that altered SMPD1 or ceramide may modulate the tumor microenvironment and influence responses to chemotherapy or immunotherapy [49, 50].
In the context of gliomas, Zhu et al. reported that ASM regulates both the expression and activation of MET, thereby potentially driving the malignant progression of glioblastoma, whereas inactivation of ASM markedly reduces MET receptor abundance and downstream signaling [51]. Moreover, ASM has been shown to modulate the sensitivity of glioma cells to radiotherapy and chemotherapy [52, 53], further highlighting its role as a critical determinant of therapeutic response. Several reviews have also highlighted ASM as a key node in the sphingolipid rheostat of gliomas. Gliomas are characterized by complex metabolic reprogramming [54], in which the sphingolipid subpathway represents an independently important axis [55]. Therefore, investigating the relationship between ASM and gliomas is of significant scientific and clinical relevance.
Analysis of data from the TCGA and CGGA datasets revealed that SMPD1 expression varied significantly across glioma subtypes, with higher expression observed in tumors exhibiting more malignant phenotypes. These findings were further validated by immunohistochemical staining of clinical glioma specimens. Both univariate and multivariate Cox regression analyses identified SMPD1 as an independent prognostic factor in glioma, with elevated expression associated with poorer OS. Based on the Cox model, a prognostic nomogram was constructed to predict 1‐, 3‐, and 5 ‐ year OS, and its accuracy was validated using calibration curves. Survival analysis of SMPD1 copy number variation further underscored the favorable prognostic impact of ASM deletion in early‐stage glioma (p < 0.05). In vivo, genetic deletion of Smpd1 in C57BL/6J mice significantly suppressed the growth of gliomas, further supporting the potential tumor‐promoting role of SMPD1.
Further analyses suggested that the effect of SMPD1 is closely linked to the modulation of the TME. Gliomas with high SMPD1 expression exhibited increased infiltration of Tregs, a key immunosuppressive population that inhibits antitumor immune responses and facilitates immune evasion [56]. Additionally, SMPD1 expression was positively correlated with infiltration of M0 and M2 macrophages. M2 macrophages are known to support tumor progression [57], angiogenesis, and immune suppression [58]. Notably, SMPD1 expression was positively associated with the immune checkpoint gene PDCD1 (PD‐1), suggesting a potential role in PD‐1‐mediated immune regulation. SMPD1 also showed correlations with multiple macrophage markers, including CD163, CD204, TGF‐β, CD40, CD16, CD32, CD64, CD68, IRF5, and CCL2. These markers predominantly represent pan‐macrophage and M2‐associated macrophage signatures, indicating that SMPD1 may influence macrophage function and tumor‐associated macrophage (TAM) polarization. Consistently, our cell‐based experiments provide additional support for this notion: genetic loss of ASM or pharmacological inhibition of its activity blunted M2 polarization of macrophages.
Correlation analyses with immune checkpoint genes revealed that SMPD1 was positively associated with several immunosuppressive checkpoint molecules, such as PDCD1LG2, IDO1, SIGLEC7, and SIGLEC9. This supported the notion that SMPD1 is upregulated in immunosuppressive tumor environments and may facilitate immune evasion. Conversely, SMPD1 was negatively correlated with KLRK1, a marker of immune activation [59], implying a potential role in dampening antitumor immunity. SMPD1 was also negatively correlated with VTCN1, an immune regulator gene [60], despite the biological implications of this association remaining unclear. In summary, these findings suggest that SMPD1 contributes to a shift in the TME toward an immunosuppressive state.
In previous studies, overexpression of ASM was shown to suppress immune activity by inducing apoptosis of CD4^+^ T cells [61]. In addition, ASM activation can impair T‐cell migration, polarization, and functional responses [62]. In the context of cancer, deficiency of ASM enhances antitumor immunity mediated by Th1 cells and cytotoxic T lymphocytes, thereby restraining tumor progression [10]. Moreover, studies in infectious disease models have demonstrated that reduced or absent ASM expression leads to exacerbated inflammation, characterized by increased neutrophil infiltration [7], macrophage accumulation, and elevated levels of multiple proinflammatory cytokines [63, 64, 65]. Collectively, these findings along with our results support the notion that ASM may act as a critical regulator of the TME [66].
Functional enrichment analyses further demonstrated that SMPD1 is involved in multiple immune and inflammatory signaling pathways, including innate immune response, Toll‐like receptor signaling, MAPK pathway, NF‐κB signaling, and NOD‐like receptor signaling, as well as cytokine‐related pathways involving IL‐1β and IL‐17, suggesting a role in immune regulation.
Copy number and mutation analyses revealed that SMPD1 deletions occurred more frequently in glioma and were associated with improved early survival. In LGG, patients with high SMPD1 expression had a higher frequency of TP53 mutations, a tumor suppressor gene whose loss contributes to malignant transformation [67]. In glioblastoma, high SMPD1 expression was significantly associated with PTEN mutations, which are linked to poor clinical outcomes [68, 69]. PTEN also plays an essential role in TME immunosuppression of glioblastoma [68]. These findings further support the association between high SMPD1 expression and adverse prognosis.
Translational and Therapeutic Implications
4.2
Our findings have potential therapeutic implications. Several pharmacological inhibitors of ASM, including fluoxetine, imipramine, and amitriptyline, are already clinically available and widely prescribed for psychiatric disorders [70]. These agents possess favorable blood–brain barrier permeability and well‐established pharmacokinetic and safety profiles, which substantially lower the threshold for clinical repurposing in glioma patients. Notably, previous studies have demonstrated that fluoxetine can effectively suppress glioma growth [71], supporting the translational relevance of ASM inhibition. These findings suggest that ASM represents a translationally relevant therapeutic target, with potential to improve patient outcomes.
Consistent with our observations and previous studies in gliomas, enhanced ASM activity has also been implicated in other malignancies. For instance, in non‐small cell lung carcinoma (NSCLC), elevated ASM activity was detected in both patient serum and tumor tissues, accompanied by an increased ceramide‐to‐sphingomyelin ratio for specific molecular species such as C18, C20, and C24 [10]. Importantly, these metabolic alterations were linked to immune evasion and tumor progression, further underscoring the immunomodulatory role of ASM. Together with our findings, this suggests that ASM‐driven ceramide metabolism may represent a common mechanism of tumor‐induced immune suppression across different cancer types. Such cross‐cancer evidence provides a strong rationale for targeting SMPD1/ASM not only as a biomarker but also as a therapeutic strategy. In particular, the association between SMPD1 expression and immune checkpoint pathways in gliomas raises the possibility that ASM inhibition could potentiate the efficacy of PD‐1/PD‐L1 blockade and other immunotherapies.
Given the involvement of SMPD1 in multiple immune and inflammatory pathways, future investigations may explore the integration of ASM inhibitors with existing treatment modalities such as radiotherapy or temozolomide. Preclinical evaluation using glioma organoids or orthotopic immunocompetent mouse models will be instrumental in assessing efficacy and safety, ultimately paving the way for early‐phase clinical trials.
Conclusion and Future Perspectives
4.3
In conclusion, this study suggests that SMPD1 is upregulated in malignant gliomas, may contribute to tumor progression, and correlates with poor prognosis, indicating its potential as a prognostic biomarker. SMPD1 appears to reshape the tumor microenvironment toward an immunosuppressive phenotype, potentially through the modulation of T cells and macrophages, which suggests a possible relevance in the context of glioma immunotherapy. Importantly, the prognostic and immunological relevance of SMPD1 was consistently observed across multiple independent levels of evidence. These included transcriptomic datasets (TCGA and CGGA), clinical glioma specimens, and functional validation in vivo. The consistency across multiple levels of evidence supports the robustness of our observations and highlights the potential translational relevance of SMPD1 as a therapeutic target. Although our study provides evidence linking SMPD1 to immunosuppressive microenvironmental remodeling, our analyses were based on bulk RNA‐seq datasets, and in vivo validation was performed in a subcutaneous glioma model, which may not fully recapitulate the intracranial tumor microenvironment. In addition, the number of clinical samples was relatively small (n = 34) and obtained from a single center. Such a limited sample size may reduce the statistical power and restrict the generalizability of our findings. Therefore, future work employing single‐cell transcriptomics, flow cytometry, and orthotopic glioma models will be necessary to dissect the mechanistic role of SMPD1 in shaping immune suppression. These limitations will be addressed in future investigations. What's more, larger multi‐center cohorts will be essential to validate the robustness and external applicability of our conclusions.
Funding
This work was supported by the Young Elite Scientist Sponsorship Program by BAST, BYESS2023394.
Ethics Statement
This study was approved by the Ethics Review Committee of Beijing Tiantan Hospital (No. KY2024‐345). All animal experiments were approved by the Institutional Animal Care and Use Committee of the Beijing Neurosurgical Institute (No. BNI202503006).
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Figure S1: Kaplan–Meier curves for overall survival in high‐grade gliomas stratified by high vs. low ASM expression in TCGA (A, B) and CGGA (C, D) datasets. Figure S2: Evaluation and validation of the prognostic models. Brier score assessment of the nomogram model constructed using the TCGA dataset (A). Brier score assessment of the nomogram model constructed using the CGGA dataset (B). Calibration curves validating the TCGA‐derived prognostic model in the CGGA cohort (C). Calibration curves validating the CGGA‐derived prognostic model in the TCGA cohort (D). Brier score evaluation of the TCGA‐derived model in the CGGA validation cohort (E). Brier score evaluation of the CGGA‐derived model in the TCGA validation cohort (F). Figure S3: Heat map of SMPD1 expression and expression of immune cell markers in the TCGA dataset. Figure S4: Heat map of SMPD1 expression and expression of immune cell markers in the CGGA dataset. Figure S5: Kaplan–Meier curves for overall survival in patients with mutant and wild‐type SMPD1. Table S1: Primer sequences for real‐time PCR.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1S. Gritsch , T. T. Batchelor , and L. N. Gonzalez Castro , “Diagnostic, Therapeutic, and Prognostic Implications of the 2021 World Health Organization Classification of Tumors of the Central Nervous System,” Cancer 128, no. 1 (2022): 47–58.34633681 10.1002/cncr.33918 · doi ↗ · pubmed ↗
- 2M. J. van den Bent , M. Geurts , P. J. French , et al., “Primary Brain Tumours in Adults,” Lancet 402, no. 10412 (2023): 1564–1579.37738997 10.1016/S 0140-6736(23)01054-1 · doi ↗ · pubmed ↗
- 3N. A. Mohile , H. Messersmith , N. T. Gatson , et al., “Therapy for Diffuse Astrocytic and Oligodendroglial Tumors in Adults: ASCO‐SNO Guideline,” Journal of Clinical Oncology 40, no. 4 (2022): 403–426.34898238 10.1200/JCO.21.02036 · doi ↗ · pubmed ↗
- 4D. S. Ettinger , D. E. Wood , D. L. Aisner , et al., “Non‐Small Cell Lung Cancer, Version 3.2022, NCCN Clinical Practice Guidelines in Oncology,” Journal of the National Comprehensive Cancer Network: JNCCN 20, no. 5 (2022): 497–530.35545176 10.6004/jnccn.2022.0025 · doi ↗ · pubmed ↗
- 5A. C. Huang and R. Zappasodi , “A Decade of Checkpoint Blockade Immunotherapy in Melanoma: Understanding the Molecular Basis for Immune Sensitivity and Resistance,” Nature Immunology 23, no. 5 (2022): 660–670.35241833 10.1038/s 41590-022-01141-1PMC 9106900 · doi ↗ · pubmed ↗
- 6M. Lim , Y. Xia , C. Bettegowda , and M. Weller , “Current State of Immunotherapy for Glioblastoma,” Nature Reviews. Clinical Oncology 15, no. 7 (2018): 422–442.10.1038/s 41571-018-0003-529643471 · doi ↗ · pubmed ↗
- 7C. Pinto , D. Sousa , V. Ghilas , A. Dardis , M. Scarpa , and M. Macedo , “Acid Sphingomyelinase Deficiency: A Clinical and Immunological Perspective,” International Journal of Molecular Sciences 22, no. 23 (2021): 12870.34884674 10.3390/ijms 222312870 PMC 8657623 · doi ↗ · pubmed ↗
- 8E. Assi , D. Cervia , L. Bizzozero , et al., “Modulation of Acid Sphingomyelinase in Melanoma Reprogrammes the Tumour Immune Microenvironment,” Mediators of Inflammation 2015 (2015): 370482.26101462 10.1155/2015/370482 PMC 4460251 · doi ↗ · pubmed ↗