Exploring the lactate-metabolism related characteristics during the development of medulloblastoma through single-cell and bulk RNA-seq

Introduction

Medulloblastoma (MB) is a common intracranial malignant tumor in children, accounting for about 20% of childhood brain tumors. It occurs mainly in the posterior cranial fossa, but can also spread to the entire brain and spinal cord through the cerebrospinal fluid (1). MB has the potential to occur at any age, with an overall annual incidence of approximately 5/1,000,000 in the pediatric population, and with an incidence of approximately 1.7 times higher in males than in females. MB has a genetic component and is positively associated with certain neuro-oncological syndromes, including Gorlin’s syndrome, Turcot’s syndrome, and Curry-Jones’ syndrome, among others (2). According to the World Health Organization, MB is classified into three types according to histopathology: classic, large cell/mesenchymal, nodular, and connective tissue-promoting proliferative. The 5-year overall survival rate for patients with large cell/mesenchymal MB is 62%, while it is up to 100% for patients with nodular and connective tissue-promoting proliferative MB (3). Furthermore, with the advent of genomics, MB has been classified into wingless (WNT), sonic hedgehog (SHH), group 3 (G3), and G4 types, each of which exhibits unique molecular and clinical characteristics (4).

Tumor metabolism represents a pivotal connection between the tumor microenvironment and its progression. A heightened risk of malignant tumor development has been observed to be closely associated with metabolic abnormalities, and malignant tumors have been shown to undergo metabolic reprogramming to promote malignant proliferation and adaptation to adverse survival conditions (5). A considerable body of research has documented substantial reprogramming in the metabolic processes of MB cells, exhibiting a diverse spectrum among MB subtypes (6,7). In particular, Park et al. recently recognized that several metabolic pathways can serve as relevant prognostic markers for MB subpopulations, and thus each subpopulation has a specific metabolic profile (8). However, our understanding of how different lactate-metabolic pathways interact or contribute to the aforementioned subgroup-specific features remains limited and is frustrated by the involvement of complex molecular mechanisms and the coordinated action of several signaling molecules.

MB requires metabolic reprogramming in order to survive and proliferate under harsh and nutritionally restricted conditions (9). In such conditions, mitochondrial function is known to decrease, whilst glycolytic activity increases. This phenomenon, termed the Warburg effect (10), confers a survival advantage upon cancer cells by reducing reactive oxygen species (ROS) levels, thereby mitigating the deleterious effects of oxidative stress and promoting the biosynthesis of metabolic precursors necessary for nucleic acid, amino acid, and lipid synthesis, which in turn promotes uncontrolled proliferation (11). During aerobic glycolysis, lactate dehydrogenase (LDH) converts pyruvate to lactate and produces nicotinamide adenine dinucleotide (NAD+), which is required to maintain the glycolytic pathway and adenosine triphosphate (ATP) production. The final step in this process is the export of lactate to the exterior of the cell via the monocarboxylate transporter protein 1 (MCT1), which facilitates the import/export of lactate, pyruvate and ketone bodies throughout the plasma membrane (12). The MB 3/4 group exhibited elevated levels of lactate and enhanced expression of lactate dehydrogenase A (LDHA) and MCT1, indicating that MB is sustained by a glycolytic phenotype. Inhibition of LDHA with oxalate significantly inhibited MB lactate production, aerobic glycolysis, proliferation and motility (12). Consequently, an exhaustive investigation into the lactate-metabolic characteristics of MB and its prognostic imaging can furnish a substantial foundation for the development of personalized therapeutic regimens and unveil novel pathways to enhance the survival and quality of life of patients. Research in this field not only reveals the biological characteristics of MB but also provides new targets and ideas for future therapeutic strategies. We present this article in accordance with the TRIPOD reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1008/rc).

Methods

Data collection

Single-cell RNA sequencing (scRNA-seq) data (GSE155446) was obtained from the Gene Expression Omnibus (GEO) database containing 4 tumor subtypes of MB patients, and we removed recurrence samples and mouse samples. Utilization of the R package “Seurat” was undertaken to conduct additional analysis. The criteria for quality control of cells encompassed a mitochondrial content threshold of under 30% and established limits for unique molecular identifier (UMI) counts, with gene counts delineated between the ranges of 200–40,000 and 200–8,000, respectively. In the end, 26,040 genes and 33,637 cells were retained for subsequent analysis. The ensuing data processing stage involved the application of scaling and normalization techniques to the remaining cells, employing linear models and regression models via the “NormalizeData” function. Post-normalization, the identification of the top 1,500 variable genes was executed using the “FindVariableFeatures” function. The R package “harmony” was implemented to address potential batch effects across samples, while normalization of the data was executed using the “ScaleData” function. The “FindNeighbors” function was utilized to establish cell-similarity relationships, leveraging the dimensions of principal component analysis (PCA). For the purpose of refined clustering, the resolution parameter was set to 0.3. To further select core lactate-metabolism related genes, we obtained bulk RNA dataset (GSE85217) for machine learning purposes, to shrink amount of candidate genes, further strengthening the reliability of the diagnostic model developed in this study. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Clusters annotation analysis

Initially, we identified markers for three principal cell types: MB cells [“epithelial cell adhesion molecule (EPCAM)”, “Keratin 18 (KRT18)”, “Keratin 19 (KRT19)”, “Cadherin 1 (CDH1)”]; lymphatic cells [“CD3 Delta Subunit Of T-Cell Receptor Complex (CD3D)”, “CD3-epsilon (CD3E)”, “CD3-gamma (CD3G)”, “T Cell Receptor Alpha Constant (TRAC)”, “CD79a Molecule (CD79A)”, “Immunoglobulin Heavy Constant Mu (IGHM)”, “Immunoglobulin Heavy Constant Gamma 3 (IGHG3)”, “(Immunoglobulin Heavy Constant Alpha 2 (IGHA2)”]; myeloid cells (macrophages: “CD14”, “CD86”, “CD16”, “Ficolin 1 (FCN1)”; monocytes: “S100A8 (S100 Calcium Binding Protein A8)”, “S100A9 (S100 Calcium Binding Protein A9)”). UMAP (Uniform Manifold Approximation and Projection) plots and dot plots were used to visualize.

Metabolic score analysis

The R package “gene set variation analysis” was used to assess metabolic characteristics of various cell types within the scRNA-seq data. This analysis referred to the GSVA scoring principle to evaluate 85 metabolic pathways, 8 lactate metabolism pathways and Hallmark pathways. Pathways related to glucose metabolism, lipid metabolism and lactate metabolism pathways were visualized using R package “pheatmap”. After removing duplicates of lactate-metabolism related genes, 273 lactate-metabolism related genes (LMRGs) were identified to form the lactate-metabolism gene set. To calculate the lactate-metabolism score, we applied the function “AddModuleScore”, which quantifies the activity of a gene set in individual cells. This method allows us to assess the relative expression of the lactate-metabolism gene set within each cell, resulting in a lactate-metabolism score that reflects the activity of lactate-metabolism related pathways in individual cells. The resulting lactate-metabolism scores were visualized using violin plots and UMAP plots, providing a clear comparison of lactate-metabolism activity across various cell types.

Eight gene sets pertinent to lactate metabolism were identified: “GOBP_LACTATE_METABOLIC_PROCESS”, “GOBP_LACTATE_TRANSMEMBRANE_TRANSPORT”, “GOMF_LACTATE_DEHYDROGENASE_ACTIVITY”, “GOMF_LACTATE_TRANSMEMBRANE_TRANSPORTER_ACTIVITY”, “HP_ABNORMAL_BRAIN_LACTATE_LEVEL_BY_MRS”, “HP_ABNORMAL_LACTATE_DEHYDROGENASE_LEVEL”, “HP_INCREASED_CSF_LACTATE”,”HP_INCREASED_SERUM_LACTATE”, were obtained from the Molecular Signatures Database (MSigDB, version 7.5.1), as referenced in a preceding study (13).

Trajectories analysis

Cell lineage trajectories were inferred using “Monocle2” (version 2.10.0), thus facilitating the identification of cell transitions. Additionally, genes displaying significant changes over the pseudotime were identified.

Differences among MB clusters

Differential expressed genes and pathway enrichment analysis were conducted using the “FindMarker” function and the R package “clusterProfiler” for Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology Biological Process (GOBP) enrichment analyses of MB clusters, visualized using R package “ggplot2”.

Machine learning

To identify core LMRGs associated with MB and construct a diagnostic model, we employed machine learning algorithms for screening and modeling. Random Forest was selected based on its strengths in handling high-dimensional gene expression data and diverse model requirements. Diagnostic performance was evaluated using average area under the curve (AUC) values from the training and validation sets, ensuring we selected the most predictive and reliable model for MB diagnosis.

Statistical analysis

All analyses in this study were conducted using the R software. Single-cell RNA-seq data were processed with Seurat and SCP for normalization, clustering, and differential expression analysis. Gene set variation analysis, trajectory inference, and pathway enrichment analysis were performed using GSVA, Monocle2, and clusterProfiler, respectively. For bulk RNA-seq data, a Random Forest model was applied to identify core gene sets, and model performance was evaluated using the AUC. Statistical comparisons were conducted using the t-test or Wilcoxon test, with adjusted P<0.05 considered statistically significant.

Results

Single-cell transcriptomic landscape of MB

First, according to methodological criteria, we performed quality control and integration of single-cell data, retaining 33,637 cells for downstream analysis. We selected default principal components for dimensional reduction clustering, resulting in 12 clusters labeled from 0 to 11 (Figure 1A). Due to Figure 1B,1C and Figure S1, we found MBCluster_7 and MBCluster_9 were specific to WNT subtype of MB, MBCluster_5 was specific to SHH subtype, while MBCluster_6 was specific to GP3/4 subtype. The various ratio of the clusters among different subtypes suggesting the existence of heterogeneity during tumor development as well as similarities. Then we annotated into 3 main cell types according to top3 markers resulted from FindAllMarkers function (Figure 1D). MBCluster_4 was specific to proliferating cells [Marker: DNA topoisomerase II alpha (TOP2A)/marker of proliferation Ki-67 (MKI67)].

Figure 1 Identification of cell clusters in MB single-cell data. (A) UMAP plot displaying 12 clusters (0–11) identified from the single-cell RNA sequencing data. Clusters were annotated into three main cell types based on literature. (B) UMAP plot displaying clusters under 4 subtypes of MB. (C) Bar plot depicting the ratio of cell types by each subtype. (D) Dot plot showing the top three markers with highly expression. MB, medulloblastoma; SHH, sonic hedgehog; UMAP, Uniform Manifold Approximation and Projection; WNT, wingless.

Metabolic characteristic in MB clusters

We focused on lactate metabolism process, then we used AddModuleScore function to calculate LMGScores for each cell cluster, which showing the high score distributed in MB clusters (Figure 2A,2B; Figure S2). This implies that the lactate metabolism process would participate in extensive signaling network and contributing to the onset of MB. Many studies indicated that metabolic processes played a role in tumor development, and later we evaluated the metabolic changes in various cell types, with tricarboxylic acid (TCA) cycle, oxidative phosphorylation, glycolysis/gluconeogenesis, fatty acid biosynthesis, nitrogen metabolism and steroid hormone biosynthesis with high GSVA score level, indicating there exist a strong correlation between metabolism process and MB development (Figure 2C). We also assessed other tumor related signaling pathway activity scores, as shown in Figure S1.

Figure 2 Identification of cell clusters in MB single-cell data. (A) Violin plot comparing lactate metabolism scores across different cell types, calculated using the AddModuleScore function. MB clusters show the high level of LMRGs scores. (B) UMAP plot showing lactate metabolism scores across different cell types. The higher the scores, the yellower the color. (C) Heatmap showing the expression of key metabolic pathways, including TCA cycle, oxidative phosphorylation, glycolysis/gluconeogenesis, fatty acid biosynthesis, nitrogen metabolism and steroid hormone biosynthesis. (D) Development difference of MB clusters displayed with monocle2. (E) Heatmap depicting the relative expression of all MB clusters markers. Figure (F) corresponds to the red and blue branches. (F) A heatmap shows the pseudotime dynamics of the gene expression patterns in MB clusters. LMRGs, lactate-metabolism related genes; LMGs1, lactate metabolism gene signature score; MB, medulloblastoma; SHH, sonic hedgehog; TCA, tricarboxylic acid; UMAP, Uniform Manifold Approximation and Projection.

To further investigate the difference among MB 4 subtypes, we conduct trajectory analysis with Monocle2, performed to reconstruct and characterize the relationships among MB clusters in the different subtypes and to derive reprogramming trajectories using an unbiased method. According to the change in trajectory, the MB clusters differed toward the WNT/SHH subtype from the Group3/Group4 (Figure 2D,2E). We then examined the pseudo-time dynamics of the gene expression patterns in MB clusters and arranged them into five clusters (Figure 2F). We checked LMRGs scores in the pseudo-time dynamics in Figure S2.

The expression feature of LMRGs in MB clusters

After obtaining the aforementioned profiles, we grouped these cells with the median LMRGs score. Then we performed differential gene expression analysis between different lactate metabolic score groups (Figure 3A; table available at https://cdn.amegroups.cn/static/public/tcr-2025-1008-1.xlsx) and pathway enrichment analysis of up-/down-regulated genes in each cluster. Gene Ontology (GO) characteristics about metabolism were detected in all MB clusters (0–9), while cluster 2 was enriched in lactate metabolism. Each MB cluster was distinguished by biological process: MBCluster_1 exhibited differential expression of transthyretin 1 (TTR1) and NDUFA4 mitochondrial complex associated like 2 (NDUFA41) (Figure 3A), which is related to metabolic regulation and energy metabolism. MBCluster_3 was enriched in chromosome segregation and organelle fission, and proliferating MBCluster_4 was enriched in ribonucleoprotein complex biogenesis and mitochondrial gene expression, while the same MBCluster_9 was. Up-regulated difference expression genes (DEGs) from MBCluster_5 specific to SHH subtype was enriched in regulation of neurogenesis and epithelial cell proliferation, which considered to contribute to MB development. MBCluster_6 and MBCluster_7 were mainly enriched in catabolic process. According to the down-regulated DEGs’ pathway enrichment analysis, we found that they were similar to the cell cycle phase and cell development (Figure 3B,3C).

Figure 3 LMRGs expression in MB single-cell data. (A) Differential gene expression analysis showing up- and down-regulated genes across all MB clusters. An adjusted P<0.01 is indicated in red, while an adjusted P≥0.01 is indicated in black. (B) Dotplot showing the pathway enrichment analysis of up-regulated DEGs. (C) Dotplot showing the pathway enrichment analysis of down-regulated DEGs. DEGs, difference expression genes; FC, fold change; LMRGs, lactate-metabolism related genes; MB, medulloblastoma.

Identification of candidate LMRGs in MB clusters

We identified 68 overlapping genes through single-cell transcriptomic differential analysis and LMRGs. Pathway analysis of these 68 genes revealed significant enrichment in metabolic and electron transport-related pathways (Figure 4A). To further identify core LMRGs, we employed a random forest machine learning model to screen for key lactate metabolism-promoting genes and construct a diagnostic model for MB. The results demonstrated that after variable selection using machine learning methods and interpretation via Shapley Additive Explanations (SHAP), we ultimately retained 10 core LMRGs for MB (Figure 4B-4D).

Figure 4 Identification of candidate LMRGs in MB clusters. (A) Intersection between differential expressed genes and LMRGs. (B) The pathway analysis of these intersection genes from STRING database. (C) The OOB errors from Random Forest machine learning. (D) The top 10 candidate LMRGs selected by importance. DEGs, difference expression genes; FDR, false discovery rate; LMRGs, lactate-metabolism related genes; MB, medulloblastoma; OOB, out-of-bag; SHH, sonic hedgehog.

Protein-protein interaction (PPI) network of candidate LMRGs in MB clusters

After obtaining the above 10 candidate LMRGs, we investigated their expression profiles in the transcriptome. To further understand the expression patterns of these 10 genes in MB, we utilized the STRING database to predict the PPI network of these 10 genes and performed pathway analysis (Figure 5A,5B). The results showed that the 10 candidate genes were primarily involved in pathways, related to metabolic pathways, such as Mitochondrial metabolism disease, Thermogenesis, mRNA binding, Oxidoreduction-driven active transmembrane transporter activity, Electron transfer activity, Oxidative phosphorylation, Inner mitochondrial membrane protein, Parkinson disease. Figure 5C showed the expression level of candidate LMRGs in bulk data set.

Figure 5 PPI network of candidate LMRGs. (A) The network from STRING database. (B) The pathway analysis of these LMRGs from STRING database. (C) Violin plot showing expression level of top 10 LMRGs. FDR, false discovery rate; LMRGs, lactate-metabolism related genes; PPI, protein-protein interaction; SHH, sonic hedgehog.

Discussion

MB is the most prevalent malignant brain tumour in children, with an increasing incidence (14). According to the clinical features of MB, as well as genomics, transcriptomics, methylation levels, SHH-activated/TP53 (cellular tumor antigen p53) wild-type, SHH-activated/TP53-mutated, and non-WNT/non-SHH-activated types (15), the WNT subtype of MB is characterised by the activation of the Wnt pathway, and has the best prognosis. This is evidenced by a 5-year survival rate of approximately 90% (16). The SHH subtype, on the other hand, is distinguished by the activation of the SHH pathway and is associated with an intermediate prognosis. However, patients with TP53 mutations exhibit a poor prognosis. The third and fourth subtypes have not yet been identified as major signaling pathways. In order to address the heterogeneity observed within each MB subtype, several studies have proposed the subdivision of the four subtypes into 12 subgroups (17). Normal metabolic tissues are instructed by growth factor signaling to convert food into nearly constant amounts of nutrients and energy for orderly cell proliferation and growth (18). However, studies on the role of the body’s own metabolite, lactate, and its subsequent lactoylation modifications on the immunomodulation of tumour cells are still scarce. The present study constitutes a systematic analysis of scRNA-seq data, which has revealed differences in the level of lactate metabolism in tumour cells according to their subtypes. Furthermore, it has identified and screened 10 candidate genes related to lactate metabolism.

The integration of single-cell transcriptome sequencing data has facilitated the construction of a comprehensive landscape of the multicellular ecosystem of MB, unveiling the metabolic heterogeneity of tumour cells from multiple perspectives. Firstly, the study characterised tumour cell subpopulations based on their functional features and evolutionary trajectory. Secondly, it revealed, for the first time, the distinct lactate metabolism characteristics of different MB subtypes at the single-cell level. Thirdly, the study screened MB subpopulations under various subtypes using machine learning. Finally, it identified lactate metabolism-related genes using machine learning. Metabolic reprogramming has been identified as a critical factor in the development of tumors. It has been demonstrated that this process involves genetic events that lead to the transformation of normal cells into cancerous cells. Specifically, there is an activation of genes that promote tumour growth and a subsequent inactivation of genes that suppress tumour growth. Examples of such genes include Kirsten rat sarcoma viral oncogene (Kras), Myelocytomatosis viral oncogene homolog (C-Myc), and p53. Such alterations in metabolic patterns can lead to metabolic reprogramming of cells, thereby providing the driving force for tumour growth. This metabolic reprogramming is achieved by activating the expression of key enzymes involved in metabolic pathways, such as glycolysis and glutamine metabolism, as well as by regulating intracellular signalling pathways, including extracellular signal-regulated kinase (ERK)-mitogen-activated protein kinases (MAPK), and affecting the function of mitochondria. The activation of key enzymes of glycolysis and glutamine metabolism, the regulation of intracellular signalling pathways such as ERK-MAPK, and the effect on mitochondrial function can alter cellular metabolic patterns and induce metabolic reprogramming, thus providing the driving force for tumour growth (19,20). Secondly, there is the improvement of cellular adaptability and remodelling of the tumour microenvironment: tumour cells are in a low-oxygen, acidic and nutrient-deficient environment (21). The metabolic reprogramming of active glycolysis, lipid metabolism and other alterations can generate intermediate metabolites. A proportion of these enter the biosynthetic pathway and contribute to the development of the tumour. Some of these intermediates enter the biosynthetic pathway and provide precursors and reducing equivalents for the synthesis of amino acids and other substances (22). The remaining metabolites, such as lactate, accumulate and are secreted outside the cell, thereby creating a dynamic microbial network that is hypoxic, acidic, and immunosuppressive. Finally, it has been demonstrated that metabolic reprogramming affects the invasion-metastasis cascade: intermediates of metabolic reprogramming, such as lactic acid, have been shown to remodel the extracellular matrix, induce the activation of transforming growth factor β, increase the expression of Snail protein, and promote the epithelial-mesenchymal transition, thus enhancing the invasive and metastatic ability of tumours (23). Mouse models have demonstrated that cerebellar granule neural precursors (CGNPs) are stimulated by Shh when mice undergo aerobic glycolysis, and Shh and phosphoinositide 3-kinase signalling stimulate aerobic glycolysis in CGNPs in a hexokinase-2 (HK2)-dependent manner (12). Moreover, the suppression of HK2 gene expression in MB mouse models has been shown to result in the inhibition of aerobic glycolytic activity, thereby promoting cell differentiation and suppressing proliferation, thus preventing the malignant proliferation of MB cells. Shh MBs have also been observed to exhibit elevated lipid synthesis from aerobic glycolysis (24). The use of nuclear magnetic resonance spectroscopy and 18-fluorodeoxyglucose positron emission tomography allows for the identification of MBs in patients exhibiting a glycolytic metabolic phenotype. A recent study has shown that elevated lactate levels are associated with group 3/4 subpopulations (25). MB cells have elevated expression of monocarboxylic acid transporter 1 (MCT1), which is involved in the process of minolactylation by transporting lactic acid and pyruvic acid through the plasma membrane (26). In addition, MCT1 is repressed by miR-124, which is normally downregulated in MB. Ectopic expression of miR-124 has been shown to inhibit MB proliferation by blocking cell cycle progression in G1. Cellular metabolism can be affected by many pathways, including several implicated in MB. The Wnt signalling pathway has been shown to increase lactation and decrease oxidative phosphorylation (27). c-Myc is overexpressed in the three MB group and activates the downstream of the Wnt signalling pathway, which is known to promote lactation in many cancers (28,29) and to upregulate LDHA expression (30). Despite the fact that MB consists of four distinct subgroups, these studies suggest that lactic acidification is a common feature of most MB.

In addition, metabolic reprogramming also supports the distal colonisation of tumour cells after metastasis. For instance, the up-regulation of aldolase B during liver colonisation by colorectal cancer cells enhances glycolysis, gluconeogenesis and pentose phosphate pathways, thereby providing fuel for tumour cell proliferation and promoting post-metastatic growth (31). Consequently, the homeostatic imbalance of material and energy caused by metabolic reprogramming exerts a significant influence on the biological behaviours of tumour proliferation, growth, invasion and metastasis. Consequently, targeting metabolic reprogramming has become an important strategy for tumour intervention. The presence of a small number of cancer stem cells (CSC) within the tumour has been identified as a significant contributing factor to the development, progression and treatment resistance of several tumour types, including MB (32). The intratumoural and intertumoural heterogeneity exhibited by MB may then be sustained by specific numbers and types of tumour-driven CSCs (16,33), and may also result in different metabolic demands, with (cancer) stem cell function dependent on different metabolic adjustments (34,35). Consequently, in recent years, there has been a surge of interest in the metabolic behaviour of cancer cells, due to its potential to lead to therapeutic resistance or to identify novel targets for overcoming this resistance. Furthermore, the lactate-metabolic adaptations that tumor cells undergo during tumor growth are now considered to be the hallmark of several types of cancers (36).

Conclusions

In summary, our study based on the comprehensive single-cell and bulk transcriptomic analyses presented, this study yields significant insights with direct implications for improving the clinical diagnosis and therapeutic targeting of MB. First, the identification of distinct lactate-metabolism heterogeneity across MB molecular subgroups (WNT, SHH, G3, G4) at single-cell resolution provides a metabolic stratification framework that complements existing genomic classifications. This metabolic heterogeneity correlates with tumor aggressiveness and prognosis, offering a novel dimension for risk assessment. Second, the machine learning-driven identification of ten core lactate-metabolism-related genes (LMRGs): COX4I1, CYC1, MECP2, MYC, NDUFAF3, NDUFS3, COX20, CALR, POMT1, and C1QBP, establishes a robust diagnostic gene signature. These genes, validated in bulk datasets, serve as potential non-invasive biomarkers for early detection and molecular subtyping, enhancing diagnostic precision beyond histopathology. Third, the elevated LMRG scores in specific MB clusters (e.g., G3) and their association with key pathways (e.g., oxidative phosphorylation, glycolysis) highlight LDHA and MCT1 as actionable therapeutic targets. Fourth, the discovery that MYC-driven Group 3 MB exhibits unique metabolic dependencies (e.g., LDHA overexpression) underscores the potential for subgroup-specific metabolic therapies. Finally, the integration of metabolic profiling with MRI-based techniques (e.g., MRS lactate detection) proposes a multimodal diagnostic approach, enabling real-time monitoring of treatment response and metabolic adaptation in tumors. Collectively, these findings advance the paradigm of metabolically informed precision oncology, where targeting lactate-metabolism heterogeneity and its molecular drivers offers novel avenues for improving survival in high-risk MB patients. Future clinical trials should validate LMRG-based diagnostics and evaluate metabolic inhibitors in stratified cohorts.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1008/rc

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1008/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-1008/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. The 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. Michaelsen GL, de Sousa Monteiro T, Imparato DO, et al. Medulloblastoma’s master regulators and their association with patients’ risk. Sci Rep 2025;15:16310. [Crossref] [PubMed]
2. Juraschka K, Taylor MD. Medulloblastoma in the age of molecular subgroups: a review. J Neurosurg Pediatr 2019;24:353-63. [Crossref] [PubMed]
3. Mynarek M, von Hoff K, Pietsch T, et al. Nonmetastatic Medulloblastoma of Early Childhood: Results From the Prospective Clinical Trial HIT-2000 and An Extended Validation Cohort. J Clin Oncol 2020;38:2028-40. [Crossref] [PubMed]
4. 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]
5. Faubert B, Solmonson A, DeBerardinis RJ. Metabolic reprogramming and cancer progression. Science 2020;368:eaaw5473. [Crossref] [PubMed]
6. Bonifacio-Mundaca J, Casavilca-Zambrano S, Desterke C, et al. Deciphering Medulloblastoma: Epigenetic and Metabolic Changes Driving Tumorigenesis and Treatment Outcomes. Biomedicines 2025;13:1898. [Crossref] [PubMed]
7. Ni H, Reitman ZJ, Zou W, et al. FLASH radiation reprograms lipid metabolism and macrophage immunity and sensitizes medulloblastoma to CAR-T cell therapy. Nat Cancer 2025;6:460-73. [Crossref] [PubMed]
8. Park AK, Lee JY, Cheong H, et al. Subgroup-specific prognostic signaling and metabolic pathways in pediatric medulloblastoma. BMC Cancer 2019;19:571. [Crossref] [PubMed]
9. Tech K, Gershon TR. Energy metabolism in neurodevelopment and medulloblastoma. Transl Pediatr 2015;4:12-9. [Crossref] [PubMed]
10. Wang Y, Patti GJ. The Warburg effect: a signature of mitochondrial overload. Trends Cell Biol 2023;33:1014-20. [Crossref] [PubMed]
11. Pavlova NN, Zhu J, Thompson CB. The hallmarks of cancer metabolism: Still emerging. Cell Metab 2022;34:355-77. [Crossref] [PubMed]
12. Valvona CJ, Fillmore HL. Oxamate, but Not Selective Targeting of LDH-A, Inhibits Medulloblastoma Cell Glycolysis, Growth and Motility. Brain Sci 2018;8:56. [Crossref] [PubMed]
13. Huang T, Lian D, Chen M, et al. Prognostic value of a lactate metabolism gene signature in lung adenocarcinoma and its associations with immune checkpoint blockade therapy response. Medicine (Baltimore) 2024;103:e39371. [Crossref] [PubMed]
14. Ostrom QT, Price M, Neff C, et al. CBTRUS Statistical Report: Primary Brain and Other Central Nervous System Tumors Diagnosed in the United States in 2016-2020. Neuro Oncol 2023;25:iv1-iv99. [Crossref] [PubMed]
15. Louis DN, Perry A, Wesseling P, et al. The 2021 WHO Classification of Tumors of the Central Nervous System: a summary. Neuro Oncol 2021;23:1231-51. [Crossref] [PubMed]
16. Hovestadt V, Smith KS, Bihannic L, et al. Resolving medulloblastoma cellular architecture by single-cell genomics. Nature 2019;572:74-9. [Crossref] [PubMed]
17. Cavalli FMG, Remke M, Rampasek L, et al. Intertumoral Heterogeneity within Medulloblastoma Subgroups. Cancer Cell 2017;31:737-754.e6. [Crossref] [PubMed]
18. Ždralević M, Brand A, Di Ianni L, et al. Double genetic disruption of lactate dehydrogenases A and B is required to ablate the "Warburg effect" restricting tumor growth to oxidative metabolism. J Biol Chem 2018;293:15947-61. [Crossref] [PubMed]
19. Matassa DS, Agliarulo I, Avolio R, et al. TRAP1 Regulation of Cancer Metabolism: Dual Role as Oncogene or Tumor Suppressor. Genes (Basel) 2018;9:195. [Crossref] [PubMed]
20. Saeed H, Leibowitz BJ, Zhang L, et al. Targeting Myc-driven stress addiction in colorectal cancer. Drug Resist Updat 2023;69:100963. [Crossref] [PubMed]
21. Zhou S, Sun J, Zhu W, et al. Hypoxia studies in non small cell lung cancer: Pathogenesis and clinical implications Oncol Rep 2025;53:29. (Review). [Crossref] [PubMed]
22. Xia H, Huang Z, Xu Y, et al. Reprogramming of central carbon metabolism in hepatocellular carcinoma. Biomed Pharmacother 2022;153:113485. [Crossref] [PubMed]
23. Kang H, Kim H, Lee S, et al. Role of Metabolic Reprogramming in Epithelial-Mesenchymal Transition (EMT). Int J Mol Sci 2019;20:2042. [Crossref] [PubMed]
24. Zeng LH, Tang C, Yao M, et al. Phosphorylation of human glioma-associated oncogene 1 on Ser937 regulates Sonic Hedgehog signaling in medulloblastoma. Nat Commun 2024;15:987. [Crossref] [PubMed]
25. Contenti J, Guo Y, Mazzu A, et al. The mitochondrial NADH shuttle system is a targetable vulnerability for Group 3 medulloblastoma in a hypoxic microenvironment. Cell Death Dis 2023;14:784. [Crossref] [PubMed]
26. Li KK, Pang JC, Ching AK, et al. miR-124 is frequently down-regulated in medulloblastoma and is a negative regulator of SLC16A1. Hum Pathol 2009;40:1234-43. [Crossref] [PubMed]
27. Motahari Z, Lepe JJ, Bautista MR, et al. Preclinical assessment of MAGMAS inhibitor as a potential therapy for pediatric medulloblastoma. PLoS One 2024;19:e0300411. [Crossref] [PubMed]
28. Lhermitte B, Blandin AF, Coca A, et al. Signaling pathway deregulation and molecular alterations across pediatric medulloblastomas. Neurochirurgie 2021;67:39-45. [Crossref] [PubMed]
29. Sursal T, Ronecker JS, Dicpinigaitis AJ, et al. Molecular Stratification of Medulloblastoma: Clinical Outcomes and Therapeutic Interventions. Anticancer Res 2022;42:2225-39. [Crossref] [PubMed]
30. Qin N, Paisana E, Langini M, et al. Intratumoral heterogeneity of MYC drives medulloblastoma metastasis and angiogenesis. Neuro Oncol 2022;24:1509-23. [Crossref] [PubMed]
31. Bian X, Liu R, Meng Y, et al. Lipid metabolism and cancer. J Exp Med 2021;218:e20201606. [Crossref] [PubMed]
32. Dave B, Tailor J. Human stem cell models to unravel brain cancer. BMC Cancer 2024;24:1465. [Crossref] [PubMed]
33. Vladoiu MC, El-Hamamy I, Donovan LK, et al. Childhood cerebellar tumours mirror conserved fetal transcriptional programs. Nature 2019;572:67-73. [Crossref] [PubMed]
34. Marabitti V, Giansanti M, De Mitri F, et al. Pathological implications of metabolic reprogramming and its therapeutic potential in medulloblastoma. Front Cell Dev Biol 2022;10:1007641. [Crossref] [PubMed]
35. Martín-Rubio P, Espiau-Romera P, Royo-García A, et al. Metabolic determinants of stemness in medulloblastoma. World J Stem Cells 2022;14:587-98. [Crossref] [PubMed]
36. Llibre A, Kucuk S, Gope A, et al. Lactate: A key regulator of the immune response. Immunity 2025;58:535-54. [Crossref] [PubMed]