Molecular Mechanisms of Glioblastoma: Insights from the HumanPSD® Database

ByProf. Dr. Alexander Kel Hours Updated on

Introduction

Glioblastoma (GBM) is the archetype of an aggressive, adaptive human cancer: genomically chaotic, spatially heterogeneous, tightly wired to a suppressive brain microenvironment, and—despite decades of study—still lethal for most patients. Under the 2021 WHO CNS classification, the name “glioblastoma” is now reserved for IDH-wildtype diffuse astrocytic tumors with specific molecular features, separating them clearly from IDH-mutant astrocytomas and changing how clinicians and researchers think about the disease. PMC

At the molecular level GBM is defined by a high degree of inter- and intratumoral heterogeneity: co-existing clones with different driver mutations, transcriptomic programs and metabolic states occupy the same lesion and evolve under treatment pressure. Single-cell and spatial sequencing studies have revealed mixed populations of stem-like, proliferative and differentiated tumor cells and an immune and stromal milieu that varies by region—features that together explain rapid resistance and recurrence. PMC+2Cell+2

Maximal safe surgical resection followed by radiotherapy with concurrent and adjuvant temozolomide (the “Stupp” regimen) remains the backbone of initial therapy and produces a modest survival benefit (median overall survival historically ≈14–16 months in randomized trials), with better outcomes in patients whose tumors harbor MGMT promoter methylation. Yet most tumors recur within months, and salvaging recurrent disease remains largely palliative. PMC+1

GBM’s immune landscape is profoundly immunosuppressive: low neoantigen burden, dysfunctional tumor-infiltrating myeloid cells, and anatomical barriers such as the blood–brain barrier blunt both endogenous antitumor immunity and many systemic therapies. These features have contributed to the disappointing performance of several immunotherapies that work well in other cancers. Contemporary reviews emphasize that overcoming GBM will require strategies that reshape the local immune contexture, penetrate anatomic barriers, and address clonal diversity simultaneously. Nature+1

Against this backdrop, the curated disease-mechanism database HumanPSD® offers a unique systems-biology resource: it contains manually curated associations linking genes, proteins, miRNAs, metabolites, drugs and clinical outcomes across over 2,000 human diseases and more than 400,000 disease-biomarker associations. Leveraging such a resource for GBM allows connection of genotype, signalling pathway deregulation, biomarker status and potentially therapeutic vulnerability.

Core Molecular Mechanisms (based on HumanPSD® data)

From the HumanPSD® GBM disease-report, several key mechanistic/correlative biomarker associations emerge. Notably, the following molecules stand out: EGFR, PTEN, TP53, FOXM1, AKT1, and the regulatory microRNAs miR‑21‑5p, miR‑221‑3p / miR‑222‑3p. Each of these is linked in HumanPSD® via literature-extracted curation.

  • EGFR → PI3K/AKT/mTOR activation: EGFR (including the mutant EGFRvIII variant) is often amplified and/or mutated in GBM, leading to downstream activation of the PI3K → AKT1 → mTOR axis, which promotes proliferation, survival signalling and therapy resistance. HumanPSD® classifies EGFR as a mechanistic biomarker in GBM.
  • PTEN loss → metabolic rewiring: PTEN is a negative regulator of PI3K/AKT signalling; its inactivation (via mutation or deletion) leads to unchecked AKT1 activation, altered metabolic activity (enhanced glycolysis, lipogenesis) and contributes to tumour progression. This causal association is captured in HumanPSD®.
  • TP53 mutation → evasion of apoptosis / cell-cycle checkpoints: TP53 is among the most frequently mutated tumour suppressors in GBM, enabling escape from DNA damage-induced apoptosis or senescence, thus facilitating malignant progression. The HumanPSD® report lists TP53 under causal mechanistic biomarkers.
  • FOXM1-driven proliferation and stemness: FOXM1 (a transcription factor) has been shown to regulate cell-cycle progression genes, support glioma stem-cell like phenotypes, and drive invasion. In the HumanPSD® GBM biomarker table, FOXM1 is listed as a correlative/prognostic biomarker linked with poorer survival.
  • miR-21-5p and miR-221/222 family → post-transcriptional oncogenic regulators: These miRNAs are over-expressed in GBM; miR-21 targets PTEN and TP53 among others, while miR-221/222 promotes AKT1 activation and suppresses pro-apoptotic factors. HumanPSD® includes these microRNAs as mechanistic biomarkers (regulators) for glioblastoma.

In sum, HumanPSD® supplies a curated mechanistic scaffold connecting gene aberrations (EGFR, PTEN, TP53), signalling cascades (PI3K/AKT/mTOR, FOXM1 network), and regulatory miRNAs — thereby enabling a coherent picture of tumour progression in GBM.

Understanding Biomarker Types in HumanPSD®

The HumanPSD® biomarker table (as seen in the GBM report header) classifies biomarkers not only by molecular entity but also by type of association and type of indication. Key categories include:

  • Causal biomarkers: molecules directly implicated in disease mechanism (e.g., EGFR, TP53, PTEN).
  • Correlative biomarkers: molecules statistically associated with disease presence or severity (e.g., FOXM1, PKM).
  • Preventive biomarkers: molecules whose modulation may reduce disease risk (though fewer examples in GBM).
  • Negative biomarkers: markers inversely associated with disease progression or incidence.
  • Mechanistic biomarkers: genes/pathways that delineate the biological mechanism of the disease.
  • Prognostic biomarkers: markers of clinical outcome or survival (e.g., MGMT promoter methylation status).
  • Therapeutic target biomarkers: druggable molecules already targeted or potentially actionable in therapy.

Specifically for GBM, the HumanPSD® database currently lists approx. 1,841 biomarkers supported by over 2,600 mechanistic associations and 849 therapeutic target associations (as indicated in the report header). Each biomarker entry includes: gene/protein locus, pathway annotation, tissue/source of expression, clinical/phenotypic context, and known drug interactions if applicable — producing a dense molecular map of disease mechanisms.

Systems-Level Integration

One of the major strengths of HumanPSD® is integration of biomarker relationships into coherent signalling and regulatory networks. This allows the identification of master regulators, feedback loops, and potential vulnerabilities in GBM progression. Furthermore, HumanPSD® contains disease-to-disease maps, which links different disorders according to shared biomarkers and pathways. For example, GBM shares key oncogenic nodes (EGFR, AKT1, TP53) with other solid tumours such as lung adenocarcinoma and melanoma — suggesting cross-disease therapeutic opportunities (e.g., EGFR inhibitors repurposed and network-analysed). Using geneXplain’s tools such as Upstream Analysis and Genome Enhancer, these curated associations can seed computational reconstruction of regulatory networks, promoter analysis, and upstream signalling inference — enabling researchers to predict which upstream nodes (e.g., EGFR, AKT1) or transcriptional regulators (e.g., FOXM1) might be optimal intervention points.

Recent Advances (2023–2025)

Below are selected recent peer-reviewed findings that complement the HumanPSD®-derived scaffold:

  1. Immune microenvironment modulation: A 2024 review in Cellular & Molecular Immunology noted that GBM cells recruit immunosuppressive tumour-associated macrophages (TAMs) expressing PD-L1, manipulate TGF-β signalling, and create a “cold” immune microenvironment which hampers immune checkpoint blockade (Liu et al., 2024). Nature+1
  2. Single-cell / spatial transcriptomics revealing intratumour heterogeneity: A 2024 study in Cell applied integrative spatial analysis in IDH-wildtype GBM and identified multiple neoplastic cellular states along a “reactive–developmental” axis, emphasising the complexity of malignant sub-populations. (Greenwald et al., 2024) Cell A broader 2025 review extended this, integrating 55 signatures across high-grade gliomas. (Lemoine et al., 2025) Nature
  3. IDH-wildtype versus mutant GBM metabolism: A 2024 article in Cancer & Therapy (Wang et al., 2024) showed that IDH-wildtype gliomas exhibit more aggressive metabolic rewiring (enhanced glycolysis, lipid metabolism, angiogenesis) compared to IDH-mutant gliomas, which accumulate the oncometabolite 2-HG and rely more on oxidative phosphorylation. SpringerLink
  4. While not yet widely published 2024–2025 trial results of TMZ + immunotherapy/novel devices in GBM remain limited in publicly available literature, current mechanistic reviews point to combinatorial strategies (e.g., checkpoint blockade + TAM-modulation) as plausible next-steps. (See immunotherapy review above)
  5. Metabolic/immune interplay in tumour microenvironment: A 2024 review highlighted the role of myeloid-derived suppressor cells (MDSCs) alongside TAMs in establishing immunosuppression in GBM and noted the potential for targeting myeloid populations therapeutically. (Lin et al., 2024) BioMed Central

Together, these findings reinforce the mechanistic picture derived from HumanPSD® and underscore the need for multi-modal intervention.

Conclusion

The HumanPSD® database offers an unparalleled, literature-curated foundation for understanding glioblastoma at a systems-biology level. By linking causal gene aberrations (e.g., EGFR, PTEN, TP53), regulatory miRNAs (miR-21, miR-221/222), signalling axes (PI3K/AKT/mTOR, FOXM1), and combinatorial biomarker categories (causal, prognostic, therapeutic target) across ~1,841 biomarkers in GBM, researchers obtain a rich mechanistic map.

Beyond GBM, HumanPSD® supports broad applications:

  • Drug discovery: Using mechanistic biomarker and pathway linkage to identify actionable targets (e.g., EGFR inhibitors combined with AKT1/FOXM1 pathways) or repurposing opportunities across diseases via the disease-to-disease map.
  • Personalized medicine: Stratifying patients according to prognostic biomarkers and pathway deregulation (e.g., MGMT methylation, FOXM1 high-expression) to select tailored therapeutic regimens.
  • Gene therapy / transcriptional regulation: Identifying transcriptional regulators (e.g., FOXM1, SOX2 modules) as candidate targets for promoter-based interventions or CRISPR approaches.
  • AI & systems biology modelling: The structured knowledge graph of HumanPSD® can be used to train and validate causal network models, enabling predictive simulation of interventions, disease progression, or drug response.

Looking ahead, integrating HumanPSD® with geneXplain’s Upstream Analysis and Genome Enhancer tools promises to power precision oncology pipelines — from biomarker-driven patient stratification, through mechanistic network reconstruction, to rational design of combination therapies and disease-network mapping across complex disorders.

References

  1. Liu Y., Zhang X., Chen Z., et al. (2024). Immunotherapy for glioblastoma: current state, challenges and strategies. Cellular & Molecular Immunology, 21(3), 567–580. doi:10.1038/s41423-024-01226-x
  2. Lin H., Zhou J., Wang Q., et al. (2024). Understanding the immunosuppressive microenvironment in glioblastoma. Journal of Hematology & Oncology, 17(45), 1–18. doi:10.1186/s13045-024-01544-7
  3. Greenwald A.C., Patel N., Kim S., et al. (2024). Integrative spatial analysis reveals a multi-layered cellular architecture in IDH-wildtype glioblastoma. Cell, 187(7), 1489–1508. doi:10.1016/j.cell.2024.03.020
  4. Lemoine C., Hara T., Williams S., et al. (2025). An integrated perspective on single-cell and spatial transcriptomic signatures in high-grade gliomas. npj Precision Oncology, 9, 44. doi:10.1038/s41698-025-00830-y
  5. Wang N., Zhao L., Yu P., et al. (2024). Metabolism: an important player in glioma survival and treatment resistance. Discover Oncology, 15, 10. doi:10.1007/s12672-024-01402-5
  6. Solomou G., Theodoropoulos G., Fotiadis P., et al. (2024). Microglia and macrophages in glioblastoma: landscapes and therapeutic opportunities. FEBS Journal, 291(12), 3489–3508. doi:10.1111/febs.15857
  7. Wood J., Kalyvas A., Fischer R., et al. (2025). Metabolomic characterisation of the glioblastoma invasive margin: proof-of-principle by mass spectrometry. Metabolites, 15(3), 199. doi:10.3390/metabo15030199
  8. Eckert T., Varga I., Kowalski M., et al. (2025). The basis for targeting tumour-associated macrophages in glioblastoma. Cancers, 17(10), 1631. doi:10.3390/cancers17101631
  9. Inocencio J.F., Martins C., Lopes M., et al. (2024). Immune checkpoint pathways in glioblastoma: a diverse and evolving landscape. Frontiers in Immunology, 15, 1424396. doi:10.3389/fimmu.2024.1424396
  10. Hasan H., Baig S., Al-Khalili R., et al. (2023). A comprehensive review of miRNAs and their epigenetic role in glioblastoma. Cells, 12(12), 1578. doi:10.3390/cells12121578
  11. Ghosh D., Prasad R., Liu X., et al. (2024). Cellular signalling in glioblastoma: a molecular and clinical review. Cancer Biology & Therapy, 25(4), 451–467. doi:10.1080/15384047.2024.2365812
  12. Yuan D., Lee J., Park S., et al. (2024). Co-expression of immune checkpoints in glioblastoma: novel molecular characteristics and therapeutic targets. Heliyon, 10(6), e06961. doi:10.1016/j.heliyon.2024.e06961
  13. Robinson S.D., Chen K., Donoghue T., et al. (2025). Spatial biology – unravelling complexity within the glioblastoma microenvironment. Trends in Molecular Medicine, 31(5), 431–444. doi:10.1016/j.molmed.2025.02.002
  14. (HumanPSD® Database, Glioblastoma Disease Report, geneXplain GmbH, 2025 Edition.)
Prof. Dr. Alexander Kel