Machine learning-integrated network toxicology uncovers glioma targets of DEHP
Frontiers in Toxicology · 08.05.2026 · 8
DOI: 10.3389/ftox.2026.1771011 ↗Kurz gefasst
Fachartikel in der Zeitschrift Frontiers in Toxicology, online erschienen am 08.05.2026. Verfasst von Ren Li und 4 weiteren Autor:innen. Beteiligte Einrichtung: Department of Environmental Health, School of Public Health, Shanxi Medical University, Taiyuan, China. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 56 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Background
Di (2-ethylhexyl) phthalate (DEHP), a ubiquitous environmental plasticizer, is increasingly linked to neurotoxicity and carcinogenesis. However, its role in glioma pathogenesis remains poorly understood. This study integrates network toxicology and machine learning to identify molecular targets of DEHP in glioma.
Methods
Potential DEHP targets were identified through four databases (CHEMBL, CTD, SwissTargetPrediction, PharmMapper). Glioma-related genes were screened using differential expression analysis and weighted gene co-expression network analysis (WGCNA) on GEO and TCGA datasets. Overlapping genes were subjected to functional enrichment, followed by 127 machine learning models to prioritize core genes. SHAP analysis interpreted model contributions, while COX regression assessed prognostic value. Molecular docking and dynamics simulations evaluated binding stability between DEHP and target proteins. In vitro validation was performed in U87 cells via RT-qPCR and Western blotting.
Results
A total of 77 overlapping genes were identified, enriched in neuroactive ligand-receptor interactions, GABAergic synapses, and ion channel activity. Machine learning prioritized 12 key genes (e.g., RELA, ABCA1, HIF1A), forming a parsimonious 12-gene diagnostic model with strong external discrimination (pooled validation AUC = 0.994). A high DEHP-related risk score was associated with poorer survival in the TCGA cohort and showed similar prognostic stratification across the external CGGA_325, CGGA_693, and GSE16011 cohorts. Molecular simulations confirmed stable binding between DEHP and core proteins. Experimental validation demonstrated dose- and time-dependent upregulation of RELA, ABCA1, and HIF1A in DEHP-exposed U87 cells.
Conclusion
This integrative approach provides a systems-level framework to prioritize DEHP-associated target genes and molecular signatures in glioma, extending beyond the previously reported PER3-related observation and offering candidate biomarkers for early detection and prognosis under environmental exposure.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Environmental Health, School of Public Health, Shanxi Medical University, Taiyuan, China.
- Center for Cerebrovascular Diseases Research, Shanxi Medical University, Taiyuan, Shanxi, China.
- Cerebrovascular Disease Center, First Hospital of Shanxi Medical University, Taiyuan, Shanxi, China.
- Department of Public Health Laboratory Sciences, School of Public Health, Shanxi Medical University, Taiyuan, China.
- Hacken Merdian Health, JFK University Medical Center, Neuroscience Institute, Edison, NJ, United States.
- Department of Emergency, First Hospital of Shanxi Medical University, Taiyuan, Shanxi, China.
Zitieren
APA
Li, R., Ren, C., He, L., Wang, S., & Guo, G. (2026). Machine learning-integrated network toxicology uncovers glioma targets of DEHP. Frontiers in Toxicology, 8. https://doi.org/10.3389/ftox.2026.1771011
Vancouver
Li R, Ren C, He L, Wang S, Guo G. Machine learning-integrated network toxicology uncovers glioma targets of DEHP. Frontiers in Toxicology. 2026;8. doi:10.3389/ftox.2026.1771011
BibTeX
@article{Li2026,
author = {Li, Ren and Ren, Chaomin and He, Lu and Wang, Shule and Guo, Geng},
title = {Machine learning-integrated network toxicology uncovers glioma targets of DEHP},
journal = {Frontiers in Toxicology},
year = {2026},
volume = {8},
doi = {10.3389/ftox.2026.1771011}
}
EUVIMED