A Study in Engineering reports that N-glycans from serum-derived extracellular vesicles (EVs) can serve as non-invasive biomarkers for diagnosing and classifying childhood epilepsy.
“EV-associated glycans are stable and specific liquid-biopsy biomarkers, protected within lipid bilayers and able to cross the blood-brain barrier.”
Key Methodology & Findings
-
EV Isolation Comparison: Researchers evaluated three workflows—differential ultracentrifugation, reagent precipitation, and a combined exosome purification filter column with ultrafiltration (EPF/UF). EPF/UF was determined most suitable for large-scale clinical serum samples.
-
Glycan Profiling: N-glycan profiles from EVs and matched serum samples were obtained using MALDI-TOF-MS, revealing distinct glycosylation patterns.
-
Machine Learning Framework: A two-step machine learning framework identified 47 characteristic N-glycans from EVs that distinguish healthy controls from epilepsy patients and differentiate focal from generalized epilepsy.
-
Superior Diagnostic Performance: EV-derived N-glycans showed stronger diagnostic performance than serum N-glycans across multiple machine learning models, including random forest, XGBoost, logistic regression, and multilayer perceptron.
-
Network Analysis: A glycan correlation network was constructed showing dynamic changes in EV glycosylation during epileptogenesis.
Significance
This study highlights EV-associated glycans as stable and specific liquid-biopsy biomarkers. They are protected within lipid bilayers and able to cross the blood-brain barrier, reducing interference from abundant serum proteins.
Future Work
Further investigation will focus on functional validation of identified glycan signatures and expansion to diverse cohorts to support clinical translation.