Cerebrospinal fluid (CSF) extracellular vesicles (EVs) show promise as a source of neurological disease biomarkers, although their precise origin is poorly understood. Current extraction techniques produce disappointing yield and purity. This study describes the application of ultrafiltration LC (UFLC) to CSF-EVs, compared with ultracentrifugation (UC), and explores CSF-EV origin. EVs are extracted from human CSF by UC and UFLC and characterized using nanoparticle tracking analysis, electron microscopy, and immunoblotting. EV and CSF proteomes are analyzed by LC-MS/MS. UFLC-isolated particles have size, morphology, and marker expression characteristic of EVs. UFLC provides greater EV yield (UFLC 7.90 × 108 ± SD 1.31 × 108 EVs mL-1 CSF, UC 1.06 × 108 ± 0.57 × 108 p < 0.001). UFLC enhances purity, proteomic depth (UFLC 622 ± 49, UC 298 ± 50, p = 0.001), and consistency of quantification (CV 17% vs 23%). EVs contain more intracellular proteins (Odds ratio [OR] 2.63 p < 0.001) and fewer plasma proteins than CSF (OR 0.60, p < 0.001). CSF and EV-enriched proteomes show overrepresentation of brain-specific proteins (EV OR 3.18, p < 0.001; CSF OR 3.37, p < 0.001). Overrepresentation of cerebral white matter (OR 1.99, p = 0.015) and choroid plexus proteins (OR 1.87, p<0.001) is observed in EVs. UFLC improves yield and purity of CSF-EVs. The EV-enriched proteome better reflects the intracellular and white matter proteome than whole CSF.
cerebrospinal fluids, exosomes, extracellular vesicles, extraction, origin, Biomarkers, Chromatography, Liquid, Extracellular Vesicles, Humans, Nervous System Diseases, Proteome, Ultrafiltration