Methodological insights into the isolation efficiency of human milk–derived extracellular vesicles

particles/mL) but slightly reduced vesicle integrity. Ultracentrifugation produced lower yields but maintained morphology and marker expression, especially at 100,000 × g. Filtration resulted in high counts but poor specificity and irregular morphology. Within the constraints of a single-donor design and based on the analytical endpoints assessed here, PEG precipitation (particularly when combined with deproteinization) and ultracentrifugation provided the most favorable trade-offs between yield, purity, and structural preservation, and these patterns should be confirmed in larger multidonor studies to evaluate their generalizability across donors. These findings also emphasize the critical role of preliminary purification in achieving high-quality HMEV preparations. To minimize biological variability and focus on method-dependent effects, all isolations were deliberately performed on aliquots of the same donor milk sample, allowing the observed differences to be attributed primarily to the isolation protocols rather than to the starting material.