Isolation of High-Purity Extracellular Vesicles by the Combination of Iodixanol Density Gradient Ultracentrifugation and Bind-Elute Chromatography From Blood Plasma

Abstract

Background: Extracellular vesicles (EVs) (isolated from blood plasma) are currently being extensively researched, both as biomarkers and for their therapeutic possibilities. One challenging aspect to this research is the efficient isolation of high-purity EVs from blood plasma in quantities sufficient for in vivo experiments. In accordance with this challenge, the aim of this study was to develop an isolation method in which to separate the majority of EVs from major impurities such as lipoprotein particles and the abundant plasma proteins albumin and fibrinogen. Methods: Samples of rat blood were centrifuged to remove cells, platelets, large EVs and protein aggregates without prior filtration. Density gradient ultracentrifugation was performed by loading plasma sample onto 50, 30, and 10% iodixanol layers and then centrifuged at 120,000 ×g for 24 h. Ten fractions (F1-10) were collected from top to bottom. Fractions with the highest EV content were further purified by ultracentrifugation, size exclusion, or bind-elute chromatography. Efficiency and purity were assessed by Western blots. Morphology and size distribution of particles were examined by dynamic light scattering and electron microscopy (EM). Results: The highest band intensities of EV markers Alix, Tsg101 and CD81 were detected by Western blot in F6 of small-scale DGUC (61.5 ± 10.4%; 48.1 ± 5.8%; 41.9 ± 3.8%, respectively) at a density of 1.128-1.174 g/mL, where the presence of vesicles with a mean diameter of 38 ± 2 nm was confirmed by EM and DLS. Only 1.4 ± 0.5% of LDL and chylomicron marker, 3.0 ± 1.3% of HDL marker, and 9.9 ± 0.4% of albumin remained in the EV-rich F6. However, 32.8 ± 1.5% of the total fibrinogen beta was found in this fraction. Second-step purification by UC or SEC did not improve EV separation, while after BEC on HiScreen Capto Core 700 albumin and lipoprotein contamination were below detection limit in EV-rich fractions. However, BEC decreased efficiency of EV isolation, and fibrinogen was still present in EV-rich fractions. Conclusion: This is the first demonstration that DGUC is able to markedly reduce the lipoprotein content of EV isolates while it separates EVs with high efficiency. Moreover, isolation of lipoprotein- and albumin-free EVs from blood plasma can be achieved by DGUC followed by BEC, however, on the expense of reduced EV yield.