An ultracentrifugation - hollow-fiber flow field-flow fractionation orthogonal approach for the purification and mapping of extracellular vesicle subtypes.

In the course of their life span, cells release a multitude of different vesicles in the extracellular matrix (EVs), constitutively and/or upon stimulation, carrying signals either inside or on their membrane for intercellular communication. As a natural delivery tool, EVs present many desirable advantages, such as biocompatibility and low toxicity. However, due to the complex biogenesis of EVs and their high heterogeneity in size distribution and composition, the characterization and quantification of EVs and their subpopulations still represents an enticing analytical challenge. Centrifugation methods allow to obtain different subpopulations in an easy way from cell culture conditioned medium and biological fluids including plasma, amniotic fluid and urine, but they still present some drawbacks and limitations. An unsatisfactory isolation can limit their downstream analysis and lead to wrong conclusions regarding biological activities. Isolation and characterization of biologically relevant nanoparticles like EVs is crucial to investigate specific molecular and signaling patterns and requires new combined approaches. Our work was focused on HF5 (miniaturized, hollow-fiber flow field-flow fractionation), and its hyphenation to ultracentrifugation techniques, which are the most assessed techniques for vesicle isolation. We exploited model samples obtained from culture medium of murine myoblasts (C2C12), known to release different subsets of membrane-derived vesicles. Large and small EVs (LEVs and SEVs) were isolated by differential ultracentrifugation (UC). Through an HF5 method employing UV, fluorescence and multi-angle laser scattering as detectors, we characterized these subpopulations in terms of size, abundance and DNA/protein content; moreover, we showed that microvesicles tend to hyper-aggregate and partially release nucleic matter. The quali-quantitative information we obtained from the fractographic profiles was improved with respect to Nano Tracking Analysis (NTA) estimation. The SEV population was then further separated using density gradient centrifugation (DGC), and four fractions were submitted again to HF5-multidetection. This technique is based on a fully orthogonal principle, since F4 does not separate by density, and provided uncorrelated information for each of the fractions processed. The "second dimension" achieved with HF5 showed good promise in sorting particles with both different size and content, and allowed to identify the presence of fibrilloid nucleic matter. This analytical bidimensional approach proved to be effective for the characterization of highly complex biological samples such as mixtures of EVs and could provide purified fractions for further biological characterization.