Four Considerations for Enriching Extracellular Vesicles

from Genetic Engineering News by Amedeo Cappione III and Chandreyee Das

The Lack of Standardized Methods Is a Significant Bottleneck to Accurately Elucidating the Function of EVs

Given the importance of exosomes and other extracellular vesicles (EVs) in therapeutic development and mechanistic studies of cancer, neuroscience, and immunology, there are many researchers seeking the “best” protocol for isolating these vesicles. On ResearchGate.net alone, over a hundred questions related to exosome isolation have been posted, with some threads viewed over 10,000 times.

The past year has witnessed rapid changes in EV research, as studies uncover differences in function and cargo between extracellular particle subtypes. In particular, the International Society for Extracellular Vesicles (ISEV) has published a number of recommendations for EV research methods because the lack of standardized methods for EV enrichment and characterization represents a bottleneck to accurately elucidating the function of EVs.1,2

Our survey of EV-related research presented at the Annual Meetings of the American Association for Cancer Research (AACR) demonstrated that while, in 2014 a large number of researchers were using ultracentrifugation, by 2015 the majority are using sucrose gradient fractionation. We saw few, if any, instances in which precipitation methods were used for EV enrichment. Finally, we observed an increase in the thoroughness with which EV preps were being characterized, including Western blots for positive and negative EV markers.

Research Trends Driving Need for EV Enrichment: EVs Shape Cell Environment and Carry exRNAs

In presentations at recent conferences, including the AACR meetings and the 2014 meeting of the American Society for Exosomes and Microvesicles, the most commonly cited reason for studying EVs is that they mediate cell-cell communication. EVs carry out this function via diverse mechanisms, such as cell-independent processing of miRNAs by exosomes, stroma remodeling by vesicle-borne proteases, and suppression of immune activity in the tumor microenvironment.3,4,5

The role of EVs in transporting extracellular RNAs (exRNAs) is also of great interest, although recent reports suggest the possibility that circulating miRNAs may not necessarily be contained within EVs.6 The National Cancer Institute has funded an exRNA research program, out of which has emerged a useful web resource compiling protocols for EV enrichment and RNA isolation. Concurrently, the NIH Extracellular RNA Communication Consortium agrees that in order to accurately assess exRNA function, there is a need to work with more pure, well-characterized, and relatively quantified preparations of EVs.7

Four Considerations for Designing an EV Enrichment Protocol

1. What is the source of your sample?

Cell culture supernatant. If the EV source is cell culture supernatant, there are many options for preparation. One of the most popular methods is density gradient ultracentrifugation. A sample protocol would involve overlaying precleared, concentrated cell culture supernatant onto a solution of sucrose- or iodixanol-based density gradient medium. Ultracentrifugation at 100,000 x g for 2-3 hours, followed by a wash step, has been shown to provide >1 g/mL EVs, with relatively high purity (as measured by particle-to-protein ratio and absence of non-EV proteins (such as Ago2).8,9

Affinity purification using beads conjugated to antibodies specific to vesicle-specific proteins (such as anti-CD63, anti-CD9, or anti-MHC beads) are also quite effective for cell culture supernatant samples.10 However, current affinity-based approaches lack efficient ways to elute EVs from the solid phase, and are mainly used for surface phenotyping and cargo analysis.

Size exclusion-based purification, either by ultrafiltration, tangential flow filtration, or chromatography, may be a simple, versatile, cost-effective option for vesicle enrichment from cell culture supernatant (Figure 1). Although exact yields and purity may be dependent on the cell type and cell culture conditions, some studies show that modern ultrafiltration systems are faster and may be higher-yielding than density gradient methods.11 Another advantage is that, given the ultrafiltration device’s capacity for buffer exchange and sample concentration, purified fractions can be easily formatted to meet the requirements of any downstream analysis platform.

Proteinaceous biological samples (Serum, Plasma, CSF). EVs are found in nearly every biological fluid, as listed in a recent review of EV function, ranging from blood and breast milk to bile and synovial fluid.12 Bodily fluids are, in general, more viscous than cell culture supernatants. Therefore, these samples should be diluted and centrifuged for longer times and at higher speeds, as reported.13 If ultrafiltration is chosen as the enrichment method, undiluted biological samples may clog filter membranes. Also, note that certain commercially available immunoaffinity-based kits for vesicle purification are incompatible with biological samples. Consult manufacturer instructions before using any kit.

2. What information do you ultimately need to obtain from your vesicle preparation?

If EVs are being prepared solely to obtain RNA profile or protein cargo information, some degree of damage to the EV membrane (due to shearing forces from repeated, high-g-force ultracentrifugation, for example) might be acceptable. Charge neutralization (“salting out”) generates EV fractions that are suitable for transcriptomics, but not for other analyses, because they are contaminated with non-EV proteins.14 Density gradient ultracentrifugation, when compared to traditional ultracentrifugation and polymeric precipitation, showed the most consistent RNA profiles.9 In one study comparing enrichment of exosome-associated proteins, immunoaffinity capture provided twofold greater enrichment than did ultracentrifugation or density gradients.15

If the goal is enriching EVs for therapeutic application or to analyze their function, gentler techniques such as tangential flow filtration (TFF) and density gradient ultracentrifugation are more likely to preserve the membrane and, by extension, the biological activity.16

Exosomes (~30 µm EVs) have been successfully separated from microvesicles (larger, more heterogeneous EVs) using sequential centrifugal ultrafiltration (SCUF) of conditioned media.17 Not only were the protein profiles of the two fractions distinct, but the two fractions also showed differing degrees of invasiveness as measured by a Boyden chamber assay.17

One factor complicating size-based enrichment of EVs is that vesicles of differing function and biogenesis pathway may overlap in size.18 Because surface markers have, thus far, been a relatively reliable proxy for EV subtype, it is likely that affinity purification approaches hold the greatest promise as an enrichment technique as long as the EVs can be effectively eluted from the solid phase.

3. Can you tolerate interference from non-EV sample components?

Non-EV particles may be co-purified with EVs in polymeric precipitation-based techniques.19 Other methods, even if using serum-depleted medium, may yield fractions containing non-EV proteins, such as albumins, lipoproteins, argonaute complexes, or uromodulins. EV samples also contain high amounts of lipids, which can interfere with quantitation of total protein. Proper characterization of such fractions may require assay-free protein analysis, such as by infrared-based quantitation, and bead-based flow cytometric quantitation (Figure 2). Combining protein analysis and relative quantitation can yield the ratio between protein concentration and number of EVs. If this ratio is very high, it may indicate the co-purification of abundant proteins.

4. How precisely can you control experimental parameters?

Because vesicles may form as a function of sample collection and preparation, it is important to control the temperature, time, speed, and cell health parameters with which samples are collected and processed. The ISEV published recommended criteria for judging whether a reported EV preparation truly contained functional EVs.2 These criteria called for detailed experimental methods for isolating EVs, including reporting on health and viability of cells of origin.

Future directions

Given the heterogeneity of EVs and the subtle differences in their properties, many exciting new technologies are being explored for EV isolation, involving microfluidics, nanointerfaces, electrodialysis, and immunoaffinity capture followed by elution. In addition, methods thought of as “gold standard” today are likely to need further optimization. For example, EVs of diverse physical properties, bearing diverse surface markers, and having diverse mechanisms of biogenesis (Rab-dependent vs. Rab-independent) can only be separated based on kinetics of density gradient flotation because their equilibrium flotation is identical.20 Altering the ultracentrifugation speed can also change the composition of the EV fraction, and refinement of affinity-based approaches may further enable discrimination of EV subtypes.21 Fortunately, close collaborations have emerged among EV researchers and among suppliers of research tools, facilitating coordinated, documented adoption of EV enrichment methods.

Source – Genetic Engineering News