Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2024 Mar 4;9(3):1239–1251. doi: 10.1021/acssensors.3c02070

Effect of Sample Preprocessing and Size-Based Extraction Methods on the Physical and Molecular Profiles of Extracellular Vesicles

Lucile Alexandre †,‡,§, Molly L Shen †,‡, Lorenna Oliveira Fernandes de Araujo †,‡, Johan Renault †,‡, Philippe DeCorwin-Martin †,‡, Rosalie Martel †,‡, Andy Ng †,‡, David Juncker †,‡,*
PMCID: PMC10964911  PMID: 38436286

Abstract

graphic file with name se3c02070_0008.jpg

Extracellular vesicles (EVs) are nanometric lipid vesicles that shuttle cargo between cells. Their analysis could shed light on health and disease conditions, but EVs must first be preserved, extracted, and often preconcentrated. Here we first compare plasma preservation agents, and second, using both plasma and cell supernatant, four EV extraction methods, including (i) ultracentrifugation (UC), (ii) size-exclusion chromatography (SEC), (iii) centrifugal filtration (LoDF), and (iv) accousto-sorting (AcS). We benchmarked them by characterizing the integrity, size distribution, concentration, purity, and expression profiles for nine proteins of EVs, as well as the overall throughput, time-to-result, and cost. We found that the difference between ethylenediaminetetraacetic acid (EDTA) and citrate anticoagulants varies with the extraction method. In our hands, ultracentrifugation produced a high yield of EVs with low contamination; SEC is low-cost, fast, and easy to implement, but the purity of EVs is lower; LoDF and AcS are both compatible with process automation, small volume requirement, and rapid processing times. When using plasma, LoDF was susceptible to clogging and sample contamination, while AcS featured high purity but a lower yield of extraction. Analysis of protein profiles suggests that the extraction methods extract different subpopulations of EVs. Our study highlights the strengths and weaknesses of sample preprocessing methods, and the variability in concentration, purity, and EV expression profiles of the extracted EVs. Preanalytical parameters such as collection or preprocessing protocols must be considered as part of the entire process in order to address EV diversity and their use as clinically actionable indicators.

Keywords: extracellular vesicles, extraction, characterization, preanalytical parameters, comparison, subpopulation

Extracellular vesicles (EVs) are nanometric lipid bilayer vesicles released by cells. They play a central role in cell-to-cell communication via ligand signaling and shuttling of cargo between cells, including proteins, lipids, and nucleic acids between the cells.1 EVs are considered highly promising biomarkers, notably for cancer,2 as they can be directly collected by liquid biopsy, without the need for invasive procedures or surgeries. EVs are commonly characterized and identified by the expression of proteins; the tetraspanins CD63, CD9, and CD81 are “canonical” EV markers.3 The expression of these proteins is analyzed using a variety of methods, including Western blots, mass spectrometry, flow cytometry, microarray, etc. Previous publications have shown that detailed protein analysis of EVs can reveal essential information on the evolution of a disease, such as the development of a tumor,4 Alzheimer’s disease propagation,5 or the spread of Parkinson’s disease.6

Prior to analysis, EVs must be collected, extracted, and preconcentrated from the sample and the complex matrix. Blood, and by extension plasma and serum, is the most widely used biofluid for analysis and is readily collected via a phlebotomy. Indeed, the complex matrix of plasma can interfere with the analysis of EVs and produce false-positive signals, as numerous nanometric particles such as lipoproteins, nucleosomes, or vaults. This puts stringent demands on the extraction methods, which must content with high concentrations of potential contaminants. Plasma has been found to be a rich source of EVs and the number of EVs can be of the order of 1010–1012 particles/mL. While it is a large number, it corresponds to picomolar concentrations, considered as very low for proteins or metabolites. Hence, the preservation of EVs and their purification and enrichment constitute critical preanalytical steps to subsequent EV analysis.

Centrifugation remains the gold standard thanks to the high purity achieved by several steps7,8 and sometimes using gradient methods.9 However, it is time-consuming, and requires expensive instruments and consumables. Size-exclusion chromatography is now widely used as a simpler, less expensive, and faster option.10,11 Several other techniques have already proven their effectiveness, such as immuno-affinity12,13 or polymeric precipitation.14 Microfluidics offers the option to work with smaller volumes, reduce costs, and automatize processes. New solutions have been developed on-chip, based on filtration,15 acoustics,16 viscosity,17 or thermophoresis.18 The need for simplified extraction is served by instruments that automate the workflow, such as the centrifugal dual-filtration system called Exodisc19 or the microbead-assisted acoustic trapping of EVs, named AcouTrap.20,21

The heterogeneity of EVs constitutes a significant challenge to preprocessing and analytical procedures. EVs from different origins and types cover several orders of magnitude in volume, with size ranging from 50 to ∼500 nm.22 The membrane composition and cargo of EVs, either on their surface or encapsulated inside the lipid bilayer, are also heterogeneous, preventing the identification of a unique biological marker for targeting EVs. Efforts have been made to establish protein expression and identify subpopulations; however, variations between studies and publications are apparent and are often ascribed to the lack of standardization of preprocessing and extraction steps. However, for membrane protein-based enrichment, there is a voluntary bias toward a particular subpopulation; other methods also enrich subpopulations of EVs based on physical (e.g., size), chemical (e.g., phospholipid composition), or molecular features. While not well understood, it is now widely recognized that each extraction method is biased in some way based on the molecular content or the size of the EVs, and, as a result, may fail to enrich certain subpopulations that then remain undetectable.

The standardization of EVs’ concentration, extraction, and analysis remains a significant challenge. It limits our ability to compare the results from different studies and to support the understanding of EVs and their biological function and clinical potential. The EV community promotes the need for standardization and multiple papers have been published, notably by the International Society of Extracellular Vesicles (ISEV),23 leading to the creation of new reporting technologies such as Exocarta24 or EV-track.25 Hence, the performance of various EV concentration and extraction methods and their biases have been the subject of comparative studies,26,27 especially the EV extraction from plasma.28 However, in the absence of a ground truth with regard to what the pristine EV composition is, characterizing the absolute performance is not possible. Therefore, comparative methods are primarily used, and new methods benchmarked to existing ones to understand their performance and biases. The study by Paget et al.29 introduced a comparison between acoustic extraction and standard isolation techniques, while the work conducted by Dong et al.30 compared the extraction performances of a lab-on-a-disc filtration to more established techniques. However, these two new techniques were not directly compared with each other. The conclusions drawn from both studies emphasize the need for additional information to assist scientists in constructing their own extraction platforms by understanding the advantages and disadvantages of each method. Moreover, while previous studies29−31 have delved into the diversity of extraction methods and provided tentative comparisons, our approach in this study involves correlating the information obtained from standardized instruments with the data generated by an in-house microarray immunoassay32 to study the protein profile of EVs. We share here a global view incorporating not only the impact of each method of extraction but also the influence of the selected preprocessing protocols.

In this paper, using cell supernatant and plasma as samples, we evaluate four label-free EV enrichment methods: ultracentrifugation (UC), size-exclusion chromatography (SEC) columns, and two emerging microfluidic technologies: centrifugal filtration on a lab-on-a-disc (LoDF)19 and acoustic enrichment (AcS).1 As no biological marker allows to target all EVs, we decided to focus on separation techniques based on the size of the particle. EVs extracted from cell supernatants from human squamous carcinoma cell line (A431) and pooled plasma samples were collected. Next, the extraction and enrichment of spiked-in and native EVs were characterized for three replicates. We also report the effect of blood collection tubes and preprocessing protocols. EVs in pre- and postenrichment samples were characterized by transmission electron microscopy (TEM), tunable resistive pulse sensing (TRPS), nanoparticle tracking analysis (NTA), and antibody microarrays. The enrichment methods are benchmarked in terms of EV integrity, size distribution, concentration, sample throughput, time-to-result, ease-of-use, purity, and cost, as well as by comparing the expression profiles of eight membrane proteins. Our study highlights the impact of preanalytical steps and enrichment biases in terms of EV subpopulations, contamination, and changes in protein expression.

Materials and Methods

EVs from Cell Supernatant

Human epidermoid carcinoma cell line A431 with CD63-GFP transgene33 was maintained in Dulbecco’s modified Eagle medium (Gilbco) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (PS) (Gilbco) under American Type Culture Collection’s (ATCC) recommendation. To initiate EV collection, the normal growth media was replaced with DMEM media containing 5% EV-depleted FBS and 1% PS. When 80% confluency was obtained, the cell supernatant was collected, centrifuged at 400g for 15 min, and filtered using 0.22 μm membranes (Filter Millex 0.22 μm #SLGPR33RS). The cell supernatant was concentrated to the desired volume via centrifugation. For SEC and AcS extraction, this step was repeated, as part of the supplier protocol.

Plasma Collection

Plasma was collected from healthy patients with informed consent form. The collection was approved by the McGill research ethics office (IRB study #A05-M27-16B). Blood samples were collected by a certified nurse and anonymized. The blood was collected in EDTA (VWR, CABD367861L) and citrate (VWR, CABD369714L) tubes, stored at 4 °C, and processed within 4 h of collection. The collected blood was then centrifuged (Thermo Fischer Scientific, Sorvall RT1 centrifuge) for 15 min at 1100g, then frozen at −80 °C. For extraction assessment, plasma samples were spiked with A431 EVs at the adjusted concentration of 1010 parts/mL, measured by TRPS.

Preanalytical Parameters

Three different sample preprocessing methods are compared in this paper: first, samples were thawed at room temperature before use, with no additional centrifugation or filtration step; the second method relies on a second step of centrifugation at 1100g for 15 min before storage at −80 °C. After storage, samples were thawed at room temperature, centrifuged at 2500g for 15 min and filtered through a 0.22 μm membrane; finally, the samples were thawed at room temperature, centrifuged first at 4700g for 15 min then at 14,000g for 30 min, and finally filtered through a 0.22 μm membrane. Collection of samples was performed under the ethics approval of the study A04-M36-13B (IRB number).

Extraction Methods

Ultracentrifugation extraction was performed by pouring a 1–2 mL sample into a polycarbonate tube and diluting in PBS up to 8 mL. The diluted sample was then centrifuged using a Beckman ultracentrifuge (Optima L90K) with two cycles of 1 h 10 min at 110,000g at 4 °C (rotor type 70Ti, Beckman Floor Ultracentrifuge Optima L90K). Between the cycles, the supernatant was discarded and the pellet was resuspended in 8 mL of PBS. The final pellet was resuspended in 150 μL of PBS. SEC was performed using qEV SEC columns and purchased from Izon (qEVoriginal/70 nm/500 μL), following their commercial protocol. For both microfluidic systems, samples were processed precisely according to the manufacturer’s protocol. 500 μL of the preconcentrated samples was flown through the columns and only fraction 8 was collected as eluate. The acoustic extraction was performed with the AcouTrap system developed by AcouSort. 50 μL of a suspension of 1 μm diameter beads was injected through the system at 50 μL/min followed by 90 μL of preconcentrated cell supernatant or 50 μL of plasma diluted 1:2 at 15 μL/min and by 200 μL of PBS at 25 μL/min. The extracted EVs were released in 50 μL of PBS at 1 mL/min. The process was repeated to match the volume of the sample to analyze. A last step of centrifugation at 500g for 2 min was performed to separate the sample from the beads. Finally, the LoDF system was purchased from LabSpinner, Exodisc C, and P. The disc filters were first wetted by 1 mL of PBS; then, 2 mL of cell supernatant or 100 μL of the plasma sample was flown through the system followed by a washing step with 100 μL of PBS. The 100 μL supernatant block in-between the filters was then collected, and the chamber was washed with 100 μL of PBS. Volumes of the samples analyzed are summarized in Table S1.

Characterization Methods

EV characterization included NTA (NanoSightNS500), TRPS (Izon), Nanodrop (spectrometerND1000), TEM (FEI Tecnai 12 120 kV), and a custom antibody microarray for protein profiling (description in the Supporting Information).

Analysis

All results are plotted with each data point, means, and standard deviation (SD). Drawing and analysis were performed using GraphPad Prism. For plasma samples, the normalized final concentration of extraction is the final concentration after extraction related to the initial and final volumes of the samples (eq 1 in Supporting Information). For the heatmaps, the fluorescent signal corresponding to a protein (capture on the slide) is normalized by the signal corresponding to the same capture spot recorded in buffer. For each graphic, clustered analysis is run to organize the columns’ order on Matlab (function clustergram), based on Euclidean distance, with the drawing of the corresponding dendrogram; then, the signal is normalized by the highest measured signal to obtain a percentage value.

Results

EV Extraction Methods

We evaluated four EV extraction methods using the cell supernatant and subsequently plasma samples, including (i) UC, (ii) SEC, (iii) LoDF, and (iv) AcS. The protocols are detailed in the section Materials and Methods and illustrated in Figure. 1. Briefly, UC was conducted using a standard protocol including two cycles of centrifugation with dilution in order to minimize contamination. SEC was conducted according to the manufacturer’s protocol and fraction 8, identified as the purest fraction based on supplier data, was collected for further analysis. AcS uses the so-called AcouTrap for trapping 1-μm-diameter beads. The waves generated by the electrodes that went through the conformation of the beads induced acoustic streaming and vortexes that in turn trapped the EVs, the latter being too small for direct acoustic trapping. LoDF was done using centrifugal flow across a spinning disk and by flowing the sample through a first filter with 600 nm pores that retained large particulates, followed by a second filter of 20 nm pores that retained EVs. For both microfluidic systems, the manufacturers’ protocol was followed.

Figure 1.

Figure 1

Open in a new tab

Description of the methods of extraction of EVs: (A) UC extraction was performed by two successive cycles of centrifugation and dilution of the sample; the final pellet was resuspended in 150 μL of PBS. (B) SEC was performed on 500 μL samples following the manufacturer’s protocol. The final volume was 500 μL. (C) For AcS samples, EVs were trapped in-between 1 μm diameter beads. The extracted EVs were released in 50 μL of PBS. (D) The LoDF system is composed of two filters with pores of 600 and 20 nm. Extracted EVs were collected in-between filters for a final sample of 200 μL.

The characterization was performed starting with cell supernatant samples to limit the number of contaminants and extended to human plasma, which has a complex matrix and is clinically more relevant.

Characterization of EVs Extracted from the Cell Supernatant

Comparison of the Size and Concentration of Extracted EVs

EV supernatant samples from the A431 cell line were collected and processed with one of the four extraction methods on the same day. The total protein concentration of the sample was measured by UV absorption at 280 nm (Figure 2A). The concentration of proteins was comparable for UC, SEC, and AcS, while LoDF samples had much higher UV absorption as a result of higher protein concentration. Next, the EV concentration and size distribution were measured by TRPS.34 The particle concentration for LoDF was about 10-fold higher compared to UC and SEC, while AcS yielded an overall lower concentration of particles along with larger variability, Figure 2B. For LoDF the average size of the extracted particles was ∼15% smaller, Figure 2C,D. The smaller average size could be related to better yield in extracting smaller EVs, or alternatively, reflected contamination by other particles possibly comprising proteins, nucleic acids, lipoproteins, and mixtures. To evaluate which one of the two scenarios was applicable, we analyzed the samples by TEM (Figure S1, Supporting Information). We observed a higher number of contaminants with diameters under 30 nm for LoDF consistent with higher particle counts and smaller average diameter. These results are consistent with the literature on Exodisc LoDF.26,30

Figure 2.

Figure 2

Open in a new tab

Characterization of (A) protein content of the sample; (B) final concentration and volumes of samples after extraction; and (C, D) size distribution of EVs extracted from the cell supernatant using UC, SEC, AcS, and LoDF by TRPS. Panel (D) is an enlargement of the data from panel (C). Samples processed using the LoDF have the highest protein content and the highest concentration of EVs, while EV size is smaller than with the three other methods. AcS samples had a lower concentration of particles along with larger variability (4 replicates showing mean and SD as error bars).

Characterization of the Protein Expression of Extracted EVs

The protein expression profile of the extracted EVs was evaluated by an in-house microarray immunoassay (Figure 3 A,B). This assay was used to scan a high number of protein targets (17 per assay) and select the relevant ones. The selection of protein targets encompassed transmembrane proteins (CD63, CD9, CD81, Integrin α6, Integrin α2, Integrin β1, Integrin β4, ADAM10, EGFR, EpCAM, CD82, CD44, CD133, E-cadherin, HER2), phosphatidylserine (identified by TIM4), and cytosolic proteins employed as negative controls (Alix, HSC70, HSP90β, and TSG101). The microarray was incubated with a sample containing 1.5 × 109 particles/mL, and bound EVs incubated with a cocktail of biotinylated antibodies against the three tetraspanins CD63, CD9, and CD81, followed by fluorescent-labeled streptavidin (Figure 3 C). Targets that did not yield a signal above the background or gave inconsistent signals were excluded from further analysis, resulting in the refined selection of the first nine targets: CD63, CD9, CD81, integrin α6, ADAM10, EGFR, EpCAM, Tim4, and Alix. Transmembrane and cytosolic proteins, as advised by the MISEV2018 guidelines,23 were included in the panel of capture antibodies to help confirm the integrity and nature of EVs.

Figure 3.

Figure 3

Open in a new tab

Antibody microarray-based profiling of the protein expression of EVs extracted from A431 cell supernatant. (A) Schematic of antibody spots, (B) of array layout, and (C) of immobilized EVs with a fluorescently labeled detection antibody cocktail against CD9, CD63, and CD81. Fluorescent labeling is achieved through a biotin–streptavidin attachment. (D) Heatmap of protein expression in EVs. Tetraspanins and integrins are generally highly expressed, while the expression levels can vary according to the extraction method. The low expression of Alix indicates that EV integrity was preserved. The binding of EVs to TIM4 was observed for UC, SEC, and LoDF, as expected, but not for AcS. Conversely, EGFR expression was low everywhere except in AcS EVs, suggesting that AcS extracts a markedly different subpopulation compared to other methods. These results underline that the method is enriched for different subpopulations of EVs. As the signals are normalized for each line, the information is related to the ratio between the subpopulations extracted by each technique rather than the expression of a specific protein.

Assuming that the expression of tetraspanins is conserved across the four methods, if we neglect the influence of changes in average EV size, then the relative fluorescent signals for each method will reflect the relative enrichment or depletion of EVs immobilized by the capture antibody. The lower expression of those proteins in LoDF samples, Figure S2, would be consistent with higher levels of contamination or a bias toward smaller EVs that may express a lower number of proteins. The integrin α6, involved in cellular adhesion,35 is highly expressed by EVs from the A431 cell line. The epidermal growth factor receptor EGFR and the metalloproteinase ADAM-10 were selected based on the reported high expression in this cell line.36 EpCAM is also a recommended marker for EVs and is known to be overexpressed in A431 cells. Interestingly, we observed significant differences in expression of these proteins in the EVs as a function of the extraction method. TIM4 (T cell immunoglobulin and mucin domain containing 4) binds phosphatidylserine, which is exposed on many EVs37 and was previously used for EV isolation.38,39 Here we used it to assess the binding of EVs independently of protein expression. The low phosphatidylserine expression in EVs extracted by AcS combined with difference in expression of other proteins suggests that AcS may be extracting different subpopulations of EVs. The integrity of EVs was confirmed by the absence of signals for the cytosolic protein Alix (ALG-2-interacting protein X).

A significant variation in the protein profile of EVs extracted through size-based techniques has already been found in previously published articles29 and could be related to the cutoff points with regard to small-sized vesicles.

Characterization of EVs Extracted from Human Plasma Samples

We evaluated the performance of the four extraction methods to enrich EVs from plasma collected from healthy patients in EDTA and citrate. These tubes were spiked with A431 EVs, extracted by SEC, using the processing protocol established above. The objective was to investigate whether the analysis could reveal the specific signature of this precharacterized subpopulation of EVs within the extracted samples.

Characterizing Sample Preprocessing on the EV Collection

To better understand the impact of preanalytical variables, we compared three sample preprocessing (SPP) protocols, from simpler to more complex:

SPP1: Collect plasma → freeze → thaw

SPP2: Collect plasma → centrifuge (1100g) → freeze → thaw → centrifuge at low g (2500g) → filter through a 0.22 μm membrane

SPP3: Collect plasma → freeze → thaw → centrifuge at moderate g (4700g) → centrifuge at high g (14,000g) → filter through a 0.22 μm membrane

Figure 4 shows the protein content measured by UV absorption, number of nanoparticles, and average nanoparticle size for SPP1–SPP3 using either EDTA or citrate tubes for blood collection and plasma separation. A general trend observed was that the more the steps the protocol includes, the more the loss of EVs observed, while the purity increases, as one might expect. UV absorption measures the EV-associated proteins and free proteins, and reflects this trend. Again, LoDF yields an ∼10 times higher protein concentration, likely due to contaminants, while AcS shows a low protein concentration consistent with the low yield but high purity, both replicating the trends observed in cell supernatant extraction. TRPS measurements reveal a low particle concentration for AcS that falls below 109 part/mL, which is below the threshold of particle numbers needed for some analytical tools and applications. The reduction in average particle size following the additional centrifugation step for SPP2 and SPP3 reflects the removal of contaminants such as cell debris as well as of EVs.

Figure 4.

Figure 4

Open in a new tab

Characterization of the effect of sample collection tubes, preprocessing, and extraction method on the EV protein content, concentration, and size for plasma from health volunteers. Total protein content as measured by UV absorption at 280 nm (left column, A, D, G, and J) on the nanodrop, concentration (middle column, B, E, H, and K,) and average size (right column. C, F, I, and L) of the particles measured by TRPS. The EVs were extracted from both EDTA and citrate tube samples in three conditions: thawed, unprocessed samples (SPP1); additional centrifugation before storage, thawing, low g centrifugation (SPP2); additional moderate g centrifugation following thawing (SPP3). SPP3 yielded the highest purity (mean shown as black bar; data and SD in color).

The overall protein concentration, particle number, and average diameter are not significantly different whether citrate or EDTA is used as anticoagulant, except for UC extraction. EDTA samples seem more variable among the replicates whereas citrate samples are consistent, especially for UC; however, more experiments are needed to confirm this trend.

Based on these results, and because contamination of EV samples by non-EV proteins and particles is a major concern for EV analysis, the SPP3 protocol was selected for subsequent experiments as it yields purer EV samples.

TEM Imaging of EVs

To evaluate the integrity of plasma EVs following SPP3 and extraction, we imaged them by TEM, Figure S4. For AcS samples with low EV concentration, a larger volume was used to obtain a comparable number of EVs. Negative staining, using uranyl acetate, was used to enhance the contrast between the particles and the background, including contaminants. We identified round and intact EVs identified by their “cub-shape”40 of different sizes in samples obtained by every extraction (see the white arrows); the particles were smaller in LoDF samples.

TEM images can be used as a proxy for qualitatively assessing the concentration, but with the caveat that they constitute only a small sample that may not be representative, and that EVs may not be distributed homogeneously on the TEM grid. TEM images of LoDF samples revealed a higher concentration of particles relative to other samples, consistent with the literature.30 The images of AcS-extracted EVs are comparable to the particles observed in other publications,20 where the same difference in size was also observed. Paget et al.29 observed by TEM and measured by NTA a smaller size of particles extracted by acoustic trapping compared to UC or SEC, similar to our observations.

Contamination of samples can also be evaluated via the TEM images, Figure S4. Differences in the level of gray in the background of TEM images are related to the density of contaminants such as proteins, salt, and aggregates. In the images of samples extracted by SEC columns, long filaments were observed, which are known residues from sepharose resins.41 Samples extracted by LoDF had a high concentration of contaminants, consistent with previous publications.30 Round and bright particles that are distinct from the “cup-shaped” EVs were identified as lipoproteins. However, this affirmation was altered by the emergence of exomeres (i.e., EVs < 50 nm), which have a similar round and bright appearance as lipoproteins and do not adopt a cup shape. There is now ambiguity as to whether the bright spots are contaminants or exomeres.

Comparison of Protein Concentration, EV Concentration, and Size Distribution

We evaluated the total protein concentration, particle concentration, and size distribution by UV absorption, see Figure S5, Supporting Information. TRPS and NTA measurements were done of plasma samples collected in citrate and EDTA tubes, processed according to SPP3, and extracted using each of the 4 methods. As plasma samples are more complex than cell supernatant ones, we measured the concentration and size by both TRPS and NTA. NTA tracks particle diffusion, thus deriving the coefficient of diffusion and the particle diameter.42 Coupling TRPS and NTA provides a more complete characterization of the EV profile.

A normalization process was necessary because unlike for the cell supernatant, the starting sample volume of plasma for each method was not the same, Table S1. The final concentration was corrected by the ratio of final to initial volumes to produce a result that represents the yield of extraction. The final concentrations of the extracted samples before and after normalization by volume, as explained in Materials and Methods section, are shown.

UC produces a high concentration in the extracted samples, while SEC offers a higher yield of extraction, after normalization. Differences in measured concentrations between NTA and TRPS are significant only for AcS samples, whose final concentrations are lower than the ones from other samples. The discrepancy could be explained by the differences in limits of detection of NTA and TRPS. LoDF extraction resulted in a high concentration of proteins, more than 50 times higher than those obtained by other methods. TRPS and NTA measurements revealed a high concentration of nanoparticles after extraction for LoDF compared to UC, SEC, and AcS (Figure 5A,B), for which the concentrations were preserved after normalization (Figure 5C,D). These observations are close to the ones by Dong et al.,30 who also observed a higher protein content in samples extracted by LoDF, with a high recovery rate. However, both for EDTA and for citrate plasma, a high concentration of contaminants was observed by TEM, which could contribute to an increase in both protein concentration and nanoparticle count, and a broad size distribution. The distribution and ratio of contaminant nanoparticles vs EVs could not be established and will require further characterization.

Figure 5.

Figure 5

Open in a new tab

Characterization of the concentration, yield, and size distribution of EVs extracted from EDTA and citrate plasma samples using UC, SEC, AcS, and LoDF by TRPS (A, C, E) and NTA (B, D, F). A high yield of extraction was measured for LoDF and SEC extraction. For all samples except AcS, the particle concentrations as measured by TRPS and NTA are consistent; for AcS, TRPS yielded a lower concentration than NTA that can be related to a low final concentration compared to other samples. The measured size of the particles is higher when measured by TRPS, as observed on the measured profile (E, F) between NTA and TRPS. Experiments included 3 replicates, with error bars representing SD and showing a line for the mean.

Whether EDTA or citrate tubes were used did not affect the composition of extracted EVs for SEC AcS and LoDF samples in our hands. However, for UC, the protein concentration appeared to be higher (almost twice as high, although not meeting statistical significance) in citrate samples compared to EDTA ones, Figure S5. The concentrations of EVs were also higher in UC citrate samples compared to EDTA ones, Figure 5A,B.

No obvious difference in size of extracted EV was noticed between the two anticoagulants used for those experiments. The measured size distribution of the particles is shifted toward bigger sizes when measured by TRPS compared to NTA for all samples, Figures 5E,F, S6 and S7. It correlates with the fact that NTA detects a higher number of small EVs than TRPS.43 Our data suggest that UC allowed to extract populations of bigger EVs than SEC, AcS, and LoDF, as can also be observed on TEM imaging, Figure S4.

Protein Expression Characterization of Extracted EVs

An antibody microarray was used to assess the multiplexed protein expression of EVs captured using a cocktail of anti-CD9, -CD63, and -CD81 antibodies, Figure 6 (see Materials and Methods section for details). The analysis for AcS EDTA and AcS citrate samples was performed at a lower concentration. Plasma samples were spiked with the cell supernatant A431 EVs analyzed previously, with the expectation of detecting both spiked and native EVs.

Figure 6.

Figure 6

Open in a new tab

Protein expression in EVs from the A431 cell supernatant spiked into plasma profiled by antibody microarray. A431 cell supernatant EVs were spiked into plasma; then, the samples were extracted by UC, SEC, AcS, and LoDF from the EDTA and citrate plasma tubes and incubated on the microarray. EVs were immobilized by capture antibodies (top labels), and then probed with a cocktail of fluorescent detection antibodies against CD9, CD63, CD81, and after rinsing, the fluorescence signal was detected. Log values of the signals were normalized by the highest protein expression for each sample. For UC, SEC, and LoDF, the samples were adjusted to 1010 part/mL and AcS 108 part/mL. As expected, CD9 was highly expressed in all plasma samples, and differences in protein expression were less visible due to the high number of EVs in plasma. The protein profile of EVs from A431 cell supernatant prior to spiking into plasma is shown at the bottom as a guide. Note that both spiked EVs and native EVs are detected in the plasma samples.

As explained before, the panel of protein targets included transmembrane proteins, phosphatidylserine (recognized by TIM4), and cytosolic proteins serving as negative control. Tetraspanins and integrins are highly expressed in all plasma samples. However, it is interesting to note that clustering may differ between the cell supernatant and plasma sample. CD9, ADAM-10, Integrin α6, and Tim4 are highly expressed in all samples, whereas no signal was measured for Alix. However, CD63 and CD81 are clustered in cell supernatant samples but not in plasma samples. EVs extracted from plasma samples highly expressed CD944 compared to EVs extracted from the cell supernatant. We conclude that the signal from native EVs dominated the one from spiked-in EVs to form this cancer cell line. EpCAM and EGFR expression specific to the A431 cell line was decreased on the other hand, which along with other EVs was subdued by the signal from the plasma sample. As expected, a low signal was recorded from LoDF samples and ACS samples, related to high contamination and lower concentration, respectively, as well as lower consistency (see Figure S7). When using EDTA compared to citrate, UC and SEC had a slightly higher protein expression, while AcS and LoDF had a lower one.

Paget et al.29 similarly employed a microarray technique to characterize EVs post extraction. While both studies agree on the high expression of CD9 in extracted particles from plasma samples, disparities arise concerning the profiles of other tetraspanins. These variations could be attributed to the selected cohort or differences in microarray preparation. It is noteworthy that in both studies, the signal for Alix is low, underlining the integrity of the EVs.

Discussion

Previous publications have evaluated new technologies of extraction,19,20 extraction from different media,30 or compared preanalytical parameters for EV collection and storage.45,46 We provide here a comparative study that evaluates the sample process flow, sample nature and composition, and extraction protocols and methods, to identify important parameters separately as well within a process flow and identifying codependencies and optimal process flows. We compared collection protocols and four methods of extraction: (i) UC, (ii) SEC, (iii) AcS, and (iv) LoDF for isolation of EVs from cell supernatant and human plasma. We optimized preanalytical parameters for plasma preprocessing, such as collection tubes and various centrifugation steps, to match the usual hospital protocols and purity requirements, finding that two steps of centrifugation after thaw and a 0.22 μm filtration allow reduction of contamination. Finally, we evaluated the impact on the EV protein profile based on our home-built immunoassay.

Our study corroborated the high purity of EV samples obtained by UC, while also showing a lengthy process time of this method and difficulty to gain access in our case. To qualitatively compare the strengths and weaknesses of each of the four purification methods for cell supernatant and plasma, we benchmarked their performance according to five and six criteria, respectively, including (i) final concentration, (ii) speed, (iii) volume preconcentration factor, (iv) purity, (v) accessibility, and in the case of plasma, (vi) yield, Figure 7. The performance of different methods for processing the cell supernatant showed a significant variation for most parameters, Figure 7A. No individual method outperforms the others, and each method has advantages (high purity for UC, accessible for SEC, fast for AcS, and high yield for LoDF) and drawbacks (lengthy for UC, low preconcentration for SEC, low final concentration for AcS, and high contamination for LoDF). The analysis of the protein expression in the purified EVs revealed differences between all methods, suggesting bias and selective enrichment or depletion of subpopulations of EVs. The bias introduced by the extraction method warrants comparative studies across sites to account for it, harmonize extraction protocols and methods, and develop strategies to benchmark and compensate for possible bias.

Figure 7.

Figure 7

Open in a new tab

Evaluation of the four methods of extraction: UC, SEC, AcS, and LoDF for (A) cell supernatant samples in terms of speed, volume preconcentration factor (ratio of the final to initial volume), recovery rate, purity, and accessibility; (B) EDTA and (C) citrate plasma according to the same criteria plus yield. UC shows excellent performance in purity and preconcentration, but accessibility, yield, speed, and final concentration are not optimal. SEC, AcS, and LoDF have interesting characteristics that make them challenging. However, all radar charts are unbalanced, showing a remaining need for a simple, automated EV extraction method.

A preanalytical protocol is necessary prior to the extraction of EVs from plasma; three different ones were tested, Figure 3, and their influence on the purity and profile of the extracted EVs were characterized for each extraction method and the best combinations were identified. The method of extraction also impacts the properties of the extracted EVs, Figure 7B,C. UC provides a high volume preconcentration factor, as the pellet from a large sample can be resuspended in a volume of buffer as low as 150 μL, inducing a high final concentration with exceptional purity. However, it is a lengthy process, the yield is low considering the high initial volume, and accessibility can be a challenge if an ultracentrifuge is not available. SEC is an accessible alternative with midrange performance on many criteria. The final concentration is lower than with UC but with a higher yield of extraction. SEC is interesting for studies where the volume of the samples is small (from 500 μL to 2 mL) and is most easily implemented as it does not require additional equipment. AcS offers a fast extraction with high purity, and if the equipment is available, with low-cost consumables, easy to integrate in a laboratory and to operate, but the final concentration, yield, and preconcentration factor were found to be comparatively low. It is worth mentioning that the yield of extraction is close to the one from UC. LoDF also depends on a specialized, compact equipment along with proprietary consumables, and produces high yields of proteins and nanoparticles, with reasonable time-to-result, an easy-to-use range speed, and easy-to-use technology, but suffers from a low volume preconcentration factor and a comparatively high contamination. The choice of EDTA or citrate tubes for blood collection overall had little impact on the performance of the various enrichment methods, except for UC, and possibly for SEC. Indeed, citrate plasma compared to EDTA plasma yielded a higher final concentration of EVs with UC, see Figure 5, while the purity of EVs extracted by SEC was higher when judged based on TEM images, see Figure S4.

This study further documented how the choice of a size-based method of extraction could change the measured protein profile of the extracted EVs, suggesting that specific subpopulations of EVs are preferentially enriched. One plausible explanation for the observed change of protein profile is the variations in protein expression as a function of EV size, underscoring the potential impact of the size cutoff of the size-based purification and extraction techniques. Lacking a standard and a “ground truth”, it is not possible to identify whether a method introduces a bias for any given protein; comparisons could only be made relative to one another and hence were not included in Figure 7. Interestingly, the subpopulation (by protein expression) of EVs enriched depended on the extraction protocol and on the sample types (cell supernatant or plasma), as revealed by the different hierarchical self-clustering based on protein expression (compare Figure 3 and Figure 6). These observations were also made by Dong et al.30 regarding the expression of CD81 in EVs in the cell supernatant versus plasma. Paget et al.29 noticed differences in the expression of common EV proteins in-between the methods of extraction. A high expression of proteins, such as tetraspanins and integrins, can hide disparity in the expression of rarer proteins. Quantifying protein expression and coexpression beyond the canonical tetraspanins is important to consider when evaluating a processing pipeline as it can enrich or deplete certain subpopulations. Differences in protein expression and subpopulation enrichment were observed for the cell supernatant based on the antibody microarray heatmap profiles. For plasma samples, the heatmap profiles showed less variation. Note that the microarray measured both spiked EVs and native EVs in the plasma that could not be distinguished, see Figure 6. Hence, the high CD63 expression of the spiked-in cell supernatant was not reflected in the measurements in plasma that showed high CD9, and in many cases high integrin α6, suggesting that native EVs were more abundant or possible matrix effects. Matrix effects could also reduce the overall sensitivity and account for other differences. The effect of EDTA or citrate collection tubes did also lead to heatmap intensity changes in protein expression profiles, but visually the overall pattern across methods of extraction was preserved, except for LoDF. When using citrate tubes, the expression of CD9 was markedly reduced compared to all other measurements. Further experiments will be needed to confirm these results and to identify the mechanisms that may lead to differences in the abundance of EV subpopulations.

A limitation to our study is that we did not consider or evaluate the impact of sample storage and of different storage conditions. The four methods of extraction were performed on the same day following the same initial sample toward minimizing any influence of storage, which will need to be studied given that many samples are stored prior to analysis. A better understanding and comparison of the available methods of extraction will help standardization of protocols, a need that is also reflected in analysis of the impact of storage conditions on EV populations. Moreover, this study could be expanded to additional separation techniques. This worked allowed us to study in depth the principle of each method, but could be completed in a second study with the exploration of combination of techniques.

Conclusions

This study demonstrates the importance of preanalytical parameters in EV studies when working with either cell supernatant or plasma samples, and their impact on subsequential analysis, with a focus on protein expression analysis. We worked with multiple pipelines for sample collection and preparation, designed to match the usual protocols from hospitals and laboratories, combined with four systems of extraction. We provide here new data on the comparison of different combinations of collection and extraction parameters, showing that preanalytical parameters directly influence the studied EVs.

We introduce here new keys for characterization and comparison of the preanalytical parameters, to help scientists build their own pipeline of EV extraction, while being aware of the compromises that need to be made and their impact on subsequent analysis. We believe that this step is essential to understand the bias introduced by the preparation of samples and it should be part of any EV analysis. The optimal sample collection protocol may not be a general characteristic but one that depends on the purification method, and of course on the intended use and measurement method. Ignoring the possible effects of sample processing could lead to inconsistent results even when only apparently minor changes are made, and could possibly lead to irreproducible or contradictory results. Preanalytical parameters must be considered as part of the entire process, including which extraction method and which type of EV subpopulation are to be considered.

Acknowledgments

The authors would like to thank the team of J. Rak for supplying the A431-CD63-GFP cell line used in this work.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acssensors.3c02070.

  • Detailed materials and methods for the characterization of EVs, supplementary table on the volumes of samples and figures (TEM, proteins assay, absorption and size distribution) (PDF)

This project has received funding from the European Union’s Horizon 2020 research and innovation program under the Marie Skodowska-Curie grant agreement no. 896313.

The authors declare no competing financial interest.

Notes

A preliminary version of this paper appeared as a preprint: L.A.; M.L.S.; L.O.F.d.A.; J.R.; P.D.-M.; R.M.; A.N.; D.J. Effect of sample preprocessing and extraction methods on the physical and molecular profiles of extracellular vesicles biorxiv, 2023 10.1101/2023.07.17.549204

Supplementary Material

se3c02070_si_001.pdf (1MB, pdf)

References

  1. Colombo M.; Raposo G.; Théry C. Biogenesis, Secretion, and Intercellular Interactions of Exosomes and Other Extracellular Vesicles. Annu. Rev. Cell Dev. Biol. 2014, 30 (1), 255–289. 10.1146/annurev-cellbio-101512-122326. [DOI] [PubMed] [Google Scholar]
  2. Skog J.; Würdinger T.; van Rijn S.; Meijer D. H.; Gainche L.; Sena-Esteves M.; Curry W. T.; Carter B. S.; Krichevsky A. M.; Breakefield X. O. Glioblastoma Microvesicles Transport RNA and Proteins That Promote Tumour Growth and Provide Diagnostic Biomarkers. Nat. Cell Biol. 2008, 10 (12), 1470–1476. 10.1038/ncb1800. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Xue V. W.; Yang C.; Wong S. C. C.; Cho W. C. S. Proteomic Profiling in Extracellular Vesicles for Cancer Detection and Monitoring. Proteomics 2021, 21 (13–14), 2000094 10.1002/pmic.202000094. [DOI] [PubMed] [Google Scholar]
  4. Rontogianni S.; Synadaki E.; Li B.; Liefaard M. C.; Lips E. H.; Wesseling J.; Wu W.; Altelaar M. Proteomic Profiling of Extracellular Vesicles Allows for Human Breast Cancer Subtyping. Commun. Biol. 2019, 2 (1), 1–13. 10.1038/s42003-019-0570-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Lee S.; Mankhong S.; Kang J.-H. Extracellular Vesicle as a Source of Alzheimer’s Biomarkers: Opportunities and Challenges. Int. J. Mol. Sci. 2019, 20 (7), 1728 10.3390/ijms20071728. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Zhao Y.; Yang G. Potential of Extracellular Vesicles in the Parkinson’s Disease – Pathological Mediators and Biomarkers. Neurochem. Int. 2021, 144, 104974 10.1016/j.neuint.2021.104974. [DOI] [PubMed] [Google Scholar]
  7. Andaluz Aguilar H.; Iliuk A. B.; Chen I.-H.; Tao W. A. Sequential Phosphoproteomics and N-Glycoproteomics of Plasma-Derived Extracellular Vesicles. Nat. Protoc. 2020, 15 (1), 161–180. 10.1038/s41596-019-0260-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Vagner T.; Spinelli C.; Minciacchi V. R.; Balaj L.; Zandian M.; Conley A.; Zijlstra A.; Freeman M. R.; Demichelis F.; De S.; Posadas E. M.; Tanaka H.; Vizio D. D. Large Extracellular Vesicles Carry Most of the Tumour DNA Circulating in Prostate Cancer Patient Plasma. J. Extracell. Vesicles 2018, 7 (1), 1505403 10.1080/20013078.2018.1505403. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Minciacchi V. R.; You S.; Spinelli C.; Morley S.; Zandian M.; Aspuria P.-J.; Cavallini L.; Ciardiello C.; Reis Sobreiro M.; Morello M.; Kharmate G.; Jang S. C.; Kim D.-K.; Hosseini-Beheshti E.; Tomlinson Guns E.; Gleave M.; Gho Y. S.; Mathivanan S.; Yang W.; Freeman M. R.; Di Vizio D. Large Oncosomes Contain Distinct Protein Cargo and Represent a Separate Functional Class of Tumor-Derived Extracellular Vesicles. Oncotarget 2015, 6 (13), 11327–11341. 10.18632/oncotarget.3598. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Sidhom K.; Obi P. O.; Saleem A. A Review of Exosomal Isolation Methods: Is Size Exclusion Chromatography the Best Option?. Int. J. Mol. Sci. 2020, 21 (18), 6466 10.3390/ijms21186466. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Yang Y.; Wang Y.; Wei S.; Zhou C.; Yu J.; Wang G.; Wang W.; Zhao L. Extracellular Vesicles Isolated by Size-Exclusion Chromatography Present Suitability for RNomics Analysis in Plasma. J. Transl. Med. 2021, 19 (1), 104 10.1186/s12967-021-02775-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Zhang P.; Zhou X.; He M.; Shang Y.; Tetlow A. L.; Godwin A. K.; Zeng Y. Ultrasensitive Detection of Circulating Exosomes with a 3D-Nanopatterned Microfluidic Chip. Nat. Biomed. Eng. 2019, 3 (6), 438–451. 10.1038/s41551-019-0356-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Kanwar S. S.; James Dunlay C.; Simeone D. M.; Nagrath S. Microfluidic Device (ExoChip) for on-Chip Isolation, Quantification and Characterization of Circulating Exosomes. Lab Chip 2014, 14 (11), 1891–1900. 10.1039/C4LC00136B. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Konoshenko M. Y.; Lekchnov E. A.; Bryzgunova O. E.; Kiseleva E.; Pyshnaya I. A.; Laktionov P. P. Isolation of Extracellular Vesicles from Biological Fluids via the Aggregation–Precipitation Approach for Downstream miRNAs Detection. Diagnostics 2021, 11 (3), 384 10.3390/diagnostics11030384. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Liang L.-G.; Kong M.-Q.; Zhou S.; Sheng Y.-F.; Wang P.; Yu T.; Inci F.; Kuo W. P.; Li L.-J.; Demirci U.; Wang S. An Integrated Double-Filtration Microfluidic Device for Isolation, Enrichment and Quantification of Urinary Extracellular Vesicles for Detection of Bladder Cancer. Sci. Rep. 2017, 7 (1), 46224 10.1038/srep46224. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Wu M.; Ouyang Y.; Wang Z.; Zhang R.; Huang P.-H.; Chen C.; Li H.; Li P.; Quinn D.; Dao M.; Suresh S.; Sadovsky Y.; Huang T. J. Isolation of Exosomes from Whole Blood by Integrating Acoustics and Microfluidics. Proc. Natl. Acad. Sci. U.S.A. 2017, 114 (40), 10584–10589. 10.1073/pnas.1709210114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Liu C.; Guo J.; Tian F.; Yang N.; Yan F.; Ding Y.; Wei J.; Hu G.; Nie G.; Sun J. Field-Free Isolation of Exosomes from Extracellular Vesicles by Microfluidic Viscoelastic Flows. ACS Nano 2017, 11 (7), 6968–6976. 10.1021/acsnano.7b02277. [DOI] [PubMed] [Google Scholar]
  18. Liu C.; Zhao J.; Tian F.; Cai L.; Zhang W.; Feng Q.; Chang J.; Wan F.; Yang Y.; Dai B.; Cong Y.; Ding B.; Sun J.; Tan W. Low-Cost Thermophoretic Profiling of Extracellular-Vesicle Surface Proteins for the Early Detection and Classification of Cancers. Nat. Biomed. Eng. 2019, 3 (3), 183–193. 10.1038/s41551-018-0343-6. [DOI] [PubMed] [Google Scholar]
  19. Woo H.-K.; Sunkara V.; Park J.; Kim T.-H.; Han J.-R.; Kim C.-J.; Choi H.-I.; Kim Y.-K.; Cho Y.-K. Exodisc for Rapid, Size-Selective, and Efficient Isolation and Analysis of Nanoscale Extracellular Vesicles from Biological Samples. ACS Nano 2017, 11 (2), 1360–1370. 10.1021/acsnano.6b06131. [DOI] [PubMed] [Google Scholar]
  20. Ku A.; Lim H. C.; Evander M.; Lilja H.; Laurell T.; Scheding S.; Ceder Y. Acoustic Enrichment of Extracellular Vesicles from Biological Fluids. Anal. Chem. 2018, 90, 8011. 10.1021/acs.analchem.8b00914. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Bryl-Górecka P.; Sathanoori R.; Al-Mashat M.; Olde B.; Jögi J.; Evander M.; Laurell T.; Erlinge D. Effect of Exercise on the Plasma Vesicular Proteome: A Methodological Study Comparing Acoustic Trapping and Centrifugation. Lab Chip 2018, 18 (20), 3101–3111. 10.1039/C8LC00686E. [DOI] [PubMed] [Google Scholar]
  22. Zijlstra A.; Di Vizio D. Size Matters in Nanoscale Communication. Nat. Cell Biol. 2018, 20 (3), 228–230. 10.1038/s41556-018-0049-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Théry C.; Witwer K. W.; Aikawa E.; Alcaraz M. J.; Anderson J. D.; Andriantsitohaina R.; Antoniou A.; Arab T.; Archer F.; Atkin-Smith; et al. Minimal Information for Studies of Extracellular Vesicles 2018 (MISEV2018): A Position Statement of the International Society for Extracellular Vesicles and Update of the MISEV2014 Guidelines. J. Extracell. Vesicles 2018, 7 (1), 1535750 10.1080/20013078.2018.1535750. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Mathivanan S.; Simpson R. J. ExoCarta: A Compendium of Exosomal Proteins and RNA. Proteomics 2009, 9 (21), 4997–5000. 10.1002/pmic.200900351. [DOI] [PubMed] [Google Scholar]
  25. Van Deun J.; Mestdagh P.; Agostinis P.; Akay Ö.; Anand S.; Anckaert J.; Martinez Z. A.; Baetens T.; Beghein E.; Bertier L.; et al. EV-TRACK: Transparent Reporting and Centralizing Knowledge in Extracellular Vesicle Research. Nat. Methods 2017, 14 (3), 228–232. 10.1038/nmeth.4185. [DOI] [PubMed] [Google Scholar]
  26. Jeon H.; Kang S.-K.; Lee M.-S. Effects of Different Separation Methods on the Physical and Functional Properties of Extracellular Vesicles. PLoS One 2020, 15 (7), e0235793 10.1371/journal.pone.0235793. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Liang Y.; Lehrich B. M.; Zheng S.; Lu M. Emerging Methods in Biomarker Identification for Extracellular Vesicle-Based Liquid Biopsy. J. Extracell. Vesicles 2021, 10 (7), e12090 10.1002/jev2.12090. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Clayton A.; Boilard E.; Buzas E. I.; Cheng L.; Falcón-Perez J. M.; Gardiner C.; Gustafson D.; Gualerzi A.; Hendrix A.; Hoffman A.; et al. Considerations towards a Roadmap for Collection, Handling and Storage of Blood Extracellular Vesicles. J. Extracell. Vesicles 2019, 8 (1), 1647027 10.1080/20013078.2019.1647027. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Paget D.; Checa A.; Zöhrer B.; Heilig R.; Shanmuganathan M.; Dhaliwal R.; Johnson E.; Jørgensen M. M.; Bæk R.; Wheelock C. E.; Channon K. M.; Fischer R.; Anthony D. C.; Choudhury R. P.; Akbar N. Comparative and Integrated Analysis of Plasma Extracellular Vesicle Isolation Methods in Healthy Volunteers and Patients Following Myocardial Infarction. J. Extracell. Biol. 2022, 1 (11), e66 10.1002/jex2.66. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Dong L.; Zieren R. C.; Horie K.; Kim C.-J.; Mallick E.; Jing Y.; Feng M.; Kuczler M. D.; Green J.; Amend S. R.; Witwer K. W.; de Reijke T. M.; Cho Y.-K.; Pienta K. J.; Xue W. Comprehensive Evaluation of Methods for Small Extracellular Vesicles Separation from Human Plasma, Urine and Cell Culture Medium. J. Extracell. Vesicles 2020, 10 (2), e12044 10.1002/jev2.12044. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Droste M.; Tertel T.; Jeruschke S.; Dittrich R.; Kontopoulou E.; Walkenfort B.; Börger V.; Hoyer P. F.; Büscher A. K.; Thakur B. K.; Giebel B. Single Extracellular Vesicle Analysis Performed by Imaging Flow Cytometry and Nanoparticle Tracking Analysis Evaluate the Accuracy of Urinary Extracellular Vesicle Preparation Techniques Differently. Int. J. Mol. Sci. 2021, 22 (22), 12436 10.3390/ijms222212436. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Martel R.; Shen M. L.; DeCorwin-Martin P.; de Araujo L. O. F.; Juncker D. Extracellular Vesicle Antibody Microarray for Multiplexed Inner and Outer Protein Analysis. ACS Sens. 2022, 7 (12), 3817–3828. 10.1021/acssensors.2c01750. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Montermini L.; Meehan B.; Garnier D.; Lee W. J.; Lee T. H.; Guha A.; Al-Nedawi K.; Rak J. Inhibition of Oncogenic Epidermal Growth Factor Receptor Kinase Triggers Release of Exosome-like Extracellular Vesicles and Impacts Their Phosphoprotein and DNA Content. J. Biol. Chem. 2015, 290 (40), 24534–24546. 10.1074/jbc.M115.679217. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Weatherall E.; Hauer P.; Vogel R.; Willmott G. R. Pulse Size Distributions in Tunable Resistive Pulse Sensing. Anal. Chem. 2016, 88 (17), 8648–8656. 10.1021/acs.analchem.6b01818. [DOI] [PubMed] [Google Scholar]
  35. Stewart R. L.; O’Connor K. L. Clinical Significance of the Integrin A6β4 in Human Malignancies. Lab. Invest. 2015, 95 (9), 976–986. 10.1038/labinvest.2015.82. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Garnier D.; Magnus N.; Meehan B.; Kislinger T.; Rak J. Qualitative Changes in the Proteome of Extracellular Vesicles Accompanying Cancer Cell Transition to Mesenchymal State. Exp. Cell Res. 2013, 319 (17), 2747–2757. 10.1016/j.yexcr.2013.08.003. [DOI] [PubMed] [Google Scholar]
  37. Miyanishi M.; Tada K.; Koike M.; Uchiyama Y.; Kitamura T.; Nagata S. Identification of Tim4 as a Phosphatidylserine Receptor. Nature 2007, 450 (7168), 435–439. 10.1038/nature06307. [DOI] [PubMed] [Google Scholar]
  38. Xu H.; Liao C.; Zuo P.; Liu Z.; Ye B.-C. Magnetic-Based Microfluidic Device for On-Chip Isolation and Detection of Tumor-Derived Exosomes. Anal. Chem. 2018, 90 (22), 13451–13458. 10.1021/acs.analchem.8b03272. [DOI] [PubMed] [Google Scholar]
  39. Nakai W.; Yoshida T.; Diez D.; Miyatake Y.; Nishibu T.; Imawaka N.; Naruse K.; Sadamura Y.; Hanayama R. A Novel Affinity-Based Method for the Isolation of Highly Purified Extracellular Vesicles. Sci. Rep. 2016, 6 (1), 33935 10.1038/srep33935. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Rikkert L. G.; Nieuwland R.; Terstappen L. W. M. M.; Coumans F. A. W. Quality of Extracellular Vesicle Images by Transmission Electron Microscopy Is Operator and Protocol Dependent. J. Extracell. Vesicles 2019, 8 (1), 1555419 10.1080/20013078.2018.1555419. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Gaspar L. S.; Santana M. M.; Henriques C.; Pinto M. M.; Ribeiro-Rodrigues T. M.; Girão H.; Nobre R. J.; Pereira de Almeida L. Simple and Fast SEC-Based Protocol to Isolate Human Plasma-Derived Extracellular Vesicles for Transcriptional Research. Mol. Ther.--Methods Clin. Dev. 2020, 18, 723–737. 10.1016/j.omtm.2020.07.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Kim A.; Ng W. B.; Bernt W.; Cho N.-J. Validation of Size Estimation of Nanoparticle Tracking Analysis on Polydisperse Macromolecule Assembly. Sci. Rep. 2019, 9 (1), 2639 10.1038/s41598-019-38915-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Akers J. C.; Ramakrishnan V.; Nolan J. P.; Duggan E.; Fu C.-C.; Hochberg F. H.; Chen C. C.; Carter B. S. Comparative Analysis of Technologies for Quantifying Extracellular Vesicles (EVs) in Clinical Cerebrospinal Fluids (CSF). PLoS One 2016, 11 (2), e0149866 10.1371/journal.pone.0149866. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Mathieu M.; Névo N.; Jouve M.; Valenzuela J. I.; Maurin M.; Verweij F. J.; Palmulli R.; Lankar D.; Dingli F.; Loew D.; Rubinstein E.; Boncompain G.; Perez F.; Théry C. Specificities of Exosome versus Small Ectosome Secretion Revealed by Live Intracellular Tracking of CD63 and CD9. Nat. Commun. 2021, 12 (1), 4389 10.1038/s41467-021-24384-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Bæk R.; Søndergaard E. K. L.; Varming K.; Jørgensen M. M. The Impact of Various Preanalytical Treatments on the Phenotype of Small Extracellular Vesicles in Blood Analyzed by Protein Microarray. J. Immunol. Methods 2016, 438, 11–20. 10.1016/j.jim.2016.08.007. [DOI] [PubMed] [Google Scholar]
  46. Gelibter S.; Marostica G.; Mandelli A.; Siciliani S.; Podini P.; Finardi A.; Furlan R. The Impact of Storage on Extracellular Vesicles: A Systematic Study. J. Extracell. Vesicles 2022, 11 (2), e12162 10.1002/jev2.12162. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

se3c02070_si_001.pdf (1MB, pdf)

Articles from ACS Sensors are provided here courtesy of American Chemical Society

RESOURCES