Comparison of Generic Fluorescent Markers for Detection of Extracellular Vesicles by Flow Cytometry.
We considered including calcein acetoxymethyl ester (calcein AM), calcein acetoxymethyl ester violet (calcein violet), carboxyfluorescein succinimidyl ester (CFDA-SE, here and commonly referred to as CFSE), 4-(2-[6-(dioctylamino)-2-naphthalenyl]ethenyl)-1(3-sulfopropyl)pyridinium (di-8-ANEPPS), annexin V, lactadherin, and Paul Karl Horan 67 (PKH67) (8, 9, 11-15). Table 1 shows an overview of the reported properties and results for these markers, except calcein violet and lactadherin, which have not been used as generic markers previously. Calcein AM (495/515 nm absorption/emission) and calcein violet (400/452 nm) are both membrane permeable and nonfluorescent, requiring hydrolysis to be converted into a membrane-impermeable fluorescent analog (11, 16). Because intravesicular esterases catalyze the hydrolysis, both markers label intact EVs only (11). Furthermore, calcein violet has limited spectral overlap with commonly applied fluorophores, which simplifies combining calcein violet with immunofluorescent labeling. Like calcein AM, CFDA-SE is also nonfluorescent and requires hydrolysis to be converted to CFSE (492/517 nm), which then binds covalently to free amines (17). Di-8-ANEPPS (467/631 nm) is nonfluorescent in solution and becomes fluorescent when bound to phospholipid membranes and possibly proteins (9, 18). Annexin V binds to membrane phosphatidylserine in the presence of calcium. Because calcium may trigger clotting ofplasma or blood, we used lactadherin since (a) both annexin V and lactadherin bind to phosphatidylserine, (b) the affinity of lactadherin for phosphatidylserine is higher than that of annexin V (19), and (c) binding of lactadherin to phosphatidylserine is calcium independent. Both annexin V and lactadherin, however, are not true generic markers because phosphatidylserine is not exposed on all EVs (20). This also holds true for calcein AM, calcein violet, and CFSE, as it is uncertain whether all EVs efficiently hydrolyze the markers. Therefore, for convenience, all markers tested are termed "generic markers." Finally, PKH67 was not evaluated because the labeling protocol is longer than 8 h and not clinically applicable (8, 21).
The aforementioned generic markers for labeling EVs were evaluated on various samples and flow cytometers (Table 1). Using flow cytometry to detect epithelial cell adhesion molecule positive ([EpCAM.sup.+]) EVs from MCF7 cell culture and integrin /33 positive (CD61) EVs from human plasma, we compared the performance of the most promising generic markers, namely, calcein AM, calcein violet, CFSE, di-8-ANEPPS, and lactadherin, and used side scatter as a reference. We also evaluated the influence of common coisolates of EVs and marker micelles or aggregates on the detection of EVs with generic markers.
Materials and Methods
Calcein AM and calcein violet were obtained from eBioscience. CFSE, Pluronic F-127, RPMI-1640 medium, fetal bovine serum, L-glutamine, penicillin, and streptomycin were all from ThermoFisher Scientific. Di-8-ANEPPS and BSA (purity [greater than or equal to]98%) were from Sigma-Aldrich, and lactadherin--fluorescein isothiocyanate (FITC) was from Hematologic Technologies. PBS was from Fresenious Kabi, and VLDLs were from Merck Millipore. Allophycocyanin (APC)-conjugated EpCAM (mouse IgC1[kappa], clone HEA-125) was from Miltenyi Biotec, CD61-APC (mouse IgG1[kappa], clone Y2/51) was from Dako, and IgG1 isotype control was from BD. The buffer used in all experiments, including size-exclusion chromatography (SEC), was PBS [154 mmol/L NaCl, 1.24 mmol/L [Na.sub.2]HP[O.sub.4].2[H.sub.2]O, 0.2 mmol/L Na[H.sub.2]P[O.sub.4].2[H.sub.2]O, pH 7.4; 50 nm filtered (Nuclepore, GE Healthcare)], supplemented with 0.32% trisodium citrate.
MCF7 breast cancer cells were cultured in RPMI-1640 medium supplemented with 10% (v/v) fetal bovine serum, 2 mmol/L L-glutamine, 10 U/mL penicillin, and 10 [micro]g/mL streptomycin. Cells were cultured in a T75 culture flask (Corning) at 37 [degrees]C, 5% CO2. At 80% to 90% confluence, cells were washed with PBS and cultured in fetal bovine serum-free medium before harvesting EVs. After 48 h, the conditioned culture medium was centrifuged at 1000gfor 30 min (Rotina46 RS, Hettich) to remove cells. We call the resulting supernatant MCF7 EVs throughout the text.
To prepare platelet-free plasma, citrate-anticoagulated blood (0.32% final concentration) from 10 healthy donors was collected as described elsewhere (22). Informed consent and approval from the ethics committee was obtained. The blood was pooled, and plasma was prepared by centrifuging the blood twice at 1560g, 20 [degrees]C for 20 min.
Both MCF7 EVs and the plasma sample were snapfrozen in liquid nitrogen, stored at -80 [degrees]C, and thawed at 37 [degrees]C before use. Figs. 1 and 2 in the Data Supplement that accompanies the online version of this article at http:// www.clinchem.org/content/vol64/issue4 show the characterization of the MCF7 EVs and plasma sample.
PREPARATION OF NON-EV COMPONENTS
Non-EV components in plasma include proteins and lipoproteins. Because liposomes are currently being investigated as reference particles for EV measurements (9), we also studied whether generic markers stained liposomes. We found, however, that the composition of our liposomes was unsuitable for evaluation of generic markers (see Table 1 of the online Data Supplement) and, therefore, excluded the results from any further discussion.
Non-EV components studied here include marker in buffer, a protein solution containing 40 g/L BSA and lipoproteins. Fig. 3 of the online Data Supplement provides the size and concentration information of the non-EV samples.
To investigate the impact of protein on staining of EVs by the generic markers, protein was removed using SEC before generic marker staining and compared with untreated samples. SEC columns (qEVsingle, Izon Science; 3.5 mL Sepharose CL-2B) were washed once with 6 mL of buffer before loading 100 [micro]L of plasma, followed by buffer elution. The first 1 mL of eluate was discarded, after which the 0.5-mL fraction containing EVs was collected (23).
Sample incubation conditions were based on literature (9, 11, 12): 20 min at 37 [degrees]C for calcein AM and calcein violet, 60 min at 37 [degrees]C for CFSE, and 60 min at room temperature for di-8-ANEPPS and lactadherin. Generic markers were titrated on MCF7 EVs (see Fig. 4 of the online Data Supplement). A 10-[micro]L sample containing EVs or non-EV components was incubated with either 10 [micro]L of calcein AM (final concentration, 20 pmol/L), calcein violet (80 pmol/L), CFSE (40 pmol/L), di-8ANEPPS (0.5 pmol/L) with 0.05% Pluronic F-127, lactadherin (8.3 pg/mL), or buffer.
MCF7 EVs were identified by exposure of EpCAM (24); platelet EVs, by CD61 (25). APC-conjugated antibodies were used for immunostaining because of negligible spectral overlap with the evaluated generic markers. Next, 2.5 [micro]L of EpCAM-APC (22 pg/mL), CD61-APC (25 pg/mL), or IgG1 isotype control at a matching concentration was added to 10 [micro]L of generic marker-stained sample and incubated for 2 h at room temperature in the dark. To remove aggregates, antibodies and generic markers were centrifuged at 19000g for 5 min before use.
After incubation, samples were diluted in buffer to event rates below approximately 5000/s to prevent swarm when triggering on side scatter (see Table 2 and Fig. 5 of the online Data Supplement). Aseparate control experiment demonstrated that >96% of antibody positive ([mAb.sup.+]) particles in the EV samples lysed upon detergent treatment, indicating that most [mAb.sup.+] particles were probably EVs (see Figs. 1 and 2 of the online Data Supplement).
All samples were analyzed on an A60-Micro (Apogee) at a flow rate of 3.0 [micro]L/min. Samples were measured for 4 min when triggering on EpCAM-APC, CD61-APC, or generic marker fluorescence and for 1 min when triggering on side scatter (405-nm laser). Fluorescence-triggered measurements were longer to ensure a statistically relevant number of detected particles. Trigger thresholds were set to a value resulting in 10 to 20 counts/s in buffer. Gates were based on isotype controls without generic marker. For both mAb and generic marker, positive (+) is defined as a fluorescent signal exceeding the gate. Particle concentrations were calculated by compensating the number of detected particles for flow rate, measurement time, and sample dilution. All concentrations were corrected using acquisition time-matched plain buffer or IgG1 controls to estimate the number of sample-specific events. A full description of the flow cytometer configuration, operating conditions, and data analysis is found in the MIFlowCyt list of the online Data Supplement.
For each generic marker, 6 repeated measurements were performed for both EV samples and 2 measurements for the non-EV samples. In all, 115 of 124 EV sample measurements were successful, with a minimum of 3 successful measurements per generic marker. Reported results are the mean of individually stained samples.
CHARACTERIZATION OF FLOW CYTOMETER ANALYTICAL SENSITIVITY
The resolution limit of the Apogee A60-Micro was 74 phycoerythrin (PE) molecules ofequivalent soluble fluorochrome (MESF), 304 FITC MESF, and 16 APC MESF. We considered the resolution limit for APC unreliable (details provided in the online Data Supplement). Scatter analytical sensitivity was characterized by reporting the smallest polystyrene bead distinguishable from noise, which was 100-nm polystyrene on our flow cytometer.
Throughout this study, the performance of the generic markers was expressed in terms of sensitivity and positive predictive value (PPV) for EV detection. Unless stated otherwise, sensitivity is defined as the percentage of [mAb.sup.+] EVs detectably stained by the generic marker out of the total [mAb.sup.+] EV population. This definition hereby differs from those used for clinical and analytical sensitivity. The PPV is defined here as the percentage of [mAb.sup.+] EVs detectably stained by the generic marker out of the total [marker.sup.+] population. Sensitivity and PPV were calculated as:
sensitivity = [[mAb.sup.+][marker.sup.+]]/ [mAb.sup.+] * 100% (1)
PPV = [[mAb.sup.+][marker.sup.+]]/ [marker.sup.+] * 100% (2)
with [[mAb.sup.+][marker.sup.+]] the detected concentration of [mAb.sup.+] particles while triggering on the generic marker, [[mAb.sup.+]] the detected concentration of [mAb.sup.+] particles while triggering on the antibody, and [[marker.sup.+]] the detected concentration of [marker.sup.+] particles while triggering on the generic marker. It was not possible to calculate the specificity of the generic markers because markerparticles (i.e., true negatives) were not detected when triggering on fluorescence. Instead, the PPV was calculated and used as a measure for specificity.
To evaluate the significance of differences between plain buffer (n = 10) and non-EV components (n = 2), a 1-sided (because buffer has the fewest particles), 1-sample t-test was applied to determine P.
GENERIC MARKER PERFORMANCE ON CELL CULTURE EVs
First, the performance of the generic markers was evaluated on EVs from MCF7 cells. EVs were identified by EpCAM-APC, and we assumed that all EpCAM binding ([mAb.sup.+]) particles were true EVs. Fig. 1, A-E, shows representative scatter plots of MCF7 EVs triggered by fluorescence of the 5 generic markers. Side scatter triggered measurements are included for reference (Fig. 1F). Fig. 6 in the online Data Supplement shows the results for buffer only, marker in buffer, and isotype controls.
Fig. 1, G-L, shows Venn diagrams that were calculated using the concentrations of particles in the upper and lower right quadrants of the scatter plots, combined with the concentration of total [EpCAM.sup.+] particles from EpCAM-APC triggered measurements (Table 4 of the online Data Supplement shows the raw numbers, and Fig. 7 of the online Data Supplement provides a calculation of the Venn diagrams). The diagrams show the concentrations of missed EVs ([mAb.sup.+][marker.sup.-]), detected EVs ([mAb.sup.+][marker.sup.+]), and possibly falsely detected EVs ([mAb.sup.-][marker.sub.+).] The sensitivity to detect EpCAM particles was calculated from Fig. 1, G-L, as shown in Fig. 1M and previously described in the Methods section.
Table 2 shows the sensitivity of each generic marker and side scatter to detect EVs. CFSE suffered from swarm at all dilutions, making these results unreliable (see Fig. 5 of the online DataSupplement). Ideally, a generic marker would exclusively stain EVs. Because not all [marker.sup.+] particles are also [EpCAM.sup.+] (Fig. 1, A-E), the PPV of generic markers was also calculated. For interpretation, a PPV of 29% indicates that 29% of the [marker.sup.+] events are also [mAb.sub.+,] and thus EVs by our definition. Table 2 shows that all generic markers had a PPV <100%. Taken together, di-8-ANEPPS and lactadherin were the most promising generic markers for MCF7 EVs based on sensitivity and PPV. Subsequently, the ability of generic markers to detect EVs in plasma was studied because plasma contains non-EV particles that may interfere with detection of EVs by generic markers.
GENERIC MARKER PERFORMANCE ON PLASMA EVs
Plasma contains various populations of EVs. Here we chose to study platelet EVs, assuming that detection of platelet EVs was representative of detection of all EV populations in plasma. Platelet EVs were immunostained with CD61-APC, assuming that all CD61-APC-binding ([mAb.sup.+]) particles were true EVs.
In analogy to MCF7 EVs (Fig. 1), we prepared Fig. 2 for platelet EVs in plasma and Table 4 in the online Data Supplement for the raw numbers. The sensitivity for detecting [CD61.sup.+] particles was calculated from Fig. 2, G--L, as shown in Fig. 2S and described in the Methods section, and presented in Table 2. Again, CFSE results were unreliable because of swarm. Approximately 30% of all plasma EVs expose CD61 (20). Therefore, a PPV of 30% for [CD61.sup.+] particles in plasma was considered a perfect performance, i.e., 30% of all [marker.sup.+] particles would also be [CD61.sub.+]. Table 2 shows the PPV for each generic marker and side scatter, as calculated from the Venn diagrams (Fig. 2, G--L). In summary, lactadherin and side scatter detected [CD61.sup.+] particles in plasma. The low PPVs indicated that the generic markers also stained mAb particles, which could be EVs with mAb fluorescence below the detection limit of our flow cytometer, other EV populations, or non-EV particles. Therefore, we studied whether generic markers stained non-EV particles.
INFLUENCE OF NON-EV COMPONENTS
Suspensions of EVs isolated from plasma often contain non-EV components (26). Table 3 shows the particles detected when the indicated marker was added to buffer, proteins, and lipoproteins. Because the particle concentration differed in each sample, the concentrations cannot be compared between columns. All generic markers and side scatter detected particle concentrations higher than the background (plain buffer) in the marker in buffer sample (P < 0.001) and in the sample containing protein (P < 0.0001). Only CFSE and side scatter detected particles in excess ofbackground in the lipoprotein sample (P = 0.002 and P < 0.001, respectively). Our protein sample contained a physiological concentration of albumin (27). Therefore, binding of the generic markers to plasma proteins may explain their low sensitivity to detect EVs in plasma.
DETECTION OF PLASMA EVs BY GENERIC MARKERS AFTER PROTEIN REMOVAL
Because proteins were stained by the generic markers (Table 3), plasma proteins were removed by SEC before labeling. Fig. 2, M--R, shows the resulting Venn diagrams, from which the sensitivity and PPV for [CD61.sup.+] particles were calculated (Table 2). Table 2 shows that protein removal before staining improved the sensitivity and PPV by 10- and 9-fold for calcein AM, 1- and 6-fold for CFSE, and 3- and 6-fold for di-8-ANEPPS, whereas the sensitivity and PPV reduced 0.3- and 0.2-fold for calcein violet and 0.1- and 0.7-fold for lactadherin. As a reference, for side scatter the sensitivity reduced 0.2-fold and the PPV increased 3-fold after protein removal. Thus, protein removal improved the detection of [CD61.sup.+] particles in plasma for calcein AM, CFSE, and di-8-ANEPPS. Again, CFSE suffered from swarm; therefore, obtained results are unreliable.
COMPARISON OF BEST PERFORMING GENERIC MARKERS WITH SIDE SCATTER
Fig. 3 shows Venn diagrams comparing side scatter of a single sample with the best performing generic markers in this study: di-8-ANEPPS and lactadherin. For MCF7 EVs (Fig. 3, A and B), 94% of [mAb.sup.+] particles were detected using di-8-ANEPPS, 79% using lactadherin, and 100% using side scatter. Also, Fig. 3, A and B, shows that fluorescence-based triggering resulted in 3- to 4-fold lower detection rates of mAb particles compared with side scatter. In plasma (Fig. 3C), lactadherin and side scatter detected 32% and 50% of mAb + particles, respectively, with lactadherin triggering resulting in a 500-fold lower detection rate of [mAb.sup.-] particles compared with side scatter.
The ideal generic marker would be 100% sensitive and specific to EVs. Our results show that none of the tested generic markers fulfills this requirement in plasma. In fact, there was almost no overlap between [marker.sup.+] and [CD61.sup.+] populations (Fig. 2, G--L), likely because of the large number of labeled non-EV components. Our findings (Table 2) emphasize large differences in the performance of generic markers.
To summarize the results, triggering on di-8ANEPPS, lactadherin, or side scatter on our flow cytometer resulted in comparable concentrations of [EpCAM.sup.+] EVs in MCF7 EV samples. In contrast, [CD61.sup.+] EVs in plasma were best detected by side scatter. Of the generic markers, [CD61.sup.+] EVs in plasma were best detected by lactadherin. All generic markers detected particles in "stained" buffer, which could be marker aggregates/ micelles and/or unbound marker that causes an increase in background fluorescence above the threshold. All PPVs were <30%, mostly owing to detection of these marker aggregates/micelles, unbound marker, and stained soluble protein. Removal of proteins from plasma increased the PPV for di-8-ANEPPS and side scatter but, surprisingly, reduced the sensitivity and PPV for lactadherin. The cause of this reduction is unknown. Side scatter triggering resulted in a detected concentration of 5.6 [+ or -] 1.2 * [10.sup.7] [CD61.sup.+] EVs/mL plasma of healthy controls, comparable with the previously reported concentration of platelet EVs (20).
Calcein AM and calcein violet had a low sensitivity and PPV in both EV samples. The low sensitivity may be because of insufficient brightness of the markers or because EVs have insufficient esterase activity. A previous study reported successful staining of EVs with calcein AM (11). The scattered intensity of these EVs, however, was comparable with 1-[micro]m polystyrene beads, which correspond to the scattered intensity of 2- to 3-[micro]m EVs or even cells (28). Consequently, the previously detected EVs were much larger than the population of EVs detected in our present study.
Samples stained with CFSE suffered from swarm detection at low dilutions or insufficient event rates at higher dilutions in both EV samples (see Fig. 5 of the online Data Supplement), an artifact discussed elsewhere (29). Previous reports did not investigate swarm in the CFSE channel (12, 14, 15). Swarm detection may be caused by stained soluble protein or lipoprotein particles, marker micelles/aggregates, and/or unbound marker (Table 3) (14). Application of SEC to reduce soluble proteins or unbound marker did not eliminate the observed swarm (see Fig. 5 of the online Data Supplement).
In this study, we defined EVs as all [mAb.sup.+] particles. This definition excluded EVs that were mAb or mAbdlm, although they may be more abundant than [mAb.sup.+] EVs, as is the case for plasma (20). In addition, non-EV particles that are [mAb.sup.+] and detergent resistant would be considered EVs. Furthermore, the composition of our liposomes was unsuitable for evaluation of generic markers. First, the lipid transition temperature was above room temperature, making the membrane more impermeable to these markers, and second, at least for di-8ANEPPS, the high concentration of cholesterol caused a blue shift in the emission spectrum (see Table 1 of the online Data Supplement) (30).
None of the tested generic markers fulfilled the requirements described earlier for an ideal generic fluorescent marker. Di-8-ANEPPS and lactadherin stained 100% of EVs in MCF7 EV samples, but both markers did not specifically stain EVs. All markers were compatible with fluorescent immunostaining but posed a challenge when designing multicolor panels. Only lactadherin was directly applicable to plasma; the other markers required isolation of EVs before staining. Another way to improve marker performance would be to increase marker concentration and/or remove unbound marker after staining.
Several publications suggest that fluorescence-based triggering is superior to scatter-based triggering (8-10, 13). The outcome of such comparisons, however, depends on the properties of the used flow cytometer and has little bearing on the fundamental superiority of one triggering strategy over the other. Furthermore, although in this study side scatter was most efficient to detect [mAb.sup.+] particles in plasma, the PPV of side scatter is, on average, 200-fold lower than fluorescence triggering on lactadherin (Fig. 3, Table 2).
Our flow cytometer has a high analytical sensitivity on scatter, a PE/FITC fluorescent sensitivity comparable with high sensitivity flow cytometers (9), and is never used to measure whole plasma or blood to keep background counts low. Because most flow cytometers are either analytically sensitive on fluorescence or scatter, compromises must be made. For an informed decision, we suggest the following roadmap: (a) quantify the limit of detection on scatter, e.g., the smallest polystyrene bead that is distinguishable from noise, (b) quantify the limit of detection on fluorescence by determining Qand B, (c) take the sample composition into account, e.g., the presence of proteins, (d) compare limits of detection with reported limits of detection for scatter-sensitive and/or fluorescence-sensitive flow cytometers (9, 14) to select suitable trigger strategies, and (e) test selected trigger strategies on the sample and flow cytometer to determine which strategy detects the highest number of EVs.
Taken together, although the roadmap above can be used to select the optimal trigger strategy and marker for your application, the search for a generic EV marker continues.
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition ofdata, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.
Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:
Employment or Leadership: Z. Varga, Centre for Natural Sciences, HAS and Semmelweis University.
Consultant or Advisory Role: None declared.
Stock Ownership: E. van der Pol, Exometry B.V.
Honoraria: None declared.
Research Funding: L. de Rond, Perspectief CANCER-ID funded by Netherlands Organisation for Scientific Research - Domain Applied and Engineering Sciences (NWO-TTW), Perspectief CANCER-ID 14195; E. van der Pol, Perspectief CANCER-ID funded by Netherlands Organisation for Scientific Research - Domain Applied and Engineering Sciences (NWO-TTW); F.A.W. Coumans, Research programVENI 13681.
Expert Testimony: None declared.
Patents: E. van der Pol, PCT application 15192403.2;T.G van Leeuwen, PCT/EP2016/076238.
Role of Sponsor: The funding organizations played a direct role in the final approval of manuscript. The funding organizations played no role in the design of study, choice of enrolled patients, or review and interpretation of data.
Acknowledgments: The authors thank Linda Rikkert from the Department of Medical Cell BioPhysics, University of Twente, Enschede, the Netherlands, for making the transmission electron microscopy images in this study.
(1.) Yuana Y, Sturk A, Nieuwland R. Extracellularvesicles in physiological and pathological conditions. Blood Rev 2013;27:31-9.
(2.) LoyerX, Vion A-C, Tedgui A, Boulanger CM. Microvesicles ascell-cell messengersincardiovasculardiseases. CircRes 2014;114:345-53.
(3.) Robbins PD, Morelli AE. Regulation of immune responses by extracellular vesicles. Nat Rev Immunol 2014;14:195-208.
(4.) Lowry MC, Gallagher WM, Driscoll L. The role of exosomes in breastcancer. Clin Chem 2015;61:1457-65.
(5.) BabicA, Wolpin BM. Circulating exosomes in pancreatic cancer: will they succeed on the long, littered road to earlydetection marker? Clin Chem 2016;62:307-9.
(6.) Lovren F, Verma S. Evolving role of microparticlesinthe pathophysiology of endothelial dysfunction. Clin Chem 2013;59:1166-74.
(7.) Nolan JP. Flow cytometry of extracellular vesicles: potential, pitfalls, and prospects. Curr Protoc Cytom 2015; 73:13.4.1-6.
(8.) van der Vlist EJ, Nolte-'t Hoen ENM, Stoorvogel W, Arkesteijn GJA, Wauben MHM. Fluorescent labeling of nano-sized vesicles released by cells and subsequent quantitative and qualitative analysis by high-resolution flow cytometry. Nat Protocols 2012;7:1311-26.
(9.) Stoner SA, Duggan E, Condello D, Guerrero A, Turk JR, Narayanan PK, Nolan JP. High sensitivity flow Cytometry of membrane vesicles. Cytometry A2016;89:196- 206.
(10.) Kormelink TG, Arkesteijn GJA, Nauwelaers FA, vanden Engh G, Nolte-'t Hoen ENM, Wauben MHM. Prerequisites forthe analysis and sorting of extracellularvesicle subpopulations by high-resolution flow cytometry. Cytometry A2016;89:135-47.
(11.) Gray WD, Mitchell AJ, Searles CD.An accurate, precise method for general labeling of extracellular vesicles. Methods X2015;2:360 -7.
(12.) Pospichalova V, Svoboda J, Dave Z, Kotrbova A, Kaiser K, Klemova D, et al.Simplified protocol for flow cytometry analysis of fluorescently labeled exosomesand microvesicles using dedicated flow cytometer. J Extracell Vesicles 2015;4:25530.
(13.) Arraud N, Gounou C, Turpin D, Brisson AR. Fluorescence triggering: a general strategy for enumerating and phenotyping extracellular vesicles by flow cytometry. Cytometry A2016;89:184-95.
(14.) Morales-Kastresana A, Telford B, Musich TA, McKinnon K, Clayborne C, Braig Z, et al. Labeling extracellular vesicles for nanoscale flow cytometry. Sci Rep 2017;7:1878.
(15.) Nolte-'t Hoen ENM, van der Vlist EJ, Aalberts M, Mertens HCH, Bosch BJ, Bartelink W, et al. Quantitative and qualitative flow cytometric analysis of nanosized cell-derived membrane vesicles. Nanomedicine 2012; 8:712-20.
(16.) Viability and cytotoxicity assay reagents. In: Johnson I, Spence MTZ, editors. The molecular probes handbook: a guide to fluorescent probes and labeling technologies, Vol. 11. Eugene (OR): Life Technologies Corporation; 2010. p. 656-73.
(17.) Parish CR. Fluorescent dyes for lymphocyte migration and proliferation studies. Immunol Cell Biol 1999;77: 499-508.
(18.) Fast-response probes. In: Johnson I, Spence MTZ, editors. The molecular probes handbook: a guide to fluorescent probes and labeling technologies, Vol. 11. Eugene (OR): Life Technologies Corporation; 2010.p.926-8.
(19.) Albanyan A-M, Murphy MF, Rasmussen JT, Heegaard CW, Harrison P. Measurement of phosphatidylserine exposure during storage of platelet concentrates using the novel probe lactadherin: a comparison study with annexin V. Transfusion 2009;49:99-107.
(20.) Arraud N, Linares R, Tan S, Gounou C, Pasquet JM, Mornet S, Brisson AR. Extracellular vesicles from blood plasma: determination of their morphology, size, phenotype and concentration. J Thromb Haemost 2014; 12:614-27.
(21.) Takov K, Yellon DM, Davidson SM. Confounding factors in vesicle uptake studies using fluorescent lipophilic membrane dyes. J Extracell Vesicles 2017;6:1388731.
(22.) Berckmans RJ, Sturk A, van Tienen LM, Schaap MCL, Nieuwland R. Cell-derived vesicles exposing coagulant tissuefactor insaliva. Blood 2011;117:3172-80.
(23.) Boing AN, van der Pol E, Grootemaat AE, Coumans FAW, SturkA, Nieuwland R. Single-step isolation of Extracellular vesicles by size-exclusion chromatography. J Extracell Vesicles2014;3.
(24.) Gool EL, Stojanovic I, Schasfoort RBM, Sturk A, van Leeuwen TG, Nieuwland R, et al. Surface plasmon resonance is an analytically sensitive method for antigen profiling of extracellular vesicles. Clin Chem 2016;63:1633-41.
(25.) Gasecka A, Boing AN, Filipiak KJ, Nieuwland R. Platelet extracellular vesicles as biomarkers for arterial thrombosis. Platelets 2017;28:228-34.
(26.) Coumans FAW, Brisson AR, Buzas EI, Dignat-George F, Drees EEE, El-Andaloussi S, et al. Methodological guidelines to study extracellular vesicles. Circ Res 2017;120:1632-48.
(27.) Hortin GL. Amino acids, peptides, and proteins. In: Burtis CA, Ashwood ER, Bruns DE, editors. Tietz textbook of Clinical chemistry and molecular diagnostics, Vol.5. St. Louis (MO): Elsevier Saunders; 2012. p. 509-65.
(28.) van der Pol E, van Gemert MJC, Sturk A, Nieuwland R, Van Leeuwen TG. Single vs. swarm detection of microparticles and exosomes by flow cytometry. J Thromb Haemost 2012;10:919-30.
(29.) Nolan JP, Stoner SA. A trigger channel threshold artifact in nanoparticle analysis. Cytometry A2013;83:301-5.
(30.) Gross E, Bedlack RS, Loew LM. Dual-wavelength ratiometric fluorescence measurement of the membrane dipole potential. Biophys J 1994;67:208 -16.
Leonie de Rond, [1,2,3] * Edwin van der Pol, [1,2,3] Chi M. Hau, [2,3] Zoltan Varga, [4,5] Auguste Sturk, [2,3] Ton G. van Leeuwen, [1,2] Rienk Nieuwland, [2,3] and Frank A.W. Coumans [1,2,3]
 Biomedical Engineering and Physics, Academic Medical Center, University of Amster dam, Amsterdam, the Netherlands;  Laboratory Experimental Clinical Chemistry, Aca demic Medical Center, University of Amsterdam, Amsterdam, the Netherlands;  Vesicle Observation Center, Academic Medical Center, University of Amsterdam, Amsterdam, the Netherlands; Biological Nanochemistry Research Group, Institute of Materials and Environmental Chemistry, Research Centre for Natural Sciences, Hungarian Academy of Sciences, Budapest, Hungary;  Department of Biophysics and Radiation Biology, Semmelweis University, Budapest, Hungary.
* Address correspondence to this author at: Academic Medical Center, University of Amsterdam, Department of Biomedical Engineering and Physics, Meibergdreef 9,1105AZ Amsterdam, the Netherlands. Fax +31-(0)-20-5664440; e-mail firstname.lastname@example.org.
Received July 3,2017; accepted January 12,2018.
Previously published online at DOI: 10.1373/clinchem.2017.278978
 Nonstandard abbreviations: EV, extracellular vesicle; AM, acetoxymethyl; CFSE, carboxyfluorescein succinimidyl ester; di-8-ANEPPS, 4-(2-[6-(dioctylamino)-2-naphthalenyl]ethenyl)-1 (3-sulfopropyl)pyridinium; PKH67, Paul Karl Horan 67; EpCAM, epithelial cell adhesion molecule; CD61, integrin, [beta]3; FITC, fluorescein isothiocyanate; APC, allophycocyanin; SEC, sizeexclusion chromatography; mAb, monoclonal antibody; PE, phycoerythrin; MESF, molecules of equivalent soluble fluorochrome; PPV, positive predictive value.
Caption: Fig.1. Generic marker performance on MCF7 EVs.
Generic marker fluorescence-triggered scatter plots of EpCAM vs generic marker fluorescence (A-E) or side scatter (F). Data of 1 representative minute are shown. Numbers in quadrants indicate percentage of total population. Venn diagrams show mean concentrations (x106 particles/ mL) of antibody+ ([mAb.sup.+]), [marker.sup.+] or scatter+, and double+ particles (n = 3-6) (G-L). Calculation of sensitivity and PPVfrom Venn diagrams (M).
Caption: Fig.2. Generic marker performance on plasma.
Generic marker fluorescence triggered scatter plot of CD61 vs generic marker fluorescence (A-E) or side scatter (F). Data of 1 representative minute are shown. Numbers in quadrants indicate percentage of total population. Venn diagrams show mean concentrations (x-6 particles/ mL) of antibody+ ([mAb.sup.+]), [marker.sup.+] or scatter+, and double+ particles in plasma (G-L) and plasma SEC (M-R) (n = 3-6). Calculation of sensitivity and PPV from Venn diagrams (S).
Caption: Fig.3. Venn diagrams comparing side scatter triggering with di-8-ANEPPS and lactadherin triggering for MCF7 EVs or plasma.
Shown are mean concentrations (x[10.sup.6] particles/mL) of antibody+ ([mAb.sup.+]), [marker.sup.+] or scatter+, double+ and triple+ particles in MCF7 EVs (A and B) and plasma (C).
Table 1. Published uses of generic markers for fluorescence triggering of extracellular vesicles. Marker Mechanism EV sample Calcein AM Fluoresces upon hydrolysis Cell culture, PFP,C erythrocyte EVs CFSE Fluoresces upon hydrolysis Cell culture, malignant and then binds to amines ascites Di-8-ANEPPS Voltage-sensitive membrane PRP marker Ann exin V Binds to PFP phosphatidylserine PKH67 Intercalates in membrane Cell culture, seminal fluid Marker Flow cytometer Fluorescence resolution limit, MESF Calcein AM BD LSRII -- CFSE Apogee A50 -- Di-8-ANEPPS Customized FITC236 FACSCanto PE 71 Ann exin V BC Gallios Cy51263 PKH67 Customized BD influx -- Marker Smallest bead Conclusion (b) measured (a) nm Calcein AM 760 Labels only intact EVs CFSE 100 Labels more EVs than annexin V, protocol <12 h Di-8-ANEPPS 530 Allows EV sizing Ann exin V 300 Fluorescence triggering detects 77x more platelet EVs than scatter triggering PKH67 200 Fluorescence triggering distinguishes EVs from noise Marker Reference number Calcein AM (id CFSE (12) Di-8-ANEPPS (9) Ann exin V (13) PKH67 (8) (a) Smallest polystyrene bead diameter distinguished in indicated reference, 180-nm refractive index 1.42 bead in reference 12 converted to polystyrene bead equivalent using scatter model (28). (b) Conclusion as drawn by the authors in reference. (c) PFP, platelet-free plasma; PRP, platelet-rich plasma; Cy5, cyanine 5. Table 2. Summary of trigger strategies for detecting extracellular vesicles. Trigger Channel Spectral overlap with Fluorescence Calcein AM FITC FITC Calcein violet Violet DAPI CFSE (d) FITC FITC Di-8-ANEPPS PerCP PerCP, PE, FITC Lactadherin FITCe FITCe Scatter SSC -- MCF7 EVs Trigger Sensitivity (a), % PPV (b), % Fluorescence Calcein AM 16 [+ or -] 8 3 Calcein violet 47 [+ or -] 8 1 CFSE (d) 49 [+ or -] 9 0.2 Di-8-ANEPPS 122 [+ or -] 25 29 Lactadherin 113 [+ or -] 39 19 Scatter 105 [+ or -] 30 13 Platelet EVs Trigger Sensitivity (a), % PPV (b), % Fluorescence Calcein AM -0.1 [+ or -] 0.3 -0.2 Calcein violet 9 [+ or -] 4 7 CFSE (d) 3 [+ or -] 4 0.3 Di-8-ANEPPS 0.3 [+ or -] 2 0.2 Lactadherin 33 [+ or -] 11 17 Scatter 61 [+ or -] 25 0.1 Platelet EVs after SEC Trigger Detects also Sensitivity PPV (b), % (a), % Fluorescence Calcein AM M, Pc 1 [+ or -] 1 2 Calcein violet M, P 3[+ or -]1 1 CFSE (d) M, P, LP 3[+ or -]1 2 Di-8-ANEPPS M, P 1[+ or -]1 1 Lactadherin M, P 3 [+ or -] 1 1 Scatter M, P, LP 12[+ or -]3 0.3 (a) Antibody+marker+ particles relative to antibody+ particles [+ or -] SD. (b) Antibody+marker+ particles relative to marker+ particles [+ or -] SD. (c) M, marker in buffer; P, proteins; LP, lipoproteins; PerCP, peridinin-chlorophyll-protein; SSC, side scatter channel. (d) Results obtained with CFSE suffered from swarm (see Fig. 5 of the online Data Supplement). (e) PE if lactadherin-PE is used. Table 3. Detected particles in non-EVsamples using different trigger strategies. Detected particles ([10.sup.6]/mL) Trigger Marker in buffer Proteins Lipoproteins Fluorescence Calcein AM 90 400 0 Calcein violet 400 300 0 CFSE 20 14 000 800 Di-8-ANEPPS 4 50 0 Lactadherin 100 200 0 Scatter 300 370000 Concentrations are means of 2 repeated measurements. All concentrations were corrected for background in buffer without generic marker. The reported concentra-tions in the different columns cannot be mutually compared.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Automation and Analytical Techniques|
|Author:||de Rond, Leonie; van der Pol, Edwin; Hau, Chi M.; Varga, Zoltan; Sturk, Auguste; van Leeuwen, Ton G.|
|Date:||Apr 1, 2018|
|Previous Article:||State of the Science in Dried Blood Spots.|
|Next Article:||Specific Substrate for the Assay of Lysosomal Acid Lipase.|