Plasma extracellular vesicle test sample to standardize flow cytometry measurements

Background Extracellular vesicles (EVs) in body fluids are explored as disease biomarkers, but EV concentrations measured by flow cytometers (FCMs) are incomparable. Objectives To improve data comparability, new reference materials with physical properties resembling EVs and reference procedures are being developed. The validation of new reference materials and procedures requires biological test samples. We developed a human plasma EV test sample (PEVTES) that i) resembles subcellular particles in plasma, ii) is ready-to-use, iii) is flow cytometry–compatible, and iv) is stable. Methods The PEVTES was prepared from human plasma of 3 fasting donors. EVs were immunofluorescently stained with antibodies against platelet-specific (CD61) and erythrocyte-specific (CD235a) antigens or lactadherin. To reduce the concentration of soluble proteins, lipoproteins, and unbound reagents, stained EVs were isolated from plasma by size-exclusion chromatography. After isolation, the PEVTES was filtered to remove remnant platelets. PEVTESs were diluted in cryopreservation agents, dimethyl sulfoxide, glycerol, or trehalose and stored at −80 °C for 12 months. After thawing, stained EV concentrations were measured with a calibrated FCM (Apogee A60-Micro). Results We demonstrate that the developed PEVTES resembles subcellular particles in human plasma when measured using FCM and that the concentrations of prestained platelet-derived, erythrocyte-derived, and lactadherin+ EVs in the PEVTES are stable during storage at −80 °C for 12 months when stored in trehalose. Conclusion The PEVTES i) resembles subcellular particles in plasma, ii) is ready-to-use, iii) is flow cytometry–compatible, and iv) is stable. Therefore, the developed PEVTES is an ideal candidate to validate newly developed reference materials and procedures.


| I N T R O D U C T I O N
The term "extracellular vesicles" (EVs) is an umbrella term for naturally released cell-derived particles with a phospholipid bilayer. EVs are present in body fluids, such as blood and urine [1,2]. Properties of EVs, such as the cellular origin, concentration, composition, and function, are disease-dependent. Therefore, EVs are being explored as biomarkers for diseases, including cancer and cardiovascular disease [1,3,4]. However, to explore EVs as biomarkers, reliable and reproducible measurements of EVs are needed. At present, a major hurdle in EV biomarker research is that measurement results are incomparable between instruments and institutes, which hampers setting up multicenter studies. Multicenter studies are a prerequisite for clinical biomarker studies.
At present, most clinical laboratories use flow cytometry to measure the concentration of immunofluorescently stained EVs in body fluids because flow cytometry is capable of detecting and characterizing particles at a throughput of thousands/s [5][6][7]. A flow cytometer (FCM) detects fluorescence and light scattering signals of single particles, such as EVs, when these signals exceed the lower detection limit of the detectors. There are at least 4 major reasons why EV concentration measurement results are incomparable. First, compared with cells, stained EVs emit little fluorescence and scatter light inefficiently, which makes EVs hard to detect [8,9]. Second, fluorescence and light scattering signals of a part of the total EV population are below the detection limit of commercially available FCMs. Third, signals measured by FCM have arbitrary units, which hampers the quantification and comparison of detection ranges between FCMs [10,11]. Fourth, most EV samples contain non-EV particles, such as lipoproteins in plasma, which outnumber EVs, overlap in size range with EVs, and could be falsely identified as EVs [10].
To standardize flow cytometry measurements, the measured fluorescence and light scattering signals require calibration to convert arbitrary units into standard units. For example, the arbitrary units of fluorescence intensity can be related to the standard unit molecules of equivalent soluble fluorochrome (MESF) [11,12]. The arbitrary units of light scattering intensity can be related to the diameter of EVs in nm. Calibrated fluorescence and light scattering signals allow to express EV concentrations within similar detection ranges.
Calibration requires stable, traceably characterized certified reference materials (RMs) with optical signal levels similar to EVs.
Please note that RMs can be divided into (i) quality control materials and (ii) certified RMs [13,14]. With certified RM, we mean a sample containing reference particles, of which the physical property that is intended for calibration purposes is homogeneous, metrologically traceably characterized, and stable. With traceable, we mean that the measurement results of the property intended for calibration (eg, particle size) can be related to the International System of Units (SI) through an unbroken chain of comparisons with known uncertainties [15][16][17] In a previous interlaboratory comparison study, aliquots of unstained frozen plasma samples were distributed as an EV-containing test sample. Participants had to fluorescently stain the plasma EVs themselves, which led to preanalytical and interuser variations in the measurement results [18].
Our aim was to develop a plasma EV test sample (PEVTES) that i) resembles subcellular particles in plasma, ii) is ready-to-use, iii) is flow cytometry-compatible, and iv) is stable. By stable, we mean that the concentration of EVs within a certain fluorescence or light scattering intensity range does not change over a time period corresponding to the intended purpose of the sample, such as an interlaboratory comparison study or quality control. In this article, we provide an overview of how to achieve each of the beforementioned aims (i-iv) and how to develop a PEVTES. The developed PEVTES is being used in an upcoming METVES II-organized interlaboratory comparison study to validate developed certified RMs and procedures.

Essentials
• Extracellular vesicle (EV) concentration measurements by flow cytometry require calibration.
• Calibration requires reference materials and test samples.
• We developed a stable, ready-to-use EV test sample from human plasma.
• The plasma EV test sample is an ideal candidate to validate reference materials and procedures.

| Study design
The aim of the study was to develop a PEVTES. The PEVTES should i) resemble subcellular particles in plasma because plasma is the most widely studied biofluid for EV biomarker research [19]. Therefore, plasma was chosen as a starting material. To be ii) ready-to-use, EVs present in plasma were immunofluorescently stained with antibodies against cell type-specific proteins. This makes the PEVTES a prestained sample. To ensure that the PEVTES iii) is flow cytometrycompatible, swarm detection of particles below the detection limit was minimized and unbound reagents were removed by size-exclusion chromatography (SEC). Furthermore, discoid-shaped residual platelets and tubular particles [8] were removed by filtration. Removing non-spherical particles is important because Mie theory, which relates scatter signals to particle diameter and refractive index, only applies to spherical particles [20]. To iv) stably store the samples, different cryopreservation agents were evaluated. To evaluate stability, the concentration of stained EVs was measured before freezing and after 1, 3, 6, and 12 months of storage at −80 • C with an FCM, of which the flow rate, fluorescence signals, and light scattering signals were calibrated. A schematic overview of the procedure can be found in
The first 3.5 mL eluate containing the void volume was discarded, after which the 1-mL fraction containing most EVs was collected and pooled.

| Platelet removal with polycarbonate filters
To remove remaining platelets from plasma, plasma was filtered using a 0.8-μm pore-size polycarbonate membrane filter (ATTP02500, Isopore, Merck Millipore) with a diameter of 25 mm. Typically, this step reduces the residual platelet concentration 1.5 × 10 2 -fold [22].

| Stabilization
To improve the stability of the prestained and SEC-isolated EVs, the cryopreservation agents, dimethyl sulfoxide (DMSO), glycerol, and trehalose, were selected based on the literature and tested [23][24][25][26][27][28]. The optimal concentration of each cryopreservation agent for long-term stability was investigated in preliminary experiments (data not shown

| Storage and thawing
Samples were frozen in liquid nitrogen and stored at −80 • C. Frozen samples were thawed in a water bath at 37 • C for 1 minute before use and kept at room temperature prior to measurements.

| Procedural controls
Besides buffer-only and isotype controls, procedural controls were included to confirm the absence of particles introduced by the sample preparation procedure [29]. As a procedural control, the entire PEVTES procedure was applied to dPBS instead of cell-depleted plasma as a starting material.   Furthermore, the PEVTES does not contain detectable concentrations of platelets, which confirms that the procedure to remove platelets by filtration after SEC is effective. As the fluorescence signals vs the diameter of the PEVTES resembles EVs present in human plasma, we show that it is feasible to prepare and store stained plasma EVs.

| Stability of the plasma EV test sample during storage
During storage at −80 • C, the concentration and staining of EVs may be affected by the freezing and thawing process itself and/or by the storage time. Both processes are known to cause EV loss [31,32]. With the aim to investigate these processes independently, we measured EV concentrations freshly and at 4 time-points after storage.   We also compared the EV concentration in the PEVTES after 12 months of storage to 1 month of storage. Please note that both 12 months of storage and 1 month of storage include a single freeze-thaw cycle, and therefore, the most prominent variable is the storage time itself rather than the influence of freeze-thawing on the sample.
The measured CD61 + EV concentration decreased 31% when stored in DMSO, 1% when stored in glycerol, and 1% when stored in trehalose, as shown in Figure 3A-C. The measured CD235a + EV concentration stored for 12 months decreased 58% when samples were stored in DMSO, 7% when stored in glycerol, and 2% when stored in trehalose, compared with 1 month of storage at −80 • C, as shown in Figure 3D-F. Furthermore, the measured lactadherin + EV concentration stored for 12 months decreased 5% when samples were stored in DMSO and 22% when stored in glycerol and increased 2% when stored in trehalose, compared with 1 month of storage at −80 • C, as shown in Figure 3G-I, respectively.
All in all, we show that the developed PEVTES can be stored stably for at least 12 months at −80 • C in the presence of a cryopreservation agent. Furthermore, we show that trehalose is the most suitable cryopreservation agent for our PEVTES. Figure 4 shows the bench stability of the PEVTES. With bench stability, we mean the stability of the test sample after thawing when stored at room temperature. Figure 4A shows that the concentration of CD61 + EVs, over time in plasma, increased by 25% (coefficient of variation [CV], 5%), which we attribute to evaporation of the medium.

| Plasma EV test sample bench stability
Please note that the samples were not covered during measurement.   This result emphasizes the relevance of cooling or covering the sample before measurement. Figure 4B shows that the concentration of CD61 + EVs in the PEVTES decreased 44% (CV, 15%) during the measurement time of 340 minutes at room temperature. We attribute the decrease in particle concentration over time to the adhesion of EVs to the plastic of the well plates in the absence of proteins [33].
Therefore, we developed a procedure to stabilize the developed PEVTES by adding BSA. Similarly, to plasma, Figure 4C shows a decrease in the CD61 + EV concentration in the PEVTES with BSA over 340 minutes (7% decrease; CV, 7%). This shows that the addition of BSA prevents adhesion of EVs to the surface of the well plates.
Taken together, the developed procedure is capable of producing prestained human plasma EVs that are stable upon storage at −80 • C for at least 12 months. Please note that prestained means that the PEVTES is a ready-to-use sample, which only requires dilution prior to measurement and no additional staining step. EV-containing test sample [18]. Therefore, we developed the PEVTES that i) resembles subcellular particles in plasma, ii) is readyto-use, iii) is flow cytometry-compatible, and iv) is stable during storage.
The Table shows an overview of currently available test samples for EV flow cytometry, which include plasma, recombinant EVs (rEVs) [34], and engineered retroviruses [35]. Our results show that within the detection range of an FCM, EVs and non-EV particles in the PEVTES resemble particles found in human plasma (Figure 2A-D). To develop a flow cytometry-compatible sample, we decreased the concentration of particles <70 nm to reduce swarm detection by SEC, which will also deplete the bulk of soluble proteins, including unbound antibodies. After SEC, we applied a 0.8-μm pore-size polycarbonate membrane filter to remove particles with a diameter >800 nm, which include discoid-shaped platelets and tubular particles [8]. By removing the discoid and nonspherical particles, we ensure the validity of Mie theory to calibrate light scattering signals. While frozen human plasma still contains cells [22], PEVTES presents a cell-depleted sample. By removal of platelets before freezing, cell fragmentation can be excluded.
Freezing may cause membrane damage and loss of EV function [36]. Generally, rEVs, engineered retrovirus samples, and PEVTESs are  all stable during and after storage, whereas this might not be the case for frozen human plasma [37]. To prepare a stable sample, different cryopreservation agents (DMSO [23,27], glycerol [24,25], and trehalose [25,26,28]) were added to the PEVTES sample before freezing.
The use of cryopreservation agents is a commonly accepted procedure to prevent osmotic damage and preserve protein stability [38].
Generally, we showed that the 3 different cryopreservation agents had different degrees of stabilizing effectiveness, expressed as the percentage decrease in the plasma-EV concentration over time ( Figure 4). DMSO and glycerol, both penetrating cryopreservation agents, had lower stabilizing effectiveness than trehalose [24]. Nonpenetrating cryopreservation agents, such as trehalose and other sugars, may represent more biocompatible cryopreservation agents [39]. In our experiments, trehalose best preserved prestained EVs present in the PEVTES as measured by a minimal decrease in EV concentration over 12 months of storage at −80 • C.
Besides storage stability, we found that the bench stability of the PEVTES decreased with measurement time (Figure 4). First, we observed that the total particle concentration and CD61 + EV in human plasma increased with measurement time. The increase in particles can be attributed to evaporation of the surrounding medium studies also warned about possible adherence of EVs to surfaces [33,36,42]. Multiple studies suggested that the addition of a soluble protein, such as BSA, might reduce the loss of EVs to "normal" tubes (eg, Eppendorf) or the use of low-binding plastics. Evtushenko et al. [33] showed that particle loss could be reduced by 18% when the surface of plastic tubes was blocked with BSA. In line with our first observation that the total particle concentration in human plasma increased with measurement time, we observed a similar trend for the PEVTES sample containing BSA. Evaporation and adhesion of EVs to plastic are processes that may occur concurrently.
In sum, we can conclude that the prestained EVs present in the PEVTES are stable during storage for at least 12 months but require the addition of BSA to improve bench stability. All in all, the new and improved PEVTES is stable after storage for at least 12 months at −80 • C and for 6 hours on bench.
The developed PEVTES has multiple potential applications, including a quality control sample for monitoring day-to-day variability of instruments measuring EV properties such as light scattering intensity, fluorescence intensity, size, number concentration, and refractive index. Furthermore, the PEVTES may be useful to optimize settings of instruments measuring EV properties or to develop new techniques measuring EV properties.
Taken together, the developed PEVTES i) resembles subcellular particles in plasma, ii) is ready-to-use, iii) is flow cytometrycompatible, and iv) is stable. The developed PEVTES will be used in an interlaboratory comparison study to validate newly developed certified RMs and procedures. In the future, traceable characterization of the size distribution, number concentration, fluorescence intensity, and refractive index of EVs in the PEVTES is planned. This will be the first time in the field of EV research that all these characteristics are traceably measured for a biological test sample. Together with EVdedicated certified RMs, the developed PEVTES sample is an essential part of the infrastructure needed to facilitate multicenter studies in the field of EV research in the future.

ACKNOWLEDGMENTS
The authors thank C.M. Hau and N. Hajji for their excellent technical support.