The effect of a freeze-thaw cycle and residual platelets on the
particle size distribution of extracellular vesicles
Student: Isis de Lange (11693762) Daily supervisor: Naomi Buntsma Senior supervisor: Rienk Nieuwland 2nd corrector: Mark Hink
Date: July 6th 2020
Abstract
Extracellular vesicles (EVs) are small subcellular vesicles ranging from 30-1000 nm in size and are present in almost all body fluids. EVs have proven to play a potential role in biomedical applications. However, there are many challenges regarding EV research, two of these challenges are sample storage and the presence of residual platelets. The aim of this study was to provide information about the effect of a freeze-thaw cycle and the presence of residual platelets on the particle size distribution of EV samples. And as a result, know more about the EV sample quality. Flow cytometry was used to determine the concentration of EVs and residual platelets in either 200, 50 or 10 nm gates. Curve fitting was done to analyze the change in PSD either after freezing and thawing or in EV samples with different residual platelet concentrations. The results showed that both freezing and thawing as well as the presence of residual platelets can affect the PSD of EVs. These results show the importance of removing residual platelets from EV samples as well as how the PSD of EVs can be affected by a freeze-thaw cycle. Overall these results give insight into the quality of EV samples.
Introduction
Extracellular vesicles (EVs) are small subcellular particles surrounded by a lipid bilayer, which get released from both prokaryotic and eukaryotic cells. EVs are present in body fluids like blood plasma, urine and milk where they function as cell-cell messengers or as messengers to the environment (Lacroix et al., 2012). In addition, EVs have been shown to play an important role in processes like apoptosis, coagulation, cancer and inflammation (Lacroix et al., 2012). Furthermore, EVs have potential applications in diagnostics as biomarkers or function as therapeutic drug carriers (Coumans et al., 2017). Unfortunately, EV research has some limitations, mainly due to the stability and the size of the particles, which ranges between 30-1000 nm (Park et al., 2018). One of these limitations is sample storage. Since EVs can be fragile the correct storage conditions are important to keep the EV samples consistent between measurements and to make reliable comparisons between different laboratory results (Arraud et al., 2014).
Therefore, knowledge of the composition of the EV samples before and after storage is important to be able to say something about the sample quality.
In order to store the EVs, the samples get frozen via snap freezing in liquid nitrogen and stored at -80C (Coumans et al., 2017). Previous research has shown that -80C is the optimal temperature for storing EVs. With higher temperatures both a shift in EV concentration and a change in EV particle size distribution (PSD) have been observed (Lőrincz et al., 2014). The study of Lőrincz et al. (2014) showed that after a single freeze-thaw cycle the PSD of EV samples increased. Fresh EVs follow a power-law function (Van der Pol et al., 2014). Ice crystal formation, as well as other forms of mechanical stress, can lead to EV destruction or EV fragmentation (Qin et al., 2020; Arraud et al., 2014). This was more apparent in the larger EVs (Arraud et al., 2014).
But not only storage temperature affects the EV sample composition, other studies have shown that the use of different anticoagulants, as well as the presence of residual platelets, can also affect the EV concentration (Mitchell et al., 2016; Artoni et al., 2012). Residual platelets have been found to be the main cause for unstable EV counts in platelet-derived EV samplessince residual platelets continue to release EVs during storage. This effect is increased after a freeze-thaw cycle (Lawrie et al., 2008; Kriebardis et al., 2016). Furthermore, was observed that the EV concentration in the samples increases when the residual platelet concentration increases (Subliminal 1, Figure A; Gasecka et al. 2020; Jamalay et al., 2018).
Since there are no detailed reports regarding the effects of a freeze-thaw cycle and the presence of residual platelets on the PSD of EVs, this study tries to answer the following question: How does a freeze-thaw cycle, different anticoagulants and the presence of residual platelets affect the PSD of EVs?
The hypothesis is that if there is a change in PSD after a freeze-thaw cycle, there are three possible causes for this change: aggregation, breakdown and shrinkage of EVs. Aggregation of EVs would show an increase in the larger diameter EVs and a decrease in the smaller EVs and thus a PSD shift towards the larger EVs. Breakdown of EVs would show the opposite effect: less large EVs and more small EVs, hence an overall PSD shift toward the smaller EVs. Shrinking would show a decrease in all EVs, not affecting the PSD. Combinations of these effects are possible as well. The hypothesis for the presence of residual platelets in EV samples is that high concentration residual platelets will result in higher concentrations of all EVs. The effect of residual platelet on the PSD is still unknown.
This study will focus on EVs which were derived from platelets (CD61+), activated platelets (CD62p+), leukocytes (CD45+) and erythrocytes (CD235a+). And this study looked at the effect of a freeze-thaw cycle on the PSD of EVs in two different anticoagulants as well as the effect of different residual platelet concentrations of the PSD of EVs.
Sample preparation. Single freeze-thaw cycle and EDTA or citrate. Samples were prepared and measured by Naomi Buntsma.
Citrate and EDTA anticoagulated blood was collected from 13 healthy donors, according to the guidelines for EV research (Lacroix et al., 2012). Platelet-poor plasma (PPP) was prepared by double centrifugation according to the ISTH guidelines, snap-frozen in liquid nitrogen and stored at -80C (Lacroix et al 2013). To examine the stability of EV concentrations over a freeze-thaw cycle, plasma samples were measured fresh and after freeze-storage by freeze-thawing at 37C. The erythrocyte (CD235a+), leukocyte (CD45+), platelet (CD61+) and activated platelet (CD62p+) derived EVs were labelled. EV concentrations were determined using a flow cytometer for nanoparticle tracking (A60-Micro, Apogee Flow Systems, Hertfordshire, UK).
Sample preparation. Residual platelets. Samples, protocol and data from the study of Gasecka et al. (2018)
Blood from 30 healthy volunteers was collected according to the guidelines for EV research, set up by Yuana et al (2015). Preparation of PPP as well as the determination of the EV and residual platelet concentration was the same as the freeze-thaw analysis samples. Platelet (CD61+) and activated platelet (CD62p+) EV samples with low (concentration < 7.2 * 105 mL),
medium (concentration < 1.2 * 107 mL), and high (concentration > 5.2 * 107 mL) concentrations
of residual platelets were compared to see how the different concentrations affect the EV PSD.
Sample analysis. Rosetta Calibration was used to relate scatter measured by forward or side scatter to the scattering cross-section and diameter of EVs. The scattering cross-sections and EV diameters were added to the flow cytometry data-files by custom-build software (MATLAB R2018b, MathWorks, US).
To analyze the change in particle size distribution after freezing and thawing, the EV population (microparticles <1000 nm) was divided into 200 nm groups (200-399 nm, 400-599 nm, 600-799 nm and 800-999 nm) and 50nm groups (200-249 nm, 250-299 nm, 300-349 nm, 350-399 nm, 400-449 nm, 450-499 nm, 500-549 nm and 550-599 nm).
For PSD analysis was decided to leave out all EVs <200, to make sure the detection limit of the flow cytometer (~160 nm) was exceeded (Van der Pol et al., 2014). For the 50 nm analysis, EVs >600 nm were left out because there were not enough counts to make a reliable estimate of the EV concentration. A systemic measuring error of 50% was tolerated (= 4 counts).
In order to examine the underlying cause of the shift of the EV population further divided in 10 nm groups ranging from 200-800 nm. In order to fit a power-law function on the 10 nm gate data, all data was log-transformed. The following function was fit on the log-transformed data:
Y = log(A*X^B)
A theoretical curve was added on the log-transformed data, resembling the linear power-law function. The theoretical curve followed the following function:
The shift of the linear power-law function fit will give information about whether the EVs aggregate, shrink of breakdown.
The PSD analysis for the freeze-thaw cycle were fit with a semi-log function since this was a better fit for the data (Subliminal 2, figure B):
Y = 10^(Slope*X + Yintercept)
The concentrations of the EVs before and after freezing and graphs made were compared using GraphPad Prism 5.
Statistical analysis. Statistical tests were executed in R studio. It was established that the data for the anticoagulant comparison was not normally distributed (Shapiro-Wilk test, p0.05, data does not follow a normal distribution). Thus, median values and interquartile ranges were used to plot the EV concentrations for the 200 nm and 50 nm groups. To compare the EV concentrations before and after a freeze-thaw cycle a Wilcoxon signed-rank test (paired) was executed. For the residual platelet analysis, the mean values and SD were used to plot the EV concentrations (Shapiro-Wilk, p > 0.05). And a paired t-test was used to analyze the statistical differences. Bar graphs were used to display changes in the PSD after a freeze-thaw cycle in the 200 nm and 50 nm groups for EDTA and citrate. Scatter plots were used to show the PSD for the 10 nm group for EDTA and citrate. Differences in function fits, between samples, were determined by a paired t-test and considered significant when p<0.05.
Results
Freeze-thaw cycle analysis: 200 nm gates
First, the effects of a single freeze-thaw cycle on the EV PSD were evaluated for erythrocyte, leukocyte and platelet-derived EVs in both 200 nm and 50 nm gates. The results of the 200 nm analysis for both Citrate and EDTA are shown in Figure 1.
The 200-400 nm EVs show a decrease in EV concentration after a freeze-thaw cycle for all EVs, except erythrocyte EVs in EDTA. Furthermore, in citrate was observed that for erythrocyte, leukocyte and platelet EVs there was a significant decrease in the EV concentration in both of the outer categories (200-400 nm, 600-800 nm and 800-1000 nm). Whilst the 400-600 nm EVs remained unchanged after freezing and thawing. For EDTA, only one side of the EV population was affectedby freezing and thawing as opposed to both sides for Citrate. For erythrocyte and platelets EV only the 600-1000 nm EVs were effected and showed a significant decrease in EV concentration. For leukocyte EVs only the 200-600 nm EVs were effected and showed a significant decrease in EV concentration.
Figure 1. Concentration of EVs before and after a freeze thaw cycle in citrate versus EDTA, 200 nm gates. Values are
medians with the interquartile range as error bar (n = 13 in each group). The effect of freeze-thaw cycle on the PSD was evaluated for both citrate and EDTA anticoagulated blood, by analyzing EV samples in PPP after freezing at -80C and thawing at 37C. (A) The effect of a freeze-thaw cycle on the PSD of citrate anticoagulated blood. Both the small (200-400 nm) and the large (600-1000 nm) EVs were effected by freezing and thawing and showed a significant decrease in concentration. The middle group (400-600 nm) did not show a significant change in concentration. (B) The effect of a freeze-thaw cycle on EDTA anticoagulated blood. Leukocyte and non activated platelet EVs showed a significant decrease in concentration for the small to medium EVs (200-600 nm). But not for the large EVs (600-1000 nm). Erythrocyte EVs showed a significant decrease in the larger EVs (600-1000 nm) (p <0.05) but not in the 200-600 nm EVs. EV counts for the 800-1000 nm EVs was < 20.
Freeze-thaw cycle analysis: 50 nm gates
To see whether there were any changes within the 200 nm groups after a single freeze-thaw cycle,
the EV populations were analyzed in 50nm groups (Figure 2). 200-250 nm and 250-300 nm EVs showed a decrease in concentration for all EVs except for erythrocyte EVs.
Erythrocyte EVs in citrate anticoagulated plasma showed a decrease in concentration after freezing and thawing for 200-300 nm EVs. All EVs sized 300-600 nm did not show a change in concentration before and after freezing and thawing. Erythrocyte EVs in EDTA anticoagulated
20 0-400n m 40 0-600n m 60 0-800n m 80 0-1000 nm 0.0 5.0×106 1.0×107 1.5×107 C o n ce n tr at io n ( m L -1 ) Leukocyte EV (Citrate) Before After * * 20 0-400n m 40 0-600n m 60 0-800n m 80 0-1000 nm 0.0 5.0×106 1.0×107 1.5×107 C o n ce n tr at io n ( m L -1 ) Leukocyte EV (EDTA) Before After * * 20 0-400n m 40 0-600n m 60 0-800n m 80 0-1000 nm 0 1×107 2×107 3×107 C o n ce n tr at io n ( m L -1 ) Erythrocyte EV (Citrate) Before After * * * 20 0-400n m 40 0-600n m 60 0-800n m 80 0-1000 nm 0 1×107 2×107 3×107 C o n ce n tr at io n ( m L -1 ) Erythrocyte EV (EDTA) Before After * * 20 0-400n m 40 0-600n m 60 0-800n m 80 0-1000 nm 0 2×107 4×107 6×107 8×107 C o n ce n tr at io n ( m L -1 ) Platelet EV (Citrate) Before After * * 20 0-400n m 40 0-600n m 60 0-800n m 80 0-1000 nm 0.0 5.0×106 1.0×107 1.5×107 2.0×107 2.5×107 C o n ce n tr at io n ( m L -1 ) Platelet EV (EDTA) Before After * *
A
B
blood did not show a change in concentration after freezing and thawing. For leukocyte EVs in citrate, there was a decrease in the 200-250 nm, 250-300 nm, 350-400 nm and 450-500 nm EVs, all other EVs remained unchanged after freezing and thawing. Leukocyte EVs in EDTA the 200-250 nm, 250-300 nm and the 400-450 nm groups showed a decrease in concentration after freezing and thawing. All other groups remained unchanged. Lastly, platelet EVs, in citrate, showed a decrease in concentration after freezing and thawing for 250-450 nm EVs. EDTA anticoagulated platelet EVs samples showed a decrease in 200-250 nm, 300-350 nm, 350-400 nm and 450-500 nm EVs, the other EVs remained unchanged after freezing and thawing.
Figure 2. Concentration of EVs before and after freeze thaw cycle in citrate versus EDTA, 50nm gates. Values are medians
with the interquartile range as error bar (n = 13 in each group). The effect of freeze-thaw cycle on the PSD was evaluated for both citrate and EDTA anticoagulated blood, by analyzing EV samples in PPP after freezing at -80C and thawing at 37C.
(A,B) For erythrocyte EVs in citrate, only the 200-300 nm showed a decrease in EV concentration the 300-600 nm EVs
remained unchanged after freezing and thawing. Erythrocyte EVs in EDTA did not show a decrease of EV concentration in any of the 50 nm groups. (C) Leukocyte EVs in citrate showed a decrease in 200-250 nm, 250-300 nm, 350-450 nm and 450-500 nm groups. (D) EDTA anticoagulated leukocyte EV samples showed a decrease in 200-250 nm, 250-300 nm and 400-450 nm EVs. The 300-400 nm and 450-600 nm EV groups remained unchanged. (E) for platelet EVs in citrate the 250-450 nm groups showed a decrease in concentration after freezing and thawing. The 200-250 nm and 450-600 nm groups remained unchanged. (F) Platelet EVs in EDTA showed a decrease in the 200-250 nm, 300-350 nm, 350-400 nm and 450-500 nm groups after freezing and thawing. The 250-300 nm, 400-450 nm and 500-600 nm groups remained unchanged.
Freeze-thaw cycle analysis: semi-log fit and PSD
Next, to determine the underlying cause of the change in PSD, a semi-log function was fit on the EV data. The results of the analysis are shown in Figure 3. A change in “Yintercept” of the
semi-20 0-250n m 25 0-300n m 30 0-350n m 35 0-400n m 40 0-450n m 45 0-500n m 50 0-550n m 55 0-600n m 0.0 5.0×106 1.0×107 1.5×107 C o n ce n tr at io n ( m L -1 ) Erythrocyte EV (Citrate) Before After * * 20 0-250n m 25 0-300n m 30 0-350n m 35 0-400n m 40 0-450n m 45 0-500n m 50 0-550n m 55 0-600n m 0.0 5.0×106 1.0×107 1.5×107 C o n ce n tr at io n ( m L -1 ) Erythrocyte EV (EDTA) Before After 20 0-250n m 25 0-300n m 30 0-350n m 35 0-400n m 40 0-450n m 45 0-500n m 50 0-550n m 55 0-600n m 0 1×106 2×106 3×106 4×106 5×106 C o n ce n tr at io n ( m L -1 ) Leukocyte EV (Citrate) Before After * * * * 20 0-250n m 25 0-300n m 30 0-350n m 35 0-400n m 40 0-450n m 45 0-500n m 50 0-550n m 55 0-600n m 0 1×106 2×106 3×106 4×106 5×106 C o n ce n tr at io n ( m L -1 ) Leukocyte EV (EDTA) Before After * * * 20 0-250n m 25 0-300n m 30 0-350n m 35 0-400n m 40 0-450n m 45 0-500n m 50 0-550n m 55 0-600n m 0 1×107 2×107 3×107 C o n ce n tr at io n ( m L -1 ) Platelet EV (Citrate) Before After * * * * 20 0-250n m 25 0-300n m 30 0-350n m 35 0-400n m 40 0-450n m 45 0-500n m 50 0-550n m 55 0-600n m 0 1×107 2×107 3×107 C o n ce n tr at io n ( m L -1 ) Platelet EV (EDTA) Before After * * * * A C E B D F
log function indicates that there is a change in EV concentration for all EVs. An increase in the slope indicates that there is a shift in PSD towards the small EVs. A decrease in the slope indicates that there is an increase in the large EVs and as a consequence a shift in de PSD towards the large EVs. The results of the analysis are shown in Figure 5. Slope and intercept of the semi-log fits are shown in subliminal 4 (table A).
Erythrocyte EVs in citrate did not show a difference in either Yintercept (p=0.174) or slope (p=0.263) of the semi-log fit. For EDTA anticoagulated samples, an increase in Yintercept (p<0.0001) and an increase in the slope (p<0.0001) after freezing and thawing was observed. Leukocyte EVs in citrate also did not show a difference in either Yintercept (p=0.264) or slope (p=0.083) of the semi-log fit after freezing and thawing. EDTA anticoagulated samples did not show a decrease in the slope after freezing and thawing (p=0.159). However, the Yintercept of the semi-log showed a decrease (p=0.030). Lastly, platelet EVs in citrate anticoagulated plasma did not show a change in the slope of the semi-log (p=0.540), but did show a decrease in Yintercept (p=0.013). For EDTA anticoagulated samples neither a change in slope (p=0.360) nor Yintercept (p=0.116) was observed after freezing and thawing.
Figure 3: semi-log fits on 10nm gate EV data before and after freezing. The values are median concentrations (n=13). And
the data was fit with a semi-log function (black and red line). (A) For citrate anticoagulated erythrocyte samples no difference
Erythrocyte EV (Citrate) 0 200 400 600 800 4.0 4.5 5.0 5.5 6.0 6.5 Diameter (nm) lo g (c o n ce n tr at io n m L -1 ) Before After Erythrocyte EV (EDTA) 0 200 400 600 800 4.0 4.5 5.0 5.5 6.0 6.5 7.0 Diameter (nm) lo g (c o n ce n tr at io n m L -1 ) Before After Leukocyte EV (Citrate) 0 200 400 600 4.0 4.5 5.0 5.5 6.0 Before After Diameter (nm) lo g (c o n ce n tr at io n m l-1) Leukocyte EV (EDTA) 0 200 400 600 3 4 5 6 7 Before After Diameter (nm) lo g (o n ce n tr at io n m L -1 ) Platelet EV (Citrate) 0 200 400 600 800 3 4 5 6 7 Diameter (nm) lo g (c o n ce n tr at io n m L -1 ) Before After Platelet EV (EDTA) 0 200 400 600 800 3 4 5 6 7 Diameter (nm) C o n ce n tr at io n ( m L -1 ) Before After A B D C E F
in either the slope (p=0.220) of Yintercept (p=0.174) was observed after freezing ant thawing. (B) Erythrocyte EV samples in EDTA did show an increase in both the slope (p<0.0001) and the Yintercept (p<0.001) of the semi-log function after freezing and thawing. (C) The semi-log fit remained the same slope (p=0.08) and Yintercept (p=0.263) after freezing and thawing for leukocyte EVs in citrate. (D) For EDTA anticoagulated leukocyte EVs no change in slope was observed (p=0.159), the Yintercept before and after freezing showed a decrease (p=0.030). (E) Platelet EVs in citrate anticoagulated plasma did not show a change in the slope of the fit (p=0.540), but did show a decrease in Yintercept (p=0.013). (F) Lastly, platelet EVs in EDTA anticoagulated plasma did not show a difference in either the slope (p=0.360) or Yintercept (p=0.116) of the semi-log function after freezing in thawing.
Residual platelets analysis: 200 nm and 50 nm gates
Observing the effect on the PSD of different concentrations residual platelets (High: concentration > 5.2 * 107 , medium: concentration < 1.2 * 107 , low: concentration < 7.2 * 105) in
platelet and activated platelet derived EV samples.
All platelet EVs showed a significant difference compared to each other for high, medium and low concentration of residual platelets. This effect was observed for each 200nm group, as well as for each 50 nm group (Figure 4A, B)
.
For activated platelets EVs was observed that the difference between high, medium and low residual platelet samples is bigger for the large EVs than for the small EVs. For the 200 nm analysis, the 200-400 nm and 800-1000 nm EVs showed a concentration difference between high-low and medium-high-low residual platelet concentrations (Figure 4C, D). The 400-800 nm EVs only showed a difference between the high and low concentration residual platelets. Further analyzing the activated platelet EVs in 50 nm gates showed that the 500-550 nm and 550-600 nm EV concentration differ significantly for high, medium and low concentration residual platelets. The 350-400 nm, 400-450 nm and 500-550 nm EVs only high-low and medium-low concentrations of residual platelets showed a significant difference in EV concentrations. Lastly, the 250-300 nm and 300-350 nm EVs only showed a significant difference in EV concentration between high and low concentration residual platelets.
Residual platelets: power-law function and PSD
Next, to determine the underlying cause of the change in PSD, a power-law function was fit on the EV data, as it has been shown in previous studies to best describe the PSD of EVs (Van der Pol et al., 2014). A change in “A” parameter of the power-law function indicates that there is a change in EV concentration for all EVs. An increase in the “B” parameter indicates that there is a shift in PSD towards the small EVs. A decrease in the “B” parameter indicates that there is an increase in the large EVs and as a consequence a shift in de PSD towards the large EVs. The results of the analysis are shown in Figure 5.
Platelet EV samples with low concentration residual platelets followed the power-law function. The platelet EVs for medium and high concentration residual platelets did not longer follow the power-law function (p<0.0001; sum of squares F-test). Furthermore, samples with medium and high concentration residual platelets showed an increase in 300-450 nm EVs.
The activated platelet EVs followed for high, medium and low concentrations the power-law function (figure 5B). High, medium and low residual platelet EV samples all showed a significant difference between each other in the “A” parameter of the power-law function (table B). When comparing high and low residual platelet samples, a significant decrease was observed for the high residual platelet samples (p<0.0001). Medium residual platelet samples also showed a decrease in slope when compared to the low residual platelet samples (p=0.013). High and medium concentration residual platelet samples however did not show a difference in slope of the power-law fit (p=0.135).
Figure 4. Concentration of EVs from platelets and activated platelets with high, medium or low concentrations of residual platelets. Values are means with the SD as error bar (n = 10 in each group). The effect of residual platelets on the PSD of non
activated platelet and activated platelet EV. EV concentrations were measured with high, medium and low concentrations of residual platelets (A) Platelets EVs showed a difference in concentration for the entire EV population in the 200 nm group analysis for high-medium, high-low and medium-low residual platelets concentrations (200-1000 nm). (B) The platelet EVs also showed a significant difference between all groups for the 50 nm group analysis for high-medium, high-low and medium-low residual platelets concentrations (200-600 nm). (C)(D) For activated platelet EVs the difference in EV concentrations between high, medium and low concentration residual platelets increases when the diameter of the vesicle increases. 250-350 nm EVs, EV concentrations with high and low concentrations differ from each other. For 250-350-500 nm EVs samples the
A B
concentration differs between high-low and medium-low concentration residual concentrations. For 500-600 nm EVs samples, the concentration differs between high-low, medium-low and high-medium concentration residual concentrations.
Figure 4. Platelet and activated platelet EV concentrations with with high, medium or low concentrations of residual platelets, 10 nm gates. The values are mean concentrations (n=10) and the data was fit with a log power-law function (black,
read and blue line) and fits were compared with a linear power-law function (not shown). (A) Platelet EVs with low concentrations residual platelets followed power-law function. Medium and high concentrations residual platelets did not follow a power-law function (p<0.0001: sum of squares F-test). (B) The power-law-fit of the activated platelet EVs shows a vertical shift downwards when comparing high-low residual platelet samples (p<0.0001), high-medium (p=0.0001) and medium-low (p<0.0001) residual platelet concentrations. When comparing the slope, it was shown that high-low residual platelet samples had a different slope (p<0.0001). Comparing the slope of medium and low residual platelet samples also showed a significant difference (p=0.0130). But comparing high and medium residual platelet samples did not show a significant difference in slope (p=0.1350).
Discussion
In this study, the effect of a freeze-thaw cycle and residual platelets on the PSD of EVs was evaluated. The results showed that 200-400 nm EVs are most often affected by freezing and thawing as well as the presence of residual platelets. This was observed for erythrocyte, leukocyte and platelet EVs. Furthermore, freezing and thawing can cause a change in PSD in erythrocyte, leukocyte and platelet EVs. The PSD of citrate anticoagulated samples were more often affected after freezing and thawing than EDTA anticoagulated samples. For the residual platelet analysis was observed that high concentrations of residual platelets in platelet EV samples resulted in a change in PSD, and the platelet EVs do not longer follow a power-law function.
These findings lead to the following conclusion: Freezing and thawing, as well as residual platelets, can result in a change in PSD of EVs.
0 200 400 600 800 2 4 6 8 10 Diameter (nm) lo g (c o n ce n tr at io n m l-1) Platelet EV High Medium Low 0 200 400 600 800 2 4 6 8 10 Diameter (nm) lo g (c o n ce n tr at io n m L -1 ) Activated platelet EV High Medium Low A B
The observation that high concentrations of residual platelets correlated with more EVs is in line with previous research (Jamaly et al., 2018: Gasecka et al. 2020). As well as the fact that the EVs followed a power-law distribution (Van der Pol et al., 2014). The results of this study however showed that with medium and high concentrations of residual platelets the EVs did no longer follow a power-law distribution. Earlier, three possible ways which can cause a change in PSD were explained: aggregation, breakdown and shrinking. Higher residual platelet concentrations showed higher concentrations of EVs sized 300-400 nm, and lower concentrations of EVs sized < 300 nm (figure 4A). This could be explained if aggregation of the <300 nm platelet EVs occurred. But since it is not possible to distinguish an EVs from residual platelets when the residual platelets are <1000 nm, the increase in 300-400 nm particles could also be due to fragmentation of residual platelets. Activated platelet EVs showed a decrease in power-law slope when comparing high concentration residual platelet samples to low concentrations residual platelet samples. This could be explained if there is aggregation of small vesicles, resulting in a shift in PSD in the direction of larger EVs.
For the freeze-thaw analysis with different anticoagulants erythrocyte EVs in EDTA showed both an increase in slope and Yintercept of the semi-log fit after freezing and thawing. This can indicate breakdown of larger vesicles and as a consequence, a shift in PSD towards the small vesicles. Leukocyte EVs in EDTA as well as platelet EVs in citrate showed a vertical shift downward after freezing and thawing. This indicates that the shape of the PSD remained consistent before and after freezing and thawing. According to the hypothesis this would indicate shrinking of the vesicles, since a decrease in EV concentration over the entire EV population was observed. Erythrocyte and leukocyte EVs in citrate as well as platelet EVs in EDTA did not show a change in function after freezing and thawing. This means that neither aggregation, breakdown or shrinking of vesicles occurred after freezing and thawing. These results show that for erythrocytes EVs in EDTA anticoagulated blood possibly breakdown of large vesicles occurs, but that the vesicles remain stable in citrate.
The 200-400 nm EVs most often showed a decrease in EV concentration after freezing and thawing. These results are in contrast with previous findings which say that small EVs are more stable than large EVs (Coumans et al., 2017). The observation that freezing and thawing can result in a shift in PSD, is in line with previous research. Since it has been shown that the PSD can increase after a freeze-thaw cycle (Park et al., 2018). However, the results from the semi-log analysis and the 200 and 50 nm gate analysis were conflicting. All EVs, except erythrocyte EVs in EDTA, showed a decrease in 200-400 nm EVs, but these EVs did not all show a decrease in slope in the semi-log analysis, which would be expected. This could be due to the fact that for 600-800 nm EVs <10 EVs per 10 nm gate were measured by the flow cytometer. The low counts result in inaccurate calculation of the EV concentration in the 600-800 nm EV groups. As a consequence, the semi-log fit was not accurate, possibly resulting in no significant differences in slope. In addition, the counts per size group were not equal, for the small >500 counts per 10 nm gate were measured, as opposed to <10 counts for the large EVs. This also affects the accuracy of statistical tests. An other observation was that the freeze-thaw data did not follow a power-law function, which is in contrast with previous research with shows that EVs follow a power-law function (Van der Pol et al., 2014). This is possibly due to the sensitivity of the flow cytometer which was used. The flow cytometer measured both forward and side scatter, instead of just side scatter. This results the
detection of more large vesicles and less small vesicles, hence the data not following a power-law function.
Overall these results give insight in the sample quality of EVs and can help with keeping samples consistent. Maintaining consistency between samples is important to be able to accurately compare results between laboratories. As well as intra-laboratory comparisons. Future research could focus on EM to visualize what happens with the vesicles. To see whether aggregation of vesicles occurs, or that it is the fragmentation of residual platelets which causes the change in PSD.
Because EVs are involved in processes like inflammation, coagulation and cancer it is important to know more about how EVs behave. Knowing more about the sample quality of EVs can support future research in the field.
References
1. Arraud, N., Linares, R., Tan, S., Gounou, C., Pasquet, J. M., Mornet, S., & Brisson, A. R. (2014).
Extracellular vesicles from blood plasma: determination of their morphology, size, phenotype and concentration. Journal of Thrombosis and Haemostasis, 12(5), 614-627.
2. Artoni, A., Merati, G., Padovan, L., Scalambrino, E., Chantarangkul, V., & Tripodi, A. (2012). Residual platelets are the main determinants of microparticles count in frozen-thawed
plasma. Thrombosis research, 130(3), 561–562.
https://doi.org/10.1016/j.thromres.2012.04.012
3. Coumans, F. A., Brisson, A. R., Buzas, E. I., Dignat-George, F., Drees, E. E., El-Andaloussi, S., ... & Lacroix, R. (2017). Methodological guidelines to study extracellular vesicles. Circulation research, 120(10), 1632-1648.
4. Gasecka, A., Nieuwland, R., Budnik, M., Dignat-George, F., Eyileten, C., Harrison, P., ... & Pluta, K. (2020). Randomized controlled trial protocol to investigate the antiplatelet therapy effect on extracellular vesicles (AFFECT EV) in acute myocardial infarction. Platelets, 31(1), 26-32.
5. Jeyaram, A., & Jay, S. M. (2018). Preservation and storage stability of extracellular vesicles for
therapeutic applications. The AAPS journal, 20(1), 1.
6. Kriebardis, A. G., Antonelou, M. H., Georgatzakou, H. T., Tzounakas, V. L., Stamoulis, K. E., &
Papassideri, I. S. (2016). Microparticles variability in fresh frozen plasma: preparation protocol and storage time effects. Blood transfusion = Trasfusione del sangue, 14(2), 228–237. https://doi.org/10.2450/2016.0179-15
7. Lacroix, R., Judicone, C., Poncelet, P., Robert, S., Arnaud, L., Sampol, J., & DIGNAT‐GEORGE, F. (2012). Impact of pre‐analytical parameters on the measurement of circulating microparticles: towards standardization of protocol. Journal of Thrombosis and Haemostasis, 10(3), 437-446.
8. Lacroix, R., Judicone, C., Mooberry, M., Boucekine, M., Key, N. S., & Dignat-George, F. (2013). Standardization of pre-analytical variables in plasma microparticle determination: results of the International Society on Thrombosis and Haemostasis SSC Collaborative workshop. Journal of thrombosis and haemostasis: JTH.
9. Lawrie, A. S., Harrison, P., Cardigan, R. A., & Mackie, I. J. (2008). The characterization and impact of microparticles on haemostasis within fresh‐frozen plasma. Vox sanguinis, 95(3), 197-204.
10. Lőrincz, Á. M., Timár, C. I., Marosvári, K. A., Veres, D. S., Otrokocsi, L., Kittel, Á., & Ligeti, E. (2014). Effect of storage on physical and functional properties of extracellular vesicles derived from neutrophilic granulocytes. Journal of extracellular vesicles, 3(1), 25465.
11. Mitchell, A. J., Gray, W. D., Hayek, S. S., Ko, Y. A., Thomas, S., Rooney, K., ... & Searles, C. D. (2016). Platelets confound the measurement of extracellular miRNA in archived plasma. Scientific reports, 6, 32651.
12. Park, S. J., Jeon, H., Yoo, S. M., & Lee, M. S. (2018). The effect of storage temperature on the biological activity of extracellular vesicles for the complement system. In Vitro Cellular & Developmental Biology-Animal, 54(6), 423-429.
13. Qin, B., Zhang, Q., Hu, X. M., Mi, T. Y., Yu, H. Y., Liu, S. S., ... & Xiong, K. (2020). How does temperature play a role in the storage of extracellular vesicles?. Journal of Cellular Physiology. 14. Van der Pol, E., Coumans, F. A. W., Grootemaat, A. E., Gardiner, C., Sargent, I. L., Harrison, P., ... & Nieuwland, R. (2014). Particle size distribution of exosomes and microvesicles determined by transmission electron microscopy, flow cytometry, nanoparticle tracking analysis, and resistive pulse sensing. Journal of Thrombosis and Haemostasis, 12(7), 1182-1192.
15. Yuana, Y., Böing, A. N., Grootemaat, A. E., van der Pol, E., Hau, C. M., Cizmar, P., ... & Nieuwland, R. (2015). Handling and storage of human body fluids for analysis of extracellular vesicles. Journal of extracellular vesicles, 4(1), 29260.
Subliminal 1
Figure A: Platelet EV concentration related to the concentration of residual platelets in sample. The concentration of platelet EVs (CD61+) increases when there are more residual platelets present in the sample (slope=8.399).
(Data from Gasecka et al. 2020)
0 5×107 1×108
0.0 5.0×108 1.0×109 1.5×109
Residual platelet concentration (mL-1)
E V c o n ce n tr at io n ( m L -1 )
Platelet EV concentration related to concentration residual platelets
Subliminal 2
Figure B: power-law fit on erythrocyte EV data. Values are median concentrations (n=13) and the values were fitted with a log power-law function (red and black line). Sum of square F-test shows that the data does not follow the power-law function (p<0.0001).
Subliminal 3
Table A: Results of semi-log fit on EV PSD data.
EV type Anticoagulant Freeze-thaw cycle Paired t-test
Before After P-value
Erythrocyte Citrate B* -0.0002237 B -0.0002346 0.220 A** 0.8509 A 0.8447 0.174 EDTA B -0.0001925 B -0.0002525 <0.0001 A 0.8340 A 0.8594 <0.0001 Leukocyte Citrate B -0.0002849 B -0.0002595 0.083 A 0.8255 A 0.8093 0.264 EDTA B -0.0003371 B -0.0002981 0.159 A 0.8516 A 0.8258 0.030 Platelet Citrate B -0.0003285 B -0.0003034 0.540 A 0.8927 A 0.8773 0.013 EDTA B -0.0003209 B -0.0003057 0.360 A 0.8586 A 0.8461 0.116
*B: slope, **A: Yintercept.
Table B: Results of power-law fit on EV samples with high, medium and low concentrations residual platelets.
EV type Power-law* parameter
Residual platelet concentration Paired t-test
High medium p-value
Activated platelet A 0.7864 0.7614 0.0001 B -7.133*10^-5 -8.823*10^-5 0.135 EV type Power-law parameter
Residual platelet concentration Paired t-test
High low p-value
Activated platelet A 0.7864 0.7234 <0.0001 B -7.133*10^-5 -1.232*10^-4 <0.0001 EV type Power-law parameter
Residual platelet concentration Paired t-test
medium low p-value
Activated platelet
A 0.7614 0.7234 <0.0001
B -8.823*10^-5 -1.232*10^-4 0.013
