Variation in activation marker expression within the platelet population – a new parameter for evaluation of platelet flow cytometry data

Abstract

In flow cytometry, individual cells are investigated. Platelet activation is normally reported in form of percentage of platelets expressing the marker (positive platelets) and/or mean/median fluorescence intensity (MFI) for the entire analyzed population. None of these take into account the variance of the marker expression between individual platelets. This can be obtained as data on coefficient of variation (CV). This study explores if CV provides additional information regarding platelet function. Samples from platelet concentrates (PCs) prepared by apheresis- (n = 13) and interim platelet unit (IPU) technique (n = 26) and stored for 6–7 days were included and compared. Spontaneous- and agonist-induced expression of activation markers (fibrinogen binding and exposure of P-selectin, LAMP-1, and CD63) was investigated as percentage positive platelets, MFI and CV. Spontaneous expression of P-selectin as percentage positive platelets and MFI was higher for IPU PCs than apheresis PCs, which in contrast had higher agonist-induced activation. CV for spontaneous fibrinogen binding and P-selectin exposure was larger for apheresis PCs, while IPU PCs generally had larger CV for P-selectin, LAMP-1, and CD63 after agonist stimulation. Our findings show that CV adds additional information when assessing platelet activation by providing data on the variation in activation responses within the platelet population.

Keywords:

• Blood platelets
• coefficient of variation
• CV
• flow cytometry
• platelet activation

Introduction

Flow cytometry allows multiple factors regarding blood platelets to be tested simultaneously with high sensitivity [1]. The low amount of blood needed is an advantage in pediatric patients, and the analysis of properties of single platelets enables testing of thrombocytopenic patients [2]. In addition, testing can be performed in diluted whole blood [3], which removes the risk of losing bigger or denser platelets during centrifugation to platelet-rich plasma [4,5].

However, although flow cytometry includes analysis of fluorescence signals from individual cells, data regarding platelet activation is traditionally reported either as percentage positive platelets or as mean/median fluorescence intensity for the total platelet population (MFI) [6]. A limitation with both percentage positive platelets and MFI is that they do not take into account the spread in binding of the activation marker between the individual platelets. As a possible solution, many software programs for analysis of flow cytometry data may also extract information regarding the coefficient of variation (CV) for the fluorescence signal. This information might add an additional dimension of information regarding the platelet function, which is the variation in response between the individual platelets that makes up the analyzed platelet population (see orange boxes in Figure 1).

Figure 1. The potential and limitations of the different ways to report platelet activation responses, the percentage positive platelets (“%Gated,” red boxes), median fluorescence intensity (“X-Med,” blue boxes) and variation of the fluorescence response (“X-CV,” Orange boxes), as illustrated with (A) a sample with low activation (resting platelets, P-selectin as activation marker), (B) a sample with moderately activated platelets (fibrinogen binding as activation marker) and (C) a sample with strongly activated platelets (P-selectin as activation marker). Results for “All” refers to all particles within the “All platelets” gate shown in Figure 2.

The purpose of this study was to explore if analysis of the CV would be able to add any additional information to the analysis of platelet function by flow cytometry. For this purpose, we compared samples from 6 to 7 days stored platelet concentrates (PCs) prepared by two different methods, apheresis technique (n = 13) and interim platelet unit (IPU) (n = 26) technique. We demonstrate that platelets in IPU PCs show higher pre-activation, lower responses to platelet stimulation and a larger activation response variation within the platelet population than platelets in apheresis PCs.

Methods

The study was conducted according to the guidelines of the Declaration of Helsinki and was approved by the regional ethical review board in Linköping (Dnr 2017/568-31).

Platelet concentrates

Subsamples were collected using sterile sampling bags (MacoPharma, Mouvaux, France) from PCs produced by apheresis- [7] or “interim platelet unit” (IPU) [8] technique after storage for 6–7 days in conjunction with transfusion. The maximum storage time in Linköping is 7 days. The samples were transferred gently from the PC to the sampling bag to limit any pre-activation of the samples and flow cytometry carried out directly. The apheresis PCs were prepared using two different machines, but with no difference in their platelet collection programs, Trima Accel® (Version 7. Terumo BCT, INC. Lakewood, CO, USA), and Spectra Optia® (Terumo BCT). The IPU PCs were manufactured by Reveos® (Terumo BCT) from whole blood donations from four to six different donors.

All PCs were prepared from the regular donor populations. All blood donors and apheresis platelet donors met standard donation criteria. This included no use of NSAID prior to donation and the apheresis donors should have a platelet count above 230 and blood group A or O. All apheresis PCs included in the study had been donated by male donors, and there was no selection based on blood group. Each donation by apheresis yielded a median of 2 PCs. The apheresis platelets were stored with 100% plasma and the IPU platelets in a combination of plasma (approximately 38% plasma) and SSP+ (MacoPharma).

The platelet concentrate subsamples were diluted 1:4 in HEPES buffer prior to flow cytometry.

Antibody staining

Antibodies used were anti-fibrinogen-FITC (Diapensia AB, Linköping, Sweden), anti-GPIb-PE (clone AN51, Dako, Glostrup, Denmark), anti-P-selectin-APC (clone AK-4), isotype IgG1κ-APC (clone MOPC-21), anti-LAMP-1-PE-Cy7 (clone H4A3), anti-CD63-PE-Cy7 (clone H5C6), and isotype IgG1κ-PE-Cy7 (clone MOPC-21), all from BD, Franklin Lakes, NJ, USA. Platelet agonists used were PAR1-activating peptide (PAR1-AP, sequence SFLLRN, JPT Peptide Technologies, Berlin, Germany), PAR4-activating peptide (PAR4-AP, sequence AYPGKF, JPT Peptide Technologies) and cross-linked collagen-related peptide (CRP-XL, University of Cambridge, Cambridge, UK). The HEPES buffer (137 mM NaCl, 2.7 mM KCl, 1 mM MgCl₂, 20 mM HEPES, 1 g/L bovine serum albumin and 5.6 mM glucose, pH 7.4) was used as is, or with added CaCl₂ (2.5 mM) or EDTA (10 mM).

Master mixes containing antibodies, platelet agonists, and dilution buffer were prepared beforehand and kept at −80°C. Final concentrations (during platelet activation) were the following: anti-fibrinogen-FITC (2.5 µg/mL), anti-GPIb-PE (2.5 µg/mL), anti-P-selectin-APC (0.17 µg/mL), anti-CD63-PE-Cy7 or anti-LAMP-1-PE-Cy7 (0.5 µg/mL), PAR1-AP (10 µM), PAR4-AP (100 µM), and CRP-XL (0.15 µg/mL). For setting of the positivity gates, two negative control samples were used. These contained anti-fibrinogen-FITC and anti-GPIb-PE in the same concentrations as stated above, but isotype IgG1κ-APC (0.17 µg/mL) and isotype IgG1κ-PE-Cy7 (0.5 µg/mL) instead of the APC- and PE-Cy7-labeled antibodies mentioned above. One of these samples contained HEPES with EDTA (10 mM), and served as the negative control for anti-fibrinogen-FITC, as previously described [9], while the other contained ordinary HEPES buffer and served as negative control for the APC- and PE-Cy7-labeled antibodies. In all other samples, the dilution buffer was HEPES with CaCl₂ (2.5 mM). The approach to report the percentage of positive platelets requires a careful selection and titration of the negative (background fluorescence) control used for gating. Both antibody subclass, concentration and fluorescence per protein (F/P) ratio were therefore thoroughly matched to ensure that the peaks of the negative control and main population of resting platelets show a good overlay [10] (see example in Figure 1a).

On the day of sample analysis, the master mixes were thawed at 37°C in a water bath for 3 min, and 33 µL master mix was added to flow cytometry tubes. After this, 3 µL thoroughly mixed, diluted PC was added to all tubes with 15 seconds intervals. After exactly 10 min, the platelet activation was stopped by addition of 2 mL HEPES buffer to tube 1–2 and HEPES buffer with 2.5 mM CaCl₂ to the other tubes. After every addition of PC or buffer, the tubes were gently mixed with a vortex-mixer at low speed.

Analysis with flow cytometry

Samples were analyzed on a Gallios® flow cytometer (Beckman Coulter, Brea, CA, USA) equipped with three lasers and a 10 color configuration. The ultra-wide angle of detection (submicron particle setting) was used for forward scatter (FSC) and a fluorescence threshold was applied (here on FL2 (GPIb-PE)) to allow reliable detection of all platelet-derived particles (as previously reported in [11]). Data from 5000 platelet-sized, GPIb-positive events were collected using a temporary platelet gate in the acquisition protocol.

The data analysis was performed using the software Kaluza (Version 2.1, Beckman Coulter Life Sciences, Indianapolis USA). The manual for Kaluza, including all data on formulas used for calculation of the different variables, is available at the Beckman Coulter web page [12]. The reporting of percentage positive platelets is a good alternative in cases with moderate platelet activation [6] (see red box in Figure 1b), and in situations where only a subpopulation of platelets become positive, which is the case for markers such as annexin V detecting phosphatidylserine exposure [13,14]. However, a drawback with the percentage positive platelets is that it loses sensitivity when the majority of platelets become activated, as in Figure 1c. In these cases, the average fluorescence for the total platelet population (MFI) is a better representative of the true difference between the analyzed samples (see blue boxes in Figure 1b, c). The use of median is recommended over mean, as fluorescence intensity increases logarithmically [15].

In Kaluza, the single platelets were identified according to the relation between forward scatter height and area, as illustrated in Figure 2a. Platelets were distinguished from platelet fragments based on size (forward scatter) and PE fluorescence, as illustrated in Figure 2b. This gate included in median 98% GPIb positive platelets for the apheresis PCs and 97% for the IPU PCs. Figure 2c shows how the negative control sample was used to determine the gate for P-selectin positive platelets, being placed as close to 1.5% as possible, as the recommendation is to place it between 1 and 2% [16], while Figure 2d shows the same blood sample in the presence of the P-selectin antibody, and demonstrates the data extracted and analyzed, namely the percentage positive platelets (%Gated in “All plt anti-P-sel pos”), as well as the median fluorescence (X-Med), and the coefficient of variation for the fluorescence (X-CV) for the total platelet population (Gate “MFI P-selectin”).

Figure 2. The gating strategy used for analysis of the PCs in this study. PE-positive particles were collected using a threshold on FL2 at acquisition. The single platelets were identified based on FS height and area (A) and then distinguished from platelet fragments through PE fluorescence (B) with a platelet gate big enough to include also platelets with a change in forward scatter due to activation and procoagulant transformation. The setting of the gate for P-selectin positive platelets as close as possible to 1.5% with the use of the negative (isotype antibody) control sample (C). Data obtained from an activated sample containing P-selectin antibody when the above presented gating strategy was used (D).

Effect of platelet events on X-CV

A separate flow cytometry test was performed when 5000, 10 000, 20 000, and 50 000 platelet events were collected in order to assess if X-CV changed with increasing platelet counts. For this experiment, platelets from four PCs were assessed and fibrinogen binding, P-selectin exposure and LAMP-1 exposure determined to illustrate the impact of varying the platelet events collected.

Statistical analysis

Statistical analysis was performed using GraphPad (Prism 8.4.2, GraphPad Software, San Diego, CA, USA). As not all data were normally distributed, the Mann-Whitney test was used to compare the data for apheresis and IPU units for the same platelet activation readout. A P-value of <0.05 was considered statistically significant.

Results

Platelet activation is measured as the amount of certain markers on the surface, indicating activation of e.g. the fibrinogen receptor or exocytosis of granules and lysosomes. Platelets from PCs prepared by IPU- and apheresis technique were incubated with antibodies detecting fibrinogen binding (indicating activation of the fibrinogen receptor GPIIb/IIIa [9]), and markers indicating degranulation of platelet granules (P-selectin from alpha and dense granules, lysosomal associated membrane protein 1 (LAMP-1) from lysosomes and CD63 from dense granules and lysosomes [16]). For platelet activation, we used peptides as specific agonists to induce activation via three different pathways, through the thrombin receptors PAR1 [17] and PAR4 [18] and through the collagen receptor GPVI [19]. These three pathways also include secondary feedback activation through release of endogenous ADP by activated platelets [1]. Samples without platelet agonist were also analyzed to determine the level of spontaneous platelet activation induced by the preparation and storage of the PCs and expression of the constitutively expressed GPIb receptor. Data for percentage positive platelets, median fluorescence intensity (MFI) and variation in fluorescence signal (CV) for the different platelet markers are shown in Table 1.

Table 1. Median, confidence interval (CI) and p-value for apheresis and IPU with all tested agonists and markers. Significant results are marked in bold.

Agonist (final concentration)	Activation marker, investigated parameter	Apheresis PCs (n = 13)	IPU PCs (n = 26)	P-value	Significance
HEPES buffer (spontaneous activation)	Fibrinogen binding, % pos	4.1 (3.4–4.4)	4.2 (3.6–4.5)	0.6011	No
	Fibrinogen binding, MFI	0.54 (0.48–0.57)	0.60 (0.59–0.67)	0.0006	Yes
	Fibrinogen binding, CV	241 (159–334)	133 (118–159)	0.0250	Yes
	P-selectin exposure, % pos	48.6 (42.5–61.8)	64.5 (61.1–68.6)	0.0001	Yes
	P-selectin exposure, MFI	1.43 (1.10–2.05)	2.17 (1.86–2.37)	0.0010	Yes
	P-selectin exposure, CV	231 (192–269)	189 (180–202)	0.0005	Yes
	LAMP-1 exposure, % pos	8.4 (6.2–9.1)	8.1 (6.7–8.6)	0.8083	No
	LAMP-1 exposure, MFI	0.20 (0.18–0.20)	0.19 (0.18–0.19)	0.3002	No
	LAMP-1 exposure, CV	60 (53–98)	62 (58–67)	0.4008	No
	GPIb expression, MFI	24.37 (21.28–28.49)	23.40 (21.90–25.24)	0.4484	No
	GPIb expression, CV	44 (40–47)	49 (48–54)	<0.0001	Yes
PAR1-AP (10 µM)	Fibrinogen binding, % pos	30.7 (22.8–45.8)	20.1 (14.6–27.0)	0.0022	Yes
	Fibrinogen binding, MFI	1.82 (1.46–1.98)	1.37 (1.12–1.51)	0.0002	Yes
	Fibrinogen binding, CV	113 (93–199)	102 (97–114)	0.4525	No
	P-selectin exposure, % pos	98.3 (97.0–99.0)	96.6 (95.8–97.4)	0.0001	Yes
	P-selectin exposure, MFI	26.83 (23.44–30.26)	21.65 (19.41–24.37)	0.0009	Yes
	P-selectin exposure, CV	74 (66–85)	84 (78–87)	0.0020	Yes
	CD63 exposure, % pos	84.5 (81.7–87.8)	74.25 (67.20–78.70)	<0.0001	Yes
	CD63 exposure, MFI	1.00 (0.88–1.10)	0.62 (0.51–0.74)	<0.0001	Yes
	CD63 exposure, CV	85 (79–152)	112 (105–156)	0.0009	Yes
PAR4-AP (100 µM)	Fibrinogen binding, % pos	33.7 (23.3–50.7)	19.8 (13.8–23.3)	0.0007	Yes
	Fibrinogen binding, MFI	1.77 (1.63–2.20)	1.20 (1.10–1.57)	<0.0001	Yes
	Fibrinogen binding, CV	96 (69–118)	95 (89–98)	0.7353	No
	P-selectin exposure, % pos	98.0 (97.0–99.0)	95.0 (94.0–96.0)	<0.0001	Yes
	P-selectin exposure, MFI	24.54 (20.11–26.34)	14.81 (12.87–15.88)	<0.0001	Yes
	P-selectin exposure, CV	73 (70–77)	91 (86–96)	0.0003	Yes
	LAMP-1 exposure, % pos	48 (38–51)	19 (15–24)	<0.0001	Yes
	LAMP-1 exposure, MFI	0.35 (0.30–0.37)	0.21 (0.20–0.23)	<0.0001	Yes
	LAMP-1 exposure, CV	66 (64–72)	70 (67–76)	0.1188	No
CRP-XL (0.15 µg/mL)	Fibrinogen binding, % pos	61.9 (41.6–69.1)	41.3 (37.2–50.0)	0.0063	Yes
	Fibrinogen binding, MFI	2.89 (2.27–3.49)	2.18 (1.86–2.63)	0.0288	Yes
	Fibrinogen binding, CV	97 (80–131)	101 (91–109)	0.8485	No
	P-selectin exposure, % pos	96.5 (92.7–97.7)	94.8 (93.3–96.4)	0.2220	No
	P-selectin exposure, MFI	24.92 (18.18–27.68)	20.93 (17.80–24.20)	0.2944	No
	P-selectin exposure, CV	81 (75–97)	89 (83–97)	0.1117	No
	LAMP-1 exposure, % pos	44.0 (31.7–57.0)	33.2 (28.2–35.8)	0.0084	Yes
	LAMP-1 exposure, MFI	0.33 (0.25–0.39)	0.25 (0.24–0.29)	0.0064	Yes
	LAMP-1 exposure, CV	71 (69–74)	77 (73–81)	0.0313	Yes

Results are presented as median and 95% CI within parenthesis for PCs stored for 6–7 days. Significant difference between apheresis PCs and IPU PCs was considered when the p-value was <0.05 (Mann-Whitney test). PC = platelet concentrate, % pos = percentage positive platelets, MFI = median fluorescence intensity, CV = coefficient of variation, PAR1-AP = PAR1-activating peptide, PAR4-AP = PAR4-activating peptide, CRP-XL = cross-linked collagen-related peptide.

Without any added platelet agonist, the platelets showed a relatively low binding of fibrinogen as well as low exposure of LAMP-1, which indicates a low degree of spontaneous activation of the fibrinogen receptor and lysosomal exocytosis. Using our standardized gating strategy with 1.5% positive in the isotype sample, the exposure of P-selectin was relatively high, and IPU PCs showed significantly higher exposure, both for percentage positive platelets and MFI, indicating that spontaneous alpha granule exocytosis has occurred in the PCs. As a reference, we assessed the basal expression in PCs stored for 2 days for comparison. It was shown to be 44% for IPU PCs (n = 2) and 22% for apheresis PCs (n = 3).

The CV revealed a wider distribution of GPIb in IPU concentrates, while the distributions for fibrinogen binding and P-selectin exposure were wider in apheresis PCs.

For all samples stimulated with platelet agonists, the fibrinogen binding was significantly higher on platelets produced by apheresis technique than IPU technique when expressed as percentage positive platelets and MFI, whereas CV was similar. In contrast, both PAR receptor activating peptides induced significantly higher P-selectin exposure (both percentage positive platelets and MFI) in apheresis PCs, but showed a higher CV in IPU PCs. A similar pattern with higher degree of activation in apheresis PCs but higher CV in IPU PCs was seen for the other markers of degranulation, LAMP-1 and CD63 upon stimulation with platelet agonists (as they were labeled with the same fluorophore, they could not be investigated in the same samples). There was no difference in the results for Trima Accel® and Spectra Optia® (data not shown). None of the PCs had any visible aggregates and all exhibited well-maintained swirling capacity.

Our data also show that X-CV was not systematically altered by increasing the number of platelet events in the assay above 5000 (Table 2). Hence, 5000 platelets should be enough to ensure stable X-CV data. It is anyway advisable to collect the same number of events in all samples to be compared. However, as could be seen for LAMP-1, very low MFI could cause a larger discrepancy in X-CV (see PC A in Table 2).

Table 2. Impact of the number of platelet events investigated on the flow cytometry variables.

Activation marker, investigated parameter	Platelet events investigated	PC A	PC B	PC C	PC D
Fibrinogen binding, % pos	5000	42.2	10.7	33.2	19.0
	10 000	43.8	11.4	33.5	20.4
	20 000	44.4	11.8	34.6	22.0
	50 000	45.5	12.9	35.9	22.9
Fibrinogen binding, MFI	5000	1.2	0.7	1.1	0.8
	10 000	1.3	0.7	1.1	0.8
	20 000	1.3	0.7	1.2	0.9
	50 000	1.3	0.8	1.2	0.9
Fibrinogen binding, CV	5000	125	102	370	395
	10 000	126	85	384	382
	20 000	128	120	372	402
	50 000	133	151	381	420
P-selectin exposure, % pos	5000	69.7	17.3	67.5	59.9
	10 000	71.9	16.3	68.5	61.0
	20 000	71.5	17.7	69.6	62.3
	50 000	71.5	17.2	69.6	62.4
P-selectin exposure, MFI	5000	4.2	0.8	3.7	2.7
	10 000	4.6	0.8	3.6	2.7
	20 000	4.7	0.8	3.9	2.8
	50 000	4.5	0.8	3.9	2.8
P-selectin exposure, CV	5000	128	172	138	143
	10 000	123	158	130	150
	20 000	127	172	132	140
	50 000	124	177	132	141
LAMP-1 exposure, % pos	5000	15.3	5.1	11.9	4.7
	10 000	16.0	4.9	11.6	4.9
	20 000	15.8	5.5	12.6	5.4
	50 000	16.5	5.1	13.2	5.5
LAMP-1 exposure, MFI	5000	0.6	0.5	0.6	0.5
	10 000	0.6	0.5	0.6	0.5
	20 000	0.6	0.5	0.6	0.5
	50 000	0.6	0.5	0.6	0.5
LAMP-1 exposure, CV	5000	75	43	56	52
	10 000	59	64	78	55
	20 000	431	42	76	56
	50 000	89	50	62	55

Results from 4 separate PCs investigated (A-D). PC = platelet concentrate, % pos = percentage positive platelets, MFI = median fluorescence intensity, CV = coefficient of variation.

Discussion

One often stated advantage with flow cytometry is that it is able to analyze the fluorescence of each individual cell separately instead of giving a bulk average value for all cells in the sample, as is the case for many other platelet activation analysis methods such as aggregometry and western blotting. Considering that, it is surprising that most studies using flow cytometry for platelet activation analysis only reports one single value for the total platelet population, either the percentage positive platelets or the average fluorescence for the total platelet population, such as median fluorescence intensity (MFI). The consensus protocol published in 1998 states that the percentage positive platelets will be the most sensitive readout in situations with low platelet activation, but that the average fluorescence intensity (or a calibrated equivalent) should be used in situations where the majority of platelets express the antigen [16]. However, in this study, we have reported both percentage positive platelets and MFI for the activation-induced markers, as it allows us to illustrate and discuss the differences and limitations of these two approaches.

As mentioned above and illustrated in Figure 1, the data show that although the percentage positive platelets and the MFI generally follows the same pattern, the percentage loses sensitivity in samples with high activation, and although significant differences are present, these are probably too small to be safely interpreted as relevant. A similar situation occur in samples with very low platelet activation, where it is common to find studies reporting on significant differences in marker expression, but where the absolute difference is only a few percent. It is important to remember that even though these differences might be statistically significant, the clinical relevance of differences of such a minute magnitude needs to be critically reviewed, also taking into account the sampling uncertainty and the imprecision of the specific analysis protocol at the relevant range of activation [20,21]. In addition, the placement of the positivity gate may have a large impact on the results, as has been illustrated in a systematic study for P-selectin [22].

Our aim with this study was to explore another parameter available from most flow cytometry analysis software, the coefficient of variation (CV) of the fluorescence signal for a specific marker. In comparison with percentage positive platelets or MFI, the CV adds another dimension as it reports on the variation in fluorescence in the platelet population and can be obtained in standard flow cytometry programs as a supplement to percentage positive platelets and MFI. It provides an objective measure on variation in responses in a population, in contrast to manual categorization of events into groups. This allows the capture of at least some of the potential of the flow cytometer to evaluate individual platelets. To the best of our knowledge, the CV has not been reported before in studies of platelet function by flow cytometry.

To evaluate the usefulness of CV, we choose single donor PCs and PCs with platelets from multiple donors that were close to expiry date (stored for 6–7 days) and therefore most likely would show variable activation response caused by varying degree of platelet storage lesions.

Here, we show that CV detects differences between apheresis and IPU PCs that are not just mirroring the differences in MFI or percentage positive platelets. On the contrary, the CV shows the opposite pattern with larger variation in spontaneous expression of activation markers for platelets from apheresis PCs, but larger variation for platelets from IPU PCs after agonist stimulation for most activation markers and activation pathways tested. The large variation within the IPU PCs indicate that the activation potential of the platelets is very different within one PC. This could possibly be a consequence of the method for preparation (such as centrifugation speed) or storage conditions (including a difference in storage medium), or by the fact that these PCs are composed of platelets from several different donors, or a combination of these. If and how this affects the platelet’s ability to circulate and activate after donation to the recipient is yet unknown. The apheresis PCs had been prepared by two different machines, but both machines have similar collection principle for platelets, and the results did not differ between these.

One limitation of CV is in situations with a mean close to zero which could result in a highly variable CV. This could occur for platelets with a low expression of the markers of interest (exemplified with LAMP-1 in Table 2). In such situations, the percentage positive platelets is a more appropriate variable. CV is thus most suitable in the setting of medium-to-high expression of a marker. It could be of interest in situations where one is interested in knowing how homogenous a population of platelets is in its responsiveness to activation or in variation in expression of surface receptors with potential shedding or internalization, such as GPIb. In the future, it could be a useful variable if one is interested in tailoring platelet transfusions using platelets with specific characteristics, such as situations with acute bleeding or in prophylactic transfusions. In some situations, a homogenous population might be preferable whereas in others a heterogenic might give a better outcome.

Further studies including the CV as parameter will give platelet researchers a tool to investigate these and many other questions regarding platelet activation responses in the future. This is definitely not limited to the field of transfusion medicine, but is likely as interesting to explore in other areas of clinical and experimental platelet research.

Conclusions

We conclude that the inclusion of CV as a parameter for analysis of platelet flow cytometry data is a way to capture some of the technique’s potential to investigate individual platelet responses. We suggest that the CV, together with either the percentage positive platelets (in cases with medium platelet activation) or MFI (in cases with high platelet activation), could be a valuable tool for platelet research in both clinical and experimental laboratories in the future.

Author contributions

SR and NT designed and supervised the study. FM analyzed the data and wrote the initial draft of the manuscript and AA re-worked it for submission with a new focus. All authors took part in the revision and editing of the final version of the manuscript, and have read and agreed to the final version.

Acknowledgements

The authors wish to thank Maria Wallstedt and Anna Södergren for their contributions in the setup of the study and practical analysis of the samples and Helena Lindblom and Ana Maria Vasilache for arrangements surrounding sample collection.

Data availability statement

The raw data used for this publication can be provided by the authors upon reasonable request.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research was funded by ALF Grants, Region Östergötland, grants from Region Östergötland, and strategic research grants from Örebro University.