Available via license: CC BY-NC
Content may be subject to copyright.
ORIGINAL RESEARCH
Comparison between manufacturing sites shows differential
adhesion, activation, and GPIbαexpression of cryopreserved
platelets
Katrijn R. Six,
1,2
Willem Delabie,
1
Katrien M.J. Devreese,
2,3
Lacey Johnson,
4
Denese C. Marks,
4,5
Larry J. Dumont,
6,7
Veerle Compernolle,
1,2,8
and Hendrik B. Feys
1,2
BACKGROUND: Transfusion of cryopreserved
platelets (cryoplatelets) is not common but may replace
standard liquid-preserved platelets (PLTs) in specific
circumstances. To better understand cryoplatelet
function, frozen concentrates from different
manufacturing sites were compared.
STUDY DESIGN AND METHODS: Cryoplatelets from
Denver, Colorado (DEN); Sydney, Australia (SYD); and
Ghent, Belgium (GHE) were compared (n = 6). A paired
noncryopreserved control was included in Ghent.
Microfluidic-flow chambers were used to study PLT
adhesion and fibrin deposition in reconstituted blood.
Receptor expression was measured by flow cytometry.
Coagulation in static conditions was evaluated by
rotational thromboelastometry (ROTEM).
RESULTS: Regardless of the manufacturing site, adhesion
of cryoplatelets under shear flow (1000/sec) was
significantly (p < 0.05) reduced compared to control.
Expression of GPIbαwas decreased in a subpopulation of
cryoplatelets comprising 45% 11% (DEN), 63% 9%
(GHE), and 94% 6% (SYD). That subpopulation
displayed increased annexin V binding and decreased
integrin activation. PLT adhesion, agglutination, and
aggregation were moreover decreased in proportion to that
subpopulation. Fibrin deposition under shear flow was
normal but initiated faster (546 163 sec GHE) than control
PLTs (631 120 sec, p < 0.01), only in the absence of
tissue factor. In static conditions, clotting time was faster, but
clot firmness decreased compared to control. Coagulation
was not different between manufacturing sites.
CONCLUSION: Cryopreservation results in a subset of
PLTs with enhanced GPIbαshedding, increased
phosphatidylserine expression, reduced integrin
response, and reduced adhesion to collagen in
microfluidic models of hemostasis. The proportion of this
phenotype is different between manufacturing sites. The
clinical effects, if any, will need to be verified.
ABBREVIATIONS: CRP-XL = cross-linked collagen-related
peptides; cryoplatelet(s) = cryopreserved platelet(s); DEN =
cryoplatelets from Denver, Colorado; GHE = cryoplatelets from
Ghent, Belgium; PC(s) = platelet concentrate(s); ROTEM =
rotational thromboelastometry; RT = room temperature; TF = tissue
factor; SYD = cryoplatelets from Sydney, Australia.
From the
1
Transfusion Research Center, Belgian Red Cross-
Flanders, Ghent, Belgium; the
2
Faculty of Medicine and Health
Sciences, Department of Diagnostic Sciences, Ghent University,
Ghent, Belgium; and the
3
Coagulation Laboratory, Department of
Laboratory Medicine, Ghent University Hospital, Ghent, Belgium;
4
Research & Development, Australian Red Cross Blood Service,
Sydney, Australia; and the
5
Sydney Medical School, University of
Sydney, Sydney, Australia;
6
Blood Systems Research Institute,
Denver, Colorado; the
7
Geisel School of Medicine at Dartmouth,
Lebanon, New Hampshire, USA; and the
8
Blood Service of the
Belgian Red Cross-Flanders, Mechelen, Belgium.
Address reprint requests to: Hendrik B. Feys, Ottergemsesteen-
weg 413, Ghent 9000, Belgium; e-mail: hendrik.feys@rodekruis.be.
This is an open access article under the terms of the Creative
Commons Attribution-NonCommercial License, which permits use,
distribution and reproduction in any medium, provided the original
work is properly cited and is not used for commercial purposes.
This research was supported by the Foundation for Scientific
Research of the Belgian Red Cross Flanders. KS is a fellow of the
Special Research Fund of Ghent University. The Australian govern-
ment funds the Australian Red Cross Blood Service to provide blood
and blood products to the Australian community.
The views, opinions, and/or findings contained in this report
are those of the author (s) and should not be construed as an offi-
cial Department of Defense position, policy, or decision unless so
designated by other documentation.
Received for publication December 22, 2017; revision received
March 22, 2018; and accepted March 22, 2018.
doi:10.1111/trf.14828
© 2018 The Authors. Transfusion published by Wiley
Periodicals, Inc. on behalf of AABB.
TRANSFUSION 2018;9999;1–12
TRANSFUSION 1
The demand for platelet concentrates (PCs) has
increased progressively over the years
1
due to an
aging population and an increased number of
patients suffering from chronic thrombocytope-
nia. Platelets (PLTs) are transfused for prophylaxis in
patients with hematologic malignancies or to treat active
bleeding.
2
Room temperature (RT) storage of PCs results in
better transfusion yields than cold-stored PLTs.
3
However, it
also limits the shelf life to between 4 and 7 days, resulting
in a logistic challenge that all blood banks continuously
face.
4
In addition, these conditions give rise to a substantial
risk of bacterial transmission. Consequently, cold storage
and cryopreservation have remained of interest as an alter-
native to standard PC banking
5–7
for nonprophylactic use,
particularly for remote areas with low-quality transportation
infrastructure or distant rural communities or during mili-
tary operations when transport of liquid-stored PCs is ham-
pered.
8
In selected cases, cryopreserved platelets
(cryoplatelets) could serve as backup product in emergen-
cies for an acutely bleeding patient when supply of regular
liquid stored PC is exhausted.
Several cryopreservation methods have been published
aiming to minimize PLT damage. Dimethyl sulfoxide
(DMSO) is the most commonly used cryoprotectant,
although several others have been investigated.
5,9–11
Most
protocols include hyperconcentration of PLTs before freez-
ing to reduce the amount of DMSO transfused. This
requires reconstitution after thawing, most commonly in
plasma or saline.
12
Cryoplatelets have been transfused in healthy volun-
teers, generally yielding lower count increments compared
to liquid-stored PLTs.
5,13,14
This is caused by biochemical
changes during cryopreservation that can be detected
in vitro including decreased PLT aggregation, deficient sig-
nal transduction, altered morphology, differential expression
of certain membrane markers, and microparticle
release.
14–18
Recent studies suggest that cryoplatelets are
procoagulant,
19,20
but these experiments were performed in
static conditions when hemostasis was not restricted by
convective hydrodynamic forces. Studies of PLT adhesive
and procoagulant function of cryoplatelets under shear flow
are lacking. In addition, comparison between cryopreserva-
tion procedures and cryopreserved products manufactured
in different facilities is not available. This study combines
both questions in a comparison of cryoplatelets from three
unrelated international blood institutions, using
microfluidic-flow chamber perfusion and real-time video
microscopy.
MATERIALS AND METHODS
Study design
Platelet concentrates were prepared and cryopreserved in
three different manufacturing sites: Denver, CO, (DEN);
Sydney, Australia (SYD); and Ghent, Belgium (GHE). Pri-
mary products from DEN and SYD were from apheresis and
stored in plasma with ACD-A. Those from GHE were from
pooling of six buffy coats in CPD-plasma with 65% (vol/vol)
additive solution (SSP+, Macopharma). Cryopreservation
protocols were similar for all three manufacturing sites and
were described previously.
14,21
In brief, a 27% (vol/vol)
DMSO with 0.9% NaCl (wt/vol) in water was added to the
PC to achieve a final concentration of 5% to 6% (vol/vol)
DMSO. These PCs were hyperconcentrated by centrifuga-
tion, resuspended in the remaining supernatant, and then
stored at −80C. Cryopreserved DEN and SYD products
were shipped (World Courier) on dry ice to GHE where
these were immediately transferred to a −80C freezer
until use.
Concentrates from DEN were thawed for 8 minutes at
37C and were then held undisturbed for 30 minutes before
resuspension in 25 mL of 0.9% (wt/vol) NaCl in water at
RT. Those from SYD and GHE were resuspended in 250 mL
of ABO/D-matched plasma (30C) immediately after thaw-
ing for 5 minutes at 37C. After resuspension, SYD and GHE
PLTs were held undisturbed for 1 hour at RT before experi-
ments. Paired liquid-preserved control PCs from DEN and
SYD could not be transported to GHE because of the logistic
challenge. Therefore, paired control PCs were only included
in Ghent (GHE CTR). These were “sham”treated with all
preparation and handling methods but without freezing,
therefore controlling for confounding effects of DMSO,
hyperconcentration, resuspension, and incubation. Prelimi-
nary experiments using microfluidic-flow chambers found
no difference between sham-treated and untreated control
PLTs (data not shown). GHE CTR PLTs were resuspended
in 250 mL of ABO/D plasma and incubated for 1 hour at
RT, as for the paired GHE cryoplatelets. From every
manufacturing site, six unrelated cryoplatelet units were
included (n = 6).
Blood reconstitution for microfluidic-flow chambers
and thromboelastometry
Fresh blood samples were taken from healthy consenting
nonmedicated volunteers in hirudin vacutainers (REF
08128812001, Diapharma Group Inc.) or in sodium citrate
vacutainers (REF 366575, BD Diagnostics). Hirudin vacutai-
ners were used for microfluidic-flow chambers studying
PLT adhesion alone, in the absence of coagulation.
22
Citrate
was used for microfluidic-flow chambers with recalcification
to study PLT adhesion in combination with fibrin deposition
(i.e., coagulation).
23
Blood cells were separated by centrifugation (13 min at
250 ×gwithout brake) generating PLT-rich plasma and
concentrated red blood cells (RBCs). PLT-poor plasma was
obtained by centrifugation of that PLT-rich plasma for
10 minutes at 4500 ×g. Blood reconstitution was performed
by mixing the RBCs with the corresponding PLT-poor
2 TRANSFUSION
SIX ET AL
plasma and with PLTs from the various cryo- or control
groups, aiming for a mean of 40% hematocrit and 250 ×10
9
PLTs/L. Complete blood and PLT counts after reconstitution
were determined with an automated hematology analyzer
(PocH-1000i, Sysmex, Kobe, Japan). The PLT count before
cryopreservation was with the hematology analyzers from
the respective manufacturing sites: PocH-1000i, Sysmex
(GHE); XE 2100D, Sysmex (DEN); and CellDyn Ruby, Abbott
Diagnostics (SYD).
Perfusion-flow chamber experiments
Adhesion of PLTs to immobilized collagen was examined by
measuring surface coverage as a function of time during
perfusion of reconstituted, hirudin anticoagulated blood in
microfluidic channels coated with collagen as described.
22
Briefly, microfluidic-flow chambers were coated with 50 μg/
mL of collagen (horm collagen Type III, Takeda Pharma) in
an isotonic glucose solution (pH 2.7) overnight at 4C.
Reconstituted blood was labeled for 10 minutes at 37C with
3,30-dihexyloxacarbo-cyanine iodide in a final concentration
of 1 μmol/L (Sigma-Aldrich, St Louis, MO). Labeling effi-
ciency was tested using flow cytometry and was not differ-
ent between PCs. A microfluidic pump (Mirus evo
nanopump, Cellix Ltd.) was used to generate a wall shear
stress of 50 dyne/cm
2
corresponding to a wall shear rate of
1000/sec. PLT deposition was recorded as a function of per-
fusion time using real-time video microscopy and accompa-
nying acquisition software (Carl Zeiss, Oberkochen,
Germany). PLT surface coverage (%) was retrieved by image
data analysis (Zen 2 [blue edition], Carl Zeiss). A linear
regression of PLT surface coverage with time was used to
calculate the PLT adhesion rate (Prism Version 6.07,
GraphPad Software Inc.).
Adhesion of PLTs in combination with fibrin deposition
was examined by measuring median fluorescence increase
as a function of time during perfusion of reconstituted and
recalcified blood under perfusion flow, as described.
23
Instead of hirudin, citrate was used to anticoagulate fresh
blood for reconstitution. The reconstituted blood was mixed
during perfusion with one tenth volume of recalcification
buffer (HBS; 10 mmol/L HEPES, 155 mmol/L NaCl, pH 7.4)
containing 100 mmol/L CaCl
2
and 37.5 mmol/L MgCl
2
. Sep-
arately operated syringe pumps (Exigo syringe pump, Cellix)
were used. PLT and fibrin deposition under flow was exam-
ined in channels coated with only collagen for contact acti-
vation or with collagen plus purified recombinant human
lipidated tissue factor (TF, Dade Innovin) at approximately
100 pmol/L in HBS (Siemens Healthcare GmbH) for TF-
based activation. Before perfusion, the reconstituted blood
was spiked with 70 μg/mL of Alexa Fluor 405-labeled fibrin-
ogen (Sigma-Aldrich) in addition to 3,30-dihexyloxacarbo-
cyanine iodide to label PLTs as above. For investigation of
TF-mediated coagulation, reconstituted blood samples were
supplemented with 4 μmol/L corn trypsin inhibitor
(Enzyme Research Laboratories) to inhibit FXIIa-based con-
tact activation. Image acquisition and analysis were as
described.
23
In brief, PLT adhesion rate (/sec) was mea-
sured by linear regression of the green fluorescent signal
increase as a function of perfusion time. The variables
retrieved for fibrin deposition included coagulation rate
(/sec), which is the linear portion of fibrin deposition kinet-
ics and clotting time (sec), which is the lag time indicating
the moment of coagulation onset. This analysis takes into
account thrombus growth in the z-plane. The outcome vari-
ables were extracted from the raw fluorescence data using a
software plugin developed in MatLab (MathWorks). All flow
chamber studies were performed in duplicate and the mean
of two technical repeats was used.
Flow cytometry
Expression of P-selectin (phycoerythrin-anti-CD62P, Life
Technologies), phosphatidylserine (peridinin-chlorophyll-
Cy5.5 Annexin V, BD Biosciences), activated integrin α
IIb
β
3
(fluorescein-labeled PAC1, BD Biosciences) and GPIbα
(fluorescein-labeled anti-CD42b, Life Technologies) was
analyzed with an acoustic focusing flow cytometer (Attune,
Life Technologies). PLTs were incubated with the labeled
antibodies for 10 minutes in buffer 10 mmol/L
HBS/1 mmol/L MgSO
4
at RT. Samples were diluted 20-fold
immediately before readout. Phosphatidylserine measure-
ment was in buffer supplemented with 2 mmol/L CaCl
2
.
Integrin α
IIb
β
3
activation was measured with PAC1 for
resting PLTs or after activation using cross-linked collagen-
related peptides (CRP-XL; 0.25 μg/mL, University of Cam-
bridge). Threshold gates were set including 0.5% of 10,000
events incubated with corresponding isotype antibody con-
trols. For phosphatidylserine controls, a sample containing
labeled annexin V was prepared without CaCl
2
. The per-
centage of positive events or median fluorescent intensities
were determined for 10,000 events staining positive for
CD61 (allophycocyanin-labeled anti-CD61, Life Technolo-
gies). The number of microparticles was determined using
calibration beads (Biocytex) and the accompanying gating
strategy provided by the manufacturer. Microparticles were
defined as GPIbαpositive events with sizes smaller than
0.9 μm. The result is expressed as the fraction of microparti-
cles per 10,000 all-size events positive for GPIbα.
Western blotting to determine GPIbαectodomain
shedding
Samples (1 mL) were taken from the PLT bag immediately
before freezing at −80C and immediately after thawing.
These experiments were performed on GHE PLTs only for
logistic reasons. Metalloproteinase activity was quenched
immediately with 50 mmol/L EDTA. Centrifugation was per-
formed at 4500 ×gfor 1 minute at RT. The PLT pellet was
resuspended in reducing sample buffer (60 mmol/L Tris,
10% [vol/vol] glycerol, 2% [wt/vol] SDS, 0.01% [wt/vol]
TRANSFUSION 3
CRYOPRESERVED PLATELETS IN SHEAR FLOW
bromophenol blue, and 40 mmol/L dithiothreitol). The
supernatant was 0.2 μmfiltered first and then prepared for
electrophoresis by addition of reducing sample buffer. Equal
volumes of sample were loaded onto 4% to 15% polyacryl-
amide Tris-Glycine TGX gels (Bio-Rad). The GPIbαprotein
or its ectodomain were detected with the mouse monoclo-
nal anti-GPIbα(Clone 8H211) antibody (Antibodies Online)
and a secondary peroxidase conjugated polyclonal antibody
(Cell Signaling Technologies). Antibodies were prepared in
Tris-buffered saline (pH 7.4, 25 mmol/L Tris with
150 mmol/L NaCl, and 2 mmol/ KCl) with 5% (wt/vol)
skimmed milk. Membranes were developed in a imaging
system (ChemiDoc MP, Bio-Rad). Densitometry was per-
formed using (ImageLab v4.0.1, Bio-Rad).
PLT aggregation
Platelet aggregation and agglutination were examined with
light transmission at 37C (Chrono-Log, Helena Laborato-
ries). Ristocetin-induced PLT agglutination was with
1.5 mg/mL ristocetin (American Biochemical and Pharma-
ceuticals). Aggregation was with a combination of 10 μmol/L
protease activated receptor-1 activating peptide (PAR1AP,
Sigma-Aldrich), 20 μmol/L 2-(methylthio)adenosine
5-diphosphate trisodium salt hydrate (MeSADP, Santa Cruz
Biotechnology) and 5 μmol/L epinephrine (Sigma-Aldrich).
Aggregation cuvettes contained 250 ×10
9
PLTs/L in a final
volume of 500 μL, diluted in their respective supernatant
(obtained by centrifugation at 4500 ×gfor 10 min). The sig-
nal was calibrated using the respective PLT-free superna-
tants. Maximal amplitude (%) is reported.
Thrombin generation assay
Generation of thrombin in vitro was analyzed with the
thrombin generation assay kit from Technoclone GmbH
according to the manufacturer’s instructions with minor
modifications. Samples were prepared in 96-well micro-
plates. PCs were diluted to 10 ×10
9
,50×10
9
,or
250 ×10
9
/L in a fixed 40% (vol/vol) volume of heterologous
human pooled plasma and 5% (vol/vol) of saline (0.9%
[wt/vol] NaCl in water) and supplemented with 1 pmol/L
TF, 4.0 μmol/L corn trypsin inhibitor, 0.5 μmol/L fluoro-
genic substrate (Z-G-G-RAMC) and 7.5 mmol/L CaCl
2
(final
concentrations). Samples were immediately analyzed in a
microplate reader (Infinite F200PRO, Tecan Group Ltd.)
with filter settings for excitation at 360 nm and emission at
460 nm. The fluorescent signal was recorded as a function
of time for a total of 120 minutes at 37C. The raw signal
was converted to thrombin concentrations based on a cali-
bration kit and a script in Excel (Microsoft) provided by the
manufacturer.
24
Rotational thromboelastometry
Rotational thromboelastometry (ROTEM) was performed on
reconstituted and citrate anticoagulated blood using a
ROTEM delta analyzer (Tem Innovations) according to the
manufacturer’s recommendations. ROTEM provides
dynamic information on the speed of coagulation initiation,
kinetics of clot growth, clot strength, and breakdown of the
clot. Contact activation is with ellagic acid and phospho-
lipids. Activation of the TF pathway with lipidated TF was
performed as described previously.
25
The following vari-
ables were analyzed: clotting time (sec) to examine speed of
fibrin formation, clot formation time (sec) to examine clot
formation kinetics, and maximum clot firmness (mm) to
examine firmness of the clot. Time between in vitro recon-
stitution of blood and analysis was 120 minutes at
maximum.
Statistical analysis
Results are reported as mean with standard deviation (SD).
Comparison between sites was by one-way analysis of vari-
ance (ANOVA) with Tukey’s multiple comparisons correc-
tion. Comparison of paired GHE and GHR-CTR
concentrates was with a two-tailed paired t-test. Compari-
son of thrombin generation output data by varying PLT con-
centration and production site was with two-way ANOVA
with Tukey’s multiple comparison’s correction. Results from
statistical analysis are depicted on top of the panels. All sta-
tistical analyses were performed using computer software
(Prism, Version 6, GraphPad Software Inc.).
RESULTS
PLT recovery and storage time
Table 1 shows the mean PLT content of the products in the
study before and after cryopreservation. Mean recovery was
more than 90% for all manufacturing sites. It should be
noted that the PLT count after thawing was determined
using the GHE hematology analyzer while for the PLT count
before freezing cell counters from the respective
manufacturing sites were used. Storage duration was lon-
gest for SYD and shortest for GHE cryoplatelets.
PLT adhesion and fibrin deposition under
perfusion flow
In reconstituted blood, the adhesion rate of cryoplatelets
measured by surface coverage was decreased at least three-
fold compared to control (p < 0.05; green signal in Video S1
[available as supporting information in the online version of
this paper] and Fig. 1A). SYD PLTs were more affected
(p < 0.05) than GHE or DEN. Similar observations were
made in the presence of restored calcium levels both in
conditions with TF (green signal in Video S2 [available as
supporting information in the online version of this paper]
and Fig. 1B) and without TF (green signal in Video S3
[available as supporting information in the online version
of this paper] and Fig. 1C). Cryoplatelets also bound to
control surfaces without collagen, leading to localized
4 TRANSFUSION
SIX ET AL
fibrin deposition (Fig. 1D). This was not observed with
control PLTs.
The rate of fibrin deposition following adhesion of
cryoplatelets was variable, but not different from control
PLTs in the presence of TF (violet signal in Video S2 and
Fig. 2A). In the absence of TF, fibrin deposition rates were
slightly decreased for cryoplatelets (violet signal in
Video S3, Fig. 2B). The clotting time was not different
when TF was present (Fig. 2C). In the absence of TF,
however, clotting time decreased for cryoplatelets (GHE,
546 163 sec) compared to control PLTs (GHE-CTR,
631 120 seconds; Fig. 2D). This indicates that cryoplate-
lets enhance contact activation of coagulation under flow
compared to control. There were no significant differ-
ences between SYD, DEN, and GHE PLTs for fibrin
deposition under hydrodynamic flow, despite the differ-
ences in PLT binding (Fig. 1).
GPIbαectodomain shedding in a subset of
cryoplatelets
Almost all events in the cryopreserved PCs were positive for
GPIbα(Fig. 3A), similar to control PLTs. Although control
PLTs had uniformly bright GPIbαexpression, the cryoplate-
lets also contained a dim GPIbαsubpopulation, comprising
45% 11% (DEN), 63% 9% (GHE), and 94% 6% (SYD)
of all GPIbαpositive events (Figs. 3B and 3C). The dim and
bright subpopulations were distinguishable but not always
entirely resolved in different cryoplatelet preparations
(Fig. S1, available as supporting information in the online
TABLE 1. PLT recovery and storage duration*
Center
PLT count (×10
9
)
Recovery (%) Storage (days)Before cryo After cryo
DEN 374.6 40.7 354.9 71.4 94.1 10.8 558 138
SYD 524.2 17.3 504.9 67.5 98.6 7.7 758 12
GHE 345.8 29.1 316.7 17.2 92.1 8.6 21 12
*Data are given as mean SD (n = 6).
Fig. 1. Platelet adhesion under shear flow. Perfusion of blood reconstituted with cryopreserved and control PLTs was at 1000/sec. (A) The PLT
surface coverage as a function of perfusion time (% SC/Time) in channels coated with only collagen (Video S1). This condition was with hirudin-
anticoagulated blood without TF. (B, C) The rate of PLT adhesion in citrated blood with restored Ca
2+
levels. These channels were coated with
(B) collagen and TF (Video S2) or with (C) collagen only (Video S3). (D) PLT (green) and fibrin (violet) accumulation in a part of the channel not
coated with collagen. Images were acquired at endpoint (magnification, 100×). Comparison was between DEN, SYD, and GHE (open bars). A
separate paired comparison between GHE and control noncryoplatelets was also performed (GHE CTR, closed bars). Data are shown as mean SD
(n = 6). Statistical analysis results are shown as *p < 0.05, **p < 0.01, or not significant (ns). [Color figure can be viewed at wileyonlinelibrary.com]
TRANSFUSION 5
CRYOPRESERVED PLATELETS IN SHEAR FLOW
version of this paper). Ristocetin induced PLT agglutination
was significantly decreased in cryoplatelets (Fig. 3D) and
SYD PLTs were more affected (P < 0.001) than GHE or DEN.
The ristocetin response was associated with the fraction of
GPIbα-bright PLTs (Fig. 3C).
Western blotting of PLT lysate and of the PLT-free
supernatant indicated significant GPIbαectodomain shed-
ding after cryopreservation (Fig. 4). The signal of full-length
GPIbαin PLT cell lysates significantly decreased (Fig. 4A).
The signal of GPIbαectodomain in PLT-free supernatant
correspondingly increased three- to sevenfold in cryoplate-
lets compared to before cryopreservation (Figs. 4A and 4B).
This significant variation in the level of ectodomain shed-
ding confirms the variation of GPIbαexpression found in
flow cytometry (Fig. S1).
Platelet aggregation was significantly reduced for cryo-
platelets (Fig. 5A) compared to control, despite a strong ago-
nist blend of PAR1AP, MeSADP, and epinephrine. Integrin
α
IIb
β
3
activation by CRP-XL was also significantly reduced
compared to control (Fig. 5B). Again, SYD PLTs were more
affected (p < 0.01) than GHE or DEN. The aggregation and
integrin activation response furthermore followed the fraction
of GPIbα-bright PLTs (Fig. 3C). In addition, the bright GPIbα
subpopulation had normal integrin activation, while the
GPIbα-dim subpopulation did not (representative data in
Fig. 5C). Similarly, the bright subpopulation had low levels of
annexin V binding, while the dim one had high levels of
annexin V binding (representative data in Fig. 5D). This
resulted overall in higher number of annexin V–positive
events (Fig. S2A). In line with this, the percentage of micro-
particles was increased after cryopreservation (Fig. S2B). The
highest number was found in SYD cryoplatelet preparations.
The refractive properties of the GPIbα-dim subpopulation
were altered. In forward mode, less light (median intensity,
1.9 ×10
5
0.2 ×10
5
) was scattered compared to the GPIbα-
bright subpopulation either in GHE cryoplatelets (median
intensity, 3.3 ×10
5
0.5 ×10
5
)orGHE-CTRcontrolPLTs
(median intensity, 3.3 ×10
5
0.4 ×10
5
;Fig.S3).Collectively,
these data suggest that the GPIbα-dim subpopulation is indic-
ative of the partial damage caused by cryopreservation.
Function of cryoplatelets in reconstituted whole
blood under static conditions
Coagulation under static conditions in vitro was examined with
thromboelastometry. Reconstituted whole blood containing
Fig. 2. Coagulation under shear flow. Perfusion of blood reconstituted with cryopreserved and control PLTs was at 1000/sec with real-time
restoration of Ca
2+
levels. Fibrin accumulation was followed over time. Accumulation rate (/sec) of fibrin in channels coated with (A) collagen
and TF or with (B) only collagen. The clotting time (sec) indicates the moment of coagulation onset in channels coated with (C) collagen and
TF or with (D) only collagen. Comparison was between DEN, SYD, and GHE (all open symbols). A separate paired comparison between GHE
and control noncryoplatelets was also performed (GHE CTR, closed symbols). Data are shown as mean SD (n = 6). Statistical analysis results
are shown as *p < 0.05, **p < 0.01, or not significant (ns).
6 TRANSFUSION
SIX ET AL
GHE cryoplatelets showed decreased clotting times com-
pared to GHE-CTR indicating that initiation of coagulation
was faster irrespective of TF (47 6 sec vs. 58 5 sec;
Fig. 6A) or contact activation (158 30 sec vs. 203 11 sec;
Fig. 6B). In line with this, the lag time between initiating
thrombin generation and initial thrombin formation was
short for all cryoplatelet preparations (Fig. 6C). Peak throm-
bin concentrations were significantly higher (Fig. 6D). Both
lag time and peak thrombin were dependent on the dose of
PLTs as well, irrespective of cryopreservation. Data between
SYD, DEN, and GHE were not different for these parame-
ters, in line with the coagulation experiments under shear
flow. Despite shortening lag times with cryoplatelets, how-
ever, the subsequent clot formation rate and clot firmness
were reduced compared to control (Fig. 7). This was inde-
pendent of TF.
DISCUSSION
Freezing cells damages their integrity, often irreversibly
decreasing the cell’s ability to function optimally.
26
Excessive release of microparticles and decreased responses
to agonists in vitro are known indicators of PLT
cryopreservation-induced damage.
4,18
Despite these
changes, no serious adverse events were documented in a
recent Phase I dose-escalation trial in hematology-oncology
patients.
27
Older studies even suggested clinical efficacy of
cryoplatelets, although count increments were consistently
lower in comparison to liquid-preserved PLTs.
28,29
Cryopla-
telets are promising as they can be shipped over great dis-
tances, but this will require significant standardization of
the production process. In addition, a deep understanding
of cryoplatelet biochemical quality and its markers is
required. This study has examined cryoplatelets produced
in different countries and shipped to one investigation site,
with a view to understanding cryoplatelet variability and
function under conditions of shear flow.
A key finding in this study was the presentation of a
GPIbα-dim subpopulation representing 45% to 94% in cryo-
platelets, but less than 5% in controls. This observation was
reported first by Barnard et al.
17
in a seminal paper on
receptor density of cryoplatelets. Cryoplatelet preparations
with an abundant GPIbα-dim subpopulation displayed
Fig. 3. GPIbαexpression and PLT activation. (A) The percentage of GPIbα-positive events determined in flow cytometry. (B) Representative
histogram of GPIbαsignals in flow cytometry in paired control GHE CTR (red) and cryopreserved GHE (black) PCs denoting dim and bright
GPIbαsubpopulations. (C) Within the GPIbα-positive population, events were gated for dim or bright GPIbαexpression based on the
respective histograms as per B. The population of dim GPIbα(hatched bars) is overlaid with the population of bright GPIbα(open bars)
expressing events. (D) Ristocetin-induced PLT agglutination was measured and data are shown as maximal amplitude (%). Comparison was
between DEN, SYD, and GHE (all open symbols). A separate paired comparison between GHE and control noncryoplatelets was also
performed (GHE CTR, closed symbols). Data are shown as mean SD (n = 6). Statistical analysis results are shown as **p < 0.01,
***p < 0.001, ****p < 0.0001, or not significant (ns). [Color figure can be viewed at wileyonlinelibrary.com]
TRANSFUSION 7
CRYOPRESERVED PLATELETS IN SHEAR FLOW
reduced or absent aggregation, poor integrin activation, and
reduced PLT adhesion under shear. In addition, the GPIbα-
dim subpopulation itself was unresponsive to agonists and
had increased annexin V binding and altered refractive
properties.
Platelets can actively decrease the GPIbαreceptor
number on the plasma membrane by internalization and by
ectodomain shedding. For instance, PLT activation by
PAR1AP decreases GPIbαlevels to below 60% of resting
control PLTs. This does not affect their subsequent adhesion
to collagen in flow chambers because internalized GPIbα
receptors are quickly restored upon contact with the immo-
bilized collagen surface.
30
These observations imply that
GPIbαinternalization is not sufficient to sustainably
decrease PLT adhesion to collagen. The GPIbαdecrease
during cryopreservation therefore is not governed by recep-
tor internalization but by ectodomain shedding,
15
in a sub-
set of PLTs. Contrary to GPIbαinternalization by PLT
activation, ectodomain shedding is irreversible. Moreover,
the GPIbα-dim PLTs have GPIbαlevels well below 50% of
normal, which is a critical value for accomplishing adhesion
Fig. 4. Ectodomain shedding of GPIbα. (A) A representative
Western blot of the GPIbαreceptor in the PLT cell lysate (full
length) or the PLT supernatant (ectodomain) is shown. All samples
were loaded on the same gel, but not adjacent to each other as
indicated by the panel borders. Molecular weight markers are in
kilodalton (kDa). Paired samples before and after cryopreservation
are shown. (B) Densitometry of the GPIbαectodomain signal was
performed for samples before and after cryopreservation. The signal
after cryopreservation is expressed relative to that before. Data from
individual biologic repeats (n = 9) are shown. Mean is shown as a
horizontal line and error bars represent the SD.
Fig. 5. Platelet activation and GPIbαexpression levels. (A) PLT aggregation in response to a mix of PAR1AP, MeSADP, and epinephrine.
Maximal amplitude (%) is shown. (B) The percentage of PLTs expressing activated integrin α
IIb
β
3
in response to stimulation with CRP-XL
measured by binding of PAC1. (C) A representative dot plot of PAC1 binding versus GPIbαexpression. The plot is divided by a quadrant
indicating subpopulations expressing high or low signals of PAC1, GPIbα, or both. (D) A representative dot plot of annexin V binding versus
GPIbαexpression. The plot is divided as in C. Comparison was between DEN, SYD, and GHE (all open symbols). A separate paired
comparison between GHE and control noncryoplatelets was also performed (GHE CTR, closed symbols). Data are shown as mean SD
(n = 6). Statistical analysis results are shown as *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, or not significant (ns).
8 TRANSFUSION
SIX ET AL
to collagen under flow.
30–32
Consequently, these PLTs con-
tribute most if not all to the decrease in adhesion and by
extrapolation to the decrease in overall cryoplatelet function.
Furthermore, that fraction was remarkably variable
between production sites as well as between PLT prepara-
tions within a given site (Fig. S1). This suggests that cryopla-
telet quality may be improved if the number of
dysfunctional PLTs can be consistently controlled at low
levels. It is currently not clear, however, what makes certain
PLTs more susceptible to cryopreservation-induced dam-
age. Maybe differences in primary PLT concentration and
age, cryoprocess steps, (preliminary) exposure to tempera-
ture cycles
33
, and/or (bio)chemicals contribute. Whether
shedding inhibitors could prevent cryopreservation-induced
damage depends on what comes first. If GPIbαshedding is
secondary to a damaging event, then shedding inhibitors
will not help but in the case that shedding sparks down-
stream lesions, inhibitors may be able to mitigate damage.
For age-related storage lesion of liquid PLTs, specific inhibi-
tion of GPIbαshedding rescued the rapid clearance of aged
PLTs to some extent.
34
Future research can always exploit
the GPIbα-dim subpopulation as a biomarker for the extent
of damage. The impact of directed modifications in the
cryoprocess can be easily monitored using this particular
marker in flow cytometry.
Cryopreservation-induced damage was particularly
obvious in assays that specifically study PLTs, like integrin
activation, microparticle release, and aggregation. This is in
line with previous publications
14–18
, which were mostly
under static conditions (or extremely low shear) that do not
always mimic the biophysical environment of blood vessel
injury. Our data now confirm that in shear flow, cryoplatelet
adhesion rates to immobilized collagen were also signifi-
cantly decreased compared to control. This was both with
and without anticoagulation, indicating that cryoplatelets
are less adhesive irrespective of simultaneous thrombin and
fibrin formation. Unlike controls, cryoplatelets also bound
to uncoated parts of the channel during perfusion. This sug-
gests that cryoplatelets are less selective than control PLTs
for binding to nonbiologic substrates. This requires further
investigation because cryoplatelets may be lost by nonspeci-
fic adhesion to transfusion bags, lines, or needles.
Fig. 6. Coagulation initiation in static conditions. Blood was reconstituted with cryopreserved (open symbols) or control PLTs (closed symbols)
and studied in the presence of Ca
2+
by ROTEM using (A) TF or (B) ellagic acid to initiate coagulation. The clotting time (sec) is the lag time
from adding the activating agent to the blood until the elastogram tracing reached 2 mm. (C, D) Thrombin generation was measured in the
presence of TF and increasing PLT concentrations of 10 ×10
9
/L (filled bars), 50 ×10
9
/L (hatched bars), and 250 ×10
9
/L (open bars). (C) The
lag time (min) measures the time until threshold thrombin concentrations are reached. Statistics are indicated for the conditions with
250 ×10
9
PLTs/L. (D) Peak thrombin (nmol/L) measures the peak amount of thrombin formed. Statistical significance is indicated for the
conditions with 250 ×10
9
PLTs/L. Cryoplatelets from DEN, SYD, and GHE were used and the latter was paired to control noncryoplatelets
GHE CTR (dashed line). Data are shown as mean SD (n = 6). Statistical analysis results are shown as *p < 0.05, ***p < 0.001,
****p < 0.0001, or not significant (ns).
TRANSFUSION 9
CRYOPRESERVED PLATELETS IN SHEAR FLOW
Cryopreservation-induced damage was less prominent
in assays that focus on coagulation. In ROTEM, cryoplatelets
had more rapid clotting times compared to controls even
though clot formation times and maximum clot firmness
were decreased. Further, in terms of thrombin generation,
cryoplatelets demonstrated faster clotting times than con-
trols and higher peak thrombin, all in a dose-dependent
manner. Johnson and colleagues
19,35
had previously shown
that the presence of cryoplatelet supernatant shortens time
to coagulation presumably by the activity of microparticles,
but these experiments were conducted in static conditions,
allowing coagulation factors to diffuse freely in bulk. In this
case, the cascade can quickly assemble tenase and pro-
thrombinase complexes on the abundant negatively charged
phospholipid surface provided by damaged cryoplatelets
and microparticles. Under shear flow, factor availability
depends on convective hydrodynamic forces rather than dif-
fusion.
36
Our data confirm, however, that also under shear
flow, coagulation is not affected by cryopreservation of
PLTs. In contrast, contact-induced coagulation initiation in
the absence of TF was even shorter than control. This is
remarkable because with or without TF, significantly fewer
PLTs were adhering to the surface during perfusion. Clinical
trials are required to investigate whether the procoagulant
role of cryoplatelets is sufficient to stop bleeding in patients.
The effect of cryopreservation for SYD cryoplatelets was
greater than that for the other manufacturing sites particu-
larly in PLT assays. The GPIbα-dim subpopulation was
more abundant than in the other sites, the aggregation
response was lower, integrin activation was less, and PLT
adhesion under shear was the lowest with SYD cryoplate-
lets. Despite this, the differences between production sites
were less obvious in PLT-based coagulation assays. It is not
clear why SYD PLTs were particularly more damaged than
those from the other sites. SYD and GHE PLTs were resus-
pended the same way with similar plasma so differences in
reconstitution procedure are not likely. Irregularities in
deep-freeze conditions could be one of the reasons, but dif-
ferences in primary product preparation are also possible.
The SYD PLTs were collected by apheresis in plasma, while
GHE PLTs were prepared by pooling of buffy coats in plasma
with SSP+. The SYD PLTs were frozen for 2 years, which is
the maximum reported shelf-life.
37
Although preliminary data
suggest that cryoplatelets stored for 4 years are equivalent to
those stored for 2 years,
38
the effect of extended storage on
the GPIbαsubpopulation and PLT function have not been
reported. The SYD PLTs also had the highest initial PLT con-
tent in comparison to the other sites, which effectively
reduces the amount of DMSO per PLT unit. These variables
might have contributed to the observations.
Fig. 7. Coagulation propagation in static conditions. Blood was reconstituted with cryopreserved (open symbols) or control PLTs (closed
symbols) and studied in the presence of Ca
2+
by ROTEM using (A, C) TF or (B, D) ellagic acid to initiate coagulation. (A, B) The clot formation
time (sec) was measured between the clotting time (Figs. 5A and 5B) and the moment a clot firmness of 20 mm has been reached. (C, D)
Maximum clot firmness (MCF) reflects the absolute strength of the clot. Comparison was between DEN, SYD, and GHE (all open symbols). A
separate paired comparison between GHE and control noncryoplatelets was also performed (GHE CTR, closed symbols). Data are shown as
mean SD (n = 6). Statistical analysis results are shown as *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, or not significant (ns).
10 TRANSFUSION
SIX ET AL
We conclude that cryoplatelets are less adhesive in
shear flow conditions, in proportion to a subset of damaged
GPIbα-dim PLTs, which decrease the effective PLT dose.
The size of the affected subpopulation may be indicative of
the extent of damage. Different production sites may have
different cryoplatelet qualities especially in PLT function
assays. Finally, even though fewer cryoplatelets bind colla-
gen under flow, these nonetheless enhance contact activa-
tion, provide sufficient support for TF activation and normal
coagulation under shear flow.
ACKNOWLEDGMENTS
The authors thank Michael Luypaert and Esha Wauters for their
technical support in performing ROTEM analyses.
AUTHOR CONTRIBUTIONS
KS, VC, LJ, LJD, DCM, and HBF designed research, provided
cryopreserved platelet products, and performed quality con-
trol experiments; KS and WD performed experiments; KD
was responsible for the thromboelastometry assays; HBF
and KS wrote the manuscript; VC, LJ, LJD, DCM, and HBF
supervised the study; and all authors critically reviewed and
amended the manuscript.
CONFLICT OF INTEREST
The authors have disclosed no conflicts of interest.
REFERENCES
1. Desborough MJ, Smethurst PA, Estcourt LJ, et al. Alternatives
to allogeneic platelet transfusion. Br J Haematol 2016;175:
381-92.
2. Liumbruno G, Bennardello F, Lattanzio A, et al. Recommenda-
tions for the transfusion of plasma and platelets. Blood Trans-
fus 2009;7:132-50.
3. Murphy S, Gardner FH. Effect of storage temperature on main-
tenance of platelet viability—deleterious effect of refrigerated
storage. N Engl J Med 1969;280:1094-8.
4. Johnson L, Tan S, Wood B, et al. Refrigeration and cryopreser-
vation of platelets differentially affect platelet metabolism and
function: a comparison with conventional platelet storage con-
ditions. Transfusion 2016;56:1807-18.
5. Valeri CR, Feingold H, Marchionni LD. A simple method for
freezing human platelets using 6% dimethylsulfoxide and stor-
age at −80C. Blood 1974;43:131-6.
6. Cap AP. Platelet storage: a license to chill! Transfusion 2016;56:
13-6.
7. Reddoch KM, Pidcoke HF, Montgomery RK, et al. Hemostatic
function of apheresis platelets stored at 4C and 22C. Shock
2014;41(Suppl 1):54-61.
8. Hess JR, Lelkens CC, Holcomb JB, et al. Advances in military,
field, and austere transfusion medicine in the last decade.
Transfus Apher Sci 2013;49:380-6.
9. Balduini CL, Mazzucco M, Sinigaglia F, et al. Cryopreservation
of human platelets using dimethyl sulfoxide and glycerol-glu-
cose: effects on "in vitro" platelet function. Haematologica
1993;78:101-4.
10. Nie Y, de Pablo JJ, Palecek SP. Platelet cryopreservation using a
trehalose and phosphate formulation. Biotechnol Bioeng 2005;
92:79-90.
11. Taylor MA. Cryopreservation of platelets: an in-vitro compari-
son of four methods. J Clin Pathol 1981;34:71-5.
12. Valeri CR, Ragno G, Khuri S. Freezing human platelets with
6 percent dimethyl sulfoxide with removal of the supernatant
solution before freezing and storage at −80 degrees C without
postthaw processing. Transfusion 2005;45:1890-8.
13. Spector JI, Yarmala JA, Marchionni LD, et al. Viability and
function of platelets frozen at 2 to 3 C per minute with 4 or
5 per cent DMSO and stored at −80 C for 8 months. Transfu-
sion 1977;17:8-15.
14. Dumont LJ, Cancelas JA, Dumont DF, et al. A randomized con-
trolled trial evaluating recovery and survival of 6% dimethyl
sulfoxide-frozen autologous platelets in healthy volunteers.
Transfusion 2013;53:128-37.
15. Johnson LN, Winter KM, Reid S, et al. Cryopreservation of
buffy-coat-derived platelet concentrates in dimethyl sulfoxide
and platelet additive solution. Cryobiology 2011;62:100-6.
16. Valeri CR, Macgregor H, Ragno G. Correlation between in vitro
aggregation and thromboxane A2 production in fresh, liquid-
preserved, and cryopreserved human platelets: effect of ago-
nists, pH, and plasma and saline resuspension. Transfusion
2005;45:596-603.
17. Barnard MR, MacGregor H, Ragno G, et al. Fresh, liquid-pre-
served, and cryopreserved platelets: adhesive surface receptors
and membrane procoagulant activity. Transfusion 1999;39:880-8.
18. Waters L, Padula MP, Marks DC, et al. Cryopreserved platelets
demonstrate reduced activation responses and impaired sig-
naling after agonist stimulation. Transfusion 2017;57:2845-57.
19. Johnson L, Coorey CP, Marks DC. The hemostatic activity of
cryopreserved platelets is mediated by phosphatidylserine-
expressing platelets and platelet microparticles. Transfusion
2014;54:1917-26.
20. Cid J, Escolar G, Galan A, et al. In vitro evaluation of the hemo-
static effectiveness of cryopreserved platelets. Transfusion
2016;56:580-6.
21. Johnson L, Reade MC, Hyland RA, et al. In vitro comparison of
cryopreserved and liquid platelets: potential clinical implica-
tions. Transfusion 2015;55:838-47.
22. Van Aelst B, Feys HB, Devloo R, et al. Microfluidic flow cham-
bers using reconstituted blood to model hemostasis and plate-
let transfusion in vitro. J Vis Exp 2016;109:e53823.
23. Six KR, Devloo R, Van Aelst B, et al. A microfluidic flow cham-
ber model for platelet transfusion and hemostasis measures
platelet deposition and fibrin formation in real-time. J Vis Exp
2017;120:e55351.
TRANSFUSION 11
CRYOPRESERVED PLATELETS IN SHEAR FLOW
24. Hemker HC, Kremers R. Data management in thrombin gener-
ation. Thromb Res 2013;131:3-11.
25. Chitlur M, Sorensen B, Rivard GE, et al. Standardization of
thromboelastography: a report from the TEG-ROTEM working
group. Haemophilia 2011;17:532-7.
26. Iwatani M, Ikegami K, Kremenska Y, et al. Dimethyl sulfoxide
has an impact on epigenetic profile in mouse embryoid body.
Stem Cells 2006;24:2549-56.
27. Slichter SJ, Dumont LJ, Cancelas JA, et al. Treatment of bleed-
ing in severely thrombocytopenic patients with transfusion of
dimethyl sulfoxide (DMSO) cryopreserved platelets (CPP) is
safe - report of a phase 1 dose escalation safety trial. Blood
2016;128:1030.
28. Daly PA, Schiffer CA, Aisner J, et al. Successful transfusion of
platelets cryopreserved for more than 3 years. Blood 1979;54:
1023-7.
29. Towell BL, Levine SP, Knight WA 3rd, et al. A comparison of
frozen and fresh platelet concentrates in the support of throm-
bocytopenic patients. Transfusion 1986;26:525-30.
30. van Zanten GH, Heijnen HF, Wu Y, et al. A fifty percent reduc-
tion of platelet surface glycoprotein Ib does not affect platelet
adhesion under flow conditions. Blood 1998;91:2353-9.
31. Cauwenberghs N, Ajzenberg N, Vauterin S, et al. Characteriza-
tion of murine anti-glycoprotein Ib monoclonal antibodies that
differentiate between shear-induced and ristocetin/botrocetin-
induced glycoprotein Ib-von Willebrand factor interaction.
Haemostasis 2000;30:139-48.
32. Kageyama S, Matsushita J, Yamamoto H. Effect of a humanized
monoclonal antibody to von Willebrand factor in a canine
model of coronary arterial thrombosis. Eur J Pharmacol 2002;
443:143-9.
33. Hoffmeister KM, Felbinger TW, Falet H, et al. The clearance
mechanism of chilled blood platelets. Cell 2003;112:87-97.
34. Chen W, Liang X, Syed AK, et al. Inhibiting GPIbalpha shed-
ding preserves post-transfusion recovery and hemostatic func-
tion of platelets after prolonged storage. Arterioscler Thromb
Vasc Biol 2016;36:1821-8.
35. Raynel S, Padula MP, Marks DC, et al. Cryopreservation alters
the membrane and cytoskeletal protein profile of platelet
microparticles. Transfusion 2015;55:2422-32.
36. Zhu S, Lu Y, Sinno T, et al. Dynamics of thrombin generation
and flux from clots during whole human blood flow over col-
lagen/tissue factor surfaces. J Biol Chem 2016;291:23027-35.
37. Valeri CR, Srey R, Lane JP, et al. Effect of WBC reduction and
storage temperature on PLTs frozen with 6 percent DMSO for
as long as 3 years. Transfusion 2003;43:1162-7.
38. Noorman F, Strelitski R, Badloe J. Frozen platelets can be
stored for 4 years at −80C without affecting in vitro recovery,
morphology, receptor expression, or coagulation profile. Trans-
fusion 2014;54:74A.
SUPPORTING INFORMATION
Additional Supporting Information may be found in the
online version of this article.
Fig. S1 Platelet subpopulations by GPIbαexpression.
GPIbαexpression was determined by flow cytometry on
10,000 platelets. Isotype negative control (blue) was com-
pared to (A) non cryopreserved control platelets of GHE-
CTR (red) and to all GHE cryopreserved platelets (black).
Three different GHE products are shown in panels B to
D. The dotted line represents the arbitrary gate determined
for dim (left) and bright (right) GPIbαsubpopulations. The
histograms were normalized to display relative count.
Fig. S2 Annexin V binding and microparticles. (A) The
percentage of GPIbαpositive events binding annexin V.
(B) Microparticles were defined as events smaller than
0.9 μm. The relative number of microparticles to the total
number of GPIbαpositive events is depicted. Comparison
was between cryopreserved platelets from Denver, CO
(DEN), Sydney (Australia, SYD) and Ghent (Belgium, GHE)
(open bars). A separate paired comparison was between
cryopreserved platelets (GHE) and control non-
cryopreserved platelets (GHE CTR, filled bars). Data are
shown as mean SD (n = 6). Statistical analysis results are
shown as ** P < 0.01, *** P < 0.001, **** P < 0.0001 or not
significant (ns).
Fig. S3 Scatter properties of cryopreserved platelets. The
median intensity (MI) of the forward scattered light (FSC) in
flow cytometry for subpopulations of cryoplatelets (open
bars) or control platelets (closed bars) based on the GPIbα
signal as shown in Fig. S1. All samples were from the Bel-
gian (GHE) site. Data are shown as mean SD (n = 6). Sta-
tistical analysis results are shown as ** P < 0.01 or not
significant (ns).
12 TRANSFUSION
SIX ET AL