646
Shear-Induced Platelet Activation and Platelet
Microparticle Formation at Blood Flow
Conditions as in Arteries With a Severe Stenosis
Pal A. Holme, Una 0rvim, Maria J.A.G. Hamers, Nils O. Solum, Frank R. Brosstad,
R. Marius Barstad, Kjell S. Sakariassen
Abstract In the present study, we investigated whether high
arterial shear stresses at various exposure times or a sudden
increase in shear stress introduced by a stenosis affect platelet
activation and platelet microparticle formation in native hu-
man blood. We used a parallel-plate perfusion chamber device
through which nonanticoagulated human blood was drawn (10
mL/min) by a pump directly from an antecubital vein through
the flow channel of a perfusion chamber at wall shear rates of
420, 2600, and 10 500 s~'. In another set of experiments, an
eccentric stenosis was introduced into the flow channel. Wall
shear rates of 2600 or 10 500 s"1 at the stenosis apex were
maintained at the same flow rate. The wall shear rate upstream
and downstream of these stenoses was 420 s~'. A shear rate of
420 s"1 is within the range of those encountered in healthy
small coronary arteries, whereas those of 2600 and 10 500 s"1
are representative for vessels with various degrees of stenotic
lesions. The blood was exposed to these shear rates for periods
Downloaded from http://ahajournals.org by on August 19, 2020
varying from 0.075 to 3.045 seconds. Platelet activation was
assessed as activated glycoprotein (GP) Ilb/IIIa by FITC-
labeled monoclonal antibody (MAb) PAC-1 and aminophos-
pholipid translocation by FITC-labeled annexin V. Micropar-
ticle formation was quantified by FITC-labeled MAb Y2/51
directed against GP Ilia. Significant platelet activation and
formation of microparticles were observed at 10 500 s"1 only
(P<.008). This shear-induced platelet activation and micropar-
A therosclerotic plaques that develop into stenotic
lesions have profound impact on the local
blood flow behavior. 1 3 Wall shear rate and
wall shear stress may increase by one or two orders of
magnitude within a fraction of a second at such lesions.
Indeed, wall shear rates exceeding 40 000 s _1 and wall
shear stresses >300 dynes/cm 2 have been reported 1 2 for
mechanically constricted epicardial arteries in dogs. 1 In
comparison, the physiological ranges of these physical
flow parameters vary from 20 to 2000 s"1 and from 1.4 to
60 dynes/cm 2 , 4 respectively. Sustained high shear condi-
tions are not maintained at focal lesions but may prevail
in vessels with diffuse lesions. It is apparent that the
shape and degree of luminal occlusion introduced by
such stenoses are determining the blood flow character-
istics at the lesion.
Received June 6, 1996; revision accepted July 15, 1996.
From the Research Institute for Internal Medicine, Rikshospi-
talet, University of Oslo (P.A.H., N.O.S., F.R.B.), and Nycomed
Pharma AS, Oslo (U.0., M.J.A.G.H., R.M.B., K.S.S.), Norway.
Correspondence to Pal Andre Holme, MD, Research Institute
for Internal Medicine, Rikshospitalet, Pilestredet 32, 0027 Oslo,
Norway. E-mail: p.a.holme@klinmed.uio.no
© 1997 American Heart Association, Inc.
tide formation were enhanced by introduction of a thrombus-
promoting surface consisting of type III human collagen fibrils.
Introduction of the most severe stenosis at 10 500 s"1 further
increased platelet activation (P<.017). The collagen-induced
thrombus formation increased the platelet thrombus volume at
10 500 s"1 from 16.5 to 33.8 p,m3/ju.m2 (P<.003) on the stenosis
apex when the most severe stenosis was used. A correlation
(P<.0001) between platelet thrombus volume and platelet
microparticle formation was observed in the presence of the
eccentric stenoses. Apparently, high shear stress (315 dynes/
cm2 at 10 500 s~'), as encountered in severe atherosclerotic
arteries, activated platelets and triggered platelet microparticle
formation. In contrast, no significant platelet activation or
formation of platelet microparticles was observed at physiolog-
ical shear (420 s"1) or at the shear condition simulating shear in
arteries with a less severe stenosis (2600 s"'). The data imply
that platelets are activated and form microparticles in native
blood at very high shear stresses. These events are potentiated
by prolonged exposure to the high shear or by a sudden change
of increasing shear due to the stenosis. The latter situation
apparently enhances platelet thrombus formation at the steno-
sis. (Arterioscler Thromb Vase Biol. 1997;17:646-653.)
Key Words • platelet activation • microparticles •
shear rate • stenosis • flow cytometry
Rapid and dramatic changes in blood flow behavior
may activate passing blood platelets. Whether shear-
induced platelet activation occurs in vivo remains to be
established, and if so, its consequences for thrombus
formation should be elucidated. This is of great impor-
tance, since arterial thrombus formation is generally
triggered by plaque rupture at stenoses, where high
shear or disturbed blood flow may prevail. Nevertheless,
in vitro experiments have shown platelet activation
elicited by shear stresses and wall shear rates as low as 50
dynes/cm 25 and 100 s"1,6 respectively. Furthermore, the
exposure time to the shear is important, since at least 7
ms is required to trigger platelet activation by a shear
stress of 170 dynes/cm 2 in a viscometer. 7
Platelet-derived microparticles are formed from the
surface membrane by an exocytotic budding process 19 on
platelet activation by agonists such as thrombin or
collagen. 8 -" Microparticles have an average diameter of
0.1 /u,m.10 They express negatively charged phospholip-
ids 1 2 1 3 and factor Va 14 and Xa activity15 and thus possess
procoagulant surface properties. 81116 " 18 Their potential
significance in hemostasis and thrombosis remains un-
known, although they have been detected in blood from

patients with activated coagulation and fibrinolysis,20
with autoimmune thrombocytopenias, 21 ' 22 and after car-
diopulmonary bypass. 23 Recently, it was reported that
microparticle formation correlates well with exposure of
the platelet procoagulant surface 11 and that activation of
glycoprotein (GP) Ilb/IIIa may be of some importance
for this process with certain agonists. 11 ' 24
The aim of the present work was to study platelet
activation and microparticle formation at different shear
conditions as encountered in healthy and atherosclerotic
arteries. This was performed by an ex vivo flow model
using nonanticoagulated human blood and a panel of
parallel-plate perfusion chambers with or without the
presence of a cosine-shaped eccentric stenosis in the
flow channel. 25 " 27 Platelet activation and formation of
microparticles were studied by flow cytometry. The wall
shear rates used ranged from a physiological value of 420
s"1 to values above the physiological range of 2600 and
up to 10 500 s~'. The latter two shear conditions are
representative of those in atherosclerotic vessels, thus
simulating flow conditions in vessels with various de-
grees of stenotic lesions. Platelet activation was assessed
as activated GP Ilb/IIIa complexes by binding of a
monoclonal antibody (MAb, PAC-1, and by expression
of negatively charged phospholipids on the surface as
measured by annexin V. Platelet-derived microparticles
were quantified by an MAb against GP Ilia. These
parameters were assessed in the presence or in the
absence of a thrombus-promoting surface (placed at the
apex of the stenoses) consisting of human type III
collagen fibrils.25'27
Methods
Downloaded from http://ahajournals.org by on August 19, 2020
Reagents
FITC was purchased from Molecular Probes, Inc. Recombi-
nant annexin V was a generous gift from Bender Wien
(Vienna, Austria).
Antibodies
The MAb PAC-1 (IgM), which recognizes an epitope on
activated GP Ilb/IIIa, was purchased from The Cell Center,
University of Pennsylvania.28
The FITC-conjugated monoclonal antibody Y2/51 directed
against GP Ilia, and X927, a FITC-conjugated negative control
of the same subtype (IgGl), were purchased from DAKO A/S.
FITC Labeling of Antibodies and Proteins
FITC conjugation of annexin V and PAC-1 was performed
according to Goding29 and as previously described in detail.11
Parallel-Plate Perfusion Chambers With or Without
an Eccentric Stenosis
Two types of parallel-plate perfusion chambers were used
(Table 1). These included four perfusion chambers with unob-
structed blood flow25-26-30 and two perfusion chambers with an
eccentric stenosis in the flow channel.2732
The chambers with unobstructed flow have been well char-
acterized regarding blood flow behavior and thrombus forma-
tion.2526 The blood flow is laminar in these chambers. The
chambers selected for this investigation had wall shear rates of
420, 2600, and 10 500 s _l at a flow rate of 10 mL/min and a
Reynolds number of 20. Geometrical dimensions of the flow
channels as well as corresponding shear rates and stresses are
summarized in Table 1.
A miniature chamber with a rectangular flow channel
18.0 mm long was constructed for the present study. The width
and height of the flow channel were chosen such that the wall
Holme et al Shear-Induced Platelet Activation 647
shear rate was 10 500 s"1 at a flow rate of 10 mL/min. The
length of the flow channel was identical to the axial length of
the stenosis apex of the perfusion chamber, with an eccentric
stenosis having an apex wall shear rate of 10 500 s"1.27-31'32
Thus, the exposure time to this wall shear rate was identical in
the two chambers. The exposure time to the respective shears
is given in Table 1.
The two chambers with a stenosis in the flow channel were
previously characterized.27'3132 Briefly, a cosine-shaped eccen-
tric stenosis is introduced into the rectangular flow channel of
the original parallel-plate perfusion chamber2526 with a wall
shear rate of 420 s"1. The cosine-shaped stenosis step has an
axial length of 0.5 mm, whereas the axial length of the stenosis
is 18.0 mm. The protrusion of the stenosis into the flow channel
determines the wall shear rate at the apex, which in the present
investigation was 2600 and 10 500 s~" at a flow rate of 10
mL/min. The corresponding stenotic occlusions of the flow
channels were 60% and 80%, respectively. The exposure time
to these shear conditions corresponds to 0.075 and 0.151
second, respectively (Table 1). The wall shear rate upstream
and downstream of the stenoses is 420 s"1. However, shear
overshoots of 2200 and 4000 s_1 at the rear end of the upstream
half-cosine-shaped step of the 60% and 80% occluding steno-
ses, respectively, were established by numerical analysis with a
finite-element program developed by Fluid Dynamics Interna-
tional Inc (FIDAP).27'31'32 The perfusion chambers with these
two stenoses are intended to simulate blood flow in coronary
arteries with "advanced" single eccentric stenosis with a long
axial dimension. Blood perfusion experiments were performed
at 2600 and 10 500 s -1 with exposure to purified human type
III collagen fibrils spray-coated on Thermanox coverslips
(Miles Laboratories).25 Uncoated Thermanox coverslips were
exposed to blood in all perfusion chambers and thus at all
shear conditions. The Thermanox surface interacts poorly with
flowing blood, whereas the collagen fibrils trigger rapid and
pronounced thrombus formation.25
Preparation of the Collagen Surface
Type III collagen was purified from human placenta by
pepsin digestion and selective salt precipitation.33 Collagen
fibrils were prepared by dialysis against 20 mmol/L Na 2 HP0 4 ,
pH 7.5, for 48 hours at 4°C and coated on Thermanox plastic
coverslips as previously described.2527 The collagen fibrils do
not activate coagulation,25 and the coating gives a maximal
thrombogenic stimulus at densities of 210 /ug/cm2.
Blood Donors and Blood Sampling
Each of the blood donors gave informed consent to donate
55 mL blood per perfusion experiment. According to their
statements, none of the individuals had taken aspirin or other
drugs for at least 14 days before the perfusion experiments.
Immediately before each perfusion experiment, 4 mL blood
was collected into EDTA-containing tubes for determination
of hemoglobin, hematocrit, and white cell and platelet counts
(Auto Counter AC 920, Swelab Instruments). Individual he-
moglobin, hematocrit, platelet, and leukocyte values were
within the normal range for all subjects studied.
Human Ex Vivo Perfusions, Fixation,
and Embedding
Ex vivo perfusion experiments were performed at 37°C with
parallel-plate perfusion chambers34 with or without a stenosis
in the flow channel and in the presence of a collagen-coated or
an uncoated Thermanox coverslip. Venipuncture was per-
formed with a No. 19 butterfly infusion set (Abbott Laborato-
ries). Nonanticoagulated blood from 6 to 17 healthy volunteers
was drawn through each perfusion chamber. The blood flow
rate was maintained at 10 mL/min for 5 minutes by an occlusive
roller pump (Minipuls 3, Gilson) placed distal to the chamber.
Blood perfusions were immediately followed by a 20-second
perfusion (10 mL/min) with a buffer containing (in mmol/L)
 648 Arteriosclerosis, Thrombosis, and Vascular Biology
Characterization of Different Chamber Types With Respect to Wall Shear Rates,
Wall Shear Stresses, Dimensions of Blood Flow Channels, and Shear Exposure
Times*
Wall Shear Wall Shear
Rate, Stress,
Chamber s" 1 dyne/cm 2 t
Plane 420 13
Plane 2600 80
Plane 10 500 315
Plane 10 500 315
Stenosis 2600 80
Stenosis 10 500 315
"The time it takes the blood to pass the area with the actual shear rate is defined as exposure time at a
constant blood flow rate of 10 mL/min and a Reynold's (Re) number of 20.
tAt a constant blood flow rate of 10 mL/min and an assumed blood viscosity of 3.0 mPa.
JFIow-channel dimensions are (heightxwidthxlength).
§Percent occlusion of blood-flow channel is given as percent occlusion of the cross-sectional area.
NaCl 130, KC1 2, NaHC0 3 12, CaCl2 2.5, and MgCl2 0.9 at pH
7.4, 37°C and by a 40-second perfusion with fixation solution
(2.5% glutaraldehyde in 0.1 mol/L cacodylate buffer, pH 7.4,
22°C). Subsequently, the coverslip was removed from the
chamber and kept in freshly prepared fixation solution for 1
hour. Specimens were stored in 7% sucrose/0.1 mol/L cacody-
late at 4°C until they were embedded in epoxy resin.
Morphometry
Evaluation of thrombotic deposits was performed on epoxy
resin-embedded semithin sections (1 ysn) prepared perpen-
dicular to the direction of the blood flow and 1 mm down-
stream from the upstream end of the coverslip.34-35 Platelet
thrombus volume (ixm2/ixm2) was derived from the sectional
thrombus area measured by computer-assisted morphometry
(Kontron Vidas, Eching).35
Downloaded from http://ahajournals.org by on August 19, 2020
Platelet Isolation
For flow cytometry analysis, 1 mL of blood anticoagulated
with 0.1 vol 0.129 mol/L trisodium citrate was sampled distal to
the perfusion chamber after 4.5 minutes of perfusion. Platelet-
rich plasma was then immediately prepared by centrifugation
for 5 minutes at 160g at 20°C. Subsequently, platelets were fixed
by addition of paraformaldehyde in PBS to 0.33%. Fixation
could not be used for studies of annexin V binding, because this
was strongly impaired by the fixation procedure.
Preparation of Platelet Samples for Flow Cytometry
Platelets (1X106) were added to polystyrene tubes containing
filtered PBS (pH 7.4) at a final volume of 100 /uL after addition of
fluorescent probes. The various FITC-labeled probes were added
in final concentrations of 5 /u,g/mL FITC-Y2/51, 6 jxg/mL FITC-
X-927, 39 ng/mL FITC-PAC-1, or 25 /xg/mL FITC-annexin V.
The mixtures were incubated in the dark at 4°C for 30 minutes.
Subsequently, they were diluted with 1 mL filtered PBS.
Flow Cytometry
Platelets labeled with the FITC-conjugated probes were
analyzed in a FACScan flow cytometer (Becton Dickinson)
equipped with a 15-mW air-cooled 488-nm argon laser as
previously described." The light scatter and fluorescence chan-
nels were set at logarithmic gain. The platelets were analyzed
by the FACScan in two ways. To study the amounts of the
probes (annexin V and PAC-1) bound to platelets after perfu-
sion, platelets were gated on the basis of the forward- and
side-scatter properties. The percentages of platelets activated
during perfusion were calculated from the fluorescence inten-
sity of the platelets before and after perfusion through the
chamber. A fluorescence threshold was set at the upper limit of
the prechamber sample, and cells with a fluorescence intensity
exceeding this threshold level after perfusion through the
chamber were considered positive.
Vol 17, No 4 April 1997
Flow-Channel Occlusion of Exposure
Dimensions, Flow Channel, Time,
mmt %§ Seconds*
0.70X5.0X145 3.045
0.28X5.0X145 1.218
0.14X5.0X145 0.609
0.14X5.0X18 0.075
0.28X5.0X18 60 0.151
0.14X5.0X18 80 0.075
To study platelet microparticle formation and to resolve
platelet-derived microparticles from background light scatter,
acquisition was gated so as to include only positive events for
antibody bound to GP Ilia (Y2/51). Consequently, a fluores-
cence threshold was set to analyze only platelets and
microparticles.
Microparticles and platelets were separated analytically on
the basis of their characteristics in forward and side scatter. To
quantify and discriminate between platelets and microparticles,
the lower limit of the platelet gate was set at the left border of
the forward scatter profile of unperfused platelets. The number
of microparticles present was expressed as the number of
particles below this limit in percent of the total number of
fluorescent particles counted (ie, platelets plus microparticles).
Altogether, 10 000 positive events were analyzed each time,
and the Cellquest program (Becton Dickinson) was used for
data processing on an Apple computer.
Statistical Analysis
The Mann-Whitney test was used for statistical analysis. The
data shown represent mean±SEM. Correlations between pa-
rameters were studied by linear regression analysis.
Results
Microparticle Formation in the Absence of Stenosis
When the platelet-specific monoclonal antibody Y2/51
against GP Ilia was used to detect platelets and micro-
particles by flow cytometry, the difference in micropar-
ticle concentration before and at 4.5 minutes of perfu-
sion was quantified. Fewer than 1.5% microparticles
were detected in blood samples collected from the arm
vein. No significant increase in platelet-derived micro-
particles was detected during perfusion at shear rates of
420 and 2600 s~\ not even during collagen-induced
thrombus formation (Fig 1A). However, a significant
increase in the number of microparticles was measured
at 10 500 s"1 (/ , <.008), both with collagen-coated cover-
slips ( 2 . 8 ± 1 . 0 % ) and with noncoated coverslips
(1.6±0.3%) (Fig 1A). The difference in formation of
microparticles between collagen-coated and noncoated
coverslips was not significant.
Microparticle Formation in the Presence
of Stenosis
Introduction of an eccentric stenosis in the flow
channel did not increase the microparticle formation at
2600 s"1 (60% stenosis occlusion). However, at 10 500
s _1 (80% stenosis occlusion), the microparticle forma-
tion was increased (P<.006) even in the absence of
 Holme et al Shear-Induced Platelet Activation 649

Without stenosis With an eccentric stenosis

B
4
| Prechamber Pre chamber
icl

3.5 3.5-
t, TO Thermanox '€ 3
Thermanox
a .* es
a ^ Collagen a Collagen
o 2.5 2.5-
o
l-H | § Mini chamber
2 2
c hi
a.
a 1.5
8 1.5
u 1 H
•c
a. 0.5 0.5
o
u 0
u • * T o
-0.5 -0.5-^
Pre chamber 420 2600 10500 Pre chamber 2600 St 10500 St
Shear rate (sec" ) Shear rate (sec" )
FIG 1. Microparticle formation after perfusion at (A) wall shear rates of 420 s" 1 (n=6), 2600 s"' (n=6), or 10 500 s~1 (n=6) without a
stenosis and at (B) wall shear rates of 2600 s" 1 (60% stenosis) (n=6) or 10 500 s" 1 (80% stenosis) (n=15 with collagen, n=9 with
Thermanox) at an eccentric stenosis with collagen-coated or noncoated Thermanox coverslips. Platelet microparticle formation is given
as increase in number of microparticles per 100 counted events (ie, microparticles and platelets) measured in blood samples collected
from arm vein before onset of perfusion (prechamber) and distal to perfusion chamber after 4.5 minutes of perfusion. Further processing
of the samples is described in "Methods." Miniature chamber: A perfusion chamber with unobstructed flow (no stenosis). Flow channel
wall shear rate is 10 500 s" 1 (n=6), and time of shear exposure is identical to 10 500 s _1 exposure time at 80% stenosis apex. Values
are mean±SEM.

collagen-induced thrombus formation (Fig IB). How-
ever, more microparticles were formed during the col-
lagen-induced thrombus (P<.018) (Fig IB). Micropar-
ticle formation at 10 500 s"1 was larger in the perfusion
chamber without a stenosis. However, the exposure time
Downloaded from http://ahajournals.org by on August 19, 2020
increased activation (P<.03) was measured (Fig 2A).
About 9% to 16% of the platelets were activated
whether perfused over noncoated or collagen-coated
coverslips (Fig 2A). Although no significant differences
in activation were observed between collagen-coated

to the high shear, or the time it takes for platelets to pass
the length of the flow channel, or the stenosis apex at the
shear rate of 10 500 s~', was different in the two cham-
bers (Table 1). The exposure time was calculated to
0.609 second for the chamber without the stenosis and
0.075 second for the perfusion chamber with the eccen-
tric stenosis. Taking this difference into consideration, it
was estimated that «<15 times more microparticles were
produced per unit time in the presence of the stenosis
compared with the chamber without the stenosis, indi-
cating that the stenosis itself and/or the stenosis geom-
etry was much more efficient in producing microparticles
than the shear rate itself. To confirm this hypothesis, a
novel perfusion chamber was constructed. This chamber
had no stenosis, but the blood exposure time to 10 500
s"1 was kept similar to that in the chamber with the
stenosis by shortening the length of the flow channel
(Table 1). The shear conditions in this parallel-plate
perfusion chamber (noncoated Thermanox coverslip)
elicited only a slight increase in microparticle formation
during perfusion, not statistically different from the low
levels measured at 420 and 2600 s"1 (Fig 1A).
Activation of GP Ilb/IIIa in the Absence
of Stenosis
Activation of GP Ilb/IIIa was studied by flow cytom-
etry using the MAb PAC-1 directed against activated GP
Ilb/IIIa and the same blood samples as used for quan-
tification of microparticles (see above). GP Ilb/IIIa
activation paralleled the formation of microparticles. No
significant increase in GP Ilb/IIIa activation was ob-
served at 420 or 2600 s"1. However, at 10 500 s~\
and noncoated coverslips, there was a tendency toward
more activation in the presence of collagen-induced
thrombus formation.
Activation of GP Ilb/IIIa in the Presence
of Stenosis
The same set of experiments as reported above was
repeated in the presence of eccentric stenoses of 60%
(2600 s_1) or 80% (10 500 s_1) occlusion. No significant
GP Ilb/IIIa activation was detected at 2600 s"1 (Fig 2B).
However, pronounced GP Ilb/IIIa activation (P<.003)
was measured at 10 500 s"1, thus at a flow condition
simulating flow in a severely stenosed artery. Although
more activation was measured during collagen-induced
thrombus formation (10.7±2.5%) than in the absence of
collagen (6.7±2.9%), the difference was not statistically
significant. Considering the difference in shear exposure
time as described above, approximately five times more
activated platelets were measured in the presence of the
stenosis than in its absence. Experiments with the min-
iature perfusion chamber, in which the 10 500 s~' shear
exposure time was identical to the exposure time at the
stenosis, showed a tendency of increased PAC-1 binding;
however, this was not significantly different from the
values measured in blood samples collected from the
arm vein (P<.06) (Fig 2A).
Annexin V Binding in the Absence of Stenosis
Translocation of aminophospholipids was measured
as binding of FITC-labeled annexin V to the surface of
the activated platelets by flow cytometry.36-'7 Significant
annexin V binding was not detected during perfusions at
 650 Arteriosclerosis, Thrombosis, and Vascular Biology Vol 17, No 4 April 1997

Without stenosis g With an eccentric stenosis

20
20H
• Pre chamber Pre chamber
£
a
1 Thermanox
§3 Collagen

r Thermanox

Collagen

r
& g Mini chamber

* 5-
•9
• 6
Pre chamber 420 2600 10500 Pre chamber 2600 St 10500 St
1^
Shear rate (sec"A)
o Shear rate (sec" ) 1

FIG 2. Activation of GP lib/Ilia measured as FITC-PAC-1 binding after perfusion at (A) wall shear rates of 420 s"' (n=6), 2600 s" 1 (n=6),
or 10 500 s""1 (n=6) without a stenosis and at (B) wall shear rates of 2600 s" 1 (60% stenosis) (n=6) or 10 500 s" 1 (80% stenosis) (n=17
with collagen, n=9 with Thermanox) at the eccentric stenoses. Thrombus formation on either collagen-coated or noncoated Thermanox
coverslips. Number of PAC-1-positive platelets in percent of all platelets analyzed in each perfusion. Blood samples collected during
perfusion were withdrawn after 4.5 minutes of perfusion. Miniature chamber as in Fig 1 (n=6). Values are mean±SEM.

420 or 2600 s~'. Collagen-induced thrombus formation
did not affect the binding of annexin V to the perfused
platelets. However, at 10 500 s"1, the annexin V binding
increased (P<.01) (Fig 3A), being «3.8% and 2.8% in
the presence and absence of collagen-induced thrombus
formation, respectively.
Annexin V Binding in the Presence of Stenosis
Downloaded from http://ahajournals.org by on August 19, 2020
exposure time, it was found that «10-fold more activa-
tion per unit time occurred with the stenosis than
without the stenosis. Again, the results imply that the
stenosis itself and/or its geometry was much more im-
portant for platelet activation than the shear rate itself.
To confirm this hypothesis, perfusion studies with the
miniature chamber at 10 500 s"1 were performed. A
tendency toward increased annexin V labeling during

No significant annexin V binding was measured at
2600 s~' either in the presence or in the absence of
collagen-induced thrombus formation. However, signif-
icant annexin V binding was measured at 10 500 s_1
(P<.006). The binding was 3.8±0.8% in the presence of
collagen-induced thrombus formation and 3.7±0.8% in
the absence of collagen (Fig 3B). In evaluation of the
annexin V binding with regard to the 10 500 s -1 shear
perfusion was detected; however, this was not signifi-
cantly different from the values measured in blood
samples collected from the arm vein (P<.07) (Fig 3A).
Platelet Thrombus Volume in the Absence
of Stenosis
Platelet thrombus volume on collagen and Thermanox
was determined by computer-assisted morphometry.35

Without stenosis With an eccentric stenosis

Pre chamber 420 2600 10500 Pre chamber 2600 St 10500 St

1
Shear rate (sec"1) Shear rate (sec -1)
FIG 3. Platelet activation detected as surface expression of aminophospholipids that bind FITC-annexin V after perfusion at (A) wall
shear rates of 420 s" 1 (n=6), 2600 s" 1 (n=6), or 10 500 s" 1 (n=6) without a stenosis and (B) wall shear rates of 2600 s" 1 (60% stenosis)
(n=6) or 10 500 s~1 (80% stenosis) (n=12 with collagen, n=8 with Thermanox) at eccentric stenoses. Thrombus formation on either
collagen-coated or noncoated Thermanox coverslips was studied. Annexin V binding is given as number of annexin V-positive platelets
in percent of all platelets analyzed in each perfusion. Blood samples collected during perfusion were withdrawn after 4.5 minutes of
perfusion. Miniature chamber as in Fig 1 (n=5). Values are mean±SEM.
 Holme et al Shear-Induced Platelet Activation 651

Without stenosis B With an eccentric stenosis

-40 .
N N
| Thermanox | Thermanox
35 35- T
r> J H Collagen | § Collagen
E30 o 30-
a.
w 25 25-
E
3, 20
1 E 20-
T
i•
i

o 15-1
1
i —

M 15

•
9 10:
• • • • i — i....

•fio o
u 5-
o
JS 5 1 0-
H 2600 St 10500 St
0 420 2600 10500
Shear rate (sec'1) Shear rate (sec )
FIG 4. Platelet thrombus volume (/xm3//xm2) on coverslip after 5-minute perfusions at (A) wall shear rates of 420 s~1 (n=6), 2600 s" 1
(n=6), or 10 500 s" 1 (n=6) without a stenosis and at (B) wall shear rates of 2600 s~1 (60% stenosis) (n=6) or 10 500 s" 1 (80% stenosis)
(n=17 with collagen, n=11 with Thermanox) at eccentric stenoses, with collagen-coated or noncoated Thermanox coverslips. Values
are mean±SEM.

Virtually no thrombotic material was detected on non-
coated Thermanox coverslips at 420, 2600, or 10 500 s_1
(Fig 4A). In contrast, pronounced thrombus formation
on collagen was observed at 2600 and 10 500 s"1. The
platelet thrombus volumes averaged 16.5±2.5 and
20.9±4.0 /j.m3/ju,m2, respectively.
Platelet Thrombus Volume in the Presence
of Stenosis
Downloaded from http://ahajournals.org by on August 19, 2020
microparticles by complex blood flow behavior, as may be
encountered in atherosclerotic vessels. It is apparent that
high shear and a sudden increase in shear leads to activa-
tion of the platelets, with the exposure time to high shear
also being important. These factors may act simultaneously
and synergistically on the platelet in diseased arteries.
It appeared that flow conditions that activate platelets
are nonphysiological. Evidence of platelet activation at
physiological shear was not detected, not even at a wall

Platelet thrombus volume on collagen-coated cover- shear rate of 2600 s"1, which is slightly above the
slips at the 80% stenosis (10 500 s"1) was significantly physiological range. Also, inclusion of a thrombus-pro-
higher (33.8±2.4 fimV^m2) (P<.0002) than at the 60% moting surface in the form of collagen did not trigger
stenosis (2600 s"1) (11.7±2.5 ju,m3/ju,m2) (Fig 4B). Also, detectable platelet activation at these shear conditions.
with noncoated coverslips, a statistically significant in- However, a nonphysiological wall shear rate of 10 500
crease in thrombus volume was detected at 10 500 s"1 s"1 triggered significant platelet activation and micropar-
(12.6±4.6 ju,m3//xm2) compared with the volume at 2600 ticle formation, and more so in the presence of collagen-
s"1 (2.0+1.2 /xm3//xm2) (P<.04). induced thrombus formation.
It is interesting to note that the thrombus volume in It was not possible to differentiate between chemical
the collagen-induced thrombus formation at 10 500 s -1 activation by substances released from platelets in collag-
was twofold increased on the stenosis apex relative to
the volume measured in the absence of stenosis 75-
(P<.003). __ U
Linear regression analysis was used to see whether E
5- 60-
statistical correlations existed between the platelet acti-
%
vation markers measured and the platelet thrombus a 45-
°a a ^ ^ a
volume on collagen-coated and noncoated coverslips. E
o D
The thrombus volume was correlated to the formation of > 30-
D
n
U) rP 3
microparticles after perfusion through the two perfusion D
U D a.
chambers with the eccentric stenosis (r=.66, ),<.0001) o 15-
a. D

(Fig 5). No such correlation was observed in the absence £

of the stenosis, nor was there any significant correlation
between PAC-1 or annexin V labeling of the platelets
and the thrombus volume. Significant correlation was
found both for annexin V and PAC-1 labeling compared
with the formation of microparticles in the different
perfusion chambers (r=.38, P<.003 and r=.42, P<.002,
respectively).
Discussion
The present investigation provides the first evidence that
human platelets in native blood are activated and form
1-
0- B-# •B-r°-
Microparticle increase
FIG 5. Regression analysis of platelet thrombus volume and
platelet microparticle formation at eccentric stenoses. Platelet
thrombus volume is given as /xm3//Lim2. Platelet microparticle
formation is given as increase in number of microparticles per
100 counted events of platelets and microparticles during per-
fusion. Data include experiments both with uncoated Thermanox
and collagen-coated Thermanox coverslips. Wall shear rates
were 2600 s" 1 (60% stenosis) and 10 500 s" 1 (80% stenosis)
(r=.66, P-c.0001, n=31).
 652 Arteriosclerosis, Thrombosis, and Vascular Biology Vol 17, No 4 April 1997
en-attached thrombi, by local thrombin formation at the
stenosis, or by shear-induced activation introduced by the
stenosis and the gradually increasing shear imposed by the
growing thrombi at the stenosis apex. The possibility of
high and low shear in close proximity to the thrombotic
deposits, and thus the potential sudden change from low to
high shear in and around the growing thrombi, may
support activation and microparticle formation. Also, flow
channels through the platelet-rich thrombi at 10500 s"1
may represent regions of very high shear27 that may
activate passing platelets. On the other hand, activation of
platelets and coagulation by the butterfly infusion set is less
likely, since the coagulation activation markers F 1+2,
thrombin-antithrombin, and fibrinopeptide A and the plate-
let activation marker j3-thromboglobulin are still within the
normal range within 5 minutes of perfusion.38 However,
prolonging the perfusion time more than 5 minutes results
in significant chemical activation of both coagulation and
platelets and should thus be avoided.
The significant platelet activation observed at 10 500 s"1
in the presence or in the absence of the eccentric stenosis
coincides with the insensitivity of aspirin to thrombus
formation at this shear condition.39 Shear-induced platelet
activation is apparently insensitive to aspirin, which previ-
ously was shown in in vitro studies.4042 This is in accor-
dance with the observation that aspirin significantly re-
duces thrombus formation at 2600 s"1, both in the presence
and absence of the stenosis,3943 thus at flow situations in
which platelet activation and microparticle formation were
not detected in the present study. Clinical angiographic
studies of reocclusion after coronary thrombolysis is in
accordance with our observations as well, since aspirin
Downloaded from http://ahajournals.org by on August 19, 2020
administration is reported not to be effective in preventing
reocclusion of lesions with >90% stenosis, whereas protec-
tion was observed at less severe stenosis.44
Whereas no platelet activation or microparticle forma-
tion was observed at 420 or 2600 s~' or with the 60%
stenosis at 2600 s"', a significant increase in microparticles
and PAC-1 and annexin V labeling was detected at 10 500
s"1 both in the presence and in the absence of the eccentric
stenosis. There was no apparent difference in platelet
activation whether the stenosis was present or not. How-
ever, because the exposure time to the 10 500 s"1 shear rate
was 8.2 times longer (Table 1) in the absence of the
stenosis, the ratio between the degree of activation and the
exposure time was calculated. These calculations showed
the presence of 5 to 15 times more activated platelets per
unit time when calculated for the binding of PAC-1 or
annexin V or as formation of microparticles at the eccen-
tric stenosis. The experiments with the miniature perfusion
chamber, in which the time of shear exposure was identical
to the exposure time at the stenosis, confirmed these
calculations, since only a slight platelet activation or micro-
particle formation was detected. These results confirm that
shear-dependent platelet activation is coupled to the expo-
sure time in native blood, as was shown in anticoagulated
blood by Konstantopoulos et al.42 The data further imply
that the stenosis geometry is important for the degree of
platelet activation. The sudden change in shear stress that
occurs over a very short distance at the stenosis (in this
model, over an axial length of 0.5 mm), including the
transient shear overshoot of 4000 s_1 at the stenosis inlet,
appears to be more important for the activation and
microparticle formation than the wall shear rate itself.
It is plausible to suggest that the twofold-increased
thrombus volume on the stenosis apex at 10 500 s"' was
due to the sudden increase in shear introduced by the
stenosis and thus due to the resulting platelet activation.
This observation suggests that the geometry of the stenosis,
and not only the shear rate itself, must be considered in the
evaluation of the resulting platelet activation and thrombus
formation. It is well known that geometric alterations in the
flow path, such as those produced in the present study at
the respective stenoses, can produce wide variations in
local shear levels, which are further complexed as throm-
bus formation proceeds, especially since initial deposits
seem to favor the apex area.45
A relatively low percentage of the platelets exiting from
the flow chamber was activated, implying that most plate-
lets were still unactivated after the passage through the
perfusion chamber. However, the number of platelets
actually activated may be higher, because some of these
platelets are incorporated into the growing thrombi. Obvi-
ously, these activated platelets would not be included in the
flow cytometric analysis. In contrast, platelets transiently in
contact with the growing thrombi without being incorpo-
rated into the thrombus but being chemically and/or shear-
activated would be detected by this method.
A significant correlation was found between the throm-
bus volume on Thermanox or collagen at the eccentric
stenosis and the number of microparticles found (r=.66,
F<.0001). No such correlation was observed between
PAC-1 or annexin V labeling and the thrombus volume.
This may indicate that most of the microparticles were
released from activated platelets that were incorporated
into growing thrombi, whereas the binding of PAC-1 or
annexin V to platelets represented platelets that generally
had been activated in the streaming blood.
We conclude that high shear stress in the parallel-plate
perfusion chambers simulating those encountered in se-
verely stenosed arteries activates platelets and produces
platelet-derived microparticles and that the sudden change
in shear stress induced by the stenosis triggers more
platelet activation and formation of microparticles than the
shear itself. It is apparent that much remains to be learned
about the relationship of intravascular platelet activation
and thrombus formation in atherosclerotic vessels. Never-
theless, this study has provided new knowledge about flow
conditions promoting platelet activation, microparticle for-
mation, and concomitant enhanced thrombus formation in
native human blood.
Acknowledgments
This study was supported by The Research Council of
Norway, Nycomed Pharma AS, The Norwegian Council on
Cardiovascular Diseases, the Anders Jahres fund, and the
Professor Paul A. Owrens fund.
References
1. Roux SP, Sakariassen KS, Turitto VT, Baumgartner HR. Effect of
aspirin and epinephrine on experimentally induced thrombogenesis
in dogs: a parallelism between in vivo and ex vivo thrombosis
models. Arterioscler Thromb. 1991;11:1182-1191.
2. Strony J, Beaudoin A, Brands D, Adelman B. Analysis of shear
stress and hemodynamic factors in a model of coronary artery
stenosis and thrombosis. Am J Physiol. 1993;265:H1787-H1796.
3. Sakariassen KS, Barstad RM. Mechanisms of thromboembolism at
arterial plaques. Blood Coag Fibrinol. 1993;4:615-625.
4. Goldsmith HL, Turitto VT. Rheological aspects of thrombosis and
haemostasis: basic principles and applications. Thromb Haemost.
1986;55:415-435.

5. Moritz MW, Reimers RC, Baker RK, Sutera SP, Joist JH. Role of
cytoplasmic and releasable ADP in platelet aggregation induced by
laminar shear stress. J Lab Clin Med. 1983;101:537-544.
6. Rhodes NP, Zuzel M, Williams DF, Derrick MR. Granule
secretion markers on fluid-phase platelets in whole blood perfused
through capillary tubing. J Biomed Mater Res. 1994;28:435-439.
7. Wurzinger LJ, Opitz R, Blasberg P, Schmid-Schonbein H. Platelet
and coagulation parameters following millisecond exposure to
laminar shear stress. Thromb Haemost. 1985;54:381-386.
8. Sims PJ, Faioni EM, Wiedmer T, Shattil SJ. Complement proteins
C5b-9 cause release of membrane vesicles from the platelet surface
that are enriched in the membrane receptor for coagulation factor
Va and express prothrombinase activity. / Biol Chem. 1988;263:
18205-18212.
9. Sims PJ, Wiedmer T, Esmon CT, Weiss HJ, Shattil SJ. Assembly of
the platelet prothrombinase complex is linked to vesiculation of the
platelet plasma membrane: studies in Scott syndrome: an isolated
defect in platelet procoagulant activity. J Biol Chem. 1989;264:
17049-1757.
10. Wiedmer T, Shattil S, Cunningham M, Sims PJ. Role of calcium
and calpain in complement-induced vesiculation of the platelet
plasma membrane and in the exposure of the platelet factor Va
receptor. Biochemistry. 1990;29:623-632.
11. Holme PA, Solum NO, Brosstad F, Egberg N, Lindahl TL. Stim-
ulated Glanzmann's thrombasthenia platelets produce
microvesicles: microvesiculation correlates better to exposure of
procoagulant surface than to activation of GPllb-lIIa. Thromb
Haemost. 1995;74:1533-1540.
12. Comfurius P, Senden JMG, Tilly RHJ, Schroit AJ, Bevers EM,
Zwaal RFA. Loss of membrane phospholipid asymmetry in plate-
lets and red cells may be associated with calcium-induced shedding
of plasma membrane and inhibition of aminophospholipid trans-
locase. Biochim BiophysActa. 1990;1026:153-160.
13. Dachary-Prigent J, Freyssinet JM, Pasquet JM, Carron JC, Nurden
AT. Annexin V as a probe of aminophospholipid exposure and
platelet membrane vesiculation: a flow cytometry study showing a
role for free sulfhydryl groups. Blood. 1993;81:2554-2565.
14. Monkovic DD, Tracy PB. Functional characterization of human
platelet-released factor V and its activation by factor Xa and
thrombin. J Biol Chem. 1990;265:17132-17140.
Downloaded from http://ahajournals.org by on August 19, 2020
15. Holme PA, Brosstad F, Solum NO. Platelet-derived microvesicles
and activated platelets express factor Xa activity. Blood Coag
Fibrinol. 1995;6:302-310.
16. George JN, Pickett EB, Sauceman S, McEver RP, Kunicki TJ,
Kieffer N, Newman PJ. Platelet surface glycoproteins: studies on
resting and activated platelets and platelet membrane micropar-
ticles in normal subjects, and observations in patients during adult
respiratory distress syndrome and cardiac surgery. J Clin Invest.
1986;78:340-348.
17. Sandberg H, Bode AP, Dombrose FA, Lentz BR. Expression of
coagulant activity in human platelets: release of membranous
vesicles providing platelet factor 1 and platelet factor 3. Thromb
Res. 1985;39:69-79.
18. Bode AP, Sandberg H, Dombrose FA, Lentz BR. Association of
factor V activity with membranous vesicles released from human
platelet: requirement for platelet stimulation. Thromb Res. 1985;39:
49-61.
19. Holme PA, Brosstad F, Solum NO. The difference between platelet
and plasma F XIII used to study the mechanism of platelet
microvesicle formation. Thromb Haemost. 1993;70:681-686.
20. Holme PA, Solum NO, Brosstad F, R0ger M, Abdelnoor M. Dem-
onstration of platelet-derived microvesicles in blood from patients
with activated coagulation and fibrinolysis using a filtration
technique and Western blotting. Thromb Haemost. 1994;72:
666-671.
21. Jy W, Horstman L, Arce M, Ann YS. Clinical significance of
platelet microparticles in autoimmune thrombocytopenias. J Lab
Clin Med. 1992;119:334-345.
22. Khan I, Zucker-Franklin D, Karpatkin S. Microthrombocytosis and
platelet fragmentation associated with idiopathic/autoimmune
thrombocytopenic purpura. Br J Haematol. 1975;31:449-460.
23. Abrams CS, Ellison N, Budzynski AZ, Shattil SJ. Direct detection
of activated platelet and platelet-derived microparticles. Blood.
1990;75:128-138.
24. Gemmell CH, Sefton MV, Yeo EL. Platelet-derived microparticle
formation involves glycoprotein Ilb-IIIa: inhibition by RGDS and a
Glanzmann's thrombasthenia defect. J Biol Chem. 1993;268:14586-
14589.
Holme et al Shear-Induced Platelet Activation 653
25. Sakariassen KS, Joss R, Muggli R, Kuhn H, Tschopp TB, Sage H,
Baumgartner HR. Collagen type III induced ex vivo thrombo-
genesis in humans: role of platelets and leukocytes in deposition of
fibrin. Arteriosclerosis. 1990;10:276-284.
26. Sakariassen KS, Aarts PAMM, de Groot PG, Houdijk WPM, Sixma
JJ. A perfusion chamber developed to investigate platelet inter-
action in flowing blood with human vessel wall cells, their extra-
cellular matrix and purified components. J Lab Clin Med. 1983;102:
522-533.
27. Barstad RM, Roald HE, Cui Y, Turitto VT, Sakariassen KS. A
perfusion chamber developed to investigate thrombus formation
and shear profiles in flowing native human blood at the apex of
well-defined stenoses. Arterioscler Thromb. 1994;14:1984-1991.
28. Shattil SJ, Hoxie JA, Cunningham M, Brass LF. Changes in the
platelet membrane glycoprotein Ilb-IIIa complex during platelet
activation. / Biol Chem. 1985;260:11107-11114.
29. Goding JW. Conjugation of antibodies with fluorochromes: modi-
fication to the standard methods. / Immunol Methods. 1976; 13:
216-226.
30. Sakariassen KS, Roald HE, Aznar Salatti J. Ex vivo models for
studying thrombosis: special emphasis on shear dependent blood
coagulation interactions. In: Hwang NHC, Turitto VT, Yen MRT,
eds. Advances in Cardiovascular Engineering. New York, NY:
Plenum Press; 1992:151-174.
31. Cui W, Sakariassen KS, Barstad RM, Turitto VT. Velocity and
shear profiles in a stenotic perfusion system. Tfiromb Haemost.
1993;69:2234. Abstract.
32. Barstad RM, Kierulf P, Sakariassen KS. Collagen-induced
thrombus formation at the apex of eccentric stenoses: a time course
study with non-anticoagulated blood. Thromb Haemost. 1996;75:
685-692.
33. Miller EF, Rhodes RK. Preparation and characterisation of the
different types of collagen. Methods Enzymol. 1982;82:32-64.
34. Sakariassen KS, Muggli R, Baumgartner HR. Measurements of
platelet interactions with components of the vessel wall in flowing
blood. Methods Enzymol. 1989;169:37-70.
35. Sakariassen KS, Kuhn H, Muggli R, Baumgartner HR. Growth and
stability of thrombi in flowing citrated blood: assessment of platelet
surface interactions with computer-assisted morphometry. Thromb
Haemost. 1988;60:392-398.
36. Andree HAM, Reutelingsperger CPM, Hauptmann R, Hemker
HC, Hermens WT, Willems GM. Binding of vascular anticoagulant
(VACa) to planar phospholipid bilayers. J Biol Chem. 1990;265:
4923-4928.
37. Tait JF, Gibson D, Fujikawa K. Phospholipid binding properties of
human placental anticoagulant protein-I, a member of the
lipocortin family. / Biol Chem. 1989;264:7944-7949.
38. Diquelou A, Lemozy S, Dupouy D, Boneu B, Sakariassen K,
Cadroy Y. Effect of blood flow on thrombin generation is
dependent on the nature of the thrombogenetic surface. Blood.
1994;84:2206-2213.
39. Barstad RM, 0rvim U, Hamers MJAG, Tj0nnfjord GE, Brosstad
F, Sakariassen KS. Reduced effect of aspirin on thrombus for-
mation at high shear and disturbed laminar blood flow. Thromb
Haemost. 1996;75:827-832.
40. O'Brien JR, Salmon GP. Shear stress activation of platelet glyco-
protein Ilb/IIIa plus von Willebrand factor causes aggregation:
filter blockage and the long bleeding time in von Willebrand's
disease. Blood. 1987;70:1354-1361.
41. Hardwich RA, Heliums JD, Peterson DM, Moake JL, Olson JD.
The effects of ASA, PGE b PGI2 and theophylline on response of
platelets subjected to shear stress. Blood. 1981;58:678-681.
42. Konstantopoulos K, Wu KK, Udden MM, Banez EI, Shattil SJ,
Heliums JD. Flow cytometric studies of platelet responses to shear
stress in whole blood. Biorheology. 1995;32:73-93.
43. Roald HE, 0rvim U, Bakken IJ, Barstad RM, Kierulf P,
Sakariassen KS. Modulation of thrombotic responses in moderately
stenosed arteries by cigarette smoking and aspirin ingestion. Arte-
rioscler Thromb. 1994;14:617-621.
44. Veen G, Meyer A, Verheugt FW, Werter CJ, de Swart H, Lie KI,
van der Pol JM, Michels J, van Eenige MJ. Culprit lesion mor-
phology and stenosis severity in the prediction of reocclusion after
coronary thrombolysis: angiographic results of the APRICOT
study. J Am Coll Cardiol. 1993;22:1755-1762.
45. Badimon L, Badimon JJ. Mechanisms of arterial thrombosis in
nonparallel streamlines: platelet thrombi grow on the apex of ste-
notic severely injured vessel wall. J Clin Invest. 1989;84:1134-1144.