British Journal of Haematology, 2003, 120, 535–541
HemolinkTM, an o-raffinose cross-linked haemoglobin-based
oxygen carrier, does not affect activation and function
of human platelets in whole blood in vitro
Valery Leytin, 1,3 David Mazer, 2 Meera Mody, 1,3 Bernadette Garvey 1,3 and John Freedman 1,3
Departments of 1Transfusion Medicine and 2Anaesthesia at St. Michael’s Hospital, Toronto and the 3Department of
Laboratory Medicine and Pathobiology at the University of Toronto, Toronto, Ontario, Canada
Received 7 May 2002; accepted for publication 26 July 2002
Summary. Haemoglobin-based oxygen carriers (HBOCs)
are anticipated to be safe and efficient alternatives to RBC
transfusions. Haemoglobin (Hb) raffimer (HemolinkTM;
Hemosol, Toronto, ON, Canada) is polymerized human Hb,
cross-linked with o-raffinose. As administration of cell-free
Hb may affect blood cells and tissues, this study was focused
on evaluating effects of Hb raffimer on human platelets in
whole blood in vitro. Citrated blood from healthy donors was
incubated with Hb raffimer to achieve raffimer concentra-
tions of 2–50 vol percentage (2–50 g/l). Platelet activation,
phosphatidylserine exposure and microparticle generation
were measured by flow cytometry. Aperture closure time on
collagen/ADP- and collagen/epinephrine-coated mem-
branes was determined by a platelet function analyser (PFA-
100
). We found that addition of Hb raffimer to blood
Over the last two decades, considerable progress has been
made in the development of haemoglobin-based oxygen
carriers (HBOCs) that are anticipated to be safe and efficient
alternatives to red blood cell (RBC) transfusions. The
current generation of HBOCs is characterized by enhanced
stability and storage time, prolonged intravascular persist-
ence, and efficient oxygen delivery to tissues (Winslow,
1992; Dietz et al, 1996; Chang, 1997; Winslow et al, 1997;
Adamson & More, 1998; Rudolph et al, 1998; Blajchman,
2000; Mazer, 2000; Carmichael, 2001; Cheng, 2001; Hill,
2001; Stowell et al, 2001).
Haemoglobin (Hb) raffimer (HemolinkTM; Hemosol,
Toronto, ON, Canada) is a polymerized human Hb cross-
linked with oxidized trisaccharide o-raffinose (Adamson &
More, 1998; Mazer, 2000; Carmichael, 2001; Cheng, 2001;
Correspondence: John Freedman, MD, Department of Transfusion
Medicine, St. Michael’s Hospital, 30 Bond Street, Toronto, Ontario
M5B 1W8, Canada. E-mail: freedman@smh.toronto.on.ca
 2003 Blackwell Publishing Ltd
samples up to 50 vol % did not affect human platelets as
measured by various markers of platelet activation (CD42b,
CD41, PAC-1, CD62, CD63), procoagulant activity (annexin
V) and microparticle formation; differences between Hb
raffimer- and lactated Ringer’s-diluted blood were not sig-
nificant. Similarly, no adverse effect of Hb raffimer on clo-
sure time was observed at concentrations up to 50 vol %, in
comparison with Ringer’s solution. These data indicate that
exposure of human blood to high concentrations of Hb
raffimer in vitro did not cause platelet activation nor affect
platelet function.
Keywords: blood substitutes, haemoglobin-based oxygen
carrier HemolinkTM, haemoglobin raffimer, platelet
activation and function, flow cytometry, PFA-100
.
Stowell et al, 2001). The manufacturing process of this
HBOC includes Hb extraction from fully tested washed
outdated human RBCs approved for transfusion, pasteur-
ization for viral and bacterial inactivation, filtration for
removal of potential contaminations, purification by cation
and anion exchange chromatography, and treatment with
o-raffinose (Cheng, 2001). The final product contains ‡ 95%
cross-linked Hb with a proportion of Hb tetramer (64 kDa)
and polymers (128–576 kDa) of 34–43% and 54–65%,
respectively, and has an intravascular half-life in the range
of 14–20 h (Carmichael et al, 2000; Mazer, 2000; Carmi-
chael, 2001; Cheng, 2001).
Several clinical trials of Hb raffimer have been completed
in Canada and the UK, including one phase I safety study in
normal healthy volunteers (Carmichael et al, 2000), six
phase II studies in cardiac and orthopaedic surgery and
renal anaemia, and one pivotal phase III study in coronary
artery bypass grafting (Mazer, 2000; Carmichael, 2001).
Almost 700 patients have been enrolled in this phase III
study and more than 300 have been treated with Hb
535
536 V. Leytin et al
raffimer (Carmichael, 2001). Hb raffimer was infused
intravenously as a 10% (w/v) solution, in doses of 0Æ25–
6Æ0 ml/kg (Carmichael et al, 2000), 750 ml or 250–
1000 ml (Mazer, 2000; Cheng, 2001); a control group
received an equal volume of lactated Ringer’s solution
(Carmichael et al, 2000) or 10% pentastarch (Pentaspan
)
(Mazer, 2000; Cheng, 2001). These studies have suggested
that the use of Hb raffimer may reduce allogeneic donor
RBC transfusions, and may be safe for use in orthopedic and
cardiac surgery and patients with anaemia due to renal
disease (Mazer, 2000; Carmichael, 2001; Cheng, 2001).
In vivo, Hb is localized within RBCs and does not come in
direct contact with tissues and cells. Replacement of RBCs
with HBOCs might, however, lead to unpredicted conse-
quences (Stowell et al, 2001) and the use of Hb raffimer or
other HBOCs raises a possibility of their interaction with
blood cells, including platelets. Enhancement of platelet
deposition and accelerated thrombosis have been shown in
rat and rabbit models following administration of HBOCs
(Olsen et al, 1996; Lee et al, 2000). These in vivo effects of
HBOCs on platelet function can be related to the nitric oxide
sequestration by the heme iron of HBOCs (Motterlini et al,
1996) or to generation of platelet-reactive free radicals
(Moro et al, 1994; Alayash, 1999). The few studies
performed for evaluating effects of HBOCs on platelet
aggregation in vitro have yielded conflicting results (Lalla
et al, 1989; Reiss et al, 1992; Mondoro et al, 1998; Lieber-
thal et al, 1999; Lee et al, 2000; Leytin et al, 2001).
We studied the direct effects of Hb raffimer on human
platelets in vitro in the whole blood environment. Platelet
activation, procoagulant activity and generation of platelet-
derived microparticles were evaluated using flow cytometry.
In addition, we tested the effect of Hb raffimer on platelet
‘haemostatic’ function in vitro by measuring aperture
closure time on collagen/epinephrine- and collagen/ADP-
coated membranes using the platelet function analyser
PFA-100
.
MATERIALS AND METHODS
Flow cytometric analysis of platelet activation, phosphatidyl-
serine externalization and microparticle formation. After dis-
carding the first 2 ml of blood, 4Æ5 ml of venous blood was
collected from seven drug-free healthy volunteers, using a
21 G needle, without tourniquet, into 3Æ8% (0Æ129 mol/l)
buffered sodium citrate vacutainers (Becton Dickinson,
Rutherford, NJ, USA; 1 part anticoagulant to 9 parts blood).
Blood samples were kept at 22C and flow cytometry was
performed within 30 min of blood withdrawal. Hb raffimer
(HemolinkTM) was supplied in lactated Ringer’s solution
containing 10 g Hb/dl. Test 1 ml aliquots of the blood were
prepared to achieve final concentrations of Hb raffimer of 2,
10, 20, 40 and 50 vol percentage (2, 10, 20, 40 and
50 g/l). The samples were incubated with gentle mixing for
15 min at 37C. For each experiment, the following
controls were included: (i) undiluted blood as a baseline
control; (ii) aliquots of blood diluted up to 50% with lactated
Ringer’s solution (Baxter, Toronto, ON, Canada) and
processed in the same manner as the Hb raffimer dilutions
 2003 Blackwell Publishing Ltd, British Journal of Haematology 120: 535–541
(negative controls); and (iii) an aliquot of blood activated
with 0Æ05 U/ml human a-thrombin (Fibrindex; Ortho
Diagnostic Systems, Raritan, NJ, USA) or 20 lmol/l ADP
(Biodata, Horsham, PA, USA) for 10 min at 22C (positive
control for submaximal platelet activation).
After 15 min incubation, 5 ll aliquots of either undiluted,
diluted and activated blood were supplemented with 40 ll
phosphate-buffered saline (PBS) and stained with 5 ll of
saturating concentrations of antibodies: anti-CD41-fluoresc-
ein isothiocyanate (FITC), anti-CD42b-FITC (Immunotech,
Beckman Coulter, Westbrook, ME, USA), PAC-1-FITC, anti-
CD62-phycoerythrin (PE) (BD Biosciences, San Jose, CA,
USA), anti-CD63-PE (Immunotech) and annexin-V-PE (BD
Biosciences). Each sample contained anti-CD41-FITC or anti-
CD42b-FITC and one of the PE-conjugated antibodies for
two-colour analyses. Annexin-V staining was performed in
the presence of 2Æ0 mmol/l Ca2+ as described (Metcalfe et al,
1997). Staining mixtures were incubated for 30 min at 22C
in the dark, fixed with 100 ll of 1% (w/v) paraformaldehyde
in PBS, pH 7Æ4, at 22C for 10 min and diluted with 1 ml of
PBS.
Flow cytometric evaluation of platelets and microparticles
was performed as previously described (Wang et al, 1999;
Leytin et al, 2000). Ten thousand CD41- or CD42-positive
events were acquired on a FACSCalibur flow cytometer (BD
Biosciences) equipped with a 15 mW argon ion laser and
analysed using cellquest software (BD Biosciences), with all
settings in log mode. Analysis was performed on a gated
platelet population identified by characteristic forward and
90 light scatter properties as well as fluorescence with a
platelet marker (anti-CD41 or anti-CD42). The expression of
CD62, CD63 and phosphatidylserine (annexin-V binding),
and activation of glycoprotein (GP) IIbIIIa (PAC-1 binding),
was determined as the percentage positive cells, whereas the
expression of CD41 and CD42b was defined as the mean
channel fluorescence. Platelet-derived microparticles were
quantified on the basis of a characteristic flow cytometric
profile of forward light scatter (smaller than 0Æ8 lm beads)
versus anti-CD41 fluorescence, and expressed as the
percentage CD41- or CD42-positive events. As measured
baseline levels of platelet activation markers, phosphatidyl-
serine expression and microparticle formation in undiluted
resting blood samples differ between different donors, the
results were normalized by calculating the ratio of Hb
raffimer or Ringer’s lactate-diluted blood samples and
thrombin/ADP-stimulated samples to undiluted untreated
blood samples from the same donor. The ratio ¼ 1Æ0 when
platelet activation, phosphatidylserine expression and
microparticle formation in a tested sample are equal to that
in an undiluted sample.
Determination of aperture closure time by PFA-100
. Forty
millilitres of blood was collected as described above and
closure time was measured within 4 h of blood with-
drawal. Aperture closure time was measured on collagen/
epinephrine and collagen/ADP cartridges using the PFA-
100
platelet function analyser (Dade Behring, Miami, FL,
USA) and reported in seconds (Carcao et al, 1998). To
evaluate the effect of blood dilution on closure time assay,
1 ml aliquots of blood were diluted with 0, 20, 50, 100,

200, 300, 400 and 500 ll of PBS, pH 7Æ4, and directly
used for PFA-100
measurements. For measuring the
effect of Hb raffimer on closure time, the following
protocol was developed in order to compensate for the
decrease in platelet and RBC counts and anticoagulant
concentration resulting from the blood dilution. Three
millilitres of blood was placed in polypropylene tubes and
spun at 600 g for 25 min to pellet the cells, including
platelets. Then 66 ll of platelet-poor plasma was discarded
and 6 ll of sodium citrate anticoagulant (0Æ129 mol/l)
added followed by 60 ll of Hb raffimer, to prepare a test
blood sample containing 2 vol percentage Hb raffimer.
Similarly, to prepare diluted blood samples with 10, 20,
40 and 50 vol percentage Hb raffimer, blood was spun,
appropriate amounts of plasma withdrawn and replaced
with anticoagulant and Hb raffimer. In parallel, as a
control for each Hb raffimer dilution, aliquots of blood
were diluted with same volumes of anticoagulant and
Ringer’s solution. Undiluted blood was used as a baseline
control. Diluted and undiluted samples were gently but
thoroughly mixed, incubated at 37C for 30 min and the
PFA-100
closure time measured using 800 ll aliquots of
blood samples.
Data analyses. Means of seven experiments for flow
cytometric and PFA-100
assays are presented. Paired
t-test was used to analyse the effect of Hb raffimer, Ringer’s
solution and ADP; P < 0Æ05 was regarded as significant.
Non-linear regression was used to evaluate the effect of
blood dilution with PBS on the closure time.
RESULTS
Platelet activation
The effect of Hb raffimer on platelet activation was studied
in resting whole blood samples. Figure 1 shows the result of
exposure to different concentrations of Hb raffimer on the
expression of the constitutive platelet membrane receptors,
GPIIbIIIa (CD41) and GPIb (CD42b). For CD42b (Fig 1A),
there was no difference between the expression of CD42b in
blood incubated with any concentration of Hb raffimer up to
50 vol percentage, in comparison with undiluted blood
(ratios ¼ 0Æ97–1Æ00; SD ¼ 0Æ04–0Æ08; P ¼ 0Æ17–1Æ00). For
the CD41 marker (Fig 1B), a small but statistically signifi-
cant decrease in binding of anti-CD41 antibody to platelets
in blood diluted with 50 vol percentage Hb raffimer was
observed compared with undiluted blood (ratio ¼
0Æ93 ± 0Æ05; P ¼ 0Æ03). However, this difference likely
reflects the effect of the twofold blood dilution rather than
downregulation of CD41 by Hb raffimer, as no difference in
CD41 expression was observed between blood diluted to
50% with Hb raffimer versus blood diluted to 50% with 50%
Ringer’s solution (P ¼ 1Æ00). A submaximal concentration
of human a-thrombin (0Æ05 U/ml) yielded significant down-
regulation of CD42b (ratio ¼ 0Æ54 ± 0Æ05; P < 0Æ0001
versus undiluted blood; Fig 1A) and upregulation of CD41
(ratio ¼ 1Æ50 ± 0Æ11; P < 0Æ0001; Fig 1B). Light scatter
characteristics of the CD41- or CD42b-positive platelet
populations suggested that Hb raffimer did not cause
platelet aggregation.
 2003 Blackwell Publishing Ltd, British Journal of Haematology 120: 535–541
HemolinkTM: Effects on Platelets In vitro 537
Fig 1. Effect of Hb raffimer on the expression of constitutive platelet
surface receptors. (A) CD42b (GPIb) and (B) CD41 (GPIIbIIIa). Blood
was incubated with Hb raffimer (vol percentage are shown),
lactated Ringer’s solution (50 vol percentage) or 0Æ05 U/ml human
a-thrombin (Thr). In each experiment, the levels of plasma mem-
brane CD42b and CD41 were determined by flow cytometry and
expressed as the ratio to a baseline level in undiluted untreated
blood. The results of seven experiments are presented; ratio of 1 (no
change from untreated sample) is shown as dotted line and means
as solid lines.
There was no significant increase in the expression of
a-granule protein P-selectin (CD62) when blood was treated
with Hb raffimer (Fig 2A). The blood diluted to 50% with
Hb raffimer showed the highest increase in CD62 expression
(ratio ¼ 1Æ57 ± 1Æ10), but the difference was not significant
compared with undiluted blood (P ¼ 0Æ26) and to 50%
Ringer’s-diluted blood (ratio ¼ 0Æ82 ± 0Æ48, P ¼ 0Æ14); in
contrast, the weak platelet agonist ADP resulted in a ratio of
9Æ27 ± 3Æ87 (P ¼ 0Æ003). While 10% Hb raffimer-diluted
blood showed the greatest increase in CD63 expression
(ratio ¼ 1Æ20 ± 0Æ68, Fig 2B), the difference versus undi-
luted blood was not significant (P ¼ 0Æ47). There was no
difference between the blood diluted to 50% with Hb
raffimer and with 50% Ringer’s solution (ratio ¼
1Æ02 ± 0Æ45 and 0Æ82 ± 0Æ57 respectively; P ¼ 0Æ41).
ADP-stimulated blood, on the other hand, showed a
significant increase in CD63 expression (ratio ¼ 4Æ18 ±
2Æ68; P ¼ 0Æ03).
Figure 2C shows that there was no increase in activation-
induced conformational change in platelet surface GPIIbIIIa
(PAC-1 staining) for blood treated with Hb raffimer. The
50% Hb raffimer-diluted blood showed the highest increase
in PAC-1, but this was similar to that seen with 50%
Ringer’s-diluted blood (ratio ¼ 1Æ48 ± 1Æ00 and 1Æ46 ±
0Æ92 respectively; P ¼ 0Æ65) and did not differ from
undiluted blood (P ¼ 0Æ29 and 0Æ33 respectively). ADP-
stimulated blood showed a significant increase in PAC-1
binding (ratio ¼ 4Æ55 ± 1Æ85; P ¼ 0Æ003).
538 V. Leytin et al
Fig 2. Effect of Hb raffimer on platelet activation as measured by
(A) CD62 and (B) CD63 expression, and (C) conformational changes
of GPIIbIIIa (PAC-1 binding). Blood was incubated with Hb raffi-
mer, Ringer’s solution (50 vol percentage) or ADP (20 lmol/l).
Phosphatidylserine exposure and microparticle formation
Annexin-V binding and microparticle generation were not
increased following exposure of platelets to the different Hb
raffimer concentrations (Fig 3). The changes were highest
in the 10% Hb raffimer sample, but were not significant in
comparison with undiluted blood samples (for annexin-V:
ratio ¼ 1Æ17 ± 0Æ34, P ¼ 0Æ29; for microparticles: ratio ¼
1Æ13 ± 0Æ17, P ¼ 0Æ09). The positive control, ADP-stimu-
lated platelets, however, showed a significant increase in
annexin-V binding (ratio ¼ 8Æ33 ± 6Æ54, P ¼ 0Æ04) and
microparticle formation (ratio ¼ 9Æ77 ± 6Æ02, P ¼ 0Æ02).
Aperture closure time on collagen/epinephrine- and collagen/
ADP-coated membranes
Aperture closure time, as measured by PFA-100
, was
employed for evaluating the effect of Hb raffimer on platelet
‘haemostatic’ function in vitro in whole blood samples.
Closure time is inversely dependent on platelet and RBC
counts, as well as affected by concentrations of anticoagu-
lant (Kundu et al, 1996). Therefore, preliminary experi-
ments were performed to assess the effect of blood dilution
on closure time measurements. As shown in Fig 4, with
increasing dilution using PBS to the values ‡ 16Æ7 vol
percentage, closure time increased significantly (r2 ¼
0Æ965, P < 0Æ0001); at the two higher dilutions (28Æ6%
 2003 Blackwell Publishing Ltd, British Journal of Haematology 120: 535–541
Fig 3. Effect of Hb raffimer on (A) phosphatidylserine externaliza-
tion (annexin V binding) and (B) platelet-derived microparticle (MP)
formation. Blood was incubated with Hb raffimer, Ringer’s solution
(50 vol percentage) or ADP (20 lmol/l).
Fig 4. Effect of blood dilution on platelet ‘haemostatic’ function
in vitro. Blood samples were diluted with PBS, pH 7Æ4, and aperture
closure times were directly measured by PFA-100
on collagen/
epinephrine cartridges without any adjustment of platelet count,
haematocrit and anticoagulant concentration; concentrations of
PBS (vol percentage) in blood–PBS mixtures are shown.
and 33Æ3% of PBS), the aperture was not plugged by
platelets (closure time > 300 s).
Based on this result, for the subsequent dose–response
studies, blood was diluted with Hb raffimer in such a
manner as to avoid the changes in platelet count, haem-
atocrit and anticoagulant concentration (see Materials and
methods). As shown in Fig 5, the progressive dilution
effect seen in Fig 4 was reduced by the modified tech-
nique employed. Although closure time had a trend to
prolongation as dilution of blood with Hb raffimer or
Ringer’s solution increased up to 50%, the differences
between 0 vs 50% dilutions were not significant (Fig 5A:

Fig 5. Effect of Hb raffimer and lactated Ringer’s solution on
platelet ‘haemostatic’ function. Closure time was determined on (A)
collagen/epinephrine and (B) collagen/ADP cartridges. Means ±
SEM for seven experiments are presented. In each experiment, blood
was incubated with either Hb raffimer or Ringer’s solution; see
Materials and methods for experimental design: adjusting platelet
count, haematocrit and anticoagulant concentration to resemble an
undiluted blood sample.
P ¼ 0Æ45–0Æ75; Fig 5B: P ¼ 0Æ07–0Æ13). No difference
between closure times for test (Hb raffimer-diluted) and
control (Ringer’s-diluted) samples was observed at any
dilution with collagen/epinephrine cartridges (e.g. P ¼ 0Æ36
for 50% Hb raffimer vs 50% Ringer’s; Fig 5A). The same
was true for collagen/ADP cartridges, with the exception of
the 40% dilution (P ¼ 0Æ04). The results with the 40%
dilution could, however, be considered as anomalous as this
difference was not observed at either lower or higher
dilutions (e.g. P ¼ 0Æ50 for 50% Hb raffimer vs 50%
Ringer’s; Fig 5B). The trend for increasing closure time at
high dilutions of blood with Ringer’s solution or Hb raffimer,
probably reflects decreasing concentrations of plasma
coagulation factors.
DISCUSSION
Flow cytometry of platelets in whole blood enables the
evaluation of surface platelet membrane glycoproteins
using monoclonal antibodies, and is the current method
of choice to study the functional aspects of platelets.
Different activation mechanisms result in different expres-
 2003 Blackwell Publishing Ltd, British Journal of Haematology 120: 535–541
HemolinkTM: Effects on Platelets In vitro 539
sions of platelet activation markers. In addition, the
presence of different activation molecules on the platelet
may be indicative of different clinical consequences of the
effects of platelet activation, i.e. hypercoagulability, throm-
bosis or bleeding. Hence it is advantageous to measure a
variety of markers of platelet activation. The most widely
studied types of activation-dependent monoclonal antibod-
ies are those directed against conformational changes in
the platelet surface GPIIbIIIa complex (PAC-1) and those
directed against granular proteins which are translocated
to the platelet surface with activation (P-selectin, CD62,
from a-granules and lysosomal protein, CD63, from
lysosomes). P-selectin is the major surface receptor for
neutrophils and monocytes on activated platelets which
mediates leucocyte adhesion. Upon activation, there is also
modulation of expression of normal platelet membrane
glycoproteins, e.g. upregulation of CD41 (GPIIbIIIa, fibrin-
ogen receptor) and downregulation of CD42b (GPIbIX,
vWF receptor). Platelet activation also leads to the
formation of platelet-derived membrane particles that
provide a catalytic surface for the activation of coagulation
factors and promote thrombosis on the platelet surface.
Flow cytometry, providing single-cell data with great
rapidity, accuracy and sensitivity, is well-suited for
quantitative evaluation of platelet microparticles. Another
marker of the activated platelet is expression of annexin-V,
a Ca++-dependent phospholipid binding protein; this
measure of platelet procoagulant activity is linked to
increased exposure of phosphatidylserine on the outer
leaflet of the plasma membrane, providing a catalytic
surface where coagulation factors Va and Xa can bind and
interact.
The effects of Hb raffimer on human platelets in vitro were
analysed, including (i) activation-induced modulation of
constitutive platelet membrane glycoproteins GPIIbIIIa
(CD41) and GPIb (CD42b); (ii) release from a-granules
(CD62) and lysosomes (CD63); (iii) activation-induced
conformational change of GPIIbIIIa (PAC-1 binding); (iv)
phosphatidylserine externalization (annexin V binding); (v)
microparticle formation; and (vi) in vitro ‘haemostatic’
function (closure time). These platelet tests were all
performed in whole blood, in the presence of normal
concentrations of plasma proteins, RBCs and WBCs. None
of these platelet characteristics were adversely affected by
Hb raffimer, even at high concentrations up to 50 vol
percentage. The concentrations of Hb raffimer tested are
3–150-fold higher than those that were used in the clinical
trials: 17–1000 ml per subject (Carmichael et al, 2000;
Mazer, 2000; Cheng, 2001) or 0Æ3–16Æ7 vol percentage,
assuming a blood volume of 5 l.
Cross-linking of Hb dimers into tetramers and oligomers
with o-raffinose in Hb raffimer could potentially enhance the
ability of cross-linked Hb to interact with platelets and cause
platelet aggregation. However, the observed absence of an
effect of Hb raffimer on activation (Figs 1–3) and aggrega-
tion [flow cytometry in this study and aggregometry (Lee
et al, 2000)] of resting platelets in vitro suggests that resting
platelets do not contain Hb binding sites. Other HBOCs [bis
3,5-dibromosalicyl fumarate (DBBF) a-a cross-linked Hb
540 V. Leytin et al
and pyridoxalated polyhaemoglobin] have also been repor-
ted to not induce aggregation of resting platelets (Lalla et al,
1989; Reiss et al, 1992; Mondoro et al, 1998).
The effect of Hb raffimer on agonist-activated platelets,
however, is controversial. Using a whole-blood impedance
aggregometer, Lee et al (2000) demonstrated that Hb
raffimer potentiated aggregation of rabbit platelets induced
by collagen, arachidonic acid and the thromboxane-
mimetic U46619, but not aggregation induced by 10 mol/l
ADP. In contrast, Lieberthal et al (1999) found that
Hb raffimer enhanced aggregation of gel-filtered human
platelets induced by 5 mol/l ADP. However, using the
PFA-100
assay, we found that Hb raffimer did not
potentiate nor inhibit adhesion and aggregation of human
platelets on the collagen/ADP- or collagen/epinephrine-
coated surfaces (Fig 5). The differences may be attributable
to the conditions of measuring platelet reactions in these
models, including differences in (i) shear rate values [low
shear £ 100 s)1 in stirred suspensions in the aggregometer
(Frojmovic, 1998) versus high shear of 5000–6000 s)1 in
the PFA-100
device (Kundu et al, 1996)]; (ii) deposition of
platelets on thrombogenic collagen surface (PFA-100
)
versus agonist-induced aggregation in suspension; or (iii)
accumulation of platelet-released pro-aggregatory (e.g.
thromboxane, ADP and serotonin) and anti-aggregatory
(e.g. nitric oxide) factors in the aggregometer cuvette versus
removal of these factors during aspiration of blood through
a PFA-100
cartridge membrane.
In studying the effects of other HBOCs on agonist-induced
platelet aggregation in vitro, Reiss et al (1992) showed that
10% DBBF a-a cross-linked Hb in 1:1 mixtures with fresh
human blood did not affect aggregation of collagen-, ADP-
and ristocetin-stimulated platelets, as measured by whole-
blood aggregometry. In contrast, Mondoro et al (1998) found
that DBBF-cross-linked Hb did potentiate platelet aggregation
induced by submaximal concentrations of agonists that
cause release of platelet arachidonic acid, but did not
potentiate aggregation induced by ADP; agonist-induced
serotonin release was not potentiated by DBBF-cross-linked
Hb (Mondoro et al, 1998). Pyridoxalated polyhaemoglobin
inhibited aggregation of rat platelets in platelet-rich plasma at
high concentrations of ADP (Lalla et al, 1989).
In accord with the presented in vitro data on platelet
phosphatidylserine externalization (Fig 3A), in vivo studies
in animals and in humans have failed to demonstrate an
effect of Hb raffimer infusion on activation of coagulation or
development of coagulopathy (Kim et al, 1992; Adamson &
More, 1998). Infusion of human-derived Hb raffimer to
rabbits did, however, shorten a prolonged bleeding time and
decreased blood loss in thrombocytopenic and anaemic
animals, and decreased the time to complete carotid
occlusion in normal rabbits (Lee et al, 2000), suggesting
enhancement of platelet functional activity by Hb raffimer
in vivo, possibly via nitric oxide sequestration (Motterlini
et al, 1996; Lee et al, 2000), although other mechanisms
may also be involved (Moro et al, 1994; Alayash, 1999).
Haemodilution is common in the surgical setting as a
result of the administration of intravenous fluids or crystal-
loid priming of the cardiopulmonary bypass apparatus in
 2003 Blackwell Publishing Ltd, British Journal of Haematology 120: 535–541
the case of cardiac surgery. Although haemodilution causes
a decrease in platelet count and coagulation factor levels,
excessive bleeding does not usually occur and a hyperco-
agulable state may exist (Mazer et al, 1995; Iselin et al,
2001; Santoso et al, 2001; Ng et al, 2002). While dilutional
coagulopathy could occur when the large amounts of Hb
raffimer are administered in vivo, preliminary data from
clinical studies in patients undergoing coronary artery
bypass graft (CABG) show no significant effects of Hb
raffimer on blood loss, platelet count, prothrombin time and
partial thromboplastin time (Hardy et al, 2000). However,
the effect of Hb raffimer in other clinical settings should be
further evaluated.
In summary, we have demonstrated that exposure to Hb
raffimer, even at high concentrations (3–150-fold higher
than those used in clinical trials), does not, in the presence
of normal concentrations of plasma proteins, RBC and WBC,
adversely affect human platelet activation or function. The
absence of an effect of Hb raffimer on activation and
aggregation of resting platelets in vitro suggests that resting
platelets do not contain haemoglobin binding sites. Hb
raffimer does not potentiate agonist-induced adhesion and
aggregation of human platelets on the collagen/ADP- or
collagen/epinephrine-coated surfaces under high shear
conditions in the PFA assay.
ACKNOWLEDGMENTS
Source of support: This work was supported by an
unrestricted grant from the Hemosol Inc., Toronto, Ontario,
Canada.
REFERENCES
Adamson, J.G. & More. C. (1998) Hb raffimer, an o-raffinose cross-
linked hemoglobin-based oxygen carrier. In: Blood Substitutes:
Principles, Methods, Products, and Clinical Trials, Vol. 2 (ed. by T.M.
Chang), pp. 62–81. Karger Landes Systems, Basel.
Alayash, A.I. (1999) Hemoglobin-based blood substitutes: oxygen
carriers, pressor agents, or oxidants? Nature Biotechnology, 17,
545–549.
Blajchman, M.A. (2000) Hemoglobin based oxygen carriers: a
backgrounder. Transfusion Alternatives in Transfusion Medicine, SI,
5–7.
Carcao, M.D., Blanchette, V.S., Dean, J.A., He, L., Kem, M.A., Stain,
A.M., Sparling, C.R., Stephens, D., Ryan, G., Freedman, J. &
Rand, M.L. (1998) The platelet function analyzer (PFA-100
): a
novel in-vitro system for evaluation of primary haemostasis in
children. British Journal of Haematology, 101, 70–73.
Carmichael, F.J.L. (2001) Recent developments in hemoglobin-
based oxygen carriers B an update on clinical trials. Transfusion
and Apheresis Science, 24, 17–21.
Carmichael, F.J.L., Ali, A.C.Y., Campbell, J.A., Langlois, S.F., Biro,
G.P., Willan, A.R., Pierce, C.H. & Greenburg, A.G. (2000) A
phase I study of oxidized raffinose cross-linked human
hemoglobin. Critical Care Medicine, 28, 2283–2292.
Chang, T.M.S. (ed.) (1997) Blood Substitutes: Principles, Methods,
Products and Clinical Trails, Vol. 1. Karger Landes Systems, Basel.
Cheng, D.C.H. (2001) Safety and efficacy of o-raffinose cross-linked
human hemogobin (Hb raffimerTM) in cardiac surgery. Canadian
Journal of Anesthesia, 48, s41–s48.

Dietz, N.M., Joyner, M.J. & Warner, M.A. (1996) Blood substitutes:
fluids, drugs, or miracle solutions? Anesthesia and Analgesia, 82,
390–405.
Frojmovic, M.M. (1998) Platelet aggregation in flow: differential
roles for adhesive receptors and ligands. American Heart Journal,
135, s119–s131.
Hardy, J.F., Martineau, R., Cheng, D. & Carmichael, L. (2000)
O-raffinose crosslinked hemoglobin (Hb raffimerTM) reduces
allogeneic blood requirements in patients undergoing CABG.
Proceedings of the 7th International Congress of Cardiothoracic and
Vascular Anesthesia, p. 176a, The Society of Cardiovascular
Anesthesiologists, Quebec City.
Hill, S.E. (2001) Oxygen therapeutics B current concepts. Canadian
Journal of Anesthesia, 48, s32–s40.
Iselin, B.M., Willimann, P.F., Seifert, B., Casutt, M., Bombeli, T.,
Zalunardo, M.P., Pasch, T. & Spahn, D.R. (2001) Isolated
reduction of haematocrit does not compromise in vitro blood
coagulation. British Journal of Anaesthesia, 87, 246–249.
Kim, H.W., Stubal, H. & Greenburg, A.G. (1992) Coagulation
dynamics after hemodilution with polyhemoglobin. Surgery,
Gynecology and Obstetrics, 175, 219–226.
Kundu, S.K., Heilmann, E.J., Sio, R., Garcia, C. & Ostgaard, R.A.
(1996) Characterization of an in vitro platelet function analyzer,
PFA-100
. Clinical and Applied Thrombosis/Hemostasis, 2, 241–
249.
Lalla, F.R., Ning, J. & Chang, T.M. (1989) Effects of pyridoxalated
polyhemoglobin and stroma-free hemoglobin on ADP-induced
aggregation. Biomaterials, Artificial Cells and Artificial Organs,
17, 363–369.
Lee, D.H., Bardossy, L., Peterson, N. & Blajchman, M.A. (2000)
o-Raffinose cross-linked hemoglobin improves the hemostatic
defect associated with anemia and thrombocytopenia in rabbits.
Blood, 96, 3630–3636.
Leytin, V., Mody, M., Mazer, D. & Freedman, J. (2001) The
hemoglobin-based oxygen carrier Hb raffimerTM does not affect
human platelet activation and function in vitro. Blood, 98, 61a.
Leytin, V., Mody, M., Semple, J.W., Garvey, B. & Freedman, J.
(2000) Flow cytometric parameters for characterizing platelet
activation by measuring P-selectin (CD62) expression: theoretical
consideration and evaluation in thrombin-treated platelet popu-
lations. Biochemical and Biophysical Research Communications,
269, 85–90.
Lieberthal, W., Fuhro, R., Freedman, J.E., Toolan, G., Loscalzo, J. &
Valeri, C.R. (1999) o-Raffinose cross-linking markedly reduces
systemic and renal vasoconstrictor effects of unmodified human
hemoglobin. Journal of Pharmacology and Experimental Thera-
peutics, 288, 1278–1287.
Mazer, C.D. (2000) Review of clinical trials of oxygen therapeutics.
Transfusion Alternatives in Transfusion Medicine, SI, 27–31.
 2003 Blackwell Publishing Ltd, British Journal of Haematology 120: 535–541
HemolinkTM: Effects on Platelets In vitro 541
Mazer, C.D., Hornstein, A. & Freedman, J. (1995) Platelet activation
in warm and cold heart surgery. Annals of Thoracic Surgery, 59,
1481–1486.
Metcalfe, P., Williamson, L.M., Reutelingsperger, C.P.M., Swann,
L., Ouwehand, W.H. & Goodall, A.H. (1997) Activation during
preparation of therapeutic platelets affects deterioration
during storage: a comparative flow cytometric study of
different production methods. British Journal of Haematology,
98, 86–95.
Mondoro, T.H., Alayash, A.I., Ryan, B.A., Terle, D.A. & Voatal, J.G.
(1998) Hemoglobin A0 and "-crosslinked hemoglobin ("-DBBF)
potentiate agonist-induced platelet aggregation through the
platelet thromboxane receptor. Artificial Cells, Blood Substitutes
and Immobilization Biotechnology, 26, 1–6.
Moro, M.A., Darley-Usmar, V.M., Goodwin, D.A., Read, N.G.,
Zamora-Pino, R., Feelisch, M., Radomski, M.W. & Moncada, S.
(1994) Paradoxical fate and biological action of peroxynitrite on
human platelets. Proceedings of the National Academy of Sciences of
the United States of America, 91, 6702–6706.
Motterlini, R., Vandegriff, K.D. & Winslow, R.M. (1996) Hemoglo-
bin–nitric oxide interaction and its implications. Transfusion
Medicine Reviews, 10, 77–84.
Ng, K.F., Lam, C.C. & Chan, L.C. (2002) In vivo effect of haemo-
dilution with saline on coagulation: a randomized controlled
trail. British Journal of Anaesthesia, 88, 470–472.
Olsen, S.B., Tang, D.B., Jackson, M.R., Gomez, E.R., Ayala, B. &
Alving, B.M. (1996) Enhancement of platelet deposition by cross-
linked hemoglobin in a rat carotid endarterectomy model. Cir-
culation, 93, 327–332.
Reiss, R.F., Caballero, R. & Hess, J. (1992) Effects of X-linked
hemoglobin on in-vitro platelet function. Biomaterials, Artificial
Cells and Immobilization Biotechnology, 20, 651–655.
Rudolph, A.S., Rabinovici, R. & Feuerstein, G.Z. (eds) (1998) Red
Blood Cell Substitutes. Basic Principles and Clinical Applications.
Marcel Dekker, New York.
Santoso, J.T., Hannigan, E.V., Levine, L., Solanki, D.R. & Mathru, M.
(2001) Effect of hemodilution on tissue perfusion and blood
coagulation during radical hysterectomy. Gynecologic Oncology,
82, 252–256.
Stowell, C.P., Levin, J., Spies, B.D. & Winslow, R.M. (2001) Progress
in the development of RBC substitutes. Transfusion, 41, 287–299.
Wang, C., Mody, M., Herst, R., Sher, G. & Freedman, J. (1999) Flow
cytometric analysis of platelet function in stored platelet con-
centrates. Transfusion Science, 20, 129–139.
Winslow, R.M. (1992) Hemoglobin-Based Red Cell Substitutes. John
Hopkins University Press, Baltimore.
Winslow, R.M., Vandegriff, K.D. & Intaglietta, M. (eds) (1997)
Advances in Blood Substitutes: Industrial Opportunities and Medical
Challenges. Birkhuser, Boston.