Nephrol Dial Transplant (2003) 18: 2088–2096
DOI: 10.1093/ndt/gfg348
Original Article
Increased platelet–monocyte aggregates and cardiovascular disease in
end-stage renal failure patients
Neil Ashman1, Marion G. Macey2, Stanley L. Fan1, Urooj Azam2 and Muhammad M. Yaqoob1
1
Department of Renal Medicine and Transplantation and 2Department of Haematology, St Bartholomew’s and the
Royal London Hospitals, London, UK
Abstract
Background. Atherosclerotic cardiovascular disease is
a major cause of morbidity and mortality in patients
with end-stage renal disease. This excess morbidity
cannot be entirely explained by well-recognized conventional and novel risk factors alone, and occurs
irrespective of dialysis modality. Recent evidence
suggests that the activation of platelets and their
interaction with circulating cells are important independent risk factors for atherosclerosis in non-uraemic
patients. We therefore studied platelet activation and
circulating platelet–leucocyte aggregates in stable
patients without evidence of cardiovascular disease on
continuous ambulatory peritoneal dialysis (CAPD) and
haemodialysis and investigated an association with
cardiovascular events.
Methods. Immunofluorescent flow cytometry was used
to measure the percentage of P-selectin- (CD62P)
positive platelets, the percentage of platelet–neutrophil
and platelet–monocyte aggregates, and the expression
of the P-selectin ligand, P-selectin glycoprotein ligand-1
(PSGL-1, CD162) on leucocytes in haemodialysis and
CAPD patients and normal controls. The platelet count
and the mean platelet component (MPC, a measure of
platelet activation) were determined on the ADVIATM
120 Haematology System (Bayer, NY).
Results. Platelet activation as assessed by MPC or
CD62P expression was significantly increased in
haemodialysis but not CAPD patients compared with
controls. Circulating platelet–monocyte aggregates
were significantly increased in parallel with a significant
reduction in PSGL-1 expression on monocytes in both
patient groups compared with normal controls. The
presence of higher platelet–monocyte aggregates in
dialysis patients was associated with increased cardiovascular events.
Correspondence and offprint requests to: Dr Marion G. Macey,
Department of Haematology, The Royal London Hospital,
Whitechapel, London E1 1BB, UK. Email: marion.macey@
bartsandthelondon.nhs.uk
Conclusion. We describe increased platelet–monocyte
aggregates with reduced leucocyte PSGL-1 expression
in patients with end-stage renal disease irrespective of
dialysis modality, associated with an increased risk of
cardiovascular disease. These findings may suggest a
novel mechanism by which accelerated atherosclerosis
occurs in uraemic patients.
Keywords: CD62P; dialysis; platelet–monocyte aggregates; platelets; PSGL-1
Introduction
Atherosclerotic cardiovascular disease is a major cause
of morbidity and mortality in patients with end-stage
renal disease (ESRD) undergoing renal replacement
therapy [1]. Despite significant progress in dialysis
technology and in the prevention and treatment of renal
and coronary artery disease, the prevalence of
cardiovascular disease has remained static during the
last decade, and is only partly explained by conventional risk factors [2].
In the non-uraemic state, vascular inflammation
plays an essential role in advancing endothelial injury
and atherogenesis [3], as well as in the formation and
propagation of platelet-dependent thrombi in acute
coronary syndromes [3,4]. Long recognized as having a
role in inflammation, platelets and platelet–leucocyte
aggregates are now known to contribute to ongoing
injury at atheromatous sites, and in plaque disruption
[5]. Platelet P-selectin (CD62P) interacts with its natural ligand on neutrophils and monocytes, P-selectin
glycoprotein ligand-1 (PSGL-1), to allow formation of
heterotypic aggregates, thus providing an anchoring
source for inflammatory cells on activated platelets [6].
These bioactive platelet–monocyte aggregates have
been shown to be important in evolving coronary
syndromes [4,7] in humans. As they are involved in
ongoing vascular inflammation and thrombosis in
Nephrol Dial Transplant 18(10) ß ERA–EDTA 2003; all rights reserved
Platelet–monocyte aggregates in ESRF
2089
potentially unstable plaques, they may provide a more
useful marker of cardiovascular disease than markers
of downstream myocyte injury [6].
Uraemia and dialysis induce a pro-inflammatory
state, with widespread microvascular and circulatory
changes [8]. This is reflected in elevated acute phase
proteins and inflammatory molecules, and implicated in
the high incidence of cardiovascular disease in ESRD
[9]. Investigation into platelet dysfunction in uraemia
has focused largely on prolonged bleeding time, and the
role of nitric oxide in platelets of chronic and end-stage
renal failure (ESRF) patients, and on surface receptor
abnormalities on haemodialysis [10–13]. Platelet activation and platelet–leucocyte interactions have been
studied in haemodialysis patients as a parameter of
membrane bioincompatibility, but these abnormalities
have not been investigated in patients on chronic
ambulatory peritoneal dialysis (CAPD).
The cardiovascular mortality in patients on CAPD is
similar to that on haemodialysis. In light of the recent
reported association between platelet activation,
platelet–leucocyte aggregates and cardiovascular morbidity in non-uraemic patients, we investigated circulating platelet–leucocyte interactions in patients on CAPD
and haemodialysis as a potential cardiovascular risk
factor.
Subjects and methods
Patient population
Patients (48) with ESRF who were treated on the out-patient
maintenance renal replacement therapy programme at
St Bartholomew’s and the Royal London hospitals were
recruited into the study in 2001. Of these patients, 23 had
been maintained on haemodialysis and 25 on CAPD. Blood
samples were obtained for measurement on the ADVIATM120
haematology system as detailed below. A group of 10 patients
from each dialysis group also had samples analysed for
platelet–leucocyte aggregates and leucocyte PSGL-1 expression by flow cytometry. A control group of 40 individuals with
normal renal function and no medical history of vascular
disease was also enrolled: within this group, 20 were analysed
for platelet–leucocyte aggregates and PSGL-1 expression.
Patient characteristics are shown in Table 1. Because the
normal controls were younger than the dialysis cohort, the
latter were stratified into two age-related groups, a young
group (aged 29.6 ± 3.2 years) and an older group (aged
55.5 ± 4.0 years, mean ± SEM). Analysis of these two groups
showed no difference for any experimental result. For
example, percentage platelet–monocyte aggregates in the
younger group was 25.8 ± 6.5%, whilst in the older group it
was 16.8 ± 3.6% (mean ± SEM, P ¼ NS). Both age groups
were significantly different from normal controls (P < 0.001).
Amongst those on dialysis, the cause of renal failure varied,
and included glomerulonephritis, adult polycystic kidney
disease, reflux nephropathy, chronic pyelonephritis and
unknown cause (scarred and shrunken kidneys on ultrasound
at presentation). Patients with ESRF due to diabetes mellitus,
renovascular disease or active vasculitis were excluded, as
were all patients on medication for ischaemic heart disease, or
with history of angina or coronary artery bypass grafting,
or known left ventricular hypertrophy. No subject had
been treated with aspirin, clopidogrel, HMG-co-A reductase
inhibitors or any non-steroidal anti-inflammatory agent for at
least 1 month prior to the study. All subjects were on a stable
erythropoeitin dose for the preceding 3 months. All subjects
were well when included in the study, with no concurrent
infection within 3 months of the study.
All haemodialysis patients were dialysed three times per
week with a HemophanTM modified cellulose hollow fibre
dialysis membrane delivering a Kt/V according to UK Renal
Association guidelines (Kt/V of >1.2). All were stable and
not prone to intradialytic events such as hypotension or
cardiovascular instability. Venepuncture for samples was
performed before any given haemodialysis session after the
short interdialytic interval. All samples were obtained prior to
any heparin bolus. All CAPD patients used a BaxterTM Minisolo system and performed four 2–2.5 l exchanges per day,
again delivering adequate dialysis according to UK Renal
Association guidelines (weekly Kt/V of >2.0). Informed
consent was obtained from all patients as per local ethical
protocols in accordance with the current (October, 2000)
revision of the Declaration of Helsinki. Blood was collected
into vacutainers containing EDTA and sodium citrate in all
cases and analysed immediately.
Materials
Tyrodes salt solution (TS; CaCl22H2O 0.265 g/l, MgCl2
6H2O 0.214 g/l, KCl 0.2 g/l, NaH2CO3 1.0 g/l, NaCl 8.0 g/l,
Na2HPO4 0.05 g/l, glucose 1.0 g/l) was from Sigma (Poole,
Dorset, UK). K3EDTA and sodium citrate in Vacutainers
were from BD Biosciences (Cowley, Oxford, UK).
Table 1. Baseline characteristics of the three cohorts
Age (years)
Gender
Time on dialysis (months)
C-reactive protein (mg/ml)
Serum albumin (g/l)
Intact PTH (pmol/l)
Calcium/phosphate product (mmol/l)
Normal controls
Haemodialysis
CAPD
33.7 (26–55)
12 M/8 F
48.2 (23–72)
7 M/3 F
44.2 (4–165)
12.4 (0–37)
41.8 (34–47)
64.4 (11–189)
4.49 (2.83–7.8)
42.4 (24–69)
6 M/4 F
38.5 (8–131)
8.6 (0–28)
38 (32–44)
48.1 (3–168)
4.51 (2.26–8.21)
No significant difference between CAPD patients or haemodialysis patients for any variable was found. Although the normal controls were
younger than the dialysis cohorts, no age-related difference in experimental outcome was shown (see Subjects and methods).
2090
Antisera
Fluorescein isothiocyanate (FITC)-conjugated mouse IgG1,
FITC–CD62P (CLBThromb/6), FITC–CD42a (SZ1) and
phycoerythrin (PE)-conjugated CD45 were from Immunotech
(Luton, Bedfordshire, UK). PE-conjugated CD162 (KPL-1
clone) was from BD (Oxford, UK).
Assessment of platelet count and platelet
activation on the ADVIATM 120
Whole blood samples were taken into Vacutainers that
contained K3EDTA. Samples were held at ambient temperature, and analysed at 30 min after venesection. The platelet
count (PLT), platelet crit and mean platelet component
(MPC) concentration were determined using the
ADVIATM120 Haematology system (Bayer Corporation,
Tarrytown, NY). The system was calibrated and standardized
prior to use with ADVIATM-SETpoint Haematology Control
and ADVIATMOPTIpoint, respectively (Bayer Corporation).
The ADVIATM120 system has a laser optical assembly that
consists of a laser diode, a flow cell and detector assemblies,
and is essentially a flow cytometer. A laser diode is used to
produce monochromatic light at 675 nm. The light from the
laser diode is focused onto the flow cell. The sample/sheath
stream in the flow cell contains platelets and red cells that are
iso-volumetrically sphered with sodium dodecyl sulfate (SDS)
using a procedure first described by Kim and Ornstein [14].
Platelets passing through the flow cell scatter light. The
scattered light in the forward direction at low and high angles
is detected by photodiodes and generates two signals. Using
the Mie theory of light scattering for homogeneous spheres,
the high angle light scatter measurement is converted into
refractive index or as presented on the ADVIATM120 system
as the MPC. The high angle light scatter is analogous to the
side scatter detected by flow cytometers, and so a decrease
in high angle light scatter may be associated with a decrease
in granularity. This led us to consider whether the
ADVIATM120 system could be used to measure platelet
activation based upon changes in refractive index. Since
platelet activation is associated with degranulation, we
speculated that the increase in CD62P cell surface expression
would correlate with a fall in refractive index (RI). We have
shown that the ADVIATM120 system may be used to measure
changes in light scatter due to changes in RI in activated
platelets. In vitro stimulation of normal platelets in whole
blood by bovine thrombin resulted in activation leading to
increased platelet CD62P expression and a concomitant
decrease in RI. This response was dose and time dependent
and could be inhibited by ridogrel, a specific inhibitor of
thromboxane synthesis, at levels known to be occur in blood
(10–7 M) and tissues (10–5 M) [15].
Measurement of the percentage of platelets expressing
CD62P and of the percentage of leucocytes that had
platelets attached (platelet–leucocyte aggregates)
Sodium citrate anticoagulated blood (5 ml) was labelled at
ambient temperature with 5 ml of one of the following: (i)
FITC–isotype control; (ii) FITC–CD62P; (iii) PE–CD45 and
FITC–isotype control; or (iv) PE–CD45 and FITC–CD42a,
in 90 ml of TS for 5 min. Previous studies have shown that
antibody binding is complete within this time. Samples were
N. Ashman et al.
diluted to 1 ml with TS and analysed immediately by flow
cytometry.
Measurement of leucocytes expressing CD162
Sodium citrate-anticoagulated blood (5 ml) was labelled at
ambient temperature with PE–CD162 (5 ml) in 90 ml of TS for
5 min. Samples were then diluted to 1 ml with TS and analysed
immediately by flow cytometry.
Flow cytometry
Blood cells were analysed on a FACScan (BD Biosciences)
equipped with CellQuestÕ software. The flow cytometer was
calibrated and standardized prior to use with fluorochromelabelled beads (Fluorospheres; Dako, Ely, Cambridgeshire,
UK). For the analysis of CD62P expression, data were
acquired in real time with a primary gate set on a dual
parameter histogram of forward light scatter (FLS) logarithmic scale (abscissa) and side light scatter (SLS) logarithmic
scale (ordinate). This facilitated identification of the platelets
within the blood and was confirmed by the analysis of CD42a
expression. Background fluorescence was assessed with
platelets labelled with the FITC-conjugated isotype control
antibody. Cursors were set in a single parameter histogram
of frequency (ordinate) and green fluorescence intensity
(abscissa), so that <1% of the platelets stained positively with
the control antibody. Changes in CD62P expression (green
fluorescence logarithmic scale), together with FLS and SLS,
were then recorded on the gated platelets.
For the analysis of CD162 expression, data were acquired
in real time with a primary gate set on a dual parameter
histogram of orange fluorescence (FLS) logarithmic scale
(abscissa) and SLS linear scale (ordinate); this allowed
identification of CD162-positive leucocytes within the blood
based upon their PE fluorescence and granularity. The
leucocytes were then gated to a dual parameter histogram of
FLS linear scale (abscissa) and SLS linear scale (ordinate).
This facilitated identification of the granulocytes within the
blood. CD162 expression (orange fluorescence logarithmic
scale), together with those of FLS and SLS, was then recorded
on the gated granulocytes.
For the analysis of platelet–leucocyte aggregates, cells
were analysed first in a histogram of side scatter (logarithmic
scale ordinate) and orange fluorescence (logarithmic scale
abscissa). Leucocytes identified by their positive staining with
PE CD45 were gated to a dot plot of green fluorescence
(logarithmic scale ordinate) and orange fluorescence (logarithmic scale abscissa). Events that were both green and
orange were considered to be platelet–leucocyte aggregates
and recorded as a percentage of a total of 10 000 gated
leucocytes. Platelet–leucocyte aggregates could then be gated
to a histogram of side scatter (logarithmic scale ordinate) and
orange fluorescence (logarithmic scale abscissa) to identify,
by their characteristic SLS, which leucocytes were forming
platelet–leucocyte aggregates.
Statistical analysis
Results from the flow cytometer and the ADVIATM120
Haematology system were compared using analysis of
variance (ANOVA) with appropriate post hoc tests to allow
Platelet–monocyte aggregates in ESRF
for the confounding potential of multiple comparisons. Data
for these analyses are expressed as mean ± SEM. Biochemical
parameters in the CAPD and haemodialysis cohorts were
compared using the unpaired t-test, two-tailed for independent variables, to test for significant differences between
the patient groups and the control group when normally
distributed. Non-parametric data were analysed using the
Mann–Whitney U-test. Correlation was sought using Pearson
rank correlation. Significance was considered at P < 0.05.
Results
2091
as channel fluorescence units) was significantly lower in
the CAPD patients (597 ± 11.2; n ¼ 10; P < 0.05) and
haemodialysis patients (605 ± 10.3; n ¼ 10; P < 0.05)
than in the control group (631 ± 6; n ¼ 20) (Figure 2).
Expression of CD162 on monocytes
Similarly, all monocytes in the blood samples from both
the control and patient groups expressed CD162.
However, the median fluorescence intensity was lower
in the CAPD patients (639 ± 13; n ¼ 10; P < 0.01) and
haemodialysis patients (647 ± 8; n ¼ 10; P < 0.05) than
in the control group (678 ± 6; n ¼ 20) (Figure 2).
Blood counts
PLT (mean ± SEM; n ¼ 40) for the normal controls
was 264 ± 7 109/l, that for the CAPD patients was
277 ± 20 109/l (n ¼ 25) whilst that for the haemodialysis patients was 216 ± 14 109/l (n ¼ 23). The PLT
for the haemodialysis patients was significantly lower
(P < 0.01) than in the normal controls (Figure 1).
MPC
The MPC, a measure of platelet activation, for the
normal controls was (mean ± SEM; n ¼ 40)
27.5 ± 0.17 g/dl, that for the CAPD patients was
26.8 ± 0.33g/dl (n ¼ 25) whilst that for the haemodialysis patients was 26.4 ± 0.26 g/dl (n ¼ 23). The
MPC for the haemodialysis patients was significantly
reduced compared with the controls (P < 0.05), but no
statistically significant difference was found between
controls and CAPD patients (P ¼ 0.045, NS with post
hoc adjustment for multiple comparisons).
Expression of CD62P on platelets
Shortly after venesection, there was a low percentage
of platelets that expressed CD62P. The mean ± SEM
for the normal controls (n ¼ 40) was 0.985 ± 0.19%
CD62P-expressing platelets, that for the CAPD
patients was 1.53 ± 0.32% (n ¼ 25) whilst that for the
haemodialysis patients was 2.12 ± 0.34% (n ¼ 23). The
percentage CD62P-positive platelets in blood from
the haemodialysis patients was significantly higher
(P < 0.01) than in the normal controls. There was
no difference in percentage CD62P-positive platelets
between normal controls and CAPD patients (P ¼ 0.08,
NS with post hoc adjustment for multiple comparisons),
nor a difference between the haemodialysis and CAPD
groups. The increased platelet P-selectin expression and
decrease in MPC, both markers of platelet activation,
showed a significant inverse correlation in haemodialysis patients (r ¼ –0.51, P ¼ 0.019).
Expression of CD162 on neutrophils
All neutrophils (100%) in the blood samples from both
the control and patient groups expressed CD162.
However, the median fluorescence intensity (measured
Platelet–neutrophil aggregate formation
Immediately after venesection, a small percentage of
neutrophils associated with platelets were found in
blood from all controls and patients. The percentage of
platelet–neutrophil aggregates (mean ± SEM; n ¼ 20)
for the normal controls was 7.4 ± 0.8%, that for the
CAPD patients was 13.5 ± 5.8% (n ¼ 10) and that for
the haemodialysis patients was 8.3 ± 1.9% (n ¼ 10).
There were no statistically significant differences
between CAPD patients and the controls. In the
haemodialysis cohort, platelet–neutrophil aggregates
were significantly increased (P < 0.05) compared with
controls, but not compared with CAPD patients
(Figure 3).
Platelet–monocyte aggregate formation
Immediately after venesection, a small percentage of
monocytes were to be found associated with platelets in
blood from all patient samples, but were found in only a
few normal controls. The percentage platelet–monocyte
aggregates (mean ± SEM; n ¼ 20) for the normal
controls was 3.72 ± 1.39%, that for the CAPD patients
was 17.4 ± 4.7% (n ¼ 10) and that for the haemodialysis patients was 22.2 ± 4.7% (n ¼ 10). The percentage platelet–monocyte aggregates found in the CAPD
and haemodialysis patients was significantly greater
(P < 0.05 and P < 0.001 respectively) than in the
normal control group (Figure 3).
Effect of parathyroid hormone on platelets and
aggregate formation
We have shown that platelet calcium is influenced by
ambient parathyroid hormone (PTH) concentration in
ESRD [16]: this may then affect platelet reactivity. We
thus analysed two dialysis groups by PTH concentration by arbitrarily examining a 10-fold change in PTH.
The low PTH group (n ¼ 8) had a mean PTH of
9.4 pmol/l, whilst the high PTH group (n ¼ 12) had a
mean of 90.4 pmol/l. In the two groups, the percentage
of platelets expressing P-selectin was similar (low PTH
group, 1.65 ± 0.6%; high PTH group 1.39 ± 0.3%;
mean ± SEM, P ¼ 0.68), as was mean platelet component (low PTH, 26.3 ± 0.45 g/dl; high PTH,
2092
N. Ashman et al.
Fig. 1. The analysis of platelet–leucocyte aggregates and leucocyte expression of CD162 in whole blood. Blood was stained with
PE-conjugated CD45 and FITC-conjugated CD42a. Leucocytes were identified by their positive staining with PE–CD45 in a plot of side
scatter (logarithmic scale, ordinate) vs PE fluorescence (logarithmic scale, abscissa) (dot plots A and E). Back-gating these events to a plot of
side scatter (logarithmic scale, ordinate) vs forward scatter (logarithmic scale, abscissa) (dot plots B and F) showed the typical light scatter
characteristics of neutrophils, monocytes and lymphocytes, regions R1, R2 and R3, respectively. Gated events were displayed in a plot of
FITC–CD42a fluorescence (logarithmic scale, ordinate) and PE–CD45 fluorescence (logarithmic scale, abscissa) (dot plots C and G). Events
that were both CD42a and CD45 positive (upper right quadrant, C and G) were considered to be platelet–leucocyte aggregates. Histograms
(D) and (H) show the cell count (logarithmic scale, ordinate) vs CD162 PE (logarithmic scale, abscissa) and illustrate the analysis of the
expression of CD162 on gated leukocytes. An example of the analysis performed on blood from a haemodialysis patient analysed within
30 min after venesection is illustrated. Plots (C) and (D) were obtained from an aliquot of a peripheral blood sample that had been analysed
gating on granulocytes (region R1), and plots (G) and (H) from the same aliquot that had been analysed gating on monocytes (region R2).
Platelet–monocyte aggregates in ESRF
2093
Fig. 2. P-selectin glycoprotein ligand-1 (PSGL-1, CD162) expression on leucocytes. PSGL-1 was significantly reduced in both CAPD patients
(P < 0.05 for neutrophils, P < 0.01 for monocytes) and HD patients (P < 0.05 for neutrophils, P < 0.05 for monocytes) against normal
controls. Results are expressed as mean ± SEM. Statistical analysis using ANOVA with post hoc Bonferroni test, * ¼ significance achieved.
Fig. 3. Platelet–leucocyte aggregates. Platelet–monocyte aggregates were significantly increased in both cohorts (P < 0.05 for CAPD patients,
P < 0.001 for HD patients). Platelet–neutrophil aggregates were significantly increased in patients on HD (P < 0.05). Although platelet–
neutrophil aggregates were increased against normal controls in CAPD patients, this difference was non-significant. Results are expressed as
mean ± SEM. Statistical analysis using ANOVA with post hoc Bonferroni test, * ¼ significance achieved.
26.6 ± 0.44 g/dl; mean ± SEM, P ¼ 0.62). No difference was found for any other parameter. In the low
PTH group, the percentage of platelet–monocyte
aggregates was 20.8 ± 4.5%, whilst in the high PTH
group it was 16.0 ± 4.7 (mean ± SEM, P ¼ 0.52).
Platelet–monocyte aggregates and
cardiovascular events
In non-uraemic subjects, platelet–monocyte aggregates
have been associated with the presence of cardiovascular disease. Given the excess of these aggregates in
2094
N. Ashman et al.
Fig. 4. Cardiovascular disease and platelet–monocyte aggregates (PMA). Amongst ESRF patients, higher platelet–monocyte aggregates
(31.7 ± 2.5%; n ¼ 11) are associated with an increased incidence of cardiovascular events after at least 1 year follow-up as defined by cardiac
death, proven vascular disease or de novo left ventricular hypertrophy against patients with lower platelet–monocyte aggregates (6.5 ± 0.8%;
n ¼ 9). Open bars, number of dialysing patients free of cardiovascular disease; shaded bars, number with cardiovascular disease, P ¼ 0.02,
Yates-corrected 2 test.
both dialysis cohorts, we sought clinical evidence of
cardiovascular disease after a mean of 16.2 months
(range 12–22) of follow-up. The group had been
selected on the basis of not having clinically apparent
cardiovascular disease at recruitment. The composite
end points were cardiac death, a vascular event
(myocardial infarct, the development of angina
pectoris) or left ventricular hypertrophy (LVH) on
echocardiogram or electrocardiogram. The 20 dialysis
patients were thus stratified into a group with lower
platelet–monocyte aggregates (6.5 ± 0.8%; n ¼ 9, range
2.9–10%), and a higher group (31.7 ± 2.5%; n ¼ 11,
range 16.7–47%). The lower group was defined as
falling within the normal range (mean ± 2 SD for
normal controls, or <15.5%). After at least 1 year’s
follow-up, eight patients of 11 within the higher group
had evidence of cardiovascular disease or had died a
cardiac death (Figure 4). Two patients had died of
myocardial infarction, one had a non-fatal infarct and
one had a cerebrovascular accident (with documented
LVH), and a further four patients had evidence of
LVH. In the lower platelet–monocyte aggregates
group, one patient had had non-transmural infarcts
and developed an ischaemic cardiomyopathy. A further
patient had died of overwhelming sepsis. The latter
patient, and the remaining seven had no evidence of
LVH, nor any history of angina. This achieved
significance at P ¼ 0.02 (Yates-corrected 2 test).
Discussion
We selected patients with ESRF and no evidence of
cardiovascular disease on haemodialysis or CAPD to
examine the activation of platelets and the formation of
platelet–leucocyte aggregates. We selected a group free
of apparent atherosclerosis to allow prospective followup, and to discover whether circulating cell changes
preceded clinical vascular syndromes in uraemia.
Platelet interactions with other circulating cells and
with the endothelium play a key part in the pathogenesis of atherosclerosis and infarction in non-uraemic
models. As accelerated vascular disease is found in
excess regardless of dialysis modality in ESRF patients,
we examined platelet surface changes in CAPD and
haemodialysis patients.
P-selectin is translocated to the surface of activated
platelets, where it contributes to platelet-assisted
enhancement of thrombosis at sites of endothelial
injury. Platelet–platelet aggregates are stabilized on
subendothelial matrix, and then leucocyte ‘rolling’ and
recruitment on this damaged, platelet-rich surface is
enabled by efficient binding to leucocyte PSGL-1—this
leads to tighter platelet–leucocyte binding, increased
local fibrin deposition [17] and supports more stable
interactions between platelets and leucocytes. Crosslinking of PSGL-1 by recruited monocytes may then
amplify production of pro-coagulant tissue factor,
tumour necrosis factor-, monocyte chemoattractant
protein-1 and chemokines. In flow conditions, platelet
P-selectin also binds PSGL-1 on the surface of
monocytes and neutrophils to form circulating mixed
cell aggregates which are stable over many hours, in
contrast to the transient rolling interactions. PSGL-1
on leucocytes may also bind endothelial P-selectin in
early engagement to enable tethering then contributing
to platelet sequestration into injured tissue with an
intensification of the local inflammatory response
[4,18,19].
Platelet–monocyte aggregates in ESRF
Platelet P-selectin expression is increased during
haemodialysis in association with increased platelet–
leucocyte aggregation, an effect replicated in vitro
by stimulating platelet activation with ADP [10–12].
This aggregate formation has since been reported to
occur via an interaction between platelet P-selectin
and leucocyte sialyl-Lewis x (CD15s) in haemodialysis
[11,13]. We confirm platelet activation as measured by
CD62P expression or reduced MPC to be significantly
increased in haemodialysis. In CAPD patients,
although a similar trend is apparent, platelet activation
is not statistically significantly increased. This may
reflect the regular exposure of platelets to relatively
bioincompatible membranes in haemodialysis patients.
It is also worth acknowledging that degranulated
platelets rapidly lose their surface P-selectin, but
continue to circulate and function, capable of cell–cell
interaction.
Leucocyte PSGL-1 is the natural ligand for platelet
P-selectin: expression is reduced on both monocytes
and neutrophils in both dialysis cohorts. Again, this
may represent PSGL-1 redistribution or shedding in
activated monocytes or neutrophils. Importantly, surface expression does not necessarily correlate with
function [20]. We find that despite this reduction in
expression, and the lack of significant platelet activation in CAPD patients, platelet–monocyte aggregation is enhanced in both CAPD and haemodialysis.
Although most aggregation occurs through P-selectin–
PSGL-1 binding, a recent report has identified
significant platelet–monocyte binding through other
mechanisms [21]: this may explain the findings in our
study.
In patients with stable coronary artery disease,
platelet–monocyte aggregates are significantly increased
compared with normal controls [22]. Furthermore,
high aggregates discriminate patients with myocardial
infarction from other causes in those presenting with
chest pain [7,23]: this was not the case with platelet Pselectin alone. Furman showed that when patients had
15.3 ± 3.0% platelet-positive monocytes, they were
proven subsequently to have a myocardial infarct [7];
dialysis patients with no history of coronary artery
disease, or risk factors for ischaemic heart disease
other than uraemia had 17.4 ± 4.7% (CAPD) and
22.2 ± 4.7% (haemodialysis) platelet-positive monocytes.
Recently, other ligands potentially involved in this
cell–cell interaction have been implicated in cardiovascular risk in non-uraemic patients [24], and a potential
mechanism by which platelet–monocyte aggregates
might be implicated in atherogenesis described in
animals [25]. It is clear that bioactive platelet–monocyte
aggregates are found in ESRF patients on both
haemodialysis and CAPD, where they are associated
with cardiovascular events, as occurs in the nonuraemic setting. They may contribute to local
thrombotic changes in acute plaque erosion and in the
evolution of atherosclerosis, but may yet prove an
epiphenomenon reflecting widespread inflammatory
changes in the microcirculation associated with
2095
uraemia. Larger scale prospective studies are needed
to assess the predictive value of these novel parameters
in uraemic cardiovascular morbidity and mortality.
Conflict of interest statement. None declared.
References
1. Herzog CA, Ma JZ, Collins AJ. Poor long-term survival after
acute myocardial infarction among patients on long-term
dialysis. N Engl J Med 1998; 339: 799–805
2. Foley RN, Parfrey PS, Sarnak MJ. Epidemiology of
cardiovascular disease in chronic renal disease. J Am Soc
Nephrol 1998; 9 [Suppl 12]: S16–S23
3. Ross R. Atherosclerosis—an inflammatory disease. N Engl J
Med 1999; 340: 115–126
4. McEver RP. Adhesive interactions of leukocytes, platelets and
the vessel wall during hemostasis and inflammation. Thromb
Haemostasis 2000; 86: 746–756
5. Van der Wal AC, Becker AE. Atherosclerotic plaque rupture—
basis of plaque stability and instability. Cardiovasc Res 1999;
41: 344–354
6. Freedman JE, Loscalzo J. Platelet–monocyte aggregates:
bridging thrombosis and inflammation. Circulation 2002; 105:
2130–2132
7. Furman MI, Barnard MR, Krueger LA et al. Circulating
monocyte–platelet aggregates are an early marker of
acute myocardial infarction. J Am Coll Cardiol 2001; 38:
1002–1006
8. Schomig M, Eisenhardt A, Ritz E. The microinflammatory
state of uremia. Blood Purif 2000; 8: 327–332
9. Zimmermann J, Herrlinger S, Pruy A et al. Inflammation
enhances cardiovascular risk and mortality in haemodialysis
patients. Kidney Int 1999; 55: 648–658
10. Gawaz MP, Mujasi SK, Schmidt B, Gurland HJ. Platelet–
leukocyte aggregation during hemodialysis. Kidney Int 1994; 46:
489–495
11. Carreno MP, Stuard S, Bonomini M et al. Cell-associated
adhesion molecules as early markers of bioincompatibility.
Nephrol Dial Transplant 1996; 11: 2248–2257
12. Bonomini M, Sirolli V, Stuard S, Settefrati N. Interactions
between platelets and leukocytes during hemodialysis. Artif
Organs 1999; 23: 23–28
13. Stuard S, Carreno MP, Poignet JL et al. A major role for
CD62P/CD15s interaction in leukocyte margination during
hemodialysis. Kidney Int 1995; 48: 93–102
14. Kim YR, Ornstein L. Isovolumetric sphering of erythrocytes
for more accurate and precise cell volume measurements by
flow cytometry. Cytometry 1983; 3: 419–427
15. Macey MG, Carty E, Webb L et al. Use of mean
platelet component to measure platelet activation on
the ADVIA 120 haematology system. Cytometry 1999; 38:
250–255
16. Fluck RJ, McMahon AC, Alameddine FM, Dawnay AB,
Baker LRI, Raine AE. Platelet cytosolic free calcium
concentration and parathyroid hormone: changing relationships
with haemodialysis in end-stage renal disease. Clin Sci (Lond)
1992; 82: 651–658
17. Palabrica T, Lobb R, Furie RC et al. Leucocyte accumulation
promoting fibrin deposition is mediated in vivo by P-selectin on
adherent platelets. Nature 1992; 81: 62–69
18. Yang J, Furie BC, Furie B. The biology of P-selectin
glycoprotein ligand-1: its role as a selectin counterreceptor in
leukocyte–endothelial and leukocyte–platelet interaction.
Thromb Haemostasis 1999; 81: 1–7
19. Hoffman M, Monroe DM. A cell-based model of hemostasis.
Thromb Haemostasis 2001; 85: 958–965
2096
20. Furie B, Furie BC, Flaumenhaft R. A journey through Pselectin: the molecular basis of granule secretion, signalling and
cell adhesion. Thromb Haemostasis 2001; 86: 214–221
21. Sarma J, Laan CA, Alam S et al. Increased platelet binding to
circulating monocytes in acute coronary syndromes. Circulation
2002; 105: 2166–2171
22. Furman MI, Benoit SE, Barnard MR et al. Increased platelet
reactivity and circulating monocyte–platelet aggregates in
patients with stable coronary artery disease. J Am Coll Cardiol
1998; 31: 352–358
N. Ashman et al.
23. Michelson AD, Barnard MR, Krueger LA et al. Circulating
monocyte–platelet aggregates are a more sensitive marker of in
vivo platelet activation than platelet surface P-selectin.
Circulation 2001; 104: 1533–1537
24. Heeschen C, Dimmeler S, Hamm CW et al. Soluble CD40
ligand in acute coronary syndromes. N Engl J Med 2003; 348:
1104–1111
25. Huo Y, Schober A, Forlow SB et al. Circulating activated
platelets exacerbate atherosclerosis in mice deficient in
apolipoprotein E. Nat Med 2003; 9: 61–67
Received for publication: 12.11.02
Accepted in revised form: 5.5.03