Articles in PresS. Am J Physiol Heart Circ Physiol (August 26, 2004). doi:10.1152/ajpheart.01078.2003
AJP Manuscript ID: H-01078-2003 Revision 2
1
Flow-Conditioned HUVEC Support Clustered-Leukocyte Adhesion by
Co-expressing ICAM-1 and E-Selectin
Michael P. Burns and Natacha DePaola*
Department of Biomedical Engineering
Rensselaer Polytechnic Institute, Troy, NY 12180.
Revision 2
August 17, 2004
*Address for correspondence:
Natacha DePaola, PhD.
Department of Biomedical Engineering
Rensselaer Polytechnic Institute
110 8th Street, Troy, NY 12180.
Tel: (518) 276-2170
Fax: (518) 276-3035
e-mail: depaola@rpi.edu
Short Title: Flow-Induced Clustered-Leukocyte Adhesion
Copyright © 2004 by the American Physiological Society.
AJP Manuscript ID: H-01078-2003 Revision 2
2
ABSTRACT
Endothelial sequestration of circulating monocytes is a key event in early atherosclerosis.
Hemodynamics is proposed to regulate monocyte-endothelial cell interactions by direct
cell activation and by establishment of flow environments which are conducive or
prohibitive to cell-cell interaction. We investigated fluid shear regulation of monocyteendothelial cell adhesion in vitro using a disturbed laminar shear system that models in
vivo hemodynamics characteristic of lesion-prone vascular regions. Human endothelial
cell monolayers were flow-conditioned for 6 hours prior to evaluating monocyte adhesion
under static and dynamic flow conditions. Results revealed a distinctive, clustered-cell
pattern of monocyte adhesion which strongly resembles in vivo leukocyte adhesion in
early and late stage atherosclerosis. Clustered-monocyte cell adhesion correlated with
endothelial cells co-expressing ICAM-1 and E-selectin, as result of a flow-induced
selective up-regulation of E-selectin expression in a subset of ICAM-1 expressing cells.
Clustered-monocyte cell adhesion assayed under static conditions exhibited a spatial
variation in size and frequency of occurrence demonstrating a differential regulation of
endothelial cell adhesiveness by the local flow environment. Dynamic adhesion studies
conducted with circulating monocytes resulted in clustered-cell adhesion only within the
disturbed flow region, where the monocyte rate of motion is sufficiently low for cell-cell
interaction. These studies provide evidence and reveal mechanisms of local
hemodynamic regulation of endothelial adhesiveness and endothelial-monocyte
interaction leading to localized monocyte adhesion, potentially contributing to the focal
origin of arterial diseases such as atherosclerosis.
AJP Manuscript ID: H-01078-2003 Revision 2
3
Key Words: Leukocyte adhesion, cell clusters, endothelial cells, flow, ICAM-1, ESelectin
AJP Manuscript ID: H-01078-2003 Revision 2
4
INTRODUCTION
A key event in early atherosclerosis is monocyte adhesion and extravasation at
predictable vascular sites [13, 27, 30]. Hemodynamics may contribute to atherogenesis
as lesion-prone vascular regions are associated with low wall shear stresses, temporal and
spatial shear stress gradients and flow recirculation [3, 44]. The fluid dynamics
characteristic of lesion-prone regions may favor monocyte adhesion by direct alteration
of endothelial cell adhesiveness and by establishment of monocyte transport
environments conducive to arterial wall adhesion.
Adhesion molecules mediate endothelial cell sequestration of circulating leukocytes by
physically supporting adhesive interactions and by participating in cell signaling [34, 35].
The endothelial cell adhesion molecules of interest in adhesion with leukocytes are
intercellular adhesion molecule-1 (ICAM-1), vascular cell adhesion molecule-1 (VCAM1), E-selectin, and P-selectin. VCAM-1 has long been proposed to be a principal
endothelial cell adhesion molecule in the selective recruitment of monocytes and
lymphocytes to atherosclerotic lesions [6] since the integrin ligand of VCAM-1, VLA-4,
is expressed by monocytes and lymphocytes but not neutrophils. However, current
knowledge of the leukocyte adhesion cascade suggests the endothelium uses selectinmediated and cytokine-mediated cell signaling pathways in addition to adhesion molecule
pairing to obtain specificity in leukocyte recruitment [33, 43].
AJP Manuscript ID: H-01078-2003 Revision 2
5
There is supporting evidence that the endothelium uses ICAM-1 and selectin coexpression in its selective sequestration of circulating monocytes. Histological
evaluation of human atherosclerosis consistently reports the co-localized expression of
ICAM-1 and E-selectin at the borders of developing lesions [17, 41] and this coexpression has been correlated with recent monocyte infiltrates [11, 17]. Growing
evidence indicates that selectins are also involved in intracellular signaling and potentiate
leukocyte-endothelial cell-cell interaction. Leukocyte adhesion to E-selectin activates
2
integrin-dependent adhesion in leukocytes [19, 33] and triggers the mitogen-activated
protein kinase (MAPK) signaling pathway in endothelial cells [15]. Likewise, ICAM-1
ligation also induces cell signaling in leukocytes and endothelial cells [42]. Interestingly,
neutrophil adhesion to activated endothelial cells has been shown to induce the colocalization of ICAM-1 and E-selectin in the endothelial cell plasma membrane and the
cytoplasmic domains of these focal adhesion complexes is associated with src, a
molecule involved in cell-signaling [38, 39].
Numerous in vitro studies have shown fluid shear stress upregulates human endothelial
cell adhesion molecule expression. Increased ICAM-1 expression has been detected as
early as 6 hours and is sustained for up to 48 hours [26, 28]. Fluid shear modulation of
E-selectin expression remains unclear with increased expression reported for oscillatory
laminar shear stress [5] while unidirectional laminar shear did not alter E-selectin
expression [26, 28]. In the present study, we report that endothelial cell expression of Eselectin is significantly increased in a selective small group of cells after 6 hour exposure
to laminar shear.
AJP Manuscript ID: H-01078-2003 Revision 2
6
Previous in vitro investigations have reported that endothelial adhesiveness for
leukocytes increases after exposure to fluid shear stress [5, 14, 25, 26, 28] due to elevated
expression of ICAM-1 [5, 26, 28] or VCAM-1 [5, 14, 25]. However, the heterogeneous
nature of leukocyte adhesion to flow-conditioned endothelial layers has never before
been addressed. The present study is, to our knowledge, the first detailed characterization
of heterogeneous leukocyte adhesion to flow-conditioned endothelial layers. We
developed a novel method of quantifying leukocyte adhesion which allows the two
patterns of cell adhesion (single and clustered) to be characterized.
Our experimental model [9, 10] permits endothelial layers to be exposed to a welldefined flow field in which a portion of the layer is exposed to disturbed flows, with wall
shear stresses and shear stress gradients similar to lesion-prone regions of the human
carotid sinus, while adjacent areas of the layer experience a flow field characteristic of
lesion-resistant regions. The adhesion assays were performed with U937 cells which
express
1
and
2
integrins, sialylated Lewis-x antigen, and L-selectin [29]. This array of
adhesion molecules allows U937 cells to bind ICAM-1, VCAM-1, E-selectin and Pselectin. We evaluated leukocyte adhesion under both static and flow conditions to
assess the functional adhesiveness of the flow-conditioned endothelial layer. The
findings here reported provide an important contribution to understanding fluid shear
modulation of endothelial adhesiveness for leukocytes and its potential role in the
development of vascular lesions.
AJP Manuscript ID: H-01078-2003 Revision 2
7
MATERIALS AND METHODS
Cell culture
Single-donor HUVEC were obtained from Clonetics Corporation (San Diego, CA) and
maintained with manufacturer’s Endothelial Growth Medium-2® (EGM-2) supplemented
with 10% fetal bovine serum (GIBCO, Green Island, NY). HUVEC were grown in tissue
culture-treated polystyrene flasks (Corning, Corning, NY) coated with 0.1% (vol/vol)
aqueous gelatin solution (Sigma) and kept in an incubator which provided a 37°C ,
humidified, 95% air/5% CO2 atmosphere. Confluent HUVEC layers were subcultured at
a 1:3 ratio, using trypsin 0.025%/EDTA 0.01% (Clonetics Corp.) to detach cells from
their substratum, and were used for experiments up to sixth passage. For shear
experiments, HUVEC were seeded onto glass slides coated with a 0.5% gelatin subbing
solution containing 0.05% potassium chromium sulfate and cell layers were used within
12 hours of reaching confluency. Control monolayers were similarly seeded and
maintained. U937 cells [37] were obtained from American Type Culture Collection
(Rockville, MD) and maintained with RPMI Medium 1640 (GIBCO) supplemented with
10% heat-inactivated FBS (GIBCO), 1 mM MEM sodium pyruvate (GIBCO), and 2.5
mg/ml D-glucose (Fisher Scientific, Pittsburgh, PA). U937 cell suspensions were kept at
a density of 0.5-2.0×106 cells/ml by periodic dilution with fresh growth medium.
Flow system
The parallel-plate flow chamber used to expose HUVEC layers to disturbed laminar shear
has been described previously [10]. Briefly, glass slides with HUVEC monolayers are
assembled into a parallel-plate flow chamber and shear stress is imposed on the HUVEC
AJP Manuscript ID: H-01078-2003 Revision 2
8
by viscous medium flowing through the chamber. A flow disturbance is created by a 0.5
mm thick rectangular step (0.5 mm x 2mm x 25mm) fixed to the glass slide, upstream of
the HUVEC monolayer. The flow rate of the circulating fluid is adjusted so that the shear
stress distribution imposed on the HUVEC layer is similar to the lesion-prone region of
the lateral wall of the internal carotid sinus [44]. Figure 1 illustrates the four distinct
hemodynamic regions that were investigated within the disturbed laminar shear
environment imposed on the HUVEC layer: A) flow recirculation, a region where the
local flow direction is opposite to the bulk flow and wall shear stress ranges between -11
dynes/cm2 and -4 dynes/cm2, B) flow reattachment, a narrow region near flow stagnation
where shear stress gradients are high and wall shear stress ranges between -7 dynes/cm2
and +6 dynes/cm2, C) flow recovery, a region where flow recovers to uniform laminar
shear and wall shear stress ranges between +6 and +10 dynes/cm2, and D) laminar flow,
the region of fully-recovered flow where the wall shear stress is spatially uniform and of
highest magnitude (+11 dynes/cm2). The location of each of the flow zones relative to the
step disturbance, their dimension, and corresponding shear stress are summarized in
Table 1. Fluid is circulated through the chamber by connecting the chamber to a flow
loop consisting of a fluid reservoir, a variable speed peristaltic pump (Cole-Parmer;
Vernon Hills, IL) and a trapped-air flow damper which removes pressure fluctuations
created by peristaltic pumping. The circulating medium is maintained at 37°C in a
humidified, 95% air/5% CO2 atmosphere. After 6 hours flow exposure, glass slides are
removed from the flow chamber and processed for in situ immunocytochemical
evaluation of adhesion molecule expression or used in static adhesion assays with U937
cells. In some studies functional U937 cell adhesion is assessed under flow conditions by
AJP Manuscript ID: H-01078-2003 Revision 2
9
injecting leukocytes into the circulating medium and visually monitoring U937 cell
adhesion to the flow-conditioned endothelial layer. A detailed description of static and
dynamic U937 cell adhesion assays is presented in a subsequent section.
In situ immunocytochemistry
HUVEC surface expression of the adhesion molecules ICAM-1, VCAM-1 and E-selectin
was evaluated by in situ immunocytochemistry using the following antibodies: 1) mouse
monoclonal IgG1 anti-human ICAM-1 (BBA3 from R&D Systems or MAB1379 from
Chemicon), 2) mouse monoclonal IgG1 anti-human VCAM-1 (BBA5 from R&D Systems
or NCL-VCAM from Novocastra Laboratories) and 3) mouse monoclonal IgG1 antihuman E-selectin (CBL-180 or MAB2150 from Chemicon). ICAM-1 binding was
blocked using MAB1379 which recognizes the D1 domain of ICAM-1 and blocks
binding of LFA-1 to ICAM-1. E-selectin binding was blocked using CBL-180 which has
been shown to inhibit neutrophil adhesion to E-selectin when used as F(ab')2 fragments
(manufacturer's information). Simultaneous staining of ICAM-1 and E-selectin
expression was done using a goat polyclonal anti-human ICAM-1 antibody (BBA17 from
R&D Systems) which was visualized using a Texas Red labeled rabbit anti-goat
secondary antibody (Zymed Laboratories).
Immunocytochemistry was performed using two techniques; HUVEC layers were either
fixed in 100% methanol (10 minutes at -20°C) or stained live (unfixed). For both
methanol-fixed and live HUVEC layers the immunostaining protocol was as follows:
HUVEC layers were blocked with PBS containing 3% BSA, incubated with 10 µg/ml of
AJP Manuscript ID: H-01078-2003 Revision 2
10
the respective primary antibody (1 hour at 37°C), incubated with 15 µg/ml biotinylated
rabbit anti-mouse IgG antibody (Zymed Laboratories, San Francisco, CA) for 30 minutes
at room temperature, incubated with 25 µg/ml Alexa 488-conjugated streptavidin
(Molecular Probes, Eugene, OR) for 30 minutes at room temperature, and coverslipped
with Vectashield (Vector Laboratories, Burlingame, CA). Immunoblocking studies were
done by performing the live immunostaining protocol prior to the U937 cell adhesion
assay.
Monocyte cell adhesion assay
U937 cells, a human histiocytic lymphoma cell line [37] which expresses ligands for
ICAM-1, VCAM-1 and E-selectin [29] were used as a model of peripheral blood
monocytes. U937 cells were fluorescently labeled by centrifuging (500 g for 5 minutes)
and incubating the resuspended pellet with 40 µM BCECF-AM in 1% BSA/PBS (pH 7.4)
for 45 minutes at 37°C. Control adhesion assays were performed with both labeled and
unlabeled U937 cells to verify that centrifugation and BCECF-AM incubation did not
affect U937 cell adhesion. For the static U937 cell adhesion assay, flow-conditioned or
control HUVEC layers were briefly rinsed with PBS and then incubated with labeled
U937 cells (5×105 cells/ml) under static conditions for 45 minutes at 37°C. Unbound
U937 cells were removed by gently submerging the glass slide in warm RPMI 1640
medium containing 5% FBS and slowly moving the slide back and forth in the direction
parallel to the long dimension of the step disturbance. The layer was visualized by phasecontrast microscopy to assess the amount of unbound leukocytes remaining on top of the
layer. If necessary, washing was repeated until the unbound leukocytes were removed. In
AJP Manuscript ID: H-01078-2003 Revision 2
11
average, two washes were required to remove leukocytes that were not adhered to the
endothelial layer. Wash removal of leukocytes was also performed by gravity
detachment. For this method of wash removal the glass slide was flipped over and rested
inverted on a rack submerged in RPMI media containing 5% FBS. The layer was left on
the rack for 15 minutes to allow unbound leukocytes to fall away from the layer. Similar
leukocyte adhesion data were obtained with manual washing and the gravity detachment
method. The leukocyte adhesion results reported here were obtained from studies using
the manual washing procedure.
HUVEC layers with bound U937 cells were fixed overnight in 1% glutaralaldehyde/PBS
at 4°C. U937 cell adhesion was quantified by cell counting. Digital images of the
HUVEC layer were systematically acquired using a Cooke Sensicam cooled digital
camera connected to the side port of an Olympus IX70 fluorescence microscope with the
movement of the microscope stage controlled by an H107 motorized stepper stage (Prior
Scientific, Rockland, MA). Series of digital images were taken along lines parallel to the
direction of flow, with the first image of each series being located immediately
downstream of the step (flow disturbance) and each successive image located
downstream of the preceding image. For each sample, 4 to 6 series of images were taken,
with each series containing 20 images, resulting in an average of 100 images per sample.
In this way a continuous documentation of U937 cell adhesion to the HUVEC layer was
obtained and the spatial location of each image was known. For each digital image, the
following information was recorded: a) the number of U937 cells which adhered as single
cells, b) the number of U937 cells involved in each clustered cell adhesion, and c) the
distance of the image from the downstream edge of the step. Total number of U937 cell
AJP Manuscript ID: H-01078-2003 Revision 2
12
adhesions was determined by adding up single and clustered-cell adhesions. A cluster of
U937 cells was defined to be any group of 4 or more cells which attached to the HUVEC
layer within 10 µm of another U937 cell.
For the dynamic U937 cell adhesion assay, the HUVEC layers were
preconditioned for 6 hours, as previously described, prior to the incorporation of
monocytes to the circulating medium. Standard cell culture medium (viscosity 3 times
lower than blood viscosity) was used as perfusion fluid in the preconditioned phase of the
studies to recreate physiological shear stresses on the endothelial monolayer.
Once the monocytes were added to the circulating medium, the perfusion rate was
adjusted (reduced) to achieve a physiological near wall velocity (10 µm from wall) of 0.3
cm/s [44] for the study of monocyte-endothelial cell-cell interactions. As the flow rate
was reduced from the preconditioning value, the recirculation region became smaller (one
third the length of the vortex during preconditioning) redefining the dynamic conditions
within the four flow zones of interest. Only the recirculation and fully recovered laminar
flow zones can be accurately recreated, with the current experimental model, to evaluate
monocyte adhesion under physiologically relevant dynamic conditions. In the
recirculation region, monocytes circulating within the smaller vortex (recirculation region)
interact with endothelial cells that have also been preconditioned under flow recirculation.
In the fully recovered flow region, monocytes circulating within unidirectional laminar
flow interact with endothelial cells also preconditioned with unidirectional laminar flow
at a physiological wall shear stress. In contrast, in both the flow recovery and the flow
reattachment regions, circulating monocytes interact with endothelial cells
preconditioned under flow recirculation, since the reduced flow rate positions these zones
AJP Manuscript ID: H-01078-2003 Revision 2
13
within the preconditioned recirculation region. Therefore, for the analysis of dynamic
U937 cell adhesion the endothelial layers were divided into two regions only; "the region
of flow disturbance" including recirculation, reattachment and recovery, and the
“recovered laminar shear region" located far downstream from the flow disturbance.
The concentration of U937 cells in the circulating medium was 106 cells/ml. U937 cells
were allowed to circulate over the flow-conditioned layer for 45 minutes while
continuously visualized by phase-contrast microscopy. Digital images of the endothelial
layer were periodically taken with the time and location of each image recorded.
Dynamic adhesion assays were concluded by gently flushing the flow chamber with
monocyte-free medium and processing the endothelial layers for in situ
immunocytochemistry as described above.
Statistical analysis
U937 cell static adhesion data was obtained from 5 independent shear experiments. In
some experiments, multiple endothelial layers were simultaneously exposed to shear. A
total of 14 shear samples and 10 control samples were analyzed and reported in this study.
Significant differences in cell adhesion were determined using Tukey’s multiple
comparison test ( =0.05) to compare respective adhesion types between various flow
regions. Static control samples (containing a flow-disturbance step but not exposed to
flow) were similarly evaluated and found to have no spatial variation in U937 cell
adhesion, thus, static control adhesion is represented by a single value in the analysis.
AJP Manuscript ID: H-01078-2003 Revision 2
14
RESULTS
Flow-induced clustered-leukocyte cell adhesion
U937 cells adhere to all regions of the flow-conditioned HUVEC monolayer when cellcell interaction is allowed to occur under static conditions (Figure 1). As can be seen in
Figure 1, U937 cells adhere to flow-conditioned HUVEC in two distinct patterns; as
single U937 cells or as a cluster of U937 cells. U937 cell adhesion to control endothelial
cell layers (Figure 1E) appears similar to the region of the layer exposed to flow
reattachment (Figure 1B) and is characterized by a low number of U937 cell adhesions,
with most adhesions occurring as single U937 cells. The pattern of U937 cell adhesion in
the region of the layer exposed to recovered laminar shear (Figure 1D) shows the most
difference when compare to control layers. Adhesion in the region of recovered laminar
shear is characterized by a large number of adherent cells, with clustered-U937 cell
adhesions occurring more frequently and involving more U937 cells compared to any
other region in the monolayer. These findings indicate that flow-conditioning increases
the adhesiveness of the endothelial layer. Furthermore, this fluid-shear induced response
appears to be sensitive to the local characteristics of the preconditioning flow
environment.
Fluid shear alteration of endothelial cell adhesiveness was evaluated by quantifying static
U937 cell adhesion for each region of the flow-conditioned HUVEC layer (Figure 2).
Significant differences in adhesion were evaluated by comparing respective adhesion
types of various regions using Tukey’s multiple comparison test ( =0.05). ClusteredU937 cell adhesion is significantly higher in the region of laminar shear compared to all
AJP Manuscript ID: H-01078-2003 Revision 2
15
other regions except the region of flow recirculation. Single U937 cell adhesion is
consistently higher in flow-conditioned layers when compared to control, but not
significantly so, and does not appear to be influenced by the pre-conditioning
environment. These data suggest that the observed fluid shear alteration of HUVEC
adhesiveness has two components; one component is shear-magnitude dependent and
supports clustered-cell adhesion while the other component is a general response of the
layer to flow and supports single cell adhesion.
U937 cell clusters vary in the number of cells participating in the cluster formation.
Figure 3 shows the distribution of cluster sizes in flow-conditioned and control HUVEC
layers. This plot was used to classify the U937 cell clusters in three sizes defined as
follows: small (4-6 U937 cells), medium (7-10 U937 cells), and large (> 10 U937 cells).
The spatial distribution of each cluster size is shown for flow-conditioned and control
HUVEC layers in Figure 4. The region of laminar shear has a significantly higher number
of medium-sized U937 cell clusters compared to all other regions, and also has more
large-sized U937 cell clusters compared to flow-reattachment and static control (Figure
4A). U937 cell adhesion in the fully recovered laminar shear region of layers
conditioned with disturbed-flow is indistinguishable from that of HUVEC layers
conditioned with uniform laminar shear (no step disturbance) of same magnitude (data
not shown).
The occurrence of small clusters is consistently higher in flow-conditioned layers
compared to control, but not significantly so, and it does not appear to be significantly
influenced by the pre-conditioning environment (Figure 4B). This finding suggests that
AJP Manuscript ID: H-01078-2003 Revision 2
16
small clusters are not indicative of the same alteration in endothelial cell adhesiveness
that supports medium and large cluster formation. Thus, from now on, we refer to cluster
formations to designate only clusters involving more than six U937 cells.
Fluid shear selectively induces E-selectin expression
Unfixed endothelial layers were immunostained for ICAM-1, VCAM-1 and E-selectin to
investigate which adhesion molecules modulate the observed flow-induced leukocyte
adhesion. While E-selectin is not detected in control layers, flow-conditioned HUVEC
monolayers reveal a significant up-regulation of E-selectin expression (Figure 5A, B).
ICAM-1 expression is present in both control and flow-conditioned layers and the level
of expression appears unchanged after 6 hours of flow-conditioning (Figure 5C, D).
VCAM-1 expression was not detected in control or flow-conditioned HUVEC layers
(Figure 5E, F) but could be induced by TNF- stimulation (10 ng/ml, 6 hours) indicating
that the HUVEC used in these experiments are capable of expressing VCAM-1 but do not
do so in response to the flow stimulus used in this study. Simultaneous staining of Eselectin and ICAM-1 (Figure 6) shows that flow-induced E-selectin expression always
co-localizes with existing ICAM-1 expression, but not all ICAM-1 expressing cells
express E-selectin. Thus, flow selectively upregulates E-selectin expression in a subset
of endothelial cells already expressing noticeable levels of ICAM-1. The number or
endothelial cells that upregulate their E-selectin expression is very small, about 5 % of
the population, and no statistically significant spatial variation in E-selectin expression
was found when the number of immunoreactive cells was quantified in situ by visual cell
counting (data not shown). E-selectin expressing endothelial cells were observed in each
AJP Manuscript ID: H-01078-2003 Revision 2
17
of the four flow regions and were most numerous in the laminar shear region, however
the quantification data did not support any statistically significant difference in the
number of E-selectin expressing cells between any of the various flow regions.
Endothelial cell adhesion molecule expression in HUVEC layers conditioned with 11
dynes/cm2 uniform laminar shear (no step disturbance) was similar to the fully recovered
laminar shear region of layers conditioned with disturbed-flow.
U937 cell clusters adhere to HUVEC co-expressing ICAM-1 and E-selectin
Immunoblocking ICAM-1 or E-selectin alone does not inhibit cluster formation but
indirectly indicates that U937 cell clusters associate with HUVEC co-expressing ICAM-1
and E-selectin (Figure 7A, B). Simultaneously immunoblocking flow-conditioned layers
for both ICAM-1 and E-selectin does inhibit clustered-cell adhesion (Figure 7C).
Furthermore, when ICAM-1 alone is blocked immunoreactivity is sometimes
unassociated with cluster formations (Figure 7A) while E-selectin immunoreactivity is
always associated with cluster formation (Figure 7B). This agrees with our
immunocytochemical findings that flow-induced E-selectin upregulation occurs only in a
subset of ICAM-1 expressing cells and that only co-expressing cells support cluster cell
adhesion. Dual staining HUVEC layers for ICAM-1/E-selectin after performing U937
cell adhesion assays (under both static and dynamic conditions) directly indicates that
cluster adhesions are associated exclusively with endothelial cells co-expressing ICAM-1
and E-selectin (Figure 8). In this way U937 cells can be viewed as reporters of
endothelial adhesion molecule expression, in particular clustered U937 cell adhesion is
indicative of endothelial ICAM-1/E-selectin co-expression.
AJP Manuscript ID: H-01078-2003 Revision 2
18
Functional U937 cell adhesion under dynamic conditions
Under dynamic conditions, functional U937 cell adhesion to flow-conditioned endothelial
layers is limited to the disturbed flow region (Figure 9A). Circulating monocytes adhere
within the region of flow recirculation and the vicinity of flow reattachment in the same
clustered-cell pattern observed in static adhesion assays. The laminar shear region does
not support U937 cell adhesion under flow conditions (Figure 9B). Adhesion in areas of
lowest but not highest shear supports the idea of a near wall velocity threshold for cellcell interaction. Control layers, which have a step disturbance but were not
preconditioned with flow, do not support leukocyte adhesion under dynamic conditions in
any region of the layer. These results demonstrate that arrest and adhesion of circulating
monocytes is determined by shear stress regulation of endothelial adhesiveness and by
the local flow dynamics regulating monocyte transport at the interface with the
endothelial layer.
AJP Manuscript ID: H-01078-2003 Revision 2
19
DISCUSSION
The present study is the first report of a detailed characterization of flow-induced
endothelial cell adhesiveness for leukocytes demonstrating that shear stress selectively
induces the co-expression of ICAM-1 and E-selectin in a limited number of endothelial
cells supporting clustered-leukocyte adhesion. This cluster pattern of adhesion is not
uniformly distributed in the monolayer as shear stress up-regulates E-selectin expression
only in a subset of ICAM-1 expressing cells. Flow mediated E-selectin expression is
never observed in the absence of ICAM-1. These findings support the idea of a naturally
heterogeneous endothelium in which individual cells, or small group of cells, selectively
respond to the local flow environment, potentially contributing to the focal origin of
arterial diseases such as atherosclerosis.
Clustered-leukocyte adhesion has been reported in vivo, in humans with advanced
atherosclerosis and in hypercholesterolemic animal models [7, 30], but never before in
vitro. The leukocyte adhesion patterns in those in vivo studies are remarkably similar to
those observed in our in vitro flow-preconditioned HUVEC layers, suggesting that the
mechanisms of adhesion may also be similar. The observed clustered-leukocyte adhesion
is not a consequence of leukocyte aggregation but is clearly a localized adhesion of
individual leukocytes to an adhesive cell or group of cells in the endothelial layer. As
mentioned above, our results demonstrate that clustered-leukocyte adhesion is supported
by a flow–mediated, localized co-expression of the adhesion molecules ICAM-1 and Eselectin.
AJP Manuscript ID: H-01078-2003 Revision 2
20
Co-localized selectin and ICAM-1 expression has interesting implications with respect to
the leukocyte adhesion cascade. It is well recognized that selectins mediate leukocyte
capture and rolling adhesion while ICAM-1 is associated with monocyte arrest and firm
adhesion [4, 20, 23, 24, 34, 35]. Relevant synergistic effects of selectin-ICAM-1 have
been reported in animal models demonstrating that ICAM-1 expression optimizes
selectin-mediated rolling resulting in lower leukocyte rolling velocity [36]. Selectinmediated rolling has also been reported to activate leukocyte
2 integrins
to bind ICAM-1
in vitro [20, 33] and most recently it has been shown that E-selectin and ICAM-1 cluster
in endothelial plasma membrane lipid rafts following leukocyte adhesion to these
molecules [38, 39]. Furthermore, the cytoplasmic regions of these selectin-ICAM-1
adhesion complexes are associated with cortactin and src suggesting the potential
involvement in cell-signaling mechanisms.
To our knowledge, the present study is the first in vitro evaluation of in situ ICAM-1 and
E-selectin co-expression in flow-conditioned endothelial cells. The
immunocytochemistry presented here is a detailed in situ investigation of fluid shear
effect on endothelial adhesion molecule expression which allows for the identification of
localized small amounts of protein expression heterogeneously distributed on the
monolayer. Previous in vitro work investigating flow-induced changes in endothelial
adhesion molecule expression removed cells from the monolayer and used flow
cytometry to quantify protein expression [5, 14, 26, 28]. Results from those studies were
often reported in the form of percent changes relative to baseline expression, without
AJP Manuscript ID: H-01078-2003 Revision 2
21
accounting for any heterogeneity in the monolayer response. The detailed in situ
immunocytochemistry presented in this study revealed that in response to flow select
ICAM-1 expressing cells up-regulate their E-selectin expression. However, the number
of cells which up-regulate E-selectin expression in response to flow is small (<5%) which
may explain the difficulties in detecting it with flow-cytometry as a minimal shift in the
mean fluorescence would be expected, and may pass unnoticeable. In this sense, flowcytometry is an effective and sensitive technique to detect homogeneous cell responses in
protein expression but less able to detect heterogeneous ones.
The observed flow-mediated clustered-leukocyte adhesion and underlying cellular
mechanisms potentially involved in the synergistic effects of selectin-ICAM-1 associated
with this phenomenon are not simple to explore nor to explain. Our results demonstrate
that only cells which co-express ICAM-1 and E-selectin support leukocyte adhesion in
the form of clusters and that simultaneous immunoblocking of ICAM-1 and E-selectin is
required to inhibit clustered-cell adhesion in flow-conditioned monolayers. Interestingly,
immunoblocking of E-selectin or ICAM-1 in the ICAM-1/E-selectin co-expressing cells
is not sufficient to inhibit clustered-leukocyte adhesion yet flow-conditioned cells which
express ICAM-1 alone are not associated with cluster formations. Both clusteredleukocyte cell adhesion and endothelial co-expression of ICAM-1 and E-selectin are
absent in control (no flow) HUVEC layers. In fact, no detectable amounts of E-selectin
expression were found by immunocytochemistry in controls layers.
AJP Manuscript ID: H-01078-2003 Revision 2
22
In interpreting the results of our immunoblocking studies we suggest that blocking both
ICAM-1 and E-selectin removes both ligands which leukocytes can use to bind our flowconditioned endothelial cells and thus adhesion is inhibited. Our finding that blocking of
ICAM-1 alone does not inhibit clustered-leukocyte adhesion can be explained by the
ability of leukocytes to bind flow-conditioned endothelial cells using an E-selectin ligand.
On the other hand, our finding that immunoblocking E-selectin alone does not inhibit
clustered-leukocyte adhesion is a difficult scenario to explain as endothelial cells which
express only ICAM-1 (in both flow-conditioned and static control layers) are not
associated with clustered-cell adhesion. One possible explanation for this phenomenon is
that flow-conditioning also selectively induces chemokine expression in these same cells
which express ICAM-1 and E-selectin offering a pathway for activating leukocyte
2
integrins which are then able to bind ICAM-1. In this scenario, blocking E-selectin alone
in flow-conditioned endothelial layers may only remove one leukocyte ligand while
leukocyte
2
integrins could still become activated by endothelial secreted chemokine
and then bind endothelial cells by ICAM-1. As leukocyte
2
integrins require activation
in order to bind ICAM-1, endothelial cells which express ICAM-1 but not E-selectin
would presumably not be secreting chemokine and thus would not support leukocyte
adhesion. This argument once again supports the idea of a heterogeneous endothelium in
which individual cells or small group of cells selectively respond to shear stress.
This proposed explanation that flow-mediated chemokine secretion activates leukocyte
2
integrins is supported by earlier studies demonstrating fluid shear modulation of HUVEC
AJP Manuscript ID: H-01078-2003 Revision 2
23
mRNA levels and protein expression of both IL-8 [18] and MCP-1 [31]. While IL-8 was
suppressed by laminar shear (7 dynes/cm2), MCP-1 demonstrated a biphasic response,
with mRNA levels initially increasing in response to laminar shear before returning to
basal levels after 4 hours of flow-conditioning. The evaluative techniques used in those
studies (ELISA, Northern blot) provide no information about potential heterogeneity of
the flow-induced endothelial cell response.
Another potential mechanism that may offer pathways for activating the leukocyte
2
integrins is flow-mediated E-selectin signaling. It has been demonstrated that incubation
of leukocytes with E-selectin increases the affinity of
2
integrins for ICAM-1 [33]. In
our static adhesion assay, monocytes are in contact to endothelial cells expressing Eselectin which may potentially activate the
2
integrins contributing to the observed
localized clustered-adhesion. However, as indicated by our immunoblocking findings, Eselectin-mediated cell signaling cannot be the only pathway for activating leukocyte
2
integrins. Adhesion studies with immobilized ICAM-1 substrates have demonstrated that
leukocyte
2
integrins can be activated by prolonged contact with ICAM-1 [32]. This is
unlikely to be the mechanism in the clustered adhesion we report here since it would not
explain the lack of clustered-cell adhesion in cells expressing only ICAM-1 in both flowconditioned and static control layers.
We have linked ICAM-1 and E-selectin co-expression to our observed flow-induced
leukocyte adhesion, while VCAM-1 expression was not detected using either
immunocytochemistry or Western blot (data not shown). VCAM-1 involvement in
AJP Manuscript ID: H-01078-2003 Revision 2
24
leukocyte adhesion is well-accepted. In vivo, VCAM-1 expression has been reported in
early atherogenesis [6, 8, 11, 16]. In vitro, although flow-modulation of VCAM-1
expression has been extensively investigated, the findings are not in full agreement [2, 5,
14, 25, 28, 40]. In the present study, exposure of endothelial layers to flow fields
representative of lesion-prone hemodynamics did not induce VCAM-1 expression.
VCAM-1 expression could be induced by cytokine stimulation (TNF- ), but flowregulation of adhesion molecule expression was found to be limited to E-selectin.
Fluid dynamic modulation of leukocyte-endothelial cell-cell interaction has been
demonstrated in this study by direct alteration of endothelial adhesiveness resulting in
clustered-leukocyte adhesions similar to those reported in association with early and late
stage atherosclerosis. The data revealed that fluid shear alteration of HUVEC
adhesiveness is shear magnitude dependent, demonstrating that leukocyte adhesion
patterns are influenced by characteristics of the local fluid dynamic environment imposed
on the endothelial layer during flow-conditioning. We also demonstrated that
hemodynamics regulate monocyte-endothelial cell interactions by establishing flow
environments which may be conducive or prohibitive to cell-cell interaction. In the
disturbed flow model investigated, monocytes circulating over flow-conditioned
endothelial layers adhered only within the recirculation region and in the vicinity of flow
reattachment. Monocytes in the flow recirculation region benefit from an increased
residence time for interaction with the endothelial layer. In the vicinity of flow
reattachment, the flow streamlines are directed almost perpendicular to the endothelial
layer surface enhancing monocyte transport to that region. In both recirculation region
AJP Manuscript ID: H-01078-2003 Revision 2
25
and vicinity of flow reattachment, the near wall velocity (10 µm from wall), which is
representative of the speed of motion of the circulating monocytes near the endothelial
layer, is lower than in the downstream undisturbed flow region. The adhesion of
circulating monocytes was not supported in the laminar shear region in spite of the high
adhesiveness of the underlying endothelial layer. This is the region of highest wall shear
stress and near wall velocity in the experimental model which, although representative of
physiological values, it appears to be beyond the near wall velocity threshold for cell-cell
interaction leading to adhesion. Here our dynamic adhesion results are consistent with
earlier studies reporting that leukocyte adhesion through selectins is sensitive to
hydrodynamic shear levels [12, 21].
The cluster-cell pattern of monocyte adhesion reported in this study does not appear to
involve L-selectin-mediated secondary (leukocyte-leukocyte) capture mechanisms. This
is evidenced by the fact that the clusters we report can form under static conditions where
L-selectin-mediated secondary capture is not expected to occur since L-selectin bond
formation requires a threshold hydrodynamic shear [12, 21]. Furthermore, under flow
conditions we recognize cluster formation mechanisms that are distinct from those
reported for L-selectin-mediated secondary capture [1, 22]. We observe cluster
formation that does not always involve leukocyte-leukocyte contact. Rather, leukocyte
accumulation appears to be associated with particular endothelial cells as demonstrated
by the fact that often leukocyte detachment from an endothelial cell is followed by the
arrival and arrest of another leukocyte on the same endothelial cell. This mechanism of
adhesion is distinct from L-selectin-mediated secondary capture which has been reported
AJP Manuscript ID: H-01078-2003 Revision 2
26
to involve an initial homotypic leukocyte contact followed by a rolling transfer of the
leukocyte to the substrate [1, 22].
The studies here reported provide an important contribution to understanding fluid shear
modulation of endothelial adhesiveness and endothelial-monocyte interaction leading to
localized monocyte adhesion. Our in vitro findings, paralleling in vivo observations,
demonstrate the value of a model system that may significantly aid in unraveling the
mechanisms underlying the focal origin of arterial diseases and early atherosclerotic
lesion development.
AJP Manuscript ID: H-01078-2003 Revision 2
27
ACKNOWLEDGEMENTS
This work was supported by National Science Foundation Grant NSF9624991 CAREER
and National Institutes of Health Grant HL64728 (to NDP). We are grateful to Michael A.
Gimbrone, Jr. of the Brigham and Women’s Hospital (Boston, MA) for discussion of the
manuscript.
AJP Manuscript ID: H-01078-2003 Revision 2
28
REFERENCES
1. Alon, R., R.C. Fuhlbrigge, E.B. Finger, T.A. Springer. Interactions through Lselectin between leukocytes and adherent leukocytes nucleate rolling adhesions on
selectins and VCAM-1 in shear flow. J. Cell Biol. 135:849-865, 1996.
2. Ando J, Tsuboi H, Korenaga R, Takada Y, Toyama-Sorimachi N, Miyasaka M,
Kamiya A. Shear stress inhibits adhesion of cultured mouse endothelial cells to
lymphocytes by downregulating VCAM-1 expression. Am J Physiol 267:C679-C687,
1994.
3. Asakura T, Karino T. Flow patterns and spatial distribution of atherosclerotic lesions
in human coronary arteries. Circ Res 66:1045-1066, 1990.
4. Butcher EC. Leukocyte-endothelial cell recognition: three (or more) steps to
specificity and diversity. Cell 67:1033-1036, 1991.
5. Chappell DC, Varner SE, Nerem RM, Medford RM, Alexander RW. Oscillatory
shear stress stimulates adhesion molecule expression in cultured human endothelium.
Circ Res 82:532-539, 1998.
6. Cybulsky MI, Gimbrone MA Jr. Endothelial expression of a mononuclear leukocyte
adhesion molecule during atherogenesis. Science 251:788-791, 1991.
AJP Manuscript ID: H-01078-2003 Revision 2
29
7. Davies MJ, Woolf N, Rowles PM, Pepper J. Morphology of the endothelium over
atherosclerotic plaques in human coronary arteries. Br Heart J 60:459-64, 1988.
8. Davies MJ, Gordon JL, Gearing AJH, Pigott R, Woolf N, Katz D, Kyriakopoulos
A. The expression of the adhesion molecules ICAM-1, VCAM-1, PECAM, and Eselectin in human atherosclerosis. J Pathol 171:223-229, 1993.
9. DePaola N, Gimbrone MA Jr, Davies PF, Dewey CF Jr. Vascular endothelium
responds to fluid shear stress gradients. Arteriosclerosis and Thrombosis 12:1254-1257,
and 13:465, 1992.
10. DePaola N, Davies PF, Pritchard WF, Florez L, Harbeck N, Polacek DC. Spatial
and temporal regulation of gap junction connexin43 in vascular endothelial cells exposed
to controlled disturbed flows in vitro. Proc Natl Acad Sci USA 96:3154-3159, 1999.
11. Duplaa C, Couffinhal T, Labat L, Moreau C, Petit-Jean M, Doutre M,
Lamaziere JD, Bonnet J. Monocyte/macrophage recruitment and expression of
endothelial adhesion proteins in human atherosclerotic lesions. Atherosclerosis 121:253266, 1996.
AJP Manuscript ID: H-01078-2003 Revision 2
30
12. Finger EB, Puri KD, Alon R, Lawrence MB, von Andrian UH, Springer TA.
Adhesion through L-selectin requires a threshold hydrodynamic shear. Nature 379:266269, 1996.
13. Gerrity RG. The role of the monocyte in atherogenesis: I. Transition of blood-borne
monocytes into foam cells in fatty lesions. Am J Pathol 103:181-190, 1981.
14. Gonzales RS and Wick TM. Hemodynamic modulation of monocytic cell adherence
to vascular endothelium. Annals of Biomedical Engineering 24:382-393, 1996.
15. Hu Y, Kiely JM, Szente BE, Rosenzweig A, Gimbrone MA Jr. E-selectindependent signaling via the mitogen-activated protein kinase pathway in vascular
endothelial cells. J Immunol 165:2142-2148, 2000.
16. Iiyama K, Hajra L, Iiyama M, Li H, DiChiara M, Medoff BD, Cybulsky MI.
Patterns of vascular cell adhesion molecule-1 and intercellular adhesion molecule-1
expression in rabbit and mouse atherosclerotic lesions and at sites predisposed to lesion
formation. Circ Res 85:199-207, 1999.
17. Johnson-Tidey RR, McGregor JL, Taylor PR, Poston RN. Increase in the
adhesion molecule P-selectin in endothelium overlying atherosclerotic plaques;
coexpression with intercellular adhesion molecule-1. Am J Pathol. 144:952-961, 1994.
AJP Manuscript ID: H-01078-2003 Revision 2
31
18. Kato H, Uchimura I, Nawa C, Kawakami A, Numano F. Fluid shear stress
suppresses interleukin 8 production by vascular endothelial cells. Biorheology 38:347353, 2001.
19. Kunkel EJ, Dunn JL, Ley K. Leukocyte arrest during cytokine-dependent
inflammation in vivo. J Immunol 164:3301-3308, 2000.
20. Lawrence MB, Springer TA. Leukocytes roll on a selectin at physiologic flow rates:
distinction from and prerequisite for adhesion through integrins. Cell 65:859-873, 1991.
21. Lawrence MB, Kansas GS, Kunkel EJ, Ley K. Threshold levels of fluid shear
promote leukocyte adhesion through selectins (CD62L,P,E). J Cell Biol 136:717-727,
1997.
22. Lim, Y.-C., K. Snapp, G.S. Kansas, R. Camphausen, H. Ding, F.W. Luscinskas.
Important contributions of P-selectin glycoprotein ligand-1-mediated secondary capture
to human monocyte adhesion under flow in vitro. J. Immunol. 161:2501-2508, 1998.
23. Luscinskas FW, Kansas GS, Ding H, Pizcueta P, Schleiffenbaum BE, Tedder TF,
Gimbrone MA Jr. Monocyte rolling, arrest and spreading on IL-4-activated vascular
endothelium under flow is mediated via sequential action of L-selectin,
2-integrins.
J Cell Biol 125:1417-1427, 1994.
1-integrins,
and
AJP Manuscript ID: H-01078-2003 Revision 2
32
24. Luscinskas FW, Ding H, Tan P, Cumming D, Tedder TF, Gerritsen ME. L- and
P-selectins, but not CD49d (VLA-4) integrins, mediate monocyte initial attachment to
TNF-alpha-activated vascular endothelium under flow in vitro. J Immunol 156:326-335,
1996.
25. Mohan S, Mohan N, Valente AJ, Sprague EA. Regulation of low shear flowinduced HAEC VCAM-1 expression and monocyte adhesion. Cell Physiol 45:C1100C1107, 1999.
26. Morigi M, Zoja C, Figliuzzi M, Foppolo M, Micheletti G, Bontempelli M,
Saronni M, Remuzzi G, Remuzzi A. Fluid shear stress modulates surface expression of
adhesion molecules by endothelial cells. Blood 85:1696-1703, 1995.
27. Munro JM, van der Walt JD, Munro CS, Chalmers JA, Cox EL. An
immunohistochemical analysis of human aortic fatty streaks. Hum Pathol 18:375-380,
1987.
28. Nagel T, Resnick N, Atkinson WJ, Dewey CF, Gimbrone MA Jr. Shear stress
selectively upregulates intercellular adhesion molecule-1 expression in cultured human
vascular endothelial cells. J Clin Invest 94:885-891, 1994.
AJP Manuscript ID: H-01078-2003 Revision 2
33
29. Prieto J, Eklund A, Patarroyo M. Regulated expression of integrins and other
adhesion molecules during differentiation of monocytes into macrophages. Cellular
Immunology 156:191-211, 1994.
30. Ross R. The pathogenesis of atherosclerosis: a perspective for the 1990s. Nature
362:801-809, 1993.
31. Shyy YJ, Hsieh HJ, Usami S, Chien S. Fluid shear stress induces a biphasic
response of human monocyte chemotactic protein 1 gene expression in vascular
endothelium. Proc Natl Acad Sci USA 91:4678-4682, 1994.
32. Sigal A, Bleijs DA, Grabovsky V, van Vliet SJ, Dwir O, Figdor GG, van Kooyk,
Alon R. The LFA-1 integrin supports rolling adhesions on ICAM-1 under physiological
shear flow in a permissive cellular environment. J Immunol 165:442-452, 2000.
33. Simon SI, Hu Y, Vestweber D, Smith CW. Neutrophil tethering on E-selectin
activates B2 integrin binding to ICAM-1 through a mitogen-activated protein kinase
signal transduction pathway. J Immunol 164:4348-4358, 2000.
34. Smith CW. Possible steps involved in the transition to stationary adhesion of rolling
neutrophils: a brief review. Microcirculation 7:385-394, 2000.
AJP Manuscript ID: H-01078-2003 Revision 2
34
35. Springer TA. Traffic signals for lymphocyte recirculation and leukocyte emigration:
the multistep paradigm. Cell 76:301-314, 1994.
36. Steeber DA, Campbell MA, Basit A, Ley K, Tedder TF. Optimal selectin-mediated
rolling of leukocytes during inflammation in vivo requires intercellular adhesion
molecule-1 expression. Proc Natl Acad Sci USA 95:7562-7567, 1998.
37. Sundstrom C and Nilsson K. Establishment and characterization of a human
histiocytic lymphoma cell line (U-937). Int J Cancer 17:565-577, 1976.
38. Tilghman RW, Hoover RL. E-selectin and ICAM-1 are incorporated into detergentinsoluble membrane domains following clustering in endothelial cells. FEBS Letters
525:83-87, 2002.
39. Tilghman RW, Hoover RL. The Src-cortactin pathway is required for clustering of
E-selectin and ICAM-1 in endothelial cells. FASEB J 16:1257-1259, 2002.
40. Tsao PS, Lewis NP, Alpert S, Cooke JP. Exposure to shear stress alters endothelial
adhesiveness: role of nitric oxide. Circulation 92:3513-3519, 1995.
41. van der Wal AC, Das PK, Tigges AJ, Becker AE. Adhesion molecules on the
endothelium and mononuclear cells in human atherosclerotic lesions. Am J Pathol
141:1427-1433, 1992.
AJP Manuscript ID: H-01078-2003 Revision 2
35
42. Wang Q, Doerschuk CM. The signaling pathways induced by neutrophil-endothelial
cell adhesion. Antioxid Redox Signal 4:39-47, 2002.
43. Weber C. Novel mechanistic concepts for the control of leukocyte transmigration:
specialization of integrins, chemokines, and junctional molecules. J Mol Med 81:4-19,
2003.
44. Zarins CK, Giddens DP, Bharadvaj BK, Sottiurai VS, Mabon RF, Glagov S.
Carotid bifurcation atherosclerosis: quantitative correlation of plaque localization with
flow velocity profiles and wall shear stress. Circ Res 53:502-514, 1983.
AJP Manuscript ID: H-01078-2003 Revision 2
36
FIGURE LEGENDS
Figure 1. Static U937 cell adhesion. Flow-conditioned HUVEC (6 hours) were
incubated with U937 cells under static conditions (45 minutes). Adherent fluorescently
labeled U937 cells show as white circles while underlying endothelial cells appear gray.
Micrographs (A-D) correspond to respective flow regions indicated in the schematic
diagram; (A) flow recirculation, (B) flow reattachment, (C) flow recovery, and (D)
recovered laminar shear. (E) Control HUVEC layer not exposed to flow. Magnification
200x. Scale bar indicates 50 µm.
Figure 2. Spatial distribution of static U937 cell adhesion. Number of adhered U937
cells per square centimeter of HUVEC layer is shown for the flow regions marked in Fig.
1 (schematic flow diagram). Total cell adhesion (
adhesion (
bars) and clustered cell adhesion (
bars) is the sum of single cell
bars). Bar plot values are mean +/-
S.E.M. (n=5). Significant differences were evaluated using Tukey’s multiple comparison
test ( =0.05). (*) indicates significant differences compared to static control, (+)
indicates significant differences compared to both reattachment and recovery regions.
Figure 3. Frequency of occurrence for each cluster size. Cluster adhesions formed under
static conditions were categorized according to the number of U937 cells involved in the
cell cluster. The number of clusters of each size are shown as closed squares ( ) for
flow-conditioned monolayers and as open circles ( ) for static control monolayers.
AJP Manuscript ID: H-01078-2003 Revision 2
37
Clusters were classified in three groups for further analysis; small (4-6 U937 cells),
medium (7-10 U937 cells) and large (>10 U937 cells).
Figure 4. Spatial distribution of cluster sizes on flow-conditioned HUVEC layer. (A)
number of medium clusters (
bars) and large clusters (
bars) formed under static
conditions per square centimeter area of HUVEC layer. Using Tukey’s multiple
comparison test ( =0.05), (*) indicates significant differences compare to static control
and all other flow regions, (+) indicates significant differences compare to static control
and flow reattachment. (B) number of small clusters (
bars) adhered per square
centimeter area of HUVEC layer. Bar plots show mean values +/- S.E.M (n=5).
Figure 5. Endothelial cell surface expression of adhesion molecules. HUVEC layers
maintained under static culture conditions (A, C, E) or preconditioned with disturbed
flow for 6 hours (B, D, F) were evaluated for E-selectin (A, B), ICAM-1 (C, D) or
VCAM-1 (E, F) expression using in situ immunocytochemistry. Immunoreactivity
indicates adhesion molecule surface expression as cells were not fixed prior to
immunostaining. Photomicrographs B, D, and F are from the fully recovered laminar
shear region.
Magnification 200x. Scale bar indicates 50 µm.
Figure 6. Flow-induced E-selectin expression co-localizes with ICAM-1 expression.
Flow-conditioned HUVEC (6 hours) were fixed in methanol and stained for ICAM-1 (red)
and E-selectin (green). Phase-contrast (A), and corresponding fluorescent micrographs
are shown (B). Red and green fluorescence are superimposed with cells co-expressing
AJP Manuscript ID: H-01078-2003 Revision 2
38
ICAM-1 and E-selectin appearing yellow. Magnification 100x. Scale bar indicates 100
µm.
Figure 7. Simultaneous ICAM-1 and E-selectin immunoblocking inhibits cluster
adhesion. Prior to performing a static U937 cell adhesion assay, flow-conditioned
HUVEC layers were immunoblocked for either (A) ICAM-1 alone, (B) E-selectin alone,
or (C) both ICAM-1 and E-selectin simultaneously. U937 cell clusters associate with
immunoreactivity when either ICAM-1 or E-selectin is blocked independently (A and B).
Clusters do not form when both ICAM-1 and E-selectin are blocked simultaneously (C).
Magnification 200x. Scale bar indicates 50 µm.
Figure 8. HUVEC co-expressing ICAM-1 and E-selectin support cluster adhesions.
Flow-conditioned HUVEC were incubated with U937 cells under static conditions then
immunostained for ICAM-1 and E-selectin expression. (A) U937 cells viewed in phase
contrast adhere to HUVEC layer as cluster formations (yellow box) or as single cells
(black arrowheads). U937 cell cluster adhesion is found exclusively on endothelial cells
which co-express (B) E-selectin (green fluorescence) and (C) ICAM-1 (red fluorescence).
Magnification 200x. Scale bar indicates 50 µm.
Figure 9. Dynamic U937 cell adhesion. Cell-cell interaction of flow-conditioned
HUVEC layers (6 hours) with circulating U937 cells [106 cells/ml] is visually monitored
by phase contrast microscopy. (A) Circulating U937 cells arrest and adhere, in a
clustered-cell pattern, in the region of flow disturbance (includes flow recirculation,
AJP Manuscript ID: H-01078-2003 Revision 2
39
reattachment and recovery). (B) Circulating U937 cells do not adhere to the flowconditioned layer in the region of fully recovered laminar shear. Magnification 100x.
Scale bar indicates 100 µm.
AJP Manuscript ID: H-01078-2003 Revision 2
40
Table 1. Investigated flow regions. Location is measured from the downstream edge of
the step disturbance. Each region is 25 mm long in the direction parallel to the long axis
of the step disturbance. Negative shear stress indicates direction of the local flow
opposite to the bulk flow in the flow chamber.
Flow Region
Recirculation (A)
Reattachment (B)
Recovery (C)
Laminar Shear (D)
Location downstream
(mm)
Shear stress range
(dynes/cm2)
0.6 – 2.4
2.4 – 3.0
3.0 – 4.8
> 6.0
-7 to -4
-7 to +6
+6 to +10
+11
AJP Manuscript ID: H-01078-2003 Revision 2
41
Figure 1
AJP Manuscript ID: H-01078-2003 Revision 2
42
Figure 2
AJP Manuscript ID: H-01078-2003 Revision 2
43
Figure 3
AJP Manuscript ID: H-01078-2003 Revision 2
44
Figure 4
AJP Manuscript ID: H-01078-2003 Revision 2
45
Figure 5
AJP Manuscript ID: H-01078-2003 Revision 2
46
Figure 6
AJP Manuscript ID: H-01078-2003 Revision 2
47
Figure 7
AJP Manuscript ID: H-01078-2003 Revision 2
48
Figure 8
AJP Manuscript ID: H-01078-2003 Revision 2
49
Figure 9