* Speros P. Martel Section of Leukocyte Biology, and
Section of Cardiovascular Sciences, Baylor College of Medicine, Houston, Texas; and
Cox Laboratory for Biomedical Engineering, Rice University, Houston, Texas
Correspondence: C. Wayne Smith, M.D., Section of Leukocyte Biology, Baylor College of Medicine, CNRC, Room 6014, 1100 Bates, Houston, TX 77030-2399. E-mail: cwsmith{at}bcm.tmc.edu
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: inflammation • adhesion molecules • transmigration
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
An in vitro model of neutrophil rolling using a parallel
plate flow chamber has revealed that neutrophils interacting with
IL-1-stimulated endothelium at venular shear rates roll for varying
distances before arresting and transmigrating [3
]. This
rolling is selectin-dependent [15
, 16
], and
chemokines present at the endothelial surface (e.g., IL-8) appear to
stimulate CD18 integrin adhesion [17
]. Transendothelial
migration of neutrophils may occur rapidly after the leukocyte stops
rolling, and this process is almost completely inhibited in the
presence of antibodies that block adhesive functions of LFA-1 and
Mac-1. However, LFA-1 appears to be distinctly dominant, because
antibodies to Mac-1 alone are marginally effective under static
conditions in vitro [18
] and in
vivo [19
, 20
]. The dominance of LFA-1
is most clearly demonstrated in mice with a targeted deletion of CD11b
[21
], where the rate of neutrophil influx into the
peritoneal cavity in response to thioglycollate is not significantly
reduced, and emigration is markedly inhibited by anti-CD11a antibodies.
The factors that determine the site of transmigration are also poorly
defined, and our recent studies [22
, 23
]
raise some potentially important questions. In experiments performed
under static conditions in vitro, we found that 
75% of
the transmigrating neutrophils moved through corners where three
endothelial cells meet. We found these to be sites of discontinuity in
some aspects of the junctional complexes between endothelial cells.
Whether the mechanisms that focus neutrophils at tricellular corners
function when neutrophil adhesion occurs under conditions of flow is
unknown.
Here we investigate neutrophil arrest and transendothelial migration under conditions of flow. We examine the relative contributions of Mac-1, LFA-1, and ICAM-1 to these events, and we investigate the site of transendothelial migration after neutrophils arrest on activated endothelial monolayers.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Monoclonal antibodies
The blocking antibodies R3.1, R15.7, and R6.5 were provided by
Dr. Robert Rothlein (Boehringer-Ingelheim Pharmaceuticals, Ridgefield,
CT). R3.1 blocks CD11a [5
], R15.7 blocks CD18
[25
], and R6.5 maps to domain 2 of ICAM-1 and blocks the
binding of LFA-1 and Mac-1 [26
, 27
].
F(ab')2 fragments of R6.5, prepared with an ImmunoPure
F(ab')2 preparation kit (Pierce, Rockford, IL), were used
in the adhesion assay. The anti-CD11b monoclonal antibody (mAb), 60.1
[28
], was provided as a F(ab')2 fragment
by Dr. James Ruschie (Repligen Corporation, Cambridge, MA). TS2/4, a
nonblocking mAb that binds to CD11a, was provided by Dr. Tim Springer
(Center for Blood Research, Harvard Medical School, Boston, MA).
F(ab')2 fragments, prepared with an ImmunoPure
F(ab')2 preparation kit (Pierce), were used in the
adhesion assay. R15.7, 60.1 F(ab')2, and TS2/4
F(ab')2 were used at 20 µg/ml, R3.1 was used at 10
µg/ml, and R6.5 F(ab')2 was used at 100 µg/ml. Cells
were preincubated with the antibodies for 10 min at 37°C, and all
mAbs were present in the perfusion buffer during the entire experiment.
Preparation of endothelial cell monolayers
Human umbilical vein endothelial cells (HUVEC) were isolated by
collagenase digestion as previously described [15
].
Cells from 3 to 10 umbilical cords were pooled and plated on flasks
coated for at least 30 min with 0.2% gelatin (Difco Laboratories,
Detroit, MI). Monolayers were cultured in M199 (GIBCO BRL), containing
10% fetal bovine serum and 10% bovine calf serum (Hyclone
Laboratories, Logan, UT), 1% penicillin-streptomycin (GIBCO BRL), 1%
fungizone (GIBCO BRL), 1% HEPES buffer (GIBCO BRL), 1 µg/ml heparin
(Sigma Chemical Co., St. Louis, MO), and endothelial cell growth factor
(prepared as per Maciag and Weinstein [29
]). Cultures
were grown in a 37°C humidified atmosphere with 5% CO2
and reached confluence after 4–6 days. HUVEC were then passaged onto
35 mm tissue culture dishes (Corning Glass Works, Corning, NY), which
were coated with glutaraldehyde crosslinked gelatin, as described by
Burns et al. [22
]. Two to four days later,
HUVEC were stimulated with 10 U/ml recombinant human IL-1ß (Genzyme
Diagnostics, Cambridge, MA) at 37°C for 4 h prior to the
adhesion assays.
In some adhesion experiments described below, canine jugular vein endothelial cell (CJVEC) monolayers were used. These cells were isolated as previously described [30 ] and grown as confluent monolayers on glutaraldhyde crosslinked gelatin.
Adhesion assay under flow conditions
Neutrophil adhesion under shear flow was performed as previously
described [15
, 31
]. Briefly, confluent
IL-1-stimulated HUVEC monolayers were mounted in a parallel plate flow
chamber, and DPBS was perfused over the monolayer for 2–3 min. The
isolated neutrophils were then diluted to 106/ml in DPBS
containing Ca++, Mg++, and 11 mM glucose
(DPBS + Glu) and were perfused over the monolayer at a range of
flow rates that resulted in shear stresses between 0.2 and 3.0
dynes/cm2. Temperature was maintained at 37°C. The
interaction between the HUVEC and neutrophils from the time the
neutrophils first passed over the monolayer was recorded on videotape
for 11 min by phase-contrast video microscopy (Diaphot-TMD microscope,
Nikon Inc., Garden City, NY; CCD Video Camera, Sony Corp., Park Ridge,
NJ) and was quantified with Optimas image analysis software (BioScan,
Edmonds, WA). For each experiment, a single field of view was recorded
for the first 9.5 min, and five additional fields of view were taped
during the last 1.5 min (i.e., 9.5–11 min after neutrophils first
interacted with the monolayer).
ICAM-1 was blocked on some HUVEC monolayers by incubation with R6.5 Fab for 30 min in a 37°C, 5% CO2 environment. In some experiments at 2.0 dynes/cm2, adhesion molecules on neutrophils were blocked by incubation with R3.1, 60.1 F(ab')2, TS2/4, or R15.7 for 10 min at 37°C. Neutrophils were then diluted and perfused through the flow chamber. R15.7, R3.1, 60.1, R6.5, and TS2/4 were present in the neutrophil suspension during the entire experiment.
Adhesion assay under static conditions
For the static adhesion assays, confluent IL-1-stimulated HUVEC
monolayers were mounted in a parallel plate flow chamber, and isolated
neutrophils (106/ml in DPBS + Glu) were perfused over
the monolayer at 2.0 dynes/cm2 for 2–3 min. The flow was
stopped for 4 min and three fields of view and recorded under a 20x
objective. The flow was resumed, and unbound cells were washed away at
2.0 dynes/cm2, and firmly adherent cells were determined in
three fields were recorded under a 20x objective. Neutrophils were
incubated with R3.1 or 60.1 (whole) for 20 min at room temperature in
some of the static adhesion experiments. Neutrophils were then diluted
and perfused through the flow chamber. R3.1 and 60.1 were present in
the neutrophil suspension during the entire experiment.
Determination of the site of neutrophil adhesion to stimulated endothelial monolayers in vitro was determined under static conditions. HUVEC monolayers were stimulated for 4 h with IL-1, as described above, or CJVEC monolayers were stimulated with 30 ng/ml of endotoxin for 4 h. Isolated human or canine neutrophils were allowed to contact the appropriate species of monolayer for 4 min. The monolayers were then fixed and stained with silver nitrate, as previously described [22 ]. The position of attached neutrophils was then determined by light microscopy. Neutrophils overlying a stained border and attached neutrophils not overlying a border were counted.
Quantitation of adhesion
The accumulation of neutrophils on the monolayer was determined
by counting the total number of rolling, arrested, and transmigrated
neutrophils on one field of view at each minute of flow during the
first 9 min. After the initial 9.5 min of interaction, the number of
neutrophils interacting with the monolayer at 2.0 dynes/cm2
was determined by averaging the total number of rolling, arrested, and
transmigrated neutrophils on the monolayer from five fields of view
from the last 1.5 min of the experiment. For the experiments where
shear stress was varied, the number of interacting neutrophils was
calculated by averaging the number of neutrophils in three fields of
view.
Arrested cells were defined as those that remained stationary during a 10-sec interval, and rolling cells were those that moved during the 10-sec interval. Transmigrating neutrophils included only those that had completely transmigrated at the beginning of the 10-sec interval. Rolling and arrested neutrophils appeared phase-bright and could be easily discerned from transmigrated cells, which appeared phase-dark [5 , 26 ]. Transmigrated cells were included in the adherent population, because they had arrested before they began to transmigrate.
For the static adhesion experiments, the average number of neutrophils remaining adherent (including transmigrated cells) in three fields of view after the flow was resumed was divided by the average number of neutrophils in three fields of view contacting the monolayer during the static interval.
The time and distance that individual neutrophils rolled from capture
until arrest at 2.0 dynes/cm2 were measured during the
first 9 min of the experiment (Fig. 1A
). The field of view was 500 µm long, and only those cells that
initially interacted directly with the monolayer in the first 100 µm
of the field of view were counted. The distance for cells that did not
stop in the field of view was denoted as >400 µm. The rolling
velocity of each neutrophil was calculated by dividing the distance by
the time that the neutrophil rolled from its capture until it stopped
or rolled off the field of view. The data from at least three
experiments for each condition, representing at least 40 total
neutrophils, were pooled for the distance and velocity histograms. The
transmigration time—the interval from arrest until the neutrophil had
completely transmigrated—was also observed in a frame-by-frame
analysis of the video tapes. Transmigration was judged to be complete
when the phase-bright uropod became phase dark. The distance that a
neutrophil moved from arrest (i.e., after it was stationary for 10 sec)
to where it transmigrated was measured (Fig. 1B)
. The location of
transmigration was determined, as previously described
[22
], to be the position of its uropod above the
endothelial monolayer as it transmigrated. The resolution of movement
possible using a 20x objective with a numerical aperture of 0.4 was
0.8 µm. The data from at least three experiments of at least 30
neutrophils were pooled for the transmigration time and distance.
 View larger version (26K): [in a new window] |
Figure 1. Measurement of neutrophil rolling (distance and velocity) and migration
after stationary adhesion. (A) The distance that individual neutrophils
rolled until they arrested or rolled out of the field of view. The flow
adhesion assays were performed on HUVEC monolayers as described in
Materials and Methods. Neutrophils captured within the first 100 µm
of the field of view were observed until they stopped rolling. Rolling
velocity was determined for these individual neutrophils. Neutrophils
that immediately stopped without rolling (*) were denoted as having
zero rolling distance and velocity, and those that rolled off the field
of view had a rolling distance >400 µm. (B) The distance that
individual neutrophils migrated from the point of arrest to the point
of transendothelial migration was measured from the center of the
spherical arrested cell to the tip of the phase-bright uropod just
prior to the cell moving beneath the monolayer. The elapsed time for
this migration was also recorded.
|
Neutrophil transmigration between two adjacent endothelial cells or between three endothelial cells at tricellular corners (the site where the margins of three endothelial cells meet) was quantitated from the silver-stained HUVEC monolayers under the 20x objective of the Nikon Diaphot phase contrast microscope. One hundred cells that were in the process of transmigrating (i.e., a phase-bright and a phase-dark portion could be observed) were counted, and the number that transmigrated between two cells or through tricellular corners was noted. The data are reported as the average of three separate experiments.
Statistical analysis
Results are presented as mean ± SEM. For the analysis of
adhesion under different shear stresses, the effects of different
interventions on the percent of abruptly arresting neutrophils and the
comparison of the actual and predicted transmigration distances,
one-way analyses of variance (ANOVA), were performed using InStat
(GraphPad Software, San Diego, CA). The probability of statistical
significance between two interventions was determined by the Tukey
test. For the velocity and distance analyses, Kruskal-Wallis tests were
performed, and Dunn’s tests determined significance between two
interventions. P values < 0.05 were considered
significant.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
 View larger version (12K): [in a new window] |
Figure 2. Neutrophil accumulation on HUVEC monolayers stimulated for 4 h
with IL-1ß under a shear stress of 2.0 dynes/cm2. The
flow adhesion assays were performed on HUVEC monolayers as described in
Materials and Methods. The number of neutrophils was counted at each
minute during the first 9 min of flow in the same field of view. The
total number of neutrophils includes all interacting neutrophils (i.e.,
rolling, arrested, and transmigrated cells). A linear regression was
performed with R2 = 0.984.
|
 View larger version (14K): [in a new window] |
Figure 3. Neutrophil arrest at 2.0 dynes/cm2 on stimulated HUVEC
monolayers. The flow adhesion assays were performed on HUVEC monolayers
as described in Materials and Methods. Neutrophils were untreated or
preincubated with 10 µg/ml of the anti-CD11a blocking mAb, R3.1; 20
µg/ml of the anti-CD11b mAb, 60.1 F(ab')2; 20 µg/ml
of anti-CD11, a nonblocking mAb TS2/4 (@); or 20 µg/ml of the
anti-CD18 mAb, R15.7. HUVEC monolayers were untreated or preincubated
for 30 min at 37°C with 100 µg/ml of the anti-ICAM
F(ab')2, R6.5. All mAbs were present in the perfusion
buffer during the entire experiment. The percent of control adhesion
refers to the total arrested and transmigrated neutrophils after 9.5
min of flow. Data are expressed as the mean ± SEM from at least
three separate experiments. **P < 0.01 compared with
untreated neutrophils on untreated HUVEC. *P < 0.05
compared with untreated neutrophils on untreated HUVEC.
|
 View larger version (17K): [in a new window] |
Figure 4. Effects of anti-CD11a antibody on neutrophil adhesion to stimulated
HUVEC. (A) Static adhesion assay was performed as described in
Materials and Methods. Neutrophils were untreated or preincubated with
10 µg/ml of the anti-CD11a-blocking mAb, R3.1, or 20 µg/ml of the
anti-CD11b mAb, 60.1. All mAbs were present with the cells during the
entire experiment. The percent total interacting polymorphonuclear
neutrophil (PMN) refers to the proportion of the cells that were firmly
adherent contacting the monolayer. *P < 0.01 compared
with control; **P < 0.01 compared with control or with
the condition with added anti-CD11a antibody; n = 3,
mean ± SEM. Both, a combination of anti-CD11a and anti-CD11b
antibodies added. (B) The flow adhesion assays were performed on HUVEC
monolayers as described in Materials and Methods. Neutrophils were
untreated or preincubated with 10 µg/ml of the anti-CD11a-blocking
mAb, R3.1; 20 µg/ml of the anti-CD11b mAb, 60.1; or a combination of
these antibodies (Both). All mAbs were present with the cells during
the entire experiment. The percent of control adhesion refers
neutrophils that had arrested or transmigrated after 9.5 min of flow in
the absence of mAb treatment. Data are expressed as the mean ±
SEM from at least three separate experiments.
|
|
View this table: [in a new window] |
Table 1. Rolling Behavior of Isolated Neutrophils on IL-1-Stimulated HUVEC
Monolayersa
|
Neutrophil migration after firm adhesion to HUVEC
As noted earlier, 58% of neutrophils associated with the
monolayer after 9.5 min of flowing a neutrophil suspension over the
endothelial cells had transmigrated. This meant that 450 neutrophils
were beneath each mm2 of endothelial monolayer. The
addition of blocking mAbs to only ICAM-1, LFA-1, or Mac-1 did not
significantly reduce this number (Fig. 5
). Blocking LFA-1 and Mac-1 significantly decreased the number of
transmigrated neutrophils from 469 ± 29
neutrophils/mm2 in the absence of mAb treatment to 146 ± 31/mm2. Blocking CD18 or CD18 and ICAM-1 similarly
decreased the number of transmigrated cells (Fig. 5)
. This effect
appeared to be a result of reductions in the number of arrested
neutrophils, because transmigration appeared to proceed favorably for
those neutrophils that were able to firmly adhere under control or
antibody-blocking conditions. In each treatment condition, >40
individual neutrophils were followed as described in Figure 1
.
Regardless of the blocking antibody combination used, >90% of
arrested cells migrated beneath the endothelial monolayer during a
3-min observation time.
 View larger version (14K): [in a new window] |
Figure 5. Transmigrated neutrophils on 4 h IL-1b-stimulated HUVEC monolayers
at 2.0 dynes/cm2. The flow adhesion assays were performed
on HUVEC monolayers as described in Materials and Methods. Neutrophils
were untreated or preincubated with 10 µg/ml of the
anti-CD11a-blocking mAb, R3.1; 20 µg/ml of the anti-CD11b mAb, 60.1
F(ab')2; or 20 µg/ml of the anti-CD18 mAb, R15.7. HUVEC
monolayers were untreated or pretreated with 100 µg/ml of the
anti-ICAM F(ab')2, R6.5. All mAbs were present in the
perfusion buffer during the entire experiment. The percent control
migration refers to the number of transmigrated neutrophils at the end
of 9.5 minutes of flowing neutrophils over the IL-1-stimulated
monolayer. Data are expressed as the mean ± SEM from at least
three separate experiments. **P < 0.01 compared with
untreated, transmigrated neutrophils.
|
The distance that neutrophils moved from the site of arrest to the site of transmigration was also measured. In the absence of blocking mAbs, the mean (SEM) migration distance was calculated to be 5.5 ± 0.7 µm (Fig. 6 ). Regression analysis revealed that this migration distance was independent of the distance each neutrophil rolled before arresting. Although blocking ICAM-1 on the monolayer, blocking CD11a and CD11b on the neutrophil, or blocking CD18 and ICAM-1 in combination reduced the number of cells arresting, there was no effect on the distance that the subset of arrested neutrophils moved before transmigrating (Fig. 6) . In addition, the mean distance (6.5 µm) migrated by cells that abruptly stopped was not significantly different from the total population. These data suggest that flowing or rolling neutrophils arrested close to their site of transmigration, and those cells that arrested in the presence of antibody inhibition of CD18 and ICAM-1 still arrested close to the site of transmigration.
 View larger version (12K): [in a new window] |
Figure 6. Distance neutrophils moved from the site of arrest to the site of
transmigration under flow conditions on 4 h IL-1b-stimulated
HUVEC. The flow adhesion assays were performed on HUVEC monolayers as
described in Materials and Methods. Neutrophils were untreated or
preincubated with 20 µg/ml of the anti-CD18 mAb, R15.7, or with a
combination of 10 µg/ml of the anti-CD11a-blocking mAb, R3.1, and 20
µg/ml of the anti-CD11b mAb, 60.1 F(ab')2. HUVEC
monolayers were untreated or pretreated with 100 µg/ml of the
anti-ICAM F(ab')2, R6.5. All mAbs were present in the
perfusion buffer during the entire experiment. Data are expressed as
the mean ± SEM from at least 30 neutrophils.
|
75%) occurs at tricellular corners, the site where
the margins of three endothelial cells meet [22
]. The
remaining 25% transmigrate at the border between two adjacent
endothelial cells. By silver staining the HUVEC monolayers after each
flow experiment, we were able to determine the location of neutrophils
in the process of transmigration at the time the monolayers were fixed.
We now report that under flow conditions (2 dynes/cm2),
69.9 ± 1.7% of neutrophils transmigrated at tricellular corners,
and the remaining neutrophils transmigrated at the junction of two
adjacent endothelial cells. We compared the observed behavior of arrested neutrophils with a stochastic model that assigned random sites of arrest on endothelial cells under flow. Previously, we determined that the average perimeter of an endothelial cell under the culture conditions used in these studies was 142 µm and that each endothelial cell formed an average of 5.3 tricellular corners with adjacent cells in the monolayer under the culture conditions used in the current studies [22 ]. For our model of predicted transmigration distance, we approximated each endothelial cell as a pentagon with a perimeter of 142 µm, with the center of a tricellular corner at each vertex (Fig. 7 ). Each pentagon was assigned a different shape to reflect the heterogeneity in the observed shapes of endothelial cells in a HUVEC monolayer. The center of each neutrophil was represented by a point. To determine the predicted average distance from the point of arrest to the site of transmigration, 150 points (neutrophils) were randomly generated within the pentagons (endothelial cells), and the distance from each point to the nearest vertex (i.e., tricellular corner) and the minimum distance to the nearest margin were determined. The assumption that neutrophils migrated in a straight line was the most conservative comparison with our empirical data. We calculated the average weighted distances to reflect the distribution of transmigration at tricellular corners and between two endothelial cells (i.e., 70% x distance to nearest tricellular corner + 30% x distance to nearest margin). The frequency histograms of the distances generated from the model suggested that randomly arrested neutrophils would migrate significantly farther than the actual transmigration distances observed in our experiments (Fig. 7B) . In most instances, the body of the neutrophil appeared to remain at the site of initial arrest, simply extending a pseudopod to the point of eventual transmigration. As noted above, >90% of the arrested cells transmigrated, and neutrophils were rarely seen migrating over the apical surface of the endothelial cells. The rare neutrophils that migrated across the apical surface of the endothelial monolayer often did not locate a site for transmigration during the observation period.
 View larger version (37K): [in a new window] |
Figure 7. Comparison of actual and predicted transmigration distances. (A) The
distance that individual neutrophils moved from their site of arrest to
their site of transmigration was predicted as described in Materials
and Methods. Endothelial cells were approximated as pentagons of
perimeter 142 µm, with each vertex representing the center of a
tricellular corner. (a) Neutrophils were modeled as points randomly
distributed in a region of three adjacent endothelial cells
(pentagons). (b) (1) The distance to the nearest tricellular corner was
the distance from each point to the nearest vertex. (2) The distance to
the nearest cell margin was the minimum distance from each point to the
margin between two adjacent endothelial cells (pentagons). (B)
Transmigration distance is the distance moved from the site of arrest
to the site of transmigration (see Fig. 1
). The histograms are
frequency distributions, expressed as percentage of the total number of
neutrophils observed. For the actual migration distance, the flow
adhesion assays were performed with neutrophils on IL-1-stimulated
HUVEC monolayers in the absence of mAbs as described in Materials and
Methods. Data are expressed as a frequency histogram of at least 40
neutrophils. The predicted transmigration distances were calculated as
described in Results and represent data from 150 neutrophils randomly
arrested on a region of three adjacent endothelial cells. The histogram
shows the weighted distribution of transmigration distance to cell
margins and tricellular corners. The median actual distance migrated
was 5.5 µm, and the median predicted distance was 9.4 µm,
P < 0.001. (C) The predicted transmigration distance
was calculated by changing the model in only one aspect. The sites of
random arrest were confined to endothelial borders of an assigned width
of 2 µm. These new predicted distances are plotted along with the
same actual values shown in (B) for comparison. This model appears to
be consistent with the empirical data.
|
 View larger version (34K): [in a new window] |
Figure 8. Site of neutrophil adhesion on activated endothelial monolayers. HUVEC
or CJVEC (DEC) monolayers were grown to confluence and then stimulated
with IL-1 (10 U/ml) or endotoxin (10 ng/ml), respectively, for 3 h. Isolated human or canine neutrophils were allowed to adhere for 4
min, and then the monolayers were rinsed gently, fixed, and stained
with silver nitrate to define the interendothelial junctions. Using
light microscopy, the percent of neutrophil overlying borders or
overlying the body of the endothelial cells (Body) was determined. The
mean ± SEM for five separate experiments is plotted.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The relative contributions of LFA-1 and Mac-1
Although the contribution of CD18 integrins to the arrest of
rolling neutrophils has been established in many laboratories using
cell culture models and intravital techniques, the relative
contributions of LFA-1 and Mac-1 remain uncertain. The in
vitro model used in the present studies reveals that either of
these ß2 integrins can stop rolling neutrophils, a
conclusion supported by the finding that neither anti-CD11a nor
anti-CD11b alone affected any reduction in the number of arrested
cells, whether in the percent of leukocytes that abruptly arrested, the
rolling velocity, or the total that were firmly adherent at any of the
observation times. The mAbs used in these studies have been shown to
block the functions of LFA-1 and Mac-1 [5
], and when
used in combination in the current experiments, these mAbs were as
inhibitory as anti-CD18.
As shown in Figure 4 , we found a significant distinction between adhesion under static conditions and under flow. Blocking LFA-1 was without effect on the arrest of neutrophils moving in response to fluid shear, but blocking LFA-1 produced almost 50% inhibition of adhesion to the HUVEC monolayers when neutrophils contacted the endothelial cells in the absence of shear. Thus, shear stress appears to increase the contribution of Mac-1 to neutrophil/endothelial cell adhesion. The limited involvement of Mac-1 in static adhesion to cytokine-stimulated endothelium has been previously reported [18 ], but the potentiating effect of shear is undefined.
One possible distinction between adhesion under static and flow conditions is that selectins may interact with the neutrophils differently in these two conditions. L-selectin has been shown to exhibit a shear threshold (i.e., an inability to function as an adhesion molecule until the wall shear stress applied to the neutrophil reaches a threshold of >0.5 dynes/cm2 [38 , 39 ]). Isolated E- and P-selectin also exhibit similar shear thresholds at site densities of 200/µm2 or less [39 ]. It appears that shear-induced torque is needed to maintain selectin-dependent adhesion (rolling) [39 ], because gravitational force is sufficient to detach cells from L-, E-, and P-selectin binding. The shear threshold of selectins is unexplained in molecular terms but functionally could mean different forms of selectin interaction under static and flow conditions. Of potential relevance to this discussion is the fact that L-selectin can function as a signaling molecule on neutrophils [40 , 41 ], triggering, among other things, an increase in the adhesive function of Mac-1 [42 ]. L-selectin ligation can also significantly potentiate the effect of chemotactic stimulation on Mac-1 function [24 ], and current evidence indicates that chemotactic stimulation [via IL-8 and platelet-activating factor (PAF)] occurs when neutrophils contact cytokine-stimulated endothelial cells under static or flow conditions [43 44 45 ]. In recent unpublished studies, we have shown that E-selectin binding to neutrophils under shear, but not static conditions, can also activate Mac-1 adhesion (Simon et al., unpublished results), and Evangelista et al. [46 ] have provided evidence that P-selectin interacting with its ligand PSGL-1 can trigger enhanced Mac-1 adhesion. The E- and P-selectin effects on Mac-1 adhesion were sensitive to tyrosine kinase inhibitors. Thus, the functional significance of selectin-dependent capture under flow may extend beyond simply the support of rolling adhesion to include potentiation of chemotactic activation of Mac-1 adhesion.
The use of anti-ICAM-1 antibody in these studies significantly diminished the ability of the neutrophils to stop rolling. However, the effect is minimal compared with that of anti-CD18 or combined anti-CD11a and -CD11b. Although ICAM-1 has been shown to be a ligand for LFA-1 and Mac-1 on neutrophils [27 ], the binding of Mac-1 to ICAM-1 is relatively weak [4 ], and Mac-1 apparently recognizes another uncharacterized ligand on endothelial cells [14 , 47 ]. Because our data indicate that Mac-1 or LFA-1 can support stationary adhesion of neutrophils under flow, it is likely that Mac-1 is primarily interacting with a ligand other than ICAM-1 under these flow conditions on stimulated HUVEC. This point is supported by the finding that combined use of anti-CD11b and anti-ICAM-1 antibodies resulted in reductions in firm adhesion that were as marked as combinations of anti-CD11b and anti-CD11a antibodies.
In contrast to the important role played by the CD18 integrins in
arrest of rolling cells, the role for these integrins in transmigration
was not distinguishable from their role in stable adhesion. When
individual arrested neutrophils were continuously observed, >90%
transmigrated regardless of antibody treatment. It appeared that if a
neutrophil could arrest in the presence of anti-CD18 antibody, it could
transmigrate. Our analysis raises the possibility that CD18 functions
are limited to adhering the leukocyte to endothelial cells, and that
the process of transendothelial migration involves other mechanisms.
CD18-independent mechanisms of emigration have been demonstrated
in vivo [48
]. Jung et al.
[49
] found that transendothelial migration was highly
efficient for arrested leukocytes in TNF
-stimulated cremasteric
vessels in CD18-deficient mice. It appears that reducing adhesion by
blocking or removing CD18 integrins indirectly reduces transmigration,
and it may be that CD18 integrins are functional in bringing the
neutrophil to the site of transmigration but not contributing to the
interaction of neutrophils with the endothelial membranes in the
interendothelial cleft. Anti-CD18 would thus reduce transmigration
under flow (the present study) or under static conditions (numerous
earlier studies [5
, 50
]). The observation
that anti-CD18 reduced the number of transmigrated neutrophils to
<20% of the control level indicates the major importance of these
integrins in promoting stationary adhesion under flow.
Efficiency of neutrophil arrest and transmigration
We found that under a shear stress of 2.0 dynes/cm2,
neutrophils are able to interact with cytokine-stimulated endothelial
cells at significant levels. The accumulation of neutrophils was linear
over time (Fig. 2)
. However, recent studies of neutrophil recruitment
on purified E- and P-selectin monolayers have reported an exponential
accumulation of neutrophils [33
, 34
]. These
and other studies have proposed that bound neutrophils are able to
support the capture (tethering) of neutrophils from the free stream,
and thus amplify neutrophil recruitment [33
34
35
,
51
]. Our results suggest that the recruitment of
neutrophils from the free stream by bound neutrophils did not
significantly increase cell capture on cytokine-stimulated endothelial
monolayers in our experimental model. The topography of the monolayer
of endothelial cells may be such that bound neutrophils are not as
accessible to those in the free stream as neutrophils bound to a
monolayer of purified E- or P-selectin on a planar surface. Neutrophils
are able to transmigrate under HUVEC monolayers within 90 sec, a
timeframe that may be too rapid to enable many neutrophils to be
recruited before transmigration.
In experiments where the behavior of individual neutrophils was followed from the time of the initial observed contact, a remarkably high percentage of cells arrested and transmigrated. The observed efficiency was consistently >90% for both events, and the median rolling distance (5 µm) was less than the diameter of an individual neutrophil. If data collection were to be limited simply to categorizing the behavior of all cells in a field at a given time in the flow experiment, it would be impossible to realize this high efficiency. Such data collection has characterized many previous studies in vivo and in vitro [52 ], leaving the question of the efficiency with which rolling cells are captured by activated endothelium unresolved.
Rainger et al. [17 , 53 ] followed in vitro the behavior of individual neutrophils on endothelial monolayers that had been stimulated with hypoxia/reoxygenation or exposure to hydrogen peroxide. These stimuli provoke mobilization of P-selectin to the apical surface of endothelial cells from Weibel-Palade bodies [54 , 55 ]. Although many rolling neutrophils were observed, arrest was undetected. Stationary adhesion could be induced, however, by pretreating the stimulated endothelium with IL-8 or PAF and washing away the unbound chemotactic factor. A significant percentage ( 30%) of rolling cells then converted to stable adhesion under flow. The rate of this conversion was remarkably fast, consistently occurring within less than a second.
Given the current assumption that IL-1- or TNF-stimulated endothelial
cells present IL-8 and PAF [44
, 56
] as
stimuli that activate CD18 integrin-dependent stable adhesion, and
given the observations of Rainger et al.[53
]
that conversion occurred in a median rolling duration of 720 msec, it
is possible that these chemotactic substances contribute to the
high-efficiency transition of neutrophils from flowing or rolling to
arrested cells. In our experiments, the median time for this transition
was 2 sec. Our current model differs from the experiments of Rainger
et al. in that the primary selectins supporting rolling on
cytokine-stimulated endothelial monolayers are L- and E-selectin
[15
, 43
], and the wall shear stress used
was higher. Under similar conditions and site densities, neutrophils
appear to roll more slowly on E-selectin than P-selectin
[57
]. Moreover, the rolling velocities of neutrophils at
a wall shear stress of 2 dynes/cm2 on P-selectin-expressing
monolayers (10–15 µm/s) [53
, 58
] may be
somewhat higher than on IL-1- or TNF-stimulated HUVEC monolayers
[3
, 15
]. Rolling velocity is affected by
the level of cytokine stimulation of endothelial cells
[59
] and the site density of selectins
[60
].
Under control conditions, nearly 30% of the interacting neutrophils
were able to arrest immediately after capture at 2.0
dynes/cm2 without observable rolling. It appears that
neutrophil rolling is not essential for arrest. However, it is possible
that rolling was more rapid than the minimum time period for detection
in our analysis system, 30 msec. Lawrence and Springer
[61
] estimated that a rolling neutrophil would contact
an E-selectin-containing lipid bilayer for only 0.1–1 sec before
arrest. They reasoned that adhesion strengthening must occur within
this time period. For the neutrophils that appeared to abruptly arrest
in our system, adhesion strengthening might have occurred within 30
msec. Another possibility is that some neutrophils became stimulated
while flowing across the activated monolayer before capture and arrest.
The complete inability to abruptly arrest when CD18 or ICAM-1 was
blocked confirms the importance of CD18 and ICAM-1 in mediating rapid
adhesion strengthening and arrest.
The arrest of rolling leukocytes in vivo has been evaluated in many studies by continuous, direct, microscopic observation of a single vessel segment [36 ]. The rolling flux appears very high compared with the number of leukocytes that stop within the single field of view, leaving the impression that the efficiency of conversion from rolling to arrest is quite low. Lack of information, though, on the type and fate of each passing cell makes it impossible to directly answer the question of efficiency. What seems likely at the present time is that the rate of conversion to stable adhesion after endothelial contact is much higher in the in vitro models than in vivo. Hydrodynamic and technical factors probably account for some of this apparent difference, as well as some possible activation of leukocytes, during the isolation for in vitro studies, but direct evidence about mechanisms is not available.
Neutrophils arrest very near the site of transendothelial migration
Neutrophils move less than the diameter of the leukocyte from the
point of arrest to the site of transendothelial migration. This appears
to be true whether neutrophils abruptly stop from the flow stream or
roll for relatively long distances, because no correlation was found
between the distance rolled and the distance migrated after arrest. The
efficiency of transendothelial migration in our in vitro
model is very high, exceeding 90%, and distinctly higher than the
60% seen under static conditions when cells are allowed to simply
settle onto monolayers making random contact with endothelial cells
[22
]. The site of arrest under flow appears to be
spatially and functionally associated with the site of transmigration.
The distance that an arrested neutrophil would migrate before migration
was consistent with a model in which they adhered preferentially at
endothelial borders. Such a condition would likely require the
colocalization of adhesion molecules and chemotactic stimuli, but
little is known about the topography of IL-8 and PAF presentation, and
ICAM-1 and E-selectin appear to be distributed uniformly over the
surface of cytokine-stimulated endothelial cells. We have recently
obtained evidence that P-selectin is concentrated at endothelial cell
junctions following histamine stimulation [37
], but
there is little evidence that this adhesion molecule participates in
rolling or stationary adhesion on HUVEC monolayers stimulated for
4 h with IL-1. ICAM-2 appears to be concentrated at the junctions
of unstimulated endothelial cells (unpublished results). There is
evidence that Mac-1 and LFA-1 can interact with this molecule
[7
, 62
], and Issekutz et al.
[50
] have shown that ICAM-2 can participate in
transmigration of neutrophils under static experimental conditions.
Another feature of the monolayer that may contribute to the stationary
adhesion is the surface contour of the monolayer. Davies and colleagues
[63
, 64
] using atomic force microscopy have
mapped the 3-D shape of the luminal endothelial surface and found it to
be an important determinant of spatial distribution of stress gradients
from the elevated central regions of endothelial cells to the lower
junctional areas. Such gradients could favor stationary adhesion at the
cellular regions of lower shear stress (i.e., at cell junctions).
The high percentage of neutrophils migrating at tricellular corners
where three endothelial cells meet is consistent with observations
under static conditions. Burns et al. [22
]
have shown that this region is the site of some discontinuity in the
junctional complexes between endothelial cells (e.g., occludin) and
have argued that this discontinuity obviates the need for disrupting
intercellular adhesion of endothelial cells. Walker et al.
[65
] and Barry et al. [66
]
have provided ultrastructural and immunohistochemical evidence that
junctional structures are discontinuous at endothelial tricellular
corners in vivo. The static experiments allowed
neutrophils to settle at unit gravity onto monolayers of
cytokine-stimulated HUVEC, and the location of transmigrating
neutrophils was determined. It is assumed that there was random contact
of settling cells with endothelial cells and that migrating neutrophils
would seek the sites for transmigration. Under flow conditions,
neutrophils stopped rolling within a few µm from the site of
transmigration and because for 70% of the cells, this site was at
the tricellular corner, there seems to be a localized endothelial
determinant that promotes stationary adhesion at distinct sites. It
remains to be determined if neutrophil-activating factors and/or
ligands for CD18 integrins are spatially organized to target neutrophil
arrest near tricellular corners. The possible significance of this site
may lie in the observation that intercellular junctions are inherently
discontinuous at tricellular corners, and leukocyte migration at this
point may minimize the disruption of integrity of the endothelial
monolayer.
Received October 30, 1999; revised January 18, 2000; accepted January 20, 2000.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
4ß7 and LFA-1 in lymphocyte homing to peyer’s patch-HEV in situ: the multistep model confirmed and refined Immunity 3,99-108[Medline]This article has been cited by other articles:
 |
|
 |
|
  M.-H. Kim, F.-R. E. Curry, and S. I. Simon Dynamics of neutrophil extravasation and vascular permeability are uncoupled during aseptic cutaneous wounding Am J Physiol Cell Physiol, April 1, 2009; 296(4): C848 - C856. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. R. Sarantos, H. Zhang, U. Y. Schaff, N. Dixit, H. N. Hayenga, C. A. Lowell, and S. I. Simon Transmigration of Neutrophils across Inflamed Endothelium Is Signaled through LFA-1 and Src Family Kinase J. Immunol., December 15, 2008; 181(12): 8660 - 8669. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. L. Gonzalez, W. El-Bjeirami, J. L. West, L. V. McIntire, and C. W. Smith Transendothelial migration enhances integrin-dependent human neutrophil chemokinesis J. Leukoc. Biol., March 1, 2007; 81(3): 686 - 695. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Ortolan, E. V. Tibaldi, B. Ferranti, L. Lavagno, G. Garbarino, R. Notaro, L. Luzzatto, F. Malavasi, and A. Funaro CD157 plays a pivotal role in neutrophil transendothelial migration Blood, December 15, 2006; 108(13): 4214 - 4222. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. N. Mundhekar, D. C. Bullard, and D. F. Kucik Intracellular heterogeneity in adhesiveness of endothelium affects early steps in leukocyte adhesion Am J Physiol Cell Physiol, July 1, 2006; 291(1): C130 - C137. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. E. Green, U. Y. Schaff, M. R. Sarantos, A. F. H. Lum, D. E. Staunton, and S. I. Simon Dynamic shifts in LFA-1 affinity regulate neutrophil rolling, arrest, and transmigration on inflamed endothelium Blood, March 1, 2006; 107(5): 2101 - 2111. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Yang, J. R. Kowalski, X. Zhan, S. M. Thomas, and F. W. Luscinskas Endothelial Cell Cortactin Phosphorylation by Src Contributes to Polymorphonuclear Leukocyte Transmigration In Vitro Circ. Res., February 17, 2006; 98(3): 394 - 402. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Cook-Mills and T. L. Deem Active participation of endothelial cells in inflammation J. Leukoc. Biol., April 1, 2005; 77(4): 487 - 495. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. W. Calderwood, J. M. Williams, M. D. Morgan, G. B. Nash, and C. O. S. Savage ANCA induces {beta}2 integrin and CXC chemokine-dependent neutrophil-endothelial cell interactions that mimic those of highly cytokine-activated endothelium J. Leukoc. Biol., January 1, 2005; 77(1): 33 - 43. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. K. Shaw, S. Ma, M. B. Kim, R. M. Rao, C. U. Hartman, R. M. Froio, L. Yang, T. Jones, Y. Liu, A. Nusrat, et al. Coordinated Redistribution of Leukocyte LFA-1 and Endothelial Cell ICAM-1 Accompany Neutrophil Transmigration J. Exp. Med., December 20, 2004; 200(12): 1571 - 1580. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. L. Deem and J. M. Cook-Mills Vascular cell adhesion molecule 1 (VCAM-1) activation of endothelial cell matrix metalloproteinases: role of reactive oxygen species Blood, October 15, 2004; 104(8): 2385 - 2393. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. H. Birdsall, W. J. Porter, D. M. Green, J. Rubio, J. Trial, and R. D. Rossen Impact of Fibronectin Fragments on the Transendothelial Migration of HIV-Infected Leukocytes and the Development of Subendothelial Foci of Infectious Leukocytes J. Immunol., August 15, 2004; 173(4): 2746 - 2754. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X. Zhang, A. Chen, D. De Leon, H. Li, E. Noiri, V. T. Moy, and M. S. Goligorsky Atomic force microscopy measurement of leukocyte-endothelial interaction Am J Physiol Heart Circ Physiol, January 1, 2004; 286(1): H359 - H367. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. R. Burns, C. W. Smith, and D. C. Walker Unique Structural Features That Influence Neutrophil Emigration Into the Lung Physiol Rev, April 1, 2003; 83(2): 309 - 336. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. L. Cuvelier and K. D. Patel Shear-dependent Eosinophil Transmigration on Interleukin 4-stimulated Endothelial Cells: A Role for Endothelium-associated Eotaxin-3 J. Exp. Med., December 10, 2001; 194(12): 1699 - 1709. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. M. Seo, L. V. McIntire, and C. W. Smith Effects of IL-8, Gro-alpha , and LTB4 on the adhesive kinetics of LFA-1 and Mac-1 on human neutrophils Am J Physiol Cell Physiol, November 1, 2001; 281(5): C1568 - C1578. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. M. Mariscalco, W. Vergara, J. Mei, E. O'B. Smith, and C. W. Smith Mechanisms of decreased leukocyte localization in the developing host Am J Physiol Heart Circ Physiol, February 1, 2002; 282(2): H636 - H644. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
