| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Physiology, The Medical School, University of Birmingham, Birmingham, United Kingdom
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
for 4 h and evaluated changes caused by exogenously
added chemotactic agents. When HUVEC were treated with 2 U/ml TNF,
flowing neutrophils adhered, with the majority rolling and relatively
few migrating through the monolayer. If fMLP, IL-8, zymosan-activated
plasma (a source of activated complement factor C5a), epithelial
cell-derived neutrophil-activating peptide (ENA-78), or
growth-regulating oncogene, GRO-, was perfused over these
neutrophils, they stopped rolling and rapidly migrated over the
monolayer, but did not penetrate it. When HUVEC were treated with 100
U/ml TNF, the majority of adherent neutrophils transmigrated. If
neutrophils were treated with fMLP, IL-8, C5a, ENA-78, or GRO- just
before perfusion over this HUVEC, transmigration, but not adhesion, was
abolished. However, when platelet-activating factor was used to
activate neutrophils, migration through HUVEC treated with 100 U/ml TNF
was not impaired, and migration through HUVEC treated with 2 U/ml TNF
was actually increased. Transmigration required ligation of CXC
chemokine receptor-2 on neutrophils, and differential desensitization
of this receptor (e.g., by fMLP but not platelet-activating factor) may
explain the pattern of disruption of migration. Thus, transmigration
may require presentation of the correct activators in the correct
sequence, and inappropriate activation (e.g., by systemic activators)
could cause pathological accumulation of neutrophils in the vessel
lumen. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and IL-1 induce the endothelial cells to express E-selectin and
P-selectin, which mediate capture and rolling adhesion of flowing
neutrophils (see Refs. 1, 2, 3 for review). Rolling adhesion
allows interaction of neutrophils with surface-presented chemotactic
agents such as platelet-activating factor
(PAF)3 and IL-8, which
induce activation of ß2 integrin receptors. The
ß2 integrins bind to endothelial members of the Ig gene
superfamily (IgSF; e.g., intercellular adhesion molecule-1) to
immobilize the rolling cells and also support onward migration
(4). It remains uncertain how neutrophils are guided to
move through the endothelial monolayer and into tissue, although it
seems that chemotactic agents work in concert with the IgSF member CD31
to regulate migration (5, 6, 7). Uncontrolled adhesion and migration of neutrophils may be pathogenic (8, 9, 10). Two main scenarios can be envisaged: continuous or untimely entry of neutrophils into tissue may lead to degradative damage, or failure of transmigration of neutrophils bound to the endothelial surface could cause vascular occlusion and damage if the cells released toxic oxidants and granule contents. The second of these may occur in conditions such as vasculitis and endotoxic shock, where systemic activation of neutrophils may occur along with deposition in the lumen of microvessels (11, 12). This concept gains support from findings that infusion of IL-8 or fMLP into rabbits actually impaired neutrophil transmigration induced by separate localized application of chemotactic agents (13, 14), although these effects were linked to the ability of these agents to inhibit adhesion to stimulated endothelium (15, 16) rather than to disruption of migration per se. IL-8 is generated by endothelial cells exposed to cytokines or to hypoxia and reoxygenation, and in these cases it has been shown to promote neutrophil activation, adhesion, and migration (17, 18, 19). Thus, it may be hypothesized that for efficient migration, activating signals must have a correct sequence or mode of presentation (e.g., on the endothelial surface rather than in the fluid phase), and that exposure of neutrophils to activating agents in the blood might cause them to fail to cross the endothelial barrier.
We set out to investigate these possibilities by characterizing the
migration of neutrophils through endothelial monolayers stimulated with
TNF, with and without exposing the neutrophils to exogenously added
activating peptides (IL-8, fMLP, activated complement fragment (C5a),
epithelial neutrophil-activating peptide (ENA-78), growth-regulating
oncogene protein (GRO-)) or phospholipid-derived PAF. Judging
from measurements of changes in intracellular
Ca2+, the peptides IL-8, fMLP, and C5a have been
shown previously to cause homologous desensitization of responses
through their own receptors, and heterologous desensitization of
responses through each other’s receptors and through the PAF receptor
(20, 21). Effectiveness was in the order fMLP >
C5a > IL-8, and recent studies suggest that these peptides can
also inhibit adhesion and migration induced by each other with a
similar hierarchy (22, 23). Desensitization of the
response to IL-8 may arise from down-regulation of its two CXC
chemokine receptors, CXCR1 and CXCR2 (24, 25, 26). However,
the effects of such agents on responses to endogenous surface-presented
activators, expressed by cytokine-treated endothelial cells, and on the
different steps during transendothelial migration have not been
described. Here, we used a flow-based model (19, 27) to
allow separate investigation of the steps in neutrophil migration, to
distinguish between the effects of fluid phase and surface-presented
activating agents, and to avoid the possibility of build-up of agents
released by endothelial cells or neutrophils that might influence
responses.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Blood was collected from healthy volunteers into EDTA tubes (Sarstedt, Numbrecht, Germany), and neutrophils were isolated using two-step density gradients, as previously described (19, 27). The neutrophils were washed twice in PBS containing 1 mM Ca2+ and 0.5 mM Mg2+ (Sigma, St. Louis, MO), 0.1% BSA (Sigma), and 5 mM glucose (PBS/BSA) and adjusted to 1 x 106/ml in the same medium. In chosen experiments, neutrophils were treated with mAb against CXCR1 or CXCR2 (9H1 and 10H2, respectively; 10 µg/ml; gift from Dr. K. Jim Kim, Genentech, San Francisco, CA) for 20 min before assay.
To prepare zymosan-activated plasma (ZAP) as a source of activated
complement fragment C5a, blood was collected from healthy volunteers
into sodium heparin (CP Pharmaceuticals, Wrexham, U.K.). Platelet-poor
plasma was separated by centrifugation at 1000 x g for
5 min. The plasma was then incubated with zymosan (8 mg/ml) for 30 min
at 37°C, followed by centrifugation at 10,000 x g
for 5 min and filtration through a sterile 0.2-µm pore size filter.
ZAP was used at 1% in PBS/BSA to activate neutrophils. Other
activating agents were PAF (Sigma), fMLP (Sigma), and recombinant human
ENA-78, GRO-, and monocyte-derived IL-8 (R&D Systems,
Abingdon, U.K.; see Results for concentrations).
Endothelial cell culture
Endothelial cells were isolated from the veins of human
umbilical cords as previously described (28) and cultured
in medium 199 (ICN Flow Laboratories, Thame, U.K.) containing 20%
heat-inactivated normal human serum (National Transfusion Service,
Birmingham, U.K.), 50 U/ml sodium heparin (CP Pharmaceuticals, Wrexham,
U.K.), 25 µg/ml gentamicin (David Bull Laboratories, Warwick, U.K.),
and 1 mg/ml glutamine (Sigma) until confluent (
4 days). Adhesion
assays were conducted using second-passage endothelial cell cultures
grown on one inner surface of glass capillary tubes (microslides) as
previously described (27, 28). Microslides (Camlab,
Cambridge, U.K.) have a rectangular cross-section of 0.3 x 3 mm,
a length of 5 cm, and good optical qualities. They were treated with
3-aminopropyltriethoxysilane (Sigma) and coated with a mixture of
gelatin (1% type B from bovine skin; Sigma) and collagen (2 x
10-4 M human type IV; Sigma) before seeding with
HUVEC at a level that gave a confluent monolayer within 24 h
(27). Confluent HUVEC in microslides were stimulated with
2 or 100 U/ml of recombinant human TNF- (Genzyme, West Malling,
U.K.) for 4 h before assays (27). Each experiment was
conducted using HUVEC from a single primary culture from a single
donor.
Adhesion and migration of neutrophils under conditions of flow
The adhesion and migration assay was recently described (27). Microslides containing confluent HUVEC were mounted on the stage of a phase-contrast, video microscope enclosed in a Perspex chamber (Wolfson Applied Technology Laboratory, Birmingham, U.K.) at 37°C. They were attached at one end, via flexible silicon tubing, to a Harvard syringe pump (Harvard Apparatus, Natick, MA), and washing buffer or a suspension of neutrophils was selectively drawn through the microslides via an electronic switching valve (Lee Products, Gerards Cross, U.K.) attached to the other end. A constant wall shear stress of 0.1 Pascal unit was maintained in the microslide by choice of the appropriate flow rate. This stress is comparable to that found in postcapillary venules (see Ref. 29 for review) and does not allow direct adhesion of neutrophils to endothelial cells in the absence of selectins (19).
HUVEC was perfused with PBS/BSA to remove excess TNF. Neutrophil suspension was flowed over the HUVEC for 4 min, followed by a period of washout with cell-free PBS/BSA. A video record of adhesive behavior was made during inflow of neutrophils and at intervals during washout for a series of fields along the center line of the microslide. After 1 min of washout, the number of adherent cells was counted in at least four complete microscopic fields of known dimension and converted to cells per square millimeter per 106 perfused.
Video microscopic recordings of neutrophils adherent to the endothelial
cell layer were analyzed off-line. Three populations of adherent
neutrophils could be distinguished: 1) rolling adherent cells, which
were spherical, and rotated slowly over the surface at a velocity
(10 µm/s) far below that of free flowing cells (700 µm/s); 2)
activated neutrophils, which were phase bright with distorted shape,
either stopping or migrating on the top of the endothelial monolayer;
and 3) transmigrated neutrophils, which were phase dark and migrated
underneath the monolayer. The percentage of adherent cells in each
category was determined, and the velocity of rolling cells was measured
from digitized images and custom software (30). The
velocity of migrating cells was also measured (7). The
outlines of neutrophils were traced onto acetate sheets overlying the
video monitor. This process was repeated as the video record was moved
forward in 1-min intervals, and the distance advanced by the leading
cellular margin was measured. The screen size was calibrated using a
stage micrometer, and velocity (microns per minute) was averaged over
3–5 min of movement.
Disruption of neutrophil adhesion and migration by activating agents
Experiments with 2 U/ml TNF.
To examine the effect of activating agents on neutrophils already
adherent to HUVEC but predominantly rolling, a low concentration of TNF
(2 U/ml) was used to activate endothelial cells. Neutrophils were
perfused over HUVEC for 4 min. After 1 min of washout, activating
agents (fMLP, IL-8, C5a, PAF, ENA-78, or GRO-) in PBS/BSA were
perfused continuously over the already adherent neutrophils, and their
behavior was recorded. To test whether activation was reversible, fMLP
was perfused for 1 min over adherent neutrophils and then the
microslides were perfused with PBS/BSA alone for up to 60 min. In some
experiments, IL-8 was perfused over the adherent neutrophils for 1 min,
and then IL-8 and PAF were perfused together for 10 min.
Experiments with 100 U/ml TNF.
To examine the effect of preactivation of neutrophils on subsequent
migration, HUVEC were stimulated with 100 U/ml TNF, which induced a
high proportion of adherent neutrophils to migrate. Activating agents
(fMLP, IL-8, C5a, PAF, ENA-78, or GRO-) were added to neutrophils
30 s before perfusion over the endothelial cell monolayer and were
perfused along with the neutrophils. fMLP, IL-8, C5a, and PAF were also
included in the PBS/BSA used to washout neutrophils for 10 min after
the bolus. Because of the scarcity of ENA-78 and GRO- and the need
to use relatively high concentrations (see Results), these
agents were not included in the washout PBS/BSA. To specifically test
reversibility, fMLP was perfused with the 4-min bolus of neutrophils,
but then PBS/BSA was perfused alone for up to 60 min. In some
experiments where neutrophils were treated with the mAb 10H2 against
CXCR2 before perfusion, PAF was added to the washout PBS/BSA after
completion of the 4-min bolus.
Surface expression of CD11b and CXC chemokine receptors
The effect of activation of neutrophils on their surface expression of CD11b was analyzed by direct immunofluorescence and flow cytometry using an R-PE-conjugated mAb against CD11b (R841, Dako, High Wycombe, U.K.) as previously described (31). The expression of CXCR1 and CXCR2 was analyzed by indirect immunofluorescence, using 9H1 and 10H2 as primary Abs and FITC-conjugated goat anti-mouse (Dako) as secondary Ab. In every case, the intensity of fluorescence labeling and the percentage of cells positively labeled were measured relative to those of cells labeled using a nonspecific conjugated or unconjugated mouse Ab (Dako) as appropriate. Binding of IL-8 to the surface of neutrophils was quantified by flow cytometry using biotin-conjugated IL-8 and avidin-FITC, supplied as a kit (Fluorokine, R&D Systems), which was used according to the manufacturer’s instructions.
Statistical analysis
Effects of time and treatments were tested using two-way ANOVA. Comparison of individual treatments at single time points were made using paired t test when appropriate. All tests were performed using the computer program Minitab (Minitab, State College, PA).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
), but did not attach to untreated HUVEC
(<1% of binding to TNF-treated cells). At 2 U/ml TNF, the majority of
adherent neutrophils were rolling, and about 20% gradually migrated
through the monolayer (Fig. 1). At 100
U/ml TNF, relatively few adherent neutrophils rolled, and 50–60%
typically transmigrated within about 5 min of attachment (Fig. 1).
These results agree with our previous reports (27, 32).
|
|
).
Transendothelial migration was thus markedly reduced. If neutrophils
were treated with fMLP, IL-8, or C5a immediately before perfusion over
HUVEC that had been treated with 100 U/ml TNF, the cells immobilized
immediately on landing, and rolling was absent. The activated cells
migrated over the surface, but again, transmigration was almost
abolished (Fig. 3). When fMLP was washed
out from neutrophils adherent to HUVEC treated with either 2 or 100
U/ml TNF, during the first 10 min the neutrophils still failed to
transmigrate (Fig. 4) but remained
activated and did not recommence rolling. If washout was prolonged for
30–60 min, transmigration still failed to occur with either 2 or 100
U/ml TNF, but an increasing number of cells tended to round up in shape
and recommence rolling (data not shown).
|
|
|
). When IL-8 was
perfused over neutrophils adherent to HUVEC treated with 2 U/ml, 1
ng/ml had no evident effect on the behavior of adherent neutrophils,
while 100 ng/ml inhibited transmigration in a similar manner as 10
ng/ml (data not shown).
|
A). When
neutrophils were treated with PAF just before perfusion over HUVEC
treated with 100 U/ml TNF, transmigration was not significantly
affected (Fig. 6B).
|
). Moreover, the agents actually increased the speed of migration of
neutrophils on top of the HUVEC treated with either dose of TNF (Fig. 7).
|
). A combination of
the Abs had no additional effect (data not shown). mAb 9H1 and 10H2
have previously been shown to block approximately 90% of the binding
of IL-8 to cells transfected with CXCR1 or CXCR2 individually and
blocked 20 and 50% of IL-8 binding to neutrophils, respectively
(34, 35). We achieved similar levels of inhibition of
binding of biotinylated IL-8 with the individual mAbs (24 and 44%;
mean from three experiments), and maximal inhibition of binding was
67% when they were used in combination. By immunofluorescence, we
verified that 98 and 86% of neutrophils were positive for binding of
9H1 and 10H2, respectively (means from three experiments). We also
examined their ability to block activation by IL-8 itself, judged by an
increase in the expression of CD11b. IL-8 (10 ng/ml for 20 min)
increased the expression of CD11b by 102 ± 26%, while after
treatment with a combination of both Abs the increase was 77 ±
19% (mean ± SEM from four experiments; p >
0.05). The individual Abs did not inhibit the increase in expression of
CD11b in response to IL-8. Finally, we checked whether the mAbs
themselves caused neutrophil activation. Neither mAb caused an increase
in CD11b expression (expression was 94 or 97% of the untreated control
value for 9H1 or 10H2 respectively; means from two experiments with
each mAb).
|
are known to be able to act through this receptor, with a
Kd lower than that for CXCR1, and to
be chemotactic for neutrophils (36, 37, 38). Their effects on
adhesion of flowing neutrophils have not been described. First, we
tested their ability to up-regulate the expression of CD11b. At 1
µg/ml (100 nM), each caused an increase in CD11b similar to that
found with 10 ng/ml IL-8 (99, 94, and 96% increase with IL-8, ENA-78,
and GRO-, respectively; means after 20-min treatment in three
comparative experiments). We also checked whether mAb 9H1 or 10H2
inhibited this up-regulation. Perhaps surprisingly, either Ab reduced
up-regulation by about 50% for each chemokine (pooling data for two
tests with each Ab on each chemokine; inhibition average, 53%).
However, it should be noted that at 100 nM, each chemokine is at a
concentration greater than its Kd for
CXCR1 or CXCR2 (36), so receptor selectivity may not be
expected. A combination of Abs reduced up-regulation of CD11b by 80%
for ENA-78 and 56% for GRO- (means from two experiments each).
When either ENA-78 or GRO- was perfused at 1 µg/ml over
neutrophils already adherent to HUVEC that had been treated with 2 U/ml
TNF, rolling rapidly converted to stationary adhesion, and cells
changed shape and migrated over, but not through, the monolayer (Table II). When neutrophils were pretreated and
perfused with these agents over HUVEC that had been treated with 100
U/ml TNF, they adhered without rolling and were motile, but failed to
transmigrate efficiently (Table II). It was notable in the experiments
with HUVEC treated with 100 U/ml TNF that when the agents were washed
out after the neutrophil bolus, some rolling and transmigration was
re-established within 10 min, especially for ENA-78 (Table II).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
were each capable of causing such disruption, PAF was not. The
results imply that successful transmigration may depend on the order or
mode of presentation of chemotactic agents and on the specific agent
presented, and that incorrect presentation in vivo could cause
neutrophils to be trapped in vessels.
The adhesion molecules by which neutrophils bind to cytokine-treated
endothelial cells have been well described (2, 3), but the
signals that regulate migration over and through the monolayer are less
clear. In the present model using TNF-stimulated HUVEC, initial
attachment from flow and subsequent rolling adhesion were supported by
endothelial E- and P-selectin, and immobilization and migration were
dependent on neutrophil ß2 integrin
(27). Migration on endothelial cells has previously been
shown to be supported by interaction between ß2
integrins and ICAM-1, ICAM-2, and possibly additional receptors on
endothelium (4, 40). In general, neutrophils can migrate
along a concentration gradient of chemotactic agent (41)
at a rate that depends on the steepness of the gradient
(42), although a gradient of a soluble agent(s) can hardly
exist in the lumen of a perfused blood vessel. In flow models, the rate
and direction of migration driven by a uniform concentration of
chemotactic agents can be modulated by ligation of neutrophil CD31 and
vß3 integrin,
respectively (7, 43). Others have shown that endothelial
transmigration is regulated through ligation of CD31 both in vitro and
in vivo (5, 6). Thus migration over and through
endothelium may require adhesive signals acting in concert with
chemotactic or chemokinetic stimulus. Moreover, it is clear that
activation per se is not adequate to cause transmigration, because
exogenous agents used here caused immobilization of rolling neutrophils
and increased the speed of subsequent migration, but disallowed
diapedesis. Our results raise questions about the specificity,
sequence, mode of presentation, and concentration of agent(s) required
for this step.
Considering undisturbed migration through TNF-treated HUVEC first,
others have implicated IL-8 and PAF as neutrophil activators (17, 44, 45). However, in our previous flow-based studies using
TNF-stimulated HUVEC, treatment of neutrophils with PAF receptor
antagonists or of HUVEC with neutralizing Ab against IL-8 had no effect
on migration (27). Here, we found that blockade of CXCR2
(IL8RB), but not CXCR1 (IL8RA), greatly reduced both transmigration and
conversion to stationary adhesion. CXCR-2 is the higher affinity IL-8R
(36) and has been specifically implicated in the promotion
of cell migration (46). ENA-78 and GRO- can also
activate neutrophils through CXCR2 and induce migration, and have
higher affinity for CXCR2 than CXCR1 (36, 37, 38). Both ENA-78
and GRO- can be produced by activated endothelial cells (47, 48). We found that either could cause up-regulation of CD11b and
immobilization of rolling neutrophils as efficiently as IL-8, albeit at
a much higher concentration (1 µg/ml). At this concentration their
up-regulation of CD11b did not appear to be specifically through CXCR2,
judged from blockade with Abs against CXCR1 and CXCR2. However, if
presented on HUVEC at low concentration, ENA-78 or GRO- should
selectively act through CXCR2 and might then induce neutrophil
activation and migration. To date, we have not been able to determine
whether they are present or active on TNF-treated HUVEC. We conclude
that when endothelial cells are exposed to TNF, neutrophils are
captured by selectins and then use CXCR2 to ligate a surface-bound,
unidentified CXC-chemokine(s) that induces integrin activation. We
cannot rule out the possibility that separate agents promoted
immobilization and transmigration. It was notable, however, that of
those neutrophils that did become activated after blockade of CXCR2,
few went on to transmigrate.
Considering disruption of migration by exogenous peptide agents next,
this may be explained by their desensitization of the response of
neutrophils through CXCR2. Each of fMLP, IL-8, and C5a has been shown
to desensitize the response of neutrophils to IL-8 (20, 21), which implies loss of function of both CXCR1 and CXCR2.
GRO- has been shown to reduce surface expression of CXCR2, but not
CXCR1 (24). The concentration of fMLP or IL-8 that we
found to be required for disruption of transmigration
(
10-9 M) is close to that previously found to
induce effective homologous and heterologous desensitization
(21). PAF does not desensitize the response of neutrophils
through any of the peptide receptors (21), and this would
explain its failure to inhibit transmigration. All the exogenous
activators themselves clearly enabled immobilization and migration of
neutrophils on the endothelial surface, while only transmigration
failed. Thus, if desensitization is the cause of disrupted
transmigration, CXCR2 must have been required for this later step as
well as for immobilization.
While desensitization may explain why exogenous agents inhibited response to the endogenous inducer(s) of transmigration, it does not explain why added agents did not themselves cause transmigration. This also raises the question of how migration through the endothelium is guided, as opposed to migration over it. Each of the disrupting agents tested here can induce neutrophil motility and chemotaxis (7, 37, 38, 41). While the ability to induce transmigration through a monolayer might be specific to certain agents, several tested here are able to act through CXCR2, which appears to be the endogenous inducer of transmigration. An additional degree of specificity must exist, arising, for instance, from the ability to bind to the endothelial surface. This might explain why fMLP or C5a failed to induce transmigration. However, this ability alone may not be sufficient. We previously found that IL-8 could bind to HUVEC when added exogenously and subsequently induce activation of rolling neutrophils (49), but this agent did not allow transmigration here. The order of the signals might also be critical in the sense that early exposure to an agent may desensitize the response to others or even to the same agent required for transmigration. However, since transmigration was disrupted even when agents were added to neutrophils already adherent, this cannot be the full explanation. It seems therefore that an agent presented from solution is unable to cause transmigration, even if that agent has the potential to induce migration, and can bind to the surface. One possibility is that presentation of an endogenous agent(s) is localized, for instance to endothelial cell margins, and that this guides transmigration. In this scenario, uniform ambient or surface presentation of any chemotactic agents would not be able to induce transmigration.
The concentration of agents may also be critical. For example, at 10-9 M, fMLP could activate most neutrophils, as judged by cessation of rolling and shape change, but did not affect the percentage transmigrated. At 10-7 M fMLP, transmigration was lost. With increasing concentrations of agents, different thresholds may be reached where cells first become activated but not refractory to other stimuli, and then become refractory at a higher concentration. The loss of response might be through concentration-dependent desensitization of receptors, but it is also possible to envisage that exposure to a high level of ambient chemoattractant simply over-rules a directional stimulus that is still in existence. A linked possibility is that a guidance system into the wall would be disrupted because the presence of soluble agents in the lumen would effectively set up a chemotactic gradient in the opposite direction. However, inhibition of transmigration was still evident when fMLP was washed away and indeed appeared unrecoverable. Moreover, recent studies show that at least in the case of IL-8, neutrophils can move down a concentration gradient when exposed to a positive gradient of a second chemotactic agent such as leukotriene B4 (50). fMLP did, however, inhibit migration toward agents such as IL-8 in that (50) and other (22) studies. Thus, disruption of migration may depend on the position of competing agents in a hierarchy of potency of heterologous desensitization, to date defined as fMLP > C5a > IL-8 >> PAF (21, 22, 50), on the concentrations of the agents, and on the order in which they are presented.
The failure of PAF to disrupt migration may be explained by its position in this hierarchy of desensitization. In fact, there is reason to believe that PAF can positively promote transmigration. For instance, when generated and presented by endothelium, PAF appears to induce transmigration as opposed to immobilization of flowing neutrophils (19, 51). In the present studies exogenous PAF actually augmented diapedesis through HUVEC treated with 2 U/ml TNF, but had little effect with 100 U/ml TNF. If added PAF can itself induce transmigration, its failure to do so at 100 U/ml TNF may have because 1) a strong migratory stimulus was already present, and PAF could not add to this; 2) the endogenous activator actually desensitized the migratory ability of PAF; 3) PAF in the soluble phase activated neutrophils but could not induce transmigration (i.e., mode of presentation was crucial); or 4) PAF was already present, as suggested by some studies (44, 45). The last possibility is not consistent with failure of the residual immobilized cells to migrate after Ab blockade of CXCR2 or of PAF receptor antagonist to affect migration in our hands. To investigate the potency of PAF further, we treated neutrophils with anti-CXCR2 to block their activation and transmigration on HUVEC treated with 100 U/ml TNF, but added PAF to the adherent cells. We were surprised to find that PAF did not induce transmigration. This is not consistent with explanations 1 and 2 above and strongly suggests that exogenous PAF added in solution cannot induce transmigration. We also added PAF to neutrophils adherent to HUVEC treated with 2 U/ml TNF shortly after IL-8 had been added to inhibit transmigration. Here, however, the inability of PAF to cause transmigration may be explained by the previously described ability of IL-8 to desensitize the PAF receptor (21). The finding that exogenous PAF did promote transmigration through HUVEC treated with 2 U/ml TNF may be explained if the PAF did not itself drive the migration but added to the effectiveness of an existing agent. Thus, PAF caused immobilization that may have aided the action of the endogenous inducer of transmigration (i.e., the percentage of cells becoming immobilized was increased by PAF, but the percentage of immobilized cells that then transmigrated was not). Hence, even in the case of PAF, which can induce transmigration under the correct circumstances (19, 51), a specific surface distribution might be important.
Others have previously shown that infusion of IL-8 or fMLP into rabbits can cause a loss of ability of neutrophils to migrate into tissue (13, 14), although interpretation of these results is made difficult because inhibition occurred at a stage after the agent itself, and the original circulating neutrophils had been cleared from the blood. Results from stationary adhesion assays indicated that added IL-8 actually inhibited the adhesion of neutrophils to cytokine-treated endothelial cells (15, 16), although various studies using stationary and flow assays indicate that endothelial-generated IL-8 promotes adhesion (17, 18, 19). In the present study capture of neutrophils on TNF-treated HUVEC was selectin mediated (27), and pretreatment of neutrophils with IL-8, fMLP, C5a, or PAF had little effect on this process. In addition, already adherent cells treated with these agents remained attached. In either case, neutrophils actually migrated more rapidly on the endothelial monolayer after addition of the agents. Thus, disruption of transmigration cannot be attributed to loss of adhesive or motile capability. The apparent discrepancy between studies of the effects of IL-8 on adhesion may be explained by variations in time course. Activated neutrophils gradually shed L-selectin and also down-regulate adhesion through P-selectin glycoprotein ligand 1 after activation (52, 53) and may thus decrease their capture by endothelium with time. We detected relatively little loss of adhesion, presumably because cells either contacted HUVEC within a short period of exposure to exogenous activators or were already bound at the time of exposure. In either case neutrophils could use integrins to become firmly bound. We did not use prolonged pre-exposure in our studies, as we considered a short time scale more appropriate to the circulation time of the blood in vivo of about 1 min.
Disordered migration may be pathogenic. Neutrophils adherent to the surface of endothelial cells but unable to transmigrate out of microvessels will increase resistance to flow (54) and also have the potential to release lytic enzymes and reactive oxygen species. We did not test whether these potentially damaging compounds were released from cells migrating over HUVEC, but our recent studies in a flow model did show that elastase was released from adherent cells exposed to fMLP, C5a, and IL-8, but not PAF (33). Agents such as fMLP and C5a should not normally be encountered in the vascular lumen, and their immunological roles may reside in their ability to cause extravascular chemotaxis and to encourage bacterial killing and tissue remodeling. However, local or systemic release into the blood may occur in situations such as septic shock or autoimmune disease. High levels of IL-8 have been associated with a range of vascular pathologies (55, 56), while anti-neutrophil cytoplasm Abs associated with vasculitis may cause trapping of neutrophils in microvessels (12) and can inhibit chemotaxis of neutrophils by causing them to release IL-8 (57). Thus, a range of immune mediators required for host protection may actually cause tissue damage if they are released at times, concentrations, or sites that are inappropriate and then cause intravascular trapping of neutrophils.
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. G. B. Nash, Department of Physiology, The Medical School, University of Birmingham, Birmingham, U.K. B15 2TT.
3 Abbreviations used in this paper: PAF, platelet-activating factor; IgSF, Ig gene superfamily; C5a, activated complement fragment 5; CXCR1, CXC-chemokine receptor type 1; ENA-78, epithelial neutrophil-activating peptide-78; GRO-, growth-regulating oncogene protein ; PBS/BSA, PBS with albumin; ZAP, zymosan-activated plasma.
Received for publication April 30, 1999. Accepted for publication March 21, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-induced transendothelial neutrophil migration is IL-8 dependent. Am. J. Physiol. 266:L238.-induced neutrophil transcellular migration. Am. J. Respir. Cell Mol. Biol. 13:323.[Abstract]
. J. Immunol. 160:4518.. Cell Adhes. Commun. 6:491.[Medline]
(TNF). J. Vasc. Res. 36:477.[Medline]
, GRO ß, GRO 
, neutrophil-activating peptide-2, and epithelial cell-derived neutrophil-activating peptide-78 are potent agonists for the type B, but not the type A, human interleukin-8 receptor. J. Biol. Chem. 271:20545.vß3-integrin. Am. J. Physiol. 276:H858.This article has been cited by other articles:
|
|
 |
|
  J. A. Riffell and R. K. Zimmer Sex and flow: the consequences of fluid shear for sperm egg interactions J. Exp. Biol., October 15, 2007; 210(20): 3644 - 3660. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. M. McGettrick, J. M. Lord, K.-Q. Wang, G. E. Rainger, C. D. Buckley, and G. B. Nash Chemokine- and adhesion-dependent survival of neutrophils after transmigration through cytokine-stimulated endothelium J. Leukoc. Biol., April 1, 2006; 79(4): 779 - 788. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. D. Buckley, E. A. Ross, H. M. McGettrick, Chloe. E. Osborne, O. Haworth, C. Schmutz, P. C. W. Stone, M. Salmon, N. M. Matharu, R. K. Vohra, et al. Identification of a phenotypically and functionally distinct population of long-lived neutrophils in a model of reverse endothelial migration J. Leukoc. Biol., February 1, 2006; 79(2): 303 - 311. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. D. Ciornei, T. Sigurdardottir, A. Schmidtchen, and M. Bodelsson Antimicrobial and Chemoattractant Activity, Lipopolysaccharide Neutralization, Cytotoxicity, and Inhibition by Serum of Analogs of Human Cathelicidin LL-37 Antimicrob. Agents Chemother., July 1, 2005; 49(7): 2845 - 2850. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. H. Tarlowe, A. Duffy, K. B. Kannan, K. Itagaki, R. F. Lavery, D. H. Livingston, P. Bankey, and C. J. Hauser Prospective Study of Neutrophil Chemokine Responses in Trauma Patients at Risk for Pneumonia Am. J. Respir. Crit. Care Med., April 1, 2005; 171(7): 753 - 759. [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]
|
|
|
|
 |
|
 G. Cinamon, V. Shinder, R. Shamri, and R. Alon Chemoattractant Signals and {beta}2 Integrin Occupancy at Apical Endothelial Contacts Combine with Shear Stress Signals to Promote Transendothelial Neutrophil Migration J. Immunol., December 15, 2004; 173(12): 7282 - 7291. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Sheikh, R. S. Parhar, R. Bakheet, S. Saleh, K. Collison, and F. Al-Mohanna Immobilization of rolling NK cells on platelet-borne P-selectin under flow by proinflammatory stimuli, interleukin-12, and leukotriene B4 J. Leukoc. Biol., September 1, 2004; 76(3): 603 - 608. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Sheikh, G. E. Rainger, Z. Gale, M. Rahman, and G. B. Nash Exposure to fluid shear stress modulates the ability of endothelial cells to recruit neutrophils in response to tumor necrosis factor-{alpha}: a basis for local variations in vascular sensitivity to inflammation Blood, October 15, 2003; 102(8): 2828 - 2834. [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]
|
|
|
|
|
|
N. Fazal, W. M. Al-Ghoul, M. A. Choudhry, and M. M. Sayeed PAF receptor antagonist modulates neutrophil responses with thermal injury in vivo Am J Physiol Cell Physiol, October 1, 2001; 281(4): C1310 - C1317. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Cinamon, V. Grabovsky, E. Winter, S. Franitza, S. Feigelson, R. Shamri, O. Dwir, and R. Alon Novel chemokine functions in lymphocyte migration through vascular endothelium under shear flow J. Leukoc. Biol., June 1, 2001; 69(6): 860 - 866. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) or 100 (
) U/ml TNF. Neutrophils were perfused for 4 min
at a wall shear stress of 0.1 Pascal, followed by washout with PBS/BSA;
the time scale origin coincides with the start of their perfusion. Data
are the mean ± SEM from three experiments.
