| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
* Experimental Cell Research Program, Methodist Research Institute, Indianapolis, Indiana 46202, USA;
Department of Physiology and Biophysics, Indiana University School of Medicine, Indianapolis, Indiana 46202, USA;
Department of Microbiology-Immunology and R. H. Lurie Cancer Center, Northwestern University Medical School, Chicago, Illinois 60611, USA;
§ Department of Biochemistry, Lipid and Lipoprotein Research Group, University of Alberta, Edmonton, Alberta, T6H 5M3, Canada; and
¶ Department of Medicine, Division of Pulmonary and Critical Care Medicine, Johns Hopkins University, Baltimore, Maryland 21205, USA
1Correspondence: Experimental Cell Research Program, Methodist Research Institute, 1701 N. Senate Ave., Indianapolis, IN 46202, USA. E-mail: dkenglish{at}msn.com
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
i-coupled receptor. After S1P-induced
migration, endothelial cells proliferated avidly and differentiated
forming multicellular structures suggestive of early blood vessel
formation. S1P was strikingly effective in enhancing the ability of
fibroblast growth factor to induce angiogenesis in the avascular mouse
cornea. Our results show that blood coagulation initiates endothelial
cell angiogenic responses through the release of S1P, a potent
endothelial cell chemoattractant that exerts its effects by activating
a receptor-dependent process.—English, D., Welch, Z., Kovala, A. T., Harvey, K., Volpert, O. V., Brindley, D. N., Garcia,
J. G. N. Sphingosine 1-phosphate released from platelets
during clotting accounts for the potent endothelial cell chemotactic
activity of blood serum and provides a novel link between hemostasis
and angiogenesis.
Key Words: lipid mediators • vascular endothelium • angiogenesis • endothelial cell migration • hemostasis • S1P
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3
4
5
6
7
8
9
10)
. A complete understanding of endothelial cell
chemotaxis and the factors that induce it are essential for a thorough
understanding of the basic mechanisms of angiogenesis as well as for
the development of methods designed to regulate the response. Several
proteins that mediate endothelial cell chemotaxis, including vascular
endothelial growth factor (VEGF) and fibroblast growth (FGF) factor,
have been identified (11
, 12)
. These proteins are thought
to play a key role in the vascularization and resultant survival of
growing tumors since neutralizing antibodies effect a marked decrease
in cancer progression as a result of inhibition of blood vessel
formation (6
, 13)
. Furthermore, inhibition of angiogenesis
and tumor survival has been observed by antagonizing the specific
receptors for these endothelial cell chemoattractants (7
, 8)
. Consequently, there is tremendous interest in the
possibility of manipulating the angiogenic response in growing tumors
for therapeutic purposes, and several candidate pharmacological agents,
including angiostatin, thrombospondin and endostatin, have been
identified (1
2
3)
.
Much less is known about mediators of the angiogenic response in
injured and traumatized tissue, a response necessary for normal wound
repair. While fluids extracted from human surgical wounds possess
potent angiogenic factors, the response they induce is not consistently
inhibited by neutralizing antibodies to known angiogenic proteins
(9)
. In addition, the physiological role of biologically
active lipids, including those released by activated platelets
(14
15
16
17
18)
, in inducing angiogenic responses of endothelial
cells has not been assessed. Moreover, the factors in blood serum that
are responsible for its ability to attract endothelial cells have not
been identified. We therefore directly examined the hypothesis that
factors released into serum during blood clotting play an important
role in neovascularization by inducing endothelial cell migration
and subsequent angiogenic differentiation. The results implicate
platelet-derived sphingosine 1-phosphate (S1P) as a novel angiogenic
modulator that accounts for the ability of blood serum to attract
endothelial cells and promote angiogenic differentiation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Rat-2 fibroblasts and human embryonic kidney
epithelial cells (HEK-293) were obtained from the American Tissue
Culture Collection (ATCC, Manassas, Va.). Transwell chemotaxis chambers
(6.5 mm diameter, 8 µm pore size) were obtained from Costar, Inc.
(Cambridge, Mass.); Blind Well chemotaxis chambers and 8 µm pore size
filters were obtained from Neuroprobe, Inc. (Cabin John, Md.). Lipids
were dissolved in ethanol at 1 mM. Lyophilized pertussis toxin (List
Biochemicals, Campbell, Calif.) was dissolved in water at 500 µg/ml.
Fetal bovine serum was from Hyclone, Inc. (Logan, Utah).
Charcoal-stripped fetal bovine serum (FBS-C) was prepared by incubating
100 ml FBS with 1.5 g of washed, activated charcoal (Sigma) for
24 h at 4°C to remove S1P and other lipid growth factors
(20)
. After incubation, charcoal was removed by
centrifugation and ultrafiltration. Chemotaxis filters were treated
with 0.2 µg rat tail collagen (Boehringer Mannheim, Indianapolis,
Ind.) and dried overnight, unless indicated otherwise. Tissue culture
media and growth supplements were obtained from Life Technologies, Inc.
(Grand Island, N.Y.). Bovine plasma fibronectin was obtained from Sigma
and dissolved in saline at a concentration of 1 mg/ml. Growth
factor-reduced Matrigel was from Collaborative Biomedical Products
(Bedford, Mass.).
Migration assays
For most experiments, BPAECs between passages 5 and 15 were
dislodged after brief trypsinization and dispersed into homogeneous
cell suspensions, which were washed extensively and resuspended in DMEM
at a concentration of 106 cells/ml. To assess
migration from established monolayers, cells
(105) were dispersed onto collagen-coated
chemotaxis filters within Transwell inserts and allowed to adhere for
30 min at 37°C, when they were challenged by adding 300 µl of a
chemoattractant solution to the lower compartments. Where indicated,
chemotaxis was monitored using Blind Well chemotaxis chambers and 8
µm collagen-coated pore-size filters. In the experiments shown, the
chemotactic activity of lipids, cell extracts and thin-layer
chromatography (TLC) fractions was assayed after reconstitution of
indicated amounts into medium containing 5% FBS-C, which itself
possessed little chemotactic activity (see below). Unless otherwise
indicated, migration was allowed to proceed for 2 h at 37°C.
Cells remaining attached to the upper surface of the filters were
carefully removed with a cotton swab; cells that had migrated to the
lower surface were fixed with formaldehyde, stained with hematoxylin,
and enumerated by microscopic examination. The average number of
migrating cells/field was assessed by counting at least 4 random
fields/filter at 200x magnification. Data points indicate the mean (±
SD) obtained from three separate chambers within one
representative experiment that was repeated at least twice. Although
the relative activity of different samples was consistent among
experiments, the magnitude of optimal responses expressed as the
average number of cells/field varied.
Where indicated, HUVEC, human neutrophils, rat-2 fibroblasts or human embryonic kidney epithelial cells were tested for chemotaxis using methods similar to those described above. HUVECs, fibroblasts and epithelial cells grown on monolayers were dislodged by trypsinization and seeded onto 8 µm collagen-coated filters prior to chemotactic challenge. After 2 h incubation, filters were removed and stained as described above, and migration was compared to that observed with BPAEC. Neutrophils were suspended in Hanks’ balanced salt solution at a concentration of 2 x 107 cells/ml prior to addition to 5 µm Transwell inserts and immediately placed above lower chambers. After incubation for 1 h at 37°C, cells in lower compartments were recovered and enumerated.
In some experiments, migrating cells fixed to the lower surface of filters were permeabilized and stained with FITC-conjugated NBD phalloidin to visualize F-actin. Cells were fixed by immersing the filters in 5% formaldehyde at room temperature. After 15 min, cells were permeabilized by adding 1 mg/ml egg yolk L-lysophosphatidylcholine (Sigma). Permeabilized cells were incubated for 20–30 min with 500 nM NBD-phalloidin (Molecular Probes, Eugene, Oreg.) delivered from a freshly prepared 6.6 µM stock solution in methanol in the dark at room temperature. Filters were then rinsed in absolute ethanol, mounted under immersion oil and examined at 1000x magnification. Photomicrographs were obtained using an Olympus BM-2 fluorescent microscope equipped with a PM-BV3 photomicroscopy system.
In vivo assessment of angiogenesis
Angiogenic responses in the avascular cornea of 7–9 wk old
C57/B16 mice induced by indicated factors embedded in hydron pellets in
the presence of sodium sucralfate were assessed as described (21
, 22)
. Corneal micropockets were created in both eyes of
anesthetized animals with a modified von Graefe cataract knife. Pellets
of 
0.4 x 0.4 x 0.2 mm coated with type NCC hydron
polymer (Interferon Sciences, New Brunswick, N.J.) containing indicated
amounts of test substances or vehicle were implanted 1.2 mm from the
vascularized limbus. Erythromycin ointment was then applied to each eye
and responses were recorded 5 days after insertion of hydron pellets
into corneal micropockets by slit lamp photography. All substances were
tested in three separate experiments.
Endothelial cell morphogenesis on Matrigel
The differentiation of HUVECs into capillary-like networks was
examined using Matrigel supports, as described (23)
. After
trypsinization and washing, 2.5 x 104 cells
in 200 µl RPMI 1640 were plated onto 200 µl of growth factor
reduced Matrigel (12.4 mg protein/ml) in the presence and absence of
2% FBS-C and 100 µM S1P at 37°C in 24-well plastic plates. Sixteen
to 24 h later, cultures were examined microscopically for evidence
of tube formation and photographed (40x). The extent of formation of
capillary-like structures was quantified by analysis of digitized
images to determine the relative area covered by the tube-like network,
using a commercially available image analysis program (Un-scan-it, Silk
Software). Each sample was assayed in triplicate and experiments were
repeated twice.
Serum growth factor levels
Human serum obtained from clotted blood of healthy individuals
was assayed for the presence of various polypeptide growth factors
including VEGF, hepatocyte growth factor (HGF), basic FGF (FGF-b), and
platelet-derived growth factor AB (PDGF) using Quantikine immunoassay
kits from R&D Systems (Minneapolis, Minn.). Levels of each growth
factor were determined in serial dilutions of sera incubated overnight
at 4°C in the presence and absence of activated charcoal (15 mg/ml).
Each sample was spiked with 100 ng/ml FGF-b prior to assay since,
unlike other factors tested, FGF-b was undetectable in fresh serum
samples.
S1P release
Platelet-rich plasma (PRP) was prepared by centrifuging
anticoagulated human blood at 200 x g for 20 min at
20°C The resulting straw-colored supernatants were devoid of
erythrocytes and leukocytes but contained platelets in numbers similar
to those found in unseparated blood (200,000–300,000/µl).
Platelet-poor plasma (PPP) was prepared by centrifuging PRP at
1250 x g for 5 min. PPP and PRP were clotted in glass
tubes by adding 10 mM CaCl2. After clot
retraction at 37°C, samples were centrifuged at 1250 x
g and supernatants were harvested for analysis.
Lipid extraction and chromatography
Lipids were extracted from clotted human plasma with 2 volumes
of 1-butanol in the presence of 0.01 M HCl. Extraction efficiency,
determined with 32P-S1P, was over 80%. Dried
extracts were resuspended in ethanol at 1/30th the volume of the sample
extracted. Solubilized extracts were assayed for chemotactic activity
or further processed by TLC. Lipids spotted on 20 x 20 cm silica
gel G plates were initially separated using a solvent system consisting
of chloroform:methanol:20% methylamine (36:60:10, v/v/v). Plates were
sectioned to obtain lipids covering the entire area of migration and
resolved lipids were eluted with chloroform:methanol:water (10:10:2,
v/v/v). To effect better separation of PA, LPA and S1P, a two-phase
system based on a method described previously (24)
was
used. Extracts spotted in the center of a plastic-backed TLC plate were
chromatographed twice in the first direction using
chloroform/methanol/NH4OH (65:35:7.5, v/v/v),
resulting in the separation of PA, LPA, and S1P from other detectable
lipids. After cutting the plates above the origin to remove other
lipids, phosphatidate, lysophosphatidate, and S1P that remained at the
origin were chromatographed three times in the opposite direction with
chloroform/methanol/acetic acid (8:2:1, v/v/v). This effected clear
separation of phosphatidate, lysophosphatidate, and S1P standards,
which were visualized with molybdenum blue. Areas of the plates
corresponding to the migration position of these standards were excised
and lipids were eluted as described above. Eluates were dried and
resuspended in ethanol for analysis of chemotactic activity or
confirmatory chromatographic analyses, as described below.
Statistical analyses
Data are presented as means ± standard deviation.
Statistical comparison between groups of data were carried out using
the Student’s t test.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
illustrates the release by platelets of a potent
endothelial cell chemoattractant during blood clotting. Supernatants of
human plasma clotted by recalcification in the presence of platelets
induced vigorous endothelial cell migration, even when diluted
100-fold. Plasma clotted in the absence of platelets exhibited
significantly less activity. Optimal levels of migration were induced
when concentrations of activated plasma of 5% were used in lower
compartments of chemotaxis chambers. Heating at 60°C for 30 min did
not significantly decrease the endothelial cell chemoattractive
activity of supernatants of clotted platelet-rich plasma (Fig. 1b
). The majority of this activity was lost, however, when
samples were treated with activated charcoal, a procedure known to
remove lipid growth factors, including LPA and S1P (20)
.
FBS also possessed strong endothelial cell chemotactic activity; this
activity was completely depleted by treating the serum with activated
charcoal (Fig. 1b
).
|
Butanol extracts of chemotactically active plasma were examined for
their ability to restore chemotactic activity to FBS-C, since butanol
extraction, like charcoal treatment, effectively removes biologically
active lipid factors from tissues and fluids (20)
. When
butanol extracts of active plasma were reconstituted into FBS-C, the
resultant mixtures were strongly chemotactic (Fig. 1c
). As
little as 0.1 µl of butanol extract (equivalent to 3 µl of clotted
platelet-rich plasma) promoted migration when introduced into
chemoattractant compartments with a final volume of 300 µl (Fig. 1d
). A butanol extract of clotted platelet-poor plasma
prepared in parallel was far less active. In the experiment shown,
activity observed with 0.5 µl of the active extract was similar to
that obtained with a 5% solution of active plasma prior to extraction.
The chemotactic activity of the butanol extracts was therefore
30-fold stronger than that of the original extracted plasma. Since
extracts were concentrated 30-fold compared to their original volume,
activity within butanol extracts accounted for most, if not all, of the
activity released by platelets during clotting.
Thin-layer chromatographic analysis of the platelet-derived
endothelial chemoattractant
We purified the endothelial cell attractant from active butanol
extracts by 1-dimensional TLC. Over 80% of the reconstituted
chemoattractant activity was recovered from the area of silica gel that
corresponded to the position of an S1P standard (Fig. 2a
). Low levels of activity were recovered in fractions
corresponding to the elution position of phosphatidate, but other
fractions covering the entire range of the resolved plate were devoid
of activity. Markedly less activity was recovered from corresponding
fractions after chromatography of extracts of supernatants of clotted
platelet-poor plasma. Since lysophosphatidate was not clearly separated
from S1P by the solvent system used, we resolved butanol extracts with
a two-phase system designed to separate more effectively phosphatidate,
lysophosphatidate, and S1P while removing other lipids. This approach
demonstrated that more than 95% of the endothelial cell attractant
factor extracted from plasma clotted in the presence of platelets
comigrated with S1P (Fig. 2b
). This activity was clearly
resolved from both lysophosphatidate and phosphatidate. To confirm
further the identity of the resolved entity as S1P, we
rechromatographed the active material recovered from the two-phase
solvent system with two additional solvent systems with different
resolving properties as described in Fig. 2
. In both cases, activity
was recovered only in fractions at the migration position of the S1P
standard (Fig. 2c
). Thus, we established by chromatography
in four different solvent systems that S1P is the predominant
endothelial cell attractant released from platelets during blood
clotting.
|
Chemoattractive activity of other serum lipids
We examined the endothelial cell chemotactic activity of
lipids that may be present in serum or released by platelets in order
to validate the conclusions reached above. Nanomolar concentrations of
synthetic S1P potently stimulated endothelial cell chemotaxis
(Fig. 3a
); concentrations of synthetic S1P as low as 12.5 nM induced
a vigorous response. The chemotactic effect of S1P was not due to a
contaminant in the synthetic S1P preparation used, since S1P recovered
after thin-layer chromatography of the synthetic standard was similarly
active (not shown). Of several nonphosphorylated lipids that may be
present in serum, only anadamide (arachidonylethanolamide) possessed
substantial activity (Fig. 3b
); sphingosine,
12-hydroxyeicosatetranoic acid (HETE), 15-HETE, sphingomyelin,
ceramide, and ceramide-1-phosphate were inactive when tested between
0.5 and 50 µM (Fig. 3b
). Certain lysophospholipids,
including lysophosphatidylinositol and lysophosphatidylethanolamine,
exhibited slight activity at low concentrations (1 µM), but
lysophosphatidate was completely inactive at 1 µM (not shown) and
exhibited very low activity when tested at higher levels. Phosphatidate
and platelet-activating factor possessed weak activity at relatively
high levels. Sphingosylphosphorylcholine, a lipid that induces
responses similar to those induced by S1P when used at higher
concentrations (25)
, was also an effective endothelial
cell chemoattractant, exerting effects 65–70% as well as S1P (500 nM)
when tested at an optimal concentration of 10 µM (data not shown).
|
Effect of charcoal stripping on protein growth factor levels in
serum
Experiments were undertaken to confirm that the sharp
reduction in the endothelial cell chemoattractive activity observed
upon treatment of serum with activated charcoal did not result from
adsorption and removal of protein growth factors. As shown in
Table 1
, overnight treatment with activated charcoal had little effect on
levels of VEGF, FGF-b, PDGF-AB, and HGF in human serum as determined by
immunoassay. In addition, overnight treatment with charcoal had little
influence on the chemotactic activity of solutions of VEGF, FGF-b, and
HGF under conditions where the treatment markedly diminished the
chemotactic potential of solutions of both S1P and dilute
serum (PDGF solutions were found not to possess detectable endothelial
cell chemotactic activity and were thus excluded from this analysis).
In contrast, whereas heating overnight at 60°C had only a moderate
inhibitory effect on the chemotactic potential of either isolated S1P
or dilute human serum, overnight heating almost completely inhibited
the chemotactic activity of solutions of VEGF, FGF-b, and HGF. These
results are consistent with our conclusion that the endothelial cell
chemotactic activity of human serum is due to its content of S1P and
not a result of protein growth factors that may be released from
platelets during blood clotting.
|
Kinetics, characteristics, and specificity of the endothelial cell
response to S1P
Figure 4a
illustrates the dynamics of the vigorous endothelial cell
response to S1P. Endothelial cells migrated rapidly from an established
monolayer through collagen in response to 500 nM S1P reconstituted into
FBS-C. Within 5 h, responding cells formed an almost confluent
monolayer on the chemoattractant side of the micropore filters as actin
stress fibers developed and converged into a contiguous net (Fig. 4b
). The responding cells then began to form multicellular
structures that were well developed by 20 h incubation (Fig. 4a
), suggestive of early angiogenic differentiation.
Consistent with this conclusion, S1P was found to effectively confer on
FBS-C the ability to vigorously promote endothelial cell morphogenesis
into capillary-like structures on Matrigel supports (Fig. 4c
), an assay widely used to define mechanisms and mediators
of angiogenic responses. S1P alone was effective in inducing
endothelial cell migration, but the response was markedly weaker than
that observed when S1P was assayed in the presence of FBS-C (Fig. 4a
), as noted above for assays of activity within butanol
extracts of serum (Fig. 1c
). This difference was ascribed to
factors that enhanced endothelial cell interactions with the
substratum, since the requirement for serum cofactors for optimal
migration was lost when assays were carried out using filters coated
with bovine plasma fibronectin (2 µg) instead of collagen or by
allowing endothelial cells to firmly adhere to and interact with the
substratum for 2 h or longer before chemotactic challenge (data
not shown).
|
The response to S1P in the presence of FBS-C was strongly attenuated
when either S1P or whole FBS was included in both the lower and the
upper compartments of chemotaxis chambers, thereby preventing the
formation of a gradient of chemoattractant across the chemotaxis
filters (Fig. 5a
). Charcoal-treated FBS, however, did not attenuate
responses when added to upper compartments. In addition, S1P did not
induce migration when added to upper compartments of chambers with
FBS-C in the lower compartments. Finally, addition of both S1P and
FBS-C to upper compartments did not induce migration when no additional
factors were present in lower compartments. These experiments
demonstrate that a gradient of S1P is required for optimal induction of
endothelial cell migration and that S1P, but not its serum cofactor,
desensitizes the response. Thus, S1P possesses properties of a true
chemoattractant and induces directed, chemotactic endothelial cell
migration.
|
Experiments were performed to assess the specificity of the chemotactic
response induced by S1P for endothelial cells, species specificity of
the response, and the sensitivity of the response to pertussis toxin.
Preincubation of endothelial cells with pertussis toxin (300 ng/ml) for
3 h at 37°C almost completely inhibited chemotaxis to S1P,
consistent with the involvement of a Gi-linked receptor (Fig. 5b
). HUVECs responded to S1P as vigorously as BPAECs
(Fig. 5b
). Of several cell types tested, including human
neutrophils, rat-2 fibroblasts, and human embryonic kidney epithelial
cells, only endothelial cells migrated in response to S1P (Fig. 5b
). Thus, S1P is a potent and specific chemoattractant that
induces migration of human as well as bovine endothelial cells by
activating a pertussis toxin-sensitive cellular process.
Angiogenic activity of S1P in vivo
Experiments were undertaken to assess the angiogenic activity of
S1P in the mouse cornea micropocket assay, a well-established assay for
assessing the angiogenic potential of various factors in
vivo (21
, 22)
. As shown in Fig. 6
, S1P alone induced little vessel formation in the avascular mouse
cornea. Responses to 10 nmol S1P were similar to those observed with
vehicle (BSA) in the absence of angiogenic factors. FGF-b consistently
induced an angiogenic response and this response was markedly
potentiated by the addition of S1P. Thus, S1P is a potent endothelial
cell stimulus that possesses the ability to promote angiogenic
responses to protein growth factors in vivo.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Activity within butanol extracts of human
serum was recovered after TLC developed with four different solvent
systems in fractions that contained S1P but not LPA, PA, or other
lipids. This activity accounted for virtually all of the activity
present in the original extract. These results strongly indicated that
the major chemoattractant for endothelial cells produced during blood
clotting—and, indeed, the major endothelial cell chemoattractant in
serum—is platelet-derived S1P.
To confirm the conclusion reached above, we tested the action of
synthetic S1P in our system. Purified, synthetic S1P was found to be a
potent endothelial cell chemoattractant. The chemotactic activity of
purified S1P, like that of dilute serum, was heat stable and readily
extracted with activated charcoal. In addition, of several lipids that
may be present in serum, including LPA, PA, and platelet-activating
factor, we found that none possessed the endothelial cell chemotactic
activity that was observed with nanomolar concentrations of S1P. The
results of this report thereby reveal that S1P is the major endothelial
cell chemoattractant released into serum by platelets during blood
clotting. Its activity accounts for almost all of the very potent
endothelial cell chemotactic activity of FBS and a large portion of the
substantial activity within human serum. Though previous reports have
traced other biological activities of serum to the presence of LPA
(14
, 26)
, we believe this study is the first to directly
trace a biological activity generated in serum or other tissue fluid to
platelet-released S1P. The veracity of the response and its cellular
specificity indicate that endothelial cell migration is an important,
physiologically relevant consequence of S1P release from platelets. The
mechanisms responsible for coagulation-initiated S1P release have not
been explored in detail, but we have not been able to attain similar
levels of release by stimulating washed platelets with thrombin, ADP,
Ca2+ ionophore, or other agonists in the absence
of clotting (not shown). Thus, clotting is an extremely effective
trigger of platelet S1P release and may thereby play an important role
in initiating angiogenesis.
Whereas platelet-released S1P may play an important role in the angiogenic response by attracting endothelial cells to the sites of injury and trauma, the lipid may exert additional effects in angiogenic differentiation of the cells after chemotactic migration. We demonstrated that cells that responded chemotactically to S1P began to differentiate into multicellular structures, suggestive of new blood vessel formation. These results must be carefully interpreted because optimal responses were observed in the presence, but not the absence, of FBS-C. Thus, it appears that that serum cofactors play an important role along with S1P in inducing optimal angiogenic responses. Consistent with this conclusion, S1P effectively restored to charcoal-treated FBS the ability to promote endothelial cell morphogenesis into capillary-like structures on Matrigel basement membrane supports. Whereas S1P and FBS-C alone promoted morphogenesis in this assay system, the two agents added together induced an optimal response, consistent with our conclusion that serum cofactors accentuate angiogenic responses to S1P. In vivo, S1P only weakly induced angiogenesis in the avascular mouse cornea, but markedly enhanced responses to FGF-b. It therefore appears likely that S1P is an angiogenic cofactor that exerts optimal effects in conjunction with protein or other lipid growth factors generated in serum and other biological fluids.
Our conclusion that S1P produced during blood clotting acts as an
endothelial cell attractant and angiogenic factor is supported by
recent results demonstrating the induction of endothelial cell
angiogenic responses by purified S1P (27
28
29)
. S1P appears
to induce these responses through activation of members of the EDG
(endothelial differentiation gene) family of receptors. The prototype
of this G-protein-coupled receptor family, EDG-1, was found to be
induced in endothelial cells during an in vitro model of
angiogenesis several years ago (30)
. The subsequent
demonstration that the EDG-1 receptor binds S1P with high affinity
(20
, 31)
led to the hypothesis that S1P plays an important
role in angiogenesis (20)
, but its mechanism of action is
only recently beginning to emerge. In this respect, Wang et al.
recently demonstrated that overexpression of EDG-1 by nonresponsive
cells confers on them chemotactic responsiveness to S1P
(27)
. Endothelial cells that expressed EDG-1 proliferated
in response to S1P. Both human and bovine endothelial cells that
expressed EDG-1 migrated chemotactically to S1P, as shown by
checkerboard analysis, and this response was blocked by pretreatment of
cells with pertussis toxin, consistent with an involvement of EDG-1.
Consistent with this conclusion, a recent report from our group
demonstrated that migration of endothelial cells to S1P was almost
completely inhibited when the cells were preincubated with antisense
oligonucleotides formulated to blunt expression of EDG-1
(29)
. In another recent study, antisense oligonucleotides
formulated to inhibit expression of both EDG-1 and EDG-3, a homologue
of EDG-1 that also binds S1P with high affinity, attenuated the
ability of S1P to induce VE-cadherin localization at endothelial cell
adherens junctions, an early event in angiogenic differentiation
(28)
. Whereas antisense to EDG-1 prevented the formation
of cortical actin structures in S1P-stimulated HUVECs—a process
dependent on activation of the Rac pathway—the formation of stress
fibers (a Rho-dependent process) was specifically inhibited by
antisense to EDG-3. Antisense to EDG-1 and EDG-3 inhibited
endothelial cell morphogenic differentiation induced by S1P in an
additive manner, indicating that this process results from activation
cascades evoked by more than one cellular pathway. Finally, similar to
the results of our study, S1P was found to possess the ability to
potentiate angiogenic responses induced by polypeptide growth
factors in vivo, as assessed using the mouse Matrigel assay.
The authors concluded that endogenous production of S1P may be an
important aspect of the angiogenic response. Our results add an
important extension to this work and demonstrate a novel link between
blood clotting—or hemostasis—and endothelial cell angiogenic
responses.
Although the critical role of angiogenesis in tissue repair after lacerations, abrasions, and traumatic injury appears obvious, the essential mediators of the response have not been identified. The present work demonstrates a direct link between an early yet key component of angiogenesis, directed endothelial cell migration, and blood coagulation. This link is mediated by S1P released from platelets during clotting. Thus, S1P produced by platelets during clotting may play a pivotal role in attracting established endothelial cells to the injured site, an early but essential step in neovascularization and resultant wound healing.
|
ACKNOWLEDGMENTS |
|---|
Received for publication March 30, 2000.
Revision received May 10, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. Heller, Q. Chang, G. Ehrlich, S. N. Hsieh, S. M. Schoenwaelder, P. J. Kuhlencordt, K. T. Preissner, E. Hirsch, and R. Wetzker Overlapping and distinct roles for PI3K{beta} and {gamma} isoforms in S1P-induced migration of human and mouse endothelial cells Cardiovasc Res, July 2, 2008; (2008) cvn159v2. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X. Ye Lysophospholipid signaling in the function and pathology of the reproductive system Hum. Reprod. Update, June 25, 2008; (2008) dmn023v2. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. V. Miller, S. E. Alvarez, S. Spiegel, and D. A. Lebman Sphingosine Kinases and Sphingosine-1-Phosphate Are Critical for Transforming Growth Factor {beta}-Induced Extracellular Signal-Regulated Kinase 1 and 2 Activation and Promotion of Migration and Invasion of Esophageal Cancer Cells Mol. Cell. Biol., June 15, 2008; 28(12): 4142 - 4151. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. K. Wacker, S. K. Alford, E. A. Scott, M. Das Thakur, G. D. Longmore, and D. L. Elbert Endothelial Cell Migration on RGD-Peptide-Containing PEG Hydrogels in the Presence of Sphingosine 1-Phosphate Biophys. J., January 1, 2008; 94(1): 273 - 285. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Roviezzo, A. Di Lorenzo, M. Bucci, V. Brancaleone, V. Vellecco, M. De Nardo, D. Orlotti, R. De Palma, F. Rossi, B. D'Agostino, et al. Sphingosine-1-Phosphate/Sphingosine Kinase Pathway Is Involved in Mouse Airway Hyperresponsiveness Am. J. Respir. Cell Mol. Biol., June 1, 2007; 36(6): 757 - 762. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Zhao, S. K. Kalari, P. V. Usatyuk, I. Gorshkova, D. He, T. Watkins, D. N. Brindley, C. Sun, R. Bittman, J. G. N. Garcia, et al. Intracellular Generation of Sphingosine 1-Phosphate in Human Lung Endothelial Cells: ROLE OF LIPID PHOSPHATE PHOSPHATASE-1 AND SPHINGOSINE KINASE 1 J. Biol. Chem., May 11, 2007; 282(19): 14165 - 14177. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. S. Novgorodov, M. El-Alwani, J. Bielawski, L. M. Obeid, and T. I. Gudz Activation of sphingosine-1-phosphate receptor S1P5 inhibits oligodendrocyte progenitor migration FASEB J, May 1, 2007; 21(7): 1503 - 1514. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Kunisawa, Y. Kurashima, M. Gohda, M. Higuchi, I. Ishikawa, F. Miura, I. Ogahara, and H. Kiyono Sphingosine 1-phosphate regulates peritoneal B-cell trafficking for subsequent intestinal IgA production Blood, May 1, 2007; 109(9): 3749 - 3756. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Ochi, Y. Momose, K. Oyama, and W. R. Giles Sphingosine-1-phosphate effects on guinea pig atrial myocytes: Alterations in action potentials and K+ currents Cardiovasc Res, April 1, 2006; 70(1): 88 - 96. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Kobayashi, T. Nishi, T. Hirata, A. Kihara, T. Sano, Y. Igarashi, and A. Yamaguchi Sphingosine 1-phosphate is released from the cytosol of rat platelets in a carrier-mediated manner J. Lipid Res., March 1, 2006; 47(3): 614 - 621. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Gonzalez, R. Kou, and T. Michel Rac1 Modulates Sphingosine 1-Phosphate-mediated Activation of Phosphoinositide 3-Kinase/Akt Signaling Pathways in Vascular Endothelial Cells J. Biol. Chem., February 10, 2006; 281(6): 3210 - 3216. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Kisucka, C. E. Butterfield, D. G. Duda, S. C. Eichenberger, S. Saffaripour, J. Ware, Z. M. Ruggeri, R. K. Jain, J. Folkman, and D. D. Wagner Platelets and platelet adhesion support angiogenesis while preventing excessive hemorrhage PNAS, January 24, 2006; 103(4): 855 - 860. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Mehta and A. B. Malik Signaling Mechanisms Regulating Endothelial Permeability Physiol Rev, January 1, 2006; 86(1): 279 - 367. [Abstract] [Full Text] [PDF]
|
|
) and absence (•) of
platelets. Indicated amounts of each extract were assayed for
endothelial cell chemotactic activity in the presence of 5% FBS-C in a
final volume of 300 µl.
