1149
High-Density Lipoprotein Stimulates
Endothelial Cell Movement by a Mechanism
Distinct From Basic Fibroblast Growth Factor
Gurunathan Murugesan, Gaurisankar Sa, Paul L. Fox
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
Abstract Endothelial cell (EC) migration is a regulatory
event in the formation and repair of blood vessels. Although
serum contains substantial promigratory activity, the responsible components and especially the role of lipoproteins have
not been determined. We examined the effect of plasma
high-density lipoprotein (HDL) on the movement of ECs in
vitro. Confluent cultures of bovine aortic ECs in serum-free
medium were "wounded," and migration was measured after
24 hours. HDL stimulated migration in a concentrationdependent manner with a half-maximal response at 25 to 40 ,ug
cholesterol per milliliter and a maximal twofold stimulation at
=150 ,ug cholesterol per milliliter. HDL-stimulated migration
was not due to cell proliferation, since migration was increased
in the presence of hydroxyurea at a concentration that blocked
proliferation. At optimal concentrations, HDL was at least as
stimulatory as basic fibroblast growth factor (FGF). However,
the activity of HDL was not due to contamination by basic
FGF, since antibodies to basic FGF did not block HDLstimulated movement and since the maximum promigratory
activities of basic FGF and HDL were additive. These results
indicate that HDL and basic FGF may use distinct signaling
pathways to initiate EC movement. This possibility was confirmed by results showing that pertussis toxin suppressed basic
FGF-stimulated but not HDL-stimulated EC motility and
that inhibitors of phospholipase A2, aristolochic acid and
ONO-RS-082, also blocked the promigratory activity of basic
FGF but had no effect on the activity of HDL. The promigratory activity of HDL may accelerate the regeneration of the
endothelium after a denuding injury in vivo and suggests a new
mechanism by which HDL may be protective against cardiovascular disease. (Circ Res. 1994;74:1149-1156.)
Key Words * basic fibroblast growth factor * cell motility
* endothelial cells * high-density lipoprotein * wound
healing
igration and proliferation of vascular endothelial cells (ECs) are initiating events in the
formation of capillaries during normal development and tumor angiogenesis.' In addition, a continuous and intact monolayer of ECs is required for the
maintenance of normal vessel wall properties, including
impermeability and nonthrombogenicity. However, the
integrity of the endothelium may be transiently or
chronically disrupted by EC turnover, traumatic injury
after balloon angioplasty, vascular grafting, organ implantation, or other pathological damage. Rapid regeneration of the endothelium under these circumstances is
a critical process in the response to the injury. It is
believed that migration is the rate-limiting process in
the repair of endothelium, since it is an early event that
is followed by proliferation.1-4 In fact, increased proliferation may be due to the release from contact inhibition caused by cell migration.25
A number of agents are known to regulate EC
movement in vitro, including vascular permeability factor,6 scatter factor,7 and acidic8 and basic fibroblast
growth factor (FGF).9 Among these, basic FGF is
believed to play a physiological role, since ECs themselves contain and release the factor and since antibodies to basic FGF substantially block migration in vitro,
suggesting that endogenous basic FGF is responsible for
basal EC motility.9 Both basic10 and acidic" FGFs
accelerate reendothelialization of denuded blood vessels after balloon catheterization in vivo. The signaling
pathways that regulate basic FGF-stimulated EC migration are less well understood than those involved in
proliferation. Activation of FGF-receptor tyrosine kinase activity is an early event in the mitogenic response
of cells to basic FGF,12 but its role in migration is not
clear. Among the FGF-stimulated processes that may
be involved in EC movement are transient gene expression, production of plasminogen activator and metalloproteinases, modulation of synthesis of extracellular
matrix components, and cytoskeletal reorganization
(see Reference 13 for review).
Migration of ECs and other cells is markedly stimulated by serum.314 Whole blood serum contains more
EC promigratory activity than plasma-derived serum,
suggesting that platelet factors may be involved,2'3 but
this result is controversial.15 Although the role of highdensity lipoprotein (HDL) and other lipoproteins on
EC movement has not been extensively investigated,
several observations suggest that such an activity is
plausible. The results of Burk et al14 suggest that
lipoproteins may regulate 3T3 cell migration, since
hypercholesterolemic serum has less activity than normal serum. Low-density lipoprotein (LDL) has been
shown to increase reendothelialization of wounds in
vitro, but the stimulation is blocked by x-ray irradiation
of the cells, showing that the increase is due strictly to
the mitogenic activity of LDL.16 HDL has also been
shown to increase reendothelialization during a 5-day
interval after wounding, but the roles of cell prolifera-
M
Received November 15, 1993; accepted February 24, 1994.
From the Department of Cell Biology, Cleveland Clinic Research Institute (Ohio).
Correspondence to Paul L. Fox, PhD, Department of Cell
Biology, Cleveland Clinic Research Institute/NC10, 9500 Euclid
Ave, Cleveland, OH 44195.
1150
Circulation Research Vol 74, No 6 June 1994
tion and migration have not been distinguished.17 In
view of the known mitogenic activity of HDL,18,19 this
distinction is critical. In our previous studies, we have
shown that oxidized LDL is a potent inhibitor of EC
migration in vitro, whereas unmodified LDL is somewhat stimulatory. Given these studies, the continuous
interaction of EC with plasma lipoproteins in vivo, and
the known inverse correlation of HDL with coronary
heart disease,20 we have examined in the present study
the effects of HDL on EC motility and compared its
mechanisms of action with basic FGF, the prototypical
agonist of EC movement.
Materials and Methods
Cell Culture and Media
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
ECs were isolated from adult bovine aortas essentially as
described.21 They were subcultured by trypsinization and
grown to confluence in Dulbecco's modified Eagle medium
and Ham F-12 medium (GIBCO-BRL) containing 5% fetal
bovine serum (Hyclone Laboratories). The identity of the cells
was confirmed by their nonoverlapping cobblestone morphology and by factor VIII-related antigen immunofluorescence.
All EC incubations were at 37°C in a humidified atmosphere of
air containing 5% CO2. For migration studies, the cells were
grown to confluence in serum-containing medium in 12-well
tissue culture clusters (Costar). The cells were made quiescent
by changing the medium to serum-free Dulbecco's modified
Eagle medium (Sigma Chemical Co) containing 1 mg/mL
gelatin (Sigma) for at least 24 hours before use. Cell proliferation was measured as incorporation of [3H]thymidine (73
Ci/mmol, DuPont NEN) into labeled nuclei detected by autoradiography.22 Cellular protein synthesis was measured as
incorporation of [3H]leucine (116 Ci/mmol, DuPont NEN) into
a cell monolayer fixed with trichloroacetic acid.21
Measurement of EC Migration
EC migration was measured by the razor-wound method
essentially as described earlier.23 Confluent and quiescent EC
cultures were wounded with a razor pressed gently through the
cell layer into the plastic well to mark the origin and drawn
through the monolayer to remove cells on one side of the
wound. The medium was replaced with medium containing 1
mg/mL of gelatin plus lipoproteins or other test materials in a
total volume of 0.5 mL. Cell migration was permitted for up to
24 hours and terminated by fixing and staining with WrightGiemsa stain (modified, Sigma). Migration was quantified by a
semiautomated computer-assisted procedure performed by a
person blinded with respect to the experimental treatments. A
256-gray level 640 x480 pixel image of the cultures was
captured by a Sony CCD digital camera mounted on a
phase-contrast microscope and was transferred to a Macintosh
computer. Image and data analysis were done using the IMAGE
software package provided by Dr Wayne Rasband, National
Institutes of Health. In each well, two randomly chosen fields,
each consisting of a 1500-am length of original wound edge,
were chosen; the number of cells that crossed the origin line
was determined; and the results were summed. The data from
duplicate wells were expressed as the mean (±+SEM) number
of migrating cells per 3000 ,um; nearly identical results were
found if expressed as the mean distance.23 All experiments
were done two or more times, and representative results are
shown.
Preparation of Lipoproteins
Lipoproteins were prepared from freshly drawn citrated
normolipemic human plasma to which EDTA was added
before ultracentrifugation. LDLs (density, 1.019 to 1.063
g/mL) and HDLs (density, 1.063 to 1.210 g/mL) were isolated
by sequential ultracentrifugation.24 Some preparations of
100
*
0
o
Serum
LPDS
80
z60
40
20
0
1
2
3
4
5
6
[Serum] (%)
FIG 1. Graph showing the promigratory activity of serum and
lipoprotein-deficient serum (LPDS). The effect of human serum
(c) and serum-matched LPDS (o) on endothelial cell migration
during a 22-hour period was determined as described in "Materials and Methods." The results are expressed as the mean
(+SEM) number of migrating cells per 3000 gm; endothelial cell
migration of untreated cells was 276±9 cells per 3000 ,um.
HDL were further purified by a second ultracentrifugation at
the higher density limit; these lipoproteins yielded results that
were essentially identical to those obtained without the additional step. The homogeneity of lipoproteins was verified by
the appearance of a single band after agarose electrophoresis
and staining with Coomassie blue and oil red 0. All lipoproteins were sterilized by passage through a 0.22-,um filter
(Millipore) before use, and protein concentration,25 total
cholesterol (Boehringer Mannheim), and endotoxin (Pierce
Chemical) were determined for all preparations. Lipoprotein
oxidation was minimized by the addition of 0.3 mmol/L EDTA
to all preparative solutions; the thiobarbituric acid reactivity26
of lipoproteins was low, generally <0.09 and 0.03 nmol malondialdehyde per milligram cholesterol for HDL and LDL,
respectively. Lipoprotein-deficient serum (LPDS) was prepared by adjusting the density of the serum to 1.25 g/mL,
followed by ultracentrifugation.24
Other Materials
Human recombinant basic and acidic FGF and a neutralizing monoclonal anti-bovine basic FGF antibody were from
Upstate Biotechnology. Pertussis toxin, aristolochic acid, and
ONO-RS-082 were from Biomol, and all other reagents were
from Sigma.
Results
We first compared the promigratory effect of serum
with that of LPDS to determine the total contribution of
lipoproteins. Serum at an optimal concentration stimulated EC migration up to 80% more than the control
value, but LPDS had only half the maximum activity of
serum, suggesting that a promigratory agent in serum
was absent in LPDS (Fig 1). To determine the relative
promigratory activities of LDL and HDL, LPDS at a
maximal stimulatory concentration was supplemented
with both lipoproteins at concentrations proportional to
their plasma abundance. Both HDL and LDL increased
the promigratory activity of LPDS (Fig 2). HDL was
significantly more potent than LDL, and at a level
equivalent to that in 7.5% serum, it almost completely
restored EC motility to the level stimulated by whole
serum. However, the difference in activity between
2.5% serum and lipoprotein-deficient serum was not
completely accounted for by the activity of HDL at
2.5% serum equivalents, and it is possible that some
Murugesan et al HDL Stimulates Endothelial Cell Movement
80
c
.2
70
*Basic FGF
0 A~~~~~cidIc FOF
E
c
.5
0
1151
60
U HDL
A BSA
1
.(50
.g50
U)
£ 40
0)
C
CD
0
I._.~
S
30
2 20
c)
._
W101
500
250
2.5%
LPDS Serum
2.5%
2.5% 7.5% 25%
HDL as SerumEquivalents
(+ 2.5% LPDS)
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
FIG 2. Bar graph comparing the relative promigratory activities
of high-density lipoprotein (HDL) and low-density lipoprotein
(LDL). Wounded endothelial cell (EC) cultures were incubated
with 2.5% lipoprotein-deficient serum (LPDS, open bar), 2.5%
serum (black bar), or 2.5% LPDS in the presence of 2.5%, 7.5%,
or 25% of the serum-equivalent of HDL (dark striped bars) or LDL
(light striped bars). EC migration was measured after 22 hours;
migration of serum-free untreated cells was 269+8 cells per
3000 gm.
activity of either HDL was lost during preparation or
that other lipoprotein fractions might account for the
remaining activity.
The promigratory activity of HDL by itself was examined in serum-free medium and compared with an
optimal concentration of basic FGF. Phase-contrast
micrography showed that the stimulation by HDL was
concentration dependent, with a maximum stimulation
comparable to that of basic FGF at 10 ng/mL (Fig 3).
Quantification of these results (Fig 4A) showed that the
maximum stimulation by HDL was =80% at 150 gg
cholesterol per milliliter (or 600 gg protein per milliliter, expressed for comparison with other proteins), with
a half-maximum activity at =25 gg cholesterol per
milliliter (100 gig protein per milliliter). There was an
apparent morphological difference between HDL-stimulated and basic FGE-stimulated ECs: the former
maintained the polygonal appearance of cells in a
confluent monolayer, whereas the latter were distinctly
elongated (Fig 3 and data not shown). In many repeti-
:
L:
750
1000 1250 0
[Protein] (ggmi)
2.5% 7.5% 25%
LDL as SerumEquivalents
(i. 2.5% LPDS)
20
40
60
80
100
[Growth Factor] (nglml)
FIG 4. Graphs showing the effect of high-density lipoprotein
(HDL) and basic fibroblast growth factor (FGF) on woundinduced migration of endothelial cells. A, Wounded endothelial
cell cultures were incubated with HDL (i) or bovine serum
albumin (BSA, A), and cell migration was determined after 22
hours. Migration of unstimulated endothelial cells was 304+34
per 3000 gm. B, Conditions were the same as in panel A, but
cells were incubated with basic FGF (e) or acidic FGF (o).
Migration of unstimulated endothelial cells was 222-- 10 cells per
3000 zm.
tions of this experiment, with at least eight different
HDL preparations, a maximum stimulation of 40% to
100% was observed with a half-maximum concentration
between 25 and 40 gg cholesterol per milliliter. Nearly
identical results were seen with several distinct isolates
of bovine aortic ECs between passages 6 and 19. Bovine
serum albumin at a protein concentration comparable
to HDL did not significantly stimulate EC movement,
indicating that the effect of HDL was specific and not
just a result of additional protein (Fig 4A). The latter
possibility was also made less likely in view of the
presence of 1 mg/mL gelatin during all experiments.
The maximal stimulation was comparable to the 75%
stimulation by basic FGF at 10 ng/mL (Fig 4B), an
increase consistent with previous reports.927 Acidic
FGF was considerably less effective, with a maximal
stimulation of only 30% at 50 ng/mL; the comparatively
low activity of acidic FGF may be due to the absence of
the synergistic agent heparin.
To determine the time required for HDL to stimulate
EC movement, a time course was examined during a
24-hour period. The amount of migration increased
A':5-'
s
.i *
.4
-i
No Agonist
HDL
25 jg/ml
HDL
50 jig/mi
-4f
I
HDL
100 kg/mi
HDL
200kg/mi
bFGF
10 ng/mi
FIG 3. Photomicrographs showing the effect of high-density lipoprotein (HOL) and basic fibroblast growth factor (bFGF) on
wound-induced migration of endothelial cells. Wounded endothelial cell cultures were incubated for 22 hours with HDL or bFGF in
serum-free medium at the concentrations indicated. The cells were fixed and stained with Wright-Giemsa and visualized by
phase-contrast microscopy.
Circulation Research Vol 74, No 6 June 1994
1152
~~~~0-HDL
+-HDL
300
E
Z 200
10
0t
15
10
5
Time
20
25
(h)
FIG 5. Time course of stimulation of endothelial cell migration by
high-density lipoprotein (HDL). Endothelial cells were incubated
with (e) or without (o) HDL (100 jig cholesterol per milliliter), and
the migration was stopped at indicated intervals of time by fixing
and staining with Wright-Giemsa. Endothelial cell migration is
expressed as the number of migrating cells per 3000 ,um.
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
almost linearly during this period, with the HDLtreated cells showing a higher migration rate than
untreated cells as early as 3 hours, the minimum interval required to accurately measure migration by this
method (Fig 5). Time-lapse videomicrography showed
that cell movement was clearly visible as early as 1 hour
after treatment with HDL and basic FGF but not
detectable until 2 to 3 hours in untreated cells (data not
shown). When HDL was removed 1 hour (or 4 hours,
data not shown) after the injury was made, there was
little stimulation of movement, indicating that HDL was
continuously required (Fig 6); a similar requirement was
observed for basic FGF. This result showed, in addition,
that the stimulation by HDL was not due to a coating of
the plastic surface after the injury was made, suggesting
that an interaction with the cells themselves may be
responsible for its activity.
Since HDL is a known mitogen for ECs and other
cells,18,19 we considered the possibility that cell proliferation, in addition to migration, contributed to the
appearance of cells in the wound region. To differentiate these activities of HDL, EC migration was measured
in the presence of hydroxyurea, an inhibitor of the
250
200
0
C)
50
HDL
EC alone
(22 h)
(1 h)
Basic FGF
(1 h)
(22 h)
FIG 6. Bar graph showing requirements for the continuous
presence of agonists. Wounded endothelial cell (EC) cultures
treated with high-density lipoprotein (HDL, 100 ,ug cholesterol per milliliter) and basic fibroblast growth factor (basic FGF,
10 ng/mL). The agonists remained for the entire 22-hour migration interval (black bars) or were removed 1 hour after the wound
was made (dense striped bars). Unstimulated EC migration (light
striped bars) was 196-ti cells per 3000 gm.
were
G1/S-phase transition of the cell cycle. Phase-contrast
photomicrography clearly showed that the stimulation
of EC migration by HDL was not at all diminished by
hydroxyurea (Table); basal (unstimulated) and basic
FGF-stimulated migration was also unaffected by hydroxyurea. HDL (and basic FGF) increased the migration of ECs by >50% whether or not they were treated
with hydroxyurea (which had little effect by itself). To
show that hydroxyurea effectively blocked EC proliferation, the cells were incubated with [3H]thymidine, and
labeled nuclei were identified by autoradiography. The
Table shows that cell division was completely blocked by
1 mmol/L hydroxyurea. These data also confirm the
mitogenic activity of HDLs (and basic FGF) observed
by others, even in the absence of serum factors. Interestingly, in all treatments the cells behind the cutting
edge were much less actively proliferative than the
migrating cells. It is possible, if not likely, that these
processes are related and that cell migration leads to
increased proliferation by decreasing contact inhibition
or by decreasing the cell's requirements for growth
stimulators.4,5
Multiple peptide growth factors, and possibly basic
FGF,28 are present in serum and may be associated with
HDL. To test whether HDL-bound basic FGF was
responsible for the activity of HDL, we examined the
ability of a neutralizing antibody to basic FGF to block
HDL-stimulated migration. An antibody to human basic FGF inhibited 52% of basic FGF-stimulated EC
migration (Fig 7). In contrast, the inhibition of the
activity of HDL was much less (28%) and consistent in
magnitude with the decrease in migration due to inactivation of cell-derived basic FGF. This result suggested
that the active agent in HDL was not basic FGF but did
not eliminate the possibility that HDL and basic FGF
shared similar promigratory signaling pathways. To test
this possibility, the effect of basic FGF (at an optimal
stimulatory concentration of 10 ng/mL) on HDL-stimulated movement was examined. The maximum stimulation by basic FGF alone was =65%, and that of HDL
alone was :80% (Fig 8). Addition of basic FGF increased the stimulatory activity by 60% to 80% at all
concentrations of HDL, with a maximum stimulation of
150%, higher than that observed by us under any other
conditions. These data indicate that HDL and basic
FGF stimulate EC migration by distinct mechanisms,
and furthermore, if the promigratory agent in HDL is a
peptide growth factor, then it follows a pathway distinct
from basic FGF.
To distinguish between the promigratory mechanisms
of basic FGF and HDL, specific inhibitors of signal
transduction pathways were investigated. We recently
found that pertussis toxin, a factor that ADP-ribosylates
and inactivates susceptible Gi-like G proteins,29 blocked
EC movement in response to basic FGF but not to
serum.30 In agreement with those results, Fig 9 shows
that basic FGF-stimulated movement was markedly
suppressed by pertussis toxin (77% inhibition). Pertussis toxin also blocked "unstimulated" EC motility (by
64%), consistent with the role of endogenous basic FGF
in this process.9 In contrast, EC motility stimulated by
HDL, as well as by serum, was only marginally inhibited
by the toxin (18% and 14%, respectively). These data
indicate that HDL may be the serum component responsible for pertussis toxin-insensitive serum-stimu-
Murugesan et al HDL Stimulates Endothelial Cell Movement
1153
Effect of Hydroxyurea on High-Density Lipoprotein- and Basic Fibroblast Growth
Factor-Stimulated Endothelial Cell Migration
EC Migration,
No. of Cells
305±15
340±5
Treatment
None
Hydroxyurea
HDL
HDL+hydroxyurea
11
EC Proliferation,
No. of Cells
91±6
0+0
57
55
242±13
0±0
Stimulation, %
...
480±26
473±1
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
231±12
54
470+2
Basic FGF
0±0
52
464±13
Basic FGF+hydroxyurea
EC indicates endothelial cell; HDL, high-density lipoprotein; and FGF, fibroblast growth factor. The
results are expressed as mean±SEM for the number of migrating or proliferating cells per 3000 ,um.
Wounded EC cultures were treated with HDL (100 jig cholesterol per milliliter) or basic FGF (10
ng/mL) in serum-free medium containing [3H]thymidine (1 uCi/mL) and in the presence or absence of
hydroxyurea (1 mmol/L). After 22 hours, the medium was aspirated, and the cells were washed with
phosphate-buffered saline, fixed with methanol, and stained with hematoxylin. The [3H]thymidinelabeled nuclei of proliferating cells were identified by autoradiography after coating the cultures with
photographic emulsion. Proliferating cells in only the wound region were counted.
lated EC movement.30 These results also demonstrate
that basic FGF and HDL stimulate movement by different pathways, with only the former requiring the
activity of pertussis toxin-sensitive G proteins.
Basic FGF-mediated EC movement requires the
release of arachidonic acid by G protein-mediated
activation of phospholipase A2.30 To test whether HDLstimulated migration likewise requires mobilization of
arachidonic acid (albeit by a pertussis toxin-insensitive
pathway), the effects of specific inhibitors of phospholipase A2 were determined. Aristolochic acid and ONORS-082 blocked basic FGF-stimulated EC migration by
up to 85% (Fig 10); the concentrations at which these
inhibitors blocked basic FGF-stimulated movement
corresponded to the concentrations reported for their
specific inhibition of phospholipase A2.31,32 In contrast,
ONO-RS-082 had no effect on HDL-stimulated cell
movement, and aristolochic acid had only a moderate
suppressive activity. Measurement of protein synthesis
using [3H]leucine showed that the inhibition of migration by the inhibitors was not due to cytotoxicity, since
aristolochic acid did not inhibit the rate of protein
synthesis, and the inhibition by ONO-RS-082 was small
(and only at the highest concentration). These data
show that HDL-stimulated EC movement, unlike FGFstimulated movement, does not require phospholipase
A2 activity.
Discussion
The experiments described in the present study demonstrate that HDL is a potent promigratory agent that
stimulates the movement of EC up to twofold at concentrations below the normal physiological range of 250
to 500 gg HDL cholesterol per milliliter. The stimulation is specific, since other serum proteins, eg, albumin,
are without activity, and the magnitude of the stimulation is comparable to that of basic FGF, the prototypical
2Cnn
0
1501
-
M
+
anti bFGF
HDL alone
+ basic FGF
* ~~~~~~~~HDL
0
0
c
0
T
0
,
,
200
anti bFGF
c
°150
-
E
100 _
'E
50s-
50
0
EC alone
Basic FGF
HDL
FIG 7. Bar graph showing the effect of anti-basic fibroblast
growth factor (anti bFGF) on the promigratory activity of highdensity lipoprotein (HDL). Wounded endothelial cell (EC) cultures were preincubated for 1 hour in the absence (striped bars)
or presence (black bars) of mouse monoclonal antibody to
bovine basic fibroblast growth factor (basic FGF, 10 ,ug/mL)
before the addition of HDL (100 gg cholesterol per milliliter) or
basic FGF (10 ng/mL) for a total of 22 hours. Migration of
untreated cells was 274±34 cells per 3000 ,m.
OC
0
50
100
I_
150
200
[HDL] (gg cholesterolml)
FIG 8. Graph showing additive promigratory activity of highdensity lipoprotein (HDL) and basic fibroblast growth factor
(basic FGF). Wounded endothelial cell cultures were incubated
for 22 hours with (c) or without (o) a maximal promigratory
concentration of basic FGF (10 ng/mL) and with the indicated
concentration of HDL. Unstimulated endothelial cell migration
was 232±9 cells per 3000 ,um.
1154
Circulation Research Vol 74, No 6 June 1994
the migration period was generally limited to 22 hours;
nevertheless, a significant number of migrating cells
were observed to go through S phase (Table). To
demonstrate that the increased number of cells in the
wound in HDL-treated cultures was due to cells moving
250
No Agonist Basic FGF
HDL
Fetal Bovine
Serum
FIG 9. Bar graph showing the effect of pertussis toxin on
endothelial cell movement. Confluent endothelial cell cultures
were injured, and the cells were preincubated for 75 minutes
with medium alone (light striped bars) or 15 ng/mL of pertussis
toxin (solid bars). Basic fibroblast growth factor (basic FGF, 10
ng/mL), high-density lipoprotein (HDL, 100 gg cholesterol per
milliliter), or fetal bovine serum (5%) was added as shown, and
cell migration was measured after an additional 22 hours.
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
agonist of EC movement. These results represent, to
knowledge, the first report of stimulation of cell
movement by normal lipoproteins (with the exception of
our previous studies of oxidized LDL, which showed
that native LDL was mildly promigratory23). Monocyte
chemotaxis induced by the EC product, monocyte
chemotactic protein-1, has been shown to be stimulated
by oxidized LDL.33 Similarly, other laboratories have
shown that specific modified lipoproteins, including
oxidized LDL and acetoacetylated LDL but not native
LDL, induce chemotactic responses in smooth muscle
cells.34'35
Several laboratories have shown that HDL stimulates
the proliferation of ECs and other cells.18'9 The increased movement of cells in response to HDL was
observed as early as 1 hour by videomicrography, thus
suggesting that the migratory response was an early
event that was independent of, or at least preceded,
proliferation. To minimize the fraction of cells dividing,
our
200
2i-0150
0
Q
0
.>
0)
c
g
50
.2
0
0
2
4
6
8
[ONO-RS-0821 (gM)
L
10
A,
12
0
20 40 60
80 100 120
[Aristolochic Acid] (gM)
FIG 10. Graphs showing the effect of phospholipase A2 inhibitors on endothelial cell migration. A, Wounded endothelial cell
cultures were pretreated with ONO-RS-082 for 1 hour before
addition of high-density lipoprotein (HDL, o, 100 jig cholesterol
per milliliter), basic fibroblast growth factor (basic FGF, A, 10
ng/mL), or medium alone (no agonist, o) for 22 hours. Migration
of untreated endothelial cells was 210±8 cells per 3000 /tm. At
the highest concentration used, ONO-RS-082 was cytotoxic. B,
Conditions were the same as in panel A, but aristolochic acid
was used in place of ONO-RS-082. Migration of untreated
endothelial cells was 201 +16 cells per 3000 gm.
into the wound rather than division of cells already in
the wound, the latter process was specifically blocked by
hydroxyurea. The results clearly show that movement of
ECs into the wound does not require proliferation and
that HDL can act as a promigratory agent independent
of its action as a growth promoter. Tile ability of
hydroxyurea36 and other agents37,38 to block cell proliferation, but not migration, has been well documented
and used as evidence of the independence of migration
and proliferation. In contrast, the high fraction of
rapidly migrating ECs in the wound area that incorporate [3H]thymidine, compared with the more slowly
moving cells behind the wound edge, suggests that
migration and replication may be related, as has been
observed previously.5,38,39
When HDL remained on the EC cultures only long
enough to coat the plastic dish, the rate of cell movement was essentially identical to that of unstimulated
cells. This result suggests that HDL does not function as
a matrixlike attachment or spreading factor but that a
direct interaction of HDL with the cells may be required
for promigratory activity. One possibility is that HDL
stimulates a specific intracellular signaling pathway after binding to a cell surface receptor or after nonreceptor-mediated uptake and internalization. Highaffinity binding sites for HDLs have been identified on
bovine aortic ECs, and a 110-kD protein that binds to
HDL and has several receptor-like features has been
cloned and expressed.40 Binding of HDLs to ECs may
be facilitated by apolipoprotein AI and apolipoprotein
All, which share a common receptor.41 An alternate
uptake mechanism has been advanced in which HDLs
bind to ECs primarily through lipid-lipid interactions
rather than a specific polypeptide receptor.42 HDL may
also exert its promigratory activity without directly
interacting with cells. As one possibility, HDL, acting as
a "lipid sink," may remove from the cells an endogenous
antimigratory molecule. Although we have no direct
evidence for this mechanism, the interaction of lipids
with HDL apolipoproteins and with native HDL has
been long appreciated, and our own results suggest that
specific oxidized lipids have substantial antimigratory
activity.23
We do not yet know the molecular mechanisms
underlying the promigratory activity of HDL. We have
shown that an antibody to basic FGF does not block the
activity of HDL, showing that the active agent in HDL
activity is not basic FGF. In agreement with this result,
the promigratory activities of HDL and basic FGF have
been found to be completely additive, indicating that
they involve different mechanisms. A similar additive
relation has been reported by others for the mitogenic
activities of the two factors.19 The rapid induction of
migration by HDL is consistent with a signal transduction mechanism. The signaling pathways used by HDL
are not well understood (see Reference 43 for review),
but others have reported a role for G protein-coupled
pathways in the stimulation by HDL of muscarinic
receptors in chick heart cells44 and a pertussis toxinsensitive activation of phospholipase C in platelets.45
Murugesan et al HDL Stimulates Endothelial Cell Movement
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
Our experiments indicate that the stimulation of EC
movement by HDL, unlike basic FGF, does not require
a pertussis toxin-sensitive G protein. We have recently
shown that the stimulation of movement by basic FGF
requires G protein-mediated release of arachidonic
acid.30 This finding is consistent with evidence showing
that arachidonic acid release is an early response of ECs
to basic FGFM and that activation of phospholipase A2
by multiple agonists may be coupled to pertussis toxinsensitive G proteins.47 Our results with two specific
inhibitors of phospholipase A2 show that this enzyme is
not required for HDL-stimulated EC motility and are
further evidence for the differential utilization of signaling pathways by HDL and basic FGF. HDL has also
been shown to stimulate phosphorylation of 27-kD
proteins,48 to block G protein interaction with its receptor,44 to induce a transient release of Ca 2+49 and to
activate protein kinase C.50 The role that these and
other intracellular signaling pathways play in HDLmediated migration is not known. The effect of HDL on
other processes required for EC migration such as
cytoskeletal rearrangement and secretion of matrixdegrading proteases are likewise unknown.
Epidemiologic and clinical studies have shown that
plasma HDL levels are inversely related to the incidence of coronary artery disease.20 Recent studies with
transgenic animals suggest that the apolipoprotein A151
component of HDL may be protective against atherogenesis. Several mechanisms have been suggested for
the putative antiatherogenic activity of HDL. One suggested mechanism is the transfer of excess cholesterol
from peripheral tissues to the liver.52 According to a
second proposed mechanism, HDL prevents the oxidation of LDL53 and suppresses several of the potentially
damaging activities of oxidized LDL, including toxicity
to vascular cells54 and stimulation of monocyte transmigration across the endothelium.33
Our own results on EC movement invite speculation
on an additional mechanism by which HDL may resist
the initiation of atherogenesis. Studies in animal models
have shown that deendothelialization of the rat aorta
with a balloon catheter is followed by a reproducible
series of events, beginning with rapid platelet aggregation, followed by slow reendothelialization, and culminating in marked intimal thickening due to SMC migration and proliferation that is apparent within 4 weeks. A
correlation between the duration of endothelial denudation and the degree of intimal thickening has been
demonstrated; in fact, injured regions that are covered
by regenerated endothelium within 1 week after injury
are completely spared.55 We suggest that high plasma
levels of HDL may stimulate the migration (and subsequent proliferation) of ECs into a wound region, thus
reducing the duration of exposure of a wound to bloodborne elements. This process could be involved in the
rate of recovery of the endothelium during normal
vessel wall processes, such as denuding endothelial
injury, or after clinical procedures, including percutaneous transluminal angioplasty, vascular reconstruction, or organ transplantation.
Acknowledgments
This work was supported by grants HL-40352 and HL41178 from the National Heart, Lung, and Blood Institute,
National Institutes of Health. Dr Fox is an Established Inves-
1155
tigator of the American Heart Association. Dr Sa is a Fellow
of the American Heart Association, Northeast Ohio Affiliate,
Inc. We thank Richard F. Marquardt for assistance with the
image and data analysis and James Lang for photography.
References
1. Ausprunk DH, Folkman J. Migration and proliferation of endothelial cells in preformed and newly formed blood vessels during
tumor angiogenesis. Microvasc Res. 1977;14:53-65.
2. Wall RT, Harker LA, Striker GE. Human endothelial cell
migration: stimulation by a released platelet factor. Lab Invest.
1978;39:523-529.
3. Schwartz SM, Gajdusek CM, Selden SC. Vascular wall growth
control: the role of the endothelium. Arteriosclerosis. 1981;1:
107-126.
4. Coomber BL, Gotlieb Al. In vitro endothelial wound repair: interaction of cell migration and proliferation. Arteriosclerosis. 1990;10:
215-222.
5. Dulbecco R, Stoker MG. Conditions determining initiation of
DNA synthesis in 3T3 cells. Proc Natl Acad Sci USA. 1970;66:
204-210.
6. Connolly DT, Heuvelman DM, Nelson R, Olander JV, Eppley BL,
Delfino JJ, Siegel NR, Leimgruber RM, Feder J. Tumor vascular
permeability factor stimulates endothelial cell growth and angiogenesis. J Clin Invest. 1989;84:1470-1478.
7. Rosen EM, Carley W, Goldberg ID. Scatter factor regulates
vascular endothelial cell motility. Cancer Invest. 1990;8:647-650.
8. Terranova VP, DiFloria R, Lyall RM, Hic S, Friesel R, Maciag T.
Human endothelial cells are chemotactic to endothelial cell growth
factor and heparin. J Cell Biol. 1985;101:2330-2334.
9. Sato Y, Rifkin DB. Autocrine activities of basic fibroblast growth
factor: regulation of endothelial cell movement, plasminogen
activator synthesis, and DNA synthesis. J Cell Biol. 1988;107:
1199-1205.
10. Lindner V, Majack RA, Reidy MA. Basic fibroblast growth factor
stimulates endothelial regrowth and proliferation in denuded
arteries. J Clin Invest. 1990;85:2004-2008.
11. Bjornsson TD, Dryjski M, Tluczek J, Mennie R, Ronan J, Mellin
TN, Thomas KA. Acidic fibroblast growth factor promotes
vascular repair. Proc NatlAcad Sci USA. 1991;88:8651-8655.
12. Lee PL,Johnson DE, Cousens LS, Fried VA, Williams LT. Purification and complementary DNA cloning of a receptor for basic
fibroblast growth factor. Science. 1989;245:57-60.
13. Montesano R. Regulation of angiogenesis in vitro. EurJ Clin Invest.
1992;22:504-515.
14. Burk RR, Clopath P, Muller K. Deficiency in the ability of serum
from pigs on an atherogenic diet to stimulate the migration of 3T3
cells in culture. Artery. 1979;6:205-219.
15. Thorgeirsson G, Robertson AL, Cowan DH. Migration of human
vascular endothelial and smooth muscle cells. Lab Invest. 1979;41:
51-62.
16. Gotlieb AI, Spector W, Desai K, Steiner G. The effect of low
density lipoprotein on in vitro porcine endothelial cell reendothelialization. Artery. 1983;12:132-144.
17. Tauber JP, Goldminz D, Gospodarowicz D. Up-regulation in
vascular endothelial cells of binding sites of high density lipoprotein induced by 25-hydroxycholesterol. Eur J Biochem. 1981;
119:327-339.
18. Tauber JP, Cheng J, Gospodarowicz D. Effect of high and low
density lipoproteins on proliferation of cultured bovine vascular
endothelial cells. J Clin Invest. 1980;66:696-708.
19. Libby P, Miao P, Ordovas JM, Schaefer EJ. Lipoproteins increase
growth of mitogen-stimulated arterial smooth muscle cells. J Cell
Physiol. 1985;124:1-8.
20. Castelli WP, Garrison RJ, Wilson PWF, Abbott RD, Kalousdian S,
Kannel WB. Incidence of coronary heart disease and lipoprotein
cholesterol levels: the Framingham study. JAMA. 1986;256:
2835-2838.
21. Fox PL, DiCorleto PE. Regulation of production of a plateletderived growth factor-like protein by cultured bovine aortic endothelial cells. J Cell Physiol. 1984;121:298-308.
22. Howe PH, Draetta G, Leof EB. Transforming growth factor p1
inhibition of p34Cdc2 phosphorylation and histone Hi kinase activity
is associated with G1/S-phase growth arrest. Mol Cell Biol. 1991;
11:1185-1194.
23. Murugesan G, Chisolm GM, Fox PL. Oxidized low density lipoprotein inhibits the migration of aortic endothelial cells in vitro.
J Cell Biol. 1993;120:1011-1019.
1156
Circulation Research Vol 74, No 6 June 1994
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
24. Hessler JR, Morel DW, Lewis LJ, Chisolm GM. Lipoprotein oxidation and lipoprotein-induced cytotoxicity. Arteriosclerosis. 1983;
3:215-222.
25. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein measurement with the Folin phenol reagent. J Biol Chem. 1951;193:
265-275.
26. Schuh J, Fairclough GF, Haschemeyer RH. Oxygen-mediated heterogeneity of apo-low-density lipoprotein. Proc Natl Acad Sci
USA. 1978;75:3173-3177.
27. Hasselaar P, Sage EH. SPARC antagonizes the effect of basic
fibroblast growth factor on the migration of bovine aortic endothelial cells. J Cell Biochem. 1992;49:272-283.
28. Baird A, Bohlen P. Fibroblast growth factors. In: Sporn MB,
Roberts AB, eds. Peptide Growth Factors and Their Receptors I.
New York, NY: Springer-Verlag; 1991:369-418.
29. Ui M, Katada T. Bacterial toxins as probe for receptor-Gi
coupling. Adv Second Messenger Phosphoprotein Res. 1990;24:63 - 69.
30. Sa G, Fox PL. Basic fibroblast growth factor-stimulated endothelial
cell movement is mediated by a pertussis toxin-sensitive pathway
regulating phospholipase A2 activity. J Biol Chem. 1994;269:
3219-3225.
31. Rosenthal MD, Vishwanath BS, Franson RC. Effects of aristolochic acid on phospholipase A2 activity and arachidonate
metabolism of human neutrophils. Biochim Biophys Acta. 1989;
1001:1-8.
32. Banga HS, Simons ER, Brass LF, Rittenhouse SE. Activation of
phospholipases A and C in human platelets exposed to epinephrine: role of glycoproteins lIb/IIla and dual role of epinephrine. Proc NatlAcad Sci USA. 1986;83:9197-9201.
33. Navab M, Imes SS, Hama SY, Hough GP, Ross LA, Bork RW,
Valente AJ, Berliner JA, Drinkwater DC, Laks H, Fogelman AM.
Monocyte transmigration induced by modification of low density
lipoprotein in cocultures of human aortic wall cells is due to
induction of monocyte chemotactic protein 1 synthesis and is
abolished by high density lipoprotein. J Clin Invest. 1991;88:
2039-2046.
34. Naito M, Hayashi T, Funaki C, Kuzuya M, Asai K, Kuzuya F.
Chemotactic response of vascular smooth muscle cells to acetoacetylated low-density lipoprotein. Heart Vessels. 1988;4:
100-103.
35. Autio I, Jaakkola 0, Solakivi T, Nikkari T. Oxidized low-density
lipoprotein is chemotactic for arterial smooth muscle cells in
culture. FEBS Lett. 1990;277:247-249.
36. Lipton A, Klinger I, Paul D, Holley RW. Migration of mouse 3T3
fibroblasts in response to a serum factor. Proc NatlAcad Sci U SA.
1971;68:2799-2801.
37. Sholley MM, Gimbrone MA, Cotran RS. Cellular migration and
replication in endothelial regeneration: a study using irradiated
endothelial cultures. Lab Invest. 1977;36:18-25.
38. Selden SC, Rabinovitch PS, Schwartz SM. Effects of cytoskeletal
agents on replication of bovine endothelium. J Cell Physiol. 1981;
108:195 -211.
39. Casscells W. Migration of smooth muscle and endothelial cells:
critical events in restenosis. Circulation. 1992;86:723-729.
40. McKnight GL, Reasoner J, Gilbert T, Sundquist KO, Hokland B,
McKernan PA, Champagne J, Johnson CJ, Bailey MC, Holly R,
O'Hara PJ, Oram JF. Cloning and expression of a cellular high
density lipoprotein-binding protein that is up-regulated by cholesterol loading of cells. J Biol Chem. 1992;267:12131-12141.
41. Vadiveloo PK, Fidge NH. The role of apoproteins AI and All in
binding of high-density lipoprotein3 to membranes derived from
bovine aortic endothelial cells. Biochem J. 1992;284:145-151.
42. Tabas I, Tall AR. Mechanism of the association of HDL3 with
endothelial cells, smooth muscle cells, and fibroblasts: evidence
against the role of specific ligand and receptor proteins. J Biol
Chem. 1984;259:13897-13905.
43. Schmitz G, Lackner KJ. High-density lipoproteins and atherosclerosis. Curr Opin Lipidol. 1993;4:392-400.
44. Subers EM, Nathanson NM. Regulation by high density lipoproteins of muscarinic acetylcholine receptor function in chick
heart cells cultured in defined medium. J Biol Chem. 1989;264:
19685-19693.
45. Nazih H, Devred D, Martin-Nizard F, Clavey V, Fruchart JC,
Delbart C. Pertussis toxin sensitive G-protein coupling of HDL
receptor to phospholipase C in human platelets. Thromb Res.
1992;67:559-567.
46. Fafeur V, Ping Jiang Z, Bohlen P. Signal transduction by bFGF,
but not TGF,81, involves arachidonic acid metabolism in endothelial cells. J Cell Physiol. 1991;149:277-283.
47. Xing M, Mattera R. Phosphorylation-dependent regulation of
phospholipase A2 by G-proteins and Ca21 in HL60 granulocytes.
J Biol Chem. 1992;267:25966-25975.
48. Darbon JM, Tournier JF, Tauber JP, Bayard F. Possible role of
protein phosphorylation in the mitogenic effect of high density
lipoproteins on cultured vascular endothelial cells. J Biol Chem.
1986;261:8002-8008.
49. Porn MI, Akerman KE, Slotte JP. High-density lipoproteins
induce a rapid and transient release of Ca 2+ in cultured fibroblasts.
Biochem J. 1991;279:29-33.
50. Voyno-Yasenetskaya TA, Dobbs LG, Erickson SK, Hamilton RL.
Low density lipoprotein- and high density lipoprotein-mediated
signal transduction and exocytosis in alveolar type II cells. Proc
Natl Acad Sci U S A. 1993;90:4256 - 4260.
51. Rubin EM, Krauss RM, Spangler EA, Verstuyft JG, Clift SM.
Inhibition of early atherogenesis in transgenic mice by human
apolipoprotein AI. Nature. 1991;353:265-267.
52. Badimon JJ, Fuster V, Badimon L. Role of high density lipoproteins in the regression of atherosclerosis. Circulation. 1992;
86(suppl III):III-86-III-94.
53. Parthasarathy S, Barnett J, Fong LG. High-density lipoprotein
inhibits the oxidative modification of low-density lipoprotein.
Biochim Biophys Acta. 1990;1044:275 -283.
54. Hessler JR, Robertson AL, Chisolm GM. LDL-induced cytotoxicity and its inhibition by HDL in human vascular smooth
muscle and endothelial cells in culture. Atherosclerosis. 1979;32:
213-229.
55. Haudenschild CC, Schwartz SM. Endothelial regeneration, II: restitution of endothelial continuity. Lab Invest. 1979;41:407-418.
High-density lipoprotein stimulates endothelial cell movement by a mechanism distinct
from basic fibroblast growth factor.
G Murugesan, G Sa and P L Fox
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
Circ Res. 1994;74:1149-1156
doi: 10.1161/01.RES.74.6.1149
Circulation Research is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1994 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7330. Online ISSN: 1524-4571
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circres.ahajournals.org/content/74/6/1149
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Circulation Research can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further information
about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation Research is online at:
http://circres.ahajournals.org//subscriptions/