jf. Cell Sci. 73, 19-32 (1985)
Primed in Great Britain © The Company of Biologists Limited 1985
19
EXTRACELLULAR MATRIX MODULATION OF
ENDOTHELIAL CELL SHAPE AND MOTILITY
FOLLOWING INJURY IN VITRO
W I L L I A M C . Y O U N G A N D IRA M .
HERMAN1
Department of Anatomy and Cellular Biology, Tufts University School of Medicine, 136
Harrison Avenue, Boston, MA 02111, L'.SA.
SUMMARY
We utilized fluorescence microscopy and affinity-purified antibodies to probe the form and
function of cytoplasmic actin in endothelial cells (EC) recovering from injury and grown on
extracellular matrices in vitro. Bovine aortic EC were seeded onto glass microscope coverslips that
had been coated with either BSA, fibronectin, type I and III (interstitial) collagens, type IV
(basement membrane) collagen or gelatin. After EC that had been grown on glass, glass-BSA or
extracellular matrix-coated coverslips reached confluence, a 300-400 fim zone of cells was
mechanically removed to stimulate EC migration and proliferation. Post-injury EC movements
were monitored with time-lapse, phase-contrast videomicrography before fixation for actin
localization with fluorescence microscopy using affinity-purified antibodies. We found that the
number of stress fibres within EC was inversely proportional to the rate of movement; and, the
rates of movement for EC grown on glass or glass-BSA were approximately eight times faster than
EC grown on gelatin or type IV collagen (X velocity = 0-5 fim/min versus 006/xm/min). EC
movements on fibronectin and interstitial collagens were similar (X velocity = 0 2 /xm/min).
These results suggest that extracellular matrix molecules modulate EC stress fibre expression,
thereby producing alterations in the cytoskeleton and the resultant EC movements that follow
injury in vitro. Moreover, the induction of stress fibres in the presence of basement membrane
(type IV) collagen may explain the failure of aortic EC to migrate and repopulate wounded regions
of intima during atherogenesis in vivo.
INTRODUCTION
In healthy blood vessels, endothelial cells (EC) function to provide a
blood-compatible inner lining of the vessel wall (Zetter, 1981). Repeated injury to
this squamous epithelial layer, resulting in a failure to maintain continuity between
EC, may give rise to atherogenesis (Ross & Glomset, 1976). EC reside on and
synthesize a basement membrane composed of numerous macromolecules,
including collagens, proteoglycans and glycosaminoglycans (GAGs), as well as other
glycoproteins (Hay, 1981). Because of this permeability barrrier, EC may depend
upon diffusion of nutrients through the basement membrane for their normal
metabolism. Thus, the contact and relationship between the EC abluminal surface
and the basement membrane plays an important role in normal EC metabolism
including the maintenance of intimal integrity.
•To whom all correspondence should be addressed.
Key words: anti-actin, endothelium, injury.
20
W. C. Young and I. M. Herman
In vivo, stability of the intima may be dependent upon the stress fibres of the EC
cytoskeleton. Since stress fibres terminate in adhesion plaques that associate with the
luminal and abluminal plasma membranes, these microfilament bundles may
promote attachment of EC to the underlying basement membrane (Herman, Pollard
& Wong, 1982; Wong, Pollard & Herman, 1983). For this reason the EC
cytoskeleton has been proposed as a cellular mediator of the blood-shearing forces
that occur at the luminal aspect of the intima in regions where blood velocity and
turbulence are excessive (Dewey, Bussolari, Gimbrone & Davies, 1981; Herman e/
al. 1982). EC stress fibres in these areas are aligned parallel to blood flow and have
recently been shown to modulate during states of hypertension (White, Gimbrone &
Fujiwara, 1983) or following EC injury in vivo (Gabbiani, Gabbiani, Lombardi &
Schwartz, 1983). Consequently, it seems likely that EC actomyosin serves to
maintain intimal integrity by promoting EC-basement membrane adhesion via the
plasma membrane, thereby anchoring the EC to the basal laminae of large vessels;
but, modulation of the EC cytoskeleton occurs in response to vascular
pathophysiology.
A role has also been proposed for contractile proteins in the EC response to injury
in vitro (Selden & Schwartz, 1979; Gottlieb, Heggeness, Ash & Singer, 1979).
Removal of EC from confluent monolayers results in spreading, migration and
proliferation (Selden & Schwartz, 1979; Scholey, Gimbrone & Cotran, 1977). The
degree to which each of these types of behaviour is expressed is dependent upon the
extent of injury inflicted in vitiv. The EC migratory response is dependent upon the
form and arrangement of actin filaments, since cytochalasin B inhibits the
post-injury migration of EC at the wound edge /';/ vitiv (Selden & Schwartz, 1979).
Thus, an understanding of the EC contractile machinery present in migratory
versus stationary cells may reveal how EC utilize contractile proteins to effect
intimal stability and, or, post-injury motile force generation.
Since EC stress fibres interact with the basement membrane of blood vessels via
membrane attachment sites, molecules in the extracellular matrix may directly or
indirectly effect the overall configuration of the cytoskeleton. We were interested to
learn what influence extracellular matrix molecules might have on the overall
arrangement of the EC cytoskeleton and whether this cell-matrix interaction could
influence response to injury in vitiv. We used time-lapse, phase-contrast
videomicrography to measure the rate and extent of living EC movements following
injury as a function of extracellular matrix molecules, and characterized the form
and distribution of cytoplasmic actin in the wounded monolayers using
affinity-purified and labelled antibodies. Results of these experiments indicate that
extracellular matrix molecules profoundly affect the rates of post-injury EC motility;
and, coincident with these events there are significant alterations in the EC
cytoskeleton.
Anti-actin localization in wounded endothelium
21
MATERIALS AND METHODS
Antibody preparation
Affinity-purified rabbit anti-actin immunoglobulin Gs (IgGs) were prepared against chicken
gizzard (smooth muscle) actin as previously reported (Herman & Pollard, 1979). Non-adherent
IgG fractions that did not bind to the affinity columns were taken as immune IgG not directed
against actin and were used, as in the past, for control experiments (Herman & Pollard, 1979;
Herman, Crisona & Pollard, 1981).
Endothelial cell isolation and preparation
Bovine aortas were obtained from a local abbatoire on ice in Dulbecco's minimum essential
medium (DMEM) cooled to 0 °C on ice. EC were isolated via collagenase irrigation as described by
Jaffe, Hoyer & Nachman (1973) and transferred to 35 mm Petri dishes. EC colonies were isolated
and cloned into 24-well plates. All EC clones were tested for factor VIII antigen by
immunofluorescence using anti-factor VIII antibodies (Jaffeel til. 1973). A single clone of the cells
between 5 and 10 passages was subsequently grown to confluence on either: (1) uncoated glass
microscope coverslips; (2) glass coverslips coated with bovine serum albumin (BSA); (3) glass
coverslips coated with fig quantities of bovine fibronectin (Biomedical Technologies); (4) glass
coverslips coated with a dispersion of types I and III collagen (Ethicon Inc., Somerville, NJ); (5)
glass coverslips coated with gelatin (Difco Lab., Detroit, MI); or (6) glass coverslips coated with
type IV collagen obtained from bovine kidney cortex, a generous gift from Dr Sanyu Bixit
(Nashville, T N ) . Concentrations of extracellular matrix molecules adsorbed to the glass coverslips
were determined using the BioRAD protein assay and were found to be in the microgram range for
all molecules tested (1-5 /xg/coverslip). Cells were grown attached to the coverslips submerged in
DMEM supplemented with 5 % (v/v) calf serum buffered with 2-OmM-Hepes ( p H 7 3 ) .
Wounding of endothelial cell mono/avers
EC monolayers were grown to confluence on glass, glass-BSA, fibronectin, collagen types I and
III, gelatin and type IV collagen. Two days post-confluence, EC on various substrates were refed
with DMEM plus 5% calf serum buffered with 2-0mM-Hepes (pH7-3). A 300-400 fim wound was
mechanically inflicted in each monolayer using a sterile scalpel blade. The sterile procedure was
performed under observation using an inverted Olympus microscope equipped with
phase-contrast optics.
Quantitative analysis of endothelial cell movements following injury in vitro
The wounded cell monolayers attached to the coverslips were placed in a 35 mm Petri dish,
submerged in growth media, sealed with parafilm to maintain an atmosphere containing 5 % COj,
and placed onto the temperature-controlled (37 °C) stage of an Olympus light microscope
equipped with interference and phase-contrast optics. A microscope-interfaced MTI-Dage 65 TV
camera, coupled to an NEC video tape recorder with built-in time/date generator was used to
record living EC movements at the wound site. EC movements were recorded for 3 h following
injury, at which time the cells were removed from the microscope stage and prepared for
fluorescent antibody staining. For analysis of EC post-injury movements a scaled grid from the
camera image of a stage micrometer was used to make a vinyl overlay for the monitor. Thirty to
fifty cells from each run of every experimental group were tracked by single frame analysis. The
positions of nuclei and leading edges were marked and instantaneous velocities and net movements
plotted. We did this for cells as deep as eight rows away from the wound edge where directional
movement was nil.
Preparation of endothelial cells for fluorescent antibody staining
Each wounded monolayer was prepared for fluorescent antibody staining using formaldehyde
fixation followed by acetone extraction as described previously (Herman el til. 1981).
22
W. C. Young and I. M. Herman
Staining ofwounded endothelial cell monolayers with affinity-purified anti-actin
The fixed monolayers of aortic endothelium were incubated with a 12/u.g/ml solution of purified
anti-actin for 60min at room temperature. Each monolayer was then washed in 100 ml of
phosphate-buffered saline (PBS = 0015 M-sodium phosphate, pH7-5, 0-15 \i-NaCl) three times
for S min. After buffer equilibration, fixed and permeabilized cells were incubated in 50/xg/ml of
rhodamine-labelled goat anti-rabbit IgGs (Cappel Labs), for 60min at room temperature. Each
monolayer was again rinsed three times for 5 min in PBS. Each coverslip was then mounted onto a
glass microscope slide using a 9:1 (v/v) dilution of glycerol: PBS and scaled with nail polish.
Photomicrography
Fixed and stained EC monolayers were observed in phase-contrast and fluorescence microscopy
using a Zeiss Photomicroscope III equipped with the appropriate narrow band-pass barrier and
excitation filters needed for rhodamine fluorescence. The microscope is equipped with a 40X
planeofluor (NA 1-0) and a 63x planapochromat objective lens (NA 1-4). For photomicrography,
Tri-X negative film was rated as ASA = 1000 and developed in Acufinc Developer for 4-75 min at
20 °C.
Quantification of endothelial cell motility and stress fibre expression
Living cells. The time-lapse video tape recording of EC motility following injury was used to
measure the rate and extent of EC motility following wound infliction. An image of a stage
micrometer was projected onto the TV screen in order to scale post-injury migration.
Fixed cells. After fluorescence micrography of fixed and stained cell cultures was complete, the
negatives were projected onto a screen at XC80. Several parameters of EC stress fibres were then
measured. Fibre width was measured at the mid-point of each stress fibre with calipers. Fibre
length was also measured, in the same clearly disccrnable stress fibres. A scaled grid made up of
100(im2 units was used to measure stress fibre density. The number of fibres appearing within a
100fim2 area was digitized for cells grown on each substrate. A count of total fibres per cell
(;; > 420) was also performed. Cell size was determined, as a function of surface area, by
computing mean cell width and length for cells in which fibres were counted. A statistically
significant number of EC cultures grown on different substrates were analysed (/; = 24).
RESULTS
We examined the influence of various extracellular matrix molecules on the
motility of EC following injury. Using time-lapse video micrography to monitor the
rate and extent of living EC movements and fluorescent anti-actin staining of these
same migrating cells after fixation, we were able to correlate rates of motility with
specific actin arrangements. Our results indicate that: (1) post-injury EC movements are modulated by extracellular matrix molecules; (2) the distribution of actin
within the cytoplasmic matrix of wounded EC monolayers is profoundly affected by
varying the composition of the extracellular matrix; and (3) the density of stress
fibres in EC cytoplasm varied inversely with the velocity of the post-injury
migration.
Effect of extracellular matrix molecules on endothelial cell motility
Time-lapse video micrography proved to be an excellent method to monitor the
movements of EC remaining in wounded monolayers. Cell movements were
recorded for 3h following injury, before fixation and antibody staining.
Anti-actin localization in wounded endothelium
Fig. 1. EC migratory response to injury in vilm. Phase-contrast micrographs (A) and
fluorescent micrographs ( B - D ) of bovine aortic EC cultured on glass. After mechanical
wounding, cells migrate into the denuded region (A). Motile cells denoted in \ as • and
* are shown in B and C after staining with fluorescent anti-actin. Note the diffuse
fluorescence and relative lack of stress fibres in these migrating cells. Arrows in c
indicate prominent advancing pseudopodia. Cells that have repopulated the wound
(within 150min after injury) are shown in o. A, X180; H-D, X530.
When bovine aortic EC were grown on glass or glass-BSA substrates (Fig. 1),
mechanically induced wounds in the monolayer precipitated characteristic cell
movements. EC at the immediate wound edge extended a variety of lamellar and
pseudopodial cytoplasmic processes into the wounded region; and, with time, these
cells became polarized, often with their long axes oriented normal to the wound
edge. These cells exhibit directed migration. Once cell-cell contact was disrupted
due to the probing movements of an EC moving into the wound, the adjacent
cells moved forward and, or, laterally into the space vacated by its neighbour. 'Phis
24
W. C. Young and I. M. Herman
Fig. 2
Anti-actin localization in wounded endothelium
25
movement slowed down considerably and became increasingly less directional in EC
further away from the wound edge. Cellular movements, which appeared as
random, rapid surface blebs in time-lapse, were witnessed up to eight cell rows back
from the wound. EC along the wound edge, which were cultured on glass, exhibited
an average post-injury migration rate of 05/i.m/min. EC move most quickly
directly after removal of neighbouring cells (2-0/u.m/min) with little directional
movement once cell-cell contact is established. EC two and three cells deep (away
from the wound edge) exhibited rates of 0-25 and 0-12/u.m/min, respectively. These
monolayers (grown on glass or glass-BSA) were the only EC in our study capable of
re-covering the denuded region during the 3 h between injury and fixation.
The aortic EC cultures grown on various extracellular substrates expressed
varying rates of post-injury migration. Cell morphology also varied according to the
rate of cell movement. EC cultured on fibronectin and collagens I and III (Fig. 2)
migrated into wound sites at approximately equivalent rates (02/Lim/min). Cells
grown on glass moved roughly 2-5 times faster than the cells that were cultured on
fibronectin or interstitial collagens. EC grown on these substrates rarely extended
spindle-shaped processes into the wound and the membranes exhibited less ruffling,
seen by observing the injury site in time-lapse. EC cultured with fibronectin, and
collagens I and III, which were two and three cells deep from the wound edge,
moved at rates of 0-06 and 0-04/u.m/min, respectively.
When EC monolayers were grown on gelatin or type IV collagen (Fig. 3) and then
injured, the cells at the wound edge moved at a rate of 0-07 and 0-06/xm/min,
respectively; a net movement that was eight times slower than EC cultured on glass
(see Table 1). The EC grown on basement membrane collagen and gelatin also had
the most regular, uninterrupted borders. These cells were flattened, spread out and
very rarely exhibited long, cellular extensions perpendicular to the wound edge.
Motility of these EC populations was restricted to membrane ruffling at the free edge
along the wound site.
Effect of extracellular matrix molecules on cytoplasmic act in arrangement
After the bovine aortic EC had been allowed to recover for 3 h following injury,
the monolayers were fixed and stained with anti-actin. Measurements of cell size
(surface area), stress fibre length, stress fibre width, and stress fibre density
(number) were made from anti-actin-stained EC grown on different substrates.
When EC were grown to confluence on glass, injured, and then stained with
fluorescent anti-actin 3h later, two distinct fluorescent patterns were observable:
one component was fibrous, and the other was diffuse. Bright, marginal diffuse
Fig. 2. Post-injury analysis of bovine aortic EC cultured on fibronectin and interstitial
collagens. Areas at the wound edge ( * and *) seen with phase-contrast microscopy (A,IJ)
are seen in B with prominent fluorescence indicating membrane ruffling (arrows), c
demonstrates the typical fluorescent actin localization seen in EC at the wound edge
when grown on fibronectin. Similar results are obtained when aortic EC are cultured on
interstitial collagens ( D - F ) . Note the numerous stress fibres seen in H,C,I:,F. \, X400;
B-C, X695; D, X135; E - F , X625.
W. C. Young and I. M. Herman
Fig. 3. Influence of collagenous components on EC migratory response to injury.
Phase-contrast (A) and fluorescence ( B - C ) micrographs of EC cultured on type IV
collagen. Cells shown at the wound edge in A as * are seen in 11 exhibiting numerous
fluorescent stress fibres. Similar fluorescent patterns are seen in EC at the edge of a
denuded region in D when gelatin is used as the extracellular matrix. \, x 170; 11, X530;
C - D , X750.
Anti-actin localization in wounded endothelium
27
Table 1. Effects of extracellular matrix on post-injury motility and stress fibre
expression in cultured aortic endothelial cells
Glass
Fibronectin
Collagens I, III
Gelatin
Collagen IV
Stress fibre density*
(no. fibres/100 /xm2 field)
Endothelial cell motility
(fim/min)
1-08
1-74
1-90
2-41
2-71
5-1 x 10"'
2-2 X 10"'
1-7x10"'
0-7 X 10"'
0-6 X 10~'
• For each group: no. cells at wound edge used in analysis = 56< x < 105; no. 100/xm2 fields
used in analysis = 390 < X < 725: total no. fibres counted within individual 100 fim2
fields = 422<x<1901.
fluorescence was observed in the majority of the migrating EC along the wound edge
and was especially intense in membrane ruffles and pseudopodia (Fig. 1). Regions of
living, motile cytoplasm were noted from the tape recordings; these cellular regions
possessed the most diffuse and least intense fibrous fluorescence after staining with
anti-actin (Fig. l c ) . EC cultured on glass or glass-BSA exhibited a fine meshwork
of actin fluorescence in slowly moving regions of the cells along the wound edge.
Once migrating cells, which were characterized by diffuse anti-actin fluorescence,
had re-established contact with other EC in the previously denuded region, the fine,
fibrous meshwork of anti-actin fluorescence reappeared (Fig. I D ) . In the repopulated region of the wound, cell-cell borders were not completely contiguous
since irregular spaces between the cells could be detected in fluorescence (Fig. I D ) .
When EC grown on extracellular matrix molecules were wounded and the
post-injury motility observed .with time-lapse video micrography before anti-actin
staining, we found that the actin staining pattern was markedly different from the
EC grown on uncoated glass or glass-BSA. Cells that were grown on fibronectin had
roughly twice the surface area of those cells cultured on glass (Fig. 2). These
FN-cultured EC also demonstrated an increase in fibrous staining. The stress fibres
within the EC grown on fibronectin were 1-2 times as long (XL = 24-9/Am), 1-5
times as wide ( x l l * = 0-19/im), and 1-7 times as densely distributed as the fibres in
EC cultured on glass or glass-BSA. Bright, diffuse anti-actin fluorescence was
restricted to membrane ruffles along the wound site.
When the injured EC monolayer that had been cultured on interstitial collagens I
and III was stained with rhodamine-labelled anti-actin, a similar fluorescent pattern
to that of EC grown on fibronectin could be observed. Here, the cells spread out to
1-6 times the surface area (xS/1 = 150^im 2 ) of the cells grown on glass, and the actin
fibres measured 1-4 times the width, 1-1 times the length and 1-9 times the density
of the actin fibres from cells cultured on glass. In Fig. 2ie a non-contacted EC
exhibited extensive fibrous fluorescence, whereas an isolated cell on a glass matrix,
as seen in Fig. l c , exhibited an almost total lack of fibrous fluorescent staining.
28
W. C. Young and I. M. Herman
When EC were plated onto gelatin-coated glass before injury and staining, the
living cells were shown to have a rate of motility of 0-07/xm/min. Fluorescent
micrographs of these cells after fixation reveal numerous stress fibres (Fig. 3). These
cells, which were roughly 3-3 times (X&4 = 310/u.m ) the size of the cells grown on
glass, contained fibres that were 1-4 times as long, 2-2 times as wide and 2-4 times
(XSF/cell = 10-6) as densely expressed compared to the cells grown on glass.
Bovine aortic EC cultured on type IV or basement membrane collagen (Fig. 3)
also showed a high degree of fibrous fluorescence. As in the EC cultured in
fibronectin, collagens I and III, and gelatin, diffuse anti-actin fluorescence was
restricted to ruffling membranes in the EC cultured in collagen IV, along the wound
edge. These cells spread out to 3-6 times the surface area seen for the cells grown on
glass. The stress fibres in these cells were broad, measuring 2-3 times the width and
1-5 times the length of the stress fibres found in the cells grown on glass. Stress fibre
density was 2-7 times that of EC cultured on glass (xSA'/cell = 11-9).
It is evident that a relationship exists between the composition of the extracellular
matrix, the rate of migration, and stress fibre density (Table 1). We found that the
rate of EC motility was inversely related to the number of stress fibres, i.e. EC with
few fibres migrate the fastest and post-injury movements can be significantly altered
as a function of test substrates. This relationship between fibre density and rate of
motility can also be found within portions of individual cells since a graded
phenomenon of fibre expression occurs even in regions of motile cytoplasm.
DISCUSSION
Our experiments tested the effect of extracellular matrix molecules on the
response of EC to injury in vitw. We found that the rate and extent of post-injury
EC motility were modulated by the matrix on which the monolayers were grown and
that the intracellular arrangement of actin was also profoundly altered as a function
of culture substrate. EC grown on uncoated glass or glass-BSA microscope
coverslips expressed the greatest degree of post-injury migration into the wounded
region. When these migratory cells were fixed and stained with fluorescent anti-actin
(Fig. 1), few actin-rich stress fibres were seen; and, those fibres that were apparent
were fine (<0-lpim wide). In contrast, EC grown on type IV (basement
membrane) collagen exhibited a significant increase in stress fibre expression; and
this population of EC was virtually immobilized. Thus, the extracellular matrix
modulated both the rate of post-injury EC movement and the density of EC stress
fibres. The fact that EC motility and stress fibre density were found to be inversely
related is consistent with previous studies relating cell motility and the configuration
of the cytoskeleton (Couchman & Rees, 1979; Herman el al. 1981, 1982; Lewis et
al. 1982).
Our findings corroborate the results of earlier studies and show that EC remaining
at the site of an inflicted wound in vitw can migrate into the 'denuded zone' before
proliferation (Selden & Schwartz, 1979; Schwartz & Benditt, 1973, 1975; Schwartz,
Haudenschild & Eddy, 1978). Earlier studies in which cytochalasin B was used to
Anti-actin localization in wounded endothelium
29
inhibit post-injury migration (Selden & Schwartz, 1979) suggested a role for F-actin
in the motile response to injury. Our fluorescent antibody probe of intracellular actin
indicates that not only is actin localized in these regions of post-injury cytoplasm,
but specific cytoplasmic arrangements of actin may be needed for wound repair.
Only when actin is distributed diffusely throughout a particular cellular zone do the
cells express motile behaviour in that region. Thus, upon EC injury, a situation in
which EC motility is needed, contractile proteins become organized in a fashion that
must allow the generation of motile force production. On the basis of this
fluorescence work, we predict that short, individual actin filaments and oligomers of
actin are present in these regions of motile cytoplasm along the wound edge, and
electron microscopic analyses with labelled antibodies are needed to solve this
problem.
Our findings suggest a dual role for aortic EC actin. During quiescence or in an
uninjured state, bundles of coaxially aligned contractile proteins serve largely to
maintain intimal integrity and must explain the presence of stress fibres that are
aligned parallel to blood flow in regions of intima prone to injury (Herman et al.
1982; Wong et al. 1983). In the uninjured, non-motile regions of our monolayers we
found stress fibres to be randomly arranged, suggesting that blood flow may
modulate cell and stress fibre orientation (White et al. 1983; Franke et al. 1984).
Moreover, it has been suggested that stress fibres, because of their location within
the vascular tree and alignment to blood flow, serve to distribute blood shearing
forces across the intima (Herman et al. 1982; Wong^f al. 1983). These studies also
suggest that stress fibres promote adhesion of the EC to the extracellular substrate,
thereby strengthening attachment between EC and the basal laminae. Perhaps an
earlier finding of purported stress fibre involvement in EC motility following injury
in vivo (Gabbiani et al. 1983) was confounded by auto-fluorescence from medial
collagen fibres since the fluorescein-labelled fibres seen in situ were orthogonally
oriented, i.e. normal to blood flow.
Because our experiments used clones of EC in each group, and cell conditions
were held constant for each EC monolayer except for the substrate on which the cells
were cultured, we feel that the extracellular matrix may directly influence the
cellular response to injury. We have found that the extracellular matrix modulates
the rate of cellular motility and cytoplasmic arrangement of actin, perhaps by acting
through the plasma membrane via specific connecting proteins (Hay, 1981, 1982).
This, in turn, may affect the three-dimensional array of the EC cytoskeleton, which
could determine the rate at which EC could migrate into the wound and effect
recovery. Perhaps the expression of adhesion-plaque proteins is modulated following
injury. These molecules could participate in mediating an interaction between the
extracellular matrix and the cytoskeleton (Geiger, 1981; Burridge & Feramisco,
1981).
Other researchers have found that cell motility, cell shape and substrate are
inter-related (Folkman & Moscona, 1975; Hynes & Destree, 1978; Hay, 1981). A
direct relationship between the extracellular fibronectin arrangement and the
intracellular array of actin in fibroblasts has been suggested by Hynes & Destree
30
W. C. Young and I. M. Herman
(1978). We found that cells grown on type IV collagen had a greater surface area and
were thus more spread out than cells grown on glass or other substrates, suggesting
enhanced adhesion of the EC to the underlying substrate. It has been shown that the
phenotype and behaviour of rat epidydimal EC cultures are regulated by the
composition of the extracellular matrix. Cells grown on collagens I and III
proliferate and rarely assume a tube-like configuration, whereas cells grown on type
IV collagen aggregated, assumed a tube-like configuration, and showed little
evidence of proliferative capability (Madri & Williams, 1983). The pattern of EC
sprouting and rate of proliferation have also been found to be influenced by the
composition of the extracellular matrix and the particular conformation of collagen,
which may regulate cell behaviour (Schor, Schor & Allen, 1983; Macarak &
Howard, 1983). A lack of wound-induced migration across basement membrane
collagen may be an EC-specific response to injury since this matrix stimulates the
motility and wound closure in epithelial cells in situ (Radice, 1980). Clearly, more
work is needed to define better the role of EC attachment to the underlying substrate
and how this association relates to growth control as a function of the cytoskeleton.
The role of the proteoglycans in the pathogenesis of atherosclerosis has also been
discussed and studied (Wight, 1980); however, extracellular matrix-EC interactions
have not been previously related with respect to the EC motile response to injury.
Clearly, future experiments must ascertain if any cellular synthesis and, or,
exporting of glycosaminoglycans or glycoproteins occurs during the migratory
response to injury in EC cultured on different extracellular matrix molecules in
vitro; and, if this can influence the EC contractile protein apparatus and, or,
motility. Our findings represent the first attempt to link the composition of the
extracellular substrate with the EC migratory response to injury. Even in conditions
(growth on glass) that allow extensive post-injury cell motility, coverage of the
injury site may not be complete (Fig. ID); and, as a result, a 'leaky' endothelial
lining is seen in vitro. Moreover, under circumstances of limited EC motility in
vitro, as we found with wounded monolayers grown on gelatin or type IV collagen,
extremely large areas of what would be analogous to uncovered basement membrane
are found, thereby allowing platelets and other formed elements of the blood access
to this 'subintimal space' (Ross & Glomset, 1976). These findings may help explain
why EC adjacent to regions of injury in vivo are prevented from rapid repopulation
of the wound site.
We thank Dr Pat D'Amore for helpful discussions concerning the development of the
experiments and critical reading of the manuscript, and Dr Bruce Batten for the use of his
microscope. This work was supported in part by an American Heart Association (AHA)
Grant-In-Aid to I.M.H. and an Alpha Omega Alpha Research Award to W.C.Y.; l.M.H. is an
Established Investigator of the AHA.
REFERENCES
BURRIDGE, K. & FERAMISCO, J. R. (1981). Alpha actinin and vinculin from non-muscle cellscalcium sensitive interactions with actin. Cold Spring Hai-bor Svmp. quant. Biol. 46 (P2),
587-597.
Anti-actin
localization
in wounded
endothelium
31
COUCIIMAN, J. R. & REES, D. A. (1979). Behaviour of fibroblasts migrating from chick heart
explants: Changes in adhesion, locomotion and growth, and in the distribution of actomyosin
and fibronectin. J. Cell Sci. 39, 149-165.
DEWEY, C. F. J R , BUSSOLARI, S. R., GIMBRONE, M. A. JR & DAVIES, P. F. (1981). The
dynamic responses of vascular endothelial cells to fluid shear stress. J. biomech. Eng. 103(3),
173-185.
FOLKMAN, J. & MOSCONA, A. (1978). The role of cell shape in growth control. Nature, Land. 27,
345-349.
FRANKE, W. W., GRAFE, M., SCIINITTLER, H., SEIFFGE, D. & MITTKRMAYER, C. (1984).
Induction of human vascular endothelial stress fibers by fluid shear stress. Nature, Land. 307,
648-650.
GABBIANI, G., GABBIANI, F., LOMBARDI, D. & SCHWARTZ, S. M. (1983). Organization of actin
cytoskeleton in normal and regenerating arterial endothelial cells. Proc. natn. Acad. Sci. U.SA.
80, 2361-2364.
GEIGER, B. (1981). Involvement of vinculin in contact induced cytoskeletal interactions. Cold
Spring Harbor Symp. quant. Biol. 46(P2), 671-682.
GOTTLIEB, A., HEGGENESS, M. H., ASH, J. F. & SINGER, J. (1979). Mechanochemical proteins,
cell motility and cell-cell contacts: The localization of mechano-chemical proteins inside
cultured cells at the edge of an in vitro wound. J. cell. Phys. 100, 563-578.
HAY, E. D. (1981). Extracellular matrix. J . Cell Biol. 91, 2055-2275.
HAY, E. D. (1982). Interaction of embryonic cell surface and cytoskeleton with extracellular
matrix. Am. J. Anat. 165, 1-12.
HERMAN, I. M., CRISONA, N. J. & POLLARD, T. D. (1981). Relation between cell activity and
the distribution of cytoplasmic actin and myosin. J. Cell Biol. 90, 84-91.
HERMAN, I. M. & POLLARD, T . D. (1979). Comparison of purified anti-actin and
fluorescent-heavy meromyosin staining patterns in dividing cells. J. Cell Biol. 80, 509-512.
HERMAN, I. M., POLLARD, T . D. & WONG, A. J. (1982). Contractile protein in endothelial cells.
Ann. N.Y. Acad. Sci. 401, 50-60.
HYNES, R. 0 . & DESTREE, A. T . (1978). Relationship between fibronectin (LETS protein) and
actin. Cell 15, 875-886.
JAFFE, E. A., HOYER, L. W. & NACIIMAN, R. L. (1973). Synthesis of antihemophilic factor
antigen by cultured human endothelial cells. J. din. Invest. 52, 2757-2764.
LEWIS, L., VERNA, J. M., LEVINSTON, S. S., MAREK, L. & BELL, E. (1982). The relationship of
fibroblast translocations to cell morphology and stress fibre density. J. Cell Sci. 53, 21-37.
MACARAK, E. J. & HOWARD, P. S. (1983). Adhesion of endothelial cells and
extracellular matrix proteins. .7. cell. Phys. 116, 76-86.
MADRI, J. A. 8C WILLIAMS, S. K. (1983). Capillary endothelial cell cultures: phenotypic
modulation by matrix components. J . Cell Biol. 97, 153-166.
RADICE, G. P. (1980). Locomotion and cell-substratum contacts of Xenopus epidermal cell in
vitro and in situ. J. Cell Sci. 44, 201-223.
Ross, R. & GLOMSET, J. A. (1976). The pathogenesis of atheroscerosis. New Engl.J. Med. 295,
369-377.
SCHOLEY, M. M., GIMBRONE, M. A. & CORTRAN, R. S. (1977). Cellular migration and
replication. Lab. Invest. 36, 18-25.
SCHOR, A. M., SCIIOR, S. L. & ALLEN, T . D. (1983). Effects of culture conditions in the
proliferative morphology and migration of bovine aortic endothelial cells. J. Cell Sci. 62,
276-287.
SCHWARTZ, S. M. & BENDITT, E. P. (1973). Cell replication in the aortic endothelium: A new
method for studying of the problem. Lab. Invest. 28, 699-707.
SCHWARTZ, S. M. & BENDITT, E. P. (1975). The aortic intima. II. Repair of the aortic lining
after mechanical denudation. Am. J. Path. 81, 15-21.
SCHWARTZ, S. M., HAUDENSCIIILD, C. C. & EDDY, E. M. (1978). Endothelial regeneration. I.
Quantitative analysis of initial stages of endothelial regeneration in rat aortic intima. Lab. Invest.
38, 568-580.
SELDEN, S. C. & SCHWARTZ, S. M. (1979). Cytochalasin B inhibition of endothelial proliferation
at wound edges in vitro. J. Cell Biol. 81, 348-354.
32
W. C. yoMM£ W /. M. Herman
W H I T E , G. E., GIMBRONE, M. A. & FUJIWARA, K. (1983). Factors influencing the expression of
stress fibres in vascular endothelial cells in situ. jf. Cell Biol. 97, 416-424.
W I G H T , T . N. (1980). Vessel proteoglycans and thrombogenesis. Prog. Heniost. Thromb. 5, 1-39.
WONG, A. J., POLLARD, T . D. & HERMAN, I. M. (1983). Actin filament stress fibres in vascular
endothelial cells in vivo. Science 219, 867-869.
ZETTER, B. (1981). The endothelial cells of large and small blood vessels. Diabetes'30 (suppl. 2),
24-28.
(Received 25 April 1984 -Accepted, in revised form, 20 July 1984)