Original Contributions |
From the Departments of Pathology (S.P.E., T.N.W.) and Biochemistry (J.C.A.), University of Washington, Seattle.
Correspondence to Thomas N. Wight, University of Washington School of Medicine, Department of Pathology, Box 357470, Seattle, WA 98195. E-mail tnw{at}u.washington.edu
| Abstract |
|---|
|
|
|---|
Key Words: cell adhesion atherosclerosis mitosis locomotion
| Introduction |
|---|
|
|
|---|
Versican is a major chondroitin sulfate PG (CSPG) produced by SMCs11 and has been shown to bind to HA.12 13 Versican synthesis is upregulated during SMC proliferation in vitro.14 Both HA and versican are major components of primary atherosclerotic plaques and the loose extracellular matrix of human restenotic lesions.5 15 16 17 18 Colocalization of HA and versican around SMCs in atherosclerotic plaques17 suggests that these molecules function together during processes involved in lesion formation. Although HA and versican are abundant in lesions of atherosclerosis, little is known regarding the capability of SMCs to organize these molecules in the pericellular matrix or regarding their functional role(s) during SMC proliferation and migration.
Therefore, we have examined the organization of HA and versican in the pericellular matrix of human SMCs in vitro, using a particle-exclusion assay, time-lapse video imaging, and electron microscopy. We show that formation of HA- and versican-rich pericellular matrices by SMCs is dynamic and rapid, and occurs preferentially during the detachment phase of cell migration and premitotic cell rounding. We also show that HA oligosaccharides, which block formation of the pericellular matrix and binding of HA to the cell surface, can inhibit proliferation and migration and alter cellular morphology.
| Methods |
|---|
|
|
|---|
Cell Culture
Arterial SMCs were derived from the media of normal
newborn human thoracic aortas by explant and were maintained in DMEM
containing 10% FBS.19 These cells were kindly provided by
Dr Russell Ross, Department of Pathology, University of Washington.
Cells stained positively for SMC
-actin (data not shown) had the
characteristic hill-and-valley morphology of cultured SMCs and were
used between passages 7 and 14.
Visualization of Pericellular Matrix
Visualization and morphometric analysis of
HA/versican matrices were done by using a particle-exclusion
assay as previously described.20 With this assay, an
otherwise invisible matrix that surrounds a number of cell types in
vitro can be visualized by virtue of the steric exclusion properties of
HA-PG gels. In brief, 1x105 cells were plated in
35-mm tissue culture dishes in DMEM containing 10% FBS. After 24
hours, 750 µL of a suspension of fixed and washed human erythrocytes
(
108/mL) was added to the cells and allowed to
settle for 15 minutes. Pericellular matrix formation was quantitated
from photographs by using a digitizing graphics pad to measure the area
delimited by red blood cells and the area delimited by the cell
membrane to give a coat-to-cell ratio.20 A ratio of
1.0 indicates no matrix. For quantitating the proportion of
cells with a matrix, cells with a ratio of
2.0 were considered to
have pericellular matrices.
Cells were treated with HA-specific Streptomyces
hyaluronidase at 1 U/mL of medium for 1 hour before visualization with
the particle assay, to demonstrate the structural dependence of the
pericellular matrix on HA. In some cases, the enzyme was added after
the particles had settled, to view individual pericellular coats before
and after digestion of HA (see Figure 1
).
To examine the effect of HA oligosaccharides on pericellular
matrix formation, cells were incubated with the
oligosaccharides at 30 µg of uronate/mL for 1 hour before
addition of the red blood cells.
|
The kinetics of pericellular matrix formation was observed by using time-lapse video microscopy in combination with the particle-exclusion method. Time-lapse video analysis was done by using a Nikon inverted microscope equipped with an extralong phase condenser and a heated incubator (Nikon, Inc). The cells were imaged through a video camera (Series 65; Dage-MTI, Inc) equipped with a high-resolution vidicon tube and connected to a time-lapse recorder (Model TLC2051R, Gyyr Products). The videotape was viewed on a monochrome monitor (Model 925; Lenco). HEPES (50 mmol/L) was added to buffer the medium. To evaluate the length of time required for pericellular matrix formation, cells with little or no matrix at the beginning of the assay were selected and monitored for 4 hours.
Preparation of Hyaluronan Fragments
Two hundred milligrams of HA was dissolved in 20 mL of 0.1 mol/L
ammonium acetate, pH 6.0, and digested for 72 hours with 50 U of
Streptomyces hyaluronidase. The enzyme was
inactivated by boiling for 20 minutes and the HA fragments
were fractionated by ultrafiltration through Centricon
microconcentrators (Amicon) with molecular weight cutoffs of 50 kDa,
followed by 3 kDa. HA fragments passing through the Centricon 50 but
retained by the Centricon 3 are designated
HA350k and fragments passing through the
Centricon 3 are designated HAoligo.
Analysis by chromatography on Biogel P4
(Bio-Rad) showed that the oligosaccharide fraction consisted
primarily of hexasaccharides (
45%) and
tetrasaccharides (
30%) with small amounts of octa-, deca-,
and dodecasaccharides comprising the remainder. Samples were
lyophilized, resuspended in PBS, and filter-sterilized. Uronic acid
concentration was measured by the orcinol method.21
As a control to verify that the effect of the oligosaccharides
was not caused by a low molecular weight contaminant in the HA
preparation, disaccharides were prepared by exhaustive
digestion of HA with chondroitin ABC lyase (ICN
Pharmaceuticals), and filtration through Centricon 3 membranes as
described above. This preparation consisted of 95% HA
disaccharide, with only minor amounts of
tetrasaccharides and hexasaccharides. High molecular
weight HA was dissolved in PBS and sterilized by overnight exposure to
UV light.
Growth and Migration Studies
For growth-inhibition studies with HA and
HAoligo, SMCs were plated in 24-well plates in
the presence of 10% FBS. On the following day, cells were switched to
medium containing 1% human plasma derived serum for 2 days before
addition of PDGF and HA fragments (above). Alternatively,
HAoligo was added to cells maintained
continuously in 10% FBS. Cells were harvested with trypsin at the
appropriate times and counted with a Coulter counter. The effects of HA
and HAoligo on SMC migration on collagen-coated
membranes in response to PDGF was assessed by using a modified Boyden
chamber.22 The number of cells that migrated to the
underside of the membrane over a 4-hour period was determined by
counting nuclei in 6 fields from each of 3 replicate wells. Data are
expressed as number of cells per field. Statistical significance
was evaluated with a Student's t test.
Immunocytochemistry
Cells were seeded in 4-well chamber slides and allowed to adhere
for 24 hours. After rinsing with PBS, the cultures were fixed with 4%
paraformaldehyde for 20 minutes at 22°C, then rinsed
in PBS. To localize HA, cells were stained with biotinylated HA-binding
region from cartilage PG at a dilution of 2 µg/mL. This probe
specifically recognizes HA.23 Controls for HA staining
included Streptomyces hyaluronidase digestion of the cells
before fixation (4 U/mL in DMEM, 1 hour, 37°C) and preincubation of
the biotinylated HA-binding region with 100 µg of HA. Rabbit
polyclonal antibody specific for human versican, VC-E, was used at a
1:250 dilution, and was kindly provided by Dr Richard LeBaron
(University of Texas at San Antonio).24 Normal rabbit
IgG was used as a negative control. Antibodies and the biotinylated
HA-binding region were applied in PBS containing 1% BSA for 1 hour at
22°C. After rinses, the versican antibody was detected by using a
biotinylated rat anti-rabbit 2o antibody
at a 1:1000 dilution. The samples were rinsed and then incubated with a
1:500 dilution of a streptavidinhorseradishperoxidase conjugate.
After rinses, color was developed for 10 minutes by using
3-amino-9-ethylcarbazole in 0.05 mol/L sodium acetate, pH 5.0. The
substrate solution was prepared by dissolving 2 mg
3-amino-9-ethylcarbazole in 0.5 mL of DMSO and adding 10 mL of 0.05
mol/L sodium acetate, pH 5.0; 10 µL of 30%
H2O2 was added just before
use. Cells were mounted with Crystal Mount (Biomeda
Inc).
Pericellular Matrix Ultrastructure
Cells grown on coverslips were rinsed in serum-free medium and
fixed with Karnovsky's fixative in 0.1 mol/L sodium cacodylate, 2
mmol/L CaCl, and 5% sucrose, containing 0.2% ruthenium
red25 (to precipitate PGs), for 1 hour at 22°C with or
without prior digestion by Streptomyces hyaluronidase. Cells
were rinsed with the cacodylate buffer containing 0.1% ruthenium red,
and then postfixed in 1% OsO4 in cacodylate
buffer, 0.05% ruthenium red for 1 hour at 22°C. Coverslips were
rinsed gently by dipping several times in PBS, and then
H2O. Cells were air-dried to maximize
preservation of the fragile HA network, then coated for scanning
electron microscopy. Other cultures were fixed as above, embedded in
Epon, and examined by transmission electron microscopy.
HA Binding Assay
Fluorescein-labeled HA was prepared according to
published methods.26 For the HA binding assay, SMCs were
plated in 60-mm culture dishes and allowed to attach for 24 hours.
Cells were rinsed with PBS and incubated with competitor (unlabeled
high molecular weight HA, or HA fragments and oligosaccharides)
at 300 µg of uronate/mL in 1.5 mL serum-free DMEM for 20 minutes at
4°C. Fluoresceinated HA (2 µg of uronate/mL) was then
added and cells were incubated for 10 minutes at 4°C with rocking.
Cells were then washed 3 times with cold PBS, and solubilized with 0.25
mol/L NaOH. Bound fluorescence was determined by using a
fluorometer and compared with a standard curve of
fluorescein-labeled HA.
| Results |
|---|
|
|
|---|
40% of the cells displayed HA-dependent
matrices.
HA was localized histochemically in these cultures by using a
biotinylated protein (HA-binding region from bovine cartilage aggrecan)
that specifically binds HA (Figure 1C
). Staining for HA was
positive in a flocculent material that localized to the cell surface
and culture dish in a pattern similar to the pericellular gel outlined
by the particle assay in living cells. Pericellular staining was most
prominent in cells with a motile phenotype, ie, having a
lamellipodium and a thin trailing uropod. Staining was usually
concentrated around the trailing process. Thin cellular processes
resembling retraction fibers were often seen extending into the areas
where the HA-rich matrix was visible. Some staining was present in
areas of membrane ruffling. Digestion of the cultures with
Streptomyces hyaluronidase before fixation abolished
virtually all of the pericellular staining (Figure 1D
). However,
there was prominent staining within vesicular structures that was not
affected by hyaluronidase digestion. These results are
consistent with recent observations of HA in caveolae in
vascular cells in vivo.28 All staining in our studies
was abolished by preincubation of the biotinylated HA-binding protein
with an excess of HA, confirming its specificity for HA (data not
shown).
Immunostaining with an antibody specific for versican
was similar to HA staining, and was detected both in the flocculent
material on the culture dish surrounding the cells and in material
closely apposed to the cell surface on the main cell body (Figure 1E
). Staining was also prominent on thin trailing processes.
Digestion of the cultures with Streptomyces hyaluronidase
removed most of the pericellular versican
immunostaining (data not shown), thus confirming
previous observations that CSPGs are tethered to cells via association
with HA.29 30 31 Versican was not detected
intracellularly under these conditions. No staining was seen with
normal rabbit IgG (Figure 1F
).
Cells grown on coverslips for 24 hours were fixed in the presence of
ruthenium red to precipitate the PGs,25 air-dried to
preserve the cell coat, and examined by scanning electron microscopy
(Figure 2
). Fine filaments (10 to 40 nm
in thickness) extending away from the cell surface were prominent on
the thin trailing processes (Figure 2A
and 2B
). The filaments
often exceeded 10 µm in length and were decorated with ruthenium
red granules, and were similar in size (50 to 200 nm) to the large PG
granules visualized by transmission electron microscopy of the SMC
extracellular matrix in vitro (Figure 2
, inset) and in
vivo.25 The strands and most of the granules were removed
by treatment with Streptomyces hyaluronidase before fixation
(Figure 2D
). This is consistent with the abolishment of
staining by the enzyme as described above, and suggests that these
structures represent HA/versican aggregates.
|
The thin filaments occurred either as individual strands (Figure 2A
) or as a tangled network (Figure 2B
). The ruthenium
redpositive granules were often seen at intersection points in the
network, suggesting that the PGs serve as points of entanglement for
the strands of HA. In some cells, dense clusters of tightly
"beaded" PG granules and HA filaments were seen on the cell surface
(Figure 2C
), in areas of membrane ruffling, and in the
pericellular space (not shown). These results are consistent
with earlier observations in chondrocytes,30 31 and
suggest that the pericellular matrix gel of SMCs, which is made
apparent by particle exclusion, forms as a result of the accumulation
and entanglement of HA and aggregated PG in the pericellular space.
Association of HA/Versican Matrices With Motile and Mitotic
Cells
On closer examination of pericellular matrices in live cultures,
as shown by the particle-exclusion assay, SMCs with a flattened and
spread morphology displayed very little if any surrounding matrix, and
those with a motile morphology displayed prominent pericellular
matrices (Figure 3A
and 3B
). The
pericellular matrix was present at predictable places on locomoting
SMCs, specifically around the trailing process, which was detached from
the substrate and retracted. Very little or no matrix was seen at the
leading lamellipodium except in areas of membrane ruffling,
consistent with HA and versican staining (see Figure 1
).
This suggests that HA secretion or pericellular matrix organization may
be directionally controlled in locomoting cells.
|
In a model for cell migration, synthesis of HA was reported to be
upregulated in SMC monolayers after scratch wounding.10 At
24 hours after wounding, the particle assay showed large deposits of
pericellular matrix around and between cells at the wound edge (Figure 3C
). Approximately 90% of the migrating and proliferating cells
at the wound edge displayed prominent pericellular matrices. In
contrast, only
40% of the cells in sparse cultures display
pericellular matrices under normal growth conditions. This suggests
that active organization of the HA/versican-rich matrix is important
for wound-induced migration.
Virtually all mitotic SMCs (identified by a rounded shape and clearly
visible chromosomes) had an investment of pericellular matrix
surrounding or immediately adjacent to them (Figure 4A
and 4B
). The thick, abundant matrix
surrounding dividing cells stained intensely for HA and versican
(Figure 4C
and 4D
). Figure 4D
shows an example of a late
telophase cell with abundant versican staining in the matrix in the
cleavage furrow, suggesting that the matrix may facilitate separation
of daughter cells.
|
Time-lapse video microscopy, in conjunction with the particle assay,
was used to obtain information on the kinetics of pericellular matrix
formation and its relation to cell migration and proliferation. These
data provided clear evidence that HA/versican-rich matrix organization
is a rapid process occurring in migrating and mitotic cells, and that
there is a dynamic interdependence between pericellular matrix
formation, cell shape, and detachment from the substrate (Figure 5
). In as little as 14 minutes, a matrix
coat several micrometers thick was organized around a
trailing process, just before its retraction.
Simultaneously, pericellular matrix formed around the main
body of the cell as it detached and rounded up. Cell division was
completed
1 hour after pericellular matrix formation, after which
the matrix appeared to dissolve into the medium. Cells that remained
stationary during filming did not display any organized pericellular
matrix. In motile cells, pericellular matrix appeared to form
episodically at sites of membrane ruffling at the leading edge and
continuously around the trailing process as the cell moved (Figure 5
, lower right corner). Observations that some of the matrix
moved with the cells and some was sloughed and left behind suggest that
this is a highly malleable matrix that facilitates detachment and
sliding of cellular processes across the substrate. In several other
cases, a large investment of pericellular matrix appeared to impede
cells from reestablishing firm adhesion to the plastic substrate. These
data are consistent with the reported involvement of HA and
versican in proliferating cells and with a proposed role for HA in cell
detachment and rounding.2 7 32
|
HA Oligosaccharides Inhibit Proliferation and Migration
of SMCs
It has been previously shown that a hexasaccharide was the
smallest fragment of HA capable of binding to cell surface
receptors33 and that HA hexasaccharides were
effective at preventing formation of HA-dependent pericellular matrices
in chondrocytes and also in other cells.20 Therefore, we
used a preparation of small HA oligosaccharides
(HAoligo) to determine if disruption of the
association of the pericellular matrix with the cell surface could
affect cell proliferation or migration. Cells treated with
HAoligo did not display pericellular matrices
(Figure 6
). The
HAoligo competed efficiently for binding of
fluorescein-labeled HA to the cell surface (Figure 6
).
|
In cell proliferation assays, HAoligo
significantly inhibited cell growth as measured by cell number (Figure 7A
). In contrast, slightly larger
fragments of HA (HA350k) and high molecular
weight HA had no inhibitory effect, or even a slight
stimulatory effect, on cell growth. The inhibitory effect
of HAoligo was dose dependent (Figure 7B
),
and occurred after PDGF stimulation of serum-starved, quiescent
cultures (Figure 7C
) and in cells cultured continuously in the
presence of serum (Figure 7D
). There was no decrease in cell
number in quiescent cells treated with HAoligo,
suggesting that this was not a cytotoxic effect (Figure 7C
). The
release of lactate dehydrogenase into the medium was not altered by
treatment with HAoligo, also confirming that this
was not a cytotoxic effect (data not shown).
|
The lack of an effect by high molecular weight HA suggests that the growth inhibition was not caused by a contaminant in the commercial HA preparation. To further verify that there was not a low molecular weight contaminant with antiproliferative activity in the HA preparation, disaccharides of HA were prepared by digestion of HA with chondroitinase ABC and fractionated by the same procedure used for the oligosaccharides. The HA disaccharides had no effect on cell growth (data not shown). Furthermore, protease treatment of the HAoligo preparation did not alter the inhibitory activity (data not shown).
After
2 days of incubation in the presence of the HA
oligosaccharides, clear morphological changes in the cells
became apparent. Control cells typically displayed an elongated,
spindle-shaped morphology and cells treated with
HAoligo became flattened and highly spread, with
prominent stress fibers apparent even under phase-contrast microscopy
(Figure 8A
and 8B
). Area measurements
confirmed that HAoligo altered cell shape (Figure 8C
), suggesting that the presence of the pericellular matrix at
the cell surface is important for SMCs to maintain their fusiform
morphology in vitro.
|
As described above, pericellular matrix formation was correlated
preferentially with detachment and sliding of the trailing uropod
during cell locomotion, as observed by time-lapse video microscopy.
Therefore, we examined the effect of disruption of pericellular matrix
formation by HAoligo on PDGF-induced cell
migration by using a modified Boyden chamber assay.
HAoligo inhibited migration on collagen-coated
membranes in response to PDGF in a dose-dependent manner (Figure 9
). In contrast, similar concentrations
of larger HA fragments or high molecular weight HA did not have a
significant effect.
|
It was also observed in freshly plated cultures that most of the cells
tended to elongate uniaxially into the characteristic motile shape
whereas some cells tended to spread out radially and uniformly,
maintaining a more flattened appearance. The particle-exclusion assay
showed that the pericellular matrix formed around extending
filopodia in the cells that tended to elongate uniaxially after
plating, suggesting that the matrix may also facilitate extension of
cellular processes along the substrate (Figure 10A
). Treatment of freshly plated SMCs
with HAoligo significantly inhibited the
elongation ("spreading") of the cells into their characteristic
spindle shape (Figure 10B
).
|
| Discussion |
|---|
|
|
|---|
Our observations of pericellular matrix formation and its relation to SMC proliferation and migration in vitro are consistent with observations of atherosclerotic lesion formation in vivo. For example, colocalization of HA with proliferating cells after balloon injury of rat carotid arteries5 and double-injured rabbit arteries (T.N. Wight, unpublished data, 1998). correlates well with our observations of HA/versican-rich pericellular matrix formation around mitotic cells and in migrating cells after wounding in vitro (present study and that by Savani et al10 ). Furthermore, HA and versican colocalize around SMCs in both early and advanced lesions of primary atherosclerosis in nonhuman primates,17 and in the loose, myxoid extracellular matrix of human restenotic arteries.5 16 Although these molecules are found in the media of normal vessels, the staining is usually patchy and much less intense than in developing lesions.17 These observations suggest that formation of HA/versican-rich pericellular matrices in SMCs is crucial for processes leading to lesion formation.
There is evidence that HA can regulate cell growth and migratory activity in SMCs and other cells.9 10 34 35 36 37 Our results disagree with a previous study in which bovine aortic SMCs did not respond to HA oligosaccharides, but in bovine endothelial cells, DNA synthesis was stimulated.34 This may be because of differences between species, the amounts of endogenously produced HA or HA-binding molecules, or other experimental variables. We show that rapid, active elaboration of the HA/versican-rich matrix at the cell surface occurs at a specific time in motile cells and in premitotic SMCs, ie, just before detachment and/or rounding. This extends previous observations in fibroblasts that HA synthesis was high during cytokinesis and was necessary for cell detachment and mitosis.7 This is significant because detachment and mitotic cell rounding would be expected to take place during G2/M, and it has been shown that exogenous, soluble, recombinant RHAMM blocks ras-transformed fibroblasts in the G2/M phase of the cell cycle, possibly by competition for binding of HA to cell surface receptors.36 The inhibition of growth and morphological changes in response to HA oligosaccharides suggests that the presence of high molecular weight HA or, perhaps, the entire organized pericellular matrix at the cell surface is required for progression through the cell cycle. It is possible that the HA oligosaccharides affect HA receptormediated signaling, in view of recent evidence for regulation of mitogen-activated protein kinase activity and other intracellular signaling events by RHAMM38 and CD44.39 40 41 42 CD44 is known to be important for pericellular matrix formation in other cells,20 31 and dimerization and clustering of CD44 variants reportedly increases hyaluronate-binding capacity.43 In our experiments, only the small HA oligosaccharides (<3000 Da) were effective inhibitors of growth and migration whereas larger fragments (>3000 Da) were ineffective. One possible explanation that we are currently investigating is that the smaller HA fragments may prevent dimerization and clustering of CD44 and that this plays a role in SMC proliferation.
Time-lapse video microscopy allowed us to view the relation between pericellular matrix formation and cell behavior in individual cells. These data provide compelling evidence that one function of the HA/versican-rich pericellular matrix is to destabilize cellular attachment to the substrate during migration and rounding. The HA-dependent pericellular matrix may also help maintain the fusiform morphology in vitro. HA and versican (or related CSPGs) have been reported to have both adhesive and antiadhesive properties.32 44 45 46 Thus, the mechanism(s) by which HA or CSPGs mediate their effects on adhesion or migration remain speculative. However, one proposed mechanism by which this matrix may mediate cell detachment is based on the steric exclusion properties of HA-PG gels.2 30 This is the basis of the particle-exclusion assay, and was clearly evident in the rapid expulsion of the red blood cells away from the SMCs during matrix formation, and the subsequent detachment of the cells from the substrate. This property may also be important in regulating cellcell contact, such as during limb morphogenesis where an HA-rich extracellular matrix can regulate mesenchymal cell condensation,2 47 or in facilitating separation of daughter cells after division (Tammi and Tammi48 and the present study). A role for HA in cell detachment is also supported by the finding that loosely adherent Chinese hamster ovary cell variants, isolated on the basis of ease of detachment, exhibited higher levels of HA synthesis compared with more tightly adherent variants,49 and that exogenous HA stimulated detachment of BALB 3T3 cells.32 Previous immunocytochemical studies have demonstrated in cultured fibroblasts that HA, versican, and CD44 are localized over the entire cell surface and concentrated in areas of membrane ruffling and in the perinuclear vicinity, but were excluded from focal contacts, suggestive of a role for these molecules in "de-adhesion."50 51 Our results are entirely consistent with these observations and indicate that actual organization of the matrix at the cell surface is important for this function. These studies suggest that HA and versican are essential by-products, inseparable from and necessary for specific events during SMC migration or proliferation, ie, cell detachment. It is noteworthy that a preexisting HA- and versican-rich matrix may not be required because the cells appear to elaborate and use it simultaneously. We speculate that the relative amount and material properties of the nonfibrillar matrix in the pericellular milieu plays a role in the ability of cells to detach from each other and from more adhesive matrix components, to round up or migrate in vivo.
Tight control of the physical mucoid properties of the pericellular matrix may be an important regulatory mechanism for SMCs during atherogenesis or the response to injury and could affect HA receptor signaling. HA is capable of self-association, and forms networks even at low concentrations.52 The length of HA may be critical for the integrity, viscosity, and malleability of the HA network. Inclusion of an aggregating CSPG could act to increase the viscosity of the gel,53 54 and add swelling pressure to the matrix because of the CS chains.20 30 CSPGs also decrease the permeability of the pericellular coat.55 CSPGs can potentially stiffen the HA network because of the shortening of HA, which results from the interaction during aggregate formation,56 enhancing the steric exclusion property. This could influence organization of other matrix components, or directly modify the effect of HA on cell behavior by altering cell/matrix tensegrity and, in turn, mechanically coupled signaling.57 These ideas may relate to previous studies showing inhibition of neural crest cell migration in response to exogenous CSPGs, thought to be mediated by interaction with cell surface HA.58
The idea that aggregate formation may occur in atherosclerotic lesions and that HA and versican function together is supported by electron microscopic studies of the extracellular matrix of lesions that show abundant ruthenium redpositive PG granules interconnected by a network of fine filaments directly adjacent to SMCs.25 As shown ultrastructurally in the present study, the network of hyaluronidase-sensitive filaments and PG granules surrounding SMCs in vitro is consistent with this, and is highly similar to endogenous pericellular matrices and those reconstituted from exogenous CSPG and HA in chondrocytes.31 It is clear that the structure of the pericellular matrix is dependent on HA.2 3 27 Furthermore, it has also been shown that aggregating CSPGs are required to visualize an expanded matrix by particle exclusion or other assays,20 30 although it is likely that other molecules are present. This suggests that the relative amounts of HA, versican, and stabilizers such as link protein, and the degree of aggregate formation, may be important for the general stability and material properties of the vascular extracellular matrix in vivo, as has been demonstrated for HA-PG aggregates in articular cartilage.59
An HA/CSPG-rich pericellular matrix could function to mediate cell detachment from other matrix components caused by the steric exclusion properties, and simultaneously facilitate cell rounding because of its apparently high malleability, which we have noted in the present study. Cells plated on highly malleable substrates tend to round up because of the traction forces that they place on the surrounding matrix.60 Relative to the tension-bearing elements of the extracellular matrix such as collagen fibrils, the nonfibrillar components are presumably much more malleable, providing less resistance to cellular traction forces, and therefore have direct bearing on cell shape and mechanically coupled signaling.61 Thus, in addition to integrin inactivation,22 HA/versican pericellular matrix formation potentially serves as another important, but infrequently considered, mechanism for controlling and counterbalancing the traction aspect of cell locomotion or other processes through this "lubricative" mechanism and the simultaneous steric interference with adhesion to fibrous matrix components. These findings may have important ramifications regarding the cell/matrix interface, adhesion and cell shaperelated signaling events, and in developing treatments for atherosclerosis and restenosis.
| Acknowledgments |
|---|
Received August 3, 1998; accepted September 21, 1998.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
S. Meran, D. W. Thomas, P. Stephens, S. Enoch, J. Martin, R. Steadman, and A. O. Phillips Hyaluronan Facilitates Transforming Growth Factor-{beta}1-mediated Fibroblast Proliferation J. Biol. Chem., March 7, 2008; 283(10): 6530 - 6545. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Kastner, G. J. Thomas, R. H. Jenkins, M. Davies, and R. Steadman Hyaluronan Induces the Selective Accumulation of Matrix- and Cell-Associated Proteoglycans by Mesangial Cells Am. J. Pathol., December 1, 2007; 171(6): 1811 - 1821. [Abstract] [Full Text] [PDF] |
||||