| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
vß3 and 5ß1 integrins
Craniofacial Developmental Biology and Regeneration Branch, National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, Maryland 20892, USA; and
* Graduate School of Environmental Earth Science, Hokkaido University, Sapporo, Japan
1Correspondence: CDBRB, NIDCR, NIH, Bldg. 30, Room 433, 30 Convent Dr., Bethesda, MD 20892, USA. E-mail address: hkleinman{at}dir.nidcr.nih.gov
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1ß1
1) by screening
559 overlapping synthetic peptides. C16, the most potent 1 chain
peptide, blocked laminin-1-mediated adhesion and was the only 1
chain peptide to block attachment to both collagen I and fibronectin.
This suggested that C16 was acting via a receptor common to these
substrates. We demonstrated that C16 is angiogenic in vivo. Affinity
chromatography identified the integrins 5ß1 and vß3 as
surface receptors. Blocking antibodies confirmed the role of these
receptors in C16 adhesion. C16 does not contain an RGD sequence and, as
expected, an RGD-containing peptide did not block C16 adhesion nor did
C16 act via MAP kinase phosphorylation. Furthermore, we identified a
C16 scrambled sequence, C16S, which antagonizes the angiogenic activity
of bFGF and of C16 by binding to the same receptors. Because the
laminin 1 chain is ubiquitous in most tissues, C16 is likely an
important functional site. Since the biological activity of C16 is
blocked by a scrambled peptide, C16S may serve as an anti-angiogenic
therapeutic agent.—Ponce, M. L., Nomizu, M., Kleinman, H. K.
An angiogenic laminin site and its antagonist bind through the vß3
and 5ß1 integrins.
Key Words: angiogenesis • laminin-1 • bFGF • endothelium • peptides
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. During angiogenesis, there is
initial protease degradation of the underlying endothelial cell matrix,
followed by cell migration, proliferation, and resynthesis of the
extracellular matrix. The formation of blood vessels is highly
regulated by different stimulators and inhibitors. Various proteinases
secreted from stromal cells during this process degrade the
extracellular matrix, thus releasing angiogenic promoters and
inhibitors such as vascular endothelial growth factor (VEGF), basic
fibroblast growth factor (bFGF), and several matrix-derived protein
fragments including angiostatin (plasminogen-derived) and endostatin (a
collagen XVIII fragment) (2
, 3)
.
In blood vessels, endothelial cells are in contact with a basement
membrane that contains laminin, a large ubiquitous glycoprotein that
exists in 12 different isoforms. Laminin is composed of three chains
(, ß, and ); five different , three ß, and three
chains have been identified. Ten of the 12 different heterotrimeric
isomers contain the 1 chain (4
5
6)
. The identity of the
laminin isoforms present in the endothelial cell matrix has not been
determined; nevertheless, polyclonal antibodies to laminin-1 (1,
ß1, and 1) recognize this matrix, suggesting the presence of at
least one of these three chains. Laminin-1 promotes the attachment of
endothelial cells in vitro, and the cells differentiate into
capillary-like structures when plated on a laminin-1-rich basement
membrane, Matrigel (7)
. Multiple binding sites for tumor
cells have been identified on laminin-1 (8
9
10)
and an
angiogenic sequence, IKVAV (ile-lys-val-ala-val), of the 1 chain has
been reported (11
, 12)
.
Since laminin is highly protease-sensitive and basement membranes are
degraded during angiogenesis, especially when tumor spread and growth
occur, we wanted to determine whether additional active sites exist for
endothelial cells. Multiple binding sites on laminin-1 have been
identified for tumor cells (8
9
10)
. Our laboratory has
recently duplicated all three laminin-1 chains using 12-mer overlapping
synthetic peptides; our goal was to identify sites on laminin that are
active for endothelial cells and angiogenesis. We used a synthetic
peptide approach because laminin can be cleaved into several large
fragments (> 200 kDa) with elastase. However, proteolytic cleavage of
these fragments gives rise to small products that cannot be isolated
for further analysis (4)
. We tested 559 peptides in
various in vivo and in vitro assays with endothelial cells, and 20
active sites were identified, including eight from the 1, five from
the ß1, and seven from the 1 chain (13
, 14)
. Four of
the active laminin 1 chain peptides (C25, C38, C75, and C102) showed
endothelial cell specificity from their inability to promote adhesion
with two different types of tumor cells, salivary gland cells, or to
induce neurite outgrowth on PC 12 cells, NG108–15, or cerebellar
granule cells. C16 (KAFDITYVRLKF) was the most potent 1 chain
peptide with a variety of cell types, including endothelial cells
(14)
. C16 blocked cell adhesion to laminin-1 more
effectively than other 1 chain peptides. Unexpectedly, C16 was also
the only 1 chain peptide to block cell attachment to collagen I,
fibronectin, and plastic (14)
. These latter data suggested
that C16 might be acting via a receptor that is common to these
substrates, such as an integrin.
In this study, we investigated the angiogenic activity of C16 in the chick chorioallantoic membrane (CAM) and rat aortic ring assays, and isolated its cell surface receptors by peptide affinity chromatography. Most important, we have identified a scrambled C16 peptide, C16S, that can inhibit the angiogenic activity of the parent peptide and bFGF. The scrambled peptide can also inhibit cell attachment to laminin-1 and to C16 in a dose-dependent manner. Furthermore, the scrambled peptide can bind to the same receptors as C16, suggesting that its inhibitory mechanism is mediated by competing with C16.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. All peptides used in this study were manually
synthesized as described previously (8
9
10)
, and included
C16 (KAFDITYVRLKF), C16S (DFKLFAVTIKYR), C16 Rev (FKLRVYTIDFAK), C15
(SINLTLHLGKAAFD), and C57 (APVKFLGNQVLSY). Cycled
RGD-containing peptide EMD 121974 (c-RGD-D-F-NMeV), a specific vß3
integrin antagonist, and its control peptide EMD 135981 (c-RAD-D-FV)
were a gift from Dr. F. Mitjans (Merck Farma y Quimica, S.A.,
Barcelona, Spain). Blocking integrin antibodies were obtained from
Chemicon (Temicula, CA) and included 1 MAB 1973, 2 MAB 1950, 3
MAB 1952, 6 MAB 1972, V MAB 1980, ß1 MAB 1965, ß3 MAB 1957,
ß4 MAB 2058, and vß3 MAB 1976 (LM609). vß5 clone P1F6 was
from Life Technologies (Gaithersburg, MD). 5 Integrin blocking
antibody Mab 16 was a kind gift from Dr. K. Yamada.
Cells and culture
Human umbilical vein endothelial cells (HUVECs) were obtained
from freshly delivered umbilical cords by treatment with 0.1%
collagenase (16)
. Cells were grown in RPMI 1640 medium
containing 20% defined and supplemented bovine calf serum (BCS)
(HyClone Laboratories, Inc., Logan, UT), 5 U/ml of heparin (Fisher
Scientific, Pittsburgh, PA), 200 µg/ml of endothelial cell growth
factor (ECGS) (Collaborative Research, Bedford, MA), 100 units/ml
penicillin, 100 µg/ml streptomycin, 50 µg/ml gentamicin, and 2.5
µg/ml amphotericin B (Life Technologies). Only early cell passages
(3
4
5)
were used.
Cell adhesion assays
Adhesion assays were performed in 96-well plates coated with
laminin-1 or synthetic laminin peptides as described (14)
.
Wells were coated overnight with either 5 µg of peptide or 0.5 µg
of laminin-1 in 100 µl of phosphate-buffered saline (PBS). Wells were
rinsed three times with PBS, blocked with 0.2% bovine serum albumin
(BSA) in PBS for 2 h, and rinsed three times with PBS. Uncoated
wells blocked with BSA served as controls. Confluent HUVECs were
detached with Versene (Life Technologies) and 35,000 endothelial cells
were plated per well in 100 µl of RPMI 1640 medium. Competition
experiments were performed using 0–50 32 µg of peptide/ml or as
otherwise specified. After a 1.5 h incubation at 37°C, unbound
cells were removed and the attached cells were fixed and stained with
20% methanol containing 0.2% crystal violet. After washing
extensively with distilled water, bound dye was solubilized with 2%
sodium dodecyl sulfate (SDS) and quantitated in an ELISA Emax plate
reader at 600 nm (Molecular Devices, Palo Alto, CA). All assays were
performed at least three times in triplicate.
Aortic ring sprouting assays
Aortas were harvested from 6-wk-old Sprague-Dawley rats and
cleaned of fatty tissue (17)
. The aortas were
cross-sectioned into thin rings with a scalpel. The rings were placed
on 150 µl of gelled Matrigel on 48-well dishes, overlaid with 50 µl
of Matrigel that was allowed to gel for 30 min, then incubated in the
presence 200 µg/ml of test peptide in 200 µl of human endothelial
serum-free medium (Life Technologies). On day 4, an additional 20 µg
of peptide in 100 µl of fresh serum-free medium was added, and the
assay was stopped on the fifth or sixth day after sprouts had
developed. Assays were repeated three times in quadruplicate and scored
by a blinded observer.
Cell membrane labeling and peptide affinity chromatography
HUVECs membrane proteins were biotin-labeled with EZ-Link
Sulfo-NHS-Biotin (Pierce, Rockford, IL). Confluent 150 mm dishes were
rinsed five times with cold PBS and incubated with 4 ml of 1 mg/ml of
Sulfo-NHS-Biotin reagent in PBS at 4°C for 15 min with shaking. Cells
were removed by scraping and 4 ml of cold Tris-buffered saline was
added to neutralize the biotin reagent. After a 5 min centrifugation at
50 g, the cell pellet was collected and incubated for 1 h on ice with 1 ml 0.5% Nonidet P-40 in PBS (NP/PBS) containing 1x
Complete protease inhibitors (Boehringer Mannheim Corporation,
Indianapolis, IN). The supernatant was collected after centrifugation
for 15 min at 11,000 g at 4°C.
Peptide affinity columns were prepared using Affi-gel 10 as specified
by the manufacturer (Bio-Rad Laboratories, Hercules, CA). Peptide C16
or C16S was coupled to 4 ml of Affi-gel beads using 1 mg/ml of resin in
0.1 M sodium carbonate buffer, pH 8.5. Remaining active groups were
blocked with 0.1 M diethanolamine. Biotin-labeled membrane proteins
equivalent to one confluent dish (
3.5–4.0x106 cells) were added to the column and
allowed to incubate for 30 min. The column was washed extensively with
50 ml of NP/PBS running buffer and eluted successively with 20 ml each
of 5 mM EDTA, 1 M NaCl, and 4 M urea; 0.5 ml fractions were collected
and aliquots (20 µl) were tested in 96-well plates for the presence
of biotin using a 1:1000 dilution of streptavidin-horseradish
peroxidase conjugate (Life Technologies) and developed with 1-Step
Turbo TMB-ELISA substrate (Pierce). Fractions containing biotin were
pooled and dialyzed against water; 50 µl aliquots were concentrated
fivefold in a Speed-Vac before SDS-PAGE electrophoresis. Gels were
transferred to nitrocellulose membranes and biotin-labeled proteins
were visualized by enhanced chemiluminescence (ECL) (Amersham Life
Science, Buckinghamshire, England) after incubating with
streptavidin-horseradish peroxidase.
Immunoprecipitation
Immunoprecipitation of proteins isolated from peptide affinity
chromatography columns were performed using integrin antibodies as
specified by the manufacturer. Protein G beads were incubated twice,
2 h each, with 0.5 ml of a 1:5 NP-PBS dilution of unlabeled HUVEC
protein extract prepared as described above. Beads were then incubated
with integrin antibodies for 1 h in NP-PBS containing 0.1% BSA
and overnight with 0.3 ml of column eluate containing 0.1% BSA and
0.5% Nonidet P-40. The beads were washed five times with RIPA buffer
(0.15 M NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, 50 mM
Tris, pH 8.0) and once in Tris buffer. Immunoprecipitated proteins were
run on SDS-PAGE, blotted onto nitrocellulose membranes, and visualized
by ECL as described above.
Tube assay and Western blotting
Confluent endothelial cells incubated for 1 h in serum-free
media (human endothelial growth medium, Life Technologies) were treated
with 20 µM of MAP kinase pathway inhibitors PD98059 (New England
BioLabs, Beverly, MA) or I0126 (Promega, Madison, WI) for an additional
hour. Tube assays were then performed as described previously
(14)
. Briefly, 24,000 cells/well were seeded onto 48-well
dishes coated with 200 µl of Matrigel in the presence or absence
(control) of 0.1 mg/ml of peptide overnight. Cells were fixed with
methanol and stained with 0.625 g/l each of azure A and methylene blue
(Diff-Quik solutions I and II, Baxter Scientific Products, Morton
Grove, IL).
Six-well dishes containing confluent HUVEC were incubated in serum-free media for 2 h. The cells were incubated for 15 min in the presence of 0.1 mg/ml of C16, C16S, a mixture of C16 and C16S (0.1 mg/ml each), 100 ng/ml of bFGF, and bFGF containing either 0.1 mg/ml of C16 or C16S. Media were removed and cells were solubilized in 150 µl of 50 mM HEPES containing 1% Triton X-100, 1% sodium deoxycholate, 0.1% sodium dodecyl sulfate, 10% glycerol, 1 mM sodium orthovanadate, 0.15 M NaCl, 1.5 mM MgCl2, 1 mM EGTA, 10 mM sodium pyrophosphate, 100 mM NaF, pH 7.5, and 1x Complete protease inhibitor (Boehringer Mannheim). Proteins were separated on 10% SDS-PAGE before Western blotting with monoclonal anti-phospho-p44/42 MAP kinase (New England Biolabs, Inc.) and polyclonal anti-ERK1 antibody (Santa Cruz Biotechnology, Santa Cruz, CA). The experiment was repeated more than three times.
CAM assay
The CAM assay was performed (14)
using 3-day-old
embryonated eggs (Truslow Farms, Charlestown, MD). On the same day, 4
ml of ovalbumin was removed from each egg and windows were opened. On
embryonal day 10, 5 µl of peptide dissolved in distilled water
(0.1–1 µg) was dried on 13 mm-diameter quartered plastic coverslips
(Thermanox, Nalge, NUNC International) and placed on the
chorioallantoic membrane. Three days later, the eggs were scored and
photographed. For competition experiments, 0.5 µg of C16 or 50 ng of
bFGF were mixed with either 0.5 or 1.0 µg of C16S or control peptides
in a total volume of 5 µl. Experiments were repeated three times
using at least 11 eggs each time.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. This suggested that the cells were attaching to
either laminin or some other component present on the matrix. To
investigate whether C16 could promote adhesion, attachment experiments
were performed with endothelial cells using peptides C16, C16S, C16Rev
(two scrambled C16 sequences), C15, and C57. The latter two peptides
had previously been shown not to affect endothelial cell tube formation
(14)
. As expected, the inactive peptides did not promote
cell attachment; the reverse sequence peptide C16Rev showed low binding
activity whereas C16 was active (Fig. 1A
). Unexpectedly, cells also bound to the scrambled C16S
peptide. Inhibition experiments performed with the same peptides to
determine their effect on adhesion to laminin-1 showed that C16 and
C16S both inhibited attachment to laminin-1, whereas the other peptides
where inactive (Fig. 1B
). These results suggest that C16 and
C16S promote cell attachment when coated directly on plates and
specifically inhibit the adhesion of cells to laminin-1. Dose-response
experiments in which C16 and C16S were used to inhibit endothelial cell
attachment to laminin-1 demonstrated that both peptides have similar
activity (data not shown).
|
C16 induces sprouting of aortic rings
Since it was an unexpected finding that C16S inhibited attachment
to laminin-1 and sustained cell adhesion, aortic ring assays were
performed to determine whether the peptide induced cell sprouting. As
expected (14)
, C16 induced endothelial cell sprouting from
aortic rings (Fig. 2C
). In contrast, C16S had no effect on the rings (Fig. 2B
). Neither of the control peptides, C15 (Fig. 2A
), C57 (not shown), nor media alone (not shown) promoted
sprouting at a concentration of 200 µg/ml. These results suggested
that even though C16S can affect cell attachment, it does not induce
cell sprouting in ring explants.
|
Identification of membrane proteins that bind to C16 and C16S
Peptide affinity chromatography of biotinylated HUVEC proteins was
used to identify the putative membrane receptor molecule(s) that binds
to angiogenic peptide C16. Biotinylated proteins eluted from a control
column or a C16 column were run on SDS-PAGE electrophoresis and
detected with streptavidin-HRP (Fig. 3
). Controls on the gels included the total biotinylated membrane protein
extract loaded into the columns (lane 1 of Fig. 3A
, B
, and
lane 5 of Fig. 3B
) and an aliquot from the Nonidet P-40/PBS
wash just before elution with 1 M NaCl began (lane 2 of Fig. 3A
, B
, and lane 6 of Fig. 3B
). After extensive washing, no
proteins were eluted from the control column with either 1 M NaCl or 4
M urea (Fig. 3A
, lanes 3 and 4, respectively). Under the
same conditions, C16 affinity chromatography showed three major bands
ranging from 110 to 150 kDa that eluted with 1 M NaCl (Fig. 3B
, lane 3), but not with 4 M urea (Fig. 3B
, lane
4). Elution of the C16 column with 5 mM EDTA (lane 7) and 1 M NaCl
(lane 8) demonstrated a similar band pattern (Fig. 3B
) that
was not observed in the control column.
|
Since many of the bound proteins from the C16 columns eluted with EDTA
and had a molecular mass (110–150 kDa) similar to integrins, we
used anti-integrin antibodies to identify the bands.
Immunoprecipitation of the EDTA eluate with antibodies to several
integrins known to bind to laminin revealed that four different
integrin chains were present, including ß1, ß3, 5, and v
(Fig. 4A
). The presence of the integrin dimer 5ß1 was indicated
since antibodies to either 5 (Fig. 4A
) or ß1 chains
immunoprecipitated a doublet. Another doublet with proteins of
different molecular size was observed with antibodies to either v,
ß3, or vß3 chains (Fig. 4A
). When an antibody to
vß5 dimer was used, these molecules were not detected (not shown),
suggesting the presence of the vß3 dimer but not that of vß5.
Other integrins present on the HUVEC cell surface but not in the C16
column eluate included 1, 2, (Fig. 4A
, bottom) 3,
6, and ß4 (not shown). These results strongly suggest that the
membrane molecules that bind to the laminin 1 peptide C16 are the
integrin dimers 5ß1 and vß3.
|
Since the scrambled C16S peptide could inhibit cell attachment to
laminin-1 (Fig. 1B
) but was not angiogenic in the aortic
ring assay (Fig. 2B
), we investigated whether C16S could
also bind to these integrin molecules. Such binding may explain the
ability of C16S to compete with C16 for its binding site. Similar
affinity chromatography experiments as with C16 were performed with
C16S (not shown), followed by immunoprecipitation with several integrin
antibodies. The results showed that integrins 5ß1 and vß3
(Fig. 4B
) can bind to C16S. Immunoprecipitation with
anti-integrin antibodies to other chains—ß4 (Fig. 4B
),
1, 2, 3, and 6 (not shown) or beads alone (Fig. 4B
)—did not show the presence of any other integrin
subunits. These data confirm that the scrambled C16 peptide binds to
the same integrins as the C16 peptide.
Integrin antibodies inhibit cell adhesion to C16 and C16S
We next determined the functional significance of the integrin
binding to C16 and C16S. Inhibition of cell adhesion to both C16- and
C16S-coated plates in the presence of various blocking integrin
antibodies confirmed that adhesion is mediated through integrins
vß3 and 5ß1 (Table 1
). The greatest inhibition of adhesion to C16 and C16S was observed with
vß3 blocking antibodies (adhesion of 44.1±5.7 and 25.4±3.6%,
respectively). Lesser inhibition to both peptides was observed with
5 and ß1 blocking antibodies, suggesting that this is a less
significant interaction. Finally, we observed inhibition with 1
blocking antibodies that was considerably lower than that observed with
the other blocking antibodies. Since integrin 1 did not bind to the
affinity column, it most likely is not important for C16 binding.
Integrin 6 blocking antibodies did not inhibit cell attachment to
C16, as expected, since this integrin does not bind to peptide C16.
These data demonstrate the functional importance of integrins 5ß1
and vß3 in endothelial cell adhesion to laminin peptide C16 and
its scrambled sequence.
|
MAP kinase signaling is not activated by C16
Since integrin vß3 is known to mediate its activity with
matrix proteins via RGD sequences and MAP kinase, we tested whether a
specific vß3 antagonist, cyclic RGD-containing peptide EMD 121974,
could block HUVEC adhesion to C16 and found that it could not (not
shown). We also investigated the effect of C16 and C16S on MAP kinase
activation in HUVEC by Western blotting analysis with an anti-ERK-1/2
monoclonal antibody. Our results showed no increase in ERK-1/2
phosphorylation above its negative control without peptide (Fig. 5
, top panel, lane 1) by either C16S (lane 2), C16 (lane 3), or a
C16/C16S mixture (lane 6). As previously reported, bFGF activated both
ERK-1 and 2 (lane 4); however, the presence of C16S (lane 5) or C16
alone (lane 7) did not inhibit bFGF-induced phosphorylation. As a
control, we found that cycled RGD containing peptide EMD 121974 did
inhibit bFGF-induced phosphorylation (lane 9) as compared to nonactive
control peptide EMD 135981 (lane 8). The bottom panel of Fig. 5
shows
even gel loading of total MAP kinase. In addition, activation of
ERK-1/2 human foreskin fibrobasts was not observed by C16 or C16S.
Furthermore, MAP kinase inhibitors PD98059 and UO126 did not block the
activity of C16 on tube formation (data not shown). Since C16 and C16S
do not contain an RGD sequence, it was not surprising that their
activity was not affected by RGD or MAP kinase, suggesting that this
part of the pathway is not involved or that signaling occurs through
another pathway.
|
C16- and bFGF-induced angiogenesis is inhibited by C16S in vivo and
in vitro
Lack of activity in the aortic ring assay by C16S and binding to
the same integrin molecules as C16 indicated that C16S might be
interfering with the proper binding of C16 to either one or both
integrin molecules. A competition experiment on the CAM assay of
10-day-old chicks was performed (Fig. 6
) to determine the ability of C16S to block the angiogenic activity of
C16. Both vehicle alone (dH2O) and control
peptide C15 (0.5 µg/egg) showed angiogenic activity in only 14% of
the eggs, whereas C16 was 80% angiogenic when the same amount of
peptide was used (Fig. 6A
, B
, panels 1 and 3, respectively).
In contrast, C16S did not show any angiogenic activity at either 0.5 or
1 µg of peptide (Fig. 6B
, panel 2). When equal amounts of
C16 and C16S were mixed together (0.5 µg each/egg), the angiogenic
response was lost (Fig. 6B
, panel 4), similar to C16S alone
(25%). In contrast, C15, which was not angiogenic in either the CAM or
the aortic ring assay, did not inhibit the angiogenic activity of C16
(77.8%) when mixed together, even when tested with double the amount
of C16 (Fig. 6A
). These results were corroborated in the rat
aortic ring assay (Fig. 6C
). Figure 6C
(panel 2)
shows aortic ring sprouting induced by 100 ng/ml of bFGF, whereas, the
presence of 0.1 mg/ml of C16S inhibits the sprouting effect of bFGF
(panel 3) and C16 (panel 4). Control vehicle did not promote sprouting
(panel 1). Along with the attachment competition data, these results
suggest that C16S inhibits the biological activity of C16 by binding to
the 5ß1 and vß3 integrins.
|
Previously it was demonstrated that the effect of growth factors on
angiogenesis depends on the signaling events that are transduced by
vß3 (18
, 19)
. Since C16S binds to vß3, we tested
whether the angiogenic activity of bFGF could be blocked by C16S in the
CAM assay (Fig. 6)
. The results indicate that indeed C16S can inhibit
the effect of bFGF in vivo and suggest that C16S is interfering with
the signaling pathways necessary for angiogenesis by binding to
integrin vß3.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Laminin is present in the endothelial
basement membrane, and we have recently identified 20 new angiogenic
sites in the laminin-1 molecule (13
, 14)
. Here we have
studied the most active 1 chain peptide, C16, and identified the
cell surface ligands on endothelial cells. C16 is also the most active
1 chain peptide with a number of other cell types, including B16F10
melanoma and HT 1080 fibrosarcoma cells. We confirmed its angiogenic
activity in the aortic ring sprouting assay and in the chick CAM assay.
These data suggest that C16 is a potent and important site on laminin.
Although the peptides are angiogenic in vivo and in vitro, they do not
affect cell migration or proliferation (unpublished observations).
However, it is not clear whether it is cryptic in vivo and becomes
exposed during development or basement membrane degradation.
When HUVEC biotin-labeled membrane proteins were passed over a C16 or
C16S affinity column, the bound proteins eluted with EDTA and were
similar in molecular size to integrins. The integrins vß3 and
5ß1 were identified based on immunoprecipitation from the EDTA
eluate with specific antibodies. The same integrin antibodies inhibited
adhesion of the cells to both peptides. The identification of integrins
as the receptors for C16 was not unexpected as several integrins are
known to function as laminin receptors, including 1ß1, 2ß1,
3ß1, 6ß1, 7ß1, 9ß1, 1ß8, and 6ß4
(20)
. Of the two integrins identified here as binding to
C16, only vß3 had previously been reported as a laminin receptor
(21)
. Binding of the 5ß1 to C16 and not to the entire
laminin molecule in vitro suggests that this is a cryptic site in the
intact molecule. Integrin vß3 appears promiscuous in its ability
to bind various molecules such as fibronectin, vitronectin, fibrinogen,
von Willebrand factor, and others via different sequences (22
, 23)
. Our results suggest that vß3 can bind to laminin via
the C16 sequence located on the first globular domain of the 1
chain.
The 5ß1 integrin has been previously shown to bind fibronectin,
tenascin, invasin, L1 cell adhesion molecule (L1-CAM), and several ADAM
family members, including echistatin, flavoridin, and metargidin,
(22
, 24
25
26
27)
. All of the 5ß1 ligands appear to
interact through the RGD sequence with the receptor. It has been shown
that echistatin, flavoridin, metargidin, and L1-CAM can bind to both
integrins 5ß1 and vß3 (22
, 27
, 28)
. Although our
results show that C16 (KAFDITYVRLKF) and C16S (DFKLFAVTIKYR) can also
bind to both of these integrins, neither peptide contains the RGD
sequence. In a previous study we had identified by systematic
truncation the active core cell attachment sequence of C16 for HT-1080
cells as YVRL (9)
. The finding of integrins as the
receptors for peptide C16 is consistent with the potent activity of the
peptide with many cell types and its being the most active 1 chain
peptide. We had previously observed that C16 could block adhesion not
only to laminin, but also to fibronectin and collagen I
(14)
. The ability to block adhesion to multiple integrin
binding substrates again supports the identification of integrins as
the functional C16 receptors. Our findings appear to be specific since
other laminin peptide affinity columns, such as AG73 (RKRLQVQLSIRT from
the G domain of the 1 chain), did not bind to integrins but rather
yielded syndecan-1 as its cell surface ligand (29)
.
Furthermore, affinity columns from the angiogenic IKVAV peptide from
the 1 chain and the anti-angiogenic YIGSR peptide from the ß1
chain have identified 67 and 110 kDa receptors, respectively (30
, 31)
. Control inactive peptides and blank columns have not
yielded any integrins. The finding that C16 is the most active peptide
on the 1 chain and that one of its receptors is the integrin
vß3 is significant since it is well known that this receptor plays
a major role in angiogenesis (23
, 32)
.
It has been reported that v integrin signaling during angiogenesis
is mediated through the MAP kinase pathway (19
, 32
, 33)
.
We investigated the role of C16 in the activation of MAP kinase in
endothelial cells and found that neither C16 nor C16S increase its
phosphorylation level, nor do they inhibit its activation mediated by
bFGF. To confirm these results, MAP kinase inhibitors PD98059 and UO126
did not inhibit the effect of C16 on tube formation. When C16 was
tested on human foreskin fibroblast, no significant activation of MAP
kinase was observed, suggesting that this part of the pathway is not
affected or that a different, yet unidentified pathway is being used.
This observation is consistent with the fact that C16 is not binding to
the integrins through an RGD sequence and that RGD-containing peptides
do not compete with C16 for integrin binding. Although we do not yet
know the signaling mechanism, preliminary results indicate that
serine/threonine kinases, including PKC, PKA, and PKG, are not
involved.
It was unexpected that the scrambled C16 peptide C16S would bind to the
5ß1 and vß3 integrins. Many scrambled peptides of other
active sites have been previously tested and none have been found to
have significant biological activity (9
, 10)
. The reverse
sequence of C16, C16Rev, was inactive, indicating that the integrin
binding is not due to the charge of the amino acids. It is well known
that vß3 is the most highly interactive of all the integrins
identified to date. This ‘promiscuity’ appears to include the
scrambled sequence of C16. Alternatively, it is highly likely that due
to the small size of the peptide, C16S can adopt a conformation in
solution closely resembling that of C16, which allows it to bind to the
integrin receptors. Our results from the adhesion studies demonstrate
that C16S can compete with C16 for its attachment site. C16S by itself
is unable to promote endothelial cell aortic ring sprouting and
angiogenesis in the CAM assay. However, C16S can compete with C16 in in
vivo angiogenesis studies. These data suggest that C16S can bind to the
receptor, but does not appear to transduce a signal. It has been
demonstrated that the angiogenic effect of bFGF depends on the
long-term activation of MAP kinase mediated through vß3
(19)
. Here we demonstrated that C16S inhibits the
angiogenic effect of bFGF, suggesting that the signaling events of
vß3 have been altered. This is further evidence that C16S
interacts with vß3.
Degradation of the extracellular matrix is one of the first steps
during tumor invasion, wound healing, and tissue remodeling. During
this process, laminin, which is not angiogenic when intact, becomes
cleaved, possibly allowing its active sequences (which might include
C16 and others) to become exposed, enabling them to induce an
angiogenic response. The 1 chain, and therefore the active sequence
within C16, are found in 10 of the 12 laminins known to date. This
strongly suggests that the 1 chain plays an important physiological
role since it is located in most tissues.
The existence of cryptic sites with biological activity within larger
molecules is not unusual. It was recently reported that different
regulatory anti-angiogenic peptides are latent within the extracellular
matrix proteins collagen IV (NC1 domains), collagen XVIII (endostatin),
plasminogen (angiostatin), and calreticulin (vasostatin) (2
, 3
, 34
, 35)
. It has also been shown that when anti-thrombin is
cleaved, it undergoes a conformational change that exposes an area of
the molecule with anti-angiogenic activity (36)
.
Taken together, these data suggest that angiogenesis is tightly regulated by several extracellular fine tuning molecules including growth factors, circulating molecules, and extracellular matrix proteins present within the adjacent microenvironment. In this study, we not only isolated the receptor for the most active of the laminin-1 chain peptides, but more important, we identified a sequence that can compete the binding of laminin to its receptors. The scrambled version of this peptide may serve as an anti-angiogenic agent for the treatment of angiogenesis-related diseases and cancer. The ability of C16S to interact with these two integrins without promoting angiogenesis makes this sequence a strong candidate as an angiogenic inhibitor.
Received for publication November 2, 2000.
Revision received February 27, 2001.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
chains: expression, developmental transitions, and chromosomal locations of 1–5, identification of heterotrimeric laminins 8–11, and cloning of a novel 3 isoform. J. Cell Biol. 137,685-7011 chain carboxyl-terminal globular domain by systematic screening of synthetic peptides. J. Biol. Chem. 270,20583-205901 chain by systematic peptide screening. J. Biol. Chem. 272,32198-322051 chain. J. Biol. Chem. 273,32491-324991 and ß1 chain peptides active for endothelial cell adhesion, tube formation, and aortic sprouting. FASEB J 13,53-621 chain. Circ. Res. 84,688-694vß3 requirement for sustained mitogen-activated protein kinase activity during angiogenesis. J. Cell Biol. 140,1255-1263IIbß3, vß3 and 5ß1 and in cell adhesion inhibition. Cell Adhes. Commun. 2,491-501[Medline]
vß3 for angiogenesis. Science 264,569-5711 integrin-dependent T lymphocyte adhesion to fibronectin through the binding of its fnIII 1–5 repeats to fibronectin. Eur. J. Immunol. 5,1435-1447
vß3 and 5ß1 integrins on different haemopoietic cells. J. Cell Sci. 112,579-587[Abstract]
vß3. J. Cell Biol. 132,475-485v integrins. Science 270,1500-1502v integrins during angiogenesis. Cancer J 6,S245-S249
This article has been cited by other articles:
|
|
 |
|
  C. Guruvayoorappan and G. Kuttan {beta}-Carotene Inhibits Tumor-Specific Angiogenesis by Altering the Cytokine Profile and Inhibits the Nuclear Translocation of Transcription Factors in B16F-10 Melanoma Cells Integr Cancer Ther, September 1, 2007; 6(3): 258 - 270. [Abstract] [PDF]
|
|
|
|
 |
|
 M. U. Naik and U. P. Naik Junctional adhesion molecule-A-induced endothelial cell migration on vitronectin is integrin {alpha}v{beta}3 specific J. Cell Sci., February 1, 2006; 119(3): 490 - 499. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G Eisenhofer, T-T Huynh, K Pacak, F M Brouwers, M M Walther, W M Linehan, P J Munson, M Mannelli, D S Goldstein, and A G Elkahloun Distinct gene expression profiles in norepinephrine- and epinephrine-producing hereditary and sporadic pheochromocytomas: activation of hypoxia-driven angiogenic pathways in von Hippel-Lindau syndrome Endocr. Relat. Cancer, December 1, 2004; 11(4): 897 - 911. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Dixelius, L. Jakobsson, E. Genersch, S. Bohman, P. Ekblom, and L. Claesson-Welsh Laminin-1 Promotes Angiogenesis in Synergy with Fibroblast Growth Factor by Distinct Regulation of the Gene and Protein Expression Profile in Endothelial Cells J. Biol. Chem., May 28, 2004; 279(22): 23766 - 23772. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. L. Ponce, S. Hibino, A. M. Lebioda, M. Mochizuki, M. Nomizu, and H. K. Kleinman Identification of a Potent Peptide Antagonist to an Active Laminin-1 Sequence That Blocks Angiogenesis and Tumor Growth Cancer Res., August 15, 2003; 63(16): 5060 - 5064. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Okazaki, N. Suzuki, N. Nishi, A. Utani, H. Matsuura, H. Shinkai, H. Yamashita, Y. Kitagawa, and M. Nomizu Identification of Biologically Active Sequences in the Laminin alpha 4 Chain G Domain J. Biol. Chem., September 27, 2002; 277(40): 37070 - 37078. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. C. Cid, J. Hernandez-Rodriguez, M.-J. Esteban, M. Cebrian, Y. S. Gho, C. Font, A. Urbano-Marquez, J. M. Grau, and H. K. Kleinman Tissue and Serum Angiogenic Activity Is Associated With Low Prevalence of Ischemic Complications in Patients With Giant-Cell Arteritis Circulation, September 24, 2002; 106(13): 1664 - 1671. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. G. Stupack and D. A. Cheresh ECM Remodeling Regulates Angiogenesis: Endothelial Integrins Look for New Ligands Sci. Signal., February 12, 2002; 2002(119): pe7 - pe7. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||
