Content uploaded by Keith A Houck
Author content
All content in this area was uploaded by Keith A Houck
Content may be subject to copyright.
Available via license: CC BY 4.0
Content may be subject to copyright.
THE
JOURNAL
OF
BIOL~CICAL
CHEMISTRY
0
1994 by The American Society
for
Biochemistry and Molecular Biology, Inc. Vol. 269,
No.
41, Issue
of
October 14,
pp.
2564625654, 1994
Printed
in
U.S.A.
Placenta
Growth
Factor
POTENTIATION OF VASCULAR ENDOTHELIAL GROWTH FACTOR BIOACTIVITY,
IN VITRO
AND
IN VIVO,
AND
HIGH
AFFINITY BINDING TO
Flt-1
BUT NOT
TO
Flk-l/KDR*
(Received for publication, April 12, 1994, and in revised form, July 18, 1994)
John
E.
Park, Helen
H.
Chen, Jane Winer, Keith
A..
HouckS, and Napoleone FerraraO
From Genentech, Inc., South San Francisco, California
94080
The recently identified placenta growth factor (PlGF)
is a member
of
the vascular endothelial growth factor
(VEGF) family of growth factors. PlGF displays a
53%
identity with the platelet-derived growth factor-like re-
gion of VEGF.
By
alternative splicing of
RNA,
two PlGF
isoforms are generated PlGF,,, (PlGF-1) and PlGF,,,
(PlGF-2). Relative to PIGF,,,, PlGF,,, has a 21-amino acid
insertion enriched in basic amino acids. Little
is
known
at the present time about the significance and function
of these proteins.
To
assess their potential role, we
cloned the cDNAs coding for both isoforms, expressed
them in mammalian cells, and purified to apparent ho-
mogeneity the recombinant proteins. Like VEGF, the
PlGF isoforms are homodimeric glycoproteins. PlGF,,,
is a non-heparin binding protein, whereas PIGFIG,
strongly binds to heparin. We examined the ability of
PlGF to bind to soluble VEGF receptors, Flt-1 and
Flk-l/KDR, and characterized the binding of PlGF to en-
dothelial cells. While the PlGF proteins bound with high
affinity to Flt-1, they failed to bind to Flk-l/KDR. Bind-
ing of l2%P1GF to human endothelial cells revealed two
classes of sites, having high and low affinity. The high
affinity site is consistent with Flt-1; the identity of the
low affinity site remains to be determined. Purified
PlGF isoforms had little or no direct mitogenic or per-
meability-enhancing activity. However, they were able
to significantly potentiate the action of low concentra-
tions of VEGF
in vitro
and, more strikingly,
in vivo.
The vascular endothelium is one of the most versatile sys-
tems
in
the body, serving
a
variety of essential exchange and
regulatory functions. A fundamental property of vascular en-
dothelial cells
is
the ability to proliferate and form
a
network of
capillaries. This process, known as angiogenesis,
is
prominent
during embryonic development (1,2). In a normal adult, angio-
genesis occurs only following injury or, in
a
cyclical fashion, in
the endometrium and
in
the ovary (3). Angiogenesis, however,
is known to play
a
critical role
in
the pathogenesis of a variety
of disorders, most notably growth and metastasis of solid
tumors (4).
Several potential positive regulators of angiogenesis have
been described including acidic and basic fibroblast growth
factors, transforming growth factors
a
and
p,
tumor necrosis
*
The costs of publication
of
this
article
were
defrayed
in
part
by
the
payment
of
page charges.
This article must therefore
be
hereby marked
"a'aduertisement"
in
accordance
with
18
U.S.C. Section
1734
solely
to
indicate this
fact.
27717.
$
Present address: Sphinx Pharmaceuticals Corp., Durham,
NC
Dept.
of
Cardiovascular Research, Genentech Inc.,
460
Point San
Bruno
8
To whom correspondence and reprint requests should
be
addressed:
Blvd., South San Francisco,
CA
94080. Tel.: 415-225-2968; Fax: 415-
225-6327.
factor
a,
and angiogenin (5-14). The recently described vascu-
lar endothelial growth factor (VEGF)'
has
several attractive
features
as
a regulator of normal and pathological angiogenesis
(15).
VEGF is an endothelial cell-specific mitogen
in vitro
and an
angiogenic inducer
in uiuo.
In addition, VEGF is able to induce
vascular leakage in the Miles assay (7,8). On the basis of
this
activity, VEGF
has
been also implicated
as
a mediator of the
abnormal permeability properties of tumor blood vessels
(7,
8).
By alternative splicing of mRNA, VEGF may exist in four
different homodimeric molecular species, each monomer hav-
ing, respectively, 121,
165,
189, or 206 amino acids (VEGF,,,,
VEGF,,,,VEGF,,,, VEGF,,,). VEGF,,, and VEGF,,, are diffus-
ible proteins, whereas VEGF,,, and VEGF,,, are mostly bound
to heparin-containing proteoglycans in the extracellular matrix
(16, 17). Recent studies have suggested
that
VEGF
is
an im-
portant regulator of both developmental and ovarian angiogen-
esis (18-21). Furthermore, inhibition of VEGF action results in
marked suppression
of
the growth of several human tumor cell
lines
in
nude mice (22).
Two
tyrosine kinases have been identified
as
putative VEGF
receptors (23,241. The fms-like tyrosine kinase (Flt-1) (25) and
the kinase domain region
(KDR)
(26) proteins have been shown
to bind VEGF with high affinity. The murine homologue of
KDR, known
as
fetal liver kinase-1 (Flk-1) (271, has been also
shown to bind VEGF (28, 29). Flt-1 and Flk-l/KDR receptors
have
a
single signal sequence, one transmembrane domain,
seven immunoglobulin-like domains in their extracellular do-
main, and
a
consensus tyrosine kinase sequence domain. In
addition,
a
cDNAencoding
a
truncated form of Flt-1 lacking the
seventh immunoglobulin-like domain, the cytoplasmic domain,
and transmembrane sequence has been identified in human
umbilical vein endothelial cells (30).
Recently,
a
cDNA encoding
a
protein having a 53% identity
with the platelet-derived growth factor-like region of VEGF has
been isolated from a human placental cDNA library
(31).
The
encoded protein, named placenta growth factor (PlGF), was
expected to have 149 amino acids. Subsequently,
a
longer PlGF
cDNA was identified (32,331. Compared to the originally iden-
tified PlGF species, the long PlGF protein was expected to have
a
21-amino acid insertion highly enriched in basic residues.
The two isoforms, resulting from alternative splicing of RNA,
were named, respectively, PlGF-1 and PlGF-2 (32, 33). In con-
trast
to the widespread distribution of the VEGF mRNA
(151,
expression of PlGF mRNA appears to be restricted
to
placenta,
trophoblastic tumors, and cultured human endothelial cells
(31-33). Based on the homology with VEGF, PlGF was pro-
The abbreviations
used
are: VEGF, vascular endothelial growth fac-
tor;
Flt-1, fms-like tyrosine
kinase;
Flk-1, fetal liver
kinase
1;
KDR,
kinase
domain
region;
PlGF,
placenta
growth
factor;
PAGE,
polyacryl-
nant human;
b,
basic;
PBS, phosphate-buffered
saline;
ACCE, adrenal
amide
gel
electrophoresis; FGF, fibroblast growth factor;
rh,
recombi-
cortex-derived capillary
endothelial;
HUVE,
human
umbilical
vein
endothelial.
25646
PlGF Binds to Flt-1 and Potentiates VEGF Activity
25647
.~
Flt-1
IgG:
GAATTC C ATG GTC AGC TAC.
.
. . . .
.
. .
. . . . . .
.
EcoRI
MVSY4
BstBI
4
TCT AAT
TTC
Ge GAC
AAA
ACT.
. .
S
N
E
E
D
K
T761
Original
receptor:
MVSY4
E758
~
KDR
IgG:
ClaI
sper
ATCG ATG GAG AGC AAG..
.
.TCA CTA GTT.
. .
.TTG GAT CCG TTC G.
BamHI
BStBI
4
”-
“ESK4
S
L
v416
L
2
2
E
ID
K
T770
Original
receptor:
s
K2
s
L
v414
L
E
762
FIG.
1.
Partial amino acid and
cDNA
sequence
of
Flt-1
IgG
and
KDR
IgG. Arrows
point to the junction between the extracellular domain
of each receptor and the IgG. Below each receptor-IgG, sequence of the original receptor is shown for comparison. Underlined amino acids or
nucleotides indicate modifications of the original sequences.
IgG
residues are enclosed in boxes.
posed to be an angiogenic factor, and initial evidence suggested
that
conditioned media of transfected cells expressing PlGF
were weakly mitogenic to endothelial cells derived from bovine
pulmonary artery
or
aorta
(31,331.
To
date, PlGF has not been
reported to have been purified to homogeneity.
In
the present study, we purified
to
apparent homogeneity
the two PlGF isoforms and tested them
in
vitro
for mitogenic
activity
on
bovine and human vascular endothelial cells and in
an
in
vivo
system, the Miles vascular permeability assay. Fur-
thermore, we examined their ability
to
bind to soluble VEGF
receptors,
Flt-1
and Flk-l/KDR, and characterized the binding
of
lZ5I-PlGF to endothelial cells. PlGF was able to bind with
high affinity to Flt-1 but not to Flk-liKDR,
a
receptor thought
to be critically involved in mediating WGF biological actions.
PlGF had little
or
no direct mitogenic or permeability-enhanc-
ing activity. Intriguingly, however, it was able to significantly
potentiate the effects of low concentrations of WGF, both
in
vitro
and
in
vivo.
EXPERIMENTAL PROCEDURES
Materials-Tissue culture reagents and media were obtained from
Life Technologies, Inc. through the Genentech media facility. Restric-
tion enzymes were from New England Biolabs except for Pfu polymer-
ase which was from Stratagene. Q-Sepharose Resource fast protein
liquid chromatography column
(6
ml), protein A-Sepharose, wheat germ
agglutinin agarose, heparin-Sepharose, and S-Sepharose fast flow res-
ins were from Pharmacia Biotech. The C4 reversed phase high perform-
ance liquid chromatography column (4.6
x
100 mm) was purchased from
SynChrom (Lafayette, IN). The preparative TSK 3000 GS column (21
x
300 mm) was from Hewlett-Packard. Trifluoroacetic acid was from
Pierce. Acetonitrile was purchased from Fisher Scientific. Tissue cul-
ture plates were from Costar except for large scale Nunc plates (24.5
x
24.5 cm), which were from Applied Scientific. Molecular weight stand-
ards for gel chromatography and prestained low molecular weight
markers for SDSPAGE gel were purchased from Bio-Rad. Recombinant
human (rh) basic FGF (bFGF) was from R
&
D Systems (Minneapolis,
MN). rhVEGF,,, was purified from transfected Chinese hamster ovary
cells
as
described (34). Iodination of VEGFle, and PlGF proteins was
performed by the indirect IODOGEN method
(35).
Specific activities
ranged from
18,000
to 32,000 cpdng for lZ5I-P1GF and 48,000 to 97,000
cpdng for lZ5I-VEGF. Rabbit polyclonal antibodies to Flk-l/KDR and
Flt-1
were produced by immunizing New Zealand White rabbits
monthly with purified Flk-1 IgG or Flt-1
IgG.
GCCGGTCA TGAGGCTGT 3’ and
5’
GAATTCTCTAGAGGTTACCTC-
Cloning
of
PLGF-Oligonucleotide primers
(5’
GAATTCTCTAGAT-
CGGGGAACAGCATC 3’) were designed to amplify
a
full-length PlGF
cDNA clone
(31).
The polymerase chain reaction was performed using
human placenta cDNA as the template. The polymerase chain reaction
products were subjected
to
polyacrylamide gel electrophoresis.
Ethidium bromide staining of the gel revealed two reaction products
at
475 and 538 base pairs, respectively. The lower molecular weight prod-
uct is consistent with PlGF-1, the higher with the longer isoform,
PlGF-2 (31-33). The bands were cut, electroeluted, and subcloned into
the XbaI site
of
pHEB023 for expression in CEN4 cells. Such cells are
a derivative of the human embryonic kidney 293 cell line that stably
expresses the Epstein-Barr virus nuclear antigen-1, required for episo-
mal replication of pHEB023 vector (36). Transfections and selection of
transformants were performed
as
described previously
(16,
17). The
authenticity of all clones was verified by DNA sequencing.
Construction of VEGF Receptor-IgG Chimeras-The extracellular do-
mains of Flt-l and KDR were cloned by polymerase chain reaction using
Pfu polymerase. Human placental cDNA served
as
the template. Prim-
ers encompassed the entire extracellular domains, including signal pep-
tides (25, 26). The cDNAs for each receptor extracellular domain were
cloned in two pieces to facilitate sequencing.
Two
sets of primers (shown
below) were used, and the resulting bands
(-1
kilobase) were digested
with the appropriate enzymes and subcloned into pBluescript I1
or
pSL301. The cDNAs produced encoded the
first
758 amino acids
of
Flt-1
and the first
762
amino acids of KDR. Since the
5’
end of the published
KDR sequence
is
incomplete (26), the initiator methionine and gluta-
mate were added to enable secretion of the fusion protein, resulting in
a 764-amino acid extracellular domain. The KDR primers add a silent
internal
SpeI
site for cloning purposes (Fig.
1).
Full-length KDR extra-
cellular domain cDNA was created by ligating the two parts at the novel
SpeI site, while full-length Flt-1 extracellular domain cDNA was cre-
ated by ligating the two
Flt-1
polymerase chain reaction clones
at
a
unique natural MunI site.
Flt-1 Set
1:
5‘
TCTAGAGAATTCCATGGTCAGCTACTGGGACACC
3’
5’
CCAGGTCATTTGAACTCTCGTGTTC
3’
Flt-1 Set
2
5‘
TACTTAGAGGCCATACTCTTGTCCT
3’
5‘
GGATCCTTCGAAATTAGACTTGTCCGAGGTTC
3’
KDR
Set
1
5’
GAATTCATCGATGGAGAGCAAGGTGCTGCTGGCCGTC
3’
5’
ACACAACTAGTGAGACCACATGGCTCTGCTTCTC
3‘
KDR
Set
2
5: GGTCTCACTAGTTGTGTATGTCCCACCCCAGATT
3‘
5‘
GAATTCGGATCCAAGTTCGTCTTTTCCGGGCA
3‘
These primers changed amino acid 757
to
phenylalanine and introduced
a
BstBI site
at
the 3’ end of the
Flt-1
extracellular domain (Fig.
1).
The
KDR primers added
a
BamHI site to the
3’
end of the KDR extracellular
domain. A BstBI mutation which eliminated any linker sequences was
introduced
at
the
5’
end of CH,CH,, an IgGyl heavy chain cDNA clone
(37). Flt-1 extracellular domain sequences were fused to the coding
sequences for amino acids 216443
of
this IgGyl heavy chain clone via
the unique BstBI site
at
the 3’ end of the extracellular domain coding
region while KDR extracellular domain sequences were fused
to
the
upstream BamHI site (see Fig. 1). Both constructs were then subcloned
into pHEB023 for expression in CEN4 cells (17). The authenticity of all
clones was verified by DNA sequencing. The strategies for construction
KDR IgG.
and cloning of Flk-1
IgG
were essentially similar to those applied to
Purification
of
VEGF Receptor-ZgG Chimeras-Conditioned media of
transfected CEN4 cells expressing the three chimeric receptors were
concentrated
5-
to 10-fold by ultrafiltration and then applied onto pro-
tein A-Sepharose columns
(10
ml) that had been pre-equilibrated in
PBS. The flow
rate
was 2 mumin. Absorbance was monitored at 280 nm.
The columns were washed with PBS until the absorbance at 280 nm
became negligible and then eluted with 100 mM citric acid, pH
3.0.
Sufficient 2
M
Tris
base was added
to
tubes to immediately neutralize
the acid. A silver-stained SDSPAGE gel revealed the presence of a
single major band
at
>300 kDa in nonreducing and
-180
kDa in reduc-
ing conditions. The authenticity of the proteins was verified by NH,-
terminal amino acid sequence.
Purification
of
PZGF Isoforms-Serum-free conditioned media from
transfected CEN4 cells expressing each PlGF molecular species
(-2.5
liters) were concentrated
5-
to 10-fold by ultrafiltration using Amicon
stir cells
CYM
10
membrane). For purification of PlGF,,JPlGF-2 (pre-
dictably
a
basic protein), the concentrated conditioned medium was
extensively dialyzed against 25 mM sodium phosphate, pH
6.0,
and then
applied onto
a
cation exchange S-Sepharose fast flow resin packed in a
25648
PlGF
Binds to Flt-1 and Potentiates
VEGF
Activity
glass column (Omni,
10
x
100 mm) pre-equilibrated with the same
buffer. The flow rate was 2 mumin. After loading, the column was
washed with 22 ml of 25
m~
sodium phosphate, pH
6.0,
containing 0.4
M
NaC1. This wash resulted in a major peak of absorbance at 280 nm
that contained less than 10% of the activity able to compete with lZ5I-
VEGF,,, for binding to Flt-IgG. Such radioreceptor assay was used at
all steps to monitor the purification. Elution of bound PlGF was with a
linear gradient
(0.4-1.0
M)
of NaCl over 30 min. Fractions containing
the highest competing activity
(0.5-0.7
M
NaCl) were pooled, diluted
6-fold in water containing 0.1% trifluoroacetic acid, and applied by
multiple injections into a C4 reversed phase column that had been
pre-equilibrated with 20% acetonitrile/
0.1%
trifluoroacetic acid. The
flow rate was
0.6
mumin. The column was washed with
5
ml of equil-
ibration buffer and was then eluted with a linear gradient of acetonitrile
(2045%) in 95 min. The activity eluted as a broad and asymmetric peak
of absorbance at 210 nm between 30 and 32% acetonitrile.
A
silver-
stained SDSffAGE gel of the most active fractions revealed the pres-
ence of a single band at -32 kDa in reducing conditions. Such fractions
were pooled and subjected
to
microsequencing. The yield in purified
protein was approximately
0.5
mgfliter, as estimated by amino acid
analysis. The recovery, as assessed by radioreceptor assay, was -40%.
For purification of PlGF,,,/PlGF-l, concentrated conditioned medium
was dialyzed against 20
m~
Tris,
pH 8.0, and then applied onto a
Q-Sepharose Resource anion exchange column. The flow rate was 2
mumin. The column was eluted with a linear gradient of NaCl(0-0.5
M)
in 30 min. The activity capable of competing with lZ5I-VEGF for binding
to
Flt-IgG was eluted in the presence of -0.2
M
NaC1. The most active
fractions were pooled and then applied onto a TSK 3000 GS column
equilibrated with
100
mM KH,PO,, pH 6.8. The flow rate was 3 mumin.
The activity eluted with an apparent molecular weight of
-60,000.
Fractions containing such activity were further purified by reverse
phase chromatography, exactly as described for PlGF-2. Purified pro-
tein was quantified by amino acid analysis. Protein yield and recovery
were very similar to those described for PlGF-2.
Binding Assays with Soluble Receptors-Ninety-six-well breakaway
enzyme-linked immunosorbent assay plates (Nunc, Kampstrup, Den-
mark) were coated overnight at 4 "C with
2
pg/ml affinity-purified goat
anti-human Fc IgG (Organon-Teknika) in 50 mM Na,CO,, pH
9.6.
Plates
were blocked for
1
h
with 10% fetal bovine serum in PBS (buffer
B).
Blocking buffer was then removed.
To
each well, IgG fusion protein
(-1
ng), "'1-VEGF or 1251-P1GF, and cold competitor were added
to
a final
volume of 100
1.11
in buffer B. Unless noted otherwise, 12,000 cpm
of
radioligand were added to the assay. When noted, binding experiments
contained
1-10
pg/ml heparin. Binding was carried out at room tem-
perature for
3.5
h followed by 6 washes with buffer B. Binding was
determined by counting individual wells in a
y
counter. Data were
analyzed by a 4-parameter nonlinear curve-fitting program developed
at Genentech.
Endothelial Cell Culture and Mitogenic Assays-Bovine adrenal cor-
tex-derived capillary endothelial (ACCE) cells
or
human umbilical vein
endothelial (HUVE) cells were used for binding and mitogenic assays.
Stock plates of ACCE cells were maintained in the presence of low
glucose Dulbecco's modified Eagle's medium supplemented with 10%
calf serum, 2 mM glutamine (growth medium) plus bFGF at a final
concentration of 2 ng/ml (6, 17). For proliferation assays, ACCE cells
were seeded at 7,00O/well in 6-multiwell plates in the presence of
growth medium. HUVE cells were maintained in endothelial growth
medium (Clonetics) supplemented with
5%
fetal bovine serum and bo-
vine pituitary extract, according to the instruction of the manufacturer.
For proliferation assays,
HUVE
cells were seeded at 12,00O/well in
6-multiwell plates in the presence
of
endothelial growth medium con-
taining 2% serum, without pituitary extract. Heparin
(1
or 10 pg/ml)
was added when noted. After
5-8
days, cells were dissociated by trypsin,
and cell numbers were determined with a Coulter counter.
Binding to Endothelial Cells-ACCE or
HUVE
cells were grown
to
confluence in 6-well plates in the appropriate growth media. The media
were removed, and binding was carried out in low glucose Dulbecco's
modified Eagle's medium containing 10% fetal bovine serum, 0.2% gela-
tin, 10 &ml heparin, and 10 mM HEPES, pH 7.4 (binding medium) on
ice. One ml of binding medium containing
50,000
cpm of lZ5I-VEGF (or
'zsI-PIGF) and various competitive agents were added
to
each well.
When noted, amounts of radioligand were varied. Monolayers were
washed three times with binding medium and then extracted with
1
ml
of
1
N
NaOH
to
recover bound lZ5I-VEGF (or '261-P1GF). The entire
NaOH extract was counted in a
y
counter. Data were analyzed by the
method of Scatchard (38).
Cross-linking ofLigands-HUVE cells were grown
to
confluence in
10-cm dishes. All subsequent steps were performed at 4 "C
or
on ice.
Media were replaced with the binding medium (detailed above) contain-
ing 1261-VEGF (110
PM)
or '251-P1GF
(230
PM)
?
2 pgiml cold VEGF, as
indicated, for
2
h. Cells were washed twice with cold PBS and incubated
for 20 min at
4
"C with
0.5
mM BS3 cross-linking agent (Pierce) dissolved
in PBS. The reaction was quenched with one-tenth volume of 200 mM
glycine,
10
mM Tris-HC1, pH
8.
The dishes were washed twice with cold
PBS and extracted with 1.25 ml of lysis buffer containing
1%
Triton
X-100,
137 mM NaCl, 10% glycerol, 2
m~
EDTA,
1
mM vanadate,
1
m~
phenylmethylsulfonyl fluoride,
0.15
unit/ml aprotinin, 20
p~
leupeptin,
20 mM
Tris
HC1, pH 8.0 (39). The extracts were spun to remove debris,
and the supernatants were saved. The supernatants were split into
equal volumes and then immunoprecipitated with rabbit antisera to
Flt-1, Flk-l/KDR, or with control antisera at the final dilution of
150
for
1
h. Protein A-agarose was added and mixed for an additional hour at
4 "C. The beads were pelleted and washed twice with lysis buffer before
addition of double-strength Laemmli sample buffer to elute proteins
from the beads. The eluates were boiled for
1
min and electrophoresed
on 620% SDS-polyacrylamide gels. The gels were dried and exposed
to
a phosphor-imaging plate for
-36
h. The plate was read on a BAS-2000
image analyzer (Fuji, Japan).
nrosine Phosphorylation Assay-ACCE cells were grown to conflu-
ence in 60-mm dishes without added growth factors. Confluent cultures
were preincubated for 24 h in the presence of low glucose Dulbecco's
modified Eagle's medium supplemented with
0.5%
calf serum (deprived
medium). Media were then removed and cells were incubated for
5
min
at 37 "C in the presence of deprived medium containing 50 ng/ml
VEGF,,,
or
varying amounts of PlGF-2 (PIGF,,,). At the end of the
incubation, media were removed and cells were washed with cold PBS
containing
1
mM sodium vanadate and then extracted with 0.25 ml
of
lysis buffer (see above). This and all subsequent steps were performed
at 4 "C. The extracts were centrifuged briefly in a Microfuge
to
pellet
debris. The supernatants were mixed for
1
h with
50
1.11 of wheat germ
agglutinin agarose beads to collect membrane proteins. The beads were
pelleted and washed once with lysis buffer. Double-strength Laemmli
sample buffer was added to elute proteins, boiled for
1
min, and subjected
to electrophoresis on 4-20% SDS-polyacrylamide gels. The gels were
electroblotted to polyvinylidene difluoride membranes. Following incu-
bation with 4G10 anti-phosphotyrosine monoclonal antibody (UBI, Lake
Placid,
NY),
immunoreactive bands were visualized with an ABC kit,
according
to
the instructions of the manufacturer (Vector Laboratories).
Miles Vascular Permeability Assay-Induction
of
vascular permeabil-
ity was determined by the Miles assay essentially as described previ-
ously (8, 40). One ml of 0.5% (w/v) Evans blue was injected intracardi-
ally into anesthetized guinea pigs. One hour later, PBS alone, different
concentrations of VEGF or PlGF, individually or in combination, were
injected intradermally in 200-1.11 aliquots in the dorsal area. Leakage of
protein-bound dye was detected by
a
blue spot surrounding the injection
site.
RESULTS
Biochemical Characterization
of
PlGF-The two PlGF pro-
teins were purified to apparent homogeneity. Anion and cation
exchange chromatographies were used
as
initial purification
steps for PIGF,,l/PIGF-l and PlGF,,@GF-2, respectively. The
final purification step
for
both isoforms was reversed phase
chromatography on
a
C4
column. Analysis of the purified ma-
terials by silver-stained SDS-PAGE gel in reducing conditions
revealed major bands (Fig. 2)
at
-32 kDa (PlGF,,,, lane
4)
and
-28 kDa (PlGF,,,, lane
61,
respectively. In nonreducing condi-
tions, bands of approximately twice that size were evidenced
(PIGF,,,
(lune
1
)
and PlGF,,, (lane
3)).
This
is
consistent with
a
homodimeric structure. Microsequencing
of
the purified ma-
terial revealed in both cases
a
single NH,-terminal amino acid
sequence: LPAW. Therefore, the PlGF proteins are generated
following cleavage of an 18-amino acid signal sequence (31,321.
The mature PlGF monomers are expected
to
have, respectively,
131
and 152 amino acids. By analogy with the WGF polypep-
tides, they may be called PlGF,,, and PlGF,,,. Both PlGF
iso-
forms reveal higher apparent molecular weight than WGFlG5
(Fig. 2), in spite of the fact that their predicted sequence
is
shorter.
A
higher level
of
glycosylation probably accounts for
such differences. This
is
consistent with the presence
of
two
putative glycosylation sites in each PlGF monomer (31-33) uer-
sus
only one in VEGF
(6).
Furthermore, the finding
that
both
PlGF Binds to Flt-1 and Potentiates VEGF Activity
25649
PlGF isoforms are eluted
as
broad and asymmetric peaks by
reversed phase HPLC (data not shown)
is
consistent with heav-
ily glycosylated proteins. The PlGF proteins were then com-
pared for their ability to bind to heparin-Sepharose
or
S-Sepha-
rose.
As
shown in Fig. 3, P1GFl3, did not bind appreciably to
S-Sepharose
or
heparin-Sepharose. In contrast, PIGF,,, bound
strongly to both matrices and was eluted in the presence of
20.5
M
and
0.9
M
NaCl, respectively. The differential chromato-
graphic behavior of PlGF,,, and PlGF,,, is clearly reminiscent
of VEGF,,, and VEGF,,, (16) and is consistent with the primary
amino acid sequence of the PlGF proteins
(31-33).
PlGF Binds to Flt-1 IgG but Not to
KDR
ZgG-Flt-1 IgG
or
KDR IgG were
first
tested in competition binding assays using
I2,I-VEGF as the radioligand (Fig.
4,
A
and B). The IC,, for
VEGF,,, binding to
Flt-1
IgG was 22
-c
3
PM,
while that
for
VEGF,,, binding to KDR IgG was 125
2
20
PM.
In both cases,
123456
106
-
80
-
495-
M
I
32
5-
27
5-
18
5-
FIG.
2.
Silvepstained SDS-polyacrylamide gel
of
purified
PlGF
isoforms and
VEGF,,.
PIGF,,,, VEGF,,,, or PlGFIs2 (approximately
200
ng) were electrophoresed on a 10-20% polyacrylamide gradient gel
at
40
d.
Lunes
13
were under nonreducing conditions while
lanes
4-6
were under reducing conditions.
Lanes
1
and
4
contain PIGF,,,
lunes
2
and
5,
VEGF,.,, and
lanes
3
and
6,
PIGF,,,. Molecular sizes of
I
111"
prestained staiJards are in kilodaltons:
of
PlGF,,, (A
and
B)
or
PlGF,,,
(C
and
FIG.
3.
Chromatographic behavior
D)
on S-Sepharose
(A
and
C)
or
hep-
arin-Sepharose
(B
andD) columns.
In
A
and
C,
concentrated conditioned me-
dium from transfected cells expressing ei-
ther PlGF
isoform
was equilibrated in 25
mM sodium phosphate, pH
6.0.
The mate-
rial was loaded onto S-Sepharose columns
(1
ml), washed with starting buffer, and
eluted stepwise with increasing concen-
trations of NaCI, as indicated. In
B
and
D,
concentrated conditioned medium was
equilibrated in 10 mM
Tris
HCI, pH
7.2,
containing
50
mM NaCI. The material was
loaded onto heparin-Sepharose columns
(1
ml), washed with starting buffer, and
eluted with increasing concentrations
of
NaCI. PlGF was detected by inhibition of
'2sI-VEGF binding to Flt-1 IgG.
100
80
60
0
40
e
5
C
20
LL
Eo
?
100
s
80
60
'
40
(u
7
.-
.-
c
C
.-
20
0
*251-VEGF,,, bound to the soluble receptor with an affinity very
similar to that of the full-length receptor expressed in trans-
fected cells (23, 24, 28). Therefore, these soluble receptor-IgGs
provide suitable in vitro model systems to study each VEGF
receptor independently of the other.
In preliminary experiments, we determined that conditioned
medium derived from transfected CEN4 cells expressing either
PlGF,,,
or
PIGF,,, was able to compete for 1251-VEGF,,, binding
to
Flt-1
IgG. Such activity provided
a
convenient assay for the
purification of both PlGF species. Like the conditioned me-
dium, highly purified PlGF was able to compete with lZ5I-
VEGF,,, for binding to
Flt-1
IgG (Fig.
5A).
In contrast, PlGF
(up to
48
nM) was unable to displace 12,1-VEGF bound to KDR
IgG (Fig.
5A)
or
Flk-1 IgG (not shown). Both '251-P1GF,,2 and
1251-PIGF13, bound to Flt-1 IgG with high affinity (Fig.
5,
B-D).
The
ED,, values were 250
?
35
PM
and 186
2
40
PM,
respectively.
However, lZ5I-P1GF failed
to
bind to KDR IgG (Fig.
5E).
In
panels D and E, binding experiments were performed in the
presence of
1
pg/ml heparin, although experiments without
heparin
or
using 1251-P1GF,31, which fails to bind heparin, gave
similar results. Similar results were also obtained when we
tested the binding of lZ5I-P1GF on transfected COS cells ex-
pressing full-length Flt-1
or
KDR
receptors (data not shown).
Binding of VEGF
or
PlGF to soluble receptors, endothelial cells,
or
transfected cells was not affected by bFGF, hepatocyte
growth factor,
or
insulin (data not shown), indicating that the
binding was specific. Thus, Flt-1, but not Flk-l/KDR,
is
a po-
tential high affinity receptor for PlGF.
PlGF Binds to Endothelial Cells-Ligand binding to endo-
thelial cells was performed to determine whether binding sites
consistent with those observed in vitro were duplicated on en-
dothelial cells.
As
shown in Fig.
6A,
1251-VEGF,6, bound
to
ACCE and could be only partially displaced by PlGF, indicating
the presence of both PlGF-sensitive and -insensitive VEGF re-
ceptors on ACCE cells. Similar results were observed in HUVE
cells (data not shown). Since HUVE cells consistently displayed
A
B
DA
0
5
10
15
20
25
5
10
15
20
25
Fraction
25650 PlGF Binds to Flt-1 and Potentiates VEGF Activity
higher expression of PlGF-sensitive VEGF binding sites than
ACCE cells, we chose to characterize PlGF binding to
HUVE
cells in greater detail (Fig. 6B). Binding of 1251-PlGFl,, to
100
A
1
10
100 1000
100
B
"i
20",
n
-.
1
10
100
1000
10000
VEGF165
(pM)
FIG.
4.
Binding
of
VEGF
to
Flt-1 IgG
or
to
KDR
IgG.
Microtiter
plates were coated with goat anti-human Fc IgG overnight and blocked
VEGF,,, (-12,000 cpdwell) and either Flt-1
IgG
or
KDR
IgG
and
with 10% fetal bovine serum in PBS. Wells were incubated with
lZ5I-
various concentrations
of
cold VEGF as described under "Experimental
Procedures." Wells were washed and individually counted in a
y
counter
to
determine binding. Nonspecific binding (determined in the absence
of
receptor-IgG) was subtracted.
LL
W
i3
>
100
50
0
1
10
100
1000
10000
100000
PlGF
(pM)
$
oL
10
HUVE
cells demonstrated two classes of sites, having affinities
of
230
2
16
PM
and
>2
nM.
HWE
cells express approximately
15,000 high affinity and
>30,000
low affinity PlGF receptors
per cell. These results demonstrate that endothelial cells ex-
press displaceable cell surface PlGF receptors. The high affin-
ity PlGF binding site is consistent with
Flt-1,
while the identity
of the low af'finity site remains to be determined.
Cross-linking
of
1251-VEGF,,S or '251-P1GFl,, to
HUVE
cells,
followed by immunoprecipitation of these complexes with an-
tibodies to
Flt-1
or Flk-l/KDR, was performed in the attempt to
identify the receptors involved in the binding events. Immuno-
precipitation
of
'251-VEGF,,,-cross-linked
complexes with anti-
serum directed against
Flt-1
yielded a relatively faint band
at
-190,000 and lower molecular weight bands consistent with
un-cross-linked VEGF (Fig.
7,
lane
1).
The anti-Flt-1 antiserum
also immunoprecipitated 12,1-P1GF (Fig.
7,
lane
3).
However,
cross-linked 1251-P1GFl,2~Flt-1 complexes
at
-
190,000 could not
be clearly visualized. Lower molecular weight complexes con-
sistent with possible oligomers of PlGF were observed (Fig.
7,
lane
3). Immunoprecipitation of cross-linked 1251-VEGF1,, with
anti-Flk-1lKDR antiserum yielded
a
strong receptor-ligand
complex band
at
-210 kDa. Un-cross-linked VEGF was also
observed (Fig.
7,
lane
5).
In contrast, the anti-Flk-1/KDR anti-
serum failed to precipitate significant lZ5I-P1GF (Fig.
7,
lane
7).
All of the observed labeling events could be inhibited by an
excess of cold VEGF (Fig.
7,
lanes
2,4,
and
6).
Preimmune sera
failed to precipitate lZ5I-VEGF or lZ5I-P1GF (data not shown).
PlGF Fails to Induce Tyrosine Phosphorylation in Endothe-
lial Cells-To
determine whether PlGF might play
a
direct role
in intracellular signal transduction pathways, we tested VEGF
or PlGF for their ability
to
trigger tyrosine phosphorylation of
membrane proteins in ACCE cells. In agreement with previous
studies (41), VEGF,,, stimulated tyrosine phosphorylation of
an -200-kDa membrane protein. In contrast, P1GFl,2,
at
all
concentrations tested, failed
to
stimulate tyrosine phosphoryl-
ation (Fig.
8).
PlGF Potentiates VEGF Mitogenic Activity in Cultured En-
dothelial Cells-We
next examined the effects of PlGF alone, or
w
60-
0
1
10
100
1000
Radioligand
(pM)
1
1
10000
=
50
5
40
5
30
;?
20
-
a,
v
U
0
D
G
;
10
ln
-
'Io
1
10
100
1000
10000
100000
Competitor added
(pM)
1
10
100
1000
10000
Radioligand
(pM)
FIG.
5.
PlGF
binds
to Flt-1 IgG
but not
to
KDR
IgG.
In
AX,
wells were incubated with radioligand, as indicated (-12,000 cpdwell),
receptor-IgG, and various concentrations
of
cold VEGF
or
PlGF, as described under "Experimental Procedures."
Panel
A
shows that PlGF
(548
m)
can displace '261-VEGF,,, bound to Flt-1 IgG but not
to
KDR IgG. In
B,
'251-P1GFl,, binding
to
Flt-1 IgG is competed by native PlGF,,,. In
C,
'251-P1GF,,, binding
to
Flt-1 IgG is competed by cold PlGF,,,
or
by VEGF,,,. Flt-1 IgG shows concentration-dependent binding of both 1251-VEGF,,,
and '251-P1GF,,,
(panel D).
In contrast, KDR
IgG
binds only 1251-VEGF,6,
(panel
E).
PlGF Binds to Flt-1 and Potentiates VEGF Activity
25651
0
i
IO
100
1000
loooo
101
600,
h
-
400
\
u)
n
U
c
v
g
200
m
0
B
Bound
(ImoVWell)
/
1
100
1000
10
125l-PLGF152
(pM)
30
10
FIG.
6.
Identification
of
1261-VEGF,, binding sites displaceable
by PlGF on ACCE cells and concentration dependence
of
lz5I-
PlGF,,, binding
to
HUVE
cells.
In
A,
confluent ACCE cells were
incubated with 12sI-VEGF,65 and varying concentrations
of
cold PlGF,5p.
Binding conditions were as described under "Experimental Proce-
dures." In
B,
confluent
HUVE
cells were incubated with varying con-
centrations of 1251-P1GF15p. Data plotted by the method of Scatchard
(38)
are shown in the
inset.
Nonspecific binding was determined in each case
with
5
pg of VEGF,, and was always
515%
of the total binding.
in combination with VEGF, on cultured endothelial cells. Con-
ditioned media of transfected CEN4 cells
or
partially purified
PlGF isoforms had little
or
no mitogenic activity for ACCE
or
HUVE cells (data not shown). Likewise, highly purified
PlGF,,,, tested up
to
530
ng/ml, did not promote the growth of
ACCE cells
(Fig.
9A). PlGF,,, also failed to stimulate ACCE cell
growth, in the presence
or
in the absence of heparin (Fig.
9B).
HUVE cells showed
a
"20% growth stimulation
at
concen-
trations of PlGF
>
100
ng/ml. In contrast, rhVEGF,,,, at the
concentrations of
2.5
or
10
ng/ml, induced
a
3-
to
4-fold increase
in final cell count
(Fig.
9, A-D). Purified PlGF did not affect the
maximal mitogenic response (efficacy)
to
VEGF16, in ACCE
cells (Fig. 9C). However, when PlGF was added in the presence
of low, marginally efficacious concentrations of VEGF,,,, this
resulted in
a
dose-dependent potentiation of the mitogenic ac-
tivity of VEGF. In ACCE cells, addition of 190 ng/ml PlGF,,2 to
0.03
ng/ml VEGF resulted in
a
4040% increase in cell number
over that observed with VEGF alone. However, PIGF,,, failed
to
potentiate low doses of bFGF, demonstrating that the observed
potentiation
is
specific for VEGF. Fig. 9D illustrates
a
repre-
sentative experiment. A similar potentiation of VEGF bioactiv-
ity by PlGF was observed in
HUVE
cells (data not shown).
PlGF Does Not Directly Induce Vascular Permeability but
It
Dramatically Potentiates VEGF Action-We then tested PlGF
12
3
4 5
678
14.3
-
or
llbI-PIGF,,, bound to
HUVE
cells.
HUVE
cells were incubated in
FIG.
7.
Cross-linking and immunoprecipitation
of
1261-VEGF,,
medium containing 1""IVEGF,6,
(lanes
1,
2,
5,
and
6)
or 12sII-PIGF,,,
(lanes
3,4,
7,
and
8).
Lanes
2,4,6,
and
8
reflect the addition of
2
pg/ml
cold VEGF,,,. Cells were washed and incubated with
0.5
mM BE?' cross-
linking agent. After cross-linking, the reaction was quenched, and then
the dishes were washed and extracted. The extracts were equally di-
vided and immunoprecipitated with rabbit antisera
to
Flt-1
(lanes
14)
or
Flk-l/KDR
(lanes
5-8).
The immunoprecipitation products were sub-
jected to electrophoresis on
4-20%
SDS-polyacrylamide gels. The gels
were dried and exposed to
a
phosphor-imaging plate and read on BAS-
2000
image analyzer.
D
ng/d
PlGF
200
c
116.5
-
49.5
c
FIG.
8.
Effect
of
VEGF or PlGF on tyrosine phosphorylation
of
ACCE cell surface glycoproteins.
ACCE cells cultured in 6-cm
dishes were stimulated with VEGF,, or various concentrations of
PlGF,,,
for
5
min at
37
"C, as indicated. Cell extracts were prepared and
then precipitated with wheat germ agglutinin agarose, followed by elu-
tion with Laemmli sample buffer. The eluates were subjected
to
elec-
trophoresis, transferred to polyvinylidene difluoride membranes, and
immunoblotted with an anti-phosphotyrosine monoclonal antibody.
in an in vivo system, the Miles vascular permeability assay
(Fig. 10).
This
assay provides
a
rapid and reproducible measure
of
one of the
known
activities of VEGF and is suitable for the
screening
of
multiple samples in
a
single animal. Tested alone
up to
500
ng/site, PlGF,,, showed no activity (Fig. 10,
24,
while VEGF16, was able to induce Evans blue extravasation in
a
dose-dependent manner (Fig. 10,
8-12),
in agreement with
previous studies
(7,
8,
40). Once again, PlGF was able to po-
tentiate the activity of minimally effective doses of VEGF. In
the presence of
500
ng of PIGF,,,,
a
dose of VEGF,,, (10 ng) that
alone gave
a
barely detectable response induced
a
near-maxi-
mal increase in dye extravasation, comparable to that induced
by 250 ng of VEGF,,, (Fig. 10,
14-18).
This potentiation was
dose-dependent. However, PlGF did not increase the maximal
25652
PlGF
Binds
to Flt-1
and
Potentiates VEGF Activity
100
I
60
A
0 0
1
10
100
1000
PIGF,,, (ng/ml)
0
no addition
2.5
nglrnl
VEGF+PIGF
A
21
0
1
10
100
PIGF,,,
(ng/ml)
0
PIGF152
+10
pg/ml
heparin
PIGF152
no heparin
2
ng/ml
FGF
401
2ot
ok
I
I I
0
1
1
10
100
1000
PIGF,,, (ng/ml)
""
Dl
0
10
ng/ml
VEGF
10
0
1
10
100
1000
PIGF,,;! (ng/ml)
FIG.
9.
Effects
of
PlGF on endothelial cell growth.
ACCE
cells were seeded into 6-well plates at
7,000
cells per well in the presence
of
VEGF,
PlGF,
or
bFGF, as indicated. After
5-7
days, cells were trypsinized and cell counts were determined using a Coulter counter. In
A,
cells were
cultured in the presence
of
varying amounts
of
PIGFIB,. In
B,
varying amounts
of
PlGF,,, were added, in the presence
or
absence
of
10
pg/ml
heparin. In
C,
each well received
2.5
ng/ml VEGF,,, in the presence
or
in the absence
of
varying amounts
of
PlGF,,,. In
D,
each well received either
0.03
ng/ml VEGF,,,
or
0.015
ng/ml bFGF and varying amounts
of
PIGF,,,. VEGF,,, at
10
ng/ml and bFGF at
2
ng/ml served as positive controls.
response (efficacy) of VEGF nor could bFGF potentiate the
activity of VEGF in this assay (data not shown).
DISCUSSION
Angiogenesis
is
a
tightly regulated process, integral to nor-
mal and pathologic conditions, including wound healing, cycli-
cal growth of the corpus luteum, rheumatoid arthritis, and
growth and metastasis
of
solid tumors. Recent evidence impli-
cates VEGF as
a
major regulator of these events
(18-22, 42,
43).
In addition, VEGF has been proposed to play
a
role in the
regulation of microvascular permeability
(7,8).
Yet, the signals
which regulate or modulate the actions of VEGF on the vascu-
lar endothelium are largely unknown. The recently identified
PlGF may share some of the functions of the VEGF polypep-
tides since: (i)
it
belongs to the same gene family, (ii) displays
significant sequence homology to WGF, and (iii) the PlGF gene
undergoes alternative exon splicing events reminiscent of the
VEGF gene. Therefore, it
is
possible that VEGF and PlGF
interact with similar pathways. To better understand this po-
tential interaction, we chose to purify to homogeneity the two
PlGF isoforms and test whether they are able to bind to the
known WGF receptors. The construction of soluble chimeric
Flt-1, KDR, and Flk-1 receptors represented
a
very useful tool
as
it
allowed
us
to individually characterize the properties of
each receptor both
in vitro
and, potentially,
in vivo,
without
interference from the other receptor
or
other factors. The bind-
ing characteristics of such soluble receptors are very similar to
those of the full-length receptors, indicating that the extracel-
lular domain contains all the information required for high
affinity binding.
PlGF bound with high affinity to Flt-1 IgG. That Flt-1
is
truly expressed in HUVE cells and available to bind PlGF
is
demonstrated by the identification of a cross-linked receptor-
lz5I-VEGF band of the predicted size (-190 kDa) following im-
munoprecipitation with an anti-Flt-1 antiserum. Such anti-
serum also immunoprecipitated Flt-1-bound '251-P1GF. Our
inability to directly visualize PlGF-Flt-1 cross-linked com-
plexes
is
likely to reflect both the lower affinity of PlGF com-
pared to VEGF as well as a low cross-linking efficiency of both
ligands to Flt-1. That such a receptor may also be present in
ACCE cells is suggested by the finding that these cells express
high affinity VEGF binding sites consistent with
Flt-1
(44).
Also,
PlGF
is
able to compete about half of the lz5I-VEGF bind-
ing to such cells. Furthermore, recent
in
situ hybridization
studies have demonstrated the ubiquitous expression of Flt-1
mRNA in endothelial cells
(45).
However, a recent study
(44)
failed to detect expression
of
the Flt-1 gene in ACCE cells by
Northern blot analysis of total RNA using
a
heterologous cDNA
probe. The RNA blotting methods employed in that study may
be less sensitive
or
specific than the radioligand binding
or
in
situ
hybridization studies. Alternatively, it
is
possible that
ACCE cells express
a
novel Flt-1-like VEGFPIGF receptor.
Remarkably, the PlGF proteins did not demonstrate high
affinity binding to Flk-1 IgG
or
to
KDR
IgG.
In agreement with
these findings, an anti-Flk-l/KDR antiserum failed to precipi-
tate '251-P1GF bound to
HUVE
cells. InFontrast, the antiserum
PlGF Binds to Flt-1 and Potentiates
VEGF
Activity
25653
FIG.
10.
Potentiation
of
VEGF
action
by
PlGF
in
the
Miles
vas-
cular
permeability assay.
Evans Blue was injected intracardially into
guinea pigs. After
1
h, various doses
of
VEGF1,,, PlGF,,,
or
a combi-
nation
of
both, were administered intradermally in 0.2 ml
of
PBS. PBS
(sites
1,
7,
and
13)
was used as control.
Sites
2-6
reflect the response to
500-,
250-,
125-,
50-,
and 25-ng doses of PIGF,,,.
Sites
8-12
show the
response
to
250-,
125-,50-, 20-, and 10-ng doses
of
VEGF,6,.
Sites
14-18
show the effect
of
the administration
of
a constant dose
of
VEGF,,, (10
ng) in the presence
of
500-, 250-, 125-,
50-,
and 25-ng doses
of
PIGF,,,.
efficiently immunoprecipitated 1251-VEGF-receptor complexes.
Although we cannot rule out the possibility that PlGF has
a
very low affinity for Flk-l/KDR
(Kd
>
100 nM), such a low
affinity binding would unlikely be physiologically relevant. A
recent report suggests that heparin enhances cross-linking of
VEGF to Flk-1 receptors (46). We observed only a modest in-
crease in binding of VEGF
or
PlGF to soluble receptors in the
presence of heparin (data not shown). However, since all bind-
ing assays described here were performed in the presence of
serum (which may contain heparin sulfate proteoglycans), the
need for exogenous heparin may have been obviated.
Purified PlGF showed little
or
no mitogenic activity for
ACCE and HUVE cells. Likewise, PlGF failed to induce extrav-
asation in the Miles vascular permeability assay. These find-
ings suggest that binding to
Flt-1
(and/or other unidentified
receptor) is not sufficient to trigger an effective mitogenic re-
sponse
or
to induce vascular permeability. Unlike the Flk-1
receptor (28, 291, Flt-1 fails to demonstrate VEGF-dependent
tyrosine phosphorylation (23). PIGF, tested over
a
wide dose
range, did not induce tyrosine phosphorylation in ACCE cells.
Interaction with Flk-l/KDR may be a critical requirement to
induce the full spectrum of VEGF actions. In this context,
a
negative-dominant Flk-1 mutant has been recently shown to
suppress tumor angiogenesis
in vivo
(43).
Intriguingly, however, PlGF was able to potentiate the action
of low, marginally efficacious, concentrations of VEGF on en-
dothelial cell growth. At such concentrations, VEGF is expected
to preferentially occupy
its
higher affinity receptors
(i.e.
Flt-1).
However, PlGF did not affect the mitogenic response to VEGF
at concentrations where the latter would be expected
to
occupy
both Flt-1 and Flk-lKDR receptors. Potentiation of VEGF ac-
tion by PlGF was replicated
in vivo
in the Miles vascular per-
meability assay. In fact, the potentiation
of
such activity of
VEGF by PlGF was much more striking than the potentiation
of the mitogenic action. Such an effect required approximately
a 10-20-fold molar excess of PlGF over VEGF. Interestingly,
this
is
similar to the difference in the relative affinities of
VEGF and PlGF for Flt-1. As mentioned above, recent
in situ
hybridization studies have documented the widespread expres-
sion of the Flt-1 mRNA in endothelial cells in normal adult
tissues, including the skin (45). This suggests’that interaction
of PlGF with Flt-1 may take place
in
vivo.
It will be of signifi-
cant interest to determine whether PlGF is also able
to
enhance
major
in vivo
actions of VEGF such
as
the ability to promote
angiogenesis in tumors (22)
or
in ischemic limbs (47).
One possible explanation for the effects that we report here
is that Flt-1 (or other unidentified high affinity VEGFPIGF
receptor) behaves as
a
“decoy,” having little
or
no transducing
activity alone. Therefore, binding to this receptor would limit
the bioactivity of VEGF by preventing its binding to a signal
transducing receptor. According to this hypothesis, PlGF would
act
to
release VEGF from Flt-1 and increase its availability
to
the more relevant Flk-VKDR. Such
a
mechanism would be
similar to that recently described for IL-1 receptors type I and
type I1 (48). Type I receptor is responsible for mediating IL-1
biological activities, whereas type I1 receptor has been shown
to
play an inhibitory role by acting as a “decoy” target for IL-1.
Regulation of Flt-1 and/or PlGF expression would provide
a
means to adjust the sensitivity of the endothelium to VEGF. In
this context, recent studies have shown that the PlGF mRNAis
expressed in HUVE cells (33). Therefore, PlGF may modulate
VEGF action by an autocrine mechanism.
Alternatively, the formation of heterodimers between KDR
and Flt-1 might confer new properties
or
ligand specificities
upon these receptors. Both possibilities are currently being
examined. In addition, we attempted to ascertain whether KDR
(or
other receptor) could acquire high affinity binding for PlGF
following stimulation with VEGF.
So
far, we have found no
evidence for such
a
mechanism in HUVE
or
ACCE cells. Fur-
thermore, low concentrations of VEGF spiked into binding as-
says did not reveal cryptic, VEGF-dependent, binding sites for
1251-P1GF,,2 in KDR IgG (data not shown).
The lack of pronounced effects
of
PlGF on endothelial cell
growth and vascular permeability suggest that this protein
may not be responsible for the direct induction of angiogenesis/
permeability and thus differs from VEGF. In this respect, it is
interesting to point out that high PlGF expression has been
reported in trophoblastic hydatiform mole and choriocarcino-
mas (321, conditions characterized by paucity or absence of blood
vessels (49). Rather, PlGF may function to enhance the activity
ofVEGF in situations where the concentrations of the latter are
limiting. Interestingly, the PlGF mRNAis expressed in the pla-
centa at substantially higher levels than the VEGF mRNA.’
Therefore, the molar excess of PlGF
versus
VEGF required
to
elicit the effects that we describe here may occur
in vivo.
The PlGF isoforms share several biochemical properties with
the VEGF polypeptides. PIGF,,, lacks the highly basic domain
characteristic of PIGF,,,. Accordingly, PlGF,,, fails to bind to
heparin
or
to
cation-exchange resins and is
a
diffusible protein.
This behavior is very similar to that displayed by VEGF,,, (16).
Although the 21-amino acid insertion confers strong heparin-
’
J.
E.
Park, H. H. Chen,
J.
Winer,
K.
A.
Houck, and
N.
Ferrara,
unpublished observations.
25654
PlGF Binds to Flt-1 and Potentiates VEGF Activity
binding properties on PIGF,,,, this protein differs from
VEGF18, or VEGF,,, since
it
is
substantially diffusible. In this
respect, the behavior
of
PIGF,,, is reminiscent of that of
VEGF,,, (16), a diffusible protein that nevertheless displays
significant binding
to
heparin-containing proteoglycans. Fur-
ther studies are required to assess whether PIGF,,, binds to the
extracellular matrix or whether proteolysis may play
a
role in
regulating PlGF bioavailability, by analogy with the VEGF
polypeptides (16, 17).
It has been observed that PlGF sequences are present in the
genome
of
a variety of species including
Drosophila melano-
gaster,
suggesting important or even essential functions
(32).
The availability of highly purified recombinant PlGF should
facilitate addressing such questions and may also permit the
identification of novel receptors which could in turn shed fur-
ther light on the significance of this protein. Furthermore, by
virtue of its selective interaction with VEGF receptors, PlGF
may constitute an important molecular tool
to
analyze the dif-
ferential role of such receptors in mediating the various biolog-
ical actions of this family
of
proteins and to dissect critical
structural domains involved in receptor binding.
Acknowledgments-We thank William J. Henzel
for
protein microse-
quencing, Allan Padua
for
amino acid analysis, Brian Fendly
for
anti-
bodies, Eileen Soriano-Szatkowski
for
performing the Miles vascular
permeability assay, and Louis Tamayo for graphics. We are grateful
to
Joffre Baker
for
helpful discussions and advice.
REFERENCES
1.
Ramsey, E. M., and Donner, M. W.
(1980)
Placental Vasculature and Circula-
2.
Gilbert, S.
F.
(1988)
Developmental Biology, 2nd Ed, Sinauer, Sunderland,
MA
3.
Bassett, D. L.
(1943)Am.
J.
Anat.
73, 251
4.
Klagshrun, M., and DAmore,
P.
A.
(1991)
Annu.
Reu.
Physiol.
35,217-239
5.
Folkman, J., and Klagsbrun, M.
(1987)
Science
235,442447
6.
Leung,
D.
W., Cachianes, G., Kuang, W. J., Goeddel, D. V., and Ferrara, N.
(1989)
Science
246, 1306-1309
7.
Keck,
P.
J.,
Hauser,
S.
D.,
Krivi,
G.,
Sanzo, K., Warren, T., Warren, T., Feder,
J., and Connolly, D. T.
(1989)
Science
246, 1309-1312
8.
Connolly, D.
T.,
Heuvelman, D. M., Nelson, R., Olander,
J.
V., Eppley,
B.,
Inuest. 84,
1470-1478
Delfino,
J.
J., Siegel, N. R., Leimgruber,
R.
S.,
and Feder,
J.
(1989)
J.
Clin.
tion, Georg Thieme, Stuttgart
9.
Basilico, C., and Moscatelli, D.
(1992)
Ado. Cancer
Res.
59, 115-165
10.
Yang, E. Y., and Moses, H. L.
(1990)
J.
Cell
Biol.
111,731-741
11.
Massague,
J.
(1990)
J.
Biol.
Chem.
265,21393-21396
12.
Klagsbrun, M., and DAmore,
PA
(1991)
Annu.
Reu.
Physiol.
53, 217-239
13.
Amento, E. P., and Beck, L.
S.
(1991)
Ciba Found. Symp.
157,115-123
14.
Leibovich,
S.
J., Polverini,
P.
J., Shepard, H. M., Wiseman,
D.
M., Shively, V.,
15.
Ferrara, N., Houck, K.
A,,
Jakeman, L., and Leung, D. W.
(1992)
Endocr.
Reu.
16.
Houck,
K.
A,,
Leung, D. W., Rowland,
A.
M., Winer, J., and Ferrara, N.
(1992)
17.
Park, J. E., Keller, G. A., and Ferrara, N.
(1993)
Mol.
Bid.
Cell
4, 1317-1326
18.
Ravindrath, N., Little-Ihrig, L., Phillips, H.
S.,
Ferrara, N., and Zeleznick, A.
and Nuseir, N.
(1987)
Nature
329,630432
13, 18-32
J.
Bid.
Chem.
267,26031-26037
J.
(1992)
Endocrinology
94,
1192-1199
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
Phillips, H.
S.,
Hains, J.,
Leung,
D. W., and Ferrara, N.
(1990)
Endocrinology
127,965-967
Jakeman, L. B., Armanini, M., Phillips, H.
S.,
and Ferrara,
N.
(1993)
Endo-
crinology
133, 848459
Breier, G., Albrecht, U., Sterrer,
S.,
and Risau,
W.
(1992)
Development
114,
521-532
Kim, K. J., Li, B., Winer, J., Armanini, M., Gillett, N., Phillips, H.
S.,
and
Ferrara,
N.
(1993)
Nature
362,8414344
deVries, C., Escobedo,
J.
A,, Ueno, H., Houck, K., Ferrara, E, and Williams, L.
T.
(1992)
Science
256,989-991
Terman,
B.
I.,
Dougher-Vermazen, M., Carrion, M. E., Dimitrov, D., Armellino,
D. C., Gospodarowicz, D., and Bohlen,
P.
(1992)
Biochem. Biophys. Res.
Shibuya, M., Yamaguchi,
S.,
Yamane,
A,,
Ikeda, T.,
Tojo,
A,, Matsushime, H.,
Commun.
187,1579-1586
Terman,
B.
I.,
Carrion, M. E., Kovacs, E., Rasmussen, B. A., Eddy, R. L., and
and Sato, M.
(1990)
Oncogene
5,519-524
Matthews, W., Jordan, C. T., Gavin, M., Jenkins, N.
A,,
Copeland, N. G., and
Shows,
T.
B.
(1991)
Oncogene
6,1677-1683
Quinn, T. P., Peters, K. G., De Vries, C., Ferrara, N., and Williams, L.
T.
(1993)
Lemischka,
I.
R.
(1991)
Proc. Natl. Acad.
Sci.
U.
S.
A.
88,
902G9030
Millauer, B., Wizigmann-Voos,
S.,
Schnurch, H., Martinez,
R.,
Moller,
N.
P. H.,
Proc. Natl. Acad.
Sci.
U.
S.
A.
90, 7533-7537
Kendell, R. L., and Thomas, K.
A.
(1993)
Proc. Natl. Acad.
Sei.
U.
S.
A.
90,
Risau, W., and Ullrich,
A.
(1993)
Cell
72, 8354346
10705-10709
Maglione, D., Guerriero, V., Vigliette, G., Delli-Bovi, P., and Persico, M. G.
Maglione, D., Guerriero, V., Viglietto, G., Ferrara, M. G., Aprelikova,
O.,
Ali-
(1991)
Proc. Natl. Acad. Sci.
U.
S.
A.
88,9267-9271
talo,
K.,
Del Vecchio,
S.,
Lei, K. J., Chou,
J.
Y., and Persico, M. G.
(1993)
Oncogene
8,
925-931
Hauser,
S.,
and Weich, H. A.
(1993)
Growth Factors
9, 259-268
Ferrara,
N.,
Leung, D. W., Cachianes, G., Winer, J., and Henzel, W.
J.
(1991)
Jakeman,
L.
B., Winer,
J.,
Bennett, G. L., Altar, C. A., and Ferrara, N.
(1992)
Methods Enzymol.
198, 391405
Su,
W., Middleton, T., Sudgen,
B.,
and Echols, H.
(1991)
Proc. Natl. Acad. Sei.
J.
Clin. Inuest.
89,
244-253
Capon, D. J., Chamow,
S.
M., Mordenti,
J.,
Marsters,
S.
A.,
Gregory, T., Mit-
U.
S.
A.
88,
10870-10874
and Smith, D.
H.
(1989)
Nature
337,525-531
suya, H., Bym, R. A,, Lucas, C., Wurm,
F.
M., Groopman,
J.
E., Broder,
S.,
Scatchard, G.
(1949)
Ann. N.
I:
Acad. Sci.
61,660472
Fantl, W. J., Escobedo,
J.
A,,
Martin,
G.
A,,
Turck, C. W., del Rosario, M.,
Kim, K. J., Li, B., Houck, K., Winer, J., and Ferrara, N.
(1992)
Growth Factors
Myoken, Y., Kayada, Y., Okamoto,
T.,
Kan, M., Sato, G. H., and Sato,
J.
D.
Brown, L. F., Yeo, K. T., Berse,
B.,
Yeo,
T.
K., Senger, D. R., Dvorak, H. F., and
(1991)
Proc. Natl. Acad. Sci. U.
S.
A. 88,
5819-5823
Millauer, B., Shawver, L. K., Plate, K. H., Risau, W., and Ullrich, A.
(1994)
van de Water, L.
(1992)
J.
Exp. Med.
176, 1375-1379
Barleone, B., Hauser,
S.,
Schollmann,
C.,
Weindel, K., Marme, D., Yayon,
A.,
Nature
367,57&579
Peters, K. G., De Vries,
C.,
and Williams, L.
T.
(1993)
Proc. Natl. Acad. Sci.
and Weich, H.
A.
(1994)
J.
Cell.
Biochem.
54,5646
Tessler,
S.,
Rockwell, P., Hicklin, D., Cohen, T., Levi, B.
Z.,
Witte, L.,
U.
S.
A.
90,89154919
Takeshita,
S.,
Zheng, L. P., Brogi, E., Kearney,
M.,
Pu, L.-Q., Bunting,
S.,
Lemischka,
I.
R., and Neufeld, G.
(1994)
J.
Biol.
Chem.
269, 1245612461
Colotta, F., Re, F., Muzio, M., Bertini, R., Polentarutti, N., Sironi, M., Gin,
J.
Ferrara, N., Symes,
J.
F., and Isner,
J.
M.
(1994)
J.
Clin.
Inuest.
93,662470
G., Dowker,
S.
K., Sirns,
J.
E.,
and Mantovani,
A.
(1993)
Science
261,
Kraus,
F.
T. (1990jinAnderson’s Pathology (Kissane,
J.
M., ed) pp.
1706-1710,
472475
The C. V. Mosby Co., St. Louis, MO
McCormick, F., and Williams, L. T.
(1992)
Cell
69, 413423
7,5344
