Download Human Endothelial Cells Express Integrin Receptors on the Luminal

Human Endothelial Cells Express Integrin Receptors on the Luminal Aspect of
Their Membrane
By Grazia Conforti, Carmen Dominguez-Jimenez, Adriana Zanetti, Michael A. Gimbrone Jr, Ottavio Cremona,
Pier Carlo Marchisio, and Elisabetta Dejana
Endothelial cells (EC) form a dynamic interface between
blood and the rest of the body. EC surface properties
promoting adhesion of reactive plasma proteins and/or
circulating cells might be of pivotal importance for the
homeostasis of blood and tissues. EC express multiple
integrin receptors that promote their attachment to the
subendothelialmatrix proteins. Among these receptors, 0 4 3 3
is of particular relevance on EC, since it is abundantly
expressed and can bind many different matrix and plasma
proteins. It is still unknown whether integrin receptors are
selectively located to the basal side of EC membrane or may
also be exposed on the cell surface in contact with blood.
This issue was addressed using different experimental approaches. First, selective surface radioiodination using lactoperoxidase (LP0)-latex beads and immunoprecipitation
analysis were performed. We found that cultured EC, similarly to human skin fibroblasts (HSF), expose a& on both
their apical (free) and basal (substratum-attached)surfaces.
qp1, mp1, and
This held also for other integrins such as
Immunoprecipitationdata were verified by morphological techniques. Immunofluorescenceand immunogold-staining of EC with a&, as well as with p1 subfamily antibodies,
showed a diffuse and granular distribution of these integrins
on EC surface. ~ y p 3and integrins were also detected on the
apical membrane of EC at higher magnification by scanning
electron microscopy (SEMI. Finally, data obtained on cultured EC were confirmed in vivo on immunogold-labeled
ultrathin cryosections of human vessels by transmission
electron microscopy (TEM). Data indicate, that in addition to
their role in promoting EC attachment t o extracellular matrix
proteins, integrin receptors of EC can be exposed to bloodstream and eventually be available for binding of plasma
proteins and circulating cells.
o 1992by The American Society of Hematology.
V
surface suitable for the binding of adhesive plasma proteins.
In the present study, we addressed this issue by studying
the distribution of integrins on EC membrane, and in
particular that of the vitronectin (vn) receptor ~433.This
protein is one of the major integrins expressed by cultured
EC. Its major ligand is vn, but it can also promote EC
adhesion to von Willebrand factor,z0,21
fibrinogen,22.23thromb o s p ~ n d i nand
, ~ ~thrombin.z
We show by multiple experimental approaches, that EC,
similarly to “nonpolarized” cells such as human skin
fibroblasts (HSF), expose a&, as well as other integrins, on
both their abluminal and luminal surface. The distribution
of integrins on endothelial blood interface might represent
a relevant mechanism for localization of plasma proteins on
their surface.
ASCULAR ENDOTHELIUM is a simple squamous
epithelial lining that faces blood with the luminal
surface and tissues of the body with the abluminal surface.
In view of its apparently simple structure, the endothelium
is devoid of the highly specialized and distinctive morphology of apical and basal membrane domains shown by many
epithelia. However, there is evidence that endothelial cells
(EC) express specialized functional and structural microdomains differentially distributed on the apical and basal
membrane.’ Endothelial luminal surface can modulate
important processes such as binding and activation/
inactivation of many elements of the coagulation and
fibrinolytic
binding of hormones or lipoproteins,6
and expression of adhesive receptors for leukocytes and
tumor
EC surface and reactivity can be modulated
by different stimuli, including thrombin,I0 histamine,“ or
inflammatory ~ytolcines.~~J~
Some studies indicate that E C
can express an apical/basal asymmetry of structure and
function. Muller and GimbroneI4 could directly show an
asymmetric distribution of glycoproteins on the EC luminal
and abluminal membrane.
EC express adhesive receptors for matrix proteins that
belong to the “integrin” family of membrane glycoprotein~.’~-’~
These receptors have several structural and functional homologies and consist of two noncovalently linked
subunits. Integrins located to the basal side of the membrane recognize and bind different components of the
basement membrane extracellular matrix with their
ectodomain. Conversely, their cytoplasmic domain links a
chain of proteins involved in F-actin microfilament assembly and cytoskeleton organization.’g Considering their major role in promoting matrix adhesion, it is natural to
assume that these receptors are localized to the basal
aspect of the cells. However, it is not yet known whether
integrins could also be exposed on the apical part of EC
membrane. This is particularly relevant since EC are in
constant contact with blood and might expose a reactive
Blood, Vol80, No 2 (July 15), 1992: pp 437-446
From the Istituto di Ricerche Farmacologiche Mario Negri, Milano,
Italy; the Vascular Research Division, Department of Pathology,
Brigham & Women’sHospital, Harvard Medical School, Boston, MA;
and the Dipartimento di S c i e w Biomediche e Oncolo@, Universith
di Torino, Torino, Italy.
Submitted August 1,1991; accepted March 20,1992.
Supported by C.N.R., Italian National Research Council (P.F.
Biotecnologie e Biostrumentazione, P.F. Tecnologie Biomediche e
Sanitarie e Progetto Speciale “Peptidi Bwattivi”), by Associuzwne
Italiana per la Ricerca sul Cancro (AIRC), and by NATO Research
Grant No. CT-042918%
Address reprint requests to Dr Grazia Conforti, Istituto di Ricerche
Farmacologiche Mario Negri; via Eritrea 62,20157, Milano, Italy.
The publication costs of this article were defrayed in part by page
charge payment. This article must therefore be hereby marked
“advertisement”in accordance with 18 U.S.C.section I734 solely to
indicate this fact.
0 1992 by The American Society of Hematology.
0006-497119218002-0018$3.00/0
437
438
CONFORTI ET AL
MATERIALS AND METHODS
Materials
The following reagents were used and are listed below with their
source: carboxylate-modified polystyrene latex beads (0.7 pm in
diameter) (Polysciences,Wamngton, PA); lactoperoxidase (LPO)
(Calbiochem, San Diego CA); poly-L-lysine hydrobromide, glucose
oxidase (GO) type VI1 fromdspergillur niger, phenylmethylsulfonyl
fluoride (PMSF), leupeptin, trypsin inhibitorfrom soybean (SBTI),
protein G-Sepharose 4B, and insulin-transferrin-sodium selenite
media supplement for cell culture (Sigma Chemical, St Louis,
MO); carrier-free Na151 (New England Nuclear, Boston, MA);
Triton X-114 (TX-114) and glutaraldehyde (Fluka Biochemicals,
Hauppague, NY);Trasylol (Bayer, Leverkusen, Germany); protein
A-Sepharose CL-4B (Pharmacia LKB Biotechnology, Uppsala,
Sweden); gold particles conjugated to goat anti-rabbit or antimouse IgGs (5 and 20 nm in diameter) and silver-enhancing kit
(Bio Cell Research Laboratory, Cardiff, UK); paraformaldehyde
and sucrose (Merck-Schuchardt, Darmstadt, Germany); mowiol
4-88 (Hoechst, Frankfurt/Main, Germany); 14C-methylated protein molecular weight standards (Amersham International, Buckinghamshire, UK); all electrophoretic reagents (Bio-Rad, Richmond, CA); culture reagents (GIBCO, Paisley, UK); secondary
antibodies used in immunofluorescence experiments (Dakopatts,
Glostrup, Denmark). Human plasma vn and fibronectin (fn) were
prepared as previously described.%
Antibodies
Anti-glycoprotein (Gp)IIIa (anti+,) rabbit serum was prepared
in this laboratory as describedz6; anti-ps (cytoplasmic domain)
rabbit serum and affinity-purifiedrabbit IgGs anti-a,,p,pswere a gift
Dr E. Ruoslahti (La Jolla Cancer Research Foundation, La Jolla,
CA); anti-human plasma angiotensin-converting enzyme (ACE)
rabbit IgGs were donated by Dr J.J. Lanzillo (Tufts University,
Boston, MA), and anti-a5 (cytoplasmicdomain) rabbit serum was
donated by Dr G. Tarone (University of Torino, Torino, Italy).
Monoclonal antibodies (MoAbs) and the scientists who kindly
provided them are listed below: affinity purified LM 142 and LM
609, directed, respectively, to the a subunit and to an a/p
complex-dependentepitope of the vn receptor, Dr D.A. Cheresh
(Research Institute of Scripps Clinic, La Jolla, CA); VIPI-2 (ascitic
fluid), anti-human platelet GpIIIa (p3), Dr W. Knapp (University
of Vienna, Vienna, Austria); B1E5 (sovranatant) anti-VLA5 a
subunit and AIIB2 (sovranatant) directed to the VLA p subunit,
Dr C. Damsky (Universityof California, San Francisco, CA); A1A5
(ascitic fluid), directed to the VLA p subunit, Dr M.E. Hemler
(Dana Farber Cancer Institute, Boston, MA); P1E6 (sovranatant),
anti-VLA2 a subunit, Dr E. Wayner (Fred Hutchinson Cancer
Center, Seattle, WA); TS2/7 (sovranatant) anti-VLAl a subunit
and HP2/1 (sovranatant) anti-VLA4 a subunit (both cleaved and
uncleaved a chain are recognized), Dr Sanchez-Madrid (Hospital
de la Princesa, Madrid, Spain); 5143 (mouse serum) anti-VLA3 a
subunit, Dr L.J. Old (Memorial Sloan-Kettering Cancer Center,
New York, NY); MT78 (affinity-purified IgGs) anti-VLA6 a
subunit, Dr Eberhard Klein (Universityof Ulm, Ulm, Germany).
Cell Cultures
Human umbilical vein endothelial cells (HWEC) were isolated
from normal term umbilical cord veins by collagenase perfusion as
previously d e s ~ r i b e d . ’Cells
~ ~ ~were
~ grown on tissue culture dishes
coated with gelatin (0.5%) in 20% newborn calf serum (NCS)medium 199 (M199) supplemented with endothelial cell growth
supplement (ECGS, prepared from bovine brain) (50 pg/mL),
heparin (100 pg/mL), and penicillin (100 U/mL)/streptomycin
(100 pg/mL) (pedstrep) and used between the second and third
passage.
Human saphena vein EC were kindly provided by Dr C. de
Castellarnau (Fundacio d’Investigacio Sant Pau, Barcelona, Spain).
Cells were isolated from specimens of normal veins removed
during coronary bypass surgery from patients who were in overall
good health. Briefly, after the vessel was opened and rinsed with
phosphate-buffered saline (PBS), pH 7.2, EC were isolated by
incubation with trypsin-EDTA solution and removed by gently
scraping. Three hours after cell seeding, debris and possible
contaminating smooth muscle cells were removed by vigorous
washing. Cells were characterized as EC by their “cobblestone”like morphology and by the presence of von Willebrand factor as
described.n Cells were cultured in M199, 20% human serum,
supplemented with ECGS (50 pg/mL), heparin (100 pg/mL), and
pen/strep, on a 0.5% gelatin coating and used at the sixth passage.
Bovine microvascular EC from adrenal cortexB were kindly
donated by Dr R. Montesano (University of Geneva, Geneva,
Switzerland). Cells were cultured in minimum essential medium15% NCS containing penlstrep and used at the tenth passage.
HSF were obtained from forearm skin biopsies following the
procedure described by Simpson and S t ~ l b e r g ?grown
~
in 20%
NCS-M199, pen/strep, and used between the eighth and tenth
passage.
Selective Radiowdination of Apical, Basal, and Total Cell
Surfaces
LPO-beads were used to selectively label the apical (A), basal
(B), and total (T) surfacesof cells. The entire procedure followed a
previously described detailed method.14This technique is based on
the fact that radiolabeling with LPO-latexbeads is restricted to the
externally disposed membrane proteins. When cells are labeled in
monolayer,LPO-beadsdo not have access to the basal (substratumattached) cell surface, because of the beads’ large size.
To selectively radioiodinate the basal surface of living EC, we
inverted the H W E C monolayer on poly-L-lysine-coated coverslips. Briefly, tissue culture plastic coverslips were precoated with
poly-L-lysine (1 mg/mL in Hanks’ balanced salt solution [HBSS]
for 5 minutes at room temperature), followed by several washes in
distilled water, and air-dried. Confluent HUVEC were washed
three times in HBSS at 3TC, then a poly-L-lysine-coatedcoverslip
was gently placed down onto the monolayer; several seconds later it
was lifted up using a fine forceps. Confluent sheet of HUVEC
remained attached by their apical surfaces with their original basal
surfaces now in contact with the surrounding medium. The
inverted monolayers were immediately placed in cold Dulbecco’s
PBS (DPBS) containing calcium and magnesium and protease
inhibitors (0.3 mg/mL SBTI and 1.4 KIU Trasylol). This inversion
technique was used on primary cultures of HUVEC grown on
uncoated culture dishes.
Total cell surface labeling was performed on viable populations
of nonenzymatically resuspended cells using 30 pmol/L EDTA in
HBSS without calcium or magnesium and removed by gentle
scraping.Resuspended cells were transferred to a culture dish well,
mixed with LPO-latex suspension, and allowed to settle on ice
(under these conditions, some LPO-latex would be under the
cells).
Confluent monolayers of H W E C and HSF were used for
radioiodinationof the apical surface. The culture plates containing
the A-, B-, or T-cell preparations and LPO-latex (1:400 dilution
from LPO-latex stock)14 were centrifuged (3,100 rpm, 4°C for 5
minutes) in microtiter plate carriers in a Minifuge T (Heraeus,
Osteroole, Germany). This brought the beads down in intimate
contact with the cell surface,where they were adherent. Radioiodi-
LUMINAL INTEGRINS IN HUMAN ENDOTHELIAL CELLS
nation was performed using 300 pCi/mL carrier-free NaIzI and 12
pU glucose oxidase (G0)lmL in DPBS-20 mmol/L dextrose as
previously described.I4
In some experiments, HUVEC were seeded on vn- and fn-coated
wells in MI99 supplemented with media supplement ( 1 9 dilution). After 3 hours, at maximum cell spreading, monolayers were
selectively radioiodinated on the apical surface as described above.
&paration of Samplesfor Immunoprecipitation and
SDS-PAGE
Radioiodinated samples were lysed by scraping with 1% TX-I 14
in IO mmollL Tris. 150 mmol/L NaCI, pH 7.4. containing protease
inhibitors (1 mmollL PMSF, IO pglmL leupeptin, 20 KlUlmL
Trasylol) and spinningout LPO-latex beads and insoluble material
in a microfuge for 5 minutes at 4°C. The resulting supernatant was
spun again in a microfuge for 25 minutes at 4°C. and aliquots used
for trichloroacetic acid (TCA) precipitation analysi~.’~
Total cell
lysate samples were mixed directly with 4x sodium dodecyl sulfate
(SDS)-sample buffer containing 2.5% 2-mercaptoethanol and fractionated by SDS-polyacrylamidegel electrophoresis (PAGE) or
kept frozen (-20T) until used. Labeled cell lysates were immunoprecipitated with the following polyclonal antibodies or MoAbs:
anti-GPllla (40 pLlmL), anti-ps (40 pLlmL), anti-ACE (80
pglmL). antias (30 pLlmL), and normal rabbit serum (40
pL/mL) coupled to protein A-Sepharose CL-4B (50 pL); LM 142
(20 KglmL), BIE5 (100 pL/mL), and an irrelevant MoAb (20
pg/mL) were coupled to protein G-Sepharose4B (50 pL): TS2/7
(200 pLlmL), PIE6 (200 pL/mL), 5143 (4 pLlmL), HP2/1 (200
pL/mL), MT78 (4 pLlmL), and an irrelevant MoAb (4 pglmL)
were coupled to protein A-Sepharose CL-4B (50 pL) that was
previously incubated with rabbit anti-mouse lgGs (5 KglmL).
Immunoprecipitation was performed as previously described.”
lmmunocomplexeswere subjected to SDS-PAGE and the electrophoresed gels were fixed, dried, and analyzed by autoradiography.
In some experiments, quantitative data on Iz1-aJ33 or 1151asf3~
were achieved by removing the gel that contained radiolabeled
integrins (the position of the radiolabeled chains was determined
by placing an x-ray film on the dried gel) and counting them in a
gamma counter.
Immunofluorescence
Immunofluorescence tests were performed essentially as described.ZLBriefly, cells were washed twice with DPBS and fixed
with 3% paraformaldehyde and 2% sucrose in PBS (15 minutes,
room temperature), washed in Tris-buffered saline (TBS)-0.2%
bovine serum albumin (BSA). and then incubated with a p3 MoAb
(VIPI-2) for 45 minutes at 3PC without any previous detergent
permeabilization. After rinsing in TBS-0.2% BSA, rhodaminetagged rabbit anti-mouse IgGs were applied for 30 minutes at 3PC
in TBS-I% BSA. Coverslips were then mounted in mowiol 4-88.
Routine observationswere performed in a Zeiss Axiophot photomicroscope equipped for epifluorescence and plan-apochromatic
lenses (Carl Zeiss, Oberkochen, Germany). Fluorescence images
were recorded on Kodak T-Max 400 films exposed at 1,OOO IS0 and
developed in T-Max Developer (Kodak Ltd, Hemel, UK) for IO
minutes at 20°C.
Light Microscopy
EC (human umbilical vein, human saphena, and bovine capillary) were grown to confluency in 96-well plates. After a DPBS
wash, cells were fixed for 15 minutes at room temperature (3%
paraformaldehyde and 2% sucrose in PBS) then washed three
times with DPBS, and DPBS-I% BSA treated for 20 minutes at
3PC, followed by two washes with TBS-03% BSA, pH 7.2. The
439
specific antibody (Ab) was then added in the same buffer for 40
minutes at 3PC, followed by two washes with TBS-0.5% BSA, pH
7.2, and two washes with TBS-0.5% BSA, pH 8.2. Finally, immunogold particles (20 nm diameter) conjugated to goat anti-rabbit or
anti-mouse lgGs (k70 dilution) were added for45 minutes at room
temperature. After several washes, beads were fixed by 2%
glutaraldehyde in PBS for 20 minutes at room temperature and
revealed by silver enhancement. Gold particles are not visible in
the light microscope, unless they are grown in situ using silverenhancing. The final signal (brownlblack) gives a precise localization that can be observed under a common light microscope. This
last step was performed following the manufacturer’s procedure.
Pictures were taken in a Zeiss IM35 photomicroscope on Kodak
plus-X pan 125 films.
Immunoelectron Microscopy
Scanningelecmnmicmscopy. Cells were plated on 0.5% gelatinprecoated glass coverslips to reach confluency overnight. After two
washes in DPBS followed by two saturating incubations with
DPBS-I% BSA (IO minutes each), cells were labeled, prior to
fixation, with 100 pglmL of anti%& affinity-purifiedantibodies3’
or the
MoAb AIIB2 in DPBS-I% BSA-0.01% NaN3 for 40
minutes at 3PC. After thorough rinsing with DPBS. cells were
fixed in 3% paraformaldeyde and 2% sucrose in PBS, rinsed with
TBS-0.2% BSA, pH 8.2, and finally incubated with gold particles
conjugated with goat anti-rabbit IgGs. After dehydration with
ethanol, specimens were COz critical-point dried and coated with
carbon, then examined in a Cambridge Stereoscan 360 scanning
Mr x 10-3
200
A
B
A
B
A
B
-
100
92.5
-
Fig 1. Selactlve rodlolodinationof the apical (A) and b a u l (B) EC
surface. The iodlnatlon p m e m of three independent experiments h
compared. Cells were radiolabeled as described in Materials and
Methods. Before cell extraction, the A and B surfaces of EC monolayers were treated with 30 pmol/L EDTA for 10 minutes at 37% to
remove any adsorbed protein from cell culture serum. Each lane was
obtained by a pool of cells from two 10”.
wells for the A surface and
from four 2zmZ wells for the B surface. Each set of AB surface-labeled
cells was performedfrom a different batch of cells. Each lane contains
300,000cmp of TCA-precipitablematerial. Cell extracts were mixed
with 4 x sample buffer and fractionated in a gel gradient 5% to 12%
acrylamlde-SDS under reducing conditions.
CONFORTI ET AL
440
C
B
A
EC
EC
A B
1
..
HSF
A T B
A
EC
HSF
A T B
EC
B
A
A
0
.
)
-
R
R
NR
R
Fig 2 Selective radlo1adin.tlon of a p k l (A), bmal (E), and total (1)cell surface, followed by extradon and lmmunopreclpltatlonanalysis using
A b . (A) Each lane contains 200.000 cpm of TCA-precipitable cell lysate. Note that the asymmetric distribution of apical and basal surface
proteins includes both integral and membrane-associated membrane proteins, because apical and basal layers were processed for SDS-PAGE
analysis without subjecting them to EDTA washing"; 5% to 12% acrylamlde gel gradient was used and samples were run under reducing
conditions (2.5-hour exposure). (B) 150,000 cpm of TCA-precipitable HUVEC and HSF lysates were immunoprecipitatedby anti-Gpllla rabbit
serum (40rLlmL), followed by SDS-PAGE fractionation (7.5% gel) under reducing (R) and nonreducing (NR) conditions. (C) 300,000 cpm of
TCA-precipitablecell lysate from A and B surface of EC were immunoprecipitatedby three successive cycles using 40 r L / mL of anti-Gp 111a serum
(anti-&) each time. Lanes 1, 2, and 3 show the immunoprecipitated products obtained in each cycle, respectively. This sequential
immunoprecipltationanalysis shows that anti-Gpllla serum was used at concentrationable to produce a complete depletion of the antigen a t the
first cycle for all samples (B and C). 14C-labeiedmolecular mass markers are on the left lane of each gel (M,200,000 myosin; M, 100.000-92.500
phosphorylase b; M, 69,000 BSA; M, 46,000 ovalbumin; M, 30,000 carbonic anhydrase; M, 14,300 lysozyme).
electron microscope (SEM) equipped with a solid-statebackscattering device and a tungsten cathode (Cambridge Instruments,
Cambridge, UK).
Tmnsmivion electron microscopy. Samples of normal blood
vessels from leg peripheral circulation (essentially femoral artery
and vein branches) obtained from a vascular surgery department,
were fixed in 3% paraformaldehyde and 2% sucrose in PBS for 10
minutes at room temperature, followed by thorough rinsing for 20
minutes in PBS. The technique was essentially that reported by
T o k u y a s ~ . ~Briefly,
U~
each sample was frozen in liquid Nz after 2
hours of infiltration in 2.3 mol/L sucrose. Ultrathin cryosections
were obtained with a Reichert-Jung FC4D cryocutting device
mounted on a Ultracut E ultramicrotome (C. Reichert, Wien,
Austria). The cutting temperature was -110°C (sample) and
- 1 W C (blade). Sections were collected on Ni grids coated with a
15 nm Formvarcarbon film (Polaron Equipment Ltd. Watsord,
UK) and incubated with the primary Ab (anti-Gpllla rabbit
serum),26 followed by incubation with 5-nm gold particles conjugated with goat anti-rabbit IgGs. Finally, cryosections were embedded in LR-White" and observed i n a Philips EM410 electron
microscope (Philips, Eindhoven, The Netherlands).
Mr x 10-3
ACE
VN-rec
NI
-200
100
92.5
69
46
RESULTS
Detection of Integrin Receptors by Radioiodination of the
Apical Surface of EC
LPO-latex beads were used for selective radioiodination
Of the apical and basal domains, as well as that of the whole
surface of HUVEC. This technique14 is based on the fact
that LPO-latex beads adhere to the cell surface and
radiolabeling is restricted to beadcontacted stretches of
the Cell membrane* As previously
the resulting
radioiodination patterns of the apical and basal HUVEC
I,-.A
B
A
B
A
B
3. meoc& a d ACE d i d b d o m on a p k l and haul H
~
surface am shown. Labeled cell extract was prepared as described in
Fig 1; 300.000 cpm of TCA-precipitable material was lmmunoprecipitated, respectively, by anti-Gpllla serum, anti-ACE IgGs, or normal
as dmribed in Materials
rabbit sarum. lmmunocomplexes
and Methods were fractionated by 7.5% SDS-PAGE under reducing
conditions and revealed by autoradiography.
C
441
LUMINAL INTEGRINS IN HUMAN ENDOTHELIAL CELLS
Fig 4. Intogrin characteriza-
-
.-. --
1
-
tion on H W E C apkal surface by
immunoprecipitation analysis.
Selective radioiodination of HUVEC apical surface and cell lysate
was obtained as described in Materials and Methods:200.000 cpm
of TCA-precipitable material for
each sample was coupled to integrin Ab,. Immunocomplexes
were run in a 7.5% gel under
nonreduclng(NR) or reducing (R)
conditions and revealed by autoradiography. (A) Anti-B, (rabbtt
serum); anti- (LM 142); a n t i g
(BlE5); NI (irrelevant MoAb). (B)
Anti-a, (TS2/7); anti-a2 (PlE6);
a n t i s (J143); anti(HP2/1);
a n t i g (BlE5); anti- (MT78);NI
(irrelevant MoAb). 14C-labeled
molecular mass markers are on
the right side of each gel (see
legend to Fig 2).
---
B
Mr
Mix
al a 2 a3 a4 a 5 a6 NI
.
..
200
200
100
92.5
100
92.5
69
69
46
46
NR
membranes were distinct (Fig 1). such as specific bands
were markedly enriched on one surface domain and absent
or weak on the opposite side. The radioiodination pattern
of the apical/basal membrane was comparable in different
independent experiments as shown in Fig 1. Some difference was found comparing cells that had been washed with
EDTA after labeling (Fig 1) with cells that had not been
subjected to EDTA washing (Fig 2A), this can be explained
by absorption of serum-derived proteins. By immunoprecipitation analysis, a,p3was found both on the apical and basal
domains of HUVEC. N o obvious differenceswere observed
comparing the M,of basal and apical a,p3 under reducing
and nonrcducing conditions (Fig 2B). This receptor was
also present on the apical surface of “nonpolarized” cells
such as HSF, and showed the same Mr as that of HUVEC
(Fig 2B). The results reported in Fig 2 were obtained using
p3 polyclonal Abs; comparable results were obtained with
the a,MoAbLM 142 (see also Fig 4A). For comparison, the
membrane distribution of ACE was analyzed. This molecule has been described as distributed in a polarized
fashion.I4 Using cell extract from the same radioiodinated
material,
was detected by immunoprecipitation analysis both on the apical and basal side of HUVEC, while ACE was
only found on the apical surface running at 175 Kd (Fig 3).
To know whether other integrins were also exposed on
HUVEC apical membrane, immunoprecipitation was performed with Abs directed to different integrin chains.
Mainly a&, asplrand a& were located to the luminal
HUVEC surface as shown by immunoprecipitation analysis
with MoAbs to a2, as, and as,while small amounts of a&
were detectable. In contrast, a1 and cq MoAbs did not
precipitate any detectable material. Since the a, MoAb
(LM142) coprecipitated a single band corresponding to f33
in M, (Fig 4A), we assumed that a,p~and a,Bs integrins
were not detectably exposed on HUVEC apical surface; ps
was also undetectable on this surface (Fig 4A) using a
polyclonal anti-ps chain Ab. A n t i e MoAb coprecipitated
R
an additional band running at 185 Kd, which is currently
under investigation. To establish whether inclusion of ~ $ 3
or as& in adhesion plaques (on the basal aspect of the cell
membrane) would prevent their recovery on the apical
surface, selective apical iodination of cells seeded either on
vn or fn was performed. Previous immunofluorescence
data26showed that
and aspIare clustered in adhesion
plaques when cells are adherent on vn or on fn, respectively. Immunoprecipitation data using anti-GpIIIa rabbit
serum and anti-as (cytoplasmic domain) rabbit serum, show
that a,p3 and asplwerc detected on the apical surface of
the cells irrespectiveof the substrata used. For instance, the
radioactivity associated to a,p3 complex (see Materials and
Methods) was 2,012 cpm on vn and 1,732 cpm on fn,
respectively; the radioactivity associated to a& complex
was 1,492 cpm on fn and 2,052 cpm on vn, respectively.
Topographyof Integins
Light micmscop. As shown in Fig 5, immunofluorescence staining of HUVEC, using the BJ MoAb (VIPI-2),
Fig 5. Immunofluorescencedetection of a,& on the apical surface
of a nonpermeabilized HUVEC. The signal generated by the Ab is in
small diffuse granules. Bar = 5 pm.
CONFORTI ET AL
442
showed a diffuse and granular distribution on HUVEC
surface. A comparable pattern was obtained using a, (LM
142) and ~ $ (LM
3
609) MoAbs, whereas no staining was
observed with irrelevant MoAbs (not shown).
We used immunogold labeling of EC monolayers, followed by analysis with light microscopy, as a routine
method for the detection of integrinson many EC types. As
shown in Fig 6, this system allowed us to identify a , $ 3
receptor on the apical membranc of bovine microvascular
EC (Fig 6a) using an anti-GpIIIa polyclonal Ab, and of
HUVEC (Fig 6b) and human saphena EC (Fig 6c)using the
:
:
0
.
~
MoAb LM 609. For human saphena, the topographic
distribution of a2 (Fig 6e) and the as (Fig 69 chains, using
P1E6 and B1E5 MoAbs, respectively, are also shown. A
marked staining was obtained when cells were labeled with
$, MoAb A1A5 (Fig 6g). Some background could be
obscrved on HUVEC (Fig 6d) labeled with an irrelevant
MoAb or with saphena EC (Fig 6h) labeled with nonimmunc rabbit serum.
Electron microscopy. The fine location of a,& to the
apical domain, was also detected by SEM immunocytochemistry. As shown in Fig 7a and b, a polyclonal a , $ 3 Ab
-c
. .,..,
L.
Fig 6. Immunogold-labeling
and sihrw-enhancing of EC, analysis by lQht microscopy. Confluent EC monolayers from different sources (a. bovine adrenal
cortex microvascular EC; b and
d, HUVEC; e, e, 1, g, and h, saphena EC) were fixed and incubated with integrln Ab8 (seeMaterials and Methods). (a) AntiGpllla serum (1:lOO); (b end c)
anti*@, (LM 609-50 Ng/mL); (d)
irrelevant MoAb (50 pg/mL); (e)
anti-a2 (PlE6-undiluted); (1)
anti-a, (ElES-undiluted); (g)
anti-@, (AlA5-1:50); (h) i m l e vant MoAb (undiluted). Ear = 20
pm.
LUMINAL INTEGRINS IN HUMAN ENDOTHELIAL CELLS
Fig7. SEM lmmunogoldcytochemical loccllltatlonof Integrim
on HUVEC and HSF. a, c, e, and g
represent secondaryelectron micrographa, while b, d, f, and h are
the corrmponding backscattering pattems produced by gold
particles. Affinity-purified lgGa
anti%& were used in a and b
(HUVEC) and in c and d (HSF);
anti-p, MoAb waa used in g and
h, and nonimmune rabbit lgGs in
e and f (HUVEC). Bar = 1pm.
.
t
.
443
--
t
1.
,
1
produced small gold particle clusters on HUVEC apical
membrane. This pattern was similar to that observed on
HSF (Fig 7c and d). Occasionally, some staining with
Ab was observed on the cell matrix (Fig 7a and b). This was
attributed to footprints of cells detached during washing
procedures. No gold particles were observed with a nonimmune serum eithcr on HUVEC (Fig 7c and f) or on HSF
(not shown). Comparable results were obtained using
MoAb LM 609 to the a 4 3 3 complex (not shown). The
topography of 6, chain was also studied with the same
technical approach using MoAb AIIBZ. Figure 7g and h
shows the pattern of distribution of 6, chain on HUVEC
apical membrane.
The data reported above suggest that integrins are
exposed on the nonadhering apical domain of cultured EC,
but do not provide information as to whether this is also the
case in vascular endothelia in situ. To address this issue, we
also studied the distribution of 4 3 3 in ultracryosectionsof
intact human vessels. Figure 8 shows one example (from
human femoral artery branch EC) of the location of a 4 3 3
CONFORTI ET AL
444
Fig 8. E M lmmunogoldcytochemical localization of a,& on
an ukracryosectlon of a human
femoral artery branch endothelium. Ftve-nanometer gold partlcles coupled to the secondary
antibody detected bound @, Abs
(anti-Gpllla rabbit serum) on the
luminal surface of the endothelium either as single particles
(small arrowheads) or in small
clusters (largearrowhead). *Web
bel-Palade body. Bar = 0.2 pm.
obtained in "EM. Five-nanometer gold particles coupled to
goat anti-rabbit IgGs detect
on the apical surface of
EC occasionally in small aggregates. The density of gold
particles is consistent with the density of gold particles
detected by SEM in cultured cells at lower magnification.
DISCUSSION
This report provides evidence, obtained by different
experimental procedures, including biochemical and morphological techniques, that EC can expose integrin receptors and, in particular, u,,f3~on their apical surface. We first
performed a selective iodination of HUVEC apical membrane by LPO-beads followed by immunoprecipitationwith
Abs directed to different integrin chains. This technique
showed that u&. in contrast to other HUVEC membrane
proteins, was not topographically polarized, but was exposed on both sides of the HUVEC membrane. No obvious
qualitative difference was found in such receptor extracted
from both surfaces. The luminal localization of ~ $ 3 in
HUVEC appeared to be a fcature common to the other
major integrins expressed by cultured HUVEC, ie,
u ~ f 3 ~15,f 3 1 ,and u&. These immunochemical observations
were validated by direct topographical identificationof wf33
and PI integrins at the lcvcl of resolution of light and
electron microscopy (SEM) on HUVEC apical membrane.
A nonpolarized cell type such as HSF presented a comparable distribution of ~ $ 3 on the luminal surface. The apical
localization of ~ $ 3 and ~ 5 f 3 did
l
not seem to be affected by
the substratum on which the cells have been seeded (vn or
fn). Because ~ $ 3 and ~ 5 f 3 1 are clustered in adhesion
plaques on vn and fn.
these data indicate that
clustering of the basal receptors does not significantly
modify their apical distribution.
Integrin exposure at the apical aspect of the membrane
was not a specific property of HUVEC, because endothelium cultured from saphena vein and adrenal gland microvessels also presentc? these characteristics. In addition,
when tissue sections of human vessels were studied, a&
was detected on the luminal surface of the endothelium in
situ, thus indicating that this type of distribution was not
artifactually related to the culture conditions of EC.
The biological role of integrin molecules on EC surface
in contact with blood remains to be established. In this
study, we focused our attention on ~ $ 3 . This integrin is of
particular interest, since it can bind different plasma
proteins able to modulate thrombosis and immune reactions includingvn, von Willebrand factor, fibrinogen, thrombospondin, and thrombin.'-L5 The possibility that the luminal membrane of the vascular endothelial lining might bind
and expose vn has important implications. In addition to its
potential adhesive properties for bloodstream circulating
cells, vn may link and modulate the activity of several
biologically active plasma proteins such as thrombin/
antithrombin 111:53 pla~minogen?~
plasminogen activator
i n h i b i t ~ r ,and
~ ~ .the
~ ~terminal components of the complement system.4"
We have previously r e p ~ r t e d ~that
l . ~ ~fibrinogen could
bind to EC in monolayers and that this is inhibited by f33
antibodies. Preissner et alSRshowed that vn could bind to
confluent EC in addition to the extracellularmatrix. Furthermore, it was recently found43that SruphyrocoCcus uuwus
attachment to EC was mediated by fibrinogen bound from
plasma onto the endothelial apical surface. Even if these
data are suggestive, direct evidence that ~ $ 3 or the other
integrins on EC luminal side can bind soluble plasma
ligands is still lacking. In platelets, GpIIb-IIIa can promote
cell adhesion to fibrinogen-coated surfaces, but it cannot
bind the soluble form of this protein unless the platelets are
activated. It remains to be established whether integrins on
EC luminal side, in a way comparable to platelets, need cell
activation to bind appropriate soluble ligands.
Monoclonal antibodies to platelet GpIIb-IIIa are currently under study as antithrombotic agents. One of these,
7E3,44is particularly promising. This antibody recognizes
~ f 3 and
3
GpIIb-IIIa. From the results of this study, it is
therefore conceivable that EC, as well as platelets, react
with this antibody in vivo. In in situ staining of human
tissues, ~ $ 3 is present in large vessel endothelium, but is
less expressed in the microcirculation.Is In preliminary
studies (L. Ruco et al, unpublished results), we found a
similar pattern of distribution for 7E3. Further studies are
required to clarify whether the binding to the endothelium
could significantly affect the blood levels and/or the activity
of this antibody.
445
LUMINAL INTEGRINS IN HUMAN ENDOTHELIAL CELLS
REFERENCES
1. Gimbrone MA Jr: Vascular endothelium: Nature’s blood
container, in Gimbrone MA Jr (ed): Vascular Endothelium in
Hemostasis and Thrombosis. Edinburgh, UK, Churchill Livingstone, 1986, p 1
2. Hajjar KA, Harpel PC, JaBe EA, Nachman R L Binding of
plasminogen to cultured human endothelial cells. J Biol Chem
261:11656,1986
3. Miles LA, Levin EG, Plescia J, Collen D, Plow EF: Plasminogen receptors, urokinase receptors, and their modulation on
human endothelial cells. Blood 72628,1988
4. Nawroth PP, Stern D M Endothelial cells as active participant in procoagulant reactions, in Gimbrone MA Jr (ed): Vascular
Endothelium in Hemostasis and Thrombosis. Edinburgh, UK,
Churchill Livingstone, 1986,p 14
5. Rosenberg RD, Rosenberg JS: Natural anticoagulantmechanisms. J Clin Invest 741,1984
6. Miles LA, Fless GM, Levin EG, Scanu AM, Plow E F A
potential basis for the thrombotic risks associated with lipoprotein
(a). Nature 339301,1989
7. Bevilacqua MP, Pober JS, Wheeler ME, Cotran RS, Gimbrone MA Jr: Interleukin 1 acts on cultured human vascular
endothelium to increase the adhesion of polymorphonuclear leukocytes, monocytes, and related leukocyte cell lines. J Clin Invest
76:2003,1985
8. Dejana E, Bertocchi F, Bortolami MC, Regonesi A, Tonta A,
Breviario F, Giavazzi R: Interleukin 1 promotes tumor cell
adhesion to cultured human endothelial cells. J Clin Invest
821466,1988
9. Osborn L Leukocyte adhesion to endothelium in inflammation. Cell 623,1990
10. Zimmerman GA, McIntyre TM, Prescott S M Thrombin
stimulates the adherence of neutrophils to human endothelial cells
in vitro. J Clin Invest 76:2235,1985
11. Geng JG, Bevilacqua MP, Moore KL, McIntyre TM, Prescott SM, Kim JM, Bliss GA, Zimmerman GA, McEver R P Rapid
neutrophil adhesion to activated endothelium mediated by GMP140. Nature 343:757,1990
12. Mantovani A, Dejana E: Cytokines as communication signals between leukocytes and endothelial cells. Immunol Today
10370,1989
13. Pober JS, Cotran RS: Cytokines and endothelial cell biology.
Physiol Rev 70427,1990
14. Muller WA, Gimbrone MA Jr: Plasmalemma1 proteins of
cultured vascular endothelial cells exhibit apical-basal polarity:
Analysis by surface-selectiveiodination. J Cell Biol103:2389, 1986
15. Albelda SM, Buck C A Integrins and other cell adhesion
molecules. FASEB J 4:2868,1990
16. Dejana E, Needham L, Gordon J: Endothelial cell adhesive
interactions: Mechanism and consequences, in Simionescu N,
SimionescuM (eds): Endothelial Cell Dysfunction.New York, NY,
Plenum, 1992, p 153
17. Hynes RO: Integrins: A family of cell surface receptors. Cell
48549,1987
18. Ruoslahti E: Integrins. J Clin Invest 871,1991
19. Burridge K, Fath K, Kelly T, Nuckolls G, Turner C: Focal
adhesions: Transmembrane junctions between the extracellular
matrix and the cytoskeleton. Annu Rev Cell Biol4487,1988
20. Cheresh D A Human endothelial cells synthesize and express an Arg-Gly-Asp-directed adhesion receptor involved in
attachment to fibrinogen and von Willebrand factor. Proc Natl
Acad Sci USA 845471,1987
21. Dejana E, Lampugnani MG, Giorgi M, Gaboli M, Federici
AB, Ruggeri ZM, Marchisio P C von Willebrand factor promotes
endothelial cell adhesion via an Arg-Gly-Aspdependent mechanism. J Cell Biol109367,1989
22. Charo IF, Bekeart LS, Phillips DR: Platelet glycoprotein
IIb-IIIa like proteins mediate endothelial cell attachment to
adhesive proteins and the extracellular matrix. J Biol Chem
2629935,1987
23. Dejana E, Lampugnani MG, Giorgi M, Gaboli M, Marchisio
P C Fibrinogen induces endothelial cell adhesion and spreadingvia
the release of endogenous matrix proteins and the recruitment of
more than one integrin receptor. Blood 75:1509,1990
24. Lawler J, Weinstein R, Hynes R O Cell attachment to
thrombospondin: The role of Arg-Gly-Asp, calcium and integrin
receptors. J Cell Biol107:2351,1988
25. Bar-Shavit R, Sabbah V, Lampugnani MG, Marchisio PC,
Fenton J
W 11, Vlodavsky I, Dejana E An Arg-Gly-Asp sequence
within thrombin promotes endothelial cell adhesion. J Cell Biol
112335,1991
26. Dejana E, Colella S, Conforti G, Abbadini M, Gaboli M,
Marchisio P C Fibronectin and vitronectin regulate the organization of their respective Arg-Gly-Aspadhesion receptors in cultured
human endothelial cells. J Cell Biol107:1215,1988
27. Barbieri B, Balconi G, Dejana E, Donati MB: Evidence that
vascular endothelial cells can induce the retraction of fibrin clots.
Proc SOCExp Biol Med 168:204,1981
28. Furie BM, Cramer EB, Naprstek BL, Silverstein S C Cultured endothelial cell monolayers that restrict the transendothelial
passage of macromolecules and electrical current. J Cell Biol
98:1033,1984
29. Simpson WF, Stulberg CS: A reliable and a rapid procedure
for development of human fibroblastic cells strains. J Tissue
Culture Methods 1:211,1975
30. Conforti G, Zanetti A, Pasquali-Ronchetti I, Quaglino D Jr,
Neyroz P, Dejana E: Modulation of vitronectin receptor binding by
membrane lipid composition. J Biol Chem 265:4011,1990
31. Freed E, Gailit J, van der Geer P, Ruoslahti E, Hunter T A
novel integrin p subunit is associated with the vitronectin receptor
ct subunit (ay) in a human osteosarcoma cell line and is a substrate
for protein kinase C. EMBO J 82955,1989
32. Tokuyasu K T Immunochemistry of ultrathin frozen sections. Histochem J 12381, 1980
33. Tokuyasu K T Applications of cryoultramicrotomyto immunochemistry. J Microsc 143:139,1986
34. Keller GA, Tokuyasu KT, Dutton AH, Singer S J An
improved procedure for immunoelectron microscopy: Ultrathin
plastic embedding of immunolabeled plastic sections. Proc Natl
Acad Sci USA 815744,1984
35. I11 CR, Ruoslahti E: Association of thrombin-antithrombin
I11 complex with vitronectin in serum. J Biol Chem 26015610,1985
36. Lane DA, Flynn AM, Pejler G, Lindahl U, Choay J,
Preissner K T Structural requirements for the neutralization of
heparin-like saccharides by complement S proteinlvitronectin. J
Biol Chem 26216343,1987
37. Preissner K T Specific binding of plasminogen to vitronectin. Evidence for a modulatory role of vitronectin on fibrin(ogen)induced plasmin formation by tissue plasminogen activator. Biochem Biophys Res Commun 168:966,1990
38. Preissner KT, Grulich-Henn J, Ehrlich HJ, Declerck P,
Justus C, Collen D, Pannekoek H, Muller-Berghaus G: Structural
requirements for the extracellular interaction of plasminogen
activator inhibitor 1with endothelial cell matrix-associatedvitronectin. J Biol Chem 265:18490,1990
446
39. Seiffert D, Wagner NN, Loskutoff DJ: Serum-derived vitronectin influences the pericellulardistributionof type 1plasminogen activator inhibitor. J Cell Biol111:1283,1990
40. Dablback B, Podack ER: Characterization of human S
protein, an inhibitor of the membrane attack complex of complement. Demonstration of a free reactive thiol group. Biochemistry
242368,1985
41. Dejana E, Languino LR, Polentarutti N, Balconi G, Ryckewaert JJ, Lameu MJ, Donati MB, Mantovani A, Marguerie G
Interaction between fibrinogen and cultured endothelial cells. J
Clin Invest 7511,1985
42. Languino LR, Colella S, Zanetti A, Andrieux A, Ryckewaert
CONFORTI ET AL
JJ, Charon MH, Marchisio PC, Plow EF, Ginsberg MH, Marguerie
G, Dejana E Fibrinogen-endothelial cell interaction in vitro: A
pathway mediated by an Arg-Gly-Asp recognition specificity.
Blood 73:734,1989
43. Cheung AL, Krishnan M, Jaffe EA, Fischetti V A Fibrinogen acts as a bridging molecule in the adherence of Staphylococcus
aureus to cultured human endothelial cells. J Clin Invest 87:2236,
1991
44. Coller BS, Scudder LE, Beer J, Gold HK, Folts JD,
CavagnaroJ, Jordan R, Wagner C, Iuliucci J, Knight D, Ghrayeb J,
Smith C, Berger H: Monoclonal antibodies to platelet GpIIb/IIIa
as antithromboticagents. Ann NY Acad Sci 614:193,1991