RESEARCH ARTICLE
993
Development 137, 993-1002 (2010) doi:10.1242/dev.045377
© 2010. Published by The Company of Biologists Ltd
Endothelial b1 integrins regulate sprouting and network
formation during vascular development
Daniela Malan1,*, Daniela Wenzel1,*, Annette Schmidt2, Caroline Geisen1, Axel Raible3, Birgit Bölck2,
Bernd K. Fleischmann1,† and Wilhelm Bloch2,†
SUMMARY
b1 integrins are important regulators of vascular differentiation and development, as their endothelial-specific deletion results in
embryonic lethality. In the present study, we investigated the molecular mechanisms underlying the prominent vascular
abnormalities that occur in the absence of b1 integrins. Because of the early embryonic lethality of knockout mice, we studied
endothelial cell and vessel development in b1-integrin-deficient murine embryonic stem cells to gain novel insights into the role of
b1 integrins in vasculo-angiogenesis. We found that vessel development was strongly defective in the mutant embryoid bodies
(EBs), as only primitive and short sprouts developed from clusters of vascular precursors in b1 integrin–/– EBs, whereas complex
network formation of endothelial tubes was observed in wild-type EBs. The vascular defect was due to deficient b1 integrin
expression in endothelial cells, as its endothelial-specific re-expression rescued the phenotype entirely. The mechanism responsible
for defective vessel formation was found to be reduced endothelial cell maturation, migration and elongation. Moreover, the lower
number of endothelial cells in b1 integrin–/– EBs was due to an increased apoptosis versus proliferation rate. The enhanced apoptosis
and proliferation of b1 integrin–/– endothelial cells was related to the elevation of peNOS and pAKT signaling molecules,
respectively. Our data demonstrate that endothelial b1 integrins are determinants of vessel formation and that this effect is
mediated via different signaling pathways.
KEY WORDS: b1 integrin, Embryonic stem cells, Vasculogenesis, Basement membrane, Mouse
1
Institute of Physiology I, University of Bonn, Sigmund-Freud-Str. 25, 53127 Bonn,
Germany. 2German Sport University, Department of Molecular and Cellular Sport
Medicine, Am Sportpark Müngersdorf 6, 50933 Cologne, Germany. 3Department of
Urology, University Medical Center of Cologne, 50924 Cologne, Germany.
*These authors contributed equally to this work
†
Authors for correspondence ([email protected]; [email protected])
Accepted 20 January 2010
particular, b1 integrins have pleiotropic and important functions in
vascular development as they were shown to be necessary for
teratoma growth and vasculogenesis in embryoid bodies (EBs)
(Bloch et al., 1997). These data are supported by the endothelialspecific deletion of b1 integrins causing early embryonic lethality
of mice due to vascular defects (Carlson et al., 2008; Lei et al., 2008;
Tanjore et al., 2008).
To investigate the cellular mechanisms underlying these
abnormalities, we chose the embryonic stem (ES) cell system, which
is an excellent model to examine the cell biological effects of
mutations resulting in early embryonic lethality. We found that b1
integrins regulate key steps of vascular development and network
formation, such as endothelial cell maturation, migration and
elongation. Proliferation and apoptosis were enhanced by activation
of the signaling molecules AKT (AKT1 – Mouse Genome
Informatics) and eNOS (NOS3), respectively. Thus, our study
provides novel mechanistic insight into the role of b1 integrins for
neo-angiogenesis/vascularization.
MATERIALS AND METHODS
Cell culture
The mouse blastocyst-derived ES cell lines D3 (wild-type) and b1 integrin–/–
were maintained in culture as described by Fassler (Fassler et al., 1996). For
EB formation cells were cultured in hanging drops for 2 days. Subsequently,
the aggregates were incubated for 3 days in suspension and plated on day 5
in a 24-well plate on gelatine-coated coverslips. Experiments were
performed after 7, 12, 14 and 18 days, referred to as 5+7, 5+12, 5+14 and
5+18.
Vital microscopy or time-lapse microscopy
In order to live monitor early stages of vasculogenesis, ES cells of the D3 or
b1 integrin–/– line were transfected with a PECAM/EGFP construct; after
differentiation, clusters of ES cell-derived endothelial precursors were
continuously observed for up to 10 hours (Kazemi et al., 2002) with an
inverted microscope (Axiovert 100, Zeiss, Goettingen, Germany) equipped
DEVELOPMENT
INTRODUCTION
Vasculogenesis and angiogenesis are the main processes responsible
for embryonic vascular development. The formation of new vascular
tubes comprises endothelial cell proliferation and migration as well
as the inhibition of apoptosis. Moreover, the surrounding
extracellular matrix (ECM) undergoes degradation and remodeling
during vascular sprouting. The molecules involved in ECM-cell
signaling are the integrin receptors, which are characterized by their
bidirectional (outside-in and inside-out) signaling. Integrins bind to
ECM proteins such as collagen, laminin and fibronectin and thereby
regulate stabilization, migration and the morphology of vascular
structures (Senger et al., 2002; Ruoslahti and Engvall, 1997).
Furthermore, integrins interact with intracellular cytoskeletal
proteins, providing structure and morphology to the cells (Pozzi and
Zent, 2003). Thus, the role of integrins in matrix adhesion and
cytoskeletal reorganization is crucial for cell growth, survival and
migration. These key processes are regulated via several integrindependent cross-linked intracellular pathways involving AKT and
NO (Abair et al., 2008; Basile et al., 2007; Viji et al., 2009).
Moreover, there is also evidence that integrins are involved in the
development of vessels, as deletion of the a5 chain reduced capillary
plexus formation (Francis et al., 2002) and avb3 integrins were
found to be upregulated in angiogenesis (Brooks et al., 1994). In
RESEARCH ARTICLE
with a computer-controlled motorized stage (Merzhaeuser, Nikon,
Germany), and a temperature/CO2-controlled chamber (Buehler, Nikon).
Fluorescence pictures (conventional 100 W halogen lamp) were recorded at
20-minute intervals using a fluorescein isothiocyanate (FITC) filter set as
well as an oil-immersion objective 40⫻ (Plan-Neofluar, Zeiss). Images were
acquired by a color video camera (Sony, DXC 950P, AVT Horn, Tuebingen,
Germany), digitized online via a video frame grabber card (Matrox Corona,
Nikon) and controlled by the Lucia software (Nikon). Fluorescence intensity
was enhanced employing a video-based buffer device (Sony MPU-F100P,
AVT Horn) and the integration mode of the camera.
Magnetic-activated cell sorting
Differentiated ES cells (5+7 days) were dissociated with Accutase. Single
cells were stained with the endothelial-specific antibody rat anti-mouse
PECAM (1:800, Pharmingen, San Diego, CA, USA) and sorted by
magnetic-activated cells (MAC) sorting as previously described (Schmidt
et al., 2004). Briefly, the bead-conjugated secondary antibody, goat anti-rat
IgG, was used as described by the manufacturer (Miltenyi Biotec, Bergisch
Gladbach, Germany). The positive fraction was cultured on gelatine-coated
dishes. Cells were used until passage 8. The purity of cells was tested with
PECAM staining and amounted to 89% (data not shown). For analysis the
number of cells in 40 defined areas was calculated using a 40⫻ objective on
an Axiophot (Zeiss, Germany).
Immunohistochemistry, detection of endothelial cells and vessels
For immunofluorescence, single cells or whole EBs (5+7, 5+12, 5+18 days)
plated on glass coverslips were fixed with 4% paraformaldehyde (PFA),
washed several times with 0.05 M TBS and permeabilized with 0.25%
Triton-X 100 and 0.5 M NH4Cl in 0.05 M TBS. After blocking with 5%
BSA in TBS endothelial cells were stained with the antibody rat anti-mouse
PECAM (1:800). Alternatively, double staining was performed with b1
integrin purified mouse anti-CD29 antibody (1:500; BD Bioscience,
Erembodegem, Belgium), polyclonal rabbit anti-collagen IV (1:500; Acris
Antibodies, Hiddenhausen, Germany), monoclonal mouse anti-fibronectin
(1:500; Sigma-Aldrich, Germany), polyclonal rabbit anti-laminin (1:500,
Sigma-Aldrich, Munich, Germany) or polyclonal rabbit anti-CD34 (Santa
Cruz Biotechnology, Santa Cruz, CA, USA). Next, EBs were stained with
Cy3- (or Cy2-) conjugated goat anti-rabbit (or goat anti-mouse) IgG (1:1500,
and 1:500, Dianova, Hamburg, Germany) or with a biotinylated goat antimouse (or goat anti-rabbit) IgG (1:400, Dako, Glostrup, Denmark) or a
biotinylated anti-rat antibody (Amersham Biosciences Europe, Freiburg,
Germany) followed by streptavidin-conjugated Cy2 (1:200, Amersham) or
Cy3 (1:1500, Amersham). The number of PECAM-positive endothelial
sprouts composed of at least five endothelial cells was counted (Schmidt et
al., 2004). For each clone (wild-type or b1 integrin–/–) at least n=9 complete
EBs were analyzed. For DAB stainings EBs were incubated overnight at 4°C
with the b1 integrin antibody. The secondary antibody was a biotinylated
goat anti-mouse antibody (Dako, Glostrup, Denmark), then EBs were
labeled with streptavidin-conjugated peroxidase (Amersham) and stained
with 3,39-diaminobenzidine as chromogen.
In parallel to the quantification of endothelial sprouts we also classified
PECAM-positive structures in four different EBs at early and late stages. We
determined ‘endothelial precursors’ as single endothelial cell precursors,
‘clusters’ as aggregates of endothelial precursors and ‘vessel-like structures’
as a network of tube-like structures with at least five branches. The whole
EB was screened and results are given as percentage of all PECAM-positive
structures.
For single cell counting, EBs were dissociated with trypsin and cells were
plated on coverslips. After 24 hours cells were fixed and stained with antiPECAM antibody and Hoechst dye to label the nuclei. The relative amount
of PECAM-positive cells was determined in five fields of sight.
Generation of transgenic ES cell clones
A b1 integrin-mRFP fusion plasmid containing the murine b1 integrin
cDNA (2.5 kb EcoRI-AgeI fragment) in frame with mRFP was kindly
provided by R. Fässler. By cutting the construct with SalI and AgeI, the b1
integrin cDNA was excised and subsequently cloned into the XhoI and AgeI
sites of pEGFP-1 (BD-Biosciences-Clontech, Heidelberg, Germany)
containing EGFP under the control of the human PECAM-1 promoter
Development 137 (6)
[–4500/+65 PstI fragment in pEGFP-1 (Kazemi et al., 2002)]. The resulting
PECAM-1 promoter/b1 integrin-EGFP fusion construct was used for
electroporation of b1 integrin–/– ES cells. Of the construct, 36 mg was
linearized by digestion with HindIII restrictase and used for coelectroporation (240 V/500 mF, Bio-Rad Gene Pulser) with 4 mg of the
pPWL512 plasmid containing the hygromycin resistance gene under the
control of the PGK promoter in 5⫻106 ES cells. Selection was performed
by addition of 1000 mg/ml hygromycin.
Tube-forming assay
The spontaneous formation of vessel-like structures by either wild-type or
b1 integrin–/– MAC-sorted endothelial cells was performed on a basement
membrane matrix preparation (Matrigel, BD Bioscience, Heidelberg,
Germany). Plates were coated with Matrigel (10 mg/ml) according to the
manufacturer’s instructions. Tube formation was monitored during a period
of 24 hours using the Axio Observer apparatus (Zeiss, Germany).
Plastic embedding and electron microscopy
EBs were fixed with 4% PFA, rinsed in cacodylate buffer three times and
then treated with 1% uranyl acetate in 70% ethanol for 8 hours to enhance
the contrast. The specimens were subsequently dehydrated in a graded series
of ethanol and then embedded in Araldite (Serva, Heidelberg, Germany).
Semithin sections (500 mm) were cut with a glass knife on an ultramicrotome
(Reichert, Bensheim, Germany) and stained with Methylene Blue. Ultrathin
sections (30 to 60 nm) for electron microscopy were processed on the same
microtome with a diamond knife and placed on copper grids. Transmission
electron microscopy was performed using a 902A electron microscope from
Zeiss.
Morphological analysis and confocal microscopy
The distribution pattern of extracellular matrix proteins was analyzed by
confocal microscopy using the LSM 510 META Zeiss microscope (Zeiss
Microimaging, Goettingen, Germany). Images were acquired in the multitrack mode, collected at 512⫻512 pixels. Cy3 was excited at 488 nm (AI
laser), Cy2 at 543 nm (HeNe laser). After the main dichroic beam splitter
(HFT 488/543 for Cy3 and Cy2), the fluorescence signal was divided by a
secondary dichroic beam splitter (NFT 543) and detected in the separate
channels using the appropriate filters: BP 505-530 (for Cy2), LP560 (for
Cy3). In this set-up, no crosstalk between the green and red channels was
observed. Pictures of whole EBs were acquired with an Axiovert 200 (Zeiss)
and assembled by the mosaix function. Automated area analysis of whole
EBs was performed by the AutMess program of the AxioVision 4.6 software
(Zeiss).
Western blot
For western blotting, samples were lysed in SDS sample buffer and
submitted to SDS-PAGE on a 4-15% gradient TRIS-HCl gel (Criterion, BioRad Laboratories, Germany). Proteins were transferred to a polyvinylidene
difluoride (PVDF) membrane and incubated with the antibodies described
above. Immunoreactive proteins were detected by the enhanced
chemiluminescence detection system (Amersham Biosciences Europe).
Phosphorylated proteins were detected by rabbit anti-pAKT (Ser473)
(1:1000, Cell Signaling, Danvers, MA, USA) and rabbit anti-peNOS
(Ser1177) (1:500, Upstate, Hamburg, Germany). These were compared to
AKT (1:500, Biomol, ENZO LIFE SCIENCES, Lörrach, Germany) and
eNOS (1:500, Millipore, Billerica, MA, USA) expression. For the detection
of b1 integrins a polyclonal antibody (1:500, Abcam, Cambridge, UK) was
applied. As secondary antibody a polyclonal anti-rabbit antibody was used
(1:500, Biomol, Hamburg, Germany).
Proliferation assay
EBs from wild-type and b1 integrin–/– clones were fixed with 4% PFA and
used for immunohistochemistry. After identifying the endothelial cells by
PECAM staining, proliferating cells were detected with the primary
antibody rabbit anti-mouse-Ki67 (1:150, pAb, Dianova, Hamburg,
Germany); as secondary antibody goat anti-rabbit CY2 (1:200, Dianova,
Hamburg, Germany) was used. To determine the number of proliferating
endothelial cells Ki67-positive cells in 50 randomly chosen vessel-like tubes
were analyzed (Muller-Ehmsen et al., 2006).
DEVELOPMENT
994
b1 integrins in vascular development
RESEARCH ARTICLE
995
Fig. 1. Formation of vascular structures in b1 integrin–/– and wild-type embryoid bodies. (A-C) PECAM-positive tubes (red) form clusters of
endothelial precursors (arrow, B) and network-like structures in wild-type embryoid bodies (EBs). Arrowheads indicate endothelial tubes (D-F) In b1
integrin–/– EBs, only short sprouts and primitive endothelial tubes were detected. (G) Quantitative analysis of vascular structures at different
developmental stages: day 7 (early stage), days 12 and 18 (late stage). (H) Distribution of PECAM-positive structures in EBs. (I,J) PECAM-positive
(red) structures of wild-type and b1 integrin–/– whole EBs (5+14 days). Insets show areas determined by automated area analysis. (K) Relative area
covered by PECAM-positive cells in EBs. (L) Relative proportion of PECAM-positive cells in EBs. *P<0.05, Student’s t-test. Scale bars: 5 mm in A,B,D,E;
1 mm in I,J.
Wild-type and b1 integrin–/– EBs were fixed and stained with PECAM
antibody as described above. Apoptotic cells were detected with a rabbit
anti-PARP p85 fragment antibody (1:250 Promega, Madison, WI, USA); as
secondary antibody goat anti-rabbit Cy2 was used. To evaluate the number
of apoptotic endothelial cells the number of PARP-cleavage-positive cells
within 50 randomly chosen vessel-like tubes was counted.
Migration assay
Migration-assay was performed in a modified Boyden chamber by using a 24well HTS Fluoro Blok insert system (Falcon Becton Dickinson, Heidelberg,
Germany). The inserts of the chamber consist of a polyethylene membrane
with 8.0 mm pores blocking >99% of the transmission light (490-700 nm).
Inserts were coated with 0.1% gelatine solution for 1 hour. EBs (5+7),
dissociated with Accutase (PAA Laboratories, Linz, Austria), were MAC
sorted with a PECAM antibody. Thereafter, cells were put on top of an insert
and incubated for 8 hours in DMEM with 15% FCS. Cells were counted after
8 hours of migration to exclude proliferation as potential mechanism (Schmidt
et al., 2004). For analysis, the membrane was cut out of the insert, cells were
stained with anti-PECAM antibody and covered in DAPI mounting medium
(4⬘,6-diamidino-2-phenylindole; Vectashield, Vector Laboratories,
Burlingame, CA, USA). The number of migrating PECAM-positive cells was
counted. Because the rescue clone did not tolerate MAC sorting, whole EBs
were dissociated and the number of PECAM-positive cells was determined
before and after migration. Before migration, the amount of PECAM-positive
cells was equal in wild type (8100±220, n=3), b1 integrin–/– (7520±440, n=3)
and the rescue clone (7364±640, n=3, P>0.05); therefore we directly compared
absolute values after migration.
Inhibitor treatment
In the case of EB treatment with inhibitors, the PI3K blocker wortmannin
(25 nM) was applied for 30 minutes, and the eNOS inhibitor L-NAME (100
mM) for 12 hours.
Statistical analysis
All data are presented as mean±s.e.m. Data analysis was performed using
analysis of variance with Bonferroni post-hoc test and/or Student’s t-test for
unpaired data. P<0.05 was considered to be significant.
DEVELOPMENT
Apoptosis assay
996
RESEARCH ARTICLE
Development 137 (6)
Fig. 2. Time-lapse microscopy of ES cellderived endothelial cells expressing EGFP
under the control of the PECAM promoter.
(A-C) In wild-type EBs, clusters of sprouting
endothelial precursors (arrow) were detected
(observation time, 8 hours). (D-F) By contrast,
endothelial precursors derived from b1 integrin–/–
cells were not able to form clusters and sprouts
during similar observation periods. Scale bars:
15 mm in A; 10 mm in D.
PECAM and CD34-positive/PECAM-positive cells: 50.0±2.2%,
n=7, b1 integrin–/– versus 24.3±2.0% n=8, wild-type, P<0.001).
Moreover, the area of EBs covered by endothelial cells was equal at
the early stage (5+7) but declined at the late stage (5+14) in b1
integrin–/– EBs (Fig. 1I-K). These results were corroborated by
endothelial cell counting after single-cell dissociation (Fig. 1L).
Thus, in b1 integrin–/– EBs there is a high proportion of early-stage
endothelial cells, and this clearly shows that b1 integrin expression
is a prerequisite for endothelial cell maturation and vascular tube
formation. As these defects persisted in late-stage EBs and the
number of endothelial cells declined, the observed phenotype is not
due to a mere delay in cell maturation.
b1-integrin-deficient EBs display morphological
defects in endothelial tube formation
To directly monitor and determine the dynamics of endothelial cell
development in b1 integrin–/– EBs, we established stable transgenic
wild-type and b1 integrin–/– ES cell lines, in which the endothelialspecific promoter PECAM drives the expression of EGFP (Kazemi
et al., 2002). Time-lapse microscopy revealed that sprouting
endothelial tubes developed from endothelial cell clusters within 8
hours in wild-type EBs by elongation and migration of endothelial
cells (Fig. 2A-C, arrow). By contrast, the endothelial cells in the b1
integrin–/– EBs remained round and displayed less migration,
resulting in a lack of tube formation and branching during the same
period of observation (Fig. 2D-F) (n=4). These findings are in full
agreement with earlier reports in vivo, and our results in EBs where
the amount of vascular tubes was strongly reduced (see above), and
indicate that the deletion of b1 integrin inhibits the development of
endothelial tubes in the early phases of vascular development
(Carlson et al., 2008; Lei et al., 2008; Tanjore et al., 2008).
b1 integrins regulate ECM formation
Integrins are important receptors for the interaction of endothelial
cells with the ECM, which is a prerequisite for adhesion, migration
and differentiation of early endothelial cells. Our ultrastructural
analysis revealed an intact basal membrane (BM) around endothelial
cells in wild-type EBs, whereas only fragments of BM in b1
integrin–/– EBs were found (arrows, Fig. 3A). Next, we examined in
detail the deposition of BM proteins in wild-type and b1 integrin–/–
DEVELOPMENT
RESULTS
Lack of b1 integrin impairs vessel formation and
inhibits maturation of endothelial cells during
development
To gain mechanistic insight into the role of b1 integrins in embryonic
vessel development, the endothelial cells and the formation of tubelike structures in wild-type and b1 integrin–/– EBs were analyzed and
quantified at different time points. Wild-type EBs showed clusters of
endothelial precursors (arrow) as well as complex networks of
endothelial tubes at early (5+7) and late (5+14) developmental stages
(arrowhead) (Fig. 1A-C). By contrast, in the b1 integrin–/– EBs only
primitive and short sprouts were detected at both stages (Fig. 1D-F).
Although the total amount of endothelial cells per EB was not different
at day 5+7 (see below), the number of endothelial tubes in the b1
integrin–/– EBs was strongly reduced compared with wild-type EBs at
all differentiation stages (Fig. 1G). To analyze in detail the differences
in vascular development, we distinguished three developmental stages
of endothelial structures in EBs: (1) single endothelial precursor cells;
(2) clusters of endothelial precursors; and (3) vessel-like structures
(see also Materials and methods). In early stage (5+7) wild-type EBs
(n=4), 13% of PECAM-positive cells were single endothelial
precursors, 36% were arranged in clusters and 51% were in vessellike structures, whereas in b1 integrin–/– EBs (n=4) 81% were single
endothelial precursors, 15% were found in clusters and 4% were in
vessel-like structures. In late stage (5+14) wild-type EBs (n=4), a low
percentage of endothelial precursors (4%) and similar amounts of
clusters (47%) and vessel-like structures (49%) were detected,
whereas in b1 integrin–/– EBs (n=4), high numbers (87%) of
endothelial precursors persisted and again only low numbers of more
differentiated forms (clusters 13%, no vessel-like structures) could be
found. These results clearly demonstrate that the formation of vessellike structures from endothelial precursors is impaired at all stages in
the absence of b1 integrins.
The increased percentage of endothelial precursors in b1
integrin–/– EBs suggested a defect in endothelial cell differentiation.
Therefore we investigated in more detail the maturation stage of the
MAC-sorted endothelial cells by co-staining with PECAM and
CD34, a marker for cells of an early maturation stage. We found a
strong increase of the PECAM-positive/CD34-positive cells in the
b1 integrin–/– compared with wild-type endothelial cells (ratio
Fig. 3. Basal membrane deposition and ECM formation in b1
integrin–/– EBs. (A) Ultrastructural analysis of basal membrane (BM)
formation in wild-type and b1 integrin–/– endothelial cells. Wild-type
endothelial cells displayed an intact BM (arrows), whereas in b1
integrin–/– cells only fragments of BM were observed (arrows).
(B-D) Collagen IV, laminin and fibronectin (all green) were deposited
around PECAM-positive (red) structures in wild-type EBs (left). In b1
integrin–/– EBs (right) no collagen IV or laminin was found, whereas
fibronectin was deposited throughout the EB in a punctate manner.
Scale bars: 100 nm in A; 40 mm in B.
EBs at 5+7 days. Collagen IV, laminin and fibronectin (all green)
were found to be highly expressed around PECAM-positive tubelike structures (red) in wild-type EBs (Fig. 3B,C,D left panel).
However, in b1 integrin–/– EBs, collagen IV and laminin were
completely absent (Fig. 3B,C middle panel), whereas fibronectin
deposition was found throughout the whole EB but not around the
vascular structures (Fig. 3D). Thus, b1 integrins in endothelial cells
are a prerequisite for ECM deposition around vascular structures.
Impaired vascular tube formation is due to the
lack of b1 integrins in endothelial cells
We next determined whether the defect of vasculo-angiogenesis in
b1 integrin–/– EBs was due to b1 integrin deficiency in endothelial
cells or secondary to a lack of b1 integrins in surrounding cells. For
RESEARCH ARTICLE
997
this purpose we used MAC-sorted endothelial cells and transferred
them to Matrigel. Whereas wild-type endothelial cells formed
aggregates and tube-like structures (Fig. 4A-C), b1 integrin–/–
endothelial cells did not assemble or develop any tube-like structures
(Fig. 4D-F). These data suggest that the lack of expression of b1integrins in endothelial cells underlies the massive cell biological
defects in tube formation and that this defect cannot be compensated
by an intact ECM.
To further corroborate this crucial point, we designed a rescue
approach and generated stable transgenic ES cell lines in which b1
integrin was re-expressed as fusion protein with EGFP under
control of the endothelial-specific PECAM promoter. In these
rescue EBs we found EGFP-positive clusters of endothelial cells
(Fig. 5A), which further developed into sprouting endothelial tubes
(Fig. 5B), recapitulating the qualitative and quantitative features
of vessel formation in wild-type EBs. In fact, large and elongated
tube-like structures were found in these rescued EBs; note that
EGFP fluorescence was decreased in more differentiated vessel
structures, most likely because of the lower rate of protein
synthesis (Fig. 5C). We also assessed b1 integrin expression with
DAB immunostaining and found specific re-expression in vessellike structures of rescue EBs (Fig. 5G,H), whereas no signal was
found in b1 integrin–/– EBs (data not shown). In addition, we
investigated the subcellular distribution of b1 integrins in
endothelial cells with double immunofluorescence and detected
their preferential membrane localization in wild-type and rescue
cells (Fig. 5D-F). Western blot experiments revealed that b1
integrin protein expression was completely absent in b1 integrin–/–
EBs and partly restored in the rescue clone (Fig. 5I). The number
of vessel-like structures in the rescue clones at day 5+7 was similar
to that found in wild-type EBs (Fig. 5J); also the distribution of
endothelial precursors, clusters and vessel-like structures was very
similar to wild-type EBs: 15, 27 and 58% (n=4), respectively (Fig.
5K).
In the rescue clones collagen IV, laminin and fibronectin
deposition around vascular structures was restored, indicating that
the b1 integrin expression in endothelial cells is sufficient for the
BM assembly around PECAM-positive tube-like structures (Fig.
5L,M,N). Conversely, the Matrigel experiments demonstrate that an
intact matrix surrounding endothelial cells is not sufficient for
vascular network formation in the absence of b1 integrins in
endothelial cells.
Thus, the reduced number of vessel-like structures, the defective
sprouting and the lack of ECM protein deposition around endothelial
cells in b1 integrin–/– EBs is clearly due to the absence of endothelial
b1 integrins and independent from the b1 integrin expression in
surrounding cells.
Lack of endothelial b1 integrins enhances
proliferation and apoptosis but reduces migration
in endothelial cells
Vascular development strongly relies on proliferation, apoptosis
and migration of endothelial cells (Kalluri, 2003; Davis and
Senger, 2005). Co-staining with antibodies against the
proliferation marker Ki67 and the endothelial cell marker
PECAM revealed a significant increase in the proliferation rate of
endothelial cells in b1 integrin–/– EBs (Fig. 6A; see also Fig. 6D
for immunohistochemistry). As the high proliferation rate does
not match with the lower numbers of endothelial cells in latestage EBs in b1 integrin–/– EBs, we also determined the rate of
apoptosis. Double staining with antibodies against the apoptosis
marker Poly(ADP-Ribose)-polymerase cleavage fragment (PARP,
DEVELOPMENT
b1 integrins in vascular development
998
RESEARCH ARTICLE
Development 137 (6)
Fig. 4. Tube formation assay in Matrigel. (AC) MAC-sorted PECAM-positive cells of wild-type
EBs built capillary-like structures in Matrigel
within 7 hours. (D-F) MAC-sorted PECAMpositive cells of b1 integrin–/– EBs did not form
any vascular sprouts. Scale bars: 100 mm.
Activation of PI3K/AKT is responsible for increased
proliferation, whereas the induction of eNOS/NO
promotes apoptosis in b1 integrin–/– endothelial
cells
The PI3K/AKT/eNOS signaling pathway is central for endothelial
cell biology. PI3K/AKT has been reported to increase endothelial
cell proliferation and migration and to inhibit apoptosis (Ackah et
al., 2005; Deregibus et al., 2003; Radisavljevic et al., 2000). In
addition, NO is known to induce in a dose-dependent manner
proliferation and apoptosis in endothelial cells (Papapetropoulos et
al., 1997; Shen et al., 1998). We have examined the expression of
these key signaling molecules using western blot analysis and found
an increase in the activated pAKTSer473 and peNOSSer1177 in b1
integrin–/– EBs (Fig. 7A) when compared with wild-type EBs; in the
rescue clone, that increase was reverted. Thus, the deletion of b1
integrins in endothelial cells increases the amount of pAKTSer473
and peNOSSer1177. Interestingly, the expression of pERK1/2, a
molecule that is known to promote FGF-dependent migration
(Pintucci et al., 2002), was found to be reduced in b1 integrin–/– EBs,
whereas its expression returned to normal in the rescue clone (data
not shown).
To investigate the relevance of the of PI3K/AKT and eNOS/NO
pathways in wild-type and b1 integrin–/– endothelial cells, we
determined proliferation, apoptosis and migration in the presence of
the PI3K blocker wortmannin (WN, 25 nM) or the eNOS inhibitor
L-NAME (100 mM). We also tried LY294002, another blocker of the
PI3K, but this substance had toxic effects and could therefore not be
used.
Our data reveal that the proliferation rate was independent from
PI3K/AKT in wild-type EBs, as blockade with wortmannin showed
no effect. However, in b1 integrin–/– EBs wortmannin led to a
significant decrease of proliferation rates in endothelial cells (Fig.
7B). Moreover, proliferation turned out to be independent from
eNOS/NO, as L-NAME had no effect in both wild-type cells and b1
integrin–/– cells (Fig. 7C).
The rate of apoptosis was not affected by the activation of the
PI3K/AKT pathway in wild-type and b1 integrin–/– endothelial cells
(Fig. 7D). In wild-type endothelial cells there was also no influence
of eNOS/NO on. By contrast, inhibition of eNOS/NO decreased
apoptosis in b1 integrin–/– endothelial cells (Fig. 7E).
Finally, there was no evidence for a contribution of the
PI3K/AKT or eNOS/NO pathway to decreased migration rates in b1
integrin KO endothelial cells.
These results illustrate that high proliferation levels are induced
by PI3K/AKT activation and that high apoptosis rates are a result of
eNOS/NO increase in b1 integrin–/– endothelial cells.
DISCUSSION
In the present study we demonstrate a central role for b1 integrins
in embryonic vasculogenesis and angiogenesis. We have found
intact endothelial cells in b1 integrin–/– EBs at different time
points during development; however, their sprouting and vascular
network formation was strongly reduced. Even though the number
of PECAM-positive cells is unaltered at the early stage in b1
integrin–/– EBs, the increased amount of CD34/PECAM-positive
endothelial cells and the reduction of vessel-like structures
suggest that the endothelial cells arrest at earlier developmental
stages and cannot form/maintain vessel tubes, which indicates
impaired endothelial cell differentiation and defective blood
vessel formation. These novel findings are in full agreement with
the strongly defective vascularization reported for b1 integrin–/–
DEVELOPMENT
p85 fragment) and PECAM also showed a strong increase
of apoptotic nuclei in endothelial cells per 50 vascular tubes
in the b1 integrin–/– EBs (Fig. 6B; see also Fig. 6E for
immunohistochemistry). At late stage the ratio of apoptotic/
proliferating cells was 2.38 times higher in b1 integrin–/–
endothelial cells compared with wild type, clearly explaining their
reduction in b1 integrin–/– EBs over time. The high levels of
proliferation and apoptosis in b1 integrin–/– EBs were reverted to
wild-type levels in the rescue clones (Fig. 6F,G).
As migration of endothelial cells is crucial for sprouting in
vasculogenesis and angiogenesis (Schmidt et al., 2004), we
determined this parameter in a modified Boyden chamber. We
found decreased migration of PECAM-positive MAC-sorted b1
integrin–/– endothelial cells (Fig. 6C). The rescue clones showed
a recovered migratory function (Fig. 6H). Similar results were
obtained at 5+14 days. Thus, the lack of b1 integrins in
endothelial cells decreases cell survival and migration of the
developing vasculature.
b1 integrins in vascular development
RESEARCH ARTICLE
999
embryos (Carlson et al., 2008; Lei et al., 2008; Tanjore et al.,
2008). In contrast to these earlier in vivo studies, our approach
using ES cells allowed us to investigate the underlying
mechanism by studying endothelial cell differentiation and vessel
formation at different time points using a variety of cell biological
assays.
We have identified several crucial defects of vascular
development – reduced migration and elongation and enhanced
apoptosis/proliferation ratio of endothelial cells, as well as
defective ECM protein deposition – in b1 integrin–/– endothelial
cells that mechanistically explain the defective vascularization. An
important finding was that these defects were entirely dependent on
the lack of endothelial b1 integrins, as our rescue approach restored
the wild-type phenotype. Migration, elongation and proliferation of
endothelial cells are crucial for vascular sprouting and branching,
and this process is known to rely on the contact of the cells with
ECM molecules via integrins (Davis and Senger, 2005; Qutub and
Popel, 2009; Senger et al., 2002). Accordingly, Boyden chamber
assay and time-lapse microscopy revealed a strong reduction of
migration and a lack of elongation of endothelial cells. This result
in combination with the enhanced apoptosis/proliferation ratio is
most likely responsible for defective sprouting and tube formation
in b1 integrin–/– EBs. Enhanced proliferation rates are also
supported by an earlier study in which b1 integrin deletion
increased proliferation in smooth muscle cells (Abraham et al.,
2008). Similarly, apoptosis rates are expected to increase in
endothelial cells because of the lack of adhesion to the matrix
(Frisch and Francis, 1994).
DEVELOPMENT
Fig. 5. Endothelial-specific b1 integrin re-expression in b1 integrin–/– EBs. (A,B) b1 integrin–/– EBs transfected with PECAM/b1 integrin-EGFP
showed (A) formation of endothelial precursor cell clusters and (B) elongated tube-like structures. (C) PECAM-positive networks (red) formed,
whereas EGFP expression (green) declined over time. (D-F) Immunofluorescence staining of b1 integrins in PECAM-positive dissociated single
endothelial cells. (G,H) DAB immunostaining revealed b1 integrin expression in tube-like structures of rescue EBs. (I) Western blot analysis of b1
integrin expression. A quantitative analysis (right) of the blot (left) is shown. *P<0.05, ANOVA. (J) Quantitative analysis of vascular structures in EBs
in the rescue clone. (K) Distribution of PECAM-positive structures in the rescue clone. (L-N) Redistribution of the ECM proteins collagen IV, laminin
and fibronectin (all green) around PECAM-positive (red) sprouts (arrows). Scale bars: 10 mm in D; 15 mm in H; 40 mm in L,M; 25 mm in N.
1000 RESEARCH ARTICLE
Development 137 (6)
In contrast to our data, Carlson et al. found no alterations in the
proliferation rate of endothelial cells after b1 integrin deletion
(Carlson et al., 2008). These differences could be explained by
compensatory effects occurring in the in vivo situation. At the early
stage we found no difference in the amount of endothelial cells in
b1 integrin–/– and wild-type endothelial cells. As in late-stage b1
integrin–/– PECAM-positive cells the balance of proliferation versus
apoptosis shifts to the latter, the overall number of endothelial cells
declined, which explains their strong reduction at day 5+14.
Besides the reduced total amount of PECAM-positive cells,
vascular network formation was strongly impaired at all
differentiation stages in the b1 integrin–/– EBs, and this is most
likely due to the strongly altered migration rates, which prevent
sprouting and tube formation. As high relative numbers of
endothelial precursors and low relative numbers of vessel-like
structures persist at all differentiation stages in the b1 integrin–/–
EBs, this is not due to a delay in vascular development but
indicative of a general maturation defect. This finding is further
supported by the increased amount of CD34-positive endothelial
cells in b1 integrin–/– EBs; these data indicate that endothelial cell
assembly and maturation are strongly disturbed. Our data also
provide a novel mechanistic explanation for altered survival of
endothelial cells in b1 integrin–/– EBs. We have identified increased
PI3K/AKT activation as underlying the enhanced proliferation of
endothelial cells and augmenting eNOS/NO activity to cause the
elevated apoptosis. The PI3K/AKT cascade has been shown in
earlier studies to be an important regulator of proliferation
(Madeddu et al., 2008; Steinle et al., 2002), whereas apoptosis is
well known to be promoted by high levels of NO (Lopez-Farre et
al., 1998). The fact that inhibitors of both molecules had no impact
on proliferation and apoptosis rates of wild-type endothelial cells
demonstrates that these molecules determine endothelial growth
and survival in our system only at the high expression levels found
in the knockout. Whether defective sprouting and vascular
development is purely a result of the above reported alterations of
signaling pathways and/or also because of the lack of adhesion to
ECM molecules remains to be investigated in future. Previous
studies reported that the absence of b1 integrins leads to defects in
BM assembly (Aumailley et al., 2000; Li and Yurchenco, 2006;
Lohikangas et al., 2001). In particular, the lack of fibronectin, a
major ligand at b1a5 integrins, has been implicated in severe
vascular defects (Yang et al., 1993). In accordance with those
findings we demonstrate that lack of endothelial b1 integrins results
in the complete absence of such important matrix proteins as
collagen IV, laminin and fibronectin around endothelial cells. This
clearly suggests that the deposition of these proteins is b1 integrin
dependent. In addition, our Matrigel experiments also show that
even intact ECM is not sufficient for the formation of vascular
networks from b1 integrin–/– endothelial cells but that adhesion to
the ECM via b1 integrins is crucial for endothelial tube formation.
DEVELOPMENT
Fig. 6. Proliferation, apoptosis
and migration of ES cell-derived
endothelial cells. (A) Proliferation
of b1 integrin–/– endothelial cells was
strongly increased compared with
wild-type cells at all stages
investigated. (B) The rate of
apoptosis in b1 integrin–/–
endothelial cells was significantly
enhanced compared with wild-type
cells at the early stage (5+7 days)
and at the late stage (5+14 days).
(C) Migration of b1 integrin–/–
endothelial cells was reduced
compared with wild-type cells.
(D) Proliferation was quantified by
double immunohistochemistry for
Ki67 (green) and PECAM (red) in
wild-type (top) and b1 integrin–/–
(bottom) EBs. (E) Apoptosis was
quantified by co-staining for PARP
cleavage protein (green) and PECAM
(red) in wild-type (top) and b1
integrin–/– (bottom) EBs.
(F-H) Comparison of proliferation,
apoptosis and migration in rescue,
wild-type and b1 integrin–/– EBs.
*P<0.05, Student’s t-test (A-C),
ANOVA (F-H). Scale bars: in D,E,
20 mm upper panels, 40 mm lower
panels.
b1 integrins in vascular development
RESEARCH ARTICLE 1001
Taken together, our current study provides a novel mechanistic
explanation for defective vascularization after b1 integrin
deletion. We found that the lack of endothelial b1 integrins
inhibited endothelial cell maturation, migration and sprouting,
which resulted in defective vascular network formation.
Moreover, the lack of b1 integrins induced endothelial cell
proliferation and apoptosis via PI3K/AKT and eNOS/NO
activation, respectively. This led to an enhanced apoptosis/
proliferation ratio, explaining the decreased number of
endothelial cells during the differentiation. These new insights
into the mechanism of how b1 integrins promote vascularization
could be also exploited for therapeutic purposes to re-vascularize
ischemic tissues or to inhibit tumor angiogenesis.
Acknowledgements
We thank S. Grünberg, A. Voss and M. Ghilav for excellent technical assistance
and R. Fässler for providing the b1 integrin–/– ES cells. This work was funded by
DFG grant 1086 419/1-1/2BL.
Competing interests statement
The authors declare no competing financial interests.
References
Abair, T. D., Bulus, N., Borza, C., Sundaramoorthy, M., Zent, R. and Pozzi,A.
(2008). Functional analysis of the cytoplasmic domain of the integrin {alpha}1
subunit in endothelial cells. Blood 112, 3242-3254.
Abraham, S., Kogata, N., Fassler, R. and Adams, R. H. (2008). Integrin beta1
subunit controls mural cell adhesion, spreading, and blood vessel wall stability.
Circ. Res. 102, 562-570.
Ackah, E., Yu, J., Zoellner, S., Iwakiri, Y., Skurk, C., Shibata, R., Ouchi, N.,
Easton, R. M., Galasso, G., Birnbaum, M. J. et al. (2005). Akt1/protein kinase
Balpha is critical for ischemic and VEGF-mediated angiogenesis. J. Clin. Invest.
115, 2119-2127.
Aumailley, M., Pesch, M., Tunggal, L., Gaill, F. and Fassler, R. (2000). Altered
synthesis of laminin 1 and absence of basement membrane component
deposition in (beta)1 integrin-deficient embryoid bodies. J. Cell Sci. 113, 259268.
Basile, J. R., Gavard, J. and Gutkind, J. S. (2007). Plexin-B1 utilizes RhoA and
Rho kinase to promote the integrin-dependent activation of Akt and ERK and
endothelial cell motility. J. Biol. Chem. 282, 34888-34895.
Bloch, W., Forsberg, E., Lentini, S., Brakebusch, C., Martin, K., Krell, H. W.,
Weidle, U. H., Addicks, K. and Fassler, R. (1997). Beta 1 integrin is essential
for teratoma growth and angiogenesis. J. Cell Biol. 139, 265-278.
Brooks, P. C., Clark, R. A. and Cheresh, D. A. (1994). Requirement of vascular
integrin alpha v beta 3 for angiogenesis. Science 264, 569-571.
Carlson, T. R., Hu, H., Braren, R., Kim, Y. H. and Wang, R. A. (2008). Cellautonomous requirement for beta1 integrin in endothelial cell adhesion,
migration and survival during angiogenesis in mice. Development 135, 21932202.
Davis, G. E. and Senger, D. R. (2005). Endothelial extracellular matrix:
biosynthesis, remodeling, and functions during vascular morphogenesis and
neovessel stabilization. Circ. Res. 97, 1093-1107.
Deregibus, M. C., Buttiglieri, S., Russo, S., Bussolati, B. and Camussi, G.
(2003). CD40-dependent activation of phosphatidylinositol 3-kinase/Akt
pathway mediates endothelial cell survival and in vitro angiogenesis. J. Biol.
Chem. 278, 18008-18014.
Fassler, R., Rohwedel, J., Maltsev, V., Bloch, W., Lentini, S., Guan, K.,
Gullberg, D., Hescheler, J., Addicks, K. and Wobus, A. M. (1996).
Differentiation and integrity of cardiac muscle cells are impaired in the absence
of beta 1 integrin. J. Cell Sci. 109, 2989-2999.
Francis, S. E., Goh, K. L., Hodivala-Dilke, K., Bader, B. L., Stark, M., Davidson,
D. and Hynes, R. O. (2002). Central roles of alpha5beta1 integrin and
fibronectin in vascular development in mouse embryos and embryoid bodies.
Arterioscler. Thromb. Vasc. Biol. 22, 927-933.
Frisch, S. M. and Francis, H. (1994). Disruption of epithelial cell-matrix
interactions induces apoptosis. J. Cell Biol. 124, 619-626.
Kalluri, R. (2003). Basement membranes: structure, assembly and role in tumour
angiogenesis. Nat. Rev. Cancer 3, 422-433.
Kazemi, S., Wenzel, D., Kolossov, E., Lenka, N., Raible, A., Sasse, P., Hescheler,
J., Addicks, K., Fleischmann, B. K. and Bloch, W. (2002). Differential role of
bFGF and VEGF for vasculogenesis. Cell Physiol. Biochem. 12, 55-62.
Lei, L., Liu, D., Huang, Y., Jovin, I., Shai, S. Y., Kyriakides, T., Ross, R. S. and
Giordano, F. J. (2008). Endothelial expression of beta1 integrin is required for
embryonic vascular patterning and postnatal vascular remodeling. Mol. Cell.
Biol. 28, 794-802.
Li, S. and Yurchenco, P. D. (2006). Matrix assembly, cell polarization, and cell
survival: analysis of peri-implantation development with cultured embryonic
stem cells. Methods Mol. Biol. 329, 113-125.
DEVELOPMENT
Fig. 7. Protein expression of pAKT, peNOS and
proliferation, apoptosis and migration after
blockade of the PI3K/pAKT and peNOS/NO
pathways. (A) PeNOSSer1177 and pAKTSer473
expression were increased in b1 integrin–/– compared
with wild-type EBs. After re-expression of b1 integrin
in the rescue clone these alterations were reverted to
wild-type levels. (B,C) Proliferation was decreased by
wortmannin in b1 integrin–/– endothelial cells (B),
whereas L-NAME was without effect (C).
(D,E) Apoptosis was decreased by L-NAME in b1
integrin–/– endothelial cells, whereas wortmannin was
without effect, *P<0.05, ANOVA.
Lohikangas, L., Gullberg, D. and Johansson, S. (2001). Assembly of laminin
polymers is dependent on beta1-integrins. Exp. Cell Res. 265, 135-144.
Lopez-Farre, A., Rodriguez-Feo, J. A., Sanchez, de M. L., Rico, L. and Casado,
S. (1998). Role of nitric oxide in the control of apoptosis in the microvasculature.
Int. J. Biochem. Cell Biol. 30, 1095-1106.
Madeddu, P., Kraenkel, N., Barcelos, L. S., Siragusa, M., Campagnolo, P.,
Oikawa, A., Caporali, A., Herman, A., Azzolino, O., Barberis, L. et al.
(2008). Phosphoinositide 3-kinase gamma gene knockout impairs postischemic
neovascularization and endothelial progenitor cell functions. Arterioscler.
Thromb. Vasc. Biol. 28, 68-76.
Muller-Ehmsen, J., Schmidt, A., Krausgrill, B., Schwinger, R. H. and Bloch, W.
(2006). Role of erythropoietin for angiogenesis and vasculogenesis: from
embryonic development through adulthood. Am. J. Physiol. Heart Circ. Physiol.
290, H331-H340.
Papapetropoulos, A., Desai, K. M., Rudic, R. D., Mayer, B., Zhang, R.,
Ruiz-Torres, M. P., Garcia-Cardena, G., Madri, J. A. and Sessa, W. C.
(1997). Nitric oxide synthase inhibitors attenuate transforming-growth-factorbeta 1-stimulated capillary organization in vitro. Am. J. Pathol. 150, 18351844.
Pintucci, G., Moscatelli, D., Saponara, F., Biernacki, P. R., Baumann, F. G.,
Bizekis, C., Galloway, A. C., Basilico, C. and Mignatti, P. (2002). Lack of ERK
activation and cell migration in FGF-2-deficient endothelial cells. FASEB J. 16,
598-600.
Pozzi, A. and Zent, R. (2003). Integrins: sensors of extracellular matrix and
modulators of cell function. Nephron Exp. Nephrol. 94, e77-e84.
Qutub, A. A. and Popel, A. S. (2009). Elongation, proliferation and migration
differentiate endothelial cell phenotypes and determine capillary sprouting. BMC
Syst. Biol. 3, 13.
Development 137 (6)
Radisavljevic, Z., Avraham, H. and Avraham, S. (2000). Vascular endothelial
growth factor up-regulates ICAM-1 expression via the phosphatidylinositol 3
OH-kinase/AKT/Nitric oxide pathway and modulates migration of brain
microvascular endothelial cells. J. Biol. Chem. 275, 20770-20774.
Ruoslahti, E. and Engvall, E. (1997). Integrins and vascular extracellular matrix
assembly. J. Clin. Invest. 99, 1149-1152.
Schmidt, A., Wenzel, D., Ferring, I., Kazemi, S., Sasaki, T., Hescheler, J.,
Timpl, R., Addicks, K., Fleischmann, B. K. and Bloch, W. (2004). Influence
of endostatin on embryonic vasculo- and angiogenesis. Dev. Dyn. 230, 468480.
Senger, D. R., Perruzzi, C. A., Streit, M., Koteliansky, V. E., de Fougerolles, A.
R. and Detmar, M. (2002). The alpha(1)beta(1) and alpha(2)beta(1) integrins
provide critical support for vascular endothelial growth factor signaling,
endothelial cell migration, and tumor angiogenesis. Am. J. Pathol. 160, 195-204.
Shen, Y. H., Wang, X. L. and Wilcken, D. E. (1998). Nitric oxide induces and
inhibits apoptosis through different pathways. FEBS Lett. 433, 125-131.
Steinle, J. J., Meininger, C. J., Forough, R., Wu, G., Wu, M. H. and Granger,
H. J. (2002). Eph B4 receptor signaling mediates endothelial cell migration and
proliferation via the phosphatidylinositol 3-kinase pathway. J. Biol. Chem. 277,
43830-43835.
Tanjore, H., Zeisberg, E. M., Gerami-Naini, B. and Kalluri, R. (2008). Beta1
integrin expression on endothelial cells is required for angiogenesis but not for
vasculogenesis. Dev. Dyn. 237, 75-82.
Viji, R. I., Kumar, V. B., Kiran, M. S. and Sudhakaran, P. R. (2009). Modulation
of endothelial nitric oxide synthase by fibronectin. Mol. Cell. Biochem. 323, 91100.
Yang, J. T., Rayburn, H. and Hynes, R. O. (1993). Embryonic mesodermal
defects in alpha 5 integrin-deficient mice. Development 119, 1093-1105.
DEVELOPMENT
1002 RESEARCH ARTICLE