Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Dev Biol. Author manuscript; available in PMC 2009 January 15.
Published in final edited form as:
PMCID: PMC2239237

Vascular sprout formation entails tissue deformations and VE-cadherin dependent cell-autonomous motility


Embryonic and fetal vascular sprouts form within constantly expanding tissues. Nevertheless, most biological assays of vascular spouting are conducted in a static mechanical milieu. Here we study embryonic mouse allantoides, which normally give raise to an umbilical artery and vein. However, when placed in culture, allantoides assemble a primary vascular network. Unlike other in vitro assays, allantoic primordial vascular cells are situated on the upper surface of a cellular layer that is engaged in robust spreading motion. Time-lapse imaging allows quantification of primordial vascular cell motility as well as the underlying mesothelial tissue motion. Specifically, we calculate endothelial cell-autonomous motion by subtracting the tissue-level mesothelial motion from the total endothelial cell displacements. Formation of new vascular polygons is hindered by administration of function-blocking VE-Cadherin antibodies. Time-lapse recordings reveal that: 1) cells at the base of sprouts normally move distally “over” existing sprout cells to form new tip-cells; and 2) loss of VE-Cadherin activity prevents this motile behavior. Thus, endothelial cell-cell-adhesion-based motility is required for the advancement of vascular sprouts within a moving tissue environment. This is the first study that couples endogenous tissue dynamics to assembly of vascular networks in a mammalian system, to the best of our knowledge.

Keywords: allantois, mouse, vasculogenesis, VE-cadherin, time-lapse imaging, computational biology


Vasculogenesis, the de novo assembly of vessels from endothelial precursors, is a fundamental process common to both embryonic development and certain pathophysiologies. Embryonic, or primary vasculogenesis is the initial process by which a characteristic “polygonal” network of endothelial tubes forms, including larger caliber vessels such as the aortae (Coffin and Poole, 1988; Noden, 1989). In warm-blooded embryos vasculogenesis is an emergent process, which involves extensive movements of both individual cells and cell groups (Rupp et al., 2004). Primary vasculogenic cords form and re-model at a time when all embryonic cells are subjected to tissue expansions acting across the entire embryo (Keller et al., 2003). Thus, a cardinal characteristic of primary vasculogenesis is that it takes place in a moving environment. The formation of vascular cords from isolated clusters of angioblasts was found to involve extensive invasive activity, termed vasculogenic sprouting. In the vasculogenic sprouting process groups of endothelial cells invade hundreds of micrometers into avascular areas (see Figure 2 and Movie 2 in Rupp et al (2004)).

Figure 2
Vascular segment morphogenesis in the allantoic explant model at successively older stages

Vascular cords, formed by vasculogenic sprouting, are multicellular structures (Drake et al., 1997). Thus, vascular sprouting is envisioned as a collective outcome of a multicellular assembly process, during which cells must establish and maintain contact (Merks et al., 2006). Vascular endothelial (VE)-cadherin is the primary candidate for endothelium-specific cell-cell attachment, and is known to be an essential component of vascular development since a null mutation of the gene is lethal in mice at embryonic day (E) 9.5 (Carmeliet et al., 1999). To study the collective motility of endothelial cells during vasculogenic sprouting — and the role of VE-cadherins in the process — we employed explanted mouse allantoides.

The allantois model is an optically accessible high-fidelity in vitro replica of embryonic vascular patterning dynamics (Downs et al., 1998; Drake and Fleming, 2000). A particularly favorable aspect of allantoic explants is that unlike many in vitro systems — comprised of static substrates — the endothelial cells organize into polygonal networks while being subjected to large-scale tissue movements, similar to vasculogenesis in situ. The tissue motion arises from the fact that vascular polygons are separated from the culture dish by a continuous, intervening, motile sheet of mesothelial cells. The mesothelium rapidly expands across the substratum — always in advance of the incipient vasculogenic network.

In situ, primordial endothelial cells live within a constantly deforming tissue scaffold from the time of their specification to full differentiation — tissue motion and related forces are present at every stage of primary vasculogenesis. Despite this fact few studies address large-scale or bulk tissue movements when attempting to elucidate vascular pattern formation. The present biophysical analyses of vascular patterns were made possible by wide-field time-lapse scanning microscopy allowing automated imaging of entire embryos, or tissues, at cellular level resolution (Czirok et al., 2002). This integrated biophysical approach recently allowed visualization of endothelial cell dynamics during vasculogenesis, in vivo (Rupp et al., 2004) and permitted the calculation of cell autonomous versus tissue motion during gastrulation (Zamir et al., 2006). Here we demonstrate that VE-cadherin mediates multicellular vasculogenic sprouting in mouse allantoic explants. The empirical and computational data significantly alter our understanding of sprout formation dynamics, and emphasize the need for quantitative multi-scale analyses.

Materials and Methods


Wild-type timed-pregnant CD1 mice (Mus musculus) were purchased from Charles River Laboratories (Raleigh, NC) and embryos were harvested at embryonic day (E) 7.5–8.5. Animal care and experimental protocols were performed in compliance with the institutional and federal animal care guidelines.

Mouse Allantois Culture

Mouse embryos are dissected free of uterine muscle and decidual tissue in cold embryonic phosphate buffered saline (ePBS) and washed in fresh batch of cold ePBS (Crosby et al., 2005; Drake and Fleming, 2000). The allantois is extracted using fine-tipped forceps, washed in cold ePBS and pipetted into fibronectin-coated (5μg/ml) Delta T culture dishes (Bioptechs, Butler, PA) containing high-glucose phenol red-free Dulbecco’s modified Eagles’ medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 1% penicillin-streptomycin, and 1% L-glutamine (GibcoBRL, Grand Island, NY). Typically, four explants were cultured in each Delta T culture dish. Explants are then housed either in a standard tissue culture incubator or in a microscope-attached incubation chamber (see below) for 12–24 hours (Rupp et al., 2003). Allantoic explants were rinsed in PBS prior to fixation in 3% paraformaldehyde for 20 minutes at room temperature and washed with PBS-Azide.

Antibodies and Other Reagents

Purified rat anti-mouse CD34 monoclonal antibody (mAb), purified rat anti-mouse CD34-FITC mAb, and purified rat anti-mouse VE-cadherin (CD144) mAb were purchased from BD PharMingen (San Diego, CA). Supernatant containing hec 1.2 antibody was a gift from Dr. William A. Muller (Cornell University, New York, NY). For most perturbations, reagents at appropriate concentrations were added to fresh medium coincident with the start of image acquisition and were present, unless otherwise indicated, for the entire length of image acquisition. CD144 was used at a concentration of 20μg/ml and hec1.2 hybridoma supernatants were used full strength, approximately 100μg/ml (personal communication, Dr. W.A. Muller).

The CD34 antibody was directly conjugated to Cy3 with a CyDye kit (Amersham Biosciences). CD34-Cy3 was introduced to allantoic explants by addition to the cell culture medium (5μl Ab/1ml medium).

Dynamic Imaging

The allantoic cultures were maintained using standard culture conditions (37°C and 5% CO2/95% air atmosphere) in a custom-designed incubation system, consisting of a 4-well heated aluminum chamber connected to a calibrated peristaltic perfusion pump (Rupp et al., 2003). The cultures were observed with the 10X objective (0.30 N.A.) on an inverted automated epifluorescence/differential interference contrast (DIC) microscope (Leica DMIRE2, Leica Microsystems, Germany). Images (608×512 pixels spatial and 12 bit intensity resolution) were recorded with a cooled Retiga 1300 camera (QImaging, Burnaby, British Columbia) in 2×2 binned acquisition mode, using 100–300 ms exposure times. Image acquisition and microscope settings were controlled by software described (Czirok et al., 2002). Briefly, 4–10 pre-selected microscopic fields were visited in each scanning cycle. In each of these fields images were taken in two optical modes: DIC and epifluorescence. For each field and optical mode, 3–10 images were acquired in multiple focal planes, separated by 10 μm. The acquisition of the corresponding z-stacks was accomplished within a short period of time (typically a minute) ensuring the correct spatial registration of the DIC/epifluorescence images. The practical result of this automated technology is that no feature moves out of focus during the extended (18–24 hours) recording time.

During image processing, a mosaic image was created as described in Czirok et al. (2002). Thermal camera noise was reduced in the low intensity fluorescence images by applying a median filter (Chen et al., 1995). All images were locally normalized (Czirok et al., 2002) to maximally enhance image details. Temporal changes in fluorescence intensity along extending vascular branches were visualized with space-time plots of fluorescence intensities as described in Czirok et al. (2006). The corresponding image manipulations were performed with ImageJ (

Tracking of Image Details

A custom made program used in previous cell tracking studies (Rupp et al., 2004) allowed browsing through stored images (time-lapse z-stacks) and monitoring features, such as endothelial cells and vascular segments labeled with fluorescent CD34-Cy3, through consecutive image frames. Although this tracking is performed in three dimensions (x, y, and z), due to the large depth of field of the microscope objective (≈10μm), the resolution along the vertical (z) direction is limited. Therefore, we confine our studies to the two-dimensional (x–y) projections of the empirical position data.

Vascular segment length

Tips and bases of elongated vascular segments were identified as the distal-most extent (see arrow in Figure 2b) and as a proximal region, usually coinciding with the sprout’s origin (see arrowhead in Figure 2b), respectively. At a given time t, segment length d(t) was obtained as the distance between the tip and base markers. Segment expansion speed at time t was calculated as d(t+τ)d(tτ), with τ =1h. The significance of differences between segment expansion data sets (pooled in time categories of Fig 4b) were compared by Student’s t-test.

Figure 4
VE-cadherin dependent vascular segment growth

Endothelial cell motility

Fluorescence-labeled cells were traced throughout the recorded image sequence, resulting in the position xi(t) of cell i at various time points t. The non-autonomous (tissue convection) cell velocity component, Vi(t), was estimated by two methods: either by (1) performing a particle image velocimetry (PIV) analysis on the the DIC image sequence, or (2) applying a low-frequency (moving average) filter as Vi(t)= [xi(t+Δ)−xi(t−Δ)]/2Δ, with the choice of Δ=2.5h (Czirok et al., 2004; Sepich et al., 2005; Zamir et al., 2005). Both methods yielded very similar estimates of V (data not shown). The autonomous cell velocity was then calculated as vi(t)=[xi(t+τ)−xi(t)]/τ − Vi(t), where τ=1h, the recorded frame rate. The cell-autonomous net displacement was calculated as the magnitude of the vectorial sum composed by the vi(t)τ segments. The significance of differences in autonomous velocity was established, time point-by-time point, by Wilcoxon tests (at p=0.05 significance level). The statistical independence of distinct cells was assumed.


Vascular network formation in allantoic explants

Allantodies change shape dramatically when placed in planar culture: Instead of forming an umbilical chord-like homologue, containing one artery and one vein embedded in connective tissue, explanted allantoides form a polygonal endothelial cellular network (Crosby et al., 2005; Downs et al., 2001). Allantoic explants attach to the fibronectin-covered substrate over a 3–4 hour period, after which the mesothelial tissue layer begins to spread across the fibronectin-coated substratum (Figure 1a, Movie 1). This robust spreading behavior continues for 15–18 hours (Figure 1c), during which time the mesothelial surface area increased approximately 10–20 fold. To visualize endothelial cells, E8.0 allantoic explants were pulse-labeled with CD34-Cy3, a non-perturbing antibody specific for endothelial cells. Wide-field time-lapse imaging began after the CD34 pulse label, when the explants resembled Figure 1a′. The recordings typically lasted 18 to 24-hours (see Figures 1a′ and 1c′). Thus, we recorded the dynamics of primary vascular network formation while the supporting cell layer expanded by more than an order of magnitude.

Figure 1
Normal morphogenesis of a mouse allantoic explant

The immunoreactive extracellular CD34 epitopes are largely preserved during the recorded time period. However, most CD34 epitopes quickly cluster into bright foci, and the cell surface is visible only by a much fainter fluorescence. Control studies show that the binding of CD34 antibodies to endothelial cells did not cause detectable perturbations in vascular pattern formation. The final vascular pattern of allantoic explants pulse-labeled with CD34 (n=44) is indistinguishable from those of control explants immunolabeled after fixation (n=10). As a further quality control measure, antibodies to platelet/endothelial cell adhesion molecule (PECAM/CD31), another definitive marker of endothelial cells (Drake and Fleming, 2000) were used to immunolabel fixed cultures; CD31-labeled cultures (n=8) were also indistinguishable from live CD34 cultures or fixed CD34 cultures.

As the circled area in Fig. 1 demonstrates, some immunoreactive endothelial cells appear during the recordings. The gradual appearance of “new” immunopositive cells reveals differentiation of (initially) unlabeled endothelial precursors. The cells that manifested latent immunoreactivity appear in large unstructured clusters, which later give rise to linear vascular segments (Movie 1).

New polygonal elements form by two distinct cellular behaviors. The first mechanism entails a gradual enlargement of a small avascular area (Figure 2a). This process is highly reminiscent of vascular intussusception (Caduff et al., 1986; Djonov et al., 2000) whereby the avascular area gradually expands — possibly driven by the spreading mechanics of the underlying mesothelium. The second mechanism entails formation of new polygons via vasculogenic sprouting, i.e., the initiation, protrusion and subsequent expansion of a multicellular vascular segment or cord. Figure 2b shows a sprout that expands and joins another vascular segment in the circled area, thereby forming two sides of a “new” polygon.

Perturbed VE-cadherin function alters vascular pattern

To probe VE-cadherin adhesive function during vasculogenic dynamics, experiments were performed in the presence of 20 μg/ml CD144, a VE-cadherin function-blocking antibody that binds to extracellular domain (EC) 1 (Corada et al., 2002). Control explants were exposed a non-function blocking antibody, hec1.2 that binds VE-cadherin between EC3 and EC4 (Corada et al., 2001) at approximately 100μg/ml (see Methods).

The CD144 antibodies were added 4 hours after explants were placed in culture, coincident with the onset of time-lapse monitoring (n=25). During the 4-hour incubation, prior to addition of CD144, the explants attached to the substrate and commenced spreading. The presence of function-blocking VE-cadherin antibodies results in overt vascular pattern abnormalities such as a reduction in the number and length of vascular segments and a paucity of polygons — becoming apparent approximately 8 hours after addition of inhibitory VE-cadherin antibodies (Figure 3 and Movie 2). After 24 hours, the vascular pattern breaks up into disconnected clusters, each containing a reduced presence of CD34+ epitopes. Despite the fragmented polygons the cells in these clusters remain viable as demonstrated by propidium iodide exclusion (Dr Chris Drake, personal communication; see Figure 5, Crosby et al, 2005).

Figure 3
Inhibition of VE-cadherin receptor function results in failure of vascular pattern formation, i.e., primary vasculogenesis
Figure 5
Vasculogenic sprouts are replenished with new endothelial cells via VE-cadherin dependent migration toward the tip

VE-cadherin is involved in vasculogenic sprouting

To quantify vascular patterning failure in the presence of inhibitory VE-cadherin antibodies, the growth of vasculogenic sprouts were analyzed using image sequences from a subset of recordings. Tracking the positions of multicellular vascular segment “tips” and “bases” within each frame of a time-lapse movie sequence was conducted (see Figure 2b and Materials and Methods for definitions). Monitoring the distance of tip-base position pairs allowed calculation of segment length and expansion speed — parameters that characterize normal (n=15), non-function-blocking hec1.2 mAb treated (n=10) and VE-cadherin compromised (n=12) sprouting behavior.

In contrast to normal vascular segments that continue to elongate up to 30 hours, CD144-treated segments begin to grow initially (in the first 3–4 hours after exposure to antibody) but then retract, resulting in shorter segments with increasing time in culture (Figure 4a). Quantification of vascular segment expansion demonstrates that exposure to CD144 results in as much as a 70% reduction (p<0.05) in the average speed of vascular segment expansion (Figure 4b). The effect is detectable throughout the recordings, but it is most conspicuous 10 hours after introduction of the inhibitory antibodies. In contrast, no statistically significant differences (p>0.05) are observed in allantoic explants treated with the non function-blocking VE-cadherin antibodies, hec1.2 mAb. Thus, the on-time advancement of vasculogenic sprouts across a moving mesothelial sheet requires normal VE-cadherin function.

Impaired endothelial cell-autonomous motility

The failure of vasculogenic sprouting is accompanied by abnormal endothelial cell motility along the vascular segments. As the representative examples in Figure 5 and Movie 3 reveal, successful maintenance of segment expansion is coupled to a continuous supply of endothelial cells, streaming along the vascular segment, towards the sprout tip. During the unperturbed sprout expansion process, the constituent cells often rearrange, with tip cells sometimes being overtaken and passed to become the penultimate cell. Such cellular motility is absent when VE-cadherin function is blocked: no cells were observed to enter the newly formed sprout at the sprout base during the 11h recording period (Figure 5e versus 5e′). Thus, treatment with inhibitory VE-cadherin antibodies prevents “new” endothelial cells from advancing from a segment base towards the tip.

To quantify statistically the differences demonstrated in Fig. 5, we determined the autonomous motility of endothelial cells in untreated (n=21), and CD144-treated (n=15) allantoic explants. The rationale behind distinguishing cell autonomous motility versus passive tissue motion is the concept that the extension of vascular branches can be partially driven by the expansion of the entire explant — This is analogous to the lateral motion of mesodermal cells in a gastrulating embryo that is partially determined by large-scale tissue deformations (Czirok et al., 2004; Sepich et al., 2005; Zamir et al., 2005, 2006). To quantify cell-autonomous motility we subtracted the smooth, slowly varying motion component, i.e., the spreading of the mesothelium, from the observed or “apparent” total cell displacements (Figure 6). The difference between the tissue-derived (mesothelial) motion and total motion is the cell-autonomous component (Figure 6c).

Figure 6
Components of endothelial cell motion during vascular segment formation

Compared to untreated allantoic cultures, inhibition of VE-cadherin function resulted in a substantial, up to 50%, reduction in cell-autonomous motility (p<0.05) — whether characterized by velocity (Figure 7a) or by total distance moved (displacement), a measure of sustained cell motion (Figure 7b). These inhibitory effects are detectible immediately after optical tracking of CD34 epitopes begins, i.e., within a few hours after exposure to CD144 inhibitory VE-cadherin antibodies. Thus, embryonic mouse endothelial cells engage in autonomous, VE-cadherin-dependent motility along vascular sprouts — a motion that is independent of mesothelial spreading. We conclude that a combination of both cell-autonomous motility and passively driven tissue convection contribute to vascular pattern formation.

Figure 7
Endothelial cell-autonomous motility is VE-cadherin dependent


The gradual spreading of the allantoic mesothelium is the in vitro homologue of global tissue expansion during embryonic vasculogenesis: Allantoic vascular branches are subjected to tissue strain (discussed below) just as vascular sprouts in embryos are under the influence of global tissue deformations. The allantoic mesothelium is positioned between the glass substrate and the forming vascular network. As the mesothelium spreads the collective tissue motion imposes strain on the nascent vascular structures (polygons and sprouts). Similarly, during embryonic vasculogenesis in amniotes the forces of gastrulation, neurulation, and axis elongation impose long-range strain fields on the nascent vascular networks, which manifest in the correlated drift or motion of an entire vascular network (Rupp et al., 2004, Czirok et al., 2007). Presumably similar forces would act on angiogenic fronts during organogenesis or wound healing. In accord, allantoic vascular segments are expanded by 1) an active, sprouting process, and 2) by a passive strain generated by the gradual spreading of the explant mesothelium.

Both the passive and the autonomous components of cell motion can be estimated based on their respective statistical characteristics: Tissue motion involves deformation of a physical body, thus it is a smooth function of space and time (Czirok et al., 2004). In contrast, correlations in cell-autonomous movements decay rapidly and are statistically distinct from those characterizing tissue motion (Sepich et al., 2005; Zamir et al., 2005). We present here for the first time empirical time-lapse data showing that mouse embryonic endothelial cells engage in local cell-autonomous motility within vascular sprouts — in a dynamically coherent replica of vasculogenesis.

VE-cadherin is an exceptionally well-studied molecule as it is the principal junctional molecule on endothelial cell surfaces (Dejana et al., 1999). VE-cadherin ligation alters biochemical signaling pathways, such as VEGF and cell survival signaling (Carmeliet et al., 1999). VE-cadherins also modulate the permeability (Dejana et al., 1999) and mechanical state of blood vessels by altering cytoskeletal organization (Lampugnani et al., 2002) and the local transfer of mechanical stress (Nelson et al., 2004; Shay-Salit et al., 2002). VE-cadherin plays a crucial role in highly aggressive melanoma tumors through enabling vascular mimicry, the formation of vascular-like networks from melanoma cells (Hendrix et al., 2001).

However, the coalescence of angioblasts that normally occurs during early vasculogenesis involves VEGFR2-expressing vascular cells that do not yet express VE-cadherin (Drake and Fleming, 2000). In accord, rudimentary vascular structures are observed in the VE-cadherin knock out mouse (Carmeliet et al., 1999). The mutant embryos exhibit vascular defects, but display normal angioblast differentiation. While important aspects of VE-cadherin signaling are understood (Spagnuolo et al., 2004), the molecule’s mechanistic role in early vasculogenesis remains an area of active inquiry (Crosby et al., 2005).

Previous work with VE-cadherin antibodies suggested that VE-cadherin inhibition caused the disassembly of established blood vessels. This was inferred from experiments using allantoic explant cultures fixed at two time points (Crosby et al., 2005). Allantoides from VE-cadherin null embryos also failed to form polygonal vascular networks. Based on analysis of explants fixed after 9 and 18 hours of culture, Crosby et al (2005) suggested that initial events of vasculogenesis including the formation of nascent endothelial tubes are not dependent on VE-cadherin activity. Our dynamic imaging studies show, however, direct unequivocal evidence that endothelial cell motion along vascular segments is reduced within hours after blocking VE-cadherin function. We also see substantially reduced expansion speeds of vascular branches in early stages of allantoic explant spreading. Interestingly, time-lapse data reveal that the early inhibitory effects do not initially manifest themselves as a disconnected vascular network; therefore such an alteration is difficult, if not impossible to detect in specimens fixed at 9 and 18hrs. With the advantage of time-lapse imaging, however, these anomalies are readily discerned.

It is important to note that mesothelial (tissue) spreading proceeds readily in VE-cadherin inhibited cultures. However, the allantoic endothelial cells deprived of normal VE-cadherin function are limited in their ability to accommodate strain — because additional “new” endothelial cells cannot be incorporated into vascular cords. Thus, as VE-cadherin inhibition blocks the supply of “new” additional cells advancing toward the tip, vascular branches eventually either retract or break up — resulting in the state described by Drake and colleagues (Crosby et al., 2005).

Carmeliet et al (Carmeliet et al., 1999) showed that under certain conditions loss of VE-cadherin function leads to reduced cell survival. During the investigated time frame, however, lack of VE-cadherin function does not result in detectable changes in cell viability (Carmeliet et al., 1999). We find that CD144 treated endothelial cells continue to move for 9–18 hours, albeit at reduced velocities. Thus, inhibition of cell-autonomous motility in VE-cadherin inhibited cultures cannot be attributed to the presence of immobile dead cells (See Movies 13). Further, dye exclusion studies showed no decrease in viability in the case of endothelial cells treated with inhibitory VE-cadherin antibodies (Figure 5, Crosby et al, 2005).

Based on the above results and data obtained in avian embryos (Rupp et al., 2004), we suggest that vascular sprouting is initiated by the invasive behavior of a single “tip” cell — which engages the surrounding ECM using an integrin-dependent process. The advancing sprout, however, needs a continuous supply of endothelial cells, which may overtake, and pass by, the branch-initiating cell at the tip. Here we show that endothelial cells forming the future vascular cords are streaming along the vascular sprout, and this motion is VE-cadherin dependent. Thus, it is reasonable to postulate that VE-cadherin is required for the autonomous motility of endothelial cells that glide along vascular structures. We propose, therefore, that VE-cadherin-based motility is necessary to sustain vasculogenic sprouting.

There are multiple ways VE-cadherin can influence cell motility. The cytoplasmic tail of VE-cadherin binds to β-catenin and consequently alters the actin cytoskeleton (Cattelino et al., 2003; Wright et al., 2002). Although cadherins are often thought to mediate stable cell interactions and to decrease motility, cadherin-mediated adhesion is critical for gastrulation movements of X. laevis (Zhong et al., 1999), or growth cone motility (Matsunaga et al., 1988). Moreover, dynamic regulation of cell-cell adhesion is also expected to drive cell movements. For example, cadherin-mediated cell-cell contacts are hypothesized to perform an intercellular motility (Gumbiner, 2005) function that is analogous to the function of integrin-extracellular matrix contacts on moving cells. On the other hand, it is also known that VE-cadherin is involved in VEGF signaling. Thus, VEGF-induced mitogenic and survival signals are attenuated in the presence of blocked VE-cadherin (Carmeliet et al., 1999).

Our experiments cannot directly distinguish whether VE-cadherin functions as an intercellular migration adhesion receptor, or as a VEGF co-receptor. However, the established role of integrins in VEGF-induced motility (Hood et al., 2003) together with our failure to significantly reduce motility along vascular cords using inhibitory antibodies against α4 and α5 integrins (Perryn, 2006) indirectly favors the interpretation that VE-cadherin functions as an intercellular motility adhesion receptor. This assertion is most clearly supported by the data showing the prompt effect of VE-cadherin antibodies on the motility of endothelial cells along vascular segments. We did not detect reductions in endothelial cell motility when allantoides were treated with inhibitory integrin α4 and α5 antibodies (Perryn, 2006) — although it is important to stress that our efforts to perturb integrins were neither exhaustive, nor comprehensive.

The quantitative data described above advance our understanding of how tissue dynamics impacts the construction of a new vascular pattern. In particular, the tissue drift data bring into question reports of purported vascular malformations in mutant mice where the surrounding soma is malformed. Thus if a “vascular” anomaly is observed in an embryo that is overtly malformed it is not possible to state, with certainty, whether the vascular abnormality is primary, or is due to the fact that the underlying organ or tissue deformed improperly thereby causing a secondary vascular malformation. We speculate this may be particularly troublesome when analyzing vascular amomalies in early to mid-gestation mouse embryos.

Our data define normal and perturbed parameters of murine vasculogenesis using a high-fidelity model system — cultured allantoides. Argraves and Drake (2005) point out that the allantoic culture vasculogenesis assay is a useful and potentially important tool for understanding and quantifying endothelial cell behavior. Our empirical data show that endothelial cells differentiate from cultured allantoides and go on to participate in vascular network pattern formation; and neither CD34 antibodies nor wide-field time-lapse imaging cause detectable perturbation. This convenient culture method, when combined with our time-resolved analysis, allows computational characterization of mutant phenotypes present in the large repertoire of knockout and transgenic mice manifesting vascular defects (Argraves and Drake, 2005). Many of these knockout and transgenic mice are embryonic lethal, therefore complicating study of late-stage embryos, fetuses or neonates. By harvesting mutant allantoides prior to lethality, the dynamics of endothelial cell behavior, and the formation of “abnormal” primary vascular networks, can be studied in the context of a precise genotype.

Supplementary Material


Movie 1: Vascular network formation in a normal allantoic explant. Mouse allantoides were explanted at E8 and endothelial cells were labeled with fluorescent CD34 antibodies. DIC (upper left) and epifluorescence (lower left) images were collected over a 14h period. The panel on the right show the CD34 fluorescence superimposed with red on the DIC images. The field of view is 840 μm × 680 μm.


Movie 2: Disruption of Vascular network formation in an allantoic explant treated with CD144, a function-blocking antibody directed against VE cadherin. Endothelial cells are labeled with fluorescent CD34 antibodies, and images collected over a 21h long time period. The field of view is 670 μm × 450 μm.


Movie 3: Degree of autonomous endothelial cell motility along vascular sprouts in normal and treated cultures. The sprout base is to the left the tip to the right. When VE-cadherin function is normal (left panel), cells vigorously move relative to each other — specifically, the “red” marked cell joins the sprout early in the movie and eventually moves past the other cells to assume the distal-most (right-most) position. In contrast, when VE-cadherin function is blocked (right panel), the sprout extends to the right; however, the cells maintain the same relationships respect to each other. In both panels the tissue motion characteristic of explant expansion was compensated for — thus, a particle that appears stationary in the movies is moving in step with mesothelial expansion (Movie 1).


The authors thank Dr. William A. Muller (Cornell University, New York, NY) for the kind gift of the hec 1.2 VE-cadherin Ab. We also thank Dr. Chris Drake and Mr. Paul Fleming (Medical University of South Caroline, Charleston, SC) for valuable discussions.

Sources of Funding

Supported by grants from the American Heart Association SDG 0535245N and Hungarian Research Fund OTKA T047055 (AC); American Heart Association Pre-Doctoral Fellowship 0410084Z (EDP); NIH R01 HL068855, and an award from the G. Harold and Leila Y. Mathers Charitable Foundation (CDL).


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


  • Argraves WS, Drake CJ. Genes Critical to Vasculogenesis as Defined by Systematic Analysis of Vascular Defects in Knockout Mice. The Anatomical Record Part A. 2005;286A:875–884. [PubMed]
  • Caduff J, Fischer L, Burri P. Scanning electron microscope study of the developing microvasculature in the postnatal rat lung. Anat Rec. 1986;216:154–164. [PubMed]
  • Carmeliet P, Lampugnani MG, Moons L, Breviario F, Compernolle V, Bono F, Balconi G, Spagnuolo R, Oosthuyse B, Dewerchin M, Zanetti A, Angellilo A, Mattot V, Nuyens D, Lutgens E, Clotman F, de Ruiter MC, Gittenberger-de Groot A, Poelmann R, Lupu F, Herbert JM, Collen D, Dejana E. Targeted deficiency or cytosolic truncation of the VE-cadherin gene in mice impairs VEGF-mediated endothelial survival and angiogenesis. Cell. 1999;98:147–157. [PubMed]
  • Cattelino A, Liebner S, Gallini R, Zanetti A, Balconi G, Corsi A, Bianco P, Wolburg H, Moore R, Oreda B, Kemler R, Dejana E. The conditional inactivation of the b-catenin gene in endothelial cells causes a defective vascular pattern and increased vascular fragility. Journal of Cell Biology. 2003;162:1111–1122. [PMC free article] [PubMed]
  • Chen H, Swedlow J, Grote M, Sedat J, Agard D. Handbook of Biological Confocal Microscopy. 1995. The collection, processing, and display of digital three-dimensional images of biological specimens; pp. 197–210.
  • Coffin J, Poole TJ. Embryonic vascular development: immunohistochemical identification of the origin and subsequent morphogenesis of the major vessel primordia in quail embryos. Development. 1988;102:735–748. [PubMed]
  • Corada M, Liao F, Lindgren M, Lampugnani MG, Breviario F, Frank R, Muller WA, Hicklin DJ, Bohlen P, Dejana E. Monoclonal antibodies directed to different regions of vascular endothelial cadherin extracellular domain affect adhesion and clustering of the protein and modulate endothelial permeability. Blood. 2001;97:1679–1684. [PubMed]
  • Crosby CV, Fleming PA, Argraves WS, Corada M, Zanetta L, Dejana E, Drake CJ. VE-cadherin is not required for the formation of nascent blood vessels but acts to prevent their disassembly. Blood. 2005;105:2771–2776. [PubMed]
  • Czirok A, Rongish BJ, Little CD. Extracellular matrix dynamics during vertebrate axis formation. Developmental Biology. 2004;268:111–122. [PubMed]
  • Czirok A, Rupp PA, Rongish BJ, Little CD. Multi-field 3D scanning light microscopy of early embryogenesis. Journal of Microscopy. 2002;206:209–217. [PubMed]
  • Czirok A, Zach J, Kozel B, Mecham R, Davis E, Rongish BJ. Elastic fiber macro-assembly is a hierarchical, cell motion-mediated process. J Cell Physiol. 2006;207:97–106. [PubMed]
  • Czirok A, Zamir EA, Szabo A, Little CD. Multicellular sprouting during vasculogenesis. Current topics in developmental biology. 2007 in press. [PMC free article] [PubMed]
  • Dejana E, Bazzoni G, Lampugnani MG. Vascular endothelial (VE)-cadherin: only an intercellular glue? Experimental Cell Research. 1999;252:13–19. [PubMed]
  • Djonov V, Schmid M, Tschanz S, Burri P. Intussusceptive angiogenesis: its role in embryonic vascular network formation. Circ Res. 2000;86:286–292. [PubMed]
  • Downs KM, Gifford S, Blahnik M, Gardner RL. Vascularization in the murine allantois occurs by vasculogenesis without accompanying erythropoiesis. Development. 1998;125:4507–4520. [PubMed]
  • Downs KM, Temkin R, Gifford S, McHugh J. Study of the murine allantois by allantoic explants. Developmental Biology. 2001;233:347–364. [PubMed]
  • Drake CJ, Brandt S, Trusk T, Little CD. TAL1/SCL is expressed in endothelial progenitor cells/angioblasts and defines a dorsal-to-ventral gradient of vasculogenesis. Developmental Biology. 1997;192:17–30. [PubMed]
  • Drake CJ, Fleming PA. Vasculogenesis in the day 6.5 to 9.5 mouse embryo. Blood. 2000;95:1671–1679. [PubMed]
  • Gumbiner BM. Regulation of cadherin-mediated adhesion in morphogenesis. Nat Rev Mol Cell Biol. 2005;6:622–634. [PubMed]
  • Hendrix MJC, Seftor E, Meltzer P, Gardner L, Hess A, Kirschmann D, Schatteman G, Seftor R. Expression and functional significance of VE-cadherin in aggressive human melanoma cells: role in vasculogenic mimicry. PNAS. 2001;98:8018–8023. [PubMed]
  • Hood J, Frausto R, Kiosses W, Schwartz M, Cheresh DA. Differential alphav integrin-mediated Ras-ERK signaling during two pathways of angiogenesis. J Cell Biol. 2003;162:933–943. [PMC free article] [PubMed]
  • Keller R, Davidson L, Shook D. How we are shaped: the biomechanics of gastrulation. Differentiation. 2003;71:171–205. [PubMed]
  • Lampugnani MG, Zanetti A, Breviario F, Balconi G, Orsenigo F, Corada M, Spagnuolo R, Betson M, Braga V, Dejana E. VE-cadherin regulates endothelial actin activating Rac and increasing membrane association of Tiam. Molecular Biology of the Cell. 2002;13:1175–1189. [PMC free article] [PubMed]
  • Matsunaga M, Hatta K, Nagafuchi A, Takeichi M. Guidance of optic nerve fibres by N-cadherin adhesion molecules. Nature. 1988;334:62–64. [PubMed]
  • Merks R, Brodsky S, Goligorksy M, Newman S, Glazier J. Cell elongation is key to in silico replication of in vitro vasculogenesis and subsequent remodeling. Dev Biol. 2006;289:44–54. [PMC free article] [PubMed]
  • Nelson CM, Pirone DM, Tan JL, Chen CS. Vascular Endothelial-Cadherin Regulates Cytoskeletal Tension, Cell Spreading, and Focal Adhesions by Stimulating RhoA. Molecular Biology of the Cell. 2004;15:2943–2953. [PMC free article] [PubMed]
  • Noden D. Embryonic origins and assembly of blood vessels. Am Rev Respir Dis. 1989;140:1097–1103. [PubMed]
  • Perryn ED. Anatomy and Cell Biology. University of Kansas; Kansas City: 2006. Time-lapse imaging of vasculogenesis: Studies on the dynamics of endothelial cell behavior using a mouse allantois model; p. 129.
  • Rupp PA, Czirok A, Little CD. Novel approaches for the study of vascular assembly and morphogenesis in avian embryos. Trends Cardiovasc Med. 2003;13:283–288. [PubMed]
  • Rupp PA, Czirok A, Little CD. alphavbeta3 integrin-dependent endothelial cell dynamics in vivo. Development. 2004;131:2887–2897. [PubMed]
  • Sepich DS, Calmelet C, Kiskowski M, Solnica-Krezel L. Initiation of convergence and extension movements of lateral mesoderm during zebrafish gastrulation. Developmental Dynamics. 2005;234:279–292. [PubMed]
  • Shay-Salit A, Shushy M, Wolfovitz E, Yahav H, Breviario F, Dejana E. VEGF receptor 2 and the adherens junction as a mechanical transducer in vascular endothelial cells. PNAS. 2002;99:9462–9467. [PubMed]
  • Spagnuolo R, Corada M, Orsenigo F, Zanetta L, Deutsch U, Sandy P, Schneider C, Drake CJ, Breviario F, Dejana E. Gas1 is induced by VE-cadherin and vascular endothelial growth factor and inhibits endothelial cell apoptosis. Blood. 2004;103:3005–3012. [PubMed]
  • Wright TJ, Leach L, Shaw PE, Jones P. Dynamics of Vascular Endothelial-Cadherin and b-Catenin Localization by Vascular Endothelial Growth Factor-Induced Angiogenesis in Human Umbilical Vein Cells. Experimental Cell Research. 2002;280:159–168. [PubMed]
  • Zamir EA, Czirok A, Cui C, Little CD, Rongish BJ. Mesodermal cell displacements during avian gastrulation are due to both individual cell-autonomous and convective tissue movements. PNAS. 2006;103:19806–19811. [PubMed]
  • Zamir EA, Czirok A, Rongish BJ, Little CD. A Digital Image-Based Method for Computational Tissue Fate Mapping during Early Avian Morphogenesis. Ann Biomed Eng. 2005;33:854–865. [PubMed]
  • Zhong Y, Brieher W, Gumbiner BM. Analysis of C-cadherin regulation during tissue morphogenesis with an activating antibody. J Cell Biol. 1999;155:351–359. [PMC free article] [PubMed]