During angiogenesis, anastomosing capillary sprouts align to form complex three-dimensional networks of new blood vessels. Using an endothelial cell spheroid model that was developed to study endothelial cell differentiation processes, we have devised a novel collagen gel-based three-dimensional in vitro angiogenesis assay. In this assay, cell number-defined, gel-embedded endothelial cell spheroids act as a cellular delivery device, which serves as a focal starting point for the sprouting of lumenized capillary-like structures that can be induced to form complex anastomosing networks. Formation of capillary anastomoses is associated with tensional remodeling of the collagen matrix and directional sprouting of outgrowing capillaries towards each other. To analyze whether directional sprouting is dependent on cytokine gradients or on endothelial cell-derived tractional forces transduced through the extracellular matrix, we designed a matrix tension generator that enables the application of defined tensional forces on the extracellular matrix. Using this matrix tension generator, causal evidence is presented that tensional forces on a fibrillar extracellular matrix such as type I collagen, but not fibrin, are sufficient to guide directional outgrowth of endothelial cells. RGD peptides but not control RAD peptides disrupted the integrity of sprouting capillary-like structures and induced detachment of outgrowing endothelial cells cultured on top of collagen gels, but did not inhibit primary outgrowth of endothelial cells. The data establish the endothelial cell spheroid-based three-dimensional angiogenesis technique as a standardized, highly reproducible quantitative assay for in vitro angiogenesis studies and demonstrate that integrin-dependent matrix tensional forces control directional capillary sprouting and network formation.

Angiogenesis, the sprouting of new capillaries from pre-existing blood vessels, is characterized by a complex morphogenetic cascade of events during which quiescent resting endothelial cells become activated to proteolytically degrade their underlying extracellular matrix, directionally migrate towards the angiogenic stimulus, proliferate and align into new three-dimensional capillary networks (Augustin, 1998; Risau, 1997). The complexity of the angiogenic cascade requires endothelial cells to perform a number of distinct microenvironmental interactions with their surrounding extracellular matrix. As a consequence, angiogenic endothelial cells have a distinct gene expression pattern that is characterized by a switch of the cells’ proteolytic balance towards an invasive phenotype as well as the expression of specific adhesion molecules such as the integrin heterodimers αvβ3 and αvβ5 (Augustin, 1998; Bischoff, 1997; Preissner et al., 1997). Furthermore, angiogenic endothelial cells themselves synthesize components of the extracellular matrix and, thus, contribute to providing a provisional matrix along which sprouting capillaries continue to grow (Sephel et al., 1996; Haralabopoulos et al., 1994; Ingber et al., 1986).

A number of assays have been established to study angiogenesis and the properties of angiogenic endothelial cells in culture. Two-dimensional lateral sheet migration and proliferation assays have been used widely to study angiogenesis in vitro (Kim et al., 1998; Augustin and Pauli, 1992; Pepper et al., 1989; Sato and Rifkin, 1988). Reductionist as these assays may be, they clearly reflect some aspects of the angiogenic cascade, as indicated by the fact that several inhibitors of angiogenesis have been identified using migration and proliferation inhibition assays as screening systems (e.g. Ingber et al., 1990). Two-dimensional assays have been established based on the observation that endothelial cells cultured as monolayers on collagen gels or Matrigel align to form a network of endothelial cell cords (Vernon et al., 1992; Grant et al., 1989; Kubota et al., 1988; Montesano et al., 1983). This cord-forming process reflects morphogenetic properties of endothelial cells and has been successfully used to study endothelial cell and extracellular matrix interactions. As an angiogenesis assay, however, the Matrigel assay appears to be of limited use, because cord formation on Matrigel is not limited to only endothelial cells; it has also been observed in nonendothelial cells (Vernon et al., 1992).

In addition to planar assays, a number of three-dimensional cell culture systems have been developed to study specific steps of the angiogenic cascade. These include the capillary invasion of a collagen or fibrin gel from a monolayer of endothelial cells cultured on top of the gel (Montesano and Orci, 1985), the overlay of an endothelial cell monolayer cultured on top of a collagen or fibrin gel by another collagen or fibrin gel (Chalupowicz et al., 1995), the outgrowth of capillaries from an aortic ring embedded in a gel (Nicosia and Ottinetti, 1990), and the sprouting of endothelial cells from freshly isolated, collagen-embedded microvessels (Hoying et al., 1996). Some studies have described the formation of capillary networks originating from single gel-embedded endothelial cells (Ment et al., 1997; Madri et al, 1988). In these experiments, a large number of cells (1-2×106 cell/ml gel) need to be seeded in the gel, suggesting that this approach may be suitable for the analysis of capillary alignment and remodeling studies, but does not really reflect sprouting angiogenesis. More recently, a number of experimental systems have been described that are aimed at focally delivering aggregates of endothelial cells from which sprouting can occur. Pepper et al. (1991) first described an in vitro angiogenesis assay in which aggregates of endothelial cells are embedded in collagen or fibrin gels. This assay has been modified to deliver endothelial cells in fibrin gels by growing them on microcarrier beads (Nehls and Drenckhahn, 1995) or by embedding endothelial cell aggregates in collagen matrices, which are supported by annuli of nylon mesh (Vernon and Sage, 1999).

We have recently described a three-dimensional spheroid model of endothelial cell differentiation (Korff and Augustin, 1998). This study revealed that single nonadherent endothelial cells are destined to undergo apoptosis and that they are not responsive to the activities of cytokines that act as survival factors. In contrast, spheroidal aggregation stabilizes endothelial cells and renders them responsive to survival factors (Korff and Augustin, 1998). These observations most likely account for the fact that the embedding of single suspended endothelial cells in collagen or fibrin gels leads to massive apoptosis (Pollman et al., 1999; Satake et al., 1998; T. Korff and H. G. Augustin, unpublished observations), whereas the embedding of cellular aggregates or microcarrier adherent cells leads to radial capillary sprouting. We therefore conducted experiments aimed at developing an endothelial cell (EC) spheroid-based in vitro angiogenesis assay that would be a simple, highly reproducible, three-dimensional in vitro angiogenesis assay. During these experiments we observed that sprouting capillary-like structures grow directionally towards each other when embedded in collagen gels. We hypothesized that endothelial cell-derived matrix-transduced tractional forces are responsible for this directionality effect and set up experiments to investigate whether tensional forces are sufficient to induce directional capillary sprouting or if paracrine cytokine signaling phenomena are responsible for this effect.

Antibodies, growth factors and reagents

FGF-2 was obtained from Promega (Mannheim, Germany). VEGF was from Upstate Biotechnology (Lake Placid, NY). Carboxymethylcellulose (4.000 centipoises) and thrombin (bovine plasma) were from Sigma (Deisenhofen, Germany). Fibrinogen (bovine Plasma, clottable P>95%) was purchased from Calbiochem (Bad Soden, Germany). RGD-containing peptides (GRGDSP) as well as control RAD-peptides (GRADSP) were from Biomol (Hamburg, Germany).

Cell culture

Endothelial cell growth medium (ECGM) and endothelial cell growth supplement (human umbilical vein endothelial cell culture) were purchased from Promocell (Heidelberg, Germany). Dulbecco’s modified Eagle’s medium (DMEM) and other cell culture media were from Life Technologies (Gibco BRL, Eggenstein, Germany). Fetal calf serum (FCS) was obtained from Biochrom (Berlin, Germany). Bovine aortic endothelial (BAE) cells were isolated from thoracic aortas of healthy cattle by collagenase digestion following standard protocols. Cells were cultured at 37°C in 75-cm2 tissue culture dishes in DMEM containing 10% heat-inactivated fetal calf serum and frozen in liquid nitrogen at passage 2 or 3. Cells were routinely split in a 1:5 ratio and cultured for up to 50 passages. Only BAE cells cultured from passage 15-30 were used for experiments. Human umbilical vein endothelial (HUVE) cells were freshly isolated from human umbilical veins of newborn babies by collagenase digestion. Cells were cultured at 37°C in 75-cm2 tissue culture dishes in ECGM containing 10% heat-inactivated fetal calf serum and frozen in liquid nitrogen at passage 2 or 3. Only HUVE cells cultured from passage 4-8 were used for experiments.

Generation of endothelial spheroids

Endothelial cell spheroids of defined cell number were generated as described previously (Korff and Augustin, 1998). In order to generate endothelial cell spheroids of defined size and cell number, a specific number of BAE or HUVE cells (750) were suspended in corresponding culture medium containing 0.25% (w/v) carboxymethylcellulose and seeded in nonadherent round-bottom 96-well plates (Greiner, Frickenhausen, Germany). Under these conditions all suspended cells contribute to the formation of a single EC spheroid. These standardized spheroids were harvested within 24 hours and used for the corresponding experiments.

In vitro angiogenesis assay

In vitro angiogenesis in collagen gels was quantified using spheroids of endothelial cells in a modification of the microcarrier-bead angiogenesis assay (Nehls and Drenckhahn, 1995). In brief, HUVE or BAE cell spheroids, containing 750 cells, were generated overnight, after which they were embedded into collagen gels. A collagen stock solution was prepared prior to use by mixing acidic collagen extract of rat tails (equilibrated to 2 mg/ml, 4°C; 8 vol.) with 10× EBSS (Gibco BRL, Eggenstein, Germany; 1 vol.) and 0.1 N NaOH (approx. 1 vol.) to adjust the pH to 7.4. This stock solution (0.5 ml) was mixed with 0.5 ml room temperature medium [BAE cells: DMEM with 20% FCS; HUVE cells: ECGM basal medium (PromoCell, Heidelberg, Germany) with 40% FCS (Biochrom, Berlin, Germany)] containing 0.5% (w/v) carboxymethylcellulose to prevent sedimentation of spheroids prior to polymerization of the collagen gel, 50-100 HUVE or BAE cell spheroids, and the corresponding test substance. The spheroid-containing gel was rapidly transferred into prewarmed 24-well plates and allowed to polymerize (for 1 minute), after which 0.15 ml ECGM basal medium was pipetted on top of the gel. The gels were incubated at 37°C in 5% CO2 at 100% humidity.

To generate spheroid-containing fibrin gels, BAE cell spheroids (750 cells/spheroid) were suspended in DMEM containing 10% FCS and 2.5 mg/ml fibrinogen. Polymerization was induced by the addition of 1 U thrombin/ml fibrinogen solution (Sigma, Deisenhofen, Germany) after which the solution was rapidly mixed and transferred into 24-well plates. When the gelation was finished (1 minute) 0.15 ml DMEM was pipetted on top of the gel.

Two different techniques were applied to quantify in-gel angiogenesis. For rapid screening experiments, it proved to be sufficient to measure the length of the three longest capillary-like sprouts that had grown out of each spheroid after 3 days (ocular grid at 100× magnification), analyzing at least 10 spheroids per experimental group and experiment. For a more detailed quantitative analysis of in-gel angiogenesis, the cumulative length of all capillary-like sprouts originating from the central plain of an individual spheroid was measured at 100× magnification using a digitized imaging system (DP-Soft, Olympus) connected to an inverted microscope (IX50, Olympus). Again, at least 10 spheroids per experimental group and experiment were analyzed. This analysis takes into consideration that the angiogenic response induced by a specific substance is more appropriately reflected by the length of individual capillary-like sprouts as well as the number of capillary-like sprouts. Both quantitation techniques give similar results. The cumulative measurement of all capillary-like sprouts, however, has a higher level of resolution that permits assessment of smaller differences than the crude measurement of the three longest capillary-like sprouts.

Directional endothelial cell migration assay

Defined tensional forces on an underlying collagen or fibrin matrix were applied with a specialized mechanical device that served as a matrix tension generator (Fig. 1). The matrix tension generator was designed and built in cooperation with the Laboratory of Medical Mechanics at the University of Göttingen. For experiments with the matrix tension generator, a collagen or fibrin gel (1.0 ml) is poured into one well of a 24-well plate and allowed to polymerize. The matrix tension generator is inserted into the plate as shown in Fig. 1. The device holds two needles, which are fixed into the gel 6.0 mm from each other. One of the needles is laterally moved with a screw that drives a calibrated thread. By moving the needle 1.5 mm, 2.0 mm or 3.0 mm, the gel is stretched by 25%, 33% or 50%, respectively. Next, a single BAE cell spheroid was placed onto the stretched gel between the two needles. The spheroid was allowed to adhere, after which the cells grew out radially. Radial outgrowth was quantified microscopically after 24 hours. To quantify directionality of endothelial cell outgrowth as a consequence of matrix-transduced tensional forces, the ratio of length to width of endothelial cell outgrowth (directionality index) was calculated.

Fig. 1.

Design of the matrix tension generator used for the mechanical exertion of defined tensional forces on collagen and fibrin gels. (A) The matrix tension generator fits into a 24-well plate (1) and holds two needles in place (2), which are inserted in a collagen or fibrin gel. The needle holders (3) are flexibly inserted into the device and their heights can be adjusted manually (4). The screw on the right (5) drives a calibrated thread through which the needle holder of the right needle can be moved laterally to transmit tensional stress onto the gel. (B) Lateral view of the matrix tension generator showing a close up of the lateral needle moving device.

Fig. 1.

Design of the matrix tension generator used for the mechanical exertion of defined tensional forces on collagen and fibrin gels. (A) The matrix tension generator fits into a 24-well plate (1) and holds two needles in place (2), which are inserted in a collagen or fibrin gel. The needle holders (3) are flexibly inserted into the device and their heights can be adjusted manually (4). The screw on the right (5) drives a calibrated thread through which the needle holder of the right needle can be moved laterally to transmit tensional stress onto the gel. (B) Lateral view of the matrix tension generator showing a close up of the lateral needle moving device.

Morphological analysis

For morphological analysis, collagen or fibrin gels were fixed for 24 hours in HBSS containing 4% formaldehyde, after which they were processed for paraffin embedding. Following dehydration (in a graded series of ethanol and isopropanol, 24 hours each; 4°C), the gels were immersed with paraffin I (melting temperature 42°C) for 24 hours at 60°C and paraffin II (melting temperature 56°C) for 36 hours at 70°C. Finally, the resulting paraffin block was cooled to room temperature and trimmed for sectioning. Sections were stained with Hematoxylin.

Ultrastructural analysis of fibrin and collagen gels

Collagen or fibrin gels (24 hours after polymerization) were cut into pieces, washed in phosphate buffer (0.1 M, pH 7.4) and fixed in 1.0% osmium tetroxide, dehydrated in a graded series of ethanol, and embedded in epon. Sections of 0.5 μm were cut and stained with Azure 11 Methylene Blue for light microscopic evaluation. Ultrathin (50-80 nm) sections were cut, collected on copper grids, and automatically stained with uranyl acetate and lead citrate for observation with a Zeiss EM 10 electron microscope.

Endothelial cell spheroid based in vitro angiogenesis assay

In order to explore the suitability of EC spheroids as focal starting points for in-gel-based three-dimensional in vitro angiogenesis experiments, EC spheroids of defined cell number (750 cells/spheroid) were seeded in collagen gels and the outgrowth of capillary-like structures was assessed qualitatively and quantitatively. Endothelial cells originating from the embedded spheroids invade the gel to form complex networks of capillary-like structures (Fig. 2A). Cross sections of collagen gels revealed numerous capillary-like structures that differentiate to form a true capillary lumen (Fig. 2B-E). These lumenized structures are lined by a single layer of flattened endothelial cells. Occasionally, unicellular sprouts originating from these lumenized tubes can be identified in the gel (Fig. 2D). Cells that fail to integrate into the endothelial monolayer exhibit a condensed and fragmented nuclear morphology indicative of their programmed cell death (apoptosis) (Fig. 2E).

Fig. 2.

Lumenized capillary sprouts originating from collagen gelembedded spheroids of BAE cells. Gel-embedded spheroids give rise to radially outgrowing capillary sprouts. Outgrowing sprouts of neighboring spheroids grow directionally towards each other to establish networks of anastomosing capillary-like structures, as shown by phase-contrast microscopic analysis of three adjacent spheroids (A). Cross sections of gels with sprouts originating from EC spheroids show capillary-like structures of varying size with lining endothelial cells that form a true lumen throughout the gel. The morphological appearance of these capillary sprouts ranges from very small vessels lined by few EC (B-C) to larger structures lined by numerous EC that form a bigger lumen (D-E). Depending on the plain of section, sprouting EC originating from lumenized capillary sprouts can be identified (D, arrow). Integration of EC into the lining monocellular surface layer stabilizes the cells. Cells that are not integrated into the monolayer become apoptotic (E, arrows). Bars, 500 μm (A); 20 μm (B-E).

Fig. 2.

Lumenized capillary sprouts originating from collagen gelembedded spheroids of BAE cells. Gel-embedded spheroids give rise to radially outgrowing capillary sprouts. Outgrowing sprouts of neighboring spheroids grow directionally towards each other to establish networks of anastomosing capillary-like structures, as shown by phase-contrast microscopic analysis of three adjacent spheroids (A). Cross sections of gels with sprouts originating from EC spheroids show capillary-like structures of varying size with lining endothelial cells that form a true lumen throughout the gel. The morphological appearance of these capillary sprouts ranges from very small vessels lined by few EC (B-C) to larger structures lined by numerous EC that form a bigger lumen (D-E). Depending on the plain of section, sprouting EC originating from lumenized capillary sprouts can be identified (D, arrow). Integration of EC into the lining monocellular surface layer stabilizes the cells. Cells that are not integrated into the monolayer become apoptotic (E, arrows). Bars, 500 μm (A); 20 μm (B-E).

Quantitative assessment of three-dimensional in vitro angiogenesis was performed by microscopically measuring the length of outgrowing capillary sprouts with an ocular grid (Fig. 3). Collagen gel-embedded HUVE cell spheroids had a low level of spontaneous angiogenesis. Even in the presence of 20% serum, on average they only gave rise to few capillary sprouts with less than 50 μm length over a 3-day period (Fig. 3A,B). This behavior reflected the low autocrine activity of HUVE cells. HUVE cells were, however, readily responsive to exogenous stimulation by angiogenic growth factors. Addition of either FGF-2 or VEGF induced capillary sprouting of HUVE cells (Fig. 3A,C,D). In contrast to the low spontaneous angiogenic activity of HUVE cells, spheroids of BAE cells gave rise to intense sprouting, resulting in the formation of up to 300 μm long capillary-like structures within 3 days (Fig. 3E,F). In fact, even a reduction of serum concentrations to 2.0% did not significantly affect the high baseline angiogenic activity of BAE cells (data not shown), indicating and extending previous findings in different bioassays (Korff and Augustin, 1998; Villaschi and Nicosia, 1993; Mignatti et al., 1991) that BAE cells are strongly regulated by autocrine activity. Addition of exogenous FGF-2 or VEGF stimulated capillary sprouting of BAE cells, albeit to a much lower degree compared with HUVE cells (Fig. 3E,G,H).

Fig. 3.

Quantitative three-dimensional in vitro angiogenesis assay based on collagen gel-embedded endothelial cell spheroids. Capillary sprouting originating from the spheroids was quantified with a digitized imaging system as described in Materials and Methods, using human umbilical vein endothelial (HUVE) cells (A) and bovine aortic endothelial (BAE) cells (E). Representatives of each experimental group are shown in (B-D) (HUVE cells) and F-H (BAE cells). (B) Control HUVE cells; (C) HUVE cells + FGF-2 (30 ng/ml); (D) HUVE cells + VEGF (50 ng/ml); (F) control BAE cells; (G) BAE cells + FGF-2 (30 ng/ml); (H) BAE cells + VEGF (50 ng/ml). HUVE cells have a low baseline level of capillary sprouting and respond strongly to exogenous FGF-2 and VEGF. In contrast, BAE cells have a two times higher baseline sprouting activity, reflecting a much higher degree of autocrine activity (***, P<0.001 compared to corresponding control). Bars, 150 μm.

Fig. 3.

Quantitative three-dimensional in vitro angiogenesis assay based on collagen gel-embedded endothelial cell spheroids. Capillary sprouting originating from the spheroids was quantified with a digitized imaging system as described in Materials and Methods, using human umbilical vein endothelial (HUVE) cells (A) and bovine aortic endothelial (BAE) cells (E). Representatives of each experimental group are shown in (B-D) (HUVE cells) and F-H (BAE cells). (B) Control HUVE cells; (C) HUVE cells + FGF-2 (30 ng/ml); (D) HUVE cells + VEGF (50 ng/ml); (F) control BAE cells; (G) BAE cells + FGF-2 (30 ng/ml); (H) BAE cells + VEGF (50 ng/ml). HUVE cells have a low baseline level of capillary sprouting and respond strongly to exogenous FGF-2 and VEGF. In contrast, BAE cells have a two times higher baseline sprouting activity, reflecting a much higher degree of autocrine activity (***, P<0.001 compared to corresponding control). Bars, 150 μm.

Capillary sprouting from gel-embedded EC spheroids leads to directional outgrowth of anastomosing capillaries

When seeding EC spheroids at different densities in the collagen gel, we observed that beyond a critical spheroid density, capillary sprouts originating from the spheroids gave rise to complex anastomosing capillary-like networks (Fig. 2A). This observation prompted us to study systematically directional sprouting in three-dimensional collagen gels and the mechanisms that are responsible for this directionality.

Upon seeding of EC spheroids into collagen gels, capillary sprouts grow radially in all three dimensions (Fig. 4A,B). After about 2 days in the gel, some sprouts change their direction to grow towards a neighboring spheroid if this spheroid is in close proximity. Analysis of a large number of gel-embedded spheroids indicated that endothelial cell sprouts can sense directionality in collagen gels over a distance of approximately 600-800 μm (Fig. 4C-E). Eventually, as the distance between different capillary sprouts becomes smaller, several sprouts align to span the gel between the two neighboring spheroids (Fig. 4F). The directionality effect was identified as a specific phenomenon of collagen matrices, since no signs of directional capillary sprouting were observed in fibrin gels (data not shown).

Fig. 4.

Directional sprouting of capillary-like structures towards each other, originating from two neighboring gel-embedded spheroids with a distance of approximately 1.6 mm (measured from the center of each spheroid, A). Sprouts grow radially out of the spheroid for the first 2 days. After 3 days, capillary sprouts start to change their direction to grow towards each other, which becomes even more evident after 4 and 5 days. Note that the centers of the two spheroids have moved closer together after 4 and 5 days, reflecting the tractional forces exerted by the outgrowing endothelial cells. Bar, 200 μm.

Fig. 4.

Directional sprouting of capillary-like structures towards each other, originating from two neighboring gel-embedded spheroids with a distance of approximately 1.6 mm (measured from the center of each spheroid, A). Sprouts grow radially out of the spheroid for the first 2 days. After 3 days, capillary sprouts start to change their direction to grow towards each other, which becomes even more evident after 4 and 5 days. Note that the centers of the two spheroids have moved closer together after 4 and 5 days, reflecting the tractional forces exerted by the outgrowing endothelial cells. Bar, 200 μm.

Outgrowing capillary sprouts exert tractional forces on the extracellular matrix

When spheroid-derived sprouts invade the collagen, they exert tractional forces onto the gel. These tractional forces are responsible for the shrinkage effect frequently observed when cells are embedded in floating collagen gels (Vernon and Sage, 1996; Gullberg et al., 1990; Harris et al., 1981). In the present study, we used nonfloating collagen gels that were solidly anchored in their tissue culture well. Nevertheless, regional differences in tractional forces could be identified when observing neighboring spheroids over several days. Fig. 4A shows two collagen gel-embedded neighboring spheroids approximately 1.6 mm apart. After 5 days in the collagen gel and extensive directional sprouting, the centers of the two spheroids were only approximately 1.2 mm apart, indicating that the two spheroids have moved 25% closer together.

Based on this observation, we analyzed the fibrillar structure of the collagen gels that contained capillary-like structures. The collagen fibrils between two neighboring spheroids were found to align along the axis that connects the two spheroids (Fig. 5A,C). The distance over which this alignment was found corresponded to the distance up to which we observed

Fig. 5.

(A,B) Cross sections of Hematoxylin-stained collagen gels containing capillary-like structures at different distances from each other (approx. 400 μm, A; approx. 70 μm, B). (C, D) Schematic representations of the alignment of collagen fibrils shown in A and B, respectively. Tractional forces exerted by sprouting endothelial cells lead to an alignment of collagen fibrils, which can be observed up to distances of 500-700 μm (A,C). When the capillary sprouts get into closer proximity, essentially all of the collagen fibrils between the sprouts have aligned (B,D). Bars, 100 μm (A); 20 μm (B).

Fig. 5.

(A,B) Cross sections of Hematoxylin-stained collagen gels containing capillary-like structures at different distances from each other (approx. 400 μm, A; approx. 70 μm, B). (C, D) Schematic representations of the alignment of collagen fibrils shown in A and B, respectively. Tractional forces exerted by sprouting endothelial cells lead to an alignment of collagen fibrils, which can be observed up to distances of 500-700 μm (A,C). When the capillary sprouts get into closer proximity, essentially all of the collagen fibrils between the sprouts have aligned (B,D). Bars, 100 μm (A); 20 μm (B).

directional capillary sprouting (500-700 μm). When the sprouts come into even closer proximity (<200 μm and less), essentially all collagen fibrils between the sprouts are aligned to connect the tips of the two adjacent sprouts.

Matrix-transduced tensional forces control directional outgrowth of endothelial cells in collagen gels but not in fibrin gels

The observations made thus far indicate that (1) invading endothelial cells exert tractional forces on the matrix, (2) collagen fibrils align along the axis of the matrix tension, and (3) capillary sprouts grow directionally along these aligned collagen fibrils to form anastomosing capillary-like networks. We next set up experiments to investigate whether matrix-transduced tensional forces exerted by invading endothelial cells are responsible and sufficient for directional endothelial cell outgrowth. In order to exclude possible paracrine cytokine signaling effects of endothelial cells sprouting from two neighboring spheroids, we performed experiments with single spheroids, using a mechanical device that served as a matrix-tension generator (Fig. 1). By controlling the lateral movement of two needles that are inserted into a collagen or fibrin gel, this device can be used to apply defined tensional forces onto the gel (Fig. 6A). After applying tensional stress onto the gel, a single EC spheroid is seeded on the gel between the two needles. If the applied tension is sufficient to control directional outgrowth of the cells, this should lead to a nonradial, ellipsoid pattern of outgrowth along the axis of the tensional stress, which can be used to calculate a directionality index (Fig. 6B). Additionally, this experiment would demonstrate that outgrowing cells do not just exert tractional forces onto the extracellular matrix, but rather that they can ‘read’ the direction of tension-aligned fibrils in the matrix.

Fig. 6.

Analysis of directional endothelial cell outgrowth by mechanical manipulation of matrix tension with a matrix-tension generator. (A) Conceptual design of the matrix-tension generator. Two needles are inserted into a collagen or fibrin gel and controlled tensional forces are exerted by lateral movement of one of the needles, after which a single EC spheroid is seeded between the needles (for details compare Fig. 1 and Materials and Methods). (B) Quantification of a directionality index. The length and the width of the front of outgrowing endothelial cells is measured microscopically and a directionality index calculated as the ratio of length to width. (C) Radial outgrowth of BAE cells from a single spheroid seeded on top of a nonstretched collagen gel. Directional outgrowth of BAE cells from a single spheroid seeded on top of a stretched collagen gel. (E) Quantification of directional endothelial cell outgrowth after seeding of a single BAE cell spheroid on top of a stretched collagen or fibrin gel. There is significant, dose-dependent directional outgrowth of BAE cells grown on top of a stretched collagen gel (P<0.001 at all points compared to control). In contrast, stretched fibrin gels do not support directional outgrowth. Bar, 100 μm.

Fig. 6.

Analysis of directional endothelial cell outgrowth by mechanical manipulation of matrix tension with a matrix-tension generator. (A) Conceptual design of the matrix-tension generator. Two needles are inserted into a collagen or fibrin gel and controlled tensional forces are exerted by lateral movement of one of the needles, after which a single EC spheroid is seeded between the needles (for details compare Fig. 1 and Materials and Methods). (B) Quantification of a directionality index. The length and the width of the front of outgrowing endothelial cells is measured microscopically and a directionality index calculated as the ratio of length to width. (C) Radial outgrowth of BAE cells from a single spheroid seeded on top of a nonstretched collagen gel. Directional outgrowth of BAE cells from a single spheroid seeded on top of a stretched collagen gel. (E) Quantification of directional endothelial cell outgrowth after seeding of a single BAE cell spheroid on top of a stretched collagen or fibrin gel. There is significant, dose-dependent directional outgrowth of BAE cells grown on top of a stretched collagen gel (P<0.001 at all points compared to control). In contrast, stretched fibrin gels do not support directional outgrowth. Bar, 100 μm.

As shown in Fig. 6C, a perfectly radial outgrowth of endothelial cells was observed, when a single EC spheroid was seeded on top of a nonstretched collagen gel. With increasing tensional stress, outgrowth of endothelial cells became more asymmetric along the direction of the tensional forces exerted by the matrix tension generator (Fig. 6D), leading to a dose-dependent increase of the directionality index (Fig. 6E). In contrast to the directional outgrowth of endothelial cells cultured on top of stretched collagen gels, no such directionality was observed when EC spheroids were placed on top of fibrin gels. Even when the fibrin gels were prestretched by 50%, no significant directional outgrowth of endothelial cells was observed (Fig. 6E), confirming that fibrin gels do not support directional capillary sprouting in three-dimensional gels.

In order to further analyze the difference between collagen and fibrin gels in supporting directional endothelial cell outgrowth, collagen and fibrin gels were ultrastructurally analyzed by electron microscopy. Analysis of in vitro polymerized collagen confirmed the complex structure of collagen gels with twisted, irregularly assembled collagen fibrils several μm long, which lack the regular striated appearance of in vivo assembled collagen fibrils (Fig. 7A) (Fratzl et al., 1997; Ploetz et al., 1991). In contrast, analysis of in vitro polymerized fibrin revealed the compact, densely meshed structure of fibrin gels with short fibrils only a few hundred nm long.

Fig. 7.

Ultrastructural analysis of a collagen (A) and a fibrin (B) gel. (A) Longitudinal and transverse section of a collagen gel fibrils. Collagen forms twisted fibrils several μm long. (B) In contrast, fibrin fibrils are short and have a compact, densely meshed structure. Bar, 200 nm.

Fig. 7.

Ultrastructural analysis of a collagen (A) and a fibrin (B) gel. (A) Longitudinal and transverse section of a collagen gel fibrils. Collagen forms twisted fibrils several μm long. (B) In contrast, fibrin fibrils are short and have a compact, densely meshed structure. Bar, 200 nm.

Integrin-mediated cell matrix contacts are involved in regulating three-dimensional capillary sprouting and two-dimensional EC outgrowth

The tractional forces exerted by invading endothelial cells on the surrounding extracellular matrix involve complex adhesive interactions between the cells and the matrix. In order to analyze the contribution of integrin-mediated cell matrix contacts on sprouting and directional endothelial cell migration, three-dimensional in vitro angiogenesis experiments and two-dimensional on-gel cellular outgrowth experiments were performed in the presence of RGD or control RAD peptides. A concentration of 30 μM RGD disrupted capillary morphogenesis in three-dimensional collagen gels (Fig. 8A). Interestingly, invasion of EC into the gel was not inhibited by RGD peptides, but addition of the peptide disrupted the integrity of the capillary sprouts. On two-dimensional collagen gels, addition of RGD peptides induced the rounding and detachment of the cells from the matrix, but corresponding to the experiments in three-dimensional gels, they did not inhibit cellular outgrowth as such (Fig. 8B). In all of these experiments, RAD peptides (30 μM) served as controls. RAD peptides neither inhibited in vitro angiogenesis in three-dimensional collagen gels (Fig. 8C) nor affected EC outgrowth or monolayer integrity on top of collagen gels (Fig. 8D).

Fig. 8.

Effect of RGD peptides on in vitro angiogenesis. BAE cell spheroids were either embedded in collagen gels (A,C) or cultured on top of collagen gels (B,D) and incubated for 3 days in the presence of 30 μM RGD peptides or with the same concentration of control RAD peptides. (A) Integrin-dependent contacts inhibiting RGD peptides disrupt capillary-like sprouting in collagen gels. (B) Addition of RGD peptides to laterally migrating endothelial cells from a spheroid cultured on top of a collagen gel leads to rounding and detachment of the cells. (C,D) Control RAD peptides have no effect on in gel angiogenesis (C) or on cellular outgrowth on top of a collagen gel (D). Bars, 100 μm.

Fig. 8.

Effect of RGD peptides on in vitro angiogenesis. BAE cell spheroids were either embedded in collagen gels (A,C) or cultured on top of collagen gels (B,D) and incubated for 3 days in the presence of 30 μM RGD peptides or with the same concentration of control RAD peptides. (A) Integrin-dependent contacts inhibiting RGD peptides disrupt capillary-like sprouting in collagen gels. (B) Addition of RGD peptides to laterally migrating endothelial cells from a spheroid cultured on top of a collagen gel leads to rounding and detachment of the cells. (C,D) Control RAD peptides have no effect on in gel angiogenesis (C) or on cellular outgrowth on top of a collagen gel (D). Bars, 100 μm.

The complexity of the angiogenic cascade is increasingly recognized. Originally supported by a simplified model of an invasive, migratory, proliferating, angiogenic endothelial cell, angiogenesis is now widely viewed as a complex morphogenetic event that includes discrete steps of capillary organization, tubular branching, network formation and vessel maturation (Darland and D’Amore, 1999; Hanahan, 1997), leading to the development of in vitro assays of angiogenesis to facilitate the study of individual steps of the angiogenic cascade under defined conditions. Towards this end a number of two-dimensional and three-dimensional assays have been developed (Vernon and Sage, 1999; Hoying et al., 1996; Nehls and Drenckhahn, 1995; Augustin and Pauli, 1992; Nicosia and Ottinetti, 1990; Grant et al., 1989; Madri et al, 1988; Montesano and Orci, 1985; Montesano et al., 1983) and all of these assays legitimately reflect some aspects of angiogenesis. Nevertheless, the limitations of in vitro systems necessitate caution when interpreting findings from such in vitro models to the in vivo situation (Passaniti, 1992; Madri and Basson, 1992; Grant et al., 1992; Vernon and Sage, 1992)

EC spheroid-based three-dimensional in vitro angiogenesis assay

The aim of the present study was to develop a novel three-dimensional in vitro angiogenesis assay, taking advantage of a recently established spheroid model of endothelial cell differentiation (Korff and Augustin, 1998). EC spheroids can be produced at any size with a defined number of cells. The EC spheroid-based in vitro angiogenesis assay proved to be simple and highly reproducible. The assay has a number of advantages over other EC aggregate-based in vitro angiogenesis assays. By utilizing cell number-defined EC spheroids, standardized experimental conditions can be applied. In comparison to the microcarrier bead angiogenesis assay (Nehls and Drenckhahn, 1995), larger numbers of endothelial cells can be focally applied into the gel, avoiding the possibility that EC proliferation may become rate-limiting for the sprouting angiogenesis process.

The very distinct advantage of all cell aggregate-based angiogenesis assays is the fact that endothelial cells can be focally delivered into a three-dimensional matrix, allowing sprouting angiogenesis to occur by invasion into the extracellular matrix. This process is very much in contrast to the alignment of large numbers of gel-embedded single endothelial cells and appears to be the closest in vitro representation of the angiogenic invasion of the extracellular matrix as it occurs during angiogenesis in vivo. If sprouting angiogenesis originating from gel-embedded EC spheroids is allowed to proceed for several days, capillary sprouts lead to the formation of complex three-dimensional networks, which can be used to analyze individual steps of the angiogenic cascade sequentially and the interactions of the forming endothelial cell network with mural cells (T. Korff and H. G. Augustin, manuscript in preparation).

We have applied the EC spheroid-based in vitro angiogenesis assay towards the analysis of the angiogenic capacity of two different cell populations, human umbilical vein endothelial (HUVE) cells and bovine aortic endothelial (BAE) cells. The two cell populations show distinct differences in their angiogenic capabilities, making them suitable target cell populations for different specific questions. HUVE cells have a low baseline angiogenic activity requiring stimulation with exogenous angiogenic cytokines such as VEGF or FGF-2 for angiogenesis to occur. This makes them suitable target cells for the analysis of angiogenesis-promoting substances. We recently exploited this capacity in the characterization of the newly identified Orf virus-encoded VEGF variant VEGF-E (Meyer et al., 1999). In contrast, BAE cells have a high baseline angiogenic activity in the absence of exogenous cytokines, even under strongly serum-reduced culture conditions. This high degree of autocrine activity corresponds to the high intensity of autocrine activity of these cells that has been observed in a number of other bioassays and appears to be largely mediated by endogenous expression of FGF-2 (Korff and Augustin, 1998; Villaschi and Nicosia, 1993; Mignatti et al., 1991). As a consequence, the autocrine activity makes BAE cells highly suitable for the study of angiogenesis-inhibiting substances.

Directional capillary sprouting in collagen versus fibrin gels

Gel-embedded EC spheroids gave rise to complex capillary-like networks. This network-forming activity appears not to be a random process, but rather the consequence of directional outgrowth of capillary sprouts towards each other. Directional capillary sprouting appears to be a critical morphogenetic determinant during vessel formation in vivo. During vasculogenesis, a primitive capillary plexus is formed as a consequence of the tractional coalescence of in situ differentiating angioblastic cells (Risau and Flamme, 1995). Likewise during angiogenesis, capillary sprouts form anastomoses that give rise to capillary networks (Benjamin et al., 1998).

Capillary-like network formation has been associated with mechanical forces exerted by angiogenic endothelial cells onto its surrounding extracellular matrix (Vernon and Sage, 1995; Ingber and Folkman, 1989). Alternatively, or acting in concert with mechanical forces, paracrine cytokine signaling processes may be responsible for the directed growth of two capillary sprouts towards each other. Tractional forces of cells exerted onto its surrounding fibrillar extracellular matrix have been most extensively studied in fibroblasts. Fibroblast-mediated tractional forces are believed to be primary morphogenetic determinants in the formation of tendons as well as during wound contraction (Harris et al., 1981). Endothelial cells have been reported to exert similar tractional forces on type I collagen as dermal fibroblasts (Vernon and Sage, 1996). Analysis in planar two-dimensional systems has given good evidence that mechanical forces exerted by endothelial cells actively modulate the underlying extracellular matrix and that this matrix modulatory effect may contribute to capillary morphogenesis (Vernon et al., 1992). Most of these experiments were performed with Matrigel as extracellular matrix. Matrigel is rich in laminin, type IV collagen and fibronectin and may not be very representative of the type I collagen-rich interstitial matrix that endothelial cells are mostly exposed to during physiological angiogenesis in vivo. Correspondingly, planar tube formation assays have similarly shown that endothelial cell-derived tractional forces can modulate the directional alignment of individual collagen fibrils (Vernon and Sage, 1995).

In the present study we observed that sprouts originating focally from collagen gel-embedded EC spheroids give rise to complex three-dimensional networks of directionality towards each other, growing anastomosing, capillary-like structures. Morphological analysis of the gels demonstrated that collagen fibrils between adjacent capillary-like sprouts align over distances of up to 1 mm. This distance can be reduced or increased somewhat by changing the mechanical properties of the gel, but it appears to correspond well with the geometrical properties of experimental models of capillary network formation in vivo, as can be observed, for example, during rabbit cornea angiogenesis (Kozian et al., 1997) or postnatal retinal vascularization (Benjamin et al., 1998).

It has long been speculated that matrix-transduced tensional forces are responsible for directional capillary sprouting (Vernon and Sage, 1995). To date, however, no causal evidence has been presented to demonstrate that matrix-transduced tensional forces are responsible and sufficient for this effect. This prompted us to develop a mechanical device that is capable of generating defined mechanical tension on to the matrix and allowed us to study the outgrowth of endothelial cells on top of stretched collagen and fibrin gels to exclude paracrine cytokine-mediated signaling effects which might be responsible for the directional capillary sprouting towards each other. These experiments showed unambiguously for the first time that matrix-transduced tensional forces in stretched collagen gels are sufficient to control the directional outgrowth of endothelial cells. In turn, the experiments also show that matrix-transduced tensional forces are, thus, not just a byproduct of tractional forces exerted by invading endothelial cells, but rather that outgrowing endothelial cells can ‘read’ the direction of tension-aligned fibrils in the extracellular matrix.

Directional capillary sprouting was only observed in type I collagen gels, but not in fibrin gels. Likewise, directional outgrowth on stretched gels was only seen in type I collagen gels and not in fibrin gels. As confirmed by direct electron microscopic analysis and corresponding well with data in established literature, fibrin forms much shorter fibrils than collagen, which appears to limit its mechanotransducing capacity. The different behaviors of fibrin and collagen gels in supporting directional capillary sprouting and network formation may not merely reflect in vitro behavior of cells in gels, but rather suggests some thought-provoking hypotheses on the mechanisms of capillary network formation in vivo. Physiological, developmental angiogenesis occurs mostly through a collagen-rich extracellular matrix and leads to the formation of a regular capillary network. In contrast, pathological angiogenic processes in the adult, as they are associated with tumor growth or wound healing, lead to the formation of an irregular, highly tortuous capillary network, as was shown by numerous casting experiments (Konerding et al., 1995). Tumor and wound healing angiogenesis, however, are capillary sprouting processes known to primarily occur in a fibrin-rich matrix (Dvorak, 1986).

Taken together, the present study presents causal evidence that matrix-transduced tensional forces are critical determinants in capillary morphogenetic network formation processes. Matrix-transduced tensional forces are sufficient to control directional EC outgrowth. Invasive endothelial cells do not just create tractional forces on the matrix, but they can ‘read’ the tension-aligned orientation of fibrillar extracellular matrices. Directional capillary sprouting processes are limited to large fibrillar matrices such as type I collagen and are not observed in short fibrillar matrices such as fibrin. This finding may have implications for the differences in the capillary morphogenetic processes that occur in collagen- and fibrin-rich matrices in vivo, e.g. the differences observed between regular developmental angiogenesis in a collagen-rich matrix and the irregular, tortuous capillaries that form in fibrin matrices as they are associated with pathological angiogenic processes such as the growth of tumors or the healing of a wound. Furthermore, the angiogenesis assay described in this study may prove to be a versatile tool for quantitative and qualitative angiogenesis studies in vitro. It is simple, highly reproducible, easy to quantify, and allows the establishment of complex three-dimensional networks in culture. This may prove useful for studies aimed at sequentially analyzing individual steps of the angiogenic cascade, including detailed morphogenetic studies involving branchpoint analyses and the formation of anastomoses as well as endothelial cell and mural cell interactions.

The authors would like to acknowledge the excellent technical assistance of Mrs Renate Dietrich and Mrs Cathleen Lakoma. We thank Dr Franz-J. Kaup (DPZ, Göttingen, Germany) for assistance with the electron microscopic analyses and Mr Wegener (Laboratory for Medical Mechanics, University of Göttingen, Germany) for support in designing and constructing the matrix tension generator. This work was supported by a grant from the Deutsche Forschungsgemeinschaft (SFB500, C3).

Augustin
,
H. G.
(
1998
).
Antiangiogenic tumour therapy: will it work?
Trends Pharmacol. Sci
.
19
,
216
222
.
Augustin
,
H. G.
and
Pauli
,
B. U.
(
1992
).
Quantitative analysis of autocrineregulated, matrix-induced and tumor cell-stimulated endothelial cell migration using a silicon template compartmentalization technique
.
Exp. Cell Res
.
198
,
221
227
.
Benjamin
,
L. E.
,
Hemo
,
I.
and
Keshet
,
E.
(
1998
).
A plasticity window for blood vessel remodelling is defined by pericyte coverage of the preformed endothelial network and is regulated by PDGF-B and VEGF
.
Development
125
,
1591
1598
.
Bischoff
,
J.
(
1997
).
Cell adhesion and angiogenesis
.
J. Clin. Invest
.
99
,
373
376
.
Chalupowicz
,
D. G.
,
Chowdhury
,
Z. A.
,
Bach
,
T. L.
,
Barsigian
,
C.
and
Martinez
,
J.
(
1995
).
Fibrin II induces endothelial cell capillary tube formation
.
J. Cell Biol
.
130
,
207
215
.
Darland
,
D. C.
and
D’Amore
,
P. A.
(
1999
).
Blood vessel maturation: vascular development comes of age
.
J. Clin. Invest
.
103
,
157
158
.
Dvorak
,
H. F.
(
1986
).
Tumors: wounds that do not heal. Similarities between tumor stroma generation and wound healing
.
New Engl. J. Med
.
315
,
1650
1659
.
Fratzl
,
P.
,
Misof
,
K.
,
Zizak
,
I.
,
Rapp
,
G.
,
Amenitsch
,
H.
and
Bernstorff
,
S.
(
1997
).
Fibrillar structure and mechanical properties of collagen
.
J. Struct. Biol
.
122
,
119
122
.
Grant
D.
,
Cid
,
M.
,
Kibbey
,
M. C.
and
Kleinman
,
H.
(
1992
).
Extracellular matrix-cell interactions: Matrigel and complex cellular pattern formation
.
Lab. Invest
.
67
,
805
806
.
Grant
,
D. S.
,
Tashiro
,
K.
,
Segui-Real
,
B.
,
Yamada
,
Y.
,
Martin
,
G. R.
and
Kleinman
,
H. K.
(
1989
).
Two different laminin domains mediate the differentiation of human endothelial cells into capillary-like structures in vitro
.
Cell
58
,
933
943
.
Gullberg
,
D.
,
Turner
,
D. C.
,
Borg
,
T. K.
,
Terracio
,
L.
and
Rubin
,
K.
(
1990
).
Different beta 1-integrin collagen receptors on rat hepatocytes and cardiac fibroblasts
.
Exp. Cell Res
.
190
,
254
264
.
Hanahan
,
D.
(
1997
).
Signaling vascular morphogenesis and maintenance
.
Science
277
,
48
50
Haralabopoulos
,
G. C.
,
Grant
,
D. S.
,
Kleinman
,
H. K.
,
Lelkes
,
P. I.
,
Papaioannou
,
S. P.
and
Maragoudakis
,
M. E.
(
1994
).
Inhibitors of basement membrane collagen synthesis prevent endothelial cell alignment in matrigel in vitro and angiogenesis in vivo
.
Lab. Invest
.
71
,
575
582
.
Harris
,
A. K.
,
Stopak
,
D.
and
Wild
,
P.
(
1981
).
Fibroblast traction as a mechanism for collagen morphogenesis
.
Nature
290
,
249
251
.
Hoying
,
J. B.
,
Boswell
,
C. A.
and
Williams
,
S. K.
(
1996
).
Angiogenic potential of microvessel fragments established in three-dimensional collagen gels
.
In Vitro Cell. Dev. Biol. Anim
.
32
,
409
419
.
Ingber
,
D.
,
Fujita
,
T.
,
Kishimoto
,
S.
,
Sudo
,
K.
,
Kanamaru
,
T.
,
Brem
,
H.
and
Folkman
,
J.
(
1990
).
Synthetic analogues of fumagillin that inhibit angiogenesis and suppress tumour growth
.
Nature
348
,
555
557
.
Ingber
,
D. E.
and
Folkman
,
J.
(
1989
).
How does extracellular-matrix control capillary morphogenesis
.
Cell
58
,
803
805
.
Ingber
,
D. E.
,
Madri
,
J. A.
and
Folkman
,
J.
(
1986
).
A possible mechanism for inhibition of angiogenesis by angiostatic steroids: induction of capillary basement membrane dissolution
.
Endocrinology
119
,
1768
1775
.
Kim
,
C. S.
,
Wang
,
T.
and
Madri
,
J. A.
(
1998
).
Platelet endothelial cell adhesion molecule-1 expression modulates endothelial cell migration in vitro
.
Lab. Invest
.
78
,
583
590
.
Konerding
,
M. A.
,
Miodonski
,
A. J.
and
Lametschwandtner
,
A.
(
1995
).
Microvascular corrosion casting in the study of tumor vascularity: a review
.
Scanning Microsc
.
9
,
1233
1243
.
Korff
,
T.
and
Augustin
,
H. G.
(
1998
).
Integration of endothelial cells in multicellular spheroids prevents apoptosis and induces differentiation
.
J. Cell Biol
.
143
,
1341
1352
.
Kozian
,
D. H.
,
Ziche
,
M.
and
Augustin
,
H. G.
(
1997
).
The activin-binding protein follistatin regulates autocrine endothelial cell activity and induces angiogenesis
.
Lab. Invest
.
76
,
267
276
.
Kubota
,
Y.
,
Kleinman
,
H. K.
,
Martin
,
G. R.
and
Lawley
,
T. J.
(
1988
).
Role of laminin and basement membrane in the morphological differentiation of human endothelial cells into capillary-like structures
.
J. Cell Biol
.
107
,
1589
1598
.
Madri
,
J. A.
and
Basson
M. D.
(
1992
).
Extracellular matrix-cell interactions: Matrigel and complex cellular pattern formation
.
Lab. Invest
.
67
,
804
805
.
Madri
,
J. A.
,
Pratt
,
B. M.
and
Tucker
,
A. M.
(
1988
).
Phenotypic modulation of endothelial cells by transforming growth factor-beta depends upon the composition and organization of the extracellular matrix
.
J. Cell Biol
.
106
,
1375
8134
.
Ment
,
L. R.
,
Stewart
,
W. B.
,
Scaramuzzino
,
D.
and
Madri
,
J. A.
(
1997
).
An in vitro three-dimensional coculture model of cerebral microvascular angiogenesis and differentiation
.
In Vitro Cell. Dev. Biol. Anim
.
33
,
684
691
.
Meyer
,
M.
,
Clauss
,
M.
,
Lepple-Wienhues
,
A.
,
Waltenberger
,
J.
,
Augustin
,
H. G.
,
Ziche
,
M.
,
Lanz
,
C.
,
Büttner
,
M.
,
Rziha
,
H. J.
and
Dehio
,
C.
(
1999
).
A novel vascular endothelial growth factor encoded by Orf virus, VEGF-E, mediates angiogenesis via signalling through VEGFR-2 (KDR). but not VEGFR-1 (Flt-1). receptor tyrosine kinases
.
EMBO J
.
18
,
363
374
.
Mignatti
,
P.
,
Morimoto
,
T.
and
Rifkin
,
D. B.
(
1991
).
Basic fibroblast growth factor released by single, isolated cells stimulates their migration in an autocrine manner
.
Proc. Natl. Acad. Sci. USA
88
,
11007
11011
.
Montesano
,
R.
and
Orci
,
L.
(
1985
).
Tumor-promoting phorbol esters induce angiogenesis in vitro
.
Cell
42
,
469
477
.
Montesano
,
R.
,
Orci
,
L.
and
Vassalli
,
P.
(
1983
).
In vitro rapid organization of endothelial cells into capillary-like networks is promoted by collagen matrices
.
J. Cell Biol
.
97
,
1648
1652
.
Nehls
,
V.
and
Drenckhahn
,
D.
(
1995
).
A novel, microcarrier-based in vitro assay for rapid and reliable quantification of three-dimensional cell migration and angiogenesis
.
Microvasc. Res
.
50
,
311
322
.
Nicosia
,
R. F.
and
Ottinetti
,
A.
(
1990
).
Growth of microvessels in serum-free matrix culture of rat aorta. A quantitative assay of angiogenesis in vitro
.
Lab. Invest
.
63
,
115
122
.
Passaniti
,
A.
(
1992
).
Extracellular matrix-cell interactions: Matrigel and complex cellular pattern formation
.
Lab. Invest
.
67
,
804
.
Pepper
,
M. S.
,
Montesano
,
R.
,
Vassalli
,
J. D.
and
Orci
,
L.
(
1991
).
Chondrocytes inhibit endothelial sprout formation in vitro: evidence for involvement of a transforming growth factor-beta
.
J. Cell. Physiol
.
146
,
170
179
.
Pepper
,
M. S.
,
Spray
,
D. C.
,
Chanson
,
M.
,
Montesano
,
R.
,
Orci
,
L.
and
Meda
,
P.
(
1989
).
Junctional communication is induced in migrating capillary endothelial cells
.
J. Cell Biol
.
109
,
3027
3038
.
Ploetz
,
C.
,
Zycband
,
E. I.
and
Birk
,
D. E.
(
1991
).
Collagen fibril assembly and deposition in the developing dermis: segmental deposition in extracellular compartments
.
J. Struct. Biol
.
106
,
73
81
.
Pollman
,
M. J.
,
Naumovski
,
L.
and
Gibbons
,
G. H.
(
1999
).
Endothelial cell apoptosis in capillary network remodeling
.
J. Cell. Physiol
.
178
,
359
70
.
Preissner
,
K. T.
,
May
,
A. E.
,
Wohn
,
K. D.
,
Germer
,
M.
and
Kanse
,
S. M.
(
1997
).
Molecular crosstalk between adhesion receptors and proteolytic cascades in vascular remodelling
.
Thromb. Haemost
.
78
,
88
95
.
Risau
,
W.
(
1997
).
Mechanisms of angiogenesis
.
Nature
386
,
671
674
.
Risau
,
W.
and
Flamme
,
I.
(
1995
).
Vasculogenesis
.
Annu. Rev. Cell. Dev. Biol
.
11
,
73
91
.
Satake
,
S.
,
Kuzuya
,
M.
,
Ramos
,
M. A.
,
Kanda
,
S.
and
Iguchi
,
A.
(
1998
).
Angiogenic stimuli are essential for survival of vascular endothelial cells in three-dimensional collagen lattice
.
Biochem. Biophys. Res. Commun
.
244
,
642
646
.
Sato
,
Y.
and
Rifkin
,
D. B.
(
1988
).
Autocrine activities of basic fibroblast growth factor: regulation of endothelial cell movement, plasminogen activator synthesis and DNA synthesis
.
J. Cell Biol
.
107
,
1199
1205
.
Sephel
,
G. C.
,
Kennedy
,
R.
and
Kudravi
,
S.
(
1996
).
Expression of capillary basement membrane components during sequential phases of wound angiogenesis
.
Matrix Biol
.
15
,
263
279
.
Vernon
,
R. B.
,
Angello
,
J. C.
,
Iruela-Arispe
,
M. L.
,
Lane
,
T. F.
and
Sage
,
E. H.
(
1992
).
Reorganization of basement membrane matrices by cellular traction promotes the formation of cellular networks in vitro
.
Lab. Invest
.
66
,
536
547
.
Vernon
,
R. B.
and
Sage
,
E. H.
(
1999
).
A novel, quantitative model for study of endothelial cell migration and sprout formation within three-dimensional collagen matrices
.
Microvasc. Res
.
57
,
118
133
.
Vernon
,
R. B.
and
Sage
,
E. H.
(
1996
).
Contraction of fibrillar type I collagen by endothelial cells: a study in vitro
.
J. Cell. Biochem
.
60
,
185
197
.
Vernon
,
R. B.
and
Sage
,
E. H.
(
1995
).
Between molecules and morphology. Extracellular matrix and creation of vascular form
.
Am. J. Pathol
.
147
,
873
883
.
Vernon
R. B.
and
Sage
,
E. H.
(
1992
).
Extracellular matrix-cell interactions: Matrigel and complex cellular pattern formation
.
Lab. Invest
.
67
,
807
808
.
Villaschi
,
S.
and
Nicosia
,
R. F.
(
1993
).
Angiogenic role of endogenous basic fibroblast growth factor released by rat aorta after injury
.
Am. J. Pathol
.
143
,
181
190
.