Vascular Network Formation in Expanding versus Static Tissues: Embryos and Tumors

Abstract

In this perspectives article, we review scientific literature regarding de novo formation of vascular networks within tissues undergoing a significant degree of motion. Next, we contrast dynamic pattern formation in embryos to the vascularization of relatively static tissues, such as the retina. We argue that formation of primary polygonal vascular networks is an emergent process, which is regulated by biophysical mechanisms. Dynamic empirical data, derived from quail embryos, show that vascular beds readily form within a moving extracellular matrix (ECM) microenvironment-which we analogize to the de novo vascularization of small rapidly growing tumors. Our perspective is that the biophysical rules, which govern cell motion during vasculogenesis, may hold important clues to understanding how the first vessels form in certain malignancies.

Keywords: cell motility; multicellular sprouting; tissue movements; vascular patterning.

Figure 1.

Tagging of endogenous vascular endothelial growth factor (VEGF) with fluorescent probes. Live quail embryos (Hamburger Hamilton [HH] 7) were injected, at 4 positions (asterisks), with 20-nL boluses of recombinant human VEGFR2 coupled to human IgG Fc (rVEGFR2), approximately 1 ng/nL. After 45 minutes, the embryo was fixed, made permeable, and incubated with secondary antibodies to human Fc (green) and QH1 antibodies directly conjugated with Alexa 555 (red). The red signal thus depicts quail endothelial cells. The image shows that upon injection, the rVEGFR2 binds to dense punctate foci of ligand near the injection sites. The diffusion of the rVEGFR2 is spatially limited by the high concentration of reactive VEGF ligand. No network-like pattern is discernible when viewing the green signal. Note that endothelial cells near the injection sites are not configured in a normal polygonal-like pattern when compared to the contralateral side. That rVEGFR2 can perturb vascular patterns, in vivo, is a well-established observation. This is direct evidence that in a live amniote embryo, VEGF, which is physically accessible, is not distributed in a polygonal pattern, nor is VEGF distributed in a manner that suggests a higher order chemotactic prepattern. n = notochord; scale bar: 100 µm.

Figure 2.

Endothelial cell movements at the onset of vasculogenesis in a Hamburger Hamilton (HH) stage 7 (5 somite) transgenic quail embryo. Motion of TIE1+ nuclei is shown in a somite-attached reference system by projecting 4 consecutive frames, the first 3 in red and the most recent time point in yellow. The asterisk marks an area where mediocranial motion is especially apparent. H = Hensen’s node; n = notochord. Scale bar: 100 µm.

Figure 3.

Tissue displacement vector components can be estimated from various microscopy modes, including differential interference contrast (DIC) and extracellular matrix (ECM) immunofluorescence. Embryonic development was recorded using multiple optical modes: DIC and two epifluorescent channels, visualizing fibronectin and fibrillin-2, two distinct ECM components. Particle image velocimetry analysis was performed on all 3 image sequences, yielding 3 estimates for tissue movement—one from each imaging mode. The correlation plot compares corresponding components of tissue displacement vectors derived either from DIC and fibrillin-2 immunofluorescence (A) or fibronectin and fibrillin-2 immunofluorescence (B). Blue, green, and red indicate progressively higher data densities in the correlation plot. High densities along the diagonal of the correlation plots indicate that analysis of all 3 optical modes yields consistent estimates.

Figure 4.

Active movements of endothelial cells. Data presented in Figure 2 were digitally transformed to remove the effect of tissue movements. The difference between the apparent cell motion and local tissue motion is the active cell movement, which is more random. Marked locations correspond to those in Figure 2.

Figure 5.

Active cell movements in the nascent vascular plexus of HH8 (6 somite) embryos. Red and green colors were assigned depending on the magnitude of active cell movements to a QH1-immunolabeled image. Fluorescence sources that move with the tissue environment are colored green. In contrast, QH1 foci that move relative to the surrounding tissue (i.e., display active motility) are colored red. Vascular sprouts are locations of intense cellular motility.

Figure 6.

Active movement of TIE1+ nuclei, obtained after digitally correcting for the deformations associated with tissue motion (left). Two consecutive frames, separated by 8 minutes, are shown—the first as red, the second as green. Autonomous cell movement is inhomogeneous: some nuclei do not move (appear as yellow; some are marked by circles), whereas most cells move in a chain-migration fashion (indicated by arrows). At this stage of development, movement directions are highly variable: even in the same vascular segment, groups/chains are seen moving in opposite directions. The right panel marks the location of the area shown in the embryo. The fourth somite and Hensen’s node is marked. Scale bar: 100 µm. After Sato et al.; see also Supplementary Video 3.

Figure 7.

Computational model of multicellular sprout elongation. A leader cell (yellow) is assumed to move randomly with a persistent polarity; remaining cells (red) are assumed to prefer adhesion to elongated cells instead of to well-spread cells. This preference helps cells to leave the initial aggregate and enter the sprout. After Szabo and Czirok; see also Supplementary Video 4.