Dynamic cell rearrangements shape the cranial vascular network of developing Zebrafish embryos

To form the complex network of endothelial tubes making up the vasculature, a number of vessels have to interact and connect to each other during development. This involves the transformation of blunt-ended angiogenic sprouts into interconnected functional tubes, a process called vessel fusion or anastomosis. While much is known about vessel sprouting, little is known about vessel fusion at the cellular and molecular levels. Most of the vessels in the developing vertebrate embryo form in the presence of stable blood flow in adjacent tubes, suggesting the importance of flow and/or blood pressure for angiogenic sprouting and anastomosis. For this reason, my analyses focused on the head vasculature where many vessels form in the presence of stable blood flow in the zebrafish embryo.
In this study I performed detailed analyses of different fusion events in cranial vessels of the developing zebrafish embryo. Using novel transgenic tools and high resolution live imaging I defined a multistep model of vessel fusion and showed that it is conserved in various vascular beds, regardless of vessel shape and the age of the embryo. I also showed that in all the cranial vessels I studied, the initial fusion steps are the same and involve de novo deposition of junctional proteins, ZO-1 and VE-cadherin, in a form of a junctional spot, which subsequently elaborates into a ring, followed by de novo apical membrane insertion. Lumen formation in the newly formed vessel takes place through blood pressure-dependent luminal/apical cell membrane invagination and fusion of apical membranes, leading to a continuous lumen. During this process, the tip cells become unicellular/seamless tubes with transcellular lumen. I found that such newly connected vessels subsequently undergo dynamic cellular rearrangements that lead to the transformation of the unicellular tubes into multicellular ones. This transformation involves cell splitting, a novel cellular mechanism that, to our knowledge, has not been described before in branching morphogenesis of any organ. Additionally, I analyzed the fusion process in VE-cadherin deficient embryos and showed that this adhesion molecule is necessary for formation of a single contact surface between the fusing vessel sprouts and thus, has an important role in coordinating anastomosis.
I have also analyzed vessel regression during vascular pruning and I showed that it follows a multistep process involving dynamic cell rearrangements that resemble “reversed” vessel fusion. These analyses represent the first studies of vessel remodeling at the cellular level in an in vivo system.