Vessel sprouting by migrating tip and proliferating stalk endothelial cells (ECs) is controlled by genetic signals (such as Notch), but it is unknown whether metabolism also regulates this process. Here, we show that ECs relied on glycolysis rather than on oxidative phosphorylation for ATP production and that loss of the glycolytic activator PFKFB3 in ECs impaired vessel formation. Mechanistically, PFKFB3 not only regulated EC proliferation but also controlled the formation of filopodia/lamellipodia and directional migration, in part by compartmentalizing with F-actin in motile protrusions. Mosaic in vitro and in vivo sprouting assays further revealed that PFKFB3 overexpression overruled the pro-stalk activity of Notch, whereas PFKFB3 deficiency impaired tip cell formation upon Notch blockade, implying that glycolysis regulates vessel branching.

Patterning of functional blood vessel networks is achieved by pruning of superfluous connections. The cellular and molecular principles of vessel regression are poorly understood. Here we show that regression is mediated by dynamic and polarized migration of endothelial cells, representing anastomosis in reverse. Establishing and analyzing the first axial polarity map of all endothelial cells in a remodeling vascular network, we propose that balanced movement of cells maintains the primitive plexus under low shear conditions in a metastable dynamic state. We predict that flow-induced polarized migration of endothelial cells breaks symmetry and leads to stabilization of high flow/shear segments and regression of adjacent low flow/shear segments.

SummaryThe correct development of blood vessels is crucial for all aspects of tissue growth and physiology in vertebrates. The formation of an elaborate hierarchically branched network of endothelial tubes, through either angiogenesis or vasculogenesis, relies on a series of coordinated morphogenic events, but how individual endothelial cells adopt specific phenotypes and how they coordinate their behaviour during vascular patterning is unclear. Recent progress in our understanding of blood vessel formation has been driven by advanced imaging techniques and detailed analyses that have used a combination of powerful in vitro, in vivo and in silico model systems. Here, we summarise these models and discuss their advantages and disadvantages. We then review the different stages of blood vessel development, highlighting the cellular mechanisms and molecular players involved at each step and focusing on cell specification and coordination within the network.

Sprouting angiogenesis drives blood vessel growth in healthy and diseased tissues. Vegf and Dll4/Notch signalling cooperate in a negative feedback loop that specifies endothelial tip and stalk cells to ensure adequate vessel branching and function. Current concepts posit that endothelial cells default to the tip-cell phenotype when Notch is inactive. Here we identify instead that the stalk-cell phenotype needs to be actively repressed to allow tip-cell formation. We show this is a key endothelial function of neuropilin-1 (Nrp1), which suppresses the stalk-cell phenotype by limiting Smad2/3 activation through Alk1 and Alk5. Notch downregulates Nrp1, thus relieving the inhibition of Alk1 and Alk5, thereby driving stalk-cell behaviour. Conceptually, our work shows that the heterogeneity between neighbouring endothelial cells established by the lateral feedback loop of Dll4/Notch utilizes Nrp1 levels as the pivot, which in turn establishes differential responsiveness to TGF-β/BMP signalling.

23 24How vascular tubes build, maintain and adapt continuously perfused lumens to meet 25 local metabolic needs remains poorly understood. Recent studies showed that blood 26 flow itself plays a critical role in the remodelling of vascular networks 1,2 , and 27 suggested it is also required for lumenisation of new vascular connections 3,4 . 28 However, it is still unknown how haemodynamic forces contribute to the formation of 29 new vascular lumens during blood vessel morphogenesis. 30Here we report that blood flow drives lumen expansion during sprouting angiogenesis 31 in vivo by inducing spherical deformations of the apical membrane of endothelial 32 cells, in a process that we termed inverse blebbing. We show that endothelial cells 33 react to these membrane intrusions by local and transient recruitment and contraction 34 of actomyosin, and that this mechanism is required for single, unidirectional lumen 35 expansion in angiogenic sprouts. 36Our work identifies inverse membrane blebbing as a cellular response to high external 37 pressure. We show that in the case of blood vessels such membrane dynamics can 38 drive local cell shape changes required for global tissue morphogenesis, shedding 39 light on a pressure-driven mechanism of lumen formation in vertebrates. 40 41Blood vessels form a vast but highly structured network that pervades all organs in 42 vertebrates. During development as well as in pathological settings in adults, vascular 43 networks expand through a process known as sprouting angiogenesis. New blood 44 vessels form from the coordinated migration and proliferation of endothelial cells into 45 vascular sprouts. Subsequent fusion of neighbouring sprouts, defined as anastomosis, 46 then leads to the formation of new vascular loops, whose functionality relies on their 47 successful lumenisation and perfusion 5 . During anastomosis, endothelial lumens form 48 both through apical membrane invagination into single anastomosing cells 49 (unicellular lumen formation), and through de novo apical membrane formation at 50 their nascent junction (multicellular lumen formation) 3,4 . Since the tip of endothelial 51 sprouts can be occupied by either one or several cells as they compete for the tip 52 position 6,7 , we asked whether similar mechanisms of lumen formation apply to 53 unicellular and multicellular endothelial sprouts prior to anastomosis. 54Using a zebrafish transgenic line expressing an mCherry-CAAX reporter for 55 endothelial plasma membrane (Tg(kdr-l:ras-Cherry) s916 ), we imaged lumen formation 56 in tip cells as they sprout from the dorsal aorta (DA) to form the intersegmental 57 vessels (ISVs) from 30 hours post-fertilisation (hpf). We found that lumens expand in 58 sprouting ISVs prior to anastomosis, and do so by invagination of the apical 59 membrane either into single tip cells, or along cell junctions when the tip of a 60 sprouting ISV is shared between several cells (Fig. 1a,b). 61To test if this mechanism of lumen formation is conserved in other vertebrates, we 62 perfo...