Zebrafish Vascular Development: General and Tissue-Specific Regulation

Blood vessels form a network to deliver blood cells and plasma to peripheral organs and tissues, through arteries and arterioles, and function as drainage system from peripheral organs and tissues, through capillaries, venules, and veins.1, 2 The initial blood vessel network is formed by endothelial cells (ECs). Subsequently, mural cells, including pericytes and smooth muscle cells, emerge close to the EC-derived network to support blood vessels.3 Furthermore, the adventitia surrounds the smooth muscle cell layer in arteries.

EC precursor cells, or angioblasts, first assemble and differentiate into ECs to form a primary vascular plexus through a process called vasculogenesis.4 Those primitive ECs then give rise to either arterial ECs or venous ECs. Two types of EC precursors appear in the trunk and in the tail in zebrafish (Fig. 1, unpublished data). These 2 types of ECs emerge distinctly from the lateral plate mesoderm and the tail bud.5, 6 There is another type of segregation: artery and vein. The primitive vascular cord is divided into the dorsal aorta (DA) primordium and cardinal vein primordium in the trunk vessel of zebrafish.5, 7 In addition, the other type of arterial EC differentiation from venous ECs is reported in the tail fin of zebrafish.8 The differentiation of angioblasts into ECs and that of primitive ECs into veins/arteries are predetermined by the hierarchy of gene expression. The gene-encoded molecules determine the fate and character of ECs.

Fig. 1
Two types of emergence of endothelial precursors in the trunk and tail of the zebrafish embryo. A Tg(fli1:EGFP) embryo is imaged at 22 hpf. The endothelial cell precursors in the trunk and those in the tail appear simultaneously, but separately. The boxed region of the scheme of the embryo is imaged.
hpf, hours post-fertilization.

• Click for larger image
• Download as PowerPoint slide

Zebrafish are widely used to clarify vascular development because of their translucency, the early growth of the circulatory system, their small size, and their extra-maternal growth. Importantly, many studies have shown that the morphological and molecular mechanisms involved in vascular development are conserved between zebrafish and humans.9, 10 These advantages enable developmental biologists to study circulatory system development using several kinds of live-imaging microscopes, including confocal, spinning disk, light-sheet, and super-resolution microscopes. In addition, multiple useful transgenic zebrafish lines for the fluorescent imaging of EC movement and signaling have been developed. These imaging techniques enable simultaneous investigations of the morphological changes and physiological functions of ECs using living embryos; these changes are triggered by EC-specific ligand/receptor-mediated intracellular signaling cascades. Thus, we review these EC-specific signaling and signal-dependent responses shown in zebrafish.

Recently, EC heterogeneity and organ/tissue-specific ECs have become a topic of interest in vascular biology and regenerative medicine.11 For example, structural and cell biological differences are present among sinusoidal ECs, fenestrated ECs, and high ECs, which are found in the liver, endocrine organs, and lymph nodes, respectively.12 Venous ECs and arterial ECs have specific characteristics in veins and arteries. During embryogenesis, tissue/organ-specific cell-cell interactions between parenchymal cells and ECs are regulated by angiocrine and paracrine mechanisms.13 Moreover, the blood flow in lumenized blood vessels also affects the heterogeneity of ECs, indicating the importance of the environment of the surrounding blood vessels in addition to cell-autonomous regulation of ECs.

Beyond organ-specific blood vessel pathways, blood vessel formation follows a common pattern (i.e., angiogenesis). ECs sprout from pre-existing vessels; form branches by proliferation, migration, and intercellular adhesions; and thereby form new lumenized vessels.14 Previous studies have tried to identify the molecular mechanisms underlying angiogenesis and maintenance/maturation in various tissues/organs.15, 16, 17, 18 However, some issues have yet to be fully clarified, including the origin of ECs, vascular niches, and tissue-specific mechanisms of angiogenesis.19 We summarize the current understanding of vascular development and highlight the unanswered issues of vascular development in zebrafish.

Circulation requires the heart, vessels, and blood as a pump, conduit, and carrier/content, respectively. The cardiomyocytes of the heart, ECs and mural cells of blood vessels, and blood cells are derived from mesodermal cells. The cooperative development of these 3 cell types efficiently and promptly starts the circulation to deliver oxygen and nourish parenchymal cell growth in the peripheral tissues and organs. At the very early stage of embryogenesis before the start of circulation, oxygen delivery depends upon diffusion.

The cooperation of blood cells and ECs is shown in hemogenic ECs, which are the origin of both hematopoietic stem and progenitor cells (HSPCs) and ECs.20 Another aspect of the cooperation between these 2 cell types is that vascular ECs form a vascular niche for HSPCs in the caudal hematopoietic tissue (CHT) and the kidney marrow in zebrafish.21 The molecular mechanism underlying the interaction of HSPCs and ECs has become a topic of interest in stem cell biology.

The coordinated growth of the myocardium and coronary vessels of the heart that supply oxygen to cardiomyocytes has been explored (Fig. 2, left). Endocardial ECs sprout and grow on the surface of the heart according to chemokine secretion from the myocardium.22 Endocardial ECs also become the source of blood cells as the hemogenic endothelium in the heart. Following coronary vessel development, lymphatic vessels grow along pre-existing coronary vessels on the surface of the heart.23 Of note, cxcr4a mutants, which lack coronary arteries, fail to exhibit lymphatic vessel growth. Therefore, the development of the myocardium, coronary vessels, and lymphatic vessels is regulated in a coordinated manner in the developing heart.

Fig. 2
Vessels of the heart. The surface coronary vessels of a zebrafish heart at 1.5 mpf is visualized by fluorescence of Tg(fli1:EGFP) (left). The outflow tract (single arrow) and inflow tract (double arrows) of endocardial endothelial cells (red) are imaged with myocytes (green) using Tg(fli1:myristoylated-mCherry);(myl7:GFP-tagged with CAAX motif of small GTPase Ras) (right).
Ao, aortic root; CCV, common cardinal vein; mpf, months post-fertilization; hpf, hours post-fertilization; dpf, days post-fertilization.

• Click for larger image
• Download as PowerPoint slide

To pump out blood from the heart, the heart and great vessels must be connected (Fig. 2). However, it remains unclear how ECs connect to the endocardial cells of the heart (Fig. 2, right). The endocardial ECs of the outflow of the heart are likely to be connected to the ECs of the aortic arch, while those of the inflow tract are connected to the common cardinal vein (CCV). Although the migration and lumenization of ECs in the CCV have been investigated, as a process referred to as “lumen ensheathment,”24 it remains elusive how endocardial ECs are connected to CCV ECs. These 2 connections allow the heart to pump out blood and establish the closed circulation in zebrafish.

Previous studies mainly focused on the trunk blood vessels, such as the DA and posterior cardinal vein (PCV) (Fig. 3, top right). Angiogenic sprouting from the DA and PCV is categorized as primary sprouting and secondary sprouting.10 The primary sprouts initially form inter-somitic vessels (ISVs) and the vessels that reach the most dorsal part form dorsal lateral anastomosis vessels (DLAVs). The secondary sprouts connecting to the preformed ISVs become venous ISVs (vISVs), whereas unconnected ISVs become arterial ISVs (aISVs). Thus, blood flow follows the path of DA-aISV-DLAV-vISV-PCV. The primary sprouts are mainly regulated by vascular endothelial growth factor (VEGF)-A and VEGF receptor (VEGFR) 2, while the secondary sprouts are promoted by VEGF-C and VEGFR3.25, 26, 27

Fig. 3
Vessels in the brain, PAA, trunk, and CHT. The vessel images were obtained from the transgenic fish embryos indicated at the top or bottom of the panels.
Yellow arrowheads point to vessel protrusion. White arrowheads and white arrows indicate the primary sprouts from the DA and secondary sprouts from the PCV, respectively.

PAA, pharyngeal aortic arch; DA, dorsal aorta; PCV, posterior cardinal vein; ISV, inter-somitic vessels; CHT, caudal hematopoietic tissue; dpf, days post-fertilization; hpf, hours post-fertilization.

• Click for larger image
• Download as PowerPoint slide

The vessels in the posterior part (i.e., the caudal vessels) have also been characterized. The caudal vein plexus (CVP) changes its shape and differentiates into the caudal vein. ECs of the CVP proliferate in response to bone morphogenetic protein (Bmp).28 We have demonstrated that Bmp-induced β-catenin-dependent signaling is activated in the ECs that preferentially differentiate into the caudal vein.29 During somitogenesis, both arteries and veins need to extend posteriorly to oxygenate the posterior trunk and tail. We have previously shown that new ECs emerge in the tail during somitogenesis in addition to the proliferation of the pre-existing ECs that are located anteriorly.30 Of particular note, the extension of the DA towards the tail depends upon the proliferation of the pre-existing ECs of the DA (unpublished data), whereas the origin of new venous ECs that contribute to the extension of trunk vein remains unclear. Interestingly, the CHT develops in parallel with venous development in the tail (Fig. 3, bottom right). Venous, but not arterial, ECs of the most dorsal CVP function as a vascular niche for HPSCs.31 HPSCs that originate from the hemogenic ECs of the DA move to the CHT and finally migrate into the kidney marrow, which mirrors the function of the bone marrow in mammals, and remain there to produce various kinds of blood cells in juveniles.21 Therefore, the transient vascular niche for HPSCs in the CHT consists of characteristic ECs that appear only for this period. The difference between these dorsal venous ECs and the ventral venous ECs of the CVP, which are β-catenin signaling-positive cells, is unclear. Therefore, the origin and function of these vascular niche ECs should be characterized in the future.

The blood vessels in the brain form a network to oxidize and nourish neural tissues. In the brain, a paired lateral DA is established and connected to the posterior DA.32 During this period, the first primitive veins, the primordial hindbrain channels (PHBCs), develop. Following PHBC formation, the basilar artery and central arteries assemble in the hindbrain (Fig. 3, bottom left).15, 18 The blood-brain barrier (BBB) is a specific anatomical and functional structure between ECs and neurons that restricts the infiltration of inflammatory cells into the nervous system. In contrast, it efficiently exchanges molecules that are required for active metabolism in neurons and glial cells, including astrocytes, oligodendrocytes, and microglia. Astrocytes participate in forming the BBB, whereas other cells—including pericytes and capillary ECs—need to assemble together with astrocyte feet to form the BBB. We have previously shown that pericytes in the brain emerge after the establishment of the network of ECs.33 The feet of pericytes seem to seal the junctions of ECs. However, we have not simultaneously monitored the assembly of ECs, pericytes, and astrocytes in the brain. Thus, more advanced in vivo imaging will help understand how the BBB functions to exchange molecules in the neural tissues.

In addition to exchanging molecules in the brain, the glymphatic system in the mouse has become an attractive focus of research into the clearance of wastes from the brain.34 In zebrafish, waste from the central nervous system (CNS) can be removed through 2 pathways, the meningeal lymphatic vessels and the glymphatic system, which drains glial water and functions like the lymphatic system.35 Upon injury of cerebrovascular system, lymphatic vessel invasion precedes the regeneration of blood vessels.36 During the investigation of meningeal lymphatic vessels, mural lymphatic ECs (muLECs), also termed brain lymphatic ECs, were identified as a novel lymphatic cell type.35, 37 The muLECs express lymphatic markers, but do not form vessels. Therefore, the function of mural lymphatic vessels should be clarified in the clearance of waste products from the CNS.38

The aortic and pharyngeal arch artery (PAA) of zebrafish is unique, because zebrafish lack pulmonary circulation, which is required for mammals. The ECs of the PAA were shown to come from Nkx2.5 and Tcf21 double-positive mesodermal cells.39, 40 However, depletion of these progenitors of the pharyngeal arch does not result in deformation of the PAA, suggesting a compensatory function of ECs from other vasculature. The shape of the PAA is different from that found in sprouting angiogenesis, which is commonly regulated by VEGF-A. The branching or sprouting of pre-existing vessels is not observed during PAA formation, whereas protrusion of ECs is found (Fig. 3, top left, unpublished data). Edn1 and hand2 might be involved in the outgrowth of ECs to form a bulging protrusion of pre-existing vessels.41

The development of the eye requires 2 vascular systems: the intraocular hyaloid vessels and the superficial choroidal vessels.42 The former vessels are transient and become retinal vessels. On the surface of the zebrafish eye, venous-derived superficial vessels develop prior to the formation of an elaborate network of choroidal vessels. Live imaging has revealed that ECs sprouting from the venous primordial midbrain channel form arteries through Notch signaling and chemokine signaling.43

Amputation and crush injury models have been developed to study fin regeneration.8, 44 Injury triggers tissue vascularization to oxygenate tissues. Like other organs and tissues, revascularization of injured bones has been investigated. In the caudal fin amputation model, the tip cells are derived from venous ECs and move backward to form the plexus. Therefore, the arteries of regenerated fins are derived from vein-derived tip cells. There is a striking difference between amputated and crushed fins. In the bone crush injury model, the main arteries in the intra-ray tissues are damaged. At a later stage, mis-patterned blood vessels are found in the crushed ray, probably due to the prolonged bone repair. These fin regeneration models may provide crucial insights into the relationship between tissue repair and vascular regeneration.

Internal organs such as the intestine, liver, and pancreas must be vascularized during development. The gut needs to efficiently absorb nutrition, which is transferred to the liver through the portal vein. The sinusoidal ECs have specific structures for nutrition absorption.12 Moreover, as an endocrine organ, the pancreas has endocrine organ-specific capillary characteristics. In zebrafish, blood vessels in these organs originate from the floor of the PCV. The initial plexus that covers the internal organs is referred to as the sub-intestinal plexus.17 Sub-intestinal vessel growth is first regulated by Bmp and secondarily by VEGF. Interestingly, the blood vessels in the internal organs are derived from venous ECs, like the superficial vessels of the eye and arteries of the fin after amputation.

ECs are required for parenchymal cell development and, conversely, parenchymal cells produce paracrine factors to stimulate maturation of the blood vessel network.12 For example, tissue hypoxia induces angiogenesis by producing angiogenic signaling (VEGF). In turn, the tissue-specific paracrine system, from ECs to parenchymal cells, is important for organ and tissue development or maintenance. Therefore, this system is called angiocrine.11 In zebrafish, ECs control hepatocyte polarization during liver development.45 There are many angiocrine cues from ECs toward parenchymal cells in several organs.11 For this reason, the importance of organotypic ECs has been highlighted in recent research.

Several conserved signals are mediated by ligand-receptor or mechano-transduction in ECs (Fig. 4). Some are commonly used beyond the vascular beds, while others are tissue-specifically or EC type-specifically regulated. In this section, we summarize those angiogenic and flow-mediated signals in zebrafish.

Fig. 4
The main signals that regulate vascular development and stabilization. These signals include tyrosine kinase-activating signaling (VEGF-VEGFR and Ang-Tie) and GPCR-mediated signaling (Apln-Aplnr and Cxcl12–Cxcr4). In addition to these receptor-mediated signaling pathways, mechanosignaling induced by blood flow (shear stress) also regulates the intracellular signaling of ECs, including Klf2 expression and nuclear translocation of Yap. PDGF-b released from ECs activates its PDGFR expressed on mural cells. Of note, Ang1 is released from mural cells to activate Tie2 expressed on ECs to regulate mutual interactions between ECs and mural cells. VE-cadherin is an endothelial-specific cadherin that is essential for adherens junction formation. Adherens junctions are strengthened by the actin cytoskeleton to maintain tissue integrity. The ECM can activate various signaling pathways directly to regulate vascular development.
VEGF, vascular endothelial growth factor; VEGFR, vascular endothelial growth factor receptor; Ang, angiopoietin; Apln, apelin; Aplnr, apelin receptor; EC, endothelial cell; VE, vascular endothelial; ECM, extracellular matrix; PDGF, platelet-derived growth factor; PDGFR, platelet-derived growth factor receptor; GPCR, G protein-coupled receptor.

• Click for larger image
• Download as PowerPoint slide

The usefulness of zebrafish when studying embryogenesis has been widely accepted in vascular biology. Herein, we summarize many reports demonstrating the essential roles of molecular pathways involved in general or tissue-specific vascular development. We emphasize again the important point that using zebrafish embryos enables us to visualize morphological alterations and cell signaling simultaneously. Molecular probes can be easily introduced into the genome, enabling the establishment of useful transgenic zebrafish. In addition, recent advancement of gene manipulation using TALEN or CRISPR/Cas9 systems has allowed researchers to test requirements and compensatory mechanisms in a tissue- or cell-type-specific manner. Using this excellent model system, we can continue to investigate how creatures live from fertilization to the end of life under conditions of environmental changes.