Open Access

Michele Marass

• Cardiovascular development requires two main steps, vasculogenesis and angiogenesis. During vasculogenesis, angioblasts, the precursors of endothelial cells (ECs), specify from the mesoderm and coalesce to form the axial vessels of the vertebrate embryo. Many questions regarding the transcriptional waves initiating and sustaining angioblast specification are still unanswered. The identity of cloche, a gene essential for EC differentiation in zebrafish, was only recently discovered by our group, and very little is known about its upstream regulators or its molecular mechanism of action. I described the molecular players involved in orchestrating npas4l expression, upstream of angioblast specification. By using genetic models and chemical treatments, I identified FGF-Erk axis and BMP signaling to be involved in npas4l regulation. I also showed that eomesa is a potent inducer of npas4l expression. In addition, in vitro experiments indicated that murine Eomes promotes EC specification, acting upstream of Etv2 and Tal1. Using a combination of gain-of-function and loss-of-function models for npas4l, I identified primary and secondary downstream effectors of npas4l. I showed that Npas4l binding sites are present in the promoter of genes involved in hematoendothelial specification, such as tal1, lmo2 and etv2. Importantly, I reported that npas4l is sufficient and necessary to promote the EC specification program. By performing a combined analysis of the developed datasets, I recovered putative genes with a potential role in EC specification. One of the most promising candidates was tspan18b. I generated a mutant allele for tspan18b and observed angiogenic defects in tspan18b-/- embryos, confirming a role for this gene in zebrafish cardiovascular development. I showed that Npas4l binds etv2 promoter in zebrafish. In mammalian embryonic stem cells, however, Etv2 promoter is bound by HIF-1α, a transcription factor homolog to Npas4l. Interestingly, Eomes knockdown in vitro lead to a significant reduction of Hif-1α expression. To test the function of Hif-1α in vivo, I took advantage of a murine loss-of-function model. Hif-1α mouse mutant embryos exhibit a significant decrease in Etv2 expression, when compared with WT siblings. These data suggest a model where mammals lost npas4l during evolution and HIF-1α acquired a new function, replacing npas4l role in EC specification. I compared the phenotype of Hif-1α mouse mutant with zebrafish hif-1α loss-of- function models. Importantly, zebrafish hif-1α mutant did not show defects in vasculogenesis or EC specification, but in EC specialization, during HSC development. I showed that hypoxia is a potent inducer of HSC formation, and hif-1α as well as hif-2α act upstream of notch1, vegfaa and evi1 in hemogenic endothelial specification. Conclusions In this work, I explored the molecular mechanisms underlying EC specification in vertebrates, analyzing the role of bHLH-PAS transcription factors in this biological process. I identified the upstream regulators and the downstream effectors of npas4l, describing a novel role for tspan18b in zebrafish cardiovascular development. Npas4l is a transcription factor necessary and sufficient for angioblast differentiation in zebrafish, but the gene was lost in the mammalian lineage. hif-1α and hif-2α, paralogous genes of npas4l, are involved in the establishment of EC heterogeneity and specifically in the specification of hemogenic endothelium in zebrafish. Murine Hif-1α, however, is responsible for Etv2 regulation, indicating a role for hypoxia inducible factor in initiating the EC specification program in mouse, similarly to npas4l function in zebrafish.