Previous reports have described ESC

Previous reports have described ESC-based differentiation in collagen gels to study the developmental events of vasculogenesis and angiogenesis (Feraud et al., 2001; Hermant et al., 2007). However, these were not optimized for the assessment of more fundamental aspects of vessel induction, patterning, and remodeling, nor were these assays standardized into a 96-well plate format suitable for screening. In some previous reports, EB size was not controlled (Feraud et al., 2001; Hermant et al., 2007), which is essential for obtaining the consistency and reproducibility required in a screen. Additionally in these assays, multiple growth factors for vascular induction were used (Feraud et al., 2001; Hermant et al., 2007), or high concentrations of VX-765 were used for EB formation (Jakobsson et al., 2006), which increases variability. Building on the foundation of these studies, we have standardized and simplified the culture system and employed a fluorescent reporter to allow easy monitoring of morphogenesis, thus producing a more robust assay suitable for drug screens in the mouse system. Future studies using human pluripotent cell lines, aided by the advances in genome-editing technologies, will allow the use of more robust reporter lines for endothelial differentiation in the human system. Our assay was validated using NOTCH and FLK-1 inhibitors, since disruption of these pathways results in visible alterations in angiogenesis (Hellstrom et al., 2007; Shalaby et al., 1995). By screening a small-molecule kinome library we expected a large number of hits, given that the vasculature is very sensitive to signaling pathway disruption. We identified many kinase targets with well-established roles in angiogenesis, including RTKs (VEGFR, PDGFR, FGFR, TIE2, FLT-3, c-MET, and IGF1R) as well as their downstream effectors including RAF, MEK, and ERK, further validating our screen. JAK, ALK, ALK5, and AURORA were also hits and have well-established roles in regulating angiogenesis. It is important to note that despite the fact that our screen is designed to detect both promoters and inhibitors of angiogenesis, all of our validated hits inhibit angiogenesis. It is possible that the NOTCH pathway may be unique in causing excessive sprouting. Interestingly, inhibition of ALK1 has also been shown to lead to excessive angiogenic sprouting, which was attributed to cooperation of ALK1 with the NOTCH pathway (Kerr et al., 2015). An additional screen of a more broad-based library similarly showed that only NOTCH inhibitors resulted in excessive angiogenic sprouting (data not shown). Screening of this second library showed that our assay is sensitive to phenotypic changes that were measurable beyond just increases or decreases in the number of FLK-1+ sprouts (i.e., retinoids had no major effect on sprout number but caused morphological changes in vessel shape). This suggests that the complete landscape of target space that can be explored with this assay is still to be determined. We identified RSK and TTK as angiogenic modulators. We showed that treatment of EBs or HUVECS with BI-D1870 and BIX-RSK2, the selective RSK inhibitors, or with AZ3146, the selective TTK inhibitor, inhibited angiogenic sprouts in EBs and network formation in HUVECs, and disrupted the preformed HUVEC tubes and the preformed EB angiogenic sprouts. It is important to note that these inhibitors disrupted network formation in HUVECs induced by bFGF without VEGF supplementation, suggesting that they are downstream of multiple proangiogenic pathways. Western blot analysis of EBs showed that TTK and RSK inhibitors resulted in a significant decrease in phosphorylation of the downstream targets, SMAD2 and LKB1, respectively, in association with the observed decrease in angiogenesis. A previous report has suggested the involvement of RSK in angiogenesis, although no direct evidence was provided (Hayashi et al., 2005). Our study provides direct proof that RSK and TTK regulate angiogenesis.