Endothelial cell migration and proliferation are essential for the establishment of a hierarchical organization of blood vessels and optimal distribution of blood. However, how these cellular processes are quantitatively coordinated to drive vascular network morphogenesis remains unknown. Here, using the zebrafish vasculature as a model system, we demonstrate that the balanced distribution of endothelial cells as well as the resulting regularity of vessel caliber, is a result of migration of cells from veins connected to arteries and cell proliferation in veins. We identify the Wiskott-Aldrich Syndrome protein (WASp) as an important molecular regulator of this process and show that loss of coordinated migration from veins to arteries upon wasb depletion results in aberrant vessel morphology and the formation of persistent arteriovenous shunts. We demonstrate that WASp achieves its function through the coordination of junctional actin assembly and PECAM1 recruitment and provide evidence that this is conserved in human. Together, we demonstrate that functional vascular patterning in the zebrafish trunk utilizes differential cell movement regulated by junctional actin, and that interruption of differential migration may represent a pathomechanism in vascular malformations.

Segmentation and tracking of endothelial nuclei

Segmentation and tracking of nuclei from live imaging in zebrafish and cultured HUVECs was performed with Imaris Image Tracking Package (Bitplane). Tracking data was formatted according to community standards for open cell migration data (Gonzalez-Beltran et al., 2020). For each zebrafish the cell tracking data was oriented in three-dimensional space such that the aorta is parallel to the x-axis and the DLAV in the positive-y half plane. In this way, the y-component of the transformed trajectories serve as a read-out for the ventral to dorsal positions of the nuclei. In order to determine the orientation of the aorta we used positional data of all nuclei in the aorta over the whole time period and performed linear regression. The tracking was subsequently rotated and translated in three-dimensional space (see https://github.com/wgiese/zebrafish_ec_migration/wiki for details). We extracted ventral to dorsal velocities over the developmental process, by calculating the signed displacement in y-direction over time windows of 2 h. This process was repeated over the whole time period from 26 h to 44 h in 10 min time steps for all cell tracks. In HUVECs cell trajectory data was aligned such that the y-axis followed the outer edge of the scratch-wound, while the cell-free space is on the positive-x half plane. We calculated the effective speed as the quotient of the distance from start to end point and the tracking time. Furthermore, we determined the directionality of each trajectory from the vector pointing from start to end point and calculated the angle with the y-axis. The data that support the findings of this study are available from the corresponding author upon reasonable request.