The Ciona notochord has emerged as a simple and tractable in vivo model for tubulogenesis. Here, using a chemical genetics approach, we identified UTKO1 as a selective small molecule inhibitor of notochord tubulogenesis. We identified 14-3-3εa protein as a direct binding partner of UTKO1 and showed that 14-3-3εa knockdown leads to failure of notochord tubulogenesis. We found that UTKO1 prevents 14-3-3εa from interacting with ezrin/radixin/moesin (ERM), which is required for notochord tubulogenesis, suggesting that interactions between 14-3-3εa and ERM play a key role in regulating the early steps of tubulogenesis. Using live imaging, we found that, as lumens begin to open between neighboring cells, 14-3-3εa and ERM are highly colocalized at the basal cortex where they undergo cycles of accumulation and disappearance. Interestingly, the disappearance of 14-3-3εa and ERM during each cycle is tightly correlated with a transient flow of 14-3-3εa, ERM, myosin II, and other cytoplasmic elements from the basal surface toward the lumen-facing apical domain, which is often accompanied by visible changes in lumen architecture. Both pulsatile flow and lumen formation are abolished in larvae treated with UTKO1, in larvae depleted of either 14-3-3εa or ERM, or in larvae expressing a truncated form of 14-3-3εa that lacks the ability to interact with ERM. These results suggest that 14-3-3εa and ERM interact at the basal cortex to direct pulsatile basal accumulation and basal-apical transport of factors that are essential for lumen formation. We propose that similar mechanisms may underlie or may contribute to lumen formation in tubulogenesis in other systems.

Flow patterns exert significant effects on vascular endothelial cells (ECs) to lead to the focal nature of atherosclerosis. Using a step flow chamber to investigate the effects of disturbed shear (DS) and pulsatile shear (PS) on ECs in the same flow channel, we conducted single-cell RNA sequencing analyses to explore the distinct transcriptomic profiles regulated by DS vs. PS. Integrated analysis identified eight cell clusters and demonstrated that DS induces EC transition from atheroprotective to proatherogenic phenotypes. Using an automated cell type annotation algorithm (SingleR), we showed that DS promoted endothelial-to-mesenchymal transition (EndMT) by inducing the transcriptional phenotypes for inflammation, hypoxia responses, transforming growth factor-beta (TGF-β) signaling, glycolysis, and fatty acid synthesis. Enolase 1 (ENO1), a key gene in glycolysis, was one of the top-ranked genes in the DS-induced EndMT cluster. Pseudotime trajectory analysis revealed that the kinetic expression of ENO1 was significantly associated with EndMT and that ENO1 silencing repressed the DS- and TGF-β-induced EC inflammation and EndMT. Consistent with these findings, ENO1 was highly expressed in ECs at the inner curvature of the mouse aortic arch (which is exposed to DS) and atherosclerotic lesions, suggesting its proatherogenic role in vivo. In summary, we present a comprehensive single-cell atlas of ECs in response to different flow patterns within the same flow channel. Among the DS-regulated genes, ENO1 plays an important role in DS-induced EC inflammation and EndMT. These results provide insights into how hemodynamic forces regulate vascular endothelium in health and disease.

Characterizing blood flow dynamics in vivo is critical to understanding the function of the vascular network under physiological and pathological conditions. Existing methods for hemodynamic imaging have insufficient spatial and temporal resolution to monitor blood flow at the cellular level in large blood vessels. By using an ultrafast line-scanning module based on free-space angular chirped enhanced delay, we achieved two-photon fluorescence imaging of cortical blood flow at 1,000 two-dimensional (2D) frames and 1,000,000 one-dimensional line scans per second in the awake mouse. This orders-of-magnitude increase in temporal resolution allowed us to measure cerebral blood flow at up to 49 mm/s and observe pulsatile blood flow at harmonics of heart rate. Directly visualizing red blood cell (RBC) flow through vessels down to >800 µm in depth, we characterized cortical layer–dependent flow velocity distributions of capillaries, obtained radial velocity profiles and kilohertz 2D velocity mapping of multifile blood flow, and performed RBC flux measurements from penetrating blood vessels.