Piezo1-Mediated Thromboprotection Is Markedly Influenced By Laminar Shear Forces in Vitro and in Vivo

A system of mechanosensing mechanisms promote a fully-differentiated, quiescent endothelium through stimulation of flow-sensitive transcriptional activation mediated in part by endothelial calcium channels. Loss of laminar flow can interrupt these mechanisms leading to a partial endothelial-mesenchymal transition and promoting a prothrombotic phenotype. Piezo1 is the most abundant calcium channel in the endothelium. It can be activated by the pharmacological agonist Yoda1. We evaluated the thromboprotective effect of Yoda1 on inflammation-induced prothrombotic changes in endothelium in static conditions, laminar flow, non-laminar flow, and in vivo. (1) In static conditions, exposure to as little as 0.25 μM Yoda1 reduced TNFα-mediated factor Xa generation, while concentrations greater than 1 μM increased caspase-3 expression and cell rounding consistent with apoptosis. Yoda1 (0.25 μM) reduced TNFα-mediated factor Xa generation in human umbilical vein endothelial cells by 79% (p<0.001) and in human microvascular endothelial cells of lung origin by 85% (p<0.001), induced thrombomodulin expression, and prevented TNFα-induced tissue factor (TF) and NF-κB expression. This Yoda1 effect is specific for Piezo1. In particular, following siRNA Piezo1 knockdown, endothelial cells exposed to TNFα and Yoda1 failed to potentiate thrombomodulin, showing only 16% (p<0.001) of thrombomodulin expression compared to control siRNA transfected cells. Conversely, the Piezo1 siRNA/Yoda1/TNFα exposed cells potentiated pro-thrombotic factors NF-κB (2.2-fold, p<0.0083) and TF (4.8-fold, p<0.0002) compared to control siRNA exposed cells. Knockdown of KLF4 similarly reversed the protective effects of Yoda1 in the TNFα-induced factor Xa generation assay, indicating that KLF4 mediates thromboprotection downstream of Piezo1. (2) Under the influence of laminar flow, dose finding studies demonstrated that 6-fold higher Yoda1 concentrations were required to attenuate TNFα-mediated FXa generation and TF expression in endothelial cells exposed to mechanical flow compared to static conditions. Under flow, Yoda1 (1.5 μM) exposure resulted in a 63% reduction in FXa generation (p<0.05) and attenuated TF protein levels by 10-fold (p<0.05). Interestingly, under flow, this higher Yoda1 concentration had no effect on endothelial cell morphology or survival. (3) In non-laminar flow conditions, still higher concentrations of Yoda1 were required to achieve protection against the thromboinflammatory transformation. Specifically, 2.5 μM Yoda1 (10-fold higher than static conditions) resulted in a 68% reduction (p<0.05) in TNFα-mediated FXa generation and TF reduction trended towards attenuation (2-fold reduction). No apoptosis was observed. (4) To assess the impact of Yoda1 on thromboprotection in vivo, we used thrombomodulin and EPCR as endothelial-specific readouts for Yoda1-sensitive antithrombotic endothelial changes. Mice were injected with vehicle or Yoda1 at 213 μg/kg, 444 μg/kg, or 533 μg/kg. At 533 μg/kg, but not lower concentrations, we found significant (p<0.05) increases in the expression of thrombomodulin (5-fold) and EPCR (2.4 fold) in pulmonary tissue samples. In mice exposed to both Yoda1 and LPS, pulmonary tissue retained increased levels of thrombomodulin (2-fold, p<0.05) and liver fibrin levels trended lower (0.71-fold) compared to mice exposed to LPS and vehicle. Yoda1 injection inhibited fibrin formation (p<0.01) in a laser-injury model. These findings support our hypothesis that pharmacologic Piezo1 agonism has a potent effect in enhancing the quiescent, antithrombotic endothelial cell barrier. The data also demonstrate that shear forces dramatically modify the effect of agonist-mechanoreceptor pharmacological interactions on endothelial morphology, viability, and prothrombotic phenotypes. Such shear-dependent changes in agonist sensitivity are likely related to modulation of Piezo1 conformation and/or cellular distribution, which are sensitive to shear, although alternative explanations cannot be ruled out. These studies underscore the importance of flow-mediated alterations in agonist-channel interactions in developing antithrombotics targeting mechanoreceptors.