Factor VIIa Induces Biased Cytoprotective Signaling through the Cleavage of Protease Activated Receptor 1 at Canonical Arg41 Site

Endothelial cell protein C receptor (EPCR) interacts with diverse ligands, in addition to its known ligands protein C and activated protein C (APC). We reported earlier that procoagulant clotting factor VIIa (FVIIa) binds EPCR with the same affinity as APC. FVIIa binding to EPCR leads to the down regulation of the EPCR-mediated anticoagulation pathway. Our recent studies showed that FVIIa, like APC, induces EPCR-dependent cytoprotective signaling through activation of protease activated receptor 1 (PAR1). Recent studies of Griffin, Mosnier and their colleagues revealed that APC noncanonical cleavage of PAR1 at Arg46 site that generates a novel tethered ligand is responsible for APC-induced β-arrestin2-dependent PAR1 biased signaling. It is unknown at present whether FVIIa follows a similar mechanism as APC in inducing PAR1 biased signaling. PAR1 reporter constructs - wild-type and cleavage site-specific mutants - were routinely used to investigate PAR1 cleavage by thrombin, APC, or other proteases, and to determine protease-specific cleavage sites in PAR1. Unfortunately, this approach was not useful in determining any FVIIa cleavage site in PAR1. In contrast to thrombin or APC, FVIIa treatment failed to show a detectable cleavage (over the background) of transfected wild-type PAR1 reporter constructs expressed in cultured endothelial cells. However, in other studies, FVIIa was shown to cleave endogenous PAR1 in endothelial cells as assessed by the loss of cleavage-specific PAR1 mAb binding. The recent generation of transgenic mice strains carrying R41Q or R46Q homozygous point mutations in PAR1 has allowed us in the present study to investigate in vivo mechanisms for PAR1-dependent cytoprotective signaling of FVIIa. We employed two murine injury models, LPS-induced inflammation and VEGF-induced barrier disruption. Murine brain endothelial cells isolated from the PAR1 mutated strains and primary human endothelial cells were used to validate in vivo findings and extend the mechanistic studies. Our studies show that administration of rFVIIa (250 µg/kg body weight) reduced LPS-induced cytokine elaboration and neutrophil infiltration in the lung tissues of wild-type (WT) PAR1 and QQ46-PAR1 mice but not in QQ41-PAR1 mice. Similarly, FVIIa suppression of the VEGF-induced barrier disruption was abolished in the QQ41-PAR1 mice but not in WT and the QQ46-PAR1 mice. Parallel experiments conducted with APC showed, as expected, that it protected WT and QQ41-PAR1 mice but not QQ46-PAR1 mice against LPS-induced inflammation and VEGF-mediated barrier destabilization. In vitro signaling studies performed with brain endothelial cells isolated from WT, QQ41-PAR1 and QQ46-PAR1 mice showed that FVIIa activation of Akt in endothelial cells required Arg41 in PAR1. Additional studies showed that FVIIa-cleaved endogenous PAR1 was readily internalized, whereas APC-cleaved PAR1 remained on the cell surface. Very low concentrations of thrombin (< 1 nM) mimicked FVIIa in inducing PAR1-dependent cytoprotective signaling. However, very low concentration thrombin-induced cytoprotective signaling differed from EPCR-FVIIa-induced cytoprotective signaling in the isoform of ß-arrestin required for the protective effect. EPCR-FVIIa-induced PAR1-mediated cytoprotective signaling was soley mediated via the β-arrestin1-dependent pathway whereas very low dose thrombin-induced cytoprotective effects appear to be mediated by either β-arrestin1 or β-arrestin2. Silencing of ß-arrestin1 or ß-arrestin2 alone did not affect the protective effects of very low doses of thrombin, but the silencing both ß-arrestin1 and ß-arrestin2 together completely prevented a low dose of thrombin-induced protective effect. In summary, our data strongly support the hypothesis that the in vivo mechanism of action for FVIIa's pharmacologic benefits in curbing inflammation and endothelial barrier disruption involves biased signaling of PAR1 due to cleavage at Arg41. Moreover, our studies demonstrate that very low concentrations of thrombin are also capable of inducing PAR1 biased beneficial -cytoprotective signaling by coupling to either ß-arrestin1 or 2. These results emphasize the striking diversity of PAR1's conformational states and interactomes that initiate canonical and biased signaling.