The endothelial cell lining constitutes the interface between the vascular tissue and the circulating blood. In their quiescent state, endothelial cells provide a non-thrombogenic surface and express receptors that participate in the anticoagulant mechanism, such as EPCR, but are devoid of procoagulant receptors, such as TF. Tissue factor, which is present normally on non-vascular but not vascular cells (9
), is the only known cellular receptor for the coagulation proteins FVII/FVIIa, and the formation of TF-FVIIa triggers activation of the coagulation cascade. Although there is no evidence that the native endothelium can support the initiation of the coagulation cascade, there is evidence that unperturbed endothelial cells do support coagulation by generating FVIIa, FXa, and thrombin (27
). Our recent studies showed that non-stimulated endothelial cells support a low level of factor X activation by FVIIa independent of TF (23
). In addition to supporting the various coagulation reactions on their cell surfaces, the endothelial cells are positioned optimally to play a role in the clearance of circulating plasma proteins. However, there is little information on how FVIIa interacts with non-stimulated EC. The studies reported here reveal that FVIIa binds specifically to non-stimulated HUVEC, and that EPCR, a known receptor for protein C/APC, serves as the binding site for FVII/FVIIa. The formation of FVIIa-EPCR complexes neither support the activation of coagulation on non-stimulated endothelial cells nor modulate TF-FVIIa activation of factor X on stimulated endothelial cells. FVIIa binding to EPCR is shown to facilitate FVIIa endocytosis.
The present study is the first to identify that EPCR serves as a major binding site for FVIIa on endothelial cells. The existence of binding sites for FVII/FVIIa on human EC had been reported earlier (13
), and our present data are consistent with some but not all of the observations made in the earlier report. For example, the affinity of FVIIa binding to non-stimulated EC obtained in the present study (Kd
, 32 nM) is in good agreement with the published finding (Kd
, 45 nM) whereas the calculated number of the binding sites differ by about 10-fold (2.8 × 105
the reported value of 3.75 × 106
sites/cell). A major difference between the present study and the earlier report (13
) was the nature of the FVIIa binding site. The earlier study found that, in addition to FVIIa, other vitamin K-dependent proteins, including prothrombin, reduced the binding of 125
I-FVIIa to EC. Based on this, the authors concluded that the binding site was shared by other vitamin K-dependent proteins and did not exhibit specificity for FVIIa. In contrast, in the present study other vitamin K-dependent proteins, except protein C and factor X, failed to reduce the 125
I-FVIIa binding to EC, suggesting that the binding site is not a common binding site for vitamin K-dependent proteins. The reason for the difference between our present study, and the earlier report is unclear. One could speculate that a potential contamination of prothrombin with substantial amounts of FVII/FVIIa was partly responsible for the earlier observation (13
). Alternatively, differences in cell culture conditions and assay buffers could be a reason.
Our conclusion that EPCR serves as a binding site for FVIIa is supported by a number of observations: (i) a 100-fold molar excess of protein C reduced the 125
I-FVIIa binding to EC by 50%, (ii) monoclonal EPCR antibody effectively blocked the 125
I-FVIIa binding to EC, and (iii) transfection of CHO cells with an EPCR expression plasmid increased FVIIa binding to the transfected cells by severalfold compared with the wild-type, and this increased binding was markedly reduced by the presence of protein C, and completely attenuated by the presence of EPCR antibody. While the present article was in preparation, Preston et al.
), as a part of characterizing the protein C gla domain interaction with EPCR using BIAcore, reported that FVIIa bound to soluble EPCR with a comparable affinity to protein C (Kd
~ 150 nM). In these studies, human prothrombin was found to have essentially no sEPCR affinity. These data are consistent with our observation that prothrombin had no effect on FVIIa binding to EPCR on the endothelial cell surface.
At present, it is unclear whether FVIIa binds to non-stimulated EC at sites other than EPCR. As with protein C, a 100-fold molar excess of factor X also reduced the 125
I-FVIIa binding to EC. However, it had no effect on 125
I-FVIIa binding to EPCR transfected cells. Further, FX does not possess Leu-8 in the gla domain that is present in both FVIIa and protein C. This residue is important for EPCR recognition (31
). These data suggest that FVIIa may bind to EC at sites other than EPCR and this site(s) could be a common binding site(s) for both FVIIa (FVII) and FX. Further experiments are needed to support this contention.
It is interesting to note that both FVII and FVIIa bind to EPCR on non-stimulated EC with a similar affinity to that of protein C/APC (32
). This supports the notion that EPCR acts a true receptor for FVII/FVIIa. The affinity of FVII as well as protein C to EPCR was relatively weak (Kd
, ~40–50 nM range) in comparison with TF-FVIIa interaction on cell surfaces (Kd
, ~100 pM to 3 nM) (18
). Therefore, it is unlikely that at plasma concentrations of protein C (70 nM) (35
), all EPCR sites on the endothelial cell surface are occupied by protein C. When both FVII and protein C were present, the amount of each ligand associated with the EPCR would be approximately proportional to their respective concentrations. Thus at their plasma concentrations, about 15% of EPCR will be occupied by FVII and the rest of it by protein C. However, this would ultimately depend upon many other factors, such as their on- and off-rates, and how their respective plasma inhibitors modulate their interaction with the EPCR.
Unlike TF-FVIIa, EPCR-FVIIa complexes neither activate factor X nor induce cell signaling through activation of PAR1 or PAR2. This suggests that EPCR fails to induce the necessary allosteric conformational changes in FVIIa to enhance its catalytic activity. Because FVII/FVIIa binds to TF with a much higher affinity than to EPCR, it is unlikely that FVIIa binding to EPCR influences the formation of TF-FVIIa complexes on activated endothelial cells. At present, it is unknown whether FVIIa binding to EPCR either on non-stimulated or stimulated EC facilitates TFPI inhibition of FVIIa. In this context, it is interesting to note that recent studies showed that protein S stimulates the inhibition of TF pathway of coagulation by TFPI (36
). Although protein S appears to support the TF inhibition by reducing the Ki
of the FXa-TFPI complex formation, it will be interesting to see whether protein S and EPCR affect the formation of the quaternary complex of TF-FVIIa-FXa-TFPI.
At present, the physiological importance of FVII/FVIIa binding to EPCR is unclear. Since FVII circulates in plasma at a 7-fold lower concentration than protein C and FVIIa binds to EPCR with a similar affinity to that of protein C, it is unlikely that FVII acts as the major ligand for EPCR. However, in therapeutic conditions, FVIIa levels may be elevated close to protein C levels in blood, and thus may compete with protein C for EPCR binding. However, under these conditions one would not expect a severe reduction in protein C binding to EPCR since FVIIa would first bind to unoccupied EPCR sites before it could compete with protein C. Nonetheless, therapeutic concentrations of FVIIa will have a significant effect on protein C activation as well as on APC-mediated anticoagulant and anti-inflammatory effects.
The relevance of FVII/FVIIa binding to EPCR may not be in modulating the activation of FVII or protein C, and the activities of FVIIa and APC, but in the clearance of FVII and FVIIa from the circulation. The data presented in this manuscript show that the rate of FVIIa internalization is much higher in EPCR-transfected CHO-K1 cells in absence of EPCR mAb than that observed in presence of the antibody or the wild type cells. Similarly, EPCR mAb also reduced FVIIa internalization in HUVEC. This suggests that the internalization of FVIIa is dependent on EPCR. Interestingly, in contrast to FVIIa internalized via TF (19
), FVIIa internalized via EPCR was not readily degraded. This raises the possibility that FVIIa endocytosed with EPCR is either retained intracellularly or recycles backs to the cell surface. Further experiments are required to evaluate the fate of endocytosed FVIIa. The plasma inhibitors TFPI and AT were shown to influence TF-dependent FVIIa internalization (37
). It would be interesting to investigate whether these inhibitors also influence EPCR-dependent FVIIa internalization and how APC-protease inhibitor complexes affect FVIIa endocytosis.
Overall our present data provide strong evidence that FVII/FVIIa binds to EPCR on endothelial cell surface and suggest that this interaction may play a role in FVIIa endocytosis and could influence the activation of protein C and APC-mediated cell signaling. However, further studies are required to establish the functional significance and consequences of FVII/FVIIa binding to EPCR.