von Willebrand factor (VWF) and factor VIII (FVIII) circulate in the plasma as a non-covalent complex. VWF can influence FVIII activity by stabilizing FVIII plasma levels and transporting FVIII to the site of thrombus formation. We have previously generated a murine model of type 2N VWD through hydrodynamic expression of the severe R816W VWF variant that exhibits a 90% decrease in FVIII-binding ability and a failure to stabilize endogenous FVIII when expressed in a VWF-deficient mouse. Intravital studies of arteriole platelet thrombus formation demonstrated that FVIII knockout and type 2N VWD mice produced smaller, less stable thrombi.

Aim: In this study, we assess the contribution of VWF and FVIII interactions to the formation and stability of hemostatic thrombi in a murine tail vein transection (TVT) bleeding model.

Method: Under isoflurane anaesthesia the left lateral tail vein was transected using standardized measuring and transection templates (provided by Novo Nordisk, described in detail in Johansen et al. Hemophilia. 2016). Bleeding time was observed over the course of 60 minutes. Blood was collected into saline for hemoglobin quantification. For experimental conditions, 300 U/Kg wild type or type 2N (R816W) VWF was administered IV to VWF KO mice 2 hours prior to TVT in order to allow for endogenous FVIII stabilization.

Results: Consistent with previous reports, VWF (44.15 min; p<0.0001) and FVIII-deficient mice (45 min; p<0.0001) exhibited a significantly prolonged total bleeding time relative to normal mice (4.6 min). Infusions of plasma-derived murine VWF (FVIII-free) into VWF deficient mice have demonstrated that wild type (WT) VWF is capable of restoring FVIII:C levels to ~75% 2 hours post-infusion, while the R816W variant does not influence FVIII stabilization and levels remain at ~15%. Mice infused with WT VWF had a shorter bleeding time (16.2 min) relative to mice infused with the severe type 2N VWD variant (44.67 minutes, p<0.0001).

For FVIII KO mice, the primary bleeding time (1.53 min), defined as the period of time until first bleeding cessation, was not different from normal mice (1.2 min) but significantly extended for VWF KO mice (35.97 min, p<0.0001). Primary bleeding time was markedly reduced for VWF KO mice infused with WT VWF (4.97 min, p=0.07) confirming the restoration of primary hemostasis by infusion of murine plasma-derived VWF.

Hemoglobin quantification confirmed that FVIII (243.8 mg, p<0.0001) and VWF KO mice (198.8 mg, p<0.0001) had increased blood loss compared to normal mice (9.91 mg). Similarly, VWF KO mice infused with type 2N VWF (308.04 mg) had increased blood loss compared to mice infused with WT VWF (73.78 mg, p=0.0008). Total blood loss and bleeding times were positively correlated across all groups (r2=0.625, p<0.0001).

Thrombus stability was characterized by the frequency of spontaneous re-bleeding events. 11% of normal mice experienced one or more spontaneous re-bleeding events during the course of injury, while 100% of FVIII deficient mice experienced spontaneous re-bleeding (p=0.0004). A similar trend was observed for VWF KO mice infused with WT VWF (0%) and type 2N VWF (66.7%, p=0.021).

Conclusions: These studies suggest that in the murine TVT model of physiological hemostasis, the initial hemostatic thrombus formation is predominantly driven by platelet plug formation and is VWF-dependent. In contrast, thrombus stability over the course of the model is reliant on the processes responsible for fibrin formation, and is FVIII-dependent. This suggests that patients with type 2N VWD and hemophilia A may have bleeding that is the result of impaired stabilization of the primary platelet plug.


Lillicrap:Baxalta: Research Funding; Biogen-Idec: Research Funding; Bayer: Research Funding; Octapharma: Research Funding.

Author notes


Asterisk with author names denotes non-ASH members.