Comprehensive In Vitro and In Vivo Characterization of Loss and Gain-of-Function Von Willebrand Factor Collagen Binding Variants Using a Mouse Model System,

Abstract 3266

Von Willebrand Factor (VWF) is a large multimeric glycoprotein that mediates platelet adhesion to the damaged blood vessel wall and subsequent platelet aggregation at the site of injury. Rare mutations in the VWF A3 domain, that disrupt collagen binding, have been found in patients with a mild bleeding phenotype. However, the analysis of these aberrant VWF-collagen interactions has been relatively limited. Thus, in this study, we have developed mouse models of collagen binding mutants and analyzed the function of the A3 and A1 domains using comprehensive in vitro and in vivo approaches. All of the collagen binding variant AAs are conserved in mice. 6 loss-of-function (S1731T, W1745C, S1783A, H1786D, A1 deletion, A3 deletion) and 1 gain-of-function (L1757A) variant was generated in the context of the mouse VWF cDNA. The 4 loss-of-function missense mutants have all been described in patients with mild bleeding phenotypes. The recombinant mouse VWFs (rmVWF) were synthesized in HEK293T cells and analyzed for type I and III collagen binding in both a static assay (CBA) and a flow-based assay at 2,500s⁻¹ in which VWF is bound to collagen on a surface, and labeled platelet adhesion is quantified. The multimer profile of all the rmVWFs was normal. The expression level of the rmVWF derived from HEK293T cells was quantified. W1745C and the A3 deletion showed significantly lower levels of expression and the A1 deletion mutant showed strong intracellular retention. In the static collagen binding assay, S1731T showed almost normal binding to collagen type I and a 50% reduction in binding to collagen type III. The other 3 missense variants, W1745C, S1783A and H1786D, showed reduced binding to both collagens I and III, and the A3 deletion mutant showed absent binding. In the in vitro flow assay, the sensitivity to detect defects in collagen binding was superior to the static assay, although the patterns of binding defects were similar. W1745C showed similar low levels of platelet adhesion to both types of collagen, while S1783A and H1786D showed a lack of platelet binding on the collagen III surface similar to the A3 deletion mutant, and a reduced binding to collagen type I similar to W1745C. The gain-of-function mutant showed consistent enhanced collagen binding and platelet adhesion in the static and flow assays, respectively. In vivo studies delivered the mVWF cDNAs with a strong liver specific promoter by hydrodynamic injection. At 7 days post-delivery, the VWF:Ag levels in the WT and collagen binding variant mice were similar, apart from the W1745C mutant, that showed 14.6% levels compared to WT. Platelet counts and multimer patterns were normal with the collagen binding variants. In vivo intravital microscopy studies were performed using the cremaster arteriolar model when VWF levels were in a physiological range. Thrombosis was induced by 10%FeCl₃ applied for 3 mins. Platelets were labeled in vivo by Rhodamine 6G and the thrombus development was analyzed by spinning disc confocal microscopy. Loss-of-function mutants showed transient platelet adhesion at the site of injury, however the adhesion was unstable and vessel occlusion was not observed. Using three complementary experimental systems we have been able to confirm the collagen binding defects in this group of variant VWFs. There is a differential sensitivity to the two forms of collagen and of the three experimental systems. The A3 deletion mutant consistently resulted in the most severe phenotype while the missense mutants showed variable degrees of functional deficit.