Identification of clearance mechanisms of von Willebrand factor (VWF) has been the subject of an increasing number of studies in the past few years. Indeed, clearance, as well as biosynthesis, contributes to the maintenance of steady-state plasma levels of VWF. Furthermore, VWF variants with increased clearance have now been identified that can lead to von Willebrand disease (VWD) . Among potential determinants of VWF clearance kinetics, the glycosylation profile of the molecule and the presence of some mutations have been reported to play a role. Another candidate potentially regulating VWF removal is ADAMTS-13 (A Disintegrin and Metalloprotease with ThromboSpondin motif), the protease that specifically cleaves VWF to prevent accumulation of very high-molecular weight multimers in plasma. It is not uncommon that proteolysis of proteins is taken by the body as a sign for removal and it has been suggested that this may apply to VWF as well . However, data available so far are not in support of this hypothesis. First, no difference in survival was observed for plasma-derived human VWF fractions enriched in either high (predominantly 14-mers and higher) or low (predominantly dimers and tetramers) molecular weight multimers in VWF-deficient mice . Second, Stoddart et al.  compared the survival of wild-type VWF and a proteolysis-sensitive VWF mutant in rats and could not detect differences in removal of both proteins. Third, the relationship between ADAMTS-13 levels and VWF half-life has also been examined in patients with VWD-type 1 or hemophilia A . This last study could not establish a correlation between ADAMTS-13 levels and post-desmopressin survival of VWF, further suggesting that ADAMTS-13 is not involved in VWF clearance. However, these studies are either indirect or hampered by cross-species issues that could introduce some bias in the results. Indeed human VWF only displays low sensitivity to murine ADAMTS-13. In the present study, we therefore addressed this issue using a more direct approach with ADAMTS-13-deficient mice, generously provided by Dr David Motto (University of Iowa) . The mice were back-crossed eight times on a C57Bl/6 background. First, these mice were compared with their wild-type littermates for blood counts using an automated animal blood cell counter (Scil Vet abc) and for VWF multimeric composition on a 2% sodium dodecyl sulfate-agarose gel electrophoresis followed by immunoblotting. Consistent with the original description of the mice and although the genetic background was different from that report, we did not detect any difference between the VWF multimeric pattern of ADAMTS+/+ or ADAMTS−/− mice (Fig. 1B). In terms of hematologic data, we observed that ADAMTS−/− mice had significantly lower platelet counts than ADAMTS+/+ mice (728 ± 21 vs. 842 ± 42 109 L−1, P = 0.036), potentially reflecting the prothrombotic state observed in these mice . In order to get first insight into the ADAMTS-13 role into VWF clearance, we analyzed steady-state levels of VWF antigen levels in ADAMTS-13-deficient mice and their wild-type littermates, assuming that levels should be higher in ADAMTS-13-deficient mice if proteolysis enhanced clearance. Although there was a slight trend towards higher VWF antigen levels in the knockout mice, it did not reach statistical significance (117 ± 13% vs. 103 ± 9% for wild-type mice, P = 0.36) (Fig. 1C). This result appears different from the data reported by Chauhan et al.  where a significant 20% increase in VWF levels was observed in ADAMTS-13−/− mice. However, even higher VWF levels are not absolutely indicative of a lower clearance and as suggested by Chauhan et al., it might also result from endothelial activation and increased Weibel-Palade bodies exocytosis. We thus decided to study the clearance of endogenous VWF in mice expressing or not ADAMTS-13. Mice, ADAMTS-13+/+ or ADAMTS-13−/−, were injected in the tail vein with 500 μg of biotin-N-hydroxysuccinimide ester (NHS; Merck Chemicals Ltd, Nottingham, UK), dissolved in saline buffer. After allowing the biotinylation reaction to complete for 10 min, blood (100 μL) was collected and plasma prepared at different time points (0.3, 1, 2, 6 and 24 h). The amount of biotynylated VWF measured at time 0 (10 min after NHS-biotin injection) was considered as 100% for each individual mouse. Residual biotinylated VWF was measured by an immunosorbent assay using polyclonal anti-human VWF (Dako; Dako France SAS, Trappes, France) and horseradish peroxidase-labeled streptavidin (R&D systems, R&D systems Europe, Lille, France) and was expressed as the percentage of biotinylated VWF levels at t = 0 . As shown on Fig. 1A, no difference could be detected in VWF disappearance kinetics in the presence or absence of ADAMTS-13. Calculated half-lives were 1.5 ± 0.2 and 1.9 ± 0.3 h, respectively (P = 0.24). We also measured VWF clearance kinetics using a slightly different approach. VWF−/−/ADAMTS-13−/− mice or VWF−/−/ADAMTS-13+/+ mice were injected with 100 μg of pLIVE vector (Mirus Bio LLC, Madison, WI, USA) encoding murine VWF cDNA via hydrodynamic injection as described previously . After this procedure, liver hepatocytes start to express murine VWF, which is secreted in the circulation. Expression is stable for 2 weeks. Four days after injection, mice were injected with NHS-biotin and VWF half-life was measured as described above. Similar to endothelial cell derived-VWF, hepatocyte derived-VWF was eliminated from the circulation at the same rate in the presence or absence of ADAMTS-13 (not shown).
Our results thus clearly show that ADAMTS-13 does not play a role in VWF removal from the circulation in an in vivo model, whether the experiment was performed on a VWF-deficient background or not. This indicates that conversion of high-molecular weight multimers to low-molecular weight multimers by ADAMTS-13 is not a pre-requisite for VWF clearance. Whether other proteases may exist which are able to modify VWF survival is yet unknown but deserves further studies.