VWF Cleavage Products Inhibit Shear-Induced Self-Association

Background

Von Willebrand factor (VWF) self-association is key to hemostasis in high-shear regions of the vasculature. Further, in pathologic small vessel thrombosis, VWF can also self-associate, even on an intact endothelium in the absence of vessel injury (Dong et al. Blood, 2002, 100:40339). VWF self-association preferentially occurs between large multimers, which are more sensitive to shear-induced unfolding. This process is regulated by ADAMTS13 proteolysis, which converts circulating ultra-large VWF multimers to smaller, less-adhesive multimers(Sadler. Proc Natl Acad Sci USA, 2002, 99(18):11552-4). ADAMTS13 not only cleaves large multimers, in the process it also generates small VWF cleavage fragments. This process is sometimes excessive, such as in von Willebrand disease (VWD) type 2A, in which large amounts of cleavage products may be present. However, the role of these small VWF fragments in regulating VWF function has not been studied. Here, we propose that small VWF fragments inhibit VWF self-association by competitively inhibiting self-association sites without providing functional substrates to propagate VWF strand formation.

Methods

Device: We developed a technique based on a device recently used by the Diamond laboratory to visualize VWF self-association in vitro (Zhu et al. Biorheology, 2015, 52:303; Herbig and Diamond. J. Thromb. Haem. 2015, 13:1699). The device consists of a PDMS microfluidic channel of width 60 µm with a 30 µm square pillar in the center. Flow through the channel produces high shear gradients next to the block. The inlet shear rate was set at 5000 s^-1.

VWF fragment preparation: Purified plasma VWF was incubated with purified recombinant ADAMTS13 (with a biotin tag) at a ratio of 5:1 at 37°C in 1.5 M urea for 16 hr. Recombinant ADAMTS13 was then removed with streptavidin-coated beads, and urea was removed with a PD10 column. This produced fragments corresponding to N- and C-termini VWF dimers.

Assay: VWF fragments were mixed with purified plasma VWF (5 µg/mL) at a molar ratio (fragment to VWF monomer) of 1:1 and 5:1 and the mixture was immediately perfused into the channel. Results were compared to those from samples that contained either only fragments or only purified VWF (30 µg/mL). A total sample volume of 200 µL was perfused over 10 minutes, and VWF strand formation was monitored in real-time with differential interference contrast and immunofluorescence microscopy.

Patient Samples: Samples from patients with type 2A VWD and thrombotic thrombocytopenic purpura (TTP) were obtained from consented adults under a protocol approved by the University of Washington Institutional Review Board. A total of 200 µL of sample was perfused in the channel over 10 minutes.

Results

We were able to visualize VWF strand formation induced by shear stress alone when plasma was perfused through the channel. The strands formed from the initial adhesion of VWF to the pillar and subsequent self-association of VWF with the adherent VWF. Strands formed at physiologic VWF concentration and shear stress. The addition of small VWF fragments at 1:1 and 5:1 ratios delayed the formation and elongation of VWF strands and reduced the final extent of strand formation by approximately 20% and 60%, respectively. Compared to normal, strand formation was more pronounced in the absence of small VWF fragments (TTP) and less pronounced in the presence of larger amounts of small VWF fragments (VWD 2A).

Summary

We present a microfluidic assay to visualize in real-time VWF strand formation by generating elongational flow (shear gradients). VWF strands thus formed were reduced in the presence of small VWF fragments generated by ADAMTS13 both in a purified system and in patient samples. This result suggests that the hemostatic defect associated with type 2 VWD is a consequence not only of the deficiency of large multimers, but also of the presence of excessive quantities of proteolytic fragments, and that the profound thrombotic phenotype of TTP results not only from excess ULVWF, but also from the deficiency of smaller fragments.