Von Willebrand Factor is not Required for Red Blood Cell Retention in Clots in Mice

Graphical Abstract

Venous thrombosis and pulmonary embolism (collectively venous thromboembolism) affect 1–2/1000 people annually.¹ Venous thromboembolism has ~30% mortality, usually associated with pulmonary embolism, and is the third leading cause of cardiovascular death worldwide.

Venous thrombosis is initiated by blood stasis and activation of coagulation, resulting in intravascular thrombin generation and fibrin deposition. Trapping of red blood cells (RBCs) within the fibrin network and platelet-mediated contraction of the thrombus culminates in a consolidated mass that obstructs blood flow. Factor XIII (FXIII) is a protransglutaminase in plasma and cells. Activated FXIII (FXIIIa) catalyzes the formation of crosslinks between γ- and α-chains of fibrin and between fibrin and other plasma proteins. These crosslinks enhance clot mechanical and biochemical stability. FXIIIa-mediated crosslinking also enhances RBC retention in contracting clots.^2,3 Thus, FXIIIa is a determinant of venous thrombus composition and size in mice and in in vitro clots from humans and mice.^2,3

Von Willebrand Factor (VWF) is a multimeric adhesive glycoprotein. VWF binds and stabilizes factor VIII (FVIII) in circulation and promotes platelet recruitment and adhesion to the endothelium. Using a flow-based system with surface-immobilized plasma and extracellular matrix proteins, Smeets et al showed immobilized VWF can bind RBCs in a leukocyte- and platelet-independent manner.⁴ VWF also colocalizes with RBCs within venous thrombi.⁴ Since VWF can be incorporated in, and crosslinked to, the fibrin network, these observations led to speculation that VWF mediates the FXIIIa-dependent retention of RBCs in venous thrombi.⁴

To investigate the contribution of VWF to RBC retention in clots, we collected blood from 18–22 week old wild-type (4 male, 3 female) or VWF-deficient (B6.129S2-Vwf [tm1Wgr], 9 male, 4 female) C57BL6/J mice (The Jackson Laboratory, Bar Harbor, ME) via inferior vena cava puncture into 3.2% sodium citrate (10% v/v, final). Data from males and females were not different and were pooled for each genotype. Hematocrit was similar between groups (39.2±1.1 versus 40.4±1.7%, VWF-sufficient versus -deficient blood, respectively). We recalcified blood and triggered clotting with tissue factor in the absence and presence of the FXIIIa inhibitor T101, as described^2,3. After 2 hours, we weighed contracted clots and quantified serum RBC content by absorbance at 575 nm. As anticipated^2,3, inhibition of FXIIIa with T101 reduced clot mass by ~50% and increased RBC extrusion from WT clots (Panels A-B). T101 also reduced clot mass and increased RBC extrusion from VWF-deficient clots (Panels A-B), demonstrating similar ability of clots to retain RBCs in the presence and absence of VWF, and in a FXIIIa-dependent manner.

Compared to VWF-sufficient clots, VWF-deficient clots were slightly, but significantly, smaller. Since VWF-deficient mice have modestly reduced FVIII (0.42±0.13 IU/mL), we speculated that the smaller clots reflected reduced thrombin generation causing slightly decreased fibrin network density and RBC retention. To test this possibility, we enlisted two experimental strategies. In the first, we recalcified blood and bypassed effects of reduced FVIII by triggering clotting with bovine thrombin. Since the thrombin concentration present during fibrin formation determines fibrin structure, clot structure differs somewhat when clotting is initiated by tissue factor versus thrombin and RBC retention is somewhat less sensitive to effects of T101 (compare panels A-B with C-D). Nonetheless, in these reactions, there was no difference in contracted clot mass or RBC extrusion from VWF-sufficient versus -deficient clots (Panels C-D). In the second experiment, we supplemented VWF-deficient blood with human plasma-derived VWF (Humate P, concentrations indicated) prior to triggering clotting with tissue factor. Since human VWF does not bind mouse glycoprotein 1b, this experiment specifically tested platelet-independent functions of VWF during clot contraction. Since Humate P also contains FVIII, in parallel reactions we spiked blood with recombinant human FVIII (Xyntha) to identify any effects of FVIII. Neither VWF/FVIII nor FVIII enhanced clot mass, suggesting the slight difference in mass seen in VWF-sufficient and -deficient clots triggered with tissue factor (Panel A) was independent of VWF or FVIII. Importantly, addition of T101 significantly decreased clot mass for all reactions (Panel E).

VWF-deficient mice are protected from venous thrombosis following inferior vena cava ligation.⁵ Intriguingly, supplementation with FVIII does not rescue this phenotype, suggesting effects of VWF are independent of its FVIII carrier function.⁵ Although VWF can bind RBCs⁴ and be crosslinked to fibrin, our data indicate crosslinked fibrin can promote RBC retention in clots in a VWF-independent manner. Our approach differed from Smeets et al in their use of low flow⁴ versus our assay conducted in static blood. Reduced wall shear stress enhances RBC binding to VWF⁴ and stasis increases venous thrombosis risk; however, the extent of flow reduction during venous thrombosis is unknown. The clot contraction assay conducted in static blood recapitulates fibrin- and FXIII-dependent mechanisms that promote RBC retention in venous thrombi in vivo.^2,3 Moreover, roles for VWF following partial (stenosis) and full (stasis) flow reduction⁵ suggest a common, flow-independent function for VWF during venous thrombosis in both models. Collectively, these findings suggest interactions between VWF and RBCs are not required to retain RBCs in venous thrombi.

Fig. 1. VWF is not required for RBC retention in clots.

Blood was recalcified (10 mM, final) and clot formation was initiated in the absence and presence of the FXIIIa inhibitor T101 (50 mM, Zedira GmbH, Darmstadt, Germany). After 2 hours, contracted clots were weighed, and percent of RBCs extruded into the serum was quantified by absorbance at 575 nm using a 384 Plus Spectromax Plate Reader. (A) Mass and (B) RBC extrusion from clots initiated with tissue factor (TF, 1:12,000 dilution of Dade Innovin, Siemens, New York, NY). (C) Mass and (D) RBC extrusion from clots initiated with bovine thrombin (IIa, 2 U/mL Enzyme Research Laboratories, South Bend, IN). Note that y-axes differ in panels A-B versus C-D. (E) Clots were formed in VWF-deficient blood supplemented with VWF/FVIII (Humate-P, CSL Behring, King of Prussia, PA) or FVIII (Xyntha, Pfizer, New York, NY). Statistical analysis was performed using GraphPad Prism v8.4.2. For A-D, paired data within genotypes (±T101) were compared by Student’s t-tests. Unpaired data across genotypes (±T101) were compared by ANOVA with Šídák’s post-hoc test. For E, planned comparisons were made by ANOVA with Šídák’s post-hoc test. Comparisons were restricted to only those shown. Lines indicate means ± standard deviation; dots represent individual mice. *P<0.05; **P<0.01; ***P<0.001; ns, not significant.