Normal Cleavage of von Willebrand Factor by ADAMTS13 in the absence of Factor VIII in Patients with Severe Hemophilia A

Von Willebrand factor (VWF) is required for hemostasis, recruiting platelets from rapidly flowing blood to sites of vessel injury and protecting factor VIII (FVIII) from degradation. The adhesive activity of VWF correlates with its size: large VWF multimers bind more avidly to platelets [1]. VWF multimer distribution is regulated by the metalloprotease ADAMTS13 in plasma [2,3]. In the absence of ADAMTS13 activity, ultra-large VWF (ULVWF) multimers accumulate and induce spontaneous platelet clumping [4] and cause the life-threatening disease thrombotic thrombocytopenia purpura [5].

Recent studies by Cao et al. [6] showed that under shear stress in a system using purified proteins, exogenous FVIII enhanced the cleavage of high-molecular-weight VWF multimers by ADAMTS13. Based on this result, the authors proposed that FVIII is a cofactor for ADAMTS13, suggesting that reduced VWF processing and increased platelet adhesion could represent a form of hemostatic compensation in patients with severe hemophilia A. This finding predicts that absence of FVIII in hemophilia A patients would reduce VWF proteolysis, which would be normalized by infusion of FVIII.

Here, we assessed VWF multimer distribution, VWF antigen levels, and ADAMTS13 activity in the plasmas of seven patients with severe hemophilia A before recombinant FVIII infusion. In two patients, we also examined the VWF cleavage by endogenous ADAMTS13 before and after FVIII infusion.

FVIII levels in patients with severe hemophilia A

The FVIII levels measured in the clinical laboratory were less than 1% in six patients and 4% in one (patient E). The post-infusion FVIII levels were 19% (Post-1) and 94% (Post-2) for patient C, and 59% for patient E.

VWF multimer distribution, antigen concentration and ADAMTS13 activity

The VWF multimer patterns of the hemophilia A patients were similar to those from pooled normal plasma (Figure 1A). No ULVWF multimers were detected in the plasma of any of the patients. Using the same control plasma, we easily detected ULVWF in plasma from patients with thrombotic thrombocytopenic purpura (right panel of Figure 1A) and patients with sickle cell disease [7]. VWF antigen levels, although still within the normal range, were decreased in all patients, ranging from 33% to 77% of the concentration in pooled normal plasma (Figure 1B). The ADAMTS13 activities, measured with a peptide substrate [8], were elevated in all of the patients, ranging from 1.1- to 1.8-fold the activity of pooled normal plasma (Figure 1C).

Figure 1. VWF multimer distribution, antigen level, and ADAMTS13 activity in patients with severe hemophilia A.

Studies on citrated plasma samples from patients with severe hemophilia A were approved by the Western IRB. Pooled normal plasma (normal reference plasma; Precision BioLogic Inc., Dartmouth, Nova Scotia, Canada) used in our studies has VWF antigen level of 1.22 U/ml. A. VWF multimer distribution. The VWF multimer distribution was examined by electrophoresis followed by Western blotting with a polyclonal VWF antibody (Dako North America, Inc., Carpinteria, CA) as previously described [17]. For each hemophilia patient, 1 µl of plasma was analyzed. For pooled normal plasma (PNP), 1 µl and 0.5 µl of plasma were analyzed, respectively. B. VWF antigen levels. VWF antigen was measured by sandwich ELISA using a polyclonal VWF antibody as the capture antibody, a horseradish peroxidase (HRP) conjugated polyclonal VWF antibody to detect the captured VWF, and PNP as a standard. C. ADAMTS13 activity. ADAMTS13 activity in plasma was determined using an HRP conjugated-peptide substrate as described previously [8] with PNP as a standard. D. Ratio of ADAMTS13 activity to VWF antigen. The ratio was calculated by dividing the ADAMTS13 activity by the VWF antigen and is expressed as a value relative to the value in normal plasma. In panels B, C, and D, data from different patients are presented in different colors; patients C and E provided plasma samples from pre- and post-FVIII infusion. The values expressed are relative to those obtained with PNP, which was arbitrarily assigned a value of 1, as indicated by the dashed lines in each panel. The assays were performed independently 2–3 times per sample; average values are shown. E. Cleavage of VWF in plasma by endogenous ADAMTS13. Pooled normal plasma, or plasma from hemophilia patients C and E was diluted 10-fold with a buffer containing 10 mM HEPES, 6.5 mM BaCl₂, and 1.5 M urea, incubated at 37°C, and ADAMTS13 cleavage was stopped at the indicated times with EDTA. VWF multimer patterns were examined on a 1.5% agarose gel and detected by an HRP-conjugated VWF antibody on Western blots.

Cleavage of VWF in plasma by endogenous ADAMTS13

In two patients, we compared the cleavage of endogenous VWF by endogenous ADAMTS13 before and after infusion of FVIII to the cleavage pattern of pooled normal plasma. In the hemophilia A plasma samples, endogenous ADAMTS13 cleaved VWF as efficiently as in pooled normal plasma, even in the absence of FVIII (Figure 1E). The pattern did not change after FVIII levels were increased to 94% or 59% by factor infusion.

These results indicate that the FVIII level does not influence VWF processing in vivo to a significant extent. These results are inconsistent with the findings of Cao and coworkers [6,9,10], who proposed a cofactor role for FVIII in ADAMTS13-mediated proteolysis of VWF. Although the inconsistency could be attributed to differences in assay systems, it is reasonable to expect that if FVIII were an important cofactor for regulating VWF proteolysis in vivo, its absence in severe hemophilia A would manifest in larger, more adhesive circulating VWF multimers and that infusion of FVIII would change the multimer distribution. We did not observe either of these phenomena.

Our results are consistent with a study by Grünewald et al. [11], who studied 21 hemophilia patients, 12 with severe hemophilia A and 9 with severe hemophilia B, and found decreased collagen-induced platelet aggregation, reduced ristocetin-induced platelet aggregation (at a ristocetin concentration of 1.0 mg/ml), and prolonged closure times by Platelet Functional Analyzer (PFA) using both collagen/epinephrine and collagen/ADP cartridges, all consistent with depressed VWF antigen or activity [12–14]. ADP- and epinephrine-induced platelet aggregation was normal in these patients.

Another interesting finding from our study was that all hemophilia patients had decreased VWF antigen levels and increased ADAMTS13 activity with respect to pooled normal plasma. In fact, when normalized to the generally low levels of VWF in the hemophilia A plasma, the ADAMTS13 activities were elevated in all of the patients, ranging from 1.4- to 5.2-fold the activity of pooled normal plasma (Figure 1D). Whether this is a consistent finding in hemophilia A will have to await study of a larger cohort of patients, especially given the very wide distribution of VWF antigen levels in the normal population. This result is, however, consistent with previous findings that low VWF antigen levels correlated with high ADAMTS13 activity [15,16]. The observed decrease in VWF levels could be a consequence of defective generation of thrombin associated with FVIII deficiency, a known agonist that stimulates VWF secretion from endothelial cells.

In conclusion, patients with severe hemophilia A have normal cleavage of VWF by ADAMTS13 in vivo and ex vivo in the absence of FVIII.