Von Willebrand Factor Binding to Myosin Assists in Coagulation

Introduction: Von Willebrand factor (VWF) binds to platelets and collagen as a means of facilitating coagulation at sites of injury. Recent evidence has shown that myosin can serve as a surface for thrombin generation and binds to activated factors V and X (Deguchi et al, Blood 2016;128:1807-1878). We hypothesized that VWF could also bind myosin as a means of bringing factor VIII (FVIII) to sites of clot formation. Methods: Recombinant wild-type (WT) was transfected into HEK293T cells and supernatants collected for VWF to use in experiments. VWF variants containing point mutations at specific sites were constructed via site directed mutagenesis and expressed as above. Myosin from rabbit skeletal muscle was plated on maleic anhydride plates, recombinant VWF or plasma samples added and presence of VWF detected using anti-VWF monoclonal antibodies. Competition assays were performed with antibodies blocking various sites on VWF. FVIII activity was measured using a chromogenic substrate (Chromogenix Coatest). Thrombin generation was performed using a fluorogenic substrate from the Technothrombin TGA kit with either tissue factor +/- myosin or the RCL reagent +/- myosin added to platelet poor plasma from healthy controls. Results: WT VWF and human plasma VWF from healthy controls bound myosin, while plasma lacking VWF failed to demonstrate binding. Binding was multimer dependent, with a dose dependent decrease in binding seen with increasing loss of high molecular weight multimers. When myosin binding to VWF:Ag ratios were compared, a solution of ultra high molecular weight multimers demonstrated a mean ratio of 1.6, a solution of high molecular weight multimers a ratio of 1.4, a solution of medium molecular weight multimers a ratio of 0.4 and binding was undetectable with only low molecular weight multimers. A polyclonal anti-VWF antibody (DAKO) completely blocked VWF binding to myosin as did an antibody directed against the VWF A1 domain (AVW-3). Antibodies directed against other VWF sites including the N and C terminal ends failed to affect VWF binding to myosin. Since collagen IV also binds in this region, collagen IV was added to the assay as competition and also completely blocked VWF binding to myosin. VWF variants p.R1395A and p.R1399A showed undetectable binding, while a variant in the A3 domain (that abrogates type III collagen binding) showed normal binding to myosin (95% of WT). However, additional residues that affect VWF-collagen IV interactions were tested and had minimal effect on myosin binding, including p.1392A and p.1402A. Taken together, these results suggest a binding site for myosin in the VWF A1 domain similar but not identical to that of collagen IV. FVIII activity was detected when a VWF concentrate containing both VWF and FVIII was bound to myosin, with a dose-dependent increase in activity. However, no increase in FVIII activity was seen using recombinant FVIII alone in the absence of additional VWF. Because previous results from J. Griffin and coworkers showed a role for myosin in enhancing thrombin generation, we also looked at thrombin generation in platelet-poor plasma with and without myosin, but no difference was seen in the absence of additional phospholipids. Peak thrombin with myosin was 80 nM vs 62 nM without myosin (p=NS), and the area under the curve was 1022 vs 775 (p=NS). With additional phospholipids, there was a trend towards increased thrombin generation with myosin but the difference was again not statistically significant. Peak thrombin with myosin was 70 nM vs 64 nM without myosin (p=NS), and the area under the curve was 974 vs 867 (p=NS). Discussion: Myosin can also serve as a surface for VWF binding. This may help facilitate delivery of FVIII to sites of injury and improve thrombin generation, although in our hands myosin did not substantially increase thrombin generation in the presence of additional phospholipid. Unlike factor V, FVIII does not intrinsically interact with myosin to accelerate thrombin generation. However, myosin, similar to collagen, can increase local accumulation of VWF-FVIII and indirectly accelerate thrombin generation and clot formation. Disclosures Montgomery: BCW: Patents & Royalties: GPIbM assay patent to the BloodCenter of Wisconsin.

Improved Values for Acquired Von Willebrand Syndrome and Better Clinical Outcome in Patients with Heartmate III Compared to Heartmate II

Objectives: Impaired binding of Von Willebrand factor (VWF) to platelets and to collagen due to an acquired Von Willebrand syndrome (AVWS) has been observed in patients with left ventricular assist device (LVAD) support. AVWS seems to enhance bleeding symptoms in LVAD patients. The onset of AVWS occurs a few hours after LVAD implantation and affects platelet counts and function as well. The new HeartMate III (HM III) is a left ventricular assist device featuring several design-improvements that show promise to ameliorate the severity of AVWS in comparison to its predecessor, the Heartmate II (HM II). In this study, our aim was to analyze AVWS-parameters and platelet function in patients with HM II compared to patients with HM III support. Methods: Data sets of 31 patients under HM II support and 20 patients under HM III support were analyzed pre-surgery as well as 1, 3, 7, and 30 days post-surgery. Collagen binding capacity (VWF:CB), VWF antigen (VWF:Ag) as well as VWF:CB/VWF:Ag-ratios were determined. Presence of high molecular weight multimers of VWF was investigated. Platelet counts were monitored and platelet function was tested using light transmission aggregometry. Results: The VWF:CB/VWF:Ag ratios were significantly higher in patients with HM III than HM II at day 1, 3, and 7 after implantation. More HM III patients had intact high molecular weight multimers of VWF compared to HM II during the whole study-period (HMII: 9%; HMIII: 25%). Platelet counts and functions were comparable in both study groups. The mean duration in the Intensive Care Unit was shorter in the HM III-group than in the HM II-group (HM II: 43±52 days; HM III: 28±23 days). Conclusion: Severity of AVWS was milder in patients with HM III compared to HM II, especially during the days after surgery, when usually most hemorrhagic events occur. Loss of high molecular weight multimers 30 days after VAD-implantation was observed in a lower percentage of HMIII-patients compared to HMII-patients suggesting a long-lasting beneficial effect. Taken together, these data show that design-improvements in the new HM III compared to the HM II result in an amelioration of AVWS that could result in lower bleeding diathesis. Disclosures Geisen: Roche: Research Funding. Berchtold-Herz:Novartis: Consultancy; Thoratec: Consultancy. Zieger:CSL Behring: Research Funding; Baxalta: Research Funding; Bayer: Research Funding; German Research Foundation, Bonn, Germany [ZI 486/7-1]: Research Funding.

A Comparison of Von Willebrand Factor Indexes in Heartmate II Compared to Heartware Left Ventricular Assist Devices

Background: Von Willebrand factor (VWF) is a multimeric glycoprotein synthesized in endothelium and megakaryocytes. Elevated total VWF antigen has been associated with arterial thrombosis, while in patients with left ventricular assist devices (LVAD), loss of high molecular weight multimers is associated with spontaneous bleeding. We hypothesized that the lower rotational speed of the Heartware LVAD would be associated with less severe disruption of VWF. Methods: A total of 22 patients were included. We compared samples from a) 17 patients with Heartmate II LVAD (HM II), 76% male, median age 67 (54-73) years, 24% non-white, 45 (33-59) months after implant, with b) 5 patients with Heartware (HW) LVAD, all male, age 64 (59-65) years, 20% non-white, 27 (14-29) months after implant. All patients received warfarin. All HW patients took aspirin, while 82% of HM II patients took aspirin. Samples were analyzed for VWF antigen, activity, platelet function analyzer-100 collagen adenosine diphosphate closure time (PFA) and VWF multimers by gel electrophoresis. Multimers were assessed qualitatively and quantitatively by dividing mers > 15 by mers 2-15 (multimer ratio) and with a normalized multimer ratio (multimer ratio of patient / normal plasma control from the same gel). Results: Although flow rates did not differ significantly between HM II and HW, 5.2 (4.6-6.5) versus 4.3 (4.2-5.4) liters per minute respectively, rotational speeds were very different, mean 9472 versus 2680 rpms, p<0.0001). There were no differences in hemoglobin, hematocrit, or platelet count at the time of sampling. Qualitative VWF multimers were abnormal in all patients, and PFA was non-clotting in 88% of HM II patients and 80% of HW patients. VWF activity to antigen ratio was not different: HM II 0.68 (0.63-0.74) versus HW 0.70 (0.63-0.71). However, VWF antigen was significantly higher in HM II patients, 199 (174-227) compared to HW, 139 (122-160), p=0.039 and the normalized VWF multimer ratio was significantly lower, suggesting greater loss of high molecular weight multimers in HM II, 0.43 (0.41-0.48) compared to HW, 0.53 (0.48-0.56), p=0.0025 (figure). Conclusion: The lower rotational speed associated with the Heartware LVAD appears to produce less elevation of VWF antigen, and slightly less loss of VWF high molecular weight multimers compared to the Heartmate II LVAD. Additional studies are needed to describe whether or not lower pump thrombosis rates and rates of spontaneous bleeding are associated with these differences. Figure 1. Figure 1. Disclosures No relevant conflicts of interest to declare.

Laboratory Predictors Of Loss Of High Molecular Weight Multimers In Patients With Cardiovascular Disease-Associated Acquired Von Willebrand Syndrom

Background Association of acquired von Willebrand syndrome (AVWS) with various cardiovascular (CV) disorders such as cardiac valve disease and hypertrophic cardiomyopathy (HCM) is well known and documented. The mechanism is thought to be related to shear stress induced loss of high molecular weight multimers (HMWM). The gold standard test to assess for loss of HMWM is von Willebrand protein electrophoresis and then visual assessment of loss of high molecular weight bands. This is both a costly and subjective test. Ratio of von Willebrand factor activity to antigen level is useful in patients with type IIA Von Willebrand Disease caused by loss of HMW multimer, but its sensitivity to detect CV-associated AVWS is unknown. Aim Our aim was to test whether routine VWF laboratory tests could be used to predict which patients with CV conditions are going to have high molecular weight multimer loss. We also aimed to assess whether these tests could be used to predict bleeding risk in patients with CV disorders. Methods We prospectively collected laboratory data of 234 patients with cardiovascular disorders known to be associated with AVWS: aortic stenosis (66), aortic insufficiency(22), aortic and mitral valve prostheses(38), mitral valve regurgitation (51) and hypertrophic cardiomyopathy(57). All patients had Von Willebrand factor antigen (VWF:Ag), Von Willebrand factor activity by latex method (VWF:Ltx), platelet function testing via PFA-100 CADP as well as von Willebrand factor multimers tested. All patients also completed a bleeding questionnaire. We used logistic regression model to calculate the relationship between the VWF:Ltx/VWF:Ag ratio and loss of high molecular weight multimers. Same analysis was performed for PFA-100. We also tested these associations for bleeding risk. Results Mean value for VWF:Ag was 142 IU/dL, VWF:Ltx 121%, PFA-100 151 seconds and 0.86 for the VWF:Ltx/Ag ratio. Over a half of patients (56%) had VWF multimer loss noted on protein electrophoresis testing and a quarter reported bleeding on bleeding questionnaire. The ratio of VWF:Ltx to VWF:Ag had strong correlation with HMW multimer loss (p<0.001) with AUC of 0.77. Correlation with PFA-100 was even stronger with AUC of 0.83. The ratio cut off value of 0.83 had sensitivity of 60% and specificity of 83% in predicting multimer loss. With the cut off of 0.77, specificity reached 95%. With PFA 100 value of 118 seconds, specificity was 76% and sensitivity was 80%. Increasing the cut off to 198 seconds improved the specificity to 95%. The association with bleeding was present for PFA-100 (p=0.01), but did not exist for the Ltx/Ag ratio. Conclusions PFA-100 CADP as well as VWF:Ag and VWF:Ltx can be used to detect acquired Von Willebrand disease in patients with cardiovascular disorders and may decrease the need for costly and time consuming testing of multimers. PFA-100 CADP also correlates with the bleeding risk in these patients. Disclosures: No relevant conflicts of interest to declare.

Von Willebrand Factor Collagen Binding Provides a Sensitive Screen for Identification of Type 2A and 2B Von Willebrand Disease

Abstract 379FN2 Collagen binding is an easily performed test of von Willebrand factor (VWF) function but its role in clinical evaluation is still debated. Analysis of multimer distribution, on the other hand, is time-consuming and technically challenging. We hypothesized that VWF antigen (VWF:Ag), ristocetin cofactor activity (VWF:RCo), and collagen binding (VWF:CB) could identify the subset of von Willebrand disease (VWD) cases in which multimer analysis would be informative. Subjects from the Zimmerman Program for the Molecular and Clinical Biology of VWD were analyzed for VWF:Ag, VWF:RCo, VWF:CB (with type III human placental collagen), multimer distribution, and full VWF exon sequencing. Normal controls as well as patients with type 1, 2A, 2B, 2M, and 2N VWD were analyzed. The mean VWF:CB/VWF:Ag ratio for subjects with normal multimers was 1.10, while the mean ratio for subjects with abnormal multimers was 0.51 (p<0.001). When results were restricted to those subjects with confirmed type 2A or type 2B mutations, however, the mean ratio for subjects with abnormal multimers decreased to 0.41 (p<0.001 compared to those with normal multimers). For the 146 normal controls with multimer results available, 2 had absence of the highest molecular weight multimers, but normal collagen binding, normal bleeding scores, and no evidence of a VWF gene mutation, suggesting that the multimer results represented assay artifact. 354 type 1 subjects were examined; of those, 12 had abnormal multimer patterns. 7 had loss of the high molecular weight multimers. Of these, 5 had known type 1 VWD mutations and normal VWF:CB/VWF:Ag ratios, possibly representing sample artifacts rather than a true multimer abnormality, as no multimer issues have been previously reported for these mutations. One had no mutation found and one had a type 2A mutation. 2 had a full spectrum of multimers with relatively increased staining of the lower molecular weight bands; both with novel A1 domain mutations that are currently under investigation. 3 had larger than normal multimers observed, all with normal VWF:CB/VWF:Ag ratios. Of the 342 type 1 subjects with normal multimers, only one had a VWF:CB/VWF:Ag ratio of <0.7, likely due to very low values (VWF:CB of 2 and VWF:Ag of 4). There were 36 type 2A subjects available for analysis. 27 had loss of high molecular weight multimers. Only 3 of those had VWF:CB/VWF:Ag ratios >0.7, but none of those subjects had VWF mutations consistent with type 2A VWD. 7 subjects had a shift from high to low molecular weight multimers, 4 with VWF:CB/VWF:Ag ratios >0.7 and either known type 1 mutations or novel VWF gene mutations. 2 subjects had normal multimer distribution, one with a type1 VWD mutation and one with a novel mutation. Characterization of these novel mutations is in progress. All the 17 type 2B subjects had loss of high molecular weight multimers and abnormal collagen binding, with a VWF:CB/VWF:Ag ratio <0.7. Interestingly, however, not all had a reduced VWF:RCo/VWF:Ag ratio, suggesting VWF:CB would be required in addition to VWF:RCo if multimer distribution was omitted in initial evaluation of this type of VWD. Of 18 type 2M subjects, only one had an abnormal multimer distribution. That subject had no mutations in the VWF coding sequence and normal VWF:CB, although the VWF:RCo/VWF:Ag ratio was low at 0.53. Repeat analysis of a new sample from this subject is pending. All 7 type 2N VWD subjects had normal multimers and VWF:CB/VWF:Ag ratios >0.7. In our population, with the exception of mutations that are yet to be characterized, the combination of VWF:Ag, VWF:RCo and VWF:CB was sufficient to categorize patients as normal, type 1, type 2A, 2B or 2M in the before multimer analysis. These findings suggest that VWF:CB is a sensitive screen for detection of an abnormal multimer distribution. Collagen binding is technically much easier to perform, allowing multimer analysis to be reserved for those cases with low VWF:RCo/VWF:Ag or low VWF:CB/VWF:Ag ratios. Disclosures: No relevant conflicts of interest to declare.