Towards a novel therapy to treat bleeding after left ventricular assist device implantation in patients with heart failure.

Left ventricular assist devices (LVADs) are successfully used to treat patients waiting for a donor heart or as an alternative for heart transplantation. LVADs are mechanical circulatory support devices, which pump blood from the left ventricle to the aorta. Although LVAD implantation improves the quality of life of advanced stage heart failure patients, bleeding complications are a serious side effect. Indeed, gastrointestinal bleeding is one of the leading severe adverse events of LVAD therapy with an incidence around 20-30%. Bleeding in LVAD patients is caused by a loss of function of the haemostatic protein von Willebrand factor (VWF). VWF is a multimeric protein and its high molecular weight multimers (HMW) are crucial to prevent bleeding. However, HMW VWF multimers disappear in almost all LVAD patients within minutes after implantation of the device. Due to the high shear stress caused by the LVAD, HMW VWF unfolds, exposing its cleavage site for the VWF-cleaving protease ADAMTS13 (A Disintegrin And Metalloprotease with ThromboSpondin type 1 repeats, number 13). The resulting loss of HMW VWF multimers and function, and concomitant bleeding phenotype is known as acquired von Willebrand syndrome (aVWS). Importantly, there is an unmet clinical need to develop a targeted therapy that tackles the underlying cause of bleeding in LVAD-induced aVWS, which is the ADAMTS13-mediated proteolysis of VWF.

Hence we aimed to develop a targeted therapy to treat the bleeding disorder aVWS, which occurs after LVAD implantation in patients with heart failure (AIM 1). The novel therapy is based on blocking the ADAMTS13-induced proteolysis of the haemostatic protein VWF which occurs after LVAD implantation using an in-house developed inhibitory anti-ADAMTS13 monoclonal antibody (mAb). We first aimed to study if blocking ADAMTS13 (using the in-house developed inhibitory anti-ADAMTS13 mAb) could prevent the loss of HMW VWF multimers and VWF function in short- and long-term in vitro LVAD systems. We indeed demonstrated that blocking ADAMTS13 using the inhibitory anti-ADAMTS13 mAb prevented the shear-induced proteolysis of VWF by ADAMTS13 when perfusing human blood through both Impella® (short-term) and Heartmate II™ (long-term) devices. Moreover, we could show that blocking bovine ADAMTS13 also prevented the shear-induced proteolysis of VWF in an in vitro Impella system, similar as what was observed with human blood. Next, we aimed to investigate if blocking ADAMTS13 could reverse aVWS in a preclinical LVAD-induced aVWS animal model and also aimed to evaluate the safety of the novel therapy. Importantly, blocking ADAMTS13 in a preclinical LVAD-induced aVWS calf model, could reverse the loss of HMW VWF multimers. Moreover, ADAMTS13 inhibition in the calves did not lead to thrombotic thrombocytopenic purpura (TTP) signs and symptoms (thrombocytopenia, hemolytic anemia and organ damage). On the contrary, blocking ovine ADAMTS13 could not prevent the loss of HMW VWF multimers in an in vitro LVAD system or reverse aVWS in a preclinical LVAD-induced aVWS sheep model, indicating that in sheep other mechanisms (besides ADAMTS13) are responsible for the VWF degradation.

Although the presence of aVWS in LVAD patients on short term (<4 months following device implantation) follow-up is well described, studies regarding the VWF profile in patients on long-term support (>4 months to years after device implantation) are less well studied. Therefore, we aimed to perform a long-term follow up of VWF (VWF multimer size, antigen and activity) and ADAMTS13 (ADAMTS13 activity and antigen) parameters in LVAD patients (AIM 2). We demonstrated that aVWS is not only present early after LVAD implantation but persists over time and that LVAD patients have a decreased ADAMTS13 activity/ADAMTS13 antigen ratio. Hence, our novel therapy would be beneficial for the treatment of bleedings occurring in LVAD patients not only shortly after device implantation but during the complete period after device implantation.

Besides an acquired defect in VWF (aVWS), inherited qualitative or quantitative abnormalities in VWF are associated with congenital von Willebrand disease (VWD). Patients with VWD type 2A or aVWS as a consequence of LVAD implantation are both characterized by a decrease of HMW VWF multimers and a loss of VWF function (VWF collagen binding activity (VWF:CB/VWF:Ag) and VWF ristocetin cofactor activity (VWF:RCo/VWF:Ag)). Loss of VWF function is however more severe in VWD type 2A than in LVAD patients and aVWS detection is thus less straightforward in LVAD patients compared to VWD type 2A patients. Hence, we aimed to compare VWF function in patients with VWD type 2A and LVAD-induced aVWS to highlight the differences in VWF function and to stress the importance of VWF multimer analysis for correct diagnosis of aVWS in LVAD patients (AIM 3). We showed that VWF:CB/VWF:Ag or VWF:RCo/VWF:Ag analysis allows detection of impaired VWF function in VWD type 2A but not always in LVAD-induced aVWS patients. In contrast, VWF multimer analysis allowed detection of the loss of HMW VWF multimers in both groups of patients. Hence, performing VWF multimer analysis is the gold standard to detect aVWS in LVAD patients.

In conclusion, inhibiting ADAMTS13 function could become a promising therapeutic strategy to rescue aVWS-induced bleeding in patients with both short- and long-term LVAD support and ultimately in any disorder where increased VWF proteolysis leads to the loss of HMW VWF multimers, like aortic stenosis, extracorporeal membrane oxygenation (ECMO) support and VWD type 2A.