Warfarin Exposure Leads to Loss of Binding of Emicizumab to Factor IX

Introduction

Coagulation factor IX (FIX) is a serine protease critical in hemostasis. We previously comprehensively characterized the impact of FIX missense variants on FIX secretion and carboxylation-sensitive antibody phenotypes using multiplexed surface-tethering of extracellular proteins (MultiSTEP), our mammalian cell display system for interrogating secreted proteins. We hypothesized that MultiSTEP could be used to study binding of emicizumab to FIX. Emicizumab is a therapeutic approved for use in individuals with hemophilia A. Emicizumab is a humanized bi-specific antibody that mimics some functions of coagulation factor VIII (FVIII) by binding both FIX / activated factor IX (FIXa) and factor X (FX), facilitating the conversion of FX to FXa. However, the mechanisms important for emicizumab’s procoagulant activity are only partially understood. We tested the capacity of our FIX MultiSTEP system to detect FIX-emicizumab binding and studied the importance of FIX carboyxlation in FIX-emicizumab binding under varying conditions of vitamin K supplementation and warfarin exposure.

Methods

We cloned an F9 cDNA representing the common F9 human nucleotide sequence into our MultiSTEP expression cassette. The F9 cassette encodes a C-terminal flexible linker, StrepII tag, and CD28 transmembrane domain, resulting in tethering of expressed FIX to the surface of the cell that made it (FIX-TM). We recombined the FIX-TM transgene into a doxycycline-inducible, genomically-integrated landing pad in HEK293 Freestyle cells. The integration event also results in loss of expression of the blue fluorescence protein landing pad control and, in cells expressing the FIX-TM cassette, gain of mCherry expression controls. Following induction of FIX-TM expression, cell cultures were supplemented with 50nM vitamin K without warfarin or vitamin K plus 100nM warfarin sodium. We tested antibody binding to displayed FIX under all conditions to assess the expression of FIX, carboxylation of the FIX Gla domain, and emicizumab-FIX binding by incubating cells with monoclonal antibodies to FIX light chain (GMA-124), heavy chain (GMA-102), carboxylated FIX Gla domain (GMA-001), and emicizumab at at concentrations ranging from 0.15-1.5 mg/mL. Antibody binding was detected with fluorescently-conjugated secondary antibodies and analyzed by fluorescence intensity via flow cytometry.

Results

FIX is robustly expressed in our system in the presence of both vitamin K and warfarin, as determined by antibody binding to FIX heavy and light chains. As expected, Gla domain carboxylation-sensitive antibody showed robust signal in the vitamin K-supplemented cells but minimal (near background) signal after warfarin exposure. Emicizumab-FIX binding is well detected in our system. Surprisingly, we only observed emicizumab binding to vitamin K-supplemented FIX-expressing cells, while cells incubated with vitamin K and warfarin showed profound loss of emicizumab binding similar to negative controls.

Conclusions

Our data show that MultiStep is useful to interrogate emicizumab-FIX binding events and indicates that emicizumab is capable of robustly binding FIX prior to activation to FIXa. We unexpectedly find that warfarin exposure ablated the ability of emicizumab to bind to FIX. This indicates FIX carboxylation is a key post-translational modification for emicizumab binding in our system. This finding offers new insights into the interaction of emicizumab with FIX that would not be able to be detected with conventional FIX activity assays. Additionally, this work suggests there is the potential for reduced efficacy if emicizumab were to be used clinically in settings of perturbed vitamin K metabolism.