Cooperative Interaction Between Classical and Non-Classical Factor VIII C2 Domain Antibody Epitopes.

Abstract 219 The immune response to factor VIII (fVIII) currently is the most significant complication in the management of patients with hemophilia A. In addition, antibodies to fVIII can develop in non-hemophiliacs, producing acquired hemophilia A, which frequently produces life- or limb-threatening bleeding. In either congenital or acquired hemophilia A, the majority of inhibitory antibodies are directed at either the 40-kDa A2 or the 15-kDa C2 domains of fVIII. Classical anti-C2 antibodies inhibit the binding of fVIII and activated fVIII to phospholipid membranes. These antibodies bind to hydrophobic feet in the C2 domain that include loops containing M2199-F2000 and L2251-L2252. Non-classical anti-C2 antibodies inhibit the activation of fVIII by thrombin and factor Xa. They bind to the beta sheet that comprises the front face of the C2 domain that is remote from the phospholipid binding region. In the Bethesda assay, non-classical anti-C2 antibodies inhibit fVIII incompletely at saturating concentrations and thus are type II inhibitors. However, they have inhibitory titers on an equimolar basis that are usually at least 10-fold higher than classical anti-C2 antibodies. von Willebrand factor (VWF) competes with classical but not non-classical anti-C2 bodies for binding to fVIII. We recently found that the binding of classical anti-C2 antibodies to fVIII in a direct ELISA was increased in the presence non-classical anti-C2 antibodies, suggesting the presence of a cooperative interaction between fVIII antibody epitopes. In the present study, we characterized the interaction of highly-purified recombinant full-length human fVIII with a classical anti-C2 antibody, ESH4, in the presence and absence of a non-classical anti-C2 antibody, G99, by surface plasmon resonance spectroscopy. The ligand, ESH4, was covalently attached to a Biacore CM5 biosensor chip and the kinetics of binding of the analyte, fVIII, at concentrations of 10, 20 and 30 nM were measured at 25 °C. When present, G99 (60 nM) was pre-mixed with fVIII for 30 min before the start of the kinetic experiment. In both the presence and absence of G99, global nonlinear least-squares analysis indicated that the binding of fVIII to ESH4 was first-order in both the adsorption and desorption phases, consistent with a 1:1 reversible bimolecular interaction (Fig. 1). Both the association and dissociation rates were slower in the presence of G99. Estimated phenomenological association and dissociation rate constants were 0.5 × 10 5 M −1 sec −1 and 1.5 × 10 −4 sec −1 and 1.7 × 10 5 M −1 sec −1 and 4.4 × 10 −4 sec −1 in the presence and absence of G99, respectively. The clinical implication of these results is that infusion of recombinant fVIII (which does not contain VWF) or plasma-derived fVIII (which is bound to VWF) may result in complex kinetic interactions that are significantly influenced by cooperative interactions between anti-C2 antibodies and competitive interactions involving VWF. Disclosures: No relevant conflicts of interest to declare.