Epitope Mapping of a Monoclonal Antibody to Factor VIII That Inhibits Factor IXa:Factor VIIIa Interaction and Thrombin Activation of Factor VIII

Abstract 1177

Factor VIII (FVIII) circulates in plasma as a noncovalent heterodimer consisting of a heavy chain (HC, A1-a1-A2-a2-B domains) and a light chain (LC, a3-A3-C1-C2 domains) in a noncovalent complex with von Willebrand factor (wVF). Thrombin (IIa) cleaves FVIII between the A1-a1/A2 domains at Arg372, A2-a2/B domains at Arg740 and B-a3/A3 domains at Arg1689 generating FVIIIa that consists of an A1-a1/A2-a2/A3-C1-C2 heterotrimer. FVIIIa increases the efficiency of Factor IXa (FIXa) catalyzed activation of Factor X (FX) in a Ca²⁺ and phospholipid (PL) dependent manner. The A3-C1-C2 segment of FVIIIa plays an important role in FIXa:FVIIIa interaction. Here, we describe a series of experiments to map the epitope of a monoclonal antibody (mAb) that is reported to inhibit FVIII clotting activity in a one stage clotting assay (Brown et al; J Lab Clin Med, 101: 793–805, 1983).

The binding of mAb to FVIII, B-domain deleted FVIII and isolated LC was assessed using surface plasmon resonance. In these experiments, mAb captured on a protein A/G coupled CM5 sensor chip served as the ligand, and FVIII and its isolated fragments served as the analytes. The Kd of binding of LC (∼40 nM) was similar to FVIII and the B-domain deleted FVIII. No binding was observed for isolated A1 and A2 domains. Further, in plasma based inhibition assays, the Kd of binding of mAb to FVIII-vWF complex and to FVIII was ∼30 nM. This suggests that the mAb epitope does not significantly overlap with the vWF binding site in the acidic a3 region of LC. Western blot analysis confirmed that the mAb is specific for the LC of FVIII. Moreover, IIa-cleaved LC starting at residue 1690 gave only a weak signal and FXa-cleaved LC starting at residue 1721 did not react with the mAb in Western blots. These data suggest that the epitope for this mAb spans the IIa-cleavage site in the LC. Consistent with these observations, the A3-C1-C2 fragment but not the C1-C2 fragment expressed in COS cells reacted with the mAb. To further define a part of the epitope in the IIa-cleaved LC, twelve A3 domain deletion fragments were constructed and expressed in E. coli. Western blot analysis of these fragments restricted the partial epitope to 1690–1710 residues of the IIa-cleaved LC. In additional experiments, the mAb did not inhibit mouse, rabbit or canine plasma FVIII in a one stage clotting assay. It did however inhibit porcine plasma FVIII with ∼40 nM Kd, sheep plasma FVIII with ∼ 68 nM Kd, and bovine plasma FVIII with ∼300 nM Kd. Analysis of the sequence alignment of residues 1680 to 1710 of FVIII from each species indicated that residues 1681 to 1694 of human FVIII most likely constitute the epitope of this mAb. The dissimilarity and the charge differences in amino acids suggest that residues Asp1681, Glu1684, Asn1685, and Ser1687 on the N terminal side and Lys1693 on the C terminal side of the IIa-cleavage site Arg1689-Ser1690 may be important for this epitope.

Fluorescence energy transfer (FRET) experiments indicated that the mAb inhibits FIXa interaction with the IIa-cleaved LC consisting of A3-C1-C2 domains. In these experiments, A3-C1-C2 subunit was labeled with acrylodan (fluorescence donor) and FIXa was labeled with fluorescein-Phe-Phe-Arg-chloromethylketone (fluorescence acceptor). In the presence of FIXa, the acrylodan fluorescence was quenched indicating a biomolecular complex formation. Addition of 1.2 μM mAb abolished the acrylodan fluorescence quenching suggesting inhibition of the FIXa:LC interaction. Notably, the mAb did not inhibit activation of FX by FIXa/Ca²⁺/PL and FXa-cleaved FVIIIa (instead of IIa-cleaved FVIIIa). This suggests that the mAb inhibits FIXa:LC interaction by a steric hindrance and not by a direct blockage of the FIXa:LC interactive sites.

In summary, the mAb inhibits clotting by preventing FVIII activation by IIa. The epitope of the mAb appears to be restricted to residues 1681–1694 of FVIII. Notably, in some of the hemophilia A patients, the epitope of the inhibitory antibodies is confined to the IIa-cleavage site including the a3 acidic domain of LC. To locate the epitope for such antibodies, one of the approaches used was to construct porcine and human FVIII hybrids. Our strategy may represent a simplified approach to locate the epitope of similar antibodies in hemophilia A patients. Such antibodies may bind strongly to LC and weakly to IIa-cleaved LC. Further, these antibodies may not bind to FXa-cleaved LC or A1/A2 subunits.