Rational humanization of the powerful antithrombotic anti-GPIbalpha antibody: 6B4.

Fab-fragments of the monoclonal antibody 6B4, raised against human glycoprotein Ibalpha (GPIbalpha), have a powerful antithrombotic effect in baboons by blocking the GPIbalpha binding site for von Willebrand factor (VWF), without significant prolongation of the skin bleeding time. In order to bring this antibody to the clinic,we here humanized for the first time an anti-human GPIbalpha by variable-domain resurfacing guided by computer modeling. First, the genes coding for the variable regions of the heavy and light chains of 6B4 were cloned and sequenced. Based on this, a three-dimensional structure of the Fv-fragment was constructed by using homology-based modeling, and with this and comparison with antibodies with known structure,"murine" putative immunogenic residues which are exposed, were changed for "human-like" residues. The humanized Fab-fragment, h6B4-Fab, was constructed in the pKaneo vector system, expressed and purified and showed in vitro an unaltered, even slightly higher binding affinity for its antigen than the murine form as determined by different ELISA set-ups and surface plasmon resonance. Finally, injection of doses of 0.1 to 1.5 mg/kg of h6B4-Fab in baboons showed that both pharmacokinetics and ex-vivo bio-activity of the molecule were to a large extent preserved. In conclusion,the method used here to humanize 6B4 by resurfacing resulted in a fully active derivative, which is now ready for further development.

Humanization by variable domain resurfacing and grafting on a human IgG4, using a new approach for determination of non-human like surface accessible framework residues based on homology modelling of variable domains.

Many antithrombotic agents have only a small therapeutic window, frequently leading to bleeding problems. However, interfering with platelet adhesion through the collagen-VWF-GPIbalpha axis is expected to cause less bleeding problems. Our group developed a monoclonal antibody, 82D6A3, directed against the von Willebrand factor (VWF) A3-domain, which inhibits the VWF-interaction to fibrillar collagen. 82D6A3 has antithrombotic effects in vivo without bleeding time prolongation. To further investigate the promising features of 82D6A3, we have humanized it by variable domain resurfacing and grafting on the constant regions of a human IgG4. First, the sequence of the variable domains was determined and the murine scFv was constructed. The expressed scFv had a comparable activity as the IgG of 82D6A3, and its DNA was thus used in subsequent humanization procedures. For this, a new approach was introduced to identify non-human like framework surface residues, since the general distribution of accessible residues described for human and murine heavy and light chain variable domains showed several discrepancies with the homology modelled Fv of 82D6A3. Identification of non-human like framework residues and evaluation of their surface accessibility within the context of the homology modelled Fv of 82D6A3, revealed 10 residues that need to be humanized without influencing the conformation of the CDR loops. Indeed, the humanized scFv of 82D6A3, obtained by mutating all 10 residues to their human counterpart, was still binding with high affinity to VWF and retained the inhibitory properties of the murine scFv. Next, in order to increase its half life and to decrease its immunogenicity, the humanized variable domains were grafted on the constant regions of a human IgG4, resulting in h82D6A3 with an in vitro activity comparable to that of the murine IgG.