Engineering a humanized bispecific F(ab')2 fragment for improved binding to T cells

We recently constructed a humanized bispecific antibody (BsF(ab')2v1) by separate E. coli expression of each Fab' arm followed by directed chemical coupling in vitro. BsF(ab')2 v1 (anti-CD3/anti-p185HER2) was demonstrated to retarget the cytotoxic activity of human CD3+ CTL in vitro against the human breast-tumor cell line, SK-BR-3, which over-expresses the p185HER2 product of the proto-oncogene HER2. Our minimalistic humanization strategy is to install as few murine residues as possible into a human antibody in order to recruit antigen-binding affinity and biological properties comparable to that of the murine parent antibody. This strategy proved very successful for the anti-p185HER2 arm of BsF(ab')2 v1. In contrast BsF(ab')2 v1 binds to T cells via its anti-CD3 arm much less efficiently than does the chimeric BsF(ab')2 which contains the variable domains of the murine parent anti-CD3 antibody. Here we have constructed additional BsF(ab')2 fragments containing variant anti-CD3 arms with selected amino acid replacements in an attempt to improve antibody binding to T cells. One such variant, BsF(ab')2 v9, was created by replacing 6 residues in the second hypervariable loop of the anti-CD3 heavy chain variable domain of BsF(ab')2 v1 with their counterparts from the murine parent anti-CD3 antibody. BsF(ab')2 v9 binds to T cells (Jurkat) much more efficiently than does BsF(ab')2 v1 and almost as efficiently as the chimeric BsF(ab')2. This improvement in the efficiency of T-cell binding of the humanized BsF(ab')2 is an important step in its development as a potential therapeutic agent for the treatment of p185HER2 over-expressing cancers.