Characterizing and engineering antibodies against the epidermal growth factor receptor

Abstract

Epidermal growth factor receptor (EGFR) signaling leads to cellular proliferation and migration, and thus EGFR dysregulation can significantly contribute to the survival of tumor cells. Aberrant EGFR signaling due to receptor overexpression, mutation, or autocrine ligand has been observed in a wide variety of malignancies, and antibody drugs which inhibit EGFR signaling have been developed. However, the epitopes of most EGFR antibodies have not been characterized, and the marginal efficacies of current antibodies underscore the need for improved therapeutics. In this thesis work, we have created a novel method of epitope mapping, which is the determination of antigen residues responsible for mediating an antibody-antigen interaction. In our technique, a library of random mutants of the EGFR antigen is displayed on the surface of yeast, and the library is combinatorially selected for loss of binding to the antibody being mapped. If a mutant shows loss of binding to an antibody, then that residue is a potential contact residue. In addition, we found that many mutants caused a global misfolding of the antigen, requiring the use of high-throughput sorting to remove the misfolded mutants. The development of our epitope mapping method using random mutagenesis and yeast surface display enabled the successful mapping of four different antibodies and three designed ankyrin repeat proteins binding to EGFR. In addition, we continued work on engineering novel antibodies against EGFR domains II and IV. Antibodies against these domains are hypothesized to directly inhibit receptor dimerization and subsequent activation, as opposed to traditional anti-EGFR antibodies which block ligand binding. To accomplish this, peptide mimics of EGFR loops were used as antigens; however, antibodies generated using both yeast surface display and rabbit monoclonal technology were peptide-specific, but did not bind to EGFR protein.

(cont) Finally, we developed a mathematical model to describe equilibrium EGFR ligand binding and dimerization. Based on observations that polyclonal antibodies against EGFR domain II or IV eliminate high affinity EGF binding normally observed on the surface of cells, we hypothesized that preformed inactive dimers were the high affinity component. Our model incorporating this hypothesis successfully reproduced experimental data, resulting in characteristic concave-up Scatchard plots.