The epidermal growth factor receptor (EGFR) is a validated target in oncology. Yet the two FDA-approved antibodies against EGFR exhibit only modest therapeutic efficacy. We hypothesize that these deficiencies result from incomplete receptor inhibition that arises from autocrine ligand stimulation, untargeted EGFR variant III, insufficient receptor downregulation, and endocytic antibody consumption. We aim to engineer binders to multiple epitopes on the EGFR ectodomain for use in the study of these mechanisms and the development of improved therapies. The fibronectin type III domain is used as the scaffold because its single-domain, cysteine-free architecture facilitates the creation of bispecific constructs and enables biophysical FRET studies, and its small size (10 kDa) yields a pharmacokinetic profile that is well-suited for in vivo imaging.

Yeast surface display libraries were created with designed diversity in three solvent exposed loops of the tenth type III domain of human fibronectin. These libraries were screened for binders to the EGFR ectodomain using magnetic bead and flow cytometry sorting. Affinity maturation of lead clones was conducted using error-prone PCR and loop shuffling. Numerous clones were identified with affinities in the picomolar to nanomolar range. Antibody competition assays indicate that, collectively, these binders target multiple epitopes. The effect of internalization and downregulation of EGFR using both monovalent, bivalent, and bispecific constructs will be presented. The impact of affinity, binding epitope, and valency will be elucidated through the use of multiple binder combinations and formats. Implications for therapeutic development will be discussed.