Abstract
Human epidermal growth factor receptor 3 (HER3) has been increasingly scrutinized as a potential drug target since the elucidation of its role in mediating tumor growth and acquired therapy resistance. Affibody molecules are so-called scaffold proteins with favorable biophysical properties, such as a small size for improved tissue penetration and extravasation, thermal and chemical stability, and a high tolerance to modifications. Additionally, affibody molecules are efficiently produced in prokaryotic hosts or by chemical peptide synthesis. We have previously evaluated the biodistribution profiles of five mono- and bivalent anti-HER3 affibody molecules (designated as 3) fused to an albumin-binding domain (designated as A), 3A, 33A, 3A3, A33, and A3, that inhibit ligand-dependent phosphorylation. In the present study, we examined the therapeutic efficacy of the three most promising variants, 3A, 33A, and 3A3, in a direct comparison with the HER3-targeting monoclonal antibody seribantumab (MM-121) in a preclinical BxPC-3 pancreatic cancer model. Xenografted mice were treated with either an affibody construct or MM-121 and the tumor growth was compared to a vehicle group. Receptor occupancy was estimated by positron emission tomography/computed tomography (PET/CT) imaging using a HER3-targeting affibody imaging agent [68Ga]Ga-(HE)3-Z08698-NODAGA. The affibody molecules could inhibit ligand-dependent phosphorylation and cell proliferation in vitro and demonstrated tumor growth inhibition in vivo comparable to that of MM-121. PET/CT imaging showed full receptor occupancy for all tested drug candidates. Treatment with 3A and 3A3 affibody constructs was more efficient than with 33A and similar to the anti-HER3 antibody seribantumab, showing that the molecular design of affibody-based therapeutics targeting HER3 in terms of the relative position of functional domains and valency has an impact on therapeutic effect.