Engineered Multivalency for Enhanced Efficacy of HER3-Targeted Affibodies

HER3 (also known as ErbB3), a member of the HER receptor tyrosine kinase (RTK) family, is a potent mediator of tumor progression in breast, lung, ovarian, and other cancers. Additionally, HER3 has been implicated in mediating resistance to EGFR-, HER2-, and PI3K-targeted drugs (among others). Yet, while some candidate molecules have entered clinical trials, there are currently no FDA-approved HER3-targeted therapeutics. Even if one or several drugs targeting HER3 do eventually receive approval, there is still likely to be a large proportion of refractory patients; for example, currently approved HER2-targeted therapeutics typically show efficacy in only ~24%-64% of patients with HER2+ metastatic breast cancer in clinical trials. Thus, development of new HER3-targeted therapeutics would fill a critical clinical need in cancer therapy.

HER3 is unique amongst the HER family in that it has an inactive or very weak kinase domain, necessitating its partnership with an active RTK via dimerization to induce downstream signaling. In addition, HER3 exists in homo-oligomeric clusters on cancer cell membranes, with binding of the HER3 ligand Neuregulin 1B (NRG) leading predominantly to formation of potent pro-neoplastic HER3-HER2 heterodimers. Based on this knowledge, we hypothesized that multivalent ligands with multiple high affinity HER3 binding domains separated by a flexible peptide spacer could sequester HER3 into homotypic configurations that prevent its interaction with HER2 or other potential partners (e.g. EGFR, c-MET, IGF-1R), thus inhibiting pro-neoplastic signaling. To test this hypothesis, we first engineered a novel class of bivalent HER ligands and demonstrated that bivalent NRG can inhibit HER3 signaling in cancer cells from diverse tissue origins, inducing apoptosis and reducing chemokinesis. We have now developed a second generation of multivalent HER3 ligands based on engineered affibody domains that have increased translational potential compared to bivalent NRG due to reduced risk of pro-neoplastic side effects.

Crucially, multivalent HER3 affibodies are more effective at inhibiting HER3 signaling (HER3 and Akt phosphorylation) and reducing cell survival across multiple cancer cell types than monovalent versions of the same affibody. Mechanistically, multivalency endows HER3 affibodies with the ability to downregulate or degrade HER3, whereas minimal degradation is observed with monovalent affibodies. Current investigations are assessing whether this multivalent effect is specific to HER3 or can be applied to other RTK targets as well. Multivalent HER3 affibodies are well tolerated by mice and work most effectively in combination with other drugs, specifically trastuzumab and small molecule chemotherapeutics such as buparlisib. In summary, using an approach fundamentally different than the HER3-targeted therapies that are currently in clinical trials, we have established that HER3 sequestration via engineered multivalent ligands is a versatile approach with broad therapeutic potential in cancer.