Directing Reversible Cell-Cell Interactions with Evolved Fibronectin Domains

Targeted molecular recognition continues to be a cornerstone of contemporary therapeutics and diagnostics. Though antibodies are commonly employed as targeting scaffolds, their large size, long plasma half life, and lack of facile bioconjugation make antibody-based ligands less than ideal for many basic and clinical applications. As the repertoire of clinically relevant biomarkers continues to grow, there is an increasing need to develop new agents capable of selectively identifying and targeting these markers.

To meet this need, we have engineered a targeted protein scaffold based upon the human tenth type III fibronectin domain that binds to the highly overexpressed carcinoma antigen epithelial cell adhesion molecule (EpCAM). Using yeast surface display, mammalian cell panning, and a novel titratable avidity-reduction selection technique, we successfully evolved fibronectin clones exhibiting high affinity (11 nM) and robust selectivity for cellular EpCAM. These new selection methodologies can be applied to other cellular biomarkers and should increase the success rate of isolating scaffolds whose binding efficacy translates to in vivo applications.

Furthermore, we incorporated these EpCAM-targeting ligands into a multivalent chemically self-assembled nanoring (CSAN) for use as a cell-directing scaffold. The CSAN is formed when bivalent dihydrofolate reductase (DHFR²) fusion proteins are spontaneously oligomerized by a chemical dimerizer, bis-methotrexate. When the CSAN is functionalized with our anti-EpCAM fibronectins and monovalent streptavidin, heterobifunctional CSANs are formed. These bispecific nanorings are capable of interfacing between EpCAM-expressing tumor cells and biotinylated surfaces – including other cells. Importantly, the CSAN scaffold can be disassembled via exposure to the FDA-approved antibiotic trimethoprim, providing a pharmacologic mechanism for reversing the interactions. Using this system, we have successfully directed and reversed targeted cell-cell interactions in vitro.

In conclusion, we have engineered novel, EpCAM-targeted fibronectin domains and incorporated them into a modular cell-directing scaffold. This CSAN scaffold affords a versatile, non-genetic, and pharmacologically reversible platform for modifying mammalian cell membranes with specific targeting elements. Thus, our CSAN platform is directly applicable to therapeutic regimes such as cell-based immunotherapy and regenerative medicine and may over several advantages over established cell-directing methods.