Engineered Fn3 protein has targeted therapeutic effect on mesothelin-expressing cancer cells and increases tumor cell sensitivity to chemotherapy.

Mesothelin is a protein expressed at high levels on the cell surface in a variety of cancers, with limited expression in healthy tissues. The presence of mesothelin on tumor tissue correlates with increased invasion and metastasis, and resistance to traditional chemotherapies, through mechanisms that remain poorly understood. Molecules that specifically recognize mesothelin and interrupt its contribution to tumor progression have significant potential for targeted therapy and targeted drug delivery applications. A number of mesothelin-targeting therapies are in preclinical and clinical development, although none are currently approved for routine clinical use. In this work, we report the development of a mesothelin-targeting protein based on the fibronectin type-III non-antibody protein scaffold, which offers opportunities for applications where antibodies have limitations. We engineered protein variants that bind mesothelin with high affinity and selectively initiate apoptosis in tumor cells expressing mesothelin. Interestingly, apoptosis does not occur through a caspase-mediated pathway and does not require downregulation of cell-surface mesothelin, suggesting a currently unknown pathway through which mesothelin contributes to cancer progression. Importantly, simultaneous treatment with mesothelin-binding protein and chemotherapeutic mitomycin C had a greater cytotoxic effect on mesothelin-positive cells compared to either molecule alone, underscoring the potential for combination therapy including biologics targeting mesothelin.

Fn3 proteins engineered to recognize tumor biomarker mesothelin internalize upon binding.

Mesothelin is a cell surface protein that is overexpressed in numerous cancers, including breast, ovarian, lung, liver, and pancreatic tumors. Aberrant expression of mesothelin has been shown to promote tumor progression and metastasis through interaction with established tumor biomarker CA125. Therefore, molecules that specifically bind to mesothelin have potential therapeutic and diagnostic applications. However, no mesothelin-targeting molecules are currently approved for routine clinical use. While antibodies that target mesothelin are in development, some clinical applications may require a targeting molecule with an alternative protein fold. For example, non-antibody proteins are more suitable for molecular imaging and may facilitate diverse chemical conjugation strategies to create drug delivery complexes. In this work, we engineered variants of the fibronectin type III domain (Fn3) non-antibody protein scaffold to bind to mesothelin with high affinity, using directed evolution and yeast surface display. Lead engineered Fn3 variants were solubly produced and purified from bacterial culture at high yield. Upon specific binding to mesothelin on human cancer cell lines, the engineered Fn3 proteins internalized and co-localized to early endosomes. To our knowledge, this is the first report of non-antibody proteins engineered to bind mesothelin. The results validate that non-antibody proteins can be engineered to bind to tumor biomarker mesothelin, and encourage the continued development of engineered variants for applications such as targeted diagnostics and therapeutics.