Proteases play a crucial role in biological functions such as tumor progression and tissue homeostasis. Recently, protease-activated prodrugs have gained attention for their potential to enhance selectivity in tumor-targeted therapies. In this study, we report the engineering of substrate sequences for matriptase, a protease overexpressed in tumors and previously explored for prodrug activation in vivo. A peptide library containing millions of potential substrates was displayed on Escherichia coli, and flow cytometric sorting was used to isolate improved substrates based on cleavage efficiency. Hits were ranked by flow cytometry, and the top substrates exhibited kcat /KM values over 40-fold higher than previously reported sequences. These substrates were further evaluated in an antibody-prodrug format, demonstrating exceptional activation. The matriptase substrates hold broad potential for applications such as cleavable linkers in next-generation antibody prodrugs. Furthermore, the developed bacterial display platform shows promise for discovering substrates of other proteases.

The application of monoclonal antibodies as targeted drugs has revolutionized cancer therapy, yet their efficacy is sometimes limited by toxicities arising from target expression in healthy tissues. Various strategies have emerged to overcome these issues, including the design of antibody-based prodrugs . Here, we report the generation of affibody masking domains designed to specifically bind and mask the paratope of the anti-epidermal growth factor receptor (EGFR) antibody, cetuximab. Bacterial display and cell sorting techniques were employed to isolate suitable affibody masking domains. Both experimental and computational analyses confirmed that several of these selected affibodies effectively blocked cetuximab from binding EGFR. A cetuximab prodrug was created by fusing the most promising masking candidate to the antibody heavy chains. This prodrug demonstrated over 400-fold difference in functionality between its masked and unmasked states. This study is the first to describe the use of an affibody masking domain to create an antibody-based prodrug, demonstrating significant potential for further development of safer and more efficacious cancer therapies. 

Proteases play a crucial role in various biological functions, including tumor progression and homeostasis. Recently, protease-activated prodrugs have gained attention for their potential to increase selectivity in tumor-targeted therapies. In this study, we report the engineering of novel substrate sequences for matriptase, a protease overexpressed in tumors and previously explored for prodrug activation in vivo. A combinatorial peptide library containing millions of potential substrates was displayed on the outer membrane of Escherichia coli, and flow-cytometric cell sorting was employed to isolate improved substrates based on their cleavage efficiency by matriptase. Hundreds of hits from the sorting process were ranked by flow cytometry, and the top substrates exhibited kcat/Km values more than 40-fold higher than previously reported substrates. These new substrates were further evaluated in an antibody-prodrug format, demonstrating exceptional prodrug activation. The matriptase substrates identified in this study hold broad potential for various applications, including their use as cleavable linkers in next-generation antibody prodrugs. Furthermore, the developed bacterial display platform offers promise for discovering other protease substrates.  

Antibody prodrugs provide a strategy to reduce systemic toxicity by masking therapeutic antibodies until activation in specific environments. Nivolumab, an anti-PD-1 antibody used in cancer immunotherapy, can lead to immune-related adverse events. In a first step towards creating a prodrug version of nivolumab, we screened an Escherichia coli affibody library using MACS and FACS, identifying affibodies that effectively mask the PD-1-binding regions. Deep sequencing revealed an unexpected enrichment of proline-rich affibodies, which likely mimic a loop in PD-1. Structural modeling using AlphaFold suggested that these proline-rich affibodies form stable interactions with nivolumab, despite their lower alpha-helical content. Biosensor assays confirmed effective masking by these affibodies in a nivolumab prodrug format, with PD-1 binding restored upon specific proteolytic cleavage. These results warrant further investigation into the use of such PD-1-mimicry peptides as masking domains, paving the way for the development of more effective and safer nivolumab prodrugs.