T cell display of masked protease-activatable receptor for mapping the protease substrate repertoire

Introduction: : Proteolysis plays a central role in normal physiology and disease. Dysregulated protease activity drives cancer progression, autoimmune disorders, and inflammatory diseases, highlighting proteases both as biomarkers for early diagnosis and as therapeutic targets for safer, more precise treatments. Advances in biomedicine increasingly rely on drugs that activate selectively at disease sites in response to endogenous or exogenous cues, enabling fields such as optogenetics, thermogenetics, biological switches, and conditional biologics. Among these, conditionally activated agents that exploit proteases as local triggers—such as prodrug forms of antibodies, chimeric antigen receptors, and bispecific T cell engagers—use peptide mimotopes that block antigen binding until cleaved by disease-associated proteases, minimizing off-target toxicity. These systems depend critically on selecting appropriate protease substrates to achieve spatially restricted activation. Traditional methods for substrate selection, including genetically encoded peptide libraries displayed on phage, yeast, or bacteria, have largely focused on binding to tissues rather than cleavage by proteases. To address this gap, we developed MAsked Protease-activatable Receptors (MAPRs), which present randomized peptide substrates on T cells for direct functional screening of protease activity. Using MAPRs, we profile the substrate specificity of recombinant proteases in vitro by next-generation sequencing of reporter-positive T cells. Substrates identified through this activity-based platform offer a new resource for designing protease-activatable agents with improved selectivity and therapeutic potential.

Materials and

Methods: : We engineered an anti-human epidermal growth factor receptor 2 (αHER2) MAPR connected via a protease-cleavable linker. Upon linker cleavage by extracellular protease activity, the scFv is unmasked, enabling HER2 binding, activation, and BFP reporter expression. To assess the ability of MAPR T cells as the platform for protease substrate mapping, we first tested protease-selective cleavage. MAPR T cells encoding substrates selectively cleaved by thrombin (Thrb) or a GGS control linker were incubated with recombinant Thrb and HER2+ cancer cells and BFP activation was measured by flow cytometry. For substrate profiling, we generated a lentiviral αHER2 MAPR library with randomized four-amino acid substrate linkers to produce a theoretical diversity of 160,000 unique sequences. T cells co-expressing this library and a BFP reporter were incubated with individual proteases (100 nM) from a panel of 25 enzymes spanning all major catalytic classes (serine, cysteine, threonine, aspartic, metalloprotease). After 24 hours of co-culture with HER2+ cancer cells, BFP reporter-expressing T cells were sorted and analyzed by next-generation sequencing (NGS) to quantify fold-changes in substrate frequencies relative to untreated controls. Bootstrapping analysis identified and removed substrates affected by sampling bias. For validation, we ranked substrates by log2 fold-change and selected one sequence from the top 20 for each protease. We then engineered 25 monoclonal MAPR T cells, each expressing a distinct substrate, and tested receptor activation against each individual protease to confirm functional specificity.

Results, Conclusions, and Discussions:: We designed T cells to express MAPRs that contain both a protease substrate domain and a target binding domain such that a BFP, fluorescent reporter protein is produced only when the substrate is cleaved by a protease and the binder engages with its cognate ligand to allowed FACS isolation of reporter-positive T cells which were then identified by NGS sequencing to identify cleavage substrate (Fig 1A). To validate the system, primary human T cells expressing Thrb-cleavable or control (GGS) αHER2 MAPRs showed BFP activation only when both Thrb and HER2-positive cancer cells were present (+/+ condition, Fig. 1B). We then explored whether MAPR T cell display could map protease substrate preferences. A library of αHER2 MAPRs with fully randomized four–amino acid linkers was introduced into primary human T cells (Fig. 1C). T cells were incubated with 25 individual proteases across all five major catalytic classes, alongside HER2-positive cancer cells. After 24 hours, BFP+ T cells were sorted and subjected to NGS, revealing distinct substrate profiles for each protease (Fig. 1D). For validation, we selected one top substrate from each protease cluster, engineered monoclonal MAPR T cells, and tested their activation across on- and off-target proteases. Impressively, all 25 sensors were significantly activated by their target protease (P < 0.0001), and 14 showed high specificity with no cross-reactivity (P > 0.05) (Fig. 1E). These included substrates cleaved by collagenases, cathepsins, and caspases despite the redundancy in protease selectivity for native protein substrates within these families. Together, these results highlight MAPR T cell display as a powerful, activity-based screening platform for comprehensive mapping of human protease substrate repertoires, with broad applications in drug development and precision biologics.

Acknowledgements and/or References (Optional)::