Coiled-coil masks keep T cell engagers silent until tumor proteases cut them loose
- Anti-CD3 Fabs (OKT3, UCHT1) are sterically masked by leucine-zipper coiled-coil domains fused through protease-cleavable linkers, and paired with PSMA- or HER2-targeting nanobodies.
- Masking strongly lowered CD3 binding, and protease cleavage restored most of it, both on recombinant CD3 and on T cell-displayed CD3.
- Masked constructs did not activate Jurkat cells, whereas unmasked ones did.
- MMP2-cleavable versions drove MMP2-dependent killing of PSMA+ and HER2+ cancer cells in PBMC co-cultures, aiming to limit on-target off-tumor toxicity.
💡 Points
The appeal is that the mask is modular: swap the cleavable linker and the same leucine-zipper scaffold could, in principle, be retuned to a different protease, which is more flexible than hard-wiring one design. What I would press on is the gap between cleaved and unmasked binding, since 'most' restored is not full, and MMP2 is not strictly tumor-restricted. Only Jurkat and PBMC co-cultures here, so the off-tumor safety claim is still an aim, not data.