Artificial intelligence-generated “minibinder” protein enhanced T cell binding to and killing of cancer cells.
Cancer immunotherapies largely depend on T cells, which recognise cancer antigens via their receptors interacting with the major histocompatibility complex (MHC). While scientists aim to enhance this interaction, a high-throughput approach has been lacking. In a Science study, data scientist Timothy Jenkins and immunologist Sine Hadrup at the Technical University of Denmark used AI to design de novo T cell receptor (TCR)-like proteins termed mini-binders that bind specifically to predetermined cancer antigens based on their MHC-bound crystal structures.
Using their platform, the team rapidly developed a minibinder that helped T cells kill cancer in vitro taking only weeks compared to the typical 18 months. As proof of concept, they targeted NY-ESO-1, a cancer-associated antigen with a known crystal structure, generating multiple minibinders. AlphaFold2 predictions for MHC-binding were validated in vitro, and cryoelectron microscopy confirmed close alignment with computational models. Engineered T cells expressing the lead mini-binder killed NY-ESO-1-positive melanoma cells more effectively than unmodified T cells.
The platform also succeeded with a metastatic melanoma neoantigen lacking a solved crystal structure, producing proteins that bound strongly in vitro.
Jenkins and Hadrup envision future applications in cancer therapy and beyond, including autoimmune diseases, by developing soluble binders to block MHC–T cell interactions. “This is just the beginning of many cool applications of de novo protein design,” said Jenkins, noting the potential for clinical translation.
Image credit: Monica Fernandez-Quintero and Johannes Loeffler, (The scientist)
Written by, Andrea Luis, Ph.D.
The article can be accesed on: The Scientist
