The “mini-bind” protein, which produced artificial intelligence, enhanced T-cell binding and killing of cancer cells.
Image credits:Monica Fernandez Quintero and Johannes Loffler
cAncer immunotherapy is widely dependent on T cells. To become active, T cells interact through their receptors with major histocompatibility complexes (MHCs), which express fragments of intracellular proteins, such as cancer antigens, on the cell surface. Scientists aim to enhance the interaction between T cells and cancer antigens, but there is no high throughput approach to doing so.
Sine Hadrup, an immunologist at the Institute of Technology in Denmark, studies the interactions between T cells and cancer.
Sign Had Wrap
To solve this problem, Timothy Jenkins, a data scientist at the Institute of Technology of Denmark, and Sign Hadrup, an immunologist, used AI to design de novo T-cell receptor (TCR)-like proteins that specifically bind to regular cancer antigens based on known crystal structures of complexes with MHC.1 recently Science Study, Jenkins, Hadrup, and their colleagues discovered a designer protein called Minibinder, which helped T cells kill cancer in vitro. Their discovery took only a few weeks, in contrast to a typical year and a half. Researchers also demonstrated that their platform could function on targets with unknown crystal structures.
“This is a very cool paper,” said Christopher Klebanov, an oncologist and immunobiologist at Memorial Sloan Kettering Cancer Center, who was not involved in the study. “Very very very cool application for generation [artificial intelligence] Solving truly, truly challenging problems in this field. ”
Christopher Klebanov, an oncologist and immunobiologist at Memorial Sloan Kettering Cancer Center, is studying T-cell-based cancer immunotherapy. Klebanoff was not involved in the study.
Christopher Klebanov
As proof of their tool's concept, the team used the platform to design a protein that can bind to the cancer-associated antigen NY-ESO-1, and killed the cells that express it.2 Hadrup said the team chose NY-ESO-1 as the target as the crystal structure of the complex with MHC was resolved. “We wanted to make sure the structure of the target is not a limitation of the computational process,” she said.
This platform has brought several minibinders. “What generative AI came up with was appealing to me because it didn't look like what T-cell receptors usually look like,” Klebanoff said. The next step was to check whether these structurally diverse minibinders matched T cell receptor function.
The researchers used Alphafold2 to predict the MHC binding properties of 44 minibinders and tested the binding and specificity in vitro. They observed that one mini-binder, whose computational model was shown as the most promising, is most strongly bound to MHC. Furthermore, extreme electron microcopy of the crystal structure of the MHC and complex minibinding showed high overlap with the computational model. The team also observed that T cells designed to express superficial minivines killed NY-ESO-1 positive melanoma cells more effectively than their unmodified counterparts.
“What's very encouraging is that what we saw in the wet lab correlates very well with the computational predictions we made,” Hadrup said.
Timothy Jenkins, a data scientist at the Institute of Technology of Denmark, is developing a new de novo protein design-based approach for translational research applications.
Timothy Jenkins
The team then tested the platform on even more difficult tasks. They designed minibinders to target metastatic melanoma neotigens whose MHC-bound crystal structure has not been resolved. They ended up with some proteins that were strongly linked to neoantigen in vitro.
“There was a lot of skepticism about the design of de novo proteins. [2024] The Nobel Prize has awakened many people in terms of how powerful this technology is,” Jenkins said.
“The driving vision here is not just about designing something academically interesting, but also about seeing something work in the clinic,” Jenkins said of plans for the future.
Hadrup added that researchers can develop these molecules as soluble binders to block interactions between MHC and T cells in autoimmune diseases. “These other areas of treatment have not yet been explored, and that's something we want to demonstrate in the future,” she said.
