A new study published Thursday (July 24th) in Journal Science shows how generative AI can create small proteins that dramatically enhance the cancer killing power of human immune cells. Lab tests show that these proteins are quickly placed and killed T cells, a fast, deadly form of skin cancer.
Dr. Timothy Jenkins, a medical biotechnologist at the Institute of Technology of Denmark, said: “We help T cells find tumors much more efficiently than they do.”
This study is an early stage form of immunotherapy, a therapeutic approach that boosts the body's natural defenses to combat disease. The idea is based on successful techniques like CAR T-cell therapy, but it gives a big twist. It can take months to isolate and test rather than rely on naturally occurring cell receptors.
Using AI Tools
To create a molecular guide, researchers used three AI tools. First, we analyzed the structure of a specific cancer target known as NY-ESO-1, a protein found in many tumor cells, using a generative model called RFDiffusion. Next, another model proposed an amino acid sequence that folds into the correct shape to bind the target. The third model helped narrow the only 44 options tested in the lab to just 44. One design stood out and it worked. This study is based on a wave of recent breakthroughs in computational biology, including the technology that led to the 2024 Nobel Prize in Chemistry for Protein Structure Prediction. Earlier this year, Jenkins' team also developed AI-designed proteins to improve Snakebite anti-vite. Clinical trials are still years away, and many questions remain about how proteins function in the human body. However, researchers are optimistic.
