What's Happening?
Researchers at the Perelman School of Medicine at the University of Pennsylvania have developed an AI framework to identify new targets for CAR T cell therapy, focusing on solid tumors. The system integrates large language models with single-cell RNA
sequencing datasets to generate and refine target lists, which are then evaluated experimentally. The framework was tested on skin cancer, filtering over 10,000 potential antigens to select Glycoprotein non-metastatic melanoma protein B (GPNMB) as a top candidate. The engineered GPNMB-directed CAR T cells showed efficacy in preclinical models, eliminating tumors in melanoma, monoblastic leukemia, and colorectal adenocarcinoma.
Why It's Important?
This development addresses a significant challenge in expanding CAR T therapies beyond blood cancers to solid tumors, which has been a slow and labor-intensive process. The use of AI to streamline target identification could accelerate the development of new cancer treatments, potentially benefiting a wide range of patients with different tumor types. The framework's modular design allows for adaptation to new datasets and evolving AI models, indicating a scalable approach to future cancer therapies.
What's Next?
The Penn team plans to apply this AI framework to additional cancer types and advance the GPNMB CAR T candidate toward clinical trials. The framework's success could lead to broader adoption in the field of cell and gene therapy, potentially transforming how new cancer treatments are developed and tested.
