AI-Driven Design Boosts Efficacy of Bi-specific CAR T Cells in Cancer Therapy

Researchers at St. Jude Children's Research Hospital have developed a novel computational approach that significantly streamlines the design of chimeric antigen receptor (CAR) T cells, particularly those targeting two cancer-related proteins simultaneously. CAR T-cell therapy, a form of immunotherapy, involves reprogramming immune cells to recognize and attack tumors by targeting specific surface proteins. However, targeting only one antigen often proves insufficient due to tumor heterogeneity and the potential for cancer cells to evade detection.

To overcome these challenges, the team created a computational pipeline that screens and ranks numerous theoretical tandem CAR constructs. This method predicts factors such as protein folding, stability, and surface expression, thus identifying the most promising designs for further laboratory validation. The top candidates were then experimentally tested in animal models, demonstrating improved tumor clearance and overcoming previous obstacles related to poor surface expression and limited anti-cancer activity.

One of the key applications involved designing a tandem CAR targeting B7-H3 and IL-13Rα2, two proteins associated with pediatric brain tumors. The computationally optimized CARs expressed effectively on T cell surfaces and successfully eradicated tumors in mouse models, outperforming unoptimized versions. These results suggest that this approach can accelerate development timelines, reduce costs, and improve the functionality of bi-specific CAR T cells.

In addition, the researchers constructed a versatile AI-informed tool capable of evaluating up to 1,000 different CAR designs within days. By training on structural and biophysical features of effective CARs, this program predicts the expression and anti-tumor efficacy of new constructs, facilitating a more systematic and efficient design process.

This innovation holds promise for tackling tumors with heterogeneous antigen expression, such as certain solid tumors and brain cancers, where current CAR therapies have limited success. Overall, the integration of computational modeling with experimental validation exemplifies a collaborative approach that could usher in a new era of precision immunotherapy against complex cancers.

Source: https://medicalxpress.com/news/2025-08-ai-approach-car-bi-specific.html