Graph Neural Network and Molecular Docking Simulations of Aptamer-Mediated CAR T-Cell Therapy

CAR T-cell therapy is a revolutionary way to treat cancer. It specifically targets and destroys cancer cells through CARS attached to a patient’s immune cells. Advances in computational simulations have made developing and perfecting CAR T-cell treatments possible. Aptamers are small, single-stranded nucleotide sequences that can bind selectively to target proteins. This study used various computation simulation techniques to find aptamers that bind to cancer cell receptors and CAR T-cells, which could increase the effectiveness of CAR T-cell treatments. AlphaFold 3, a sophisticated machine-learning algorithm intended for protein structure prediction, was used to elucidate the structures of the receptors (PDL1, CD19, CD4, and CD8) under study. Statistical thermodynamic simulations were then run to ascertain the aptamers’ three-dimensional architectures. Afterward, molecular docking simulations were used to investigate the binding positions of aptamers and receptors. As the graph neural network (GNN) proposed, our findings show that the aptamers bind selectively to the receptor binding sites. The PDA-Pred software used scoring techniques to evaluate the interactions and predict the binding energies of the receptor-aptamer complexes. These findings will also aid in developing bispecific aptamers that can identify cancer cells and CAR T-cells simultaneously, opening the door to developing tunable CAR T-cell therapies for various cancer forms.