UCSF

Engineered T cell therapies have emerged as a powerful therapeutic modality, but advancing their efficacy and accessibility requires a deeper understanding of the molecular mechanisms that govern T cell function and new approaches for precise genetic manipulation. This dissertation explores how synthetic receptor signaling can be systematically optimized and how targeted genome engineering can enable new modes of cellular immunotherapy.Through comprehensive mutational analysis of synthetic receptor signaling domains, we identified specific amino acid substitutions that substantially enhance T cell persistence and antitumor activity across multiple disease models. We then integrated these experimental datasets with protein language models, showing that limited functional measurements can markedly improve prediction of synthetic signaling phenotypes and enable identification of high-performing receptor variants. Finally, we developed strategies for targeted gene integration directly in T cells in vivo using engineered adeno-associated virus (AAV) vectors and genome editing technologies, establishing a foundation for cellular engineering without ex vivo manufacturing.Together, these studies connect functional genomics, machine learning, and in vivo genome engineering to advance a more predictive and scalable approach to T cell therapy design. The work provides new insights into the molecular determinants of T cell function while enabling technologies that may accelerate the development and deployment of next-generation cellular immunotherapies.