(205c) Logical Computation By Chimeric Antigen Receptors Increases Tumor-Targeting Capability of Engineered T Cells

T cells expressing tumor-targeting chimeric antigen receptors (CARs) have shown exciting promise in clinical trials, particularly in the treatment of B-cell leukemia. However, important challenges remain in the use of CAR-modified T cells, including off-target toxicity toward normal cells and susceptibility to mutational escape by targeted tumors. Due to a general lack of truly tumor-exclusive antigens, CARs typically target tumor-associated antigens that are also present on a subset of healthy cells, leading to on-target, off-tumor toxicity such as B-cell aplasia observed in patients treated with T cells expressing anti-CD19 CARs. Furthermore, patients who attain complete remission with anti-CD19 CAR-expressing T cells remain susceptible to relapse from the outgrowth of CD19- blast cells—i.e., mutational escape by tumor cells. One strategy to prevent mutation escape is to engineer CARs that recognize multiple antigen inputs and execute killing upon finding any one of the targeted antigens. Alternatively, CARs can be programmed to require the correct combination of input signals before triggering T-cell effector functions, thereby increasing targeting specificity. Here, we present an “OR-Gate” CAR that, when expressed in primary human T cells, efficiently lyse target cells expressing either CD19 or CD20, both of which are common markers of leukemic B cells. We demonstrate this dual-antigen targeting capability through quantitative measurements of target cell lysis and cytokine production. We will also present our designs for a “NOT-Gate” CAR and discuss its implications on our understanding of CAR signaling in relation to natural T-cell receptor signaling. Ongoing work focuses on widening the target differentiation response for the NOT-gate CAR and expanding the technology to additional antigen combinations for improved tumor-targeting specificity in T-cell immunotherapy.