Th39 - The Role of Chimeric Antigen Receptor (CAR) Versus Endogenous Co-stimulation in the Function of Alloantigen-specific Tregs

Cell therapy with regulatory T cells (Tregs) can diminish rejection and their efficacy can be improved by expression of an alloantigen-specific CAR. In pursuit of optimized CARs for use in Tregs, we and others explored the effect of CAR-encoded co-stimulatory domains finding that inclusion of a CD28 co-stimulatory domain is optimal for human CAR-Treg. However, a limitation was the use of immunodeficient mice which lack a full complement of antigen presenting cells. Here we aimed to re-visit the question of optimal CAR design in an immunocompetent Bl/6-based mouse model of skin transplantation. CARs specific for HLA-A2 encoding CD3z with or without intracellular domains from CD28, PD1, ICOS, GITR, OX40 or 41BB were generated and expressed in mouse Tregs. In vitro proliferation and cytokine production revealed superior effects of the CD28-encoding CAR, confirming data from human CAR-Tregs. Surprisingly, in vivo CAR-Treg function showed no significant difference between CD28 and CD3zeta alone CAR-Tregs to extend allograft survival. Moreover, we found that CD28 and CD3zeta CAR-Tregs were equal in their ability to suppress expression of CD86 and CD80 on HLA-A2+ splenic DCs. Hypothesizing that co-stimulation from the DCs could influence CAR-Treg function, we tested effects of CD86+ or CD86- HLA-A2+ cells, finding that CD86-mediated stimulation through endogenous CD28 can partially replace the requirement for a CAR-encoded CD28 domain. Our results demonstrate that CAR-Tregs require co-stimulation via CD28, which can be delivered via the CAR or DC-CAR-Treg interactions. Optimization of CAR-design should consider signals mediated by physiologically-relevant cell-cell interactions and use fully immunocompetent models.