Chimeric antigen receptor (CAR) T cells have shown clinical efficacy in refractory B cell malignancies. Despite these exciting clinical results, our fundamental understanding of CAR-T cell biology is limited, possibly limiting the broader application of CAR-T cells to additional haemopoetic and solid cancers. To date, the mechanism of CAR-T immunological synapse formation with tumor cells and kinetics of subsequent serial killing by CAR-T cells has not been explored. Here, we investigated the kinetics of CAR-T cell activation and cytotoxicity, including immune synapse formation, and kinetics of tumor cell killing, closely comparing to activation and cytotoxicity via the T cell receptor (TCR). To address this, we developed a mouse model in which the CD8+ T cells (termed CAR.OT-I cells) co-expressed two antigen receptors, the clonogenic OT-I TCR, and a second generation CAR comprising a scFV to human HER2, CD28 and CD3ζ signaling domains. Effector CAR.OT-I cells were activated via their antigen receptors using either SIINFEKL-pulsed or HER-2 expressing tumor cells, the interactions between the effector CAR.OT-I cells and tumor cells were then assessed by time lapse live microscopy. CAR.OT-I cell activation via the endogenous TCR or the CAR did not affect tumor killing kinetics, except the time taken from CAR.OT-I activation to detachment (from the dying tumor cell) was significantly slower when the endogenous TCR was engaged. Subsequently, we showed for the first time, that CAR.OT-I cells have serial killing capacity, which is important to consider when therapeutic numbers of CAR-T cells are likely to be outnumbered by tumor targets. Individual CAR.OT-I cells killed multiple tumor cells, whether activated via the endogenous TCR or the CAR. We further explored whether these findings have implications for killing of tumor cells using low effector:target cell ratio in short versus long-term killing assays, chromium release and xCELLigence killing assays respectively. We observed, no matter which antigen receptor was activated, the effector CAR.OT-I cells were equivalent killers of tumor cells in short term assays (4-8 hours). However, over a period of 50 hours, CAR.OT-I cells activated via the CAR killed tumor cells at a lower rate than when activated via the TCR. This was due to CAR.OT-I CAR expression down-regulation from 20-50 hours. This study highlights that fundamental differences occur in the way CAR-T cells kill tumor cells, depending on how the effector CAR-T cell is activated. Furthermore, the study provides important insights for CAR-T cell activation in vivo with implications for single- or dual-receptor-focused CAR-T cell therapy and improved clinical benefit.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.