A Novel Revcar Platform for Switchable and Gated Tumor Targeting

Hematological malignancies are successfully treated with chimeric antigen receptor (CAR) armed T cells. Despite the clinical success, CAR T cell therapy struggles still with some problems including the selection of tumor escape variants and on-target, off-tumor side reactions as well as massive cytokine release and uncontrollability of CAR T cell activity in the patients. In order to enable controllability of CAR T cells and to avoid unspecific side effects, we established a novel switchable, split and adaptable CAR platform technology, termed RevCAR system.

The novel RevCARs lack the single chain variable fragment (scFv) commonly used as extracellular domain in conventional CARs. Instead of the scFv, RevCARs contain only a small peptide epitope as extracellular portion. This design reduces the CAR size, avoids unspecific antigen binding and prevents antigen independent tonic signaling caused by scFv dimerization. As RevCAR T cells do not recognize anything, they are per se inert. Only in the presence of a corresponding bispecific antibody based target module (RevTM) they can be specifically redirected to tumor cells. Therefor RevTMs consist of two scFvs. One recognizes the RevCAR peptide epitope and the other one simultaneously binds to a tumor associated antigen (TAA). By dosing of the RevTM, which has a very short half-life, the reactivity of RevCAR T cells can be switched on and off reversibly. Another advantage is that the RevCAR system can be flexibly adapted to any tumor antigen simply by exchanging the RevTM. Furthermore, the small RevCAR size is favorable for inserting more than one RevCAR in the same T cell thus facilitating the mode of gated targeting which is a highly attractive approach to minimize the risk for on-target, off-tumor toxicities against healthy tissues and to increase tumor specificity of conventional CAR T cells. For 'AND' gate targeting via the RevCAR system, two different RevCARs were constructed and expressed simultaneously in the same T cell. The two RevCARs differed with respect to the extracellular peptide epitope and the intracellular signaling domain. Moreover, the respective transmembrane domain was selected to isolate the respective RevCAR signal. The first RevCAR is designed to transmit the activation signal, the second RevCAR to deliver a costimulatory signal. For efficient RevCAR T cell activation, both RevCARs must be engaged via their respective RevTM which on the one hand binds to one of the two RevCAR epitopes and on the other hand to one of two TAAs expressed on the same target cell. Here, we present two RevCAR/RevTM systems for retargeting of AML cells as well as solid tumor cells including via gated targeting.

In summary, we show proof of concept for a novel switchable RevCAR system that can be used for retargeting of AML cells as well as solid tumors. The novel modular RevCAR platform is characterized by small size, lacks unwanted tonic signaling effects, allows the control of RevCAR T cell activity, enables gated targeting strategies, and can be adapted to any tumor antigen and tumor type.