Immunotherapy using chimeric antigen receptor (CAR)‐engineered lymphocytes has shown impressive results in leukemia. However, for solid tumors such as colorectal cancer (CRC), new preclinical models are needed that allow to test CAR‐mediated cytotoxicity in a tissue‐like environment. Here, we developed a platform to study CAR cell cytotoxicity against 3‐dimensional (3D) patient‐derived colon organoids. Luciferase‐based measurement served as a quantitative read‐out for target cell viability. Additionally, we set up a confocal live imaging protocol to monitor effector cell recruitment and cytolytic activity at a single organoid level. As proof of principle, we demonstrated efficient targeting in diverse organoid models using CAR‐engineered NK‐92 cells directed toward a ubiquitous epithelial antigen (EPCAM). Tumor antigen‐specific cytotoxicity was studied with CAR‐NK‐92 cells targeting organoids expressing EGFRvIII, a neoantigen found in several cancers. Finally, we tested a novel CAR strategy targeting FRIZZLED receptors that show increased expression in a subgroup of CRC tumors. Here, comparative killing assays with normal organoids failed to show tumor‐specific activity. Taken together, we report a sensitive in vitro platform to evaluate CAR efficacy and tumor specificity in a personalized manner.

Chimeric antigen receptor (CAR)‐engineered natural killer (NK) cells represent a promising effector cell type for adoptive cancer immunotherapy. Both, genetically modified donor‐derived NK cells as well as continuously expanding NK‐92 cells are currently under clinical development. To enhance their therapeutic utility for the treatment of pre‐B‐cell acute lymphoblastic leukemia (B‐ALL), we engineered NK‐92 cells by lentiviral gene transfer to express a FMS‐like tyrosine kinase 3 (FLT3)‐specific CAR that contains a composite CD28‐CD3ζ signaling domain. FLT3 has primarily been described as a therapeutic target for acute myeloid leukemia, but overexpression of FLT3 has also been reported in B‐ALL. Exposure of FLT3‐positive targets to CAR NK‐92 cells resulted in conjugate formation between NK and leukemia cells, NK‐cell degranulation and selective cytotoxicity toward established B‐ALL cell lines and primary blasts that were resistant to parental NK‐92. In a SEM B‐ALL xenograft model in NOD‐SCID IL2R γnull mice, treatment with CAR NK‐92 but not parental NK‐92 cells markedly inhibited disease progression, demonstrating high antileukemic activity in vivo. As FLT3 is known to be also expressed on precursor cells, we assessed the feasibility of incorporating an inducible caspase‐9 (iCasp9) suicide switch to enhance safety of our approach. Upon addition of the chemical dimerizer AP20187 to NK‐92 cells coexpressing the FLT3‐specific CAR and iCasp9, rapid iCasp9 activation was observed, precluding further CAR‐mediated cytotoxicity. Our data demonstrate that B‐ALL can be effectively targeted by FLT3‐specific CAR NK cells which may complement CD19‐directed immunotherapies, particularly in cases of inherent or acquired resistance to the latter.