274. One-Step Generation of Universal CAR T Cells

Abstract

Adoptive immunotherapy using chimeric antigen receptors (CARs) has shown remarkable clinical results in the treatment of leukemia and is one of the most promising new strategies to treat cancer. Current clinical protocols utilize autologous T cells that are collected by apheresis and engineered with retroviral vectors to stably express the CAR. This approach therefore requires patient-specific cell manufacturing, which unavoidably results in patient-to-patient variability in the final cell product. Widespread implementation of this approach will further require progress in automation and miniaturization of cell manufacturing to meet the demand for CAR T cells. Furthermore, current approaches utilize randomly integrating vectors, including gamma-retroviral, lentiviral and transposons, which all result in semi-random integration and variable expression of the CAR owing to transgene variegation. Position effects may result in heterogeneous T cell function, transgene silencing and, potentially, insertional oncogenesis. Thus, the conjunction of autologous cell sourcing and random vector integration is prone to generating cell products with variable potency. Here we utilize gene editing to generate histocompatible T cell products with consistent and homogeneous CAR expression. Different tailored nucleases, including CRISPR/Cas9 system, Zinc Finger Nucleases or TAL effector nucleases (TALENs), have been previously used for gene disruption in a wide range of human cells including primary T cells. In some instances, these nucleases have been used to generate so-called “universal T cells” for allogeneic administration, by disrupting T cell receptor (TCR) or HLA class I expression, but viral vectors or the sleeping beauty transposon were used to deliver the CAR cDNA, all of which result in semi-random transgene integration and its downstream consequences. We present here a novel strategy for one-step generation of universal CAR T cells. We first compared the efficiency of TALEN and CRISPR/Cas9 to promote homologous recombination using AAV6 donor template in T cells and established conditions yielding more than 50% of universal CAR T cells combining target gene disruption and CAR insertion in a single single step. We molecularly confirmed the targeted integration of the CAR transgene, which results in highly homogeneous and stable CAR expression in human peripheral blood T cells. These T cells exhibited the same in vitro tumor lysis activity and proliferation than retrovirally transduced CAR T cells, which augur favorably for their in vivo anti-tumor activity. Deep sequencing analyses to evaluate off-target effects of the nucleases and random AAV integration are in progress, as are in vivo experiments comparing the anti-tumor activity and graft-versus-host disease potential of edited T cells vs conventional CAR T cells. The process we describe here, which combines the scalability of universal T cell manufacturing with the uniformity and safety of targeted CAR gene integration, should be useful for the development of off-the-shelf CAR therapy.