Engineering CD19-Redirected T Lymphocytes to Enhance Their Safety and Efficacy

Modification of primary T cells to express chimeric antigen receptors (CARs) is an attractive strategy for the generation of tumor-specific T cells for adoptive therapies. However, their efficacy is limited by poor expansion within the tumor microenvironment. The addition of co-stimulatory endodomains, such as CD28, to the CAR may enhance cell expansion in response to the antigen, but cell growth and survival remain suboptimal since the co-stimulation provided through these endodomains cannot recapitulate the spatiotemporal sequence of signals T cells receive during physiologic activation. To further potentiate the expansion and survival of CAR-modified T lymphocytes, we generated a new vector encoding 3 molecules: CAR.19 incorporating the CD28 endodomain, codon optimized hIL15 to enhance cell survival and growth, and an inducible suicide gene based on the expression of Caspase9 (iCasp9) to increase the margin of safety associated with transgenic expression of an autocrine growth factor. These three sequences were linked using 2A-like peptide sequences. We compared the proliferative capacity, cytotoxic activity and in vivo anti-tumor effects of T lymphocytes expressing either CAR.19-28z alone or CAR.19-28z, IL15 and the suicide gene. T lymphocytes were activated with OKT3/CD28 antibodies and then transduced with retroviral supernatants. Phenotypic analysis showed 70±10% and 75±5% transduction efficiency for iCasp9/CAR19-28z/IL15 and CAR19-28z T cells, respectively. Only the iCasp9/CAR19-28z/IL15 T cells produced IL15 (>100pg/mL) after stimulation with CD19+ tumor cells. In co-culture experiments, T cells expressing CAR.19-28z alone or iCasp9/CAR19-28z/IL15 both completely eliminated CD19+ tumor cells within 72 hours. However, labeling with carboxyfluorescein diacetate succinimidyl ester (CFSE) showed that the proliferation of the iCasp9/CAR19-28z/IL15+ T cells in response to CD19+ tumor cells was greater than that of CAR.19-28z control+ T cells. Finally, the activation of the suicide gene iCasp9 with a small-molecule dimerizer (CID 50ng/mL) rapidly induced >90% apoptosis of T cells expressing iCasp9/CAR19-28z/IL15. Hence all three transgenes were functional. To assess the antitumor effects of the modified cells in vivo, we used a xenograft SCID mouse model and an in vivo bioluminescence system. CD19+ Daudi cells (1×10⁶) expressing firefly Luciferase (FL) were injected i.p., and on day 4, mice received i.p iCasp9/CAR19-28z/IL15+ or CAR19-28z+ or control T cells (10×106). By day 30 the tumor signal was significantly reduced in mice receiving iCasp9/CAR19-28z/IL15+ T cells (ROI<7×107) compared to mice receiving CAR19-28z or control T cells (ROI>2.0×109). In conclusion, our data indicate that a tricistronic vector can effectively be expressed in tumor-redirected human T cells, improving their survival and allowing their destruction should unwanted effects occur.