Abstract 3505: Preclinical testing of mRNA engineered chimeric antigen receptor T cells: Comparison of a multi-infusion protocol to a single injection of stably expressed CARs in ALL xenograft models

Abstract

Abstract T cells engineered to express chimeric antigen receptors (CARs), consisting of an anti-tumor targeting domain fused to Signal 1 and 2 activation domains, are a promising approach to cell therapy. We compared T cells engineered to stably express CARs using integrating viral vectors with T cells transiently expressing CARS via direct transduction with non-integrating mRNA. In early clinical testing, this CAR design using lentiviral transduction has been tested in 3 patients with advanced CLL. In these refractory patients with very high disease burdens, dramatic clinical responses have been observed, with two patients achieving CRs and one a PR, all sustained at >12 months. Transient expression has several advantages: short-term expression may allow initial clinical testing while limiting toxicity, and expediting manufacturing and scaling to GMP of CAR modifications, allowing rapid iterative improvements of CAR design. To evaluate the in vivo efficacy of mRNA CAR+T cells, we used xenograft models of human ALL (Nalm-6 and a primary patient leukemia, both CD19+) and high-level human T cell engraftment. NSG mice were injected with ALL at a dose producing 100% BM engraftment by 5d with 100% mortality by 25d (Nalm-6) or 36d (Primary). T cells were expanded ex vivo using CD3/CD28 beads and underwent lentiviral gene transfer or mRNA electroporation. mRNA transfection produced high but transient CAR surface expression, peaking at 4 hours and then declining over 8d. Using a novel method of simultaneous bioluminescent tracking of CARs and leukemia, we observed rapid migration of CAR T cells to all sites of disease and 100-fold reduction in ALL burden after only 24h. Using low dose cyclophosphamide (CTX; 60 mg/kg IP) 24h before the next infusion, we were able to deplete non-CAR expressing T cells and observe continued effects on leukemia from subsequent mRNA CAR infusions resulting in significantly prolonged survival in mice treated with mRNA CARs. Short-term efficacy of a single dose of mRNA vs. lentiviral CAR+ T cells was similar. Neither CTX nor untargeted repeat CAR T cells infusions alone had an effect on leukemia burden or survival. While lentiviral CAR+ T cells persisted and were active longer, median overall survival was not different between repeat mRNA CARs (52d Nalm-6, 76d primary) and a single dose of lentiviral transduced CARs (54d Nalm-6, 87d primary). Only stably expressed CARs resulted in any long term cures in this model. Relapse patterns suggest that the efficacy of mRNA and lentivirus CAR T cells are similar, as peripheral disease can be repeatedly cleared with subsequent mRNA CAR infusions but immune privileged xenograft artifact sites such as the tooth root are more resistant. These pre-clinical results demonstrate an important advance in T cell engineering, and are leading directly to clinical trials testing mRNA CAR CTLs. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 3505. doi:1538-7445.AM2012-3505