From TCR Gene Transfer to TCR Gene Editing of Central Memory T Lymphocytes for Immunotherapy of Leukemia.

Abstract 374

The transfer of a T cell receptor (TCR) from a high-avidity tumor-specific lymphocyte to polyclonal lymphocytes as been proposed as an attractive approach to overcome the major limitations of adoptive immunotherapy of cancer, such as the difficulty in expanding rare, high-avidity tumor-specific CTLs from cancer patients, in conditions that can preserve their function and prevent exhaustion. However, complete exploitation of this approach is limited by technical and safety issues, like the low and transient transgene expression, unpredictable pairing of the exogenous and endogenous TCR chains and poor survival and expansion potential of gene-modified effector lymphocytes. We developed a strategy aimed at overcoming these limitations. First, to increase TCR expression and facilitate proper TCR pairing we used a codon-optimized high-avidity TCR, specific for an HLA-A2-restricted peptide from the oncogenic Wilms tumor antigen 1 (WT1), modified with point mutations to introduce cysteines into the constant regions of the α and β chains. We cloned genes encoding the TCR in third generation lentiviral vectors under the control of a bi-directional PGK or a bi-directional EF1α promoter. Second, to increase the expected life span and expansion potential of gene modified cells, we transduced lymphocytes upon activation with anti-CD3 and anti-CD28 antibody-conjugated beads (bCD3/CD28) and culture with low doses of IL-7/IL-15. Human T lymphocytes were efficiently transduced by both vectors. However, PGK promoter was superior to EF1α in sustaining stochiometric expression of WT1-specific TCR chains, at levels appropriate for efficient HLA-A2/WT1 pentamer binding (16%), for up to 70 days, in the absence of further T cell stimulation. The phenotype of transduced cells was consistent with early T cell differentiation (CD62L+, CD28+CD27+, IL7Rα+, IL-2+ γIFN±), a phenotype associated with long-term persistence of adoptively transferred lymphocytes. Most importantly, TCR transduced cells were able to specifically produce γIFN and exhibited cytotoxic activity against WT1+HLA-A2+ primary leukemic blasts from AML patients. Third, to further improve the safety of the strategy and ensure predictable levels of transgene expression, we developed, for the first time, a protocol for site-specific integration of transgenes into primary T lymphocytes by exploiting zinc finger nucleases (ZFNs). Delivery in primary T lymphocytes of ZFNs targeting the CCR5 gene, in association with a GFP expression cassette flanked by CCR5 homology arms, resulted in efficient (6%) site-specific integration of GFP into the CCR5 locus. Extensive molecular analysis on sorted gene-modified cells and clones confirmed site-specific integration and showed a biallelic CCR5 disruption in 87% of GFP+ clones, due to a combined site-specific integration in one CCR5 allele, and non-homologous end joining (NHEJ) repair in the other allele. Finally, we exploited ZFNs to address the major limitation of TCR gene transfer, represented by the co-expression in the same cell of the endogenous and tumor-specific TCR. Such co-expression results in a dilution of tumor-specific TCR, thus reducing the efficacy of gene-modified cells, and might result in autoreactive specificities due to misparing between endogenous and exogenous α and β TCR chains. To this purpose, we used a new set of ZFNs specific for the constant region of the TCR β chain and knock down the expression of endogenous TCR in 20% of Jurkat cell line and 3.5% of primary human lymphocytes. These results show that a complete editing of T cell specificities can be achieved in human lymphocytes. (E.P. and P.G. equally contributed to this work).