Retargeting of Citokine-Induced Killer (CIK) Cells with Molecularly Engrafted T-Cell Receptors (TCR): A Preclinical in Vitro and In Vivo Study

Abstract 1917

Cytokine Induced Killer (CIK) cells are a heterogeneous population of T lymphocytes sharing NK phenotype and functional properties: they are CD3+/CD56+ and have a potent MHC-unrestricted antitumor activity. We hypothesized that the therapeutic potential of CIK cells might be increased if they acquired the ability to recognize MHC-restricted tumor associated antigens. To this end, we transduced CIK cells with an HLA-A2 restricted T-Cell Receptor (TCR) directed against the melanocyte associated antigen Mart-1. PBMC were incubated with IFN-γ on day 0 and supplemented with anti-CD3 and IL-2 on day +1 to generate CIK cells. Cultures were transduced at day 4 with concentrated lentiviral particles and successfully expanded over a 4 week period. This allowed to generate CIK cells that contained 11±9% Mart-1 TCR positive cells, as detected by staining with a Mart-1 specific tetramer. Transduced CIK cultures contained 61±19% CD3+/CD56+ cells. Tetramer positive cells were both CD3+/CD56- and CD3+/CD56+ (31±8% and 59±9%, respectively), indicating that both MHC-restricted T-cells and MHC-unrestricted CIK cells could be targeted by lentiviral transduction. TCR-transduced CIK cells specifically recognized tumor cells presenting the relevant peptide and maintained their MHC-unrestricted tumor activity at the same time. The cytotoxic activity of Mart-1 redirected CIK against HLA-A2+ melanoma cell lines was 2.8 fold higher than the untransduced counterparts (62%±9 vs 22±6% lysis at an effector/target ratio of 20:1), while the cytotoxic activity against a Mart-1+/HLA-A2- melanoma cell line was similar in transduced and untransduced CIK cells (24%±8 vs 22±6% lysis), indicating that the increased activity was due to HLA-restricted recognition. This was confirmed by blocking experiments with an HLA-Class I antibody. At the end of the culture, the majority of both unmodified and transduced CIK cells expressed an effector memory phenotype, with few residual central memory cells. In TCR redirected cells there was a slight increase of cells with a naive phenotype compared to unmodified cells (19±5% vs 9±4%). These data suggest that the naive and central memory pool of redirected CIK cells might efficiently expand in vivo and support a long lived memory response, whereas the terminally differentiated pool might mediate short lived but potent MHC-restricted and unrestricted activity.

To demonstrate that TCR transduced CIK cells display an increased antitumor activity also in vivo, we have conducted preparative experiments in humanized immunocompromised mice (NOD/scid/γ(c)(−/−), NSG). Results obtained so far have shown that: i) When 5×10⁶ or 10×10⁶ CIK cells (both TCR-transduced and unmodified) were injected intravenously, they stably engrafted NSG mice, homing predominantly to spleen and liver and also, to a lesser extent, to bone marrow and kidney (36±9%, 39±12%, 4±3%, 1.6±3% of human CD3+/CD45+ cells at 3 months in the spleen, liver, bone marrow and kidney, respectively); circulating cells were also detected in the peripheral blood. Engrafted CIK cells maintained high expression of CD8 but progressively downregulated and finally lost CD56 expression. When 10×10⁶ CFSE-marked CIK cells where injected intravenously, they similarly engrafted and proliferated in NSG mice, reaching a peak of proliferation 2–3 weeks after injection; at 4 weeks, the CFSE dye was already completely diluted out. ii) Differently from normal PBMC, CIK cells did not induce any appreciable clinical sign of acute or chronic Graft versus Host Disease (GVHD), as determined by the general appearance of the fur, mobility and weight loss. Mice were observed up to 5 months, and both unmodified and TCR transduced CIK cells displayed the same behavior. iii) NSG accepted the graft of as few as 50.000 matrigel-resuspended melanoma cells that were injected subcutaneously, with the appearance of a measurable mass (>3mm) ten days after inoculation. From these in vivo experiments we conclude that: i) human CIK cells engraft and proliferate in NSG mice; ii) CIK cells do not cause GVHD; iii) the human melanoma cell lines used in vitro can grow in NSG mice. We are now testing whether TCR transduced CIK cells have superior antitumor activity than unmodified CIK in the NSG mice model. Taken together, our data suggest that TCR transfer into CIK cells is feasible and greatly improves their antitumor potential in vitro and possibly in vivo.