Ex Vivo Expansion of Cord Blood Natural Killer Cells Overcomes Impaired Immune Synapse Formation and Effector Function in Acute Myeloid Leukemia

Natural killer (NK) cells play an important role in host immunity by eradicating pathogen-infected and tumor cells. Peripheral blood-derived NK cell therapy has shown promise in clinical trials for acute myeloid leukemia. Cord blood (CB) is another potentially rich source of NK cells. However NK cells isolated directly from CB have poor cytolytic activity. We investigated the mechanism for the low cytolytic activity of CB NK cells and whether the defect could be overcome by ex vivo expansion. NK cell killing of the target cells is achieved by formation of a mature immune synapse, followed by secretion of lytic granules containing perforin and granzymes. F-actin polymerization at the NK-tumor cell conjugates is a hallmark of immune synapse formation which can be detected and quantitated by confocal microscopy. We hypothesized that CB NK cells exhibit low cytotoxicity against leukemia target cells due to a defect in the formation of the immune synapse. We confirmed that unmanipulated CB NK cells exhibit low cytotoxicity against AML blasts (5% at an E:T ratio 20:1) in comparison to peripheral blood (PB) NK cells (35% at an E:T ratio 20:1). Evaluation of the natural cytotoxicity receptors (NCRs) showed normal expression of the NKp30 and the NK46 receptors. We then investigated whether this poor cytolytic capability is accompanied by poor immune synapse formation with tumor cells. Surprisingly, both CB and PB NK cells expressed comparable levels of perforin and demonstrated similar perforin polarization to the immune synapse. F-actin polarization was observed in only 12% (range 9– 21%) of the CB NK cell/K562 conjugates versus 85% (range 68–87%) of PB NK cell/K562 conjugates (P<0.001). Remarkably, this impairment could be reversed by ex vivo expansion of CB NK cells with rIL-2. Expanded CB NK cells formed an increased percentage of immune synapses with K562 tumor cells 65% (range 60–71%) and primary human AML blasts 48% (range 39 –55%) which was comparable to the levels generated with PB NK cells. The present data reinforce the conclusion that formation of the activating NK cell immune synapse is required for cytotoxic activity. Furthermore, natural cytotoxicity receptors (NCR) and Kir receptors including KIR2DL1/S1 and KIR2DL2 were preserved on the expanded CB NK population. Moreover, the expanded CB NK cells were able to lyse AML targets in vitro (29% at 20:1 E:T ratio). Finally, we demonstrated that ex vivo expanded CB NK cells efficiently kill human AML in an NOD/SCID ^null mouse model. A 50 % reduction in AML tumor burden was documented in comparison to control groups by 6 weeks post NK infusion (P<0.05). Our results suggest that ex vivo expansion of CB NK cells is a feasible and potentially effective strategy for the treatment of AML.