Natural killer (NK) cells play an essential role in early innate killing of virally-infected and tumor targets. NK cell-mediated activity is regulated by a repertoire of activating and inhibitory receptors that recognize ligands on diseased, stressed, or tumor targets. Killer cell immunoglobulin-like receptors (KIRs) are a family of polymorphic receptors that can be inhibitory or activating based on their intracellular signaling motifs. Expression of certain KIR haplotypes plays a key role in survival and relapse prevention for patients with acute myelogenous leukemia (AML) who receive allogeneic hematopoietic cell transplantation. Therefore, KIR haplotypes are an important consideration in selecting allogeneic donors for patients with AML. However, it is unclear if KIRs play a role in adoptive transfer of NK cells that are becoming more routinely utilized to treat refractory AML and other malignancies. To better address this question we used umbilical cord blood to isolate both CD34+ hematopoietic stem cells and CD45+CD56+ NK cells (UCB-56-NK) from the same umbilical cord blood unit. The CD34+ hematopoietic stem cells were then differentiated in vitro into CD56+ NK cells (UCB-34-NK). Despite originating from the same donor and sharing the same genetic background, as well as comparable expression of Fas ligand, TRAIL, NKp46, NKp44, NKG2A, and NKG2D, the UCB-34-NK cells have characteristically low KIR expression, whereas the UCB-56-NK cells have high KIR expression. This phenotype was further confirmed by mass cytometric (CyTOF) analysis of UCB-56-NK cells and UCB-34-NKcells with a panel of 36 phenotypic and functional NK cell markers. This unique system allows us to study the role of KIR expression independent of any other variations in donor or cell characteristics. The cytotoxicity and NK cell activation of UCB-34-NK cells and UCB-56-NK cells are compared to control NK cells isolated from peripheral blood (PB-NK cells) with standard in vitro cytotoxicity assays against neuroblastoma lines with varying HLA genotypes and a control K562 leukemic targets. Our data demonstrates that there is no statistical difference in NK cytotoxicity and activation of UCB-34-NK cells and UCB-56-NK cells across a spectrum of target cell HLA types, despite the differences in KIR expression. For example, at effector to target (E:T) ratios of 1:5 and 1:20 against neuroblastoma line IMR32, UCB-34-NK cells (KIR-low) demonstrated 68.5% and 84.1% maximal Caspase 3/7 activation, compared to 81.3% and 89.6% by UCB-56-NK cells (KIR-high). Additionally, we have used human induced pluripotent stem cells to derive NK cells (iPSC-NK cells) that vary in KIR expression levels. These CD45+CD56+ iPSC-NKs are differentiated from the same well-established iPSC line in the laboratory and therefore again share the same genetic background, and they have similar NK cell surface receptor expression of Fas ligand, TRAIL, NKp46, NKp44, and NKG2D, but differ in levels of KIR expression. Again in vitro cytotoxicity against hematopoietic tumor targets such as K562 and MOLM13 do not demonstrate a significant difference in killing, despite these KIR differences. For example, in targeting erythroleukemia K562 cells, iPSC-NK cells with high levels of KIR expression at E:T ratios of 1:2.5, 1:5, and 1: 10 have Caspase 3/7 activation of 21.1%, 28.2%, and 41.0%, compared to 20.1%, 22.0%, and 31.2% by iPSC-NK cells with low KIR expression. Together, these studies demonstrate that in vitro-derived NK cells do not require KIR expression to become licensed for anti-tumor activity and these cells are able to kill tumor targets whether or not they express KIRs. These studies better enable use of these allogeneic NK cell populations for off-the-shelf NK cell-based therapies without the need to optimize for KIR profiles for patients of differing HLA haplotypes.


Malmberg:Fate Therapeutics Inc.: Consultancy, Research Funding. Kaufman:Fate Therapeutics: Consultancy, Research Funding.

Author notes


Asterisk with author names denotes non-ASH members.