Kaufman, Dan S.

Abstract NK cell-mediated killing is inversely proportional to the amount of MHC class I expression on certain tumor targets. Two hypotheses have been proposed to explain why class I-bearing targets are more resistant to NK cell-mediated lysis: 1) the presence of class I prevents NK cell recognition of a triggering molecule on the target cell surface, or 2) class I recognition transmits a separate inhibitory signal to the NK cell. To differentiate between these potential mechanisms, we have used cloned human CD16+/CD3- NK cells, the class I-deficient cell line C1R, and C1R cells expressing high levels of transfected HLA class I gene products. If class I expression blocks NK cell recognition of the targets, then proximal cell signaling events such as phospholipase C-mediated hydrolysis of membrane phosphoinositides should be decreased in the NK cells interacting with the class I transfectants. However, we found that increases in the level of target cell class I expressions did not decrease phosphoinositide turnover or calcium signaling in NK cells. We also examined the effect of treating HLA-transfected C1R cells with mAb specific for the transfected MHC class I gene product. If class I expression has a negative regulatory influence on NK cell activation, then treating the targets with anti-HLA mAb should block the transmission of this negative signal. Consistent with this notion, addition of anti-HLA mAb (either whole Ig or F(ab’)2 fragments) led to increased lysis of the class I-transfected targets. In contrast, addition of isotype-matched mAb specific for other cell surface markers did not alter sensitivity to lysis. All of these results suggest that MHC class I expression on target cells can initiate inhibitory signals in NK cells without blocking access to target structures.

Abstract Abstract 4034 Natural Killer (NK) cells are lymphocytes of the innate immune system with anti-viral and anti-cancer activity. Over the past decade, they have gained interest as a promising cellular source for use in adoptive immunotherapy for the treatment of cancer. Most notably, NK cells play an important role in the graft-vs-tumor effect seen in allogeneic hematopoietic stem cell transplantation (allo-HSCT), and a better understanding of NK cell biology has translated into improved transplant outcomes in acute myelogenous leukemia (AML). Small studies have demonstrated a role for NK cell activity in multiple myeloma (MM) patients receiving allo-HSCT. Investigators have also utilized haplo-identical killer immunoglobulin-like receptor (KIR) mismatched NK cells for adoptive immunotherapy in patients with multiple myeloma (MM). Our group has focused on the development of NK cells from human embryonic stem cells (hESCs) and induced pluripotent stem cells (iPSCs) as a novel starting source of lymphocytes for immunotherapy. We have previously demonstrated potent anti-tumor activity of hESC-derived NK cells in vitro and in vivo against a variety of different targets. We have also shown that iPSC-derived NK cells from a variety of different somatic cell starting sources posses potent anti-tumor and anti-viral activity. Here, we demonstrate hESC- and iPSC-derived NK cell development in a completely defined, feeder-free system that is amenable to clinical scale-up. These cultures contain a pure population of mature NK cells devoid of any T or B cell contamination, which are common adverse bystanders of cellular products isolated and enriched from peripheral blood. Our cultures are homogenous for their expression of CD56 and express high levels of effector molecules known to be important in anti-MM activity, including KIR, CD16, NKG2D, NKp46, NKp44, FasL and TRAIL. We have now tested the activity of hESC- and iPSC-derived NK cells against MM tumor cells in order to provide a universal source of lymphocytes for adoptive immunotherapy in patients with treatment refractory disease. We find that similar to peripheral blood NK cells (PB-NK), hESC- and iPSC-derived NK cells are cytotoxic against 3 distinct MM cell lines in a standard chromium release cytotoxicity assay. Specifically, activated PB-NK cells killed 48.5% of targets at 10 to 1 effector to target ratios, whereas hESC (46.3%) and iPSC (42.4%) derived NK cells also demonstrated significant anti-MM activity. Also, hESC- and iPSC-derived NK cells secrete cytokines (IFNγ and TNFα) and degranulate as demonstrated by CD107a surface expression in response to MM target cell stimulation. When tested against freshly isolated samples from MM patients, hESC- and IPSC-derived NK cells respond at a similar level as activated PB-NK cells, the current source of NK cells used in adoptive immunotherapy trials. These MM targets (both cell lines and primary tumor cells) are known to express defined ligands (MICA/B, DR4/5, ULBP-1, BAT3) for receptors expressed on NK cells as well as a number of undefined ligands for natural cytotoxicity receptors (NCRs) and KIR. As these receptor-ligand interactions drive the anti-MM activity of NK cells, we are currently evaluating expression of each of these molecules on the surface of both the effector and target cell populations. Not only do hESC- and iPSC-derived NK cells provide a unique, homogenous cell population to study these interactions, they also provide a genetically tractable source of lymphocytes for improvement of the graft-vs-myeloma effect and could be tailored on a patient specific basis using banks of hESC-or iPSC-derived NK cells with defined KIR genotypes for use as allogeneic or autologous effector cells. Disclosures: No relevant conflicts of interest to declare.

Many studies have characterized the transmembrane signaling events initiated after T-cell antigen receptor recognition of major histocompatibility complex (MHC)-bound peptides. Yet, little is known about signal transduction from a set of MHC class I recognizing receptors on natural killer (NK) cells whose ligation dramatically inhibits NK cell-mediated killing. In this study we evaluated the influence of MHC recognition on the proximal signaling events in NK cells binding tumor targets. We utilized two experimental models where NK cell-mediated cytotoxicity was fully inhibited by the recognition of specific MHC class I molecules. NK cell binding to either class I-deficient or class I-transfected target cells initiated rapid protein tyrosine kinase activation. In contrast, whereas NK cell binding to class I-deficient targets led to inositol phosphate release and increased intracellular free calcium ([Ca2+]i), NK recognition of class I-bearing targets did not induce the activation of these phospholipase C-dependent signaling events. The recognition of class I by NK cells clearly had a negative regulatory effect since blocking this interaction using anti-class I F(ab')2 fragments increased inositol 1,4,5-trisphosphate release and [Ca2+]i and increased the lysis of the targets. These results suggest that one of the mechanisms by which NK cell recognition of specific MHC class I molecules can block the development of cell-mediated cytotoxicity is by inhibiting specific critical signaling events.

Target structures important for natural killer (NK) cell recognition of virally infected cells are not well defined. Since major histocompatibility complex (MHC) class I molecules bind viral peptides during acute infection, we evaluated whether an interaction between MHC and virus might influence the susceptibility of infected cells to NK cell-mediated lysis. To control for MHC class I expression on target cells, we used either HLA class I-deficient C1R cells or C1R sublines expressing transfected HLA class I gene products. Human NK cells were unable to preferentially lyse class I-deficient C1R cells after infection with herpes simplex virus (HSV). In contrast, HLA class I transfectants were significantly more susceptible to NK cell-mediated cytotoxicity after HSV infection. This occurred for HSV-infected C1R cells expressing any of the three HLA class I gene products tested (i.e., HLA-B27, HLA-A3, or HLA-Aw68), indicating that NK cell recognition in this system does not require "self" MHC and is not unique for a single haplotype. Productive HSV infection is required for the increased killing, since inoculation with UV-inactivated virus did not lead to increased lysis. In addition, since HSV infection of the transfectants did not significantly alter the level of class I expression, the change in susceptibility appears to be due to qualitative changes in the target structures on HSV-infected, HLA class I+ targets. These results demonstrate a role for MHC class I in regulating NK cell-mediated killing of virus-infected cells.