Blocking Inhibitory KIR Is Insufficient for Optimal Killing of AML and ALL Targets: Additional Requirements for NKG2A and LIR-1 Blockade

NK cell alloreactivity is determined by the interaction between inhibitory NK cell receptors and their MHC class I cognate ligands. The best characterized NK cell receptor that recognizes MHC I is KIR. Mismatching donors and recipients for KIR ligands has led to transplant strategies that can result in less relapse and better survival in myeloid leukemia patients. However, other MHC recognizing receptors may also be involved in leukemia eradication. We developed a model using primary AML and ALL targets to explore the hierarchy of different class I MHC recognizing inhibitory NK cell receptors. In addition to KIR, NKG2A is expressed on 54±14% of PB NK cells and is the dominant NK cell receptor expressed in the first six months after transplant. LIR-1 is expressed on 35±17% of PB NK cells and is more likely to be co-expressed on KIR+ NK cells than KIR- NK cells. The goal of our study was to determine optimal conditions to enhance NK alloreactivity against primary AML and ALL targets. A pan-HLA blocking antibody (HP-1F7, binding to HLA, B, C, E and G) was used as a positive control to interrupt all signals from MHC class I. Receptor blockade with anti-KIR (1-7F9, blocking KIR 2DL1/L2/L3, Novo Nordisk), anti-NKG2A and anti-LIR-1 was tested either alone or in combination to understand the hierarchy of inhibitory receptors that determine NK cell reactivity of leukemia targets. Resting NK cells were tested fresh or activated with low concentrations of IL-2 (5 IU for 42 hours) to mimic conditions achievable by pharmacologic IL-2 administration. Six of ten primary AML targets were sensitive to allogeneic NK cell killing only when HLA-A, B, C, E and G were blocked (7.8±5.7 vs. 24±9.6, n=6; p<0.005). Six of seven primary ALL targets also responded to resting allogeneic NK cells but only after class I blockade (8.2±5.7 vs. 25±9.7, n=6; p<0.005). Low concentrations of IL-2 enhanced mean killing with class I blockade by approximately 10% for AML and ALL targets. CD107a degranulation yielded similar results. We next evaluated the class I recognizing inhibitory receptors on NK cells responsible for the enhanced killing of primary leukemia targets. Experiments were performed with KIR, NKG2A and LIR-1 blockade. In both cytotoxicity and CD107a degranulation assays, single inhibitory receptor blockade resulted in enhanced killing slightly. However, addition of LIR-1 blockade to either KIR blockade or NKG2A blockade consistently increased killing of all targets sensitive to pan-HLA blockade. Purification of KIR- NK cells, expressing NKG2A or LIR-1, showed they were potently alloreactive against primary leukemia targets only upon dual blockade with NKG2A and LIR-1 mAbs. Blocking of more than one receptor was always more effective than single receptor blockade. All leukemia targets expressed HLA-E (the ligand for NKG2A) but none expressed HLA-G (the high affinity ligand for LIR-1) by Western Blot analysis, suggesting that other HLA class I (low affinity ligands for LIR-1) are involved. Our data show that:

1. Both KIR⁺ and KIR- NK cells exhibit significant potential for cytotoxicity of killing primary AML and ALL blasts but only when more than one class I receptor was blocked,

2. KIR- NK cells can be licensed through receptors other than KIR, and

3. More than half of primary AML and ALL targets tested are susceptible to allogeneic NK cell killing following multiple inhibitory receptor blockade.

Therefore, it is not surprising that the current strategies of adoptive NK cell therapy may be limited. Clinical trials with mAb blocking of multiple MHC class I inhibitor receptors (KIR, NKG2A, or LIR-1) to enhance NK cell killing of AML and ALL is warranted.