Natural Killer (NK) function in patients with MDS as measured by non-MHC-restricted cytotoxicity and activation-dependent cell cytotoxicity (ADCC) are reduced in patients with MDS, however, the mechanisms of the functional impairment are not known. Tumor cytolysis occurs through orchestrated control by inhibitory NK receptors (NKRs) and activating NKRs, which control signaling events that lead to polarized movement of perforin-containing granules toward the NK-tumor contact area. We found that NK cells from 23 out of 35 patients with MDS (66%) displayed reduced lysis of K562 tumor cells compared to age-matched normal controls (p<0.01). To better characterize this defect, we evaluated patient NK function against differential tumor targets including the MDS1 cell line established from an MDS patients. We found that MDS1 incited non-MHC-restricted lysis. Unactivated PBMCs, unactivated NK cells, NK cell lines (NK92 and NKL) but not purified unactivated T cells from normal donors killed MDS1 in 4-hr 51Cr-release assays. Normal NK cells and NK cell lines were also found to rapidly redistrubute perforin granules after exposure to MDS1suggesting that a perforin-dependent lytic pathway was activated. We then performed simultaneous cytolytic assays with K562, MDS1, and the 721.221 B cell lymphoma cell line as target cells. We found that NK cells from MDS patients had greater lytic activity against MDS1 (average 24% vs. average 8% at 50:1 Effector:Target ratio, respectively, p<0.01) Antibody-blocking experiments demonstrated that the NKL cell line and PBMCs from 8 out of 10 MDS patients predominantly used the NKG2D activating receptor to kill MDS1. Consistent with this finding, we showed that MDS1 cells express the major human stress-inducible endogenous proteins MICA and MICB, which are NKG2D ligands. In contast, lysis by NK92 cells and normal PBMCs was not appreciably reduced by NKG2D blocking antibodies suggesting that other unidentified NKR(s) also mediate lysis. To identify the NKRs expressed in MDS patients, we performed immunophenotyping for both the activating NKRs and inhibitory NKRs compared to age-matched normal controls. We found that two activating receptors, NKp30 and CD244 (2B4), were significantly reduced on NK cells from all MDS patients regardless of their ability to lyse NK targets. Inhibitory NKR expression and function were normal. Interestingly, NKG2D expression correlated with reduced cytolytic function. Similar to studies on normal NK cells with low NKp30 and NKp46 (NCRdull) phenotypes, these results suggest that low NKp30 expression leads to predominant NKG2D utilization for tumor cell lysis, which is reduced in MDS patients with defective NK function. Our findings provide critical information about potential importance for immunosurviellance through NKG2D-NKG2D ligands.