Reduced natural killer (NK) function associated with high-risk myelodysplastic syndrome (MDS) and reduced expression of activating NK receptors

Abstract

Myelodysplastic syndromes (MDS) are characterized by ineffective hematopoiesis with potential for progression to acute myeloid leukemia (AML). We compared natural killer (NK) cytolytic function in 48 MDS patients with 37 healthy donors and found reduced activity in the patient population (K562 cytolysis, 19% +/- 21% SD versus 40% +/- 17%) (P < .001). NK cytotoxicity in MDS patients was reduced against 3 disparate tumor targets with differential activating receptor requirement, suggesting global defects in NK function. Reduced NK function in MDS was significantly associated with higher International Prognostic Score (P = .01), abnormal karyotype (P = .05), the presence of excess blasts (P = .01), and age-adjusted bone marrow hypercellularity (P = .04). MDS patients had a display of the activating receptor NKp30, and NKG2D down-regulation closely correlated with impaired NK function (P = .001). NKG2D ligands (MICA and MICB) were expressed on CD34(+) cells from bone marrow of 30% of MDS patients and a leukemic cell line derived from an MDS patient (MDS1). Collectively, these findings suggest that impairment of NK cytolytic function derives in part from reduced activating NK receptors such as NKG2D in association with disease progression. Evasion of NK immunosurveillance may have importance for MDS disease progression.

Figures

Figure 1

K562 lysis is reduced in patients with MDS by direct cytotoxicity of target cells. Results are shown for 5-hour 51Cr-release assays using K562 as a target. (A) The graphic representation of the percent specific lysis at a 50:1 E/T ratio is shown using normal PBMCs (n = 37) and PBMCs from patients with MDS (n = 48). (B) Percent specific lysis of K562 at a 50:1 E/T ratio versus percentage of NK cells in the sample for 26 patients and the percent specific lysis of K562 at a 50:1 E/T ratio versus the absolute number of NK cells in the peripheral blood. (C) Percent specific lysis by highly enriched NK cells (NK) and PBMCs at 50:1, 25:1, 12:1, and 6:1 E/T ratios are shown from 3 healthy controls and 2 MDS patients. Samples used for these assays were never frozen. The graphs represent the average of triplicate samples; standard deviation is indicated by the error bars, and the asterisks represent statistical significance at P ≤ .001 as determined by the nonparametric Wilcoxon rank sum test.

Figure 2

NK receptor phenotype analysis in MDS patients and healthy controls. (A) The percentage of NK cells that coexpressed NKp46, NKp30, and NKG2D was determined relative to isotype control antibody staining. Receptor expression was compared between group A versus normal and group B versus healthy controls (NL) using the Wilcoxon rank sum test. P values are shown only for those comparisons that were statistically significant. Horizontal lines represent the mean of that group. (B) Twenty-six MDS patients were divided into 2 equal groups based on the median expression of NKG2D and median expression of NKp30. The percent specific lysis of K562 at a 50:1 E/T ratio was then compared between groups with lower NKG2D plus lower NKp30 (n = 6), lower NKG2D plus higher NKp30 (n = 6), lower NKp30 plus higher NKG2D, and higher NKp30 plus higher NKG2D. Error bars indicate SD. These groups were compared using the Wilcoxon rank sum test.

Figure 3

Generalized impairment in NK signaling in patients with MDS. (A) Using K562, 721.221, and MDS1 cells as the target, antibody blocking experiments were performed using PBMCs from healthy donors at a 50:1 effector-target (E/T) ratio in the presence of medium alone or 5 μg/mL of the following antibodies: isotype control antibody, anti-NKG2D, anti-NKp30, anti-NKp46, anti-NKG2D plus anti-NKp30, anti-NKG2D plus anti-NKp46, and anti-NKp30 plus anti-NKp46. (B) Direct cytotoxicity of K562, MDS1, and 721.221 tumor cells in 5-hour 51Cr-release assays using PBMCs at a 50:1 E/T ratio from MDS group A patients with low K562 lysis (less than 13%) (n = 8), MDS group B patients with normal lysis of K562 (13% or more) (n = 11), and healthy control donors (n = 12). Horizontal lines represent the mean of that group. (C) Antibody blocking experiments were performed using PBMCs from MDS patients at a 50:1 E/T ratio. Effector cells were incubated in the absence of blocking antibodies (medium, black bars) or in the presence of 5 μg/mL isotype control antibody (hatched bars) and anti-NKG2D antibody (gray bars). NKL and NK92 cells were used at a 10:1 E/T ratio. The graphic representation of the percent specific lysis at a 50:1 effector-target (E/T) ratio is shown. The graphs represent the average of triplicate samples; SD is indicated by the error bars, and asterisks indicate statistical significance at P ≤ .01 as determined by a Wilcoxon rank sum test.

Figure 4

MICA and MICB expression on K562, MDS1, and 721.221 cell lines and on CD34+ bone marrow cells from MDS patients. Flow cytometry histograms for MICA/MICB expression on K562 cells, 721.221 tumor cells, and MDS1 cells. Isotype control expression (dotted line) was compared with samples stained with an equal amount of anti-MICA or anti-MICB antibody (solid line). Histograms represent results from individual patients (UPN). SFI was calculated using the equation described in “Patients, materials, and methods,” and specific value is shown in the histograms. The determination of positive (+) or negative (-) was based on an SFI value of at least 1.5 (indicated in the upper right hand corner of the histograms).

Figure 5

IL-2 restores impaired NK function in MDS patients. PMBCs from patients with MDS and PBMCs from healthy donors were cultured for 3 days in the absence (−IL-2) or the presence of 100 IU/mL IL-2 (+IL-2). (A) The graphic representation of the percent specific lysis at a 50:1 effector-target (E/T) ratio is shown for 5-hour 51Cr-release assays using K562, MDS1, and 721.221 cells as targets. From a subset of these samples, the (B) percentage and the (C) median fluorescence intensity (MFI) of NKG2D, NKp30, NKp46, NKp44, and CD69 was determined in CD56+/CD3− NK cells by flow cytometry. The number of patients and healthy controls included in each group is shown at the bottom of each graph. The graphs represent the average of duplicate samples; SD is indicated by the error bars, and asterisks indicate statistical significance as determined by a paired t test.