Characterization of natural killer cells expressing markers associated with maturity and cytotoxicity in children and young adults with sickle cell disease

Abstract

Background: Sickle cell disease (SCD) is increasingly recognized as a red blood cell disorder modulated by abnormally increased inflammation. We have previously shown that in patients with SCD not on a disease-modifying therapy (hydroxyurea or chronic transfusions), natural killer (NK) cell numbers are increased. In the current study, we further investigated the NK cell function to determine if there was evidence of increased activation and cytotoxicity.

Procedure: We conducted a cross-sectional study of 44 patients with HbSS/HbSβ⁰ thalassemia at steady state (hydroxyurea = 13, chronic transfusion = 11, no disease-modifying therapy = 20) and 23 healthy controls. Using a fresh blood sample, NK immunophenotyping was performed as follows: NK cells (CD3^- CD56⁺ lymphocytes) were evaluated for makers associated with activation (NKG2D, NKp30, NKp44, and CD69) and maturity (CD57, killer immunoglobulin-like receptors (KIR), and CD56dim). Degranulation and cytotoxicity assays were performed to evaluate NK cell function.

Results: Patients with SCD who were not on disease-modifying therapy had a higher number of NK cells with an immunophenotype associated with increased cytotoxicity (NKG2D⁺ , NKp30⁺ , CD56dim⁺ , and KIR⁺ NK cells) compared with healthy controls and patients on hydroxyurea. NK cells from SCD patients not on disease-modifying therapy demonstrated significantly increased cytotoxicity (measured by assaying NK cell killing of the K562 cell line) compared with healthy controls (P = 0.005). Notably, NK cell cytotoxicity against K562 cells in the hydroxyurea or chronic transfusion patients was not significantly different from that in healthy controls.

Conclusion: SCD is associated with increased NK cell function as well as increased NK cell numbers, which appears to be normalized with disease-modifying therapy.

Keywords: NK cells; hydroxyurea; sickle cell disease.

Figure 1.. Comparison of absolute NK cell numbers in peripheral blood of patients with sickle cell disease based on treatment category and healthy controls.

Absolute NK cell numbers were quantified (A) and further characterized by phenotype: markers associated with activation (B-E) and maturity (F-H). Mean and standard error of the mean (SEM) are displayed. Kruskal-Wallis test *** P ranges >0.0001 to <0.001, ** P ranges >0.001 to <0.01, * P ranges >0.01 to <0.05.

Figure 2.. Correlation between clinical laboratory parameters and absolute NK cell count in patients with sickle cell disease including those not on disease modifying therapy and those on hydroxyurea.

Linear regression lines and Pearson’s correlation coefficients, r, of hemoglobin F (HbF) percentage (A), hemoglobin (B) and absolute reticulocyte count (C) on day of NK cell assessment in patients on supportive care or hydroxyurea are shown.

Figure 3.. Assessment of NK cell degranulation and interferon-gamma (IFN-gamma) production in patients with sickle cell disease.

Shown is a representative plot of a patient with sickle cell disease not on disease modifying therapy (A). NK cells were co-incubated with K562 cells and CD107a surface expression as a surrogate for degranulation (B) and intracellular interferon-gamma (C) measured in response to the stimulus. Mean and standard error of the mean (SEM) are displayed in (B) and (C). Kruskal-Wallis test *** P ranges >0.0001 to <0.001, ** P ranges >0.001 to <0.01, * P ranges >0.01 to <0.05.

Figure 4.. Comparison of NK cell cytotoxicity in patients with sickle cell disease not on disease modifying therapy with healthy controls.

NK cell cytotoxicity was determined at different effector to target ratios (10:1, 5:1, 2.5:1, 1.25:1) for each group and standardized to the cytotoxicity of the reference sample NK cell line NK92 MI. A statistically significant difference between patients in the no disease modifying therapy group (n=8) and healthy controls (n=5) is denoted (**). No difference was seen between those on therapy (hydroxyurea or transfusion, n=3) and the control group. Mean and standard error of the mean (SEM) are displayed. Ordinary two-way ANOVA with Turkey’s multiple comparison test. *** P ranges >0.0001 to <0.001, ** P ranges >0.001 to <0.01, * P ranges >0.01 to <0.05.