Primary adrenal insufficiency is associated with impaired natural killer cell function: a potential link to increased mortality

Abstract

Objective: Mortality in patients with primary adrenal insufficiency (PAI) is significantly increased, with respiratory infections as a major cause of death. Moreover, patients with PAI report an increased rate of non-fatal infections. Neutrophils and natural killer (NK) cells are innate immune cells that provide frontline protection against invading pathogens. Thus, we compared the function and phenotype of NK cells and neutrophils isolated from PAI patients and healthy controls to ascertain whether altered innate immune responses could be a contributory factor for the increased susceptibility of PAI patients to infection.

Design and methods: We undertook a cross-sectional study of 42 patients with PAI due to autoimmune adrenalitis (n = 37) or bilateral adrenalectomy (n = 5) and 58 sex- and age-matched controls. A comprehensive screen of innate immune function, consisting of measurements of neutrophil phagocytosis, reactive oxygen species production, NK cell cytotoxicity (NKCC) and NK cell surface receptor expression, was performed on all subjects.

Results: Neutrophil function did not differ between PAI and controls. However, NKCC was significantly reduced in PAI (12.0 ± 1.5% vs 21.1 ± 2.6%, P < 0.0001). Phenotypically, the percentage of NK cells expressing the activating receptors NKG2D and NKp46 was significantly lower in PAI, as was the surface density of NKG2D (all P < 0.0001). Intracellular granzyme B expression was significantly increased in NK cells from PAI patients (P < 0.01).

Conclusions: Adrenal insufficiency is associated with significantly decreased NKCC, thereby potentially compromising early recognition and elimination of virally infected cells. This potential impairment in anti-viral immune defense may contribute to the increased rate of respiratory infections and ultimately mortality in PAI.

Figure 1

Natural killer cell cytotoxicity and conjugate formation. Panel A, cytotoxicity of resting NK cells isolated from PAI patients (n = 41) and age- and sex-matched healthy controls (n = 29) toward the erythroleukemic K562 cell line at an effector:target cell ratio of 10:1. Horizontal line depicts the median value. Panel B, percentage of NK cells bound to K562 tumor cells as a measure of conjugate formation between effector (NK) and target (K562) cells. Conjugate formation was assessed following a 4-h co-culture at an effector:target cell ratio of 10:1. Data were obtained from 41 PAI patients and 28 age- and sex-matched healthy controls. Horizontal line depicts the median value.

Figure 2

Intracellular perforin and granzyme B expression in resting NK cells. Panels A and C, NK cells isolated from PAI patients (n = 42) and age- and sex-matched healthy controls (n = 34) were analyzed for intracellular perforin expression. Data are presented as percentage positive NK cells (Panel A) and staining intensity (Panel C). Horizontal line depicts the median value. Panels B and D, NK cells isolated from PAI patients (n = 39) and age- and sex-matched healthy controls (n = 30) were analyzed for intracellular granzyme B expression. Data are presented as percentage positive NK cells (Panel B) and staining intensity (Panel D). Horizontal line depicts the median value. MFI, mean fluorescence intensity. Panels E and F, representative flow cytometry plots depicting intracellular perforin (Panel E) and granzyme B (Panel F) expression in NK cells isolated from a single PAI patient (blue line) and healthy control (red line). The black line represents the isotype control.

Figure 3

Natural killer cell cytotoxicity (NKCC), DHEA replacement therapy and cause of primary adrenal insufficiency (PAI). Panel A shows NKCC (median and individual data points) as assessed in healthy controls (n = 29) and patients with PAI (n = 41), separated into patients with (n = 11) and without (n = 30) chronic DHEA replacement therapy. Panel B shows NKCC (median and individual data points) in healthy controls (n = 29), patients with PAI due to autoimmune adrenalitis (n = 36) and patients with PAI after bilateral adrenalectomy (n = 4). Statistical comparisons were made with a non-parametric Kruskal–Wallis test with Dunn’s multiple comparison test. *P < 0.05; **P < 0.01; ***P < 0.001.