Natural killer cell dysfunction is a distinguishing feature of systemic onset juvenile rheumatoid arthritis and macrophage activation syndrome

Abstract

Macrophage activation syndrome (MAS) has been reported in association with many rheumatic diseases, most commonly in systemic juvenile rheumatoid arthritis (sJRA). Clinically, MAS is similar to hemophagocytic lymphohistiocytosis (HLH), a genetic disorder with absent or depressed natural killer (NK) function. We have previously reported that, as in HLH, patients with MAS have profoundly decreased NK activity, suggesting that this abnormality might be relevant to the pathogenesis of the syndrome. Here we examined the extent of NK dysfunction across the spectrum of diseases that comprise juvenile rheumatoid arthritis (JRA). Peripheral blood mononuclear cells (PBMC) were collected from patients with pauciarticular (n = 4), polyarticular (n = 16), and systemic (n = 20) forms of JRA. NK cytolytic activity was measured after co-incubation of PBMC with the NK-sensitive K562 cell line. NK cells (CD56+/T cell receptor [TCR]-alphabeta-), NK T cells (CD56+/TCR-alphabeta+), and CD8+ T cells were also assessed for perforin and granzyme B expression by flow cytometry. Overall, NK cytolytic activity was significantly lower in patients with sJRA than in other JRA patients and controls. In a subgroup of patients with predominantly sJRA, NK cell activity was profoundly decreased: in 10 of 20 patients with sJRA and in only 1 of 20 patients with other JRA, levels of NK activity were below two standard deviations of pediatric controls (P = 0.002). Some decrease in perforin expression in NK cells and cytotoxic T lymphocytes was seen in patients within each of the JRA groups with no statistically significant differences. There was a profound decrease in the proportion of circulating CD56bright NK cells in three sJRA patients, a pattern similar to that previously observed in MAS and HLH. In conclusion, a subgroup of patients with JRA who have not yet had an episode of MAS showed decreased NK function and an absence of circulating CD56bright population, similar to the abnormalities observed in patients with MAS and HLH. This phenomenon was particularly common in the systemic form of JRA, a clinical entity strongly associated with MAS.

Figure 1

NK cell cytolytic activity in patients with juvenile rheumatoid arthritis (JRA). Activity of NK cells was determined after co-incubation of peripheral blood mononuclear cells with the NK-sensitive K562 cell line and expressed in cytolytic units (LU, y-axis). NK activity values in 20 patients with systemic JRA, 20 patients with other subtypes of JRA and 41 healthy children are shown as dots. The difference between results for patients with systemic JRA and other JRA is statistically significant (Wilcoxon two-sample test, P = 0.0062). For comparison, mean ± 2SD values determined in healthy individuals are shown as horizontal lines.

Figure 2

Flow cytometric analysis of CD56^bright and CD56^dim natural killer cells. (a) In healthy individuals, most circulating NK cells are CD56^dim (about 90%) and express high levels of perforin. In contrast, CD56^bright NK cells comprise about 10% of all human NK cells and express low levels of perforin. (b, c) Two distinct patterns of CD56 staining observed in patients with systemic juvenile rheumatoid arthritis (sJRA): (b) normal pattern (observed in 17 of 20 patients with sJRA as well as in all controls, and all other JRA patients); (c) almost complete disappearance of CD56^bright subset of NK cells (observed in 3 of 20 patients with sJRA only).

Figure 3

Flow cytometric analysis of perforin expression in NK cells and cytotoxic CD8⁺ cells. (a) Normal control: 95% of NK cells and 14% of CD8⁺ cells are perforin-positive. (b) A patient with systemic juvenile rheumatoid arthritis (sJRA) with a normal pattern of perforin expression: normal mean channel fluorescence (MCF) in NK cells with 92% of NK cells and 10% of CD8⁺ T cells being perforin-positive (this pattern was observed in all controls, in 18 of 20 other JRA patients, and in 15 of 20 sJRA patients). (c) A patient with sJRA with low MCF in NK cells and mildly decreased proportions of perforin-positive NK cells (74%) and CD8⁺ T lymphocytes (5%) (this pattern was observed in 2 of 20 other JRA patients and in 5 of 20 sJRA patients). (d) A patient with sJRA with low MCF in NK cells and very low proportions of perforin-positive NK cells (20%) and no perforin-positive CD8⁺ T lymphocytes (observed in 1 of 20 sJRA patients only).