Autoantibodies to Killer Cell Immunoglobulin-Like Receptors in Patients With Systemic Lupus Erythematosus Induce Natural Killer Cell Hyporesponsiveness

Abstract

Natural killer (NK) cell cytotoxicity toward self-cells is restrained by the inhibitory HLA class I-binding receptors CD94/NKG2A and the killer cell immunoglobulin-like receptors (KIRs). CD94/NKG2A and KIRs are also essential for NK cell education, which is a dynamic functional maturation process where a constitutive binding of inhibitory receptors to cognate HLA class I molecules is required for NK cells to maintain their full cytotoxic capacity. Previously, we described autoantibodies to CD94/NKG2A in patients with systemic lupus erythematosus (SLE). In this study we analyzed sera from 191 patients with SLE, 119 patients with primary Sjögren's syndrome (pSS), 48 patients with systemic sclerosis (SSc), and 100 healthy donors (HD) for autoantibodies to eight different KIRs. Anti-KIR autoantibodies were identified in sera from 23.0% of patients with SLE, 10.9% of patients with pSS, 12.5% of patients with SSc, and 3.0% of HD. IgG from anti-KIR-positive SLE patients reduced the degranulation and cytotoxicity of NK cells toward K562 tumor cells. The presence of anti-KIR-autoantibodies reacting with >3 KIRs was associated with an increased disease activity (p < 0.0001), elevated serum levels of IFN-α (p < 0.0001), nephritis (p = 0.001), and the presence of anti-Sm (p = 0.007), and anti-RNP (p = 0.003) autoantibodies in serum. Together these findings suggest that anti-KIR autoantibodies may contribute to the reduced function of NK cells in SLE patients, and that a defective NK cell function may be a risk factor for the development of lupus nephritis.

Keywords: autoantibody; killer cell immunoglobulin-like receptor; natural killer cells; nephritis; primary Sjögren's syndrome; systemic lupus erythematosus.

Figure 1

Autoantibodies to KIRs are present in sera from patients with systemic lupus erythematosus, primary Sjögren's syndrome and systemic sclerosis. (A) IgG-binding to KIR transfectants in sera from 100 healthy donors (HD), 191 systemic lupus erythematosus (SLE), 119 primary Sjögren's syndrome (pSS), and 48 systemic sclerosis (SSc) patients. Dotted lines denote a z-score of ±4. (B) Frequency of HD and patients displaying anti-KIR reactivity to at least 1 KIR. (C) The number of KIRs recognized by the anti-KIR-positive sera. (D) A heatmap visualizing the KIR-specificity in individual SLE patients. Each column represent serum from one patient and each row one KIR. The number of sera reacting with each KIR is denoted to the right. (E) The strongest KIR-reactivity in each sera in SLE patients reacting with 1, 2, 3, or >3 KIRs. Dotted line indicates a z-score of 10. Differences between groups were assessed using the (B) Fischer's exact test or (E) Mann-Whitney U-test.

Figure 2

Anti-KIR autoantibodies from SLE patients block the binding of monoclonal anti-KIR antibodies. (A,B) Flow-cytometric KIR stainings of healthy donor (HD) cells following 30 min incubation with 50% serum from patients harboring >3 anti-KIRs (n = 10), KIR-positive HD (n = 3), and KIR-negative (n = 8) HD (A) Flow cytometry plots from one representative HD and the three SLE patients that blocked the binding of anti-KIR2DL1/DS1. (B) The frequency of KIR-positive cells in serum-treated cells relative to the mean frequency of KIR-positive cells in eight non-treated cells. (C,D) Flow-cytometric KIR stainings of CD3⁻CD56^dim NK cells in PBMCs from anti-KIR-negative (n = 34), and anti-KIR-positive SLE patients (n = 2). Data are from two separate experiments. (C) Flow cytometry plots from SLE3, SLE138 and one representative HD from each experiment. (D) Data for all patients presented as boxplots, with the median, interquartile range, and range denoted.

Figure 3

Anti-KIR autoantibodies induce hyporesponsiveness in NK cells. (A) NK cell degranulation and (B) viability of IL-2-activated PBMCs from 3 HDs following incubation with K562 cells, rituximab (RTX, anti-CD20), or IgG as indicated. (C) K562-induced degranulation in NK cells from 6 HDs exposed to IgG from patient SLE3, SLE138, or HDs (IVIG) relative to no IgG. P-values from a repeated measures one-way ANOVA comparing the relative degranulation in IgG-treated cells to untreated cells are shown. (D) Correlation between the difference in degranulation and the difference in cytotoxicity of IgG-treated NK cells compared to non-treated NK cells. (E) Comparison of the education level and the inhibitory effect of SLE138-IgG on degranulation compared to no IgG. Education level was defined as the difference in degranulation in indicated NK cell subsets relative to KIR-negative NK cells (KIR2DL1/DS1⁻KIR2DL2/DL3/DS2⁻KIR3DL1/DS1⁻) in the absence of IgG. (A–E) Data from patient SLE3 and SLE138 are colored in red and blue, respectively.

Figure 4

NK cells retune following wash-out of IgG. (A) Surface stainings of indicated NK cell receptors, and (B) degranulation of NK cells from 10 HDs was determined before (0 h), and after wash-out of IgG (48 and 96 h). Degranulation of NK cells treated with IgG from SLE3 relative to IVIG-treated cells. Data are shown as (A) mean values with error bars representing the standard error of the mean, and (B) boxplots with median, interquartile range, and range. (B) The one-sample t-test was used to assess whether the relative degranulation differed from 100%. (C) Correlation between data in (A,B). The p-values and Pearson's correlation coefficients (r) from linear regressions are denoted.

Figure 5

The presence of anti-KIR autoantibodies to >3 KIRs is associated with an increased risk for nephritis and the presence of anti-Sm antibodies. (A,B,D) Forest plots depicting the relative risk and 95% confidence intervals of fulfilling clinical characteristics as specified. (A,B) The relative risk of SLE patients presenting with 1, 2, 3, or >3 anti-KIR autoantibodies compared to anti-KIR-negative patients. (C) The strongest levels of anti-KIR autoantibodies in anti-KIR-positive SLE patients stratified by the presence (n = 21) or absence (n = 23) of nephritis. The median values are denoted with vertical bars. (D) The relative risk of fulfilling the ACR criteria for nephritis in patients positive relative to negative for the indicated autoantibodies. (A,B,D) Fischer's exact test with p < 0.05 denoted in the graphs. (C) Mann-Whitney U-test.