Elevated levels of cell-free NKG2D-ligands modulate NKG2D surface expression and compromise NK cell function in severe COVID-19 disease

Abstract

Introduction: It is now clear that coronavirus disease 19 (COVID-19) severity is associated with a dysregulated immune response, but the relative contributions of different immune cells is still not fully understood. SARS CoV-2 infection triggers marked changes in NK cell populations, but there are contradictory reports as to whether these effector lymphocytes play a protective or pathogenic role in immunity to SARS-CoV-2.

Methods: To address this question we have analysed differences in the phenotype and function of NK cells in SARS-CoV-2 infected individuals who developed either very mild, or life-threatening COVID-19 disease.

Results: Although NK cells from patients with severe disease appeared more activated and the frequency of adaptive NK cells was increased, they were less potent mediators of ADCC than NK cells from patients with mild disease. Further analysis of peripheral blood NK cells in these patients revealed that a population of NK cells that had lost expression of the activating receptor NKG2D were a feature of patients with severe disease and this correlated with elevated levels of cell free NKG2D ligands, especially ULBP2 and ULBP3 in the plasma of critically ill patients. In vitro, culture in NKG2DL containing patient sera reduced the ADCC function of healthy donor NK cells and this could be blocked by NKG2DL-specific antibodies.

Discussion: These observations of reduced NK function in severe disease are consistent with the hypothesis that defects in immune surveillance by NK cells permit higher levels of viral replication, rather than that aberrant NK cell function contributes to immune system dysregulation and immunopathogenicity.

Keywords: ADCC - antibody dependent cellular cytotoxicity; COVID-19; NKG2D (natural killer group 2 member D); natural killer cells; soluble NKG2D ligands.

Figure 1

NK cells from patients critically ill with COVID-19 disease are less effective mediators of ADCC than NK cells from patients with mild disease. (A) Serum samples from healthy individuals collected pre-pandemia, SARS-CoV-2 infected individuals with mild disease and SARS-CoV-2 infected individuals with life-threatening disease were tested in ADCC assays using PBMC samples from healthy controls. The NK cell response was assayed by measuring degranulation and MIP1β production. (B) The ADCC response mediated by NK cells in PBMCs obtained from mild/asymptomatic cases (squares) or from critically ill patients (circles) was assayed using selected serum samples (see Supplementary Figure 1 ). Samples from patients who died due to COVID-19 disease are coloured red. Statistical significance was assessed using two-tailed Mann-Whitney tests and p values are indicated.

Figure 2

Differences in the numbers and proportions of NK cells (A–C) and various NK cell subpopulations (D–F) between patients with severe or mild COVID-19 disease. Absolute numbers and proportions of selected NK and NK-cell subsets were measured from COVID-19 patients with either mild or life-threatening disease. Counts were based on full blood count analysis to obtain lymphocyte numbers and flow cytometric analysis to identify population frequencies and fluorescence intensities of staining. Normal ranges, based on analysis of more than 40 anonymous healthy controls, are indicated by the light grey shading. ns: p-value > 0.05.

Figure 3

Significant differences in NK cell phenotype between patients with severe or mild COVID-19 disease. Expression intensity of markers that distinguish bulk CD56dim CD16+ NK cells (black line, light grey shaded background) from the cells identified in clusters 4486 (A), 3687 (B), or 3156 (C) (dark grey shading). x-axis data represent marker expression intensity; y-axis data represent frequency. The frequency of cells within clusters identified by CITRUS to differentiate COVID-19 patients with mild (light grey) or severe disease (dark grey) subject groups are also shown. Each point represents an individual subject, and horizontal lines represent the group median. Significance was calculated using the Kruskal Wallis test. P values were further adjusted by multiplying by the number of comparisons made (Bonferroni correction).

Figure 4

The levels of various factors able to modulate expression of the NKG2D receptor were evaluated and significantly elevated levels of cell-free NKG2D ligands ULBP2 and ULBP3 were found in COVID-19 patients with severe disease. The levels of TGFβ (A), MICA (B); ULBP1, ULBP2/5/6 and ULBP3 (C), all of which can downregulate expression of the NKG2D receptor, present in plasma samples from healthy controls and COVID-19 infected individuals were determined by ELISA. Absorbance data were converted to amounts in ng/ml, by reference to standard curves established using recombinant proteins. Statistical significance was assessed using two-tailed Mann-Whitney tests and p values are indicated.

Figure 5

Culture in serum from COVID-19 patients leads to reduced NK cell ADCC activity and this can be partially blocked by NKG2D-L specific mAbs. (A) PBMC from healthy individuals were cultured overnight in medium alone, medium supplemented with human serum from healthy controls, or medium with serum from patients who had developed severe COVID-19 disease. NK cell ADCC activity was measured by flow cytometry for LAMP1 expression on CD3-CD56+ NK cells stimulated after exposure to either P815 cells or P815 cells loaded with agonistic CD16-specific mAb (3G8). (B) PBMC from healthy donors were cultured in medium to which serum from patients who had developed severe COVID-19 disease, in the presence or absence of a cocktail of MICA/B, ULBP1, ULBP2 and ULBP3-specific mAbs. NK cell ADCC activity was measured as described. Statistical significance was assessed using two-tailed Mann-Whitney tests and p values are indicated. ns: p-value > 0.05.

Figure 6

ULBP2 and ULBP3 expression is specifically upregulated in lung tissue from COVID-19 patients. The expression of different NKG2D ligands, and metalloproteases known to be involved in NKG2D-ligand shedding, was analysed using publicly available, pre-processed RNA-seq data obtained on analysis of autopsy specimens from patients who died due after developing COVID-19, or non-COVID, pneumonia. SARS-CoV-2 infected and non-infected lungs were compared using the Log2 normalised counts (plus pseudocount). (A) Expression of ULBP2 and ULBP3 mRNAs in these data. (B) A heatmap to show expression of various NKG2D ligands and metalloproteases in these samples.