An altered natural killer cell immunophenotype characterizes clinically severe pediatric RSV infection

Abstract

Respiratory syncytial virus (RSV) infects nearly all children by 2 years of age and is a leading cause of pediatric hospitalizations. A subset of children with RSV infection (RSV⁺ children) develop respiratory failure requiring intensive care, but immune mechanisms distinguishing severe pediatric RSV infection are not fully elucidated. Natural killer (NK) cells are key innate immune effectors of viral host defense. In this study of 47 critically ill RSV⁺ children, we coupled NK cell immunophenotype and cytotoxic function with clinical parameters to identify an NK cell immune signature of severe pediatric RSV disease. Airway NK cells were increased in intubated RSV⁺ children with severe hypoxemia and prolonged duration of mechanical ventilation and were correlated with clinical severity scores. Peripheral blood NK cells were decreased in RSV⁺ patients and had altered activating receptor expression, with increased expression of CD69 and decreased expression of NKG2D. Ex vivo, circulating NK cells from RSV⁺ patients exhibited functional impairment characterized by decreased cytotoxicity as well as aberrant immune synapse assembly and lytic granule trafficking. NK cell frequency and phenotype correlated with clinical measures that defined disease severity. These findings implicate a role for NK cells in mediating RSV immunopathology and suggest that an altered NK cell immunophenotype is associated with severe RSV disease in young children.

Fig. 1.. Flowchart of study participants.

Study participants (n = 69) included 47 RSV⁺ patients and 22 uninfected control patients. Biospecimens were available for airway leukocyte analysis in ETAs from 14 ventilated patients and for circulating leukocyte analysis in peripheral blood from 55 patients. ETAs were collected from ventilated RSV⁺ patients with moderate hypoxemia (gray, n = 5) or severe hypoxemia (black, n = 5) and uninfected ventilated control patients (white, n = 4). Peripheral blood was collected from RSV⁺ patients supported with HFNC (light orange, n = 15), NIV (dark orange, n = 17), or mechanical ventilation (Vent; red, n = 5) and from uninfected HCs (white, n = 18).

Fig. 2.. Airway NK cells are increased in severe pediatric RSV.

(A) Chest radiographs of an uninfected ventilated control patient and ventilated RSV⁺ children with moderate (SPO₂/FiO₂ ratio = 217) or severe hypoxemia (SpO₂/FiO₂ ratio = 85). (B) RSV⁺ patients were categorized by degree of hypoxemia as moderate (SpO₂/FiO₂ ratio > 200; n = 5, gray bar) or severe (SpO₂/FiO₂ ratio < 200; n = 5, black bar). Uninfected ventilated controls were intubated for noninfectious reasons (n = 4; white). (C) pSOFA scores on day 1 of PICU admission are shown. (D) Lymphocytes (Lymphs) were identified in ETAs by FSC/SSC using flow cytometry, and T cells (CD3⁺) and NK cells (CD3⁻NKp46⁺) were gated. (E and F) Endotracheal T cells (E) and NK cells (F) were enumerated as the percent of total lymphocytes. (G) Number of days requiring mechanical ventilation for uninfected controls, patients with moderate RSV, and patients with severe RSV. Bars depict the means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.0005, and ****P < 0.0001 by one-way ANOVA with Tukey’s multiple comparison test. (H to J) Relationships between the frequency of endotracheal NK cells and days ventilated (H), SpO₂/FiO₂ ratio (I), and pSOFA scores (J) are shown. Pearson correlation coefficient (r) and significance (P) are noted.

Fig. 3.. Circulating NK cells are decreased in children with RSV and exhibit altered subset distribution.

(A) RSV⁺ children with respiratory failure were supported with either HFNC (n = 15), NIV (n = 17), or intubation and mechanical ventilation (Vent, n = 5). (B) pSOFA scores were calculated for each RSV⁺ patient in the first 24 hours of admission. (C) Circulating NK cells were identified by flow cytometry as CD3⁻CD56⁺ lymphocytes and enumerated per microliter of blood in HCs (white, n = 18) and RSV⁺ patients (HFNC, light orange; NIV, dark orange; Vent, red). (D and E) The frequency of CD3⁺ T cells (D) and NK cells (E) as the percentage of circulating lymphocytes was calculated for HCs (white, n = 18) and all RSV⁺ patients (yellow, n = 37). (F) CD56^bright and CD56^dim NK cell subsets were determined by CD56 and CD16 expression. Plots are shown for representative HC (left) and RSV⁺ (right) samples, and quadrant numbers denote the percentage of total NK cells. (G to J) Absolute numbers [(G) and (H)] and relative frequencies [(I) and (J)] of CD56^bright [(G) and (I)] and CD56^dim [(H) and (J)] NK cell subsets were calculated for HC and RSV⁺ peripheral blood samples. Bars denote the means ± SEM. *P < 0.05, **P < 0.01, and ****P < 0.0001 by one-way ANOVA with Tukey’s multiple comparison test [(B) and (C)] or two-tailed unpaired Student’s t test [(D), (E), and (G) to (J)].

Fig. 4.. Severe RSV is distinguished by altered NK cell activating receptor expression.

(A to E) NK cells were identified in whole blood by flow cytometry as CD3⁻CD56⁺ lymphocytes, and expression levels of activation marker CD69 [(A) and (B)] and activating receptors NKG2D [(C) and (D)] and NKp46 [(E) and (F)] were determined as the percentage of total NK cells. Histograms [(A), (C), and (E)] show the percentage expression of the indicated receptor in representative HC (white) and RSV⁺ (yellow) samples. Bars [(B), (D), and (F)] depict the means ± SEM in the HC (white, n = 18), RSV⁺ HFNC (light orange, n = 15), RSV⁺ NIV (dark orange, n = 17), and RSV⁺ Vent (red, n = 5) subgroups. *P < 0.05 and ***P < 0.0005 by one-way ANOVA with Tukey’s multiple comparison test.

Fig. 5.. NK cells are shifted toward a CD69⁺NKG2D⁻ subphenotype in severe RSV.

Coexpression of CD69 and NKG2D was evaluated on peripheral blood NK cells by flow cytometry two-dimensional Boolean gating analysis. (A) Representative flow plots are shown from HC (left) and RSV⁺ (right) samples, with the CD69⁺NKG2D⁻ subphenotype outlined in blue (top left quadrant) and the CD69⁻NKG2D⁺ subphenotype outlined in green (bottom right quadrant). Quadrant numbers denote the percentage of total CD3⁻CD56⁺ NK cells. (B) Pie charts show the mean percentage of each subphenotype in HC (white label, n = 18), RSV⁺ HFNC (light orange label, n = 15), RSV⁺ NIV (dark orange label, n = 17), and RSV⁺ Vent (red label, n = 5) subgroups. CD69⁻NKG2D⁺ (green), CD69⁺NKG2D⁻ (blue), CD69⁺NKG2D⁺ (dark gray), and CD69⁻NKG2D⁻ (light gray) subphenotypes are shown. (C to F) The frequencies of CD69⁻NKG2D⁺ (C), CD69⁺NKG2D⁻ (D), CD69⁺NKG2D⁺ (E), and CD69⁻NKG2D⁻ (F) subphenotypes as the percentage of total circulating NK cells were determined for each participant. Bars depict the means ± SEM. *P < 0.05, ***P < 0.0005, and ****P < 0.0001 by one-way ANOVA with Tukey’s multiple comparison test.

Fig. 6.. NK cells from children hospitalized with RSV have impaired cytotoxicity despite increased cytotoxic mediator release.

(A) Experimental schema. NK cells were negatively selected from fresh PBMCs from RSV⁺ patients or HCs and cocultured with K562 target cells at a 2:1 effector:target ratio for 2 hours. (B) NK cell–directed apoptosis of target cells was evaluated by flow cytometry staining for annexin V and 7-AAD in cocultures with NK cells from HC (left) or RSV⁺ (right) samples. Quadrant numbers show the percentage of target cells. (C) Specific target cell apoptosis was evaluated by total annexin V⁺ staining in HC (white, n = 4) and RSV⁺ samples (yellow, n = 4, including one RSV⁺ HFNC patient, two RSV⁺ NIV patients, and one RSV⁺ Vent patient). (D) Concentrations of the NK cell cytotoxic mediators perforin, granulysin, granzyme A, granzyme B, and soluble Fas (sFas) were measured in the coculture supernatant. Bars show the means ± SEM. *P < 0.05 by Mann-Whitney U test.

Fig. 7.. NK cells in RSV have disrupted cytoskeletal rearrangement at the IS.

(A and B) NK cells from HCs (A) or RSV⁺ children (B) were isolated from PBMCs and cocultured with K562 target cells for 2 hours. Confocal microscopy was used to image NK cell–target conjugates for intracellular F-actin (red), perforin (green), and α-tubulin (blue). Representative images are shown. (C) F-Actin accumulation at the IS was calculated (integrated F-actin density = area × MFI) at the NK cell–target interface in HC (white, n = 4) or RSV⁺ samples (yellow, n = 2, including one RSV⁺ HFNC and one RSV⁺ NIV patient). au, arbitrary units. (D) α-Tubulin was used to identify the NK cell MTOC, and the distance from the MTOC to the IS was calculated for each NK cell–target conjugate. (E) Pie charts depict the percentage of NK cells with the MTOC polarized to the IS (< 2 μm) in the HC and RSV⁺ cohorts. Twenty to 30 NK cell–target conjugates were imaged for each sample. ***P < 0.0005 and ****P < 0.0001 by unpaired two-tailed Student’s t test.

Fig. 8.. NK cells are associated with clinical measures of critical illness in pediatric RSV.

(A and B) Two-dimensional loading plots and score plots from multivariate PCA were determined for clinical severity indicators and NK cell immunophenotyping markers in the ventilated endotracheal NK cell cohort [n = 5 RSV⁺ moderate (gray), n = 5 RSV⁺ severe (black) (A)] and circulating NK cell cohort {n = 37 RSV⁺ patients [n = 15 HFNC (light orange), n = 17 NIV (dark orange), and n = 5 Vent (red) (B)]}. Clinical severity indicators include ventilator days, ICU lengths of stay, need for vasopressor support, need for respiratory support escalation, blood gas parameters (pH and pCO₂), pSOFA scores, and SPO₂/FiO₂ ratios. NK cell immunophenotyping markers include ETA percentage NK cells (A), and peripheral blood NK cell number and CD69 and NKG2D expression (B) are highlighted in light blue. PC, principal component.