Natural Killer Cells Dampen the Pathogenic Features of Recall Responses to Influenza Infection

Abstract

Despite evidence of augmented Natural Killer (NK) cell responses after influenza vaccination, the role of these cells in vaccine-induced immunity remains unclear. Here, we hypothesized that NK cells might increase viral clearance but possibly at the expense of increased severity of pathology. On the contrary, we found that NK cells serve a homeostatic role during influenza virus infection of vaccinated mice, allowing viral clearance with minimal pathology. Using a diphtheria toxin receptor transgenic mouse model, we were able to specifically deplete NKp46+ NK cells through the administration of diphtheria toxin. Using this model, we assessed the effect of NK cell depletion prior to influenza challenge in vaccinated and unvaccinated mice. NK-depleted, vaccinated animals lost significantly more weight after viral challenge than vaccinated NK intact animals, indicating that NK cells ameliorate disease in vaccinated animals. However, there was also a significant reduction in viral load in NK-depleted, unvaccinated animals indicating that NK cells also constrain viral clearance. Depletion of NK cells after vaccination, but 21 days before infection, did not affect viral clearance or weight loss-indicating that it is the presence of NK cells during the infection itself that promotes homeostasis. Further work is needed to identify the mechanism(s) by which NK cells regulate adaptive immunity in influenza-vaccinated animals to allow efficient and effective virus control whilst simultaneously minimizing inflammation and pathology.

Keywords: NK cell; diphtheria toxin; influenza; lung; mice; vaccine.

Figure 1

Primary influenza infection induces rapid weight loss and NK cell activation in lung but vaccination reduces weight loss and lung viral burden. (A) C57BL/6 female mice were challenged intranasally with 0.5 hemagglutination units (HAU) of influenza A/California/4/2009 (Flu) or mock treated with DPBS (Mock). Four weeks prior to challenge, mice were vaccinated intraperitoneally with the trivalent Sanofi influenza vaccine (Vac). (B) Weight loss (mean ± SEM) over 4 days. (B–F) At day 4 post infection, lungs were excised and cell-free supernatant was analyzed by qPCR for influenza viral burden (plotted against a dose curve of Flu with known HAU, giving HAU equivalents) (C) and plotted against weight loss (D). Data fitted to a non-linear regression line with R square value shown (D). (E,F) Lung cell pellets were analyzed by flow cytometry for (E) cellular abundance and (F) Natural Killer (NK) activation markers. Data is compiled from two independent experiment (n = 5–11/group), with each independent experimental data shown in Figures S1, S2. Dots represent individual mice with bars showing mean. Significance determined by Mann–Whitney U-test, ns, not significant.

Figure 2

NK cell depletion reduces lung viral burden and increases weight loss in vaccinated, influenza-challenged mice. Transgenic C57BL/6 mice with NKp46 driven expression of diphtheria toxin (DT) receptor were vaccinated 28 days prior to intranasal influenza (flu) challenge, as in Figure 1. (A) Immediately prior to infection, a subset of mice received two intraperitoneal injections of DT (1.25 μg). (B) Levels of NK1.1+, NKp46+ NK cells in the lung, as a proportion of singlet, live leukocytes at 4 days post infection (necropsy, nx). (C) At 4 days post infection, lung cell-free supernatants were analyzed by qPCR for influenza viral burden (plotted against a dose curve of IFA with known HAU, giving HAU equivalents). (D) Weight loss (mean ± SEM) followed for 14 days post challenge, one experiment (n = 4–5/group) (E–F) Weight loss at day 4 post challenge. Data from males {M} (n = 9–10/group) is compiled from two independent experiments and females {F} (n = 9–13/group) from three independent experiments, with each independent experimental data shown in Figure S3. Dots represent individual mice with bars showing mean. Significance determined by Mann–Whitney U-test, ns, not significant. ^*p < 0.05.

Figure 3

NK cell depletion in vaccinated, influenza-challenged mice does not change inflammatory transcripts. Four days post infection in the model described in Figure 2A, (A) Lung RNA was analyzed by qPCR for influenza viral burden (plotted against a dose curve of influenza with known HAU, giving HAU equivalents per 5 μg RNA tested). (B,C) Transcript levels of inflammatory cytokine genes (B) Il6 and Ifnγ and (C) neutrophil-related chemokines Cxcl1 and Cxcl2, along with neutrophil lipocalin protein (Lcn2). RNA induction normalized to housekeeping gene β-actin and displayed as induction over mock-treated control mice. (D) Plasma levels of IL-6 (pg/mL). (E) IL-6 levels in lung supernatants after whole lung enzymatic digestion for single cell isolation and viral burden quantification. Data from females {F} (n = 5–9/group) and males {M} (n = 7–10/group) is compiled from two independent experiments, with each independent experimental data shown in Figures S4, S5. Dots represent individual mice with bars showing mean. Significance determined by Mann–Whitney U-test, ns, not significant.

Figure 4

NK cell depletion in vaccinated, influenza-challenged mice does not alter pathology. Four days post infection in the model described in Figure 2A, whole lungs were excised, stored in 10% formalin, and embedded on paraffin for hematoxylin and eosin staining. Pathology was scored for: (A) Inflammation (vasculitis, bronchiolitis, and alveolitis), Edema (perivascular, peribronchiolar, and alveolar), Leukocytes and Neutrophils (in perivascular space, peribronchiolar space, and alveolar wall). Full scoring details in Supplementary Files. Pathology scores from females {F} (n = 4–8/group) is compiled from two independent experiments, with each independent experimental data shown in Figure S6. Dots represent individual female {F} mice with bars showing mean. Significance determined by Mann–Whitney test, ns, not significant. (B) Representative photomicrographs of pathological changes observed. Arrows, bronchiolitis with exudates in bronchiolar lumina. ^*Perivascular and peribronchiolar edema. Scale bar in mock equals 50 μm.

Figure 5

NK cell depletion in vaccinated, influenza-challenged mice increases lung CD3+ T cell and neutrophil infiltration. Four days post infection in the model described in Figure 2A, whole lungs were excised and single cells isolated for flow cytometry. Proportion (%) of (A) CD3+ T cells, (B) CD3+CD4+ and CD3+CD8+ T cells, and (C) active (CD69+) CD3+ T cells, as determined from singlet, live lung leukocytes. Proportion (%) of (B) CD19+ B cells, (C) Ly6C-high inflammatory monocytes, and (D) Ly6G+ neutrophils. Data in (A) from males {M} (n = 7–14/group) is compiled from three independent experiments and data in (B–F) from males {M} (n = 7–10/group) is compiled from two independent experiments, with each independent experimental data shown in Figure S9. Dots represent individual mice, with bar representing mean. Significance determined by Mann-Whitney U test, ns, not significant.

Figure 6

Depletion of NK cells after vaccination, and subsequent repopulation, does not alter lung viral burden, disease severity or systemic inflammation after challenge. (A) Transgenic C57BL/6 mice with NKp46 driven expression of diphtheria toxin (DT) receptor were vaccinated 42 days (d) prior to intranasal influenza (Flu) challenge and treated with DT (NK-depleted) 21 days prior to challenge with necropsy (nx) at 4 days post influenza challenge. (B) At 3 and 21 days post DT treatment, lungs were excised and single cells isolated for flow cytometry for the proportion (%) of NK1.1+, NKp46+ NK cells. (C) Weight loss at 4 days post influenza challenge. (D) Lung cell-free supernatants were analyzed by qPCR for influenza viral burden (plotted against a dose curve of Flu with known HAU, giving HAU equivalents). (E) Plasma levels of IL-6 (pg/mL). Data from males {M} (n = 4–10/group) is compiled from two independent experiments, with each independent experimental data shown in Figure S10. Dots represent individual mice with bars showing mean. Significance determined by Mann-Whitney test, ns, not significant.

Figure 7

Summary model of role of NK cells in influenza vaccination and challenge. (A) Influenza virus infection in naïve mice results in an influx of inflammatory monocytes (Ly6C^high) into the lung, with evident histopathology and rapid weight loss. (B) Vaccination reduces lung virus load and associated pathology (weight loss). (C) In unvaccinated animals, NK cell depletion reduces viral load and associated inflammation after influenza infection compared to NK cell replete animals, but animals still lose a significant amount of weight. (D) NK depletion of vaccinated mice further reduces influenza viral load and weight loss after infection, yet pathology scores are similar to those of NK intact mice. However, vaccination combined with NK depletion is accompanied by an increase in infiltration of neutrophils and activated T cells into the lung after influenza infection, suggesting that NK cell depletion enhances the control of virus replication as a result of enhanced adaptive immunity but that this comes at the price of significantly increased morbidity. Model created with BioRender.