NADPH Oxidase 1 Is Associated with Altered Host Survival and T Cell Phenotypes after Influenza A Virus Infection in Mice

Abstract

The role of the reactive oxygen species-producing NADPH oxidase family of enzymes in the pathology of influenza A virus infection remains enigmatic. Previous reports implicated NADPH oxidase 2 in influenza A virus-induced inflammation. In contrast, NADPH oxidase 1 (Nox1) was reported to decrease inflammation in mice within 7 days post-influenza A virus infection. However, the effect of NADPH oxidase 1 on lethality and adaptive immunity after influenza A virus challenge has not been explored. Here we report improved survival and decreased morbidity in mice with catalytically inactive NADPH oxidase 1 (Nox1*/Y) compared with controls after challenge with A/PR/8/34 influenza A virus. While changes in lung inflammation were not obvious between Nox1*/Y and control mice, we observed alterations in the T cell response to influenza A virus by day 15 post-infection, including increased interleukin-7 receptor-expressing virus-specific CD8+ T cells in lungs and draining lymph nodes of Nox1*/Y, and increased cytokine-producing T cells in lungs and spleen. Furthermore, a greater percentage of conventional and interstitial dendritic cells from Nox1*/Y draining lymph nodes expressed the co-stimulatory ligand CD40 within 6 days post-infection. Results indicate that NADPH oxidase 1 modulates the innate and adaptive cellular immune response to influenza virus infection, while also playing a role in host survival. Results suggest that NADPH oxidase 1 inhibitors may be beneficial as adjunct therapeutics during acute influenza infection.

Fig 1. Nox1*^/Y mice have improved survival and delayed weight loss following intranasal challenge with IAV.

B6 and Nox1*^/Y mice were challenged with 50 MID₅₀ IAV. Mortality is plotted in A as percentage of mice surviving over time. A significant difference (p<0.01) was observed between the B6 and Nox1*^/Y mortality curves by both the Log-rank test and the Gehan-Breslow-Wilcoxon Tests. Morbidity is plotted in B as percentage of weight on day of infection (day 0). Results represent mean values ± standard deviation from the mean (S.D.). Average percent original body weights were compared between B6 and Nox1*^/Y mice at individual days p.i. by a Student’s t test. Data compiled from two independent experiments; n = 15–16 mice per group (*, p < 0.05; **, p = 0.01).

Fig 2. Lack of Nox1 catalytic activity alters frequencies of CD8⁺ T cells after IAV infection.

Single-cell suspensions of dLN (A), lung (B, D, E) and spleen (C) harvested from B6 and Nox1*^/Y mice at day 9 and 15 p.i. were stained with antibodies and D^b-NP-IAV pentamer then analyzed by flow cytometry. CD19⁺ cells were excluded from singlet lymphocyte events before differential CD4⁺ and CD8⁺ gating. NP multimer⁺ cells were gated from CD8⁺ cells as demonstrated in the representative FACS dot plots in E. E gated on the day 15 lung CD8⁺ T cell populations. Results in A-D plotted as individual mice with the horizontal bar indicating the mean. 2-way ANOVA indicates that the Nox1*^/Y genotype affects the percentage of CD8⁺ T cells plotted in A (p < 0.05), C (p < 0.01) and D (p < 0.05). Bonferonni post-test indicates significant difference between B6 and Nox1*^/Y mice at day 15 p.i. for C and D. Data compiled from two independent experiments; n = 9–11 mice per group (*, p < 0.05; **, p < 0.01).

Fig 3. Nox1*^/Y mice have a greater percentage of CD127⁺ NP-specific CD8⁺ T cells at day 15.

Single-cell suspensions of lung (A, C, E) and dLN (B, D, F) harvested from B6 and Nox1*^/Y mice at day 9 and day 15 p.i. were stained with antibodies and D^b-NP-IAV pentamer then analyzed by flow cytometry. CD19⁺ cells were excluded from singlet lymphocyte events before differential CD4⁺ and CD8⁺ gating. CD127 and KLRG-1 staining was analyzed on D^b-NP-IAV pentamer⁺ cells among CD8⁺ cells as demonstrated in the representative FACS dot plots taken from day 15 p.i. samples in A and B. Results in C-F plotted as individual mice with horizontal bar indicating mean. 2-way ANOVA indicates that the Nox1*^/Y genotype contributes to increased lung CD127⁺ cells (C, p < 0.001) and decreased double-negative CD127^-KLRG-1^- expressing cells (E, p < 0.0001; F, p < 0.05). Bonferonni post-test indicates significant difference between B6 and Nox1*^/Y mice at day 15 p.i. for C and D, and at both day 9 and day 15 for E. Data compiled from two independent experiments; n = 9–11 mice per group (*, p < 0.05; **, p < 0.01; ****, p < 0.0001).

Fig 4. Nox1*^/Y mice have increased IAV-specific cytokine-producing T cells at day 15.

Single-cell suspensions of lung (A, E) and spleen (B-D, F-N) harvested from B6 and Nox1*^/Y mice at day 15 p.i. were incubated overnight in the presence or absence of IAV at 1 MOI. Next, cells were stained extracellularly and intracellularly with antibodies and analyzed by flow cytometry. CD4⁺ and CD8⁺ populations gated from singlet lymphocyte events. A-D, representative dot plots of virus-stimulated B6 or Nox1*^/Y cells gated on the indicated T cell population. E-H, individual cytokines presented as a percent of the CD4⁺ or CD8⁺ populations. Results represent mean values ± S.D. I-J, representative dot plots of virus-stimulated B6 or Nox1*^/Y cells gated on the indicated T cell population. Red dots are IL-2⁺. K-N, data from Boolean gating of cytokines gated as in A-D presented as a percent of total cytokine (IFN-γ, TNF-α or IL-2)-producing CD4⁺ or CD8⁺ T cells. 2-way ANOVA indicates that the Nox1*^/Y genotype contributes to increased cytokine-producing cells after ex vivo virus stimulation in E-H (p < 0.01). Bonferroni post-test indicates significant difference between stimulated cells from B6 and Nox1*^/Y mice for E-H. Data compiled from two independent experiments; n = 11 mice per group (**, p < 0.01; ***, p < 0.001).

Fig 5. Germinal center B cell frequencies and HAI titers unaffected by lack of Nox1 catalytic activity.

Single-cell suspensions of dLN (A-B) harvested from B6 and Nox1*^/Y mice at day 3 and 6 p.i. were stained with antibodies and analyzed by flow cytometry. A: B220⁺ IgD^- cells were gated from singlet leukocyte events. B: GC B cells were identified as the CD38^- GL7⁺ population among IgD^- B cells. C: day 9 and 15 serum titers of HA-binding antibody as determined by HAI assay. Results in A-C plotted as individual mice with horizontal bar indicating mean. Data compiled from two independent experiments; in A & B, n = 8 mice per group; in D, n = 9–11 mice per group.

Fig 6. Lack of Nox1 catalytic activity does not affect lung inflammatory protein levels or viral titers.

Supernatant harvested from whole lung homogenate was tested for cytokine and chemokine protein levels by Bioplex assay (A-G) and ELISA (H). Supernatant MPO protein levels were determined by ELISA (I). IAV titers were determined by EID₅₀ assay (J). Assay limits of detection depicted by dashed horizontal line. Results in A-I represent mean values ± S.D. All samples run in duplicate; average of two wells used to generate final data. Result in J plotted as individual mice with horizontal bar indicating mean. Data compiled from two independent experiments, n = 7–8 mice per group.

Fig 7. A greater percentage of cDCs and iDCs in the dLN of Nox1*^/Y mice express CD40.

Single-cell suspensions of dLN harvested from B6 and Nox1*^/Y mice at day 3 and 6 p.i. were stained with antibodies and analyzed by flow cytometry. Among SSC^lo singlet lymphocytes, CD11c^hi B220^hi cells were gated for the pDC population. All other DC populations were gated from the B220^- cells using the gating strategy described by Ballesteros-Tato et al. [43], including tDCs (CD11c⁺ MHCII^hi) from which were gated both the CD103⁺ and CD103^- tDC populations, cDCs (CD11c^hi MHCII^int) from which were gated the CD8α⁺, CD4⁺ and double negative cDC populations, and iDCs (CD11c^int MHCII^int). CD40 expression was analyzed on individual DC populations as shown in representative histogram plots (A-F). Results in G-I plotted as individual mice with horizontal bar indicating mean. 2-way ANOVA indicates the Nox1*^/Y genotype affects the percentage of populations graphed in G (p < 0.01), H and I (p < 0.001). Bonferonni post-test indicates significant difference between B6 and Nox1*^/Y mice at day 3 p.i. for G. Data compiled from two independent experiments; n = 9–11 mice per group (*, p < 0.05).