Dual oxidase 1 promotes antiviral innate immunity

Abstract

Dual oxidase 1 (DUOX1) is an NADPH oxidase that is highly expre-ssed in respiratory epithelial cells and produces H₂O₂ in the airway lumen. While a line of prior in vitro observations suggested that DUOX1 works in partnership with an airway peroxidase, lactoperoxidase (LPO), to produce antimicrobial hypothiocyanite (OSCN^-) in the airways, the in vivo role of DUOX1 in mammalian organisms has remained unproven to date. Here, we show that Duox1 promotes antiviral innate immunity in vivo. Upon influenza airway challenge, Duox1^-/- mice have enhanced mortality, morbidity, and impaired lung viral clearance. Duox1 increases the airway levels of several cytokines (IL-1β, IL-2, CCL1, CCL3, CCL11, CCL19, CCL20, CCL27, CXCL5, and CXCL11), contributes to innate immune cell recruitment, and affects epithelial apoptosis in the airways. In primary human tracheobronchial epithelial cells, OSCN^- is generated by LPO using DUOX1-derived H₂O₂ and inactivates several influenza strains in vitro. We also show that OSCN^- diminishes influenza replication and viral RNA synthesis in infected host cells that is inhibited by the H₂O₂ scavenger catalase. Binding of the influenza virus to host cells and viral entry are both reduced by OSCN^- in an H₂O₂-dependent manner in vitro. OSCN^- does not affect the neuraminidase activity or morphology of the influenza virus. Overall, this antiviral function of Duox1 identifies an in vivo role of this gene, defines the steps in the infection cycle targeted by OSCN^-, and proposes that boosting this mechanism in vivo can have therapeutic potential in treating viral infections.

Keywords: DUOX1; Dual oxidase 1; hypothiocyanite; influenza; lactoperoxidase.

Fig. 1.

Duox1^−/− mice demonstrate increased mortality, morbidity, and impaired viral clearance following influenza airway infection. (A) Kaplan–Meier survival curves of Duox1^−/− (n = 28) and WT (n = 29) mice infected i.n. with A/Puerto Rico/8/1934 (H1N1) (PR8) influenza strain (50 PFU) followed for 12 d (chi-square test). Weight loss of Duox1^−/− and WT mice infected as in A is presented as B, the daily average for the 12-d duration of the study or (C) individual animal data on 4 to 7 dpi, before and on the day of the onset of mortality on 8 dpi (Mann–Whitney U test). Duox1^−/− and WT mice were infected i.n. with (D) PR8 (50 PFU) or (E) the nonlethal A/Swine/Illinois/02860/2009 H1N2 (10⁵ PFU) influenza virus strain, and viral titers in whole-lung homogenates were determined using PFU assay on MDCK cells at the indicated time points. The dotted lines indicate the limit of detection of the assay (unpaired t test). (F) Influenza virus NP expression was measured by immunohistochemistry in the lungs of Duox1^−/− and WT mice infected i.n. with PR8. Representative images (n = 10 WT and n = 15 Duox1^−/− mice) are shown. (G) Quantification of influenza NP expression determined as in F in Duox1^−/− and WT mice 7 dpi using the Fiji software (n = 10 WT and n = 15 Duox1^−/− mice) (Mann–Whitney U test). (H) Gene expression of H1N1 influenza virus was determined in the lungs of WT and Duox1^−/− animals by qPCR. Bar graphs display mean. n = 7. (I) Quantification of H₂O₂ concentrations in the BAL of uninfected Duox1^−/− (n = 14) and WT (n = 16) mice by Amplex Red fluorescence (Mann–Whitney U test). (J) Duox1^−/− and WT mice were i.n. infected with PR8 (50 PFU) as in A, lungs were fixed, and immunofluorescence staining was performed using an anti-DUOX1 antibody (red), an anti-NP antibody identifying influenza virus NP (green), and DAPI (blue) to detect DNA. Merged images indicate the overlay of red, green, and blue. The white line and text “lumen” on the blue images indicate the border between the airway lumen and the epithelium. (Scale bars, 20 μm.) (K) Gene expression of Duox1 was measured in the lungs of WT and Duox1^−/− animals by qPCR. Bar graphs display mean. n = 3. (L) Total protein levels were determined in the BAL of PR8-infected (n = 9 WT and n = 5 Duox1^−/−) mice on 7 dpi using the BCA assay (Mann–Whitney U test). (M) Duox1^−/− and WT mice were either uninfected (0 dpi) or i.n. infected with PR8 (50 PFU), and BAL concentrations of LPO were determined by ELISA at indicated days (n = 32 WT, n = 30 Duox1^−/− animals with a minimum number of 8 mice per day). (N) SCN⁻ concentrations were determined by mass spectrometry in BAL fluid of PR8-infected mice (n = 28 WT and n = 21 Duox1^−/−, with a minimum number of 4 mice per time point) throughout the infection (1 to 5 dpi). Significance levels are indicated as *P < 0.05, **P < 0.01, and ***P < 0.001. Abbreviations: AU, arbitrary unit; BAL, bronchoalveolar lavage; DAB, 3,3′-diaminobenzidine; dpi, day postinfection; Duox1, Dual oxidase 1; i.n., intranasally; LPO, lactoperoxidase; NP, nucleoprotein; ns, nonsignificant; PFU, plaque-forming unit; SCN⁻, thiocyanate; WT, wild-type.

Fig. 2.

Duox1^−/− mice have an altered cytokine response following influenza lung infection. Concentrations of indicated cytokines/chemokines in the BAL of WT (n = 19) and Duox1^−/− mice (n = 20) infected with the A/Puerto Rico/8/1934 (H1N1) (PR8) influenza virus strain (50 PFU) were determined by a multiplex bead-based ELISA array or classical, microplate-based ELISAs (IFN-α, IFN-λ) at 3, 5, and 7 dpi. Data are presented as mean ± SEM. Statistical comparisons were done with Mann–Whitney U test at each dpi between Duox1^−/− and WT mice. Significance levels are indicated as *P < 0.05, **P < 0.01, and ***P < 0.001. Abbreviations: dpi, day postinfection; Duox1, Dual oxidase 1; WT, wild-type.

Fig. 3.

Enhanced epithelial apoptosis in influenza virus–infected Duox1^−/− mice. (A) Histopathological micrographs of hematoxylin and eosin-stained lung tissue sections of Duox1^−/− and WT mice that were left uninfected or were infected i.n. with the A/Puerto Rico/8/1934 (H1N1) (PR8) influenza virus strain (50 PFU) at 3 and 7 dpi are shown. Representative images of at least five similar results are shown. Magnification: 40×. (Scale bars, 20 μm.) (B) Clinical scores of alveoli, bronchioles, and blood vessels in PR8-infected Duox1^−/− and WT mice at 3 and 7 dpi (n = 5 to 12). Statistical comparisons were done by Mann–Whitney U test at each dpi between genotypes. (C) High magnification images of affected lungs in WT and Duox1^−/− mice at 12 dpi. WT: The alveolar spaces are obscured due to accumulation of histiocytes (arrows) and syncytial cells (dashed arrows). There is also some residual, mostly lymphocytic inflammatory infiltrate within alveolar spaces. The histiocytes represent residual inflammation, while the syncytial cells are formed by fusion of parenchymal cells (pneumocytes). Duox1^−/−: The alveolar spaces are infiltrated with inflammatory cells throughout the image. There are also numerous outlines of small open areas (arrows) in the parenchyma (where the normal alveoli used to be). These open areas are lined by cuboidal epithelial cells (type II pneumocyte hyperplasia). Type II pneumocyte hyperplasia signifies a regenerative process unique to the lungs where, after necrosis of type I pneumocytes (flat epithelial cells that normally line the alveoli), type II pneumocytes (cuboidal epithelial cells) proliferate. Hematoxylin and eosin staining, 200× original magnification. (Scale bar, 50 μm.) (D) Representative immunofluorescence staining of cleaved caspase-3 (CC3, red), α-tubulin (green), and DAPI (blue) in uninfected (two Top Rows) and PR8-infected (two Bottom Rows, 3 dpi) WT and Duox1^−/− mouse lungs (n = 5). (E) Comparative statistics of the quantitation of the immunofluorescence signals (shown in D) of PR8-infected Duox1^−/− and WT mice at 3 and 7 dpi. Mean fluorescent intensity (MFI) ± SEM of cleaved caspase-3 (CC3, Left) and of α-tubulin (Right) immunofluorescence is shown (Mann–Whitney U test, n = 5 WT and n = 5 Duox1^−/− mice). (F) Immunoblot analysis of lung lysates collected from Duox1^−/− and WT mice (uninfected or infected with influenza PR8 strain 7 dpi) using specific antibodies detecting α-tubulin, viral NP, cleaved CC3 and full-length caspase-3 (FL-C3). Each column represents a separate animal. Four animals per genotype were infected with influenza, and two animals per genotype were used as uninfected controls. (G) Densitometry data of the immunoblot shown in F. Each dot represents a separate animal. One-way ANOVA, Holm–Šídák’s multiple comparisons test. (H) Indicated densitometric ratios (NP/α-tubulin versus CC3/FLC3) were calculated from both Duox1^−/− and WT animals and correlated with each other based on the results of the immunoblot shown in F. Pearson’s correlation analysis. r, correlation coefficient. (I) TUNEL assay was used to detect apoptosis in the bronchioles of influenza-infected Duox1^−/− and WT mice (7 dpi). One representative result is shown (n = 7). (J) Apoptotic bodies per field were counted in TUNEL-strained bronchioles of Duox1^−/− and WT mice inflected with influenza (n = 7). Mann–Whitney U test. Significance levels are indicated as *P < 0.05 and **P < 0.01. Abbreviations: AU, arbitrary unit; Br, bronchus; CC3, cleaved caspase-3; dpi, day postinfection; Duox1, Dual oxidase 1; FL-C3, full-length caspase-3; INF, infected; MFI, mean fluorescence intensity; NP, viral nucleoprotein; ns, nonsignificant; UI, uninfected; WT, wild-type.

Fig. 4.

OSCN⁻ generated by LPO and DUOX1-derived H₂O₂ reduces influenza virus replication in host cells. (A) DUOX1 and DUOX2 mRNA expressions were measured in differentiated (diff) and undifferentiated (undiff) NHBE cells using real-time PCR. GAPDH was used as reference. Mean ± SD. (B) NHBE cells differentiated on air–liquid interface (ALI) were fixed and processed for immunofluorescence to detect α-tubulin (green) and DUOX1 (red) using specific antibodies. DNA is stained by DAPI (blue). One representative result of three similar ones. (C) ALI cultures of differentiated human NHBE cells were exposed to 5 × 10⁴ PFUs of the indicated influenza virus strains in the presence or absence of LPO, SCN⁻, or catalase in the indicated combinations. After 60 min, infectious virus titers were determined by PFU assay using MDCK cells. Data are expressed as mean ± SEM of independent experiments: A/California/07/2009 (H1N1) (n = 2), A/turkey/KS/4880/1980 (H1N1) (n = 4), A/swine/Illinois/02860/2009 (H1N2) (n = 6), A/Wisconsin/67/2005 (H3N2) (n = 3), and B/Florida/4/2006 Yamagata (n = 3), performed in duplicates. One-way ANOVA and Tukey’s multiple comparisons test. (D) The indicated influenza virus strains were exposed to OSCN⁻ in the cell-free system for 1 h in the presence or absence of catalase, and viral titers were determined by PFU assay on MDCK cells. Data are expressed as mean ± SEM of the following numbers of independent experiments: A/California/07/2009 (H1N1) (n = 5), A/Puerto Rico/8/1934 (H1N1) (n = 4), A/Aichi/2/1968 (H3N2) (n = 4), B/Brisbane/33/2008 Victoria (n = 2), B/Florida/4/2006 Yamagata (n = 5), the NA inhibitor (oseltamivir)-resistant influenza virus strain, A/Mississippi/3/2001 (H1N1) H275Y, and its NA inhibitor-sensitive, parental strain A/Mississippi/3/2001 (H1N1) 275H (n = 2). One-way ANOVA and Tukey’s multiple comparisons tests. (E) The A/California/07/2009 (H1N1) strain was exposed to OSCN⁻ for 1 h in the presence or absence of catalase in the cell-free system and transferred to MDCK cells. Cells were fixed and stained for the viral NP (green) 24 h later. DNA is detected by DAPI. One representative experiment of three similar ones is shown. (F) Growth curve of influenza virus A/Puerto Rico/8/1934-PA-NanoLuciferase (H1N1) (PR8-NanoLuc) in MDCK cells infected with 128 hemagglutination units (HAUs) was measured in a microplate luminometer and expressed as relative luminescence. Presented data are derived from one experiment performed in triplicates. (G) MDCK cells were infected with 128 HAUs of PR8-NanoLuc and exposed to OSCN⁻ before or at the same time of viral infection, as explained in the scheme. Luminescence was measured at 8 h postinfection in a microplate luminometer. Data are expressed as mean ± SEM of n = 4 independent experiments performed in triplicates. One-way ANOVA and Tukey’s multiple comparisons test. (H) Indicated influenza strains were exposed to OSCN⁻ in the cell-free system for 1 h in the presence or absence of catalase, used subsequently to infect MDCK cells, and viral RNA levels were detected by qPCR 6 h postinfection and normalized on host cell GAPDH gene expression. Mean ± S.E.M., n = 3. One-way ANOVA and Tukey’s multiple comparisons test. Significance levels are indicated as *P < 0.05, **P < 0.01, and ***P < 0.001. Abbreviations: ALI, air–liquid interface; hpi, hours postinfection; IAV, influenza A virus; LPO, lactoperoxidase; NHBE, normal human bronchial epithelial cells; NP, nucleoprotein; PFU, plaque-forming unit; RLU, relative luminescence unit; SCN⁻, thiocyanate.

Fig. 5.

LPO-generated OSCN⁻ reduces influenza virus binding and entry into host cells. (A) A modified hemagglutination inhibition assay using human RBCs was performed with the indicated influenza virus strains in vitro in the presence of OSCN⁻, with or without catalase. Data are expressed as mean ± SEM on a log₂ scale of the following numbers of independent experiments: A/California/07/2009 (H1N1) (n = 6), A/Puerto Rico/8/1934 (H1N1) (n = 4), A/Aichi/2/1968 (H3N2) (n = 4), B/Brisbane/33/2008 Victoria (n = 7), and B/Florida/4/2006 Yamagata (n = 6). One-way ANOVA and Tukey’s multiple comparisons tests. (B) Attachment of Alexa 488–labeled A/Aichi/2/1968 (H3N2) influenza viruses to MDCK cells was measured in the presence or absence of OSCN⁻ and catalase as indicated by a fluorescent microplate reader. Data are expressed as a percentage relative to the virus input control and mean ± SEM of n = 5 independent experiments performed in triplicates. One-way ANOVA and Tukey’s multiple comparisons tests. (C) A/Aichi/2/1968 (H3N2) or B/Brisbane/33/2008 Victoria influenza virus strains were exposed to OSCN⁻ (in presence or absence of catalase) for 1 h in vitro before (prebinding) or after (postbinding) incubation with human RBCs. RBC lysis (hemolysis) indicative of virus-induced endosomal membrane fusion was initiated by lowering the pH and measured as a percentage of untreated controls. Data are expressed as mean ± SEM of n = 5 independent experiments performed in triplicates. One-way ANOVA and Tukey’s multiple comparisons tests. (D–F) The indicated influenza virus strains were allowed to adhere to MDCK cells at 4 °C for 1 h. Unbound virions were washed out, and MDCK cells were exposed to OSCN⁻ in the presence or absence of catalase for 1 h at 37 °C as indicated. Cells were fixed and subjected to the detection of viral (D) NP protein (n = 3, 6 hours postinfection [hpi]), (E) M1 matrix protein, or (F) M2 ion channel (n = 3, 3 hpi) by immunofluorescence (green). Host cell DNA was labeled by DAPI (blue). Amantadine was used as a control. Merged images of the viral protein and DAPI-stained nuclei are shown. Images are representative of two or three independent experiments performed in duplicates. (Scale bars, 25 μm.) Significance levels are indicated as *P < 0.05, **P < 0.01, and ***P < 0.001. Abbreviations: HAU, hemagglutination unit; LPO, lactoperoxidase; MDCK, Madin–Darby canine kidney cells; ns, nonsignificant; RBC, red blood cell; RFU, relative fluorescence unit; SCN⁻, thiocyanate.