Murine parainfluenza virus persists in lung innate immune cells sustaining chronic lung pathology

Abstract

Common respiratory viruses, including the human parainfluenza viruses, threaten human health seasonally and associate with the development of chronic lung diseases. Evidence suggests that these viruses can persist, but the sources of viral products in vivo and their impact on chronic respiratory diseases remain unknown. Using the murine parainfluenza virus Sendai, we demonstrate that viral protein and RNA persist in lung macrophages, type 2 innate lymphoid cells (ILC2s) and dendritic cells long after the infectious virus is cleared. Cells containing persistent viral protein expressed Th2 inflammation-related transcriptomic signatures associated with the development of chronic lung diseases, including asthma. Lineage tracing demonstrated that distinct functional groups of cells contribute to the chronic pathology. Importantly, targeted ablation of infected cells significantly ameliorated chronic lung disease. Overall, we identified persistent infection of innate immune cells as a key factor in the progression from acute to chronic lung disease after infection with parainfluenza virus.

Extended Data Fig. 1 |. Gating strategy for spectral flow cytometry analysis and sorting of ILC2s from SeV-infected lungs.

(a) Representative plots indicating the following cell subsets defined from live cells for spectral flow cytometry analysis: Epithelial cells (CD45⁻EpCAM⁺PeCAM⁻), endothelial cells (CD45⁻EpCAM⁻PeCAM⁺), polymorphonuclear cells (PMNs) (CD45⁺Ly6G⁺CD3⁻), T CD4⁺ lymphocytes (CD45⁺Ly6G⁻CD3⁺CD4⁺CD8a⁻), T CD8⁺ lymphocytes (CD45⁺Ly6G⁻CD3⁺CD4⁻CD8a⁺), B cells (CD45⁺Ly6G⁻CD3⁻B220^hi), natural killer cells (NK) (CD45⁺Ly6G⁻CD3⁻B220⁻NK1.1^hi), alveolar macrophages (AMs) (CD45⁺Ly6G⁻CD3⁻F4/80⁺CD11c^hiCD11b^low), tissue macrophages (TMs) (CD45⁺Ly6G⁻CD3⁻F4/80⁺CD11c^lowCD11b^hi), dendritic cells (DCs) (CD45⁺Ly6G⁻CD3⁻F4/80⁻B220⁻NK1.1⁻CD11c^hiCD11b^hi), type 2 innate lymphoid cells (ILC2s) (CD45⁺Ly6G⁻CD3⁻B220⁻NK1.1⁻F4/80⁻CD11c⁻CD11b⁻Thy1.2⁺CD25⁺). (b) Representative plots indicating the gating strategy for ILC2 sorting from sm13 mouse lungs. ILC2s were sorted as CD45⁺Lin⁻Thy1.2⁺huCD4(sm13-IL13)⁺. Lineage markers included mouse CD3, CD4, CD8a, CD11b, Ly6G/Ly6C, NK1.1, CD49b (Pan-NK), CD19, Ter-119, CD11c, TCR gamma/delta, and Siglec-F.

Extended Data Fig. 2 |. Generation of a dual reporter recombinant Sendai virus.

(a) Genome schematics showing the insertion of genes coding for eGFP and the recombinase Cre as independent reading frames in the SeV Cantell genome. EGFP and Cre were inserted between the viral genes N and P, to generate a dual reporter recombinant virus (rSeV-C^eGFP-Cre). (b) Schematic design showing Cre recombination of the LoxP-STOP-tdTomato (tdTom) reporter gene cassette leading to constitutive expression of the tdTom fluorescent protein. (c) Murine embryonic fibroblasts (MEFs) from either WT C57BL/6 mice or tdTom reporter mice were infected in vitro at a multiplicity of infection (MOI) of 0.01 TCID₅₀/cell to test the robustness of the reporter system. (d) Representative images from two independent experiments were taken at 24 hpi using a widefield fluorescence microscope. eGFP signal (green) and tdTom signal (red) were overlaid on brightfield images of WT and tdTom⁻infected MEFs. 20x magnification. Scale bars: 50 μm. (e) tdTom reporter mice were infected with rSeV-C^eGFP-Cre and at days 5 and 49 post-infection lungs were analyzed by immunofluorescence. (f) Mock- and rSeV-C^eGFP-Cre-infected lungs were stained for SeV NP (white staining). Immunofluorescence shows eGFP signal (green), tdTom signal (red), nuclear staining (Hoechst) in blue. Upper panels: Widefield tiled microscopy images, 20x magnification. Scale bars: 500 μm. Lower panels: tiles from the upper panels. Arrowhead indicates an eGFP⁺tdTom⁺SeV NP⁺ cell. Data representative of 3 independent experiments.

Extended Data Fig. 3 |. rSeV-C^eGFP-Cre reaches proximal and distal lung airways and can be used as a robust model to track SeV infection over time.

Whole lungs from tdTom reporter mice intranasally infected with either PBS (Mock) or rSeV-C^eGFP-Cre were harvested at 3 and 49 days post-infection, and imaged using a light sheet microscope, 4x magnification, with a 1.66x magnifier objective, and dual side excitation. 3D reconstruction of infected lungs was done by using the autofluorescence from the 405 nm laser channel (gray) as the tissue scaffold. The eGFP signal is depicted in green, while tdTom signal is depicted in red.

Extended Data Fig. 4 |. Persistent and cleared SeV infection of innate immune cells induces long-lasting activation transcriptomic changes.

(a) B6.Cg-Gt(ROSA)^tdTom (tdTom) mice were inoculated intranasally with 5×10⁵ TCID₅₀ rSeV-C^eGFP-Cre. Lungs were harvested at 49 dpi for cell sorting. ILC2s, macrophages, and dendritic cells (DCs) were individually sorted based on tdTom expression (tdTom⁺ and tdTom⁻), from 6 total infected mice pooled in pairs (S1–3) and subjected to bulk RNAseq. (b) Volcano plots indicating differentially expressed genes (DEGs) in tdTom⁺ILC2s, macrophages, and DCs against their tdTom⁻ counterparts. Individual genes displayed in the gene expression scattered dot plots from Fig. 4 and Figure S4 subpanel D are indicated in the volcano plots. (c) Scattered dot plots showing expression of selected activation, phagocytosis, antigen presentation and type 2 inflammation-related genes from sorted tdTom⁺DCs, in comparison with tdTom⁻DCs. Each dot corresponds to an individual pool of cells. Statistical significance was estimated with non-parametric t tests, corrected using the Holm-Sídak method. P values are indicated. (d) Bubble chart showing gene set enrichment analysis (GSEA) of transcriptomes from tdTom⁺DCs in comparison with tdTom⁻DCs. Bubble size indicates gene set size per GSEA pathway, while bubble color gradient indicates normalized enrichment scores (NES).

Extended Data Fig. 5 |. Diphtheria toxin administration efficiently depletes SeV infected cells.

(a) Experimental design showing diphtheria toxin (DT) regime treatment and timepoints for tissue harvesting and analysis. Representative dot plots of rSeV-C^eGFP-Cre -infected lungs analyzed by flow cytometry for SeV NP expression at (b) 5 dpi and (d) 49 dpi (50,000 events acquired). Quantification of SeV NP⁺ cell frequency (% of live) detected by flow cytometry at (c) 5 dpi and (e) 49 dpi. Individual values, mean ± SD are displayed, data are representative of 2 individual experiments, 4 mice per group. Kruskal-Wallis non-parametric test was used to estimate statistical significance between groups. All box boundaries represent the 25th and 75th percentiles while whiskers represent minimum and maximum values.

Fig. 1 |. Viral antigens and RNA persist in mouse lungs after SeV infection.

a, Mice were inoculated with either phosphate-buffered saline (mock) or 5 × 10⁴ TCID₅₀ of SeV-52. Lungs were analysed on days 3 and 49 post infection. b, Disease progression was monitored by measuring weight loss throughout the experiment. Data are representative of 4 independent experiments (mean ± s.d.). For both groups, the AUC was calculated and two-sided t-tests were performed for statistical significance analysis. c, Whole lung homogenates were collected at 3 and 49 dpi, and SeV NP RNA expression (left) and infectious virus titres (right) were quantified by qPCR and infectivity assays, respectively. Relative RNA quantitation by qPCR was normalized to mouse GAPDH and β-actin. Multiple two-sided Mann–Whitney tests were used to estimate statistical significance of differences between groups. N = 5 animals per group. d, Lungs collected at 49 dpi were sectioned and stained with H&E for analysis of histological changes. Black stars and blue stars indicate bronchioles and blood/lymphatic vessel structures, respectively. Top: tiling images, ×5 magnification. Scale bar, 500 μm. Bottom: digital zooms from the ×5 sections. Scale bars, 100 μm. e, SeV-infected lungs were stained for SeV NP (white staining) using immunofluorescence and for SeV NP RNA using RNAscope (white staining). Nuclear staining (Hoechst for immunofluorescence and DAPI for RNAscope) is shown in blue. Representative images of 3 independent experiments. Scale bars, 100 μm. f, Mock- and SeV-infected lungs were stained for basal stem cells (Krt5⁺, green) and SeV NP (magenta) to localize SeV NP⁺ cells in relation to areas displaying chronic lesions (dashed areas, subpanels f1–f3) and unaffected areas (subpanel f4). Arrowheads, SeV NP⁺ cells. Widefield microscopy. Left: tiling image, ×5 magnification. Right: ×20 magnification. Right insets: digital zooms from the ×20 magnification images. Scale bars, 500 μm (left panel), 100 μm (subpanels). Images are representative of 3 independent experiments, 5 mice per group. g, Quantification of Krt5-stained area (%) over total lung section area. Mean ± s.d. are shown. Data summarizes 3 individual experiments, 4–7 mice per condition. Statistical significance was estimated using two-sided non-parametric Mann–Whitney test.

Fig. 2 |. Diverse lung immune cells express SeV NP during chronic infection.

a, Characterization by immunofluorescence of immune cells expressing SeV NP from cryopreserved mouse lungs at 49 dpi. Tissue sections were stained for SeV NP (magenta) in combination with the surface markers CD3 (T lymphocytes), CD11c and CD11b (dendritic cell subsets), F4/80 (macrophages) or Thy1.2 (innate lymphoid cells and some T lymphocyte subsets) in the indicated colours. Nuclear staining is displayed in blue. White arrows indicate individual SeV NP⁺ cells. Images were taken in a widefield fluorescence microscope using ×20 magnification. Scale bars, 25 μm. Representative images from 3 independent experiments, 5 mice per condition. Right: insets from the dashed areas. b–e, Lungs from SeV-infected mice were collected at 3 and 49 dpi, enzymatically digested and analysed by multiplex spectral flow cytometry with a panel of 16 antibodies to quantify (b,c) and characterize (d,e) SeV NP⁺ cells. b, Representative dot plots of SeV NP⁺ cells (% of live) comparing acute (3 dpi) with long-term (49 dpi) SeV infection. SeV NP⁺ gates were drawn on the basis of the isotype control and the mock-infected samples. c, Frequency of SeV NP⁺ cells gated on total live cells from SeV 3 days-, SeV 49 days- and mock-infected lungs. Data shown as mean ± s.d. d, Representative histograms of SeV NP⁺ fluorescence intensity from 9 individual cell subsets: B cells, ILC2s, T CD4⁺ lymphocytes, T CD8⁺ lymphocytes, NK cells, polymorphonuclear cells (PMNs), alveolar macrophages (AMs), tissue macrophages (TMs) and dendritic cells (DCs) at 49 dpi. Histograms from SeV-infected animals are displayed in red, while histograms from mock-infected animals are displayed in grey. e, Frequency and mean fluorescence intensity (MFI) of SeV NP⁺ cells within lymphoid- and myeloidorigin cell subsets. Means ± s.d. are shown. All multiple comparisons were done with two-sided non-parametric Kruskal–Wallis with Dunn’s post test. Data representative of 2 independent experiments, 4 animals per condition, total of 1 million events acquired per animal.

Fig. 3 |. ILC2s and macrophages are activated during SeV chronic lung disease.

Lung type 2 innate lymphoid cells (ILC2s) (a–c) and macrophages (d–f) were isolated either from SeV- or mock-infected mouse lungs on day 49 post infection and subjected to bulk RNA-seq. a, Volcano plot indicating differentially expressed genes detected in ILC2s from SeV-infected lungs over mock. P < 0.05, logFC > 2. b,c, Scattered dot plots showing expression of ILC2 hallmark genes (b) and virus-related interferon-stimulated genes (ISGs) (c). Each dot corresponds to an individual pool of cells (n = 6 animals pooled in pairs per condition). d, Volcano plot indicating differentially expressed genes detected in macrophages from SeV-infected lungs over mock. P < 0.05, logFC > 2. e,f, Scattered dot plots indicating expression of Th2 polarization (e) and phagocytic activity (f) genes from macrophages. Each dot corresponds to cells obtained from an individual animal (minimum of n = 3 animals per condition). Means ± s.d. are displayed. Multiple two-sided non-parametric t-tests corrected using the Holm–Sídak method were used to estimate statistical significance of differences between conditions. CPM, copies per million. g, Bubble chart showing GSEA of transcripts in ILC2s and macrophages sorted from SeV 49 dpi lungs. Bubble size indicates gene set size per GSEA pathway, while bubble colour gradient indicates normalized enrichment score (NES) values.

Fig. 4 |. Persistent imprinting of lung cells after parainfluenza virus infection.

a, B6.Cg-Gt(ROSA)^tdTom (tdTom) mice were inoculated intranasally with either PBS (mock) or 5 × 10⁵ TCID₅₀ rSeV-C^eGFP-Cre. Lungs were collected at 3 or 49 dpi for flow cytometry analysis, cell sorting and imaging. b, Weight loss was recorded up to 21 dpi to monitor disease progression. Data (mean ± s.d.) are representative of 2 individual experiments (minimum of 3 mice per group). c, 3D lung reconstruction of terminal bronchioles from mock- and rSeV-C^eGFP-Cre-infected (3 and 49 dpi) mouse lungs. Autofluorescence (grey) from the 405 nm channel was used to mount the tissue scaffold. eGFP fluorescence is indicated in green. tdTom fluorescence is indicated in red. d, Higher magnification images from the areas displayed in c. e, Representative dot plots comparing percentage of tdTom⁺ cells in mouse lungs during acute and chronic rSeV-C^eGFP-Cre infection. f, Frequency of tdTom⁺ cells gated on total live cells from rSeV-C^eGFP-Cre 3 days-, rSeV-C^eGFP-Cre 49 days- and mock-infected lungs. Data are representative of 3 independent experiments, 3–4 mice per group, shown as mean ± s.d. Statistical significance was estimated using two-sided Kruskal–Wallis non-parametric test. g,h, Characterization of immune (CD45⁺) and non-immune (CD45⁻) cell proportions within tdTom⁺ cells. Representative dot plots (g) and quantification (h) of CD45 staining in tdTom⁺ cells during acute and chronic rSeV-C^eGFP-Cre infection. Statistical significance was estimated using two-sided non-parametric Mann–Whitney tests, corrected using the Holm–Sídak method. Data from 3 independent experiments, 3–4 mice per condition are shown as mean ± s.d. Box boundaries represent the 25th and 75th percentiles, while whiskers represent minimum and maximum values. i, Scattered dot plots showing expression of selected activation and type 2 inflammation-related genes from sorted tdTom⁺ ILC2s and macrophages, in comparison with each cell type tdTom⁻ counterpart. Each dot corresponds to an individual pool of cells (n = 6 animals pooled in pairs). Means ± s.d. are indicated. Statistical significance was estimated using non-parametric two-sided t-tests, corrected using the Holm–Sídak method. P values are displayed. j, Bubble chart showing GSEA of transcriptomes from tdTom⁺ in comparison with tdTom⁻ ILC2s and macrophages sorted from rSeV-C^eGFP–Cre 49 dpi lungs. Bubble size indicates gene set size per GSEA pathway, while colour gradient indicates NES.

Fig. 5 |. Transcriptional programmes in long-term SeV-infected cell populations.

Combination of tdTom and SeV NP detection enables sorting of two distinct subsets of viral infected cells: SeV-infected cells persistently expressing viral antigens (tdTom⁺NP⁺) and SeV-infected/survivor cells only (tdTom⁺NP⁻). a, Representative dot plots indicating the gating strategy used for sorting 3 cell subpopulations from rSeV-C^eGFP–Cre-infected lungs: tdTom⁺NP⁻, tdTom⁺NP⁺ and negative. Negative cells from mock-infected lungs were also sorted. Data representative of 2 individual experiments, 6 mice per condition. Viral infected cells and cells derived from them sorted from tdTom mice infected with rSeV-C^eGFP–Cre at 49 dpi were subjected to bulk RNA-seq and host transcriptome analysis. b, Volcano plots indicating DEGs in tdTom⁺NP⁻, tdTom⁺NP⁺ and negative cells against mock negative cells (P < 0.05 and logFC > 2). c, Venn diagram showing overlapping DEGs from tdTom⁺NP⁻ and tdTom⁺NP⁺, as well as exclusive DEGs from each cell subset. d,e, Bar graphs showing GO enrichment analysis of each of the cell subsets’ (tdTom⁺NP⁻ (d) and tdTom⁺NP⁺ (e)) exclusive DEGs. f, Heat maps of selected gene collections from the GO pathways in d and e. FC values of tdTom⁺NP⁻ and tdTom⁺NP⁻ are displayed. g, GSEA bubble chart indicating the most significant enriched pathways of tdTom⁺NP⁺ transcriptome signatures in comparison with tdTom⁺NP⁻ cells. Bubble colour gradient indicates NES and bubble sizes correspond to gene set size for each pathway.

Fig. 6 |. Surviving and persistently infected cells impact chronic lung pathology.

a, Diphtheria toxin regime treatment and timepoints for tissue collection and analysis. b, Disease progression was assessed by monitoring animal weight loss until 26 dpi. Graphs are representative of 3 independent experiments and depict mean ± s.d. weight loss values, minimum of 4 mice per condition. c, Mouse lungs were collected at 49 dpi and tissue sections were stained with H&E to compare pathological changes between the analysis groups. Top: representative images of whole lung sections (brightfield, ×20 magnification tiled images; scale bars, 500 μm). Bottom: zoomed-in images from the indicated areas, 5 animals per condition. d, Lung sections from SeV +DT and SeV −DT groups were blindly scored for histopathological changes. Total area affected, percentage of airway structures affected, and intensity of alveolitis, peribroncholitis, vasculitis and BALT expansion were determined for every individual lung sample. Individual weighted scores ± s.d. are indicated. Data represent 3 pooled individual experiments, 3–4 mice per condition. e, Lung sections were stained for the tissue remodelling and chronic lung lesion markers Krt5 (green) and Krt8 (magenta) with immunofluorescence to check for chronic lung lesion progression. Nuclear staining is displayed in blue. Dashed areas indicate chronic lung lesions and areas of intense tissue remodelling. Images were taken with a widefield microscope; tiling images, ×20 magnification; scale bars, 500 μm. f, Quantification of chronic lung lesion area (%) over total lung section area. Data represent 3 pooled individual experiments, 3–5 mice per condition. Statistical significance was estimated using non-parametric two-sided Mann–Whitney test. All box boundaries represent the 25th and 75th percentiles, while whiskers represent minimum and maximum values.