TRAIL+ monocytes and monocyte-related cells cause lung damage and thereby increase susceptibility to influenza-Streptococcus pneumoniae coinfection

Abstract

Streptococcus pneumoniae coinfection is a major cause of influenza-associated mortality; however, the mechanisms underlying pathogenesis or protection remain unclear. Using a clinically relevant mouse model, we identify immune-mediated damage early during coinfection as a new mechanism causing susceptibility. Coinfected CCR2(-/-) mice lacking monocytes and monocyte-derived cells control bacterial invasion better, show reduced epithelial damage and are overall more resistant than wild-type controls. In influenza-infected wild-type lungs, monocytes and monocyte-derived cells are the major cell populations expressing the apoptosis-inducing ligand TRAIL. Accordingly, anti-TRAIL treatment reduces bacterial load and protects against coinfection if administered during viral infection, but not following bacterial exposure. Post-influenza bacterial outgrowth induces a strong proinflammatory cytokine response and massive inflammatory cell infiltrate. Depletion of neutrophils or blockade of TNF-α facilitate bacterial outgrowth, leading to increased mortality, demonstrating that these factors aid bacterial control. We conclude that inflammatory monocytes recruited early, during the viral phase of coinfection, induce TRAIL-mediated lung damage, which facilitates bacterial invasion, while TNF-α and neutrophil responses help control subsequent bacterial outgrowth. We thus identify novel determinants of protection versus pathology in influenza-Streptococcus pneumoniae coinfection.

Keywords: C‐C‐chemokine receptor type (CCR) 2; Streptococcus pneumoniae; TNF‐related apoptosis‐inducing ligand (TRAIL); influenza; neutrophil.

Figure 1

Influenza A predisposes mice to Streptococcus pneumoniae coinfection

1. A–C Mortality (A), weights (B) and clinical scores (C) following intranasal infection of C57BL/6 mice with 400 TCID₅₀/30 μl IAV X31, 2 × 10⁵ CFU/30 μl S. pneumoniae D39 or mock (PBS) (data shown are pooled from five independent experiments, n = 6–12; this dosing regimen hereafter referred to as “low dose”; for clarity, means shown include euthanized mice at endpoint weight or clinical score).

2. D, E Pneumococcal load in the lung (D) from 6 to 8 dpi and in the spleen (E) at 8 dpi during low-dose coinfection (6–7 dpi data shown are pooled from four independent experiments, 8 dpi data (after dashed line) are pooled from eight independent experiments, and dotted line indicates detection limit, n = 3–10).

3. F Mortality following coinfection with high (8 × 10³ TCID₅₀ IAV, 2 × 10⁷ CFU S. pneumoniae) or low (as above) coinfection doses (n = 9).

4. G Comparison of lung pneumococcal load in mice harvested upon reaching endpoint or concurrently harvested recovering (gaining weight) mice, at low and high dose from 8 to 10 dpi. All mice at high dose reached endpoint; all low-dose mice are grouped (left panels) and then separated into recovering and endpoint groups (right panels) (dotted line indicates detection limit, n = 13–21).

5. H, I Quantification of inflammatory monocytes (H) or neutrophils (I) at 7 dpi by flow cytometry during low-dose coinfection (data shown are pooled from two independent experiments; n = 2–6).

6. J Multiplex quantification of TNF-α, KC and MIP2 in the airways at 7 dpi during low-dose coinfection (data shown are pooled from two independent experiments, n = 2–6).

Data information: Data are displayed as percentage survival (mortality), geometric means (pneumococcal loads) or arithmetic means ± SEM (weights, clinical scores, cells and cytokines). Significance was assessed by log-rank (Mantel–Cox) test (mortality), two-way ANOVA (weights and clinical scores) or Mann–Whitney U-test (pneumococcal loads, cells and cytokines).

Mortality in coinfection is linked to outgrowth of live bacteria

1. A–C Mortality (A), weights (B) and clinical scores (C) following infection with 8 × 10³ TCID₅₀/30 μl IAV X31, 2 × 10⁷ CFU/30 μl S. pneumoniae D39 or mock (PBS) (data shown are pooled from four independent experiments, n = 6–9; this dosing regimen hereafter referred to as “high dose”; for clarity, means shown include euthanized mice at endpoint weight or clinical score).

2. D, E Pneumococcal load in the lung (D) from 6 to 8 dpi and in the spleen (E) at 8 dpi during low-dose coinfection (dotted line indicates detection limit, n = 4–5).

3. F Quantitative PCR for influenza matrix RNA in the lung during high-dose coinfection from 6 to 8 dpi (n = 5).

4. G Quantitative PCR for influenza matrix RNA in the spleen at 8 dpi during high-dose coinfection, compared to influenza-infected positive control lung (6 dpi 8 × 10³ TCID₅₀ IAV) (n = 5).

5. H Mortality following high-dose coinfection with live S. pneumoniae, heat-killed S. pneumoniae or Pam3CSK4 (15 μg) (representative of two independent experiments, n = 6–9).

6. I Pneumococcal load in the airways early during high-dose coinfection from 5 dpi + 4 h to 5 dpi + 16 h (dotted line indicates detection limit, n = 5).

7. J Quantitative PCR for pneumococcal 16 s rRNA in the lung during high-dose coinfection from 5 dpi+4 h to 5 dpi + 48 h (n = 10).

Data information: Data are displayed as percentage survival (mortality), geometric means (viral and bacterial loads, bacterial RNA) or arithmetic means ± SEM (weights and clinical scores). Significance was assessed by Mann–Whitney U-test (viral and bacterial loads, bacterial RNA), two-way ANOVA (weights and clinical scores) or log-rank (Mantel–Cox) test (mortality). n.s. = not significant.

Quantification of cells and cytokines during high-dose coinfection

1. A, B Quantification of inflammatory monocytes (A) and neutrophils (B) during high-dose coinfection at 7 dpi by flow cytometry (data shown are pooled from two independent experiments, n = 2–3).

2. C Quantification of CD4 T cells (CD3⁺CD4⁺), CD8 T cells (CD3⁺CD8⁺), NK cells (CD3⁻CD4⁻CD8⁻NKp46⁺) and alveolar macrophages during high-dose coinfection at 7 dpi (n = 2–5).

3. D H&E staining of lung tissue sections at 8 dpi during high-dose coinfection. Scale bar indicates 200 μm (n = 2–3).

4. E Multiplex quantification of TNF-α, KC and MIP2 in the airways at 7 dpi during high-dose coinfection (n = 2–6).

Data information: Data are displayed as arithmetic means ± SEM. Significance was assessed by Mann–Whitney U-test (viral and bacterial loads, bacterial RNA). n.s. = not significant.

Figure 2

Mice deficient in inflammatory monocytes and related myeloid populations are resistant to coinfection and have reduced early lung damage

1. A Mortality of CCR2^−/− and wild-type (WT) mice during low-dose coinfection (data shown are pooled from two independent experiments, n = 6–9).

2. B, C Pneumococcal load in the lung (B) and spleen (C) at 8 dpi during low-dose coinfection in CCR2^−/− and wild-type (WT) mice (data shown are pooled from two independent experiments; dotted line indicates detection limit, n = 4–9).

3. D Quantitative PCR for influenza matrix RNA in the lung during low-dose coinfection at 8 dpi in CCR2^−/− and wild-type (WT) mice (n = 5–9).

4. E, F Airway protein (E) and airway LDH activity relative to wild-type IAV-infected group mean (defined as 100%) (F) at 5 dpi in CCR2^−/− and wild-type (WT) mice (8 × 10³ TCID₅₀) (data shown are pooled from three independent experiments, n = 2–3).

5. G H&E staining of lung tissue sections at 5 dpi in CCR2^−/− and wild-type (WT) mice (8 × 10³ TCID₅₀). Scale bar indicates 200 μm (n = 2–3).

6. H Quantification of lung cells by flow cytometry at 5 dpi in CCR2^−/− and wild-type (WT) mice (8 × 10³ TCID₅₀); AM = alveolar macrophages (Siglec F⁺CD11c⁺CD64⁺Ly6C⁻), IM = inflammatory monocytes (Siglec F⁻CD11b⁺MHCII⁻Ly6C⁺Ly6G⁻CD64⁺), Mono d. DC = monocyte-derived dendritic cells (Siglec F⁻CD11b⁺MHCII⁺CD11c⁺CD64⁺Ly6C⁺Ly6G⁻), Inter. Mac = interstitial macrophages (Siglec F⁻CD11b⁺MHCII⁺CD11c⁻CD64⁺Ly6C⁺), CD103⁺ DC = CD103⁺ dendritic cells (CD103⁺CD3⁻CD11c⁺CD24⁺Siglec F⁻CD11b⁺Ly6G⁻CD64⁻), pDC = plasmacytoid dendritic cells (PDCA-1⁺Ly6C⁺CD11c^intCD11b⁻Siglec F⁻Ly6G⁻) (data shown are pooled from three independent experiments, n = 3–4).

7. I Quantification of lung inflammatory monocytes by flow cytometry at 7 dpi in CCR2^−/− and wild-type (WT) mice during high-dose coinfection (data shown are pooled from two independent experiments, n = 3).

8. J Quantification of blood inflammatory monocytes (as proportion of live cells) by flow cytometry at 5 dpi in CCR2^−/− and wild-type (WT) mice (8 × 10³ TCID₅₀) (n = 3).

Data information: Data are displayed as percentage survival (mortality), geometric means (viral and pneumococcal loads) or arithmetic means ± SEM (damage and cells). Significance was assessed by Mann–Whitney U-test (viral and pneumococcal loads, damage and cells) or log-rank (Mantel–Cox) test (mortality). n.s. = not significant.

Myeloid flow cytometry gating strategy Flow cytometry gating strategy used for myeloid cells (neutrophils, alveolar macrophages, inflammatory monocytes and inflammatory monocyte-derived cells); pre-gated for live cells (Death Stain⁻ and exclusion of debris by size). Representative 5 dpi wild-type non-lavaged whole lung shown (8 × 10³ TCID₅₀) (n = 4).

Monocytes and monocyte-related cells express TRAIL and are absent in CCR2^−/− mice

1. A–F CCR2 and TRAIL expression in different myeloid cell populations as assessed by flow cytometry; pre-gated for live cells (Death Stain⁻ and exclusion of debris by size). Representative 5 dpi wild-type non-lavaged whole lung shown (8 × 10³ TCID₅₀) (n = 4; note this lung is the same as shown in FigEV3).

2. G–L CCR2 and TRAIL expression in different myeloid cell populations as assessed by flow cytometry. Representative 5 dpi non-lavaged whole CCR2^−/− lung shown (8 × 10³ TCID₅₀) (n = 4).

Data information: Gating of each myeloid population is shown in the left panels, and histograms of CCR2 and TRAIL expression are shown on the right. Grey histogram represents unstained, and coloured histogram represents stained.

Figure 3

Blockade of TRAIL ameliorates coinfection

1. A Quantification of TRAIL⁺ lung cells by flow cytometry at 5 dpi in CCR2^−/− and wild-type (WT) mice (8 × 10³ TCID₅₀); abbreviations as in Fig2H (n = 4).

2. B Quantification of the DR5⁺ proportion of lung epithelial cells (E-cadherin⁺Ep-Cam⁺) by flow cytometry at 5 dpi in CCR2^−/− and wild-type (WT) mice (8 × 10³ TCID₅₀) (n = 5).

3. C Mortality during low-dose coinfection following treatment with anti-TRAIL or vehicle control (PBS) every 48 h from 1 to 9 dpi (in wild-type mice) (data shown are pooled from two independent experiments, n = 6–9).

4. D Mortality during low-dose coinfection following treatment with anti-TRAIL at 1 and 3 dpi (early), 6 and 8 dpi (late) or vehicle control (PBS) at 1, 3, 6 and 8 dpi (data shown are pooled from two independent experiments, n = 8–9).

5. E Lung pneumococcal load at 5 dpi + 16 h during high-dose coinfection, following treatment with anti-TRAIL at 1 and 3 dpi or vehicle (PBS) (dotted line indicates detection limit, n = 7–9).

6. F Quantitative PCR for influenza matrix RNA in the lung during low-dose coinfection at 8 dpi following treatment with anti-TRAIL at 1 and 3 dpi (early), 6 dpi (late) or vehicle control (PBS) at 1, 3 and 6 dpi (n = 5–10).

7. G, H Airway protein (G) and airway LDH activity relative to wild-type IAV-infected group mean (defined as 100%) (H) at 5 dpi (8 × 10³ TCID₅₀) following treatment with anti-TRAIL or vehicle (PBS) at 1 and 3 dpi (n = 3–6).

8. I Quantification of lung inflammatory monocytes by flow cytometry at 5 dpi (8 × 10³ TCID₅₀) following anti-TRAIL treatment at 1 and 3 dpi (n = 3–6).

Data information: Data are displayed as mortality (survival), geometric mean (viral and bacterial loads) or arithmetic means ± SEM (damage and cells). Significance was assessed by log-rank (Mantel–Cox) test (mortality) or Mann–Whitney U-test (viral and bacterial loads, damage and cells). n.s. = not significant.

Figure 4

Neutrophils from coinfected mice are functional and contribute to survival and bacterial control

1. ROS production was assessed by luminol assay of PDBu (50 nM)-stimulated lung neutrophils purified by MACS from high-dose-coinfected or S. pneumoniae-infected mice at 6 dpi (neutrophils from nine mice pooled into three replicates/group).

2. ELISA quantification of TNF-α and KC produced by Pam3CSK4 (1 μg/ml)-stimulated neutrophils purified by MACS from high-dose-coinfected or S. pneumoniae-infected mice at 6 dpi (neutrophils from three mice/group).

3. Percentage of NET-forming cells was assessed by microscopy of C. albicans-stimulated (5 × 10⁵ CFU) neutrophils purified by MACS and Percoll gradient from high-dose-coinfected or S. pneumoniae-infected mice at 6 dpi (neutrophils from three mice pooled/group).

4. ELISA quantification of airway myeloperoxidase at 6 dpi during high-dose coinfection (n = 3–5).

5. Quantification of lung neutrophils (without staining for Ly6G—CD11b⁺SSC^>lowCD11c⁻MHCII⁻Ly6C^lowF4/80⁻) by flow cytometry at 6 and 7 dpi during low-dose coinfection following treatment with anti-Ly6G or vehicle control every 24 h from 4 dpi (n = 3–5).

6. Mortality during low-dose coinfection following treatment with anti-Ly6G or isotype control every 24 h from 4 to 12 dpi (data shown are pooled from two independent experiments, n = 9).

7. Lung pneumococcal load at 8 dpi during low-dose coinfection following treatment with anti-Ly6G or isotype control (data shown are pooled from two independent experiments; dotted line indicates detection limit, n = 5–10).

8. Quantitative PCR for influenza matrix RNA in the lung during low-dose coinfection at 8 dpi following treatment with anti-Ly6G or isotype control every 24 h from 4 to 7 dpi (n = 4–10).

Data information: Data are displayed as arithmetic means ± SEM (ROS, cytokine production, myeloperoxidase and neutrophil numbers), percentage of neutrophils (NET formation), percentage survival (mortality) or geometric means (viral and pneumococcal loads). Significance was assessed by Mann–Whitney U-test (myeloperoxidase, neutrophil numbers, viral and pneumococcal loads) or log-rank (Mantel–Cox) test (mortality). n.s. = not significant.

Neutrophils phagocytose streptococci but do not produce NETs during coinfection

1. Confocal microscopy of lung tissue sections at 8 dpi during high-dose coinfection stained for cell nuclei (DAPI), streptococci (αStrep) and neutrophils (αMPO). Black text indicates infection condition; coloured text indicates staining. Right column shows the Z-projection of 10 individual focal planes; other columns show a single plane. Arrows indicate bacteria phagocytosed by neutrophils. Scale bars indicate 5 μm (n = 3).

2. Confocal microscopy of lung tissue sections at 8 dpi during high-dose coinfection (or during C. albicans infection as positive control) stained for cell nuclei (DAPI), neutrophils (αMPO) or the NET constituent citrullinated histone H3 (αcitH3). Black text indicates infection condition; coloured text indicates staining. Scale bars indicate 20 μm (n = 3).

Figure 5

TNF-α contributes to survival and bacterial control in coinfected mice

1. Mortality during low-dose coinfection following treatment with anti-TNF-α or isotype control at 5 and 7 dpi (n = 6–9).

2. Pneumococcal load in the lung at 8 dpi during low-dose coinfection following anti-TNF-α or isotype control treatment (data shown are pooled from two independent experiments; dotted line indicates detection limit, n = 3–9).

3. Quantitative PCR for influenza matrix RNA in the lung during low dose at 7 dpi following treatment with anti-TNF-α or vehicle control at 5 and 7 dpi (n = 5–6).

Data information: Data are displayed as percentage survival (mortality) or geometric means (viral and pneumococcal loads). Significance was assessed by Mann–Whitney U-test (viral and pneumococcal loads) or log-rank (Mantel–Cox) test (mortality).

Figure 6

Summary of coinfection pathogenesis Influenza infection leads CCR2-dependent recruitment of monocytes and other monocyte-related immune cell populations into the lung. These cells express TRAIL and induce epithelial apoptosis causing lung damage, thus facilitating bacterial invasion due to breakdown of barrier function. The subsequent bacterial outgrowth drives strong TNF-α and neutrophil responses which are on balance protective as they contribute to the (limited) control of bacteria.