GM-CSF in the lung protects against lethal influenza infection

Abstract

Rationale: Alveolar macrophages contribute to host defenses against influenza in animal models. Enhancing alveolar macrophage function may contribute to protection against influenza.

Objectives: To determine if increased expression of granulocyte/macrophage colony-stimulating factor (GM-CSF) in the lung increases resistance to influenza.

Methods: Wild-type mice and transgenic mice that expressed GM-CSF in the lung were infected with influenza virus, and lung pathology, weight loss, and mortality were measured. We also administered GM-CSF to the lungs of wild-type mice that were infected with influenza virus.

Measurements and main results: Wild-type mice all died after infection with different strains of influenza virus, but all transgenic mice expressing GM-CSF in the lungs survived. The latter also had greatly reduced weight loss and lung injury, and showed histologic evidence of a rapid host inflammatory response that controlled infection. The resistance of transgenic mice to influenza was abrogated by elimination of alveolar phagocytes, but not by depletion of T cells, B cells, or neutrophils. Transgenic mice had far more alveolar macrophages than did wild-type mice, and they were more resistant to influenza-induced apoptosis. Delivery of intranasal GM-CSF to wild-type mice also conferred resistance to influenza.

Conclusions: GM-CSF confers resistance to influenza by enhancing innate immune mechanisms that depend on alveolar macrophages. Pulmonary delivery of this cytokine has the potential to reduce the morbidity and mortality due to influenza virus.

Figure 1.

Pulmonary granulocyte/macrophage colony-stimulating factor (GM-CSF) expression protects against influenza. (A) Wild-type (WT), GM-CSF–deficient (GM^−/−), and SPC-GM mice that overexpress GM-CSF by alveolar epithelial type II cells and were generated from GM^−/− mice on a C57BL/6 background, by expressing a chimeric gene containing GM-CSF under the control of the human SP-C promoter (n = 10 per group, result representative of three experiments) were infected with 5 LD₅₀ of influenza A virus PR8. WT and SPC-GM mice were infected with lethal doses of H3N2 HK68 (n = 5 per group) or mouse-adapted H1N1 swine influenza (n = 4 per group). All mice were followed until death or recovery. (B) WT, GM−/−, and SPC-GM mice (n = 8 per group) were infected with 5 LD₅₀ of PR8, and weighed daily until death or recovery. Means ± SE are shown. (C) WT mice were given a polyethylenimine-coated GM-CSF expression vector or a control empty vector (n = 7–8 per group) by retro-orbital injection. Four to 6 weeks later, mice were infected with 2 LD₅₀ of PR8, and followed until death or recovery (P = 0.0008, comparing treatment with the GM-CSF–expressing vector and empty vector). In an independent experiment, mice were treated with PBS (n = 5 per group) or rGM-CSF intranasally for 7 days (n = 6 per group), then infected 1 day later with 2 LD₅₀ of PR8 (P = 0.008, comparing rGM-CSF–treated and PBS-treated mice). Results are representative of two to five experiments.

Figure 2.

Pulmonary GM-CSF expression reduces lung injury and viral burden, and elicits early inflammation. (A) SPC-GM and WT mice (n = 20 per group) were infected with 5 LD₅₀ of influenza virus PR8. Three to 6 days after infection (DPI), four mice in each group were killed daily, and albumin levels in bronchoalveolar lavage fluid were measured by enzyme-linked immunoassay. Means ± SEs are shown. *P = 0.01, comparing SPC-GM and WT mice. (B) SPC-GM and WT mice (n = 12 per group) were infected with 5 LD₅₀ of PR8. One, 3, and 6 days after infection, four mice in each group were killed, and viral loads were determined by measuring the TCID₅₀, as outlined in the methods. *P = 0.03, comparing SPC-GM and WT mice. (C) Histology of the lungs of SPC-GM and WT mice, both uninfected and 1–6 days after infection with PR8. Representative sections, stained with hematoxylin and eosin, are shown, at ×40 magnification. (D) Cytokine levels were measured by enzyme-linked immunoassay in lung homogenates from SPC-GM and WT mice, before and after infection with PR8 (n = 3 per time point). Means ± SEs are shown. *P < 0.01, comparing SPC-GM and WT mice.

Figure 3.

Expression of CD200R in bronchoalveolar lavage cells. WT mice were treated with 1.34 mg/kg of recombinant murine GM-CSF or with PBS for 7 days, prior to infection with PR8 H1N1 influenza. Mice were killed (n = 3 per time point), and bronchoalveolar lavage cells were stained with anti-CD200R. Mean values and SEs for the percentages of CD200R⁺ cells are shown. *P < 0.001.

Figure 4.

Resistance to influenza of SPC-GM mice does not require neutrophils, T cells, or B cells. (A) SPC-GM mice were treated intraperitoneally with anti-GR1 or isotype control IgG (n = 7 per group) at Days −1, 0, 3, and 6 after infection with 5 LD₅₀ of influenza virus PR8, and followed until death or recovery. Anti-GR1 reduced the percentages of neutrophils from 15% to 1% in blood. A representative result of two experiments is shown. (B) CD4⁺ and/or CD8⁺ T cells were depleted with monoclonal antibodies (n = 5–7 per group), at −3, 0, and 3 days after infection with 5 LD₅₀ of PR8, and followed until death or recovery. Monoclonal antibodies reduced the percentages of CD4⁺ and CD8⁺ cells in the mediastinal lymph nodes from 45% to 0.2%, and from 24% to 2%, respectively. Survival rates did not differ significantly in all groups. (C) SPC-GM mice were treated intraperitoneally with anti-CD90.2 or isotype control IgG (n = 7 per group) at Days −1, 0, 3, and 6 after infection. Mice were infected with 5 LD50 of influenza virus PR8 on Day 0 and followed until death or recovery. (D) SPC-GM mice were treated with N-acetyl-γ-calicheamicin dimethylhydrazide, conjugated either to anti-CD22 or isotype control antibody (n = 5 per group) at Days −5 and 0 after infection with 5 LD₅₀ of PR8, and followed until death or recovery. A representative result of two experiments is shown.

Figure 5.

Alveolar macrophages (AM) are required for resistance of SPC-GM mice to influenza. (A) SPC-GM mice were given clodronate-liposomes or PBS-liposomes (n = 5 per group). Eighteen to 24 hours later, mice were infected with 5 LD₅₀ of influenza virus PR8, and followed until death or recovery. (B) SPC-GM mice were given clodronate-liposomes or PBS-liposomes 1–6 days after infection with 5 LD₅₀ of PR8 (n = 3–5 per group), and followed until death or recovery. (C) SPC-GM mice were treated with clodronate-liposome to deplete alveolar phagocytes. Three days later, bronchoalveolar lavage cells (99% AM) from naïve SPC-GM mice were collected. Clodronate-treated SPC-GM mice each received 2 × 10⁶ of these AM or PBS intratracheally, and were infected with PR8 16 hours later. A representative of two experiments with identical results is shown. (D) WT mice (n = 5 per group) received 2 × 10⁶ bronchoalveolar lavage cells (99% AM) from naïve SPC-GM mice or PBS intratracheally, and were infected with PR8 16 hours later. A representative result of two experiments is shown.

Figure 6.

AM of SPC-GM mice are more resistant to apoptosis than those of WT mice. SPC-GM and WT mice were infected with 5 LD50 of influenza virus PR8. Bronchoalveolar lavage cells were collected at 0–3 days after infection (DPI) and stained with anti-F4/80 and annexin V or anti-Fas. (A) Mean values and SEs for the percentages of Annexin V⁺ cells (n = 5–6 per time point) are shown. *P < 0.05, comparing SPC-GM and WT mice. (B) We gated on F4/80⁺ cells and measured the mean fluorescence intensity (MFI) of Fas. Mean values and SEs for the net MFI of Fas (n = 5–6 per time point) are shown. *P < 0.05, **P < 0.01, comparing SPC-GM and WT mice. (C) In vitro analysis of apoptosis. AM from naïve SPC-GM and WT mice (n = 4 per group) were incubated with influenza virus PR8, as detailed in Methods. Cells were stained with anti-F4/80, followed by FITC–Annexin V, and analyzed by flow cytometry. A representative result of two experiments is shown.

Figure 7.

Phagocytic capacity of AM from SPC-GM and WT mice. (A) Mice were intranasally inoculated with FITC-labeled PR8 (n = 2), FITC stock in PBS (n = 1), or PBS alone (n = 1). Two hours later, AM from bronchoalveolar lavage were stained with allophycocyanin anti-F4/80, and subjected to flow cytometry. We show the MFI of FITC-labeled PR8 in F4/80⁺ cells, and the phagocytic index (PI), which is the percentage of FITC⁺ F4/80 cells multiplied by the MFI of FITC-labeled PR8 in F4/80⁺ cells. (B) AM from SPC-GM and WT mice (n = 3 per group) were cultured with yellow-green FluoSphere Carboxylate-Modified beads. Thirty minutes later, AMs were stained with anti-F4/80 and analyzed by flow cytometry. MFI and PI were calculated, as in Panel A.