Leukemia inhibitory factor signaling is required for lung protection during pneumonia

Abstract

Lung infections represent a tremendous disease burden and a leading cause of acute lung injury. STAT3 signaling is essential for controlling lung injury during pneumonia. We previously identified LIF as a prominent STAT3-activating cytokine expressed in the airspaces of pneumonic lungs, but its physiological significance in this setting has never been explored. To do so, Escherichia coli was intratracheally instilled into C57BL/6 mice in the presence of neutralizing anti-LIF IgG or control IgG. Anti-LIF completely eliminated lung LIF detection and markedly exacerbated lung injury compared with control mice as evidenced by airspace albumin content, lung liquid accumulation, and histological analysis. Although lung bacteriology was equivalent between groups, bacteremia was more prevalent with anti-LIF treatment, suggestive of compromised barrier function rather than impaired antibacterial defense as the cause of dissemination. Inflammatory cytokine expression was also exaggerated in anti-LIF-treated lungs, albeit after injury had ensued. Interestingly, alveolar neutrophil recruitment was modestly but significantly reduced compared with control mice despite elevated cytokine levels, indicating that inflammatory injury was not a consequence of excessive neutrophilic alveolitis. Lastly, the lung transcriptome was dramatically remodeled during pneumonia, but far more so following LIF neutralization, with gene changes implicating cell death and epithelial homeostasis among other processes relevant to tissue injury. From these findings, we conclude that endogenous LIF facilitates tissue protection during pneumonia. The LIF-STAT3 axis is identified in this study as a critical determinant of lung injury with clinical implications for pneumonia patients.

Figure 1

LIF expression and biological activity in the lungs during pneumonia. (A) LIF protein concentrations were quantified in bronchoalveolar lavage fluid (BALF) 0–24 hrs after intratracheal (i.t.) instillation of Escherichia coli. Values are expressed as means ± SEM (n = 4–5). * p < 0.05 compared to uninfected (0h) controls. (B) Immunohistochemistry was used to visualize Y705-phosphorylated STAT3 in histological lung sections prepared from mice treated for 1h with or without i.t. recombinant murine LIF (rmLIF). Background staining was undetectable on sections from i.t. rmLIF mice exposed to an isotype control antibody (not shown). Representative images are shown at 10× and 20× magnification. (C) LIF protein concentrations were quantified in BALF 24 hrs after i.t. E. coli co-instilled with anti-LIF or control IgG. Values are expressed as means ± SEM. * p < 0.05 compared to mice treated with control IgG (n = 4–5). (D) Y705- phosphorylated STAT3 immunoreactivity was measured by immunoblot in lung homogenates collected from mice 1h after i.t. rmLIF in the presence of 0–10 µg anti-LIF.

Figure 2

Effect of LIF neutralization on acute lung injury. Lungs were collected from mice 24 hrs after intratracheal inoculation of Escherichia coli co-instilled with anti-LIF or control IgG. (A) Representative images are shown for intact freshly isolated lungs and hematoxylin/eosin-stained lung sections. Red circles are used to denote infected left lung lobes. Bronchoalveolar lavage fluid concentrations of (B) mouse albumin and (C) receptor for advanced glycation end products (RAGE), as well as (D) lung wet:dry weight ratios were determined and expressed as means ± SEM. * p < 0.05 compared to mice treated with control IgG (n = 3–5).

Figure 3

Effect of LIF neutralization on bacterial clearance. Escherichia coli colony-forming units (CFU) were enumerated in blood and lungs after 24 hrs of pneumonia in the presence of anti-LIF or control IgG. CFU/ml (blood) and total lung CFU are shown for individual mice with horizontal lines indicating the median value within each experimental group. * p < 0.05 compared to mice treated with control IgG (n = 9–13).

Figure 4

Effect of LIF neutralization on innate immunity. (A–B) Neutrophil numbers and (C–D) cytokine protein concentrations were quantified in bronchoalveolar lavage fluid (BALF) harvested from mouse lungs 24 or 30 hrs after intratracheal Escherichia coli. Values are expressed as means ± SEM. * p < 0.05 compared to mice treated with control IgG (n = 4–5).

Figure 5

Effect of LIF supplementation on IL-6 induction in vitro. IL-6 mRNA induction was determined in alveolar macrophages (AM), RAW 264.7 cells and MLE-12 cells following 2 hrs of LPS stimulation in the absence and presence of recombinant murine LIF (rmLIF). Data are expressed as fold-induction compared to cells not treated with LPS. Means ± SEM were calculated by combining data from 3 separate experiments or 3 individual mice (n = 3).

Figure 6

Effects of pneumonia and LIF blockade on the lung transcriptome and markers of cell death. (A–B) Microarrays were performed on total lung RNA 24 hrs after intratracheal (i.t.) instillations of the following 3 combinations: 1) saline and control IgG; 2) Escherichia coli and control IgG; and 3) Escherichia coli and anti-LIF (n = 3). (A) Values indicate the number of differentially expressed genes in each group compared to uninfected mice treated with control IgG alone. (B) A heat map illustrates the expression profile of the 1313 genes significantly different between pneumonic mice treated with control IgG (left 3 columns) or anti-LIF (right 3 columns). For each row (gene), green and red annotations denote decreased or increased expression, respectively. Data in panels (A) and (B) include significantly altered transcripts (FDR < 0.05) with an expression difference of at least 2 fold. qRT-PCR was used to interrogate mRNA fold-induction (compared to control IgG-treated mice) for (C) Fas and (D) VEGF, both of which were significantly affected by anti-LIF treatment as determined by microarray. Individual heat map data are shown for both transcripts to illustrate relative expression levels determined by microarray analysis. * p < 0.05 compared to mice treated with control IgG (n = 3).