Neutrophils and galectin-3 defend mice from lethal bacterial infection and humans from acute respiratory failure

Abstract

Respiratory infection by Pseudomonas aeruginosa, common in hospitalized immunocompromised and immunocompetent ventilated patients, can be life-threatening because of antibiotic resistance. This raises the question of whether the host's immune system can be educated to combat this bacterium. Here we show that prior exposure to a single low dose of lipopolysaccharide (LPS) protects mice from a lethal infection by P. aeruginosa. LPS exposure trained the innate immune system by promoting expansion of neutrophil and interstitial macrophage populations distinguishable from other immune cells with enrichment of gene sets for phagocytosis- and cell-killing-associated genes. The cell-killing gene set in the neutrophil population uniquely expressed Lgals3, which encodes the multifunctional antibacterial protein, galectin-3. Intravital imaging for bacterial phagocytosis, assessment of bacterial killing and neutrophil-associated galectin-3 protein levels together with use of galectin-3-deficient mice collectively highlight neutrophils and galectin-3 as central players in LPS-mediated protection. Patients with acute respiratory failure revealed significantly higher galectin-3 levels in endotracheal aspirates (ETAs) of survivors compared to non-survivors, galectin-3 levels strongly correlating with a neutrophil signature in the ETAs and a prognostically favorable hypoinflammatory plasma biomarker subphenotype. Taken together, our study provides impetus for harnessing the potential of galectin-3-expressing neutrophils to protect from lethal infections and respiratory failure.

Fig. 1. Pre-exposure to LPS protects mice from a lethal infection by PA14 with reduced bacterial burden and a regulated inflammatory response.

a Schematic of PA14 infection with or without LPS pre-treatment. Bacterial load showing reduced bacterial burden in the lung (b) and decreased peripheral dissemination (c) in LPS + PA14 mice compared to PA14 mice (4 h post-infection). n = 9 (PA14), 9 (LPS(1d)+PA14), and 10 (LPS(3d)+PA14) mice. Data were log transformed using log base 10 to adjust for differences in standard deviation prior to analysis. Data pooled from 3 independent experiments and analyzed using one-way ANOVA with the Dunn’s test. d Kaplan-Meier survival curve showing 100% mortality of PA14 infected mice within 18–20 h of infection, whereas 100% survival of LPS + PA14 group as monitored up to 14 days post infection. n = 8 (PA14 and LPS + PA14) mice per group. Data representative of 3 independent experiments and analyzed using Log-rank test. e Higher albumin levels in BALF of PA14 mice compared to LPS ±  PA14 mice (mice pre-exposed to LPS for 3 days), BALF being collected 4 h post-infection from both groups of mice and also from naïve mice and mice only treated with LPS. n = 7 (Naïve and LPS), and 8 (PA14 and LPS + PA14) mice. Data pooled from 2 independent experiments and analyzed using ordinary one-way ANOVA with the Tukey post-hoc test. f RT-qPCR analysis of gene expression of Il1b, Il6, Ifit1, Isg15, Ifng, Il10, Stat1 and Aoah in the lung tissue of the four groups of mice. n = 3 mice per group. Data representative of 3 independent experiments and analyzed using ordinary one-way ANOVA with the Dunnett’s test. g MPO and ELA2 protein levels associated with neutrophil influx and chemokine and cytokine levels in the BALF of the two groups of mice. n = 7 (Naïve and LPS), and 8 (PA14 and LPS + PA14) mice. Data pooled from 2 independent experiments and analyzed using ordinary one-way ANOVA with the Tukey post-hoc test. All data are presented as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Source data and exact P values are provided as a Source Data file.

Fig. 2. Pre-exposure to LPS promotes an increase in neutrophils and interstitial macrophages in mouse lungs.

a Flow cytometry analysis showing total numbers of lung cells, AMs, CD11b⁺ cells, neutrophils, classical monocytes, and IMs. n = 6 (Naïve and LPS), and 9 (PA14 and LPS + PA14) mice. Data pooled from 2 independent experiments and analyzed using ordinary one-way ANOVA with the Tukey post-hoc test. b Representative flow plots for identification of the cell types in the lungs of the two groups of mice. c Representative flow plots and total counts of IFN-γ-expressing neutrophils in the lungs of the two groups of mice from 2 independent experiments. n = 3 (Naïve and LPS), and 4 (PA14 and LPS + PA14) mice. Data analyzed using ordinary one-way ANOVA with the Tukey post-hoc test. All data are presented as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Source data and exact P values are provided as a Source Data file.

Fig. 3. Identification of immune cell populations in the lungs of LPS + PA14 and PA14 mice by scRNA-seq.

a UMAP embedding of the integrated expression profile of 46,826 cells from two conditions (LPS + PA14 and PA14), each comprised of three biological replicates. Distinct clusters are annotated and color-coded. b Expression of canonical markers used to annotate clusters, in conjunction with automated annotation using SingleR v2.2.0 (ImmGen reference scRNA-seq dataset). c Proportions of cell types represented in the two conditions, including the total count of cells in each cell type and the experimental condition. d Feature plot visualizing the expression of Aoah, the color of points (cells) represents the expression level of Aoah. Cell types are labeled, and the plot is split by experimental condition to highlight differences in expression levels between cells from LPS + PA14 and PA14 mice. n = 3 mice per group.

Fig. 4. Differential gene expression in cell clusters in LPS + PA14 versus PA14 mice.

a Volcano plot of genes differentially expressed genes in the DESeq2-based pseudo-bulk analysis of LPS + PA14 versus PA14 cells. Red points indicate significance in both fold-change (log2FC > 0.5) and Benjamini-Hochberg (BH) adjusted p-value (p < 0.05). b Stacked violin plots highlighting select genes that were found to be differentially expressed between the two conditions (LPS + PA14 versus PA14). c A dotplot comparing the expression levels of core genes of select pathways of interest enriched in LPS + PA14 against PA14 cells. The size of the dot indicates the gene ratio, and the color of the dot shows the normalized enrichment score (NES). The dendrogram in the y-axis represents the clustering of the pathways by the NES. d Feature plot visualizing expression of Lgals3, and the color represents the average expression level across cell types. e Chord diagrams showing the pathways of interest from the GSEA. In each diagram, enriched clusters are shown on the right and genes contributing to the enrichment are shown on the left. Genes are colored by log fold-change value, and the color of the chords represent distinct pathway/ontology Terms.

Fig. 5. Pre-treatment with LPS educates neutrophils for efficient phagocytosis and cell killing.

a Quantitative fluorescence intravital lung microscopy (qFILM) images of the same field of view (FOV) in the lung of an LPS + PA-GFP mouse at four different time points showing neutrophils crawling intravascularly (direction shown by white arrow) towards PA-GFP and finally phagocytosis of PA-GFP. Complete time series for (a) shown in Supplementary movies 1 and 2. b Magnified images showing phagocytosis in three different mice using merged channel pictures and isolated red and green channels. c Relative neutrophil counts in FOV in the two groups of mice. d qFILM data were analyzed as described in Methods. Phagocytosis of PA by neutrophils is shown as the average colocalization of eGFP and Pacific Blue (PB) signals within the FOV. e The percentage of eGFP⁺ neutrophils combining all fields of view. Representative images (a, b) and pooled data (c–e) from 3 independent experiments with n = 2 mice per group per experiment. Data analyzed using two-tailed unpaired t test with Welch’s correction. f Dot plots showing the expression levels of phagocytosis-related genes in all cell types, separated by condition. Flow cytometry analysis of BAL cells (g) with neutrophils (h, i) comprising more than 90% of the BAL cells in PA14 and 98% in LPS + PA14 mice and AMs (j, k) comprising around 10% and 0.2% respectively. g–k n = 3 mice per group per experiment. Data pooled from 2 independent experiments and analyzed using two-tailed unpaired t test with Welch’s correction. l Bacterial burden in BALF. m Intracellular bacterial load in BAL cells assessed by permeabilizing BAL cells. n Galectin-3 protein associated with BAL cells (primarily neutrophils). l–n Data pooled from 2 independent experiments with n = 7 mice per group. Data were log transformed using log base 10 to adjust for differences in standard deviation prior to analysis for (l, m). Data analyzed using two-tailed unpaired t test with Welch’s correction. All data are presented as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Source data and exact P values are provided as a Source Data file.

Fig. 6. Galectin-3 levels in lower respiratory tract specimens predict survival and correlate with markers of neutrophil activation in patients with acute respiratory failure.

a Survivors of acute respiratory failure showing significantly higher galectin-3 levels in supernatant fluid of endotracheal aspirates (ETAs) compared to non-survivors by 30-days post intubation (p = 0.026). b ARDS survivors showing significantly higher galectin-3 levels in ETA supernatants compared to ARDS non-survivors (p = 0.0046). c Based on plasma biomarkers, patients classified as prognostically favorable hypoinflammatory subphenotype showed higher LRT galectin-3 levels compared to patients classified to the prognostically adverse hyperinflammatory subphenotype (p = 0.00039). a–c Data are represented as boxplots with median as the line inside the box, interquartile range (25th–75th percentile) as the box itself, whiskers extend to 1.5 times the interquartile range, and individual dots beyond whiskers signify outlier observations. P values from Wilcoxon test. All tests were two-sided. d Galectin-3 levels were significantly correlated with neutrophil elastase levels in ETA supernatants among acute respiratory failure survivors only. e A significant correlation between ETA levels of galectin-3 and neutrophil elastase was observed only among ARDS survivors. f LRT galectin-3 levels correlated with neutrophil elastase levels in ETA supernatants in hypoinflammatory subphenotype patients. P values from Pearson’s correlation tests. n = 81 independent subjects. Source data is provided as a Source Data file.

Fig. 7. Lack of galectin-3 in mouse increases lung bacterial burden with suppression of IFN-γ-expressing neutrophils.

a Bacterial load showing increased bacterial burden in the lungs of KO mice in both PA14 and LPS + PA14 groups as compared to the respective groups in WT mice (assayed 4 h post-infection). Data were log transformed using log base 10 to adjust for differences in standard deviation prior to analysis. Data pooled from 3 independent experiments with n = 10 mice per group. Data analyzed using Brown Forsythe and Welch ANOVA test with Dunnett’s T3 comparisons. Flow cytometry analysis showing total numbers of lung cells (b), CD11b⁺ cells (c), and neutrophils (d) in both WT and galectin-3 KO LPS + PA14 mice. Total cell counts (e) and representative flow plots (f) of IFN-γ-expressing neutrophils in the lungs of both WT and KO LPS + PA14 mice. b–f Data pooled from 3 independent experiments where n = 10 mice per group. g Flow cytometry analysis showing total BAL cells in WT and KO mice. h Intracellular bacterial load in BAL cells assessed by permeabilizing BAL cells from WT and KO mice. Data pooled from 2 independent experiments with n = 3 mice per group. i Bacterial burden in lungs of LPS + PA14 infected KO mice without or with adoptive transfer of BAL neutrophils isolated from LPS treated WT mice. g–i Data pooled from 2 independent experiments with n = 7 mice per group. b–e, g–i Data analyzed using two-tailed unpaired t test with Welch’s correction. All data are presented as mean ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Source data and exact P values are provided as a Source Data file.

Fig. 8. Schematic depicting the protective role of galectin-3 in mice and humans, as revealed in our study.

A single low-dose LPS administration in the lungs of mice results in 100% survival of mice infected with a lethal dose of the virulent strain of Pseudomonas aeruginosa, PA14. scRNA-seq and downstream bioinformatic of lung cells of PA14-infected mice that were pre-exposed to LPS or not revealed expansion of a specific neutrophil population, we named N3, and an IM population resembling IM2 macrophages described in the literature, in the LPS pre-exposed mice. Among the four neutrophil populations, compared to the N3 population in PA14 mice, the N3 population in the LPS-treated mice was significantly more enriched in pathways associated with bacterial phagocytosis and cell killing. The cell-killing pathway comprised multiple genes with well-documented anti-bacterial functions including Lgals3, Cxcl9, Cxcl10, and Ifng. BAL cells were 100x more abundant in LPS-exposed mice compared to PA14 mice, in both >95% being neutrophils. BAL-neutrophil-associated levels of the anti-bacterial protein galectin-3 (encoded by Lgals3), were also significantly higher in the LPS-treated mice. The host-protective role of galectin-3 was also evident in critically ill hospitalized patients in the ICU with acute respiratory failure (ARF). High galectin-3 levels and a high neutrophil signature in endotracheal aspirates of patients with ARF were associated with greater survival, the data being primarily driven by patients with a diagnosis of ARDS. The schematic was created with BioRender (www.biorender.com).