Ketogenesis promotes tolerance to Pseudomonas aeruginosa pulmonary infection

Abstract

Pseudomonas aeruginosa is a common cause of pulmonary infection. As a Gram-negative pathogen, it can initiate a brisk and highly destructive inflammatory response; however, most hosts become tolerant to the bacterial burden, developing chronic infection. Using a murine model of pneumonia, we demonstrate that this shift from inflammation to disease tolerance is promoted by ketogenesis. In response to pulmonary infection, ketone bodies are generated in the liver and circulate to the lungs where they impose selection for P. aeruginosa strains unable to display surface lipopolysaccharide (LPS). Such keto-adapted LPS strains fail to activate glycolysis and tissue-damaging cytokines and, instead, facilitate mitochondrial catabolism of fats and oxidative phosphorylation (OXPHOS), which maintains airway homeostasis. Within the lung, P. aeruginosa exploits the host immunometabolite itaconate to further stimulate ketogenesis. This environment enables host-P. aeruginosa coexistence, supporting both pathoadaptive changes in the bacteria and the maintenance of respiratory integrity via OXPHOS.

Keywords: OXPHOS; Pseudomonas aeruginosa; bioenergetics; disease tolerance; infection; inflammation; itaconate; ketogenesis; ketogenic diet; pneumonia.

Figure 1.. P. aeruginosa LPS display inhibits pulmonary ATP synthase –

Mice were treated with PBS, WT PAO1, or ΔlptD PAO1. The following was measured: A) OXPHOS (ATP5A, MTCO1, SDHB) and glycolysis (HK2) in lung; B) BAL glucose. C) NADH and FADH₂ abundance in WT and ΔlptD PAO1 via Biolog plates. D) WT and ΔlptD PAO1 growth in succinate-rich minimal media. E) Lung ATP5A expression in mice exposed to either PBS, WT PAO1, or ΔlptD PAO1. ATP5/β-actin ratio is shown. F-G) BDMDs oxygen consumption rate (OCR). Data are shown as average of 2-3 independent experiments +/− SEM. B-C: t-Student test. D, F-G: Two-Way ANOVA. *: P<0.05; **: P<0.01; ****: P<0.0001; ns: non-significant.

Figure 2.. Host OXPHOS impairment worsens lung inflammation and P. aeruginosa burden –

Mice were administered either vehicle or oligomycin (ATP synthase inhibitor) and exposed to PBS or WT PAO1. The following were analyzed: A) BAL cytokines; B) bacterial burden in BAL and lung. Animals were treated with either vehicle or etomoxir (FAO blocker) and exposed to PBS, WT PAO1, or ΔlptD PAO1. The following were examined: C) BAL cytokines for WT PAO1; D) bacterial burden in BAL and lung for WT PAO1; E) BAL cytokines for ΔlptD PAO1; F) bacterial burden in BAL and lung for ΔlptD PAO1. Data are shown as average of 2-3 independent experiments +/− SEM, 4-7 mice in total. A-F: One-Way ANOVA. *: P<0.05; **: P<0.01; ns: non-significant. See also Figure S1.

Figure 3.. P. aeruginosa isolates preserve airway OXPHOS and enable host-pathogen coexistence –

LptD mRNA levels in P. aeruginosa strains from: A) 13 patients at ICU; B) tolerant people with CF: CF#1: 14 strains; CF#2: 17 strains. Animals were exposed to PBS, WT PAO1, or a mixture of the 17 P. aeruginosa isolates from CF#2. The following were analyzed: C) OXPHOS (ATP5A, MTCO1, SDHB) and glycolysis (HK2) in lung; D) pulmonary cell subsets identified by scRNA-Seq; E-F) ΔOXPHOS transcriptomic score per lung population in response to each pathogen; G) mouse survival; H) BAL cytokines; I) bacterial burden in lung and BAL (black cross indicates all animal are dead). J-K) Survival of animals exposed to either ~10⁶ or ~10⁷ CFUs of the P. aeruginosa isolates and then treated with either a sublethal (J) or lethal dose (K) of oligomycin. WT and ΔlptD PAO1 were also studied. L) Bacterial endpoint growth in increasing succinate. Data are shown as average +/− SEM. C-I: 2-3 independent experiments, with 3-7 animals in total. J-K: 2 independent experiments, with 5 mice in total. I, L: Two-Way ANOVA; G, J-K: Kaplan-Maier; H: One-Way ANOVA. *: P<0.05; **: P<0.01; ***: P<0.001; ****: P<0.0001; ns: non-significant. See also Figure S1, Figure S2, Table S1 and Table S2.

Figure 4.. Dietary fats modulate host tolerance to P. aeruginosa lung infection –

Mice fed with either a fatty acid-rich (F. Acids) or a carbohydrate-rich (Carbs) diet were exposed to PBS, WT PAO1, ΔlptD PAO1 or the P. aeruginosa isolates. The following were examined: A) BAL glucose; B-C) bacterial burden; D) glycemia; E-G) BAL cytokines enrichment; H) blood BHB (beginning of the light cycle); I-J) phagocyte recruitment in BAL and lung; K) Lung H&E; L-M) body temperature; N-O) animal survival; Data are from 2-3 independent assays, 4-8 mice in total, and shown as average +/− SEM. A-D, H-J: One-Way ANOVA. L-M: Two-Way ANOVA. N-O: Kaplan-Maier. E-G: t-Student test. *: P<0.05; **: P<0.01; ***: P<0.001; ns: non-significant. See also Figure S3.

Figure 5.. Airway ketones drive P. aeruginosa LPS patho-adaptation –

Mice were exposed to PBS, WT PAO1, or ΔlptD PAO1. The following were measured at the beginning of the light cycle: A) BHB in blood; B-D) BHB and AcAc abundance in BAL. In D, pathway enrichment analysis was performed via Metaboloanalyst 5.0 using the KEGG platform as database. E) Ketones in sputum of HC and individuals with CF. F-H) Number of non-synonymous mutations (NSM) found in each P. aeruginosa isolate in routes linked to LPS biogenesis. I) mRNA levels of LPS assembly genes. WT PAO1 was grown in glucose-rich minimal media and complemented with AcAc, BHB, or both ketones. The following were measured: J) mRNA levels of gene clusters involved in LPS assembly; K) lipid A phosphorylation via MALDI-TOF; L) O-antigen abundance. Data are shown as average +/− SEM from 2-3 independent assays. A-D: 5-6 mice in total per group. A-C, J: One-Way ANOVA. D-E: t-Student test: *: P<0.05; **: P<0.01; ***: P<0.001; ****: P<0.0001; ns: non-significant. See also Figure S4.

Figure 6.. Itaconate promotes ketone enrichment in the lung during P. aeruginosa infection –

Irg1^+/+ and Irg1^−/− mice were exposed to PBS, WT PAO1, or the P. aeruginosa isolates. The following were measured: A-B) BAL metabolites; C) mRNA expression of ketogenic clusters in each lung cell subset identified by scRNA-Seq; D) Volcano plot of Δketogenesis scores (“Irg1^+/+” – “Irg1^−/−”) for each lung cell subset during infection with the P. aeruginosa isolates; E) ketogenesis transcriptomic score in fibroblasts; F) Expression of ketogenic clusters in fibroblasts; G) BAL FGF21; H-I) respiratory pathogen burden during infection with the P. aeruginosa isolates; J) animal survival; K) BAL cytokines; L) weight change. Data are shown as average +/− SEM from 2-3 independent experiments, with 3-10 animals in total. A-B, E, G-I: One-Way ANOVA; L: Two-Way ANOVA; D: t-Student test. J: Kaplan-Maier. *: P<0.05; **: P<0.01; ***: P<0.001; ****: P<0.0001; ns: non-significant. See also Figure S5, Figure S6, and Table S2.