Hepcidin sustains Kupffer cell immune defense against bloodstream bacterial infection via gut-derived metabolites in mice

Abstract

Bloodstream bacterial infections cause one-third of deaths from bacterial infections, and eradication of circulating bacteria is essential to prevent disseminated infections. Here, we found that hepcidin, the master regulator of systemic iron homeostasis, affected Kupffer cell (KC) immune defense against bloodstream bacterial infections by modulating the gut commensal bacteria-derived tryptophan derivative indole-3-propionic acid (IPA). Hepcidin deficiency impaired bacterial capture by KCs and exacerbated systemic bacterial dissemination through morphological changes in KCs. Gut microbiota depletion and fecal microbiota transplantation revealed that the gut microbiota mediated the alteration of KCs volume. Mechanistically, hepcidin deficiency led to a decreased abundance of the IPA-producing commensal Lactobacillus intestinalis and a concomitant reduction in the gut-to-liver shuttling of its metabolite IPA. IPA supplementation or L. intestinalis colonization restored the KC volume and hepatic immune defense against bloodstream bacterial infection in hepcidin-deficient mice. Moreover, hepcidin levels in patients with bacteremia were associated with days of antibiotic usage and hospitalization. Collectively, our findings highlight a previously unappreciated role of hepcidin in sustaining KC-mediated hepatic defense against bloodstream bacterial infections through the gut commensal L. intestinalis and its tryptophan derivative IPA. More importantly, we show that restoring the crosstalk between the gut microbiota and liver through IPA-inspired therapies may offer a promising strategy for enhancing the host defense against bloodstream bacterial infections in those with low hepcidin levels and a high risk for bacterial infections.

Keywords: Bacterial infections; Immunology; Infectious disease.

Figure 1. Hepcidin deficiency impairs hepatic immune defense against bloodstream bacterial infection.

(A) Representative confocal intravital microscopy (IVM) images of the liver microcirculation of Hamp1^–/– and WT mice at 2 h after E. coli–GFP (green) infection. Scale bars: 20 μm. (B) Quantitative analysis of E. coli–GFP sequestered in the liver microcirculation (per field of view [FOV]) by confocal intravital microscopy. n = 5 per group; data are presented as mean ± SEM. (C and D) Bacterial load in the liver (C) and peripheral blood (D) at 2 h after E. coli infection in Hamp1^–/– and WT mice. n = 8 per group; data are presented as median ± interquartile range. (E) Representative IVM images of the liver microcirculation of Hamp1^–/– and WT mice at 24 h after E. coli–GFP infection. Scale bars: 20 μm. (F) Quantitative analysis of residual E. coli–GFP in the liver microcirculation (per FOV) at 24 h after E. coli–GFP infection by confocal intravital microscopy. n = 6–7 per group; data are presented as mean ± SEM. (G and H) Bacterial load in the liver (G) and peripheral blood (H) at 24 h after E. coli infection in Hamp1^–/– and WT mice. n = 6 per group; data are presented as median ± interquartile range. (I) Survival rate of Hamp1^–/– and WT mice after E. coli infection. n = 10 per group. *P < 0.05, **P < 0.01, by 2-tailed Student’s t test (A, B, E, and F), Mann-Whitney U test (C, D, G, and H), and Kaplan-Meier log-rank test (I). Data presented are from 5 (A and B), 8 (C and D), 4 (E and F), 3 (G and H), and 6 (I) independent experiments. Each symbol represents an individual mouse.

Figure 2. Hepcidin deficiency impairs KCs to clear invading bacteria.

(A) Representative intravital microscopy (IVM) images showing KCs (red) capturing circulating E. coli–GFP (yellow) in Hamp1^–/– and WT mice. Scale bars: 20 μm. (B) Number of E. coli–GFP captured per KC. n = 6 per group; data are presented as mean ± SEM. (C) Representative IVM images of KCs (red) in Hamp1^–/– and WT mice. Scale bars: 50 μm. (D) Total number of KCs (per field of view) in Hamp1^–/– and WT mice. n = 8 per group; data are presented as mean ± SEM. (E) Flow cytometry analysis of liver CD45⁺F4/80⁺CD11b^int cells of Hamp1^–/– and WT mice with t-distributed stochastic neighbor embedding dimension reduction. (F) Quantitative analysis of subsets of KCs (KC1 and KC2) in liver CD45⁺F4/80⁺CD11b^int cells of Hamp1^–/– and WT mice. n = 7–8 per group; data are presented as mean ± SEM. (G–J) IVM images (G) combined with 3-dimensional reconstruction (H) to analyze KC volume (I) and surface area (J) in Hamp1^–/– and WT mice. Scale bars: 20 μm. n = 6 per group; data are presented as mean ± SEM. *P < 0.05, ****P < 0.0001, by 2-tailed Student’s t test. Data presented are from 6 (A and B), 8 (C and D), 4 (E and F), and 6 (G–J) independent experiments. Each symbol represents an individual mouse (A–F). Symbols represent individual KCs from 6 mice with 5 fields of view per mouse (I and J).

Figure 3. The gut microbiota mediates the malfunction of KC antibacterial defense in Hamp1^–/– mice.

(A) Representative intravital microscopy (IVM) images showing KCs (red) capturing circulating E. coli–GFP (yellow) in Hamp1^–/– and WT mice treated with antibiotic cocktail (4Abx) or sterile water (control). Scale bars: 20 μm. (B) Number of E. coli–GFP captured per KC. n = 6–7 per group; data are presented as mean ± SEM. (C–F) IVM images (C) combined with 3-dimensional reconstruction (D) to analyze KC volume (E) and surface area (F) in Hamp1^–/– and WT mice treated with 4Abx or sterile water (control). Scale bars: 20 μm. n = 6–7 per group; data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, by 1-way ANOVA followed by Šidák’s multiple-comparison test. Data presented are from 6 independent experiments (A–F). Each symbol represents an individual mouse (A and B). Symbols represent individual KCs from mice, and data are from 6–7 mice with 5 fields of view per mouse (E and F).

Figure 4. Fecal microbiota transplantation rescues malfunction of KC bacterial clearance in Hamp1^–/– mice.

(A) Representative intravital microscopy (IVM) images showing KCs (red) capturing circulating E. coli–GFP (yellow) in Hamp1^–/– and WT mice that received fecal microbiota transplantation. Scale bars: 20 μm. (B) Number of E. coli–GFP captured per KC. n = 6 per group; data are presented as mean ± SEM. (C–F) IVM image (C) combined with 3-dimensional reconstruction (D) to analyze KC volume (E) and surface area (F) in Hamp1^–/– and WT mice that received fecal microbiota transplantation. Scale bars: 20 μm. n = 6 per group; data are presented as mean ± SEM. *P < 0.05, ***P < 0.001, ****P < 0.0001, by 1-way ANOVA followed by Šidák’s multiple-comparison test. Data presented are from 6 independent experiments (A–F). Each symbol represents an individual mouse (B), and symbols represent individual KCs from 6 mice with 5 fields of view per mouse (E and F).

Figure 5. IPA supplementation restores hepatic immune defense against bloodstream bacterial infection in Hamp1^–/– mice.

(A) Hepcidin deficiency downregulated metabolites involved in the tryptophan metabolism pathway. n = 4 per group. (B) Volcano plot showing the fold change and P value of gut microbiota–derived metabolites in portal blood from Hamp1^–/– and WT mice. n = 4 per group. (C) Diagram of tryptophan metabolized by the gut microbiota to ILA, IA, and IPA. (D and E) Portal blood levels of ILA, IA, and IPA in Hamp1^–/– and WT mice treated with either antibiotic cocktail (4Abx) or sterile water (control) (D) and in Hamp1^–/– and WT mice that received fecal microbiota transplantation (E). n = 6 per group; data are presented as mean ± SEM. (F) Portal blood level of IPA in Hamp1^–/– mice pretreated with IPA or sterile PBS. n = 5 per group; data are presented as mean ± SEM. (G) Representative intravital microscopy (IVM) images showing KCs (red) capturing circulating E. coli–GFP (yellow) in Hamp1^–/– mice pretreated with IPA or sterile PBS. Scale bars: 20 μm. (H) Number of E. coli–GFP captured per KC. n = 6 per group; data are presented as mean ± SEM. (I–L) IVM images (I) combined with 3-dimensional reconstruction (J) to analyze KC volume (K) and surface area (L) in Hamp1^–/– mice pretreated with IPA or sterile PBS. Scale bars: 20 μm. n = 6 per group; data are presented as mean ± SEM. *P < 0.05, ****P < 0.0001, by 1-way ANOVA followed by Šidák’s multiple-comparison test (D and E) or 2-tailed Student’s t test (F–L). Data presented are from 2 (D and E) and 4 (G–L) independent experiments. Each symbol represents an individual mouse (D–H). Symbols represent individual KCs from 6 mice with 5 fields of view per mouse (K and L). Trp, tryptophan.

Figure 6. Colonization by IPA-producing L. intestinalis restores hepatic immune defense against bloodstream bacterial infection in Hamp1^–/– mice.

(A) IPA concentration in the culture supernatant of L. intestinalis and Man, Rogosa, and Sharpe (MRS) medium. n = 6 per group; data are presented as mean ± SEM. (B) Portal blood level of IPA in mice pretreated with L. intestinalis or sterile PBS (control). n = 6 per group; data are presented as mean ± SEM. (C) Representative intravital microscopy (IVM) images showing KCs (red) capturing circulating E. coli–GFP (yellow) mice pretreated with L. intestinalis or sterile PBS (control). Scale bars: 20 μm. (D) Number of E. coli–GFP captured per KC. n = 6 per group; data are presented as mean ± SEM. (E–H) IVM image (E) combined with 3-dimensional reconstruction (F) to analyze KC volume (G) and surface area (H) in Hamp1^–/– and WT mice pretreated with L. intestinalis or sterile PBS (Control). Scale bars: 20 μm. n = 6 per group; data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, by 2-tailed Student’s t test (A) or 1-way ANOVA followed by Šidák’s multiple-comparison test (B–H). Data presented are from 6 independent experiments (A–H). Each symbol represents an individual mouse (B–D). Symbols represent individual KCs from 6 mice with 5 fields of view per mouse (G and H).

Figure 7. Dietary iron restriction improves KC bacterial clearance and liver immune defense in Hamp1^–/– mice.

(A) Portal blood level of IPA in Hamp1^–/–-Std (Hamp1^–/– mice fed a standard diet) and Hamp1^–/–-Low (Hamp1^–/– mice fed a low-iron diet) mice. n = 6 per group; data are presented as mean ± SEM. (B–E) Intravital microscopy (IVM) image (B) combined with 3-dimensional reconstruction (C) to analyze KC volume (D) and surface area (E) in Hamp1^–/–-Std and Hamp1^–/–-Low mice. Scale bars: 20 μm. n = 6 per group; data are presented as mean ± SEM. (F) Representative IVM images showing KCs (red) capturing circulating E. coli–GFP (yellow) in Hamp1^–/–-Std and Hamp1^–/–-Low mice. Scale bars: 20 μm. (G) Number of E. coli–GFP captured per KC. n = 6 per group; data are presented as mean ± SEM. (H) Representative IVM images of the liver microcirculation at 24 h after E. coli–GFP infection. Scale bars: 20 μm. (I) Quantitative analysis of residual E. coli–GFP in the liver microcirculation (per field of view) at 24 h after E. coli–GFP infection. n = 5–6 per group; data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ****P < 0.0001, by 2-tailed Student’s t test. Data presented are from 2 independent experiments (A–I). Each symbol represents an individual mouse (A and F–I). Symbols represent individual KCs from 6 mice with 5 fields of view per mouse (D and E).

Figure 8. Hepcidin levels in patients with bacteremia correlate with their clinical status.

(A) Plasma hepcidin levels in noninfectious controls (n = 25) and patients with E. coli bacteremia (n = 27) or with S. aureus bacteremia (n = 28). (B–D) Correlations of plasma hepcidin level with supersensitive C-reactive protein (SCRP) (B), platelet count (C), and days of antibiotic usage (D) in patients with E. coli bacteremia. (E–G) Correlations of plasma hepcidin level with SCRP (E), platelet count (F), and days of antibiotic usage (G) in patients with S. aureus bacteremia. (H) Correlation of plasma hepcidin level with hospital stay in all patients with bacteremia (n = 55). (I–K) Correlations of plasma ferritin level with plasma hepcidin level (I), days of antibiotic usage (J), and days of hospital stay (K) in all patients with bacteremia (n = 55). **P < 0.01, ***P < 0.001, by 1-way ANOVA followed by Tukey’s multiple-comparison test (A) or Pearson’s correlation analysis (B–K).