Heme oxygenase 1-mediated ferroptosis in Kupffer cells initiates liver injury during heat stroke

Abstract

With the escalating prevalence of global heat waves, heat stroke has become a prominent health concern, leading to substantial liver damage. Unlike other forms of liver injury, heat stroke-induced damage is characterized by heat cytotoxicity and heightened inflammation, directly contributing to elevated mortality rates. While clinical assessments have identified elevated bilirubin levels as indicative of Kupffer cell dysfunction, their specific correlation with heat stroke liver injury remains unclear. Our hypothesis proposes the involvement of Kupffer cell ferroptosis during heat stroke, initiating IL-1β-mediated inflammation. Using single-cell RNA sequencing of murine macrophages, a distinct and highly susceptible Kupffer cell subtype, Clec4F⁺/CD206⁺, emerged, with heme oxygenase 1 (HMOX-1) playing a pivotal role. Mechanistically, heat-induced HMOX-1, regulated by early growth response factor 1, mediated ferroptosis in Kupffer cells, specifically in the Clec4F⁺/CD206⁺ subtype (KC2), activating phosphatidylinositol 4-kinase beta and promoting PI4P production. This cascade triggered NLRP3 inflammasome activation and maturation of IL-1β. These findings underscore the critical role of targeted therapy against HMOX-1 in ferroptosis within Kupffer cells, particularly in Clec4F⁺/CD206⁺ KCs. Such an approach has the potential to mitigate inflammation and alleviate acute liver injury in the context of heat stroke, offering a promising avenue for future therapeutic interventions.

Keywords: Early growth response factor 1; Ferroptosis; Heat stroke; Heme oxygenase 1; Kupffer cells; Liver injury; NLRP3; Phosphatidylinositol 4-kinase beta.

Graphical abstract

Figure 1

Key role of KCs ferroptosis in HS mice mortality. (A) ALT, (B) AST, and (C) TBIL levels in serum from HC (n = 32) and HS (n = 32) individuals. (D) TBIL levels in the serum of human HS patients on the first- and second-days. (E) Survival time of HS patients was categorized by low (0–23 μmol/L) and high (>23 μmol/L) serum TBIL levels on the first day. (F) Core body temperature (T_c) of HS mice during HS and the recovery phase (n = 4 mice per group). (G) ALT, (H) AST, (I) TBIL, (J) DBIL, (K) LDH, and (L) Heme levels in the plasma of HS mice recovery 24 h (n = 6 mice per group). (M) Representative histological images of liver paraffin sections stained with Hematoxylin and Eosin (H&E) (scale bar: 200 μm). (N) Representative immunofluorescence staining images of KCs labeled with Clec4F (purple) and DAPI (blue) in the liver of mice after HS (scale bar: 50 μm), with statistical results of quantification of Clec4F fluorescent positive regions (n = 4) (O). (P) Plasma IL-1β, IL-6, IL-10, and TNF-α content measurement by ELISA (n = 3–4 mice per group). (Q) Representative immunofluorescence staining images labeled with IL-1β (red), Clec4F (purple), and DAPI (blue) in the liver of mice after HS, depicting co-localization (scale bar: 50 μm), with statistical results of quantification of IL-1β fluorescent positive regions (n = 5) (R). (S) Cell death detection (PI positive) in KCs pretreated with different inhibitors followed by heat stress (n = 3). (T) Schematic diagram illustrating the determination of KCs depletion through Clodronate Liposomes (CLP) and Control Liposomes (CTR). (U) Survival curves of mice pretreated with 200 μL CTR or CLP followed by HS on Day 7 (n = 15 mice per group). (V) Representative histological images of liver paraffin sections from HS mice after KCs depletion stained with H&E (scale bar: 200 μm) and labeled with IL-1β (red), Clec4F (purple), and DAPI (blue) (scale bar: 10 μm) in the liver. (W) Representative images of BODIPY 581/591 C11 staining in liver tissue of HS mice pretreated with CLP or CTR (scale bar: 200 μm). (X) Representative images from the transmission electron microscope showing mitochondrial changes in the liver of HS mice, with white arrows indicating smaller mitochondria, crest breaks, and deep staining (scale bar: 1 μm). Data are presented as the mean ± SEM. Paired t-tests were used in TBIL of Day 1 and Day 2 experiments. Significance was assessed by the Student t test or one-way ANOVA with Tukey's post hoc test. Significance in (E) and (U) determined log-rank (Mantel–Cox) test.

Figure 2

Unbiased approaches reveal macrophage heterogeneity with single-cell RNA sequencing. (A) tSNE plot displaying single-cell RNA sequencing data, with each color representing a distinct cluster. (B) tSNE plots illustrating cell subtype annotation. (C) Key marker genes for the respective cell subtypes. (D) tSNE plot showing the cell status of each sample. (E) Histogram depicting the proportion of cell subtypes in four sample groups: NC, HS at 0, 6, and 24 h. (F) KEGG pathway analysis was specific to the KC2 subtype. (G) tSNE graph and (H) violin graph displaying Clec4f and Mrc1 (CD206) expression in all macrophage subclusters. (I) Immunofluorescence technology monitoring the expression of Clec4F and CD206 in KCs of normal mice. Clec4F⁺/CD206^– cells labeled as “a” belong to KC1, while Clec4F⁺/CD206⁺ cells labeled as “b” belong to KC2 (scale bar: 10 μm). (J) Flow cytometry monitoring the frequency of CD206⁺ KCs in HS mice recovered at 0, 6, and 24 h. (K) Statistical results (n = 3 mice per group) for CD206⁺ KC frequency. Data are presented as the mean ± SEM, and statistical significance was determined by one-way ANOVA with Tukey's test.

Figure 3

KC2 shows distinct susceptibility to ferroptosis. (A) Heat maps illustrate the expression of genes related to iron metabolism, lipid metabolism, and oxidant metabolism in KC1 and KC2. (B) Gene Set Enrichment Analysis (GSEA) plot showing the enrichment of gene sets in KC1 and KC2. (C) Gene Ontology (GO) pathway analysis of KC1 and KC2 from NC and HS-recovered mice at 6 h. (D) Measurement of relative mRNA levels of genes from (A) in KC2 subjected to heat stress (n = 4). (E) Representative images showing Clec4F (red), CD206 (cyan blue), and HMOX-1 (green) in liver paraffin sections of HS mice (scale bar: 10 μm), with statistical results (n = 3) (F). (G)Flow cytometry was used to detect HMOX-1 expression in Clec4F⁺/CD206⁺ and Clec4F⁺/CD206^– cells in liver tissues, with statistical analysis shown in (n = 3 mice per group) (H). Data are presented as the mean ± SEM, and statistical significance was determined using a one-way ANOVA with Tukey's post hoc test.

Figure 4

HMOX-1-specific targeting in KCs induces ferroptosis. (A) Schematic diagram illustrating the influence of HMOX-1 through ZnPP, CoPP, and CoPP + DFO. (B) Survival curves of mice pretreated with 10 mg/kg ZnPP, 5 mg/kg CoPP, or 5 mg/kg CoPP + 100 mg/kg DFO followed by HS (n = 15 mice per group). Evaluation of hepatocellular function by AST (C), ALT (D), and TBIL (E) (n = 3–5 mice per group). (F) Representative H&E staining of liver paraffin sections from HS mice (scale bar: 200 μm). Detection of MDA (G), tissue non-heme iron (H) (n = 3–5 mice per group), and BODIPY 581/591 C11 (I) in liver tissue of HS mice (scale bar: 200 μm). (J) Schematic diagram illustrating the influence of HMOX-1 through specific Hmox1 knockout in KCs. (K) Survival curves of Clec4f-crexHmox1^flox/flox mice or Hmox1^flox/flox mice, followed by HS (n = 15 mice per group). Measurement of ALT (L), AST (M), and TBIL (N) (n = 3–5 mice per group). (O) Representative H&E staining of liver paraffin sections of Clec4f-crexHmox1^flox/flox mice or Hmox1^flox/flox mice (scale bar: 200 μm). Measurement of MDA (P), tissue non-heme iron (Q) (n = 3–5 mice per group), and BODIPY 581/591 C11 (R) in liver tissue (scale bar: 200 μm). Summary data are presented as the mean ± SEM. Significance was calculated using a one-way ANOVA with Tukey's post hoc test. Significance in (B) and (K) is determined by the log-rank (Mantel–Cox) test.

Figure 5

HMOX-1 affects NLRP3 inflammasome activation. (A) Representative immunofluorescence staining images displaying the expression of Clec4F (purple), HMOX-1 (green), NLRP3 (red), and in the liver of mice after HS, with magnified insets (scale bar: 10 μm). (B) Western Blotting analysis of Caspase-1 in liver tissues of HS mice and statistical results (n = 3) (C). (D) Representative images of immunofluorescence staining for Clec4F (purple), HMOX-1 (green), and NLRP3 (red) in liver tissues of HS mice pretreated with PBS, 10 mg/kg ZnPP, 5 mg/kg CoPP, or 5 mg/kg CoPP + 100 mg/kg DFO, with magnified insets (scale bar: 10 μm). (E) estern blotting analysis of caspase-1 in liver tissues of HS mice and statistical results (n = 3) (F). (G) Measurement of plasma IL-1β content by ELISA (n = 3–5 mice per group). (H) Representative images of immunofluorescence staining for Clec4F (purple), HMOX-1 (green), and NLRP3 (red) in liver tissues of Clec4f-crexHmox1^flox/flox mice or Hmox1^flox/flox mice, with magnified insets (scale bar: 10 μm). (I) Western blotting analysis of caspase-1 in liver tissues and statistical results (n = 3) (J). (K) Measurement of plasma IL-1β content by ELISA (n = 3–5 mice per group). Significance was calculated using a one-way ANOVA with Tukey's post hoc test.

Figure 6

HMOX-1 affects PI4Kβ activation in vitro. (A) Representative images of immunofluorescence staining for OSBP-PH-GFP (green), TGN38 (red), and DAPI (blue) in KC2 treated at 43 °C for 3 h and recovered at 37 °C for 0, 6, or 24 h (scale bar: 5 μm), and statistical analysis (B) of co-localization of OSBP-PH-GFP and TGN38 (n = 20). (C) Representative images of immunofluorescence staining for OSBP-PH-GFP (green), TGN38 (red), and DAPI (blue) in KC2 pretreated with 4 μmol/L ZnPP for 12 h or 4 μmol/L CoPP for 12 h (scale bar: 5 μm), and statistical analysis (D) of co-localization of OSBP-PH-GFP and TGN38 (n = 20). (E) Representative images of immunofluorescence staining for OSBP-PH-GFP (green), PI4Kβ (red), and DAPI (blue) in KC2 (scale bar: 5 μm), and statistical analysis (F) of co-localization of OSBP-PH-GFP and PI4Kβ (n = 20). (G) Representative images of immunofluorescence staining for OSBP-PH-GFP (green), TGN38 (red), and DAPI (blue) in KC2 pretreated with 5 μmol/L PI4Kβ inhibitor for 1 h (scale bar: 5 μm), and statistical analysis (H) of co-localization of OSBP-PH-GFP and TGN38 (n = 20). (I) Representative images of immunofluorescence staining for NLRP3 (red), TGN38 (green), and DAPI (blue) in KC2 (scale bar: 5 μm), and statistical analysis (J) of co-localization of OSBP-PH-GFP and TGN38 (n = 20). (K) Representative images of immunofluorescence staining for PI4Kβ (red), TGN38 (green), and DAPI (blue) in KC2 pretreated with 4 μmol/L ZnPP for 12 h, 4 μmol/L CoPP for 12 h or 4 μmol/L CoPP for 12 h + 1 μmol/L Fer-1 for 1 h (scale bar: 5 μm), and statistical analysis (L) of co-localization of PI4Kβ and TGN38 (n = 20). (M) Western blotting analysis of caspase-1 in KC2 pretreated with DMSO or PI4Kβ inhibitor and (N) statistical results (n = 3). (O) Representative images of immunofluorescence staining for PI4Kβ (red), TGN38 (green), and DAPI (blue) in vector, sh-Hmox1 or OE-Hmox1 treated with 43 °C for 3 h and recovered at 37 °C for 6 h (scale bar: 5 μm) and (P) statistical analysis of co-localization of PI4Kβ and TGN38 (n = 20). Summary data are presented as the mean ± SEM. Significance was calculated using a one-way ANOVA with Tukey's post hoc test.

Figure 7

EGR1 regulates the transcription of Hmox1. (A) Representative images of immunofluorescence staining for HMOX-1 (red), EGR1 (green), and DAPI (blue) in KC2 treated at 43 °C for 3 h and recovered at 37 °C for 0, 6, or 24 h (scale bar: 5 μm). (B) Western blotting analysis of the expression of HMOX-1 after knocking down Egr1 (sh-Egr1) in ImKCs and statistical analysis (n = 3) (C). (D) Representative images of immunofluorescence staining for PI4Kβ (red), TGN38 (green), and DAPI (blue) in vector or sh-Egr1 (scale bar: 5 μm). (E) Representative images of immunofluorescence staining for NLRP3 (red), TGN38 (green) and DAPI (blue) (scale bar: 5 μm). (F) Western blotting analysis of caspase-1. (G) Cellular supernatant IL-1β content detected by ELISA (n = 3 mice per group). (H) Dual-luciferase experiments in HEK-293T cells transfected with different truncated regions of the Hmox1 promoter (n = 3). (I) Dual-luciferase experiments in HEK-293T cells transfected with different mutational regions of the Hmox1 promoter (n = 3). ChIP-qRT-PCR analysis of EGR1 binding to −103 to −93, −1300 to −1290, −1849 to −1839 sites in the Hmox1 promoter region (n = 3) (J) and fold change of signals in ImKC treated with 43 °C for 3 h and recovered at 37 °C for 6 h (n = 3) (K). (L) A hypothetical model for KC2 ferroptosis in HS mice. Summary data are presented as the mean ± SEM. Significance was calculated using a one-way ANOVA with Tukey's post hoc test.