Hepatic pannexin-1 mediates ST2+ regulatory T cells promoting resolution of inflammation in lipopolysaccharide-induced endotoxemia

Abstract

Sepsis remains the most lethal infectious disease and substantially impairs patient prognosis after liver transplantation (LT). Our previous study reported a role of the pannexin 1 (PANX1)-interleukin-33 (IL-33) axis in activating innate immunity to protect against methicillin-resistant Staphylococcus aureus infection; however, the role of PANX1 in regulating adaptive immunity in sepsis and the underlying mechanism are unclear. In this study, we examined the role of the PANX1-IL-33 axis in protecting against sepsis caused by a gram-negative bacterial infection in an independent LT cohort. Next, in animal studies, we assessed the immunological state of Panx1^-/- mice with lipopolysaccharide (LPS)-induced endotoxemia and then focused on the cytokine storm and regulatory T cells (Tregs), which are crucial for the resolution of inflammation. To generate liver-specific Panx1-deficient mice and mimic clinical LT procedures, a mouse LT model was established. We demonstrated that hepatic PANX1 deficiency exacerbated LPS-induced endotoxemia and dysregulated the immune response in the mouse LT model. In hepatocytes, we confirmed that PANX1 positively regulated IL-33 synthesis after LPS administration. We showed that the adenosine triphosphate-P2X7 pathway regulated the hepatic PANX1-IL-33 axis during endotoxemia in vitro and in vivo. Recombinant IL-33 treatment rescued LPS-induced endotoxemia by increasing the numbers of liver-infiltrating ST2⁺ Tregs and attenuating the cytokine storm in hepatic PANX1-deficient mice. In conclusion, our findings revealed that the hepatic PANX1-IL-33 axis protects against endotoxemia and liver injury by targeting ST2⁺ Tregs and promoting the early resolution of hyperinflammation.

Keywords: ATP; IL-33; Panx1; Tregs; endotoxemia.

FIGURE 1

The PANX1–IL‐33 axis was associated with sepsis in liver transplantation (LT) patients. (A) Experimental design of this study. (B and C) Representative images and quantification of immunohistochemistry (IHC) staining for PANX1, which was evaluated blindly by two pathologists, in an independent cohort comprising 13 donor grafts of septic recipients and 26 grafts of uninfected controls that were obtained during liver graft procurement (scale bar = 50 μm). (D and E) Immunoblot and relative quantification of PANX1 GLY2 in donor graft tissue samples from patients with sepsis (n = 6) and control subjects (n = 6) prior to liver transplantation (LT). (F and G) Serum IL‐33 levels measured one week after diagnosis in the sepsis group and at matched time points in the control group. (H) Correlation between circulating IL‐33 levels after LT and the PANX1 staining score of donor graft tissue samples prior to LT (n = 13). (I) Peak alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels in patients with sepsis (n = 13) compared with control subjects (n = 26). (J) Correlation between the peak AST levels measured during sepsis after LT and the PANX1 staining score of donor graft tissue samples prior to LT (n = 13). (K) Correlation between circulating IL‐33 levels after LT and peak AST levels measured during sepsis after LT (n = 13). *p < 0.05 and ***p < 0.001 by Student's t‐test or one‐way analysis of variance (ANOVA). Correlations were analyzed using Pearson's correlation test

FIGURE 2

Immune cell and cytokine levels in septic patients after liver transplantation (LT). (A) Blood leukocyte numbers and proportions of (B) neutrophils, (C) lymphocytes, and (D) monocytes from septic patients in the deceased group (n = 8) and the discharged group (n = 16). Serum levels of (E) IL‐6, (F) TNF‐α, (G) IL‐10, (H) IL‐1β, (I) IL‐2R and (J) IL‐33 in septic patients in the deceased group (n = 8) and discharged group (n = 16). Proportions of (K) CD3⁺ T cells, (L) CD3⁺CD4⁺ T cells, (M) CD3⁺CD8⁺ T cells, (N) CD3⁺HLA‐DR⁺ T cells, (O) CD4⁺CD25⁺CD127⁻ regulatory T cells, (P) CD3⁻CD19⁺ B cells, and (Q) CD3⁻CD56⁺ NK cells in the blood of septic patients in the deceased group (n = 8) and discharged group (n = 16). *p < 0.05 by Student's t‐test

FIGURE 3

Protective role of hepatic PANX1 in endotoxemia and the liver in vivo. (A) Survival (n = 6 mice per group), (B) rectal temperatures (n = 5 mice per group), and (C) serum ALT and AST levels (n = 5 mice per group) of Panx1^−/‐ and WT mice after injection with 15 mg/kg lipopolysaccharide (LPS). (D) Survival (n = 6 mice per group), (E) rectal temperatures, and (F) serum ALT and AST levels of WT → WT, Panx1^−/− → wild‐type (WT), WT → Panx1^−/−, and Panx1^−/− → Panx1^−/− mice at 24 h after injection with 15 mg/kg LPS. (G and H) Representative images and quantification of hematoxylin and eosin staining and (I) quantification of TUNEL staining and IHC staining for cleaved caspase‐3, caspase‐8 and caspase‐9 in liver tissue samples from the four groups of mice 24 h after injection with 15 mg/kg LPS (scale bar = 50 μm). The quantified results are the average of 10 images per mouse (n = 5 mice per group). *p < 0.05, **p < 0.01, and ***p < 0.001 by Student's t‐test, one‐way ANOVA, or log‐rank test

FIGURE 4

Hepatic PANX1‐IL‐33 axis deficiency results in dysregulated immune responses. Serum levels of (A) interleukin 6 (IL‐6), (B) tumour necrosis factor‐alpha (TNF‐α), (C) IFN‐γ, and (D) IL‐33 in wild‐type (WT) → WT, Panx1^−/− → WT, WT → Panx1^−/−, and Panx1^−/− → Panx1^−/− mice at 24 h after injection with 15 mg/kg LPS (n = 5 mice per group). (E–G) Representative images and quantification of IHC staining for IL‐33 in the livers of the four groups at 24 h after injection with 15 mg/kg lipopolysaccharide (LPS). Staining of the livers from IL‐33^−/‐ mice was used as a negative control. The quantified results are the average of 10 images per mouse (scale bar = 50 μm) (n = 5 mice per group). (H and I) Quantification of IHC staining for F4/80⁺ and Foxp3⁺ in the livers of the four groups at 24 h after injection with 15 mg/kg LPS. The quantified results are the average of 10 images per mouse (n = 5 mice per group). (J and K) Proportions of CD4⁺ Foxp3⁺ cells on CD3⁺ cells in the blood of mice from the four groups at 24 h after injection with 15 mg/kg LPS (n = 5 mice per group). *p < 0.05, **p < 0.01, and ***P < 0.001 by Student's t‐test or one‐way ANOVA

FIGURE 5

PANX1 positively regulates IL‐33 synthesis in hepatocytes after LPS administration. (A and B) Proportions of IL‐33⁺ primary mouse hepatocytes at 12 h after lipopolysaccharide (LPS) stimulation (1 μg/ml) (n = 5 per group). (C) Immunoblot and quantification of IL‐33 in primary WT hepatocytes and Panx1^−/− hepatocytes from mice at 12 and 24 h after LPS stimulation (1 μg/ml) (n = 4 per group). (D‐N) Immunoblots and quantification of the levels of IL‐33 and intermediates in the PANX1‐P2Xs‐NLRP3‐caspase‐1‐IL‐1 pathway in L02 cells treated with ¹⁰Panx (200 μM), scrambled ¹⁰Panx (200 μM), probenecid (50 μM) or vehicle at 0, 4, 8, 12, 24 and 48 h after LPS stimulation (1 μg/ml) (n = 4 per group). *p < 0.05, **p < 0.01 by Student's t‐test or ANOVA. Correlations were analyzed using Pearson's correlation test

FIGURE 6

The ATP/P2X7 pathway regulated the hepatic PANX1–IL‐33 axis in endotoxemia in vitro. (A) Extracellular adenosine triphosphate (ATP) concentrations in the supernatants of primary WT hepatocytes and Panx1^−/− hepatocytes at various time points after lipopolysaccharide (LPS) stimulation (1 μg/ml) (n = 5 per group). (B) ATP induced IL‐33 secretion by wild‐type (WT) hepatocytes stimulated with LPS (1 μg/ml) for 12 h (n = 4). (C) IL‐33 levels in the supernatants of LPS‐primed (1 μg/ml) primary WT and Panx1^−/− hepatocytes treated with ATP (0.25 mM), P2X1 inhibitor (10 μM), P2X2/3 inhibitor (50 μM), P2X3 inhibitor (50 μM), P2X4 inhibitor (10 μM), or P2X7 inhibitor (50 μM) alone or in combination for 12 h (n = 4 per group). (D–H) Immunoblots and quantification of IL‐33 expression in LPS‐primed (1 μg/ml) primary WT and Panx1^−/− hepatocytes stimulated with ATP (0.25 mM), P2X7 inhibitor (50 μM) or both for 12 h (n = 4 per group). **p < 0.01 and ***p < 0.001 by Student's t‐test or ANOVA

FIGURE 7

The ATP‐P2X7 pathway regulated the hepatic PANX1–IL‐33 axis in endotoxemia in vivo. (A) Adenosine triphosphate (ATP) concentrations in liver tissue interstitial fluid from wild‐type (WT) and Panx1^−/− mice at 24 h after injection with 15 mg/kg lipopolysaccharide (LPS) (n = 5 mice per group). (B) mRNA levels of P2xs in wild‐type (WT) and Panx1^−/− mice at 24 h after injection with 15 mg/kg LPS (n = 4 per group). (C) Quantification of IHC staining for P2Xs in liver tissue samples from WT and Panx1^−/− mice at 24 h after injection with 15 mg/kg LPS. The quantified results are the average of 10 images per mouse (n = 5 mice per group). (D–F) Immunoblots and quantifications of P2X7 and PANX1 expression in the liver and kidney of WT mice treated with adeno‐associated virus (AAV)‐Panx1‐shRNA, AAV‐P2x7‐shRNA or both at 2 weeks. (G) Survival (n = 6 mice per group) and (H) serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels (n = 5 mice per group) of control mice and mice treated with AAV‐P2x7‐shRNA, AAV‐Panx1‐shRNA, or both at 24 h after injection with 15 mg/kg LPS. (I and J) Representative images of hematoxylin and eosin staining and pathology scoring (scale bar = 50 μm) (n = 5 mice per group). (K and L) Representative images and quantification of IHC staining of IL‐33 in the liver tissue samples of control mice and mice treated with AAV‐P2x7‐shRNA, AAV‐Panx1‐shRNA, or both at 24 h after injection with 15 mg/kg LPS. The quantified results are the average of 10 images per mouse (scale bar = 50 μm) (n = 5 mice per group). **p < 0.01 and ***p < 0.001 by Student's t‐test or ANOVA

FIGURE 8

rIL‐33 rescued LPS‐induced endotoxemia by increasing liver‐infiltrating ST2⁺ Treg numbers in hepatic PANX1‐deficient mice. (A) Survival and (B) serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels of wild‐type (WT) → WT mice (n = 6) and Panx1^−/− → WT mice treated with vehicle (n = 6) or rIL‐33 (n = 7) at 24 h after injection with 15 mg/kg LPS. (C–E) Representative images and quantification of hematoxylin and eosin staining and TUNEL staining of the liver tissue samples of mice from the three groups at 24 h after LPS injection. The quantified results are the average of 10 images per mouse (scale bar = 50 μm) (n = 5 mice per group). Proportions of liver‐infiltrating CD4⁺ Foxp3⁺ Tregs (F and G) and ST2⁺ Tregs (H and I) in mice from the three groups at 24 h after injection with 15 mg/kg LPS (n = 5 mice per group). Serum levels of IL‐6 (J), TNF‐α (K), and IFN‐γ (L) in mice from the three groups at 24 h after injection with 15 mg/kg LPS (n = 5 mice per group). *p < 0.05, **p < 0.01, and ***p < 0.001 by Student's t‐test, one‐way ANOVA, or log‐rank test

FIGURE 9

Schematic of the mechanism by which the hepatic PANX1‐IL‐33 axis is regulated by the ATP‐P2X7 pathway and targets ST2⁺ Tregs to attenuate LPS‐induced endotoxemia