Gasdermin D inhibition prevents multiple organ dysfunction during sepsis by blocking NET formation

Abstract

Multiple organ dysfunction is the most severe outcome of sepsis progression and is highly correlated with a worse prognosis. Excessive neutrophil extracellular traps (NETs) are critical players in the development of organ failure during sepsis. Therefore, interventions targeting NET release would likely effectively prevent NET-based organ injury associated with this disease. Herein, we demonstrate that the pore-forming protein gasdermin D (GSDMD) is active in neutrophils from septic humans and mice and plays a crucial role in NET release. Inhibition of GSDMD with disulfiram or genic deletion abrogated NET formation, reducing multiple organ dysfunction and sepsis lethality. Mechanistically, we demonstrate that during sepsis, activation of the caspase-11/GSDMD pathway controls NET release by neutrophils during sepsis. In summary, our findings uncover a novel therapeutic use for disulfiram and suggest that GSDMD is a therapeutic target to improve sepsis treatment.

Graphical abstract

Figure 1.

Gasdermin D contributes to organ dysfunction in sepsis by NET release. (A) The MPO/DNA-NET concentrations in the plasma from WT and Gsdmd^−/− mice were determined at 6, 12, or 24 hours after sepsis induction by CLP. (B) Bone marrow neutrophils from WT and Gsdmd^−/− mice were primed with PAM3CSK4 (1 µg/mL) for 4 hours and then transfected with ultrapure LPS (10 µg/mL) for 4 hours. Representative fluorescence images of NETs stained for DNA (DAPI, blue), myeloperoxidase (MPO, green), and the gasdermin D cleaved fraction (GSDMD, red) are shown. Scale bar, 50 µm at ×630 magnification. (C) The concentrations of MPO/DNA-NETs in the neutrophil culture supernatants after 4 hours of stimulation were determined using the picogreen test. (D-E) The plasma levels of the organ injury markers CK-MB and BUN and (F-G) TNF-α and IL-1β were determined at 6, 12, and 24 hours after sepsis induction by CLP. (H) Representative histopathology images of lung tissue sections 24 hours after sepsis induction by CLP are shown at ×200 magnification. The square insets represent the image at ×400 magnification. (I) Blood pressure was continuously monitored by telemetry, which was used to calculate the MAP for 24 hours after sepsis induction by CLP. (J) WT and Gsdmd^−/− mice were subjected or not to CLP, and survival was recorded for 7 days. The data are expressed as means ± SEM. *P < .05; WT CLP vs Gsdmd^−/− CLP, Student t test (A,D-I), 1-way ANOVA followed by Tukey's test (C); H-Mantel-Cox log-rank test (J). The data are representative of ≥2 independent experiments, each including 5-7 animals per group.

Figure 2.

Gasdermin D expression in neutrophils is essential for organ dysfunction during polymicrobial sepsis. (A) Schematic representation of the cell transfer experiment in which neutrophils (1 × 10⁷ cells per mouse) or monocytes (1 × 10⁷ cells per mouse) from WT animals were transferred to Gsdmd^−/− mice via IV injection 1 hour after the mice were subjected to CLP and 12 hours after the animals were euthanized. (B) Circulating concentrations of MPO/DNA-NETs were quantified 12 hours after sepsis induction by CLP. (C-F) The plasma levels of organ injury markers and (G-I) cytokines TNF-α, IL-6, and IL-1β were determined 12 hours after sepsis induction by CLP. (B-I) The dashed lines indicate the upper and lower levels of the mediators measured in the control sham groups. The data are expressed as means ± SEM. *P < .05; 1-way ANOVA followed by Tukey’s test (B-I). The data are representative of ≥2 independent experiments, each including 4 to 6 animals per group.

Figure 3.

Caspase-11 contributes to NET release through gasdermin D. (A-D) The survival of WT, Aim2^−/−, Nlrc4^−/−, Casp1^−/−, and Casp11^−/− mice subjected or not to CLP was followed for 7 days. (E) Bone marrow neutrophils from WT and Casp11^−/− mice were primed with PAM3CSK4 (1 μg/mL) for 4 hours and then transfected with ultrapure LPS (10 μg/mL) for 4 hours. The cell lysates were harvested for immunoblot analysis of GSDMD total and its cleaved fraction (GSDMD-N). Each band represents 1 different animal. Actin (β-actin) was used as a loading control. (F) Quantification of GSDMD-N relative to β-actin. Bone marrow neutrophils from WT and Casp11^−/− mice were primed with PAM3CSK4 (1 µg/mL) for 4 hours and then transfected with ultrapure LPS (10 µg/mL) for 4 hours. (G) Representative fluorescence images of NETs stained for DNA (DAPI, blue), myeloperoxidase (MPO, green), and the gasdermin D cleaved fraction (GSDMD, red) are shown. Scale bar, 50 µm at ×630 magnification. (H) The concentration of MPO/DNA-NETs was determined in the culture supernatants. The data are expressed as means ± SEM. *P < .05; H-Mantel-Cox log-rank test (A-D), 1-way ANOVA followed by Tukey’s test (F,H). The data are representative of ≥2 independent experiments, each including 5 to 7 animals per group.

Figure 4.

Caspase-11 deletion prevents NET release and organ dysfunction during polymicrobial sepsis. (A) Circulating amounts of MPO/DNA-NETs were quantified 24 hours after sepsis induction by CLP. (B-F) Plasma levels of organ injury markers. (G) Representative images of H&E staining of lung tissue sections from WT and Casp11^−/− mice 24 hours after sepsis induction by CLP are shown at ×200 magnification. The square insets represent the image at ×400 magnification. (H-J) The plasma levels of the cytokines TNF-α, IL-6, and IL-1β (G-I) were determined 24 hours after sepsis induction by CLP. (K-L) WT and Casp11^−/− mice were implanted using a telemetric pressure transmitter probe to determine the MAP and HR 24 hours after CLP-induced sepsis. The data are expressed as means ± SEM. *P < .05; 1-way ANOVA followed by Tukey’s test (A-F,H-J), CLP WT vs CLP Casp11^−/− Student t test (K,L). The data are representative of ≥ 2 independent experiments, each including 4 to 7 animals per group.

Figure 5.

Disulfiram, a gasdermin D inhibitor, protects against CLP-induced sepsis by reducing NET release. (A) Bone marrow neutrophils were treated with disulfiram (30 µM) or vehicle 1 hour before stimulation with PAM3CSK4 (1 µg/mL) for 4 hours and then were transfected with ultrapure LPS (10 µg/mL) for an additional 4 hours. Representative fluorescence images of NETs stained for DNA (DAPI, blue), myeloperoxidase (MPO, green), and the gasdermin D cleaved fraction (GSDMD, red) are shown. Scale bar, 50 µm at ×630 magnification. (B) The mice were pretreated with disulfiram (80 mg/kg) or vehicle by subcutaneous injection 24 and 4 hours before surgery and at 6 and 18 hours after CLP. The MPO/DNA-NET concentration in the plasma was determined 24 hours after CLP-induced sepsis. (C-F) The circulating levels of organ injury markers were determined 24 hours after sepsis induction by CLP. (G) Representative images of the histologic staining of the lung sections performed 24 hours after sepsis induction are shown. Scale bar 200 μm at ×400 magnification. (H-J) The systemic levels of TNF-α, IL-6, and IL-1β were measured by enzyme-linked immunosorbent assay 24 hours after CLP-induced sepsis. (K) The mice were pretreated with disulfiram (80 mg/kg) or vehicle by SC injection 24 and 4 hours before CLP-induced sepsis, 6 hours after CLP, and every 12 hours for 2 days and then were followed for survival analysis. The data are expressed as means ± SEM. *P < .05; 1-way ANOVA followed by Tukey’s test (B-F,H-J), H-Mantel-Cox log-rank test (K). The data are representative of ≥2 independent experiments, each including 5 to 7 animals per group.

Figure 6.

Neutrophils from septic patients who undergo NETosis express gasdermin D. (A) Gene set enrichment analysis identified “neutrophil degranulation,” “neutrophil-mediated immunity,” and “neutrophil activation involved in immune response gene sets” as the top 3 GO terms with the highest normalized enrichment score in nonsurvivor patients with septic shock at day 1 after ICU admission compared with healthy volunteers (adjusted P < .1). (B) Expression levels of Gsdmd in healthy volunteers and patients with septic shock (adjusted P < .1). (C) Heatmap representation of differentially expressed genes in surviving and nonsurviving patients with septic shock compared with healthy volunteers at days 1, 2, and 3 after ICU admission (adjusted P < .1). (D) Circulating amounts of MPO/DNA-NETs were quantified from the plasma of patients with sepsis and healthy volunteers. (E) Neutrophils from patients with sepsis and healthy volunteers were isolated and cultured for 4 hours at 37°C. Representative immunostaining images for DNA (DAPI, gray), myeloperoxidase (MPO, green), and the gasdermin D cleaved fraction (GSDMD, red) are shown. Scale bar, 50 µm at ×630 magnification. (F) GSDMD expression was quantified by Median Fluorescence Intensity (MFI) per field. (G) Magnified fluorescence microscopy images of panel E for DNA (DAPI, gray) and the gasdermin D cleaved fraction (GSDMD, red) are shown. Scale bar, 50 µm at ×630 magnification. Furthermore, to demonstrate the details of GSDMD in neutrophil membranes and its association with NETs (G, top), we performed a 4× digital zoom of the inset square (G, bottom). (H) Neutrophils from patients with sepsis and healthy volunteers were isolated, treated with disulfiram (30 µM) or vehicle, and cultured for 4 hours at 37°C. Representative fluorescence images of NETs stained for DNA (DAPI, gray), myeloperoxidase (MPO, green), and the gasdermin D cleaved fraction (GSDMD, red) are shown. Scale bar, 50 µm at ×630 magnification. (I) The concentrations of MPO/DNA-NETs in the neutrophil culture supernatant after 4 hours were determined using the picogreen test. The data are expressed as means ± SEM. *P < .05; Student t test (D), 1-way ANOVA, followed by Tukey’s test (I). The data are representative of ≥2 independent experiments.