Inhibition of the NLRP3/IL-1β axis protects against sepsis-induced cardiomyopathy

Abstract

Background: Septic cardiomyopathy worsens the prognosis of critically ill patients. Clinical data suggest that interleukin-1β (IL-1β), activated by the NLRP3 inflammasome, compromises cardiac function. Whether or not deleting Nlrp3 would prevent cardiac atrophy and improve diastolic cardiac function in sepsis was unclear. Here, we investigated the role of NLRP3/IL-1β in sepsis-induced cardiomyopathy and cardiac atrophy.

Methods: Male Nlrp3 knockout (KO) and wild-type (WT) mice were exposed to polymicrobial sepsis by caecal ligation and puncture (CLP) surgery (KO, n = 27; WT, n = 33) to induce septic cardiomyopathy. Sham-treated mice served as controls (KO, n = 11; WT, n = 16). Heart weights and morphology, echocardiography and analyses of gene and protein expression were used to evaluate septic cardiomyopathy and cardiac atrophy. IL-1β effects on primary and immortalized cardiomyocytes were investigated by morphological and molecular analyses. IonOptix and real-time deformability cytometry (RT-DC) analysis were used to investigate functional and mechanical effects of IL-1β on cardiomyocytes.

Results: Heart morphology and echocardiography revealed preserved systolic (stroke volume: WT sham vs. WT CLP: 33.1 ± 7.2 μL vs. 24.6 ± 8.7 μL, P < 0.05; KO sham vs. KO CLP: 28.3 ± 8.1 μL vs. 29.9 ± 9.9 μL, n.s.; P < 0.05 vs. WT CLP) and diastolic (peak E wave velocity: WT sham vs. WT CLP: 750 ± 132 vs. 522 ± 200 mm/s, P < 0.001; KO sham vs. KO CLP: 709 ± 152 vs. 639 ± 165 mm/s, n.s.; P < 0.05 vs. WT CLP) cardiac function and attenuated cardiac (heart weight-tibia length ratio: WT CLP vs. WT sham: -26.6%, P < 0.05; KO CLP vs. KO sham: -3.3%, n.s.; P < 0.05 vs. WT CLP) and cardiomyocyte atrophy in KO mice during sepsis. IonOptix measurements showed that IL-1β decreased contractility (cell shortening: IL-1β: -15.4 ± 2.3%, P < 0.001 vs. vehicle, IL-1RA: -6.1 ± 3.3%, P < 0.05 vs. IL-1β) and relaxation of adult rat ventricular cardiomyocytes (time-to-50% relengthening: IL-1β: 2071 ± 225 ms, P < 0.001 vs. vehicle, IL-1RA: 564 ± 247 ms, P < 0.001 vs. IL-1β), which was attenuated by an IL-1 receptor antagonist (IL-1RA). RT-DC analysis indicated that IL-1β reduced cardiomyocyte size (P < 0.001) and deformation (P < 0.05). RNA sequencing showed that genes involved in NF-κB signalling, autophagy and lysosomal protein degradation were enriched in hearts of septic WT but not in septic KO mice. Western blotting and qPCR disclosed that IL-1β activated NF-κB and its target genes, caused atrophy and decreased myosin protein in myocytes, which was accompanied by an increased autophagy gene expression. These effects were attenuated by IL-1RA.

Conclusions: IL-1β causes atrophy, impairs contractility and relaxation and decreases deformation of cardiomyocytes. Because NLRP3/IL-1β pathway inhibition attenuates cardiac atrophy and cardiomyopathy in sepsis, it could be useful to prevent septic cardiomyopathy.

Keywords: Heart failure; Interleukin-1 beta; NLR family, pyrin domain-containing 3 protein; Sepsis.

Figure 1

Nlrp3 KO mice show less cardiac atrophy and a preserved cardiac function during sepsis. 12‐ to 16‐week‐old male Nlrp3 KO and WT mice were subjected to CLP or sham surgery. (A) Heart weight normalized to tibia length and expressed as percent‐wise change compared with the respective sham group 96 h after CLP or sham surgery. CLP‐treated Nlrp3 KO (n = 16); sham Nlrp3 KO (n = 8), CLP Nlrp3 WT (n = 12), sham Nlrp3 WT (n = 13). (B) Mean myocyte cross‐sectional area (MCSA) determined in haematoxylin and eosin‐stained histological cross sections from hearts of CLP‐treated Nlrp3 KO (n = 6), sham Nlrp3 KO (n = 6), CLP Nlrp3 WT (n = 6) and sham Nlrp3 WT (n = 6) mice (n = 100 myocytes per mouse). Data are presented as mean ± SEM. *P ≤ 0.05; **P ≤ 0.01; ^*** P ≤ 0.001. (C) Results of transthoracic echocardiography of Nlrp3 WT sham (n = 13), Nlrp3 WT CLP (n = 13), Nlrp3 KO sham (n = 18) and Nlrp3 KO CLP (n = 18) mice 24 h after CLP or sham operation. (D) Serum IL‐1β concentration in sham and CLP mice. Nlrp3 WT sham (n = 15), CLP Nlrp3 WT (n = 12), Nlrp3 KO sham (n = 14), Nlrp3 KO CLP (n = 14). Data are presented as mean ± SEM. *P ≤ 0.05; **P ≤ 0.01. qRT‐PCR analysis of Il1b (E) and Il6 (F) expression in hearts of Nlrp3 WT sham (n = 13), Nlrp3 WT CLP (n = 12), Nlrp3 KO sham (n = 8) and Nlrp3 KO CLP (n = 15) mice. mRNA expression was normalized to Gapdh. Data are presented as mean ± SEM. *P ≤ 0.05; **P ≤ 0.01; ^*** P ≤ 0.001; n.s., not significant.

Figure 2

IL‐1β via the IL‐1 receptor reduces contractility, relaxation and deformation of cardiomyocytes. (A–D) Adult rat ventricular cardiomyocytes were treated with vehicle (PBS), recombinant IL‐1β (10 ng/mL) or the IL‐1 receptor antagonist (IL‐1RA) (10 μg/mL), as indicated, and IonOptix measurements were performed. Quantitation of effects on cell shortening (A) and Ca²⁺ transients (B) are shown. The threshold is defined as a reduction of cell shortening by −7.70% and is displayed by the red line. It equates to the mean effect on cell shortening induced by vehicle minus twice the standard deviation of this effect. (C) Time‐to‐50% relengthening. (D) Change of relaxation speed. Values are expressed as mean percentage change from baseline ± SEM. *P ≤ 0.05; **P ≤ 0.01; ^*** P ≤ 0.001. (E–I) Mechanical high‐throughput characterization of HL‐1 cardiomyocytes by RT‐DC was used to investigate the effects of IL‐1β to cell size and deformation. (E) RT‐DC scatter plots of cell size (area, μm²) and deformation of vehicle (PBS, control, n = 2948 cells) and IL‐1β (n = 2554 cells) treated HL‐1 cells. Isoelasticity lines in grey highlight areas of equal elastic Young's modulus. Colour code indicates red (maximum) to blue (minimum) cell density. (F) Contour plot showing 50% (dashed) and 90% (solid) of maximum event density in vehicle (black) and IL‐1β (brown) treated HL‐1 cells. Statistical analysis comparing cells size (median area; G), median deformation (H) and Young's modulus (I) of experimental triplicates from three different days using linear mixed models. Error bars represent SEM of the distribution. *P ≤ 0.05; ^*** P ≤ 0.001.

Figure 3

Principal component analysis and hierarchical clustering of RNA sequencing data. (A) Principal component analysis (PCA) performed using DESeq2 rlog‐normalized RNA‐seq data. Loadings for principal components 1 (PC1), PC2 and PC3 are shown in graph on x‐, y‐ and z‐axis. Genotypes and treatments are indicated. (B) Hierarchical clustering analyses performed using DESeq2 rlog‐normalized RNA‐seq data. Colour code (from white to dark blue) refers to the distance metric used for clustering (dark blue corresponds to the maximum of correlation values). Genotypes and treatments are indicated. (C) Venn diagrams showing the number of twofold upregulated (top panel) and twofold downregulated (bottom panel) genes and their overlap in hearts of septic Nlrp3 WT and Nlrp3 KO mice. (D) Selected enrichment scores of GO term analysis of genes, which were upregulated (left panel) or downregulated (right panel) in hearts of septic Nlrp3 KO compared with Nlrp3 WT mice. (E) Heatmaps of normalized expression values of genes involved in the NF‐κB signalling pathway (left panel) and the autophagy and lysosomal pathway of protein degradation (right panel). Genotypes and treatments as well as the z‐score are indicated.

Figure 4

Sepsis‐induced expression of autophagy genes is attenuated in hearts of Nlrp3 KO mice. (A) qRT‐PCR analysis of Sqstm1, Map 1LC3b, Ctsl, Atg13 and Bnip3 in hearts of Nlrp3 WT sham (n = 13), Nlrp3 WT CLP (n = 12), Nlrp3 KO sham (n = 8) and Nlrp3 KO CLP (n = 15) mice. mRNA expression was normalized to Gapdh. (B) Western blot analysis with anti‐p62 and anti‐LC3B antibody. Actin was used as loading control. (C) qRT‐PCR analysis of Trim63 and Fbxo32 in hearts of Nlrp3 WT sham, Nlrp3 WT CLP, Nlrp3 KO sham and Nlrp3 KO CLP mice. mRNA expression was normalized to Gapdh. Data are presented as mean ± SEM. *P ≤ 0.05; **P ≤ 0.01; ^*** P ≤ 0.001.

Figure 5

The IL‐1β signalling pathway is contained and active in H9c2 cardiomyocytes. (A) Seven days differentiated H9c2 myotubes were treated with recombinant vehicle, IL‐1β (50 ng/mL) and TNF (5 ng/mL), as indicated, for 25 min. Western blot analysis was performed with anti‐phospho‐NF‐κB p65 and anti‐NF‐κB p65 antibody. GAPDH was used as loading control. (B) Differentiated H9c2 cells were treated with recombinant vehicle, IL‐1β (50 ng/mL) and TNF (5 ng/mL), as indicated, for 15 and 35 min, and NF‐κB‐DNA‐response element complex formation was analysed by EMSA. (C) Differentiated H9c2 cells were treated with recombinant IL‐1β (50 ng/mL), vehicle, the IL‐1 receptor antagonist (IL‐1RA) (10 μg/mL) and the NF‐κB inhibitor BMS‐345541 (25 μM), as indicated. qRT‐PCR analysis of Il6 (left panel) and Nlrp3 (right panel) is shown. mRNA expression was normalized to Gapdh. (D) Differentiated H9c2 cells were treated with solvent or increasing amounts of recombinant IL‐1β (10, 20 and 50 ng/mL) for 72 h. Western blot analysis of isolated proteins with anti‐myosin heavy chain (MyHC)‐slow antibody is shown. GAPDH was used as loading control. (E) Frequency distribution histograms of cell width of vehicle, IL‐1β (50 ng/mL), IL‐1RA (10 μg/mL) and BMS‐345541 (25 μM)‐treated myotubes, as indicated, are shown, n = 100 cells per condition. (F) Mean myotube width. (G) Differentiated H9c2 cells were treated with vehicle or recombinant IL‐1β (50 ng/mL), as indicated, for 72 h. qRT‐PCR analysis of Sqstm1, MapLC3b and Ctsl is shown. mRNA expression was normalized to Gapdh. Data are presented as mean ± SEM. *P ≤ 0.05; **P ≤ 0.01; ^*** P ≤ 0.001.

Figure 6

Proposed mechanism of septic cardiomyopathy. (Left panel) Sepsis, caused by polymicrobial infection leads to an activation of the NLRP3 inflammasome, resulting in increased conversion of inactive pro‐IL‐1ß in active IL‐1ß, which acts on cardiomyocytes via the IL‐1 receptor complex (IL‐1R1/IL‐1RAcP). Binding of IL‐1 to its receptor causes an activation of NF‐κB, which leads to increased protein degradation by the autophagy–lysosomal pathway (ALP), a decrease in myosin heavy chain content (MyHC), a reduced deformability and a decreased cardiomyocyte contraction and relaxation. This in turn causes septic cardiomyopathy with cardiac atrophy, systolic and diastolic cardiac dysfunction and increased expression of pro‐inflammatory cytokines in the heart. (Right panel) In Nlrp3 knockout mice, sepsis can no longer activate the NLRP3 inflammasome, generation of IL‐1ß is much reduced, and activation of NF‐κB as well as NF‐κB signalling is attenuated, which maintains cardiomyocyte size, deformability, contraction and relaxation. As a result, Nlrp3 knockout mice show a normal cardiac function in sepsis without cardiac atrophy and preserved expression of pro‐inflammatory cytokines. Inhibitors of the NLRP3 inflammasome (e.g. haemin, scutellarin, glyburide and MCC950), IL‐1ß (anti‐IL‐1 antibody; anti‐IL‐1 Ab), IL‐1 receptor antagonist (IL‐1 RA) and NF‐κB (IKKβ‐directed NF‐κB inhibitor BMS‐345541) that hold therapeutic potential are indicated.