The Inflammasome Drives GSDMD-Independent Secondary Pyroptosis and IL-1 Release in the Absence of Caspase-1 Protease Activity

Abstract

Inflammasomes activate the protease caspase-1, which cleaves interleukin-1β and interleukin-18 to generate the mature cytokines and controls their secretion and a form of inflammatory cell death called pyroptosis. By generating mice expressing enzymatically inactive caspase-1^C284A, we provide genetic evidence that caspase-1 protease activity is required for canonical IL-1 secretion, pyroptosis, and inflammasome-mediated immunity. In caspase-1-deficient cells, caspase-8 can be activated at the inflammasome. Using mice either lacking the pyroptosis effector gasdermin D (GSDMD) or expressing caspase-1^C284A, we found that GSDMD-dependent pyroptosis prevented caspase-8 activation at the inflammasome. In the absence of GSDMD-dependent pyroptosis, the inflammasome engaged a delayed, alternative form of lytic cell death that was accompanied by the release of large amounts of mature IL-1 and contributed to host protection. Features of this cell death modality distinguished it from apoptosis, suggesting it may represent a distinct form of pro-inflammatory regulated necrosis.

Keywords: ASC; IL-1; caspase-1; caspase-8; gasdermin; inflammasome; pyroptosis; regulated necrosis.

Graphical abstract

Figure 1

Caspase-1 Protease Activity Is Required for Canonical IL-1 Secretion and Pyroptosis (A) Unprimed BMDCs derived from B6.129-Casp1^+/+ and B6.129-Casp1^mlt/mlt mice were stimulated for 6 hr with different TLR and Dectin-1 agonists as indicated or left unstimulated (medium), and IL-6 and TNF secretion were measured in the supernatants by ELISA (data representative of 3 independent experiments). (B) BMDCs from the indicated mouse strains (B6.129-Casp1^mlt/mlt and B6.129-Casp1/11^−/−) were primed with LPS (50 ng/mL for 3 hr) and subsequently stimulated with agents activating the NLRP3 (nigericin and imiquimod), AIM2 (poly(dA:dT)), and NLRC4 (Salmonella enterica serovar Typhimurium) inflammasomes. IL-1β, pro-IL-1β, and IL-1α (top) and LDH (bottom) were quantified from cell-free supernatants by ELISA and a colorimetric assay, respectively. (C) Cleavage and secretion of caspase-1 and IL-1β in BMDCs following inflammasome activation as in (B) were determined by immunoblotting (B6.129-Casp1^mlt/mlt). In (A) and (B), mean ± SEM are shown. In (B and (C), data are representative of >10 independent experiments.

Figure 2

Caspase-1 Protease Activity Is Required for Innate Immunity to Francisella (A–D) Bacterial loads in spleens (A and C) and livers (B and D) of mice infected with Francisella novicida for 48 hr were determined by plating serial dilutions of organ homogenates on selective medium plates. In (C) and (D), the wild-type group consists of littermates of B6.129-Casp1^mlt/mlt mice; B6.129-Casp1^−/− animals were matched by age and sex and were bred in the same isolator. The Mann-Whitney test was used for statistical analysis (^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, ^∗∗∗∗p < 0.0001). (A and B) Spleen: Casp1^−/−, p = 0.0317; Casp1^mlt/mlt, p = 0.0175; liver: Casp1^−/−, p = 0.0303; Casp1^mlt/mlt, p = 0.00012; B6-Casp1^−/− and littermate B6-Casp1^+/+: n = 5; B6-Casp1^mlt/mlt and littermate B6-Casp1^+/+: n = 7. (C and D) Spleen: Casp1^mlt/mlt, p < 0.0001; liver: Casp1^mlt/mlt, p = 0.0076; B6.129-Casp1^mlt/mlt and littermate B6.129-Casp1^+/+: n = 9 and n = 10, respectively; B6.129-Casp1^−/−: n = 4.

Figure 3

Caspase Inhibitors Vary in Their Ability to Prevent GSDMD Cleavage and Pyroptosis (A) BMDCs were pretreated with the indicated doses of the caspase-1 inhibitors Ac-YVAD-cmk and VX-765 for 30 min and stimulated with 5 μM nigericin for 45 min, and secretion of IL-1β and release of LDH into the supernatant were measured (mean ± SEM are shown). (B) BMDCs were pretreated with caspase inhibitors as indicated and stimulated with nigericin. Caspase-1 and GSDMD cleavage were assessed by immunoblot analysis. (C) BMDCs from wild-type, B6.129-Casp1^−/−, and B6.129-Casp1^mlt/mlt mice were stimulated with nigericin, and GSDMD cleavage was analyzed in cytoplasmic fractions by immunoblotting. (D) BMDCs from the indicated mouse strains (B6.129-Casp1^mlt/mlt) were pretreated with caspase inhibitor Z-VAD-fmk (20 μM) or Ac-YVAD-cmk (30 μM) and stimulated with nigericin for 1.5 hr. Secretion of the cleaved and uncleaved forms of IL-1β was assessed by immunoblot analysis from cell-free cell culture supernatants. (A–C) Data representative of ≥3 independent experiments.

Figure 4

Caspase-1^mlt Accumulates at the Inflammasome (A) BMDCs from B6.129-Casp1^mlt/mlt or wild-type mice on culture slides were pretreated with 20 μM Ac-YVAD-cmk where indicated and subsequently stimulated with nigericin. Cells were fixed with paraformaldehyde, stained by immunofluorescence for caspase-1 (green) and ASC (magenta), and analyzed by confocal microscopy. DAPI (blue) localizes with the nuclei. Scale bar represents 20 μm (data representative of >5 independent experiments). (B) BMDCs from B6.129-Casp1^mlt/mlt or wild-type mice on culture slides, the latter pretreated with 20 μM Ac-YVAD-cmk, were stimulated with nigericin, fixed, and stained by immunofluorescence for caspase-1 (green) and ASC (magenta). ASC specks were visualized at high resolution by STED imaging. (C) BMDCs of the indicated genotypes (B6.129-Casp1^−/− and B6.129-Casp1^mlt/mlt) were primed with LPS and subsequently stimulated with nigericin. An IGEPAL CA-630-insoluble fraction was isolated by centrifugation, and cell lysates, insoluble fractions, and soluble fractions were analyzed for the presence of ASC, caspase-1, and vimentin as fractionation control by immunoblotting. c.l., cell lysate; i.f., insoluble fraction. (D) Four replicates each of BMDCs from B6.129-Casp1^mlt/mlt, Nlrp3^−/−, and B6.129-Casp1^−/− were stimulated and prepared as in (C), and insoluble fraction samples were analyzed by mass spectrometry. The relative enrichment of proteins detected in B6.129-Casp1^mlt/mlt over Nlrp3^−/− or B6.129-Casp1^−/− samples is plotted against the inverse relative SD (RSD) within the B6.129-Casp1^mlt/mlt quadruplet (mean/SD). (C and D) Data representative of 2 independent experiments with n = 4 replicates per genotype.

Figure 5

Enhanced Activation of Caspase-8 in Cells Expressing Caspase-1^mlt (A) BMDCs of the indicated genotypes (B6.129-Casp1^mlt/mlt) on culture slides were stimulated with nigericin, fixed, and stained by immunofluorescence for caspase-1 (cyan), caspase-8 (yellow), and ASC (magenta), and then confocal imaging was performed. DAPI (blue) localizes with the nuclei. Scale bar represents 20 μm (data representative of >5 independent experiments). (B) BMDCs from different Casp1 genotypes as indicated (B6-Casp1^mlt/mlt) and ASC-deficient cells (Pycard^−/−) were investigated for recruitment of caspase-8 into IGEPAL CA-630-insoluble fractions following LPS priming and inflammasome activation by nigericin (for 0.5 or 8 hr) (data representative of 2 independent experiments). c.l., cell lysate; s.f., soluble fraction; i.f., insoluble fraction; h.e., high exposure. (C) Cleavage and secretion of IL-1β and caspase-1 and generation of caspase-8 p18 were monitored by immunoblot analysis 1, 3, and 9 hr after stimulation of BMDCs from the indicated genotypes (B6-Casp1^mlt/mlt) with nigericin, imiquimod (70 μM), poly(dA:dT) (1 μg/mL), and Salmonella enterica serovar Typhimurium (MOI 20). (D) Secretion of mature IL-1β in conditions as in (C) was measured by ELISA (mean ± SEM are shown). In (C) and (D), data are representative of 5 independent experiments.

Figure 6

GSDMD-Dependent Pyroptosis Suppresses Caspase-8 Activation (A) BMDCs of wild-type, Gsdmd^−/−, B6-Casp1^−/−, and B6-Casp1^mlt/mlt mice were analyzed by immunoblot for caspase-8 processing, as well as maturation and secretion of IL-1β after inflammasome activation by nigericin or poly(dA:dT) for 3, 9, or 18 hr. (B) Measurement of released IL-1β by ELISA and LDH by an enzymatic assay from samples in (A) (mean ± SEM are shown). (C) Following infection with F. novicida for 48 hr, bacterial loads in spleens (left) and livers (right) of wild-type, Gsdmd^−/−, and B6.129-Casp1 mice were determined by plating serial dilutions of organ homogenates on selective medium plates. The Mann-Whitney test was used for statistical analysis (^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, ^∗∗∗∗p < 0.0001). Spleen: Gsdmd^−/− versus wild-type, p = 0.0017; Gsdmd^−/− versus B6.129-Casp1^−/−, p = 0.0014; wild-type versus B6.129-Casp1^−/−, p < 0.0001; liver: Gsdmd^−/− versus wild-type, p = 0.0004; Gsdmd^−/− versus B6.129-Casp1^−/−, p = 0.0189; wild-type versus B6.129-Casp1^−/−, p < 0.0001; wild-type, B6.129-Casp1^−/−, and Gsdmd^−/−: n = 16. (D) BMDCs from B6.129-Casp1^mlt/mlt mice on chambered coverslips were stimulated with nigericin or 10 μM raptinal and monitored for morphological changes over time by differential interference contrast (DIC) and fluorescence microscopy. Loss of membrane integrity was indicated by DRAQ7 (red) staining of DNA. (E) BMDCs of wild-type, B6.129-Casp1^−/−, and Pycard^−/− mice were pretreated with the caspase inhibitors Z-VAD-fmk (20 μM), IETD-fmk (30 μM), or DEVD-fmk (30 μM) and subsequently stimulated with nigericin. Lytic cell death was quantified by LDH release as measured by an enzymatic assay after 4, 8, 12, and 16 hr. In (A) and (B), data are representative of 2 independent experiments.