Gasdermin pores permeabilize mitochondria to augment caspase-3 activation during apoptosis and inflammasome activation

Abstract

Gasdermin E (GSDME/DFNA5) cleavage by caspase-3 liberates the GSDME-N domain, which mediates pyroptosis by forming pores in the plasma membrane. Here we show that GSDME-N also permeabilizes the mitochondrial membrane, releasing cytochrome c and activating the apoptosome. Cytochrome c release and caspase-3 activation in response to intrinsic and extrinsic apoptotic stimuli are significantly reduced in GSDME-deficient cells comparing with wild type cells. GSDME deficiency also accelerates cell growth in culture and in a mouse model of melanoma. Phosphomimetic mutation of the highly conserved phosphorylatable Thr6 residue of GSDME, inhibits its pore-forming activity, thus uncovering a potential mechanism by which GSDME might be regulated. Like GSDME-N, inflammasome-generated gasdermin D-N (GSDMD-N), can also permeabilize the mitochondria linking inflammasome activation to downstream activation of the apoptosome. Collectively, our results point to a role of gasdermin proteins in targeting the mitochondria to promote cytochrome c release to augment the mitochondrial apoptotic pathway.

Fig. 1

GSDME regulates caspase-3 activation and pyroptosis in CEM-C7 cells. a Immunoblots of GSDME, caspase-3, and β-actin in cell lysates (Lys), or high-mobility group box 1 (HMGB1) released in culture media (media) of CEM-C7 cells untreated (UT) or triamcinolone acetonide (TA) treated for the indicated times. b Cytotoxicity of TA as measured by lactate dehydrogenase (LDH) release in the culture supernatants of untreated (Untreated) or TA-treated (TA) CEM-C7 cells for the indicated times. c, d Propidium iodide (PI) uptake (c), and active caspase-3 staining (d–g) in wild-type (WT) and GSDME-knockout (KO) CEM-C7 cells treated with TA (d), etoposide (e), ultraviolet (UV) (f), or serum starvation (g) as measured on the IncuCyte over time. h Immunoblots of GSDME, caspase-3, and β-actin in combined cell lysates plus culture media  of WT and GSDME-KO (KO) CEM-C7 cells untreated (UT) or treated with TA, etoposide, serum starvation, or UV as indicated. Results are representative of at least three independent experiments performed in duplicate or triplicate. Error bars represent standard deviation (S.D.). Student’s t-test, *p < 0.05, **p < 0.005, ***p < 0.0005, ****p < 0.00005, ******p < 0.0000005

Fig. 2

T6E mutation inhibits the pyroptotic and self-oligomerization activities of GSDME. a, b Cytotoxicity and pyroptotic activity of the indicated GSDME-N constructs as measured by lactate dehydrogenase (LDH) release (a) or propidium iodide (PI) uptake (b), respectively, in 293T cells transfected with empty vector (Ctrl) or the indicated GSDME. Expression of these constructs in cell lysates is shown in (a, top panel) as visualized by immunoblot analysis with anti-GSDME antibody. *Internal translation. **Non-specific band. c Representative IncuCyte images showing pyroptosis induction in cells expressing GSDME-N-EGFP (left) or GSDME-N-T6A-EGFP (right), but not in cells expressing GSDME-N-T6E-EGFP (middle). Scale bar, 50 µm. d Confocal images of GSDME-N-EGFP variants T6E (top panels), T6A (middle panels) and wild-type (WT; lower panels) transiently expressed in 293T cells. The perinuclear localization of WT or T6A GSDME-EGFP is largely due to the association of GSDME-N-EGFP with the mitochondria (see Fig. 4), which coalesce around the nucleus in pyroptotic cells. Scale bar, 10 µm. e Immunoblots showing WT or T6E GSDME-N after incubation with purified cell membranes and fractionation on sodium dodecyl sulfate (SDS) polyacrylamide gel in the presence or absence of β-mercaptoethanol (2-ME), or incubation with purified cell membranes followed by cross-linking with disuccinimidyl suberate (DSS) and fractionation on SDS–polyacrylamide gel in the presence of 2-ME. Results are representative of at least three independent experiments performed in duplicate or triplicate. Error bars represent S.D. Student’s t-test, *p < 0.05, ****p < 0.00005

Fig. 3

Pyroptosis and caspase-3 activation are compromised in GSDME-T6E-expressing CEM-C7 cells. a–d Propidium iodide (PI) uptake (a, c), and active caspase-3 staining (b, d) in CEM-C7 cells (wild-type (WT), GSDME-knockout (KO), or GSDME-KO reconstituted with WT GSDME-EGFP or GSDME-T6E-EGFP) treated with triamcinolone acetonide (TA) (a, b) or ultraviolet (UV) (c, d) as measured on the IncuCyte over time. e, f Immunoblots of GSDME, active caspase-3 p17/p19 and high-mobility group box 1 (HMGB1) in combined cell lysates plus media, or HMGB1 released in culture media (media) of WT (WT), GSDME-KO (KO), or GSDME-KO reconstituted with WT GSDME-EGFP (KO + WT) or GSDME-T6E-EGFP (KO+T6E) CEM-C7 cells treated with TA (e) or UV (f). *Indicates GSDME-EGFP degradation product. Results are representative of at least three independent experiments performed in duplicate or triplicate. Error bars represent S.D. Student’s t-test, *p < 0.05, **p < 0.005, ***p < 0.0005

Fig. 4

GSDME-N localizes to mitochondria and releases proapoptotic proteins. a Confocal live-cell imaging of GSDME-N-EGFP (1st and 2nd rows), GSDME-EGFP (3rd row), and EGFP (4th row) expressed in HeLa cells stably reconstituted with mCherry-tagged mitochondrial HtrA2. The green arrows indicate non-pyroptotic cells, and the red arrows indicate pyroptotic cells. b Confocal live-cell imaging of MitoTraker red-stained 293T cells after transfection with GSDME-N-EGFP (upper panels), or full-length GSDME-EGFP (lower panels). The green channels show the expression of GSDME-N-EGFP (upper left), or GSDME-EGFP (lower left). The red channels show mitochondrial staining with MitoTracker red. The merged channels are shown on the right. The arrows (upper right) indicate the co-localization of GSDME-N-EGFP with MitoTracker red. Results are representative of at least three independent experiments. Scale bar, 10 µm

Fig. 5

GSDME-N induces cytochrome c (Cyt c) release from the mitochondria. a Immunoblots of Cyt c (1st panel) released from purified mitochondria incubated with recombinant caspase-3, and S100 lysates from control (Ctrl) HEK293T cells or from HEK293T cells stably expressing wild-type (WT) GSDME-EGFP (WT) or GSDME-T6E-EGFP (T6E). (Lys+casp-3): S100 lysates after incubation with caspase-3 and mitochondria followed by removal of mitochondria by centrifugation. (Mito+casp-3): mitochondrial pellet obtained at the end of the reaction. (Lys): S100 lysates before incubation with caspase-3 or mitochondria. b, Upper Immunoblots of Cyt c (1st panel) released into the reaction supernatant (Sup) from purified mitochondria (Mito) incubated with or without caspase-3 and WT GSDME-N (WT) or inactive GSDME-N-T6E (T6E) mutant (2nd panel). b, Lower Coomassie-stained sodium dodecyl sulfate (SDS)-gel of the purified full-length WT GSDME-N-EGFP (WT) or inactive GSDME-N-T6E (T6E) used in this experiment. c Immunoblots of Cyt c (1st and 2nd panels) released into the culture media (media) or cytosol (Cyto) from 293T cells transfected with constructs encoding GSDME, GSDME-N, GSDME-N-T6E, or Bax. d Immunoblots of Cyt c (1st and 2nd panels) released into the culture media (media) or cytosol (Cyto) from stable 293T cells expressing WT GSDME (293T-GSDME) or the uncleavable GSDME-D270E (293-GSDME-D270E) transfected with constructs encoding constitutively active caspase-3 (Casp-3) or inactive caspase-3 mutant (Casp-3-C285A). e Immunoblots of Cyt c (1st and 2nd panels) released into the culture media (media) or cytosol (Cyto) from WT (WT) or GSDME-KO (KO) CEM-C7 cells treated with ultraviolet (UV) for 18 h. Voltage-dependent anion channel (VDAC) and cytochrome c oxidase IV (Cox IV) blots were used as loading controls for mitochondria. β-Actin blots were used as loading controls for cytosolic extracts or lysates. Quantitative analyses of Cyt c release in (a–e) are shown in Supplementary Fig. 16. Results are representative of at least three independent experiments

Fig. 6

GSDME augments death receptor-induced pyroptosis and caspase-3 activation. a, b Propidium iodide (PI) uptake (a) and active caspase-3/7 staining (b) in CEM-C7 cells (wild-type (WT), GSDME-knockout (KO), or GSDME-KO reconstituted with WT GSDME-EGFP or GSDME-T6E-EGFP) treated with tumor necrosis factor α (TNFα) plus actinomycin D (ActD) as measured on the IncuCyte over time. c Immunoblots of GSDME, active caspase-3 p17/p19, and high-mobility group box 1 (HMGB1) in combined cell lysates plus media, or HMGB1 released in culture media (media) of WT (WT), GSDME-KO (KO), or GSDME-KO reconstituted with WT GSDME-EGFP (KO+WT) or GSDME-T6E-EGFP (KO+T6E) CEM-C7 cells treated with TNFα plus ActD as indicated. d Immunoblots of cytochrome c (Cyt c) (1st and 2nd panels) released into the culture media (media) or cytosol (Cyto) from WT (WT) or GSDME-KO (KO) CEM-C7 cells treated with TNFα plus ActD for 18 h. e Active caspase-3/7 staining in WT, GSDME-KO, Bid-KO, or GSDME/Bid-dKO CEM-C7 cells treated with TNFα plus ActD as measured on the IncuCyte over time. f Immunoblots of Cyt c (1st and 2nd panels) released into the culture media (media) or cytosol (Cyto) from WT, GSDME-KO (GSE-KO), Bid-KO, or GSDME/Bid-dKO (dKO) CEM-C7 cells treated with TNFα plus ActD. Quantitative analyses of Cyt c release in (d, f) are shown in Supplementary Fig. 17. Results are representative of at least three independent experiments performed in duplicate or triplicate. Error bars represent S.D. Student’s t-test, *p < 0.05, **p < 0.005, ***p < 0.0005, ****p < 0.00005

Fig. 7

GSDME functions as a feed-forward amplifier of caspase-3 activation and pyroptosis. a Active caspase-3 in 293T cells transfected with empty vector (Ctrl) or the indicated GSDME constructs as measured on the IncuCyte over time. b Immunoblots of lysates from 293T cells transfected with the indicated constructs as in (a). c Immunoblots of lysates from stable 293T-GSDME-EGFP and 293T-GSDME-D270E-EGFP cells transfected with the indicated constructs encoding C-terminally T7-tagged GSDME proteins. The upper panel (probed with anti-EGFP) shows generation of the cleaved GSDME-C-EGFP domain in 293T-GSDME-EGFP, but not 293T-GSDME-EGFP-D270E cells in response to activation of caspase-3 by the transfected GSDME-N fragment. d, e Active caspase-3 staining (d) and propidium iodide (PI) uptake (e) in 293T cells transfected with the indicated GSDME constructs as measured on the IncuCyte over time. f Immunoblots of cytochrome c (Cyt c) (1st and 2nd panels) released into the culture media (media) or cytosol (Cyto) from 293T cells transfected with the indicated GSDME constructs. GSDME, full-length GSDME, GSDME-N, cleaved GSDME-N domain, Mutant GSDME, deafness-associated GSDME mutant. Quantitative analyses of Cyt c release in (f) are shown in Supplementary Fig. 18. Results are representative of at least three independent experiments performed in duplicate or triplicate. Error bars represent S.D. Student’s t-test, **p < 0.005, ***p < 0.0005, *****p < 0.000005

Fig. 8

GSDMD functions downstream of the inflammasome to activate caspase-3. a Active caspase-3 staining in PamCSK4-primed wild-type (WT), GSDMD-knockout (KO), or ASC-KO immortalized bone marrow-derived macrophages (iBMDMs) transfected with lipopolysaccharide (LPS; caspase-11 activator) as measured on the IncuCyte over time. b Immunoblots of GSDMD, active caspase-3 p17/p19, and β-actin in cell lysates of PamCSK4-primed WT and GSDMD-KO (KO) iBMDMs untransfected (−) or transfected (+) with LPS as indicated. c Active caspase-3 in 293T cells transfected with empty vector (Ctrl) or the indicated GSDMD constructs as measured on the IncuCyte over time. d Immunoblots of cytochrome c (Cyt c; 1st and 2nd panels) released into the culture media (media) or cytosol (Cyto) from 293T cells transfected with the indicated GSDMA, GSDMD, or GSDME constructs. Quantitative analyses of Cyt c release in (d) are shown in Supplementary Fig. 19. Results are representative of at least three independent experiments performed in duplicate or triplicate. Error bars represent S.D. Student’s t-test, *p < 0.05, ***p < 0.0005

Fig. 9

GSDME suppresses tumor cell growth in vitro and in vivo. Growth of wild-type (WT) and GSDME-KO CEM-C7 (a), immortalized bone marrow-derived macrophages (iBMDMs) (b), and B16-Ova (c) cells as measured on the IncuCyte over time. Growth curves are representative of at least three independent experiments performed in duplicate or triplicate. Error bars represent S.D. d Survival curve for mice inoculated with 3 × 10⁵ WT (n = 11 mice) or GSDME-KO (n = 9 mice) B16-Ova cells. Mice were killed once tumors reached volumes greater than 450 mm³. e A representative image of early tumor volume difference between WT and GSDME-KO tumors in mice. f A model representing how GSDMD and GSDME permeabilize the mitochondria to release cytochrome c (Cyt c) and augment caspase-3 activation and apoptosis. Cell growth curve significance was determined by Student’s t-test, **p < 0.005, ***p < 0.0005. Tumor survival curve significance was determined by log-rank test, *p < 0.05