Gasdermin pores are not limited to the plasma membrane: the N-terminal fragments of GSDME and GSDMD can also permeabilize mitochondria

GSDME executes pyroptosis after cleavage by caspase-3, an established downstream function. This paper asks: Does GSDME also influence the apoptotic pathway itself in return? The answer is yes: besides forming pores in the plasma membrane, GSDME-N permeabilizes mitochondria to release cytochrome c and amplifies caspase-3 activation through the APAF1 apoptosome. GSDMD-N has the same capability.

Step 1: Discover that GSDME influences caspase-3 activation (Figs. 1–3). The authors induced apoptosis with a glucocorticoid in CEM-C7 leukemia cells. After CRISPR deletion of GSDME, not only was pyroptosis reduced, but the speed and magnitude of caspase-3 activation also decreased markedly. This phenotype was reproducible in immortalized macrophages, melanoma cells, thymoma cells and with other apoptotic stimuli. The authors also found that a phosphomimetic mutation of threonine at position 6 of GSDME, T6E, disrupted oligomerization and thereby abolished pore-forming activity; caspase-3 activation was likewise weaker in cells reconstituted with the T6E mutant. These results show that GSDME's enhancement of caspase-3 depends on its pore-forming activity itself, rather than other functions of the cleaved fragment.

Figure 1Complete original figure

View complete figure ↗

Compare figure ↗

Step 2: Demonstrate that GSDME-N directly permeabilizes mitochondria (Figs. 4–5). Imaging showed that GSDME-N colocalized with mitochondria, and the mitochondrial protein HtrA2 was released before plasma membrane rupture. In vitro, the authors incubated purified mitochondria with caspase-3-cleaved GSDME and directly observed cytochrome c release, whereas the T6E mutant had no effect. These results show that GSDME-N can directly permeabilize the outer mitochondrial membrane, releasing cytochrome c and other pro-apoptotic factors.

Figure 5Complete original figure

View complete figure ↗

Compare figure ↗

Step 3: Confirm an amplification pathway operating in parallel with tBid (Figs. 6–7). In death receptor signaling, caspase-8 activates the mitochondrial pathway by cleaving Bid. The authors tested whether GSDME was redundant with this mechanism. In TNFα-treated CEM-C7 cells, deleting either Bid or GSDME slowed caspase-3 activation, and deleting both weakened it further. In APAF1-knockout cells, ectopically expressed GSDME-N could not activate caspase-3 at all, showing that this amplification loop must act through the apoptosome. These results show that GSDME-N amplifies caspase-3 activation through the mitochondria–apoptosome pathway, resulting in cleavage of more GSDME and creating a positive feedback loop that operates in parallel with the tBid pathway.

Figure 7Complete original figure

View complete figure ↗

Compare figure ↗

Step 4: Extend the findings to GSDMD and in vivo (Figs. 8–9). In macrophages transfected with cytosolic LPS, the authors found that GSDMD deletion markedly delayed and weakened caspase-3 activation, whereas ASC deletion had no effect. Overexpression of either GSDMD-N or GSDMA-N also released cytochrome c and activated caspase-3. In vivo, GSDME-knockout B16 melanomas grew faster after subcutaneous implantation in mice. These results show that mitochondrial permeabilization is a shared property of the gasdermin family and that GSDME's pro-death function has tumor-suppressive significance.

Figure 8Complete original figure

View complete figure ↗

Compare figure ↗

1. Changes in apoptotic markers after gasdermin deletion should not be surprising: GSDME/GSDMD deletion can slow caspase-3 activation itself, a factor to consider when interpreting caspase-3 cleavage bands.
2. Distinguish the timing of mitochondrial release from plasma membrane rupture: Mitochondrial proteins are released before plasma membrane rupture; readouts that conflate the two into a single time point would miss the direct action of GSDME-N.
3. Pay attention to phosphorylation-dependent regulation: The T6E phosphomimetic mutation allows GSDME-N to reach the membrane but prevents oligomerization. Detecting "membrane localization" is not the same as detecting "activity".

• Extension | GSDME mediates secondary necrosis following apoptosis (Rogers 2017, Nat Commun): The 2017 paper showed that GSDME determines whether apoptotic cells undergo membrane rupture; this paper shows that GSDME also amplifies upstream caspase-3 activation. Together, the two papers provide a complete picture of GSDME function.
• Comparison | Caspase-1 initiates apoptosis in the absence of GSDMD (Tsuchiya 2019, Nat Commun): Tsuchiya's paper reports that, in the absence of GSDMD, caspase-1 connects to mitochondrial apoptosis through Bid. This paper reports direct mitochondrial permeabilization by GSDMD-N. Both routes lead to the mitochondria.
• Extension | GSDMD permeabilizes mitochondria to amplify pyroptosis (Miao 2023, Immunity): This paper shows that gasdermin-N permeabilizes mitochondria to amplify caspase-3 during apoptosis. Miao shows that GSDMD-NT permeabilizes both mitochondrial membranes and amplifies pyroptosis itself in a cardiolipin-dependent manner—the mechanism of "gasdermin attacking mitochondria" serves two cell death pathways.
• Extension | Effector caspases amplify apoptotic signaling through feedback (McComb 2019, Sci Adv): Positive feedback in apoptotic signaling operates at two levels. McComb shows that effector caspases amplify upstream caspase-8/-9 and mitochondrial permeabilization through protein cleavage; this paper shows that GSDME-N amplifies caspase-3 activation itself by permeabilizing mitochondria.

(The cross-paper comparisons above are a synthesis, rather than direct conclusions of any one original paper.)