Pyroptosis

Why pyroptosis matters

• It clears the intracellular pathogen niche. Bacteria like Salmonella Typhimurium, Shigella flexneri, Legionella pneumophila, Listeria, and Francisella hide inside cells where antibodies and complement cannot reach. A macrophage that senses the intruder in its cytosol and then ruptures destroys the replicative niche and ejects the bacteria into the open, where neutrophils and complement can finish them.
• Pore-induced intracellular traps (PITs). When a pyroptotic macrophage bursts, it does not scatter live bacteria freely — the bacteria stay trapped in the membrane remnant, a structure named a PIT. Incoming neutrophils recognize and clear these traps by efferocytosis, so pyroptosis and neutrophil killing operate as a relay, not a scatter-shot.
• It matures and releases IL-1β and IL-18. These cytokines have no signal peptide and cannot leave through the conventional secretory route. They are synthesized as inactive precursors, cleaved by caspase-1, and released through the gasdermin D pore — a form of unconventional secretion that couples cytokine export directly to the death program.
• It is a druggable inflammatory hub. Gain-of-function NLRP3 mutations cause the cryopyrin-associated periodic syndromes (CAPS); the IL-1 blockers anakinra, canakinumab, and rilonacept treat them. The CANTOS trial (2017) showed the anti-IL-1β antibody canakinumab cut recurrent cardiovascular events, direct clinical proof that inflammasome output drives atherosclerosis.
• It sculpts sterile inflammatory disease. Cholesterol crystals in plaques, monosodium urate crystals in gout, silica and asbestos in the lung, and islet amyloid in type 2 diabetes all activate NLRP3, linking pyroptosis to some of the most common chronic diseases rather than only to infection.
• It drives CD4 T-cell loss in HIV. Abortive HIV infection of resting lymphoid CD4 T cells triggers caspase-1-dependent pyroptosis; this — not direct viral killing of productively infected cells — accounts for the bulk of the CD4 depletion that defines AIDS, which is why caspase-1 inhibition was proposed as an anti-inflammatory adjunct.
• Gasdermins are an ancient pore-forming family. Gasdermin D is one of at least six human gasdermins (GSDMA–E, plus PJVK/DFNB59). Gasdermin E is cleaved by apoptotic caspase-3 to convert stalled apoptosis into secondary pyroptosis, and gasdermin-like pore formers exist in bacteria and fungi, marking this as a deeply conserved defense mechanism.

Common misconceptions

• "Pyroptosis is just necrosis." Early observers grouped the swelling, lytic morphology with accidental necrosis. It is not: pyroptosis is genetically programmed, requires specific inflammatory caspases and gasdermin D, and produces a defined pore of controlled size. Genetic deletion of caspase-1 or gasdermin D abolishes it, which no accidental death would allow.
• "Caspase-1 kills by cleaving the same substrates as caspase-3." Inflammatory and apoptotic caspases have almost no overlap in substrates. Caspase-1's decisive death substrate is gasdermin D — not the ICAD/CAD, lamin, or PARP substrates of executioner caspases-3/7. A cell with active caspase-1 but no gasdermin D releases little and survives longer, dying by a slower apoptosis-like route.
• "The gasdermin D pore is what kills the cell directly." The pore itself does not shred the cell; it collapses the ion gradients. Potassium leaves, sodium and calcium enter, water follows osmotically, and the cell swells until the membrane fails. The ninjurin-1 (NINJ1) protein was shown in 2021 to drive the final plasma-membrane rupture, so lysis is itself a regulated step downstream of the pore.
• "IL-1β is secreted like a normal protein." Pro-IL-1β lacks a signal peptide and never enters the ER–Golgi secretory pathway. It is matured in the cytosol by caspase-1 and exits through the gasdermin D pore. Block the pore and the cytokine stays trapped inside — pore formation and cytokine release are mechanically the same event.
• "Non-canonical pyroptosis makes its own IL-1β." Caspase-4/5/11 bind cytosolic LPS and cleave gasdermin D for lysis, but they cannot process pro-IL-1β directly. The potassium efflux through the resulting pore activates NLRP3 secondarily, recruiting caspase-1 to mature the cytokine. The lytic and cytokine arms are split between two caspases.
• "Only immune cells undergo pyroptosis." Macrophages, monocytes, and dendritic cells are the classic responders, but intestinal and airway epithelium, keratinocytes, and endothelium also express gasdermin D and inflammasome sensors and can pyroptose — epithelial pyroptosis expels infected enterocytes into the gut lumen as a barrier-defense mechanism.

How pyroptosis works, step by step

Pyroptosis begins with a cytosolic sensor detecting a pathogen- or danger-associated signal that has no business being in the cytoplasm. Different sensors recognize different threats: NLRP3 is a general danger sensor triggered downstream of potassium efflux, lysosomal rupture, mitochondrial damage, extracellular ATP (via the P2X7 channel), and particulate crystals; NLRC4, aided by NAIP proteins, recognizes bacterial flagellin and the rod and needle proteins of type III secretion systems; AIM2 binds cytosolic double-stranded DNA; and Pyrin senses toxins that inactivate host RhoA. On activation, the sensor nucleates the adaptor ASC, whose PYD and CARD domains self-polymerize into filaments that condense into a single micron-scale "ASC speck" — a one-per-cell hallmark of inflammasome assembly.

The ASC speck clusters many molecules of pro-caspase-1 through CARD–CARD interactions. Forced into proximity, pro-caspase-1 undergoes proximity-induced dimerization and autoprocessing to become active caspase-1. This is the canonical route. In the non-canonical route, the inflammatory caspases-4 and -5 in humans (caspase-11 in mice) skip the sensor entirely: their CARD domains are themselves direct receptors for cytosolic lipopolysaccharide (LPS) from Gram-negative bacteria, so contact with LPS oligomerizes and activates them.

Active caspase-1 (or -4/-5/-11) then cleaves gasdermin D at Asp275 in the linker between its lipid-binding N-terminal domain and its autoinhibitory C-terminal domain. Freed from autoinhibition, the N-terminal fragments diffuse to the membrane, bind acidic inner-leaflet lipids (phosphatidylserine, phosphoinositides) and, on bacterial membranes, cardiolipin. There, roughly 27 to 33 protomers oligomerize into a ring and insert a beta-barrel transmembrane pore of about 10–20 nm inner diameter — one of the largest channels in cell biology, wide enough to pass a folded 17 kDa cytokine.

Two things now happen at once. First, the mature cytokines IL-1β and IL-18 — cleaved from their precursors by the same caspase-1 — escape through the pore, alarming and recruiting neutrophils, driving fever, and priming neighbors. Second, the pore dissipates the ionic gradients: potassium floods out, sodium and calcium rush in, water follows osmotically, and the cell balloons. The final plasma-membrane rupture is executed by the membrane protein NINJ1, spilling danger-associated molecular patterns (DAMPs) such as HMGB1, ATP, and IL-1α. The whole program — from speck to lysis — plays out in tens of minutes, converting one compromised cell into a loud, coordinated immune alarm.

Pyroptosis vs apoptosis vs necroptosis vs necrosis

Feature	Pyroptosis	Apoptosis	Necroptosis	Necrosis (accidental)
Trigger	Inflammasomes (NLRP3, NLRC4, AIM2), cytosolic LPS	BH3-only proteins / death receptors	Death receptors + caspase-8 block	Energy failure, severe injury
Key effectors	Caspase-1/-4/-5/-11, gasdermin D	Caspases-3/7/8/9	RIPK1, RIPK3, MLKL	None (passive)
Membrane fate	Gasdermin D pores → swelling → lysis	Intact, blebs outward	MLKL pores → lysis	Ruptures, swells
Inflammation	Very high (IL-1β, IL-18, DAMPs)	Silent (anti-inflammatory)	High (DAMPs)	High (DAMPs)
DNA	Nicked, no laddering	180-bp laddering	Random degradation	Random degradation
Programmed?	Yes	Yes	Yes	No
Main role	Anti-intracellular-pathogen defense	Homeostasis, tissue sculpting	Backup death when apoptosis is blocked	Injury response
Drug target	NLRP3 (MCC950), GSDMD (disulfiram), IL-1 (canakinumab)	BCL-2 (venetoclax)	RIPK1 (necrostatins)	—

Canonical vs non-canonical pyroptosis

Property	Canonical	Non-canonical
Upstream trigger	PAMP/DAMP sensed by NLRP3, NLRC4, AIM2, or Pyrin	Cytosolic LPS binding the caspase CARD directly
Adaptor / platform	ASC speck (for most sensors)	None — caspase is its own sensor
Active caspase	Caspase-1	Caspase-4 / -5 (human); caspase-11 (mouse)
Cleaves gasdermin D	Yes	Yes
Cleaves pro-IL-1β / IL-18 directly	Yes	No (cannot process the cytokines)
Cytokine maturation	Same caspase-1 event	Indirect: K⁺ efflux → NLRP3 → caspase-1
Pathogen class targeted	Broad (flagellated, DNA, toxin, crystals)	Gram-negative bacteria (LPS-bearing)

Famous experiments and history

• Zychlinsky and Shigella (1992–1996). Arturo Zychlinsky's group found that Shigella-infected macrophages died in a way that required the protease then called ICE (interleukin-1-converting enzyme, now caspase-1) and released IL-1β — the first hint that a caspase could drive inflammatory, not apoptotic, death.
• Cookson and Brennan coin "pyroptosis" (2001). Studying Salmonella-infected macrophages, Brad Cookson and Molly Brennan named the caspase-1-dependent, pro-inflammatory death "pyroptosis" — Greek for "falling by fire" — to distinguish it firmly from apoptosis in Trends in Microbiology and Cellular Microbiology.
• Gasdermin D unmasked (2015). Two teams — one led by Feng Shao, one by Vishva Dixit and Judy Lieberman — used forward-genetic and CRISPR screens to show that gasdermin D is the essential caspase-1/-11 substrate; cleaving it releases an N-terminal fragment that is necessary and sufficient for pyroptotic lysis. This solved a mechanism that had been open since 2001.
• The pore is visualized (2016). Cryo-EM and atomic-force microscopy from the Shao, Wu, and Ruan groups resolved the gasdermin D N-terminal fragment as a large membrane-spanning ring of ~27–33 subunits with a ~10–20 nm lumen, explaining how a folded cytokine can pass through and how ions collapse.
• Non-canonical LPS sensing (2013–2014). Work from the Dixit and Shao labs showed that caspase-11 (and human caspase-4/5) bind intracellular LPS directly through their CARD domains — the caspases are themselves the pattern-recognition receptors, a striking exception to the sensor-adaptor-caspase paradigm.
• NINJ1 executes rupture (2021). A genome-wide screen from the Kayagaki/Dixit group identified ninjurin-1 (NINJ1) as the protein that drives the final plasma-membrane rupture in pyroptosis and other lytic deaths, showing that even cell bursting is genetically programmed rather than passive.