To die or not to die: Gasdermins in intestinal health and disease

Abstract

Intestinal homeostasis is achieved by the balance among intestinal epithelium, immune cells, and gut microbiota. Gasdermins (GSDMs), a family of membrane pore forming proteins, can trigger rapid inflammatory cell death in the gut, mainly pyroptosis and NETosis. Importantly, there is increasing literature on the non-cell lytic roles of GSDMs in intestinal homeostasis and disease. While GSDMA is low and PJVK is not expressed in the gut, high GSDMB and GSDMC expression is found restrictively in intestinal epithelial cells. Conversely, GSDMD and GSDME show more ubiquitous expression among various cell types in the gut. The N-terminal region of GSDMs can be liberated for pore formation by an array of proteases in response to pathogen- and danger-associated signals, but it is not fully understood what cell type-specific mechanisms activate intestinal GSDMs. The host relies on GSDMs for pathogen defense, tissue tolerance, and cancerous cell death; however, pro-inflammatory milieu caused by pyroptosis and excessive cytokine release may favor the development and progression of inflammatory bowel disease and cancer. Therefore, a thorough understanding of spatiotemporal mechanisms that control gasdermin expression, activation, and function is essential for the development of future therapeutics for intestinal disorders.

Introduction

The first member of the gasdermin family was identified in the upper gastrointestinal tract in 2000 [1]. Up until now, the GSDM family has six members in human, including GSDMA, GSDMB, GSDMC, GSDMD, GSDME (DFNA5) and PJVK (DFNB59, GSDMF) [2]. In mice, GSDMA has three homologues: GSDMA1-3. GSDMB has no homologue in mice. There are four homologues of GSDMC in mice: GSDMC1-4. GSDMD, GSDME, and PJVK each possesses a single homologue in mice. All GSDMs are capable executors of pyroptosis, except for PJVK. While they have conserved N-terminus (NT) and C-terminus (CT), the linker regions differ, which leads to distinct regulations and functions of GSDMs.

GSDMs play pivotal roles in maintaining intestinal homeostasis. They have been shown to induce inflammatory lytic cell death (e.g., pyroptosis and NETosis), rapid cytokine release without cell death, and other non-cell-lytic functions. The aim of this review is to: (1) provide a comprehensive summary of the expression pattern of GSDMs in the intestine based on current literature and single cell RNA sequencing (scRNA-seq) databases; (2) summarize the regulation, cleavage, and functions of GSDMs; and (3) discuss the pathophysiological roles of GSDMs in pathogen-host interactions, inflammatory bowel disease (IBD), and gastrointestinal cancer.

1. Gasdermin Expression in the Intestine

The gut epithelium is a highly regenerative tissue and is replenished every 3-5 days. Considering the translation rate and protein lifetimes, one would predict discordances between mRNA and protein expression in the intestine. Indeed, many genes exhibit mRNA zonated toward the crypt, while proteins are zonated toward the villus, due to delayed protein synthesis [3]. Therefore, caution should be taken when drawing conclusions on intestinal epithelial protein expression and function solely based on mRNA results. For instance, while mouse Gsdmc2-4 mRNA expression is enriched in the transit amplifying (TA) region shown by scRNA-seq and RNAscope, its protein is much more abundant in the villus than the crypt [4]. In addition, we would like to point out that many expression results discussed below were obtained from intestine at homeostatic conditions. Infectious, inflammatory, and tumorigenic environments may vary the expression levels and locations of GSDMs [5].

GSDMs, except PJVK, have been reported or detected at some level in the gastrointestinal (GI) tract. Their expression is mainly found in the epithelium and immune cells, but can also be detected in other cell types including endothelial cells and fibroblasts. In order to decipher the role of individual GSDMs in the intestine, we have to first know which cells express them, at what conditions.

Human GSDMA/ Mouse GSDMA1-3

Although the human GSDMA gene was suggested to be expressed in the upper GI tract and suppressed in gastric cancer [1], further study is needed to determine what subpopulations of human intestinal epithelial cells (IECs), if any, express GSDMA.

Mouse Gsdma genes are primarily expressed in epidermal keratinocytes and hair follicle stem cells [6]. In the mouse intestine, based on scRNA-seq results, Gsdma1 is expressed in a cluster of Chgb⁺ enteroendocrine cells (EECs), Gsdma2 is expressed in a small subset of goblet cells, while Gsdma3 is barely detectable (Table1, 2) [7, 8].

Table 1.

Expression of human gasdermin genes

	GSDMA	GSDMB	GSDMC	GSDMD	GSDME	PJVK
Epithelial cells
Stem cells	−	++	−	++++	−	−
Progenitor cells	−	+++	−	+++	−	−
Transit amplifying cells	−	+++	−	+++	−	−
Goblet cells	−	++	++	+++	+	−
Tuft cells	−	−	−	+	−	−
Paneth cells	−	+	−	+++	−	−
Microfold cells (M cells)	−	−	−	+	−	−
Enterocytes	−	+++	−	++++	++	−
Enteroendocrine cells (EECs)	−	+	−	++	−	−
Immune cells
B cells	−	−	−	+	+	−
T cells	−	+	−	++	−	−
Mast cells	−	−	−	+	−	−
ILCs	−	−	−	+	−	−
Macrophages	−	+	−	++++	+	−
Dendritic cells (DCs)	−	−	−	+++	−	−
Monocytes	+	−	−	++++	++	−
NK cells	−	−	−	++	−	−
Other cells
Endothelial cells	−	−	−	+++	−	−
Smooth muscle cells	−	−	−	++	−	−
Glial cells	−	−	−	++	−	−
Fibroblasts	−	−	−	+++	−	−
Pericytes	−	−	−	+	−	−

−: not detected; + : low; ++ : medium; +++ : high; ++++ : very high.

− indicates less than 1% cells express GSDM genes; + indicates 1%-10% cells express low levels of GSDM genes; ++ indicates 10%-30% cells express low or medium levels of GSDM genes; +++ indicates 30%-50% cells express medium or high levels of GSDM genes; ++++ indicates more than 50% cells express high levels of GSDM genes.

The scRNA-seq data is from https://www.ebi.ac.uk/gxa/sc/ and https://singlecell.broadinstitute.org [8]

Table 2.

Expression of mouse gasdermin genes

	Gsdm a1	Gsdm a2	Gsdm a3	Gsdm c1	Gsdm c2	Gsdm c3	Gsdm c4	Gsd md	Gsd me	Pjv k
Epithelial cells
Stem cells	−	−	−	−	−	−	−	+	−	−
Progenit ors	−	−	−	−	++	+	++	+	−	−
Goblet cells	−	+	−	−	+	−	+	+	+++	−
Tuft cells	−	−	−	−	+	−	+	+	−	−
Paneth cells	−	−	−	−	−	−	−	+	−	−
Enterocyt es	−	−	−	−	+++	++	++++	+++	++++	−
EECs	+	−	−	−	−	−	−	+++	++++	−
Immune cells
B cells	−	−	−	−	−	−	−	+	+	−
T cells	−	−	−	−	−	−	−	++	+	−
Mast cells	−	−	−	−	−	−	−	++++	+++	−
Macroph ages	−	−	−	−	−	−	−	++++	++	−
Neutrop hils	−	−	−	−	−	−	−	++	+	−
Dendritic cells	−	−	−	−	−	−	−	++	+	−
Monocyt es	−	−	−	−	−	−	−	++++	+	−
NK cells	−	−	−	−	−	−	−	++	−	−
Other cells
Endothelial cells	−	−	−	−	−	−	−	++	−	−
Smooth muscle	−	−	−	−	−	−	−	++	−	−
Glial cells	−	−	−	−	−	−	−	++	+	−
Fibroblas	−	−	−	−	−	−	−	+++	+	−
Pericytes	−	−	−	−	−	−	−	+	+	−

−: not detected; + : low; ++ : medium; +++ : high; ++++ : very high.

− indicates less than 1% cells express Gsdm genes; + indicates 1%-10% cells express low levels of Gsdm genes; ++ indicates 10%-30% cells express low or medium levels of Gsdm genes; +++ indicates 30%-50% cells express medium or high levels of Gsdm genes; ++++ indicates more than 50% cells express high levels of Gsdm genes.

The scRNA-seq data is from https://bis.zju.edu.cn/MCA/ and https://singlecell.broadinstitute.org [7]

Human GSDMB

The predominant expression site of human GSDMB is the epithelium of lung, esophagus, intestine, and stomach [9-12]. scRNA-seq shows that intestinal GSDMB is expressed in stem cells, TA cells and almost all types of differentiated IECs, expect tuft cells and microfold cells (Table 1). In the immune compartment, GSDMB is expressed by macrophage and T cells [13, 14].

Human GSDMC/ Mouse GSDMC1-4

Human GSDMC expression has been reported to be restricted in the epithelium, not any immune cells, of trachea, esophagus, and intestine [4, 12, 15]. According to human scRNA-seq data, GSDMC high-expressing intestinal cells are goblet cells (Table 1), which is consistent with recent reports in mice [4, 16].

In mice, Gsdmc1 is not expressed in the intestine due to epigenetic silencing (Table 2) [4, 16, 17]. Gsdmc4, Gsdmc2, and Gsdm3 (in the order of their expression levels) can be detected in several types of IECs, including goblet cells, tuft cells, progenitor cells, and enterocytes. It is worth noting that although Gsdmc 2-4 mRNA is enriched in the crypt base and TA region based on sequencing and RNA in situ hybridization [4,16,17], immunostaining revealed that GSDMC protein is only abundant expression in villous enterocytes [4]. Interestingly, the duplicated Gsdmc2-4 genes in mice display an expression pattern analogous to the human-only GSDMB, suggesting that they might be adopted by the two hosts for similar gut functions (see below).

Human GSDMD/Mouse GSDMD

Human GSDMD (hGSDMD) exhibits global expressions in epithelial cells, neurons, cardiomyocytes, and immune cells [2,18]. Almost all cell types in the intestine express some level of hGSDMD, including epithelial cells, fibroblasts, smooth muscle cells, as well as innate and adaptive immune cells (Table 1, 2).

The expression pattern of mouse GSDMD (mGSDMD) is similar to hGSDMD, but some notable differences exist between mice and humans. Mouse GSDMD expression in mast cells is significantly higher than hGSDMD, while in goblet cells, tuft cells, epithelial stem cells and progenitor cells is much lower than that of human. When investigating the function of mGSDMD using mouse models, these expression differences should be considered.

Human GSDME/Mouse GSDME

Human GSDME (hGSDME), also known as DFNA5, was first discovered as a deafness associated gene in the epithelium of cochlea [19]. In the epithelium of intestine, stomach, and esophagus, hGSDME is also highly expressed [20, 21]. In the IECs, human GSDME gene expression is only found in enterocytes and goblet cells (Table 1). Outside of the epithelium, hGSDME is expressed in the endothelium of blood vessels and various immune cells, including B cells, macrophages, and monocytes [2, 22, 23].

Mouse Gsdme is expressed in enterocytes and goblet cells, at a level seemingly higher that of human GSDME, as well as in EECs. Most murine immune cells, including B cells, T cells, mast cells, neutrophils, dendritic cells (DCs), macrophages, and monocytes, express Gsdme. Intriguingly, both Gsdmd and Gsdme are abundantly expressed by mast cells in mice but not in humans, which may contribute to species-specific roles of mast cells in various biological responses. Furthermore, mouse Gsdme is also expressed by glial cells, fibroblasts, and pericytes in mice, implicating their broad and impactful roles.

The cell-type specific expression will provide a new perspective to dissect the roles of GSDMs in the intestine. While gene expression data become more abundantly available, we still lack knowledge of the protein expression and localization of most GSDMs, particularly during disease states. Moreover, the gut is highly compartmentalized in terms of structure (duodenum, jejunum, ileum, and colon), microbial and immune compositions, and function. Many studies on GSDMs only focused on one specific part of the intestine. Spatiotemporal analysis of GSDM distribution and expressional regulation will shed light on their functional roles.

2. Gasdermin cleavage and function

GSDMs play a fundamental role as pore formers of the cell membrane and serve as the executors of pyroptosis [24-26]. Additionally, they are implicated in the promotion of NETosis, a program for formation of neutrophil extracellular traps (NETs) [27, 28]. However, the GSDM pores do not always lead to cell death. Under many conditions, the GSDM pores facilitate the rapid cytokine release and ions fluxes without inducing cell death [4, 29, 30]. The concept of rapid cytokine release was first developed for inflammasome activation in immune cells [31-35]. These pores could be repaired by the endosomal sorting complexes required for transport (ESCRT) or caspase-7 activated acid sphingomyelinase [31, 36]. Furthermore, GSDMs exhibit multifaceted functions beyond pore formation, for example cell proliferation, mucin production, and the maintenance of immune tolerance to food [10, 37, 38].

GSDMA/GSDMA1-3

For cell-lytic function, human GSDMA N-terminus (GSDMA_NT) has been demonstrated to be a pyroptosis executor in vitro as early as hGSDMD [26]. However, the enzymes that cleave GSDMA in vivo remained unknown for an extended period. Recent studies by Deng et al. and LaRock et al. have shed light on this issue, revealing that Streptococcal pyrogenic exotoxin B (SpeB) from Group A Streptococcus (GAS) cleaves GSDMA and triggers pyroptosis [39, 40] (Figure 1A, B). These results were substantiated through experiments using Gsdma1 KO and Gsdma1-3 KO mice. GAS is a pathogen usually found on the skin and throat. Human GSDMA and mouse Gsdma1-2 may be expressed in specific populations of EECs and goblet cells. It presents an intriguing avenue for research to investigate whether SpeB or other bacterial toxins can cleave GSDMA/GSDMA1-3 in intestinal cells. Until now, the physiological role of hGSDMA/mGSDMA1-3 in the intestine is still unknown and needs further study.

Figure 1. Cleavage of GSDMs.

Human (A) and mouse (B) GSDMs are cleaved in the linker region by microbial and host-derived proteases (listed on top of the proteins in the figure) to release the N-termini (NT), allowing their oligomerization and pore formation on membranes. GSDMB and GSDMD can be inactivated by cleavages in the NT domain by Caspases and viral protease 3C (drawn below the proteins). While GSDMA is not expressed in the intestine, Streptococcal pyrogenc exotoxin B (SpeB) from Group A Streptococcus cleaves GSDMA and triggers pyroptosis. Human GSDMC can be activated by Cas-8 in cancer cells; however, such site is not conserved in all mouse GSDMCs and GSDMC cleavage enzymes in the intestine have not been reported. Differential cleavage events have been reported between human and mouse GSDMD, suggesting species-specific activation and function.

GSDMB

While GSDMB is a well-known asthma related gene, its role in the intestine remains largely unexplored. Recent reports have suggested that GSDMB expressed in epithelial cells can be cleaved at K229 and K244 residues by Granzyme A (GZMA) from cytotoxic CD8⁺ T cells or NK cells [41, 42] (Figure 1A). The cleaved GSDMB N-terminus triggers pyroptosis and cytokine release [41]. This process can be inhibited by IpaH7.8 effector protein secreted by enteroinvasive Shigella flexneri [42]. It is important to note that different GSDMB isoforms exhibit varying cell-lytic activities. These isoforms are equally targeted by IpaH7.8, but the presence of exon 6 determines their pore-forming activity [43-46]. Notably, some epithelial cell lines used in these reports originate from the intestine, such as Caco2, SW837 and SK-CO-1. However, in vivo evidence supporting the pathophysiological cell-lytic role of GSDMB in response to pathogen invasion is still lacking.

Full-length GSDMB can enhance the activity of Caspase-4 to promote non-canonical pyroptosis in THP1 cells [13]. The functional region mediating this signaling pathway has been identified within the 1-91 amino acids of GSDMB [47]. The activation of apoptotic caspase-7 inhibits non-canonical pyroptosis by cleaving GSDMB at D91 [47, 48] (Figure 1A). After cleavage, the region spanning from GSDMB 92-417 amino acids binds to the 1-91 amino acid sequence and blocks its function. This pathway was confirmed using GSDMB BAC (Bacterial Artifact Chromosome)-transgenic mice that exhibit a severe septic phenotype. The inhibition or deficiency of caspase-7 in GSDMB-transgenic mice exacerbate disease severity [47]. As GSDMB is also expressed in T cells and macrophages in the intestine, albeit at a lower level than in epithelial cells, it is plausible that GSDMB may function as a non-canonical pyroptosis promoter in these immune cells.

GSDMB also exhibits non-cell-lytic function. High levels of GSDMB are found in intestinal epithelial stem cells and TA cells. Recently, Rana et al. reported that GSDMB enhances epithelial restitution in human intestine organoids, and this effect is dependent on PDGF-A-mediated FAK phosphorylation [10] (Figure 2A).

Figure 2. Cell type-specific function of GSDMs in pathogenic infections and intestinal inflammation.

(A) In intestinal epithelial cells (IECs), human full-length GSDMB (B_FL) regulates proliferation and restitution for tissue repair. GSDMB N-terminus (B_NT) and GSDMD N-terminus (D_NT) can directly kill bacteria by forming pores on their membrane. Through not fully understood mechanisms, GSDMC pore (C_NT) formation upon helminth infections or IL-10 knockout exacerbates colitis. GSDMD executes diverse functions in IECs by inducing pyroptosis and releasing cytokines to restrain bacterial and viral infections while promoting colitis. Interestingly, a 13-kD fragment of GSDMD (D_13KD) promotes CIIA and MHC-II transcription to support Tr1 induction and food tolerance. Activated GSDME (E_NT) leads to pyroptosis and facilitates enteroviral pathogenesis and colitis. (B) In goblet cells, cell membrane GSDMC_NT and GSDMD_NT pores mediate unconventional secretion of IL-33 to expulse helminth and calcium influx for mucin exocytosis, respectively. (C) In macrophages, full-length GSDMB (B_FL) activates caspase-4 and −11 to promote non-canonical pyroptosis. GSDMD_NT pores promotes cytokine release, PIT and pyroptosis to prevent bacterial pathogens and colitis. (D) In neutrophils, GSDMD_NT-dependent pyroptosis facilities while NETosis inhibits bacterial infections. Conversely, GSDME_NT-mediated pyroptosis and IL-1β secretion defends Yersinia infection. (E) In innate lymphoid cell type 3 cells, GSDMD_NT trigger pyroptosis to defend Salmonella infection in mice.

GSDMC/GSDMC1-4

Human GSDMC can be cleaved by caspase-8 at D240 or D365, under the stimulation of α-KG or hypoxia, respectively (Figure 1A). The cleaved GSDMC N-terminus forms pores on cell membrane and trigger pyroptosis in human breast cancer cells and Hela cells [49, 50]. However, in mice, only the GSDMC4 could be cleaved by caspase-8 at D233 (Figure 1B). This residue is not conserved in mouse GSDMC1-3. Moreover, D365 site is not conserved in murine GSDMCs [50]. In IECs, murine GSDMC2-4 expression is induced by helminth infections [4, 17], which in turn enhances host type 2 immunity for helminth expulsion. While GSDMC cleavage can be detected during helminth infections, protease(s) that cleave intestinal GSDMC in vivo remain enigmatic.

Does GSDMC2-4 activation trigger intestinal pyroptosis during helminth infections? While Xi et al. reported that murine GSDMC2-4 expression is induced by IL-4 and IL-13 [17], type 2 cytokines have never been shown to induce cell death and thus their reported lytic cell death requires a greater scrutiny. On the other hand, Zhao et al. demonstrated that murine GSDMCs undergo cleavage and membrane localization during helminth infections using a validated antibody [4]. Activated by type 2 cytokines and STAT6 signaling, GSDMC pores serve as a conduit for unconventional IL-33 secretion in goblet cells (Figure 2B), without significant lytic cell death [4]. It is worth noting that murine GSDMCs is expressed in all villous IECs besides goblet cells, suggesting that GSDMC proteins may have other yet-to-be defined, non-lytic functions in addition to IL-33 release.

Recently, Du et al. found that Lgr5⁺ stem cell apoptosis and premature death of mice resulted from Mettl14 depletion are associated with loss of Gsdmc2-4 [51]. GSDMC RNAi knockdown resulted in apoptosis in human colonic organoids. However, these results are counterintuitive and could be off-target effects of RNAi experiments. Instead, independently generated whole-body or IEC-specific Gsdmc1-4 KO mice develop normally [4, 52], suggesting that targets other than Gsdmc genes mediate Mettl14’s function on stem cell survival. Finally, human GSDMC is very low expressed in the intestine and its regulation, cleavage, and function are almost completely unknown.

Human GSDMD/Mouse GSDMD

GSDMD is the most broadly expressed and well-studied member of the GSDM family. It can be cleaved by various enzymes, generating different pore-forming N-termini (Figure 1). In canonical and non-canonical pyroptosis, hGSDMD is cleaved at D275 by caspase-1 and caspase-4/5 (mGSDMD is cleaved at D275 by caspase-1 and caspase-11), respectively, and then becomes the most prominent executor of pyroptosis [24-26]. When infected by Yersinia, caspase-8 cleaves mGSDMD and induces pyroptosis in macrophages [23, 53]. mGSDMD in neutrophils and monocytes can be cleaved by cathepsin G at L274 [54]. mGSDMD in neutrophils can also be cleaved the neutrophil elastase at V251 site [27, 55]. These cleaved GSDMD N-terminus form pores mainly on the cell membrane, as well as on the membrane of other organelles, like mitochondria, azurophilic granules, and autophagosomes [56-58]. Conversely, activated caspase-3 and -7 can cleave hGSDMD at D87 site (mGSDMD at D88) to generate a non-pore-forming 13kD N-terminus [59, 60]. This apoptotic cleavage of hGSDMD/mGSDMD blocks pyroptosis in vitro [59, 60], but is not potent enough to inhibit pyroptosis in vivo [61].

Although mGSDMD is dispensable for intestinal epithelium development [12, 62], it plays a multifaceted role in homeostatic responses to external stimulations. mGSDMD in goblet cells promotes mucin granule exocytosis through calcium influx-mediated F-actin disassembly [63], thus protecting the host from microbial and pathogen invasion (Figure 2B). Recently, He et al. found that dietary antigens activate Caspase-3 and -7 to generate the apoptotic 13KD mGSDMD fragment (mGSDMD_13kD) in duodenal IECs. mGSDMD_13kD then localizes to nucleus and upregulates CIITA and MHCII transcription, leading to Tr1 cell induction and immune tolerance to food antigens [38] (Figure 2A). Furthermore, hGSDMD/mGSDMD N-terminus also form pores on the mitochondrial membrane and promote pyroptosis and necroptosis [64, 65].

Glial cells are closely associated with enteric neurons, and their interaction modulates internal reflexes within the intestine. Enteric neuronal pyroptosis is reported to be a contributing factor to colonic dysmotility induced by a Western diet [66]. Although GSDMD is expressed in glial cells, it is so far unknown if GSDMD is responsible for the pyroptosis of enteric glial cells.

Human GSDME/Mouse GSDME

The pyroptosis induced by GSDME is also called the secondary necrosis. In GSDME highly expressed cells, apoptotic activated caspase-3 cleaves GSDME and leads to pyroptosis prior to apoptosis. Both human and mouse GSDME is cleaved by caspase-3 and −7 at D270 [21, 67] (Figure 1). In addition, granzyme B from cytotoxic cells can cleave hGSDME/mGSDME at the same site [68].

Although GSDME N-terminus forms pores on the cell membrane, it does not always lead to lytic cell death. The hGSDME/mGSDME pores are also important for IL-1β release independent of its pyroptotic ability in macrophages [69]. Furthermore, in TH17 cells stimulated by Candida albicans, hGSDME pores are important for rapid IL-1α release and NLRP3 inflammasome activation [70]. Notably, TH17 cells resist pyroptosis despite hGSDME pores. However, whether GSDME mediates IL-1α secretion in intestinal TH17 requires further investigation. Finally, hGSDME/mGSDME N-terminus not only form pores on the cell membrane, but also permeabilize mitochondria and activate inflammasomes [57].

3. Gasdermins in Microbe-Host Interactions

The intestinal microbiota is a diverse community comprising bacteria, virus and fungi. The interactions among microbiota, IECs and immune cells maintain the intestinal homeostasis. When the physical and chemical barriers formed by IECs are compromised, the commensal microbiota can cause inflammation. The majority of isolated human gut commensal bacterial strains can induce innate immune responses and promote the secretion of cytokines. Particularly, Proteobacteria have been observed to stimulate IL-1β release when co-cultured with innate immune cells [71]. Consistently, mGSDMD targets Proteobacteria in the colon of high-fat diet-feeding mice to protect host from systemic endotoxemia, suggesting GSDMs and pyroptosis may be involved in these immune responses [72].

One of the foremost roles of GSDMs in the intestine is to function as a crucial defense mechanism against pathogen invasions. The studies about Salmonella and Shigella infected mice reveal the roles of GSDMs in the intestines, which is reviewed recently [73]. Although GSDMA is not reported to participate in these processes largely due to its low expression in the intestine, other GSDMs are important for defending various pathogen invasion.

GSDMB

During bacterial infections, GSDMB is cleaved by GZMA from NK cells in IECs. The cleaved N-terminus can target cardiolipin and lipid A, then directly form pores in membranes of Gram-negative bacteria (Figure 2A), such as E.coli, Citrobacter rodentium, Enterobacter cloacae, Shigella and Salmonella. On the other hand, pathogenic microbes are able to target GSDMB as a virulence strategy. For examples, IpaH7.8 effector protein secreted by enteroinvasive Shigella flexneri ubiquitinates GSDMB and inhibits its cleavage [42]. Furthermore, during the infection of E.coli and Salmonella, GSDMB promotes the non-canonical pyroptosis to defend pathogens in macrophages [47] (Figure 2C). Collectively, GSDMB plays an important role in defending bacterial infections.

GSDMC/GSDMC1-4

In mice, Gsdmc genes, not other Gsdm genes, can be drastically induced by parasitic helminths that infect the gut [4, 74]. Type 2 cytokines and STAT6 activation via coordinated phosphorylation and O-GlcNAcylation are absolutely required for the transcriptional upregulation of Gsdmc genes [4]. GSDMC cleavage and cell membrane localization upon helminth infections can be observed in IECs; however, proteases that activate GSDMC remain unidentified. In goblet cells, GSDMC_NT pores facilitate the unconventional secretion of IL-33, which is a pleiotropic “alarmin” that amplify type 2 inflammation for effective antihelminth immunity (Figure 2B) [4]. So far, no evidence suggests that GSDMC responds to other pathogens such as bacteria and viruses, indicative the unique role of GSDMC in antihelminth responses.

Human GSDMD/ Mouse GSDMD

Human and mouse GSDMD mediated pyroptosis and rapid cytokine release in innate immune cells and IECs are crucial for the host to defend against various pathogens, including Staphylococcus aureus, E. coli, Salmonella, Francisella, Shigella, Yersinia and rotavirus (Figure 2A)[23, 53, 75-83].

Shiga toxin 2 (Stx2) /LPS complex from pathogenic enterohaemorrhagic E. coli (EHEC) activated non-canonical pyroptosis in human macrophages in vitro (Figure 2C), although this process does not occur in mouse macrophages, as they lack the Stx receptor CD77 [84]. EHEC can also trigger hGSDMD-mediated pyroptosis in IECs. But the pathophysiological function of this process is not yet confirmed [75]. It is important to note that the role of GSDMD in neutrophils differs from that in macrophages (Figure 2D). Kambara et al. reveal that mGSDMD deficiency in neutrophils delays their death and enhanced host defense to E. coli in mice [55].

During Yersinia infection, a bacterial protein named YopJ can block TAK1 and trigger caspase-8 activation. Casapase-8 cleaves mGSDMD and induce pyroptosis in mouse macrophages (Figure 3C), but not in human macrophage [23, 53, 61]. Furthermore, apoptotic caspases cleaved mGSDMD at D88 also promote Yersinia defense [61]. Although mGSDME is also cleaved by caspase-8 during Yersinia infection, loss of mGSDME does not confer further protection [23].

Figure 3. Upregulated GSDMC/Gsdmc gene expression in IBD.

(A) Expression of GSDM genes, determined by RNA-seq, in a prospective cross-sectional cohort consisting of adult IBD patients and controls (GEO accession # GSE193677, n = 498 Crohn's disease, n = 421 UC and 243 controls) [105]. (B) Raw counts (left) and relative expression (right) of mouse Gsdm genes during DSS colitis (GEO accession # GSE131032). (C) Relative expression of mouse Gsdm genes during TNBS colitis (GEO accession # GSE35609).

hGSDMD/mGSDMD in various cell types responds to Salmonella infection. Recently, Fattinger and colleagues reported that only mGSDMD protect mice from acute Salmonella gut infection by using mouse lines lacking individual GSDMs or all GSDMs (Gsdma1-3, Gsdmc1-4, Gsdmd, and Gsdme). The role of GSDMB was not studied since it is not expressed in mice [85]. Zuo et al. reported that mGSDMD-dependent pyroptosis in macrophage protects against infection, which can be counteracted by Salmonella via its plasmid virulence C gene [86]. In IECs, Salmonella infection activates the NAIP-NLRC4 inflammasome, followed by hGSDMD/mGSDMD pore formation for epithelial contractions [77, 87]. However, the expulsion of IECs was also suggested not to depend on hGSDMD/mGSDMD [87-89]. Furthermore, mGSDMD triggers NETosis in neutrophils [27, 28] and pyroptosis in innate lymphoid cell type 3 cells (ILC3s) [90] to defend Salmonella infection in mice (Figure 2D, E). NETosis elicits NETs (Neutrophil Extracellular Traps), those consist of a DNA scaffold adorned with granule-derived proteins, including enzymatically active proteases and antimicrobial peptides. GSDMD-dependent pyroptosis in macrophage participates in the formation of a structure called pore-induced intracellular trap (PIT), which is different from NETs.

mGSDMD-dependent pyroptosis in macrophage participates in the formation of a structure called pore-induced intracellular trap (PIT), which is different from NETs. PIT refers to that the pores on membrane are small enough to retain organelles and bacteria. PIT traps intracellular bacteria and promotes recruitment of and efferocytosis by neutrophils [91]. IL-1β, IL-18, eicosanoids and virulence factors are important for promoting neutrophil recruitment in this process [91-93]. Furthermore, hGSDMD/mGSDMD N-terminus can also directly bind to the membrane of Gram-negative and Gram-positive bacteria, make pores, and kill the bacteria in vitro [76, 94] (Figure 2A, C). However, whether it can be functional during pathogen infection in vivo is not reported.

mGSDMD promotes AIM2 inflammasome activation to protect the host against Francisella novicida, but the downstream of AIM2 is unclear [81]. Nlrp9b is specifically expressed in IECs and rotavirus-infected IECs exhibit Nlrp9b-hGSDMD/mGSDMD dependent pyroptosis, which helps host restrict rotavirus infection [83]. On the other hand, in Stat1-deficient mice, the Nlrp3/Caspase-1/mGSDMD canonical pyroptosis pathway is immunopathogenic [95].

Human GSDME/Mouse GSDME

hGSDME/mGSDME also protects the host from pathogens. mGSDME is cleaved by apoptotic caspase-3 and activate pyroptosis to induce severe disease during enterovirus 71 (EV71) infection [96] (Figure 2A). mGSDME in neutrophils defends Yersinia infection through inducing pyroptosis and IL-1β release (Figure 2D), but not dependent on NETosis [97]. The cleavage of hGSDME/mGSDME can be activated by TNFα or bacteria toxins, such as cytolethal distending toxin (CDT) from Campylobacter jejuni [98, 99].

To summarize, GSDMs are key regulators of infection (Figure 2). They can directly kill bacteria by pore-forming on bacterial cell membrane and trigger rapid inflammation through host cell pyroptosis and non-cell-lytic cytokine release. They also contribute to the formation of PIT and NET to trap pathogens and disrupt the niche of pathogen replication in IECs. However, an excessive pyroptosis triggered by GSDMs can be detrimental, impeding the host's defense and promoting disease severity during certain viral infections. Therefore, the outcome of host-pathogen interactions can be determined by the location, expression, and activity of GSDMs.

4. Gasdermins in Inflammatory Bowel Disease (IBD)

GSDMs serve as pivotal regulators of intestinal inflammation. The genome-wide association study (GWAS) [100] and data mining efforts [101] have revealed that SNPs of GSDMA, GSDMB, GSDMD and GSDME are associated with IBD. Molecular mechanisms of GSDMs in IBD pathogenesis; however, are incompletely understood.

GSDMA/GSDMA1-3

Although SNPs of GSDMA are associated with IBD, action mechanisms of GSDMA in the intestine remain poorly characterized. Intestinal GSDMA seems to be only expressed by EECs and goblet cells, both of which contribute critically to barrier function and mucosal homeostasis. The GSDMA gene is located near GSDMB on human chromosome 17q12-21, a region highly relevant to asthma and IBD [100, 102]. One plausible hypothesis is that GSDMA SNPs contribute to IBD by regulating GSDMB expression.

GSDMB

GSDMB is one of the most well-known genes associated with autoimmune disease, including asthma and IBD [10, 48, 100, 102-104]. However, mixed results have been reported regarding GSDMB expression by disease variants and in IBD patients [10, 13, 14, 100]. Rana et al. discovered that mutant GSDMB in IBD patients loses the ability to enhance epithelial restitution using cultured human intestine organoids [10] (Figure 2A). Functional characterization of genetic loci linked to GSDMB and whether GSDMB induced immune responses are involved in IBD etiology require more studies.

GSDMC/GSDMC1-4

Although GWAS analyses have not indicated GSDMC’s association with IBD, GSDMC is the only gasdermin gene showing significant upregulation in both ulcerative colitis (UC) and Crohn’s disease (CD) patients (Figure 3A), based on a recent transcriptome analysis on hundreds of intestinal biopsies [105]. In mice, Gsdmc genes (except for the silenced Gsdmc1) not only show higher expression in the intestine than other GSDMs, but also is the only being induced by experimental colitis (Figure 3B, C) [106, 107]. Moreover, mouse GSDMC cleavage is observed in chemically (DSS) and genetically (Il10^−/−) induced colitis [4]. Deletion of mouse GSDMCs attenuates DSS colitis exacerbated by helminth infections or IL-10 knockout, suggesting that mouse GSDMC-supported cytokine release contributes to inflammation [4] (Figure 2A). Gsdmc1-4 KO mice showed no significant differences of DSS colitis in naïve wildtype mice, indicating that helminth infection or IL-10 KO induced immune changes facilitate the expression and pro-inflammatory actions of mouse GSDMCs. Further investigations are required to determine the functional involvement of GSDMC in human IBD.

Human GSDMD/Mouse GSDMD

hGSDMD is upregulated in some IBD tissues, especially IECs [108, 109]. Canonical pyroptosis drives experimental colitis in mice. As the executor of canonical pyroptosis, the activation of mGSDMD promotes IL-18 release and colitis [29], while Gsdmd^−/− mice exhibit attenuated intestinal inflammation [108]. mGSDMD is also required for IECs FADD deficiency induced ileitis [110] (Figure 2A).

While the prominent role of hGSDMD/mGSDMD in macrophages is to induce pyroptosis and rapid release of IL-1 family cytokines [29, 111], mGSDMD-expressing macrophages in the intestine have been reported to inhibit cGAS-mediated inflammation. As a result, Gsdmd^−/− or Gsdmd^fl/fl;lyz2^cre mice display more severe colitis [112]. We suspect that these conflicting results on Gsdmd^−/− mice could be attributed to potential cell-specific roles of mGSDMD – epithelial mGSDMD promotes pyroptosis while macrophage mGSDMD pores mediate potassium efflux to inhibit cGAS-dependent type 1 interferon response [80, 112].

Human GSDME/Mouse GSDME

Epithelial hGSDME/mGSDME is predominant expressed in enterocytes and goblet cells. Tan et al. found that hGSDME participates in the pathogenesis of Crohn’s disease (CD). mGSDME mediates IECs pyroptosis and promotes mucosal inflammation in TNBS-induced colitis [98]. mGSDME KO mice are resistant to intestinal barrier loss [113, 114]. Furthermore, Gsdmd and Gsdme double knockout mice showed almost total resistance to DSS induced colitis [114] (Figure 2A).

As discussed above in section 3 and 4, actions of GSDMs in intestinal homeostasis are intricately balanced. Insufficient GSDM activity permits pathogen invasion, while excessive activation triggers inflammation (Figure 2). Therefore, when investigate the roles of GSDMs in IBD, it is imperative to exercise caution given their double-edged effects.

5. Gasdermins in Intestinal Cancers

Through a pan-cancer analysis, expression of all gasdermin genes has been demonstrated to be extensively perturbed and correlated with patient survival [115]. In colorectal cancer (CRC), GSDMA, GSDMB and GSDMD could negatively regulate cell migration, while GSDME is associated with tumor invasion and metastasis [115, 116]. However, studies of individual GSDM family members in gastrointestinal cancers are still scarce.

GSDMA is a target of LMO1 in TGFβ-dependent apoptotic signaling in gastric cancer, where it is typically suppressed [117]. But it remains undetermined if GSDMA actively contributes to intestinal malignancies.

GZMA cleaved GSDMB induces pyroptosis in cancer cells and promotes the clearance of cancer cells in CRC (Figure 4), suggesting GSDMB can be a tumor suppressor [41].

Figure 4. GSDMs in intestinal tumors and anti-tumor immunity.

In IECs, GSDMC and GSDME can act like oncogenes to promote colorectal cancer development. However, in cancer cells, GSDMB-, C-, D- and E-mediated pyroptotic cell death suppresses tumorigenesis, primarily by activating anti-tumor immunity. Conversely, tumor killing ability of CD8 T cells can be dampened by GSDMD and pyroptosis in dendritic cells and macrophages. Note that GSDMC in driving tumor pyroptosis was observed in breast cancers, not intestinal cancers.

GSDMC was firstly discovered in cancer cells [1, 118]. In a mouse CRC model with simultaneous loss of APC and TGFBR2, Gsdmc2 and Gsdmc4 genes were the most highly regulated [119]. By promoting cell proliferation, GSDMC may function as an oncogene for colorectal carcinogenesis. However, molecular mechanisms driving cancer proliferation remain unknown (Figure 4). In breast cancer, hypoxia facilitates the nuclear translocation of PD-L1, which cooperates with p-STAT3 to transcriptionally activate hGSDMC expression [49]. Subsequently, hGSDMC_NT pores are generated by TNFα-activated caspase-8 to induce pyroptotic cell death in breast cancer. Nonetheless, it is unknown if GSDMC_NT pores form in intestinal cancers and mediate tumor pyroptosis (Figure 4).

In gastric cancer, low hGSDMD expression promotes cell proliferation through regulating cell cycle-related proteins [120]. In CRC, hGSDMD levels are lower than non-cancerous control tissues [121]. Furthermore, lower hGSDMD levels correlate with poorer prognosis of CRC patients [122]. All these data suggest that hGSDMD suppress intestinal tumorigenesis. However, hGSDMD/mGSDMD induces the cell death of tumor associated antigen presenting cells, such as dendritic cells and macrophage, and impairs the function of CD8⁺ T cells in tumor [123] (Figure 4). When studying GSDMD in intestinal cancers, cell-specific actions shall be carefully considered.

hGSDME/mGSDME is downregulated in most cancer cell lines [20, 124] and demonstrated to suppress tumor growth by activating anti-tumor immunity [68]. However, in the intestine, mGSDME has been suggested to promote CRC in mice via HMGB1 released from pyroptotic IECs [125] (Figure 4). Surprisingly, Gsdme^−/− mice showed no differences compared with wildtype mice in chemically induced and in genetic intestinal cancer mouse models [126], possible due to mGSDME’s dual function in anti-tumorigenesis and pro-inflammation. Cell-specific knockout mouse model will in the future help delineate the precise role of hGSDME/mGSDME in intestinal cancers.

These contrasting results underscore the complex roles of GSDMs in intestinal cancers as well other tumors (Figure 4). No unified mechanism can be generalized to all GSDMs, all cells, or all tumorigenic stages. Just like many other cancer-related pathways, pyroptosis and mediated inflammation can be double-edged swords.

Conclusions and Perspectives

To die or not to die, that is the decision made by GSDMs based on their expression locations and the external or internal stimulations. GSDMs can induce inflammatory cell death of IECs and immune cells to regulate mucosal immunity and anti-tumor immune responses. Meanwhile, GSDM pores allow rapid cytokine and pathogen trap release, as well as intracellular PIT formation, which collectively help host defend pathogenic invasions. However, overactivation of GSDMs leads to inflammatory diseases like IBD and can sometimes promote pathogen infections. Finally, GSDMs non-canonically promote the cell proliferation and gene transcription.

Many open questions remain regarding GSDM in intestinal health and disease. What regulatory mechanisms control GSDM expression in time and space, particularly at the protein level? What proteases activate or inactivate GSDMs in different intestinal cell types in vivo, upon what microbial, immunological, damage, or stress signals? When it comes to pyroptosis and cytokine release, does compensation exist between different GSDM pores? Duplicated Gsdmc2-4 genes in mice demonstrate the similar expression pattern as human GSDMB gene, which was lost in mice. But do mouse GSDMC2-4 function similarly as human GSDMB in the gut? Future studies are needed to address species-specific as well as evolutionarily conserved properties of GSDMs. Answering these questions will provide additional mechanistic insight into GSDM biology and inform therapeutic interventions for intestinal disorders.