Molecular mechanisms and functions of pyroptosis in inflammation and antitumor immunity

Abstract

Canonically, GSDMD cleavage by caspase-1 through inflammasome signaling triggers immune cell pyroptosis (ICP) as a host defense against pathogen infection. However, cancer cell pyroptosis (CCP) was recently discovered to be activated by distinct molecular mechanisms in which GSDMB, GSDMC and GSDME, rather than GSDMD, are the executioners. Moreover, instead of inflammatory caspases, apoptotic caspases and granzymes are required for gasdermin protein cleavage to induce CCP. Sufficient accumulation of protease-cleaved gasdermin proteins is the prerequisite for CCP. Inflammation induced by ICP or CCP results in diametrically opposite effect on antitumor immunity due to the differential duration and released cellular contents, leading to contrary impacts on therapeutic outcomes. Here, we focus on the distinct mechanisms of ICP and CCP, and discuss the roles of ICP and CCP in inflammation and antitumor immunity representing actionable targets.

Introduction

Pyroptosis, whose Greek roots “pyro” and “ptosis” denote fire and falling off, respectively, has received increasing interest recently owing to several newly revealed, pivotal molecular events. Currently, pyroptosis is interpreted to be a result of an imbalance between intracellular and extracellular osmotic pressure generated by caspases or other proteolytic enzymes-dependent pore formation of gasdermins in the plasma membrane. These pores dissipate cellular ionic gradients, leading to spillage of cellular contents such as interleukin 1β (IL-1β) and high-mobility group box 1 (HMGB1) (Erkes et al., 2020; Shi et al., 2017). Pyroptotic cell death was originally observed by Arthur Friedlander in 1986, who described the atypical death of anthrax lethal toxin-treated mouse macrophages (Friedlander, 1986). From the 1980s onwards, pyroptosis has been recognized as an inflammatory macrophage death due to pathogen insults for the protection of the host. However, intensive activation of pyroptosis may cause severe tissue damage, organ failure, and lethal sepsis (Deng et al., 2018; Linkermann et al., 2014; Mandal et al., 2018; Wang et al., 2017). Interestingly, emerging evidence reveals that inflammasome-mediated pyroptosis in myeloid cells links inflammation with tumor development and immunity (Kantono and Guo, 2017; Kolb et al., 2014). At the same time, inflammasome-independent pyroptosis in cancer cells, initiated by hypoxia (Hou et al., 2020), anticancer drugs, and cytotoxic lymphocyte killing through gasdermin protein cleavage, provides another promising strategy for cancer treatment (Erkes et al., 2020; Wang et al., 2020c; Zhang et al., 2020; Zhou et al., 2020). A better understanding of the role of pyroptosis in both pro-cancer and anti-cancer potential would favor the exploitation and clinical translation of new and improved therapeutic approaches.

Molecular mechanisms of pyroptosis

Despite great progress in understanding the signaling mechanisms of pyroptosis, the pyroptosis executioner remained unknown until gasdermin D (GSDMD) was discovered as the substrate of caspase-1 and caspase-11, leading to increased attention to this inflammatory cell death of immune cells. Moreover, recent studies revealed that pyroptosis can also occur in normal tissue and cancer cells via distinct mechanisms from pyroptosis of immune cells. Here, we review the recent advances in molecular mechanisms of pyroptosis.

Pyroptotic cell death in immune cells

Many pathogens invade phagocytes such as macrophages and neutrophils for survival and replication within the host cells. Canonically, these intracellular pathogens can be detected by inflammasomes, which subsequently induce pyroptotic cell death of infected cells, releasing chemoattractants to recruit phagocytes to kill these cells (Kovacs and Miao, 2017) (Figure 1). Inflammasomes are large oligomeric complexes consisting of sensor and adaptor proteins and serve as the molecular platform to trigger activation of inflammatory caspases (Deets and Vance, 2021). Sensor proteins such as pattern recognition receptors (PRRs) are activated by specific pathogen-associated molecular patterns (PAMPs) such as lipopolysaccharide (LPS) and damage-associated molecular patterns (DAMPs) such as cytosolic DNA to form distinct inflammasomes (Guo et al., 2015; Lamkanfi and Dixit, 2014). For example, the NLR family pyrin domain-containing 1 (NLRP1) is activated by anthrax lethal toxin and Toxoplasma gondii (Chavarría-Smith and Vance, 2015; Hollingsworth et al., 2021); absent in melanoma 2 (AIM2) recognizes cytosolic double-stranded DNA (Lammert et al., 2020); and NLRP3 senses diverse stimuli such as extracellular ATP, particulate and crystalline matter, and pore-forming toxins (Elliott and Sutterwala, 2015; Magupalli et al., 2020). Furthermore, pyrin indirectly senses the inactivation of host Rho GTPases caused by bacterial toxins (Xu et al., 2014), and the NLR family of apoptosis inhibitory proteins (NAIP) detects the bacterial flagellum and type III secretion apparatus (Vance, 2015). Typically, following activation, these sensor proteins oligomerize with an adapter protein to bridge the inflammasome to pro-caspase-1. The sensor proteins AIM2, NLRP1, NLRP3, and pyrin signal to the adapter protein apoptosis-associated speck-like protein containing a CARD (ASC, also known as PYCARD) (Linder et al., 2020; Wang et al., 2021), while the NAIP recruits NLR family CARD domain-containing protein 4 (NLRC4) as an adapter (Broz and Dixit, 2016). The inflammasome acts as a scaffold for pro-caspase-1 autocleavage followed by IL-1β/IL-18 maturation processed by activated caspase-1 (Man and Kanneganti, 2016). However, a noncanonical inflammasome pathway was reported based on the observation that cytosolic LPS directly binds to the caspase activation and recruitment domain (CARD) of murine caspase-11 (human caspase-4) and then forms an oligomeric complex for caspase-11 activation and pyroptosis induction, in which inflammasomes are not required (Kayagaki et al., 2011; Rathinam et al., 2019; Shi et al., 2014).

Figure 1. Pyroptotic pathway in immune cells.

Canonically, pathogens drive inflammasome formation in macrophages to activate caspase-1 which subsequently cleaves GSDMD and promotes IL-1β/IL-18 maturation. GSDMD-N terminal domain oligomerizes and forms pores in cell membrane, leading to cell lysis and IL-1β/IL-18 release. However, gram-negative bacteria-derived lipopolysaccharide (LPS) is able to directly activate caspase-11 to cleave GSDMD and form GSDMD pores. Potassium efflux though GSDMD pores then triggers canonical inflammasome-induced pyroptotic pathway. Activated caspase-8 by TNFα signaling or Yersinia-derived YopJ-induced TAK1 inhibition has also been shown to cleave GSDMD to cause cell lysis, which could be negatively regulated by cFLIPL. GSDMD succination by fumarate prevents its cleavage by caspases, limiting its capacity of pyroptosis induction. Mg²⁺ blocks pyroptosis by inhibiting Ca²⁺ channel P2X7 which is required for GSDMD pore formation. Pyroptotic process could be reversed by ESCRT-III machinery-dependent plasma membrane repair.

The substrate hunting for inflammatory caspases culminated in the breakthrough discovery that GSDMD is a direct target of inflammatory caspases and the executioner of pyroptosis (Kayagaki et al., 2015; Shi et al., 2015). GSDMD belongs to a gasdermin family that includes four paralogs, termed GSDMA, GSDMB, GSDMC, and GSDMD, and two extended family members, GSDME and DFNB59 (also called PJVK). Mouse GSDMD contains a 242-amino-acid N-terminal domain and a 199-amino-acid C-terminal domain, connected by a 43-amino-acid linker loop. The N-terminal domain harbors the oligomerization and pore-forming activity that is held in check by the C-terminal domain (Shi et al., 2015). Upon mouse GSDMD cleavage by caspase-1 or caspase-11 at the residue Asp276 within the linker loop, the inhibitory C-terminal domain is removed to liberate the N-terminal domain, which then oligomerizes and forms functional pores with an inner diameter of 10-20 nm in the cell membrane, which allows IL-1β/IL-18 to be released from the cell easily (Ding et al., 2016; Liu et al., 2016; Liu et al., 2020b; Wang et al., 2020b). This permeabilization of the cell membrane disrupts its osmotic potential, resulting in cell swelling and eventual lysis (Ding et al., 2016). Besides plasma membrane, GSDMD pores insert into multiple organelle membranes, including mitochondria (Huang et al., 2020; Rogers et al., 2019), lysosome, autophagosome and azurophilic granules (Karmakar et al., 2020). At the early stage of pyroptosis, cells with GSDMD pores stay alive and some intracellular contents such as IL-1β and IL-18 are selectively released (Bulek et al., 2020; de Vasconcelos et al., 2019; Evavold et al., 2018; Karmakar et al., 2020; Tsuchiya et al., 2021; Volchuk et al., 2020).

Importantly, early pyroptosis could be negatively regulated via multiple mechanisms. Ca²⁺ influx through its channel P2X7 is required for pyroptosis induction, however, Mg²⁺ protects cells against lysis by inhibiting P2X7 (Wang et al., 2020a). Although GSDMD pores impair cell membrane integrity, endosomal sorting complexes required for transport (ESCRT) machinery is recruited to repair cell membrane upon GSDMD activation, limiting pyroptosis (Rühl et al., 2018). Interestingly, succination of GSDMD by fumarate blocks pyroptosis through converting the critical cysteine residues into S-(2-succinyl)-cysteines to prevent its interaction with caspases and subsequent GSDMD cleavage (Humphries et al., 2020; Pickering and Bryant, 2020). In addition, cellular FLICE-like inhibitory protein (cFLIP_L) inhibits LPS-induced pyroptosis in macrophage through suppression of respiratory complex II (succinate dehydrogenase) formation (Muendlein et al., 2020). The negative regulation of pyroptosis restricts excessive inflammation and ensures immune homeostasis during pathogen infection.

Pyroptotic cell death in cancer cells

Most gasdermins have been shown to harbor a pyroptotic pore-forming N-terminal domain (Ding et al., 2016), indicating that gasdermins other than GSDMD may also be able to induce pyroptosis under certain conditions when their gasdermin N-terminal domains are liberated. Increasing evidence shows that cleavage of gasdermins by various caspases triggers CCP (Table 1). In 2017, Wang et al. reported that chemotherapy-induced caspase-3 activation can trigger pyroptosis by cleavage of GSDME at the conserved site Asp270, causing normal-tissue toxicity (Wang et al., 2017). Soon after, GSDME/caspase-3-mediated pyroptosis was reported in chemotherapeutic drug-treated gastric cancer cells (Wang et al., 2018) and predicted palmitoylation sites C407 and C408 in GSDME C-terminal domain were shown to be involved in chemotherapeutic drug-induced CCP (Hu et al., 2020), suggesting that gasdermin-mediated pyroptosis may occur not only in immune cells but also in normal tissue and cancer cells. Since then, there has been a surge of studies of gasdermin-induced pyroptotic cell death in cancer. For example, small-molecule inhibitors of kinases, including MEK, BRAF, EGFR, ALK, and KRAS, promoted caspase-3-induced cleavage of GSDME and caused durable tumor regression in lung cancer and melanoma (Erkes et al., 2020; Lu et al., 2018). In response to iron-elevated ROS, oxidized Tom20 activates caspase-3 through the mitochondrial pathway, culminating in GSDME-dependent pyroptosis in melanoma (Zhou et al., 2018). Moreover, Hou et al. established that nuclear PD-L1–mediated GSDMC expression switched TNFα-induced apoptosis to pyroptosis under hypoxia in breast cancer through caspase-8 cleavage of GSDMC, leading to tumor necrosis, which was also observed in lung cancer, liver cancer, ovarian cancer, and melanoma (Hou et al., 2020). Existing evidence demonstrate that clearance of tumor cells by killer lymphocytes via apoptosis or ferroptosis (Martínez-Lostao et al., 2015; Wang et al., 2019). Interestingly, two recent studies demonstrated that granzyme B (GZMB) derived from activated natural killer (NK) cells and cytotoxic T lymphocytes triggers caspase-dependent pyroptosis in target cells by direct cleavage of GSDME at the same site as caspase-3 (Liu et al., 2020a; Zhang et al., 2020), while another recent study similarly demonstrated that granzyme A (GZMA) released by NK cells and cytotoxic T lymphocytes directly cleaved GSDMB and activated caspase-independent pyroptosis and that introducing GZMA-cleavable GSDMB into murine cancer cells caused tumor rejection (Zhou et al., 2020). In addition to caspase-1, GSDMD has also been shown to be cleaved by caspase-4 under α-NETA (2-naphthoylethyltrimethylammonium) treatment to induce pyroptosis in ovarian cancer (Qiao et al., 2019). Chao et al. reported that apoptotic caspase-3/6/7, but not the inflammatory caspases, cleave GSDMB at ₈₈DNVD₉₁ within the N-terminal domain, indicating that GSDMB is a potential executioner in apoptosis stimuli-treated cancer cells (Chao et al., 2017). These studies of pyroptosis in cancer cells clearly show that cancer cells are provided with a distinct molecular mechanism of pyroptosis from immune cells in at least five characteristics: (i) agents or factors classically inducing apoptosis serve as pyroptotic triggers in cancer cells with sufficient expression of gasdermins; (ii) inflammasomes are not required for caspase activation and pyroptosis induction; (iii) apoptotic but not inflammatory caspases are involved in gasdermin cleavage; (iv) pyroptosis occurs in a caspase-independent manner in cancer cells in the presence of other active proteases that are able to liberate the gasdermin N-terminal domain, such as granzymes; (v) pyroptosis occurs with only rare IL-1β maturation and release. Evidently, there is no uniform pyroptosis pathway in cancer cells; however, sufficient expression of cleavable gasdermin proteins by an active protease is required.

Table 1.

Gasdermin cleavage by various proteases triggers CCP.

Gasdermin	Protease	Inducer	Reference
GSDMA	N/A	N/A	N/A
GSDMB	Caspase-3	N/A	(Chao et al., 2017)
	Caspase-6	N/A
	Caspase-7	N/A
	GZMA	Killer lymphocytes	(Zhou et al., 2020)
GSDMC	Caspase-6	N/A	(Hou et al., 2020)
	Caspase-8	chemotherapeutic drugs, TNFα
GSDMD	Caspase-4	α-NETA	(Qiao et al., 2019)
GSDME	Caspase-3	Chemotherapeutic drugs, TNFα	(Wang et al., 2017; Wang et al., 2018)
		ROS, iron	(Zhou et al., 2018)
		Kinase inhibitors	(Lu et al., 2018; Wu et al., 2019)

N/A: not applicable

Pyroptosis of immune cell and cancer cell in tumorigenesis and cancer development

Both ICP and CCP affect tumorigenesis and cancer development. The effect of ICP on tumorigenesis and cancer development depends on inflammasome signaling and pyroptotic cytokine production primed by PAMPs and DAMPs in the tumor microenvironment. While CCP primes or suppresses antitumor immunity as a result of different intensity and duration of CCP. Here, we highlight recent progress in our understanding of the importance of ICP and CCP in cancer.

Diametric roles of inflammasomes of immune cells in tumorigenesis and cancer development

The NLRP3 inflammasome is important in protection against tumor progression. Mice with a genetic absence of NLRP3 in hematopoietic cells show high susceptibility to colitis-associated colon cancer induced by the DNA-damaging agent azoxymethane (AOM) and dextran sulfate sodium (DSS) (Allen et al., 2010; Zaki et al., 2010a; Zaki et al., 2010b), which may be attributed to the limitation of colitis by GSDMD via controlling cGAS (cyclic GMP-AMP synthase)-mediated inflammation (Ma et al., 2020). However, another report argues that there is no significant difference in susceptibility to AOM- and/or DSS-induced colon cancer between wild-type and NLRP3-knockout mice treated with AOM and DSS (Hu et al., 2010). In addition, activation of the NLRP3 inflammasome in dendritic cells by dying tumor cell-released ATP primes IFN-γ-producing CD8+ T cells and boosts IL-1β-dependent antitumor immunity (Ghiringhelli et al., 2009).

Other inflammasome sensors also act as tumor suppressors in the initiation and progression of cancer. The NLRC4/NAIP5 inflammasome, together with TLR5 in macrophages recognizes over-expressed flagellin in tumor cells, leading to tumor clearance by the innate immune response and tumor-specific CD4+ and CD8+ T cells in vivo (Garaude et al., 2012). Moreover, NLRC4 plays a critical role in chemokine and cytokine production in tumor-associated macrophages (TAMs) and in the generation of protective IFN-γ–producing CD4+ and CD8+ T cells (Janowski et al., 2016). AIM2 reduces the phosphorylation of AKT by limiting activation of DNA-dependent protein kinase, which results in decreased cell proliferation and tumor burden in AOM/DSS-induced colorectal cancer models (Wilson et al., 2015). Clearly, the studies above provide the direct evidence for the protective roles of inflammasome in suppressing tumor development.

Despite the protective functions of inflammasomes against cancer, their detrimental role in tumorigenesis and cancer progression has been increasingly reported. Caspase-1 promotes TAMs differentiation by peroxisome proliferator-activated receptor gamma (PPARγ) cleavage-mediated medium-chain acyl-CoA dehydrogenase (MCAD) suppression, fueling tumor progression (Niu et al., 2017). Moreover, the caspase-1 activity of myeloid-derived suppressor cells (MDSCs) contributes to tumor-intrinsic MyD88-dependent carcinogenesis although caspase-1 as p63 target in cancer cells suppresses tumor growth (Celardo et al., 2013). A group using multiple animal models found that mice lacking NLRP3 are resistant to carcinogenesis and metastases (Chow et al., 2012). The authors demonstrated that a subgroup of MDSCs, CD11b+Gr-1^intermediate, is expanded in the lung tumor microenvironment of mice deficient for NLRP3 and secretes remarkable levels of CCL5 and CXCL9 chemokines, which attract and activate NK cells to suppress tumor growth and metastases (Chow et al., 2012). In contrast, vaccination of NLRP3−/− mice with a dendritic cell vaccine resulted in a 4-fold improvement in survival compared with control animals due to the reduced number of tumor-associated MDSCs in the tumor microenvironment, whereas adoptive transfer of NLRP3−/− MDSCs was less efficient in reaching the tumor site (van Deventer et al., 2010), implicating a requirement of NLRP3 for accumulation of MDSCs in a tumor microenvironment that promotes tumor growth. Due to the fact that DSS could disrupt the composition of commensal bacterial in gut (Wirtz et al., 2017), differences in the intestinal microbiome might contribute to the diametric effect of NLRP3 in DSS-induced colorectal cancer (Zhen and Zhang, 2019).

Diametric roles of pyroptotic cytokine IL-18 from immune cells in tumorigenesis and cancer development

IL-18, a member of the IL-1 family, was first discovered for its capability to enhance IFN-γ production from anti-CD3-stimulated T cells (Okamura et al., 1995). It is broadly considered as a pro-inflammatory cytokine and a suppressor of cancer (Figure 2). IL-18 is one of the main cytokines released by pyroptotic immune cells (Figure 1). The capability of inflammasomes to facilitate tumor rejection relies at least partially on inflammasome-mediated secretion of IL-18, which promotes epithelial barrier repair against damage in colitis-associated colon cancer. Multiple studies report that conditional deletion of IL-18 in mice increases their susceptibility to DSS-induced intestinal inflammation and development of adenocarcinomas of the colon (Salcedo et al., 2010; Takagi et al., 2003; Zaki et al., 2010b). Administration of recombinant IL-18 into inflammasome-deficient mice significantly reduces the incidence of colon tumors treated with AOM and DSS (Dupaul-Chicoine et al., 2010; Zaki et al., 2010b), indicating that IL-18 secretion downstream of the inflammasome activation confers protection against colon tumorigenesis and is a potential anticancer drug candidate for certain types of colon cancer lacking inflammasomes. The NLRP3 inflammasome-mediated production of IL-18 dampens metastasized colonic tumor cells in the mouse liver by inducing NK cell maturation and tumoricidal activity (Dupaul-Chicoine et al., 2015). IL-22 has been reported to facilitate intestinal stem cell-mediated epithelial regeneration against intestinal injury (Lindemans et al., 2015). A study reported that IL-18 upregulates the blood level of IL-22 by downregulation of IL-22-binding protein, which protects against intestinal damage induced by severe inflammation or later stages of tumor progression (Huber et al., 2012). However, injection of recombinant IL-18 into mice twice a week exacerbated the progression of B16-F10 metastases in the lung, whereas daily administration of recombinant IL-18 for 5 days reduced tumor prevalence (Terme et al., 2011). Moreover, IL-18-knockout C57BL/6 mice are more susceptible to B16-F10 tumor metastasis (Chow et al., 2012). These observations suggest that a sustained circuit of IL-18 may fuel antitumor immunity and induce an anti-tumorigenic effect, yet long-interval or occasional exposure to IL-18 may cause inflammation and worsen metastasis. Additionally, IL-18 positively regulates the production of IFN-γ from T helper 1 lymphocyte and macrophage activation and enhances the cytotoxicity of NK cells and CD8+ T cells, inhibiting tumor development in colitis-associated colon cancer (Novick et al., 2013; Okamura et al., 1995). The antitumor activity of IL-18 well exemplifies the tumor rejection effect of ICP. However, in established melanoma, colon cancer, and multiple myeloma models, IL-18 promotes tumor progression (Cho et al., 2000; Nakamura et al., 2018; Terme et al., 2011). A study shows that IL-18 drives generation of MDSCs, leading to accelerated tumor progression in multiple myeloma (Nakamura et al., 2018). In established B16-F10 model, IL-18 decreases intracellular reactive oxygen intermediate production and Fas ligand expression to fuel tumor growth (Cho et al., 2000). Terme et al found that low-level IL-18 acts to suppress the NK cell arm of tumor immunosurveillance by upregulating PD-1 expression (Terme et al., 2011). Therefore, IL-18 exhibits distinct impacts on tumor development and established tumors.

Figure 2. Diverse, cell-context dependent impacts of ICP-released IL-1β and IL-18 on tumor cells and tumor immune microenvironment.

IL-1β and IL-18 are the products of ICP, through which anti-tumor immunity and tumor growth are affected in a cell context-dependent manner. IL-1β promotes VEGFA expression and angiogenesis, and thus attracts MDSC into tumor microenvironment. It is well-known for its enhancement in tumor progression, oncogenesis, lymphangiogenesis, and metastasis and suppression in antitumor effect of chemotherapy drugs. Instead, the role of IL-18 in tumor prevention and suppression is impressive. It increases IFN-γ production of Th1 cells, boosts antitumor activity of macrophage, and fuels the cytotoxicity of killer lymphocytes. It protects intestine from damage, preventing intestine inflammation. Administration of IL-18 reduces tumor incidence and kills tumor cells in metastasis sites, leading to inhibition of tumor progression.

Diametric roles of pyroptotic cytokine IL-1β from immune cells in tumorigenesis and cancer development

IL-1β and IL-18, which both belong to IL-1 family, are the only cytokines processed by caspase-1 in pyroptotic immune cells. Unlike IL-18, IL-1β is well known for its tumor-promoting functions (Figure 2). IL-1β exerts the deleterious impact of inflammasome-induced pyroptosis on cancer by its relationship with tumor-infiltrating myeloid cells. Stomach-specific expression of IL-1β in mice attracts and activates MDSCs through an IL-1 receptor (IL-1R)/NF-κB pathway, leading to spontaneous gastric inflammation and cancer (Tu et al., 2008). Mice lacking IL-1R have delayed accumulation of MDSCs and inhibited primary and metastatic tumor progression (Bunt et al., 2007). The production of NLRC4 inflammasome–mediated IL-1β is elevated in tumor-infiltrating MDSCs induced by obesity, which drives cancer progression through adipocyte-mediated vascular endothelial growth factor A (VEGFA) expression and angiogenesis (Kolb et al., 2016). Importantly, IL-1β favors the resistance of tumor cells to chemotherapeutic drugs such as doxorubicin, gemcitabine, and 5-fluorouracil, underlining the significance of the inflammatory environment in chemotherapy for cancer patients (Bruchard et al., 2013; Feng et al., 2017; Mendoza-Rodríguez et al., 2017). Interestingly, the pro-tumorigenic effect of IL-1β is enhanced by hypoxia. TAM-derived IL-1β is increased by HIF-1α under hypoxia, which promotes hepatoma epithelial-mesenchymal transition and metastasis (Zhang et al., 2018). IL-1β acts through the COX2-HIF-1α pathway to promote lung oncogenesis (Jung et al., 2003; Wang et al., 2014). Moreover, S1P receptor 1 (S1PR1) on TAMs promotes lymphangiogenesis and metastasis via NLRP3-induced IL-1β production, implicating IL-1β plays an unappreciated role in promoting metastasis via the lymphatics downstream of S1PR1 signaling in TAMs (Weichand et al., 2017). Despite the well-known deleterious role of IL-1β in tumor growth, some studies show its protective function against tumor development. Absence of IL-1R fails priming of CD8+ T cells, which could be rescued by providing exogenous IL-1β (Ghiringhelli et al., 2009). In NLRP1 (−/−) mice, attenuated levels of IL-1β increases inflammation and tumor burden (Williams et al., 2015). Importantly, a recent study reported that blanket inactivation of IL-1R does not reduce progression of colorectal cancer, but deletion of IL-1R in T cells and epithelium does (Dmitrieva-Posocco et al., 2019). Whereas knockout of IL-1R in myeloid cells accelerates progression of colorectal cancer (Dmitrieva-Posocco et al., 2019), indicating that the roles of IL-1β in control of tumor growth is cell-type-specific.

Chronic and acute pyroptotic deaths of cancer cells exhibit distinct outcomes

CCP was first described in NCI-H522 lung cancer cells treated with chemotherapy drugs, including doxorubicin, actinomycin-D, bleomycin, and topotecan (Wang et al., 2017). In this study, GSDME was shown to be cleaved by the apoptotic caspase-3, inducing robust pyroptosis; since then, GSDME-induced pyroptosis in cancer cells has been reported in extensive studies. Currently, most studies show that inducing pyroptotic death of cancer cells contributes to tumor rejection. Among methods for inducing pyroptosis in cancer, kinase inhibition is emerging as a promising strategy (Erkes et al., 2020; Lu et al., 2018; Wu et al., 2019). Similarly, treatment of tumor cells with decitabine to elevate the GSDME expression level enhances the chemosensitivity of breast tumors (Fan et al., 2019). In particular, Zhang et al., from the standpoint of immunology, uncovered a mechanism of GSDME-mediated tumor inhibition in which GZMB from killer cytotoxic lymphocytes induces the cleavage and activation of GSDME in cancer cells, which promotes the phagocytosis of tumor cells by TAMs and increases the number and activation of tumor-infiltrating NK and CD8+ T cells (Zhang et al., 2020). Similarly, Zhou et al. showed that GZMA from cytotoxic NK and T lymphocytes cleaves another gasdermin, GSDMB, to trigger pyroptosis in target cells, causing tumor clearance in animal tests (Zhou et al., 2020). Of note, controlled induction of pyroptosis by nanoparticle-mediated selective delivery of active gasdermin proteins into cancer cells revealed that pyroptotic cell death of only 10% of tumor cells is sufficient to clear an entire 4T1 mammary tumor graft (Wang et al., 2020c). Overall, the introduction of intense pyroptosis of tumor cells, even of a small population, may suppress tumor growth.

Although mounting evidence shows that induction of pyroptosis slows tumor growth, the opposite function of pyroptosis in cancer development has also been observed in different experimental contexts. Large tumor size or rapid tumor growth leads to tumor necrosis due to hypoxia in the center of the tumor (Brown, 2007; Pollheimer et al., 2010; Tomes et al., 2003). Chronic tumor necrosis both accelerates tumor growth (Kono and Rock, 2008) and suppresses antitumor immunity (Karin and Greten, 2005). A recent study showed that PD-L1 translocates into the nucleus in response to hypoxia and transcriptionally turns on the expression of GSDMC, which is subsequently cleaved by TAM-derived TNFα-activated caspase-8, causing pyroptosis in breast cancer cells, resulting in tumor necrosis (Hou et al., 2020). GSDMC-mediated CCP accelerates tumor growth in immuno-deficient mice. Moreover, high expression of GSDMC correlates with poor prognosis in breast cancer patients (Hou et al., 2020). A similar clinical observation in lung adenocarcinoma showed that overexpression of GSDMC is a promising predictive factor for the poor prognosis of lung adenocarcinoma patients (Wei et al., 2020). In addition, GSDMC expression is upregulated by inactive mutation of TGF-β receptor type II, which fuels xenograft tumor growth in vivo (Miguchi et al., 2016). Furthermore, overexpression of GSDMB increases tumorigenesis and metastasis in breast cancer (Hergueta-Redondo et al., 2014). Consistently, clinical analysis of GSDMB expression in breast cancer patients demonstrates that its expression is associated with trastuzumab resistance in a patient-derived xenograft model and predicts poor clinical outcome in ERBB2-positive breast cancer (Hergueta-Redondo et al., 2016). A recent study analyzed the correlation of patient survival with gasdermin expression in The Cancer Genome Atlas (TCGA) database and showed that overexpression of GSDMB is associated with poor survival in adenoid cystic carcinoma, clear cell renal cell carcinoma, and prostate adenocarcinoma, while overexpression of GSDMD or GSDME predicts poor prognosis in uveal melanoma (Zhou et al., 2020). In an agreement with the notion that chronic inflammation and tumor necrosis favor tumor progression, these studies evidence that spontaneous and enduring pyroptosis of cancer cells facilitates tumor growth.

Pyroptosis, inflammation, and antitumor immunity

Pyroptosis from cancer cells and immune cells may result in different clinical outcome. Even with pyroptosis of the same cell type, the anti- or pro- tumor effects of pyroptosis are also determined by the interactions between pyroptotic cells and the surrounding microenvironment. Below, we discuss the connections between pyroptosis, inflammasome, and antitumor immunity (Figure 3).

Figure 3. CCP-induced acute inflammation enhances, however, ICP-mediated chronic inflammation inhibits anti-tumor immunity.

Killer lymphocytes including NK cells and T cells induce apoptosis or ferroptosis in gasdermin-negative cancer cells but trigger pyroptosis of cancer cells expressing GSDME and GSDMB (CCP), which promotes acute (short-term) inflammation and enhances killer lymphocytes cytotoxicity, leading to tumor growth suppression. Instead, pyroptosis of immune cells (ICP) including TAM and MDSC causes chronic (long-term) inflammation that inhibits killer cells activity and fuels tumor growth. The difference of inflammation duration between ICP and CCP causes distinct outcomes in tumor growth.

Cancer is often preceded by long-term inflammation and a micronecrotic setting, which originates from rounds of unscheduled or disordered necrotic cell death (Vakkila and Lotze, 2004). Multiple cancer types, including colon, lung, bladder, prostate, cervical, gastric, pancreatic, and esophageal, are associated with chronic inflammation (Balkwill and Mantovani, 2001). Microbial infection is considered one of the main contributors to chronic inflammation and contributes to 16.1% of cancer incidence worldwide (Balkwill and Mantovani, 2001; de Martel et al., 2012). For example, Helicobacter pylori, whose colonization stimulates a persistent inflammatory response, is the strongest known risk factor for gastric cancer (Polk and Peek, 2010). Recently, Nejman et al. generated and characterized an exhaustive catalog of tumor microbiomes across seven cancer types, including melanoma and lung, breast, ovarian, pancreatic, brain, and bone cancers (Nejman et al., 2020). It is interesting to note that microbiome composition is distinct according to tumor type, which is associated with clinical outcomes in immune checkpoint inhibitor-treated patients, and more importantly, bacteria within tumors are mostly localized inside both cancer and immune cells. Given that pyroptosis is an inflammatory form of cell death that is against pathogen infection, it is reasonable to speculate that the bacteria within tumors likely cause pyroptotic cell death and chronic inflammation, promoting tumor growth. Indeed, it has been reported that microbiome composition of pancreatic adenocarcinoma influences the host immune response and patient survival by cross-talk with the gut microbiome (Riquelme et al., 2019). Moreover, several inflammasomes, including NLRP3, NLRP6, and NLRC4, which are sensors of intracellular pathogens, have been implicated in tumorigenesis, although their role in cancer development is still controversial (Janowski et al., 2013). In addition to bacteria, hypoxia also induces pyroptosis of cancer cells in the central regions of large tumors, which leads to tumor necrosis that inhibits antitumor immunity and promotes tumor progression (Hou et al., 2020).

In agreement with the notion that inflammasome-induced pyroptosis of immune cells inside the tumor may contribute to tumorigenesis, a heavy infiltration of leukocytes, either intratumorally or peritumorally, has been reported in intensive cancer studies. The accumulation of leukocytes may, on one hand, be attributed to the inflammatory origins of a tumor, or may, on the other hand, result from a tumor per se that recruits leukocytes succeeding the release of cytokines and chemokines, which conversely accelerates tumor growth by expediting secretion of angiogenic and growth-promoting factors (Arenberg et al., 2000; Mantovani et al., 2002; Vakkila and Lotze, 2004). Typically, it has been well documented that TAMs and MDSCs are frequently associated with suppressive antitumor immunity and tumor progression, predicting poor prognosis in cancer patients (Gabrilovich and Nagaraj, 2009; Noy and Pollard, 2014).

Additional substantial evidence that chronic inflammation is a causative factor in tumor etiology comes from the findings that sustained use of non-steroidal anti-inflammatory drugs, such as aspirin, and of natural products that suppress inflammasome signaling to relieve inflammation, such as green-tea extracts, ginseng extracts, curcumin, and resveratrol, lowers the incidence of various cancer types, including lung, colorectal, esophageal, stomach, and ovarian cancers, as well as lymphoma (Chang et al., 2004; Garber, 2004; Ness and Cottreau, 1999; Wang et al., 2003). Furthermore, as described above, inhibitors targeting the inflammasome signaling process to modulate chronic inflammatory response exhibit promising antitumor effects (Kolb et al., 2014).

In contrast to the pro-tumor effects of ICP, several investigators have independently reported that acute pyroptotic cell death of tumor cells stimulates an inflammatory response in the tumor microenvironment and mobilizes potently anticancer immunity. Wang et al. showed that CCP, induced by the delivery of active GSDMA3 protein directly and specifically into tumor cells by a nanoparticle, resulted in tumor regression (Wang et al., 2020c). The authors further found that tumor cell pyroptosis markedly enhanced the infiltration of both the CD4+ and CD8+ T cell populations. However, the tumor regression was abrogated in immunodeficient mice or upon depletion of T cells but correlated with augmented anticancer immunity, suggesting that active gasdermin protein–induced tumor clearance is immune dependent (Wang et al., 2020c).

Zhang and colleagues observed a similar phenomenon in GSDME-induced tumor rejection, in which GSDME mediates CCP in vivo through direct cleavage by killer lymphocyte-released GZMB, which in turn increases infiltration of killer lymphocytes, including macrophages, NK cells, and CD8+ cells, and thus enhances CCP (Zhang et al., 2020). This positive feedback loop relies on pyroptosis-induced recruitment of killer lymphocytes, indicating the occurrence of immunogenic cell death. According to the gold-standard criterion of immunogenic cell death, mice should be protected from secondary tumor challenge after vaccination with cancer cells undergoing immunogenic cell death (Galluzzi et al., 2017). Indeed, when the researchers re-challenged mice vaccinated with wild-type or GSDME-overexpressing tumor cells, mice vaccinated with GSDME-overexpressing tumor cells, compared with the wild-type group, showed substantially delayed tumor growth and better tumor-free survival as well as significantly reduced tumor incidence. Moreover, authors of one recent study showed that CAR T cell-released GZMB induces caspase 3-mediated GSDME cleavage in target tumor cells and causes pyroptosis (Liu et al., 2020a).

Nonetheless, GSDME-induced immunogenic cell death has been also been confirmed in BRAFi + MEKi-driven melanoma regression by another group. BRAFi + MEKi treatment was shown to trigger GSDME/caspase-3-induced pyroptosis of melanoma cells, causing release of HMGB1 and increased cell surface expression of calreticulin, both of which are markers for immunogenic cell death (Erkes et al., 2020). HMGB1 and calreticulin then activated dendritic cells to expand CD4+ and CD8+ T cells, especially the activated (CD44+) and proliferating (Ki-67+) T cells, leading to durable tumor regression and extended overall survival (Erkes et al., 2020). Defective HMGB1 release or depletion of T cells resulted in more frequent tumor relapse and shortened overall survival, suggesting that BRAFi + MEKi caused durable tumor clearance in an immune-dependent manner. Importantly, for treatment of BRAFi + MEKi-resistant melanoma, the group screened a panel of therapeutic modalities, including radiation, chemotherapy, targeted inhibitors, and epigenetic inhibitors, and confirmed that only chemotherapeutic drugs such as etoposide and doxorubicin, which can cause GSDME cleavage and HMGB1 release to stimulate antitumor immunity, reverse resistance of melanoma to BRAFi + MEKi, suggesting a salvage therapy for BRAFi + MEKi-resistant melanoma. Furthermore, Zhou et al. identified GZMA, which is also derived from killer lymphocytes, as the critical factor that induces pyroptosis and tumor rejection by direct cleavage of GSDMB in GSDMB+ cancer cells (Zhou et al., 2020). The killing effect of GZMA is further enhanced through upregulated expression of GSDMB by killer lymphocyte-produced IFN-γ.

It may appear contradictory that inhibition of inflammasome signaling–mediated ICP and activation of CCP have both demonstrated antitumor effects, at least in experimental models. We propose that this seeming contradiction is due to differences in inflammation duration and contents of cellular spillage. Inflammasome signaling–mediated pyroptosis is generally chronic in the tumor microenvironment. Chronic inflammation resulting from chemical irritants, infectious microorganisms, or hypoxia contributes to inhibition of antitumor immunity and tumor growth (Hou et al., 2020; Vakkila and Lotze, 2004). Therefore, long-term use of non-steroidal anti-inflammatory drugs for relief of potential chronic inflammation benefits cancer prevention (Ulrich et al., 2006). In addition, the major pyroptotic cytokine IL-1β, resulting from the inflammasome signaling process, is well known for its ability to mobilize TAMs and MDSCs to promote tumor progression (Kaler et al., 2009; Tu et al., 2008). However, introducing pyroptosis in cancer cells by either elevating active gasdermin protein levels or inducing gasdermin cleavage leads to short-term, acute inflammation, which is able to counteract cancer development (Philip et al., 2004), and massive HMGB1 release, which has been shown to attract and activate dendritic cells to prime T cells (Dumitriu et al., 2005).

So far, most studies of tumor clearance induced by CCP show heavy infiltration of CD4+ and CD8+ T cells and cause tumor regression in an immune-dependent manner. In fact, pyroptosis as a form of “dirty death” could prime T cells, at least partially, in either chronic or acute inflammation, because immune checkpoint blockade enhances antitumor immunity not only in gasdermin+ tumors (Wang et al., 2020c; Zhou et al., 2020) but also in tumors with a specific intratumor microbial signature (Nejman et al., 2020). In addition, along with tumor growth, tumor necrosis is initiated and gradually aggravated in hypoxic regions of tumors, which results from nuclear PD-L1–mediated CCP in response to hypoxia (Hou et al., 2020). Furthermore, some apoptotic caspases such as caspase-8, which activates pyroptosis in GSDMC+ cancer cells (Hou et al., 2020), have also been shown to cleave GSDMD in immune cells to induce pyroptosis (Aizawa et al., 2020; de Vasconcelos et al., 2020; Doerflinger et al., 2020; Orning et al., 2018; Schwarzer et al., 2020; Zheng et al., 2020). Therefore, it is probable that inflammasome-mediated ICP and CCP are concurrent during tumor development or clinical treatment and that the clinical outcome depends on the contest between immunosuppressive factors and immune-active factors.

Concluding remarks

In this review, we summarize the mechanisms and functions of pyroptosis in cancer development and antitumor immunity. Clearly, our current knowledge of pyroptosis is just the tip of the iceberg. Many questions are still open to be answered. Considering that gasdermins are the executioners of pyroptosis, their expression level would be a critical factor that determines cell death patterns. However, it is still unclear the extent to which gasdermin proteins should be expressed to override apoptosis. Moreover, existing evidence suggests that the proteases responsible for gasdermin cleavage are not limited to caspases. It remains to be determined, however, what other types of proteases cleave gasdermins under certain conditions. Furthermore, is the oligomeric pore formed immediately after monomeric GSDMD-N domain insertion in the membrane, or is it assembled in the cytoplasm first and then trafficked as a single unit into the lipid bilayer? Does protein modification of gasdermins affect their membrane insertion or organelle specificity? How do spilled cellular contents differ between immune cells and cancer cells, and what is the influence of this spillage on antitumor immunity? Overall, more comprehensive mechanistic insights into pyroptosis in the tumor microenvironment will be instrumental in developing novel, efficacious therapeutics to eliminate cancer cells.