Background

One of the crucial functions of the innate immune system is to recognize DAMPs and PAMPs by pattern recognition receptors (PRRs) during microbial infection and sterile damage [1]. Some PRRs, such as Toll-like receptors (TLRs), are located in the cytoplasm membrane and endosome membrane to supervise extracellular and endosomal dangers [2]. In the cytosol, nucleotide-binding leucine-rich repeat receptors (NLRs), absent in melanoma 2 (AIM2), and pyrin are able to recognize cytosolic DAMPs and PAMPs [3]. Distinct from TLRs that eventually elevate pro-inflammatory cytokines, type I interferons, and chemokines at the transcription level, NLRs (NACHT, leucine-rich repeat and pyrin domain-containing 1 (NLRP1), NOD-, LRR- and pyrin domain-containing 3 (NLRP3), and NLR family apoptosis inhibitory protein (NAIP)/NLR family CARD domain-containing 4 (NLRC4)), AIM2, and pyrin initiate posttranslational mechanisms by assembling inflammasomes, a group of multicomponent complexes [4, 5]. Briefly, inflammasome sensors recruit caspase-1 family members with or without the assistance of apoptosis-associated speck-like protein-containing CARD (ASC) to initiate auto-cleavage of caspase-1. The activated caspase-1 cleaves precursors of GSDMD and IL-1 family members to release these cytokines and induce pyroptosis. The canonical and non-canonical inflammasome pathways are summarized in Fig. 1.

Fig. 1
figure 1

Overview of inflammasome activation

Full size image

Canonical inflammasomes are composed of sensors, ASC, and caspase-1 [5]. Once activated, these inflammasome sensors oligomerize and recruit ASC to form an “ASC speck” through pyrin–pyrin (PYD–PYD) interaction [6, 7]. Then caspase-1 is recruited to ASC through CARD–CARD interaction [6, 7]. However, exceptions have been reported in the activation of NLRP1 and NLRC4. CARD domain of the NLRP1 directly recruits caspase-1 through CARD–CARD interaction without ASC [8, 9]. However, human NLRP1 also recruits ASC through the PYD domain [10]. For NLRC4, caspase-1 can be recruited to NLRC4 with (through CARD–CARD interaction between NLRC4 and ASC [11]) or without (through CARD–CARD interaction between NLRC4 and caspase-1 [12]) ASC, although differences in size and duration of activated inflammasomes have been observed between these two kinds of NLRC4 inflammasomes [13]. The recruited caspase-1 (also known as caspase-11 in mice) dimerizes and autoclaves to generate p33/p10 species with full protease activity [13]. The cleaved caspase-1 is able to process pro-IL-1β at D26 and D116 and pro-IL-18 at D36 to produce active IL-1β and IL-18 [14]. GSDMD is also cleaved by the caspase-1 to release the amino-terminal domain of GSDMD, which inserts into the plasma membrane to form GSDMD pores leading to pyroptosis [15,16,17]. The GSDMD oligomerization relies on mitochondria reactive oxygen species (mtROS) provoked by the Ragulator-Rag complex and its downstream mTORC1 [18]. In some cases, cleaved GSDMD can insert into mitochondrial membranes [19, 20]. The leakage of mtROS switches pyroptosis into necroptosis [19]. Through GSDMD pores, are mature IL-1β and IL-18 released into the extracellular environment [21, 22]. A good question is how the GSDMD pores distinguish mature IL-1β and IL-18 from their precursors. A recently published cryo-electron microscopy analysis shows a predominantly negatively charged conduit of the GSDMD pore that favors the passage of mature IL-1β and IL-18 and sequestrates negatively charged IL-1 precursors containing acidic domain [159, 203, 246]. These agents inhibit NLRP3 in different mechanisms, some of which remain elusive.

The most commonly used NLRP3 inhibitor in preclinical experiments is MCC950 (CRID3), which inhibits NLRP3 with nM potency without interfering with other inflammasome sensors [246]. Mechanistically, MCC950 interacts with Walker B motif of NACHT domain that is close to the ATP binding pocket, thereby blocking the hydrolysis of ATP and suppressing NLRP3 activation [247]. This specific blockage is consistent no matter whether in wild-type or mutated NLRP3 [248]. Another structural research has illustrated that the sulfonylurea group of MCC950 interacts with the Walker A motif of NLRP3 and it is sandwiched between Arg351 and Arg578 resulting in stabilized NACHT and LLR domains relative to each other [63]. MCC950 is initially developed as a potential therapeutic agent for CAPS, as well as other autoinflammatory and autoimmune diseases [246]. Later research works illustrate the potential anti-tumor effect of MCC950 against pancreatic cancer and head and neck squamous cell carcinoma [159, 203]. Similar to MCC950, the target of CY-09 is Walker A motif of NACHT, which binds ATP [249]. Another inhibitor that targets NACHT domain is tranilast [250]. However, it also suppresses TGF-β, MAPK, and NF-κB signals [251]. Present results have demonstrated that tranilast inhibits malignant behaviors of NSCLC and gastric cancer [252, 253], but the authors do not clarify whether these effects are related to inhibited inflammasome signal. Other inhibitors targeting NACHT domain include 3,4-methylenedioxy-β-nitrostyrene (MNS) [254] and oridonin [255]. MNS also binds to LRR domain [254]. Interestingly, isoliquiritigenin and glycyrrhizin are able to inhibit NLRP3 through both signal 1 (TLR4) and signal 2 (NLRP3) [256], but their inhibitory potency is not as powerful as that of MCC950. There is also an indirect NLRP3 inhibitor, ibrutinib [257]. Ibrutinib inhibits the generation of phosphorylated Bruton tyrosine kinase (BTK) that directly interacts with NLRP3 and ASC leading to the formation of inflammasomes [257]. Another indirect NLRP3 modulator is resveratrol that suppresses the expression of NLRP3 in renal cancer cells [258]. In a word, present findings imply that NACHT domain is the key target for inhibitors.

Although NLRP3 inhibitors alone have shown anti-tumor effects, some attempts highlighted the potential combination of NLRP3 inhibitors with other therapeutic methods. OLT1177 disrupts IL-1β/IL-6/STAT3 axis in TME resulting in reduced tumor growth through attenuating immunosuppressive activities in MDSCs [202], and the anti-tumor effect is further enhanced in combination with anti-PD-1 [245]. In addition to enhancing therapeutic effects, NLRP3 inhibitors may protect against the side effects of chemotherapy and radiotherapy. Resveratrol reduces doxorubicin-induced cardiac injury and systemic inflammation through downregulating NLRP3 inflammasomes [259]. Likewise, bowel inflammation after irradiation is also suppressed by resveratrol through a similar mechanism [228]. More research works are needed to explore other possibilities of such a combination.

It is suppressing that although many kinds of NLRP3 inhibitors have been invented, only a few of these agents have entered clinical trials for tumor therapy. ACT001 combined with anti-PD-1 or ACT001 alone has been applied for a phase I/II trial against glioblastoma. This agent is primarily developed for Parkinson’s disease [260]. Others are some agents that have been reported to inhibit NLRP3, including glycyrrhizin and andrographolides. Whether their anti-tumor effects are underpinned by NLRP3 inhibition remains to be further evaluated.

AIM2 inhibitors

Compared with NLRP3, limited AIM2 inhibitors have been found. Two agents are able to inhibit AIM2, but they are not specific inhibitors. Glycyrrhizin suppresses both AIM2 and NLRP3 [256]. Methylene blue is a broad-spectrum inflammasome inhibitor against NLRP3, NLRC4, AIM2, and non-canonical inflammasomes [261]. Fortunately, andrographolide has shown a promising effect for the future clinical application that it reduces radiation-induced lung inflammation and fibrosis by preventing AIM2 from entering the nucleus and sensing DNA damage [224]. Although potential anti-tumor effects of glycyrrhizin, andrographolide, and methylene blue would be evaluated in colon cancer, breast cancer, and liver cancer during clinical trials, specific AIM2 inhibitors are in need.

NLRC4 inhibitors

Few specific NLRC4 inhibitors are available. Instead, two inflammasome inhibitors with limited selectivity have been reported to suppress NLRC4. Sulforaphane attenuates the activation of both NLRC4 and NLRP3 at μM potency, which limits inflammation during peritonitis [262]. Methylene blue, a broad-spectrum inflammasome inhibitor, blocks NLRC4, NLRP3, AIM2, and non-canonical inflammasomes, which improves the survival rate of mice challenged with LPS [261]. Considering that many NLRP3 inhibitors target NACHT and that NACHT domain also exists in the NLRC4 inflammasomes, future selective NLRC4 inhibitors might be the derivate of the NLRP3 inhibitors.

Caspase-1 inhibitors

Owing to the upsurge of the study in caspase-related signals, many inhibitors targeting caspase family members have been developed, some of which are able to inhibit caspase-1. As the inhibitors of the common downstream protein of inflammasomes, caspase-1 inhibitors restrain not only NLRP3-derived but also AIM2-derived inflammasome signals. For example, VX-765 inhibits NLRP3/caspase-1/GSDMD-induced pyroptosis in NSCLC [263], and it also attenuates AIM2-mediated cell migration in NSCLC [264]. An interesting question is whether caspase-1 inhibitors promote or suppress cancer cell growth. Direct inhibition of pyroptosis has been reported in NSCLC by VX-765 [263], liver cancer by Ac-YVAD-CMK [265], and prostate cancer by Z-YVAD-fmk [266]. On the contrary, the caspase-1 inhibitor, thalidomide, impedes tumor growth in melanoma by suppressing caspase-1 in MDSCs [267]. Thus, non-selective administration of caspase-1 inhibitors may promote tumor growth, while selective caspase-1 inhibition in MDSCs may attenuate tumor development. Of note, in the early years when inflammasome signal was not intensively studied, caspase-1 mediated cell death was regarded as apoptosis [268, 269]. These findings should be updated to clarify the kind of cell death. Additionally, the relationship between pyroptosis of cancer cells and tumor growth should be further studied. Because DAMPs from dead cancer cells may elicit inflammasomes in adjacent myeloid cells and probably cancer cells resulting in the recruitment of MDSCs that facilitate tumor growth. At present, thalidomide alone or plus other agents have entered clinical evaluation against multiple myeloma, prostate cancer, and other advanced cancer.

ASC inhibitors

Although MCC950 has been demonstrated to selectively block the NACHT domain of NLRP3, it is also able to downregulate protein expression of ASC, caspase-1, IL-1β, and IL-18 [270]. This is an in vivo study that tests protein expression in tissues. Thus, it is possible that MCC950 directly inhibits NLRP3-mediated pyroptosis and IL-1β and IL-18 secretion causing reduced infiltration of macrophages [270]. Decreased number of macrophages in tissues may explain the downregulated protein levels of the inflammasome components. However, another research has found that MCC950 inhibits both NLRP3 and AIM2-derived inflammasome formation [271]. The MCC950-mediated ASC suppression is possibly through Glutathione S-Transferase Omega 1 (GSTO1), a putative target of MCC950 [271]. In a word, there is a lack of selective and direct ASC inhibitors.

GSDMD inhibitors

GSDMD pores are the direct cause of pyroptosis and the exit for intracellular mature IL-1β and IL-18. In addition to blocking pyroptosis and secretion of pro-inflammatory cytokines, two GSDMD inhibitors, LDC7559 [272] and Disulfiram [273], also restrain inflammation through curbing NETosis, a special kind of cell death of neutrophils. Although GSDMD inhibitors, disulfiram and Bay 11–7082, potently suppress pyroptosis [274], they show anti-tumor effects through inducing ferroptosis [275] (by disulfiram) or apoptosis [276, 277] (by Bay 11–7082). In a word, the anti-inflammation effects of GSDMD inhibitors have been repeatedly proven, but their applications in tumor development remain to be further evaluated.

IL-1 signal inhibitors

Four targets of the IL-1 signal have been developed, including IL-1 receptor, IL-1α, IL-1β, and IL-18 that can be intervened by antagonists, antibodies, and binding proteins. These potent anti-inflammatory inhibitors are pleiotropic agents applied in various inflammation-related diseases, for example, rheumatoid arthritis [278] (by canakinumab), autoimmune disorders [279, 280] (by rilonacept), and cardiac remodeling [281] (by gevokizumab). Blocking IL-1 signals might promote or inhibit tumor development. IL-18BP, a binding protein targeting IL-18, limits anti-tumor immunity [282]. However, anakinra, an IL-1 receptor antagonist, reduces IL-1β and downstream production of cancer-promoting IL-22 [283]. Similarly, anti-inflammatory therapy in patients with atherosclerosis by canakinumab reduces lung cancer incidence [284]. This effect has been proven to be underpinned by the reduced tumor-promoting inflammation [285]. For tumor therapies, anakinra gives rise to cytotoxic/NK cell transcriptional pathways and hampers innate inflammation in breast cancer patients receiving chemotherapy [286]. Additionally, anakinra is reported to limit the mucosal barrier injury and the accompanying clinical symptoms induced by melphalan [287]. Although more frequent fatal infections and sepsis are recorded in the canakinumab treatment group, all-cause mortality does not differ significantly between the placebo and the canakinumab group [284]. On the contrary, the anakinra treatment seems to be more safety [286], and no adverse events or dose-limiting toxicities have been observed [287]. In another phase 2 clinical trial, anakinra is applied in patients receiving 5-FU plus bevacizumab therapy, and no grade 4/5 toxicity related to therapy occurs during the study [288]. Interestingly, anakinra abrogates cytokine release syndrome during CAR-T therapy implying its compelling clinical application [289]. A series of clinical trials testing the prevention of CAR-T cell-mediated toxicity by anakinra have been launched, such as NCT04432506, NCT04150913, and NCT04148430. At present, most interventions targeting inflammasome pathways for cancer therapies listed in Table 3 are based on IL-1 signal inhibitors, possibly owing to the ready-made agents for other non-malignant diseases. For example, the therapeutic effects of canakinumab in lung cancer, colon cancer, breast cancer, pancreatic cancer, renal cancer, and leukemia would be evaluated in a number of clinical trials. It is a compelling topic to test whether the combination of these IL-1 signal inhibitors with other therapies can be beneficial for patients.

Inflammasome activators

Although many results support that activated inflammasomes show anti-tumor effects directly through inducing pyroptosis and indirectly through stimulating immune cells, limited inflammasome activators are developed at present. Polyphyllin VI induces pyroptosis by activating NLRP3 in NSCLC cells [263]. Similarly, 17β-estradiol provokes pyroptosis via NLRP3 in liver cancer cells [265, 290]. Another NLRP3 activator, BMS-986299, shows potential anticancer effects, but the details are largely unknown [3]. BMS-986299 have entered a phase I trial to explore its safety and effectiveness in patients with solid tumor or advanced tumor. An alternative strategy is to supply the downstream IL-1 cytokines directly. Since IL-18 is likely to be beneficial for anti-tumor immunity [182, 282], recombinant IL-18 has been applied in several clinical trials such as NCT00659178, NCT00107718, and NCT00500058. In the clinical trial NCT04684563, CAR-T cells targeting CD19 and expressing IL-18 are applied in patients with chronic lymphocytic leukemia or non-Hodgkin lymphoma. Present results indicate that more efforts should be paid to develop inflammasome activators. Considering that inflammasomes may initiate pyroptosis in tumor cells and that IL-1β and IL-18 have been shown to activate T cells and NK cells, inflammasome activators may improve the effects of immune checkpoint inhibitors.

Conclusions

In this review, we summarize the mechanisms that activate canonical and non-canonical inflammasome pathways. More importantly, we discuss the roles of canonical and non-canonical inflammasomes in tumorigenesis, tumor cell death, tumor metastasis, immune evasion, chemotherapy, and radiotherapy. Finally, we review the interventions targeting the inflammasome pathways in preclinical and clinical stages.

A good question is how the inflammasomes are activated in TME. Expression levels of inflammasome components have been compared between healthy and tumor tissues [117, 134, 138, 170, 183]. Mice deficient in certain components of the inflammasome pathway [138, 182, 183] or inflammasome inhibitors [196, 291, 292] have been applied to reveal the various influences of inflammasomes on tumor behaviors. Inflammasome activators (such as ATP, H2O2, monosodium urate, and Mycoplasma hyorhinis) have been used in vitro to confirm that inflammasomes can be activated in certain cell subsets [117, 172, 175, 184]. However, little is known about the direct activators of inflammasomes in TME during tumor progression. The activators may be bacteria [130], cell debris [117], ATP [206], PKR [199], other unknown factors in TME, or more complicated cross-talk between cells. Novel techniques such as single-cell sequencing may improve our understanding of the details during inflammasome activation.

The inflammasome signal seems to be a conserved pathway, which even exists in bacteria [293]. Although similar mechanisms have been identified in different species, discrepancies in NLR homologous genes and inflammasome pathways between humans and mice have been found [294]. For example, Francisella tularensis activates NLRP3 in humans instead of mice [295]. Thus, more detailed comparisons are needed to answer the question of to what extent can the findings from mouse models be extended to human patients.

It seems fuzzy that inflammasome signals have conflicting effects in different research works. A possible explanation is that inflammasomes can be activated at different extents, which may result in distinct inflammation responses [16, 42]. Future research works should compare the outcomes of different extents of inflammasome activation in various cell subsets in TME. Through this way, we can make accurate decisions about whether and how inflammasomes should be activated or inhibited.

GSDMD-mediated pyroptosis is involved in cancer cell death during chemotherapy and radiotherapy; however, secreted IL-1β may recruit immunosuppressive cell subsets and initiate inflammation-related side effects. Thus, the combination of IL-1R signal inhibitors and chemotherapy or radiotherapy may improve outcomes. On the other hand, NLRP3 in DCs [206] and AIM2 in macrophages [196] have been shown to facilitate anti-tumor immunity. The combination of NLRP3 or AIM2 activators and immune checkpoint inhibitors is a compelling strategy for immunotherapy.