Introduction

Pyroptosis is a programmed cell death process associated with inflammation. It is characterized by apoptosis and necrosis in morphology [1, 2] (Table 1). Cells that undergo pyroptosis will form pyroptotic bodies similar in size to apoptotic bodies, with pyknosis and chromatin damage [3]. Large numbers of pores formed on the cell membrane causes it to lose integrity. Eventually, the membrane lysis arises, and the intracellular content is released to induce inflammation. In apoptotic cells, eflux of potassium and chlorine resulting in cell contraction is often observed, while necrosis is a selective ion overload that causes water influx and then cells to swell and die [4]. When the cell undergoes pyroptosis, the nonselective pores of the GSDMD formed on the cell membrane lose the natural ion gradient, which may be due to a slight swelling of the cell caused by the intracellular nonionic penetrant driving water into the cytoplasm [3, 5], and there are reports revealing that this swelling can be blocked by extracellular osmoprotectant or glycine [6]. A recent report suggested that cell death and membrane lysis are uncoupled, inhibiting cell membrane lysis but not preventing cell death [7]. Compared to pyroptosis, necrosis is more like a process of cell explosion while pyroptosis undergoes cytoplasm flattening caused by plasma membrane leakage [3, 8].

Table 1 Comparison of pyroptosis, apoptosis, necrosis and NETosis
Full size table

Pyroptosis is first detected in macrophages and its related diseases [9], and then a large number of reports confirmed that pyroptosis may also occur to neutrophils [24, 25]. The research on neutrophil pyroptosis has attracted more and more attention in recent years, but the role of neutrophil pyroptosis in sepsis still did not cause enough attention. As we all know, sepsis is a life-threatening organ dysfunction caused by a dysregulated host response to infection [26]. It is an imbalance between the pro-inflammatory and anti-inflammatory mediators. However, early sepsis mortality is caused by an acute, deleterious pro-inflammatory response. Neutrophils are the most abundant natural immune cells in a human body that play paradoxical roles in the progression of sepsis. In the early stage of sepsis, neutrophils first arrive at the site of infection [27], secrete important cytokines and chemokines, and obliterate pathogenic microorganisms by phagocytosis, degranulation, and release of ROS and Nets [28, 29]. And in severe sepsis, the release of various enzymes, inflammatory mediators in heart, lung, kidney and other vital organs by a large number of activated neutrophils, but with chemotactic dysfunctions, leads to tissue cell damage, and ultimately to the development of multiple organs function failure [30]. Some evidences suggest that the pyroptosis of neutrophils is an important way for neutrophils to function during sepsis and that neutrophils continuously synthesize and secrete IL-1β and IL-18 during pyroptosis. Eventually, neutrophils swell slightly and develop a membrane lysis, producing cytoplasmic DAMP (damage-associated molecular patterns) [10] and releasing a large number of immunomodulatory cytokines such as IL-10, IL-13, chemokines such as IL-8, MIP-1α, and myeloperoxidase (MPO), cathepsin G and other granzymes [31]. After the neutrophil membrane lysis, intracellular pathogens form a pore-induced intracellar trap (PIT) though they have not been released directly extracellularly. Via the complement and the scavenger receptor to coordinate the innate immune response, PITs promote the recruitment of neutrophils to release ROS or secondary phagocytosis to kill pathogens [21, 22]. Under the combined effects of these cytokines and structures, neutrophil pyroptosis plays the role of inflammatory signal amplifiers in the recruitment of immune cells, There is evidence that neutrophil pyroptosis is beneficial to bacterial clearance during infection [21] and that neutrophils pyroptosis is even more essential for certain types of pathogen infections such as Salmonella infection [32]. However, because the neutrophils have been proved to be the main source of IL-1β in the infection [24, 25], excessive neutrophil pyroptosis is obviously harmful in the early hyperinflammatory state in sepsis. Numberous studies have confirmed that Caspase-1/11 knockout and IL-1β/IL-18 knockout can improve the survival rate in CLP mice model and septic shock mice model [33,34,35].However, another septic shock model shows that knocking out IL-1beta and IL-18 is useless [36]. These seemingly contradictory conclusions suggest that we should have different regulations on neutrophil pyroptosis in different stages of sepsis and different types of bacterial infection. Therefore, the study and summarization of the main mechanism of pyroptosis in sepsis and the potential value of neutrophil pyroptosis in sepsis is of great significance for the treatment of sepsis.

Neutrophils in sepsis

In the early phase of sepsis, neutrophils are essential for pathogen control. Studies on neutrophils obtained that most neutrophils normally undergo apoptosis within 24 h. Interestingly, due to the increased release of immature neutrophils and the delayed apoptosis of circulating neutrophils, a large number of circulating neutrophils of various degrees of maturation can be detected in sepsis patients. Investigation of sepsis patients and animal models of sepsis revealed disrupted neutrophil functions, including impaired neutrophil migration, impaired clearance of bacteria, reduced production of reactive oxygen species (ROS) and disordered release of cytokines. [37] Recent studies have shown that some subsets of neutrophils in sepsis patients can secrete a large amount of IL-10, which is an immunosuppressive cytokine that can suppress the proliferation of T lymphocyte [38]. Other evidences show complex interactions between the neutrophils and complement system, which cause complement-induced innate immune damage during sepsis. This review focuses on another important way in which neutrophils regulate immunity in sepsis: pyroptosis.

Pyroptosis is a programmed cell death process dependent on caspase-1 and -11, considered to be an immune response involved in the acute bacterial and viral infections, as well as the exposure of bacterial toxins and intracellular pathogens eradication [14] (Fig. 1). Caspases, a class of cysteine proteases widely expressed in mammals, are mainly associated with programmed cell death and inflammatory responses, some of which play a key role in antiviral immunity. Among them, Caspase-1, Caspase-4, Caspase-5 and Caspase-11 have proved to dominate the process of pyroptosis, and the role of Caspase-13 and Caspase-14 in pyroptosis is still controversial. Caspase-4, Caspase-5 and Caspase-11 are homologous proteins, Caspase-4 and Caspase-5 are expressed in mice and Caspase-11 is expressed in human [14, 39]. When pathogens exist, natural immune cells, including neutrophils, macrophages, dendritic cells and other immune participating cells such as epithelial cells, endothelial cells and fibroblasts, detect microorganisms by identifying, via pattern recognition receptors (PRRs), highly conserved pathogen-associated molecular patterns (PAMPs), including flagellin, peptidoglycan (PGN) and lipopolysaccharide (LPS). Other evidences reveal that PRRs are also responsible for identifying endogenous molecules released from damaged cells, called danger-associated molecular patterns (DAMPs) [40]. The activation of innate immunity depends on the recognition of both PAMPs and DAMPs, which is beneficial to the immune system’s tolerance to commensal probiotics and non-pathogenic bacteria [41]. At the same time, there is a growing number of evidence that DAMPs is essentially cytotoxic and not only depends on the involvement of inflammasomes in this process [42]. The PRR family consists of three categories: TLRs, NLRs and RLRs. Some NLRs can form a complex with specific functions called inflammasome, including NLRP1, NLRP3, NLRC4 and AIM2 involved in signal recognition during pyroptosis. NLRP1, NLRP3 and NLRC4 mRNAs were expressed in both human and mouse neutrophils at similar or greater levels than other cell types whose function is well studied, such as marrow-derived macrophages (BMDMs) or bone marrow-derived dendritic cells (BMDCs) [32]. The NLR family is characterized by the presence of a central nucleotide-binding and oligomerization domain (NACHT), the repeat domain of caspase recruitment (LRRs) and N-terminal (CARD) or PYRIN domain (PYD) [43,44,45]. GSDMD (gasdermin D), a member of the gasdermin protein family, is essential for pyroptosis in humans and mice. The structure of GSDMD has both the N terminal domain and the C terminal domain. Studies have shown that its N terminal domain plays a major role in inducing pyroptosis. The first loop on GSDMD-C that inserts into the N-terminal domain (GSDMD-N) helps stabilize the conformation of the full-length GSDMD [46], whereas the over expressed C end blocks GSDMD-N dependent pyroptosis [5, 47, 48].

Fig. 1
figure 1

In sepsis, there exist the neutrophil recognition of flagellin, exotoxins, other PAMPs and DAMPs by caspase-1 and the recognition of cytosolic LPS by caspase 4/5/11. In this process, NLRC4 is responsible for the detection of bacterial flagellin; NLRP1 is responsible for the detection of deadly toxins, and NLRP3 is responsible for the detection of other PAMPs and DAMPs as well as the cytoplasmic low potassium induced by ATP-mediated P2X7 signaling. Activation of caspase-1 and production of mature IL-1, beta and IL-18 are carried out in ASC focus. The pro-caspase-1 in the form of the dimer is cleaved into P10 and P20 subunits to form catalytically active caspase-1, and LPS triggers Caspase-11 oligomerization and activates its proteolytic activity. Caspases of the two active states mentioned above cleave the GSDMD and generate the oligomerization to form pores on the cell membrane, eventually leading to membrane lysis and pyroptosis. Other reports suggest that Caspase-11 induces pyroptosis via ATP-mediated P2X7 signaling, resulting in cytoplasmic low K+ activation of NLRP3 inflammasome and Caspase-1 induced IL-1 beta and IL-18 maturation and release. There are probably five ways of extracellular LPS entering cytoplasm: (a) through an unknown endocytosis. (b) Gram-negative bacteria in the intracellular vacuoles release LPS directly into the host cytoplasm by IFN-γ induced GTPase action. (c) Gram-negative bacteria secrete outer membrane vesicles (OMV) wrapped LPS that enters the cytoplasm via endocytosis. (d) Extracellular LPS enters the cytoplasm via the carrier of CTB. (e) HMGB1 destabilizes phagolysosomes for the transfer of LPS to cytosolic caspase-11. In neutrophils and other cells, three possible different ways have been found for mature IL-1β and IL-18 to release extracellularly: (a) mature IL-1β and IL-18 release via extracellularly through the membrane pores. (b) Monocyte encapsulates the activated caspase-1 and cytokine substrates into lysosomes and releases the processed cytokines via the lysosome to the cell surface. (c) Dendritic cells, microglia and macrophages can release the cytokine-containing vesicles after ATP stimulates Caspase-1. There are two different ways of release here: the release of cell surface fusions of multivesicular bodies and the release of microbubbles directly from the cell membrane

Full size image

Canonical inflammasome induced pyroptosis

In bacterial infection, pattern recognition receptors (NLRP1b, NLRP3, NLRC4) recognize microbial PAMPs, and then, via ASC (apoptosis-associated speck-like protein adaptor protein containing a CARD) indirectly connected with procaspase-1, form a Caspase-1-dependent inflammasome. ASC docks onto the inflammasome hub via pyrin–pyrin interactions and then recruits Caspase-1 via CARD–CARD interactions. However, ASC can also interact with itself, recruiting additional ASC molecules via pyrin–pyrin and CARD–CARD interactions, where all of the ASC within a cell is recruited via the cascade effect to a single subcellular location, which has been called the ‘ASC focus’ or ‘ASC speck’. In the ASC focus, the pro-caspase-1 in the form of the dimer is cleaved into P10 and P20 subunits to form catalytically active caspase-1, where pro-IL-1β is activated into mature IL-1β [15]. In addition, NLRP1 and NLRC4 can also directly connect and activate Caspase-1 without depending on ASC (Fig. 2). Different inflammasomes detect different intracellular contamination and perturbation: (a) Gram negative bacteria such as S. typhimurium [32], S. flexneri [49] and P. aeruginosa [50] or L. pneumophila [51] and Brucella [52] can transport, respectively, through Type III secretion system or through Type IV secretion system, the flagellin into the neutrophil cytoplasm, which are to be recognized by NLRC4 inflammasome. (b) Exotoxins secreted by Gram-positive bacteria and some Gram-negative bacteria can enter the cytoplasm by ionophore, pore former [53], protease [54] and other mechanisms, which are to be recognized by NLRP1 or NLRP3 inflammasome. (c) Other PAMPs, DAMPs and cytosolic low K+ can be recognized by NLRP3 inflammasome. These inflammasomes recruit pro-caspase -1 via ASC and activate Caspase-1 ultimately [11, 15, 55]. The activated Caspase-1 cleaves the connecting part between N-terminal and C-terminal of GSDMD rapidly, removes the inhibitory effect of C-terminal on the N-terminal and releases the GSDMD N-terminal to connect with the phosphoinositide of cell membrane and generate the oligomerization. Then the formation of pores disrupts the osmotic balance of the cell membrane, leading to cell swelling and membrane lysis, and releasing large amounts of cell contents causing inflammation [5, 47, 48]. Interestingly, activation of inflammasomes does not imply that the cells are destined to lyse and that certain toxins, such as melittin, can prevent cell lysis but induce NLRP3 inflammasome activation and IL-1β release [56]. In some cases, such as acute Salmonella infection, NLRC4 inflammasome sustained activation of neutrophils to release IL-1β without pyroptosis [32]. There are still controversies in previous reports on the pore diameters formed on the membrane of GSDMD, such as 15 nm inner and 32 nm outer [47], 12–14 nm [5], 13 nm [57], 20 nm [58] and 1.1–2.4 nm [6], respectively. Caspase-1 activated GSDMD can promote proIL-1β and proIL-18 maturation, and the continuous production of active IL-1β and IL-18 [11, 15], which are to be released extracellularly through the membrane pores formed by GSDMD or other ways [59, 60]. In addition, studies have shown that activated GSDMD not only induces the host cell membrane to form pores, but also form pores on the surface of the bacteria infecting the host to kill the bacteria. The GSDMD N-terminal released after lysis can also kill bacteria directly outside the host, including Escherichia coli, Staphylococcus aureus and Listeria monocytogenes [47].

Fig. 2
figure 2

(1) NLRC4 and mouse NLRP1b contain the CARD domain that binds directly to the CARD domain of Caspase-1 to trigger Caspase-1. (2) The ASC contains the Pyrin signal domain and the CARD domain, NLRP3 that contains the Pyrin signal domain can be combined with ASC’s Pyrin signal structure and the CARD domain of ASC can connect with the CARD domain of Caspase-1

Full size image

Non-canonical inflammasome induced pyroptosis

Lipopolysaccharide is a major component of the cell wall of gram-negative bacteria. It can not only induce inflammation [61, 68] and autophagy [62], but can also induce cell pyroptosis [13, 63,64,65]. Under the stimulation of intracellular LPS, Caspase-11 can specifically bind to the lipid A of LPS, which triggers Caspase-11 oligomerization, activates its proteolytic activity and cleaves the GSDMD to form a large number of pores on the cell membrane, eventually leading to membrane lysis and pyroptosis [13, 63, 64]. Several reports also reveal that the binding of intracellular LPS to Caspase-11 triggers Caspase-11 oligomerization, and the activated Caspase-11 can cleave Pannexin-1 to make it open [65]. Since intracellular ATP concentration is one million times extracellular ATP concentration [66], the intracellular ATP released through the Pannexin-1 channel to the extracellular activation of P2X7 receptor leads to P2X7 channel opening, rapidly triggers cytoskeleton destruction and membrane PS flip, blebbing, microvesicle shedding and eventually leads to the destruction and lysis of the cell membrane [67]. P2X7 receptor activation mediates membrane blebbing in at least two different ways: calcium-dependent mechanisms and calcium-independent RhoA and ROCK-1 mechanisms [68]. In the pannexin-1 channel to open the release of ATP, under the concentration gradient of K+ inside and outside, K+ is released in large quantities extracellularly, resulting in cytoplasmic low K+ activation of NLRP3 inflammasome and inducing pyroptosis via the ASC activation of Caspase-1 to release mature IL-1β and IL-18 [69,70,71]. Other reports reveal that not only can the Pannexin-1 channel release K+ to the extracellular matrix, the P2X7 receptor also has this ability, and the latter plays a major role in the process of intracellular low potassium activation of NLRP3 [71, 72]. Previous reports indicate that the expression and function of P2X7 receptors on human and mice neutrophils are contradictory [70, 73] or P2X7 is present in the cytoplasm rather than in the human neutrophil surface [74]. However, recent reports have demonstrated that P2X7 receptors are widely expressed on human and mice neutrophils and exhibit a certain racial differences [70, 71].

The source of cytosolic LPS in pyroptosis

In severe sepsis, a great deal of LPS is released into the blood circulation, activating a large number of neutrophils via TLR2 and TLR4 receptor. At the same time, LPS inhibits neutrophil chemotaxis through autocrine ATP signaling pathway [75], resulting in extensive damage to tissues and organs [30]. At present, the source of cytosolic LPS is still not entirely clear, and obviously the great deal of extracellular LPS does not directly penetrate through the membrane into the cytoplasm during sepsis. It may be that the extracellular LPS enters cytoplasm via a special transport way or through the channel generated during pyroptosis, or it may originate from the release of bacteria in the inner niche under specific conditions. Neutrophils are a key member of the immune system with a strong phagocytic capacity, so neutrophils in the sepsis can swallow the pathogen to form phagocytosis to kill the pathogen, where LPS cannot be released into the cytoplasm [76]. Many microbes have been found to survive in neutrophils, such as Salmonella [77], Neisseria gonorrhoeae [59, 78], S. aureus [79], Chlamydia pneumoniae [80], Burkholderia pseudomallei [60], Anaplasma phagocytophilum [81] and L. monocytogenes [82]. Another report suggests that the survival of Gram-negative bacteria in the intracellular vacuoles can directly release LPS into the host cytoplasm, which requires IFN-γ to induce GTPase action to cause bacteria to escape from the vacuole or to break the vacuole [83]. Several reports suggest that ATP, which is abundant in inflammation sites, activates the P2X7-dependent Pannexin-1 half channel pore formation and may allow extracellular LPS to promote the LPS entry into the cytoplasm via either Pannexin-1 channels or Pannexin-1 mediated indirect effects [44, 84]. P2X7 is a selective cation channel in the presence of ATP stimulation, and when the divalent cation level is low, the cationic channel can be converted into small pores and ions that can penetrate up to 900 Da [73]., Because the LPS molecular weight is about 5–15 kD, and Pannexin-1 channel can penetrate about 1kD [85], it is obvious that LPS cannot directly penetrate through the Pannexin-1 or P2X7 channel. According to a previously reported LPS electroporation model, the extracellular LPS that gets free access to the cytoplasm via the membrane pore can cause the vast majority of cells to undergo pyroptosis [86]. As stated above, the extracellular LPS stimulation of neutrophils can also activate TLR4-P38-Cx43 pathway to autocrine ATP extracellularly [75]. Therefore, the existence and efficiency of this mechanism where the LPS directly or indirectly enters the cytoplasm via the Pannexin-1 channel remains to be further studied and discussed.

The other two distinct extracellular LPS transport mechanisms require a specific vector. The first is cholera toxin B (CTB), which can be used as a carrier binding LPS and transported to cytoplasmic induction of non-canonical inflammasome pyroptosis [87,88,89], and this combination is required for LPS type. There are data showing that only LPS O111:B4 can be combined with CTB to induce non-canonical inflammasome pyroptosis, but LPS O55: B5 and LPS O127: B8 have no such ability [87]. Another mechanism reported in recent years is Gram-negative bacteria secrete outer membrane vesicles (OMV) wrapped LPS that enter the cytoplasm via endocytosis. OMVs are vesicles between 20 and 250 nm produced in a programmatic way; they are not by-products of bacterial cell wall damage or bacterial dissolution. Bacteria that poorly produce OMVs elicit low-level pyroptosis and IL-1 maturation [63]. OMVs can carry LPS, phospholipids, peptidoglycan, outer membrane proteins (OMPs), cell wall components, proteins (periplasmic, cytoplasmic, and membrane-bound), ion metabolites and signaling molecules [90], and many bacterial OMVs also carry nucleic acids (DNA, RNA) [91]. Following the clathrin-mediated endocytic uptake, OMVs transmit LPS into the cytosol from early endocytic compartments and eventually activate Caspase-11 to trigger pyroptosis [63]. Early endosomal escape allows OMV-bound LPS to reach cytosol functionally intact and avoid complete degradation in the lysosomes. There is an evidence that OMVs can strongly stimulate neutrophils to secrete IL-1β.Therefore, the neutrophil pyroptosis caused by cytosolic LPS is very likely transport by OMVS too. A recent report confirms that HMGB1 destabilizes phagolysosomes for the transfer of LPS to cytosolic caspase-11 [92]. The author’s ample evidence in the article seems to have brought us closer to solving the mystery.

Roles and release mechanisms of IL-1β and IL-18 in pyroptosis

The interleukin-1 (IL-1) cytokine family, an important regulator of innate immunity and adaptive immunity, plays an important role in the host’s defense against infection and inflammatory injury. The family includes IL-1, IL-18 and IL-33 and the recently found IL-36 and IL-37, in which both IL-1β and IL-18 mediate the inflammatory response [93, 94]. IL-1β and IL-18 have been shown to be the most important cytokines during pyroptosis [69, 95]. IL-1β is usually produced by tissue macrophages, DC cells, blood mononuclear cells, neutrophils, B lymphocytes and NK cells, but generally not by fibroblasts and epithelial cells [94]. In the early stage of sepsis, the resident macrophages in the site of infection secrete IL-1β to recruit neutrophils. Then the partial neutrophils to the site of infection undergo pyroptosis and become the main source of IL-1 beta secretion, mediated by the positive signal amplification circuit to recruit more neutrophils to the infection sites [32]. In the meantime, one report showed that neutrophils were the main source of IL-1β secretion in bone marrow during infection [24] and another reported that neutrophils were the main source of IL-1β and IL-18 in BALF during infection [25].Therefore, we can assume that neutrophil pyroptosis in sepsis controls the level of IL-1β and IL-18. Another important cytokine, IL-18, can be detected in many cells including Kuppfer cells, monocytes, dendritic cells (DCs), macrophages, keratinocytes, chondrocytes, intestinal epithelial cells, Synovial fibroblasts and osteoblasts [96,97,98]. In a case of sepsis, LPS first stimulates the production of pro-IL-1β and Pro-IL-18 by TLR or RLR. Then, NLR mediates inflammasome activation and facilitates post-translational processing, which is necessary for its secretion and biological activity [99]. IL-1β is one of the most potent proinflammatory cytokines known at present. It has an obviously protective effect in acute infection, including rapid recruitment of neutrophils to inflammatory sites, activation of endothelial adhesion molecules, induction of cytokines and chemokines, induced febrile reaction and stimulating specific immune responses such as Th17 response [99]. IL-1β has a protective effect in several bacterial, viral and fungal infection models, and the use of IL-1 receptor antagonist, IL-1R, to antagonize IL-1 increases the susceptibility to bacteria. The most special function of IL-18 is to promote Th1 cells, NK cells and cytotoxic T lymphocytes to produce interferon-γ(IFN-γ) and promote CD8+ T cells and NK cell proliferation. In addition, IL-18 can also stimulate other inflammatory cells to secrete cytokines such as tumor necrosis factor α (TNF-α), IL-1β, IL-8 and GM-CSF [100].

The mechanism of the secretion of IL-1β and IL-18 is also controversial in pyroptosis. In addition to the release of IL-1β and IL-18 via extracellularly through the membrane pores, there are two more possible release mechanisms that have been found in neutrophils and other cells: (a) monocyte encapsulates the activated caspase-1 and cytokine substrates into lysosomes and releases the processed cytokines via the lysosome to the cell surface [101]. Although this mechanism of secretion can avoid pyroptosis and continue to secrete cytokines, it has been considered as not dominant [102]. (b) Dendritic cells, microglia and macrophages can release the cytokine-containing vesicles after ATP stimulates Caspase-1 [103, 104]. There are two different ways of release here: the release of cell surface fusions of multivesicular bodies and the release of microbubbles directly from the cell membrane [104,105,106]. The specific mechanism for the release of IL-1β and IL-18 in the case of neutrophil pyroptosis is still controversial because of the short life of neutrophils and they cannot be proliferate, it is difficult to study the release of granules and cytokines. Direct release of IL-1β and IL-18 after cell membrane dissolution is present. However, the continuous maturation of IL-1β and mature IL-18 release require the continuous activation of Caspase-1, so this way is obviously not the major one. A recent article that attracted wide attention showed that the macrophage IL-1β was released actively through the holes formed by GSDMD [107], but it has not been proved directly in neutrophils and require further study.

Pore-induced intracellular traps triggered by pyroptosis

Caspase-4 and Caspase-4/5/11 trigger the pathway of pyroptosis, eventually causing cell membrane lysis, that is the release of intracellular contents and cytokine, which is an important natural immune response. Recent studies have shown that natural immune cells such as macrophages and neutrophils have shown other specific ways to function in pyroptosis: the formation of PITs, rather than the mere release of intracellular pathogens and contents to the extracellular environment. These reports reveal that organelles and cytoskeleton do not have mere separation and diffusion after the membrane lysis, but remain in the corpse of the cells to form an intracellular trap-like structure to restrict surviving pathogens, which is a structure called pore-induced intracellular traps(PITs) [21, 22]. Then, the pathogens trapped in PITs were subsequently killed by neutrophil ROS or by secondary phagocytosis of other neutrophils [10]. PITs are conceptually similar to Nets, both by preventing bacteria from spreading and promoting clearance. The difference between the two is that Nets are directed against extracellular bacteria, while PITs are targeted to intracellular niche. Nets and PITs are independent of each other. In Nets-deficient mice (MPO −/− and Elane −/−) the formation of PITs was not affected [21]. Nets are essentially bactericidal because they contain antibacterial peptides and enzymes and produce ROS [108,109,110,111,112], while PITs damage, but do not kill bacteria. After the formation of PITs, the killing or secondary phagocytosis of neutrophils ROS apparently poses a question: how does neutrophil discover PITs?

Apoptosis or necrotic cells will release the “Find-me” signal, expose the “Eat-me” signals and lose “Don’t eat me” signals, which combine to promote the clearance of apoptotic and necrotic cells [23, 113]. When PIT is formed, “Find-me” signals, such as microbial surface C5a or chemokines and LPC released from pyroptosis cells, are first released. Then the various “Eat-me signals” such as PS and DNA are rapidly exposed, while several “Don’t eat me” signals such as CD47, CD31 are lost or modified. Then together with some extra elements of expression, they eventually reach a new balance between “Eat-me” signals and “Don’t eat me “signals, causing neutrophils to eventually reach the environs of PITs to undergo efferocytosis and destroy the pathogens entrapped inside by a secondary killing [114, 115]. After the membrane lysis of phagocytic cells, PITs are bound to form, and PITs are small enough to retain organelles and bacteria [21].The clearance of pyroptotic neutrophils is actually equal to the clearance of PITs and it seems to induce a pro-inflammatory reaction.

Conclusions and perspectives

Pyroptosis has been shown to be an important natural immune response. Although a variety of classical and non-classical inflammasome-induced pyroptosis have been described, there are still many controversies concerning the specific details of pyroptosis. With the research of pyroptosis deepening gradually, the possibility of pyroptosis applied to sepsis treatment is adequately amplified. Many studies have confirmed that Caspase-1/11 knockout and IL-1β/IL-18 knockout can improve the survival rate in CLP and septic shock mice model (Table 2), but clinical trials failed [116]. It seems to confirm that the inhibition of IL-1 and IL18 are harmful. However, due to the pro-inflammatory and anti-inflammatory responses in sepsis, the same treatment may lead to different outcomes for patients at different stages of sepsis. Furthermore, differences in sepsis-causing bacteria may also affect outcomes.

Table 2 Effects of pyroptosis-related molecule-deficient mice in sepsis
Full size table

On the basis of preclinical studies, some clinical trials have been carried out. Most of these clinical trials focused on IL-1β, the downstream of pyroptosis. However, attempts to improve sepsis survival by blocking IL-1β, although effective against animal models [36], failed in some previous human trials [121, 122]. Encouragingly, a recent III phase clinical trial supports the possibility that anakinra (recombinant interleukin-1 receptor antagonist) treatment provides a survival benefit in septic patients with features of Macrophage Activation Syndrome [123]. A prospective trial has been initiated to validate these findings (NCT03332225; clinicaltrials.gov). Another precise clinical study shows that recombinant IL-1 receptor antagonists are beneficial to septic patients with high plasma interleukin-1 receptor antagonist (IL-1RA) [124].Since IL-1β induces gene expression of both itself and IL-1RA, significantly increased IL-1RA may inhibit IL-1β, stopping the cycle of IL-1β-IL-1RA amplification. Therefore, both subgroups of sepsis patients above actually occur to the context of cytokine storm, in which IL-1β plays a major role. In addition, there is another report that yielded similar results, but the sample size of sepsis patients is too small, and the results are still debatable [125]. At present, there are still few studies on the upstream molecules of pyroptosis in sepsis. A published study reveals an association of NLRP3 and Caspase-1 mRNA levels with severity of sepsis caused by cytomegalovirus (CMV) [126]. Another study suggested that caspase-1 is a potential marker for predicting the development of sepsis after severe trauma [127]. These preclinical and clinical studies have revealed that the regulation of neutrophil pyroptosis may benefit in the treatment of sepsis. Given the important role of neutrophil pyroptosis in sepsis, the regulation of neutrophil pyroptosis may have more potential value.

Neutrophils are the most abundant immune cells in sepsis and also the main source of IL-1β and IL-18. Considering the short life of neutrophils and the characteristics of its inability to proliferate, targeting the neutrophil pyroptosis specifically, which is the upstream of IL-1 and IL-18 may be better to avoid the chain reaction caused by Caspase-1/11 and IL-1β/IL-18 blocking. There is evidence that in the Pao1 mice lung infection model, the pyroptosis level of neutrophils in bronchoalveolar lavage fluid is less than 2% in 24 h [25], but contributes most of IL-1β production. At the same time, due to the low level of neutrophils pyroptosis, the tissue damage caused by released cytosolic contents is far from the abnormal infiltration of neutrophils in tissues and organs. All these suggest us that the great potential for the regulation of neutrophil pyroptosis in sepsis. For some types of infections such as Salmonella infection, specific regulation of neutrophil pyroptosis may bring more benefits. On the other hand, because neutrophils are relatively short-lived and cannot proliferate, the regulation of neutrophil pyroptosis seems to be easier to avoid the unpredictable risks of sepsis treatment. Therefore, the study of neutrophil pyroptosis will be a highly valuable and promising treatment for sepsis.