Reprogramming of Cell Death Pathways by Bacterial Effectors as a Widespread Virulence Strategy

ABSTRACT

The modulation of programmed cell death (PCD) processes during bacterial infections is an evolving arms race between pathogens and their hosts. The initiation of apoptosis, necroptosis, and pyroptosis pathways are essential to immunity against many intracellular and extracellular bacteria. These cellular self-destructive mechanisms are used by the infected host to restrict and eliminate bacterial pathogens. Without a tight regulatory control, host cell death can become a double-edged sword. Inflammatory PCDs contribute to an effective immune response against pathogens, but unregulated inflammation aggravates the damage caused by bacterial infections. Thus, fine-tuning of these pathways is required to resolve infection while preserving the host immune homeostasis. In turn, bacterial pathogens have evolved secreted virulence factors or effector proteins that manipulate PCD pathways to promote infection. In this review, we discuss the importance of controlled cell death in immunity to bacterial infection. We also detail the mechanisms employed by type 3 secreted bacterial effectors to bypass these pathways and their importance in bacterial pathogenesis.

CELL DEATH AS AN ANTIMICROBIAL DEFENSE MECHANISM

The ability to program and control cell death is essential to many physiological processes in all living organisms. While cell death is part of organismal homeostasis during nonvirulent encounters (1, 2), its roles in infection are context and pathogen dependent. Cell death can be broadly divided into lytic and nonlytic forms, both of which participate in the innate response to infection. Apoptosis was the first identified programmed cell death (PCD) pathway (3). Immunologically silent, apoptosis of infected cells eliminates the replicative niche of the pathogen and does not lead to the release of proinflammatory mediators (4). In contrast, pyroptosis and necroptosis are lytic and proinflammatory PCDs, characterized by the release of cytokines (pyroptosis) or damage-associated molecular patterns (DAMPs; necroptosis) that recruit immune cells, such as neutrophils, to clear infiltrating pathogens and infected cells from the site of infection (5, 6). Uncontrolled, these inflammatory responses cause devastating consequences on tissue homeostasis (7). For example, the recruitment and influx of neutrophils at infection foci disrupt the integrity of epithelial barriers (8), furthering the access of bacteria to distally located tissues (9, 10).

Bacterial pathogens have evolved strategies that can inhibit and/or promote cell death during infection to support their pathogenesis. Gram-negative bacteria manipulate their hosts using specialized machineries called secretion systems that are categorized by types according to their function and composition (11). Type III, IV, and VI secretion systems (T3SS/T4SS/T6SS) are widely distributed across Gram-negative bacterial pathogens and inject virulence factors, also called effector proteins, into targeted host cells (12–16). The combined activities of effectors manipulate host cellular homeostasis to promote and sustain bacterial infection (17, 18). Broadly, effectors target host cellular pathways to manipulate cytoskeletal dynamics (to elicit bacterial invasion of nonprofessional phagocytes or to prevent phagocytosis), cell-intrinsic surveillance mechanisms that orchestrate inflammatory signaling, and the molecular drivers of cell death processes (19–21).

In this review, we discuss a subset of reported T3SS effectors which manipulate the three major host cell death pathways; apoptosis, necroptosis, and pyroptosis (Table 1). Further, we highlight literature pertaining to the importance of these virulence factors in mammalian infection models.

TABLE 1.

Inhibitors of programmed cell death pathways

Pathway(s) targeted	Pathogen	Virulence factor	Molecular target/mechanism	Function	Species defineda	Relevance in vivo	Reference(s)
Apoptosis/ necroptosis	EPEC/ EHEC/C. rodentium	NleH	BI-1	Inhibition of intrinsic apoptosis via activation of the negative regulator BI-1	Human (Ep)	Yes	72, 73
		NleF	Caspase 8 and 9	Inhibition of FasL-mediated apoptosis	Human (Ep), mouse (in vivo)	No	76, 77
		NleB1/2	TRADD, FADD, RIPK1	N-Acetylglucosamine modification of death receptors, inhibition of extrinsic apoptosis and necroptosis	Human (Ep), mouse (MEF)	Yes	77–82, 84
	S. enterica	SseK1 and SseK3	TRADD, FADD, RIPK1	N-Acetylglucosamine modification of death receptors, inhibition of extrinsic apoptosis and necroptosis	Mouse (Mφ), human (Ep)	Unclear	84, 89, 90, 181–183
		SopB	Unknown	Activation of Akt	Human (Ep)	Not tested	91
		AvrA	MAPK kinases	Inhibition of JNK signaling	Mouse (in vivo)	Yes	92
		IpgD	PtdIns(4,5)P2	Activation of Akt, inhibition of p53	Human (Ep)	Not tested	98, 101
		VirA	p53	Inhibition of p53-mediated proapoptosis signaling in response to calpain activation	Human (Ep)	Not tested	98
		OspC1	Unknown	Inhibition of apoptosis via blockade of caspase 8	Human (Ep)	Not tested	49
		OspD3	RIP1/RIP3	Inhibition of necroptosis caused by OspC1-mediated inhibition of caspase 8	Human (Ep)	Not tested	49
Pyroptosis	EPEC/ EHEC	NleA	NLRP3	Inhibition of inflammasome	Human (Ep, Mφ)	Not tested	131
		NleF	Caspase 4	Inhibition of noncanonical inflammasome	Human (Ep)	No effect on virulence	76, 132
	Yersinia spp.	YopE/H	Unknown	Inhibition of integrin-mediated NLRP3 activation	Mouse (Mφ), Human (Ep)	Not tested	167
		YopK	T3SS translocon	Blockade of recognition of the T3SS translocon by NLRP3	Human (Ep)	Yes	164–166, 184
		YopM	Pyrin, RSK, PRK	Inhibition of pyrin inflammasome, activated in response to RhoA targeting by YopT/E	Mouse (Mφ)	Yes	174, 175
	Shigella spp.	IpgD	PtdIns(4,5)P2	Blocks NLRP3 inflammasome via inhibition of ATP release	Human (Ep)	Not tested	146
		IpaH9.8	GBP1, -2, -4	Degradation of GBPs, inhibition of initiation of noncanonical inflammasome	Human (Ep), mouse (Mφ)	Yes	152 – 154
		OspC3	Caspase 4/11	Binding and ADP riboxanation of caspase 4/11; inhibition of noncanonical inflammasome	Human (Ep)	Yes	112, 155–157
		IpaH7.8	Gasdermin D	Ubiquitination of gasdermin D, causing its degradation; inhibition of inflammasome	Mouse (Mφ) and human (Ep, En)	Not tested	158

^aEp, epithelial cell; En, endothelial cell; MEF, mouse embryonic fibroblast; Mφ, macrophage.

APOPTOSIS AND NECROPTOSIS

Apoptosis occurs following detection of a proapoptotic signal, leading to death of the cell. The engulfment of the cell corpse by macrophages and its lysosomal degradation constitute a cellular process referred to as efferocytosis (22, 23). Proapoptotic signaling can be triggered by intrinsic and extrinsic stimuli, which activate distinct pathways relying on the activation and engagement of caspases (24) (Fig. 1). Caspases belong to a family of cysteine proteases conserved in higher eukarya: human genomes encode 12 caspases, while those of mice encode 10. These zymogens, inactive in their steady state and activated upon their cleavage, are categorized into initiator caspases (caspases 2, 9, 8, and 10 in humans) (25–28) and executioner caspases (caspases 3, 6, and 7) (29, 30), which regulate apoptosis, inflammatory caspases (caspases 1, 4, 5, and 12; discussed below), and caspase 14, which regulates cell differentiation.

FIG 1.

Initiation and execution of apoptosis and necroptosis and their modulation by bacterial virulence factors. Apoptosis is triggered by both intrinsic (left) and extrinsic (right) stimuli. Components of the cell death machinery specific to apoptosis are shown in light blue rectangles. Necroptosis is initiated predominantly from death receptors shared with extrinsic apoptosis pathway. Key components of necroptosis are depicted in dark yellow rectangles. The EHEC effector NleH inhibits intrinsic apoptosis by promoting the activity of Bax inhibitor 1 (BI-1). Shigella employs the effector VirA to activate calpains via the inhibition of calpastatin. Calpains degrade the proapoptotic master regulator p53. p53 activity is also negatively regulated by the kinase Akt. Salmonella SopB and Shigella IpgD promote the activity of Akt to modulate p53 activity. Necroptosis and apoptosis pathways converge at the level of caspase 8. Shigella OspC1 effector can inhibit caspase 8 via an unknown intermediate (displayed as a circled question mark). Caspase 8 inhibition induces necroptosis, which is subsequently blocked by Shigella OspD3 by targeting RIPK1 and -3. The EPEC/EHEC OspD3 homolog, EspL targets death domain-containing proteins to inhibit apoptosis and necroptosis. Finally, Salmonella SseK1/3 and EPEC/EHEC NleB1/2 also inhibit death receptor signaling, respectively.

Upon receipt of proapoptotic stimuli, immature forms of initiator caspases (procaspases) are processed into mature enzymes. Following their activation by initiator caspases, executioner caspases mediate the degradation of intracellular products, leading to the characteristic cell shrinkage, nuclear condensation, chromatin fragmentation, and formation of apoptotic bodies (31). Cellular stressors, including DNA damage, nutrient starvation, osmotic stress, and free radicals, promote intrinsic apoptosis. These stimuli lead to depolarization of the mitochondrial membrane, causing Ca²⁺ and cytochrome c release (32–38). Together, the apoptotic protease activating factor 1 (apaf-1) and the cytosolic cytochrome c assemble into the apoptosome, which activates procaspase 9 (39, 40). In turn, mature caspase 9 activates the executioner caspases 3 and 7 (Fig. 1).

Extrinsic apoptosis is engaged through the receptor-mediated activation of initiator caspases by ligands comprising tumor necrosis factor alpha (TNF-α), Fas ligand (FasL), and TNF-related apoptosis-inducing ligand (TRAIL). These extrinsic stimuli promote the assembly of receptor-mediated signaling platforms (41–43). This signaling leads to the recruitment and activation of initiator caspase 8 (44, 45). Operating at the interface of both extrinsic and intrinsic apoptosis, caspase 8 promotes mitochondrial depolarization and activation of the intrinsic pathway, while also directly activating caspase 3/7 (46).

Apoptosis is considered a less inflammatory PCD, integral to the cellular turnover supporting tissue homeostasis. Many pathogens have evolved mechanisms targeting apoptosis signaling to maintain their replicative niche during host infection. Recent host-pathogen studies exploiting genetically engineered mice and refined biochemical characterization have indicated that caspase 8 activities extend beyond apoptosis and intersect with those of necroptosis, an inflammatory PCD (47, 48). Indeed, the targeting of apoptosis pathways by virulence factors or by caspase pharmacological inhibitors has been shown to favor necroptosis. Thus, necroptosis is emerging in the field as a fail-safe mechanism to induce cell death, evidenced by its intertwined signaling pathway with apoptosis (47–51).

Necroptosis is activated by ligands common to extrinsic apoptosis, including TNF-α, FasL, and TRAIL (51–53). In addition, necroptosis is activated downstream of a few immune stimuli such as toll-like receptors (TLRs) or cytokines (54–57). Receptor engagement leads to the activation of receptor-interacting serine/threonine-protein kinase 1 (RIPK1). RIPK1 then recruits RIPK3, forming the necrosome complex (Fig. 1). The necrosome recruits and phosphorylates the executioner of necroptosis, the protein mixed-lineage kinase domain-like pseudokinase (MLKL), causing its oligomerization and localization to cell membranes, leading to the eventual disruption of plasma membrane integrity (58–60). In contrast to apoptosis, necroptosis is a lytic PCD and promotes inflammation.

Caspase 8 is increasingly demonstrated as a pivotal switch in the cellular fate, operating at the nexus of apoptosis and necroptosis (61). Following receipt of a death signal, caspase 8 activation will induce apoptosis but inhibit necroptosis via cleavage of RIPK1/3 (62). However, if caspase 8 activity is compromised by chemical or pathogen-mediated inhibition, necroptosis can be engaged (49, 50, 63–65).

This section discusses examples through which the key stages of apoptosis and necroptosis are modulated by bacterial virulence factors (Fig. 1). A comprehensive list of bacterial effectors interfering with PCD factors and processes can be found in Table 1.

Modulation of apoptosis/necroptosis by attaching-effacing pathogens.

The infection mechanisms deployed by the bacterial pathogens described below will inevitably cause cellular stresses. To promote the survival of their cellular niche, pathogens can counteract these deleterious effects with T3SS effectors. However, some T3SS effectors also exert cytotoxic activities. For example, apoptosis can be induced during host cell infection by enteropathogenic and enterohemorrhagic Escherichia coli (EPEC and EHEC, respectively) and by Citrobacter rodentium, a closely related murine pathogen (66, 67). These enteric bacteria belong to the attaching-effacing (A/E) family of extracellular gastrointestinal pathogens. These use their T3SS to attach intimately to intestinal epithelial cells and manipulate PCD pathways (Fig. 1; Table 1). For example, the T3SS effectors EspF and Map destabilize mitochondrial integrity, leading to membrane depolarization and subsequently apoptosis (68–71). However, many EPEC/EHEC effectors bear well-defined biochemical activities inhibiting apoptosis, and the induction and/or inhibition of PCD in response to EPEC/EHEC infection is likely cell type and context dependent.

For example, the EPEC/EHEC effector NleH, which shares homology with the effector OspG in Shigella spp., was demonstrated to potently inhibit intrinsic apoptosis (72, 73). While NleH exhibits a kinase activity, this catalytic role was demonstrated to be dispensable for apoptosis inhibition. The identification of host targets phosphorylated by this effector would increase our understanding of its role in apoptosis signaling. However, NleH was shown to bind and activate Bax inhibitor 1 (BI-1), a negative regulator of intrinsic apoptosis (74). This observation is further supported during mouse infection, wherein the C. rodentium nleH mutant caused enhanced caspase 3 cleavage in the murine gut (73). Importantly, the role of NleH in apoptosis regulation remains contentious, in part due to its additional function in the regulation of NF-κB activation (75). Further biochemical studies are required to dissect this protein’s direct role in the regulation of the apoptotic signaling cascade.

Interestingly, EPEC/EHEC also use several effectors whose activities can inhibit extrinsic apoptosis and necroptosis, as both pathways share an initiating death receptor signaling axis, triggered by exogenous stimuli such as TNF-α, FasL, and TRAIL. Binding of these ligands to their cognate receptors, TNF-α receptor (TNFR), Fas, and TRAIL receptor, instigates the oligomerization of a multiprotein complex comprising the Fas-associated protein with death domain (FADD). TNFR also promotes the engagement of tumor necrosis factor receptor type 1-associated DEATH domain protein (TRADD). In extrinsic apoptosis, FADD recruits and activates procaspase 8, which in turn activates caspases 3 and 7. In the absence of caspase 8 activity, the same signals trigger a distinct pathway that involves RIPK1, RIPK3, and MLKL to induce necroptosis (Fig. 1) (49, 50, 63–65).

A/E pathogens inhibit multiple components of the above-mentioned signaling pathways using effectors with distinct biochemical properties. For example, NleF is a caspase inhibitor that blocks FasL-mediated apoptosis by directly inhibiting the activities of caspases 8 and 9 (76, 77). Additionally, the arginine N-acetylglucosamine transferase NleB1 was shown to target, modify, and inactivate proteins containing death domains, including TRADD, FADD, TNF receptor, and RIPK1. As a result, NleB1 interferes with both apoptosis and necroptosis. In the absence of NleB1, C. rodentium displayed reduced colonization of the murine gut, highlighting the pivotal role of cell death inhibition during enteric infection (77–82). Interestingly, the NleB1 paralog NleB2 was recently shown to also posttranslationally modify RIPK1 and TRADD. However, the nucleotide-sugar modification differs from that of NleB1, as NleB2 was shown to preferentially utilize UDP-glucose instead of UDP-GlcNAc (83, 84).

Modulation of apoptosis/necroptosis by Salmonella and Shigella.

A/E pathogens tightly interact with the extracellular surface of epithelial cells and have evolved multiple strategies to mitigate the cellular stresses caused by their localized attachment. Instead, intracellular pathogens such as Salmonella and Shigella are well adapted to life within host cells, where their metabolic activity and rapid replication can generate proapoptotic stress signals and trigger their detection by host cell pattern recognition receptors (PRRs). Salmonella enterica serovar Typhimurium, for example, resides predominantly in the Salmonella-containing vacuole, where it replicates in epithelial cells and macrophages. In some instances, S. Typhimurium has also been reported to replicate in the cytosol of epithelial cells (85–88). Occupation of these different cellular compartments exposes Salmonella to multiple surveillance mechanisms, against which it deploys multiple strategies to circumvent cell death responses triggered by its presence. S. Typhimurium encodes three homologs of NleB1: SseK1, -2, and -3. SseK1 and -3 were shown to target TRADD and/or FADD for GlcNAcylation in infected macrophages to prevent necroptosis, while the role of SseK2 is still unclear (89, 90). Moreover, Salmonella translocates two effector proteins which interfere with apoptosis (Fig. 1; Table 1). The phosphatase activity of SopB blocks apoptosis by stimulating the activity of the prosurvival Akt serine/theonine kinase (91). In addition, the effector AvrA is an acetyltransferase that modulates apoptosis by inhibiting c-Jun NH₂-terminal kinase (JNK) (92).

Shigella spp. are major etiological agents of bacillary dysentery (shigellosis) (93–96). Shigella is responsible for significant mortality in lower-income countries. In developed countries, Shigella sonnei is now emerging as a sexually transmitted agent that is increasingly resistant to classical antibiotic therapies (97). Following T3SS-mediated invasion of epithelial cells, Shigella ruptures its vacuole and undergoes rapid cytosolic replication. These events cause cellular stress, inducing damage to the mitochondria, genotoxic stress, and the production of oxidative species, all of which can induce cell death (98). To mitigate these effects and maintain its replicative niche, Shigella deploys a panel of effector proteins with distinct activities to delay cell death while permitting its intracellular replication (Fig. 1; Table 1).

Shigella exploits the effector IpgD to modulate a range of host cell processes (1, 2). IpgD is a phosphoinositide phosphatase that converts PtdIns(4,5)P2 to PtdIns(5)P during cell invasion, promoting actin reorganization and limiting the propagation of Ca²⁺ signals (99). Furthermore, the PtdIns(5)P molecules generated by IpgD activate the PI3K/Akt pathway and induce the degradation of the transcription factor p53, interfering with these pathways to promote host cell survival (98, 100, 101). Additionally, IpgD activates the small GTPase Arf6 to enhance bacterial entry (102). Since one of the many functions of Arf6 is to promote cell viability, one could speculate that IpgD-mediated activation of Arf6 may also promote cell survival (103–105). However, the IpgD inhibitory roles on apoptosis need to be examined in the context of multiple T3SS effectors, as infection with ipgD-deficient bacteria did not result in a sharp increase in apoptotic cell numbers, suggesting that the pathogen employs various strategies to prevent the demise of infected cells (106).

The invasion of host cells by intracellular bacteria can cause massive fluctuations in Ca²⁺ levels that could lead to apoptosis if left unchecked (107). As this ion is a potent secondary messenger whose concentration is oscillatory within the cytosol and organelles, its control by bacterial effectors is necessary to maintain cell host viability during infection (99, 108). Shigella’s first line of attack is to initially limit the induction of global Ca²⁺ responses by using IpgD, which prevents the release of Ca²⁺ from intracellular stores (99). However, the global levels of cytosolic Ca²⁺ become elevated at later infection stages (109). Calpains are Ca²⁺-activated proteases and function as both positive and negative regulators of apoptosis (110). Calpains are activated by increases in cellular Ca²⁺ concentrations, which are predominantly the result of mitochondrial damage, a potent trigger of intrinsic apoptosis. The proteolytic degradation of cytosolic proteins and components of the host cell membrane by calpains has been proposed as one of their functions in the induction of apoptosis (111). Shigella uses its VirA effector to stimulate the degradation of calpastatin, an inhibitor of calpain. Derepressed, calpain promotes the early degradation of the cell cycle regulator p53, which facilitates context-dependent apoptotic signaling (94). Thus, the combined activities of VirA and IpgD aim at multiple cellular targets to abrogate the early initiation of apoptosis and to prolong cell viability.

Finally, the effectors OspC1 and OspD3 further the manipulation of apoptosis and necroptosis by Shigella. OspC1, which shares 64% sequence homology with the antipyroptosis effector OspC3 (discussed below), was recently shown to inhibit caspase 3/7-dependent apoptosis in epithelial cells (49). Although the molecular target of OspC1 is yet to be identified, its activity inhibits caspase 8. Recent work has shown that the OspC family proteins share a previously undescribed enzymatic activity called ADP-riboxanation (112). It will be interesting to determine if this sophisticated process contributes to OspC1 antiapoptotic activity. Importantly, the high caspase 8 activity, normally observed during apoptosis, inhibits necroptosis signaling via cleavage of RIPK1/3 (113). One can hypothesize that the inhibition of caspase 8 by OspC1 is sensed by the host cell and shifts the engagement of cell death toward the necroptosis pathway. Interestingly, Shigella is equipped with another effector to counteract this host response. OspD3 is a protease that cleaves RIPK1 and RIPK3, leading to their degradation (49, 114) (Fig. 1). As a result, by preventing the activation of MLKL and necroptosis, cells infected with wild-type (WT) Shigella remain alive, while cells infected with an ospD3 mutant die. However, cell death can be prevented if an ospC1 deletion is introduced in an ospD3 mutant background, and infection with an ospC1 ospD3 double mutant does not lead to MLKL phosphorylation. These fascinating findings highlight the activation of necroptosis as a fail-safe mechanism following inhibition of apoptosis by a bacterial pathogen (49). Furthermore, OspD3 is homologous to the EPEC/EHEC effector EspL, a cysteine protease which had previously been shown to prevent necroptosis by targeting the RHIM domain of RIPK1/3, leading to their degradation (115). Whether EspL has been evolutionarily acquired to thwart the host fail-safe mechanisms engaged by the inhibition of apoptosis by EPEC effectors, such as the caspase inhibitor NleF, is yet to be examined.

PYROPTOSIS AND INFLAMMASOMES

The execution of apoptosis and necroptosis may not be fast enough to eradicate the niche of rapidly growing intracellular bacteria (59, 116). Epithelial cells and phagocytes are also equipped with the cellular machinery to execute pyroptosis, a rapid form of lytic and inflammatory PCD. Activated following detection of microbial components by specialized sensors, pyroptosis is characterized by the release of the highly inflammatory cytokines interleukin 1β (IL-1β) and IL-18. These chemoattractants recruit phagocytes and inflammatory cells such as neutrophils to the site of infection. While contributing to host defense, their abundant influx could disrupt the integrity of tissues, promoting bacterial dissemination. Bacterial pathogens can exploit this fine balance to their advantage by inducing or blocking pyroptosis, according to their infection modalities (Fig. 2).

FIG 2.

Pyroptosis induction by different inflammasomes and their manipulation by pathogens. Inflammasomes sensors are shown in dark blue rectangles. Other inflammasome components are shown in light blue rectangles. Yersinia, predominantly extracellular, injects effectors from outside the cell. Detection of its type III secretion system (T3SS) induces the NLRP3 inflammasome. This is inhibited by the effector YopK. The Yersinia effectors YopE and YopT modify the small Rho GTPase RhoA to prevent phagocytosis. The resulting activation of the pyrin inflammasome is inhibited by YopM. Shigella is an intracellular pathogen that blocks each step of the noncanonical inflammasome caspase 4/11. The Shigella IpaH9.8 effector degrades GBP proteins, preventing their signaling to caspase 4/11, itself blocked by the effector OspC3. Shigella IpaH7.8 also targets the executioner gasdermin D for degradation. Effectors from Yersinia, Shigella, EPEC/EHEC, and Salmonella are shown in related colors (see key on the right). Proteolytic cleavage events are indicated by dotted lines.

Pyroptosis is triggered upon detection of microbes in the cytosol of epithelial and immune cells. Pyroptosis is controlled by inflammasomes, macromolecular platforms whose assembly and composition are defined by the nature of their pathogenic ligands. Some inflammasome sensors detect the presence of pathogen-associated molecular patterns (PAMPs). These molecules comprise conserved bacterial factors (e.g., lipopolysaccharide [LPS], peptidoglycan, and flagellin) and are specifically recognized as nonself by PRRs. For example, Nod-like receptor (NLR) family CARD domain-containing protein 4 (NLRC4) detects the presence of bacterial flagellin or various components of the T3S machinery (117). Other inflammasome sensors detect damage or danger-associated molecular patterns (DAMPs) which arise from the disruptions of host cell homeostasis caused by infection. These include ATP and ion fluxes, abundance of reactive oxygen species, and membrane damage (118, 119). Thus, inflammasomes are versatile and efficient sensors capable of directly detecting structural components of bacterial pathogens as well as the damage caused by bacterial virulence factors. For example, NLR family pyrin domain containing 3 (NLRP3) senses a range of stimuli that include membrane damage induced by pore-forming toxins or the insertion of the T3SS translocon into cell membranes (120–122). Interestingly, the pyrin inflammasome, much like a plant immunity guard protein, detects molecular modifications to small GTPases that are key to host cell function and a common target of invasive bacteria (123–126).

Inflammasomes are classically activated by two signals. In a first step, often referred to as priming, PAMP detection by PRRs leads to the transcriptional upregulation of pro-IL-1β and pro-IL-18, as well as other subunits of the inflammasome, including NLRP3. Further PAMP or DAMP detection by inflammasome sensors induces the oligomerization of large signaling platforms that recruit caspase 1. Following its proteolysis and maturation, caspase 1 processes IL-1β and IL-18 into mature cytokines. A noncanonical inflammasome was more recently identified (further described below) whereby caspase 11 in mice or caspase 4 and 5 in humans (hereafter referred to as caspase 4/11) sense the presence of LPS in the cytosol and directly process pro-IL-1β and IL-18.

In all cases, inflammasome activation culminates in the processing of the pyroptosis executioner gasdermin D, which oligomerizes as pores into the host plasma membranes. The insertion of gasdermin D pores lyses cells, eradicating the pathogen niche while enabling the release of leaderless IL-1β and IL-18, inducing a proinflammatory response (127–129).

In contrast to extracellular PRRs (such as some TLRs), which indiscriminately detect commensals and pathogens, inflammasomes are specific intracellular pathogen sensors. As such, they are ideal targets for manipulation by pathogenic bacteria (Fig. 2). Proinflammatory signaling that drives expression of pro-IL-1β and pro-IL-18 (signal 1) has long been a known target of bacterial pathogens (19, 20, 130). However, many bacterial pathogens also directly target inflammasome assembly and signaling (signal 2) (Fig. 2). For example, the EPEC effector NleA binds and inhibits NLRP3 in macrophages (131), while NleF blocks caspase 4 activation in intestinal epithelial cells (76, 132).

Modulation of inflammasomes by Salmonella and Shigella.

Salmonella is known to block the activation of inflammasomes by using one of its two T3SSs (133), but only a limited number of death-antagonistic effectors have been identified so far (Table 1). For instance, the inositol phosphatase activity of SopB is required to prevent the activation of NLRC4 (134). Salmonella also downregulates NLRC4 expression, by an unknown mechanism (135). Additional reports have demonstrated that Salmonella can downregulate the expression of its flagella to evade NLRC4 detection, and Salmonella serovar Paratyphi A strains exhibit long O-antigen chains which interfere with the activation of multiple inflammasomes (136, 137). Considering the successful adaptation of Salmonella to intracellular environments and its ability to target other inflammatory pathways, the discovery of additional effectors targeting PCDs can be expected.

The importance of inflammasomes in antibacterial defense was brought to light by infecting NLRC4-deficient mice with Salmonella. These mice displayed reduced survival and increased bacterial loads following orogastric infection (138, 139). Several groups have since demonstrated that NLRC4 in intestinal epithelial cells was specifically required for immunity against invasive bacteria (140, 141). There, NLRC4 inflammasome activation leads to extrusion or expulsion of the infected cell from the monolayer (140, 141). Interestingly, NLRC4-deficient mice are also highly susceptible to colitis induced by Shigella, a human-adapted pathogen (142). Mitchell et al. demonstrated that Shigella was able to antagonize human NLRC4 but not the murine counterpart. Importantly, ablation of NLRC4 rendered mice susceptible to Shigella infection. These findings highlight the NLRC4 inflammasome as a key determinant of Shigella host tropism while providing a long-awaited murine model to study shigellosis (142).

While the specific effector(s) antagonizing NLRC4 is unknown, other effectors have been shown to manipulate different inflammasomes. Interestingly, Shigella spp. appear to induce pyroptosis in phagocytes while blocking it in epithelial cells. It is tempting to speculate that this is an evolved trait, whereby epithelial cells must be preserved as a replicative niche, while pyroptosis of macrophages is used to reduce the cellular pool of antimicrobial phagocytes. Using the ubiquitin ligase IpaH7.8, Shigella activates the NLRP1 and NLRP3 inflammasomes in macrophages, inducing pyroptosis and promoting severe inflammation (143–145). In epithelial cells, Shigella inhibits the NLRP3 and noncanonical (caspase 4/11) inflammasomes. IpgD is another such effector which targets the NLRP3 inflammasome. The generation of PtdIns(5)P by IpgD (see above) modulates the activity of connexin hemichannels, which, uncontrolled, would promote ATP flux and activation of NLRP3 (146).

Shigella is also equipped with a panel of effectors that inhibit the noncanonical inflammasome (Fig. 2; Table 1). This inflammasome is activated upon cytosolic detection of LPS by caspase 11 in mice or caspase 4/5 in humans. An essential initial step to noncanonical inflammasome activation is mediated by interferon-inducible guanylate binding proteins (GBP1 to -4) (147–150). These were initially shown to promote lysis of the pathogen-containing vacuole (150), delivering LPS to caspase 11 in the cytosol of murine cells. More recently, human GBP1 was shown to directly bind LPS, which initiates the assembly of a signaling platform containing GBP1 to -4, resulting in the activation of caspase 4 (147–149). Caspase 4 then catalyzes the cleavage and activation of the pyroptosis executor gasdermin D. It is noteworthy that Shigella has evolved a set of effectors interfering negatively with each signaling step of this cascade (151). Firstly, the ubiquitin ligase IpaH9.8 ubiquitylates GBPs to target them for degradation (147, 152–154). Subsequently, OspC3 targets caspase 4 (and caspase 11) to prevent the initiation of pyroptosis (112, 155–157). Recent work has shown that this inhibition is mediated through ADP-riboxanation (112). Finally, IpaH7.8 targets gasdermin D (as well as gasdermin B) for degradation (158, 159). Importantly, GBP1-, caspase 11-, and gasdermin D-deficient mice were highly susceptible to Shigella infection. However, while Shigella strains lacking IpaH9.8 or OspC3 displayed attenuated phenotypes in WT mice, they were no longer attenuated in GBP1- and/or caspase 11-deficient mice (112, 147, 155, 157, 158). Together, these findings strongly reflect the pivotal role of pyroptosis in the defense against cytosolic bacteria while also highlighting the exquisite functions of Shigella T3SS effectors and their importance in pathogenesis.

Modulation of inflammasome activation by Yersinia.

Finally, Yersinia species, including the etiological agent of plague, Yersinia pestis, represent a fascinating example of the interplay between extracellular pathogens and inflammasomes (Fig. 2; Table 1). Yersinia species have evolved complex mechanisms to disarm host immune responses. All strains deploy T3SS effectors to hamper their phagocytosis by macrophages and inhibit innate sensing pathways. For example, the effector YopJ inhibits proinflammatory signaling while promoting cell death (160–163). In response to this virulence mechanism, the host cell can detect the insertion of the T3SS translocon into its plasma membrane, triggering the activation of the NLRP3 inflammasome. The effector YopK circumvents this response and has been found to block NLRP3-mediated caspase 1 processing and subsequent pyroptosis (164). This is mediated by the direct interaction between YopK and components of the translocon, shielding them from detection by host sensors and preventing hyperinjection of translocon proteins and lysosomal damage (164–166). Furthermore, the effectors YopE and YopH also inhibit NLRP3 activation (167). Therefore, while the Yersinia T3SS is detected as a danger signal by the NLRP3 inflammasome, this pathogen has evolved effectors to block the downstream execution of pyroptosis.

Another elegant example of the interplay between Yersinia effectors and the host innate immune response occurs at the level of the pyrin inflammasome. This inflammasome is particularly interesting, as it is activated in response to molecular changes in host proteins brought upon by pathogenic virulence factors themselves. This phenomenon, coined “effector-triggered immunity,” was identified in plants and called the “guard hypothesis” (168). Indeed, plant immunologists were the first to discover that pathogen-induced molecular modifications to host proteins could be sensed by the plant cells. This phenomenon has since been described in metazoans and is characterized by the protection or “guarding” of host proteins that are commonly targeted by pathogens (123, 124).

Many pathogenic bacteria target Rho GTPases to induce changes to the actin cytoskeleton to favor or prevent cell internalization. In the case of invasive bacteria such as Salmonella, this promotes their invasion of nonphagocytic cells. Other bacteria like Yersinia spp. inhibit Rho GTPases to avoid their phagocytosis. Yersinia inactivates RhoA using YopE and YopT, two T3SS effectors targeting Rho GTPase activities differently. YopE is a GTPase-activating protein (GAP) (169, 170), while YopT is a cysteine protease that targets the localization domain of Rho GTPases, thereby disrupting their function (171). Such changes are detected by the pyrin inflammasome (125). Under physiological conditions, pyrin is phosphorylated and locked in an inactive state. This phosphorylation is mediated by ribosomal S6 kinases (RSKs) and protein kinase C-related kinase (PRK), themselves activated by RhoA. Therefore, inactivation of RhoA leads to inactivation of RSK and PRK, leading to the subsequent activation of pyrin (125, 172). Thus, pyrin can be considered a sentinel or guard of RhoA activity that can detect the modifications of Rho GTPases caused by T3SS effectors. Therefore, the combined antiphagocytic activities of YopE and YopT effectors are detected by the cell via the activation of the inflammasome (173). However, Yersinia has evolved a pyrin inhibitor, YopM. YopM not only binds and inhibits pyrin directly, it also hijacks RSKs and PRKs, forcing the phosphorylation and therefore the inactivation of pyrin (174–177). The importance of YopM in Yersinia pathogenesis is further supported in vivo, wherein yopM-deficient Yersinia is attenuated in WT mice. This deficiency is abrogated in mice lacking pyrin (174). However, upon infection with a strain lacking yopE and yopT, which does not activate pyrin, YopM was no longer essential for virulence (174, 175). These findings further underline the interplay between these three effectors and the pyrin inflammasome. Yersinia is a fascinating example of the tug of war between a pathogen and its host. YopE and YopT inhibit Rho GTPases and block bacterial phagocytosis. This modification would trigger the pyrin inflammasome if not for the presence of YopM, which blocks pyrin activation. Fascinatingly, carriage of these effector genes on the Yersinia virulence plasmid has occurred by different evolutionary routes across the genus. Convergence of these factors in a single mosaic plasmid reiterates their importance in the evolution of the Yersinia genus (178). This adaptation to host defenses is a testament to the success of Yersinia as a pathogen in human history.

CONCLUSIONS AND FUTURE PERSPECTIVES

During coevolution with their human hosts, bacterial pathogens have evolved a diverse repertoire of secreted proteins which modulate multiple arms of host cell death to facilitate infection. The study of these antagonists has taught us key lessons about the initiation of host cell death and the role of these germ line-encoded pathways in innate immunity. While analyses of the cell death-antagonistic functions of individual effectors are required, future work should examine their network-like interactions (17, 18, 179, 180) in the stratified manipulation of host cell death. This review has highlighted both synergistic and antagonistic activities of T3SS effectors in modulating the host cell death mechanisms. For example, Shigella utilizes OspC1 to inhibit apoptosis, which shifts the cell to a necroptotic program, whereas OspD3 inhibits this “backup” PCD. The numerous effectors with PCD-antagonistic activities in phylogenetically distinct Gram-negative human pathogens indicate that inhibition of cell death confers a strong evolutionary advantage to bacteria. Why, therefore, do some pathogens have such a diverse repertoire of PCD-modulating effector networks, including effectors which appear to initiate cell death? One possibility is that effector repertoires mirror the ongoing evolution of human pathogens in the face of their host’s highly sophisticated innate immune processes. This is best exemplified by EPEC and Yersinia, both extracellular pathogens that have evolved distinct effector repertoires to target the same host PCD pathways. Also, different PCD modes may be beneficial to pathogens during the infection of their targeted host cells, thus requiring different effector sets depending on the cellular context. This latter point is exemplified by Shigella, which causes the death of macrophages to promote its dissemination to the basolateral side of epithelia, whereas inhibition of PCD in epithelial cells enables bacterial replication to high numbers. Could the induction of PCD in one cell type and its inhibition in another represent an adaptive trait emerging in the face of host cell selection pressures, or is it an evolutionary trade-off in the human host? We anticipate that answering these fundamental questions will shed key insights into the evolution of PCD in multicellular organisms and in the acquisition of bacterial effectors.

Biographies

Joseph J. Wanford pursued his undergraduate and doctoral training at the University of Leicester Department of Genetics and Genome Biology. His Ph.D. with Professor Marco Oggioni focused on the tissue tropism of two major human pathogens, Klebsiella pneumoniae and Streptococcus pneumoniae, during infection. There, he developed multiple preclinical infection models to define a critical role of macrophages in disease outcome. After a brief sabbatical working in a diagnostics lab as part of the COVID-19 pandemic response, he took up a Postdoctoral Research Associate position at King’s College London with Dr. Charlotte Odendall. There, he is studying the role of the Shigella sonnei type 3 secretion system (T3SS) in modulating innate responses, including the production of cytokines and induction of programmed cell death. Joe is fascinated by the lessons we can learn about immunity from host-adapted bacterial pathogens, which he believes to be nature’s best cell biologists.

Abderrahman Hachani (known as Abdou) received his Ph.D. degree from Université Libre de Bruxelles in Belgium, where he carried his work out on Shigella T3SS machinery and effectors under the supervision of Abdelmounaim Allaoui in collaboration with Philippe Sansonetti. He joined the laboratory of Alain Filloux, first at CNRS-Marseille in France and then at Imperial College London in the United Kingdom, to work on the multiple Pseudomonas aeruginosa virulence mechanisms, including the T3SS and the T6SS and their effectors. As a mobile scientist and with a Global Marie-Sklodowska Curie Fellowship, Abdou moved to the Peter Doherty Institute in Melbourne, Australia, to expand his understanding in bacterial effector biology with Elizabeth Hartland. His “training without borders” reflects the passion of Abdou for host-pathogen and interbacterial interactions. He is now exploring these fields in the broader context of clinical disease to illuminate the mechanisms of pathoadaptation that support the success of bacterial pathogens.

Charlotte Odendall received her undergraduate and postgraduate degrees from Imperial College London. She carried out her graduate work on Salmonella pathogenesis under the supervision of David Holden. She then joined the lab of Jonathan Kagan at Harvard Medical School/Boston Children’s Hospital to study host innate immune processes, in particular the regulation of type I and III interferons. Charlotte then started her lab as a Research Fellow, now funded by the Wellcome Trust and Royal Society, at King’s College London. The Odendall lab investigates immune responses to enteric bacteria, as well as how bacterial pathogens manipulate host processes to colonize the host. Charlotte has been working in the field of host-pathogen interactions since the first experiment she did in a lab as an undergraduate student. She is fascinated by the evolutionary arms race between pathogens and their host and how studying the virulence mechanisms of bacteria has taught her about host processes.