The unexpected link between infection‐induced apoptosis and a T_h17 immune response

Short abstract

Review on the regulatory relation between apoptosis during infection induction of T_h17 responses.

Abstract

Microbial pathogens can initiate MOMP in host cells and as such, initiate the mitochondrial pathway of apoptosis. Innate immune recognition of cells dying in this way by infection‐induced apoptosis would involve recognition of ligands derived from the apoptotic host cell simultaneously with those derived from the infecting pathogen. The resultant signal transduction pathways engaged direct DCs to concomitantly synthesize TGF‐β and IL‐6, two cytokines that subsequently favor the differentiation of naïve CD4 T cells into T_h17 cells. Citrobacter rodentium is one rodent pathogen that targets mitochondria and induces apoptosis, and blockade of apoptosis during enteric Citrobacter infection impairs the characteristic T_h17 response in the intestinal LP. Here, we review these original findings. We discuss microbial infections other than Citrobacter that have been shown to induce T_h17 responses, and we examine what is known about the ability of those pathogens to induce apoptosis. We also consider types of cell death other than apoptosis that can be triggered by microbial infection, and we highlight how little we know about the impact of various forms of cell death on the ensuing adaptive immune response.

Abbreviations

ΔΨ_m

mitochondrial transmembrane potential

ADPRT

ADP ribosyl transferase

A/E

attaching and effacing

ASC

apoptosis‐associated speck‐like protein containing a caspase recruitment domain

BAI1

brain‐specific angiogenesis inhibitor 1

Cag

cytotoxin‐associated gene

CD95L

CD95 ligand

EHEC

enterohemorrhagic strains of Escherichia coli

EPEC

enteropathogenic Escherichia coli

EspF

enteropathogenic Escherichia coli‐secreted protein F

Exo

exoenzyme

Foxp3

forkhead box p3

GAP

GTPase‐activating protein

GAS6

growth arrest‐specific 6

gld

generalized lymphoproliferative disease

HMGB1

high‐mobility group box 1 protein

KC

keratinocyte‐derived chemokine

LEE

locus of enterocyte effacement

LP

lamina propria

lpr

lymphoproliferation

MER

c‐mer proto‐oncogene tyrosine kinase

MFG‐E8

milk fat globule EGF8

MMP

mitochondrial membrane permeabilization

MOMP

mitochondrial outer membrane permeabilization

MPT

mitochondrial permeability transition

NLR

Nod‐like receptor

PAI

pathogenicity island

PS

phosphatidylserine

Reg III

regenerating islet‐derived 3

RIP1/3

receptor‐interacting serine‐threonine kinase 1/3

ROR

retinoic acid‐related orphan receptor

T3SS/T4SS

type III/IV secretion system

TIM‐4

T cell Ig domain and mucin domain‐4

TIR

Toll/IL‐1R homology

T_reg

T regulatory cell

TRIF

Toll/IL‐1R homology domain‐containing adaptor‐inducing IFN‐β

VacA

vacuolating cytotoxin A

WT

wild type

Introduction

The immune system defends against infection by responding rapidly to microbial pathogens through a process of recognition that is based on the detection of conserved molecular structures uniquely expressed by the invading microbe. These structures are collectively known as PAMPs and are recognized by inflammatory PRRs expressed on the surface of phagocytic cells such as DCs, macrophages, and neutrophils [1]. Among the PRRs that recognize microbial pathogens, TLRs are the best characterized. TLRs are type 1 integral membrane glycoproteins, characterized by an ectodomain composed of leucine‐rich repeats responsible for recognition of PAMPs and a cytoplasmic signaling domain homologous to that of the IL‐1R, termed the TIR domain [2]. Recognition of pathogens by PRRs activates conserved host defense signaling pathways that control the expression of a variety of immune response genes, the majority of which leads to inflammatory responses [3, 4]. The TIR domain functions in the recruitment of adaptor proteins, of which, there are five: MyD88, MyD88 adaptor‐like (also known as TIR domain‐containing adaptor protein), TRIF (also known as TLR adaptor molecule 1), TIR domain‐containing adaptor‐inducing IFN‐β‐related adaptor molecule, and sterile‐α and armadillo motif‐containing protein. The ability of TLRs to recruit specific combinations of adaptors results in the activation of specific transcription factors, giving appropriate and effective responses to particular pathogens [5]. Engagement of TLRs mobilizes the activation of MAPK p38, ERK, and JNK and transcription of NF‐κB and IFN regulatory factor 1‐responsive genes, pivotal for the transcriptional initiation of a number of inflammatory, antimicrobial defense, and immune response genes [4, 5].

Importantly, PRRs also play a critical role in the recognition of dying cells in a process that is crucial for the maintenance of normal tissue turnover. Programmed cell death describes the regulated process of cell death, executed through specific intracellular signaling pathways, a process that is central to development and tissue homeostasis. Apoptotic cells express ligands, known as ″eat me″ signals, such as externalized PS [6] or ICAM‐3 [7, 8], and existing molecules that are altered by oxidation, such as oxidized low‐density lipoprotein‐like moieties [9, –, 11]. These molecules engage PRRs directly, or they can do so indirectly via specific “bridging” molecules (reviewed in refs. [12, 13]). The indirect recognition of PS can occur via bridging molecules, such as MFG‐E8, which links PS to phagocyte α_vβ₃ integrin [14], and GAS6 as well as Protein S, which link PS to the TYRO3, AXL, and MER receptor tyrosine kinases [15, –, 17]. Three other receptors that recognize PS include a TIM‐4 molecule on macrophages and DCs [18, 19], BAI1 [20], and stabilin‐2 [21]. Innate recognition of apoptotic cells induces the synthesis of anti‐inflammatory mediators, such as TGF‐β, PGE2, and platelet‐activating factor, by macrophages [22]. As a result, PRRs mediating the clearance of apoptotic cells are classified as noninflammatory PRRs.

APCs can discriminate between an apoptotic cell and a pathogen, according to the types of PRRs that are engaged during the activation of the innate immune response ( Fig. 1 ) [23, 24]. When pathogenic bacteria such as EPEC or Pseudomonas aeruginosa induce cell death [25, 26], the hostˈs immune system is exposed to signals from the microbial pathogen but also to signals derived from the infected dying cells [27]. Given that APCs are capable of responding to different signals by activating different PRRs, it might be expected that combinatorial signals derived from inflammatory and noninflammatory PRRs would have a large impact on host defense against microbial pathogens that induce cell death. We asked the question: what would be the resultant T cell response following simultaneous activation of inflammatory and noninflammatory PRRs in vitro and in the context of infection? Here, we discuss the evidence leading up to our hypothesis that the phagocytosis of infected apoptotic cells by DCs can induce the simultaneous production of IL‐6 and TGF‐β and lead to the polarization of T_h17 cells [28]. We summarize our studies that associated the characteristic T_h17 cell response, induced in response to Citrobacter rodentium infection, with the ability of Citrobacter to cause host cell apoptosis [28]. We discuss other bacterial infections known to induce T_h17 and IL‐17 responses, and we consider the evidence linking these infections to the induction of host cell apoptosis. Finally, we briefly review different types of cell death and emphasize that the nature of adaptive T cell responses that might be mobilized in response to these forms of cell death is still not known.

Figure 1.

Inflammatory infection and immunosuppressive apoptosis. At steady‐state, DCs phagocytose apoptotic cells during normal tissue turnover. This figure depicts the recognition of apoptotic epithelial cells by DCs at the epithelial surface. Here, the apoptotic cell expresses the eat me signal PS, which is recognized by PRRs expressed on the cell surface of DCs, such as TIM‐4, stabilin‐2, and BAI‐1. It has been shown that the regulatory cytokine TGF‐β is induced in response to uptake of apoptotic cells, and as such, this type of phagocytosis is known as noninflammatory. In contrast, microbial pathogens, as illustrated by bacteria here, activate TLRs on the surface of DCs by presenting ligands known as PAMPs, which are associated with particular bacteria. In this case, activated DCs produce proinflammatory cytokines, such as IL‐12, IL‐6, and TNF‐α, ultimately inducing an inflammatory, antimicrobial response required to fight the infection.

THE CYTOKINES SECRETED BY ACTIVATED APCs INSTRUCT SPECIFIC ADAPTIVE IMMUNE RESPONSES

Following activation by APCs, αβ T cells differentiate in a process first described for naïve CD4 T cells on the basis of their function and the cytokines they secrete [29]. The fate of naïve T_h cells is determined by three signals provided by stimulated APCs. First, the APC presents peptide on MHC molecules to the T cell through the TCR and delivers signal 1 to T cells. A second signal, referred to as ″costimulation″, is an accessory signal, required in addition to signal 1, to induce immunity [30] and is often equated to interactions among CD28, CTLA‐4, ICOS, and members of the B7 family, such as CD80 and CD86 [31]. In the absence of signal 2, T_h cells become anergic, which can lead to tolerance. The third signal delivered from the APC to the T cell is through the release of proteins, primarily cytokines, which determine its differentiation into an effector cell, such as T_h1 or T_h2 cells [32], as well as T_h17 or T_reg cell populations [33].

The polarization of a T cell response strongly depends on the conditions under which the APC is stimulated, as this determines the balance of the cytokines that are produced [34]. In addition, the fate of the T cell is controlled by lineage‐specific transcription factors, and the resulting cell phenotype is characterized by a unique cytokine expression profile and effector function [35]. IL‐12 signaling through STAT4, which leads to the activation of the transcription factor T‐bet, is essential for the induction of T_h1 cells [32]. IL‐4 is the critical cytokine that controls T_h2 cell polarization [32]. Activation of STAT6 by IL‐4 enhances the expression of IL‐4, IL‐13, and the transcription factor GATA‐binding protein 3 [36]. TGF‐β is required for the generation of T_reg cells—the natural and inducible populations [37]—both of which express the transcription factor Foxp3 [38]. IL‐6 and TGF‐β are required for murine T_h17 cell differentiation from naïve CD4⁺ T cells [39, –, 41], and IL‐23 supports T_h17 cell expansion and maintenance [41]. A role for IL‐21 induced by IL‐6 in the presence of TGF‐β has also been shown to drive T_h17 cell generation [42, 43]. It was shown originally that human T_h17 cell differentiation required IL‐1 plus IL‐23 or IL‐6 [44, 45]. Studies in IL‐1R1‐deficient mice demonstrated that IL‐1 was essential for promoting IL‐17 production by T cells that mediate autoimmune disease [46]. More recently, studies by Chung et al. [47] showed that IL‐1 is required for the early differentiation of murine T_h17 cells and T_h17 cell‐mediated autoimmunity. STAT3, the major STAT protein activated by IL‐6, is essential for T_h17 cell differentiation in the mouse and human [48]. The orphan nuclear receptors RORγt (encoded by Ror‐related orphan receptor C) and to a lesser extent, RORα, are expressed specifically by mouse and human T_h17 cells [49, 50], which not only play an important role in host defense against infection but also in autoimmunity [45]. Other T_h cell fates have been described, including T_h9 [51, 52], T_h22 [53, 54], and follicular T_h cells [55, –, 57, 58].

γδ T cells are a relatively rare population of T cells (2–4%) that are evolutionary conserved and found in all jawed vertebrates, but in contrast to αβ T cells, which reside primarily in secondary lymphoid organs, γδ T cells form the majority of intraepithelial lymphocytes. In the mouse, each tissue has its own specific subset of γδ T cells, each displaying a limited TCR diversity [59]. Resident γδ T cells are thought to represent a first line of defense against infection and provide certain other tissue‐specific functions, including the regulation of wound healing [60]. γδ T cells are thought to function largely like their αβ counterparts; however, there is evidence to suggest that γδ T cells do not routinely rely on APCs for antigen recognition, and they themselves have recently been shown to efficiently process and present peptide antigen to αβ T cells [61]. Along with T_h17 cells, γδ T cells are an innate source of IL‐17. Studies by Sutton et al. [62] showed that αβ T cells express IL‐23R and RORγt and produce IL‐17, IL‐21, and IL‐22 in response to IL‐1β and IL‐23 without TCR engagement. Furthermore, γδ T cells activated by IL‐1β and IL‐23 promoted IL‐17 production by CD4⁺ T cells and increased susceptibility to experimental autoimmune encephalomyelitis, a murine model for multiple sclerosis.

APOPTOSIS: ONLY ONE OF MANY FORMS OF CELL DEATH

The term apoptosis was first used to describe a mode of cell death based on morphological features that include cell shrinkage, membrane blebbing, chromatin condensation, and formation of apoptotic bodies [63]. Apoptotic cell death is executed by members of the family of caspases, cysteinyl aspartate‐directed proteases [64]. In mammalian cells, there are two pathways that result in the activation of caspases that are involved in executing apoptosis: the extrinsic and the intrinsic pathways. The extrinsic pathway can be mediated by cell surface receptors of the TNFR family, such as TNFR1, FAS, and TRAIL‐R1 [65]. Initiator caspases, including caspase‐2, ‐8, ‐9, and ‐10, become activated, leading to the subsequent activation of downstream executioner caspases, such as caspase‐3, ‐6, and ‐7. The intrinsic pathway is triggered by various stress stimuli, such as growth factor withdrawal, UV irradiation, DNA damage, and cytotoxic drugs [66, 67]. Apoptosis occurs as a result of MMP, in particular, as a consequence of MOMP, which causes the leakage of molecules, including cytochrome c; second mitochondria‐derived activator of caspase/direct inhibitor of apoptosis‐binding protein with low isoelectric point; endonuclease G; and apoptosis‐inducing factor, among many others, which are normally confined within the space between the inner and outer mitochondrial membranes [68, 69]. Release of these molecules initiates a cascade of signaling events that results in the phenotypic characteristics associated with apoptosis, including DNA fragmentation, nuclear condensation, the formation of cytoplamic blebs and of apoptotic bodies, and exposure of eat me signals, such as PS, on the plasma membrane [66].

A form of cell death that is distinct from apoptosis is a process known as necrosis, which is caspase‐independent and characterized by cytoplasmic and organelle swelling, followed by membrane breakdown and release of cellular contents into the surrounding extracellular space [70]. Unlike apoptosis, necrosis was thought to be an unregulated process, but recent studies have shown that the kinase activities of RIP1 and RIP3 initiate a cellular pathway downstream of TNFR1 [71, –, 73] to mediate programmed necrosis, also known as necroptosis [74, 75]. Interdependent phosphorylation of RIP1 and RIP3 was linked to increased energy metabolism and production of ROS, which precede necrosis [71, 72], also likely to involve MMP; this has been proposed to occur as a result of MPT, as distinct from the MOMP, which is more typical of apoptosis [75]. MPT involves a multiprotein pore called the permeability transition pore complex, a complex that functions in exchanging small metabolites between the cytosol and the mitochondrial matrix [69].

Pyroptosis is a more recently defined form of cell death described in monocytes, macrophages, and DCs (reviewed in ref. [76]). DNA damage, as a result of pyroptosis, is observed as marked nuclear condensation without intranucleosomal fragmentation of DNA or resultant DNA laddering morphology. In contrast to apoptosis, there is no loss of MOMP, no activation of the executioner caspases 3 and 6, and no release of cytochrome c. Notably, unique to this type of cell death is the activation of the inflammatory caspase, caspase‐1, which is activated in a complex initiated by NLRs and called the inflammasome [76]. Rupture of the plasma membrane is observed as a result of caspase‐1‐mediated formation of pores, which allows efflux of ions and influx of water, resulting in osmotic lysis and release of inflammatory intracellular contents.

APOPTOSIS IN THE CONTEXT OF INFLAMMATION

Noninflammatory phagocytosis of apoptotic cells by macrophages results in the synthesis of TGF‐β [22, 77]. In contrast, the activation of APCs in response to TLR ligation induces IL‐6, a prototypical, proinflammatory cytokine [78]. Our laboratory showed that uptake of infected apoptotic cells, apoptotic cells that express TLR ligands, induced the synthesis of IL‐6 and biologically active TGF‐β, as well as IL‐23 [28]. Given the critical role that IL‐6 and TGF‐β play in the induction of T_h17 cells, we hypothesized that innate recognition of an infected apoptotic cell would result in a cytokine milieu conducive for T_h17 differentiation. We demonstrated that TGF‐β and IL‐6 produced by DCs in response to infected apoptotic cells promoted the differentiation of naïve CD4 T cells to the T_h17 lineage ( Fig. 2 ) [28]. T_h17 cell differentiation was impaired severely when TGF‐β and IL‐6 were neutralized with antibodies or when Il6 ^−/− or Trif ^−/− Myd88 ^/− DCs were used, indicating that these cytokines were the major drivers of T_h17 cell differentiation in response to the TLR ligand carrying apoptotic cells in vitro. Furthermore, in the absence of TLR engagement, innate recognition of apoptotic cells induced the synthesis of TGF‐β, resulting in the preferential induction of T_reg cells (Fig. 2). These results emphasized the activation of distinct signaling pathways in DCs in response to apoptotic cells, versus infected apoptotic cells that harbor TLR ligands and ligands derived from apoptotic cells.

Figure 2.

Innate immune recognition of apoptotic cells carrying TLR ligands instructs T_h17 cell differentiation. Apoptotic cells express eat‐me signals, including exposure of PS at the outer leaflet of the plasma membrane, which activates PRRs such as TIM‐4, stabilin‐2, BAI‐1, MER, and the α_vβ₃ integrin expressed on the surface of phagocytes, such as DCs. The receptors MER or the α_vβ₃ integrin require interaction with bridging molecules such as GAS6 or MFG‐E8, respectively, to recognize PS and initiate phagocytosis. DCs that phagocytose apoptotic cells induce a noninflammatory response, characterized by the production of TGF‐β, resulting in the polarization of T_reg cells that secrete anti‐inflammatory cytokines such as TGF‐β and IL‐10. Bacterial invasion of target cells (bacteria are shown here in pink) can induce the release of cytochrome c (green), a characteristic of apoptosis. In this case, DCs are activated in response to apoptotic cell‐derived and TLR‐derived signals. Antigens derived from the apoptotic cell and the microbial pathogen end up in the phagosome, resulting in their presentation on MHC molecules to T cells. This combination of signals induces IL‐6, TGF‐β, and IL‐23, a unique cytokine profile that triggers T_h17 differentiation.

When might apoptosis occur in the context of TLR activation? Pathogenic enteric bacteria, such as EPEC, can disrupt intestinal epithelial barrier structure and function, as well as mediate apoptosis of intestinal mucosal epithelial cells [26]. Many of these bacteria possess virulence factors, such as pore‐forming toxins and adhesins. EPEC belong to a family of closely related pathogens, including EHEC, known collectively as A/E pathogens, which cause characteristic lesions when they adhere to intestinal epithelial cells, leading to effacement of brush border microvilli [26]. A 35,624‐bp genetic element known as the LEE is necessary and sufficient for the process of A/E [79, 80]. The LEE is a conserved PAI that is restricted to EPEC, EHEC, as well as the rodent pathogen C. rodentium, and encodes components of the T3SS, which mediates firm adhesion of the bacteria to the host cell and injection of various bacterial effectors directly into the host cell cytosol to alter cellular functions and survival. One of these effectors, EspF, translocates to the mitochondria to initiate MOMP and subsequent apoptosis [81, 82]. More recently, EspF has been shown to bind membrane‐ and actin‐remodeling host factors, namely SNX9 and neuronal Wiskott‐Aldrich syndrome protein, and alter their activity [83]. EPEC are known to disrupt intestinal epithelial tight junctions, a factor considered to be essential to EPEC pathogenesis [26, 83], and have been shown to induce apoptosis, as demonstrated by the exposure of PS, subsequent DNA fragmentation, and release of ATP [84, –, 86, 87, 88, 89]. C. rodentium is a commonly used rodent infection model for the study of the human infections with EPEC and EHEC.

BLOCKADE OF APOPTOSIS IMPAIRS T_h17 DEVELOPMENT IN VIVO DURING A BACTERIAL INFECTION KNOWN TO TRIGGER A T_h17 RESPONSE

Although it was well established that infection with C. rodentium induced T_h1 immunity [90], studies by the laboratory of C. T. Weaver [39] showed that infection of mice with C. rodentium induced a robust T_h17 response within the intestinal LP. The characteristic development of colitis that is associated with Citrobacter infection resolves within 2–3 weeks as a result of clearance of bacteria and development of a protective humoral and CD4 T cell‐mediated immune response [91, 92]. IL‐17A and IL‐17F are important for host defense against C. rodentium infection, such that deficiency in IL‐17A or IL‐17F resulted in splenomegaly, colonic hypertrophy, inflammation, and greater bacterial burdens in the colons of these mice compared with WT mice [93]. In addition to IL‐17A and IL‐17F, studies by the laboratory of W. Ouyang [94] showed that IL‐22 mediated protection against C. rodentium infection, such that infected Il22 ^−/− mice suffered 80–100% mortality in the second week postinfection. These mice exhibited mucosal hyperplasia, inflammation, colonic ulceration, and increased numbers of bacteria in the colonic crypts and within the mesenteric LNs, spleen, and liver [94]. The protective role for IL‐22 was mediated through the production of antimicrobial peptides of the Reg III family, Reg IIIβ and Reg IIIγ, as well as S100A8 and S100A9.

Having shown that innate recognition of an infected apoptotic cell by DCs induced IL‐6 and TGF‐β production and subsequent T_h17 cell differentiation, we tested whether this may be the relevant physiological stimulus in vivo that induces T_h17 responses. Our laboratory found that infection with C. rodentium induced T_h17 immunity and that Citrobacter induction of host cell apoptosis was necessary to induce this response [28]. Previous studies by the laboratories of B.B. Finlay, M.S. Donnenberg, and C. Sasakawa [81, 82, 95, 96] had all reported that C. rodentium induced apoptosis in vitro as well as in vivo in the intestinal epithelium of infected animals. When we infected mice with the mutant strain ΔEspF, which is unable to induce host cell apoptosis, we observed a significantly reduced number of TUNEL‐positive cells in the distal colon of ΔEspF‐infected mice compared with mice infected with WT C. rodentium [28]. Furthermore, mice infected with ΔEspF, which does not induce apoptosis, had a reduction in IL‐17‐producing CD4⁺ T cells as compared with mice infected with WT C. rodentium. Importantly, the percentage of IFN‐γ⁺ CD4⁺ T cells was similar in the small intestinal LP of mice infected with WT or ΔEspF C. rodentium, indicating that unlike the T_h17 response, the ability of ΔEspF to drive T_h1 responses was not impaired [28]. Consistent with these findings, blocking apoptosis with the pan‐caspase inhibitor Q‐VD‐OPH throughout the course of Citrobacter infection, again led to a significant reduction in the number of TUNEL‐positive cells in colonic sections of treated animals compared with infected animals that were not treated with the inhibitor [28]. This correlated with a decrease in the population of IL‐17‐producing CD4 T cells in the colon of infected mice that were treated with the inhibitor. Our studies thus highlighted a novel role for host cell apoptosis during infection in determining the differentiation of T_h17 cells. Whether this pathway also plays a role in triggering IL‐17 production by γδ T cells remains to be seen.

OTHER MICROBIAL PATHOGENS KNOWN TO TRIGGER IL‐17 RESPONSES AND THEIR LINK TO THE INDUCTION OF APOPTOSIS

Bacteria

Like C. rodentium, many other bacterial pathogens have been shown to elicit IL‐17 responses, including Klebsiella pneumoniae, Streprococcus pneumoniae, Borrelia species, P. aeruginosa, Helicobacter pylori [97], and Staphylococcus aureus [93]. The relationship between the ability of these pathogens to induce apoptosis and to trigger IL‐17‐associated responses is not yet clear. In many of the infections associated with these pathogens, a role for other cytokines involved in T_h17 responses, such as IL‐23, IL‐22, and IL‐21, has been investigated. However, not all of these studies have shown a definitive involvement of CD4 T cells that have differentiated into the T_h17 lineage, and as such, other cell types, such as γδ T cells [98, –, 100, 101, 102, 103], NKT cells [104, –, 106, 107, 108], NK cells [109, –, 111], neutrophils [112, 113], and eosinophils [114], could be contributing to the IL‐17 response being measured.

Helicobactor pylori.

H. pylori, a Gram‐negative bacterium, colonizes the proximal duodenum or distal esophagus in ∼50% of the human population [115]. Colonization is asymptomatic but has been associated with increased risk for peptic ulcers, gastritis, gastric carcinomas, and mucosa‐associated lymphomas [115]. Studies examining a link between T_h17 and Helicobacter reported a consistently higher level of IL‐17 RNA transcripts in gastric mucosa and LP mononuclear cells isolated from gastric biopsies from H. pylori‐positive compared with H. pylori‐negative patients presenting with gastritis [116, 117]. Treatment of patients for H. pylori correlated with a decrease in the levels of IL‐17 protein detected within the mucosal tissue [116]. CD4 and CD8 T cells appeared to be the major source of IL‐17 in H. pylori‐positive biopsies, but notably, there was also a higher percentage of CD4 T cells that produced IFN‐γ alone in H. pylori‐positive biopsies [118]. Together with IL‐17, significantly increased levels of IL‐23 and IL‐21 were detected in biopsies of H. pylori‐positive patients compared with H. pylori‐negative patients [118, 119]. Thus, these studies collectively suggest a role for the IL‐17/IL‐23 axis in diseases associated with H. pylori colonization [120].

Has Helicobacter been reported to induce apoptosis? H. pylori secrete VacA, which has been shown to induce apoptosis of gastric epithelial cells [121, –, 123, 124, 125] ( Fig. 3 ; reviewed in ref. [126]). Addition of VacA to cells and intracellular expression of VacA cause dissipation of mitochondrial ΔΨ_m [125] and release of cytochrome c [122, 125], consistent with MOMP. Indeed, VacA has been reported to translocate to mitochondria, suggesting a direct role in MOMP [122]. Transient transfection of VacA resulted in activation of caspase‐3, as determined by cleavage of poly(ADP‐ribose) polymerase, and this was inhibited by cotransfection of B cell lymphoma 2 [122]. Although all H. pylori strains isolated from the stomachs of rodent model animals or humans are VacA⁺ and should theoretically induce apoptosis, successful induction of apoptosis in the end is determined by whether other virulence factors, such as the Helicobacter cag PAI‐encoded T4SS and the effector protein CagA, are coexpressed [115]. Therefore, studies examining IL‐17 responses in human subjects should thus take into account the CagA status of the H. pylori present in those subjects. Caruso et al. [118, 119] found no differences in the levels of IL‐21 and IL‐23 proteins among patients carrying CagA⁺ versus CagA^– strains. On the other hand, all 51 H. pylori strains from gastric ulcer and nonulcer patients studied by Mizuno et al. [117] were CagA^– and VacA⁺ and showed increased IL‐17 transcripts, making it tempting to speculate that the IL‐17 responses observed in those patients correlate with the lack of CagA expression and possibly an intact ability to induce apoptosis.

Figure 3.

Induction of host cell apoptosis during infection with H. pylori or P. aeruginosa. H. pylori, a Gram‐negative bacterium, secretes a VacA, which causes the dissipation of ΔΨ_m and release of cytochrome c from gastric epithelial cells, resulting in their death by apoptosis. P. aeruginosa is another Gram‐negative bacterium, which is known to induce apoptotic cell death of lung epithelial cells. The ability P. aeruginosa to induce apoptosis relies on its expression of a T3SS, as well as several protein effectors that are injected into host cells, including ExoS, as shown in the diagram. LL‐37 is a peptide produced by epithelial cells in response to infection and inflammation and mediates important antimicrobial functions. Invasion of P. aeruginosa into epithelial cells along with the release of LL‐37 can result in the induction of epithelial cell apoptosis. Although we have established a link between C. rodentium‐induced apoptosis and T_h17 immunity, little is known about the impact of H. pylori‐ or P. aeruginosa‐induced apoptosis on the IL‐17 responses reported to occur in these infections.

Klebsiella pneumoniae.

Studies have shown that IL‐23, IL‐17, and IL‐22 are critical players in the early phases of the host response to K. pneumoniae infection, even before IL‐12 and IFN‐γ responses are fully developed. Increased mRNA levels of the IL‐23 p19 subunit were detected in alveolar macrophages and DCs sooner than those for the IL‐12 p40 or p35 subunits within the BAL of mice infected with K. pneumoniae [127]. IL‐23 expression in the lung of infected mice was required for the expression of IL‐17A and IL‐17F, and IL‐12 was required for IFN‐γ. IL‐23 regulated the expression of a number of IL‐17‐dependent inflammatory mediators, including G‐CSF, IL‐6, MIP‐1α, MIP‐2, LPS‐induced CXC chemokine, and KC, within the lungs. IL‐23 was additionally important for the induction of IL‐22, which was also detectable early after infection. Deficiency in IL‐12 or IL‐23 led to reduced survival of mice during a pulmonary challenge with K. pneumoniae [127]. In a separate study, CD90 (Thy‐1)⁺ T cells isolated from the lungs of infected mice produced significant levels of IL‐22 [128]. Neutralization of IL‐22 in vivo led to increased mortality of mice as a result of the infection, as compared with WT or Il17a ^−/− mice, indicating a more critical role for IL‐22 than IL‐17 in host defense against K. pneumoniae infection.

Although there is some evidence that K. pneumoniae can induce cytotoxicity of respiratory epithelial cells [129], there has not been extensive characterization of the mode of cell death involved. Microcin E492, a 7.9‐kDa, channel‐forming bacteriocin produced by K. pneumoniae [130], induced cell shrinkage, dissipation of mitochondrial ΔΨ_m, release of cytochrome c, and activation of caspases 1 and 3, events that were blocked by the pan‐caspase inhibitor z‐VAD‐fmk [131]. High doses of microcin E492, which may not reflect the physiological levels of microcin E492 produced by bacteria at the infected site, induced necrosis, and a low dose was shown to induce apoptosis [131]. These data suggest that the mode of cell death induced by K. pneumoniae is more likely to be apoptosis rather than pyroptosis. However, further studies are needed to confirm whether K. pneumoniae can induce apoptosis of respiratory epithelial cells and whether this apoptosis is mediated through microcin E492 and/or dependent on the mitochondrial pathway.

Pseudomonas aeruginosa.

P. aeruginosa is a Gram‐negative, opportunistic bacterium that causes serious pneumonia by colonizing and damaging lung epithelial surfaces in immunocompromised patients and patients with cystic fibrosis [25]. P. aeruginosa is also associated with infections of skin, wounds, burns, urinary tract, and bloodstream. In searching for an association between IL‐17 and chronic neutrophilic infiltrates that are characteristic of cystic fibrosis [132], J.K. Kolls and colleagues [133] found elevated levels of IL‐23, IL‐17A, and IL‐17F in the sputum of cystic fibrosis patients colonized with P. aeruginosa. To model airway infections in cystic fibrosis patients, this group challenged mice with agarose beads containing a clinical, mucoid isolate of P. aeruginosa and found that IL‐17, KC, the metalloproteinase MMP‐9, and IL‐6 were decreased significantly in the BAL and lung homogenates of p19 ^−/− mice compared with WT controls [134]. As a result, significantly less airway inflammation and neutrophilic infiltrates were observed in p19 ^−/− mice, and the levels of IFN‐γ and TNF‐α were similar to those in WT mice. The increases in IL‐17 and IL‐23 observed in the human samples and the murine model suggested that these cytokines may be important for neutrophil recruitment and that IL‐23 could potentially be targeted in the immunotherapy of cystic fibrosis.

There is plenty of evidence pointing to the ability of P. aeruginosa to induce apoptosis (Fig. 3). P. aeruginosa possesses a T3SS that encodes several protein effectors that are injected into host cells (reviewed in ref. [25]). Of these, ExoS, ExoT, and ExoU have been linked to the induction of host cell death. ExoS and ExoT are toxins expressed by invasive strains of P. aeruginosa, and both exhibit dual GAP and ADPRT activities [25]. The GAP activity enables Pseudomonas to disrupt tight junctions within the epithelial barrier, and the ADPRT activity results in cell rounding, inhibition of DNA synthesis, vesicular traffic, and eventual cell death. The ADPRT activity of ExoT causes a form of cell death that resembles apoptosis [135]. Similarly, ExoS‐induced cell death is also associated with hallmarks of apoptosis, including membrane blebbing, cell shrinkage, caspase‐3 cleavage, DNA fragmentation, and chromatin condensation [136]. Finally, ExoU possesses PLA₂ activity that causes rapid cell death mediated by the C terminus of the protein [137, –, 139, 140] and is characterized by loss of plasma membrane integrity similar to that seen in necrosis [140, 141].

A study by E. Gulbins and colleagues [142] showed a different mode of inducing cell death by P. aeruginosa, alternative to activation by T3SS proteins, and mediated by CD95/CD95L (Fas/Fas ligand) interactions. P. aeruginosa‐induced apoptosis of cultured human and murine cells was blocked by a CD95‐Fc fusion protein [142]. Similarly, fibroblasts derived from lpr mice, genetically deficient in CD95, or gld mice, genetically deficient in CD95L, were protected from P. aeruginosa‐mediated apoptosis [142]. Consistent with these results, intranasal infection with P. aeruginosa triggered apoptosis of lung epithelial cells that was absent in lpr or gld mice [142]. Although the P. aeruginosa virulence factors mediating apoptosis here were not identified, apoptosis played an important role in host defense against P. aeruginosa infection, as evidenced by increased splenic and lung bacterial burdens and increased mortality of infected lpr or gld mice.

A surprising new player in mediating apoptosis of infected epithelial cells is the human cathelicidin LL‐37, which is a cationic, amphipathic peptide produced by epithelial cells in response to infection and inflammation and mediates important antimicrobial functions (reviewed in refs. [143, 144]). D.J. Davidsonˈs group [145] showed that treatment of P. aeruginosa‐infected human bronchial epithelial cells with physiological levels of LL‐37 resulted in epithelial cell apoptosis, as determined by nuclear DNA fragmentation, cleavage of caspase‐3 and ‐9, loss of mitochondrial ΔΨ_m, and release of cytochrome c into the cytosol. These effects were blocked with the pan‐caspase inhibitor z‐VAD‐fmk and were not observed upon treatment of the epithelial cells with P. aeruginosa alone or LL‐37 alone, indicating an essential synergistic activity of LL‐37 and P. aeruginosa [145]. Curiously, the synergistic LL‐37/P. aeruginosa‐mediated apoptosis was lost only when a mutant of P. aeruginosa, which was incapable of invading the epithelial cells, was used. These data demonstrated that invasion of P. aeruginosa into epithelial cells, in combination with LL‐37, resulted in the induction of epithelial cell apoptosis. Whether this mode of cell death initiated by the infected epithelial cell itself can lead to T_h17 responses remains to be investigated (Fig. 3).

Fungi

Candida albicans.

IL‐17A and as such, T_h17 cells are considered to be protective in candidiasis and in general, against infections with fungi, findings that stemmed mainly from studies with experimental candidiasis and aspergillosis. C. albicans is a fungal pathogen that causes opportunistic infections in humans. It has been reported that T_h17 cells are important for host defense against C. albicans, whereby IL‐17RA‐deficient and p19 ^−/− mice showed increased susceptibility to mucosal candidiasis [146] and to systemic C. albicans infection [147]. Futhermore, evidence for a requirement for a T_h17 response to protect against infection in humans comes from the observations that patients with impaired Candida‐specific T_h17 responses, including patients with hyper‐IgE syndrome or chronic mucocutaneous candidiasis, are highly susceptible to mucosal C. albicans infections [148]. A role for Dectin‐2 in the recognition of α‐mannans and in the induction of T_h17 cell differentiation was recently shown to be essential for host defense against C. albicans [147]. The ability of C. albicans to invade and damage oral epithelial cells has been shown to be essential to establish oral infection. A recent study by Villar and Zhao [149] showed that C. albicans induce early apoptosis in oral epithelial cells, demonstrated by the activation of caspases, followed by necrotic cell death. This is in line with studies showing that systemic infection with C. albicans induced apoptosis, followed by necrosis in macrophages isolated from the peritoneal exudates of mice 30 min–2 h postinfection [150]. Whether the induction of apoptosis in response to C. albicans infection is required for the T_h17 response is still unknown. Interestingly, a recent study by Martin et al. [100] showed that the number of IL‐17‐producing γδ T cells expanded rapidly following injection of mice with C. albicans hyphae, indicating that γδ T cells may also contribute to the IL‐17 response to C. albicans infection.

INNATE RECOGNITION OF OTHER FORMS OF CELL DEATH—IMMUNOLOGICAL CONSEQUENCES

Although our studies have shown that infections associated with significant apoptosis can induce T_h17 immunity, the consequences of other types of cell death on the adaptive immune system are currently not well defined [151] ( Fig. 4 ). We will only briefly review what is known here. As mentioned above, apart from apoptosis, other forms of programmed cell death that have been described include necrosis and pyroptosis. Many intracellular bacterial pathogens such as Legionella, Listeria, Shigella, Salmonella, and Francisella trigger pyroptosis [152, –, 154, 155, 156, 157, 158]. Unlike apoptosis, the consequences of pyroptosis on the adaptive immune system are currently not defined [151]. In particular, whether innate immune recognition of pyroptotic cells can lead to the differentiation of T_h17 cells has not been investigated. Caspase‐1 activation, which is a hallmark of pyroptosis, results in the cleavage of inflammatory cytokines such as IL‐1β and IL‐18, and their release from the pyroptotic cell. Given that pyroptosis and apoptosis are mediated by different cellular pathways and result in the release of different cytokines, the prediction would be that pyroptosis would not favor development of a T_h17 response. Instead, IL‐18, in combination with IL‐12, has been reported to synergistically induce IFN‐γ production by T_h1 cells [159]. Indeed, infections with Salmonella and Listeria tend to induce T_h1 rather than T_h17 cell responses [160, 161].

Figure 4.

Although uptake of apoptotic cells carrying TLR ligands by phagocytes instructs T_h17 differentiation, the consequence of innate recognition of other modes of cell death is yet to be defined. Phagocytosis of infected apoptotic cells by DCs induces IL‐6, TGF‐β, and IL‐23 production, a unique cytokine profile that triggers T_h17 differentiation. Other types of cell death, such as necrosis and pyroptosis, also exist. Although caspase‐1 activation is typical of pyroptosis, the RIP1 and RIP3 kinases mediate death receptor‐mediated necrosis such as that downstream of TNFR1. Activation of the kinase activities of RIP1 and RIP3 targets key enzymes in bioenergetic pathways mediating glycogen breakdown, glutaminolysis, and ROS production. Caspase‐1 activation is controlled by the assembly of an inflammasome complex consisting of NLRs and the adaptor protein ASC. Activated caspase‐1 in turn mediates the processing of pro‐IL‐1β to IL‐1β. Cells undergoing pyroptosis thus release cellular contents and inflammatory cytokines such as IL‐1β, which can then activate DCs. Similarly, necrotic cells also release intracellular contents and have most typically been associated with the release of HMGB1, which can cause DC maturation. The consequence of phagocytosis of necrotic and pyroptotic cells on CD4 T cell differentiation is not yet known.

On the other hand, a study by Lin et al. [162] showed a critical role for the IL‐23‐T_h17 pathway in the immune response to a live vaccine strain of Francisella tularensis, where IL‐17A but not IL‐17F or IL‐22 induced IL‐12 production by DCs and macrophages and thus, regulated the IL‐12‐T_h1 pathway that is required for protective immunity against F. tularensis. In this study, IL‐17A was produced by γδ T cells as well as TCRβ⁺ CD4⁺ T cells, suggesting the development of T_h17 cells. Notably, and as mentioned above, Francisella induces pyroptosis, the process of caspase‐1‐dependent programmed cell death. Activation of caspase‐1 and ASC signaling, which controls the protoelytic cleavage of caspase‐1 [163] in response to Francisella, resulted in host cell death and the release of IL‐1β and IL‐18 [164]. Futhermore, F. tularensis‐infected caspase‐1‐ and ASC‐deficient mice showed markedly increased bacterial burdens and mortality as compared with WT mice [164]. These studies demonstrate an important role for caspase‐1 and ASC in innate defense against infection by this pathogen. Given the critical role that IL‐1 plays in the induction of T_h17 responses, the ability of Francisella to cause pyroptosis and as such, IL‐1β release may be important in the IL‐17‐mediated response to infection with this intracellular bacterium. However, a direct causal relationship between pyroptosis and T_h17 cell differentiation in general remains to be formally demonstrated.

A number of studies have demonstrated the occurrence of necrotic cell death during infection with microbial pathogens including Shigella, HIV‐1, West Nile virus, and Coxackievirus B (reviewed in ref. [165]). Interestingly, although Shigella mediates pyroptosis in macrophages, it was found to mediate necrosis in nonepithelial cells [166]. Consistent with the notion that the presence of RIP3 permits the cells to undergo necrosis [71, –, 73], Rip3^−/− mice failed to initiate vaccinia virus‐induced tissue necrosis and inflammation, which resulted in increased mortality as a result of exacerbated viral replication [71]. These results illustrate the importance of necrosis in host defense, particularly against infections with microbes that interfere with the ability of infected cells to undergo apoptosis. Necrotic cells have been shown to release proteins such as HMGB1 during primary and secondary necrosis [151], and HMGB1 can engage receptors, such as receptor for advanced glycation endproducts, TLR2, and TLR4 [167]. Given the potentially different PRRs that can be engaged by necrotic cells, the type of adaptive immune response induced in response to necrotic cell death is likely to be distinct and remains to be elucidated.

CONCLUDING REMARKS

Our studies have established a link between the differentiation of T_h17 cells and apoptosis of bacterially infected cells [28]. Importantly, in this case, two signals are provided to the APC during recognition of the infected apoptotic cell. Apoptotic cell‐derived signals activate noninflammatory and likely PS‐specific PRRs on the surface of the APC, and TLR ligands provided by the invading pathogen engage TLRs and induce a proinflammatory response. Although we have shown this to occur during C. rodentium infection, this situation could conceivably occur in a number of other bacterial infections that also induce host cell apoptosis, such as H. pylori and P. aeruginosa infections. The realization that apoptosis can form an important component of the innate immune triggers which impact CD4 T_h cell differentiation leads to the obvious next question as to how other forms of cell death shape the ensuing adaptive T cell response. The type of cell death induced by a pathogen is dependent on the virulence factors and mechanisms of pathogenicity unique to that pathogen. Therefore, although one pathogen induces apoptosis by targeting the intrinsic mitochondrial pathway of cell death, another might induce pyroptosis by targeting the inflammasome and caspase‐1 activation. Innate immune recognition of the particular form of cell death within the context of infection can control the induction of a specific, tailored T_h cell response against the infection. Given that many pathogens have evolved mechanisms to modulate cell death, there is a clear need to better understand how cell death shapes the host immune responses to microbes.