| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
THEMES
Department of Pathology, Emory University School of Medicine, Atlanta, Georgia
Submitted 20 June 2006 ; accepted in final form 7 July 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT REFERENCES |
|---|
Toll-like receptor 5; inflammation; apoptosis; bacterial pathogenesis; colitis; epithelia
20,000 subunits to assemble the extracellular filament structure. Flagellar filaments and monomers are shed into the environment; indeed, active secretion of flagellin may occur when flagellated bacteria make initial contact with eukaryotic cells. In the prototypical enteropathogen Salmonella typhimurium, the flagellar subunits are 55-kDa monomers encoded by the two similar, but not identical genes, FliC and FljB, whereas in other bacteria a single open reading frame encodes the protein (21). From the perspective of the host, flagellin is a microbial-associated molecular pattern (MAMP), a microbial structural determinant that can be perceived by the innate immune system, generally via pattern recognition receptors. Flagellin is increasingly recognized as an important inducer of innate and adaptive immunity in mammals and likely plays a significant role in the ability of mammals to perceive bacterial threats. Flagellin can also stimulate cellular defensive responses in invertebrates and plants, suggesting a wide range of eukaryotic life has evolved mechanisms to perceive the protein and react to it as a danger signal (21).
Flagellin is the only MAMP that is purely protein; this property permits ablation of the determinant by bacterial genetic techniques and thus presents a unique opportunity to study MAMP loss of function in in vivo and in vitro model systems. These attributes have allowed several recent advances in delineating the role of flagellin in mammalian infection.
Flagellin Receptors
Flagellin is detected by mammalian cells via the action of pattern recognition receptors, of which several are known (Fig. 1).
|
TLR5 transduces signals into the cell by means transient interactions with cytoplasmic adaptor proteins in the manner of all TLRs. As currently understood, binding of an appropriate MAMP ligand to a TLR results in dimerization and formation of a cytoplasmic TIR domain competent to bind a class of adaptor proteins (1). The original member of this family is known as myeloid differentiation primary response gene 88 (MyD88), and additional family members have been described in recent years, including Mal/TIR domain-containing adaptor protein (TIRAP), translocation-associated membrane protein (TRAM), and TIR domain-containing adaptor inducing interferon-
(TRIF). Evidence is emerging that these different adaptor proteins (or combinations of them) may preferentially interact with specific TLRs or groups of TLRs and presumably have a role in tailoring the most appropriate signaling pathways for a given MAMP/TLR recognition event (1). In the case of TLR5 specifically, to date only MyD88 has been implicated in directly interacting with the TLR5 TIR domain (9, 20). MyD88 interacts with a second adaptor molecule, IL-1 receptor-associated kinase (IRAK), of which several family members are known. IRAK, a serine kinase, then activates the cytoplasmic signaling intermediate TNF receptor-associated factor (TRAF)6. In turn, TRAF6 is then thought to activate the I
B kinase (IKK) complex and set in motion downstream signaling pathways, resulting in transcriptional upregulation of proinflammatory effector molecules.
When polarized T84 cells or nontransformed human epithelial cells are stimulated with flagellin robust activation of the NF-B pathway, p38 and JNK kinases of the MAPK signaling group, as well as ERK and AKT kinase activation, are observed within 5 min of treatment (20, 25). To date, flagellin has not been shown to activate interferon-regulatory factor (IRF) pathways, which are known to be stimulated TLR3 and TLR7. Presumably, these TLRs utilize distinct adaptor molecules that feed into IKK-
pathways and thus induce gene regulatory pathways more appropriate for the control of viral pathogens.
IL-1-converting enzyme protease-activating factor.
Recently, a second LRR-containing pattern recognition receptor, IL-1-converting enzyme (ICE) protease-activating factor (Ipaf), has been implicated in flagellin perception by macrophages (5, 11). This protein is a member of the nucleotide-binding oligomerization domain (Nod)-LRR family of intracellular pattern recognition receptors [which includes Nod1, Nod2, neuronal apoptosis inhibitory protein (Naip), and LRR and pyrin domain (Nalp) proteins] that are defined by a COOH-terminal LRR, a central Nod, and an NH2-terminal effector domain. The effector domain of Ipaf is a caspase recruitment domain that interacts with the adaptor molecule apoptosis-associated speck-like protein (ASC). In macrophages, intracytoplasmic, but, importantly, not extracellular, flagellin stimulates Ipaf association with ASC and results in the formation and activation of the so-called "inflammosome," which serves to activate capsase-1 and cleave the pro form of IL-1, an important proinflammatory effector. Ipaf has not been demonstrated to activate NF-B or proinflammatory gene transcription. Thus, present data support the notion that Ipaf represents an intracellular monitor of flagellin, and, because murine macrophages do not express TLR5 and do not respond to extracellular flagellin, Ipaf may be the only means by which these cells perceive flagellin. A major unanswered question is whether Ipaf or other cytoplasmic sensors function in epithelial cells, if indeed they are expressed at all in this cell type.
Naip5. Genetic evidence has implicated a third potential flagellin sensor. Reverse genetic analysis of a murine Legionella susceptibility locus identified Naip5, another member of the Nod-LRR family. It is one of an iterated series of duplicated genes in the murine genome (only 1 is present in the human genome). Intriguingly, the effector domain of Naip5 is a baculovirus inhibitory repeat, which places it in the same family of the antiapoptotic cellular inhibitor of apoptosis (IAP) family. Naip5 was recently shown to be involved in perception of cytosolic flagellin and flagellated Legionella bacteria, which resulted in caspase-1 activation and macrophage apoptosis in vitro (12, 18). More work is necessary to define the role of Naip5 (and Ipaf) in the overall events of bacterial pathogenesis in vivo.
Topology of Flagellin Perception
The expression patterns of TLR5 are still somewhat controversial, bedeviled by lack of quality antibody reagents. Most workers have described immunoreactivity on the basolateral aspect of polarized cultured intestinal epithelial cells. In mice and human biopsy specimens, TLR5 protein expression is generally found basolaterally in colonic epithelia (presumably sheltered from the bacteria- and flagellin-rich environment of the colonic lumen) but is seen apically in ileal and respiratory epithelia as well on the whole peripheral plasma membrane of dendritic cells (21).
Functional responses to flagellin are also polarized. Purified flagellin or flagellated nonpathogens such as Escherichia coli applied basolaterally to polarized cultured epithelia or human mucosa ex vivo potently activates proinflammatory pathways. However, the same stimuli do not activate when applied apically, presumably because the protein cannot be translocated across the apical epithelial barrier to access basolateral receptors (19, 24). However, pathogens such as Salmonella can potently stimulate proinflammatory response when they interact with the apical surface of epithelial monolayers. Plausibly, additional signaling events between enteropathogenic Salmonella bacteria and the apical epithelium could induce transcytosis of flagellin to the basolateral aspect of the cells, or pathogen interactions may influence tight junctions and allow paracellular "leaking" of soluble flagellin. It is also possible that in vivo, epithelial basolateral TLR5 serves as a sentinel against flagellate pathogens that have gained access to the lamina propria via invasion of M cells or dendritic cell processes. Finally, epithelial cells may possess functional Ipaf and/or Naip5 and may be able to perceive intracytoplasmic flagellin from invasive pathogens such as Salmonella.
Flagellin as a Proinflammatory and Antiapoptotic Regulator
Early work has demonstrated that purified flagellin could activate the transcription and secretion of proinflammatory chemokines in in vitro epithelial cell culture systems (6, 9, 23). Subsequent experiments with cDNA microarray expression profiling in cultured epithelial cells has elucidated the full epithelial responses to this protein (24, 25). In these studies, we utilized cDNA microarrays in model epithelia that were stimulated for 1, 2, 3, 4, 5, and 6 h with purified flagellin, TNF-
, and flagellated pathogenic S. typhimurium. This time frame was chosen to evaluate the primary, largely transcriptional changes in epithelial gene expression. A striking aspect of this series of experiments was the degree to which transcriptional programs in response to S. typhimurium could be recapitulated by stimulation with purified flagellin (or, for that matter, TNF-). Most of the activated genes have recognized roles in innate immune and inflammatory reactions, and many/most of them are regulated by the NF-B pathway. These results underscore the role of acute inflammation in the control of flagellated bacterial pathogens, especially from an environmental (e.g., luminal) invader, where a rapid neutrophilic response is paramount.
It is now well established that the process of inflammation is functionally intertwined with the complex network of apoptotic activation. Many genes transcriptionally activated during inflammatory responses, including flagellin-induced responses, are now recognized to have antiapoptotic properties (10). These rapidly induced genes can inhibit biochemical events necessary for caspase activation and subsequent proapoptotic signaling. For example, IAPs are a family of ubiquitin ligases that can bind to and induce degradation of activated caspases. Several members of the Bcl2 (e.g., A1 and Bcl-xL) family are antiapoptotic at the level of cytochrome c release, thus acting as specific inhibitors of the intrinsic pathway. Cellular Fas-activated death domain (FADD)-like ICE-interacting protein (cFLIP) is a specific inhibitor the extrinsic pathway acting by blocking the interaction of FADD to procaspase 8. Thus, these gene products provide a multiply redundant "failsafe" system to control caspase activation. The activation of cytoprotective/survival factors may have evolved to protect the cells from inadvertent cellular damage from exogenous stressors and also to allow local inflammation to proceed without cell death.
Furthermore, antioxidant gene upregulation is another invariant feature of the inflammatory agonist-induced transcriptional program. These effector genes include Mn-superoxide dismutase, gluathionine-S-transferase, thioredoxins, and others. This is likely a second, complementary cytoprotective response because these proteins quench ROS generated endogenously during signaling events or exogenously as a byproduct of neutrophilic inflammation that characteristically attends acute inflammation. Our microarray data have demonstrated that flagellin-induced gene transcription in epithelial cells induces a clearly recognizable pattern of antiapoptotic and antioxidant gene upregulation.
An emerging idea in intestinal biology is that TLR signaling can be cytoprotective. "Cytoprotection" indicates a given agent or gene product can reduce apoptosis or necrotic cell death in response to a stressor. Rakoff-Nahoum et al. (16) have demonstrated that mice with intestines cleared of normal flora, and thus MAMPs, were markedly more sensitive to dextran sodium sulfate (DSS)-induced colitis and that the mucosal injury mediated by this compound could be ameliorated by the oral administration of MAMPS such as LPS and lipoteichoic acid. Additionally, this study demonstrated that these protective effects were lost in TLR2- and TLR4-null mice, implicating TLR signaling in the protective mechanism. Other investigators have found that the beneficial effects of probiotics can be mediated by isolated MAMPs; purified unmethylated probiotic DNA has been shown to ameliorate DSS-induced colitis. Furthermore, the protective effects were lost in TLR9-null mice, directly implicating TLR signaling in intestinal cytoprotection (15). Other MAMP-inducible genes, such as stromal growth factors and angiogenic factors, have reparative functions and may be protective in the sense of accelerating restitution. Interestingly, flagellin, but not other MAMPs, can induce fibroblast proliferation and cell cycle entry (8).
Experimentally, we have shown that pretreatment of model epithelia with flagellin could protect cells from subsequent proapoptotic stimuli, including staurosporine or direct bacterial contact (25), and this cytoprotective effect could be recapitulated by transfection of constitutively active TLR5 (A. S. Neish, unpublished data). Exogenous flagellin administered to live mice can also ameliorate intestinal tissue injury. It has been shown that oral or systemic pretreatment with flagellin could reduce the subsequent intestinal inflammation induced by infection with wild-type Salmonella bacteria (22) and DSS- and radiation-induced colitis (A. S. Neish and A. M. Gewirtz, unpublished observations). Taken together, it is apparent that the mucosa, either epithelial cells, immunomodulatory cells resident in the lamina propria, or both, can perceive flagellin and induce a transcriptional response eventuating in enhanced cytodefenses. Potentially, exogenous flagellin or other MAMPs could be used a therapeutic agents, hypothetically stimulating a cytoprotective response before an epithelial challenge, such as irradiation or hypoxic or chemotherapeutic insults.
Flagellin as a Proapoptotic Activator
While TLRs can clearly transduce proinflammatory signals and induce antiapoptotic and cytoprotective genes, they can also initiate proapoptotic signaling in a similar fashion to TNF receptors and other "death receptors," at least in cell culture systems. TLR2-bacterial lipoprotein interactions can activate the extrinsic pathway of apoptosis (caspase-8) via signaling involving MyD88 and the subsequent recruitment of FADD (2). TLR4-LPS interactions have been demonstrated to induce apoptotic pathways in macrophages under conditions of proteasomal blockade, and TLR4-deficient macrophages have been shown to be resistant to Yersinia-induced apoptosis (7). Similarly, it has been shown that flagellin can mediate caspase activation in cultured epithelial cells, and, when proinflammatory signaling is inhibited by pharmacological blockade of NF-B pathways or with dominant negative overexpression of IB- (thus preventing the upregulation of the cytoprotective component of the inflammatory response), flagellin can stimulate epithelial cells to undergo apoptosis (25). In a different model, utilizing in vitro bacteria-macrophage cocultures, cytosolic flagellin as well as flagellate Legionella and Salmonella bacteria were shown to potently stimulate macrophage apoptotic activation and cell death via the Ipaf-ASC-caspase-1 pathway, whereas aflagellate organisms were nearly devoid of proapoptotic potential (5, 11, 12, 18). Whether flagellin-dependant Salmonella- or Legionella-mediated macrophage apoptosis plays a role in in vivo infection remains to be established.
Intriguingly, in bacteria-epithelium coculture systems, the same aflagellate Salmonella mutant (FliC–/FljB–) that failed to activate macrophage apoptosis exhibited the opposite effect: accelerated and enhanced activation of caspases and progression to overt apoptosis, indicating that Salmonella-induced epithelial apoptosis is flagellin independent (22). These striking differences can be reconciled by noting that murine macrophages do not have TLR5 and do not respond to extracellular flagellin, rendering them unable to activate NF-B and MAPK pathways and thus the antiapoptotic/survival proteins necessary to arrest proapoptotic pathways, systems that are fully functional in epithelia. At least in epithelial cells, the cytoprotective effects of flagellin-stimulated proinflammatory pathways outweigh any proapoptotic signaling activated by flagellin. The differential susceptibility of macrophages and epithelial cells to bacterially induced apoptosis may influence bacterial pathogenesis in whole animals.
Flagellate Bacteria in Model Infection
The complexity and variability of responses to flagellated pathogens in different cell types obviously requires hypothesis testing in vivo. Using the streptomycin-treated mouse model, Barthel et al. (3) showed a marked reduction of enteric inflammation over the first day of infection with aflagellate Salmonella bacteria. However, in a previous study, whereas there was a confirmed loss or diminution of tissue inflammation at 6–12 h postinoculation, by 48 h strikingly increased clinical, gross, and histopathological evidence of local and systemic inflammation was induced by aflagellate Salmonella as well as increased epithelial apoptosis (22). These in vivo observations may reflect the increased epithelial apoptosis seen in vitro. Plausibly, aflagellate Salmonella bacteria (indeed, any invasive bacteria) are able to initiate caspase activation in cells of the epithelial barrier, but the reduced proinflammatory response (necessarily including survival gene upregulation) permits the progression to local epithelial apoptosis/damage and greater exposure of bacteria to the lamina propria. Additionally, a reduced initial epithelial proinflammatory reaction may allow bacteria to "gain a footing" in the mucosa, allowing more time for proliferation and exposure to lamina propria macrophages and resultant IL-1 release.
Molofsky et al. (12) studied the intratracheal inoculation of wild-type and aflagellate Legionella bacteria (FlaA) into C57Bl6 mice (12). The numbers of wild-type bacteria in the lung progressively fell over 3 days; in contrast, infection with aflagellate Legionella bacteria (and in Naip5-null mice) resulted in increased bacterial counts for the first 2 days. Whether increased apoptosis occurred in pulmonary cell types is not known, but these data are also consistent with the idea that flagellin-induced signaling plays an important role in the control of mucosal pathogens.
Flagellin as an Avirulence Factor in Infection: the Benefits of Inflammation
It is evident that epithelial cells exposed to a microbial challenge can ultimately respond with cellular proinflammatory activation or in programmed cell death. I propose that epithelial cells under stress by bacterial pathogens activate apoptotic pathways in parallel with proinflammatory pathways. Flagellin is apparently the dominant activator of protective pathways during Salmonella infection (and perhaps Legionella and others). In most cases, a proinflammatory upregulation of antiapoptotic effectors serves to arrest apoptotic pathways before the activation of executioner caspases and irreversible cell damage occurs (Fig. 2). In the event of inhibition of proinflammatory signaling and subsequent transcription, caspase activation could occur unimpeded and terminate in the dismantling of the cell. Such "deliberate" inhibition is now recognized to play an important role in the pathogenesis of certain bacterial infections where macrophage apoptosis has been identified (e.g., Yersinia plague and systemic anthrax). Yersinia bacteria possesses a translocated effector protein, YopJ, with well-known inhibitory effects against NF-B and MAPK survival pathways. This allows them to induce macrophage apoptosis, efficiently eliminating an immunomodulatory cell (13). Analogous events may occur in epithelial cells. Salmonella bacteria can induce apoptosis in HT-29 epithelial cell cultures after at least 24 h of coculture (14). As previously shown, aflagellate Salmonella bacteria are strikingly more potent activators of caspases and frank apoptosis (22). Salmonella bacteria possess a YopJ homolog, AvrA, which has been demonstrated to inhibit NF-B and promote apoptosis in epithelial cells (4); deletion of this protein reverses the proapoptotic effects of aflagellate Salmonella bacteria (A. S. Neish, unpublished observations). Thus, during infection with wild-type Salmonella bacteria, flagellin-induced proinflammatory signaling reduces cellular proapoptotic responses to bacterial effectors, whereas the absence of this potent proinflammatory determinant in mutants allows unimpeded apoptosis. Plausibly, "failure to signal" in naturally aflagellate pathogens may result in more severe tissue injury. For example, Shigella sp. bacteria are aflagellate and are known to induce extensive mucosal apoptosis and epithelial damage in experimental Shigella infection (17).
|
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTREFERENCES |
|---|
in salmonella-infected macrophages. Nat Immun 7: 576–582, 2006. via Ipaf. Nat Immun 7: 569–575, 2006.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
