Children’s Hospital of Pittsburgh, Pittsburgh, Pennsylvania, USA.
Address correspondence to: Jay K. Kolls, Children’s Hospital of Pittsburgh, Suite 3765, 3705 Fifth Avenue, Pittsburgh, Pennsylvania 15213, USA. Phone: (412) 648-7457; Fax: (412) 692-6645; E-mail: firstname.lastname@example.org .
Published April 3, 2006 - More info
Host genetic factors that regulate innate immunity determine susceptibility to sepsis. Disruption of nuclear factor-erythroid 2–related factor 2 (
Rajesh K. Thimmulappa, Hannah Lee, Tirumalai Rangasamy, Sekhar P. Reddy, Masayuki Yamamoto, Thomas W. Kensler, Shyam Biswal
Sepsis and sepsis syndrome are leading causes of mortality throughout the world. It is widely held that sepsis represents a dysregulated innate immune response to an offending pathogen. This immune response is often initiated via microbial products signaling through TLRs expressed on host immune cells. There is increasing evidence that this innate response can be dramatically influenced by the cellular redox state, and thus a better understanding of oxidative regulation of innate immunity could lead to new treatments for sepsis. In this issue of the JCI, Thimmulappa et al. show that nuclear factor-erythroid 2–related factor 2 (Nrf2), a member of the “cap’n’collar” family of basic region–leucine zipper transcription factors, which has previously been shown to be involved in the transcription of antioxidant gene expression in response to xenobiotic stress, is also a critical regulator of cellular oxidative stress in sepsis (see the related article beginning on page 984).
Despite decades of advances in antibiotic treatment, sepsis remains an elusive killer, with over 750,000 cases per year in North America (1) with a 40–50% mortality rate in adults. Sepsis is mediated by infectious stimuli, and many of the clinical findings of sepsis can be replicated in experimental animal models using specific bacterial products such as LPS (2). The last decade of immunological research has revolutionized how scientists understand the initiation of the innate immune response to invading pathogens. For many offending agents, the TLR family of proteins functions as the host sentinel to invading pathogens. This was first demonstrated in Drosophila melanogaster in 1996 (3), where Toll was shown to regulate the production of the antifungal molecule dorsomycin, and later in mammals, when positional cloning revealed Tlr4 to be the Lps gene product (4). An additional 10 human TLRs have been described that recognize other bacterial products as well as fungi and viruses. These receptors signal via their Toll/IL-1 receptor (TIR) domains using 4 adaptor proteins: (a) MyD88; (b) TIR domain–containing adaptor inducing IFN-β (TRIF); (c) MyD88 adaptor–like/TIR-associated protein (MAL/TIRAP); and (d) TRIF-related adaptor molecule (TRAM) (2). In the case of LPS signaling through TLR4, the MyD88-dependent pathway is critical for NF-κB activation and the production of TNF-α whereas the MyD88-independent, TRIF-dependent pathway is required for type I IFN production. Based on the fact that 10 human TLRs signal via 4 adaptors and 2 predominant kinases to subsequently regulate the expression of hundreds of genes, the innate immune response has been proposed to be shaped like an hourglass (Figure1). The top of the hourglass is wide, indicating that 10 TLR proteins recognize a variety of potential offending pathogens, then the hourglass narrows to represent a smaller number of highly conserved TLR adaptor proteins and initial kinases, and then it widens again to reflect the increased number of genes that are transcriptionally activated by NF-κB and other transcription factors (2). This notion is further supported by the fact that this signaling pathway is markedly conserved among mammalian species, and mutations in this pathway in humans that lead to defective TLR signaling are associated with the development of invasive meningococcal (5) or Legionella infections (6). The work by Thimmulappa et al. (7) in this issue of the JCI suggests that this hourglass may not be so narrow at its center, as many of the kinases active downstream of TLR signaling can be regulated through oxidant (redox)–dependent posttranslational modifications, resulting in an additional level of control of TLR signaling (Figure1).
A schematic representation of ROS- and TLR-mediated gene expression. Despite the diversity of microbes that are potential pathogens, there is precise molecular recognition of microbial products by molecules of the innate immune system, with the TLR family being 1 of the most intensely studied mediators of this recognition. There are 10 human TLRs that signal via 4 adaptor proteins and 2 initial kinases; this signaling is followed by the activation of distal kinases that subsequently regulate transcription factors such as NF-κB and activator protein 1 (AP-1), which control gene expression. Although the signaling cascade is quite narrow at its center, posttranslational modifications of kinase activity by ROS likely contribute to the diversity and intensity of gene expression after microbial activation of the innate immune system.
Redox-dependent control of TLR4 signaling stems from the fact that many of the kinases, transcription factors, and subsequent gene products induced by the TLR4 ligand, LPS, can be posttranslationally modified by ROS (Figure 1). In fact, this has been well studied in the context of NF-κB translocation to the nucleus and activation of activator protein 1 (AP-1) — 2 transcription factors that regulate gene expression after LPS stimulation of macrophages (8, 9). Due to the many sources of ROS within the cell, it has been difficult to take a reductionist approach to identify the critical regulatory factors that control oxidant production at the subcellular level. Moreover, the molecular targets of ROS are broad and pleiotropic and include proinflammatory events, such as increased production of cytokines, but also activation of antiinflammatory molecules, such as IL-10 and soluble TNF receptors (10). The study by Thimmulappa et al. shows that nuclear factor-erythroid 2–related factor 2 (Nrf2, also known as Nfe2l2), which encodes a basic region–leucine zipper transcription factor, is a key regulator of the cellular redox state in sepsis, in part by regulating levels of intracellular glutathione, a key antioxidant (7). Nrf2 activity is held in check by Kelch-like, erythroid cell–derived protein with cap’n’collar homology–associated protein 1 (Keap1), but upon cellular activation, such as oxidative stress or MAPK activation, it dissociates and translocates to the nucleus where it binds to cis-acting antioxidant response elements, which regulate the expression of antioxidant and phase II detoxification genes (11) (Figure 2). Studies of mice with a homozygous deletion of Nrf2 show that Nrf2 is critical for the hepatic induction of glutathione S-transferase (GST) and NAD(P)H:quinone oxidoreductase (NQO1) in response to phenolic antioxidants (12). Moreover, these mice develop hemolytic anemia, presumably due to increased lipid peroxidation (13). Although Nrf2 is clearly critical for regulating the antioxidant response and phase II detoxification (Figure 2) of certain xenobiotics, such as acetaminophen (14), its role in innate immunity had not been previously investigated.
Potential interactions of Nrf2 and TLR4 signaling. By regulating glutathione S-transferase (GST) and intracellular glutathione (GSH) levels, Nrf2 controls the level of ROS in the cell induced by external stressors such as xenobiotic or electrophilic stress. In this issue of the JCI, Thimmulappa et al. show that the level of ROS regulated by Nrf2 also influences TLR4 signaling at the level of IKK activation, resulting in increased nuclear translocation of NF-κB. In addition, the authors show that IRF-3–mediated gene transcription is also regulated by Nrf2; however, at what level this occurs remains to be determined. Other potential kinase targets of ROS modification include IRAK/IRAK4 and MAPKs as well as TRAF-associated NF-κB activator–binding kinase 1 (TBK1). ARE, antioxidant response element; CBP, CREB-binding protein; HO-1, heme oxygenase I; Keap1, Kelch-like erythroid cell–derived protein with cap'n'collar homology–associated protein; Maf, mammary cell–activating factor; NQO1, quinone oxidoreductase; TIRAP, TIR-associated protein; TRAM, TRIF-related adaptor molecule.
In the current study, mice deficient in Nrf2 displayed increased mortality in both a sterile (LPS administration) and a nonsterile (cecal ligation and puncture) model of sepsis (7). Moreover, Nrf2–/– mice had increased expression levels of TNF and lung injury after systemic LPS administration. Using gene expression profiling, Thimmulappa et al. demonstrated that Nrf2 regulates a number of key proinflammatory cytokines, including IL-23p19, IL-1F9, and IL-6, as well as chemokines, including CCL8, CCL6, CCL9, CCL2 (also known as MCP-1), and CXCL10. These mice also demonstrated increased expression of genes coding for proinflammatory molecules, such as TNF, IL-1β, and IL-6, after stimulation with TNF in vivo. However, it remains unclear what proportion of this dysregulated gene expression is a direct effect of Nrf2 deficiency versus a downstream effect of the dysregulation of, for example, TNF. However, some of the data suggest that Nrf2 may be acting at the level of kinase activation downstream from TLR4 (Figure 2). Fibroblasts from Nrf2–/– mice showed higher levels of inhibitor of κB kinase (IKK) activity, increased IκB-α phosphorylation, and more rapid nuclear translocation of NF-κB (7). Moreover, the effect of Nrf2 was not just relegated to the MyD88-dependent (Figure 2) pathway; Nrf2 also had effects on the MyD88-independent, TRIF-dependent pathway as evidenced by the upregulation of IFN regulatory factor 3–dependent (IRF-3–dependent) genes as well as transactivation of an IRF-3–dependent reporter gene in vitro in Nrf2–/– fibroblasts. Based on these data, it would be important to determine whether Nrf2 is regulating the upstream kinases IL-1 receptor–associated kinase (IRAK) and IRAK4, which are critical for MyD88-dependent signaling, or TNF receptor-associated factor–associated NF-κB activator–binding kinase 1 (TBK1) (Figure 2) (15), which is critical for IRF-3–mediated transcription, or if Nrf2 is asserting its regulatory effects downstream from these kinases at the level of the MAP kinases, in addition to its described effects on distal kinases, such as IKK (Figure 2). Since IL-23p19 is upregulated in Nrf2–/–mice, it will also be important to determine if Nrf2 also regulates adaptive Th1 and ThIL-17 responses (16, 17), as has been reported for Th2 immune responses (18).
In support of a role for Nrf2 as a critical regulator of antioxidant gene expression, restoration of intracellular glutathione with N-acetyl cytseine (NAC) decreased LPS- and TNF-induced NF-κB activation and reduced LPS-induced lung injury in Nrf2–/– mice (7). These data confirm the critical role of the cellular redox state in regulating innate immune responses and support the contention that the transcriptional regulation of the antioxidant response is critical in regulating the cellular response to external stressors. Thus, polymorphisms may exist in the Nrf2 gene that may identify subjects at risk for more severe sepsis. In a recent clinical trial, NAC was shown to reduce NF-κB activation as well as IL-8 secretion in patients with sepsis (19). However, in a subsequent randomized trial of 34 patients, NAC failed to improve end-organ function or microalbuminuria (20). This underlying reason for the failure of antioxidant therapy may be similar to that observed for the failure of other antiinflammatory approaches: the timing of these interventions may be too late to adequately interfere with the induction of the inflammatory cascade. Thus, early administration of antioxidant-based therapy is likely critical. A potential advantage of an antioxidant approach is that restoration of normal cellular glutathione levels should leave basal innate immunity intact. However, it remains unclear which subcellular stores of glutathione need to be restored. Guidot et al. have shown that NAC preferentially repletes cytosolic glutathione stores, but not mitochondrial stores (21). Therefore, we need to better understand redox regulation of TLR signaling at the subcellular level in order to propose a rational redox-based therapy for a disease as complex as sepsis.
The author would like to thank Bruce Beutler for his critical review of the manuscript.
Nonstandard abbreviations used: IKK, inhibitor of κΒ kinase; IRAK, IL-1 receptor–associated kinase; IRF-3, IFN regulatory factor 3; NAC, N-acetyl cysteine; Nrf2, nuclear factor-erythroid 2–related factor 2; TIR, Toll/IL-1 receptor; TRIF, TIR domain–containing adaptor inducing IFN-β.
Conflict of interest: The author has declared that no conflict of interest exists.
See the related article beginning on page 984.