QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Hayden, M. S. Articles by Ghosh, S. Search for Related Content PubMed PubMed Citation Articles by Hayden, M. S. Articles by Ghosh, S. Related Content Chromatin and Gene Expression Immunology Signal Transduction Social Bookmarking                   What's this?

REVIEW

Signaling to NF-B

Section of Immunobiology and Department of Molecular Biophysics & Biochemistry, Yale University School of Medicine, New Haven, Connecticut 06520, USA

	Abstract

Top Abstract Overview of the NF-{kappa}B... Signaling pathways to NF... Regulation of IKK Regulation of NF-{kappa}B... Perspective and summary Acknowledgments References

 
The transcription factor NF-B has been the focus of intense investigation for nearly two decades. Over this period, considerable progress has been made in determining the function and regulation of NF-B, although there are nuances in this important signaling pathway that still remain to be understood. The challenge now is to reconcile the regulatory complexity in this pathway with the complexity of responses in which NF-B family members play important roles. In this review, we provide an overview of established NF-B signaling pathways with focus on the current state of research into the mechanisms that regulate IKK activation and NF-B transcriptional activity.

[Keywords: NF-B; I-B; signal transduction; transcription factors; protein-serine-threonine]

The biological system in which NF-B plays the most important role is the immune system (for reviews, see Ghosh et al. 1998; Li and Verma 2002; Bonizzi and Karin 2004). Careful regulation of the transcriptional responses to many different stimuli is crucial to the proper functioning of the mammalian immune system. NF-B regulates the expression of cytokines, growth factors, and effector enzymes in response to ligation of many receptors involved in immunity including T-cell receptors (TCRs) and B-cell receptors (BCRs), TNFR, CD40, BAFFR, LTR, and the Toll/IL-1R family (for reviews, see Ghosh et al. 1998; Silverman and Maniatis 2001; Bonizzi and Karin 2004). NF-B also regulates the expression of genes outside of the immune system and, hence, can influence multiple aspects of normal and disease physiology. Recent work has highlighted the role of NF-B in embryonic development and in the development and physiology of tissues including mammary gland, bone, skin, and central nervous system. However, such varied biological roles for NF-B raise the intriguing question of whether one common mechanism regulates signaling to NF-B in all systems or whether discrete inputs create a diversity of transcriptional responses that are tailored to particular tissues and organs. Understanding how NF-B integrates multiple stimuli in multiple systems to generate a unified outcome suitable for specific situations is a challenge that faces researchers in this area. In keeping with the enormous progress that has been made in the study of NF-B, there has been a veritable explosion of review articles that have elegantly summarized progress in different aspects of NF-B regulation and biology (Ghosh and Karin 2002; Kucharczak et al. 2003; Ruland and Mak 2003; Ben-Neriah and Schmitz 2004; Bonizzi and Karin 2004; Chen and Greene 2004; Karin et al. 2004). Therefore, to avoid duplication, we have decided to focus in this review on a few areas of current activity. The main question that will be discussed in this review is how the different inducers activate NF-B and the mechanisms that underlie the regulation of NF-B transcriptional activity. The choice of these areas for discussion is, of course, idiosyncratic and we apologize for the narrow focus of this review. However, to help an uninitiated reader delve right into these areas of current research, we have provided a brief overview of the current state of knowledge about this transcription factor. We hope that interested readers will find a sufficiently comprehensive listing of the relevant literature in this article such that they will be able to go on and explore the biology of this fascinating transcription factor in depth.

	Overview of the NF-B pathway

Top Abstract Overview of the NF-{kappa}B... Signaling pathways to NF... Regulation of IKK Regulation of NF-{kappa}B... Perspective and summary Acknowledgments References

 
The five members of the mammalian NF-B family, p65 (RelA), RelB, c-Rel, p50/p105 (NF-B1), and p52/p100 (NF-B2), exist in unstimulated cells as homo- or heterodimers bound to IB family proteins. NF-B proteins are characterized by the presence of a conserved 300-amino acid Rel homology domain (RHD) that is located toward the N terminus of the protein and is responsible for dimerization, interaction with IBs, and binding to DNA (Fig. 1). Binding to IB prevents the NF-B:IB complex from translocating to the nucleus, thereby maintaining NF-B in an inactive state. NF-B signaling is generally considered to occur through either the classical or alternative pathway (for review, see Bonizzi and Karin 2004). In the classical pathway of NF-B activation, for example, upon stimulation by the proinflammatory cytokine tumor necrosis factor (TNF), signaling pathways lead to activation of the subunit of the IB kinase (IKK) complex, which then phosphorylates IB proteins on two N-terminal serine residues (Fig. 2A). In the alternative pathway, IKK is activated and phosphorylates p100 (Fig. 2B). Phosphorylated IBs are recognized by the ubiquitin ligase machinery, leading to their polyubiquitination and subsequent degradation, or processing in the case of p100, by the proteasome (for review, see Karin and Ben-Neriah 2000). The freed NF-B dimers translocate to the nucleus, where they bind to specific sequences in the promoter or enhancer regions of target genes. Activated NF-B can then be down-regulated through multiple mechanisms including the well-characterized feedback pathway whereby newly synthesized IB protein binds to nuclear NF-B and exports it out to the cytosol.

View larger version (42K): [in this window] [in a new window]   Figure 1. Schematic representation of NF-B, IB, and IKK proteins family of proteins. Members of the NF-B, IB, and IKK proteins families are shown. The number of amino acids in each protein is indicated on the right. Presumed sites of cleavage for p100 (amino acid 447) and p105 (amino acid 433) are shown. Phosphorylation and ubiquitination sites on p100, p105, and IB proteins are indicated. (RHD) Rel homology domain; (TAD) transactivation domain; (LZ) leucine zipper domain on IKK/ and Rel-B; (GRR) glycine-rich region; (HLH) helix-loop-helix domain; (Z) zinc finger domain; (CC1/2) coiled-coil domains; (NBD) NEMO-binding domain; () -helical domain.

View larger version (67K): [in this window] [in a new window]   Figure 2. Classical and alternative pathways of NF-B activation. A model depicting the two signaling pathways to NF-B. (A) One pathway is the classical pathway mediated by IKK and leading to phosphorylation of IB. Inputs for the classical pathway include TNFR1/2, TCR and BCR, TLR/IL-1R, and many others. (B) The alternative pathway involves NIK activation of IKK and leads to the phosphorylation and processing of p100, generating p52:RelB heterodimers. Input signals for the alternative pathway follow ligation of LTR, BAFFR, and CD40R. These alternative pathway stimuli also activate the classical pathway.

Degradation of IB is a tightly regulated event that is initiated upon specific phosphorylation by activated IKK. The IKK activity in cells can be purified as a 700-900-kDa complex, and has been shown to contain two kinase subunits, IKK (IKK1) and IKK (IKK2), and a regulatory subunit, NEMO (NF-B essential modifier) or IKK (for reviews, see Rothwarf and Karin 1999; Ghosh and Karin 2002). In the classical NF-B signaling pathway, IKK is both necessary and sufficient for phosphorylation of IB on Ser 32 and Ser 36, and IB on Ser 19 and Ser 23. The role of IKK in the classical pathway is unclear, although recent studies suggest it may regulate gene expression in the nucleus by modifying the phosphorylation status of histones. The alternative pathway, however, depends only on the IKK subunit, which functions by phosphorylating p100 and causing its inducible processing to p52 (Fig. 2B). The alternative pathway is activated in response to a subset of NF-B inducers including LT and BAFF.

Upon phosphorylation by IKKs, IB proteins are recognized and ubiquitinated by members of the Skp1-Culin-Roc1/Rbx1/Hrt-1-F-box (SCF or SCRF) family of ubiquitin ligases (for review, see Ben-Neriah 2002). TrCP (E3RS or Fbw1a), the receptor subunit of the SCF family ubiquitin ligase machinery, binds directly to the phosphorylated E3 recognition sequence (DSGXXS) on IB (Yaron et al. 1997, 1998; Fuchs et al. 1999; Hatakeyama et al. 1999; Kroll et al. 1999; Spencer et al. 1999; Suzuki et al. 1999; Winston et al. 1999; Wu and Ghosh 1999). Recognition of IB leads to polyubiquitination at conserved residues, Lys 21 and Lys 22 on IB, by the E3 SCF^-TrCP and the E2 UbcH5 (Alkalay et al. 1995; Scherer et al. 1995; DiDonato et al. 1996). Although it is commonly believed that degradation of IB is a cytoplasmic event, TrCP1 is almost exclusively nuclear with its receptor site occupied by hnRNP-U (Davis et al. 2002). This finding has led to the suggestion that phosphorylated IB must out-compete hnRNP-U for binding to TrCP1. However, TrCP2, a highly homologous isoform of TrCP1, is localized in the cytoplasm and can bind IB, although it has lower ligase efficiency than TrCP1 (Suzuki et al. 2000; Davis et al. 2002). Cells from TrCP1-deficient mice display partially decreased rates of IB and IB degradation, indicating that this TrCP1 function can be partially compensated for by TrCP2 (Nakayama et al. 2003). Determination of the actual role of -TrCP isoforms in IB degradation will have to await the generation of mice lacking both isoforms.

The precursor protein p105 undergoes constitutive processing, as opposed to degradation, via the proteasome, through a cotranslational mechanism (Fan and Maniatis 1991; Palombella et al. 1994; L. Lin et al. 1998). Limited proteolysis of the precursor protein, which generates p50, is dependent on the presence of a glycine-rich region (GRR) between amino acids 376 and 404 that serves as a stop signal for proteolysis (Lin and Ghosh 1996; Orian et al. 1999). Whether p105 can also undergo inducible processing remains contentious. Multiple reports have demonstrated IKK-dependent, and IKK-independent, phosphorylation of p105 C-terminal serines including Ser 923 and Ser 927, followed by inducible degradation, although there is no definitive evidence that this degradation of p105 has a functional role (Fujimoto et al. 1995; MacKichan et al. 1996; Heissmeyer et al. 1999, 2001; Orian et al. 2000; Salmeron et al. 2001; Lang et al. 2003; Cohen et al. 2004). It has been suggested that similar to IB, phosphorylation of p105 leads to SCF^-TrCP binding, resulting in polyubiquitination and degradation of the protein. Polyubiquitination of p105 leading to degradation involves multiple lysine residues, and is dependent on an acidic region between residues 445 and 453 (Harhaj et al. 1996; Orian et al. 1999; Cohen et al. 2004). In contrast, partial processing leading to p50 generation may be carried out by a different ubiquitin ligase machinery (Amir et al. 2002; Lang et al. 2003; Cohen et al. 2004). It was recently demonstrated that SCF^TrCP is not responsible for signal-induced processing, and it has been suggested that this event is independent of ubiquitination (Cohen et al. 2004). The p105 precursor protein forms multiple heterodimeric and homodimeric complexes (Ghosh 2004), and this association with NF-B proteins inhibits constitutive processing (Harhaj et al. 1996; Cohen et al. 2001). Most likely, processing of p105 is partially determined by the milieu of NF-B dimers present in any given tissue.

Processing of p100 bears some similarity to both IB degradation and p105 processing. The phosphorylation of p100 has, so far, only been shown to be catalyzed by IKK acting downstream from NF-B inducing kinase (NIK; Fig. 2B; Senftleben et al. 2001a). Unlike p105, p100 processing is a tightly regulated event, with only minimal constitutive processing in unstimulated cells (Heusch et al. 1999). Interestingly, it appears that NIK can act both as an IKK-activating kinase as well as a docking protein linking IKK to p100 (Xiao et al. 2004). The C-terminal death domain (DD) of p100 has been shown to function as a processing inhibitory domain (PID), and expression of a p100 lacking this domain results in increased constitutive processing in a manner dependent on nuclear shuttling (Xiao et al. 2001; Liao and Sun 2003). Similar to IB, phosphorylation of p100 leads to the recruitment of SCF^TrCP, polyubiquitination of Lys 855 in a region with sequence homology to Lys 22 of IB, and subsequent processing to p52 (Fong and Sun 2002; Amir et al. 2004; Cohen et al. 2004). Like p105, the p100 GRR is also required for partial processing yielding p52 and release of active p52:NF-B complexes (Heusch et al. 1999).

Following degradation of IB, the released NF-B is able to bind promoter and enhancer regions containing B sites with the consensus sequence GGGRNNYYCC (N = any base, R = purine, and Y = pyrimidine). The crystal structures of NF-B dimers bound to the B enhancer reveal how both immunoglobulin-like domains that make up the RHD contact the DNA of the B site. The N-terminal Ig-like domain is primarily responsible for sequence specificity of NF-B, whereas hydrophobic residues within the C-terminal domain form the dimerization interface (Ghosh et al. 1995; Muller et al. 1995; F.E. Chen et al. 1998; Y.Q. Chen et al. 1998; Huxford et al. 1998; Huang et al. 2001). RelB, c-Rel, and p65 contain a transactivation domain (TAD) located toward the C terminus that is necessary for transactivation by these proteins. Homodimers of p52 and p50 lack TADs and hence have no intrinsic ability to drive transcription. In fact, binding of p52 or p50 homodimers to B sites of resting cells leads to repression of gene expression (Zhong et al. 2002). The repressive function of p50 or p52 homodimers may provide a threshold for NF-B transactivation that can be regulated through the expression and processing of the p100 and p105 precursors. The TADs on p65, c-Rel, and RelB promote transcription by facilitating the recruitment of coactivators and the displacement of repressors. The function of TADs is enhanced through direct modifications of NF-B including phosphorylation, and represents another layer of regulation of the NF-B-mediated transcriptional response (for review, see Chen and Greene 2004).

Biological roles of NF-B and IB proteins

All of the mammalian Rel and IB family members have been knocked out in mice, and many conditional knockouts and multigene knockouts have also been generated. The phenotype of these various mice with regard to their immune system function and roles in the regulation of apoptosis has recently been reviewed in detail (Kucharczak et al. 2003; Bonizzi and Karin 2004). Therefore, we provide here a succinct overview of the function of each pathway component based on the relevant genetic experiments in support of such a role.

Genetic studies first revealed the crucial role of p65 in mediating protection from apoptosis during TNF signaling. p65^-/- mice exhibit lethality caused by liver degeneration at gestational day 15-16 (Beg et al. 1995b). The insult leading to the observed liver apoptosis was shown to be TNF signaling in the developing liver, as crossing p65^-/- mice with either TNF^-/- (Doi et al. 1999) or TNFR^-/- (Alcamo et al. 2001) mice rescues the liver phenotype and allows the study of the effects of p65 knockout on other tissues. p65/TNFR1 double knockouts have increased susceptibility to bacterial infection, highlighting the role of p65 in innate immune responses and the initiation of innate immune responses by nonhematopoietic cells (Alcamo et al. 2001). The knockouts of p65 were also used to show its importance in class switching in B cells and lymphocyte proliferation following various stimuli in chimeric mouse models (Doi et al. 1997).

In contrast to p65-deficient mice, p50 (NF-B1)-deficient mice do not show any developmental defects (Sha et al. 1995). Instead, p50^-/- mice exhibit decreased immunoglobulin production and defective humoral immune responses. B cells from these mice do not respond efficiently to LPS, underscoring the role of the classical p65/p50 heterodimer in Toll/IL-1 signaling pathways, whereas p50 is redundant for antiapoptotic pathways.

Mice lacking p52 (NF-B2) fail to develop normal B-cell follicles and germinal centers and have additional defects in their splenic architecture and Peyer's patch development (Caamano et al. 1998; Franzoso et al. 1998; Paxian et al. 2002). These mice display normal B-cell maturation and undergo class-switch recombination, but generate inadequate humoral responses to various T-cell-dependent antigens. In addition, p52 knockouts fail to generate appropriate class-switched antibodies following influenza infection. Rescue of the T-cell-dependent response by adoptively transferring p52-deficient cells into rag-1^-/- mice indicates that this deficit is not intrinsic to the B cells (Franzoso et al. 1998). Overall, the p52 knockout mice have a phenotype that is consistent with a role for p52 in LTR signaling, particularly with regard to development of secondary lymphoid organs (for reviews, see Mebius 2003; Bonizzi and Karin 2004).

RelB is unique in that it does not homodimerize, and further, is unable to heterodimerize with c-Rel or p65. RelB forms heterodimers with p100, p52, and p50 (Ryseck et al. 1992; Dobrzanski et al. 1994) and does not interact with any IB molecules other than p100 (Solan et al. 2002). Therefore, RelB/p52 heterodimers exhibit constitutive nuclear localization (Ryseck et al. 1992; Dobrzanski et al. 1993, 1994; Lernbecher et al. 1994). RelB has been implicated in constitutive NF-B activity in multiple tissues, and RelB-deficient mice have decreased baseline NF-B activity in the thymus and spleen. Interestingly, loss of RelB results in increased inflammatory infiltration in multiple organs as well as severe deficits in adaptive immunity (Weih et al. 1995). Loss of cellular immune responses in these mice is consistent with the observed deficit in thymic medullary epithelial cells and the loss of functional dendritic cells (Burkly et al. 1995). This phenotype is exaggerated when p50 is also knocked out, indicating that p50 and RelB cooperate in the regulation of genes that limit inflammation (Weih et al. 1997). relB^-/- B cells, although crippled in their proliferative response, undergo normal IgM secretion and class switching in response to various stimuli (Snapper et al. 1996). The importance of RelB in lymphoid organogenesis is evident in RelB knockouts, particularly in the development of Peyer's patches (Paxian et al. 2002; Yilmaz et al. 2003). These findings highlight the role of RelB/p52 dimers (probably in response to LTR signaling) during lymphoid organogenesis.

c-Rel forms homodimers and heterodimers with p50 and p65. c-Rel-deficient mice fail to generate a productive humoral immune response, and peripheral T and B cells fail to respond to typical mitogenic stimuli (Kontgen et al. 1995). Mice lacking the c-Rel TAD, but with an intact RHD display a more severe phenotype, with marked hyperplasia in secondary lymphoid organs and hypoplastic bone marrow, and support a role for c-Rel in B-cell class switching (Zelazowski et al. 1997; Carrasco et al. 1998).

Transcriptional specificity is partially regulated by the ability of specific NF-B dimers to preferentially associate with certain members of the IB family. For example, the classical NF-B heterodimer (p65/p50) is predominantly, although not exclusively, regulated by IB. Therefore, IB-deficient mice display increased but not constitutive p65/p50 DNA-binding activity, demonstrating that additional mechanisms regulate the activity of this dimer (Beg et al. 1995a; Klement et al. 1996). IB is known to participate in a feedback loop where newly synthesized IB inhibits the activity of NF-B. Consistent with such a role, in the absence of IB the termination of NF-B activation in response to TNF- is significantly delayed. In fact, if one places IB, which otherwise does not participate in the early down-regulation of NF-B activity, under control of the IB promoter, the normal kinetics of NF-B activation and down-regulation are restored (Cheng et al. 1998). Analysis of cell lines in which multiple isoforms of IBs have been knocked out demonstrates that the functional characteristics of IB, IB, and IB are primarily the result of temporal differences in their degradation and re-synthesis that is attributable to differential transcriptional regulation of their promoters (Hoffmann et al. 2002). Therefore, although IB is the primary regulator of p65:p50 complexes, it remains to be seen whether there are intrinsic properties of the IB protein, in addition to its transcriptional regulation, that contribute to this central function in classical NF-B signaling pathways.

The role of the other IB family members is less well understood. The phenotype of IB-deficient mice has not been published, and, consequently, the biological role of this family member remains to be clarified. Early studies suggested that BCL-3 interacts with and removes p50 and p52 homodimers from B sites, thereby removing the repressive effects of dimers lacking TADs (Hatada et al. 1992; Kerr et al. 1992; Wulczyn et al. 1992; Franzoso et al. 1993; Naumann et al. 1993). However, subsequent reports suggest BCL-3 forms transcriptionally active complexes with p52 and p50 homodimers (Bours et al. 1993; Fujita et al. 1993). Although these reports have demonstrated that the ability of BCL-3 complexes to promote transcription depends on BCL-3 phosphorylation, the upstream components of this pathway and their regulation have not been well characterized (Nolan et al. 1993). The BCL-3 knockout mouse is viable but is unable to generate an appropriate humoral immune response to multiple pathogens. These mice lack spleen germinal centers, and although they exhibit normal serum antibody levels, they fail to develop antigen-specific responses (Franzoso et al. 1997; Schwarz et al. 1997). This phenotype is partially shared by p52^-/- mice, supporting the idea that BCL-3 cooperates with p52 to form a transcriptionally active complex (Sha et al. 1995; Caamano et al. 1996; Franzoso et al. 1997; Schwarz et al. 1997; Brasier et al. 2001; Kuwata et al. 2003).

IB has a selective role in the regulation of p65 homodimers and c-Rel:p65 heterodimers (Li and Nabel 1997; Simeonidis et al. 1997; Whiteside et al. 1997). IB degradation occurs in response to many NF-B inducers, albeit with considerably slower kinetics than that for IB. Similar to IB, IB has some nuclear-cytoplasmic shuttling ability, although markedly less than IB, and hence, IB-containing complexes are predominantly cytoplasmic (Simeonidis et al. 1997; Tam et al. 2000; Lee and Hannink 2002). IB has a noncanonical NES that is necessary for the cytoplasmic localization of c-Rel and IB itself (Lee and Hannink 2002). As IB is induced in response to NF-B activation (Simeonidis et al. 1997; Whiteside et al. 1997), it appears that it may function in regulating later phases of NF-B gene activation by p65:c-Rel complexes. IB is actually the C-terminal region of mouse p105 (Inoue et al. 1992; Gerondakis et al. 1993; Grumont and Gerondakis 1994). It is still unclear whether IB, which is synthesized as an alternate mRNA using a separate promoter, has a defined biological role. Finally, IB (also called MAIL/INAP), perhaps the eighth member of the mammalian IB family, is found in the nucleus and is expressed in response to IL-1 and TLR stimulation of NF-B but not TNF (Kitamura et al. 2000; Haruta et al. 2001; Yamazaki et al. 2001; Eto et al. 2003; Muta et al. 2003). Although IB has weak homology to BCL-3 and the other IBs, there is not yet any direct evidence that it is functionally an IB family member, as it has not yet been shown to interact with any Rel family members.

	Signaling pathways to NF-B

Top Abstract Overview of the NF-{kappa}B... Signaling pathways to NF... Regulation of IKK Regulation of NF-{kappa}B... Perspective and summary Acknowledgments References

 
Numerous pathways lead to the activation of NF-B. Almost universally, these pathways proceed via activation of IKK, degradation of IB, and enhancement of the transcriptional activity of NF-B. However, there is significant variability in fundamental aspects of these pathways, for example, TNF signaling via TNFR1 results in the rapid activation of IKK and nearly complete degradation of IB within 10 min, whereas signal-induced degradation of IB through the TCR takes nearly 45 min. Furthermore, individual NF-B responses can be characterized as consisting of waves of activation and inactivation of the various NF-B family members (Hoffmann et al. 2002, 2003). This is because of sustained activation of the IKK complex, and its selectivity for different IBs, as well as the differential regulation of IB expression by NF-B dimers. The baseline complement of NF-B complexes present in a cell can therefore influence the nature of the transcriptional response to a given stimulus. The transcriptional activity and specificity of NF-B are also differentially regulated. Summation of these effects results in unique transcriptional responses to a given stimulus in a particular cell type. A careful examination of the most-studied pathways—TNFR, TLR/IL-1R, TCR, and BCR—highlights some of these differences, but also reinforces the commonalities that characterize signaling to NF-B in general (Fig. 3). The TNFR superfamily consists of at least 19 ligands and 29 receptors and exhibits a remarkable diversity in tissue distribution and physiology (for review, see Aggarwal 2003). These receptor/ligand pairs initiate a variety of biological responses, primarily through activation of inducible transcription factors such as NF-B and AP-1 (for review, see Wajant et al. 2003). TNF is probably the most widely studied member of this family of cytokines and performs multiple roles in innate and adaptive immune responses. Most notably, through activation of NF-B, TNF signaling directly regulates the expression of antiapoptotic genes such as cIAP1/2 and Bcl-XL (for review, see Kucharczak et al. 2003). If NF-B signaling is blocked, then exposure to TNF induces rapid apoptosis in most cell types. Aberrant TNF signaling has been implicated in many disease states, and anti-TNF antibodies are in clinical use for the treatment of inflammatory conditions such as rheumatoid arthritis (for review, see Aggarwal 2003). Therefore, a detailed understanding of IKK activation through this pathway may provide an opportunity to develop targeted inhibitors with the potential to treat such conditions. However, despite the tremendous clinical relevance of this pathway, and enormous effort in numerous laboratories, there are still significant gaps in our understanding of the mechanisms that underlie TNF-mediated activation of NF-B.

View larger version (95K): [in this window] [in a new window]   Figure 3. Major signaling pathways that lead to NF-B activation. Signal transduction pathways emanating from TNF receptor (A), Toll/IL-1 receptor (B), the BCR (C), and intermediary proteins involved. (Lower left) An outline of pathways leading to NF-B as a consequence of cell stress and DNA damage is also indicated. The activation of the TCR is shown using cross-linking anti-CD3/CD28 antibodies.

Multiple members of the TRAF family including TRAF2, TRAF3, and TRAF5 have been implicated in TNF signaling, however elucidating their functional roles has been complicated because of diverse homotypic and heterotypic interactions that can occur between members of this family. TRAF2 is inducibly recruited to the TNFR and interacts efficiently with TRADD (Hsu et al. 1996). Surprisingly, TRAF2-deficient mice have intact TNF signaling to NF-B; instead, TRAF2 appears to mediate signaling to AP-1 (Yeh et al. 1997). TRAF5 knockouts also exhibit normal NF-B activation by TNF; however, TRAF2/5 double knockout cells have substantially reduced TNF-induced IKK activation (Yeh et al. 1997; Nakano et al. 1999; Tada et al. 2001). Therefore, TRAF2 and TRAF5 appear to play a redundant role in TNF signaling to NF-B, although the nature of this role remains to be determined. It has been suggested that TRAF2 recruits the IKK complex following TNF treatment through a direct interaction with the leucine zipper of IKK or IKK (Devin et al. 2001). TRAF2 also interacts with the serine/threonine kinase, Receptor interacting protein 1 (RIP1), which can also be independently recruited to TRADD. RIP1 is essential for TNF-induced NF-B activation (Hsu et al. 1996), and RIP-deficient cells exhibit increased apoptosis following TNF stimulation owing to their inability to activate NF-B (Kelliher et al. 1998). Intriguingly, RIP may function as a scaffold molecule in TNF signaling because its kinase activity is dispensable for NF-B activation (Hsu et al. 1996; Ting et al. 1996; Devin et al. 2000). RIP1 can bind directly to NEMO and thereby recruit IKK to the TNFR1 signaling complex independent of TRAF2 (Zhang et al. 2000). Interestingly, IKK recruitment in the absence of RIP, presumably through TRAF2, is not sufficient for IKK activation (Devin et al. 2001), demonstrating that RIP has a role in addition to or independent of simple recruitment of the IKK complex. RIP may nucleate the assembly of a signaling complex that induces IKK activation through oligomerization of NEMO and subsequent autophosphorylation of IKK (Delhase and Karin 1999). Alternatively, the assembly of a signaling complex may facilitate NF-B signaling by bringing the IKK complex into close proximity with an IKK kinase.

An intriguing feature of signaling by TNF is its ability to stimulate both death and survival. As noted above, the balanced induction of both pathways usually leads to activation of cells; however, if activation of NF-B, and consequently the survival pathways, are blocked, TNF becomes a potent apoptosis-inducing factor. In general, direct up-regulation by NF-B of factors such as IAPs is responsible for the antiapoptotic effect of NF-B (for review, see Kucharczak et al. 2003). It appears, however, that the two main target transcription factors of TNF signaling, NF-B and AP-1, also participate directly in arbitrating the choice between survival and death. It has been observed that prolonged activation of JNK/AP-1 leads to apoptosis (Guo et al. 1998; De Smaele et al. 2001; Tang et al. 2001). NF-B induction can lead to the synthesis of certain effectors that inhibit the JNK pathway, thereby limiting JNK/AP1 activation. One such JNK-inhibiting factor is GADD45, which inhibits JNK signaling by blocking the upstream kinase MKK7 (De Smaele et al. 2001; Jin et al. 2002; Papa et al. 2004). Although a more recent study showed that activation of JNK is similar to wild-type cells in GADD45^-/- MEFs (Jin et al. 2002), it is possible that GADD45 may cooperate with some other NF-B-induced inhibitor of JNK. For example, X-chromosome-linked inhibitor of apoptosis (XIAP) is induced by NF-B, and, in addition to inhibiting multiple proapoptotic caspases, it also blocks JNK activation (Tang et al. 2001). Another NF-B-regulated gene, cFLIP (Casper), inhibits JNK at the level of TRAF/RIP (Shu et al. 1997; Kreuz et al. 2001). Furthermore, the roles of JNK and NF-B are not entirely antagonistic, as expression of the antiapoptotic protein cIAP1 is coregulated by JNK and NF-B, and consequently TNF-induced apoptosis is increased in JNK double-knockout cells (Lamb et al. 2003; Ventura et al. 2003). Interestingly, TRAF-2-deficient cells are more sensitive to TNF-induced cell death although NF-B signaling remains intact. This may reflect the requirement for TRAF-2 binding to cIAP in order for it to be localized to its site of function or the requirement of JNK-mediated JunD activation for appropriate expression of cIAP (Shu et al. 1996; Wang et al. 1998; Lamb et al. 2003).

Lymphotoxin-R, BAFF-R, and CD40: TNFR family members that signal to NF-B through the alternative pathway

The recent discovery of the alternative signaling pathway leading to inducible p100 processing has provided a novel addition to the well-studied pathway of NF-B activation through the degradation of IB proteins (for review, see Bonizzi and Karin 2004). This represents an important step forward in understanding how NF-B family members might be differentially regulated. The alternative pathway is unique in that it is independent of IKK and NEMO (Claudio et al. 2002; Dejardin et al. 2002). Instead, the functional IKK is thought to be IKK homodimers, which selectively phosphorylate p100 associated with RelB. Therefore, processing of p100 releases a subset of transcriptionally active NF-B dimers, consisting mainly of p52:RelB (Dejardin et al. 2002; Xiao et al. 2004). In contrast to the other pathways discussed here, in the alternative pathway the mechanism of IKK activation is known. NF-B inducing kinase (NIK) is responsible for directly phosphorylating and activating IKK (Senftleben et al. 2001a; Xiao et al. 2001). However, events that occur upstream of NIK are unclear. The key question concerning signaling by these stimuli is how they are channeled to NIK and IKK, even though their receptor signaling domains resemble those of other TNF family members. Because LT, BAFF, and CD40L also activate the classical pathway, it would appear that the intracellular signaling domains of these receptors possess additional sequence motifs that allow their coupling to NIK and the alternative pathway. The creation and systematic analysis of chimeras between these receptors and those that only signal through the classical pathway will probably be necessary to elucidate the mechanism by which the alternative pathway is activated.

Toll/IL-1 receptor signaling to NF-B

Signaling to NF-B mediates multiple aspects of innate and adaptive immunity (for review, see Ghosh et al. 1998; Silverman and Maniatis 2001; Bonizzi and Karin 2004). NF-B plays an essential role in early events of innate immune responses through Toll-like receptor (TLR) signaling pathways. TLRs are evolutionarily conserved Pattern recognition receptors (PRRs) that recognize conserved Pathogen-associated microbial patterns (PAMPs) present on various microbes (for review, see Janeway and Medzhitov 2002; Kopp and Medzhitov 2003; Takeda et al. 2003). The role of TLRs as arbitrators of the self-non-self decision means that they play a central role in innate immunity as well as in the initiation of the adaptive immune responses. To date, 11 mammalian TLRs have been described, and each of these signals to NF-B. These receptors have varied tissue distribution and recognize many different PAMPs including LPS, dsRNA, nonmethylated CpG DNA, and flagellin. Some members of the TLR family are also capable of heterodimerization, thereby further expanding the repertoire of molecules that are recognized. Significant progress has been made over the past couple of years in deciphering the relevant signaling pathways that operate downstream of TLRs (for review, see Barton and Medzhitov 2003). The intracellular domain of Toll-like receptors bears strong homology with the intracellular domain of the IL-1 receptor, and it is this shared Toll-IL-1R (TIR) domain that mediates interaction with downstream signaling adapters that lead to activation of three key transcription factors, NF-B, AP-1, and IRF3.

TLR signaling is initiated by the recruitment of cytosolic adapters that all share the TIR domain (for review, see Dunne and O'Neill 2003). MyD88 was the first TIR domain containing adapter protein characterized and was shown to interact with the TIR domain on TLR/IL-1R cytoplasmic tails by homotypic interaction. MyD88 is crucial for normal NF-B induction in response to IL-1, IL-18, and LPS (TLR4; Adachi et al. 1998; Kawai et al. 1999). MyD88 recruitment to TLR4 following receptor aggregation leads to recruitment of another TIR-domain-containing adapter, TIRAP or Mal (Fig. 3B). TIRAP mediates NF-B activation downstream of TLR2 and TLR4, but not IL-1R or other TLRs (Horng et al. 2002; Yamamoto et al. 2002). Currently it is not known how TIRAP selectively acts only in a subset of MyD88-mediated signaling pathways (for review, see McGettrick and O'Neill 2004). MyD88 also contains an N-terminal death domain (DD) that recruits downstream DD-containing signaling molecules such as the serine/threonine kinase IRAK (for review, see Janssens and Beyaert 2003). The exact role played by IRAK family members is somewhat enigmatic because like RIP, IRAK-1 kinase function is also dispensable for its role in TLR/IL-1 signaling (Knop and Martin 1999; X. Li et al. 1999). IRAK-4 is also required for signaling from TLR and IL-1, likely upstream of IRAK-1 (Suzuki et al. 2002). IRAK recruitment and activation are necessary to bring TRAF6 into the signaling complex, although exactly how this is achieved remains mysterious (Qian et al. 2001; Takaesu et al. 2001). TRAF6 is recruited to TLR/IL-1R and is required for MyD88-dependent activation of NF-B (Cao et al. 1996; Wesche et al. 1997; Wu and Arron 2003). TRAF6-deficient cells exhibit a complete loss of NF-B DNA binding induced by IL-1R as well as loss of the early NF-B response downstream from TLR4 (Lomaga et al. 1999).

The link between TRAF6 and the IKK complex has remained controversial. Two sets of adapters have been proposed to link TRAF6 with IKK. The first involves the Transforming growth factor activated kinase 1 (TAK1), and two associated adapter proteins TAB1 and TAB2 (Takaesu et al. 2000). The major support for this signaling module came from studies that indicated that they could be copurified with TRAF6 in cells (Wang et al. 2001). The presence of a RING finger domain in TRAF6 led to the suggestion that TRAF6 could function as a ubiquitin ligase (Deng et al. 2000). Upon stimulation, TRAF6 would ubiquitinate itself or components of the TAK1/TAB1/TAB2 complex, thereby priming them for IKK activation (Sun et al. 2004). However, as of yet there is no genetic evidence in mammalian systems supporting such a critical role for TAK1 in TLR/IL-1 signaling. Intriguingly, a recent study indicated that RNAi-mediated knock-down of TAK1 inhibited signaling from both IL-1 and TNF receptors, which was surprising because TRAF6 is not involved in the TNF pathway (Takaesu et al. 2003). One possibility is that TAK1 function is not limited to TRAF6, but rather may extend to other TRAFs that also possess RING finger domains. A more recent report has implicated TAK1 as an essential component in TCR signaling, raising the possibility that contrary to current belief, TRAF proteins are also involved in antigen-receptor signaling (Sun et al. 2004). However, the phenotypes of TAB1 and TAB2 knockouts fail to support any role for these proteins in TLR/IL1 and TNF signaling. Knockout of TAB1 and TAB2 both lead to embryonic lethality at embryonic days 17 (E17) and 12.5 (E12.5), respectively (Komatsu et al. 2002; Sanjo et al. 2003). Fibroblasts isolated from the embryos display completely normal signaling in response to TNF and IL-1 (Sanjo et al. 2003; M.S. Hayden and S. Ghosh, unpubl.). Knocking out TAK1 causes embryonic lethality at E10, thus preventing signaling studies. Therefore, although it remains possible that future development and characterization of conditionally knocked out TAK1 cells will support the proposed role in TLR/IL-1, TNF, and TCR signaling, it is clear that there are significant complexities yet to be resolved.

Besides the TAK1/TAB1/TAB2 proteins, another protein that has been reported to act as a bridge between TRAF6 and the IKK complex is ECSIT (evolutionarily conserved signaling intermediate in toll pathways). ECSIT was identified in a yeast two-hybrid screen as a binding partner of TRAF6 and was shown to be required for TLR and IL-1 signaling, but not TNF-signaling, using RNAi knock-down techniques (Kopp et al. 1999; Xiao et al. 2003). Knockout of ECSIT leads to very early embryonic lethality, thereby preventing isolation of cells suitable for signaling studies. However, the knockouts of ECSIT display phenotypes that are identical to those seen in knockouts of Bmp receptor-1, a member of the TGF-receptor family that is known to function in early development. In fact, subsequent analysis showed that ECSIT can function as a coactivator for effectors of Bmp/TGF signaling, namely, the Smad transcription factors (Xiao et al. 2003). This surprising finding raises the possibility that ECSIT may regulate both TLR and TGF/Bmp pathways, and hence may help provide an explanation for why these pathways cross-repress one another (Moustakas and Heldin 2003). It is also important to point out that the TAK/TAB proteins were also initially identified and characterized as intermediates in TGF signaling, and hence it is possible that this link between adapters in TLR and TGF signaling pathways may be more extensive than currently imagined.

The initial studies with ECSIT suggested that it might function in TLR signaling by recruiting and activating the kinase MEKK1, which is also involved in TGF signaling (Kopp et al. 1999; Zhang et al. 2003). However, knockouts of MEKK1 do not display any overt phenotype in TLR or TNF signaling, suggesting that some other kinase might be involved (Xia et al. 2000; Yujiri et al. 2000). More recently, MEKK3 has been strongly implicated in TLR signaling. MEKK3-deficient cells do not transcribe IL-6 following TLR4 or IL-1R stimulation and exhibit delayed and weak NF-B DNA binding following LPS stimulation (Huang et al. 2004). Although TRAF6 and MEKK3 inducibly associate in TLR4 signaling, the mechanisms that regulate this process are not known (Huang et al. 2004). Therefore, it is possible, although not proven, that ECSIT exerts its role in TLR signaling by somehow modulating the function of MEKK3.

Multiple TLRs are also capable of signaling in the absence of MyD88. LPS stimulation in MyD88^-/- cells results in NF-B activation with slower kinetics than from normal TLR signaling and leads to expression of only a subset of target genes (Kawai et al. 1999). TRIF (TICAM-1), a TIR-domain-containing adapter, mediates activation of NF-B in the absence of MyD88 when cells are stimulated through TLR3 and TLR4 (Oshiumi et al. 2003). TRIF expression is regulated by NF-B and is therefore induced by TLR and IL-1R signaling (Hardy et al. 2004). Studies using cells from TRIF-deficient mice demonstrate that TRIF is required for early and late NF-B responses and IRF3 responses to LPS, but not for JNK activation (M. Yamamoto et al. 2003a; Akira 2004). Reconstitution of TRIF^-/- cells with a mutant form lacking the TRAF-binding domain restores induction of IFN-responsive genes, via activation of IRF3, but not NF-B activation, indicating that TRIF is the point of divergence in signaling to NF-B and IRF3 by TLR4. Recently it has been reported that TRIF binds RIP1 and that RIP1^-/- MEFs have decreased NF-B signaling from TLR3-poly(I:C) (Meylan et al. 2004). Finally, another TIR-domain-containing adapter, TRAM (TRIF-related adapter molecule), functions upstream of TRIF in MyD88-independent signaling from TLR4. TRAM is required for IRF-3 activation and for the delayed phase of NF-B activation following TLR4 engagement. TLR4-induced IRAK activation by MyD88, however, is unaffected by the absence of TRAM (M. Yamamoto et al. 2003b). TRAM does not function in TLR3 or IL-1R signaling pathways (Fitzgerald et al. 2003; M. Yamamoto et al. 2003b).

Two divergent members of the IKK family, IKKi (IKK) and TBK1 (T2K), have been implicated downstream of TRIF in signaling to IRF-3 and IRF-7, following engagement of TLR3 and TLR4 (Hemmi et al. 2004; Sharma et al. 2003). Although initially both of these kinases were implicated in regulation of NF-B activity, recent results have cast doubt on those conclusions. Studies using knockout cells indicate that neither is required for NF-B activation by LPS or TNF (Akira 2004; Hemmi et al. 2004). Also in contrast to the study reporting the knockout of TBK1, a recent study reported that there was no discernible effect on the transcriptional activity of NF-B in TBK1 knockout cells following stimulation with multiple PAMPs (McWhirter et al. 2004). The reason for this discrepancy remains to be resolved. IKKi^-/- cells have normal induction of IRF3 following stimulation with LPS (hence TLR4), whereas TBK1^-/- cells do not (Hemmi et al. 2004). IKKi expression is regulated by NF-B, and it appears that IKKi may be constitutively active once expressed (Mercurio 2004). IKKi may facilitate CCAAA/enhancer-binding protein (C/EBP) pathways that contribute to the expression of a subset of genes that are induced by IL-1, LPS, TNF, or PMA (Kravchenko et al. 2003). Therefore, TBK1 but not IKKi mediates TLR signaling to IRF-3, likely through direct phosphorylation of IRF-3, and neither kinase appears to be directly involved in TLR signaling to NF-B.

The Toll-signaling pathway can also be negatively regulated by proteins that are induced or activated upon TLR signaling and therefore may help to limit signaling from these receptors. For example, Tollip is an adapter protein found in association with IRAK in the IL-1R signaling pathway (Burns et al. 2000; Zhang and Ghosh 2002). Tollip binds to IRAK, and to TLR2, TLR4, or IL-1R upon signaling. Activation of IRAK by MyD88 results in autophosphorylation of IRAK and phosphorylation of Tollip, causing their dissociation (Bulut et al. 2001; Zhang and Ghosh 2002). Tollip can suppress TLR/IL-1R signaling when overexpressed