Oncogene (2006) 25, 67586780

& 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00
www.nature.com/onc

REVIEW

NF-jB and the immune response

One of the primary physiological roles of nuclear factorkappa B (NF-jB) is in the immune system. In particular, NF-jB family members control the transcription of cytokines and antimicrobial effectors as well as genes that regulate cellular differentiation, survival and proliferation, thereby regulating various aspects of innate and adaptive immune responses. In addition, NF-jB also contributes to the development and survival of the cells and tissues that carry out immune responses in mammals. This review, therefore, describes the role of the NF-jB pathway in the development and functioning of the immune system. Oncogene (2006) 25, 67586780. doi:10.1038/sj.onc.1209943 Keywords: NF-kB; T-cell receptor; B-cell receptor; inammation; TLR; hematopoiesis

Introduction The discovery and characterization of the nuclear factor-kappa B (NF-kB) family of transcription factors resulted from studies in two major areas of research: immunology and cancer biology. Although the role of NF-kB in cancer biology is becoming progressively better established, historically much of our current knowledge of NF-kB resulted from efforts directed at understanding the regulation and function of the immune response. In keeping with the critical role played by NF-kB in different areas of immunology, numerous excellent reviews have been published covering the role of NF-kB in Toll-like receptor (TLR) and antigen receptor (AgR) signaling, lymphoid organogenesis and hematopoiesis (Mebius, 2003; Bonizzi and Karin, 2004; Hayden and Ghosh, 2004; Lin and Wang, 2004; Siebenlist et al., 2005; Akira et al., 2006). This review will, therefore, attempt to provide a more comprehensive, if less detailed, review of the diverse functions of NF-kB in immunology, with the goal of illuminating how it is that so much in immunology seems to revolve around this family of transcription factors.

Correspondence: Professor S Ghosh, Yale University School of Medicine, Department of Immunobiology, 300 Cedar Street, New Haven, CT 06510, USA. E-mail: [email protected]

Considered broadly, mammalian immune responses can be divided into innate and adaptive responses. The immune response begins with the host recognizing the presence of foreign pathogens, followed by responses at the cellular, tissue and organismal levels, that ultimately lead to the clearance of the pathogen. As such, immune responses can be broken down into individual signal transduction events through which changes in the extracellular environment elicit altered gene expression at the cellular level. In a remarkable number of instances, NF-kB is the transcription factor that mediates these transcriptional changes. The gene products characteristic of early events in immune responses include cytokines and other soluble factors that propagate and elaborate the initial recognition event. The activation and modulation of NF-kB is also a common target of these factors. Thus, in a surprising number of situations NF-kB mediates the critical changes that are characteristic of innate and adaptive immune responses. In mammals, the NF-kB family is composed of ve related transcription factors: p50, p52, RelA (aka p65), c-Rel and RelB (see Gilmore, 2006). These transcription factors are related through an N-terminal DNAbinding/dimerization domain, called the Rel homology domain, through which they can form homodimers and heterodimers, which bind to a variety of related target DNA sequences called kB sites to modulate gene expression. RelA, c-Rel and RelB also contain C-terminal transcription activation domains (TADs), which enable them to activate target gene expression. In contrast, p50 and p52 do not contain C-terminal transactivation domains; therefore, p50 and p52 homodimers can repress transcription unless they are bound to a protein containing a TAD, such as Bcl-3. Alternatively, p50 and p52 often form heterodimers with RelA, c-Rel or RelB and act as transcriptional activating dimers. In most cells, NF-kB complexes are inactive, residing primarily in the cytoplasm in a complex with any of the family of inhibitory IkB proteins. When the pathway is activated, the IkB protein is degraded and the NF-kB complex enters the nucleus to modulate target gene expression. In almost all cases, the common step in this activating process is mediated by an IkB kinase (IKK) complex, which phosphorylates IkB and targets it for proteasomal degradation (see Scheidereit, 2006). The IKK complex consists of two catalytically active kinases (IKKa and IKKb) and a regulatory scaffold protein, NEMO. In what is called the canonical (or classical)

NF-jB and immunology MS Hayden et al

6759

pathway, IKKb and NEMO are required for the activation of complexes such as p50/RelA, p50/c-Rel, etc., whereas IKKa is relatively dispensable. Conversely, in the non-canonical (or alternative) pathway IKKa alone controls the activation of complexes that are inhibited by the IkB protein p100. These two NF-kB pathways can be activated by overlapping but distinct sets of stimuli, and also target activation/repression of overlapping but distinct sets of target genes. One of the most conserved functions of the NF-kB signaling pathway is the regulation of the immune system; indeed, NF-kB is even the primary regulator of innate immunity in insects such as Drosophila and mosquitoes (see Minakhina and Steward, 2006). This review focuses on the vast role played by NF-kB in mammalian immunity.

Development and formation of the immune system The mammalian immune system consists of a functionally linked group of anatomically disparate tissues and cell types. The dispersed cellular components of the immune system that arise from the bone marrow receive much of the attention in immunology and the study of NF-kB has likewise focused on its role in leukocytes. However, lymphoid organs that facilitate coordination and dissemination of immune responses carried out by immune cells are also key sites of NF-kB function. Therefore, whereas this section is largely concerned with the role of NF-kB in hematopoiesis, the role of NF-kB in lymphoid organogenesis is also discussed briey. NF-kB and hematopoiesis Most cells of the immune system are subject to rapid turnover. This process requires the regulation of the competing forces of cell proliferation and cell death processes heavily inuenced by NF-kB-regulated genes. Bone marrow-derived hematopoietic cells in particular are subject to high levels of turnover and consequently are particularly sensitive to changes in rates of apoptosis or proliferation. Likewise, during immune responses immune cells selectively undergo rapid expansion that must be resolved by targeted cell death. Although the role of NF-kB in development and homeostasis of hematopoietic cells has focused largely on B-cell and T-cell maturation, it is likely that as our understanding increases of the pathways responsible for the development of natural killer (NK) cells, dendritic cells (DCs), macrophages, etc. our appreciation for the role of NF-kB in the biology of these cell types will also expand. Hematopoietic components of the immune system include cells of the lymphoid, myeloid and granulocytic lineages. These lineages give rise to T cells, B cells, monocytes, macrophages, DCs (both myeloid and lymphoid), NK cells, basophils, eosinophils, neutrophils and mast cells (Figure 1). Many cells of the body can contribute to immune responses; however, these bone marrow-derived cells are the core constituents of both the innate and adaptive immune responses. Although NF-kB generally plays a prosurvival role in these cells,

its function during hematopoiesis is far more nuanced than one might expect. In the current review, we limit our discussion to those instances where the role for NF-kB is illustrative of its broader functions in the immune system. Before delving into specic aspects of NF-kB function in hematopoiesis, it is worthwhile to discuss the shortcomings of the experimental approaches that are used. For example, embryonic lethality of RelA knockout mice prevents straightforward analysis of the hematopoietic events that are relevant to adult animals. In other instances, severe defects in lymphoid organogenesis in the absence of NF-kB make it difcult to determine whether the observed defects are intrinsic to the hematopoietic lineage or are due to alterations in the relevant organ, for example, stromal tissues, within which hematopoietic development occurs. In some instances, such as for RelA or IKKb knockouts, embryonic lethality can be rescued by deletion of the tumor necrosis factor-receptor (TNF-R) or tumor necrosis factor-alpha (TNFa), which permits analysis of hematopoiesis in these mice, but potentially distorts certain aspects of the hematopoietic pathway. Similar concerns apply to adoptive transfer experiments that can be inuenced by the cytokine milieu to which transferred cells are exposed. In each case, therefore, one must ask whether the defect exhibited by a cell lacking some component of the NF-kB pathway is relevant to the course of normal hematopoiesis or simply to the experimental system being employed. Finally, there are numerous instances where one NF-kB family member can complement the function of another member, or alternatively, where the absence of one family member impedes the function or expression of other family members. Nevertheless, despite these limitations, genetic analysis has unequivocally illustrated a key role for NF-kB in the development and survival of hematopoietic cells. NF-kB in development of innate immune cells DC development is largely dependent on canonical NF-kB complexes, although a particular subset appears to require only the non-canonical RelB containing NF-kB complexes. RelB is known to facilitate the development of DCs (Burkly et al., 1995; Weih et al., 1995); specically development of CD8a, but not CD8a , DCs (Wu et al., 1998). Conversely, double knockout studies have shown that canonical p50/RelA complexes are required for the development of both CD8a and CD8a DCs, but not other myeloid and lymphoid lineages, most likely by mediating the response of DCs to TNFa (Ouaaz et al., 2002; Abe et al., 2003). Survival of DCs in the periphery following activation tends to be short, but can be prolonged upon CD40L expression on T cells. CD40L activates both the canonical and non-canonical NF-kB pathways and hence DCs decient in both p50 and c-Rel, or DCs overexpressing a mutant super-repressor form of IkBa, demonstrate signicantly decreased survival (Ouaaz et al., 2002; Kriehuber et al., 2005).
Oncogene

NF-jB and immunology MS Hayden et al

6760
Pluripotent HSC

Myeloid Stem Cell

Lymphoid Stem Cell

Platelets Basophil Erythrocytes Progenitor Granulocyte-Monocyte Progenitor T cell B cell NK cell

Basophil

Mast Cell

Eosinophil Monocyte Neutrophil Dendritic Cells

Macrophage
Figure 1 Schematic of NF-kB in hematopoiesis. Red arrows indicate stages in which NF-kB activation is thought to contribute negatively and green arrows indicate a positive function in the development of the indicated lineages. Curved arrows indicate examples in which NF-kB contributes to the survival of the mature cell population, either in the resting state or during immune responses. Gray arrows indicate developmental events for which NF-kB plays no role or for which the role of NF-kB has not being clearly demonstrated. See the text for details.

IkBa knockout mice display robust granulocytosis (Beg et al., 1995), and suggest an antiapoptotic role for NF-kB during granulocyte development. Likewise, NF-kB has an antiapoptotic role in mature granulocytes. For example, neutrophils, which undergo daily turnover and rapidly apoptose in vitro, exhibit accelerated apoptosis as well as sensitization to proapoptotic stimuli following NF-kB inhibition. Unlike lymphocytes, which are relatively long-lived in the absence of activation, protection from apoptosis in neutrophils is more important during the inammatory response than in homeostasis. Indeed, many TLR ligands increase neutrophil survival in vitro, likely due to NF-kBmediated expression of antiapoptotic genes (Francois et al., 2005). Although neutrophils are capable of activating NF-kB in response to many pro-inammatory stimuli (McDonald, 2004), they lack p52 and RelB (McDonald et al., 1997), the very subunits that are crucial for the maintenance of long-lived lymphocytes. Thus in the case of neutrophils, NF-kB fullls its predicted role as a prosurvival and proinammatory factor. The general granulocytosis observed in IkBa knockout mice suggested that an antiapoptotic role could be broadly assigned to NF-kB in this lineage. However, chimeras generated with cells from ikba/ikbe/mice (i.e., lacking IkBa and IkBe) instead display a modest defect in both myelopoiesis and granulopoiesis of transferred cells (Goudeau et al., 2003), and a
Oncogene

pronounced defect in NK cells and lymphoid lineages (Samson et al., 2004). Therefore, elevated levels of NF-kB activity in these cells appears to exert a proapoptotic effect. Thus, it appears that the role of NF-kB in granulopoiesis is selective and cell-type-specic. Furthermore, as described for lymphocyte development (below), the requirement for individual NF-kB subunits is not uniform at different developmental stages. NF-kB in development of B and T cells As in cells of the innate immune system, NF-kB is vital for the development and function of adaptive immune cells (Siebenlist et al., 2005). Although lymphocytes may exhibit great longevity in the periphery, their selection in the bone marrow and thymus is characterized by a high rate of apoptosis. As a consequence, the antiapoptotic properties of NF-kB play a key role in lymphopoiesis. The centrality of its antiapoptotic function is supported in part by the demonstration that most of the requirements for NF-kB during T-cell development can be overcome by transgenic expression of the prototypical antiapoptotic factor Bcl-2 (Sentman et al., 1991). The necessity of NF-kB for lymphopoiesis is strikingly illustrated in human genetic diseases wherein the gene encoding NEMO is inactivated by mutation (see Courtois and Gilmore, 2006). Because the NEMO gene is located on the X-chromosome, it is usually subject to random inactivation in individual cells in females.

NF-jB and immunology MS Hayden et al

6761

However, in female patients who are heterozygous for a mutant version of NEMO all peripheral lymphocytes possess an intact NEMO gene, rather than the 50% predicted by random inactivation, suggesting that in the absence of NEMO-dependent NF-kB signaling, B and T cells fail to develop. The effects of NEMO inactivation in both mice and humans solidify the role of NF-kB in lymphopoiesis, even though the details by which NF-kB functions in this process remain obscure. NF-kB plays diverse roles in lymphocyte development that can be grouped according to when in development it functions that is, before, during or after pre-AgR signaling. Although no single NF-kB subunit knockout mouse has as severe of a phenotype as NEMO deciencies with regard to the generation of mature lymphocytes, double knockouts demonstrate that the antiapoptotic function of NF-kB is important in the maturation and survival of lymphocytes. For example, loss of both of the canonical NF-kB family members p50 and RelA, or both RelA and c-Rel, halts development early in lymphopoiesis, before expression of the pre-AgRs (Horwitz et al., 1997; Grossmann et al., 1999), suggesting that NF-kB is involved in the expression of antiapoptotic factors required for early lymphoid cell survival in response to proapoptotic stimuli (Figure 2). In fact, early CD34 bone marrow cells can activate NF-kB in response to TNFa, and in these cells NF-kB acts as a prosurvival factor (Pyatt et al., 1999). Evidence suggests that expression of the pre-AgR leads to survival signals that depend, at least in part, on NF-kB. For example, pre-T-cell receptor (pre-TCR) expression in double negative (DN; CD8CD4) thymocytes coincides with high levels of NF-kB activity, and NF-kB activity at this stage is necessary for DN survival and maturation (Figure 2). Therefore, enforced IKKb activation eliminates the requirement for TCR recombination, whereas inhibition of NF-kB by expression of an IkBa super-repressor decreases DN thymocyte maturation and survival (Voll et al., 2000). Signaling through the pre-B-cell receptor (pre-BCR) also likely induces antiapoptotic signals through NF-kB. Consequently, the reduced pre-B-cell population seen upon expression of the IkBa super-repressor in bone marrow cells can be rescued by overexpression of the antiapoptotic NF-kB target gene Bcl-XL (Feng et al., 2004; Jimi et al., 2005). However, whereas evidence points toward NF-kB-mediated production of antiapoptotic factors it remains unclear how NF-kB is activated downstream of the pre-AgR. Selection of DP (double positive; CD4 CD8 ) thymocytes depends on the ability of their TCR to recognize peptide:MHC (major histocompatibility complex) complexes. Thymocytes that express a TCR that is unable to bind MHC die in a process termed death by neglect, whereas those that bind peptide:MHC are either positively or negatively selected depending on the strength of this interaction. Thymocytes that bind selfpeptide:MHC with very high afnity are likely to be self-reactive, and hence are deleted through negative selection. Thus, only DP thymocytes that recognize

Common Bone Marrow Lymphoid Precursor Pro-B cell

Periphery Thymus

DN Thymocyte

Pre-B cell

Immature B cell

DP Thymocyte

Transitional B cell Spleen Naive B cell

SP Thymocyte

Naive T cell

Figure 2 NF-kB function during lymphopoiesis. NF-kB plays a prosurvival role in common lymphoid precursor cells that give rise to B- and T-cell lineages. B-cell development occurs in the bone marrow, where NF-kB protects pre-B cells from proapoptotic stimuli including TNFa. Signaling to NF-kB through the pre-B-cell receptor mediates survival of pre-B-cells that then undergo light chain recombination to produce a functional B-cell receptor. Expression of BCR leads to NF-kB-dependent differentiation into immature B cells. High levels of BCR signaling, that is, through recognition of self-antigen, results in negative selection through the loss of NF-kB activity. Transitional B cells exit the bone marrow and migrate to the spleen where they mature and differentiate, a process that also requires NF-kB. T-cell development occurs following migration of precursor cells into the thymus. Stimulation of NF-kB through pre-TCRa provides a prosurvival signal allowing recombination of the TCRa chain and maturation to the DP stage. Optimal signaling through the TCRa/b complex induces NF-kB-dependent survival pathways, whereas a failure to signal or high-level signaling results in death by neglect or negative selection, respectively. NF-kB activity is required for the maintenance of long-lived B and T cells.

self-peptide:MHC with an afnity that falls within a dened range are positively selected to become singlepositive T cells. Somewhat counterintuitively, it appears that NF-kB functions in both positive and negative selection of thymocytes. During negative selection, NF-kB facilitates the induction of apoptosis following high-afnity TCR ligation (Hettmann et al., 1999; Mora et al., 2001b), perhaps by facilitating expression of proapoptotic genes and the consequent sensitization to proapoptotic signals (French et al., 1996; Kishimoto et al., 1998). The role of NF-kB in positive selection of thymocytes is more in keeping with the better-established role of NF-kB as an inducer of antiapoptotic genes. Unlike in thymocytes, however, NF-kB functions as a prosurvival factor during negative selection of B cells. Immature B cells display constitutive NF-kB activity that is down-regulated following BCR ligation
Oncogene

NF-jB and immunology MS Hayden et al

6762

(Wu et al., 1996). Decreased NF-kB activity might then sensitize these cells to proapoptotic signals. Interestingly, some signaling components required for NF-kB activation in mature B and T cells can be genetically disrupted without affecting their development, suggesting that pathways leading to activation of NF-kB in developing B or T cells differ signicantly from the pathways engaged following AgR ligation in mature lymphocytes. Following positive and negative selection, DP thymocytes must make a lineage commitment and become single positive (SP) thymocytes (CD4 CD8 or CD4CD8 ), which shortly thereafter emigrate from the thymus. Analyses of kB-site luciferase transgenic reporter mice have shown that CD8 SP cells have signicantly higher levels of NF-kB activity than CD4 SP thymocytes (Voll et al., 2000). Conversely, the antiapoptotic factor Bcl-2 is more highly expressed in CD4 than CD8 cells, suggesting that CD8 SP thymocytes are more dependent on NF-kB for survival. However, during the course of the development from the DP stage to emigration from the thymus, both DP and SP lineages require NF-kB. Targeted deletion of oxed-NEMO using cd4-promoter-driven Cre recombinase expression, or overexpression of kinase dead IKKb, results in loss of mature peripheral T cells (Schmidt-Supprian et al., 2004). These data strongly suggest that NF-kB activation is required for late stages of T-cell development; however, ikkb/,tnfr1/ double knockouts, ikkb/ chimeras or cd4-Cre IKKb conditional knockouts are not defective in the production of na ve T cells (Senftleben et al., 2001b; Schmidt-Supprian et al., 2004), suggesting a requirement for NEMO but not IKKb. Immature B cells exit the bone marrow, becoming transitional B cells, and complete development into either follicular or marginal zone B cells. NF-kBregulated expression of prosurvival factors is important to these nal steps of B-cell development (Grossmann et al., 2000). Interestingly, the activation of NF-kB in late B-cell maturation is the result of signaling by both canonical and non-canonical NF-kB pathways. Thus deciency in NEMO, IKKa or IKKb decreases the numbers of mature B cells (Kaisho et al., 2001; Senftleben et al., 2001b; Pasparakis et al., 2002). Likewise, either p50/p52 or RelA/c-Rel double knockout progenitor cells are defective in their ability to mature beyond the transitional B-cell stage (Franzoso et al., 1997; Grossmann et al., 1999). A requirement for both the canonical and non-canonical NF-kB pathways may explain why deletion of p50 and p52 produces a more complete block in B-cell development than loss of RelA and c-Rel. Although both canonical and noncanonical NF-kB pathways are functional during B-cell development, recent work (Batten et al., 2000; Schiemann et al., 2001; Claudio et al., 2002) has underscored the importance of BAFF ligation in selectively activating the non-canonical NF-kB pathway and the consequent expression of antiapoptotic Bcl-2 family members in transitional B cells (see Mackay et al., 2003). Indeed, BAFF knockout mice exhibit a complete failure
Oncogene

of transitional B-cell maturation, which mirrors that seen in Bcl-XL knockout mice (Motoyama et al., 1995; Gross et al., 2001; Schiemann et al., 2001). Thus, only those knockouts that target both the canonical and noncanonical NF-kB pathways have an effect that approximates the phenotype seen in BAFF or Bcl-XL deciency.

NF-jB and lymphoid organogenesis In addition to its role in the development of cells that directly mediate immune responses, NF-kB also plays an important role in the development and function of primary and secondary lymphoid tissues. Primary (central) lymphoid organs include the bone marrow and thymus whereas secondary (peripheral) lymphoid organs include lymph nodes (LNs), Peyers patches, mucosal-associated lymphoid tissue (MALT) and the spleen. Among the primary lymphoid organs, the bone marrow remains active throughout life, whereas thymic activity dwindles with the onset of adulthood. The secondary lymphoid tissues are associated with the maintenance and activation of mature lymphocytes, and provide an environment within which the interaction of lymphocytes and other leukocytes can be carefully orchestrated. Although there is clearly a role for NF-kB in the development and regulation of bone, this role has not yet been clearly correlated with effects on hematopoiesis (Jimi and Ghosh, 2005). Careful anatomical and histological examination of NF-kB-decient mice has resulted in NF-kB being assigned an increasingly prominent role in lymphoid organogenesis. Secondary lymphoid organs The secondary lymphoid organs have highly characteristic structural features that are crucial to the development and activation of lymphocytes. Analysis of the role of NF-kB in lymphoid organogenesis in knockouts has been complicated by the necessity of interfering with the TNF response to rescue the lethality associated with NF-kB deciency. The initial events of lymphoid organogenesis involve the association of lymphotoxin (LT)a1b2-expressing hematopoietic cells and vascular cell adhesion molecule-1 (VCAM1)expressing stromal cells (for a review see Mebius, 2003). This interaction initiates a positive feedbacksignaling loop in which NF-kB plays a prominent role (Figure 3). Cytokines implicated in this signaling loop LTa1b2, RANKL (receptor activator of NF-kB ligand) and TNFa are known to activate NF-kB. Also, mediators of lymphoid organogenesis and homeostasis, such as the adhesion molecules ICAM (intercellular adhesion molecule), VCAM, PNAd (peripheral node addressin), GlyCAM-1 (glycosylation-dependent cell adhesion molecule) and MadCAM (mucosal addressin cellular adhesion molecule), cytokines including TNFa, and organogenic chemokines such as CXCL12 (GRO/ MIP-2), CXCL13 (BLC), CCL19 (ELC) and CCL21 (SLC), are regulated by NF-kB.

NF-jB and immunology MS Hayden et al

6763
VCAM-1 Hematopoietic cell 41 Chemokines

LT12

LTR

RANK

RANKL

Local stromal cell

Figure 3 NF-kB function in the early events of lymphoid organogenesis. NF-kB is a vital part of the positive feedback loop between hematopoietic and stromal cells that comprises the early events of lymphoid organogenesis. LTa1b2-expressing hematopoietic cells induce production of VCAM-1 through the canonical NF-kB pathway and chemokines through the non-canonical (IKKa-dependent) pathway in LTbR-expressing stromal cells. Stromal expression of chemokines induces the upregulation of integrins (a4b1) on hematopoietic cells resulting in increased recruitment of LTa1b2-expressing cells and signaling through stromal LTbR. RANKL stimulation of NF-kB through TRAF6 is also crucial for the upregulation of LTa1b2 in hematopoietic cells.

Lymphoid organogenesis exhibits distinct requirements for both the canonical and non-canonical NF-kB pathways. Signaling through TNF-R, LTbR and RANK activates canonical RelA-containing complexes and, hence, it is not surprising that rela//tnfr1/ double knockout mice lack Peyers patches and LNs and exhibit disorganized spleens (Alcamo et al., 2002). The requirement for RelA in development of these tissues lies with the stromal cells and is likely due to a combination of effects: regulation of apoptosis (e.g., that induced by TNF); regulation of expression of organogenic factors including VCAM and LTa1b2; and enhancement of the non-canonical p52/RelB pathway through the LTbR signaling pathway. Several lines of evidence highlight the importance of the non-canonical pathway and activation of p52/RelB complexes in LN development. Mice with a point mutation in nik (aly/aly mice) lack multiple secondary lymphoid organs (Miyawaki et al., 1994; Koike et al., 1996; Shinkura et al., 1999) and share several phenotypic similarities with lymphotoxin and IKKa single knockout animals (Mebius, 2003; Bonizzi and Karin, 2004). p52/RelB, which is activated downstream of NIK and IKKa, is thought to be the primary transcriptional mediator of several key organogenic factors including CXCL12, CXCL13, CCL19, CCL21 and MadCAM-1 (Yilmaz et al., 2003). The p52 single knockout lacks

normal B-cell follicles, germinal centers (GCs) and Peyers patch development (Caamano et al., 1998; Franzoso et al., 1998; Paxian et al., 2002); RelB is likewise also required for Peyers patch development (Yilmaz et al., 2003). Although LN development occurs in RelB knockout mice, the nodes are small at birth and are resorbed perinatally. In addition to LTbR, knockouts of RANK, which likewise signals through the noncanonical pathway, also lack peripheral LNs (Dougall et al., 1999). Splenic architecture is crucial for B-cell development as well as for the initiation and maturation of B-cell responses. The spleen is divided histologically into white and red pulp zones. Macrophages in the red pulp are responsible for destroying erythrocytes that are damaged or have reached the end of their lifespan. The white pulp is populated by splenic lymphocytes and consists of B-cell follicles and T-cell zones. Splenic architecture allows for dynamic changes, most notably in the formation of GCs, during the initiation and maturation of B-cell responses. Multiple NF-kB knockouts exhibit defects in some aspect of splenic architecture; however, as for other lymphoid organs, the analysis of splenic architecture has been complicated by defects that occur upon deletion of TNF-R used to rescue the embryonic lethality. Nevertheless, there has been considerable progress in deciphering the role of NF-kB family members in development and maintenance of splenic architecture. Mice in which RelA has been targeted for deletion exhibit aberrant segregation of B- and T-cell areas and defects in one particular macrophage population, the metallophilic marginal zone macrophages. In addition, rela//tnfr1/ spleens have a more pronounced defect in GC generation following immunization, than do tnrf1/ mice (Alcamo et al., 2002). However, it is worth emphasizing that defects observed in tnfr1/ animals may, in fact, be due to changes in RelA-dependent responses, and it is possible that the role of the RelA/canonical NF-kB pathway in the spleen is underappreciated. The importance of the non-canonical pathway in the spleen has been observed in multiple circumstances. Mice in which non-canonical pathway components, RelB, NIK or IKKa, have been inactivated demonstrate severe defects in splenic architecture, similar to that seen in ltbr/ spleens. These defects largely reect deciencies in splenic stromal cells (Miyawaki et al., 1994; Koike et al., 1996). Mice decient in the non-canonical pathway fail to segregate B-cellT-cell zones and FDC networks, and they fail to form GCs following immunization. Marginal zone macrophages, which line the border between red and white pulp areas, are also absent or disorganized in RelB, p52, NIK or IKKa knockouts (Franzoso et al., 1998; Weih et al., 2001). Some splenic defects are also attributable to effects on hematopoietic cells. For example, the presence of metallophilic marginal zone macrophages depends on p52 (Franzoso et al., 1997). Finally, knockout of the atypical IkB family member BCL-3 also leads to alterations in lymphoid architecture that are reminiscent of those seen in the absence of p52, with which BCL-3
Oncogene

NF-jB and immunology MS Hayden et al

6764

forms a transcriptionally active complex. BCL-3 knockout mice lack splenic GCs, and although they exhibit normal serum antibody levels, they fail to develop antigen-specic humoral responses (Sha et al., 1995; Caamano et al., 1996; Franzoso et al., 1997; Schwarz et al., 1997). In summary, both the canonical and non-canonical NF-kB pathways are required for the development of most secondary lymphoid organs. However, the role of the non-canonical pathway, as assessed by examining mice decient for IKKa, p52, NIK or RelB, is especially important both during organogenesis and maintenance of splenic architecture. Recent data suggest that the non-canonical pathway is also important in thymic development and organization (Burkly et al., 1995; Weih et al., 1995; Kajiura et al., 2004; Kinoshita et al., 2006). However, it is important to note that canonical NF-kB pathway function in these events may be underappreciated owing to embryonic lethality and complicated by the defects introduced by crossing them onto the tnfr/ background. Nevertheless, our understanding of noncanonical pathway function in secondary lymphoid organs is consistent with the ability of RelB-containing complexes to regulate genes encoding key organogenic chemokines and adhesion molecules that direct leukocyte trafcking. The functional consequences of defects in these processes are severe and have direct ramications for the hosts ability to mount a robust immune response. Alterations in lymphoid architecture likewise impede the initiation of the adaptive response as well as the ne-tuning of this response through processes such as B-cell afnity maturation.

have been shown to alter TLR expression under inammatory conditions (Mueller et al., 2006). However, sentinel cells of the innate immune system, particularly tissue resident DCs and macrophages, express a more complete complement of PRRs, and thus are likely to bear the largest portion of the burden in the earliest events of pathogen recognition. Toll-like receptors TLRs are evolutionarily conserved PRRs that recognize unique, essential molecules characteristic of various classes of microbes (Akira et al., 2006). The function of TLRs as arbitrators of self/non-self discrimination highlights their central role in innate immunity as well as in the initiation of the adaptive immune response. The 11 characterized mammalian TLRs have varied tissue distribution and serve as recognition receptors for pathogen-associated molecular patterns (PAMPs) present on bacteria, viruses, fungi and parasites. Perhaps due to the multimeric nature of the TLR extracellular domain (ED), which consists of multiple leucine-rich repeats (LRRs), several receptors are capable of recognizing more than one microbial molecule (Figure 4 and below). Heterodimerization of some TLRs and the use of co-receptors (e.g., CD14 and MD-2) further expand the repertoire of PAMPs recognized. As we shall see below, the ability of TLRs to distinguish between pathogen types is translated into appropriate innate and adaptive responses through the selective activation of NF-kB and other inducible transcription factors. Signicant progress has been made over the past few years in deciphering the relevant signaling pathways that operate downstream of TLRs in particular. TLR signaling to NF-kB Ligand binding to TLRs is just now beginning to be understood at the molecular level. Extracellular LRRs bind to ligand and, either through receptor oligomerization and/or induction of a conformational change across the plasma membrane, induce the recruitment/activation of adapter proteins through the Toll/IL-1 Receptor (TIR) domain. These adapters lead to the activation of canonical IKKb-dependent complexes, degradation of IkBa and IkBb, and liberation of, primarily, RelA and c-Rel containing NF-kB complexes. TLR signaling to NF-kB is divided into two pathways: those that are MyD88 (myeloid differentiation primary response gene 88)-dependent and those that are MyD88-independent (Figure 5). We will base our discussion primarily on signaling events emanating from TLR4, which despite having the most complex downstream pathways is the most thoroughly studied TLR. Clear differences exist in signaling from other TLRs as noted throughout our discussion, and it is likely that further specializations will become apparent as individual TLR signaling pathways are investigated more thoroughly. MyD88-dependent signaling to NF-kB. TLR4 signaling is relatively unique amongst TLRs in that the effector adaptors are one step removed from the receptor. For

Role of NF-jB in the innate response Pattern recognition receptors To activate an appropriate immune response, the host must rst recognize the presence of pathogens. This discrimination between self and non-self is an absolute requirement for the initiation of effector functions, such as the secretion of cytokines and antimicrobial peptides, carried out by the cells of the innate immune system. A number of pattern recognition receptors (PRRs) have evolved to recognize microbial invaders. These PPRs include TLRs, members of the CATERPILLAR/NOD family of cytoplasmic receptors, scavenger receptors and the complement system. Although epithelial cells are frequently the rst to encounter pathogens, they are also constantly exposed to non-pathogenic microbes. Therefore, whereas a variety of TLRs are differentially expressed in epidermis, gut, pulmonary, urinary and reproductive epithelium, in many cases it is thought that both TLR expression and responsiveness is tightly controlled in these cells. For example, keratinocytes upregulate TLRs expression and responsiveness following transforming growth factor-alpha (TGFa) exposure (Miller et al., 2005); renal epithelial cells increase expression of TLR2 and TLR4 in response to IFNg or TNFa (Wolfs et al., 2002) and intestinal epithelial cells
Oncogene

NF-jB and immunology MS Hayden et al

6765
Lipopolysaccharide (LPS) Glycoinositolphospholipids Glucuronoxylomannan Mannan

Triacyl Lipopeptides Peptidoglycan (PGN)

Lipoteichoic Acid (LTA) Diacyl Lipopeptides Yeast Zymosan

Flagellin

TLR-2 TLR-6 TLR-1 TLR-2 Gamma-D-glutamyl-mesodiaminopimelic acid (iE-DAP) Muramyl Dipeptide (MDP) dsRNA dsRNA

TLR-4

TLR-5

TLR-10 TLR-12

NOD1

dsRNA ssRNA

PFTG UPEC CpG-DNA Hemozoin

TLR-11

NOD2 TLR-3

RIG-I

MDA5

TLR-7

TLR-8

TLR-9

Figure 4 PRRs that signal to NF-kB and their cognate ligands. TLRs 3, 7, 8, 9 and 11 have been reported to exhibit endosomal or intracellular localization whereas NOD1, NOD2, RIG-I and MDA5 function in the cytoplasm.

example, MyD88 recruitment to the receptor complex depends upon the TIR-domain containing adapter protein (TIRAP, also known as Mal) (Fitzgerald et al., 2001; Horng et al., 2002; Yamamoto et al., 2002). TLR2 also requires TIRAP to bridge MyD88 to the receptor; however it is believed that other MyD88-utilizing TLRs directly recruit MyD88. Recent reports suggest that the requirement for these intermediatory adapters is related to localization of the TLRs to certain domains in the plasma membrane (Kagan and Medzhitov, 2006; Rowe et al., 2006). The N-terminal domain of MyD88 contains a death domain (DD) that recruits the DDcontaining serine/threonine kinase interleukin-1associated kinase-4 (IRAK-4). IRAK-4 and IRAK-1 form an active complex capable of recruiting the TNF receptor-associated factor TRAF6 (Figure 5a). The link between TRAF6 and the IKK complex remains somewhat enigmatic, although a few key players are known. The kinase TAK1 (TGFb-activated kinase-1) is required for NF-kB, as well as AP-1 and extracellular signalrelated kinase (ERK), activation downstream of MyD88 (Sato et al., 2005; Shim et al., 2005). Although it is widely accepted that ubiquitination is a key switch at this crucial step of NF-kB activation, considerable work at the molecular level remains to be done to understand how ubiquitination leads to activation. In addition to TAK1, another protein, termed ECSIT (evolutionarily conserved signaling intermediate in Toll

pathways), was identied because of its interaction with TRAF6. ECSIT binds to TRAF6 and is required for TLR and interleulin-1 (IL-1) signaling, but not TNFsignaling (Kopp et al., 1999; Xiao et al., 2003). Although these studies suggested that ECSIT functions by recruiting and activating the kinase MEKK1 (mitogen activated protein kinase or ERK kinase (MEK) kinase 1) (Kopp et al., 1999; Xiao et al., 2003), the role of MEKK1 in TLR signaling remains unclear (Xia et al., 2000; Yujiri et al., 2000). MEKK3-decient cells, however, do not transcribe IL-6 following TLR4 or IL-1R stimulation and exhibit delayed and weakened NF-kB DNA binding following lipopolysaccharide (LPS) stimulation (Huang et al., 2004). ECSIT interacts with both TAK1 and MEKK3 (AP West and S Ghosh, unpublished observations) and it is, therefore, possible that ECSIT exerts its role in TLR signaling by modulating the function of TAK1 and/or MEKK3. TRIF-dependent signaling to NF-kB Somewhat unexpectedly, when exposed to LPS, MyD88/ cells display partial NF-kB activation, albeit with slower kinetics than in wild-type cells (Kawai et al., 1999). When cells are stimulated through TLR3 and TLR4, TRIF (TICAM-1), a TIR-domain containing adapter, mediates activation of NF-kB in the absence of MyD88 (Oshiumi et al., 2003). Furthermore, in the case
Oncogene

NF-jB and immunology MS Hayden et al

6766

LPS TLR4

a
MyD88

TRAM

TIRAP

b c
TRIF dsRNA

RIGI
IPS-1

TAK1

Figure 5 PRR signaling to NF-kB. Signaling through LPS/TLR4 via the MyD88-dependent (b) and TRIF-dependent (a) pathways converge on IKK activation through TRAFs. Signaling through dsRNA/RIG-I (c) proceeds through ISP1 to IKKi/TBK1 and through RIP1 to IKK. Signaling from NOD to NF-kB (d) is thought to involve oligomerization of RIP2 and activation of IKK through induced proximity. See the text for details.

of TLR3, all downstream signaling appears to be TRIFdependent. In TLR4 signaling, TRIF is required for late-phase NF-kB and IRF3 responses, but is not required for activation of JNK (Yamamoto et al., 2003). TRIF signaling to NF-kB and IRF3 also appear to be separately regulated (Figure 5b). Signaling to IRF3 occurs through two divergent members of the IKK family, IKKi (IKKe) and TBK1 (T2K) (Fitzgerald et al., 2003; Sharma et al., 2003); however, neither kinase is required for NF-kB activation by LPS or TNFa (Hemmi et al., 2004; McWhirter et al., 2004). Furthermore, reconstitution of trif/ cells with mutant TRIF lacking the TRAF-binding domain selectively restores induction of IRF3 but not NF-kB. Increasingly, therefore, it appears that the events leading from TRIF to IKK activation share a common set of intermediates as seen in other NF-kB activation pathways. TRIF interacts with receptor interacting protein (RIP)1 and RIP3 through their RIP homotypic interaction motif (RHIM), and rip1/ embryonic broblasts have decreased NF-kB activation following TLR3poly(I:C) signaling (Meylan et al., 2004). Finally, another TIR-domain containing adapter TRAM
Oncogene

IRAK1

TRAF6

IRAK4
Ub

IKKi

FADD

d
NOD2
Ub

RIP1

TBK1 MDP

ECSIT P
Ub NEMO

TRAF6 RIP1

IRF3 NEMO
Ub

RIP2
P

PP NF- B

Gene Transcription

(TRIF-related adapter molecule) functions upstream of TRIF in MyD88-independent signaling from TLR4. TRAM is required for IRF3 activation and for the delayed phase of NF-kB activation following TLR4 engagement. TLR4-induced IRAK activation by MyD88, however, is unaffected by the absence of TRAM and TRAM does not function in TLR3 TRIFdependent signaling pathways (Fitzgerald et al., 2003; Yamamoto et al., 2003). Therefore, it appears that TRAM is only needed for TRIF signaling downstream of TLR4. Adding further complexity, it has recently been suggested that TLR4, but not TLR3, TRIFdependent NF-kB activation is largely due to IRF3induced TNFa rather than to direct signaling to IKK (Covert et al., 2005; Werner et al., 2005). Although the applicability of these nding to other cell types is, as of yet, unclear, these results may be explained by differences in the recruitment of TRIF to the receptor; that is, by TRAM in the case of TLR4 versus directly to TLR3, resulting in changes in the availability of TRAF binding site or the availability of additional signaling intermediates at distinct subcellular localizations.

NF-jB and immunology MS Hayden et al

6767

Negative regulation of TLR signaling Inammatory responses are built upon waves of cytokine production and positive feedback mechanisms. As a result, tight control must be placed on the initiation and maintenance of these responses. Multiple negative feedback loops have been described that involve proteins that are induced or activated upon TLR signaling. In a number of instances, the target of these regulatory mechanisms is the IRAK family of proteins. For example, IRAK-M (IRAK3) inhibits signaling to TRAF6 by xing IRAK-1/4 to the TLR/MyD88 signaling complex; irakm/ knockouts exhibit enhanced signaling to NF-kB (Kobayashi et al., 2002). Tollip, an adapter protein constitutively associated with IRAK, is phosphorylated and dissociates following IRAK4 activation (Burns et al., 2000; Zhang and Ghosh, 2002). Negative regulation of TLR signaling by Tollip in the intestinal epithelium may prevent inammatory responses to commensal bacteria (Melmed et al., 2003). However, Tollip-decient cells demonstrate only minor defects in the production of NF-kB-regulated cytokines (Didierlaurent et al., 2006). Therefore, it is unclear whether Tollip indeed functions as initially thought. SIGIRR (TIR8), a member of the IL-1R family, binds to Toll/IL-1 receptors, IRAK and TRAF6 and may also function by inhibiting the association of IRAK with TLRs (Thomassen et al., 1999; Wald et al., 2003). SIGIRR deciency yields prolonged activation of NF-kB by Toll/IL-1 stimulation consistent with a regulatory function. Interestingly, SIGIRR is also highly expressed in epithelial cells, suggesting that it too may suppress signaling at sites of constitutive microbial exposure. Finally, suppressor of cytokine signaling-1 (SOCS-1) has been reported to negatively regulate LPS signaling to NF-kB and socs1/ mice exhibit an inammatory phenotype that is consistent with this prediction (Kinjyo et al., 2002; Nakagawa et al., 2002). SOCS-1 may function by directly targeting TIRAP/Mal, and selectively inhibit TLR4 signaling through the TIRAP/Mal/MyD88 pathway (Mansell et al., 2006). In addition to these TLR-specic regulators of signaling to NF-kB, other proteins function to control the extent and duration of NF-kB activation. These factors both set thresholds for activation and help to prevent uncontrolled, and potentially deleterious, innate immune responses. The broad array of PAMPs recognized by the TLR system affords the host the ability to mount responses against many pathogens. Nevertheless, for some pathogens, TLRs alone are not sufcient, and some physical spaces, most notably the cytosol, are not effectively monitored by TLRs. Cytoplasmic PRRs that activate NF-kB PRRs that recognize bacterial PAMPs are expressed at the plasma membrane or with LRRs projecting into the lumen of vesicles that are topologically related to the extracellular space. However, in such a system, intracellular pathogens are uniquely protected from detection. Furthermore, viral infection and the resulting induction of interferon occurs in many cells that do not

express the full panoply of antiviral TLRs suggesting that other PRRs must be at work. In fact, cells do have at their disposal families of cytoplasmic PRRs that are capable of activating NF-kB and other transcriptional mediators of the innate immune response. Interestingly, many of these PRRs contain caspase activation and recruitment domains (CARDs) that are required for activation of NF-kB following ligand binding. Here, we provide a brief description of two classes of cytoplasmic PRRs CARD-containing members of the CATEPILLAR and DExD/H-box helicase families. CATEPILLER-NODs. Nucleotide oligomerization domain proteins (NOD) 1 and 2 are part of a large family termed the CATEPILLER family, which is named for CARD, transcription enhancer, R (purine)binding, pyrin, lots of leucine repeats (Figure 4). The NOD-LRR subfamily is typied by the presence of LRRs and nucleotide oligomerization domains. NOD1, NOD2 and IPAF have CARDs and can signal to NFkB (for a review see Inohara and Nunez, 2003). NOD1 recognizes a peptidoglycan containing meso-diaminopimelic acid (meso-DAP) and induces NF-kB through a canonical pathway that includes activation of IKKb. NOD2 recognizes muramyl dipeptide, a ubiquitous component of nearly all bacterial cell walls. Relatively, few signaling intermediates downstream of NOD-LRRs are known; however, there is growing evidence that the CARD-containing kinase RIP2 (RICK) is required for NF-kB activation. Intriguingly, the ATP-binding cassette of both NOD1 and NOD2 is needed for signaling (Tanabe et al., 2004). RIP2 binds to NEMO and therefore is thought to directly mediate activation of the IKK complex by induced proximity (Inohara et al., 2000). In this model, ligand-dependent oligomerization of NOD-LRRs, which is dependent on the ATP-binding cassette, leads to a scaffold containing multiple RIP2 molecules, which allows trans-autophosphorylation of neighboring IKK complexes (Figure 5d). Retinoic acid inducible gene I and melanoma differentiation-associated gene 5 Two members of the DExD/H-box RNA helicase family stand out because of the presence of N-terminal CARD domains. Retinoic acid inducible gene I (RIG-I) and melanoma differentiation-associated gene 5 (MDA5) are RNA helicase-containing cytoplasmic proteins. The RNA helicase domains of RIG-I and MDA5 bind directly to double-stranded RNA (dsRNA) and induce production of type I interferons (Kang et al., 2002; Andrejeva et al., 2004; Yoneyama et al., 2004). Upon binding to dsRNA, representing either the viral genome or viral replication intermediate, RIG-I and MDA5 induce the activation of IRF3 and NF-kB. Interestingly, initiation of these signaling cascade is abrogated by point mutations in the Walker-type ATP-binding site, suggesting that their ATPase activity is required for signaling (Yoneyama et al., 2004). The link between these two proteins and NF-kB remains somewhat unclear (Figure 5c); however overexpression of the
Oncogene

NF-jB and immunology MS Hayden et al

6768

N-terminal CARD domain alone is sufcient to induce signaling. Recently, a CARD-containing protein, variably named CARDIF, IPS1, MAVS and VISA, has been implicated downstream of RIG-I; however, the link between this protein and IKKb is unclear (Kawai et al., 2005; Meylan et al., 2005; Seth et al., 2005; Xu et al., 2005). It appears, however, that there are similarities to TRIF mediated signaling, in that RIG-I activation of NF-kB requires FADD and RIP-1 (Balachandran et al., 2004; Yoneyama et al., 2005). Recently, it was shown that RIG-I and MDA5 differentially recognize various groups of RNA viruses and are thus critical for a robust antiviral response (Kato et al., 2006).

Pathogen recognition in innate immunity Bacterial recognition Pathogens recognized by PRRs can be categorized as bacterial, viral or eukaryotic. In each of these categories, PAMPs have been described that more or less t with existing hypotheses of how pathogen recognition by the innate immune system should occur (Janeway, 1989). Both in terms of accessibility and uniqueness to prokaryotes, the bacterial cell well is a logical source of PAMPs for TLRs and other PRRs. LPS was originally thought to be the ligand for TLR2, but subsequent studies revealed that contaminating bacterial lipoprotein in LPS preparations is the actual ligand (Wetzler, 2003). TLR2 also mediates responses to several Gram-positive bacterial cell wall components as well as Staphylococcus aureus peptidoglycan (Takeuchi et al., 2000). Additional work has shown that TLR2 is involved in the recognition of a wide range of microbial products and generally functions as a heterodimer with either TLR1 or TLR6 (Ozinsky et al., 2000; Wyllie et al., 2000). The TLR2/ TLR1 heterodimer recognizes a variety of lipoproteins, including those from mycobacteria and meningococci (Takeuchi et al., 2002; Wetzler, 2003), whereas the TLR2/TLR6 complex recognizes mycoplasma lipoproteins and peptidoglycan (Takeuchi et al., 2001). Recent reports have demonstrated that triacylated lipoproteins from bacteria are preferentially recognized by the TLR1/TLR2 complex, whereas diacylated lipoproteins are recognized by the TLR2/TLR6 complex (Takeuchi et al., 2002). However, additional TLR2 ligands do not require TLR1 or TLR6 for signaling, implying that TLR2 recognizes some ligands as a homodimer or heterodimer with other non-TLR molecules. Such TLR2 ligands include the Gram-positive cell wall component lipoteichoic acid; the mycobacterial cell wall component lipoarabinomannan; atypical LPS produced by Legionella, Leptospira interrogans, Porphyromonas gingivitis and Bordetella; and porins present in the outer membrane of Neisseria (Massari et al., 2003; Wetzler, 2003). The TLR4 ligand LPS, a glycolipid component of the outer membrane of Gram-negative bacteria, is the most
Oncogene

thoroughly studied and the most potent TLR ligand known. Trace amounts of LPS activate the innate immune system via TLR4, leading to the production of numerous proinammatory mediators, such as TNFa, IL-1 and IL-6. TLR4-mediated responses to LPS require CD14 and MD-2. Other bacterial TLR4 ligands include Lipid A analogs (Lien et al., 2001) and mycobacterial components (Means et al., 1999). TLR5 recognizes agellin, a protein component of Gram-negative bacterial agella and virulence factor for multiple human pathogens (Hayashi et al., 2001). In light of the fact that it was thought that proteins would be too mutable to serve as PAMPs, it is notable that TLR5 recognizes a highly conserved, central core structure of agellin that is essential for protolament assembly (Smith et al., 2003). Interestingly, the TLR5 recognition site is masked in the lamentous agellar structure, thus indicating that TLR5 recognizes only monomeric agellin (Smith et al., 2003). Furthermore, agellin appears to bind directly to TLR5 at residues 386407, as TLR5 mutants lacking this domain are unable to interact with agellin in biochemical assays (Mizel et al., 2003). Recent articles have demonstrated TLR5-independent recognition of cytosolic Salmonella typhimurium agellin via Ipaf, a member of the NODLRR family (Franchi et al., 2006; Miao et al., 2006). Ipaf-mediated recognition of cytosolic agellin induces caspase-1 activation and subsequent IL-1b secretion by macrophages. TLR11 recognizes a protein PAMP that is present on uropathogenic Escherichia coli (Zhang et al., 2004). Although the identity of this ligand is unknown, its ability to stimulate in a TLR11-dependent manner is destroyed by proteinase treatment. Conserved differences in bacterial nucleic acid structures can also be recognized by the innate immune system. TLR9 recognizes bacterial DNA containing unmethylated CpG motifs, and TLR9-decient mice are not responsive to CpG DNA challenge (Hemmi et al., 2000). The low frequency and high rate of methylation of CpG motifs prevent recognition of mammalian DNA by TLR9 under physiological circumstances. A recent report indicated that the intracellular, endosomal restriction of TLR9 is critical for discriminating between self and nonself DNA, as host DNA, unlike microbial DNA, does not usually enter the endosomal compartment (Barton et al., 2006). Viral recognition Although viruses are composed entirely of host products they, nevertheless, have unique components that readily serve as PAMPs. Nucleic acids are also key viral PAMPs, and are recognized by TLRs 3, 7, 8 and 9, as well as by cytoplasmic receptors of the RIG family (as described above). TLR3 recognizes dsRNA, a common viral replicative intermediate (Alexopoulou et al., 2001). TLR3 signaling results in the activation of NF-kB and IRF3, ultimately leading to the production of antiviral molecules, such as type I interferons (IFN-a/b) (Alexopoulou et al., 2001). The importance of RIG-Iand MDA5-mediated viral recognition is further

NF-jB and immunology MS Hayden et al

6769

supported by gene-targeting experiments demonstrating that TLR3 and its adaptor TRIF are not required for type I IFN production in some virally infected cells, such as broblasts and conventional DCs (Honda et al., 2003). However, plasmacytoid DCs exclusively utilize TLR3/TRIF signaling for type I IFN production in response to RNA viruses and poly(I:C) (Kato et al., 2005). Although initially found to recognize synthetic antiviral compounds, namely imidazoquinolines, the azoquinoline R-848 and loxoribine (Hemmi et al., 2002; Jurk et al., 2002), TLR7 and human TLR8 are now known to recognize guanosine- or uridine-rich singlestranded RNA derived from RNA viruses (Diebold et al., 2004; Heil et al., 2004; Lund et al., 2004). Interestingly, mammalian RNA, which contains many modied nucleosides, is signicantly less stimulatory via TLRs 7 and 8 than bacterial RNA, suggesting that nucleoside modication allows mammals to distinguish between endogenous and pathogen-derived RNA (Kariko et al., 2005). Similar to TLR3, engagement of these receptors leads to the production of type I IFNs. TLR9 recognizes viral CpG sequences and induces the induction of IFN-a (Takeshita et al., 2001; Lund et al., 2003; Krug et al., 2004). However, as membrane restriction prevents TLRs from sampling the cytosol where much of the viral life cycle occurs, cytosolic PRRs provide comprehensive innate immune recognition. For example, recognition of cytoplasmic dsDNA leading to NF-kB activation and type I interferon production has also been reported, although the relevant receptor has not yet been identied (Ishii et al., 2006; Stetson and Medzhitov, 2006). This receptor(s) is predicted to be important for type I IFN production in response to viruses and intracellular pathogens, such as Listeria monocytogenes and Shigella exneri. Finally, there have been some reports suggesting that certain viral proteins function as PAMPs. For example, TLR4 may recognize respiratory syncytial virus (RSV) F protein (Kurt-Jones et al., 2000). Recognition of other pathogens MyD88-decient cells demonstrate that many fungal species are capable of activating TLR pathways, although the receptors have not always been identied. TLR4 has been shown to recognize Aspergillus hyphae (Mambula et al., 2002), and Cryptococcus neoformans capsular polysaccharide (Shoham et al., 2001). TLR2 and TLR6 are required for recognition of yeast zymosan, whereas TLR4 is thought to recognize certain yeast mannans (for a review see Levitz, 2004). The identication of parasite PAMPs has been more elusive, and their existence is somewhat controversial. However, TLR2 heterodimers reportedly recognize various parasite GPI-anchored proteins and glycoinositolphospholipids from the parasitic protozoa Trypanosoma cruzi (Campos et al., 2001). Some TLR knockout mice have been shown to have variable defects in their ability to defend against various parasites (for a review see Gazzinelli et al., 2004). Recently, TLR9 has been

reported to recognize the malarial pigment hemozoin, a byproduct of heme metabolism in infected erythrocytes (Coban et al., 2005) whereas TLR11 recognizes a prolin-like protein that is conserved in apicomplexan parasites including Toxoplasma gondii (Yarovinsky et al., 2005). Immediate antimicrobial responses PRRs initiate a complex series of events following exposure to certain microbial components: the rst is the mounting of immediate antimicrobial responses at the cellular level. This is an effective and evolutionarily conserved function of PRRs, and one in which NF-kB has an important role. The liberation of products with direct antimicrobial activity occurs early at sites of pathogen entry. TLR ligation is at least partly responsible for the NF-kB-dependent expression of defensins cationic peptides that exert direct bactericidal activity by inducing membrane permeabilization. Small intestinal Paneth cells, for example, release large amounts of a-defensins into the intestinal lumen following exposure to a variety of bacteria/bacterial products (Ayabe et al., 2000). The production of antimicrobial nitrogen and oxygen species, which are acutely toxic to a variety of microbes, augments the activity of antimicrobial peptides. Production of nitric oxide is mediated in part by inducible nitric oxide synthase (iNOS), which is partially regulated by NF-kB. Consequently, iNOS production results from TLR or NOD-LRR ligation by PAMPs. Much of the early innate response has been demonstrated to depend on the canonical NF-kB pathway. Thus, rela//tnfr1/ double knockout mice have increased susceptibility to bacterial infection (Alcamo et al., 2001). Likewise, B cells from p50/ mice do not respond efciently to LPS, emphasizing the importance of p50-containing complexes, that is, p50/RelA, p50/ p50/BCL-3 and p50/c-Rel, in TLR signaling (Sha et al., 1995). As might be expected, TNFR/IKKb double knockouts show a more pronounced defect in innate responses owing to the more complete block in canonical NF-kB pathways, and succumb to infection more rapidly than rela//tnfr1/ mice (Li et al., 1999a, b; Senftleben et al., 2001b). Furthermore, MEFs from nemo/ mice do not exhibit NF-kB activation by LPS or IL-1 (Rudolph et al., 2000). Therefore, activation of NF-kB responsive genes by the innate immune system depends on NEMO and likely progresses through the canonical NF-kB signaling pathway. Inammation There is a staggering amount of literature that correlates NF-kB activation with inammation in a wide array of diseases and animal models. There are, likewise, numerous studies using gene targeting and inhibitors of NF-kB that have established that NF-kB plays a causative role in inammatory processes. We have already discussed the role of NF-kB in the survival of
Oncogene

NF-jB and immunology MS Hayden et al

6770

leukocytes, and how this role is particularly important during the responses that include inammation. Here, we briey discuss a few of the additional ways in which NF-kB regulates inammation. Inammation begins with epithelial or stromal cells of the infected tissue or tissue resident hematopoietic cells such as mast cells or DCs recognizing an inammatory stimulus and propagating proinammatory signals. These signals lead to the recruitment and activation of effector cells, initially neutrophils and later macrophages and other leukocytes, resulting in the tissue changes characteristic of inammation rubor, calor, dolor and tumor (redness, heat, pain and swelling, respectively). As for the immediate antimicrobial products discussed above, NF-kB is responsible for the transcription of the genes encoding many proinammatory cytokines and chemokines. One important early target of these effectors is the vascular endothelium. Changes in vascular endothelial cells both recruit circulating leukocytes and provide them with a means of exiting the vasculature into the infected tissue. NF-kB regulates the expression of adhesion molecules, both on leukocytes and endothelial cells, which allow the extravasation of leukocytes from the circulation to the site of infection (Eck et al., 1993). Indeed, RelA-decient mice display a severe defect in the recruitment of circulating leukocytes to sites of inammation (Alcamo et al., 2001). Recruited neutrophils are the key mediators of local inammation and NF-kB is important for the survival of these cells, which must function in relatively toxic conditions (Ward et al., 1999). NF-kB is important for the production of the enzymes that generate prostaglandins and reactive oxygen species (e.g., iNOS and Cox, both NF-kB target genes) and may, furthermore, be involved in the signaling induced by prostaglandins (Poligone and Baldwin, 2001; Catley et al., 2003). NF-kB has also been implicated in the response to leukotrienes, which like prostaglandins are short-lived paracrine effectors, although it is unclear whether this represents a direct signaling event. Finally, matrix metalloproteinases (MMPs) also are crucial mediators of local inammation and leukocyte chemotaxis; and their expression is also regulated by NF-kB (Vincenti et al., 1998; Vincenti and Brinckerhoff, 2002; Lai et al., 2003). The pathway from pathogen recognition to proinammatory cytokine production demonstrates a particular reliance on NF-kB. The immediate targets of NF-kB-dependent proinammatory cytokines, such as TNFa, tend to be receptors that, in turn, activate NF-kB. Therefore, NF-kB is crucial to the propagation and elaboration of cytokine responses. TNFa is particularly important for both local and systemic inammation, and it is a potent and well-studied inducer of NF-kB. TNF-R superfamily signaling The TNF-R superfamily is remarkably diverse with more than two-dozen receptors and nearly as many ligands that are variably expressed throughout the body. However, despite the physiological diversity of
Oncogene

responses, in most cases signaling converges on the activation of NF-kB and AP-1. NF-kB activation in response to TNF signaling induces expression of antiapoptotic genes such as cIAP1/2 and Bcl-XL (see Dutta et al., 2006). TNF family receptors lack intrinsic enzymatic activity. Instead, signaling is achieved by recruitment of intracellular adapter molecules that associate with the cytoplasmic tail of the TNF-R in a signal-dependent manner (Figure 6a). The recruitment of TNF-R1 to membrane microdomains, referred to as lipid rafts, with subsequent assembly of the signaling complex, is necessary for signaling to NF-kB and prevention of apoptosis (Hueber, 2003; Legler et al., 2003). Ligation of TNF-R1 by trimeric TNFa causes aggregation of the receptor allowing binding of the TNF-R-associated death domain protein (TRADD). TRADD subsequently recruits adapter molecules including TRAF2

FADD TRADD

RIP1 Ub

RIP1 TRAF2/5/6

Ub

TRAF2/5 TRAF3
TAK1
MEKK3

NIK

Ub P NEMO

P NF- B

P p100 RelB

p52 RelB

Gene Transcription

Figure 6 TNF receptor superfamily signaling to NF-kB. Activation of the canonical NF-kB pathway downstream of TNF-RI is initiated by trimerization through ligand binding and recruitment of FADD, TRADD, RIP1 and TRAF2/5 to the receptor (a). TAK1 and MEKK3 are subsequently recruited to the receptor complex through RIP1, and with TRAF2/5, mediate the activation of IKK. The phosphorylation and degradation of classical IkBs require IKKb and NEMO. The non-canonical pathway is mediated by IKKa through NIK (b). In the resting state, NIK is inactivated/ degraded through an interaction with TRAF3. Upon stimulation, TRAF3 is inactivated/degraded resulting in the accumulation of NIK, activation of IKKa, phosphorylation of p100 and liberation of p100-inhibited NF-kB complexes. Simultaneous activation of the canonical NF-kB pathway through either TRAF2, 5, or 6 also commonly occurs downstream of receptors that activate the non-canonical pathway. See the text for additional details.

NF-jB and immunology MS Hayden et al

6771

(Hsu et al., 1996); however, TRAF2 and TRAF5 appear to play redundant roles in TNF signaling to NF-kB. TRAF2 or TRAF5-decient mice have intact TNF activation of NF-kB, whereas TRAF2/5 double knockout cells have substantially reduced TNF-induced IKK activation (Yeh et al., 1997; Nakano et al., 1999; Tada et al., 2001). TRAFs may either recruit the IKK complex directly (Devin et al., 2001) or indirectly through the serine/threonine kinase, RIP1. RIP1 can also interact independently with TRADD and is an essential adapter for TNF-induced NF-kB activation and protection from apoptosis (Hsu et al., 1996; Ting et al., 1996; Kelliher et al., 1998; Devin et al., 2000). Upon ubiquitination, RIP1 can bind directly to NEMO and recruit IKK independent of TRAF2 (Zhang et al., 2000). Signaling downstream of RIP1 requires TAK1 for the activation of IKK (Sato et al., 2000; Shim et al., 2005). Whether TAK1 directly activates IKK or this process proceeds through an intermediary such as MEKK3 is not yet clear (Takaesu et al., 2003; Li et al., 2006). However, it does not appear that TAK1 is involved in the activation of the non-canonical NF-kB pathway (Shim et al., 2005). The non-canonical NF-kB pathway is unique in that it is independent of IKKb and NEMO and instead requires IKKa which is phosphorylated by NF-kB inducing kinase (NIK) (Xiao et al., 2001; Senftleben et al., 2001a; Claudio et al., 2002; Dejardin et al., 2002; see Scheidereit, 2006). The key question concerning signaling by these stimuli is how they are channeled to NIK and IKKa, even though their receptor signaling domains resemble those of other TNF family members (Figure 6b). Because RANKL, BAFF and CD40L may also activate the canonical pathway through TRAF2/6, it would appear that the intracellular signaling domain of these receptors possess additional sequence motifs that allow their signaling to NIK. This function is mediated by TRAF3, which interacts with these receptors. TRAF3 negatively regulates NIK, and undergoes signal-dependent degradation resulting in the activation of the non-canonical pathway (Liao et al., 2004). Resolution of inammation may also involve NF-kB Resolution of inammation and subsequent tissue repair is a crucial event, and its failure is a common source of pathology. It is believed that reversal of inammation is an active process that is as complex as the inammatory response itself, and involves numerous pathways that are not all directly relevant to NF-kB (Serhan and Savill, 2005). Although the traditional view of NF-kB would lead one to imagine that it would primarily function by being turned off during the resolution phase of inammation, recent work has suggested that NF-kB also has a more active role. During acute inammation, there are multiple negative feedback pathways that help to rein in inammatory responses. It has long been known that cells such as macrophages become resistant to repeated proinammatory stimuli. BCL-3, which is induced late following

LPS stimulation, in combination with p50 dimers has been shown to have a role in the inhibition of repeated LPS responses in macrophages, a phenomenon also referred to as LPS tolerance (Wessells et al., 2004). Furthermore by selectively affecting chromatin remodeling, BCL-3 mediates repression of proinammatory genes, but also facilitates expression of the antiinammatory gene IL-10. NF-kB p50 also appears to negatively regulate IFNg production and proliferation by NK cells (Tato et al., 2006). In addition to these and other negative feedback pathways, it was recently found that inhibition of NF-kB during the resolution phase can prolong the inammatory process and prevent proper tissue repair (Lawrence et al., 2001). It was subsequently found that IKKa-decient mice display increased inammatory responses in models of local and systemic inammation (Lawrence et al., 2005). Macrophages, in particular, show increased production of proinammatory chemokines and cytokines in the absence of IKKa (Lawrence et al., 2005; Li et al., 2005). It was suggested from these studies that IKKa negatively regulates proinammatory gene expression, perhaps through mediating degradation of RelA and c-Rel following macrophage activation by LPS.

Initiation of adaptive responses Although innate responses alone can bring about potent antimicrobial activities, alerting and activating the adaptive immune system remains a crucial step for robust and durable immune responses. This process is largely mediated by activation and maturation of antigen-presenting cells (APCs), which can, in turn instruct T and B cells to carry out the adaptive response. DC maturation mediated by pathogen recognition is crucial for the initiation of the adaptive immune response. To activate na ve T cells, DCs must undergo multiple changes. First, DCs must gain the ability to interact with T cells by changing their chemokine receptor expression and migrating into lymphoid tissues. Second, DCs must alter their antigen processing machinery to favor the presentation of pathogen epitopes on MHC. Third, APCs must upregulate the expression of costimulatory molecules B7.1/B7.2 (or CD80/CD86), which are regulated by NF-kB and ligate CD28 providing the second signal necessary to induce T-cell activation. Finally, as progress is made in exploring these events it is becoming increasingly clear that the responses to different pathogens are tailored based on the distribution of PRRs in different cell types and the ability of different cell types to, in turn, interact with T cells in a biasing manner (Iwasaki and Medzhitov, 2004). Maturation of DCs following viral infection depends on nucleic acid-binding PRRs, including both TLRs and cytoplasmic RIG family molecules. Indeed, DC maturation during viral infection occurs normally in the absence of MyD88 or TLR3, as reported previously
Oncogene

NF-jB and immunology MS Hayden et al

6772

(Lopez et al., 2003). Bacterial responses are either mediated through TLRs DCs express TLRs 1, 2, 5 and 6 or other classes of PRRs. Murine CD8a DCs, which tend to induce TH1 responses important in clearance of viral and parasitic infections, express TLR1, 2, 6, 9 and 11. In the absence of TLR11, for example, mice fail to mount a TH1 response against T. gondii, owing to the failure of CD8a DCs to recognize the TLR11 ligand (Yarovinsky et al., 2005). It remains unclear whether the ability of distinct TLR ligands to induce TH1 versus TH2 responses is intrinsic to the specic TLR/ligand, dose of ligand or the cell type within which this activation occurs; evidence to date points towards the latter. Even less clear are the roles of other PRRs in APC maturation and the initiation of adaptive responses. Finally, there is the question of the role of innate recognition of pathogens by lymphocytes themselves. Recently, it has been suggested that TLR signaling in B cells is also required for optimal response to certain antigens (Pasare and Medzhitov, 2005). Both B and T cells express TLRs, although exactly which TLRs are expressed is debatable. At the level of mRNA

expression, human peripheral B cells express TLR1, 2, 4, 6 and 9 (Hornung et al., 2002).

Role of NF-jB in the adaptive response AgR signaling to NF-kB The hallmark of the adaptive immune response is antigen specicity. In the section on hematopoiesis, we discussed how NF-kB plays an important role in the selection of lymphocytes bearing somatically generating AgRs. Signaling through these antigen-specic B-cell and T-cell receptors is therefore the central event of the adaptive immune response. Activation of NF-kB downstream of BCR and TCR ligation facilitates antigenspecic proliferation and maturation of lymphocytes into effector cells. Signaling through these two AgRs appears to be functionally analogous, although some of the components utilized differ. BCR and TCR signaling are analogous in many aspects, in particular with respect to the activation of NF-kB (Figure 7). The T-cell receptor complex consists

a
Ag

b
Ag

TCR

BCR

P ZAP70 PDK1 CARMA1 P BCL10 MALT1 P Ras IP3 DAG P NFAT PP NF- B
Ub NEMO

CARMA1 BCL10 MALT1 P

Syk

Lyn

Vav

Vav

PI3K

RIP2

TRAF6
DAG IP3

Ras

AP-1 NFAT

AP-1

Gene Transcription

Figure 7 AgR signaling to NF-kB. Binding of TCR to peptide:MHC and co-stimulation through CD28:B7 interaction activates NF-kB (a). Through multiple steps, initial phosphorylation events at the TCR complex lead to PI3K activation and recruitment of PDK1. PDK1, in turn, mediates recruitment of the IKK complex through PKCy and the CARMA1 directly. PKCy phosphorylates CARMA1 leading to the activation of IKK through BCL10, MALT1 and TRAF6. Binding to TI antigens or TD antigens in the presence of co-stimulatory cytokines results in the activation of NF-kB through the BCR complex (b). Receptor proximal events of BCR activation are highly analogous to those of the TCR. Although the role of PDK1 in B-cell signaling is not established, BCR signaling to NF-kB requires PKCb as well as the CBM complex to achieve activation of IKK.
Oncogene

NF-jB and immunology MS Hayden et al

6773

of a/b subunits that are associated with the CD3 protein heterodimers g/e, d/e, and either Z/Z, Z/z or z/z, for a total of eight membrane proteins. The BCR is, likewise, a multiprotein complex consisting of the surface immunoglobulin receptor associated with a heterodimer of Iga and Igb. The AgR complexes associate with Src family tyrosine kinases (SFKs), Lck and Fyn in T cells and Lyn in B cells, which phosphorylate immunoreceptor tyrosine activation motifs (ITAMs) on CD3 and Iga/b chains. The cytoplasmic tyrosine kinases ZAP70 or Syk are then recruited via SH2 domains to the phosphorylated ITAMs and initiate activation of the IP3 and Ras family pathways. The IKK complex is rapidly recruited to the immunological synapse and can be colocalized to the TCR (Khoshnan et al., 2000; Weil et al., 2003). In ZAP-70-decient T cells, NF-kB activation can be rescued by directly targeting a chimeric NEMO to the immunological synapse, suggesting that signaling downstream of ZAP-70 may largely function for IKK recruitment (Weil et al., 2003). The signaling pathway from the receptor to NF-kB requires PKCy (PKCb in B cells), CARMA1/CARD11, BCL10 and MALT1 (Sun et al., 2000; Ruland et al., 2001, 2003; Saijo et al., 2002; Hara et al., 2003; RueiBrasse et al., 2003;). Although there has been some controversy, available data suggest that PKCy is largely essential for activation of NF-kB via T-cell stimulation (Sun et al., 2000; Pfeifhofer et al., 2003) and can mediate the activation of IKK (Lin et al., 2000). PKCy is specically recruited to the immunological synapse; although how PKCy but not other PKC isoforms is selectively recruited remains a mystery. In B cells PKCb is, likewise, required for recruitment of the IKK complex to lipid rafts following BCR ligation (Su et al., 2002). PKCy is capable of directly interacting with the IKK complex in primary T cells (Khoshnan et al., 2000) and might, therefore, function by bringing IKK to the receptor complex and into proximity with other essential components in this pathway; namely CARMA1, BCL10 and MALT1 (collectively known as the CBM complex). Recently, studies have provided evidence that the protein kinase PDK1 recruits PKCy and the IKK complex to lipid rafts. In addition, PDK1 can simultaneously recruit the CBM complex through binding to CARMA1 (Lee et al., 2005). The induced proximity of PKC and the CBM may allow phosphorylation of CARMA1 by PKCy/b (Matsumoto et al., 2005; Sommer et al., 2005). Interestingly, T-cell-specic PDK1 conditional deletion results in a defect in T-cell development, preventing the production of peripheral T cells (Hinton et al., 2004). On the other hand, PCKy knockouts do not display defects in thymocyte development. The CBM complex is essential for both AgR signaling in mature B and T cells. What is surprising however, as mentioned above, is the lack of a role for this complex in developing lymphocytes, and by extrapolation signaling through the pre-AgRs. The MAGUK family protein CARMA1 is required for activation of NF-kB in T cells following TCR ligation, but its loss has no effect on the development of thymocyte (Gaide et al., 2002; Egawa

et al., 2003; Hara et al., 2003). Similarly, BCL10 is critical for NF-kB activation via the BCR and TCR, yet normal numbers of peripheral T cells are seen in BCL10 knockouts, and no clear defects in B-cell development is observed (Ruland et al., 2001). BCL10 interacts with CARMA1 leading to BCL10 phosphorylation, although CARMA1 lacks kinase activity (Bertin et al., 2001; Gaide et al., 2001). Interestingly, genetic evidence of a role for RIP2 has recently been reported in T-cell signaling (Ruei-Brasse et al., 2004). RIP2 associates with BCL10 and is necessary for TCR-induced BCL10 phosphorylation and IKK activation. It is not yet clear how TCR signaling regulates RIP2 and, in turn, how RIP2 might affect IKK activity. BCL10 oligomerization has been implicated in IKK activation through a process that involves ubiquitination of NEMO (Zhou et al., 2004). This ubiquitination event appears to be mediated by MALT1, and perhaps TRAF6, although no TCR signaling decits have been reported in TRAF6-decient mice (Lomaga et al., 1999; Sun et al., 2004). If this is true however, it is possible that the effect is mediated through the kinase TAK1, which functions in IKK activation downstream of TRAF6 in other pathways. However, B cells in which TAK1 has been conditionally inactivated have normal BCR signaling to NF-kB. More work must be carried out with TAK1 conditional knockouts in both the BCR and TCR pathways to address the role of this kinase. T-cell responses mediated by NF-kB To become activated, na ve T cells must receive two distinct signals: antigen-specic and co-stimulatory. Antigen-specic activation signals emanate from the binding of the TCR to cognate antigenic peptides presented in the binding cleft of MHC. Co-stimulatory signaling is provided through ligation of CD28 by B7 molecules expressed on activated APCs. Stimulation of na ve T cells results in the production of IL-2, which is necessary for their proliferation and survival. These activated na ve T cell blasts proliferate rapidly and simultaneously undergo differentiation into effector cells. In the case of TH cells, proliferation leads to differentiation into immature effector cells, TH0, which subsequently differentiate into TH1 or TH2 cells depending on the predominant cytokine milieu. CD8 T cells are likewise activated by professional APCs, although they may receive secondary signals from activated TH1 cells. The unavailability of CD8 conditional knockouts, and the selective loss of CD8 cells in the absence of NF-kB activity has prevented a thorough characterization of the role of NF-kB in these cells. Rapidly proliferating activated T cells rely on NF-kB activity for protection from apoptosis as well as for the production of cytokines supporting proliferation and differentiation. As expected, inhibition of NF-kB in activated T cells facilitates progression towards AICD or apoptosis (Ivanov and Nikolic-Zugic, 1997; Jeremias et al., 1998). Indeed, stimulation of RelA-decient na ve T cells induces cell death (Wan and DeGregori, 2003).
Oncogene

NF-jB and immunology MS Hayden et al

6774

Peripheral T cells lacking c-Rel do not undergo apoptosis, but nevertheless, fail to proliferate in response to typical mitogenic stimuli (Ko ntgen et al., 1995), and both RelA and c-Rel containing complexes accumulate in the nucleus following TCR/CD28 stimulation (Ghosh et al., 1993). Interestingly, c-Rel-decient T cells appear to have a defect in TH1 proliferation and production of IFNg, indicating a selective role for NF-kB family members in TH1/TH2 differentiation, independent of that mediated by the innate response. Multiple transcriptional activators and repressors regulate expression of IL-2. Among these, members of the NF-kB family play multiple roles. In na ve T cells, which do not express IL-2, repressive p50 homodimers are found associated with the IL-2 promoter (Grundstrom et al., 2004). Failure of T-cell proliferative responses in c-Rel knockout mice is attributable to a failure to produce IL-2 (Ko ntgen et al., 1995). In na ve T cells, c-Rel is responsible for mediating chromatin remodeling across the IL-2 locus following CD3/CD28 co-stimulation (Rao et al., 2003). Na ve T cells can be primed by exposure to inammatory cytokines such that they generate a more robust response to CD3/CD28 costimulation. Overexpression of an IkBa super-repressor suggested that NF-kB is required for this priming event in T cells (Mora et al., 2001a). More recent data indicate that c-Rel is necessary for na ve helper T-cell priming by pro-inammatory cytokines elicited following stimulation with TLR ligands (Banerjee et al., 2005). NF-kB RelA-containing complexes, on the other hand, appear to function more traditionally in mediating transactivation of IL-2 gene expression, and overexpression of RelA with c-Jun can overcome the requirement for co-stimulation in na ve T cells (Parra et al., 1998). However, as discussed below, these complexes may also be the targets of negative regulation following T-cell differentiation. Recent work in TH1/TH2 differentiation has focused on the induction of specic transcription factors in these two effector cell types T-bet and GATA3, respectively. Interestingly, mice lacking p50 are unable to mount an asthma-like airway TH2 response, and do not induce GATA-3 expression during T-cell stimulation under TH2 differentiating conditions (Das et al., 2001). Consistent with this nding, BCL-3-decient T cells also fail to undergo TH2 differentiation. Furthermore, BCL-3 can induce expression of a reporter gene from a gata-3 promoter, suggesting that p50/BCL-3 complexes are crucial for TH2 differentiation (Corn et al., 2005). Conversely, the same authors found that RelB-decient T cells are decient in TH1 differentiation and IFNg production, and show decreased expression of T-bet; likely through a failure to upregulate STAT4, which functions in signaling from IFN to T-bet induction. Therefore, it appears that NF-kB activation during TCR stimulation may render cells competent for both proliferative and differentiating stimuli. As TH cells differentiate into TH1 or TH2, they decrease their expression of IL-2 and, instead, become dependent on TH1 and TH2 cytokines (e.g., IFNg and IL-4). As a corollary, NF-kB transactivation of the IL-2
Oncogene

gene is repressed. Direct binding of T-bet to p65 that is associated with the IL-2 gene enhancer may mediate the repression of IL-2 production in TH1 cells (Hwang et al., 2005). Alternatively, in TH2 cells the lack of IL-2 transcription may be due to the decreased levels of RelA activation in TH2 cells (Lederer et al., 1994). B-cell responses mediated by NF-kB B-cell responses can be classied into two groups: thymus-dependent (TD) or thymus-independent (TI). In response to T-dependent antigens, B cells require costimulatory signaling from TH cells expressing CD40L and cytokines, such as IL-4. B cells from individuals with a mutation in CD40L are unable to undergo class switch recombination in response to T-dependent antigens (Aruffo et al., 1993). Signaling through CD40 activates both canonical and non-canonical NF-kB pathways, although it is unclear which is operative in the response to T-dependent antigens. For example, whereas B cells from p52/ mice mount inadequate humoral responses to various T-dependent antigens, they exhibit a normal response following adoptive transfer into rag-1/ mice indicating that this decit is not intrinsic to B cells (Franzoso et al., 1998). Furthermore, B cells from relB/ mice, although crippled in their proliferative response, undergo normal IgM secretion and class switching in response to various stimuli (Snapper et al., 1996a). Therefore, non-canonical NF-kB pathway activation downstream of CD40 is probably not required for class switching during T-dependent antigen responses. Analysis of the canonical NF-kB pathway is complicated by more generalized defects in lymphocyte response owing to the requirement for this pathway in AgR signaling. Nevertheless, whether downstream of CD40 or other stimuli, evidence supports canonical pathway activation in the process of class switch recombination. Following adoptive transfer, B cells from rela/ mice exhibit markedly diminished class switching, despite a modest loss of lymphocyte proliferation following various stimuli (Doi et al., 1997). Likewise, c-Rel-decient mice, or mice lacking the c-Rel C-terminal transactivation domain, fail to generate a productive humoral immune response suggesting a requirement for c-Rel in class switch recombination (Ko ntgen et al., 1995; Zelazowski et al., 1997; Carrasco et al., 1998). B cells from nfkb1/ mice exhibit decreased proliferation in response to mitogenic stimulation, and p50/RelA double knockout B cells exhibit greater defects in proliferation and class switching (Snapper et al., 1996b; Horwitz et al., 1999). Therefore, analyses of knockout animals suggest that the canonical NF-kB pathway likely has a role in maturation of the B-cell response in addition to directly mediating proliferative responses following BCR ligation. As discussed above, signaling through TLRs has an important role in the initiation of the adaptive immune response via APCs of the innate immune system. In recent years, however, there has also been increasing interest in the ability of TLR signaling to directly

NF-jB and immunology MS Hayden et al

6775

modulate the adaptive response. For example, it has been observed that homeostatic polyclonal activation of B cells, which results in the so-called serological memory, that is, detectable antibody to antigens that are no longer present in the host, can be induced/ maintained by TLR ligation (Bernasconi et al., 2002). Analogously, TLR2 is upregulated in CD4 T cells following TCR stimulation, and TLR2 ligands may thus provide an activation/maintenance signal in these cells (Komai-Koma et al., 2004). That signaling through TLRs in these aspects of B-cell responses requires NF-kB seems likely, but has yet to be demonstrated. TI antigens have an intrinsic ability to activate B-cell responses in the absence of T-cell help by acting as B-cell mitogens, for example by acting as TLR ligands or by binding with high avidity to the BCR through repetitive structural features. In such cases, it is expected that B-cell responses are more dependent on members of the canonical pathway that have well-documented roles in TLR signaling, or BCR signaling (as discussed above). For example, c-Rel-decient B cells are highly sensitive to apoptosis following BCR cross-linking (Grumont et al., 1998, 1999; Owyang et al., 2001). As mentioned above, p50 and p50/pRelA double knockout B cells are decient in responses to TI stimulation. Likewise, IKKb-decient B cells fail to mount TI or TD responses (Li et al., 2003). These IKKb-decient B cells also exhibit increased spontaneous apoptosis, suggesting that NF-kB is important in survival of B cells. Maintenance and memory: a role for NF-kB in lymphoid cell survival Lymphocyte homeostasis is dependent on the survival of mature lymphocytes in addition to replenishment of the peripheral lymphocyte pool through lymphopoiesis. Consequently, there is increasing interest in the possible role of NF-kB in the survival of mature lymphocytes. It is widely accepted that lymphocyte survival is mediated through tonic stimulation downstream of the AgR, as well as certain cytokine receptors. As discussed above, genetic targeting experiments support an important role for NF-kB family members in lymphocyte survival. Na ve T cells require continued contact with MHC:selfpeptides, most likely expressed on lymphoid DCs, to generate the tonic TCR signal that is essential for continued survival. Survival of memory cells, on the other hand, is independent of continued contact with self-peptide:MHC complexes. B cells are formed at a far higher rate than T cells and therefore B cells also undergo a signicantly higher rate of turnover. Nonetheless, B cells too require maintenance signals to achieve peripheral homeostasis. The AgR on B cells most likely provides a basal level of signaling, albeit independent of the presence of antigen, which is required for maintenance of mature B cells. Not surprisingly, B cells from RelA/, p100/, p105/ and c-Rel/ mice display increased sensitivity to apoptosis and/or decreased survival ex vivo (Grumont et al., 1998; Claudio et al., 2002; Prendes et al., 2003).

In large part, these defects appear to be due to a loss of BCR signaling, as demonstrated in an elegant study that demonstrated that deletion of the BCR from mature B cells led to a complete loss of the peripheral B-cell pool (Kraus et al., 2004). Most likely this was due to the loss of signaling to NF-kB in these cells because loss of IKKb, NEMO or components of the CBM complex in mature B cells also results in a complete loss of peripheral B cells (Pasparakis et al., 2002; Li et al., 2003; Thome, 2004). The non-canonical NF-kB pathway is also relevant to B-cell survival, as the loss of IKKa results in striking defects in B-cell survival (Kaisho et al., 2001; Senftleben et al., 2001a). However, rather than acting downstream of tonic BCR signaling, recent studies have implicated signaling from BAFFR in this aspect of the B lymphocyte survival (reviewed in Mackay et al., 2003). Together, these data suggest that a subset of Bcl-2 family members, for example, the antiapoptotic factor A1, are regulated by p52/RelB-containing complexes and are necessary for the maintenance of mature B cells.

Concluding remarks NF-kB was originally described as a transcriptional regulator in the adaptive immune response; however, subsequent studies have revealed its importance in hematopoiesis, lymphoid organogenesis and innate immunity. Investigations of mice with targeted deletions and mutations have shed considerable light on the role of NF-kB in lymphoid organogenesis. For example, characterization of the aly/aly mouse in part led to the discovery of the non-canonical NF-kB pathway and its role in organogenesis. Progress continues to be made in understanding AgR signaling to NF-kB, and this work has led to the appreciation of regulatory ubiquitination events in IKK activation (see reviews by Perkins, 2006; Scheidereit, 2006). The specicity of the requirement for the CBM complex and PKCy/b in signaling by T-cell and B-cell receptors opens the door to the development of equally specic NF-kB inhibitors. However, there remain fundamental gaps in our understanding of how these components mediate IKK activation and the possible role of regulatory ubiquitination in this process. The role of NF-kB in hematopoiesis remains partially dened, although there is little doubt about its importance. Genetic targeting in mice has allowed the enumeration of multiple steps in hematopoiesis at which various NF-kB pathway components are required; however, a cohesive picture of how NF-kB functions in these steps remains elusive. For example, whereas we know that canonical NF-kB activity is required in early lymphopoiesis, it is unclear whether this is in the regulation of TNFa production or in mediating a survival signal in the developing lymphocyte. NF-kB is required in the process of positive and negative selection, but again, its mechanism(s) of action remains poorly dened. Progress using in vitro systems for studying lymphocyte development may allow a more
Oncogene

NF-jB and immunology MS Hayden et al

6776

rigorous assessment of NF-kB function in these processes. Likewise, advances in conditional gene targeting approaches and the generation of animals in which defective versions of NF-kB genes have been knocked-in may help to overcome the problems of embryonic lethality and functional redundancy that have sometimes made existing studies difcult to interpret. T- and B-cell responses require NF-kB as a prosurvival factor as well as for the regulation of genes involved in differentiation to effector cells. More recently, a limited number of studies have suggested a role for PRRs in lymphocyte activation, and the role of NF-kB in this process remains to be elucidated. The role of NF-kB in the differentiation of TH cells is incompletely understood, and requires further clarication. We also know very little about the role of NF-kB in CD8 activation, differentiation and function; progress in this area awaits development of better tools for genetically manipulating this cell population. Upon successful clearance of pathogen, regulation of NF-kB allows resolution of the response and facilitates the development of memory cells. To date, relatively little is known
References Abe K, Yarovinsky FO, Murakami T, Shakhov AN, Tumanov AV, Ito D et al. (2003). Blood 101: 14771483. Akira S, Uematsu S, Takeuchi O. (2006). Cell 124: 783801. Alcamo E, Hacohen N, Schulte LC, Rennert PD, Hynes RO, Baltimore D. (2002). J Exp Med 195: 233244. Alcamo E, Mizgerd JP, Horwitz BH, Bronson R, Beg AA, Scott M et al. (2001). J Immunol 167: 15921600. Alexopoulou L, Holt AC, Medzhitov R, Flavell RA. (2001). Nature 413: 732738. Andrejeva J, Childs KS, Young DF, Carlos TS, Stock N, Goodbourn S et al. (2004). Proc Natl Acad Sci USA 101: 1726417269. Aruffo A, Farrington M, Hollenbaugh D, Li X, Milatovich A, Nonoyama S et al. (1993). Cell 72: 291300. Ayabe T, Satchell DP, Wilson CL, Parks WC, Selsted ME, Ouellette AJ. (2000). Nat Immunol 1: 113118. Balachandran S, Thomas E, Barber GN. (2004). Nature 432: 401405. Banerjee D, Liou H-C, Sen R. (2005). Immunity 23: 445458. Barton GM, Kagan JC, Medzhitov R. (2006). Nat Immunol 7: 4956. Batten M, Groom J, Cachero TG, Qian F, Schneider P, Tschopp J et al. (2000). J Exp Med 192: 14531466. Beg AA, Sha WC, Bronson RT, Baltimore D. (1995). Genes Dev 9: 27362746. Bernasconi NL, Traggiai E, Lanzavecchia A. (2002). Science 298: 21992202. Bertin J, Wang L, Guo Y, Jacobson MD, Poyet JL, Srinivasula SM et al. (2001). J Biol Chem 276: 1187711882. Bonizzi G, Karin M. (2004). Trends Immunol 25: 280288. Burkly L, Hession C, Ogata L, Reilly C, Marconi LA, Olson D et al. (1995). Nature 373: 531536. Burns K, Clatworthy J, Martin L, Martinon F, Plumpton C, Maschera B et al. (2000). Nat Cell Biol 2: 346351. Caamano JH, Perez P, Lira SA, Bravo R. (1996). Mol Cell Biol 16: 13421348. Caamano JH, Rizzo CA, Durham SK, Barton DS, RaventosSuarez C, Snapper CM et al. (1998). J Exp Med 187: 185196.
Oncogene

about the role of NF-kB in memory cells, although recent advances in identifying and characterizing memory precursors bodes well for future progress in this area. In summary, research to date has highlighted the importance of NF-kB in regulating genes that prevent apoptosis and that promote differentiation and development in cells of the innate and adaptive immune systems. A large body of work has elucidated many of the molecular mechanisms governing regulation of NF-kB by engagement of innate or lymphocyte antigen-specic receptors. In turn, these data provide a trove of information that will likely prove useful in attempts to manipulate the immune system to prevent and treat disease.
Acknowledgements Research in the authors laboratory was supported by grants from the National Institutes of Health (to SG). MSH was supported by NIH/National Institute of General Medical Sciences Medical Scientist Training Grant GM07205.

Campos MA, Almeida IC, Takeuchi O, Akira S, Valente EP, Procopio DO et al. (2001). J Immunol 167: 416423. Carrasco D, Cheng J, Lewin A, Warr G, Yang H, Rizzo C et al. (1998). J Exp Med 187: 973984. Catley MC, Chivers JE, Cambridge LM, Holden N, Slater DM, Staples KJ et al. (2003). FEBS Lett 547: 7579. Claudio E, Brown K, Park S, Wang H, Siebenlist U. (2002). Nat Immunol 3: 958965. Coban C, Ishii KJ, Kawai T, Hemmi H, Sato S, Uematsu S et al. (2005). J Exp Med 201: 1925. Corn RA, Hunter C, Liou H-C, Siebenlist U, Boothby MR. (2005). J Immunol 175: 21022110. Courtois G, Gilmore TD. (2006). Oncogene 25: 68316843. Covert MW, Leung TH, Gaston JE, Baltimore D. (2005). Science 309: 18541857. Das J, Chen CH, Yang L, Cohn L, Ray P, Ray A. (2001). Nat Immunol 2: 4550. Dejardin E, Droin NM, Delhase M, Haas E, Cao Y, Makris C et al. (2002). Immunity 17: 525535. Devin A, Cook A, Lin Y, Rodriguez Y, Kelliher M, Liu Z. (2000). Immunity 12: 419429. Devin A, Lin Y, Yamaoka S, Li Z, Karin M, Liu Z. (2001). Mol Cell Biol 21: 39863994. Didierlaurent A, Brissoni B, Velin D, Aebi N, Tardivel A, Kaslin E et al. (2006). Mol Cell Biol 26: 735742. Diebold SS, Kaisho T, Hemmi H, Akira S, Reis e Sousa C. (2004). Science 303: 15291531. Doi TS, Takahashi T, Taguchi O, Azuma T, Obata Y. (1997). J Exp Med 185: 953961. Dougall WC, Glaccum M, Charrier K, Rohrbach K, Brasel K, De Smedt T et al. (1999). Genes Dev 13: 24122424. linas C. (2006). Oncogene Dutta j, Fan Y, Gupta N, Fan G, Ge 25: 68006816. Eck SL, Perkins ND, Carr DP, Nabel GJ. (1993). Mol Cell Biol 13: 65306536. Egawa T, Albrecht B, Favier B, Sunshine MJ, Mirchandani K, OBrien W et al. (2003). Curr Biol 13: 12521258.

NF-jB and immunology MS Hayden et al

6777 Feng B, Cheng S, Hsia CY, King LB, Monroe JG, Liou H-C. (2004). Cell Immunol 232: 920. Fitzgerald KA, Palsson-McDermott EM, Bowie AG, Jefferies CA, Mansell AS, Brady G et al. (2001). Nature 413: 7883. Fitzgerald KA, Rowe DC, Barnes BJ, Caffrey DR, Visintin A, Latz E et al. (2003). J Exp Med 198: 10431055. Franchi L, Amer A, Body-Malapel M, Kanneganti TD, Ozoren N, Jagirdar R et al. (2006). Nat Immunol 7: 576582. Francois S, El Benna J, Dang PM, Pedruzzi E, GougerotPocidalo MA, Elbim C. (2005). J Immunol 174: 36333642. Franzoso G, Carlson L, Poljak L, Shores EW, Epstein S, Leonardi A et al. (1998). J Exp Med 187: 147159. Franzoso G, Carlson L, Xing L, Poljak L, Shores EW, Brown KD et al. (1997). Genes Dev 11: 34823496. French LE, Hahne M, Viard I, Radlgruber G, Zanone R, Becker K et al. (1996). J Cell Biol 133: 335343. Gaide O, Favier B, Legler DF, Bonnet D, Brissoni B, Valitutti S et al. (2002). Nat Immunol 3: 836843. Gaide O, Martinon F, Micheau O, Bonnet D, Thome M, Tschopp J. (2001). FEBS Lett 496: 121127. Gazzinelli RT, Ropert C, Campos MA. (2004). Immunol Rev 201: 925. Ghosh P, Tan TH, Rice NR, Sica A, Young HA. (1993). Proc Natl Acad Sci USA 90: 16961700. Gilmore TD. (2006). Oncogene 25: 66806684. Goudeau B, Huetz F, Samson S, Di Santo JP, Cumano A, Beg A et al. (2003). Proc Natl Acad Sci USA 100: 1580015805. Gross JA, Dillon SR, Mudri S, Johnston J, Littau A, Roque R et al. (2001). Immunity 15: 289302. Grossmann M, Metcalf D, Merryfull J, Beg A, Baltimore D, Gerondakis S. (1999). Proc Natl Acad Sci USA 96: 1184811853. Grossmann M, OReilly LA, Gugasyan R, Strasser A, Adams JM, Gerondakis S. (2000). EMBO J 19: 63516360. Grumont RJ, Rourke IJ, Gerondakis S. (1999). Genes Dev 13: 400411. Grumont RJ, Rourke IJ, OReilly LA, Strasser A, Miyake K, Sha W et al. (1998). J Exp Med 187: 663674. Grundstrom S, Anderson P, Scheipers P, Sundstedt A. (2004). J Biol Chem 279: 84608468. Hara H, Wada T, Bakal C, Kozieradzki I, Suzuki S, Suzuki N et al. (2003). Immunity 18: 763775. Hayashi F, Smith KD, Ozinsky A, Hawn TR, Yi EC, Goodlett DR et al. (2001). Nature 410: 10991103. Hayden MS, Ghosh S. (2004). Genes Dev 18: 21952224. Heil F, Hemmi H, Hochrein H, Ampenberger F, Kirschning C, Akira S et al. (2004). Science 303: 15261529. Hemmi H, Kaisho T, Takeuchi O, Sato S, Sanjo H, Hoshino K et al. (2002). Nat Immunol 3: 196200. Hemmi H, Takeuchi O, Kawai T, Kaisho T, Sato S, Sanjo H et al. (2000). Nature 408: 740745. Hemmi H, Takeuchi O, Sato S, Yamamoto M, Kaisho T, Sanjo H et al. (2004). J Exp Med 199: 16411650. Hettmann T, DiDonato J, Karin M, Leiden JM. (1999). J Exp Med 189: 145158. Hinton HJ, Alessi DR, Cantrell DA. (2004). Nat Immunol 5: 539545. Honda K, Yanai H, Mizutani T, Negishi N, Shimada N, Susuki N et al. (2003). Proc Natl Acad Sci USA 101 : 1541615421. Horng T, Barton GM, Flavell RA, Medzhitov R. (2002). Nature 420: 329333. Hornung V, Rothenfusser S, Britsch S, Krug A, Jahrsdorfer B, Giese T et al. (2002). J Immunol 168: 45314537. Horwitz BH, Scott ML, Cherry SR, Bronson RT, Baltimore D. (1997). Immunity 6: 765772. Horwitz BH, Zelazowski P, Shen Y, Wolcott KM, Scott ML, Baltimore D et al. (1999). J Immunol 162: 19411946. Hsu H, Shu HB, Pan MG, Goeddel DV. (1996). Cell 84: 299308. Huang Q, Yang J, Lin Y, Walker C, Cheng J, Liu ZG et al. (2004). Nat Immunol 5: 98103. Hueber AO. (2003). Cell Death Differ 10: 79. Hwang ES, Hong JH, Glimcher LH. (2005). J Exp Med 202: 12891300. Inohara N, Nunez G. (2003). Nat Rev Immunol 3: 371382. Inohara N, Koseki T, Lin J, del Peso L, Lucas PC, Chen FF et al. (2000). J Biol Chem 275: 2782327831. Ishii KJ, Coban C, Kato H, Takahashi K, Torii Y, Takeshita F et al. (2006). Nat Immunol 7: 4048. Ivanov VN, Nikolic-Zugic J. (1997). J Biol Chem 272: 85588566. Iwasaki A, Medzhitov R. (2004). Nat Immunol 5: 987995. Janeway Jr CA. (1989). Cold Spring Harbor Symp Quant Biol 54(Part 1): 113. Jeremias I, Kupatt C, Baumann B, Herr I, Wirth T, Debatin KM. (1998). Blood 91: 46244631. Jimi E, Ghosh S. (2005). Immunol Rev 208: 8087. Jimi E, Phillips RJ, Rincon M, Voll R, Karasuyama H, Flavell R et al. (2005). Int Immunol 17: 815825. Jurk M, Heil F, Vollmer J, Schetter C, Krieg AM, Wagner H et al. (2002). Nat Immunol 3: 499. Kagan JC, Medzhitov R. (2006). Cell 125: 943955. Kaisho T, Takeda K, Tsujimura T, Kawai T, Nomura F, Terada N et al. (2001). J Exp Med 193: 417426. Kajiura F, Sun S, Nomura T, Izumi K, Ueno T, Bando Y et al. (2004). J Immunol 172: 20672075. Kang DC, Gopalkrishnan RV, Wu Q, Jankowsky E, Pyle AM, Fisher PB. (2002). Proc Natl Acad Sci USA 99: 637642. Kariko K, Buckstein M, Ni H, Weissman D. (2005). Immunity 23: 165175. Kato H, Sato S, Yoneyama M, Yamamoto M, Uematsu S, Matsui K et al. (2005). Immunity 23: 1928. Kato H, Takeuchi O, Sato S, Yoneyama M, Yamamoto M, Matsui K et al. (2006). Nature 441: 101105. Kawai T, Adachi O, Ogawa T, Takeda K, Akira S. (1999). Immunity 11: 115122. Kawai T, Takahashi K, Sato S, Coban C, Kumar H, Kato H et al. (2005). Nat Immunol 6: 981988. Kelliher MA, Grimm S, Ishida Y, Kuo F, Stanger BZ, Leder P. (1998). Immunity 8: 297303. Khoshnan A, Bae D, Tindell CA, Nel AE. (2000). J Immunol 165: 69336940. Kinjyo I, Hanada T, Inagaki-Ohara K, Mori H, Aki D, Ohishi M et al. (2002). Immunity 17: 583591. Kinoshita D, Hirota F, Kaisho T, Kasai M, Izumi K, Bando Y et al. (2006). J Immunol 176: 39954002. Kishimoto H, Surh CD, Sprent J. (1998). J Exp Med 187: 14271438. Kobayashi K, Hernandez LD, Galan JE, Janeway Jr CA, Medzhitov R, Flavell RA. (2002). Cell 110: 191202. Koike R, Nishimura T, Yasumizu R, Tanaka H, Hataba Y, Watanabe T et al. (1996). Eur J Immunol 26: 669675. Komai-Koma M, Jones L, Ogg GS, Xu D, Liew FY. (2004). Proc Natl Acad Sci USA 101: 30293034. Ko ntgen F, Grumont RJ, Strasser A, Metcalf D, Li R, Tarlinton D et al. (1995). Genes Dev 9: 19651977. Kopp E, Medzhitov R, Carothers J, Xiao C, Douglas I, Janeway CA et al. (1999). Genes Dev 13: 20592071. Kraus M, Alimzhanov MB, Rajewsky N, Rajewsky K. (2004). Cell 117: 787800.
Oncogene

NF-jB and immunology MS Hayden et al

6778 Kriehuber E, Bauer W, Charbonnier AS, Winter D, Amatschek S, Tamandl D et al. (2005). Blood 106: 175183. Krug A, French AR, Barchet W, Fischer JA, Dzionek A, Pingel JT et al. (2004). Immunity 21: 107119. Kurt-Jones EA, Popova L, Kwinn L, Haynes LM, Jones LP, Tripp RA et al. (2000). Nat Immunol 1: 398401. Lai WC, Zhou M, Shankavaram U, Peng G, Wahl LM. (2003). J Immunol 170: 62446249. Lawrence T, Bebien M, Liu GY, Nizet V, Karin M. (2005). Nature 434: 11381143. Lawrence T, Gilroy DW, Colville-Nash PR, Willoughby DA. (2001). Nat Med 7: 12911297. Lederer JA, Liou JS, Todd MD, Glimcher LH, Lichtman AH. (1994). J Immunol 152: 7786. Lee KY, DAcquisto F, Hayden MS, Shim JH, Ghosh S. (2005). Science 308: 114118. Legler DF, Micheau O, Doucey MA, Tschopp J, Bron C. (2003). Immunity 18: 655664. Levitz SM. (2004). Microbes Infect 6: 13511355. Li H, Kobayashi M, Blonska M, You Y, Lin X. (2006). J Biol Chem 281: 1363613643. Li Q, Lu Q, Bottero V, Estepa G, Morrison L, Mercurio F et al. (2005). Proc Natl Acad Sci USA 102: 1242512430. Li Q, Van Antwerp D, Mercurio F, Lee KF, Verma IM. (1999a). Science 284: 321325. Li ZW, Chu W, Hu Y, Delhase M, Deerinck T, Ellisman M et al. (1999b). J Exp Med 189: 18391845. Li ZW, Omori SA, Labuda T, Karin M, Rickert RC. (2003). J Immunol 170: 46304637. Liao G, Zhang M, Harhaj EW, Sun S-C. (2004). J Biol Chem 279: 2624326250. Lien E, Chow JC, Hawkins LD, McGuinness PD, Miyake K, Espevik T et al. (2001). J Biol Chem 276: 18731880. Lin X, Wang D. (2004). Semin Immunol 16: 429435. Lin X, OMahony A, Mu Y, Geleziunas R, Greene WC. (2000). Mol Cell Biol 20: 29332940. Lomaga MA, Yeh WC, Sarosi I, Duncan GS, Furlonger C, Ho A et al. (1999). Genes Dev 13: 10151024. Lopez CB, Garcia-Sastre A, Williams BR, Moran TM. (2003). J Infect Dis 187: 11261136. Lund J, Sato A, Akira S, Medzhitov R, Iwasaki A. (2003). J Exp Med 198: 513520. Lund JM, Alexopoulou L, Sato A, Karow M, Adams NC, Gale NW et al. (2004). Proc Natl Acad Sci USA 101: 55985603. Mackay F, Schneider P, Rennert P, Browning J. (2003). Annu Rev Immunol 21: 231264. Mambula SS, Sau K, Henneke P, Golenbock DT, Levitz SM. (2002). J Biol Chem 277: 3932039326. Mansell A, Smith R, Doyle SL, Gray P, Fenner JE, Crack PJ et al. (2006). Nat Immunol 7: 148155. Massari P, Ram S, Macleod H, Wetzler LM. (2003). Trends Microbiol 11: 8793. Matsumoto R, Wang D, Blonska M, Li H, Kobayashi M, Pappu B et al. (2005). Immunity 23: 575585. McDonald PP. (2004). Adv Immunol 82: 148. McDonald PP, Bald A, Cassatella MA. (1997). Blood 89: 34213433. McWhirter SM, Fitzgerald KA, Rosains J, Rowe DC, Golenbock DT, Maniatis T. (2004). Proc Natl Acad Sci USA 101: 233238. Means TK, Wang S, Lien E, Yoshimura A, Golenbock DT, Fenton MJ. (1999). J Immunol 163: 39203927. Mebius RE. (2003). Nat Rev Immunol 3: 292303. Melmed G, Thomas LS, Lee N, Tesfay SY, Lukasek K, Michelsen KS et al. (2003). J Immunol 170: 14061415.
Oncogene

Meylan E, Burns K, Hofmann K, Blancheteau V, Martinon F, Kelliher M et al. (2004). Nat Immunol 5: 503507. Meylan E, Curran J, Hofmann K, Moradpour D, Binder M, Bartenschlager R et al. (2005). Nature 437: 11671172. Miao EA, Alpuche-Aranda CM, Dors M, Clark AE, Bader MW, Miller SI et al. (2006). Nat Immunol 7: 569575. Miller LS, Sorensen OE, Liu PT, Jalian HR, Eshtiaghpour D, Behmanesh BE et al. (2005). J Immunol 174: 61376143. Minakhina S, Steward R. (2006). Oncogene 25: 67496757. Miyawaki S, Nakamura Y, Suzuka H, Koba M, Yasumizu R, Ikehara S et al. (1994). Eur J Immunol 24: 429434. Mizel SB, West AP, Hantgan RR. (2003). J Biol Chem 278: 2362423629. Mora A, Youn J, Keegan A, Boothby M. (2001a). J Immunol 166: 22182227. Mora AL, Stanley S, Armistead W, Chan AC, Boothby M. (2001b). J Immunol 167: 56285635. Motoyama N, Wang F, Roth KA, Sawa H, Nakayama K, Negishi I et al. (1995). Science 267: 15061510. Mueller T, Terada T, Rosenberg IM, Shibolet O, Podolsky DK. (2006). J Immunol 176: 58055814. Nakagawa R, Naka T, Tsutsui H, Fujimoto M, Kimura A, Abe T et al. (2002). Immunity 17: 677687. Nakano H, Sakon S, Koseki H, Takemori T, Tada K, Matsumoto M et al. (1999). Proc Natl Acad Sci USA 96: 98039808. Oshiumi H, Matsumoto M, Funami K, Akazawa T, Seya T. (2003). Nat Immunol 4: 161167. Ouaaz F, Arron J, Zheng Y, Choi Y, Beg AA. (2002). Immunity 16: 257270. Owyang AM, Tumang JR, Schram BR, Hsia CY, Behrens TW, Rothstein TL et al. (2001). J Immunol 167: 49484956. Ozinsky A, Smith KD, Hume D, Underhill DM. (2000). J Endotoxin Res 6: 393396. Parra E, McGuire K, Hedlund G, Dohlsten M. (1998). J Immunol 160: 53745381. Pasare C, Medzhitov R. (2005). Nature 438: 364368. Pasparakis M, Schmidt-Supprian M, Rajewsky K. (2002). J Exp Med 196: 743752. Paxian S, Merkle H, Riemann M, Wilda M, Adler G, Hameister H et al. (2002). Gastroenterology 122: 18531868. Perkins ND. (2006). Oncogene 25: 67176730. Pfeifhofer C, Koer K, Gruber T, Tabrizi NG, Lutz C, Maly K et al. (2003). J Exp Med 197: 15251535. Poligone B, Baldwin AS. (2001). J Biol Chem 276: 3865838664. Prendes M, Zheng Y, Beg AA. (2003). J Immunol 171: 39633969. Pyatt DW, Stillman WS, Yang Y, Gross S, Zheng JH, Irons RD. (1999). Blood 93: 33023308. Rao S, Gerondakis S, Woltring D, Shannon MF. (2003). J Immunol 170: 37243731. Rowe TM, Rizzi M, Hirose K, Peters GA, Sen GC. (2006). Proc Natl Acad Sci USA 103: 58235828. Rudolph D, Yeh WC, Wakeham A, Rudolph B, Nallainathan D, Potter J et al. (2000). Genes Dev 14: 854862. Ruei-Brasse AA, French DM, Dixit VM. (2003). Science 302: 15811584. Ruei-Brasse AA, Lee WP, Hurst S, Dixit VM. (2004). J Biol Chem 279: 15701574. Ruland J, Duncan GS, Elia A, del Barco Barrantes I, Nguyen L, Plyte S et al. (2001). Cell 104: 3342. Ruland J, Duncan GS, Wakeham A, Mak TW. (2003). Immunity 19: 749758. Saijo K, Mecklenbrauker I, Santana A, Leitger M, Schmedt C, Tarakhovsky A. (2002). J Exp Med 195: 16471652.

NF-jB and immunology MS Hayden et al

6779 Samson SI, Memet S, Vosshenrich CA, Colucci F, Richard O, Ndiaye D et al. (2004). Blood 103: 45734580. Sato S, Fujita N, Tsuruo T. (2000). Proc Natl Acad Sci USA 97: 1083210837. Sato S, Sanjo H, Takeda K, Ninomiya-Tsuji J, Yamamoto M, Kawai T et al. (2005). Nat Immunol 6: 10871095. Scheidereit C. (2006). Oncogene 25: 66856705. Schiemann B, Gommerman JL, Vora K, Cachero TG, Shulga-Morskaya S, Dobles M et al. (2001). Science 293: 21112114. Schmidt-Supprian M, Tian J, Ji H, Terhorst C, Bhan AK, Grant EP et al. (2004). J Immunol 173: 16121619. Schwarz EM, Krimpenfort P, Berns A, Verma IM. (1997). Genes Dev 11: 187197. Senftleben U, Cao Y, Xiao G, Greten FR, Krahn G, Bonizzi G et al. (2001a). Science 293: 14951499. Senftleben U, Li ZW, Baud V, Karin M. (2001b). Immunity 14: 217230. Sentman CL, Shutter JR, Hockenbery D, Kanagawa O, Korsmeyer SJ. (1991). Cell 67: 879888. Serhan CN, Savill J. (2005). Nat Immunol 6: 11911197. Seth RB, Sun L, Ea CK, Chen ZJ. (2005). Cell 122: 669682. Sha WC, Liou H-C, Tuomanen EI, Baltimore D. (1995). Cell 80: 321330. Sharma S, tenOever BR, Grandvaux N, Zhou GP, Lin R, Hiscott J. (2003). Science 300: 11481151. Shim JH, Xiao C, Paschal AE, Bailey ST, Rao P, Hayden MS et al. (2005). Genes Dev 19: 26682681. Shinkura R, Kitada K, Matsuda F, Tashiro K, Ikuta K, Suzuki M et al. (1999). Nat Genet 22: 7477. Shoham S, Huang C, Chen JM, Golenbock DT, Levitz SM. (2001). J Immunol 166: 46204626. Siebenlist U, Brown K, Claudio E. (2005). Nat Rev Immunol 5: 435445. Smith KD, Andersen-Nissen E, Hayashi F, Strobe K, Bergman MA, Barrett SL et al. (2003). Nat Immunol 4: 12471253. Snapper CM, Rosas FR, Zelazowski P, Moorman MA, Kehry MR, Bravo R et al. (1996a). J Exp Med 184: 15371541. Snapper CM, Zelazowski P, Rosas FR, Kehry MR, Tian M, Baltimore D et al. (1996b). J Immunol 156: 183191. Sommer K, Guo B, Pomerantz JL, Bandaranayake AD, Moreno-Garcia ME, Ovechkina YL et al. (2005). Immunity 23: 561574. Stetson DB, Medzhitov R. (2006). Immunity 24: 93103. Su TT, Guo B, Kawakami Y, Sommer K, Chae K, Humphries LA et al. (2002). Nat Immunol 3: 780786. Sun L, Deng L, Ea CK, Xia ZP, Chen ZJ. (2004). Mol Cell 14: 289301. Sun Z, Arendt CW, Ellmeier W, Schaeffer EM, Sunshine MJ, Gandhi L et al. (2000). Nature 404: 402407. Tada K, Okazaki T, Sakon S, Kobarai T, Kurosawa K, Yamaoka S et al. (2001). J Biol Chem 276: 3653036534. Takaesu G, Surabhi RM, Park KJ, Ninomiya-Tsuji J, Matsumoto K, Gaynor RB. (2003). J Mol Biol 326: 105115. Takeshita F, Leifer CA, Gursel I, Ishii KJ, Takeshita S, Gursel M et al. (2001). J Immunol 167: 35553558. Takeuchi O, Hoshino K, Akira S. (2000). J Immunol 165: 53925396. Takeuchi O, Kawai T, Muhlradt PF, Morr M, Radolf JD, Zychlinsky A et al. (2001). Int Immunol 13: 933940. Takeuchi O, Sato S, Horiuchi T, Hoshino K, Takeda K, Dong Z et al. (2002). J Immunol 169: 1014. Tanabe T, Chamaillard M, Ogura Y, Zhu L, Qiu S, Masumoto J et al. (2004). EMBO J 23: 15871597. Tato CM, Mason N, Artis D, Shapira S, Caamano JC, Bream JH et al. (2006). Int Immunol 18: 505513. Thomassen E, Renshaw BR, Sims JE. (1999). Cytokine 11: 389399. Thome M. (2004). Nat Rev Immunol 4: 348359. Ting AT, Pimentel-Muinos FX, Seed B. (1996). EMBO J 15: 61896196. Vincenti MP, Brinckerhoff CE. (2002). Arthritis Res 4: 157164. Vincenti MP, Coon CI, Brinckerhoff CE. (1998). Arthritis Rheum 41: 19871994. Voll RE, Jimi E, Phillips RJ, Barber DF, Rincon M, Hayday AC et al. (2000). Immunity 13: 677689. Wald D, Qin J, Zhao Z, Qian Y, Naramura M, Tian L et al. (2003). Nat Immunol 4: 920927. Wan YY, DeGregori J. (2003). Immunity 18: 331342. Ward C, Chilvers ER, Lawson MF, Pryde JG, Fujihara S, Farrow SN et al. (1999). J Biol Chem 274: 43094318. Weih DS, Yilmaz ZB, Weih F. (2001). J Immunol 167: 19091919. Weih F, Carrasco D, Durham SK, Barton DS, Rizzo CA, Ryseck R-P et al. (1995). Cell 80: 331340. Weil R, Schwamborn K, Alcover A, Bessia C, Di Bartolo V, Israe l A. (2003). Immunity 18: 1326. Werner SL, Barken D, Hoffmann A. (2005). Science 309: 18571861. Wessells J, Baer M, Young HA, Claudio E, Brown K, Siebenlist U et al. (2004). J Biol Chem 279: 4999550003. Wetzler LM. (2003). Vaccine 21(Suppl 2): S55S60. Wolfs TG, Buurman WA, van Schadewijk A, de Vries B, Daemen MA, Hiemstra PS et al. (2002). J Immunol 168: 12861293. Wu L, DAmico A, Winkel KD, Suter M, Lo D, Shortman K. (1998). Immunity 9: 839847. Wu M, Lee H, Bellas RE, Schauer SL, Arsura M, Katz D et al. (1996). EMBO J 15: 46824690. Wyllie DH, Kiss-Toth E, Visintin A, Smith SC, Boussouf S, Segal DM et al. (2000). J Immunol 165: 71257132. Xia Y, Makris C, Su B, Li E, Yang J, Nemerow GR et al. (2000). Proc Natl Acad Sci USA 97: 52435248. Xiao C, Shim JH, Kluppel M, Zhang SS, Dong C, Flavell RA et al. (2003). Genes Dev 17: 29332949. Xiao G, Harhaj EW, Sun S-C. (2001). Mol Cell 7: 401409. Xu LG, Wang YY, Han KJ, Li LY, Zhai Z, Shu HB. (2005). Mol Cell 19: 727740. Yamamoto M, Sato S, Hemmi H, Sanjo H, Uematsu S, Kaisho T et al. (2002). Nature 420: 324329. Yamamoto M, Sato S, Hemmi H, Uematsu S, Hoshino K, Kaisho T et al. (2003). Nat Immunol 4: 11441150. Yarovinsky F, Zhang D, Andersen JF, Bannenberg GL, Serhan CN, Hayden MS et al. (2005). Science 308: 16261629. Yeh WC, Shahinian A, Speiser D, Kraunus J, Billia F, Wakeham A et al. (1997). Immunity 7: 715725. Yilmaz ZB, Weih DS, Sivakumar V, Weih F. (2003). EMBO J 22: 121130. Yoneyama M, Kikuchi M, Matsumoto K, Imaizumi T, Miyagishi M, Taira K et al. (2005). J Immunol 175: 28512858. Yoneyama M, Kikuchi M, Natsukawa T, Shinobu N, Imaizumi T, Miyagishi M et al. (2004). Nat Immunol 5: 730737. Yujiri T, Ware M, Widmann C, Oyer R, Russell D, Chan E et al. (2000). Proc Natl Acad Sci USA 97: 72727277. Zelazowski P, Carrasco D, Rosas FR, Moorman MA, Bravo R, Snapper CM. (1997). J Immunol 159: 31333139.
Oncogene

NF-jB and immunology MS Hayden et al

6780 Zhang D, Zhang G, Hayden MS, Greenblatt MB, Bussey C, Flavell RA et al. (2004). Science 303: 15221526. Zhang G, Ghosh S. (2002). J Biol Chem 277: 70597065. Zhang SQ, Kovalenko A, Cantarella G, Wallach D. (2000). Immunity 12: 301311. Zhou H, Wertz I, ORourke K, Ultsch M, Seshagiri S, Eby M et al. (2004). Nature 427: 167171.

Oncogene