roles of the nf-kb pathway in lymphocyte development and...

Roles of the NF-kB Pathway in LymphocyteDevelopment and Function

Steve Gerondakis1 and Ulrich Siebenlist2

1The Burnet Institute, 85 Commercial Rd, Prahran, Victoria 3004, Australia2Laboratory of Immunoregulation, National Institutes of Allergy and Infectious Diseases,National Institutes of Health, Bethesda, Maryland 20892

Correspondence: [email protected], [email protected]

This article focuses on the functions of NF-kB that vitally impact lymphocytes and thus adap-tive immunity. NF-kB has long been known to be essential for many of the responses ofmature lymphocytes to invading pathogens. In addition, NF-kB has important functions inshaping the immune system so it is able to generate adaptive responses to pathogens. Inboth contexts, NF-kB executes critical cell-autonomous functions within lymphocytes aswell as within supportive cells, such as antigen-presenting cells or epithelial cells. It isthese aspects of NF-kB’s physiologic impact that we address in this article.

CELL-AUTONOMOUS ROLES OF NF-kB INLYMPHOCYTE DEVELOPMENT

NF-kB makes numerous cell-autonomous con-tributions to the development of mature Tand B lymphocytes. Given the importance ofNF-kB in adaptive immune responses mediatedby mature lymphocytes, it seems prudent for de-veloping lymphocytes to have adopted a strategyin which their maturation hinges on a properlyfunctioning NF-kB system. As discussed later,the primary, though not exclusive, contributionof NF-kB to lymphocyte development is to as-sure cell survival. These antiapoptotic functionsof NF-kB remain crucial for the health of lym-phocytes even after they mature. Unfortunately,these functions also aid tumorigenesis whenNF-kB is dysregulated (Vallabhapurapu andKarin 2009). Most of the insights about the

role of NF-kB in development of lymphocyteshave come from analyses of genetically manipu-lated mice in which NF-kB components aremissing or in which NF-kB activation has beencompromised or is constitutively induced. Band T lymphocytes will be discussed in parallelto highlight similarities at related stages of theirdevelopment. Figures 1 and 2 summarize someof the findings described later.

Early Lymphocyte Progenitors

Despite a clear role for the NF-kB homolog Dor-sal in early Drosophila development (Hong et al.2008), in mammalian development no such rolefor NF-kB has emerged, including developmentof early lymphocyte precursors. NF-kB can,however, play a protective role in precursors to

Editors: Louis M. Staudt and Michael Karin

Additional Perspectives on NF-kB available at www.cshperspectives.org

Copyright # 2010 Cold Spring Harbor Laboratory Press; all rights reserved; doi: 10.1101/cshperspect.a000182Cite this article as Cold Spring Harb Perspect Biol 2010;2:a000182

1

on March 29, 2021 - Published by Cold Spring Harbor Laboratory Press http://cshperspectives.cshlp.org/Downloaded from

http://cshperspectives.cshlp.org/

protect them from TNFa-induced apoptosis.Artificially high levels of TNFa arise duringadoptive transfers of hematopoietic stem cellsinto lethally irradiated hosts, so when donor cellswere compromised in their ability to activateNF-kB, reconstitution of lymphocytes failed

(Grossmann et al. 2000; Senftleben et al.2001b; Claudio et al. 2006; Gerondakis et al.2006; Igarashi et al. 2006). It is possible that aminimum of NF-kB activity may yet be neces-sary even during normal development (normallevels of TNFa). Female mice heterozygous for

DN III b

DN IV

DP

DN III a

gd TCR

Pre-TCR(pTa/Tb)

Classical NF-kBsurvival

ab TCR

Tarearrangement

Positive selection

MHC Iself-AGs

MHC IIself-AGsNegative

selection

TSAs

Classic

al NF-k

B - su

rvival

CD8SP

High self-reactivityhigh NF-kB

CD4SP

CD4 SPFoxP3+

treg

CD1dself-AGs

Intermediate self-reactivity

classical NF-kBNKT

No self-reactivity(neglect)

Thymuscortex

Medulla

NF-kB?

High self-reactivity

Artificial

high NF-kB

RankCD40LTbR

Classical andalternative NF-kB

Self-AGson DCs

mTEC

PeripheryNF-kB required to maintain all T cells m

TEC positive selection?

DN II

Figure 1. NF-kB in thymic T-cell development. Shown is a schematic and simplified representation of thymicT-cell development, highlighting stages at which NF-kB contributes in a cell-autonomous fashion. Alsohighlighted is the requirement of NF-kB for generation of medullary thymic epithelial cells (mTECs). gd Tcells and Tb-expressing thymocytes can be distinguished at the (CD4, CD8) double negative (DN) stage IIIb. The pre-TCR (pTa/Tb) drives development of thymocytes into DN IV cells, which in turn give rise todouble positive (DP) cells (abTCR). Positive selection of DP thymocytes to become CD4 or CD8 single-positive (SP) thymocytes is driven by weak recognition of self-AGs presented on cortical thymic epithelialcells in the context of MHC class II or class I, respectively. T-regulatory cells (Tregs, FoxP3þ) and NKT cellsmay develop from DP thymocytes by recognition of self-AGs with intermediate strength (lipids presented onCD1d in the case of NKT cells). Failure to recognize self-AGs leads to elimination of thymocytes (death byneglect); strong recognition of self-AGs also leads to elimination (negative selection). Negative selectionbegins in the cortex but may occur predominantly in the medulla, where self-AGs are presented on dendriticcells (DCs) and on mTECs. mTECs produce tissue-specific (self )-AGs (TSAs) and can cross-prime DCs withthese antigens. See text for further details.

S. Gerondakis and U. Siebenlist

2 Cite this article as Cold Spring Harb Perspect Biol 2010;2:a000182



Pro-B Late pro-B/large pre-B

Pre-BCR(m/VpreB+l5)

Small pre-BClassical NF-kB

survival

k rearrangementCl

assic

al NF

-kB

surv

ival?

l rearrangement

Immature B

BCR(m/k or m/l)

Self-AG reactivitynegative selection – editing

or elimination

Editing correlates withNF-kB activation

‘Mature’immature B

Classical and

alternative NF-kBsurvival

Bone marrow

BAFFR +?

T1 BT2 B

Spleen(white pulp)

Marginalzone B Cell

follicle

FDC

LTbRFM (B2)

MZB

Stromal cell

Classical and alternative NF-kBBAFFR + BCR +?

survival +

PeripheryNF-kB required tomaintain all B cells

B1 B cells

Fetalliver

Classical(alternative ?)

NF-kB

Lymph nodesPeyer’s patches

classical and alternativeNF-kB in stromal cells

(Rank, LTbR)

Classical and alternativeNF-kB

Self-AGreactivity

LTbRTNFR +

Figure 2. NF-kB in B cell development. A schematic and simplified representation of bone marrow and splenicB-cell development, highlighting stages at which NF-kB contributes in a cell-autonomous fashion to formationof marginal zone (MZ B) and follicular mature (FM) B cells; the latter are also known as B2 B cells and enter theperipheral circulation. Also highlighted is the requirement for NF-kB in B1 B-cell development, a peripherallyself-renewing population with precursors in fetal liver and possibly bone marrow. Also highlighted is theimportance of NF-kB in stromal cells/follicular dendritic cells (FDCs) in forming a proper splenicarchitecture (B-cell follicles, marginal zone) and in forming Peyer’s patches and lymph nodes. B-celldevelopment commences in the bone marrow, where the pre-BCR on large pre-B cells (a.k.a. late pro-B)drives development into small (late) pre-B cells, which in turn give rise to immature B cells (first to express afull BCR [IgM]). Self-antigen (AG)-reactive immature B cells edit their receptors by further light chain generearrangements or they are eliminated (negative selection). Surviving immature and “more mature”immature B cells (T2-like) then migrate to the spleen (white pulp), where they progress through thetransitional 1 (T1) and T2 stages to become FM (located in B cell follicles) and MZB cells (located inmarginal zones). Early transitional-staged cells continue to be subject to negative selection. Generation ofFMs and MZBs from early transitional stages is driven by signals from the BCR, BAFFR, and other receptors(“þ”) to assure survival, but also to regulate cell differentiation (“þ”) in part via NF-kB. (Additional minorpathways and populations have been postulated, but are not shown here.) See text for further details.

NF-kB and Lymphocytes

Cite this article as Cold Spring Harb Perspect Biol 2010;2:a000182 3



loss of X-chromosome-encoded NEMO onlygenerated NEMO sufficient, but not NEMO-de-ficient lymphocytes, even though random lyoni-zation should have generated equal numbers(Makris et al. 2000; Schmidt-Supprian et al.2000). NEMO (IKKg) is an essential componentof the classical pathway for NF-kB activation,and absolutely required for NF-kB activationby TNFa (reviewed in Hayden and Ghosh2008; Vallabhapurapu and Karin 2009). It re-mains to be shown, however, if NEMO-deficientlymphocyte precursors were indeed eliminatedby “normal” levels TNFa

Pre-antigen Receptor Expressing LargePre-B Cells and Double Negative (DN)Thymocytes

The appearance of pre-TCRs on DN thymocytes(stage III) and pre-BCRs on developing bonemarrow large pre-B cells provides important,ligand-independent signals for expansion andprogression to the DN stage IV/DP stage andto small pre-B cells, respectively (T- and B-celldevelopment reviewed in Bommhardt et al.2004; Hardy et al. 2007; Allman and Pillai2008; Northrup and Allman 2008; Taghon andRothenberg 2008). These cells contain signifi-cant levels of nuclear NF-kB activity, presum-ably because of activation by the pre-antigenreceptors (Voll et al. 2000; Jimi et al. 2005; Der-udder et al. 2009). Failure to assemble a pre-TCR receptor (pTa paired with rearrangedTRCb) or pre-BCR (VpreB and g5 surrogatelight chains paired with rearranged m heavychain) eliminates these cells.

IkB super-repressor-mediated interferencewith NF-kB activation in pre-TCR expressingthymocytes led to their apoptosis (Voll et al.2000), because of interference with NF-kB-mediated induction of the antiapoptotic Bcl-2family member A1 (Mandal et al. 2005; Aifantiset al. 2006). Furthermore, exogenous expressionof a constitutively active IKKb allowed some DNthymocytes to progress to the double positive(DP) stage, even in the absence of a pre-TCR inRAG-deficient mice (Voll et al. 2000). Similarly,in pre-BCRþ large pre-B cells, suppression ofNF-kB by the IkB super-repressor also resulted

in apoptosis, which could be overcome byectopic expression of Bcl-xL, another NF-kB-induced antiapoptotic member of the Bcl-2family (Feng et al. 2004a; Jimi et al. 2005). At aminimum, these data imply a survival role forNF-kB in both pre-BCR and pre-TCR depend-ent development of B and T lymphocytes (thishas previously been reviewed [Denk et al. 2000;Siebenlist et al. 2005; Claudio et al. 2006; Vallab-hapurapu and Karin 2009]).

Small (Late) Pre-B Cells

Following cellular activation via pre-BCRs, cellsno longer express the receptor and enter thesmall pre-B cell stage. On successful rearrange-ment of their light chains, they begin to expressa full BCR (IgM) on their surface and aretermed immature B cells (Hardy et al. 2007;Northrup and Allman 2008). Circumstantialevidence has implicated NF-kB in demethyla-tion of the k light chain (LC) locus, an obligatestep during k gene rearrangement (Goldmitet al. 2005). However, loss of the intronic k en-hancer containing the first known kB bindingsite had no effect on this process (Inlay et al.2004; Sen 2004), and a recent report impliesno or only a minor role for NF-kB in k locusdemethylation (Derudder et al. 2009). mb1promoter-driven Cre-induced conditional abla-tion of NEMO (eliminating the classical path-way) or of both IKK kinase subunits (IKK1and IKK2; eliminating classical and alternativeNF-kB pathways) at this stage of developmentdid not significantly reduce k LC expressing(immature) B cells. Surprisingly though, l LCexpressing immature B cells were reduced. Rear-rangement of the l locus is temporally delayedrelative to that of k, so it is possible that theseB cells require NF-kB activity to allow them tosurvive longer. In support, the antiapoptoticprotein kinase Pim2 was down-modulated inNEMO/IKK deficient small B cells and ectopicexpression of a Bcl-2 transgene resurrected l LCrearrangements.

It is not known what signals activate NF-kB atthis stage. CD40, Bcl-10, MyD88, and ATM kin-ase were not required (Derudder et al. 2009);the latter was thought to be potentially involved





in NF-kB activation and developmental pro-gression in response to double-stranded DNAbreaks associated with DNA rearrangements(Bredemeyer et al. 2008). Because Bcl-10 wasnot required, one might conclude that antigen-receptor signaling is not involved, but it is possi-ble that this receptor can activate NF-kB by anunconventional pathway (see further discussionlater). Loss of TRAF6 did reduce the number ofl-expressing B cells (Derudder et al. 2009), butmany signaling receptors in addition to the anti-gen receptor avail themselves of this adaptor.

Immature (Bone Marrow) B Cells andNegative Selection

Autoreactive immature (bone marrow) B cellsare subject to negative selection, but ratherthan undergo immediate apoptosis, a portionof these cells reactivate the RAG recombinaseto further rearrange (edit) their light chain inhope of generating a nonautoreactive BCR(Nemazee 2006). Strong engagement of self-antigens was reported to down-regulate BCRcell surface expression and de-differentiate im-mature B cells back to the small pre-B cell stage(Schram et al. 2008). Contrary to an earlier re-port that concluded a critical role for NF-kB inRAG regulation during negative selection (Ver-koczy et al. 2005), a recent report instead at-tached an at best only minor role to NF-kB:receptor editing for k light chains and RAG ex-pression levels were not significantly affected inB cells conditionally ablated for NEMO or forboth IKK1 and IKK2, although as noted previ-ously, l light chain rearrangement was affected(Derudder et al. 2009). On the other hand, an-other recent study found a strong correlation be-tween elevated NF-kB activity (induced byautoreactive BCRs) and various indicators of re-ceptor editing, suggesting a causal relationship,even though these authors also failed to find alink between NF-kB and RAG expression (Ca-dera et al. 2009). These divergent conclusionsmight yet be reconciled if for example someNF-kB could be activated by BCRs via aNEMO/IKK independent mechanism.

NF-kB may contribute to the survival ofimmature cells during the negative selection/

editing phase and thereafter. Initial in vitrostudies with immature B-like WEHI321 cellssuggested as much; a-IgM induced apopotisof these cells could be prevented if cells werealso stimulated with a-CD40, which led to sus-tained NF-kB activity, and in turn, higherc-myc levels (Schauer et al. 1998). CD40 acti-vates both the classical and the alternative (non-classical, noncanonical) pathways for NF-kB(Mineva et al. 2007; Hayden and Ghosh 2008;Vallabhapurapu and Karin 2009). We recentlydocumented a partial reduction in numbers ofimmature bone marrow B cells in mice lackingboth NF-kB1 and NF-kB2 (this partiallyeliminates classically and alternatively activatedcomplexes) (Claudio et al. 2009). This losscorrelated with impaired survival of mutant im-mature B cells. Activation of the alternativepathway in wild-type immature B cells is likelyto occur via BAFF receptor signaling (Claudioet al. 2009), which was also suggested by anotherstudy that implicated BAFF receptor signalingin the de novo generation of a small subset of“more mature” immature B cells in bone mar-row (Lindsley et al. 2007). The previously citedreport on conditional ablation of NEMO or ofIKK1/IKK2 in B cells also showed a reductionin the more mature immature B cells (Derudderet al. 2009).

DP and SP “Conventional” Thymocytes:Positive and Negative Selection

After “b” selection of pre-TCRþDN thymocytes(stage III), these cells rearrange their TCRa gene(DN stage IV) and become DP (CD4þ CD8þ)TCRabþ thymocytes (Bommhardt et al. 2004;Taghon and Rothenberg 2008). At this time,the thymocytes are subjected first to positiveand then, in an overlapping fashion, to negativeselection, while they physically move from thethymic cortex to the medulla (Boehm 2008).TCRs that fail to recognize self-antigens pre-sented by MHCs are deleted (neglect), whereasweakly self-reactive TCRs are positively selectedand strongly self-reactive TCRs are negatively se-lected (death by apoptosis). Surviving thymo-cytes become CD4 or CD8 single positive (SP)(depending on recognition of MHC class II or I,





respectively), undergo further maturation, andeventually exit the thymus to enter the peripheralcirculation as mature naı̈ve T cells (Bommhardtet al. 2004; Taghon and Rothenberg 2008).

The role of NF-kB in selection remainssomewhat controversial; NF-kB appears tohave both positive and negative effects, probablybecause TCR signal strengths can be sensedwithin cells via the level of NF-kB activation;too much or too little activity may be selectedagainst. Early studies with mice expressing IkBsuper-repressor transgenes have suggested arole for NF-kB in positive selection of CD8,and, to a lesser degree, of CD4 SP thymocytes,including studies with mouse models express-ing TCR transgenes that weakly recognize self-peptides presented on class I or class II MHCs,respectively (positive selection) (Boothby et al.1997; Hettmann and Leiden 2000; Mora et al.2001). On the other hand, NF-kB has also beenimplicated in promoting apoptosis of DP thy-mocytes mimicked by a-CD3 treatment in vivo(Hettmann et al. 1999; Ren et al. 2002) and inmouse models with TCRs that strongly recognizeself-peptides presented on class I or class II (neg-ative selection) (Mora et al. 2001), although onestudy failed to see a role for NF-kB in class I neg-ative selection (Hettmann and Leiden 2000).

In these studies, it is difficult to differentiateroles of NF-kB during the initial positive/nega-tive selection from roles during the subsequent,SP stage, although a more recent study impli-cates NF-kB in the latter stage, especially inCD8 SP thymocytes. Mice conditionally ablatedin thymocytes for NEMO or that conditionallyexpress a kinase deficient (potentially domi-nant negative-acting) form of IKKb in thymo-cytes generated significantly fewer CD8 SPthymocytes, and had no peripheral T cells, in-cluding CD4 T cells (Schmidt-Supprian et al.2003). However, straight loss of IKKb was welltolerated, suggesting compensation by IKKa.This study only informed on IKK/NF-kB inSP but not DP thymocytes, because the lck- orCD4-driven genetic changes only became fullypenetrant after thymocytes had traversed theDP stage. Those SP thymocytes that remainedin this NF-kB impaired model exhibitedincreased expression of apoptotic markers,

consistent with an antiapoptotic role forNF-kB. This study was further able to concludethat NF-kB was continuously required in T cellseven after they emerged from the thymus, i.e.,the maintenance of the peripheral populationdepended on NF-kB. A very recent study hasadded further insight, strongly implicatingNF-kB in positive and negative selection ofCD8 (DP and SP stage), but surprisingly, notof CD4 thymocytes. This study relied on mousemodels with transgenes featuring positivelyor negatively selectable TCRs as well as anIkBa super-repressor or a constitutively activemutant IKK2 kinase (Jimi et al. 2008). IkBasuper-repressor mediated inhibition of NF-kBeliminated cells expressing TCRs that weaklyrecognized MHC class I-presented self-peptides(CD8 positive selection), while also preventingthe elimination of cells expressing TCRs thatstrongly recognized class I-presented self-pepti-des (CD8 negative selection). Confirming a rolefor strong NF-kB activity in negative selection,high NF-kB activity induced by constitutivelyactive IKK2 eliminated some cells during posi-tive selection for CD8, presumably mimickinga strong TCR signal that pushed cells into a neg-ative selection mode. Thus, the level of NF-kBactivity appeared to reflect TCR signal strengthduring CD8 selection, thereby setting thresh-olds for positive and negative selection. A lowlevel of NF-kB activity during CD8 positive se-lection allowed cells to avoid death by neglect,whereas a high level was needed during CD8negative selection to induce cell death. Interest-ingly, CD8 but not CD4 thymocytes expressedsignificant levels of NF-kB, which may explainwhy the IkBa super-repressor had little to no ef-fect on CD4 selection processes. However, con-stitutively active IKK2 also eliminated cellsduring positive selection for CD4, presumablypushing these cells into a “pseudo”-negative se-lection mode. CD8 T cells may depend on aminimum of NF-kB for survival because theyhave very low levels of Bcl-2, whereas CD4 Tcells express high levels of this antiapoptoticregulator, rendering them independent ofNF-kB’s antiapoptotic functions. It is surpris-ing that high levels of NF-kB activity promotedcell death during negative selection, given the





almost universal linkage of NF-kB with survivalfunctions; how this is accomplished remains aninteresting, yet open question. It will also be ofinterest to determine how DP thymocytes dif-ferentially regulate for NF-kB depending onwhether their TCRs recognize self-peptides inthe context of class I or class II MHCs; thismay involve other signals coming from anti-gen-presenting cells. The IkBa super-repressortransgene models cited previously do have oneinherent shortcoming; NF-kB inhibition islikely to be only partial, with the degree of in-hibition dependent on the level of expressionof the transgene, which may vary between cellsand models.

How NF-kB is activated in thymocytes dur-ing positive and negative selection is not fullyunderstood. If TCR signaling is responsible,then it must activate NF-kB by an “unconven-tional” pathway, given obligate componentsof the “conventional” pathway (as defined inmature T cells) could be largely dispensed withfor generation of more mature SP thymocytes/peripheral T cells. Regardless of this finding, theconventional pathway is likely to contribute toNF-kB activation during pre-TCR signaling(Sun et al. 2000; Lin and Wang 2004; Thome2004; Felli et al. 2005; Jost et al. 2007). Obligatecomponents of the conventional pathways in-clude PKCu and “CBM” complex componentsCarma1, Bcl-10, and Malt1 (PKCu may havean NF-kB-independent role though [Morleyet al. 2008]). Interestingly, a recent study reportsthe existence of an unconventional signalingpath for TCR-mediated activation of NF-kB inmature CD8, but not CD4 T cells (Kingeterand Schaefer 2008); if such a pathway was alsofunctional during selection of CD8 thymocytes,it would be consistent with the role for NF-kB inthe development of these cells as discussed pre-viously. Alternatively, NF-kB may also beactivated by other signals, independent of theTCR. In either case, Tak1 is most likely involved;Tak 1 functions just upstream of the IKK com-plex in response to many NF-kB activating sig-nals, and loss of this kinase reduced thenumber of SP thymocytes, and especially pe-ripheral T cells, because of increased apoptosis(Liu et al. 2006; Sato et al. 2006; Wan et al. 2006).

Nonconventional T Cells

T-regulatory (Treg) cells recognize self-antigens,yet escape negative selection, possibly becausethey recognize self-antigens with intermediatestrength (Caton et al. 2004; Lu and Rudensky2009). Invariant natural killer T cells (iNKT;a.k.a. NKT) also recognize self-antigens (alongwith foreign ones), although in this case theantigens are lipids presented to developingNKT cells by CD1d on thymocytes (MacDonaldand Mycko 2007; Burrows et al. 2009). gd Tcells are also thought to recognize certain self-antigens, but not in the context of MHC orCD1d (Thedrez et al. 2007; Xiong and Raulet2007; Taghon and Rothenberg 2008). Auto-antigen-mediated TCR signaling may thus pro-mote the development of all of these cell types,and it may do so in an NF-kB dependent manner.

The conditional loss of IKKb in thymocytesprevented the emergence of Treg and NKT cells(Schmidt-Supprian et al. 2003, 2004a); there-fore, IKKa was unable to compensate, eventhough it did during development of conven-tional T cells. Furthermore, a compound defi-ciency of c-Rel and NF-kB1 abrogated Treggeneration (Zheng et al. 2003), while loss ofRelA or the presence of the IkBa super-repressorblocked NKT development (Sivakumar et al.2003; Vallabhapurapu et al. 2008).

The commitment of early thymocytes tothe Treg lineage (as determined by FoxP3 ex-pression) was fully dependent on TCR-inducedconventional signaling for activation of NF-kB(PKCu, CBM, and Tak1 were all required)(Schmidt-Supprian et al. 2004a; Sato et al.2006; Wan et al. 2006; Barnes et al. 2009; Medoffet al. 2009; Molinero et al. 2009). For NKT cells,PKCu and the CBM adaptor complex were notabsolutely required, although PKCu contributedto thymic and Bcl-10 to peripheral NKT-cellnumbers, respectively (Schmidt-Supprian et al.2004a; Medoff et al. 2009). On the other hand,stimulation of thymocytes via NKT-specificVa14i TCRs was reported to activate RelA/NF-kB complexes and thereby induce exp-ression of IL-15 receptor a and the commongchain (gc), which in turn allowed for IL-15and IL-7 mediated expansion of NK1.12 NKT





precursors and maturation to NK1.1þ NKTcells (Vallabhapurapu et al. 2008). It is possiblethat TCRs on NKT activated NF-kB by a non-conventional pathway; alternatively, TCR inde-pendent signals might also have contributedto NF-kB activation.

Like NKTs, gd T cells have innate and adap-tive functions; they are able to rapidly producegIFN or IL-17 in response to a variety of stimuli(Taghon and Rothenberg 2008; Jensen andChien 2009). Relatively little is known abouttheir development, but recently their differen-tiation into gIFN producing cells was reportedto be imprinted early in developing gd thymo-cytes, correlating with expression of TNF recep-tor members CD27 and LTbRs (Ribot et al.2009). The latter is thought to be engaged byLTab-expressing DP thymocytes. The presenceof CD27 and LTbR suggests roles for the classi-cal and alternative pathways in generating gIFNproducing gd T cells, although this remains tobe shown.

Transitional and Follicular Mature B Cellsand Nonconventional Marginal Zone andB1 B Cells

Immature B cells eventually leave the bone mar-row and migrate to the spleen to complete theirmaturation process. There they undergo severalphenotypic and functional changes to becomefollicular mature (FM) B cells (a.k.a. B2 B cells),which recirculate in the periphery, and marginalzone B cells (MZB), which are largely sessile,although they may shuttle between the marginalzone and follicles to transport and present anti-gens (Thomas et al. 2006; Hardy et al. 2007;Allman and Pillai 2008). In this way, MZBs con-tribute to T-dependent antigen responses, butMZBs can also rapidly respond to pathogensin a T-independent, innate-like fashion. MZBsare a first line of defense to blood-borne patho-gens, as splenic marginal zones filter thesepathogens out of the blood stream (Lopes-Car-valho et al. 2005). B1 B cells represent a distinctlineage from B2 B cells; they populate the peri-toneal and pleural cavities and are thought tooriginate from fetal liver precursors that self-re-new in the periphery, although some precursors

may reside in the bone marrow (Allman andPillai 2008). Like MZBs, B1 B cells participatein rapid T-independent innate-like antibodyproduction and appear to recognize some self-antigens.

Newly arrived immature B cells in the spleenare referred to as transitional-1 B cells (T1);these cells continue to be subject to negativeselection pressures, as antigen-receptor stimu-lation induces apoptosis. At some time duringthe T2 stage, B cells instead begin to respondpositively to antigen-receptor stimulation,whereupon they become FM and MZB cells.The mechanisms underlying the divergence ofFM and MZB remain poorly understood. Insome studies, an additional third transitionalB cell stage has been distinguished phenotypi-cally, but this population may also includeanergic B cells (Thomas et al. 2006; Hardyet al. 2007; Allman and Pillai 2008).

NF-kB is absolutely essential for the sur-vival of developing B cells in the spleen. Com-pound deficiencies in NF-kB1/NF-kB2 andc-Rel/RelA profoundly arrested B-cell develop-ment at or shortly after the transition from T1to T2, and resulted in a near-complete absenceof FM and MZB cells (Franzoso et al. 1997b;Grossmann et al. 2000; Claudio et al. 2002,2006; Gerondakis et al. 2006); a milder reduc-tion of FM B cells occurred with compounddeficiencies of NF-kB1/c-Rel (Pohl et al. 2002;Gerondakis et al. 2006) and NF-kB2/Bcl-3 orNIK/Bcl-3 (U.S., unpubl. results). These resultssuggest involvement of the classical and alterna-tive pathways. Loss of alternative pathway com-ponents only (NF-kB2, NIK, or IKKa) resultedin partial reductions of more mature B cells(Franzoso et al. 1998; Yamada et al. 2000; Kaishoet al. 2001; Senftleben et al. 2001a; Siebenlistet al. 2005; Claudio et al. 2006). Bcl-3 is an atyp-ical member of the IkB family and its involve-ment was unexpected. Bcl-3 may be inducedby classical NF-kB activation and it is likely tomodulate gene transcription via associationwith DNA-binding p50 homodimers in thenucleus, although biologic targets during B-celldevelopment are unknown (Palmer and Chen2008; Yamamoto and Takeda 2008). The severeblock in B-cell development in the compound





deficient NF-kB models cited previously wasbecause of loss of survival; early transitional Bcells (T1) from these mutant mice expressedlower levels of the antiapoptotic proteins Bcl-2and A1, and exhibited increased spontaneousapoptosis in culture. Ectopic expression ofBcl-2 allowed the transitional B cells to surviveand progress further; however, these cellsdid not fully mature phenotypically or func-tionally, remaining unable to produce antibod-ies. Therefore, NF-kB is critical not onlyfor survival of transitional B cells, but also forthe complete development of functions thataccompanies differentiation.

During the transitional phases, BAFF (a.k.a.BLyS, TNFSF13B) induces the alternative path-way in B cells via the BAFF receptor (BAFFR,a.k.a. BR3, TNFRSF13C) (Claudio et al. 2002;Stadanlick and Cancro 2008; Mackay andSchneider 2009; Rauch et al. 2009). Becauseloss of BAFFR blocks B-cell maturation moreprofoundly than loss of the alternative pathway,BAFFR must contribute additional signals forcell survival. Although the alternative pathwaydoes promote Bcl-2 expression, and, indirectly,the retention in the cytoplasm of otherwiseapoptotic nuclear PKCd (Sasaki et al. 2006),BAFFR also activates AKT, independently ofNF-kB (Otipoby et al. 2008). Interestingly, amutant form of the alternative pathway compo-nent p100/NF-kB2 that cannot be processed top52 caused a substantial block in T1 to T2/Mmaturation (Tucker et al. 2007); the mutantp100 eliminates the alternative NF-kB activa-tion, but, in addition, it also inhibits p65(RelA) complexes, which are normally activatedby the classical pathway (Kanno et al. 1994).

Although the alternative pathway contrib-utes to B-cell maturation, the classical pathwayis absolutely required. This conclusion was al-ready apparent from the results with c-Rel,RelA compound deficiency (see previous dis-cussion) and was confirmed with NEMO(IKKg) deficient B cells. Conditional loss ofNEMO in B cells with CD19- or mb1-Crecaused significant arrest at or near the T1 toT2 transition with a near complete absence ofmature B cells; similarly induced conditionalexpression of a kinase deficient form of IKKb

(see previous discussion) caused a milder block,whereas outright (conditional) loss of IKKbcaused no discernable block at the T1 to T2transition (Pasparakis et al. 2002; Sasaki et al.2007; Derudder et al. 2009). However, all three(B-cell specific) genetic changes nearly elimi-nated or at least severely reduced numbers ofmature recirculating B cells. The most straight-forward interpretation of these results is thatduring the transitional phases, IKKa may parti-ally compensate for the outright loss of IKKb toactivate the classical pathway (a similar situa-tion was seen in conventional thymocyte devel-opment discussed previously). On the otherhand, once B cells have matured, they absolutelyrequire IKK2-mediated classical activation forlong-term survival. The most severe block atthe T1 to T2 transition occurred in mice condi-tionally ablated with mb1-Cre for both IKK1and IKK2; the block was more complete thanthat observed in the absence of NEMO (Der-udder et al. 2009). This suggests that IKK1also contributes in a NEMO-independent fash-ion, presumably via the alternative pathway. Ofnote, a constitutively active form of IKK2 wasable to completely overcome the absence ofBAFFR, indicating that the classical pathway isboth necessary and sufficient to allow full matu-ration of B cells in the absence of BAFFR/alter-native activation (Sasaki et al. 2006). However,because the alternative pathway was requiredfor optimal generation of B cells under normalconditions, it must be inferred that the classicalpathway is not normally sufficiently activated.

It has been suggested that weak tonic BCRsignaling activates the classical pathway, whichin turn is required to set up the alternative path-way by inducing expression of BAFFR andNF-kB2 via c-Rel (Stadanlick et al. 2008; Castroet al. 2009). However, BCR-independent signalsmust also be considered, because loss of CBMcomplex components (Carma1, Bcl-10, orMalt1) only mildly reduced FM B-cell numbers(CBM is part of the conventional signaling pathfrom the BCR to NF-kB; see TCR signaling dis-cussed previously) (Thome 2004; Ferch et al.2007). This situation is somewhat reminiscentof what was noted for development of naı̈ve Tcells, and again raises the possibility that BCRs





and TCRs may activate NF-kB in developinglymphocytes by an unconventional pathway.Whatever the pathway or signal, TAK1 was re-ported to be required for B-cell maturation(Schuman et al. 2009), although an earlier re-port appears to be at odds with this finding(Sato et al. 2005), while TRAF6 was reportedto be at least partially required (Kobayashiet al. 2009). However, these proteins can be partof many NF-kB signaling pathways, so theirinvolvement does not inform on the nature ofthe signal.

The generation of MZB cells is particularlysensitive to perturbations in NF-kB activity.Single deficiency in NF-kB1, NF-kB2 or RelBand to a lesser extent c-Rel or RelA in B cellsalready reduced numbers of MZBs (Caamanoet al. 1998; Franzoso et al. 1998; Cariappaet al. 2000; Weih et al. 2001; Guo et al. 2007),although there may be an unexplained reboundwith advancing age in NF-kB1 knockouts (Fer-guson and Corley 2005). Furthermore, loss ofIKK2 essentially eliminated these cells (Paspar-akis et al. 2002). NF-kB1 reportedly synergizeswith Notch2 receptor signaling during genera-tion of MZBs, but mechanisms remain to beelucidated (Moran et al. 2007; Allman and Pillai2008). The exquisite dependence of MZBs onNF-kB activity is also highlighted by mousemodels in which the numbers of these cellswere augmented. CYLD is deubiquitinase forK63-linked ubiquitin chains, which down-modulates the classical pathway by targetingproteins such as NEMO and RIP (Courtois2008). The loss of CYLD in B cells led to amarked increase in constitutive NF-kB activityand a significant rise in MZBs (Jin et al.2007). Overactivation of the alternative pathwayalso led to a selective increase of MZBs, as is thecase in mice with a BAFF transgene (Mackayand Schneider 2009) or in mice deficient inthe IkB-like part of p100/NF-kB2, leaving p52to form complexes with RelB and enter nuclei(Guo et al. 2007). Finally, loss of Bcl-3 unex-pectedly increased MZB cell numbers; thisimplicates Bcl-3 as a gatekeeper during MZBformation (U.S., unpubl. observations). Asdiscussed previously, Bcl-3 is an unusual IkBfamily member; how it normally suppresses

MZB development remains to be determined,although it might do so by dampening NF-kBtarget gene expression.

MZB cells have been suggested to be selectedby expression of weakly self-reactive BCRs thatavoid negative selection but are possibly strongenough to activate some NF-kB to aid in theirsurvival and/or expansion (Allman and Pillai2008). Such a scenario would be consistentwith the fact that components of theconventional BCR signaling path to NF-kBwere required for MZB development (e.g.,Carma1, Malt1, Bcl-10, TRAF6, and Tak1)(Thome 2004; Pappu and Lin 2006; Kobayashiet al. 2009; Schuman et al. 2009).

Formation of B1 B cells is highly dependenton the function of the conventional BCR signal-ing pathway to NF-kB and on NF-kB activity,possibly because these cells are thought to beselected for fairly strong recognition of self-antigens (Allman and Pillai 2008). Loss of com-ponent parts of this pathway (Thome 2004; Satoet al. 2005; Kobayashi et al. 2009; Schuman et al.2009) or interference with NF-kB activity blo-cked development of B1 B cells; for example,deficiency in NF-kB1 already reduced theirnumbers, whereas a deficiency in both c-Reland NF-kB1 nearly eliminated them (FM cellnumbers were only mildly affected) (Pohlet al. 2002; Gerondakis et al. 2006). c-Rel is par-ticularly important for cell-cycle entry and pro-liferation in response to BCR stimulation(Gilmore et al. 2004), suggesting the possibilitythat B1 B cells may expand in response to BCRsignaling induced by self-antigens. As expected,B1 B cells also failed to develop in the absence ofIKK2 (Pasparakis et al. 2002). Interestingly, lossof NF-kB1 and NF-kB2 severely diminished B1B-cell numbers (Claudio et al. 2002), implying apossible role for the alternative pathway of acti-vation, even though BAFFR was not required(Mackay and Schneider 2009; Rauch et al.2009).

NF-kB IN THYMIC EPITHELIAL CELLS

NF-kB activity in epithelial cells is criticalfor central tolerance, i.e., tolerance of self-antigens enforced in the thymus via elimination





(negative selection) of autoreactive TCRs (seeprevious discussion). Negative selection of con-ventional thymocytes with strongly autoreactiveTCRs begins during interaction with self-anti-gen presented on cortical thymic epithelial cells(cTECs; also responsible for positive selection),but the bulk of negative selection is thought totake place in the medulla, where self-antigensare presented by thymic dendritic cells (DCs)and medullary thymic epithelial cells (mTECs)(Kyewski and Klein 2006; Boehm 2008; Nittaet al. 2008). mTECs promiscuously expressmany tissue specific self-antigens (TSAs), suchas insulin and salivary protein 1, driven inpart via the AIRE regulator; mTECs also effi-ciently cross-prime thymic DCs with these self-antigens (Mathis and Benoist 2009). mTECs arelikely required for accumulation of thymic DCsin the medulla (Tykocinski et al. 2008). Finally,mTECs may help promote Treg development, atleast those with specificities to TSAs, althoughthis remains to be shown (Zhang et al. 2007;Koble and Kyewski 2009).

Development/differentiation of mTECs isunder control of NF-kB. In the prenatal stage,generation of mTECs is induced by interactionwith hematopoietic lymphoid tissue inducercells (LTi) (Kim et al. 2009). These cells expressligands for RANK and LTb receptors and there-by activate NF-kB in mTEC progenitors viamatching receptors (Akiyama et al. 2008).Thereafter, continued generation and mainte-nance of the mTEC network is controlled bypositively selected DP thymocytes, which ex-press ligands for RANK, LTb, and CD40 tostimulate mTECs (Akiyama et al. 2008; Hiko-saka et al. 2008). Autoreactive thymocytes inparticular have been suggested to promote thegeneration of mTECs (Irla et al. 2008). mTECsturn over rapidly, express self-antigens opti-mally on terminal differentiation, and cross-prime thymic DCs, most likely in a processlinked to their turnover (Yano et al. 2008; Kobleand Kyewski 2009). The nuclear protein AIREfacilitates expression of a significant portion ofthe promiscuously produced TSAs and it mayalso have a role in terminal differentiationand antigen presentation (Yano et al. 2008;Koble and Kyewski 2009; Mathis and Benoist

2009). Expression of some TSAs is independentof AIRE and instead may be produced by adistinct subclass of mTECs, dependent onLTbR signaling (Martins et al. 2008; Seachet al. 2008).

RANK, CD40, and LTbR are TNF receptorfamily members that activate both classicallyand alternatively activated NF-kB complexes(Ghosh and Hayden 2008; Hayden and Ghosh2008; Vallabhapurapu and Karin 2009). Consis-tent with a critical role for the alternative path-way, loss of NIK, IKKa, RelB (Weih et al. 1995;Kajiura et al. 2004; Kinoshita et al. 2006), orcompound loss of NF-kB2 and Bcl-3 abrogatedthe appearance of differentiated mTECs (Zhanget al. 2007). This in turn engendered multiorganinflammation and early death, directly medi-ated by autoreactive T cells emerging from thethymus (Zhang et al. 2007) (U.S., unpubl. ob-servations). Loss of NF-kB2 alone caused onlymild impairment of mTECs, presumably be-cause of p50/RelB complexes that remain activein the absence of p100/NF-kB2, the primaryinhibitor of RelB (Zhu et al. 2006; Zhang et al.2007). As discussed previously, Bcl-3 is notpart of the alternative pathway and is most likelyinduced by the classical pathway. It is not knownhow the combined loss of both NF-kB2 andBcl-3 leads to such a profound defect. Bcl-3most likely functions by interaction withNF-kB1; this is consistent with a complete blockin mTEC development in mice lacking NF-kB1and NF-kB2 (Zhang et al. 2007).

The classical NF-kB pathway is also indis-pensable. TRAF6 knockout mutants failed todevelop mTECs and exhibited multiorgan in-flammation (Inoue et al. 2007; Akiyama et al.2008). TRAF6 is engaged by numerous signalsand functions just upstream of the IKK complexin the classical pathway (Ghosh and Hayden2008; Hayden and Ghosh 2008; Vallabhapurapuand Karin 2009). RelB and NF-kB2 are likelycritical targets of the classical pathway, becausecontinuous production of these components ofthe alternative pathway would be required forlong-term activation of this pathway (Inoueet al. 2007; Akiyama et al. 2008). Chemokines,such as CCR7 ligands, are important targets ofNF-kB in epithelial cells that permit the





attraction of positively selected DP thymocytes(Förster et al. 2008; Nitta et al. 2008). But howthe alternative and/or classical pathways controlthe generation of mTECs themselves is an openquestion.

The alternative pathway and RelB may alsohave a specific role in nonhematopoietic cellsrequired for thymic iNKT formation (Elewautet al. 2003; Sivakumar et al. 2003; Franki et al.2005), although recent reports suggest that LTbR-mediated activation of this pathway may beprimarily required for thymicexport/peripheralcolonization of these cells and not their dev-elopment (Franki et al. 2006; Vallabhapurapuet al. 2008).

SECONDARY LYMPHOID ORGANS: NF-kBIN STROMAL CELLS

The classical and alternative activation ofNF-kB in stromal cells is instrumental in estab-lishing secondary lymphoid organs and/orlymphoid organ architecture. The lymphoid or-gans provide the anatomical niches required foroptimal initiation and expansion of adaptiveimmune responses (Drayton et al. 2006; Muel-ler and Ahmed 2008; Randall et al. 2008). Anumber of mouse mutants in which the alterna-tive pathway was fully or partially blocked failed,to variable degrees, in forming splenic B-cellfollicles, proper marginal zones, lymph nodes,Peyer’s patches, and germinal centers. Thesemutants include those deficient in RelB orNIK, or with compound deficiency in NF-kB2and Bcl-3, or harboring inactivating mutationsof IKKa or NIK (Matsushima et al. 2001; Senft-leben et al. 2001a; Weih et al. 2001; Paxian et al.2002; Yilmaz et al. 2003; Weih and Caamano2003; Zhang et al. 2007). Loss of NF-kB2 alonehad milder defects, consistent with only partialabrogation of RelB complexes (see previousdiscussion) (Franzoso et al. 1998; Weih andCaamano 2003; Guo et al. 2007). Interestingly,loss of Bcl-3 also displayed some mild defectsin lymphoid architecture (Franzoso et al.1997b; Poljak et al. 1999).

The absence of splenic B-cell follicles inmice impaired in alternative NF-kB activationin the models cited previously correlated with

a lack of follicular dendritic cell (FDC) networks(Weih and Caamano 2003; Zhang et al. 2007).FDCs present antigen–antibody immune com-plexes to facilitate clonal selection/differentia-tion of antigen-activated B cells in germinalcenter reactions that generate plasma cellsand memory B cells (Allen and Cyster 2008;Mueller and Ahmed 2008; Batista and Harwood2009). Apart from the ability of both FDCs andthymic mTECs to present antigens, these cellshave many other parallels. They attract theirlymphocyte partners by expressing chemo-attractants—in the case of FDCs, CXCL13(a.k.a. BLC) to attract B and organize theminto follicles. These cells also require continu-ous stimulation by lymphocytes to maintaintheir differentiated network (Allen and Cyster2008; Mueller and Ahmed 2008; Batista andHarwood 2009). B cells express LTab (andLight) and signal FDCs via LTbRs to activateRelB complexes via the alternative pathway, inturn inducing expression of chemokines, e.g.,CXCL13 (Suto et al. 2009). LTbR signalingalso activates RelA complexes, either by directactivation of the classical pathway or by liberat-ing RelA complexes inhibited by p100 (Mullerand Siebenlist 2003; Basak et al. 2007). Miceblocked in activation of the alternative pathwayshare many of the same defects apparent in micedeficient in LTbR signaling (Matsushima et al.2001; Hehlgans and Pfeffer 2005; Draytonet al. 2006; Ruddle and Akirav 2009). The alter-native pathway is also required within stromalcells to form proper marginal zones. Impairedsignaling via this pathway decreased the num-bers of MZB cells, even if the pathway was intactwithin B cells (where it also plays a role; see pre-vious discussion). RelB and the alternativepathway appear to control integrin-ligand in-teractions and chemokine expression requiredfor MZB retention (Weih et al. 2001; Weihand Caamano 2003; Guo et al. 2007).

Activation of the alternative pathway viaLTbR in stromal (mesenchymal organizer) cellsis essential for lymph node and Peyer’s patchorganogenesis (Drayton et al. 2006; Elewautand Ware 2007; Vondenhoff et al. 2007). Similarto the situation in prenatal mTEC development,hematopoietic LTi cells initiate the process in





part via expression of ligands for LTbR andRANK (Evans and Kim 2009); subsequently, Bcells provide the necessary ligands (Elewautand Ware 2007). Interestingly, only organogen-esis of lymph nodes, but not of Peyer’s patches,requires RANK activation on stromal cells, apotent activator of both the alternative andclassical pathway for NF-kB (see previous dis-cussion) (Yoshida et al. 2002).

In addition to RANK and LTbR, the TNFreceptor I may also play a role in lymphoidorganogenesis and maintenance, although itonly stimulates the classical pathway (Ruddleand Akirav 2009). The importance of the classi-cal pathway is also underscored by the role ofthe adaptor protein TRAF6, which mediatesclassical activation by many receptors, includ-ing RANK, though not TNF receptor I (Ghoshand Hayden 2008; Hayden and Ghosh 2008;Vallabhapurapu and Karin 2009); TRAF6 wasrequired within stromal cells for proper splenicarchitecture (e.g., B-cell follicles, marginalzones) as well as for lymph node and Peyer’spatch organogenesis (Qin et al. 2007). Similarto the situation with mTECs, the classical path-way is likely required to induce expression ofRelB and NF-kB2 to provide the componentsneeded for continuous signaling via the alter-native pathway (see previous discussion); inaddition, the classical pathway may induce ex-pression of Bcl-3. Little is known about howNF-kB controls the development/differentia-tion of stromal cells, such as FDCs, but as inmTECs, NF-kB does induce expression of che-mokines essential for cellular interactions andcompartmentalization in secondary lymphoidorgans (Vondenhoff et al. 2007; Mueller andAhmed 2008; Randall et al. 2008; Evans andKim 2009; Suto et al. 2009).

NF-kB AND B-CELL ACTIVATION

Mature B-cell populations respond to twobroad categories of antigen: so-called thymus-independent (TI) and thymus (T)-dependentantigens. Polyvalent antigens that typicallypossess multiple repeat determinants foundon macromolecules such as polysaccharidescan activate B cells without T-cell help. These

T-independent humoral responses, dominatedby CD5þ and marginal zone B cells, typically in-volve the secretion of relatively low affinity IgMantibodies (McHeyzer-Williams 2003). Follicu-lar B cells on the other hand, which serve as pre-cursors of T-dependent B-cell responses toprotein antigens are typically triggered by acti-vated helper T cells through CD40 and receptorsfor cytokines such as IL-4 (McHeyzer-Williamsand McHeyzer-Williams 2005). This leadsto iso-type switching and germinal center formationin secondary lymphoid organs, during whichsomatic hyper-mutation and antigen-drivenselection produce long-lived affinity-maturedplasma cells, plus memory B cells that can re-spond rapidly to antigen rechallenge (McHeyz-er-Williams 2003; McHeyzer-Williams andMcHeyzer-Williams 2005). A dichotomy be-tween T-dependent and T-independent B-cellresponses is unlikely to be as clear-cut in thecontext of a microbial infection, in which multi-valent antigens are encountered in conjunctionwith protein antigens. Figure 3 outlines thosestages during follicular B-cell activation and dif-ferentiation that involve NF-kB signaling.

Mature B cells express different forms ofNF-kB. In resting cells, NF-kB1/c-Rel hetero-dimers and NF-kB1 homodimers predominate(Grumont and Gerondakis 1994), with RelA(NF-kB1/RelA, c-Rel/RelA) and NF-kB2 (NF-kB2/RelB, NF-kB2 homodimers) dimers alsopresent, albeit at lower levels. Constitutive, low-grade nuclear expression of NF-kB factorsdownstream of both the classical and alternatepathway is a feature of nonactivated B cellsthat is thought to mainly serve a survival role(Gerondakis and Strasser 2003; Claudio et al.2006). Low intensity, nonmitogenic BCR sig-nals sustain the production of p100 NF-kB2,which in turn serves to transmit survival signalsthrough the BLyS receptor, BR3 (a.k.a. BAFFR)(Stadanlick et al. 2008).

Functional and developmental outcomes forB cells, including proliferation, differentiation,and survival that arise from specific types ofactivation signals, are reflected in the qualitativeand quantitative nature of NF-kB responses.BCR, CD40, and TLR receptors all individu-ally engage NF-kB during B-cell activation





(Vallabhapurapu and Karin 2009). Strong BCRsignals rapidly lead to the IKKb-dependent nu-clear recruitment of NF-kB1/c-Rel, NF-kB1/RelA, and NF-kB1 homodimers from cytoplas-mic reserves (Kontgen et al. 1995; Grumontet al. 1998). Lipopolysaccharide and CpG doublestranded DNA, respective ligands for TLR4 andTLR9, also rapidly activate these same NF-kBcomplexes (Grumont and Gerondakis 1994;Krieg 2002). In contrast, CD40 signals activateboth IKKa and IKKb (Hayden and Ghosh2008; Vallabhapurapu and Karin 2009), result-ing in the nuclear expression of both RelB andc-Rel complexes. Although both NF-kB path-ways participate in CD40-dependent humoralimmunity (Gerondakis et al. 2006; Haydenand Ghosh 2008; Vallabhapurapu and Karin2009), impaired antibody responses that arisefrom a block in the alternate NF-kB pathway arenot caused by B-cell intrinsic defects (Franzoso

et al. 1998), highlighting the distinct roles servedby the two arms of the NF-kB pathway. Crosstalk among these receptors is crucial in determin-ing how NF-kB influences functional outcomesduring B-cell activation. BCR signals alone,although not able to efficiently trigger B-cell dif-ferentiation, nevertheless provide essential costi-mulatory signals that influence survival anddifferentiation. BCR induction ofNF-kB protectsCD40-stimulatedB cells fromactivation-inducedcell death (Rothstein et al. 2000) (see later dis-cussion) and prevent B-cell hypo-responsivenessthat can accompany continuous TLR signaling(Poovassery et al. 2009). Finally, the duration ofB-cell stimulation also influences the temporalregulation of NF-kB activity. In addition to theinitial rapid nuclear induction of NF-kB, sus-tained TLR4 (Grumont and Gerondakis 1994)and BCR (S.G., unpubl. results) signals leadto multiple, transient waves of nuclear NF-kB

NF-kB dependent

BCR

NF-kB independent NF-kB dependent NF-kB?

Highaffinity

IL-4T

B

Th

Th

T

IgM

B

B

BCD40LCD40

BCR

B

B

B

BLowaffinity

ApoptosisApoptosis

APC

Ag

FDC

IgG1

Plasma cell

B

B B B B

MemoryB cell

Figure 3. NF-kB in B-cell activation. The activation of follicular B cells by CD4 T-cell-dependent CD40 and BCRsignals promotes a rapid antigen-driven expansion of B cells in the germinal centers (GC) of secondarylymphoid organs that is accompanied by isotype switching and affinity maturation. These antigen andcytokine driven events lead to the development of long-lived plasma cells and memory B cells. Shown are therequirements for NF-kB during the various phases of B-cell differentiation and proliferation.





expression that are dependent on the “de novo”synthesis of these transcription factors (GrumontandGerondakis1994).Itremainstobedeterminedwhich functions during B-cell activation are linkedto the specific phases of NF-kB expression.

NF-kB and B-cell Division

Both the IKKa and IKKb-dependent NF-kBpathways contribute to mitogen-induced B-cellproliferation by regulating multiple B-cell intrin-sic mechanisms. BCR, CD40, and TLR4 or TLR9signals each use the classical NF-kB pathwayto promote B-cell proliferation (Ghosh andHayden 2008; Hayden and Ghosh 2008; Vallab-hapurapu and Karin 2009); only CD40 appearsto require the alternate pathway for this function(Vallabhapurapu and Karin 2009). All IKKb-dependent B-cell mitogenic responses requirec-Rel (Gerondakis et al 1998), albeit to varyingdegrees, with BCR activation being most de-pendent on c-Rel (Grumont et al. 1998).NF-kB1 on the other hand is only crucial for

TLR4-dependent B-cell proliferation (Sha et al.1995; Grumont et al. 1998).

The classical NF-kB pathway serves severaldistinct roles in the cell cycle during BCR-induced proliferation (Fig. 4). c-Rel is requiredfor progression from G1 to S-phase (Grumontet al. 1998 ). Although c-Rel regulates cell sur-vival during BCR activation (Grumont et al.1998, 1999; Cheng et al. 2003), the failure ofc-Rel-deficient B cells to proliferate is a directconsequence of a cell cycle block and notmerely a secondary consequence of cell death(Grumont et al. 1998; Cheng et al. 2003). Theexpression of E2F3, an E2F transcription factorinvolved in G1/S phase progression (Humbertet al. 2000) was shown to be impaired in BCRactivated c-Rel-deficient B cells (Cheng et al.2003). With E2F3 known to contribute tothe suppression of p21Cip1 function (Wuet al. 2001), c-Rel may promote cell cycle pro-gression by serving to inactivate Rb-dependentrepression of S-phase entry. Whether E2F3 isthe only direct c-Rel target important in cell

Go

Early

BCR

TLR

Cellgrowth

G1progression

CD40

Cellcycle

NF-kB dependent NF-kB independent

Cellsurvival

Late

G1 S G2/M

Figure 4. Roles of NF-kB in B-cell division. Following stimulation through the BCR, CD40, or TLR4/TLR9,mature quiescent B cells in G0 enter G1 and undergo an NF-kB dependent phase of growth. B-cell growthcontinues until late G1, at which point NF-kB is required for entry into S-phase. Subsequent steps in theB-cell cycle appear to be NF-kB independent. NF-kB also regulates survival signals associated with B-cellactivation and division.





cycle progression remains to be determined.The expression of IRF4, a transcription factorshown to influence normal B cell (Mittruckeret al. 1997) and malignant B cell (Shaffer et al.2008) proliferation is also impaired in mitogenactivated c-Rel-deficient B cells (Grumont andGerondakis 2000). Despite c-Rel being neces-sary for BCR or TLR4 entry into S phase (Gru-mont et al. 1998), these combined signalspermit a majority of B cells to progress throughthe cell cycle in the absence of c-Rel. These find-ings are consistent with c-Rel setting a thresholdfor cell cycle progression by modulating thestimulus-dependent transcription of cell cycleregulators. Transcriptional targets of NF-kB1involved in TLR4-dependent G1 to S phase pro-gression remain to be identified.

c-Rel and NF-kB1 function redundantly topromote cell growth during G0 to G1 progres-sion (Pohl et al. 2002; Grumont et al. 2002).This step in the cell cycle, associated with ribo-some biogenesis (Stocker and Hafen 2000), isessential for assembling the cellular machinerynecessary for DNA replication. This NF-kB-dependent control of B-cell growth, which is re-quired for all activation signals, involves c-Reland NF-kB1 inducing the expression of c-Myc(Lee et al. 1995; Grumont et al. 2002), whichserves a key role in the growth of eukaryotic cells(Levens 2002). Despite c-Myc transgene expres-sion promoting B-cell growth in the absence ofc-Rel and NF-kB1, it is insufficient to drive themitogen-dependent proliferation of B cells lack-ing these NF-kB transcription factors (Grumontet al. 2002), further highlighting the multiplerole(s) c-Rel and possibly NF-kB1 serve inB-cell division.

In vivo, B cells undergo extensive antigen-dependent proliferation in germinal centerslinked with BCR affinity maturation that occursduring the development of memory B cells andplasma cells (McHeyzer-Williams 2003). Re-markably, germinal center (GC) B cells fail toexpress most NF-kB target genes and NF-kBpathway components (Shaffer et al. 2001), indi-cating that NF-kB signaling appears to be dis-pensable for GC B-cell proliferation. A strongbias of the GC B-cell gene expression profiletoward G2/M regulators rather than cell growth

regulators such as ribosomal components(Shaffer et al. 2001) is consistent with a short-ened G1 that is characteristic of rapidly dividingGC B cells (McHeyzer-Williams 2003). Thefindings that the NF-kB pathway appears to beimportant for initiating the proliferation ofnaı̈ve B cells, but not the division of GC B cellsduring differentiation, reinforces the link be-tween NF-kB and G0/G1 control of the cellcycle. It also highlights how selective engage-ment of the NF-kB pathway during differentstages of B-cell activation and differentiationcan be used to control the rate of B-cell division.

NF-kB Control of Activated B-cell Survival

Elevated levels of apoptosis typically accompanyB-cell activation. The prevailing models favorapoptosis in this instance as a mechanismused to eliminate B cells that are not pro-grammed to undergo normal cell division.Such models are consistent with c-Rel-control-ling BCR and TLR4-induced division and sur-vival through the regulation of distinct targetgenes (Grumont et al. 1998, 1999; Cheng et al.2003). The viability of BCR-activated B cells isdependent on c-Rel directly inducing the tran-scription of genes encoding the Bcl-2 prosur-vival homologs, A1/Blf1 (Grumont et al.1999) and Bcl-xL (Cheng et al. 2003). Althoughthe coexpression of two related antiapoptoticgenes has been viewed as a redundant, fail-safesurvival mechanism (Lee et al. 1999), neitheroverexpression of A1 (Grumont e al. 1999) orBcl-xL (S.G., unpubl. results) alone is sufficientto confer complete protection to BCR-activatedB cells lacking c-Rel. However, with enforcedBcl-2 transgene expression able to block theBCR induced death of c-Rel deficient B cells(Grumont et al. 1998), thereby ruling out anyinvolvement of the death receptor pathway,this reinforces the notion that A1 and Bcl-xLserve distinct survival roles. With the c-Rel in-duction of A1 preceding Bcl-xL in activated Bcells (Grumont et al. 1999; Cheng et al. 2003),it is conceivable that A1-mediated B-cell surviv-al is linked to early events in the cell cycle, suchas cell growth, whereas Bcl-xL prevents death as-sociated with the G1 to S phase transition.





Unlike cell death associated with BCR acti-vation, which is limited by c-Rel alone, bothNF-kB1 and c-Rel cooperate to protect TLR4-stimulated B cells from apoptosis (Gerondakiset al. 2007). This cell death, driven by activationof the BH3-only proapoptotic protein Bim isblocked by c-Rel and NF-kB1 regulating distinctsurvival pathways (Gerondakis et al. 2007; Bane-rjee et al. 2008). c-Rel induction of A1 andBcl-xL coincide with both survival proteinsbinding to Bim, thereby preventing it from en-gaging the cell intrinsic death pathway. NF-kB1indirectly inhibits TLR4-induced death by con-trolling ERK phosphorylation of Bim (Banerjeeet al. 2008), a posttranslational event that targetsBim for degradation (Ewings et al. 2007).NF-kB1 and ERK activation are linked throughthe NF-kB1 precursor, p105 serving as a scaffoldfor the ERK pathway specific MAP3 kinase, Tpl2(Belich et al. 1999). Following TLR4 activation,IKKb-induced phosphorylation and degrada-tion of p105 serves as an essential step in Tpl2 ac-tivation (Waterfield et al. 2003; Banerjee et al.2005). It remains to be determined if these dis-tinct NF-kB-regulated survival mechanismsneutralize Bim in a spatial or temporal manner.

CD40 serves opposing roles in B-cell survi-val. In isolation, CD40 signals are able to pro-mote B-cell survival through the NF-kBinduction of A1 and Bcl-xL (Lee et al. 1999).However, under physiological conditions inwhich antigen-stimulated helper T cells expresshigh levels of CD40 ligand, CD40 signals sensi-tize B cells to Fas-induced apoptosis (Rothsteinet al. 2000). This death, which is prevented bycostimulating B cells through the BCR, ensuresthat only appropriate antigen-specific B cellsand not bystander cells are activated by CD40(Rothstein et al. 2000). This BCR-induced pro-tection of CD40-sensitized B cells is independ-ent of c-Rel (Owyang et al. 2001) and requiresthe PI3K/AKT-dependent induction of c-FLIP(Moriyama and Yonehara 2007), a caspase-8 in-hibitor essential for death-receptor-inducedapoptosis (Strasser et al. 2009). Although theNF-kB pathway activated through CD40 hasbeen shown in certain cell line models toprevent Fas-induced B-cell death (Lee et al.1999; Zazzeroni et al. 2003), these seemingly

paradoxical findings may be reconciled withCD40 alone serving to protect B cells from tran-sient, but not sustained Fas signals (Lee et al.1999). With death-receptor-induced apoptosislargely resistant to prosurvival signals emanat-ing from the cell intrinsic survival pathway(Strasser et al. 2009), CD40 induction of Bcl-xLand A1 most likely serves to prevent deathlinked to CD40-induced cell division.

Isotype Switching

NF-kB signaling is essential for isotype switch-ing (Hayden et al. 2006), a mechanism used tofine-tune humoral immune responses that in-volves the transposition of the assembled VDJgene located upstream of Cm to a downstreamheavy chain (CH) gene. T-independent antigensand T-cell-dependent CD40 signals, in combi-nation with specific cytokines, direct switchingto specific classes of immunoglobulin (Maniset al. 2002). In the mouse, for example, LPS pro-motes switching to IgG3, whereas CD40 signalsand IL-4 produced by Th2 T cells promoteswitching to IgG1 and IgE (Manis et al. 2002).Switch recombination occurs between regionsof repeat sequence (S-regions) located upstreamof each CH gene. S-region transcription activatedby mitogen and cytokine-induced binding ofproteins to cryptic promoters (I regions) flank-ing S-regions facilitates S-region access to theDNA recombination machinery (Manis et al.2002). NF-kB transcription factors downstreamof the classical and alternate pathways partici-pate in switching through a variety of mecha-nisms (Hayden et al. 2006). NF-kB factorsinduced by mitogens and cytokines promoteS-region transcription by binding to specific Iregions (Delphin and Stavnezer 1995; Agrestiand Vercelli 2002; Bhattacharya et al. 2002;Wang et al. 2006). Switching is also dependenton NF-kB factors binding to the 30 IgHenhancer,a locus control region located downstream of theCa gene (Zelazowski et al. 2000; Laurencikieneet al. 2001). Although NF-kB does not regulatethe expression of the switch recombination en-zymes (Chaudhuri et al. 2007), NF-kB1 hasbeen implicated in the DNA rearrangementprocess (Kenter et al. 2004), and given DNA





synthesis is essential for S-region rearrangement(Chaudhari et al. 2007), NF-kB signaling alsoprobably indirectly influences switching throughthe promotion of B-cell proliferation.

NF-kB AND T-CELL ACTIVATION

The successful activation of naı̈ve T cells re-quires two distinct signals: an antigen-specificsignal arising from TCR engaging MHC boundpeptides expressed on APC, plus a costimula-tory signal in the form of the B7 ligand up-regulated on APC that binds to CD28. Thesesignals, in addition to inducing the rapid ex-pression of IL-2 and activation-associated cellsurface molecules, promote T-cell division andeffector T-cell differentiation along differentlineages, the later of which is dictated by cyto-kine signals. The careful regulation of T-cellnumbers before, during, and following immuneresponses by controlling the competing mecha-nisms of cell survival and apoptosis, serve toprevent immune-related pathology associatedwith inappropriate T-cell expansion and func-tion. Those stages during T-cell activation anddifferentiation that involve NF-kB functionare summarized in Figure 5.

Before antigen activation, naı̈ve CD4 andCD8 T cells express very low levels of NF-kB1in the nucleus. In contrast to B cells, the absenceof constitutive NF-kB activity in these T cells isconsistent with this pathway being dispensablefor the homeostatic survival of naı̈ve T cells. Fol-lowing TCR and costimulatory signals, differentNF-kBproteins enter the nucleus in atemporallyregulated fashion (Molitor et al. 1990). Initially,pre-exisiting NF-kB1/RelA dimers bound toIkBa are rapidly transported to the nucleus(Molitor et al. 1990; Venkataraman et al. 1996).Although c-Rel is present in the cytoplasm ofnaı̈ve T cells, it is not rapidly mobilized by TCRsignals (Venkataraman et al. 1995; Rao et al.2003) due to its association with IkBb (Banerjeeet al. 2005), which is relatively resistant to degra-dation triggered by the initial TCR stimulus. Sig-nificant nuclear levels of c-Rel, which followNF-kB1/RelA expression, depend on the TCR-dependent induction of c-rel transcription(Grumont and Gerondakis 1990; Venkataraman

et al. 1996; Rao et al. 2003). TCR-induced c-Relexpression is NFAT dependent (Venkataramanet al. 1995), consistent with c-Rel being a directtranscriptional target of NFAT (Grumont et al.2004). Although c-Rel is necessary for the induc-tion ofIL-2(see laterdiscussion) followingT-cellactivation, the role served by the initial wave ofNF-kB1/RelA expression remains unclear,although it may be related to cell-cycle controlas reflected by the impaired proliferation ofTCR-activated RelA-deficient T cells (Doi et al.1997).

NF-kB and T-cell Proliferation

Both the classical and alternate NF-kB pathwaysregulate T-cell proliferation in culture and “invivo” via T-cell autonomous and nonautono-mous mechanisms (Gerondakis et al. 2006; Val-labhapurapu and Karin 2009). NF-kB signalingis pivotal in controlling the proliferation of naı̈veT cells (Gerondakis et al. 2006). NF-kB1, c-Rel,and RelA serves unique (Kontgen et al. 1995;Gerondakis et al. 2006; Doi et al. 1997; Sriskan-tharajah et al. 2009) and overlapping roles (Ger-ondakis et al. 2006) in T-cell proliferation duringdifferent stages of the cell cycle. In contrast to Bcells, which require NF-kB1 or c-Rel to inducemitogen-dependent T-cell growth (Grumontet al. 2002), it is RelA and c-Rel that function re-dundantly to promote c-Myc-induced T-cellgrowth following TCR activation (Grumontet al. 2004). Although c-Myc expression is bothnecessary and sufficient to promote T-cellgrowth in the absence of NF-kB function, it isnot to rescue the TCR-induced proliferation ofNF-kB-deficient T cells (Grumont et al. 2004).Therefore, the combined activity of these tran-scription factors, as well NF-kB1 and c-Rel(Zheng et al. 2003), serve functions in additionto growth, which are required for T-cell prolifer-ation. The finding that IKK2-deficient CD4 Tcells proliferate in culture as effectively as normalcells in response to a range of strong polyclonalstimuli (Schmidt-Supprian et al. 2003) is seem-ingly at odds with the findings for T-cell prolifer-ation that emerge from the analysis of thedifferent NF-kB compound mutant mice. Thiscan be reconciled by the ability of the residual





NEMO/IKKa complexes in IKKb-deficient CD4T cells to phosphorylate IkBa and induce anormal pattern of NF-kB complexes followingT-cell activation (Schmidt-Supprian et al. 2003).However, the inability of OVA or KLH antigen-specific IKKb-deficient T cells to proliferate inculture following restimulation with thesesame antigens that had previously been used toimmunize mice in which IKKb was selectivelydeleted in CD4 T lineage cells (Schmidt-Suppr-ian et al. 2004b) suggests that IKKa can only

activate NF-kB complexes normally inducedby IKKb in response to strong stimuli.

IL-2, which is involved in autocrine-medi-ated T-cell proliferation, is dependent on c-Relfor its induction following TCR and costimula-tory signaling (Kontgen et al. 1995). c-Rel ap-pears to induce il2 transcription in response toTCR and CD28 stimulation by remodelingchromatin in the vicinity of the CD28 responseregion of the IL-2 promoter, so that it is nowconducive to transcription (Rao et al. 2003;

Ag

APC

IL-4

IL-12 IL-2

CD4

CD8

Antigenpresentation

(NF-κB)

Activation(NF-κB)

Proliferation&

maturation(NF-κB)

Effectorfunctions(NF-κB)

EffectorCTL perforin

AICD(NF-κB)

Differentiation

Ag

Ag

Ag B7

TGF-bIL-6IL-21

IL-17IL-21

IL-4IL-5Il-13

gIFN

gIFN

gIFN

MemoryT cells

Th17

Th2

Th1

Figure 5. NF-kB in conventional T-cell activation. TCR-dependent activation of naı̈ve CD4 and CD8 T cells byantigen presenting cells (APC) leads to T-cell activation, division, and effector T-cell differentiation. Thevarious CD4 T helper cell subsets and CD8 cytotoxic T lymphocytes that develop in response to differentcytokine and costimulatory signals delivered by APC and T cells undergo rapid antigen-driven expansionduring the course of an infection. Following pathogen elimination, most effector T cells undergo activation-induced cell death, with the few surviving antigen-specific T cells differentiating into long-lived memorycells. Highlighted are roles for NF-kB in APC function, T-cell activation, CD4 Th differentiation, and T-cellsurvival and proliferation.





Chen et al. 2005). Despite c-Rel being essentialfor IL-2 dependent CD4 T-cell proliferation inculture (Kontgen et al. 1995), this cytokine isdispensable for conventional T-cell proliferationin vivo (Schimpl et al. 2002). Instead, IL-2-defi-cient mice develop lymphoproliferative andautoimmune disease caused by impaired CD4regulatory T-cell development (Liston and Ru-densky 2007). Although c-Rel-deficient micehave reduced numbers of CD4 regulatory T cells,this is caused by IL-2-independent mechanisms(S.G., unpubl. results). The failure of c-Rel-defi-cient mice to develop autoimmune disease indi-cates that IL-2 expression can be controlled in ac-Rel-independent fashion in vivo.

Although the mechanisms by which NF-kBcontrols T-cell proliferation in culture are rea-sonably well established, NF-kB regulation ofT-cell proliferation in vivo is complex, with anin depth overview beyond the scope of thisarticle. Nevertheless, it is clear that the extentto which the NF-kB pathway contributes toT-cell proliferation in vivo is dictated by thetype of T cell, its state of differentiation, plusthe nature and strength of the T-cell stimulatorysignals. For example, the homeostatic expan-sion of naı̈ve CD4 and CD8 T cells in a lympho-penic environment is absolutely dependenton IKKb (Schmidt-Supprian et al. 2004b).Although IKKb is not essential for polyclonalT-cell expansion in vivo that is driven by strongmitogenic stimuli such as super antigens(Schmidt-Supprian et al. 2004b), it does servea nonredundant T-cell-intrinsic role in theantigen-dependent proliferation that accompa-nies a recall response (Schmidt-Supprian et al.2004b). These results, supported by in vitrofindings, indicate that IKKb function is onlydispensable for T-cell proliferation in vivo thatis induced by strong TCR signals. Despitesome confusion about the roles of NF-kB inTCR-induced T-cell proliferation that arisefrom the overlap of IKKa and IKKb-dependentfunctions, individual NF-kB transcription fac-tors do serve select roles in the T-cell prolifera-tive responses triggered by infectious agents. Inthe case of c-Rel, it serves a T-cell-intrinsic rolein the clonal expansion of Th1 effector cells re-sponding to Toxoplasma gondii (Mason et al.

2004), whereas it is dispensable for the expan-sion of antigen-specific CD8 cytotoxic effectorT cells during the primary response to influenzavirus (Harling-McNabb et al. 1999). AlthoughNF-kB is involved in the proliferative responseof effector T cells, it remains to be determinedwhether this pathway features in the rapid clo-nal expansion of CD4 and CD8 memory T cellsfollowing antigen recall.

NF-kB and T-cell Survival

NF-kB regulates T-cell survival during antigen-dependent activation and proliferation bycontrolling both the cell intrinsic and deathreceptor pathways. Controversy that surroundshow some aspects of how T-cell survival is con-trolled appears to reflect the complexity of aprocess that is intimately linked to the strengthand duration of immune signals.

TCR and costimulatory signals exploit theNF-kB pathway to promote T-cell survival fol-lowing antigen-dependent activation. NF-kBactivation triggered through antigen-depend-ent TCR recruitment of the IKK complex to asignaling hub at the immunological synapsethat requires PKCu and the signaling compo-nents CARM1, Bcl-10, and MALT1 (Weil andIsrael 2006) appears to afford different degreesof protection to CD4 and CD8 T cells. Follow-ing TCR activation, PKCu is critical for CD8T-cell survival, which is mediated via c-Rel (Sai-bil et al. 2007). Although c-Rel is necessary forthe induction of Bcl-xL, expression of this anti-apoptotic molecule alone appears to be insuffi-cient to confer protection from TCR-inducedcell death (Saibil et al. 2007). In contrast,PKCu only serves a modest role in CD4 T-cellsurvival, which surprisingly is not dependenton either NF-kB1 or c-Rel (Saibil et al. 2007).However, the finding that a combined defi-ciency of c-Rel and NF-kB1 has a profound im-pact on TCR-induced CD4 T-cell survival(Zheng et al. 2003) points to these transcriptionfactors serving overlapping survival roles, pre-sumably mediated through PCKu dependentand independent pathways. This differenceserves to highlight that the NF-kB survival signalsdelivered through the TCR are programmed





differently in CD4 and CD8 T cells. The NF-kBpathway also exploits or is used by the T-cell cos-timulatory molecules OX40 (CD134) and gluco-corticoid induced TNF receptor (GITR) topromote the survival of activated T cells (Songet al. 2008; Zhan et al. 2008). In the case ofGITR, NF-kB-dependent expression of this core-ceptor on TCR-stimulated CD4 and CD8 T cells,appears to serve as a mechanism for promotingT-cell survival in response to weak, rather thanstrong TCR signals (Zhan et al. 2008).

NF-kB also serves as a lynch pin in coordi-nating cell survival during T-cell proliferation.Following TCR-dependent antigen stimulation,successful S-phase entry associated with the up-regulation of E2F target gene expression andcyclin E-dependent kinase activity requiresNF-kB to neutralize p73-induced apoptosis(Wan and DeGregori 2003). In the absence ofappropriate signals that activate NF-kB duringG1, although E2F activation and the capacityto enter S phase remains intact, cell death is trig-gered by CDK2-dependent p73 expression. Torestrict the number of activated T cells duringimmune responses, ongoing T-cell proliferationassociated with persistent TCR stimulation pro-motes activation-induced T-cell death (AICD)by engaging the death receptor Fas (Strasseret al. 2009). TCR-dependent activation of theNF-kB pathway through its inhibition ofcaspase-8 activation helps counter Fas-inducedT-cell death (Jones et al. 2005). It remains to bedetermined how in the face of continuous TCRsignals, T cells sense a need to prevent or suc-cumb to AICD by regulating NF-kB activation.Once pathogenic agents have been eliminated, asmall number of antigen-specific CD4 or CD8effector T cells survive and differentiate intolong-lived memory T cells capable of mountingrapid immune responses on recountering thepathogen. It remains to be determined ifNF-kB signaling is required for memory T-cellsurvival.

NF-kB and Effector T-cell Differentiation

Naı̈ve T can adopt a number of developmentaloptions that are dictated by the types of signalsthese cells encounter during antigen-dependent

activation. In the case of CD4 T helper cells(Th), these include Th1, Th2, Th17, and T reg-ulatory cells, all of which are distinguished bythe types of cytokines these cells produce (Fig.5). The NF-kB pathway regulates CD4 Th dif-ferentiation through APC-dependent andT-cell intrinsic mechanisms. Aside from therole NF-kB plays in antigen-dependent and in-dependent differentiation of certain dendriticcell (DC) subsets (Ouaaz et al. 2002; O’Keeffeet al. 2005), the NF-kB-regulated expressionof specific cytokines in these key APC, influen-ces Th cell differentiation and function. c-Rel,for example, which is required for Th1 immuneresponses (Gerondakis et al. 2006), is crucial forDC expression of p35, a subunit of the cytokineIL-12 that is essential in Th1 development(Grumont et al. 2001). The finding that in acti-vated DC, c-Rel controls the expression of IL-12(Grumont et al. 2001), whereas RelA is moreimportant in regulating the production of in-flammatory cytokines such as IL-1, IL-6, andTNF (Wang et al. 2007) that are involved inshaping Th17 responses (Laurence and O’Shea2007), may indicate that the antigen-dependentactivation of specific NF-kB complexes in DC,in part influences the outcome of Th cell devel-opment. However, a flexible rather than strictbifurcation of inflammatory versus regulatorycytokine production controlled by differentNF-kB factors is more likely, with c-Rel in cer-tain instances implicated in the expression ofinflammatory cytokines such as TNF (Grigoria-dis et al. 1996) and IL-23 (Mise-Omata et al.2007). In contrast to the role of c-Rel in Th1but not Th2 differentiation, NF-kB1 selectivelyregulates Th2 cell differentiation (Das et al.2001). T cells lacking NF-kB1 fail to induceGATA3 expression, a transcription factor crucialin the differentiation of Th2 cells. The role ofNF-kB1 in Th2 cells appears to be restricted tothe development rather than the function ofcommitted Th2 cells, with NF-kB1-deficientmice failing to mount airway inflammatory re-sponses, a defect that coincides with the inabil-ity to produce the Th2 signature cytokines IL-4,IL-5, and IL-13 (Das et al. 2001).

In addition to helping polarize T-cell re-sponses linked with the production of specific





cytokines, NF-kB transcription factors also reg-ulate cytokine gene expression in activated Tcells. Aside from NF-kB having a direct role inthe TCR and costimulatory signal dependenttranscription of cytokine genes that includeIL-2, GM-CSF, IL-3, and IFN-g (Gerondakiset al. 2006), cytokine production by naı̈ve T cellscan be enhanced by the TLR-dependent induc-tion in APC of T-cell costimulatory moleculesand cytokines that alter c-Rel responsivenessduring TCR activation (Banerjee et al. 2005).In naı̈ve T cells, which are notoriously refractoryfor cytokine production, c-Rel is primarilybound to IkB-b, which is relatively resistant todegradation triggered by TCR signals. T cells ex-posed to the pro-inflammatory cytokinesTNF-a and IL-1, shift c-Rel to IkB-a-associatedcomplexes that are readily targeted by the TCR.As a consequence, in cytokine-primed T cells,IL-2 and IFN-g mRNA are now produced rap-idly, and at higher levels.

Finally, little is currently known aboutthe genetic programs that regulate the develop-ment of memory cells from the pool of effectorT cells. Despite an essential role for NF-kB func-tion in memory T-cell development (Schmidt-Supprian et al. 2003), it remains unclear whetherthis dependence on the NF-kB pathway reflects adirect role in memory cell differentiation or amaintenance role through the control of mem-ory cell survival.

CONCLUDING REMARKS

Despite remarkable progress in understandingthe roles of the NF-kB pathway in normal lym-phocyte development and function, much stillremains to be learned. Areas in need of greaterattention include gaining a better understand-ing of how NF-kB controls the developmentof the innate T-cell lineages, determining whatroles the NF-kB pathway serves in memoryT-cell differentiation, and cataloging the manyand varied roles the NF-kB pathway serves inlymphocytes during different types of infec-tious diseases. With NF-kB involvement beingdocumented in an increasing number of lym-phoid-associated pathologies, answers to theseand a range of other outstanding questions

will serve as a basis for preventing and treatingthese diseases.

REFERENCES

Agresti A, Vercelli D. 2002. c-Rel is a selective activator of anovel IL-4/CD40 responsive element in the human Igg4 germline promoter. Mol Immunol 38: 849–859.

Aifantis I, Mandal M, Sawai K, Ferrando A, Vilimas T. 2006.Regulation of T-cell progenitor survival and cell-cycleentry by the pre-T-cell receptor. Immunol Rev 209:159–169.

Akiyama T, Shimo Y, Yanai H, Qin J, Ohshima D, MaruyamaY, Asaumi Y, Kitazawa J, Takayanagi H, Penninger JM, etal. 2008. The tumor necrosis factor family receptorsRANK and CD40 cooperatively establish the thymic me-dullary microenvironment and self-tolerance. Immunity29: 423–437.

Allen CD, Cyster JG. 2008. Follicular dendritic cell networksof primary follicles and germinal centers: phenotype andfunction. Semin Immunol 20: 14–25.

Allman D, Pillai S. 2008. Peripheral B cell subsets. Curr OpinImmunol 20: 149–157.

Banerjee D, Liou HC, Sen R. 2005. c-Rel-dependent primingof naive T cells by inflammatory cytokines. Immunity 23:445–458.

Banerjee A, Grumont R, Gugasyan R, White C, Strasser A,Gerondakis S. 2008. NF-kB1 and c-Rel cooperate to pro-mote the survival of TLR4-activated B cells by neutraliz-ing Bim via distinct mechanisms. Blood 112: 5063–5073.

Banerjee A, Gugasyan R, McMahon M, Gerondakis S. 2006.Diverse Toll-like receptors utilize Tpl2 to activate extrac-ellular signal-regulated kinase (ERK) in hemopoieticcells. Proc Natl Acad Sci 103: 3274�

roles of the nf-kb pathway in lymphocyte development and...

Documents