Thymus Medulla Formation and Central Tolerance Are ... · polymorphic MHC class II molecule I-O,...

10
of February 1, 2019. This information is current as Thymic Epithelial Cells + 5 Transgene in Keratin α That Express an IKK Mice - / - α Tolerance Are Restored in IKK Thymus Medulla Formation and Central Hu and Ellen R. Richie Dakshayani Lomada, Bigang Liu, Lezlee Coghlan, Yinling http://www.jimmunol.org/content/178/2/829 doi: 10.4049/jimmunol.178.2.829 2007; 178:829-837; ; J Immunol References http://www.jimmunol.org/content/178/2/829.full#ref-list-1 , 21 of which you can access for free at: cites 50 articles This article average * 4 weeks from acceptance to publication Fast Publication! Every submission reviewed by practicing scientists No Triage! from submission to initial decision Rapid Reviews! 30 days* Submit online. ? The JI Why Subscription http://jimmunol.org/subscription is online at: The Journal of Immunology Information about subscribing to Permissions http://www.aai.org/About/Publications/JI/copyright.html Submit copyright permission requests at: Email Alerts http://jimmunol.org/alerts Receive free email-alerts when new articles cite this article. Sign up at: Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists All rights reserved. Copyright © 2007 by The American Association of 1451 Rockville Pike, Suite 650, Rockville, MD 20852 The American Association of Immunologists, Inc., is published twice each month by The Journal of Immunology by guest on February 1, 2019 http://www.jimmunol.org/ Downloaded from by guest on February 1, 2019 http://www.jimmunol.org/ Downloaded from

Transcript of Thymus Medulla Formation and Central Tolerance Are ... · polymorphic MHC class II molecule I-O,...

of February 1, 2019.This information is current as

Thymic Epithelial Cells+5 Transgene in KeratinαThat Express an IKK

Mice−/−αTolerance Are Restored in IKKThymus Medulla Formation and Central

Hu and Ellen R. RichieDakshayani Lomada, Bigang Liu, Lezlee Coghlan, Yinling

http://www.jimmunol.org/content/178/2/829doi: 10.4049/jimmunol.178.2.829

2007; 178:829-837; ;J Immunol 

Referenceshttp://www.jimmunol.org/content/178/2/829.full#ref-list-1

, 21 of which you can access for free at: cites 50 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists All rights reserved.Copyright © 2007 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on February 1, 2019http://w

ww

.jimm

unol.org/D

ownloaded from

by guest on February 1, 2019

http://ww

w.jim

munol.org/

Dow

nloaded from

Thymus Medulla Formation and Central Tolerance AreRestored in IKK��/� Mice That Express an IKK� Transgenein Keratin 5� Thymic Epithelial Cells1

Dakshayani Lomada, Bigang Liu, Lezlee Coghlan, Yinling Hu, and Ellen R. Richie2

Medullary thymic epithelial cells (mTECs) play an essential role in establishing central tolerance due to their unique capacity topresent a diverse array of tissue restricted Ags that induce clonal deletion of self-reactive thymocytes. One mTEC subset expresseskeratin 5 (K5) and K14, but fails to bind Ulex europaeus agglutinin-1 (UEA-1) lectin. A distinct mTEC subset binds UEA-1 andexpresses K8, but not K5 or K14. Development of both mTEC subsets requires activation of the noncanonical NF-�B pathway.In this study, we show that mTEC development is severely impaired and autoimmune manifestations occur in mice that aredeficient in I�B kinase (IKK)�, a required intermediate in the noncanonical NF-�B signaling pathway. Introduction of an IKK�transgene driven by a K5 promoter restores the K5�K14� mTEC subset in IKK��/� mice. Unexpectedly, the K5-IKK� transgenealso rescues the UEA-1 binding mTEC subset even though K5 expression is not detectable in these cells. In addition, expressionof the K5-IKK� transgene ameliorates autoimmune symptoms in IKK��/� mice. These data suggest that 1) medulla formationand central tolerance depend on activating the alternative NF-�B signaling pathway selectively in K5-expressing mTECs and2) the K5-expressing subset either contains immediate precursors of UEA-1 binding cells or indirectly induces theirdevelopment. The Journal of Immunology, 2007, 178: 829 – 837.

T he thymus provides a unique microenvironment that isresponsible for the development of T cells that respond toforeign peptides presented by self-MHC molecules, yet

are tolerant of self-peptide-MHC complexes. Interactions betweenthymocytes and thymic epithelial cells (TECs)3 are required toshape the TCR repertoire during thymocyte differentiation (1, 2).Cortical TECs (cTECs) display peptide-MHC complexes that in-duce positive selection of CD4�CD8� double positive thymocytesexpressing TCRs with relatively low peptide binding affinity. Sig-naling pathways activated during positive selection promote con-tinued differentiation of double positive thymocytes to theCD4�CD8� or CD4�CD8� single positive stage and initiate thy-mocyte migration into the medulla. Medullary TECs (mTECs)play a key role in negatively selecting potentially autoreactive thy-mocytes. One primary mechanism for establishing central toler-ance to autoantigens is the induction of apoptosis in thymocytesthat express high affinity TCRs for peptide-MHC complexes on thesurface of mTECs. Medullary TECs are also required for the gen-

eration of CD4�CD25� T regulatory cells (Tregs) and NKT cells,both of which actively repress self-reactive T cells (3, 4).

The requirement for mTECs in generating central tolerance isprimarily a function of their unique ability to express genes en-coding a wide range of Ags many of which were previously con-sidered to be expressed only in peripheral tissues (reviewed in Ref.5). This phenomenon, referred to as promiscuous gene expressionis regulated in part by the autoimmune regulator (AIRE) gene,which is preferentially expressed in mTECs. The role of AIREgene in establishing central tolerance is illustrated by a rare syn-drome, termed APECED (autoimmune-polyendocrinopathy-candi-diasis-ectodermal dystrophy), in which individuals with AIREmutations develop severe, multiorgan autoimmune disease (6).Animal models of AIRE deficiency have verified that self-reactivethymocytes persist due to defective clonal deletion, although ad-ditional undefined factors regulate tissue-restricted Ag (TRA) ex-pression because certain TRAs are expressed even in the absenceof AIRE (7, 8). In addition to direct presentation, mTECs supplyself-peptides to medullary dendritic cells (DCs), which process andcross-present them to augment deletion of self-reactive T cells (9).DCs and mTECs both express CD40, CD80, and CD86 costimu-latory molecules that enhance negative selection as well as pro-mote development of Tregs (10–13). Thus, mTECs contribute di-rectly and indirectly to establishing central tolerance and avertingthe development of autoimmunity.

The medullary compartment contains phenotypically diverse ep-ithelial cells that can be categorized into two major subsets. OnemTEC subset displays high levels of surface MHC class II mole-cules, binds the lectin Ulex europaeus agglutinin-1 (UEA-1), ex-presses keratin 8 (K8) but not K5, and has a compact, globularappearance (14–16). A distinct mTEC subset expresses the non-polymorphic MHC class II molecule I-O, fails to bind UEA-1,expresses K5 but not K8, and has a more stellate morphology (15,16). Although both medullary subsets express CD80, a positivecorrelation has been found between the level of UEA-1 bindingsites and CD80 expression (8). Interestingly, AIRE and TRAs are

Department of Carcinogenesis, Science Park Research Division, University of TexasM. D. Anderson Cancer Center, Smithville, TX 78957

Received for publication July 14, 2006. Accepted for publication October 26, 2006.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This work was supported by Grant CA16672 from the National Institutes of Healthand Grant ES07784 from the National Institute on Environmental Health Sciences.2 Address correspondence and reprint requests to Dr. Ellen Richie, Department ofCarcinogenesis, University of Texas M. D. Anderson Cancer Center, Science ParkResearch Division, P.O. Box 389, 1808 Park Road 1C, Smithville, TX 78957. E-mailaddress: [email protected] Abbreviations used in this paper: TEC, thymic epithelial cell; mTEC, medullaryTEC; cTEC, cortical TEC; UEA-1, Ulex europaeus agglutinin-1; AIRE, autoimmuneregulator; K5, keratin 5; IKK, I�B kinase; Treg, regulatory T cell; DC, dendritic cell;TRA, tissue-restricted Ag; NIK, NF-�B-inducing kinase; LT�R, lymphotoxin �receptor; ANA, antinuclear Ab.

Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00

The Journal of Immunology

www.jimmunol.org

by guest on February 1, 2019http://w

ww

.jimm

unol.org/D

ownloaded from

expressed at higher levels in UEA-1high or CD80high cells prompt-ing the suggestion that these cells represent a more advanced stageof mTEC differentiation (8, 17). However, a direct precursor-prog-eny relationship between phenotypically or functionally distinctmTEC subsets has not been demonstrated.

The development of both mTEC subsets as well as thymic DCsdepends on activation of the NF-�B signaling pathway, specifi-cally the noncanonical, alternative pathway that culminates inRelB activation (18). Ligand engagement of certain receptors inthe TNFR family such as lymphotoxin � receptor (LT�R) acti-vates NF-�B-inducing kinase (NIK), which phosphorylates ho-modimers of the downstream I�B kinase (IKK)�. Activated IKK�in turn phosphorylates the C-terminal region of NF-�B2 (p100)leading to ubiquitin-dependent degradation and release of the N-terminal polypeptide, p52. The formation of RelB/p52 het-erodimers permits shuttling of RelB from the cytoplasm into thenucleus where it functions as a transcriptional regulator (18). Arange of medullary defects occurs in mice that are deficient invarious components of the alternative NF-�B activation pathway.Targeted disruption of the LT�R gene results in disorganized med-ullary regions that contain reduced numbers of both major mTECssubsets (19). Mice that have a naturally occurring mutation in NIK(NIKaly/aly) not only fail to develop lymph nodes and Peyer’spatches, but also have severe defects in medulla formation, includ-ing a deficiency of UEA-1 binding mTECs and reduced AIREexpression (20). A similar medullary phenotype is found in micethat are deficient in TNFR-associated factor-6 (21). Furthermore,DCs as well as UEA-1 binding TECs are absent in RelB-deficientmice (22, 23). Interestingly, these mutant strains also develop se-vere inflammation and autoimmune disease, indicating a break-down in the establishment of central tolerance.

Although there is ample experimental evidence that RelB acti-vation is essential for mTEC development and medullary organi-zation, it is not clear whether activation of the alternative NF-�Bpathway is required for the differentiation and function of bothmajor mTEC subsets. To address this issue we analyzed mice inwhich a targeted mutation in IKK� prevents noncanonical NF-�Bactivation (24, 25) and asked whether restoration of IKK� expres-sion exclusively in the K5� mTEC subset is sufficient to restorethe generation of UEA-1 binding cells and prevent development ofautoimmune disease. Medulla formation and mTEC developmentare severely impaired in IKK��/� mice, and autoimmune-like pa-thology develops when IKK��/� fetal thymuses are transferredinto athymic recipients. Introduction of an IKK� transgene regu-lated by a K5 promoter rescues development of the K5�UEA-1high

as well as the K5�UEA-1neg mTEC and prevents the developmentof autoimmunity. This outcome suggests that 1) medulla formationand central tolerance depend on activating the alternative NF-�Bsignaling pathway selectively in K5-expressing mTECs and 2) theK5-expressing subset either contains immediate precursors ofUEA-1 binding cells or indirectly induces their development.

Materials and MethodsMice

NCrnu/nu and BALB/cnu/nu mice were purchased from National Cancer In-stitute. IKK��/� mice were generated and genotyped as previously de-scribed (24). The IKK��/� mice were originally on a C57BL6/129Svbackground and backcrossed onto C57BL6/J for more than five genera-tions. Embryos were obtained by setting up timed matings for 16 h andconsidering the morning of finding the vaginal plug as embryonic day 0.5.

To produce the K5-IKK� transgenic line, a human IKK� cDNA frag-ment tagged with hemagglutinin was generated by PCR from a DNA tem-plate of an IKK�-expressing vector (26). The IKK� cDNA PCR fragmentwas inserted into the pGEM-T vector (Promega) sequening, and subse-quently subcloned into the NotI and SnaB1 sites of the BK5 vector con-

taining a bovine K5 promoter as previously described (27). The constructDNA was linearized by XhoI digestion and used to generate K5-IKK�transgenic mice in the transgenic core located at the Science Park ResearchDivision of the University of Texas M. D. Anderson Cancer Center (Smith-ville, TX). The PCR primers 5�-AAAGTGTGGGCTGAAGCAGTG-3�and 5�-GCCCAACAACTTGCTCAAATG-3� (1273–1819 bp) generate a546 bp DNA fragment for genotyping. IKK��/� and K5-IKK� transgenicmice were crossed to obtain K5-IKK�/IKK��/� mice that were identifiedby genotyping. Unless stated otherwise, age-matched embryonic day 18.5or newborn mice were used for all experiments.

Immunohistochemistry

Serial sections (5 �m) from OCT-embedded frozen tissue were air driedand fixed in cold acetone for 2 min at room temperature. After washing inHBSS, they are blocked for 15 min in HBSS containing 1% BSA. Theslides were incubated at 4°C with optimal dilutions of primary Abs. Thefollowing Abs and reagents were used: polyclonal anti-mouse K5 and K14(Covance), Troma-1 mAb (rat anti-K8; Developmental Studies HybridomaBank) and hamster anti-mouse CD11c (clone HL3; BD Biosciences). Bi-otinylated and FITC-conjugated UEA-1 were obtained from Vector Lab-oratories. Control slides were incubated with nonimmune serum or isotype-matched Ig. After washing, sections were incubated 30 min with directlyconjugated secondary reagents. For CD11c, immunoreactivity was en-hanced by tyramide amplification (PerkinElmer Life Sciences). Micro-scopic analysis was performed with an Olympus ProVis AX70 microscope.

Thymic stromal cell suspensions

Single-cell suspensions from thymuses of 4- to 6-wk-old K5-IKK� micewere prepared as described by Gray et al. (28) with minor modifications.Thymuses were finely minced and gently stirred in a beaker in cold RPMI1640 containing 1–5% FBS. Passing the suspension through a 70-�m cellstrainer (BD Falcon) depleted thymocytes. Fresh medium was added to thestromal fragments recovered from the strainer and the process was repeatedthree or four times. Stromal fragments were washed in Ca2� and Mg2� freePBS and digested with 0.25% trypsin (Sigma-Aldrich) in 0.02% EDTA for45 min at 37°C during which they were dispersed several times by pipet-ting. Digestion was stopped by washing in RPMI 1640 containing 5% FBS.The resulting suspension was further depleted of thymocytes by incubationwith anti-mouse CD45 bound magnetic beads (Dynal Biotech) and deplet-ing the bound cells with a magnet.

Flow cytometric analysis and sorting

Flow cytometric analysis and sorting was performed Beckman Coulter Al-tra flow cytometer. The following mAbs were purchased from BD Bio-sciences: anti-CD4 (RM4-5) conjugated with PE, anti-CD8 (53-6.7) con-jugated with FITC, FITC-conjugated anti-CD45.2, biotinylated anti-CD62ligand (MEL-14), and biotinylated anti-CD44 (IM7). Monoclonal Ab toFoxp3 (clone FJK-16s) was purchased from eBioscience. Thymocytes inHBSS containing 1% BSA and 0.1% sodium azide were incubated withdirectly conjugated or biotinylated Abs on ice for 30 min followed by threewashes. Biotinylated Abs were detected with streptavidin-allophycocyanin.The cells were fixed in 1% paraformaldehyde and stored at 4°C until anal-ysis. For sorting thymus stromal cells, the CD45-depleted cell suspensionswere incubated with allophycocyanin-conjugated anti-CD45 (clone30-F11) and FITC-conjugated UEA-1 in the presence of 2 �g/ml pro-pidium iodide to exclude dead cells. CD45 and propidium iodide-positivecells were excluded by electronic gating and UEAneg-low, UEAint, andUEAhigh cells were collected using the sort gates (see Fig. 5).

Fetal thymus grafts

Fetal thymic lobes from embryonic day 15.5 IKK��/�, K5-IKK�/IKK��/�, and their corresponding wild-type littermates were placed onNucleopore filters (Whatman) and cultured for 4 days in complete RPMI1640 containing 10% FBS and 1.35 mM 2-deoxyguanosine. The thymo-cyte-depleted lobes from IKK��/� and K5-IKK�/IKK��/� mice weregrafted under the renal capsule of NCrnu/nu and BALB/cnu/nu mice, respec-tively, as previously described (29). After 6–8 wk, the grafted thymuseswere recovered for immunohistochemical analysis and peripheral lymphnodes were obtained for analysis of T cell development by flow cytometricanalysis. Peripheral lymph node cells obtained from K5-IKK�/IKK��/�

thymus graft recipients were electronically gated for CD45.2 and then an-alyzed for activation phenotype.

830 THYMUS MEDULLA FORMATION AND CENTRAL TOLERANCE

by guest on February 1, 2019http://w

ww

.jimm

unol.org/D

ownloaded from

Pathology

Antinuclear Abs (ANA) were detected by using mouse ANA ELISA kit(Alpha Diagnostic International). To examine tissues for lymphocytic in-filtration, formalin-fixed slides were stained with H&E.

Semiquantitative RT-PCR analysis

RNA was prepared using the Absolutely RNA Microprep kit (Stratagene)from thymic stromal cell preparations depleted of CD45� hemopoieticcells by anti-CD45 bound magnetic beads or from isolated stromal subsetsobtained by cell sorting. Equal amounts of RNA were subjected tooligo(dT)-primed reverse transcription using SuperScript II (InvitrogenLife Technologies), following the manufacturer’s protocol. Equal amountsof cDNA from each sample were added to a final 25-�l PCR mix using theQiagen TaqPCR kit. The following PCR conditions were used: initial de-naturation of 94°C 5 min, followed by either 40 cycles (transgene), 35cycles (promiscuously expressed genes and AIRE), or 33 cycles (RelB, K5,chemokines, and actin) of 94°C for 0.5–1 min, 55–60°C for 0.5–1 min, and72°C for 1 min, and final extension of 72°C for 5 min. The followingprimer sequences were used: AIRE, (forward) 5�-GAC CTA AAC CAGTCC CGG AA-3�, (reverse) 5�-ATC CCT TCC ACG GCC CCT-3�; sali-vary protein 1, (forward) 5�-GGC TCT GAA ACT CAG GCA GA-3�,(reverse) 5�-TGC AAA CTC ATC CAC GTT GT-3�; C-reactive protein,(forward) 5�-CCA TGG AGA AGC TAC TCT G-3�, (reverse) 5�-CCCAAG ATG ATG CTT GC-3�; fatty acid binding protein, (forward) 5�-AGACGG AAC GGA GCT CAC-3�, (reverse) 5�-GCT CTT CAG CGT TGCTCC-3�; insulin, (forward) 5�-AGA CCA TCA GCA AGC AGG TC-3�,(reverse) 5�-CTG GTG CAG CAC TGA TCC AC-3�; actin, (forward) 5�-GTT TGA GAC CTT CAA CAC C-3�, (reverse) 5�-GTG GCC ATC CCTGCT CGA AGT C-3�; CCL19, (forward) 5�-GCT AAT GAT GCG GAAGAC TG-3�, (reverse) ACT CAC ATC GAC TCT CTA GG-3�; CCL21,

(forward) 5�-GCT GCC TTA AGT ACA GCC AG-3�, (reverse) 5�-GTGTCT GTT CAG TTC TCT TGC-3�; RelB, (forward) 5�-CAA GAA GTCCAC CAA CAC ATC-3�, (reverse) 5�-GGA AGT GGT CCA AGA ACACTG-3�; IKK� transgene, (forward) 5�-GCC ATG TAC CCA TAC GATGTT CC-3�, (reverse) 5�-GCT CCA ATA ATC AAC AGT GGC TG-3�;and K5, (forward) 5�-GAT GCT GCC TAC ATG AAC AAG-3�, (reverse)5�-TCC AGC TCT GTC AGC TTG TT-3�.

ResultsIKK� is required for development of mTECs and DCs

IKK��/� mutant mice die shortly after birth and display numerousdevelopmental defects including failure of limb development, ab-normal skeletal patterning and hyperplasia of the suprabasal epi-dermal layer resulting from a block in keratinocyte differentiation(24, 25, 30). In addition, we invariably observed thymus hypopla-sia in newborn IKK��/� mice (Fig. 1A). Despite a marked reduc-tion in cellularity, (3.3 � 0.3 � 106 IKK��/� compared with8.9 � 0.4 � 106 IKK��/� thymocytes), the relative percentage ofthymocyte subsets defined by CD4 and CD8 expression was notaltered in the absence of IKK� (data not shown). However, in-spection of H&E stained sections revealed an absence of distinctmedullary regions in the IKK��/� thymus (Fig. 1A). Immunohis-tochemical staining with Abs to mouse K8 and K5 showed a wellorganized network of K8�K5� cTECs in the IKK��/� newbornthymus that appeared comparable to cTECs in IKK��/� littermatecontrols. In contrast, mTEC development was severely impaired inthe IKK��/� thymus as shown in Fig. 1B. There was a markedreduction in K8�K5� mTECs, and those that were present ap-peared to be condensed into atypical, compact clusters. In addition,there was an almost complete absence of the UEA-1 bindingmTEC subset. These findings are consistent with a recent reportdescribing the medullary phenotype in an independently derivedIKK��/� line (31). Overall, the TEC phenotype in IKK��/� miceis similar to that reported in LT�R�/�, NIKaly/aly, and RelB�/�

mice (19, 20, 22). However, in contrast to LT�R�/� and NIKaly/aly

mice, there is a marked depletion of thymic DCs in IKK��/� andRelB�/� mice.

Expression levels of AIRE and TRA transcripts were comparedin IKK��/� and IKK��/� thymic stromal cells that were depletedof CD45� hemopoietic cells (Fig. 2). As predicted from the scar-city of mTECs in the IKK��/� thymus, semiquantitative RT-PCR

FIGURE 1. IKK� is required for normal development of mTECs andDCs. A, H&E stained frozen sections (5 �m) revealed thymus hypoplasiaand absence of medullary regions in the IKK��/� thymus compared withwild-type littermate thymus at original magnification �40 (top) and atmagnification �100 (bottom). B, Serial frozen sections were stained withAbs to K8, K5, and CD11c and fluorochrome-conjugated second step re-agents or biotinylated UEA-1 and streptavidin-FITC as indicated. Imagesshown are representative of five different thymuses of each genotype atoriginal magnification of �100.

FIGURE 2. Reduced AIRE, TRA, and chemokine expression inIKK��/� thymic stromal cells. Semiquantitative RT-PCR analysis wasperformed using RNA obtained from CD45-depleted stromal cells fromIKK��/� and IKK��/� newborn thymuses. Three-fold serial dilutions ofcDNA were used. �-actin served as a loading control for each sample. Arepresentative RT-PCR analysis from three experiments is shown.

831The Journal of Immunology

by guest on February 1, 2019http://w

ww

.jimm

unol.org/D

ownloaded from

analysis of RNA obtained from thymic stromal cells revealed areduction in expression of AIRE and the AIRE-dependent TRAsinsulin, salivary protein 1, and fatty acid binding protein. Interest-ingly, expression of C-reactive protein, an AIRE-independent TRAwas also reduced, consistent with the general depletion of mTECsin IKK��/� mice. CCL19 and CCL21, chemokines that are pro-duced predominantly by mTECs and attract positively selectedthymocytes into the medulla, were expressed at low levels in theIKK��/� thymus (32, 33). Finally, RT-PCR analysis of IKK��/�

stromal cells also revealed diminished expression of RelB, a down-stream target of IKK�.

Impaired mTEC development in IKK��/� embryos is a cellautonomous defect

Because thymocyte-derived signals are required for developmentand maintenance of the medullary compartment (34, 35), themTEC deficiencies in IKK��/� mice could be an indirect conse-quence of defective cross-talk between thymocytes and mTECprogenitors. To explore this issue, embryonic day 15.5 fetal thy-mus lobes from IKK��/� or IKK��/� littermates were depletedof thymocytes by 2-deoxyguanosine treatment and transplantedunder the kidney capsule of athymic nude (NCrnu/nu) mice. IKK�sufficient hemopoietic precursors in the recipient bone marrow mi-grate to and differentiate within the stromal microenvironment ofthe thymus graft. As shown in Fig. 3, the IKK��/� thymus graftsrecovered 8 wk after transplantation contained well-organizedmedullary regions with an abundant network of stellate K5�

mTECs that coexpress K14, a typical K5 binding partner. Thepresence of UEA-1 binding mTECs intermingled among the

K5�K14� cells was further evidence of normal medullary epithe-lial development in the IKK��/� thymus graft. In striking con-trast, the transplanted IKK��/� thymus contained only small med-ullary-like regions that contained few K5�K14� mTECs. UEA-1binding mTECs were undetectable. These data demonstrate thatthe deficiency of mTEC development in IKK��/� mice is an ep-ithelial cell autonomous defect. Conversely, DCs were present inthe IKK��/� thymus grafts suggesting that the absence of IKK�in hemopoietic precursors was a major factor responsible for theDC deficiency in IKK��/� mice.

IKK� is required for development of self-tolerance

Recent studies have shown that mTECs play a critical role in es-tablishing central tolerance by promiscuous expression of TRAsthat negatively select autoreactive T cell clones (7, 8, 17, 36).Consistent with this notion, impaired negative selection and auto-immunity are associated with experimental models in whichmTEC development is defective due to a block in the nonclassical

FIGURE 3. Impaired mTEC development in IKK��/� mice is a cellautonomous defect. Embryonic day 15.5 thymic lobes from IKK��/� andIKK��/� embryos were incubated for 4 days in the presence of 2-deoxy-guanosine and transplanted under the kidney capsule of athymic recipients.Eight weeks later, the transplanted thymuses were recovered and stainedwith H&E, Abs to K5, K8, K14, and CD11c, and the lectin UEA-1 asindicated. The cortex (C) and medulla (M) are indicated. IKK��/� thymusgrafts contained well-organized medullary regions with K5� and UEA-1binding mTEC subsets, whereas IKK��/� thymus grafts contained fewK5� and undetectable UEA-1 binding mTECs. The original magnificationwas �100. The images shown are representative of four different thymusgrafts.

FIGURE 4. Activated peripheral T cells and impaired central tolerancein athymic recipients of IKK��/� fetal thymus grafts. A, Flow cytometricanalysis of CD4� and CD8� T cells in peripheral lymph nodes from re-cipients of IKK��/� or IKK��/� fetal thymus lobes. B, Electronic gateswere set on CD4� lymph node T cells to determine the frequency ofCD44high and CD62Llow cells in recipients of IKK��/� (f) or IKK��/�

(�) fetal thymus grafts. C, H&E stained formalin-fixed sections of liverand pancreas from recipients of IKK��/� and IKK��/� fetal thymusgrafts. Arrows indicate lymphoid cell infiltrates. The results shown arerepresentative of four experiments.

832 THYMUS MEDULLA FORMATION AND CENTRAL TOLERANCE

by guest on February 1, 2019http://w

ww

.jimm

unol.org/D

ownloaded from

NF-�B signaling pathway (e.g., LT�R�/�, NIKaly/aly, RelB�/�

mice) (19, 20, 22, 23). Thus, it was not surprising to find activatedperipheral T cells and autoimmune manifestations in NCrnu/nu re-cipients of IKK��/� fetal thymus grafts. As shown in Fig. 4A bothIKK��/� and IKK��/� fetal thymus grafts support the develop-ment of CD4� and CD8� lymph node T cells. However, CD4�

peripheral T cells from recipients of IKK��/�, but not IKK��/�,grafts displayed an activated phenotype based on increased CD44and decreased CD62L expression (Fig. 4B). The IKK��/� thymusgrafts also appeared to be inefficient in supporting the generationof peripheral Tregs because expression levels of Foxp3 were re-duced in isolated CD4�CD25� splenic T cells from recipients ofIKK��/� compared with wild-type thymus grafts (compare withFig. 7D). Furthermore, there was a low frequency of Foxp3 ex-pressing cells detected by immunohistochemistry in IKK��/� thy-mus grafts (data not shown). IKK��/� graft recipients producedhigh levels of serum ANA (see Fig. 7C) and developed prominentperivascular lymphocytic infiltrates in liver and pancreas (Fig. 4C).These results suggest that impaired mTEC development in theIKK��/� thymus results in the generation of activated peripheralT cells, reduced Tregs and a profound deficit in the establishmentof central tolerance.

A K5 promoter-driven IKK� transgene rescues medullaformation in IKK��/� mice

It has been suggested that increased expression of CD80, MHCclass II, and UEA-1 binding molecules correlates with increasedmTEC maturation, although a direct lineage relationship betweenphenotypically distinct mTEC subsets has not been established(17). Moreover, it is not clear whether maturation and/or functionof mTECs depends on activating the alternative NF-�B pathway inboth the K5� and UEA-1 binding mTEC subsets, or alternatively,whether RelB activation in the K5� mTEC subset is sufficient todirectly or indirectly induce development of UEA-1 bindingmTECs. To address this issue, we asked whether restoring IKK�expression exclusively in the K5� mTEC subset would rescuegeneration of UEA-1 binding mTECs. To this end, we generated anovel transgenic mouse model in which expression of a humanIKK� transgene is expressed under the control of a K5 promoter.The K5-IKK� transgenic mice are healthy and fertile. Fig. 5Ashows that transgene expression in the thymus is restricted toCD45-negative thymic stromal cells and is not detected by RT-PCR analysis in CD45-positive hemopoietic cells. CD45-negativethymic stromal cells from K5-IKK� transgenic mice were incu-bated with FITC-conjugated UEA-1 and FACS sorted into UEA-1neg-low, UEA-1int, and UEA-1high populations (Fig. 5B). Semi-quantitative RT-PCR analysis revealed that endogenous K5 aswell as the IKK� transgene were highly expressed in the UEA-1neg-low population (Fig. 5C). Importantly, neither K5 nor the

FIGURE 5. K5 and TRA expression in K5-IKK� thymic stromal cellsisolated by differential UEA-1 binding. A, RT-PCR analysis of IKK� trans-gene expression in FACS sorted CD45� thymic stromal cells and CD45�

hemopoietic cells from 4-wk-old K5-IKK� transgenic mice. A represen-tative result from three individual animals is shown. B, CD45� thymicstromal cells from 4-wk-old K5-IKK� transgenic thymuses were pooledand incubated with FITC-conjugated UEA-1 and FACS sorted using theindicated gates to obtain UEA-1neg-low (59% of stromal cells), UEA-1int

(26% of stromal cells), and UEA-1high (14% of stromal cells) populations.C, RT-PCR analysis of isolated UEA-1neg-low, UEA-1int, and UEA-1high

cells for IKK� transgene, K5, AIRE, and TRA expression. A representativeresult from three similar experiments is shown.

FIGURE 6. The K5-IKK� transgene rescues medullary development inIKK��/� mice. A, Frozen thymus sections from newborn IKK��/�, K5-IKK�, and K5-IKK�/IKK��/� littermates were stained with H&E (upper)at original magnification of �40 for Abs specific for K8, K5, K14, andbiotinylated UEA-1 as indicated. Middle and lower panels, Original mag-nification, �100. Images shown are representative of thymuses from fiveindividual mice of each genotype. B, Semiquantitative RT-PCR analysiswas performed using RNA obtained from CD45 depleted stromal cellsfrom K5-IKKa transgenic and K5-IKKa/IKKa�/� mice newborn thy-muses. Three-fold serial dilutions of cDNA were used. �-actin served as aloading control for each sample. A representative RT-PCR analysis fromtwo experiments is shown.

833The Journal of Immunology

by guest on February 1, 2019http://w

ww

.jimm

unol.org/D

ownloaded from

IKK� transgene were expressed at detectable levels in UEA-1high cells. Consistent with a previous report (8), AIRE andAIRE-dependent TRAs salivary protein 1 and fatty acid bindingprotein were expressed at highest levels in UEA-1high TECs. Incontrast, C-reactive protein, an AIRE-independent TRA was ex-pressed at similar levels in both the UEA-1neg-low and UEA-1high cells. These data are consistent with recent detailed anal-yses showing a positive correlation between the frequency anddiversity of AIRE-dependent TRA expression and high levelexpression of markers such as CD80, UEA-1 binding, and classII expression by mTECs (8, 17).

There were no obvious abnormalities in thymocyte or TEC de-velopment in newborn K5-IKK� transgenic mice. Analysis ofH&E stained sections revealed that the K5-IKK� transgenic thy-muses contained well demarcated cortical and medullary regionscontaining cTEC and mTEC subsets (Fig. 6A). Because thymuscellularity and thymocyte subset distribution were comparable inK5-IKK� transgenic and nontransgenic newborn littermates (datanot shown), the K5-IKK� transgene was introduced into IKK��/�

mice, which were then intercrossed to generate IKK��/� progenythat express the K5-IKK� transgene (referred to as K5-IKK�/IKK��/� mice). Introduction of the K5-IKK� transgene into new-born IKK��/� mice resulted in a modest increase in thymus cel-lularity to �50% (5.1 � 0.7 � 106) of wild-type numbers.However, expression of the K5-IKK� transgene in the IKK��/�

background restored medullary development as shown in H&Estained thymus sections from newborn K5-IKK�/IKK��/� mice.These medullary regions contain DCs (data not shown) and TECs(Fig. 6A). The recovery of morphologically normal K5�K14�

mTECs was not surprising given that a K5 promoter directs ex-pression of the IKK� transgene. Interestingly, the K5-IKK� trans-gene also rescues development of the UEA-1 binding mTEC sub-set. Moreover, Fig. 6B shows that thymocyte depleted stromal cellsfrom K5-IKK�/IKK��/� newborn mice express TRAs at levelscomparable to those of K5-IKK� cells.

Central tolerance is restored in IKK��/� mice that expressa K5-IKK� transgene

To determine whether the K5-IKK� transgene also rescued mTECfunction with respect to induction of central tolerance, we trans-planted K5-IKK�/IKK��/� fetal thymuses under the kidney cap-sule of athymic recipients. This approach was necessary becausethe K5-IKK�/IKK��/� mice die shortly after birth. Fig. 7A showsthat well organized medullary regions containing K5� and UEA-1binding mTEC subsets were generated in the thymus grafts recov-ered after 6 wk. Furthermore, analysis of CD44 and CD62L ex-pression on lymph node T cells revealed that in contrast to recip-ients of IKK��/� thymus grafts, peripheral T cells generated inrecipients of K5-IKK�/IKK��/� thymus grafts did not display an

FIGURE 7. Central tolerance is restored in IKK��/�

mice that express a K5-IKK� transgene. A, Embryonicday 15.5 thymic lobes from K5-IKK�/IKK��/� andwild-type embryos were transplanted under the kidneycapsule of athymic recipients. Six weeks later the trans-planted thymuses were recovered and stained as indi-cated. Staining for TEC subsets in the K5-IKK�/IKK��/� transplanted thymus is shown; TEC deve-lopment in the wild-type transplanted thymus was com-parable to the results in Fig. 3 (data not shown). Theexperiment was repeated twice with the same results.B, Flow cytometric analysis of CD4� lymph node Tcells from athymic recipients that were grafted with ei-ther K5-IKK�/IKK��/� (f) or wild-type (�) embry-onic day 15.5 thymus lobes. Electronic gates were seton CD4� lymph node cells to analyze the frequency ofCD44high and CD62Llow CD4� T cells. The valuesshown are the average of three different experiments.C, Serum ANA were detected by ELISA in serum fromathymic recipients grafted with wild-type (f), IKK��/�

(u), or K5-IKK�/IKK��/� (�) embryonic day 15.5thymus lobes. Absorbance is directly proportional to theamount of ANA present in each sample. Results areexpressed as mean � SD of values obtained in two in-dependent experiments. D, RT-PCR analysis of Foxp3expression in FACS sorted CD4�CD25� splenocytes.E, H&E staining of formalin-fixed liver and pancreassections from athymic recipients of K5-IKK�/IKK��/�

thymic lobes. Note the absence of lymphoid cell infil-trates in comparison to the marked infiltration observedin recipients of IKK��/� thymus lobes (compare withFig. 4C).

834 THYMUS MEDULLA FORMATION AND CENTRAL TOLERANCE

by guest on February 1, 2019http://w

ww

.jimm

unol.org/D

ownloaded from

activated phenotype (Fig. 7B). Similar expression of Foxp3 tran-scripts in CD4�CD25� splenic T cells from recipients of K5-IKK�/IKK��/� and wild-type thymus grafts suggested compara-ble production of peripheral Tregs (Fig. 7D). This notion wassupported by immunohistochemical analysis of Foxp3-expressingcells in K5-IKK�/IKK��/� fetal thymus grafts (data not shown).The level of ANA detected in serum from recipients of K5-IKK�/IKK��/� was reduced by �60% compared with the titer of ANAin serum from recipients of IKK��/� thymus grafts (Fig. 7C).Importantly, there was no evidence of multiorgan inflammatoryinfiltrates in recipients of K5-IKK�/IKK��/� thymus grafts (Fig.7E). Taken together, these data demonstrate that expression of aK5-directed IKK� transgene in IKK��/� mice restores develop-ment of UEA-1 binding as well as K5� mTECs and provides amicroenvironment that supports central tolerance.

DiscussionThis investigation demonstrates that expression of a K5 promoter-driven IKK� transgene in IKK��/� mice rescues development ofthe mTEC compartment and restores the ability of the thymic stro-mal microenvironment to generate self-tolerant T cells. Previousstudies using various mutant mouse strains have shown that acti-vation of the nonclassical NF-�B signaling pathway in the thymicepithelial compartment is not only necessary for development ofmTECs, particularly the UEA-1 binding subset, but also is re-quired to induce central tolerance (19, 20, 22, 23, 31). However,the earlier studies did not determine whether the requirement forNF-�B signaling pertains to one or both major mTEC subsets. Thepresent report shows that selective expression of an IKK� trans-gene in the K5-expressing mTEC subset rescues development ofthe UEA-1 binding mTEC subset and restores central tolerance inIKK��/� mice. These data implicate the K5-positive subset in thegeneration of UEA-1 binding mTECs, although precise lineagerelationships among TEC subsets remain obscure. In a pivotalstudy involving lineage tracing analysis and pharyngeal endodermtransplantation, Gordon et al. (37) demonstrated that cTECs andmTECs are derived entirely from third pharyngeal pouchendoderm confirming classical chick-quail chimera experiments(37, 38). An endoderm-derived epithelial progenitor in the fetalthymus that coexpresses MTS20 and MTS24 surface Ags wasshown to generate cortical and medullary epithelial compartments(39, 40), and recent reports showing cTEC and mTEC develop-ment from single cells strengthen the argument for a common bi-potent progenitor (41, 42). The data presented in this report sug-gest that K5-expressing mTECs are either direct precursors ofUEA-1 binding cells or indirectly induce the maturation of UEA-1binding mTECs from immature progenitors. In either case, acti-vation of the NF-�B signaling pathway is mandatory only in theK5-expressing mTEC subset.

The mTEC deficiencies that we observed in IKK��/� mice aresimilar to those found in a recent report describing an indepen-dently generated IKK��/� mouse line (31). The present investi-gation extends the phenotypic analysis by showing that CD11c�

thymic DCs are absent or severely depleted in IKK��/� mice.Impaired TEC differentiation and autoimmune manifestations aretypical features in experimental models that lack components up-stream of IKK� in the alternative NF-�B pathway that results inRelB activation or in mice with a targeted deletion in RelB (19, 20,22, 23). The mechanistic basis linking medullary developmentwith the generation of self-tolerant T cells was recently clarified bystudies showing that mTECs express high levels of AIRE and adiverse array of TRAs that are required to eliminate self-reactivethymocytes by negative selection (7, 8, 17, 43). Given that AIREand TRAs are highly expressed by mTECs (8, 17), the paucity of

mTECs in the IKK��/� thymus is likely responsible for the ob-served reduction in AIRE and TRA expression. This interpretationis consistent with a recent report showing that NIK does not exertdirect transcriptional control on genes encoding TRAs (31). Ex-pression of CCL19 and CCL21 was also reduced in the IKK��/�

thymus. Again, this finding is not surprising because mTECs arethe primary source of CCL19 and CCL21 in the thymus (33).Moreover, RelB:p52 dimers have been shown to bind the promoterregions of various chemokine loci including CCL19 (18). Thymo-cytes from mice that are deficient in CCL19 and CCL21 or theircorresponding receptor, CCR7, accumulate in the cortex after pos-itive selection (33). Although single positive thymocytes fromthese mice undergo normal maturation and are exported to theperiphery, exocrinopathy develops in lacrimal and salivary glandsindicating that impaired migration of single positive thymocytesinto the medulla interferes with the establishment of central toler-ance (44). Therefore, it is possible that a deficiency in chemokineproduction interferes with migration of positively selected thymo-cytes into the poorly developed medullary areas of IKK��/� miceand is an additional factor that may further compromise centraltolerance in this model.

IKK��/� fetal thymus grafts placed under the kidney capsule ofathymic recipients consistently generated autoimmune-like symp-toms including activated peripheral T cells, high titers of ANA andprominent inflammatory infiltrates in liver and pancreas. The au-toimmune phenotype developed despite the generation of DCs inthe IKK��/� thymus grafts. DCs have been shown to play animportant role in establishing central tolerance to self-Ags by ac-quiring and cross-presenting peptides from mTECs (9). Cross-pre-sentation is likely to be inefficient in the IKK��/� thymus giventhe paucity of mTECs and corresponding reduction in TRA ex-pression. Deficient cross-presentation could explain the failure toachieve central tolerance even though DCs develop in IKK��/�

thymus grafts. DCs also play a role in the generation of Tregs, andmTECs have recently been shown to regulate this process. In thehuman thymus, epithelial aggregates in the medulla (Hassall’s cor-puscles) produce thymic stromal lymphopoietin that activates im-mature DCs to up-regulate CD80 and CD86 expression (45). Acti-vated thymic DCs in turn promote the proliferation of CD4single positive thymocytes, a subset of which differentiate togenerate Foxp3-expressing Tregs (45). The poorly developedmedullary regions in IKK��/� thymus grafts contain fewTregs. Moreover, the diminished Foxp3 expression levels inperipheral CD4� T cells in recipients of IKK��/� thymusgrafts suggest a deficiency of peripheral Tregs. Thus, a reduc-tion in the output of Tregs as a consequence of impaired mTECdevelopment is another factor that could contribute to the au-toimmune phenotype experienced by recipients of IKK��/� fe-tal thymus grafts.

Two models have been proposed to explain the expression of awide ranging spectrum of TRAs in medullary epithelial cells. TheKyewski group (17, 46) proposes a terminal differentiation modelin which TRA expression is induced as a result of derepressedgene expression during mTEC maturation. This model is based onthe assumption that increasing expression of MHC class II, CD80,and UEA-1 binding molecules is coincident with mTEC matura-tion as well as with data showing increased expression and com-plexity of TRAs expressed by CD80high or UEA-1high mTECs (8,17, 46). If this model is correct, then the recovery of UEA-1 bind-ing cells in K5-IKK�/IKK��/� mice implies that IKK�-mediatedNF-�B signaling in the K5� mTEC subset initiates a maturationprocess culminating in the generation of mature UEA-1high

mTECs that express a diverse array of TRAs. It is not possible to

835The Journal of Immunology

by guest on February 1, 2019http://w

ww

.jimm

unol.org/D

ownloaded from

ascertain from the present data whether K5� mTECs are immedi-ate precursors of UEA-1high cells. However, a recent study show-ing that medullary epithelium is composed of discrete islets eachof which is derived from a single precursor is consistent with thenotion of a precursor-progeny relationship between the K5� andUEA-1 binding mTEC subsets (47).

The Farr group (48, 49) proposes a developmental model pos-iting that mTECs contain a population of epithelial progenitors thathave not committed to a thymic epithelial lineage. As a conse-quence of developmental plasticity, these progenitors maintain thepotential to differentiate into various epithelial lineages and there-fore, express lineage-specific TRAs in response to proximal mi-croenvironmental signals. In support of this model, Dooley et al.(50) found cystic structures in the normal thymus medulla that aresurrounded by ciliated epithelial cells and express TRAs associ-ated with respiratory epithelium. The data in the present report canbe interpreted within the context of the developmental model con-sidering our previous studies demonstrating that TEC progenitorsare present in a unique subset of K8�K5� TECs localized at thecorticomedullary junction (15, 29). Assuming that multipotentialprecursors are also included in the K8�K5� subset, expression ofthe K5-IKK� transgene and consequent IKK�-mediated NF-�Bsignaling could promote the induction of individual epithelial pre-cursors into one of several potential differentiation programs re-sulting in concomitant TRA expression. Presumably, this processwould also induce differentiation of epithelial cells committed to aTEC fate and in the process generate UEA-1 binding cells, al-though not necessarily through a K5�K8� mTEC intermediate.

The terminal differentiation and developmental models are notmutually exclusive. Regardless of the correct scenario, this reportreveals new information on the differential requirement for IKK�-mediated NF-�B signaling in generating the two major medullarysubsets distinguished by K5 expression and UEA-1 binding. Weconclude that medulla formation and central tolerance depend onactivation of the IKK�-dependent alternative NF-�B signalingpathway in K5� epithelial cells. Activation of this signaling path-way induces K5-expressing progenitors to directly or indirectlygenerate the UEA-1high mTEC subset that expresses high levels ofAIRE and a diverse TRA repertoire. UEA-1high mTECs may alsoplay a role in thymic DC activation and generation of Tregs.

AcknowledgmentsWe gratefully acknowledge the helpful advice and assistance provided byCarla Carter, Monica Zamisch, and Ann Griffith. We thank Kent Claypoolfor performing flow cytometric analysis, Joi Holcomb for assistance withfigures, and Becky Brooks for manuscript preparation.

DisclosuresThe authors have no financial conflict of interest.

References1. Anderson, G., and E. J. Jenkinson. 2001. Lymphostromal interactions in thymic

development and function. Nat. Rev. Immunol. 1: 31–40.2. Hogquist, K. A., T. A. Baldwin, and S. C. Jameson. 2005. Central tolerance:

learning self-control in the thymus. Nat. Rev. Immunol. 5: 772–782.3. Bluestone, J. A., and Q. Tang. 2005. How do CD4�CD25� regulatory T cells

control autoimmunity? Curr. Opin. Immunol. 17: 638–642.4. Kronenberg, M., and A. Rudensky. 2005. Regulation of immunity by self-reac-

tive T cells. Nature 435: 598–604.5. Kyewski, B., and L. Klein. 2006. A central role for central tolerance. Annu. Rev.

Immunol. 24: 571–606.6. Villasenor, J., C. Benoist, and D. Mathis. 2005. AIRE and APECED: molecular

insights into an autoimmune disease. Immunol. Rev. 204: 156–164.7. Anderson, M. S., E. S. Venanzi, L. Klein, Z. Chen, S. P. Berzins, S. J. Turley,

H. von Boehmer, R. Bronson, A. Dierich, C. Benoist, and D. Mathis. 2002.Projection of an immunological self shadow within the thymus by the AIREprotein. Science 298: 1395–1401.

8. Derbinski, J., A. Schulte, B. Kyewski, and L. Klein. 2001. Promiscuous geneexpression in medullary thymic epithelial cells mirrors the peripheral self. Nat.Immunol. 2: 1032–1039.

9. Gallegos, A. M., and M. J. Bevan. 2004. Central tolerance to tissue-specificantigens mediated by direct and indirect antigen presentation. J. Exp. Med. 200:1039–1049.

10. Buhlmann, J. E., S. K. Elkin, and A. H. Sharpe. 2003. A role for the B7-1/B7-2:CD28/CTLA-4 pathway during negative selection. J. Immunol. 170:5421–5428.

11. Guiducci, C., B. Valzasina, H. Dislich, and M. P. Colombo. 2005. CD40/CD40Linteraction regulates CD4�CD25� T reg homeostasis through dendritic cell-pro-duced IL-2. Eur. J. Immunol. 35: 557–567.

12. Salomon, B., D. J. Lenschow, L. Rhee, N. Ashourian, B. Singh, A. Sharpe, andJ. A. Bluestone. 2000. B7/CD28 costimulation is essential for the homeostasis ofthe CD4�CD25� immunoregulatory T cells that control autoimmune diabetes.Immunity 12: 431–440.

13. Williams, J. A., S. O. Sharrow, A. J. Adams, and R. J. Hodes. 2002. CD40 ligandfunctions non-cell autonomously to promote deletion of self-reactive thymocytes.J. Immunol. 168: 2759–2765.

14. Farr, A. G., and S. K. Anderson. 1985. Epithelial heterogeneity in the murinethymus: fucose-specific lectins bind medullary epithelial cells. J. Immunol. 134:2971–2977.

15. Klug, D. B., C. Carter, E. Crouch, D. Roop, C. J. Conti, and E. R. Richie. 1998.Interdependence of cortical thymic epithelial cell differentiation and T-lineagecommitment. Proc. Natl. Acad. Sci. USA 95: 11822–11827.

16. Surh, C. D., E.-K. Gao, H. Kosaka, D. Lo, C. Ahn, D. B. Murphy, L. Karlsson,P. Peterson, and J. Sprent. 1992. Two subsets of epithelial cells in the thymicmedulla. J. Exp. Med. 176: 495–505.

17. Derbinski, J., J. Gabler, B. Brors, S. Tierling, S. Jonnakuty, M. Hergenhahn,L. Peltonen, J. Walter, and B. Kyewski. 2005. Promiscuous gene expression inthymic epithelial cells is regulated at multiple levels. J. Exp. Med. 202: 33–45.

18. Bonizzi, G., and M. Karin. 2004. The two NF-�B activation pathways and theirrole in innate and adaptive immunity. Trends Immunol. 25: 280–288.

19. Boehm, T., S. Scheu, K. Pfeffer, and C. C. Bleul. 2003. Thymic medullary epi-thelial cell differentiation, thymocyte emigration, and the control of autoimmu-nity require lympho-epithelial cross talk via LT�R. J. Exp. Med. 198: 757–769.

20. Kajiura, F., S. Sun, T. Nomura, K. Izumi, T. Ueno, Y. Bando, N. Kuroda, H. Han,Y. Li, A. Matsushima, et al. 2004. NF-�B-inducing kinase establishes self-tol-erance in a thymic stroma-dependent manner. J. Immunol. 172: 2067–2075.

21. Akiyama, T., S. Maeda, S. Yamane, K. Ogino, M. Kasai, F. Kajiura,M. Matsumoto, and J. Inoue. 2005. Dependence of self-tolerance on TRAF6-directed development of thymic stroma. Science 308: 248–251.

22. Burkly, L., C. Hession, L. Ogata, C. Reilly, L. A. Marconi, D. Olson, R. Tizard,R. Cate, and D. Lo. 1995. Expression of relB is required for the development ofthymic medulla and dendritic cells. Nature 373: 531–536.

23. Weih, F., D. Carrasco, S. K. Durham, D. S. Barton, C. A. Rizzo, R. P. Ryseck,S. A. Lira, and R. Bravo. 1995. Multiorgan inflammation and hematopoieticabnormalities in mice with a targeted disruption of RelB, a member of the NF-�B/Rel family. Cell 80: 331–340.

24. Hu, Y., V. Baud, M. Delhase, P. Zhang, T. Deerinck, M. Ellisman, R. Johnson,and M. Karin. 1999. Abnormal morphogenesis but intact IKK activation in micelacking the IKK� subunit of I�B kinase. Science 284: 316–320.

25. Takeda, K., O. Takeuchi, T. Tsujimura, S. Itami, O. Adachi, T. Kawai, H. Sanjo,K. Yoshikawa, N. Terada, and S. Akira. 1999. Limb and skin abnormalities inmice lacking IKK�. Science 284: 313–316.

26. DiDonato, J. A., M. Hayakawa, D. M. Rothwarf, E. Zandi, and M. Karin. 1997.A cytokine-responsive I�B kinase that activates the transcription factor NF-�B.Nature 388: 548–554.

27. Matsumoto, T., J. Jiang, K. Kiguchi, L. Ruffino, S. Carbajal, L. Beltran,D. K. Bol, M. P. Rosenberg, and J. DiGiovanni. 2003. Targeted expression ofc-Src in epidermal basal cells leads to enhanced skin tumor promotion, malignantprogression, and metastasis. Cancer Res. 63: 4819–4828.

28. Gray, D. H., A. P. Chidgey, and R. L. Boyd. 2002. Analysis of thymic stromalcell populations using flow cytometry. J. Immunol. Methods 260: 15–28.

29. Klug, D. B., E. Crouch, C. Carter, L. Coghlan, C. J. Conti, and E. R. Richie. 2000.Transgenic expression of cyclin D1 in thymic epithelial precursors promotesepithelial and T cell development. J. Immunol. 164: 1881–1888.

30. Hu, Y., V. Baud, T. Oga, K. I. Kim, K. Yoshida, and M. Karin. 2001. IKK�controls formation of the epidermis independently of NF-�B. Nature 410:710–714.

31. Kinoshita, D., F. Hirota, T. Kaisho, M. Kasai, K. Izumi, Y. Bando, Y. Mouri,A. Matsushima, S. Niki, H. Han, et al. 2006. Essential role of I�B kinase � inthymic organogenesis required for the establishment of self-tolerance. J. Immu-nol. 176: 3995–4002.

32. Ueno, T., K. Hara, M. S. Willis, M. A. Malin, U. E. Hopken, D. H. Gray,K. Matsushima, M. Lipp, T. A. Springer, R. L. Boyd, et al. 2002. Role for CCR7ligands in the emigration of newly generated T lymphocytes from the neonatalthymus. Immunity 16: 205–218.

33. Ueno, T., F. Saito, D. H. Gray, S. Kuse, K. Hieshima, H. Nakano, T. Kakiuchi,M. Lipp, R. L. Boyd, and Y. Takahama. 2004. CCR7 signals are essential forcortex-medulla migration of developing thymocytes. J. Exp. Med. 200: 493–505.

34. Shores, E. W., W. Van Ewijk, and A. Singer. 1994. Maturation of medullarythymic epithelium requires thymocytes expressing fully assembled CD3-TCRcomplexes. Int. Immunol. 6: 1393–1402.

35. Surh, C. D., B. Ernst, and J. Sprent. 1992. Growth of epithelial cells in the thymicmedulla is under the control of mature T cells. J. Exp. Med. 176: 611–616.

36. Liston, A., S. Lesage, J. Wilson, L. Peltonen, and C. C. Goodnow. 2003. Aireregulates negative selection of organ-specific T cells. Nat. Immunol. 4: 350–354.

836 THYMUS MEDULLA FORMATION AND CENTRAL TOLERANCE

by guest on February 1, 2019http://w

ww

.jimm

unol.org/D

ownloaded from

37. Gordon, J., V. A. Wilson, N. F. Blair, J. Sheridan, A. Farley, L. Wilson,N. R. Manley, and C. C. Blackburn. 2004. Functional evidence for a singleendodermal origin for the thymic epithelium. Nat. Immunol. 5: 546–553.

38. Le Douarin, N. M., and F. V. Jotereau. 1975. Tracing of cells of the avian thymusthrough embryonic life in interspecific chimeras. J. Exp. Med. 142: 17–40.

39. Bennett, A. R., A. Farley, N. F. Blair, J. Gordon, L. Sharp, and C. C. Blackburn.2002. Identification and characterization of thymic epithelial progenitor cells.Immunity 16: 803–814.

40. Gill, J., M. Malin, G. A. Hollander, and R. Boyd. 2002. Generation of a completethymic microenvironment by MTS24� thymic epithelial cells. Nat. Immunol. 3:635–642.

41. Bleul, C. C., T. Corbeaux, A. Reuter, P. Fisch, J. S. Monting, and T. Boehm.2006. Formation of a functional thymus initiated by a postnatal epithelial pro-genitor cell. Nature 441: 992–996.

42. Rossi, S. W., W. E. Jenkinson, G. Anderson, and E. J. Jenkinson. 2006. Clonalanalysis reveals a common progenitor for thymic cortical and medullary epithe-lium. Nature 441: 988–991.

43. Halonen, M., M. Pelto-Huikko, P. Eskelin, L. Peltonen, I. Ulmanen, andM. Kolmer. 2001. Subcellular location and expression pattern of autoimmuneregulator (Aire), the mouse orthologue for human gene defective in autoimmune-polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED). J. Histochem.Cytochem. 49: 197–208.

44. Kurobe, H., C. Liu, T. Ueno, F. Saito, I. Ohigashi, N. Seach, R. Arakaki,Y. Hayashi, T. Kitagawa, M. Lipp, et al. 2006. CCR7-dependent cortex-to-me-dulla migration of positively selected thymocytes is essential for establishingcentral tolerance. Immunity 24: 165–177.

45. Watanabe, N., Y. H. Wang, H. K. Lee, T. Ito, Y. H. Wang, W. Cao, and Y. J. Liu.2005. Hassall’s corpuscles instruct dendritic cells to induce CD4�CD25� regu-latory T cells in human thymus. Nature 436: 1181–1185.

46. Derbinski, J., and B. Kyewski. 2005. Linking signalling pathways, thymic stromaintegrity and autoimmunity. Trends Immunol. 26: 503–506.

47. Rodewald, H. R., S. Paul, C. Haller, H. Bluethmann, and C. Blum. 2001. Thymusmedulla consisting of epithelial islets each derived from a single progenitor.Nature 414: 763–768.

48. Farr, A. G., J. L. Dooley, and M. Erickson. 2002. Organization of thymic med-ullary epithelial heterogeneity: implications for mechanisms of epithelial differ-entiation. Immunol. Rev. 189: 20–27.

49. Gillard, G. O., and A. G. Farr. 2005. Contrasting models of promiscuous geneexpression by thymic epithelium. J. Exp. Med. 202: 15–19.

50. Dooley, J., M. Erickson, and A. G. Farr. 2005. An organized medullary epithelialstructure in the normal thymus expresses molecules of respiratory epithelium andresembles the epithelial thymic rudiment of nude mice. J. Immunol. 175:4331–4337.

837The Journal of Immunology

by guest on February 1, 2019http://w

ww

.jimm

unol.org/D

ownloaded from