T Cell Development from Kit-Negative Progenitors in the Foxn1∆/∆ Mutant Thymus Shiyun Xiao, Dong-ming Su and Nancy R. Manley This information is current as J Immunol 2008; 180:914-921; ; of September 30, 2021. doi: 10.4049/jimmunol.180.2.914 http://www.jimmunol.org/content/180/2/914 Downloaded from References This article cites 33 articles, 8 of which you can access for free at: http://www.jimmunol.org/content/180/2/914.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/ • Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on September 30, 2021 Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2008 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

T Cell Development from Kit-Negative Progenitors in the Foxn1⌬/⌬ Mutant Thymus1

Shiyun Xiao, Dong-ming Su,2 and Nancy R. Manley3

Foxn1⌬ is a hypomorphic allele of the nude gene that causes arrested thymic epithelial cell differentiation and abnormal thymic architecture lacking cortical and medullary domains. T cells develop in the Foxn1⌬/⌬ adult thymus to the double- and single- positive stages, but in the apparent absence of double-negative 3 (DN3) cells; however, DN3 cells are present in the fetal thymus. To investigate the origin of this seemingly contradictory phenotype, we performed an analysis of fetal and adult DN cells in these mutants. Neither adult bone marrow-derived cells nor fetal liver cells from wild-type or Rag1؊/؊ mice were able to differentiate to the DN2 or DN3 stage in the Foxn1⌬/⌬ thymus. Our data suggest that thymopoiesis in the Foxn1⌬/⌬ adult thymus proceeds from .CD117؊ atypical progenitors, while CD117؉ DN1a cells are absent or blocked in their ability to differentiate to the T lineage

Wild-type cells generated by this pathway in the postnatal thymus were exported to the periphery, demonstrating that these Downloaded from atypical cells contributed to the peripheral T cell pool. The Foxn1⌬/⌬ adult (but not fetal) thymus also preferentially supports B cell development, specifically of the B-1 type, and this phenotype correlated with reduced Notch ligand expression in the adult stroma. The Journal of Immunology, 2008, 180: 914–921.

he adult mouse thymus receives bone marrow (BM)4-de- recently generated a Foxn1 hypomorphic allele, Foxn1⌬, in which

rived hemopoietic progenitors, supports thymocyte pro- TEC differentiation initiates, but subsequently appears to arrest at http://www.jimmunol.org/ T liferation, differentiation, and repertoire selection, and an intermediate progenitor phenotype (8) corresponding to a thy- then releases the mature T cells to the thymic periphery (1). Most, mocyte-dependent stage of TEC differentiation (9). The resulting if not all, stages of thymocyte differentiation are mediated by the adult thymus does not have identifiable cortical or medullary re- thymic stroma, which is mainly composed of thymic epithelial gions, and most of the thymic epithelial cells retain a fetal pro- cells (TECs) (2). Although initial TEC differentiation in the fetal genitor-like phenotype. This TEC phenotype results in blocks at thymus is TEC intrinsic, continued TEC differentiation, and main- both the DN1 (CD44ϩCD25Ϫ) and DP thymocyte stages, with a tenance of adult thymic architecture is dependent on interactions striking loss of CD25ϩ DN2 and DN3 thymocytes. Interestingly, between TECs and thymocytes (2). This interdependence of thy- this thymocyte development phenotype was seen in adult, but not mocyte and TEC differentiation is referred to as thymic cross-talk fetal, stages; in the Foxn1⌬/⌬ fetal thymus, TEC differentiation was by guest on September 30, 2021 (3, 4). Although the stages of thymocyte differentiation, from the Ϫ Ϫ blocked, but thymocyte differentiation was only slightly delayed. immature double-negative (DN; CD4 CD8 DN), to double-pos- These results suggested that Foxn1 is required for both initial and itive (DP; CD4ϩCD8ϩ DP), then to functional single-positive (SP; later stages of TEC differentiation, and further showed that the CD4ϩ or CD8ϩ SP) T cells are well-known (5), the stages of TEC fetal and adult phenotypes were strikingly different. The difference differentiation are just beginning to be described (6). Additionally, in the fetal and adult thymocyte phenotypes and the extremely low the molecular mechanisms through which these stepwise interac- number of SP T cells in the adult thymus further raises the possi- tions mediate thymocyte and TEC differentiation are largely bility that the cells generated via this abnormal pathway in the unknown. The Foxn1 gene is known to be required for most, if not all, adult thymus might not be exported to the periphery at all. This TEC differentiation. TEC differentiation in the Foxn1null mutant would suggest that the majority of cells in the adult mouse may mouse, nude (␯), is arrested at a very immature stage (7). We originate from persisting cells that had developed in the initial postnatal period from fetal-derived progenitors. DN1 cells, as defined by the cell surface markers Department of Genetics, Coverdell Center, University of Georgia, Athens, GA 30602 CD44ϩCD25Ϫ, are considered heterogenous prothymocytes. Received for publication February 21, 2007. Accepted for publication November Since our initial report of the Foxn1⌬/⌬ mutant phenotype, five 11, 2007. subsets (DN1a, b, c, d, and e) of DN1 cells have been identified in The costs of publication of this article were defrayed in part by the payment of page the adult thymus, based on the expression of CD24 (HSA) and charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. CD117 (c-kit) markers (10). These subsets retain the distinct ca- 1 This work was supported by National Institutes of Health, National Institute of pacity to differentiate into several different types of cells, including Allergy and Infectious Disease Grant AI055001 (to N.R.M.). T cells, B cells, NK cells, and dendritic cells (10–12). Although all 2 Current address: Department of Biomedical Research, University of Texas Health five subsets have the capacity to generate T cells when cultured on Center, Tyler, TX 75708. the OP9-DL1 coculture system, only DN1a (CD117highCD24Ϫ) 3 Address correspondence and reprint requests to Dr. Nancy R. Manley, Department and DN1b (CD117highCD24ϩ) do so efficiently and with high pro- of Genetics, S270 Coverdell Center for Biomedical and Health Sciences, University of Georgia, Athens, GA 30602. E-mail address: [email protected] liferative capacity. The data support the conclusion that DN1a are 4 Abbreviations used in this paper: BM, bone marrow; TEC, thymic epithelial cell; apparently the “canonical” hemopoietic progenitor cells that mi- DN, double negative; SP, single positive; DP, double positive; FL, fetal liver; FTOC, grate in from BM, whereas DN1b cells are likely derived from fetal thymic organ culture. DN1a, and then take the normal T cell development pathway to Copyright © 2008 by The American Association of Immunologists, Inc. 0022-1767/08/$2.00 DN2. DN1a and b are also the only DN1 subsets to demonstrate www.jimmunol.org The Journal of Immunology 915

NK cell potential. The other subsets are able to take a T cell de- Ly5.2ϩ host cells were gated on Ly5.1-negative cells. To characterize the ⌬ ⌬ ϩ ⌬ velopmental pathway in vitro, albeit with low proliferation capac- profile of DN1 cells, thymocytes from adult Foxn1 / and Foxn1 / lit- ity and an abnormal development profile; DN1c and d also have B termate controls were deleted for most DP cells by gradient density cen- ϩ trifugation in 13.6% Opti-Prep (Greiner Bio-One) solution at 2000 rpm ϫ lineage potential. Thus, consistent with other reports, CD117 20 min. All lineage markers (anti CD3, CD4, CD8, CD11b, CD19, B220, cells define the primary progenitors that produce most, if not all, of Gr-1, TER-119, NK1.1) and anti-CD25 were PE conjugated (BD Pharm- the T cells in a wild-type adult thymus, but are a small minority of ingen or BioLegend). Biotin anti-CD44 following streptavidin-PerCP, the total DN1 population. As the abnormal T cell differentiation FITC-HSA, and allophycocyanin c-Kit Abs (BD Pharmingen) was used. Ϫ Four-color staining was used to analyze DN1 cells. DN1 cells were profiles shown by some of these CD117low/ populations in ⌬/⌬ gated on all lineages and CD25-negative and CD44-positive subpopula- vitro resemble the DN1 profile in the Foxn1 adult thymus, tions, and analyzed with HSA and c-Kit staining. For analysis of thymic B these results raise the possibility that differences in the differ- cells, single-cell suspensions were prepared from adult Foxn1⌬/⌬ and ϩ ⌬ entiation of the various DN1 subsets could underlie the Foxn1 / thymi and stained with PE-CD19, allophycocyanin-B220, and Foxn1⌬/⌬ thymocyte differentiation phenotype. However, pre- FITC-IgM or FITC-IgD or FITC-CD5 (BioLegend). All cells were ac- quired by dual-laser FACSCalibur system and analyzed by CellQuest soft- vious studies were unable to show that these other DN1 cell ware (BD Biosciences). populations can generate T cells in vivo (10). In the current study, we investigated the origins of the DN dif- Reconstitution of DN1 cells in 2dGuo-treated fetal thymic organ ferentiation defects in the Foxn1⌬/⌬ hypomorphic postnatal thy- culture (FTOC) Ϫ/Ϫ mus. Neither adult nor fetal wild-type or Rag1 BM-derived DN1 cells were enriched and sorted by MoFlo cytomation using PE-con- progenitors were able to differentiate to the DN2 or DN3 stages in jugated lineage markers as above (anti CD3, CD4, CD8, CD11b, CD19, ⌬ ⌬ ⌬ ⌬ the Foxn1 / thymus, supporting the conclusion that the Foxn1 / B220, Gr-1, TER-119, NK1.1), CD25-FITC and CD44 allophycocyanin. Downloaded from E15.5 day thymi were isolated and treated with 10% FBS plus RPMI 1640 thymic microenvironment cannot support the normal thymocyte medium containing 1.35 mM 2dGuo (Sigma-Aldrich) in a high oxygen development pathway via the DN3 stage. Wild-type T cells dif- culture system at 37°C for 5 days to deplete hemopoietic cells. Each dGuo- ferentiating by this pathway were exported to the periphery. treated fetal thymic lobe was put in a “hanging drop” coculture with 1000 CD177high DN1a/b “canonical” T cell progenitors did not contrib- sorted DN1 cells in total volume 25 ␮l of 10% FBS plus RPMI 1640 medium at 37°C, 5% CO for 48 h. Fetal thymi were then washed and ute to the T lineage. The thymocytes that develop into DP cells 2 transferred to a membrane floating on 10% FBS plus RPMI 1640 medium appear to arise primarily from atypical progenitors (DN1d and/or http://www.jimmunol.org/ and incubated at 37°C, 5% CO2 for 7 or 10 days. Cultured thymi were DN1e), which do not normally contribute significantly to thymo- ground through a mesh screen and single-cell suspensions were analyzed as poiesis in vivo. Other recent data from our laboratory shows that above. the peripheral T cell phenotype of cells developing in this pathway Immunofluorescence staining is also atypical (13). B and NK cell development was much more ϩ/⌬ ⌬/⌬ normal, although B cells preferentially developed along the B-1 Freshly isolated Foxn1 and B cell-deleted Foxn1 thymocytes were seeded on the coverslips coated with 150 ␮g/ml poly-L-lysine (Sigma- phenotype. Taken together, these data suggest that the microenvi- Ϫ ⌬/⌬ Aldrich). After washing, attached cells were fixed with methanol at 20°C ronment in the Foxn1 adult thymus generates T cells via an for 20 min. Cells were washed with PBS for three times, blocked by 5% atypical differentiation pathway that may represent a minor path- donkey serum plus 1% BSA in PBS at room temperature for 1 h. Cells were way in the normal thymus. incubated with rabbit anti-␥-H2AX Ab (1/1,000; Sigma-Aldrich) for 1 h. by guest on September 30, 2021 Rabbit serum at the same dilution was used as negative control. Cells were Materials and Methods washed and then incubated with donkey anti-rabbit Cy3 (1:300; Jackson ImmunoResearch Laboratories) for 1 h following nuclear staining with Mice 4Ј,6Ј-diamidino-2-phenylindole (1/10,000; Sigma-Aldrich) for 2 min. The Foxn1⌬/⌬ mice were generated and genotyped by PCR as described (8). coverslips were mounted with Aqua Poly/mount (Polysciences). Images Ϫ Ϫ ϩ ϫ Rag1 / Ly5.1 mice were provided by E. V. Rothenberg (California were acquired with a Zeiss ApoTome microscope, at a 40 objective, Institute of Technology, Pasadena, CA). All experiments were performed using Axio Vision Rel.6 software. using mice on a mixed 129SvJ, C57BL6/J genetic background. For timed Semiquantitative RT-PCR for Notch ligands and TCR␣ matings to generate embryos, the day of the vaginal plug was designated E0.5. All experiments using animals were performed with the approval of TRIzol (Invitrogen Life Technologies) was used to isolate total RNA from the University of Georgia Institutional Animal Care and Use Committee. the E15.5 thymi by FTOC with 1.35 mM dGUO (depleting hemopoietic Chimera generation cells), or from sorted adult DN thymocytes. Reverse transcription and PCR amplification were done by using the SuperScript II (Invitrogen Life Tech- Donor adult BM cells and fetal liver (FL) cells were isolated by lysing nologies) and Taq (Qiagen) or SuperScript III One-step RT-PCR with Plat- RBC in ACK-lysing buffer (Cambrex Bio Science). CD45ϩ BM cells (and inum Taq (Invitrogen Life Technologies). Each reaction contained 0.5 ␮g FL cells at age fetal day 15.5–18.5) from Rag1Ϫ/Ϫ Ly5.1ϩ mice were of RNA for Notch ligand genes or 0.025 ␮g for control CD45 or GAPDH purified by anti-CD45 microbeads and passing through a column (Miltenyi genes. Reactions were repeated at least twice using RNA from different Biotec). Sorted BM cell (5 ϫ 106/recipient) and FL cells (1ϳ2 ϫ 106/ fetal litters and different adult animals. Primers (14, 15) for Notch ligands recipient) were injected i.v. into sublethal irradiated (600 rad) Ly5.2ϩ were Delta-like-1: (forward) 5Ј-GTC ACA GAG CTC TGC AGG AG-3Ј; Foxn1⌬/⌬ or Foxn1ϩ/⌬ littermate control recipients (see Fig. 1). In Fig. 2, (reverse) 5Ј-TGT GGG CAG TGC GTG CTT CC-3Ј; Jagged-1: (forward) Adult BM cells from Bl6-Ly5.1 mice were first deleted for mature T cells 5Ј-CAT TAC GTG TTG CCT GTA AGC C-3Ј; (reverse) 5Ј-GTG GTT by incubating with rat-anti CD3, CD4, and CD8 Abs (BD Pharmingen), CAG CAT TAC ATA CG-3Ј; Delta-like-4: (forward) 5Ј-CAG AGA CTT followed by anti-rat IgG Dynal bead subtraction (Invitrogen Life Technol- CGC CAG GAA AC-3Ј; (reverse) 5Ј-ATC CAT TC TGC ACG GAG ogies). The T cell-deleted cells were further purified by anti-CD45 mi- AG-3Ј and Jagged-2: (forward) 5Ј-GTC CTT CCC ACA TGG GAG TT- crobeads and passed through a column for CD45ϩ cells. The T cell-deleted 3Ј; (reverse) 5Ј-GTT TCC ACC TTG ACC TCG GT-3Ј. Primers for pre- CD45ϩ BM cells were transferred into sublethally irradiated mice. Three TCR␣ were pT␣-F: 5Ј-CAG AGC CTC CTC CCC CAA CAG-5Ј;pT␣-R: weeks later, profiles of thymocytes and spleen cells from donor and host 5Ј-GCT CAG AGG GGT GGG TAA GAT-3Ј (16). cells were analyzed as above. Southern blot analysis of TCR␤ rearrangement Flow cytometric analysis and Abs TCR␤ gene rearrangement was assayed at the genomic level. Sorted DP A total of 0.5–1 ϫ 106 cells in suspension were used for each sample. Cells and DN adult thymocytes (5 ϫ 104 per PCR), and FL cells (for germline were blocked by anti-CD16/32 Ab (BioLegend) plus rat serum before control) were lysed in 5,000 cells/␮l PCR lysis buffer (17) with 0.1 mg/ml staining. For transfer experiments, PE anti-Ly5.1, biotin anti-mouse CD25 proteinase K at 55°C for 2 h, then 100°C for 10 min. PCR was performed following streptavidin-PerCP, and allophycocyanin anti-mouse CD44 (BD with primers of: (3Ј primer) J␤2: 5Ј-TGA GAG CTG TCT CCT ACT ATC Pharmingen), FITC anti-CD4 and CD8, or PE anti-CD4 (BioLegend) were CAT T-3Ј (18); (5Ј primer) D␤2: 5Ј-GTA GGC ACC TGT GGG GAA used. Donor cells were analyzed by gating live Ly5.1-positive cells; GAA ACT-3Ј (19); (5Ј primer) V␤5.1: 3Ј-CCC AGC AGA TTC TCA 916 T CELL DEVELOPMENT FROM kit-NEGATIVE PROGENITORS

Table I. Cell numbers recovered from FTOC day 10

Cell Number (1 ϫ103)

Seeded Cells Total Thymic Cells Total T Cells B Cells

ϩ/⌬ DN1 100 Ϯ 12 96.7 Ϯ 11.5 0.3 Ϯ 0.05 ⌬/⌬ DN1 14 Ϯ 1.3 1.3 Ϯ 0.25 2.4 Ϯ 0.41 None 7 Ϯ 0.8 0 0

Foxn1⌬/⌬ thymus is unable to support the normal thymocyte dif- ferentiation pathway to the DN3 stage (possibility no. 3 above). This result also confirms that DN1 cells from these mutants require TCR rearrangement to progress to the DN4 stage. Our previous analysis of Foxn1⌬/⌬ mutants had also shown that DN2 and DN3 cells were present in relatively normal numbers in the fetal thymus, even though TEC differentiation was blocked. To test the capacity of fetal thymocytes to differentiate in the adult ⌬/⌬ Foxn1 microenvironment, we performed a similar experiment Downloaded from using FL-derived CD45ϩLy5.1ϩ Rag1Ϫ/Ϫ progenitors. All of FIGURE 1. The thymic microenvironment in adult Foxn1⌬/⌬ mutant these cells were also blocked at the DN1 stage (Fig. 1b) (three mice failed to support progenitor cell development to DN2 and DN3 stages. experiments, total of eight mice injected with E15.5 or E18.5 cells, Sorted CD45ϩ progenitor cells derived from BM and FL of RagϪ/Ϫ Ly5.1 or with mixed E14.5, 15.5, and 17.5 cells). These results showed ⌬ ⌬ mice were i.v. transferred into sublethal irradiated adult Foxn1 / mutant that the Foxn1⌬/⌬ thymic microenvironment is completely nonper- or Foxn1ϩ/⌬ control mice. After 3 wk, donor (Ly5.1ϩ) and host (Ly5.1Ϫ)

missive for the DN2/3 stages of differentiation. http://www.jimmunol.org/ cells were analyzed. a, Expression of CD25 and CD44 on gated Ϫ Ϫ ϩ CD4 CD8 BM-derived cells. b, Similar analysis of CD45 FL-derived SP thymocytes develop in the Foxn1⌬/⌬ thymus and are cells. exported to the periphery This atypical differentiation pathway was associated with a reduc- Ј tion in SP cell production by at least 500-fold in the postnatal GTC CAA CAG-3 (19). PCR products were run on a 1.8% agarose gel, ⌬/⌬ denatured, neutralized, and transferred to Hybond Nϩ nylon membrane, Foxn1 thymus (Table I; Refs. 8 and 13). Other data from our then hybridized with a [␥-32P]ATP 5Ј-end labeled J␤2 probe 5Ј-TTT CCC laboratory showed that peripheral T cells in the adult have a phe- TCC CGG AGA TTC CCT AA-3Ј (18). notype similar to peripheral T cells in early postnatal stages (13). ⌬ ⌬ Furthermore, the Foxn1 / postnatal thymus is highly disorga- by guest on September 30, 2021 Results nized with no discernible cortical or medullary compartments ⌬/⌬ The adult Foxn1 thymic microenvironment fails to support (8). These results raised question of whether SP cells were ma- progenitor differentiation to DN2 and DN3 turing in the absence of an organized medulla, and whether they Our previously published data (8) showed that thymocyte devel- could actually exit the postnatal thymus. ⌬ ⌬ opment in adult Foxn1⌬/⌬ thymi is blocked at the DN1 To determine whether the postnatal Foxn1 / thymus was able (CD44ϩCD25Ϫ) stage. Despite a near total absence of DN2 and to export new naive T cells into the periphery, we performed sim- DN3 cells, DN4, DP, and SP cells were present, although in re- ilar transfer experiments using T cell-depleted wild-type donor Ϫ Ϫ duced numbers. There are several possible explanations for this BM cells (Fig. 2). Like the Rag1 / cells, these cells did not steady-state profile: 1) DN2/3 cells could be present, but in too low produce CD25high DN2/3 cells after transfer, only DN1, DN4, and numbers to reliably detect due to low efficiency of generation; 2) a minor CD25low subpopulation (Fig. 2a). The differentiation pro- thymocytes could be transiting the DN2/3 stages with unusually file of these cells was similar to the host thymocytes, and did rapid kinetics, then accumulating at DN4; or 3) thymocytes could produce DP and SP thymocytes (Fig. 2a), although very few of the progress directly from the DN1 to DN4 stages via an atypical transfer-derived cells were recovered. Analysis of CD4 and CD8 differentiation pathway. expression in the spleen also suggested that these cells were ex- To distinguish between these possibilities, we compared the ported to the periphery (Fig. 2b), although we cannot completely ability of Rag1Ϫ/Ϫ and wild-type progenitors to differentiate in the exclude the possibility that very low numbers of undepleted T cells Foxn1⌬/⌬ thymus. If DN cells could develop to the DN2/3 stages, in the donor marrow contributed to this result. Taken together, even at low efficiency, Rag1Ϫ/Ϫ thymocytes should accumulate at these data suggest that despite its highly abnormal thymic archi- the DN3 stage, revealing their presence in the Foxn1⌬/⌬ thymus. tecture and the atypical thymocyte differentiation profile, the ⌬ ⌬ Sorted CD45ϩ progenitors derived from Ly5.1ϩ BM were i.v. in- Foxn1 / adult thymus continues to export a low level of new jected into sublethally irradiated Foxn1ϩ/⌬ controls or Foxn1⌬/⌬ naive SP cells to the periphery. ϩ mutants (Ly5.2 congenic recipients). Differentiation of the donor high ⌬/⌬ and host cells in the thymus was analyzed at 3 wk postinjection by Absence of CD117 cells in Foxn1 adult thymus gating on the Ly5.1- and Ly5.2-positive populations and staining The failure of transferred Rag1Ϫ/Ϫ and wild-type or endogenous for CD44 and CD25. By 3 wk postinjection, Ͼ90% of Ly5.1ϩ progenitors to differentiate to the DN2/3 stages could be due to thymocytes in the control thymus had progressed to the DN2 or either failure of progenitor immigration or an inability of these DN3 stages (Fig. 1a). In contrast, all of the progenitors from progenitors to differentiate in the mutant thymic microenviron- Rag1Ϫ/Ϫ BM (two experiments, total six mice) failed to generate ment. To determine which DN1subpopulation is affected in DN3 cells, and were blocked at the DN1 stage in the Foxn1⌬/⌬ Foxn1⌬/⌬ mutants, we characterized the profile of freshly isolated, thymus (Fig. 1a). This result clearly demonstrated that the enriched DN thymocytes from adult Foxn1⌬/⌬ and their littermate The Journal of Immunology 917 Downloaded from

FIGURE 3. Absence of CD117high cells in Foxn1⌬/⌬ adult thymus. a, Characteristic plot of DN1a–e subsets. DN-enriched thymocytes were stained with nine PE-conjugated lineage markers and PE-CD25, CD44- http://www.jimmunol.org/ ϩ FIGURE 2. The thymic microenvironment in adult Foxn1⌬/⌬ mutant biotin plus avidin PerCP, CD117-allophycocyanin, HSA-FITC. CD44 , Ϫ Ϫ mice produces DP and SP thymocytes that contribute to the peripheral T lin , CD25 cells were gated as DN1. b, Reconstitution of sorted DN1 cell pool. Sorted T cell-depleted CD45ϩ progenitors derived from BL6 cells in FTOC. Cocultured thy1.2-positive cells were collected and ana- Ly5.1 BM cells were i.v. transferred into sublethally irradiated Foxn1⌬/⌬ lyzed by flow cytometry after 7 or 10 days of culture. Top panel, Expres- and Foxn1ϩ/⌬ adult mice. After 3 wk, gated donor Ly5.1ϩ cells (right sion of CD4, CD8; bottom panel, expression of CD44, CD25. c, Expression column) and host (Ly5.1Ϫ) cells (left column) were analyzed. a, Expression of thy1.2 and B220 at FTOC day 10. of CD25 and CD44 (top panels), and CD4 and CD8 (bottom panels)on thymocytes. b, Expression of CD4 and CD8 on peripheral T cells. ⌬ ⌬ ϩ ϩ Foxn1 / DN1 cultures were CD4 CD8 DP. This rapid devel- by guest on September 30, 2021 opment of DP cells was consistent with the previously reported controls. The DN1 population of control mice displayed five sub- developmental profile of the DN1c, d populations in vitro, which populations based on the expression of CD117 (c-kit) and CD24 should have been present in both cultures (10). On day 10, most ⌬ ⌬ (HSA) as previously described (Fig. 3a) (10). In contrast, very few DN cells in both control and Foxn1 / cultures had developed to or no CD117high cells were present in the Foxn1⌬/⌬ adult thymus, DN3 and DN4 stages (Fig. 3b), and DP and SP cells appeared in although some HSAlowckitlow cells were present in the DN1a both cultures, although the cell numbers in the mutant cultures quadrant defined in the control mice (Fig. 3a, right panels,R3 were extremely low. This profile is consistent with the conven- gate). The c-kitlow“DN1a” cells in the mutants could reflect an tional thymocyte developmental pathway from DN1 through DN3 inability of the ⌬/⌬ microenvironment to up-regulate or maintain to DP. The low cell numbers recovered are not entirely consistent high levels of c-kit on these cells, or reflect an absence of true with the reported high level of proliferative capacity in true DN1a DN1a cells and presence of separate class of atypical cells in these cells, although this reduced proliferation could reflect the low lev- mutants. As expected, the percentage of CD117high cells in BM els of c-kit expression on DN1a cells from the mutants. This result was not significantly different between Foxn1⌬/⌬ and their supports the possibility that at most a very few true DN1a precur- ⌬ ⌬ Foxn1ϩ/⌬ littermate controls (data not shown). Although the very sors exist in Foxn1 / mutant DN1 subpopulation. This low cell low numbers of DN1a cells made it difficult to determine conclu- recovery at FTOC day 10 (Table I) was also consistent with the sively whether they were present by this approach, DN1b cells low proliferative capacity of isolated DN1c and d subpopulations were clearly absent (Fig. 3a). These results suggested that progen- after coculture with the OP9-DL1 cell line in vitro (10). In addi- ϩ itors for the “canonical” DN1a-DN1b-DN2 pathway were either tion, ϳ17% of B220 cells appeared in total recovered cells from ⌬ ⌬ ϩ⌬ not present or were blocked in their differentiation along the usual reconstitution of Foxn1 / DN1 cells (only 0.3% in Foxn1 DN1 pathway, via the DN2 and DN3 stages. DN1c cells were also sig- cells control) (Fig. 3c), consistent with a higher percentage of B ⌬ ⌬ nificantly decreased, resulting in the majority of DN1 cells being committed precursors in the Foxn1 / DN1 subpopulation.

DN1d and DN1e subsets, which were previously shown to differ- ⌬/⌬ entiate in the T lineage with atypical kinetics and with marker Foxn1 DN4 cells contain DP precursors profiles similar to thymocyte differentiation in the Foxn1⌬/⌬ adult The data above suggest that DP cells in the Foxn1⌬/⌬ thymus are thymus (10). generated from atypical DN1d or DN1e progenitor cells, differen- To test the differentiation potential of the DN1 cells in the mu- tiating from DN1 to DN4, then through to DP and SP stages. As tants, and as a functional assay for the presence of DN1a cells, we DN4 cells are primarily defined by the absence of markers, we cocultured sorted DN1 cells from control or mutant adult thymus determined whether the DN4 cells in the Foxn1⌬/⌬ thymus gener- in an FTOC reconstitution assay (Fig 3b). On day 7, most cells ated via this atypical pathway represented authentic pre-DP cells from both genotypes were CD4ϪCD8Ϫ DN, but Ͼ10% of cells in by testing whether they could differentiate directly to the DP stage 918 T CELL DEVELOPMENT FROM kit-NEGATIVE PROGENITORS

Table II. Cell numbers of lymphoid cell subpopulations in the thymus

Cell Number (1 ϫ104)

Total Thymic Total T B-1 B-2 Genotype Cells Cells Cells Cells

Foxn1ϩ/⌬ 24,120 Ϯ 324 23,200 Ϯ 256 8.6 Ϯ 4.8 24.3 Ϯ 1.9 Foxn1⌬/⌬ 10.62 Ϯ 3.5 1.87 Ϯ 0.42 4.35 Ϯ 0.55 1.72 Ϯ 0.34

profile and at a frequency similar to that of control cultures (Fig. 4a). The low level of DP cells produced in this assay was due to the low density of seeding required by the poor recovery of cells from the mutant thymi, because increasing the density of seeding for control DN4 cells progressively increased DP cell production (Fig. 4a). If the thymocytes in the Foxn1⌬/⌬ thymus were differentiating to DN4 without transiting through a normal DN2 or DN3 stage, this

raises the question of when events that normally take place during Downloaded from these stages are occurring. TCR␤ rearrangement and pre-T␣ ex- pression normally initiate during the DN2 stage of thymocyte dif- ferentiation, with selection for surface expression of ␤TCR occur- ring at the DN3 stage. DP cells in the Foxn1⌬/⌬ thymus do have surface expression of ␣␤TCR, although at a lower level than in ␤ ␤ ␤ ␤ controls (8) (13). Analysis of D -J and V -J rearrangements in http://www.jimmunol.org/ sorted total DN thymocytes showed similar degrees of rearrange- ments in cells from control and Foxn1⌬/⌬ thymi (Fig. 4b). Al- though we set the negative gate for these cells conservatively, we cannot absolutely exclude the possibility that some SP cells con- taminated this sample. In contrast, D␤-J␤ rearrangement and a low level of Rag1 expression has been previously shown for DN1e cells (10), which are strongly represented in Foxn1⌬/⌬ DN1 cells. To further confirm that TCR rearrangements were taking place in the postnatal thymus, we used an Ab against histone ␥-H2AX, by guest on September 30, 2021 which form nuclear foci at sites of double-strand breaks associated with TCR recombination (21). Nuclei with these punctate accu- mulations of ␥-H2AX were found in Foxn1⌬/⌬ thymocytes (Fig. 4c). Pre-T␣ expression was also clearly present by RT-PCR in sorted DN thymocytes (Fig. 4d). Furthermore, total DN cells from Foxn1⌬/⌬ mutants had approximately the same level of cytoplas- mic TCR␤ protein as controls, even in the absence of DN2 and ⌬ ⌬ FIGURE 4. Characteristics of DN cells in Foxn1 / mutant mice. a, DN3 cells (data not shown). These results show that DN cells Ϫ Ϫ Ϫ ⌬/⌬ Sorted DN4 cells (Lin , CD25 , CD44 cells) derived from Foxn1 developing via this atypical differentiation pathway undergo many ϩ/⌬ mutant and Foxn1 mice were cultured in medium only. The expression of the same processes normally occurring during DN thymocyte of CD4 and CD8 were analyzed after 18 h. Top row, Results before and differentiation, including TCR rearrangement, without develop- after culture of control and mutant-derived cells seeded at 5 ϫ 103 cells/ well. Second row shows that the number of DP cells recovered from control ment through the DN3 stage. cultures depends on initial seeding density. b, TCR␤ gene rearrangement in ⌬ ⌬ B and NK cells in the Foxn1 / adult thymus sorted thymocytes. Southern blot analysis of TCR␤ chain D␤2-J␤2(left ⌬ ⌬ panel) and V␤5.1-J␤2(right panel) rearrangement in FL cells (lanes 1 and Because the results presented above showed that the Foxn1 / 6), sorted DP cells from Foxn1ϩ/⌬ (lanes 2 and 7) and Foxn1⌬/⌬ (lanes 4 mutant adult thymic microenvironment is not favorable to canon- ϩ ⌬ and 9) thymi, and sorted DN cells from Foxn1 / (lanes 3 and 8) and ical T cell differentiation, we investigated whether B and NK cell ⌬/⌬ Foxn1 (lanes 5 and 10) thymi. The 1.8-kb D␤2-J␤2 germline band for development in the thymus was affected as well. Overall, the per- ␤ ␤ ␥ D 2-J 2 is indicated with an arrow. c, Presence of punctate -H2AX stain- centage of B cells was greatly increased in Foxn1⌬/⌬ mice, reflect- ing in control and mutant thymocytes from 1-mo-old adult mice. Examples ing the reduction in thymocyte number, although variation in the of cells with one or two foci of H2AX staining indicative of active TCR ⌬/⌬ ϩ rearrangement are indicated with arrowheads. d, Expression of pT␣ mRNA percentage of thymic B cells in Foxn1 mice was large (CD19 ϳ in sorted DN thymocytes. cells: 4 90%). The total number of B cells was reduced, although not as severely as for thymocytes (Table II). We have recently shown that many of the SP thymocytes in the in suspension culture (20). We cultured 5 ϫ 103 sorted DN4 cells/ Foxn1⌬/⌬ adult mice are likely recirculated into the thymus from well (linϪ, CD44Ϫ25Ϫ, Thy1.2ϩ). To obtain sufficient numbers of the periphery (13). Therefore, the thymic B cells could also result DN4 cells from Foxn1⌬/⌬ mutants, cells from 10 adult mutant from peripheral B cells reentering the thymus, rather than differ- thymi were pooled. Controls contained the same number of sorted entiating in situ. To investigate this possibility, we analyzed the DN4 cells from individual mice. The pooled cells from the phenotypes of thymic and peripheral B cells in the Foxn1⌬/⌬ mice. Foxn1⌬/⌬ mutants did differentiate to the DP stage with a similar There are two subpopulations of B cell in the periphery. B-1 cells The Journal of Immunology 919 Downloaded from

FIGURE 5. B cells in adult and E17.5 embryonic thymi. Freshly iso- lated adult (a–c) or E17.5 fetal (d) thymic lymphocytes and splenocytes were analyzed for surface expression of B cell markers (CD19, B220 and CD5, IgM, IgD) as indicated. Each data point on the graphs represents one http://www.jimmunol.org/ p Ͻ 0.001. a, Expression of CD19 ,ء .thymus or spleen. Bar is the mean and B220 on thymocyte and splenocytes. b, Cell size on forward scatter (FSC) linear scale, and expression of CD5, IgM, and IgD on gated thymic FIGURE 6. Expression of Notch ligands in fetal and adult TECs and CD19med, B220high (R3, thin line) and CD19high, B220Ϫ/low (R2, bold line) cleaved Notch1ϩ cells in DN subsets. a, mRNA expression of Notch li- subpopulations only from Foxn1⌬/⌬ mice. c, Ratio of B-1 cells to B-2 cells gands Delta-like 1, Delta-like 4, Jagged-1, and Jagged-4 in collagenase in thymus and periphery. d, Analysis of total B cell number in E17.5 fetal treated TEC-enriched adult thymi. b, Expression of Notch ligands Delta- thymus. like 1 and Jagged-1 in dGUO-treated E15.5 fetal thymi by RT-PCR.

are a minor population of B cells that are reported to originate and mus, but cell numbers were decreased by ϳ50% (Fig. 5d). This is by guest on September 30, 2021 exist in peripheral organs like peritoneum and gut, and are very similar to the results for B-1 cells in the adult thymus, and CD5ϩIgMϩIgDϪCD19highB220low. B-2 cells are the conventional suggest that this phenotype remains constant throughout the fetal CD5ϪIgMlowIgDhighB220high B cells, which develop from BM and adult stages. and are the major B cell population in BM, spleen, and blood (22). As the only DN1 populations reported to have NK lineage po- We characterized the profile of thymic B cells by expression of tential are the DN1a and b subsets (10), absence of these cells ϩ CD19 and B220. CD19ϩ cells could be separated into B220high might result in a decrease in thymic NK cells. NK1.1 cells were and B220Ϫ/low subpopulations both in control and Foxn1⌬/⌬ thy- present in variable but similar numbers in the control and mutant mus (Fig. 5a). The B220high cells had lower expression of CD19 thymi, although the relative percentage was increased in the mutants than B220Ϫ/low cells, and were CD5ϪIgMlowIgDhigh (Fig. 5b), due to the general hypocellularity (data not shown). Given that both B ⌬ ⌬ consistent with a B-2 cell phenotype. Because the profile of thymic and T cells were reduced in the Foxn1 / adult thymus, this result is B220highCD19med B cells in both control and Foxn1⌬/⌬ thymus consistent with normal generation of thymic NK cells, possibly from were similar to the major B cell subpopulation in the periphery the DN1a/b lineage (10).

(Fig. 5a), these cells could represent recirculated B cells. The ⌬/⌬ B220Ϫ/lowCD19high cells were larger and CD5ϩ, and IgMϩ but not Reduced Notch ligand expression in the Foxn1 mutant IgDϩ(Fig. 5b), similar to B-1 cells. These cells were dispropor- thymus tionately represented within thymic B cells compared with the pe- Notch signals have been strongly implicated in the choice between riphery in Foxn1⌬/⌬ mice (Fig. 5c). B or T lineage commitment in the thymus (23–26). Reduced Notch The relative increase in B-1 cells raised the possibility that thy- signaling has been suggested to favor B-lineage commitment early mic B-1 cells represent those produced within the thymus itself. in differentiation (27), and the presence of DL-1 on OP9 stromal Although the total number of B cells was reduced ϳ5-fold, B-1 cells is required to support T cell development in coculture (28– cells were only reduced 2-fold (Table I), and the ratio of B-1 to B-2 31). There are four Notch ligands expressed on mouse TECs: Del- cells was dramatically increased in Foxn1⌬/⌬ thymus compared ta-like 1, Jagged-1, Delta-like 4, and Jagged-2 (14, 15, 32). We with control thymus and both mutant and control peripheral B cells assayed the expression of these four Notch ligands in fetal and (Fig. 5c). This result suggested that the Foxn1⌬/⌬ thymic environ- adult Foxn1⌬/⌬ mutant thymic epithelial cells by semiquantitative ment preferentially supports B cell development, specifically in the RT-PCR. Expression of all four ligands was strongly decreased in B-1 pathway. adult Foxn1⌬/⌬ thymus RNA enriched for stromal cells, with DL-4 As we have previously reported that fetal thymocyte differenti- being most markedly reduced (Fig. 6a). There were no differences ation is less severely affected than adult stages in Foxn1⌬/⌬ mu- in any of the ligands in the fetal thymus compared with controls tants, we analyzed thymic B cell development at fetal stages. The (Fig. 6b, and data not shown), again consistent with the milder percentage of B220ϩ cells was increased in the fetal mutant thy- thymocyte phenotype at this stage (Fig. 5d). Thus, Notch ligands 920 T CELL DEVELOPMENT FROM kit-NEGATIVE PROGENITORS are either down-regulated or fail to be up-regulated in the postnatal c-kit expression on thymocytes could result in much lower effec- Foxn1⌬/⌬ thymus, and may contribute to the observed phenotypes. tive availability of kitl to DN1a cells, resulting in loss of or dif- ferentiation defects in these cells. Discussion The conclusion that the T cells generated in the Foxn1⌬/⌬ adult Our initial analysis of the Foxn1⌬/⌬ phenotype showed that thy- thymus are derived from atypical or noncanonical progenitors mocytes in the Foxn1⌬/⌬ thymus did not contain DN2 and DN3 raises the question of whether the T cells produced in these mice cells, although DP and SP cells were produced (8). As stated contribute to the peripheral T cell pool, and have normal function. above, in the current study we postulated three possibilities that These atypical DN1 populations do not proliferate or differentiate could account for this phenotype in the adult mutant thymus: in the context of a normal thymic microenvironment (10), and DN2/3 cells could be present, but in too low numbers to reliably therefore it is unknown what type of T cells, if any, these progen- detect; thymocytes could be transiting the DN2/3 stages with un- itor populations can generate in a normal thymus. The Foxn1⌬/⌬ usually rapid kinetics; or thymocytes could progress directly from mice, therefore, provide an in vivo situation in which these pop- the DN1 to DN4 stages. The results showing that Rag mutant DN1 ulations can differentiate, albeit inefficiently, to produce mature T cells in the Foxn1⌬/⌬ thymus accumulate at the DN1 stage (Fig. 1) cells which are exported to the periphery. Other data from our are inconsistent with the first two possibilities, and further show laboratory has shown that the peripheral T cells in Foxn1⌬/⌬ mice that progression to the DN4 stage in these mutants requires gene have an atypical phenotype (13). Taken together, these data raise rearrangement. Furthermore, we show that the DN4 cells in these the possibility that under some conditions, these cells may give rise mice can differentiate to the DP stage (Fig. 4a), that total DN cells to a minor subpopulation of T cells which may have different func- that do not appear to contain any DN2 or DN3 cells have rear- tional characteristics from those arising from canonical precursors, Downloaded from ranged TCR (Fig. 4b), that thymocytes are actively undergoing even in a wild-type mouse. TCR rearrangement in the postnatal thymus (Fig. 4c), and that As DN1c and d cells were previously shown to have significant ϩ ⌬ ⌬ ⌬ sorted DN1 cells from both Foxn1 / control and Foxn1 / mu- B cell potential, it is perhaps not surprising that B cell development tant mice rapidly produce DP thymocytes in day 7 FTOCs (Fig. is much less affected than T cell development in these mice. Al- 3b). These data are all consistent with our conclusion that thymo- though total B cell numbers are reduced, the number of B-1 cells ⌬/⌬ cytes and peripheral T cells in the Foxn1 mutants are generated is least affected, and the ratio of B-1 to B-2 cells is increased http://www.jimmunol.org/ Ϫ via differentiation of atypical CD117 progenitors. Although it is 5-fold. Our data suggest that these B-1 cells are mostly produced possible that some thymocytes that arose earlier from fetal precur- inside the thymus, and not recirculated from the periphery. Inter- sors persist in adult mutants, especially in the early postnatal pe- estingly, the ratio of B-1 to B-2 cells was also increased in the riod, the evidence presented here supports the conclusion that new spleen, suggesting that the thymus may be a significant source of SP cells originating from the atypical pathway are exported to the B-1 cells in the periphery as well. In thymocytes, Notch signaling periphery from the postnatal thymus. It is possible that a very small has been linked to both cell lineage commitment and proliferation ϩ ⌬ ⌬ number of c-kit DN1a cells are present in the Foxn1 / thymus, (32, 33). Notch1 is present in murine thymocytes mostly at the and differentiate into T cells along an atypical path (i.e., not via DN2 and DN3 stages, with the highest levels of expression at DN2 DN1b, DN2, and DN3) due to the influence of an abnormal mi- stage. Thus, even if a low number of DN2 cells is produced, de- by guest on September 30, 2021 croenvironment. However, because of the similarity between the creased Notch ligand availability in the Foxn1⌬/⌬ adult thymus ⌬ ⌬ phenotype of developing thymocytes in the Foxn1 / thymus in could contribute to the loss of these cells. In addition, this altered vivo and in FTOC and the published behavior of DN1d/e subsets microenvironment is likely to selectively promote thymic B cell in vitro on OP9-DL1 cells, our data suggest that T cells in the development. Reduced availability of Notch ligands might also be ⌬ ⌬ Foxn1 / adult thymus arise from the DN1d or DN1e subsets. a contributing factor in the lower proliferation of Foxn1⌬/⌬ T lin- The extremely low number of DN1a cells even in wild-type eage thymocytes. Taken together, all of the above could contribute thymus makes it difficult to confirm whether they are physically to the thymic B and T cell phenotypes in Foxn1⌬/⌬ adult thymus. ⌬ ⌬ present, especially in the very hypocellular Foxn1 / thymus. One The differences in thymocyte development profiles between fe- possibility is that the DN1a progenitors never enter the postnatal tal and adult Foxn1⌬/⌬ thymus remain an intriguing aspect of this thymus, although clearly the DN1c, d and e cells can. Previously phenotype. However, as the DN1a–e subsets defined in adult published data (8) and our more recent analysis of SP cells in stages are not clearly present in the fetal thymus, the difference in ⌬ ⌬ Foxn1 / mice (13) have also shown that peripheral T cells can fetal and adult T cell phenotypes may not be as simple as presence easily recirculate into the postnatal thymus, so there is no physical or absence of DN1a/b cells. One clear difference is in Notch ligand barrier to their entry. However, we cannot exclude the possibility availability in the fetal vs adult mutant thymus. This difference in ⌬ ⌬ that DN1a progenitor cells never enter the postnatal Foxn1 / thy- Notch ligand levels may be compounded by a difference in the mus. If they do enter the thymus, what is the fate of these DN1a response of fetal vs adult progenitor cells to the Foxn1⌬/⌬ micro- cells? The normal number of NK cells raises the possibility that environment. Although the ratio of B to T cells changes from the fetal these cells may still be present and able to differentiate along this to adult stages, the relative numbers of B cells produced in the thymus ⌬/⌬ lineage in the Foxn1 thymus. Alternatively, they could fail to at fetal and adult stages are affected to a similar degree (both are proliferate, survive, migrate, or differentiate properly, due to re- reduced by ϳ50%). Alternatively, it is entirely possible that changes duced availability of kitl. The low levels of c-kit on the putative in Notch ligand availability play a role in the phenotypic differences ⌬/⌬ DN1a subset could represent an inability of the Foxn1 micro- in T cell differentiation at fetal and adult stages. environment to up-regulate or maintain high levels of c-kit on these cells, or reflect an absence of true DN1a cells and presence or expansion of a novel class of progenitor cells in these mutants. Acknowledgments Signaling through the Kit receptor is dosage sensitive, and kitl ⌬/⌬ We thank Drs. Howard Petrie (Scripps/Florida Research Institute, Jupiter, expression in Foxn1 adult thymic stromal cells is reduced by FL) and Juan-Carlos Zuniga Pflucker (Sunnybrook and Women’s College 50% (our unpublished data). This reduction, and/or changes in the Health Sciences Centre, Toronto, Ontario, Canada) for helpful discussions. location and/or presentation of kit signals in the highly disorga- We also thank Dr. E. V. Rothenberg (California Institute of Technology, ⌬ ⌬ nized and abnormal Foxn1 / thymus combined with low levels of Pasadena, CA) for RagϪ/Ϫ LY5.1 mice, and Julie Nelson in the Center for The Journal of Immunology 921

Tropical and Emerging Global Diseases Flow Cytometry Facility (Univer- 17. Oosterwegel, M. A., M. C. Haks, U. Jeffry, R. Murray, and A. M. Kruisbeek. sity of Georgia, Atlanta, GA) for help in cell sorting. 1997. Induction of TCR gene rearrangements in uncommitted stem cells by a subset of IL-7 producing, MHC class-II-expressing thymic stromal cells. Immu- nity 6: 351–360. Disclosures 18. Levin, S. D., S. J. Anderson, K. A. Forbush, and R. M. Perlmutter. 1993. A The authors have no financial conflict of interest. dominant-negative transgene defines a role for p56lck in thymopoiesis. EMBO J. 12: 1671–1680. References 19. Anderson, S. J., K. M. Abraham, T. Nakayama, A. Singer, and R. M. Perlmutter. 1992. Inhibition of T-cell receptor ␤-chain gene rearrangement by overexpression 1. Miller, J. F. 1961. Immunological function of the thymus. Lancet 2: 748–749. lck 2. Anderson, G., and E. J. Jenkinson. 2001. Lymphostromal interactions in thymic of the non-receptor protein tyrosine kinase p56 . EMBO J. 11: 4877–4886. development and function. Nat. Rev. Immunol. 1: 31–40. 20. Wilson, A., H. T. Petrie, R. Scollay, and K. Shortman. 1989. The acquisition of 3. Ritter, M. A., and R. L. Boyd. 1993. Development in the thymus: it takes two to CD4 and CD8 during the differentiation of early thymocytes in short-term culture. tango. Immunol. Today 14: 462–469. Int. Immunol. 1: 605–612. 4. van Ewijk, W., E. W. Shores, and A. Singer. 1994. Crosstalk in the mouse thy- 21. Chen, H. T., A. Bhandoola, M. J. Difilippantonio, J. Zhu, M. J. Brown, X. Tai, mus. Immunol. Today 15: 214–217. E. P. Rogakou, T. M. Brotz, W. M. Bonner, T. Ried, and A. Nussenzweig. 2000. 5. Kisielow, P., and H. von Boehmer. 1995. Development and selection of T cells: Response to RAG-mediated VDJ cleavage by NBS1 and ␥-H2AX. Science 290: facts and puzzles. Adv. Immunol. 58: 87–209. 1962–1965. 6. Blackburn, C. C., and N. R. Manley. 2004. Developing a new paradigm for 22. Sato, S., N. Ono, D. A. Steeber, D. S. Pisetsky, and T. F. Tedder. 1996. CD19 thymus organogenesis. Nat. Rev. Immunol. 4: 278–289. regulates B lymphocyte signaling thresholds critical for the development of B-1 7. Nehls, M., B. Kyewski, M. Messerle, R. Waldschutz, K. Schuddekopf, lineage cells and autoimmunity. J. Immunol. 157: 4371–4378. A. J. Smith, and T. Boehm. 1996. Two genetically separable steps in the differ- 23. Osborne, B., and L. Miele. 1999. Notch and the immune system. Immunity 11: entiation of thymic epithelium. Science 272: 886–889. 653–663. 8. Su, D. M., S. Navarre, W. J. Oh, B. G. Condie, and N. R. Manley. 2003. A 24. Pear, W. S., and F. Radtke. 2003. Notch signaling in lymphopoiesis. Semin. domain of Foxn1 required for crosstalk-dependent thymic epithelial cell differ- Immunol. 15: 69–79. entiation. Nat. Immunol. 4: 1128–1135. Downloaded from 9. Klug, D. B., C. Carter, E. Crouch, D. Roop, C. J. Conti, and E. R. Richie. 1998. 25. Maillard, I., S. H. Adler, and W. S. Pear. 2003. Notch and the immune system. Interdependence of cortical thymic epithelial cell differentiation and T-lineage Immunity 19: 781–791. commitment. Proc. Natl. Acad. Sci. USA 95: 11822–11827. 26. Radtke, F., A. Wilson, and H. R. MacDonald. 2004. Notch signaling in T- and 10. Porritt, H. E., L. L. Rumfelt, S. Tabrizifard, T. M. Schmitt, J. C. Zuniga-Pflucker, B-cell development. Curr. Opin. Immunol. 16: 174–179. and H. T. Petrie. 2004. Heterogeneity among DN1 prothymocytes reveals mul- 27. Pui, J. C., D. Allman, L. Xu, S. DeRocco, F. G. Karnell, S. Bakkour, J. Y. Lee, tiple progenitors with different capacities to generate T cell and non-T cell lin- T. Kadesch, R. R. Hardy, J. C. Aster, and W. S. Pear. 1999. Notch1 expression eages. Immunity 20: 735–745. in early lymphopoiesis influences B versus T lineage determination. Immunity 11: 11. Shortman, K., and L. Wu. 1996. Early T lymphocyte progenitors. Annu. Rev. 299–308. http://www.jimmunol.org/ Immunol. 14: 29–47. 28. Schmitt, T. M., and J. C. Zuniga-Pflucker. 2002. Induction of T cell development 12. Allman, D., A. Sambandam, S. Kim, J. P. Miller, A. Pagan, D. Well, A. Meraz, from hematopoietic progenitor cells by ␦-like-1 in vitro. Immunity 17: 749–756. and A. Bhandoola. 2003. Thymopoiesis independent of common lymphoid pro- 29. Schmitt, T. M., M. Ciofani, H. T. Petrie, and J. C. Zuniga-Pflucker. 2004. Main- genitors. Nat. Immunol. 4: 168–174. tenance of T cell specification and differentiation requires recurrent notch recep- 13. Xiao, S., D. M. Su, and N .R. Marley. 2007. Atypical memory phenotype T cells tor-ligand interactions. J. Exp. Med. 200: 469–479. with low homeostatic potential and impaired TCR signaling and regulatory T cell ⌬ ⌬ 30. Schmitt, T. M., R. F. de Pooter, M. A. Gronski, S. K. Cho, P. S. Ohashi, and function in Foxn1 / mutant mice. J. Immunol. 179: 8153–8163. J. C. Zuniga-Pflucker. 2004. Induction of T cell development and establishment 14. Anderson, G., J. Pongracz, S. Parnell, and E. J. Jenkinson. 2001. Notch ligand- of T cell competence from embryonic stem cells differentiated in vitro. Nat. bearing thymic epithelial cells initiate and sustain Notch signaling in thymocytes Immunol. 5: 410–417. independently of T cell receptor signaling. Eur. J. Immunol. 31: 3349–3354. 15. Harman, B. C., E. J. Jenkinson, and G. Anderson. 2003. Entry into the thymic 31. Zuniga-Pflucker, J. C. 2004. T-cell development made simple. Nat. Rev. Immu- nol. 4: 67–72.

microenvironment triggers Notch activation in the earliest migrant T cell pro- by guest on September 30, 2021 genitors. J. Immunol. 170: 1299–1303. 32. Radtke, F., A. Wilson, S. J. Mancini, and H. R. MacDonald. 2004. Notch regu- 16. Trop, S., M. Rhodes, D. L. Wiest, P. Hugo, and J. C. Zuniga-Pflucker. 2000. lation of lymphocyte development and function. Nat. Immunol. 5: 247–253. Competitive displacement of pT␣ by TCR-␣ during TCR assembly prevents sur- 33. Lehar, S. M., and M. J. Bevan. 2002. T cell development in culture. Immunity 17: face coexpression of pre-TCR and ␣␤ TCR. J. Immunol. 165: 5566–5572. 689–692.