Commitment Expression, and T Cell Lineage Α Receptor Leads To

Growth Factor Independence-1B Expression Leads to Defects in T Cell Activation, IL-7 Receptor α Expression, and T Cell Lineage Commitment This information is current as of September 26, 2021. Loretta L. Doan, Mary Kate Kitay, Qing Yu, Alfred Singer, Sabine Herblot, Trang Hoang, Susan E. Bear, Herbert C. Morse III, Philip N. Tsichlis and H. Leighton Grimes J Immunol 2003; 170:2356-2366; ; doi: 10.4049/jimmunol.170.5.2356 Downloaded from http://www.jimmunol.org/content/170/5/2356

References This article cites 41 articles, 16 of which you can access for free at: http://www.jimmunol.org/ http://www.jimmunol.org/content/170/5/2356.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 26, 2021 • Fast Publication! 4 weeks from acceptance to publication

*average

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 © 2003 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Growth Factor Independence-1B Expression Leads to Defects in T Cell Activation, IL-7 Receptor ␣ Expression, and T Cell Lineage Commitment1

Loretta L. Doan,2*† Mary Kate Kitay,2* Qing Yu,‡ Alfred Singer,‡ Sabine Herblot,¶ Trang Hoang,¶ Susan E. Bear,ʈ Herbert C. Morse III,§ Philip N. Tsichlis,# and H. Leighton Grimes3*†

T cell differentiation in the thymus is dependent upon signaling through the TCR and is characterized by the resulting changes in expression patterns of CD4 and CD8 surface coreceptor molecules. Although recent studies have characterized the effects of proximal TCR signaling on T cell differentiation, the downstream integration of these signals remains largely unknown. The growth factor independence-1 (GFI1) and GFI1B transcriptional repressors may regulate cytokine signaling pathways to affect Downloaded from lymphocyte growth and survival. In this study, we show that Gﬁ1 expression is induced upon induction of the T cell program. Gﬁ1B expression is low and dynamic during T cell development, but is terminated in mature thymocytes. Transgenic expression of GFI1 and GFI1B in T cells allowed us to determine the functional consequences of constitutive expression. GFI1 potentiates response to TCR stimulation and IL-2, whereas GFI1B-transgenic T cells are defective in T cell activation. Moreover, GFI1B- transgenic thymocytes display reduced expression of the late-activation marker IL-7R␣, and a decrease in CD4؊8؉ single-positive http://www.jimmunol.org/ T cells that can be mitigated by transgenic expression of BCL2 or GFI1. These data show that GFI1 and GFI1B are functionally unique, and implicate a role for GFI1 in the integration of activation and survival signals. The Journal of Immunology, 2003, 170: 2356–2366.

cell differentiation in the thymus is dependent upon sig- positive (CD4 SP) or CD4Ϫ8ϩ (CD8 SP) T cells is triggered by the naling through the TCR and is characterized by the re- engagement of TCRs on the immature TCRhighCD4ϩCD8ϩ thy- T sulting changes in expression patterns of CD4 and CD8 mocytes by self-peptide/MHC complex on thymic epithelial cells surface coreceptor molecules. Early thymocyte precursors are (1, 2). Negative selection eliminates immature DP thymocytes Ϫ Ϫ 4 CD4 CD8 (double negative; DN) and are signaled to differen- through clonal deletion of those T cells that have high affinity for by guest on September 26, 2021 ϩ ϩ tiate first into CD4 CD8 (double positive; DP) thymocytes. Fur- self-peptide and thus are potentially autoreactive (3). Positive se- ϩ Ϫ ther differentiation of DP thymocytes into mature CD4 8 single- lection occurs when low-affinity TCR-ligand interactions trigger a signal for survival and results in termination of one or the other CD4 or CD8 coreceptor molecule. The choice of which coreceptor *Institute for Cellular Therapeutics and Department of Surgery, University of Lou- to extinguish is referred to as lineage commitment. isville School of Medicine, Louisville, KY 40202; †Department of Biochemistry and Molecular Biology, ‡Experimental Immunology Branch, National Cancer Institute, The growth factor independence-1 (GFI1) and GFI1B proteins and ¤Laboratory of Immunopathology, National Institute of Allergy and Infectious are closely related nuclear oncoproteins that may regulate cytokine Diseases, National Institutes of Health, Bethesda, MD 20892; ¶Clinical Research Institute of Montreal, Quebec, Canada; ʈKimmel Cancer Center, Thomas Jefferson pathways. Gfi1 was originally identified as the gene up-regulated University, Philadelphia, PA 19107; and # Molecular Oncology Research Institute, by insertion of Moloney murine leukemia virus in a thymic lym- Tufts-New England Medical Center, Boston, MA 02111 phoma that was selected for its ability to grow in the absence of the Received for publication April 30, 2002. Accepted for publication December T cell cytokine IL-2 (4). Forced expression of GFI1 in the IL-2- 17, 2002. dependent parental cell line potentiates the outgrowth of IL-2-in- The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance dependent cell lines, without inducing IL-2 (4, 5). Gfi1B was iden- with 18 U.S.C. Section 1734 solely to indicate this fact. tified by low stringency hybridization screening with a cDNA 1 This work was supported by Hope Street Kids, the Jennifer Sacco Memorial Fund, probe encoding the zinc-finger region of Gfi1 (6). GFI1 and GFI1B the University of Louisville School of Medicine Grant-in-Aid, a University of Lou- isville Research Initiation Grant, and in part by the Commonwealth of Kentucky are 97% homologous in the carboxyl-terminal 165 aa that code for Research Challenge Trust Fund and the Jewish Hospital Foundation. L.L.D. is sup- six Cys-His zinc fingers. An amino-terminal 20-aa snail and Gfi1 ported by a National Science Foundation Graduate Research Fellowship. This work was supported in part by PHS CA56110 (to P.N.T.). (SNAG) domain, responsible for nuclear localization and transcriptional repressor function, is also highly conserved (5). In con- 2 L.L.D. and M.K.K. contributed equally to this work. trast, the 236 intervening amino acids between the GFI1 SNAG 3 Address correspondence and reprint requests to Dr. H. Leighton Grimes, Institute for Cellular Therapeutics, University of Louisville, Baxter Biomedical Research and zinc-finger domains bear no homology to the corresponding Building, Suite 404-F, 570 South Preston Street, Louisville, KY 40202-1760. E-mail 145 aa of GFI1B. Both proteins bind to virtually identical DNA address: [email protected] consensus sequences and function as transcriptional repressors in a 4 Abbreviations used in this paper: DN, CD4ϪCD8Ϫ (double negative); DP, CD4ϩCD8ϩ (double positive); CD4 SP, CD4ϩ8Ϫ single positive; CD8 SP, CD4Ϫ8ϩ SNAG-dependent manner (5, 6). GFI1 is mildly antiapoptotic and single positive; GFI1, growth factor independence-1; MFI, mean fluorescence inten- inhibits growth arrest of IL-2-dependent T cell lines under condi- sity; ISP, intermediate single positive; PϩI, phorbal-12-myristate-13-acetate and ionomycin; SNAG, snail and Gfi1 repressor domain; GH, human growth hormone; tions of limiting IL-2 (5, 7), while GFI1B inhibits both IL-6-in- RAG, recombination-activating gene; IRF, IFN regulatory factor; WT, wild type. duced differentiation and growth arrest of M1 myelomonocytic

Copyright © 2003 by The American Association of Immunologists, Inc. 0022-1767/03/$02.00 The Journal of Immunology 2357 cells (6). Mice deleted for GFI1 have altered inflammatory re- perfusion with 10 ml of medium 199. Both thymocyte and splenocyte cell sponses and differentiation in the myeloid lineage (8), while mouse debris was depleted by passage through Nytex nylon-mesh screens. embryos deleted for GFI1B die in utero due to a lack of definitive Splenocytes were treated with ammonium chloride-potassium bicarbonate solution (150 mM NH4Cl and 10 mM KHCO3) to lyse RBCs. For exper- erythropoiesis (9). iments requiring isolation of T cells, splenic cell preps were depleted of Gfi1 and Gfi1B are differentially expressed in lymphoid com- other cell types by the use of T cell enrichment columns (R&D Systems, partments. Northern analysis reveals that Gfi1 is expressed in the Minneapolis, MN). All cells were counted with a Coulter Counter model bone marrow and thymus, with low-level expression in the spleen, Z2 (Coulter, Miami, FL), and viability was assayed by trypan blue exclusion. whereas Gfi1B is expressed in the bone marrow and spleen, with low-level expression in the thymus (6). Both Gfi1 and Gfi1B show Flow cytometric analysis regulated expression during T cell development, but Gfi1B expression is terminated in mature thymocytes. Gfi1 message is not ex- Cell surface staining was performed by incubating 1 ϫ 106 cells with pressed in G splenic T cells, but is induced upon T cell activation mAbs at varying concentrations in FACS medium (HBSS with 0.1% BSA, 0 0.1% sodium azide, and 0.036% sodium bicarbonate) for 20 min on ice. (4, 10). Transgenic expression of GFI1 and GFI1B in T cells al- Stained cells were washed twice with FACS medium and fixed in 1% lowed us to determine the functional basis for differential expres- formaldehyde (Polysciences, Warrington, PA). For intranuclear staining, sion of these factors. Transgenic expression of GFI1 potentiates T cells were fixed in 2% formaldehyde in PBS and permeabilized and stained cell activation. In contrast, ectopic expression of GFI1B in T cells in PBS plus 5% FBS and 0.5% Triton X-100. Flow cytometry was performed on a FACSCalibur, FACSVantageSE, or FACStar flow cytometer results in defective T cell activation, lower numbers of peripheral using standard CellQuest acquisition. Data were analyzed using CellQuest T cells, a reduction in IL-7R␣ expression, and a developmental (BD Biosciences, Mountain View, CA) and FlowJo (Tree Star, San Carlos, block to CD8 SP T cells. The block to CD8 SP development is CA) software. The absolute cell numbers of gated cells per thymus or Downloaded from mitigated by forced expression of BCL2 or GFI1. These data in- spleen were calculated by multiplying the percentage of each population dicate that GFI1 and GFI1B are not redundant for T cell activation with the total number of cells per thymus or spleen respectively. function, and implicate integration of activation and survival sig- Northern, Western, and RT-PCR analyses naling in CD8 lineage commitment. RT-PCR analyses were performed as previously published (13, 14). The sequences of the Gfi1-specific primers were 5Ј-CACACCTTCATCCACA Materials and Methods http://www.jimmunol.org/ Ј Ј Ј Mice CAGG-3 and 5 -GATGAGCTTTGCACACTGGA-3 , and the probe was 5Ј-TACCGTGAGGATGTCTTCCC-3Ј. The sequences of the Gfi1B-specific The generation of the CD2-GFI1-transgenic mice has been described (10). primers were 5Ј-AGCACAGAGTCTCCCTTGGA-3Ј and 5Ј-CAAAGGTTT The lck-Gfi1-transgenic mouse (line 3A) was generated by cloning the rat TGCCACAGACA-3Ј, and the probe was 5Ј-ACCCCTCATGGGCTAGAA Gfi1 cDNA into the BamHI site of the TLC vector (11). This vector con- GT-3Ј. The Gfi1B pattern was confirmed with the primers 5Ј-GAGCAG tains a 3.2-kb fragment of the mouse lck proximal promoter and a 2.2-kb CATACTCACGTCCA-3Ј and 5Ј-TTCATGTCCGACTTCTGGTG-3Ј, and fragment of the human growth hormone (GH) gene, which provides exons the probe was 5Ј-CAAAGCCTTCAAGCGTTCAT-3Ј. and introns for splicing and polyadenylation sequences. A 2.2-kb fragment Western blotting with Abs against GFI1 (sc-6357), GFI1B (sc-8559; of the 3Ј locus control region of the human CD2 gene is located down- Santa Cruz Biotechnology), GFI1 and GFI1B (sc-6357), p27 (554069; BD stream of the GH to obtain copy-number- and insertion-site-independent PharMingen), and IFN regulatory factor (IRF)1 Ab (sc-640) was performed levels of expression. The GFI1B-transgenic mice (lines 5B and 5C) were as follows. Single-cell suspensions of primary thymocytes were lysed at a by guest on September 26, 2021 constructed in an identical manner to the GFI1 transgenic except the con- concentration of 10Ð20 ϫ 106 cells/100 ␮l SDS lysis buffer (25 mM Tris struct contained the cDNA for mouse Gfi1B inserted into the BamHI site. (pH 7.5), 10% glycerol, and 2% SDS) supplemented with protease inhib- The GFI1 or the GFI1B transgene (1Ð5 ng/␮l) was microinjected into itors complete (Roche, Basel, Switzerland) and 2 mM PMSF. Protein con- C57BL/6J (The Jackson Laboratory, Bar Harbor, ME) eggs according to centration was determined using the BCA protein assay reagent (Pierce, standard methodology in the Laboratory of Immunopathology (National Rockford IL), and 75 ␮g of cellular extract was run on a 10% SDS-poly- Institute of Allergy and Infectious Diseases, National Institutes of Health). acrylamide gel, transferred to Immobilon-polyvinylidene fluoride (Milli- Recombination-activating gene (RAG)2Ϫ/Ϫ/HY mice (Taconic Farms, pore, Bedford, MA), and blocked for1hatroom temperature in blocking Germantown, NY), and E␮-BCL2-25-transgenic mice (The Jackson Lab- buffer (5% milk, 20 mM Tris (pH 7.3), 6.85 mM NaCl, 0.1% (v/v) Tween oratory) (12) were purchased from commercial vendors. All mice were on 20, and 0.5 g/L MgCl2). Membranes were incubated overnight at 4¡Cin a C57BL/6 J background, bred in the Baxter Barrier animal care facility at primary Ab diluted in 5% protease-free BSA (Fisher Biotech, Pittsburgh, the University of Louisville School of Medicine, and housed under specific PA), and then HRP-conjugated secondary Ab (Amersham, Piscataway, NJ) pathogen-free conditions. GFI1- and GFI1B-transgenic daughters who for1hatroom temperature. Blots were developed using ECL reagents were heterozygous for the RAG mutation, as well as the HY TCR trans- (Amersham). For Western analysis of sorted thymocytes and purified T genes were then backcrossed with their RAG2Ϫ/Ϫ,HYTCRϩ fathers. Col- cells, 1 ϫ 106 cells were resuspended in 15 ␮l of lysis buffer (50 mM onies were expanded by intercrossing of littermates. BCL2/GFI1B and HEPES (pH 7.8), 450 mM NaCl, 0.2 mM EDTA, 25% glycerol, 1% Non- CD2-GFI1/GFI1B bitransgenics were generated in a similar manner. All idet P-40 (15), protease inhibitors complete (Roche), and 2 mM PMSF), animal work performed was reviewed and approved by the University of then sonicated using a Misonix (Farmingdale, NY) sonic dismembrator Louisville Institutional Animal Care and Use Committee. with microprobe tip. Loading buffer (4ϫ) was added, and the lysates were boiled. The entire contents of the lysate were loaded onto a denaturing Antibodies SDS-polyacrylamide gel, and Western blotting was performed as described above. Abs with the following specificities were used for cell stimulations: CD3⑀ (145.2C11) and CD28 (37.51). Abs with the following specificities were used for staining of thymocytes and splenocytes: CD4 (RM4-5 and Cell stimulation and proliferation ␣ ␣ GK1.5), IL-7R (A7R34; eBiosciences, San Diego, CA), CD8 (53-6.7), Single-cell suspensions of spleen cells in RPMI 1640 (Life Technologies) CD3⑀ (145-2C11), TCR␣␤ (H57-597), TCR V␤8 (F32), CD24␣ (M1/69), supplemented with 5% FBS, L-glutamine, penicillin, streptomycin, genta- CD69 (H1.2F3), and CD25 (PC61). Abs were purchased from BD Phar- micin, 2% HEPES (all from Life Technologies) and 0.1% 2-ME (Sigma- Mingen (San Diego, CA) unless otherwise noted. Intranuclear staining was Aldrich, St. Louis, MO) were plated in 96-well round-bottom plates (Corn- performed using anti-GIF1B goat polyclonal IgG (sc-8559), anti-GFI1 goat ing, Corning, NY) at a density of 1 ϫ 105/well in 100 ␮l. Stimuli were polyclonal IgG (sc-8558), normal goat IgG (sc-2028) control, and second- added as indicated at a range of concentrations to assess dose dependency. ary bovine anti-goat IgG-FITC (sc-2348), all from Santa Cruz Biotechnol- The stimuli were low-endotoxin, no-azide anti-CD3 (145.2C11) and anti- ogy (Santa Cruz, CA). CD28 (37.51) (both Abs from BD PharMingen), and recombinant human Preparation of cell suspensions IL-2 (Chiron, Emeryville, CA). Cells were cultured for 48 or 72 h, then pulsed with [3H]thymidine (1 ␮Ci; Amersham), and harvested 18 h later Thymuses from 4- to 6-wk-old mice were removed and disrupted between using a TOMTEC-Harvester 96/Model Mach II (Wallac, Akron, OH). Pro- frosted ends of glass slides and washed twice with medium 199 (Life liferation was determined by measuring radioactivity (Wallac 1205-SP2 Technologies, Gaithersburg, MD). Cells were obtained from spleens by betaplate counter). 2358 INTEGRATION OF THYMOCYTE ACTIVATION AND LINEAGE COMMITMENT

Coreceptor reversal Purified DP thymocytes (Ͼ96%) were obtained by panning with IgM anti- CD8 (83-12-5)-coated plates. DP thymocytes (5 ϫ 106/ml) were first placed into signaling cultures and stimulated for 12Ð18 h with a combina- tion of phorbal-12-myristate-13-acetate (0.6 ng/ml) and ionomycin (0.6 ␮g/ml) (PϩI; Calbiochem, La Jolla, CA) (16). At the conclusion of signaling culture, cells were harvested, washed, and placed into nonstimula- tory recovery cultures for an additional 12Ð16 h. Cells were stained for CD4 and CD8 expression, and CD4ϩCD8Ϫ cells were obtained by elec- tronic sorting of the stained cells. The purified CD4ϩ8Ϫ cells were further cultured in postrecovery cultures in the presence or absence of 6 ng/ml recombinant mouse IL-7 (Calbiochem) overnight, after which they were harvested and stained for CD4 and CD8 expression. All cultures were performed at 37¡Cin5%CO2 humidified air atmosphere in RPMI 1640 supplemented with 5 ϫ 10Ϫ5 M 2-ME and 10% FCS that had been depleted of endogenous steroids by pretreatment with 0.5% Norit A charcoal and 0.05% dextran for 30 min at 56¡C. Results Gfi1 and Gfi1B are regulated during T lymphopoiesis

We examined the steady-state mRNA levels of Gfi1 and Gfi1B Downloaded from during T lymphopoiesis (Fig. 1A). Thymocyte populations were sorted, RNA was extracted, and RT-PCR was performed to detect ribosomal S16 expression (13, 14). The products of the reaction were analyzed by Southern blot with a radiolabeled S16-specific oligonucleotide probe. The signal was quantified by phosphor im-

ager, and the samples were normalized to obtain equivalent S16 http://www.jimmunol.org/ signal from each template. Subsequent analysis of Gfi1 expression in the S16-normalized cDNA templates revealed low-level signal in CD4ϪCD8ϪCD44ϩCD25Ϫ cells (DN1) and 10-fold greater levels in CD4ϪCD8ϪCD44ϩCD25ϩ (DN2) thymocytes. The transition between DN1 and DN2 corresponds to T lymphocyte lineage commitment. Signal intensity from the Gfi1 RT-PCR product gradually increases to double the DN2 levels at the CD4ϩCD8ϩ TCRhigh stage, which contains cells that have been recently posi- tively selected and are about to undergo lineage commitment. Gfi1 by guest on September 26, 2021 RT-PCR product levels then decrease 10-fold in CD4 and CD8 SP FIGURE 1. Expression of endogenous GFI1 and GFI1B. A, Blotted and probed RT-PCR analyses of Gfi1 and Gfi1B in reverse-transcribed thy- thymocytes. In striking contrast, the signal levels of probed and mocyte-subset cDNA templates. RT-PCR was first performed for S16, and quantified Gfi1B RT-PCR product were low, but increased at ␤ the products of the reaction were blotted, probed, and quantified. Next, the stages corresponding to TCR -chain selection (DN3), and posi- cDNA templates were normalized to obtain equivalent signal from the high tive selection (DP TCR ). In the thymus, Gfi1 expression is blotted and probed S16 products. Finally, primer/probe pairs for Gfi1 and gradually induced upon induction of the T cell differentiation pro- Gfi1B were applied to the normalized cDNA templates to determine the gram, whereas low-level Gfi1B expression correlates with positive Gfi1 and Gfi1B RT-PCR product levels in thymocyte subsets relative to selection events. S16. B, Western blot analysis of sorted thymocyte subsets. A single-thymocyte suspension from 4-wk-old WT mice was stained and sorted as GFI1 and GFI1B are expressed in normal thymocytes indicated. Sorted cells (1 ϫ 106) of each subpopulation were lysed, and the To examine the level of GFI1 in thymocyte subsets, we performed total lysate was probed with anti-GFI1 (top panel) or anti-p27 (bottom Western analysis on one million sorted DP, CD8 SP, CD4 SP, and panel) as a loading control. In contrast to the expression pattern of Gfi1 ϩ low message, GFI1 protein levels are moderate in DP thymocytes and inter- CD4 CD8 thymocytes (Fig. 1B). The level of GFI1 does not ϩ mediate CD4 CD8low cells, but increase in mature CD4 and CD8 SP cells. differ between the bulk of DP thymocytes and those poised to C, Flow cytometric analysis of GFI1 expression in thymocyte subsets. ϩ low make a lineage commitment step (CD4 CD8 ); however, the Permeabilization of surface-stained, formaldehyde-fixed thymocytes and level of GFI1 is dramatically higher in SP thymocytes. To confirm subsequent incubation with Abs to GFI1, followed by a secondary Ab these data, we examined thymocyte expression of GFI1 by intranu- conjugated to FITC, yielded an expression profile for GFI1 similar to that clear staining and flow cytometry. Like the Western blot data, flow observed by standard Western blotting techniques. Specifically, the level of cytometric analyses revealed higher levels of GFI1 protein in SP GFI1 protein increases in SP thymocytes. The pattern of GFI1 expression thymocytes (Fig. 1C). The up-regulation of Gfi1 message in DP in DN thymocytes has been published (15). D, Flow cytometric analysis of thymocytes, with subsequent increase in GFI1 protein in SP cells, GFI1B expression in thymocyte subsets. Intranuclear staining of surface- stained, formaldehyde-fixed thymocytes revealed expression of GFI1B in suggests that GFI1 may play a role in the transition between these Ϫ Ϫ Ϫ two developmental stages. the relatively rare CD4 CD8 CD44 populations of cells. The average change in MFI is indicated below each graph, and is defined as follows: RT-PCR analyses revealed restricted expression of Gfi1B in rel- ⌬ ϭ ϫ Ϯ Ϫ MFI (MFIGFI1B/MFIIgG) 100 ( SEM). GFI1B-transgenic CD4 atively rare thymocyte subsets (Fig. 1A). Not surprisingly, Western CD8ϪCD44Ϫ thymocytes have a higher MFI in comparison to WT cells, analysis and intranuclear stains for GFI1B failed to reveal GFI1B whereas RAG2Ϫ/Ϫ thymocytes, which lack the capacity for TCR signaling expression in bulk thymocytes (data not shown). Therefore, we (42), do not express detectable levels of GFI1B. Although available re- Ϫ focused on a flow-cytometric analysis of the relatively rare CD44 agents for detection of GFI1B are not ideal, and the protein is likely ex- DN3 and DN4 thymocytes that appear to express the highest levels pressed at very low levels in normal cells, the observed shift is significant of Gfi1B message (Fig. 1A). We first examined the DN3 and DN4 and reproducible. The Journal of Immunology 2359 cells from GFI1B-transgenic mice (detailed below, Fig. 1D). A comparative 66% shift in mean fluorescence intensity (MFI) ⌬ ϭ ϫ ( MFI MFIGFI1B/MFIIgG 100) between the control IgG antisera and GFI1B-reactive antisera indicates that GFI1B protein is present (Fig. 1D). In a similar manner, analysis of the DN3 and DN4 cells from nontransgenic littermates revealed a 41% shift in MFI between control and GFI1B-specific antisera stains (Fig. 1D). The RT-PCR data (Fig. 1A) indicate that Gfi1B expression correlates with positive selection events. To explore this correlation, we examined thymocytes from RAG2Ϫ/Ϫ mice, which are arrested at the DN3 stage because they lack the pre-TCR selection signal that follows RAG-mediated rearrangement of the TCR ␤-chain. Flow-cytometric analyses reveal no difference in MFI between control and GFI1B-reactive antisera in RAG2Ϫ/Ϫ thymocytes (Fig. 1D). These data indicate that signals from the pre-TCR may be required for GFI1B expression. Transgenic expression of GFI1 and GFI1B

GFI1 and GFI1B bind the same DNA sequence and repress tran- Downloaded from scription in a manner dependent on the SNAG repressor domain (5, 6), suggesting the possibility that these factors are redundant. High-level transgenic expression of GFI1 in the thymus results in a block to T cell development at a stage corresponding to selection of cells after successful formation of the TCR ␤-chain (15). Given

the thymic phenotype of GFI1 overexpression, we constructed http://www.jimmunol.org/ transgenic mice expressing GFI1 and GFI1B in developing and mature T cells (Fig. 2A). Transgene-specific Northern analysis re- FIGURE 2. Expression of transgenic GFI1 and GFI1B. A, Schematic vealed that GFI1-transgenic founders had moderate expression representation of the GFI1- and GFI1B-transgene constructs. The rat while GFI1B-transgenic founders had higher levels of expression; cDNA of Gfi1 (4) and mouse cDNA of Gfi1B (6) were placed under the representative lines are shown (Fig. 2B). Western analysis of total control of the lck proximal promoter and 2.2-kb human CD2 enhancer, thymocytes and of column-enriched splenic T cells indicates that with human growth hormone (GH) gene (11). B, Northern blot analysis of total thymus or spleen RNA from representative GFI1- (line 3A) and the transgenic GFI1B protein is expressed in both the thymus and GFI1B- (line 5B) transgenic lines hybridized with transgene-specific the periphery (Fig. 2C). Finally, flow cytometric analysis of the probes. Total RNA from transgenic animals revealed moderate levels of by guest on September 26, 2021 GFI1B-transgenic thymocyte populations revealed transgenic transgenic Gfi1 in both thymus and spleen, while transgenic Gfi1B is ex- GFI1B expression in DN, DP, CD8 SP, and CD4 SP thymocytes pressed at higher levels. C, Western blot analysis of transgenic GFI1B in (Fig. 2D). thymic or column-purified splenic T cell whole-cell protein extracts. Cell To explore the subcellular localization of transgenic GFI1 and extracts from two control (WT) and two GFI1B-transgenic mice were an- GFI1B, the cytoplasm of permeabilized and stained thymocytes alyzed by Western blot with an Ab specific for the last 20 aa of GFI1 was removed using Nonidet P-40. Stripping the cytoplasm resulted (cross-reactive to GFI1B). Densitometric analysis revealed that transgenic in a loss of signal for the mitochondria-localized BCL2, but did not GFI1B is expressed at 5-fold higher levels than is endogenous GFI1 in alter the ⌬MFI of thymocytes stained for GFI1 or GFI1B (data not nontransgenic littermate thymocytes. Lysates probed with antiserum against IRF-1 act as a control for loading. D, Flow cytometric analysis of shown), indicating that the targets of the respective antisera are GFI1B expression in GFI1B-transgenic thymocyte subsets. Transgenic ex- nuclear. Given that these antisera are specific for GFI1 and GFI1B pression of GFI1B is observed in all thymocyte subsets, although the level in Western analysis, our flow cytometric data indicate that trans- of expression is somewhat lower in CD8 SP cells. genic GFI1 and GFI1B are nuclear in T cells. Transgenic expression of GFI1 enhances T cell response to CD3 cross-linking and IL-2 we limited the amount of CD3 Ab, and titrated IL-2. Again, we GFI1 was previously shown to confer IL-2 independence to rat T found that cells from GFI1-transgenic mice proliferated more vig- cell lymphomas (4, 7). IL-2 is a critical T cell cytokine during orously in response to stimulation than cells from nontransgenic activation, and though peripheral lymphocytes do not express de- littermates (Fig. 3B). A flow cytometric analysis of splenocytes tectable levels of Gfi1, activation signals induce Gfi1 within 30 min from control and GFI1-transgenic mice showed equivalent abso- (10). Moreover, transgenic expression of GFI1 was previously lute cell numbers of total splenocytes and T cell subsets as delin- shown to mildly increase [3H]thymidine uptake given a fixed eated by the markers CD4, CD8, TCR␤, and CD3⑀ (Fig. 3C), or amount of CD3-cross-linking Ab (10). To further examine the ef- CD62 ligand and CD44 (data not shown). Therefore, GFI1 poten- fect of GFI1 on T cell activation potential, spleen cells from 4- to tiates the response to CD3 and IL-2 stimulation. 6-wk-old mice were stimulated by titration of a CD3-cross-linking Ab. T cells from mice expressing GFI1 from either the Lck pro- GFI1B-transgenic mice display peripheral T lymphopenia and a moter-driven transgene (Fig. 2A) or a previously published CD2 profound defect in activation after CD3 cross-linking promoter-driven transgene (10) proliferated at a higher rate than T We next looked at the response of GFI1B-transgenic splenic T cells from control mice as evidenced by enhanced [3H]thymidine cells to stimulation with anti-CD3⑀ and found that they neither uptake (Fig. 3A) and by increased numbers of cells in each cellular died (data not shown) nor proliferated substantially. Spleen cells division as evidenced by CSFE staining (data not shown). To dis- from 4- to 6-wk-old mice from two GFI1B-transgenic lines (5B or sect the response of GFI1-transgenic splenic T cells to stimulation, 5C; Fig. 3D) were stimulated by titration of a CD3-cross-linking 2360 INTEGRATION OF THYMOCYTE ACTIVATION AND LINEAGE COMMITMENT Downloaded from http://www.jimmunol.org/

FIGURE 3. GFI1 potentiates, whereas GFI1B inhibits, T cell activation. A and B, Spleen cells from 4- to 6-wk-old GFI1-transgenic mice (line 3A or a CD2 promoter-driven GFI1-transgenic (10)) and nontransgenic littermate control mice were isolated and stimulated with either a dose titration of plate-bound Ab to CD3⑀ (A) or a low level of anti-CD3⑀ (0.06 ␮g/ml) and increasing amounts of human rIL-2 (B) and cultured for 48 h. [3H]Thymidine (1 ␮Ci) was added, and plates were incubated for 18 h. Proliferation was measured as cpm of [3H]thymidine incorporation. Error bars indicate SEM. C, by guest on September 26, 2021 Spleen cells isolated from 4- to 6-wk-old GFI1-transgenic mice, as well as littermate controls, were stained with Abs to CD4 and CD8, and then analyzed by flow cytometry. A representative FACS profile is shown. No difference was observed in total number of splenocytes or in splenic T cell numbers between either GFI1 transgenics of either line and littermate controls. D, Spleen cells from 4- to 6-wk-old GFI1B-transgenic mice (lines 5B and 5C) were activated as in A for either 48 h (left panel)or72h(right panel). [3H]Thymidine (1 ␮Ci) was added, and plates were incubated for 18 h. Proliferation was measured as cpm of [3H]thymidine incorporation. E, Spleen cells isolated from 4- to 6-wk-old GFI1B-transgenic line 5B, as well as littermate controls, were stained with Abs to CD8 and CD4 and analyzed by flow cytometry. A representative FACS profile is shown. Absolute spleen cell numbers from all mice examined were determined and are expressed in the table below as cell numbers ϫ106 Ϯ SEM in the table below.

Ab. T cells from GFI1B-transgenic mice proliferated at a substan- addition, GFI1B transgenics demonstrated decreased proportions tially lower rate than T cells from control mice as evidenced by of cells expressing activation markers CD69 and CD25 (IL-2R [3H]thymidine uptake, even when the activated cells were given an ␣-chain) as compared with WT cells (Fig. 4B), as well as a de- additional 24-h incubation (Fig. 3D;48vs72h). crease in fluorescence intensity of these markers on positive cells An unanticipated explanation for this observation came from the (data not shown). Therefore, GFI1B-transgenic T cells are pro- finding that spleen cells from GFI1B-transgenic mice show signif- foundly impaired in response to activation signals because of an icant reduction in the numbers of CD4 and CD8 T cells that could intrinsic signaling defect, and not because of the overall reduction respond to CD3 stimulation (Fig. 3E). Mature CD4 cells were in CD3⑀ϩ T cell numbers. These data are diametrically opposed to reduced to 32% of wild-type (WT) levels, while mature CD8 cells our findings for GFI1-transgenic mice. were reduced to 46% of WT levels (Fig. 3E). However, neither the TCR-expression level on splenic T cells (data not shown) nor the total number of splenocytes (Fig. 3E) was significantly reduced in Transgenic GFI1B expression alters thymocyte lineage GFI1B-transgenic mice. Because both the GFI1B-transgenic lines commitment and maturation gave equivalent data, we focused on the 5B line for further studies. To determine the cause of peripheral T lymphopenia, we examined The unresponsiveness of GFI1B lymphocyte populations to the thymus. Thymocytes were stained with Abs against CD4, CD8, TCR-mediated activation signals could be due to either T lym- and TCR␤ and analyzed by flow cytometry. In GFI1B-transgenic phopenia or a defect in signaling. To determine the mechanism, T mice, the number of CD4 SP cells was considerably enhanced cells were purified by negative selection, normalized to CD3⑀ϩ T while there was a severe reduction in the development of CD8 SP cell numbers, and stimulated simultaneously with both anti-CD3⑀ T cells (Fig. 5A). The CD8 SP compartment contains both mature and anti-CD28 in a coreceptor activation assay. Costimulated and immature intermediate single-positive (ISP) cells (CD8 ISP). GFI1B-transgenic cells showed a marked inability to proliferate as CD8 ISP cells, in contrast to CD8 SP cells, do not have high-level compared with cells from nontransgenic littermates (Fig. 4A). In TCR expression. Therefore, the analysis was repeated through a The Journal of Immunology 2361 Downloaded from http://www.jimmunol.org/

FIGURE 4. The GFI1B-induced inhibition of activation is cell autonomous and includes defective up-regulation of early activation markers. A, Four-wk-old GFI1B-transgenic and control splenic T lymphocytes were ϩ

purified by negative selection, normalized for numbers of CD3 T cells, by guest on September 26, 2021 and then costimulated with Abs to CD3⑀ and CD28. Proliferation was measured by [3H]thymidine incorporation, and the stimulation index was calculated. A representative of three experiments is shown. B, Splenocytes from GFI1B-transgenic (E and ᭛, black background) and littermate controls (F and ᭜, white background) were stimulated with Abs to CD3 for 24 h then stained with Abs to CD4, CD8, CD25 (E and F), and CD69 (᭛ and ᭜), and analyzed by flow cytometry. CD4ϩ or CD8ϩ events were gated, and the percentages of CD25ϩ or CD69ϩ cells are depicted. Fewer GFI1B-transgenic splenocytes exhibit induction of either CD25 or CD69. Bars ϭ mean. FIGURE 5. Transgenic expression of GFI1B in the thymus results in enhanced numbers of mature CD4ϩ thymocytes and decreased CD8ϩ populations. A and B, Thymocytes from GFI1B-transgenic and control littermates were stained with Abs to TCR␤, CD4, and CD8, and analyzed by int-high TCR gate, which would include CD4 SP and CD8 SP thy- flow cytometry. The average number of total thymocytes ϫ 106 Ϯ SEM is ϩ mocytes and their immediate precursors (17). The ratio of CD4 to expressed above the plots, with percentage of cells in each gate in A.InB, ϩ CD8 cells was increased in the TCRint-high population from a cells are gated on TCRint-high events, and the CD4 vs CD8 FACS profiles normal ratio of 7:1 to a ratio of 32:1 in the GFI1B-transgenic mice are depicted. Thymocyte subsets are expressed as absolute cell numbers ϫ (Fig. 5B). Therefore, few mature CD8 SP cells are generated in the 106 per thymus Ϯ SEM. C, Thymocytes from 4-wk-old GFI1B-transgenic GFI1B-transgenic mice. Moreover, the lower numbers of CD8 SP and control littermates were stained with Abs to CD24␣, CD4, and CD8, cells provide a potential explanation for peripheral CD8 lymphopenia. and analyzed by flow cytometry. FACS profiles depicted are gated on CD4ϩ events. Absolute thymocyte cell numbers from all mice examined We next examined the CD4 SP population. As CD4 SP thymo- were determined and are expressed as cell numbers ϫ106 Ϯ SEM. cytes mature, they down-regulate CD24/heat-stable Ag. Therefore, the most mature CD4 SP thymocytes express low levels of CD24 while less mature CD4 SP thymocytes express higher levels of this marker (17Ð19). We found that the number of immature CD4 SP on a class II MHC-null background to find that no CD4 SP cells thymocytes (CD24high) was relatively normal in GFI1B-transgenic were generated (data not shown). These data exclude the possibil- mice; however, the number of mature CD4 SP thymocytes was ity that the increased CD4 SP cells are due to redirection of MHC dramatically increased by 5-fold (Fig. 5C). These data are sup- class I-restricted CD8 SP development into the CD4 lineage. Al- ported by the surface expression of TCR␤, Qa-2, and CD62 ligand tered signaling during activation in GFI1B-transgenic thymocytes (data not shown), which also indicate that GFI1B-transgenic mice may lead to the accumulation of cells that should not normally be have greater numbers of phenotypically mature CD4 SP thymo- selected (defective negative selection) or the accumulation of cells cytes. We have also examined the effects of the GFI1B transgene unable to mature and egress. 2362 INTEGRATION OF THYMOCYTE ACTIVATION AND LINEAGE COMMITMENT

The development of CD8 SP T cells in GFI1B-transgenic mice is not rescued by expression of the HY class I-restricted transgene It is unlikely that GFI1B represses CD8 or MHC class I expression because thymocytes from GFI1B-transgenic mice reveal normal surface expression of CD8␣ and CD8␤ in the DP fraction, and are not class I deﬁcient (data not shown). To determine whether the GFI1B block to CD8 development involved an alteration in repertoire selection, GFI1B-transgenic mice were mated to RAG2Ϫ/Ϫ/HY mice, and resulting progeny were backcrossed to generate GFI1B/ RAG2Ϫ/Ϫ/HY mice. The HY transgene encodes a class I-restricted TCR that selects large numbers of V␤8ϩ thymocytes into the CD8ϩ T cell lineage in female mice (20). Because the Rag2 gene product is necessary for TCR rearrangement, all RAG2Ϫ/Ϫ/HY/ GFI1B-transgenic thymocytes express only the HY TCR as evidenced by V␤8 staining (data not shown). Positive selection of CD8ϩ T cells by the HY TCR was severely reduced in the RAG2Ϫ/Ϫ/HY/GFI1B female mice as compared with controls (Fig. 6). Therefore, the critical defect in GFI1B-transgenic mice is Downloaded from not simply an inability to form a class I-restricted TCR.

GFI1B expression overcomes a block to DP development FIGURE 7. Expression of GFI1B overcomes a block to development in imposed by autoreactive TCR signaling male HY-transgenic T cells. Representative FACS plots showing expression of CD4 and CD8 (A) or CD8 and TCR V␤8(B) in male HY-transgenic To determine whether GFI1B-transgenic thymocytes display de- mice. Male RAG2Ϫ/Ϫ/HY/GFI1B mice show a 4-fold increase in the num- fective TCR signaling and/or activation in vivo, we next examined ber of CD4ϩCD8ϩ and TCR V␤8ϩ thymocytes when compared with http://www.jimmunol.org/ Ϫ Ϫ Ϫ Ϫ the male RAG2Ϫ/Ϫ/HY/GFI1B mouse thymus. The HY TCR rec- RAG2 / HY littermates. The total numbers of RAG2 / /HY and Ϫ/Ϫ ognizes the male HY Ag when presented by H-2Db. Male HY- RAG2 /HY/GFI1B thymocytes are shown above the plots. transgenic thymocytes are blocked at the DN stage by autoreactive HY TCR signaling that mimics negative selection (21). Mutant and in vivo (Fig. 7) data support a model in which ectopic expres- mice with defective intracellular signaling overcome this block sion of GFI1B leads to defective T cell activation. ϩ and accumulate V␤8 DP cells (22Ð24). GFI1B-transgenic periph- GFI1B impairs CD8 SP development in vitro and decreases eral T cells are defective in T cell activation (Figs. 3D and 4A). In IL-7R␣ expression

agreement with these data, GFI1B-transgenic thymocytes over- by guest on September 26, 2021 We have previously described an in vitro model system in which come the block to development imposed by the autoreactive HY DP thymocytes can be signaled to differentiate into CD8 SP T cells transgene, as shown by a modest 4-fold accumulation of ϩ ϩ (16). In this experimental system, signaled DP thymocytes initially CD8 TCR␤ cells (11.1 vs 47.9%) and a doubling of total thymus terminate CD8 transcription, differentiate into CD4ϩ8Ϫ interme- cellularity (Fig. 7, A and B). Therefore, our in vitro (Figs. 3 and 4) diate thymocytes, and up-regulate surface expression of IL-7R. In the presence of IL-7, CD4ϩ8Ϫ intermediate thymocytes terminate CD4 transcription and reinitiate CD8 transcription (events referred to as coreceptor reversal) and ultimately differentiate into CD8 SP T cells (16). Consequently, we assessed the ability of signaled DP thymocytes from WT and GFI1B-transgenic mice to differentiate in vitro into CD8 SP T cells (Fig. 8A). We isolated WT and GFI1B-transgenic DP thymocytes (Fig. 8A, D0) and stimulated them with PϩI as previously described (16). Signaled DP thymocytes from both WT and GFI1B-transgenic mice were induced to differentiate into CD4ϩ8Ϫ intermediate cells (Fig. 8A, D2). Nota- bly, in vitro-generated intermediate CD4ϩ8Ϫ cells from GFI1B- transgenic mice expressed lower surface levels of IL-7R␣ compared with cells from WT mice (Fig. 8A, D2). We then added IL-7 to both populations of in vitro-generated intermediate CD4ϩ8Ϫ thymocytes, and, after 24 h, pronase stripped the cells to remove pre-existing CD4/CD8 surface proteins so that we could determine the CD4/CD8 proteins that the cells were actively synthesizing. Addition of IL-7 (Fig. 8A, D3) followed by pronase stripping and re-expression culture revealed that 69.3% of WT cells had under- FIGURE 6. Expression of GFI1B inhibits the development of female gone coreceptor reversal and differentiated into CD8 SP T cells. In HY-transgenic CD8-SP T cells. Thymocytes from 4-wk-old female Ϫ Ϫ Ϫ Ϫ contrast, only 21.0% of GFI1B-transgenic cells had undergone co- RAG2 / /HY and RAG2 / /HY/GFI1B mice were examined for the ex- ␤ receptor reversal to become CD8 SP T cells (Fig. 7A, D4). Thus, pression of CD4, CD8, and TCR V 8 (used in the HY transgene) by ﬂow ϩ Ϫ cytometry. Representative FACS plots are shown with the total number of GFI1B-transgenic thymocytes at the CD4 8 intermediate stage RAG2Ϫ/Ϫ/HY and RAG2Ϫ/Ϫ/HY/GFI1B thymocytes above the plots. of development are quantitatively deﬁcient in their ability to un- Thymocyte subsets are expressed as absolute cell numbers ϫ 106 per dergo coreceptor reversal in response to IL-7 and are impaired in thymus Ϯ SEM. their ability to undergo in vitro differentiation into CD8 SP T cells. The Journal of Immunology 2363

expression of activation-induced genes such as CD25, CD69, and IL-7R␣. The lower level of IL-7R␣ on GFI1B-transgenic thymocytes provides a potential explanation for altered CD8 SP development in the GFI1B-transgenic mice. Moreover, examination of peripheral T cell subsets revealed lower levels of IL-7R␣ (data not shown). Because IL-7 has been shown to regulate the survival of naive CD4 T cells (25Ð28), the lack of IL-7R␣ expression is a possible explanation for peripheral CD4 T cell lymphopenia.

Bitransgenic BCL2/GFI1B mice generate CD8 SP thymocytes. GFI1B-transgenic thymocytes express lower levels of IL-7R␣ (Fig. 8). IL-7 maintains the expression of endogenous BCL2 in T-lineage cells (29). BCL2 does not support thymic positive selection in the absence of MHC (30Ð32); however, transgenic expression of BCL2 can substitute for survival signals induced by cytokines such as IL-7 (32). GFI1B-transgenic mice were mated to E␮-BCL2Ð25-transgenic mice, in which the human BCL2 transgene is expressed mainly in T-lineage cells (12). As expected by Downloaded from the presence of the BCL2 transgene, total thymocyte cellularity was increased in bitransgenic mice (12); however, we found that BCL2/GFI1B bitransgenics show increased numbers of TCRint-high CD8 SP thymocytes and a normal ratio of CD4 SP to CD8 SP thymocytes (Fig. 9). Although BCL2 may have pleiotropic effects on T cell development (31), the ability of BCL2 to rescue CD8 SP http://www.jimmunol.org/ development in GFI1B-transgenic mice is consistent with the role of BCL2 as a downstream target of IL-7. by guest on September 26, 2021

FIGURE 8. GFI1B impairs coreceptor reversal. A, D0, Purified CD4ϩCD8ϩ cells were obtained from pooled thymocytes of four 4-wk-old GFI1B-transgenic or four control littermates by panning with anti-CD8 Ab, and subsequently stimulated with PϩIfor16hat37¡C. Stimulated cells were stripped of their surface coreceptors with pronase protease, cultured in complete medium overnight, and stained for CD4 and CD8 expression, and purified CD4ϩ8Ϫ cells were obtained by cell sorting. D2, Control (dashed line) and GFI1B-transgenic (solid line) cells were analyzed by flow cytometry for IL-7R␣ expression. Purified CD4ϩ8Ϫ cells (2 ϫ 106) were cultured overnight in the presence of IL-7. Flow analysis of samples of each culture indicated CD4 and CD8 surface levels before (D3) and after pronase protease treatment and culture in complete medium overnight (D4). B, Four-week-old GFI1B-transgenic (solid line) and normal littermate controls (dashed line) were stained for CD4, CD8, and IL-7R␣, and analyzed by flow cytometry. Representative histograms comparing the IL- 7R␣ expression in thymocyte subpopulations are shown.

To determine whether the lower in vitro expression pattern of IL-7R␣ on GFI1B-transgenic thymocytes was a possible explanation for the defective CD8 SP development in vivo, we next ex- FIGURE 9. Transgenic BCL2 restores CD8 SP thymocytes in GFI1B- amined the expression of IL-7R␣ in unmanipulated thymocytes. transgenic mice. Thymocytes from GFI1B and BCL2/GFI1B-bitransgenic Cells were stained for CD4, CD8, and IL-7R␣, then gated on thy- littermates were stained with Abs to TCR␤, CD4, and CD8, and analyzed int-high mocyte subpopulations and analyzed for IL-7R␣ expression (Fig. by flow cytometry. Cells are gated on TCR events, and representative CD4 vs CD8 FACS profiles are depicted. A scatter plot of CD8 SP cells in 8B). The level of IL-7R␣ on DN thymocytes was not altered by the BCL2-transgenic, GFI1B-transgenic, and BCL2/GFI1B-bitransgenic mice presence of the GFI1B transgene. However, in both CD4 SP and reveals the increase in the CD8 SP population in the bitransgenic animals ␣ CD8 SP thymocytes, in which IL-7R expression is regulated by (in comparison to GFI1B-transgenic mice). Each diamond represents the TCR signaling, GFI1B-transgenic mice had lower levels of IL- percentage of TCRint-high CD8 SP cells per thymus of an individual mouse 7R␣ than did WT littermate controls (Fig. 8B). Thus, GFI1B-trans- (five to seven mice per group). The black bar is the mean value for each genic thymocytes are defective in both T cell activation and the group. 2364 INTEGRATION OF THYMOCYTE ACTIVATION AND LINEAGE COMMITMENT

Bitransgenic GFI1/GFI1B mice generate CD8 SP thymocytes expression of IL-7R␣ (data not shown) indicate that both BCL2 We next determined whether GFI1 could alter the defects engen- and GFI1 must act to increase CD8 SP generation by a mechanism dered by GFI1B expression. GFI1 enhances T cell activation, that is independent of the induction of IL-7 signaling. However, it whereas GFI1B impairs this process (Fig. 3). Because GFI1 and is formally possible that GFI1, like BCL2, acts downstream of GFI1B bind to the same DNA sequence, it is possible that some of IL-7 signaling and that transgenic expression of either protein the defects in GFI1B-transgenic thymocytes are the result of an mimics the effects of IL-7. imbalance between DNA-bound GFI1B vs GFI1. In fact, 6-wk-old GFI1/GFI1B-bitransgenic mice generate twice as many TCRint-high Discussion CD8 SP thymocytes than do littermate GFI1B transgenics (Fig. 10, The phenotypes observed in GFI1- and GFI1B-transgenic mice p ϭ 0.0003). The lower level of expression of the GFI1 transgene can be understood in the context of the normal expression pattern in comparison to the GFI1B transgene may explain the modest of either factor. GFI1 is induced at the transition between DP and ability of GFI1 transgene to compete with transgenic GFI1B. Nev- SP cells. GFI1 may thus play a role in the activation-induced de- ertheless, GFI1 expression increases the number of GFI1B-trans- velopmental steps between DP and SP thymocytes. GFI1B is induced at steps in thymocyte development in which thymocytes are genic CD8 SP cells. Ϫ/Ϫ Transgenic expression of GFI1 does not alter other GFI1B-in- activated. In fact, RAG2 cells that lack the ability to activate do duced defects. The GFI1B-induced abnormal generation of phe- not express GFI1B (Fig. 1). GFI1 and GFI1B have opposite effects notypically mature CD4 SP thymocytes remains in the GFI1/ on T cell activation. Peripheral GFI1-transgenic T cells activate GFI1B-bitransgenic thymus (Fig. 10). Moreover, peripheral T better than do those from WT littermates, whereas GFI1B-trans- lymphopenia in GFI1B transgenics is not altered by BCL2 or genic T cells are impaired in activation. Both phenotypes are cell Downloaded from GFI1. It is known that BCL2 transgene-induced elevation in thy- autonomous, as they occur in purified T cells from either trans- mocyte numbers does not alter peripheral T cell numbers (29). genic (Fig. 4, and data not shown). Likewise, we find that neither BCL2 nor GFI1 increases the num- The impairment of T cell activation function in GFI1B-trans- ber of peripheral CD8 SP T cells in the GFI1B-transgenic spleen. genic T cells appears to be linked to an inability to signal properly GFI1, like BCL2, appears to ameliorate only the GFI1B effect on after TCR engagement and costimulation. First, purified GFI1B- the production of CD8 SP cells. The inability of either the BCL2 transgenic T cells do not activate given CD28- and CD3-cross- http://www.jimmunol.org/ or GFI1 transgene to cancel the GFI1B-induced alteration in the linking Abs (Fig. 4A). Second, the TCR signaling-dependent RAS/ mitogen-activated protein kinase-induced expression of early activation markers CD25 and CD69 (33) is impaired in GFI1B- transgenic T cells (Fig. 4B). Finally, in male HY-transgenic mice, GFI1B expression rescues the generation of DP thymocytes that are normally deleted due to strong autoreactive TCR signals (21). Similar results have been obtained in male HY-transgenic mice that are rendered defective in TCR signaling by deletion of intracellular proteins that participate in the TCR-signaling cascade (22Ð by guest on September 26, 2021 24). However, the GFI1B-induced defect in T cell activation can- not be restricted to proximal signaling molecules in the TCR- and costimulatory-signaling pathways, because GFI1B-induced impairment of CD8 SP cell formation is also observed after in vitro drug-stimulated activation that bypasses the need for proximal TCR- and costimulatory-signaling events (Fig. 8). The ability of GFI1B-transgenic thymocytes to respond to cytokine signaling may also be affected. Although activated GFI1B-transgenic thymocytes fail to induce IL-7R␣ expression, transgenic expression of IL-7R␣ (34) did not change the GFI1B-induced phenotypes (data not shown), whereas expression of the IL-7 downstream effector BCL2 increased CD8 generation. Thus, GFI1B expression in T cells induces cell-autonomous defects in intracellular signaling that may impair T cell activation at multiple steps. GFI1B may serve as a negative regulator of GFI1-enhanced activation. Both GFI1 and GFI1B are expected to repress genes to alter the kinetics or activation potential of intracellular signals. The induction of GFI1B in activated thymocytes could result in com- FIGURE 10. Transgenic GFI1 increases the generation of CD8 SP thy- petition between GFI1 and GFI1B for DNA binding at specific mocytes in GFI1B-transgenic mice. Thymocytes from 6-wk-old GFI1B promoters. Alternatively, GFI1 and GFI1B may regulate different and GFI1/GFI1B-bitransgenic littermates were stained with Abs to TCR␤, promoters. Both scenarios appear to be relevant to the GFI1B- CD4, and CD8, and analyzed by flow cytometry. Cells are gated on induced T cell defects. Because transgenic GFI1 expression dou- int-high TCR events, and representative CD4 vs CD8 FACS profiles are bles the generation of GFI1B-transgenic CD8 SP cells, GFI1 and depicted. A scatter plot of CD8 SP cells in GFI1-transgenic, GFI1B-trans- GFI1B may compete at promoters to regulate genes important for genic, and GFI1/GFI1B-bitransgenic mice shows the partial restoration of CD8 development. Target genes responsible for this phenotype CD8 SP thymocytes observed in bitransgenic animals (in comparison to GFI1B-transgenic mice). Each diamond represents the percentage of may be properly regulated by GFI1, but improperly regulated by TCRint-high CD8 SP cells per thymus of an individual mouse (four to six GFI1B in cells about to undergo lineage commitment. Interest- mice per group). The black bar is the mean value for each group. The ingly, transgenic BCL2 or GFI1 rescued CD8 SP development, but increase is statistically significant with a p value of 0.0003 as determined did not affect GFI1B-induced changes in CD4 SP development, the using two-tailed Student’s t test. expression of IL-7R␣, or peripheral T lymphopenia. Target genes The Journal of Immunology 2365 responsible for the latter GFI1B-induced phenotypes should be for Cellular Therapeutics for helpful discussions, and we give special independent of GFI1 regulation and instead uniquely regulated by thanks to Tarik Moroy and Holger Karsunky for the CD2-GFI1 mice, and GFI1B. GFI1 and GFI1B differ in amino acid sequence in the Thomas Malek for the generous gift of IL-7R␣-transgenic mice. region between the SNAG repressor domain and the zinc-finger DNA-binding domain. These dissimilar regions may mediate in- References teraction with other proteins (such as other transcription factors or 1. Lo, D., and J. Sprent. 1986. Identity of cells that imprint H-2-restricted T-cell specificity in the thymus. Nature 319:672. adapter proteins), leading to differential regulation of target pro- 2. Anderson, G., E. J. Jenkinson, N. C. Moore, and J. J. Owen. 1993. MHC class moters by GFI1 and GFI1B. The normal role of GFI1B in the II-positive epithelium and mesenchyme cells are both required for T-cell devel- thymus may be to regulate the extent of GFI1-mediated thymocyte opment in the thymus. Nature 362:70. 3. Schwartz, R. H. 1989. Acquisition of immunologic self-tolerance. Cell 57:1073. activation during development. Our data are the first to demon- 4. Gilks, C. B., S. E. Bear, H. L. Grimes, and P. N. Tsichlis. 1993. Progression of strate that GFI1 and GFI1B are not redundant for T cell activation interleukin-2 (IL-2)-dependent rat T cell lymphoma lines to IL-2-independent growth following activation of a gene (Gfi-1) encoding a novel zinc finger pro- functions, providing a potential biological explanation for the ter- tein. Mol. Cell. Biol. 13:1759. mination of GFI1B expression in mature T cells. 5. Grimes, H. L., T. O. Chan, P. A. Zweidler-McKay, B. Tong, and P. N. Tsichlis. IL-7 is critical for the survival and proliferation of immature 1996. The Gfi-1 proto-oncoprotein contains a novel transcriptional repressor domain, SNAG, and inhibits G1 arrest induced by interleukin-2 withdrawal. Mol. thymocytes into mature CD8 SP cells (16, 32, 35). IL-7 is consti- Cell. Biol. 16:6263. tutively present in the thymus; however, IL-7R␣ expression is not 6. Tong, B., H. L. Grimes, T. Y. Yang, S. E. Bear, Z. Qin, K. Du, W. S. El Deiry, WAF1 induced until TCR stimulation ceases (16). Thus, whether TCR and P. N. Tsichlis. 1998. The Gfi-1B proto-oncoprotein represses p21 and inhibits myeloid cell differentiation. Mol. Cell. Biol. 18:2462. signaling continues or ceases determines the ability to differentiate 7. Zornig, M., T. Schmidt, H. Karsunky, A. Grzeschiczek, and T. Moroy. 1996. Zinc into CD8 SP T cells. This new perspective on lineage commitment finger protein GFI-1 cooperates with myc and pim-1 in T-cell lymphomagenesis Downloaded from by reducing the requirements for IL-2. Oncogene 12:1789. is referred to as the kinetic signaling model. In accordance with 8. Karsunky, H., H. Zeng, T. Schmidt, B. Zevnik, R. Kluge, K. W. Schmid, this model, GFI1B-transgenic thymocytes, in which GFI1 function U. Duhrsen, and T. Moroy. 2002. Inflammatory reactions and severe neutropenia is compromised, fail to integrate activation signals, resulting in in mice lacking the transcriptional repressor Gfi1. Nat. Genet. 30:295. 9. Saleque, S., S. Cameron, and S. H. Orkin. 2002. The zinc-finger proto-oncogene impaired expression of activation-induced survival genes such as Gfi-1b is essential for development of the erythroid and megakaryocytic lineages. IL-7R␣ (16). Models of lineage commitment which require quan- Genes Dev. 16:301. 10. Rodel, B., K. Tavassoli, H. Karsunky, T. Schmidt, M. Bachmann, F. Schaper, titative or temporal differences in TCR signals to direct T cell http://www.jimmunol.org/ P. Heinrich, K. Shuai, H. P. Elsasser, and T. Moroy. 2000. The zinc finger protein development (36, 37) would predict that the reduced activation in Gfi-1 can enhance STAT3 signaling by interacting with the STAT3 inhibitor GFI1B-transgenic T cells should lead to improved CD8 SP devel- PIAS3. EMBO J. 19:5845. 11. Perez, P., S. A. Lira, and R. Bravo. 1995. Overexpression of RelA in transgenic opment (38). Transgenic expression of GFI1B in the thymus in- mouse thymocytes: specific increase in levels of the inhibitor protein I ␬B ␣. Mol. stead results in a severe decrease in the generation of CD8 SP cells. Cell. Biol. 15:3523. A similar phenotype is exhibited by mice made deficient for IFN 12. Strasser, A., A. W. Harris, and S. Cory. 1991. bcl-2 transgene inhibits T cell death ␣␤ and perturbs thymic self-censorship. Cell 67:889. signaling by knocking out the transcription factor IRF-1 (39). 13. Herblot, S., P. D. Aplan, and T. Hoang. 2002. Gradient of E2A activity in B-cell These mice are devoid of CD8 SP T cells even though they contain development. Mol. Cell. Biol. 22:886. normal numbers of CD4 SP T cells. We have established that the 14. Herblot, S., A. M. Steff, P. Hugo, P. D. Aplan, and T. Hoang. 2000. SCL and LMO1 alter thymocyte differentiation: inhibition of E2A-HEB function and pre-T by guest on September 26, 2021 GFI1B-transgenic thymus defects are not due to GFI1B repression ␣-chain expression. Nat. Immunol. 1:138. of IRF-1 (Fig. 2B), CD8, or MHC class I (data not shown), the 15. Schmidt, T., H. Karsunky, B. Rodel, B. Zevnik, H. P. Elsasser, and T. Moroy. 1998. Evidence implicating Gfi-1 and Pim-1 in pre-T-cell differentiation steps absence of a functional TCR rearrangement (Fig. 6), or a TCR associated with ␤-selection. EMBO J. 17:5349. signal-independent redirection of CD8 cells to a CD4 lineage 16. Brugnera, E., A. Bhandoola, R. Cibotti, Q. Yu, T. I. Guinter, Y. Yamashita, choice (data not shown). However, in agreement with the kinetic S. O. Sharrow, and A. Singer. 2000. Coreceptor reversal in the thymus: signaled CD4ϩ8ϩ thymocytes initially terminate CD8 transcription even when differen- signaling model, a signal-independent survival cue from the down- tiating into CD8ϩ T cells. Immunity 13:59. stream target of IL-7, BCL2, increases CD8 SP generation in the 17. Lucas, B., and R. N. Germain. 1996. Unexpectedly complex regulation of CD4/ ϩ ϩ context of GFI1B expression. Therefore, the defect in generation CD8 coreceptor expression supports a revised model for CD4 CD8 thymocyte differentiation. Immunity 5:461. of CD8 SP cells in GFI1B-transgenic mice appears to be linked to 18. Robey, E., and B. J. Fowlkes. 1994. Selective events in T cell development. an inability to transduce a postactivation signal-dependent Annu. Rev. Immunol. 12:675. 19. Punt, J. A., H. Suzuki, L. G. Granger, S. O. Sharrow, and A. Singer. 1996. survival cue. Lineage commitment in the thymus: only the most differentiated (TCRhibcl-2hi) GFI1 functions to inhibit T cell death induced by specific stim- subset of CD4ϩCD8ϩ thymocytes has selectively terminated CD4 or CD8 syn- uli. Specifically, GFI1 increases survival of explanted thymocytes thesis. J. Exp. Med. 184:2091. 20. Kisielow, P., H. Bluthmann, U. D. Staerz, M. Steinmetz, and H. von Boehmer. (40) and decreases apoptosis induced by TCR ligation (41). In 1988. Tolerance in T-cell-receptor transgenic mice involves deletion of nonma- GFI1B-transgenic T cells, GFI1 expression does not correct ture CD4ϩ8ϩ thymocytes. Nature 333:742. GFI1B-impaired expression of IL-7R␣; however, like BCL2, 21. Takahama, Y., E. W. Shores, and A. Singer. 1992. Negative selection of precur- sor thymocytes before their differentiation into CD4ϩCD8ϩ cells. Science 258: transgenic GFI1 increases CD8 SP generation in the GFI1B-trans- 653. genic thymus. The salient function of GFI1 in CD8 SP generation 22. Yoder, J., C. Pham, Y. M. Iizuka, O. Kanagawa, S. K. Liu, J. McGlade, and A. M. Cheng. 2001. Requirement for the SLP-76 adaptor GADS in T cell de- may be GFI1 mediation of cytokine-induced survival signaling velopment. Science 291:1987. that is impeded in GFI1B-transgenic thymocytes. Taken together, 23. Saijo, K., S. Y. Park, Y. Ishida, H. Arase, and T. Saito. 1997. Crucial role of Jak3 these data are the first to suggest a role for GFI1 in the integration in negative selection of self-reactive T cells. J. Exp. Med. 185:351. 24. Rincon, M., A. Whitmarsh, D. D. Yang, L. Weiss, B. Derijard, P. Jayaraj, of activation signals from the TCR with survival signals from cy- R. J. Davis, and R. A. Flavell. 1998. The JNK pathway regulates the in vivo tokines such as IL-7. deletion of immature CD4ϩCD8ϩ thymocytes. J. Exp. Med. 188:1817. 25. Schluns, K. S., W. C. Kieper, S. C. Jameson, and L. Lefrancois. 2000. Interleu- kin-7 mediates the homeostasis of naive and memory CD8 T cells in vivo. Nat. Acknowledgments Immunol. 1:426. 26. Webb, L. M., B. M. Foxwell, and M. Feldmann. 1999. Putative role for inter- We thank Jennifer Punt, Gordon Ross, and Thomas Mitchell for critically leukin-7 in the maintenance of the recirculating naive CD4ϩ T-cell pool. Immu- reviewing the manuscript. We gratefully acknowledge the following indi- nology 98:400. viduals: Avedis Kazanjian, Amy Barber Shonk, Rachel Rivoli, Natalie 27. Vivien, L., C. Benoist, and D. Mathis. 2001. T lymphocytes need IL-7 but not Claudio, and Yajuan Jiang for excellent technical assistance; Barry Udis IL-4 or IL-6 to survive in vivo. Int. Immunol. 13:763. 28. Tan, J. T., E. Dudl, E. LeRoy, R. Murray, J. Sprent, K. I. Weinberg, and and Michael Tanner for expert flow cytometry; and Michael Alexander for C. D. Surh. 2001. IL-7 is critical for homeostatic proliferation and survival of initial T cell activation assays. We also thank the members of the Institute naive T cells. Proc. Natl. Acad. Sci. USA 98:8732. 2366 INTEGRATION OF THYMOCYTE ACTIVATION AND LINEAGE COMMITMENT

29. Freeden-Jeffry, U., N. Solvason, M. Howard, and R. Murray. 1997. The earliest 36. Yasutomo, K., C. Doyle, L. Miele, C. Fuchs, and R. N. Germain. 2000. The T lineage-committed cells depend on IL-7 for Bcl-2 expression and normal cell duration of antigen receptor signalling determines CD4ϩ versus CD8ϩ T-cell cycle progression. Immunity 7:147. lineage fate. Nature 404:506. 30. Strasser, A., A. W. Harris, L. M. Corcoran, and S. Cory. 1994. Bcl-2 expression 37. Wilkinson, B., and J. Kaye. 2001. Requirement for sustained MAPK signaling promotes B- but not T-lymphoid development in scid mice. Nature 368:457. in both CD4 and CD8 lineage commitment: a threshold model. Cell. Immunol. 211:86. 31. Linette, G. P., M. J. Grusby, S. M. Hedrick, T. H. Hansen, L. H. Glimcher, and ϩ ϩ 38. Matechak, E. O., N. Killeen, S. M. Hedrick, and B. J. Fowlkes. 1996. MHC class S. J. Korsmeyer. 1994. Bcl-2 is upregulated at the CD4 CD8 stage during II-specific T cells can develop in the CD8 lineage when CD4 is absent. Immunity positive selection and promotes thymocyte differentiation at several control 4:337. points. Immunity 1:197. 39. Penninger, J. M., C. Sirard, H. W. Mittrucker, A. Chidgey, I. Kozieradzki, 32. Akashi, K., M. Kondo, U. Freeden-Jeffry, R. Murray, and I. L. Weissman. 1997. M. Nghiem, A. Hakem, T. Kimura, E. Timms, R. Boyd, et al. 1997. The inter- Bcl-2 rescues T lymphopoiesis in interleukin-7 receptor-deficient mice. Cell 89: feron regulatory transcription factor IRF-1 controls positive and negative selec- 1033. tion of CD8ϩ thymocytes. Immunity 7:243. 33. D’Ambrosio, D., D. A. Cantrell, L. Frati, A. Santoni, and R. Testi. 1994. In- 40. Grimes, H. L., C. B. Gilks, T. O. Chan, S. Porter, and P. N. Tsichlis. 1996. The volvement of p21ras activation in T cell CD69 expression. Eur. J. Immunol. Gfi-1 protooncoprotein represses Bax expression and inhibits T-cell death. Proc. 24:616. Natl. Acad. Sci. USA 93:14569. 41. Karsunky, H., I. Mende, T. Schmidt, and T. Moroy. 2002. High levels of the 34. Porter, B. O., P. Scibelli, and T. R. Malek. 2001. Control of T cell development onco-protein Gfi-1 accelerate T-cell proliferation and inhibit activation induced ␣ in vivo by subdomains within the IL-7 receptor -chain cytoplasmic tail. J. Im- T-cell death in Jurkat T-cells. Oncogene 21:1571. munol. 166:262. 42. Shinkai, Y., G. Rathbun, K. P. Lam, E. M. Oltz, V. Stewart, M. Mendelsohn, 35. Maraskovsky, E., L. A. O’Reilly, M. Teepe, L. M. Corcoran, J. J. Peschon, and J. Charron, M. Datta, F. Young, and A. M. Stall. 1992. RAG-2-deficient mice A. Strasser. 1997. Bcl-2 can rescue T lymphocyte development in interleukin-7 lack mature lymphocytes owing to inability to initiate V(D)J rearrangement. Cell receptor-deficient mice but not in mutant rag-1Ϫ/Ϫ mice. Cell 89:1011. 68:855. Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021