CD8 Lineage-specific Regulation of Interleukin-7 Receptor Expression by the Transcriptional Repressor Gfi1*

Background: Expression of the IL-7Rα gene is up-/down-regulated during T/B-lymphocyte development. Results: IL-7Rα gene transcription is repressed by the transcription factor Gfi1, specifically in CD8+ T-lymphocytes. Conclusion: Treatment by dexamethasone down-regulates Gfi1, which contributes to glucocorticoid receptor mediated up-regulation of IL-7R expression. Significance: The mechanism by which the IL-7R gene gets turned on and off during development is a critical issue in biology. Interleukin-7 receptor α (IL-7Rα) is essential for T cell survival and differentiation. Glucocorticoids are potent enhancers of IL-7Rα expression with diverse roles in T cell biology. Here we identify the transcriptional repressor, growth factor independent-1 (Gfi1), as a novel intermediary in glucocorticoid-induced IL-7Rα up-regulation. We found Gfi1 to be a major inhibitory target of dexamethasone by microarray expression profiling of 3B4.15 T-hybridoma cells. Concordantly, retroviral transduction of Gfi1 significantly blunted IL-7Rα up-regulation by dexamethasone. To further assess the role of Gfi1 in vivo, we generated bacterial artificial chromosome (BAC) transgenic mice, in which a modified Il7r locus expresses GFP to report Il7r gene transcription. By introducing this BAC reporter transgene into either Gfi1-deficient or Gfi1-transgenic mice, we document in vivo that IL-7Rα transcription is up-regulated in the absence of Gfi1 and down-regulated when Gfi1 is overexpressed. Strikingly, the in vivo regulatory role of Gfi1 was specific for CD8+, and not CD4+ T cells or immature thymocytes. These results identify Gfi1 as a specific transcriptional repressor of the Il7r gene in CD8 T lymphocytes in vivo.

A critical issue in biology is the mechanism by which genes get turned on and off during development and differentiation. Because IL-7 receptor (IL-7R) proteins provide critical survival signals to developing lymphocytes, the expression of the Il7r gene that encodes the IL-7R␣ receptor protein is tightly regulated at different stages of T and B lymphocyte development and precisely timed to stages when selection and programmed cell death occur in the immune system (1)(2)(3). The expression of IL-7R␣ follows an on-off-on pattern in the thymus at the CD4 Ϫ CD8 Ϫ double negative (DN), 8 CD4 ϩ CD8 ϩ double positive (DP), and CD4 ϩ CD8 Ϫ or CD4 Ϫ CD8 ϩ single positive (SP) stages, respectively (4). Thus, developmental cues during thymocyte differentiation control IL-7R␣ expression. During CD8 ϩ memory cell generation in the peripheral immune system, Il7r gene expression again correlates with developmental outcome, in that long-lived memory cell precursors up-regulate IL-7R␣ expression and short-lived CD8 ϩ cells lose IL-7R␣ expression (5). Notably, up-regulated IL-7R␣ expression is not sufficient to drive long-lived memory CD8 ϩ T cell generation, even though IL-7R␣ up-regulation clearly marks progenitors of this T cell subset (6,7). Importantly, the differentiation signals that match IL-7R␣ expression to CD8 T cell fate remain unknown.
In T cells, IL-7R␣ expression is thought to be primarily regulated at the transcriptional level through an array of nuclear factors whose expression is also tightly controlled during development and activation. Several transcription factors that control Il7r gene expression have been identified. The promoter of Il7r contains binding sites for the PU.1 transcription factor, which is necessary for the IL-7R␣ expression in developing B cells (8,9). The same site is occupied in T cells by another ETS family transcription factor, GABP (10). Promoter occupancy by these factors likely prevents CpG methylation of promoter sequences and subsequent down-regulation of expression in mature T cells (11). Additionally, in human thymopoiesis, Notch may be complementing these ETS family proteins by acting through a conserved RBP-Jk/CSL binding site close by in the promoter of the Il7r gene (12). Therefore, down-regulation of Notch expression at the DP stage may be causative of the complete loss of Il7r gene transcription in murine DP thymocytes. Also, the potential role of microRNAs acting on the Il7r gene locus, specifically at the DP stage has not been addressed and needs to be tested. Furthermore, the zinc finger protein Gfi1, for which a regulatory role was originally proposed in T cells and more recently confirmed in pro-B cells, was shown to bind to a putative intronic silencer (13)(14)(15). Additionally, glucocorticoid receptor (GR), Runx1/3, FoxOA1/3, and FoxP1 were all shown to bind to a putative enhancer in an evolutionarily conserved region 3.5 kb upstream of the gene (16 -21). Finally FoxP3 was found to bind near the promoter in T reg cells to suppress IL-7R␣ transcription (22). Importantly, however, how these factors interact with each other and what controls the mechanism of developmental stage-specific differences in Il7r gene transcription remains ill defined.
In the present study, we addressed this issue first by profiling gene expression in 3B4. 15 T hybridoma cells that respond to dexamethasone (Dex) treatment by up-regulating IL-7R␣ expression (23). We identified Gfi1 as a novel target of Dex and we further documented that either Gfi1 overexpression or treatment with the glucocorticoid receptor (GR) inhibitor RU486 (Mifepristone) in 3B4.15 cells prevented IL-7R␣ up-regulation by Dex. These results indicate that Gfi1 is either controlled by GR or cooperates with it to down-regulate IL-7R␣ expression. To further assess the role of Gfi1 in vivo, we then generated a novel bacterial artificial chromosome (BAC) transgenic mouse that reports transcriptional activity of the Il7r gene locus. We show that Gfi1 is a transcriptional repressor of the Il7r gene locus, but only in CD8 lineage cells, by assessing Il7r reporter activity in Gfi1-deficient and Gfi1-transgenic thymocytes and T cells. Our observations place Gfi1 as a lineagespecific and developmental stage-dependent transcriptional repressor of IL-7R␣ in vivo.

EXPERIMENTAL PROCEDURES
Animals-C57BL/6 and RAG2-deficient mice were obtained from The Jackson Laboratory (Bar Harbor, ME). Gfi1-deficient (Gfi1 KO ) and Gfi1-transgenic (Gfi1 Tg ) mice have been previously described (24,25). Animal experiments were approved by the National Cancer Institute Animal Care and Use Committee, and all mice were cared for in accordance with National Institutes of Health (NIH) guidelines.
Generation of 7RIG-BAC Tg Transgenic Mice-A BAC clone (RP23-365P6) containing the Il7r gene locus was modified by recombineering an IRES-EGFP cassette into the 3Ј UTR region of the gene in Escherichia coli (26). Briefly, a targeting vector was generated containing (1) an HincII fragment of the pIRES2EGFP plasmid (Clontech) (2), an SV40 late poly(A) signal sequence PCR amplified from the pGL3Basic plasmid (Promega), (3) a KpnI fragment of the pLTM260 plasmid containing an Frt and a loxP-flanked Neomycin resistance gene with a PGK promoter and a bGH poly(A) signal and (4) two flanking regions (210 and 300 bp long) homologous to the Il7r 3Ј UTR PCR amplified from BAC DNA. This reporter BAC DNA was purified by sucrose gradient centrifugation and was injected into fertilized B6 oocytes to generate transgenic mice as described (27). Founder mice were identified by flow cytometric detection of GFP expression on peripheral blood lymphocytes. One transgenic line out of 4 founders that recapitulated IL-7R␣ expression patterns on peripheral T and B lymphocytes was selected for further study and named for this study "7RIG-BAC Tg ." Note that this transgene is unique compared with a recently reported BAC transgene, as we utilized an internal ribosome entry site (IRES) element in the 3Ј UTR to ensure GFP reporter expression coincident with transgenic IL-7R␣ expression (28). The transgene was introduced into a Gfi1 Tg or Gfi1 KO background by breeding.
Genotyping of IL-7R␣ KO Allele by Quantitative Real Time-PCR-Originally, the IL-7R␣-deficient (IL-7R␣ KO ) mouse was generated by inserting a 1-kb MC1 neo cassette into a HindIII site within the third exon of the Il7r gene, around position 90 of the 180-amino acid long extracellular domain (1). The originally inserted neo r gene is a modified neo r gene from pMC1Neo as described in Thomas and Capecchi (29). In this altered neomycin resistance gene, a synthetic translation initiation sequence "5Ј-gccaatatgggatcggcc-3Ј" is introduced. We used the reverse sequence of this synthetic translation initiation sequence in combination with an Il7r exon3-specific primer to amplify a short PCR fragment. The exon 3-specific primer corresponds to the amino acid sequence "GSSNICV" of the IL-7R␣ extracellular domain. Copy numbers of the IL-7R␣ KO allele was determined by real time-PCR using primers IL7Rex3GSSNICV and MC1neo-R.
Adoptive Transfer-Purified LNT cells from 7RIG-BAC Tg and Gfi Tg 7RIG-BAC Tg mice were labeled with CellTrace Violet (Invitrogen) and adoptively transferred into RAG2-deficient host mice. 4 ϫ 10 6 -Labeled cells were intravenously injected, and spleen and lymph node cells from host mice were harvested 5 days later. Single cell suspensions were stained for surface IL-7R␣ and TCR␤ expression and analyzed by flow cytometry.
RNA Isolation and Northern Blot Analysis-Total RNA was isolated using TRIzol (Invitrogen). Equal amounts of RNA were resolved in a 1.5% agarose gel under denaturing conditions and blotted onto Hybond-N ϩ nylon membranes (Amersham Biosciences). Radioactive probes for detecting specific gene expression were generated using the EZ-strip DNA kit (Ambion) and used to hybridize with RNA-blotted membrane in UltraHyb hybridization solution (Ambion) at 42°C. The next day, membranes were washed two times with 2ϫ SSC, 0.1% SDS for 30 min and two more times with 0.1ϫ SSC, 0.1% SDS at 55°C. Membranes were exposed to a PhosphorImager screen (Amersham Biosciences) and analyzed.
Expression Plasmids and Gene Transfer-Full-length and truncation mutants of murine Gfi1 cDNAs were C-terminal FLAG epitope-tagged and cloned into the pBluescript II plasmid using oligonucleotides incorporating 5Ј XhoI and 3Ј NotI restriction sites. All cDNAs were transferred from pBluescript II to a retroviral expression plasmid, LZRSpBMN-linker-IRES-EGFP, using XhoI and NotI restriction enzymes. This resulted in bicistronic expression of FLAG epitope-tagged Gfi1 cDNA variants with an EGFP reporter gene. The following oligonucleotide pairs were used to amplify Gfi1 truncations: D-ZF, M13ϩdZFsrev; mGfi1-ZF, ZFsforϩT7; and D-SNAG, dSNAGforϩT7. LZRSpBMN-linker-IRES-EGFP with fulllength or truncated Gfi1 cDNAs were transfected into Phoenix-Eco retroviral packaging cell lines with a plasmid encoding ecotropic retrovirus envelope proteins (pCL-Eco Addgene plasmid 12371) (30) and supernatants were collected for 2 days, pooled, and filtered through 45-m filters. 3B4.15 cells were infected by spin infection in the presence of 6 g/ml of Polybrene (Sigma).
Oligonucleotides Used in this Study-The following oligonucleotides were used to PCR amplify Gfi1 cDNA constructs. Microarray Analysis-Expression analysis was performed on 3B4.15 T hybridoma cells, either untreated or treated with 1 M dexamethasone (Sigma) for 16 h. Total RNA was extracted using TriReagent (Sigma), RNA quality was confirmed on an RNA 6000 Nano chip (AGT-5067-1511) in an Agilent Bioanalyzer. Double-stranded cDNA was generated using a Super-Script cDNA Synthesis Kit (Invitrogen), and the cDNA was then labeled with Cy-3, cleaned, quantified, and hybridized according to the manufacturer's protocols (Roche-Nimblegen). Nimblegen full genome Mouse Expression arrays (12X135K RO5543797) were washed and scanned at the Sabanci University Nanotechnology Research and Application Center-SUNUM. Results were processed using the ANAIS software (31). Array quality was assessed at the probe level. Values for 3 probes for each gene in each array were combined to summarize gene expression from probe sets. Robust Multi-Array Analysis background normalization and quantile normalization were performed for intra-and inter-array normalization, respectively. Genes with signal intensities above a 95% random threshold were chosen for further studies. Differentially expressed genes were obtained based on the following criteria: fold-change Ն2.5 and analysis of variance p value Յ 0.01. Hierarchical clustering was applied to the top 500 differentially expressed genes with Genesis software (32). The data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus (33) and are accessible through GEO Series accession number GSE39296.

Glucocorticoids Induce IL-7R␣ Expression by Down-regulat-
ing Expression of Gfi1-Treatment of primary T lymphocytes and T cell lines with glucocorticoids, such as Dex, results in the up-regulation of surface IL-7R␣ expression (17,34). T cells normally express high levels of surface IL-7R␣ but the I-E k -restricted, PCC-specific T cell hybridoma 3B4.15 expresses only low levels of IL-7R␣ (23). Nevertheless, when treated with glucocorticoids such as Dex, 3B4.15 cells dramatically up-regulate both IL-7R␣ mRNA and cell surface protein expression (Fig. 1,  A and B). These results parallel those previously obtained using the transformed murine T cell line KKF, which responds to Dex by up-regulating IL-7R␣ protein expression (16). Dex treatment induces nuclear localization and binding of the glucocorticoid receptor transcription factor to an evolutionarily conserved region 3.5 kb upstream of the transcriptional start site of the Il7r gene (16). Thus, Dex-induced IL-7R␣ up-regulation has been considered to be a direct transcriptional effect of activated GRs. To understand other gene regulators that control this phenotypic change, we compared the gene expression profiles of mock treated versus Dex-treated 3B4.15 cells by high coverage Nimblegen expression arrays with 135,000 features.
To confirm the specificity of Dex signaling in 3B4.15 hybridomas, first, we examined the expression profiles of the known Dex-regulated genes, GILZ (Tsc22d3) and GITR (Tnfrsf18) and found that these genes were positioned in the top 520 differentially expressed genes (supplemental Fig. S1) (35)(36)(37). Next, we further analyzed the expression profiles of all transcription factors that have previously been reported to regulate Il7r transcription. Among these were: glucocorticoid receptor, Gfi1, its close homolog Gfi1b, GABP␣ and its partners GABP␤1 and GABP␤2, PU.1 (Sfpi1), Runx1/3, NF-B, FoxOA1/3, FoxP1, and FoxP3. Notably, we found that among these transcription factors, Gfi1 was the only one that passed our differentially expressed gene criteria of fold-change Ն 2.5 and analysis of variance p value Յ 0.01. Thus, Dex treatment results in an up-regulation of Il7r transcription and a dramatic down-regulation of the zinc finger repressor protein Gfi1 mRNA expression ( Fig. 1B and supplemental Fig. S1).
Gfi1 was previously proposed to repress IL-7R␣ transcription during lymphocyte development (14). Consequently, we wished to assess whether Gfi1 down-regulation would contribute to IL-7R␣ up-regulation in 3B4.15 cells. To test this idea, we retrovirally overexpressed Gfi1 in 3B4.15 cells. Strikingly, Gfi1 overexpression inhibited IL-7R␣ up-regulation by dexamethasone, and Gfi1 overexpressing 3B4.15 cells remained IL-7R␣ low (Fig. 1C). Thus, Dex induces the down-regulation of endogenous Gfi1 expression, but retroviral overexpression of Gfi1 is maintained even in the presence of Dex, and results in these cells remaining IL-7R␣ low. This effect was indeed directly dependent on GR, because Dex treatment in the presence of RU-486, which is a competitive inhibitor of Dex for GR binding, completely inhibited IL-7R␣ up-regulation (Fig. 1D).
Next, to understand the mechanism of Gfi1-mediated repression of IL-7R␣ expression, we retrovirally overexpressed a series of truncated Gfi1 cDNAs and assessed their effects on the Dex response of 3B4.15 cells (supplemental Fig. S2, A and B). Although expression of full-length Gfi1 significantly suppressed IL-7R␣ re-expression, truncated Gfi1 lacking the demethylase-recruiting Snail-Gfi (SNAG) domain or the DNA binding zinc finger (ZF) domain failed to do so (Fig. 1E). Thus, repression of IL-7R␣ expression by Gfi1 requires both its transcriptional repressor and the DNA binding domains and uncover Gfi1 to be a potent transcriptional inhibitor of IL-7R␣ expression that acts downstream of Dex signaling, which de-represses Il7r transcription.
Assessing IL-7R␣ Transcription in Vivo Using a Novel IL-7R␣ Reporter Mouse-Untreated 3B4.15 cells express high levels of Gfi1 and low levels of IL-7R␣. Mature resting T cells, on the other hand, express only low levels of Gfi1 and high levels of IL-7R␣ (38). Therefore, Gfi1 levels in T cells in vivo do not correspond to those in 3B4.15 T hybridomas. To test whether Gfi1 can also control IL-7R␣ expression in vivo, we generated a novel IL-7R␣ transcriptional reporter transgenic mouse. We used a 210-kb BAC fragment containing the Il7r gene locus and inserted an IRES-EGFP cassette in the 3Ј untranslated region (UTR) of the Il7r gene. We used this construct in pronuclear injections to generate transgenic mice (7RIG-BAC Tg ) (supplemental Fig. S3).
Because the BAC construct contained the full Il7r gene locus, including putative transcriptional control regions, we expected the transgenic Il7r gene to faithfully reproduce expression of endogenous Il7r. To test this, we assessed IL-7R␣ surface levels on freshly isolated 7RIG-BAC Tg thymocytes. Endogenous IL-7R␣ displays a characteristic on-off-on pattern during the progression through the DN-DP-SP stages of thymocyte development (3,39). Indeed, 7RIG-BAC Tg thymocytes showed the expected on-off-on pattern for IL-7R␣ expression in the thymus. These data suggest that transgenic IL-7R␣ transcription is appropriately controlled in 7RIG-BAC Tg thymocytes ( Fig. 2A). Notably, IL-7R␣ expression by 7RIG-BAC Tg differed from that of a human CD2 promoter/enhancer-driven IL-7R␣ transgene (IL-7R␣ Tg ), in its ability to down-regulate IL-7R␣ expression on

Gfi1 Controls IL-7R Transcription
DP thymocytes ( Fig. 2A). These results support the expectation that the BAC transgene retained all endogenous regulatory elements for correct IL-7R␣ expression. Notably, IL-7R␣ surface levels on 7RIG-BAC Tg transgenic cells were significantly higher compared with WT controls, presumably because Il7r gene transcription was active both from the endogenous and the transgenic Il7r locus. To demonstrate that the 7RIG-BAC Tg transgene faithfully reports Il7r gene transcription, we assessed GFP expression in thymocytes. GFP expression also followed the on-off-on pattern of IL-7R␣ expression in DN, DP, and SP thymocytes, respectively, indicating that transcription from the transgenic locus was correctly inhibited in DP thymocytes (Fig.  2B). We also assessed IL-7R␣ reporter expression in peripheral LN cells. Mature B cells do not express IL-7R␣. Accordingly, we found that 7RIG-BAC Tg DN lymph node cells, which include all mature B cells, were negative for GFP expression (Fig. 2C). CD4 and CD8 lymph node T cells, on the other hand, correctly expressed high levels of GFP. Interestingly, CD8 T cells reported higher levels of Il7r transcription than CD4 T cells based on their GFP levels (Fig. 2C). These data confirm that 7RIG-BAC Tg reporter mice faithfully represent endogenous IL-7R␣ expression in vivo and reveal a hitherto unappreciated difference in IL-7R␣ transcription levels in CD4 and CD8 T cells.
Cytokine-induced Regulation of 7RIG-BAC Tg Expression-To further document distinct IL-7R␣ expression in CD4 and CD8 T cells, we quantified surface IL-7R␣ and intracellular GFP levels in freshly isolated 7RIG-BAC Tg LN T cells. We confirmed statistically significant higher levels of both IL-7R␣ expression and transcription in CD8 T cells in multiple experiments (Fig. 3, A and B). Moreover, such lineage-specific IL-7R␣ expression was developmentally set as CD4 and CD8 T cells incubated overnight in medium in the absence of in vivo signals still displayed distinct levels of IL-7R␣ expression (data not shown).
The Il7r gene locus is exquisitely sensitive to cytokine signaling. For instance, in vivo IL-7 signaling down-regulates Il7r transcription and steady-state levels of IL-7R␣ mRNA in T cells (13). To determine whether the transgenic IL-7R␣ gene locus also responds to cytokine treatment, we incubated LN T cells from either WT or 7RIG-BAC Tg mice overnight in medium or in the presence of IL-7. The next day, total RNA was extracted and IL-7R␣ mRNA signals were assessed and compared with those from freshly isolated T cells. In both WT and 7RIG-BAC Tg T cells, overnight release from the cytokine-rich in vivo environment highly up-regulated IL-7R␣ mRNA expression (Fig. 3C). Furthermore, overnight IL-7 signaling potently suppressed IL-7R␣ mRNA expression in both WT and transgenic T cells. These results suggest that both the endogenous and the transgenic Il7r loci are regulated in a cytokine-dependent manner. More importantly, GFP mRNA levels also faithfully replicated IL-7R␣ expression, which confirms the validity of this transgenic model as an in vivo IL-7R␣ reporter (Fig. 3).
Restoring T Cell Development in IL-7R␣-deficient Mice by 7RIG-BAC Tg -IL-7R␣ deficiency results in severely impaired T cell development and peripheral T cell homeostasis. Because 7RIG-BAC Tg replicates expression of endogenous Il7r, we wished to know if the BAC transgene could restore T cell  defects in IL-7R␣ KO mice. To test this, we crossed 7RIG-BAC Tg into IL-7R␣ KO mice to generate IL-7R␣ KO 7RIG-BAC Tg mice (supplemental Fig. S4A). In these mice, thymic ␣␤ T cell development was largely restored as demonstrated by thymocyte CD4 versus CD8 profiles and TCR␤ surface expression (Fig.  4A). Also, thymic NKT cell development and ␥␦ T cell generation was dramatically improved compared with IL-7R␣ KO mice (supplemental Fig. S4, B and C). Notably, total thymocyte numbers were also restored compared with IL-7R␣ KO mice, but they did not fully recover to WT levels (Fig. 4B). To further under-stand this, we analyzed surface IL-7R␣ levels on thymocytes, and we found that IL-7R␣ KO 7RIG-BAC Tg mice expressed significantly higher levels of IL-7R␣ than WT mice (Fig. 4C). This is presumably caused by insertion of multiple copies of the BAC transgene into the genome as usually observed in BAC transgenesis. Elevated levels of IL-7R␣, however, have been shown to increase IL-7 consumption and competition for IL-7, which results in an overall decrease in thymocyte numbers (40,41). Thus, decreased total thymocyte numbers might reflect quantitative differences in surface IL-7R␣ expression between WT

Gfi1 Controls IL-7R Transcription
and IL-7R␣ KO 7RIG-BAC Tg thymocytes. Interestingly, however, no major differences were observed in peripheral LN T cell numbers and CD4/CD8 profiles (Fig. 4D), despite surface IL-7R␣ levels on transgenic LN T cells being still higher than WT counterparts (Fig. 4E). Thus, thymocytes and LN T cells are differently affected by IL-7R␣ levels, likely because of the different mechanisms of proliferation and homeostasis in these two organs. Collectively, we find that BAC transgenic IL-7R␣ is expressed and regulated in a developmentally correct fashion, and consequently restores T cell development and maintenance in IL-7R␣-deficient mice.
Gfi1 Overexpression Suppresses Il7r Gene Transcription in CD8 T Cells-Using these reporter mice, next we asked whether Gfi1 can suppress IL-7R␣ transcription and expression in vivo. To this end, we introduced the 7RIG-BAC Tg onto mice expressing a T lineage-specific, human CD2 promoter driven Gfi1 transgene (25) to generate Gfi1 Tg 7RIG-BAC Tg mice and determined their IL-7R␣ transcription by analyzing GFP expression. Strikingly, in Gfi1 Tg 7RIG-BAC Tg thymocytes, we found that GFP levels were significantly down-regulated by Gfi1, but that this was specific and restricted to CD8SP thymocytes (Fig. 5A). These results suggest that transgenic Gfi1 expression fails to suppress IL-7R␣ transcription in any other thymocyte subpopulation than in CD8SP cells. Such a specific effect of Gfi1 on CD8 lineage cells was further confirmed in peripheral CD8 T cells. We found both surface IL-7R␣ and intracellular GFP levels selectively down-regulated in CD8, but not in CD4 T cells (Fig. 5B). Thus, these data document that Gfi1 only down-regulates IL-7R␣ gene transcription and surface protein expression in CD8 T lymphocytes, and that all other thymocyte subpopulations and CD4 T cells are not susceptible to repression of this gene by Gfi1.
To further assess the specificity of Gfi1 transcriptional repression, we examined expression of other members of the ␥c cytokine receptor family. The ␥c cytokine family is composed of IL-2, IL-4, IL-7, IL-9, IL-15, and IL-21 (42), and antibodies are available for surface staining of each member of the ␥c receptor family with the exception of IL-9. Consequently, we examined expression of ␥c cytokine receptors for IL-2R␣, -2R␤, ␥c, IL-4R␣, IL-7R␣, and IL-21R on WT and Gfi1 Tg CD8 T cells (Fig. 5C). Strikingly, IL-7R␣ was the only cytokine receptor that was significantly affected by Gfi1 overexpression (Fig. 5D), which revealed a highly selective effect of Gfi1 on IL-7R␣ and reaffirmed its specificity for CD8 T cells.
Gfi1 Deficiency Up-regulates Il7r Transcription in CD8 T Cells-Such a CD8 lineage-specific effect of Gfi1 was intriguing. One potential explanation could be that Gfi1 is expressed in all thymocytes and T cells, but that its repressor activity is only limited to CD8 T cells. Consequently, we wished to test whether removal of Gfi1 would up-regulate IL-7R␣ expression in T cells other than in CD8 lineage cells. To this end, we generated Gfi1 KO 7RIG-BAC Tg mice and analyzed the GFP levels in thymocytes and mature LN T cells (supplemental Fig. S5A and Fig.  6A) (24). Notably, Gfi1 deficiency failed to de-repress IL-7R␣ transcription in most thymocytes, with the exception of CD8SP thymocytes (supplemental Fig. S5A). DP thymocytes, which express high levels of Gfi1 and are completely silent for IL-7R␣ transcription, were still negative for GFP expression, when Gfi1 expression was ablated. These results indicate that Gfi1 is not required to suppress IL-7R␣ expression in this particular subset. Rather, the effect of Gfi1 was highly restricted to post-selection CD8 lineage cells as GFP expression was quantitatively increased in both CD8SP thymocytes and LN CD8 T cells only (supplemental Fig. S5A and Fig. 6B).
Because absent Gfi1 was sufficient to up-regulate IL-7R␣ expression on CD8 T cells, next, we wished to know whether this would correlate with the Dex effect that we observed in 3B4.15 cells (Fig. 1, A and B). To this end, we stimulated WT and Gfi1 Tg T cells with Dex and assessed their surface IL-7R␣ expression after overnight culture (Fig. 6C). Although WT CD8 ϩ T cells significantly up-regulated IL-7R␣ as previously observed (17), IL-7R␣ levels on Gfi1 KO CD8 ϩ T cells were unaffected by stimulation with Dex (Fig. 6C). Notably, IL-7R␣ expression on Gfi1 KO CD4 ϩ T cells was still up-regulated by Dex indicating that Gfi1 effect is CD8 lineage specific (supplemental Fig. S5B). These data strongly suggest that a major role for Dex is to suppress Gfi1, and that Gfi1 directly controls IL-7R␣ expression in CD8 T cells.
Gfi1 deficiency has been proposed to promote CD8 memory phenotype cell generation (43). Because memory CD8 T cells express high levels of IL-7R (42), we wished to test whether increased IL-7R␣ is a consequence of memory cell differentiation or a direct effect of Gfi1 deficiency. GFP levels in CD8 T cell subsets demonstrated that IL-7R␣ transcription was significantly increased in both naive (CD44 low ) and activated/memory (CD44 high ) phenotype CD8 T cells (Fig. 6D). Thus, we conclude that absent Gfi1 expression de-represses IL-7R␣ transcription and expression in all CD8 T cells independently of their differentiation status.
In Vivo Analysis of IL-7R␣ Transcription Using 7RIG-BACTg on a Single Cell Basis-We wished to know if using the 7RIG-BAC Tg could provide us with new insights on IL-7R␣ expression that so far has not been experimentally feasible to assess. Adoptive transfer of CD8 T cells into chronic lymphopenic mice results in lymphopenia-induced homeostatic proliferation (LIP). Slowly dividing homeostatic proliferation is dependent on IL-7, whereas rapid proliferation is IL-7 independent and driven by commensal antigens (44). To further understand the role of IL-7 in this process, it would be important to assess how IL-7R␣ expression is regulated during LIP. To do so, we assessed surface IL-7R␣ and intracellular GFP expression on day 5 adoptively transferred 7RIG-BAC Tg CD8 T cells. Analyzing IL-7R␣ transcription in adoptively transferred cells on a single cell basis has not been possible so far. IL-7 signaling suppresses IL-7R␣ expression under steady-state conditions (13). Surprisingly, however, surface IL-7R␣ levels remained largely unchanged during IL-7-driven LIP as assessed on slowly dividing cells (Fig. 7, A, left, and B). Strikingly, in the same cells, Il7r transcription was dramatically down-regulated upon cell division, as demonstrated by reduced GFP expression in cell tracediluted cells (Fig. 7, A, right, and B). These data indicate that surface IL-7R␣ expression is not a reliable marker for Il7r transcription, at least during homeostatic proliferation. They further suggest the operation of a transcription-independent mechanism of surface IL-7R expression, which could be either increased recycling of endocytosed IL-7R␣, stabilization of pre-Gfi1 Controls IL-7R Transcription OCTOBER 5, 2012 • VOLUME 287 • NUMBER 41 existing IL-7R␣ proteins or a yet unknown post-transcriptional mechanism. We are currently in the process of addressing these possibilities.
Because Gfi1 suppresses Il7r transcription, next we wished to assess the effect of Gfi1 on IL-7R␣ expression during LIP. Day 5 adoptively transferred Gfi1 Tg 7RIG-BAC Tg cells displayed a comparable pattern of surface IL-7R␣ and intracellular GFP expression to WT cells, in that GFP levels steadily decreased upon proliferation (Fig. 7C, right) but surface IL-7R␣ levels remain largely unaffected (Fig. 7C, left). Thus, the dichotomoy of IL-7R␣ transcription and surface protein expression during LIP still remained distinct in Gfi1 Tg 7RIG-BAC Tg cells. Moreover, LIP further down-regulated IL-7R␣ transcription in Gfi1 Tg cells, which is presumably mediated by a mechanism independent of Gfi1. Importantly, Gfi1 overexpression significantly impaired CD8 T cell LIP, which correlated with lower Cell surface IL-7R␣ and GFP expression were determined by flow cytometry. Bar graph shows mean Ϯ S.E. from three independent experiments with 3 WT and 5 Gfi1 Tg 7RIG-BAC Tg mice. C, expression of ␥c cytokine receptor families on Gfi1 Tg CD8 ϩ LNT cells. Surface expression of the indicated ␥c cytokine receptors were determined on gated CD8 T cells from WT and Gfi1 Tg mice. Data are representative of three independent experiments. D, Gfi1 Tg specifically suppresses IL-7R␣ expression on CD8 ϩ T cells. Surface cytokine receptor levels on Gfi1 Tg CD8 ϩ T cells were quantified and normalized to levels on WT CD8 ϩ T cells. Bar graph shows mean Ϯ S.E. of three independent experiments.  OCTOBER 5, 2012 • VOLUME 287 • NUMBER 41

JOURNAL OF BIOLOGICAL CHEMISTRY 34395
IL-7R␣ levels on Gfi1 Tg CD8 T cells (Fig. 5B). Thus, Gfi1 suppresses IL-7R␣ expression and also IL-7-dependent proliferation. Taken together, these results uncover Gfi1 as a critical regulator of IL-7R␣ transcription and expression in vivo, but exclusively in CD8 lineage thymocytes and T cells.

DISCUSSION
To understand the molecular mechanisms of IL-7R␣ expression during T cell development and activation, here we generated a novel BAC GFP-reporter transgene and utilized this tool to assess IL-7R␣ transcription in vivo. The BAC transgene was constructed by inserting a GFP-reporter and an IRES element into the 3Ј UTR of the murine Il7r gene. Consequently, BAC reporter mice overexpressed full-length IL-7R␣ proteins in addition to GFP. We affirmed the developmentally correct expression of reporter transgenes by assessing lymphocyte differentiation in IL-7R␣-deficient mice reconstituted with the BAC construct (IL-7R␣ KO 7RIG-BAC Tg ). In these mice, development of all IL-7R␣-dependent lymphoid cell populations, including ␣␤-, ␥␦-T cells, B-cells, and NKT cells, were restored. Thus, the transgenic Il7r gene locus in 7RIG-BAC Tg mice is correctly and lineage specifically regulated, and it equipped us with a new tool to assess IL-7R␣ transcription and expression in vivo.
IL-7R␣ is the ligand-specific subunit of the functional IL-7 receptor, which is composed of the IL-7R␣ chain and the ␥c-chain (4). In contrast to the ␥c-chain, IL-7R␣ expression is dynamic and is actively regulated during T cell development and differentiation. All developing thymocytes and all mature T cells express IL-7R␣, albeit at varying degrees. However, immature DP thymocytes are unique in that they have completely terminated IL-7R␣ expression. Such peculiar absence of IL-7R␣ on pre-selection DP cells was proposed to reflect a critical thymic selection mechanism that ensures a random but self-MHC-specific TCR repertoire (39,45,46). Accordingly, absent IL-7R␣ expression renders DP thymocytes dependent on selecting TCR signals and not on nondiscriminatory IL-7 signaling for survival. However, the molecular mechanisms that terminate IL-7R␣ expression in DP cells remain unclear. Along this line, the molecular circuitry that re-induces IL-7R␣ expression upon positive selection also remains unmapped. This is even more intriguing as TCR signaling in mature T cells downregulates IL-7R␣ expression, but in immature DP thymocytes, TCR signaling induces IL-7R␣ expression. Recent studies have suggested that re-expression of IL-7R␣ on DP cells is dependent on positive selecting TCR signals in a NFAT-and MAPKdependent fashion and to a lesser degree on Akt (47). However, the detailed mechanism and molecular events remain unknown. Using 7RIG-BAC Tg mice, here we establish that absent IL-7R␣ expression in DP cells is a transcriptionally regulated event. Pre-selected DP cells failed to express GFP, while positive selection and maturation resulted in re-expression of GFP. With the 7RIG-BAC Tg reporter mouse, it is now feasible

Gfi1 Controls IL-7R Transcription
to test nuclear factors that regulate IL-7R␣ transcription in vivo. We anticipate further applications of these reporter mice in dissecting the thymic signals that lead to termination as well as re-induction of Il7r transcription under various in vitro and in vivo settings.
In this regard, we were able to assess the effect of a transcriptional repressor on Il7r gene expression in vivo. Gfi1 has been considered as a potential Il7r regulator in immature thymocytes because of its high level of expression in pre-selection DP thymocytes and its potent suppression of IL-7R␣ in pre-B cells. As such, Gfi1 KO 7RIG-BAC Tg mice offered a unique opportunity to assess the effect of Gfi1 on IL-7R␣ transcription in DP thymocytes in vivo. The complete absence of GFP signals in Gfi1 KO 7RIG-BAC Tg DP thymocytes strongly suggested that transcriptional silencing of Il7r gene expression does not require Gfi1, at least in pre-selection DP thymocytes. In mature T cells, Gfi1 clearly plays a more critical role. In fact, Gfi1 was first reported to suppress IL-7R␣ expression by analyzing peripheral mature T cells (13). Nevertheless, a definitive and direct role of Gfi1 in IL-7R␣ transcription has remained highly controversial. In some cases, surface IL-7R␣ expression on Gfi1 KO T cells was found to be not, or insignificantly different (38), and also increased IL-7R␣ levels in Gfi1 KO mice had been suggested to be indirect, i.e. due to high proportions of memory cells under Gfi1 deficiency (43). Moreover, increased Gfi1 mRNA levels in human IL-7R␣ high effector memory CD8 T cells contradicted a repressor role for Gfi1 in IL-7R␣ expression (48). Our current data showing significantly increased GFP expression in Gfi KO CD8 T cells, however, clearly documents Gfi1 as an in vivo repressor for IL-7R␣ transcription.
Nevertheless, contradictory observations still keep open the possibility of a more complex regulatory network, with several Gfi1 co-factors being involved. As such, the cis-elements participating in the Gfi1 control of Il7r gene transcription are still not well defined. In B lymphocytes, Gfi1 can bind to the second intron of the Il7r gene to suppress transcription (14). In myeloid cells, Gfi1 can also associate with PU.1, an ETS family transcription factor that binds to the promoter of IL-7R␣ and activates its transcription (49). However, PU.1 is not expressed in T lymphocytes. Rather another ETS family transcription factor, GABP, is thought to occupy the ETS site in the promoter of the Il7r gene (50). Whether Gfi1 can interact with the GABP protein is not known. But in T cells, GABP and Gfi1 were shown to have opposing roles so that a direct interaction and mutual suppression cannot be excluded (51). Thus, in addition to its role in binding the Il7r gene promoter, GABP may control Gfi1 expression as GABP-deficient splenocytes lack Gfi1 expression, presumably due to the positively acting GABP binding sites in the Gfi1 promoter (52).
With Gfi1 emerging as a key control factor in IL-7R␣ expression and also in many aspects of T and B cell immunology, it is critical to know what controls Gfi1 expression. So far, an autoregulatory role for Gfi1 as well as a transregulatory role for Gfi1b has been documented (53). Also, downstream signaling of the GTPase Cdc42 has been proposed to suppress Gfi1 expression in resting T cells (54). The current study now proposes that glucocorticoids are a new family of molecules that can control Gfi1 expression. Glucocorticoids have been known to take part in thymocyte development and selection (55)(56)(57)(58). Recently, conditional deletion of GR in pre-selection thymocytes documented significantly reduced thymus cellularity with an altered TCR repertoire and increased negative selection (59). Mechanistically, glucocorticoids have been shown to intersect with TCR signaling and increase TCR signaling thresholds to promote thymic selection (60). Our current data now also proposes a role for Gfi1 downstream of GR signaling, and it would be informative to assess the contribution of Gfi1 to the GR-deficient phenotype.
In peripheral T cells, glucocorticoids are immunosuppressive, but they also up-regulate IL-7R␣ transcription and expression. The identification of Gfi1 as a novel target of Dex sheds new light onto pre-existing observations on the effects of glucocorticoids and in this regard, the discovery of Gfi1 as an intermediary of Dex in IL-7R␣ expression suggests that other Dexinduced events might also employ such a mechanism. Although the up-regulation of IL-7R␣ expression in Dex-treated B cells, which normally do not express IL-7R␣, could be another case of an Gfi1-mediated Dex effect (34), inhibition of cytokine production in Dex-treated cells also could be explained along this line. Further studies are planned to address these possibilities and to delineate a direct Dex effect versus Gfi1-mediated Dex effects.
Although Gfi1 expression is down-regulated by Dex signaling, and overexpression of Gfi1 can overcome Dex-mediated up-regulation of IL-7R␣, we found, as expected, that GR is critical for Dex-mediated IL-7R␣ up-regulation. GR binds to a putative Il7r enhancer in an evolutionarily conserved sequence 3.5 kb upstream of the transcriptional start site (16). The GR site in this evolutionarily conserved sequence is 50 base pairs away from a FoxO1 transcription factor binding site, which positively regulates gene expression in both naive CD4 ϩ and CD8 ϩ T cells (20,21). Whether this enhancer, the PU.1/GABP binding promoter, and the putative Gfi1-binding intronic silencer are the only functional control elements in the Il7r gene locus is not known. In fact, a series of nuclear factors are known to regulate IL-7R␣ expression. In addition to Gfi1, two other transcriptional repressors have been identified that down-regulate Il7r gene expression in T cells. The forkhead family transcription factor FoxP1 suppresses IL-7R␣ expression by competing with the Il7r transactivator FoxO1 for enhancer occupancy (19). On the other hand, FoxP3, which is specifically expressed in regulatory T cells, directly suppress IL-7R␣ expression (22). Overexpression of a Gfi1 family member, Gfi1b, can also repress IL-7R␣ expression in T lymphocytes, but presumably this effect is through the same intronic Gfi1 binding site, because the DNA binding domains of these two factors are very similar (61). Notably, in our expression profiling experiments, Gfi1 was the only nuclear factor that showed a significant difference in expression upon Dex treatment.
Collectively, the present study identified and tested a novel Il7r transcriptional control mechanism in vivo using a newly established IL-7R␣ reporter mouse. We validated these reporter mice by complementing IL-7R␣ deficiency, and we utilized this tool to assess Il7r transcription downstream of Gfi1. Using a lymphopenia-induced homeostatic proliferation model, we further documented the superiority of these reporter mice in assessing IL-7R␣ gene transcription in T cells to traditional methods. Finally, the current study not only resolves the controversies surrounding the role of Gfi1 as a transcriptional repressor of IL-7R␣ in CD8 T cells but also demonstrates that IL-7R␣ expression in any other T cell population is independent of Gfi1. The molecular basis for such lineage-specific regulation of cytokine receptor expression is intriguing and important, and we think that our findings will provide a new venue to identify critical players for controlling IL-7R␣ expression.