An Enhancer of the IL-7 Receptor α-Chain Locus Controls IL-7 Receptor Expression and Maintenance of Peripheral T Cells

Interleukin-7, a cytokine essential for lymphocyte development and homeostasis, controls the survival, proliferation, and differentiation of early T and B cells (1–3) as well as V(D)J recombination of the TCRγ and IgH loci (4, 5). In the periphery, IL-7 supports survival and homeostasis of naive and memory T cells (6). IL-7 exerts its effects through interaction with the IL-7R, which consists of the IL-7R α-chain (IL-7Rα) and a common cytokine receptor γ-chain. IL-7Rα also dimerizes with the thymic stromal lymphopoietin receptor. IL-7 binding to the IL-7R activates JAK-1 and JAK-3, which then activates STAT5 and PI3K.

IL-7Rα expression is dynamically controlled in T and B cell development. During T cell differentiation, IL-7Rα is expressed in double-negative (DN) thymocytes, downregulated in double-positive (DP) thymocytes, upregulated in single-positive (SP) thymocytes, and maintained in naive T cells in the periphery (2, 7). When naive T cells receive Ag stimuli, the IL-7Rα is downregulated in effector T cells (6, 8, 9). However, a small subset of effector T cells, referred to as memory precursor effector cells, upregulates IL-7Rα and differentiates into memory CD8 T cells (9). During B cell development, IL-7Rα is expressed on common lymphoid progenitors (CLPs) and pro– and pre–B cells and then is downregulated during the transition from pre–B to B cells (7, 10, 11). Thus, IL-7Rα expression is precisely regulated during lymphocyte development and the immune response.

IL-7Rα expression on T cells is regulated at transcriptional levels and suppressed by IL-7 and other prosurvival cytokines, such as IL-2, IL-4, IL-6, and IL-15 (12). Additionally, TCR signaling also represses IL-7Rα transcription (13). Furthermore, both glucocorticoids (GC) and TNF-α upregulate IL-7Rα transcription (14–16). Additionally, several transcriptional factors interact with the IL-7Rα locus and control its transcription. Two Ets family proteins, PU.1 and GA binding protein α, interact with the IL-7Rα promoter and are crucial for IL-7Rα transcription in early B and T cells, respectively (17, 18). We previously showed that a conserved non-coding sequence 1 (CNS1) element found 3.6 kb upstream of the IL-7Rα promoter is a GC-responsive enhancer of the IL-7Rα locus in vitro (16). Consistently, a DNase I hypersensitivity site was identified 3.8 kb upstream of that promoter (19). Consensus motifs for NF-κB, the GC receptor (GR), Evi-1, and forkhead transcription factors are conserved in the CNS1 element, and Foxo1 and Foxp1 interact with the element to control IL-7Rα expression in naive T and early B cells (20–22). However, it is largely unknown how the element controls IL-7Rα expression during lymphocyte development and the immune response.

To determine its function in vivo, we established a mouse line carrying targeted deletion of the CNS1 element and analyzed IL-7Rα expression at different differentiation stages. IL-7Rα expression was significantly reduced in peripheral T cells of CNS1^−/− mice, and naive CNS1^−/− T cells showed impaired survival and homeostatic proliferation. IL-7Rα upregulation by GC and TNF-α was also impaired in CNS1^−/− T cells. We also report that IL-7Rα upregulation after Ag stimulation was independent of the CNS1 element. Overall, this study demonstrates that the CNS1 element controls IL-7Rα expression and maintenance of peripheral T cells.

B6.CD45.1 congenic, B6.OT-1 TCR transgenic (23), and Rag2^−/− mice were used. All mice were maintained under specific pathogen-free conditions in the Experimental Research Center for Infectious Diseases in the Institute for Virus Research, Kyoto University. All mouse protocols were approved by Kyoto University.

To construct a targeting vector, the following DNA fragments were assembled in the pBluescript KS (+) vector: a diphtheria toxin A cassette, a 1822-bp fragment of the 5′ CNS1 element, a neomycin resistance gene cassette flanked by loxP sites, and a 7683-bp fragment of the 3′ CNS1 element. That linearized vector was introduced into the KY1.1 embryonic stem (ES) cells (C57BL/6 × 129S6/SvEvTac F₁ background) by electroporation, and homologous recombinants were screened by PCR. Targeted ES clones were confirmed by Southern blot analysis with 5′ and 3′ probes (see Fig. 1A, 1B). The neomycin resistance gene cassette was removed from the recombinant allele by infecting targeted ES clones in vitro with adenovirus expressing Cre recombinase (a gift of Dr. Izumu Saito of the Institute of Medical Science, University of Tokyo) (24). Resultant ES clones contained one copy of the loxP site with an 839-bp deletion of the CNS1 element. ES clones were injected into ICR eight-cell embryos. Chimeric mice were bred with C57BL/6 mice, and CNS1-deficient mice were backcrossed onto C57BL/6 mice for eight generations.

To sort thymocyte subpopulations, triple-negative and CD8SP thymocytes were first enriched by negative separation with anti-CD4 microbeads and LS columns (Miltenyi Biotec, Auburn, CA). DN (CD3⁻CD4⁻CD8⁻), DP (CD4⁺CD8⁺), CD4SP (CD3^hiCD4⁺CD8⁻), and CD8SP (CD3^hiCD4⁻CD8⁺) thymocytes were sorted with a FACSAria II cell sorter (Becton Dickinson, Franklin Lakes, NJ) (purity >98%). CD4 and CD8 T cells were purified from lymph nodes using an EasySep CD4⁺ and CD8⁺ T cell enrichment kits (Stemcell Technologies, Vancouver, BC, Canada; purity >98%). Naive CD4 (CD3⁺CD4⁺CD25⁻CD44^loCD62L^hi or CD3⁺CD4⁺CD25⁻CD44^loNK1.1⁻) T cells, naive CD8 (CD3⁺CD8⁺CD44^loCD62L^hi) T cells, and regulatory T (Treg) cells were sorted from lymph node cells or enriched CD4 T cells (purity >98, >97, and >95%, respectively). In some experiments, enriched CD8 T cells were stained with PE–anti-CD44, followed by incubation with anti-PE microbeads (Miltenyi Biotec), and negatively separated on MS columns to purify naive CD8 T cells (CD3⁺CD8⁺CD44^lo) (purity >98%). γδ T cells (CD3⁺γδTCR⁺) were sorted from lymph node cells (purity >92%). Intraepithelial lymphocytes (IELs) were prepared from small intestine as described (25). Intrahepatic lymphocytes (IHLs) were prepared as described with minor modification (26). Briefly, mice were perfused with PBS. IHLs were obtained by dispersing the liver through a strainer, followed by centrifugation through a 33% Percoll gradient and erythrocyte lysis. NKT (CD3⁺NK1.1⁺) and NK (CD3^–NK1.1⁺) cells were sorted from IHLs (purity >98%).

The following fluorescent dye– or biotin-conjugated Abs were purchased from BD Biosciences (San Jose, CA), eBioscience (San Diego, CA), BioLegend (San Diego, CA), and Wako Pure Chemical Industries: CD3ε (145-2C11), TCRβ (H57-597), CD4 (RM4.5), CD8α (53-6.7), CD8β (H35-17.2), CD25 (7D4), Foxp3 (FJK-165), CD44 (IM7), γδTCR (GL3), NK1.1 (PK136), CD45R/B220 (RA3-6B2), CD43 (S7), BP-1 (6C3), CD24 (30F1), Igμ (M41), CD11c (N418), CD11b (M1/70), Gr-1 (RB6-8C5), TER-119, Sca-1 (E13-161.7), c-Kit (2B8), Thy-1 (53-2.1), IL-7Rα (A7R34), CD45.1 (A20), CD45.2 (104), Ly-6C (HK1.4), IFN-γ (XMG1.2), phosphorylated STAT5 (Tyr⁶⁹⁴) (47), Bcl-2 (A19-3), annexin V, Ki-67 (SolA15). H-2K^b OVA G4 tetramer–SIIGFEKL-PE was purchased from Medical and Biological Laboratories (Nagoya, Japan). Biotinylated mAbs were detected with PE-, allophycocyanin-, or PE-Cy7–conjugated streptavidin (eBioscience). Viable cells were analyzed with FACSCalibur or FACSCanto II flow cytometers (BD Biosciences) with CellQuest and FlowJo software. Values in quadrants, the gated area, and interval gates indicate the percentages in each population in all figures.

Intracellular staining of phosphorylated STAT5 was performed as described (27). In brief, sorted naive CD4 (CD3⁺CD4⁺CD25⁻CD44^loCD62L^hi) and CD8 (CD3⁺CD8⁺CD44^loCD62L^hi) T cells were stimulated for 20 min with 5 ng/ml IL-7 (R&D Systems, Minneapolis, MN) at 37°C, stained for surface Ags, fixed, permeabilized in ice-cold methanol, and stained with anti-phosphorylated STAT5 Ab using a Foxp3 staining buffer set (eBioscience). For intracellular staining of Foxp3, Bcl-2, and Ki-67, cells were first stained for surface Ags, fixed, permeabilized, and stained with relevant Abs using a Foxp3 staining buffer set according to the manufacturer’s instructions. Intracellular staining of IFN-γ in OT-1 T cells was carried out using the same buffer set. In brief, lymphocytes obtained from spleen were incubated with 1 μg/ml OVA_257–264 peptide (Medical and Biological Laboratories), 2 μg/ml anti-CD28 (PV-1), and 10 μg/ml brefeldin A (Biomol, Hamburg, Germany) for 4 h at 37°C. Cells were stained for surface Ags, fixed, and permeabilized with fixation/permeabilization solution (eBioscience) for 30 min at 4°C and stained with FITC–anti-IFN-γ Ab for 30 min at 4°C.

T cells were cultured in RPMI 1640 medium supplemented with 10% FBS and 50 μM 2-ME. The following reagents were used at the indicated concentrations: 0.5 and 5 μg/ml anti-CD3ε (2C11), 2 μg/ml anti-CD28 (PV-1), 5 ng/ml IL-7 (R&D Systems), 10⁻⁹ M dexamethasone (Sigma-Aldrich, St. Louis, MO), and 1.5 ng/ml TNF-α. For T cell stimulation, purified CD4 T cells (1 × 10⁶ cells/ml) were stimulated with IL-7 or plate-bound anti-CD3 in the presence of fixed doses of soluble anti-CD28 for 20 h, or dexamethasone or TNF-α for 12 h at 37°C. For an in vitro survival assay, sorted naive CD4 (CD3⁺CD4⁺CD25⁻CD44^loCD62L^hi) and CD8 (CD3⁺CD8⁺CD44^loCD62L^hi) T cells (2 × 10⁵ cells/ml) were cultured with and without IL-7, and the proportion of propidium iodide⁻ living cells was measured by flow cytometry. For in vitro apoptosis assay, purified T cells (1 × 10⁶ cells/ml) were cultured with and without IL-7 for 24 h, and the proportion of annexin V⁺ dead cells was measured by flow cytometry.

Total RNA was isolated and treated with RNase-free DNase to remove genomic DNA. After DNase inactivation, total RNA was reverse transcribed using random primers. Complementary DNAs were analyzed by real-time RT-PCR using SYBR Green PCR Master Mix (Qiagen, Hilden, Germany) in an ABI 7500 real-time PCR system (Applied Biosystems, Foster City, CA). PCR efficiency was normalized using cDNA made from whole thymocytes of wild-type mice. The following primer sets were used: Il7ra, 5′-GGATGGAGACCTAGAAGATG-3′ and 5′-GAGTTAGGCATTTCACTCGT-3′; bcl2, 5′-TCGCTACCGTCGTGACTTC-3′ and 5′-AAACAGAGGTCGCATGCTG-3′; bclxl, 5′-GGAGAGCGTTCAGTGATC-3′ and 5′-CAATGGTGGCTGAAGAGA-3′; mcl1, 5′-TCAAAGATGGCGTAACA-3′ and 5′-CCCGTTTCGTCCTTACAAGAAC-3′; and Gapdh, 5′-CCTCGTCCCGTAGACAAAATG-3′ and 5′-TCTCCACTTTGCCACTGCAA-3′.

Naive CD4 (CD3⁺CD4⁺CD25⁻CD44^loNK1.1⁻) and CD8 (CD3⁺CD8⁺CD44^lo) T cells were labeled with 5 μM CFSE (Dojindo Laboratories, Kumamoto, Japan) in PBS at 37°C for 10 min and washed thoroughly. Cells were resuspended in 200 μl PBS and adoptively transferred into Rag2^−/− mice by i.v. injection. After 7 d, proliferation of transferred cells was assessed by FACS-measured CFSE dilution. CFSE-labeled cells were divided into three populations: CFSE^hi (no division), CFSE^int (one to four divisions), and CFSE^lo (five or more divisions) cells.

Adoptive transfer and immunization of OT-I cells was performed as described (28, 29). In brief, OT-I cells (1.5 × 10⁶ cells) were purified from B6.OT-1 CNS1^+/+Rag2^−/− and B6.OT-1 CNS1^−/−Rag2^−/− mice. OT-I cells from CNS1^+/+ or CNS1^−/− Rag2^−/− mice were suspended in 200 μl PBS and adoptively transferred into 6- to 8-wk-old B6.CD45.1 recipient mice. After 24 h, mice were immunized by i.v. injection of 200 μg OVA (Wako Pure Chemical Industries, Osaka, Japan) and 25 μg LPS (from Esherichia coli 055:B5, Sigma-Aldrich). Transferred OT-I cells were analyzed 3, 7, 15, and 25 d after the primary immunization by flow cytometry.

Recombinant Listeria monocytogenes expressing OVA (rLM-OVA) was previously described (30, 31). For primary infection, mice were injected i.v. with 2 × 10⁴ CFU rLM-OVA at day 0. At days 7, 14, 21, 28, and 35, CD8 T cells were analyzed. To examine memory T cells, the recipient mice were rechallenged with 2 × 10⁵ CFU rLM-OVA at day 35. After 5 d, CD8 T cells were analyzed. Flow cytometry was performed after fixing surface-stained cells with 0.1% paraformaldehyde-PBS (−) for 20 min.

An unpaired two-tailed Student t test was used for all statistical analysis. A p value <0.05 was considered statistically significant.

To investigate the function of the CNS1 element in IL-7R expression in vivo, we established a mutant mouse line with targeted deletion of that element at the IL-7Rα locus (CNS1^−/− mice) (Fig. 1A, 1B). Initially, we used flow cytometry to compare thymocytes between CNS1^+/+ and CNS1^−/− mice. Thymocyte numbers were unchanged between wild-type and mutant mice (Fig. 1C). Additionally, both the proportion and absolute numbers of DN1, DN2, DN3, DN4, DP, CD4SP, and CD8SP thymocytes were unchanged between genotypes (Fig. 2A, 2B). We next analyzed surface IL-7Rα expression and found that each thymocyte population showed comparable levels of IL-7Rα expression in CNS1^+/+ and CNS1^−/− mice (Fig. 2C). To assess regulation at the transcriptional level, we compared Il7ra transcripts in thymocyte subpopulations using real-time RT-PCR and found that transcript levels were also comparable between wild-type and mutant mice (Fig. 2D). These results demonstrate that the CNS1 element is dispensable for IL-7Rα expression during thymocyte development.

We next compared numbers of Treg, γδ T, NKT, and NK cells between CNS1^+/+ and CNS1^−/− thymocytes using flow cytometry. The proportion and absolute numbers of all of these cells were unchanged between CNS1^+/+ and CNS1^−/− mice (Fig. 2E, 2F). Whereas γδ T and NKT cells showed comparable levels of surface IL-7Rα expression between wild-type and mutant mice, Treg cells showed significantly lower levels of IL-7Rα expression in CNS1^−/− compared with wild-type mice (Fig. 2G). These results strongly suggest that CNS1 controls IL-7Rα expression in Treg cells but not in γδ T, NKT, and NK cells in thymus.

Next, we asked whether the CNS1 element functions during B cell development in bone marrow. We observed that the numbers of bone marrow cells were comparable in CNS1^−/− and CNS1^+/+ mice (Fig. 1C). The proportion and absolute numbers of CLPs and cells in Hardy fractions A–F were also unchanged (Fig. 3A, 3B). Furthermore, surface IL-7Rα expression was comparable in each of these B cell fractions in wild-type and mutant mice (Fig. 3C), indicating that the CNS1 element is dispensable for IL-7Rα expression during B cell development in bone marrow.

We previously showed that GR interacts with the CNS1 element and enhances Il7ra transcription in vitro (16). Foxo1 and Foxp1 also reportedly interact with the element to control IL-7Rα expression in naive T cells (20, 21). To determine the function of this element in peripheral CD4 and CD8 T cells, we first analyzed lymph node cells in CNS1^+/+ and CNS1^−/− mice using flow cytometry. The numbers of lymph node cells were slightly decreased in CNS1^−/− compared with CNS1^+/+ mice (Fig. 1C). The proportion and absolute numbers of naive CD4 and CD8 T cells were moderately reduced in CNS1^−/− compared with wild-type mice (Fig. 4A, 4B). Surface IL-7Rα expression in peripheral naive CD4 and CD8 T cells was significantly reduced in mutant versus wild-type mice (Fig. 4C). Il7ra transcripts were decreased by 3.5-fold in CD4 and CD8 T cells in mutant compared with wild-type mice (Fig. 4D), strongly suggesting that the CNS1 element controls IL-7Rα expression in these cell populations.

We next asked whether the CNS1 element was important for IL-7Rα expression in other T cell subpopulations in the periphery. To do so, we first analyzed Treg cells in CNS1^+/+ and CNS1^−/− lymph nodes. The absolute number of Treg (Foxp3⁺CD4⁺) cells was unchanged between genotypes (Fig. 5A, 5B). Additionally, Treg cells displayed significantly reduced levels of IL-7Rα (Fig. 5C). Furthermore, Il7ra transcript levels were decreased by 3.5-fold in Treg cells of CNS1^−/− relative to wild-type mice (Fig. 5D).

Both γδ T and NKT cells reportedly required IL-7 and IL-15 for maintenance in periphery (32, 33). When we analyzed γδ T cells in lymph nodes, we found that their absolute numbers and proportion were slightly reduced in CNS1^−/− compared with CNS1^+/+ mice (Fig. 5A, 5B). Additionally, the proportion of γδ T cells showing low levels of IL-7Rα expression increased in CNS1^−/− mice (Fig. 5C). Furthermore, Il7ra transcript levels were slightly decreased in γδ T cells of CNS1^−/− compared with CNS-1^+/+ mice (Fig. 5D). When we analyzed NKT cells in liver, we found that numbers of IHLs moderately decreased in CNS1^−/− compared with CNS1^+/+ mice (Fig. 1C). Although the proportion of NKT cells was unchanged (Fig. 5A), the number of NKT cells decreased slightly in CNS1^−/− compared with CNS1^+/+ mice (Fig. 5B), and those cells displayed relatively lower levels of surface IL-7Rα expression (Fig. 5C). Furthermore, Il7ra transcripts in NKT cells decreased 2-fold in CNS1^−/− mice (Fig. 5D).

We next analyzed NK cells in liver. Absolute numbers and proportion of NK cells were slightly reduced in CNS1^−/− compared with CNS1^+/+ mice (Fig. 5A, 5B). Additionally, the proportion of IL-7Rα⁺ NK cells was reduced in CNS1^−/− mice (Fig. 5C). Furthermore, Il7ra transcript levels in NK cells decreased by 5-fold in CNS1^−/− mice (Fig. 5D).

Because IL-7 plays a critical role in conventional and plasmacytoid dendritic cell (cDC and pDC) precursors (34), we analyzed DCs in bone marrow and spleen of CNS1^+/+ and CNS1^−/− mice. Numbers of pDCs and cDCs were unchanged between CNS1^+/+ and CNS1^−/− mice (Fig. 5A, 5B) as was cell surface IL-7Rα expression in these cells (Fig. 5C).

Because IL-7 is indispensable for IEL development (35), we analyzed IL-7Rα expression in IELs of CNS1^+/+ and CNS1^−/− from the small intestine. The absolute numbers and proportion of αβ IELs were unchanged in CNS1^−/− compared with CNS1^+/+ mice (Supplemental Fig. 1A, 1B). However, the proportion of CD4⁺ αβ IELs was significantly reduced in CNS1^−/− mice, whereas the proportion of other αβ IELs remained unchanged (Supplemental Fig. 1C). Additionally, surface IL-7Rα expression in CD4⁺ and CD8αβ⁺ αβ IELs was slightly reduced in CNS1^−/− compared with CNS1^+/+ mice (Supplemental Fig. 1D). IL-7Rα expression was significantly decreased in CD8αα⁺ αβ and γδ IELs in mutant mice. However, each IEL subpopulation displayed comparable levels of surface IL-2/15Rβ expression in mutant and wild-type mice (Supplemental Fig. 1D). Overall, these analyses show that the CNS1 element controls IL-7Rα expression primarily in Treg cells, γδ T cells, NKT cells, some NK cells, and some αβ IELs but not in DCs.

Because IL-7 controls survival of T cells, we first analyzed STAT5 phosphorylation and cell survival of peripheral CD4 and CD8 T cells after IL-7 stimulation in vitro. STAT5 phosphorylation levels were moderately reduced in naive CD4 and CD8 T cells of CNS1^−/− compared with wild-type mice (Fig. 6A, 6B). Addition of IL-7 only weakly rescued naive CNS1^−/− CD4 and CD8 T cells from cell death (Fig. 6C). Consistent with the cell survival results, CNS1^−/− CD4 and CD8 T cells showed a 2- to 3-fold increase in the proportion of annexin V⁺ dead cells after culture with IL-7 (Supplemental Fig. 2A). These results suggest that CNS1 element affects IL-7 signaling and survival of naive T cells.

Because IL-7 induces Bcl-2 in T cells, we next analyzed expression of antiapoptotic proteins. Both CD4 and CD8 T cells showed comparable levels of Bcl-2 protein in CNS1^+/+ and CNS1^−/− mice (Supplemental Fig. 2B). Given that IL-7 induces bcl2 transcription in CD4 and CD8 T cells (12), we measured bcl2 transcripts and found that they were comparable between CNS1^+/+ and CNS1^−/− mice (Supplemental Fig. 2C). Because T cell survival also depends on Bcl-x_L and Mcl-1 (36–38), we also measured bclxl and mcl1 transcripts but did not detect significant changes in their levels (Supplemental Fig. 2C).

Because IL-7 is essential for homeostatic proliferation of naive T cells (6), we analyzed expansion of naive T cells transferred into lymphopenic hosts. Naive CD4 and CD8 T cells were purified from CNS1^+/+ and CNS1^−/− mice, labeled with CFSE, and transferred into Rag2^−/− recipients. After 7 d, proliferation of transferred T cells was assessed using a CFSE dilution assay. Naive T cells transferred into lymphopenic hosts are classified into rapidly and slowly dividing populations. Rapidly dividing cells are cross-reactive with environmental Ags undergoing IL-7–independent spontaneous proliferation, whereas slowly dividing cells require IL-7 for homeostatic proliferation (6, 39). As previously reported, transferred T cells split into rapidly and slowly dividing populations by day 7 (Fig. 6D). The proportion of CFSE⁺ cells in the slowly dividing population tended to be low in CNS1^−/− T cells, although that tendency was more prominent in CD4 than in CD8 T cells. Absolute numbers of cells in the CFSE⁺ slowly dividing population were significantly decreased in CD4 and CD8 T cells of CNS1^−/− relative to wild-type mice (Fig. 6E). We further divided slowly dividing cells into three fractions: CFSE^hi (no division), CFSE^int (one to four divisions), and CFSE^lo (five or more divisions). The proportion of CFSE^hi cells was significantly increased in CD4 and CD8 T cells of CNS1^−/− mice (Fig. 6F), indicating that homeostatic proliferation was partially impaired. Consistently, the absolute numbers of CFSE^int and CFSE^lo cells were significantly reduced in CNS1^−/− relative to wild-type T cells (Fig. 6G). These results demonstrate that the CNS1 element controls homeostatic proliferation of naive T cells.

Removal of T cells from their in vivo environment markedly increases IL-7Rα expression, whereas IL-7 and TCR signals downregulate it (12, 13). We next sought to determine how IL-7Rα expression is regulated by the CNS1 element in peripheral T cells. To do so, we first analyzed IL-7Rα upregulation in the absence of growth factors. CD4 T cells cultured without IL-7 showed an ∼3-fold increase in Il7ra transcripts in a time-dependent manner in CNS1^+/+ and CNS1^−/− CD4 T cells, although absolute levels were much lower in CNS1^−/− than in CNS1^+/+ cells (Fig. 7A). We next analyzed IL-7Rα downregulation by IL-7 or anti-CD3ε plus anti-CD28 Abs. Both stimuli significantly reduced levels of Il7ra transcripts in CNS1^+/+ and CNS1^−/− CD4 T cells (Fig. 7B, 7C).

Both GC and TNF-α reportedly enhance IL-7Rα transcription (14–16). Therefore, we analyzed the function of the CNS1 element on IL-7Rα upregulation by dexamethasone and TNF-α. Both stimuli induced Il7ra transcripts by 6- to 8-fold in CNS1^+/+ CD4 T cells compared with the cells grown in medium alone (Fig. 7D). However, this induction was completely lost in CNS1^−/− CD4 T cells. Overall, these results demonstrate that CNS1 is a GC- and TNF-α–responsive enhancer of the IL-7Rα locus in T cells.

To determine the function of CNS1 element in IL-7Rα expression during the immune response, we used an OT-I TCR transgenic mouse model. To do so, we transferred CD8 T cells from CNS1^+/+ or CNS1^−/− OT-I Rag2^−/− mice into CD45.1⁺ recipients and then immunized them with OVA and LPS. After immunization, transferred OT-I cells were assessed by flow cytometry. Absolute numbers and the proportion of CNS1^−/− OT-I cells were slightly reduced compared with CNS1^+/+ OT-I cells before immunization (∼70% of control) (day 0, Fig. 8A, 8B). At the peak of the primary response (day 3), the population of CNS1^+/+ OT-I cells had drastically expanded, whereas the absolute numbers and proportion of CNS1^−/− OT-I cells were significantly reduced compared with controls (∼50% of control). After the peak, the number of CNS1^−/− OT-I cells substantially declined compared with controls during contraction (∼17 and ∼7% of control at days 7 and 15, respectively) and were hardly detectable at day 25.

We next compared surface IL-7Rα expression in Ag-specific CD8 T cells during the immune response. Whereas naive CNS1^+/+ OT-I cells expressed high levels of IL-7Rα at day 0, most effector cells slightly downregulated IL-7Rα at the peak of expansion (day 3) (Fig. 8C). IL-7Rα expression slightly increased over time in CNS1^+/+ OT-I cells at later time points (from days 7 to 25). CNS1^−/− OT-I cells completely lost IL-7Rα expression by day 3, suggesting that the CNS1 element is dispensable for IL-7Rα downregulation by Ag stimulation. After the peak of expansion, CNS1^−/− OT-I cells showed IL-7Rα upregulation (from days 7 to 15). To evaluate the frequency of memory phenotype effector CD8 T cells, we next analyzed Ly-6C expression in OT-I cells (40). Both mutant and wild-type OT-I cells showed comparable frequencies of Ly-6C⁺ cells during the immune response (Fig. 8D). These results demonstrate that although IL-7Rα expression is controlled by the CNS1 element, changes in IL-7Rα expression that occur during the immune response are independent of the CNS1.

Given the relative reduction of CNS1^−/− OT-I cells compared with wild-type OT-I cells, we next compared the function of these cells. To do so, we first evaluated their cytokine production upon restimulation. Three days after immunization, spleen cells from recipients transferred with CNS1^+/+ or CNS1^−/− OT-I cells were restimulated with OVA peptide, and IFN-γ production was assessed by flow cytometry. Both CNS1^+/+ and CNS1^−/− OT-I cells produced IFN-γ at similar levels (Fig. 8E). Next, we compared Bcl-2 protein expression. At the peak of the primary response (day 3), Bcl-2 was expressed at comparable levels between CNS1^+/+ and CNS1^−/− OT-I cells (Fig. 8F). However, afterward (by day 7), Bcl-2 expression significantly declined in CNS1^−/− effector OT-I cells, suggesting that the CNS1 element is crucial for induction of Bcl-2 expression in effector T cells. To evaluate proliferation of Ag-specific CD8 T cells, we next analyzed Ki-67 expression. Whereas Ki-67 expression was comparable between CNS1^+/+ and CNS1^−/− OT-I cells at day 3, the proportion of Ki-67^hi cells significantly increased in CNS1^−/− OT-I cells after the peak of expansion (Fig. 8G). These results suggest that the CNS1 element may control survival of effector T cells.

To determine the function of CNS1 element during the immune response after infection and to exclude any possibility of rejection of transferred T cells by host, we infected CNS1^+/+ or CNS1^−/− mice with rLM-OVA and analyzed the OVA-specific CD8 T cell response. At the peak of the primary response (day 7), both CNS1^+/+ and CNS1^−/− CD44⁺ OVA_257−264/K^b MHC tetramer⁺ CD8 T cells had drastically expanded (Fig. 9A). Indeed, CNS1^−/− cells proliferated more extensively compared with CNS1^+/+ cells at day 7 (Fig. 9B). Whereas both CNS1^+/+ and CNS1^−/− OVA-specific CD8 T cells were reduced during contraction, CNS1^−/− cells tended to decrease more rapidly than did CNS1^+/+ cells. We next compared surface expression of IL-7Rα, Ly6C, and KLRG1 in OVA-specific CD8 T cells during the infection. CNS1^+/+ and CNS1^−/− OVA-specific CD8 T cells showed similar time course of IL-7Rα, Ly6C, and KLRG1 expression, although the proportion of KLRG1^lo memory precursor effector T cells tended to be lower in CNS1^−/− cells compared with CNS1^+/+ cells at days 14 and 21 (Fig. 9C). These results suggest that the changes of IL-7Rα expression during the infection are not impaired in CNS1^−/− mice.

In this study, we determined the function of the CNS1 element in IL-7Rα expression during lymphocyte development and immune response. We first showed that IL-7Rα expression was unchanged in thymocytes of CNS1^−/− mice, except that Treg cells showed significantly lower IL-7Rα expression. We also found that IL-7Rα expression was unchanged during B cell development in the bone marrow of CNS1^−/− mice. In contrast, IL-7Rα expression was significantly reduced in peripheral T cell subpopulations of CNS1^−/− mice. Naive T cells of CNS1^−/− mice showed significantly reduced survival and homeostatic proliferation. Additionally, IL-7Rα upregulation by GC and TNF-α was completely inhibited in naive CNS1^−/− T cells in vitro. Clonal expansion of naive T cells and maintenance of memory T cells were significantly impaired in CNS1^−/− CD8 T cells. Although IL-7Rα expression was significantly lower in CNS1^−/− CD8 T cells, regulation of IL-7Rα expression during the immune response was retained. Thus, this study demonstrates that the CNS1 element is an enhancer essential for IL-7Rα expression in peripheral T cells and suggests a potential role for GC and TNF-α in controlling IL-7Rα expression and immune homeostasis.

The CNS1 element controls survival and proliferation by regulating IL-7Rα expression in peripheral T cells. CNS1-deficient mice represent a new model: they exhibit a profound defect in IL-7Rα expression in peripheral T cells but do not show complications in T cell development associated with deficiencies in IL-7 or IL-7Rα. Although naive CNS1^−/− T cells showed defective cell survival in vitro, Bcl-2, Bcl-x_L, and Mcl-1 expression was unchanged in CNS1^−/− naive T cells. As reported, Bcl-2 expression is not completely abrogated in IL-7Rα–deficient mice (6). Thus, these prosurvival proteins may be maintained via the activity of other cytokines present in the normal lymphoid microenvironment. Alternatively, decreased levels of these prosurvival proteins may lead to rapid cell death.

Deletion of CNS1 element did not alter regulation of IL-7Rα by IL-7 and TCR signals: IL-7Rα transcription was induced by growth factor withdrawal and repressed by IL-7 and TCR signals in CNS1^−/− T cells, albeit at lower expression levels than seen in wild-type T cells. In contrast to CNS1^−/− T cells, Foxo1-deficient T cells failed to up-regulate IL-7Rα after growth factor withdrawal (20). IL-7 signaling activates the PI3K/Akt pathway, which then inhibits Foxo1 activity by phosphorylation (41). Foxo1 reportedly binds to the CNS1 element and enhances IL-7Rα expression in peripheral T cells (20). Although lack of Foxo1 binding to the CNS1 contributes in part to IL-7Rα reduction in CNS1^−/− T cells, Foxo1 and/or other transcription factors may further modulate IL-7Rα transcription through different control elements.

CNS1^−/− T cells failed to induce IL-7Rα transcription in response to GC and TNF-α in vitro, suggesting a potential role for these factors in controlling IL-7Rα expression in peripheral T cells. Although GR-deficient mice show normal cellular composition of thymus and peripheral lymphoid organs (42, 43), little is known about IL-7Rα expression in GR-deficient T cells. Following infection with lymphocytic choriomeningitis virus, effector CD8 T cells reportedly express the IL-7Rα at lower levels in TNF-α–deficient compared with control mice (44). The NF-κB consensus motif found in the CNS1 element may function in IL-7Rα induction by TNF-α. Further studies should elucidate how GC and TNF-α regulate IL-7Rα.

Treg cells are highly dependent on the CNS1 element for IL-7Rα expression. Compared to wild-type cells, CNS1^−/− Treg cells completely lost IL-7Rα expression in thymus and lymph nodes. Treg cells receive stronger agonistic TCR signals than do conventional CD4 T cells. Therefore, agonistic TCR signaling may exert effects through the CNS1 element in this cell type. In contrast, anti-CD3 and anti-CD28 stimulation downregulated IL-7Rα in CNS1^−/− T cells in vitro. Such stimulation might be much stronger than agonistic TCR signals in Treg cells and exert effects through control elements other than the CNS1. Furthermore, because IL-7Rα–deficient Treg cells show defects in suppressive activity (2, 45), CNS1^−/− Treg cells may show weaker suppression than CNS1^+/+ Treg cells.

The CNS1 element plays a role in clonal expansion of naive CD8 T cells. OT-I cells from CNS1^−/− mice showed impaired expansion after Ag stimulation compared with cells from CNS1^+/+ mice. Although OT-I cells from wild-type mice downregulated IL-7Rα after Ag stimulation, they expressed IL-7Rα at substantial levels at the peak of the immune response (day 3). In contrast, OT-I cells from CNS1^−/− mice completely lost IL-7Rα expression by day 3. Therefore, IL-7 signaling may transmit a competence signal for efficient clonal expansion. In agreement with this idea, we previously observed that naive T cells from CD4-Cre IL-7Rα^flox/flox mice proliferate less efficiently than do T cells from control mice after anti-CD3 and anti-CD28 stimulation (2).

The CNS1 element may be important for survival of Ag-stimulated effector T cells. CNS1^−/− Ag-specific CD8 T cells showed a substantial reduction in effector cells, which was associated with low levels of IL-7Rα and Bcl-2 protein. Consistently, the number of CNS1^−/− memory CD8 T cells was substantially reduced. Despite loss of the CNS1 element, we observed relatively normal alterations in IL-7Rα expression in CNS1^−/− CD8 T cells. These data indicate that the CNS1 element may control survival and proliferation of effector cells by regulating IL-7Rα expression.

The CNS1 element is essential for IL-7Rα expression in peripheral T cells but not thymocytes. One explanation for this difference may be that GR occupancy of the enhancer may differ. A GR binding pattern was reportedly predetermined by DNase I sensitivity of target chromatin (46). Levels of acetylated histone H3 and monomethylated histone H3K4 at the CNS1 element are comparable between SP thymocytes and peripheral T cells (our unpublished observation), suggesting that chromatin accessibility is unchanged. Thus, the CNS1 element may represent a poised enhancer in thymocytes primed for gene expression, which then becomes active in peripheral T cells following GR binding.

In summary, we have demonstrated that the CNS1 element controls IL-7Rα expression in peripheral T cells and is critical for survival and expansion of naive and memory T cells. These findings will accelerate the understanding of molecular mechanisms relevant to IL-7Rα transcription by cis-control elements at the IL-7Rα locus. CNS1^−/− mice should also serve as a powerful tool to determine mechanisms underlying how IL-7Rα regulation influences T cell homeostasis and the immune response.

We thank Drs. J. Takeda and G. Kondoh for the KY1.1 ES cell line and targeting system, Drs. I. Saito, Y. Shinkai, M. Tachibana, and Y. Matsumura for adenovirus expressing Cre recombinase, Drs. N. Minato and Y. Agata for OT-I TCR transgenic mice, H. Hayashi and S. Kamioka for excellent technical assistance, and members of the Ikuta Laboratory for discussions.

The authors have no financial conflicts of interest.