IL-7R–Dependent Phosphatidylinositol 3-Kinase Competes with the STAT5 Signal to Modulate T Cell Development and Homeostasis

Appropriate cytokine signaling is required for precise control of the immune system in vivo. IL-7, a cytokine essential for lymphocyte development and homeostasis, controls survival, proliferation, and differentiation of T cells as well as V(D)J recombination of TCR γ-chain loci and IgH in early lymphocytes. IL-7 transduces signals through binding with IL-7R, which consists of a common receptor γ-chain (CD132) and a unique IL-7R α-chain (IL-7Rα; CD127). Mice lacking either IL-7 or IL-7Rα show markedly reduced numbers of T cells, B cells, and innate lymphoid cells (ILCs) in thymus and bone marrow (1–4) and impaired maintenance of naive and memory T cells in the periphery (5, 6). Following IL-7 binding to IL-7R, signaling through IL-7Rα activates JAK1 and JAK3 tyrosine kinases and triggers STAT5 and PI3K signaling.

IL-7Rα expression is dynamically controlled during lymphocyte development. In T cell differentiation, IL-7Rα is expressed on CD4⁻CD8⁻ double negative (DN) thymocytes and completely disappears from CD4⁺CD8⁺ double positive (DP) thymocytes (7). After positive selection, transient cessation of TCR signaling induces IL-7Rα re-expression on postselected thymocytes (8). IL-7Rα levels are upregulated on CD4⁺ or CD8⁺ single positive (SP) thymocytes and maintained on naive T cells in the periphery (7, 9). Furthermore, when activation of naive T cells occurs in immune responses, IL-7Rα is downregulated in effector T cells. Following T cell activation, a subset of effector T cells, referred to as memory precursor effector cells (MPECs), upregulates IL-7Rα and differentiates into memory CD8 T cells (10, 11). During B cell development in bone marrow, IL-7Rα is continuously expressed from common lymphoid progenitors (CLPs) to pro- and pre-B cells, then downregulated in late pre-B cells and lost on mature B cells (12). In contrast, IL-7Rα is also continuously expressed from common helper-like ILC progenitors (CHILPs) to group 2 ILC (ILC2) (4). Therefore, IL-7Rα expression is precisely regulated in a variety of lymphocytes during their development.

IL-7Rα activity depends on two major signaling cascades, STAT5 and PI3K pathways, both initiated by tyrosine (Y) phosphorylation of IL-7Rα (13, 14). Although IL-7Rα exhibits three cytoplasmic tyrosine residues conserved between humans and mice, tyrosine Y449 in a distal YXXM motif is most important for both STAT5 and PI3K signaling cascades and plays a critical role in executing IL-7 function in T and B cells in vivo (15–17). However, how STAT5 and PI3K signaling is differentially regulated and contributed remains unclear. In addition to Y449, methionine residue M452 of the YXXM motif is important for PI3K recruitment, as tyrosine-phosphorylated YXXM motifs are consensus-docking sites for Scr homology 2 domains of PI3K p85 subunits (18, 19). Thus, STAT5 and PI3K signals may play distinct roles downstream of IL-7Rα depending on different levels of IL-7 or the intensity of STAT5 and PI3K signals (20).

To investigate differences between STAT5 and PI3K signals downstream of IL-7Rα and effects of both pathways on lymphocyte development at different stages in vivo, we newly generated two mouse lines with point mutations in the IL-7Rα locus: IL-7Rα–M452L and IL-7Rα–Y449F mice. PI3K signaling was impaired in T cells of IL-7R–M452L and IL-7R–Y449F mice, although IL-7Rα signaling in wild-type (WT) T cells only modestly promoted phosphorylation of Akt, which is the downstream mediator for PI3K signal. Interestingly, STAT5 activation was significantly enhanced in T cells of IL-7R–M452L mice. Thus, our study reveals, to our knowledge, a novel manner of IL-7Rα signaling (through competition between PI3K with STAT5 signals) that contributes to development of early thymocytes and maintenance of memory precursor effector T cells via T cell factor–1 (TCF-1) regulation and limits excessive survival and homeostatic proliferation of peripheral T cells.

C57BL/6 mice were purchased from CLEA Japan. CD4-Cre transgenic (21), p85α^fl/flp85β^−/− (22), and Rag2^−/− mice on a C57BL/6 background (obtained from Dr. M. Ito at Central Laboratories for Experimental Animals, Kawasaki, Japan and a kind gift of Dr. F. W. Alt, Harvard Medical School, Boston, MA) were used throughout this article. All mice were maintained under specific pathogen-free conditions in the Experimental Research Center for Infectious Diseases in the Institute for Frontier Life and Medical Sciences, Kyoto University. All procedures were carried out under anesthesia to minimize animal suffering. All mouse protocols were approved by the Animal Experimentation Committee of the Institute for Frontier Life and Medical Sciences, Kyoto University.

The IL-7R–M452L and IL-7R–Y449F mutations were made by gene targeting of the YXXM motif of exon 8 (Supplemental Fig. 1A). To construct targeting vectors, the following DNA fragments were assembled in pBluescript KS (+) Vector: a diphtheria toxin A cassette, a 6010-bp fragment upstream of the IL-7Rα stop codon containing the mutated YXXM motif, a neomycin resistance gene cassette flanked by loxP sequences, and a 2293-bp fragment of 3′UTR and downstream sequence of the IL-7Rα locus. WT, IL-7R–M452L, and IL-7R–Y449F mutant sequences in the YXXM motif were as follows: WT, 5′-TATGTCACCATG-3′; IL-7R–M452L, 5′-TATGTCACCCTG-3′; and IL-7R–Y449F, 5′-TTTGTCACCATG-3′. The linearized targeting vector was introduced into the KY1.1 embryonic stem cell line, which was derived from C57BL/6 × 129S6/SvEvTac F1 mouse embryo (a kind gift of Dr. J. Takeda, Osaka University), by electroporation, and homologous recombinants were screened by PCR. Targeted embryonic stem clones were confirmed by PCR and DNA sequencing and then microinjected into ICR eight-cell embryos. Chimeric mice were bred with CAG-Cre transgenic mice to delete the neomycin resistance gene cassette and then backcrossed to C57BL/6 mice for at least 10 generations. We confirmed correct targeting of the IL-7Rα locus by PCR and DNA sequencing (Supplemental Fig. 1B, 1C).

Cells were prepared from indicated organs and surface-stained for 20 min at 4°C in PBS containing 0.05% NaN₃ and 0.1% BSA with fluorescent dye– or biotin-conjugated Abs. The following fluorescent dye– or biotin-conjugated Abs were used: anti-CD3ε (145-2C11), TCRβ (H57-597), CD5 (53-7.3), CD4 (RM4.5), CD8α (53-6.7), CD11b (M1/70), CD11c (N418), CD19 (MB19-1), CD24 (HSA) (30F1), CD25 (7D4), CD27 (LG.3A10), CD43 (S7), CD44 (IM7), CD45.1 (A20), CD45.2 (104), CD69 (H1.2F3), NK1.1 (PK136), γδTCR (GL-3), IL-33R (DIH9), CD62L (MEL-14), Qa-2 (69H1-9-9), c-kit (2B8), CD45R/B220 (RA3-6B2), IgM (M41), CD127 (A7R34), KLRG1 (2F1), Ter119, IFN-γ (XMG1.2), IL-2 (JES6-5H4), IL-4 (11B11), IL-10 (JES5-16E3), IL-13 (eBio13A), IL-17A (eBio17B7), Gr-1 (RB6-8C5), GM-CSF (MP1-22E9), p-STAT5 (47), p-Akt (T308) (J1-223.371), p-Akt (S473) (M89-61), and Bcl-2 (A19-3). Fluorescent dye– or biotin-conjugated Abs were purchased from Thermo Fisher Scientific, BD Biosciences, BioLegend, Cell Signaling Technology, and Tonbo Biosciences. The PBS57-loaded CD1d Tetramer was provided by the National Institutes of Health Tetramer Core Facility. Biotinylated mAbs were detected with PE-conjugated streptavidin (Thermo Fisher Scientific). Viable cells were analyzed on the FACSCanto II or FACSVerse Flow Cytometer (BD Biosciences) using FlowJo Software. In figures, values in quadrants, gated areas, and interval gates indicate percentages in each population.

For intracellular Bcl-2 staining, T cells were stained for surface Ags, fixed, permeabilized, and stained using the Foxp3 Staining Buffer Set (Thermo Fisher Scientific). For intracellular p-STAT5 staining, T cells after stimulation were fixed, permeabilized in ice-cold methanol, and stained using the Foxp3 Staining Buffer Set. For intracellular p-Akt staining, T cells after stimulation were fixed, permeabilized, and stained using BD Phosflow Buffer (BD Biosciences). For intracellular cytokine staining, cultured T cells were fixed, permeabilized, and stained with relevant Abs using IC Fixation Buffer (Thermo Fisher Scientific).

Naive CD4 and CD8 T cells were isolated from lymph nodes by using EasySep Naive CD4 and CD8 T Cell Enrichment Kits (STEMCELL Technologies). For adoptive transfer, isolated cells were counted, and 1 × 10⁶ cells were resuspended in 200 μl of PBS for i.v. injection into Rag2^−/− recipients. For quantitative PCR (qPCR) analysis, early T cell precursors (ETP) (Lin⁻c-kit⁺CD44⁺CD25⁻), DN2a (Lin⁻c-kit^highCD44⁺CD25⁺) and DN2b (Lin⁻c-kit^lowCD44⁺ CD25⁺) cells were sorted from the thymus with a FACSAria II Cell Sorter (BD Biosciences) after depletion using anti-CD4, anti-CD8, and anti-Ter119 MACS Microbeads and an LD column (Miltenyi Biotec).

Isolated T cells were cultured in RPMI 1640 medium containing 10% FBS, 50 μM 2-ME, and 10 mM HEPES (pH 7.4). For T cell stimulation, naive CD4 and CD8 T cells were cultured with 10 ng/ml IL-7. For a 2-[N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino]-2-deoxyglucose (2-NBDG) uptake assay, naive CD4 T cells were cultured with 10 ng/ml IL-7 in RPMI 1640 medium containing 10% FBS without glucose and then incubated with 10 μM 2-NBDG (Thermo Fisher Scientific) at 37°C for 20 min.

For Th1 and Th2 differentiation in vitro, naive CD4 T cells were cultured as we described previously (23) and with the presence of IL-7 (20 ng/ml). For induction of Th17 subsets, naive CD4 T cells were cultured with 4 μg/ml plate-bound anti-CD3 Ab, 4 μg/ml soluble anti-CD28 Ab (PV-1, a kind gift of Dr. R. Abe at Tokyo University of Science), and 20 ng/ml hIL-2 in the presence of 20 ng/ml IL-6, 2 ng/ml TGF-β, anti–IL-4 Ab (5 ng/ml), and with or without 20 ng/ml IL-7. After 4 d, the cells were restimulated with PMA (50 ng/ml) and ionomycin (2 μg/ml) for 4 h in the presence of brefeldin A. All cells were fixed, permeabilized, and stained with relevant Abs for flow cytometry.

Isolated cells from lymph nodes were used for immunoprecipitation as described previously (24). In brief, the cells were lysed in buffer containing 1% IGEPAL-630 (Sigma-Aldrich). Cell lysates were immunoprecipitated with anti-STAT5 rabbit polyclonal Ab (Santa Cruz Biotechnology), and the proteins were electrophoresed through a SuperSep Ace 5–20% gel (Wako Chemicals), then immunoblotted with anti–IL-7Rα goat polyclonal Ab (R&D Systems). An LAS-4000 Imaging System (Fuji Film) and ImageJ Software were used to quantitate digital images.

Total RNA was isolated and reverse-transcribed using random primers. cDNA was analyzed by real-time RT-PCR using SYBR Green PCR Master Mix (Qiagen) in an ABI7500 Real-Time PCR System (Applied Biosystems). PCR results were normalized to corresponding levels of Hprt mRNA in cDNA from thymocytes from WT mice. The following primers were used for real-time RT-PCR: Hprt, 5′-GTTGGATACAGGCCAGACTTTGTTG-3′ and 5′-GATTCAACTTGCGCTCATCTTAGGC-3′; stat5a, 5′-ATTACACTCCTGTACTTGCG-3′ and 5′-GGTCAAACTCGCCATCTTGG-3′; stat5b, 5′-TCCCCTGTGAGCCCGCAAC-3′ and 5′-GGTGAGGTCTGGTCATGAC-3′; tcf7, 5′-GGAGATGAGAGCCAAGGTCATT-3′ and 5′-CTGTGGTGGATTCTTGATGTTT-3′; Gata3, 5′-TACCGGGTTCGGATGTAAGTC-3′ and 5′-CCTTCGCTTGGGCTTGATAAG-3′; Notch1, 5′-CCCTTGCTCTGCCTAACGC-3′ and 5′-GGAGTCCTGGCATCGTTGG-3′; and Bcl2, 5′-TCGCTACCGTCGTGACTTC-3′ and 5′-AAACAGAGGTCGCATGCTG-3′.

Mice were injected i.p. with 1 mg of BrdU (Sigma-Aldrich). After 9 or 12 h, thymocytes or bone marrow cells were stained with the indicated surface markers and then fixed and permeabilized with Cytofix/Cytoperm (BD Biosciences) and treated with DNase at 37°C for 1 h. BrdU was stained using FITC-conjugated anti-BrdU Ab (BD Biosciences) for 30 min at room temperature. In the figures, values in histograms defined by interval gates indicate percentages of BrdU⁺ cells in each population.

Isolated naive CD4 and CD8 T cells were labeled with 2 μM CFSE (Dojindo Laboratories) in PBS at 37°C for 10 min. For in vitro proliferation and activation, cells were cultured with 50 ng/ml IL-7, or with 4 μg/ml plate-bound anti-CD3 Ab, and 4 μg/ml soluble anti-CD28 Ab (PV-1). For in vivo homeostatic proliferation, 1 × 10⁶ cells from each genotype or mixed 2 × 10⁶ cells from WT (CD45.1) and IL-7R–M452L (CD45.1 × CD45.2) mice at a 1:1 ratio were resuspended in 200 μl of PBS and adoptively transferred into Rag2^−/− recipients by i.v. injection. At indicated days, cells were analyzed by flow cytometry.

CD45.1 WT host mice were irradiated with a single dose of 9 Gy and then i.v. injected with a 1:1 mixture of 6 × 10⁶ congenic WT (CD45.1 × CD45.2) and IL-7R–M452L (CD45.2) bone marrow cells. Thymocytes and lymph node cells were isolated and analyzed by flow cytometry 7 wk later, and ratios were normalized to the neutrophil (CD11b⁺Gr-1⁺) population in bone marrow. For bacterial infection, mice were inoculated with 2 × 10⁴ CFU of Listeria monocytogenes expressing OVA (LM-OVA) after 6–7 wk.

Recombinant LM-OVA was described previously (23). Mice were infected with 2 × 10⁴ CFU of LM-OVA by i.v. injection. At days 7, 15, and 30, spleen cells were stained with H-2K^b OVA G4 Tetramer-SIIGFEKL-APC (Medical and Biological Laboratories) and then stained with relevant Abs for flow cytometry.

All data are presented as mean ± SEM. Comparisons between two samples were performed using an unpaired two-tailed Student t test. For multiple group comparisons, one-way ANOVA analyses with multiple-comparison tests were performed using GraphPad Prism 7 Software (*p < 0.05, **p < 0.01; N.S. is not significant).

To investigate differential roles of STAT5 and PI3K signaling through IL-7Rα, we generated two mouse lines harboring point mutations in the IL-7Rα locus: IL-7R–M452L and IL-7R–Y449F mice (Supplemental Fig. 1). PI3K activation results in phosphorylation of two key Akt residues, T308 and S473 (25). To assess PI3K signaling in IL-7R–M452L and IL-7R–Y449F mice, we stimulated freshly isolated naive T cells with IL-7 and then analyzed phosphorylation of Akt T308 and S473 by flow cytometry. Although IL-7–mediated signaling modestly promoted Akt phosphorylation (p-Akt) in WT T cells, those levels were significantly reduced in naive CD4 and CD8 T cells of IL-7R–M452L and IL-7R–Y449F mice compared with WT cells (Fig. 1A, 1B). These results suggest that PI3K signaling downstream of IL-7Rα requires both Y449 and M452 and that PI3K activation is impaired in either IL-7R–M452L or IL-7R–Y449F mice.

To examine effects on STAT5 signaling in IL-7R–M452L and IL-7R–Y449F mice, we stimulated naive T cells with IL-7. Levels of p-STAT5 were severely decreased in CD4 and CD8 T cells of IL-7R–Y449F mice (Fig. 1C, 1D), consistent with a previous report (16). By contrast, STAT5 signaling was significantly enhanced in CD4 and CD8 T cells of IL-7R–M452L mice. Levels of p-STAT5 in IL-7R–M452L T cells showed ∼1.4- to 1.5-fold increase compared with that in WT cells (Fig. 1C, 1D). The enhanced STAT5 signaling was confirmed even after stimulation with IL-7 for 12 h (Fig. 1E, 1F). However, both M452L and Y449F mutations did not alter cell surface expression of IL-7Rα (CD127) on SP thymocytes and peripheral T cells (Fig. 1G). Furthermore, although the IL-7Rα expression was downregulated by IL-7 stimulation, the protein levels of IL-7Rα on cell surface and cytoplasm of IL-7R–M452L T cells were comparable to those of WT T cells after IL-7 stimulation (Supplemental Fig. 2A, 2B). In addition, the levels of p-STAT5 were also enhanced in IL-7R-WT/M452L heterozygous T cells but limited to 1.1- to 1.2-fold increase (Supplemental Fig. 2C), much weaker than in IL-7R–M452L homozygous T cells. Coincidently, the PI3K signaling seemed to be preserved in IL-7R–WT/M452L heterozygous T cells (Supplemental Fig. 2D). Thus, STAT5 signaling under IL-7Rα is enhanced in IL-7R–M452L mice without IL-7Rα expression change.

To investigate how PI3K regulates STAT5 activation under IL-7R signaling, we freshly isolated naive CD4 T cells from IL-7R–M452L, IL-7R–Y449F, and WT mice. However, Stat5a and Stat5b mRNA levels were comparable in all three genotypes of T cells, even stimulated with IL-7 (Fig. 2A). Then, we asked whether STAT5 binding to IL-7Rα is modified in IL-7R–M452L T cells by immunoprecipitation. After IL-7 stimulation, binding of STAT5 and IL-7Rα was enhanced in IL-7R–M452L compare with WT T cells (Fig. 2B, 2C). Thus, PI3K inhibits STAT5 activation and therefore suppresses STAT5 signaling by limiting the interaction between STAT5 and IL-7Rα.

We next tested the possibility that PI3K biding to IL-7Rα Y449 might suppress STAT5 signaling. Because the PI3K binding subunit is p85 (26), we analyzed CD4-Cre p85α^fl/flp85β^−/− mice to assess competition for IL-7Rα between PI3K and STAT5. Consistent with our observation in IL-7R–M452L T cells, p-STAT5 levels significantly increased after IL-7 stimulation in CD4 and CD8 T cells of CD4-Cre p85α^fl/flp85β^−/− mice (Fig. 2D, 2E), suggesting that impaired binding of PI3K via p85 allows STAT5 a greater chance to interact with IL-7Rα in IL-7R–M452L T cells.

Because IL-7Rα signaling is essential for normal T cell development in thymus and influences the ETP lineage (27), we first compared both the frequency and absolute number of DN thymocytes at different stages between IL-7R–M452L, IL-7R–Y449F, and WT mice. ETP were severely reduced in number in IL-7R–Y449F mice, but thymocyte development slightly recovered during DN3 to DN4 stages (Fig. 3A, 3B). By contrast, the number of DN thymocytes modestly decreased only at ETP and DN2a stages in IL-7R–M452L mice (Fig. 3A, 3B). This transient reduction was rescued at the DN2b stage, and cell numbers were normal at DN3 and DN4 stages (Fig. 3B). Although the reduction becomes milder, IL-7R–WT/M452L heterozygous mice also showed a slight decrease in cell numbers of ETP and DN2 populations (Supplemental Fig. 3A). We then performed a competitive bone marrow transplantation with WT (CD45.1 × CD45.2) and IL-7R–M452L (CD45.2) cells mixed at a 1:1 ratio and transferred them into irradiated congenic hosts (CD45.1). DN thymocytes were analyzed 7 wk later for relative contribution of each genotype. The IL-7R–M452L ETP and DN2a thymocytes were outcompeted by the WT cells (Fig. 3C, 3D).

We next asked how early T cell development is impaired in IL-7R–M452L or IL-7R–Y449F mice. As STAT5 induces expression of the anti-apoptotic factor Bcl-2 in thymocytes (15), we compared Bcl-2 expression in IL-7R–M452L and IL-7R–Y449F mice with that of WT. Bcl-2 expression was reduced at DN2 to DN4 stages in IL-7R–Y449F mice but not in IL-7R–M452L mice (Fig. 3E). We then carried out a BrdU uptake assay to assess cell proliferation. The frequency of BrdU⁺ cells was significantly decreased in DN2 thymocytes of IL-7R–Y449F mice compared with WT mice but was relatively unchanged in IL-7R–M452L mice (Fig. 3F, Supplemental Fig. 3B).

IL-7 signaling reportedly regulates expression of several transcription factors essential for early T cell development (28, 29). We found that transcript levels of TCF-1 (encoded by the Tcf7 gene) were significantly reduced in IL-7R–M452L mice, whereas transcript levels of other transcription factors examined (among them GATA3 and Notch1) were almost unchanged (Fig. 3G). Furthermore, flow cytometry analysis using anti–TCF-1 Ab showed slightly reduced TCF-1 protein levels in ETP and DN2 thymocytes of IL-7R–M452L mice (Fig. 3H). Taken together, these results suggest that STAT5 and PI3K signals through IL-7Rα Y449 promote cell survival and proliferation, although the PI3K suppresses STAT5 signals by competing for phosphorylated Y449 in IL-7Rα, which regulates TCF-1 expression and modulates early T cell development at ETP to DN2 stages.

We previously reported that IL-7Rα controls survival and proliferation of SP thymocytes (1). To determine how STAT5 and PI3K signals downstream of IL-7Rα Y449 alters development of SP thymocytes, we compared the frequency and absolute number of DP and SP thymocytes between IL-7R–M452L, IL-7R–Y449F, and WT mice. Although the DP and SP thymocyte frequency in IL-7R–Y449F mice was comparable to that in WT mice, absolute numbers of DP and SP thymocytes decreased compared with WT mice (Fig. 4A, 4B). By contrast, frequency and absolute numbers of DP and SP thymocytes in IL-7R–M452L mice were comparable to those in WT mice (Fig. 4A, 4B). In addition, the number of invariant NKT (iNKT) cells, regulatory T (Treg) cells, and γδ T cells significantly decreased, relative to WT mice, in IL-7R–Y449F mice but not in IL-7R–M452L mice (Fig. 4B). Thymocytes undergo positive selection during progression from TCRβ^intCD69⁺ to TCRβ^hiCD69⁺ stages. The frequencies of those thymocytes in IL-7R–M452L mice were similar to those in WT mice (Fig. 4C). After positive selection, SP thymocytes maturation process was also unchanged in IL-7R–M452L mice (Fig. 4D). Furthermore, Bcl-2 expression was reduced, relative to WT mice, in DP and SP thymocytes of IL-7R–Y449F mice but not in IL-7R–M452L mice (Fig. 4E). In addition, cell proliferation decreased compared with WT mice in SP thymocytes of IL-7R–Y449F mice but not IL-7Rα–M452L mice (Fig. 4F, Supplemental Fig. 3C). These results indicate that although IL-7Rα–dependent STAT5 and PI3K signals promote survival and proliferation in SP thymocytes, IL-7Rα–M452L mice show no alteration of SP thymocyte development.

IL-7Rα is expressed in CLPs. Both Ly6D⁻ and Ly6D⁺ CLP fractions in bone marrow showed similar absolute numbers between IL-7R–Y449F, IL-7R–M452L, and WT mice (Fig. 5A). IL-7R signaling is also involved in ILCs development (4, 30), and its expression is especially higher on the ILC2 population. We observed comparable frequency and absolute numbers of CHILPs, ILC2 progenitors (ILC2Ps), and mature ILC2s (mILC2s) in bone marrow between IL-7R–M452L and WT mice (Fig. 5B, 5C). By contrast, the mILC2 population was severely reduced in IL-7R–Y449F mice relative to WT mice, although the absolute numbers of ILC2Ps slightly decreased (Fig. 5B, 5C). We next asked how IL-7 signaling via IL-7Rα Y449 affects ILC2 development in bone marrow. The Bcl-2 expression was downregulated in ILC2s, and the frequency of BrdU⁺ cells decreased in CHILPs of IL-7R–Y449F mice relative to WT mice, whereas those numbers were unchanged in IL-7R–M452L mice (Fig. 5D, 5E). These results suggest that STAT5 and PI3K signals dependent on IL-7Rα Y449 promote ILC2 development and are indispensable for mILC2 differentiation in bone marrow.

IL-7Rα is also expressed on pro- and pre-B cells for their survival and differentiation (17). The absolute number of B cells in Hardy fractions A to F were unchanged in IL-7R–M452L mice, whereas B cells in fractions C through F were severely reduced in IL-7R–Y449F mice relative to WT mice (Fig. 5F). These results suggest that STAT5 and PI3K signals downstream of IL-7Rα Y449 are important for B cell development in bone marrow.

Apart from its role in early lymphocyte development, IL-7 is also important for maintenance of mature T cells in the periphery (6). The number of CD4 and CD8 T cells, iNKT cells, regulatory T cells, and γδ T cells markedly decreased in lymph nodes of IL-7R–Y449F mice relative to WT mice (Fig. 6A). By contrast, the number of CD4 and CD8 T cells increased in lymph nodes of IL-7R–M452L mice, whereas iNKT cells, regulatory T cells, and γδ T cells were unchanged relative to WT mice (Fig. 6A). We also found that CD4 and CD8 T cells slightly increased in IL-7R–WT/M452L heterozygous mice (Supplemental Fig. 3D). To determine mechanisms underlying these changes, we first checked expression of Bcl-2 in CD4 and CD8 T cells from lymph nodes. Relative to WT mice, levels of Bcl-2 protein and mRNA increased in IL-7R–M452L mice but decreased in IL-7R–Y449F mice (Fig. 6B, 6C). Next, we assessed effects of IL-7 or TCR stimulation ex vivo on naive T cell proliferation and activation. To do so, we freshly isolated naive CD4 T cells from lymph nodes, labeled them with CFSE, and cultured them with rIL-7 or anti-CD3 and anti-CD28 Abs for 4 d. We then separated dividing cells into three fractions: CFSE^high, CFSE^int, and CFSE^low. Following IL-7 simulation, the frequency of CFSE^high cells significantly increased in IL-7R–Y449F T cells but was unchanged, relative to WT mice, in IL-7R–M452L T cells (Fig. 6D, 6E). We obtained similar results following stimulation with anti-CD3 and anti-CD28 Abs: frequencies of CFSE^high and CFSE^int cells significantly increased, relative to WT mice, in IL-7R–Y449F T cells (Fig. 6F, 6G). Because expression of CD5 correlates with the affinity or avidity of the TCR, we checked CD5 levels in lymph node T cells by flow cytometry. CD5 expression decreased in IL-7R–Y449F relative to WT T cells (Supplemental Fig. 3E, 3F). These results suggest that IL-7– or TCR-induced proliferation and activation are impaired in IL-7R–Y449F mice but not in IL-7R–M452L mice.

IL-7 is essential for homeostatic proliferation of mature T cells (6, 16). To evaluate effects on this process, we purified naive CD4 T cells of all three genotypes from lymph nodes, labeled them with CFSE, and then transferred them separately into Rag2^−/− recipient mice. The frequency of CFSE^high cells notably increased in IL-7R–Y449F cells relative to WT T cells (Fig. 6H, 6I), indicating that homeostatic proliferation is impaired in these mutants. Conversely, the frequency of CFSE^low cells from IL-7R–M452L mice significantly increased relative to WT mice (Fig. 6H, 6I). Coincidently, we equally mixed naive CD4 T cells from WT (CD45.1) and IL-7R–M452L (CD45.1 × CD45.2) mice and transferred the cells into Rag2^−/− (CD45.2) recipient mice. The frequency of CFSE^low cells in IL-7R–M452L T cells was also significantly increased compared with WT T cells (Fig. 6J), suggesting that homeostatic proliferation is enhanced in these mutants. Moreover, we generated bone marrow chimera mice in irradiated congenic hosts (CD45.1) with WT (CD45.1 × CD45.2) and IL-7R–M452L (CD45.2) T cells mixed at a 1:1 ratio. After 7 wk, the reconstitution of IL-7R–M452L T cells were significantly elevated in lymph node than WT T cells (Fig. 6K, 6L). In addition, because IL-7 signaling promotes glucose uptake to support T cell survival (31), we measured potential differences in glucose uptake by directly incubating naive T cells of all three genotypes with IL-7 and 2-NBDG. Relative to WT cells, glucose uptake decreased in IL-7R–Y449F cells but not in IL-7R–M452L T cells (Supplemental Fig. 3G). Collectively, these results demonstrate that STAT5 and PI3K signals through IL-7Rα Y449 are important for survival and proliferation of T cells in the periphery and that STAT5 and PI3K competitive signals seem to limit excessive STAT5 activation and support normal survival and homeostatic proliferation of T cells.

Because IL-7 signaling plays a role in Th cell differentiation and Th17 cells highly express IL-7Rα (32), we analyzed Th cell differentiation under an IL-7-stimulated condition in IL-7R–Y449F, IL-7R–M452L, and WT mice. Differentiation of IFN-γ–producing Th1 cells was unchanged in T cells from all three genotypes (Supplemental Fig. 4A, 4B). We obtained similar results when we examined Th2 cell differentiation (Supplemental Fig. 4C, 4D). By contrast, the frequency of IL-17–producing Th17 cells was reduced in IL-7R–M452L relative to WT T cells but was unchanged in IL-7R–Y449F T cells (Fig. 7A, 7B). However, the frequency of Th17 cells was comparable between WT and IL-7R–M452L T cells when cultured without IL-7 (Supplemental Fig. 4E, 4F), indicating that the reduction of IL-17–producing Th17 cells in IL-7R–M452L T cells was due to the IL-7Rα signaling. Additionally, there was no difference in frequency of GM-CSF–producing cells among genotypes. To analyze Th17 differentiation in vivo, we isolated naive CD4 T cells from IL-7R–M452L and WT mice and transferred them into Rag2^−/− mice. After 6 d, we stimulated splenic CD4 T cells with PMA and ionomycin ex vivo. Consistent with in vitro results, the frequency of IL-17–producing Th17 cells moderately decreased in IL-7R–M452L T cells, whereas the frequency of IFN-γ–producing Th1 cells was unchanged in IL-7R–M452L and WT mice (Fig. 7C). These results suggest that STAT5 and PI3K signals downstream of IL-7Rα are dispensable for Th cell differentiation but that T cells mutant in M452 of IL-7Rα is impaired in Th17 differentiation.

IL-7Rα is expressed at high levels in memory T cells and critical for their survival. Thus, we asked whether competition between IL-7Rα–dependent STAT5 and PI3K signals alters development of effector and memory CD8 T cells. To do so, we infected IL-7R–M452L and WT littermate mice with LM-OVA and analyzed splenic T cells at 7, 15, and 30 d later. Ag-specific CD8 T cells were divided into KLRG1⁻CD127⁻ early effector cells, KLRG1⁺CD127⁻ short-lived effector cells (SLECs), and KLRG1⁻CD127⁺ MPECs, as described (10) (Fig. 7D). Although absolute numbers of the three fractions were unchanged in IL-7R–M452L relative to WT mice at days 7 and 15, MPEC numbers moderately decreased at day 30 (Fig. 7E). Furthermore, we mixed WT (CD45.1 × CD45.2) and IL-7R–M452L (CD45.2) bone marrow cells at a 1:1 ratio, transferred them into irradiated congenic hosts (CD45.1), then infected the mice 6–7 wk later with LM-OVA. After 30 d, the frequency of IL-7R–M452L MPECs were significantly decreased compare with WT MPECs (Fig. 7F). IFN-γ production was comparable in Ag-specific CD8 T cells of IL-7Rα–M452L and WT mice (Fig. 7G). As TCF-1 deficiency limits effector CD8 T cell expansion (33), we next assessed TCF-1 levels in Ag-specific effector and memory precursors postinfection. Relative to WT mice, TCF-1 expression decreased in MPECs but not in SLECs of IL-7R–M452L mice (Fig. 7H). These results showed that the impaired MPECs in IL-7Rα–M452L mice might be related with the reduction of TCF-1 expression.

Signal transmission through IL-7Rα Y449 is important for lymphocyte development and homeostasis. In addition to the role of individual STAT5 and PI3K signals through IL-7Rα Y449 in lymphocyte differentiation, survival, and proliferation, we identified, to our knowledge, a novel regulation of competitive STAT5 and PI3K signals downstream of IL-7Rα Y449 using IL-7R–M452L mice. That competition maintains an appropriate balance between both signaling pathways and supports normal development and maintenance of early thymocytes and memory T cells. Coincidently, this competition also controls survival and homeostatic proliferation of peripheral T cells. These findings suggest that IL-7–mediated signaling is controlled not only by surface IL-7Rα expression but also by regulation of signals of its effectors.

In this study, we first investigated differences of STAT5 and PI3K signals dependent on the IL-7Rα YXXM motif, which contains both Y449 and M452. The levels of IL-7–mediated Akt phosphorylation were modestly reduced in naive T cells of IL-7R–M452L and IL-7R–Y449F mice, suggesting that Y and M residues in the YXXM motif function in PI3K signaling. By contrast, STAT5 phosphorylation markedly decreased, relative to WT T cells, in IL-7R–Y449F T cells but was significantly elevated in IL-7R–M452L T cells, indicating that a decline in PI3K signals correlates with enhanced STAT5 signaling without altering IL-7Rα expression. As the mRNA levels of stat5a and stat5b were unchanged in either freshly isolated or IL-7 stimulated T cells, we hypothesize that PI3K reduced the chance of interaction between STAT5 and IL-7Rα in IL-7R–M452L T cells. This idea is supported by immunoprecipitation with anti-STAT5 Ab in IL-7R–M452L or WT T cells and our finding that T cells of CD4-Cre p85α^fl/flp85β^−/− mice show phenotypes similar to those of IL-7R–M452L mice. Although it is clear that PI3K competes with STAT5 for the YXXM motif, further validation may be needed for whether STAT5 is also able to compete with PI3K in a similar manner. Taken together, PI3K suppresses STAT5 signaling by competing for phosphorylated Y449 in IL-7Rα signaling.

In thymic T cell development, survival and proliferation were most affected in early thymocytes at DN2 stages in IL-7R–Y449F mice, consistent with a previous report (16, 27). Coincidently, development of early thymocytes at ETP to DN2 stages, which expressed higher levels of IL-7Rα in DN thymocyte, was transiently impaired in IL-7R–M452L mice. The impairment in IL-7R–M452L mice was restored after DN2 stage. This may be due to the downregulation of IL-7Rα, whereas the expression of TCF-1, a transcription factor critical for T-lineage specification and differentiation (34), can be induced by a pre-TCR signal other than IL-7Rα at later DN stages. PI3K signaling enhances the Wnt/β-catenin cascade by inhibiting GSK3 activation and then Wnt/β-catenin promotes TCF-1 activation for such T cell development (35), whereas IL-7 has also been considered as a positive effector for the thymocyte development. Nevertheless, intense IL-7 signaling inhibits TCF-1 expression in DN thymocytes (29, 36), and our IL-7R–M452L heterozygous mice with enhanced STAT5 signaling also exhibited reduction of ETP and DN2 thymocytes. Taken together, beyond the regulation of cell surface IL-7Rα expression, synergistic effects of decreased PI3K and increased STAT5 signals under IL-7Rα may restrict TCF-1 expression and also influence its function at ETP to DN2 stages. In contrast, although IL-7Rα is re-expressed from SP stage at high levels, IL-7R–M452L mice showed normal development of SP thymocytes. It may be due to the compensation from IL-2Rβ or TCR, which are also highly expressed in SP thymocytes and transduce STAT5 or PI3K signal. Thus, PI3K in competition with STAT5 in IL-7Rα signaling appears to keep adequate signals and has a transient effect on early T cell development.

In contrast to the context of the thymus, early ILC2 development in bone marrow was only partially altered in IL-7R–Y449F mice. Considering that ILCs also express IL-2Rβ (CD122) and that IL-15 (which, like IL-7, can activate both STAT5 and PI3K) is expressed by diverse cells in bone marrow (37, 38), ILC2 development might be compensated by IL-15 when IL-7 signal is reduced. By contrast, the mILC2 population was severely reduced in IL-7R–Y449F mice, indicating that STAT5 and PI3K signals through IL-7Rα are critical for ILC2 maturation, although further study in the relevance to TSLP signaling will be needed. Nevertheless, the competition between STAT5 and PI3K through IL-7Rα Y449 did not affect ILC2 and B cell development in bone marrow. Because distinct signals are reportedly transduced by different IL-7 levels in vitro (20), different IL-7Rα expression on the cell surface or local concentrations of IL-7 and IL-15 in the thymus and bone marrow microenvironments may ensue different effects.

Cell survival and homeostatic proliferation of peripheral T cells are impaired in IL-7R–Y449F mice but enhanced in IL-7R–M452L mice. In addition, STAT5 and PI3K signals through IL-7Rα Y449 are important for IL-7– and TCR-induced proliferation and glucose uptake in T cells. Despite the fact that IL-7Rα was unchanged in IL-7R–M452L mice, the absolute number of peripheral T cells increased with the elevated levels of Bcl-2 and p-STAT5. Furthermore, a competitive bone marrow transplantation experiment showed a significant increase of IL-7R–M452L T cells. Thus, we conclude that PI3K competes with STAT5 in IL-7Rα signaling and modulates T cell homeostasis to maintain a proper population size of peripheral T cells.

In effector and memory T cell development, IL-7 reportedly enhances Th1 but suppresses Th17 differentiation from naive CD4 T cells (32). Although cell survival was severely impaired, differentiation of the Th subset of IL-7R–Y449F T cells was comparable to that in WT mice, suggesting that STAT5 and PI3K signals through IL-7Rα Y449, per se, do not drive differentiation of particular Th subsets. By contrast, the differentiation of Th17 cells, which express higher levels of IL-7Rα among Th subsets, was impaired in IL-7R–M452L T cells. IL-17 expression is reportedly driven in an STAT3-dependent manner (39), and STAT5 competes with STAT3 for nuclear transport (40). Coincidentally, the PI3K-Akt axis enhances Th17 differentiation (41). Thus, the increased STAT5 and decreased PI3K signals may synergize to impair Th17 differentiation in IL-7R–M452L mice.

IL-7 signaling has been reported to promote survival and differentiation of memory CD8 T cells (5). Moreover, STAT5 and PI3K reportedly play differential roles in effector and memory CD8 T cells dependent on signaling mediated by cytokines, such as IL-2, IL-7, and IL-15. Similar to our findings, others report that constitutive PI3K signaling is associated with downregulation of STAT5 signaling, whereas constitutive STAT5 signals maintain effector and memory CD8 T cells by enhancing cell survival (42). To assess effects of competition between STAT5 and PI3K signals through IL-7Rα, we analyzed Ag-specific CD8 T cells after bacterial infection and found a moderate decrease in the number of MPECs 30 d later. Although STAT5 signaling was stronger in IL-7R–M452L T cells, it is insufficient to fully support the long-term maintenance of MPECs, which may be due to differentiation rather than survival. Consistent with a report that differentiation and maintenance of memory CD8 T cells depends on the transcription factor TCF-1 (33), TCF-1 expression also decreased in MPECs but not SLECs in memory CD8 T cell development of IL-7R–M452L mice. Therefore, the competition between PI3K and STAT5 in IL-7Rα signaling likely sustains development and maintenance of MPECs by inducing TCF-1 expression.

In conclusion, we found that the distal YXXM motif of IL-7Rα is important for both STAT5 and PI3K signals in vivo using IL-7R–M452L and IL-7R–Y449F mutant mice. This study indicates that PI3K competes with STAT5 in IL-7Rα signaling and maintains an appropriate signal balance for development and maintenance of early thymocytes and memory precursors by regulating TCF-1 expression. Furthermore, this competition maintains signal intensity appropriate for T cell homeostasis and helps control the population size of peripheral T cells. Thus, this study provides, to our knowledge, a new insight into complex regulation of IL-7Rα signaling, which supports immune development and responses.

We thank Drs. J. Takeda, K. Yusa, and G. Kondoh for providing the KY1.1 embroyonic stem cell line and targeting system, Dr. R. Abe for providing the anti-CD28 Ab, Dr. M. Ito for Rag2^−/− mice, Dr. L. C. Cantley for p85α^fl/flp85β^−/− mice, and members of the K.I. laboratory for discussion.

The authors have no financial conflicts of interest.