C57BL/6NCrl (C57BL/6) and BALB/cAnNCrl (BALB/c) mice were obtained from the Charles River Laboratories (Frederick, MD). IL-2Rβ^Tg mice were generated in house as previously described (20). IL-2Rβ-deficient mice (Il2rb ^–/–) and IL-2Rβ-floxed mice (Il2rb ^fl/fl) were obtained from the Jackson Laboratories (21, 22). PLZF^Cre transgenic mice were kindly provided by Dr. D. Sant’Angelo (Rutgers University, NJ) (23). CA-STAT5^Tg mice were a kind gift from by Dr. M. A. Farrar (U. Minnesota, MN) (24). T-bet-ZsGreen reporter transgenic mice (TBGR^Tg) on C57BL/6 background were a kind gift from Dr. Jinfang Zhu (NIAID, NIH) (25), and these mice were backcrossed to BALB/cAnNCrl for at least 6 generations before analysis. All experimental mice were analyzed between 6 and 12 weeks of age, except for Il2rb ^–/– and their littermate controls, which were analyzed between 4 and 6 weeks of age. All mice were cared for in accordance with NIH guidelines. Animal experiments were approved by the NCI Animal Care and Use Committee.

Single-cell suspensions were prepared from the indicated tissues and stained with fluorescence-conjugated antibodies as previously described (26). The data were acquired using LSR Fortessa or LSRII flow cytometers (BD Biosciences) and were analyzed using software platforms developed by the EIB Flow Cytometry Facility, CCR, NCI, NIH. Live cells were gated by forward scatter exclusion of dead cells stained with propidium iodide. The following antibodies were used for staining: HSA (M1/69), T-bet (4B10), Foxp3 (FJK-16s), RORγt (AKFJS-9) and isotype control antibodies, all from eBioscience; CD4 (GK1.5 and RM4.5), CD8α (53–6–7), CD69 (H1.2F3), TCRβ (H57-597) and IL2Rβ (TM-β1) from BD Biosciences; CD44 (IM7), NK1.1 (PK136), IL2Rα (PC61), CCR7 (4B12) and PLZF (9E12) from BioLegend. Fluorochrome-conjugated CD1d tetramers loaded with PBS-57 (CD1dTet) and unloaded controls were obtained from the NIH tetramer facility (Emory University, Atlanta, GA). Intracellular Foxp3, PLZF, RORγt, and T-bet proteins were detected using a Foxp3 staining kit according to the manufacturer’s instructions (eBioscience Thermo Fisher).

To identify the subset composition of iNKT cells, total thymocytes were first stained with PBS-57-loaded mouse CD1d tetramers, fixed, permeabilized and then stained for nuclear transcription factors. Briefly, thymocytes were stained with fluorochrome-conjugated CD1d tetramers in FACS buffer (0.5% BSA, 0.1% sodium azide in Ca²⁺ and Mg²⁺-free HBSS) for 20 minutes at 4°C. Without removing the tetramer reagents, the cells were incubated for an additional 30 min at 4°C with antibodies for surface markers. Excess reagents were then washed out with FACS buffer, and the cells were resuspended in a 1:3 mixture of concentrate/diluent working solution of the Foxp3 Transcription Factor Staining Buffer kit (eBioscience Thermo Fisher), followed by incubation at room temperature for 20 minutes. Next, the cells were washed twice with 1× permeabilization buffer (eBioscience Thermo Fisher) before adding antibodies for transcription factors, specifically antibodies against PLZF, RORγt and T-bet. The cells were incubated at room temperature for 1 hour before washing out excess reagents with FACS buffer and analysis by flow cytometry.

Recombinant mouse IL-15 (Peprotech) dissolved in PBS was administered into IL-2Rβ^Tg BALB/c mice using Alzet osmotic pumps (Durect) following the manufacturer’s instruction. The pumps (model 1002) were set to release IL-15 at a rate of 3 µg of recombinant IL-15 per 24 hours.

The data were shown as the means ± SEM. Statistical significance was determined by unpaired two-tailed Student’s t-test. P values of less than 0.05 were considered significant. *P < 0.05, **P < 0.01, ***P <0.001. All statistical analysis was performed using GraphPad Prism 7 software (GraphPad software).

To understand the role of IL-2Rβ in thymic iNKT cell generation, we first examined IL-2Rβ expression on iNKT cells in the context of αβ T cell development in the thymus. The process of thymocyte positive selection and maturation can be visualized using the combination of the two surface markers CD69 and CCR7 which identifies 5 distinct developmental stages, progressing from stage I to stage V ( Figure 1A , top) (27). TCR engagement of immature thymocytes (stage I) induces the expression CD69 (stage II) followed by the upregulation of the chemokine receptor CCR7, so that CD69⁺CCR7^int cells (stage III) correspond to thymocytes undergoing positive selection and lineage commitment. Stage IV and V cells, on the other hand, correspond to postselection thymocytes. We found that iNKT cells, as identified by PBS-57-loaded CD1d tetramers (CD1dTet⁺), arise as early as in stage I, and that they were then prominent in stage II ( Figure 1A , bottom). Importantly, stage I iNKT cells expressed both the cytokine receptor γc and IL-2Rβ ( Figure 1B and Supplementary Figure 1 ), indicating that the induction of IL-2Rβ is one of the earliest events in iNKT cell development. Indeed, such kinetics of IL-2Rβ expression differed substantially from that of CD1dTet-negative conventional αβ T cells, which remained mostly absent for IL-2Rβ expression throughout their development ( Figure 1B and Supplementary Figure 2 ).

To further align IL-2Rβ expression with the selection and maturation of iNKT cells, we next assessed surface IL-2Rβ expression in developmental stages of thymic iNKT cells. CD44-negative HSA^hi thymocytes that bind PBS-57-loaded CD1d tetramers correspond to preselection iNKT cells, and they are usually referred to as stage 0 (ST0) iNKT cells (28). IL-2Rβ was mostly absent on ST0 iNKT cells but was induced following differentiation into HSA^lo mature iNKT cells ( Figure 1C and Supplementary Figure 1B ). Thymic differentiation of HSA^lo iNKT cells proceeds along a well-characterized pathway that is marked by the expression of CD44 and NK1.1, whereby CD44^–NK1.1^– cells correspond to stage 1 (ST1), followed by CD44⁺NK1.1^– cells that are stage 2 (ST2), and terminally differentiate into CD44⁺NK1.1⁺ stage 3 cells (ST3) (28). Interestingly, IL-2Rβ was expressed at low levels on ST1 and ST2 cells but highly expressed on ST3 iNKT cells ( Figure 1C and Supplementary Figure 1B ). Thus, all mature thymic iNKT cells express IL-2Rβ to a certain degree, but the amount of surface IL-2Rβ proteins substantially increases with maturation.

Mature iNKT cells are also categorized into 3 distinct subsets, namely NKT1, NKT2, and NKT17 cells, based on their transcription factor expression profiles (10). Differential expression of the transcription factors PLZF versus RORγt can visualize these three subsets, which we utilized to assess IL-2Rβ expression on individual thymic iNKT subsets ( Figure 1D ). Consistent with previous observations (10), IL-2Rβ was highly induced on NKT1 cells and minimally expressed on PLZF^hiRORγt^neg NKT2 and PLZF^intRORγt⁺ NKT17 cells ( Figure 1D ). Altogether, these results demonstrated that IL-2Rβ is highly upregulated on terminally differentiated ST3 iNKT cells and that such IL-2Rβ⁺ iNKT cells correspond to NKT1 cells.

Having established that IL-2Rβ expression is associated with iNKT cell maturation, we next wished to examine whether IL-2Rβ would be also required for the development and differentiation of iNKT cells. To this end, we aimed to set up an experimental system where IL-2Rβ would be selectively deleted in iNKT cells. The zinc finger transcription factor PLZF is expressed in iNKT cells but absent in conventional T cells (29, 30). Thus, we bred PLZF-Cre transgenic mice (PLZF^Cre) with Il2rb ^fl/fl mice (21, 23) to generate Il2rb ^fl/flPLZF^Cre mice and to specifically delete IL-2Rβ in iNKT cells. In such Il2rb ^fl/flPLZF^Cre mice, the thymic development of conventional T cells was unaffected ( Supplementary Figure 3A ). Both TCRβ expression and the CD4 versus CD8 profiles of Il2rb ^fl/flPLZF^Cre thymocytes remained comparable to those of littermate Il2rb ^fl/flPLZF^WT thymocytes ( Supplementary Figure 3A ). Contrary to our expectation, however, the frequency and number of thymic iNKT cells also remained unaltered in Il2rb ^fl/flPLZF^Cre mice ( Figure 2A ).

To understand these results, we considered two possibilities. Firstly, IL-2Rβ expression in the thymus could be dispensable for iNKT cell generation. Alternatively, the PLZF-Cre-mediated deletion of IL-2Rβ could be incomplete. In the latter case, the variegated expression of the PLZF-Cre or inefficient excision of the floxed allele could result in the retention of IL-2Rβ expression. To discriminate these possibilities, it would be necessary to monitor the deletion efficiency of IL-2Rβ in Il2rb ^fl/flPLZF^Cre mice. Notably, the Il2rb-floxed allele (Il2rb ^fl) is engineered to induce the expression of Green Fluorescent Proteins (GFP) upon IL-2Rβ deletion (21), permitting the identification of iNKT cells that have deleted IL-2Rβ. Employing this feature, we next analyzed the efficiency of IL-2Rβ deletion by assessing GFP expression. Among HSA^hi immature iNKT cells (ST0), GFP was only found in a very small fraction of iNKT cells ( Figure 2B , top histogram). Among HSA^lo iNKT cells, however, a large fraction of cells turned out to be GFP⁺, reporting the successful deletion of the Il2rb ^fl allele in mature iNKT cells ( Figure 2B , bottom histogram, and Supplementary Figure 3B ). Assessing the surface IL-2Rβ expression on GFP⁺ and GFP-negative mature iNKT cells showed a substantial loss of surface IL-2Rβ on GFP⁺ cells ( Figure 2C and Supplementary Figure 3C ). Thus, GFP⁺ cells correspond to iNKT cells that have terminated IL-2Rβ expression. Importantly, the frequencies of such GFP⁺ iNKT cells increased upon their further differentiation into ST1, ST2, and ST3 iNKT cells ( Figure 2D ). These results suggested that PLZF^Cre- mediated deletion of IL-2Rβ is initiated in ST0 cells but not accomplished until later stages of iNKT cell development. Because of this delayed deletion, significant amounts of surface IL-2Rβ remained on GFP⁺ iNKT cells which could provide residual IL-2R signaling ( Figure 2C ). Along these lines, we noted that a significant fraction of HSA^lo mature iNKT cells remained GFP-negative and fully expressed IL-2Rβ ( Figure 2B bottom). Such inefficient deletion could explain why we did not find significant differences in the frequency and number of thymic iNKT cells between Il2rb ^fl/flPLZF^Cre and littermate control mice ( Figure 2A ). We also did not find any significant changes in the ST1-ST3 distribution of GFP⁺ iNKT cells which expressed dramatically lower amounts of IL-2Rβ compared to GFP-negative iNKT cells ( Figure 2E ). We also did not find differences in the ST1-3 distribution of Il2rb ^fl/flPLZF^Cre mice compared to Il2rb ^fl/flPLZF^WT littermates that express normal amounts of IL-2Rβ ( Supplementary Figure 3D ). Along these lines, the thymic iNKT subset composition and the frequency of T-bet⁺ iNKT cells which correspond to ST3 cells were virtually unaltered in Il2rb ^fl/flPLZF^Cre mice ( Supplementary Figures 4A, B ). On the other hand, the protein abundance of Bcl-2, a direct downstream molecule of IL-2Rβ signaling (31), was markedly diminished ( Figure 2F ), indicating that the lack of IL-2Rβ had physiological consequences for iNKT cells. Nonetheless, the survival of GFP⁺ thymic iNKT cells remained unaffected, because Bcl-xL and not Bcl-2 provides pro-survival signals for iNKT cells (11). In agreement the frequency and number as well as the subset composition of iNKT cells in the spleen did not differ between Il2rb ^fl/flPLZF^Cre and Il2rb ^fl/flPLZF^WT mice ( Supplementary Figure 4C ). Collectively, the undisturbed differentiation of GFP⁺ iNKT cells in Il2rb ^fl/flPLZF^Cre mice suggested that the late-stage deletion of IL-2Rβ is not detrimental for the ST1 to ST3 iNKT cell maturation.

To further delineate the role of IL-2Rβ in the generation and differentiation of iNKT cells, we next analyzed the development and subset composition of thymic iNKT cells in germline IL-2Rβ-deficient (Il2rb ^–/–) mice (22). In Il2rb ^–/– mice, all thymocytes - including ST0 immature iNKT cells - are devoid of IL-2Rβ. Thus, an IL-2Rβ requirement for iNKT cells can be assessed starting at the earliest precursor stages of iNKT cell development. Notably, Il2rb ^–/– mice are autoimmune because of the impaired generation of immunosuppressive Foxp3⁺ Treg cells that require IL-2 signaling for their maturation (32, 33). Consequently, we analyzed Il2rb ^–/– mice before 6 weeks of age to avoid the potential skewing of iNKT cell differentiation due to autoimmunity. We confirmed that the thymus of Il2rb ^–/– mice at that age did not show an overtly activated phenotype, and that T cell development remained comparable to that of WT littermate mice ( Supplementary Figure 5A ). We further confirmed the lack of in vivo IL-2Rβ signaling in Il2rb ^–/– mice by their absence of mature CD25⁺Foxp3⁺ Treg cells which require IL-2R signaling for their generation ( Supplementary Figure 5B ) (32, 33). Consistent with previous reports (18, 19), we confirmed a dramatic reduction in the frequency and number of thymic iNKT cells of Il2rb ^–/– mice compared with those of WT littermate mice ( Figure 3A ). Notably, the loss was specific for iNKT cells and did not affect conventional αβ T cells, resulting in a dramatically decreased ratio of iNKT cells over αβ T cells ( Supplementary Figure 5C ).

Most thymic iNKT cells in C57BL/6 background mice correspond to ST3 cells. Therefore, we expected that IL-2Rβ-deficiency would selectively affect the differentiation of ST3 iNKT cells. Indeed, Il2rb ^–/– thymic iNKT cells were developmentally arrested, resulting in the accumulation of ST1/ST2 iNKT cells and in a substantial decrease in ST3 iNKT cells ( Figure 3B ). We further found that the fraction of NKT1 cells was significantly diminished among Il2rb ^–/– iNKT cells, which in turn resulted in the relative overrepresentation of NKT2 and NKT17 cells ( Figure 3C ). The iNKT cell number for each subset, however, were significantly reduced ( Figure 3C ), which was consistent with the dramatic decrease in total iNKT cell numbers in Il2rb ^–/– mice ( Figure 3A ). Collectively, these results from Il2rb ^–/– mice documented a major role for the cytokine receptor IL-2Rβ in the generation, but also in the subset differentiation of thymic iNKT cells.

Because IL-2Rβ is important for NKT1 cell generation ( Figure 3C ), we next asked whether IL-2Rβ would be also sufficient to impose NKT1 lineage fate during iNKT cell development. To this end, we assessed iNKT cell development in IL-2Rβ-transgenic mice (IL-2Rβ^Tg) that overexpress mouse IL-2Rβ under the control of human CD2 promoter/enhancer elements (20, 34). In these animals, the transgenic IL-2Rβ is prematurely expressed on preselection thymocytes ( Supplementary Figure 6A ) and further overexpressed on postselection mature T cells ( Supplementary Figure 6A ). Consequently, the IL-2Rβ is abundantly expressed on both immature and mature thymocytes of IL-2Rβ^Tg mice. Among iNKT cells, immature HSA^hiCD44^– iNKT cells (ST0) normally do not express IL-2Rβ ( Supplementary Figure 6B ). However, the ST0 iNKT cells in IL-2Rβ^Tg mice displayed substantially increased abundance of IL-2Rβ ( Supplementary Figure 6B ). Importantly, such premature expression of IL-2Rβ dramatically impaired the generation of iNKT cells, so that the frequency and number of thymic iNKT cells in IL-2Rβ^Tg mice were markedly reduced ( Figure 4A ). These results suggested that IL-2 receptor signaling in immature iNKT is detrimental to the generation of thymic iNKT cells. These findings also indicated that the timing of IL-2Rβ expression during iNKT cell development needs to be carefully controlled, presumably to protect immature iNKT cells from premature IL-2 or IL-15 signaling.

Because IL-2Rβ is associated with NKT1 cell differentiation, we next aimed to interrogate if premature IL-2Rβ expression in immature ST0 iNKT cells would affect the development of other iNKT subsets. In C57BL/6 mice, most thymic iNKT cells correspond to NKT1 cells (10), making it difficult to discern how the IL-2Rβ^Tg would affect NKT2 and NKT17 cells which are underrepresented in this mouse strain. BALB/c mice differ from C57BL/6 mice in their thymic iNKT subset composition as it is enriched in NKT2 and NKT17 cells and reduced in NKT1 cells compared with that in C57BL/6 mice (10). Thus, we considered BALB/c mice as an appropriate model to assess the effects of premature IL-2Rβ expression on NKT2 and NKT17 cell generation. To this end, we backcrossed the IL-2Rβ^Tg onto the BALB/c background and analyzed the development and differentiation of thymic iNKT cells in these mice.

As expected, the thymocytes of IL-2Rβ^Tg BALB/c mice expressed substantially increased amounts of surface IL-2Rβ compared with the thymocytes of littermate mice ( Supplementary Figure 6C ). Such increase in the IL-2Rβ abundance was also observed in HSA^hiCD44^– (ST0) immature iNKT cells ( Supplementary Figure 6D ), but curiously not in HSA^lo mature iNKT cells ( Supplementary Figure 6D ). These results suggested that the expression of the IL-2Rβ^Tg is either silenced in HSA^lo iNKT cells or that the premature expression of IL-2Rβ in HSA^hiCD44^– (ST0) immature iNKT cells is detrimental for iNKT cell generation so that only iNKT cells that have escaped the transgene expression can mature. In fact, we found that the frequency and number of thymic iNKT cells were indeed dramatically reduced in IL-2Rβ^Tg BALB/c mice ( Figure 4B ). Thus, consistent with our findings in IL-2Rβ^Tg C57BL/6 mice ( Figure 4A ), the forced expression of IL-2Rβ suppresses the generation of thymic iNKT cells in BALB/c mice also ( Figure 4B ).

To understand the effects of IL-2Rβ on iNKT cell differentiation, we next examined the iNKT subset composition in IL-2Rβ^Tg BALB/c mice. Surprisingly, NKT1 cells, as identified by T-bet expression, were virtually absent in IL-2Rβ^Tg BALB/c thymocytes. Both the frequency and number of T-bet⁺ NKT1 cells were dramatically decreased in these mice ( Figures 4C, D ). Conversely, NKT2 and NKT17 cell frequencies were significantly increased in IL-2Rβ^Tg BALB/c mice compared with those of WT littermate controls ( Figures 4C, D ). Thus, transgenic IL-2Rβ selectively inhibited the development of NKT1 cells without suppressing the generation of NKT2 and NKT17 cells.

To further examine how IL-2Rβ^Tg would interfere with the induction of T-bet and impair NKT1 cell generation, we employed the T-bet-ZsGreen reporter mouse (TBGR^Tg) to monitor T-bet transcription in IL-2Rβ^Tg and WT BALB/c mice ( Figure 5A ) (25). The TBGR^Tg is designed to express ZsGreen reporter proteins from the Tbx21 gene locus which encodes for T-bet so that ZsGreen expression reflects Tbx21 transcription (25). Indeed, we identified a substantial population of ZsGreen⁺ cells among thymic iNKT cells of TBGR^Tg BALB/c mice ( Figure 5A ), and these ZsGreen⁺ iNKT cells corresponded to NKT1 cells based on their intracellular PLZF and T-bet protein expression ( Figure 5B , left). In contrast, TBGR^Tg IL-2Rβ^Tg BALB/c failed to generate ZsGreen⁺T-bet⁺ iNKT cells ( Figure 5B , right), affirming the detrimental effect of IL-2Rβ^Tg on NKT1 cell generation.

Of note, we found in both WT and IL-2Rβ^Tg thymocytes a distinct population of iNKT cells with intermediate amounts of ZsGreen and T-bet expression, i.e. ZsGreen^intT-bet^int ( Figure 5B , cells gated in blue), whose developmental status was unclear to us. To further characterize these ZsGreen^intT-bet^int iNKT cells, we divided TBGR^Tg BALB/c iNKT cells into 3 populations, i.e. I, II, and III, based on their progressive increase in ZsGreen reporter expression. Accordingly, the ZsGreen^intT-bet^int cells corresponded to population II, while ZsGreen^hiT-bet^hi cells were defined as population III ( Figure 5C ). Thymic iNKT cells that did not express ZsGreen were referred to as population I. Terminally differentiated NKT1 cells, as identified as population III, expressed large amounts of T-bet and CD44 ( Figure 5C and Supplementary Figure 6E ) (11). ZsGreen^intT-bet^int cells (population II), on the other hand, expressed significantly smaller amounts of T-bet and CD44 but showed greater abundance of TCRβ. Importantly, ZsGreen^intT-bet^int cells had not yet induced CD122 expression, suggesting that they are NKT1-lineage committed but not fully differentiated NKT1 cells ( Figures 5C, D and Supplementary Figure 6E ) (35). Along these lines, the ZsGreen^intT-bet^int population did not contain RORγt⁺ iNKT cells ( Figure 5D ), indicating that NKT17 cell differentiation is branched off before the appearance of ZsGreen^intT-bet^int population II iNKT cells. Thus, we identified a new subpopulation of thymic iNKT cells, i.e. ZsGreen^intT-bet^int iNKT cells, that are NKT1-lineage committed but are not fully differentiated into mature NKT1 cells. Because ZsGreen⁺T-bet⁺ iNKT cells were missing in IL-2Rβ^Tg thymocytes, these results further suggest that the forced and premature expression of IL-2Rβ interferes with the terminal differentiation of T-bet⁺ NKT1 cells but did not inhibit the induction of T-bet expression itself. Collectively, these results unveil an unexpected negative effect of IL-2Rβ on iNKT subset differentiation.

To understand how IL-2Rβ overexpression would affect thymic iNKT cell generation, we considered two alternative but not mutually exclusive hypotheses. First, the increased abundance of IL-2Rβ proteins in IL-2Rβ^Tg mice could increase the binding and consumption of IL-15, thus diminishing the availability of IL-15 for iNKT cell development. IL-15 is critical for the generation of iNKT cells in general and specifically for NKT1 cells (11, 36). Consequently, the detrimental effect of IL-2Rβ^Tg on iNKT cells could have been due to insufficient IL-15 availability, partly phenocopying the effect of IL-15-deficiency (37). In a second scenario, we focused on the timing of IL-2Rβ expression. IL-2Rβ is normally not induced in immature iNKT cells but forcibly expressed in IL-2Rβ^Tg mice. Premature IL-2Rβ signaling could interfere with iNKT cell generation and specifically with NKT1 cell differentiation, albeit the molecular mechanism would be unclear. To discriminate between these possibilities, we implanted IL-2Rβ^Tg mice with osmotic pumps that release recombinant IL-15 to supply excess amounts of IL-15 in vivo ( Figure 6A ). After 10 days of IL-15 pump installation in IL-2Rβ^Tg mice, we found a dramatic increase in spleen size and increased frequencies of NK cells whose maintenance depends on IL-15 (38) ( Figure 6A and Supplementary Figure 7A ). Moreover, the frequency of CD8 T cells and specifically the number of CD44^hi CXCR3⁺ memory-phenotype CD8 T cells were substantially increased in mice that were implanted with IL-15- but not with PBS-releasing pumps ( Figure 6A and Supplementary Figure 7B ). These results were consistent with the effects of increased IL-15 availability in vivo (39). The generation of thymic iNKT cells in IL-2Rβ^Tg BALB/c mice, however, remained unaffected by the IL-15 infusion ( Figure 6B ), and the iNKT subset composition also remained unaltered ( Figures 6C, D ). Thus, despite supplemented with excess amounts of IL-15, the IL-2Rβ^Tg mice were still impaired in thymic iNKT cell differentiation, with both dramatically reduced frequencies and numbers of NKT1 cells ( Figures 6C, D ). Along these lines, the thymic generation of Foxp3⁺ Treg cells which depend on IL-2 and IL-2 receptor signaling (33) was unaffected in IL-2Rβ^Tg mice ( Supplementary Figure 7C ), which further supports that the availability of intrathymic cytokines remains unaltered by IL-2Rβ overexpression. Altogether, these results suggested that it is unlikely that increased consumption and reduced availability of in vivo IL-15 would account for the lack of NKT1 cells in IL-2Rβ^Tg mice.

To gain further mechanistic insights into whether and how premature IL-2Rβ expression would impair iNKT cell generation, we considered that STAT5 phosphorylation is a major downstream event of IL-2Rβ signaling (40). The transcription factor STAT5 plays a critical role in T cell development as it induces the expression of prosurvival molecules, such as Bcl-2, and controls the activity of several key transcription factors in T cell differentiation (40, 41). Because immature ST0 iNKT cells lack IL-2Rβ ( Figure 1C ), we postulated that ST0 iNKT cells of WT mice would not induce phospho-STAT5 (pSTAT5) upon IL-15 stimulation. ST0 iNKT cells in IL-2Rβ^Tg mice, on the other hand, would have induced substantial amounts of pSTAT5 due the premature expression and signaling of IL-2Rβ.

If it would be the premature STAT5 activation which suppresses iNKT cell generation, we hypothesized that a constitutively active STAT5b transgene (CA-STAT5^Tg) would also impair iNKT cell generation (24). Indeed, we found that both the frequency and number of thymic iNKT cells were dramatically reduced in CA-STAT5^Tg mice compared to those in WT littermate controls ( Figure 6E ). Moreover, assessing the composition of thymic iNKT subsets in CA-STAT5^Tg mice ( Supplementary Figure 8A ) demonstrated that the generation of NKT1 cells was impaired by ectopic STAT5 activation ( Figure 6F ). T cell development in general, however, was not negatively affected by the constitutive activation of STAT5 ( Supplementary Figure 8B ). As such, the cell numbers of preselection DP thymocytes remained unaffected ( Figure 6F ), and the generation of memory-phenotype CD8 T cells was rather increased in CA-STAT5^Tg thymocytes ( Supplementary Figure 8C ). Thus, the detrimental effect of STAT5 activation was limited to the generation of iNKT cells. Collectively, these results identify and demonstrate STAT5 as a major player acting downstream of IL-2Rβ to control the development of thymic iNKT cells.