Semi-invariant NKT cells are thymus-derived innate-like lymphocytes that modulate microbial and tumor immunity as well as autoimmune diseases. These immunoregulatory properties of NKT cells are acquired during their development. Much has been learned regarding the molecular and cellular cues that promote NKT cell development, yet how these cells are maintained in the thymus and the periphery and how they acquire functional competence are incompletely understood. We found that IL-15 induced several Bcl-2 family survival factors in thymic and splenic NKT cells in vitro. Yet, IL-15–mediated thymic and peripheral NKT cell survival critically depended on Bcl-xL expression. Additionally, IL-15 regulated thymic developmental stage 2 to stage 3 lineage progression and terminal NKT cell differentiation. Global gene expression analyses and validation revealed that IL-15 regulated Tbx21 (T-bet) expression in thymic NKT cells. The loss of IL-15 also resulted in poor expression of key effector molecules such as IFN-γ, granzyme A and C, as well as several NK cell receptors, which are also regulated by T-bet in NKT cells. Taken together, our findings reveal a critical role for IL-15 in NKT cell survival, which is mediated by Bcl-xL, and effector differentiation, which is consistent with a role of T-bet in regulating terminal maturation.

Natural killer T cells are thymus-derived lymphocytes whose functions are regulated by foreign or self-glycolipid ligands presented by CD1d molecules. Most NKT cells recognize Ag with a semi-invariant TCR consisting of an invariant mouse Vα14Jα18 (human Vα24Jα18) α-chain that pairs with a Vβ8.2 (human Vβ11), Vβ7, or Vβ2 β-chain. Upon agonistic ligand recognition, NKT cells rapidly secrete large amounts of immunoregulatory cytokines and upregulate costimulatory molecules to initiate and modulate the effector functions of myeloid and lymphoid cells. By doing so, NKT cells regulate microbial and tumor immunity as well as autoimmune diseases (1, 2). NKT cells acquire these immunoregulatory properties during ontogeny through poorly understood processes. NKT cell frequency varies both between mouse strains, the highest in BALB/c (1–2 million in thymus, liver, and spleen) and lowest in the autoimmune disease-predisposed strains NOD and SJL (0.1–0.3 million), and between healthy human individuals, ranging from 0.01–5.0% in peripheral blood (3). Hence, understanding how NKT cells develop, regulate their frequency, and acquire functional potential is critical for designing NKT cell-based therapies.

NKT cells develop in the thymus and acquire markers typical of NK and activated/memory T cells as they mature (3). NKT cell precursors (CD24hiCD69+CD44NK1; stage [st]0) emerge from CD4 and CD8 double-positive (DP) thymocytes soon after expression of the semi-invariant TCR in a process requiring three transcription factors: HEB, a basic helix-loop-helix E protein; retinoic acid-related orphan receptor γt; and c-Myb (47). Precursor st0 cells proliferate extensively in a c-Myc–dependent manner (8, 9) and undergo positive selection through heterotypic interactions between the semi-invariant TCR and CD1d molecules expressed by DP thymocytes (1, 3). Further development and differentiation into st1 (CD24CD44NK1.1), st2 (CD24CD44+NK1.1), and to fully functional st3 (CD24CD44+NK1.1+) NKT cells require transcriptional regulation by promyelocytic leukemia zinc finger (PLZF) (10, 11), c-Myb (7), NF-κB/Rel-A (1215), Egr-2 (16), Gata-3 (1719), and T-bet (17, 20), as well as sustained homotypic interactions between the signaling lymphocyte activation molecule (SLAM) family members Slamf1 and Slamf6 (21) by a mechanism involving the SAP and Fyn signaling module (2226). Furthermore, effector differentiation, which is ontogenetically coupled in NKT cells, is regulated by Itk (27), NF-κB (28), T-bet (20, 29), Egr-2 (16), Th-POK (30), and CSF-2 (31). Upon lineage commitment and the emergence of st0 NKT precursors and through subsequent stages of differentiation and maturation, thymic and peripheral NKT cells are maintained by complex yet poorly understood homeostatic mechanisms involving NF-κB (28), RelA (15), CYLD (32), c-Myc (8, 9), CD122 (33), γc (34), IL-7Rα (34), IL-15 (35, 36), IL-15Rα (37, 38), IRF-1 (39), calcineurin (16), Egr-2 (16), and Id2 (40).

During differentiation, NKT cells acquire the ability to produce IL-4 (st1), IL-4 and IFN-γ (st1 and st2), or IL-4lo and IFN-γhi (st3) in response to TCR stimulation (27, 41), and secretion of these cytokines is regulated by CSF-2 (31). At st2 and st3, NKT cells also acquire memory T cell markers (CD44hi, CD62Llo) as well as NK cell markers (NKG2D, Ly49, and NK1.1) and cytolytic function (granzyme B and perforin) (1, 42). Although a substantial fraction of st2 cells remains in the thymus and matures to st3, a small fraction egresses from the thymus and immigrates into the spleen and liver, whereupon further maturation occurs (41, 43, 44). Migration to these peripheral sites depends on CXCR6 (45), whereas continued development and maintenance in the liver require LFA-1 (46) and Id2 (40). Even though mature thymic and splenic NKT cells are phenotypically similar, they behave as distinct functional populations at these sites because the developmental cues and perhaps the agonistic ligands are distinct at different sites (44, 47, 48). Thus, unlike conventional T lymphocytes, which undergo complete ontogenetic maturation in the thymus but commit to effector differentiation only upon Ag recognition in the periphery, NKT cells complete both these processes within and outside the thymus. The instructive signals that differentially regulate the maturation and acquisition of effector functions of NKT cells remain poorly understood.

IL-15 is a pleiotropic cytokine that regulates the development and maintenance of several subsets of innate-like lymphocytes, including γδ T, CD8αα T, NK, and NKT cells (49). Furthermore, IL-15 signals the ontogeny, effector differentiation, and Mcl-1–dependent survival of NK cells (50). Moreover, thymic and peripheral NKT cells develop poorly in IL-15–deficient (IL-150) mice (35, 36). In the periphery, this is due to impaired NKT cell homeostatic proliferation (36). In the thymus, NKT cell proliferation appears intact, yet IL-150 mice develop a severe blockade in NKT cell ontogeny at st2, with leaky progression to st3 (36). What role IL-15 signals play during thymic NKT cell ontogeny and whether IL-15 is essential for NKT cell functional maturation in the periphery remain unknown.

In this study, we provide evidence that thymic and peripheral NKT cell survival critically depends on IL-15. One important downstream target of IL-15 appears to be Bcl-xL. Furthermore, we describe a new function for IL-15 by which it regulates st2 to st3 lineage progression and terminal NKT cell maturation, which is consistent with a role of T-bet in regulating st2 to st3 transition and terminal differentiation.

B6–IL-150, B6.129–IL-7α0, B6.Jα180, and B6.129-CD1dTD mice (34, 35, 5153) were generous gifts from J. Peschon (Immunex), M. Taniguchi (RIKEN), and A. Bendelac (University of Chicago), respectively. B6–IL-150 and B6.129–IL-7Rα0 mice were bred with B6-Eμ36–Bcl-2tg (54) and C-Lck–Bcl-xLtg (55) mice. B6.C–IL-150;Bcl-xLtg mice were backcrossed to B6–IL-150 mice for six generations. All mouse crosses and experiments were performed under approval by Institutional Animal Care and Use Committee protocols.

All Abs used for the identification of NKT cells were from BD Pharmingen as described (56). Anti-BrdU–Alexa 647 (PRB-1; Invitrogen); anti-human Bcl-xL–PE and IgG3 isotype control-PE (7B25 and B10; Southern Biotechnology Associates); anti-mouse Bcl-2–PE (3F11) and Armenian hamster IgG1κ isotype control-PE (A19-3), anti-human Bcl-2–PE (6C8) and Armenian hamster IgG2λ isotype control-PE (Ha4/8), anti–Mcl-1 rabbit polyclonal (Rockland); chicken anti-rabbit–Alexa 647 (Invitrogen); anti–T-bet–Alexa 647 (eBio4B10) and mouse IgG1 isotype control-Alexa 647 (P3; eBioscience); anti–Gata-3–Alexa 647 (L50-823) and mouse IgG1κ isotype control-Alexa 647 (MOPC-21; Santa Cruz Biotechnology); and mouse IgG1 isotype control and anti-mouse IgG-FITC (eBioscience) Abs were purchased. α-galactosylceramide (αGalCer; KRN7000) was generously provided by Kirin Brewery Company (Gunma, Japan) or purchased from Funakoshi (Tokyo, Japan). Mouse αGalCer-loaded CD1d monomers were obtained from the National Institutes of Health Tetramer Facility (Emory University). Preparation of αGalCer-loaded CD1d tetramers from monomer and their use are described elsewhere (56).

Abs and staining procedures for the identification of NKT cells were as described elsewhere (56). NKT cells were identified as CD3ε+tetramer+ cells among B220lo splenocytes and hepatic mononuclear cells or CD8lo thymocytes. Four-color flow cytometry was performed with a FACSCalibur instrument (BD Biosciences), whereas seven-color flow cytometry was performed with an LSR II instrument (BD Biosciences). Data were analyzed with FlowJo software (Tree Star). Absolute NKT cell numbers were calculated from the percentage of tetramer+ cells and total number of cells recovered from each organ.

Thymocytes and splenocytes were stimulated in vitro with 100 ng/ml recombinant human (rh)IL-15 (Miltenyi Biotec) for 5 d. Cells were then collected and stained with specific mAb for intracellular expression of Bcl-2, Bcl-xL, and Mcl-1.

Individual age-matched (6–10-wk-old) mice were injected i.p. with 5 μg αGalCer or vehicle (0.1% Tween-20 in PBS) as controls. After 4 h, serum, splenocytes, and hepatic mononuclear cells were prepared. Cells were stained with anti-CD3ε mAb, tetramer, and anti–IFN-γ or –IL-4 mAb and analyzed by flow cytometry. Sandwich ELISA was performed as described (56) to monitor serum cytokine response.

Cells were fixed and permeabilized with Cytofix/Cytoperm solution for intracellular staining or with Cytofix/Cytoperm Plus (both from BD Pharmingen) for intranuclear BrdU staining according to the manufacturer’s protocol. Staining for intracellular IFN-γ and IL-4 also used GolgiStop (BD Pharmingen). Staining for intracellular T-bet and Gata-3 used the Foxp3 staining buffer kit (eBioscience). Intracellular and intranuclear staining was performed according to the manufacturer’s protocol.

Thymic NKT cells were enriched by magnetic sorting using a Pan-T cell isolation kit in combination with CD8 beads (Miltenyi Biotec) as instructed by the manufacturer. Enriched NKT cells from 5–10 mice were pooled and further purified by flow sorting CD3ε+tetramer+ cells (FACSAria; BD Biosciences). CD4 and CD8 DP thymocytes were flow sorted as above. Freshly purified NKT cells and DP thymocytes were >98% pure as judged by CD3ε+tetramer+ and CD4/CD8-specific staining (data not shown).

Freshly purified NKT cells were washed and lysed by passage through a Qiashedder column (Qiagen). Total RNA was extracted per the manufacturer’s instructions (RNeasy; Qiagen). RNA yield was quantified spectrophotometrically (Nanodrop ND-1000; Nanodrop) and aliquots electrophoresed to determine sample purity and concentration (Agilent 2100 Bioanalyzer; Agilent Technologies). Microarray hybridizations were performed using one-color hybridization on a SurePrint G3 Mouse GE 60K Microarray (Agilent Technologies) by the Functional Genomics and Shared Resources Core (Vanderbilt University). Scanning was performed using a GenePix 4000B Microarray Scanner (Molecular Devices). The resulting images were processed with Agilent Feature Extraction software (Agilent Technologies) to generate raw probe intensity data, which was subsequently normalized by the quantile method (57). Differential expression analysis was carried out with an empirical Bayes approach on a linear model using the limma package (58, 59). In this study, we defined differentially expressed genes as those for which the log2 fold change values were ≥1.5 or ≤−1.5 and for which nominal p values were <0.001 (60). Next, we performed hierarchical clustering of the differentially expressed genes using the R AMAP Package (http://cran.r-project.org/web/packages/amap/). The microarray data are available at the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=dtgxpkoosiokmpm&acc=) under accession number GSE32568.

RNA was reverse transcribed with MuLV RT Reverse Transcriptase (Applied Biosystems) according to the manufacturer’s protocol. Quantitative PCR (qPCR) was performed in the presence of IQ RealTime Sybr Green PCR Supermix and target gene-specific primers (Supplemental Table I) in an iQ5 thermocycler (Bio-Rad). Results were analyzed using iQ5 software (Bio-Rad). Data were first normalized against β-actin control. As the threshold cycle (CT) values are obtained as log2 values, the normalized Δ CT values of wild-type (wt) samples were subtracted from normalized Δ CT values of IL-150 samples. The resulting log2 ΔΔ CT values were transformed yielding relative expression, which enabled comparison between NKT cells isolated from the two strains. For this, the IL-150 sample, called the Calibrator, was set to 1 and relative expression calculated as fold change in gene expression in wt NKT cells relative to the Calibrator.

cDNA was prepared from total RNA isolated from DP thymocytes. qPCR for Vα14 to Jα18 rearrangement was performed as described above using the following forward (F) and reverse (R) primers: Vα14F (5′-AGGTCTTGTGTCCCTGACAG-3′), Jα18R (5′-CAGGTATGACAATCAGCTGAGTCC-3′), Vα8F (5′-TCACAGACAACAAGAGGACC-3′), and Jα5R (5′-AGTGAGCTGCCCCACAACCT-3′). PCR products were separated by agarose gel electrophoresis and visualized by ethidium bromide staining.

Comparisons of normally distributed continuous data were performed by one-way ANOVA with Tukey’s posttest to determine significance. The significance of gene expression analyzed by qPCR was determined by unpaired t test. Statistical significance of Vα14-Jα18 rearrangement by qPCR was calculated by two-tailed, unpaired t test. Comparisons were performed using Prism software (GraphPad).

Previous reports demonstrated that NKT cell frequency is significantly lower in mice lacking IL-15 compared with wt mice due to altered NKT cell homeostasis within the thymus and spleen (35, 36). Although IL-15 regulates proliferation of splenic NKT cells, its role in homeostasis of the thymic subset has remained unclear (36). A re-evaluation confirmed that NKT cell frequency was altered in the absence of IL-15, and, hence, their numbers were reduced in the thymus, spleen, and liver (Fig. 1A). Likewise, in vivo BrdU incorporation experiments revealed that splenic but not thymic NKT cell proliferation was impaired in IL-150 mice (Fig. 1B). These data indicate that, in addition to its role in cell proliferation, IL-15 plays another role(s) during thymic NKT cell ontogeny.

FIGURE 1.

Defective NKT cell development and maintenance in IL-150 mice. A, Thymic, splenic, and hepatic NKT cells from B6 (n = 13) and B6–IL-150 (n = 9) mice were identified as CD3ε+tetramer+ cells within electronically gated CD8LO thymic, B220LO splenic, or liver mononuclear cells. Numbers are percent NKT cells among total leukocytes within each organ. Data are representative of nine independent experiments. B, B6 and B6–IL-150 mice were injected i.p. with 2 mg BrdU daily for 3 consecutive d and sacrificed 1 d later. BrdU incorporation was determined by flow cytometry after extracellular lineage-specific staining and intracellular labeling with anti-BrdU Alexa 647 mAb. Overlaid histograms are of NKT cells identified as in A from vehicle (shaded) or BrdU injected (open) mice. The histograms show the expression levels of BrdU within NK1.1NEG (left panels) and NK1.1+ (right panels) NKT cells within the thymus and spleen. Data are representative of three independent experiments. n = 6. C and D, Thymic and splenic mononuclear cells from B6 and B6–IL-150 mice were assessed for intracellular expression of Bcl-2, Bcl-xL (C), and Mcl-1 (D) detected with specific mAb within electronically gated B220LOCD3ε+tetramer+ cells. Numbers refer to the difference in mean fluorescence intensity (ΔMFI) between isotype control (shaded) and specific mAb (open) staining. Data are representative of three independent experiments. n = 5. The flow data were acquired on two different instruments (FACSCalibur and LSR-II; BD Biosciences) in different experiments. Even though the actual MFI varied, the represented trend remained the same nonetheless. Therefore, we have chosen to represent the ratio of expression between wt and IL-150 NKT cells. The ratio of Bcl-2, Bcl-xL, and Mcl-1 expression by IL-150 (KO) and wt NKT cells is shown. A ratio of ∼1 indicates no difference in expression; a ratio <1 indicates lower expression in IL-150 NKT cells.

FIGURE 1.

Defective NKT cell development and maintenance in IL-150 mice. A, Thymic, splenic, and hepatic NKT cells from B6 (n = 13) and B6–IL-150 (n = 9) mice were identified as CD3ε+tetramer+ cells within electronically gated CD8LO thymic, B220LO splenic, or liver mononuclear cells. Numbers are percent NKT cells among total leukocytes within each organ. Data are representative of nine independent experiments. B, B6 and B6–IL-150 mice were injected i.p. with 2 mg BrdU daily for 3 consecutive d and sacrificed 1 d later. BrdU incorporation was determined by flow cytometry after extracellular lineage-specific staining and intracellular labeling with anti-BrdU Alexa 647 mAb. Overlaid histograms are of NKT cells identified as in A from vehicle (shaded) or BrdU injected (open) mice. The histograms show the expression levels of BrdU within NK1.1NEG (left panels) and NK1.1+ (right panels) NKT cells within the thymus and spleen. Data are representative of three independent experiments. n = 6. C and D, Thymic and splenic mononuclear cells from B6 and B6–IL-150 mice were assessed for intracellular expression of Bcl-2, Bcl-xL (C), and Mcl-1 (D) detected with specific mAb within electronically gated B220LOCD3ε+tetramer+ cells. Numbers refer to the difference in mean fluorescence intensity (ΔMFI) between isotype control (shaded) and specific mAb (open) staining. Data are representative of three independent experiments. n = 5. The flow data were acquired on two different instruments (FACSCalibur and LSR-II; BD Biosciences) in different experiments. Even though the actual MFI varied, the represented trend remained the same nonetheless. Therefore, we have chosen to represent the ratio of expression between wt and IL-150 NKT cells. The ratio of Bcl-2, Bcl-xL, and Mcl-1 expression by IL-150 (KO) and wt NKT cells is shown. A ratio of ∼1 indicates no difference in expression; a ratio <1 indicates lower expression in IL-150 NKT cells.

Close modal

Previous studies have suggested that IL-15 induces antiapoptotic Bcl-2 family members, including Bcl-2, Bcl-xL, and Mcl-1, to prevent apoptosis and regulate homeostasis in vivo (49). For example, IL-15 promotes the survival of NK cells by inducing the survival factor Mcl-1 (50). An examination of the expression of the Bcl-2 family of survival factors by flow cytometry revealed that intracellular Bcl-2 expression was dramatically reduced in IL-150 thymic and splenic NKT cells (Fig. 1C). Although not as dramatic, intracellular Bcl-xL and Mcl-1 expression levels were also reduced either in thymic (Bcl-xL) or in both thymic and splenic (Mcl-1) IL-150 NKT cells when compared with wt NKT cells (Fig. 1C, 1D).

To test whether IL-15 induces survival factors in NKT cells, wt thymocytes and splenocytes were cultivated in the absence or presence of exogenous rhIL-15 for 5 d. Bcl-2, Bcl-xL, and Mcl-1 expression in fresh or cultivated NKT cells (with or without rhIL-15) was evaluated using flow cytometry. Both fresh NKT cells and those cultivated without rhIL-15 were included as controls because expression of these survival factors changes during the 5-d culture. The data revealed that IL-15 upregulated the expression of Bcl-2 and Bcl-xL but not Mcl-1 in thymic NKT cells to levels above those expressed by freshly isolated NKT cells (Fig. 2A, 2B). Very few NKT cells cultivated without IL-15 expressed the three survival factors, and, hence, when compared with this control, IL-15 also appeared to induce Mcl-1 expression in thymic NKT cells (Fig. 2B). IL-15 also modestly induced Bcl-2 and Bcl-xL expression within splenic NKT cells but not to the levels found in freshly isolated NKT cells; their expression was above control NKT cells cultured in medium without rhIL-15 (Fig. 2A). These data indicate that IL-15 has the potential to induce or maintain the expression of survival factors of the Bcl-2 family in thymic or splenic NKT cells, respectively, and, hence, to support NKT cell survival during development.

FIGURE 2.

IL-15 upregulates expression of the survival factors Bcl-2, Bcl-xL, and Mcl-1 within thymic and splenic NKT cells. Thymic and splenic mononuclear cells from B6 mice were cultivated in vitro in the absence or presence of 100 ng/ml of rhIL-15. After 5 d, intracellular expression of Bcl-2, Bcl-xL (A), and Mcl-1 (B) was detected with specific mAb within electronically gated B220LOCD3ε+tetramer+ cells. Numbers refer to the ΔMFI between isotype control (shaded) and specific mAb (open). Data are representative of three independent experiments. n = 5. As for Fig. 1, the flow data were acquired on two different instruments in different experiments. Hence, the ratio of IL-15–induced Bcl-2, Bcl-xL, and Mcl-1 expression within NKT cells in comparison with unstimulated, freshly isolated NKT cells is shown. A ratio of <1 indicates no induction; a ratio >1 indicates IL-15–induced expression.

FIGURE 2.

IL-15 upregulates expression of the survival factors Bcl-2, Bcl-xL, and Mcl-1 within thymic and splenic NKT cells. Thymic and splenic mononuclear cells from B6 mice were cultivated in vitro in the absence or presence of 100 ng/ml of rhIL-15. After 5 d, intracellular expression of Bcl-2, Bcl-xL (A), and Mcl-1 (B) was detected with specific mAb within electronically gated B220LOCD3ε+tetramer+ cells. Numbers refer to the ΔMFI between isotype control (shaded) and specific mAb (open). Data are representative of three independent experiments. n = 5. As for Fig. 1, the flow data were acquired on two different instruments in different experiments. Hence, the ratio of IL-15–induced Bcl-2, Bcl-xL, and Mcl-1 expression within NKT cells in comparison with unstimulated, freshly isolated NKT cells is shown. A ratio of <1 indicates no induction; a ratio >1 indicates IL-15–induced expression.

Close modal

We took a genetic approach to further investigate the role of IL-15–induced genes in the survival of thymic and peripheral NKT cells. Thus, Bcl-2 and Bcl-xl transgenes were independently introgressed into IL-150 mice. Analysis of the resulting mice at age 6–8 wk, when thymic and splenic cellularity is similar between the different strains, revealed that enforced Bcl-xL but not Bcl-2 overexpression rescued the frequency and absolute numbers of thymic, splenic, and hepatic NKT cells in IL-150 mice (Fig. 3A, 3B).

FIGURE 3.

Bcl-xL overexpression restores NKT cell development in IL-150 mice. A, Thymic, splenic, and hepatic NKT cells from B6 (n = 8), B6–IL-150 (n = 10), B6–IL-150;Bcl-2tg (n = 6), and B6–IL-150;Bcl-xLtg (n = 16) mice were identified as CD3ε+tetramer+ cells within electronically gated CD8LO thymic, B220LO splenic, or liver mononuclear cells. Numbers are percent NKT cells among total leukocytes. B, Absolute numbers of NKT cells in the thymus and spleen of B6, B6–IL-150, B6–IL-150;Bcl-2tg, and B6–IL-150;Bcl-xLtg mice were calculated from percent NKT cells in A and total cell count. Data are representative of six independent experiments showing mean + SEM; n, as in A. The p value was calculated by one-way ANOVA with Tukey’s posttest. C, Intracellular expression of the hBcl-xL and hBcl-2 transgene within double-negative, DP, CD4+CD8NEG, and CD8+CD4 thymocytes of B6.C-Bcl-xLtg (right panels) and B6.Bcl-2tg (left panels) mice. Overlaid histograms represent isotype control (shaded) and specific mAb (open). Data are representative of three independent experiments. n = 5. D, Thymic and splenic NKT cells from IL-7Rα+/− (n = 3), IL-7Rα0 (n = 3), IL-7Rα0;Bcl-2tg (n = 3), and IL-7Rα0;Bcl-xLtg (n = 2) mice were identified as CD3ε+tetramer+ cells within electronically gated CD8LO thymocytes or B220LO splenocytes and liver mononuclear cells. Numbers are percent NKT cells among total leukocytes. Absolute NKT cell number (mean + SEM) was calculated from percent NKT cells and total mononuclear cell number. Data are representative of two independent experiments. Note that because IL-7Rα0 thymic size is small due to low cellularity, percent NKT cells appears artificially high despite extremely low NKT cell numbers. The p value was calculated by one-way ANOVA with Tukey’s posttest. E, Intracellular expression of the hBcl-xL and hBcl-2 transgenes within NKT and NK cells of B6.C-Bcl-xLtg and B6.Bcl-2tg mice. Overlaid histograms represent isotype control (shaded) and specific mAb (open) staining. Data are representative of two independent experiments; n, as in C. F, Splenic and hepatic NK and NKT cells from B6 (n = 6), B6–IL-150 (n = 7), B6–IL-150;Bcl-2tg (n = 4), and B6–IL-150;Bcl-xLtg (n = 10) mice were identified as NK1.1+tetramerNEG and NK1.1+tetramer+ cells, respectively, within electronically gated B220LO splenic or hepatic mononuclear cells. Data are representative of four independent experiments.

FIGURE 3.

Bcl-xL overexpression restores NKT cell development in IL-150 mice. A, Thymic, splenic, and hepatic NKT cells from B6 (n = 8), B6–IL-150 (n = 10), B6–IL-150;Bcl-2tg (n = 6), and B6–IL-150;Bcl-xLtg (n = 16) mice were identified as CD3ε+tetramer+ cells within electronically gated CD8LO thymic, B220LO splenic, or liver mononuclear cells. Numbers are percent NKT cells among total leukocytes. B, Absolute numbers of NKT cells in the thymus and spleen of B6, B6–IL-150, B6–IL-150;Bcl-2tg, and B6–IL-150;Bcl-xLtg mice were calculated from percent NKT cells in A and total cell count. Data are representative of six independent experiments showing mean + SEM; n, as in A. The p value was calculated by one-way ANOVA with Tukey’s posttest. C, Intracellular expression of the hBcl-xL and hBcl-2 transgene within double-negative, DP, CD4+CD8NEG, and CD8+CD4 thymocytes of B6.C-Bcl-xLtg (right panels) and B6.Bcl-2tg (left panels) mice. Overlaid histograms represent isotype control (shaded) and specific mAb (open). Data are representative of three independent experiments. n = 5. D, Thymic and splenic NKT cells from IL-7Rα+/− (n = 3), IL-7Rα0 (n = 3), IL-7Rα0;Bcl-2tg (n = 3), and IL-7Rα0;Bcl-xLtg (n = 2) mice were identified as CD3ε+tetramer+ cells within electronically gated CD8LO thymocytes or B220LO splenocytes and liver mononuclear cells. Numbers are percent NKT cells among total leukocytes. Absolute NKT cell number (mean + SEM) was calculated from percent NKT cells and total mononuclear cell number. Data are representative of two independent experiments. Note that because IL-7Rα0 thymic size is small due to low cellularity, percent NKT cells appears artificially high despite extremely low NKT cell numbers. The p value was calculated by one-way ANOVA with Tukey’s posttest. E, Intracellular expression of the hBcl-xL and hBcl-2 transgenes within NKT and NK cells of B6.C-Bcl-xLtg and B6.Bcl-2tg mice. Overlaid histograms represent isotype control (shaded) and specific mAb (open) staining. Data are representative of two independent experiments; n, as in C. F, Splenic and hepatic NK and NKT cells from B6 (n = 6), B6–IL-150 (n = 7), B6–IL-150;Bcl-2tg (n = 4), and B6–IL-150;Bcl-xLtg (n = 10) mice were identified as NK1.1+tetramerNEG and NK1.1+tetramer+ cells, respectively, within electronically gated B220LO splenic or hepatic mononuclear cells. Data are representative of four independent experiments.

Close modal

Because the Bcl2 (Igμ) and Bclxl (pLck) transgenes are under the control of different promoters (54, 55), we assessed the expression patterns of both transgenes within different thymocyte subsets. The data revealed that both transgenes were expressed as early as the CD4CD8 double-negative stage of development, through the DP stage, and within CD4 and CD8 single-positive thymocytes (Fig. 3C). Consistent with Bcl-2 expression during early thymocyte development, Bcl-2 transgene introgression into IL-7Rα null mice, which poorly if at all develop conventional T and NKT cells (34, 61), rescued the development of both lineages (Fig. 3D). Therefore, the failure of Bcl-2 overexpression to restore NKT cell numbers in IL-150 mice was not due to the lack of proper transgene expression or functionality, but most likely due to the fact that Bcl-2 plays a minor role when compared with Bcl-xL in conveying IL-15 signals for NKT cell survival. Together, the above data indicate that IL-15 relays specific, Bcl-xL–dependent survival signals during NKT cell ontogeny.

A previous study reported that IL-15–induced Mcl-1 specifically supported NK cell survival in vivo and in vitro (50) and that overexpression of Bcl-2 in mice lacking IL-2Rβ, the common β-chain of IL-2R and IL-15R, or in IL-150 mice repopulates NK cells in vivo (62, 63). Because NKT and NK cells are closely related innate-like lymphocyte lineages and because both lineages use IL-15 for survival, we ascertained whether IL-15 induces survival of the two cell types by similar mechanism(s). This question was addressed because Bcl-2 and Bcl-xL transgenes were expressed in NK cells from B6–Bcl-2tg and C-Bcl-xLtg mice (Fig. 3E). We found that, as previously reported (62, 63), Bcl-2 partially rescued hepatic but not splenic NK cells. Nonetheless, Bcl-xL overexpression rescued NKT but not hepatic or splenic NK cells in IL-150 mice (Fig. 3F), which is consistent with the reported model in which NK cells predominantly use Mcl-1 for IL-15–induced survival (50). Thus, IL-15 appears to regulate survival of distinct lineages in a unique, cell type-specific manner, with Bcl-xL serving as an IL-15–induced survival factor for NKT cells.

Because Bcl-xL is required for DP thymocyte survival and Vα14 to Jα18 rearrangement (4), it is possible that IL-150 DP thymocytes poorly rearrange these distal gene segments, thereby resulting in low NKT cell numbers. To test this, Vα14 to Jα18 rearrangement within DP thymocytes flow sorted from wt, IL-150, Jα180, and CD1dTD mice, which rearrange Vα14 to Jα18 gene segments but are deficient in positive selection of NKT cells (52), was quantified. No defect in the canonical Vα14 to Jα18 rearrangement was seen in IL-150 DP thymocytes (Fig. 4). Likewise, the more distal Vα8 to Jα5 rearrangement was also unaffected in IL-150 DP thymocytes as gauged by nonquantitative PCR (Fig. 4B). Interestingly, however, ∼1.5-fold higher Vα14 to Jα18 rearrangement was observed in DP thymocytes of IL-150;Bcl-xLtg mice (Fig. 4A), which was statistically insignificant. This result suggests that any increase in the initial pool of NKT cell precursors in DP thymocytes of IL-150;Bcl-xLtg mice less likely led to increased numbers of NKT cells in these mice.

FIGURE 4.

Normal TCRα rearrangement in IL-150 mice. A, Vα14-to-Jα18 rearrangement within DP thymocytes flow sorted from the indicated strains was assessed by real-time qPCR and represented as mean + SD of two independent experiments. β-actin was used as the internal control for normalization. The p value calculated by two-tailed, unpaired t test indicated that Vα14-to-Jα18 rearrangement in B6, B6–IL-150, and B6–IL-150;Bcl-xLtg DP thymocytes was significant only when compared with that in B6-Jα180 DP cells. The ∼1.5-fold increased Vα14-to-Jα18 rearrangement seen in B6–IL-150;Bcl-xLtg DP thymocytes compared with the others was statistically insignificant (p < 0.0997). B, RT-PCR assessment of Vα8-to-Jα5 and Vα14-to-Jα18 rearrangements. Data are representative of two experiments.

FIGURE 4.

Normal TCRα rearrangement in IL-150 mice. A, Vα14-to-Jα18 rearrangement within DP thymocytes flow sorted from the indicated strains was assessed by real-time qPCR and represented as mean + SD of two independent experiments. β-actin was used as the internal control for normalization. The p value calculated by two-tailed, unpaired t test indicated that Vα14-to-Jα18 rearrangement in B6, B6–IL-150, and B6–IL-150;Bcl-xLtg DP thymocytes was significant only when compared with that in B6-Jα180 DP cells. The ∼1.5-fold increased Vα14-to-Jα18 rearrangement seen in B6–IL-150;Bcl-xLtg DP thymocytes compared with the others was statistically insignificant (p < 0.0997). B, RT-PCR assessment of Vα8-to-Jα5 and Vα14-to-Jα18 rearrangements. Data are representative of two experiments.

Close modal

Because IL-15 deficiency reduced the numbers of thymic CD44+NK1.1st2 and CD44+NK1.1+st3 NKT cells, and because forced Bcl-xL overexpression in IL-150 mice rescued total thymic and peripheral NKT cell numbers (Fig. 3), whether IL-15 has additional roles in NKT cell ontogeny was ascertained. Therefore, we determined whether IL-15 regulates terminal maturation of developing NKT cells as well, particularly at st3, in which NKT cells acquire most of their NK cell-like effector functions. Thus, the phenotype of IL-150;Bcl-xLtg NKT cells was compared with wt and IL-150 NKT cells. Surprisingly, akin to IL-150 NKT cells, phenotypic st2 to st3 transition, as measured by acquisition of the NK1.1 surface marker, was still partially impaired in IL-150;Bcl-xLtg mice (Fig. 5A, 5B), despite near wt numbers of NKT cells in these mice. Additional phenotypic analyses revealed that expression patterns of few other mature NKT cell markers, such as NKG2D, Ly6C, CD69, and Ly49C/I, reflected those of thymic NKT cells that were blocked at st2 (Fig. 5C), as these markers are predominantly expressed by NK1.1+ NKT cells (Fig. 6A). Of those markers, the expression of Ly6C was also altered in hepatic but not splenic NKT cells in IL-150 and IL-150;Bcl-xLtg mice (Fig. 5C). Thus, IL-15 is required not only for lineage survival but also for terminal maturation of NKT cells, especially of the thymic subset. Furthermore, the differential requirement for IL-15 in terminal differentiation (e.g., acquisition of Ly6C, NK1.1) of some (thymic and/or hepatic) but not all (splenic) NKT cells suggests that distinct signals instruct NKT cell maturation in distinct lymphoid organs.

FIGURE 5.

IL-15 regulates terminal maturation of NKT cells. A, NKT cell developmental stages in B6 (n = 9), B6–IL-150 (n = 6), and B6–IL-150;Bcl-xLtg (n = 13) mice were identified as CD44NEGNK1.1NEGst1, CD44+NK1.1NEGst2, or CD44+NK1.1+st3 in the thymus or as NK1.1NEGtetramer+ or NK1.1+tetramer+ within the splenic and hepatic mononuclear cells. Numbers are percentage of cells among total NKT cells. Data are representative of six independent experiments. B, Absolute numbers of thymic NKT cells within st1, st2, and st3 were calculated as in Fig. 3D. Data are representative of six independent experiments showing the mean + SEM; n, same as in A. The p value was calculated by one-way ANOVA with Tukey’s posttest. C, Expression of lineage-specific markers by thymic, splenic, and hepatic NKT cells from B6 (n = 6), B6–IL-150 (n = 6), and B6–IL-150;Bcl-xLtg (n = 6) mice as determined by flow cytometry after staining with specific mAb. Data are representative of three independent experiments.

FIGURE 5.

IL-15 regulates terminal maturation of NKT cells. A, NKT cell developmental stages in B6 (n = 9), B6–IL-150 (n = 6), and B6–IL-150;Bcl-xLtg (n = 13) mice were identified as CD44NEGNK1.1NEGst1, CD44+NK1.1NEGst2, or CD44+NK1.1+st3 in the thymus or as NK1.1NEGtetramer+ or NK1.1+tetramer+ within the splenic and hepatic mononuclear cells. Numbers are percentage of cells among total NKT cells. Data are representative of six independent experiments. B, Absolute numbers of thymic NKT cells within st1, st2, and st3 were calculated as in Fig. 3D. Data are representative of six independent experiments showing the mean + SEM; n, same as in A. The p value was calculated by one-way ANOVA with Tukey’s posttest. C, Expression of lineage-specific markers by thymic, splenic, and hepatic NKT cells from B6 (n = 6), B6–IL-150 (n = 6), and B6–IL-150;Bcl-xLtg (n = 6) mice as determined by flow cytometry after staining with specific mAb. Data are representative of three independent experiments.

Close modal
FIGURE 6.

IL-15 induces Tbx21- and T-bet–regulated genes in developing NKT cells. A, Cluster analysis of all (left panel) or select genes (right panel) differentially expressed in B6 (n = 3) and IL-150 (n = 2) NKT cells from microarray experiments are shown. Differential expression was defined as those genes that showed log2 fold change ≥1.5/≤−1.5 with a nominal p value <0.001. B and C, Select differentially expressed genes identified in A were validated by qPCR using RNA isolated from flow-sorted B6 and B6–IL-150 thymic NKT cells. Genes upregulated in B6-derived thymic NKT cells (B) and genes upregulated in B6–IL-150–derived NKT cells (C) are shown. β-actin was used to normalize expression levels. Data represent mean + SD of two independent experiments; each qPCR was performed in duplicate for an n value of 4 to calculate the p value by unpaired t test.

FIGURE 6.

IL-15 induces Tbx21- and T-bet–regulated genes in developing NKT cells. A, Cluster analysis of all (left panel) or select genes (right panel) differentially expressed in B6 (n = 3) and IL-150 (n = 2) NKT cells from microarray experiments are shown. Differential expression was defined as those genes that showed log2 fold change ≥1.5/≤−1.5 with a nominal p value <0.001. B and C, Select differentially expressed genes identified in A were validated by qPCR using RNA isolated from flow-sorted B6 and B6–IL-150 thymic NKT cells. Genes upregulated in B6-derived thymic NKT cells (B) and genes upregulated in B6–IL-150–derived NKT cells (C) are shown. β-actin was used to normalize expression levels. Data represent mean + SD of two independent experiments; each qPCR was performed in duplicate for an n value of 4 to calculate the p value by unpaired t test.

Close modal

To define the mechanism by which IL-15 regulates terminal maturation of thymic NKT cells, microarray experiments were performed. For this, wt and IL-150 thymic NKT cells were flow sorted, and the derived RNA was subjected to hybridization to a mouse microarray chip containing 60,000 genes; hybridization was performed in duplicate (IL-150) or triplicate (wt). Select genes differentially expressed (minimally log2 fold change ≥+1.5/≥−1.5 and nominal p value < 0.001) in wt and IL-150 NKT cells (Fig. 6A) were then validated by real-time qPCR. We found that Tbx21 (T-bet) was upregulated in wt NKT cells compared with IL-150 NKT cells (Fig. 6A, 6B). This finding is consistent with the previously reported IL-15–induced Tbx21-encoded T-bet expression in NKT cells in vitro (20). Furthermore, multiple Tbx21-regulated genes, such as Ifng, Gzma (granzyme A), Gzmc, Hopx, and several NK cell receptor genes (Klra3 [NKG2E], Klrb1c [NK1.1], Klrc1 [NKG2A/B], and Klrk1 [NKG2D]) (20, 29), were upregulated in wt thymic NKT cells, which predominantly consist of st3 NKT cells (Fig. 6B). Consistent with the lack of intracellular Bcl-2 in IL-150 NKT cells (Fig. 1C), the Bcl2 gene was poorly expressed in these cells (Fig. 6B). Conversely, Zbtb16 (PLZF), Tcf7 (T cell factor 1), and Rorc (retinoic acid-related orphan receptor γt) genes, all signatures of st1 and st2 NKT cells, were upregulated in IL-150 thymic NKT cells (Fig. 6C), consistent with developmental arrest at st2. Because T-bet is a critical regulator of differentiation in other lineages with cytotoxic properties (CTL and NK cells), our results are consistent with a role of Tbx21 in regulating st2 to st3 transition.

Terminally mature NKT cells produce the highest amounts of effector cytokines. Because phenotypically this subset of NKT cells appears impaired in IL-150 and IL-150;Bcl-xLtg mice, we determined whether they were functionally competent. For this purpose, the potent NKT cell ligand αGalCer (64) was administered i.p. into IL-15+/−, IL-150, and IL-150;Bcl-xLtg mice, and 4 h later, intracellular and serum IL-4 and IFN-γ responses were monitored. Compared to IL-15+/− mice, significantly fewer splenic and hepatic NKT cells from IL-150 and IL-150;Bcl-xLtg mice expressed intracellular IL-4 and IFN-γ in response to in vivo stimulation (Fig. 7A, 7B). Furthermore, NKT cells that expressed IL-4 (splenic) and IFN-γ (splenic and hepatic) in IL-150 and IL-150;Bcl-xLtg mice expressed these cytokines at significantly lower levels compared with those in IL-15+/− mice (Fig. 7A, 7B). Consistent with this expression pattern, the serum IFN-γ and IL-4 responses were also poor in IL-150 and IL-150;Bcl-xLtg mice compared with their IL-15+/− counterparts (Fig. 7C, 7D). Thus, IL-15 signaling is required for fully competent NKT cell responses in vivo.

FIGURE 7.

IL-15 regulates functional maturation of NKT cells. A and B, B6 (n = 6), B6–IL-150 (n = 6), and B6–IL-150;Bcl-xLtg (n = 12) mice were injected i.p. with 5 μg αGalCer (open) or vehicle (shaded). After 4 h, splenic (A) and hepatic (B) B220LOCD3ε+tetramer+ cells were electronically gated, and intracellular IL-4 and IFN-γ expression was monitored. Numbers refer to ΔMFI. Data are representative of four independent experiments. C and D, Using ELISA, serum IFN-γ (C) and IL-4 (D) responses to in vivo NKT cell stimulation in the above experiment were measured. The p value was calculated by one-way ANOVA with Tukey’s posttest. E and F, Thymic and splenic mononuclear cells from B6 and B6–IL-150 mice were analyzed for intracellular T-bet (E) and Gata-3 (F) expression within electronically gated CD8LOCD3ε+tetramerNEG thymocytes or B220LOCD3ε+tetramer+ splenocytes. Numbers refer to the ΔMFI between isotype control (shaded) and specific mAb (open). Data are representative of two independent experiments. n = 5. G, The ratio of T-bet and Gata-3 expression by IL-150 and wt NKT cells is shown. A ratio of ∼1 (e.g., Gata-3) indicates no difference in expression; a ratio <1 (e.g., T-bet) indicates lower expression in IL-150 NKT cells.

FIGURE 7.

IL-15 regulates functional maturation of NKT cells. A and B, B6 (n = 6), B6–IL-150 (n = 6), and B6–IL-150;Bcl-xLtg (n = 12) mice were injected i.p. with 5 μg αGalCer (open) or vehicle (shaded). After 4 h, splenic (A) and hepatic (B) B220LOCD3ε+tetramer+ cells were electronically gated, and intracellular IL-4 and IFN-γ expression was monitored. Numbers refer to ΔMFI. Data are representative of four independent experiments. C and D, Using ELISA, serum IFN-γ (C) and IL-4 (D) responses to in vivo NKT cell stimulation in the above experiment were measured. The p value was calculated by one-way ANOVA with Tukey’s posttest. E and F, Thymic and splenic mononuclear cells from B6 and B6–IL-150 mice were analyzed for intracellular T-bet (E) and Gata-3 (F) expression within electronically gated CD8LOCD3ε+tetramerNEG thymocytes or B220LOCD3ε+tetramer+ splenocytes. Numbers refer to the ΔMFI between isotype control (shaded) and specific mAb (open). Data are representative of two independent experiments. n = 5. G, The ratio of T-bet and Gata-3 expression by IL-150 and wt NKT cells is shown. A ratio of ∼1 (e.g., Gata-3) indicates no difference in expression; a ratio <1 (e.g., T-bet) indicates lower expression in IL-150 NKT cells.

Close modal

IFN-γ and IL-4 responses by NKT cells require transcriptional regulation by T-bet and Gata-3 (17, 19). T-bet is also required for the terminal maturation of thymic NKT cells, as T-bet deficiency blocked NKT cell ontogeny at the st2 to st3 transition (20), similar to that seen with IL-15 deficiency (Fig. 5). Because Tbx21 expression is reduced in IL-150 NKT cells, and IL-15 induces Tbx21 expression in NK cells in vitro (20), we investigated whether impaired NKT cell function in IL-150 and IL-150;Bcl-xLtg mice was due to altered levels of T-bet and Gata-3 expression. For this, thymic and splenic NK1.1 and NK1.1+ NKT cells from wt and IL-150 mice were probed for intracellular expression of T-bet and Gata-3. We found that the level of T-bet expression was reduced by 30 (splenic) to 50% (thymic) in IL-150 compared with wt NKT cells (Fig. 7E, 7G). Additionally, a slightly reduced percentage of thymic, but not splenic, NKT cells expressed T-bet in IL-150 mice. Similarly, intracellular Gata-3 expression was ∼15–20% lower in thymic and splenic NK1 NKT cells but about similar in NK1+ NKT cells of IL-150 mice compared with their wt counterparts (Fig. 7F, 7G), consistent with somewhat more preserved IL-4 production in IL-150 NKT cells when compared with IFN-γ response. In wt NKT cells, Gata-3 expression level was constant, whereas T-bet levels increased with st3 differentiation (Fig. 7G). Hence, the reduced T-bet expression (Fig. 7E) in IL-150 mice is consistent with reduced st3 NKT cells in these mice. Thus, IL-15 signals are partially required for T-bet and Gata-3 expression in NKT cells and, possibly, for the robust IFN-γ and IL-4 cytokine responses seen in wt but not in IL-150 and IL-150;Bcl-xLtg mice.

We describe three unique features of IL-15 that are essential for functional NKT cell development: 1) it induces survival of thymic and peripheral NKT cells by regulating Bcl-xL and Bcl-2 expression; 2) IL-15 regulates T-bet expression and signals st3 NKT cell induction and differentiation; and 3) it regulates effector differentiation that is consistent with the role of T-bet in this process. Thus, IL-15 plays a central role in NKT cell development and function.

It appears as though NKT cells narrowly escape death at each stage of their development: immediately after positive selection, they turn on Nur77 expression in st0 NKT cells (65), perhaps owing to high-avidity interactions with their cognate ligand, namely, CD1d–self-lipid complexes. In T cells, the transcription factor Nur77, in addition to inducing an apoptotic program, when targeted to the mitochondrion binds to the otherwise antiapoptotic factor Bcl-2 and, by changing its conformation, converts it into a proapoptotic protein (66). Therefore, onward development of NKT cells perhaps requires survival signal(s) to protect the lineage from death. We have identified one such survival signal to be IL-15. It induces both Bcl-2 and Bcl-xL within thymic NKT cells in vitro. Yet, only genetic overexpression of Bcl-xL but not Bcl-2 confers survival potential to st1 and st2, but not st3, NKT cells. These data are consistent with a previous report showing that enforced Bcl-2 expression does not rescue NKT cell development in IL-15Rβ–null mice (62). A recent report suggested that forced Bcl-2 overexpression or introgression of Bim deficiency modestly rescues NKT cell development in the IL-15–presenting IL-15Rα–null mice (38). The differences in survival mechanisms in these three animal models currently remain unknown.

The global gene expression analysis of wt and IL-15–deficient NKT cells, as well as our published data (14), indicate that st3 but not st0–2 NKT cells express high levels of Bcl-2. Hence, it is possible that Bcl-2 functions as an important survival factor only upon commitment and development into st3 NKT cells. Because the st3 NKT cells are absent in IL-15–, IL-15Rα–, and IL-15Rβ–null mice, another factor must be necessary to support NKT cell survival up to this stage. Our data suggest Bcl-xL provides that function. The failure of transgenic Bcl-2 to do so in this study perhaps reflects the need for an IL-15–induced, st3-specific accessory factor to execute its function.

Previous studies have demonstrated that thymic and peripheral NKT cells represent distinct functional subsets (44, 47, 48). The factor(s) that impacts differentiation of these two subsets remained incompletely defined. We found that most thymic NKT cells in IL-150;Bcl-xLtg mice were blocked at the st2-to-st3 ontogenetic transition and lacked activation/memory marker expression, as were IL-150 NKT cells. In striking contrast, splenic and hepatic NKT cells in IL-150 and IL-150;Bcl-xLtg mice underwent almost complete phenotypic maturation, suggesting that developmental cues in the thymus and periphery are somewhat distinct. In contrast, gene expression analyses indicated reduced T-bet expression in both thymic and splenic NKT cells of IL-150 mice. T-bet was previously shown to regulate the st2-to-st3 NKT ontogenetic transition (20, 29). Thus, reduced T-bet expression is consistent with the st2-to-st3 developmental block in IL-150 NKT cells. Furthermore, low levels of T-bet may in part explain the poor expression of NK cell receptors as well as the deficiency in st3-specific IFN-γ response in IL-150 mice. Taken together, we conclude that IL-15 differentially impacts central and peripheral NKT cell maturation in vivo.

It is noteworthy that T-bet belongs to a group of transcription factors whose functions are regulated not only by their presence or absence but also by absolute levels. An example of such regulation is seen in effector CD8 T cell fate determination via graded expression of T-bet (67). Differentiation into effector and memory CD8 T cell fates was dictated by the levels of T-bet, in which high T-bet expression induced an effector cell fate, whereas lower T-bet level directed cells into the memory pool. Such a model would in principle be consistent with a past report (38) and ours suggesting that partially reduced levels of T-bet expression in IL-150 mice could indeed significantly impair the NKT cell differentiation program. Finally, because T-bet regulates expression of IL-15/IL-2Rβ (CD122) on NKT cells (this report and 29) and therefore IL-15 responsiveness, the two factors (IL-15 and T-bet) generate a positive-feedback regulatory loop. The property of all such loops is that small changes in gene expression are amplified by self-sustained and self-amplified system oscillation. Therefore, we predict that the reduced but not the complete absence of T-bet observed in IL-150 NKT cells may be sufficient to prevent complete effector differentiation.

The importance of NKT cell function(s) is implicated in multiple infectious diseases. Because NKT cells acquire most of their NK cell-like properties at st3 of their differentiation, and because IL-15 seems to regulate acquisition of these properties, we speculate that IL-15–encoded NKT cell functions will be important in controlling in vivo responses to infections. This will be an exciting area for future investigations. In conclusion, IL-15 functions by supporting survival of developing NKT cells and perhaps by inducing Tbx21 expression as an st3-specific differentiation factor in NKT cells. NKT cells share this feature with NK cells and CD8 T cells and are expected to collaborate with these lineages in executing Tbx21-induced functions in vivo.

We thank Drs. J.J. Peschon (Immunex), M. Taniguchi (RIKEN, Yokohama, Japan), and A. Bendelac (University of Chicago) for sharing B6–IL-150, B6.129–IL-7α0, B6-Jα180, and B6.129-CD1dTD mice; the National Institutes of Health Tetramer Facility for unloaded and PBS57-loaded mouse CD1d monomers; the Kirin Brewery Company for KRN7000; and A.J. Joyce for technical assistance.

This work was supported by National Institutes of Health training (HL069765 and AI007611), research (AI042284, AI016721, AI070305, HL089667, AI068149, AI074754, and AI064639), and core (CA068485 and DK058404) grants.

The microarray data presented in this article have been submitted to the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=dtgxpkoosiokmpm&acc=) under accession number GSE32568.

The online version of this article contains supplemental material.

Abbreviations used in this article:

CT

threshold cycle

DP

double-positive

αGalCer

α-galactosylceramide

IL-150

IL-15–deficient

PLZF

promyelocytic leukemia zinc finger

qPCR

quantitative PCR

rh

recombinant human

SLAM

signaling lymphocyte activation molecule

st

stage

wt

wild-type.

1
Bendelac
A.
,
Savage
P. B.
,
Teyton
L.
.
2007
.
The biology of NKT cells.
Annu. Rev. Immunol.
25
:
297
336
.
2
Van Kaer
L.
2005
.
α-Galactosylceramide therapy for autoimmune diseases: prospects and obstacles.
Nat. Rev. Immunol.
5
:
31
42
.
3
Godfrey
D. I.
,
Stankovic
S.
,
Baxter
A. G.
.
2010
.
Raising the NKT cell family.
Nat. Immunol.
11
:
197
206
.
4
Bezbradica
J. S.
,
Hill
T.
,
Stanic
A. K.
,
Van Kaer
L.
,
Joyce
S.
.
2005
.
Commitment toward the natural T (iNKT) cell lineage occurs at the CD4+8+ stage of thymic ontogeny.
Proc. Natl. Acad. Sci. USA
102
:
5114
5119
.
5
D’Cruz
L. M.
,
Knell
J.
,
Fujimoto
J. K.
,
Goldrath
A. W.
.
2010
.
An essential role for the transcription factor HEB in thymocyte survival, Tcra rearrangement and the development of natural killer T cells.
Nat. Immunol.
11
:
240
249
.
6
Egawa
T.
,
Eberl
G.
,
Taniuchi
I.
,
Benlagha
K.
,
Geissmann
F.
,
Hennighausen
L.
,
Bendelac
A.
,
Littman
D. R.
.
2005
.
Genetic evidence supporting selection of the Valpha14i NKT cell lineage from double-positive thymocyte precursors.
Immunity
22
:
705
716
.
7
Hu
T.
,
Simmons
A.
,
Yuan
J.
,
Bender
T. P.
,
Alberola-Ila
J.
.
2010
.
The transcription factor c-Myb primes CD4+CD8+ immature thymocytes for selection into the iNKT lineage.
Nat. Immunol.
11
:
435
441
.
8
Dose
M.
,
Sleckman
B. P.
,
Han
J.
,
Bredemeyer
A. L.
,
Bendelac
A.
,
Gounari
F.
.
2009
.
Intrathymic proliferation wave essential for Valpha14+ natural killer T cell development depends on c-Myc.
Proc. Natl. Acad. Sci. USA
106
:
8641
8646
.
9
Mycko
M. P.
,
Ferrero
I.
,
Wilson
A.
,
Jiang
W.
,
Bianchi
T.
,
Trumpp
A.
,
MacDonald
H. R.
.
2009
.
Selective requirement for c-Myc at an early stage of V(α)14i NKT cell development.
J. Immunol.
182
:
4641
4648
.
10
Kovalovsky
D.
,
Uche
O. U.
,
Eladad
S.
,
Hobbs
R. M.
,
Yi
W.
,
Alonzo
E.
,
Chua
K.
,
Eidson
M.
,
Kim
H. J.
,
Im
J. S.
, et al
.
2008
.
The BTB-zinc finger transcriptional regulator PLZF controls the development of invariant natural killer T cell effector functions.
Nat. Immunol.
9
:
1055
1064
.
11
Savage
A. K.
,
Constantinides
M. G.
,
Han
J.
,
Picard
D.
,
Martin
E.
,
Li
B.
,
Lantz
O.
,
Bendelac
A.
.
2008
.
The transcription factor PLZF directs the effector program of the NKT cell lineage.
Immunity
29
:
391
403
.
12
Schmidt-Supprian
M.
,
Tian
J.
,
Grant
E. P.
,
Pasparakis
M.
,
Maehr
R.
,
Ovaa
H.
,
Ploegh
H. L.
,
Coyle
A. J.
,
Rajewsky
K.
.
2004
.
Differential dependence of CD4+CD25+ regulatory and natural killer-like T cells on signals leading to NF-kappaB activation.
Proc. Natl. Acad. Sci. USA
101
:
4566
4571
.
13
Sivakumar
V.
,
Hammond
K. J.
,
Howells
N.
,
Pfeffer
K.
,
Weih
F.
.
2003
.
Differential requirement for Rel/nuclear factor kappa B family members in natural killer T cell development.
J. Exp. Med.
197
:
1613
1621
.
14
Stanic
A. K.
,
Bezbradica
J. S.
,
Park
J. J.
,
Matsuki
N.
,
Mora
A. L.
,
Van Kaer
L.
,
Boothby
M. R.
,
Joyce
S.
.
2004
.
NF-κ B controls cell fate specification, survival, and molecular differentiation of immunoregulatory natural T lymphocytes.
J. Immunol.
172
:
2265
2273
.
15
Vallabhapurapu
S.
,
Powolny-Budnicka
I.
,
Riemann
M.
,
Schmid
R. M.
,
Paxian
S.
,
Pfeffer
K.
,
Körner
H.
,
Weih
F.
.
2008
.
Rel/NF-kappaB family member RelA regulates NK1.1- to NK1.1+ transition as well as IL-15-induced expansion of NKT cells.
Eur. J. Immunol.
38
:
3508
3519
.
16
Lazarevic
V.
,
Zullo
A. J.
,
Schweitzer
M. N.
,
Staton
T. L.
,
Gallo
E. M.
,
Crabtree
G. R.
,
Glimcher
L. H.
.
2009
.
The gene encoding early growth response 2, a target of the transcription factor NFAT, is required for the development and maturation of natural killer T cells.
Nat. Immunol.
10
:
306
313
.
17
Cen
O.
,
Ueda
A.
,
Guzman
L.
,
Jain
J.
,
Bassiri
H.
,
Nichols
K. E.
,
Stein
P. L.
.
2009
.
The adaptor molecule signaling lymphocytic activation molecule-associated protein (SAP) regulates IFN-gamma and IL-4 production in V alpha 14 transgenic NKT cells via effects on GATA-3 and T-bet expression.
J. Immunol.
182
:
1370
1378
.
18
Kim
P. J.
,
Pai
S. Y.
,
Brigl
M.
,
Besra
G. S.
,
Gumperz
J.
,
Ho
I. C.
.
2006
.
GATA-3 regulates the development and function of invariant NKT cells.
J. Immunol.
177
:
6650
6659
.
19
Wang
Z. Y.
,
Kusam
S.
,
Munugalavadla
V.
,
Kapur
R.
,
Brutkiewicz
R. R.
,
Dent
A. L.
.
2006
.
Regulation of Th2 cytokine expression in NKT cells: unconventional use of Stat6, GATA-3, and NFAT2.
J. Immunol.
176
:
880
888
.
20
Townsend
M. J.
,
Weinmann
A. S.
,
Matsuda
J. L.
,
Salomon
R.
,
Farnham
P. J.
,
Biron
C. A.
,
Gapin
L.
,
Glimcher
L. H.
.
2004
.
T-bet regulates the terminal maturation and homeostasis of NK and Valpha14i NKT cells.
Immunity
20
:
477
494
.
21
Griewank
K.
,
Borowski
C.
,
Rietdijk
S.
,
Wang
N.
,
Julien
A.
,
Wei
D. G.
,
Mamchak
A. A.
,
Terhorst
C.
,
Bendelac
A.
.
2007
.
Homotypic interactions mediated by Slamf1 and Slamf6 receptors control NKT cell lineage development.
Immunity
27
:
751
762
.
22
Eberl
G.
,
Lowin-Kropf
B.
,
MacDonald
H. R.
.
1999
.
Cutting edge: NKT cell development is selectively impaired in Fyn- deficient mice.
J. Immunol.
163
:
4091
4094
.
23
Gadue
P.
,
Morton
N.
,
Stein
P. L.
.
1999
.
The Src family tyrosine kinase Fyn regulates natural killer T cell development.
J. Exp. Med.
190
:
1189
1196
.
24
Nichols
K. E.
,
Hom
J.
,
Gong
S. Y.
,
Ganguly
A.
,
Ma
C. S.
,
Cannons
J. L.
,
Tangye
S. G.
,
Schwartzberg
P. L.
,
Koretzky
G. A.
,
Stein
P. L.
.
2005
.
Regulation of NKT cell development by SAP, the protein defective in XLP.
Nat. Med.
11
:
340
345
.
25
Chung
B.
,
Aoukaty
A.
,
Dutz
J.
,
Terhorst
C.
,
Tan
R.
.
2005
.
Signaling lymphocytic activation molecule-associated protein controls NKT cell functions.
J. Immunol.
174
:
3153
3157
.
26
Pasquier
B.
,
Yin
L.
,
Fondanèche
M. C.
,
Relouzat
F.
,
Bloch-Queyrat
C.
,
Lambert
N.
,
Fischer
A.
,
de Saint-Basile
G.
,
Latour
S.
.
2005
.
Defective NKT cell development in mice and humans lacking the adapter SAP, the X-linked lymphoproliferative syndrome gene product.
J. Exp. Med.
201
:
695
701
.
27
Gadue
P.
,
Stein
P. L.
.
2002
.
NK T cell precursors exhibit differential cytokine regulation and require Itk for efficient maturation.
J. Immunol.
169
:
2397
2406
.
28
Stanic
A. K.
,
Bezbradica
J. S.
,
Park
J. J.
,
Van Kaer
L.
,
Boothby
M. R.
,
Joyce
S.
.
2004
.
Cutting edge: the ontogeny and function of Va14Ja18 natural T lymphocytes require signal processing by protein kinase C θ and NF-κ B.
J. Immunol.
172
:
4667
4671
.
29
Matsuda
J. L.
,
Zhang
Q.
,
Ndonye
R.
,
Richardson
S. K.
,
Howell
A. R.
,
Gapin
L.
.
2006
.
T-bet concomitantly controls migration, survival, and effector functions during the development of Valpha14i NKT cells.
Blood
107
:
2797
2805
.
30
Engel
I.
,
Hammond
K.
,
Sullivan
B. A.
,
He
X.
,
Taniuchi
I.
,
Kappes
D.
,
Kronenberg
M.
.
2010
.
Co-receptor choice by V alpha14i NKT cells is driven by Th-POK expression rather than avoidance of CD8-mediated negative selection.
J. Exp. Med.
207
:
1015
1029
.
31
Bezbradica
J. S.
,
Gordy
L. E.
,
Stanic
A. K.
,
Dragovic
S.
,
Hill
T.
,
Hawiger
J.
,
Unutmaz
D.
,
Van Kaer
L.
,
Joyce
S.
.
2006
.
Granulocyte-macrophage colony-stimulating factor regulates effector differentiation of invariant natural killer T cells during thymic ontogeny.
Immunity
25
:
487
497
.
32
Lee
A. J.
,
Zhou
X.
,
Chang
M.
,
Hunzeker
J.
,
Bonneau
R. H.
,
Zhou
D.
,
Sun
S. C.
.
2010
.
Regulation of natural killer T-cell development by deubiquitinase CYLD.
EMBO J.
29
:
1600
1612
.
33
Ohteki
T.
,
Ho
S.
,
Suzuki
H.
,
Mak
T. W.
,
Ohashi
P. S.
.
1997
.
Role for IL-15/IL-15 receptor β-chain in natural killer 1.1+ T cell receptor-α β+ cell development.
J. Immunol.
159
:
5931
5935
.
34
Boesteanu
A.
,
Silva
A. D.
,
Nakajima
H.
,
Leonard
W. J.
,
Peschon
J. J.
,
Joyce
S.
.
1997
.
Distinct roles for signals relayed through the common cytokine receptor γ chain and interleukin 7 receptor α chain in natural T cell development.
J. Exp. Med.
186
:
331
336
.
35
Kennedy
M. K.
,
Glaccum
M.
,
Brown
S. N.
,
Butz
E. A.
,
Viney
J. L.
,
Embers
M.
,
Matsuki
N.
,
Charrier
K.
,
Sedger
L.
,
Willis
C. R.
, et al
.
2000
.
Reversible defects in natural killer and memory CD8 T cell lineages in interleukin 15-deficient mice.
J. Exp. Med.
191
:
771
780
.
36
Matsuda
J. L.
,
Gapin
L.
,
Sidobre
S.
,
Kieper
W. C.
,
Tan
J. T.
,
Ceredig
R.
,
Surh
C. D.
,
Kronenberg
M.
.
2002
.
Homeostasis of V α 14i NKT cells.
Nat. Immunol.
3
:
966
974
.
37
Lodolce
J. P.
,
Boone
D. L.
,
Chai
S.
,
Swain
R. E.
,
Dassopoulos
T.
,
Trettin
S.
,
Ma
A.
.
1998
.
IL-15 receptor maintains lymphoid homeostasis by supporting lymphocyte homing and proliferation.
Immunity
9
:
669
676
.
38
Chang
C. L.
,
Lai
Y. G.
,
Hou
M. S.
,
Huang
P. L.
,
Liao
N. S.
.
2011
.
IL-15Rα of radiation-resistant cells is necessary and sufficient for thymic invariant NKT cell survival and functional maturation.
J. Immunol.
187
:
1235
1242
.
39
Ohteki
T.
,
Yoshida
H.
,
Matsuyama
T.
,
Duncan
G. S.
,
Mak
T. W.
,
Ohashi
P. S.
.
1998
.
The transcription factor interferon regulatory factor 1 (IRF-1) is important during the maturation of natural killer 1.1+ T cell receptor-α/β+ (NK1+ T) cells, natural killer cells, and intestinal intraepithelial T cells.
J. Exp. Med.
187
:
967
972
.
40
Monticelli
L. A.
,
Yang
Y.
,
Knell
J.
,
D’Cruz
L. M.
,
Cannarile
M. A.
,
Engel
I.
,
Kronenberg
M.
,
Goldrath
A. W.
.
2009
.
Transcriptional regulator Id2 controls survival of hepatic NKT cells.
Proc. Natl. Acad. Sci. USA
106
:
19461
19466
.
41
Benlagha
K.
,
Kyin
T.
,
Beavis
A.
,
Teyton
L.
,
Bendelac
A.
.
2002
.
A thymic precursor to the NK T cell lineage.
Science
296
:
553
555
.
42
Godfrey
D. I.
,
Berzins
S. P.
.
2007
.
Control points in NKT-cell development.
Nat. Rev. Immunol.
7
:
505
518
.
43
Pellicci
D. G.
,
Hammond
K. J.
,
Uldrich
A. P.
,
Baxter
A. G.
,
Smyth
M. J.
,
Godfrey
D. I.
.
2002
.
A natural killer T (NKT) cell developmental pathway iInvolving a thymus-dependent NK1.1(-)CD4(+) CD1d-dependent precursor stage.
J. Exp. Med.
195
:
835
844
.
44
Berzins
S. P.
,
McNab
F. W.
,
Jones
C. M.
,
Smyth
M. J.
,
Godfrey
D. I.
.
2006
.
Long-term retention of mature NK1.1+ NKT cells in the thymus.
J. Immunol.
176
:
4059
4065
.
45
Geissmann
F.
,
Cameron
T. O.
,
Sidobre
S.
,
Manlongat
N.
,
Kronenberg
M.
,
Briskin
M. J.
,
Dustin
M. L.
,
Littman
D. R.
.
2005
.
Intravascular immune surveillance by CXCR6+ NKT cells patrolling liver sinusoids.
PLoS Biol.
3
:
e113
.
46
Ohteki
T.
,
Maki
C.
,
Koyasu
S.
,
Mak
T. W.
,
Ohashi
P. S.
.
1999
.
Cutting edge: LFA-1 is required for liver NK1.1+TCR alpha beta+ cell development: evidence that liver NK1.1+TCR alpha beta+ cells originate from multiple pathways.
J. Immunol.
162
:
3753
3756
.
47
McNab
F. W.
,
Berzins
S. P.
,
Pellicci
D. G.
,
Kyparissoudis
K.
,
Field
K.
,
Smyth
M. J.
,
Godfrey
D. I.
.
2005
.
The influence of CD1d in postselection NKT cell maturation and homeostasis.
J. Immunol.
175
:
3762
3768
.
48
McNab
F. W.
,
Pellicci
D. G.
,
Field
K.
,
Besra
G.
,
Smyth
M. J.
,
Godfrey
D. I.
,
Berzins
S. P.
.
2007
.
Peripheral NK1.1 NKT cells are mature and functionally distinct from their thymic counterparts.
J. Immunol.
179
:
6630
6637
.
49
Ma
A.
,
Koka
R.
,
Burkett
P.
.
2006
.
Diverse functions of IL-2, IL-15, and IL-7 in lymphoid homeostasis.
Annu. Rev. Immunol.
24
:
657
679
.
50
Huntington
N. D.
,
Puthalakath
H.
,
Gunn
P.
,
Naik
E.
,
Michalak
E. M.
,
Smyth
M. J.
,
Tabarias
H.
,
Degli-Esposti
M. A.
,
Dewson
G.
,
Willis
S. N.
, et al
.
2007
.
Interleukin 15-mediated survival of natural killer cells is determined by interactions among Bim, Noxa and Mcl-1.
Nat. Immunol.
8
:
856
863
.
51
Calnan
B. J.
,
Szychowski
S.
,
Chan
F. K.
,
Cado
D.
,
Winoto
A.
.
1995
.
A role for the orphan steroid receptor Nur77 in apoptosis accompanying antigen-induced negative selection.
Immunity
3
:
273
282
.
52
Chiu
Y. H.
,
Park
S. H.
,
Benlagha
K.
,
Forestier
C.
,
Jayawardena-Wolf
J.
,
Savage
P. B.
,
Teyton
L.
,
Bendelac
A.
.
2002
.
Multiple defects in antigen presentation and T cell development by mice expressing cytoplasmic tail-truncated CD1d.
Nat. Immunol.
3
:
55
60
.
53
Cui
J.
,
Shin
T.
,
Kawano
T.
,
Sato
H.
,
Kondo
E.
,
Toura
I.
,
Kaneko
Y.
,
Koseki
H.
,
Kanno
M.
,
Taniguchi
M.
.
1997
.
Requirement for Valpha14 NKT cells in IL-12-mediated rejection of tumors.
Science
278
:
1623
1626
.
54
Strasser
A.
,
Harris
A. W.
,
Cory
S.
.
1991
.
bcl-2 transgene inhibits T cell death and perturbs thymic self-censorship.
Cell
67
:
889
899
.
55
Chao
D. T.
,
Linette
G. P.
,
Boise
L. H.
,
White
L. S.
,
Thompson
C. B.
,
Korsmeyer
S. J.
.
1995
.
Bcl-XL and Bcl-2 repress a common pathway of cell death.
J. Exp. Med.
182
:
821
828
.
56
Bezbradica
J. S.
,
Stanic
A. K.
,
Joyce
S.
.
2006
.
Characterization and functional analysis of mouse invariant natural T (iNKT) cells.
Curr. Protoc. Immunol.
Chapter 14
:
Unit 14.13
.
57
Bolstad
B. M.
,
Irizarry
R. A.
,
Astrand
M.
,
Speed
T. P.
.
2003
.
A comparison of normalization methods for high density oligonucleotide array data based on variance and bias.
Bioinformatics
19
:
185
193
.
58
Smyth
G. K.
2004
.
Linear models and empirical bayes methods for assessing differential expression in microarray experiments.
Stat. Appl. Genet. Mol. Biol.
3
:
Article3
.
59
Smyth
G. K.
2005
.
Limma: Linear models for Microarray Data.
Springer
,
New York
.
60
Huang
H. C.
,
Zheng
S.
,
VanBuren
V.
,
Zhao
Z.
.
2010
.
Discovering disease-specific biomarker genes for cancer diagnosis and prognosis.
Technol. Cancer Res. Treat.
9
:
219
230
.
61
Peschon
J. J.
,
Morrissey
P. J.
,
Grabstein
K. H.
,
Ramsdell
F. J.
,
Maraskovsky
E.
,
Gliniak
B. C.
,
Park
L. S.
,
Ziegler
S. F.
,
Williams
D. E.
,
Ware
C. B.
, et al
.
1994
.
Early lymphocyte expansion is severely impaired in interleukin 7 receptor-deficient mice.
J. Exp. Med.
180
:
1955
1960
.
62
Minagawa
M.
,
Watanabe
H.
,
Miyaji
C.
,
Tomiyama
K.
,
Shimura
H.
,
Ito
A.
,
Ito
M.
,
Domen
J.
,
Weissman
I. L.
,
Kawai
K.
.
2002
.
Enforced expression of Bcl-2 restores the number of NK cells, but does not rescue the impaired development of NKT cells or intraepithelial lymphocytes, in IL-2/IL-15 receptor beta-chain-deficient mice.
J. Immunol.
169
:
4153
4160
.
63
Nakazato
K.
,
Yamada
H.
,
Yajima
T.
,
Kagimoto
Y.
,
Kuwano
H.
,
Yoshikai
Y.
.
2007
.
Enforced expression of Bcl-2 partially restores cell numbers but not functions of TCRgammadelta intestinal intraepithelial T lymphocytes in IL-15-deficient mice.
J. Immunol.
178
:
757
764
.
64
Kawano
T.
,
Cui
J.
,
Koezuka
Y.
,
Toura
I.
,
Kaneko
Y.
,
Motoki
K.
,
Ueno
H.
,
Nakagawa
R.
,
Sato
H.
,
Kondo
E.
, et al
.
1997
.
CD1d-restricted and TCR-mediated activation of valpha14 NKT cells by glycosylceramides.
Science
278
:
1626
1629
.
65
Moran
A. E.
,
Holzapfel
K. L.
,
Xing
Y.
,
Cunningham
N. R.
,
Maltzman
J. S.
,
Punt
J.
,
Hogquist
K. A.
.
2011
.
T cell receptor signal strength in Treg and iNKT cell development demonstrated by a novel fluorescent reporter mouse.
J. Exp. Med.
208
:
1279
1289
.
66
Thompson
J.
,
Burger
M. L.
,
Whang
H.
,
Winoto
A.
.
2010
.
Protein kinase C regulates mitochondrial targeting of Nur77 and its family member Nor-1 in thymocytes undergoing apoptosis.
Eur. J. Immunol.
40
:
2041
2049
.
67
Joshi
N. S.
,
Cui
W.
,
Chandele
A.
,
Lee
H. K.
,
Urso
D. R.
,
Hagman
J.
,
Gapin
L.
,
Kaech
S. M.
.
2007
.
Inflammation directs memory precursor and short-lived effector CD8(+) T cell fates via the graded expression of T-bet transcription factor.
Immunity
27
:
281
295
.

The authors have no financial conflicts of interest.