We have generated a novel mouse model harboring the in-frame rearranged TCRVα specific for invariant NKT (iNKT) cells (Vα14-Jα18) on one allele by crossing the mouse cloned from NKT cells with wild-type mice. This genomic configuration would ensure further rearrangement and expression of TCRVα14-Jα18 under the endogenous promoters and enhancers. Mice harboring such an in-frame rearranged TCRVα (Vα14-Jα18 mouse) possessed an increase in iNKT cells in the thymus, liver, spleen, and bone marrow. Intriguingly, both Th1- and Th2-type cytokines were produced upon stimulation with αGalactosylceramide, an agonist of iNKT cells, and the IgE level in the serum remained unaffected in the Vα14-Jα18 mouse. These features markedly distinguish the nature of iNKT cells present in the Vα14-Jα18 mouse from that of iNKT cells found in the Vα14-Jα18 transgenic mouse. Besides these, the expression of TCRVγδ cells remained intact, and the use of the TCRVβ repertoire in iNKT cells was highly biased to TCRVβ8 in the Vα14-Jα18 mouse. Furthermore, αGalactosylceramide-CD1d dimer-reactive immature iNKT cells expressed less Rag2 as compared with the conventional immature T cells at the positive selection stage. Cell cycle analysis on the thymocytes revealed that no particular subset proliferated more vigorously than the others. Crossing the Vα14-Jα18 mouse with the CD1d knockout mouse revealed a novel population of iNKT cells whose coreceptor expression profile was similar to that assigned to iNKT precursor cells. These mice will be useful for the study on the development of iNKT cells as well as on their functions in the immune system.

Invariant NKT (iNKT)3 cells are a subset of TCRαβ T cells possessing a set of markers for the NK cell lineage, such as NK1.1 and a member of Ly-49 family, as well as the invariant Vα14-Jα18 TCR (1, 2, 3). They may function as regulatory cells because they produce a copious amount of cytokines. Defective development of iNKT cells has been linked to autoimmunity (4, 5). Intriguingly, the development of iNKT cells is not dependent on the classical MHCs, but is conditional upon CD1d, a MHC-like molecule, mainly expressed on double-positive (DP) thymocytes (6, 7). Furthermore, they occupy up to 0.7% of the thymocytes as revealed by CD1d tetramer or dimer loaded with a synthetic ligand, α-galactosylceramide (αGalCer) (8, 9). Such an unusual abundance in the invariant TCRVα is intriguing in that their development may be somehow unique as compared with that of the mainstream αβT cells. Many efforts have been devoted to elucidate the development of such cells and suggested that their commitment is “selected” during thymocyte development (10). According to this, iNKT cells diverge from CD4+CD8+ DP in which cells undertake random TCRVα rearrangement, and those succeeding in rearranging Vα14 with Jα18 are “selected” to follow the fate to become iNKT cells. The fact that there are no iNKT cells present in CD1d knockout mice argues that CD1d is indispensable for the development of iNKT cells as well as of non-iNKT cells (8, 11, 12, 13).

The study on the Vα14-Jα18 transgenic (Tg) mice demonstrates 1) a preponderance of iNKT cells in the thymocytes and 2) a biased decrease of TCRVβ7 and Vβ8 in CD8 single-positive (SP) cells that gives ground for a favorable selection of Vα14-Jα18 (8, 14). Such a bias in TCRVβ repertoire, however, does not necessarily insure the lineage commitment of iNKT cells upon assemblage of Vα14-Jα18 and TCRVβ7 or TCRVβ8, as it is conceivable that such a combination may result in an inhibition of the positive selection imposed by TCRVβ8+/CD8 SP cells. Similarly, Vα24-JαQ Tg/Cα−/− mice show a slight increase in iNKT cell number and an enhanced Vβ7 usage concomitant with a decrease in Vβ8.2 usage (15). Furthermore, the Vα14-Jα18 Tg mouse produced only IL-4, but not IFN-γ, upon the Ag stimulation, and the mice possessed Th-2 biased Ig production in the serum (14). Although Vα24-JαQ Tg/Cα−/− mice produced both cytokines, the amount of the cytokines did not reflect the increase of iNKT cells (15). Because the production of both Th1- and Th2-type cytokines is a hallmark of iNKT cells, the Vα14-Jα18 or Vα24-JαQ Tg mouse may not be suitable for certain experiments. These Tg mice have other disadvantages in that the spatiotemporal expression of the Vα14-Jα18 or Vα24-JαQ is not ensured, and that cells are not allowed to delete such a transgene by further rearrangement. It is, therefore, imperative to develop a novel mouse model in which the development and the functions of iNKT cells mimic those found in the control as faithfully as possible.

Our recent work has provided evidence that nuclei from iNKT cells are competent for reprogramming a genome in such a way that they exert the totipotency when transferred into enucleated oocytes (16). Because such cloned mice possess a rearranged set of the in-frame rearranged Vα14-Jα18 and Vβ8-D-J, some progenies from the clone will inherit the in-frame rearranged Vα14-Jα18 in one allele (hereafter referred as the Vα14-Jα18 mouse). In such circumstances, expression of the Vα14-Jα18 TCR would be under the control of the endogenous promoter(s) and enhancers that ensure their proper expression, and the Vα14-Jα18 TCR is readily subject to further rearrangement.

We found that Vα14-Jα18 mice harbored an increased population of iNKT cells in the thymus as well as liver, spleen, and bone marrow. Concomitantly, these mice produced a copious amount of Th1- and Th2-type cytokines upon an appropriate stimulus both in vitro and in vivo relative to the control. These data indicate that the Vα14-Jα18 mouse would be a novel model suitable for studying the development and the functions of iNKT cells.

The cloned mouse harboring the in-frame rearranged Vα14-Jα18 locus derived from C57BL/6 was mated with female C57BL/6 (Charles River Laboratories) and resulting progenies were analyzed by Southern blot for the inheritance of the Vα14-Jα18 locus (16). Mice harboring the in-frame rearranged Vα14-Jα18 on one allele (Vα14-Jα18 mice) were used throughout the study unless specified. Vα14-Jα18 mice were mated with CD1d knockout mice to generate mice possessing the in-frame rearranged Vα14-Jα18 on the CD1d−/− background in F2. Mouse colonies were maintained in a specific pathogen-free facility at the Research Center for Allergy and Immunology and all experiments were conducted in compliance with the protocol approved by the RIKEN Animal Care and Use committee.

Liver lymphocytes were prepared with the Percoll gradient method by centrifugation. Cells from 8- to 10-wk-old mice were stained with mAbs, anti-CD4 (L3T4 RM4-5), anti-CD8 (LY-2 53-6.7), anti-TCRVβ (H57-597), anti-TCRVβ8 (F23.1), anti-TCRγδ (GL3), anti-NK1.1 (PK136), anti-CD24 (M1/69), anti-CD44 (Im7), anti-CD69 (H1.2F3), anti-CD1d (1B1) (BD Pharmingen). αGalCer was obtained from Kirin Brewery and CD1d:Ig dimers were obtained from BD Biosciences. αGalCer-CD1d dimers were prepared according to the manufacturer’s protocol (BD Biosciences). For αGalCer-CD1d dimer staining, cells were blocked with anti-CD16/32 (2.4G2; BD Biosciences) for 5 min on ice. After brief centrifugation, the supernatant was removed and mixed with fluorescently labeled αGalCer-CD1d dimers followed with the appropriate mAbs for 30 min on ice. Stained cells were analyzed with a FACSCalibur or FACSVantage Diva (BD Immunocytometry Systems). Data were analyzed with FlowJo software (Tree Star).

Dendritic cells (DCs) (5 × 104 cells) were enriched with CD11c Microbeads (Miltenyi Biotec) from the control and Vα14-Jα18 mice spleens and pulsed with αGalCer (200 ng/ml) for 6 h before mixture with the whole spleen cells derived from both types of mice (1 × 106 cells). After 72-h coculture, the concentration of IFN-γ and IL-4 was measured with an ELISA kit (Genzyme Techne). For in vivo stimulation, mice received i.p. αGalCer (2 μg/mouse) and the serum was sampled at the indicated time. Concentration of the resultant cytokines was measured with the Bio-Plex Suspension Array system (Bio-Rad). The relative amounts of IgG1, IgE, IgG2a, IgG2b, IgG3, IgM, and IgA were measured with the Mouse Ig ELISA-Quantification kit (Bethyl Laboratories).

Three hours after i.p. αGalCer injection (2 μg/mouse), TCRVβ8+ cells from the Vα14-Jα18 mouse spleen were subject to intracellular staining with anti-IL-4 or anti-IFN-γ Abs in parallel with the isotype control (Cytofix/Cytoperm kit; BD Biosciences).

Thymocytes from Vα14-Jα18 mice were stained with anti-TCRVβ, anti-TCRVβ8 mAbs, and αGalCer-CD1d dimers, and sorted into four populations as described in the legend for Fig. 5 a. Typically, starting from 3.0 × 108 Vα14-Jα18 mouse thymocytes, 2.5 × 105 αGalCer-reactive TCRVβmed/TCRVβ8 cells and 1.0 × 106 αGalCer-nonreactive TCRVβmed/TCRVβ8 cells were obtained after cell sorting. Similarly, the number of αGalCer-reactive TCRVβmed/TCRVβ8med cells and that of αGalCer-nonreactive TCRVβmed/TCRVβ8med cells were 2.1 × 105 and 2.3 × 105, respectively. RNA was prepared with the RNeasy Micro kit (Qiagen) and the first strand cDNA was synthesized with SuperScript II reverse transcriptase (Invitrogen Life Technologies) according to the manufacturer’s instruction. The sequences of primers were: 5′-CACAggACTAgAACACCTgC-3′ and 5′-gCTggTgAAAAggACCTCT-3′ for hypoxanthine phosphoribosyltransferase (HPRT); 5′-CTCCTggTCACgAACTgTCA-3′ and 5′-gggTAgggTAgCCTTTgAgg-3′ for MyD118, and 5′-CCCAgAgAACCACAgAAAAAT-3′ and 5′-TAACCACCCACAATAACAAAT-3′ for RAG2.

FIGURE 5.

a, Less frequent further rearrangement in the immature iNKT cells. TCRVβmed/TCRVβ8 (1) and TCRVβmed/TCRVβ8med (2) thymocytes derived from Vα14-Jα18 mice were subdivided into αGalCer-CD1d dimer-reactive (dim+) and nonreactive (dim) cells, and subjected to semiquantitative RT-PCR for the expression of Rag2, Myd118, and Hprt. Each sample was analyzed with a 3-fold serial dilution. b, No massive apoptosis and vigorous expansion in the Vα14-Jα18 mice thymocytes. Thymocytes from C57BL/6 (control) and Vα14-Jα18 mice were stained with anti-TCRVβ and -TCRVβ8 Abs. Four subsets representing TCRVβhigh/TCRVβ8 (1), TCRVβmed/TCRVβ8 (2), TCRVβ/TCRVβ8 (3), and TCRVβ+/TCRVβ8+ (4) were analyzed for the cell cycle status and apoptosis with propidium iodide staining. The number indicates the percentage of cells in G2-M phase. One representative data from three experiments is shown.

FIGURE 5.

a, Less frequent further rearrangement in the immature iNKT cells. TCRVβmed/TCRVβ8 (1) and TCRVβmed/TCRVβ8med (2) thymocytes derived from Vα14-Jα18 mice were subdivided into αGalCer-CD1d dimer-reactive (dim+) and nonreactive (dim) cells, and subjected to semiquantitative RT-PCR for the expression of Rag2, Myd118, and Hprt. Each sample was analyzed with a 3-fold serial dilution. b, No massive apoptosis and vigorous expansion in the Vα14-Jα18 mice thymocytes. Thymocytes from C57BL/6 (control) and Vα14-Jα18 mice were stained with anti-TCRVβ and -TCRVβ8 Abs. Four subsets representing TCRVβhigh/TCRVβ8 (1), TCRVβmed/TCRVβ8 (2), TCRVβ/TCRVβ8 (3), and TCRVβ+/TCRVβ8+ (4) were analyzed for the cell cycle status and apoptosis with propidium iodide staining. The number indicates the percentage of cells in G2-M phase. One representative data from three experiments is shown.

Close modal

Thymocytes from C57BL/6 and Vα14-Jα18 mice were stained with anti-TCRVβ and -TCRVβ8 mAbs, and sorted into four populations as described in the legend for Fig. 5 b. Cell cycle analysis on these thymocytes was performed with propidium iodide according to the protocol provided by BD Biosciences after RNase treatment.

We have established the cloned mice from the peripheral NKT cells by direct nuclear transfer (16). In such mice, both TCRVα and TCRVβ loci are rearranged in-frame in the germline. As for the TCRVαβ combination, clone 1 had Vα14-Jα18 and Vβ8S2-Dβ1-Jβ2S5, while clone 2 possessed Vα14-Jα18 and Vβ8S3-Dβ1-Jβ1S4. Because the genomic sequence for Vα14-Jα18 in clone 1 is from C57BL/6 (16), we have focused our study on the progenies using clone 1 as a founder. We first determined how these loci are segregated in F1 progeny. Mating with C57BL/6 resulted in the pups possessing the rearranged Vα14-Jα18, Vβ8S2-Dβ1-Vβ2S5, or both (Fig. 1,a). Mating these F1 with C57BL/6 gave rise to mice harboring the in-frame rearranged Vα14-Jα18 in one allele (Fig. 1 b, lower panel). The data also demonstrate that the rearranged Vα14-Jα18 locus is inherited by the progenies without affecting the fertility.

FIGURE 1.

a, Inheritance of the in-frame Vα14-Jα18 and Vβ8-Dβ1-Jβ2S5 locus to the progeny. Genomic DNA (5 μg/lane) from the F1 progeny [((C57BL/6 × 129/Sv) × C57BL/6)] (lanes 1–9) was digested with EcoRI (for TCRVα14) or BamHI (for TCRVβ) and subjected to Southern blot with the specific probes (16 ). Arrowheads show the bands indicating the TCR rearrangement. C: control DNA from C57BL/6. b, Schematic representation of the TCRVα14 locus in the C57BL/6 and Vα14-Jα18 mouse. The TCRVα14 locus in nonrearranged and rearranged germline configuration is shown with EcoRI sites for C57BL/6 (upper panel) and Vα14-Jα18 mice (lower panel), respectively. □, Vα14 exon; ▦, Jα18 exon; ▪, probe used for Southern blot.

FIGURE 1.

a, Inheritance of the in-frame Vα14-Jα18 and Vβ8-Dβ1-Jβ2S5 locus to the progeny. Genomic DNA (5 μg/lane) from the F1 progeny [((C57BL/6 × 129/Sv) × C57BL/6)] (lanes 1–9) was digested with EcoRI (for TCRVα14) or BamHI (for TCRVβ) and subjected to Southern blot with the specific probes (16 ). Arrowheads show the bands indicating the TCR rearrangement. C: control DNA from C57BL/6. b, Schematic representation of the TCRVα14 locus in the C57BL/6 and Vα14-Jα18 mouse. The TCRVα14 locus in nonrearranged and rearranged germline configuration is shown with EcoRI sites for C57BL/6 (upper panel) and Vα14-Jα18 mice (lower panel), respectively. □, Vα14 exon; ▦, Jα18 exon; ▪, probe used for Southern blot.

Close modal

We examined the presence of iNKT cells in Vα14-Jα18 mice and found that 16% of the total thymocytes were iNKT cells (Fig. 2,a). The size of thymi from Vα14-Jα18 mice was, however, reduced compared with the control mice. Indeed, the total cell number of the thymocytes from such mice represented ∼15% of that from the control mice (Table I). Although the absolute number of the thymocytes decreased, the net iNKT cell number was augmented six times over the control (Table I and Fig. 2,a). In the spleen, the number of iNKT cells increased >20 times as compared with the control, and iNKT cells occupied >50% of the total T lymphocytes (Fig. 2,a). In the liver mononuclear cells, more than half were iNKT cells (Fig. 2,a). A similar increase of iNKT cells was also observed in the bone marrow where the number of such cells exhibited almost 20 times that of the control (Fig. 2,a). We then examined the CD4/CD8 expression profile in the thymocytes. CD4+/CD8+ DP cells showed a decrease (47%) accompanying an increase in CD4+ (20%) or CD8+ (13%) SP cells (Fig. 2,b). Concomitantly, double-negative (DN) cells represented ∼20% of the total thymocytes (Fig. 2,b). Because most of the iNKT cells reported to date are either CD4+/CD8 or CD4/CD8, we evaluated the percentage of iNKT cells within the Vα14-Jα18 mice thymocytes. More than 40% of CD4+/CD8 cells were iNKT cells, while almost half of CD4/CD8 cells represented iNKT cells. Among CD4+/CD8+ and CD4/CD8+ cells, few iNKT cells were present (Fig. 2 b). In the periphery, a similar tendency was observed except that some iNKT cells were CD4/CD8+ in the liver (data not shown). The results demonstrated that iNKT cells in Vα14-Jα18 mice are essentially either CD4+/CD8 or CD4/CD8.

FIGURE 2.

a, Predominance of αGalCer-CD1d dimer-reactive NKT (iNKT) cells in Vα14-Jα18 mice. Thymocytes, spleen cells, liver mononuclear cells, and bone marrow cells from the control and Vα14-Jα18 mouse were stained with anti-TCRVβ and αGalCer-CD1 dimers. The number in the figure represents the percentage of iNKT cells. One representative data from three experiments (three to four mice per experiment) is shown. Littermates harboring the nonrearranged TCR locus or C57BL/6 were used as control, both of which resulted in essentially the same profile. b, Distribution of iNKT cells in the Vα14-Jα18 mouse thymocytes. Thymocytes from Vα14-Jα18 mouse were tricolor-stained with anti-CD4, anti-CD8 and αGalCer- or vehicle-loaded CD1d dimers. The number in the figure represents the percentage of the each subset (upper panel, CD4 vs CD8). The percentage of iNKT cells (bold line) over the control staining (dotted line) within the each subset is shown (lower panels). c, Cell surface marker expressions on iNKT cells in Vα14-Jα18 mouse. Thymocytes from the control and from Vα14-Jα18 mouse were stained with anti-TCRVβ and αGalCer-CD1d dimers together with the indicated mAbs. Histograms display the frequency of the cells expressing the indicated markers among TCRVβ+/αGalCer-CD1d dimer+ gated cells (bold line) over the control staining (dotted line). The number in figure represents the percentage of iNKT cells expressing the indicated molecule.

FIGURE 2.

a, Predominance of αGalCer-CD1d dimer-reactive NKT (iNKT) cells in Vα14-Jα18 mice. Thymocytes, spleen cells, liver mononuclear cells, and bone marrow cells from the control and Vα14-Jα18 mouse were stained with anti-TCRVβ and αGalCer-CD1 dimers. The number in the figure represents the percentage of iNKT cells. One representative data from three experiments (three to four mice per experiment) is shown. Littermates harboring the nonrearranged TCR locus or C57BL/6 were used as control, both of which resulted in essentially the same profile. b, Distribution of iNKT cells in the Vα14-Jα18 mouse thymocytes. Thymocytes from Vα14-Jα18 mouse were tricolor-stained with anti-CD4, anti-CD8 and αGalCer- or vehicle-loaded CD1d dimers. The number in the figure represents the percentage of the each subset (upper panel, CD4 vs CD8). The percentage of iNKT cells (bold line) over the control staining (dotted line) within the each subset is shown (lower panels). c, Cell surface marker expressions on iNKT cells in Vα14-Jα18 mouse. Thymocytes from the control and from Vα14-Jα18 mouse were stained with anti-TCRVβ and αGalCer-CD1d dimers together with the indicated mAbs. Histograms display the frequency of the cells expressing the indicated markers among TCRVβ+/αGalCer-CD1d dimer+ gated cells (bold line) over the control staining (dotted line). The number in figure represents the percentage of iNKT cells expressing the indicated molecule.

Close modal
Table I.

Lymphocyte number (means± SD)

Vα14-Jα18 Mouse (n = 4), %Control Mouse (n = 4), %
Thymocytes 1.6 × 107 ± 10 1.1 × 108 ± 11 
Spleen cells 6.0 × 107 ± 12 1.1 × 108 ± 16 
Liver mononuclear cells 9.0 × 105 ± 15 2.0 × 106 ± 20 
Vα14-Jα18 Mouse (n = 4), %Control Mouse (n = 4), %
Thymocytes 1.6 × 107 ± 10 1.1 × 108 ± 11 
Spleen cells 6.0 × 107 ± 12 1.1 × 108 ± 16 
Liver mononuclear cells 9.0 × 105 ± 15 2.0 × 106 ± 20 

We analyzed the expression of the other markers on iNKT cells in the thymus. The profile for the NKRs such as Ly49A, Ly49D, and CD25 in Vα14-Jα18 mice was almost equivalent to that found in the control thymocytes (Fig. 2,c, data not shown). In contrast, CD69+ and NK1.1+ cells decreased concomitant with an increase of CD44dull cells, while CD24+ cells increased in the Vα14-Jα18 mouse thymocytes (Fig. 2 c).

We next evaluated the potential of iNKT cells to respond to αGalCer and the resulting production of the cytokines in the serum. Without any stimulation, no cytokine was produced in both control and Vα14-Jα18 mice. Upon αGalCer stimulation, however, both types of mice simultaneously produced Th1-type cytokines such as IL-2 and IFN-γ, Th2-type cytokines, IL-4, IL-5, and IL-10, and inflammatory cytokines such as IL-1β and TNF-α (Fig. 3,a). Cytokine production peaked at 4 h poststimulation for all the cytokines. Vα14-Jα18 mice produced >10 times cytokines such as IL-2, IL-4, IL-10, GM-CSF, and IFN-γ relative to the control at 4 h. In contrast, no significant increment in the amount of IL-1β and IL-5 was detected in Vα14-Jα18 mice under the same conditions (Fig. 3 a).

FIGURE 3.

a, Cytokine production upon αGalCer stimulation. The concentration of cytokines in the serum was measured at the indicated time points using 3-mo-old Vα14-Jα18 mice and the age-matched littermates (control). Four individuals were sampled at each time points. Data are means ± SD. b, IL-4 and IFN-γ production upon in vitro DC stimulation. Spleen cells from the control and from Vα14-Jα18 mice were mixed with αGalCer-pulsed DC. The concentration of IFN-γ and IL-4 in the resultant supernatant is shown. Data are means from three experiments ± SD (three wells per experiment). c, Production of IL-4 and IFN-γ from the Vα14-Jα18 mouse spleen cells. Predominance of iNKT cells in TCRVβ+/TCRVβ8+ cells (upper panels). Vα14-Jα18 mouse spleen cells were stained with anti-TCRVβ, anti-TCRVβ8 Abs, and αGalCer- or vehicle-loaded CD1d dimers, and examined for TCRVβ vs TCRVβ8 expression (upper left panel). The number in the figure indicates the percentage of TCRVβ+/TCRVβ8+ cells. The percentage of αGalCer-CD1d reactive cells (bold line) over control-stained cells (dotted line) among the TCRVβ+/TCRVβ8+ cells is shown (upper right panel). IL-4 and IFN-γ production from the TCRVβ8+ cells (lower panels). The percentage of cells positive for the indicated cytokine staining (bold line) relative to the isotype control-stained cells (dotted line) is shown. One representative data from three experiments is shown. d, Slightly Th2-biased Ig isotypes in Vα14-Jα18 mice. Serum level of the each Ig isotype was measured by ELISA in 3-mo-old Vα14-Jα18 mice and littermate control housed under the same conditions. Four individuals were examined and means ± SD are shown. e, γδ and αβT cells in the iIEL. iIEL were prepared from the control and Vα14-Jα18 mouse intestine, and stained with anti-TCRVβ and anti-TCRδ. Numbers in the quadrant represent the percentage of each subset ±SD.

FIGURE 3.

a, Cytokine production upon αGalCer stimulation. The concentration of cytokines in the serum was measured at the indicated time points using 3-mo-old Vα14-Jα18 mice and the age-matched littermates (control). Four individuals were sampled at each time points. Data are means ± SD. b, IL-4 and IFN-γ production upon in vitro DC stimulation. Spleen cells from the control and from Vα14-Jα18 mice were mixed with αGalCer-pulsed DC. The concentration of IFN-γ and IL-4 in the resultant supernatant is shown. Data are means from three experiments ± SD (three wells per experiment). c, Production of IL-4 and IFN-γ from the Vα14-Jα18 mouse spleen cells. Predominance of iNKT cells in TCRVβ+/TCRVβ8+ cells (upper panels). Vα14-Jα18 mouse spleen cells were stained with anti-TCRVβ, anti-TCRVβ8 Abs, and αGalCer- or vehicle-loaded CD1d dimers, and examined for TCRVβ vs TCRVβ8 expression (upper left panel). The number in the figure indicates the percentage of TCRVβ+/TCRVβ8+ cells. The percentage of αGalCer-CD1d reactive cells (bold line) over control-stained cells (dotted line) among the TCRVβ+/TCRVβ8+ cells is shown (upper right panel). IL-4 and IFN-γ production from the TCRVβ8+ cells (lower panels). The percentage of cells positive for the indicated cytokine staining (bold line) relative to the isotype control-stained cells (dotted line) is shown. One representative data from three experiments is shown. d, Slightly Th2-biased Ig isotypes in Vα14-Jα18 mice. Serum level of the each Ig isotype was measured by ELISA in 3-mo-old Vα14-Jα18 mice and littermate control housed under the same conditions. Four individuals were examined and means ± SD are shown. e, γδ and αβT cells in the iIEL. iIEL were prepared from the control and Vα14-Jα18 mouse intestine, and stained with anti-TCRVβ and anti-TCRδ. Numbers in the quadrant represent the percentage of each subset ±SD.

Close modal

We also examined whether stimulation with αGalCer-pulsed DCs resulted in enhanced cytokine production in vitro. DCs prepared from the spleen of Vα14-Jα18 or control mice were cocultured with the whole spleen cells from these mice. Coculture of the control spleen cells with DCs from either control or Vα14-Jα18 mice led to a modest production of IFN-γ and IL-4 (Fig. 3,b). In contrast, irrespective of the origin of DCs, use of spleen cells from Vα14-Jα18 mice resulted in a 5- and 10-fold increase for IL-4 and IFN-γ, respectively (Fig. 3,b). This increment mirrored well the abundance of iNKT cells in the Vα14-Jα18 mouse spleen. The production of IL-4 and IFN-γ from the splenic iNKT cells was confirmed by intracellular staining (Fig. 3 c).

Because Vα14-Jα18 Tg mice harbor a Th2-biased Ig isotype in serum, we examined whether this was the case for Vα14-Jα18 mice (14). A modest increase in Th2-type IgG1 (5-fold above controls on averages) concomitant with a decrease in Th1-type IgG2a production (one-tenth relative to the controls on average) was detected. Nevertheless, there was little bias in the amount of IgE between Vα14-Jα18 mice and the littermates (Fig. 3 d).

To exclude the possibility that the increase of iNKT cells in the thymus was due to the results of αβ vs γδ lineage commitment competition (17) or to exclusive transcription of the allele harboring the rearranged Vα14-Jα18 whatever be the reason, γδT cells in the intestinal intraepithelial lymphocytes (iIEL) were analyzed. It turned out that the cell number of the each subset and the ratio of αβ:γδ in Vα14-Jα18 mice were quasiequivalent to those of the control mice (Fig. 3,e), and there were γδT cells in the thymocytes (data not shown). These experiments also demonstrated that γδ T cells stemmed from the other nonrearranged allele (Fig. 1 b). When B and NK cells from Vα14-Jα18 mice were examined, little difference in the cell number and in the receptor expression profile was noticed relative to the controls (data not shown). Together, these data suggested that the in-frame rearranged Vα14-Jα18 locus mainly affects T lymphocyte development.

Mice deficient for CD1d are devoid of iNKT cells, as CD1d is responsible for the development of iNKT and non-iNKT cells (11, 12, 13). iNKT cells present in the periphery express the coreceptor such as CD44, CD69, and NK1.1. Studies on the ontogeny of iNKT cells have indicated that iNKT cells emerge from CD24high/NK1.1/CD44low cells in the thymus, and they acquire the CD24low/NK1.1+/CD44high phenotype upon emigration from the thymus (18, 19). We have mated Vα14-Jα18 mice with CD1d−/− mice to obtain mice harboring the hemizygous Vα14-Jα18 and no CD1d (Vα14-Jα18/CD1d−/− mouse). If a Vα14-Jα18 TCR signal is critical for iNKT cell development, an immature iNKT cell population would be tangible on the CD1d−/− background. Analyses of the Vα14-Jα18/CD1d−/− mouse revealed that ∼1% of the thymocytes were stained with αGalCer-CD1d dimers, harboring a little expression of CD69, but little NK1.1, denoting that NK1.1 expression is most likely dependent on CD1d. In contrast, CD24- and CD44low-expressing cells were abundant in the population (Fig. 4,a, compare Figs. 2,c and 4 b). Thus, immature iNKT cells that did not react with CD1d were CD24+CD44dullCD69+/−NK1.1. It would be intriguing to compare these cells with Vα14-Jα18 Tg DN cells in CD1d−/− mice (20).

FIGURE 4.

a, Immature iNKT cells in the thymocytes of mice harboring the Vα14-Jα18 locus on CD1d−/−. Thymocytes from the control (C57BL/6), CD1d−/− mouse, and mouse harboring a hemizygous Vα14-Jα18 on CD1d−/− background were stained as described in Fig. 2,a to visualize CD1d-nonreacted immature iNKT cells. One representative data from five mice are shown. The number in the figure represents the percentage of αGalCer-CD1d reactive iNKT cells. b, Cell surface marker expressions on immature iNKT cells. Thymocytes from Vα14-Jα18/CD1d−/− mouse were stained as described in Fig. 2,c. The number in the figure represents the percentage of immature iNKT cells expressing the indicated marker (bold line) over the isotype control (dotted line). c, CD1d-dependent TCRVβ8 usage. Thymocytes from the Vα14-Jα18 and Vα14-Jα18/CD1d−/− mouse were stained as described in Fig. 2 a together with anti-TCRVβ8 (left panel). The histogram shows the percentage of αGalCer-CD1d reactive iNKT cells (bold line) over the control-stained cells (dotted line) within the TCRVβ+/TCRVβ8+ cells (right panel).

FIGURE 4.

a, Immature iNKT cells in the thymocytes of mice harboring the Vα14-Jα18 locus on CD1d−/−. Thymocytes from the control (C57BL/6), CD1d−/− mouse, and mouse harboring a hemizygous Vα14-Jα18 on CD1d−/− background were stained as described in Fig. 2,a to visualize CD1d-nonreacted immature iNKT cells. One representative data from five mice are shown. The number in the figure represents the percentage of αGalCer-CD1d reactive iNKT cells. b, Cell surface marker expressions on immature iNKT cells. Thymocytes from Vα14-Jα18/CD1d−/− mouse were stained as described in Fig. 2,c. The number in the figure represents the percentage of immature iNKT cells expressing the indicated marker (bold line) over the isotype control (dotted line). c, CD1d-dependent TCRVβ8 usage. Thymocytes from the Vα14-Jα18 and Vα14-Jα18/CD1d−/− mouse were stained as described in Fig. 2 a together with anti-TCRVβ8 (left panel). The histogram shows the percentage of αGalCer-CD1d reactive iNKT cells (bold line) over the control-stained cells (dotted line) within the TCRVβ+/TCRVβ8+ cells (right panel).

Close modal

Previous studies showed that the usage of TCRVβ in iNKT cells is highly biased to TCRVβ8, 7, and 5 (1, 2, 3). The analysis of TCRVβ+/TCRVβ8+ cells revealed that >80% of this population comprised iNKT cells (Fig. 4,c, upper panel). A similar tendency was observed in TCRVβ+/TCRVβ7+ cells but not in TCRVβ+/TCRVβ5+ cells (data now shown). These data indicated that TCRVβ8 was the preferred repertoire of iNKT cells in Vα14-Jα18 mice. In contrast, only ∼20% of TCRVβ+/TCRVβ8+ cells were iNKT cells in the Vα14-Jα18/CD1d−/− thymocytes (Fig. 4 c, lower panel).

We next examined whether there was a difference in further rearrangement between immature iNKT and the immature conventional T cells. Rag2 expression was examined with RT-PCR using TCRVβmed/TCRVβ8med and TCRVβmed/TCRVβ8 populations, both of which represent DP cells. A population stained with αGalCer-CD1d dimers in DP is considered as immature iNKT cells under maturation, while cells not stained with this reagent are considered as the conventional T cells under positive selection. RT-PCR analysis revealed that immature iNKT cells expressed less Rag2 than the conventional T cells in both TCRVβmed/TCRVβ8med and TCRVβmed/TCRVβ8 populations (Fig. 5,a). As an internal control, there was no difference in the expression of Myd118 and Hprt between αGalCer-CD1d dimer-reactive and nonreactive cells (Fig. 5 a).

Because the above data indicated that immature iNKT cells less frequently perform further rearrangement relative to the conventional immature T cells, it was necessary to examine whether a particular subset of the thymocytes proliferated more vigorously than the others, and/or whether there was massive apoptosis to give grounds for the preponderance of iNKT cells in the thymus. We, therefore, have examined the cell cycle status of the thymocytes. When TCRVβhigh/TCRVβ8 cells (mature T cells) from the control mouse thymocytes were stained with propidium iodide, >98% of cells were in the G0/G1 phase (Fig. 5,b, 1, control). This was also true for TCRVβhigh/TCRVβ8+ cells (Fig. 5,b, 4). In TCRVβmed/TCRVβ8 cells that represent the immature T cells, ∼3% of cells were actively cycling (Fig. 5,b, 2). On the contrary, >9% of TCRVβ/TCRVβ8 cells that mostly represent DN cells were in the cycling status (Fig. 5,b, 3). In Vα14-Jα18 mice, the percentage of the cycling cells was almost equivalent to that of the control for each subset except TCRVβ/TCRVβ8 cells in which >15% were actively cycling (Fig. 5,b, 3, Vα14-Jα18). This may reflect the fact that the absolute number of DN cells increased ∼2-fold in Vα14-Jα18 mice (Fig. 2,b and Table I). The analyses revealed that no subset of the thymocytes proliferated more vigorously than the others within the Vα14-Jα18 mouse thymocytes concomitant with no extensive apoptosis (Fig. 5 b, Vα14-Jα18).

Our present data showed that Vα14-Jα18 locus has a large impact on the development of T lymphocytes in the thymus accompanying a significant increase in iNKT cells (Fig. 2 a). Cytokine production experiments suggested that the function of DC remained intact in Vα14-Jα18 mice. Analysis of DC markers such as CD40, CD80, CD86, CD1d, and I-Ab on DC revealed no significant difference in their expression profiles relative to the control mice (data not shown). From these data, we could conclude that the functions of iNKT cells in Vα14-Jα18 mice mimic those found in the control mouse. Thus, Vα14-Jα18 mice could represent a novel model to study further the roles of iNKT cells in the immune system.

In line with the assumption that iNKT cells pass through DP thymocytes, the recent fate-mapping study has demonstrated that indeed these cells derive from DP (21). Nonetheless, there remains a fundamental question as to how iNKT cells develop in the thymus. The identification of immature iNKT cells in Vα14-Jα18/CD1d−/− mice underpinned the phenotype described in the other studies (Fig. 4,b) (18, 19). CD1d may be responsible for the expression of NK1.1, CD44, and CD69, and for down-regulation of CD24 (compare Figs. 2,c and 4 b). The fact that iNKT precursor cells undergo extensive proliferation during their maturation suggests that CD1d is required for iNKT cell survival/proliferation (10, 18). Alternatively, it may serve as an extrinsic stimulus that instructs the development of iNKT cells.

Extrinsic factors can impact on the consequence of fate decision by multipotent progenitors through two distinct mechanisms. In a permissive mechanism, the cell fate decision of progenitors to differentiate into a particular lineage is made independent of extrinsic factors, and their raison-d’être is to support the survival and proliferation of the committed cells. On the contrary, in an instructive mechanism, extrinsic factors impose upon the progenitors to select one lineage at the expense of others (22). The intensive scrutiny in the neural crest stem cells has demonstrated that neural cell fate is instructively determined by the extrinsic factors such as bone marrow protein 2, neuregulin, TGFβ, and cilial neurotrophic factor (23, 24, 25).

In lymphopoiesis, AgR signals play a pivotal role in the development of lymphocytes. However, the relative contribution of the above two mechanisms in determining the fate of progenitors has been a matter of debate. It is proposed that formation of the pre-TCR complex composed of TCRVβ together with pTα and CD3ε instructively induces αβT cells over γδT cells, based on the fact that pTα knockout mice contain less in-frame rearranged TCRVβ-harboring cells relative to the control (17). Nonetheless, it can be interpreted that such a pre-TCR complex formation merely supports the survival and/or proliferation of the committed αβT cells, indicating the permissive role for the above pre-TCR complex. Another report has suggested that the αβ-γδT cell lineage diversification occurs well before the TCR rearrangement, rather supporting the permissive role for TCR signals in the αβT cell lineage commitment (26).

Similarly, the mode of the cell fate decision to be helper (CD4) or cytotoxic (CD8) T cells has also been a long-standing dispute. Transgenic expression of CD8 substantially improves MHC class I-specific CD8 selection while affecting little on class II-specific CD4 selection, arguing that the CD8 cell lineage commitment is instructively determined (27). Subsequent study on MHC class I knockout mice revealed that indeed there are no CD8 cells. However, reintroduction of CD4 using the transgene into such mice restored CD8 cells and vice versa (28, 29). These studies simply denote that the lineage choice to be CD4 or CD8 cells is not dictated by CD4 or CD8 per se; it conferred room for an interpretation that signals elicited from these molecules permissively contribute to the survival and/or proliferation of the committed CD4 or CD8 cells. These seemingly conflicting interpretations stem from the fact that transgenic expression or ablation of the signaling molecules tend to support their permissive action on the cell fate decision through providing or depriving of survival signals.

Another caveat to address the above issue is that the lymphocyte progenitors proliferate more vigorously than the neural progenitors. Because the presence of precommitted and committed cells is a prerequisite for addressing the issue, such a propensity often blunts the interpretation. It is thus imperative to examine whether the lineage precommitted cells already predominated the lymphocytes early in development. Hitherto, due to lack of appropriate assays, it has not been possible to address properly whether the AgR signals instructively or permissively affect the destiny of the progenitor cells.

The analysis on the Vα14-Jα18 mouse thymocytes revealed that iNKT cells comprise CD1-interacted and CD1d-noninteracted cells as evidenced by the presence of CD24low/NK1.1+/CD44high and CD24high/NK1.1/CD44low cells, respectively (Figs. 2,c, 4,b, and 6, a and b). This antecedent allows us to investigate further whether the development of iNKT cell is determined instructively or permissively by CD1d through the interaction with the Vα14-Jα18TCR. Our present data rather support the former possibility based on the following observations. 1) The preponderance of iNKT cells in the thymus (Fig. 2,a). 2) The CD1d-dependent preferential association of Vα14-Jα18TCR with Vβ8 (Fig. 4,c). 3) Immature iNKT cells found in DP express less Rag2 relative to the conventional T cells under positive selection (Fig. 5,a). 4) Few subsets of the thymocytes proliferate more vigorously than the others, and show a massive apoptosis (Fig. 5,b). These data are in line with the instructive mode of iNKT cell development in that upon interaction with CD1d, immature iNKT cells obey their fate to be iNKT cells without striving to change the TCRα repertoire by operating the recombination machinery, and without dying massively by apoptosis (Fig. 6 c). Although further study has yet to be performed to rigorously prove the above hypothesis, the Vα14-Jα18 mouse is an invaluable model to shed light on the mode of iNKT cell development.

FIGURE 6.

a, iNKT cells before/after CD1d interaction. The status of iNKT cells before and after interaction with CD1d is schematically depicted. Half-tone text shows that the commitment has not yet completed, thereby the nascent Vα14-Jα18 TCR may be subject to further rearrangement (left panel, CD1d-nonreacted). Plain text indicates that the commitment is over upon CD1d− (Vα14-Jα18 TCR) interaction and further rearrangement of the Vα14-Jα18 TCR that culminates in conversion to Th or cytotoxic T cell (Th/Tc) lineage is blocked (right panel, CD1d-reacted). b, Schematic representation of iNKT cells in the thymocytes. CD1d-nonreacted (half tone text and open circle) and CD1d-reacted (plain text and shaded circle) iNKT cells in the thymocytes are shown. However, the relative abundance of the CD1d-nonreacted cells over the CD1d-reacted cells is different in each thymocyte. In the control thymocytes, a few CD1d-nonreacted immature iNKT are present (left panel), while the Vα14-Jα18 thymocytes consist of CD1d-nonreacted and CD1d-reacted iNKT cells (middle panel). In contrast, no CD1d-reacted iNKT cells are present in the Vα14-Jα18/CD1d−/− mouse thymocytes (right panel). Note that the number of CD1d-nonreacted and of CD1d-reacted iNKT cells depicted in the figure does not necessarily reflect the real number of the each subset. c, Instructive vs permissive cell commitment. Instructive model (left): The (Vα14-Jα18TCR)-CD1d signal instructs the multipotent progenitors (NKT/Th/Tc) to commit to be iNKT cells at the expense of the Th and CTLs (half tone). Permissive model (right): In contrast, the (Vα14-Jα18TCR)-CD1d signal solely supports the survival and proliferation of the committed iNKT cells. Note that the commitment of iNKT cells from the multipotent progenitors is determined independent of the (Vα14-Jα18TCR)-CD1d signal. In this case, simultaneously generated Th and CTLs that happened to express the Vα14-Jα18 TCR should be deleted by apoptosis, or they should alter Vα14-Jα18 TCR to change their fate by operating recombination machinery to give ground for the predominance of the thymic iNKT cells upon (Vα14-Jα18 TCR)-CD1d signal.

FIGURE 6.

a, iNKT cells before/after CD1d interaction. The status of iNKT cells before and after interaction with CD1d is schematically depicted. Half-tone text shows that the commitment has not yet completed, thereby the nascent Vα14-Jα18 TCR may be subject to further rearrangement (left panel, CD1d-nonreacted). Plain text indicates that the commitment is over upon CD1d− (Vα14-Jα18 TCR) interaction and further rearrangement of the Vα14-Jα18 TCR that culminates in conversion to Th or cytotoxic T cell (Th/Tc) lineage is blocked (right panel, CD1d-reacted). b, Schematic representation of iNKT cells in the thymocytes. CD1d-nonreacted (half tone text and open circle) and CD1d-reacted (plain text and shaded circle) iNKT cells in the thymocytes are shown. However, the relative abundance of the CD1d-nonreacted cells over the CD1d-reacted cells is different in each thymocyte. In the control thymocytes, a few CD1d-nonreacted immature iNKT are present (left panel), while the Vα14-Jα18 thymocytes consist of CD1d-nonreacted and CD1d-reacted iNKT cells (middle panel). In contrast, no CD1d-reacted iNKT cells are present in the Vα14-Jα18/CD1d−/− mouse thymocytes (right panel). Note that the number of CD1d-nonreacted and of CD1d-reacted iNKT cells depicted in the figure does not necessarily reflect the real number of the each subset. c, Instructive vs permissive cell commitment. Instructive model (left): The (Vα14-Jα18TCR)-CD1d signal instructs the multipotent progenitors (NKT/Th/Tc) to commit to be iNKT cells at the expense of the Th and CTLs (half tone). Permissive model (right): In contrast, the (Vα14-Jα18TCR)-CD1d signal solely supports the survival and proliferation of the committed iNKT cells. Note that the commitment of iNKT cells from the multipotent progenitors is determined independent of the (Vα14-Jα18TCR)-CD1d signal. In this case, simultaneously generated Th and CTLs that happened to express the Vα14-Jα18 TCR should be deleted by apoptosis, or they should alter Vα14-Jα18 TCR to change their fate by operating recombination machinery to give ground for the predominance of the thymic iNKT cells upon (Vα14-Jα18 TCR)-CD1d signal.

Close modal

We thank H. Fujimoto for FACS sorting, Y. Nagata for technical assistance, Dr. M. Taniguchi for laboratory facility, and the members of the immune regulation laboratory for discussion.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by RIKEN and in part by grants from the Ministry of Education, Culture, Sports, Science and Technology of Japan (to A. Ogura).

3

Abbreviations used in this paper: iNKT, invariant NKT; DP, double positive; αGalCer, αGalactosylceramide; Tg, transgenic; SP, single positive; DC, dendritic cell; HPRT, hypoxanthine phosphoribosyltransferase; DN, double negative; iIEL, intestinal intraepithelial lymphocyte.

1
Bendelac, A., M. N. Rivera, S. H. Park, J. H. Roark.
1997
. Mouse CD1-specific NK1 T cells: development, specificity, and function.
Annu. Rev. Immunol.
15
:
535
-562.
2
Kronenberg, M., L. Gapin.
2002
. The unconventional lifestyle of NKT cells.
Nat. Rev. Immunol.
2
:
557
-568.
3
Taniguchi, M., M. Harada, S. Kojo, T. Nakayama, H. Wakao.
2003
. The regulatory role of Vα14 NKT cells in innate and acquired immune response.
Annu. Rev. Immunol.
21
:
483
-513.
4
Lehuen, A., O. Lantz, L. Beaudoin, V. Laloux, C. Carnaud, A. Bendelac, J. F. Bach, R. C. Monteiro.
1998
. Overexpression of natural killer T cells protects Vα14-Jα281 transgenic nonobese diabetic mice against diabetes.
J. Exp. Med.
188
:
1831
-1839.
5
Poulton, L. D., M. J. Smyth, C. G. Hawke, P. Silveira, D. Shepherd, O. V. Naidenko, D. I. Godfrey, A. G. Baxter.
2001
. Cytometric and functional analyses of NK and NKT cell deficiencies in NOD mice.
Int. Immunol.
13
:
887
-896.
6
Bendelac, A..
1995
. Positive selection of mouse NK1+ T cells by CD1-expressing cortical thymocytes.
J. Exp. Med.
182
:
2091
-2096.
7
Coles, M. C., D. H. Raulet.
2000
. NK1.1+ T cells in the liver arise in the thymus and are selected by interactions with class I molecules on CD4+CD8+ cells.
J. Immunol.
164
:
2412
-2418.
8
Benlagha, K., A. Weiss, A. Beavis, L. Teyton, A. Bendelac.
2000
. In vivo identification of glycolipid antigen-specific T cells using fluorescent CD1d tetramers.
J. Exp. Med.
191
:
1895
-1903.
9
Matsuda, J. L., O. V. Naidenko, L. Gapin, T. Nakayama, M. Taniguchi, C. R. Wang, Y. Koezuka, M. Kronenberg.
2000
. Tracking the response of natural killer T cells to a glycolipid antigen using CD1d tetramers.
J. Exp. Med.
192
:
741
-754.
10
Gapin, L., J. L. Matsuda, C. D. Surh, M. Kronenberg.
2001
. NKT cells derive from double-positive thymocytes that are positively selected by CD1d.
Nat. Immunol.
2
:
971
-978.
11
Chen, Y. H., N. M. Chiu, M. Mandal, N. Wang, C. R. Wang.
1997
. Impaired NK1+ T cell development and early IL-4 production in CD1-deficient mice.
Immunity
6
:
459
-467.
12
Mendiratta, S. K., W. D. Martin, S. Hong, A. Boesteanu, S. Joyce, L. Van Kaer.
1997
. CD1d1 mutant mice are deficient in natural T cells that promptly produce IL-4.
Immunity
6
:
469
-477.
13
Skold, M., N. N. Faizunnessa, C. R. Wang, S. Cardell.
2000
. CD1d-specific NK1.1+ T cells with a transgenic variant TCR.
J. Immunol.
165
:
168
-174.
14
Bendelac, A., R. D. Hunziker, O. Lantz.
1996
. Increased interleukin 4 and immunoglobulin E production in transgenic mice overexpressing NK1 T cells.
J. Exp. Med.
184
:
1285
-1293.
15
Capone, M., D. Cantarella, J. Schumann, O. V. Naidenko, C. Garavaglia, F. Beermann, M. Kronenberg, P. Dellabona, H. R. MacDonald, G. Casorati.
2003
. Human invariant Vα24-JαQ TCR supports the development of CD1d-dependent NK1.1+ and NK1.1 T cells in transgenic mice.
J. Immunol.
170
:
2390
-2398.
16
Inoue, K., H. Wakao, N. Ogonuki, H. Miki, K. Seino, R. Nambu-Wakao, S. Noda, H. Miyoshi, H. Koseki, M. Taniguchi, A. Ogura.
2005
. Generation of cloned mice by direct nuclear transfer from natural killer T cells.
Curr. Biol.
15
:
1114
-1118.
17
Aifantis, I., O. Azogui, J. Feinberg, C. Saint-Ruf, J. Buer, H. von Boehmer.
1998
. On the role of the pre-T cell receptor in αβ versus γδ T lineage commitment.
Immunity
9
:
649
-655.
18
Benlagha, K., T. Kyin, A. Beavis, L. Teyton, A. Bendelac.
2002
. A thymic precursor to the NK T cell lineage.
Science
296
:
553
-555.
19
Benlagha, K., D. G. Wei, J. Veiga, L. Teyton, A. Bendelac.
2005
. Characterization of the early stages of thymic NKT cell development.
J. Exp. Med.
202
:
485
-492.
20
Wei, D. G., S. A. Curran, P. B. Savage, L. Teyton, A. Bendelac.
2006
. Mechanisms imposing the Vβ bias of Vα14 natural killer T cells and consequences for microbial glycolipid recognition.
J. Exp. Med.
203
:
1197
-1207.
21
Egawa, T., G. Eberl, I. Taniuchi, K. Benlagha, F. Geissmann, L. Hennighausen, A. Bendelac, D. R. Littman.
2005
. Genetic evidence supporting selection of the Vα14i NKT cell lineage from double-positive thymocyte precursors.
Immunity
22
:
705
-716.
22
Morrison, S. J., N. M. Shah, D. J. Anderson.
1997
. Regulatory mechanisms in stem cell biology.
Cell
88
:
287
-298.
23
Shah, N. M., M. A. Marchionni, I. Isaacs, P. Stroobant, D. J. Anderson.
1994
. Glial growth factor restricts mammalian neural crest stem cells to a glial fate.
Cell
77
:
349
-360.
24
Shah, N. M., A. K. Groves, D. J. Anderson.
1996
. Alternative neural crest cell fates are instructively promoted by TGFβ superfamily members.
Cell
85
:
331
-343.
25
Johe, K. K., T. G. Hazel, T. Muller, M. M. Dugich-Djordjevic, R. D. McKay.
1996
. Single factors direct the differentiation of stem cells from the fetal and adult central nervous system.
Genes Dev.
10
:
3129
-3140.
26
Kang, J., A. Volkmann, D. H. Raulet.
2001
. Evidence that γδ versus αβ T cell fate determination is initiated independently of T cell receptor signaling.
J. Exp. Med.
193
:
689
-698.
27
Robey, E. A., B. J. Fowlkes, J. W. Gordon, D. Kioussis, H. von Boehmer, F. Ramsdell, R. Axel.
1991
. Thymic selection in CD8 transgenic mice supports an instructive model for commitment to a CD4 or CD8 lineage.
Cell
64
:
99
-107.
28
Davis, C. B., N. Killeen, M. E. Crooks, D. Raulet, D. R. Littman.
1993
. Evidence for a stochastic mechanism in the differentiation of mature subsets of T lymphocytes.
Cell
73
:
237
-247.
29
Robey, E., A. Itano, W. C. Fanslow, B. J. Fowlkes.
1994
. Constitutive CD8 expression allows inefficient maturation of CD4+ helper T cells in class II major histocompatibility complex mutant mice.
J. Exp. Med.
179
:
1997
-2004.