Intrathymic proliferation wave essential for Vα14⁺ natural killer T cell development depends on c-Myc

Abstract

The molecular requirements for invariant Vα14-bearing natural killer T cells (iNKT) in the thymus are poorly understood. A minute population of ≈500 newly selected CD69⁺CD24⁺ stage 0 (ST0) iNKT cells gives rise to ≈100 times more CD44^neg/loCD24⁻ stage 1 (ST1) cells, which then generate similar frequencies of CD44^hiCD24⁻ stage 2 (ST2) and mature iNKT cells. Although the increased number of ST1 compared with ST0 cells indicates the initiation of a proliferation wave in the very early stages of iNKT cell development, details about the controlling mechanism are currently lacking. Here, we show that the transcription factor c-Myc is required for iNKT cell development. Conditional ablation of c-Myc in double-positive thymocytes specifically impacted iNKT but not conventional T cell development. Within the iNKT population, a progressive reduction of iNKT cells was observed starting at ST1 (≈50-fold) and ST2 (≈350-fold), with a complete lack of mature cells in thymus, spleen, and liver. ST0/ST1 c-Myc-deficient iNKT cells showed reduced proliferation. In contrast, annexin V staining did not reveal increased apoptosis, and transgenic overexpression of BCL-2 did not rescue iNKT cell development in c-Myc-deficient mice. Moreover, expression of known iNKT differentiation factors such as Plzf and Gata3 was not dramatically altered. These, findings provide compelling evidence that c-Myc mediates an intrathymic proliferation wave immediately after agonist selection of iNKT cells and illustrate the importance of this expansion for the generation of mature iNKT cells in vivo.

Natural killer T cells develop in the thymus and are characterized by the expression of various surface molecules originally detected on natural killer (NK) cells (1). The majority of murine NKT cells, referred to as iNKT, express a semiinvariant T cell receptor (TCR) repertoire with invariant Vα14-Jα18 usage in combination with Vβ8, Vβ7, or Vβ2 (2). Most exogenous ligands for NKT cells identified so far are components of the cell wall of Gram-negative bacteria, implying an important role in innate immunity, but they also display autoreactivity and have been implicated in autoimmune disease and cancer (reviewed in ref. 1).

iNKT cells are positively selected on CD1d, an MHCI-like molecule expressed on cortical CD4⁺CD8⁺ [double-positive (DP)] thymocytes in an agonist selection process involving endogenous ligands, including the glycosphingolipid isoglobotrihexosyl ceramide (iGb3) (3–5). iNKT cell selection is profoundly different from positive selection of conventional T cells, and coreceptor interactions are likely involved in the process. As an example, homophilic interactions of SLAM and Ly108 surface molecules between iNKT and DP thymocytes induce activation of the Src kinase FynT leading to expansion and differentiation of immature iNKT cells (1).

The development of CD1d tetramers loaded with α-galactosyl ceramide (α-GalCer) (6, 7) as a tool to label iNKT cells allowed the identification of an extremely infrequent (≈1/10⁶) stage 0 (ST0) CD24^hiCD69⁺ precursor population thought to represent cells immediately after positive selection (8). As cells progress to the 100-fold more frequent CD44^low stage 1 (ST1), they expand in numbers and down-regulate CD24. Both ST1 and the subsequent CD44^high stage 2 (ST2) are NK1.1⁻. Most recent thymic emigrants resemble CD4⁺ ST2 cells and up-regulate NK cell markers only as they mature in the periphery (9, 10), whereas an independent set of long-lived CD44^highNK1.1⁺ stage 3 (ST3) cells develop in the thymus (11). Apart from NK1.1, DX5 that detects an epitope of α2 integrin (CD49b) can be used in conjunction with CD1d tetramers to identify later stage iNKT cells, although NK1.1⁺ and DX5⁺ populations are not identical (12, 13). iNKT cell stages display differences in cytokine mRNA expression and cytokine release on stimulation. Although ST1 cells produce predominantly IL-4, subsequent stages also acquire the capability to produce IFN-γ as they mature, as deduced from in vitro receptor stimulation experiments and cytokine reporter knockin mice (1, 14). Two recent reports describe the zinc finger transcription factor Plzf as an essential regulator of the iNKT cell effector phenotype (15, 16). Several other transcription factors have been implicated in iNKT cell development and function (17). However, despite strong evidence indicating that the infrequent CD44^low iNKT precursors undergo massive expansion as they develop in the thymus, the factors that control cell cycling and their overall impact on iNKT cell development remain unknown.

The basic region/helix–loop–helix/leucine zipper (bHLHZip) transcription factor myelocytomatosis oncogene (c-Myc) plays an integral role in proliferation, survival, and differentiation of normal and neoplastic cells. Myc binds E-box DNA motifs as a heterodimer with Max, resulting in cell cycle entry (18) and transcriptional activation or suppression of genes. c-Myc has been implicated in cell proliferation (19) as well as the control of cell growth. Its expression increases rapidly in response to growth factors, B cell receptor or TCR ligation, and conventional CD4 T cells expressing hypomorphic c-Myc alleles display profound defects in activation induced proliferation (20). We and others have shown that c-Myc is essential for development at the pre-TCR checkpoint (21, 22). c-Myc has also been reported to control the self-renewal of hematopoietic stem cells (HSCs) (23), and its conditional ablation in the bone marrow favored self-renewal over differentiation of HSCs in the stem cell niche (24).

Here, we show that c-Myc is essential for iNKT cell development. Conditional ablation of c-Myc at the DP stage in mice leads to a dramatic reduction of iNKT cells. CD44^low Myc-deficient iNKT cells are not prone to apoptosis but proliferate less, providing evidence that c-Myc is involved in controlling the proliferation wave of early iNKT cell development.

Results

Myc Deficiency Disproportionately Impairs iNKT Cell Development.

We have previously reported a strict requirement for the transcription factor c-Myc in the expansion of thymocytes undergoing β-selection (21). Here, we show that, although CD4Cre mediated deletion of c-Myc after β-selection has no significant effect on thymocyte development at large, it specifically prohibits the development of iNKT cells (Fig. 1). Cellularity and surface expression of CD4 and CD8 on thymocytes and splenocytes of BALB/c control and c-Myc-deficient thymi (2.1 ± 0.6 × 10⁸ vs. 1.3 ± 0.4 × 10⁸) and spleens (2.2 ± 0.5 × 10⁸ vs. 2.3 ± 0.7 × 10⁸) were comparable (Fig. 1A). An up to 50% reduction in thymic cellularity with a corresponding reduction in DP cells could sometimes be observed but was not statistically significant (1.7 ± 0.5 × 10⁸ vs. 1.0 ± 0.3 × 10⁸). Likewise, CD4Cre Myc^fl/fl mice tended to have fewer peripheral T cells, but only the reduction in CD4 single-positive (SP) cells was statistically significant (4.1 ± 1.3 × 10⁸ vs. 1.8 ± 0.8 × 10⁸). Thus, with the exception of mild defects, CD4Cre Myc^fl/fl mice displayed normal distribution of conventional T cell subsets in thymus and spleen.

Fig. 1.

Myc deficiency disproportionately impairs iNKT cell development. BALB/c control and CD4Cre Myc^fl/fl mice were analyzed at 4–8 weeks of age. (A) (Upper) Representative FACS plots depicting CD4 and CD8 surface expression on lymphocytes from thymus and spleen. (B) (Upper) Invariant iNKT cell populations in thymus, spleen, and liver. (A and B) (Lower) Absolute cell numbers are given as histograms. N = number of experimental mice per genotype indicated here. B, BALB/c; M, CD4Cre Myc^fl/fl. (C) Quantitative PCR for Myc mRNA expression was performed on cDNA obtained from the indicated cell populations. No Myc expression was detected in CD4 SP or iNKT ST1 cells from CD4Cre Myc^fl/fl mice. (D) iNKT cells from sublethally irradiated bone marrow chimeras were MACS enriched and analyzed by FACS. Host, Thy1.1⁺ BALB/c mice; donor, Thy1.2⁺ CD4Cre Myc^fl/fl mice. Analysis was performed 12 weeks after injection. N = 4.

To our surprise, however, we found an almost complete ablation of iNKT cells in thymus, spleen, and liver (Fig. 1B). Absolute numbers of iNKT cells were reduced almost 70-fold in the thymus (1.3 ± 0.5 × 10⁶ vs. 0.020 ± 0.015 × 10⁶) and 30-fold in the spleen (1.6 ± 0.4 × 10⁶ vs. 0.05 ± 0.03 × 10⁶), where they reached the limit of detection. Myc mRNA in CD4Cre Myc^fl/fl mice was below detection limits in both CD4 SP as well as the few remaining iNKT cells, indicating efficient deletion (Fig. 1C). We also looked at γδ-T cells as another small lymphocyte population and found increased numbers in thymus (3.7 ± 0.9 × 10⁵ vs. 9.9 ± 3.2 × 10⁵) and spleen (1.0 ± 0.4 × 10⁶ vs. 2.8 ± 0.8 × 10⁶) (Fig. 1B).

Thus, c-Myc ablation at the DP stage specifically and profoundly affected iNKT cells. To examine whether this defect was cell intrinsic or due to altered properties of c-Myc-deficient DP cells, which are essential for iNKT selection, we also analyzed sublethally irradiated Thy1.1⁺ BALB/c mice reconstituted with a 1:1 mixture of host and Thy1.2⁺ CD4Cre Myc^fl/fl donor bone marrow. These chimeras showed a disproportionate reduction of iNKT cells originating from CD4Cre Myc^fl/fl donors. Thymocyte preparations were enriched for iNKT cells with PBS57 loaded CD1d-tetramers (hereafter referred to as tetramer) using magnetic microbeads. The ratio of tetramer⁺ to tetramer⁻ host-derived (Thy1.2⁻) cells was 1:4, whereas it was 1:180 in donor-derived (Thy1.2⁺) cells (Fig. 1D), suggesting that iNKT cell development is controlled by c-Myc in a cell-intrinsic fashion.

c-Myc Expression Is Required During iNKT Cell Development.

To measure c-Myc protein levels in the different stages of iNKT cell development, we analyzed Myc^G/G mice (25). These knockin mice express an N-terminal c-Myc-GFP fusion protein and thus allow the detection of c-Myc protein expression by flow cytometry. iNKT cells were MACS enriched from pools of control and Myc^G/G mice by using tetramers and magnetic beads, and stained with B220, heat-stable antigen (HSA), CD44, and DX5 antibodies to discriminate the developmental stages 1–3 (Fig. 2A). Thymocyte suspensions were also stained for CD4, CD8, and TCRβ. As reported, c-Myc expression was highest in double-negative (DN) cells, containing the highly proliferating DN2 and DN4 cells. TCRβ⁻ DP (DP^lo) cells, mostly preselected DP cells, had the least amount of Myc-GFP, whereas ≈4% of cells expressed intermediate levels of c-Myc in TCRβ^high DP (DP^hi) and CD4 SP cells, indicating that these populations contain a small percentage of dividing cells. Interestingly, 46% of the ST1 and ≈32% of the ST2 and ST3 iNKT cells expressed intermediate levels of c-Myc-GFP. c-Myc expression levels were highest in ST1 cells (Fig. 2B), indicating that c-Myc might be up-regulated on selection to initiate a proliferative burst and then stabilize at intermediate levels. Notably and consistent with the proliferative burst required to expand the iNKT cell compartment on selection, the percentages of c-Myc-GFP intermediate-expressing cells are considerably higher in postselected iNKT cells than in postselected DP^hi thymocytes.

Fig. 2.

c-Myc expression in developing iNKT cells. (A) Flow cytometric analyses of Myc-GFP knockin mice. Overlay histograms show GFP expression compared to the corresponding population of control mice. Populations were gated as indicated. For iNKT cell subsets, thymocytes from 3 mice were pooled and magnetically enriched for CD1d-tetramer-positive cells before FACS analysis. Numbers indicate percentages of cells with intermediate and high levels of c-Myc-GFP in Myc^G/G mice. The experiment was repeated 3 times with 3 mice in each group, or pools of 3 mice for MACS enrichment. (B) Overlay histogram of c-Myc-GFP expression in iNKT stages 1–3 from Myc^G/G mice. (C) Quantitative RT-PCR was performed on cDNA obtained from sorted cells as indicated.

In agreement with the protein levels, Myc mRNA was up-regulated in the DP^hi, CD4⁺ SP, and iNKT ST1, -2, and -3 stages when compared to DP^lo cells (Fig. 2C). In summary, these data support a role for c-Myc during iNKT development, potentially on selection.

To determine the precise stage of iNKT cell development affected by c-Myc ablation, MACS-enriched iNKT cells from BALB/c and CD4Cre Myc^fl/fl mice were stained for HSA, CD44, and DX5 surface expression. Myc deficiency resulted in a complete loss of ST3 and a severe loss of ST1 and ST2 iNKT cells. More than 80% of tetramer⁺HSA⁻ c-Myc-deficient iNKT cells were ST1 cells compared with ≈30% in control mice (Fig. 3A). Whereas ST3 cells were absent, ST2 cellularity was reduced ≈350-fold (44,000 vs. 150 event counts when acquiring all cells enriched from 1 thymus) (Fig. 3B). ST1 cells were reduced in numbers >50-fold (44,000 vs. 700), whereas ST0 cells were present at comparable numbers (500 vs. 700), indicating that these cells are less affected by c-Myc ablation, presumably because they are just upstream of a proliferative burst. Alternatively, despite the absence of Myc mRNA in iNKT ST1 thymocytes from CD4Cre Myc^fl/fl animals (Fig. 1C), it is possible that some cells escape timely Cre-mediated deletion or retain c-Myc protein for a limited time. In conclusion, c-Myc ablation severely reduces the cellularity of all immature iNKT cells, with the exception of the earliest ST0 further supporting a role for c-Myc on selection of iNKT cells.

Fig. 3.

c-Myc deficiency severely reduces iNKT ST1, ST2, and ST3 cells. (A) (Upper) FACS plots after magnetic bead enrichment for tetramer⁺ cells. Numbers on plots indicate total event counts. (Lower) MACS-enriched tetramer⁺HSA⁻ cells are plotted for CD44 vs. DX5 surface expression to discern iNKT ST1 to -3. Numbers indicate percentages. (B) Total event counts in gate corresponding to acquisition of 1 thymus. Results are representative of 2 experiments with a total of 6 mice per group from 2 different litters.

Loss of Proliferating ST0/ST1 iNKT Cells on c-Myc Deletion.

c-Myc is an important mediator of cell proliferation and the loss of c-Myc in lymphocytes impairs their capacity to cycle (20, 21). In addition, in vivo BrdU uptake experiments indicated that early stage iNKT cells actively divide, whereas thymic stage 3 cells do not (9). To address whether c-Myc deficiency affected the capacity of immature iNKT cells to cycle, we injected the thymidine analog 5-ethynyl-2′-deoxyuridine (EdU) into CD4Cre Myc^fl/fl and BALB/c control mice. Three hours later, 2–3 control or 3–5 CD4Cre Myc^fl/fl thymi were pooled and magnetically enriched with tetramers and MACS beads to obtain large enough cell numbers for analysis. We observed a >10-fold reduction (10.4 ± 1.4-fold) in EdU⁺ cells on loss of c-Myc in the ST0/1 compartment (B220⁻CD44^lowDX5⁻) (Fig. 4). Thus, immature CD44^low iNKT cells depend on c-Myc for proliferation, and their impaired cycling capacity preventing exponential expansion might explain the lack of the subsequent developmental iNKT cell stages in CD4Cre Myc^fl/fl mice. We also observed a 2.5-fold reduction in proliferating c-Myc-deficient ST2 (CD44^highDX5⁻) cells; however, this was not statistically significant (2.5 ± 1.5-fold) probably because of the extremely low number of events in this population (e.g., 281 events from pools of 5 mice).

Fig. 4.

Reduced proliferation in c-Myc-deficient iNKT cells. Mice were injected with EdU and analyzed 3 h later. Totals of 2–3 BALB/c and 3–5 CD4Cre Myc^fl/fl thymi were pooled and MACS enriched for iNKT cells. Events were gated as indicated. The experiment was performed with 3 independent pools of mice. Numbers under EdU histograms are EdU⁺ events/total events in population.

No Excessive Cell Death on c-Myc Deletion.

Apart from its critical role in mediating proliferation, c-Myc has also been implicated in controlling survival in several systems. Thus, the loss of iNKT cells on ablation of c-Myc may also be caused by increased cell death. We performed annexin V staining of MACS-enriched thymocytes from CD4Cre Myc^fl/fl and Myc^fl/fl littermates. The fraction of annexin V⁺ cells was ≈2% in ST1 iNKT cells in both cases (Fig. 5A), indicating that c-Myc deficiency does not lead to excessive cell death. annexin V staining may, however, fail to identify apoptotic processes that nevertheless occur in vivo. In an independent approach, we therefore crossed CD4Cre Myc^fl/fl mice with the vav-bcl-2 transgenic strain. These mice overexpress human BCL-2 in all hematopoietic cells under the control of the vav promoter (26). FACS analysis of thymus and spleen from CD4Cre Myc^fl/fl vav-bcl-2 mice showed that transgenic expression of BCL-2 could not ameliorate the iNKT cell deficiency of CD4Cre Myc^fl/fl mice (Fig. 5B). In conclusion, we did not observe excessive cell death in c-Myc-deficient iNKT cells nor could the loss of iNKT cells on c-Myc deletion be recovered by enforced survival through transgenic BCL-2 expression. Thus, we did not find evidence that Myc deficiency predisposes developing iNKT cells to apoptosis.

Fig. 5.

c-Myc deficiency does not impair iNKT survival. (A) FACS analyses of tetramer-enriched thymocytes from pools of 4-week-old CD4Cre Myc^fl/fl and Myc^fl/fl littermates. Events were gated as indicated. SSC, side scatter. (B) FACS analyses of thymocytes from littermates of the indicated genotypes. Results are representative of 3 independently analyzed litters.

Expression of Lineage Determinants.

To investigate whether lack of c-Myc altered the expression of developmental regulators, we performed quantitative real-time RT-PCR using TaqMan probes (Fig. 6). Given the proliferative defect we observed in c-Myc-deficient iNKT cells, we measured mRNA levels of cyclin D2, a downstream target that can also be indirectly regulated by c-Myc. Cyclin D2 expression was 50% lower in c-Myc-deficient CD4⁺ SP thymocytes and ST0 iNKT cells compared with controls, but it was comparable or even slightly elevated in ST1 cells. The proliferative effect of c-Myc may therefore be independent of cyclin D2.

Fig. 6.

Expression of lineage determinants. cDNA from 2,000 cells per population was subjected to quantitative PCR. ST0 and ST1 cells were sorted after MACS enrichment. DP^hi and CD4 SP cells were sorted from the tetramer-depleted fraction. The experiment was repeated 3 times with comparable results.

Thus far, our observations have linked impaired iNKT cell development in c-Myc-deficient thymocytes to reduced proliferation in immature iNKT subsets, whereas there was no evidence for increased apoptosis. Apart from impacting proliferation and apoptosis, however, c-Myc deficiency could also alter the expression of lineage determinants controlling the differentiation of iNKT cells. To examine whether there was a deregulation of critical differentiation factors, we measured mRNA levels of factors previously implicated in T cell differentiation. These included the zinc-finger domain containing transcription factors Plzf and Th-POK (c-Krox) as well as Gata3. cDNA was obtained from 2,000 sorted cells from pools of 4 mice each. The experiment was repeated 3 times with similar results (Fig. 6). All experiments showed clearly detectable Plzf expression that was higher in ST1 than in ST0 iNKT cells. Therefore, c-Myc ablation does not appear to prevent up-regulation of Plzf. The expression of Th-POK, which is essential for CD4 T cell differentiation (27) and also expressed in iNKT cells (16), was unperturbed by Myc deletion. Gata3, which has been implicated in the development of virtually all lymphocyte lineages (28) was down-regulated in c-Myc-deficient DP^hi thymocytes but unchanged in CD4 SP and slightly elevated in ST0 and ST1 iNKT cells. Thus, critical differentiation factors are expressed in c-Myc-deficient iNKT cells at comparable or slightly altered levels.

Early-stage iNKT cells express higher levels of IL-4 mRNA than later stages and respond predominantly by IL-4 production when stimulated (9). Compared with control cells, c-Myc-deficient iNKT ST0 and ST1 cells expressed lower amounts of IL-4 mRNA, indicating that these cells may not be differentiated sufficiently to up-regulate IL-4 to wild-type levels or that c-Myc is required for proper control of IL-4 expression. Unfortunately, the extremely low number of thymic CD4Cre Myc^fl/fl iNKT cells precludes further analysis of potential defects in cytokine production on stimulation as well as broader screening approaches for differentially regulated genes.

At the current state of the art, our data do not support a specific role for c-Myc in controlling iNKT signature. However, we cannot formally rule out that c-Myc ablation affects differentiation as well as proliferation. In addition, effects on differentiation may be difficult to dissect because they might be obscured by the general lack of mature iNKT cells because of the lack of proliferative expansion.

Discussion

Here, we provide evidence that Myc ablation in DP thymocytes specifically impacts iNKT cell development but not the development of conventional T cells. In contrast to developing conventional T cells, intrathymic iNKT cell differentiation depends on agonist selection as well as costimulation (29). iNKT cells are not killed by negative selection on encounter of agonist but rather the receptor signal they receive places them in a “standby” mode, priming them for immediate response to further receptor challenge (1, 7). Selection and development of iNKT cells in the thymus have been proposed to follow a program that resembles T cell activation rather than the development of conventional T cells (14). This includes a mechanistically elusive proliferation wave after selection to account for the substantial expansion observed between ST0 and ST1 of thymic iNKT cell development (7). Our finding that c-Myc ablation reduces proliferation of ST0/ST1 cells shows that this transcription factor is a crucial mediator of proliferation in immature iNKT cells. The observation that this leads to a complete lack of mature iNKT cells highlights the significance of proliferation in iNKT cell development. c-Myc has been implicated in proliferative processes along multiple stages of hematopoietic development. Among these processes are the self-renewal of HSCs (24), the activation of peripheral CD4 T cells after TCR stimulation (20), as well as the proliferative burst after pre-TCR assembly and signaling.

A mechanism for c-Myc induction in response to TCR stimulation has been proposed by analyzing c-Rel and RelA double-deficient T cells. It was shown that growth and proliferation of activated T cells requires induction of c-Myc by Rel/NF-κB activity. This process depends on protein kinase Cθ (PKCθ)-controlled nuclear translocation of c-Rel and AP-1/NFAT-induced transcription of RelA (30). Based on the resemblance between iNKT cell development and activation of conventional T cells, it is tempting to speculate that this mechanism is also involved in iNKT development. Both ablation of PKCθ (31) and inhibition of AP-1 signaling (32, 33) resulted in loss of iNKT cells; however, the observed phenotypes in these studies are less severe and affect later developmental stages compared with c-Myc ablation. This in turn would indicate that c-Myc is required before NF-κB and AP-1 signaling at an earlier stage in iNKT cell development.

Although c-Myc clearly controls proliferation during iNKT cell development, we did not observe excessive cell death in Myc-deficient iNKT cells nor did transgenic expression of BCL-2 rescue the phenotype of CD4Cre Myc^fl/fl mice, arguing against a role for c-Myc in apoptosis of developing iNKT cells. Likewise, we did not detect severe changes in the expression of transcription factors previously implicated in iNKT cell development. Both Plzf, a critical factor for iNKT cell function, and Th-POK, a master regulator of CD4 T cell development, are detectable in c-Myc-deficient ST0 and ST1 cells, whereas Gata3 appears slightly elevated. Even small changes in the levels of transcription factors may deregulate networks in control of developmental processes (34). It remains to be determined whether the small transcriptional alterations we observed reflect a significant disturbance in the transcriptional network controlling iNKT cell development.

In conclusion, data presented here provide mechanistic depth for our understanding of the early developmental stages of the enigmatic iNKT cell lineage. Our data provide compelling evidence that c-Myc mediates an intrathymic proliferation wave immediately after agonist selection of iNKT cells and illustrate the importance of this expansion for the generation of mature iNKT cells in vivo.

Materials and Methods

Mice.

All mice were kept in the animal facilities of the University of Chicago according to protocol no. 71880 approved by the Institutional Animal Care and Use Committee. Myc^fl/fl mice (35) were crossed with CD4Cre transgenic mice (36). CD4Cre Myc^fl/wt mice were backcrossed to the BALB/c background for 6 generations and then crossed inter se to obtain CD4Cre Myc^fl/fl animals. BALB/c control mice were purchased from Charles River. The CD4Cre transgene was detected by PCR using 5′-ATCGCTCGACCAGTTTAGT-3′ (forward) and 5′-CGATGCAACGAGTGATGA-3′ (reverse), and the floxed Myc allele was detected with 5′-TAAGAAGTTGCTATTTTGGC-3′ (forward) and 5′-TTTTCTTTCCGATTGCTGAC-3′ (reverse) primers. vav-bcl2-Tg mice were a gift from A. T. Look (Dana–Farber Cancer Institute, Boston, MA) and were originally made by J. Adams and colleagues (26). Myc^G/G mice were provided by B. P. Sleckman (Washington University, St. Louis, MO).

Flow Cytometry.

Multicolor-FACS stainings were performed for analysis and cell sorting of primary thymocytes. A total of 1–5 × 10⁶ cells were stained in a total volume of 50–200 μL of FACS buffer (HBSS/2% FBS/50 μg/mL DNase I). Antibodies were from BD Pharmingen or eBioscience: B220-PacificBlue(RA3-6B2), CD4-FITC/-peridinin chlorophyll protein (PerCp)-Cy5.5/-allophycocyanin (APC)-Alexa750/-phycoerythrin (PE)-Cy7(RM4-5), -PE/-APC(GK1.5), CD8-FITC/-PerCp-Cy5.5/-APC-Alexa750/-PE/-APC/-PE-Cy7/-PacificBlue(53-6.7), TCRβ-PE/-APC/-APC-Alexa750(H57-597), TCRγδ-PE/-FITC(eBioGL3), pan-NK-FITC(DX5), CD44-PerCp-Cy5.5(IM7), CD69-FITC/-PE(H1.2F3), CD24-PE(M1/69). For analysis of iNKT cells, thymocyte single-cell suspensions from 1–10 mice were stained for 1 h on ice in 500 μL of HBSS/2% FCS with APC-labeled CD1d tetramers loaded with the α-GalCer analog PBS57 (37) obtained from the tetramer core facility of the National Institutes of Health. Tetramer-positive cells were then enriched with anti-APC magnetic microbeads by using an autoMACS cell separator (Miltenyi Biotec) as described in ref. 8 and subjected to further staining for flow cytometry. To detect apoptotic cells, the annexin V-FITC labeling kit (BD Pharmingen) was used according to the manufacturer's instructions. All samples were analyzed on an LSRII or sorted on a FACSAria instrument (BD Biosciences). Data were analyzed by using FlowJo software (Tree Star).

Irradiation Bone Marrow Chimeras.

Sublethally irradiated (650 rad; Gammacell 40) Thy1.1⁺ BALB/c mice (host) were injected with a 1:1 mixture of FACS-sorted host and Thy1.2⁺ CD4Cre Myc^fl/fl donor lineage (B220, CD19, TCRβ, CD8, Gr-1, Mac-1, DX5) negative bone marrow (2 × 10⁵ cells per mouse). Animals were treated with Bactrim in the drinking water for the entire time of observation (12 weeks).

Cell Cycle Analysis.

Mice were injected intravenously with 1 mg of EdU (Invitrogen) 3 h before analysis. Thymocytes were processed and stained for flow cytometry according to the manufacturer's instructions.

Quantitative Real-Time PCR.

A total of 2,000 cells were sorted directly into lysis buffer, and RNA was extracted by using the RNeasy Micro kit (Qiagen). cDNA was prepared with the SuperScript-III RT kit (Invitrogen). Quantitative PCR was performed on an ABI7300 machine (Applied Biosystems). All targets were determined relative to β-actin expression by using TaqMan Gene Expression Assays from Applied Biosystems. Data were analyzed and evaluated according to the relative ΔΔC_T method.