The Tec kinases Itk and Rlk are required for efficient positive selection of conventional CD4+ and CD8+ T cells in the thymus. In contrast, recent studies have shown that these Tec kinases are dispensable for the development of CD8+ T cells with characteristics of innate T cells. These findings raise questions about the potential role of Itk and Rlk in NKT cell development, because NKT cells represent a subset of innate T cells. To address this issue, we examined invariant NKT cells in Itk−/− and Itk/Rlk−/− mice. We find, as has been reported previously, that Itk−/− mice have reduced numbers of NKT cells with a predominantly immature phenotype. We further show that this defect is greatly exacerbated in the absence of both Itk and Rlk, leading to a 7-fold reduction in invariant NKT cell numbers in the thymus of Itk/Rlk−/− mice and a more severe block in NKT cell maturation. Splenic Itk−/− and Itk/Rlk−/− NKT cells are also functionally defective, because they produce little to no cytokine following in vivo activation. Tec kinase-deficient NKT cells also show enhanced cell death in the spleen. These defects correlate with greatly diminished expression of CD122, the IL-2R/IL-15R β-chain, and impaired expression of the T-box transcription factor, T-bet. These data indicate that the Tec kinases Itk and Rlk provide important signals for terminal maturation, efficient cytokine production, and peripheral survival of NKT cells.
Natural killer T cells are an innate subset of αβTCR+ T cells that express surface markers characteristic of both T and NK cells, and can rapidly produce large amounts of cytokines such as IFN-γ, IL-4, and IL-10 (1, 2). The great majority of murine NKT cells express an invariant TCR Vα and Jα combination (Vα14-Jα18) that is preferentially paired with a restricted TCR β-chain (mostly Vβ8.2 or Vβ7), and recognize lipid Ags in the context of CD1d (3, 4). These cells develop in the thymus from CD4+8+ (double-positive (DP))3 thymocyte progenitors, and are positively selected by interactions with other CD1d+/signaling lymphocytic activating molecule (SLAM)+ DP cortical thymocytes (5, 6). Upon positive selection, NKT cells acquire markers characteristic of memory T cells, such as CD44 (7).
Previous studies have identified three differentiation stages of NKT cells in the thymus subsequent to their expression of a mature TCR that binds to CD1d tetramer, as follows: stage 1, in which they have low expression of CD44 and NK1.1; stage 2, in which they acquire high expression of CD44; and stage 3, in which they acquire expression of NK1.1 as well as several other NK markers (7). Accompanying these changes in surface markers, NKT cells undergo a proliferative expansion between stages 1 and 2, during which they already possess the ability to produce IL-4. At stage 3, the cells are also competent to produce IFN-γ (8, 9, 10). Interestingly, the NK1.1− NKT cells leave the thymus to populate the periphery, where they can further mature into NK1.1+ NKT cells, whereas the more mature NK1.1+NKT cells generated in the thymus remain in the thymus to become a long-lived nondividing resident population (8, 9, 11). Furthermore, up-regulation of NK1.1 on peripheral NKT cells, which denotes their terminal differentiation, is CD1d dependent (12), whereas the survival and homeostasis of these cells depend on IL-15 (13, 14).
Several cytokines, transcription factors, and signaling molecules have been implicated in the development and survival of NKT cells. As mentioned above, IL-15 is involved in NKT survival and proliferation in the periphery, whereas Csf-2 (GM-CSF) is required for NKT cell differentiation and cytokine secretion (15). Members of the NF-κB transcription factor family have also been implicated in proliferation and survival of NKT cells (16). During development in the thymus, the SLAM family of receptors, the adaptor protein SLAM-associated protein (SAP), and the Src family tyrosine kinase Fyn have all been implicated in a signaling axis required for selection of NKT cells (5, 17, 18, 19, 20). In addition, protein kinase Cθ (PKCθ), Vav-1, IFN-regulatory factor 1, γ-ETS erythroblastosis virus E 26 oncogene homolog 1, E74-like factor 4 or myeloid ELF-1-like factor, runt-related transcription factor 1 (Runx1), and retinoic acid receptor-related orphan receptor-γt (RORγt) have been shown to have significant roles in NKT cell development, function, and survival (21, 22, 23, 24, 25, 26, 27).
Of interest, Th1 and Th2 transcriptional regulators, T-bet and GATA-3, respectively, have also been shown to have prominent roles in NKT development and homeostasis. For instance, in T-bet-deficient mice, the development of NKT cells in the thymus is blocked before up-regulation of NK1.1 (i.e., before stage 3), leading to decreased numbers of NKT cells in the thymus and periphery (28). These mice also have defects in NKT cell migration, survival, and effector functions, findings that are consistent with the prominent role of T-bet in CD122 expression and IFN-γ transcription in T cells (29). Mice deficient in GATA-3 exhibit decreases in peripheral NKT cell numbers due to increased cell death, as well as alterations in NKT cell surface marker expression both in the thymus and the periphery; GATA-3-deficient NKT cells also show decreased cytokine secretion (30). It has recently been shown that a member of the NF-AT family, NF-AT2, has a more prominent role in IL-4 cytokine secretion by NKT cells than GATA-3 (31). Interestingly, Tec family kinases have been shown to modulate NF-AT activation (32).
The Tec family of nonreceptor tyrosine kinases has been shown to play a significant role in signaling downstream of the TCR (33, 34). Three members of this family are expressed in conventional αβTCR+ T cells, as follows: Itk, Rlk, and Tec. T cells from animals deficient in Itk have defects in TCR-induced phospholipase C-γ phosphorylation, calcium flux generation, MAPK activation, and NF-AT and AP-1 activation. These defects are exacerbated in T cells deficient for both Itk and Rlk, indicating possible redundancy among the Tec family kinases. Itk−/− and Itk/Rlk−/− mice also have impaired development of conventional CD4+ and CD8+ T cells in the thymus (35, 36).
Although Itk and Rlk are dispensable for the development of CD8+ innate T cells (37, 38, 39), the role of Tec family kinases has not been thoroughly studied in other subsets of innate T cells, such as NKT cells. One previous study demonstrated a role for Itk in NKT cell development, indicating impaired progression of Itk−/− NKT cells to stage 3 and a resulting decrease in NKT cell numbers (40). Recently, another group has shown impaired function of Itk−/− NKT cells (41). In this study, we confirm the developmental defect in NKT cells in Itk−/− mice, and show that this defect is exacerbated in the absence of both Itk and Rlk. We also confirm the functional defect in Itk−/− NKT cells and demonstrate that NKT cell function is further impaired in Itk/Rlk−/− NKT cells, which produce virtually no effector cytokines in response to activation either in vitro or in vivo. Interestingly, expression of the T-box transcription factor T-bet is reduced in NKT cells from Itk−/− mice, and even further reduced in Itk/Rlk−/− NKT cells. CD122 expression is also impaired on Tec kinase-deficient NKT cells, leading to reduced NKT cell survival in the periphery. These data implicate a critical role for Tec family kinases in NKT cell development, function, and survival.
Materials and Methods
Itk−/− mice (42) have been backcrossed >13 generations to the C57BL/10 strain. Itk/Rlk−/− mice (43) were a gift from P. Schwartzberg (National Human Genome Research Institute, National Institutes of Health, Bethesda, MD) and were backcrossed >10 generations to C57BL/6 mice. Rlk−/− mice were derived from Itk/Rlk−/− mice. C57BL/10 mice were used as controls. All mice were used between 6 and 12 wk of age and were maintained at the University of Massachusetts Medical School specific pathogen-free animal facility after review and approval by the institutional animal care and use committee.
Abs, CD1d tetramer, and flow cytometric analysis
Anti-heat stable Ag (HSA) biotin, anti-HSA FITC, anti-CD1d PE, anti-SLAM FITC, anti-CD4 allophycocyanin, anti-CD4 PE, anti-CD8 allophycocyanin, anti-CD44 Cy, anti-NK1.1 PE, anti-NK1.1 allophycocyanin, anti-CD69 FITC, anti-NKG2D allophycocyanin, anti-Ly49C/I FITC, anti-Ly49G2 FITC, anti-Ly49-F biotin, anti-TCRβ Cy, anti-IL-4 PE, anti-IFN-γ FITC, annexin V FITC, 7-aminoactinomycin D (7AAD), and anti-CD122 PE were all purchased from BD Pharmingen. CFSE was purchased from Molecular Probes. CD1d-PBS44-PE tetramer used for preliminary experiments was provided by A. Bendelac (University of Chicago, Chicago, IL). CD1d-PBS57-PE and CD1d-PBS57-allophycocyanin tetramers were provided by the National Institute of Allergy and Infectious Diseases Tetramer Facility. For flow cytometric analysis, cells were analyzed on a FACSCalibur cytometer (BD Biosciences), and data were analyzed using FlowJo software (Tree Star). Thymocyte and splenocyte subsets were sorted on either a MoFlo (DakoCytomation) or a FACSVantage (BD Biosciences) cell sorter.
Quantitative real-time PCR
Cells were prepared from thymi and spleens isolated from wild-type, Itk−/−, or Itk/Rlk−/− mice and lysed for RBCs. Thymic NKT cells were first enriched by depleting with anti-CD8α magnetic microbeads (Miltenyi Biotec) on an AutoMACS machine. Splenic NKT cells were first enriched by labeling with anti-B220 biotin and anti-CD8α biotin (BD Pharmingen) and depleting with anti-biotin magnetic microbeads (Miltenyi Biotec) on an AutoMACS machine. Remaining cells were then sorted to >95% purity on either a Moflo (DakoCytomation) or a FACSVantage (BD Biosciences) cell sorter by gating on HSAlow CD1d/α-galactosylceramide (αGAL) tetramer+ cells. For controls, CD8+CD4− conventional T cells were sorted on Moflo cell sorter (DakoCytomation). RNA and cDNA were prepared, as described previously (44). Real-time quantitative PCR was performed on an i-Cycler (Bio-Rad). Primer sequences and reaction conditions are available upon request.
In vivo stimulation assays
NKT cells were activated in vivo for cytokine analysis by ELISA, as previously described (45). Briefly, wild type, Itk−/−, and Itk/Rlk−/− mice were injected i.v. with 1 μg of anti-CD3 Ab (BD Pharmingen). Spleens were harvested after 90 min, and a suspension of 107 cells was incubated at 37°C for 2 h, at which time supernatants were collected and IL-4 and IFN-γ were measured by ELISA (BD Biosciences). For intracellular cytokine staining, mice were injected with 2 μg of αGAL (Alexis Biochemicals; distributed by Axxora) i.v. in 0.025% PBS Tween 20, and spleens were harvested after 2 and 3 h. Cells were then surface stained for either anti-NK1.1 and anti-TCRβ or anti-HSA and CD1d tetramer, fixed, and permeabilized using the Cytofix/Cytoperm Kit (BD Pharmingen), and stained for anti-IFN-γ and anti-IL-4. Splenocytes were also activated with PMA (10 ng/ml (Sigma-Aldrich)) and ionomycin (2000 ng/ml; Calbiochem) for 1 h and incubated with Golgi-stop (BD Pharmingen) for 3 h, after which they were stained for cytokines, as described above.
In vitro survival and proliferation assay
Cell preparations were made from spleens of wild-type, Itk−/−, and Itk/Rlk−/− mice. Cells were then labeled with CFSE, and then 5 × 105 cells/well were plated in quadruplicate for each genotype and condition in a 96-well round-bottom plate. Cells were then incubated at 37°C in RPMI 1640/10% FBS with the following: no stimulus, IL-2 (20 ng/ml), IL-2 (2 μg/ml), IL-7 (10 ng/ml; R&D Systems), αGAL (500 ng/ml), IL-2 (20 ng/ml) plus αGAL (500 ng/ml), and IL-2 (2 μg/ml) plus αGAL (500 ng/ml). At day 2 postincubation, medium was replaced with RPMI 1640/10% FBS containing no additions (for the no stimulus and αGAL groups), or the appropriate cytokine alone (without αGAL). Cells were then harvested and counted at days 4 and 6, quadruplicate samples were pooled, and cells were stained for anti-HSA and CD1d tetramer and analyzed by flow cytometry.
Two-tailed nonparametric Mann-Whitney tests were performed using In-Stat software (GraphPad).
Itk, Rlk, and Tec are expressed in NKT cells
Three Tec family kinases are expressed in conventional αβTCR+ T cells, Itk, Rlk, and Tec. The hierarchy of expression of these kinases correlates with their known importance in T cell signaling, with Itk mRNA being expressed at the highest levels, followed by Rlk, and then Tec (Fig. 1,A) (34). To assess relative Tec kinase expression levels in NKT cells, thymic NKT cells were purified from wild-type mice based on CD1d/αGAL tetramer (hereafter referred to as CD1d tetramer) binding, and RNA was isolated and used for quantitative real-time RT-PCR to measure the levels of Itk, Rlk, and Tec mRNA. As shown in Fig. 1 A, NKT cells show a similar hierarchy of expression of Tec family kinases compared with conventional αβ T cells, with Itk mRNA being expressed at the highest levels (similar to levels found in peripheral T cells), followed by Rlk and then Tec.
Tec family kinases are known to have altered patterns of expression during differentiation of CD4+ T cells. For instance, Rlk is down-regulated following TCR stimulation, and is re-expressed in Th1 cells, but not Th2 cells. In contrast, Itk is up-regulated in Th2 cells relative to Th1 cells (35). Terminal differentiation of NKT cells is characterized by the up-regulation of NK1.1 (7). To determine whether Tec family kinases are differentially expressed following terminal differentiation of NKT cells, CD1d tetramer-binding T cells were isolated from the thymus and spleen, and separated into NK1.1− and NK1.1+ populations. Itk (Fig. 1,B), Rlk (Fig. 1,C), and Tec (Fig. 1,D) mRNA were all up-regulated in the NK1.1+ compared with the NK1.1− population in the thymus. Surprisingly, both subsets of peripheral NKT cells (NK1.1+ and NK1.1−) express similar levels of each Tec kinase mRNA, comparable to the levels seen in thymic NK1.1+ cells (Fig. 1, E–G). Because the only NKT cells known to emigrate from the thymus are the NK1.1− cells (8, 9, 11), this latter finding suggests that thymic emigration, in addition to terminal maturation, may induce Tec kinase up-regulation.
Altered distribution of NKT cell subsets in Tec family kinase-deficient mice
In a previous study using Itk−/− mice, a deficiency in NKT cell numbers was seen as early as 6–8 wk in the spleen and after 20 wk of age in both the thymus and the spleen (40). To further address the role of Tec kinases in NKT cells, we examined whether a similar deficiency occurs in the absence of Rlk, and whether the decrease in NKT cell numbers observed in the absence of Itk is exacerbated in Itk/Rlk−/− mice. We found that in our Itk−/− mice, the decrease in NKT cell numbers could be seen both in the thymus and the spleen as early as 6 wk of age (Fig. 2, A–F, and data not shown). Although this is a more profound difference than that observed in the previous study (40), it is important to note that our Itk−/− mice were analyzed after 13 generations of backcrossing to C57BL/10, whereas the Itk−/− mice used previously were backcrossed for 5 generations. We then analyzed Rlk−/− mice, and found a slight decrease in the percentage of NKT cells compared with wild-type mice in both the thymus and periphery (Fig. 2, A–D); however, this decrease in percentage did not translate into a decrease in NKT cell numbers in the absence of Rlk (Fig. 2, E and F). In contrast, Itk/Rlk double-deficient mice had reduced percentages and numbers of NKT cells in the thymus when compared with wild-type (7-fold reduced in number), as well as with Itk−/− mice (3-fold reduced in number), indicating potential redundancy between Itk and Rlk in NKT cells (Fig. 2, A, C, and E). Itk/Rlk−/− mice also had reduced percentages and numbers of splenic NKT cells compared with wild type (Fig. 2, B, D, and F). Despite a slight, but consistent decrease in NKT cell percentages in the spleen between the Itk−/− and Itk/Rlk−/− mice, there was no significant difference in absolute numbers in this comparison (Fig. 2, B, D, and F), perhaps due to the increased overall splenic and thymic cellularity seen in Itk/Rlk−/− when compared with the Itk−/− mice.
Because NKT cell progenitors in the thymus cannot be identified before the up-regulation of their invariant TCR, it is difficult to investigate the positive selection of this subset directly. Therefore, we examined whether or not molecules known to be involved in the positive selection of NKT cells are expressed at normal levels in the Itk−/− and Itk/Rlk−/− mice. To this end, we stained DP cortical thymocytes, the cells that provide positive selection interactions for NKT cells, with Abs to CD1d and SLAM (CD150) (5, 6). Both CD1d (Fig. 2,G) and SLAM (Fig. 2 H) are expressed at comparable levels on DP thymocytes from wild-type, Itk−/−, and Itk/Rlk−/− mice, indicating that altered expression of these molecules is not responsible for the decreased number of NKT cells developing in the absence of Itk, or Itk and Rlk.
Although it is not certain that CD150 is the SLAM family member involved in NKT cell selection in the thymus, it is to date the only one reported to be expressed on the surface of cortical thymocytes (46). Nonetheless, there is indirect evidence indicating that SLAM (CD150) might be dispensable for NKT cell development, because slam−/− mice can produce IL-4 quickly post-anti-CD3 activation, a response typically attributed to NKT cells (47). Alternatively, genetic evidence obtained in NOD mice, which are deficient in NKT cells, indicates that both SLAM and NK-T-B Ag may provide instructive signals for development of NKT cells (48). Other studies have shown that Ly9 (CD229), another SLAM family member, is not required for NKT cell development (49). Thus, whereas our data indicate that alterations in SLAM (CD150) expression in the absence of Itk, or Itk and Rlk, are not responsible for the observed deficiency in NKT cell selection (Fig. 2 H), we cannot exclude the possibility that other SLAM family members are involved in Itk−/− and Itk/Rlk−/− mice.
Tec family kinase-deficient NKT cells exhibit an immature phenotype
Cell surface markers distinguishing the stages of NKT cell development are well characterized (7, 40). Immature NKT cells, before the up-regulation of NK1.1 (fractions 1 and 2), have high expression of the CD4 coreceptor. Following terminal maturation, CD4 is down-regulated, and NKT cells are either CD4+CD8− or CD4−CD8−. CD44 up-regulation occurs in fraction 2, before the up-regulation of NK1.1. The final stage of maturation in NKT cells (stage 3) is accompanied by the up-regulation of NK1.1, CD69, and several additional NK cell receptors. CD4+ or CD4−, CD44high, NK1.1−NKT cells are exported from the thymus to the periphery, where they become NK1.1+ following additional stimulation requiring CD1d (7).
Thymocytes and splenocytes were isolated from wild-type, Itk−/−, and Itk/Rlk−/− mice; gated on HSAlow CD1d tetramer-positive cells; and analyzed for several developmental markers. When compared with wild-type thymic NKT cells, CD4 expression on Itk−/− and Itk/Rlk−/− NKT cells remained high, similar to that seen on immature wild-type NKT cells, and few to no CD4−CD8− NKT cells were present in the thymus of Itk−/− and Itk/Rlk−/− mice (Fig. 3,A). CD44 expression on all of the thymic NKT cells was comparable; however, expression of terminal maturation markers NK1.1 and CD69 was reduced on Itk−/− thymic NKT cells, and even more severely impaired on Itk/Rlk−/− NKT cells (Fig. 3,A). This pattern of differences between wild-type, Itk−/−, and Itk/Rlk−/− NKT cells was also observed for the surface markers NKG2D, Ly49C/I, Ly49G2, and Ly49F expression (Fig. 3,A). Recently, an earlier stage of NKT cell development has been described in which cells are either HSAhighDPlow or HSAhighCD4+/high (50). Because this stage precedes early developmental expansion, we considered the possibility that the substantial decrease in NKT cell numbers seen in Itk−/− and Itk/Rlk−/− mice was due to a block at this stage. If this were the case, we would expect to find increased numbers of HSAhigh NKT cells in the thymus of Itk−/− and Itk/Rlk−/− mice compared with wild-type mice. To assess this issue, we identified NKT cells based on CD1d tetramer staining, and compared numbers of HSAhigh and HSAlow thymocytes in wild-type, Itk−/−, and Itk/Rlk−/− mice. As shown in Fig. 3,C, we did not find increased numbers of HSAhigh thymic NKT cells in the Itk−/− or Itk/Rlk−/− mice relative to wild type, arguing against a block at this stage. Furthermore, similar decreases in the Itk−/− and Itk/Rlk−/− HSAlow and HSAhigh NKT cell populations are seen when compared with wild type, suggesting that the decrease in overall NKT cell numbers is most likely due to a block preceding the HSAhigh stage. Although the data shown in Fig. 3 C derive from mice analyzed at 6–8 wk of age, Itk−/− and Itk/Rlk−/− mice analyzed at 3 wk of age also show no apparent block at the HSAhigh stage of NKT cell development (data not shown).
Splenic NKT cells in Itk−/− and Itk/Rlk−/− mice also showed a less mature phenotype than that of wild-type NKT cells. CD4 expression on splenic NKT cells from Itk−/− and Itk/Rlk−/− remained high, indicating reduced maturation; however, all mice showed a comparable proportion of CD4−CD8− splenic NKT cells (Fig. 3,B). Of interest, CD44 expression, which did not differ between the three genotypes of thymic NKT cells, is abnormal on the splenic NKT cells, with the Itk−/− and Itk/Rlk−/− mice showing a substantial proportion of CD44low NKT cells. Accompanying this reduced proportion of CD44high NKT cells, Itk−/− and Itk/Rlk−/− mice cells also have significant numbers of splenic NKT cells with low levels of NK1.1, CD69, and NKG2D (Fig. 3,B). In general, the Itk/Rlk−/− mice exhibit a more severe defect than the Itk−/− mice. Finally, no differences were seen for the other Ly49 markers, most likely due to lower expression of these on NKT cells in the spleen (Fig. 3 B) (51). Overall, these data show that Itk−/− NKT cells fail to progress efficiently to the most mature stage of NKT cell development, and that this defect is exacerbated in the absence of Itk and Rlk, further supporting the notion of redundancy between Itk and Rlk in the maturation of NKT cells.
Tec family kinase-deficient NKT cells are functionally impaired
NKT cells are well known for their ability to rapidly produce effector cytokines following initial stimulation (1). Furthermore, the pattern of cytokines produced by NKT cells alters as the cells mature. Before NK1.1 up-regulation, NKT cells can produce IL-4, but little or no IFN-γ. Following up-regulation of NK1.1, NKT cell cytokine production shifts to favor production of IFN-γ over IL-4 (40). Given the less mature phenotype of Itk−/− and Itk/Rlk−/− NKT cells, we reasoned that these cells might produce IL-4, but little IFN-γ. To examine this issue, we used a previously characterized assay (45). Briefly, wild-type, Itk−/−, or Itk/Rlk−/− mice were injected i.v. with anti-CD3 Ab, and 90 min later splenocytes were harvested and cultured for 2 h, and supernatants were examined for cytokines by ELISA. Previous studies have shown that, in this short-term assay, the only cells capable of producing IFN-γ and IL-4 are NKT cells (45). As shown in Fig. 4, IFN-γ (Fig. 4,A) and IL-4 (Fig. 4,B) are secreted by wild-type NKT cells. Secretion of IFN-γ by Itk−/− splenocytes was substantially reduced, and was undetectable by Itk/Rlk−/− splenocytes (Fig. 4,A). Similarly, IL-4 secretion by Itk−/− splenocytes was decreased, whereas no IL-4 was detected in the cultures of Itk/Rlk−/− splenocytes (Fig. 4 B). It should be noted that decreased numbers of splenic NKT cells in the Itk−/− and Itk/Rlk−/− mice may account for a portion of the decreased cytokine production observed; however, the magnitude of the reduction in IFN-γ and IL-4 observed far outweighs the 50–60% decrease in proportions of NKT cells seen in the spleens of these mice.
To confirm these results with a second assay system that is more targeted to activation of NKT cells, we injected wild-type, Itk−/−, or Itk/Rlk−/− mice i.v. with αGAL or PBS as a control, and examined cytokine production by intracellular cytokine staining 2 and 3 h following injection (52). Splenocytes from the injected mice were stained with CD1d tetramer and anti-HSA Abs, or with anti-NK1.1 and anti-TCRβ Abs, permeabilized, and stained with Abs to IL-4 and IFN-γ. Using this staining scheme, NKT cell (CD1d tetramer+, HSAlow) and NK cell (TCRβnegative, NK1.1+) populations can each be identified, allowing us to assess NKT cell activation as well as priming of NK cells by the IFN-γ produced by the NKT cells (Fig. 4,C). After 2 h of αGAL stimulation in vivo, a significant proportion of the wild-type NKT cells is producing both IL-4 and IFN-γ (16%) or IFN-γ alone (17%). At 3 h, the wild-type NKT cells are predominantly producing IFN-γ (24%), rather than both IFN-γ plus IL-4 (10%). Cytokine-producing NKT cells are decreased in Itk−/− mice at both 2 h (IL-4/IFN-γ = 1% and IFN-γ = 2%) and 3 h (IL-4/IFN-γ = 2% and IFN-γ = 2%) post-αGAL injection. In the Itk/Rlk−/− mice, the defect in cytokine production by NKT cells is further exacerbated. It is worth noting that we did observe small decreases in TCR expression on the activated NKT cells, particularly in the wild-type mice (data not shown); however, this alteration did not impair our ability to identify the activated NKT cells for purposes of assessing cytokine production. Finally, as expected due to the lack of IFN-γ production by Itk−/− and Itk/Rlk−/− NKT cells, priming of the NK cells in these mice was also impaired, with 48% of the wild-type NK cells producing IFN-γ, whereas only 2% of Itk−/− and Itk/Rlk−/− NK cells produced IFN-γ at the 3-h time point (Fig. 4 C). In addition to the reduced IFN-γ production by Itk−/− and Itk/Rlk−/− NKT cells, it is likely that the reduced numbers of NKT cells in these mice are contributing to the impaired priming of NK cells.
To further examine the functional capabilities of Itk−/− and Itk/Rlk−/− NKT cells, thymic NKT cells were activated in vitro with anti-CD3 Ab, and cytokine production was assessed. These studies produced similar results to the ex vivo analysis described above, indicating impaired cytokine production by Itk−/− compared with wild-type NKT cells, and a more severe defect seen in Itk/Rlk−/− vs Itk−/− cells (data not shown). Bypassing any proximal defects in TCR signaling by stimulating cells with PMA and ionomycin indicated that slightly reduced proportions of Tec kinase-deficient NKT cells were capable of producing effector cytokines when compared with wild-type NKT cells (Fig. 4,D). This reduction is substantially less in magnitude than that observed when triggering Itk−/− and Itk/Rlk−/− NKT cells through the TCR using anti-CD3 Ab or αGAL, indicating that the majority of the defect in cytokine production is due to a defect in proximal TCR signaling. This experiment also confirmed that the less mature phenotype of Itk−/− and Itk/Rlk−/− NKT cells corresponds to a bias toward IL-4 production vs the predominant IFN-γ production observed from wild-type NKT cells. As can be seen in Fig. 4 D, only 58% of the cytokine-producing wild-type NKT cells made IL-4, whereas 68 and 82% of the cytokine-producing NKT cells in the Itk−/− and Itk/Rlk−/−, respectively, made IL-4; these data also indicate a corresponding decrease in IFN-γ production by Itk−/− and Itk/Rlk−/− NKT cells. These results correlate well to a recent report by Au-Yeung and Fowell (41), in which Itk−/− NKT cells were activated with αGAL and then stimulated in vitro with ionomycin, and also showed a shift in cytokine production toward IL-4 and away from IFN-γ.
Itk−/− and Itk/Rlk−/− NKT cells stimulated in vitro proliferate, but fail to accumulate
In an effort to determine the underlying cause of the decreased numbers of NKT cells in Itk−/− and the Itk/Rlk−/− mice, we examined the proliferative responses and survival of NKT cells following in vitro stimulation. For these experiments, we adapted a previously described NKT cell in vitro proliferation assay (53). Briefly, splenocytes were harvested from wild-type, Itk−/−, and Itk/RlK−/− mice; labeled with CFSE; and incubated in vitro for 4 or 6 days with a variety of stimuli. Cultures were initiated with 2 × 106 cells (pooled from three mice) of each genotype, and cell numbers were assessed at each time point. The number of NKT cells for each sample was then calculated based on surface staining for CD1d tetramer and HSA (HSAlowCD1d tetramer+). Although the magnitude of the responses varied from experiment to experiment, the relative responses of wild-type vs Itk−/− and Itk/Rlk−/− NKT cells showed a consistent pattern between experiments. In all experiments, we saw reductions in the percentages, as well as absolute numbers of NKT cells detected at day 4, for all three genotypes analyzed. This result reflects both a decrease in overall cell numbers at this time point, coupled with a down-regulation of the TCR on the NKT cells, particularly those cultured with αGAL.
The majority of NKT cells that received either no stimulus (NS), or were cultured with a low dose of IL-2 (20 ng/ml) alone, succumb to cell death (Fig. 5,A). Increasing the concentration of IL-2 to 2 μg/ml promotes the survival of wild-type NKT cells accompanied by proliferation, leading to a modest expansion of this population by day 6; in contrast, neither Itk−/− nor Itk/Rlk−/− NKT cell populations expand under these conditions, despite the fact that some proliferative response is observed (Fig. 5). Stimulation with IL-7 promotes the survival of all wild-type and Itk−/− NKT cell populations above that seen in the absence of any stimulus (NS), but is less effective at maintaining Itk/Rlk−/− NKT cells in culture. These data indicate that Itk−/− and Itk/Rlk−/− NKT cells are more responsive to IL-7 than they are to IL-2 for in vitro survival.
In vitro stimulation with αGAL, either alone or together with IL-2, promotes a dramatic expansion of the wild-type NKT cells by day 6 of culture (∼20-fold; Fig. 5,A). For Itk−/− NKT cells, αGAL alone promotes a modest expansion (∼2.5-fold), which is slightly enhanced when combined with a low dose of IL-2 (∼4-fold expansion). Itk/Rlk−/− NKT cells did not expand at all following in vitro αGAL stimulation. This lack of expansion of Itk−/− and Itk/Rlk−/− NKT cells cannot be attributed to a predominant defect in proliferation, because all three genotypes of NKT cells show comparable dilution of CFSE by day 6 (Fig. 5 B). Instead, our findings indicate that Itk−/− and Itk/Rlk−/− NKT cells exhibit a profound deficiency in survival following in vitro stimulation.
Tec family kinase-deficient NKT cells show impaired survival and reduced CD122 and T-bet expression
The data shown above suggest that Tec kinase-deficient NKT cells have a survival defect. To test this notion more directly, we examined freshly isolated ex vivo NKT cells for markers of apoptosis/death. Thymocytes and splenocytes from wild-type, Itk−/−, and Itk/Rlk−/− NKT cells were analyzed for coexpression of annexin V and 7AAD. Although no striking differences were seen in the percentages of apoptotic/dead NKT cells in the thymus (Fig. 6,A), we consistently observed an increased proportion of annexin V/7AAD+ NKT cells in the spleens of Itk−/− and Itk/Rlk−/− compared with wild-type mice (Fig. 6,B). Because IL-15 has a key role in peripheral maintenance of NKT cells (13, 14), we considered that defective survival of Itk−/− and Itk/Rlk−/− NKT cells might result from impaired expression of CD122, the IL-2R/IL-15R β-chain. As shown in Fig. 6, C and D, CD122 was dramatically decreased on the Itk−/− and the Itk/Rlk−/− NKT cells, both in the thymus and the spleen. This reduced expression of CD122 correlated well with the numbers of NKT cells in the mice, because Itk/Rlk−/− NKT cells had lower expression of CD122 than NKT cells from Itk−/− mice. Furthermore, this defect correlates with the peripheral survival data, and is consistent with the known role of IL-15 in the peripheral survival of all NKT cell subsets. Interestingly, this decrease in CD122 expression also correlates with the impaired up-regulation of NK1.1 in the Itk−/− and Itk/Rlk−/− NKT cells, and is consistent with the observation that IL-15-deficient mice have a marked decrease in mature (NK1.1+) NKT cells, with no effect on the immature NKT cells present in the thymus (13). Putting these data together may also account for why we fail to see an increase in apoptotic/dead NKT cells in the thymus of Itk−/− and Itk/Rlk−/− mice, because these thymi contain only small numbers of mature NKT cells, with few cells reaching the stage at which they would be susceptible to an absence of IL-15.
Up-regulation of CD122 has been associated with the T-box transcription factors T-bet and eomesodermin in conventional αβTCR+ T cells (54). Furthermore, T-bet-deficient mice also have severe defects in NKT cell maturation, peripheral survival, and cytokine production. Given the similarities seen between T-bet−/− and Itk−/− or Itk/Rlk−/− NKT cells, we investigated T-bet expression in these cells. To this end, NKT cells were sorted from the wild-type, Itk−/−, and Itk/Rlk−/− mice; RNA was harvested; and the samples were analyzed for T-bet mRNA expression using quantitative real-time RT-PCR. A substantial reduction in T-bet mRNA was found in the NKT cells from the Itk−/− mice (3- to 4-fold decrease) compared with wild-type NKT cells, with a more profound defect seen in NKT cells from the Itk/Rlk−/− mice (7- to 8-fold decrease) (Fig. 6 E). Because T-bet expression is up-regulated during the final stage of NKT cell development, this result is not unexpected, as most Itk−/− and Itk/Rlk−/− NKT cells do not reach this stage of maturation. However, we also found that a comparison between purified thymic NK1.1−NKT cells from the three types of mice also showed the same reduction in T-bet mRNA expression in Tec kinase-deficient cells vs wild type (data not shown). Interestingly, no decrease in T-bet expression was seen in a recent study comparing Itk−/− with wild-type NKT cells (41). Both our study and that of Au-Yeung and Fowell (41) used quantitative RT-PCR to assess T-bet mRNA levels in NKT cell populations; however, our experiments examined T-bet expression in thymic NKT cells, whereas the previous study examined splenic NKT cells. We have also analyzed splenic NKT cell populations from wild-type and Itk−/− mice for T-bet expression, and find results consistent with those reported by Au-Yeung et al. (41), that T-bet mRNA is, if anything, slightly increased in Itk−/− compared with wild-type splenic NKT cells (data not shown). In contrast, T-bet mRNA levels in the small number of Itk/Rlk−/− splenic NKT cells are decreased when compared with the wild-type NKT cells (data not shown). These findings suggest that peripheral NKT cell survival may require T-bet, leading to a strong selection for those cells able to up-regulate T-bet expression. This hypothesis could account for the differences seen between thymic vs splenic Itk−/− NKT cells regarding T-bet mRNA levels, and also, the more severe loss of Itk/Rlk−/− NKT cells compared with those lacking Itk alone.
Eomesodermin has also been shown to regulate CD122 expression in conventional T cells, similar to T-bet (54), yet is not normally expressed in NKT cells (28). Recently, we found aberrantly high expression of eomesodermin in CD8+ thymocytes and T cells from Itk−/− and Itk/Rlk−/− mice (37). Therefore, we also examined Itk−/− and Itk/Rlk−/− NKT cells for eomesodermin mRNA expression. Although no eomesodermin mRNA was detected in the wild-type or Itk/Rlk−/− NKT cells, we consistently found low levels of eomesodermin mRNA in the Itk−/− NKT cells (Fig. 6,E). Additionally, GATA-3 has recently been implicated in regulating NKT peripheral survival, function, and maturation, based on studies of GATA-3-deficient mice (30). Although no consistent differences were seen in expression of GATA-3 between the wild-type and Itk−/− NKT cells, confirming results from Au-Yeung and Fowell (41), a 15–30% decrease was seen in the Itk/Rlk−/− NKT cells compared with those of wild-type mice (Fig. 6 F). Overall, the failure of Tec kinase-deficient NKT cells to regulate expression of these key transcription factors is likely to impact NKT cell survival, due to impaired CD122 expression, NKT cell effector functions, specifically IFN-γ secretion, and NKT cell terminal maturation.
The role of Tec family kinases in conventional αβ T cell development, differentiation, and function has been well documented (33, 34, 35). Of the three family members expressed in T cells, Itk, Rlk, and Tec, Itk has the predominant role in TCR-dependent thymic selection events, as well as in TCR-dependent activation of mature T cells. In contrast, Tec appears to have no important role during T cell development or in the primary activation of naive T cells (55). Although the function of Rlk is still uncertain, a combined deficiency in Itk and Rlk leads to more severely impaired signaling than a deficiency in Itk alone, indicating some redundancy between Itk and Rlk in conventional αβ T cells (43). Unlike the well-characterized role of Itk and Rlk in conventional T cells, little is currently known about the function of these kinases in nonconventional innate T cells, such as γδTCR+ T cells, NKT cells, and other subsets of CD8+ and CD4−8− T cells with specificity for nonclassical MHC class I molecules. One previous report showed decreased numbers of NKT cells in the thymus and periphery of Itk-deficient mice, and showed that Itk−/− NKT cells had reduced expression of NK1.1 (40), whereas a more recent study focused on cytokine secretion defects of Itk−/− NKT cells (41). Based on the initial study indicating a possible role for Tec family kinases in NKT cells, we set out to determine whether Rlk could compensate for Itk in NKT cell development, and also whether the Tec kinases Itk and Rlk are required for other aspects of NKT cell biology, including effector function and peripheral homeostasis.
Overall, our data indicate that both Itk and Rlk together are critical for the terminal maturation, cytokine production, and peripheral survival of NKT cells. Interestingly, the phenotype we observe for NKT cells in Itk−/− and Itk/Rlk−/− mice shares strong similarities with defects observed in the absence of the cytokine IL-15 (13, 14), and more strikingly, in the absence of the T-box transcription factor, T-bet (28). In addition, it is likely that an earlier developmental block, similar to that seen in the SAP- and Fyn-deficient mice, is also contributing to the reduced NKT cell numbers observed in the Tec kinase-deficient mice. Because T-bet is known to regulate transcription of CD122, the IL-15R β-chain (54), the ability of NKT cells to respond to IL-15 would most likely depend on prior T-bet expression. Furthermore, a study aimed at identifying other genes regulated by T-bet in NKT cells identified a wide array of T cell effector molecules, including cytokines, chemokines, and cytolytic proteins, as well as survival factors (29). Thus, T-bet together with IL-15 is essential for terminal maturation, development of effector function, and long-term survival of NKT cells.
One important function of T-bet is regulation of IFN-γ transcription (56). This correlates well with the fact that NKT cell maturation to the NK1.1+ stage, which coincides with T-bet up-regulation, is accompanied by a switch in NKT cell effector cytokine production from IL-4 to IFN-γ (40). Thus, it is not surprising that Itk−/− and Itk/Rlk−/− NKT cells exhibit a severe defect in IFN-γ production, because few of these cells progress to the NK1.1+ stage of NKT cell maturation. Although at first glance these findings appear inconsistent with a recent study showing modest reductions in IL-4 and IFN-γ transcript levels in Itk−/− NKT cells (41), this earlier study examined only constitutive levels of IL-4 and IFN-γ transcripts, and did not examine cytokine mRNA levels following NKT cell activation. In fact, both studies agree that Itk−/− NKT cells are dramatically impaired in cytokine protein production in response to TCR stimulation, suggesting a defect in TCR-induced transcription at these cytokine loci.
As mentioned above, we also find a substantial defect in IL-4 production by Tec kinase-deficient NKT cells. This latter defect is unlikely to be due to altered NKT cell maturation, but instead, is most likely explained by impaired TCR signaling in the absence of Itk and Rlk. In fact, PMA/ionomycin activation, which bypasses proximal TCR signaling and the requirement for Tec kinases, largely restores IL-4 production to Itk−/− and Itk/Rlk−/− NKT cells. Of interest, we also see low, but detectable expression of eomesodermin mRNA in Itk−/−, but not Itk/Rlk−/−, NKT cells. This finding may in part account for the increased ability of Itk−/− over Itk/Rlk−/− NKT cells to respond to in vivo stimulation by producing some IFN-γ, and may also contribute to the IFN-γ transcripts seen by Au-Yeung and Fowell (41) in freshly isolated splenic Itk−/− NKT cells.
Based on our findings in comparison with published data, we propose that Itk and Rlk are required for the up-regulation of T-bet mRNA during the final stage of NKT cell maturation (Fig. 7). Should this be the case, it implicates TCR signaling in regulating T-bet expression, because Tec kinases are predominantly activated in T cells downstream of the TCR. This notion is at odds with current evidence regarding the regulation of T-bet transcription, which in conventional CD4+ T cells has been shown to be dependent solely on IFN-γ receptor signaling, and independent of TCR signaling (57). Thus, an alternative possibility is that TCR signaling leading to activation of Itk and Rlk is required for NKT cell maturation, but not directly responsible for regulation of T-bet transcription. Because additional signaling proteins in the TCR pathway have been shown to be important for NKT cell development, including PKCθ and Vav-1 (21, 22, 23), there is strong evidence supporting an ongoing role for TCR signaling in NKT cell maturation. However, the precise mechanism of T-bet regulation in developing NKT cells remains to be determined.
In contrast to the defects in NKT cells observed in the absence of Itk and Rlk, a deficiency in the Src kinase, Fyn, leads to a more severe and earlier block in NKT cell ontogeny (17, 58). Recently, these data have been extended to include the adapter protein SAP, which activates and recruits Fyn to SLAM family receptors (5, 18, 19, 20). Interestingly, Tec family kinases, specifically Btk and Itk, have been shown to bind to the Fyn Src homology 3 domain through an N-terminal region that is conserved between these kinases (59, 60). Although these findings suggest a connection between Tec family kinases and the SLAM-SAP-Fyn pathway, this possibility seems unlikely. Specifically, the defect in NKT cell numbers is much more severe in the Fyn-deficient mice, and occurs at an earlier stage of NKT cell maturation. Similarities between the Fyn- and SAP-deficient mice, and mice lacking CD1d (2) or the transcription factors Runx1 and RORγt (26, 27), suggest instead that TCR plus SLAM family receptor signaling is required for early steps in NKT cell positive selection, and that this developmental step is largely Tec kinase independent (Fig. 7). However, we cannot rule out the possibility that there is an additional partial block in NKT cell positive selection in Itk−/−, and more particularly Itk/Rlk−/− mice, at a stage that precedes the earliest identifiable NKT cell subset (HSAhigh). It may also be the case that, in the absence of all three Tec family kinases, Itk, Rlk, and Tec, NKT cell development would be impaired to a degree comparable to that observed in the absence of Fyn or SAP. At the current time, an Itk/Rlk/Tec triple-deficient mouse is not yet available, and in addition, the identification of early defects in NKT cell development before the up-regulation of the TCR is quite difficult.
A final clue on the role of Tec kinases in NKT cell development comes from studies of the transcription factor, B cell-activating transcription factor (BATF) (61, 62). BATF is an inhibitory component of the AP-1 family of transcription factors, which reduces AP-1-binding activity when expressed in T cells in a transgenic mouse model. In these transgenic mice, conventional T cell development is normal; however, NKT cell numbers are substantially reduced. Similar to the phenotype of NKT cells in Tec kinase-deficient mice, the BATF-transgenic NKT cells fail to undergo the final stage of maturation as measured by up-regulation of NK1.1. This finding indicates that AP-1 activation plays a crucial role in the terminal maturation of NKT cells. Because Itk and Rlk are known to be essential for optimal AP-1 activation in conventional T cells (63), these data suggest that a similar pathway operates in NKT cells, and is required for the final step of NKT cell development.
The role of Tec family kinases in innate immune cell subsets is an interesting and complex issue. In one subset of T cells, αβ CD8+ T cells, Tec family kinases are required to promote development into the conventional T cell lineage; in the absence of Itk, or Itk and Rlk, CD8+ T cell development proceeds into an innate-like lineage (36, 37, 38, 39). In this study, we show that in another subset of innate T cells, αβ NKT cells, Tec family kinases have a critical positive function in the development, maturation, function, and survival of these cells. Together, these findings indicate that Itk and Rlk have a distinct function in each lineage of innate T cells. In this context, the Tec family kinases may be required for optimal TCR signaling, and defects in this pathway may lead to alterations in lineage development in the thymus. Alternatively, it remains possible that Itk and Rlk may be important for a distinct signaling pathway, such as that mediated by a costimulatory receptor, which differentially impacts NKT cell, conventional, and innate CD8+ T cell maturation.
We thank Atsushi Tanaka and Albert Bendelac (University of Chicago, Chicago, IL) and the National Institute of Allergy and Infectious Diseases Tetramer Facility (Atlanta, GA) for CD1d/αGAL tetramers, and Paul Stein (Northwestern University, Chicago, IL) for helpful discussions. We thank Amanda Prince, Yoko Kosaka, John Evans III, and Regina Whitehead for technical assistance and helpful discussions, and Raymond Welsh for the gift of Ab reagents.
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.
This work was supported by grants from the National Institutes of Health (AI37584, AI66118) and the Centers for Disease Control and Prevention (CI000101).
Abbreviations used in this paper: DP, double positive; αGAL, α-galactosylceramide; 7AAD, 7-aminoactinomycin D; BATF, B cell-activating transcription factor; HSA, heat stable Ag; PKCθ, protein kinase Cθ; RORγt, retinoic acid receptor-related orphan receptor-γt; Runx1, runt-related transcription factor 1; SAP, signaling lymphocytic activating molecule-associated protein; SLAM, signaling lymphocytic activating molecule.