|Home | About | Journals | Submit | Contact Us | Français|
The invariant NKT (iNKT) cell lineage contains CD4+ and CD4- subsets. The mechanisms that control such subset differentiation and iNKT cell maturation in general have not been fully understood. RasGRP1, a guanine nucleotide exchange factor for T cell receptor-induced activation of the Ras-Erk1/2 pathway, is critical for conventional αβ T cell development but dispensable for generating regulatory T cells. Its role in iNKT cells has been unknown. Here we report severe decreases of iNKT cells in RasGRP1-/- mice through cell intrinsic mechanisms. In the remaining iNKT cells in RasGRP1-/- mice, there is a selective absence of the CD4+ subset. Furthermore, RasGRP1-/- iNKT cells are defective in T cell receptor induced proliferation in vitro. These observations establish that RasGRP1 is not only important for early iNKT cell development, but also for the generation/maintenance of the CD4+ iNKT cells. Our data provides genetic evidence that the CD4+ and CD4- iNKT cells are distinct sub-lineages with differential signaling requirements for their development.
Natural killer T (NKT) cells are subsets of T cells co-expressing markers found on NK cells and T cells. While rare in number, NKT cells play important roles in immune responses and pathogenesis of disease (1-3). The invariant Vα14-Jα18 T cell receptor (iVα14TCR) expressing NKT (iNKT) cells represent the major subset within the NKT cell lineage and are the best characterized (4, 5). The iVα14TCR recognizes both endogenous and synthetic glycolipids such as iGB3 and α-galactosyl ceramide (α-GalCer), respectively, presented by CD1d (6, 7). Use of CD1d tetramers loaded with α-GalCer has provided a pivotal tool to define iNKT cells and has allowed for the delineation of iNKT cell development into multiple developmental stages. The earliest iNKT cells (stage 0) are defined as CD24+CD44-NK1.1-, and such cells are extremely rare in the thymus. As iNKT cells mature, they down-regulate CD24 expression and progress sequentially through stage 1 (CD24-CD44-NK1.1-), stage 2 (CD24-CD44+NK1.1-), and finally stage 3 (CD24-CD44+NK1.1+) (8, 9). Further from these stage definitions, iNKT cells can also be divided into CD4+ and CD4- subsets that may branch out at the stage 1 and represent two different sub-lineages of iNKT cells (10). However, genetic evidence supporting such sub-lineage definition remains quite rare.
The iVα14TCR is critical for iNKT cell development. Deficiency of the receptor or its ligand CD1d results in a failure to generate iNKT cells in mice (11-13). Upon TCR engagement, PLCγ1 plays a crucial role in TCR signaling by producing diacylglycerol (DAG) and inositol 1,4,5-tris-phosphate (IP3) second messengers (14). IP3 activates the Ca++-calcineurin-NFAT pathway, which has been recently demonstrated to be crucial for iNKT cell maturation via the transcription factor Egr2 (15). DAG activates the PKCθ-Carma1/Bcl10/Malt-IKK-NFκB pathway. The NFκB pathway is critical for iNKT cell ontogeny, as deficiencies of its different components have been shown to block iNKT cell development at various stages (16-21). DAG also associates with and activates RasGRP1, a guanine nucleotide releasing factor for Ras. RasGRP1 in turn activates the downstream Ras-Erk1/2-AP1 pathway. The RasGRP1-Ras-Erk1/2 pathway is important for positive selection of conventional αβ T (cαβT) cells (22, 23). While uncontrolled DAG-mediated signaling due to absence of diacylglycerol kinases α and ζ causes severe defect of iNKT cell development (24), the role of RasGRP1 in iNKT cell development remains unclear. In this report, we demonstrate severe decreases of iNKT cells and a selective absence of the CD4+ subset of iNKT cells in RasGRP1-/- mice. Our data not only reveals a critical role of RasGRP1 for early iNKT cell development, but also provide genetic evidence that the CD4+ and CD4- subsets of iNKT cells are indeed distinct sublineages since they have differential signaling requirements for their generation/maintenance.
The C57BL6/J and TCRβ-/-δ-/- mice were all purchased from the Jackson Laboratory. The RasGRP1-/- mice were previously described (22) and were backcrossed onto B6 background for 9 generations. All mice were used according to a protocol approved by the Duke University Institute Animal Care and Use Committee. Thymocytes and splenocytes were prepared following standard procedures. Liver mononuclear cells were isolated according to a published protocol (18).
Cells were stained with PE- or APC-CD1d-Tet (NIH tetramer core facility) and fluorescence-conjugated anti-mouse CD24, CD44, NK1.1, CD4, CD8, TCRβ, CD45.1, CD45.2, Thy1.1, Thy1.2, CD1d, CD150 (SLAM), and Ly108 (SLAMF6) antibodies (BioLegend) in PBS-2% FBS on ice for 30 minutes. Cell survival/death was determined by addition of the Live/Dead® Fixable Violet Dead Cell Stain (L/D, Invitrogen) during the staining according to the manufacturer’s protocol. Dead cells stain positive for L/D. For Ki67 expression, cells were permeabolized using the Foxp3 staining kit (eBioscience) after cell surface staining, followed by staining with unconjugated anti-Ki-67 (B56, BD Biosciences). An Alexa Fluor® 488-conjugated goat anti-mouse IgG (H+L) (Invitrogen) was used to detect anti-Ki67 antibody. Stained cells were collected on FACSCanto™ II (BD Biosciences) and analyzed using the Flowjo software.
Thymocytes were resuspended in 500 μl IMDM with 10% FBS (IMDM-10) and were sequentially added with 5μl Fc-blocker from the EasySep PE selection kit (Stem Cell Technologies) and 2.5 μl of PE-CD1d-Tetramer. After incubation at room temperature (RT) for 15 minutes, cells were washed once with IMDM-10. The cells were resuspended in 500 μl of IMDM-10 and mixed with 5 μl of EasySep PE selection cocktail. After incubation at RT for 15 minutes, 5 μl of EasySep nanoparticles were added and the mixture was incubated at RT for additional 15 minutes. After addition of IMDM-10 to a total volume of 2.5 ml, cells in FACS tubes were inserted into the EasySep magnet and let stand for 5 minute. The unbound cells were discarded and the bound cells were resuspended in 2.5 ml IMDM-10. After repeating magnetic enrichment for another time, the magnetic bounding fractions were collected for staining and FACS analysis.
Thymocytes from WT and RasGRP1-/- mice were labeled with 10 μM CFSE at RT for 9 minutes as previously described (25). Cells were seed at 5 ×106 cells/ml in a 48-well plate plate and left unstimulated or stimulated with 125 ng/ml α-GalCer at 37°C for 72 hours. Cells were then stained for TCRβ and APC-conjugated CD1d-Tet before analyzed by flow cytometry.
Recipient TCRβ-/-δ-/- mice were sublethally irradiated (600 rad) one day before adoptive transfer. Ten million 1:1 mixed bone marrow (BM) cells from age- and sex-matched CD45.1+ B6 and CD45.2+ RasGRP1-/- mice were intravenously injected into the recipients. Alternatively, lethally irradiated (1100 rad) WT C57B6 mice were used as recipients and were reconstituted with Thy1.1+-C57B6 (WT) and Thy1.2+-RasGRP1-/- BM cells at 1:10 ratio. The resulting chimeric mice were analyzed 7 to 8 weeks later.
Viable CD4+CD8+ double positive (DP) thymocytes and TCRβ+CD1dTet+ iNKT cells from age- and sex-matched control or RasGRP1-/- mice were sorted on MoFlo Cell Sorter (Beckman Coulter), with post-sort purity>98%, and lysed in Trizol (Invitrogen). Total RNAs were extracted, and cDNAs were obtained using the Superscript III First-Strand Synthesis System (Invitrogen). Realtime PCR was prepared using the RealMasterMix (Eppendorf) and performed on the Mastercycler® ep realplex2 system (Eppendorf). Primers used for different genes are listed as following. SAP: forward 5’-acgcctctgcagtatccagt-3’, reverse 5’-ttcttcatggtgcattcagg-3’; Fyn: forward 5’-caagccaagcagtgtttgaa-3’, reverse 5’-acattgcacacagcccatta-3’; RORγt: forward 5’-cgactggaggaccttctacg-3’, reverse 5’-ttggcaaactccaccacata-3’; RUNX1: forward 5’-gcaggacgaatcacactgaa-3’, reverse 5’-tggcatctctcatgaagcac-3’; cMyc: forward 5’-tgaaggctggatttcctttg-3’, reverse 5’-ttctcttcctcgtcgcagat-3’; HEB: forward 5’-aggtatggatgagcgtggag-3’, reverse 5’-agccttcgtgggttcctaat-3’; PLZF: forward 5’-tgcgcagctatatttgcagt-3’, reverse 5’-tgtggctcttgagtgtgctc-3’; RasGRP1: forward 5’-agcccaccttctgtgacaac -3’, revers3 5’-cttcttgcactcgaacacca-3’; RasGRP2: forward 5’-gggcttcgtacacaacttcc-3’, reverse 5’-gtggcagttcacaccacaag-3’.
Decreasing amounts of DNA template (100ng, 33ng, 11ng) from sorted viable RasGRP1+/- and RasGRP1-/- CD4+CD8+ thymocytes were used for semi-quantitative PCR. The forward primer for Vα14 segment was 5’-acactgccacctacatctgt-3’. The reverse primers for different Jα segments were: Jα2 5’-ggttgcaaatggtgccactt-3’; Jα18 5’-gtagaaagaaacctactcacca-3’; Jα56 5’-tgtcatcaaaacgtacctggt-3’. Primers for CD14 PCR (loading control) were: forward 5’-gctcaaactttcagaatctaccgac-3’, reverse agtcagttcgtggaggccggaaatc-3’.
For statistic analysis, two-tail Student t-test was performed. *, p<0.05. **, p<0.01, ***, p<0.001.
The expression of RasGRP1 and the other RasGRP1 family members in iNKT cells has been unknown. We first examined their expression in iNKT cells. iNKT cells stained positive for both α-Galcer loaded CD1d tetramer (CD1d-Tet) and TCRβ in thymocytes from wild-type (WT) mice. These iNKT cells were sorted by FACS and mRNA levels of these genes were determined by real-time quantitative PCR. As shown in Fig 1A, both RasGRP1 and RasGRP2 can be detected in cαβT cells and iNKT cells. However, RasGRP3 and RasGRP4 were undetectable in iNKT cells (Data not shown).
To determine the role of RasGRP1for iNKT cell development, we analyzed mice deficient in RasGRP1. Thymocytes, splenocytes, and liver mononuclear cells were stained with CD1d-Tet, as well as other cell surface markers. As shown in Figure 1B-1C, CD1dTet+TCRβ+ cells were decreased about 6 – 11 fold in RasGRP1-/- mice as compared to RasGRP1+/- mice. To further determine whether RasGRP1 is required for generation of stage 0 CD24+ iNKT cells, we enriched iNKT cells from WT and RasGRP1-/- thymus using PE-CD1d-Tet and magnetic beads. As shown in Figure 1D-1E, the CD24+ iNKT cell number in RasGRP1-/- thymi was 6-fold lower than WT control, indicating that RasGRP1 is required for efficient generation of stage 0 iNKT cells. Further analysis of the CD1dTet+CD24- iNKT cells revealed relative enrichment of CD44-NK1.1- and CD44+NK1.1- populations, but a decrease of the CD44+NK1.1+ population in RasGRP1-/- mice (Fig. 1F). Together, these observations indicate that RasGRP1 is critical for early iNKT cell development and is also involved in promoting iNKT maturation at later stages. Similar to a previous report (22), there was a severe decrease of cαβT cells and about 50% decrease of total thymocyte number in RasGRP1-/- mice (Fig. 1G and data not shown). The ratios of iNKT cells to cαβT cells in RasGRP1-/- thymus and spleen were similar to those of WT controls (Fig. 1H). Thus, the defect of iNKT cell development caused by RasGRP1 deficiency was in parallel with that of cαβT cells. However, the iNKT cells to cαβT cell ratio was about seven-fold lower in RasGRP1-/- liver than in WT liver.
Since RasGRP1 is expressed in multiple cell lineages and iNKT cells are positively selected by engagement of iVα14TCR with CD1d expressed on DP thymocytes, we further investigated whether the developmental defects of RasGRP1-/- iNKT cells are intrinsic. To this end, a 1:1 mixture of CD45.2+ RasGRP1-/- and CD45.1+CD45.2+ WT BM cells were adoptively transferred to reconstitute TCRβ-/-δ-/- hosts. The recipients were analyzed seven to eight weeks after reconstitution. Although a close to 1:1 ratio of WT and RasGRP1-/- BM cells were injected into the recipients (Fig. 2A), only 15% of thymocytes from the recipients and less than 40% of total splenocytes and liver mononuclear cells were derived from RasGRP1-/- origin (Fig. 2B-2C). There were progressive decreases of representation by RasGRP1-/- thymocytes as they mature from the CD4-CD8- double negative (DN), the CD4+CD8+ DP, to the CD4+CD8- or CD4-CD8+ single positive (SP) stage. The decrease was most severe within the RasGRP1-/- CD4 SP and CD8 SP populations. Further analysis of non-T cells (DN) from splenocytes in the recipients revealed a roughly equal contributions of WT and RasGRP1-/- origins, suggesting that RasGRP1 deficiency does not globally affect hematopoietic stem cell engraftment or early lymphoid progenitor cells and that RasGRP1 may promote cαβT cell development at multiple stages.
In the thymi of recipient mice, CD45.1+CD45.2+ WT CD1dTet+ iNKT cells could be easily detected. However, the CD45.2+ RasGRP1-/- CD1dTet+ iNKT cells were virtually undetectable in the recipients (Fig. 2C). Similarly, few RasGRP1-/- iNKT cells were observed in the spleen and liver as well. Due to under representation of RasGRP1-/- thymocytes in chimeric mice reconstituted with WT and RasGRP1-/- BM cells at 1:1 ratio, we further generated and analyzed chimeric mice reconstituted with BM cells from WT (Thy1.1) and RasGRP1-/- (Thy1.2) mice at 1:10 ratio. As shown in figure 2D, Thy1.1 and Thy1.2 staining of thymocytes from recipients displayed a close to expected ratio of cells originated from WT to RasGRP1-/- BM cells. Severe decreases of Thy1.2+ RasGRP1-/- iNKT cells as well as CD4 and CD8 SP cαβT cells were observed as compared with their Thy1.1+ WT counterparts in the recipients. Together, these observations indicate that the developmental defects of RasGRP1-/- iNKT cells and cαβT cells are cell-intrinsic and cannot be rescued by WT thymocytes. CD1d expression on cortical thymocytes are critical for iNKT cell development (26, 27). No obvious difference was observed between RasGRP1-/- and control thymocytes (Fig. 2E). Together, these data indicate that RasGRP1 deficiency does not affect CD1d expression, and rule out defective presentation by cortical thymocytes as a cause of defective iNKT cell development in RasGRP1-/- mice.
Insufficient Vα14-Jα18 recombination has been shown to cause a severe developmental block early in iNKT development in some mouse models (28, 29). We detected similar levels of Vα14 to Jα18, Jα2, or Jα56 recombination in RasGRP1+/- and RasGRP1-/- CD4+CD8+ DP thymocytes (Fig. 3A), ruling out the possibility that the deficiency of RasGRP1 somehow inhibited Vα14-Jα18 recombination. Beside iVα14TCR, homotypic interactions of cell surface receptors Slamsf1 and Slamsf6 on thymocytes also play an essential role in iNKT development (30). No differences in the surface expression of these receptors were detected between RasGRP1+/- and RasGRP1-/- thymocytes (Fig. 3B). However, a significantly higher rate of cell death was observed in the RasGRP1-/- CD1d-Tet+TCRβ+ iNKT cells as well as CD1d-Tet-TCRβ+ cαβT cells than in the RasGRP1+/- controls, suggesting that RasGRP1 is important for iNKT and cαβT cell survival, and increased death of these cells may contribute to the developmental defects in RasGRP1-/- mice (Fig. 3C). Ki67 expression is usually correlated with cell division. Freshly isolated RasGRP1-/- iNKT cells displayed a higher level of Ki67 staining compared with control iNKT cells (Fig. 3D-3E), suggesting increased homeostatic proliferation of RasGRP1-/- iNKT cells in vivo, likely due to the T cell lymphopenic environment in these mice. We further used CFSE-dilution assay to examine whether RasGRP1 regulates TCR induced iNKT cell activation. As shown in Figure 3F, WT but not RasGRP1-/- iNKT cells proliferated following α-GalCer stimulation for 72 hours in vitro (Fig. 3F). Together, these data suggest that RasGRP1 is important for iVα14TCR-induced iNKT cell activation.
Signaling proteins SAP (31, 32) and Fyn (33, 34), as well as several transcription factors, such as RORγt (29, 35), Runx1 (35), cMyc (36), and HEB (28), are all critical for early iNKT cell development (37). No obvious differences in mRNA expression of these molecules were detected between RasGRP1+/- and RasGRP1-/- DP thymocytes (Fig. 3G), ruling out that RasGRP1 deficiency may affect iNKT cell development through modulating mRNA expression of these molecules. However, expression of PLZF, a transcription factor critical for the development of CD44+ iNKT cells (38, 39), was much lower in RasGRP1-/- iNKT cells than in RasGRP1+/- control, which might contribute to the relative enrichment of CD24-CD44- iNKT cells in RasGRP1-/- mice.
It has been demonstrated that, while stage 0 iNKT cells all express CD4 (8), the presence of CD4- iNKT cells can be observed in thymus at later stages or in the periphery. While accumulating evidence has revealed that the CD4+ and CD4- iNKT cells are functionally distinct (10, 40, 41), the developmental relationship between these two subsets is not well understood. Recently published data show that the CD4-NK1.1- cells appeared to be precursors of the CD4-NK1.1+ iNKT cells in the thymus. A revised model of thymic iNKT development was proposed in which the CD4- and CD4+ subsets represent two distinct sub-lineages of iNKT cells, whose divergence appears to occur at stage 1 when the CD4- iNKT cells are first observed (10). However, genetic evidence supporting such lineage definition is rare and mechanisms directing such lineage differentiation are not well defined. Strikingly, a dramatic decrease in the percentage of CD4+ subset was observed in the RasGRP1-/- iNKT cells in thymus, spleen, and liver (Fig. 4A-4B). When assessing the CD4 expression pattern at each iNKT developmental stage in RasGRP1+/- thymus, there is a progressive increase of the CD4- subset as the iNKT cells mature. About 10%, 25% and 40% of stage 1, stage 2 and stage 3 iNKT cells are CD4-. However, in RasGRP1-/- iNKT cells, about 90% of stage 2 and stage 3 iNKT cells were CD4-, yet no difference in CD4+/CD4- ratio was observed at stage 1 as compared to RasGRP1+/- (Fig. 4C-4D). Thus, besides promoting early iNKT cell development, RasGRP1 is selectively required for the maturation and/or maintenance of the CD4+CD44+ iNKT cells.
RasGRP1 promotes positive selection of cαβT cells, particularly those expressing TCR with low affinity to self-peptide-MHC complex (22). Positive selection of thymocytes with relative high affinity to self-peptide-MHC complex, including regulatory T cells and some innate CD8 T cells, is less dependent on RasGRP1 (42, 43). We have demonstrated here that RasGRP1 plays crucial roles in iNKT cell development and is important for the generation and/or maintenance of CD4+ iNKT cells (Fig. 4E). At present, it is still unclear how RasGRP1 promotes αβT and iNKT cell maturation. The increased death of RasGRP1 deficient cαβT cells and iNKT cells suggests that RasGRP1 may promote normal development of iNKT and cαβT cells by enhancing their survival. Of note, in addition to activating the Ras-Erk1/2 pathway in thymocytes following TCR engagement (22), we have recently found that RasGRP1 is also critical for TCR-induced activation of PI3K/Akt and the mammalian target of rapamycin (mTOR) (44). Both PI3K/Akt and mTOR are important regulators for cell survival, growth, and metabolism (45-47). It is likely that RasGRP1 may promote iNKT cells and cαβT cell maturation through multiple mechanisms.
The CD4+CD44+ iNKT cells are selectively or more severely affected than the CD4-CD44+ iNKT cells by RasGRP1 deficiency, suggesting that these two subsets of cells may signal differently. In RasGRP1 deficient thymocytes, TCR-induced activation of Ras/Erk1/2, PI3K/Akt, and mTOR is greatly decreased but not completely abolished (44). The exact differences of these signaling events between the CD4+iNKT T cells and CD4-iNKT T cells, as well as the effect of RasGRP1 deficiency on the activation of these signaling pathways in iNKT cells, are hard to assess since these cells are rare. At present, it is unclear whether the CD4-CD44+ iNKT cells are independent or less dependent of one or multiple signaling pathways downstream of RasGRP1 or they utilize other guanine nucleotide exchange factors such as Sos to activate these downstream signaling molecules. However, RasGRP1 promotes Sos to induce Ras activation (48). TCR induced Ras/Erk1/2 activation in RasGRP1-/- iNKT cells is likely decreased and the CD4+CD44+ and the CD4+CD44- iNKT cells probably have differential requirement for the Ras/Erk1/2 pathway. In addition to RasGRP1, deficiency of the transcription factor GATA-3 also cause a severe decrease of CD4+ iNKT cells in mice (49). Together, these observations provide genetic evidence that the CD4+ and CD4- iNKT cells are distinct sublineages with differential signaling/transcription factor requirements for their development. Further studies are required to determine whether RasGRP1 and GATA3 may regulate each other to promote CD4+ iNKT cell development.
It is important to note that our data appear to contradict a previous report that the Ras-Mek1/2-Erk1/2 pathway is dispensable for NKT cell development (50). In that study, dominant negative Ras and Mek1, specifically expressed in thymocytes, cause severe decreases of CD4+CD8- and CD4-CD8+ single positive thymocytes. However, NK1.1+TCRβ+ T cells were reported to be normal. Since CD1d-Galcer tetramer was not available at that time, the effects of dnRas/dnMek1 on iNKT cell development remain unclear. However, we did observe sharp decrease of NK1.1+TCRβ+ cells in RasGRP1-/- mice as well (data not shown). The discrepancy between these two studies could result from less complete abolishment of the Ras-Erk1/2 signaling in thymocytes of the dnRas/dnMek transgenic mice than in the RasGRP1-/- mice, some unknown effects of dnRas/dnMek1 transgenes on the cells, or variegate expression pattern due to the integration site effects on the transgenes. Additionally, RasGRP1 deficiency and dnRas or dnMek1 may differentially affect signaling pathways such as the PI3K/Akt, mTOR, and other yet to be identified signaling pathways that may play different roles for iNKT cell development.
We thank Nancy Martin in Duke Cancer Center Flow Cytometry Core Facility for providing sorting services, Li Xu for technical support, the NIH Tetramer Core Facility for providing the CD1d tetramer, and Tommy O’Brien and Dr. Kim Nichols for critical reviewing the manuscript.
2This study is supported by funding from the National Institute of Health (R01AI076357, R01AI079088, and R21AI079873), the American Cancer Society, and the American Heart Association to X-P.Z.