Article

Nature Immunology 8, 1345 - 1352 (2007)
Published online: 21 October 2007 | doi:10.1038/ni1524

NKG2D signaling is coupled to the interleukin 15 receptor signaling pathway

Tiffany Horng1,2, Jelena S Bezbradica1 & Ruslan Medzhitov1

The effector functions of natural killer cells are regulated by activating receptors, which recognize stress-inducible ligands expressed on target cells and signal through association with signaling adaptors. Here we developed a mouse model in which a fusion of the signaling adaptor DAP10 and ubiquitin efficiently downregulated expression of the activating receptor NKG2D on the surfaces of natural killer cells. With this system, we identified coupling of the signaling pathways triggered by NKG2D and DAP10 to those initiated by the interleukin 15 receptor. We suggest that this coupling of activating receptors to other receptor systems could function more generally to regulate cell type–specific signaling events in distinct physiological contexts.

Natural killer (NK) cells are important in the immune system in antiviral defense and tumor surveillance1, 2. They recognize and kill stressed or infected cells that express markers of viral infection or transformation but spare healthy cells displaying markers of normal self. Recognition of these ligands by NK cells is mediated by cell surface receptors that fall into two categories: activating and inhibitory3, 4. Inhibitory receptors consist of a single polypeptide containing cytoplasmic immunoreceptor tyrosine-based inhibitory motifs that recruit tyrosine phosphatases and abort signaling. Whereas activating receptors typically couple to signaling adaptors that contain either an immunoreceptor tyrosine-based activation motif (ITAM; such as DAP12, FcRgamma or CD3zeta) or a 'YXNM' motif, where 'X' is any amino acid (such as the signaling adaptor DAP10) through interactions specified by their transmembrane regions3. It is generally thought that integration of positive and negative signals from activating and inhibitory receptors determines 'downstream' signaling events. Signals emanating from activating and inhibitory receptors determine the repertoire of NK receptors expressed on developing NK cells and regulate effector functions such as cytotoxicity and the production of interferon-gamma (IFN-gamma) and other inflammatory mediators in mature NK cells5.

A critical component of NK cell maturation is selection of the activating and inhibitory receptor repertoire on self major histocompatibility molecules expressed on stromal cells. This 'education' process ensures that mature NK cells will not attack self but can become activated by altered self or missing self and is in many ways analogous to self antigen–mediated selection of antigen receptors during the development of T lymphocytes and B lymphocytes6, 7. Notably, NK cell activating receptors can associate with more than one adaptor and, conversely, each signaling adaptor can pair with multiple receptors. Moreover, most activating receptors are expressed by overlapping subsets of NK cells, such that NK cells are oligoclonal in their expression of these receptors. For this reason, there is some redundancy in terms of the functions of NK cell receptors and adaptors8, 9. Indeed, in contrast to the absolute requirement for antigen receptors for the development of T lymphocytes and B lymphocytes, DAP10-deficient mice10, DAP12-deficient mice11, 12, mice deficient in the activating receptor NKG2D (D. Raulet, personal communication), and even mice lacking functional DAP12, FcRgamma and CD3zeta13, have no fewer NK cells than do wild-type mice.

We were therefore interested in alternative systems for studying NK cell receptors that would address their redundancy and identify functions not obvious by typical gene-targeting approaches. We focused on DAP10, an adaptor that transmits signals for NKG2D14 and possibly for other as-yet-unidentified receptors. Of the many activating receptors expressed by NK cells, NKG2D is distinguished by its ubiquitous expression on all NK cells, its recognition of stress- and infection-inducible ligands, and its expression early during NK cell development15, 16.

We targeted DAP10 and its associated proteins for ubiquitin-mediated degradation in NK cells in vivo. To our knowledge, this is the first use of such a strategy for the downregulation of cell surface proteins. We found that elimination of the DAP10 signaling complex blocked NK cells from responding to the cytokine interleukin 15 (IL-15). Furthermore, we found that after activation by IL-15, Janus kinase 3 (Jak3) phosphorylated DAP10, thus priming DAP10 for NKG2D-dependent signaling in a transcription- and translation-independent way. Therefore, in addition to its well-established involvement in the survival and proliferation of NK cells, IL-15 is also required for the priming of NK cells for cytotoxic responses induced by NKG2D.

Top

Results

Targeting the DAP10 complex for degradation

We wanted to design a strategy to target DAP10 and its associated proteins for degradation. Studies of yeast have shown that a truncated form of the cell surface protein Ste3p lacking known signals for internalization and endocytosis becomes targeted for vacuolar degradation when fused at its carboxyl terminus to a ubiquitin monomer17. Moreover, the physiologically relevant NKG2D-DAP10 receptor complex consists of two molecules of NKG2D paired with two disulfide-bonded DAP10 dimers18, and the transmembrane regions of the receptor and the adaptor are necessary and sufficient for specifying association between the two proteins19. Therefore, we speculated that a similar strategy with fusion of DAP10 to ubiquitin might be able to target DAP10 and its associated proteins, including NKG2D, for internalization from the plasma membrane and subsequent lysosomal degradation20.

To test that hypothesis, we truncated the cytoplasmic tail of DAP10 immediately before its critical YXNM signaling motif and appended monoubiquitin to create the chimeric protein DAP10-Ub (Supplementary Fig. 1 online). We sought to determine if DAP10-Ub would be constitutively targeted for endocytosis and removed from the cell surface and, furthermore, whether DAP10-Ub would function in a dominant way to downregulate associated cell surface proteins. We transfected human embryonic kidney 293T cells with constructs encoding NKG2D alone, NKG2D and Flag-tagged DAP10 (Flag-DAP10), or NKG2D and Flag–DAP10-Ub, and used flow cytometry to measure cell surface expression of NKG2D and the adaptor proteins (Fig. 1a). Consistent with published studies, transfection of DAP10 and NKG2D together was required for cell surface expression of the two proteins. However, expression of DAP10-Ub and NKG2D together resulted in downregulation of both DAP10-Ub and NKG2D (Fig. 1a). NKG2D downregulation occurred even in the presence of coexpressed DAP10 and DAP10-Ub (Supplementary Fig. 2 online), suggesting that DAP10-Ub acted in a dominant way. Moreover, the NKD2D-S isoform that associates with both DAP10 and DAP12 (ref. 21) was also downregulated from the cell surface when expressed together with DAP12 and DAP10-Ub (Supplementary Fig. 3 online), further confirming that DAP10-Ub acts in a dominant way. Therefore, we made transgenic mice expressing DAP10-Ub under the control of a human CD2 promoter.

Figure 1: Cell surface expression of the NKG2D-DAP10 receptor complex is downregulated by DAP10-U6.

Figure 1 : Cell surface expression of the NKG2D-DAP10 receptor complex is downregulated by DAP10-U6.

(a) Flow cytometry of 293T cells transiently transfected with 'empty' plasmid (Mock) or plasmids encoding NKG2D alone, NKG2D and Flag-DAP10, or NKG2D and Flag-DAP10-Ub, then stained 24 h later with antibodies specific for NKG2D and Flag. Representative of three experiments. (b) Flow cytometry of NKG2D expression on NK cells from spleens of wild-type (WT) or transgenic (Tg) mice, either naive (–) or injected with poly(I:C), and on LAK cells prepared from wild-type or transgenic splenocyte cultures. Plots are gated on NK1.1+CD3epsilon- populations. Representative of at least three experiments with NK cells from over 50 transgenic mice.

Full size image (36 KB)

DAP10-Ub–transgenic mice were born at the expected mendelian frequency and showed no gross abnormalities. Analysis of various lymphoid populations indicated expression of the transgene in NK cells, CD8+ T cells, CD4+ T cells and immature CD19+IgM- B cells (Supplementary Fig. 4 online). With the exception of NK and NKT cells (discussed below), the main lymphocyte subsets were present in similar numbers in wild-type and transgenic spleens and thymi (Supplementary Table 1 online), although we noted consistently fewer B cells in transgenic spleens, bone marrow and lymph nodes (Supplementary Table 1 and data not shown), particularly in younger mice.

To test if DAP10-Ub was able to downregulate DAP10-associated proteins, we measured cell surface expression of NKG2D, the only known receptor associated with DAP10, on wild-type and transgenic cells (Fig. 1b and Supplementary Fig. 5 online). As expected, NKG2D was expressed on wild-type NK cells and CD8+ T cells with activated-memory phenotypes15. Notably, cell surface expression of NKG2D was much lower on the corresponding populations from DAP10-Ub–transgenic mice (Fig. 1b and Supplementary Fig. 5). Wild-type and transgenic NK cells activated with the Toll-like receptor 3 ligand poly(I:C) and in vitro–generated wild-type and transgenic lymphokine-activated killer (LAK) cells showed similar differences in extracellular NKG2D staining (Fig. 1b).

Blocking NKG2D-dependent functions

We next sought to determine the functional consequences of the lower cell surface NKG2D expression. To do this, we used an in vitro killing assay (Fig. 2a) in which we mixed target cells with either wild-type or transgenic LAK cells and assessed cytotoxicity by measuring the release of cytoplasmic lactate dehydrogenase into the culture supernatants. We noted a lower frequency of NK cells in the transgenic LAK cell cultures but adjusted all in vitro killing experiments such that at any given effector/target ratio, we compared killing of targets by the same number of wild-type or transgenic NK cells. As expected, wild-type LAK cells efficiently killed RMA mouse T cell lymphoma target cells infected with a retrovirus encoding the NKG2D ligand Rae1 and the human CD2 marker (Rae1-RMA cells) but spared target cells infected with a control retrovirus expressing only human CD2 (MIG-RMA cells). In contrast, transgenic LAK cells were impaired in their ability to eliminate Rae1-RMA cells, indicating a defect in NKG2D-triggered cytotoxicity (Fig. 2a). We further confirmed those results with an in vivo killing assay (Fig. 2b) by labeling Rae1-RMA cells with the cytosolic dye CFSE and injecting the cells into wild-type, DAP10-Ub–transgenic and perforin-deficient mice previously primed with poly(I:C) to activate NK cells. We also injected unlabeled control target cells (MIG-RMA) to normalize for the recovery of target cells from mice 24 h later. Consistent with the in vitro killing assay, transgenic NK cells were profoundly impaired in their ability to eliminate Rae1-RMA cells. In fact, this deficiency was similar to that of NK cells lacking perforin (Fig. 2b), shown before to be essential for NKG2D-dependent cytotoxicity22. Because of differences in the numbers of NK cells in wild-type versus transgenic mice (discussed below), the in vivo killing assays indicated defects in NKG2D-dependent killing at the level of the NK cell population, whereas the in vitro killing assays demonstrated that transgenic NK cells show impaired NKG2D-dependent cytotoxicity on a 'per-cell' basis. We conclude that transgenic NK cells fail to engage in cytotoxic responses triggered by NKG2D signaling, consistent with their lower surface expression of NKG2D.

Figure 2: Lower NKG2D-dependent cytotoxicity and IFN-big gamma production in transgenic cells.

Figure 2 : Lower NKG2D-dependent cytotoxicity and IFN-|[gamma]| production in transgenic cells.

(a) NKG2D-dependent killing, assessed by release of cytoplasmic lactate dehydrogenase in vitro from wild-type or transgenic LAK cells added to Rae1-RMA or MIG-RMA cells at various effector/target ratios (E:T). Representative of three experiments. (b) Killing of CFSE-labeled Rae1-RMA cells (CFSE+, human CD2+) and unlabeled MIG-RMA cells (CFSE-, human CD2+) injected together into wild-type, transgenic or perforin-deficient (Prf1-/-) mice primed 24 h earlier with poly(I:C); target cells were recovered 24 h later. Below plots, ratio of MIG-RMA to Rae1-RMA cells. Human CD2 staining identifies cells infected with control and Rae1 retroviruses. Numbers above outlined areas indicate percent CFSE-CD2+ cells (left) or CFSE+CD2+ cells (right). Data are from one representative mouse from each group of two and are representative of one experiment. (c) Flow cytometry of intracellular IFN-gamma in splenocyte samples enriched for NK cells from poly(I:C)-treated wild-type or transgenic mice, incubated with plate-bound Rae1-Fc or Fc control. Plots are gated on CD3epsilon- populations. Numbers above outlined areas indicate percent NK1.1+IFN-gamma- cells (left) or NK1.1+IFN-gamma+ cells (right); numbers in the top right corners indicate percent IFN-gamma+ NK cells in the NK1.1+CD3epsilon- population. Representative of two experiments. (d) IFN-gamma production by the cells described in c after stimulation with PMA and ionomycin. Plots are gated on NK1.1+CD3epsilon- cells; numbers in plots indicate percent IFN-gamma+ cells in the NK1.1+CD3epsilon- population. Representative of two experiments. (e) IFN-gamma production by the cells described in c after stimulation with plate-bound anti-NK1.1 (alpha-NK1.1; NK1.1 crosslinking) or isotype control antibody. Because NK cells cannot be identified by NK1.1 staining after stimulation with anti-NK1.1, the percent NK cells in these samples was calculated in duplicate samples not incubated with anti-NK1.1: wild-type, 18.8%; transgenic, 2.2%. Numbers above outlined areas indicate percent NK1.1+IFN-gamma- cells (left, isotype control) or NK1.1+IFN-gamma+ cells (anti-NK1.1, and right, isotype control); numbers in the top right and bottom right corners indicate percent NK cells making IFN-gamma relative to the NK1.1+CD3epsilon- population. Representative of two experiments.

Full size image (78 KB)

Next we evaluated IFN-gamma production by transgenic NK cells. By intracellular cytokine staining (to control for differences in NK cell numbers), we found that IFN-gamma production in response to NKG2D ligation by plate-bound Rae1-Fc fusion proteins was much lower in transgenic NK cells than in their wild-type counterparts (Fig. 2c). Notably, wild-type and transgenic NK cells produced similar amounts of IFN-gamma in response to other stimuli, such as ionomycin plus PMA (Fig. 2d) or a crosslinking antibody specific for NK1.1 (Fig. 2e). These observations indicated that the block in cytokine production was specific to NKG2D-induced signals and did not reflect a general inability of transgenic NK cells to produce IFN-gamma. Several NK receptors were present in similar quantities on the surfaces of transgenic and wild-type NK cells (Supplementary Fig. 6 online). The percentages of transgenic NK cells expressing some of the NK activating and inhibitory receptors (Ly49D, KLRG1, NKG2A, NKG2C and NKG2E) were slightly altered relative to those of wild-type NK cells, but these changes were very subtle (Supplementary Fig. 6), further indicating the specificity of DAP10-Ub–mediated receptor degradation. Therefore, whereas transgenic NK cells showed impaired NKG2D-dependent function, consistent with downregulation of NKG2D from the cell surface, other functions of transgenic NK cells were not compromised, and the cells could receive signals through at least some NK cell receptors.

DAP10-Ub renders NK cells unresponsive to IL-15

Notably, steady-state transgenic mice had many fewer NK cells (Fig. 3a and Supplementary Table 1). We recovered 70–90% fewer NK and NKT cells from transgenic spleens and livers (Fig. 3a). Furthermore, we noted much lower NK cell numbers in all other tissues analyzed, including the thymus, lymph nodes and bone marrow, as well as in the peritoneal cavity (data not shown). Notably, steady-state numbers of NK and NKT cells are also much lower in mice lacking components of the IL-15 signaling pathway23, 24, 25, and IL-15 is essential for the maintenance and survival of NK cells26, 27. Such observations suggest that IL-15 signaling might be impaired in NK cells from DAP10-Ub–transgenic mice and prompted us to test the IL-15 responsiveness of transgenic NK cells. We cultured wild-type and transgenic splenocytes for 6 d in vitro in the presence or absence of IL-15 (Fig. 3b). We recovered wild-type NK cells only from cultures containing IL-15, reflecting the effects of IL-15 on the promotion of both proliferation, as measured by BrdU incorporation (Fig. 3c), and probably also survival, through the induction of key prosurvival factors and suppression of proapoptotic proteins27. In contrast, IL-15 promoted neither the survival nor the proliferation of transgenic NK cells (Fig. 3b,c and data not shown). In our LAK cell cultures, IL-2 was very inefficient in driving the production of NK cells from transgenic splenocytes, as we consistently recovered many fewer NK1.1+, T cell receptor-beta–negative (TCRbeta-) cells from these cultures (data not shown). This defect in IL-2 responsiveness is consistent with the lower IL-15 responsiveness of transgenic NK cells, as the high dose of IL-2 used for generating LAK cells is generally thought to be a substitute for IL-15.

Figure 3: IL-15 unresponsiveness of transgenic NK cells.

Figure 3 : IL-15 unresponsiveness of transgenic NK cells.

(a) Flow cytometry to quantify NK cells in spleens and livers from wild-type and transgenic mice. Numbers above outlined areas indicate percent NK cells among total gated lymphocytes and are representative of at least 40 mice (spleen) or 6 mice (liver) in at least three experiments. (b) Flow cytometry of the proliferation and survival of wild-type or transgenic splenocytes either isolated and stained immediately after isolation (top row) or cultured for 6 d in the presence (bottom row) or absence (middle row) of IL-15. Plots are gated on CD3epsilon- populations. Representative of three experiments. (c) Flow cytometry of BrdU incorporation to assess the IL-15-induced proliferation of the cells described in a, pulsed with BrdU 30 h after start of culture and collected 36 h later. Plots are gated on NK1.1+TCRbeta- populations. Data are from one representative mouse from each group of two in one experiment. (d) Quantitative RT-PCR of IL-15-dependent gene transcription in sorted wild-type and transgenic NK cells allowed to 'rest' in vitro for several hours and then stimulated for 5 h with IL-15. Expression of transcripts (normalized to that of the 'housekeeping' gene Hprt1) in transgenic cells is presented relative to expression in wild-type cells, arbitrarily set as 1. Data are the average of two independent experiments with NK cells pooled from two wild-type mice or ten transgenic mice per experiment.

Full size image (74 KB)

The data presented above indicated that transgenic NK cells cannot respond normally to IL-15. To further confirm that conclusion, we measured IL-15-induced gene expression in sorted wild-type or transgenic NK cells (Fig. 3d). Consistent with the defect in BrdU incorporation, IL-15-treated transgenic cells had fewer transcripts encoding cyclin D2 than did IL-15-treated wild-type cells. Transgenic NK cells also had smaller amounts of other IL-15-dependent transcripts, including those encoding DAP10 and NKG2D28. In contrast, CX3CR1 mRNA, shown before to be repressed by IL-15 stimulation29, was upregulated in transgenic NK cells relative to their wild-type counterparts (Fig. 3d). Finally, whereas wild-type NK cells upregulated granzyme B in response to stimulation with high-dose IL-2, transgenic NK cells failed to do so (Supplementary Fig. 7 online). These results collectively indicate that transgenic NK cells do not respond to IL-15 and that this IL-15 insensitivity may constitute the underlying basis for the paucity of NK cells in transgenic mice.

We next considered the possibility that transgenic NK cells may fail to respond to IL-15 because of lower expression of the IL-15 receptor (IL-15R) complex. However, the IL-15R alpha-chain and beta-chain (CD122) and IL-15R gamma-chain (CD132) were expressed in similar amounts on the surfaces of wild-type and transgenic NK cells (Fig. 4a). These data suggested that transgenic NK cells might fail to transduce intracellular IL-15 signals and that DAP10 and its associated proteins may promote signaling by IL-15R. We therefore looked for an association between DAP10 and the IL-15R complex. To do this, we stably expressed Flag-DAP10 in the NK cell line KY-1 (ref. 30); expression of DAP10 in these cells was twofold higher than that in the parental cells (data not shown). In these cells, Flag-DAP10 associated with the gamma- and beta-chains of IL-15R, as shown by coimmunoprecipitation (Fig. 4b). We also analyzed interactions between endogenous DAP-10 and IL-15R proteins in LAK cells. DAP10 associated with NKG2D as well as with the p85 subunit of phosphatidylinositol-3-OH kinase, as shown before14 (Fig. 4c). Notably, endogenous DAP10 bound to the gamma- and beta-chains of IL-15R in LAK cells (Fig. 4c,d). Furthermore, the association of DAP10 with those chains was specific, as DAP10 did not coimmunoprecipitate other cell surface receptors such as 2B4 (Fig. 4c,d).

Figure 4: Impaired STAT5 phosphorylation in transgenic NK cells.

Figure 4 : Impaired STAT5 phosphorylation in transgenic NK cells.

(a) Cell surface expression of IL-15R components on wild-type and transgenic NK cells. Ictrl, isotope control. Data are gated on NK1.1+TCRbeta- populations and are representative of at least four independent experiments. (b) Analysis of the association of DAP10 with IL-15R in KY-1 cells left untransduced (Parental) or stably transduced to express Flag-DAP10; lysates were immunoprecipitated (IP) with anti-Flag and analyzed by immunoblot (IB) with anti-CD132 and anti-CD122. Representative of more than three experiments. (c,d) Analysis of the association of endogenous DAP10 with CD132 (c) and CD122 (d) in LAK cells stimulated for 2 min with 0.05 mM pervanadate (c) or left untreated (d); lysates were immunoprecipitated with various antibodies (above lanes) and analyzed by immunoblot (antibodies, left margins). NRS, nonimmune rabbit serum; As, antiserum. Representative of two independent experiments. (e) Flow cytometry of IL-15-induced STAT5 phosphorylation (p-STAT5) in wild-type or transgenic splenocytes cultured for 24 h in vitro before stimulation with IL-15 for 0 min (red), 7 min (blue) or 15 min (green), assessed in gated NK1.1+TCRbeta- (top) or CD8+ (bottom) populations. Representative of two experiments.

Full size image (52 KB)

The association of DAP10 with the IL-15R complex was suggestive of involvement of DAP10 in IL-15R signaling and prompted us to look at 'downstream' signaling events in IL-15-treated wild-type and transgenic NK cells. Studies have shown that IL-15 stimulation induces the association of the IL-15R gamma- and beta-chains and activation of Jak3 and the consequent phosphorylation of 'downstream' targets such as the transcription factor STAT5, which is essential for the transmission of IL-15 signals31. Therefore, we measured STAT5 phosphorylation induced by IL-15 stimulation in wild-type and transgenic NK cells32. IL-15 triggered robust STAT5 phosphorylation in wild-type but not transgenic NK cells (Fig. 4e). In contrast, transgenic CD8+ T cells responded normally to IL-15 (Fig. 4e), even though DAP10-Ub was expressed in CD8+ T cells (Supplementary Fig. 4). These findings suggested that coupling of IL-15R signaling to other signaling systems may be cell type specific, with an obligate function for DAP10-associated proteins in NK cells but not for the responsiveness of CD8+ T cells to IL-15.

The IL-15-unresponsive phenotype of transgenic NK cells was not due to IL-15R downregulation, and expression of Jak3 and STAT5 was similar in wild-type and transgenic NK cells, as assessed by intracellular staining (Supplementary Fig. 8 online). Hence, the precise molecular basis of the cell type–specific defect in IL-15-induced STAT5 phosphorylation remains to be determined. However, it does not seem to involve downregulation of any of the components of the classical IL-15R signaling pathway in transgenic NK cells.

Priming of NKG2D signaling by Jak3

In addition to its involvement in NK cell survival and maintenance, IL-15 has also been suggested to promote NK cell effector functions in general and NKG2D-dependent activities in particular. NK cell activation in vivo is achieved by injection of poly(I:C), which induces IL-15 through type I interferons33, and priming of NK cell killing by dendritic cells is mediated by IL-15 cross-presentation34. Moreover, NKG2D-dependent cytotoxicity is much lower in IL-15-unresponsive NK cells35, 36, and IL-15 potentiates NKG2D-dependent cytotoxicity in celiac disease28. Therefore, we hypothesized that the association between DAP10 and IL-15R is involved not only in IL-15R signaling but also in the priming of DAP10 for NKG2D-mediated signal transduction. Specifically, because the kinase responsible for phosphorylation of the critical tyrosine residue in DAP10 has not yet been identified, we considered that Jak3, the kinase activated by IL-15R signaling, might exert this function. To test this hypothesis, we pretreated the NK cell line expressing Flag-DAP10 (Fig. 5a) or LAK cells (data not shown) with either of two Jak3 inhibitors or with dimethyl sulfoxide (vehicle control) and then stimulated the treated cells with pervanadate, a general tyrosine phosphatase inhibitor. Consistent with published studies, pervanandate stimulation induced rapid DAP10 phosphorylation14. However, the Jak3 inhibitors WHI-P131 and WHI-P154 specifically suppressed the DAP10 phosphorylation noted in the presence of pervanadate (Fig. 5a and data not shown). The concentrations of the inhibitors used here have been shown before to have no nonspecific effects on closely related kinases37; moreover, the concentrations required for the inhibition of DAP10 consistently paralleled the concentrations needed to block the phosphorylation of STAT5, an established Jak3 substrate (Fig. 5a). Therefore, at least in this context, Jak3 activity was required for DAP10 signaling. We further demonstrated a direct function for Jak3 in activating DAP10 with an in vitro kinase assay (Fig. 5b). Wild-type Jak3 but not a catalytically inactive form of Jak3 phosphorylated recombinant glutathione S-transferase–linked DAP10 (GST-DAP10). Notably, substitution of the tyrosine residue in the YXNM motif of DAP10 with phenylalanine abolished Jak3-mediated DAP10 phosphorylation (Fig. 5b). These results suggest that Jak3 is a true DAP10 kinase in vitro and in vivo and that Jak3-mediated tyrosine phosphorylation of DAP10 is critical for NKG2D signaling.

Figure 5: Priming of DAP10 signaling by IL-15-activated Jak3.

Figure 5 : Priming of DAP10 signaling by IL-15-activated Jak3.

(a) Immunoassay of KY-1 cells stably transfected with Flag-DAP10, then pretreated with either of two Jak3 inhibitors or with dimethyl sulfoxide (DMSO) before being stimulated with pervanadate; lysates were immunoprecipitated with anti-Flag, followed by immunoblot analysis of DAP10 phosphorylation with an antibody specific for phosphorylated tyrosine (Phosphotyrosine). Representative of at least three experiments. (b) Immunoblot analysis of the incorporation of radiolabeled phosphate (Phospho-DAP10) by immunocomplexes of wild-type Jak3 (WT Jak3) or catalytically inactive Jak3 (KD Jak3) incubated with recombinant GST-DAP10 (DAP10 WT), or a GST-linked DAP10 mutant with substitution of phenylalanine for tyrosine in the YXNM motif (DAP10 YF), in the presence of radiolabeled gamma-phosphate (top). Blots below show the amount of Jak3 (middle) or DAP10 (bottom) in each sample. Representative of at least two experiments. (c) DAP10-mediated killing by LAK cells deprived of IL-2 overnight and then cultured for 1 h together with MIG-RMA or Rae1-RMA cells in the presence or absence of IL-15, assessed as degranulation, measured by staining for cell surface exposure of Lamp1 (CD107a). + cyclohex (middle row), addition of cyclohexamide at time of coculture; + Jak3 inhib, pretreatment with the Jak3 inhibitor WHI-154 before IL-15 stimulation. Numbers above outlined areas indicate percent CD107a+ cells. SSC, side scatter. Plots are gated on LAK cells and are representative of two experiments.

Full size image (54 KB)

IL-15 has pleiotropic effects on NK cells, regulating their survival as well as general NK cell–mediated cytotoxicity by upregulating components of the cytotoxic machinery. Thus, consistent with their IL-15 unresponsiveness, transgenic NK cells were impaired in their NKG2D-independent killing of the major histocompatibility complex class II–deficient T cell lymphoma line RMA/s (Supplementary Fig. 9 online) and upregulation of granzyme B (Supplementary Fig. 7). In addition to these IL-15-dependent defects, our results indicating Jak3 is a DAP10 kinase suggested that IL-15-mediated activation of Jak3 would be necessary to prime DAP10 signaling. To test this, we assessed NKG2D-induced DAP10-dependent NK cell degranulation in the presence or absence of IL-15. We also measured degranulation of LAK cells within a 1-hour period after culture together with Rae1-expressing target cells38, with or without the protein synthesis inhibitor cyclohexamide, to rule out the contribution of the transcription-dependent effects of IL-15 on NK cell cytotoxicity. This experimental system enabled us to address the effect of 'lateral' transcription-independent signaling by IL-15-activated Jak3 on DAP10 in NK cell priming. Treatment with IL-15 enhanced NKG2D-induced LAK cell degranulation, as indicated by upregulation of CD107a on the cell surface, independently of new protein synthesis (Fig. 5c). Pretreatment of LAK cells with a Jak3 inhibitor before IL-15 stimulation completely abolished the effects of IL-15 on DAP10-mediated signaling, further supporting the idea of a critical function for Jak3 as a physiologically relevant DAP10 kinase. The ability of the Jak3 inhibitor to block killing in all samples also suggested that the low degree of degranulation in cells not treated with exogenous IL-15 was due to residual IL-15R signaling from the IL-2 used in the LAK cell cultures. Our results collectively indicate that IL-15, by inducing Jak3-dependent DAP10 phosphorylation, is necessary for optimal NKG2D-DAP10 signaling (Supplementary Fig. 10 online).

Top

Discussion

Here we developed a new mouse model in which overexpression of a DAP10-Ub fusion protein substantially reduced cell surface expression of NKG2D. As a result, NK cells from these mice had defects in NKG2D-dependent cytotoxicity and IFN-gamma production but retained normal responsiveness to other stimuli. Studies of yeast have used a monoubiquitin fusion protein to target cell surface proteins for constitutive internalization17. We built on that strategy, using a ubiquitin fusion protein to target multisubunit receptor complexes for downregulation from the plasma membrane. Our results here suggest that this strategy could be used more generally to eliminate multisubunit receptor complexes that may not be easily targeted by more conventional methods.

We believe that the DAP10-Ub fusion protein works in a dominant way for two reasons. First, cell surface expression of NKG2D was almost completely ablated. This finding contrasts with those obtained with NK cells lacking either DAP10 (ref. 10) or DAP12 (refs. 11,12), which have substantial surface expression of NKG2D, presumably because of compensatory association of NKG2D with the remaining adaptor protein. Second, DAP10-Ub–transgenic mice showed a paucity of NK cells; in contrast, mice lacking DAP10 (ref. 10), DAP12 (refs. 11,12) or NKG2D (D. Raulet, personal communication), and even mice lacking functional DAP12, CD3zeta and FcRgamma, have wild-type numbers of NK cells13. The normal NK cell development in these gene-deletion models is postulated to be a consequence of redundancy among NK cell signaling adaptors, and the phenotype of DAP10-Ub–transgenic mice described here further emphasizes that redundancy. Specifically, our results suggest that additional signaling proteins must be eliminated in DAP10-Ub–transgenic mice to account for their considerable NK cell deficiency and to explain the differences in the phenotypes of DAP10-Ub–transgenic and DAP10-deficient mice. Biochemical studies to identify these DAP10-associated proteins should prove useful and may demonstrate previously unidentified signaling adaptors involved in IL-15 signaling.

Notably, DAP10-Ub–transgenic mice also had fewer B cells, suggesting a perturbation of B cell development and/or maintenance. We believe that the paucity of B cells is due to a cell-autonomous function of the transgene in this cell type, because DAP10-Ub was expressed in immature transgenic B cells and we detected small amounts of endogenous DAP10 mRNA in wild-type immature B cells (data not shown). As B cells do not express NKG2D, the identity of the DAP10-associated receptor in B cells remains to be determined, but our results suggest involvement of DAP10 and DAP10-associated proteins in B cell development and/or maintenance. As with NK cells, the difference between DAP10-deficient mice (which have a normal B cell compartment) and DAP10-Ub–transgenic mice in terms of steady-state B cell numbers confirmed the ability of DAP10-Ub to act in a dominant way to reveal a latent phenotype.

Many IL-15-dependent functions were dysregulated in DAP10-Ub–transgenic NK cells. Given the crucial function of IL-15 in the development, maintenance, survival and proliferation of NK cells, any or all of these factors may contribute to the paucity of NK cells in transgenic mice. These results indicate involvement of DAP10 and/or its associated protein(s) in IL-15 signaling and may explain why NKG2D signaling has been shown to lead to STAT5 activation39.

We have also demonstrated a previously unappreciated coupling of DAP10 and IL-15R signaling in NK cells. We found that DAP10 interacted with the IL-15R complex in an NK cell line and in LAK cells. Notably, another cell surface receptor, 2B4, did not associate with DAP10, and DAP12 did not immunoprecipitate together with IL-15R in KY-1 cells stably transduced with Flag-DAP12 (data not shown). This suggests that the interaction of DAP10 with the IL-15R is specific. Notably, despite the association of DAP10 with IL-15R in LAK cells and KY-1 cells, the IL-15R complex was expressed in wild-type quantities on the surfaces of transgenic NK cells and thus apparently was not targeted for downregulation by DAP10-Ub. This is most likely because the association of DAP10 with IL-15R is indirect, transient and signal dependent. We attempted to show stimulus-dependent association between DAP10 and IL-15R by depriving LAK cells of IL-2 before treating them with IL-15, but this experiment proved difficult to interpret because of downregulation of DAP10 in IL-2-deprived cells and loss of cell viability after prolonged IL-2 deprivation.

Perhaps most notably, we found that activation of IL-15R was required for NKG2D-mediated signal transduction and cytotoxicity. Published studies have shown that a phosphorylated-tyrosine motif in the cytoplasmic tail of DAP10 is essential for the recruitment of 'downstream' signaling components, but, notably, the kinase responsible for this activity has not been identified. Here we demonstrated that Jak3 phosphorylates this key tyrosine residue in the YXNM motif of DAP10 and is thus important for DAP10-mediated NK cell function. It is well established that NK cell activation requires culture in vitro with IL-15 or high doses of IL-2 (as a cheaper substitute for IL-15) or priming in vivo by poly(I:C), which induces IL-15 by eliciting production of type 1 interferon. Moreover, a study has demonstrated that efficient NK cell killing in vivo requires priming by IL-15-producing dendritic cells34. Our results provide a mechanistic basis for such observations. Although it is possible that constitutively active kinases (such as Src family kinases) can phosphorylate DAP10 to enable basal signaling, our data suggest that optimal triggering of NKG2D-dependent functions occurs in the context of infection and IL-15 production, when DAP10 is activated by IL-15 to allow signaling through NKG2D. Thus, our results suggest that NK cell cytotoxic activity is 'primed' by IL-15 produced during viral infection. Therefore, we propose a model in which IL-15 may have a permissive role for cytotoxic functions, restricting these activities to conditions of viral and intracellular bacterial infection. Furthermore, because of the intimate coupling of these two receptor signaling systems, IL-15 may direct the most efficient NK cell–mediated killing to IL-15-producing infected target cells, thus achieving specificity and preventing immunopathology. Indeed, it is possible that all NK cell activating receptors would require prior priming, by IL-15 or other cytokines, for efficient signaling. Our results emphasize the need for inflammatory signals to boost NK cell functions that may otherwise be insufficient in certain settings (for example, to eradicate some tumors).

Finally, we note that activating receptors (that signal through ITAM or ITAM-like motifs) are expressed on many cells other than NK cells, including those found in nonhematopoietic tissues40. Given that these receptors recognize diverse ligands that induce distinct biological signals, it is somewhat unexpected that these receptors should use a common, ITAM-based signaling pathway. We propose that activating receptors couple with distinct cell type–specific receptors in different contexts and that synergy between the two receptor systems serves to deliver unique signals. Consistent with that idea, CD8+ T cells from transgenic mice seemed to have no defects in IL-15 responsiveness, suggesting that naive CD8+ T cells and NK cells use distinct pathways for potentiating IL-15 signal transduction. These observations further emphasize the cell type–specific differences in the biological functions of activating receptors and signaling adaptors. Indeed, other studies have reported ligand-specific, cell type–specific and function-specific cooperation between activating receptors and other receptor systems41, 42, 43. With their modular nature, activating receptors may be readily co-opted by other receptor systems to provide a flexible, highly evolvable mechanism for signal integration in diverse physiological contexts.

Top

Methods

Mice.

Sequence encoding DAP10-Ub was subcloned into the VACD2 vector under control of the human CD2 promoter. The vector was injected into fertilized (C57BL/6 times SJL) F2 oocytes by Yale Animal Genomics Services. Most experiments used transgenic mice and wild-type littermate controls backcrossed three times to the C57BL/6 strain. The defects in NK development and IL-15 unresponsiveness were confirmed in transgenic mice backcrossed seven times to C57BL/6 mice. Animals were bred and maintained at the Yale Animal Resources Center at Yale University and all animal experiments were done with approval by and in accordance with regulatory guidelines and standards set by the Institutional Animal Care and Use Committee of Yale University. Perforin-deficient mice were from Jackson Laboratories.

In vitro killing assay.

Wild-type or transgenic LAK cells, or RMA cells transduced with control retrovirus or retrovirus encoding Rae1, were used. Target cells were plated at a density of 2 times 104 cells per well and effector cells were 'titrated' in triplicate. Then, 4 h later, culture supernatants were analyzed for release of cytoplasmic lactate dehydrogenase as a measure of cytotoxicity with the CytoTox Non-Radioactive Cytotoxicity Assay (Promega). Because the generation of LAK cells in response to high-dose IL-2, a substitute for IL-15, is impaired in transgenic cells, an aliquot of effector cells was analyzed by flow cytometry to determine the relative percentages of NK1.1+CD3epsilon- cells in wild-type and transgenic LAK cell cultures. The effector/target ratio reflects the number of NK1.1+CD3epsilon- wild-type or transgenic effector cells and thus corrects for the difference in NK cell numbers in wild-type versus transgenic LAK cell cultures. Cytotoxicity was calculated as suggested by the manufacturer according to the following formula: percent specific lysis = 100 times (experimental lysis – effector spontaneous lysis – target spontaneous lysis) / (target maximum lysis – target spontaneous lysis).

Immunoprecipitation and immunoblot analysis.

KY-1 cells and LAK cells were washed once with PBS and were lysed in lysis buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 1% (vol/vol) Triton X-100, 5 mM EDTA and 10% (vol/vol) glycerol) supplemented with protease inhibitors and phosphatase inhibitors. Cleared lysates were incubated for 4 h at 4 °C with anti-Flag M2 agarose beads (Sigma), with DAP10-specific antiserum, or with various antibodies, followed by incubation with protein A agarose beads (Invitrogen), then they were washed, resolved by SDS-PAGE and transferred onto polyvinyldifluoride membranes. Antibodies for immunoblot analysis were to NKG2D (V-14), p85 (Z-8), DAP10 (M-19 and F4-93), CD132 (M-20) and CD122 (C-20; all from Santa Cruz Biotechnology); phosphorylated tyrosine (4G10) and Jak3 (AB3807; both from Upstate Cell Signaling Solutions); phosphorylated STAT5 (9351; Cell Signaling Technology); and 2B4 (eBio244F4; eBiosciences).

In vitro kinase assay.

The 293T cells were transfected with wild-type or catalytically inactive Jak3, then 24 h later, Jak3-containing complexes were isolated by immunoprecipitation with antibody to Jak3 (anti-Jak3). Washed immunocomplexes were resuspended in kinase assay buffer (50 mM Tris, pH 7.5, 5 mM MgCl2, 5 mM MnCl2 and 1 muM ATP) in the presence of 5 muCi [gamma-32P]ATP and 5 mug of either GST-DAP10 or a GST-linked DAP10 mutant with substitution of phenylalanine for tyrosine in the YXNM motif. The kinase reaction was incubated on ice for 10 min, was stopped by the addition of 100 mM EDTA, and then was incubated with glutathione sepharose beads (Amersham). Beads were washed and were resolved by SDS-PAGE, followed by transfer to polyvinyldifluoride membranes and exposure to autoradiography film. The radioactivity in membranes was allowed to decay before immunoblot analysis.

Pervanadate stimulation of KY-1 stable transfectants.

Pervanadate was prepared by mixture for 5 min of 1 ml of 20 mM sodium orthovanadate with 330 mul of 30% (vol/vol) H2O2. For stimulation, cells were resuspended in PBS at a density of 30 times 106 cells per ml per sample and then were pretreated for 20–30 min with dimethyl sulfoxide or various concentrations of Jak3 inhibitor I or II (Calbiochem). Pervanadate was then added to the cells for 10 min at 37 °C at a dilution of 1:60 before reactions were stopped with ice-cold PBS and cells were lysed for immunoprecipitation as described above.

Degranulation assay.

Wild-type LAK cells deprived of IL-2 were incubated at effector/target ratio of 5:1 with MIG-RMA or Rae1-RMA cells (labeled with the red fluorescent dye PKH26) in the presence or absence of IL-15 (100 ng/ml). For some samples, LAK cells were pretreated for 30 min with Jak3 inhibitor II (100 muM); for others, cyclohexamide was added at the time of coculture. For maximum detection of CD107a, anti-CD107a (final dilution, 1:100) was added at time of coculture. After 1 h, cells were collected and stained and were analyzed by flow cytometry for expression of CD107a on PKH26-NK1.1+TCRbeta- populations.

Additional methods.

Information on cell lines and ex vivo cell culture, flow cytometry, transient transfection, in vivo killing assays, cellular stimulation for IFN-gamma production and quantitative RT-PCR is available in the Supplementary Methods online.

Note: Supplementary information is available on the Nature Immunology website.



Top

Acknowledgments

We thank D. Raulet (University of California, Berkeley) for DAP10-specific antiserum and communication of unpublished results; W. Yokoyama for the KY-1 cells; S. Holley and C. Anicelli for help with mouse work; G. Tokmoulina and T. Taylor for assistance with cell sorting; and S. Shih and members of the lab for discussions and for critically reviewing the manuscript. VACD2 vector was provided by D. Kioussis (National Institute for Medical Research, London, UK), and wild-type and catalytically inactive Jak3 were from J. Ihle (St. Jude Children's Research Hospital). Supported by the Howard Hughes Medical Institute and the National Institutes of Health (AI-0466888 and AI-055502).

Received 16 January 2007; Accepted 18 September 2007; Published online 21 October 2007.

Top

References

  1. Biron, C.A., Nguyen, K.B., Pien, G.C., Cousens, L.P. & Salazar-Mather, T.P. Natural killer cells in antiviral defense: function and regulation by innate cytokines. Annu. Rev. Immunol. 17, 189–220 (1999). | Article | PubMed | ISI | ChemPort |
  2. Smyth, M.J., Hayakawa, Y., Takeda, K. & Yagita, H. New aspects of natural-killer-cell surveillance and therapy of cancer. Nat. Rev. Cancer 2, 850–861 (2002). | Article | PubMed | ISI | ChemPort |
  3. Lanier, L.L. NK cell recognition. Annu. Rev. Immunol. 23, 225–274 (2005). | Article | PubMed | ISI | ChemPort |
  4. Raulet, D.H. Interplay of natural killer cells and their receptors with the adaptive immune response. Nat. Immunol. 5, 996–1002 (2004). | Article | PubMed | ISI | ChemPort |
  5. Yokoyama, W.M., Kim, S. & French, A.R. The dynamic life of natural killer cells. Annu. Rev. Immunol. 22, 405–429 (2004). | Article | PubMed | ISI | ChemPort |
  6. Kim, S. et al. Licensing of natural killer cells by host major histocompatibility complex class I molecules. Nature 436, 709–713 (2005). | Article | PubMed | ISI | ChemPort |
  7. Fernandez, N.C. et al. A subset of natural killer cells achieves self-tolerance without expressing inhibitory receptors specific for self-MHC molecules. Blood 105, 4416–4423 (2005). | Article | PubMed | ISI | ChemPort |
  8. Lanier, L.L. Natural killer cell receptor signaling. Curr. Opin. Immunol. 15, 308–314 (2003). | Article | PubMed | ISI | ChemPort |
  9. Colucci, F. et al. Natural cytotoxicity uncoupled from the Syk and ZAP-70 intracellular kinases. Nat. Immunol. 3, 288–294 (2002). | Article | PubMed | ISI | ChemPort |
  10. Gilfillan, S., Ho, E.L., Cella, M., Yokoyama, W.M. & Colonna, M. NKG2D recruits two distinct adaptors to trigger NK cell activation and costimulation. Nat. Immunol. 3, 1150–1155 (2002). | Article | PubMed | ISI | ChemPort |
  11. Bakker, A.B. et al. DAP12-deficient mice fail to develop autoimmunity due to impaired antigen priming. Immunity 13, 345–353 (2000). | Article | PubMed | ISI | ChemPort |
  12. Tomasello, E. et al. Combined natural killer cell and dendritic cell functional deficiency in KARAP/DAP12 loss-of-function mutant mice. Immunity 13, 355–364 (2000). | Article | PubMed | ISI | ChemPort |
  13. Chiesa, S. et al. Multiplicity and plasticity of natural killer cell signaling pathways. Blood 107, 2364–2372 (2006). | Article | PubMed | ISI | ChemPort |
  14. Wu, J. et al. An activating immunoreceptor complex formed by NKG2D and DAP10. Science 285, 730–732 (1999). | Article | PubMed | ISI | ChemPort |
  15. Jamieson, A.M. et al. The role of the NKG2D immunoreceptor in immune cell activation and natural killing. Immunity 17, 19–29 (2002). | Article | PubMed | ISI | ChemPort |
  16. Raulet, D.H. Roles of the NKG2D immunoreceptor and its ligands. Nat. Rev. Immunol. 3, 781–790 (2003). | Article | PubMed | ISI | ChemPort |
  17. Roth, A.F. & Davis, N.G. Ubiquitination of the PEST-like endocytosis signal of the yeast a-factor receptor. J. Biol. Chem. 275, 8143–8153 (2000). | Article | PubMed | ISI | ChemPort |
  18. Garrity, D., Call, M.E., Feng, J. & Wucherpfennig, K.W. The activating NKG2D receptor assembles in the membrane with two signaling dimers into a hexameric structure. Proc. Natl. Acad. Sci. USA 102, 7641–7646 (2005). | Article | PubMed | ChemPort |
  19. Rosen, D.B. et al. A Structural basis for the association of DAP12 with mouse, but not human, NKG2D. J. Immunol. 173, 2470–2478 (2004). | PubMed | ISI | ChemPort |
  20. Hicke, L. & Dunn, R. Regulation of membrane protein transport by ubiquitin and ubiquitin-binding proteins. Annu. Rev. Cell Dev. Biol. 19, 141–172 (2003). | Article | PubMed | ISI | ChemPort |
  21. Diefenbach, A. et al. Selective associations with signaling proteins determine stimulatory versus costimulatory activity of NKG2D. Nat. Immunol. 3, 1142–1149 (2002). | Article | PubMed | ISI | ChemPort |
  22. Hayakawa, Y. et al. Cutting edge: tumor rejection mediated by NKG2D receptor-ligand interaction is dependent upon perforin. J. Immunol. 169, 5377–5381 (2002). | PubMed | ISI | ChemPort |
  23. Kennedy, M.K. et al. Reversible defects in natural killer and memory CD8 T cell lineages in interleukin 15-deficient mice. J. Exp. Med. 191, 771–780 (2000). | Article | PubMed | ISI | ChemPort |
  24. Lodolce, J.P. et al. IL-15 receptor maintains lymphoid homeostasis by supporting lymphocyte homing and proliferation. Immunity 9, 669–676 (1998). | Article | PubMed | ISI | ChemPort |
  25. Matsuda, J.L. et al. Homeostasis of Valpha14i NKT cells. Nat. Immunol. 3, 966–974 (2002). | Article | PubMed | ISI | ChemPort |
  26. Cooper, M.A. et al. In vivo evidence for a dependence on interleukin 15 for survival of natural killer cells. Blood 100, 3633–3638 (2002). | Article | PubMed | ISI | ChemPort |
  27. Huntington, N.D. et al. Interleukin 15-mediated survival of natural killer cells is determined by interactions among Bim, Noxa and Mcl-1. Nat. Immunol. 8, 856–863 (2007). | Article | PubMed | ISI | ChemPort |
  28. Meresse, B. et al. Coordinated induction by IL15 of a TCR-independent NKG2D signaling pathway converts CTL into lymphokine-activated killer cells in celiac disease. Immunity 21, 357–366 (2004). | Article | PubMed | ISI | ChemPort |
  29. Sechler, J.M., Barlic, J., Grivel, J.C. & Murphy, P.M. IL-15 alters expression and function of the chemokine receptor CX3CR1 in human NK cells. Cell. Immunol. 230, 99–108 (2004). | Article | PubMed | ISI | ChemPort |
  30. Karlhofer, F.M., Orihuela, M.M. & Yokoyama, W.M. Ly-49-independent natural killer (NK) cell specificity revealed by NK cell clones derived from p53-deficient mice. J. Exp. Med. 181, 1785–1795 (1995). | Article | PubMed | ISI | ChemPort |
  31. Kovanen, P.E. & Leonard, W.J. Cytokines and immunodeficiency diseases: critical roles of the gammac-dependent cytokines interleukins 2, 4, 7, 9, 15, and 21, and their signaling pathways. Immunol. Rev. 202, 67–83 (2004). | Article | PubMed | ISI | ChemPort |
  32. Krutzik, P.O. & Nolan, G.P. Intracellular phospho-protein staining techniques for flow cytometry: monitoring single cell signaling events. Cytometry A 55, 61–70 (2003). | Article | PubMed | ChemPort |
  33. Zhang, X., Sun, S., Hwang, I., Tough, D.F. & Sprent, J. Potent and selective stimulation of memory-phenotype CD8+ T cells in vivo by IL-15. Immunity 8, 591–599 (1998). | Article | PubMed | ISI | ChemPort |
  34. Lucas, M., Schachterle, W., Oberle, K., Aichele, P. & Diefenbach, A. Dendritic cells prime natural killer cells by trans-presenting interleukin 15. Immunity 26, 503–517 (2007). | Article | PubMed | ISI | ChemPort |
  35. Minagawa, M. et al. Enforced expression of Bcl-2 restores the number of NK cells, but does not rescue the impaired development of NKT cells or intraepithelial lymphocytes, in IL-2/IL-15 receptor beta-chain-deficient mice. J. Immunol. 169, 4153–4160 (2002). | PubMed | ISI | ChemPort |
  36. Imada, K. et al. Stat5b is essential for natural killer cell-mediated proliferation and cytolytic activity. J. Exp. Med. 188, 2067–2074 (1998). | Article | PubMed | ISI | ChemPort |
  37. Sudbeck, E.A. et al. Structure-based design of specific inhibitors of Janus kinase 3 as apoptosis-inducing antileukemic agents. Clin. Cancer Res. 5, 1569–1582 (1999). | PubMed | ISI | ChemPort |
  38. B