Nature Immunology
3, 1119 - 1120 (2002)
doi:10.1038/ni1202-1119
Versatile signaling through NKG2DEric O. LongLaboratory of Immunogenetics, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Rockville, MD 20852, USA. eLong@nih.gov Receptors in the immune system usually specialize in transmitting specific types of signals. However, by associating with two different signaling subunits, the receptor NKG2D can transmit distinct signals for either costimulation or full-fledged activation.The signaling capability of many receptors in the immune system is controlled by their association with partner chains that carry specific sequence motifs for the recruitment of signaling molecules. Some of the activating receptors associate with subunits that carry immunoreceptor tyrosine-based activation motifs (ITAMs) in their cytoplasmic tail. These include subunits of the T cell receptor (TCR)-associated CD3 complex, immunoglobulin  (Ig) and Ig subunits of the B cell receptor, the  chain of Fc receptors and of several natural killer (NK) cell receptors and DAP12, which associates with several receptors on NK cells and myeloid cells. Tyrosine-phosphorylated ITAMs deliver activation signals by recruiting the tyrosine kinases Syk or ZAP-70. A different tyrosine motif (YxxM) is found on receptors that provide costimulation, such as CD28 in T cells and the DAP10 partner chain of the receptor NKG2D1. Tyrosine-phosphorylated DAP10 may signal for costimulation by binding to the p85 subunit of phosphatidylinositol-3 kinase and to Grb21,
2. The delivery of primary and accessory signals from receptors on immune cells must be carefully balanced not only to initiate cellular responses but also to control them by fine-tuning the extent of activation, cell death or inhibition. It came as a surprise, therefore, when Raulet and colleagues3 and Colonna and colleagues4 in this issue of Nature Immunology showed that the receptor NKG2D can provide both types of signals: costimulation through YxxM-containing DAP10 and activation through ITAM-containing DAP12.
NKG2D is an important activation receptor expressed on NK cells, CD8+ T cells,  T cells and macrophages and is triggered by its ligands on cells that have been modified by infections, transformation or stress5,
6,
7. Several ligands for NKG2D have been identified. Mouse NKG2D binds with high affinity to H-60, to murine ULBP-like transcript 1 (MULT1) and to members of the Rae1 family of molecules6,
8. MULT is expressed ubiquitously, whereas expression of H-60 and Rae1 is found on some tumor and infected cells. On the other hand, human NKG2D binds to MICA and MICB, which are induced by stress signals, and to UL16-binding proteins (ULBP). Activation of NKG2D on NK cells results in a strong signal that either bypasses or overrides the inhibitory signal from major histocompatibility complex (MHC) class I−specific receptors, with the consequent target cell killing. In contrast, engagement of NKG2D on T cells costimulates effector T cell function. How NKG2D functions to stimulate both innate and adaptive lymphocyte responses has recently become the subject of intense studies. Adding further intensity and interest to the topic, the studies by Raulet and colleagues3 and Colonna and colleagues4 converge and come to the same startling conclusion: NKG2D can deliver different signals depending on its association with signaling subunits. In their simplest interpretation, their results indicate that NKG2D provides only a costimulatory signal in T cells, but can provide both activating and costimulatory signals in innate immune cells such as NK cells and macrophages3,
4. This versatility is achieved by the selective expression of DAP10 and DAP12 on different cells and by their distinct association with alternative isoforms of NKG2D3. In addition, the multiple levels of regulation suggest an even greater complexity in NKG2D function.
Using complementary approaches, Raulet and colleagues3 and Colonna and colleagues4 showed that NKG2D function could not be explained by its association with DAP10 alone. Mice with a deletion of the gene encoding DAP10 lacked NKG2D expression and function in T cells, but expressed NKG2D on NK cells and could mount NK cell responses through NKG2D4. Conversely, mice with a mutation in the gene encoding DAP12 showed normal T cell, but impaired NK cell, responses through NKG2D3. In addition, transcription of the gene encoding mouse NKG2D yielded two alternatively spliced products encoding NKG2D with a longer (NKG2D-L) and a shorter (NKG2D-S) cytoplasmic tail3. NKG2D-L binds only DAP10, as reported previously with the human NKG2D receptor9. In contrast, Raulet's group found that NKG2D-S, which is 13 amino acids shorter at the cytoplasmic tail, binds both DAP10 and DAP123. Whether NKG2D-S provides a specific binding site for DAP12 or whether NKG2D-L interferes sterically with binding of DAP12 is not known. As the tail of DAP12 is 48 amino acids long, whereas that of DAP10 is only 23 amino acids long, it is possible that the short DAP10 tail can be more easily accommodated in a complex with NKG2D-L (Fig. 1). Thus, the pairing of NKG2D with its partner chains is regulated at the level of expression of DAP10 and DAP12 and of NKG2D-L and NKG2D-S isoforms. As a result of this complexity, engagement of NKG2D results in different outcomes in T cells, macrophages and NK cells (Fig. 1).
 | | |
The role of a costimulatory signal in T cells is clear. It enhances and sustains signals from the TCR and protects naïve cells from activation-induced cell death. By expressing only DAP10, T cells ensure that NKG2D can deliver only a costimulatory signal. Indeed, NKG2D can substitute for CD28 in cytomegalovirus-specific T cells5. The induction of both NKG2D isoforms in activated T cells seems at first unnecessary, as either one alone could signal through DAP10. However, in the hypothetical event that the cytoplasmic tail of NKG2D itself has signaling capability, having both isoforms would provide the potential for additional signals. The absence of ITAM signals via NKG2D in T cells is an important safety measure, as T cell activation must be ultimately controlled by the TCR.
Activated macrophages preferentially express the NKG2D-S isoform, which can associate with both DAP10 and DAP12. DAP12 plays a role in macrophage activation through NKG2D, as Raulet and colleagues found that nitric oxide production was defective in macrophages from DAP12-mutant mice3. The role played by DAP10-mediated signals in macrophages has not been explored yet. As DAP10 and DAP12 compete for association with NKG2D-S, it is unclear how much signaling occurs through the NKG2D-S−DAP10 complex in macrophages. DAP12 also associates with other receptors in myeloid cells10, including myeloid DAP12-associating lectin 1 (MDL-1), signal-regulatory protein 1 (SIRP-1) and members of the triggering receptor expressed on myeloid cells (TREM) family. In resting macrophages, which do not express NKG2D, DAP12 may well have signaling functions in the context of those other receptors.
The concept of costimulation as an amplifier of TCR-mediated signaling in T cells is not directly applicable to NK cells. Activation of NK cells is not controlled by a single essential receptor but, rather, by the summation of signals coming from a number of different receptors. The regulation of NKG2D expression and pairing with DAP10 or DAP12 in NK cells is rather complex (Fig. 1). Resting NK cells express NKG2D-L, DAP10 and DAP12, but only DAP10 can pair with NKG2D-L. Colonna and colleagues found that signaling through NKG2D-L−DAP10 in resting NK cells results in an increase in natural cytotoxicity4. Activated NK cells express NKG2D-L−DAP10, NKG2D-S−DAP10 and NKG2D-S−DAP12. The stimulation of NK cells by contact with cells expressing ligands of NKG2D is clearly enhanced by either DAP123 or DAP104 signals. These new data, which show that NKG2D engagement on NK cells delivers signals through both DAP10 and DAP12, may explain why soluble NKG2D ligands and plate-bound NKG2D monoclonal antibody are sufficient to activate NK cell responses7,
11. However, new questions arise. How do DAP10 and DAP12 signals integrate when delivered simultaneously? What is the specific role of the DAP10 signal: is it to amplify other signals or to overcome inhibitory signals by MHC class I−specific receptors?
As NKG2D-S can pair with both DAP10 and DAP12 adapter molecules, it is not obvious why NKG2D-L is necessary. Shouldn't expression of NKG2D-S alone provide all the functions that NKG2D-L or both isoforms together may provide? A possible reason for the existence of NKG2D-L is that some cellular functions may require signaling through DAP10 alone, without coengagement of DAP12. This could not be achieved simply by regulating DAP12 expression because DAP12 is required for signaling through other receptors. Therefore, even when both adapters are expressed (for example, in resting NK cells), NKG2D-L can ensure that DAP10 functions independently from DAP12. Another possibility is that NKG2D-S may associate preferentially with DAP12. In that case, coexpression of NKG2D-L may be necessary to provide optimal signaling through DAP10 as well as DAP12. Finally, discrete signaling functions of the cytoplasmic tails of NKG2D-S and NKG2D-L, distinct from those of DAP10 and DAP12, are not excluded.
The molecular basis for selective pairing of transmembrane receptors with adapter chains is poorly understood in any receptor system. The presence of charged amino acids in the transmembrane regions cannot alone explain specific receptor-adapter associations. A fascinating implication of the promiscuous pairing of NKG2D-S with DAP10 and DAP12 is that some of the DAP12-associated receptors may also bind DAP10. For instance, the activation receptor KIR2DS2, which signals via DAP12 in NK cells, can signal in the CD28- T cells of rheumatoid arthritis patients12. The lack of DAP12 in T cells suggests an association of KIR2DS2 with another partner chain. Promiscuous pairing of receptors with different partner chains would provide versatility in the control of biological responses by individual receptors and may be more prevalent than we have appreciated so far.
REFERENCES
-
Wu, J. et al. Science 285, 730–732 (1999). | Article | PubMed | ISI | ChemPort |
-
Chang, C. et al. J. Immunol. 163, 4651–4654 (1999). | PubMed | ISI | ChemPort |
-
Diefenbach, A. et al. Nature Immunol. 3, 1142–1149 (2002). | Article | PubMed | ISI | ChemPort |
-
Gilfillan, S., Ho, E.L., Cella, M., Yokoyama, W.M. & Colonna, M. Nature Immunol. 3, 1150–1155 (2002). | Article | PubMed | ISI | ChemPort |
-
Groh, V. et al. Nature Immunol. 2, 255–260 (2001). | Article | PubMed | ISI | ChemPort |
-
Cerwenka, A. & Lanier, L.L. Nature Rev. Immunol. 1, 41–49 (2001). | Article | PubMed | ChemPort |
-
Jamieson, A.M. et al. Immunity 17, 19–29 (2002). | Article | PubMed | ISI | ChemPort |
-
Carayannopoulos, L.N., Naidenko, O.V., Fremont, D.H. & Yokoyama, W.M. J. Immunol. 169, 4079–4083 (2002). | PubMed | ISI | ChemPort |
-
Wu, J., Cherwinski, H., Spies, T., Phillips, J.H. & Lanier, L.L. J. Exp. Med. 192, 1059–1067 (2000). | Article | PubMed | ISI | ChemPort |
-
Lanier, L.L. & Bakker, A.B. Immunol. Today 21, 611–614 (2000). | Article | PubMed | ISI | ChemPort |
-
Sutherland, C.L., Chalupny, N.J., Schooley, K., VandenBos, T. & Cosman, D. J. Immunol. 168, 671–679 (2002). | PubMed | ISI | ChemPort |
-
Namekawa, T. et al. J. Immunol. 165, 1138–1145 (2000). | PubMed | ISI | ChemPort |
|
