- Email: [email protected]
Betting on NKT and NK Cells
Previews 363
tion of technology to the experimental problem at hand, since it is difficult to imagine approaching the question in any other way. Extending their prior findings on aminopeptidase cleavage of peptides in living cells (Reits et al., 2003), Reits et al. show that as peptides exceed ⵑ15 residues the critical role of aminopeptidases in degradation is usurped by TPP II. TPP II functions as both an aminopeptidase and an endopeptidase. Curiously, its endopeptidase activity requires an unblocked amino terminus, suggesting that the two activities are linked. The judicious insertion of D amino acids (which resist proteolytic cleavage), into the polypeptide sequence, demonstrates that TPP II liberates peptides of 9 or more residues from longer polypeptide chains. To address the involvement of TPP II in antigen processing the authors used butabindide, a cell permeable TPP II inhibitor. Cells were treated with acid to denature existing cell surface class I molecules, and class I reexpression was measured in the presence of inhibitors. Surprisingly, butabindide was as effective as a proteasome inhibitor in blocking class I reexpression. Crucially, the effects of the proteasomeand TPP II-inhibitors were not at all additive, strongly implying that they act on a single proteolytic pathway. Based on these findings, Reits et al. suggest the following model for the generation of the bulk of class I peptide ligands. Proteasomes degrade polypeptides into peptides of greater than 15 residues with extensions at one or both termini. These fragments in turn are acted on by TPP II, which removes some of the N-terminal residues and also provides the last chance for peptides to acquire the proper C terminus. Final trimming of N-terminal residues is then performed by cytosolic and ER aminopeptidases. It is important to emphasize that this process is anything but efficient: thousands of endogenous proteins must be degraded to create each peptide-class I complex (Yewdell et al., 2003). Indeed, the immune system is simply sampling peptide from an ancient pathway that evolved to recycle amino acids from old proteins into new ones. Many questions remain, of course. Why is TPP II so large (even bigger than proteasomes)? Can it also degrade proteins? Do substrates simply diffuse to TPPII, or is this facilitated in some way? Does TPP II physically associate with proteasomes or TAP? Does TPP II’s role
Betting on NKT and NK Cells Natural killer T (NKT) cells, as their name implies, constitutively express markers and receptors first identified on bona fide natural killer (NK) cells, supporting a potential relationship between NKT and NK cells. In this issue of Immunity, Townsend et al. further define this relationship in terms of the transcription factor, T-bet. Although NKT cells display rearranged T cell antigen receptor (TCR) chains in association with the CD3 complex in distinction from NK cells that lack expression
change in pAPCs or cells exposed to cytokines that enhance antigen presentation? Do immunoproteasomes produce shorter peptides less dependent on TPP II cleavage? Or do PA28 proteasome regulators, whose role in antigen presentation has remained elusive, play this role? Does TPP II also participate in the generation of peptides from exogenous antigens? If so, does it colocalize with proteasomes and TAP in regions where phagosomes fuse with the ER in pAPCs? And finally the big one: have all of the major proteases involved in the generation of class I peptide ligand been identified, or do other surprises await us? Jonathan W. Yewdell and Michael F. Princiotta Laboratory of Viral Diseases National Institute of Allergy and Infectious Diseases Bethesda, Maryland 208920 Selected Reading Geier, E., Pfeifer, G., Wilm, M., Lucchiari-Hartz, M., Baumeister, W., Eichmann, K., and Niedermann, G. (1999). Science 283, 978–981. Glas, R., Bogyo, M., McMaster, J.S., Gaczynska, M., and Ploegh, H.L. (1998). Nature 392, 618–622. Princiotta, M.F., Schubert, U., Chen, W., Bennink, J.R., Myung, J., Crews, C.M., and Yewdell, J.W. (2001). Proc. Natl. Acad. Sci. USA 98, 513–518. Reits, E., Griekspoor, A., Neijssen, J., Groothuis, T., Jalink, K., van Veelen, P., Janssen, H., Calafat, J., Drijfhout, J.W., and Neefjes, J. (2003). Immunity 18, 97–108. Reits, E., Neijssen, J., Herberts, C., Benckhuijsen, W., Janssen, L., Drijfhout J.W., and Neefjes, J. (2004). Immunity 20, this issue, 495–506. Rock, K.L., Gramm, C., Rothstein, L., Clark, K., Stein, R., Dick, L., Hwang, D., and Goldberg, A.L. (1994). Cell 78, 761–771. Seifert, U., Maranon, C., Shmueli, A., Desoutter, J.F., Wesoloski, L., Janek, K., Henklein, P., Diescher, S., Andrieu, M., de la Salle, .H., et al. (2003). Nat. Immunol 4, 375–379. Serwold, T., Gonzalez, F., Kim, J., Jacob, R., and Shastri, N. (2002). Nature 419, 480–483. Shastri, N., Schwab, S., and Serwold, T. (2002). Annu. Rev. Immunol. 20, 463–493. Yewdell, J.W., Reits, E., and Neefjes, J. (2003). Nat. Rev. Immunol. 3, 952–961. York, I.A., Chang, S.C., Saric, T., Keys, J.A., Favreau, J.M., Goldberg, A.L., and Rock, K.L. (2002). Nat. Immunol. 3, 1177–1184.
of the TCR/CD3 complex, they share properties with “innate lymphocytes” that closely resemble NK cells, such as the expression of invariant, germline-encoded receptors (Bendelac et al., 2001). Although the definition of NKT cells is expanding (Kronenberg and Gapin, 2002), a specific TCR␣ chain (V␣14/J␣28) is expressed by most murine NKT cells and is strikingly conserved between mice and humans. These V␣14-invariant (V␣14i) NKT cells recognize CD1-restricted antigens, as represented by the marine sponge-derived glycolipid, ␣-galactosyl ceramide (␣-GalCer). Perhaps the best known function of NKT cells is related to their unique innate capacity to secrete several cytokines, including IL-4 and interferon-␥ (IFN␥), very quickly after TCR stimulation in vivo.
Immunity 364
NKT cells play roles in immune responses ranging from anti-pathogen effects, due to recognition of glycolipids presented by CD1, to the pathogenesis of inflammatory conditions, such as allergic asthma. Similar to conventional T cells, NKT cells fail to develop in mice with deficiencies in the recombinase machinery for antigen receptor gene rearrangement or in the absence of the restricting MHC element. By contrast, NKT cells also fail to develop in mice lacking lymphotoxin-␣ (Lt␣) or components of the IL-15 pathway, such as IL-15 and IL-15R␣, and IL-2/15R, whereas conventional T cell development is grossly intact (Kronenberg and Gapin, 2002). Such data support the thesis that NKT cells form a distinct sublineage of T cells, differing in terms of developmental requirements. Like NKT cells, NK cells express immunoreceptor tyrosine-based activation motif (ITAM)-containing signaling chains coupled to target recognition receptors that are germline encoded (though unrearranged). In C57BL/6 mice, these receptors include the lectin-like NK1.1 (Nkrp1c) molecule and several others encoded in the NK gene complex. In addition to target killing, NK cells are poised for prompt cytokine release in response to inflammation and infection, akin to NKT cell responses. Indeed, there is evidence indicating crosstalk between NKT and NK cells during acute inflammatory responses (Bendelac et al., 2001), and the shared receptors may contribute to the immediate responses of these cells. Interestingly, Lt␣ and IL-15 pathway targeted mutant mice lacking NKT cells also lack NK cells, indicating that NKT and NK cells require some of the same cytokines for functional development and highlighting the close developmental relationship of NKT and NK cells. Also, both NKT and NK cells are absent in mice deficient in certain transcription factors, such as Ets-1 and IRF-1, and their absence is relatively selective with few effects on conventional T cell development. Thus, there is a close developmental relationship between NKT and NK cells. Townsend et al. provide a very comprehensive evaluation of NKT and NK cell development in mice deficient in T-bet (T-box expressed in T cells), a transcription factor first identified in terms of T helper cell differentiation but now known to affect functions of other immune cells (Townsend et al., 2004). T-bet-deficient mice have a marked reduction in NKT cell number in peripheral tissues as detected by anti-NK1.1 and by ␣-GalCer-CD1 tetramers. This reduction is apparently not due to an effect of T-bet on CD1 expression. Detailed analyses of the few remaining NKT cells in T-bet-deficient mice indicate a block at an intermediate stage of NKT cell development. In contrast to the effect of T-bet on NKT numbers and to the profound effects of Ets-1 and IRF-1 deficiency on NK cell development (Kronenberg and Gapin, 2002), the effect of T-bet deficiency on NK cells is more subtle, with only modest reductions in NK cell numbers. Similar to the block in NKT cell development, however, the T-bet-deficient NK cells manifest defects in expression of markers associated with their final maturation steps (Yokoyama et al., 2004) and abnormalities in the effector functions of mature NK cells, such as cytokine production and to a less impressive extent, cytotoxicity. Some of the less dramatic effects on NK cells may be due to the T-bet-independent expression
of a related transcription factor, Eomesodermin, in NK but not NKT cells, in a manner reminiscent of a recently reported role of Eomesodermin in CD8⫹ T cell effector function (Pearce et al., 2003). Nevertheless, T-bet deficiency affects the development of both NKT and NK cells. In prior studies of NKT or NK cell defects in transcription factor-deficient mice, it was not always clear if the transcription factor defects were cell intrinsic (autonomous) or extrinsic (micro-environment). Such effects are often revealed when transcription factor-deficient hematopoietic stem cells are transplanted into wild-type recipients and vice versa with an allotypic marker, such as Ly5.1, to detect donor-derived mature cells in the bone marrow chimeras. By using the chimera approach, Townsend et al. demonstrate that the T-bet deficiency on NKT and NK cell development is a stem cell intrinsic effect (Townsend et al., 2004). They then go one step further and demonstrate that restoration of T-bet expression by retroviral transduction of T-bet-deficient thymocytes or hematopoietic stem cells reconstitutes NKT and NK cell development, respectively. Thus, the T-bet deficiency is one of only a few clearly documented examples of cell intrinsic transcription factor defects that affect development of both NKT and NK cells since others have either not been documented (Ets-1), are extrinsic (IRF-1, RelB), or have more widespread developmental deficits (Ikaros, PU.1, MEF, IRF-2, RelB, GATA-3) (Colucci et al., 2003). An intrinsic defect with effects on NKT and NK cells suggests that the pathways for final differentiation of NKT and NK cells could be similar but operate in already committed immature cells, as Townsend et al. favor. Alternatively, T-bet could operate at the level of a precursor cell that gives rise to either NKT or NK cell but not other lineages. Moreover, it remains possible that the precise function of T-bet may differ in different lineage contexts. Finally, T-bet may have other cell intrinsic roles in the development or terminal differentiation of other immune cells. These topics are worthy of further evaluation. Of course, a cell intrinsic defect in T-bet-deficient animals should mean that T-bet is expressed in developing NKT and NK cells. Indeed, Townsend et al. confirm this and suggest its expression is related to the level of cell maturation. For future studies, it would be of interest to evaluate T-bet expression in developing NK cells at different maturation stages that are just beginning to be defined in the bone marrow (Colucci et al., 2003; Yokoyama et al., 2004). Interestingly, T-bet transcripts can be markedly upregulated (more than 10-fold) in mature NK cells by various cytokines, especially by IL-12 plus IL-18, whereas more modest increases were noted with stimulation of ITAM-coupled NK cell activation receptors. This observation suggests another role for T-bet in regulating the final effector functions of NK cells in response to cytokines, rather than development per se, perhaps more akin to T helper differentiation that is also dependent on the cytokine milieu. If so, this possibility suggests that the NK cell defect in T-bet deficiency could also be due to defective signaling and gene activation in response to inflammation rather than primarily due to arrested development of mature NK cells as Townsend et al. favor. These two related aspects of
Previews 365
T-bet function in NK cells may prove difficult to study in isolation unless T-bet affects transcription of only selective genes, directly or indirectly. What genes are directly affected by T-bet? Townsend et al. provide novel insight with a modified chromatin immunoprecipitation experiment in which proteins are crosslinked to DNA, immunoprecipitated with anti-T-bet antibody, and the target DNA identified by PCR. These experiments were aided by CpG island microarray analysis and revealed several candidate genes whose promoters are bound by T-bet. These genes include those for perforin and granzyme B that are found in the lytic granules of NK cells, are known to be enhanced by IL12 and IL-18 stimulation as they verify, and are affected by T-bet deficiency. These results are intriguing because presumably Townsend et al. did not find significant evidence for T-bet involvement in other genes that are also induced by IL-12 and IL-18 or the genes for Mac-1 and CD43 that are poorly expressed by T-bet-deficient NK cells. Other, less defined genes were also identified, including Runx1, a transcription factor of recent interest due to the genetic linkage of polymorphisms of Runx1 binding sites to human autoimmune diseases, including systemic lupus erythematosus, psoriasis, and rheumatoid arthritis (Helms et al., 2003; Prokunina et al., 2002; Tokuhiro et al., 2003), suggesting that T-bet may function in chronic inflammatory autoimmune diseases as already noted (Neurath et al., 2002; Peng et al., 2002) by acting upstream of Runx1. There is much more that needs to be done to understand the relationship between T-bet regulation of these genes and to evaluate the significance of the putative regulation of Runx1 and other genes by T-bet in autoimmunity and other conditions. Meanwhile, it will be informative to discern the relationship of T-bet to other transcription factors that affect NKT and NK development. Townsend et al. appear to have begun this complex analysis, suggesting that IRF-2, Ets-1, MEF, PU.1, and GATA-3 are upstream of T-bet as discerned from published observations and from potential effects of these factors on T-bet expression. But there may be differences in how these factors regulate each other in the NKT and NK cell lineages.
Moreover, the putative redundant role of Eomesodermin in NK cell development requires further analysis. Nevertheless, immunologists now have significant new insights in the dissection of the molecular basis of NKT and NK development and effector function, and these findings may also lead to clues in understanding autoimmune diseases. Wayne M. Yokoyama Howard Hughes Medical Institute Rheumatology Division Box 8045 Washington University School of Medicine 660 South Euclid Avenue St. Louis, Missouri 63110 Selected Reading Bendelac, A., Bonneville, M., and Kearney, J.F. (2001). Nat. Rev. Immunol. 1, 177–186. Colucci, F., Caligiuri, M.A., and Di Santo, J.P. (2003). Nat. Rev. Immunol. 3, 413–425. Helms, C., Cao, L., Krueger, J.G., Wijsman, E.M., Chamian, F., Gordon, D., Heffernan, M., Daw, J.A., Robarge, J., Ott, J., et al. (2003). Nat. Genet. 35, 349–356. Kronenberg, M., and Gapin, L. (2002). Nat. Rev. Immunol. 2, 557–568. Neurath, M.F., Weigmann, B., Finotto, S., Glickman, J., Nieuwenhuis, E., Iijima, H., Mizoguchi, A., Mizoguchi, E., Mudter, J., Galle, P.R., et al. (2002). J. Exp. Med. 195, 1129–1143. Pearce, E.L., Mullen, A.C., Martins, G.A., Krawczyk, C.M., Hutchins, A.S., Zediak, V.P., Banica, M., DiCioccio, C.B., Gross, D.A., Mao, C.A., et al. (2003). Science 302, 1041–1043. Peng, S.L., Szabo, S.J., and Glimcher, L.H. (2002). Proc. Natl. Acad. Sci. USA 99, 5545–5550. Prokunina, L., Castillejo-Lopez, C., Oberg, F., Gunnarsson, I., Berg, L., Magnusson, V., Brookes, A.J., Tentler, D., Kristjansdottir, H., Grondal, G., et al. (2002). Nat. Genet. 32, 666–669. Tokuhiro, S., Yamada, R., Chang, X., Suzuki, A., Kochi, Y., Sawada, T., Suzuki, M., Nagasaki, M., Ohtsuki, M., Ono, M., et al. (2003). Nat. Genet. 35, 341–348. Townsend, M.J., Weinmann, A.S., Matsuda, J., Saloman, R., Farnham, P., Biron, C.A., Gapin, L., and Glimcher, L.H. (2004). Immunity 20, 477–494. Yokoyama, W.M., Kim, S., and French, A.R. (2004). Annu. Rev. Immunol. 22, 405–429.
tion of technology to the experimental problem at hand, since it is difficult to imagine approaching the question in any other way. Extending their prior findings on aminopeptidase cleavage of peptides in living cells (Reits et al., 2003), Reits et al. show that as peptides exceed ⵑ15 residues the critical role of aminopeptidases in degradation is usurped by TPP II. TPP II functions as both an aminopeptidase and an endopeptidase. Curiously, its endopeptidase activity requires an unblocked amino terminus, suggesting that the two activities are linked. The judicious insertion of D amino acids (which resist proteolytic cleavage), into the polypeptide sequence, demonstrates that TPP II liberates peptides of 9 or more residues from longer polypeptide chains. To address the involvement of TPP II in antigen processing the authors used butabindide, a cell permeable TPP II inhibitor. Cells were treated with acid to denature existing cell surface class I molecules, and class I reexpression was measured in the presence of inhibitors. Surprisingly, butabindide was as effective as a proteasome inhibitor in blocking class I reexpression. Crucially, the effects of the proteasomeand TPP II-inhibitors were not at all additive, strongly implying that they act on a single proteolytic pathway. Based on these findings, Reits et al. suggest the following model for the generation of the bulk of class I peptide ligands. Proteasomes degrade polypeptides into peptides of greater than 15 residues with extensions at one or both termini. These fragments in turn are acted on by TPP II, which removes some of the N-terminal residues and also provides the last chance for peptides to acquire the proper C terminus. Final trimming of N-terminal residues is then performed by cytosolic and ER aminopeptidases. It is important to emphasize that this process is anything but efficient: thousands of endogenous proteins must be degraded to create each peptide-class I complex (Yewdell et al., 2003). Indeed, the immune system is simply sampling peptide from an ancient pathway that evolved to recycle amino acids from old proteins into new ones. Many questions remain, of course. Why is TPP II so large (even bigger than proteasomes)? Can it also degrade proteins? Do substrates simply diffuse to TPPII, or is this facilitated in some way? Does TPP II physically associate with proteasomes or TAP? Does TPP II’s role
Betting on NKT and NK Cells Natural killer T (NKT) cells, as their name implies, constitutively express markers and receptors first identified on bona fide natural killer (NK) cells, supporting a potential relationship between NKT and NK cells. In this issue of Immunity, Townsend et al. further define this relationship in terms of the transcription factor, T-bet. Although NKT cells display rearranged T cell antigen receptor (TCR) chains in association with the CD3 complex in distinction from NK cells that lack expression
change in pAPCs or cells exposed to cytokines that enhance antigen presentation? Do immunoproteasomes produce shorter peptides less dependent on TPP II cleavage? Or do PA28 proteasome regulators, whose role in antigen presentation has remained elusive, play this role? Does TPP II also participate in the generation of peptides from exogenous antigens? If so, does it colocalize with proteasomes and TAP in regions where phagosomes fuse with the ER in pAPCs? And finally the big one: have all of the major proteases involved in the generation of class I peptide ligand been identified, or do other surprises await us? Jonathan W. Yewdell and Michael F. Princiotta Laboratory of Viral Diseases National Institute of Allergy and Infectious Diseases Bethesda, Maryland 208920 Selected Reading Geier, E., Pfeifer, G., Wilm, M., Lucchiari-Hartz, M., Baumeister, W., Eichmann, K., and Niedermann, G. (1999). Science 283, 978–981. Glas, R., Bogyo, M., McMaster, J.S., Gaczynska, M., and Ploegh, H.L. (1998). Nature 392, 618–622. Princiotta, M.F., Schubert, U., Chen, W., Bennink, J.R., Myung, J., Crews, C.M., and Yewdell, J.W. (2001). Proc. Natl. Acad. Sci. USA 98, 513–518. Reits, E., Griekspoor, A., Neijssen, J., Groothuis, T., Jalink, K., van Veelen, P., Janssen, H., Calafat, J., Drijfhout, J.W., and Neefjes, J. (2003). Immunity 18, 97–108. Reits, E., Neijssen, J., Herberts, C., Benckhuijsen, W., Janssen, L., Drijfhout J.W., and Neefjes, J. (2004). Immunity 20, this issue, 495–506. Rock, K.L., Gramm, C., Rothstein, L., Clark, K., Stein, R., Dick, L., Hwang, D., and Goldberg, A.L. (1994). Cell 78, 761–771. Seifert, U., Maranon, C., Shmueli, A., Desoutter, J.F., Wesoloski, L., Janek, K., Henklein, P., Diescher, S., Andrieu, M., de la Salle, .H., et al. (2003). Nat. Immunol 4, 375–379. Serwold, T., Gonzalez, F., Kim, J., Jacob, R., and Shastri, N. (2002). Nature 419, 480–483. Shastri, N., Schwab, S., and Serwold, T. (2002). Annu. Rev. Immunol. 20, 463–493. Yewdell, J.W., Reits, E., and Neefjes, J. (2003). Nat. Rev. Immunol. 3, 952–961. York, I.A., Chang, S.C., Saric, T., Keys, J.A., Favreau, J.M., Goldberg, A.L., and Rock, K.L. (2002). Nat. Immunol. 3, 1177–1184.
of the TCR/CD3 complex, they share properties with “innate lymphocytes” that closely resemble NK cells, such as the expression of invariant, germline-encoded receptors (Bendelac et al., 2001). Although the definition of NKT cells is expanding (Kronenberg and Gapin, 2002), a specific TCR␣ chain (V␣14/J␣28) is expressed by most murine NKT cells and is strikingly conserved between mice and humans. These V␣14-invariant (V␣14i) NKT cells recognize CD1-restricted antigens, as represented by the marine sponge-derived glycolipid, ␣-galactosyl ceramide (␣-GalCer). Perhaps the best known function of NKT cells is related to their unique innate capacity to secrete several cytokines, including IL-4 and interferon-␥ (IFN␥), very quickly after TCR stimulation in vivo.
Immunity 364
NKT cells play roles in immune responses ranging from anti-pathogen effects, due to recognition of glycolipids presented by CD1, to the pathogenesis of inflammatory conditions, such as allergic asthma. Similar to conventional T cells, NKT cells fail to develop in mice with deficiencies in the recombinase machinery for antigen receptor gene rearrangement or in the absence of the restricting MHC element. By contrast, NKT cells also fail to develop in mice lacking lymphotoxin-␣ (Lt␣) or components of the IL-15 pathway, such as IL-15 and IL-15R␣, and IL-2/15R, whereas conventional T cell development is grossly intact (Kronenberg and Gapin, 2002). Such data support the thesis that NKT cells form a distinct sublineage of T cells, differing in terms of developmental requirements. Like NKT cells, NK cells express immunoreceptor tyrosine-based activation motif (ITAM)-containing signaling chains coupled to target recognition receptors that are germline encoded (though unrearranged). In C57BL/6 mice, these receptors include the lectin-like NK1.1 (Nkrp1c) molecule and several others encoded in the NK gene complex. In addition to target killing, NK cells are poised for prompt cytokine release in response to inflammation and infection, akin to NKT cell responses. Indeed, there is evidence indicating crosstalk between NKT and NK cells during acute inflammatory responses (Bendelac et al., 2001), and the shared receptors may contribute to the immediate responses of these cells. Interestingly, Lt␣ and IL-15 pathway targeted mutant mice lacking NKT cells also lack NK cells, indicating that NKT and NK cells require some of the same cytokines for functional development and highlighting the close developmental relationship of NKT and NK cells. Also, both NKT and NK cells are absent in mice deficient in certain transcription factors, such as Ets-1 and IRF-1, and their absence is relatively selective with few effects on conventional T cell development. Thus, there is a close developmental relationship between NKT and NK cells. Townsend et al. provide a very comprehensive evaluation of NKT and NK cell development in mice deficient in T-bet (T-box expressed in T cells), a transcription factor first identified in terms of T helper cell differentiation but now known to affect functions of other immune cells (Townsend et al., 2004). T-bet-deficient mice have a marked reduction in NKT cell number in peripheral tissues as detected by anti-NK1.1 and by ␣-GalCer-CD1 tetramers. This reduction is apparently not due to an effect of T-bet on CD1 expression. Detailed analyses of the few remaining NKT cells in T-bet-deficient mice indicate a block at an intermediate stage of NKT cell development. In contrast to the effect of T-bet on NKT numbers and to the profound effects of Ets-1 and IRF-1 deficiency on NK cell development (Kronenberg and Gapin, 2002), the effect of T-bet deficiency on NK cells is more subtle, with only modest reductions in NK cell numbers. Similar to the block in NKT cell development, however, the T-bet-deficient NK cells manifest defects in expression of markers associated with their final maturation steps (Yokoyama et al., 2004) and abnormalities in the effector functions of mature NK cells, such as cytokine production and to a less impressive extent, cytotoxicity. Some of the less dramatic effects on NK cells may be due to the T-bet-independent expression
of a related transcription factor, Eomesodermin, in NK but not NKT cells, in a manner reminiscent of a recently reported role of Eomesodermin in CD8⫹ T cell effector function (Pearce et al., 2003). Nevertheless, T-bet deficiency affects the development of both NKT and NK cells. In prior studies of NKT or NK cell defects in transcription factor-deficient mice, it was not always clear if the transcription factor defects were cell intrinsic (autonomous) or extrinsic (micro-environment). Such effects are often revealed when transcription factor-deficient hematopoietic stem cells are transplanted into wild-type recipients and vice versa with an allotypic marker, such as Ly5.1, to detect donor-derived mature cells in the bone marrow chimeras. By using the chimera approach, Townsend et al. demonstrate that the T-bet deficiency on NKT and NK cell development is a stem cell intrinsic effect (Townsend et al., 2004). They then go one step further and demonstrate that restoration of T-bet expression by retroviral transduction of T-bet-deficient thymocytes or hematopoietic stem cells reconstitutes NKT and NK cell development, respectively. Thus, the T-bet deficiency is one of only a few clearly documented examples of cell intrinsic transcription factor defects that affect development of both NKT and NK cells since others have either not been documented (Ets-1), are extrinsic (IRF-1, RelB), or have more widespread developmental deficits (Ikaros, PU.1, MEF, IRF-2, RelB, GATA-3) (Colucci et al., 2003). An intrinsic defect with effects on NKT and NK cells suggests that the pathways for final differentiation of NKT and NK cells could be similar but operate in already committed immature cells, as Townsend et al. favor. Alternatively, T-bet could operate at the level of a precursor cell that gives rise to either NKT or NK cell but not other lineages. Moreover, it remains possible that the precise function of T-bet may differ in different lineage contexts. Finally, T-bet may have other cell intrinsic roles in the development or terminal differentiation of other immune cells. These topics are worthy of further evaluation. Of course, a cell intrinsic defect in T-bet-deficient animals should mean that T-bet is expressed in developing NKT and NK cells. Indeed, Townsend et al. confirm this and suggest its expression is related to the level of cell maturation. For future studies, it would be of interest to evaluate T-bet expression in developing NK cells at different maturation stages that are just beginning to be defined in the bone marrow (Colucci et al., 2003; Yokoyama et al., 2004). Interestingly, T-bet transcripts can be markedly upregulated (more than 10-fold) in mature NK cells by various cytokines, especially by IL-12 plus IL-18, whereas more modest increases were noted with stimulation of ITAM-coupled NK cell activation receptors. This observation suggests another role for T-bet in regulating the final effector functions of NK cells in response to cytokines, rather than development per se, perhaps more akin to T helper differentiation that is also dependent on the cytokine milieu. If so, this possibility suggests that the NK cell defect in T-bet deficiency could also be due to defective signaling and gene activation in response to inflammation rather than primarily due to arrested development of mature NK cells as Townsend et al. favor. These two related aspects of
Previews 365
T-bet function in NK cells may prove difficult to study in isolation unless T-bet affects transcription of only selective genes, directly or indirectly. What genes are directly affected by T-bet? Townsend et al. provide novel insight with a modified chromatin immunoprecipitation experiment in which proteins are crosslinked to DNA, immunoprecipitated with anti-T-bet antibody, and the target DNA identified by PCR. These experiments were aided by CpG island microarray analysis and revealed several candidate genes whose promoters are bound by T-bet. These genes include those for perforin and granzyme B that are found in the lytic granules of NK cells, are known to be enhanced by IL12 and IL-18 stimulation as they verify, and are affected by T-bet deficiency. These results are intriguing because presumably Townsend et al. did not find significant evidence for T-bet involvement in other genes that are also induced by IL-12 and IL-18 or the genes for Mac-1 and CD43 that are poorly expressed by T-bet-deficient NK cells. Other, less defined genes were also identified, including Runx1, a transcription factor of recent interest due to the genetic linkage of polymorphisms of Runx1 binding sites to human autoimmune diseases, including systemic lupus erythematosus, psoriasis, and rheumatoid arthritis (Helms et al., 2003; Prokunina et al., 2002; Tokuhiro et al., 2003), suggesting that T-bet may function in chronic inflammatory autoimmune diseases as already noted (Neurath et al., 2002; Peng et al., 2002) by acting upstream of Runx1. There is much more that needs to be done to understand the relationship between T-bet regulation of these genes and to evaluate the significance of the putative regulation of Runx1 and other genes by T-bet in autoimmunity and other conditions. Meanwhile, it will be informative to discern the relationship of T-bet to other transcription factors that affect NKT and NK development. Townsend et al. appear to have begun this complex analysis, suggesting that IRF-2, Ets-1, MEF, PU.1, and GATA-3 are upstream of T-bet as discerned from published observations and from potential effects of these factors on T-bet expression. But there may be differences in how these factors regulate each other in the NKT and NK cell lineages.
Moreover, the putative redundant role of Eomesodermin in NK cell development requires further analysis. Nevertheless, immunologists now have significant new insights in the dissection of the molecular basis of NKT and NK development and effector function, and these findings may also lead to clues in understanding autoimmune diseases. Wayne M. Yokoyama Howard Hughes Medical Institute Rheumatology Division Box 8045 Washington University School of Medicine 660 South Euclid Avenue St. Louis, Missouri 63110 Selected Reading Bendelac, A., Bonneville, M., and Kearney, J.F. (2001). Nat. Rev. Immunol. 1, 177–186. Colucci, F., Caligiuri, M.A., and Di Santo, J.P. (2003). Nat. Rev. Immunol. 3, 413–425. Helms, C., Cao, L., Krueger, J.G., Wijsman, E.M., Chamian, F., Gordon, D., Heffernan, M., Daw, J.A., Robarge, J., Ott, J., et al. (2003). Nat. Genet. 35, 349–356. Kronenberg, M., and Gapin, L. (2002). Nat. Rev. Immunol. 2, 557–568. Neurath, M.F., Weigmann, B., Finotto, S., Glickman, J., Nieuwenhuis, E., Iijima, H., Mizoguchi, A., Mizoguchi, E., Mudter, J., Galle, P.R., et al. (2002). J. Exp. Med. 195, 1129–1143. Pearce, E.L., Mullen, A.C., Martins, G.A., Krawczyk, C.M., Hutchins, A.S., Zediak, V.P., Banica, M., DiCioccio, C.B., Gross, D.A., Mao, C.A., et al. (2003). Science 302, 1041–1043. Peng, S.L., Szabo, S.J., and Glimcher, L.H. (2002). Proc. Natl. Acad. Sci. USA 99, 5545–5550. Prokunina, L., Castillejo-Lopez, C., Oberg, F., Gunnarsson, I., Berg, L., Magnusson, V., Brookes, A.J., Tentler, D., Kristjansdottir, H., Grondal, G., et al. (2002). Nat. Genet. 32, 666–669. Tokuhiro, S., Yamada, R., Chang, X., Suzuki, A., Kochi, Y., Sawada, T., Suzuki, M., Nagasaki, M., Ohtsuki, M., Ono, M., et al. (2003). Nat. Genet. 35, 341–348. Townsend, M.J., Weinmann, A.S., Matsuda, J., Saloman, R., Farnham, P., Biron, C.A., Gapin, L., and Glimcher, L.H. (2004). Immunity 20, 477–494. Yokoyama, W.M., Kim, S., and French, A.R. (2004). Annu. Rev. Immunol. 22, 405–429.