- Email: [email protected]
TCF-1 Flips the Switch on Eomes
Immunity
Previews a-Gal-A enzyme activity. Therefore, these findings suggested that microbial activation of DCs leads to an accumulation of glycolipids that are normally degraded by a-Gal-A (Figure 1, right panel). Inhibitor studies further suggested that iGb3 could well be a relevant glycolipid in this context. The idea that iGb3 might be the sole antigen that controls the intrathymic development of iNKT cells is particularly contentious (Gapin, 2010). Darmoise et al. (2010) were unable to identify a defect in intrathymic development of iNKT cells in Fabry mice, which is consistent with some, but not all, prior studies (e.g., Gadola et al., 2006). Nevertheless, they found that numbers of peripheral iNKT cells in these animals were substantially suppressed, which they attributed to increased death because of chronic activation in the periphery. Consistent with this possibility, the remaining iNKT cells in Fabry mice exhibited a hyporesponsive phenotype when stimulated with a-galactosylceramide, which was associated with increased expression of inhibitory NK cell receptors. Is this pathway for the generation of endogenous iNKT cell antigens in mice conserved in humans? Probably not. Studies with human iNKT cell clones have failed to find a role for the CD1d
recycling pathway (see Figure 1) in the autoreactivity of these cells (Chen et al., 2007). Further, as mentioned above, although ultrasensitive mass spectrometry studies were able to identify iGb3 in human thymus, the iGb3 synthase gene in humans appears to be defective. Lastly, patients with Fabry disease have normal numbers of iNKT cells. As with most provocative studies, the paper from Darmoise et al. (2010) raises more questions than it answers: What are the relevant iNKT cell self-antigens that accumulate in Fabry mice? Is the accumulation of iNKT cell antigens in Fabry mice unique to DCs, or does this also occur in other cell types such as the cells that mediate thymic selection of iNKT cells? What controls the transient inactivation of a-Gal-A enzyme activity in DCs stimulated with microbial products? How might the relatively modest suppression of a-Gal-A activity in response to microbes mediate such a profound effect on iNKT cell activation? Do the relevant self-antigens that accumulate in Fabry mice contribute to other pathological conditions that are influenced by iNKT cells? To what extent, if at all, is this pathway conserved in humans? Providing answers to these questions will likely keep iNKT cell biologists busy for the coming years.
REFERENCES Chen, X., Wang, X., Keaton, J.M., Reddington, F., Illarionov, P.A., Besra, G.S., and Gumperz, J.E. (2007). J. Immunol. 178, 6181–6190. Darmoise, A., Teneberg, S., Bouzonville, L., Brady, R.O., Beck, M., Kaufmann, S.H.E., and Winau, F. (2010). Immunity 33, this issue, 216–228. Gadola, S.D., Silk, J.D., Jeans, A., Illarionov, P.A., Salio, M., Besra, G.S., Dwek, R., Butters, T.D., Platt, F.M., and Cerundolo, V. (2006). J. Exp. Med. 203, 2293–2303. Gapin, L. (2010). Nat. Rev. Immunol. 10, 272–277. Li, Y., Thapa, P., Hawke, D., Kondo, Y., Kurukawa, K., Furukawa, K., Hsu, F.-F., Adlercreutz, D., Weadge, J., Palcic, M.M., et al. (2009). J. Proteome Res. 8, 2740–2751. Mattner, J., Debord, K.L., Ismail, N., Goff, R.D., Cantu, C., III, Zhou, D., Saint-Mezard, P., Wang, V., Gao, Y., Yin, N., et al. (2005). Science 434, 525–529. Muindi, K., Cernadas, M., Watts, G.F.M., Royle, L., Neville, D.C.A., Dwek, R.A., Besra, G.S., Rudd, P.M., Butters, T.D., and Brenner, M.B. (2010). Proc. Natl. Acad. Sci. USA 107, 3052– 3057. Stanic, A.K., De Silva, A.D., Park, J.-J., Sriram, V., Ichikawa, S., Hirabyashi, Y., Hayakawa, K., Van Kaer, L., Brutkiewicz, R.R., and Joyce, S. (2003). Proc. Natl. Acad. Sci. USA 100, 1849– 1854. Zhou, D., Mattner, J., Cantu, C., III, Schrantz, N., Yin, N., Gao, Y., Sagiv, Y., Hudspeth, K., Wu, Y.P., Yamashita, T., et al. (2004). Science 306, 1786–1789.
TCF-1 Flips the Switch on Eomes Michael A. Paley1 and E. John Wherry1,* 1Department of Microbiology and Institute for Immunology, University of Pennsylvania School of Medicine, Philadelphia, PA 19104, USA *Correspondence: [email protected] DOI 10.1016/j.immuni.2010.08.008
The interaction between different transcriptional pathways in CD8+ T cell memory remains incompletely understood. In this issue, Zhou et al. (2010) demonstrate an important role for T cell factor 1 in regulating CD8+ T cell memory via control of a second transcription factor, Eomesodermin. There are clear parallels between the development of adaptive immune responses and other developmental systems. The Wnt-b-catenin pathway is important in many of these development processes. Although canonical Wnt signaling has a role in thymic development and has been reported to influence CD8+ T cell memory
(Zhao et al., 2010), precisely how this morphogenic signaling pathway is linked to the transcriptional regulation of T cell memory has remained unclear. Normally, in the absence of Wnt ligands, b-catenin is sequestered and rapidly degraded in the cytoplasm. Wnt signaling stabilizes b-catenin, allowing translocation to the
nucleus and interaction with T cell factor 1 (TCF-1). b-catenin displaces TLE1, a Groucho family corepressor bound to TCF-1, converting the repressive TCF-1-TLE1 complex into a TCF-1-b-catenin transcriptional activator (MacDonald et al., 2009). In this issue, Zhou et al. (2010) demonstrate that TCF-1 activation by
Immunity 33, August 27, 2010 ª2010 Elsevier Inc. 145
Immunity
Previews Naive
Effector
Memory 3 TCF-1?
2 TCF-1?
4
TCF-1? Tem
?
TCF-1 (Eomes)
Terminal effector
1 TCF-1?
KLRG1 CD127lo CD62Llo hi
2 TCF-1?
Tcm
KLRG1lo KLRG1lo CD127hi CD127hi CD62Llo CD62Lhi
TCF-1 Eomes
Death
Figure 1. Development of the CD8 T Cell Response to Viral Infection
Upon activation, naive CD8+ T cells proliferate and differentiate into two subsets: terminal effectors (red) and memory precursors (yellow). While terminal effectors undergo cell death, memory precursors mature into Tem-like cells (green) as they enter the memory phase. Over time, these Tem cells further differentiate into Tcm cells (blue), which are efficiently maintained by IL-7- and IL-15-driven self-renewal. TCF-1 might regulate this developmental process during (1) early lineage decisions, (2) survival during the contraction phase, (3) differentiation from Tem to Tcm cells, or (4) self-renewal of Tcm cells.
Wnt-b-catenin signaling is a key event in the transcriptional induction of Eomes, a transcription factor involved in CD8+ T cell memory (Intlekofer et al., 2005). Although some alterations in the primary CD8+ T cell response to Listeria monocytogenes were observed, a major effect of TCF-1 deficiency was the failure to acquire canonical memory CD8+ T cells properties, notably homeostatic maintenance. This defect was more severe in lymphoid organs, and TCF-1-deficient memory CD8+ T cells remained CD62Llo and CCR7lo, consistent with poor development or maintenance of ‘‘central memory’’ CD8+ T (Tcm) cells. Without TCF-1, expression of IL-15Rb (CD122) was reduced and TCF-1-deficient memory CD8+ T cells had defective homeostatic proliferation consistent with impaired IL-15 responsiveness. Eomes has been previously implicated in CD8+ T cell memory via CD122 expression (Intlekofer et al., 2005). Zhou et al. (2010) now show that Eomes expression was markedly reduced in the absence of TCF-1. Retroviral overexpression of Eomes in TCF-1deficient cells restored CD122 expression and improved survival, demonstrating that TCF-1-dependent expression of Eomes contributed to memory CD8+ T cell maintenance.
One interesting set of observations from this work and other recent studies (Jeannet et al., 2010) is the potential role of TCF-1 at different points during CD8+ T cell activation and differentiation. Normally, after acute infection, distinct stages of memory CD8+ T cell differentiation exist. After initial stimulation, naive CD8+ T cells become activated, proliferate, acquire effector functions, and give rise to two subsets of effector CD8+ T cells. KLRG1hiCD127lo CD8+ T cells are potent effector CD8+ T cells but are terminally differentiated. These cells cannot give rise to self-renewing, stable populations of memory CD8+ T cells and their numbers decline rapidly during contraction, but then more slowly in the early memory phase (Joshi et al., 2007). These cells also proliferate poorly upon rechallenge (Kaech and Wherry, 2007). In contrast, KLRG1loCD127hi CD8+ T cells can further differentiate into long-lived, homeostatically maintained memory CD8+ T cells. Initially, the memory population formed by the KLRG1loCD127hi memory precursors is CD62Llo and resembles an effector memory T (Tem) cell-like population. Over time, these Tem cells can give rise to CD62Lhi Tcm-like cells capable of robust proliferative expansion after rechallenge (Kaech and Wherry,
146 Immunity 33, August 27, 2010 ª2010 Elsevier Inc.
2007). It is also these Tcm cells that can efficiently use IL-7 and IL-15 and undergo homeostatic proliferation (Kaech and Wherry, 2007). Although several other populations of memory CD8+ T cells might also exist, the circulating pool of pathogen-specific CD8+ T cells present during the first several months after infection is composed of mainly three subpopulations: residual terminally differentiated KLRG1hi CD8+ T cells, KLRG1lo Tem cells derived from memory precursors, and Tcm cells derived from Tem cells (Figure 1). Zhou et al. (2010) demonstrated that without TCF-1, lower numbers of CD62Lhi Tcm cells were formed. In addition, the TCF-1-deficient CD8+ T cells expanded poorly after rechallenge. Impaired recall responses in TCF-1-deficient mice were also recently reported by Jeannet et al. (2010). These observations, together with the defective homeostatic maintenance reported by Zhou et al. (2010), are consistent with defects in Tcm cells in the absence of TCF-1. Conversely, both Zhou et al. (2010) and Jeannet et al. (2010) report a major increase in KLRG1hi effector CD8+ T cells (though the details differed slightly depending on the infection). These early alterations in the subsets of effector CD8+ T cells and the
Immunity
Previews later defects in Tcm cells suggest that TCF-1 could impact memory CD8+ T cell differentiation at several stages (Figure 1). First, TCF-1 might be important during the first several days after infection when early fate decisions between terminal effectors and memory precursors are made. TCF-1 is highly expressed in naive CD8+ T cells, but expression decreases during the first 5 days after infection (Zhao et al., 2010). Thus, the increase in KLRG1hi effector cells in the absence of TCF-1 suggests that TCF-1 favors the generation of memory precursors at the expense of terminally differentiated effector CD8+ T cells. Such a shift could explain the defective recall responses in the absence of TCF-1. However, KLRG1hi effector CD8+ T cells survive poorly into the memory phase in the absence of IL-15 signals (Joshi et al., 2007). Presumably, the ability to use IL-15 would be defective in TCF-1-deficient effector CD8+ T cells given the impact of TCF-1 on Eomes-dependent CD122 expression in memory CD8+ T cells, though this issue warrants further investigation. However, it is unlikely that the impact of TCF-1 deficiency on recall responses is entirely due to defective IL-15 signaling, because recall responses are intact in Il15 / mice (Becker et al., 2002). A role for changes in IL-2 signaling via CD122, which can impact CD8+ T cell memory, has also not been ruled out. In addition, although increasing Eomes expression improved survival of TCF-1-deficient memory CD8+ T cells, it is unclear: (1) whether retroviral expression of Eomes in effector cells would restore the normal distribution of effector subsets, (2) whether TCF-1 has additional Eomes-independent functions in effector CD8+ T cell lineage decisions, or (3) whether Eomes expression can be controlled in a TCF-1-independent manner at some stages. A second way TCF-1 deficiency could impact the CD8+ T cell memory is via differential survival. Although KLRG1lo memory CD8+ T cells appear less dependent on IL-15 to enter the memory pool compared the KLRG1hi subset (Joshi et al., 2007), it will be interesting to examine selective survival of TCF-1-deficient effector subpopulations during this transition. It is also possible that TCF-1
has a role in the establishment of the Tcm cell pool because either Tcm (or Tem) cells fail to survive in the absence of CD122, or because TCF-1 is important for Tem/Tcm cell conversion. This latter possibility is intriguing, because little is known about the transcriptional control of Tem/Tcm cell differentiation. Finally, it will be interesting to determine whether Wnt ligand availability is greater in anatomical locations associated with early fate decision checkpoints (e.g., in contact with dendritic cells) and/or memory CD8+ T cell homeostasis (e.g., the bone marrow). Zhou et al. (2010) demonstrate that TCF-1 binds to the regulatory regions of the Eomes gene. In the presence of Wnt signaling and b-catenin activation, TCF-1 induces Eomes expression. However, TCR signaling might disrupt Wnt signaling via TCF-1 downregulation. As antigen is cleared, TCF-1 expression rises and b-catenin-TCF-1 complexes can again form. It is interesting that Zhou et al. (2010) show both the corepressor TLE1 and b-catenin present with TCF-1 at the regulatory regions of the Eomes gene in naive CD8+ T cells. Because a deficiency in TCF-1 eliminates both activating TCF1-b-catenin and repressive TCF-1-TLE1 complexes, it will be important in the future to determine whether TCF-1dependent transcriptional repression also impacts effector and/or memory CD8+ T cell differentiation. In addition, the precise role of TCF-1 in the regulating Eomes before, during, and after activation, where Eomes cooperates with T-bet and Runx3 (Cruz-Guilloty et al., 2009), will be exciting to unravel. Cascades of interdependent transcription factor pathways are central to many developmental pathways including hematapoiesis (Laslo et al., 2006). The identification of a key role for TCF-1 in CD8+ T cell memory highlights how this transcription factor can toggle between different functional states based on the presence of b-catenin and can regulate a second transcription factor important for CD8+ T cell memory (Eomes). These data indicate that defining the networks of overlapping transcriptional modules controlling CD8+ T cell memory will be complex and will depend on the applica-
tion of genomic approaches, detailed loss- and gain-of-function studies, and careful consideration of the state of the cells being analyzed. This latter point is illustrated by the distinct functional TCF-1 complexes that might exist in naive, effector, memory CD8+ T cells given the presence of different repressive (TLE) or activating (b-catenin) cofactors. Ultimately, of course, the goals of such studies are to improve the ability to manipulate immunological memory, vaccines, and immunotherapies. Transcription factors as drugable targets are notoriously difficult. However, as illustrated by Zhou et al. (2010), when the details of the upstream biological pathways are understood in sufficient detail, it might be feasible to manipulate a transcription factor-mediated influence on T cell memory. As our understanding of the transcriptional circuitry of T cell memory improves, it should be possible to investigate ways to apply this information toward clinical immunomodulation. REFERENCES Becker, T.C., Wherry, E.J., Boone, D., MuraliKrishna, K., Antia, R., Ma, A., and Ahmed, R. (2002). J. Exp. Med. 195, 1541–1548. Cruz-Guilloty, F., Pipkin, M.E., Djuretic, I.M., Levanon, D., Lotem, J., Lichtenheld, M.G., Groner, Y., and Rao, A. (2009). J. Exp. Med. 206, 51–59. Intlekofer, A.M., Takemoto, N., Wherry, E.J., Longworth, S.A., Northrup, J.T., Palanivel, V.R., Mullen, A.C., Gasink, C.R., Kaech, S.M., Miller, J.D., et al. (2005). Nat. Immunol. 6, 1236–1244. Jeannet, G., Boudousquie, C., Gardiol, N., Kang, J., Huelsken, J., and Held, W. (2010). Proc. Natl. Acad. Sci. USA 107, 9777–9782. Joshi, N.S., Cui, W., Chandele, A., Lee, H.K., Urso, D.R., Hagman, J., Gapin, L., and Kaech, S.M. (2007). Immunity 27, 281–295. Kaech, S.M., and Wherry, E.J. (2007). Immunity 27, 393–405. Laslo, P., Spooner, C.J., Warmflash, A., Lancki, D.W., Lee, H.J., Sciammas, R., Gantner, B.N., Dinner, A.R., and Singh, H. (2006). Cell 126, 755–766. MacDonald, B.T., Tamai, K., and He, X. (2009). Dev. Cell 17, 9–26. Zhao, D.M., Yu, S., Zhou, X., Haring, J.S., Held, W., Badovinac, V.P., Harty, J.T., and Xue, H.H. (2010). J. Immunol. 184, 1191–1199. Zhou, X., Yu, S., Zhao, D.-M., Harty, J.T., Badovinac, V.P., and Xue, H.-H. (2010). Immunity 33, this issue, 229–240.
Immunity 33, August 27, 2010 ª2010 Elsevier Inc. 147
Previews a-Gal-A enzyme activity. Therefore, these findings suggested that microbial activation of DCs leads to an accumulation of glycolipids that are normally degraded by a-Gal-A (Figure 1, right panel). Inhibitor studies further suggested that iGb3 could well be a relevant glycolipid in this context. The idea that iGb3 might be the sole antigen that controls the intrathymic development of iNKT cells is particularly contentious (Gapin, 2010). Darmoise et al. (2010) were unable to identify a defect in intrathymic development of iNKT cells in Fabry mice, which is consistent with some, but not all, prior studies (e.g., Gadola et al., 2006). Nevertheless, they found that numbers of peripheral iNKT cells in these animals were substantially suppressed, which they attributed to increased death because of chronic activation in the periphery. Consistent with this possibility, the remaining iNKT cells in Fabry mice exhibited a hyporesponsive phenotype when stimulated with a-galactosylceramide, which was associated with increased expression of inhibitory NK cell receptors. Is this pathway for the generation of endogenous iNKT cell antigens in mice conserved in humans? Probably not. Studies with human iNKT cell clones have failed to find a role for the CD1d
recycling pathway (see Figure 1) in the autoreactivity of these cells (Chen et al., 2007). Further, as mentioned above, although ultrasensitive mass spectrometry studies were able to identify iGb3 in human thymus, the iGb3 synthase gene in humans appears to be defective. Lastly, patients with Fabry disease have normal numbers of iNKT cells. As with most provocative studies, the paper from Darmoise et al. (2010) raises more questions than it answers: What are the relevant iNKT cell self-antigens that accumulate in Fabry mice? Is the accumulation of iNKT cell antigens in Fabry mice unique to DCs, or does this also occur in other cell types such as the cells that mediate thymic selection of iNKT cells? What controls the transient inactivation of a-Gal-A enzyme activity in DCs stimulated with microbial products? How might the relatively modest suppression of a-Gal-A activity in response to microbes mediate such a profound effect on iNKT cell activation? Do the relevant self-antigens that accumulate in Fabry mice contribute to other pathological conditions that are influenced by iNKT cells? To what extent, if at all, is this pathway conserved in humans? Providing answers to these questions will likely keep iNKT cell biologists busy for the coming years.
REFERENCES Chen, X., Wang, X., Keaton, J.M., Reddington, F., Illarionov, P.A., Besra, G.S., and Gumperz, J.E. (2007). J. Immunol. 178, 6181–6190. Darmoise, A., Teneberg, S., Bouzonville, L., Brady, R.O., Beck, M., Kaufmann, S.H.E., and Winau, F. (2010). Immunity 33, this issue, 216–228. Gadola, S.D., Silk, J.D., Jeans, A., Illarionov, P.A., Salio, M., Besra, G.S., Dwek, R., Butters, T.D., Platt, F.M., and Cerundolo, V. (2006). J. Exp. Med. 203, 2293–2303. Gapin, L. (2010). Nat. Rev. Immunol. 10, 272–277. Li, Y., Thapa, P., Hawke, D., Kondo, Y., Kurukawa, K., Furukawa, K., Hsu, F.-F., Adlercreutz, D., Weadge, J., Palcic, M.M., et al. (2009). J. Proteome Res. 8, 2740–2751. Mattner, J., Debord, K.L., Ismail, N., Goff, R.D., Cantu, C., III, Zhou, D., Saint-Mezard, P., Wang, V., Gao, Y., Yin, N., et al. (2005). Science 434, 525–529. Muindi, K., Cernadas, M., Watts, G.F.M., Royle, L., Neville, D.C.A., Dwek, R.A., Besra, G.S., Rudd, P.M., Butters, T.D., and Brenner, M.B. (2010). Proc. Natl. Acad. Sci. USA 107, 3052– 3057. Stanic, A.K., De Silva, A.D., Park, J.-J., Sriram, V., Ichikawa, S., Hirabyashi, Y., Hayakawa, K., Van Kaer, L., Brutkiewicz, R.R., and Joyce, S. (2003). Proc. Natl. Acad. Sci. USA 100, 1849– 1854. Zhou, D., Mattner, J., Cantu, C., III, Schrantz, N., Yin, N., Gao, Y., Sagiv, Y., Hudspeth, K., Wu, Y.P., Yamashita, T., et al. (2004). Science 306, 1786–1789.
TCF-1 Flips the Switch on Eomes Michael A. Paley1 and E. John Wherry1,* 1Department of Microbiology and Institute for Immunology, University of Pennsylvania School of Medicine, Philadelphia, PA 19104, USA *Correspondence: [email protected] DOI 10.1016/j.immuni.2010.08.008
The interaction between different transcriptional pathways in CD8+ T cell memory remains incompletely understood. In this issue, Zhou et al. (2010) demonstrate an important role for T cell factor 1 in regulating CD8+ T cell memory via control of a second transcription factor, Eomesodermin. There are clear parallels between the development of adaptive immune responses and other developmental systems. The Wnt-b-catenin pathway is important in many of these development processes. Although canonical Wnt signaling has a role in thymic development and has been reported to influence CD8+ T cell memory
(Zhao et al., 2010), precisely how this morphogenic signaling pathway is linked to the transcriptional regulation of T cell memory has remained unclear. Normally, in the absence of Wnt ligands, b-catenin is sequestered and rapidly degraded in the cytoplasm. Wnt signaling stabilizes b-catenin, allowing translocation to the
nucleus and interaction with T cell factor 1 (TCF-1). b-catenin displaces TLE1, a Groucho family corepressor bound to TCF-1, converting the repressive TCF-1-TLE1 complex into a TCF-1-b-catenin transcriptional activator (MacDonald et al., 2009). In this issue, Zhou et al. (2010) demonstrate that TCF-1 activation by
Immunity 33, August 27, 2010 ª2010 Elsevier Inc. 145
Immunity
Previews Naive
Effector
Memory 3 TCF-1?
2 TCF-1?
4
TCF-1? Tem
?
TCF-1 (Eomes)
Terminal effector
1 TCF-1?
KLRG1 CD127lo CD62Llo hi
2 TCF-1?
Tcm
KLRG1lo KLRG1lo CD127hi CD127hi CD62Llo CD62Lhi
TCF-1 Eomes
Death
Figure 1. Development of the CD8 T Cell Response to Viral Infection
Upon activation, naive CD8+ T cells proliferate and differentiate into two subsets: terminal effectors (red) and memory precursors (yellow). While terminal effectors undergo cell death, memory precursors mature into Tem-like cells (green) as they enter the memory phase. Over time, these Tem cells further differentiate into Tcm cells (blue), which are efficiently maintained by IL-7- and IL-15-driven self-renewal. TCF-1 might regulate this developmental process during (1) early lineage decisions, (2) survival during the contraction phase, (3) differentiation from Tem to Tcm cells, or (4) self-renewal of Tcm cells.
Wnt-b-catenin signaling is a key event in the transcriptional induction of Eomes, a transcription factor involved in CD8+ T cell memory (Intlekofer et al., 2005). Although some alterations in the primary CD8+ T cell response to Listeria monocytogenes were observed, a major effect of TCF-1 deficiency was the failure to acquire canonical memory CD8+ T cells properties, notably homeostatic maintenance. This defect was more severe in lymphoid organs, and TCF-1-deficient memory CD8+ T cells remained CD62Llo and CCR7lo, consistent with poor development or maintenance of ‘‘central memory’’ CD8+ T (Tcm) cells. Without TCF-1, expression of IL-15Rb (CD122) was reduced and TCF-1-deficient memory CD8+ T cells had defective homeostatic proliferation consistent with impaired IL-15 responsiveness. Eomes has been previously implicated in CD8+ T cell memory via CD122 expression (Intlekofer et al., 2005). Zhou et al. (2010) now show that Eomes expression was markedly reduced in the absence of TCF-1. Retroviral overexpression of Eomes in TCF-1deficient cells restored CD122 expression and improved survival, demonstrating that TCF-1-dependent expression of Eomes contributed to memory CD8+ T cell maintenance.
One interesting set of observations from this work and other recent studies (Jeannet et al., 2010) is the potential role of TCF-1 at different points during CD8+ T cell activation and differentiation. Normally, after acute infection, distinct stages of memory CD8+ T cell differentiation exist. After initial stimulation, naive CD8+ T cells become activated, proliferate, acquire effector functions, and give rise to two subsets of effector CD8+ T cells. KLRG1hiCD127lo CD8+ T cells are potent effector CD8+ T cells but are terminally differentiated. These cells cannot give rise to self-renewing, stable populations of memory CD8+ T cells and their numbers decline rapidly during contraction, but then more slowly in the early memory phase (Joshi et al., 2007). These cells also proliferate poorly upon rechallenge (Kaech and Wherry, 2007). In contrast, KLRG1loCD127hi CD8+ T cells can further differentiate into long-lived, homeostatically maintained memory CD8+ T cells. Initially, the memory population formed by the KLRG1loCD127hi memory precursors is CD62Llo and resembles an effector memory T (Tem) cell-like population. Over time, these Tem cells can give rise to CD62Lhi Tcm-like cells capable of robust proliferative expansion after rechallenge (Kaech and Wherry,
146 Immunity 33, August 27, 2010 ª2010 Elsevier Inc.
2007). It is also these Tcm cells that can efficiently use IL-7 and IL-15 and undergo homeostatic proliferation (Kaech and Wherry, 2007). Although several other populations of memory CD8+ T cells might also exist, the circulating pool of pathogen-specific CD8+ T cells present during the first several months after infection is composed of mainly three subpopulations: residual terminally differentiated KLRG1hi CD8+ T cells, KLRG1lo Tem cells derived from memory precursors, and Tcm cells derived from Tem cells (Figure 1). Zhou et al. (2010) demonstrated that without TCF-1, lower numbers of CD62Lhi Tcm cells were formed. In addition, the TCF-1-deficient CD8+ T cells expanded poorly after rechallenge. Impaired recall responses in TCF-1-deficient mice were also recently reported by Jeannet et al. (2010). These observations, together with the defective homeostatic maintenance reported by Zhou et al. (2010), are consistent with defects in Tcm cells in the absence of TCF-1. Conversely, both Zhou et al. (2010) and Jeannet et al. (2010) report a major increase in KLRG1hi effector CD8+ T cells (though the details differed slightly depending on the infection). These early alterations in the subsets of effector CD8+ T cells and the
Immunity
Previews later defects in Tcm cells suggest that TCF-1 could impact memory CD8+ T cell differentiation at several stages (Figure 1). First, TCF-1 might be important during the first several days after infection when early fate decisions between terminal effectors and memory precursors are made. TCF-1 is highly expressed in naive CD8+ T cells, but expression decreases during the first 5 days after infection (Zhao et al., 2010). Thus, the increase in KLRG1hi effector cells in the absence of TCF-1 suggests that TCF-1 favors the generation of memory precursors at the expense of terminally differentiated effector CD8+ T cells. Such a shift could explain the defective recall responses in the absence of TCF-1. However, KLRG1hi effector CD8+ T cells survive poorly into the memory phase in the absence of IL-15 signals (Joshi et al., 2007). Presumably, the ability to use IL-15 would be defective in TCF-1-deficient effector CD8+ T cells given the impact of TCF-1 on Eomes-dependent CD122 expression in memory CD8+ T cells, though this issue warrants further investigation. However, it is unlikely that the impact of TCF-1 deficiency on recall responses is entirely due to defective IL-15 signaling, because recall responses are intact in Il15 / mice (Becker et al., 2002). A role for changes in IL-2 signaling via CD122, which can impact CD8+ T cell memory, has also not been ruled out. In addition, although increasing Eomes expression improved survival of TCF-1-deficient memory CD8+ T cells, it is unclear: (1) whether retroviral expression of Eomes in effector cells would restore the normal distribution of effector subsets, (2) whether TCF-1 has additional Eomes-independent functions in effector CD8+ T cell lineage decisions, or (3) whether Eomes expression can be controlled in a TCF-1-independent manner at some stages. A second way TCF-1 deficiency could impact the CD8+ T cell memory is via differential survival. Although KLRG1lo memory CD8+ T cells appear less dependent on IL-15 to enter the memory pool compared the KLRG1hi subset (Joshi et al., 2007), it will be interesting to examine selective survival of TCF-1-deficient effector subpopulations during this transition. It is also possible that TCF-1
has a role in the establishment of the Tcm cell pool because either Tcm (or Tem) cells fail to survive in the absence of CD122, or because TCF-1 is important for Tem/Tcm cell conversion. This latter possibility is intriguing, because little is known about the transcriptional control of Tem/Tcm cell differentiation. Finally, it will be interesting to determine whether Wnt ligand availability is greater in anatomical locations associated with early fate decision checkpoints (e.g., in contact with dendritic cells) and/or memory CD8+ T cell homeostasis (e.g., the bone marrow). Zhou et al. (2010) demonstrate that TCF-1 binds to the regulatory regions of the Eomes gene. In the presence of Wnt signaling and b-catenin activation, TCF-1 induces Eomes expression. However, TCR signaling might disrupt Wnt signaling via TCF-1 downregulation. As antigen is cleared, TCF-1 expression rises and b-catenin-TCF-1 complexes can again form. It is interesting that Zhou et al. (2010) show both the corepressor TLE1 and b-catenin present with TCF-1 at the regulatory regions of the Eomes gene in naive CD8+ T cells. Because a deficiency in TCF-1 eliminates both activating TCF1-b-catenin and repressive TCF-1-TLE1 complexes, it will be important in the future to determine whether TCF-1dependent transcriptional repression also impacts effector and/or memory CD8+ T cell differentiation. In addition, the precise role of TCF-1 in the regulating Eomes before, during, and after activation, where Eomes cooperates with T-bet and Runx3 (Cruz-Guilloty et al., 2009), will be exciting to unravel. Cascades of interdependent transcription factor pathways are central to many developmental pathways including hematapoiesis (Laslo et al., 2006). The identification of a key role for TCF-1 in CD8+ T cell memory highlights how this transcription factor can toggle between different functional states based on the presence of b-catenin and can regulate a second transcription factor important for CD8+ T cell memory (Eomes). These data indicate that defining the networks of overlapping transcriptional modules controlling CD8+ T cell memory will be complex and will depend on the applica-
tion of genomic approaches, detailed loss- and gain-of-function studies, and careful consideration of the state of the cells being analyzed. This latter point is illustrated by the distinct functional TCF-1 complexes that might exist in naive, effector, memory CD8+ T cells given the presence of different repressive (TLE) or activating (b-catenin) cofactors. Ultimately, of course, the goals of such studies are to improve the ability to manipulate immunological memory, vaccines, and immunotherapies. Transcription factors as drugable targets are notoriously difficult. However, as illustrated by Zhou et al. (2010), when the details of the upstream biological pathways are understood in sufficient detail, it might be feasible to manipulate a transcription factor-mediated influence on T cell memory. As our understanding of the transcriptional circuitry of T cell memory improves, it should be possible to investigate ways to apply this information toward clinical immunomodulation. REFERENCES Becker, T.C., Wherry, E.J., Boone, D., MuraliKrishna, K., Antia, R., Ma, A., and Ahmed, R. (2002). J. Exp. Med. 195, 1541–1548. Cruz-Guilloty, F., Pipkin, M.E., Djuretic, I.M., Levanon, D., Lotem, J., Lichtenheld, M.G., Groner, Y., and Rao, A. (2009). J. Exp. Med. 206, 51–59. Intlekofer, A.M., Takemoto, N., Wherry, E.J., Longworth, S.A., Northrup, J.T., Palanivel, V.R., Mullen, A.C., Gasink, C.R., Kaech, S.M., Miller, J.D., et al. (2005). Nat. Immunol. 6, 1236–1244. Jeannet, G., Boudousquie, C., Gardiol, N., Kang, J., Huelsken, J., and Held, W. (2010). Proc. Natl. Acad. Sci. USA 107, 9777–9782. Joshi, N.S., Cui, W., Chandele, A., Lee, H.K., Urso, D.R., Hagman, J., Gapin, L., and Kaech, S.M. (2007). Immunity 27, 281–295. Kaech, S.M., and Wherry, E.J. (2007). Immunity 27, 393–405. Laslo, P., Spooner, C.J., Warmflash, A., Lancki, D.W., Lee, H.J., Sciammas, R., Gantner, B.N., Dinner, A.R., and Singh, H. (2006). Cell 126, 755–766. MacDonald, B.T., Tamai, K., and He, X. (2009). Dev. Cell 17, 9–26. Zhao, D.M., Yu, S., Zhou, X., Haring, J.S., Held, W., Badovinac, V.P., Harty, J.T., and Xue, H.H. (2010). J. Immunol. 184, 1191–1199. Zhou, X., Yu, S., Zhao, D.-M., Harty, J.T., Badovinac, V.P., and Xue, H.-H. (2010). Immunity 33, this issue, 229–240.
Immunity 33, August 27, 2010 ª2010 Elsevier Inc. 147