Robin D. Hatton and Casey T. Weaver*
T-bet or Not T-bet
T and B lymphocytes have commandeered well-established developmental strategies to ensure their efficient response to antigens. This is evident during the development of effector CD4 and CD8 T cells when antigen-naïve T cells are reprogrammed to express new genes that arm them for heightened immune defense against pathogens. Key transcription factors regulating the development of CD4 effector T cells have been identified, whereas those involved in CD8 effector T cell development are less well understood. This disparity is addressed by Pearce et al. (1) on page 1041 of this issue with their introduction of a major player in the development of CD8 T cells.
In the adaptive immune system, antigens derived from pathogens stimulate the clonal proliferation of T lymphocyte precursors that carry receptors specific for that antigen. In response to different pathogen-derived antigens, CD4 T lymphocytes become either T helper 1 (TH1) or T helper 2 (TH2) cells depending on the type of pathogen (see the figure). TH1 cells produce the cytokine interferon- (IFN-) and eradicate intracellular bacteria, viruses, and protozoa, whereas TH2 cells produce interleukin-4 (IL-4), IL-5, and IL-13 and eliminate extracellular parasites (2). Like TH1 cells, CD8 T cells are activated in response to intracellular pathogens and share some of the same effector mechanisms, particularly production of IFN-
A division of labor. Type 1 pathogens (viruses and intracellular bacteria) induce a TH1 immune response mediated by CD4 cells, and a CD8 effector T cell response. In both responses, IFN-is produced. CD8 effector T cell development is regulated by the T-box transcription factors, T-bet and Eomesodermin (Eomes), whereas CD4 TH1 cell development is regulated by T-bet alone. Type 2 pathogens (helminth parasites) induce CD4 TH2 cell development, which is regulated by the transcription factor GATA-3. Neither the signals that specifically induce expression of T-bet and Eomesodermin in developing CD8 T cells (question mark) nor the gene targets of these factors have been identified. Ag, antigen.
CREDIT: KATHARINE SUTLIFF/SCIENCE
Cytokines are the dominant factors guiding the development of TH1 and TH2 cells (3, 4), and their receptor and signaling pathways have been characterized. Important proximal signals in TH1 lineage commitment (5) include the IFN- and IL-12 receptors, which use the signaling components STAT1 and STAT4, respectively. TH2 commitment is induced by IL-4, which uses STAT6 for signaling. Cytokine signaling initiates effector T cell development through induction of transcription factors that may act as "master regulators." GATA-3 has been identified as a key regulator of TH2 development (6). Activation of GATA-3 switches on genes in the TH2 cytokine cluster (Il4, Il5, and Il13), both by initiating the remodeling of local chromatin and by acting as an acute transcription factor (5, 7). Importantly, GATA-3 autoamplifies its own expression while down-regulating expression of STAT4 and the
2 chain of the IL-12 receptor (IL-12R2). Consequently, low-level induction of GATA-3 early in TH2 development can stabilize TH2 lineage specification by simultaneously enforcing TH2 development and blunting TH1 development.
TH1 development is guided by its own master called T-bet (T-box expressed in T cells; also known as Tbx-21) (8). In contrast to GATA-3, T-bet is detected early in developing TH1 cells, but not in TH2 cells. Like GATA-3, T-bet is a member of a family of transcription factors that shares a highly conserved DNA binding domain, the T-box. The T-box family is phylogenetically ancient and includes more than 50 members (9). The highly conserved DNA binding domain of the T-box family contrasts with the great diversity of non-DNA binding domains, which are involved in transcriptional activation or repression. Because all known members bind to the same core DNA sequence, specificity is likely to be determined by association with different cofactors. These have yet to be defined for T cell development, but intriguingly there may be a direct interaction between factors from the T-bet and GATA families (10).
Like GATA-3, T-bet is either absent or expressed at low levels by naïve CD4 T cells. However, when naïve CD4 T cells recognize antigen in the presence of IFN-, coordinate signaling is activated through the T cell receptor and STAT1, respectively, resulting in increased expression of T-bet. Elevated T-bet induces IL-12R2 expression, thereby enabling induction of STAT4 activation by IL-12, which, in turn, induces the expression of IFN- and the IL-18 receptor. Because IL-12 and IL-18 receptor signaling both act to augment IFN- expression by T cells, their sequential induction is a powerful mechanism for amplifying effector differentiation initiated by T-bet. Because IFN- may feed back to augment T-bet expression either in the same cell that produces it or in developing TH1 neighbors, a cell-extrinsic amplification loop exists to stabilize T-bet expression and the TH1 differentiation program. Like GATA-3, T-bet also seems to positively regulate its own expression through a cell-intrinsic mechanism involving the homeobox factor Hlx, which is itself induced by T-bet (11). Although it remains to be determined whether T-bet can directly activate the transcription of Ifng in vivo, multiple consensus sites have been identified in the Ifng locus, and are likely targets for T-bet-mediated chromatin remodeling, acute transcription, or both (12).
Mice lacking T-bet reveal an unusual feature of T-bet ecology. Although they have a profound defect in TH1 development, the development and function of CD8 effector T cells remains relatively unperturbed (13). This hinted that a T-bet-independent mechanism might regulate IFN- expression and cytolytic activity in CD8 T cells even though paradoxically these cells usually express T-bet. CD8 T cells forced to produce more T-bet show increased production of IFN- and of the cytolytic effector proteins perforin and granzyme B. Why, then, do the CD8 cells of mice lacking T-bet show unabated cytolytic activity and IFN- production? Is T-bet not the master in the CD8 cell household?
In addressing this question, Pearce, Reiner, and their co-workers discovered a second member of the T-box gene family, Eomesodermin, which is expressed in developing CD8 T cells (1). Surprisingly, a dominant-negative form of T-bet that blocks T-bet activity, also suppresses effector responses in CD8 cells from T-bet-deficient mice. This implies that CD8 T cell development is regulated by a T-bet-independent pathway that is sensitive to dominant-negative T-bet. Reasoning that other T-box family members in CD8 cells must be suppressed by a dominant-negative form of T-bet, the authors used degenerate oligonucleotides specific for the T-box domain to amplify cDNAs from recently activated CD8 T cells. Analysis of the amplified cDNAs revealed roughly equal representation of two T-box genes: T-bet and Eomesodermin. Subsequent gain-of-function experiments established that enforced expression of Eomesodermin in T-bet-deficient CD8 cells could restore CD8 effector development, that is, IFN- production and expression of perforin and granzyme B. Interestingly, enforced expression of Eomesodermin in CD4 T cells from T-bet-deficient mice also restored IFN- production and the TH1 response. There was a similar response in CD4 TH2 cells transfected with Eomesodermin. Importantly, although activated CD4 and CD8 T cells both expressed T-bet, only CD8 cells expressed high levels of Eomesodermin transcripts. Thus, although enforced expression of Eomesodermin in T-bet-deficient CD4 T cells can restore a TH1 immune response, it seems unlikely that Eomesodermin plays a prominent part in the normal development of CD4 T cells. In contrast, Eomesodermin may play a complementary, if not dominant, role in the development of CD8 effector cells.
Although the Pearce et al. study offers an explanation for T-bet-independent regulation of CD8 cells producing IFN-, it does not preclude a role for T-bet in this process. It does, however, raise important questions about how distinct T-box factors may influence effector T cell development. Why are both T-bet and Eomesodermin expressed in developing CD8 T cells when only T-bet is expressed in developing CD4 T cells? Does this simply reflect redundancy and plasticity in the regulation of IFN-, or is there a division of labor between these T-box factors with respect to the target genes they regulate? The latter possibility is suggested by the greater reduction in expression of genes associated with cytolysis compared with those involved in IFN- production in mice expressing only half the usual amount of Eomesodermin. In addition, dominant-negative forms of both Eomesodermin and T-bet result in a greater reduction in CD8 cytolytic activity than does dominant-negative T-bet alone. Notably, CD8 T cells from STAT4-deficient mice continue to produce IFN-. Does Eomesodermin amplify and maintain expression of IFN- through a STAT4-independent mechanism? CD8 T cells acquire effector competency more rapidly than do CD4 T cells; does coexpression of Eomesodermin and T-bet facilitate this? Alternatively, does the coexpression of both Eomesodermin and T-bet reflect their activation by distinct upstream pathways that are both important for CD8 effector T cell development? For instance, although type I interferons are known to activate STAT1 signaling, they do not induce T-bet transcription (14); does Eomesodermin complement this function in CD8 cells, and, if so, to what end? Finally, is the apparent coexpression of T-bet and Eomesodermin an artifact of a population-based analysis, and is the true situation that these factors are expressed by distinct subpopulations of activated CD8 cells?
Clearly, definitive answers to these and other questions will require further study and better tools for their analysis, such as the production of Eomesodermin and T-bet conditional knockout mice. Nevertheless, the Pearce et al. study provides new insights into transcriptional events that operate during CD8 effector T cell development. Attention can now be centered on comparisons between CD8 and CD4 effector systems, which will benefit our understanding of both.
- E. L. Pearce et al., Science 302, 1041 (2003).
- T. R. Mosmann, R. L. Coffman, Annu. Rev. Immunol. 7, 145 (1989) [Medline].
- C. S. Hsieh, A. B. Heimberger, J. S. Gold, A. O'Garra, K. M. Murphy, Proc. Natl. Acad. Sci. U.S.A. 89, 6065 (1992) [Medline].
- C. S. Hsieh et al., Science 260, 547 (1993) [Medline].
- K. M. Murphy, S. L. Reiner, Nature Rev. Immunol. 2, 933 (2002) [Medline].
- W. Zheng, R. A. Flavell, Cell 89, 587 (1997) [Medline].
- S. Agarwal, A. Rao, Immunity 9, 765 (1998) [Medline].
- S. J. Szabo et al., Cell 100, 655 (2000) [Medline].
- V. Wilson, F. L. Conlon, Genome Biol. 3, 3008 (2002) [Medline].
- V. Garg et al., Nature 424, 443 (2003) [Medline].
- A. C. Mullen et al., Nature Immunol. 3, 652 (2002) [Medline].
- J. Y. Cho, V. Grigura, T. L. Murphy, K. Murphy, Int. Immunol. 15, 1149 (2003) [Medline].
- S. J. Szabo et al., Science 295,  (2002).
- A. A. Lighvani et al., Proc. Natl. Acad. Sci. U.S.A. 98, 15137 (2001) [Medline].
The authors are in the Department of Pathology, University of Alabama at Birmingham, Birmingham, AL 35294, USA.
Include this information when citing this paper.
Issue of 7 Nov 2003,
Copyright © 2003 by The American Association for the Advancement of Science. All rights reserved.