Two Receptors, Two Kinases, and T Cell Lineage Determination

+ See all authors and affiliations

Science Signaling  23 Mar 2010:
Vol. 3, Issue 114, pp. pe11
DOI: 10.1126/scisignal.3114pe11


The T cell antigen receptor (TCR) serves as a paradigm for how membrane receptors transmit signals to the cytoplasm because it controls many aspects of T cell differentiation and function by detecting atom-sized variations in the quality of the ligand that is recognized. The mechanisms that underlie the different signaling outcomes are unclear. Studies that suggest a ligand-tailored, qualitatively different signal are confronted with evidence that favors a quantitative model, and studies of TCR-dependent T cell differentiation in the thymus are no exception. Mature T cells with an αβ TCR are classified according to two major distinct subsets based on the mutually exclusive presence of the co-receptors CD4 and CD8, which play essential roles in recognition of the major histocompatibility complex (MHC) class II and I ligands, respectively, and in the recruitment of the tyrosine kinase Lck to the TCR complex. Mature CD4+ and CD8+ T cells derive from a common precursor in the thymus, a double-positive (DP) thymocyte, which has both co-receptors. Early signaling models suggested that the differential capacity of CD4 and CD8 to recruit Lck to the TCR underlay lineage decision. A study now shows that differentiation into the CD8+ lineage requires the TCR-induced increased abundance of the tyrosine kinase ζ chain–associated protein kinase of 70 kD (Zap70). This finding, together with the known importance of Lck in the determination of CD4+ and CD8+ lineages, enables us to propose that a balance between the activation of these two kinases by the TCR determines lineage decisions.

Immunologists love to give high-sounding terms to processes driven by thermodynamic laws. The recognition of ligand by the T cell antigen receptor (TCR) and the activation of intracellular signaling pathways have led to the coining of such terms as “positive selection,” “negative selection,” “serial triggering,” “instructive process,” and “kinetic signaling.” This tendency emerges from the peculiarity of ligand-recognition by the TCR. No other receptor, apart from the B cell antigen receptor, is confronted by such an intense scrutiny for fitness and signaling capabilities during cell differentiation. The ligand of the TCR for the majority of T cells (called αβ T cells) is another membrane protein found on the surface of antigen-presenting cells (APCs). This ligand is a molecule of the major histocompatibility complex (MHC) that bears an elongated cavity through which peptides, of some 8 to 14 amino acid residues that result from the degradation of self proteins and antigens, are presented to the TCR. The TCR recognizes the surface created by the MHC molecule and the presented peptide, which is sometimes antigenic. Most of the interacting surface of the antigen-MHC complex is contributed by amino acid residues of the MHC molecule. This means that all T cells are to a certain extent self-reactive—that is, they react to MHC molecules found on cells of the same individual. Indeed, during differentiation, those immature T cells (thymocytes) that do not have an appropriate TCR capable of recognizing MHC that presents self peptides die in the thymus by apoptosis because the TCR transmits survival signals to the T cell only if it binds to self peptide–MHC complexes with sufficient affinity.

Having survived this stage (the process of positive selection), the immature thymocyte undergoes additional differentiation stages in which the TCR also plays an essential role. These stages can be defined according to the presence of two other membrane proteins: the co-receptors CD4 and CD8. Mature αβ T cells belong to two major types: CD4+ T cells, which recognize antigenic peptides presented by MHC class II molecules, and CD8+ T cells, which recognize antigenic peptides presented by MHC class I molecules. Whereas CD4+ T cells have a predominantly regulatory role, CD8+ T cells are essentially responsible for killing infected cells—that is, they are cytotoxic. Immature thymocytes go through a double positive (DP) stage in which they possess both co-receptors. At this point, if the TCR provides positive-selecting signals, the DP thymocytes differentiate into either mature CD4+ thymocytes or mature CD8+ thymocytes. The “decision” to differentiate into one lineage or the other is again taken according to the signals that emanate from the activated TCR in conjunction with the signals provided by the co-receptors. The co-receptors help in the recognition of the ligand by binding to a site in MHC that is distinct from the TCR-binding site. CD4 interacts with MHC class II and CD8 interacts with MHC class I.

One of the puzzling questions that has been attracting the attention of immunologists is how an immature DP thymocyte “knows” whether its TCR is recognizing MHC class II, which would instruct the thymocyte to differentiate into a mature CD4+ T cell, or whether its TCR is recognizing MHC class I, which would result in the thymocyte becoming a mature CD8+ T cell. The simplest solution to this problem would be an instructive mechanism in which recognition of MHC by either CD4 or CD8 modulates the TCR signal in such a way that qualitatively different signaling pathways are activated that lead to differentiation to either lineage (1, 2). There are a number of transcription factors that are important for lineage determination; elimination of ThPOK, TOX, or GATA-3 selectively impairs differentiation to CD4+ T cells, whereas elimination of Runx3 or Runx1 selectively affects CD8+ T cells (39). Qualitatively different signaling pathways could differentially induce expression of these factors. However, most data do not support such a model (10, 11). Indeed, CD4 and CD8 are not known to recruit distinct cytoplasmic effectors. Instead, both molecules recruit the tyrosine kinase Lck, which plays an important role in the initiation of the TCR signaling cascade by phosphorylating tyrosine residues in the cytoplasmic tails of some of the subunits of the TCR complex.

The finding that CD4 has a higher affinity for Lck than does CD8 led to the proposal that it is the intensity of the signal provided by the TCR–co-receptor complex that determines whether the cell will adopt a CD4+ or CD8+ lineage (1215). Because of the better recruitment of Lck by CD4, the TCR signal aided by CD4 would be of higher intensity than that of the TCR-CD8 complex; however, this scenario was complicated by the finding that DP immature thymocytes do not differentiate directly into mature CD4+ or CD8+ thymocytes. Instead, DP thymocytes down-regulate CD8 to become CD4+CD8low thymocytes, an intermediate population common to both lineages (16). The existence of this intermediate population could exacerbate the difference between MHC class II and MHC class I recognition, because whereas the recognition of MHC class II would lead to an intense signal, the recognition of MHC class I would lead to a signal even weaker than that at the DP stage. This finding therefore favored the “signal strength model” but brought into play another hypothesis, that of “co-receptor reversal” or “kinetic signaling” (1618). According to this model, it is the duration of TCR signaling that matters most. DP cells that bear a TCR capable of recognizing MHC class II receive continuous signals that lead to differentiation through the CD4+CD8low intermediate stage into mature CD4+ thymocytes, whereas DP thymocytes that bear a TCR that recognizes MHC class I undergo a “break” in positive selection—that is, an interruption in TCR signaling, which results from the down-regulation of the CD8 co-receptor at the CD4+CD8low stage (16). It is thought that this interruption in TCR signaling would likely lead to commitment of the cell to the CD8+ T cell lineage, which would result in the down-regulation of CD4 and the reexpression of CD8.

With regard to the “kinetic signaling” model, the group of Singer and colleagues has shown that differentiation into the CD8 lineage is dependent on signaling through the interleukin 7 (IL-7) receptor (IL-7R), which induces expression of the lineage-specific transcription factor Runx3 (19). The authors wish to suggest that CD8 lineage commitment is determined by cytokine signaling. By artificially forcing constitutive activity of the IL-7R–dependent signaling pathway, the authors observed the differentiation of CD8+ cells in the absence of the tyrosine kinase ζ chain–associated protein kinase of 70 kD (Zap70) protein, which they interpret to mean that TCR-mediated signals are unnecessary for CD8 lineage commitment. However, these authors have not demonstrated that acquisition of the IL-7–responsive status by DP thymocytes is independent of the TCR.

Saini et al. (20) have now carefully sought to demonstrate the “signal strength” (21) and “kinetic signaling” (16) models. The authors took advantage of mice that lack the Zap70, which is essential for TCR signaling. Thymocytes in these mice are blocked at the DP stage because these immature thymocytes cannot receive the signals that are needed to induce positive selection. Through an exogenously controlled gene expression system, the authors could induce the expression of the gene encoding Zap70 at will, thus rescuing thymic differentiation beyond the DP stage. The authors found that mature CD4+ and CD8+ thymocytes emerged from the population of DP cells in a marked chronological order: CD4+ thymocytes developed first and CD8+ thymocytes second, with a lag time of 2 to 5 days. Furthermore, by a careful analysis of the abundance of CD5 and the TCR within cells of the DP population, in combination with adoptive-transfer approaches in which an ex vivo–purified cell population was injected into the thymus of a recipient mouse, Saini et al. (20) found that CD4+ thymocytes developed directly from a DP subpopulation, termed DP2, which had a CD5highTCRmed phenotype. CD8+ thymocytes, however, developed from another population, DP3, which had a CD5medTCRhigh phenotype and was derived from the DP2 population.

Perhaps the most important finding resulted from the measurement of the abundance of Zap70 in the different subpopulations. The amount of Zap70 protein increased as the DP thymocytes matured from the DP1 stage (CD5lowTCRlow) through the DP2 and DP3 stages. This result suggests that the increased abundance of Zap70, which is developmentally controlled (22), and, as shown by Saini et al., is dependent on TCR signaling, is a mechanism to endow thymocytes with higher sensitivity to ligation of the TCR with peptide. These findings therefore support both the “signal strength” and the “kinetic signaling” models, because a strong TCR-CD4 signal (resulting from an interaction with MHC class II) would rapidly lead to differentiation to the CD4+ lineage, whereas differentiation to the CD8+ lineage would require the induced accumulation of Zap70 protein. Accordingly, expression of Zap70, which is induced by TCR signaling, would set a timer for the CD8+ lineage. The authors propose that the regulation of the abundance of Zap70 by TCR signaling can serve as a proofreading mechanism for positive selection, at least for CD8+ T cells, because only those T cells that bear a TCR competent to recognize MHC class I with sufficient affinity will be able to increase the abundance of Zap70 to the extent necessary to complete differentiation.

Having said that the study of Saini et al. (20) supports both the “signal strength” and the “kinetic signaling” models, it also contains additional findings that were not predicted by these hypotheses. First, the study establishes a clear precursor-progeny relationship between DP2 and CD4+ thymocytes and between DP3 and CD8+ thymocytes. This finding questions the apparent homogeneity of the CD4+CD8low intermediate population that was previously proposed (16). If both DP2 and DP3 precursors down-regulate CD8 to become CD4+CD8low, the DP2-CD4+ and DP3-CD8+ relationships would indicate that the intermediate CD4+CD8low population is formed by thymocytes already committed to either lineage. Second, TCR triggering in the DP3 population gave stronger responses than in the DP2 population. This contradicts the “signal strength” hypothesis at least in appearance. TCR triggering in DP2 and DP3 populations was produced by antibody-mediated crosslinking, not by peptide-MHC complexes. We should wait to see whether DP3 precursors still respond better than do DP2 precursors to the physiologic TCR ligand. Third, and perhaps most important, the need for an increase in the abundance of Zap70 for commitment to the CD8+ lineage suggests that for an immature DP precursor, lineage decision is a matter not only of the intensity of the signal but also of its quality. The “signal strength” hypothesis was proposed because Lck binds more strongly to CD4 than to CD8; however, Lck-deficient mice are more defective in the generation of CD4+ T cells than that of CD8+ T cells, and restoration of Lck by transgenic approaches has a stronger impact on CD4+ cells than on CD8+ T cells (13). This indicates a special sensitivity of the CD4+ lineage to Lck signaling beyond what was predicted by a model purely based on signal strength. On the other hand, there is a form of human severe combined immunodeficiency caused by mutations in the kinase domain of Zap70 that affect the stability of the protein (2325). These patients are devoid of CD8+ T cells but have abundant, although nonfunctional, CD4+ T cells. Thus, Zap70 is more important for the development of the CD8 lineage than for development of the CD4 lineage.

Unlike in humans, depletion of Zap70 in mice leads to defective formation of cells of both lineages (24). This human-mouse divergence could indicate a different requirement for Zap70 during lineage commitment (for example, at the DP2 and DP3 stages) or at an earlier step (for instance, during positive selection). Taking into account the dependence of the CD4+ lineage on Lck and the dependence of the CD8+ lineage on Zap70 (20), one could propose a part quantitative, part qualitative model of TCR signaling in which engagement of TCR-CD4 by self peptide–loaded MHC class II at the DP2 stage delivers a strong Lck signal but a weak Zap70 signal, which drives differentiation to the CD4+ lineage. In the absence of a strong Lck signal, DP2 cells differentiate into DP3 cells, a process that is accompanied by the increased abundance of Zap70. The increased Zap70-dependent signal, together with weak Lck signal, would lead to CD8+ lineage commitment (Fig. 1).

Fig. 1

A semiqualitative model of CD4-CD8 lineage determination by TCR signals. Immature CD4+CD8+ (DP) thymocytes receive TCR signals that result from the triggering of the TCR by MHC class I or class II molecules with help from the CD8 or CD4 co-receptors, respectively. These signals are in part mediated by the activation of the tyrosine kinases Lck and Zap70 and result in the positive selection of DP thymocytes that have an appropriate TCR. According to Saini et al. (20), positive selection leads to the increased abundance of CD5 and TCR by the most immature (DP1) thymocytes and their transition to DP2 thymocytes. At this stage, if the TCR on the DP2 thymocyte recognizes MHC class II, the cells receive a strong Lck-dependent signal that is provided by engagement of MHC class II by the CD4 co-receptor. This signal drives DP2 thymocytes to fully differentiate into mature CD4+ thymocytes, presumably through a CD4+CD8low transient stage. However, if the TCR on the DP2 thymocyte recognizes MHC class I, the TCR, aided by the binding of CD8, does not provide such a strong signal, thus preventing differentiation into the CD4 lineage. During the transition from the DP2 to the DP3 stage, however, the amount of Zap70 increases, whereas the amount of CD5, a negative regulator of TCR-mediated signaling, is reduced. When the abundance of Zap70 protein is sufficiently high, this leads to a Zap70-dependent TCR signal that drives differentiation of DP3 cells to mature CD8+ thymocytes, presumably through an intermediate CD4+CD8low stage. At the DP3 stage, Zap70 might be required either as a kinase or as a scaffolding protein. In summary, the model proposes that the CD4-CD8 lineage decision is determined in part by the strength of the signals provided by the TCR and co-receptors and by the relative involvement of Lck and Zap70 in the final output. SP, single positive.

This model is, however, complicated by the functional relationship between Lck and Zap70. The best known, but not the only, substrates of Lck are the cytoplasmic tyrosines of the TCR complex and Zap70. Zap70 is recruited to the phosphotyrosines of the TCR complex and is activated by Lck-mediated tyrosine phosphorylation of its kinase domain. In addition, Zap70 also has an important scaffolding role through which it recruits other effector proteins. Therefore, a model that might emerge from all of these data is one in which the activity of Lck and the transcriptionally regulated function of Zap70, either as a kinase or as a scaffolding protein, establish a balance in which the predominance of either determines differentiation toward the CD4+ or CD8+ lineages, respectively. However, the balance between the abundance or activity of the substrates of Zap70 and Lck, other than the TCR and Zap70, might also be a determinant of lineage decision. The study by Singer and colleagues (19), which indicates a dependence on IL-7R for differentiation toward the CD8+ lineage, shows that the inhibitor of IL-7R signaling, SOCS-1, has to be down-regulated, the IL-7Rα chain has to be expressed, and the chemokine receptor CCR7 has to be induced for DP thymocytes to migrate from the thymic cortex to the medulla, where IL-7 is produced. In light of the observations by Saini et al. (20), it could be that it is the TCR, through Zap70, that enables DP thymocytes to reach the IL-7–responsive state necessary for differentiation to CD8+ T cells. More than 20 years after the concept of positive selection and 30 years since the essential steps of thymic maturation were figured out, it is not clear how the TCR directs so many different steps. The study by Saini et al. is definitively a step forward in our knowledge, but certainly many questions are still pending.


This work was supported by grants SAF2006-01391 (to B.A.) and BFU2009-08009 (to H.M.vS.) from the Comisión Interministerial de Ciencia y Tecnología. The authors declare no conflict of interest.

References and Notes

View Abstract

Related Content

Navigate This Article