The TCR-mediated signals leading to IL-2 production by T cell lines and naive T cells have been well characterized; however, the biochemical signaling mechanisms leading to disparate functions such as differentiation, effector cytokine production, death, and dysfunction due to immune pathologies are not known. In this review, we highlight recent studies that have identified specific alterations in proximal and distal TCR-coupled signaling and in the configuration of the TCR-signaling complex that occur as a result of T cell differentiation, and that have been associated with T cell dysfunctions in a number of diseases. We discuss how these specific alterations directly couple to disparate functions, and how these findings suggest that primary control of T cell functional outcomes can be traced to alterations in the activation state, configuration, and functional coupling of T cell-specific signaling intermediates.
T cell activation triggered via contact of the TCR with Ag/MHC is mediated by a series of biochemical events that transduce signals from the surface TCR to the nucleus, leading to activation of gene transcription for T cell proliferation and differentiation. TCR-coupled signaling leading to IL-2 production has been extensively characterized using the human T cell leukemia line Jurkat (1), which for over 20 years has been the model system of choice for dissecting the biochemical connections comprising TCR-coupled signal transduction. However, T cells can also mediate a variety of other functions depending on their activation/differentiation state and mode of activation, and T cells contribute to a variety of immune pathologies in autoimmune diseases. What is lacking in our knowledge of TCR-coupled signal transduction is a biochemical map of the routes that TCR-coupled signaling molecules follow to disparate functional destinations such as effector cytokine production, death, survival, anergy, memory, and T cell pathology.
TCR-coupled signaling comprises a series of phosphorylation events, kinase activation, and recruitment events, leading to the formation of molecular complexes that deliver sequential and converging biochemical messages from cytoplasm to nucleus. TCR ligation by Ag/MHC results in the phosphorylation of TCR-associated CD3 subunits, such as CD3ε and CD3ζ, by members of the src family kinase family, p56lck and p59fyn. Phosphorylated CD3 molecules subsequently recruit and activate the T cell-specific ζ-associated protein-70 (ZAP-70)3 tyrosine kinase, which then phosphorylates the critical linker/adapter molecules linker for activation of T cells (LAT) and Src homology 2 domain-containing leukocyte protein-76 (SLP-76), which serve to form molecular complexes for coupling proximal phosphorylation to distal events (2). These distal events include activation and mobilization of GTP-binding proteins for cytoskeletal reorganization, activation of mitogen-activated protein kinases (MAPKs), and calcium flux, which together culminate in nuclear gene transcription (for reviews, see Refs. 3 and 4).
In this review, we discuss how specific stages in the TCR-coupled signaling cascade serve as control points for modulation of T cell function during activation and differentiation. We will present recent studies that have identified that changes in the expression, activation, and/or phosphorylation of signaling intermediates lead to rewiring of the TCR-coupled signaling pathway, with resultant functional consequences. Furthermore, we will present evidence that suggests that abnormal wiring of the TCR signaling machinery can contribute to disease pathology associated with chronic immune activation.
Control points in the proximal TCR signaling complex
Analysis of the αβ TCR structure has revealed that the signaling unit is a multimeric complex consisting of Ag/MHC binding α- and β-chains that associate with a group of invariant chains CD3γ, -δ, -ε, and -ζ (reviewed in Ref. 5). These invariant chains stabilize TCR surface expression, and also act as signal initiators following TCR ligation by Ag/MHC by their immediate phosphorylation and recruitment of signaling molecules to the immunoreceptor tyrosine-based activation motifs (ITAMs) present on their cytoplasmic tails (6). The CD3ε, CD3δ, and CD3γ subunits each bear single ITAMs, whereas CD3ζ bears three ITAMs. Two other proteins of the CD3ζ family, CD3η (derived by alternative splicing of the CD3ζ gene (7, 8) and FcRγ, originally found associated to the high affinity IgE receptor complex (9), bear two and one ITAM(s), respectively (10). The CD3ζ, CD3η, and FcRγ proteins share significant sequence homology, can stabilize surface TCR expression, and have been shown to functionally complement each other in vivo (10, 11, 12).
The αβ TCR complex largely contains CD3γε-CD3δε-CD3ζζ hetero- and homodimers (13); however, several studies have demonstrated that homologous subunits can replace the CD3ζ homodimer in the TCR complex. Although 90% of CD3ζ exists as homodimers in murine T cell clones, tumor cells, hybridomas, and normal splenic and thymic cells, CD3ζ can also form disulfide-linked heterodimers with the CD3η chain (14). Similarly, the murine CTLL cell line bears heterodimers of CD3ζ and FcRγ, and CD3η and FcRγ in the TCR complex (11). The FcRγ subunit has also been shown to be present in the TCR complex of specific nonconventional T cell types, such as human γδ T cells (15), murine CD8αα+γδ+ intraepithelial lymphocytes (16) and Ti αβ+NK1.1+ large granular lymphocytes (17).
The potential for differential phosphorylation and expression of ITAM-containing signaling subunits serves as initial control points for TCR-mediated signaling. For example, TCR stimulation induces two tyrosine-phosphorylated forms of the CD3ζ chain of 21 and 23 kDa (reviewed in Ref. 18), which trigger recruitment and activation of the Syk family kinases, ZAP-70 and Syk (19). Differential phosphorylation of CD3ζ leads to disparate functional outcomes, with a higher level of the 23-kDa form favoring ZAP-70 recruitment and phosphorylation and resultant T cell activation, whereas a preponderance of the 21-kDa form results in T cell anergy and lack of ZAP-70 phosphorylation (20, 21). In addition, the presence of CD3ζ or FcRγ on T cells can likewise affect Syk family kinase recruitment. Although CD3ζ recruits ZAP-70, FcRγ preferentially associates with Syk (22, 23). Although the precise functions of CD3ζ and FcRγ remain unclear, the differential binding affinities of tyrosine kinases to the three ITAMs of CD3ζ (24) suggest that lack of the first two ITAMs might serve as a mechanism for divergence of signaling for FcRγ. Consistent with this idea is the observation that FcRγ+ T cell hybridomas produce different levels of IL-2 than CD3ζ+ counterparts (10).
Control points in distal TCR signaling
The activation of Syk family kinases triggers two related signaling pathways that ultimately converge in transcription activation in the nucleus. One pathway involves phosphorylation-induced activation of phospholipase C-γ1 by Lck, ZAP-70, and Tec kinases in a process that is aided by the adapter proteins LAT and SLP-76 (reviewed in Ref. 25). Activated phospholipase C-γ1 splits membrane phosphatidylinositol 4,5-bisphosphate into inositol 1,4,5-triphosphate and diacylglycerol (26), leading to calcium flux, calmodulin/calcineurin-dependent activation (27), and translocation of NFAT for IL-2 production (28). The second pathway also involves the adapter proteins LAT and SLP-76, which interface with the linker/adapter Grb2-related adaptor downstream of Shc to couple ras/raf activation to MAPKs for IL-2 production (reviewed in Refs. 25 and 29).
Alterations in these pathways have been observed in response to TCR ligation leading to signaling divergence. On the linker/adapter level, diminished LAT phosphorylation was found to be triggered by antagonist peptide ligands, suggesting that alterations in the linker/adapter signaling may contribute to anergy induction (30, 31). Additional linker/adapter molecules have also been shown to be involved in regulation of signaling following initial activation. The adapter protein Nck was recently shown to be necessary for optimal T cell activation, and acts by binding to CD3ε that is conformationally altered due tο TCR-ligand binding (32). Another adapter protein, Clnk, gets up-regulated following treatment of cells with IL-2 and can induce the transcriptional activity of NFAT and AP-1 (33). Adapter molecules of the Cbl family, cCbl and Cbl-b, negatively regulate T cell activation by regulating the dynamics of TCR down-modulation following engagement with the Ag (34). These studies suggest that differential expression of adapter proteins may be a mechanism to achieve different functional outcomes.
Differential calcium responses and MAPK activation also determine whether a T cell undergoes activation or anergy following TCR ligation. Thus, anergy induced by antagonist peptide ligands was accompanied by sustained duration of induced calcium influxes (35, 36, 37). Similarly, anergy can be induced by maintaining sustained high levels of intracellular calcium, suggesting that altering calcium fluxes could be a mechanism for signaling control of T cell activation (38). Disparate patterns of MAPK phosphorylation are also observed in activated vs anergic T cells, with anergic cells exhibiting impairments in Ras activation and activation of the MAPKs, Erk1/2 and c-Jun N-terminal kinase (JNK) (39, 40).
Additional control of TCR-mediated signaling is triggered via costimulatory molecules that act to negatively regulate T cell activation, such as CTLA-4 and programmed death-1 gene. Although the signaling pathways coupled to these molecules are not fully elucidated, both have been shown to recruit phosphatases, such as protein phosphatase 2A (41) and Src homology 2-containing protein tyrosine phosphatase 2 (42), that presumably act to dephosphorylate key signaling intermediates, attenuating the cascade of activating signals from the TCR to the nucleus (43). A deficit in either of these molecules results in unchecked T cell proliferation and activation, or autoimmunity (44, 45, 46, 47), indicating that the integrity of pathways mediated by these negative regulators is critical for controlling T cell tolerance and preventing autoimmunity.
TCR signaling changes during effector and memory T cell differentiation
The signaling control points described above have been identified during the initial activation of resting T cells leading to IL-2 production. The resultant activated cells then proliferate and differentiate to effector cells that produce effector cytokines such as IFN-γ, TNF-α, IL-4, and or IL-10 for the coordination and regulation of Ag clearance. Although most of these effector cells die after Ag is cleared, a proportion of activated cells persist in a quiescent state as long-lived memory T cells, which mediate an enhanced and highly effective recall immune response upon Ag re-encounter. The mechanisms and TCR-coupled signaling changes underlying the distinct functions, life spans, and generation of effector and memory T cells are not known. Furthermore, it has been difficult to establish roles for specific signaling intermediates in effector and memory T cell differentiation, because mice genetically manipulated to lack key signaling intermediates, such as Lck (48), ZAP-70 (49), SLP-76 (50), LAT (51), and Erk (52), all exhibit profound defects in T cell development and lack peripheral T cells.
Our laboratories have closely examined TCR-coupled signaling pathways in naive, effector, and memory CD4 T cells from human and mice and have uncovered a number of striking differences in the biochemical pathways coupled to the TCR in these subsets. In general, TCR-mediated signaling in naive T cells, which primarily produce IL-2 when activated, resembles the general TCR signaling identified in T cell lines and clones (53, 54). By contrast, effector T cells are characterized by a profound amplification of signaling, manifested by an increase in tyrosine phosphorylation, which is further increased following TCR cross-linking (53, 54, 55). Memory T cells from humans and mice, in contrast to both naive and effector cells, are characterized by a dampening of signals through the TCR/CD3 complex, as manifested by a decrease in total tyrosine phosphorylation in the resting state and an absence of specific tyrosine-phosphorylated species following TCR/CD3 cross-linking (53, 54, 55, 56). These findings by our laboratories and others indicate that the distinct functions of naive, effector, and memory T cells are controlled, in part, by the specialized TCR-coupled signaling pathways present in these subsets.
We have further dissected the specific signaling alterations underlying the unique biochemical signatures of effector and memory CD4 T cells on proximal, linker/adapter, and distal levels (Table I). In mouse effector cells, earlier studies by Bachmann et al. (57) demonstrated increased Lck kinase activity in mouse effector CD8 T cells, consistent with our findings of enhanced tyrosine phosphorylation in effector T cells. However, other proximal events such as CD3ζ phosphorylation and ZAP-70 activation are not amplified in effector cells, and do not participate in effector cell-specific TCR signaling. For example, human effector T cells generated by anti-CD3 or antigenic stimuli, exhibit a dramatic decrease in expression of the CD3ζ protein, concomitant with a reduction in CD3ε and TCR surface expression, and up-regulation of the FcRγ signaling subunit (55, 58). FcRγ in effector cells associates strongly with the TCR/CD3ε complex after anti-CD3 stimulation, which excludes CD3ζ and ZAP-70, and contains FcRγ and the Syk tyrosine kinase, whose expression is likewise up-regulated in effector T cells (55, 58) (Fig. 1). On the linker/adapter level, we found striking increases in the phosphorylation of the linker/adapter molecules LAT and SLP-76, and increased association of phosphorylated proteins to the ubiquitous linker Grb2, which couples proximal events to ras activation, calcium flux, and MAPK activation (54). Distally, effector T cell signaling is also amplified with hyperphosphorylated MAPK activity and increased calcium flux (Ref. 54 and our unpublished data). Thus, TCR signaling in effector cells is initiated by an altered proximal TCR configuration concomitant with amplification of phosphorylation and signaling at the linker/adapter and distal levels (see Fig. 1). These downstream signaling amplifications are also found in effector CD8 T cells that phosphorylate LAT and MAPKs ERK, JNK, and p38 more efficiently than naive cells (59), and correlate to enhanced functional responses of effector cells including increased cytokine production (53, 55, 60) and response kinetics (61).
|.||Signaling Intermediateb .||Naive .||Effector .||Memory .||SLE .|
|Distal intermediates||Calcium flux||+||++||+/−||++|
|.||Signaling Intermediateb .||Naive .||Effector .||Memory .||SLE .|
|Distal intermediates||Calcium flux||+||++||+/−||++|
See text for references.
Exp, Protein expression; Phos, tyrosine phosphorylated.
Indicates very low level of expression or phosphorylation.
Does the change in the proximal TCR configuration from the canonical TCR/CD3ε/CD3ζ/ZAP-70 to a new TCR/CD3ε/FcRγ/Syk complex direct increased downstream signaling in effector cells? To address this question, we transfected cDNA encoding FcRγ into resting T cells isolated from healthy individuals. Indeed, transfection of FcRγ into normal T cells results in recruitment of the Syk kinase, increased intracellular tyrosine phosphorylation, and increased calcium flux (62), strongly implicating FcRγ and/or Syk as driving the enhanced signaling observed in effector cells. The fact that the Syk kinase is 100-fold more potent than ZAP-70 on a molar level (63) suggests that recruitment of Syk enables increased signals to be propagated with reduced TCR ligation, and may be a mechanism accounting for the lower activation threshold of effector vs naive T cells (61, 64). We likewise observed increased IL-2 production in FcRγ+ vs FcRγ− T cells (62), establishing a functional consequence of a change in proximal signaling. Further studies are underway to determine the contribution of FcRγ/Syk in effector functions such as effector cytokine production, activation-induced cell death, and rapid response kinetics. Downstream MAPK activation has also been linked to effector cell function. For example, Erk1/2 and JNK1 have been demonstrated to be necessary for Th2 cell generation (65, 66, 67), whereas JNK2 and p38 MAPK are necessary for Th1 generation (67, 68). It is not known what proximal signaling changes and/or activation differences govern the type of MAPK activated during T cell activation and differentiation.
Memory CD4 T cells exhibit a unique TCR-coupled signaling pattern, distinct from naive and effector cells. Whereas CD3ζ expression and phosphorylation are decreased in effector T cells, CD3ζ is highly expressed in memory T cells and is phosphorylated following TCR triggering of memory T cells (69, 70), although ZAP-70 is phosphorylated in mouse and human memory T cells to a lesser extent than in naive T cells (56, 70). Interestingly, the TCR signaling intermediates that are specifically amplified and hyperphosphorylated in effector cells are down-regulated and/or hypophosphorylated in memory T cells. For example, we found that mouse memory CD4 T cells are remarkably deficient in SLP-76 expression, although the related linker/adapter LAT is expressed in comparable levels in naive and memory T cells (54). The memory T cell-specific decrease in SLP-76 is consistent with the decreased calcium responses and IL-2 production previously identified in memory-phenotype CD4 T cells (71, 72, 73, 74), and also found in SLP-76-deficient Jurkat T cells (75). Whether a dampening of signals controlled by decreasing SLP-76 expression facilitates memory generation from effector cells, or whether memory cell signaling dampening results from their long-term maintenance and/or homeostasis remains to be established.
The role of the negative signaling regulators in different T cell stages is currently unclear. There is evidence that memory cells from both mice and humans contain small intracellular pools of CTLA-4 expression that may resist memory T cell activation in response to suboptimal stimulation (Refs.76 and 77 ; and V. G. Warke, D. L. Farber, G. C. Tsokos, unpublished observation). Programmed death-1 expression occurs in activated and not resting T cells (78) and may therefore play a role in controlling TCR signaling in effector and/or memory T cells.
Our findings of differential signaling in effector and memory T cells at the proximal and linker/adapter levels suggests that these signaling junctures serve as control points for regulation of T cell differentiation. We favor a model whereby alterations in the expression and/or activation and phosphorylation state of specific signaling intermediates during effector and/or memory generation, results in a rewiring of the αβTCR from the cytoplasm to the nucleus, through the establishment of distinct molecular complexes. The specific connections that result in turn, govern the unique functions and properties of these subsets. Further dissection of these novel pathways in effector and memory T cells will enable the specific targeting of these pathways in therapies to modulate effector and memory cell function in vivo in autoimmune, malignant, and infectious disease and in vaccine design.
TCR rewiring in disease
Alterations in TCR-mediated signaling have been found in T cells isolated from individuals with a number of chronic diseases such as autoimmune diseases, chronic viral infections, and cancer. A reduction in the expression of CD3ζ chain has been shown to occur in T cells associated with viral infections such as HIV (79), EBV- and CMV-mediated infectious mononucleosis (80), cancer (81, 82, 83), and autoimmune diseases including rheumatoid arthritis (RA) (84) and systemic lupus erythematosus (SLE) (85). Although T cells in RA demonstrate CD3ε loss as well (86), tumor-infiltrating lymphocytes (TILs) demonstrate an additional loss of CD3γ (83). The involvement of chronic immune activation in the mediation of some of these effects is suggested by the effector phenotype of T cells in autoimmune diseases such as RA and SLE (87, 88).
The correlation of CD3ζ down-regulation with effector T cells generated from healthy individuals (55) raises the question of whether these signaling abnormalities simply indicate the presence of effector cells in these diseases marked by chronic immune activation. Our studies on peripheral T cells from patients with SLE indicate that these changes may, in part, derive from the presence of effector cells. Like effector cells, T cells from SLE patients exhibit decreases in CD3ζ expression (85, 89), FcRγ up-regulation and Syk recruitment to the TCR complex (90), increased and sustained distal signaling as measured by augmented calcium responses, and increased intracellular phosphorylation (85, 90). However, unlike healthy effector T cells, T cells from SLE patients display defects in the transcription of CD3ζ and the expression of activated Elf-1, which promotes CD3ζ transcription (85, 91) and reductions in IL-2 production (92, 93). However, effector cells have been shown to exhibit decreased IL-2 production at later times after activation (94), and activation-induced alterations in CD3ζ transcription have been demonstrated to occur in T cell lines (95), suggesting that SLE-specific T cell abnormalities may also be activation driven. The fact that these effector-like signaling alterations are found in freshly isolated peripheral blood T cells from SLE patients independent of the disease activity (85) suggests that there may be abnormal persistence of long-lived effector-like T cells in SLE. The presence of chronically activated effector cells in autoimmune diseases, in general, could also be due to impairments in negative regulation as suggested by findings that CTLA-4 polymorphisms are associated with a number of autoimmune diseases (96). Analysis of the biochemical mediators of positive and negative signaling in effector cells and SLE T cells will help resolve this issue.
TILs from a variety of sources also exhibit certain features of effector cells, yet similar to SLE T cells, show evidence of T cell impairments. Like effector cells, TILs from tumor-bearing mice exhibit replacement of CD3ζ by FcRγ in the TCR complex (83). Moreover, human TILs from multiple cancer cell types including renal cell carcinoma (81), melanoma (97), oral carcinoma (98), colorectal carcinoma (99), and ovarian carcinoma (100) exhibit strikingly reduced expression of CD3ζ. However, mouse TILs also exhibit decreased calcium responses (83), suggesting functional impairments. Furthermore, the presence of T lymphocytes with low CD3ζ expression was found to correlate to low survival outcomes in patients with oral carcinoma (101, 102) and gastric carcinoma (103), and correlated with increased disease severity in colorectal carcinoma (104) and Hodgkin’s lymphoma (105), suggesting impaired immune effector function.
When taken together, T cells from different disease states exhibit similar proximal signaling alterations, yet altered function. These findings suggest two nonexclusive explanations for the origin of these signaling changes. First, distal signaling pathways and/or differential transcriptional activation may be found in disparate diseases. SLE T cells, for example, have been shown to up-regulate an IL-2 transcriptional repressor, cyclic adenosine 5′-monophosphate response element modulator, when compared with healthy T cells (106). Second, the observed signaling alterations may not necessarily govern T cell dysfunctions, but rather reflect the differentiation state on which the pathological influences within the disease environment, such as cytokines and an altered cellular composition, are exerted. Further dissections of signaling pathways used in T cells from disparate diseases will identify disease-specific signaling abnormalities and how they may arise.
It appears that T cell differentiation carries a distinct biochemical signature that apparently precedes the functional phenotype. The work discussed herein clearly indicates T cell rewiring both at the proximal TCR receptor level and subsequent levels that determine the fate of the T cell. A theme that has arisen is that the T cells from SLE share many features with normal effector T cells including the down-regulation of the CD3ζ chain, the up-regulation of the FcRγ chain, the increased [Ca2+]i, tyrosine phosphorylation response, and MAPK activation, yet exhibit alterations in the transcriptional machinery controlling cytokine production. The fact that certain biochemical steps characterize the behavior of the T cell raises enormous possibilities first for interventions to modify the rate of appearance of memory and effector T cells following exposure to Ag and for modulation of the T cells to limit disease pathology.
This work was supported by National Institutes of Health RO1 AI42092 (to D.L.F.), RO1 AI 42269, and RO1 AI49954 (to G.C.T.).
Abbreviations used in this paper: ZAP-70, ζ-associated protein-70; LAT, linker for activation of T cells; SLP-76, Src homology 2 domain-containing leukocyte protein-76; MAPK, mitogen-activated protein kinase; ITAM, immunoreceptor tyrosine-based activation motif; JNK, c-Jun N-terminal kinase; RA, rheumatoid arthritis; SLE, systemic lupus erythematosus; TIL, tumor-infiltrating lymphocyte.