Introduction
TCRs responsible for antigen recognition by T cells are noncovalently associated with multiple CD3 signaling subunits that contain one or several ITAMs. As a result, each antigen-binding 
TCR dimer is functionally coupled to a constellation of ten ITAMs (Fig. 1). The evolutionary relationship between all CD3 ITAMs is evidenced by their shared capacity to recruit the ZAP-70 cytosolic protein tyrosine kinase (PTK) after phosphorylation by the Lck and Fyn PTKs. The large number of ITAMs associated with a single TCR is unique among immunoreceptors and is thought to assist T cells in their formidable task of detecting minute amounts of foreign peptide-MHC (pMHC) complexes on the surface of antigen-presenting cells (APCs). In this issue of Nature Immunology, Holst et al. used a powerful transgenic mouse technique and showed that the large numbers of CD3 ITAMs associated with the TCR are required for the establishment and maintenance of self-tolerance1.
Figure 1: TCR-CD3 complexes on 
T cells contain a single antigen-binding TCR heterodimer that assembles with single CD3
, CD3
and CD3
pairs.
![Figure 1 : TCR-CD3 complexes on |[alpha]||[beta]| T cells contain a single antigen-binding TCR heterodimer that assembles with single CD3|[gamma]||[epsi]|, CD3|[delta]||[epsi]| and CD3|[zeta]||[zeta]| pairs.](/ni/journal/v9/n6/thumbs/ni0608-583-F1.jpg)
The cytoplasmic segment of the CD3
, CD3
and CD3
subunits contain a single ITAM, whereas that of the CD3
subunit contains three ITAM copies. According to that model, each TCR heterodimer is functionally coupled to ten ITAMs. Once the two tyrosine residues (YY) found in a given CD3 ITAM are phosphorylated, they recruit the cytosolic protein tyrosine kinase ZAP-70. As a result, ZAP-70 becomes activated and phosphorylates a number of downstream molecules, among which the adaptor LAT plays a cardinal role in that it links the TCR to a wealth of signaling pathways. SH2, Src homology 2 domain.
Immunodeficient mice were reconstituted with hematopoietic stem cells that were isolated from mice deprived of all CD3 subunits and that were infected with multicistronic retroviral vectors encoding combinations of CD3 subunits containing functional or inactivated ITAMs. Analysis of 25 groups of these 'retrogenic' mice expressing graded numbers of functional CD3 ITAMs showed that almost all mice with fewer than seven functional ITAMs developed fulminant multiorgan inflammatory diseases triggered by CD4+ T cells. The sporadic presence of autoantibodies 6 weeks after bone marrow transfer, the mild in vitro CD4+ T cell proliferation noted in response to autologous APCs and the presence of functionally competent Foxp3+ regulatory T (Treg) cells led Holst et al.1 to conclude that in mice with fewer than seven functional ITAMS, a breakdown in central rather than peripheral tolerance triggered the generation of interferon--producing autoreactive CD4+ T cells. These findings are surprising because in previous studies based on conventional transgenesis, CD3-deficient mice complemented with ITAM-less CD3 subunits did not develop any signs of inflammation despite having six or fewer functional ITAMs in their TCRs2, 3. Whether the expression of lower amounts of TCR on the surface of T cells from retrogenic mice accounts for this discrepancy remains to be determined. However, it is important to emphasize that thymi and spleens of retrogenic mice reconstituted with wild-type CD3 chains contained only one-tenth as many T cells as corresponding organs from unmanipulated mice, suggesting that lymphopenia might contribute in part to the phenotype observed in some retrogenic mice.
Autoimmunity can result from defects in the mechanisms that keep TCR-driven T cell activation in check. For instance, the roquin (Rc3h1) gene is involved in follicular T-helper cell differentiation, and roquin-defective follicular T cells become uncontrolled and trigger the production of large amounts of autoantibodies4. As illustrated by the results of Holst et al.1 and by previous studies of mutations in several molecules involved in TCR signaling (such as ZAP-70, the adaptor LAT and the protein tyrosine phosphatase PTPN22), TCR hyporesponsiveness can result in autoimmunity. How can mutations that reduce TCR signaling output paradoxically lead to autoimmunity? Mice with a point mutation in ZAP-70 (SKG mice) develop a T cell–mediated autoimmune disease resembling human rheumatoid arthritis5. The arthritogenic CD4+ T cells found in SKG mice are hyporesponsive to TCR cross-linking but show vigorous in vitro proliferative responses to MHC molecules expressed on autologous APCs. Moreover, the SKG pathology can be prevented by neonatal infusion of wild-type Treg cells. Therefore, it is likely that in SKG mice the ZAP-70 mutation compromises the development of Treg cells to a greater degree than that of effector CD4+ T cells. In contrast, imbalanced development of regulatory and effector CD4+ cell subsets does not seem to account for the autoimmune pathology that develops in mice with point mutations in the ZAP-70 catalytic domain (Zap70mrd/mrt mice; ref. 6) or in the LAT adaptor (LatY136F mice; refs. 7,8). It has been hypothesized that the LatY136F mutation produces a failure of central tolerance but also impedes TCR-triggered inhibitory feedback pathways to a greater extent than activating pathways in peripheral CD4+ T cells. This putative intracellular signaling imbalance should result in hyperactive self-reactive T helper cells responsible for the presence of severe inflammation and autoantibodies7. Consistent with the view that the multiple manifestations of T cell activation (such as proliferation, induction of apoptosis and cytokine production) have distinct activation thresholds, Holst et al.1 showed that although there was a linear relationship between the number of ITAMs and T cell proliferative capacity, T cells possessing only a single functional ITAM were nevertheless capable of producing cytokines1.
The extent to which some autoimmune diseases are due to a failure in the establishment and maintenance of self-tolerance, as compared to abnormal antigen-independent signaling events that result in uncontrolled polyclonal activation of effector lymphocytes, remains the subject of many debates9. In the case of T cells, reduced TCR signaling during thymic selection should cause overtly self-reactive thymocytes to be positively selected instead of negatively selected and leads to the generation of peripheral T cells expressing TCR with higher affinity for self-pMHC ligands. Intuitively, however, one might expect that such gain in TCR affinity for self-pMHC ligands is cancelled out by a commensurate increase in the threshold required for antigen-driven TCR activation in the periphery, with the whole system simply readapting to a new 'calibration set-point' dictated by the nature of the mutation under study10. According to that model, the CD4+ T cells emerging in the periphery of Zap70mrd/mrt or LatY136F mice should not show any overt autoreactivity. A revisiting of the basis of the CD4+ T cell hyperactivity in LatY136F mice supported that view and suggested that LatY136F CD4+ T cell proliferation and helper function occurred in the absence of TCR-pMHC interactions11. Moreover, the autoantibody production triggered by LatY136F CD4+ T cells reflected a capacity to help B cells in a TCR-independent 'quasi-mitogenic' mode that induced a massive polyclonal B cell activation accompanied by the production of autoantibodies among other antibodies. The LatY136F mutation might thus relieve the tight control the TCR normally exerts on co-stimulators or cytokine receptors, leaving to these last, uncontrolled pathways the responsibility for the abnormal behavior manifested by the LatY136F T cells. In a way, the etiology postulated for the LatY136F pathology is reminiscent of the conditions known as autoinflammatory diseases or sterile inflammation, in which recurrent episodes of systematic inflammatory attacks unfold in the absence of exogenous infectious triggers and result from intrinsic signaling dysfunction of innate immune cells12. It will thus be interesting to analyze the relative contributions of self-tolerance defects, as opposed to imbalances in TCR-independent activating and inhibitory signals in peripheral T cells, to the etiology of pathological conditions—including the one described by Holst et al.1—that have been defined as 'autoimmune' primarily due to the presence of autoantibodies.
In summary, the technical tour de force of Holst et al.1 reinforces the view that mutations in the genes encoding molecules involved in the proximal signaling core of the TCR contribute to the development of autoimmune and inflammatory diseases by affecting thymic T cell selection, by relieving some signaling pathways from the control normally exerted by the TCR, or both. Finally, it is important to stress the reduced ability to mount immune responses and the lower lymphocyte counts are generally associated with hypomorphic mutations in the TCR signaling cassette might also result in opportunistic infections and in lymphopenia-associated abnormalities that should be taken in considerated when analyzing the trigger of the resulting disease.
