Reversal of the TCR Stop Signal by CTLA-4

The T cell coreceptor CTLA-4 has a powerful modulatory effect on the signals that induce T cell cytokine production and proliferation (1, 2). Consequently, CTLA-4–deficient (CTLA-4^–/–) mice show a profound postthymic autoimmune phenotype marked by massive tissue infiltration and organ destruction (3, 4). Proposed mechanisms for the negative influence of CTLA-4 include ectodomain competition for binding of the related costimulatory molecule CD28 to CD80/86 (5), disruption of CD28 localization at the immunological synapse (IS) (6), modulation of TCR signaling by protein phosphatase 2A and the tyrosine phosphatase SHP-2 (7–9), and interference with the expression or composition of lipid rafts on the surface of T cells (10–12). In addition, CTLA-4 engagement of CD80/86 on dendritic cells (DCs) can induce the release of indoleamine 2,3-dioxygenase (13), whereas CD4⁺CD25⁺ regulatory T cells can modulate disease in the CTLA-4^–/– mouse (14–16). Antibodies to CTLA-4 (anti–CTLA-4) and CTLA-4–immunoglobulin have been used as therapeutics in the modulation of autoimmunity, transplantation, and tumor immunotherapy (17, 18).

CTLA-4 is a potent direct activator and a TCR coactivator of LFA-1 integrin clustering and adhesion (19). To assess whether CTLA-4 could also influence integrin-dependent motility, we tracked preactivated CTLA-4^+/+ and CTLA-4^–/– T cells for movement on plates coated with the LFA-1 ligand intercellular adhesion molecule–1 (ICAM-1) (Fig. 1). In contrast to untreated CTLA-4^+/+ and CTLA-4^–/– cells, which migrated at similar speeds (Fig. 1, A, C, and E), CTLA-4^+/+ cells treated with anti–CTLA-4 increased motility (Fig. 1, B versus A). As anticipated, the effects of anti–CTLA-4 were not observed with CTLA-4^–/– T cells (Fig. 1, D versus C; E). Thus, in addition to increasing LFA-1 adhesion (19), CTLA-4 readily increased the movement of T cells.

• Download high-res image
• Open in new tab
• Download Powerpoint

TCR ligation reduces or arrests T cell motility (i.e., the stop signal), an event that is required for stable T cell conjugate formation with antigen-presenting cells (APCs) and efficient activation (20–22). Given that CTLA-4 increased T cell motility, we assessed whether the coreceptor could override this stop signal. Preactivated murine and human T cells were monitored for 20 min after exposure to anti-CD3 in the presence or absence of anti–CTLA-4. In contrast to the ligation of anti-CD3, which reduced the movement of mouse and human primary T cells (Fig. 2, B versus A and F versus E), the coligation of CTLA-4 and anti-CD3 reversed the arrest such that treated and untreated cells moved at similar speeds (Fig. 2, C versus A and G versus E; D and H). Similar effects were seen with the CTLA-4–transfected T cell hybridoma DC27.10 (fig. S1). These findings showed that CTLA-4 can override the stop signal induced by the antigen receptor complex.

• Download high-res image
• Open in new tab
• Download Powerpoint

To extend our findings to an in vivo setting, we separated preactivated CTLA-4 positive and negative CD4⁺ T cells transgenic for the DO11.10 TCR using beads coated with anti–CTLA-4 (23). These cells were then labeled with carboxyfluorescein diacetate succinimidyl ester (CFSE) and tracked in cervical lymph nodes by two-photon laser scanning microscopy (TPLSM) (Fig. 3). T cells in mice injected with antigen have been shown to undergo changes in motility involving initial transient serial encounters, followed by a phase of slowing and stable contacts and then by a return to high motility (22, 24–27). In our study, CTLA-4 positive and negative T cells were monitored at 15, 20, and 24 hours after injection with ovalbumin (OVA) peptide. The first noticeable difference between cells was a more extended, polarized appearance of CTLA-4 positive cells (fig. S2 and movies S1 to S4). In the absence of peptide, CTLA-4–expressing and –nonexpressing T cells showed a random walk pattern with similar motilities (Fig. 3, A and E versus G). After the OVA peptide injection, CTLA-4 negative cells showed the anticipated decrease in mean motility (Fig. 3, A and F versus E). By contrast, the motility of CTLA-4 positive cells remained unchanged and occasionally increased in the presence of OVA peptide (Fig. 3, A and H versus F; fig. S2, movies S1 to S4). The three-dimensional motility coefficient showed a reduction from 140 to 29.7 μm³ min^–1 for CTLA-4 negative cells and an increase from 114 to 210 μm³ min^–1 for CTLA-4 positive cells (Fig. 3B). Dot plot analysis showed some heterogeneity in the motility of individual cells, although the mean motility of CTLA-4 positive cells remained unchanged in the presence of peptide (Fig. 3, I and J). Despite this result, the two subsets had similar volumes (Fig. 3C) and meandering indices (Fig. 3D). This resistance of CTLA-4 positive cells to antigen-induced slowing was observed at all times examined. These experiments revealed that CTLA-4 positive cells fail to stop or slow down in response to the in vivo peptide challenge.

• Download high-res image
• Open in new tab
• Download Powerpoint

Because the slowing of motility is required for stable T cell–APC conjugate formation (20–22, 26, 27), the reversal of the stop signal by CTLA-4 suggested that the coreceptor might reduce the period of interaction. To test this conjecture, we incubated CTLA-4 positive and negative DO11.10 T cells with a CFSE-labeled B cell line, A20 (Fig. 4A), or labeled mature DCs (Fig. 4B). APCs were left untreated or were preincubated with OVA peptide. Individual T cells were then monitored for the duration of APC binding over a 1200-s period. Whereas CTLA-4 negative T cells incubated with APCs without peptide showed short-term interactions (75% of which were less than 400 s), the addition of OVA peptide induced longer term interactions (60 to 70% were more than 600 s). In contrast, CTLA-4 positive T cells failed to form long-term interactions even in the presence of antigen. Instead, like those cells unexposed to peptide, CTLA-4 positive cells under-took short-term interactions. We observed this effect when A20 B cells or DCs were used as APCs (Fig. 4, A and B). In turn, these shorter dwell periods (Fig. 4C) resulted in lower interleukin-2 (IL-2) production (Fig. 4D) and proliferation (Fig. 4E). CTLA-4 expression therefore led to more transient interactions with APCs and a reduction in T cell activation.

• Download high-res image
• Open in new tab
• Download Powerpoint

Overall, our observations support a reverse stop-signaling mechanism for modulating the threshold of T cell activation by CTLA-4, which is distinct from previous models of CTLA-4 function. By limiting T cell–APC contact times, CTLA-4 would reduce the efficiency of the major histocompatibility complex (MHC)–peptide presentation and the number of TCR ligation events resulting in reduced signaling and activation. Shorter dwell times could also alter receptor rearrangements at the IS and account for the reduced localization of CD28 (6). In this way, CTLA-4 may be tailored to regulate the secondary responses of effectors such as cytotoxic T lymphocytes, whose shorter dwell times are sufficient to elicit cytotoxicity. More transient interactions may in turn increase the frequency of contacts between effector and target cells. CTLA-4 positive cells may also compete with CTLA-4 negative cells to inhibit longer term conjugation with APCs and facilitate a rapid exit from lymph nodes and reentry to the circulation for migration to sites of inflammation.

The role of CTLA-4 as a gatekeeper of conjugation may also account for the connection with autoimmunity (3, 4). Reduced conjugation might protect against prolonged contact periods that allow for responses to lower-affinity autoantigen. Increased LFA-1–mediated adhesion in the absence of increased motility allows for responses to low-affinity subthreshold agonist including self-MHC molecules (28–30). This may also explain the reported requirement for anergy induction of the T cell (31, 32). In certain settings, by limiting dwell times, suboptimal or altered signaling may result in nonresponsiveness. Indeed, tolerance induction is accompanied by less stable DC–T cell interactions (26, 33), smaller clusters (27), and a rapid restoration of motility (33). Regulatory T cells can also limit T cell–APC contact times (34). Further studies will more precisely elucidate the role of altered motility and conjugation in the full range of functions regulated by the coreceptor.