Report

A Receptor/Cytoskeletal Movement Triggered by Costimulation During T Cell Activation

See allHide authors and affiliations

Science  18 Dec 1998:
Vol. 282, Issue 5397, pp. 2266-2269
DOI: 10.1126/science.282.5397.2266

Abstract

During T cell activation, the engagement of costimulatory molecules is often crucial to the development of an effective immune response, but the mechanism by which this is achieved is not known. Here, it is shown that beads attached to the surface of a T cell translocate toward the interface shortly after the start of T cell activation. This movement appears to depend on myosin motor proteins and requires the engagement of the major costimulatory receptor pairs, B7-CD28 and ICAM-1–LFA-1. This suggests that the engagement of costimulatory receptors triggers an active accumulation of molecules at the interface of the T cell and the antigen-presenting cell, which then increases the overall amplitude and duration of T cell signaling.

The central event in T cell activation is the interaction of the T cell receptor (TCR) with the antigenic peptide presented by the major histocompatibility complex (MHC) of the antigen-presenting cell (APC). However, because the number of agonist peptide–MHC complexes can be very low, in the range of 10 to 100 per APC (1), and because the TCR is continuously modulated from the T cell surface (2), sustained T cell activation is likely to require signal amplification (3,4). An important component of this amplification is thought to be provided by costimulatory molecules on the T cell, although the mechanism by which they accomplish it is unclear (5). The most important of the costimulatory receptors on T cells and their ligand on APCs are CD28-B7 (6) and LFA-1–ICAM-1 (7). Many different receptor couples, including TCR–peptide-MHC and LFA-1–ICAM-1, accumulate at the T cell–APC interface (8, 9). This accumulation has been assumed to be a passive, diffusion-limited cocapping mechanism (10). Here, we describe an active, cytoskeletal mechanism that appears to drive receptor accumulation at the T cell–APC interface. This mechanism requires the APC to express B7 and ICAM-1 and is independent of TCR signaling. We suggest that this mechanism is a central part of costimulation, as it would effectively amplify any TCR-mediated signals.

To study whether receptor accumulation at the T cell–APC interface could be actively driven by the T cell cytoskeleton, we monitored the general movement of the cortical actin cytoskeleton and linked receptors (11) using the classical technique of attaching large beads to the surfaces of antigen-specific T cells. We coated 4.5-μm beads with an antibody to the T cell surface antigen ICAM-1 (12). Cross-linking ICAM-1 by beads in this way is not expected to influence T cell function (13). As has been observed with fibroblasts (11, 14), we find that in migrating 5C.C7 transgenic T cells (9, 15) the beads translocate from the anterior to the posterior end of the cell (16). We then mixed bead-loaded 5C.C7 T cells with B cell lymphoma cells that express the appropriate MHC molecule (I-Ek) and have been pulsed with the moth cytochrome c peptide 88-103. After contact with the APC, the T cells rapidly become activated (9, 17) and the beads move from the posterior end of the T cell to the newly formed interface with the B cells (Fig. 1, movie 1) beginning 4 ± 1 min after the first rise in intracellular calcium (n = 13). This suggests that the T cell cortical actin cytoskeleton reorients toward the T cell–APC interface soon after the start of T cell activation. The ensuing cytoskeletal flow would allow receptors that are linked to the actin cytoskeleton to be transported to the newly formed T cell–APC interface.

Figure 1

Bead movement toward the interface. Single frames of a video microscopy experiment (36) of the interaction of 5C.C7 T cells, loaded with 4.5-μm anti–ICAM-1 (YN1) beads, with peptide-loaded CH27 B cell lymphoma cells (35) are shown. The CH27 cell is substantially larger than the T cells. The T cell intracellular calcium concentration is overlaid in a false color scale from blue (low concentration) to red (high concentration) to mark the onset of T cell activation, set to time 0:00 min. Although the beads, one of which is marked with an arrow, are bound to the posterior end of the central T cell before and at the time of its activation, they can be seen to move toward the T cell–B cell interface in subsequent frames. The still frames have been excerpted from movie 1 (37), which can be viewed at Science Online (www.sciencemag.org).

To rule out ICAM-1–specific effects, we also attached beads to T cells in two other ways. First, we surface-biotinylated the 5C.C7 T cells with an amine-reactive form of biotin and used 2.8-μm streptavidin beads (18). Second, we exchanged a lipid that is biotinylated at its head into the cell surface of the T cells from liposomes (19) and used the same 2.8-μm streptavidin beads (20). In both cases, we found a behavior identical to that of the antibody-coated beads (Table 1). This confirms that the beads monitor the movement of the cortical actin cytoskeleton and not that of a specific receptor. We have also obtained identical results using monocytes and dendritic cells prepared directly ex vivo as APCs (16). The time to travel half a cell circumference is 6 ± 2 min (n = 14). Given that the diameter of our activated T cells is ∼10 μm, this speed of ∼3 μm/min is similar to that reported in comparable studies on migrating fibroblasts (1 to 5 μm/min) (21). This behavior is independent of the peptide concentration on the APC as long as enough peptide is presented to activate the T cell (16). Beads bound to the B cell do not move, consistent with the lack of involvement of the B cell actin cytoskeleton in the redistribution of specific receptors (9, 22).

Table 1

Occurrence of different types of T cell bead movement (39) during the interaction of a 5C.C7 T cell with APCs loaded with an activating peptide. Unless mentioned, the APCs were CH27 B cell lymphomas. Bead movement was analyzed using three different types of beads (12, 18, 20); only two of these bead types (12, 18) were used for the rest of the experiments. Blocking antibodies and pharmacological agents were used under standard conditions. (40). The number of cells analyzed during wortmannin treatment was small because the majority of the T cells did not establish a polarized phenotype (41). When transfected CHO cells were used as APCs, we found that GFP fusion did not inhibit the function of its fusion partner (9, 22). In the last row, an activating antibody to CD28 was present during the interaction of the T cell and APC; n indicates the number of cells studied (34).

View this table:

To study the role of costimulatory receptor pairs in the regulation of the cytoskeletal movement, we added antibodies that block a number of accessory interactions to this system. We found that blocking ICAM-1 or B7-1 together with B7-2 abolishes bead movement while leaving undisturbed other aspects of T cell activation, such as the formation of a tight interface and the elevation of intracellular calcium (Table 1) (16). Antibodies to CD48 or VCAM had no effect (Table 1). To determine whether the interactions of ICAM-1 with LFA-1 or of B7 with CD28 are sufficient to trigger bead movement, we used transfected CHO cells as APCs. When transfected with I-Ek, the CHO cells were not able to induce bead movement (Table 1). When the CHO cells were transfected with I-Ekand either ICAM-1 or B7-2 gene constructs, we observed a partial restoration of the bead movement (Table 1). The best results were obtained by transfecting ICAM-1 and engaging CD28 on the T cell with an activating antibody, which allows bead movement to near B cell levels (Table 1). Preliminary experiments with CHO cells expressing all three molecules (I-Ek, ICAM-1, and B7-2) showed that 7 of 10 T-CHO couples examined moved beads toward the T cell–APC interface (16). These experiments show that the interactions of LFA-1–ICAM-1 and CD28-B7 are both necessary and sufficient to induce movement of the T cell cortical actin cytoskeleton toward the newly formed T cell–APC interface.

We have also characterized second messengers that are synergistically activated by the TCR and the accessory interactions that are important for bead movement, namely the intracellular calcium concentration for ICAM-1–LFA-1 (9,23) and the phosphatidylinositol 3-kinase (PI 3-kinase) for B7-CD28 (24). Blocking the rise in the intracellular calcium concentration with 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA) and Ni2+ blocks bead movement (Table 1) while leaving the formation of a tight interface intact (9). Using the PI 3-kinase inhibitor wortmannin (25) at 100 nM leads to a distinct phenotype in that most T cells do not establish a polarized phenotype before activation. For the minority of T cells that display a polarized phenotype, bead movement is abolished. These results are consistent with both PI 3-kinase and intracellular calcium being necessary for bead movement.

The cortical actin cytoskeleton moves either by controlled assembly and disassembly of actin filaments or by sliding actin filaments past each other using myosin motor proteins (11, 26). In particular, myosin motor proteins are thought to be involved in actin cytoskeleton movement at the cell body (21). Previously, we and others have shown that blocking actin filament assembly in the T cell (but not B cell APCs) with cytochalasin D severely disrupts many aspects of T cell activation (3, 9). Inhibiting the myosin motors with butanedione monoxime (BDM) (27) gives a more specific phenotype in this system, as shown in Fig. 2 (movie 2) and Table 2. At high concentrations of BDM, the diameter of the T cell–B cell interface is reduced, the calcium signal is less sustained, the redistribution of ICAM-1 to the T cell–B cell interface is impaired, the redistribution of I-Ek is inhibited (22), and bead movement is partially blocked (Table 2). At low BDM concentrations (2 mM), the reduction in interface diameter and the impairment of bead movement remain (Table 2), suggesting that myosin motors are most involved in these two processes. Because BDM has also been reported to block T cell potassium uptake (28), we inhibited voltage-gated potassium channels specifically with noxius toxin (29) and found no inhibition of bead movement (Table 2). The small effect we see on early T cell activation is consistent with the moderate effects reported for specific potassium channel blockers on later activation events (30). We conclude that the receptor/cytoskeletal movement that we describe here is dependent on myosin motor proteins.

Figure 2

T cell–APC interaction with blocked myosin motor proteins. (A) Single frames of a video microscopy experiment (36) of the interaction of 5C.C7 T cells, loaded with 4.5-μm anti–ICAM-1 (YN1) beads, with peptide-loaded, ICAM-1–GFP–transfected CH27 cells are shown. The bright-field image has been duplicated. In the top panels, the T cell intracellular calcium concentration is overlaid in a false color scale as in Fig. 1. In the bottom panels, the ICAM-1–GFP fluorescence is overlaid in a false color scale from green (low fluorescence) to blue (high fluorescence) to allow the simultaneous investigation of the additional cellular parameter. The accumulation of ICAM-1 at the T cell–APC interface in the absence of all pharmacological agents was described in (9); here, such an interaction is shown in the presence of 2 μM BDM. Although the calcium signal is elevated in a stable manner and ICAM-1–GFP accumulates at the interface, this interface is narrow (as most easily seen by the tightly concentrated ICAM-1–GFP accumulation), and a bead that is bound to an activating T cell, marked with an arrow, does not move. The still frames were excerpted from movie 2 (38), which can be viewed atScience Online (www.sciencemag.org). (B) An even narrower interface at 20 μM BDM, excerpted from a movie of a different experiment.

Table 2

Effects of blocking myosin motor proteins. The analysis of T cell activation by peptide loaded CH27 cells in the presence of BDM or noxius toxin (NTX) is shown (42). Narrow interface form denotes activated T cells having a T cell–B cell interface diameter that is smaller than the diameter of the T cell (Fig. 2). Calcium signal back to baseline denotes activated T cells whose intracellular calcium concentration returns to preactivation levels during an observation period of at least 5 min. ICAM cluster formation denotes CH27 cells that show accumulation of ICAM-1–GFP at the T cell–B cell interface after T cell activation (9) (Fig. 2). Bead movement toward the interface (39) refers to activated T cells that show anti–ICAM-1 bead movement. In all cases,n denotes the number of cells analyzed (43).

View this table:

In vivo, costimulation is usually required for efficient T cell activation, probably because TCR signaling is limited by small numbers of agonist peptide–MHC complexes and continuous internalization of the TCR-CD3 complex. Previously, we have shown that the accumulation of MHC-peptide–TCR and accessory receptor couples at the T cell–APC interface after T cell activation can amplify the weak TCR signal efficiently (9, 22). Here, we have described an active movement of the cortical actin cytoskeleton toward the interface, involving myosin motor proteins, that would move all molecules linked to the actin cytoskeleton toward the T cell–APC interface. In support of this possibility, we find that the increase in concentration of LFA-1–ICAM-1 and other (TCR–peptide–I-Ek and CD2-CD48) receptor couples at the interface and the movement of the cortical cytoskeleton are regulated similarly. Both processes are sensitive to the myosin inhibitor BDM, to anti-B7, and to the choice of APC (9, 22) (Table 2). Furthermore, we have shown that this movement of the actin cytoskeleton is regulated by the two most important costimulatory receptor pairs, CD28-B7 and LFA-1–ICAM-1. We suggest that this active accumulation of receptor pairs and other cytoskeleton-linked molecules at the T cell–APC interface, and the signal amplification that would result from these increased receptor densities, could be the principal basis of the costimulatory effect. This is in contrast to models in which a costimulatory signal integrates with TCR signals in the nucleus to affect gene expression.

REFERENCES AND NOTES

View Abstract

Navigate This Article