Seeing Is Believing: Sorting Out Signaling Events at the Immunological Synapse

This month’s Pillars of Immunology section features two papers that form the foundation for our current understanding of the immunological synapse, the specialized cell/cell contact site formed between a T cell and an APC (1, 2). Today, we know that this structure serves as a platform for the assembly and segregation of signaling complexes, defining the outcome of T cell activation and targeting effector function. In the mid-1990s, however, little was known about the cytoarchitecture of T cell signaling. Biochemical analysis had revealed that TCR engagement leads to rapid Ca²⁺ influx and initiates a tyrosine phosphorylation cascade, but many questions were unresolved. There was evidence that the signals produced by stimulatory Abs differed from those produced by APCs, but the basis for this was unknown. The role of other receptor/ligand pairs had not been delineated, and it was unclear how low-affinity TCR/peptide/MHC (pMHC) interactions could generate the sustained signaling events needed for full T cell activation. Finally, whereas TCR/pMHC kinetics were known to influence T cell activation, it was unclear how the T cells integrate and interpret variations in signal strength. The articles highlighted in this issue provided a structural context for understanding signaling at the T cell/APC interface. In the first, Kupfer and colleagues (1) showed that signaling molecules are spatially segregated at the T/B interface, raising the idea that individual receptor signaling complexes interact to facilitate and tune TCR signaling. In the second, Dustin and colleagues (2) showed that these structures are highly dynamic and correlated their formation with the signaling events associated with full T cell activation.

The story of the immunological synapse began with a flurry of papers in the mid-1980s, which showed that interaction of T cells with APCs involves the formation of a specialized contact site enriched in receptors and adhesion molecules, together with their respective ligands on the APC surface (reviewed in Ref. 3). Initially, much of this work focused on understanding the mechanisms through which T cell/APC engagement led to polarization of cytoskeletal elements and secretory organelles toward the bound APC, setting the stage for polarized delivery of cytokines or cytotoxins. As the field matured and new cytokines and receptor/ligand pairs were identified, investigators began to think of the T cell/APC contact site as a specialized site for cognate interactions and bidirectional signaling. In a 1994 review article, Paul and Seder (4) clearly articulated this idea and coined the term “immunological synapse.”

During the next several years, investigators began to address T cell activation by APCs and model stimulatory surfaces (5–7). Working on this backdrop, Monks et al. (1) published images that captivated the imagination of the immunology community and quite literally brought the topic into focus. Using the newly developed technique of deconvolution microscopy, they removed out-of-focus haze from images of Ag-specific T cells and peptide-loaded B cells, yielding stunning three-dimensional images of molecular organization at the cell/cell interface (1). This analysis revealed that synapse proteins are segregated into specialized domains, which the authors termed supramolecular activation clusters (SMACs), with CD3 and interacting pMHC complexes concentrated in the central region (the central SMAC [c-SMAC]), and LFA-1 organized into an outer ring (the peripheral SMAC [p-SMAC]). Importantly, Monks et al. showed that this organization extended beyond plasma membrane receptors; signaling molecules, including protein kinase Cθ, Fyn, and Lck, as well as structural proteins such as talin were also organized into c-SMAC and p-SMAC regions, respectively. By varying T cell specificity and antigenic peptides, they went on to show that molecular segregation occurs only for cognate receptor/ligand pairs and that TCR signaling is required; null peptides or antagonist peptides that engage TCR without generating productive signals are insufficient to drive segregation. Taken together, these findings revealed that T cell activation involves spatial control of protein/protein interactions, such that “regulated segregation of activated receptors and their associated proteins would generate unique signals that would not be otherwise generated by randomly coaggregated receptors” (1).

While the Kupfer laboratory was studying protein segregation with high spatial resolution, Dustin and colleagues (6) had been making new tools to study this process with high temporal resolution. Their approach, also used successfully by the Davis laboratory (7), used supported planar lipid bilayers containing fluorescently labeled T cell ligands as artificial stimulatory surfaces. This allowed quantitative, real-time imaging of receptor segregation, revealing features that were not evident in static images. Indeed, Grakoui et al. (2) showed that the pattern described by Kupfer represents a late stage of a dynamic, multistep process that occurs during the span of 10–15 min. At early times of immunological synapse formation, they observed an inside-out pattern in which the central region is enriched in LFA-1 and its ligand ICAM-1 and is surrounded by a ring of TCR. This configuration, they argued, provides anchorage as the T cell lamellipodium spreads over its substrate, sampling pMHC complexes. Following this initial spreading phase is a contraction phase during which the pMHC complexes are transported toward the center of the contact site. Finally, the process culminates in the formation of the mature immunological synapse, which is essentially the pattern described by Monks et al. The agreement between these two very different studies lent credence to both, and indicated that the bull's-eye structure is dictated by the T cell, rather than the APC.

In addition to describing the dynamics of synapse formation, Grakoui et al. made important strides toward relating this process to signaling events associated with T cell activation. As reported by Monks et al. for T/B contacts, they found that only agonist peptides could induce protein segregation, and they showed that the accumulation of pMHC complexes under the T cell is dependent on TCR binding kinetics. They compared the minimal peptide doses needed for synapse formation and proliferation and found them to be similar and consistent with peptide densities needed for T cell stimulation by APCs. Although it did not prove that synapse formation is needed for T cell activation, this evidence, similar to evidence in the Monks paper, showed a strong correlation. Interestingly, however, Grakoui et al. found that the initial rise in intracellular Ca²⁺ occurs at early times, when the synapse has an inside-out structure. This important finding shows that the mature immunological synapse forms too late to be required for initial TCR signaling, although it can certainly play a role in sustained Ca²⁺ responses and additional downstream signaling events.

By demonstrating that T cell signaling involves a high degree of spatiotemporal organization, these papers provided an essential framework for thinking about how T cells can integrate signals from rare pMHC complexes with low binding affinity, taking into account costimulatory signals and other contextual cues. In the years since this work was published, progress in the field has continued to be driven by advances in imaging technology. Just as Kupfer’s initial observations were made possible by the use of deconvolution microscopy and Dustin’s observations by the use of planar lipid bilayers, total internal reflection fluorescence imaging has revealed the existence of dynamic signaling microclusters, the functional units that make up the larger synapse domains (8, 9). Most recently, superresolution imaging is providing new insights into specific protein/protein interactions and the role of membrane dynamics in synapse function (10, 11). As a result of these and other advances, our understanding of the immunological synapse has by now been revised substantially. We now know that the canonical bull's-eye structure is only one of several patterns, which vary with, for example, T cell developmental state, APC type, agonist strength, etc. (12), and that additional variations are seen in NK cells, B cells, and regulatory T cells (13–15). The central c-SMAC region, initially thought to be a site for sustained signaling, is now thought to represent a site for signal termination and membrane trafficking, and it is now clear (16) that most early signaling events occur in newly formed microclusters near the periphery (although this, too, varies with the stimulus) (17). Models for synapse function continue to be refined and new questions can be asked. Do specific protein/protein interactions take place on the cell surface, in the cytoplasm, or on intracellular vesicles? How do the thermodynamics of receptor/ligand interactions affect the formation and persistence of signaling complexes? What are the roles of mechanical force, ligand mobility, and membrane stiffness? And how does all this play out in a tissue setting? Although many questions remain, all trace back to these pioneering papers, which defined the stage on which the captivating molecular dance of T cell signaling is performed.