[The immunological synapse: models facing facts]

The notion of immunological synapse is generally associated to a concentric structure (a core of T cell receptors surrounded by a ring of adhesion molecules) often called "mature synapse". This schematic view has been built on observations corresponding to peculiar experimental conditions: very high antigen concentration presented by surrogate APCs such as lipid bilayers or B lymphoma. These observations have been hastily constituted in a dogma that a "normal" synapse should look like this, should form only in the presence of antigen, and should trigger a "stop" signal that completely immobilizes the T cell. However, when analyzing the interaction between naive T cells and dendritic cells (DC), that are the only antigen-presenting cells able to activate naive T cells, a very different picture emerges. Firstly, T-DC synapses can form in the absence of antigen; therefore antigen recognition is not a prerequisite for synapse formation. Secondly, these antigen-independent synapses are likely to play several roles, including sensitization of T cells for later antigen detection, and delivery of survival signals. Thirdly, in vivo, naive T cells interacting with antigen-laden DC do not fully stop, but start to make transient contacts with DCs for a few minutes, before continuing their exploration. It is only after several hours of this process that T cells eventually immobilize. Fourthly, the structure of the T-DC synapse is clearly multifocal, the two cells interacting through several tens of tight appositions of a few tens of nm in diameter. These numerous tight appositions are reminiscent of the microclusters that have been recently described at the T-bilayer interface. Finally, synaptic signaling is not a transient initial event, but is sustained for hours. In particular, sustained activation of phosphatidylinositol 3-kinase allows the exclusion out of the nucleus of FoxO transcription factors, normally maintaining T cells in a quiescent state.