, 2008 and Zenisek et al., 2000]). In WT astrocytes (data not shown) and in Tnf−/− astrocytes incubated with TNFα, ( Figure 4E) the two pools underwent exocytosis in a clear biphasic temporal sequence: during the first phase (0–400 ms) most of the fusing vesicles belonged to the “resident” pool (80.6%, n = 7 cells), whereas during the second phase (500 ms–2 s), to the “newcomers” pool (82.5%). This temporal segregation reflects the different readiness to fusion of the two pools, in particular the fact that most “resident” vesicles, contrary to “newcomers,”

have already undergone the docking steps and are ready for fusion (i.e., are functionally docked [ Ohara-Imaizumi et al., 2007 and Toonen et al., 2006]). However, in Tnf−/− Bortezomib astrocytes,

the situation was very different. Events attributable to “residents” decreased in percentage (20% instead of 40%; n = 3680 vesicle fusions analyzed, n = 7 cells). Moreover, importantly, events due to “residents” IDH inhibitor and “newcomers” occurred randomly, without the expected temporal segregation. This indicates that even the residual “resident” pool seen in Figure 4A is defective in Tnf−/− astrocytes, because it is not ready/competent to fuse. Most likely, these vesicles dock only transiently and, like all the others, most in the absence of TNFα are hampered in reaching the stage of functional docking

allowing them to undergo rapid fusion ( Toonen et al., 2006). We conclude that constitutive TNFα is necessary for the correct reception of glutamatergic vesicles to release sites, a precondition for efficient exocytosis upon stimulation. In parallel TIRF experiments, we studied local submembrane Ca2+ events, previously shown to be temporally locked to exocytic events (Marchaland et al., 2008). Indeed, in WT astrocytes, 2MeSADP stimulation induced a burst of submembrane Ca2+ events whose temporal pattern mirrored the one of VGLUT1-pHluorin fusion events, with two peaks of Ca2+ events, each one slightly preceding the corresponding peak of vesicular fusions (Figure 4B, inset). Importantly, and in full agreement with the observations in situ, this pattern was totally preserved in Tnf−/− astrocytes ( Figure 4C, inset), further confirming that TNFα does not act on the coupling between GPCR and [Ca2+]i elevation, and indicating that this step of gliotransmission can be perfectly normal while the downstream signaling is dramatically defective.