Then, by analyzing pure neuronal cultures from DAP12KI P0 pups, we investigated the impact of the prenatal period in inducing late synaptic defects. Finally, as we observed that DAP12KI animals displayed a transient microglial activation at birth, we compared their synaptic phenotype with the one of WT animals having experienced a pharmacologicallyinduced LY2835219 CDK inhibitor inflammation during their fetal development. We conclude from both genetical and pharmacological models that prenatal activation of microglia has a delayed impact on synaptic function. To test whether fetal activation of microglia could be sufficient to induce delayed synaptic defects, we induced prenatal pharmacological inflammation and evaluated the consequences on synaptic function. In mice, the generation of hippocampal neurons starts at around E14, and microglia begin to invade parenchyma at around E15. We thus choose to induce an inflammation at E15. We induced inflammation by injecting a low dose of E. coli lipopolysaccharide intraperitoneally into pregnant dams. Low doses of lipopolysaccharide do not cross the placental barrier. Therefore, the fetal inflammation is not mediated by the injected agent itself, but rather by the maternal immune response, possibly in coordination with fetal immune cells. This PF-04217903 c-Met inhibitor treatment did not alter the birth date or the size of the litters. To check whether this treatment mimicked the inflammation in DAP12KI pups, we analyzed the microglial density in the offspring born to LPS-injected mother. As shown in Figure 4B, microglial density was increased in the hippocampus of P0 born to LPS-injected dams. This increase was comparable with the one observed in pups. We next evaluated the consequence of such prenatal inflammation on glutamatergic transmission. To do so, we first measured the ratio of AMPAR versus NMDAR EPSCs in an adult born to inflamed mother. As in DAP12KI mice, the relative contribution of AMPAR to evoked EPSCs was enhanced in hippocampal slices taken from adult mice that had been subjected to fetal inflammation. In order to evaluate whether LPS-induced prenatal inflammation also impairs synaptic function with delay, we analyzed the neuronal activity of neurons cultured from P0 pups born to control and to LPS-injected dams, both in the absence of microglia. In basal conditions, neuronal activity, monitored by calcium concentration fluctuations, was higher in neurons from maternally inflamed P0, as compared with controls. Yet, calcium fluctuations in the presence of CNQX were significantly lower in neurons from maternally inflamed P0 than from controls. This resembles what we observed in neurons cultured from DAP12KI hippocampus.