Growing experimental evidences highlight microglial alterations in psychiatric diseases, including depression and stress-related disorders. The central goal of my thesis was to investigate microglia in a mouse model of chronic-stress consisting of 5 days of Repeated Forced Swim Stress (RFSS). Mice exposed to RFSS the paradigm exhibit behavioral changes resembling depressive-like behavior. I first analyzed microglia by confocal imaging in the hippocampus. My analysis showed that the number of microglial cells was unaltered in mice undergoing the RFSS paradigm. However, microglia exhibited an increased ramification of the processes (hyper-ramified morphology) in the hippocampal regions DG, CA1 and CA3. By contrast, I found no change in microglial morphology or cell number in brain regions that are presumably not affected by chronic-stress (somatosensory cortex, caudate-putamen striatal nucleus, cerebellum). Since the phenotypic and functional properties of hyper-ramified microglia are unknown, I analyzed the expression of cytokines and microglial activation markers. By qPCR, I showed that the expression of the classical pro-inflammatory cytokines CCL2, IL6 and IL1β, and of the macrophage-activation marker CD11b was increased in hippocampal microglia upon the RFSS paradigm. By confocal imaging, I showed increased CD11b immunoreactivity and phagosomes volume in hippocampal microglia from RFSS mice. These data suggested that RFSS promotes a pro-inflammatory phenotype in hippocampal microglia. Next, I assessed the motility of microglia in CA1 by in vivo two-photon microscopy. Time-lapse imaging in Cx3cr1GFP/+ mice (with GFP-positive microglia) before and immediately after the RFSS paradigm showed that RFSS markedly reduced the motility of microglial processes. Together, these data showed that RFSS affects the homeostatic functions of microglia in the hippocampus, however, it was not yet clear whether microglia are important for the behavioral phenotype of the RFSS model. To answer this question, I investigated the IL34 knock-out mouse line, exhibiting a 50% reduction in the number of microglial cells. The cytokine IL34 is a ligand for CSF1-receptor, expressed in microglia and required for their survival. My data showed that IL34 knock-out mice did not develop RFSS-induced behavioral changes, whereas learning and memory skills at the Morris-Water Maze were unaffected. These data suggest that hyper-ramified microglia play a crucial role for the behavioral phenotype of the RFSS model. In line with this assumption, I could show that LPS injection in RFSS mice induced a rapid de-ramification of microglial processes, along with a partial “rescue” of the RFSS-induced behavioral changes. Further studies are needed to understand the biological underpinnings of this mechanism.
Because TNFα was found to be increased in depressed patients, and microglia are the main producer of TNFα in the bran, I hypothesized that microglia may elicit RFSS-induced behavioral changes in a TNFα-dependent manner. To test this, I analyzed mice harboring a conditional genetic deletion of the TNFα locus in microglia. My results showed that mice with TNFα-deficient microglia exhibit RFSS-induced behavioral changes indistinguishable from wild-type. These data suggested that microglia-derived TNFα does not play an important role at the RFSS paradigm. I then hypothesized that hyper-ramified microglia in the hippocampus may affect synaptic function during the RFSS paradigm. Hence, I analyzed the number of glutamatergic synapses in CA1 and somatosensory cortex in RFSS and control mice by confocal imaging. In CA1 of RFSS mice the number of glutamatergic synapses was significantly reduced whereas the somatosensory cortex was unaffected. Given that hyper-ramified microglia were found in the CA1, but not in somatosensory cortex, it is possible that microglia played a role in the synaptic loss.
It was shown that microglia can sense the activity of glutamatergic synapses via the purinergic receptors. Moreover, growing evidences suggest that the brain’s purinergic signaling is involved in chronic-stress. I then set out to investigate the role of P2Y12R (a microglial purinergic receptor) in the RFSS model. To do so, I tested both P2Y12R knock-out and wild-type mice to the RFSS paradigm. Interestingly, P2Y12R knock-out mice did not develop RFSS-induced behavioral changes. Moreover, RFSS-induced microglia hyper-ramification and synaptic loss in CA1 were partially reduced in P2Y12R-deficient mice. With a following experiment, I showed that wild-type mice treated with a P2Y12R-inhibitor did not exhibit RFSS-induced behavioral changes, indicating that pharmacological blockage of the P2Y12R-signalling in wild-types recapitulates the RFSS-resilient phenotype of the P2Y12R knock-out mice. Together, these data emphasize the importance of microglia in this model of RFSS and reveal a previously unappreciated role for the P2Y12R signaling during chronic-stress.