Stroke attracts neutrophils to the injured brain tissue where they can damage the integrity of the blood–brain barrier and exacerbate the lesion. However, the mechanisms involved in neutrophil transmigration, location and accumulation in the ischemic brain are not fully elucidated. Neutrophils can reach the perivascular spaces of brain vessels after crossing the endothelial cell layer and endothelial basal lamina of post-capillary venules, or migrating from the leptomeninges following pial vessel extravasation and/or a suggested translocation from the skull bone marrow. Based on previous observations of microglia phagocytosing neutrophils recruited to the ischemic brain lesion, we hypothesized that microglial cells might control neutrophil accumulation in the injured brain. We studied a model of permanent occlusion of the middle cerebral artery in mice, including microglia- and neutrophil-reporter mice. Using various in vitro and in vivo strategies to impair microglial function or to eliminate microglia by targeting colony stimulating factor 1 receptor (CSF1R), this study demonstrates that microglial phagocytosis of neutrophils has fundamental consequences for the ischemic tissue. We found that reactive microglia engulf neutrophils at the periphery of the ischemic lesion, whereas local microglial cell loss and dystrophy occurring in the ischemic core are associated with the accumulation of neutrophils first in perivascular spaces and later in the parenchyma. Accordingly, microglia depletion by long-term treatment with a CSF1R inhibitor increased the numbers of neutrophils and enlarged the ischemic lesion. Hence, microglial phagocytic function sets a critical line of defense against the vascular and tissue damaging capacity of neutrophils in brain ischemia. Electronic supplementary material The online version of this article (10.1007/s00401-018-1954-4) contains supplementary material, which is available to authorized users.

Inflammatory Ly6ChiCCR2+ monocytes infiltrate the brain after stroke but their functions are not entirely clear. We report that CCR2+ monocytes and CCR2+ lymphocytes infiltrate the brain after permanent ischemia. To underscore the role of CCR2+ monocytes, we generated mice with selective CCR2 deletion in monocytes. One day post-ischemia, these mice showed less infiltrating monocytes and reduced expression of pro-inflammatory cytokines, markers of alternatively macrophage activation, and angiogenesis. Accordingly, Ly6Chi monocytes sorted from the brain of wild type mice 24 h post-ischemia expressed pro-inflammatory genes, M2 genes, and pro-angiogenic genes. Flow cytometry showed heterogeneous phenotypes within the infiltrating Ly6ChiCCR2+ monocytes, including a subgroup of Arginase-1+ cells. Mice with CCR2-deficient monocytes displayed a delayed inflammatory rebound 15 days post-ischemia that was not found in wild type mice. Furthermore, they showed reduced angiogenesis and worse behavioral performance. Administration of CCR2+/+ bone-marrow monocytes to mice with CCR2-deficient monocytes did not improve the behavioral performance suggesting that immature bone-marrow monocytes lack pro-reparative functions. The results show that CCR2+ monocytes contribute to acute post-ischemic inflammation and participate in functional recovery. The study unravels heterogeneity in the population of CCR2+ monocytes infiltrating the ischemic brain and suggests that pro-reparative monocyte subsets promote functional recovery after ischemic stroke.