The inflammatory response following ischemic stroke is dominated by innate immune cells: resident microglia and blood-derived macrophages. The ambivalent role of these cells in stroke outcome might be explained in part by the acquisition of distinct functional phenotypes: classically (M1) and alternatively activated (M2) macrophages. To shed light on the crosstalk between hypoxic neurons and macrophages, an in vitro model was set up in which bone marrow-derived macrophages were co-cultured with hippocampal slices subjected to oxygen and glucose deprivation. The results showed that macrophages provided potent protection against neuron cell loss through a paracrine mechanism, and that they expressed M2-type alternative polarization. These findings raised the possibility of using bone marrow-derived M2 macrophages in cellular therapy for stroke. Therefore, 2 million M2 macrophages (or vehicle) were intravenously administered during the subacute stage of ischemia (D4) in a model of transient middle cerebral artery occlusion. Functional neuroscores and magnetic resonance imaging endpoints (infarct volumes, blood-brain barrier integrity, phagocytic activity assessed by iron oxide uptake) were longitudinally monitored for 2 weeks. This cell-based treatment did not significantly improve any outcome measure compared with vehicle, suggesting that this strategy is not relevant to stroke therapy.