NeutrophilsNeutrophils are among the first cells attracted to the brain after ischemic stroke where they are detected in the microvessels within the first hour7 and peak at 1 to 3 days.4,7,8 Neutrophils are short-lived innate immune cells containing different types of granules with antimicrobial pro-oxidant and proteolytic enzymes that can damage the tissues. Accordingly, neutrophils are regarded as detrimental following compelling evidence associating these cells with blood-brain barrier breakdown and brain injury.7,9 Also, higher blood neutrophil counts are associated with larger infarct volumes in acute ischemic stroke patients.10 Nevertheless, the pathogenic role of neutrophils in ischemic stroke is still not conclusive. For instance, there are conflicting results in the literature on the potential benefit of neutrophil depletion in experimental ischemia models.9,11 Furthermore, we lack entire demonstration that neutrophils reach the ischemic tissue before substantial neuronal death has occurred.11 Nonetheless, neutrophils can exert detrimental effects already from the vessel wall. Adhesion of neutrophils to the inflamed endothelium after ischemia/reperfusion is involved in the no-reflow phenomenon, obstructing blood flow in precapillary arterioles, postcapillary venules, and the capillary bed.7,12 In addition, neutrophils in the vessel lumen and at perivascular locations can damage the blood-brain barrier by releasing proteolytic enzymes and pro-oxidant molecules (Figure).9 Moreover, neutrophils can produce NETs (neutrophil extracellular traps) promoting clot formation.13 NETs can precipitate thrombotic events and impair tPA (tissue-type plasminogen activator)-induced thrombolysis.14 In turn, thrombolysis may exacerbate detrimental effects of neutrophils because tPA promotes neutrophil transmigration to the reperfused tissue by proteolytic activation of plasmin and matrix metalloproteinases.15 These effects might contribute to explain why neutrophilia and high neutrophil-to-lymphocyte ratio are associated with the risk of hemorrhagic transformation in ischemic stroke patients treated with tPA.16 After permanent middle cerebral artery occlusion (MCAo) in mice, we observed the formation of intravascular NETs and found NETs in perivascular locations and in the brain parenchyma.17
There is also some controversy on whether neutrophils actually reach the ischemic brain parenchyma at all. An elegant study by Enzmann et al18 noticed the massive accumulation of neutrophils in perivascular spaces surrounding venules within the ischemic tissue after ischemia/reperfusion in mice. Most neutrophils remained perivascular, and only a few were detected in the brain parenchyma.18 This is an important observation because it highlights that perivascular spaces are a niche for neutrophils where they accumulate after transient MCAo. We also detected neutrophils in perivascular locations and leptomeningeal spaces in the mouse after permanent MCAo.17 These results suggest that, besides extravasating from intracerebral venules, neutrophils extravasate from leptomeningeal vessels and migrate from the subpial space along the vessels penetrating the cortex.17 However, our study17 and previous studies7 found neutrophils in the ischemic brain parenchyma using models of permanent MCAo. Ischemic conditions involving severe endothelial damage, vessel rupture, and microbleeds or hemorrhagic transformation, are expected to facilitate the presence of neutrophils in the brain parenchyma. Other conditions, such as high blood glucose, also promote neutrophil infiltration.19 A recent study20 analyzed the postmortem brain of 16 ischemic stroke patients and confirmed the presence of neutrophils in the leptomeninges and perivascular spaces, but neutrophils were rare in the infarcted parenchyma with the exception of 1 patient deceased 3 days after stroke with no signs of infection. Interestingly, the time to death of this series of patients was 1 day in 2 cases, 3 days in the case above mentioned, and then times ranged from 8 days to 240 days poststroke.20 Neutrophils display a specific time-window of attraction to the damaged tissues after acute injuries, and they have a short life in tissues. Therefore, the time to death of ischemic stroke patients is critical to look for the presence of neutrophils in the brain parenchyma. More studies of human tissue within the first days poststroke are necessary to understand under which conditions neutrophils might gain access to the infarcted brain parenchyma.
Despite many advances, there are aspects of neutrophil behavior in stroke that are still difficult to interpret. For instance, neutrophils with anti-inflammatory and repair phenotypes were found in the ischemic brain tissue of experimental animals,21 neutrophils of ischemic stroke patients show a reduced oxidative burst and NET formation,22 and microglia surrounding blood vessels phagocyte neutrophils.23 The possibility that neutrophils were passive bystanders under some circumstances but active players in others depending on specific features of the ischemic lesion needs further consideration.
Monocyte/MacrophagesAfter brain ischemia, microglia acquire a reactive morphology resembling macrophages. Classically, immunohistochemical studies have described the presence of reactive microglia/macrophages peaking at ≈4 days postischemia in rats or mice, but it was not possible to distinguish whether these cells derived from resident microglia or they infiltrated from the periphery. Nowadays, flow cytometry, cell type-specific fluorescent reporter mice, adoptive transfer of fluorescent cells, generation of chimeras, and recently identified specific microglia markers, allow differentiating resident reactive microglia from infiltrating macrophages. Monocyte infiltration is detected within the first 24 hours postischemia, peak at 4 days, and some of these cells persist for weeks and acquire features of tissue macrophages. Immature CCR2+Ly6Chi proinflammatory monocytes are the subset of monocytes first attracted to the ischemic brain tissue.24–26 These cells might be released by the bone marrow, but a study reported that monocytes reaching the ischemic brain originate in the spleen.25
Infiltrating macrophages were classically associated with inflammation and brain damage after ischemic stroke. In mice, monocyte infiltration is largely dependent on CCR2 (C-C motif chemokine receptor type 2), the receptor of the chemokine CCL2 (C-C motif chemokine ligand 2), also known as MCP1 (monocyte chemoattractant protein 1). To investigate the role of monocytes, several studies used CCR2-deficient mice or CCR2 inhibitors, with the limitation that besides the subset of Ly6Chi monocytes other cells, like some T cells, also express CCR2. CCR2-deficiency reduced the ischemic brain lesions in mice.27 Challenging this view, CCR2 drug inhibitors exacerbated the brain lesion.28 Furthermore, anti-CCR2 blocking antibodies impaired spontaneous long-term functional recovery,29 depletion of monocytes/macrophages worsened the ischemic lesion,30 and infiltrating macrophages prevented hemorrhagic transformation of the ischemic lesion.24 By systemic injection of fluorescent monocytes after brain ischemia, we observed fluorescent cells in the subpial space, and along the vessels penetrating the cortex,26 supporting the view that a subset of infiltrating macrophages establish persistent interactions with the blood vessels (Figure).
The phenotype of activated macrophages depends on the environmental stimuli. The M1 and M2 phenotypes are prototypical states of macrophage polarization achieved in culture after exposure to certain cytokines. The M1 phenotype is proinflammatory whereas the M2 phenotype promotes resolution of inflammation and repair. Macrophages infiltrating the ischemic tissue, including the Ly6Clo population and some of the Ly6Chi monocytes, acquire features of alternatively polarized M2 macrophages during the first week postischemia.26,28–30 Studies of human ischemic infarcts reported that macrophages initially showed proinflammatory features that with lesion maturation transformed into anti-inflammatory phenotypes.20 Interestingly, a study noticed that after ischemia in mice, the expression of M2 markers increased within the first week but then decreased, whereas proinflammatory markers persisted and predominated at week 2, suggesting a long-lasting inflammatory status.31
The factors that contribute to the time-dependent changes in macrophage phenotypes in the ischemic brain tissue are not entirely identified. Increased anaerobic glycolysis and activation of the hypoxia-inducible factor-1 are associated with proinflammatory M1 phenotypes, whereas energy production in M2 phenotypes rather relies on fatty acid oxidation.32 In M1 activated macrophages, arginine metabolism occurs through inducible nitric oxide synthase leading to generation of reactive oxygen and nitrogen species that damage proteins, lipids, and DNA. In contrast, M2 macrophages metabolize arginine through arginase-1 generating polyamines involved in cell division and collagen synthesis, among other functions.32 However, under pathological conditions 1 single phenotypic feature may not be sufficient to attribute any specific phenotype to the cells. Likely, cellular metabolic adaptations to ischemia and reperfusion together with the cytokine environment and phagocytic activity have an impact on the phenotype and function of macrophages and microglia.
T CellsSevere stroke reduces the numbers of lymphocytes in the circulation and lymphoid organs.33 In contrast, T-cell numbers increase in the ischemic brain within the first 24 hours and can persist for long times.4 During the first hours after ischemia/reperfusion, T cells facilitate adhesion of platelets and leukocytes to the vascular endothelium34 causing a phenomenon called thromboinflammation35 by which molecular and cellular players in thrombosis and coagulation promote proinflammatory pathways exacerbating the brain lesion.36 However, the interaction of T cells with platelets may also have hemostatic effects preventing hemorrhagic transformation after severe ischemic stroke.37 T cells are found in subpial and cortical vessels and infiltrating the ischemic lesion.38,39 In addition, the choroid plexus is a gateway for T cells migrating to the periphery of cortical infarction.40 Importantly, CD8+ cytotoxic T cells were detected in human ischemic infarcts.20 Also, ischemic stroke patients show increased frequency of CD4+CD28null cells in blood associated with stroke severity and serum levels of proinflammatory cytokines.41 CD4+CD28null T cells are an interesting subset of T cells because they have enhanced effector functions, are associated with senescent T cells, and expand under inflammatory conditions.42 At later phases, CD4+ cells accumulate in the brain of mice peaking at day 14 and persisting at day 30 after ischemia/reperfusion.43 Furthermore, emerging evidence suggests that antigen-mediated T-cell responses take place in subacute or chronic stages after stroke and may worsen stroke outcome.43–47 However, in central nervous system trauma, protective autoimmunity mediated by T-cell responses is involved in promoting recovery.48 Overall, T cells seem to play innate functions and interact with players in thrombosis and hemostasis in the acute phase of stroke, whereas at later stages they exert adaptive functions that could affect stroke outcome in the long term.
γδ T CellsSubsets of unconventional innate T cells with invariant T-cell receptor could play a role in acute ischemic brain damage. Growing evidence supports that γδ T cells are pathogenic in experimental brain ischemia/reperfusion by secreting IL (interleukin)-17 and exacerbating the inflammatory response.49–51 Moreover, IL-17A+ lymphocytes were detected in the postmortem brain of stroke patients.50 Interestingly, γδ T cells are abundant in the gut from where they seem to traffic to the leptomeninges after brain ischemia.52 T helper 17 cells (Th17) and γδ T cells increase in the blood of stroke patients in association with increased levels of IL-17A, IL-23, IL-6, and IL-1β.53 In spite of the fact that IL-17 producing cells are a small subset of cells, they seem to play a prominent role in orchestrating the inflammatory response in acute stroke and exacerbating the lesion (Figure).
Regulatory LymphocytesRegulatory lymphocytes exert immunomodulatory and immunosuppressor functions. Several lines of evidence support beneficial effects of regulatory T cells (Treg)54 and regulatory B cells55 in experimental brain ischemia. However, other studies found acute detrimental effects of Treg in brain ischemia/reperfusion56 as previously reviewed.57 Although the number of Treg found in the ischemic brain parenchyma during the first days poststroke is low, Treg strongly accumulate in the ischemic lesion 15 days poststroke where potentially they could inhibit autoimmune responses.43 Increased apoptosis of Tregs, loss of Tregs in peripheral blood, and impaired suppressive function of the remaining Treg population has been reported in ischemic stroke patients.53,58,59 However, other studies reported upregulation of Tregs in stroke patients in spite that decreased Treg function was observed, particularly in female patients.60 Notably, an increased proportion of Treg cells was reported in the spleen of mice 4 days after transient MCAo.61
NK CellsNK innate lymphocytes show a rapid and transient increase in the ischemic brain tissue.4,8 A study reported no benefits of depleting NK cells in permanent or transient MCAo.44 In contrast, another study suggested pathogenic actions of NK cells by promoting inflammation and neuronal cytotoxicity.62 This study reported infiltration of NK cells in the ischemic brain tissue of humans and mice where NK cell numbers peaked as soon as 3 hours postischemia and then declined.62 Interestingly, the β2-nACh-R (nicotinic acetylcholine) receptor seems to be involved in the NK cell decline observed in the ischemic tissue from 3 hours postischemia.63 Induced-persistence of NK cells in the ischemic tissue achieved by interfering with this cholinergic receptor did not modify lesion size but increased systemic IFNγ (interferon γ), protected from bacterial infection, and enhanced poststroke survival.63 More studies are needed to validate the putative capacity of brain infiltrating NK cells to prevent poststroke infection, the role of central acetylcholine in this process, as well as the suggested rapid pathogenic effect of NK cells worsening the acute ischemic brain lesion.
Therapeutic InterventionStrategies designed to prevent negative actions of leukocytes have been taken to the clinic in acute ischemic stroke patients but with no success to date.64 Several studies with drugs blocking the action of neutrophils were investigated.11,64 As an example, the ASTIN trial (Acute Stroke Therapy by Inhibition of Neutrophils) investigated a compound known as UK-279 276, a recombinant neutrophil inhibitory factor that selectively binds the CD11b integrin of macrophage-1 antigen (CD11b/CD18).65 The treatment did not improve recovery above placebo. This trial followed encouraging results of a few preclinical studies with UK-279 276 in experimental models of brain ischemia/reperfusion.66 However, only a fifth of patients in the ASTIN trial received tPA.65
Monocyte/macrophages may acutely exacerbate the inflammatory responses, but experimental studies have identified their involvement in resolution of inflammation, vascular protection, and recovery of function,24,27–29 possibly linked to the phagocytic and vasculoprotective roles of these cells. These protective actions of monocytes are in line with the beneficial effects of administration of autologous bone marrow–derived mononuclear cells (MNC) after experimental ischemic stroke.67 MNC contain myeloid and lymphoid cells, as well as hematopoietic and mesenchymal stem cells. MNC administration 24 hours after MCAo improved functional recovery, reduced lesion size and proinflammatory cytokines, and enhanced vessel density and neurogenesis,68 and these benefits were long lasting.69 Furthermore, MNC reduced blood-brain barrier permeability and decreased the severity of hemorrhagic transformation after tPA in an embolic stroke model.70 However, MNC therapy did not improve outcome in hypertensive rats.71 MNC reach the periphery of brain infarction soon after administration,68 and then the cells seem to differentiate into smooth muscle cells and endothelial cells, incorporate into vessel walls, and enhance the growth of leptomeningeal anastomoses, the circle of Willis, and basilar arteries.69 Phase I trials administering MNC to ischemic stroke patients have shown safety, but the clinical efficacy of this cell therapy awaits demonstration.72
Anti-inflammatory treatments, such as minocycline, have not been successful in the clinic.64 Experimental evidences, including a cross-laboratory preclinical study in mouse models of brain ischemia,73 support the therapeutic potential of the IL-1Ra (IL-1β receptor antagonist). A recent clinical trial with subcutaneous administration of IL-1Ra showed safety and reduction of plasma IL-6.74 However, the analysis excluded a major clinical benefit of the treatment, and negative effects potentially attributable to interactions of IL-1Ra with tPA became apparent.74
Experimental studies support damaging effects of T lymphocytes in the acute phase of stroke. Accordingly, fingolimod, a drug approved for remitting-relapsing multiple sclerosis that sequesters lymphocytes in the lymph nodes preventing lymphocyte access to the inflamed tissues, showed beneficial effects in preclinical studies and small clinical trials in acute ischemic stroke patients, including patients receiving thrombolysis.64 By acting on S1P1 (sphingosine-1-phosphate receptor 1), fingolimod induces sustained lymphopenia, but current data do not show higher incidence of poststroke infection in patients receiving fingolimod. Fingolimod also acts on endothelial S1P1 receptor increasing vascular barrier function that might contribute to the observed benefits of this drug in ischemic stroke. Given that the benefits of fingolimod seem to be mediated by S1P1, whereas certain side effects are dependent on other S1P receptors, selective S1P1 agonists were studied in experimental stroke showing reduced lesion size after ischemia/reperfusion in mice.75 In contrast to the benefits of blocking T-cell trafficking, systemic administration of regulatory T lymphocytes in rodent models of ischemic stroke reduced infarct size, ameliorated the neurological functions,76 and reduced hemorrhagic transformation after tPA.77
Leukocyte recruitment to inflammatory sites is attenuated by blocking α4β1 integrin (VLA-4 [very late antigen-4]) with natalizumab, an antibody in clinical use for multiple sclerosis treatment. Blockade of VLA-4 with CD49d antibody was investigated in a multicentric preclinical study in mice using 2 different models of cerebral ischemia.78 CD49d antibody attenuated leukocyte infiltration and reduced infarct volume in small cortical lesions but not in large infarctions. Natalizumab was investigated in ischemic stroke patients in the ACTION trial (Effect of Natalizumab on Infarct Volume in Acute Ischemic Stroke).79 Natalizumab did not meet the primary end point of the study, but secondary and exploratory end points suggested improvement of clinical outcomes,64,79 encouraging the second ACTION2 trial. This phase-IIb trial was recently completed and the notes released by the sponsor (Biogen) state that natalizumab did not improve clinical outcomes compared with placebo.