Glial Cells: Role of the Immune Response in Ischemic Stroke

Microglial Activation

Microglial activation is the first step in the inflammatory response after ischemic brain injury (27). When ischemia occurs, native microglia are rapidly mobilized to the site of injury where they undergo morphological changes corresponding to the reduction of cerebral blood flow and energy deprivation (usually manifested as hypertrophy of the cell body), the development of motile branches, or migration of the somata (28–30). These morphological changes are closely associated with the function of activated microglia (31, 32).

Microglia are also associated with blood–brain barrier (BBB) disruption after ischemic stroke. Ischemic insults lead to junctional protein phosphorylation, translocation, or degradation, which then increase BBB permeability (33). Microglia release MMPs, such as MMP-9, which subsequently promote BBB breakdown (34). In addition, various cytokines, and chemokines produced by microglia can upregulate endothelial cell adhesion molecules and promote leukocyte infiltration (35). Neutrophils, infiltrated leukocytes, are also a major source of MMPs (36, 37). When reperfusion begins, activated microglia engulf endothelial cells via phagocytosis, which allows the entrance of blood serum components (38). These morphological changes are closely associated with the function of activated microglia (31, 32). Microglia can monitor the extracellular space and adjacent cell surfaces, and most of these can induce microglial activation. Damage-associated molecular pattern molecules (DAMPs), which include high-mobility group box 1 (HMGB1) (39), extracellular PRX family proteins (40, 41), and galectin-3 (42), are the main factors that activate microglia. Yenari et al. have described the signaling pathways driving microglial activation and its related transduction events (43). Several factors influence ischemia, including toll-like receptors (TLRs), particularly TLR4, HMGB1, chemokine and cytokine receptors, purinergic receptors such as P2X7 and P2Y12, glutamate receptors, and the triggering receptor expressed on myeloid cells 2 (TREM2). These factors are integrated through several signal transduction pathways, such as the mitogen-activated protein kinase (MAPK) cascade, NF-κb, peroxisome proliferator activated receptor (PPAR), and others. Further studies on microglia activation signaling pathways after ischemic stroke will help to identify effective agents for suppressing microglial activation and prevent the series of microglial activation-mediated CNS injuries, including neuroinflammation.

Microglial Polarization

Polarization is a process through which macrophages adopt different phenotypes depending on the stimulus, period, and environment (44, 45). Like macrophages, microglia also undergo polarization. As mentioned previously, activated microglia can be defined as classic (proinflammatory; M1) or alternative (anti-inflammatory or protective; M2) under pathophysiological conditions (46).

The main basis for distinguishing between M1 and M2 microglia is their biological function as well as the secreted cytokines and chemokines. M1 microglia usually have antigen-presenting and killing effects and secrete various inflammatory factors to evoke a strong inflammatory response. Inducible nitric oxide synthase (iNOS) is a commonly used marker for M1 microglia. By contrast, M2 microglia enhance phagocytic activity to remove debris and produce many anti-inflammatory and repair factors. The enzyme arginase 1 is one of the best characterized markers of the M2 microglia (47). In vitro studies have successfully used different inducers to regulate and study the polarization of microglia; stimulation with lipopolysaccharide and interferon-γ (IFNγ) promotes the differentiation of M1 microglia, whereas interleukin (IL)-4 and IL-10 induce the M2 phenotype (25, 48–50). The interferon regulatory factor (IRF) family has recently been found to have an important relationship with the polarization of microglia after stroke (51–53). For example, IRF4 negatively regulates inflammation and promote M2 polarization of macrophage (54), whereas IRF5 induces M1 polarization (55). Other IRFs, such as IRF3, IRF7, and IRF8, have also been shown to participate in the process of microglial polarization. This finding opens a new perspective for stroke treatment (53).

The association between M1 and M2 differentiation and disease progression varies among different diseases. Hu et al. (25), who used a transient focal ischemia model to reveal the dynamic changes of microglial polarization, reported a differential shift from the M2 phenotype to the M1 phenotype in the ischemic brain. Specifically, soon after ischemic injury, a majority of the microglia migrated into or infiltrated infarcted areas exhibiting the M2 phenotype, which represented an endogenous effort to clear the ischemic tissue and restrict brain damage. However, the number of M2 microglia gradually decreased within 7 days, and M1 microglia consequently began to dominate the damaged area. This finding contributes to the ability to effectively determine possible intervention methods and the optimal intervention time after ischemic stroke.

In addition to classical typing, various subtypes have emerged with the deepening of microglia research. For example, M2 microglia can be further divided into three subtypes, including M2a, M2b, and M2c, based on different stimulation processes and functions (56–58). Further investigation is required to explore the role of these cell populations in ischemic stroke or other CNS diseases.

Microglial polarization is theoretically greatly important in ischemic stroke; however, investigators have suggested that the M1/M2 framework is limited (23, 59). Ransohoff et al. (23) elaborated this perspective on microglial polarization, which is based on the current understanding of microglial polarization being influenced by macrophage polarization. This schema was adopted to simplify data interpretation at a time when the ontogeny and function of microglia had not been characterized (23). With the development of new technologies, more in-depth research on microglia is expected. In this context, the significance and concept of microglial polarization also needs to be developed simultaneously.

Function of Microglia in Ischemic Stroke

Microglia activate rapidly after ischemic stroke, as described previously. However, their role is a double-edged sword because, during different periods of stroke, they have different and sometimes opposite functions. In the acute phase, activated microglia secrete a range of inflammatory cytokines, including tumor necrosis factor (TNF), IL-1β, and IL-6 (60), which contribute to a robust inflammatory response. Investigators have reported that, after MCAO, infiltrating macrophages are also a source of inflammatory factors (61). The role of macrophages and microglia in stroke is similar in many aspects and, therefore, is often described as microglia/macrophage. However, the types of major inflammatory factors produced by microglia and macrophages are different: the former produce relatively higher levels of reactive oxygen species (ROS) and TNF-α, whereas the latter produce higher levels of IL-1β (61). The inflammatory response caused by the inflammatory factors secreted by these cells is a hallmark of the acute stage of ischemic stroke. It is generally believed that, in the acute phase, a severe inflammatory response like this is negatively correlated with prognosis (62, 63). Therefore, administering anti-inflammatory treatments in the acute stage is important.

After the acute phase of ischemic stroke, the inflammatory response gradually decreases, and ischemic stroke enters another stage during which tissue repair dominates the infarction region (64). At this stage, microglia function as a “repairer.” Tissue repair after stroke relies on neurogenesis to replace injured cells. It has been suggested that microglia actively modulate neurogenesis via guiding cell migration and influencing synaptic activity by regulating the number of functional synapses in the CNS (65). It is worth noting that microglia may have a controversial effect on neurogenesis at 2 and 16 weeks after stroke. Therefore, studies with a longer observation time and multiple time points are needed to clarify when and how microglia influence neurogenesis (14).

Microglia are the major phagocytic cells in the CNS. Their phagocytosis, mediated by TREM2, CD36, and other molecules, is critical in removing degenerating neuronal cells and debris after ischemia (66, 67). The morphological changes in microglia within the lesion core were greater than those of microglia in the periphery, which indicated a transformation from reactive microglia to dystrophic microglia. Furthermore, reduced microglial cell numbers were in the core of the infarction, and the number of neutrophils was higher in the core than in the periphery of the lesion. These results suggest that microglia can engulf and remove neutrophils after brain ischemia, and that this phagocytic function has a distinct relationship, based on the internal location of lesions (68). The phagocytic function of microglia may also be associated with the destruction of the BBB, which is an important pathological mechanism after ischemic stroke. Perivascular microglia can phagocytize vascular endothelial cells, thereby destroying vascular integrity after stroke (69).

Reactive Astrocytes

In the acute phase of stroke, injured cells in the lesion and penumbra release cytokines, including transforming growth factor (TGF)-α, ciliary neurotrophic factor, IL-1, IL-6, and Kallikrein-related peptidase 6 (100). In response, reactive astrogliosis occurs in the peri-infarct region, and a glial scar is formed to maintain CNS homeostasis and wall off the lesion (101). The hallmarks of reactive astrogliosis are astrocytic hypertrophy and overexpression of glial fibrillary acidic protein (GFAP) (102). After undergoing reactive astrogliosis, astrocytes produce and release proinflammatory mediators, such as IL-6, TNF-α, IL-1α, IL-1β, and IFNγ, and free radicals, such as NO, superoxide, and peroxynitrite (100). By contrast, astrocyte proliferation and glial scar formation restrict the diffusion of neuroinflammation (103). Through Affymetrix GeneChip arrays, Zamanian et al. studied the expression of various genes upregulated in reactive astrocytes (104). Lcn2, which may directly promote neuronal death, was induced 228-fold, and Serpina3n was induced 9.1-fold in reactive astrocytes 1 day after experimental ischemic stroke (104). Moreover, Liddelow et al. (82) termed the two subtypes of reactive astrocytes as “A1” and “A2.” They found that A1 reactive astrocytes were induced by IL-1α, TNFα, and C1q secreted by activated microglia. These astrocytes have few physiological functions but contribute to the death of neurons and oligodendrocytes. One of the most upregulated genes is C3. By contrast, A2 reactive astrocytes, which upregulate neurotrophic factors, were postulated as a neuroprotective subtype. Rakers et al. (105) recently used next-generation sequencing to investigate transcriptome change after experimental focal ischemia. The markers of reactive astrocytes, Lcn2, Gfap, Vimentin, and Timp1, were highly expressed, as expected. The genes that contribute to inflammation (e.g., Spp1, Cd52, Lcn2, and Ifi202b), cell division, and migration (e.g., Cdk1, Myo1f, and Anxa3) were upregulated. A2-specific transcripts were intriguingly predominant at 72 h after tMACO (105). An in vitro study (106) showed that IL-1β can induce reactive astrogliosis, the upregulation of inflammatory mediators, such as IL-6 and CXCL5, and the elevation of neurotrophic factor levels, such as brain-derived neurotrophic factor and nerve growth factor. These transcriptome analyses unraveled that reactive astrocytes after ischemic stroke exerted both proinflammatory and neuroprotective functions. In addition, microglia-derived cytokines are critical for determining astrocyte phenotype.

Function in the Innate Immune Response

Astrocytes participate in innate and adaptive immune responses after ischemic stroke. Once ischemic stroke occurs, cytokines, DAMPs, and ROS are generated and released by injured cells to stimulate the receptors of astrocytes and alter their phenotype, thereby inducing “reactive astrogliosis.” Astrocytes synthesize cytokines and chemokines and interact with other cells via the activation of their receptors and alteration of their intracellular signal pathways (Table 1).

The P2Y1 receptors of astrocytes that are stimulated by adenosine 5′-triphosphate (ATP), which is released by injured cells in post-ischemic stroke, promote the production of proinflammatory cytokines and chemokines through the phosphorylated-p65 subunit (RelA)-mediated nuclear factor-kappa B (NF-κB) pathway (107). Furthermore, treatment with P2Y1 receptor antagonists protects astrocytes from ischemic injury (124). Zheng et al. controversially reported that P2Y1R stimulation increased ATP production and decreased the infarction volume via the IP3R-dependent pathway (108, 109).

Toll-like receptors activated by HMGB1, PRX proteins, and other DAMPs induce the release of proinflammatory molecules (10). Among the TLR isotypes, TLR2, and TLR4 are crucial inflammatory mediators after stroke. In neurons and astrocytes, HMGB1 induces MMP-9 upregulation through activating TLR4 (125). Moreover, astrocytes treated with the TLR4 activator lipopolysaccharide upregulate the expression of SOCS-1, CXCL10, TNF-α, VCAM-1, IL-15, and IL-27 through the MyD88-independent Jak1/Stat1 pathway and the MyD88-dependent NF-κB pathway (110). Suppression of TLR2 and TLR4 by pharmacological or transgenic approaches reduces NF-κB activity, lowering the level of proinflammatory cytokines iNOS and COX2; however, the exact mechanisms remain unclear (111–113). Growth associated protein-43 inhibits the TLR4/ NF-κB pathway to decrease the production of IL-6 and TNF-α (126). TLR4 paradoxically promotes neuroblast migration and neurogenesis after stroke (114). In addition, ischemic preconditioning stimuli ameliorate the inflammatory response after stroke via the TLR/cytokine pathway (127).

The class B scavenger receptor CD36, which is primarily found in microglia and macrophages, is co-located with GFAP in the peri-infarct astrocytes at 3–7 days after stroke (128). The ligands of CD36 produced during ischemia, such as thrombospondins, oxidized lipids, and apoptotic bodies, may upregulate its expression (115, 128). In addition, CD36-mediated neuroinflammation promotes the expression of intermediate filaments, free radical production, and scar formation (115, 128). The knockdown of CD36 reduces the level of proinflammatory factors, such as IL-1, IL-6, and MCP-1, and reduces infarct volume (115). However, the pharmacological inhibition of CD36 in hyperlipidemic stroke worsens the outcome (129). An unexpected finding is that preconditioning with a CD36 inhibitor can reduce the effects of stroke (129).

After ischemic stroke, TGFβ signaling is increased in astrocytes and activated microglia/macrophages after ischemic stroke (130). Studies have shown that TGFβ exerts antiapoptotic and neuroprotective effects after stroke (131, 132). Furthermore, the inhibition of the TGFβ pathway promotes the infiltration of immune cells in the peri-infarct lesion and does not change the physical barrier formed by astrocytes (99). To date, the exact mechanism of its neuroprotective effect is not fully elucidated. Previous studies suggest that the Smads, PI3K/Akt, ERK/MAPK, or PKA pathway could be potentially involved (99, 133–136).

The Janus kinase/signal transducer and activator of transcription 3 (JAK/STAT3) signaling pathway is one of the primary regulators affecting the functional and molecular changes in reactive astrogliosis. Reactive oxygen species and other molecules produced after ischemic stroke activate STAT3 (137). The regulation of the STAT3 pathway through miRNA-31 or sinomenine attenuates neuroinflammatory and neuronal injury (116, 138). The increased p-STAT3 also promotes neuronal progenitor cell proliferation and functional rehabilitation via the STAT3-HIF1α-VEGF axis (117, 139). Investigators have postulated that IL-6 induces neuroprotection via STAT3 pathway activation (118, 140). The role of STAT3 signaling in ischemic stroke requires further exploration.

The Notch-1 pathway, which participates in reactive astrogliosis and neural stem cell differentiation, is activated after stroke (119). Activation of this pathway facilitates the proliferation of reactive astrocytes and restricts infiltration of immune cells (119).

NF-κB is another well-understood pathway that regulates proinflammatory cytokine expression. By using transgenic mice, Dvoriantchikova et al. elucidated how NF-κB is involved in proinflammatory and redox-active pathways and has a pivotal role in neurotoxicity (120). The inhibition of the NF-κB pathways may reduce the expression of proinflammatory genes, including Tnf-α, Icam1, Ccl2 (Mcp1), Cxcl10 (IP10), and Vcam1, and reduce the infiltration of CD11b+ leukocytes (120). Tumor protein 53-induced glycolysis reduces the degradation of IκBα and inhibits NF-κB translocation in astrocytes, thereby ameliorating neuroinflammation (141).

Tumor necrosis factor-like weak inducer of apoptosis (TWEAK) and its receptor fibroblast growth factor-inducible 14 (Fn14), expressed in neurons (142), may also be a proinflammatory factor. TWEAK/Fn14 can activate NF-κB pathways and increase the expression of the chemokine monocyte chemoattractant protein-1, which leads to leukocyte infiltration into the infraction lesion (121). TWEAK stimulation also contributes to reactive astrogliosis through the TGF-α/EGFR signal pathway (122). TWEAK/Fn14 inhibition therefore protects the integrity of the BBB and reduces infarct volume (143, 144). However, the TWEAK/Fn14 pathway activation in hypoxia preconditioning results in hypoxic and ischemic tolerance (145).

Hypoxia-induced glial T-cell immunoglobulin and mucin domain protein (TIM)-3 is highly expressed in astrocytes and microglia. Targeting TIM-3 for treatment decreases the infarct volume and improves neurological deficits (123).

The cytoplasmic protein, NLR family pyrin domain containing 2, which is expressed in astrocytes, forms an inflammasome with apoptosis-associated speck-like protein containing the caspase recruitment domain and caspase-17 (146, 147). Its expression is significantly increased after stroke (148). In vitro, apoptosis signal-regulating kinase 1 inhibition contributes to the reduction of proinflammatory cytokines (149).

In addition to scar formation and cytokine production, astrocytes also have a beneficial role in increasing extracellular glutamate uptake and sodium/potassium-ATP activity (150), releasing neurotrophic factors (151), and rebuilding the BBB (152). In addition, Morizawa et al. (153) reported that some reactive astrocytes transform into phagocytic cells in the penumbra after ischemic stroke. The phagocytic markers galectin-3 and LAMP-2 increased and peaked at 7 days after stroke. The ATP-binding cassette transporter ABCA1 and its pathway molecules MEGF10 and GULP1 have a critical role in the phagocytic transformation, contributing to tissue recovery (153).

Interaction With the Complement System

The complement system, which is a major effector of the innate immune system, prevents the invasion of pathogens and regulates brain function in physiological and pathological conditions. Astrocytes express complement receptors for C1q, C3a, C5a, and CR3 (8).

C1q participates in the classical pathway of complement activation as a pattern recognition molecule. C1q secreted by microglia promotes A1 phenotype transformation. This process is potentially mediated by Megf10, a scavenger receptor that is predominantly expressed by astrocytes (154, 155). C1q deficiency is neuroprotective in neonatal mice suffering from hypoxic-ischemic injury (156). A C1q inhibitor reduced leukocyte infiltration and exhibited a neuroprotective effect in an adult mice ischemic injury model; unfortunately, this effect is C1q-independent (157).

C3, a component of the complement cascade and precursor of C3a, was recently considered a marker of A1 astrocytes because of its high expression (82). Mice deficient in C3 manifested with a reduced infarction volume and decreased granulocyte infiltration (158). C3a, a derivate of C3, mediates cerebral endothelial cell activation and leukocyte recruitment (159). In previous studies, the pharmacological inhibition of the C3a-receptor promoted anti-inflammatory neuroprotection, ameliorated neurological deficits, and improved the outcome after stroke (158, 160). Oxygen-glucose deprivation unexpectedly increased the expression of the C3a-receptors of astrocytes in vitro, and C3a administration protected astrocytes from cell death in response to ischemic stress by reducing ERK signaling and caspase-3 activation (161). This protective effect was reversed by C3a-receptor deficiency (161). Analogously, C3a expressed in astrocytes exerted a neuroprotective effect in neonatal hypoxic-ischemic brain injury (162). C3a overexpression or intranasal administration stimulates neural plasticity, increases the number of the new neurons in the peri-infarct area, and promotes neuronal survival (126, 163).

C5 apparently contributes little in the response to ischemic stroke because the infarction volume of C5- deficient mice was no different from that in wildtype mice (158). However, a neuroprotective effect has been observed in C5aR1-deficient mice or in mice treated with a C5aR1 antagonist after ischemic stroke (164, 165).

Neuroinflammatory Influence

After ischemic stroke, several mechanisms mediate axonal degeneration, including energy and metabolite deficiency, calcium-regulated cell apoptosis and degeneration, and myelin-associated inhibitors of regeneration (196). Oligodendrocytes undergo a complement attack and undergo apoptosis and necroptosis induced by the toxicity of the released glutamate and ATP (9, 179). Glutamate also activates the AMPK/kainate receptors in the neighboring microglia, thereby promoting the release of proinflammatory cytokines, such as IL-1β and IFN-γ (179). Investigators have shown that inflammatory cytokines have a critical role in demyelination diseases. Tumor necrosis factor-α induces oligodendrocyte apoptosis, delays myelination, and inhibits OPC proliferation and differentiation (180). In addition, IFN-γ induces oligodendrocyte apoptosis and reduces OPC proliferation (181). However, IL-6 combined with IL-6R enhances the differentiation and survival of oligodendrocytes (182). IL-11 promotes oligodendrocyte survival via IL-11R. IL-17A also potentiates OPC differentiation through the ERK1/2 signaling pathway (183). Treatment with IL-4 promotes the regeneration and remyelination of oligodendrocytes via the IL-4/PPARγ signaling axis (197). IL-1β intriguingly promotes the survival of oligodendrocytes and the death of oligodendrocytes and OPCs via the ciliary neurotrophic factor and glutamate excitotoxicity, respectively (184, 185). Dying oligodendrocytes release HMGB1, which combines with TLR2 and exerts autocrine trophic effects (198).

The number of OPCs increases in the penumbra after ischemic stroke but decreases in the core lesion (199). In addition, they became hypertrophic (200), with altered potassium channel permeability (201), and undergo death induced by excitotoxicity (202, 203). The overexpression of the Nertin-1 or CXCL12 gene leads to OPC migration, proliferation, and remyelination (204, 205). Furthermore, NeruoD1 expression in OPC alters their phenotype toward glutamatergic and GABAergic neurons (206).