This article is mentioned in:

Introduction

According to a recent report from the World Health Organization, ischemic stroke is the second leading cause of mortality worldwide (1). In 2016 alone, ~5.5 million people succumbed to an ischemic stroke and an estimated 116.4 million people suffered from permanent disabilities following a stroke event worldwide (2). Undoubtedly, these mortalities and disabilities have a huge effect, both emotionally and financially, on families and society.

Ischemic stroke is primarily caused by the blockage of cerebrovascular blood flow. Previous studies have reported that the incidence of early mortality following ischemic stroke (between 21 days to 1 month) in both high-income and low-middle-income countries increased gradually from 2000 to 2008, from 13 to 23% and 13 to 19%, respectively (3,4). Currently, intravenous thrombolysis and mechanical thrombectomy are the only effective means of treating ischemic stroke (5,6); however, these treatment options are limited by the short-recommended treatment window (6). In addition, a series of reperfusion injuries may occur following ischemic reperfusion (6,7); in particular, inflammation and oxidative stress are the main causes of reperfusion injuries, and the latter can induce the former (6–9). An increasing amount of evidence has demonstrated that continuous neuroinflammation can damage neurons and the blood-brain barrier (BBB) during the process of cerebral infarction, and this in turn results in tissue destruction and worsens the functional outcome (10–12). In addition, it has been demonstrated that the inhibition of neuroinflammation decreased the neurological deficit in experimental stroke models (12,13). Therefore, it is evident that neuroinflammation serves a crucial role in ischemic stroke-induced brain injury.

Microglia are important cells that initiate the inflammatory response in the brain following various brain injuries (14), including ischemic stroke (15). Accumulating evidence has indicated that activated microglia are the primary source of immunomodulatory molecules, such as cytokines, chemokines and free radicals, in the brain (16–18); these molecules are closely associated with secondary brain injury and brain tissue repair in ischemic stroke (16–18). In addition, some molecules, including soluble Fas ligand (19) and endothelial nitric oxide synthase (20,21), which are generated by other cells, such as neurons and endothelial cells, have also been identified to affect the stroke outcome through interacting with microglia in the brain following ischemic stroke (19). Therefore, investigating microglial changes and their function is crucial to understanding ischemic stroke pathophysiology.

To develop new and effective therapies for ischemic stroke, it is also necessary to investigate the cellular and molecular mechanisms underlying brain damage and tissue repair in ischemic stroke. Despite the fact that microglia-induced inflammation can aggravate brain damage upon ischemic stroke, the modulation of the microglial phenotype has also been observed to promote brain repair; for example, through decreasing brain edema, improving white matter integrity and recovering motor function (13). Therefore, investigating the mechanisms underlying microglial phenotypic transformation is crucial for the development of novel treatments for patients with ischemic stroke.

The present review aimed to describe the important advances in the current understanding of the changes, functions and modulators of microglia during ischemic stroke. The roles of autophagy in the modulation of microglia polarization are discussed, in addition to the future research directions of ischemic stroke, which may help identify novel treatment strategies for ischemic stroke.

Microglia

Microglia are small, macrophage-like glial cells that account for 10–15% of cells in the central nervous system (CNS). In response to various stresses, microglia are rapidly activated to differentiate into either the M1 or the M2 phenotype, which are involved in tissue damage and repair, respectively (22–24).

M1 microglia and changes

M1 microglia exist in a proinflammatory state and secrete proinflammatory cytokines, which have been identified to promote brain damage (13,19). Following ischemic stroke, M1 microglia are activated and subsequently serve a detrimental role. Briefly, upon experiencing ischemia/hypoxia, NF-κB is activated within microglia and translocates from the cytoplasm to the nucleus; this activates the release of proinflammatory cytokines, which causes secondary brain damage (16–18), such as interleukin (IL)-1β (24), IL-6 (25) and tumor necrosis factor-α (TNF-α) (26), in addition to the production of inducible nitric oxide synthase (iNOS) (27,28). For example, in middle cerebral artery occlusion (MCAO) model mice, TNF-α secreted by M1 microglia was identified to increase endothelial necroptosis and BBB leakage following ischemic stroke (21,29). This further promotes neuroinflammation and brain edema, resulting in a poor outcome (29). Moraes et al (30) demonstrated that IL-1β, which is generated by M1 microglia, decreased the number of synapses in the hippocampus and cortex, which aggravated cognitive impairment. In oxygen-glucose-deprived (OGD) microglia, the increase in IL-17A levels aggravated neuronal death (31). In addition, C-C motif chemokine ligand 2 (CCL2) generated by microglia was observed to recruit CD8+ γδ T cells to the ischemic brain, which was identified to be the primary factor aggravating brain infarction (32,33). In fact, besides ischemic stroke, M1 microglia activation occurs in multiple other neurological diseases; for example, in an RD1 mouse model of retinal degeneration, flow cytometry analysis revealed that the percentage of T-lymphocyte activation antigen CD86 (CD86)+ M1 microglia was markedly increased in the rapid degeneration phase (34); paraquat stimulation was revealed to activate M1 microglia, which subsequently activated the toll-like receptor-4 (TLR4)-mediated NF-κB signaling pathway and resulted in the release of proinflammatory cytokines (35); and the number of low affinity immunoglobulin gamma Fc region receptor III-A/b (CD16/32)+ M1 microglia was significantly increased in the brains of subarachnoid hemorrhage model mice through the activation of the myeloid differentiation primary response protein MyD88 (MyD88)/mitogen-activated protein kinase pathway, whereas decreasing the percentage of M1 microglia significantly improved the neurological deficits (36). Other results from the literature have indicated that M1 microglia activation can also aggravate brain damage in other types of neurological disease (34–36).

The activation of M1 microglia can be determined by detecting the expression of their surface markers (Table I); these classical markers include integrin alpha-M (CD11b) (37), CD16 (38), CD32 (39) and CD86 (40). However, these markers cannot distinguish between microglia and macrophages in the brain, as they can be expressed by both M1 microglia and macrophages. For example, Liu et al (41) reported that both CD16+/ionized calcium-binding adapter molecule 1 (Iba1)+ M1 microglia/macrophages were detected using double immunofluorescence staining. To distinguish resident microglia from blood-derived macrophages in the brain, Satoh et al (42) suggested that TMEM119 may be used as a marker of resting microglia in the human brain; however, the challenge remains to discriminate activated microglia from infiltrated macrophages in the brain.

Table I. Markers of M1 and M2 microglia.

Table I.

Markers of M1 and M2 microglia.

Phenotype	Marker	Type	Effect	(Refs.)	Author, year
M1	IL-1β	Cytokine	Proinflammatory	(25)	Zhu et al, 2019
	IL-6	Cytokine	Proinflammatory	(59)	Shu et al, 2016
	TNF-α	Cytokine	Proinflammatory	(25)	Zhu et al, 2019
	iNOS	Metabolic enzyme	Oxidative damage	(59)	Shu et al, 2016
	ROS	Metabolite	Oxidative damage	(7)	Xu et al, 2018
	CD11b	receptor	Induce proinflammatory signaling	(37,81)	Liu et al, 2018; Hyakkoku et al, 2010
	CD16	Receptor	Induce proinflammatory signaling	(38)	Jiang et al, 2018
	CD32	Receptor	Induce proinflammatory signaling	(37)	Liu et al, 2018
	CD68	Receptor	Induce proinflammatory signaling	(27)	Boddaert et al, 2018
	CD86	Receptor	Induce proinflammatory signaling	(44)	Hu et al, 2012
	NF-кB	Transcription	Induce M1 phenotype, elevate the expression of proinflammatory cytokines	(16)	Yang et al, 2019
	STAT1	Transcription	Elevate the expression of NF-κB	(70)	Butturini et al, 2019
	STAT3	Transcription	Elevate the expression of NF-κB	(72)	Ding et al, 2019
	Hv1	Ion channel protein	Elevate the expression of ROS and NF-κB, induce M1 phenotype	(94,96)	Tian et al, 2016; Yu et al, 2018
	Kv1.3	Ion channel protein	Proinflammatory, induce M1 phenotype	(83,97)	Di Lucente et al, 2018; Nguyen et al, 2017
	H19	Gene	Proinflammatory, induce M1 phenotype	(98)	Wang et al, 2017
	miRNA-155	Gene	Proinflammatory, induce M1 phenotype	(18)	Zheng et al, 2018
M2	IL-4	Cytokine	Anti-inflammatory, induce M2 phenotype	(23)	Liu et al, 2016
	IL-10	Cytokine	Anti-inflammatory	(25)	Zhu et al, 2019
	TGF-β	Cytokine	Anti-inflammatory, regeneration	(44)	Hu et al, 2012
	Ym1/2	Protein	Anti-inflammatory, regeneration	(25)	Zhu et al, 2019
	Arg-1	Protein	Anti-inflammatory, regeneration	(44)	Hu et al, 2012
	CCL22	Chemokine	Recruits regulatory T cells	(44)	Hu et al, 2012
	CD206	Receptor	Induce anti-inflammatory signaling	(39)	Jin et al, 2014
	STAT6	Transcription	Inhibit the expression of NF-κB	(74)	Yang et al, 2017
	Nrf2	Transcription	Anti-inflammatory, antioxidant stress	(7)	Xu et al, 2018
	PPARγ	Transcription	Inhibit the expression of NF-κB, promote the expression of Nrf2	(26,78)	Liu et al, 2018; Cai et al, 2017
	miRNA-124	Gene	Anti-inflammatory, induce M2 phenotype	(99)	Hamzei et al, 2016
	FAM19A3	Gene	Anti-inflammatory, induce M2 phenotype	(100)	Shao et al, 2015

[i] IL, interleukin; TNF-α, tumor necrosis factor-α; iNOS, inducible nitric oxide synthase; miRNA, microRNA; ROS, reactive oxygen species; CD11b, integrin alpha-M; CD16, low affinity immunoglobulin gamma Fc region receptor III-A; CD32, low affinity immunoglobulin gamma Fc region receptor III-b; CD68, cluster of differentiation 68; CD86, T-lymphocyte activation antigen CD86; CD206, macrophage mannose receptor 1; STAT, signal transducer and activator of transcription; H19, H19 imprinted maternally expressed transcript; TGF-β, transforming growth factor-β; Ym1/2, chitinase-3-like protein 3; Arg-1, arginase-1; CCL22, C-C motif chemokine ligand 22; Nrf2, nuclear factor erythroid 2-related factor 2; PPARγ, Peroxisome proliferator-activated receptor γ; FAM19A3, chemokine-like protein TAFA-3.

Changes in the activation of M1 microglia can affect stroke prognosis and changes in the expression levels of M1 microglia biomarkers, including surface markers and proinflammatory cytokines, are also common in ischemic stroke (13,25,43). Based on these changes, it is possible to draw some useful conclusions regarding ischemic stroke. For example, in MCAO mice, the expression levels of certain proinflammatory cytokines, such as IL-1β, TNF-α and IL-6, in addition to iNOS, were observed to increase from 12 h, peak at day 14, then subsequently decrease and remain constant in the ischemic boundary zone until day 28 following MCAO (25,43). In addition, the study (25) also observed that CD32 expression levels increased in a similar manner. Another previous study drew similar conclusions, witnessing a significant increase in CD16 expression levels from day 3, which remained elevated until day 14 in the cortex and striatum at the inner boundary of the infarct (44). These conclusions were further confirmed in both a bilateral common carotid artery ligation (BCCA) model and a photothrombotic (PT) stroke model. In the hippocampal CA1 region of the BCCA ligation mice, the M1 microglia marker CD74 was observed to be expressed from day 3 post-ischemic insult and elevated from day 5 post-ischemia (45). Additionally, CD16 and iNOS expression levels were identified to be increased in infarct regions at day 42 following BCCA (46). In PT stroke model mice, CD16 and CD32 expression levels were detected at day 1 and remained elevated until day 14 following PT cortical ischemia (47). In addition, numerous inflammatory cytokines, such as TNF-α, IL-1β and IL-6, remained highly expressed at days 14 and 21 in the ischemic region following a PT stroke (48,49). Other associated studies reported that age was also associated with microglia polarization, and the trend of the expression levels of M1 microglial markers in ischemic stroke was consistent with the former studies (25,44) in both young adult and older mice (50,51).

Based on these changes, it is suggested that the proinflammatory state may be dominant in the acute stage of ischemic stroke; however, the changes in the expression levels of proinflammatory cytokines during the chronic phase of ischemic stroke, and the interactions between them and brain tissue repair, remain unclear.

M2 microglia and changes

M2 microglia exist in an anti-inflammatory state and secrete anti-inflammatory cytokines and neurotrophic factors to promote brain repair (23,25,52). M2 microglia are activated and reported to serve a beneficial role following ischemic stroke; for example, in conditions of ischemia/hypoxia, peroxisome proliferator-activated receptor γ (PPARγ), a transcription factor with anti-inflammatory properties, is reportedly activated and mobilized from the nucleus to the cytoplasm in microglia (26). This leads to the activation of M2 microglia, which release anti-inflammatory cytokines that improve stroke outcome. For example, Liu et al (23) demonstrated that IL-4 secreted by M2 microglia decreased the infarct size following ischemic stroke and improved the long-term functional recovery. Zhu et al (25) observed that M2 microglia-induced chitinase-3-like protein 3 (Ym1/2), IL-10 and transforming growth factor-β (TGF-β) secretion promoted angiogenesis, and thereby decreased the BBB leakage and improved stroke outcome. In addition, Choi et al (53) confirmed that M2 microglia promoted the proliferation and neuronal differentiation of neural stem/progenitor cells in the ipsilateral subventricular zone following ischemic stroke through upregulating the expression levels of TGF-α; this may provide an effective therapy for neurogenesis. Additionally, macrophages recruited by microglia were identified to enhance M2 microglia polarization and improve stroke outcome (54,55). Similar to M1 microglia, besides from ischemic stroke, M2 microglia activation occurs in numerous other types of neurological disease; for example, in both spinal cord injury and intracerebral hemorrhage mouse models, M2 microglia were identified to be activated, which released anti-inflammatory cytokines downstream of cAMP response element-binding protein (CREB)-associated signaling pathways (56,57). In neonatal germinal matrix hemorrhage rats, Rh-Chemerin promoted the accumulation and proliferation of M2 microglia in periventricular regions and suppressed the inflammatory response through nuclear factor erythroid 2-related factor 2 (Nrf2)-associated signaling pathways (58). Previous studies demonstrated that M2 microglia activation can also promote brain tissue repair in other neurological diseases (56–58).

The activation of M2 microglia can be assessed by determining the expression levels of their surface markers (Table I), for example, macrophage mannose receptor 1 (CD206) (59), and the secretion of anti-inflammatory cytokines, including IL-4 (23), IL-10 (59), arginase-1 (Arg-1) (60), Ym1 (28) and TGF-β (59). Changes in the percentage of activated M2 microglia can affect stroke prognosis and based on these changes, it is possible to draw some useful conclusions regarding ischemic stroke; for example, in MCAO model mice (44), the mRNA expression levels of cytokines generated by M2 microglia, including Arg-1, Ym1/2, CCL22, IL-10 and TGF-β, were detected from days 1 to 3 following MCAO and peaked at days 3 to 5 post-injury in ischemic regions (44). The expression levels of the majority of M2 microglia genes was observed to begin to decrease at days 5–7 after MCAO and by day 14, were restored to pre-injury levels in ischemic regions (44). In addition, the number of CD206+ cells followed a similar trend (44); CD206 remained highly expressed until day 21 in the cortex and striatum following ischemia (61). Previous studies have further validated these results in a PT stroke model (47–49); in the ischemic cortex, CD206 expression levels rapidly increased at day 1, peaked at day 7 (47) and remained highly expressed until day 21 following a PT stroke (49). In addition, certain anti-inflammatory cytokines were also observed to be highly expressed until days 14 and 21 following a PT stroke; for example, IL-4 and IL-10 remained highly expressed until day 14 (47) and TGF-β remained highly expressed until day 21 in ischemic regions (49). Furthermore, another previous study (44) revealed that the M2 phenotype was the dominant phenotype for the first 7 days following ischemic stroke and could promote neuronal survival following OGD.

Based on these data regarding the changes in the expression levels of M1 and M2 markers, several conclusions were drawn (Fig. 1); it was hypothesized that microglia are activated following ischemic stroke and the M1/M2-type pattern of expression changes dynamically post-injury, exhibiting an early beneficial M2 phenotype, followed by the transition to a detrimental M1 phenotype at the later stages. These results suggested that regulating microglia transformation from the M1 to M2 phenotype may be a crucial method of treating ischemic stroke.

Figure 1. Dynamic changes in microglia marker levels over time following IS. Top panel: M1 microglia markers demonstrated an increasing trend during the first 14 days following IS, after which they decreased. Expression levels of M2 microglia markers increased from day 1, peaked at days 5–7 and decreased until day 42. Bottom panel: Microglia exhibited an M2-like response as early as 1 day following IS, which manifested as increased expression levels of CD206 Arg-1. The expression levels of TGF-β increased from day 3 until day 21. With regards to the M1 markers, the expression levels of the proinflammatory cytokines, TNF-α, IL-6 and IL-1β increased from day 3. The levels of M1 surface markers CD16 and CD32 increased from day 1 and CD16 and iNOS expression levels remained increased until day 42 after IS. IS, ischemic stroke; Arg-1, arginase 1; TGF-β, transforming growth factor-β; TNF-α, tumor necrosis factor α; IL, interleukin; iNOS, inducible nitric oxide synthase; CCL, C-C motif chemokine ligand; Ym1/2, chitinase-3-like protein 3.

Modulatory mechanisms of microglia polarization

It has been suggested that ischemic stroke may be effectively treated through modulating microglia polarization; in general, these modulators include four categories: Transcription factors; receptors; ion channels; and gene modulators.

Transcription factors

Transcription factors are proteins that bind to DNA and regulate the transcriptional activity of genes. NF-κB, members of the STAT family, nuclear factor erythroid 2-related factor 2 (Nrf2) and PPARγ are 4 important transcription factors that have been associated with the mechanisms underlying microglia polarization (16,19,26,62,63) (Fig. 2). The first 2 transcription factors are closely associated with M1 polarization (16,19,62), whereas the other 2 are associated with M2 polarization (26,63).

Figure 2. Modulatory mechanisms of microglia polarization following ischemic stroke. Microglia are activated following ischemic stroke. M1 microglia activation involves different factors, including the cytokines S1P, LPS, HMGB1, S100B and CKLF1. These factors bind to membrane receptors, such as the cytokine receptors S1PR, RAGE, TLR4, IL-4R and CCR4, to trigger proinflammatory cellular signaling pathways. These pathways include the JAK2/STAT1/NF-κB, JAK2/STAT3/NF-κB and MyD88/NF-κB pathways. The downstream target NF-κB is released and translocates from the cytoplasm to the nucleus following the phosphorylation of IκB, where it initiates the transcription of proinflammatory genes. M2 microglia can be activated by IL-4, which promotes PPARγ mobilization from the nucleus to the cytoplasm. Subsequently, PPARγ inhibits the activation of NF-κB and promotes the activation of Nrf2 to induce the transcription of anti-inflammatory genes. Additionally, the activated Nrf2 promotes the expression of HO-1, which scavenges reactive oxygen species and nitric oxide. S1P, sphingosine 1 phosphate; LPS, lipopolysaccharide; IL, interleukin; HMGB1, high mobility group protein B1; S100B, protein S100-B; CKLF1, chemokine-like factor; SIPR, sphingosine 1 phosphate receptor; TRL4, toll-like receptor 4; CCR4, C-C chemokine receptor type 4; RAGE; JAK2, Janus activated kinase 2; MyD88, myeloid differentiation primary response protein MyD88; IκB, NF-κB inhibitor; PPARγ, peroxisome proliferator-activated receptor γ; Nrf2, nuclear factor erythroid 2-related factor 2; HO-1, heme oxygenase 1; Keap 1, Kelch-like ECH-associated protein 1; STAT, signal transducer and activator of transcription.

NF-κB

The transcription factor NF-κB has been identified to activate microglia, and to transform these activated cells into the M1 phenotype (16,62). In fact, it has been confirmed that NF-κB can be activated in response to detrimental stresses and induce inflammatory responses in the brain (16,62). Upon ischemia or hypoxia, the expression levels of some damage-associated molecules, including protein S100-B (11,64), chemokine-like factor 1 (65–67) and CXC motif ligand 12 (43) are reportedly increased; these molecules bind to membrane receptors, such as receptor for advanced glycation end products, C-C chemokine receptor 4 and CXC-chemokine receptor 4 to activate NF-κB (43,64–66). The activated NF-κB is subsequently translocated to the nucleus, where it triggers inflammatory responses (16). NF-κB inhibitor α (IκBα), a protein bound to NF-κB, can inhibit the activation of NF-κB; however, upon phosphorylation, it can release NF-κB, which translocates to the nucleus from the cytoplasm (16,68). The activated NF-κB can promote the secretion of proinflammatory cytokines from M1 microglia, such as IL-18, IL-6 and TNF-α (16). For example, in MCAO model mice, a significant increase in NF-κB (p65) expression levels were detected in the ischemic brain, which was accompanied by an increase in IL-18, IL-6 and TNF-α expression levels, whereas an inhibitor of NF-κB significantly inhibited the release of these cytokines (16). In addition, inhibiting the increased upregulation of activated NF-κB (p65) decreased the production of hypoxia-induced factor-1α, which then attenuated the neuroinflammatory responses in ischemic stroke (69).

STAT family members

STAT1-STAT6 exist in the cytoplasm and serve an important role in the regulation of inflammatory responses. In hypoxia-activated BV2 microglia, the phosphorylation levels of STAT1 were significantly increased, which was accompanied by increased expression levels of M1 microglia markers, such as cluster of differentiation 68 and iNOS (70); however, the expression levels of these markers were significantly decreased in STAT1−/− BV2 cells following hypoxia. Several previous studies have also identified that STAT3 is also associated with the regulation of M1 microglia polarization in both an MCAO-induced and a bilateral common carotid arteries stenosis (BCAS)-induced model of ischemic stroke (19,71–73). Ding et al (72) identified that the inhibition of the Janus kinase 2 (JAK2)/STAT3 signaling pathway promoted the transition from resting microglia to the M2 phenotype to exert anti-inflammatory effects; briefly, the phosphorylation levels of JAK2 and STAT3 were increased in the acute phase of cerebral ischemia in the mice, which was accompanied by a significant increase in M1 microglia markers, including CD16, IL-1β and TNF-α, and a mild increase in M2 microglia markers, such as CD206 and IL-10. However, SC99, an inhibitor targeting the JAK2/STAT3 signaling pathway, inhibited the production of M1 markers and increased the expression of M2 markers, exerting significant anti-inflammatory effects both in vivo and in vitro. Similarly, another study demonstrated that AG490 (a tyrosine kinase inhibitor of JAK2) decreased the expression levels of M1 markers, iNOS, IL-1β, TNF-α, CD16 and CD32, the phosphorylation of IκBα and translocation of p50/p65 in serum soluble tumor necrosis factor ligand superfamily member 6 (Fas ligand)-induced M1 microglia (19). Conversely, another previous study demonstrated that the activated STAT3 and STAT6 pathways could promote the polarization of nonactivated microglia to the M2 phenotype (74). Therefore, based on these contradictory observations, further studies are required to elucidate whether STATs and their associated signaling pathways can regulate microglia polarization.

Nrf2

The redox-sensitive transcription factor Nrf2 demonstrates both antioxidant and anti-inflammatory properties (75,76), and Nrf2 and its associated signaling pathways have been identified to exert a protective effect following ischemic stroke; for example, in Nrf2−/− mice, the neurological deficit and cerebral infarct volume were both aggravated following MCAO compared with the wild-type mice (75). Meanwhile, the beneficial effects of sirtuin 6, a member of the sirtuin family of NAD+-dependent deacetylases, were abolished in Nrf2−/− mice following ischemic stroke, in that the reactive oxygen species (ROS) levels and the number of dead neuro-2A cells in Nrf2−/− mice were increased (75). In addition, another previous study demonstrated that L-F001, a novel multifunctional Rho-associated protein kinase inhibitor, increased the expression levels of the M2 microglia marker CD206 through the activation of Nrf2 signaling pathways in vitro (63). These beneficial results have also been validated in vitro; in oxygen-glucose deprivation/reperfusion (OGDR)-activated BV2 microglia, Nrf2 was observed to be activated and dissociated from the repressor protein, Kelch ECH-associating protein 1. This facilitated its translocation from the cytoplasm to the nucleus, where it bound to the antioxidant response element (ARE) and induced the expression of downstream genes, such as NADPH quinone oxidoreductase and heme oxygenase 1, which scavenge ROS and nitric oxide (7,68,77). Additionally, the activation of Nrf2 was identified to decrease ROS, NOD like receptor family pyrin domain containing 3 (NLRP3) and IL-1β expression levels in BV2 microglia (7). However, despite the evident protective role of Nrf2, further studies are required to elucidate whether this transcription factor and its signaling pathways may regulate M2 microglia polarization and function following ischemic stroke.

PPARγ

PPARγ is a ligand-activated transcription factor that is considered to be the primary mediator of inflammatory responses. In acute ischemic stroke, PPARγ is reported to be activated, which has been demonstrated to directly decrease tissue damage through inhibiting the NF-κB pathway, decreasing inflammation and stimulating the Nrf2/ARE axis to decrease oxidative stress (78). A previous study observed that 10-O-(N,N-dimethylaminoethyl)-ginkgolide B methanesulfonate, a novel derivative of Ginkgolide B, could transform polarized BV2 microglia from the M1 to the M2 phenotype through promoting PPARγ translocation from the nucleus to the cytoplasm following OGD; this subsequently promoted the decreased expression levels of TNF-α and the increased expression levels of TGF-β1 (26). Additionally, IL-4 generated by neurons was identified to bind to the IL-4 receptor expressed on the surface of microglia and activate M2 microglia through regulating the PPARγ signaling pathway (52); the activated M2 microglia were observed to phagocytose damaged cells and secrete brain-derived neurotrophic factor (BDNF) to decrease ischemic brain damage (52). Based on these studies, it was suggested that the modulation of PPARγ and its associated signaling pathways may represent a potential target for the treatment of ischemic stroke.

Receptors

Receptors are proteins, usually cell surface receptors, that recognize and bind to ligands and serve an important role in the signaling pathways involved in inducing local inflammation, the recruitment of new effector cells, the containment of local infection and the initiation of an adaptive immune response. Previous studies have identified that 3 important membrane receptors, TLR4, sphingosine 1 phosphate receptors (S1PRs) and thromboxane A2 receptor (TXA2R), are closely associated with microglia polarization (31,70,79,80).

TLR4

TLR4, which is predominantly expressed by microglia, is an important regulator of inflammatory responses (79). Increased protein expression levels of TLR4 are reportedly associated with poor outcomes in patients with ischemic stroke (81). Furthermore, in ischemic stroke, TLR4 has been observed to recognize damage-associated molecules, such as lipopolysaccharide (LPS) and high mobility group box-1, and trigger inflammatory responses (Fig. 2) (82,83). In mice, the number of TLR4+ cells, as well as NF-κB+ and IL-1β+ cells, was significantly increased at 72 h following MCAO; the number of TLR4+ cells was identified to be closely positively correlated with the number of IL-1β+ cells and the infarct volume (84). In addition, in TLR4−/− mice, the infarct volume was decreased compared with TLR4+/+ mice following MCAO (85). These results have also been validated in vitro; in BV2 microglia, a significant increase in TLR4 expression levels were detected following OGD, which was accompanied by an increase in proinflammatory cytokines, including TNF-α, IL-1β and IL-6 (86); however, in TLR4−/− mice, expression levels of these cytokines generated by M1 microglia were observed to be significantly decreased through the inhibition of the MyD88/NF-κB signaling pathway (79,87). These results suggested that the overexpression of TLR4 may activate and promote M1 microglia to trigger neuroinflammatory responses in ischemic stroke, which subsequently aggravates the brain damage. However, in another study, TLR4 was demonstrated to be positively associated with brain tissue repair (85): TLR4 activation regulated cell proliferation and neuroblast migration in the subventricular zone on days 1, 7 and 14 following MCAO, which suggested an effective means to modulate neurogenesis following ischemic stroke. Based on this conflicting evidence, further studies on the effects of TLR4 in different cells and at different time points in ischemic stroke are required.

S1PRs

S1PR1-S1PR5 are G-protein-coupled receptors that are expressed in abundance in the microglia and are demonstrated to regulate inflammatory responses following ischemic stroke (88). S1P is the ligand for S1PR and has been identified to bind to S1PR1, S1PR2 and S1PR3 to trigger neuroinflammatory responses in ischemic stroke (Fig. 2) (89–91). An in vitro study reported that upon the addition of S1P to the culture medium of primary microglia, the expression levels of IL-17 were increased (31). Additionally, the exogenous administration of S1P to OGDR microglia aggravated OGDR-induced neuronal apoptosis (31). In MCAO model mice, sphingosine kinase 1 (Sphk1) was demonstrated to phosphorylate sphingosine to S1P, and then S1P bound to S1PR3 to activate M1 microglia, which released proinflammatory cytokines and aggravated brain damage. This culminated in an increase in the expression levels of S1P and M1 microglial markers, CD11b, CD16, CD32, CD86, TNF-α and IL-1β, as well as an increased infarct volume and neurological deficit (81). CAY10444, an inhibitor of S1PR3, and PF-543, an inhibitor of Sphk1, were identified to partially reverse the result to increase the expression levels (31,91); however, Qin et al (71) described opposing results, following the observation that Fingolimod, an S1PR agonist, inhibited M1 microglia-induced neuroinflammation and shift microglia toward M2 polarization through regulating STAT3 signaling in a BCAS-induced model of ischemic stroke. Therefore, additional studies are required to determine the role of S1P and S1PRs in the regulation of microglia polarization following ischemic stroke.

TXA2R

TXA2R, another G-protein-coupled receptor, can promote platelet activation and aggregation, which regulates thrombosis/hemostasis and inflammatory responses (92,93). In ischemia/reperfusion mice, significant TXA2R expression levels were detected in microglia/macrophages by immunofluorescence double staining (80). In addition, the expression levels of M1 microglia markers, such as CD16, CD86, IL-1β, IL-6 and TNF-α, were significantly increased in ischemic brains (80). However, SQ29548, a TXA2R antagonist, inhibited the activation of M1 microglia and subsequent inflammatory responses, which culminated in a decrease in CD16, CD86, IL-1β, IL-6 and TNF-α expression levels.

Ion channels

The expression of ion channels changes in response to voltage and pH gradients in the microenvironment, which induces intracellular signal transduction. Currently, 2 important ion channels, Hv1 and Kv1.3, have been reported to be closely associated with microglia polarization (83,94).

Hv1

Hv1 is a voltage-gated proton channel that enables NADPH oxidase (NOX) to produce ROS by sensing both voltage and pH gradients (95). ROS have been identified to be one of the main factors contributing to the induction of M1 microglia to release proinflammatory cytokines (96). For example, a previous study demonstrated that an increase in Hv1 expression levels induced neuroinflammatory responses through the Hv1/NOX/ROS pathway, thereby exacerbating brain damage in ischemic stroke (95). Furthermore, in both an MCAO-induced and a PT-induced model of ischemic stroke, large Hv1-mediated currents were detected in microglia, but not in neurons or astrocytes (94,95); this was accompanied by increased expression levels of ROS in the brain, which activated its downstream target NF-κB to trigger the secretion of proinflammatory cytokines generated by M1 microglia, further aggravating brain damage following ischemic stroke. In Hv1−/− mice, the brain infarct volume, neuronal damage and motor deficits were attenuated compared with wild-type mice following PT induction, which was accompanied by decreased expression levels of the M1 microglia markers, iNOS and CD16 (94). Additionally, the study also confirmed that more ROS could be released by M1 microglia following ischemic stroke, and in turn, these increased levels of ROS promoted M1 microglia polarization, resulting in a vicious circle of neuroinflammation, which exacerbated neuronal death. Another study demonstrated that the regenerative rate of oligodendrocyte precursor cells was increased, and subsequently the periventricular leukomalacia was reduced in Hv1−/− mice, confirming that Hv1 may aggravate brain damage (96). These results indicated that Hv1 may aggravate brain damage through increasing the expression levels of ROS and proinflammatory cytokines generated by M1 microglia; however, whether Hv1 affects M2 microglia polarization remains unclear.

Kv1.3

Previous studies have identified that changes in the expression levels of K+ channels also occur in ischemic stroke and this can regulate microglia polarization (83,97). K+ channels are membrane proteins that permit the rapid and selective flow of K+ ions across the cell membrane, thus generating electrical signals in cells. In response to brain ischemia, the expression levels of Kv1.3 were observed to be increased in microglia of both rodents and humans, which modulated Ca2+ signaling and induced neuroinflammation (97). In addition, in an LPS-induced microglia activated mouse model, M1 microglia were activated, accompanied by an increase in the current density and mRNA expression levels of Kv1.3 and inflammatory cytokines, such as TNF-α, IL-1β and IL-6, in addition to iNOS (83). Notably, Kv1.3 knockout or treatment with the Kv1.3 inhibitor 5-(4-phenoxybutoxy)psoralen (PAP-1) could partially reverse these results; in LPS-activated microglia the expression levels of Kv1.3 and M1 microglia markers, such as TNF-α, IL-1β and nitric oxide were significantly increased; however, PAP-1 was observed to decrease the polarization of M1 microglia, as well as the expression levels of proinflammatory cytokines (83). These data suggested that Kv1.3 may be one of the main mediators of M1 microglia polarization.

Modulators of gene expression

H19 imprinted maternally expressed transcript (H19), microRNA (miRNA)−155, miRNA-124 and chemokine-like protein TAFA-3 (FAM19A3) have been reported to be closely associated with inflammatory responses; the first 2 factors are closely associated with M1 polarization, whereas the latter 2 factors are associated with M2 polarization (18,98–100).

H19

H19 is a long non-coding RNA (lncRNA) that has been demonstrated to induce the onset of ischemic stroke and participate in the chronic regeneration process following ischemic stroke (98,101–103). In fact, H19 has been demonstrated to promote M1 microglia polarization and induce neuroinflammation in ischemic stroke through regulating histone deacetylases (HDACs) (98,104,105). In clinical experiments (98,101), H19 levels were identified to be significantly increased in the plasma at 3 h, 7, 30 and 90 days post-ischemic stroke, in addition to being positively correlated with the National Institute of Health Stroke Scale score (106) and TNF-α expression levels. Furthermore, in MCAO model mice, H19 expression levels were increased in both the plasma and the brain (98). Notably, H19 knockdown decreased the infarct volume and brain edema and decreased TNF-α and IL-1β expression levels in the brain following ischemic stroke (98). Furthermore, the study also demonstrated that H19 knockdown could inhibit OGD-induced M1 microglia polarization in BV2 cells, which manifested as a decrease in the expression levels of TNF-α and CD11b (98). In addition, H19 knockdown reversed the OGD-induced upregulation of HDAC1, whereas HDAC1 overexpression reversed the effects of H19 knockdown and promoted microglia polarization toward the M1 phenotype (98). These data could be used to regulate M1 microglia polarization through the H19/HDAC1 signaling pathway.

miRNA-155

Increased expression levels of miRNA-155 have been demonstrated to be associated with the neuroinflammatory responses (20,107). In LPS-activated microglia, miRNA-155 expression levels were significantly increased, which could target suppressor of cytokine signaling 1 to trigger M1 microglia-mediated inflammatory responses and aggravate brain damage (18). In addition, miRNA-155−/− mice were reported to partially reverse LPS-induced neurological deficits and microglia activation (107); however, the roles of miRNA-155 in microglia polarization in ischemic stroke require additional clarification.

miRNA-124

miRNA-124 is the most abundant miRNA present in the CNS and is involved extensively in neuroprotective mechanisms (108,109). In fact, it has been identified to regulate microglia activation and polarization in ischemic stroke (99,110). A previous study demonstrated that miRNA-124 induced neuroprotection and functional improvement through regulating M2 microglia polarization in ischemic stroke (99); upon injection of miRNA-124 into the mouse ipsilateral hemisphere, the number of M2 microglia/macrophages was markedly increased, contributing to neuronal survival by releasing the trophic factor Arg-1 following MCAO. miRNA-124 treatment also resulted in the accelerated infiltration of reactive astrocytes into the lesion area, which subsequently resulted in a tighter glial scar border; however, whether the activation and polarization of microglia affects astrocyte function through modulating miRNA-124 remains unclear.

FAM19A3

FAM19A3, a member of the TAFA family of genes (TAFA1-5), is predominantly expressed in the CNS (111). Shao et al (100) suggested that FAM19A3 expression levels were significantly increased in the microglia of MCAO model mice, which promoted M2 microglia polarization and ameliorated cerebral ischemia. Notably, within these mice, FAM19A3 expression levels in the brain were elevated 1–3 days following MCAO, peaked at day 5, began to decrease at day 7 and returned to pre-injury levels by day 14, which is consistent with the pattern of M2 microglia markers, CD206, IL-10, Ym1/2 and TGF-β discussed previously. Additionally, FAM19A3 was observed to be secreted by M2 microglia and in turn, FAM19A3 increased the expression levels of M2 microglia markers in a dose-dependent manner (100). These results suggested that FAM19A3 may be a genetic marker for M2 microglia; however, the underlying molecular mechanisms remains to be investigated.

Drug modulators

In clinical and experimental studies, a large number of drugs have been demonstrated to serve important roles as modulators of neuroinflammation and microglia polarization; for example, in both a PT-induced stroke model and BCAS-induced model, Fingolimod, a United States of America Food and Drug Administration-approved drug for the treatment of relapsing-remitting multiple sclerosis, modulated microglia toward M2 polarization, protected against ischemic damage and promoted angiogenesis through suppressing the STAT3 signaling pathway (47,71). In MCAO model mice, rosiglitazone decreased the numbers of Iba1+/CD16+ M1 microglia and increased the numbers of Iba1+/CD206+ M2 microglia following stroke, which improved the overall long-term white matter integrity (61,112). In addition to western medicines, some traditional Chinese medicines can also be used to regulate microglia polarization (16,25,113); a study by Yang et al (16) demonstrated that baicalein significantly decreased the expression levels of the M1 markers, CD16 and CD86, and increased the expression levels of the M2 markers, CD163 and CD206. Additionally, expression levels of the proinflammatory factors, IL-6, IL-18 and TNF-α were decreased upon inhibition of NF-κB signaling following baicalein treatment in MCAO model mice (16). Zhu et al (25) demonstrated that berberine could inhibit M1 polarization and promote M2 polarization to protect against ischemic stroke through the AMPK signaling pathway. Wang et al (46) reported that xinnao shutong capsules, a Chinese medicine produced in China, decreased the number of M1 microglia and increased the number of M2 microglia, which improved memory impairments in a BCCA-induced model of ischemic stroke (46). Nonetheless, despite the effects of these drugs on microglia polarization being confirmed in experimental studies, clinical studies are still lacking.

Autophagy

Autophagy is a cellular metabolic pathway through which damaged organelles and misfolded proteins are degraded and recycled to maintain cellular homeostasis. Accumulating evidence has reported that autophagy is activated in brain cells such as neurons, endothelial cells and microglia in ischemic stroke, and interfering with autophagy can aggravate brain damage (114–116). For example, in an in vitro study, the inhibition of autophagic flux with NH4Cl was demonstrated to aggravate the damage to brain cells, as it stimulated M1 microglia polarization and the release of TNF-α, iNOS and cyclooxygenase-2 (117). In addition, the expression levels of M2 markers, including IL-10, Arg-1 and BDNF were decreased following the treatment of NH4Cl in OGD microglia. mTOR, a 289-kDa serine/threonine protein kinase, was demonstrated to inhibit autophagy through regulating its downstream signaling pathways (118), whereas Li et al (119) observed that mTOR complex 1 (mTORC1), a form of mTOR, prompted microglia to shift towards the M1 phenotype. Moreover, the inhibition of mTORC1 decreased the infarct volume and improved motor function, which lead to a decrease in the production of proinflammatory cytokines, including iNOS, TNF-α and IL-1β, and chemokines, C-C motif chemokine 2 and C-C motif chemokine ligand 22, by M1 microglia. However, another previous study proposed the opposite view and suggested that autophagy may stimulate a microglial shift toward the M1 phenotype to exacerbate cerebral ischemia (120). In an in vitro study, the expression levels of the autophagy markers microtubule-associated proteins 1A/1B light chain 3B lipid modified form, beclin-1 and autophagy protein 5 were increased, which was accompanied by increased expression levels of the proinflammatory factors IL-1β, TNF-α and iNOS in primary microglia under OGD (120); however, these proinflammatory factors and the number of M1 microglia were downregulated by the autophagy inhibitor 3-methyladenine (3-MA). Moreover, 3-MA also prevented M2 microglia polarization through inhibiting the activation of CREB. Thus, based on these contradictory results, the roles of autophagy in microglia polarization in ischemic stroke require further investigation.

Conclusions

Microglia activation and polarization is considered to be important in brain damage and repair following ischemic stroke. The present review provided an overview of the activation and changes of M1 and M2 microglia following ischemic stroke. This is useful for determining the optimal time point at which to target microglia to treat ischemic stroke. Additionally, studies investigating the modulators of microglia activation and polarization were analyzed, as well as the role of autophagy in microglia polarization in ischemic stroke. This is of importance as it provides therapeutic targets to target microglia in ischemic stroke and it is also critical for our understanding of the pathophysiology and future investigations of ischemic stroke. It is worth noting that according to the American Heart Association/American Stroke Association guidelines intracerebral hemorrhage and ischemic stroke are 2 different types of stroke with different pathogenesis and treatments (106,121). Lan et al (122) have already reviewed the modulation of microglial activation and polarization in intracerebral hemorrhage from 3 aspects: Microglial activation and function in intracerebral hemorrhage; modulators of microglial polarization; and microglial interactions with other cells. However, the present review differs from the review by Lan et al (122), as the present study investigated the changes and modulation of microglia in ischemic stroke. Additionally, Lan et al (122) primarily focused on the modulation of microglial polarization by the transcription factors and receptors. In addition to these typical modulation factors, which also serve a similar role in the modulation of microglia in ischemic stroke, the present review also discussed other factors associated with the modulation of microglial polarization, such as ion channels, genes and drugs. Notably, the time and spatial changes of microglia in ischemic stroke were further examined in this review. In practice, this review suggested future research directions to provide novel strategies for the treatment of ischemic stroke.

Based on the data presented in the present review, several issues can be selected for further investigation. Firstly, the experimental models of ischemic stroke need to be improved and more data on the experimental models needs to be collected and collated, to decrease the discrepancies reported between studies. Secondly, beneficial factors identified that regulated the polarization of microglia towards the M2 phenotype in animals with ischemic stroke require further investigations. Thirdly, further investigations are required into the role of autophagy in microglia polarization to improve our understanding of the pathology of ischemic stroke. In addition, regarding the extensive roles of microglia in ischemic stroke, investigations into the interaction between microglia and other cells will be important to identify the effective pathophysiological pathways that prevent brain damage. Finally, the focus of this review was on the pathophysiological regulatory mechanisms of microglia polarization. Therefore, the range of potential drug targets has not been fully explored. Our future studies will focus on other aspects of the roles of microglia in ischemia.

Acknowledgements

Not applicable.

Funding

This work was supported by the National Natural Science Foundation of China (grant nos. 81874416 and 81774174), the Hunan Education Department's Science & Research Project (grant no. 16K063) and the Department of Integrated Traditional Chinese and Western Medicine, Hunan University of Traditional Chinese Medicine (grant no. 2018ZXYJH01).

Availability of data and materials

Not applicable.

Authors' contributions

CTJ collected the literature, performed the analysis and interpretation and wrote the manuscript. WFW, YHD and JWG contributed to literature collection, and analysis and interpretation of the results. CTJ, YHD and JWG revised the manuscript. All authors reviewed the final draft of the manuscript.

Ethics approval and consent to participate

Not applicable.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References