Microglia, the resident immune cells of the brain, constitute between 6-21% of the brain's total glial cell population (5). Historically, the understanding of microglia's origins underwent a paradigm shift in the early 1990s. Contrary to the previously held belief that microglia solely originated from peripheral macrophages, it was unveiled that they could also arise from bone marrow progenitor cells located in the yolk sac. This novel classification of macrophages was aptly christened 'microglia' by Pío del Río-Hortega, a terminology that has since gained universal acceptance (6,7).

Despite their behavioral and functional parallels with macrophages, which led to their initial identification as such (8), microglia maintain a distinct and irreplaceable role within the central nervous system (CNS). A mere h post-ischemic insult, resident microglia spring into action. These cells undergo notable morphological transformations: Retracting their external protrusions, they adopt an amoeboid form (9). Additionally, based on temporal progression post-injury, microglia can be categorized into acute, subacute and chronic phases, each presenting unique pathological characteristics (10).

Within mere h post-ischemic injury, microglia are promptly activated by the infiltration of plasma proteins. Once stimulated, they infiltrate both the infarct core and the peri-infarct region. This activation is mediated by a suite of downstream signaling effectors, notably including nuclear factor κB (NF-κB), hypoxia-inducible factor 1 (HIF-1) and activator of transcription 3 (STAT3). Morphologically, peri-infarct microglia exhibit enlarged cell bodies, short branches and amoeboid structures. Adjacent to the ischemic core, particularly large, round microglia display heightened activity, attributed to their low activation thresholds.

Historically, microglia were perceived to polarize into two distinct phenotypes: M1 and M2, in response to activation. The M1 phenotype, characterized by its pro-inflammatory and neurotoxic activities, emerges within the initial 24 h post-injury. This induction is primarily steered by the Toll-like receptor (TLR) and interferon-gamma (IFN-gamma) signaling pathways (11). M1 microglia secrete a plethora of pro-inflammatory factors such as tumour necrosis factor alpha (TNF-alpha), IL-6, IL-12 and IL-1β, alongside reactive oxidative agents such as reactive oxygen species (ROS) and inducible nitric oxide synthase (12,13). These agents, particularly ROS and TNF-alpha, can disrupt the integrity of the blood-brain barrier (BBB), inflicting severe endothelial damage and potentially leading to hemorrhagic transformation (14,15).

Conversely, while the pro-inflammatory M1 phenotype emerges, certain chemokines, including CCL2 and CXCL4, initiate the polarization of microglia towards the anti-inflammatory M2 phenotype post the 24-h mark (16,17). Distinct from M1, the M2 phenotype secretes anti-inflammatory agents such as IL-4, IL-10 and transforming growth factor beta, aiming to curb inflammation (18). These cells also release growth factors and neurotrophic agents including insulin-like growth factor 1, vascular endothelial growth factor and brain-derived neurotrophic factors to foster vascular repair and prevent neuronal cell apoptosis in the ischemic zone (19). Furthermore, the M2 phenotype is nuanced, with subcategories M2a, M2b and M2c (20). Each subtype is characterized by its induction mechanism and function (21).

However, the once clear-cut M1/M2 distinction is now acknowledged as an oversimplification. Contemporary understanding recognizes the vast functional overlaps and asserts that the complete array of microglial activation states in vivo extends beyond just the conventional M1 or M2 categories (22). For instance, post-ischemia, the presence of M1-like microglia amplifies from day one and peaks around the second week, whereas M2-like microglia peak around day five and subsequently decline (23). Notably, older mice exhibit significantly fewer M2 microglia compared with their younger counterparts, underscoring the adverse influence of advanced age on post-stroke neuroinflammation (24-26). The intricate nature of microglial responses is epitomized by findings where markers of both inflammatory resistance and promotion co-exist within the same cell. This complexity is a key factor influencing the challenges in treating IS injury with whole microglial cells (27).

While research efforts into the acute phase of post-stroke neuroinflammation have burgeoned over the years, studies exploring the subacute and chronic phases remain relatively sparse. During these later phases, damage-associated molecular patterns activate microglia and infiltrating macrophages. These activated cells subsequently release inflammatory cytokines such as TNF-alpha and IL-1beta, exacerbating neuronal cell death in the ischemic penumbra (28). Remarkably, even after six months post-injury, blood samples from stroke patients persistently exhibited elevated levels of TNF-alpha (29).

Anti-inflammatory M2-type microglia play a pivotal role during this phase by secreting cytokines, most notably IL-10, which counteract the inflammatory effects incited by IL-1beta and TNF-alpha, providing a protective barrier against further ischemic damage (30). Supporting this, a study by Ooboshi et al (31) demonstrated that administering an adenoviral vector encoding the human IL-10 gene into the lateral ventricles of stroke-affected rats substantially reduced cerebral infarct sizes and mitigated hippocampal ischemic injuries. The likely mechanism involves the capacity of IL-10 to dampen the production of inflammatory cytokines and enhance the activity of anti-inflammatory cytokine antagonists (31).

Interestingly, microglia can sustain a state of prolonged activity. While their activity tends to wane after six months within the infarct core, it often radiates to more distant regions, following the brain's corticospinal tracts (32). Modern imaging techniques such as positron emission tomography and magnetic resonance imaging have unveiled that, for certain patients, neuroinflammation migrates from the infarct's initial site to remote brain areas, encompassing the hippocampus, thalamus and basal ganglia, even those situated contralaterally (33). Echoing this, Price et al (34) observed a decline in microglial activity in the infarct core three to four weeks post-stroke. Conversely, there was a conspicuous surge of microglia in distant regions of the ipsilateral hemisphere and even in areas within the contralateral cerebral hemisphere (34).

Astrocytes, akin to microglia, contribute significantly to the brain's innate immune response. These versatile cells govern ionic homeostasis, modulate neurotransmitter dynamics and contribute to the structural integrity of the BBB, thus playing pivotal roles in safeguarding both neuronal and vascular functionalities. In the context of CI/RI, astrocytes have been linked to the release of Kruppel-like factor 4 (KLF4), which not only curtails the infarcted area and suppresses oxidative stress but also fortifies the BBB. Intriguingly, KLF4 amplifies the levels of Nrf2 and Trx1 mRNA in the oxygen glucose deprivation/reperfusion (OGD/R) model, resulting in improved BBB repair, diminished malondialdehyde levels and elevated antioxidant SOD levels (35).

Astrocytes are swiftly activated post-brain injury and exhibit a dual-functional paradigm (36). Classically activated A1-like astrocytes, stimulated by IL1 alpha, TNF alpha, and complement component 1q (C1q) from microglia, unleash pro-inflammatory molecules, manifesting neurotoxic implications (37). By contrast, alternatively activated A2-like astrocytes confer neuroprotection via the secretion of anti-inflammatory mediators (38). Notably, post-brain I/R, lipopolysaccharide (LPS)-driven neuroinflammation precipitates the transformation of astrocytes to the A1 type. Research by Li et al (39) illuminated that hydrogen sulfide (H2S) curbs this LPS-induced A1-like astrocyte metamorphosis in the murine hippocampus, simultaneously promoting the shift of reactive astrocytes to the A2 type. This transformation was hypothesized to be modulated by the upregulation of the BKCa channel (39). In a complex cascade, A1 astrocytes release LCN2, which subsequently binds to 24p3R on astrocytes, instigating focal death of astrocytes via the NLRP3 mechanism. Yet, mitigating this astrocytic focal death curbs neuronal damage, particularly through attenuating caspase-1 activation, subsequently diminishing neuronal apoptosis (40).

Reactive astrocytes, once activated in the wake of cerebral ischemic injury, orchestrate a neuroglial scar within the peri-infarct cortex (PIC). This scar acts as a protective barrier, isolating the injury site from healthy tissues and impeding the spread of deleterious agents (41). Nevertheless, as cerebral ischemia advances, this very neuroglial scar evolves into an obstacle, hampering post-ischemic brain repair. Yuan et al (42) uncovered that p-Hydroxybenzaldehyde, derived from asparagus, curtailed astrocytosis in middle cerebral artery occlusion (MCAO) rat PICs. This led to a reprogramming of astrocytes, transitioning the neuroglial scar back to neuronal progenitor cells, thus promoting neural and vascular rejuvenation within the PIC (42). Another noteworthy study revealed that miR-124, encapsulated within the small extracellular vesicles (EVs) of M2-type microglia, expedited the reprogramming of astrocytes. This reprogramming counteracted the impediments posed by the glial scar in the PIC, catalyzing stroke recuperation (43).

The BBB is a complex and multifaceted structure, comprising astrocytes, pericytes, the extracellular matrix and the endothelial cells that line the brain's microvasculature. Its pivotal role is to safeguard the CNS. Not only does the BBB facilitate the entry of vital metabolic substrates such as oxygen and glucose into the brain, but it also acts as a vigilant sentinel, preventing the infiltration of pathogens, non-CNS cells, and other potentially harmful agents. In doing so, the BBB plays an indispensable role in preserving the delicate homeostasis of the brain (44).

Astrocyte-derived Matrix metalloproteinase-2/9 (MMP-2/9) have been implicated in disrupting the integrity of BBB. Specifically, these enzymes catalyze the degradation of tight junction protein (TJP) ZO-1, thereby compromising BBB permeability (45). Corroborating this, research has revealed that MMPs degrade the BBB's TJPs. However, agents such as BB-1101 have been shown to counteract this degradation, thereby bolstering BBB integrity (46).

In the context of I/R, the cyclic RNA of FoxO3 (circ-FoxO3) has been identified as a crucial player. It interacts synergistically with mTOR and E2F1, enhancing autophagic activity through the inhibition of mTORC1. This cascade of molecular events ultimately attenuates BBB-associated neuro-pathologies stemming from damage (47).

The cGAS-STING pathway is a central molecular signaling cascade underpinned by two primary proteins: Cyclic GMP-AMP synthase (cGAS) and the stimulator of interferon genes (STING). Acting as an intrinsic immune sensor, cGAS detects an array of cytoplasmic double-stranded DNA (dsDNA) sources, encompassing viral, bacterial, mitochondrial, micronuclear and reverse transcription-derived DNA (48). Once bound to dsDNA, cGAS undergoes an enzymatic transformation, leading to the synthesis of the secondary messenger, cGAMP, from ATP and GTP. This messenger is subsequently identified by STING, a homodimeric protein anchored in the endoplasmic reticulum (ER) membrane. Upon cGAMP binding, STING undergoes a conformational shift, facilitating its oligomerization. This modified STING then translocates to the Golgi apparatus through the ER-Golgi intermediate compartment (ERGIC). Within the Golgi, a palmitoylation event occurs at STING's cysteine residues (Cys88 and Cys91), enabling its C-terminus to liaise with TANK-binding kinase 1 (TBK1). This interaction culminates in the phosphorylation of the STING protein at Ser366, leading to the recruitment and subsequent phosphorylation of interferon-regulating factor 3 (IRF3). As a result, IRF3 dimerizes, undergoes nuclear translocation, and induces target gene expression (49). Parallelly, STING also stimulates the nuclear factor kappaB (NF-kB) pathway, further influencing type 1 IFN expression (50). Experimental data derived from IS models, wherein cGAS was deleted or STING silenced, revealed reduced cerebral infarct injury and enhanced neuronal survival (51). By contrast, overactivation of this cGAS-STING axis has been linked to heightened neuroinflammation (52) (Fig. 1).

Modulating the cGAS-STING pathway has demonstrated therapeutic potential in mitigating neuroinflammation. Specifically, suppressing this pathway diminishes the inflammatory response elicited by LPS/IFN-gamma in BV-2 microglia (53). Such modulation has been noted to ameliorate learning and memory deficits in Alzheimer's disease, making cognitive impairments less pronounced. In addition, downregulation of the cGAS-STING pathway not only inhibited the attenuation of microglia M1 polarization but also reduced the expression of A1-type astrocytes to achieve an anti-inflammatory effect (54,55). The inhibitor RU.521, targeting cGAS, has shown promise in several contexts. It curbs the expression of NF-kB-associated inflammatory cytokines, thereby safeguarding neurons from potential damage. Furthermore, RU.521 effectively counteracts microglial pyroptosis, which is typically incited by the activation of cytosolic gasdermin-D and NLRP3 post cerebral venous sinus thrombosis (CVST) (52). Notably, RU.521 can also counteract the mtDNA-driven activation of the cGAS-STING pathway and modulate autophagy, further restraining the M1 polarization of microglia (56). Consequently, RU.521's modulation of the STING pathway offers a therapeutic avenue for mitigating IS. The evolving understanding and manipulation of the cGAS-STING signaling in microglia underscore its potential as a therapeutic target for managing neuroinflammation following cerebral I/R.

In the context of cerebral I/R, the downstream STING-TBK1-IRF3 pathway of cGAS is adeptly regulated by histone deacetylase 3 (HDAC3). The BCl-2-associated X-gene (BAX) is known to complex with IRF3, promoting immune cell apoptosis-a process that, if unchecked, can amplify post-stroke neuroinflammation (57,58). Both IRF3 and NF-kB, being downstream effectors of STING, can intensify neuroinflammation, especially evident in the OGD/R model (51). However, pharmacological intervention using HDAC3 inhibitors, such as Entinostat and Trichostatin A, has revealed efficacy in curtailing I/R injuries and the consequent neuroinflammation (59). Post-I/R, HDACs undergo oxidative activation. Different HDAC subclasses can either bolster or hinder cell viability and neurogenesis. For instance, inhibiting HDAC1, HDAC2, HDAC3 and HDAC6 enhances cell survival and dampens neuroinflammation (60,61). Conversely, augmenting the activity of HDAC4 and HDAC5 promotes both cell viability and neurogenesis (62).

Neutrophil extracellular traps (NETs) have emerged as central players in the vascular dynamics of stroke. They stimulate platelets, fostering a conducive environment for thrombosis, impede the repair of damaged vessels, breach the sanctity of the BBB, and underlie degenerative shifts observed in the protracted aftermath of stroke (63). Notably, tissue plasminogen activator (tPA), the primary therapeutic agent for thrombolysis and recanalization in IS, instigates neutrophils to yield NETs. Coincidentally, tPA treatment amplifies the cGAS expression, synchronously boosting STING functionality. This chain of events precipitates enhanced phosphorylation of TANK-binding kinase 1 (pTBK1) and ignites TBK1-facilitated activation of interferon regulatory factor 3 (IRF3), culminating in a surge of type I interferon (IFN-I) and interleukin-6 (IL-6) in ischemic zones. A study led by Rosell et al (64) underscored the instrumental role of NETs activation in the cGAS-STING pathway, especially post-tPA administration (64).

In a notable experiment, neutrophils treated with A23187 spawned EVs laden with DNA via NETs. These EV-DNAs, upon internalization by macrophages, activate an alternate cGAS-STING trajectory that propels NF-kB to synthesize IL-6-a process distinctly independent of the conventional cGAS-STING-TBK1-IRF3 signaling mechanism (65). These A23187-induced EVs provide invaluable insights into neutrophil-driven inflammatory cascades under ischemic conditions, offering potential therapeutic avenues to address neuroinflammation, infections, and autoimmunity in stroke patients.

Currently, tPA thrombolytic therapy is the only available treatment for reperfusion in areas affected by cerebral infarction. However, this treatment still carries the risk of cerebral hemorrhage (66,67). NETs or cGAS as a treatment could reduce the risk of hemorrhage post-tPA thrombolysis in IS, offering a new strategy to enhance the safety of tPA thrombolysis (68). By targeting NETs, it is also possible to influence cGAS-STING activation, further reducing M1-type microglia activation induced by STING/IFN-gamma (53). This reduction can alleviate neurotoxic effects, protect the BBB, and prevent further death of neurons in the ischemic penumbra (14,37). Additionally, this approach appears capable of addressing the prolonged elevated levels of TNF-alpha in the serum of CI/RI patients, which can last up to six months (29).

Additional insights suggest that LPS serves as a catalyst for NET production (69), with inflammatory markers, such as TNF-alpha, IL-6 and IFN, correlating with inflammatory intensity of IS (14,70,71). Targeted interventions, such as the JAK 1/2 inhibitor Ruxolitinib (Rux), have shown promise in curbing this inflammation via the JAK2/STAT3 pathway, ultimately reducing cerebral edema and lessening infarct volume post-stroke (72). Caspase1, as part of the NLRP3 inflammasome, contributes to increased neurodegeneration and neurological impairment post-cerebrovascular sinus thrombosis (CVST) (52). Repeated cranial magnetic stimulation, operating through TLR4-NF-kB-NLRP3 pathways, enhances TNF-alpha and IFN-gamma while reducing stroke-adjacent areas. Caspase1 alleviates neuronal stress and limb movement disorders following the reperfusion of cerebral ischemia (71). Furthermore, NLRP3 mediates astrocyte death; targeting STING can suppress NLRP3 production, and inhibiting astrocyte death can reduce neuronal damage (40,73). The efficacy of the cGAS inhibitor RU.521 in reducing nerve damage through NLRP3 inhibition further confirms this (52). Targeting the downregulation of the cGAS-STING/NLRP3, TLR4-NF-κB/NLRP3, or NF-κB/NLRP3 axes to inhibit NLRP3 can alleviate astrocyte pyroptosis. This approach represents a promising future method for treating neuroinflammation following CI/RI (74,75).

The gastrointestinal (GI) tract stands as a cornerstone of the body's immunological defense, harboring an extensive congregation of immune cells. Remarkably, over 70% of the human immune system finds its home within the GI immune system. Resident within this GI tract is a diverse community of microorganisms collectively termed the GI microbiota (GM). Advances in microbiome research have led to the evolution of the classical GBA concept, ushering in the more comprehensive GBMA. This holistic perspective encompasses the GI neuroendocrine system, the GI immune system, the enteric nervous system, and, pivotally, the GM.

Within the GBMA, GM emerges as a pivotal player, intricately influencing metabolic regulation, nutritional physiology, anticancer responses, immunomodulation, and crucially, neurological functions (76,77). Dominating the GI flora are Firmicutes (~48%) and Bacteroidetes (roughly 51%), with the residual composition being a mix of Actinobacteria, Proteobacteria, and other microorganisms (78). Notably, a marked reduction in the abundance of Proteobacteria and certain Firmicutes (such as Clostridia and Lachnospiraceae) has been observed in post-stroke cognitive impairment patients, setting them apart from non-stroke counterparts (79).

The human gut serves as a sanctuary for an array of microorganisms pivotal in either bolstering health or predisposing to disease, additionally playing a central role in immune homeostasis (47). Deviations in the GM's composition and function, termed dysbiosis, carry potent associations with a plethora of health concerns. Internally, within the GI domain, GM dysregulation correlates with inflammatory bowel diseases (IBD) such as ulcerative colitis, Crohn's disease, colorectal cancers and irritable bowel syndrome (80). Externally, beyond the confines of the gut, an imbalanced GM has links with neurodegenerative ailments including AD, IS and Parkinson's disease (PD) (81-83). Intriguingly, IBD patients exhibit a heightened prevalence of PD compared with those without IBD (84). Beyond merely being bystanders, GM can actively sway the onset of stroke by influencing risk factors including hypertension, diabetes, hyperlipidemia and atherosclerosis (85). Furthermore, the role of GM extends to modulating stroke outcomes through its interactions with neural pathways, endocrine-mediated neurotransmitter pathways, metabolite production and intricate immune signaling cascades.

Butyrate effectively inhibits the phosphorylation of both TBK1 and IRF3, a process that is activated in the presence of STING. This action in turn impedes the production of IFN-I (145). Upon stimulation by LPS, Kupffer cells release mtDNA through the voltage-dependent anion channel 1, thereby initiating the cGAS-STING cascade. Notably, within this cascade, IRF3 is revealed to elevate the expression of NLRP3, instigating inflammatory reactions and apoptosis. It is noteworthy that by suppressing STING, the inflammatory and apoptotic effects of LPS are significantly diminished (154).

However, there remains an ambiguity regarding the role of IRF3 in the LPS-driven TLR4-NF-kB-NLRP3 signaling pathway. Further comprehensive studies are essential to elucidate the interrelation between the STING-IRF3 pathway and the TLR4 signaling mechanisms.

Fluvoxamine, recognized as an SSRI, exhibits a dual inhibitory action on both cGAS and STING (155). The modulation of the cGAS-STING signaling pathway by these specific gut microbial metabolites within the GBMA framework opens promising avenues for potential targeted pharmaceutical interventions or nutritional strategies to address cerebral I/R.

Pathology NF-kB mediates neuroinflammation by modulating the expression of multiple inflammatory factors such as IL-6, TNF-α and COX2 (156,157). L-alpha-GPC, a dietary supplement, undergoes metabolic transformation in vivo, leading to the formation of TMAO, which in turn activates the NF-kB and MAPK signaling pathways (99). Notably, the effects of cholinergic and anti-cholinergic agents on IS might be attributed to their modulation of TMAO. Furthermore, LPS and homocysteine (Hcy) has been identified as activators of the NF-kB signaling cascade (158,159).

Interestingly, certain SSRIs such as Fluoxetine (160) and Tropisetron (161) have been demonstrated to inhibit the NF-kB signaling pathway in cerebral I/R scenarios. Their anti-inflammatory properties serve to confer neuroprotective benefits.

The phosphatidylinositol 3-kinase (PI3K)/protein kinase B (PKB, also known as AKT) signaling pathway has been identified as a key regulator of apoptosis-associated protein expression and cell survival in CI/RI, thus playing a significant neuroprotective role (162,163). Conversely, the mitogen-activated protein kinase (MAPK) signaling pathway is implicated in inflammatory processes in CI/RI (164). Activation of the MAPK pathway triggers a substantial release of inflammatory factors (165). Inhibiting this pathway can reduce astrocyte apoptosis, attenuate astrocyte-mediated inflammatory responses, and protect brain neurons from I/R injury (166). Additionally, blocking the MAPK pathway may alleviate oxidative stress-induced damage and reduce microglial activation and the overexpression of inflammatory factors (167).

Aloe-emodin acts as a potent modulator, attenuating the LPS-induced TLR4-NF-kB pathway while simultaneously enhancing the TLR4-PI3K-AKT signaling. This dual activity offers promising therapeutic potential for mitigating neurological damage and inflammation associated with IS (168). Several other agents have displayed similar inhibitory effects on the LPS-triggered TLR4-NF-kB pathway. Notably, the depletion of Polymerase δ-interacting protein 2 (169), reduction in Cofilin (170), neuroprotective agents (171) and Escin (172) have all shown substantial inhibitory effects on this pathway.

Furthermore, several compounds, including Esculentoside A (173), Imperatorin (174), Progesterone (175), Probucol (176), Hydroxysafflor Yellow A (177) and steppogenin (178), have been identified to impede the LPS-mediated TLR4-NF-kB/MAPK pathway, specifically in BV-2 microglia. This suggests their potential utility in protecting the brain and supporting neuronal health, either as prophylactic agents before IS or as therapeutic agents during cerebral I/R, by targeting the LPS-driven TLR4 pathway within the GBMA framework.

The MAPK signaling pathway, particularly its p38 branch, plays a crucial role in the cellular responses post-IS. A notable influencer of this pathway is Hyperhomocysteinemia (hHcy). Recognized as a standalone risk factor for stroke, hHcy's interactions with the MAPK pathway heighten its implications. Specifically, hHcy's activation of the p38 MAPK pathway exacerbates neurotoxic effects after an ischemic event (179,180). This correlation underscores the importance of understanding the mechanistic details of hHcy and the MAPK pathway. Exploring this further could unveil potential therapeutic targets that can mitigate the detrimental aftermath of stroke events.

The PI3K)-AKT signaling pathway stands as a crucial component in the cellular responses following IS. One noteworthy mediator of this pathway is SCFA. Research has illuminated SCFA's potential in mitigating the adverse effects of IS by activating the PI3K-AKT signaling route, thereby inhibiting neuronal apoptosis, curbing neuroinflammation and enhancing cognitive function (181). Additionally, experiments involving intranasal administration of sodium butyrate have revealed a significant reduction in infarct volume and overall improvement in neurological function, outcomes attributed to the activation of the PI3K-AKT pathway (182). Conversely, hHcy is a noteworthy antagonist in this context. By suppressing the PI3K-AKT pathway, hHcy amplifies neurotoxic effects post-ischemic damage (113), highlighting the pathway's double-edged nature in the realm of IS (Fig. 2). The aforementioned content is summarized in Table II.