Neuroprotective effects of electroacupuncture in ischemic stroke: from mechanisms to clinical implications

Autophagy, an evolutionarily conserved eukaryotic cellular process, sustains intracellular homeostasis through macromolecule and organelle degradation. This mechanism employs autophagosomes and lysosomes to achieve dual biological outcomes: recycling of degradation products or induction of non-apoptotic programed cell death. Under pathological conditions, excessive autophagy may trigger cell death (Levine and Kroemer, 2019; Liu S. et al., 2023; Vargas et al., 2023; Zhang et al., 2022). Autophagy is functionally categorized into two subtypes: non-selective (bulk) autophagy and selective autophagy. The latter specifically mediates targeted degradation of individual cellular components, exemplified by mitophagy in mitochondrial clearance (Klionsky et al., 2021).

Current studies report conflicting findings regarding the neuroprotective effects of EA-mediated autophagy regulation during cerebral ischemia-reperfusion. In a recent investigation, Qiu et al. quantified autophagy markers (LC3 and p62) in CIRI rat neurons, observing that EA decreased LC3 expression and increased p62 expression. Researchers administered EA at Baihui, Quchi, and Zusanli acupoints with parameters: 2/15 Hz dense-dispersive wave, 1 mA current. Results indicate that EA suppresses autophagy during reperfusion, alleviating CIRI (Mizushima and Levine, 2020; Qiu et al., 2025).

Conversely, emerging evidence suggests autophagy upregulation exerts neuroprotective effects. Xu et al. demonstrated that EA treatment enhances autophagic activity and mitigates CIRI. EA was administered at Du20 and Du26 acupoints at 5 min and 16 h post-injury, employing sparse (3.85 Hz for the 1.28-s duration) and dense (6.25 Hz for the 2.08-s duration) wave patterns. Stimulation parameters comprised a current intensity of 0.8–1.0 mA delivered over a 30-min treatment session (Xu et al., 2020). The precise mechanisms underlying EA-mediated autophagy regulation during injury repair require systematic investigation. These divergent outcomes may stem from intercellular variations in baseline autophagy levels across experimental models. Specifically, in cellular populations with diminished autophagic activity, EA potentially enhances autophagic flux, facilitating the clearance of compromised organelles and subsequent mitigation of injury. Conversely, in cellular systems exhibiting elevated basal autophagic activity, EA may exert inhibitory regulation on autophagic flux, thereby attenuating pathological consequences associated with excessive self-digestive processes.

Cellular autophagy is modulated through multiple signaling pathways, with extensive research elucidating key regulatory mechanisms. However, the precise pathway by which EA exerts its autophagic regulation remains undetermined, attributable to limited mechanistic evidence. Current investigations primarily focus on two pathways: mTOR-dependent and mTOR-independent signaling cascades. The mTOR-independent cascade, canonically designated as the Beclin1/B-cell lymphoma 2 (Bcl-2)/c-Jun N-terminal kinase 1 (JNK1) signaling axis, operates through coordinated regulation by autophagy-related genes (ATG) (Ahsan et al., 2021).

EA ameliorates CIRI through mTOR pathway upregulation, modulating autophagy activity (Wang M. et al., 2020). mTOR functions through two distinct complexes: mTOR complex 1 (mTORC1) and mTORC2. mTORC1 primarily suppresses autophagic activity, whereas this regulatory effect operates through the class I phosphatidylinositol 3-kinase (PI3K-I)/protein kinase B (AKT) signaling axis. Class 1 PI3K initiates this signaling cascade by phosphorylating plasma membrane phospholipids, thereby activating AKT. Activated AKT subsequently phosphorylates tuberous sclerosis complex 2 (TSC2), disrupting the TSC1-TSC2 interaction and ultimately relieving mTORC1 from inhibitory regulation (Ahsan et al., 2021; Tian et al., 2022).

EA modulates autophagy through dual regulatory mechanisms: controlling autophagosome generation and upstream ATGs. Experimental evidence indicates that EA pretreatment suppresses LC3 accumulation while elevating p62 expression, concurrent with significant structural restoration of neurons in the ischemic cerebral cortex of rat models. The regulation of autophagosomal components constitutes a critical control node in autophagy modulation (Mei et al., 2020; Figure 2). Experimental evidence reveals that silent information regulator 1 modulates autophagic activity via deacetylation mechanisms targeting both histones and non-histone proteins, including FOXO1 and p53. Under cerebral ischemia-reperfusion-induced oxidative stress, cytoplasmic acetylated FOXO1 accumulates and interacts with ATG7, a core autophagy executor, thereby activating autophagic flux. EA intervention effectively reverses this molecular interaction cascade (Mei et al., 2020; Figure 2).

Mitophagy serves as a critical neuroprotective mechanism in EA-mediated mitigation of CIRI. Ischemia-hypoxia during injury, disrupts mitochondrial electron transport chain function, impairing oxygen/glucose supply and adenosine triphosphate (ATP) synthesis. This oxygen delivery-consumption mismatch compromises oxidative phosphorylation, amplifying neuronal degeneration. Prolonged hypoxia depletes ATP, disrupting energy-dependent ion channel function and triggering sequential pathological events: potassium efflux, membrane depolarization, and voltage-gated calcium channel activation. Subsequent intracellular Ca²⁺ overload induces mitochondrial edema, which initiates pro-apoptotic factor release and structural disintegration of mitochondria. During reperfusion, electrons escaping from the impaired respiratory chain react with molecular oxygen, generating ROS that potentiate mitochondrial damage through oxidative stress cascades. ROS mediate dual activation of AMPK and NF-κB signaling pathways, driving proinflammatory cytokine secretion while amplifying oxidative damage (Li and Xing, 2023). Autophagy can also aggravate oxidative stress injury, and the combination of these reactions worsens cellular damage. Experimental data reveal that 20 Hz compressional wave EA at Baihui (GV20) and Zusanli (ST36) acupoints for 30 min activates PTEN-induced putative kinase 1 (PINK1)/Parkin-dependent mitophagy, effectively counteracting nitro-oxidative stress-mediated mitochondrial dysfunction in CIRI (Wang et al., 2019; Figure 2). Pink1, a central regulator of mitophagy, undergoes activation in response to mitochondrial dysfunction, subsequently recruiting Parkin to orchestrate mitochondrial clearance through selective autophagy (Ji et al., 2022). The pathophysiological triad of oxidative stress, inflammatory cascades, and autophagic dysregulation exhibits reciprocal potentiation, forming a self-amplifying circuit that escalates cytopathological damage. Strategic therapeutic intervention targeting this interconnected network demonstrates enhanced potential for achieving multidimensional neuroprotection in ischemia-reperfusion pathologies. Dysfunctional mitochondria constitute primary generators of ROS in oxidative stress. Controlled mitophagy exerts cytoprotective effects by selectively eliminating compromised mitochondria and inhibiting ROS overproduction, effectively attenuating pathological oxidative stress propagation and downstream inflammatory cascades.

The precise mechanisms underlying EA dual modulation of autophagic activity and mitigation of autophagy-mediated cytopathology remain systematically uncharacterized. Current evidence suggests these regulatory processes exhibit inherent complexity and multifactorial interactions. Deciphering such mechanisms holds significant potential to enhance the efficacy of EA-based therapeutic strategies for CIRI. Targeted investigations are required to map the molecular circuitry governing autophagic regulation and its pathophysiological crosstalk in neural injury contexts.

Regulation of autophagy by EA is one of the main targets to improve CIRI, mainly by regulating the expression of autophagy-related proteins, and autophagy-related genes to achieve the prevention of damage produced by excessive autophagy in neuronal cells, but part of autophagy, such as mitophagy, is up-regulated by EA, and it can prevent the occurrence of oxidative stress.

Oxidative stress manifests as a pathophysiological imbalance wherein pro-oxidant forces override antioxidant defenses, disrupting cellular redox equilibrium. In biological systems, endogenous antioxidant enzymes constitute the primary molecular machinery responsible for counteracting oxidative damage and maintaining metabolic homeostasis. Antioxidant defense mechanisms encompass multiple molecular regulators and signaling pathways, notably the nuclear factor E2-related factor 2, Kelch-like ECH-associated protein 1, PI3K/AKT cascade, mitogen-activated protein kinase, NF-κB inflammatory pathway, and heme oxygenase-1 enzymatic system, which collectively maintain redox homeostasis through coordinated transcriptional regulation and stress response modulation (Forman and Zhang, 2021; Wang et al., 2022b).

Oxidative stress plays a central role in CIRI; its pathological dominance constitutes a primary therapeutic target (Huang G. et al., 2022). CIRI induces redox imbalance through two principal mechanisms: ROS overproduction and compromised antioxidant enzymatic defenses. Pathological suppression of the Wnt/β-catenin pathway exemplifies this process, as it triggers mitochondrial complex I dysfunction and ROS overgeneration. This propagates mitochondrial destruction, culminating in programed neuronal death (Zhang et al., 2024).

EA demonstrates significant efficacy in ameliorating oxidative stress injury following cerebral ischemia-reperfusion (Tang et al., 2024). Experimental studies reveal that EA preconditioning attenuates redox imbalance through coordinated suppression of transient receptor potential vanilloid 1 activation, downregulation of cytochrome C, and biomarker regulation of MDA, glutathione (GSH), and SOD levels (Figure 3; Long et al., 2019). Cheng et al. demonstrated that MAPKs serve as central regulators of intracellular oxidative stress post-cerebral ischemia. Mechanistically, p38 MAPK activation upregulates survival proteins Bcl-2 and B-cell lymphoma-extra-large (Bcl-xL) through cAMP-responsive element-binding protein-dependent transcriptional regulation, while concurrently inducing mitochondrial permeability transition pore opening - a critical determinant of mitochondrial apoptotic signaling. Mitochondrial permeability transition pore opening enables X-linked inhibitors of apoptosis protein family members to bind caspases, directly modulating apoptotic signaling pathways (Figure 3). Experimental data further indicate that frequency-specific EA elevates the Bcl-2/Bcl-2-associated X protein (Bax) and Bcl-xL/Bax ratios, stabilizing mitochondrial membrane integrity post-CIRI through anti-apoptotic protein dominance, thereby conferring neuroprotective efficacy (Cheng et al., 2015). EA exerts dual regulatory effects on astrocytes and microglia through targeted molecular interventions. Specifically, EA suppresses astrocytic S100B secretion — a critical mediator that upregulates inducible nitric oxide synthase (iNOS) expression and drives pathological nitric oxide (NO) overproduction in glial cells at elevated concentrations, ultimately triggering neurotoxic NO diffusion. This coordinated modulation attenuates nitro-oxidative stress propagation and suppresses proinflammatory cytokine release, demonstrating EA’s capacity to disrupt the S100B/iNOS/NO neurotoxic axis while preserving neural homeostasis (Cheng et al., 2014).

While the precise mechanisms underlying EA regulation of redox homeostasis require further elucidation, its capacity to mitigate CIRI through coordinated ROS suppression and enzymatic antioxidant system modulation is well-established. Current evidence reveals substantial pathway convergence among EA’s antioxidant, anti-inflammatory, and anti-apoptotic effects. Critical nodal regulators like the MAPK signaling cascade demonstrate multifunctional therapeutic potential, simultaneously addressing oxidative stress propagation, neuroinflammation amplification, and programed cell death initiation post-CIRI. This mechanistic overlap highlights this regulator as a central role in EA-mediated neural protection, positioning it as a prime translational target for developing multimodal interventions against ischemia-reperfusion pathologies.

MAPK and cytochrome C are relatively important intracellular targets in recent studies, acting by affecting the production of oxidative substances and regulating oxidase production, while in glial cells EA also reduces the production of substances that contribute to oxidative stress, such as S100B.

In CIRI, inflammation is initiated through endogenous signaling, wherein damaged tissues release molecular mediators that activate innate immune responses. Plasma membrane rupture in necrotic cells triggers the extracellular release of substances, including ATP, potassium ions, uric acid, high-mobility group box 1 protein, and S100 calcium-binding proteins. The release of these substances activates multiple inflammasome complexes, including NLRP3 and NALP3, while simultaneously triggering Toll-like receptor-mediated caspase-1 activation. This signaling mechanism propagates inflammatory cascades and induces pyroptotic cell death. ROS, central to oxidative stress, further amplifies inflammation through mechanisms such as oxidized lipoprotein-induced inflammatory signaling (Jayaraj et al., 2019). Key inflammatory mediators including tumor necrosis factor-α (TNF-α), NF-κB, and interleukin-1β (IL-1β) critically regulate inflammatory cascade activation in CIRI (Cao et al., 2021).

EA significantly inhibits inflammatory responses in CIRI. Experimental studies employing focal cerebral ischemia models in male Sprague-Dawley rats demonstrated that EA suppresses NLRP3 inflammasome activation in penumbral tissues post-reperfusion, as quantified by Western blot and immunofluorescence analyses. Mechanistically, EA activates α7 nicotinic acetylcholine receptor signaling to block NLRP3 inflammasome assembly, thereby attenuating neuroinflammation (Jiang et al., 2019). EA directly inhibits the production of major inflammatory factors, including TNF-α, IL-6, IL-1β, and Ccl-2 mRNA, exerting an anti-inflammatory effect within the cell (Ren et al., 2024). It also modulates immune cells such as Tregs and γδ T-cells, which suppress other T-cells, B-cells, and pro-inflammatory factors through this pathway, thereby reducing inflammation (Wang Y. et al., 2023).

Recent studies suggest that EA is closely associated with both anti-apoptotic and anti-inflammatory mechanisms, involving pathways that simultaneously downregulate inflammatory factors and exert inflammatory-trophic effects. For example, in a brain ischemia-reperfusion model, EA treatment was evaluated using cell fluorescent dye and real-time quantitative polymerase chain reaction to assess mRNA levels. The expression of brain-derived neurotrophic factor (BDNF) was measured via western blot or enzyme-linked immunosorbent assay in vitro, revealing that EA induced delta-opioid receptor activation and improved cell morphology.

It reduced cell apoptosis and downregulated proinflammatory cytokine expression. Additionally, the delta-opioid receptor inhibited inflammation through the BDNF/tropomyosin-related kinase B pathway (Geng et al., 2020). BDNF has anti-inflammatory effects and can inhibit neuronal apoptosis (Gu W. et al., 2022). However, the specific anti-inflammatory mechanism of EA remains insufficiently studied. Future research clarifying its mechanism of action will enhance the clinical application of this traditional Chinese medicine-based treatment, enabling more targeted therapeutic effects.

EA can play a protective role against CIRI and ameliorate it through the inhibition of neuronal ferroptosis. Some studies (Liang et al., 2022; Yao et al., 2021) demonstrate that ferroptosis, an iron-dependent form of programed cell death caused by lipid peroxide accumulation, plays a critical role in the development of neurological diseases. This process is closely related to intracellular GSH deficiency and ROS production. Glutathione peroxidase 4 (GPX4) is a critical inhibitor of ferroptosis, and the GSH/GPX4 system reduces lipid hydroperoxides to 12/15-lipoxygenase (12/15-LOX) under oxidative stress (Liu S. et al., 2023), activating the transferrin receptor (TFRC) to enhance ferrous ion (Fe²⁺) uptake (Lu et al., 2021). GSH biosynthesis is critical for the functional activity of GPX4. When GSH levels are depleted, GPX4 becomes inactivated, leading to ROS accumulation and lipid peroxidation, which at a certain threshold triggers cellular ferroptosis (Ou et al., 2022). Biochemically, the Fenton reaction between excess free Fe²⁺ and hydrogen peroxide, a product of mitochondrial oxidative respiration, triggers a significant increase in ROS, which inhibits GSH synthesis and GPX4 activity. This leads to impaired antioxidant systems, redox imbalances, and a buildup of toxic lipid peroxides, ultimately contributing to the ferroptosis of cells (Jiang et al., 2021; Lei et al., 2019; Li et al., 2020; Xie et al., 2022). The approximate process of ferroptosis described above can be seen in Figure 4.

Li et al. demonstrated that EA has a ferroptosis-inhibitory effect in MCAO rats (Li et al., 2021). The study showed that EA enhances GPX4 and SOD activities, as well as GSH levels while decreasing MDA and iron accumulation. EA was also found to increase the levels of FTH1 and decrease the levels of Tf (transferrin) and TfR1. These results suggest that EA alleviates ICS (ischemic stroke) by inhibiting ferroptosis. Moreover, Wang G. et al. (2023) found that EA significantly reduced the expression of ACSL4 and TFR1, increased GPX4 levels, and inhibited ROS production in MCAO/R model rats. Furthermore, EA treatment improved neurological deficits and reduced infarct volume. These findings suggest that EA may alleviate CIRI by inhibiting ferroptosis. The results from these animal studies indicate that EA holds the potential for mitigating CIRI through ferroptosis inhibition.

GSH, GPX4, and ROS are keys to the occurrence of ferroptosis in neurons. EA can inhibit neuronal ferroptosis after cerebral ischemia/reperfusion and attenuate cerebral ischemia/reperfusion injury by decreasing the content of GSH and ROS and downregulating the expression of GPX4.

EA can exert anti-neuronal apoptosis by regulating the balance of apoptosis-related proteins, thereby demonstrating therapeutic potential in improving CIRI (Ma et al., 2016).

Firstly, EA treatment can prevent neuronal apoptosis by regulating the balance of Bcl-2 and Bax proteins. Bcl-2 and Bax are key proteins in the Bcl-2 family, closely associated with apoptosis. Bcl-2 is a membrane-stabilizing protein that is associated with organelle membranes. By inhibiting free radical production and intracellular calcium overload, this treatment prevents mitochondrial membrane permeability, maintains membrane stability, and blocks cytochrome C release and caspase activation, thereby inhibiting apoptosis. Bcl-2 overexpression reduces cellular damage, while Bax exhibits an antagonistic role, promoting apoptosis. Bcl-2 and Bax can form heterodimers. When Bcl-2 expression is increased, Bcl-2 homodimers are formed, which protect cells. Conversely, when Bax expression is increased, Bax/Bcl-2 heterodimers form, leading to apoptosis (Alam et al., 2022; Moldoveanu and Czabotar, 2020; Wolf et al., 2022). Therefore, the Bcl-2/Bax ratio plays a significant role in determining apoptosis in neurons and other cells. It has been shown that in the MCAO-induced cerebral ischemia/reperfusion rat model, EA treatment significantly enhanced the Bcl-2/Bax ratio, reduced the number of apoptotic cells, and alleviated the degree of neurological impairment. Additionally, EA improved cognitive deficits in ischemia/reperfusion-injured rats by regulating Bcl-2 and Bax expression (Liu et al., 2015). This finding is consistent with the experimental study by Ma et al. (2016), which demonstrated that EA stimulation in MCAO rats up-regulated Bcl-2 protein expression, down-regulated Bax protein expression and significantly reduced the number of apoptotic cells. These results suggest that EA may mitigate neuronal apoptosis following CIRI by modulating the balance between Bax and Bcl-2 proteins.

Secondly, EA can also reduce CIRI by regulating caspase-3 expression, an apoptotic execution factor that exists in normal cells as a zymogen. During cerebral ischemia, caspase-3 activation triggers a series of events, including nuclear factor activation, cytoskeletal protein disruption, and DNA cleavage. These processes lead to cell morphological changes, apoptotic vesicle formation, and ultimately apoptosis (Sahoo et al., 2023; Unnisa et al., 2023; Vatandoust et al., 2023). Meng et al. (2024) demonstrated that EA treatment in MCAO rats significantly upregulated the mRNA and protein expression of the Bcl-2 gene in the ischemic brain, thereby downregulating caspase-3, Bax, and cleaved caspase-3. This led to a reduction in apoptosis of the ischemic brain tissue in stroke rats and attenuation of CIRI in MCAO rats. EA treatment enhances neuroprotection by promoting the expression of protective proteins and inhibiting pro-apoptotic factors, resulting in improved outcomes for cerebral ischemia-reperfusion injury.

In addition, studies have demonstrated that EA may inhibit neuronal apoptosis by activating the PI3K/Akt pathway, thereby protecting brain tissue from CIRI (Rao et al., 2025; Wang et al., 2021). Additionally, EA enhances BDNF expression, further exerting anti-apoptotic, neuroprotective effects in CIRI (Geng et al., 2020; Li et al., 2022). These mechanisms are summarized in Figure 5.

Overexpression of Bcl-2 attenuates cell damage, whereas Bax acts in the opposite direction to Bcl-2 and has a pro-apoptotic effect. EA intervention in MCAO rats up-regulated Bcl-2 protein expression and down-regulated Bax protein expression, which significantly reduced the number of apoptotic cells and exerted neuroprotective effects. caspase-3 is an apoptotic execution factor that triggers cell apoptosis. EA can inhibit apoptosis in the ischemic brain tissue of stroke rats by reducing the expression of caspase-3. EA can also activate the PI3K/Akt pathway, causing a series of responses to inhibit neuronal apoptosis after CIRI.

CIRI leads to excitotoxic injury, primarily due to the excessive release of excitatory amino acids. This process causes neuronal apoptosis and neurotoxic injury, characterized by increased glutamate release. Glutamate (Glu), an excitatory neurotransmitter, activates NMDA receptors (NMDARs). The excessive release of glutamate into the synaptic gap triggers increased calcium influx through NMDARs (NMDA receptors), which is central to the toxic damage (Liao et al., 2023).

Under pathological conditions such as cerebral ischemia, glutamate is excessively released into the synaptic gap, which then activates the NR2B-containing NMDAR (GluN2B), leading to intracellular calcium overload. The overloaded calcium activates cytoplasmic m-calpain, which triggers the enzymatic activity of STEP61, further inhibits p38 MAPK dephosphorylation, and ultimately leads to apoptosis through signal transduction (Huang W. et al., 2022; Jang et al., 2014). This process is illustrated in Figure 6. Therefore, inhibition of glutamate over-release is very important for mitigating CIRI. It has been reported in the literature (Zhang et al., 2018) that EA administration in MCAO rat models increased NMDAR2A protein expression in the hippocampus compared to the MCAO group. It also reduced Glu and NMDAR2B protein levels and decreased intracellular Ca²⁺ levels in the MCAO rats. This helped protect the hippocampus and attenuate the CIRI caused by MCAO. The mechanism behind this effect includes the reduction of NMDAR2B expression and the decrease of Ca²⁺-related neurotoxicity by inhibiting Glu release. Zhang et al. (2020) demonstrated that EA pretreatment down-regulated the GluN2B expression, reduced m-calpain activity, and inhibited p38 MAPK phosphorylation in hippocampal CA1 neurons. These findings suggest that EA suppresses the GluN2B/m-calpain/p38 MAPK pathway, thereby reducing brain damage following CIRI. The proapoptotic pathway of GluN2B/m-calpain/p38 MAPK effectively reduced the volume of cerebral infarction and the level of neuronal apoptosis after CIRI.

Following CIRI, EA reduced excitotoxicity brought on by calcium overload in brain tissues and prevented glutamate release. Pretreatment with EA may be useful in lowering the amount of cerebral infarction and the degree of neuronal death during CIRI by blocking the GluN2B/m-calpain/p38 MAPK pro-apoptotic pathway.

In addition, NMDARs exhibit two subtypes, NMDAR2A and NMDAR2B, which have opposing roles. NMDAR2A activation promotes neuronal regeneration and protects neuronal function, while NMDAR2B mediates the free radical chain reaction, inducing neurotoxicity and apoptosis (Goldsmith, 2019; Skowrońska et al., 2019). Zhang et al. showed that EA enhanced NMDAR2A expression and down-regulated NMDAR2B expression, thereby alleviating CIRI. This modulation significantly influenced cognitive recovery post-stroke in a rat model (Zhang et al., 2018).

Previous studies have shown that while reducing glutamate release or enhancing glutamate reuptake are effective strategies to counter ischemic injury, most drugs in this category are inefficient in reducing morbidity or mortality and exhibit neurological side effects in stroke clinical trials. Targeted therapies and interventions based on the “NMDAR” subtype hypothesis have also shown limited efficacy and significant side effects in preclinical studies (Mao et al., 2022). Therefore, the current study on EA for CIRI highlights a potential mechanism to address excitotoxic injury, which could address the limitations of existing ischemic stroke therapies. Moreover, EA intervention is free of drug toxicity and has minimal neurological side effects, offering promise for broader clinical application in the foreseeable future.

EA intervention may also ameliorate CIRI by protecting the blood-brain barrier. Cerebral ischemia/reperfusion can cause and aggravate cerebral edema by increasing the expression of Aquaporin-4 (AQP4), a water channel protein closely associated with cerebral edema formation (Vandebroek and Yasui, 2020). Taniguchi et al. (2000) found that after focal cerebral ischemia in rats, cerebral edema worsened over time and the expression of AQP4 protein and mRNA was in line with the time of aggravation of cerebral edema. The changes in AQP4 mRNA expression in MCAO were closely related to the changes in edema formation and regression. Xu et al. (2014) demonstrated that EA significantly attenuated the expression of water channel protein AQP4 in the ischemic brain, effectively alleviating cerebral edema symptoms in rats following cerebral ischemia and reperfusion. Studies have also shown that AQP4 is closely related to the BBB (Mueller et al., 2023). Neural astrocyte peduncles contribute to the BBB’s structure by forming a glial border membrane around capillaries, acting as the BBB’s second barrier. This structure aids in regulating water transport, with AQP4 concentrated in the peduncles of glial cells. Under cerebral ischemia, AQP4 expression on astrocyte membranes is enhanced. AQP4 channels open abundantly, leading to cerebral edema formation and progression. Therefore, EA downregulates AQP4 expression, thereby reducing astrocyte peduncle swelling and restricting AQP4 activity. This leads to BBB protection, cerebral edema alleviation, and improved outcomes in CIRI.

Additionally, EA may improve CIRI by down-regulating Matrix metalloproteinase 2 (MMP2)/AQP4/APQ9 expression. MMPs belong to the family of Zn²⁺-dependent enzymes associated with BBB permeability and disruption (Lu and Wen, 2025; Sun et al., 2023), and MMP2 plays an important role in brain water transport homeostasis, as do AQP4 and AQP9. All three are associated with the permeability of the BBB in CIRI. BBB disruption, tissue inflammation, and MMP/AQP up-regulation together trigger cerebral edema/swelling after CIRI. EA mitigates this condition by regulating MMP2/AQP4/AQP9 activity, thereby exerting a neuroprotective effect (Xu et al., 2014).

MMP-9, a collagenase in the MMP family, degrades the extracellular matrix and contributes to BBB disruption during CIRI (Gu Y. et al., 2022). In contrast, some studies have shown that MMP-9 expression in the brain of MCAO rats pretreated with EA was significantly reduced, suggesting that EA pretreatment may maintain the integrity of the blood-brain barrier after injury by inhibiting MMP-9 expression, reduce the brain tissue damage, and improve the cerebral function (Lin et al., 2016).

Following CIRI, the microvasculature of the cerebral cortex is damaged, resulting in neuronal death in the ischemic central zone. Neurons in the ischemic hemiparetic zone are subjected to ischemia and hypoxia, leading to neuronal damage and functional impairments (Wang et al., 2018). In this case, angiogenesis in ischemic areas is crucial for reducing the damage and improving prognosis. Angiogenesis not only increases collateral circulation to reduce ischemia-induced brain damage but also plays a key role in restoring neurological function after ischemic stroke (Shi et al., 2017). Angiogenesis is regulated by several vascular growth factors, including vascular endothelial growth factor (VEGF) and erythropoietin (EPO). Previous studies have shown that VEGF is a key cytokine associated with the differentiation, proliferation, and migration of endothelial cells. By binding to its receptor (VEGFR), VEGF activates downstream signaling pathways and transmits angiogenic signals (Wang H. et al., 2020). Therefore, promoting the expression of VEGF and VEGFR is important in angiogenesis after brain injury. In this regard, studies have shown that EA can promote neovascularization after focal cerebral ischemia/reperfusion by up-regulating VEGF expression, reducing the damage of the vascular wall induced by ischemia/reperfusion, and exerting a protective effect against ischemia/reperfusion-induced brain injury (Zhang W. et al., 2023). EPO exhibits multiple functions, including erythropoiesis promotion, anti-inflammatory, antioxidant, pro-angiogenic, and neurotrophic effects (Alural et al., 2014; Si et al., 2019). EPO activates its function by binding to the EPO receptor (EPOR). Importantly, both EPO and EPOR are expressed in red lineage cells, neurons, astrocytes, and cerebrovascular endothelial cells (Tsiftsoglou, 2021). The non-receptor tyrosine kinase Src plays an important role in angiogenesis, and increased levels of Src phosphorylation have a protective effect against ischemia-induced brain damage (Sun et al., 2021). Wang et al. (2022a) showed that EA increased the expression of EPO, VEGF and P-Src proteins. This effect may involve EPO-mediated activation of the Src signaling pathway and VEGF signaling pathway, thereby promoting angiogenesis. Notably, EPO activates the Src pathway, which further increases VEGF expression (Figure 7), amplifying pro-angiogenic effects. This process facilitates vascular neogenesis, enhancing blood flow and oxygen supply in ischemic brain areas, and alleviates CIRI.

In CIRI rat brain tissue, EA stimulates the VEGF and Src signaling pathways, which are mediated by EPO, to encourage angiogenesis. Moreover, activation of the Src signaling pathway by EPO contributes to the increase in VEGF expression, which further enhances the pro-angiogenic effect and facilitates neovascularization, thereby attenuating CIRI.

Moreover, EA may also promote the mobilization of endothelial progenitor cells (EPCs) to ischemic areas and promote vascular neogenesis. EPCs are a group of naïve endothelial cells with wandering characteristics and capable of further proliferation and directional differentiation. Their main function is to participate in vascular endothelial repair and neointimal formation in ischemic organs (Kong et al., 2018; Shi et al., 2023). After the ischemic injury, EPCs in bone marrow may mobilize, enter the peripheral circulation, and migrate to the ischemic site to participate in vascular endothelialisation and angiogenesis. EA promotes the mobilization of EPCs in the bone marrow and their migration in the peripheral blood, which facilitates angiogenesis after cerebral ischemia/reperfusion and improves injury symptoms (Lu et al., 2013; Xie et al., 2016).

EA’s key advantage in CIRI treatment lies in its ability to stimulate cerebral angiogenesis at injury sites. Unlike existing ischemic stroke treatments, fewer drugs target cerebral angiogenesis. EA offers a promising new therapeutic approach for the clinical treatment of ischemic stroke.