Mechanisms of Acupuncture on the Neurovascular Unit in Migraine: A Narrative Review

ABSTRACT

Objective

The neurovascular unit (NVU) is a core structural and functional entity involved in migraine pathogenesis. Acupuncture may exert therapeutic effects by protecting NVU integrity. This narrative review aims to summarize the current evidence on the mechanisms by which acupuncture protects the NVU in migraine.

Methods

We reviewed and synthesized relevant literature focusing on the effects of acupuncture on neuronal excitation, blood–brain barrier (BBB) components, neuroinflammation, and energy metabolism in the context of migraine.

Results

Acupuncture protects the NVU through four major pathways: (1) Acupuncture inhibits pathological excitation of neurons, including activating the endogenous pain regulation system, regulating the balance of neurotransmitters and receptors, and inhibiting peripheral sensitization and inflammation; (2) Acupuncture maintains the integrity of the BBB, specifically involving improving endothelial cell function and tight junctions (TJs), regulating astrocyte reactivity, enhancing pericyte function, and inhibiting extracellular matrix (ECM) degradation; (3) Acupuncture regulates neuroinflammation mediated by microglia; (4) Acupuncture improves energy metabolism and oxidative stress.

Conclusions

Acupuncture exerts multi‐target, comprehensive protective effects on NVU damage in migraine. These findings provide a mechanistic basis for its clinical application and may offer new therapeutic directions for migraine management.

Abbreviations

5‐HT

5‐hydroxytryptamine

A1AR

adenosine A1 receptors

ACC

anterior cingulate cortex

AD

alzheimer's disease

AJs

adherens junctions

ATP

adenosine triphosphate

BBB

blood–brain barrier

BMECs

brain microvascular endothelial cells

CB1

cannabinoid receptor type 1

CBF

cerebral blood flow

CCI

chronic constriction injury

CGRP

calcitonin gene‐related peptide

CNS

central nervous system

CSD

cortical spreading depression

Cx43

connexin 43

DCE‐MRI

dynamic contrast‐enhanced magnetic resonance imaging

DES

dural electrical stimulation

dMPFC

dorsomedial prefrontal cortex

DRG

dorsal root ganglion

EA

electroacupuncture

EC

endothelial cells

ECM

extracellular matrix

ECS

endocannabinoid system

EPMS

Endogenous Pain Modulatory System

GABA

γ‐aminobutyric acid

IFN‐γR

Interferon‐γ receptor

JAMs

junctional adhesion molecules

MA

manual acupuncture

mGluRs

metabotropic glutamate receptors

MMP‐9

matrix metalloproteinase‐9

MMPs

matrix metalloproteinases

MOH

medication overuse headache

NE

norepinephrine

NF‐κB

nuclear factor κB

NMDA

N‐methyl‐D‐aspartic acid receptor

NO

Nitric Oxide

NSAIDs

nonsteroidal anti‐inflammatory drugs

NT‐3

Neurotrophic factor‐3

NTG

nitroglycerin

NVC

neurovascular coupling

NVU

neurovascular unit

PAG

periaqueductal gray

PAR‐1

protease‐activated receptor‐1

PCC

posterior cingulate cortex

PVN

paraventricular nucleus

rACC/mPFC

anterior cingulate cortex/medial prefrontal cortex

RNA

Ribonucleic Acid

ROS

reactive oxygen species

ROS/RNS

reactive oxygen/nitrogen species

rs‐fMRI

resting‐state functional MRI

RVM

rostral ventromedial nucleus

SGCs

satellite glial cells

SNL

sciatic nerve ligation

TCC

trigeminal cervical complex

TGVS

trigeminal vascular system

TJs

tight junctions

TLR4

toll‐like receptor 4

TLRs

Toll‐like receptors

TNC

trigeminal nucleus caudalis

TRPA1

Transient Receptor Potential Ankyrin 1

TRPV1

transient receptor potential vanilloid subtype 1

VEGF

vascular endothelial growth factor

vlPAG

ventrolateral periaqueductal gray

WHO

World Health Organization

1. Introduction

Migraine is a common neurological disorder characterized by recurrent episodes of unilateral or bilateral throbbing headache of moderate to severe intensity, often accompanied by symptoms of nausea, vomiting, and phonophobia. Daily activities may exacerbate the pain [1]. According to data released by the World Health Organization (WHO), migraine ranks high among the most common human diseases in terms of disability burden [2]. Migraine, especially migraine with aura, can increase the risk of cerebrovascular diseases such as stroke by 27%, posing a risk to human health [3]. Furthermore, the distribution of this disease exhibits significant demographic differences. Particularly noteworthy is that the prevalence rate among women is approximately three times that of men, with peak incidence occurring during adolescence and among individuals aged 30–39 [4]. Currently, nonsteroidal anti‐inflammatory drugs (NSAIDs) or triptans are commonly used for acute migraine treatment. However, these medications carry risks of medication overuse headache (MOH), and triptans have certain cardiovascular contraindications, limiting their use in some patients [5]. For preventive treatment, traditional medications such as beta‐blockers, antiepileptic drugs, or antidepressants are commonly used. However, these drugs carry numerous side effects, including fatigue, weight changes, and cognitive impairment, leading to poor patient tolerance and compliance [6]. In recent years, calcitonin gene‐related peptide (CGRP)‐targeting agents—including CGRP monoclonal antibodies and small‐molecule gepants, can prevent the onset of migraines from the source by antagonizing the neurogenic inflammation and vasodilation mediated by CGRP [7]. Therefore, such drugs have become a new option for preventive treatment of migraines. Recently, extensive real‐world studies have largely confirmed the long‐term effectiveness, high treatment adherence, and favorable safety profiles of CGRP monoclonal antibodies in routine clinical practice [8, 9]. However, these agents primarily act on a single vasoactive peptide located downstream of the trigeminal vascular system (TGVS), and their high cost remains a limiting factor for broad accessibility. In recent years, the NVU theory has attracted more and more attention in the study of the mechanism of cerebrovascular diseases. The concept of the NVU was formally established in 2001 by the National Institute of Neurological Disorders and Stroke at the Stroke Progress Review Panel meeting. It emphasizes that the brain and blood vessels do not function independently but interact cooperatively to dynamically regulate neurovascular coupling (NVC), maintain BBB integrity, and preserve homeostasis within the central nervous system (CNS) [10]. The pathophysiological process of migraine actually involves complex interactions among neurons, glial cells, vascular components and the BBB. These structures collectively form the NVU. The dysfunction of the NVU not only includes CGRP‐mediated vasodilation, but also involves multiple mechanisms such as cortical spreading depression (CSD), neuroinflammation, oxidative stress, and NVC disorders. Therefore, therapies targeting a single molecule such as CGRP may not fully address the multifactorial pathological processes involved in NVU dysfunction.

Acupuncture demonstrates unique advantages in treating NVC disorders, characterized by its simplicity, low cost, and minimal side effects, making it an alternative to pharmacological therapies [11]. Acupuncture demonstrates remarkable efficacy in migraine patients, but its precise mechanisms remain incompletely understood [12]. Elucidating the regulatory effects of acupuncture on the NVU may provide a novel, multi‐target approach that goes beyond conventional single‐agent strategies. This review will analyze the mechanisms by which acupuncture intervenes in migraine treatment from an NVU perspective, based on both clinical and basic experimental research, aiming to provide insights for more effective migraine management in the future.

2. Method

We conducted a search on PubMed and Web of Science for articles published between 2000 and April 2026. The search terms included combinations such as “migraine,” “hemicrania,” “sick headache,” “acupuncture,” “needling,” “stab,” “electroacupuncture,” “migraine,” “neurovascular unit,” “blood–brain barrier,” “microglia,” “astrocyte,” “neuron,” “pericyte,” “extracellular matrix,” “oxidative stress,” and “energy metabolism”. The eligible studies included original basic research and clinical trials on the effects of acupuncture on each component of the NVU. Additionally, we manually screened the references of the retrieved literature to supplement additional relevant studies. The extracted data were narratively organized and summarized according to the components of the NVU and the mechanisms.

3. Pathophysiological Mechanisms of Neurovascular Unit in Migraine

The NVU is a complex multicellular structure composed of neurons, astrocytes, vascular endothelial cells (EC), pericytes, microglia, and ECM [13]. It fundamentally maintains homeostasis in the brain microenvironment through the coupling of neurons with the vascular system [14]. Following increased neuronal activity, the NVU promptly enhances local blood flow supply to meet energy demands while forming a selective barrier to protect brain parenchyma [15].

Therefore, the functional status of each component within the NVU and the overall microenvironmental homeostasis are closely related to cerebral microcirculation status [16]. Current mechanistic hypotheses regarding migraine predominantly revolve around the cortical spreading depression theory and the trigeminal vascular theory [17]. Migraine is classified into two common subtypes based on the presence or absence of aura prior to an attack: migraine without aura and migraine with aura [18]. It is generally recognized that common triggers for migraine include endogenous and exogenous factors such as emotions, diet, stress, and genetics [19]. During the onset of migraine without aura, these factors influence hypothalamic function [20], subsequently regulating the brainstem and driving activation of the TGVS. This promotes the release of vasoactive peptides, including CGRP and substance P [21]. CGRP, in particular, is the most potent vasodilator known [22] and plays a crucial role in the pathophysiology of migraine [23]. Under the influence of these potent vasodilators, BBB permeability increases plasma fibrinogen extravasation further, which further activates endothelial protease‐activated receptor‐1 (PAR‐1) [24], inducing secretion of matrix metalloproteinase‐9 (MMP‐9) [25], which degrades type IV collagen in the basement membrane and amplifies neurogenic inflammation. Inflammatory mediators (e.g., endothelin‐1) activate pericyte endothelin‐A receptors [26], triggering pathological constriction via the Rho kinase pathway, leading to capillary collapse and local ischemia [27]. At the same time, mitochondrial reactive oxygen species (ROS) accumulation inhibits adenosine triphosphate (ATP) synthesis [28] and weakens the ability of pericytes to regulate cerebral blood flow (CBF). K⁺ and glutamate that extravasate after BBB disruption activate neuronal N‐methyl‐D‐aspartate (NMDA) receptors, inducing Ca²⁺ overload and mitochondrial oxidative stress [29], ultimately triggering depolarization that initiates CSD. ATP released during CSD activates microglia [30] which in turn activate the nuclear factor κB (NF‐κB)‐mediated inflammatory signaling pathway [31]. Meanwhile, disrupted polarity of astrocytic terminal processes impedes K⁺ clearance, sustaining neuronal hyperexcitability [32]. This pathway upregulates cyclooxygenase‐2 expression, promotes prostaglandin synthesis, and induces and exacerbates migraine symptoms [33]. The “astrocyte‐microglia crosstalk” plays a central role in neuroinflammation [34]. These cells mutually activate each other by releasing multiple cytokines and signaling molecules, forming an inflammatory amplification loop. This crosstalk may be a key mechanism sustaining central sensitization and chronicity in migraine [35]. Furthermore, numerous neurotransmitter dysregulations contribute to migraine pathogenesis. For instance, desensitization of midbrain 5‐hydroxytryptamine (5‐HT) neurons leads to internalization of 5‐HT1B/1D receptors, thereby shutting down descending pain inhibitory pathways [36] and triggering migraine attacks.

Research on the mechanisms of migraine with aura reveals that its pathogenesis resembles that of migraine without aura. Both involve CGRP/P substance release leading to a significant increase in BBB permeability, activation of microglia, cytokine release, further disruption of the BBB, and tissue damage, ultimately amplifying the inflammatory cascade. However, the difference lies in the temporal sequence of the pathological trigger points: CSD versus trigeminal neurovascular activation [37]. In migraine with aura, CSD is generally considered dominant [38]. The inflammatory mediators, Nitric Oxide (NO), ATP, and other substances it generates can disrupt the NVU balance on the dura mater [39, 40], thereby strongly activating the trigeminal neurovascular system [41]. CSD is a slowly propagating wave of altered brain activity involving changes in neurons, glial cells, and vascular function [42]. During neuronal inhibition, CSD induces extracellular K⁺ elevation and massive glutamate release [43]. Excessive glutamate accumulation strongly activates metabotropic glutamate receptors (mGluRs) on astrocytes, triggering sustained intracellular Ca²⁺ oscillations that exacerbate excitotoxicity [43]. This causes pericyte contraction and activation of trigeminal nerve endings, thereby initiating the TGVS [44], promoting massive secretion of CGRP and substance P. This disrupts blood flow regulation while simultaneously inducing multiple pathological processes within the brain, including oxidative stress [45] and neuroinflammation [46]. Regardless of subtype, key components of the NVU mechanism—including neurons, astrocytes, microglia, vascular EC, and pericytes—play pivotal roles in inducing migraine onset (Figure 1).

FIGURE 1.

Pathophysiological mechanisms of neurovascular unit in migraine. Migraine with aura is triggered by CSD, while migraine without aura is associated with factors such as direct neurogenic inflammation. Both pathways ultimately converge on multifaceted NVU damage: A1/A2 representing the pathway for migraine without aura involving direct neurogenic inflammation, and B1/B2 showing CSD‐triggered migraine with aura.

Migraine arises from abnormal neuronal excitation triggered by endogenous and exogenous stimuli, leading to BBB disruption and cross‐talk between astrocytes and microglia. Over time, this induces central sensitization, forming a vicious cycle. Extensive research demonstrates that acupuncture significantly reduces attack frequency, duration, and pain intensity [47].

4. Protective and Regulatory Mechanisms of Acupuncture Against NVU Damage in Migraine

Acupuncture has gained increasing recognition as an effective adjunct therapy for migraine prevention, and its analgesic mechanisms have been partially elucidated within multiple theoretical frameworks. Traditional research has primarily focused on regulating central pain processing, such as activating endogenous opioid systems within descending pain modulation pathways and modulating key neurotransmitters like serotonin and glutamate. Additionally, peripheral mechanisms involving the TGVS have been extensively studied, demonstrating its capacity to inhibit the release of vasoactive neuropeptides like CGRP and substance P. However, these mechanisms are often interpreted from a singular “neuronal” or “vascular” perspective, failing to fully capture migraine's essence as a NVC disorder. In recent years, the NVU has emerged as an integrative concept, viewing neurons, blood vessels, the BBB, and glial cells as a dynamically coupled functional and structural entity. This framework offers a revolutionary perspective for understanding CNS disorders. In migraine, NVU homeostasis imbalance is considered a core pathophysiological mechanism, involving the initiation of CSD, sensitization of the TGVS, and maintenance of central sensitization (Figure 2).

FIGURE 2.

Acupuncture for migraine: An NVU perspective. Acupuncture therapy alleviates pain at the neural level by suppressing pathological neuronal excitability and modulating peripheral sensitization and endogenous pain regulation systems. It stabilizes perivascular cell function and vasoactivity while maintaining metabolic homeostasis in NVU by improving cerebral energy metabolism and oxidative stress status. And comprehensively maintains the structure and function of the BBB by regulating glial‐mediated neuroinflammation, improving astrocytic reactivity, and protecting endothelial cell function and tight junction integrity. These synergistic effects collectively preserve NVU stability, thereby exerting therapeutic effects against migraine.

The following sections will explore the specific mechanisms of acupuncture intervention for migraine from the NVU perspective, based on empirical research, by examining how acupuncture regulates interactions among NVU components, elevates neuronal thresholds, and maintains BBB integrity to protect NVU function.

4.1. Acupuncture Suppresses Pathological Neuronal Excitability

As the functional core of the NVU, pathological hyperexcitability and central sensitization of neurons constitute the initiating phase of migraine attacks [48]. Pathological neuronal excitation fundamentally involves enhanced synchronized neuronal discharge due to ion channel dysfunction and impaired inhibitory neural regulation [49]. Without timely suppression of such abnormal discharge, mitochondrial calcium overload and excitotoxicity irreversibly damage synaptic structures, ultimately progressing to central sensitization [50]. Therefore, timely suppression of neuronal excitation during migraine stress states [51] and maintenance of NVU homeostasis are critical. Studies indicate that acupuncture at the “Fengchi” point modulates neuronal discharge in the trigeminal cervical complex (TCC) to alleviate migraine attacks [52]. Interferon‐γ receptor (IFN‐γR) is highly expressed on neurons, enabling IFN‐γR to directly activate surface receptors and regulate neuronal excitability [53]. Following acupuncture treatment, IFN‐γR levels on trigeminal nucleus caudalis (TNC) neurons were significantly reduced in model rats, demonstrating acupuncture's ability to decrease neuronal excitability [54]. This confirms acupuncture's capacity to suppress migraine neuronal excitation, providing an integrated protective mechanism against NVU pathology in migraine.

4.1.1. Acupuncture Activates the Endogenous Pain Modulatory System

The Endogenous Pain Modulatory System (EPMS) constitutes an intrinsic pain regulation network within the CNS. Its primary structural foundations encompass brain regions including the periaqueductal gray matter (PAG) of the midbrain, the rostral ventromedial nucleus (RVM) of the medulla, the raphe nuclei, and the anterior cingulate cortex (ACC) [55]. This system exerts top‐down inhibition on excessive activation of pain pathways by releasing endogenous opioids (e.g., β‐endorphin, enkephalin), serotonin (5‐HT), and norepinephrine (NE) [56]. Acupuncture first excites PAG neurons by activating Aδ/C‐type afferent fibers [57], thereby initiating the PAG‐RVM descending inhibitory pathway. The core mechanism of this pathway lies in its dual neurotransmitter action [58], specifically manifested through endogenous opioid pathways and monoamine neurotransmitter pathways. In the endogenous opioid pathway, β‐endorphin and enkephalin act on μ/δ receptors [56] of secondary pain‐transmitting neurons. They reduce glutamate and substance P release through presynaptic inhibition while inducing hyperpolarization via postsynaptic potassium efflux, directly lowering neuronal excitability [59]. The monoamine pathway, via 5‐HT and NE projections, activates 5‐HT1B/1D and α2 receptors respectively, inhibiting the release of pro‐pain substances (e.g., CGRP) and reducing the secretion of pro‐inflammatory factors (e.g., IL‐6, TNF‐α) by astrocytes, thereby blocking pain signal transmission [60]. Additionally, the adenosine signaling pathway is also involved. Animal studies indicate that acupuncture as a potential approach significantly elevates local adenosine concentrations at acupoints. By activating adenosine A1 receptors (A1AR), it inhibits voltage‐gated calcium channels and reduces excitatory neurotransmitter release [61].

Abnormal functional connectivity in the EPMS of migraine patients is a key factor in neuronal excitability dysregulation [62]. Resting‐state functional MRI (rs‐fMRI) studies revealed [63] significantly reduced functional connectivity between the PAG and the anterior cingulate cortex/medial prefrontal cortex (rACC/mPFC) in patients with migraine without aura, indicating impaired descending inhibitory pathways. Acupuncture can regulate these pathways by restoring such abnormal connectivity. Studies indicate that post‐treatment recovery of PAG‐rACC/mPFC functional connectivity strength, coupled with reduced hyperconnectivity between bilateral posterior cingulate cortex (PCC) and dorsomedial prefrontal cortex (dMPFC), directly restores the integrity of the “cortex‐brainstem‐spinal cord” descending inhibitory axis [64]. Furthermore, acupuncture may activate oxytocin neurons in the paraventricular nucleus (PVN) of the hypothalamus, enhancing their projection to γ‐aminobutyric acid (GABA)‐ergic neurons in pain nuclei. This increases GABA release [65], strengthens GABAergic inhibition, and significantly reverses neuronal hyperexcitability. The involvement of the endocannabinoid system (ECS) is also noteworthy. Acupuncture may stimulate cannabinoid CB1 receptor (central type) and CB2 receptor (peripheral type), jointly inhibiting pain transmission by suppressing microglial activity and inflammatory factor release [66].

4.1.2. Acupuncture Regulation of Neurotransmitter and Receptor Balance

Neurotransmitters are categorized into excitatory and inhibitory types [67]. The former induces depolarization by increasing postsynaptic membrane permeability to Na⁺, thereby promoting neuronal discharge. The latter induces hyperpolarization by enhancing Cl⁻ or K⁺ permeability, thereby reducing neuronal excitability. Receptors can specifically bind neurotransmitters to trigger ion channel opening (e.g., NMDA receptors mediating Ca²⁺ influx) [68] or intracellular signaling pathways (e.g., GABA_A receptors mediating Cl⁻ influx), thereby regulating neuronal electrical activity.

In the pathophysiology of migraine, disruption of the excitatory‐inhibitory balance is a key factor in pathological neuronal excitation [69]. This manifests as hyperactivity of excitatory neurotransmitter systems (e.g., excessive glutamate activation of NMDA receptors triggering calcium overload and CSD [70]) or dysfunction of inhibitory neurotransmitter systems (e.g., weakened GABAergic inhibition leading to neuronal hyperexcitability). This imbalance induces central sensitization, perpetuating the migraine cycle [71]. Studies indicate that in CFA‐injected mice, phosphorylation of GluA2ser880 in AMPARs and subsequent GluA2 internalization elevate AMPAR density on neuronal membranes, promoting Ca²⁺ influx and resulting in inflammatory pain hypersensitivity [72]. Additionally, electroacupuncture (EA) at Zusanli and Kunlun elevates mechanical pain threshold and thermal pain latency in sciatic nerve ligation (SNL) rats, correlates with decreased P2X4 receptor expression in the dorsal horn of the spinal cord, reduces microglial activation, and simultaneously upregulates GABA_Aγ2 receptor activity [73]. Electroacupuncture can also stimulate Zusanli and Sanyinjiao to specifically inhibit glutamatergic neuronal projections from the ACC to the ventrolateral periaqueductal gray (vlPAG), thereby blocking descending facilitation pathways for pain signals [74]. Further studies indicate that electroacupuncture may downregulate P2X7 receptor expression in the spinal dorsal horn, inhibits p38 mitogen‐activated protein kinase phosphorylation, reduces microglial activation and IL‐1β/TNF‐α release, and suppresses abnormal dendritic spine remodeling [75]. During migraine attacks, changes in 5‐HT levels affect vascular tone and neuronal excitability, while activation of 5‐HT receptors in descending pathways promotes central sensitization. Studies confirm that compared to the control group, electroacupuncture suppressed 5‐HT activation in rats, improving central sensitization and modulating receptors in the descending pain pathways of migraine animal models [76].

4.1.3. Acupuncture Suppresses Peripheral Sensitization and Inflammation

In the pathophysiology of migraine, peripheral sensitization represents a critical link in the process of neuronal pathological excitation [77]. The pathological mechanism involves nociceptors abnormally activating ion channels and voltage‐gated sodium channels upon inflammatory stimulation (e.g., ATP) at trigeminal nerve terminals. This induces Na⁺/Ca²⁺ influx [78], significantly lowering neuronal excitation thresholds and triggering spontaneous discharges. Sensitized nociceptors further release CGRP and substance P [79], triggering dural vasodilation, plasma protein extravasation [80], and mast cell degranulation, thereby establishing neurogenic inflammation. This inflammatory environment further stimulates trigeminal nerve endings, forming a self‐reinforcing repeated cycle [81]. Immune cell activation plays a pivotal amplifying role in this process. Mast cell degranulation releases proinflammatory mediators such as TNF‐α and IL‐1β, activating satellite glial cells (SGCs) within the trigeminal ganglion. These SGCs enhance neuronal excitability via the CCL2/CCR2 signaling pathway [82]. Thus, effectively intervening in peripheral sensitization and neurogenic inflammation is crucial for blocking the initiation and persistence of migraine.

Extensive evidence indicates that acupuncture exerts its therapeutic effects by inhibiting this peripheral pathological process through multiple targets. Acupuncture directly modulates the release of key neuropeptides. Studies demonstrate that acupuncture reduces levels of CGRP and P in the plasma and trigeminal ganglion of animal models of migraine, thereby alleviating neurogenic inflammation and vasodilation at their source [83] Moreover, acupuncture as a potential approach exhibits significant anti‐inflammatory effects. It downregulates the expression of multiple pro‐inflammatory mediators (e.g., TNF‐α, IL‐1β, IL‐6) while potentially promoting the release of anti‐inflammatory factors (e.g., IL‐10). This improves the inflammatory microenvironment of damaged tissues and desensitizes sensitized nociceptors [84]. Acupuncture also modulates immune responses within the trigeminal ganglion, suppressing SGCs activation and disrupting abnormal signaling between these cells and neurons, thereby stabilizing neuronal excitability [85]. From this, it can be seen that acupuncture can rapidly inhibit the release of CGRP and stabilize the pain threshold, providing a mechanistic basis for its application in acute interventions. There is an assumption that this rapid peripheral nerve regulation may be related to the immediate clinical effect of terminating episodic migraine attacks.

4.2. Acupuncture Maintains Blood–Brain Barrier Integrity

As the core structure of the NVU, the BBB comprises EC, tight junction proteins, astrocytic endfeet, pericytes, and the basement membrane. The extensive coverage formed by astrocytic endfeet is crucial for maintaining the structural integrity of the BBB [86]. By precisely regulating substance exchange, the BBB provides nutrients to the brain while effectively blocking harmful substances from the bloodstream [87], serving as a critical defense for maintaining the brain's microenvironmental homeostasis. Under migraine pathophysiology, the activation of CSD and the trigeminal neurovascular system leads to a massive release of inflammatory mediators and vasoactive neuropeptides. This activates matrix metalloproteinases, degrading TJs proteins and basement membrane components, thereby increasing BBB permeability [88]. This triggers plasma protein leakage, ion imbalance, and neuroinflammatory cascades, exacerbating central sensitization and pathological neuronal discharges, thereby forming a self‐amplifying cycle. Research indicates that acupuncture can improve clinical migraine symptoms by enhancing the function of BBB components and maintaining NVU homeostasis.

4.2.1. Acupuncture Improves Endothelial Cell Function and Tight Junction Integrity

The structural and functional core of the BBB consists of the brain microvascular endothelial cells (BMECs), which form a continuous cell layer through TJs, adherens junctions (AJs), and junctional adhesion molecules (JAMs) [89]. Transmembrane TJs proteins primarily comprise the claudin family (particularly claudin‐5), occludin, and JAMs, while adherens junctions are mainly composed of cadherins [90]. These transmembrane domains span the intercellular gaps, binding to proteins on adjacent EC to seal the intercellular spaces. Intracellularly, these proteins connect to cytoplasmically anchored proteins such as Zonula Occludens proteins, which are then stabilized by the actin‐based cytoskeletal network. Together, they maintain the integrity and low permeability of the BBB [91]. This intricate structure rigorously controls the selective passage of substances from the blood to the brain parenchyma, preserving the stability of the central nervous system's internal environment.

Migraine, particularly in migraine with aura [92], involves increased expression and activity of inflammatory mediators, neuropeptides, and matrix metalloproteinases following the CSD and activation of the TGVS [93]. Tight junction proteins and basement membrane components undergo degradation [94]. These alterations compromise EC's ability to form an intact, tight barrier, increasing BBB permeability [95]. Plasma proteins (e.g., albumin, fibrinogen) leak into brain parenchyma, triggering neurogenic inflammation and neuronal hyperexcitability, thereby exacerbating migraine attacks. Additionally, migraine patients often exhibit imbalances in vascular endothelial mediators, such as elevated plasma endothelin‐1 and NO levels. Endothelin and NO are recently discovered endothelial‐derived vasoconstrictive and vasodilatory factors, respectively. Together, they form a balanced system involved in regulating cerebral vascular tone. Metabolic abnormalities further impair endothelial cell function [96], affecting the CBF regulation and contributing to the pathological process of migraine.

Multiple studies have demonstrated through animal experiments and clinical trials that acupuncture may effectively improves endothelial cell function and TJs integrity in migraine models. Yasemin Cayir et al. found that MMP‐2 activity was clinically reduced in migraine patients treated with acupuncture, thereby decreasing BBB permeability [97]. Ke Hongkui et al. investigated that Zhulian acupuncture at points including “Ergu,” “Lianggu,” “Shousanli,” “Hegu,” “Yangbai,” “Tongzi Liao,” and “Zusanli” not only reduced headache frequency and duration but also decreased serum vascular endothelial growth factor (VEGF) and CGRP levels, thereby alleviating vascular permeability and neurogenic inflammation [98]. Yasemin Gündüztepe et al. observed that after five acupuncture sessions, migraine patients exhibited significantly reduced serum nitric oxide levels, which plays a crucial role in protecting endothelial cell function [99]. Applying these mechanisms to clinical practice, the repair of the closely connected structures seems to be a cumulative process, which might partly explain why preventive acupuncture intervention is needed to reduce the number of monthly migraine attacks. Within this translational research framework, establishing the association between the mechanisms and the results using advanced neuroimaging techniques will be highly beneficial. Recent high‐resolution neuroimaging has transitioned DCE‐MRI from a conceptual tool to a gold‐standard for quantifying BBB permeability in migraineurs [100]. Consequently, future clinical protocols should integrate DCE‐MRI and advanced PET as primary endpoints [101]. By utilizing these modalities to objectively track the restoration of endothelial integrity and the attenuation of tracer leakage, acupuncture's clinical efficacy can be validated through the lens of structural NVU preservation, providing a robust, non‐invasive standard for evaluating therapeutic response.

4.2.2. Acupuncture Modulates Astrocytic Reactivity and Neuroinflammatory Signaling

Within the BBB, astrocytes guide surrounding vascular EC to form TJs, reduce vesicular transport, and express specific transporters [102]. This induces the formation of BBB characteristics and provides physical support and stability through terminal foot processes connecting to the vascular basement membrane, helping maintain vascular structural integrity. Astrocytes also regulate CBF to adapt to neuronal demands. Moreover, upon sensing CGRP, substance P, or inflammatory mediators (e.g., TNF‐α, IL‐1β), astrocytes become hyperactivated [103]. This triggers the production of reactive oxygen/nitrogen species (ROS/RNS), inducing oxidative stress that further damages neurons and BBB EC [104]. Secreted inflammatory mediators and matrix metalloproteinases can downregulate endothelial tight junction proteins (e.g., Claudin‐5, Occludin, ZO‐1) [105], disrupting their structure and increasing BBB permeability. This dysfunction in regulating CBF may cause abnormal vasodilation or vasoconstriction, allowing K⁺ ions, NO, adenosine, or other products to reach the perivascular trigeminal afferents within the dural vessels. This sensitizes the afferents, leading to the symptoms of pulsatile headaches. Multiple studies confirm that astrocyte‐mediated neuroinflammatory responses play a significant role in the pathogenesis of migraine [106].

Research indicates that acupuncture treatment can effectively suppress astrocyte activity in both inflammatory and neuropathic pain [107]. Zhao Luopeng et al. established a migraine model using epidural electrical stimulation and found that electroacupuncture at the “Fengchi” and “Yanglingquan” acupoints reduced serum levels of inflammatory cytokines IL‐1β, IL‐6, and TNF‐α, thereby inhibiting migraine [108]. The trigeminal nucleus complex is a key brain region closely associated with migraine pathogenesis. Liu Yi et al. found significantly increased connexin 43 (Cx43) protein expression in the TNC of migraine rats, which was markedly downregulated by acupuncture treatment [54]. Neurotrophic factor‐3 (NT‐3) and IL‐1β are implicated in neuropathic pain. Using a chronic constriction injury (CCI) model, Tu Wenzhan et al. demonstrated that electroacupuncture may upregulate NT‐3 expression in the dorsal root ganglion (DRG) of CCI rats while downregulating IL‐1β expression. Electroacupuncture intervention significantly suppressed CCI‐induced activation of microglia and astrocytes [109].

4.2.3. Acupuncture Enhances Pericytic Function and Vascular Stability

Pericytes reside perivascularly, embedded within the basement membrane and in close proximity to EC. They serve as the central component within the NVU alongside EC, astrocytes, and neurons, primarily participating in microvascular constriction, maintenance of vascular stability, and ECM regulation [110]. Their role in sustaining vascular stability is particularly prominent. Pericytes interact with EC via gap junctions and adhesion plaques [111], transmitting mechanical stress signals to enhance vascular shear resistance. They secrete ECM components such as type IV collagen and laminin, reinforcing the basement membrane structure and preventing microvascular collapse within the BBB [112]. Pericytes do not directly act on the BBB but instead secrete factors like TGF‐β1 and Ang‐1 to promote the expression of TJs proteins such as Claudin‐5 and Occludin. By supporting EC and regulating their function, they indirectly maintain the barrier [113], stabilizing the microvascular system. As the central regulators of microvascular constriction and guardians of the basement membrane, pericyte dysfunction leads to blood flow disruption, BBB structural breakdown, and amplified inflammation [114], particularly in migraine aura where it correlates with CSD and white matter lesions. Therefore, modulating pericytes and endothelial cell structures to enhance vascular stability is crucial for maintaining BBB integrity and regulating CBF homeostasis.

To date, no single protein has proven to be a perfect marker for pericyte cells, and their identification typically requires the combined use of multiple markers. This may explain why research on pericyte cells using needle puncture techniques has been limited. Furthermore, since pericyte pathology directly leads to BBB disruption, neuronal death, and irreversible functional impairment [115], stroke and alzheimer's disease (AD) models exhibit dramatic pathological changes with distinct phenotypes that are easily observable and quantifiable. Most research on acupuncture's effects on pericytes has focused on cerebral ischemia and dementia models. Therefore, this review, by exploring the known mechanisms in stroke and AD models as indirect evidence, provides new ideas for future research on the effects of acupuncture on pericytes in migraine models. In ischemic stroke models, acupuncture at the Renzhong and Baihui points may enhance pericytes' vitality while reducing apoptosis and migration [116]. In chronic cerebral ischemia animal models, electroacupuncture promotes PDGFR‐β, CD31, and ZO‐1 expression. Under electroacupuncture intervention, pericytes and ECs exhibit more structurally intact forms and abundant TJs. This suggests electroacupuncture may influence the number and function of pericytes and EC [117]. In clinical trials for vascular dementia, acupuncture improves cognitive function by promoting functional neovascularization through upregulating the angiogenic factor VEGF and coordinating pericytes‐endothelial cell interactions [118]. Although they share some NVU injury mechanisms with migraine, fundamental differences exist—most notably, migraine's transient cortical spreading depolarization versus stroke's persistent ischemia. Therefore, acupuncture treatment for migraines remains a hypothetical study and no direct experiments have been conducted to verify it yet.

4.2.3.1. Current Limitations and Future Directions for Pericyte Research in Migraine

Now, no direct studies have investigated acupuncture's effects on pericytes in migraine models. However, some research shows that, during migraine attacks, CSD and activation of the TGVS trigger neuroinflammation, oxidative stress, and BBB disruption that may be similar to stroke and AD. As a critical component of the BBB, pericyte dysfunction also contributes to migraine pathogenesis [119]. Therefore, we make the assumption that the anti‐inflammatory, antioxidant, and anti‐apoptotic effects of acupuncture demonstrated in stroke and AD studies are highly likely to similarly protect pericellular cells in migraine models. Current research on acupuncture and pericellular cells in migraine remains limited. Future studies should utilize specific pericyte markers to quantitatively assess changes in pericyte coverage and their impact on vascular stability following acupuncture in migraine models.

4.2.4. Acupuncture Prevents Excessive Degradation of Extracellular Matrix

ECM, primarily the basement membrane, provides a physical scaffold for the physical support and attachment of EC, pericytes, and astrocytic endfeet. It “binds” together the various cellular components of the NVU, forming a stable, integrated structure [120]. Its diverse proteins—including laminin, fibronectin, and collagen—bind to receptors such as integrins on cell surfaces, continuously signaling to regulate cell survival, proliferation, migration, differentiation, and gene expression [121]. The ECM stores diverse growth factors and cytokines, releasing them as needed to activate or inhibit downstream signaling pathways (e.g., PI3K/Akt, MAPK/ERK), thereby influencing overall BBB function and stability. CGRP and inflammatory factors stimulate the up‐regulation of matrix metalloproteinases (MMPs) activity, and MMPs degrade ECM [122], which is different from the loss of endothelial cell function and the inability to maintain tight junctions between cells. MMPs lose the physical barrier function of ECM from the tissue and structural level, providing a direct channel for the leakage of plasma components [123]. Therefore, clinical treatment focuses on preventing excessive degradation of extracellular matrix ECM, which is essential for maintaining BBB integrity.

Migraine and stroke share significant commonalities in the mechanism of neurovascular unit (NVU) damage, both involving processes such as disruption of the blood–brain barrier (BBB), neuroinflammation, glial cell activation, and degradation of extracellular matrix (ECM) enzymes. Among these, the activation of MMP‐9 is considered a pivotal factor in compromising BBB integrity [124]. Therefore, we attempted to use the known mechanisms derived from studies on stroke and other conditions as indirect evidence, and to infer the mechanisms applicable to the migraine model. In ischemic stroke models, numerous studies have demonstrated that electroacupuncture at acupoints such as “Fengchi” and “Baihui” may effectively mitigate BBB damage and cerebral edema by inhibiting MMP‐9 activity and upregulating TIMP‐1 expression [125]. Research has revealed elevated plasma MMP levels, particularly MMP‐9, in migraine patients during headache episodes, suggesting MMP elevation plays a pathogenic role in migraine attacks [126]. This suggests acupuncture may exert protective effects in migraine by regulating the MMP‐9/TIMP‐1 balance through similar pathways. However, this assumption has not yet been directly verified in the migraine model. The above speculation was made because migraines and strokes share similar mechanisms of damage to the NVU. Activation of the TGVS induces neurogenic inflammation, which further activates MMPs and damages the ECM [127]. Acupuncture may indirectly reduce ECM destruction by modulating multiple signaling pathways to inhibit neurogenic inflammation.

4.2.4.1. Limitations and Future Directions of Extracellular Matrix Research on Migraine

Currently, direct experimental evidence for acupuncture protecting the BBB in migraine by intervening in the ECM remains relatively limited. Most mechanistic inferences derive from other models. Future research utilizing migraine animal models is needed to investigate acupuncture's effects on the expression and arrangement of ECM components and to elucidate its precise signaling pathways. This will provide a foundation for future exploration of acupuncture's impact on migraine from an ECM perspective.

4.3. Acupuncture Modulates Neuroinflammation Mediated by Microglia

As the primary innate immune cells in the CNS, microglia exert immune surveillance functions in the dynamic equilibrium of the NVU through their highly plastic somatic branches [128]. They not only clear cellular debris and abnormal proteins through phagocytosis [129], maintaining homeostasis, but also engage in bidirectional communication with surrounding neurons, astrocytes, vascular EC, and pericytes to precisely regulate synaptic plasticity, neurotransmitter balance, and BBB permeability [130]. Under steady‐state conditions, microglia exist in a resting, dendritic‐branched state (M0 type). Upon disruption of NVU homeostasis, such as tissue injury, pathogen invasion, or abnormal neural electrical activity, microglia rapidly activate [131], undergoing morphological and functional transformation into two distinct states termed M1 and M2. M1 microglia are defined by pro‐inflammatory and pro‐cytotoxic functions, while M2 microglia participate in immune regulation, control of inflammatory mechanisms, and tissue repair. M2 microglia also possess the ability to phagocytose cellular debris and contribute to neural repair [132].

For a long time, research on the pathophysiology of migraine has primarily focused on CSD and the TGVS. However, increasing evidence suggests that neuroinflammation serves as a crucial bridge connecting peripheral stimuli with central perception, with microglia playing a central role in this process [133]. Studies demonstrate that in nitroglycerin (NTG)‐induced migraine animal models, microglia in key pain‐processing brain regions such as the thalamus are significantly activated, with upregulation of their specific marker Iba1 [134]. Using RNA sequencing results and the Panglaodb database, the glial‐associated differentially expressed gene S100A8 was identified as a target for interference in migraine. S100A8 interference suppressed microglia activation in migraine rats, reduced pain perception, inhibited inflammatory factor release, and alleviated NTG‐induced neuroinflammation by suppressing glial M1 polarization [135]. The microglial inhibitor minocycline effectively alleviated hyperalgesic behavior in migraine animal models, directly demonstrating that activated microglia are active “participants” rather than passive “bystanders” in the migraine pathophysiological process [136].

During the initial phase of a migraine attack, events such as CSD or trigeminal ganglion stimulation trigger massive release of ATP, K⁺, glutamate, NO, and inflammatory mediators like CGRP. These substances act as potent stimulants for activating microglia [137]. These microglia then trigger their transition to a pro‐inflammatory phenotype (classically activated, M1 type) through pattern recognition receptors on their surface, including purinergic receptors P2X4, P2X7, P2Y12, and Toll‐like receptors (TLRs) [138]. Activated M1 microglia release a cascade of proinflammatory cytokines (e.g., TNF‐α, IL‐1β, IL‐6) and chemokines, further amplifying neuroinflammation [139]. These inflammatory mediators then trigger migraine episodes through peripheral sensitization, central sensitization, and disruption of the BBB, creating a chain reaction. Therefore, exploring strategies to inhibit excessive microglial activation, regulate their phenotypic polarization, and suppress downstream inflammatory cascades holds significant clinical relevance.

In a rat model of recurrent migraine‐like headaches, electroacupuncture at the “Fengchi” and “Yanglingquan” acupoints was administered to observe changes in CGRP and Iba1‐labeled microglial activation and the resulting inflammatory response. After analysis, it was found that the activation of microglia was reduced, and the pain hypersensitivity of the model was alleviated [140]. The toll‐like receptor 4 (TLR4) is a key trigger for inducible macrophage activation. NF‐κB is the downstream molecule of TLR4 and the main regulator of inflammatory cytokines. TLR4 can promote NF‐κB, thereby further promoting IL‐1β and TNF‐α [141], and ultimately promoting the inflammatory response. Studies have observed that acupuncture may alleviate inflammation by inhibiting the TLR4/NF‐κB signaling pathway in microglia, and TLR4 antagonists can mimic anti‐inflammatory effects [142]. After 20 sessions of electroacupuncture at the “Fengchi” point, Pei Pei et al. found that repeated dural electrical stimulation (DES) significantly suppressed microglial activation in a rat model of recurrent migraine established by SNL [143]. The increase of PD‐L1 expression promotes M2‐like polarization. Under the intervention of electroacupuncture, the pain threshold of SNL rats is increased, the expression of PD‐L1 is promoted, and the transformation of microglia polarization from M1 phenotype to M2 phenotype is promoted, and the inflammatory response is reduced [144]. Microglial activation is accompanied by increased receptor expression. Studies revealed that electroacupuncture at the “Fengchi” and “Yanglingquan” points may decrease the levels of the purinergic receptor P2X4 on microglia and restrict inflammatory responses (NLRP3/Caspase‐1/IL‐1β signaling pathway) [145]. In addition to the classic inflammatory pathways, recent studies have shown that neuronal‐derived CGRP directly regulates the M1 polarization of microglia, thereby forming a positive neuroinflammatory feedback loop [146]. Moreover, molecular switches such as the TREM2 [147] and CX3CL1/CX3CR1 axis [148] have been proven to be key regulatory factors for the transformation of microglia from the M1 to the M2 phenotype during the pathogenesis of migraine. Clinically, to advance precision acupuncture, emerging epigenetic signatures, such as circulating exosomal microRNAs reflecting microglial states, hold promise as predictive biomarkers for individual treatment responses.

4.4. Acupuncture Improves Energy Metabolism and Oxidative Stress

NVU energy metabolism serves as the ultimate power source for maintaining brain homeostasis and function. Through the NVC mechanism, the NVU precisely regulates local CBF to match neuronal energy demands, a process dependent on efficient energy metabolism [149]. Neuronal firing and neurotransmitter release/reuptake are highly energy‐intensive processes. When energy supply is insufficient, ion gradients cannot be maintained, leading to instability in the resting membrane potential, abnormally heightened excitability, and increased susceptibility to abnormal discharges [150]. The formation of TJs in cerebral microvascular EC and the integrity of the cells themselves require continuous energy supply. Insufficient energy supply compromises BBB function and increases permeability [151]. Thus, energy metabolism disorders directly cause NVC dysfunction, impairing the brain's ability to adapt to changing energy demands. Migraine onset is associated with impaired brain energy metabolism [152]. Studies indicate that in NTG mouse models, energy metabolism significantly decreases during central sensitization, accompanied by increased excitatory neurotransmitters like glutamate and a tendency toward reduced inhibitory neurotransmitters such as GABA. This demonstrates that abnormal energy metabolism and neurotransmitter imbalance are key mechanisms driving migraine chronicity [153].

Oxidative stress refers to a state where the production of reactive oxygen species (ROS) within the body exceeds endogenous antioxidant defense capabilities, leading to damage to cellular structure and function [154]. The brain's high oxygen consumption and lipid‐rich characteristics render it highly susceptible to oxidative damage. Under normal conditions, ROS participate in regulating synaptic plasticity, gene expression, and vascular tone. When energy metabolism is disrupted, inducing migraine onset, excessive ROS production [155] leads to endothelial dysfunction, increased BBB permeability [156], mitochondrial damage, and energy deficiency. Excessively produced ROS and other oxidants are sensed by the Transient Receptor Potential Ankyrin 1 (TRPA1) ion channel present in nociceptive nerve endings [157]. Activated TRPA1 can induce pain signal transduction and neurogenic inflammation, which are characteristic of migraine attacks [158]. It has been proposed that the brain energy deficiency‐mitochondria‐oxidative stress axis may represent a key pathway in migraine pathogenesis [159]. Therefore, therapeutic interventions targeting brain energy metabolism—specifically those that reduce ROS, enhance mitochondrial function, and decrease free radical production and oxidative damage—are crucial for protecting the NVU and mitigating migraine‐related pathology.

Liu Lu et al. analyzed clinical trial data using proteomics and metabolomics, revealing that after 4 weeks of acupuncture treatment, migraine patients exhibited reduced oxidative stress levels, increased glucose availability in the brain, and restored energy homeostasis. They concluded that energy metabolism pathways may provide a critical link between migraine development and the molecular changes emphasized by acupuncture treatment [160]. Luo Jianchang et al. observed that in rats with nitroglycerin‐induced migraine, ROS levels increased and mitochondrial damage occurred. Following electroacupuncture intervention, ROS decreased and migraine‐related mitochondrial dysfunction improved via the PINK1/Parkin pathway [161]. Research indicates that in a chronic inflammatory pain model established in male rats, daily electroacupuncture stimulation at the “Zusanli” acupoint may reduce transient receptor potential vanilloid subtype 1 (TRPV1) expression, decreased ROS oxidative stress, and alleviated pain [162] (Table 1).

TABLE 1.

Summary of experimental research on the key mechanism of acupuncture regulating migraine neurovascular unit (NVU).

Target in NVU	Key mechanism	Animal/cell model	Acupuncture intervention	Key findings	References
Neuron	Regulate neuronal excitability	Migraine rats	MA, “Shuaigu,” “Yanglingquan,” 20 min, once daily for 9 days	Reduced IFN‐γR levels	[54]
Neuron	Adenosine signaling pathway	Mouse models of chronic pain	MA, 1.5 mm deep in the “Zusanli,” every 5 min for 30 min	Decreased excitatory neurotransmitter release	[163]
Neuron	Repair abnormal connection	Patients with migraine	MA, “Yanglingquan,” “Qiuxu,” “Waiguan,” “Xiyangguan,” “Diwuhui” and “Sanyangluo,” “Zusanli,” “Chongyang,” “Pianli,” 5–15 mm deep, 30 min each, once per day for 5 weekdays followed by a 2‐day break.	Restored integrity of the “cortex‐brainstem‐spinal cord” descending inhibitory pathway	[164]
Neuron	GABA release	Adult male wistar rats	EA, “Neiguan,” “Jianshi,” 1–4 mA, 2‐Hz	Activated oxytocin neurons in the PVN	[165]
Neuron	Endocannabinoid system	Male C57BL/6J mice	MA, “Baihui,” “Sanyinjiao,” “Dazhui,” 0.5–1.5 mA, 2/15‐Hz, 15–30 min	Inhibited pain transmission	[66]
Neuron	Increase threshold, inhibit expression	SNL mice model	MA, “Baihui,” “Dazhui,” for 30 min, once daily for 7 days	Balanced excitation‐inhibition	[73]
Neuron	Regulating receptor balance	SNI mice model	EA, “Zusanli,” “Sanyinjiao,” 2‐Hz, 0.1 mA, 30 min	Blocking descending facilitation pathways for pain signals	[74]
Neuron	P2X7 receptor expression	SNL mice model	EA, “Zusanli,” “Kunlun,” 2‐Hz, 1.5 mA, 30 min	Inhibiting abnormal dendritic spine remodeling	[75]
Neuron	5‐HT levels	Male Sprague–Dawley rats	EA, “Fengchi,” 2‐Hz, 0.5–1.0 mA, 15 min, days 1 through 7, for a total of seven treatments	Enhancing receptors in descending pain pathways	[76]
Neuron	Release of key neuropeptides	Migraine rats	EA, “Yanglingquan,” “Taichong,” 1 mA, 2 Hz/100‐Hz, 30 min	Alleviating neurogenic inflammation and vasodilation	[166]
Neuron	Anti‐inflammatory effect	DES stimulation‐induced migraine rat model	EA, “Baihui,” “Shenting,” 1.0–3.0 mA, 2 Hz/100‐Hz, 2–3 mm deep, 30 min	Desensitizing nociceptors	[84]
Neuron	Regulate immune response	Mice model of migraine‐like pain	EA, “Fengchi,” “Yanglingquan,” 1 mA, 2/15‐Hz, 15 min per session	Interrupts abnormal signaling between nociceptors and neurons	[167]
BBB	Reduce MMP‐2 activity.	Patients with migraine	MA, “Touwei,” “Neiting,” “Hegu,” “Quchi,” “Taichong,” “SanYinjiao,” “TongZiliao,” “Yangbai,” “Fengchi,” “Yintang,” “Taiyang,” “ear Shenmen,” 10 sessions for 5 weeks, two sessions a week	Clinically observed reduction in MMP‐2 activity and decreased BBB permeability	[97]
BBB	Vascular permeability, neuroinflammation	Patients with migraine	MA, “Ear‐Liao,” “ShouSanli,” “Hegu,” “Yangbai,” “TongziLiao,” “ZuSanli,” 1 min each time, 15–30 min 4 weeks	Reduces headache frequency and duration while lowering serum VEGF and CGRP levels	[98]
BBB	Endothelial cell function	Patients with migraine	MA, “Hegu,” “ShouSanli,” “Fengchi,” 5 sessions with 2 sessions per week	Significantly decreased serum nitric oxide levels	[99]
BBB	Activation of astrocytes	Epidural electrical stimulation model	EA, “Fengchi,” “Yanglingquan,” 1.8–2.0 mA, 20‐Hz, 15 min period every other day over a total of three sessions	Reduced serum inflammatory cytokines IL‐1β, IL‐6, TNF‐α	[168]
BBB	The expression of Cx43 protein	Migraine rats	MA, “Shuaigu,” “Yanglingquan,” 5 mm deep, 20 min each time, once a day, for 9 days	Downregulates Cx43 protein expression	[54]
BBB	NT‐3, IL‐1β	CCI rats	EA, “Zusanli,” “Yanglingquan,” 2–3 mm, 2/100‐Hz, 30 min	Inhibits activation of microglia and astrocytes	[109]
Microglia	Inhibition of microglial activation	Rat model of recurrent migraine headache	EA, “Fengchi,” “Yanglingquan,” “Waiguan,” 10–50‐Hz, 30 min	Reduced microglial activation alleviates pain hypersensitivity	[140]
Microglia	Inhibition of TLR4/NF‐κB signaling pathway	Rat model of migraine	EA, “Fengchi,” “Yanglingquan,” 0.5–1.0 mA for 20 min/day	Suppresses microglial inflammatory responses	[142]
Microglia	Microglia activation	Rat model of recurrent migraine headache	EA, “Fengchi,” 0.5–1.0 mA, 2/15‐Hz, 15 min	Inhibits microglial activation	[143]
Microglia	M2 sample polarization	SNL mice model	EA, “Kunlun,” “ZuSanli,” 2–3 mm, 2/100‐Hz, once a day for 30 min each time for seven consecutive days	Promotes M2‐like polarization	[144]
Microglia	Inflammatory response	Migraine rats	EA, “Fengchi,” “Yanglingquan,” 0.5–1 mA, 2/15‐HZ, 20 min	Decreased P2X4 receptor levels limit inflammatory responses	[145]
Energy Metabolism	Improve mitochondrial dysfunction	Migraine rats	EA, “Fengchi,” “Taichong,” 0.5–1 mA, 2/15‐HZ, once a day, 15–20 min, 5 days	Improves mitochondrial function via PINK1/Parkin pathway, reducing ROS	[161]
Energy Metabolism	Oxidative stress	Male rat model of chronic inflammatory pain	EA, “Zusanli,” 2‐Hz, 1 mA, 15 min	Decreased TRPV1 expression alleviates pain	[162]

5. Parameter‐Specific Effects of Acupuncture on NVU Components

In the current intervention of acupuncture on the various components of the NVU in the migraine model, a recurring limitation is the heterogeneity of the acupuncture protocols, including the selection of acupoints, the methods, the frequency, the intensity, and the treatment duration. Most studies have not systematically compared these parameters, making it difficult to attribute the different results of the MVU to specific stimulation characteristics.

However, some preliminary patterns can still be seen from the existing literature. High frequencies such as 100‐Hz or alternating frequencies such as 2/100‐Hz EA therapy are more common in studies reporting activation of the endogenous opioid pathway, while low frequencies such as 2‐Hz stimulation are related to adenosine‐mediated effects. Some animal experiments directly compared the effects of 2 and 100‐Hz electrical stimulation therapy on neuropathic pain. The results showed that the analgesic effect of 2‐Hz EA was higher than that of 100‐Hz EA [169]. Moreover, classic studies have confirmed that 2‐Hz electroacupuncture therapy produces analgesic effects through μ and δ opioid receptors, while 100‐Hz acts through κ opioid receptors [170]. Recent research indicates that ultra‐high frequencies such as 18 kHz electroacupuncture therapy can activate different serotonergic pathways from those activated by traditional frequencies [171]. Regarding acupoint specificity, electroacupuncture therapy at the “Fengchi” acupoint has been proven to directly reduce the neuronal discharge of the trigeminal cervical complex in migraine rats [52], indicating that local cervical acupoints may preferentially act on the brainstem pain processing nuclei.

Based on these findings, we propose a framework for future research to stratify acupuncture protocols according to the dominant pathological subtypes of the neural vascular unit in migraine. For neuroinflammation‐driven migraines characterized by elevated cytokine levels, microglial cell activation, and astrocyte reactivity, parameters that can regulate neuroglia, such as 2‐Hz electroacupuncture or local acupoints such as “Fengchi”, can be prioritized. For BBBr or vascular dysfunction, including increased permeability and disruption of TJs, parameters that can upregulate occludin‐5 and occludin in stroke models are worthy of testing in migraine models. For excessive neuronal excitation or central sensitization, parameters that activate the descending inhibitory pathway may be most relevant, such as 2/100‐Hz alternating frequency or strong manual stimulation.

Ideally, future mechanism studies should include direct comparisons of at least two different acupuncture protocols on the same NVU outcome indicators, such as BBB permeability, microglial polarization, or neuronal discharge rate. In the absence of such data, the field should prioritize standardized reporting of stimulation parameters for future research (Table 2).

TABLE 2.

Comprehensive dose‐effect framework linking acupuncture parameters to NVU targets in migraine and preclinical models.

Acupuncture parameter	Comparative analysis design	Target NVU component	Dose‐effect specificity conclusion	Representative supporting studies
Stimulation frequency	2 Hz vs. 100 Hz EA	NVU neuronal component	2 Hz: μ/δ‐receptor mediated antinociception; 100 Hz: κ‐receptor specific	[172]
Stimulation frequency	2/100 Hz (alternating) vs. fixed frequencies	NVU neurotransmitter pool	Synergistic opioid release; robust, sustained suppression of hyperexcitability	[173]
Acupoint selection (spatial)	Local “Fengchi” vs. distal/sham points	NVU glial‐neuronal crosstalk	Superior inhibition of TCC nociception and regional microglial activation	[142]
Treatment timing and duration	Cumulative (pretreatment/repeated) vs. single acute session	NVU vascular endothelial component	Cumulative sessions essential for BBB structural restoration (TJs upregulation, MMP‐9 inhibition)	[174]
Stimulation intensity/modality	Low‐frequency neuromodulation (1 Hz) vs. conventional EA	NVU perivascular nerves and vascular smooth muscle	Activates PAG descending inhibition; mitigates high‐intensity‐induced CGRP release	[175]

6. Discussion

Numerous clinical studies support that acupuncture effectively treats and prevents migraine attacks and chronic progression, specifically reducing the number of migraine days [176] and alleviating pain intensity [177], while significantly improving patients' quality of life. In an 8‐week trial, the intervention of acupuncture was compared with the control group, sham acupuncture, or other intervention measures for patients with migraines. The researchers found that both the acupuncture group and the sham acupuncture group showed good results. It was hypothesized that the skin puncture in the sham acupuncture group might have produced a placebo effect or activated fibers, thereby inducing the effect of local microcirculation. However, the acupuncture group had better results than the sham acupuncture group, which were comparable to, or even better than, preventive medications, and also had fewer adverse reactions [178]. Furthermore, a study evaluating acupuncture efficacy across five pain conditions—including low back pain, migraine, fibromyalgia, neck pain, and abdominal pain—identified migraine and fibromyalgia as the two conditions yielding the most favorable outcomes post‐acupuncture [179]. In recent years, scholars have attempted to answer the question of how brief acupuncture stimulation can transform into long‐term preventive efficacy from the perspective of epigenetic remodeling. The latest research shows that EA can induce specific DNA methylation patterns [6, 180] and histone modifications in the trigeminal nerve cervical complex, effectively inhibiting the expression of pro‐inflammatory genes. Moreover, non‐coding Ribonucleic Acids (RNAs) are increasingly recognized as epigenetic mediators that can convert peripheral stimuli into stable transcriptional changes in central glial cells [181]. This epigenetic “resetting” of the NVU may reverse the established inflammatory state, providing a molecular basis for the sustained therapeutic effects observed in clinical practice. Thus, acupuncture demonstrates substantial potential for future migraine management (Table 3).

TABLE 3.

Summary of representative clinical trials on the effectiveness of acupuncture in the treatment of migraine.

Design	Sample size	Intervention group	Control group	Primary outcome measures	Primary outcome	References
A clinical trial	n = 150	MA, “Hegu,” “Taichong,” “Taiyang,” “Fengchi,” “Shuaigu,” 8 weeks 20 sessions	Sham acupuncture, usual care group	Change in migraine days and migraine attacks	MA is better	[182]
A Randomized Clinical Trial	n = 249	EA, “Fengchi,” “Shuaigu,” 5 days per week for 4 weeks, 2/100‐Hz	Sham acupuncture	The change in the frequency of migraine attacks	True acupuncture may be associated with long‐term reduction	[183]
A Randomized Controlled Trial	n = 218	“Fengchi,” “Baihui,” “Taiyang,” “Hegu,” “Taichong,” 20 sessions of acupuncture over 8 weeks	Sham acupuncture	The responder rate	Acupuncture (achieving deqi sensation) reduces the mean headache days (per month) in patients	[184]
A multicenter Randomized Clinical Trial	n = 142	Combined effects of spinal manipulation and dry needling, 1 to 2 times per week for 4 weeks	Spinal mobilization and exercise on	Headache intensity	Upper cervical and upper thoracic high‐velocity low‐amplitude thrust spinal manipulation and electrical dry needling were more effective	[185]
A randomized double‐blinded placebo‐controlled trial	n = 36	EA, “Taiyang,” “Fengchi,” “Hegu,” 2/100‐Hz at 3 s intervals	Usual care for	The effect of acupuncture‐like electrical stimulation on chronic tension‐type headache	Acupuncture‐like electrical stimulation is a safe and potentially analgesic‐sparing therapy	[186]
A Randomized Controlled Multicenter Trial	n = 114	Gall Bladder 20, 40 or 41 or 42, Du Mai—Governing Vessel 20, Liver 3, San Jiao 3 or 5, extrapoint Taiyang, at least 8 to a maximum of 15 sessions of 20 to 30 min duration, administered over a period of 12 weeks.	Metoprolol	The difference in number of days with migraine	Acupuncture group tended to be slightly better	[187]
A Randomized Clinical Trial	n = 66	24 sessions over 12 weeks	Topiramate	The efficacy and tolerability of acupuncture compared with topiramate treatment	Acupuncture provided a greater reduction in monthly moderate/severe headache days.	[188]
A Multicenter Randomized Controlled Clinical Trial	n = 960	Ten sessions in 6 weeks	Sham acupuncture	The difference in migraine days	Verum acupuncture was better	[189]
A multicenter, sham‐controlled, double‐blind randomized trial	n = 409	Ten 30‐min sessions were given over a six‐week period	Sham acupuncture	The responder rate	TTH improves after acupuncture treatment	[190]
A Randomized Controlled Trial	n = 270	Gall bladder 20 GB 21 Liver 3 12 sessions per patient over 8 weeks.	The minimal acupuncture	Difference in numbers of days with headache	Acupuncture was more effective.	[191]
A Randomized, Controlled, Crossover Trial	n = 40	Distal Acupoints Only	Sham acupuncture	Effective and sustained tension‐type headache relief from distal acupuncture.	REA, on distal points only, effectively reduces the frequency of attacks, duration, and intensity of TTH	[192]
A Randomized Trial	n = 94	Ear acupuncture	Unsuitable area	Evaluating the therapeutic value of specific auricular acupuncture points	A therapeutic specificity of auricular points probably exists	[193]
A Randomized Trial	n = 40	12 sessions of acupuncture in 4 weeks	Sham acupuncture	Headache severity, Headache impact: headache impact Test‐6 results	Migraine days reduction after acupuncture treatment	[194]
A Randomized Controlled Trial	n = 15,056	15 acupuncture sessions over 3 months	Usual medical care	The effectiveness of acupuncture in addition to routine care in patients with primary headache	Acupuncture was better	[195]
A Prospective single‐center Randomized Controlled Trial	n = 96	12 sessions of acupuncture in 6 weeks	Usual care	Pain intensity, frequency of headache, responder rate, duration of headache and use of headache medication	The intervention group showed greater improvement in efficiency.	[196]
A Multicenter single blinded Randomized Controlled Trial	n = 150	Received a treatment when having a migraine attack	Sham acupuncture	The primary outcome was visual analog scale scores for pain	The treatment group demonstrated superior outcomes in pain relief and reduction of acute medication use.	[197]

Note: It should be noted that sham acupuncture may produce non‐specific physiological effects, and therefore the magnitude of specific acupuncture effects could be underestimated in sham‐controlled trials.

This review systematically clarifies from the NVU perspective that acupuncture, as a multi‐target intervention strategy, can regulate multiple components of the migraine neural vascular unit (NVU). Increasing evidence indicates that acupuncture can inhibit pathological neural excitability, maintain the integrity of the blood–brain barrier, suppress neuroinflammation, and improve energy metabolism and oxidative stress. The traditional view that acupuncture intervention for migraines often explores the classic descending inhibitory pathways through endogenous opioid and monoamine neurotransmitter transmission. What we want to emphasize is that the NVU mechanism is not in competition with these pathways but may play a complementary role at different time stages. The rapid analgesic effect of acupuncture may mainly rely on the activation of the classic descending inhibitory pathways, while the NVU‐related mechanism is that acupuncture blocks peripheral sensitization, central sensitization, and pathological cycles of neurovascular dysfunction, achieving the effect of stabilizing neural vascular homeostasis and preventing the recurrence of migraines in the long term.

Building upon these mechanistic insights, we propose a translational framework for acupuncture in migraine management. For acute attacks, interventions should target rapid inhibition of CGRP release to block peripheral sensitization. Conversely, for preventive treatment, cumulative acupuncture sessions are required to maintain BBB integrity via tight junction upregulation and epigenetic remodeling of glial cells. Future clinical trials should prospectively incorporate objective NVU biomarkers, such as DCE‐MRI for BBB permeability and circulating exosomal microRNAs, to predict and monitor treatment responses.

However, several pressing issues remain in this field. First, existing research is largely based on animal models or small‐sample clinical trials, limiting the extrapolation of results to human populations. While animal models can simulate certain pathological features of migraine, they differ from humans in details of mechanisms such as NVC and CSD, and cannot fully reflect acupuncture's effect. One more point needs to be added, that is, in clinical experiments of acupuncture, the effect of the sham acupuncture group cannot be ignored. It may interfere with the patient's expectations, conditioned reflexes, and anxiety state, thereby causing a placebo effect. It may also activate unmyelinated (C “tactile”) afferent nerves [198], thereby influencing pain perception and achieving therapeutic effects. Clinical trials exhibit substantial inter‐individual variability, with patient genetic backgrounds, disease duration, and acupuncture protocol choices potentially affecting treatment consistency and reproducibility. Therefore, the research using animal models cannot replace clinical trials. In response to these limitations, in order to ensure that future sham control trials do not confuse variables and effectively control the specific effects of NVU, we have attempted to propose some ideas. For example, whether the sham surgery control can be selected based on the desired NVU results, for results that are not easily affected by superficial stimuli, such as systemic inflammatory markers or central neurotransmitter levels, non‐acupoint sham acupuncture can be chosen. For results closely related to the activation of skin afferent nerves, such as local CBF or neuronal excitability, a conventional care control group may be more appropriate. Additionally, it is also possible to consider incorporating objective NVU biomarkers as the main evaluation indicators. For example, measurement indicators of BBB permeability based on advanced imaging, indicators of cortical excitability in quantitative electroencephalography, or epigenetic signatures such as specific circulating exosomal microRNAs. This will help distinguish specific physiological effects from non‐specific placebo responses. Another point to note is that the characteristics of the sham needles should be reported in detail in the experiments. Many studies will omit some key information, such as the thickness of the needle, the depth of insertion, and the retention time. Standardized reporting of sham procedures will greatly improve the comparability between different trials and help determine which parts of the sham procedures are most likely to interfere with specific NVU results. Meanwhile, there are significant methodological differences in acupuncture treatment itself, such as the selection of acupoints, stimulation parameters (frequency, intensity, duration), and acupuncture techniques. Future research should also place emphasis on standardized interventions.

Most mechanistic studies on acupuncture and the migraine NVU remain exploratory. Current evidence primarily focuses on isolated neuronal or vascular responses, while the integrated regulation of the “neurovascular–glial” unit remains insufficiently understood. Furthermore, at present, the understanding of the NVU mechanism in migraine is partly derived from studies on stroke and neurodegenerative diseases, which are used as indirect evidence for speculation. However, direct evidence in migraine‐specific models is still limited. Although acupuncture appears to influence multiple NVU components simultaneously, the causal relationships and upstream–downstream signaling sequence among these components remain unclear, and current findings are therefore largely correlational rather than causal. This is particularly evident in the interaction between astrocytes and microglia. Relevant studies often report that both cell types are simultaneously inhibited, but do not clarify the temporal sequence or main targets of acupuncture intervention. Future research should further explore the dynamic interactions among the various components of the NVU endothelial network to more clearly elucidate the mechanism framework underlying acupuncture treatment for migraines.

Therefore, elucidating the regulatory mechanisms of acupuncture on causal relationships among components within the NVU should become a key direction for future research. This requires leveraging more advanced techniques, such as in vivo real‐time imaging to monitor intercellular signaling and multi‐omics integrated analysis, to systematically decipher the upstream and downstream regulatory logic of molecular‐cellular‐network interactions within the NVU under acupuncture intervention. Ultimately, this will provide a robust systems biology foundation for the precise application of acupuncture. This also prompts researchers to increase focus on the migraine NVU and consolidate high‐quality clinical data to validate acupuncture efficacy through large‐scale, multicenter studies. Concurrently, basic research should be strengthened using animal models and cellular experiments to uncover the molecular networks underlying acupuncture effects. Ultimately, by integrating evidence‐based medicine with systems biology strategies, acupuncture holds promise as a vital component of personalized migraine treatment and offers novel insights for intervening in neurovascular disorders.

Author Contributions

Yajie Wang: writing – original draft. Yuanyuan Wang: visualization. Chao Zhang: investigation. Linbo Shen: investigation. Dayong Ma: writing – review and editing. Tao Xiong: writing – review and editing, supervision. Li Wang: formal analysis. Ningle Zhang: formal analysis. Xiaocheng Wang: methodology. Linjing Song: methodology.

Funding

The authors declare that no funds were received during the preparation of this manuscript.

Ethics Statement

This article does not contain any studies with patients or animals performed by any of the authors.

Consent

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The authors have nothing to report. All data referenced are from previously published studies as cited.