Epigenetic regulators polyphenols in neurodegenerative diseases: a promising intervention strategy

Unlike histone acetylation, which generally promotes gene transcription activation, the biological effects of histone methylation depend on the specific amino acid residues being modified, the degree of methylation (i.e. mono-, di-, or tri-methylation), and the functional proteins subsequently recruited [60]. Overall, this regulatory process is coordinated by three major classes of factors: histone methyltransferases (‘writers’), demethylases (‘erasers’), and methylation-binding proteins (‘readers’) [60].

In the methylation ‘writing’ process, lysine methyltransferases (KMTs), such as the SETD1/MLL family, EZH2, EHMT1/2, and SETDB1, catalyze methylation at key sites like H3K4, H3K9, and H3K27 [61–63]. In the nervous system, these enzymes help establish finely tuned transcriptional regulation, contributing to neuronal fate determination, activity-dependent gene expression, and long-term memory formation [64]. In neurodegenerative diseases such as Alzheimer’s and Parkinson’s, these methylation patterns are often disrupted early, leading to suppressed expression of neurotrophic and synapse-related genes, and accelerating cognitive decline and neuronal damage [53,65].

Histone demethylases (KDMs), which mediate the ‘erasure’ of methylation marks, are equally essential. Members of the KDM5, KDM6 (JMJD3/UTX), and KDM4 families can dynamically remove methyl marks in response to neuronal activity, metabolic shifts, or inflammatory signals, thereby enhancing chromatin plasticity and transcriptional adaptability [66]. In the context of neurodegeneration, impaired demethylation can leave chromatin in a persistently repressive state, limiting the neuronal response to stress and injury [67].

In addition to the processes of methylation writing and erasing, the functional outcome of histone methylation also relies heavily on specific recognition mechanisms [68]. Methylation reader proteins are able to detect methylated histone tails and translate these modifications into defined transcriptional responses or changes in chromatin organization [68]. For example, heterochromatin protein 1 binds to H3K9me3 and promotes heterochromatin formation, while Polycomb group complexes associate with H3K27me3 to maintain long term gene repression [68,69]. When the function or localization of these reader proteins is disrupted, proper transcriptional regulation may fail to occur even in the presence of methylation marks, ultimately leading to synaptic dysfunction and neurodegenerative alterations [69].

Overall, histone methylation should not be viewed as a static epigenetic modification but rather as a dynamic and highly coordinated regulatory system. Disruption at any level, including methylation deposition, removal, or recognition, can interfere with normal gene expression programs in neurons and constitutes an important molecular basis for the onset and progression of neurodegenerative diseases.

In addition to covalent modifications of histones, the three-dimensional organization of chromatin and the accessibility of DNA are also tightly regulated by ATP-dependent chromatin remodeling complexes [70]. These multi-protein assemblies utilize the energy from ATP hydrolysis to drive nucleosome sliding, repositioning, or eviction along the DNA, thereby dynamically modulating the access of transcription factors and the transcriptional machinery to target genes [70]. In the nervous system, neuronal responses to synaptic plasticity, learning and memory processes, as well as environmental stimuli, rely on highly plastic and rapidly reversible transcriptional responses, making this regulatory mechanism particularly crucial [71].

The major families of chromatin remodeling complexes identified to date include SWI/SNF (referred to as BAF complexes in mammals), ISWI, CHD, and INO80. Among them, the SWI/SNF or BAF complex plays a central role in neurodevelopment and the maintenance of mature neuronal functions. During neuronal differentiation, the subunit composition of the BAF complex undergoes a specific switch to form neuron-specific nBAF complexes, which are essential for dendritic development, synapse formation, and memory consolidation [72]. Numerous studies have shown that alterations in the composition or function of the BAF complex are closely associated with impaired synaptic plasticity and neurodegenerative processes [73,74].

CHD family proteins regulate nucleosome spacing and enhancer accessibility, contributing to activity-dependent transcriptional regulation and neuronal survival. When CHD function is compromised, gene expression programs may become dysregulated, leading to cognitive deficits and various neurological disorders [75]. ISWI complexes are mainly involved in maintaining higher-order chromatin structure, ensuring stable expression of neuronal identity genes while preventing aberrant activation of non-specific gene expression [76].

An increasing body of evidence suggests that dysfunction in chromatin remodeling is an important, yet long-underestimated, pathogenic mechanism in neurodegenerative diseases [77]. As aging or disease progresses, the efficiency of chromatin remodeling tends to decline, reducing the ability of neurons to flexibly adjust gene expression in response to metabolic stress, inflammation, or toxic protein aggregates [77]. Therefore, ATP-dependent chromatin remodeling represents a key epigenetic layer that links environmental signals, transcriptional regulation, and neuronal vulnerability [78].

Regarding microRNAs (miRNAs), these small RNA molecules primarily exert negative regulation of gene expression by binding specifically to the 3′ untranslated regions (3′-UTRs) of target messenger RNAs (mRNAs), leading to mRNA degradation or translational repression [81]. In the nervous system, miRNAs are highly expressed, undergoing stepwise processing by Drosha and Dicer before becoming functionally active within the RNA-induced silencing complex (RISC) complex, where they fine-tune the expression of genes critical for neuronal activity [82].

During early neural development, miRNAs control the differentiation fate of neural progenitor cells [83]. In mature neurons, they regulate synaptic structure and gene expression to maintain functional homeostasis and feedback control [83]. For instance, miR-124 is highly expressed in neurons and maintains neuronal phenotype stability by repressing non-neuronal genes while also modulating chromatin-remodeling factors, thereby linking post-transcriptional regulation with epigenetic control [84,85].

In neural injury and regeneration, miRNAs finely modulate cell fate by regulating signaling pathways involved in apoptosis, autophagy, inflammation, and axonal regeneration [86]. Some miRNAs promote axonal regrowth by inhibiting phosphatase and tensin homolog (PTEN) and activating the mammalian target of rapamycin (mTOR) pathway, while others influence neuronal survival or death under stress conditions such as ischemia or trauma by targeting apoptotic regulators including the Caspase family and B-cell lymphoma 2 (Bcl-2)[87].

At the level of long non-coding RNAs (lncRNAs), these RNA molecules, typically longer than 200 nucleotides, generally do not encode proteins but regulate gene expression through interactions with DNA, RNA, or proteins [88]. They participate in chromatin modification, transcriptional regulation, and post-transcriptional processes such as RNA splicing, stability, and translation, thereby playing crucial roles in multilayered regulatory networks within the nervous system [89]. lncRNAs display high tissue specificity and finely tuned expression patterns in the nervous system, enabling cross-cell-type communication and regulation among neurons, glial cells, neural stem cells, and immune cells [89].

Functionally, lncRNAs can act as competing endogenous RNAs (ceRNAs) that sponge miRNAs, thereby relieving repression of neuronal mRNA targets, or they can directly participate in pathological processes [90]. For example, lncRNAs such as MEG3, NEAT1, and H19 are commonly upregulated in various neurodegenerative diseases and contribute to disease pathology by regulating tau protein phosphorylation, amyloid beta (Aβ) production, mitochondrial autophagy, inflammatory cytokine release, and microglial activation [91–94]. In addition, lncRNAs can modulate the expression of N6-methyladenosine (m6A) modification enzymes or serve as modification substrates themselves, affecting their stability and functional state, thus linking epigenetic regulation with RNA metabolism [95,96]. In central nervous system disorders such as ischemia-reperfusion injury following stroke, lncRNAs influence angiogenesis and autophagy, exerting significant effects on functional recovery and disease progression, and demonstrating considerable therapeutic potential [97,98].

Circular RNAs (circRNAs) are a class of covalently closed-loop RNA molecules formed through back-splicing of precursor mRNAs [99]. Due to the absence of a 5′ cap and a 3′ tail, they exhibit high stability and are abundantly expressed in the nervous system, displaying distinct developmental stage dependent and cell type specific expression patterns [100]. The primary mechanism of circRNA function involves acting as molecular sponges for miRNAs, competitively binding to them and thereby indirectly regulating the expression of target mRNAs [101]. Through this mechanism, circRNAs participate in processes such as synaptic plasticity, axon guidance, neuroinflammation, and cell survival [102,103].

In addition, some circRNAs interact with RNA-binding proteins to modulate mRNA transport, splicing, or local translation [104]. Under pathological conditions such as CNS injury, neurodegenerative diseases, and chronic neuropathic pain, circRNA levels undergo significant alterations. Functioning as ceRNAs that interact with miRNAs, they modulate the transcriptional activity of genes connected to inflammatory processes, oxidative stress responses, ion channel performance, and synaptic plasticity [105,106]. Furthermore, circRNAs can interact with both miRNAs and lncRNAs, collaboratively contributing to the dynamic regulation of the CNS and playing important roles in disease progression [106,107].

In summary, miRNAs, lncRNAs, and circRNAs not only exert their respective regulatory functions within the nervous system but also interact with one another to produce synergistic effects.

Neurodegenerative disorders arise from the prolonged interplay among multiple contributing factors [109]. The core mechanisms involve not only cellular pathological changes such as protein misfolding, mitochondrial dysfunction, disruption of the blood–brain barrier (BBB), and neuroinflammation, but are also closely related to epigenetic regulation [5,109,110]. In recent years, numerous studies have demonstrated that lifestyle and environmental factors, including nutritional metal imbalance, exposure to environmental toxins, smoking, and coffee consumption, as well as the aging process itself, may induce persistent transcriptional abnormalities in the nervous system by influencing epigenetic mechanisms such as DNA methylation, histone modification, and non-coding RNA (ncRNA) expression [111,112]. These changes collectively contribute to neuronal dysfunction and degenerative alterations.

Epidemiological evidence indicates that smoking and coffee consumption are closely associated with a reduced risk of PD [113,114]. However, the underlying mechanisms extend beyond a single neuroprotective effect and involve complex epigenetic regulatory processes [115,116]. Caffeine, as an adenosine A2A receptor antagonist, not only alleviates glutamate excitotoxicity at the synaptic level but also promotes the clearance of α-synuclein by enhancing autophagy and activating the AMP-activated protein kinase (AMPK) and sirtuin 3 (SIRT3) signaling pathways [116]. In addition, caffeine and the coffee component eicosanoyl-5-hydroxytryptamide (EHT) have been found to synergistically increase the methylation level of protein phosphatase 2 A (PP2A), thereby reducing the accumulation of phosphorylated pathogenic proteins [113]. This process involves the regulation of protein phosphatase methylesterase-1 (PME-1) activity and exemplifies a typical enzyme-mediated epigenetic modification pathway [113].

In contrast, the neuromodulatory mechanisms activated by nicotine in tobacco smoke may exert early-stage intervention in Parkinson’s disease pathology by inhibiting dopamine transporter (DAT) function and increasing synaptic dopamine concentrations [114]. These neurotransmitter alterations are often accompanied by long-term adjustments in histone acetylation and non-coding RNA (ncRNA) expression, indicating a close relationship between neuronal activity and chromatin plasticity [117].

In addition to neurotransmitters and metals, environmental toxins, particularly neurotoxic pesticides such as paraquat and dieldrin, heavy metals including lead, arsenic, and aluminum, as well as certain organic solvents, are also important contributors to neurodegenerative diseases [117,118]. These toxic substances can impair mitochondrial complex function, leading to ATP depletion and excessive production of ROS [119]. When acting together with dopamine and metal ions, they can trigger Fenton and Haber reactions that result in oxidative damage to DNA, proteins, and lipids [118].

Oxidation of guanine at CpG sites, producing 8-oxo-2′-deoxyguanosine (8-oxodG), inhibits methylation of neighboring cytosine residues, which may lead to site-specific hypomethylation. At the same time, abnormal promoter methylation has been observed in several disease-related genes, such as amyloid precursor protein (APP), tau protein (Tau), glycogen synthase kinase 3 beta (GSK3β), and alpha-synuclein (SNCA), with SNCA hypomethylation being a representative example [118].

It is noteworthy that early exposure to these toxicants may induce persistent epigenetic alterations during critical developmental stages, which can later be reactivated in adulthood or aging. This phenomenon may help explain why some individuals develop neurodegenerative pathology following the same toxic exposure, whereas others remain unaffected, emphasizing the individual variability of epigenetic regulation in mediating the conversion of environmental exposure into disease risk [119,120].

In addition, aging itself is one of the most significant risk factors for neurodegenerative diseases, and one of its core features is the decline in epigenetic stability [117]. As the aging process progresses, global DNA methylation levels tend to decrease, leading to the gradual opening of genomic regions that were previously transcriptionally silent. At the same time, certain actively transcribed regions may become repressed due to reduced histone acetylation and enhanced heterochromatin formation [117,121].

In summary, various risk factors for neurodegenerative diseases, including nutritional metal imbalance, environmental toxin exposure, lifestyle differences, and natural aging, can induce epigenetic abnormalities that alter neuronal transcriptional activity, thereby disrupting structural stability and functional integrity [119]. Together, these processes emphasize the central regulatory role of epigenetic mechanisms in maintaining neural homeostasis and mediating neurodegenerative changes, providing an important theoretical basis and potential targets for early prediction and intervention.

Curcumin is a representative polyphenolic compound derived from the rhizome of turmeric. Its unique chemical structure consists of two aromatic phenolic rings linked by an α,β-unsaturated diketone moiety, forming a highly conjugated electronic system with lipophilic properties [191]. The chemical structure of curcumin is shown in Figure 3A.

In neurodegenerative diseases, excessive accumulation of reactive oxygen species (ROS) can lead to mitochondrial dysfunction, enhanced lipid peroxidation, and DNA damage, thereby triggering inflammatory responses and cell apoptosis [192]. Curcumin has been shown to effectively alleviate oxidative stress-related tissue damage by increasing superoxide dismutase (SOD) activity, elevating levels of reduced glutathione (GSH), and enhancing total antioxidant capacity (TAC)[193]. In addition, curcumin can regulate the methylation status of the Nrf2 promoter and the acetylation levels of histones, thereby enhancing the transcriptional activity of antioxidant genes. Through these epigenetic mechanisms, curcumin helps maintain redox homeostasis at the molecular level [151,194].

At the same time, curcumin suppresses inflammation-related signaling pathways mediated by Toll-like receptors (TLRs), including NF-κB and mitogen-activated protein kinase (MAPK) pathways, thereby reducing the production of proinflammatory cytokines such as interleukin 1 beta (IL-1β), tumor necrosis factor alpha (TNF-α), and interleukin 6 (IL-6), and preventing excessive microglial activation [195,196]. In microglial cells, curcumin inhibits the TLR4/MyD88/NF-κB signaling pathway and promotes the phenotypic transition from the M1 proinflammatory state to the M2 anti-inflammatory state, thus modulating the development of neuroinflammation and protecting neurons from sustained injury [197].

In the regulation of protein misfolding, curcumin can form noncovalent complexes with Aβ and alpha-synuclein, thereby inhibiting the formation of toxic oligomers and preventing mitochondrial damage and endoplasmic reticulum stress [198]. In addition, curcumin induces the expression of heat shock factor 1 (HSF-1), which upregulates molecular chaperone proteins such as heat shock protein 70 (HSP70), enhancing the cellular capacity to recognize, refold, and clear misfolded proteins [199].

In addition, curcumin can modulate cellular autophagy under specific pathological conditions [200]. Studies have shown that in HT-22 cells treated with amyloid beta 1–42 (Aβ1–42), curcumin slightly downregulates the expression of the autophagy-related factor Beclin-1 and reduces the formation of autophagosomes. This suppression of excessive autophagy may help alleviate Aβ-induced cytotoxicity, thereby improving cell viability and mitigating neuronal injury and degeneration [200].

In the regulation of neurogenesis and neural stem cell fate, curcumin activates classical signaling pathways such as Wnt/β-catenin and cAMP response element-binding protein/brain-derived neurotrophic factor (CREB/BDNF), thereby promoting the proliferation and directed differentiation of neural stem cells and stimulating regenerative potential under conditions such as stroke [201–204]. In addition, curcumin enhances the expression of neurogenesis-related genes, including neurogenin 1 (Neurog1), by increasing chromatin accessibility, thereby facilitating the restoration of neural function [205,206].

On the epigenetic level, curcumin participates in the neuroprotective process through multiple mechanisms. Curcumin can decrease CpG island methylation within the promoter areas of crucial genes by modulating upstream regulatory elements such as specificity protein 1 (Sp1) [152,207].

In AD model cells, curcumin effectively reduces the hypermethylation level of the neprilysin (NEP) gene promoter, thereby restoring its expression [152]. As an enzyme closely associated with the degradation of Aβ, the upregulation of NEP not only enhances the cellular capacity for Aβ clearance but also inhibits the phosphorylation of protein kinase B (AKT), subsequently blocking the activation of the downstream NF-κB pathway. This suppression leads to reduced expression of inflammatory mediators such as cyclooxygenase-2 (COX-2) and inducible nitric oxide synthase (iNOS), collectively alleviating the intracellular inflammatory state [152].

In terms of histone modifications, curcumin exhibits bidirectional regulatory effects on chromatin structure. Acting as a natural inhibitor of HATs, curcumin suppresses the activity of HAT family members such as p300 and CREB-binding protein (CBP), thereby reducing acetylation levels at key histone sites including H3K9ac and H4K5ac [154,208–210]. This resulting hypoacetylated state promotes chromatin condensation and subsequently represses the transcriptional activation of inflammation-related genes such as cyclooxygenase-2 (COX-2) and BDNF [154].

On the other hand, curcumin can inhibit certain HDACs, including HDAC1, HDAC3, and HDAC8, leading to increased acetylation levels of histones H3 and H4, which facilitates chromatin relaxation. At the same time, curcumin is also capable of suppressing HATs such as p300 and CREB-binding protein (CBP) while promoting the restoration of HDAC2 under specific conditions, thereby exhibiting an environment-dependent bidirectional regulatory effect [151].

In neuropathic pain models, this regulatory mechanism is particularly pronounced. Chronic nerve injury induces the enrichment of p300/CBP at promoter regions of pain-related genes, enhancing H3K9 acetylation (H3K9ac) and thereby promoting the transcription of sensitization-associated genes [154]. Curcumin has been shown to inhibit the binding of p300 to target gene promoters and reduce H3K9ac levels, effectively alleviating mechanical and thermal hyperalgesia [154].

During neural development and neural stem cell differentiation, curcumin modulates the balance between HAT (histone acetyltransferase) and HDAC (histone deacetylase) activities, thereby influencing histone acetylation status [211]. This regulation promotes the differentiation of neural progenitor cells toward a neuronal lineage, while reducing their tendency to differentiate into glial cells [211].

By dynamically regulating histone acetylation states, curcumin can exert both protective and inhibitory effects depending on the cellular context [151]. Within the nervous system, it demonstrates strong target specificity and functional diversity, highlighting its significant neuroprotective and therapeutic potential [151–153,212].

Curcumin has been shown to upregulate the expression and activity of the NAD⁺-dependent deacetylase SIRT1, which in turn regulates energy metabolism, oxidative stress tolerance, and inflammatory responses through the deacetylation of transcription factors such as FOXO3a, p53, and NF-κB [151–154,208,213]. This SIRT1-associated regulatory pathway contributes to the establishment of a stable neuroprotective state during chronic disease progression.

At the post-transcriptional level, curcumin exerts fine-tuned regulation over key neural pathways by modulating microRNA (miRNA) expression. Specifically, curcumin upregulates miR-146a, leading to suppression of TRAF6 and IRAK1 expression and subsequent negative regulation of the NF-κB signaling pathway. In parallel, curcumin acts synergistically with miR-101 to inhibit the expression and translation of amyloid precursor protein (APP), thereby reducing Aβ42 production [200,201,209]. The coordinated interplay among miRNAs, DNA methylation, and histone modifications collectively constitutes a multilayered and reversible epigenetic regulatory network [156].

Overall, curcumin exerts coordinated antioxidant, anti-inflammatory, and epigenetic regulatory effects in the CNS by scavenging ROS, activating Nrf2, inhibiting TLR4/NF-κB/MAPK signaling pathways, and modulating DNA methylation, histone modifications, and microRNA (miRNA) expression. Through these mechanisms, curcumin promotes neural plasticity and repair.

Although curcumin has demonstrated notable neuroprotective potential across various models of neurodegenerative diseases, its underlying mechanisms of action remain complex and insufficiently predictable. On one hand, curcumin exhibits pronounced bidirectional epigenetic modulation, which is highly dependent on cell type and microenvironmental context. This is especially evident in its regulation of histone acetylation, where its effects on histone acetyltransferases (HATs), such as p300/CBP, and histone deacetylases (HDACs) often vary across experimental systems, thereby increasing the complexity of mechanistic interpretation [151,154,208–210].

On the other hand, although curcumin’s regulation of DNA methylation at specific gene loci (e.g. the NEP promoter) has been shown to be functionally significant, there remains a lack of comprehensive genome-wide assessments. As a result, its overall capacity to remodel the epigenomic landscape is still uncertain [152]. Related neuroprotective effects and their associated epigenetic regulatory mechanisms are systematically summarized in Table 2.

Moreover, curcumin participates in the regulation of neuroinflammation and proteostasis through modulation of microRNAs such as miR-146a and miR-101. However, these miRNAs target broad gene networks, and their regulatory effects are highly influenced by factors such as developmental stage, pathological status, and brain region specificity, introducing substantial spatiotemporal variability [156,200,201,209]. While some studies suggest that curcumin may exert multifaceted effects, including the induction of autophagy and the promotion of neuroregeneration, many of these findings are derived from short-term or acute intervention models. Whether these effects can be sustained or are functionally relevant during chronic and progressive neurodegenerative processes remains to be fully validated [204].

More importantly, curcumin faces significant pharmacokinetic limitations, including poor bioavailability, rapid metabolism, and limited central nervous system exposure. In addition, variations across studies in terms of dosage, administration routes, and duration of intervention further compromise the comparability and consistency of results [214].

Taken together, future research must leverage human-derived cellular models, long-term intervention studies, and high-throughput epigenetic profiling technologies to better elucidate the context-dependent and target-specific characteristics of curcumin’s actions. Such efforts will be critical to establishing a more robust evidence base for its clinical translation.

Epigallocatechin-3-gallate (EGCG), the principal bioactive polyphenolic compound in green tea, exerts neuroprotective effects through multiple pathways that modulate key pathological processes in neurodegenerative diseases such as AD, PD, and amyotrophic lateral sclerosis (ALS) [222–224]. Overall, EGCG has been shown to delay disease progression, improve cognitive and motor functions, promote neuronal survival, and alleviate neuroinflammation [222–224]. The molecular structure of epigallocatechin-3-gallate (EGCG) is illustrated in Figure 3B.

In the regulation of neuroinflammation, EGCG exerts systemic inhibitory effects on multiple proinflammatory signaling pathways, particularly NF-κB、MAPK、TXNIP-NLRP3 pathways [156,222]. Through these mechanisms, EGCG reduces the expression of inflammatory cytokines such as interleukin-1 beta (IL-1β), tumor necrosis factor alpha (TNF-α), and interleukin-6 (IL-6), while preventing the polarization of microglia from the homeostatic M2 phenotype to the proinflammatory M1 phenotype, thereby mitigating the neurotoxic inflammatory microenvironment [156,222].

In BV2 microglial cells, EGCG significantly suppresses inflammation induced by lipopolysaccharide (LPS) and Aβ. This effect is achieved by inhibiting the generation of ROS, restoring mitochondrial membrane potential, and blocking TXNIP-mediated activation of the NLRP3 inflammasome, which together attenuate the progression of neuroinflammation [156]. Furthermore, EGCG upregulates protein kinase D1 (PKD1) while inhibiting Parthanatos by reducing poly(ADP-ribose) polymerase 1 (PARP-1) and apoptosis-inducing factor (AIF) levels, thereby alleviating oxidative stress and inflammation in the substantia nigra, promoting dopaminergic neuronal survival, and improving motor dysfunction in PD rat models [157].

Mitochondrial dysfunction represents a critical pathological feature of neurodegenerative diseases, and EGCG has demonstrated reparative effects on multiple fronts. EGCG enhances ATP synthesis efficiency, reduces mitochondrial reactive oxygen species (mtROS) production, and promotes mitochondrial biogenesis by activating key regulators such as nuclear respiratory factor 1 (NRF1) and mitochondrial transcription factor A (TFAM), thereby improving cellular metabolic resilience [225]. In addition, EGCG activates autophagic pathways, contributing to the maintenance of protein homeostasis and mitigating oxidative stress–induced mitochondrial damage [159].

In rat models of PD, EGCG protects dopaminergic neurons in the substantia nigra and alleviates motor dysfunction by upregulating the protein kinase D1 (PKD1) signaling pathway. This regulation inhibits the occurrence of parthanatos and reduces oxidative stress and inflammatory responses, thereby exerting significant neuroprotective effects [157].

In vitro experiments, EGCG binds to the RNA recognition motif of TDP-43 and stabilizes its structure, significantly delaying the aggregation process and reducing the formation of amyloid aggregates. This finding provides molecular-level evidence supporting the potential therapeutic application of EGCG in neurodegenerative diseases such as ALS and frontotemporal dementia (FTD) [158].

In models of MS, EGCG modulates CD4⁺ T cell subsets by suppressing Th1 and Th17 cells while enhancing regulatory T (Treg) cell activity. This immune modulation attenuates myelin-specific inflammatory responses, reduces inflammatory infiltration and demyelination in the CNS, and ultimately alleviates neuronal and axonal damage, leading to improvement in disease symptoms [159].

In the regulation of protein misfolding and aggregation, EGCG exhibits a unique structural intervention capability, particularly in modulating the aggregation of key pathogenic proteins such as TDP-43, α-syn, and Aβ [158]. In models of ALS and frontotemporal dementia (FTD), EGCG directly binds to the RNA recognition motif (RRM) of TDP-43, stabilizing the conformation of the RRM1 domain, delaying its nucleation process, and preventing its conversion into amyloid aggregates, thereby inhibiting pathological aggregation. This interaction does not occupy the RNA-binding site but instead stabilizes the protein conformation through aromatic stacking and hydrogen bonding, forming non-aggregating complexes that reflect EGCG’s ability to modulate protein structural dynamics [158].

Molecular dynamics simulations and nuclear magnetic resonance (NMR) chemical shift perturbation experiments further confirm the interaction between EGCG and the flexible linker region as well as key aromatic residues of TDP-43, revealing the diversity and stability of this binding mode [158]. Similar mechanisms are observed in the regulation of α-syn and Aβ aggregation, where EGCG promotes the conversion of toxic oligomers into non-toxic, amorphous aggregates and disrupts beta-sheet formation, thereby protecting neurons from oligomer-induced cytotoxic stress [159].

In terms of epigenetic regulation, the neuroprotective effects of epigallocatechin-3-gallate (EGCG) are primarily reflected at the levels of histone modification and non-coding RNA regulation. Compared to its more clearly defined DNA demethylation effects observed in cancer models, the direct impact of EGCG on DNA methylation in the central nervous system remains controversial. Its functional significance may depend on specific cell types and pathological contexts [158].

In contrast, more definitive evidence indicates that EGCG exerts neuroprotective effects through the regulation of histone modifications. The key mechanism involves the inhibition of histone deacetylase (HDAC) activity or expression, leading to elevated histone acetylation levels, chromatin relaxation, and transcriptional activation of neuroprotective genes [160]. In AD models, EGCG exhibits properties similar to those of HDAC inhibitors, particularly showing pronounced inhibition of HDAC1. This inhibition results in a relaxed chromatin conformation in the promoter region of the neprilysin (NEP) gene, enhancing the recruitment of transcription factors and transcriptional complexes, thereby upregulating NEP mRNA and protein expression, promoting Aβ clearance, and improving cognitive function [160].

Under oxidative or toxic stress conditions, EGCG downregulates DNMT1 and HDAC2 while upregulating stress-responsive miRNAs such as miR-200b [147]. These changes promote histone acetylation and transcriptional activity at the promoters of antioxidant genes, accompanied by global DNA hypomethylation, thus creating an epigenetic environment favorable for neuroprotection [147]. Furthermore, studies suggest that EGCG-mediated inhibition of HDAC2 is closely associated with increased histone acetylation at the promoters of antioxidant-related genes, such as superoxide dismutase 2 (SOD2) and glutathione peroxidase 1 (GPX1), thereby enhancing cellular antioxidant defenses [147].

In addition to DNA methylation and histone modification, the regulation of non-coding RNAs by EGCG represents an important component of its epigenetic effects. Multiple studies have shown that EGCG can reshape miRNA expression profiles, thereby modulating key pathological processes such as Aβ deposition, oxidative stress, and inflammatory responses at the post-transcriptional level [226]. Among these, miR-125b has emerged as a central regulatory node, displaying pronounced model- and context-dependent effects. In SH-SY5Y cells and APP/PS1 transgenic mice, EGCG treatment downregulates miR-125b expression, which is accompanied by cognitive improvement. Conversely, in Neuro2a cells and in the serum of certain transgenic mouse models, an upregulation of miR-125b has been observed, suggesting that EGCG exerts direction-dependent modulation of this miRNA, potentially influenced by cell type, disease stage, and stress conditions [226].

Meanwhile, studies have also found that EGCG exerts a restorative effect on the regulation of molecules such as miR-9 and miR-191–5p in the serum of APP/PS1 mice, further highlighting its systemic impact on circulating miRNA levels [227].

This type of epigenetic modification affects not only gene expression within neurons but also the fate determination of immune cells. For example, it can enhance the expression of the transcription factor forkhead box P3 (Foxp3), which is associated with regulatory T (Treg) cells, thereby increasing immune tolerance and attenuating the spread of central inflammation in multiple sclerosis (MS) [159,161].

Overall, as a natural polyphenol, EGCG can influence the initiation and development of neurodegenerative disorders through multiple mechanisms, including epigenetic regulation and modulation of signaling pathways. Its comprehensive regulatory capacity provides strong support for future precision interventions, disease course attenuation, and functional improvement.

Table 2: Neuroprotective and Epigenetic Mechanisms of Epigallocatechin-3-Gallate (EGCG) in Neurodegenerative Diseases

As a natural epigenetic modulator with multi-target activity, epigallocatechin-3-gallate (EGCG) has attracted considerable attention for its neuroprotective potential in neurodegenerative diseases. However, current studies exhibit substantial divergence in terms of result consistency and reproducibility. On one hand, most mechanistic investigations of EGCG rely heavily on BV2 microglial cells, transgenic animal models, or acute injury paradigms. Marked differences in model construction, stress conditions, and treatment protocols across studies have led to pronounced variability in findings related to key regulatory processes such as miRNA modulation and DNA methylation [156,157,159]. For example, contradictory effects of EGCG on the expression of miR-125b and miR-132 have been reported, suggesting that its regulatory outcomes may be highly dependent on specific cellular metabolic states, pathological environments, or experimental conditions [147,226,227]. Related neuroprotective effects and their associated epigenetic regulatory mechanisms are systematically summarized in Table 3.

Furthermore, although EGCG is proposed to participate in DNA methylation regulation through DNMT1 inhibition or by altering the SAM/SAH ratio, its epigenetic mechanisms, while theoretically plausible, lack systematic validation regarding stability and disease specificity within the nervous system [147,158]. The potential of EGCG to interfere with pathological protein aggregation is largely attributed to its non-covalent interactions with misfolded proteins such as alpha-synuclein, amyloid beta, and TDP-43. However, most of these observations originate from in vitro molecular-level studies, and whether such interactions exhibit sufficient specificity and sustained efficacy within the complex intracellular environment remains to be conclusively demonstrated [158,159].

From a pharmacokinetic perspective, EGCG is limited by poor oral absorption, rapid metabolism, and restricted central nervous system exposure, which significantly constrain its translational potential. In addition, substantial heterogeneity across studies in terms of dosage, intervention duration, and experimental animal strains poses further challenges for cross-study comparison and mechanistic synthesis [156,157,159].

Therefore, future research should prioritize quantitative investigations based on dose–response relationships, combined with standardized experimental designs and multicenter validation, to enhance the scientific rigor and clinical translational value of EGCG-related studies [156,157,159].

RSV exhibits remarkable neuroprotective effects in the nervous system. It has been reported that it not only suppresses oxidative stress and inflammatory responses and enhances cell survival but also improves neuronal function by modulating mitochondrial activity and dynamics [162]. In AD models, RSV activates AMPK and inhibits mTOR, thereby promoting the clearance of abnormal proteins such as Aβ. At the same time, through SIRT1-dependent regulation, RSV improves transcriptional and metabolic responses, helping to maintain neuronal survival and cognitive function [162].

The most well-characterized epigenetic feature of resveratrol (RSV) is its ability to activate the NAD⁺-dependent deacetylase SIRT1. By enhancing SIRT1 activity, RSV regulates the acetylation status of various transcription factors and chromatin-associated proteins, thereby reshaping neuronal stress responses and inflammatory states at the transcriptional level [164].

At the level of DNA methylation, RSV inhibits the activity of DNMTs, including DNMT1, DNMT3a, and DNMT3b, thereby significantly reducing the abnormal hypermethylation of promoters of neuroprotective genes such as estrogen receptor alpha (ERα) and restoring their normal expression [163,233]. This process is directly linked to neuronal synaptic plasticity, trophic support, and survival capacity.

At the same time, RSV plays a crucial role in regulating histone modifications. RSV markedly increases active histone marks such as H3K9ac and H3K27ac, enhancing the transcriptional accessibility of anti-inflammatory genes and neurotrophic factors [163]. This action reverses chromatin repression caused by chronic inflammation or aging, thereby maintaining neuronal functional stability [163].

Resveratrol (RSV) also exhibits distinct advantages in regulating neuronal cell cycle and aging-related pathways. Studies have shown that RSV can inhibit the expression and activity of TET2, thereby increasing DNA methylation at the CDKN2A promoter region and reducing the expression of the cell cycle inhibitor p16^INK4a [165]. This regulation alleviates the inhibition of CDK4, promotes the phosphorylation and functional restoration of retinoblastoma protein (pRb), and ultimately reduces aberrant neuronal cell cycle re-entry and susceptibility to apoptosis [165]. This mechanism provides important epigenetic evidence for the neuroprotective effects of RSV in the context of neuronal aging and neurodegenerative processes.

In inflammatory conditions, resveratrol (RSV) suppresses the transcriptional activity of the NF-κB p65 (RelA) subunit through SIRT1-mediated deacetylation, leading to a significant reduction in the expression of pro-inflammatory cytokines such as IL-1β, TNF-α, and IL-6. This, in turn, inhibits microglial overactivation and alleviates the sustained stimulation of the NLRP3 inflammasome [166,167,229].

At the post-transcriptional level, resveratrol (RSV) further refines its epigenetic regulatory network by remodeling microRNA (miRNA) expression profiles. RSV upregulates miR-132 and miR-124, which are closely associated with neuronal plasticity and dendritic spine formation, thereby enhancing synaptic connectivity and cognitive function [168,234]. Concurrently, it downregulates pro-inflammatory or pro-apoptotic miRNAs such as miR-34a and miR-155, thereby relieving their suppressive effects on SIRT1, antioxidant pathways, and neurotrophic factors [169,170]. This miRNA–SIRT1 co-regulatory mechanism contributes to the maintenance of neuronal functional homeostasis under chronic pathological conditions.

Overall, resveratrol (RSV) exerts its neuroprotective effects in neurodegenerative diseases through a SIRT1-centered epigenetic regulatory axis, which coordinately modulates DNA methylation, histone modifications, inflammatory signaling, and miRNA networks. This multi-target, reversible, and highly adaptable mode of action highlights RSV’s potential in both the prevention and treatment of neurodegenerative disorders, offering promising avenues for future therapeutic development.

The therapeutic potential of resveratrol (RSV) in neurodegenerative disease intervention has been widely recognized, particularly for its ability to activate the SIRT1 signaling axis involved in chromatin remodeling and inflammation regulation [164]. However, this mechanism-dependent feature serves as both a key strength and a significant limitation [162,229–231]. On one hand, SIRT1, as a metabolism-sensitive deacetylase, exhibits considerable variability in expression levels and functional status across different brain regions, cell types, and stages of disease progression. As a result, the biological effects of RSV are characterized by substantial individual variability and context dependence [162,229–231]. Related neuroprotective effects and their associated epigenetic regulatory mechanisms are systematically summarized in Table 4.

On the other hand, although several studies have reported RSV’s influence on DNA methylation, the TET2 pathway, and key factors such as p16^INK4a, most of this evidence is confined to candidate genes or specific signaling nodes. There remains a lack of systematic, genome-wide, and temporally resolved analyses based on large sample sizes, limiting our understanding of its broader epigenetic impact [163,165,233].

At the level of non-coding RNAs, RSV modulates multiple miRNA networks to finely regulate neuroinflammation and synaptic plasticity. However, these miRNA pathways are often involved in a wide array of physiological and pathological processes, making the regulatory network highly complex. Potential off-target effects and directional reversals of regulation at different disease stages introduce further uncertainty in mechanistic interpretation [168–170,234].

From a translational medicine perspective, RSV also faces several pharmacokinetic limitations, including poor water solubility, low bioavailability, and rapid metabolism, which severely restrict its effective exposure in the central nervous system [229,232]. Although advanced delivery strategies such as liposomal encapsulation and prodrug design have been proposed in recent years, most of these efforts remain at the proof-of-concept stage. Comprehensive toxicological evaluations and validation in large animal models are still lacking.

Moreover, substantial variation across studies in terms of formulation selection, administration routes, and intervention duration further complicates the long-term assessment of RSV’s safety and efficacy.

In summary, to advance the clinical translation of RSV in neurodegenerative diseases, it is essential not only to deepen our understanding of the boundaries of SIRT1-dependent regulation, but also to pursue more systematic and refined research in areas such as delivery technology optimization, population heterogeneity analysis, and the design of long-term follow-up studies.

From a translational medicine perspective, the largely negative or limited efficacy observed in the aforementioned clinical studies may be influenced by multiple factors. First, polyphenolic compounds generally exhibit low oral bioavailability and rapid metabolism in vivo. Different polyphenols also show considerable variation in absorption efficiency, plasma half-life, and formulation dependence. Insufficient systemic exposure has been identified as one of the key factors limiting their clinical effectiveness [243].

For instance, studies have shown that curcumin has extremely low oral bioavailability due to its high hydrophobicity, poor intestinal absorption, and significant first-pass metabolism. Free curcumin is almost undetectable in plasma, and its elimination half-life is short. Although novel formulations such as nanoemulsions and liposomes can significantly enhance its relative bioavailability, systemic exposure under conventional oral preparations remains very limited [214]. Similarly, resveratrol, despite having relatively high absorption, undergoes rapid glucuronidation and sulfation in the enterohepatic circulation, resulting in very low plasma concentrations of the parent compound. Its absolute bioavailability is generally less than 1% and it is rapidly metabolized and excreted [244]. Taking EGCG as an example, its plasma exposure levels, such as AUC and Cmax, are highly influenced by dosing conditions and formulation types, indicating that its systemic availability depends heavily on intestinal absorption and the surrounding chemical environment. Overall, plasma concentrations remain low, reflecting clear limitations in its absorption and pharmacokinetics [245].

In addition, differences in physicochemical properties among various polyphenol subclasses, such as lipophilicity, molecular weight, and polarity, lead to significant differences in their ability to cross the blood-brain barrier. This further contributes to the inconsistency in clinical research outcomes [246]. Although curcumin, resveratrol, and EGCG have been reported to possess certain blood-brain barrier permeability, quantitative studies have shown that their actual concentrations in brain tissue are generally low. There is also marked heterogeneity across different polyphenol types and disease stages, making it difficult to achieve sufficient pharmacological concentrations in the central nervous system, which affects their therapeutic effects [240,247,248].

Current in vivo studies suggest that while most polyphenols or their conjugated metabolites can cross the blood-brain barrier, their actual concentrations in the central nervous system typically remain at the nanomolar level. For example, even in Alzheimer’s disease clinical studies involving high-dose, long-term oral administration of resveratrol, the concentration of the parent compound in cerebrospinal fluid remained at a low nanomolar level, with a higher abundance of conjugated metabolites [249]. This indicates that possessing central permeability does not necessarily equate to achieving adequate pharmacological accumulation in the brain. In the case of catechins, tissue distribution studies in animal models have also shown that even with strategies such as nano-delivery, EGCG concentrations in brain tissue or cerebrospinal fluid may still be as low as approximately 1 nanomolar. For instance, reported brain levels of EGCG were about 0.6 nanograms per milliliter, which converts to around 1.3 nanomolar, suggesting that there is a significant upper limit to brain exposure [250]. These in vivo exposure levels are much lower than the effective concentrations typically used in in vitro experiments. Many cell studies employ micromolar or even higher concentrations to observe antioxidant, anti-inflammatory, or anti-aggregation effects [251]. The large gap between effective concentrations in vitro and the achievable concentrations in vivo is likely one of the key limiting factors preventing the consistent clinical translation of the neuroprotective effects of polyphenols [252].

Moreover, increasing evidence indicates that polyphenols are more commonly present in the brain in the form of conjugated metabolites, such as glucuronides and sulfates. These metabolites also tend to be present at low concentrations, usually within the nanomolar range [249]. This further complicates the interpretation of central pharmacological mechanisms based solely on the parent compounds.

Microbiota-mediated metabolic regulation is a critical yet often overlooked factor influencing the systemic exposure and central delivery of polyphenols. The biological effects of polyphenols in vivo largely depend on their metabolic transformation by the gut microbiota, which generates phenolic acid metabolites that may differ significantly from the parent compounds in terms of absorption efficiency and blood-brain barrier permeability [253]. After ingestion, dietary polyphenols are typically present in conjugated forms and are only partially absorbed in the small intestine [246]. A substantial portion reaches the colon, where they serve as substrates for microbial metabolism. Through microbial enzymatic activity, polyphenols are converted into structurally simpler, low-molecular-weight phenolic metabolites that are more readily absorbed into the systemic circulation [246].

Once in the bloodstream, both polyphenols and their microbial metabolites can undergo further Phase II metabolism in the liver, affecting their distribution and exposure profiles within the body [246]. Current evidence suggests that only certain gut microbiota-derived phenolic metabolites possess the capacity to cross the blood-brain barrier, and their accessibility to brain tissues is further constrained by physicochemical properties and transport mechanisms [246]. Moreover, interindividual variability in gut microbiota composition has been identified as a potential source of the high heterogeneity observed in the clinical efficacy of polyphenols [254].

Additionally, considerable differences across clinical studies in terms of dosage, formulation types, and duration of intervention contribute to further heterogeneity in outcomes and introduce uncertainty in efficacy assessment [255].

In terms of safety, polyphenolic compounds are not entirely devoid of risk. Some studies have indicated that under high doses or specific conditions, certain polyphenols may exhibit pro-oxidant activities and potentially interfere with drug-metabolizing enzymes or transporters, thereby posing a risk of drug-drug interactions [256]. Hepatotoxicity associated with green tea extracts and EGCG has been reported in clinical settings. For instance, elevated liver enzymes were observed in the PROMESA study, suggesting that long-term or high-dose administration may carry certain safety concerns [242,257].

Furthermore, polyphenols typically exert multi-target effects, involving a wide range of biological pathways such as antioxidant defense, anti-inflammatory activity, metabolic regulation, and mitochondrial function. The complexity of these mechanisms makes it difficult to attribute observed clinical or molecular outcomes to a single epigenetic regulatory pathway, thereby increasing the challenges and uncertainties in mechanistic interpretation [258].