Natural Product-Derived Carbon Dots in Neurodegenerative Diseases: Advances in Blood–Brain-Barrier-Related Delivery, Neuroprotection, and Theranostics Authors: Kaixin Song, Xiang Gu, Na Sun, Rujia Xie, Ziyan Chen, Zili Wang, Ya Li, Lei Meng Published: 2026 ABSTRACT Simple SummaryNeurodegenerative diseases gradually damage nerve cells in the brain, leading to a decline in memory, thinking, and movement. Because a protective barrier separates the blood from the brain, many medicines cannot easily enter brain tissue or reach effective levels, making these diseases difficult to treat. Natural product-derived carbon dots are very small, carbon-based materials derived from plants or other natural substances. They can be chemically modified to emit fluorescence for tracking purposes and may help to protect nerve cells. This review summarizes current research on their use to deliver treatments to the brain, reduce oxidative damage and inflammation, interfere with harmful protein clumping, and support imaging and monitoring. Studies in cells and animals suggest that some of these materials show promise. However, there is still insufficient evidence regarding how much of the material enters brain tissue, its safety during long-term use, consistency between production batches, and applicability to humans. More rigorous quality control, long-term safety testing, and experimental models that more accurately reflect the human body are required before these materials can be considered safe and reliable for diagnosing and treating brain diseases. 1. Introduction Neurodegenerative diseases (NDDs) are a group of chronic disorders characterized primarily by the progressive dysfunction and loss of specific neuronal populations. The major NDDs include Alzheimer’s disease (AD), Parkinson’s disease (PD), Huntington’s disease and amyotrophic lateral sclerosis (ALS) [1,2,3,4,5]. While these diseases differ in the abnormal proteins involved, affected brain regions, and clinical manifestations, they share pathological processes such as protein misfolding and aggregation, oxidative stress, neuroinflammation, mitochondrial dysfunction, synaptic damage, and neuronal death [6,7,8,9,10,11]. As populations are aging faster than ever before, NDDs have become a leading cause of cognitive and motor impairment, as well as a significant long-term care burden [12,13,14,15]. Treating them presents the dual challenge of complex pathological networks and insufficient effective exposure within the brain. Abnormal protein aggregation, oxidative damage, chronic inflammation, disrupted energy metabolism, and neurovascular unit dysfunction interact, making single-target interventions generally insufficient to address the multiple pathological processes involved in disease progression [16,17,18]. Meanwhile, the highly selective blood–brain barrier (BBB) restricts the entry of most therapeutic molecules into the central nervous system (CNS), making it difficult to achieve adequate and sustained drug exposure in affected regions [19,20]. Consequently, material-based systems capable of combining brain delivery with multimodal therapeutic intervention have attracted considerable attention [21,22]. Carbon dots (CDs) are fluorescent nanomaterials characterized by carbon-based nanostructures and tunable surface chemistry [23]. In this review, ‘CDs’ is used as an umbrella term encompassing carbon quantum dots (CQDs), graphene quantum dots, carbonized polymer dots, and related materials. When the structure or nomenclature of a material is explicitly defined in the original study, the specific designation is retained. The size, surface chemistry, and charge characteristics of CDs can influence their interactions with biological interfaces and the BBB, thereby conferring potential for delivery, tracking, and imaging applications [24,25]. However, small particle size does not necessarily indicate effective BBB penetration, and fluorescence signals detected in brain tissue do not demonstrate that the material has entered the brain parenchyma or achieved therapeutically relevant exposure in disease-associated regions [26,27]. Natural product-derived CDs can be prepared from medicinal plants, naturally occurring bioactive molecules, and other natural raw materials. Nevertheless, natural origin does not guarantee a well-defined composition, a favorable safety profile, or clinical translatability of the final material. Carbonization, post-processing, and purification can alter the material’s composition and complicate the attribution of biological effects. Therefore, the structure and function of the resulting CDs cannot be inferred solely from the properties of their precursors [28,29,30,31,32]. Current studies have predominantly focused on AD-related cellular and animal models, suggesting that certain natural product-derived CDs may exert antioxidant and anti-inflammatory effects, modulate abnormal protein aggregation, and provide neuroprotection [33,34,35]. Nevertheless, evidence regarding BBB transport, brain parenchymal exposure, pharmacokinetics, and long-term safety remains insufficient. BBB permeability, brain fluorescence signals, brain parenchymal exposure, and therapeutic benefit represent distinct levels of evidence and should not be considered interchangeable [32,36,37,38,39]. Accordingly, this critical narrative review focuses on BBB-related delivery, neuroprotection, imaging applications, and the challenges of clinical translation associated with natural product-derived CDs in NDDs. Particular emphasis is placed on studies of natural product-derived CDs that provide material characterization and direct experimental evidence relating to NDDs, the BBB, or brain delivery. Other naturally derived materials, non-NDD models, model-protein studies, and composite nanosystems are included only as supplementary mechanistic or methodological references. The available evidence is critically evaluated with respect to material sources and characteristics, interactions with the BBB, brain exposure, functional outcomes, and translational feasibility. Particular attention is given to distinguishing brain-associated distribution, imaging capability, brain parenchymal exposure, and therapeutic effects while also considering key translational factors such as pharmacokinetics, long-term safety, material consistency, and manufacturability. 2. Pathophysiological Mechanisms and Therapeutic Challenges of Neurodegenerative Diseases 2.1. Common Pathological Mechanisms of Neurodegenerative Diseases and Their Therapeutic Implications NDDs are a heterogeneous group of disorders characterized by neuronal dysfunction, progressive neuronal loss, and abnormal protein deposition. While individual diseases differ in their pathological proteins, selectively vulnerable brain regions, and clinical manifestations, they commonly progress through interconnected pathological processes, including abnormal protein aggregation, oxidative stress, mitochondrial dysfunction, chronic neuroinflammation, disruption of neurovascular unit (NVU) homeostasis, and synaptic injury [17,40,41,42,43,44,45]. The core pathological features of AD include amyloid-β (Aβ) deposition and abnormal phosphorylation and aggregation of tau protein. These alterations may be accompanied by glial activation, oxidative stress, neuroinflammation, and BBB dysfunction, which collectively promote synaptic injury and neurodegeneration [46,47,48,49,50,51]. PD is primarily characterized by progressive loss of dopaminergic neurons in the substantia nigra and abnormal aggregation of α-synuclein (α-syn). Misfolding and aberrant modification of α-syn can promote the formation of toxic aggregates and Lewy bodies and, together with mitochondrial dysfunction, oxidative stress, and glial activation, can exacerbate neuronal injury [52,53,54]. Other NDDs may involve mutant huntingtin, TAR DNA-binding protein 43 (TDP-43), fused in sarcoma (FUS), or other pathological proteins, and are often associated with proteostasis disruption, cellular stress, and selective neuronal vulnerability [10,55,56,57,58,59]. Protein misfolding, aggregation, and impaired clearance are major shared mechanisms across multiple NDDs. The persistent accumulation of pathological proteins, including Aβ, tau, α-syn and TDP-43, can disrupt proteostasis and impair mitochondrial function, membrane trafficking and synaptic function. Accordingly, inhibiting abnormal protein aggregation and promoting protein clearance represent potential disease-modifying strategies [60,61,62,63,64]. Additionally, oxidative stress, mitochondrial dysfunction, and chronic neuroinflammation can reinforce each other to form self-amplifying pathological loops that further aggravate synaptic and neuronal injury. Therefore, strategies that restore redox homeostasis, preserve mitochondrial function, and alleviate neuroinflammation are also important for neuroprotection [65,66]. These interconnected pathological processes provide the principal biological rationale for investigating the anti-oxidative, anti-inflammatory, anti-aggregation, and neuroprotective effects of CDs derived from natural products. Table 1 summarizes the representative abnormal proteins, the principal affected brain regions, the characteristic pathological features, and the clinical manifestations of the major NDDs, and Figure 1 provides an overview of the associated proteinopathies and their common pathological mechanisms. 2.2. Limitations of Current Treatments and Unmet Needs NDDs are characterized by prolonged disease courses and complex pathogenic mechanisms. Current treatments primarily aim to relieve symptoms or slow disease progression, but face challenges including insufficient exposure of therapeutic molecules to the brain, limitations of single-target interventions, long-term safety concerns, and poor treatment adherence [87,88]. The BBB restricts the entry of many therapeutic molecules into the CNS, making it difficult to achieve adequate and sustained drug exposure in disease-relevant brain regions [89,90,91]. Furthermore, detecting a drug or material in brain tissue does not necessarily indicate that an effective concentration has been reached at the pathological site. Therefore, brain exposure should be evaluated comprehensively alongside pathological localization and functional outcomes [92,93]. As multiple pathological processes interact, interventions targeting a single molecule or pathway are generally insufficient to address the complex networks underlying disease progression. Therapeutic evaluation should also incorporate neuroprotective effects and disease-relevant functional outcomes [17,45,94]. NDDs often require long-term or repeated treatment, and treatment sustainability may be affected by adverse effects, the burden of administration, and patient adherence. Increasing the dose when brain exposure is insufficient may enhance peripheral exposure and systemic toxicity without necessarily producing a corresponding therapeutic benefit in the CNS [95,96,97]. For nanosystems designed for brain delivery, attention should also be given to the potential neurotoxicity and inflammatory responses associated with long-term material accumulation [98]. Therefore, emerging therapeutic systems for NDDs must improve effective brain exposure, address multiple pathological processes, and ensure safety during long-term use. 2.3. Impact of Blood–Brain Barrier and Neurovascular Unit Dysfunction on Treatment The BBB is a key interface that maintains homeostasis within the CNS. Primarily composed of brain microvascular endothelial cells and their tight junctions, it also includes pericytes, the basement membrane, astrocytic endfeet, and neurons, collectively constituting the NVU [99,100]. Under physiological conditions, the BBB restricts the entry of foreign substances into the brain parenchyma, thereby preserving CNS homeostasis via tight junctions, low transcytosis rates, and selective transport and efflux systems [101,102]. In NDDs, however, the BBB/NVU can undergo structural and functional alterations, including impaired endothelial barrier function, disrupted tight junctions, reduced pericyte support, and altered glial reactivity [103,104,105]. These changes can compromise barrier integrity and disrupt the brain microenvironment, allowing harmful components and inflammatory signals in the blood to abnormally enter brain tissue, thereby aggravating neuroinflammation, synaptic dysfunction, and neuronal injury [105,106]. Therefore, BBB/NVU dysfunction not only affects the brain distribution of therapeutics but also represents an important pathological process that contributes to disease progression. It is important to note that enhancing a therapeutic system’s ability to cross the BBB and repairing pathological BBB damage are distinct objectives. The former aims to improve effective brain exposure while preserving the barrier’s selective properties, whereas the latter focuses on restoring barrier integrity and brain homeostasis [107,108]. Pathological increases in BBB permeability are commonly accompanied by vascular inflammation and barrier injury and should not be interpreted solely as improved delivery efficiency [105]. Therefore, when evaluating brain-delivery systems such as natural product-derived CDs, assessments should extend beyond brain distribution to include effective brain exposure, localization within pathological regions, BBB/NVU integrity, and disease-relevant functional outcomes. Current in vitro BBB models cannot fully reproduce the structure and function of the NVU, and interspecies differences in barrier properties further limit the direct extrapolation of experimental findings to humans [109,110,111]. Evidence of brain delivery from animal studies should therefore be evaluated hierarchically, alongside brain parenchymal exposure and localization within pathological regions. Whole-brain fluorescence or total tissue content alone is insufficient to demonstrate that a material has effectively crossed the BBB [112,113]. 3. Natural Product-Derived Carbon Dots: Physicochemical Basis, Blood–Brain-Barrier-Related Mechanisms, and Evidence Boundaries 3.1. Physicochemical Characteristics and Biological Basis of Natural Product-Derived Carbon Dots CDs derived from natural products have attracted increasing attention in nanomedicine due to their nanoscale dimensions, tunable surface chemistry, fluorescence properties, and potential bioactivity [25,28,114]. These CDs are typically smaller than 10 nm [115]. Their small size may facilitate dispersion, cellular uptake and tissue diffusion, while their carbon core structures, surface states and heteroatom doping may confer fluorescence imaging capability and potential biological activity [116,117,118]. However, effective brain delivery requires more than just small particle size and fluorescence; their in vivo behavior is also influenced by factors such as surface chemistry, protein corona formation, route of administration, and the status of biological barriers [113,119]. The surfaces of natural product-derived CDs commonly contain functional groups, such as hydroxyl, carboxyl, and amino groups. These groups can improve water solubility and chemical modifiability and provide reactive sites for the conjugation of drugs or targeting ligands [120,121]. Defects in the carbon core, surface functional groups, and heteroatom doping may contribute to free-radical scavenging, enabling certain CDs to exhibit antioxidant and neuroprotective enzyme-like activity [122,123]. Nevertheless, most current studies rely primarily on conventional physicochemical and spectroscopic characterization. The relationships between structural heterogeneity, surface composition, defect states and the optical and biological effects of CDs are not fully understood, and their structure–property–activity relationships remain to be elucidated [124,125]. Natural product-derived CDs are commonly prepared from medicinal plants, food-derived materials, naturally occurring bioactive small molecules, or biomacromolecules using hydrothermal, solvothermal, or pyrolytic methods [126,127,128]. However, a natural origin does not necessarily indicate low toxicity, retention of the precursor’s pharmacological activity, or batch-to-batch stability. The composition of natural precursors can vary with geographical origin, harvest period, and extraction method, and carbonization can substantially alter their original constituents. Inadequate purification may also leave residual precursors or molecular by-products that interfere with material characterization and the attribution of biological effects [39,129,130,131]. Therefore, studies of natural product-derived CDs should be based on clearly defined material identity, adequate purification, consistent batch-to-batch performance, and well-established structure–function relationships. Table 2 summarizes the precursor categories, representative structural features, and reported biological functions of natural product-derived CDs, and Figure 2 depicts their multifunctional bioactivities and potential BBB-related transport mechanisms. 3.2. Interactions of Natural Product-Derived Carbon Dots with the Blood–Brain Barrier and Potential Transport Mechanisms The potential applications of CDs derived from natural products in NDDs are closely associated with their interactions with the BBB and their distribution within the brain [24]. Their water solubility, colloidal stability, protein corona formation, and interactions with brain microvascular endothelial cells can be influenced by particle size, surface charge, and surface functional groups, including hydroxyl, carboxyl, and amino groups [102,132,157,158,159,160]. This affects cellular uptake and in vivo distribution. Some natural product-derived CDs have been reported to exhibit brain-associated distribution or interactions with the BBB. However, small size, endothelial uptake, or signals detected in brain tissue do not directly demonstrate quantifiable and reproducible exposure within the brain parenchyma [112,113,161]. Receptor- or carrier-mediated transport provides a mechanistic basis for designing CD-based brain delivery systems. Transferrin-modified CDs may enhance interactions with the transferrin receptor [157], whereas receptor-mediated pathways involving the lactoferrin receptor also provide a basis for the brain-targeted functionalization of CDs [162,163,164]. Tryptophan-derived CDs may be taken up via L-type amino acid transporter 1 (LAT1), while glucose- or trehalose-functionalized carbonized polymer dots may be associated with glucose transporter 1 (GLUT1)-related pathways [161,165]. Nevertheless, receptor binding, endothelial uptake, or increased permeability in vitro alone cannot demonstrate that a material has crossed an intact BBB and reached an effective concentration in the brain parenchyma. BBB injury and altered permeability under pathological conditions may also affect the brain distribution of CDs. Neuroinflammation or acute brain injury can disrupt tight junctions and increase barrier leakage, thereby facilitating the passive entry of materials. However, such distribution should not be equated with active transport across an intact BBB under physiological conditions [102,166,167]. Overall, the interactions of natural product-derived CDs with the BBB and their distribution within the brain are influenced by multiple factors, including the characteristics of the material’s surface, interactions with the biological interface, and the pathological status of the barrier. BBB interactions, in vitro permeability, brain tissue signals, brain parenchymal exposure, and therapeutic effects represent distinct levels of evidence and should not be considered interchangeable. Future studies should integrate confirmation of material identity, quantitative assessment of brain distribution and validation of transport mechanisms, establishing an evidence chain linking ‘material characteristics–brain parenchymal exposure–functional effects’. 4. Advances in the Application of Natural Product-Derived Carbon Dots in Neurodegenerative Diseases NDDs involve multiple pathological processes, including oxidative stress, neuroinflammation, and abnormal protein aggregation. Interventions targeting a single pathway often fail to achieve sufficient therapeutic efficacy [45,168]. Natural product-derived CDs may exert neuroprotective effects by modulating oxidative stress and inflammatory responses and inhibiting abnormal protein aggregation. Their BBB-related delivery and fluorescence imaging properties also have potential applications in theranostics [33,169]. This narrative review therefore summarizes recent advances in natural product-derived CDs for NDDs, emphasizing their principal mechanisms of action and major application areas (Figure 3). 4.1. Antioxidant Effects and Regulation of Inflammation Oxidative stress and chronic neuroinflammation are two major mechanisms that underlie the progression of NDDs. These processes can reinforce each other, jointly contributing to mitochondrial dysfunction, synaptic injury, and neuronal death [170,171,172]. Natural product-derived CDs have tunable carbon core structures, surface functional groups, and defect sites that may enable them to scavenge ROS, regulate redox homeostasis, and modulate inflammation-related signaling pathways [173,174]. Accordingly, natural product-derived CDs have demonstrated preclinical potential to mitigate oxidative damage and modulate the inflammatory microenvironment associated with NDDs. However, it should be noted that reductions in ROS levels or inflammatory mediators primarily reflect mechanistic effects in specific experimental models and do not directly demonstrate effective brain exposure or the ability to delay long-term disease progression [36,38,175]. This section discusses the neuroprotective potential of natural product-derived CDs from two perspectives: their antioxidative effects and their ability to modulate inflammatory responses. 4.1.1. Antioxidant Effects In NDDs such as AD and PD, excessive accumulation of ROS can induce lipid peroxidation, oxidative protein damage, mitochondrial dysfunction, and neuronal apoptosis. This contributes to neuronal injury and disease progression [172,176]. Studies have shown that some CDs derived from natural products possess free-radical-scavenging activity and can alleviate oxidative stress by reducing ROS levels. Certain CDs have also been reported to exhibit antioxidant enzyme-mimetic activity, participating in the regulation of redox homeostasis by catalytically removing ROS [172,177,178,179]. The antioxidant activity of CDs derived from natural products is closely associated with their carbon-core structures, surface functional groups, defect sites, and heteroatom doping. For instance, Polygonatum–Gastrodia-derived CDs may exhibit free-radical-scavenging capacity and antioxidative activity due to their oxygen-containing surface groups, defect sites, and carbon-core structures, which serve as potential active centers [180]. Heteroatom doping can also modulate the electronic structure and surface active sites of CDs, enhancing their ability to scavenge ROS and related reactive species, and improving their antioxidant activity [122]. In addition to directly scavenging ROS, natural product-derived CDs may exert protective effects by regulating intracellular antioxidant defense systems. For instance, Radix Isatidis-derived CDs have been reported to cross the BBB and reduce ROS levels, suppress neuroinflammation and Aβ aggregation, and improve AD-like pathological phenotypes in a mouse model [33]. Epigallocatechin gallate (EGCG)-derived carbonized polymer dots, curcumin-derived CDs, caffeic acid-derived CDs, and vitamin C-derived CDs have also been reported to scavenge ROS, alleviate oxidative stress, modulate abnormal protein aggregation, or reduce injury in neural cells [34,174,181,182]. Overall, the antioxidant effects of natural product-derived CDs mainly involve the direct removal of excessive ROS and the regulation of endogenous antioxidant enzyme systems. These effects may alleviate oxidative stress-related cellular injury and provide a material basis for antioxidant interventions in NDDs. However, apart from a limited number of AD-related animal studies, the available evidence largely concerns changes in ROS levels, antioxidant enzyme activity, and cellular injury markers. These findings cannot be equated directly with long-term disease-modifying effects and require further validation through assessments of brain distribution, pharmacokinetics, dose–response relationships and long-term safety. 4.1.2. Regulation of Inflammatory Responses Chronic neuroinflammation is an important underlying mechanism of the sustained progression of NDDs. Aberrant microglial activation, release of pro-inflammatory cytokines, and inflammasome activation can collectively promote neuronal injury and synaptic dysfunction. CDs derived from natural products have shown potential to modulate neuroinflammatory processes by regulating inflammation-related signaling pathways, suppressing the excessive activation of inflammatory cells and reducing the expression of pro-inflammatory mediators [183,184,185]. Previous studies have demonstrated that CDs derived from natural products or constructed using naturally occurring bioactive compounds can reduce the levels of pro-inflammatory mediators such as tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), interleukin-6 (IL-6), and nitric oxide (NO), thereby alleviating inflammation-associated cellular injury [186,187,188]. For instance, red-fluorescent cerium-based CDs prepared using curcumin have been shown to reduce the expression of TNF-α and IL-6 in BV-2 microglial cells, indicating their capacity to suppress inflammatory responses in microglia [187]. EGCG-derived CQDs have also been reported to reduce inflammatory cytokine release and alleviate neuroinflammation-associated injury [188]. Regarding inflammatory signaling pathways, nuclear factor kappa B (NF-κB) is a key regulator of neuroinflammatory responses. Its activation can promote the assembly of the NOD-like receptor family pyrin domain-containing 3 (NLRP3) inflammasome and induce the release of inflammatory cytokines, such as IL-1β and interleukin-18 (IL-18) [189,190,191,192,193]. Accordingly, modulation of the NF-κB/NLRP3 inflammatory axis is an important strategy for alleviating neuroinflammation. Studies have demonstrated that purple sweet potato-derived CDs can inhibit the activation of the Toll-like receptor 4 (TLR4)/NF-κB/NLRP3 signaling axis in lipopolysaccharide (LPS)-induced inflammatory models. They can also downregulate pro-inflammatory cytokines, including IL-1β, IL-6, and TNF-α, and attenuate pyroptosis-related responses [186]. In an intracerebral hemorrhage model, EGCG-based carbon nanodots inhibited components of the inflammatory cascade, including NF-κB, NLRP3, caspase-1, and gasdermin D (GSDMD), reducing pro-inflammatory cytokine levels and ameliorating inflammation-associated neurological injury [188]. Overall, the anti-inflammatory effects of CDs derived from natural products are manifested in three ways: suppression of excessive activation of microglia and other inflammatory cells, reduction in pro-inflammatory mediators such as TNF-α, IL-1β, IL-6, and NO, and regulation of inflammation-related signaling pathways including the TLR4/NF-κB/NLRP3 axis. This attenuates inflammatory cascades and pyroptosis. Current studies provide preliminary experimental evidence that natural product-derived CDs can improve the neuroinflammatory microenvironment. However, most of the available evidence has been obtained from BV-2 cells, LPS-induced inflammatory models, epilepsy models, or acute brain injury models, and therefore cannot be equated directly with long-term therapeutic effects in chronic NDDs such as AD and PD. Future studies should integrate assessments of brain exposure, glial phenotypic transitions, dose–response relationships, and long-term safety to further clarify the relevance of their anti-inflammatory effects to disease and their translational value. 4.2. Inhibition of Abnormal Protein Aggregation and Regulation of Proteostasis Protein misfolding, aggregation, and impaired clearance are major pathological features of NDDs [194,195]. In diseases such as AD, PD, and ALS, abnormal proteins including Aβ, tau, α-syn, and TDP-43 can form toxic oligomers and fibrillar aggregates. These aggregates subsequently induce synaptic injury, mitochondrial dysfunction, oxidative stress, and neuroinflammation [196,197,198,199]. Recent studies have shown that CDs derived from naturally occurring bioactive molecules or natural products, together with related functionalized carbon-based quantum dots, may interfere with pathological protein aggregation. This is achieved by inhibiting the initiation of aggregation and fibril elongation, promoting aggregate disassembly, reducing aggregate-associated toxicity, and regulating proteostasis [34,200]. It should be noted that these studies vary considerably in the sources of the materials used, their surface structures, and the disease models employed. Some of the evidence is derived from non-natural-product-derived or functionalized CD systems and is therefore more appropriately regarded as mechanistic evidence than as direct therapeutic evidence for natural product-derived CDs in NDDs. 4.2.1. Inhibition of Protein Aggregation Initiation and Fibril Elongation The toxicity of proteins associated with NDDs is closely linked to their misfolding, oligomer formation, and subsequent fibrillization [201,202]. Some CDs can interact with protein monomers or oligomers during the early stages of aggregation, thereby preventing further conversion into mature fibrils and inhibiting abnormal protein aggregation. For instance, sodium citrate-derived CQDs have been reported to prevent the conversion of Aβ monomers and oligomers into mature fibrils and to promote oligomer dissociation, thereby suppressing amyloid aggregation [203]. While this system is not strictly classified as a natural product-derived CD, it provides evidence that CDs can interfere with the initial stages of Aβ aggregation. CDs derived from naturally occurring bioactive molecules have also shown potential to inhibit abnormal protein aggregation. For instance, EGCG-derived carbonized polymer dots can interact with Aβ aggregates through their aromatic hydrophobic structures and surface functional groups, such as hydroxyl and carboxyl groups. These interactions prevent the addition of further monomers to fibril ends, thereby inhibiting Aβ fibril elongation [182]. Additionally, quercetin-derived red-emissive carbon nanoparticles have been reported to inhibit Aβ aggregation and exhibit anti-amyloid fibrillization activity [204]. Taken together, these findings suggest that CDs derived from naturally occurring bioactive molecules may interfere with abnormal protein aggregation through multiple interfacial interactions, including hydrophobic, π–π, hydrogen-bonding, and electrostatic interactions. Overall, CDs derived from naturally occurring bioactive molecules and related carbon-based nanomaterials may interfere with the initiation of abnormal protein aggregation and fibril elongation. This provides a material basis for anti-aggregation therapeutic strategies. However, the current evidence is predominantly derived from in vitro protein aggregation assays and does not directly demonstrate stable anti-aggregation effects in the complex brain microenvironment or long-term disease-modifying activity. 4.2.2. Promotion of Aggregate Disassembly and Reduction in Aggregate Toxicity In addition to inhibiting the initiation of aggregation, some CDs derived from naturally occurring bioactive molecules, as well as related functionalized CDs, can promote the disassembly of mature aggregates and reduce aggregate-associated toxicity. Amino-functionalized CQDs have been shown to interact with amyloid proteins via their surface functional groups, thereby altering the aggregation of hen egg white lysozyme and inducing the formation of less toxic spherical aggregates [205]. While this system uses functionalized CDs and the model protein is not one associated with AD or PD, it shows that modifying the surface chemistry can significantly impact the interaction between CDs and amyloid proteins. CDs derived from naturally occurring bioactive molecules may also promote the disassembly of mature amyloid fibrils. For instance, flavonoid-derived carbonized polymer dots have been shown to inhibit the aggregation of both hen and human lysozyme amyloid proteins in a dose-dependent manner, while also promoting the degradation of mature fibrils into smaller aggregates [206]. Red-emissive quercetin-derived CDs have been reported to rapidly disassemble mature Aβ fibrils whilst also exhibiting ROS-scavenging and Aβ plaque-imaging capabilities. In an AD nematode model, these CDs reduced toxicity and prolonged lifespan, suggesting the potential integration of anti-aggregation, antioxidant, and imaging functions [204]. Additionally, plant-derived fluorescent CDs can modulate copper-induced disturbances in lipid metabolism and protein aggregation through copper-ion chelation. This suggests that they may alleviate associated neurotoxicity by interfering with the ‘metal-ion imbalance–lipid metabolic disturbance–protein aggregation’ cascade [207]. Overall, CDs derived from naturally occurring bioactive molecules may regulate multiple stages of pathological protein aggregation by disassembling mature fibrils, scavenging ROS, modulating metal-ion homeostasis, and reducing aggregate-associated cytotoxicity. However, the available findings are largely based on in vitro aggregation systems, nematode models, or short-term toxicity assessments. There is a lack of causal evidence linking these effects to brain exposure, reduced pathological protein burden, and long-term behavioral improvement. 4.2.3. Regulation of Proteostasis and Reduction in Abnormal Protein Burden The effects of CDs derived from natural products and their functionalized derivatives on pathological protein processes extend beyond inhibiting aggregation and promoting aggregate disassembly. They may also involve regulating proteostasis, restoring autophagic clearance, and reducing abnormal protein burden [127,165,195]. For instance, curcumin-derived CQDs can enhance the water solubility and bioavailability of curcumin while simultaneously reducing Aβ aggregation and tau hyperphosphorylation. This suggests a potential role in regulating AD-related proteostasis [34]. Additionally, glucose- or trehalose-functionalized carbon-based polymeric nanoparticles loaded with plasmid DNA (pDNA) can facilitate nuclear gene delivery under ROS-responsive conditions. These nanoparticles can silence the expression of SNCA, the gene encoding α-syn, restore autophagic flux, reduce α-syn aggregation, alleviate neuroinflammation, and improve motor function in PD mice [165]. This study suggests that CD-related delivery systems may facilitate the clearance of abnormal proteins by regulating genes and restoring autophagy. While some CDs have not been directly demonstrated to promote abnormal protein degradation, they may contribute to proteostasis regulation by inhibiting aggregation or reducing aggregate burden. For instance, phenylalanine-derived CDs can bind to and stabilize α-syn monomers and dimers via hydrophobic interactions, thereby preventing further aggregation. They have also been reported to enhance motor function in PD mice after crossing the BBB [208]. Congo red-derived CDs can simultaneously inhibit tau and Aβ aggregation, providing experimental evidence for dual-target intervention against abnormal protein pathology in AD [209]. Graphene quantum dots (GQDs) derived from coffee waste and CDs derived from camel milk have also been reported to inhibit α-syn fibrillization, suggesting that they may reduce the burden of PD-related abnormal protein aggregates [210,211]. However, not all of these systems are strictly classified as natural product-derived CDs, and aggregation inhibition remains the primary endpoint in most studies. Therefore, these findings cannot be equated directly with enhanced protein degradation or clearance. In summary, CDs derived from natural products or naturally occurring bioactive molecules, together with related functionalized CDs, may modulate NDD-associated protein pathology by inhibiting Aβ, tau, and α-syn aggregation, promoting the disassembly of mature fibrils, regulating metal-ion homeostasis, restoring autophagic flux, or reducing abnormal protein burden. These findings provide important directions for the design of multitarget neuroprotective and theranostic systems. However, current studies have several notable limitations. Firstly, some materials are not strictly derived from natural products, resulting in heterogeneous levels of evidence. Secondly, many studies remain limited to in vitro protein aggregation assays or short-term experimental models. Thirdly, it is unclear whether reductions in abnormal protein burden result from genuine brain exposure, direct interaction with pathological proteins, enhanced autophagy, or other indirect effects. To further clarify the actual contribution and translational value of natural product-derived CDs in regulating abnormal protein homeostasis, future studies should integrate quantitative assessments of brain distribution, measurements of pathological protein burden, validation of the autophagy–lysosomal pathway, long-term behavioral evaluation, and safety analysis. 4.3. Targeted Delivery and Controlled-Release Systems In addition to their potential intrinsic neuroprotective effects, natural product-derived CDs (nanoparticles) and their functionalized, drug-loaded systems can serve as nanodelivery platforms, enhancing the transport of drugs or nucleic acids across the BBB and improving their delivery and release within the brain [212]. In NDDs, targeted delivery systems are designed to overcome the limited entry of therapeutic molecules into the CNS and insufficient exposure to the brain, whereas controlled-release systems can extend the duration of therapeutic action and reduce exposure to non-target tissues, thereby mitigating potential adverse effects to some extent [166]. Accordingly, the value of natural product-derived CDs in NDD delivery research lies not only in their potential to cross the BBB but also in their modifiable surface, drug-loading capacity, and suitability for responsive-release designs. It should be emphasized that BBB penetration, brain distribution, accumulation at pathological sites, and therapeutic benefit are distinct levels of evidence and should not be considered equivalent. 4.3.1. Targeted Drug Delivery Systems The advantages of CDs for brain delivery are primarily associated with their small size, good water solubility, and surface-modification capability. Their small size may facilitate tissue penetration and cellular uptake, while surface functional groups, such as hydroxyl, carboxyl, and amino groups, can improve colloidal stability and provide reactive sites for the modification of targeting ligands, drug conjugation, and the modulation of biological interfaces [24,213]. Unlike tumor delivery, where the enhanced permeability and retention effect is often emphasized, brain delivery in NDDs relies more heavily on interactions between materials and the BBB’s endothelial cells, transporters, and receptor systems [212,214,215]. Therefore, the brain-delivery capacity of CDs should be evaluated in terms of in vitro BBB permeability, fluorescence signals in brain tissue, brain parenchymal exposure, and accumulation in pathological regions. Previous studies have provided some experimental evidence for the delivery of natural product-derived CDs across the BBB. For instance, fucoidan-derived CDs may retain functional groups, such as sulfate groups, and exhibit a negative surface charge. These CDs have been reported to cross the BBB and improve motor function in PD mice [216]. Radix Isatidis-derived CDs have also been shown to penetrate the BBB and produce therapeutic effects in an AD mouse model [33]. Additionally, glucose-derived fluorescent CDs have been reported to cross the BBB without additional targeting ligands, potentially via glucose transporter-mediated uptake [159]. Conversely, tryptophan-derived CDs can enter the CNS via the LAT1 transport system [161]. These findings suggest that surface groups derived from natural precursors or generated through carbonization-induced structural reconstruction may facilitate interactions between CDs and BBB transport systems. To enable active targeting, the surfaces of CDs can be functionalized with hydroxyl, carboxyl, and amino groups to which glucose, amino acids, lactoferrin, or peptide ligands can be introduced, thereby enhancing their interactions with transport systems such as GLUT1, LAT1, and the lactoferrin receptor [159,215,217]. For instance, glucose- or trehalose-functionalized carbon-based polymers loaded with pDNA were able to cross the BBB via a GLUT1-related pathway to achieve nuclear gene delivery in a PD model [165]. Phenylalanine-functionalized CDs have also been reported to cross the BBB, inhibiting α-syn aggregation and alleviating neuroinflammation in the brains of PD mice [208]. Additionally, microglial membrane-biomimetic nanovesicles constructed from caffeic acid-conjugated CQDs could bypass BBB-related limitations via intranasal administration, thereby enhancing delivery to inflamed brain regions and offering a new design strategy for targeting AD-associated lesions [218]. Overall, natural product-derived CDs and their functionalized systems may improve brain-delivery potential through the possible effects of their small size on endothelial uptake and tissue diffusion, recognition mediated by surface groups derived from natural precursors, active targeting modifications, and biomimetic nose-to-brain delivery. These strategies provide a material basis for brain-targeted drug delivery in NDDs. However, most current studies still lack systematic analyses of quantitative brain distribution, brain-to-blood ratios, localization within pathological regions, and associations with long-term pharmacodynamic outcomes. Therefore, while the available findings support the potential of these systems for brain delivery, they should not be interpreted as evidence of stable accumulation at pathological sites or clinical therapeutic superiority. 4.3.2. Stimuli-Responsive Controlled-Release Systems In addition to targeted delivery, CDs derived from natural products and their functionalized systems can be used to load and control the release of drugs, nucleic acids, or naturally occurring bioactive molecules. Their small size, good water solubility, biocompatibility, and numerous surface functional groups (including hydroxyl, carboxyl, and amino groups) provide a basis for drug adsorption, covalent conjugation, nucleic acid complexation, and the construction of responsive structures [36,219,220]. In the treatment of NDDs, brain-targeted, controlled-release systems are relevant in two main ways. Firstly, strategies involving drug protection, carrier targeting, and trans-BBB transport may improve drug delivery and exposure within the brain. Secondly, elevated ROS levels, altered activities of inflammation-associated enzymes, or acidic endosomal and lysosomal environments in pathological regions can be exploited to achieve sustained or stimuli-responsive release. This prolongs local therapeutic activity and reduces off-target exposure [221,222,223,224]. Previous studies have reported CD-based systems for sustained drug release. For instance, chitosan-derived CDs can be combined with chitosan (CS) to form a CS/CD nanocarrier loaded with dopamine, creating a dopamine@CS/CD system. This system enables sustained dopamine release and uses the fluorescence properties of CDs to track delivery, indicating its potential for drug delivery in the nervous system [220]. Additionally, a curcumin delivery system based on an Fe3O4@carbon dot nanocomposite can enhance curcumin stability and cellular uptake, setting a precedent for the nanodelivery of naturally occurring bioactive compounds [225]. However, these studies mainly demonstrate the drug-loading and release capacities of CD-based systems. Whether such systems can effectively target the brain and demonstrate long-term therapeutic efficacy in NDD models remains to be validated. For stimuli-responsive release, microenvironmental changes such as elevated ROS levels, inflammatory conditions, and acidic intracellular compartments may serve as potential triggers. A supramolecular carbon dot system constructed from lemon-derived CDs and a curcumin analog was able to load methotrexate and exhibit sustained release under acidic pH conditions. This provides a methodological reference for designing pH-responsive, controlled-release systems [226]. However, this study was not conducted in an NDD model and therefore cannot be extrapolated as therapeutic evidence for NDDs. In contrast, glucose/trehalose-functionalised carbon-based polymeric nanoparticles loaded with pDNA exhibited ROS-responsive behavior under oxidative stress, combining this property with GLUT1-mediated transport across the BBB to achieve nuclear gene delivery and functional cargo release in a PD model [165]. This study more directly demonstrates the potential of CD-related systems for responsive delivery in NDDs. In summary, natural product-derived CDs and their functionalized systems may enable the stable loading and controlled release of drugs or nucleic acids through surface functional group-mediated interactions, covalent conjugation, non-covalent assembly, and pH- or ROS-responsive structural designs. These strategies could improve brain delivery efficiency and the release profiles of therapeutic agents, offering new approaches to long-term treatment and multimodal intervention in NDDs. However, the current evidence still has several notable limitations. First, some controlled-release systems have only been evaluated in non-NDD models and therefore mainly provide methodological guidance. Second, systematic evaluations of release kinetics in the brain, local drug concentrations at pathological sites, and long-term safety remain lacking. Third, the respective contributions of the carrier itself, the loaded therapeutic agent, and surface modifications have not been adequately distinguished. Future studies should integrate quantitative pharmacokinetic analysis, regional brain distribution, release kinetics, long-term toxicological assessment, and correlations with therapeutic efficacy in disease models to further clarify the therapeutic value of controlled-release systems based on natural product-derived CDs for NDDs. 4.4. Theranostics and Visualization-Based Monitoring In NDD research, theranostic materials can combine pathological recognition and delivery tracking with therapeutic intervention. This provides researchers with tools to investigate disease mechanisms and evaluate treatment efficacy. Some CD systems have exhibited low cytotoxicity in short-term in vitro studies, and have tunable fluorescence and surfaces that can be easily functionalized. These properties enable their use for imaging pathological structures, tracking carrier distribution, and delivering therapeutic agents. They can also be combined with anti-oxidative, anti-inflammatory, or anti-aggregation functions to create integrated systems that encompass ‘imaging and tracking–targeted delivery–therapeutic intervention’ [212,227,228]. However, it should be noted that fluorescence signals primarily reflect material distribution, binding to pathological structures or local response processes, and cannot be directly equated with effective brain exposure, therapeutic concentrations at pathological sites, or long-term therapeutic benefit [113,229]. Therefore, when evaluating the theranostic value of CDs, distinct levels of evidence, including imaging capability, brain delivery, pathological lesion recognition, and therapeutic efficacy, should be clearly differentiated. 4.4.1. Fluorescence-Based Pathological Imaging The imaging value of CDs derived from natural products primarily arises from their stable and tunable fluorescence properties. Their photoluminescence behavior is generally associated with the carbon core structure, surface states, molecular states, and fluorophore groups formed during precursor conversion. Their emission wavelengths can be tuned by adjusting the precursor composition, synthesis conditions, and surface functional groups. Red- and near-infrared (NIR)-emissive CDs can reduce interference from tissue autofluorescence and light scattering to some extent. This improves tissue penetration depth and imaging contrast, making them more suitable for deep-tissue or in vivo imaging [230,231,232,233,234,235,236]. For instance, protonated o-phenylenediamine-derived CDs demonstrate high-color-purity red emission and NIR two-photon emission, offering a valuable model for grasping fluorescence-regulation mechanisms in CDs [230]. However, this system is not derived from natural products and is therefore more appropriately regarded as a reference for mechanistic studies of CD fluorescence. Among natural product-derived CDs, glutathione-derived NIR CDs exhibited low cytotoxicity in short-term in vitro assessments and possessed two-photon imaging capability, suggesting their potential for deep-tissue imaging [231]. Additionally, red- or NIR-emissive CDs derived from Epigynum auritum, spinach, lemon juice, mulberry leaves, and yew leaves have demonstrated favorable photostability, biocompatibility, and in vitro and in vivo imaging performance [232,233,234,235,236]. These studies suggest that red- and NIR-emissive, natural product-derived CDs could be used as fluorescent probes to track brain distribution, identify pathological structures, and design theranostic systems. However, most studies remain limited to general bioimaging or tissue imaging applications, and substantial evidence is lacking regarding the specific recognition of NDD-associated pathological lesions. 4.4.2. Lesion Recognition and Dynamic Monitoring CDs derived from natural products and their functionalized derivatives can interact with abnormal proteins or pathological microenvironments via their surface aromatic structures and functional groups, such as hydroxyl, carboxyl, and amino groups. This induces changes in fluorescence signals that can be used to recognize pathological structures and monitor their dynamics [209,237]. For instance, EGCG-derived carbonized polymer dots exhibit increased fluorescence upon binding to Aβ fibrils, thereby enabling Aβ fibril recognition [182]. Similarly, quercetin-derived red-emissive CDs exhibit a ‘turn-on’ red fluorescence upon contact with Aβ plaques, facilitating visualization of Aβ aggregates [204]. Similarly, resveratrol-derived CDs exhibit fluorescence responses to amyloid fibrils and can be used to monitor amyloid aggregation [238]. These findings suggest that CDs derived from naturally occurring bioactive molecules may enable recognition of pathological structures via binding-induced fluorescence changes to abnormal proteins. Some natural product-derived CDs and their functionalized systems can also be used to track and visualize lesion-associated regions in vivo. Microglial membrane-biomimetic nanovesicles constructed from caffeic acid-conjugated CQDs can bypass BBB-related limitations upon intranasal administration and accumulate in inflamed brain regions. Their fluorescence signals enable the visualization of the in vivo distribution and delivery processes [218]. Fucoidan- and Radix Isatidis-derived CDs have both been reported to cross the BBB, exhibiting brain-delivery- and neuroprotection-related effects in PD and AD models [33,216]. In PD-related studies, phenylalanine-derived CDs have been reported to bind to α-syn monomers and oligomers and to penetrate the BBB, providing a basis for tracking and intervening in α-syn-associated pathological processes [208]. Overall, natural product-derived CDs and their functionalized systems may contribute to monitoring NDD-associated lesions through fluorescence responses, abnormal protein recognition, and brain-delivery tracking. However, most current studies have focused on Aβ fibrils, Aβ plaques, or model proteins. Furthermore, fluorescence signals may be influenced by material distribution, residual intravascular signal, tissue autofluorescence, and probe concentration. Therefore, fluorescence enhancement or signals detected in the brain alone cannot demonstrate localization within the brain parenchyma or the occurrence of a therapeutic effect. 4.4.3. Integration of Theranostic Functions and Monitoring of Therapeutic Processes In addition to lesion recognition, natural product-derived CDs and their functionalized systems can combine fluorescence tracking, brain-targeted delivery, and therapeutic intervention on a single platform. Their tunable fluorescence properties facilitate visualization of carrier distribution and delivery processes, while their surface modifiability and drug-loading capacity provide a structural basis for anti-aggregation, antioxidant, and anti-inflammatory functions, as well as drug delivery [24,37,38,239]. For instance, EGCG-derived carbonized polymer dots demonstrate enhanced fluorescence when they bind to Aβ fibrils, simultaneously inhibit Aβ fibrillization, promote the disassembly of mature fibrils, and scavenge ROS. This shows that they have a certain degree of integrated ‘recognition–intervention’ capability [182]. Quercetin-derived red-emissive CDs enable red-fluorescence imaging of Aβ plaques and exhibit anti-Aβ aggregation and antioxidant activities. This suggests their potential to recognize Aβ pathology and monitor therapeutic processes [204]. For delivery and tracking purposes, nanovesicles biomimetic of the microglial membrane, constructed from caffeic acid-conjugated CQDs, can be delivered to inflamed brain regions following intranasal administration. Their fluorescence signals enable visualization of in vivo distribution, brain-delivery processes, and accumulation within inflammatory regions [218]. Curcumin-loaded Fe3O4@CQD nanocomposites combine CQD-based fluorescence tracking with magnetically targeted delivery and improve curcumin cellular uptake, inhibiting Aβ aggregation and ROS-induced neurotoxicity [225]. Additionally, fucoidan- and phenylalanine-derived CDs have exhibited BBB-penetrating ability and improved motor function in PD models, providing experimental evidence for tracking brain delivery and evaluating efficacy [208,216]. In summary, CDs derived from natural products and their functionalized systems may contribute to the development of theranostic platforms for NDDs through fluorescence tracking, brain delivery, recognition of abnormal proteins, and anti-aggregation, antioxidant, and anti-inflammatory effects. These properties provide a material basis for lesion localization, visualization of carrier distribution, and monitoring of therapeutic processes. However, the available evidence should be interpreted cautiously. Fluorescence signals primarily reflect material distribution, binding to pathological structures, or local response processes, rather than directly demonstrating effective drug concentrations in the brain, selective accumulation within pathological lesions, or long-term disease-modifying effects. Table 3 summarizes the relevant preclinical studies by preparation methods, disease models, evidence of BBB penetration or brain delivery, major outcomes, and study limitations. Taken together, the studies listed in Table 3 suggest that the brain exposure, in vivo stability, metabolism and clearance, tissue accumulation, and long-term safety of natural product-derived CDs have not yet been systematically evaluated. Given the chronic progression of NDDs and the need for long-term treatment, future studies should extend beyond functional validation and strengthen systematic assessments of pharmacokinetics, biodistribution, safety, quality control, and translational feasibility. The following section discusses the major challenges facing natural product-derived CDs with respect to absorption, distribution, metabolism, and excretion (ADME), long-term safety, material consistency, and clinical translation within these evidentiary boundaries. 5. Translational Challenges and Preclinical Evaluation of Natural Product-Derived Carbon Dots The translational potential of CDs derived from natural products cannot be assessed solely on the basis of short-term functions such as antioxidant and anti-inflammatory effects, inhibition of protein aggregation, or imaging and tracking capabilities. Instead, they should be evaluated comprehensively in conjunction with their in vivo disposition, exposure to the brain parenchyma, long-term safety, material consistency, and validation in human-relevant models [92,98,127]. As current studies largely rely on simplified BBB models and short-term animal experiments, in vitro permeability, signals detected in brain tissue, and functional improvements alone are insufficient to demonstrate effective exposure within the brain parenchyma [112,113,240,241]. 5.1. ADME Characteristics and Evaluation of Brain Exposure The circulation kinetics, biodistribution, and clearance of CDs can be markedly influenced by different routes of administration. In addition, the protein corona formed upon entry into the bloodstream can alter their surface properties, circulation time, immune recognition, and in vivo fate [119,242]. While oral administration is convenient for long-term treatment, its efficiency for brain delivery is limited by gastrointestinal stability, intestinal absorption, first-pass metabolism, and systemic bioavailability [243]. Following intranasal administration, CDs can access the CNS via the olfactory, trigeminal, and perivascular pathways. However, distribution to the brain after intranasal delivery should not be considered equivalent to conventional BBB penetration, and direct nose-to-brain transport should be distinguished from brain distribution following systemic absorption [244,245,246]. Assessment of brain exposure should not rely solely on fluorescence signals from the whole brain or brain sections. These signals may be affected by residual intravascular material, inadequate perfusion, binding to or uptake by brain microvascular endothelial cells, leakage, or degradation of fluorescent labels. Therefore, they cannot independently demonstrate entry into the brain parenchyma or attainment of an effective concentration at pathological sites [113,229,247]. Brain tissue-associated total signal, brain parenchymal exposure, and exposure at the site of action must therefore be distinguished and evaluated comprehensively using time-course measurements of brain-to-blood ratios, adequate perfusion or vascular correction, capillary depletion, regional and cellular localization, and quantitative tissue analysis [112,113,248]. For CDs containing stable characteristic elements or radiolabels, in vivo tracking and quantification may be feasible; however, the label’s in vivo behavior must remain consistent with that of the CD itself. Methods such as liquid chromatography can be used to analyze free fluorophores, residual precursors, and low-molecular-weight by-products. However, in the absence of particle-specific separation and identification, the detected signals should not be directly interpreted as quantitative measurements of the CD itself [249,250,251]. The in vivo distribution, transformation, and clearance pathways of CDs remain to be fully clarified. Their clearance behavior cannot be inferred solely from dry-state particle size measured by transmission electron microscopy (TEM) but should also be evaluated in relation to hydrodynamic size, aggregation state, surface charge, and surface modification. Well-dispersed CDs with small hydrodynamic diameters may be partially cleared via renal excretion, whereas cross-linked or aggregated multi-CD nanoparticles may be predominantly cleared via the hepatobiliary system [242,252,253,254]. In addition to tissue distribution and excretion, it is essential to determine whether fluorescent labels can stably represent the CD itself. Dye leakage, fluorescence quenching or dequenching, signal saturation, and labeling-induced changes in cellular uptake may all confound the interpretation of in vivo distribution [92,113,229]. For drug- or nucleic acid-loaded systems, the in vivo behavior of the carrier, label, and therapeutic payload should be evaluated separately to avoid inferring drug release or effective brain exposure solely from carrier fluorescence. Future studies should incorporate time-dependent measurements in blood, urine, feces, and major organs to systematically characterize the in vivo fate of CDs and their associated components. 5.2. Long-Term Safety, Immunocompatibility, and Neurotoxicology The treatment of NDDs generally requires long-term or repeated administration; therefore, favorable short-term tolerability or low short-term cytotoxicity does not guarantee long-term safety. For nanomaterials with the potential to be delivered to the brain, safety assessments should address not only toxicity in peripheral organs but also the risks of oxidative stress, deoxyribonucleic acid (DNA) damage, neuroinflammation, and intracerebral accumulation following BBB translocation [98,255,256]. Previous studies have shown that some CDs exhibit low toxicity in vitro yet cause significant in vivo toxicity following repeated administration [175]. CD-based nanoparticles characterized by prolonged retention or hepatobiliary clearance also require further evaluation of their long-term nanotoxicological risks [254]. It should be emphasized that the historical medicinal use or natural origin of a precursor cannot be regarded as evidence that the resulting CDs are non-toxic or immunocompatible. Carbonization can substantially alter the original constituents, generating new carbon cores, surface structures, and low-molecular-weight by-products. Therefore, safety studies should evaluate the final CD product after adequate purification and characterization, rather than relying on the medicinal history or natural origin of the precursor as a substitute for systematic toxicological evidence [28,230,257,258]. Future studies should systematically assess hematological parameters, coagulation function, hemolysis, complement activation, inflammatory mediators, oxidative stress, liver and kidney function, and histopathological alterations in major organs under conditions of repeated administration and long-term exposure [259,260,261]. For brain-delivery systems in particular, attention should be paid to nanocarrier-induced neuroinflammation, alterations in BBB homeostasis, intracerebral accumulation, and chronic neurotoxicity. Relevant endpoints should include neuronal injury, synaptic structure and function, glial activation, and cerebrovascular integrity [98]. Only by integrating the evaluation of peripheral toxicity, CNS safety, and long-term immunocompatibility can the suitability of CDs derived from natural products for treating chronic NDDs be assessed more accurately. 5.3. Material Reproducibility, Validation Using Human-Relevant Models, and Regulatory Translation Material reproducibility is a critical prerequisite for the translational application of CDs derived from natural products. Natural product precursors are compositionally complex, and the source of the precursors, the conditions under which they are prepared, and the purification procedures can all affect the composition, surface properties, fluorescence behavior, and biological activity of the resulting CDs [28,258,262,263]. Inadequate purification can also result in the presence of low-molecular-weight components, nanoscale aggregates, or fluorescent impurities, leading to an unclear material identity and incorrect attribution of biological effects [257,258]. Candidate systems with translational potential should therefore establish a systematic quality-control framework that standardizes precursor sources, preparation parameters, purification procedures, and batch-to-batch consistency. Key quality attributes, including particle size and size distribution, surface charge and chemical composition, optical properties, purity and stability in biological media, should be adequately characterized [264,265,266]. Biological experiments should also include controls consisting of the precursor, low-molecular-weight fractions, and adequately purified CDs in order to distinguish the relative contributions of the CD material itself, residual precursors, and by-products. Differences in gene expression, transporter profiles, and receptor-mediated transcytosis systems between the human and rodent BBB may limit the clinical predictive value of animal brain-delivery findings [267,268,269]. In addition, BBB/NVU dysfunction associated with aging and NDDs may alter the brain distribution, clearance, and safety of these materials [270,271]. Therefore, future studies should incorporate platforms such as human-induced pluripotent stem cell (iPSC)-derived BBB/NVU models, microfluidic barrier-on-chip systems, and vascularized brain organoids to evaluate the brain-delivery capacity and central nervous system safety of natural product-derived CDs [272,273,274]. Based on the available preclinical evidence, CDs derived from Radix Isatidis, fucoidan, phenylalanine, and other naturally occurring bioactive molecules have been evaluated in disease models, demonstrating therapeutic efficacy and BBB-associated delivery. These CDs therefore currently have comparatively greater preclinical support. However, existing studies generally lack quantification of brain parenchymal exposure specific to individual particles, systematic pharmacokinetic characterization, repeated-dose safety assessment, evaluation of batch-to-batch consistency, and validation in human-relevant BBB models [33,182,208,216]. Consequently, no natural product-derived CD system can yet be identified as having a clear translational advantage, and the field remains largely at the proof-of-concept stage. From a regulatory perspective, the development category of a candidate product should be defined according to its composition and intended use. This includes whether it is being developed as an active nanomaterial, a drug delivery system, an imaging probe, or a theranostic combination product. Preclinical development should also clarify the identity of the materials and the basis of the active components, establish key quality attributes and exposure–response relationships, and generate evidence to support dose selection, Good Laboratory Practice (GLP)-compliant toxicology, Good Manufacturing Practice (GMP)-compliant scale-up production, and product batch consistency [275,276,277,278,279]. Overall, the translational evaluation of natural product-derived CDs requires a continuous chain of evidence extending from material quality and BBB-associated delivery to brain parenchymal exposure, pathological localization, attribution of pharmacological effects, long-term safety, validation in human-relevant models, and standardized manufacturing. As illustrated in Figure 4, each stage should be validated sequentially. A natural precursor, in vitro BBB permeability, a signal detected in brain tissue, or therapeutic efficacy in an animal model should not be considered equivalent to a compositionally defined material, effective brain parenchymal exposure, lesion-specific targeting, or clinical translational value, respectively. 6. Conclusions Overall, CDs derived from natural products represent a class of candidate materials with exploratory value for brain delivery, neuroprotection, and theranostic research in NDDs. Preclinical studies indicate that their small size, tunable surface chemistry, and functional modifications could facilitate interaction with the BBB. Some systems have also been shown to exhibit antioxidant, anti-inflammatory, and abnormal protein aggregation-modulating activities, as well as drug delivery and fluorescence tracking in cellular and animal models. Therefore, natural product-derived CDs may support the development of multifunctional nanotheranostic strategies that integrate ‘delivery–intervention–monitoring’ for NDDs. However, it should be emphasized that in vitro BBB permeability, fluorescence signals in brain tissue, brain-associated distribution, brain parenchymal exposure, and therapeutic benefit are distinct levels of evidence and should not be considered interchangeable. The further translation of natural product-derived CDs is constrained by multiple factors. Firstly, their material composition, structural characteristics, and consistency across batches require further clarification. Because natural product precursors are compositionally complex, carbonization conditions, purification procedures, surface states, and residual small molecules may all influence the structural characteristics and biological effects of the resulting CDs. Reproducible manufacturing processes must therefore be established, and key quality attributes and impurity-control specifications must be defined. Secondly, there is insufficient evidence regarding brain delivery and ADME. Brain fluorescence signals or behavioral improvements alone cannot demonstrate effective exposure of the brain parenchyma. Future studies should therefore incorporate quantitative brain distribution measurements, brain-to-blood ratios, capillary depletion analysis, regional brain localization, pharmacokinetics, metabolic transformation, and hepatic and renal clearance to clarify the relationships among route of administration, brain exposure, and therapeutic efficacy. Thirdly, long-term efficacy and safety require systematic validation, paying particular attention to tissue accumulation, immune responses, neurotoxicity, hematocompatibility, clearance kinetics and chronic nanotoxicological risks following repeated administration. Future research should move beyond functional proof of concept toward a systematic evaluation oriented toward clinical translation. Structurally defined, adequately purified, and batch-consistent CD systems should be used to clarify the influence of particle size, surface charge, functional groups, heteroatom doping, and residual precursor components on BBB interactions and neuroprotective effects. In addition, more clinically relevant, human-derived BBB models, NVU models, organ-on-chip platforms, and multicellular co-culture systems should be introduced to assess delivery efficiency, barrier safety, and immunological effects in human-relevant barrier environments. Meanwhile, representative NDD models should be used to establish integrated evaluation frameworks encompassing brain exposure, pathological endpoints, behavioral outcomes, and long-term safety. Natural product-derived CDs can only progress beyond the proof-of-concept stage toward nanomedicine platforms with genuine translational potential for neurological applications when material composition is controllable, brain parenchymal exposure is clearly demonstrated, exposure–efficacy relationships are established, and long-term safety is acceptable.