Neurological disorders represent one of the most significant public health challenges globally, affecting hundreds of millions of individuals across all age groups [1, 2]. Conditions such as Alzheimer’s disease (AD), Parkinson’s disease (PD), amyotrophic lateral sclerosis (ALS), multiple sclerosis (MS), epilepsy, and stroke are characterized by progressive functional deterioration, high morbidity, and often a lack of curative treatment [3, 4]. Despite ongoing advances in molecular neuroscience and neuroimmunology, a critical gap remains in our understanding of the dynamic intercellular communication systems that underlie both the maintenance of central nervous system (CNS) homeostasis and the pathogenesis of neurodegeneration [5, 6]. The complexity of CNS signaling is amplified by diverse cellular constituents, including neurons, astrocytes, oligodendrocytes, microglia, and endothelial cells, which must be coordinated in a highly synchronized and spatially regulated environment. Impaired intercellular communication drives disease pathogenesis, highlighting the importance of understanding alternative signaling mechanisms [7, 8].

In recent years, there has been a paradigm shift in our understanding of how cells communicate in the CNS, with increasing recognition of extracellular vesicles (EVs) as pivotal mediators of this process [9, 10]. EVs are nano-to microscale membrane-bound structures released by nearly all cell types. EV-mediated communication, previously underappreciated compared to synaptic transmission and soluble factor signaling, now attracts significant attention due to its capacity for complex molecular messaging, including proteins, lipids, and nucleic acids, across long distances and even across biological barriers such as the blood-brain barrier (BBB) [11]. In the neural context, EVs serve as shuttles for regulatory RNA molecules, neurotransmitter-modulating proteins, and inflammatory signals, making them indispensable for both normal physiology and the propagation of disease signals [12].

EVs are classified into exosomes and microvesicles based on their size and biogenesis. Exosomes (30–150 nm) originate from the endosomal system and are secreted via the fusion of multivesicular bodies with the plasma membrane, whereas microvesicles (100–1,000 nm) bud directly from the cell membrane [13, 14]. These vesicle subtypes transport distinct, cell-specific cargo that reflects their cellular origin and physiological state. They regulate critical processes, including synaptic plasticity, myelination, immune surveillance, and tissue repair [15, 16].

EVs represent actively regulated cellular products rather than passive byproducts. Cells employ sophisticated machinery to control EV formation and cargo selection. Specific intracellular mechanisms govern EV molecular composition by selectively packaging signaling molecules into vesicles [17, 18]. Among these regulatory mechanisms, post-translational modifications (PTMs) have emerged as critical modulators of EV cargo selection. In particular, ubiquitin-like protein 3 (UBL3), a membrane-anchored PTM enzyme, has recently garnered attention for its role in promoting the incorporation of proteins into small EVs (sEVs). UBL3 is preferentially modified with geranylgeranyl moieties, facilitating membrane anchoring and cargo affinity [19]. While its functions are still being elucidated, early evidence suggests that UBL3 is a key player in shaping the immunological and signaling profiles of EVs, with potential implications in neuroinflammation and neurodegenerative disease pathways [20, 21].

The relevance of EVs to neurological disorders has become increasingly evident. Pathogenic proteins, such as amyloid-β (Aβ), tau, α-synuclein, and TAR DNA-binding protein 43 (TDP-43), have been detected in EVs, implicating them in the intercellular spread of protein aggregates in AD, PD, and ALS [22–24]. EVs also play a role in modulating the immune environment in MS and compromising the integrity of the BBB in stroke [25, 26]. These findings establish mechanistic connections between EV biology and neurological disease while highlighting their potential as diagnostic biomarkers and therapeutic tools. Their presence in biofluids, such as cerebrospinal fluid (CSF), blood, and saliva, makes them accessible for noninvasive monitoring. In contrast, their ability to cross the BBB opens exciting avenues for targeted drug delivery to the CNS [27, 28].

Despite these advances, significant challenges remain in EV research, including the heterogeneity of vesicle populations, the need for standardization of isolation protocols, and the requirement for cell-type-specific in vivo models [29, 30]. Moreover, while UBL3 and other PTMs offer intriguing molecular handles for manipulating EV cargo, their physiological relevance and druggability in neural contexts remain to be explored. Understanding the functional roles of EV-associated PTMs, particularly in the context of CNS-specific signaling and disease modulation, is crucial for translating basic EV research into clinical applications [31–33].

This review aims to provide a comprehensive synthesis of the emerging roles of EVs in neurological disorders, with special emphasis on the molecular mechanisms underlying EV biogenesis, cargo sorting, and function. Particular focus is given to the regulatory role of UBL3 and its potential as a modulator of EV composition in neurodegenerative and neuroinflammatory diseases. By integrating insights from EV biology, neuroscience, and translational medicine, this study outlines the current landscape and future directions for leveraging EVs as diagnostic and therapeutic tools in neurology.

EVs in the CNS are released by diverse neural cell types, including neurons, astrocytes, microglia, and oligodendrocytes, each contributing unique molecular cargoes that reflect their physiological or pathological states [34]. Neurons produce EVs from both the soma and dendrites, which regulate synaptic plasticity and neuronal support. For example, neuronal EVs contain microRNAs (miRNAs) (such as miR-98) that regulate microglial phagocytosis after ischemic injuries [35, 36]. Oligodendrocyte-derived EVs supply axonal support under stress by delivering enzymes such as catalase and superoxide dismutase (SOD) [37, 38]. Astrocyte-derived EVs carry apolipoprotein D and neurotransmitter transporters that influence neuronal survival [39], and microglia-derived EVs contain proteins such as glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and lipid mediators such as endocannabinoids that modulate synaptic activity and inflammatory signaling [40, 41].

Cell-specific regulatory mechanisms control the secretion of EVs from these cells. In microglia, ATP stimulation via P2X7 receptors triggers ectosome shedding, whereas serotonin and Wnt-3a signaling enhance exosome release through alterations in calcium flux and endosomal pathways [42]. Neurons rely on endosomal sorting complex required for transport (ESCRT)-dependent and ceramide-mediated pathways for EV biogenesis in response to synaptic activity or stress [43]. Oligodendrocytic EV release is enhanced by neuronal-glial interactions, suggesting bidirectional regulation [44]. Neuroanatomical studies have revealed cell-type-specific expression patterns of UBL3 and related ubiquitin-like proteins in brain regions affected by neurodegenerative diseases [45, 46].

EV populations are enriched in lipids, RNAs, and proteins that mediate intercellular communication. Lipidomic studies have shown that EVs from oligodendrocytes and microglia are enriched in ceramides, sphingomyelins, cholesterol, and glycolipids, which are molecules critical for membrane dynamics and signaling [47]. RNA cargoes include miRNAs that regulate inflammation (e.g., miR‑98 and miR‑146a), lncRNAs, and mRNAs that reflect the transcriptional state of the parent cell and can alter gene expression in recipient cells [48–50]. Proteomic analyses have revealed the presence of ESCRT proteins, tetraspanins (transmembrane proteins including CD9, CD63, and CD81 that organize membrane microdomains), heat shock proteins, metabolic enzymes, and cell-specific markers, including apolipoprotein D, GAPDH, and major histocompatibility complex (MHC) molecules [51, 52].

Several molecular mechanisms govern the specificity of EV cargo sorting. ESCRT-dependent ubiquitination pathways enable the selective packaging of proteins into exosomes by recognizing ubiquitinated membrane proteins through the recruitment of ESCRT-0 and ALIX [53, 54]. Additionally, ESCRT-independent mechanisms involving lipid rafts, ceramide, and tetraspanins play a crucial role in EV formation in neural cells [55].

PTMs of cargo proteins play a pivotal role in determining their incorporation into EVs. Beyond ubiquitination, SUMOylation and lipidation influence cargo sorting; for instance, SUMOylated heterogeneous nuclear ribonucleoprotein A2/B1 (hnRNPA2B1) binds specific miRNAs for exosomal inclusion, while palmitoylation governs exosomal protein-membrane interactions [56–58]. Most recently, UBL3 has been identified as a key regulator of EV protein sorting. UBL3 modifies proteins via S-palmitoylation-like PTMs to promote their targeting into sEVs. UBL3 knock-out cells exhibit a ~60% reduction in EV protein content, and UBL3-mediated sorting of signaling molecules, such as Ras, to EVs has been demonstrated [13, 20, 59]. While initially identified in non-neuronal systems, the role of UBL3 in neural EV biology is likely substantial but remains underexplored.

EV uptake by recipient cells occurs through multiple mechanisms, including membrane fusion, clathrin-mediated endocytosis, caveolin-dependent uptake, and macropinocytosis, with the specific pathway influenced by EV surface molecules and cell type [60, 61]. In neural cells, tetraspanins and heparan sulfate proteoglycans mediate receptor-dependent internalization, while microglia preferentially use phagocytosis compared to neurons and astrocytes [62–64].

Collectively, CNS-derived EVs exhibit highly specialized intercellular communication functions, reflecting precise cellular states shaped by regulated release mechanisms, enriched molecular cargo, and PTM-driven sorting machinery such as UBL3. Understanding these systems is critical for leveraging EVs in therapeutic and diagnostic applications in neurobiology. The diverse cellular origins of EVs in the CNS result in heterogeneous vesicle populations with distinct molecular signatures and functional properties. Table 1 provides a comprehensive overview of how neurons, astrocytes, microglia, and oligodendrocytes each contribute unique EV subtypes that collectively orchestrate complex intercellular communication networks essential for brain function.

UBL3 has emerged as a key regulator of sEV cargo sorting via PTM mechanisms. UBL3 is a membrane-anchored ubiquitinfold protein that binds to protein substrates via its C-terminal CAAX prenylation motif, facilitating targeted inclusion into sEVs [20, 33, 59]. Once attached to endosomal or plasma membranes, UBL3 promotes the packaging of select proteins into intraluminal vesicles by enhancing membrane association and potentially recruiting ESCRT or lipid-sorting machinery [20, 59]. The comparative analysis of PTM in EV cargo selection is summarized in Table 2.

The foundational understanding of UBL3 (initially termed membrane-anchored ubiquitin-fold, MUB) emerged from pioneering work by Vierstra’s group. They first characterized MUB as a novel ubiquitin-fold protein with a C-terminal CAAX prenylation motif essential for membrane localization [19]. Subsequent studies revealed that MUB localizes to both the plasma membrane and endomembrane systems, where it modifies target proteins through a mechanism distinct from classical ubiquitination [67]. Importantly, they demonstrated that MUB/UBL3 functions as a post-translational modifier controlling protein trafficking and signaling, establishing its role as a regulatory hub for cellular protein sorting [68]. This work laid the foundation for understanding UBL3’s role in EV biogenesis, as comprehensively reviewed by Vierstra [69].

While previous reviews have examined ubiquitin-like proteins in EV biology broadly, the present review specifically focuses on UBL3’s emerging role in neurological disorders, integrating recent discoveries about its interaction with neurodegenerative disease proteins and potential as a therapeutic target in CNS pathologies [70].

Proteomic analyses reveal that UBL3 interacts with more than 1,200 proteins, with approximately 30% classified as sEV cargo proteins. UBL3 knock-out cells exhibit roughly a 60% reduction in sEV protein content, underscoring its centrality in protein loading Figure 1 [20, 59]. Notably, UBL3-mediated sorting is essential for the incorporation of growth regulators, such as oncogenic Ras, which influences extracellular signal-regulated kinase (ERK) activation in recipient cells, illustrating how UBL3 can shift intercellular signaling profiles [33, 59].

In neural contexts, emerging evidence suggests that UBL3 is involved in the trafficking of proteins associated with neurodegenerative diseases. UBL3 interacts with α-synuclein in cell models, and this interaction increases under oxidative stress via microsomal glutathione S-transferase 3 [31]. UBL3 also colocalizes with mutant huntingtin (mHTT) fragments in neurons from Huntington’s disease patients, influencing their intracellular sorting [32]. Recent studies have further characterized UBL3’s role in neurodegenerative protein trafficking, demonstrating its involvement in cellular stress responses and protein quality control [71]. UBL3 expression patterns in neural tissues are dynamically regulated under pathological conditions, influencing both intracellular protein sorting and EV cargo composition [72]. While these studies focus on intracellular localization, they suggest UBL3 could direct pathogenic protein loading into sEVs, thus promoting pathological propagation across neural circuits.

Recent studies have expanded our understanding of UBL3’s roles in various pathological contexts. UBL3 has been identified as a critical mediator of inflammatory responses, with its expression significantly altered in inflammatory diseases [73]. In neural tissues, UBL3 shows distinct expression patterns during development and disease states, suggesting context-dependent functions [74]. Notably, UBL3 expression correlates with neuroinflammatory markers in neurodegenerative conditions [75]. In cancer pathology, UBL3 serves as a prognostic biomarker in certain malignancies [76, 77], while structural studies reveal UBL3’s unique membrane-targeting mechanisms that distinguish it from other ubiquitin-like modifiers [78]. These findings collectively highlight UBL3’s multifaceted roles beyond EV cargo sorting, encompassing inflammation, cancer progression, and neural development.

The brain-specific expression of UBL3 further supports a functional role in CNS EV biology. Proteomic analyses have identified additional UBL3-interacting proteins in neuronal models, expanding our understanding of its role in neurodegenerative disease pathways [79]. Transgenic mouse models overexpressing UBL3 in cortex and hippocampus identified 35 new UBL3 interactors, including RNA-binding proteins implicated in neurodegeneration [e.g., fused in sarcoma (FUS), hypoxanthine phosphoribosyltransferase 1 (HPRT1)] [75]. This raises the possibility that UBL3 participates in sEV-mediated RNA transfer, potentially influencing neural protein expression and homeostasis; however, in vivo functional data remain limited [80].

Beyond UBL3, other PTMs regulate EV cargo. SUMOylation of hnRNPA2B1 directs specific miRNAs into exosomes [58]. SUMO2 modifies α-synuclein and amyloid precursor protein (APP) to promote their sorting into EVs via interactions with ESCRT components [65]. Additionally, palmitoylation enhances the membrane targeting of cargo proteins, while ubiquitination by membrane-associated RING-CH (MARCH) E3 ligases controls the endosomal sorting of immune molecules, such as MHC-II, into EVs [81]. Glycosylation of transmembrane proteins can determine EV targeting and uptake, as shown for EWI2 [66]. These combined mechanisms suggest an intricate PTM network orchestrating EV content.

In the CNS, such PTM-regulated sorting can influence neuroinflammation, synaptic function, and neurodegeneration. UBL3-mediated inclusion of α-synuclein or mHTT into sEVs may exacerbate disease progression or modulate microglial activation. SUMOylated hnRNPA2B1 sorting of inflammatory miRNAs might enhance cytokine responses in astrocytes or microglia. Glycosylation of neuronal receptors in EVs could impact BBB integrity or neuroimmune crosstalk [56, 59, 82].

In vitro and in vivo functional studies demonstrate that EVs carrying pathogenic proteins induce aggregation and toxicity in recipient cells. In contrast, EVs loaded with regulatory RNAs modulate neuroinflammatory or synaptic gene expression [32, 83]. However, direct evidence linking UBL3-mediated sorting with disease phenotypes remains sparse. Future investigations may employ UBL3-knockdown or overexpressing models to examine effects on disease progression, such as α-synuclein accumulation in PD models or cognitive decline in Alzheimer’s models [84, 85].

In summary, UBL3 serves as a critical PTM-based gatekeeper for protein loading into sEVs, with emerging evidence pointing toward its involvement in neural disease pathways via cargo sorting of α-synuclein and mHTT. Coupled with other PTMs like SUMOylation, palmitoylation, ubiquitination, and glycosylation, these regulatory networks fine-tune EV composition and function. A deeper mechanistic understanding of UBL3 and these PTMs in neural EVs could illuminate new therapeutic targets for neurodegenerative diseases.

EVs serve as vehicles for the intercellular transfer of pathogenic factors within the brain, thereby facilitating the progression of major neurological disorders. AD, both tau and Aβ peptides are sequestered into EVs, which can propagate pathological aggregates to healthy neurons [86]. Tau-containing EVs isolated from postmortem AD brains have been shown to seed tau phosphorylation and neurofibrillary tangle formation in mouse models with as little as 300 pg of EV-associated tau [87]. Cryo-EM and proteomic studies further confirm that EVs specifically carry truncated tau filaments tethered to EV membranes, implicating active sorting mechanisms in EV-mediated propagation [88].

PD exhibits similar EV‑mediated propagation of α-synuclein. Neuron-derived EVs from PD patients contain oligomeric and phosphorylated forms of α-synuclein. These EVs elicit neuroinflammatory responses by activating microglia, amplifying disease progression [89, 90]. Conversely, mesenchymal stem cell (MSC)-derived EVs containing catalase exhibit neuroprotective properties in dopaminergic neurons. This demonstrates the dual nature of EVs as both pathological mediators and therapeutic agents in PD [91, 92].

In ALS, EVs mediate intercellular toxicity via TDP‑43 and mutant SOD1 cargo. Postmortem tissue and biofluid‑derived EVs from ALS patients contain full-length TDP‑43 and pathogenic C‑terminal fragments, which induce protein aggregation in recipient cells [93, 94]. The EV-mediated dispersal of TDP-43 correlates with disease severity in ALS and distinguishes patients at both pre- and peri-symptomatic stages [95].

Beyond neurodegeneration, EVs contribute to pathology in MS, stroke, and epilepsy through immune modulation and barrier dysfunction. In MS, EVs derived from activated microglia and immune cells carry pro-inflammatory cytokines, miRNAs, and matrix metalloproteinases that disrupt the BBB and promote demyelination [96, 97]. During a stroke, astrocyte-derived EVs deliver inflammatory miRNAs to endothelial cells, increasing tight junction permeability and exacerbating edema [98]. In epilepsy, alterations in blood-derived EV protein signatures suggest a role in seizure propagation through neurovascular unit disruption, but specific mechanisms remain to be clarified [99, 100].

Collectively, these studies indicate that EVs shuttle both toxic and inflammatory cargo across cell types and anatomical boundaries, thereby amplifying pathological cascades across the CNS. EVs enhance the seeding of misfolded proteins, promote neuroinflammation, and breach protective barriers, such as the BBB, thereby integrating multiple disease pathways. Targeting EV release or their uptake provides a promising therapeutic avenue. For instance, blockade of EV biogenesis via neutral sphingomyelinase inhibitors reduces tau spread in AD models, while reducing microglial EV release attenuates neuroinflammation in MS [101–103]. However, these approaches require fine-tuning to avoid disrupting homeostatic EV functions, which are essential for neural health. Table 3 summarizes EV involvement in neurological disorders across multiple disease categories, highlighting pathogenic mechanisms, therapeutic applications, and biomarker potential.

These findings establish EVs as critical contributors to neurological disease progression while revealing opportunities for therapeutic intervention. Table 3 consolidates current evidence demonstrating how EVs facilitate disease progression through the intercellular transfer of pathogenic cargo while simultaneously offering promising avenues for biomarker development and therapeutic applications across diverse neurological conditions.

This review summarizes current understanding of EV roles in neurological disorders, with particular emphasis on UBL3 as an essential regulator of EV biogenesis. UBL3 functions as a PTM factor governing protein sorting into sEVs, with Ubl3-knockout mice exhibiting a 60% reduction in protein content in sEVs compared to controls [59]. Recent proteomic analyses identified 1,241 UBL3-interacting proteins, including Ras, revealing extensive regulatory networks [59]. Crucially, UBL3 influences α-synuclein sorting to EVs, with this interaction upregulated under oxidative stress through microsomal glutathione S-transferase 3 [31], providing mechanistic insights into protein aggregation in PD.

EVs demonstrate remarkable therapeutic potential across neurological conditions. Pre-clinical evidence strongly supports exosome therapy for stroke, demonstrating significant improvements in neurological outcomes across multiple rodent studies [190]. Hypoxic-preconditioned MSC-derived sEVs promote spinal cord injury recovery through the miR-21/JAK2/STAT3 pathway [191], while intranasally administered MSC-EVs inhibit NLRP3-p38/MAPK signaling after traumatic brain injury [192]. miRNA cargo in EVs, particularly miR-181b, plays a crucial role in promoting angiogenesis and neurological recovery after ischemic stroke [193], and astrocyte-derived exosomal miR-138-5p provides neuroprotection through the regulation of mitophagy [194].

Despite progress, fundamental knowledge gaps persist in EV biology and therapeutic implementation. Current challenges include inconsistent nomenclature, complex characterization protocols, and underdeveloped production methodologies. Clinical trials predominantly utilize bulk EV populations rather than specific subpopulations that may offer enhanced efficacy [195]. Critical questions require investigation, including precise cargo selection mechanisms, the role of UBL3 in disease-specific protein sorting, and the determinants of EV targeting specificity [33, 196].

While EVs can cross the BBB through transcytosis (vesicular transport across endothelial cells), the precise molecular mechanisms governing BBB penetration require further investigation [197, 198]. Engineering approaches enhancing BBB penetration show promise for CNS drug delivery [197, 198]. Transport studies demonstrate differential pharmacokinetics depending on cellular origin and surface characteristics [199, 200]. In vivo validation must address optimal preconditioning protocols, administration routes, biodistribution patterns, and long-term safety assessments [199, 200]. Standardized isolation, characterization, and quality control protocols represent urgent priorities for clinical translation [162].

EV-based liquid biopsies represent a paradigm shift in the detection of neurological diseases. Brain-derived EVs from peripheral biofluids provide CNS molecular signatures through minimally invasive procedures [201]. Breakthrough studies identified plasma EV tau and TDP-43 as specific diagnostic biomarkers for frontotemporal dementia (FTD) and ALS [202]. miRNA profiles in neuron-derived EVs serve as potential biomarkers for neurodegenerative diseases [203], with specific alterations identified in psychiatric disorders, including depression and anxiety [204].

The growing demand for personalized neurological treatments drives increased investment in EV-based biomarker development and clinical applications [205]. Machine learning algorithms achieve diagnostic accuracies exceeding 0.88 ROC-AUC scores in neurological disease detection [206]. Gene editing technologies integrated with EV delivery represent the cutting edge of therapeutic development. CRISPR/Cas9 ribonucleoprotein complexes delivered via engineered EVs enable precise targeting of disease-causing mutations while minimizing off-target effects [207]. Recent advances show promise for treating AD, PD, ALS, and Huntington’s diseases through genetic modifications. Engineered EVs deliver therapeutic RNA molecules into the brain, overcoming traditional nucleic acid delivery barriers [138, 208].

EVs embody a fundamental duality in neurological disorders. As pathological “messengers”, they facilitate intercellular transfer of misfolded proteins, including α-synuclein, tau, and TDP-43, contributing to disease progression through prion-like spreading mechanisms (whereby misfolded proteins propagate between cells and template misfolding in recipient cells) [209, 210]. Altered miRNA profiles in CSF exosomes from Parkinson’s and Alzheimer’s patients demonstrate their pathological communication role [202]. Conversely, as therapeutic “modulators”, EVs offer unprecedented neuroprotection and regeneration opportunities. They attenuate reactive gliosis, reduce neuronal death, suppress pro-inflammatory signaling, and improve cognitive functions [211, 212]. Microglia-derived EVs actively participate in stroke neuroprotection and neurorepair [213].

Advanced delivery systems revolutionize EV-based therapeutics. The combination of EV delivery with CRISPR technology enables precise genome editing for neurological disorders, offering potential cures for genetic conditions [214, 215]. Bibliometric analyses reveal exponential growth in CRISPR-EV research, indicating robust scientific interest and clinical translation potential [216]. Artificial intelligence integration with EV-based biomarker discovery, combined with patient-specific engineering approaches, promises personalized neurological medicine [217].

Understanding EV subtypes and their specific roles will enable safer, more effective, disease-tailored therapies [218]. The convergence of UBL3-mediated protein sorting, advanced EV engineering, personalized diagnostics, and precision gene editing positions the field for transformative clinical applications. As we advance toward precision medicine, EVs represent both the molecular messengers defining disease pathogenesis and therapeutic modulators reshaping treatment paradigms, offering new hope for patients with intractable neurological conditions [218, 219].