Extracellular vesicles (EVs) are membranous structures that are secreted by all cell types and that are found in all body fluids, as well as tissue culture supernatants (van Niel et al., 2018). While the term `extracellular vesicles` is used to annotate all cell-derived membrane-delimited vesicles, these structures are highly divergent and can be divided into two main groups, namely exosomes and microvesicles, that are characterized based on their biogenesis, content, morphology, and functional atributes (van Niel et al., 2018). Exosomes, referred to as small EVs, are the most widely studied and comprise a size distribution ranging from 30 nm to 150 nm (Johnstone et al., 1987). Exosomes arise from the inward budding of early endosomes that enables formation of intraluminal vesicles (ILVs) within multivesicular endosomes (MVEs) (Harding et al., 1984). MVEs can fuse with the plasma membrane, releasing the ILVs, now referred to as exosomes, into the extracellular space (Pan et al., 1985; Doyle and Wang, 2019). Contrary to exosomes, microvesicles bud directly from the plasma membrane into the extracellular environment and have a size distribution ranging from 100 nm to 1000 nm (Tricarico et al., 2017). While, historically characterized as means of eliminating cellular waste (Johnstone et al., 1987), it is now established that EVs play a critical role in intercellular communication, transporting cellular components such as microRNAs, mRNAs, and DNAs, lipids and proteins, in both physiological and diseased conditions (Colombo et al., 2014; Lo Cicero et al., 2015; Yanez-Mo et al., 2015; van Niel et al., 2018). The transfer of these molecules via EVs can have a range of effects on recipient cells, including alteration of gene expression, modulation of signaling pathways, induction of phenotypic changes, and influence on the cellular behavior (Raposo et al., 1996; Valadi et al., 2007; van Niel et al., 2018). The composition of EVs is cell type-specific and determined by the physiological state of the cell, along with environmental stimuli (Minciacchi et al., 2015). Here, we compile evidence for the potential of EVs to enhance our understanding of the brain and its associated disease mechanisms, particularly in the context of viral infections. Additionally, we will briefly review the challenging potential of EVs as new disease biomarkers.

Examining EVs in the context of viral infection might shed light into disease mechanisms, particularly in inaccessible organs such as brain. Here, we focus on selected viruses that can invade the CNS and the role of EVs in their interaction with CNS target cells.

Identifying biomarkers for neurological diseases, including those that result from viral infections, is especially challenging as most neurodegenerative disorders are confined to 1) an inaccessible organ, 2) a distinct region of the brain and 3) a particular subset of cells. However, various CNS cell types have been demonstrated to secrete EVs, including oligodendrocytes, neurons and astrocytes (Bakhti et al., 2011; Wu et al., 2017), suggesting that brain-derived EVs in body fluids could offer a non-invasive means of obtaining a representation of the entire brain and sustain new biomarker characterization.

Indeed, EVs offer multiple advantages as a unique source of new biomarkers due to their specificity, stability, ability to cross biological barriers and non-invasive sampling. Because the contents of EVs are shielded from degradation by DNases, RNases, and proteases, the characterization of EVs isolated from any biofluid allows for the identification of robust biomarkers, as opposed to the analysis of plain blood, urine, or CSF (Boukouris and Mathivanan, 2015).

The potential of EVs to cross biological barriers, including the BBB, has opened avenues in which CNS-derived EVs found in the CSF and in the blood could serve as non-invasive means to gain valuable insights into the physiological state of the brain (Ramos-Zaldivar et al., 2022). In patients with Parkinson’s disease (PD), multiple system atrophy (MSA) or progressive supranuclear palsy (PSP), plasma levels of oligodendrocyte- and neuron-derived EVs were shown to be increased compared to disease control group (Ohmichi et al., 2019). In AD patients, as compared to patients with frontotemporal dementia, neuron-derived blood exosomes were found to have higher levels of lysosome-associated membrane protein 1 (LAMP1) and cathepsin D (Goetzl et al., 2015). Another study found that astrocyte-derived EVs contained elevated levels of classical and alternative complement pathway proteins in AD patients as compared to matched controls (Goetzl et al., 2018). Furthermore, the analysis of EVs might enable the enrichment of low abundant proteins that are normally masked by high abundant ones, such as albumin, in crude biofluids (Upadhya et al., 2020). Yet, protein composition within EVs can be influenced by various factors, such as sample handling and storage conditions, necessitating standardized isolation and detection methods.

Among other cargos, nucleic acids encapsulated within EVs, including microRNAs (miRNAs), or messenger RNAs (mRNAs), and DNA fragments, have been demonstrated to reflect disease-specific molecular signatures, facilitating early detection and monitoring of disease progression such as mainly reported in the context of neurodegeneration (Raghav et al., 2022; Lim et al., 2023) or primary CNS tumors (Zhang et al., 2023; Hallal et al., 2024). Despite being an attractive source of EV-associated biomarkers, accurate EV-RNA/DNA quantitation faces key challenges due to low yields, influenced by sample source and isolation methods, further impacting data interpretation (Hill et al., 2013; Saugstad et al., 2017; Gandham et al., 2020). Overcoming these hurdles will be essential for reliable EV RNA/DNA characterization, particularly in disease diagnosis and therapeutic development.

Lipids associated with EVs, even if less studied to date especially as potential biomarkers, also may play pivotal roles in cellular signaling and neurotransmission within the CNS. Alterations in lipid profiles of CNS EVs may signify cellular dysfunction and disease progression, offering novel avenues for biomarker discovery such as envisioned in PD or AD (Vanherle et al., 2020). However, comprehensive lipidomic analysis of EVs requires specialized equipment and expertise further hindering transfer into routine laboratory testing as diagnosis or prognostic markers.

Another challenge in approaching CNS specific biomarkers is establishing tissue-specific EV isolation technics. To isolate brain-derived EVs from the blood, current strategies are typically based on immunoprecipitation techniques that utilize antibodies targeting cell-type specific markers: oligodendrocyte-myelin glycoprotein (MOG) for oligodendrocytes, neural cell adhesion molecule L1 (L1CAM) for neurons, and glutamate aspartate transport protein (GLAST) for astrocytes. However, the reliability of these markers for isolating ‘brain-derived’ EVs has been questioned, as these markers are also widely expressed by cells outside the CNS (Kramer-Albers and Hill, 2016). Furthermore, in vitro studies have suggested that these markers may not consistently be incorporated into EVs (Norman et al., 2021; You et al., 2022). Nonetheless, since brain-derived EVs in the periphery might present a real-time source of information relevant to non-accessible organ such as the brain, efforts are underway to identify new brain cell-type specific membrane makers. Research about EVs as a diagnostic tool for viral infectious diseases is still in its early stages, yet it holds significant promise. CSF EVs derived from glial cells, neurons and the choroid plexus of patients with HIV-associated neurocognitive disorders were shown to comprise proteins involved in stress responses and immune/inflammatory responses, suggesting their potential as a source of biomarkers to neurological disease activity (Guha et al., 2019). In another study, plasma neuron-derived EVs from HIV-positive patients with neurological complications showed an enrichment of neurofilament-light chain (NFL), high mobility group box 1 (HMGB1) and amyloid beta (Aβ) proteins as compared to controls (Sun et al., 2017). More recently, neuron-derived EVs isolated from patients experiencing long term effects of coronavirus disease 2019 (COVID-19) were found to contain elevated levels of proteins indicative of neuronal damage as compared to pre-COVID-19 controls (Tang et al., 2024). As such, neuron-derived EVs might provide a non-invasive source of biomarkers to identify cognitive impairment related to HIV-infection. Similarly, plasma-derived EVs expressing the astrocytic marker, glial fibrillary acidic protein (GFAP), were significantly increased in patients suffering from HAND as compared to patients with normal cognition, further supporting the relevance of brain-derived EVs as indicators of neuronal injury (de Menezes et al., 2022).

Finally, the translation of EV research into clinical applications is a concern, particularly regarding their potential as a novel source of biomarkers. Indeed, before EVs can be widely accepted as disease biomarkers, it is imperative to overcome the technical challenges posed by their isolation, characterization, methodological standardization, and reporting (Van Deun et al., 2017; Roux et al., 2020; Welsh et al., 2024). The chosen method for isolating EVs significantly affects both sample yield and purity. Currently, there is no unified standard for exosome separation and purification. Indeed, it exists various methods including differential ultracentrifugation (UC), density gradient ultracentrifugation (DGUC), size exclusion chromatography (SEC), filtration, precipitation or immune-based extraction or more recently microfluidic approaches; each of those have their advantages and limitations (Xu et al., 2016; He et al., 2022). Technical challenges in isolating EVs consistently, along with complex analysis techniques, may hinder their use as powerful biomarkers. Pre- and post-analytical variability can also impact the presence of non-vesicular particles together with the surface and content of specific EVs (Clayton et al., 2019; Buzas, 2022; Görgens et al., 2022). Prioritizing reproducibility and ease-of-use is crucial for effective biomarker discovery in any diseases. All these challenges are currently addressed. Guidelines and position papers have been recently published for blood and CSF (Hendrix, 2021; Lucien et al., 2023; Sandau et al., 2024). Resolving these challenges is essential for harnessing the full diagnostic potential of EVs in clinical settings.

The intricate relationship between EVs and viruses operates on multiple levels. EVs might serve as a vehicle of viral propagation. Moreover, transported viral or cellular content might evoke distinct responses in recipient cells ( Figure 1A ). As per virus point-of view, these responses can either be advantageous, priming cells for infection, or detrimental, hindering viral replication. Nevertheless, our comprehension of the role of EVs in viral infection, especially those pertaining to the CNS, remains incomplete, primarily because of the difficulty of isolating single cell specific EVs in vivo. While significant insights have been gained through cell culture models, the complexity of the CNS cellular composition and microenvironment is not fully captured yet. Similarly, while animal models offer valuable insights into EV dynamics in a physiological context, they may fail to mimic certain developmental or pathological features that are unique to humans, including specific human host-pathogen interactions ( Figure 1A ) (Bárbara and Sónia, 2017; Swearengen, 2018). To address these challenges, future research should focus on developing more physiologically relevant in vitro human models, such as organoid cultures or microfluidic systems, that better mimic the complexity of the CNS microenvironment ( Figure 1B ). The capacity of EVs to cross biological barriers, including the BBB, means that brain-derived EVs could be identified in the plasma. Characterization of brain-derived Evs in the blood might thus be used for diagnostic and prognostic assessment of various neurological disorders and real-time monitoring of patient responses to targeted therapies. However, current techniques for evaluating brain-derived EVs in CSF and blood face challenges due to the need for sensitive and specific isolation methods. Subsequently, there is an urgent need to standardize brain-derived EV isolation and identification methods to support efficient disease diagnosis ( Figure 1C ) (Sandau et al., 2024). Nevertheless, while research into brain-derived EVs in the context of virus-induced neurological diseases is still in its infancy, it holds significant promise as EVs 1) could serve as valuable sources of information that can help to elucidate intricate cellular processes within an inaccessible organ such as the brain and 2) hold promise for advancing clinical diagnoses.

LO: Conceptualization, Writing – review & editing, Writing – original draft. RP: Writing – review & editing, Funding acquisition, Supervision. AM: Supervision, Writing – review & editing, Conceptualization, Investigation, Project administration, Visualization.