Skip to content
Frontiers
Search

Review

Frontiers in Lab on a Chip Technologies ·

Volume 5 - 2026 | https://doi.org/10.3389/frlct.2026.1915723

Extracellular vesicle-integrated brain organoid-on-a-chip systems for neuroimmune communication and precision therapy in viral encephalitis

  • Md. M. N. Azim1*corresponding author
  1. Department of Physiology, College of Veterinary Medicine, Jeonbuk National University, Iksan, Republic of Korea
Article metrics0 viewsView details
PDFCiteMetrics

Abstract

Viral encephalitis causes substantial neurological morbidity and mortality through the combined effects of direct viral injury, blood–brain barrier (BBB) dysfunction, and dysregulated communication among neural, vascular, and immune cells. Extracellular vesicles (EVs) are increasingly recognized as mediators of this communication because they transfer proteins, lipids, viral constituents, cytokine-related signals, and regulatory RNAs between neurons, astrocytes, microglia, endothelial cells, and peripheral immune cells. Depending on their cellular origin and cargo, EVs may support antiviral defense or propagate viral dissemination, neuroinflammation, barrier injury, and persistent neurological sequelae. This narrative review synthesizes evidence for an EV–neuroimmune axis in viral encephalitis and evaluates how that evidence could inform the design of brain organoid-on-a-chip (BOoC) models. The review distinguishes established findings from a proposed integrated platform that would combine patient-derived brain organoids, a BBB-interposed vascular and immune compartment, real-time EV tracking, transendothelial electrical resistance and electrophysiological sensing, and multi-omics analysis. Existing studies support several individual components of this framework, but a fully integrated EV–BOoC system with closed-loop artificial intelligence (AI) control has not yet been demonstrated for viral encephalitis. Accordingly, digital twins, autonomous treatment optimization, and patient-specific clinical prediction are presented as future research directions rather than current capabilities. This calibrated framework identifies measurable design requirements, evidence gaps, and translational priorities for mechanistic studies, biomarker discovery, and precision therapeutic screening.

Keywords

1 Introduction

Viral encephalitis is a severe inflammatory disease of the brain caused by neurotropic viruses, with an estimated annual incidence of 7–15 cases per 100,000 people and a disproportionate burden in low- and middle-income countries (Granerød et al., 2023). Untreated herpes simplex virus type 1 (HSV-1) encephalitis has a mortality approaching 70%; even with acyclovir treatment, 30%–50% of adults develop persistent seizures, cognitive impairment, or hemiparesis (Bradshaw and Venkatesan, 2016). Encephalitis-related hospitalizations in the United States are associated with annual healthcare costs exceeding $2 billion (Kiyani et al., 2020). The global burden also encompasses emerging and re-emerging neurotropic RNA viruses, such as Zika virus (ZIKV), West Nile virus (WNV), Japanese encephalitis virus (JEV), and SARS-CoV-2 (Caldwell et al., 2022).

Disease severity is not determined by viral cytotoxicity alone. It also reflects disruption of neuroimmune networks linking neurons, astrocytes, microglia, endothelial cells, and infiltrating peripheral immune populations (Pavlou et al., 2024). Pro-inflammatory mediators such as TNF-α increase BBB permeability and stimulate glial activation, creating a neuroinflammatory loop that can persist after viral clearance (Pavlou et al., 2024) (Figure 1). This sustained dysregulation contributes to the cognitive, psychiatric, and epileptic sequelae observed in survivors.

Blood–brain barrier disruption in viral encephalitis. Schematic comparison of a healthy neurovascular unit and viral encephalitis–induced BBB breakdown. In the healthy state, endothelial cells, tight junction proteins, pericytes, astrocyte end-feet, and basement membrane maintain BBB integrity and CNS homeostasis. During viral encephalitis, neurotropic viruses and pro-inflammatory cytokines disrupt tight junctions, activate endothelial cells (VCAM-1, ICAM-1), and increase BBB permeability, leading to leukocyte infiltration, microglial activation, astrocyte dysfunction, and vasogenic edema. These changes facilitate viral entry into the CNS and drive a self-amplifying cycle of neuroinflammation and neural injury.
Figure 1. Blood–brain barrier disruption in viral encephalitis. Schematic comparison of a healthy neurovascular unit and viral encephalitis–induced BBB breakdown. In the healthy state, endothelial cells, tight junction proteins, pericytes, astrocyte end-feet, and basement membrane maintain BBB integrity and CNS homeostasis. During viral encephalitis, neurotropic viruses and pro-inflammatory cytokines disrupt tight junctions, activate endothelial cells (VCAM-1, ICAM-1), and increase BBB permeability, leading to leukocyte infiltration, microglial activation, astrocyte dysfunction, and vasogenic edema. These changes facilitate viral entry into the CNS and drive a self-amplifying cycle of neuroinflammation and neural injury.

Conventional experimental models do not fully reproduce human disease. Animal models differ from humans in BBB biology, immune responses, and brain architecture, which can contribute to translational failure (Rossi et al., 2018). Two-dimensional (2D) cell cultures lack the three-dimensional (3D) cytoarchitecture and mechanical environment of brain tissue (Wang et al., 2024a), whereas conventional static organoids generally lack functional vasculature and a complete immune compartment. Consequently, these models are poorly suited to studying dynamic interactions among human neurons, astrocytes, endothelial cells, and immune cells. BOoC platforms can address some of these limitations by combining iPSC-derived 3D neural tissue with microfluidic perfusion and real-time monitoring of barrier integrity, cytokine dynamics, and electrophysiological activity (Wang et al., 2023).

Within multicellular CNS models, EVs are released by nearly all cell types and contribute to intercellular communication (Oberholster et al., 2024). EVs can cross biological barriers and participate in both protective and pathogenic neuroimmune responses (Horn and MacLean, 2021). Their dual roles as disease mediators and potential therapeutic carriers make them relevant to biomarker discovery, mechanistic investigation, and precision drug delivery (Choi et al., 2024).

Previous reviews have established that EVs can contribute to viral immune evasion, CNS invasion, and glia-associated neurodegeneration (Horn and MacLean, 2021). The distinct contribution of the present narrative review is to translate that evidence into a model-design framework that connects EV biology with BBB-competent BOoC systems, EV-omics, and AI-assisted analysis. The EV–neuroimmune axis is therefore used here as an organizing framework, not as a newly discovered biological pathway, to define measurable modules, distinguish current evidence from prospective capabilities, and identify validation priorities for viral encephalitis research.

2 Viral encephalitis as a multicellular neuroimmune disorder

2.1 Neurotropic viruses and CNS invasion mechanisms

Neurotropic viruses invade the CNS through three principal routes: hematogenous spread, in which circulating virus infects or crosses BBB endothelial cells; retrograde axonal transport along peripheral nerves, as exemplified by HSV-1 spread from trigeminal pathways to the temporal lobe; and a “Trojan horse” mechanism in which infected monocytes or macrophages cross the BBB (Bradshaw and Venkatesan, 2016; Yachou et al., 2020; Rocamonde et al., 2023). Table 1 summarizes the invasion strategies, cellular tropism, clinical features, and sequelae of six major neurotropic viruses.

TABLE 1. Neurotropic viruses: CNS invasion mechanisms and clinical features.

VirusCNS invasion routePrimary target cellPeak clinical syndromeMortality rateLong-term sequelaeReferences
HSV-1Retrograde axonal transport from trigeminal/olfactory nervesNeurons (limbic, orbitofrontal, mesiotemporal)Focal necrotizing encephalitis, fever, seizures, altered consciousness∼70% (untreated); 20%–30% with acyclovirCognitive impairment, epilepsy, memory deficits (30%–50%)(Bradshaw and Venkatesan, 2016; McGrath et al., 1997)
ZIKVHematogenous; Trojan horse (infected monocytes)Neural progenitor cells, neurons, gliaMicrocephaly (congenital); Guillain-Barré syndrome (adults)<1% adults; high neonatal mortalityDevelopmental delay, microcephaly, hearing loss(Salimi et al., 2016; Caldwell et al., 2022)
WNVHematogenous after mosquito inoculationNeurons (anterior horn), microgliaFlaccid paralysis, meningitis, encephalitis∼10% neuroinvasive (3%–15%); 35% in adults >65 yearsPersistent flaccid paralysis, cognitive decline(Petersen et al., 2013; Team et al., 2014)
SARS-CoV-2Hematogenous; olfactory nerve; enteric; choroid plexusChoroid plexus, endothelial cells, neuronsAcute encephalitis, stroke, neuropsychiatric manifestations0.5%–2% neurological forms; elevated in elderlyLong COVID: brain fog, fatigue, depression (up to 30%)(Yang et al., 2021; Deffner et al., 2020)
Rabies virusRetrograde axonal transport from peripheral nervesNeurons (brain stem, hippocampus, limbic system)Hydrophobia, agitation, autonomic failure, comaVirtually 100% once symptomaticUniformly fatal; survivors extremely rare(Davis et al., 2015; Fisher et al., 2018)
JEVHematogenous (mosquito); possible olfactory routeNeurons (thalamus, basal ganglia, brainstem)Acute encephalitis, parkinsonian features, seizures20%–30% of neuroinvasive casesNeurological sequelae in ∼50% survivors; parkinsonism, paralysis, seizures(Misra, 2020; Chaudhuri and Kennedy, 2002)

2.2 Blood–brain barrier disruption

Viral entry and localized immune infiltration disrupt the tight-junction complex, including occludin, claudin-5, and zonula occludens-1, that maintains BBB impermeability (Pediaditakis et al., 2022; Choi et al., 2023; Amartumur et al., 2024). Loss or redistribution of these proteins increases paracellular permeability, allowing plasma proteins, toxins, and inflammatory mediators to enter the neuropil (Pediaditakis et al., 2022; Choi et al., 2023). TNF-α and IL-6 can sustain barrier injury while inducing vascular cell adhesion molecule 1 (VCAM-1) and intercellular adhesion molecule 1 (ICAM-1) on endothelial cells, thereby promoting leukocyte adhesion and transendothelial migration (Pediaditakis et al., 2022; Rocamonde et al., 2023; Pavlou et al., 2024). Microengineered brain-chip studies further show that endothelial cells actively secrete chemokines and contribute to T-cell and monocyte recruitment, rather than functioning only as a passively damaged barrier (Pediaditakis et al., 2022; Choi et al., 2023; Wang et al., 2023).

2.3 Cellular participants in neuroimmune communication

Many neurotropic viruses use neurons as major sites of replication while simultaneously inducing type I interferon responses (Chen et al., 2019; Pavlou et al., 2024). Astrocytes support BBB integrity and neurotrophic function but become reactive during infection and release cytokines and chemokines that shape the innate immune response (Tavčar et al., 2021; Rocamonde et al., 2023; Pavlou et al., 2024). Microglia, the resident macrophages of the CNS, can rapidly adopt pro-inflammatory states characterized by reactive oxygen species (ROS) and inflammatory cytokines that aid viral containment but may also cause bystander neuronal injury (Chen et al., 2019; Rocamonde et al., 2023; Pavlou et al., 2024). Endothelial activation recruits peripheral CD8+ T cells and inflammatory monocytes; although these cells may contribute to viral clearance, their infiltration can also damage surrounding neural tissue (Pediaditakis et al., 2022; Pavlou et al., 2024).

2.4 Acute and chronic consequences of neuroinflammation

A cytokine storm occurs due to the uncontrolled release of TNF-α, IL-6, and type I interferons during the acute phase of neural injury, leading to cerebral edema, neuronal apoptosis, and excitotoxicity (Chen et al., 2019; Rocamonde et al., 2023; Pavlou et al., 2024). Even after viral clearance, glial reactivity and chronic signaling via EVs sustain low-level neuroinflammation that gradually disrupts synaptic plasticity and neural circuit stability, resulting in cognitive impairment, seizure vulnerability, and eventual neurodegeneration (Piret and Boivin, 2020; Rocamonde et al., 2023; Xu et al., 2023; Oberholster et al., 2024). Most importantly, disease severity is dictated not only by viral load but also by the functional status of neuroimmune communication networks. Thus, restoration of these networks ranks among the primary therapeutic imperatives (Bradshaw and Venkatesan, 2016; Piret and Boivin, 2020; Pavlou et al., 2024).

3 Brain organoid-on-a-chip systems: toward human-relevant neuroinfection models

3.1 Evolution of in vitro CNS models

Progression from 2D monolayers to EV-integrated BOoC platforms is motivated by three limitations of conventional models: limited physiological relevance, inadequate vascularization, and incomplete immune representation (Figure 2) (Pediaditakis et al., 2022; Amartumur et al., 2024; Wang et al., 2024b). Two-dimensional cultures lack 3D cytoarchitecture and extracellular-matrix mechanics, simple spheroids provide limited tissue organization, and static brain organoids may develop necrotic cores and remain immunologically incomplete. Organ-on-chip technology adds perfusion, shear stress, spatial control, and real-time sensing (Pediaditakis et al., 2022). An EV-aware BOoC would extend these capabilities by enabling vesicle sampling and trafficking analysis; however, this fully integrated configuration remains a proposed architecture rather than an established viral encephalitis model.

Evolution of experimental models for studying viral encephalitis, from conventional 2D culture to EV-integrated brain organoid-on-a-chip systems. Traditional 2D cell culture and simple 3D spheroid models offer limited physiological relevance, whereas brain organoids more closely recapitulate human neural development and cell diversity. Brain organ-on-chip platforms add microfluidic perfusion, mechanical control, and real-time monitoring, thereby improving the modeling of the neurovascular unit and the blood-brain barrier. The proposed EV-integrated brain organoid-on-a-chip system incorporates neural, vascular, and immune compartments to study extracellular vesicle-mediated neuroimmune communication during viral encephalitis.
Figure 2. Evolution of experimental models for studying viral encephalitis, from conventional 2D culture to EV-integrated brain organoid-on-a-chip systems. Traditional 2D cell culture and simple 3D spheroid models offer limited physiological relevance, whereas brain organoids more closely recapitulate human neural development and cell diversity. Brain organ-on-chip platforms add microfluidic perfusion, mechanical control, and real-time monitoring, thereby improving the modeling of the neurovascular unit and the blood-brain barrier. The proposed EV-integrated brain organoid-on-a-chip system incorporates neural, vascular, and immune compartments to study extracellular vesicle-mediated neuroimmune communication during viral encephalitis.

3.2 Brain organoids

Brain organoids are self-organizing 3D tissues derived from human pluripotent stem cells that reproduce selected features of regional brain development, cellular diversity, and tissue architecture (Azim et al., 2026). Cerebral, cortical, and forebrain organoids can model radial glial organization, inside-out cortical layering, and region-specific gene expression that are difficult to reproduce in rodent models (Rossi et al., 2018; La Rosa et al., 2026). Assembloids formed by fusing regionally specified organoids can additionally model neuronal migration and interregional circuit formation (Rossi et al., 2018; Wang et al., 2024a). These systems are relevant to the present review because they provide a human 3D context in which EV-mediated communication can be examined across neural cell types (La Rosa et al., 2026). Their limitations include absent or incomplete vasculature, restricted oxygen diffusion, necrotic cores in larger tissues, and incomplete representation of microglia and peripheral leukocytes (Ao et al., 2021; Amartumur et al., 2024; Wang et al., 2024b).

3.3 Organ-on-chip technology

Microfluidic OoC platforms address vascular and mechanical limitations by providing controlled perfusion that transports nutrients, removes metabolites, and applies physiologically relevant shear stress to endothelial cells (Pediaditakis et al., 2022; Amartumur et al., 2024). Defined biochemical gradients enable spatiotemporal analysis of viral spread and chemokine-driven immune recruitment. Integrated transendothelial electrical resistance (TEER) sensors permit non-invasive monitoring of barrier integrity, whereas multielectrode arrays (MEAs) record neuronal network activity (Pediaditakis et al., 2022; Choi et al., 2023). A recent BBB organ-on-chip study of Eastern equine encephalitis virus further demonstrated the value of these systems for modeling virus-induced barrier injury and assessing candidate countermeasures (Boghdeh-Olson et al., 2026).

3.4 Brain organoid-on-chip platforms and conceptual architecture

The proposed EV-integrated BOoC architecture comprises four functionally connected modules (Figure 3). First, a neural compartment contains an iPSC-derived brain organoid with neurons, astrocytes, and, where feasible, incorporated microglia (Ao et al., 2021). Second, a BBB partition—composed of brain microvascular endothelial cells with supporting pericytes and astrocyte end-feet—is physically interposed between the vascular and neural compartments (Choi et al., 2023). Third, the luminal vascular compartment receives T cells and monocytes so that adhesion and transendothelial migration can be measured across the BBB rather than through direct continuity with the organoid (Pediaditakis et al., 2022). Fourth, a sensing module combines EV sampling or capture, TEER electrodes spanning the barrier, and an MEA beneath the neural compartment. Individual components are supported by existing studies, but their integration into a single EV-focused viral encephalitis platform remains to be experimentally validated.

Proposed EV-integrated brain organoid-on-a-chip architecture for viral encephalitis. The BBB partition is explicitly interposed between the luminal vascular/peripheral immune compartment and the abluminal neural organoid compartment. T cells and monocytes are introduced into the vascular channel, and any movement toward the organoid therefore requires adhesion to and migration across the endothelial barrier. TEER electrodes span the BBB partition, whereas an MEA records neural activity beneath the organoid. EV capture and molecular sensing can sample both compartments to examine vesicle trafficking and neuroimmune signaling. The figure represents a conceptual design assembled from validated component technologies; the complete integrated platform has not yet been demonstrated.
Figure 3. Proposed EV-integrated brain organoid-on-a-chip architecture for viral encephalitis. The BBB partition is explicitly interposed between the luminal vascular/peripheral immune compartment and the abluminal neural organoid compartment. T cells and monocytes are introduced into the vascular channel, and any movement toward the organoid therefore requires adhesion to and migration across the endothelial barrier. TEER electrodes span the BBB partition, whereas an MEA records neural activity beneath the organoid. EV capture and molecular sensing can sample both compartments to examine vesicle trafficking and neuroimmune signaling. The figure represents a conceptual design assembled from validated component technologies; the complete integrated platform has not yet been demonstrated.

3.5 Current applications in viral infection research

BOoC and organoid models provide mechanistic information that is difficult to obtain from conventional systems (Wang et al., 2024b). ZIKV infection of neural progenitor cells in cerebral organoids reproduces radial glial loss and cortical thinning associated with congenital microcephaly (Salimi et al., 2016; Chen et al., 2019). Choroid plexus and barrier models have clarified how SARS-CoV-2 can disrupt blood–CSF interfaces without requiring widespread productive neuronal infection (McQuaid et al., 2021; Thompson et al., 2022). In HSV-1 cerebral organoids, combined single-cell RNA sequencing, spatial transcriptomics, electrophysiology, and immunostaining showed that acyclovir markedly reduced viral replication but did not fully suppress TNF–NF-κB signaling or tissue injury (Rybak-Wolf et al., 2023). More recent studies have extended the field: a BBB–cerebral organoid co-culture demonstrated that EVs from HIV-infected T cells and microglia could cross the barrier, induce IL-1β and TNF-α, and activate inflammatory and cell-death pathways in organoids (Yang et al., 2026), while a BBB chip enabled assessment of countermeasures against Eastern equine encephalitis virus (Boghdeh-Olson et al., 2026). Table 2 compares representative models and their current limitations.

TABLE 2. Brain organoid and organoid-on-chip models for viral encephalitis research.

ModelCell compositionVirus studiedAdvantagesLimitationsReferences
Traditional Cerebral OrganoidsiPSC-derived neurons, astrocytes, radial glia, neural progenitorsHSV-1, WNV, ZIKV, SARS-CoV-2Human-specific neuroanatomy; fetal-like 3D architecture; preserves patient geneticsNo functional vasculature; necrotic core; immune-sterile; no shear stress(Rossi et al., 2018; Chen et al., 2019)
Tubular Brain OrganoidsiPSC-derived neurons, astrocytes, integrated microgliaMicroglia-mediated neuroinflammation modelsIncorporates resident immune cells; microfluidic perfusion reduces necrosis; physiological glial phenotypesLacks peripheral immune components; high technical complexity(Ao et al., 2021)
Neurovascular Unit-on-a-ChipEndothelial cells, pericytes, astrocytes, neurons in compartmentalized channelsSARS-CoV-2; general neuroinflammatory insultsFunctional BBB with physiological shear stress; real-time TEER and MEA monitoringSimpler than 3D organoids; lacks layered cortical architecture(Pediaditakis et al., 2022; Choi et al., 2023)
Brain AssembloidsFusion of regional organoidsZIKV-induced developmental defects; circuit disruptionModels long-range neuronal migration and complex inter-regional circuit formationVariable fusion efficiency; difficult to engineer at scale(Rossi et al., 2018; Wang et al., 2024a)
Integrated Organoid-on-a-Chip3D organoid coupled with vascular channels and immune compartmentsSARS-CoV-2, WNV, neuroinvasive RNA virusesCombines biological complexity with microfluidic control; enables leukocyte recruitment and EV trafficking studiesHigh cost; limited scalability for clinical applications(Amartumur et al., 2024; Wang et al., 2024b)
Choroid Plexus OrganoidsSpecialized blood–CSF barrier epithelial cellsSARS-CoV-2 neurotropism and barrier disruptionModels viral CSF entry; captures barrier leakage and inflammatory signalingLacks neuronal and glial diversity of cerebral organoids(Thompson et al., 2022)

Incorporating EV biology addresses this gap (Oberholster et al., 2024). By labeling and tracking EVs within vascular, immune, and neural compartments, researchers can map the molecular messages that drive viral encephalitis toward resolution or chronic neurodegeneration (Oberholster et al., 2024).

4 Extracellular vesicles as central mediators of neuroimmune communication

4.1 EV classification and biogenesis

EVs are lipid bilayer-enclosed vesicles that do not replicate (Choi et al., 2024; Oberholster et al., 2024). They contain proteins, lipids, and nucleic acids, which provide a snapshot of the progenitor cell’s physiological state (Shi et al., 2019). The major EV subclasses are defined by their biogenesis pathway and size (Xu et al., 2023; Oberholster et al., 2024). Exosomes originate from multivesicular bodies via the ESCRT machinery, whereas microvesicles bud directly from the plasma membrane (Shi et al., 2019; Choi et al., 2024). Apoptotic bodies, which contain organellar fragments, are released during apoptosis (Xu et al., 2023). The lipid bilayer protects EV cargo against extracellular proteases and nucleases, enabling the functional delivery of genetic information across biological barriers (Oberholster et al., 2024). Figure 4 shows a schematic of EV biogenesis in CNS cell types and relevant molecular cargo in viral encephalitis.

Biogenesis, cellular sources, and molecular cargo of extracellular vesicles (EVs) in the CNS. **(A)** EV biogenesis pathways showing exosomes, microvesicles, and apoptotic bodies. **(B)** CNS cellular sources of EVs, including neurons, astrocytes, microglia, and endothelial cells, with representative cargos. **(C)** Major EV cargo components, including nucleic acids, proteins, lipids, and immune-related molecules. **(D)** Size comparison of EVs and viruses, highlighting overlapping dimensions and potential roles in cargo transfer and CNS communication.
Figure 4. Biogenesis, cellular sources, and molecular cargo of extracellular vesicles (EVs) in the CNS. (A) EV biogenesis pathways showing exosomes, microvesicles, and apoptotic bodies. (B) CNS cellular sources of EVs, including neurons, astrocytes, microglia, and endothelial cells, with representative cargos. (C) Major EV cargo components, including nucleic acids, proteins, lipids, and immune-related molecules. (D) Size comparison of EVs and viruses, highlighting overlapping dimensions and potential roles in cargo transfer and CNS communication.

4.2 EVs in healthy CNS communication

Under homeostatic conditions, neurons release EVs loaded with miR-124 and neurotrophic factors to regulate both astrocyte metabolism and microglial synaptic surveillance (Shi et al., 2019; Oberholster et al., 2024). EVs released from astrocytes help convert the inflamed brain to a neurotrophic environment by supplying anti-inflammatory mediators and metabolic substrates required for neuronal survival (Oberholster et al., 2024). EVs released from brain microvascular endothelial cells inform pericytes and astrocytes about BBB health and transmit CNS state information into the bloodstream via liquid endocrinology (Shi et al., 2019; Xu et al., 2023). This ongoing vesicular exchange maintains immune privilege while keeping the CNS coherently linked to systemic physiological signaling (Xu et al., 2023).

4.3 EV-mediated viral dissemination

Neurotropic viruses can co-opt EV biogenesis and cargo-sorting pathways to facilitate spread and immune evasion (Oberholster et al., 2024). ZIKV-associated material, including viral RNA and proteins, has been reported in EV-enriched fractions, raising the possibility that vesicular transfer complements receptor-dependent infection routes. EVs released from SARS-CoV-2-infected cells may likewise contain viral or inflammatory cargo capable of activating recipient cells even when those cells do not support productive infection (Yachou et al., 2020; Oberholster et al., 2024). Because EV preparations can overlap in size and composition with virions and non-vesicular particles, these mechanisms require rigorous isolation, infectivity controls, and orthogonal characterization before “infectious vesicle” activity can be attributed specifically to EVs.

4.4 EV-mediated neuroinflammation and BBB dysfunction

Viral infection can shift the CNS EV secretome toward pro-inflammatory cargo (Oberholster et al., 2024). EV-associated damage- and pathogen-related signals may activate innate pathways such as TLR and NF-κB signaling in recipient microglia, thereby promoting TNF-α, IL-6, and type I interferon responses (Chen et al., 2019). Astrocyte- or endothelial-derived EVs may also contribute to barrier dysfunction by transporting inflammatory mediators or matrix-remodeling enzymes that alter tight-junction proteins. Recent BBB–organoid co-culture data provide direct support for this general mechanism: HIV-associated EVs reduced TEER, traversed a multicellular BBB, and induced inflammatory cytokines and apoptosis in cerebral organoids (Yang et al., 2026). Nevertheless, the direction and magnitude of EV effects are context dependent and should be distinguished from soluble co-isolates and direct viral injury.

4.5 EVs and post-viral neurological sequelae

Pathological signaling through EVs continues after the acute infection. Chronically activated astrocytes and microglia maintain a continual exchange of EVs that carry miR-155, neurotoxic proteins, and misfolded aggregates, leading to progressive deficits in synaptic plasticity (Oberholster et al., 2024). This explains the cognitive impairment, epilepsy, and neurodegeneration-like presentations in long-COVID and post-encephalitis syndromes (Figure 5) (Yachou et al., 2020; Oberholster et al., 2024). Targeted modulation of this EV signaling loop is a novel therapeutic strategy for restoring neural function (Choi et al., 2024). EV sources, cargo, mechanisms, and clinical relevance along the spectrum of viral encephalitis are summarized in Table 3.

Extracellular vesicle-mediated mechanisms underlying viral encephalitis and post-viral neurological sequelae. Under physiological conditions, EVs maintain neuron–glia communication, BBB integrity, and CNS homeostasis. During viral infection, neurotropic viruses exploit EV biogenesis pathways to generate infectious EVs that transfer viral RNA and proteins to recipient CNS cells. These EVs activate microglial TLR7/8–NF-κB signaling and promote the release of pro-inflammatory mediators, while astrocyte-derived EVs contribute to BBB disruption through matrix metalloproteinase-mediated degradation of tight junction proteins. Subsequent upregulation of ICAM-1 and VCAM-1 facilitates leukocyte infiltration, amplifying neuroinflammation. Persistent EV exchange between activated astrocytes and microglia sustains inflammatory signaling, contributing to cognitive impairment, epilepsy, synaptic dysfunction, and neurodegeneration-like changes following viral encephalitis.
Figure 5. Extracellular vesicle-mediated mechanisms underlying viral encephalitis and post-viral neurological sequelae. Under physiological conditions, EVs maintain neuron–glia communication, BBB integrity, and CNS homeostasis. During viral infection, neurotropic viruses exploit EV biogenesis pathways to generate infectious EVs that transfer viral RNA and proteins to recipient CNS cells. These EVs activate microglial TLR7/8–NF-κB signaling and promote the release of pro-inflammatory mediators, while astrocyte-derived EVs contribute to BBB disruption through matrix metalloproteinase-mediated degradation of tight junction proteins. Subsequent upregulation of ICAM-1 and VCAM-1 facilitates leukocyte infiltration, amplifying neuroinflammation. Persistent EV exchange between activated astrocytes and microglia sustains inflammatory signaling, contributing to cognitive impairment, epilepsy, synaptic dysfunction, and neurodegeneration-like changes following viral encephalitis.

TABLE 3. Roles of extracellular vesicles in viral encephalitis.

EV sourceMajor cargoTarget cellsMechanismBiological effectClinical/Therapeutic relevanceStatusReferences
Infected NeuronsViral RNA, viral proteins, miR-124, DAMPs, neurotoxic proteinsMicroglia, astrocytesActivates TLR7/8 and innate immune pathways; transfers viral signalsMicroglial activation, cytokine production, neuroinflammatory amplificationLiquid-biopsy biomarkers for early viral neuroinvasion detectionPreclinical(Furr and Marriott, 2012; Chen et al., 2019; Rocamonde et al., 2023; Oberholster et al., 2024)
Activated AstrocytesmiR-146a, IL-1β, TNF-α, matrix metalloproteinasesEndothelial cells, neurons, microgliaDownregulates tight-junction proteins; promotes inflammatory signalingBBB disruption, vascular leakage, synaptic dysfunction, neuroinflammationTarget for RNAi-based suppression of neuroinflammatory signalingPreclinical(Tavčar et al., 2021; Pediaditakis et al., 2022; Rocamonde et al., 2023; Pavlou et al., 2024)
Reactive MicrogliamiR-155, TNF-α, IL-6, ROS-associated proteinsHealthy neurons, astrocytesActivates NF-κB; propagates inflammatory responses to uninfected cellsSecondary neuronal injury, apoptosis, synaptic loss, neurodegenerationBiomarker source for monitoring neuroinflammatory burden and cytokine stormPreclinical/Translational(Chen et al., 2019; Rocamonde et al., 2023; Pavlou et al., 2024)
Brain Endothelial CellsICAM-1, VCAM-1, inflammatory miRNAs, adhesion moleculesPeripheral T cells, monocytes, astrocytesFacilitates leukocyte adhesion and transendothelial migration into CNSEnhanced immune-cell infiltration, BBB dysfunction, neuroinflammationTargeting EV-mediated leukocyte recruitment to reduce acute neuroinflammationPreclinical(Pediaditakis et al., 2022; Choi et al., 2023; Amartumur et al., 2024)
Peripheral Monocytes/MacrophagesCytokines, viral RNA, inflammatory miRNAs, antigen-presenting moleculesEndothelial cells, microglia, neurons“Trojan horse” viral dissemination; immune activation across BBBCNS viral invasion and amplification of inflammatory cascadesTarget to prevent CNS viral entry and peripheral-to-central immune signalingExperimental(Salimi et al., 2016; Yachou et al., 2020; Rocamonde et al., 2023; Pavlou et al., 2024)
MSC-Derived EVsmiR-21, TGF-β, BDNF, GDNF, anti-inflammatory cytokinesDamaged neurons, astrocytes, microgliaSuppresses inflammatory pathways; promotes tissue repair and remyelinationNeuroprotection, axonal regeneration, and functional recoveryCandidate EV-based regenerative therapyClinical trials ongoing(Tian et al., 2022; Choi et al., 2024; Li et al., 2025a; Sanadgol et al., 2025)
Engineered iPSC-Derived EVsAntiviral siRNA, miRNA mimics, Cas9/sgRNA, mRNA therapeuticsInfected neurons, astrocytes, microgliaDelivers precision therapeutic cargo directly to CNS cellsInhibits viral replication; modulates neuroinflammatory pathwaysPersonalized EV-based precision medicine platformEarly clinical translation(Tian et al., 2022; Choi et al., 2024; Fatima et al., 2024; Sanadgol et al., 2025)
Viral “Decoy” VesiclesEngineered viral receptors (ACE2, CD4), neutralizing moleculesFree virions in circulationFunctions as nanosponge binding virions before cellular entryReduces viral infectivity and tissue disseminationBroad-spectrum antiviral therapeutic strategyPreclinical(Zhang et al., 2020; Gunnels et al., 2022)
CNS-Derived Blood EVsp-Tau, Aβ42, viral miRNAs, neurofilament proteinsPeripheral circulationTransfers molecular signatures of CNS pathology across BBBReflects ongoing neurodegeneration, neuroinflammation, and viral activityNon-invasive biomarkers for diagnosis, prognosis, and treatment monitoringTranslational/Clinical(Shi et al., 2019; Xu et al., 2023; Oberholster et al., 2024)
Brain Organoid-Derived EVsDevelopmental miRNAs, neurotrophic factors, inflammatory proteins, viral cargoEndothelial cells, immune cells, neural cellsRecapitulates human-specific neuroimmune communication networksModels viral dissemination, BBB dysfunction, neuroinflammatory signalingKey tool for biomarker discovery and therapeutic screening in BOoC systemsPreclinical research(Rossi et al., 2018; Ao et al., 2021; Amartumur et al., 2024; Oberholster et al., 2024)

5 The EV–Neuroimmune axis: a unifying conceptual framework

5.1 Defining the EV–Neuroimmune axis

The EV–neuroimmune axis functions as a systems-level framework wherein viral infection constitutes both a cytopathic assault and a critical disruption of vesicular communication within the brain (Figure 6) (Oberholster et al., 2024). This axis comprises three interlinked bidirectional networks: intra-CNS communication, where neural cells exchange EVs to coordinate antiviral responses and maintain synaptic homeostasis; CNS–BBB communication, where glial-derived EVs modulate endothelial permeability; and CNS–peripheral immune communication, involving brain-derived EVs that mobilize systemic immune cells and peripheral EVs that conversely modulate CNS neuroinflammation (Yachou et al., 2020; Rocamonde et al., 2023). Dysregulation within any segment of this loop is amplified across the others, explaining why focal viral infections often precipitate system-wide neuropathology (Oberholster et al., 2024).

EV–mediated neuroimmune axis in neurotropic viral infection. Neurotropic viruses disrupt EV communication between the CNS, the BBB, and the peripheral immune system, triggering an inflammatory response. Intra-CNS EV signaling (neurons, astrocytes, microglia) regulates homeostasis but promotes microglial activation when dysregulated. At the BBB, EVs increase permeability and leukocyte recruitment via endothelial activation. In the periphery, brain-derived EVs activate immune cells and amplify CNS inflammation, leading to neuroinflammation, cognitive impairment, and neurodegeneration-like outcomes.
Figure 6. EV–mediated neuroimmune axis in neurotropic viral infection. Neurotropic viruses disrupt EV communication between the CNS, the BBB, and the peripheral immune system, triggering an inflammatory response. Intra-CNS EV signaling (neurons, astrocytes, microglia) regulates homeostasis but promotes microglial activation when dysregulated. At the BBB, EVs increase permeability and leukocyte recruitment via endothelial activation. In the periphery, brain-derived EVs activate immune cells and amplify CNS inflammation, leading to neuroinflammation, cognitive impairment, and neurodegeneration-like outcomes.

5.2 EVs as communication currency

Unlike soluble cytokines, which generally transmit a limited set of alarm signals, EVs can transfer coordinated cargoes of proteins, lipids, and regulatory RNAs that reflect the state of the donor cell (Pitt et al., 2016; Oberholster et al., 2024). During viral encephalitis, this homeostatic communication system may be redirected by vesicles carrying viral material or inflammatory regulators, thereby altering recipient-cell behavior (Horn and MacLean, 2021). Such vesicular transfer provides a plausible mechanism by which local infection could influence distant neural and vascular compartments, although the relative contribution of EVs must be established experimentally for each virus and model.

5.3 Systems-level implications for model design

The EV–neuroimmune axis necessitates a shift in the design of viral encephalitis models. Capturing these dynamics requires real-time monitoring of EV production, cargo composition, and intercompartmental trafficking within Brain-on-a-Chip platforms (Pediaditakis et al., 2022; Amartumur et al., 2024). Furthermore, addressing the complexity of this network requires advanced analytical frameworks that integrate electrophysiological, transcriptomic, and proteomic data (Amartumur et al., 2024). Implementing AI-driven multi-omics pipelines will be essential to characterize these EV-mediated neuroimmune interactions and to validate therapeutic interventions during viral encephalitis (Oberholster et al., 2024; Amartumur et al., 2024).

6 Integrating EV biology into brain organoid-on-a-chip systems

6.1 Engineering approaches for EV tracking

Multiplexed detection strategies are required to resolve nanoscale vesicular events within microfluidic environments (Amartumur et al., 2024). Intercompartmental EV trafficking can be monitored using fluorescent membrane labels or genetically encoded markers such as CD63–GFP, provided that dye aggregates and non-vesicular fluorescence are controlled for. Integrated optofluidic biosensors can support single-particle detection and counting (Jahani et al., 2021), whereas immunoaffinity platforms such as ExoChip enable on-chip capture for downstream protein or miRNA analysis. Surface-plasmon-resonance sensors offer label-free monitoring of EV-associated surface signals (Shao et al., 2018). Together, these methods could provide complementary measurements of EV abundance, movement, and cargo, but no single method is sufficient for EV identification. Figure 7 compares representative detection approaches.

EV Monitoring Technologies in Microfluidic Systems. Schematic illustration of four microfluidic EV detection platforms. EVs are analyzed using (i) CD63-GFP fluorescence labeling with real-time confocal tracking of vesicle movement, (ii) optofluidic PDMS biosensors enabling single-vesicle counting and size profiling via integrated optical waveguides, (iii) ExoChip antibody-coated capture arrays for fluorescence-based detection of EV surface markers (CD63, CD9, CD81), and (iv) surface plasmon resonance (SPR) sensors for label-free, real-time EV cargo profiling. A performance comparison inset summarizes detection range (nm), throughput (vesicles/min), label-free capability, and suitability for live monitoring. Abbreviations: GFP, green fluorescent protein; PDMS, polydimethylsiloxane; SPR, surface plasmon resonance.
Figure 7. EV Monitoring Technologies in Microfluidic Systems. Schematic illustration of four microfluidic EV detection platforms. EVs are analyzed using (i) CD63-GFP fluorescence labeling with real-time confocal tracking of vesicle movement, (ii) optofluidic PDMS biosensors enabling single-vesicle counting and size profiling via integrated optical waveguides, (iii) ExoChip antibody-coated capture arrays for fluorescence-based detection of EV surface markers (CD63, CD9, CD81), and (iv) surface plasmon resonance (SPR) sensors for label-free, real-time EV cargo profiling. A performance comparison inset summarizes detection range (nm), throughput (vesicles/min), label-free capability, and suitability for live monitoring. Abbreviations: GFP, green fluorescent protein; PDMS, polydimethylsiloxane; SPR, surface plasmon resonance.

6.2 EV trafficking and cargo mapping

Cell-type-specific reporters combined with compartmentalized microfluidics can be used to map the origin, destination, and cargo of EVs (Amartumur et al., 2024). Brain organoids release both supernatant-derived extracellular vesicles (SuEVs) and matrix-bound nanovesicles (MBNVs) (Liu et al., 2025). In the cited organoid study, these fractions differed in lipid and RNA composition, illustrating that sampling location can influence the biological signals recovered. In a future viral encephalitis BOoC, mass spectrometry, small-RNA sequencing, or NanoString-based profiling could test whether HSV-1- or WNV-associated cargo signatures emerge across sequential vascular, BBB, and neural compartments. This is a proposed experimental capability; virus-specific signatures have not yet been established in such an integrated sequential microfluidic system.

6.3 Real-time neuroimmune communication analysis

One major advantage of BOoC integration is continuous multiparametric monitoring. Using multiplexed bead-based immunoassays, cytokines and EV secretion dynamics may be measured concomitantly from identical microfluidic effluents (Amartumur et al., 2024). MEAs integrated with the organoid floor record spontaneous local field potentials and action-potential bursts, enabling the detection of network hyperexcitability within hours of viral challenge (Amartumur et al., 2024). Concurrently, TEER sensors provide continuous, sub-ohm-resolution measurements of BBB integrity, capturing barrier degradation kinetics in real time (Choi et al., 2023). The temporal sequence can thus be reconstructed from a single experimental run, allowing researchers to separate initiating and sustaining events in neuroinflammation.

6.4 Advanced modeling and personalized applications

Advanced in vitro models, including organoid-on-chip systems, are being developed to study neurovascular pathology and host-specific responses to viral infection (Amartumur et al., 2024). Patient-derived iPSC organoids may preserve genetic variants that influence viral susceptibility, inflammatory signaling, or treatment response (Rossi et al., 2018). Within these personalized models, EV cargo profiling is an emerging approach for identifying molecular correlates of disease severity and treatment response. At present, however, patient-specific EV signatures require validation in sufficiently large, clinically annotated cohorts before they can support diagnostic or therapeutic decisions (Amartumur et al., 2024).

7 Artificial intelligence and multi-omics integration

7.1 Single-cell and spatial transcriptomics

Single-cell and spatial transcriptomic methods can resolve cell-type-specific responses and regional heterogeneity within infected organoids. In HSV-1 cerebral organoids, Rybak-Wolf et al. (2023) used scRNA-seq to show infection-associated shifts in cellular states and persistent activation of TNF-related inflammatory programs after acyclovir treatment. TNF signaling is therefore described here as a pathway-level response rather than a cell-surface marker. Spatial transcriptomics in the same study localized infection-associated changes in synaptic and stress-response programs within tissue architecture. Together, these approaches support infection-state atlases that link viral distribution, cell identity, inflammatory signaling, and neural dysfunction (Rybak-Wolf et al., 2023).

7.2 EV-omics: proteomics, lipidomics, and transcriptomics

Proteomic analysis of CNS-derived EV preparations has identified canonical EV proteins together with cargo associated with neuronal and glial origins (Gallart-Palau et al., 2016). EVs released from forebrain- and hindbrain-patterned organoids share a core proteome but also contain region-associated subsets that may influence recipient cells (Liu et al., 2025). The same work reported distinct lipid enrichment in matrix-bound nanovesicles (MBNVs), including glycerophospholipids and sphingolipids associated with membrane organization and signaling. These findings support multimodal EV profiling, but they do not yet establish a validated clinical biomarker panel for viral encephalitis (Liu et al., 2025; Oberholster et al., 2024).

7.3 AI-driven analytics and patient stratification

Machine-learning methods are increasingly being explored for integrating organoid imaging, electrophysiology, and multi-omics data. Current studies support tasks such as morphological quality control, feature extraction, disease-state classification, and hypothesis generation, but performance remains dependent on sample size, data quality, and external validation (Partiot et al., 2024; Wang et al., 2024a). Multi-omics integration may help identify molecular subgroups and candidate treatment-response signatures (Zheng et al., 2022; Liu et al., 2023). Figure 8 therefore presents a proposed analytical pipeline rather than a clinically validated prediction system.

Proposed multi-omics and AI pipeline for an EV-integrated BOoC. Candidate inputs include scRNA-seq, spatial transcriptomics, EV proteomics, lipidomics, miRNA profiles, MEA recordings, TEER measurements, and cytokine data. Statistical integration and machine-learning methods could be used for feature selection, disease-state classification, biomarker prioritization, and treatment-response modeling. Patient stratification and digital-twin outputs are prospective goals that require large, harmonized datasets and independent clinical validation; they are not established capabilities of current viral encephalitis BOoC systems.
Figure 8. Proposed multi-omics and AI pipeline for an EV-integrated BOoC. Candidate inputs include scRNA-seq, spatial transcriptomics, EV proteomics, lipidomics, miRNA profiles, MEA recordings, TEER measurements, and cytokine data. Statistical integration and machine-learning methods could be used for feature selection, disease-state classification, biomarker prioritization, and treatment-response modeling. Patient stratification and digital-twin outputs are prospective goals that require large, harmonized datasets and independent clinical validation; they are not established capabilities of current viral encephalitis BOoC systems.

7.4 Prospective digital twins of viral encephalitis

A digital twin for viral encephalitis would be a computational representation iteratively updated with data from a patient-derived BOoC, EV multi-omics, and clinical observations. Such a model could, in principle, simulate disease trajectories and rank candidate interventions. No validated digital twin of viral encephalitis currently exists, and the concept should therefore be regarded as a research objective rather than an operational platform.

A prospective workflow would be:

Patient iPSCs → personalized BOoC → EV and functional profiling → computational model → candidate treatment ranking → experimental validation → clinical outcome feedback.

In this future framework, chip experiments and treatment outcomes could be used to update model parameters and improve predictions over successive iterations. Before such a closed-loop system could support clinical translation, it would require standardized BOoC manufacturing, reproducible EV measurements, transparent model development, independent external validation, and prospective comparison with patient outcomes (Zheng et al., 2022; Wang et al., 2024b). The proposed system is therefore best viewed as a hypothesis-generating preclinical framework, not as a current surrogate for clinical decision-making.

8 EV-based precision therapeutics and drug screening

8.1 Natural EVs as therapeutic vehicles

EVs may offer advantages over some synthetic nanoparticle carriers for CNS delivery, including biocompatible membranes, opportunities for cell-derived targeting, and the capacity to protect molecular cargo (Amartumur et al., 2024; Oberholster et al., 2024). However, BBB crossing is not universal across EV populations and depends on source, dose, route, surface composition, and disease state. Unlike many liposomes, EVs can display endogenous membrane proteins that influence uptake, but their heterogeneity complicates manufacturing and pharmacokinetic control. Proposed benefits such as reduced immunogenicity or neural-tissue tropism therefore require product-specific validation. Major limitations include variable yields, heterogeneous cargo, co-isolated contaminants, and uncertain in vivo biodistribution (György et al., 2014; Liu et al., 2025).

8.2 Engineered EVs: precision cargo delivery

Bioengineering approaches allow for the conversion of naturally heterogeneous, cargo-laden EVs into molecularly programmed therapeutics. Loading methods include electroporation, sonication, and co-transfection, as well as the surface display of targeting ligands via fusion to exosomal membrane proteins (Armstrong and Stevens, 2018; Choi et al., 2024). Engineered and functional EV designs in CNS models include anti-NF-κB siRNA-carrying EVs that decrease glial activation and bystander neuronal damage (Rong et al., 2022; Li et al., 2025b); miRNA-219 delivery to promote remyelination (Pusic and Kraig, 2013; Osorio-Querejeta et al., 2020); and ACE2-or CD4-decorated nanodecoy receptors that act as nanosponges to absorb free SARS-CoV-2 or HIV virions before cellular entry (El-Shennawy et al., 2022; Kim et al., 2022). Furthermore, encapsulating conventional antivirals into EVs reduces off-target toxicities and enhances CNS delivery compared to free-drug administration (Zhou et al., 2025).

8.3 Therapeutic screening in EV-integrated chips

An EV-integrated BOoC could support simultaneous assessment of antiviral activity, inflammatory signaling, barrier integrity, and neural function within a human tissue context (Amartumur et al., 2024). Candidate drugs or engineered EVs could be introduced into the vascular compartment while viral replication, EV profiles, cytokines, TEER, and MEA activity are monitored (Maisonneuve et al., 2022; Deng et al., 2023). Post-treatment EV cargo analysis might provide a molecular readout that complements viability and viral-load endpoints. This proposed workflow requires validation against conventional assays and in vivo or clinical outcomes before it can be used to prioritize therapies (Staufer et al., 2021; Machhi et al., 2021).

8.4 Precision medicine and personalized optimization

Patient-derived iPSC-BOoC models could enable comparison to therapeutic responses across defined genetic backgrounds (Zheng et al., 2022). In principle, patient-specific chips may help investigate how host variation influences viral replication, inflammatory signaling, dose response, and EV cargo. They may also support exploratory biomarker development for longitudinal blood- or CSF-based monitoring. These applications remain prospective and will require reproducibility studies, sufficiently powered patient cohorts, and concordance with clinical outcomes (Oberholster et al., 2024).

8.5 Closed-loop therapeutic platforms

A long-term research goal is to couple BOoC sensing with AI-assisted decision support in a closed-loop experimental system:

Drug Administration → EV / Biology Response Monitoring → AI Analysis → Treatment Optimization

In such a future system, real-time EV, cytokine, TEER, and MEA data could be analyzed computationally to suggest the next experimental dose or formulation. Repeated cycles might improve preclinical optimization, but autonomous treatment selection is not currently established and would require strict human oversight, transparent algorithms, predefined safety constraints, and prospective validation (Zheng et al., 2022).

9 Translational challenges and future opportunities

Important biological, technical, and regulatory barriers must be addressed before EV-integrated BOoC platforms can approach clinical use. iPSC-derived organoids often resemble fetal rather than adult brain tissue, limiting the modeling of age-dependent viral susceptibility and mature circuit dysfunction (Zheng et al., 2022). Batch-to-batch and line-to-line variability can undermine reproducibility, whereas incomplete vascular and immune compartments restrict the study of long-term host responses (Garreta et al., 2020; Huang et al., 2021). These limitations are particularly important for viral encephalitis, in which age, immune status, endothelial function, and peripheral immune recruitment strongly influence disease severity.

EV analysis introduces additional methodological challenges, including low yields, co-purification of non-vesicular material, and functional heterogeneity that is obscured by bulk measurements (Ramirez et al., 2017). Single-vesicle multi-omics and high-resolution imaging are advancing but are not yet scalable for routine clinical throughput. Translation will also require GMP-compatible iPSC and EV manufacturing, harmonized infection and sampling protocols, biosafety controls, multicenter reproducibility studies, and prospective validation that chip-based pharmacological predictions correspond to human outcomes.

Future opportunities include non-destructive biosensors for continuous monitoring of neuronal activity, viral spread, and EV release; fluidic coupling of brain chips to peripheral organ models; and multicenter data infrastructures for model validation (Lambert et al., 2025; Amartumur et al., 2024). AI-guided control may eventually support organoid culture and experimental adaptation, but claims of autonomous maturation or clinical prediction should remain prospective until independently demonstrated (Deng et al., 2023; Zheng et al., 2022). Clinical translation will depend on standardized EV characterization, matched analysis of patient CSF and blood, and transparent validation across diverse cohorts (Kumar et al., 2024).

10 Conclusion

Viral encephalitis is best understood as the combined consequence of direct viral injury and dysregulated neuroimmune communication. EVs can carry viral components, inflammatory mediators, and regulatory molecules among neural, glial, endothelial, and peripheral immune cells, thereby linking acute infection to barrier dysfunction and persistent neurological sequelae. This review uses the EV–neuroimmune axis as an organizing framework and identifies the integration of this axis with BBB-competent BOoC models, EV-omics, and computational analysis as the principal conceptual advance. Existing evidence supports many individual components, but the complete EV-integrated BOoC, digital twin, and closed-loop therapeutic systems remain future objectives. Progress will require standardized EV characterization, reproducible organoid and vascular modules, clinically relevant immune integration, scalable manufacturing, multicenter datasets, and prospective validation. Meeting these requirements could establish human-relevant platforms for mechanistic investigation, biomarker discovery, and the development of carefully calibrated precision therapies for viral encephalitis and related neuroinflammatory and/or autoimmune conditions.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. During manuscript preparation, AI-assisted tools were used for language clarity, structural editing, and preliminary conceptual layouts of some schematic figures. The revised graphical abstract and Figure 3 were redrawn using Python-based drawing tools. The author manually verified and corrected all scientific content, labels, citations, and final figure compositions and takes full responsibility for the manuscript. No primary data were generated, analyzed, interpreted, or altered by AI.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Duc-Huy Tran Nguyen, University of Wisconsin-Madison, United States

Reviewed by: Joydeb Majumder, University of Wisconsin-Madison, United States

Sara Micheli, University of Padua, Italy

Statements

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Funding

The author(s) declared that financial support was not received for this work and/or its publication.

Author contributions

MA: Supervision, Project administration, Validation, Data curation, Formal Analysis, Methodology, Visualization, Funding acquisition, Software, Investigation, Conceptualization, Writing – original draft, Resources, Writing – review and editing.

Acknowledgements

The author thanks colleagues and institutional facilities that provided scholarly feedback and access to academic resources during manuscript preparation.

References

  1. Amartumur S., Nguyen H. M., Huynh T., Kim T. S., Woo R., Oh E. (2024). Neuropathogenesis-on-chips for neurodegenerative diseases. Nat. Commun. 15, 2219. doi:10.1038/s41467-024-46554-8
  2. Ao Z., Cai H., Wu Z., Song S., Karahan H., Kim B.-W. (2021). Tubular human brain organoids to model microglia-mediated neuroinflammation. Lab a Chip 21, 2751 – 2762. doi:10.1039/d1lc00030f
  3. Armstrong J. P. K., Stevens M. M. (2018). Strategic design of extracellular vesicle drug delivery systems. Adv. Drug Deliv. Rev. 130, 12 – 16. doi:10.1016/j.addr.2018.06.017
  4. Azim M. M. N., Bhajan S. K., Park J. H., Abass K. S., Rahman A., Choi M. (2026). Stem cell-derived organoids for cancer therapy: precision medicine and drug selection. Int. J. Mol. Sci. 27 2954. doi:10.3390/ijms27072954
  5. Boghdeh-Olson N. A., Barrera M. D., Britt C. M., Schaffer D. K., Brown J. A., Wikswo J. P. (2026). Application of a blood–brain barrier Organ-on-a-Chip model for assessment of countermeasure efficiency against eastern equine encephalitis virus. Viruses 18, 548. doi:10.3390/v18050548
  6. Bradshaw M., Venkatesan A. (2016). Herpes simplex Virus-1 encephalitis in adults: pathophysiology, diagnosis, and management. Neurotherapeutics 13, 493 – 508. doi:10.1007/s13311-016-0433-7
  7. Caldwell M., Boruah A. P., Thakur K. T. (2022). Acute neurologic emerging flaviviruses. Ther. Adv. Infect. Dis. 9, 20499361221102664. doi:10.1177/20499361221102664
  8. Chaudhuri A., Kennedy P. G. E. (2002). Diagnosis and treatment of viral encephalitis. Postgrad. Med. J. 78, 575 – 583. doi:10.1136/pmj.78.924.575
  9. Chen Z., Zhong D., Li G. (2019). The role of microglia in viral encephalitis: a review. J. Neuroinflammation 16, 76. doi:10.1186/s12974-019-1443-2
  10. Choi J., Choi H. K., Lee K. (2023). In situ detection of neuroinflammation using multicellular 3D Neurovascular‐Unit‐on‐a‐Chip. Adv. Funct. Mater. 33, 2304382. doi:10.1002/adfm.202304382
  11. Choi H. K., Chen M., Goldston L. L., Lee K. (2024). Extracellular vesicles as nanotheranostic platforms for targeted neurological disorder interventions. Nano Converg. 11, 19. doi:10.1186/s40580-024-00426-5
  12. Davis B., Rall G. F., Schnell M. J. (2015). Everything you always wanted to know about rabies virus (but were afraid to ask). Annu. Rev. Virology 2, 451 – 471. doi:10.1146/annurev-virology-100114-055157
  13. Deffner F., Scharr M., Klingenstein S., Klingenstein M., Milazzo A., Scherer S. (2020). Histological evidence for the enteric nervous system and the choroid plexus as alternative routes of neuroinvasion by SARS-CoV2. Front. Neuroanat. 14, 596439. doi:10.3389/fnana.2020.596439
  14. Deng S., Li C., Cao J., Cui Z., Du J., Fu Z. (2023). Organ-on-a-chip meets artificial intelligence in drug evaluation. Theranostics 13, 4526 – 4558. doi:10.7150/thno.87266
  15. El-Shennawy L., Hoffmann A. D., Dashzeveg N. K., McAndrews K. M., Mehl P. J., Cornish D. (2022). Circulating ACE2-expressing extracellular vesicles block broad strains of SARS-CoV-2. Nat. Commun. 13, 405. doi:10.1038/s41467-021-27893-2
  16. Fatima S., Qaiser A., Andleeb S., Hashmi A. H., Manzoor S. (2024). Navigating the brain: the role of exosomal shuttles in precision therapeutics. Front. Neurology 14, 1324216. doi:10.3389/fneur.2023.1324216
  17. Fisher C. R., Streicker D. G., Schnell M. J. (2018). The spread and evolution of rabies virus: conquering new frontiers. Nat. Rev. Microbiol. 16, 241 – 255. doi:10.1038/nrmicro.2018.11
  18. Furr S. R., Marriott I. (2012). Viral CNS infections: role of glial pattern recognition receptors in neuroinflammation. Front. Microbiol. 3, 201. doi:10.3389/fmicb.2012.00201
  19. Gallart‐Palau X., Serra A., Sze S. K. (2016). Enrichment of extracellular vesicles from tissues of the central nervous system by PROSPR. Mol. Neurodegener. 11, 41. doi:10.1186/s13024-016-0108-1
  20. Garreta E., Kamm R. D., Lopes S. M. C. de S., Lancaster M. A., Weiss R., Trepat X. (2020). Rethinking organoid technology through bioengineering. Nat. Mater. 20, 145 – 155. doi:10.1038/s41563-020-00804-4
  21. Granerød J., Huang Y., Davies N., de Sequeira P. C., Mwapasa V., Rupali P. (2023). Global landscape of encephalitis: key priorities to reduce future disease burden. Clin. Infect. Dis. 77, 1552 – 1560. doi:10.1093/cid/ciad417
  22. Gunnels T. F., Stranford D. M., Mitrut R. E., Kamat N. P., Leonard J. N. (2022). Elucidating design principles for engineering cell‐derived vesicles to inhibit SARS‐CoV‐2 infection. Small 18, 2200125. doi:10.1002/smll.202200125
  23. György B., Fitzpatrick Z., Crommentuijn M. H. W., Mu D., Maguire C. A. (2014). Naturally enveloped AAV vectors for shielding neutralizing antibodies and robust gene delivery in vivo. Biomaterials 35, 7598 – 7609. doi:10.1016/j.biomaterials.2014.05.032
  24. Horn M. D., MacLean A. G. (2021). Extracellular vesicles as a means of viral immune evasion, CNS invasion, and glia-induced neurodegeneration. Front. Cell. Neurosci. 15, 695899. doi:10.3389/fncel.2021.695899
  25. Huang Y., Huang Z., Tang Z., Chen Y., Huang M., Liu H. (2021). Research progress, challenges, and breakthroughs of organoids as disease models. Front. Cell Dev. Biol. 9, 740574. doi:10.3389/fcell.2021.740574
  26. Jahani Y., Arvelo E. R., Yesilköy F., Koshelev K., Cianciaruso C., Palma M. D. (2021). Imaging-based spectrometer-less optofluidic biosensors based on dielectric metasurfaces for detecting extracellular vesicles. Nat. Commun. 12, 3246. doi:10.1038/s41467-021-23257-y
  27. Kim H. K., Cho J., Kim E., Kim J., Yang J., Kim K. (2022). Engineered small extracellular vesicles displaying ACE2 variants on the surface protect against SARS‐CoV‐2 infection. J. Extracell. Vesicles 11, e12179. doi:10.1002/jev2.12179
  28. Kiyani M., Liu B., Charalambous L. T., Adil S. M., Hodges S. E., Yang S. (2020). The longitudinal health economic impact of viral encephalitis in the United States. J. Med. Microbiol. 69, 270 – 279. doi:10.1099/jmm.0.001153
  29. Kumar M. A., Baba S. K., Sadida H. Q., Marzooqi S. A., Jerobin J., Altemani F. H. (2024). Extracellular vesicles as tools and targets in therapy for diseases. Signal Transduct. Target. Ther. 9, 27. doi:10.1038/s41392-024-01735-1
  30. La Rosa G., Pascale E., Iazzetta M. R., Sozzi E., Fico A., Parrotta E. I. (2026). Brain organoids as models of extracellular vesicle-mediated human neural communication. Neural Regen. Res. 21, 4529 – 4537. doi:10.4103/NRR.NRR-D-25-01806
  31. Lambert G. G., Crespo E. L., Murphy J. W., Turner K. L., Gershowitz E., Cunningham M. J. (2025). CaBLAM: a high-contrast bioluminescent Ca2+ indicator derived from an engineered Oplophorus gracilirostris luciferase. Nat. Methods 23, 205 – 215. doi:10.1038/s41592-025-02972-0
  32. Li P., Sun S., Zhu X., Liu X., Yin R., Chen Y. (2025a). Intranasal delivery of engineered extracellular vesicles promotes neurofunctional recovery in traumatic brain injury. J. Nanobiotechnology 23, 229. doi:10.1186/s12951-025-03181-9
  33. Li S., Zhang J., Sun L., Yang Z., Liu X., Liu J. (2025b). Mesenchymal stem cell-derived extracellular vesicles: current advances in preparation and therapeutic applications for neurological disorders. Front. Cell Dev. Biol. 13, 1626996. doi:10.3389/fcell.2025.1626996
  34. Liu X., Hasan M. R., Ahmed K. A., Hossain M. Z. (2023). Machine learning to analyse omic-data for COVID-19 diagnosis and prognosis. BMC Bioinforma. 24, 7. doi:10.1186/s12859-022-05127-6
  35. Liu C., Chen X., Ene J., Esmonde C., Kanekiyo T., Zeng C. (2025). Engineering extracellular vesicles secreted by human brain organoids with different regional identity. ACS Appl. Mater. and Interfaces 17, 15145 – 15162. doi:10.1021/acsami.4c22692
  36. Machhi J., Shahjin F., Das S., Patel M., Abdelmoaty M. M., Cohen J. D. (2021). A role for extracellular vesicles in SARS-CoV-2 therapeutics and prevention. J. Neuroimmune Pharmacol. 16, 270 – 288. doi:10.1007/s11481-020-09981-0
  37. Maisonneuve B. G. C., Libralesso L., Miny L., Batut A., Rontard J., Gleyzes M. (2022). Deposition chamber technology as building blocks for a standardized brain-on-chip framework. Microsystems and Nanoeng. 8, 86. doi:10.1038/s41378-022-00406-x
  38. McGrath N., Anderson N. E., Croxson M. C., Powell K. F. (1997). Herpes simplex encephalitis treated with acyclovir: diagnosis and long term outcome. J. Neurology Neurosurg. and Psychiatry 63, 321 – 326. doi:10.1136/jnnp.63.3.321
  39. McQuaid C., Brady M., Deane R. (2021). SARS-CoV-2: is there neuroinvasion? Fluids Barriers CNS 18, 32. doi:10.1186/s12987-021-00267-y
  40. Misra U. K. (2020). Clinical and experimental studies in Japanese encephalitis: lessons learnt. Ann. Natl. Acad. Med. Sci. (India) 53, 202 – 209. doi:10.1055/s-0040-1712808
  41. NIAID Collaborative Antiviral Study Group West Nile Virus 210 Protocol Team Hart J., Tillman G. D., Kraut M. A., Chiang H., Strain J. F. (2014). West nile virus neuroinvasive disease: neurological manifestations and prospective longitudinal outcomes. BMC Infect. Dis. 14, 248. doi:10.1186/1471-2334-14-248
  42. Oberholster L., Pasquier R. D., Mathias A. (2024). Exploring the role of brain-derived extracellular vesicles in viral infections: from pathological insights to biomarker potential. Front. Cell. Infect. Microbiol. 14, 1423394. doi:10.3389/fcimb.2024.1423394
  43. Osorio-Querejeta I., Carregal‐Romero S., Ayerdi-Izquierdo A., Mäger I., Nash L. A., Wood M. J. A. (2020). MiR-219a-5p enriched extracellular vesicles induce OPC differentiation and EAE improvement more efficiently than liposomes and polymeric nanoparticles. Pharmaceutics 12, 186. doi:10.3390/pharmaceutics12020186
  44. Partiot E., Gorda B., Lutz W., Lebrun S., Khalfi P., Mora S. (2024). Organotypic culture of human brain explants as a preclinical model for AI-driven antiviral studies. EMBO Mol. Med. 16, 1004 – 1026. doi:10.1038/s44321-024-00039-9
  45. Pavlou A., Mulenge F., Gern O. L., Busker L. M., Greimel E., Waltl I. (2024). Orchestration of antiviral responses within the infected central nervous system. Cell. Mol. Immunol. 21, 943 – 958. doi:10.1038/s41423-024-01181-7
  46. Pediaditakis I., Kodella K., Manatakis D. V., Le C. Y., Barthakur S., Sorets A. G. (2022). A microengineered brain-chip to model neuroinflammation in humans. iScience 25, 104813. doi:10.1016/j.isci.2022.104813
  47. Petersen L. R., Brault A. C., Nasci R. S. (2013). West nile virus: review of the literature. JAMA 310, 308. doi:10.1001/jama.2013.8042
  48. Piret J., Boivin G. (2020). Immunomodulatory strategies in Herpes simplex virus encephalitis. Clin. Microbiol. Rev. 33. doi:10.1128/cmr.00105-19
  49. Pitt J. M., Kroemer G., Zitvogel L. (2016). Extracellular vesicles: masters of intercellular communication and potential clinical interventions. J. Clin. Investigation 126, 1139 – 1143. doi:10.1172/jci87316
  50. Pusic A. D., Kraig R. P. (2013). Youth and environmental enrichment generate serum exosomes containing miR‐219 that promote CNS myelination. Glia 62, 284 – 299. doi:10.1002/glia.22606
  51. Ramirez M. I., de Amorim M. G., Gadelha C., Milic I., Welsh J. A., Freitas V. M. (2017). Technical challenges of working with extracellular vesicles. Nanoscale 10, 881 – 906. doi:10.1039/c7nr08360b
  52. Rocamonde B., Hasan U., Mathieu C., Dutartre H. (2023). Viral-induced neuroinflammation: different mechanisms converging to similar exacerbated glial responses. Front. Neurosci. 17, 1108212. doi:10.3389/fnins.2023.1108212
  53. Rong Y., Wang Z., Tang P., Wang J., Ji C., Chang J. (2022). Engineered extracellular vesicles for delivery of siRNA promoting targeted repair of traumatic spinal cord injury. Bioact. Mater. 23, 328 – 342. doi:10.1016/j.bioactmat.2022.11.011
  54. Rossi G., Manfrin A., Lütolf M. P. (2018). Progress and potential in organoid research. Nat. Rev. Genet. 19, 671 – 687. doi:10.1038/s41576-018-0051-9
  55. Rybak‐Wolf A., Wyler E., Pentimalli T. M., Legnini I., Oliveras A., Glažar P. (2023). Modelling viral encephalitis caused by herpes simplex virus 1 infection in cerebral organoids. Nat. Microbiol. 8, 1252 – 1266. doi:10.1038/s41564-023-01405-y
  56. Salimi H., Cain M. D., Klein R. S. (2016). Encephalitic arboviruses: emergence, clinical presentation, and neuropathogenesis. Neurotherapeutics 13, 514 – 534. doi:10.1007/s13311-016-0443-5
  57. Sanadgol N., Abedi M., Id D. A., Kamran Z., Khalseh R., Beyer C. (2025). Exosomes as nanocarriers for brain-targeted delivery of therapeutic nucleic acids: advances and challenges. J. Nanobiotechnology 23, 453. doi:10.1186/s12951-025-03528-2
  58. Shao H., Im H., Castro C. M., Breakefield X. O., Weissleder R., Lee H. (2018). New technologies for analysis of extracellular vesicles. Chem. Rev. 118, 1917 – 1950. doi:10.1021/acs.chemrev.7b00534
  59. Shi M., Sheng L., Stewart T., Zabetian C. P., Zhang J. (2019). New windows into the brain: central nervous system-derived extracellular vesicles in blood. Prog. Neurobiol. 175, 96 – 106. doi:10.1016/j.pneurobio.2019.01.005
  60. Staufer O., Dietrich F., Rimal R., Schröter M., Fabritz S., Boehm H. (2021). Bottom-up assembly of biomedical relevant fully synthetic extracellular vesicles. Sci. Adv. 7, eabg6666. doi:10.1126/sciadv.abg6666
  61. Tavčar P., Potokar M., Kolenc M., Korva M., Avšič‐Županc T., Zorec R. (2021). Neurotropic viruses, astrocytes, and COVID-19. Front. Cell. Neurosci. 15, 662578. doi:10.3389/fncel.2021.662578
  62. Thompson D., Brissette C. A., Watt J. A. (2022). The choroid plexus and its role in the pathogenesis of neurological infections. Fluids Barriers CNS 19, 75. doi:10.1186/s12987-022-00372-6
  63. Tian T., Qiao S., Tannous B. A. (2022). Nanotechnology-inspired extracellular vesicles theranostics for diagnosis and therapy of central nervous system diseases. ACS Appl. Mater. and Interfaces 15, 182 – 199. doi:10.1021/acsami.2c07981
  64. Wang Z., Zhang Y., Li Z., Wang H., Li N., Deng Y. (2023). Microfluidic Brain‐on‐a‐Chip: from key technology to system integration and application. Small 19, 2304427. doi:10.1002/smll.202304427
  65. Wang H., Li X., You X., Zhao G. (2024a). Harnessing the power of artificial intelligence for human living organoid research. Bioact. Mater. 42, 140 – 164. doi:10.1016/j.bioactmat.2024.08.027
  66. Wang H., Ning X., Zhao F., Zhao H., Li D. (2024b). Human organoids-on-chips for biomedical research and applications. Theranostics 14, 788 – 818. doi:10.7150/thno.90492
  67. Xu X., Iqbal Z., Xu L., Wen C., Duan L., Xia J. (2023). Brain‐derived extracellular vesicles: potential diagnostic biomarkers for central nervous system diseases. Psychiatry Clin. Neurosci. 78, 83 – 96. doi:10.1111/pcn.13610
  68. Yachou Y., Idrissi A. E., Belapasov V., Benali S. A. (2020). Neuroinvasion, neurotropic, and neuroinflammatory events of SARS-CoV-2: understanding the neurological manifestations in COVID-19 patients. Neurol. Sci. 41, 2657 – 2669. doi:10.1007/s10072-020-04575-3
  69. Yang A. C., Kern F., Losada P. M., Agam M., Maat C. A., Schmartz G. P. (2021). Dysregulation of brain and choroid plexus cell types in severe COVID-19. Nature 595, 565 – 571. doi:10.1038/s41586-021-03710-0
  70. Yang J., Feng Z., McComsey G. A., Avery A., Hileman C., Xiong W.-C. (2026). Role of extracellular vesicles derived from HIV-Infected T cells and microglial cells in neuroinflammation and neuropathogenesis using a blood–brain barrier and cerebral organoid co-culture model. J. NeuroVirology. doi:10.1007/s13365-025-01297-0
  71. Zhang Q., Honko A. N., Zhou J., Gong H., Downs S. N., Vasquez J. H. (2020). Cellular nanosponges inhibit SARS-CoV-2 infectivity. Nano Lett. 20, 5570 – 5574. doi:10.1021/acs.nanolett.0c02278
  72. Zheng H., Feng Y., Tang J., Ma S. (2022). Interfacing brain organoids with precision medicine and machine learning. Cell Rep. Phys. Sci. 3, 100974. doi:10.1016/j.xcrp.2022.100974
  73. Zhou L., Godse S., Sinha N., Ma D., Mirzahosseini G., Salman M. (2025). Extracellular vesicle-liposome-darunavir formulation for the treatment of HIV neuropathogenesis. ACS Appl. Nano Mater. 8, 6857 – 6876. doi:10.1021/acsanm.4c04637

Citation

Azim, M. M. N. (2026). Extracellular vesicle-integrated brain organoid-on-a-chip systems for neuroimmune communication and precision therapy in viral encephalitis. Frontiers in Lab on a Chip Technologies, Volume 5. https://doi.org/
Received
Accepted
Published
Volume
5 - 2026

Copyright

Copyright © 2026 Azim.

License: https://creativecommons.org/licenses/by/4.0/

*Correspondence: Md. M. N. Azim

Article metrics

Views
0
Downloads
0
Citations
0
Shares
0

Related articles

ReviewMultiscale mechanobiology of circulating tumor cells: microfluidics and atomic force microscopyAyoub Glia, Muhammedin Deliorman, Mohammad A. QasaimehOriginal ResearchLaser-welded elastomeric valves for the rapid fabrication of flow control in microfluidicsC. Bishayee, C. Herrmann, M. Feinaeugle, M. Skolimowski, G. R. B. E. RömerOriginal ResearchAnalytical solutions for electroosmotic core–annular flow of power-law fluidsWai Yuen Leung, Md Mainul Islam, Xiangchun Xuan